Functions of the Nonsense-Mediated mRNA Decay Pathway in Drosophila Development

doi:10.1371/journal.pgen.0020180

Journal List > PLoS Genet > v.2(12); Dec 2006

PLoS Genet. 2006 December; 2(12): e180.

Published online 2006 December 29. Prepublished online 2006 September 6. doi: 10.1371/journal.pgen.0020180.

PMCID: PMC1756896

Copyright : © 2006 Metzstein and Krasnow. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Functions of the Nonsense-Mediated mRNA Decay Pathway in Drosophila Development

Mark M Metzstein^¤ and Mark A Krasnow^*

Howard Hughes Medical Institute and Department of Biochemistry, Stanford University School of Medicine, Stanford, California, United States of America

R. Scott Hawley, Editor

Stowers Institute for Medical Research, United States of America

* To whom correspondence should be addressed. E-mail: krasnow/at/cmgm.stanford.edu

Received July 14, 2006; Accepted September 6, 2006.

This article has been cited by other articles in PMC.

Abstract

Nonsense-mediated mRNA decay (NMD) is a cellular surveillance mechanism that degrades transcripts containing premature translation termination codons, and it also influences expression of certain wild-type transcripts. Although the biochemical mechanisms of NMD have been studied intensively, its developmental functions and importance are less clear. Here, we describe the isolation and characterization of Drosophila “photoshop” mutations, which increase expression of green fluorescent protein and other transgenes. Mapping and molecular analyses show that photoshop mutations are loss-of-function mutations in the Drosophila homologs of NMD genes Upf1, Upf2, and Smg1. We find that Upf1 and Upf2 are broadly active during development, and they are required for NMD as well as for proper expression of dozens of wild-type genes during development and for larval viability. Genetic mosaic analysis shows that Upf1 and Upf2 are required for growth and/or survival of imaginal cell clones, but this defect can be overcome if surrounding wild-type cells are eliminated. By contrast, we find that the PI3K-related kinase Smg1 potentiates but is not required for NMD or for viability, implying that the Upf1 phosphorylation cycle that is required for mammalian and Caenorhabditis elegans NMD has a more limited role during Drosophila development. Finally, we show that the SV40 3′ UTR, present in many Drosophila transgenes, targets the transgenes for regulation by the NMD pathway. The results establish that the Drosophila NMD pathway is broadly active and essential for development, and one critical function of the pathway is to endow proliferating imaginal cells with a competitive growth advantage that prevents them from being overtaken by other proliferating cells.

Synopsis

Cells possess a variety of surveillance mechanisms that detect and dispose of defective gene products. One such system is the nonsense-mediated mRNA decay (NMD) pathway, which degrades aberrant mRNAs containing nonsense mutations or other premature translation stop signals. In a genetic screen in Drosophila, the authors identified a set of mutations they call “photoshop” mutations because they increase expression of green fluorescent protein transgenes such that cells expressing green fluorescent protein are more easily visualized. They found that the photoshop mutations are mutations in three different genes implicated in NMD. Using these mutations, they show that the NMD pathway not only degrades mutant mRNAs but also influences expression of many transgenes and dozens of endogenous genes during development and is essential for development beyond the larval stage. One important function of the pathway is to provide proliferating cells with a competitive growth advantage that prevents them from being overtaken by other proliferating cells during development. Thus, the Drosophila NMD pathway has critical cellular and developmental roles beyond the classical surveillance function of eliminating mutant transcripts.

Introduction

Nonsense-mediated mRNA decay (NMD) is a cellular surveillance pathway in eukaryotes that recognizes and degrades transcripts with premature termination codons (PTCs). Such transcripts arise as a consequence of genomic mutation, as in numerous human genetic diseases [1,2], and from errors in transcription and aberrant RNA splicing. Destruction of PTC-containing transcripts by NMD prevents production of truncated, potentially harmful proteins that can interfere with normal cellular processes (e.g., [3]). The NMD pathway has also been found to influence expression of a variety of wild-type transcripts (reviewed in [4]), implying that the pathway has regulatory roles beyond its surveillance function. In this paper, we describe Drosophila mutants that affect NMD.

NMD pathway genes were discovered by genetic studies in yeast (up-frameshift suppressor [Upf] genes; [5]) and Caenorhabditis elegans (suppressor with morphogenetic effect on genitalia [smg] genes; [6]), and their functions and mechanisms of action have been characterized by molecular genetic and biochemical analysis of the proteins and target RNAs in yeast [7] and cultured mammalian and Drosophila cells [8–10]. There are three conserved core components of the pathway, Upf1 (smg-2), Upf2 (smg-3), and Upf3 (smg-4) (reviewed in [11]). Upf1 is an RNA helicase that associates with the translation termination complex at PTCs and, at least in yeast, targets the RNA to cytoplasmic RNA processing centers called P bodies [12]. Upf1 is proposed to recruit Upf2 and Upf3 to these termination complexes, which leads to activation of decapping enzymes and nucleases that degrade the target RNA.

Additionally, in metazoans, Upf1 undergoes a phosphorylation cycle (reviewed in [13]). Upf1 is phosphorylated on serine residues by Smg1, a PI3K-related kinase. The phosphates are subsequently removed by complex(es) containing Smg5, Smg6, and/or Smg7, three similar proteins that are thought to recruit the phosphatase PPA2. The Upf1 phosphorylation cycle is apparently necessary for Upf1 and NMD activity at least in some organisms, because NMD function is abrogated when Smg1, Smg5, Smg6, or Smg7 activity is reduced [6,9,10,14].

One intriguing mechanistic question is how the NMD machinery distinguishes a PTC from a normal termination codon. In mammals, an important feature appears to be the relationship between the termination codon and splice junctions in the mRNA [15]. Most normal termination codons are located beyond the last splice junction, in the final exon of the mRNA. Termination codons that lie upstream of an exon–exon boundary are generally recognized as premature and target the mRNA for destruction by NMD. Such boundaries are marked after splicing by deposition of a multiprotein complex, the exon junction complex (EJC), which includes Upf2 and Upf3. One current model proposes that EJCs along an mRNA are normally all displaced by the translocating ribosome, but if the ribosome encounters a termination codon before the last EJC, this EJC remains and promotes delivery of Upf2 and Upf3 to Upf1 at the termination complex to activate NMD [8,12]. However, this cannot be the sole mechanism of PTC recognition because there are mammalian transcripts with stop codons in the last exon that are subject to NMD, as well as transcripts with stop codons upstream of introns that are resistant to NMD (reviewed in [15]). Furthermore, PTC recognition in yeast and cultured Drosophila cells can occur in the absence of introns and splice junctions [7,10]. In yeast it appears that the distance between the stop codon and a special site in or near the 3′ UTR, or the ability of proteins bound at these sites to efficiently associate, marks a termination codon as premature and targets the mRNA for destruction by NMD [16], and this also appears important for NMD targeting of certain mammalian transcripts [17].

Although the mechanism of the NMD pathway has been studied extensively, its developmental functions have received less attention. Yeast and C. elegans NMD pathway mutants are viable as homozygotes and have only subtle or no effects on development and differentiation. The most conspicuous defects in C. elegans mutants are the swollen bursa in the tail of adult males and swollen vulva of hermaphrodites, both of which apparently result from an effect on morphogenesis rather than cell lineage [6]. By contrast, mouse UPF1^−/− mutants are not viable [18], and RNA interference knockdown of Upf1, Smg1, or other NMD pathway genes in cultured Drosophila S2 or mammalian cells causes a cell cycle arrest, implicating the NMD pathway in cell cycle progression [19,20]. However, two recently described mutations in the Drosophila homolog of Smg1, the only extant mutations in Drosophila NMD genes, are homozygous viable and do not appear to affect NMD, raising questions about the function of Smg1 and the NMD pathway in Drosophila [21].

