All fly stocks were maintained at 22–25°C on standard medium unless otherwise indicated. myc⁴, UAS-Myc, hs-FLP; Act>CD2>Gal4, UAS-GFP and the lacZ reporter transgenes have been previously described [29]–[31]. The Act^TS stock was generated by combining Tubulin-Gal80^TS (2^nd chromosome; Bloomington stock center) with Actin-Gal4 (3^rd chromosome; Bloomington stock center). Apterous-Gal4 (ap-Gal4) was obtained from the Bloomington stock center. The UAS-p35 strain was obtained from Dr. Bruce Edgar (University of Heidelberg) and is a P element insertion on the 3rd chromosome that has been used previously to inhibit cell death (e.g. Jiang et al. [32]).

To construct myc:GFP flies, P[acman] BAC CH321-88A16, covering 72 kb of genomic DNA from the X chromosome (3224568 to 3296675) [33] was used to drive Myc expression. Recombineering was performed to tag Myc in frame with the coding region as described in Jungreis et al., [34] using plasmids containing the recombineering plasmids (Donald Court, National Cancer Institute, Frederick, MD) [35]. The stop codon for myc was replaced with the coding sequence for superfolder EGFP codon-optimized for Drosophila [36] followed by sequence coding for a flag tag, a Precision site, a TEV site and the biotin ligase recognition peptide. The tagged construct was injected into a stock containing an attP site on the 3^rd chromosome (Bloomington Stock Number 24871). Integration was marked with w⁺. Transformants were validated by PCR amplifying across the attL and attR integration events as described previously by Venken et al. [37], and the final exon-tag junction (myc specific primer: ATACGATCGAGAAGCGCAAT, EGFP reverse primer: CGATGTTTCGCTTGGTGG-TCGAAT). PCR products of the expected sizes were found to support insertion of the BAC at the attP site and the retention of the tag. The line is available from the Bloomington Stock Center, stock number 38633.

To determine lacZ mutation frequency in larvae, hs-FLP; UAS-Myc (or hs-FLP control) females were crossed to lacZ #2 (or lacZ #9); Actin>CD2>Gal4, UAS-GFP males. 1^st instar larvae (24–48 hrs AED) were heat shocked at 37°C for 45 minutes to induce recombination in ∼90% of cells and 3^rd instar larvae were collected 3 days later. Larvae were sexed and then stored at −80°C until assays were carried out as described previously [27], [29]. 30 larvae were used for each lacZ experiment. For lacZ mutation frequency in adults, Tub-Gal80^TS; Actin-Gal4/TM6B females were crossed to lacZ #2; UAS-Myc (or lacZ #2; +) males at 18°C. From this cross, Tub-Gal80^TS/lacZ #2; Actin-Gal4/UAS-Myc (and Tub-Gal80^TS/lacZ #2; Actin-Gal4/+) females were placed at 25°C or 29°C to induce expression of Myc for 10 or 5 days, respectively. Adults were stored at −80°C until lacZ assays were carried out. 50 adult flies were used per lacZ experiment and 3–5 experiments were carried out for each genotype. The mutation frequency is the ratio of colonies growing on the selective plate vs. the total number of recovered plasmids from the DNA sample (as measured on the titer plate). Hence, mutation frequencies as determined with this system reflect a ratio and do not depend on the amount of DNA. They are expressed on a per locus basis as the number of mutant lacZ copies for a given number of lacZ copies isolated from the in vivo situation.

Categorizing lacZ mutations into “point mutations” and “genome rearrangements” was carried out as previously described [27], [29]. Candidate genome rearrangements were sequenced using lacZ primers as previously described [27] and breakpoints were mapped to the Drosophila genome sequence. Briefly, the Sanger sequences for each sample were aligned against the sequence of the LacZ construct using blastn from the NCBI Blast+ toolkit [38], version 2.2.26. The top hit for each sequence was taken, and BEDTools [39] was used to combine the alignments; the resulting intervals of the construct present in the Sanger sequences were reported. In addition, alignments whose start sites were at a position after their end sites were reported as inversions. Finally, the Sanger sequences were aligned to the organism genome with blastn and the top hit, if any was found, for each sequence was reported, along with its position in the construct as inferred from partial alignment of the Sanger sequence with the construct.

