Absence of X-chromosome dosage compensation in the primordial germ cells of Drosophila embryos

Dosage compensation is a mechanism that equalizes sex chromosome gene expression between the sexes. In Drosophila, individuals with two X chromosomes (XX) become female, whereas males have one X chromosome (XY). In males, dosage compensation of the X chromosome in the soma is achieved by five proteins and two non-coding RNAs, which assemble into the male-specific lethal (MSL) complex to upregulate X-linked genes twofold. By contrast, it remains unclear whether dosage compensation occurs in the germline. To address this issue, we performed transcriptome analysis of male and female primordial germ cells (PGCs). We found that the expression levels of X-linked genes were approximately twofold higher in female PGCs than in male PGCs. Acetylation of lysine residue 16 on histone H4 (H4K16ac), which is catalyzed by the MSL complex, was undetectable in these cells. In male PGCs, hyperactivation of X-linked genes and H4K16ac were induced by overexpression of the essential components of the MSL complex, which were expressed at very low levels in PGCs. Together, these findings indicate that failure of MSL complex formation results in the absence of X-chromosome dosage compensation in male PGCs.

In some reproductive animals, sex chromosomes play key roles in sex determination and differentiation 1 . In such animals, sex-chromosome constitution differs between females and males. For example, in mammals and Drosophila, females and males carry two X chromosomes (XX) and one X chromosome (XY), respectively 1 . This difference in the chromosome constitution causes an imbalance in the number of X-linked genes between the sexes; however, their expression can be equalized between XX females and XY males through a mechanism called dosage compensation 2 .
In the somatic cells of Drosophila males, expression of X-linked genes is upregulated twofold by the dosage compensation mechanism 3 . This upregulation is mainly mediated by the male-specific lethal (MSL) complex. Exclusively in males, this complex binds the X chromosome to acetylate lysine residue 16 in histone H4 (H4K16ac), which in turn hyperactivates X-linked genes 4,5 . The MSL complex contains the essential components for catalysis, MSL-1 (a scaffold protein), MSL-2 (a ubiquitin ligase), MSL-3 (a chromodomain protein), MOF (a histone acetyltransferase), and MLE (a DNA/RNA helicase), in addition to the non-coding RNAs roX1 and roX2 3 . In the soma, the transcript encoding MSL-2 is repressed translationally in females by Sex-lethal (Sxl), which is a binary switch gene that regulates sex determination and dosage compensation, allowing formation of the MSL complex only in males [6][7][8] .
In contrast to the soma, the existence of dosage compensation in the Drosophila germline remains controversial. Rastelli and Kuroda showed that in the germline cells of adult testes, MLE does not localize to the X chromosome, and MSL-3 is not expressed, although H4K16ac is present 9 . Furthermore, Gupta and colleagues performed a microarray analysis that revealed twofold upregulation of X-linked genes in the germline of adult testes 10 . However, recent transcriptome analyses do not support the idea that dosage compensation occurs in the germline of adult testis [11][12][13] . The discrepancies may result from differences in sample preparation or the gene sets selected for each analysis. Likewise, varying degrees of contamination of somatic cells in the germline samples www.nature.com/scientificreports/ could yield dissimilar conclusions. Furthermore, meiotic inactivation of X-linked genes in spermatocytes 14 raises concerns about the validity of the conclusions. In this study, we focused on primordial germ cells (PGCs) in embryos, as dosage compensation is linked with sex determination and is an early event in the soma that differs between the sexes 3,6,15,16 . In Drosophila, PGCs are formed in the posterior pole region of early embryos, and then migrate through the embryo (middle embryogenesis) to reach the embryonic gonads (late embryogenesis). In migrating PGCs, Sxl is expressed in a female-specific manner and induces female fate 17 , although its downstream targets in these cells are distinct from those in the soma [18][19][20][21][22] . By contrast, male PGCs within gonads initiate male fate in response to JAK/STAT signals from the gonadal soma 23,24 . These observations led us to investigate whether dosage compensation occurs in PGCs during these embryonic stages.