Here, we describe the isolation and characterization of Drosophila mutants that enhance expression of green fluorescent protein (GFP) and other transgenes. We demonstrate that these are loss-of-function mutations in three NMD pathway genes, Upf1, Upf2, and Smg1. We show that Upf1 and Upf2 are required for NMD pathway activity, whereas Smg1 has a variable, gene-selective role potentiating the pathway. We then use the Upf1 and Upf2 mutations to characterize the functions and endogenous RNA targets of the pathway in Drosophila development. We find that the NMD pathway is broadly active during development and required for proper expression of dozens of endogenous genes and for larval viability, and that one critical function of the pathway is to enhance the ability of cells to compete with other cells during proliferative growth.

Results

Identification of Mutations That Increase Transgene Expression in Drosophila

We conducted a genetic mosaic screen of ethane methyl sulfonate–induced mutations on the X chromosome for tracheal (respiratory) system mutants. Details of the screen will be described elsewhere. In the screen, we used the S. cerevisiase FLP1 recombinase (FLP)/FLP1 recombinase target (FRT) system [22] to generate homozygous mutant cell clones in otherwise heterozygous animals. Tracheal clones were identified by labeling them with GFP using the MARCM system [23] in which a btl-GAL4 transgene drives tracheal expression of a UAS-GFP transgene but is kept off in heterozygous cells by a btl-GAL80 transgene present on the wild-type X chromosome (Figure 1

A). In a screen of 749 ethane methyl sulfonate–induced X-linked lethal lines, we identified six mutations (13D, 14J, 25G, 26A, 29AA, and 32AP) that caused markedly increased GFP signal in homozygous mutant tracheal clones compared to control wild-type clones examined as third instar (L3) larvae (Figure 1

B). We named this the “photoshop” phenotype because it increased visualization of clones like that achieved by digital enhancement with Photoshop software (Adobe, http://www.adobe.com). The photoshop phenotype was not dependent on the btl-GAL80 repressor in the MARCM system: homozygous photoshop tracheal cell clones in btl-GAL4, UAS-GFP larvae showed a similar enhancement of GFP signal, and viable hemizygous 25G and 32AP larvae and adults (see below) carrying the same transgenes showed increased GFP signal throughout the tracheal system (Figure 1

C and and1D1

D and data not shown). The photoshop phenotype was not specific to GFP, because mutant clones showed similar enhancement of DsRed1 signal (Figure 1

E). All tracheal cells examined showed the photoshop phenotype, and all six mutations gave similar GFP enhancement.

Figure 1

Identification of Photoshop Mutants

All but two of the mutations (14J, 29AA, 13D, and 26A) lead to hemizygous male lethality before L3. One exception was 25G: 25G/Y hemizygous males and 25G homozygous females developed to L3 larvae at approximately normal frequencies and produced a few percent of escaper adults. These adults appeared morphologically normal but had greatly reduced fertility. The other exception was 32AP, which after removal of extraneous linked lethal loci (see Materials and Methods) was found to be hemizygous-male and homozygous-female viable and fertile. We generated a deficiency uncovering 32AP and found that 32AP/Df showed the same GFP enhancement as 32AP/Y (Figure 1

D), implying that 32AP is an amorphic (null) allele.

To study the effect of the mutations on GFP signal in other tissues and other stages of development, we used the viable alleles 25G and 32AP along with da-GAL4 and UAS-CD8:GFP transgenes, which give ubiquitous expression of GFP [24]. Compared to control animals, 25G/Y male and homozygous 25G female larvae and adults had greatly increased GFP in all tissues examined, including epidermis, salivary glands, fat body, and eyes (Figure 1

F and data not shown). 32AP mutants also showed enhanced GFP signal in all tissues examined, although enhancement appeared slightly weaker than in 25G mutants. 32AP also enhanced levels of a nuclear localized DsRed2 construct expressed in the larval epidermis (Figure 1

G). Thus, photoshop mutations increase transgene signal in many, if not all, larval and adult tissues.

To test whether the increased signal seen in the photoshop mutants was due to increased expression of the transgene, we performed quantitative RT-PCR experiments on RNA derived from hemizygous 25G, hemizygous 32AP, or control wild-type male L3 larvae, each also containing da-GAL4 and UAS-CD8:GFP. GFP RNA levels were increased ~5-fold in 25G mutants and ~2.5-fold in 32AP mutants compared to the wild type (Figure 1

H). Hence, the increased GFP signal in photoshop mutants is due at least in part to increased GFP RNA levels.

Photoshop Mutations Are Loss-of-Function Alleles of Three NMD Pathway Genes

We used meiotic recombination to map the lethality and the mosaic tracheal GFP enhancement phenotype of the photoshop mutations. 14J and 29AA mapped between ct and v (Figure 2

A). Complementation analysis (see below) showed that 14J and 29AA and the semi-lethal allele 25G formed a single complementation group. 13D and 26A mapped between v and g (Figure 2

D), and complemented 14J for lethality. 32AP was mapped by its enhancement of GFP expression in hemizygous males, and localized between cv and ct (Figure 2

F). Thus, the six photoshop mutations define at least three loci.

Figure 2

Photoshop Mutations Are Alleles of Upf2, Upf1, and Smg1

We further refined the position of mutation 14J by single nucleotide polymorphism (SNP) mapping and localized it to an ~200-kb interval (Figure 2

A). DNA sequencing of predicted genes in this interval identified changes in the Drosophila homolog of the NMD pathway gene Upf2 in all three alleles of the 14J complementation group. The 14J mutation is a 14-bp deletion approximately halfway through the 1,241-residue coding sequence of Upf2 (Figure 2

B), which causes a frame shift at codon 562 followed by a stop 40 codons later. 29AA is a nonsense mutation at codon 324. Both 14J and 29AA truncate the Upf2 coding sequence and are likely to be null alleles. A transgene comprising the Upf2 gene (Figure 2

B) rescued 14J hemizygous males and homozygous females to viability and fertility; 29AA could not be tested because of a linked lethal mutation. We conclude that 14J and 29AA are loss-of-function mutations in Upf2.

25G hemizygous males and homozygous females survived to L3 and beyond (see above), whereas 25G/14J trans heterozygotes were not viable after L2, suggesting that 25G is a hypomorphic allele of Upf2. The Upf2⁺ transgene rescued 25G hemizygous males and homozygous females to viability and fertility, confirming this assignment. The 25G allele is a curious mutation that alters the natural Upf2 stop codon to an arginine codon (TGA to AGA). The next in-frame termination codon is 45 bp downstream, so 25G encodes a Upf2 protein with a 15-residue C-terminal extension (Figure 2

C). The C-terminus of yeast Upf2p is required for its interaction with Upf1p [25], so the Upf2 extension might interfere with this function. The 25G mutation also creates a consensus 3′ splice site, but RT-PCR experiments on 25G mutant RNA did not detect any novel splice forms involving this site.

The map positions of 13D and 26A were refined using SNP markers, and the lethality associated with each allele localized to an ~400-kb interval that contained the Drosophila homolog of Upf1 (Figure 2

D). Sequencing of the Upf1 gene in 26A revealed a nonsense mutation (CAG to TAG; Q637stop) in the middle of the 1,180-residue coding sequence (Figure 2

E), and sequencing of 13D identified a missense mutation (TGC to TAC) that alters a conserved cysteine (C186Y) in a domain required for interaction between yeast Upf1p and Upf2p [25] (Figure 2

E). We also found that 26A and 13D failed to complement each other for lethality. Thus, 26A and 13D appear to be loss-of-function alleles of Upf1, and we refer to them as Upf1^26A and Upf1^13D.