For antibody staining, third instar larvae were dissected, fixed and stained as previously described [40]. Anti-γ-H2A.X was obtained from Active Motif and used at 1∶500. Secondary antibodies were obtained from Invitrogen. For Western bot analyses, larval wing imaginal discs or adult heads of the appropriate genotype were homogenized in SDS load buffer, sonicated and run on a 4–12% gel (Invitrogen). Western blots were carried out using standard procedures and Myc protein detected using anti-Myc mouse monoclonal antibody that has been described previously [41] (used at 1∶500). Anti-gamma-tubulin antibody was obtained from Sigma and used at 1∶2000. Westerns were analyzed using infrared conjugated secondary antibodies (LiCOR) and Odyssey scanner and software (v3.0) to quantitate levels of protein using γ-tubulin to normalize loading. Quantitation was carried out from three independent Westerns per genotype using samples from independent crosses.

Flies of the specified genotype were placed at low density (20–25 flies per vial) at the temperature indicated (with 50% humidity and 12 hour light/dark cycle). The number of dead flies was scored every 2–3 days and the living flies transferred to new food. Lifespan experiments were carried out at least three times from independent parents using ∼100 flies per genotype. Representative lifespan experiments are shown in the Figures. The myc⁴ null allele and the myc:GFP transgene were backcrossed into the w¹¹¹⁸ background for five generations to limit genetic background effects. The UAS-Myc and lacZ transgenes were both generated in w¹¹¹⁸.

The statistical significance of lacZ mutation frequency data and γ-H2A.X staining were determined using a student’s t-test (Microsoft excel). Statistical significance of lifespan data were determined with Log-rank (Mantel-cox) test using Prism (GraphPad software).

In mammalian cells, overexpression of c-Myc increases the number of DNA double strand breaks (DSBs), as assayed by the level of γ-H2A.X, a phospho-histone marker of DSBs [9], [11]–[13], [15]. To determine if this is a conserved feature of Myc family proteins, we overexpressed Drosophila Myc in the dorsal compartment of the larval wing imaginal disc using apterous-Gal4 (ap-Gal4) and showed that this led to a ∼2-fold increase in the number of γ-H2A.X positive cells (Figure S1). Myc overexpression can cause apoptosis in wing disc and other mitotically dividing (but not endoreplicating) cells, which also induces phosphorylation of H2A.X [30], [41]. We therefore co-expressed Myc with the cell death inhibitor p35 and re-quantified the number of γ-H2A.X positive cells. Jointly overexpressing Myc and p35 increases the number of γ-H2A.X positive cells ∼2.5-fold compared to overexpression of p35 alone (Figure 1). Excess Myc therefore indirectly or directly increases the frequency of DSBs.

To establish whether Drosophila Myc’s ability to increase the number of DSBs leads to an increased mutation load, we overexpressed Myc during larval development and assessed mutation frequency using an in vivo lacZ mutation reporter transgene (Figure 2A) [27], [29]. We ubiquitously overexpressed Myc using the UAS/Gal4 system and the established “FLP on” strategy [42]. By using a strong (45 min) heat shock during the first larval instar more than 90% of larval cells overexpress Myc as assessed by anti-Myc staining and by examining the number of GFP positive cells, which is co-expressed with Myc (data not shown). Using this system, Myc is overexpressed an average of 9-fold as assessed by quantitative anti-Myc Western blot (Figure 2B). Heat shocked animals were then allowed to develop into 3^rd instar larvae, at which point they were separated by sex, and their lacZ mutation frequency determined using total genomic DNA as described previously [27], [29]. Mutation frequencies are expressed per locus, i.e., the number of mutant lacZ genes over the number of lacZ copies recovered from a DNA sample. As shown in Figure 2C, the spontaneous mutation frequency of both male and female larvae overexpressing Myc is approximately double that of control larvae that do not contain a UAS-Myc transgene. Myc overexpression during larval development therefore increases mutation load.

Our initial mutation analysis was carried out using a lacZ reporter inserted at cytological location 36F on the 2^nd chromosome (lacZ #2) [29]. To ensure that our increased mutation frequency was not specific to this genomic location, we quantitated the mutation frequency of control and Myc overexpressing larvae carrying an independent lacZ reporter inserted at 46D on the 2^nd chromosome (lacZ line #9) [29]. As shown in Figure 2D, although the baseline number of spontaneous mutations is higher at this genomic locus [29], both male and female larvae overexpressing Myc show a ∼2-fold increase in the number of mutations in the lacZ reporter. Myc therefore increases the mutation load at more than one genomic locus, so is likely to influence the frequency of mutations genome-wide.

