Skip to main content
PMC home page
eLife logoLink to eLife
. 2018 Feb 13;7:e32421. doi: 10.7554/eLife.32421

Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells

Kevin L Lu 1,2,3,†, Jonathan O Nelson 1,4,†, George J Watase 1,4, Natalie Warsinger-Pepe 1,5, Yukiko M Yamashita 1,2,4,6,✉
Editor: Allan C Spradling7
PMCID: PMC5811208  PMID: 29436367

Abstract

rDNA loci, composed of hundreds of tandemly duplicated arrays of rRNA genes, are known to be among the most unstable genetic elements due to their repetitive nature. rDNA instability underlies aging (replicative senescence) in yeast cells, however, its contribution to the aging of multicellular organisms is poorly understood. In this study, we investigate the dynamics of rDNA loci during aging in the Drosophila male germline stem cell (GSC) lineage, and show that rDNA copy number decreases during aging. Our study further reveals that this age-dependent decrease in rDNA copy number is heritable from generation to generation, yet GSCs in young animals that inherited reduced rDNA copy number are capable of recovering normal rDNA copy number. Based on these findings, we propose that rDNA loci are dynamic genetic elements, where rDNA copy number changes dynamically yet is maintained through a recovery mechanism in the germline.

Research organism: D. melanogaster

Introduction

The ribosomal DNA (rDNA) loci consist of tandem repetitive arrays of the rRNA genes, which code for the mature RNA components of ribosomes, flanked by intergenic spacer sequences (IGS) (Figure 1A). rDNA loci are considered to be one of the most unstable regions of the genome due to their repetitive nature and high transcriptional activity. First, as an inherent characteristic of repetitive DNA, rDNA can undergo intrachromatid recombination leading to copy number loss and generation of circularized repeat units (extrachromosomal rDNA circles (ERCs)) (Figure 1A) (Sinclair and Guarente, 1997). Second, the rDNA is highly transcribed even during S phase, leading to possible collisions between replication and transcription machineries. This collision can result in double strand breaks, further contributing to genomic instability of the rDNA (Helmrich et al., 2013; Takeuchi et al., 2003).

Figure 1. Drosophila male GSCs exhibit perturbations in nucleolar morphology with age.

Figure 1.

Open in a new tab

(A) Illustration of rDNA destabilization through intra-chromatid recombination. (B) Apical tip of the testes stained for Fibrillarin (red, nucleolus), Vasa (green, germ cells), Fas III (white, hub) and DAPI (blue). The hub, a major component of the GSC niche, is denoted by the asterisk. GSCs with representative nucleolar morphologies are outlined. Bar: 5 μm. (C) Distribution of GSC nucleolar morphology during aging, as a percentage of total GSCs scored (n, number of GSCs scored). P-values from chi-squared test (see methods) is shown.

Consistent with the notion that rDNA copy number is unstable, it is widely known that rDNA copy number varies considerably among individuals within a population. For example, it has been shown that Drosophila melanogaster exhibits 6-fold range of rDNA copy number (Lyckegaard and Clark, 1989; Lyckegaard and Clark, 1991), and another study estimated copy number range to be 80–600, where individuals with >130 copies are asymptomatic (Mohan and Ritossa, 1970). Similarly, rDNA copy number variation has been observed in many species including budding and fission yeasts (James et al., 2009; Maleszka and Clark-Walker, 1993; Pasero and Marilley, 1993), planktonic crustaceans Daphnia (Eagle and Crease, 2012), plants (barley [Zhang et al., 1990], bell bean [Rogers and Bendich, 1987]), mouse (Gibbons et al., 2015) and humans (Gibbons et al., 2015; Gibbons et al., 2014). Despite striking copy number variation among individuals, the range of copy number is well maintained within the population, suggestive of a mechanism that maintains rDNA copy number through generations. However, underlying molecular mechanisms responsible for maintenance of rDNA copy number are poorly understood.

In yeast, rDNA instability (i.e. reduction of rDNA copy number and associated accumulation of ERCs) is a major cause of replicative aging/senescence (Ganley and Kobayashi, 2014; Kobayashi, 2011; Sinclair and Guarente, 1997). The state of rDNA stability in various mutants that either stabilize or decrease rDNA copy number correlates well with lengthened or shortened life span, respectively (Ganley and Kobayashi, 2014). Despite clear involvement of the rDNA in the replicative lifespan of yeast, whether and how rDNA instability contributes to aging of multicellular organisms remains unclear. In particular, stem cells proliferate throughout the life of multicellular organisms, and their declining number and/or function during aging is proposed to be an underlying cause of organismal aging, due to inability to replenish essential cell populations (López-Otín et al., 2013). Despite the clear relationship between rDNA and aging, and that between aging and stem cells, little is known about whether or how the rDNA may change in stem cells during organismal aging.

In this study, we investigate the dynamics of the rDNA loci during aging in Drosophila male germline stem cells (GSCs). Drosophila male GSCs serve as an excellent experimental paradigm to study rDNA stability during aging, by providing a genetically tractable system to examine the aging of stem cells at a single cell resolution. GSCs undergo continuous asymmetric divisions throughout adulthood, producing one self-renewed GSC and one differentiating cell (Inaba and Yamashita, 2012). Additionally, Drosophila rDNA loci are limited to the sex chromosomes (X and Y), providing a simplified system to assess rDNA stability compared to other animal models that have multiple rDNA loci spread across many chromosomes (McStay, 2016). We show that male GSCs undergo destabilization of rDNA during aging. This rDNA destabilization manifests cytologically as atypical morphology of the nucleolus, the site of rDNA transcription (Boisvert et al., 2007; Pederson, 2011). We find that rDNA transcription is normally restricted to the Y chromosome in GSCs, as has been observed in other male Drosophila cell types (Greil and Ahmad, 2012; Zhou et al., 2012), but GSCs with atypical nucleolar morphology exhibit transcriptional activation of the normally-silent X rDNA locus. Our results indicate that X rDNA activation is likely to compensate for reduction in rDNA copy number that progresses during aging. We further show that such destabilization of rDNA is heritable, impacting the rDNA copy number of the germline and GSC nucleolar morphology in the next generation. Strikingly, we found that GSCs in the F1 generation are capable of recovering rDNA copy number in the early ages of adulthood, revealing the likely presence of a mechanism that maintains rDNA copy number through generations. We further show that this recovery requires the same factors needed for the phenomenon known as ‘rDNA magnification’, where rDNA copy number increases in the male germline in the animals with large rDNA deletions (Hawley and Tartof, 1983; Hawley and Tartof, 1985; Ritossa, 1968; Tartof, 1974). Taken together, we propose that the rDNA represents dynamic genetic loci that undergo degeneration and recovery during the aging of individuals and through generations. We further propose that rDNA copy number reduction during aging of parents and its recovery in the subsequent generation may contribute to widely observed copy number variation among individuals within species, which is nonetheless maintained within a certain range.

Results

Nucleolar morphology is perturbed during the aging of Drosophila male germline stem cells

To investigate the potential destabilization of rDNA in Drosophila male GSCs, we first examined nucleolar morphology during aging. The nucleolus is organized by transcription of the rDNA, thus its morphology is expected to reflect the transcriptional activity of rDNA loci (Boisvert et al., 2007; Pederson, 2011). Immunofluorescence staining of the nucleolar component Fibrillarin (Ochs et al., 1985) showed that most GSCs in young males contain a single, round nucleolus approximately 2 μm in diameter (89.2%, n = 408, Figure 1B, ‘typical’). However, we found that the frequency of GSCs with typical nucleolar morphology progressively decreased during aging throughout 40 days of adulthood (Figure 1C), as the GSCs with atypical nucleolar morphology increased. Atypical nucleolar morphology can be categorized into two types: (1) fragmented, where more than one distinct nucleolar foci were observed, and (2) deformed, where the nucleolus lost its typical compact, round morphology (Figure 1B, ‘fragmented’ and ‘deformed’). Although fragmented and deformed nucleoli are sometimes difficult to distinguish from each other and may indeed represent the same population of cells, we scored them separately to protect against losing information. However, these two populations behaved similarly in most assays performed in this study. These results show that nucleolar morphology is progressively perturbed during the aging of Drosophila male GSCs, prompting us to further investigate the underlying causes.

Perturbed nucleolar morphology is associated with transcriptional activation of the normally silent X chromosome rDNA locus

To investigate the underlying cause(s) of perturbed nucleolar morphology, we first examined the spatial relationship between the nucleolus and the rDNA loci that organize nucleolus formation. In Drosophila melanogaster, the rDNA loci are located on the sex chromosomes (X and Y), each containing ~200–250 copies of rDNA (Figure 2A) (Ritossa et al., 1971). The X and Y rDNA loci can be detected by DNA fluorescence in situ hybridization (FISH) using chromosome-specific probes against the (AATAAAC)n satellite and the 359 bp repeat satellite, which are adjacent to the Y and X rDNA loci, respectively (Figure 2A). By combining DNA FISH with immunofluorescence staining to detect nucleoli, we found that the Y rDNA locus is always associated with the nucleolus when GSCs have typical nucleolar morphology, whereas the X rDNA locus was not, irrespective of age (Figure 2B). Because the assembly of the nucleolus reflects rDNA transcription, these results suggest that Y rDNA is actively transcribed, whereas X rDNA is not. This likely reflects a phenomenon known as ‘nucleolar dominance’, where only certain rDNA loci (Y rDNA in the case of D. melanogaster males) are actively transcribed while the others (X rDNA in D. melanogaster males) are silent (see below) (Greil and Ahmad, 2012; Zhou et al., 2012).

Figure 2. Transcriptional activation of X rDNA in GSCs with atypical nucleolar morphology.

(A) Illustration of rDNA loci on Drosophila X and Y chromosomes. Y rDNA locus is juxtaposed to (AATAAAC)n satellite repeats (cyan), whereas X rDNA locus is juxtaposed to 359 bp satellite repeats (magenta). (B–D) DNA FISH for the 359 bp satellite repeats (magenta) and (AATAAAC)n satellite repeats (cyan), combined with immunofluorescence staining for Vasa (white), FasIII/Fibrillarin (green), DAPI (blue). B: typical nucleolus, C: fragmented nucleoli, D: deformed nucleolus. The hub is denoted by (*). GSCs with typical nucleoli are indicated by yellow dotted lines, GSCs with atypical nucleoli are indicated by red dotted lines. Arrows indicate the position of nucleoli. (E) SNP in situ hybridization with Y and X chromosome-specific rRNA probes combined with Nopp140-GFP to mark nucleolar morphology. Y rRNA (green), X rRNA (red), Nopp140-GFP (white), DAPI (blue). Bar: 7.5 μm. (F, G) SNP in situ hybridization in the testes from 0 day (F) and 40 day (G) old flies. GSCs with only Y rRNA (yellow outline) and with both X and Y rRNA transcription (red outline). Y rRNA (green), X rRNA (red), DAPI (blue). (H) XY rRNA transcription during aging of GSCs, as a percentage of total GSCs scored (n, number of GSCs scored). Mean ±SD (p-value of t-test is indicated). Note that ‘X-only’ rRNA transcription was never observed.

Figure 2.

Figure 2—figure supplement 1. SNP-FISH is highly specific for rRNA transcribed from the Y vs X chromosomes.

Figure 2—figure supplement 1.

(A) Schematic of the principle behind SNP-FISH for detecting rRNA transcripts. Shown is an example of a SNP in the 18S coding region of the X and Y rDNA. (B) Detection of X and Y rRNA using SNP-FISH in X/O (containing only X rDNA) and C(1)DX/Y (containing only Y rDNA) flies shows minimal cross-hybridization between probes. DAPI (blue), Y rRNA (green), X rRNA (red). Bars: 25 μm.
Open in a new tab

Interestingly, when the nucleolus is fragmented, the ectopic nucleolus (typically the smaller one) was almost always closely located near the X rDNA locus, irrespective of age (Figure 2C, n = 46/46, 100.0% in 0–1 day old flies, n = 46/47, 97.9% in 40 day old flies). X rDNA was also associated with deformed nucleoli (Figure 2D). These results suggest that the X chromosome has gained nucleolar organizing activity (i.e. became transcriptionally active) in GSCs when the nucleolar morphology is atypical.

To directly test the idea that atypical nucleolar morphology is associated with transcriptional activation of the normally inactive X rDNA locus, we adapted single nucleotide polymorphism (SNP) RNA in situ hybridization to differentially visualize the X and Y rDNA transcripts (Levesque et al., 2013) (Figure 2—figure supplement 1A). By genetically isolating and sequencing rDNA arrays from the X and Y chromosomes, we identified four SNPs in the coding and ITS sequences between the X and Y rDNA loci of the wild type strain used in this study (yw) (see methods). We designed probes (‘SNP probes’) utilizing these four SNPs to distinguish X- vs. Y-derived rRNA transcripts (Supplementary file 1). Specificity of SNP probes was confirmed by SNP in situ hybridization in X/O males (containing only X rDNA), and C(1)DX/Y females (containing only Y chromosome rDNA), where only the expected SNP signals were observed (Figure 2—figure supplement 1B).

We combined SNP in situ hybridization with the nucleolar marker Nopp140-GFP (McCain et al., 2006) to correlate X- vs. Y-derived rRNA transcription with nucleolar morphology. We found that most GSCs only transcribed Y rRNA (Figure 2E,F,H), demonstrating that nucleolar dominance indeed occurs in the male germline, as observed in male larval neuroblasts (Greil and Ahmad, 2012; Zhou et al., 2012). In all GSCs with fragmented nucleoli, one nucleolus (typically the larger one) showed a Y SNP signal, whereas the other showed an X SNP signal, supporting the idea that nucleolar fragmentation is associated with transcriptional activation of the X rDNA locus (Figure 2E,G, n > 40). As flies age, many GSCs exhibited expression of rRNA from both the Y and X chromosomes (Figure 2G,H). The number of GSCs expressing X rRNA increased from 18.2 ± 3.0% at 0 days old to 37.2 ± 2.4% at 40 days old (Figure 2H). These results, together with the above result that showed association of the X rDNA locus with fragmented/deformed nucleoli (Figure 2A–F), strongly suggest that atypical nucleolar morphology that accumulates in GSCs during aging is due to activation of the normally silent X rDNA locus, causing rDNA transcription from two separate chromosomes, each forming distinct nucleoli.

Y chromosome rDNA copy number decreases in the male germline during aging

Why does the X rDNA locus activate in aging GSCs? We hypothesized that rDNA copy number might be reduced during aging due to the inherent instability of the repetitive locus, requiring compensatory activation of a normally silent rDNA locus (i.e. X rDNA) to meet the cellular requirement for rRNA transcription. It has been extensively shown in yeast that stability of the rDNA is compromised during aging, with intrachromatid recombination leading to loss of rDNA on the chromosomes (Ganley et al., 2009; Ganley and Kobayashi, 2014; Kobayashi, 2008), leading us to hypothesize that a similar process may underlie the aging of Drosophila male GSCs.

To address the possibility that rDNA copy number might be decreased during aging in male GSCs, we first isolated genomic DNA from testes of young and old flies and quantified their rDNA copy number using a previously published qPCR-based method for quantifying rDNA copy number (Supplementary file 2) (Aldrich and Maggert, 2014), and found a significant reduction in rDNA copy number (Figure 3A). This copy number loss was observed across all the mature rRNA genes in the 45S cistron (18S, 5.8S and 28S rRNA genes). Interestingly, the copy number of R1 and R2 retrotransposable elements did not decrease with age (Figure 3A). R1 and R2 retrotransposable elements selectively insert into the 28S rRNA gene, and inserted rDNA units are transcriptionally repressed (Ye and Eickbush, 2006). These results suggest that the rDNA copy number loss primarily occurs in uninserted, actively-transcribed rDNA. These results imply that loss of rDNA copy number is associated with its transcriptional activity, consistent with the well-established notion that collision between transcription and replication machineries causes rDNA instability (Helmrich et al., 2013; Takeuchi et al., 2003).

