2023 Volume 98 Issue 4 Pages 201-206
Many organisms with heteromorphic sex chromosomes possess a mechanism of dosage compensation (DC) in which X-linked genes are upregulated in males to mitigate the dosage imbalance between sexes and between chromosomes. However, how quickly the DC is established during evolution remains unknown. In this study, by irradiating Drosophila miranda male flies, which carry young sex chromosomes (the so-called neo-sex chromosomes), with heavy-ion beams, we induced deletions in the neo-Y chromosome to mimic the condition of Y-chromosome degeneration, in which functional neo-Y-linked genes are nonfunctionalized; furthermore, we tested whether their neo-X-linked gametologs were immediately upregulated. Because the males that received 2-Gy iron-ion beam irradiation exhibited lower fertility, we sequenced the genomes and transcriptomes of six F1 males derived from these males. Our pipeline identified 82 neo-Y-linked genes in which deletions were predicted in the F1 males. Only three of them showed a one-to-one gametologous relationship with the neo-X-linked genes. The candidate deletions in these three genes occurred in UTRs and did not seriously affect their expression levels. These observations indirectly suggest that DC was unlikely to have operated on the neo-X-linked genes immediately after the pseudogenization of their neo-Y-linked gametologs in D. miranda. Therefore, the dosage imbalance caused by deletions in the neo-Y-linked genes without paralogs may not have effectively been compensated, and individuals with such deletions could have exhibited lethality. Future studies on sex chromosomes at different ages will further reveal the relationship between the age of sex chromosomes and the stringency of DC.
Sex chromosomes are thought to have originated from a pair of autosomes (Vicoso, 2019). Meiotic recombination between the X chromosome (termed X henceforth) and the Y chromosome (termed Y henceforth) is then suppressed in many cases, to maintain stable sex determination, which results in massive pseudogenization of genes on the Y, with the exception of the genes involved in male determination and sexual antagonism (Charlesworth et al., 2005). Because the losses of Y-linked genes cause a dosage imbalance between sexes (i.e., one copy and two copies of X-linked genes in males and females, respectively), the X in many organisms developed a mechanism, the so-called dosage compensation (DC), to mitigate it (Ohno, 1967). In Drosophila melanogaster, for example, the protein–RNA complex termed male-specific lethal (MSL) globally recruits histone acetylation to the entire male X, thus triggering the doubling of the expression of X-linked genes in males (Lucchesi and Kuroda, 2015).
However, because many Y-linked genes remain functional at the early phase of sex chromosome differentiation, the global DC on the X-linked genes with Y-linked gametologs that are functional seems to cause overexpression of the genes. Thus, DC on X-linked genes may need to operate more locally only when their Y-linked gametologs are nonfunctionalized in the early stage of sex chromosome evolution. Young sex chromosomes, the so-called neo-sex chromosomes, which are formed by the fusion of an autosome with an ordinary sex chromosome, have been used to elucidate the early stage of sex chromosome evolution. One such neo-sex chromosome emerged in D. miranda about 1.1 million years ago via the fusion of its third chromosome with the Y (Steinemann and Steinemann, 1998; Bachtrog and Charlesworth, 2002). Previous studies reported that global DC via the MSL complex is already established on the neo-X chromosome (termed neo-X henceforth) in D. miranda (Alekseyenko et al., 2013; Zhou et al., 2013); however, the global DC is incomplete (Nozawa et al., 2018, 2021). These studies also found that the DC on the neo-X-linked genes with pseudogenized neo-Y-linked gametologs is greater than that on the neo-X-linked genes with functional neo-Y-linked gametologs (Nozawa et al., 2018, 2021). These observations suggest that not only global DC, but also more localized DC (termed gene-by-gene DC henceforth), operate on the neo-X by recognizing the functionality of neo-Y-linked genes in D. miranda, although the mechanism underlying this process and the immediacy of gene-by-gene DC remain unknown.
Therefore, in this study we assessed how quickly such gene-by-gene DC can be established during the evolution of sex chromosomes. For this purpose, we mimicked Y-chromosome degeneration by irradiating D. miranda with heavy-ion beams, which have been utilized as a mutagen and are known to induce larger deletions compared with X-rays and gamma-rays (e.g., Tanaka et al., 2010). This method has also been applied to D. melanogaster (Tsuneizumi and Abe, 2014; Tsuneizumi et al., 2017, 2018); however, to the best of our knowledge, the present study was the first to apply this method to other Drosophila species. Irradiation of D. miranda males with heavy-ion beams may cause the deletion of some genomic regions in their sperms; moreover, the F1 males derived from a cross of the irradiated males with wild-type females may carry deletions on the Y/neo-Y (and/or on autosomes, but not on the X or the neo-X, Fig. 1), which can be examined using whole-genome sequencing. If a deletion in a functional neo-Y-linked gene was detected in a F1 male, the expression level of the neo-X-linked gametolog in the F1 and the wild-type males was compared to test if the neo-X-linked gametolog was immediately upregulated in the F1 male.