32AP was mapped to a 2-Mb interval containing 122 predicted genes, one of which is the Drosophila homolog of Smg1 (Figure 2

F). We generated an ~46-kb deficiency that removes ten genes including Smg1, and found that this deficiency failed to complement the photoshop phenotype of 32AP (Figure 1

D). DNA sequencing identified a nonsense mutation (CAG to TAG) in codon 1651 of the 3,218-residue Smg1 coding sequence, which truncates the Smg1 protein before most of the conserved domains, including the kinase domain (Figure 2

G). Thus, 32AP is most likely a null Smg1 allele, consistent with the gene dosage experiments described above (Figure 1

D). We refer to it as Smg1^32AP.

Photoshop Mutations Abolish or Reduce NMD of a Mutant Transcript In Vivo

Upf1, Upf2, and Smg1 are required for NMD in yeast and C. elegans and have been shown to be involved in NMD in cultured Drosophila cells [10], although recent data question the in vivo role of Smg1 in Drosophila [21]. To determine whether photoshop genes are required for NMD in vivo, we tested the effects of photoshop mutations on mRNA levels of Adhⁿ⁴, a nonsense mutation in Adh [26] that subjects the mRNA to NMD in S2 cells [10]. We isolated RNA from adult wild-type males (y w FRT^19A/Y; Adhⁿ⁴/Adh⁺) and from Upf2 (Upf2^25G/Y; Adhⁿ⁴/Adh⁺) and Smg1 (Smg1^32AP/Y; Adhⁿ⁴/Adh⁺) mutant males, amplified Adh mRNA by RT-PCR, and quantitated the Adh⁺ and Adhⁿ⁴ products to assess the relative levels of the two transcripts (Figure 3

). In wild-type animals, steady state transcript levels from the Adhⁿ⁴ allele were reduced 13-fold relative to Adh⁺ transcripts, presumably because of increased turnover of the mutant transcript by the NMD pathway. In Upf2^25G animals, Adhⁿ⁴ levels increased 13-fold, such that Adhⁿ⁴ and Adh⁺ transcript levels were equivalent, implying that the 25G mutation abolishes NMD pathway function. By contrast, Smg1^32AP had only a modest effect, increasing Adhⁿ⁴ levels by 30%.

Figure 3

Effect of Photoshop Mutations on Steady State Levels of an Adh Transcript Containing a Nonsense Mutation

The NMD Pathway Targets the SV40 3′ UTR of Drosophila Transgenes

To investigate how the NMD genes influence transgene expression, we examined the effect of photoshop mutations on transgenes containing different reporter genes, promoters, and 3′ UTRs (Table 1; Figure 4

). All of the transgenes whose steady state expression levels were found to be upregulated in photoshop mutants were constructed in the pUAST vector, the standard vector for gene misexpression in Drosophila [27]. This vector contains multiple binding sites for the yeast transcription factor GAL4 followed by a core promoter derived from the hsp70 gene upstream (5′) of the reporter insertion site, and the SV40 3′ small t antigen intron and polyadenylation signal (henceforth referred to as the SV40 3′ UTR) downstream (3′) of the insertion site [28]. Hence, it was possible that the photoshop phenotype was due to increased GAL4 expression or activity, increased hsp70 promoter activity, an interaction with the SV40 3′ UTR, or some special feature of the GFP and DsRed coding sequences. Upf2^25G enhanced expression of the pUAST-eGFP transgene (Figure 4

B), which has a metazoan optimized codon bias, suggesting that the unusual codon bias of native GFP and DsRed genes is not critical for the photoshop effect. The GFP and DsRed reporters do not contain introns in their coding sequences, so we inserted a 61-bp intron from Drosophila gene CG3585 into pUAST-eGFP to make pUAS-eGFP+I. This transgene did not show significantly increased eGFP signal in wild type and was still sensitive to photoshop mutations (Figure 4

C), suggesting that the photoshop effect is not through surveillance of intronless coding regions.

Table 1

Transgenes Tested for Enhanced Expression in Photoshop Mutants

Figure 4

Effect of a Photoshop Mutation on Expression of GFP Transgenes

One tested transgene, src:GFP, which was not upregulated in Smg1^32AP animals, contains the same GAL4 binding sites and hsp70 promoter as the photoshop-sensitive pUAST constructs, but its 3′ UTR and polyA signal are derived from Drosophila hsp70 instead of SV40 [29]. This implicated the SV40 3′ UTR in the photoshop effect. To test the role of the SV40 3′ UTR directly, we replaced the SV40 3′ UTR in pUAST-GFP and pUAST-GFP+I constructs with the hsp70 3′ UTR. Expression of these constructs was insensitive to photoshop mutations (Figure 4

D and and4E),4

E), confirming the importance of the 3′ UTR. Interestingly, expression levels of GFP from the constructs containing the hsp70 3′ UTR were somewhat lower than those of the corresponding pUAST-eGFP or pUAST-eGFP+I transgenes in a wild-type background and much lower than those of pUAST-eGFP or pUAST-eGFP+I transgenes in a photoshop mutant background. The significance of this finding is discussed below.

A small intron present in the SV40 3′ UTR was an appealing candidate for sensitizing transcripts to the photoshop effect. A model for recognition of PTCs in vertebrates is that they occur 5′ of the site of an intron, the same arrangement of the termination codons of GFP and DsRed with respect to the SV40 intron. RT-PCR experiments confirmed that the SV40 intron was indeed recognized and spliced in Drosophila larvae. However, deletion of this intron in the pUAST-eGFP and pUAST-eGFP+I constructs (to make pUASTΔI-eGFP and pUASTΔI-eGFP+I) did not eliminate the photoshop effect (Figure 4

F and and4G).4

G). Thus, targeting of the SV40 3′ UTR by the NMD pathway does not require splicing of this intron, the only known intron in the UTR.

Targets of the NMD Pathway during Development

To identify candidate endogenous targets of the NMD pathway, we compared steady state RNA levels in hemizygous Upf2^25G male larvae to those in wild-type male larvae using a whole genome microarray. To avoid confounding effects of the Upf2 mutation on developmental progression, we focused our analysis on a set of 954 genes that do not undergo significant changes in expression levels during development or differ in expression between male and female larvae (see Materials and Methods). Among this set, we found 14 genes whose expression was upregulated 2-fold or more, and 26 genes that were downregulated 2-fold or more in the mutant compared to wild type (Tables S1 and S2). Genes whose expression was downregulated could be novel targets whose expression is paradoxically enhanced by the NMD pathway, or they could be indirect effects of NMD pathway inactivation. The affected genes encode proteins of diverse classes, including signal transduction molecules, proteases, and proteins involved in cell metabolism. Most of the affected genes differed from ones identified recently in Drosophila S2 cells depleted of NMD-gene function by RNA interference [19], suggesting tissue-specific regulation.

Two well characterized genes upregulated in the photoshop mutant were analyzed further. The mRNA of orthinine decarboxylase antizyme (oda, also called gut feeling) contains a naturally occurring coding sequence frame shift [30] that causes the transcript to contain early termination codons, and is an NMD target in cultured Drosophila S2 cells [19]. oda transcript is also a target of the NMD pathway during development: oda RNA levels were increased in hemizygous Upf2^25G larvae as determined by microarray analysis (3-fold) and by quantitative RT-PCR (2-fold; Figure 5

A). However, oda transcript levels were not significantly increased in Smg1^32AP mutants (Figure 5

A).

Figure 5

Effect of Photoshop Mutations on Expression of Endogenous Genes

The sex determination gene transformer (tra) [31] is also an NMD pathway target during development, as in S2 cells [19]. tra transcript levels were increased 3- to 4-fold in hemizygous Upf2^25G larvae, as determined by microarray analysis and quantitative RT-PCR (Figure 5

A). Smg1^32AP also increased tra levels, but to a lesser extent (Figure 5

A). tra might be expected to be an NMD pathway target because the primary transcript is spliced in males to produce a long mRNA (traL) that contains an early termination codon that would likely be recognized as premature, whereas in females some of the primary transcript is alternatively spliced to produce a shorter transcript (traS) lacking the early termination codon and encoding full-length protein (Figure 5

B). Indeed, RT-PCR experiments indicate that the increase in tra levels in Upf2^25G mutant males and females was due to selective stabilization of traL (Figure 5

B and data not shown).

