Three recent sex chromosome-to-autosome fusions in a Drosophila virilis strain with high satellite DNA content

Abstract

The karyotype, or number and arrangement of chromosomes, has varying levels of stability across both evolution and disease. Karyotype changes often originate from DNA breaks near the centromeres of chromosomes, which generally contain long arrays of tandem repeats or satellite DNA. Drosophila virilis possesses among the highest relative satellite abundances of studied species, with almost half its genome composed of three related 7 bp satellites. We discovered a strain of D. virilis that we infer recently underwent three independent chromosome fusion events involving the X and Y chromosomes, in addition to one subsequent fission event. Here, we isolate and characterize the four different karyotypes we discovered in this strain which we believe demonstrates remarkable genome instability. We discovered that one of the substrains with an X-autosome fusion has an X-to-Y chromosome nondisjunction rate 20 × higher than the D. virilis reference strain (21% vs 1%). Finally, we found an overall higher rate of DNA breakage in the substrain with higher satellite DNA compared to a genetically similar substrain with less satellite DNA. This suggests that satellite DNA abundance may play a role in the risk of genome instability. Overall, we introduce a novel system consisting of a single strain with four different karyotypes, which we believe will be useful for future studies of genome instability, centromere function, and sex chromosome evolution.

Introduction

Genome instability is characterized by large-scale mutations or rearrangements that result from repair of DNA breaks, and is a fundamental driver of karyotype evolution, chromosomal disorders, and cancer rearrangements (Black and Giunta 2018; Mayrose and Lysak 2021). The initiating event is spontaneous DNA damage such as double-stranded DNA breaks, one of the most dire events to occur in a cell (Featherstone and Jackson 1999). Even if DNA breaks are repaired, they often result in large-scale genome rearrangements. Robertsonian translocations, one of the most common rearrangements in medical genetics and evolution, occur when there are breaks near the centromere of acrocentric chromosomes and when repaired are fused to each other (Mayrose and Lysak 2021). Robertsonian translocations are associated with repeated miscarriages in humans and aneuploidy disorders like Patau and Down syndromes (Braekeleer and Dao 1990), with increased rates of aneuploidy driven by increased nondisjunction of Robertsonian or fused chromosomes (Schulz et al. 2006). Spontaneous Robertsonian translocation events, or chromosome fusions, are rarely detected in experimentally-tractable systems. Several consecutive steps must occur in order to detect a spontaneous fusion: (1) DNA breakage near the centromere; (2) repair and fusion with another acrocentric chromosome; (3) resolution of a single functional centromere; and (4) retention and increase in frequency of the fusion in the line. Natural genetic variation and environmental conditions that can affect the rate or probability of any of the steps are not understood.

Satellite DNA consists of long arrays of tandemly repeated sequences, and is often located near centromeres in heterochromatin (reviewed in Thakur et al. 2021). Satellite DNA varies greatly in sequence and abundance within and among species (Subirana et al. 2015; Wei et al. 2018; Cechova et al. 2019). Although satellite DNA differences between some species have been linked to reproductive incompatibilities (Ferree and Barbash 2009; Jagannathan and Yamashita 2021), the biological implications of intraspecies abundance variation have not been explored. Satellite DNA can vary in abundance by several megabases among individuals of the same species, including in flies and humans (Miga et al. 2014; Wei et al. 2014; Flynn et al. 2020). Satellite DNA appears to be constrained by maximum limits, with no species studied so far having more than about half of their genome made up of satellite DNA (Gall and Atherton 1974; Petitpierre et al. 1995; Mora et al. 2023). In past modeling efforts, satellite DNA arrays have been proposed to be weakly deleterious until they reach a maximum length beyond which they are not tolerated by selection (Charlesworth et al. 1986). Slow DNA replication or development time has been suggested as a mechanism to enforce strong negative selection against long satellite arrays, however, empirical evidence for this has been limited (but see Bilinski et al. 2018). Most karyotype rearrangements detected in both humans and other organisms originated from breaks near the centromere, especially in or near satellite DNA (Black and Giunta 2018; Balzano et al. 2021). This may be due to the intrinsic instability of satellite DNA arrays caused by replication stress of polymerases progressing through highly repetitive sequences, or the formation of unstable DNA topology (Barra and Fachinetti 2018).

Drosophila virilis is an excellent model for studying satellite DNA variation and karyotype stability. D. virilis has the highest relative abundance of simple satellite DNA (defined as satellites with unit length ≤20 bp) compared to any other studied species. Three 7 bp satellites, AAACTAC, AAACTAT, and AAATTAC take up over 40% of the genome in D. virilis, and they form arrays tens of megabases long in the pericentromeric region (Gall et al. 1971; Gall and Atherton 1974; Flynn et al. 2020). Other species such as Triatoma delpontei (Mora et al. 2023), Drosophila albomicans (de Lima and Ruiz-Ruano 2022), and Tenebrio molitor (Petitpierre et al. 1995) have similarly high amounts of satellite DNA, but are dominated by more complex satellites with unit lengths >20 bp. For simple satellites, to our knowledge, Drosophila melanogaster and Drosophila sechellia have the next highest abundance of simple satellite DNA outside of the virilis group, composing about 20% of the genome, or only half the relative amount of D. virilis (Wei et al. 2018). The D. virilis reference strain has retained the ancestral karyotype of the Drosophila genus, made up by five pairs of acrocentric chromosomes and one small chromosome pair referred to as “the dot” (Powell 1997, Schaeffer et al. 2008). In our 2020 work (Flynn et al. 2020) describing inter and intraspecies satellite DNA variation, we found that one strain in particular, vir00 (15010–1051.00) contained 15% higher satellite DNA abundance than other strains. The elevated satellite DNA abundance in vir00 prompted us to study it further, which we will present here.