We next assessed the type of mutation induced by Myc overexpression during larval development. This classification is based on the fact that point mutations and small alterations to the lacZ gene will not change the size of the rescued plasmid, while larger deletions and inversions (genome rearrangements) alter the size of the plasmid (Figure 2A). Restriction enzyme digestion-based size analyses of rescued lacZ-containing plasmids results in two broad classes of mutation: point mutations and genome rearrangements [27]–[29]. As seen in Figure 2E, Myc specifically increases the number of genome rearrangements that are caused by erroneous repair of DNA double strand breaks in both male and female larvae.

In addition to quantifying the number and type of mutation caused by Myc overexpression, we also determined the precise molecular basis of lacZ reporter mutations from control and Myc overexpressing larvae. To do this, we sequenced a subset of lacZ-containing plasmids from control and Myc overexpressing larvae containing lacZ insertion #2. Because Myc overexpression increases the frequency of DSB-induced mutations and not point mutations, we restricted our sequence analyses to plasmids that were classified as genome rearrangements above (Figure 2E).

Sequencing our lacZ-containing plasmids revealed breakpoints in the lacZ gene, which we then mapped to the Drosophila genome or sequences within the P-element by BLAST searches. Based on these data, we classified our DSB-mediated mutations as either indels (insertions, deletions and inversions) or translocations (sequences from other chromosomes). As a measure of mutation severity, we subdivided our indels into “small” or “large”: small indels involve DNA sequences less than 10 kb from the lacZ breakpoint site, and large indels involve sequences more than 10 kb away. While the cutoff of 10 kb to distinguish small indel from large is arbitrary, it serves a useful purpose in comparing mutations found in control and Myc overexpressing larvae. For these analyses, we pooled samples from male and female larvae since both showed an increase in the frequency of DSB-mediated mutations. As seen in Table 1, 61.9% of spontaneous DSB-mediated mutation events in control lacZ #2-containing larvae were small indels and 28.6% were large. In contrast, in Myc overexpressing larvae only 19.2% of indels were small and the vast majority (69.3%) were large. The frequency of translocations does not vary significantly between control and Myc overexpressing larvae. From our sequence analyses of lacZ breakpoints, we also observe regions of microhomology in 85% of breakpoints in both control and Myc overexpressing larvae (Table S1). We have defined these regions of microhomology as stretches of at least 5 identical base pairs immediately flanking (within 50 bp) the breakpoint. While Myc and controls have regions of microhomology, the key difference is that while control animals repair their DSBs using sequences close to the original breakpoint, Myc overexpressing larvae repair using sequences that are more distant. Taken with our γ-H2A.X data, Myc overexpression increases both the frequency and the severity of spontaneous DSB-induced mutations.

Accumulation of mutations, particularly genome rearrangements that are caused by the misrepair of DSBs, correlates with aging across metazoans [20], [22], [43], [44], although a causal link between these observations has not been established. We therefore tested whether overexpressing Myc increases mutation frequency in adults, and whether it affects adult lifespan. Because ubiquitous Myc overexpression during embryonic and/or larval development causes pupal lethality (data not shown), we generated a fly strain that combined a ubiquitously expressed Gal4 driver (Actin-Gal4) and a temperature sensitive form of the Gal80 inhibitor protein (Tubulin-Gal80^TS; referred to as Act^TS when combined). At 18°C Act^TS is off (due to Gal80^TS activity), at 29°C the Gal80^TS repressor is inactive thus Act^TS is active, and at the in-between temperature of 25°C intermediate levels of expression are observed [45].

At 29°C, Myc levels are increased ∼9-fold using Act^TS compared to controls, and this increases the mutation frequency in adult females ∼2-fold after 5 days (Figure 3A, B). In addition, although we observe a higher proportion of point mutations in this genetic background than previously observed [29], Myc overexpression in adults increases the number of DSB-mediated genome rearrangements (Figure 3B). To assess whether overexpression of Myc in the adult affects lifespan, we used our Act^TS system to overexpress Myc in adult female flies and assessed their lifespan at 29°C where ∼9-fold induction of Myc is observed by Western analyses (Figure 3A). As shown in Figure 3C, Myc overexpression shortened median lifespan from 38 days in controls (Act^TS crossed to w¹¹¹⁸) to 22 days.