Drosophila NMD Genes Are Dispensable for Many Developmental Processes but Provide Cells a Competitive Growth or Survival Advantage

The above results show that the core NMD gene Upf2 is broadly active during development and influences expression of dozens of genes, including a key sex determination gene, and is required for larval viability. To identify specific cellular and developmental functions of the NMD pathway, we analyzed the effect of photoshop mutations on sex determination and cell growth and differentiation. We did not detect any defects in sex determination in adult males hemizygous for Upf2^25G, despite the observed increase in traL levels: sex-specific splicing of downstream gene dsx was normal, as was that of the sex determination genes Sxl and msl-2; sex combs and genitalia appeared normal; and males made sperm and were capable of mating. Homozygous Upf2^25G females also appeared normal. We also did not detect any defects in larval cell differentiation. Larval tracheal cell clones lacking Upf1 or Upf2 displayed wild-type morphology at all levels of branching, including tracheal terminal cell clones, which showed normal branching patterns and luminal structures (Figure 6

A and and6B).6

B). Likewise, type IV da neurons, another morphologically complex cell type [32], appeared normal when homozygous for Upf1 or Upf2 null mutations (Figure 6

C and and6D6

D and data not shown).

Figure 6

Effects of Photoshop Mutations on Tracheal System, Nervous System, and Eye Development

The only prominent defect we detected in cell clones mutant for NMD genes was an inability to contribute to adult (imaginal) structures. Whereas homozygous clones of all photoshop mutants were readily obtained in the larval tracheal system, we did not recover clones in the adult tracheal system, which is generated by proliferation of imaginal tracheal precursor cells during metamorphosis. We also did not recover large clones in adult epidermis or eyes, although small peripheral eye clones were occasionally observed (Figure 6

E and and6F).6

F). However, when the GMR-hid technique [33] was used to eliminate all heterozygous and wild-type cells in the developing eye, we found that the remaining photoshop mutant cells could proliferate, differentiate, and form an eye, although the eyes were smaller and more disorganized (rougher) than those of controls (Figure 6

G and and6H).6

H). These results suggest that Drosophila NMD genes are not required for cell proliferation, survival, or differentiation, but provide proliferating cells with a competitive growth or survival advantage during development.

Discussion

We have isolated to our knowledge the first mutations affecting NMD in Drosophila based on their ability to enhance expression of a GFP transgene, an effect we call the photoshop phenotype. Mapping of the mutations, complementation tests, and molecular analysis demonstrate that the photoshop mutations identify three genes, the Drosophila orthologs of NMD pathway genes Upf1, Upf2, and Smg1. The results show that Upf1 and Upf2 are essential genes, required for NMD and, at least in the case of Upf2, for proper expression of dozens of native mRNAs during development, including oda and the sex-nonspecific form of tra that contain early termination codons. By contrast, Smg1 is dispensable and only potentiates the NMD pathway. Genetic mosaic analysis of the Upf genes showed that they are not required for cell proliferation, survival, or complex cell differentiation events such as tracheal and neuronal growth and sprouting, but they provide proliferating imaginal cells with a competitive growth or survival advantage during development. We also mapped the cis-acting signal that confers sensitivity to the NMD pathway in the transgenic reporter assay, and discovered that it resides in the heterologous 3′ UTR present in the reporter construct. Below, we discuss the implications of these results for our understanding of the functions and mechanism of the Drosophila NMD pathway during development, and compare and contrast them with what has been found for NMD pathway function in other organisms.

Roles of NMD Genes in Drosophila Development

The finding that mutations in Drosophila NMD genes Upf1 and Upf2 cause lethality during larval development contrasts with the minor effect of mutations in the homologous genes in yeast (mutations have almost no discernable effect on growth or survival [5]) and in C. elegans (mutants are viable and have only morphogenetic defects late in development [6]). Why are Upf1 and Upf2 essential in Drosophila? One possibility is that they are required to eliminate mutant transcripts with PTCs that encode truncated, deleterious protein products. Such PTC-containing alleles could be present in the background of our Upf1 and Upf2 mutants, and in the absence of NMD activity, these mutations become lethal. However, all our Upf1 and Upf2 mutations were independently isolated, and the lethality in these lines segregates with the Upf mutations, not with other genomic regions. Thus, if this explanation is correct, there would have to be multiple, potentially lethal PTC mutations distributed throughout the genome.

A second possibility is that the NMD pathway has a more general surveillance function that also eliminates naturally occurring transcripts resulting from aberrant splicing events or repetitive DNA elements, and accumulation of such transcripts is toxic. However, many aspects of cell biology and development appear normal in Upf mutants. Indeed, sensitive assays examining individual tracheal cells and neurons show that loss of Upf function does not lead to cell death or impairment of complex cell morphogenesis events, implying that NMD inactivation does not cause general cellular toxicity.

A third possibility, which we favor, is that the NMD pathway modulates the activity of specific native transcripts, whose misregulation leads to lethality. An initial microarray survey identified several dozen genes of diverse functional classes whose expression was altered in NMD mutant larvae (Tables S1 and S2). Some of the affected transcripts contain early stop codons that are interpreted as bona fide PTCs, as for tra and oda genes (Figure 5

A); other affected transcripts could harbor other cis-acting signals that are interpreted as aberrant by the NMD machinery, like the SV40 3′ UTR discussed below. In the absence of the NMD pathway, overexpression of such transcripts would perturb the development or function of select cells or tissues and lead to lethality. Indeed, although many cells appear to develop normally in photoshop mutants, we never observed Upf1⁻ and Upf2⁻ clones in adult tracheae or epidermis, and only small clones were found in the eye, implying that the NMD pathway promotes growth or survival of the proliferating imaginal cells that give rise to these tissues. Because Upf genes are broadly expressed [34] and broadly active (Figure 1

F and and1G1

G and unpublished data) throughout development, identification of the tissue focus of Upf lethality will be an important first step towards identifying the critical cellular targets of NMD gene regulation.

The sole cellular defect we identified in NMD mutants was the absence or small size of mutant clones in the adult tissues described above, which is reminiscent of the cell cycle arrest observed in cultured Drosophila S2 cells depleted of Upf function [19]. Although these results suggest a function for the NMD pathway in cell proliferation or survival, the requirement for the pathway in these processes is not absolute: Upf1⁻ and Upf2⁻ cells were able to proliferate and form an eye when competing wild-type eye progenitor cells were eliminated (Figure 6

). This implies that the NMD pathway provides proliferating imaginal cells with a competitive growth advantage that prevents them from being overtaken by other proliferating cells during development. This could be a cell autonomous effect, like Minute mutations [35], or a cell nonautonomous effect, e.g., if the pathway influences expression of a signal secreted from proliferating cells that affects growth of neighboring cells (e.g., [36]). The NMD pathway may play a similar, or more extreme, role in mice, because a mouse UPF1 knockout is lethal and attempts at establishing UPF1^−/− embryonic stem cells were unsuccessful [18].

Recently, it has been suggested that Upf1 and Upf2 participate in other aspects of gene regulation besides NMD, such as stimulating translation [37] and in translational termination [38] (reviewed by [39]). It is important to note that for neither mouse nor our Drosophila mutants is it established that the lethality associated with Upf1 and Upf2 mutations derives from their roles in NMD. Indeed, our analysis of Upf2^25G suggests the opposite possibility. This allele appears completely compromised for NMD, as assessed by expression of an Adh mRNA carrying a PTC (Figure 3

), yet unlike Upf2 null alleles, some Upf2^25G mutants survived to adulthood. Thus, either the essential function of Upf2 is in a process other than NMD, or this allele retains residual NMD function sufficient for regulation of its essential target genes but not of others like the mutant Adh mRNA.