Evolution of the sex chromosome arrangement has received abundant attention in both empirical and theoretical studies. Sex chromosome evolutionary studies have mainly made use of sex chromosomes that arose in different time periods (Charlesworth and Charlesworth 2000). In Drosophila, when an autosome fuses to either an X or Y chromosome, a so-called neo-Y chromosome is formed. Because either the fused or unfused version of the chromosome will only be present in males, and male Drosophila do not undergo recombination, mutations immediately begin to accumulate through Hill–Robertson interference and other linked-selection processes (Charlesworth and Charlesworth 2000). In the genus Drosophila, autosomes have fused to sex chromosomes multiple independent times (Nozawa et al. 2021): D. pseudoobscura (10 million years), D. miranda (1 million years; Bachtrog 2013), D. albomicans (0.24 million years; Wei and Bachtrog 2019), and D. americana (29 thousand years; Vieira et al. 2006). Neo-sex chromosomes formed by sex chromosome to autosome fusions are actually under-represented in Drosophila compared to the null model that all chromosomes have an equal probability of being involved in a fusion (Anderson et al. 2020), and have never been discovered at their infancy before detectable divergence has occurred. How new sex chromosome fusions become stable and how they compete with the ancestral karyotype within a species are unknown.

Here, we report the discovery of three independent and extremely recent sex chromosome to autosome fusions in one D. virilis strain, vir00. Two fusions involved the X chromosome and one involved the Y chromosome. We isolated four different versions of the same strain that differ in their karyotype (and satellite DNA abundance). We hypothesize that vir00 has been more prone to chromosome fusions, because of an increased risk of any or multiple biological processes that lead to fusions occurring and rising to detectable frequency. Our findings suggest that satellite DNA abundance is a factor that could affect the risk of DNA breakage events under stress, the first event that could lead to a chromosome fusion.

Materials and methods

Scripts required to reproduce the computational results are available here: https://github.com/jmf422/D-virilis-fusion-chromosomes.

Fly stock maintenance

The National Drosophila Species Stock Center typically maintains each stock in approximately 12 replicate vials, which are only merged if deaths make it necessary. For each strain ordered, two replicate vials are sent from the stock center. With the exception of vir00-Yfus, we assayed the karyotype within three generations of receiving it. Both sets of segregating chromosome fusions we found (vir00-Xfus-a/vir00-Nofus and vir00-Yfus/vir00-Yfus + Xfus-b) appeared to be segregating within a vial, rather than between replicate vials. Stocks were maintained on standard Drosophila medium at room temperature.

Neuroblast squashes and satellite DNA FISH

We dissected brains from wandering 3rd-instar larvae and performed the fixation steps as in Larracuente and Ferree (2015). Specifically, we placed brains in sodium citrate solution for 6 minutes before fixation. After fixation and drying of slides, we applied the Vectashield DAPI mounting medium. We performed DNA fluorescence in situ hybridization (FISH) on vir00-Yfus, which allowed us to confidently identify the Y chromosome based on its unique satellite DNA composition. We used the same fixation and staining protocol as Flynn et al. (2020). We imaged metaphase cells using a 100 × oil objective on an Olympus fluorescent microscope and Metamorph capture system at the Cornell Imaging Facility.

vir00-Yfus Y-autosome fusion validation

We designed an experiment that would both validate the Y chromosome fusion (Fig. 1a) and to distinguish which autosome is fused. We first designed primers flanking microsatellite loci on all four autosomes that met the following criteria: (1) had 100% conserved nonrepetitive and unique priming sites between D. virilis and D. novamexicana; (2) amplicon length differed between the species by at least 15 bp as to be distinguished on an agarose gel; (3) locus contained a di or trinucleotide repeat; and (4) locus length ∼200 bp. We next set up a two-generation crossing scheme (Fig. 1d). We crossed D. novamexicana virgin females with vir00-Yfus males and selected the male progeny, which we backcrossed to D. novamexicana virgin females. We then genotyped the male F2 progeny from this cross at the 4 sets of primers corresponding to the four non-dot autosomes (Chr2, 3, 4, and 5) (Supplementary Table 2). We performed single-fly DNA extraction in strip tubes with Tris-EDTA buffer and 0.2 mg/mL proteinase K. We did 12 μL standard PCR reactions (3 min at 95, 30 cycles of 30 sec 95, 30 sec 55, 50 sec 72, and final extension 5 min). Each primer on each PCR plate had a homozygous (D. novamexicana) and heterozygous (D. novamexicana–D. virilis F1 hybrid) control. We then ran the PCR product on 2.5% agarose gels. If there was indeed an autosome fused to the Y chromosome, we would expect to see 100% of the male progeny being heterozygous for the virilis and novamexicana alleles (except for rare cases of nondisjunction). For the autosomes that are not fused, we would expect to see 50% of the progeny being homozygous for the novamexicana allele, and half heterozygous, due to Mendel's law of random segregation. We successfully validated the existence of the Y fusion, and found that it is fused to chromosome 3 (Muller D) (Supplementary Table 2 and Fig. 1d). There were 3 male progeny that were homozygous for the Chr3 novamexicana allele. We verified that these were cases of nondisjunction (opposed to the Y fusion not being fixed in this subline) by finding that the Y chromosome was absent in controlled Y chromosome PCR assays (Supplementary Fig. 5).

Fig. 1.

Discovery of three independent fusion events in vir00, and genetic validation of two of them. All metaphase images are of male larval neuroblasts. a) DNA FISH image staining three satellites and demonstrating the Y fusion in vir00-Yfus. b) DAPI staining image demonstrating an X fusion in vir00-Xfus-a. c) DAPI staining image demonstrating a different X fusion in vir00-Yfus + Xfus-b. d) Two-generation crossing experiment to validate the Y-3 fusion in vir00-Yfus and e) the X-4 fusion in vir00-Xfus-a. Red chromosomes, Drosophila novamexicana; orange, wildtype Drosophila virilis autosomes; green, autosome containing a GFP marker; light blue, X chromosome; dark blue, Y chromosome.