In this study, we found that dosage compensation was undetectable in PGCs during middle to late embryogenesis. Consistent with this, H4K16ac was undetectable in male PGCs. All transcripts encoding the components of the MSL complex, except for MOF, were expressed at very low levels in PGCs. Moreover, overexpression of msl-1, msl-2, msl-3, and roX2 induced hyperactivation of X-linked genes and H4K16ac in male PGCs. Our observations strongly suggest that dosage compensation is absent in male PGCs, due to the failure of MSL expression.
To produce flies carrying UAS-msl-3 and UAS-roX2, the msl-3 ORF and full-length roX2 (RD-type isoform) were amplified from cDNA obtained from 4 to 8 h AEL embryos using primers msl-3Fw/msl-3Rv and roX2Fw/ roX2Rv (Table S1), respectively. The msl-3 ORF and roX2 fragment were cloned into KpnI-digested pUASp-K10-attB using the In-Fusion HD Cloning Kit. DNA spanning from the 5´ terminus of the GAGA-binding site to the 3´ terminus of K10, which contains the msl-3 ORF, was amplified using primers pUASp-K10-attBFw/ pUASp-K10-attBRv (Table S1), and the amplicon was cloned into PsiI-digested pUASp-K10-attB containing the roX2 fragment. The resultant vector was injected to M{3xP3-RFP.attP}ZH-2A embryos to produce flies carrying UAS-msl-3 and UAS-roX2 on the X chromosome. w + transformants were collected and used to establish homozygous stocks. As described above, the roX2 transgene (UAS-roX2) was inserted into the X chromosome, because its expression from an autosome causes assembly of the MSL complex at the expression site and ectopic upregulation of the surrounding autosomal genes 30 .
Transcriptome analysis. For RNA-seq of female and male PGCs at stages 12-16, 1.9-2.3 × 10 5 female and male PGCs (Table S2) were isolated by fluorescence-activated cell sorting (FACS), as described 21,31 , from 8 to 16 h AEL embryos derived from vasa-EGFP/vasa-EGFP; nos-Gal4/nos-Gal4 females mated with UAS-RFP/Y males (Fig. S1a). EGFP and RFP double-positive cells and EGFP-positive cells were isolated as female and male PGCs, respectively. Total RNA was extracted from isolated PGCs using the RNeasy mini Kit (QIAGEN, 74104). Library creation using the TruSeq Standard mRNA Library Prep Kit (Illumina, 20020594) and RNA-seq using the HiSeq 2500 platform (Illumina) were carried out at the University of Minnesota Genomics Center (UMGC), and approximately 40 million reads per sample (125-bp paired-end reads) were obtained.
Raw reads obtained from both stage 12-16 and stage 15-16 samples were processed using Trimmomatic-0. 36 32  www.nature.com/scientificreports/ Microarray data analysis. To compare gene expression profiles of somatic tissues between males and females, we used microarray data of thoraxes dissected from adult males and females 11 . These data have been deposited in the Gene Expression Omnibus (GEO) under accession No. GSE30850. Raw values were normalized across samples using quantile normalization, and fold changes in female vs. male thorax were calculated for each gene.
To select zygotically expressed genes, we used microarray data obtained from PGCs at 11 different embryonic stages 21 . These data have been deposited in GEO under Accession No. GSE83460. Raw values were normalized across samples using quantile normalization. Genes expressed at low levels in PGCs at stage 4 (log 2 expression values < 7) and at high levels in PGCs at stage 16 (log 2 expression values > 8) were selected as zygotically expressed genes.

Fixation of embryos.
For immunostaining and in situ hybridization, embryos were collected and dechorionated in a sodium hypochlorite solution. The dechorionated embryos were fixed in 1:1 heptane:fixative [4% paraformaldehyde in PBS (130 mM NaCl, 7 mM Na 2 HPO 4 , and 3 mM NaH 2 PO 4 )] for 30 min. Vitelline membranes of the fixed embryos were removed by vigorous shaking in 1:1 methanol:heptane. The embryos were then rinsed with methanol and stored in methanol at − 20 °C until use.