Figure 3. rRNA gene copy number decreases in germ cells during aging.

(A) rRNA gene copy number quantification by qPCR from 0 day and 40 day old testes. Mean ±SD (p-value *≤0.05 t-test). (B–D) FISH on testis mitotic chromosome spreads. DAPI (white), 18S rDNA (red), 240 bp IGS (green), (AATAAAC)n (blue). Bar: 2.5 μm. Y chromosome, identified by the presence of (AATAAAC)n, is indicated by cyan outline, and X chromosome is indicated by yellow outline. (E) Y:X signal intensity ratio for the 18S rDNA in mitotic germ cells in day 0, day 40 old wild type (yw) testes, and X/Df(YS)bb testes. Bracket indicates mean ±SD. p-values from Student’s t-test is shown. (F) Y:X signal intensity ratio for the IGS in mitotic germ cells in day 0, day 40 old wild type (yw) testes, and X/Df(YS)bb testes. Bracket indicates mean ±SD. p-values from Student’s t-test is shown. Note that different Y:X ratios for 18S vs. IGS probes indicates that Y rDNA locus might have higher number of IGS repeats per rDNA unit. (G) Examples of GSCs with atypical nucleolar morphology from X/Df(YS)bb flies at 0 days (red outline). Fibrillarin (red), DAPI (blue), Vasa (green), FasIII (white). The hub is denoted by (*). Bar: 5 μm. (H) Distribution of GSC nucleolar morphologies in X/Df(YS)bb flies compared to 0 and 40-day-old WT flies, as a percentage of total GSCs scored (n, number of GSCs scored). Chi-squared test, p-values listed.

Figure 3—source data 1. Pixel intensity measurement and its ratio from DNA FISH plotted in Figure 3E,F.
elife-32421-fig3-data1.xlsx (54.2KB, xlsx)
DOI: 10.7554/eLife.32421.007

Figure 3.

Figure 3—figure supplement 1. The Df(YS)bb- from Kyoto Stock Center chromosome does not have any detectable 240-IGS sequence.

Figure 3—figure supplement 1.

(A) Chromosome spread from animals containing Df(YS)bb- chromosome shows a near complete loss of Y rDNA. The Y chromosome is marked with probes for the Y-specific AATAC (blue) and AATAAAC (red) satellite sequences. rDNA is marked by probes for the 240-IGS sequence (green). DAPI is shown in white. (A’) 240-IGS channel only.
Open in a new tab

Given that Y rDNA is predominantly transcribed in most young GSCs (Figure 2H), and that rDNA copies with retrotransposon insertion, which are known to be mostly silenced (Ye and Eickbush, 2006), are not lost during aging (Figure 3A), we hypothesized that the transcriptionally active Y rDNA copies are more frequently lost than normally silent X rDNA copies during aging. To assess this possibility, we used a quantitative DNA fluorescence in situ hybridization (FISH) method to examine changes in relative copy number of rDNA on X and Y chromosomes during aging (see methods). DNA FISH was performed on chromosome spreads from mitotic spermatogonia and meiotic spermatocytes (at stages when the X and Y chromosomes are not paired) using differentially-labelled 18S rDNA, 240-IGS and (AATAAAC)n probes, where (AATAAAC)n marked Y chromosomes. Then, relative fluorescence intensity of 18S rDNA and 240-IGS signals between X and Y chromosomes were determined (Figure 3B–F). By using this method, we found that Y:X ratio of the 18S rRNA gene was 1.35 ± 0.31 and that of the 240 bp repeat intergenic spacer (240-IGS) was 2.88 ± 0.60 in wild type (yw) at day 0, indicating that Y chromosome harbors slightly more rDNA than X chromosome in the wild type strain (yw) used in this study. To confirm the sensitivity of this quantitative DNA FISH method, we conducted DNA FISH on the chromosomes from animals with large deletions of Y rDNA. First, when DNA FISH was performed using a strain harboring a Y chromosome with a near complete loss of rDNA (Df(YS)bb-, obtained from Kyoto Stock Center, described in [Endow, 1982]) (Figure 3—figure supplement 1A–B), no signal was detected on Y chromosomes, suggesting that our FISH method has a very low background. Second, we were able to detect a reduction in rDNA of another Y chromosome harboring an rDNA deletion (Df(YS)bb, obtained from Bloomington Stock Center, described by Cline [2001]). When this Y chromosome was combined with the X chromosome of the wild type strain (yw), we detected that Df(YS)bb:X ratio of the 18S rDNA was reduced to 0.46 ± 0.08 from 1.35 ± 0.31 of the wild type (Figure 3E), and Df(YS)bb:X 240-IGS ratio was dropped to 0.42 ± 0.08 from 2.88 ± 0.60 of the wild type (Figure 3F), demonstrating that this method can detect partial deletion of rDNA. These results suggest that our DNA FISH method is sensitive enough to distinguish differences in the relative copy number of X and Y chromosome rDNA loci between different conditions, although it might not be fully quantitative.

By using this method, we compared Y:X rDNA ratio in day 0 vs. day 40 old testes. We found that the Y:X 18S rRNA gene ratio dropped from 1.35 ± 0.31 to 0.70 ± 0.25 by 40 days (Figure 3E) and that the Y:X 240-IGS ratio reduced from 2.88 ± 0.60 at 0 days to 1.86 ± 0.58 by 40 days (Figure 3F). Although the quantitative FISH method is only capable of detecting the ratio between X and Y, but not the absolute amount on each chromosome, the fact that overall germline rDNA copy number decreases during aging as demonstrated by qPCR (Figure 3A) suggests the change in Y:X ratio reflects Y rDNA loss, instead of X rDNA expansion. It should be noted that these data do not exclude the possibility that rDNA copies are also lost from the X chromosome. However, decreased Y:X ratio during aging suggests preferential loss of Y rDNA.

Since all adult germ cells are derived from GSCs, the loss of Y rDNA copies in the germline (detected by qPCR and quantitative FISH) suggests that Y rDNA copy number is reduced in GSCs. The loss of Y chromosome rDNA may lead to compensatory activation of X rDNA and the atypical nucleolar morphology observed during aging in GSCs. To directly address the causal relationship whether reduced rDNA copy number on the Y chromosome causes disrupted nucleolar morphology in GSCs, we examined GSC nucleolar morphology in X/Df(YS)bb flies, which harbors a partial deletion of Y rDNA. Even at a young age, GSCs in these flies exhibited atypical nucleolar morphology at a frequency comparable to 40-day-old wild type flies (Figure 3G,H). Taken together, we propose that rDNA copy number decreases during aging in male GSCs, which more profoundly occurs to the Y chromosome rDNA likely due to its transcriptionally active state. The reduction in rDNA copy number on the Y chromosome then leads to compensatory activation of X rDNA in GSCs, causing atypical nucleolar morphology.

GSC nucleolar morphology and rDNA loss is heritable

Because GSCs are the progenitors of gametes, we reasoned that rDNA copy number loss during the aging of the germline may be heritable to the next generation. To test this idea, we examined the F1 sons from old (day 40) parents (P0) (Figure 4A). qPCR on genomic DNA collected from testes of newly-eclosed F1 flies showed a significant reduction in the rDNA copy number when compared to young P0 flies (0d P0), but similar to 40 day old P0 (40d P0) (Figure 4B). Again, the number of R1- and R2-inserted rDNA copies was not significantly affected. These results suggest that old parents transmit reduced rDNA copy number to their offspring.

Figure 4. GSC nucleolar morphology and rDNA copy number decrease is heritable.

Figure 4.

Open in a new tab

(A) Scheme for aging of flies and collection of F1 progeny from old parents. (B) rDNA quantification from testes by qPCR in P0 at 0 and 40 days, and F1 at 0 days. Mean ±SD (p-value *≤0.05, t-test). (C) GSC nucleolar morphology in young P0, old P0 and young F1, as a percentage of total GSCs scored (n, number of GSCs scored). p-values from chi-squared test are shown. (D) Nucleolar dominance assessed by SNP in situ in GSCs from young P0, old P0 and young F1 (n, number of GSCs scored). Mean ±SD. p-value of t-test is shown. Note that ‘X-only’ rRNA transcription was never observed. (E) Y:X signal intensity ratio for the 18S rDNA and 240-IGS in mitotic germ cells comparing day 0 P0 Y and day 0 F1 Y (from day 40 father). Day 0 vs. day 40 fathers (P0) were mated to day 0 old females to yield P0 Y/X vs. F1 Y/X ratio, where X comes from the same source (day 0 yw female). Bracket indicates mean ±SD. p-values from Student’s t-test is shown. (F) Effect of X and Y chromosome inheritance from young vs. old parents on nucleolar morphology, as a percentage of total GSCs scored (n, number of GSCs scored). P-value from chi-squared test is shown.

Figure 4—source data 1. Pixel intensity measurement and its ratio from DNA FISH plotted in Figure 4E.
elife-32421-fig4-data1.xlsx (45.5KB, xlsx)
DOI: 10.7554/eLife.32421.009

Consistent with the reduced copy number of rDNA in F1 sons measured by qPCR, nucleolar morphology in F1 GSCs from old parents was also perturbed, with only 59.1% of GSCs displaying typical nucleolar morphology in 0-day-old F1 sons (0d F1), similar to 40-day-old P0 fathers (40d P0) (Figure 4C). Furthermore, SNP in situ hybridization demonstrated similar levels of activation of the X rDNA locus in 40-day-old P0 and 0-day-old F1 GSCs from the sons of old fathers, consistent with disrupted Y rDNA copy number in these flies (Figure 4D). Reduction in Y rDNA copy number in F1 sons was further confirmed by DNA FISH (Figure 4E): 0 day vs. 40-day-old P0 (father) was mated to young female and Y: X ratio of 18S and 240-IGS was determined in young (day 0) F1 sons. Because X chromosomes were inherited from young females in both cases, the Y: X ratio allowed direct comparison of Y rDNA in P0 vs. F1. These data clearly suggest reduction in rDNA copy number on the Y chromosome of F1 sons from old fathers.

Examining GSC nucleolar morphology in progeny from multiple genetic crosses provided insight into the dynamics of rDNA copy number on X and Y chromosomes in aged fathers. When 40 day old fathers (contributing Yold to their sons) were crossed to young mothers (contributing Xyoung to their sons), the newly-eclosed F1 sons (Xyoung/Yold) displayed disrupted GSC nucleolar morphology, suggesting that Y chromosomes from aged fathers are compromised (Figure 4F, column three compared to column 1). In contrast, when young fathers (contributing Yyoung) were crossed to 40 day old virgin mothers (contributing Xold), their offspring containing Xold and Yyoung chromosomes did not show perturbed nucleolar morphology (Figure 4F, column one vs. column 4), suggesting that male GSC nucleolar morphology is determined solely by the age of the inherited Y chromosome. This may be in part due to the dominance of the Y chromosome rDNA locus over the X rDNA locus, and does not necessarily reveal the state of the maternally inherited X chromosome, or whether the X chromosome undergoes degeneration during aging in females. However, the state of the X chromosome (Xyoung vs. Xold) passed from the mother appears to affect the degree of disruption of nucleolar morphology in the context of the Y chromosome from the old father (Yold) (Figure 4F, column 2 and column 3). This might suggest that the X chromosome rDNA might also undergo degeneration in the female germline during aging. Taken together, we conclude that rDNA copy number reduction is heritable and passed to the offspring from old fathers.

Germline rDNA recovers in the F1 generation

Although F1 sons from old fathers started with disrupted GSC nucleolar morphology, we unexpectedly found that nucleolar morphology recovered as these F1 sons age (Figure 5A). Interestingly, the recovery was specifically observed during the first 10 days after eclosion. At 10 days, F1 sons from old fathers recovered to the point where they were comparable to F1 sons of the same age from young fathers, after which nucleolar morphology in these two populations worsened at a similar rate (Figure 5A, compare to Figure 1D). Concomitant with the recovery of nucleolar morphology, X rDNA expression in F1 sons from old fathers diminished until 10 days, after which it again increased (Figure 5B,C). During this period, Y: X rDNA ratio also recovered (Figure 5D), revealing remarkable ability of the Y chromosome rDNA to expand in copy number. Together, the recovery of GSC nucleolar morphology and X rDNA repression, combined with the increase in Y:X rDNA ratio during this recovery period in germ cells, suggests that rDNA copies are expanded in GSCs to restore normal state of rDNA transcription (i.e. Y dominant). Interestingly, while we found many germ cells to have recovered Y:X rDNA ratio, there was also a large number of germ cells that apparently did not recover (Figure 5D), suggesting that rDNA expansion may be a stochastic event that occurs in individual GSCs.

Figure 5. Recovery of GSC nucleolar morphology, Y rDNA dominance and Y rDNA copy number in F1 flies.

Figure 5.

Open in a new tab

(A) Changes in GSC nucleolar morphology in F1 flies from old parents, as a percentage of total GSCs scored (n, number of GSCs scored). p-value from chi-squared test is shown. (B) Nucleolar dominance in GSCs from day 0 and day 10 old F1 testes assessed by SNP in situ hybridization. DAPI (blue), Y rRNA (green), X rRNA (red). The hub is denoted by (*). Bars: 5 µm. Co-dominant GSCs are indicated by red dotted lines, Y-dominant GSCs are indicated by yellow dotted lines. Arrows indicate X rRNA signal, thus co-dominance. (C) Nucleolar dominance in F1 GSCs during aging, as a percentage of total GSCs scored (n, number of GSCs scored). Mean ±SD. P-value of t-test is shown. Note that X-only rRNA transcription was never observed, except for once (out of 770 cells) at day 0, which is included in the graph. (D) Ratio of Y:X signal intensity for the 18S rDNA from mitotic chromosome spread of germ cells in F1 flies. Mean ±SD, t-test.

Figure 5—source data 1. Pixel intensity measurement and its ratio from DNA FISH plotted in Figure 5D.
elife-32421-fig5-data1.xlsx (59.9KB, xlsx)
DOI: 10.7554/eLife.32421.011

Recovery of GSC nucleolar morphology depends on the homologous recombination repair pathway

The observed recovery of GSC rDNA in the sons of aged fathers resembles the phenomenon called ‘rDNA magnification.’ Animals with rDNA insufficiency due to large deletions of rDNA develop a cuticular and bristle length defect called the bobbed phenotype (Ritossa et al., 1966). rDNA magnification is the observation that this defect is reverted to a wild type cuticle and bristle in a small subset of offspring from bobbed animals, due to the rapid expansion of rDNA copies (de Cicco and Glover, 1983; Ritossa, 1968). Although the molecular mechanisms of rDNA magnification are not fully understood (Bianciardi et al., 2012; Paredes and Maggert, 2009; Ritossa, 1968; Robbins, 1996), it has been shown that the genes involved in the homologous recombination repair are required for rDNA magnification. Specifically, mutations in mus-101 (Drosophila homolog of TOPBP1 (DNA topoisomerase 2-binding protein 1)) and mei-41 (Drosophila homolog of ATR), factors necessary for the resolution step of the homologous recombination-mediated repair of DNA double-strand breaks have severely reduced rates of rDNA magnification (Hawley and Tartof, 1983; Hawley and Tartof, 1985). Given the resemblance between rDNA magnification from large rDNA deletion and rDNA recovery in young sons from old fathers, we hypothesized that the same mechanism might underlie these phenomena. Indeed, we found that mus-101 and mei-41 mutants failed to recover typical nucleolar morphology in the GSCs of animals that inherited their Y chromosome from aged fathers (Figure 6A–C, compare to Figure 5A). Conversely, mutations in mus-102, which is required for DNA damage repair but not for rDNA magnification (Hawley and Tartof, 1985), had no effect on the recovery of GSC nucleolar morphology (Figure 6D). These results indicate that similar mechanisms might underlie these two phenomena, and that rDNA magnification may be the manifestation of the mechanisms that normally maintains rDNA copy number.

Figure 6. Recovery of GSC nucleolar morphology requires mus-101 and mei-41.

Figure 6.