Isolation of vir00-Xfus-a and vir00-Nofus

The first X fusion was found to be segregating with a no fusion substrain in the 2019 stock of vir00. We wanted to isolate these into two separate substrains where the karyotype is fixed. From the progeny of the three original crosses in which we found the X fusion, we made 10 single-pair crosses and did neuroblast squashes of 6–8 larval progeny per cross, including both sexes. By chance, we should be able to find a cross in which the mother had two copies of the fusion and the father had a single copy—in which the derived line would be fixed for the fusion. If all progeny imaged contained the fusion (and females contained two copies of the fusion), then it is likely that this was the case. We isolated this line, and called it vir00-Xfus-a. We maintained a line isolated from the 2019 stock that had low frequency of the X fusion (∼25%), and called it vir00-Nofus/Xfus-a. We used this strain for several experiments (Supplementary Table 1). By the time we completed the nondisjunction assays, we had isolated a pure Nofus substrain (vir00-Nofus, Supplementary Table 1), so we only performed nondisjunction assays on fixed versions of the vir00 substrains.

vir00-Xfus-a X-autosome fusion validation

We obtained transgenic strains with GFP (or Blue) insertions which are expressed in the eye and larval brain from the National Drosophila Species Stock Center (vir95, vir121, and vir117). Stern et al. (2017) found the insertion sites of these lines. We chose lines which contained the GFP marker on candidate autosomes Chr2, Chr4, and Chr5. Chr3 was not a candidate because it is fused to the Y in vir00-Yfus and contains a different centromeric satellite. Before setting up crosses, we screened 10–20 larvae of each line with a fluorescent microscope to ensure that the transgene had not drifted to low frequency. Larvae containing the transgene demonstrated the GFP signal in their brain. We chose to phenotype at the larval stage since we would be crossing GFP strains to wildtype red-eyed flies, and the visibility of GFP in the adult eye would be low. We then designed a crossing scheme which would allow us to both validate that the X chromosome was fused and distinguish which autosome it was fused to (Fig. 1b, e). We crossed GFP-line males to vir00-Xfus virgin females. We then selected the male F1 progeny and crossed them to virilis genome strain virgin females. We then phenotyped F2 larvae, classifying each as either GFP positive or negative. When the phenotyped flies emerged, we sexed and counted them. If the candidate autosome is fused to the X, we would expect sex to segregate with the GFP marker: all female progeny will be GFP negative, and all male progeny will be GFP positive (except for phenotyping errors or rare nondisjunction events). For all other lines, sex should not segregate with GFP status. We performed negative control crosses in which the parental cross was replaced by genome strain virgin females, to ensure the crossing scheme produced the expected results (Supplementary Table 3). We found that the X chromosome is fused to Chr4 (Muller B).

Validation of the three versions of vir00 with private fixed indels

We first resequenced the same sequencing libraries produced from Flynn et al. (2020) to obtain high enough coverage for genotyping. Raw data, including the original sequencing and the resequencing, are available at NCBI SRA PRJNA548201. We then used genome analysis toolkit (GATK) recommended practices to do genotyping of our low-coverage whole-genome sequencing data. We used vcftools to subset singletons present only in vir00, which was, in hindsight, vir00-Yfus. We then used GATK's SelectVariants to select only nonreference homozygous indels 12 bp or more with a depth of at least 10 in vir00 and at least 2 in the other strains. We then manually inspected each potential candidate in IGV to ensure: no reads in other strains supported the indel, all reads in vir00 supported the indel, and there were no nearby indels in other strains. We then designed primers for the four loci (on Chr 2, 3, 5, and 6) that met these criteria and also had enough SNP-free sequence flanking for the indel in order to design primers that would amplify a locus 100–200 bp equally in all strains (Supplementary Table 5). We performed PCR and gel electrophoresis (2.5% gel, 98 V, 90 min). Later, when we obtained sequencing data for other vir00 substrains, we confirmed that substrains of vir00 were very closely related by using SNPhylo (Lee et al. 2014) to estimate the relative genetic divergence within vir00 substrains and between vir00 and the genome strain (Supplementary Fig. 8).

DNA damage assays

We chose two stressors that would moderately increase the rate of DNA breaks and allow us to potentially detect differences between strains. Gemcitabine is a nucleoside analog that induces replication stress by stalling polymerases, and also sensitizes cells to radiation via the RAD51 pathway (Kobashigawa et al. 2015). We selected dosage and a fly-feeding regime based on Kislukhin et al. (2012). Ionizing radiation has long been used to increase the rates of DNA breaks in flies for mutagenesis. We chose a dose 1/4–1/2 of what has been typically used in mutagenesis (Carlson and Southin 1962).

We collected male flies 0–1-days old and fed them with gemcitabine (0.718 mM) mixed with liquid food in vials with 8–12 adult flies as in Kislukhin et al. (2012). Liquid food consisted of 12.5 g sucrose, 17.5 g dry yeast, 5 mL corn syrup, and 95 mL PBS (autoclaved for 30 min immediately after adding the yeast). Flies were fed with the drug for 7–9 days before radiation. Controls were fed with the same liquid food for 7–9 days, except no gemcitabine was added. Flies were moved to fresh vials every 3–4 days. For radiation treatment, we transferred flies into 50 mL conical tubes with 5 mL agar because these tubes were compatible with the radiation source. Control flies were also transferred to new tubes. We used a J.L. Shepherd & Associates Mark I Irradiator with 1,100 Ci of Cs-137, and flies were irradiated at approximately 400 rad/min for a total of 10 Gy. In one case, for the GDvir stress treatment, the radiation was not stopped on time so 4 extra Gy was applied. We believe that this did not affect our results, especially because the GDvir strain had the lowest DNA damage increase with gemcitabine and radiation stress. We did comet assays to measure DNA damage (Angelis et al. 1999) over three different dates (Supplementary Table 7), but ensured experimental conditions were practically identical each time. For some samples, we had to combine results from two different dates to have enough nuclei for statistical analysis (Supplementary Table 7). We dissected testes from approximately 8 flies from each treatment within 1 hour of radiation treatment to minimize the opportunity for breakage repair (Shetty et al. 2017). We then homogenized the testis tissue using a Dounce, filtered the homogenate through a 40-micron filter to remove debris, and centrifuged and resuspended the cells to approximately 10⁵ cells/mL.