Testes were dissected from adult flies 3-5 days after eclosion and fixed for 15 min. Fixed testes were rinsed with PBSTr (PBS containing 0.1% Triton X-100) and stored in PBSTr at 4 °C until use.
Whole-mount in situ hybridization combined with immunostaining was performed as described 21 . The following experiments were conducted at room temperature unless otherwise stated. Fixed embryos derived from nos-Gal4/nos-Gal4 females mated with UAS-RFP/Y males were rinsed with ME [50 mM EGTA (pH 8.0) in 90% methanol] and incubated in 7:3, 5:5, and 3:7 ME:fixative for 5 min each. The embryos were re-fixed with fixative for 20 min, and then washed with PBSTw (PBS containing 0.1% Tween 20) three times for 5 min each. The embryos were digested with 50 µg/ml Proteinase K in PBSTw for 3 min, and the digestion was stopped by incubation in fixative for 20 min. The digested embryos were washed with PBSTw three times for 10 min each, and then incubated in pre-hybridization solution [pre-HS; 50% formamide, 5 × SSC (750 mM NaCl and 75 mM sodium-citrate), 100 µg/ml heparin, 100 µg/ml yeast tRNA, 10 mM DTT, and 0.1% Tween 20] for 60 min at 60 °C. They were then hybridized with 2 ng/µl RNA probe in hybridization solution (pre-HS containing 10% dextran sulfate) overnight at 60 °C. After hybridization, the embryos were washed with a washing solution (50% formamide, 5  www.nature.com/scientificreports/ overnight at 4 °C in TNB containing chick anti-Vasa (1:500) 37 and rabbit anti-RFP (1:1000; Thermo Fisher Scientific, R10367) and washed with TNT three times for 15 min each. For detection of RNA probes and antibodies, embryos were incubated overnight at 4 °C in TNB containing fluorescein-conjugated streptavidin (1:1000; PerkinElmer, S869), Alexa Fluor 633-conjugated goat anti-chick (1:500, Thermo Fisher Scientific, A21103), and Alexa Fluor 546-conjugated goat anti-rabbit (1:500, Thermo Fisher Scientific, A11035). The embryos were washed with TNT three times for 15 min each, and then mounted in VECTASHIELD Mounting Medium. The sex of each embryo was determined based on RFP staining, except in the case of stage 5 embryos.

Results and discussion
Dosage compensation is indiscernible in the PGCs. To investigate whether dosage compensation occurs in male PGCs, we separately isolated female and male PGCs and compared the expression of X-linked genes between the two sexes. For this purpose, we used females carrying nos-Gal4 and vasa-EGFP mated with males carrying UAS-RFP on the X chromosome 21 (Fig. S1a). In embryos derived from these mothers, female and male PGCs were double-positive for EGFP and RFP and single-positive for EGFP, respectively (Fig. S1a), as nos-Gal4-driven UAS-RFP is activated only in female PGCs from stage 9 onward, whereas vasa-EGFP is expressed throughout germline development in both sexes 25,26,37 . RNA-seq data were obtained from female and male PGCs at stages 12-16. Our data revealed that female-biased expression was significantly more common in transcripts from X-linked genes than in those from autosomal genes, although sex-biased expression was also observed in autosomes (Fig. S2a-f). Expression of the transcripts from X-linked genes was twofold higher on average in female PGCs than in male PGCs, whereas the female/male expression ratio of transcripts from autosomal genes was ~ 1 (Fig. 1a). Similar female-biased expression was observed when zygotically expressed genes were selected from X-linked genes (Fig. S3), indicating female-biased zygotic expression from X-linked genes.