The differing molecular, cellular, and developmental requirements for NMD- pathway genes in yeast, C. elegans, Drosophila, and mice make clear that the function of this pathway has diversified during evolution. Perhaps the ancestral function of the pathway was in some general process like translation termination, and only later did the pathway evolve roles in monitoring transcripts for PTCs and more specialized regulatory roles. Alternatively, the ancestral function could have been regulation of RNAs involved in a specific cellular process such as cell growth regulation, and only later did the pathway acquire a more general role in RNA surveillance.

Smg1 Potentiates the Drosophila NMD Pathway

Our genetic analysis demonstrated a striking difference in the developmental requirements of Smg1 compared to those of Upf1 and Upf2. First, Upf1 and Upf2 are essential genes, whereas an amorphic Smg1 allele resulted in viable and fertile animals. Second, a Upf2 mutation abolished NMD of an Adh PTC allele, whereas the amorphic Smg1 mutation only modestly reduced NMD efficiency. Third, the magnitude of the Smg1 mutant effect differed at different targets. At some targets, such as oda, there was little or no effect of the Smg1 mutation, whereas at other targets, such as tra and a GFP transgene, the Smg1 mutant effect was up to half that of the Upf2 mutant. The small and gene-selective effect of Smg1 could explain why a recent genetic analysis failed to detect a role for Drosophila Smg1 in NMD [21], whereas earlier Drosophila cell culture studies suggested an important role for the gene [10]. The small and gene-selective function of Drosophila Smg1 contrasts with genetic results in C. elegans, which did not identify differences in the requirements of smg-1 and the Upf1 and Upf2 homologs smg-2 and smg-3 [6]. One possibility is that Drosophila has another protein with activity similar to that of Smg1. This seems unlikely because Smg1 is the only sequence ortholog of Smg1-family genes in the Drosophila genome, although there are other genes that encode proteins with PI3K-related kinase domains. Another possibility is that phosphorylation of Upf1 by Smg1 is not absolutely required for Upf1 activity in Drosophila but only enhances its activity or reactivates spent protein after a catalytic cycle. This would be more similar to the NMD pathway in yeast, which lacks a Smg1 ortholog and is thought to function without a Upf1 phosphorylation cycle, than to the NMD pathways in C. elegans and vertebrates, where the Upf1 phosphorylation cycle is thought to be essential for pathway activity.

Targeting of a Specific 3′ UTR by the Drosophila NMD Pathway

We found that a variety of reporter constructs in the pUAST transformation vector were upregulated when NMD pathway function was abrogated (Table 1), implying that transcripts derived from this vector are recognized as aberrant by the RNA surveillance machinery. Strictly speaking, this is not an NMD process, because all of the transgenic constructs contain full-length coding sequences with no PTCs. However, because multiple NMD genes are involved in this regulation, and because the observed increase in reporter activity is associated with increased reporter mRNA, it suggests that an NMD-related RNA decay process normally limits expression of pUAST transgenes in Drosophila. The signal that targets transgenes for regulation by the NMD pathway appears to lie in the SV40 3′ UTR of the pUAST vector: all reporter constructs that were sensitive to the NMD pathway contain this UTR, and swapping it for one derived from the hsp70 gene rendered the transcript insensitive to NMD, the first example to our knowledge of a change in NMD pathway sensitivity due solely to swapping intact 3′ UTRs. Transgenes containing three other endogenous 3′ UTRs (K10, drs, and His2AvD; see Table 1) were also insensitive to NMD, supporting the conclusion that the SV40 3′ UTR harbors a critical targeting element. Although we have not identified the specific sequence or structural characteristic (e.g., length [17,40]) within the SV40 3′ UTR responsible for targeting by NMD machinery, it does not require the small, naturally occurring intron within the UTR, nor does it require any specific sequences in the translated region of the targeted mRNA. We conclude that a 3′ UTR can provide a critical signal for regulation by the NMD pathway in Drosophila, as has been observed in yeast [7] and humans [17].

Our analysis also indicates that the SV40 3′ UTR has a second, positive effect on transgene expression in Drosophila, similar to one noted in insect cell culture [41]. In wild-type Drosophila, this positive effect is partially offset by destruction of the RNA by the NMD pathway, resulting in an intermediate level of reporter expression (Figure 7

). However, in NMD pathway mutants, the positive effect of the SV40 3′ UTR is unmasked, resulting in enhanced reporter expression and the photoshop phenotype. This contrasts with results for transgenes carrying the 3′ UTR from hsp70, which does not strongly enhance transgene expression and is not targeted by the NMD pathway, resulting in a lower and photoshop-mutation-insensitive level of reporter expression.

Figure 7

Model for Drosophila NMD Pathway Action on SV40 3′ UTR

Because expression of UAS-GFP and other reporters is affected by mutations in homologs of all three NMD genes on the chromosome we screened, the assay can likely be used to identify and characterize additional NMD pathway genes on other chromosomes. The assay has two important features. First, because it can be carried out in single cells in genetically mosaic animals, a requirement of candidate NMD genes for organismal viability can be bypassed. Second, the assay is very sensitive to perturbations in the NMD pathway. For example, loss of Smg1 activity leads to only a modest increase in stability of PTC-containing transcripts but a readily detectable enhancement of GFP reporter expression. Together these features suggest that the transgenic assay system can be used to test requirements in vivo of candidate NMD genes and drugs that influence pathway activity, which could be useful in modulating expression of human disease genes carrying PTCs [42,43].

Materials and Methods

Fly stocks and genetics.

GAL4/UAS system [27] drivers used were btl-GAL4 [44], e22c-GAL4 [45], ppk-GAL4 [32], and da-GAL4 [24]. GFP and DsRed transgenes are referenced in Table 1. Df(Smg1)exe2B was generated by using FLP-mediated recombination between the FRTs in P{XP}C3G[d00589] and PBac{WH}CG3044[f02328] as described in Thibault et al. [46]. Marker mutations and balancer chromosomes are described at http://www.flybase.org. Flies were reared at 25 °C on cornmeal/dextrose medium.

The photoshop mutations were obtained by mutagenesis of an isogenized y w FRT^19A chromosome [22] with 25 mM ethane methyl sulfonate overnight [47] in a tracheal mutant screen (to be described elsewhere). The mutations used were on this chromosome unless otherwise noted. To generate homozygous mutant clones, 2- to 6-h-old embryos were collected at 25 °C from a cross of y w * FRT^19A/FM7 females to gal80 FRT^19A, hsFLP122/Y; btl-GAL4, UAS-GFP males. After a 45-min heat shock at 38 °C to induce FLP expression, embryos were returned to 25 °C to continue development. L3 larvae of genotype y w * FRT^19A/gal80 FRT^19A, hsFLP122; btl-GAL4, UAS-GFP/+ were identified by GFP mosaicism within the tracheal system and scored for the photoshop phenotype.

The original Smg1^32AP chromosome (designated 32AP†) carried lethal mutations not associated with the photoshop phenotype. Lethals were removed by crossing 32AP†/y w FRT^19A females to w/Y; btl-GAL4, UAS-GFP males and identifying L3 larvae with enhanced GFP throughout their tracheal systems. These Smg1^32AP/Y; btl-GAL4, UAS-GFP/+ larvae developed into viable adult males. We also found a viable wing morphology mutation on 32AP† that is allelic to wavy. Existing wavy alleles do not show a photoshop phenotype, and the wing phenotype is separable from the photoshop phenotype, so wavy does not seem to contribute to the photoshop phenotype.