Open in a new tab

(A) Mating scheme to assess the ability to recover nucleolar morphology after inheriting compromised Y chromosome from old fathers. yw males (0 or 40 days old) were mated to 0 day old mus-101D1, mei-41D12, or mus-102D1 mutant females. GSC nucleolar morphology in F1 mutant males was examined by anti-Fibrillarin antibody at day 0 or 10. (B–D) GSC nucleolar morphology in 0 and 10 day-old F1 mus-101D1 (B), mei-41D12 (C), and mus-102D1 (D) mutants, as a percentage of total GSCs scored (n, number of GSCs scored). P-values from chi-squared test between indicated conditions are shown.

mus-101 is required for normal germline rDNA maintenance during aging

Since we found that mei-41 and mus-101 are required for the recovery of GSC nucleolar morphology in F1 sons from old fathers, we wondered if these factors also contribute to the maintenance of rDNA during aging. However, wild type control (yw), mei-41, and mus-101 mutants exhibited unequal rates of atypical nucleolar morphology in GSCs at day 0, likely due to background differences in baseline X rDNA copy number (Figure 6B–C, first columns). This variation in initial nucleolar morphology made it difficult to directly compare changes in GSC rDNA content during aging between these genotypes. To control these background differences in rDNA copy number, we crossed a standard Y chromosome from our wild type strain (yw) into wild type X (Xyw) or mus-101D1 mutant background (note that mus-101 is an X-linked gene) (Figure 7A,F1). Then these Y chromosomes were allowed to age for 40 days in their respective genetic backgrounds. These Y chromosomes were isolated by crossing the F1 males to yw females, and the state of F1 Y chromosomes were assessed by nucleolar morphology of F2 GSCs (Figure 7A). This scheme allowed the comparison of Y chromosomes that have undergone aging in yw or mus-101D1 mutant background using the same source of X and Y chromosomes. This scheme effectively eliminated the effects of background variation, as we found that there was no difference in the fraction of GSCs with atypical nucleolar morphology between the sons of 0 day old yw and mus-101D1 mutants (Figure 7B, n.s (p=0.1647)). However, there was a significant difference between the sons of 40-day-old yw and mus-101D1mutant (Figure 7B, p=0.0070), suggesting that the mus-101 mutant fathers suffer more Y chromosome rDNA loss during aging compared to the yw fathers. This result suggests that the same molecular machineries might underlie normal germline rDNA maintenance and the phenomenon of rDNA magnification.

Figure 7. mus-101 is required for rDNA maintenance during aging.

Figure 7.

Open in a new tab

(A) Mating scheme to compare rDNA loss during aging between yw control and mus-101D1 mutant flies. Males with the same Y chromosome are mated to either yw or mus-101D1 females. mus-101D1 and yw F1 males are mated to young females with the same X chromosome at 0 and 40 days old. All F2 males have X and Y chromosomes from the same source, independent of paternal age or genotype. (B) GSC nucleolar morphology in the sons of 0 and 40-day-old yw and mus-101D1 F1, as a percentage of total GSCs scored (n, number of GSCs scored). P-values from chi-squared test between indicated conditions are shown.

Discussion

This study provides evidence that rDNA loci are highly unstable but actively maintained genetic loci. Our data shed light onto a few longstanding questions and also raise new questions for future investigation.

Destabilization of rDNA loci during aging in Drosophila male GSCs

Our data show that rDNA copy number decreases during aging of Drosophila male GSCs. Although early observations indicated that rDNA content may decrease during aging in mammals (Johnson and Strehler, 1972; Strehler, 1986), it was observed using bulk tissues containing mostly post-mitotic cells, and its implication in aging of multicellular organisms has been poorly explored. It was shown that in mouse hematopoietic stem cells, replicative stress is a major driver of stem cell aging (Flach et al., 2014). Curiously, they observed signs of replication stress (accumulation of γ-H2Ax) mainly in the nucleolus, but it remained unclear why the nucleolus specifically accumulates replication stress. We speculate that destabilization of rDNA loci may underlie age-associated accumulation of replication stress in the nucleolus. No matter how many rDNA copies are present in the genome, approximately, the same number of copies must be transcribed to support the cellular demands for ribosome biogenesis. This constant requirement for rDNA transcription means that those cells with fewer rDNA copies will have a larger proportion of their rDNA being transcribed than cells with more copies, as has been shown in yeast. Replication through actively transcribing rDNA creates the possibility for collision between replication and transcription machineries. Large rDNA arrays may be able to avoid such collision by selectively transcribing rDNA copies that are not undergoing DNA replication at the moment. However, smaller arrays would be limited in their ability to avoid collision due to their requirement to transcribe a larger portion of their rDNA. Increased collisions between replication and transcription machineries has been observed in a yeast strain with reduced rDNA copy number, leading to replication stress (Takeuchi et al., 2003). Consistent with the idea that transcription increases the probability of collision and thus destabilization, we observed that the dominantly-transcribed Y chromosome rDNA preferentially underwent destabilization.

Our results revealed a decrease in chromosomal rDNA copy number (predominantly on the Y chromosome) during aging. However, our study does not exclude the potential contribution of ERCs. In yeast, ERCs are specifically segregated to mother cells, whereas daughter cells are devoid of ERCs, potentially explaining the mechanism by which daughter cells reset their age (Shcheprova et al., 2008; Sinclair and Guarente, 1997). Although detection of ERCs in GSCs (compared to their daughter cells, gonialblasts) at a single cell resolution is not currently possible, it will be of future interest to investigate whether GSCs do accumulate ERCs and, if so, whether their inheritance is asymmetric.

We also found that uninserted copies are selectively lost during aging, as R1 and R2 retrotransposon abundance was maintained. Mathematical modeling by Zhou and Eickbush (Zhou et al., 2013) using D. simulans X rDNA locus suggested that transcription of rDNA preferentially occurs in contiguous blocks of uninserted copies, and rDNA copy loss preferentially occurs in such blocks. This model is consistent with our finding that uninserted copies are selectively lost, potentially providing mechanistic basis to our observations and suggesting that transcribed rDNA copies are preferentially destabilized.

Nucleolar dominance in Drosophila melanogaster

Nucleolar dominance is a phenomenon originally discovered in interspecific hybrids, wherein entire rDNA loci from one species are predominantly activated and those from the other species are silenced (Chen et al., 1998; Chen and Pikaard, 1997; Durica and Krider, 1977; Preuss and Pikaard, 2007). However, it was shown that nucleolar dominance also occurs within the D. melanogaster males, where Y chromosome rDNA is predominantly expressed, whereas X rDNA is silent (Greil and Ahmad, 2012; Zhou et al., 2012). It remains unknown whether the nucleolar dominance that occurs within a species vs. that in interspecific hybrids represent the same phenomena or share similar molecular mechanisms. Our study may shed light onto the significance of nucleolar dominance. This study reveals the preferential loss of rDNA copy number from the transcriptionally active Y rDNA locus, whereas the X remains silent. There are a few potential explanations why this might be advantageous: (1) by limiting the transcription of rDNA (thus potential DNA breaks) to one chromosome (i.e. Y chromosome), cells can reduce the risk of deleterious recombination events between the X and Y chromosomes, (2) by maintaining a stable chromosomal locus (X rDNA) for later use, stem cells might be able to extend their life span, delaying the timing of collapse and thus the overall aging. It is tempting to speculate that cells evolved nucleolar dominance to protect their rDNA loci, which are distributed among multiple chromosomes, from deleterious recombination.

It was reported that female neuroblasts exhibit co-dominance between two X chromosome rDNA loci (Greil and Ahmad, 2012). We were not able to assess the state of nucleolar dominance in female GSCs for several reasons: (1) after examining multiple D. melanogaster strains, we did not detect sufficient SNPs among different X chromosomes and thus we could not perform SNP in situ hybridization in females, (2) female GSCs barely showed atypical nucleolar morphology in young or old ovaries: however this could be attributed to constant pairing of two rDNA loci in female GSCs (Joyce et al., 2013), thus we could not rely on nucleolar morphology to infer the state of nucleolar dominance. Nonetheless, activating both rDNA loci on two X chromosomes might not impose as serious a risk as in male GSCs, as recombination between two X chromosomes would not lead to deleterious chromosomal rearrangements. Indeed, our inability to find SNPs among X chromosomes from many strains might reflect homogenization of rDNA sequences among X chromosomes through homologous recombination within the population.

In this study, we have adapted SNP in situ hybridization (Levesque et al., 2013) to assess nucleolar dominance. Previous studies on nucleolar dominance have had to rely on significant sequence differences such as those found in interspecific hybrids, and/or mitotic chromosome spreads (where active rDNA loci can be detected as secondary constrictions of the chromosome or histone H3.3 incorporation) (Chandrasekhara et al., 2016; Greil and Ahmad, 2012; Lawrence and Pikaard, 2004; McStay, 2006; McStay and Grummt, 2008). These approaches have limited the study of nucleolar dominance to hybrids or certain cell types. Our approach can open up the study of nucleolar dominance to a significantly broader range of species/cell types.

rDNA copy number maintenance through generations

We showed that rDNA copy number is heritable, wherein old fathers pass a Y chromosome with reduced rDNA copy number to their sons. However, our study revealed that rDNA copy number can recover in the next generation. Although it may be logically deduced that rDNA copy number should not continuously decrease from generation to generation without eventually resulting in complete collapse, the present study is the first to show that rDNA copy number is indeed actively maintained through generations. Our data provide a few important implications in the mechanism of rDNA copy number maintenance. First, it is of particular interest to note that the visible recovery in rDNA copy number was limited to young adults (~first 10 days of adulthood) (Figure 5). These results suggest that the rDNA recovery mechanism operates only under certain conditions. It remains unclear if such conditions are developmentally programmed or reflect the limitation of certain cell biological processes that underlie rDNA copy number recovery.

Our study also indicates that the phenomenon classically regarded as ‘rDNA magnification’ might be a manifestation of a general ‘maintenance’ mechanism that operates in the population that experiences normal fluctuations in rDNA copy number. We found that mutants that are known to be defective in rDNA magnification also exhibit signs of accelerated destabilization of rDNA during aging (e.g. atypical nucleolar morphology) and fail to restore rDNA in the subsequent generation. These findings suggest that the same molecular mechanisms might underlie rDNA magnification and maintenance.

It has been shown that rDNA copy number changes in response to nutrient conditions, and such copy number changes are inherited to the next generations (Aldrich and Maggert, 2015). It has been unknown how this inheritance is achieved. Our results on rDNA copy number changes in germline potentially provide explanation on how rDNA copy number changes are transmitted to the next generation. It awaits future investigation on how nutrient sensing operates in the germline to influence the rDNA copy number to be transmitted to the next generation.

Our findings reveal that tandem rDNA repeats are unstable in Drosophila male GSCs, similar to their well characterized instability in yeast, suggesting rDNA loss may occur in other metazoan stem cell populations. Although this instability in germ cells can cause the inheritance of reduced rDNA copies, the germline of young animals has the capacity to restore the lost rDNA copies. These findings suggest that the dynamic contraction and expansion of rDNA loci across generations normally maintains sufficient rDNA copies throughout a population.

Materials and methods

Key resources table.

Reagent type (species) or resource Designation Source or reference Identifiers
Strain, strain background (D. melanogaster) yw Bloomington Stock Center ID_BSC: 1495
strain, strain background (D. melanogaster) Nopp140-GFP PMCID: 16158326
strain, strain background (D. melanogaster) C(1)RM/C(1;Y)6, y1 w1 f1/0 Bloomington Stock Center ID_BSC: 9640
strain, strain background (D. melanogaster) FM6/C(1)DX, y* f1 Bloomington Stock Center ID_BSC: 784
strain, strain background (D. melanogaster) Df(YS)bb/w1sn1bb*/C(1)RM, y1v1f1 Bloomington Stock Center ID_BSC: 4491
strain, strain background (D. melanogaster) y[1] eq[1]/Df(YS)bb[-] Kyoto Stock Center DGRC#: 101–260
strain, strain background (D. melanogaster) w[1] mus-101[D1] Bloomington Stock Center ID_BSC: 2310
strain, strain background (D. melanogaster) w[1] mei-41[D12] Bloomington Stock Center ID_BSC: 6789
strain, strain background (D. melanogaster) w[1] mus-102[D1] Bloomington Stock Center ID_BSC: 2317
antibody anti-Fibrillarin Abcam ID_abcam: ab5821
antibody anti-Fibrillarin [38F3] Abcam ID_abcam: ab4566
antibody anti-H3K9 dimethyl Abcam ID_abcam: ab32521
antibody anti-vasa Santa Cruz Biotechnology ID_SCB: d-26
antibody anti-Adducin-like 1B1 Developmental Studies Hybridoma Bank
antibody anti-vasa Developmental Studies Hybridoma Bank
antibody anti-Fasciclin III Developmental Studies Hybridoma Bank
Open in a new tab

Fly husbandry and strains

All fly stocks were raised on standard Bloomington medium at 25°C. Unless otherwise stated, flies used for wild-type experiments were the standard lab wild-type strain yw (y1 w1). C(1)RM/C(1;Y)6, y1 w1 f1/0 (Bloomington Stock Center), FM6/C(1)DX, y* f1 (Bloomington Stock Center)(Novitski, 1954), Nopp140-GFP (McCain et al., 2006) (a gift of Pat DiMario, Louisiana State University), Df(YS)bb/w1sn1bb*/C(1)RM, y1v1f1 (Bloomington Stock Center), y1eq1/ Df(YS)bb- (Kyoto Stock Center). Note that we used two Y rDNA deletion chromosomes from different sources: we found Df(YS)bb- chromosome from Kyoto stock center has no detectable rDNA on the Y chromosome, whereas the one from Bloomington stock center has reduced rDNA copy number.

Immunofluorescence staining and microscopy

Immunofluorescence staining of testes was performed as described previously (Cheng et al., 2008). Briefly, testes were dissected in PBS, transferred to 4% formaldehyde in PBS and fixed for 30 min. Testes were then washed in PBS-T (PBS containing 0.1% Triton-X) for at least 60 min, followed by incubation with primary antibody in 3% bovine serum albumin (BSA) in PBS-T at 4°C overnight. Samples were washed for 60 min (three 20 min washes) in PBS-T, incubated with secondary antibody in 3% BSA in PBS-T at 4°C overnight, washed as above, and mounted in VECTASHIELD with DAPI (Vector Labs). The following primary antibodies were used: rat anti-vasa (1:20; DSHB; developed by A. Spradling), rabbit anti-vasa (1:200; d-26; Santa Cruz Biotechnology), mouse anti-Fasciclin III (1:200; DSHB; developed by C. Goodman), rabbit anti-Fibrillarin (1:200; Abcam ab5821), mouse anti-Fibrillarin (1:200; Abcam [38F3] ab4566). Images were taken using a Leica TCS SP8 confocal microscope with 63x oil-immersion objectives (NA = 1.4) and processed using Adobe Photoshop software.

DNA fluorescence in situ hybridization

Testes were prepared as described above, and optional immunofluorescence staining protocol was carried out first. Subsequently, fixed samples were incubated with 2 mg/ml RNase A solution at 37°C for 10 min, then washed with PBS-T +1 mM EDTA. Samples were washed in 2xSSC-T (2xSSC containing 0.1% Tween-20) containing increasing formamide concentrations (20%, 40%, then 50% formamide) for 15 min each. Hybridization buffer (50% formamide, 10% dextran sulfate, 2x SSC, 1 mM EDTA, 1 μM probe) was added to washed samples. Samples were denatured at 91°C for 2 min, then incubated overnight at 37°C. Probes used included Cy5-(AATAAAC)6 for detection of the Y chromosome and Cy-3-CCACATTTTGCAAATTTTGATGACCCCCCTCCTTACAAAAAATGCG (a part of 359 bp repeats) for detection of the X chromosome.