First, we examined the effects of heavy-ion beam irradiation on male fertility in D. miranda using nonirradiated males as controls (Fig. 2, see also Supplementary Fig. S1 for the detailed scheme of the crossing experiments). The effects of three heavy-ion beams, i.e., iron (Fe), argon (Ar), and carbon (C) ions, were examined separately. Irradiation using the Fe-ion beam at 0.5 and 1 Gy was unlikely to affect male fertility severely, although the males that received 1-Gy irradiation showed a somewhat lower fertility compared with the control males at 4–6 days after irradiation (Fig. 2A, Supplementary Table S1). In contrast, the males that received Fe-ion beam irradiation at 2 Gy apparently showed lower fertility, particularly at 4–6 days after irradiation (Fig. 2A, Supplementary Table S1). Regarding the Ar-ion beam, male fertility was reduced at 7–9 days after irradiation (Fig. 2B, Supplementary Table S2), whereas the fertility of males was low immediately after 10-Gy irradiation with the C-ion beam (Fig. 2C, Supplementary Table S3).
Because the 2-Gy irradiation with the Fe-ion beam clearly reduced male fertility and the Fe-ion was the largest heavy ion among the three, which may have induced larger deletions compared with the Ar and C ions, the genomes of six F1 males derived from the males that received this irradiation condition were sequenced, to identify deletions (Supplementary Table S4). As controls, the genomes of four nonirradiated males were also sequenced. By comparing the mapping depth between the irradiated and the control genomes, candidate deletion regions were predicted. Briefly, the genomic regions that exhibited a mapping depth of zero in an irradiated male genome and of at least one in all nonirradiated male genomes were regarded as deletions (please refer to the “Prediction of deletions” subsection of the Supplementary Materials and Methods for details). The results of this experiment revealed the presence of deletions not only on the Y/neo-Y, but also on the neo-X and the X (Table 1). It should be noted that any deletion on the neo-X and the X caused by irradiation was unexpected, because these chromosomes in the F1 males were inherited from the nonirradiated females (Fig. 1). Therefore, our pipeline for detecting deletions produced false positives. We currently speculate that the indel variations on the neo-X and the X among individuals within a strain could explain this observation. This speculation is plausible because nucleotide polymorphism is expected to be three times greater on the X vs. the Y at equilibrium. In fact, the nucleotide polymorphism based on SNPs on the neo-X is much greater than that on the neo-Y (Bachtrog, 2004; Nozawa et al., 2018).
| Chr. | Chr. length (Mbp) | Number of candidate deletions | |||||||
|---|---|---|---|---|---|---|---|---|---|
| #13 | #2 | #3 | #4 | #5 | #6 | Avg.5 ± SE6 | per Mbp8 | ||
| Y/Neo-Y1 | 101.54 | 177 | 141 | 79 | 66 | 111 | 66 | 106.67 ± 18.45 | 1.05 |
| (70.52)2 | (35)4 | (28) | (11) | (13) | (23) | (5) | (19.17 ± 4.65)7 | (0.27)9 | |
| Neo-X | 25.30 | 28 | 22 | 11 | 18 | 24 | 9 | 18.67 ± 3.05 | 0.74 |
| (17.93) | (9) | (5) | (2) | (4) | (3) | (2) | (4.17 ± 0.98) | (0.23) | |
| X | 77.74 | 109 | 141 | 65 | 74 | 121 | 57 | 94.50 ± 13.87 | 1.22 |
| (53.65) | (32) | (24) | (17) | (12) | (26) | (11) | (20.33 ± 3.12) | (0.38) | |
An insufficient mapping depth may be another possible explanation for the detection of false positives (Supplementary Table S5). The mapping depth in each sample ranged from 14 to 23, and the proportion of the genome to which at least 10 reads were mapped was only 61%–80%. In this case, the depth for some genomic regions in a male may be zero, just by chance rather than because of deletions.
Nevertheless, our prediction also included true positives. The deletion density was higher for the neo-Y than the neo-X (Table 1), which was unexpected if most of the predicted deletions were caused by polymorphism. In addition, when we randomly selected seven regions in which deletions were predicted on the Y/neo-Y for PCR, cloning, and Sanger sequencing, we found that two out of the seven candidate regions were confirmed to contain deletions in regions located near our prediction (Supplementary Table S6, Supplementary Fig. S2; see also the “Confirmation of candidate deletions” subsection of the Supplementary Materials and Methods for details). It should be mentioned that repetitive sequences accumulated on the Y/neo-Y, which severely hampered the design of primers and rendered the off-target amplification inevitable. In this case, the sequencing of target regions was also difficult, even after cloning. Therefore, the accuracy of our pipeline was likely underestimated.
The candidate deletions were distributed throughout the Y/neo-Y in all F1 males (Supplementary Fig. S3), indicating that irradiation with the Fe-ion beam affected the entire chromosome more or less uniformly. The size of the largest candidate deletion on the Y/neo-Y was 1,568 bp, which was much smaller than the deletions found in plants, in which the deletion size is about several hundred kb on average (Hirano et al., 2015). It is known that plant species are, in general, highly tolerant of irradiation. In fact, individuals with large deletions (>600 kb at maximum) were obtained in Arabidopsis thaliana after ~400-Gy irradiation with a C-ion beam (Kazama et al., 2011, 2017). Notably, the genome size, number of genes, and number of chromosomes are comparable between Drosophila and Arabidopsis (https://www.ncbi.nlm.nih.gov/genome). Therefore, the performance of further trials using greater amounts of irradiation may be a worthwhile approach in the future to obtain large deletions in D. miranda.