The Upf2^29AA chromosome carries a linked lethal mutation. When recombined away from Upf2^29AA, the mutation had no effect on tracheal development or reporter expression. However, we have not obtained a recombinant containing Upf2^29AA without the extraneous mutation.

For complementation tests of Upf2, we used genomic rescue transgenes located on the autosomes to generate males of genotype y w Upf2^14J v g f FRT^19A/Y; P{w⁺, Upf2⁺}/+ and crossed these to y w * FRT^19A/FM7c females, where the asterisk indicates the tested mutation. Absence of Bar⁺, white-eyed female progeny indicated failure to complement. For complementation tests of Upf1, we used the Y-linked duplication Dp(1;Y)BSC1, y⁺, which covers the Upf1 locus. Males of genotype Upf1^13D/Dp(1;Y)BSC1, y⁺ were crossed to Upf1^26A/FM7c females, and the absence of female Bar⁺ progeny indicated a failure to complement.

Mapping of photoshop mutations.

Identification of SNPs, construction of the SNP map of the X chromosome, and details of their use in mapping X chromosome mutations will be described elsewhere. Briefly, the location of the lethality associated with a photoshop mutation was mapped by crossing y w * FRT^19A/sc cv ct v g f FRT^19A females (where the asterisk indicates the lethal mutation) to FM7c/Y males and scoring the viable male progeny for the visible markers to determine the lethal interval. To refine the map position, we collected male progeny in which a recombination event had occurred within the mapped interval and scored them for SNPs. For each mutation we typically scored 300–400 males for SNPs. For Upf1^26A and Upf1^13D, which map between v and g, we crossed the y w * FRT^19A/sc cv ct v g f FRT^19A females to males of genotype Df(1)64c18, g¹ sd¹/Dp(1;2;Y)w⁺ to distinguish y sc⁺ w cv⁺ ct⁺ v⁺ g f recombinants, which we could not otherwise identify because of epistasis of w over v and g. We also crossed the y w * FRT^19A/sc cv ct v g f FRT^19A females to sc cv ct v g f FRT^19A/Y males to identify recombinant females, which were then tested for the photoshop phenotype in genetic mosaics to confirm that the lethality and photoshop phenotype mapped to the same interval. For the viable mutation 32AP, we followed a similar strategy as for the lethals, except we scored recombinant males for the presence or absence of 32AP by testing the enhancement of GFP in btl-GAL4, UAS-GFP transgenic animals.

Transgene construction.

For the Upf2 genomic rescue construct, Drosophila BAC 24A2 [48], which contains Upf2, was transformed into Escherichia coli strain EL250, which harbors heat-shock-inducible homologous recombination machinery [49]. We then cloned a 200-bp fragment located upstream, and a 300-bp fragment located downstream, of Upf2 coding sequence and UTRs based on the cDNA RE04053 (rather than the canonical Upf2 cDNA SD07232, which appears to be defective as it lacks a conserved portion of Upf2 coding sequence) tandemly into the Drosophila transformation vector pCaSpeR4 [50] with a unique NotI site between the fragments to give pMM#200. pMM#200 was linearized with NotI and transformed into EL250[24A2], in which the recombination machinery had been induced. Transformants were plated onto LB plates containing carbenicillin to select for gap repair of pMM#200, which can occur by homologous recombination with the BAC and result in transfer of Upf2 into pMM#200. The resultant plasmids were analyzed by restriction digestion, and one with the expected pattern (pMM#201) was used to establish transgenic lines on the second and third chromosomes by P-element transformation. Six lines were tested and all six rescued hemizygous male Upf2^14J mutants to viability. The degree of rescue varied based on insertion site, but for the strongest lines (P{w⁺, Upf2⁺}11A and 24) rescued animals appeared indistinguishable from Upf2⁺ animals and were readily maintained as stocks of genotype Upf2^25G; P{w⁺, Upf2⁺}11A or 24/+ or Upf2^14J; P{w⁺, Upf2⁺}11A or 24/+.

For GFP reporter constructs, coding sequence of eGFP (Clontech; http://www.clontech.com) was amplified by PCR using KpnI linker primers Kpn5GFP and Kpn3GFP (see Table S3 for primer sequences). The product was digested with KpnI and cloned into the KpnI site of pUAST to generate pUAST-eGFP. A PCR-based strategy was used to insert a 61-bp intron derived from gene CG3585 between nucleotides 330 and 331 of the eGFP coding sequence to make eGFP+I constructs.

The intron in the SV40 3′ UTR of pUAST was deleted by a PCR-based strategy to give pUASTΔI. To replace the SV40 3′ UTR with the hsp70 3′ UTR, primers XbaHsp70 and Hsp70Stu were used to amplify the hsp70 3′ UTR from pGATB [27]. The PCR product was digested with XbaI and StuI and cloned between the XbaI and StuI restriction sites of pUAST-eGFP (to make pUAST.h-eGFP) or pUAST-eGFP+I (to make pUAST.h-eGFP+I).

Constructs made using PCR were sequenced to confirm that mutations had not been introduced. Constructs were transformed into Drosophila using standard microinjection techniques using the Δ2–3 helper plasmid as the transposase source.

RNA analysis and quantitation.

L3 larvae of the appropriate genotype were identified by enhancement of GFP for mutant alleles or by sexing using gonad morphology for wild-type controls. Total larval RNA was prepared using Trizol reagent (Invitrogen; http://www.invitrogen.com), and genomic DNA contamination was eliminated with DNAse (DNA-free, Ambion; http://www.ambion.com). RNA concentration was determined spectrophotometrically and normalized before reverse transcription with MuMLV reverse transcriptase (Retroscript kit, Ambion). Real-time quantitative PCR was done with a thermocycler (iCycler) and real-time PCR mix (iQ SYBR Green Supermix; Bio-Rad Laboratories; http://www.bio-rad.com). Primers used in the qRT-PCR experiments were designed to amplify 60- to 100-bp fragments within single exons; amplification of Drosophila genomic DNA containing a UAS-GFP transgene showed that primers gave a linear amplification response at concentrations ranging over four orders magnitude. Control reactions performed on RNA without reverse transcription or with primers against nontranscribed regions of genomic DNA gave negligible signals compared to experimental reactions. Experimental reactions were carried out in duplicate (except EK161155 was done once) on two RNA samples that were derived independently from the RNA used for microarray analysis. Results were normalized to results with rp18LA transcript, a gene that is not developmentally regulated [51].

To make Upf2^25G homozygotes for amplifying tra transcript in females, we used Upf2 genomic rescue construct 11A on the second chromosome. w/Y; btl-GAL4, UAS-GFP males were crossed to Upf2^25G; P{w⁺, Upf2⁺}11A/+ females; the resultant Upf2^25G/Y; P{w⁺, Upf2⁺}11A/btl-GAL4, UAS-GFP males were crossed to Upf2^25G/FM7i, ACT-GFP females, and female larvae with enhanced tracheal GFP expression were collected. For control larvae, we crossed the Upf2^25G/Y; P{w⁺, Upf2⁺}11A/btl-GAL4, UAS-GFP males to y w FRT^19A females and collected female larvae with tracheal GFP expression. cDNA from these larvae was prepared as above.

For analysis of Adh RNA levels, cDNA was prepared from adult males of the appropriate genotype. Adh transcripts were amplified with primer AdhL and the primer AdhR (for agarose gel analysis and sequencing) or AdhR2Fam (for capillary electrophoresis). PCR products were sequenced directly or digested with PvuII to distinguish the Adh⁺ and Adhⁿ⁴ alleles. Capillary electrophoresis was performed on an ABI 3730x1 (Applied Biosystems; http://www.appliedbiosystems.com) and analyzed using GeneMapper v3.0 software (Applied Biosystems).

Microarray analysis.