In scoring, association of FISH signals (either 359 bp repeat next to the X rDNA or (AATAAAC)n repeat next to the Y rDNA) with the nucleolus, FISH signal was typically found in the direct proximity to the nucleolus or within the distance smaller than the diameter of FISH signal itself, which was typically less than 0.5 µm. The ‘non-associated’ FISH signal was far away from the nucleolus. Thus, the distance between associated FISH signal and the nucleolus was always clearly smaller than that between non-associated FISH signal and the nucleolus.

Determination of X and Y chromosome SNPs

The X chromosome was isolated by crossing experimental XY males with C(1)RM females, generating X/O males lacking the Y chromosome (and the Y rDNA). The Y chromosome rDNA was isolated by crossing experimental XY males with C(1)DX/Y females, which generated C(1)DX/Y females containing our experimental Y and no rDNA on the compound X chromosome. 45S rRNA genes were sequenced using the following primers to identify single nucleotide variants between the two consensus sequences. ITS region: 5’-CTTGCGTGTTACGGTTGTTTC-3’ (forward) and 5’- ACAGCATGGACTGCGATATG-3’ (reverse). 18S region: 5’-GAAACGGCTACCACATCTAAGG-3’ (forward) and 5’- GGACCTCTCGGTCTAGGAAATA-3’ (reverse). 28S region: 5’- AGCCCGATGAACCTGAATATC-3’ (forward) and 5’- CATGCTCTTCTAGCCCATCTAC-3’ (reverse). Sequence alignment was done using ClustalW2.

SNP RNA in situ hybridization

For SNP RNA in situ hybridization, all solutions used were RNase-free. Testes were collected in PBS and fixed in 4% formaldehyde in PBS for 30 min. Then testes were washed briefly in PBS, and permeabilized in 70% ethanol overnight at 4°C. Following overnight permeabilization, testes were briefly rinsed in 2xSSC with 10% formamide. Hybridization buffer (prepared according to protocol by LGB Biosearch for Stellaris probes) was prepared with probe (50 nM final concentration) and incubated overnight at 37°C. Following hybridization, samples were washed twice in 2x SSC with 10% formamide for 30 min each and mounted in VECTASHIELD with DAPI (Vector Labs).

Final concentration of each SNP probe was 100 nM, and each mask oligo was 300 nM. Sequences of SNP probes and oligos are provided in Supplementary file 1.

qPCR

Quantitative PCR was carried out using cycling conditions previously described (Aldrich and Maggert, 2014) and Power SYBR Green reagent (Applied Biosystems). All numbers were normalized to tRNA-K-CTT, a multicopy tRNA gene known to be interspersed throughout the genome, and GAPDH. Primers used are listed in Supplementary file 2:

Mitotic chromosome spreads and fluorescence quantification

Testes were squashed according to previously described methods (Larracuente and Ferree, 2015). Briefly, testes were dissected into 0.5% sodium citrate for 5–10 min and fixed in 45% acetic acid/2.2% formaldehyde for 4–5 min. Fixed tissues were firmly squashed with a cover slip then slides were submerged in liquid nitrogen. Following liquid nitrogen, slides were dehydrated in 100% ethanol for at least 5 min. Slides were then treated with 0.1 μg/ml RNase A for 1 hr at room temperature, then dehydrated in 100% ethanol again. Hybridization mix (50% formamide, 2x SSC, 10% dextran sulfate) with 100 ng each probe was applied directly to the slide and allowed to hybridize overnight at room temperature. Then slides were washed 3x for 15 min in 0.2x SSC, and mounted with VECTASHIELD with DAPI (Vector Labs). Sequences for probes used are listed in Supplementary file 3.

Images were taken using Leica SP8 confocal microscope, using the setting to detect saturation to ensure that acquired signals were not saturated. Fluorescence quantification was done on merged z-stacks using ImageJ using the Maximum Entropy plugin for automatic thresholding based on the histogram to automatically determine real signal from noise. Using this method, fluorescent probe signal was measured as Integrated Density and compared between the X and Y chromosomes.

Statistical analysis

For comparison of nucleolar morphologies, significance was determined by chi-squared test using a 2 × 3 contingency table (Typical; Deformed; Fragmented). For X rDNA activation by SNP-FISH, because X-only transcription was virtually never detected we simplified the comparison to Y-only rRNA vs both X and Y-rRNA and performed Student’s t-tests.

Acknowledgements

We thank Pat DiMario, Bloomington Drosophila Stock Center, Kyoto Drosophila Stock Center and Developmental Studies Hybridoma Bank for reagents. We thank the Yamashita lab members, Sue Hammoud and Lei Lei for discussion and comments on the manuscript, and Craig Pikaard for consultation. This research was supported by Howard Hughes Medical Institute. Kevin Lu is supported by National Institute of Aging (F30 AG050398-01A1). Natalie Warsinger-Pepe is supported by NIH Career Training in Reproductive Biology (5T32HD079342-04).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Yukiko M Yamashita, Email: yukikomy@umich.edu.

Allan C Spradling, Howard Hughes Medical Institute, Carnegie Institution for Science, United States.

Funding Information

This paper was supported by the following grants:

  • Howard Hughes Medical Institute to Yukiko M Yamashita.

  • National Institute of General Medical Sciences to Yukiko M Yamashita.

  • National Institute on Aging to Kevin L Lu.

  • National Institute of Child Health and Human Development to Natalie Warsinger-Pepe.

Additional information

Competing interests

Reviewing editor, eLife.

No competing interests declared.

Author contributions

Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing—original draft, Writing—review and editing.

Formal analysis, Validation, Investigation, Writing—original draft, Writing—review and editing.

Investigation, Contributed the data (Figure 3—figure supplement 1) and edited the manuscript.

Formal analysis, Validation, Writing—review and editing.

Conceptualization, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing.

Additional files

Supplementary file 1. Probe sequences for rRNA SNP in situ hybridization.
elife-32421-supp1.docx (69.5KB, docx)
DOI: 10.7554/eLife.32421.014
Supplementary file 2. Primer sequences for genomic rDNA qPCR.
elife-32421-supp2.docx (73.6KB, docx)
DOI: 10.7554/eLife.32421.015
Supplementary file 3. Probe sequences for rDNA DNA FISH.
elife-32421-supp3.docx (96.3KB, docx)
DOI: 10.7554/eLife.32421.016
Transparent reporting form
elife-32421-transrepform.docx (244.8KB, docx)
DOI: 10.7554/eLife.32421.017

References

  1. Aldrich JC, Maggert KA. Simple quantitative PCR approach to reveal naturally occurring and mutation-induced repetitive sequence variation on the Drosophila Y chromosome. PLoS One. 2014;9:e109906. doi: 10.1371/journal.pone.0109906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Aldrich JC, Maggert KA. Transgenerational inheritance of diet-induced genome rearrangements in Drosophila. PLOS Genetics. 2015;11:e1005148. doi: 10.1371/journal.pgen.1005148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bianciardi A, Boschi M, Swanson EE, Belloni M, Robbins LG. Ribosomal DNA organization before and after magnification in Drosophila melanogaster. Genetics. 2012;191:703–723. doi: 10.1534/genetics.112.140335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Boisvert FM, van Koningsbruggen S, Navascués J, Lamond AI. The multifunctional nucleolus. Nature Reviews Molecular Cell Biology. 2007;8:574–585. doi: 10.1038/nrm2184. [DOI] [PubMed] [Google Scholar]
  5. Chandrasekhara C, Mohannath G, Blevins T, Pontvianne F, Pikaard CS. Chromosome-specific NOR inactivation explains selective rRNA gene silencing and dosage control in Arabidopsis. Genes & Development. 2016;30:177–190. doi: 10.1101/gad.273755.115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen ZJ, Comai L, Pikaard CS. Gene dosage and stochastic effects determine the severity and direction of uniparental ribosomal RNA gene silencing (nucleolar dominance) in Arabidopsis allopolyploids. PNAS. 1998;95:14891–14896. doi: 10.1073/pnas.95.25.14891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chen ZJ, Pikaard CS. Epigenetic silencing of RNA polymerase I transcription: a role for DNA methylation and histone modification in nucleolar dominance. Genes & Development. 1997;11:2124–2136. doi: 10.1101/gad.11.16.2124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cheng J, Türkel N, Hemati N, Fuller MT, Hunt AJ, Yamashita YM. Centrosome misorientation reduces stem cell division during ageing. Nature. 2008;456:599–604. doi: 10.1038/nature07386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cline TW. A lesson from flex: consider the Y chromosome when assessing Drosophila sex-specific lethals. Development. 2001;128:1015–1018. doi: 10.1242/dev.128.6.1015. [DOI] [PubMed] [Google Scholar]
  10. de Cicco DV, Glover DM. Amplification of rDNA and type I sequences in Drosophila males deficient in rDNA. Cell. 1983;32:1217–1225. doi: 10.1016/0092-8674(83)90304-5. [DOI] [PubMed] [Google Scholar]
  11. Durica DS, Krider HM. Studies on the ribosomal RNA cistrons in interspecific Drosophila hybrids. I. Nucleolar dominance. Developmental Biology. 1977;59:62–74. doi: 10.1016/0012-1606(77)90240-8. [DOI] [PubMed] [Google Scholar]
  12. Eagle SH, Crease TJ. Copy number variation of ribosomal DNA and Pokey transposons in natural populations of Daphnia. Mobile DNA. 2012;3:4. doi: 10.1186/1759-8753-3-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Endow SA. Molecular characterization of ribosomal genes on the Ybb- chromosome of Drosophila melanogaster. Genetics. 1982;102:91–99. doi: 10.1093/genetics/102.1.91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Flach J, Bakker ST, Mohrin M, Conroy PC, Pietras EM, Reynaud D, Alvarez S, Diolaiti ME, Ugarte F, Forsberg EC, Le Beau MM, Stohr BA, Méndez J, Morrison CG, Passegué E. Replication stress is a potent driver of functional decline in ageing haematopoietic stem cells. Nature. 2014;512:198–202. doi: 10.1038/nature13619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ganley AR, Ide S, Saka K, Kobayashi T. The effect of replication initiation on gene amplification in the rDNA and its relationship to aging. Molecular Cell. 2009;35:683–693. doi: 10.1016/j.molcel.2009.07.012. [DOI] [PubMed] [Google Scholar]
  16. Ganley AR, Kobayashi T. Ribosomal DNA and cellular senescence: new evidence supporting the connection between rDNA and aging. FEMS Yeast Research. 2014;14:49–59. doi: 10.1111/1567-1364.12133. [DOI] [PubMed] [Google Scholar]
  17. Gibbons JG, Branco AT, Godinho SA, Yu S, Lemos B. Concerted copy number variation balances ribosomal DNA dosage in human and mouse genomes. PNAS. 2015;112:2485–2490. doi: 10.1073/pnas.1416878112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gibbons JG, Branco AT, Yu S, Lemos B. Ribosomal DNA copy number is coupled with gene expression variation and mitochondrial abundance in humans. Nature Communications. 2014;5:4850. doi: 10.1038/ncomms5850. [DOI] [PubMed] [Google Scholar]
  19. Greil F, Ahmad K. Nucleolar dominance of the Y chromosome in Drosophila melanogaster. Genetics. 2012;191:1119–1128. doi: 10.1534/genetics.112.141242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hawley RS, Tartof KD. The effect of mei-41 on rDNA redundancy in Drosophila melanogaster. Genetics. 1983;104:63–80. doi: 10.1093/genetics/104.1.63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hawley RS, Tartof KD. A two-stage model for the control of rDNA magnification. Genetics. 1985;109:691–700. doi: 10.1093/genetics/109.4.691. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Helmrich A, Ballarino M, Nudler E, Tora L. Transcription-replication encounters, consequences and genomic instability. Nature Structural & Molecular Biology. 2013;20:412–418. doi: 10.1038/nsmb.2543. [DOI] [PubMed] [Google Scholar]
  23. Inaba M, Yamashita YM. Asymmetric stem cell division: precision for robustness. Cell Stem Cell. 2012;11:461–469. doi: 10.1016/j.stem.2012.09.003. [DOI] [PubMed] [Google Scholar]
  24. James SA, O'Kelly MJ, Carter DM, Davey RP, van Oudenaarden A, Roberts IN. Repetitive sequence variation and dynamics in the ribosomal DNA array of Saccharomyces cerevisiae as revealed by whole-genome resequencing. Genome Research. 2009;19:626–635. doi: 10.1101/gr.084517.108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Johnson R, Strehler BL. Loss of genes coding for ribosomal RNA in ageing brain cells. Nature. 1972;240:412–414. doi: 10.1038/240412a0. [DOI] [PubMed] [Google Scholar]
  26. Joyce EF, Apostolopoulos N, Beliveau BJ, Wu CT. Germline progenitors escape the widespread phenomenon of homolog pairing during Drosophila development. PLoS Genetics. 2013;9:e1004013. doi: 10.1371/journal.pgen.1004013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kobayashi T. A new role of the rDNA and nucleolus in the nucleus--rDNA instability maintains genome integrity. BioEssays. 2008;30:267–272. doi: 10.1002/bies.20723. [DOI] [PubMed] [Google Scholar]
  28. Kobayashi T. How does genome instability affect lifespan?: roles of rDNA and telomeres. Genes to Cells : Devoted to Molecular & Cellular Mechanisms. 2011;16:617–624. doi: 10.1111/j.1365-2443.2011.01519.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Larracuente AM, Ferree PM. Simple method for fluorescence DNA in situ hybridization to squashed chromosomes. Journal of Visualized Experiments. 2015;95:52288. doi: 10.3791/52288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lawrence RJ, Pikaard CS. Chromatin turn ons and turn offs of ribosomal RNA genes. Cell Cycle. 2004;3:878–881. doi: 10.4161/cc.3.7.983. [DOI] [PubMed] [Google Scholar]
  31. Levesque MJ, Ginart P, Wei Y, Raj A. Visualizing SNVs to quantify allele-specific expression in single cells. Nature Methods. 2013;10:865–867. doi: 10.1038/nmeth.2589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging. Cell. 2013;153:1194–1217. doi: 10.1016/j.cell.2013.05.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lyckegaard EM, Clark AG. Ribosomal DNA and Stellate gene copy number variation on the Y chromosome of Drosophila melanogaster. PNAS. 1989;86:1944–1948. doi: 10.1073/pnas.86.6.1944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lyckegaard EM, Clark AG. Evolution of ribosomal RNA gene copy number on the sex chromosomes of Drosophila melanogaster. Molecular Biology and Evolution. 1991;8:458–474. doi: 10.1093/oxfordjournals.molbev.a040664. [DOI] [PubMed] [Google Scholar]
  35. Maleszka R, Clark-Walker GD. Yeasts have a four-fold variation in ribosomal DNA copy number. Yeast. 1993;9:53–58. doi: 10.1002/yea.320090107. [DOI] [PubMed] [Google Scholar]
  36. McCain J, Danzy L, Hamdi A, Dellafosse O, DiMario P. Tracking nucleolar dynamics with GFP-Nopp140 during Drosophila oogenesis and embryogenesis. Cell and Tissue Research. 2006;323:105–115. doi: 10.1007/s00441-005-0044-9. [DOI] [PubMed] [Google Scholar]
  37. McStay B, Grummt I. The epigenetics of rRNA genes: from molecular to chromosome biology. Annual Review of Cell and Developmental Biology. 2008;24:131–157. doi: 10.1146/annurev.cellbio.24.110707.175259. [DOI] [PubMed] [Google Scholar]
  38. McStay B. Nucleolar dominance: a model for rRNA gene silencing. Genes & Development. 2006;20:1207–1214. doi: 10.1101/gad.1436906. [DOI] [PubMed] [Google Scholar]
  39. McStay B. Nucleolar organizer regions: genomic 'dark matter' requiring illumination. Genes & Development. 2016;30:1598–1610. doi: 10.1101/gad.283838.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mohan J, Ritossa FM. Regulation of ribosomal RNA synthesis and its bearing on the bobbed phenotype in Drosophila melanogaster. Developmental Biology. 1970;22:495–512. doi: 10.1016/0012-1606(70)90165-X. [DOI] [PubMed] [Google Scholar]
  41. Novitski E. The compound X chromosomes in Drosophila. Genetics. 1954;39:127–140. doi: 10.1093/genetics/39.1.127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Ochs RL, Lischwe MA, Spohn WH, Busch H. Fibrillarin: a new protein of the nucleolus identified by autoimmune sera. Biology of the Cell. 1985;54:123–133. doi: 10.1111/j.1768-322X.1985.tb00387.x. [DOI] [PubMed] [Google Scholar]
  43. Paredes S, Maggert KA. Expression of I-CreI endonuclease generates deletions within the rDNA of Drosophila. Genetics. 2009;181:1661–1671. doi: 10.1534/genetics.108.099093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Pasero P, Marilley M. Size variation of rDNA clusters in the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe. Molecular & General Genetics : MGG. 1993;236:448–452. doi: 10.1007/BF00277147. [DOI] [PubMed] [Google Scholar]
  45. Pederson T. The nucleolus. Cold Spring Harbor Perspectives in Biology. 2011;3:a000638. doi: 10.1101/cshperspect.a000638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Preuss S, Pikaard CS. rRNA gene silencing and nucleolar dominance: insights into a chromosome-scale epigenetic on/off switch. Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression. 2007;1769:383–392. doi: 10.1016/j.bbaexp.2007.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Ritossa F, Malva C, Boncinelli E, Graziani F, Polito L. The first steps of magnification of DNA complementary to ribosomal RNA in Drosophila malanogaster. PNAS. 1971;68:1580–1584. doi: 10.1073/pnas.68.7.1580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Ritossa FM, Atwood KC, Spiegelman S. A molecular explanation of the bobbed mutants of Drosophila as partial deficiencies of "ribosomal" DNA. Genetics. 1966;54:819–834. doi: 10.1093/genetics/54.3.819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Ritossa FM. Unstable redundancy of genes for ribosomal RNA. PNAS. 1968;60:509–516. doi: 10.1073/pnas.60.2.509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Robbins LG. Specificity of chromosome damage caused by the Rex element of Drosophila melanogaster. Genetics. 1996;144:109–115. doi: 10.1093/genetics/144.1.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Rogers SO, Bendich AJ. Heritability and variability in ribosomal RNA genes of Vicia faba. Genetics. 1987;117:285–295. doi: 10.1093/genetics/117.2.285. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Shcheprova Z, Baldi S, Frei SB, Gonnet G, Barral Y. A mechanism for asymmetric segregation of age during yeast budding. Nature. 2008;454:728–734. doi: 10.1038/nature07212. [DOI] [PubMed] [Google Scholar]
  53. Sinclair DA, Guarente L. Extrachromosomal rDNA circles--a cause of aging in yeast. Cell. 1997;91:1033–1042. doi: 10.1016/S0092-8674(00)80493-6. [DOI] [PubMed] [Google Scholar]
  54. Strehler BL. Genetic instability as the primary cause of human aging. Experimental Gerontology. 1986;21:283–319. doi: 10.1016/0531-5565(86)90038-0. [DOI] [PubMed] [Google Scholar]
  55. Takeuchi Y, Horiuchi T, Kobayashi T. Transcription-dependent recombination and the role of fork collision in yeast rDNA. Genes & Development. 2003;17:1497–1506. doi: 10.1101/gad.1085403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Tartof KD. Unequal mitotic sister chromatin exchange as the mechanism of ribosomal RNA gene magnification. PNAS. 1974;71:1272–1276. doi: 10.1073/pnas.71.4.1272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Ye J, Eickbush TH. Chromatin structure and transcription of the R1- and R2-inserted rRNA genes of Drosophila melanogaster. Molecular and Cellular Biology. 2006;26:8781–8790. doi: 10.1128/MCB.01409-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Zhang QF, Saghai Maroof MA, Allard RW. Effects on adaptedness of variations in ribosomal DNA copy number in populations of wild barley (Hordeum vulgare ssp. spontaneum) PNAS. 1990;87:8741–8745. doi: 10.1073/pnas.87.22.8741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zhou J, Eickbush MT, Eickbush TH. A population genetic model for the maintenance of R2 retrotransposons in rRNA gene loci. PLoS Genetics. 2013;9:e1003179. doi: 10.1371/journal.pgen.1003179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zhou J, Sackton TB, Martinsen L, Lemos B, Eickbush TH, Hartl DL. Y chromosome mediates ribosomal DNA silencing and modulates the chromatin state in Drosophila. PNAS. 2012;109:9941–9946. doi: 10.1073/pnas.1207367109. [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision letter