Among the 9,775 genes located on the Y/neo-Y, 82 genes were predicted to contain deletions in at least one of the six F1 males, and candidate deletions within genes were found on average every 1.24 (101.54/82) Mbp on the Y/neo-Y. Among these 82 genes, only three genes exhibited a one-to-one gametologous relationship with the neo-X-linked genes (Table 2, Supplementary Fig. S3). The proportion of genes with deletions was lower for one-to-one gametologs than for other neo-Y-linked genes, with marginal significance (P = 0.041 and 0.085 by one- and two-sided Fisher’s exact tests, respectively). This observation suggests that individuals with disruptions of single-copy neo-Y-linked genes are likely to exhibit lethality, possibly because of the absence of immediate DC on the neo-X-linked gametologs. In contrast, disruptions of other genes, including many multi-copy neo-Y-linked genes, are unlikely to be deleterious because the expression of other copies (i.e., paralogs) would mask the effect of the disruptions, which would render the immediate DC dispensable.
| Category | No. of genes2 | No. of genes with deletions |
|---|---|---|
| All | 9,775 | 82 |
| One-to-one gametologs | 927 | 3 |
The three neo-Y-linked genes with a one-to-one gametologous relationship with the neo-X-linked genes were homologs of staufen (stau), back seat driver (bsd), and alicorn (alc) in D. melanogaster. In all of them, the deletions predicted were detected in UTRs (Fig. 3A) and did not affect the expression level considerably in the F1 males (Fig. 3B–3D), as determined using RNA-seq (see the “Gene expression analysis” subsection of the Supplementary Materials and Methods for details). Moreover, the expression levels of their neo-X-linked gametologs was not significantly changed (Fig. 3B–3D). Therefore, we were unable to evaluate directly whether DC operated immediately in response to the disruption of the neo-Y-linked genes, because no neo-Y-linked gene was disrupted in terms of coding integrity or gene expression by 2-Gy irradiation with the Fe-ion beam.
However, the observations that deletions in the neo-Y-linked genes with a one-to-one gametologous relationship with the neo-X-linked genes were less frequent than those in other neo-Y-linked genes, and that there was no deletion that affected the coding integrity or expression of these neo-Y-linked genes suggest that DC did not operate immediately on the neo-X in response to the pseudogenization of the neo-Y-linked genes. However, further experiments using different irradiation conditions are apparently needed, although individuals with large deletions may exhibit lethality if there is no immediate DC. The application of the CRISPR-Cas9 system to D. miranda may also represent another approach to test immediate DC further. Although CRISPR-Cas9 cannot induce large deletions, this method is able to delete a gene of interest, which is impossible to achieve using any of the irradiation methods available currently. Using CRISPR-Cas9, Xia et al. (2021) actually reported that the knockout of vismay on the second chromosome induced the upregulation of its neighboring homolog, achintya, in D. melanogaster.
We will also need to compare our pipeline with others based on different approaches. For example, Ishii et al. (2017) detected deletions in the Arabidopsis thaliana genome if the distance between the forward and reverse reads of a pair on the reference genome was significantly longer than expected. Because this approach seems inappropriate to detect small deletions, we were unable to apply it in the present study. Nevertheless, once the conditions necessary to obtain large deletions are established, we will use this method and compare it with our pipeline. We will also improve our method to detect heterozygous deletions on autosomes and compare the effects of deletions on autosomes and sex chromosomes in the future.
In conclusion, gene-by-gene DC on the neo-X in D. miranda may not have operated within one generation after the nonfunctionalization of neo-Y-linked genes, although our observations remain indirect. Because global DC is also known to operate only partially (Nozawa et al., 2018, 2021), it is possible that the neo-sex chromosomes in D. miranda are too young to allow the establishment of the stringent mechanism of DC. The investigation of other species with different sex chromosome ages is crucial for understanding the relationship between the age of the sex chromosomes and the stringency of immediate DC.
We thank Yasuko Ichikawa and Yukari Kobayashi for their help on the experiments. We also thank Aya Takahashi, Reika Sato, and Takehiro K. Katoh for their comments on the manuscript. We appreciate Koichiro Tamura, Yasukazu Okada, and Yuuya Tachiki for critical discussion. We also appreciate Yusuke Kazama and Takuya Abe for their advices on the experiments. This work was supported by JSPS KAKENHI Grant Numbers 21H02539 and 19K22460 to M.N. Author contributions are as follows: M.N. and M.O. designed the research. M.O., M.N., K.T., and T.A. conducted experiments. M.O. and M.N. analyzed the data. M.N. and M.O. wrote the manuscript. All authors carefully checked the manuscript and approved the research contents. All sequence data generated in this study have been submitted to the DDBJ Sequence Read Archive (https://www.ddbj.nig.ac.jp/dra/index-e.html) under the accession number DRA015898.