RNA was isolated from Upf2^25G/Y; btl-GAL4, UAS-GFP/+ and y w FRT^19A/Y; btl-GAL4, UAS-GFP/+ L3 larvae using Trizol as described above. cDNA labeled with Cy3 or Cy5 was prepared from each RNA sample and hybridized to microarrays containing ~14,000 gene probes [52,53]. Hybridizations were performed with two independently isolated and labeled RNA samples. Analysis was carried out using the Stanford Microarray Database (http://genome-www5.stanford.edu). During analysis we noted that the Upf2^25G mutants were delayed in development. To avoid confounding effects of changes in gene expression that result from developmental regulation rather than more direct effects of Upf2 loss of function, we used the available wild-type developmental gene expression time course [51] to filter out genes whose transcription changed more than 25% from their maximal value during hours 72–96 of larval development. The wild-type dataset includes ~33% of genes, and we used just this subset for our analysis. Furthermore, the wild-type dataset is for mixed sex populations, while our microarray was performed only on males. To compensate for sex differences, we also excluded from analysis genes that differed between males and females by more than 50% based on data from male and female larvae (E. Johnson and M. A. K., unpublished data). Our analysis of genes regulated by NMD is therefore conservative, covering only non-developmentally regulated and non-sex-regulated genes whose expression was affected by a hypomorphic Upf2 allele, and thus provides only a lower estimate of genes regulated by NMD.

Supporting Information

Table S1: Genes Upregulated More than 2-Fold in Microarray Analysis of Upf2^25G Larval RNA

(22 KB DOC)

Click here for additional data file.^{(23K, doc)}

Table S2: Genes Downregulated More than 2-Fold in Microarray Analysis of Upf2^25G Larval RNA

(36 KB DOC)

Click here for additional data file.^{(36K, doc)}

Table S3: Sequences of DNA Primers Used

(31 KB DOC)

Click here for additional data file.^{(31K, doc)}

Accession Numbers

The FlyBase (http://flybase.net)/Entrez Gene (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene) accession numbers for the genes described in the text are Adh (CG3481/Gene I.D. 3771877), oda (CG16747), Smg1 (CG32743/Gene I.D. 31625), Upf1 (CG1559/Gene I.D. 32153), and Upf2 (CG2253/Gene I.D. 31724). Microarray data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo) and are accessible through accession number GSE5585.

Acknowledgments

We thank Nathan Trinklein, Doug Menke, and Melissa Marks for technical advice; Alana O′Reilly, Rebecca Yang, Christiane Nüsslein-Volhard, the Bloomington Drosophila Stock Center, and members of the lab for fly stocks; Molly Weaver for help examining adult tracheal clones; Stefan Luschnig, Inga Spiess, Eric Johnson, and Michelle Arbeitman for help with microarray experiments; and Doris Chen for SNP data. We thank Gillian Stanfield and members of the lab for comments on the manuscript. Photoshop is a registered trademark of Adobe Systems Incorporated, used with permission.

Abbreviations


EJC	exon junction complex

FLP	Saccharomyces cerevisiae FLP1 recombinase

FRT	FLP1 recombinase target

GFP	green fluorescent protein

L3	third larval instar

NMD	nonsense-mediated mRNA decay

oda	orthinine decarboxylase antizyme

PTC	premature termination codon

smg	suppressor with morphogenetic effect on genitalia

SNP	single nucleotide polymorphism

tra	transformer

Upf	up-frameshift suppressor

Footnotes

¤ Current address: Department of Human Genetics, University of Utah, Salt Lake City, Utah, United States of America

A previous version of this article appeared as an Early Online Release on September 6, 2006 (DOI: 10.1371/journal.pgen.0020180.eor).

Author contributions. MMM and MAK conceived the experiments. MMM performed the experiments. MMM and MAK analyzed the data and wrote the paper.

Funding. MMM was supported by a Helen Hay Whitney Foundation postdoctoral fellowship. MAK is an investigator of the Howard Hughes Medical Institute.

Competing interests. The authors have declared that no competing interests exist.