Editor: Allan C Spradling1

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors and the evaluation has been overseen by K VijayRaghavan as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Kami Ahmad (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

Loss of rDNA in aging yeast cells and the presence of damaged rRNA gene copies in many genomes testifies to the challenges of keeping highly active repeated rDNA genes in good working order. Drosophila rDNAs are a complex mix of active and inserted genes located on both X or Y chromosomes whose functional gene number varies between strains, with severe deficiencies causing a "bobbed" phenotype. rDNA homogeneity and copy number are maintained at least in part by unequal crossing over. Under certain conditions, rDNA-deficient male germ cells can restore copy number by a "magnification" process whose mechanism and regulation remain poorly understood. In the manuscript, the authors analyze rDNA expression in male germ cells using imaging, FISH, and qPCR to document Y chromosome nucleolar dominance that weakens with age in association with reductions in rDNA copy number. Progeny of old males inherit reduced Y chromosome rDNA copy number but, restore it to normal levels based on cytological assays during the first 10 days of development by an unknown process the authors argue may be magnification.

The work is interesting and if fully validated would represent a significant advance, but at present, the authors' claims are inadequately documented.

Essential revisions:

1) A major problem is that the changes in Y chromosome rDNA copy number are not simply and directly documented. The authors quantify rDNA by Q-PCR in genomic DNA purified from whole testes. Testes contain germ cells, meiocytes, GSCs, gonadal mesoderm (including polyploid cells), etc. This is a very insensitive way to quantify the amount of Y-linked rDNA in GSCs, given the other cell types and the X-linked array signals. Although they attempt a second method to confirm the results, the authors present no data that their allele-specific FISH are quantitative. To critically test the central issue or rDNA copy number changes, the authors should cross young X/Y males with a full complement of rDNA as well as increasingly older males that seem to be losing Y-linked rDNA genes, to rDNA deficient C(1)DX/Y,B females or Df(1)sc4sc8/FM7 females and measure the rDNA levels in young C(1)DX/Y or Df(1)sc4sc8/Y offspring containing only Y-linked rDNA genes. They should observe if these animals begin to show a bobbed phenotype and progressively less rDNA as the age of the X/Y father progressively increases. The same test of Y chromosome rDNA copy number should be done using F1 male offspring of an old father, of various ages. In this case, the exclusively Y chromosome rDNA should be deficient when the F1 male was young but should return to normal levels during the 1st 10 days of F1 adult life.

2) With respect to the FISH assays designed to identify X from Y-linked nucleoli, the criteria used to determine "association" must be described (on and around the second paragraph of subsection “Perturbed nucleolar morphology is associated with transcriptional activation of the normally silent X chromosome rDNA locus.”), especially because the few images presented in Figure 2 show a variance in locations of FISH, fibrillarin staining, etc. This is a concern because the active rDNA are thought to loop into the FC of the nucleolus, while the interspersed inactive cistrons are at the FC/DFC border, and linked chromosome material (including the heterochromatin) may not be found in the nucleolus. Probes to neighboring DNAs (e.g., the 359 satellite) may be some distance away from active cistrons, so proximity is not an adequate indicator of transcriptional activity.

3) The relational of this work to magnification is currently exaggerated and needs to be accurately discussed. First, a summary of previous work on rDNA magnification including recent work should be given in the introduction. Second, the authors must recognize that there are no known "rDNA magnification" genes exclusively involved in this process. Hence, conclusions such as: "This result indicates that rDNA magnification contributes to germline rDNA maintenance during normal aging" are unjustified. Third, the authors should specifically include a discussion reconciling the differences in frequencies of magnification observed in previous studies (>1%) from those reported here (100%). Fourth, the authors should include a discussion of the previous observations that only some rDNA array-containing chromosomes are subject to magnification (and only in some genetic conditions), and how that would bear on their model.

4) It seems to be a misnomer to describe the changes in nucleolar appearance as aberrant nucleolar morphology. The FISH mapping shows that these changes are not a nucleolar defect but are simply the activation of two loci.

5) The authors do not address the very interesting question of whether the nucleolar phenomena and rDNA copy number changes they observed are confined to germ cells. Is there nucleolar dominance, loss of genes with aging, recovery in F1 adults in somatic hub cells and cyst stem cells, like in GSCs? In particular, does the reduced level of Y rDNA inherited from aged fathers persist in the soma of F1 animals, even though rDNA copy number is restored in the germline?

6) Figure 5: Can't see the X and Y signals in panel B, F1 10d.

7) Subsection “Destabilization of rDNA loci during aging in Drosophila male GSCs.” Why would activation of additional rDNA genes to compensate for lost genes, leaving the total number of active genes intact, "increase the chance of replication transcription collisions and generate a vicious cycle? The argument is not clear here.

8) The structure of the Y and X chromosome rDNAs is not fully described and affect some of the inferences made in this manuscript. The authors describe loss of 45S cistron copies with age, but no loss of R1/R2 retroposons. Are there R1/R2 insertions on the Y chromosome (implying preferential loss), or are all the copies on the X chromosome?

9) The data in Figure 3 and the fourth paragraph of subsection “Y chromosome rDNA copy number decreases in the male germline during aging.” details the loss of rDNA cistrons by qPCR. It would be worth converting these numbers into assignments for each chromosome, i.e. in a wildtype male 34% of 18S sequence is on the X and 66% on the Y. Then, after 40 days, there is (if the X doesn't change) 34% on the X and 18% remaining on the Y (loss of two thirds of the locus).

9) Multiple graphs are interrupted to emphasize the small changes in nucleolar morphology frequencies. This may overstate the size of the effect. There seems to be some discordance of the measurements of rDNA loss by qPCR and the more subtle changes in nucleolar appearance. Is this possibly due to a threshold at which nucleolar dominance is lost?

10) The age effect described predicts that there would be loss or rDNA cistrons from the X chromosome when combined with a Df(YS)bb; is this the case? Overall, more details on the structure of the "partially deleted" Df(YS)bb would be useful.

11) The authors under-explain the results by Ahmad and Hartl and Eickbush, whose groups all saw that not all chromosomes or genetic background manifest nucleolar dominance. Further, they do not mention that in plants nucleolar dominance is a transient phenomenon, often lost after the first week of development. These findings may bear on the authors' interpretation and certainly bear on the "null hypotheses" that loss of Y-dominance is due to Y-linked rDNA loss.

12) Eickbush has done structural studies of the R2 distribution within the arrays and finds them mostly distributed throughout. How can all the copies of active rDNA be lost without reducing the copies of the interspersed R2 inserted copies?

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells" for further consideration at eLife. Your revised article has been favorably evaluated by K VijayRaghavan (Senior editor), a Reviewing editor, and three reviewers.

The manuscript has been significantly improved and no further experiments are needed. However, some important changes in the presentation need to be addressed before acceptance and publication, as outlined below:

The revised version of the paper "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells" has addressed most of the issues raised previously in the opinion of a majority of the reviewers. In particular, the assays clearly demonstrate that rDNA copy number and nucleolar usage changes in a regular way within male germ cells over the course of a lifecycle. This discovery is significant because rDNA copy number is currently thought to be much more stable, and all rDNA genes, if not inserted, are currently thought to be equivalently regulated. The work reported here is inconsistent with one or both of these assumptions and will stimulate new interest and advancements in this area.

However, the authors do not demonstrate how the observed germline changes relate to somatic rDNA levels and phenotype. They assume that the changes they observe in germ cell rDNA copy number also affect somatic cell rDNA copy number and could in an extreme case generate a defective NO. Since somatic cells could compensate significantly for the observed low level rDNA copy number variation by differential replication, and since no mutant chromosome derivatives or bobbed flies have been recovered, this is unjustified. Consequently, the final version of the paper should focus on the observed changes in germline rDNA copy number and nucleolar usage. How cyclic age dependent losses and recovery affect the soma and directional rDNA evolution, if it does so at all, remains unclear. As a result, it is premature to speculate so extensively on how these small germline rDNA copy number changes might relate to rDNA magnification. Such speculation should be cut back to 1 or two sentences and saved for a publication in which the copy number and phenotypic expression of rDNA in somatic cells is experimentally addressed.

In addition, the line-breaks in the graphs should be removed as the authors agreed in their response.

Second, the authors should acknowledge that the FISH experiments have not been shown to be fully quantitative.

Reviewer #1:

In the revised version and the response to reviewers, the authors have cogently addressed my concerns regarding some weaknesses and reviewer misunderstandings in the first version of this paper. This is a highly original work that addresses a central but neglected topic, namely the role of rDNA instability in stem cells during aging. The authors developed cytologically based methods that allow X and Y nucleolar activity to be visualized independently in stem cell. The in-situ hybridization data provide convincing corroboration of the PCR data that rDNA gene copy number is changing during aging and undergoing restoration in you F1 males. This paper will stimulate research in this field to answer many remaining questions concerning the developmental timing and mechanism of rDNA restoration, and on the epigenetic mechanisms that program rRNA gene activation and repression.

Reviewer #2:

The revised manuscript "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells" extends analysis of changes in rDNA expression using SNPs in transcripts, and includes new data using FISH to determine gene copy number on deletion Y chromosomes. These data do help document changes in gene copy number but there are some details that should be included. The authors use two deletion Y chromosomes to demonstrate that FISH is sensitive to copy number differences. However, copy numbers of these chromosomes (determined by some other method) should be reported. The authors seem to be using two different deletion Y chromosomes, but the names of these chromosomes are confusing (Df(YS)bb is different from Df(YS)bb-?).

A central issue that was raised in the first review was whether aging throws bobbed-deficient Y chromosomes, that should be apparent by some frequency of bobbed phenotypes in progeny. The authors allude that they have some data on this, but this is not clearly described. They should include a more thorough description and data supporting the statement in subsection “rDNA copy number maintenance through generations” that "we barely observed bobbed phenotype even among the sons of very old fathers". Given the extent of copy number loss the authors are estimating, such bobbed progeny are expected.

Reviewer #3:

This is a re-review of "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells," eLife, submitted by Yukiko Yamashita and colleagues.

My concerns remain, even after the authors sought to clarify their findings.

First, I remain very concerned that the vast literature on rDNA magnification remains unaddressed, even though those data are at odds with the observations made here. That concern is more profound because the measurements of rDNA loss here are not direct and simple: the authors use nucleolar shape, Y:X ratio, unvalidated FISH, and qPCR quantification on one tissue relatively late in development (eclosed adult testes) as proxy for rDNA copy number in sperm; more direct and more reliable assays are available.

The authors assert that rDNA counts drop with aging males, that those lower counts are still detectable (and are to the same degree) in testes of newly-eclosed males, they recover soon, and are lost again (presumably due to aging). The essential experiment that was called for in the first review was to measure loss in the whole soma of the progeny (in some type of rDNA[0]/Y flies – males of sc[4]sc[8]Y or females of DX,rDNA[0]/Y). The authors state that they did not observe any bobbed phenotype, which is perhaps expected if the original Y had a large number of rDNA copies, however losses should be easily detectable in whole animals using qPCR (as Aldrich and Maggert did), or the authors could start with shorter rDNA array containing Ys, such as one of the bobbed alleles to which they have access, and measure loss based on bobbed phenotype. The justification for asking this is simple: the Ys from old males should have reduced rDNA copy number, obvious in all of the soma of the offspring. In the case of the bobbed literature, this would have been noticed (by > dozens of reports) as an increase in the bobbed expressivity or penetrance, which was never reported; in fact rDNA copy number is remarkably robust fly-to-fly and generation-to-generation, which is at odds with the observations here. Either those previous studies are wrong, or they did not do the crosses in a way to detect such a remarkable variation in rDNA copy number, or the observations here are not detecting rDNA loss. Whichever the answer is, this has to be resolved prior to accepting the author's conclusion of their data (Points 1, 3, 10).