We next performed the alkaline comet assay as directed by the Enzo comet kit (ADI-900–166), which provides higher sensitivity than the neutral comet assay (Angelis et al. 1999). We imaged slides on a Metamorph imaging system at 10 × magnification using a fluorescent green filter to detect the CyGreen dye included in the comet kit. We quantified damage levels using the software OpenComet as a plugin in ImageJ (Gyori et al. 2014). We filtered scored nuclei that were not comet shapes or contained background interference. We used the measure of “olive moment,” which is the product of the percent of DNA in the tail and distance between intensity-weighted centroids of head and tail (Gyori et al. 2014) as the statistic to compare between strains and treatments.

Resequencing sublines to determine differences in their satellite abundance

Pools of 6 male flies were DNA extracted with Qiagen DNeasy blood and tissue kit. PCR-free libraries were then prepared with Illumina TruSeq PCR-free library prep. Libraries were sequenced on a NextSeq 500 single-end 150 bp. We removed adapters and poly-G signal with fastp and then ran k-Seek to count satellite abundances (Wei et al. 2014). We used average read depth to normalize the kmer counts. We also mapped the reads to the D. virilis rDNA consensus sequence (Stage and Eickbush 2007) to estimate the rDNA copy number in the three vir00 substrains as well as a vir08 as a control (Supplementary Tables 4, 8).

Using sequencing data to estimate the age of the Y-3 fusion

Scripts for this section are available here: https://github.com/jmf422/D-virilis-fusion-chromosomes/tree/main/simulate_degradation. We used the sequencing data from Flynn et al. (2020) and the resequencing of the same strains, in addition to data produced here for vir00-Yfus, vir00-Xfus-a, and vir00-Nofus/Xfus-a and 10 other D. virilis strains. All raw data are available at NCBI SRA PRJNA548201. We mapped the data to the RS2 genome assembly using bowtie2. We then genotyped with GATK following standard procedures (McKenna et al. 2010). We extracted heterozygous singleton sites for vir00, and counted how many occurred on each autosome. We calculated the enrichment on Chr3 in vir00-Yfus based on the difference from the average SNP density on the other autosomes (excluding the dot chromosome Chr6). To determine whether this enrichment of SNPs was significant based on the size of the chromosome and the number of mutations, we randomly permuted the total number of heterozygous singleton SNPs on all autosomes and calculated the proportion falling on Chr3, and repeated this 1,000 times.

We then performed simple simulations to determine approximately how many generations of mutation accumulation without recombination or selection would result in the enrichment we observed. Since heterozygous singletons are challenging for the genotyper to call with moderate coverage sequencing data, we incorporated this into our simulation. First, we made the genome assembly diploid then used mutation-simulator (Kühl et al. 2020) to simulate random mutations (transition/transversion ratio 2.0) at a rate of 2 × 10⁻⁹ per bp per generation for 500, 1,000, 2,000, and 5,000 generations on one copy of Chr3 only. We then simulated Illumina reads with ART (Huang et al. 2012) at the same depth as we have for vir00-Yfus in our real data (23×haploid or 11.5×diploid). We next used standard GATK genotyping and selected out heterozygous singletons on Chr3. We repeated the simulation 10 times for each number of generations to get a range of values. The empirical enrichment fell in between what we found in the simulations for 1,000 and 2,000 generations.

Nondisjunction assays

We crossed a single male from vir00-Yfus, vir00-Xfus-a, vir00-Nofus (fixed), and GDvir (control) to one or two GDvir (genome strain vir87) females. We collected the virgin progeny from each cross, extracted DNA with a squish-proteinase K prep, and genotyped with PCR and gel electrophoresis for the presence or absence of the Y chromosome in up to 16 female and 16 male progeny. We amplified a locus unique to the Y chromosome (primers designed by Yasir Ahmed-Braimah for a different project, Supplementary Table 5). For a subset of individuals, we also performed multiplex controls with an autosomal locus. Otherwise, we performed DNA extractions in large batches with the same proteinase K mixture to minimize the chance of DNA extraction failure. A very small quantity of DNA is required for a standard PCR with robust primers. To control for the completeness of the PCR mastermix, we included male samples in the same batch as female samples. If a male lacked a Y chromosome, we inferred that the father's sperm was missing the Y chromosome (nullisomic), and if a female contained a Y chromosome, we inferred that the father's sperm contained both X and Y. We used R prop.test to evaluate whether there were any differences between nondisjunction proportions for the different strains. After finding this highly significant, we used pairwise.prop.test in R with Holm–Bonferroni multiple test correction to determine which pairs of substrain nondisjunction rates were significantly different from each other.

Results

Three novel sex chromosome Robertsonian translocations in D. virilis strain vir00

In summer 2019, we performed DNA FISH on larval neuroblast nuclei of the vir00 strain that we had obtained in Fall 2018 from the National Drosophila Species Stock Center. We discovered that it contained a Y-autosome fusion (Fig. 1a). The Y chromosome is recognizable in D. virilis because it has a distinct DAPI staining intensity pattern, and contains a distinct arrangement of satellite DNA (Flynn et al. 2020). However, all other chromosomes are difficult to distinguish in metaphase spreads, so we could not immediately determine the fusion partner. We observed that the fused chromosome contained the same centromere-proximal satellite as the Y chromosome, AAACTAT (staining faintly with DAPI). To determine if this chromosome fusion was present in other strains from similar geographical locations, we imaged larvae from strains vir08, vir86, and vir48, which were collected from different localities in California and Mexico (vir00 was collected from California). All larvae screened from these other strains contained the same karyotype as the D. virilis genome strain (Flynn et al. 2020). We designate the vir00 substrain with the Y fusion as vir00-Yfus.

In Fall 2019, we obtained a second copy of vir00 from the National Drosophila Species Stock Center. We set up single-pair crosses and performed larval neuroblast squashes to karyotype multiple male progeny of each cross. Surprisingly, we found that the Y-autosome fusion was not present in any of the larvae karyotyped. However, 3/10 crosses karyotyped contained a different fusion (Fig. 1b). This new fusion contained the bright-DAPI-staining AAATTAC satellite of both ancestral centromeres, indicating that it is completely distinct from the Y fusion and involved different chromosomes. We hypothesized that this new fusion involved the X chromosome for several reasons; (1) based on centromere satellite arrangement, it had a 67% probability (Supplementary Fig. 1); (2) it was found only in single copy in male larvae, but sometimes, two copies in female larvae; and (3) we observed karyotypes arising from X–Y nondisjunction events in most crosses from this line, such as an XXYY female (Supplementary Fig. 2). We performed single-pair crosses and screened the resulting progeny until we isolated a substrain fixed for the X fusion, and called this substrain vir00-Xfus-a. We retained the stock that contained vir00-Xfus-a segregating with a no-fusion karyotype (vir00-Nofus/vir00-Xfus-a). Additionally, we later isolated a substrain fixed for the no fusion karyotype, vir00-Nofus, by screening single-pair crosses as described above.

In Fall 2021, we obtained a third copy of the vir00 stock from a colleague who had obtained it from the stock center circa 2016 (Yasir Ahmed-Braimah, personal communication). The original purpose of checking this stock copy was to aid in inferring the timeline of the chromosome fusion events, specifically whether the Y fusion was ancestral in the vir00 lineage. We confirmed that this stock contained the above-described Y fusion, but it also had an additional and different fusion segregating, present in about 50% of the larvae imaged (Fig. 1c). The new fusion contained the bright-staining AAATTAC satellite on only one of the two ancestral Robertsonian centromeres, in contrast to the other two fusions we previously identified. We performed single-pair crosses and screened the resulting progeny until we isolated a substrain fixed for this new fusion. We also inferred that this fusion involved the X chromosome as well, since once fixed, there were always two copies of the fusion in females (Supplementary Fig. 3), and one copy in males (Fig. 1c). Therefore, we designate this substrain as vir00-Yfus + Xfus-b. We inferred that all three fusions we discovered likely represent canonical Robertsonian translocations, in which two acrocentric chromosomes that underwent DNA breakage were fused together at the centromere during repair. We did further experiments on multiple versions of vir00, summarized in Supplementary Table 1.

Genetic validation revealed Y-3 and X-4 fusions

We designed two separate two-generation crossing experiments to validate the initially-discovered fusions (vir00-Yfus and vir00-Xfus-a) and identify the autosome each sex chromosome is fused to. Both experiments exploited autosomal markers which we used to determine whether they were segregating non-independently of sex. The first experiment to validate the Y-autosome fusion used crosses between vir00-Yfus and D. novamexicana and scoring of polymorphic microsatellite loci on each candidate autosome. We scored 82 F2 male progeny for the Chr3 marker, and 79/82 contained both alleles, so the null hypothesis of 50% is rejected at P < 0.00001 with a Chi-square test (Fig. 1d and Supplementary Fig. 4). We determined that the three progeny that did not contain both alleles (Fig. 1d observed-homozygous) did not contain a Y chromosome because of a nondisjunction event (Supplementary Fig. 5). We concluded that the Y chromosome is fused to Chr3. We also inferred that the Y-3 is fixed in this substrain since no Y chromosomes in the 82 male progeny we assayed segregated independently of Chr3. The other markers on Chr2, Chr4, and Chr5 segregated independently of sex and acted as negative controls (Supplementary Table 2). We also did a negative control with the same crossing scheme except with vir08 instead of vir00 (Chr3 Chi-square P = 0.39, N = 22, Supplementary Table 2).

The second experiment validated the X-autosome fusion in vir00-Xfus-a by crossing the substrain to D. virilis transgenic lines containing eye-expressing GFP markers on one of the candidate autosomes. For the crosses to vir95 (which contained a GFP marker inserted on Chr4), we phenotype 304 F2 female progeny, and found that progeny containing a GFP signal were significantly depleted compared to the Mendelian expectation of 50% (Fig. 1e). This indicated that the X chromosome is fused to Chr4. The same crossing scheme against two other strains containing the GFP marker on Chr2 and Chr5 did not show association of GFP signal with sex (Chi-square P > 0.1, Supplementary Table 3). We also performed a negative control with the D. virilis genome strain instead of vir00-Xfus-a crossed to the vir95 line and found that the GFP signal was independent of sex (Chi-square P = 0.92, Supplementary Table 3).

We validated that the three fusions are the same genetic line and not the result of contamination from other lines either in our lab or the stock center. We made use of medium-coverage whole-genome sequencing from Flynn et al. (2020) to design primers to amplify singleton insertion/deletion variants present only in vir00 (the version that was sequenced was, in hindsight, vir00-Yfus) and not in any other wildtype D. virilis strain present in the stock center (Supplementary Table 4). We designed primers to amplify four loci on chromosomes 2, 3, 5, and 6 which contain a homozygous 12–13 bp deletion in vir00 compared to the reference and the other strains (Supplementary Table 5). We found that vir00-Yfus, vir00-Xfus-a, vir00-Nofus, and vir00-Yfus + Xfus-b all contained the deletion at each of these loci, supporting that these substrains indeed originated from the same strain vir00 (Supplementary Fig. 6). Although we performed the above indel experiment first, the whole-genome sequencing SNP analysis described in a later section was also concordant with these substrains being the same genetic line.

Nondisjunction between the X and Y chromosomes is 21% in vir00-Xfus-a

Nondisjunction occurs when homologous chromosomes fail to separate at meiosis, and results in aneuploidy in the progeny. Autosomal and X chromosome aneuploidy is lethal in flies; but Y chromosome aneuploidy is viable: females that have a Y chromosome (XXY) are fertile, males with no Y chromosome (XO) are sterile, and males with two Y chromosomes (XYY) are fertile. Elevated rates of X–Y nondisjunction represent a fitness cost because zygotes with infertile or lethal karyotypes will form at increased frequency. We found some evidence of nondisjunction in the vir00-Yfus genetic validation (3/82 males, Fig. 1d and Supplementary Fig. 5), and also common Y chromosome aneuploidy in the stock of vir00-Xfus-a (Supplementary Fig. 2). Since the fusions involve the sex chromosomes, we tested the rate of primary X–Y nondisjunction in males in vir00-Yfus, vir00-Xfus-a, and vir00-Nofus, along with the D. virilis genome strain as a control.

We crossed individual males of each strain we tested to genome strain females and genotyped for the presence or absence of the Y chromosome in progeny. Female progeny containing a Y chromosome indicate XY sperm from the father, and male progeny lacking a Y chromosome indicate nullisomic sperm from the father. Since nondisjunction was extremely high in vir00-Xfus-a, 2/7 fathers tested were of XYY karyotype, which we could infer if more than half of his female progeny contained a Y chromosome (Maggert 2014). We eliminated these fathers’ progeny from the primary nondisjunction rate calculation. No fathers tested from other sublines were determined to be XYY. Males without a Y chromosome would not produce progeny. We found that vir00-Yfus had a slightly elevated nondisjunction rate of 4.5%, compared to the genome strain control of 1.2%, but it was not statistically significant with the sample sizes we used (Table 1, P = 0.123). vir00-Xfus-a had an extremely high primary nondisjunction rate of 21% (e.g. Supplementary Fig. 7), which was significantly higher than that of all other substrains (P < 0.004, pairwise proportion test, Table 1). Surprisingly, vir00-Nofus had a significantly elevated nondisjunction rate compared to the genome strain control, at 5.7% (P = 0.004). This was not statistically different from vir00-Yfus (P = 0.231).

Table 1.

Nondisjunction of the X and Y chromosomes in males is elevated in chromosome fusion lines.

Line	Nondisjunction rate	Fathers with nondisjunction/total fathers	Aneuploid progeny/total progeny	Significance group (P < 0.004)
GDvir	1.2%	1/10	3/248	a
vir00-Yfus	4.5%	4/7	9/198	ab
vir00-Nofus	5.7%	4/6	11/192	b
vir00-Xfus-aa	21%	5/5	27/127	c

Only including primary nondisjunction.

Satellite DNA abundance has evolved between vir00 substrains

We used Illumina to resequence males from three vir00 substrains: vir00-Yfus, vir00-Xfus-a, and the mixed stock of vir00-Nofus/Xfus-a. We then used k-Seek to quantify the satellite abundance in each substrain (Wei et al. 2014). Satellite abundance was highest in vir00-Yfus compared to vir00-Xfus-a and vir00-Nofus/Xfus-a (Supplementary Table 6). vir00-Xfus-a had 8% less AAATTAC, which occupies the centromere region of ChrX and Chr4. Overall, vir00-Nofus/Xfus-a contained 12% less pericentromeric satellite DNA compared to vir00-Yfus, with 10% less in pericentromeric AAACTAC, 8% less in AAATTAC, and 13% less in AAACTAT. Given the difference between vir00-Nofus/Xfus-a and the pure vir00-Xfus-a stock, we infer that satellite DNA would be even less abundant in vir00-Nofus but unfortunately, the X fusion was still segregating with the unfused chromosomes at the time of sequencing. We use these data to map the satellite abundance evolution to the timeline of chromosome fusion and fission events, described in the section below. We expect that chromosome fusion/fission events can result in losses of satellite DNA, so the genome ancestral to all the fusion events would contain the highest amount of satellite DNA.

Estimated timeline of chromosome fusion events

Since we discovered three different chromosome fusion events in the vir00 stock at different times, we propose here the most likely and parsimonious timeline of events (Fig. 2). We used the data we had available, including karyotype, nondisjunction rate, and SNP data from resequencing analysis. Since the Y-3 fusion was fixed in both the pre-2016 (vir00-Yfus + Xfus-b) and 2018 (vir00-Yfus) substrains, we inferred this fusion likely occurred ancestrally. We used resequencing data to estimate how old this fusion may be. The Y-fused version of Chr3 is expected to accumulate mutations independently of the autosomal version of Chr3 over time because of the halt in recombination in male flies. This would be represented by elevated heterozygosity on Chr3 as assayed by aligning short read sequencing data (since reads from both the Y-fused and autosomal copy will map to the same copy of Chr3 in the reference genome). Specifically, if the mutations arose after the fusion of Y-3, they would not be present in any other strains of D. virilis (assuming no recurrent mutation). Thus, we used our sequencing data for SNP genotyping and found heterozygous singletons unique to vir00-Yfus compared to other non-vir00 strains on each autosome. We found that the number of heterozygous singletons was modestly but significantly enriched on Chr3 in vir00-Yfus: 2.45 SNPs/Mb more than other autosomes (permutation test, P < 0.001). We suggest that this elevated density of singleton heterozygous sites may be due to the fusion with the Y chromosome and lack of recombination over several generations. Assuming enrichment was caused by only the Y chromosome and that the neo-Y (Chr3) evolved clonally in a single lineage with single nucleotide mutation rate of 2 × 10⁻⁹ per nucleotide per generation, we estimate with simulations that the Y-3 fusion occurred 1,000–2,000 generations ago. It is difficult to determine whether the Y-3 fusion occurred in the wild or the lab. There is no recorded collection date for vir00 (from Pasadena, California, USA). But assuming it was collected at a similar date to other virilis group strains with recorded collection dates in North America (∼1950), and it was passed 15 generations per year in the lab, the chromosome fusion likely occurred very close to the collection date (15 generations × 70 years = 1,050 generations, within our estimate).

Fig. 2.

Proposed timeline of chromosome fusion events illustrated on a cladogram (not to scale). The Y-3 fusion, shown by red bars and arrows, appears to be ancestral to the strain and likely occurred 1,000–2,000 generations ago. The X-Unknown (Xfus-b) fusion is shown by purple bars and arrows, and the X-4 (Xfus-a) fusion is shown by blue bars and arrows. Green text describes satellite abundance changes. Gray letters and arrows label chromosomes: “D” represents the dot chromosome (chromosome 6 in Drosophila virilis); “B” represents chromosomes 2, 4, and X which have bright DAPI staining at the centromere and are difficult to distinguish; and “NB” represents chromosomes 3 and 5 which have nonbright DAPI staining at the centromere and are difficult to distinguish (Marchetti et al. 2022). See Supplementary Fig. 1 for a schematic representation of the D. virilis chromosomes.

Because vir00-Xfus-a was not fixed in the substrain we discovered it in (it was segregating with vir00-Nofus), recombination in heterozygotes would prevent degeneration of the unfused (neo-Y) version of Chr4. Furthermore, we believe that the X-4 (Xfus-a) fusion occurred between 2018 and 2019 since it was not present in the stock we obtained in 2018 or from the pre-2016 stock. We did not observe an enrichment of heterozygous singletons on Chr4 in vir00-Xfus-a. However, there was still a slight enrichment on Chr3 (1.45 SNPs/Mb, permutation test, P = 0.031), supporting our assumption that the X-4 fusion occurred after the Y-3 fusion broke apart in the same lineage (Fig. 2). Furthermore, vir00-Nofus contained a significantly elevated nondisjunction rate of 5.7%, which was surprising since we expected this result only for substrains with chromosome fusions. We think that the high nondisjunction rate could be caused by an additional aberration that occurred on the same branch that the Y-3 broke apart. Further studies will be required to test this and determine the nature of the aberration. Although we did not measure the nondisjunction rate in vir00-Yfus + Xfus-b (Supplementary Table 1), it did not appear as high as vir00-Xfus-a since only one of five crosses we checked had any Y chromosome aneuploidy, supporting its more basal position on the cladogram. An alternative possibility, a tweak to our proposed timeline, is that vir00-Nofus was segregating in the stock undetected since the beginning, and that the Y-3 fusion was lost by drift rather than fission in 2018–2019. Indeed, centromeric fissions are thought to be rare (Piálek et al. 2005). However, we think that the long-term segregation of vir00-Yfus and vir00-Nofus is unlikely because the maintenance of fly stocks involves small population sizes which makes it more probable that alleles will be fixed or lost. Genetic divergence between substrains of vir00 measured by SNPs was low (9.76×lower than between vir00 and the virilis genome strain; Supplementary Fig. 8). Our data also suggest that the Y-3 fusion existed for the longest time, with significant SNPs accumulated on Chr3, making it unlikely that an unfused allele would persist segregating for 1,000–2,000 generations. Additionally, the Y fusion was completely fixed in both the stock copies we received containing it, whereas both the X fusions were segregating.

DNA damage levels in response to stress is higher in vir00-Yfus than derived substrains

The unusual commonness of chromosome fusions in vir00 could be caused by an increase in the probability of any (or multiple) of the four steps required to detect a Robertsonian translocation outlined in the Introduction. We decided to explore the idea that genetic variation in satellite abundance could contribute generally to the rate of DNA breakage after stress, the fundamental first step that could lead to a Robertsonian translocation or other large-scale rearrangement. We therefore measured DNA damage levels in 7 different D. virilis strains, including vir00-Yfus and vir00-Nofus/Xfus-a, with varying abundances of satellite DNA in response to replication stress (induced with the drug gemcitabine) and low-level radiation (10 Gy). For each of seven strains tested, we included a control which was fed with the same liquid food with no gemcitabine and did not receive radiation treatment. We used the comet assay or single-cell gel electrophoresis to measure DNA damage in the male germline in each line, treatment and control (Fig. 3). We used the metric “olive moment” to quantify the proportion of damaged DNA per nucleus, which is based on comet shape and DNA staining intensity (Supplementary Fig. 9).

Fig. 3.

DNA damage levels and satellite abundance in males of several Drosophila virilis strains. a) Olive moment, a statistic of the comet assay, measuring DNA damage for 7 different strains paired with a control and stress treatment. N indicates the number of nuclei analyzed for each treatment. Flies treated with gemcitabine and radiation are suffixed with “rad”, and control flies are suffixed with “con”. The lower panel shows the unpaired mean difference between “rad” and “con” for each fly strain (black dot) and 95% confidence intervals (black line) produced from dabestr (5,000 bootstrap method). Groups a, b, and c indicate samples with overlapping confidence intervals. b) The satellite DNA abundance of each D. virilis strain (x axis) and its unpaired mean difference between treatment and control (y axis). vir00-Nofus/Xfus-a experienced a 12% loss in satellite DNA compared to vir00-Yfus, which is associated with a significantly decreased DNA damage in response to stress.

The strain with the lowest satellite abundance, the genome strain, contained the lowest DNA damage response, and this was significantly lower than all other strains tested (Fig. 3a). The strain with the highest satellite abundance, vir00-Yfus, had among the highest DNA damage responses, but was not significantly different from two other strains vir8 and vir48 (Fig. 3a). Like most other phenotypes, there are likely multiple genetic factors contributing to the variation in DNA damage we found. Ideally, to demonstrate that satellite DNA is a causal factor, we would manipulate satellite DNA abundance and test the DNA damage phenotypes, but multimegabase long arrays of satellite DNA cannot be manipulated with traditional genome editing. However, vir00-Yfus and vir00-Nofus/Xfus-a are expected to have an almost identical genetic background except for a chromosome fusion difference and a satellite abundance difference. Thus, if there is a difference in DNA damage response between these substrains, it may be caused by differing satellite abundance. vir00-Nofus/Xfus-a, which contained 12% less satellite DNA than vir00-Yfus, had a significantly reduced DNA damage response, concordant with our expectations that satellite DNA plays a role (Fig. 3b), in addition to other factors.

Discussion

Here, we report the discovery of a D. virilis strain that has undergone an unusual number of chromosome fusion (and likely fission) events very recently. Chromosome rearrangements such as Robertsonian translocations are rare occurrences, but are often the source of karyotype evolution between species (Mayrose and Lysak 2021). Our study adds to the small collection of metazoans found to have within-species polymorphic chromosome fusions, for example: Drosophila birchii (Baimai 1969), Oryzomys (Koop et al. 1983), Mus musculus domesticus (Piálek et al. 2005), Leptidea sinapis (Lukhtanov et al. 2011), Salmo salar (Brenna-Hansen et al. 2012), and Drosophila melanogaster (Li et al. 2022). In our case though, we were able to observe karyotype changes in real-time on single-year timescales. In humans, rearrangements resulting from breaks near the centromere are associated with miscarriages and in the case of somatic rearrangements, cancer (Barra and Fachinetti 2018). Previous studies have found repetitive element involvement in chromosome rearrangements in multiple species (Paço et al. 2015; Reis et al. 2018). Here, we suggest that satellite abundance may influence the risk of DNA damage and chromosome rearrangements. There are several possible mechanisms that might cause excess satellite DNA to increase the risk of DNA breakage. Satellite DNA may form complex structures, loops, or non-B DNA, and increasing the length of the arrays may make these regions more unstable in cis (Barra and Fachinetti 2018). It may be challenging for polymerases to replicate many megabases of tandem satellite sequence, and longer arrays may have a higher risk of stalling polymerases, and thus, DNA breaks in cis (Barra and Fachinetti 2018). Finally, increased satellite DNA may titrate away binding proteins that maintain genome stability, in trans (Francisco and Lemos 2014; Brown et al. 2020a; Giunta et al. 2021).