References

Maquat LE, Kinniburgh AJ, Rachmilewitz EA, Ross J. Unstable beta-globin mRNA in mRNA-deficient ß⁰ thalassemia. Cell. 1981;27:543–553. [PubMed]
Noensie EN, Dietz HC. A strategy for disease gene identification through nonsense-mediated mRNA decay inhibition. Nat Biotechnol. 2001;19:434–439. [PubMed]
Inoue K, Khajavi M, Ohyama T, Hirabayashi S, Wilson J, et al. Molecular mechanism for distinct neurological phenotypes conveyed by allelic truncating mutations. Nat Genet. 2004;36:361–369. [PubMed]
Neu-Yilik G, Gehring NH, Hentze MW, Kulozik AE. Nonsense-mediated mRNA decay: From vacuum cleaner to Swiss army knife. Genome Biol. 2004;5:218. [PubMed]
Leeds P, Wood JM, Lee BS, Culbertson MR. Gene products that promote mRNA turnover in Saccharomyces cerevisiae. Mol Cell Biol. 1992;12:2165–2177. [PubMed]
Hodgkin J, Papp A, Pulak R, Ambros V, Anderson P. A new kind of informational suppression in the nematode Caenorhabditis elegans. Genetics. 1989;123:301–313. [PubMed]
Amrani N, Ganesan R, Kervestin S, Mangus DA, Ghosh S, et al. A faux 3′-UTR promotes aberrant termination and triggers nonsense-mediated mRNA decay. Nature. 2004;432:112–118. [PubMed]
Lykke-Andersen J, Shu MD, Steitz JA. Human Upf proteins target an mRNA for nonsense-mediated decay when bound downstream of a termination codon. Cell. 2000;103:1121–1131. [PubMed]
Kashima I, Yamashita A, Izumi N, Kataoka N, Morishita R, et al. Binding of a novel SMG-1-Upf1-eRF1-eRF3 complex (SURF) to the exon junction complex triggers Upf1 phosphorylation and nonsense-mediated mRNA decay. Genes Dev. 2006;20:355–367. [PubMed]
Gatfield D, Unterholzner L, Ciccarelli FD, Bork P, Izaurralde E. Nonsense-mediated mRNA decay in Drosophila: At the intersection of the yeast and mammalian pathways. EMBO J. 2003;22:3960–3970. [PubMed]
Conti E, Izaurralde E. Nonsense-mediated mRNA decay: Molecular insights and mechanistic variations across species. Curr Opin Cell Biol. 2005;17:316–325. [PubMed]
Sheth U, Parker R. Targeting of aberrant mRNAs to cytoplasmic processing bodies. Cell. 2006;125:1095–1109. [PubMed]
Yamashita A, Kashima I, Ohno S. The role of SMG-1 in nonsense-mediated mRNA decay. Biochim Biophys Acta. 2005;1754:305–315. [PubMed]
Yamashita A, Ohnishi T, Kashima I, Taya Y, Ohno S. Human SMG-1, a novel phosphatidylinositol 3-kinase-related protein kinase, associates with components of the mRNA surveillance complex and is involved in the regulation of nonsense-mediated mRNA decay. Genes Dev. 2001;15:2215–2228. [PubMed]
Lejeune F, Maquat LE. Mechanistic links between nonsense-mediated mRNA decay and pre-mRNA splicing in mammalian cells. Curr Opin Cell Biol. 2005;17:309–315. [PubMed]
Amrani N, Dong S, He F, Ganesan R, Ghosh S, et al. Aberrant termination triggers nonsense-mediated mRNA decay. Biochem Soc Trans. 2006;34:39–42. [PubMed]
Buhler M, Steiner S, Mohn F, Paillusson A, Muhlemann O. EJC-independent degradation of nonsense immunoglobulin-mu mRNA depends on 3′ UTR length. Nat Struct Mol Biol. 2006;13:462–464. [PubMed]
Medghalchi SM, Frischmeyer PA, Mendell JT, Kelly AG, Lawler AM, et al. Rent1, a trans-effector of nonsense-mediated mRNA decay, is essential for mammalian embryonic viability. Hum Mol Genet. 2001;10:99–105. [PubMed]
Rehwinkel J, Letunic I, Raes J, Bork P, Izaurralde E. Nonsense-mediated mRNA decay factors act in concert to regulate common mRNA targets. RNA. 2005;11:1530–1544. [PubMed]
Azzalin CM, Lingner J. The human RNA surveillance factor UPF1 is required for S phase progression and genome stability. Curr Biol. 2006;16:433–439. [PubMed]
Chen Z, Smith KR, Batterham P, Robin C. Smg1 nonsense mutations do not abolish nonsense-mediated mRNA decay in Drosophila melanogaster. Genetics. 2005;171:403–406. [PubMed]
Xu T, Rubin GM. Analysis of genetic mosaics in developing and adult Drosophila tissues. Development. 1993;117:1223–1237. [PubMed]
Lee T, Luo L. Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron. 1999;22:451–461. [PubMed]
Wodarz A, Hinz U, Engelbert M, Knust E. Expression of crumbs confers apical character on plasma membrane domains of ectodermal epithelia of Drosophila. Cell. 1995;82:67–76. [PubMed]
He F, Brown AH, Jacobson A. Upf1p, Nmd2p, and Upf3p are interacting components of the yeast nonsense-mediated mRNA decay pathway. Mol Cell Biol. 1997;17:1580–1594. [PubMed]
Chia W, Savakis C, Karp R, Ashburner M. Adhⁿ⁴ of Drosophila melanogaster is a nonsense mutation. Nucleic Acids Res. 1987;15:3931. [PubMed]
Brand AH, Perrimon N. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development. 1993;118:401–415. [PubMed]
Thummel CS, Boulet AM, Lipshitz HD. Vectors for Drosophila P-element-mediated transformation and tissue culture transfection. Gene. 1988;74:445–456. [PubMed]
Kaltschmidt JA, Davidson CM, Brown NH, Brand AH. Rotation and asymmetry of the mitotic spindle direct asymmetric cell division in the developing central nervous system. Nat Cell Biol. 2000;2:7–12. [PubMed]
Ivanov IP, Simin K, Letsou A, Atkins JF, Gesteland RF. The Drosophila gene for antizyme requires ribosomal frameshifting for expression and contains an intronic gene for snRNP Sm D3 on the opposite strand. Mol Cell Biol. 1998;18:1553–1561. [PubMed]
Boggs RT, Gregor P, Idriss S, Belote JM, McKeown M. Regulation of sexual differentiation in D. melanogaster via alternative splicing of RNA from the transformer gene. Cell. 1987;50:739–747. [PubMed]
Grueber WB, Ye B, Moore AW, Jan LY, Jan YN. Dendrites of distinct classes of Drosophila sensory neurons show different capacities for homotypic repulsion. Curr Biol. 2003;13:618–626. [PubMed]
Stowers RS, Schwarz TL. A genetic method for generating Drosophila eyes composed exclusively of mitotic clones of a single genotype. Genetics. 1999;152:1631–1639. [PubMed]
Alonso CR, Akam M. A Hox gene mutation that triggers nonsense-mediated RNA decay and affects alternative splicing during Drosophila development. Nucleic Acids Res. 2003;31:3873–3880. [PubMed]
Morata G, Ripoll P. Minutes: Mutants of Drosophila autonomously affecting cell division rate. Dev Biol. 1975;42:211–221. [PubMed]
Ryoo HD, Gorenc T, Steller H. Apoptotic cells can induce compensatory cell proliferation through the JNK and the Wingless signaling pathways. Dev Cell. 2004;7:491–501. [PubMed]
Nott A, Le Hir H, Moore MJ. Splicing enhances translation in mammalian cells: An additional function of the exon junction complex. Genes Dev. 2004;18:210–222. [PubMed]
Keeling KM, Lanier J, Du M, Salas-Marco J, Gao L, et al. Leaky termination at premature stop codons antagonizes nonsense-mediated mRNA decay in S. cerevisiae. RNA. 2004;10:691–703. [PubMed]
Wilkinson MF. A new function for nonsense-mediated mRNA-decay factors. Trends Genet. 2005;21:143–148. [PubMed]
Muhlrad D, Parker R. Aberrant mRNAs with extended 3′ UTRs are substrates for rapid degradation by mRNA surveillance. RNA. 1999;5:1299–1307. [PubMed]
Huynh CQ, Zieler H. Construction of modular and versatile plasmid vectors for the high-level expression of single or multiple genes in insects and insect cell lines. J Mol Biol. 1999;288:13–20. [PubMed]
Holbrook JA, Neu-Yilik G, Hentze MW, Kulozik AE. Nonsense-mediated decay approaches the clinic. Nat Genet. 2004;36:801–808. [PubMed]
Ainsworth C. Nonsense mutations: Running the red light. Nature. 2005;438:726–728. [PubMed]
Shiga Y, Tanaka-Matakatsu M, Hayashi S. A nuclear GFP/ß-galactosidase fusion protein as a marker for morphogenesis in living Drosophila. Dev Growth Differ. 1996;38:99–106.
Lawrence PA, Bodmer R, Vincent JP. Segmental patterning of heart precursors in Drosophila. Development. 1995;121:4303–4308. [PubMed]
Thibault ST, Singer MA, Miyazaki WY, Milash B, Dompe NA, et al. A complementary transposon tool kit for Drosophila melanogaster using P and piggyBac. Nat Genet. 2004;36:283–287. [PubMed]
Lewis EB, Bacher F. Methods of feeding ethyl methane sulfonate (EMS) to Drosophila males. Drosoph Inf Serv. 1968;43:193.
Hoskins RA, Nelson CR, Berman BP, Laverty TR, George RA, et al. A BAC-based physical map of the major autosomes of Drosophila melanogaster. Science. 2000;287:2271–2274. [PubMed]
Lee EC, Yu D, Martinez de Velasco J, Tessarollo L, Swing DA, et al. A highly efficient Escherichia coli-based chromosome engineering system adapted for recombinogenic targeting and subcloning of BAC DNA. Genomics. 2001;73:56–65. [PubMed]
Thummel CS, Pirrotta V. Technical notes: New pCasper P-element vectors. Drosoph Inf Serv. 1992;71:150.
Arbeitman MN, Furlong EE, Imam F, Johnson E, Null BH, et al. Gene expression during the life cycle of Drosophila melanogaster. Science. 2002;297:2270–2275. [PubMed]
Gerber AP, Luschnig S, Krasnow MA, Brown PO, Herschlag D. Genome-wide identification of mRNAs associated with the translational regulator PUMILIO in Drosophila melanogaster. Proc Natl Acad Sci U S A. 2006;103:4487–4492. [PubMed]
Johnston R, Wang B, Nuttall R, Doctolero M, Edwards P, et al. FlyGEM, a full transcriptome array platform for the Drosophila community. Genome Biol. 2004;5:R19. [PubMed]
Reichhart JM, Ferrandon D. Green balancers. Drosoph Inf Serv. 1998;81:201–202.
Clarkson M, Saint R. A His2AvDGFP fusion gene complements a lethal His2AvD mutant allele and provides an in vivo marker for Drosophila chromosome behavior. DNA Cell Biol. 1999;18:457–462. [PubMed]
Rorth P. Gal4 in the Drosophila female germline. Mech Dev. 1998;78:113–118. [PubMed]

Formats:

PubMed articles by these authors

PubMed related articles

Recent Activity

Links

Simple NCBI Directory

Getting Started

Resources

Popular

Featured

NCBI Information