As an aside, in terms of rDNA copy number changes, nothing in the authors' data rule out that they are observing a process of endoreduplication/polytenization in a subset of cells in the testes (as they allude to in their response to Point 5). If that is the case, the apparent "losses" and "gains" would not bear on the sperm haploid genomes, which would be consistent with both the authors' measured effects and the known genetic stability from others' work on bobbed magnification. Without looking in the soma of the offspring, they cannot know. This experiment, in my mind, remains essential.

Second, I am unconvinced by the authors' assurance that their FISH is quantitative. Their care to not saturate the signal is fine, but they do not show any data that it is quantitative (which would require determining a lowest-threshold, a linearity of response between that threshold and saturation, and ample controls/experiments showing that photobleaching, preferential binding, etc are not issues). I think that a lower threshold must exist since the authors cannot detect their hypothetical extrachromosomal rDNA (Lines 371-374). FISH is not intrinsically quantitative and cannot be used as such unless great pains are taken. Hence, treating or calling it quantitative is not appropriate (Point 1).

Related to this, I'd still like to know how exclusive and abundant these SNPs are to the two (X-linked and Y-linked) rDNA arrays (Point 8).

Third, I still do not understand how the overall structure of the Y-linked array is consistent with the authors' model. If R1- and R2- inserted rDNA copies are interspersed with uninserted rDNA, I cannot imagine a damage-and-repair mechanism (that involves intrachromosomal HR-based recombination) that does not remove intervening rDNA. Either way, the structure of the Ys (i.e., degree of interspersion/clustering of R1- and R2- elements in the entire rDNA gene cluster) being used in this study should be included, as we asked (Points 8 and 12).

In general, I am not satisfied with the discussion of magnification and how that bears on these studies. I am not convinced they are the same thing, and I do not understand how the authors envision their proposed phenomenon and magnification are related. In subsection “rDNA copy number maintenance through generations”, they propose unequal sister exchange as a mechanism, however the authors should explain the fate of the other (shortened) product of such an exchange. Are they detectable in their data? (Points 3, 11).

I continue to be uncomfortable with the data presentation, with line breaks accentuating the magnitude of effects. Breaks are conventionally used to graph data with vastly different values on the same graph, where here they are used consistently but have the effect of making smaller effects seem larger. Graphs should give a visual representation of the size of effects, but the use in this paper overemphasizes them (sometimes to a very large degree). This change remains essential (Point 9).

The question as to whether an X, in X/Ybb males, is also subject to these losses seems central to the authors' model. It can – and should be – addressed experimentally (Point 10).

These above experiments were originally denoted as "essential," but they were not done in this revision. It is my opinion that this study – if correct – could overturn decades of research on rDNA copy number determination, rDNA stability and changes, and the work on rDNA magnification. It is thus critical that the appropriate experiments be done to rule out alternatives and to pursue the specific predictions of these proposed models.

eLife. 2018 Feb 13;7:e32421. doi: 10.7554/eLife.32421.021

Author response

Summary:

Loss of rDNA in aging yeast cells and the presence of damaged rRNA gene copies in many genomes testifies to the challenges of keeping highly active repeated rDNA genes in good working order. Drosophila rDNAs are a complex mix of active and inserted genes located on both X or Y chromosomes whose functional gene number varies between strains, with severe deficiencies causing a "bobbed" phenotype. rDNA homogeneity and copy number are maintained at least in part by unequal crossing over. Under certain conditions, rDNA-deficient male germ cells can restore copy number by a "magnification" process whose mechanism and regulation remain poorly understood. In the manuscript, the authors analyze rDNA expression in male germ cells using imaging, FISH, and qPCR to document Y chromosome nucleolar dominance that weakens with age in association with reductions in rDNA copy number. Progeny of old males inherit reduced Y chromosome rDNA copy number but restore it to normal levels based on cytological assays during the first 10 days of development by an unknown process the authors argue may be magnification.

The work is interesting and if fully validated would represent a significant advance, but at present, the authors' claims are inadequately documented.

Before getting into the details of the response, first we would like to summarize two major points in addressing reviewers’ concerns.

Major point 1. We document that rDNA is reduced during aging, which is the first clear demonstration of rDNA reduction in stem cells during aging in multi-cellular organisms. Reviewers felt that we have not adequately shown that rDNA loss occurs exclusively occur from Y chromosome. We do not claim or believe that rDNA copy number reduction occurs exclusively on the Y chromosome. Our results only show ‘preferential’ loss of Y rDNA, which we speculate (based on our results and established knowledge in the literature) to be caused by transcriptionally active state of Y rDNA. This does not indicate that silent rDNA (on X) never loses the copy number. In addition, X rDNA also becomes active at some point during aging, thus likely start losing the copy number. Importantly, our major conclusion on rDNA copy number reduction in aging does not lose its impact if it’s not exclusively from Y rDNA. We edited throughout the text to clarify this point to ensure that we do not indicate that rDNA loss occurs exclusively on Y rDNA.

Major point 2. We provide evidence that reduced rDNA copy number can recover in the subsequent generation (specifically during young age), which we call a ‘maintenance mechanism’, and we propose that classically-observed ‘magnification’ phenomenon may be the manifestation of this ‘maintenance mechanism.’ Reviewers felt that it is strange that some observations do not match between rDNA magnification from previous studies and rDNA maintenance mechanism in our study. Our proposal is that magnification is an extreme case that utilizes ‘maintenance mechanism’, and that ‘regular’ maintenance mechanism operates within the range without showing extremity of magnification (i.e. extreme rDNA loss to yield bobbed phenotype). We do not intend to claim that the maintenance mechanism we describe here is a synonymous of rDNA magnification: instead, rDNA magnification is manifestation of rDNA maintenance mechanism, which may not be equally efficient when recovering from extreme degrees of rDNA loss. We have clarified this point throughout the text.

Essential revisions:

1) A major problem is that the changes in Y chromosome rDNA copy number are not simply and directly documented. The authors quantify rDNA by Q-PCR in genomic DNA purified from whole testes. Testes contain germ cells, meiocytes, GSCs, gonadal mesoderm (including polyploid cells), etc. This is a very insensitive way to quantify the amount of Y-linked rDNA in GSCs, given the other cell types and the X-linked array signals. Although they attempt a second method to confirm the results, the authors present no data that their allele-specific FISH are quantitative. To critically test the central issue or rDNA copy number changes, the authors should cross young X/Y males with a full complement of rDNA as well as increasingly older males that seem to be losing Y-linked rDNA genes, to rDNA deficient C(1)DX/Y,B females or Df(1)sc4sc8/FM7 females and measure the rDNA levels in young C(1)DX/Y or Df(1)sc4sc8/Y offspring containing only Y-linked rDNA genes. They should observe if these animals begin to show a bobbed phenotype and progressively less rDNA as the age of the X/Y father progressively increases. The same test of Y chromosome rDNA copy number should be done using F1 male offspring of an old father, of various ages. In this case, the exclusively Y chromosome rDNA should be deficient when the F1 male was young but should return to normal levels during the 1st 10 days of F1 adult life.

Concerns raised here are important, because these data are the most critical foundation of our proposal in this study. First, as we described in our ‘major point 1’ above, the importance of the present study does not rely on rDNA reduction being exclusively from Y rDNA. qPCR is not intended to show Y-specific loss of rDNA. qPCR has been used successfully by Maggert lab to detect changes in rDNA copy number (Paredes and Maggert, 2009), and we believe that the qPCR results here establishes the overall loss of rDNA in the testis, which is primarily composed of germ cells derived from GSCs.

The weakness due to the use of the whole tissue was complemented by DNA FISH, which surely focuses on GSCs and their progeny. Please note that the DNA FISH method to quantify rDNA on chromosomes were not ‘allele specific’ unlike the RNA FISH in this study, and the probes should hybridize equally to both X and Y rDNA loci (as described in the Materials and methods section and Supplementary file 3, 18S rDNA on mitotic chromosomes were detected by Stellaris probes, comprised of 48 fluorescently-labelled oligonucleotides). We distinguished X vs. Y rDNA loci by adding Y-specific marker (AATAAAC)n, which is located right next to the Y rDNA (although we can easily distinguish X and Y chromosomes by their mitotic morphology, we used this (AATAAAC)n signal as a definitive marker to tell apart X vs. Y). We used our microscope setting to detect ‘signal saturation’, and all of images for quantification were obtained under the condition where no signal was saturated (this detail was added to the method section). Thus, pixel intensity analysis comparing X vs. Y rDNA ratio is expected to be very accurate. Combined with the results that we can detect clear reduction in Y rDNA in the partial Y rDNA deletion Df(YS)bb mutant, and complete loss in the Df(YS)bb- deficiency (now in Figure 3—figure supplement 1), we believe that this method is sufficiently quantitative.

We do not believe that assessing rDNA loss by bobbed phenotype observation is a sensitive method. While this is the most well-established characterization of rDNA insufficiency, changes in rDNA copy number can occur without the bobbed phenotype arising. It is known that the Drosophila stocks can have varying amount of rDNA (80-600 copies per genome) (Mohan and Ritossa, 1970) and it was described that bobbed phenotype arises when rDNA copy number is lower than 130 in total (Ritossa et al., 1966), although this number may vary depending on genetic background. This indicates that rDNA copy number can fluctuate without showing bobbed phenotype. As the Figure 3E, F suggests, the reduction in Y rDNA is around ~50% in aged males, the level that would not necessarily show bb phenotype even when combined with Xbb chromosome (rDNA complete loss). Indeed, we have already conducted this cross (crossing aging Y chromosomes into the background of Xbb chromosomes) and we did not see detectable increase in bobbed phenotype, indicating that rDNA loss during aging is not significant enough to cause bobbed phenotypes.

Instead, to better establish the heritable effects of germline rDNA copy number loss during aging, we crossed young vs. old fathers to young mothers to conduct FISH analysis in the sons, where young mother-derived X chromosome serve as a standard to compare ‘young Y’ vs. ‘old Y’ (as the ratio of Yyoung:Xyoung vs. Yold:Xyoung, where Xyoung coming from the same source of females. The results confirmed that Y chromosomes inherited from old fathers have reduced rDNA copies compared to Y chromosomes from young fathers. This data is now included in Figure 4E. We hope that the added experiments, the clarification of the text and methodology address reviewers’ concerns.

2) With respect to the FISH assays designed to identify X from Y-linked nucleoli, the criteria used to determine "association" must be described (on and around the second paragraph of subsection “Perturbed nucleolar morphology is associated with transcriptional activation of the normally silent X chromosome rDNA locus.”), especially because the few images presented in Figure 2 show a variance in locations of FISH, fibrillarin staining, etc. This is a concern because the active rDNA are thought to loop into the FC of the nucleolus, while the interspersed inactive cistrons are at the FC/DFC border, and linked chromosome material (including the heterochromatin) may not be found in the nucleolus. Probes to neighboring DNAs (e.g., the 359 satellite) may be some distance away from active cistrons, so proximity is not an adequate indicator of transcriptional activity.

We agree that our FISH probes (359 repeat or AATAAAC repeat) do not actually represent the localization of rDNA itself. We also agree that distance between the FISH signal and the nearby nucleolus can vary to some extent as the reviewers pointed out. However, we did not have any difficulty in assigning which FISH signal (359 or AATAAAC) is closer to the nucleolus for the following reasons: most of the case, each nucleolus was juxtaposed to a FISH signal with no gap in between at all, or with the gap no larger than the FISH signal’s diameter itself (typically ~0.5µm), whereas the distance between the nucleolus and the other (non-associated) FISH signal was more than a few microns (this detail was added to the method). We did not encounter any ‘close calls’, where a nucleolus is equally far from X and Y FISH signals. If the reviewers’ concern is indeed the case, we would have observed cases where a nucleolus does not have any nearby FISH signals. Instead, we observed that every single nucleolus was found to be associated with a FISH signal, making us confident that rDNA and the juxtaposing heterochromatin locus are not separated far away. Therefore, we believe that it is highly unlikely that we are assigning a nucleolus to a wrong rDNA locus.

3) The relational of this work to magnification is currently exaggerated and needs to be accurately discussed. First, a summary of previous work on rDNA magnification including recent work should be given in the introduction. Second, the authors must recognize that there are no known "rDNA magnification" genes exclusively involved in this process. Hence, conclusions such as: "This result indicates that rDNA magnification contributes to germline rDNA maintenance during normal aging" are unjustified. Third, the authors should specifically include a discussion reconciling the differences in frequencies of magnification observed in previous studies (>1%) from those reported here (100%). Fourth, the authors should include a discussion of the previous observations that only some rDNA array-containing chromosomes are subject to magnification (and only in some genetic conditions), and how that would bear on their model.

We appreciate these constructive comments that helped improve the clarity of the manuscript. We have rewritten the section describing rDNA magnification and the involvement of mus-101 and mei-41 genes in rDNA magnification to better clarify the points mentioned here by the reviewers. The manuscript now emphasizes that these genes are not necessarily ‘rDNA magnification’ genes, but instead these genes, which also have functions in other cellular processes, have been shown to be required for rDNA magnification. We indicate that the conceptual similarities between rDNA magnification and the recovery of rDNA in F1s and the requirement of the same genes for both phenomena suggests that the same mechanisms may underlie both processes and thus germline rDNA maintenance in general. Our introduction to rDNA magnification now includes that the majority of magnification is primarily observed in the offspring of males with large deletions of Y rDNA. This is an important feature of our conclusion that the same mechanisms underlie both phenomena, as they both occur in the germline of males with reduced Y rDNA. The revisions to the text can be found in subsection “Recovery of GSC nucleolar morphology depends on the homologous recombination repair pathway”.

Regarding the third point, we realize that the text in the previous version of the manuscript was not clear enough and confused the reviewers. Our hypothesis (based on the data presented in this study) is that rDNA copy number is maintained within a ‘normal’ range (without undergoing severe reduction to the extent of showing bobbed phenotype) through generations by the ‘maintenance’ mechanism, as discussed above (our ‘major point 2’ at the beginning of the response letter). In contrast, magnification, which is defined by the recovery from the severe reduction (showing bb phenotype) to a phenotypically-normal range, is the process that likely pushes the limit of ‘maintenance’ mechanism. Therefore, it is plausible that the magnification does not happen as efficiently as regular ‘maintenance’ recovery. We have clarified this issue in the revised text in subsection “rDNA copy number maintenance through generations”.

4) It seems to be a misnomer to describe the changes in nucleolar appearance as aberrant nucleolar morphology. The FISH mapping shows that these changes are not a nucleolar defect but are simply the activation of two loci.

The term ‘abnormal’ nucleolar morphology is not meant to convey that there is a defect in the nucleolus itself, but simply that this morphology is different from the ‘normal’ morphology (i.e. single, round nucleolus morphology) that is most common. To prevent confusion, we have edited the text to instead describe the deformed and fragmented morphology as ‘atypical.’

5) The authors do not address the very interesting question of whether the nucleolar phenomena and rDNA copy number changes they observed are confined to germ cells. Is there nucleolar dominance, loss of genes with aging, recovery in F1 adults in somatic hub cells and cyst stem cells, like in GSCs? In particular, does the reduced level of Y rDNA inherited from aged fathers persist in the soma of F1 animals, even though rDNA copy number is restored in the germline?

We also find the possibility of changes in rDNA in somatic stem cells to be very interesting. However, we consider the rDNA of the germline to be the most relevant, since these are the only cells that contribute genetic information for subsequent generations and our report concerns transgenerational inheritance of the rDNA dynamics. Also, we consider that the study of somatic stem cells might not provide relevant information that is worth comparing with the results obtained from germline for a few reasons described below.