In addition, we found that expression of X-linked housekeeping genes for glycolysis and ATP synthesis was female-biased in PGCs (Fig. S4). By contrast, in the soma, female-biased expression of X-linked genes was not observed (Fig. 1b). In addition, we found that H4K16ac was undetectable in the PGCs, which were surrounded by H4K16ac-positive somatic cells in males (Fig. 1c,d). These observations led us to conclude that expression of genes on the X chromosome is subjected to dosage compensation in the soma, but not in PGCs. However, we cannot rule out the possibility that expression of a fraction of X-linked genes in male PGCs is compensated by mechanisms independent of H4K16ac. Zygotic expression of ovo and ovarian tumor (otu) genes on the X chromosome is female-biased 38 , and both genes are required for female germline development 20,39,40 . We found that these genes exhibited female-biased expression (Fig. S2d). Furthermore, transcripts from marker genes for male PGCs 38 , such as disc proliferation abnormal (dpa), CG9253, no child left behind (nclb), Ran GTPase activating protein (RanGap), Rs1, CG6693, Kinesin-like protein at 61F (Klp61F), and Minichromosome maintenance 5 (Mcm5), exhibited male-biased expression in PGCs (Fig. S2e and f).

Expression of genes encoding the components of MSL complex in male PGCs during embryogenesis. Given that the MSL complex is required for dosage compensation of the X chromosome through
H4K16ac in the soma 4,5 , we next asked whether genes encoding the components of the MSL complex are expressed in male PGCs during embryogenesis. For this purpose, we performed in situ hybridization of embryos produced from females carrying nos-Gal4 mated with males carrying UAS-RFP on the X chromosome. In the PGCs of stage 5 embryos, msl1 and mof mRNAs were detectable ( Fig. 2a and p). Expression of these mRNAs is considered to be maternal in origin, as PGCs are transcriptionally repressed at this stage 26,41,42 , and their zygotic transcription in PGCs is repressed until late gastrulation. Furthermore, we detected the msl-2, msl-3, and mle mRNAs and two roX RNAs at very low levels (Fig. 2f, k, u, z, and ee). In PGCs at the middle and late embryonic stages (stages 10-16), only mof mRNA was observed at a high level (Fig. 2q-t), and the msl-1, msl-2, msl-3, and mle mRNAs and two roX RNAs were all detected at low levels, if at all (Fig. 2b-e, g-j, l-o, v-y, aa-dd, and ff-ii). We observed no significant difference in the expression of these RNAs between male and female PGCs (Fig. 2  and Fig. S5). These observations suggest that the PGCs are depleted of some essential components of the MSL complex, which in turn eliminates the ability to induce dosage compensation in male PGCs.

Overexpression of msl-1, msl-2, msl-3, and roX2 induces H4K16ac and hyperactivation of X-linked genes in male PGCs.
To determine whether depletion of MSL components causes deficiency of dosage compensation in male PGCs, we overexpressed genes encoding MSL components in male PGCs. MLE protein is maternally supplied and expressed in male PGCs during embryogenesis, whereas MSL-1 and MSL-2 proteins are not 43 . Hence, we forced expression of msl-1, msl-2, msl-3, and roX2 in PGCs. Only roX2 was overexpressed because the roX RNAs are functionally redundant 44 . When msl-1, msl-2, msl-3, and roX2 were overexpressed under the control of nos-Gal4 (msl oe), H4K16ac became detectable in the nuclei of male PGCs (Fig. 3a,b).
Thus, overexpression of msl-1, msl-2, msl-3, and roX2 in male PGCs is insufficient to induce twofold upregulation of X-linked genes, as observed in male soma (Fig. 1b). One possible explanation for this is that dosage compensation requires factors other than the MSL complex in male PGCs. The functions of JIL-1 kinase and Topoisomerase 2 are also needed for dosage compensation in male soma [45][46][47] . The possibility that overexpression of these molecules, along with msl-1, msl-2, msl-3, and roX2, can induce twofold upregulation of X-linked genes in male PGCs remains to be tested.
Biological significance of the absence of dosage compensation in male PGCs. In Drosophila, the sexual identity of the germline is regulated by both cell-autonomous cues, which are produced depending on X-chromosome constitution, and the sex of the surrounding soma 23,[48][49][50] . In the absence of dosage compensation, X-linked genes are expressed at twofold higher levels in female (XX) PGCs than in male (XY) PGCs at stages 12-16 (Fig. 1a). This biased expression of X-linked genes may be one of the determinants of femaleness. We found that overexpression of msl-1, msl-2, msl-3, and roX2 can hyperactivate X-linked genes in XY PGCs (Fig. 4a,b, Fig. S6d-f, Fig. S7d-f, and Fig. S8), suggesting that XY PGCs acquire femaleness, but only partially. Indeed, overexpression of msl-1, msl-2, msl-3, and roX2 in XY PGCs caused upregulation of ovo (Fig. S7d) and downregulation of genes that exhibited male-biased expression in PGCs (Fig. S7e and f). Thus, it is possible that female sexual identity of PGCs can be brought about by introducing dosage compensation. However, PGCs overexpressing msl-1, msl-2, msl-3, and roX2 executed normal spermatogenesis and became functional sperm (Fig. S9). This is presumably because masculinizing signals from the surrounding male soma 38 override feminization of PGCs. Transplantation of XY PGCs overexpressing msl-1, msl-2, msl-3, and roX2 into female soma may clarify whether twofold upregulation of X-lined genes induces femaleness in XY PGCs.  -1 (a-e), msl-2 (f-j), msl-3 (k-o), mof (p-t), mle (u-y), roX1 (z-dd), and roX2 (ee-ii) in PGCs at embryonic stage 5 (St. 5; a, f, k, p, u, z, and ee) and in male PGCs at embryonic stages 10-11 (St. 10-11; b, g, l, q, v, aa, and ff), stages 12-13 (St. 12-13; c, h, m, r, w, bb, and gg), stage 14 (St. 14; d, i, n, s, x, cc, and hh), and stage 16 (St. 16; e, j, o, t, y, dd, and ii). Embryos derived from nos-Gal4/nos-Gal4 females mated with UAS-RFP/Y males were in situ hybridized with a probe for each gene (green) and immunostained for Vasa (magenta) and RFP. Because nos-Gal4 activates UAS-RFP only in female PGCs from stage 9 onward, the sexes of PGCs could be determined by RFP signal at stages 10-16. Probes were designed to detect all RNA variants identified in each gene region. Scale bar: 10 µm. White arrowheads show PGCs with high-level signals (green). ◂ Figure 3. Overexpression of msl-1, msl-2, msl-3, and roX2 induces H4K16ac in male PGCs. (a and b) H4K16ac expression in male PGCs of nos-Gal4 strain (a; control) and male PGCs overexpressing msl-1, msl-2, msl-3, and roX2 (b; msl oe) at embryonic stage 14. Embryos were stained for H4K16ac (green) and Vasa (magenta). Embryos were sexed based on H4K16ac staining in the soma. Scale bar: 10 µm.  Fig. 1a. Mean values for the X, second, and third chromosomes were 1.03, − 0.23, and − 0.21, respectively. Blue dotted lines indicate log 2 expression ratios of 0 and 1. Significance was calculated by two-sided Mann-Whitney U test (*, P < 0.05; ns, not significant). Transcripts for which TPM was zero in all samples were excluded from this analysis. N: number of the transcripts examined. (b) Log 2 expression ratio of transcripts from genes on the X (pink), second (light blue), and third chromosomes (lime green) between msl oe male PGCs and male PGCs (msl oe male/male) at embryonic stages 15-16. Each box plot represents values as in Fig. 1a. Mean values for the X, second, and third chromosomes were 0.51, − 0.06, and − 0.04, respectively. Blue dotted lines indicate log 2 expression ratios of 0 and 1. Significance was calculated by two-sided Mann-Whitney U test (*, P < 0.05; ns, not significant). Transcripts for which TPM was zero in all samples were excluded from this analysis. N: number of transcripts examined. Expression ratio of the autosomal genes in msl oe male/male was higher than in female/male in (a) (P values, calculated by two-sided Mann-Whitney U test for the second and third chromosomes, were < 0.05), but the effect sizes calculated by Cliff 's Delta for the second and third chromosomes were statistically negligible (0.07 and 0.08, respectively) 51 . Expression ratio of X-linked genes in msl oe male/male was significantly lower than in female/male in (a) (P values, calculated by two-sided Mann-Whitney U test, were < 0.05), and the effect size was non-negligible (0.21).