Because 29°C is considered a stress temperature for flies, we also examined the lifespan of flies overexpressing Myc using Act^TS at 25°C, at which temperature we see a ∼5-fold upregulation of Myc protein levels (Figure 4A). Like Myc overexpression at 29°C, at 25°C we observe that Myc increases lacZ mutation frequency 2-fold after 10 days of expression (Figure 4B). As seen in Figure 4C, Myc overexpression shortens lifespan at 25°C from a median of 58 days to 31 days. It is not known whether Myc overexpression in adult flies causes apoptosis, but based on its ability to cause cell death in diploid (but not polyploid) larval cell types, it is possible that this occurs and could contribute to our observed Myc-induced shortened lifespan. To resolve this concern, we co-expressed Myc with the cell death inhibitor p35 and compared their lifespan to flies expressing p35 alone using Act^TS. Overexpression of p35 has been previously shown to not significantly affect lifespan when expressed in the adult [46]. As shown in Figure 4D, flies co-expressing Myc and p35 have a longer median lifespan than flies expressing Myc alone, consistent with Myc-induced apoptosis in the adult influencing lifespan. However, flies co-expressing Myc and p35 still have a significantly shorter median lifespan than flies overexpressing p35 alone (50 days and 61 days, respectively). Thus, while Myc-induced cell death plays a part in the early death of Myc overexpressing flies, non-apoptotic functions also contribute. A good candidate for the early death of flies co-expressing Myc and p35 is an increased mutation frequency leading to cellular and tissue dysfunction. Importantly, at the non-permissive temperature of 18°C, we do not see any increase in Myc protein levels, nor do we observe any changes to lifespan (Figure S2).

Overexpression of Myc family proteins increases genomic instability [6], [9], [12], [24]–[26], [47]. While understanding the effects of overexpressed Myc has clear relevance for understanding diseases caused by dysregulation of Myc such as cancer, we also wish to test a role for endogenous Myc in influencing mutation load. The Drosophila Myc gene (myc, also known as diminutive and dm) is located on the X chromosome and null alleles are hemizygous lethal [31]. We therefore examined flies with reduced Myc levels using females heterozygous for the null allele myc⁴ and assessed their mutation load using our lacZ #2 reporter. myc⁴ heterozygotes have a 2-fold decrease in Myc protein levels and have a significantly reduced lacZ mutation load (Figure 5A, B).Endogenous Myc therefore influences mutation frequency. Consistent with our data linking Myc overexpression to DSB-mediated mutations, categorizing the lacZ mutations found in myc heterozygotes revealed a decrease in the incidence of genome rearrangements (Figure S3). Thus the frequency of DSB-mediated mutations is increased by Myc overexpression and decreased by reducing Myc.

Because we observed a decreased mutation load in myc⁴ heterozygous flies, we tested whether flies with reduced Myc levels also have an altered lifespan. We compared the lifespan of female flies heterozygous for lacZ #2 and those heterozygous for both myc⁴ and lacZ #2. As shown in Figure 5C, flies heterozygous for myc and lacZ #2 live longer than lacZ #2 heterozygotes alone, living a median lifespan of 67 and 59 days, respectively. To confirm that this effect is specific to reduced Myc function, we used a fly strain harboring a transgene in which a GFP tagged version of genomic Myc is expressed under the control of its endogenous promoter. We crossed this transgene into the myc⁴ null background and found that it is expressed at wildtype levels and rescues myc⁴-associated lethality, demonstrating that this transgene contains all the essential regulatory elements for myc expression (Figure 5A; data not shown). We then assessed the lacZ mutation load and lifespan of females flies heterozygous for myc⁴, lacZ #2 and the Myc:GFP rescue transgene. These flies have two copies of myc gene and are therefore expected to behave like wildtype flies. The decreased mutation load observed in myc⁴ heterozygotes is restored to wildtype levels by the addition of one copy of the myc:GFP transgene (Figure 5B). In addition, flies heterozygous for myc⁴, lacZ #2 and myc:GFP have a median lifespan indistinguishable from lacZ #2 heterozygous controls (58 days; Figure 5C). Lowering Myc levels therefore reduces mutation load and extends lifespan.