In an ongoing project in our laboratory, we found that nucleolar dominance indeed happens in somatic tissues of males during development (Y dominant in gut, fat body, larval brain etc.). However, studying nucleolar dominance during aging in these somatic tissues is not relevant to the present study, because 1) some somatic cells (such as neuroblast) do not persist into adulthood (thus the effect of aging cannot be studied), 2) other somatic cells stop proliferating, and thus any observations made in such cells cannot be interpreted in parallel with GSCs, where our interest is in aging phenotype likely caused by repeated cell division cycles, 3) yet other somatic cells undergo polyploidization, confounding the interpretation of most of experiments that were employed in this study (qPCR, FISH etc.). Because of these confounding factors, and because the knowledge obtained from the somatic cells cannot be utilized to better interpret the results of the present study, we consider this to be beyond the scope of present study.

6) Figure 5: Can't see the X and Y signals in panel B, F1 10d.

We have altered the figure to include arrows indicating the X signal. Also, we apologize that we had mistakenly circled (marked) GSCs whose nucleolus cannot be seen (being out of the focal plane). Now we removed those circles and only the relevant GSCs are marked.

7) Subsection “Destabilization of rDNA loci during aging in Drosophila male GSCs.” Why would activation of additional rDNA genes to compensate for lost genes, leaving the total number of active genes intact, "increase the chance of replication transcription collisions and generate a vicious cycle? The argument is not clear here.

The rationale for this argument is based on the assumption that the same number of rDNA copies would need to be transcriptionally active in all cells of the same type (GSCs in this case), no matter how many rDNA copies there are in each individual cell (e.g. 150 copies to be transcriptionally active out of total 200 copies (75%) vs. 150 copies to be active out of total 450 copies (33%)). In this scenario, cells with fewer rDNA copies have a higher proportion of active rDNA, as the denominator in the ratio is lower than in cells with more rDNA, while the numerator is constant. The likelihood of a collision between transcription and replication machinery is a function of the frequency that any given region is being transcribed or replicated. As the portion of rDNA that is being transcribed increases, the frequency of transcription across all rDNA, on average, increases, meaning the likelihood for collision with replication machinery also increases, even though there is no change to the rate of replication. Since collisions between transcription and replication machinery have been proposed to induce DNA damage that can cause further rDNA copy loss, this creates the potential for a ‘vicious cycle.’ We propose there is the potential that rDNA loss can become accelerated because as copies are lost, the portion of rDNA that is transcriptionally active increases, increasing the probability of transcription-replication collisions, which cause further rDNA loss, thus causing a greater increase in the portion of active rDNA, and creating more opportunity for more rDNA loss. We have edited the text at in subsection “Destabilization of rDNA loci during aging in Drosophila male GSCs” to better clarify this argument within the discussion.

8) The structure of the Y and X chromosome rDNAs is not fully described and affect some of the inferences made in this manuscript. The authors describe loss of 45S cistron copies with age, but no loss of R1/R2 retroposons. Are there R1/R2 insertions on the Y chromosome (implying preferential loss), or are all the copies on the X chromosome?

We have done qPCR from animals, which harbors Y rDNA only or X rDNA only to determine the relative copy number of R1 and R2 on X vs. Y chromosomes. Interestingly, we found that R2 copy number is similar between the X and Y chromosome, whereas R1 is much more abundant on the X chromosome than the Y (see Author response image 1). This finding indicates that the Y chromosome has many more uninserted rDNA copies than the X chromosome. However, this data is uninformative to our hypothesis that transcriptionally active rDNA copies are preferentially lost, since it is unclear if the more abundant uninserted copies on the Y may or may not influence Y dominance, and any assumption that it does influence Y dominance would be purely speculative. Since this data neither supports nor weakens our hypothesis, we consider it to be unessential to main analysis in this report and distracting from those points, and decided it is better to leave this data unpublished until the context of this result can be better understood in the future. If, however, the editor/reviewers do think that including this data would be critical for publication, we would be happy to include it with this manuscript.

Author response image 1.

Author response image 1.

Open in a new tab

9) The data in Figure 3 and the fourth paragraph of subsection “Y chromosome rDNA copy number decreases in the male germline during aging.” details the loss of rDNA cistrons by qPCR. It would be worth converting these numbers into assignments for each chromosome, i.e. in a wildtype male 34% of 18S sequence is on the X and 66% on the Y. Then, after 40 days, there is (if the X doesn't change) 34% on the X and 18% remaining on the Y (loss of two thirds of the locus).

Although it is theoretically possible to combine qPCR data and the FISH data (to assess the XY ratio) to convert to numbers such as ‘33% on X, 66% on Y’, we do not think that this conversion would provide an accurate estimation for multiple reasons. First, our data do not conclude that copy number loss is ‘exclusively’ from Y rDNA (our ‘major point 1’ described at the beginning of the response letter). We think that our data only suggest that the copy number loss is ‘preferentially’ from Y (due to its transcriptional activity). Note that transcriptional activity only accelerates instability but is not the sole cause of instability. Thus, silent X rDNA is expected to lose its copy number albeit at a lower late. During aging, as nucleolar dominance is compromised, X rDNA becomes active, and thus likely accelerates the rate of instability: this means that X rDNA’s rate of instability is not constant during the course of GSCs’ aging. Because X rDNA copy number cannot be assumed to be constant during aging for these reasons, simply combining qPCR data and FISH data would not provide an accurate estimation of the total amount of rDNA on each chromosome. We have changed the text in subsection “Y chromosome rDNA copy number decreases in the male germline during aging” to clarify this distinction.

9) Multiple graphs are interrupted to emphasize the small changes in nucleolar morphology frequencies. This may overstate the size of the effect. There seems to be some discordance of the measurements of rDNA loss by qPCR and the more subtle changes in nucleolar appearance. Is this possibly due to a threshold at which nucleolar dominance is lost?

Please note that we consistently made the break in ALL graphs describing percentage of nucleolar morphology or nucleolar dominance (X and Y expression), for ease of reading and greater transparency of our data (by making thin stacks or error bars more visible). Exact p-values are provided to all relevant data sets, and the number is clearly indicated on Y axis. Thus, we believe that these are fair and even representation of the data under all conditions. We did not do this to mislead or exaggerate the effect of our findings (if we have done this manipulation to only a subset of graphs, it would have been misleading, but it is not what we did). If it is critical to represent the data without the breaks, we can remove them, but we believe that the breaks make our data more digestible for our readers.

10) The age effect described predicts that there would be loss or rDNA cistrons from the X chromosome when combined with a Df(YS)bb; is this the case? Overall, more details on the structure of the "partially deleted" Df(YS)bb would be useful.

Indeed, X/Df(YS)bb flies would have active X rDNA, which is clearly indicated in Figure 3G, H. And we certainly predict that X rDNA would be more destabilized. However, we would like to clarify that we did not/do not claim that silent X rDNA is entirely stable. The core message of the present study is dynamic nature of rDNA copy number (decrease during aging, recovery in the next generation), and we do not think this is X or Y chromosome-specific phenomenon. Transcriptional state accelerates but not solely determines the instability - thus, Y rDNA is more affected than X rDNA, but not exclusively. Indeed, the data in Figure 4E suggest that X chromosome is also somewhat compromised during aging, and we had discussed that X rDNA is also likely affected during aging. We have now clarified this point in the revised text not to leave the impression that rDNA instability occurs exclusively on the Y chromosome.

Regarding the nature of the Df(YS)bb chromosome, it is the Y rDNA deficiency stock listed at the Bloomington Drosophila Stock Center, and is previously characterized having insufficient Y rDNA for viability when X rDNA is disrupted (Cline, 2001). This reference is now included in the text. The nature of ‘partial deletion’ of this chromosome is described in this study by DNA FISH, which shows clear reduction (but not complete deletion) of Y rDNA (Figure 3E, F).

11) The authors under-explain the results by Ahmad and Hartl and Eickbush, whose groups all saw that not all chromosomes or genetic background manifest nucleolar dominance. Further, they do not mention that in plants nucleolar dominance is a transient phenomenon, often lost after the first week of development. These findings may bear on the authors' interpretation and certainly bear on the "null hypotheses" that loss of Y-dominance is due to Y-linked rDNA loss.

Please note that (Zhou et al., 2012) by Hartl and Eickbush groups provided quite strong evidence that Y rDNA is dominant over X rDNA in all genetic backgrounds examined (although they found that position effect variegation shows variance based on Y rDNA contents). (Greil and Ahmad, 2012) have presented a few Y chromosome variants that exhibit co-dominance, where rDNA copy number cannot account for the observed co-dominance. However, considering the fact that the very same chromosome from a particular stock is undergoing aging in our aging experiments, we consider it to be highly unlikely that ‘background’ changes occur to these chromosomes, influencing the nucleolar dominance.

In regards to developmental changes in nucleolar dominance in plants, it is established that while co-dominance occurs during early development in both Arabidopsis hybrids and A. thaliana non-hybrid development, a subset of rRNA genes are silenced during early vegetative development (days 10-14) the dominance is then maintained (Earley et al., 2010; Pontes et al., 2007). This suggests that loss of established nucleolar dominance is in fact uncommon unless rDNA is disrupted.

12) Eickbush has done structural studies of the R2 distribution within the arrays and finds them mostly distributed throughout. How can all the copies of active rDNA be lost without reducing the copies of the interspersed R2 inserted copies?

Recent work from Thomas Eickbush’s lab proposes a model that may explain this bias in the rDNA copies lost during aging. (Zhou et al., 2013) suggests that transcription of rDNA preferentially occurs in contiguous blocks of uninserted rDNA copies, and that intra-chromosomal exchanges that cause rDNA loss only occur within these transcribing blocks, thus preferentially causing loss of uninserted rDNA copies. Therefore, preferential loss of uninserted copies (as shown in Figure 3A) is consistent with the current knowledge of the field. We have added to the text in subsection “Destabilization of rDNA loci during aging in Drosophila male GSCs” to include in our discussion how our data is consistent with this model for rDNA loss

[Editors' note: further revisions were requested prior to acceptance, as described below.]

The manuscript has been significantly improved and no further experiments are needed. However, some important changes in the presentation need to be addressed before acceptance and publication, as outlined below:

The revised version of the paper "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells" has addressed most of the issues raised previously in the opinion of a majority of the reviewers. In particular, the assays clearly demonstrate that rDNA copy number and nucleolar usage changes in a regular way within male germ cells over the course of a lifecycle. This discovery is significant because rDNA copy number is currently thought to be much more stable, and all rDNA genes, if not inserted, are currently thought to be equivalently regulated. The work reported here is inconsistent with one or both of these assumptions and will stimulate new interest and advancements in this area.

We are grateful that the majority of reviewers agreed on the significance of our work and its suitability for publication.

However, we are afraid that the perceived ‘inconsistency with the literature’ mentioned here might be caused by misunderstanding of the literature by reviewer #3. First, on the contrary to reviewer #3’s statement that the fly-to-fly, generation-to-generation variation in rDNA copy number is minimal, Lyckegaard and Clark have demonstrated striking 6-fold variations in rDNA copy number among D. melanogaster populations (Lyckegaard and Clark, 1989, Lyckegaard and Clark, 1991). Also, an earlier study showed similar variation in rDNA copy number (80-600 copies) (Mohan and Ritossa, 1970), where flies with <~130 copies exhibiting bobbed phenotype (meaning that the copy number variation between 130-600, ~4.5-fold variation, is asymptomatic). Furthermore, rDNA copy number fluctuation is a broadly observed phenomenon as mentioned above, arguing that our discovery is rather in line with the existing literature.

We suspect that reviewer #3’s perception of ‘robust stability of rDNA’ comes from the fact that ‘normal’ fly populations or aging flies do not exhibit bobbed phenotype. However, the earlier studies have clearly demonstrated that asymptomatic copy number variation is fairly common. We would very much appreciate if reviewer #3 could point to any literature that refuted these studies we are referring to (if s/he is aware of any).

Second, the notion that ‘all rDNA genes, if not inserted, are currently thought to be equivalently regulated’ is not supported by the body of existing literature. Although the higher rate of transposon insertion on X chromosome was speculated to be a cause of nucleolar dominance in Drosophila (Greil and Ahmad, 2012), it has not been experimentally proven (which would be extremely difficult, because one would have to remove all/many transposon insertions from X chromosomes in an isogenic background to test this hypothesis). Moreover, and more importantly, the field of ‘nucleolar dominance’ is the research area to specifically investigate how particular rDNA loci are selectively activated, whereas other rDNA loci are selectively inactivated as a means of dosage control of rRNA expression (reviewed in (McStay and Grummt, 2008, Tucker et al., 2010)). This is often described as ‘a large scale epigenetic regulation, only second to X inactivation in mammalian females’(Pontes et al., 2007). Therefore, we do not believe that our results or claims are inconsistent with the existing literature, requiring reconciliation as reviewer #3 indicates.

However, the authors do not demonstrate how the observed germline changes relate to somatic rDNA levels and phenotype. They assume that the changes they observe in germ cell rDNA copy number also affect somatic cell rDNA copy number and could in an extreme case generate a defective NO. Since somatic cells could compensate significantly for the observed low level rDNA copy number variation by differential replication, and since no mutant chromosome derivatives or bobbed flies have been recovered, this is unjustified. Consequently, the final version of the paper should focus on the observed changes in germline rDNA copy number and nucleolar usage. How cyclic age dependent losses and recovery affect the soma and directional rDNA evolution, if it does so at all, remains unclear. As a result, it is premature to speculate so extensively on how these small germline rDNA copy number changes might relate to rDNA magnification. Such speculation should be cut back to 1 or two sentences and saved for a publication in which the copy number and phenotypic expression of rDNA in somatic cells is experimentally addressed.

As we have discussed in the previous round of revision, we fully agree with these comments. The inclusion of more extensive discussion on rDNA copy number in somatic cells was in response to the specific instruction requested by reviewers in the previous round. (previous review comment ‘The authors do not address the very interesting question of whether the nucleolar phenomena and rDNA copy number changes they observed are confined to germ cells. Is there nucleolar dominance, loss of genes with aging, recovery in F1 adults in somatic hub cells and cyst stem cells, like in GSCs? In particular, does the reduced level of Y rDNA inherited from aged fathers persist in the soma of F1 animals, even though rDNA copy number is restored in the germline?’). Please also note that, as we have stated in our previous response letter, we considered rDNA copy number dynamics in somatic cells to be beyond the scope of the current study (although very interesting), for the exact same reason that the reviewers felt here: thus, we limited our discussion on this issue to within our response letter and did not include any discussion on somatic cells in the main text of the previous version. Thus, we believe there is nothing to be ‘cut back’.

In addition, the line-breaks in the graphs should be removed as the authors agreed in their response.

We have revised the figures according to this suggestion.

Second, the authors should acknowledge that the FISH experiments have not been shown to be fully quantitative.

We have added this description to the revised text, which now reads as “These results suggest that our DNA FISH method is sensitive enough to distinguish differences in the relative copy number of X and Y chromosome rDNA loci between different conditions, although it might not be fully quantitative.” (subsection “Y chromosome rDNA copy number decreases in the male germline during aging”)

Reviewer #1:

In the revised version and the response to reviewers, the authors have cogently addressed my concerns regarding some weaknesses and reviewer misunderstandings in the first version of this paper. This is a highly original work that addresses a central but neglected topic, namely the role of rDNA instability in stem cells during aging. The authors developed cytologically based methods that allow X and Y nucleolar activity to be visualized independently in stem cell. The in-situ hybridization data provide convincing corroboration of the PCR data that rDNA gene copy number is changing during aging and undergoing restoration in you F1 males. This paper will stimulate research in this field to answer many remaining questions concerning the developmental timing and mechanism of rDNA restoration, and on the epigenetic mechanisms that program rRNA gene activation and repression.

We very much appreciate this reviewer for his/her positive comments. As concisely summarized by this reviewer, we believe that the combination of cytological method (DNA FISH) and PCR provides strong evidence for rDNA copy number fluctuation. Whereas our method might not provide the resolution of rDNA copy number down to the absolute exact number, we believe that the resolution is sufficient to make our points of age-related decline and subsequent recovery.