Our study has several limitations. Firstly, the timeline of unexpectedly discovering different karyotypes in vir00 at different times resulted in some substrains not having data collected for all experiments (Supplementary Table 1). For example, vir00-Nofus/Xfus-a was segregating in our sequencing and comet assay experiment, but vir00-Nofus was isolated for the nondisjunction assay. vir00-Yfus + Xfus-b was discovered very recently and only strain validation was able to be completed at this time. Furthermore, our experiment demonstrating a higher damage level between the control and stressed flies in vir00-Yfus may not be directly applicable to the risk of DNA breaks and genome instability in natural conditions. Vir48 and vir08 had a similar olive moment difference to vir00-Yfus, but did not exhibit chromosome fusions, suggesting other factors not studied here playing an important role, possibly at other steps. Although we find a difference in the DNA damage in response to stress between vir00 substrains with different abundances of satellite DNA, their fusion status is also different. We cannot eliminate the possibility that the presence of the Y fusion itself increased the rate of DNA damage instead of the abundance of satellite DNA. Additionally, factors affecting the maintenance and rise in frequency of the fusions may also be important to our discovery, but we did not investigate this here. D. virilis in general has the ancestral karyotype of Drosophila and has not undergone any chromosome fusions since the genus evolved, despite having among the highest satellite contents. This could suggest that the probability of breaks and fusions only increases after a certain threshold of satellite abundance; and other D. virilis lineages whose satellite content expanded above the threshold became extinct because of negative consequences of the fusions in the wild. Furthermore, other factors such as variation in nonsatellite parts of the genome or selection favoring sex chromosome fusions may have additionally contributed to our observations in vir00.

We identified elevated nondisjunction in some vir00 substrains. Surprisingly, the rate of nondisjunction was 5-times higher in vir00-Xfus-a compared to vir00-Yfus. With a nondisjunction rate of 21%, a high proportion of abnormal karyotypes are produced, such as XO (sterile), XYY (viable and fertile), and XXY or XXYY (viable and fertile), all of which we found in cytological samples. Further mating between these abnormal karyotypes will produce significant proportions of sterile or inviable karyotypes like XYYY or XXX, which will further decrease the fitness of this line. Furthermore, karyotypes with extra Y chromosomes such as XYY and XXY have been found to have decreased lifespan (Brown et al. 2020b). The extreme nondisjunction rate vir00-Xfus-a but not in vir00-Yfus indicates that there may be additional rearrangements on the X and/or Y that occurred during fission of the Y-3 and disrupted proper pairing and segregation of the sex chromosomes. Furthermore, the nondisjunction rate in vir00-Nofus is still significantly higher than the genome strain, despite the same karyotype. We believe that this indicates a remaining rearrangement in vir00-Nofus affecting the pairing and/or disjunction of the X–Y. We found only a modest difference in estimated rDNA copy number between the substrains (Supplementary Table 8), which has been found to mediate pairing of the X and Y (McKee and Karpen 1990). Detailed analysis of structural rearrangements in the heterochromatin along with cytological analysis will be required to determine the mechanism of the elevated nondisjunction rates.

We have isolated four different versions of the strain vir00, each with a different karyotype: vir00-Nofus, vir00-Xfus-a, vir00-Yfus, and vir00-Yfus + Xfus-b, which will be useful for a variety of future studies. The vir00 fusion substrains will be useful for studying centromere identity. In both the Y-3 and X-4 fusions, two spherical regions of satellite DNA are present at the centromere of these fusions, representing one from each acrocentric chromosome (Fig. 1). We note that the X-4 fusion in vir00 is homologous to an independent X-4 fusion in D. americana 29 thousand years ago, a species only 4.5 million years diverged. In the X-4 fusion of D. americana, there is only one discrete region of centromeric satellite (Flynn et al. 2020), unlike what we found here. In female meiosis where chromosomes can compete to get into the oocyte rather than the polar body, “stronger” centromeres may have an advantage (Malik and Bayes 2006). A “supercentromere” resulting from a centromere–centromere fusion is one possible way to do this, and in D. americana, the X-4 fusion has biased transmission into the egg (Stewart et al. 2019). Furthermore, all the fusions we discovered involve sex chromosomes and create neo-sex chromosomes, to our knowledge, the most recently formed neo-sex chromosomes reported. We recommend to those who study this strain in the future to periodically (e.g. annually) check the karyotype in ∼10 individuals from the same substrain. We also recommend maintaining multiple independent replicates of each substrain, in case of any karyotype changes in an individual replicate. If a new chromosome fusion is discovered, we recommend performing single-pair crosses to fix the fusion as soon as possible, to eliminate the chance of a polymorphic fusion being lost. In conclusion, here we present a novel system with multiple recent chromosome fusions, providing a unique opportunity to study genome and karyotype evolutionary dynamics.

Data availability

Fly strains are available upon request. Sequencing data used in this paper are available at NCBI SRA PRJNA548201. Scripts used to analyze data and produce figures are available at https://github.com/jmf422/D-virilis-fusion-chromosomes. Supplementary Table 4 lists all strains used in this study, including the full strain name and the abbreviated strain name. Primer sequences used in this study and their purpose are listed in Supplementary Table 5. Supplementary Table 7 contains the DNA damage metric (olive moment) measurements used for statistical analysis in Fig. 3a.

Supplemental material available at GENETICS online.

Funding

This work was supported by the National Institutes of Health (NIH) R01 grant number GM119125 to AGC and Daniel Barbash.

Conflicts of interest statement

The author(s) declare no conflict of interest.