Reviewer #2:

The revised manuscript "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells" extends analysis of changes in rDNA expression using SNPs in transcripts, and includes new data using FISH to determine gene copy number on deletion Y chromosomes. These data do help document changes in gene copy number but there are some details that should be included. The authors use two deletion Y chromosomes to demonstrate that FISH is sensitive to copy number differences. However, copy numbers of these chromosomes (determined by some other method) should be reported.

We appreciate this reviewer for his/her support.

For copy number estimation of two Ybb alleles, we do not believe that any technically feasible experiments would provide meaningful answers. The most precise estimation may be achieved by qPCR: however, to obtain rDNA copy number on particular chromosome, we would have to place it in genetic background that contains no rDNA on the other chromosome. In this case, placing Ybb in the background of Xbb. This leads to lethality for both of Df(YS)bb and Df(YS)bb- strains used in our study, thus we cannot perform such qPCR. However, for Df(YS)bb (partial deletion), based on Y:X rDNA ratio shown in Figure 3E, F, we can estimate this Ybb allele contains 15-32% of wild type copy number. For Df(YS)bb-, its rDNA is under the detection limit by our rDNA FISH method (Figure 3—figure supplement 1). We believe that this level of resolution provides sufficient information for the purpose of our current study. We also stated that our method may not be fully quantitative.

The authors seem to be using two different deletion Y chromosomes, but the names of these chromosomes are confusing (Df(YS)bb is different from Df(YS)bb-?).

We fully agree and had sought for the way to name them more clearly (we had a lengthy discussion among authors), such as bbpartial vs. bbcomplete. However, Df(YS)bb- and Df(YS)bb are the terms used by the original authors who described these deficiencies and listed as such in FlyBase. We concluded that, if we rename, it would cause further confusion for the future readers who try to track down the information, corresponding our “bbpartial” to “bbcomplete”. Therefore, we decided to cite the original papers that described each allele and adhere to original nomenclature.

A central issue that was raised in the first review was whether aging throws bobbed-deficient Y chromosomes, that should be apparent by some frequency of bobbed phenotypes in progeny. The authors allude that they have some data on this, but this is not clearly described. They should include a more thorough description and data supporting the statement in subsection “rDNA copy number maintenance through generations” that "we barely observed bobbed phenotype even among the sons of very old fathers". Given the extent of copy number loss the authors are estimating, such bobbed progeny are expected.

As we discussed in the first round of revision, it was described that bobbed phenotype arises when rDNA copy number is lower than 130 in total (Ritossa et al., 1966), although this number may vary depending on genetic background. This indicates that rDNA copy number can fluctuate without showing bobbed phenotype, which is consistent with earlier work that showed copy number variation of 130-600 while being asymptomatic (Mohan and Ritossa, 1970). As the Figure 3E, F suggests, the reduction in Y rDNA is around ~50% in aged males, the level that would not necessarily show bb phenotype even when combined with Xbb chromosome (rDNA complete loss). Indeed, when 40 day old fathers are mated with bb158/FM7 females, we obtained 428 male bb158/Y off spring and 424 female bb158/X offspring, and none of them were bobbed. This indicates that rDNA copy number decrease is not sufficient to exhibit bobbed phenotype even when placed in the background of Xbb (bb158).

Reviewer #3:

This is a re-review of "Transgenerational dynamics of rDNA copy number in Drosophila male germline stem cells,", submitted by Yukiko Yamashita and colleagues.

We thank this reviewer for thoroughly reviewing our manuscript. However, we’re afraid that some of concerns raised by this reviewer are ungrounded. As detailed below, our observations are consistent with existing literature and we do not find a need of ‘reconciliation’.

My concerns remain, even after the authors sought to clarify their findings.

First, I remain very concerned that the vast literature on rDNA magnification remains unaddressed, even though those data are at odds with the observations made here. That concern is more profound because the measurements of rDNA loss here are not direct and simple: the authors use nucleolar shape, Y:X ratio, unvalidated FISH, and qPCR quantification on one tissue relatively late in development (eclosed adult testes) as proxy for rDNA copy number in sperm; more direct and more reliable assays are available.

In fact, our data are NOT at odds with the existing literature. This reviewer states that rDNA copy number is stably maintained with fly-to-fly, generation-to-generation variations being minimal. On the contrary to this statement, a large body of literature has described variations in rDNA copy number (Lyckegaard and Clark, 1989, Lyckegaard and Clark, 1991, Mohan and Ritossa, 1970) with striking 6-fold difference without showing detectable bobbed phenotype. Although these studies did not specifically consider aging as a parameter that may affect rDNA copy number (which is one of our major discovery in this study), their study unlikely controlled for the age of flies, and thus we suspect that a part of the copy number variation they detected may reflect aging.

We suspect that this reviewer assumed the fact that there is little fly-to-fly, generation-to-generation variation in the frequency of bobbed flies as indication of little copy number variation. However, previous studies and our study agree that copy number variation is not necessarily symptomatic, as bobbed phenotype only appears when copy number is extremely low (less than 130 copies, as estimated by (Ritossa et al., 1966)).

As we already described above, rDNA copy number variation is not unique to Drosophila, and broadly observed in many species including humans.

The authors assert that rDNA counts drop with aging males, that those lower counts are still detectable (and are to the same degree) in testes of newly-eclosed males, they recover soon, and are lost again (presumably due to aging). The essential experiment that was called for in the first review was to measure loss in the whole soma of the progeny (in some type of rDNA[0]/Y flies – males of sc[4]sc[8]Y or females of DX,rDNA[0]/Y). The authors state that they did not observe any bobbed phenotype, which is perhaps expected if the original Y had a large number of rDNA copies, however losses should be easily detectable in whole animals using qPCR (as Aldrich and Maggert did), or the authors could start with shorter rDNA array containing Ys, such as one of the bobbed alleles to which they have access, and measure loss based on bobbed phenotype. The justification for asking this is simple: the Ys from old males should have reduced rDNA copy number, obvious in all of the soma of the offspring. In the case of the bobbed literature, this would have been noticed (by > dozens of reports) as an increase in the bobbed expressivity or penetrance, which was never reported; in fact rDNA copy number is remarkably robust fly-to-fly and generation-to-generation, which is at odds with the observations here. Either those previous studies are wrong, or they did not do the crosses in a way to detect such a remarkable variation in rDNA copy number, or the observations here are not detecting rDNA loss. Whichever the answer is, this has to be resolved prior to accepting the author's conclusion of their data (Points 1, 3, 10).

In the previous round of revision, we have provided explanations for somatic cells’ rDNA copy number is not relevant to our discovery described here, to which the majority of the reviewers seemed to have agreed and suggested not to discuss about somatic rDNA copy number (some more details are provided at the beginning of this response).

As an aside, in terms of rDNA copy number changes, nothing in the authors' data rule out that they are observing a process of endoreduplication/polytenization in a subset of cells in the testes (as they allude to in their response to Point 5). If that is the case, the apparent "losses" and "gains" would not bear on the sperm haploid genomes, which would be consistent with both the authors' measured effects and the known genetic stability from others' work on bobbed magnification. Without looking in the soma of the offspring, they cannot know. This experiment, in my mind, remains essential.

It is in theory possible that somatic gonadal cells undergo endoreplication and polytenization, influencing our qPCR results. However, if it were the case, we would not expect to see changes in rDNA copy number in F1s from old father (obviously, somatic gonadal cells will not be transmitted to the next generation, polytenized or not).

Second, I am unconvinced by the authors' assurance that their FISH is quantitative. Their care to not saturate the signal is fine, but they do not show any data that it is quantitative (which would require determining a lowest-threshold, a linearity of response between that threshold and saturation, and ample controls/experiments showing that photobleaching, preferential binding, etc. are not issues). I think that a lower threshold must exist since the authors cannot detect their hypothetical extrachromosomal rDNA (Lines 371-374). FISH is not intrinsically quantitative and cannot be used as such unless great pains are taken. Hence, treating or calling it quantitative is not appropriate (Point 1).

According to the suggestion by reviewers, we explained that our method might not be fully quantitative (see above). We also noted that photobleaching is minimal (no detectable bleaching during our imaging). As we had explained in the previous round of revision that there would not be ‘preferential binding’ of probes, because DNA FISH was conducted with Stellaris tiled probes (48 oligos) that do not distinguish SNPs (detailed in method, and probe sequences had been provided in the supplementary file 3). Thus, it is highly unlikely there will be preferential binding between X vs Y rDNA locus.

Related to this, I'd still like to know how exclusive and abundant these SNPs are to the two (X-linked and Y-linked) rDNA arrays (Point 8).

Figure 2—figure supplement 1 had been provided for this purpose. Y SNP signal is undetectable with our method in XO flies, and X SNP signal is undetectable in Xbb/Y flies, suggesting that SNPs are quite specific (if not ‘exclusive’) to individual locus. As was described in the method, DNA sequencing to determine SNPs were conducted using genomic DNA from XO and Xbb/Y without cloning a single rDNA copy from them. Therefore, if various SNPs exist as a mixture among rDNA copies within a rDNA locus, we would not have been able to ‘read’ the sequence well (chromatogram showing peaks of mixed signals). Therefore, it is fairly reasonable to assume the SNPs are highly homogenous and robust within individual rDNA locus. This is consistent with the well-established notion of ‘concerted evolution’ of repetitive sequence, which homogenizes mutations within the repeats across the locus (reviewed in (Eickbush and Eickbush, 2007)).

Third, I still do not understand how the overall structure of the Y-linked array is consistent with the authors' model. If R1- and R2- inserted rDNA copies are interspersed with uninserted rDNA, I cannot imagine a damage-and-repair mechanism (that involves intrachromosomal HR-based recombination) that does not remove intervening rDNA. Either way, the structure of the Ys (i.e., degree of interspersion/clustering of R1- and R2- elements in the entire rDNA gene cluster) being used in this study should be included, as we asked (Points 8 and 12).

This issue has been beautifully explained by (Zhou et al., 2013) using modeling, and our results are consistent with their model. Most importantly, however, the present study is simply reporting the experimental data that uninserted copies seem to be specifically lost, which is not affected by the presence of, but happens to fit well with, the existing modeling: thus, it is not reasonable to argue that our experimental data cannot be true, because one cannot imagine how it can happen.

In general, I am not satisfied with the discussion of magnification and how that bears on these studies. I am not convinced they are the same thing, and I do not understand how the authors envision their proposed phenomenon and magnification are related. In subsection “rDNA copy number maintenance through generations”, they propose unequal sister exchange as a mechanism, however the authors should explain the fate of the other (shortened) product of such an exchange. Are they detectable in their data? (Points 3, 11).

We would like to point out that no studies to date have shown the fate of shortened copy after unequal sister chromatid exchange. (Tartof, 1974) noted the emergence of a small number of individuals with worsened bobbed phenotype, while the flies undergo magnification, leading Tartof to speculate they might represent the individuals who inherited ‘shortened’ copies. We acknowledge that we do not fully understand how rDNA copy number recovers in young flies from old fathers: however, it is also true that no evidence has been provided as to exactly how magnification happens, either, and nobody knows where the shorted copies may go during unequal sister chromatid exchange that happens during magnification. Even more critically, although gene requirements (mus-101 etc.) and other circumstantial evidence have pointed to unequal sister chromatid exchange as a likely mechanism for magnification, this is not directly proven yet. In fact, (de Cicco and Glover, 1983) concluded that ‘unequal sister chromatid exchange cannot explain magnification, unless a single cell can undergo such a process repeatedly’. In our discussion in the manuscript, we mentioned that it might be explained if GSCs (which can undergo repeated round of asymmetric division) are the place for magnification, with clear indication that this is a pure speculation. As is clear by now, the request of examining the fate of shortened copy as a proof of our model clearly goes beyond the scope of this study.

I continue to be uncomfortable with the data presentation, with line breaks accentuating the magnitude of effects. Breaks are conventionally used to graph data with vastly different values on the same graph, where here they are used consistently but have the effect of making smaller effects seem larger. Graphs should give a visual representation of the size of effects, but the use in this paper overemphasizes them (sometimes to a very large degree). This change remains essential (Point 9).

We have revised all the graphs accordingly.

The question as to whether an X, in X/Ybb males, is also subject to these losses seems central to the authors' model. It can – and should be – addressed experimentally (Point 10).

We do not believe the nature of X rDNA dynamics is central to our model. If X rDNA does not change at all, it would only mean that our data indicate a striking decrease of Y rDNA copy number, and all data presented in this study becomes more quantitative in nature (for example, when we measure Y rDNA copy number as Y/X ratio, we would underestimate Y copy number decrease, if the denominator (X) is also decreasing. This concern will be eliminated if X is not losing copy number. If X rDNA is being lost from X during aging as well, this would mean that rDNA copy number loss is likely due to inherent nature of repeat, and it is not due to anything special to Y chromosome, or Y rDNA. Our discussion and statements are inclusive to both models, and we do not believe that distinction of these two possibilities is the core impact of our discovery.

Our explanation provided in the last round of revision was as follows:

“Indeed, X/Df(YS)bb flies would have active X rDNA, which is clearly indicated in Figure 3G, H. And we certainly predict that X rDNA would be more destabilized. However, we would like to clarify that we did not/do not claim that silent X rDNA is entirely stable. The core message of the present study is dynamic nature of rDNA copy number (decrease during aging, recovery in the next generation), and we do not think this is X or Y chromosome-specific phenomenon. Transcriptional state accelerates but not solely determines the instability-thus, Y rDNA is more affected than X rDNA, but not exclusively. Indeed, the data in Figure 4E suggest that X chromosome is also somewhat compromised during aging, and we had discussed that X rDNA is also likely affected during aging. We have now clarified this point in the revised text not to leave the impression that rDNA instability occurs exclusively on the Y chromosome.

Regarding the nature of the Df(YS)bb chromosome, it is the Y rDNA deficiency stock listed at the Bloomington Drosophila Stock Center, and is previously characterized having insufficient Y rDNA for viability when X rDNA is disrupted (Cline 2001). This reference is now included in the text. The nature of ‘partial deletion’ of this chromosome is described in this study by DNA FISH, which shows clear reduction (but not complete deletion) of Y rDNA (Figure 3E, F).”

We hope that this makes it clear that knowing whether X rDNA may undergo copy number reduction or not does not impact the value of our present study.

These above experiments were originally denoted as "essential," but they were not done in this revision. It is my opinion that this study – if correct – could overturn decades of research on rDNA copy number determination, rDNA stability and changes, and the work on rDNA magnification. It is thus critical that the appropriate experiments be done to rule out alternatives and to pursue the specific predictions of these proposed models.

As we hope is clear by now (based on explanation provided above), our results are not overturning decades of research: on the contrary to this perception, our results are in fact in line with existing knowledge, while providing advancement.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Figure 3—source data 1. Pixel intensity measurement and its ratio from DNA FISH plotted in Figure 3E,F.
    elife-32421-fig3-data1.xlsx (54.2KB, xlsx)
    DOI: 10.7554/eLife.32421.007
    Figure 4—source data 1. Pixel intensity measurement and its ratio from DNA FISH plotted in Figure 4E.
    elife-32421-fig4-data1.xlsx (45.5KB, xlsx)
    DOI: 10.7554/eLife.32421.009
    Figure 5—source data 1. Pixel intensity measurement and its ratio from DNA FISH plotted in Figure 5D.
    elife-32421-fig5-data1.xlsx (59.9KB, xlsx)
    DOI: 10.7554/eLife.32421.011
    Supplementary file 1. Probe sequences for rRNA SNP in situ hybridization.
    elife-32421-supp1.docx (69.5KB, docx)
    DOI: 10.7554/eLife.32421.014
    Supplementary file 2. Primer sequences for genomic rDNA qPCR.
    elife-32421-supp2.docx (73.6KB, docx)
    DOI: 10.7554/eLife.32421.015
    Supplementary file 3. Probe sequences for rDNA DNA FISH.
    elife-32421-supp3.docx (96.3KB, docx)
    DOI: 10.7554/eLife.32421.016
    Transparent reporting form
    elife-32421-transrepform.docx (244.8KB, docx)
    DOI: 10.7554/eLife.32421.017

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES