D. magna has two Dsx orthologs, Dsx1 and Dsx2 (Kato et al., 2011). Because Dsx2 seems to have lost sex-related function (Kato et al., 2011), its characterization and potential function is not focused in this review. D. magna Dsx1 has the DM domain and oligomerization domain at the N-terminus and C-terminus respectively like Drosophila and the other insect Dsx proteins (Kato et al., 2011) (Figure 2A). The amino acid sequence of Dsx1 DM domain is well-conserved including zinc-chelating amino acids (Figure 2B). The oligomerization domain shows a more divergent sequence, but three non-polar amino acids involved in the formation of the interface are conserved (Figure 2C), suggesting that modes of transcriptional regulation of the D. magna Dsx1 are similar to those of Drosophila Dsx.

The D. magna Dsx1 gene is approximately 20 kb in length and consists of four exons (Kato et al., 2011) (Figure 3). The first and second introns are approximately 9 and 8 kb, each of which is considerably longer than the average size of introns (392 bp) on the D. magna genome, which may imply that important regulatory sequences are in the introns. Two transcripts, Dsx1α and Dsx1β, are produced from the Dsx1 locus by alternative promoter usage: Dsx1α utilizes exon 3 and exon 4, and Dsx1β is encoded in exons 1, 2, and 4. The Dsx1 ORF is included in exon 4 and is not spliced in a sex-specific manner, demonstrating that, unlike Drosophila and other insects’ Dsx, the same polypeptide is produced in both sexes. The 3′ UTR region contains four alternative polyadenylation sites, resulting in the synthesis of mRNAs with different 3′ UTR lengths (Kato et al., 2011). Even though Dsx1 gene is not regulated by sex-specific splicing, these sequence characteristics suggest that this gene is intricately regulated at transcriptional and post-transcriptional levels.

Temporal Dsx1 expression is divided into three phases: initiation, upregulation, and maintenance (Figure 4) (Mohamad Ishak et al., 2017). Before 6-hour post ovulation (hpo), Dsx1α mRNA is synthesized both in males and females (initiation phase). Around the timing of gastrulation, from 6 to 9 hpo, its expression is upregulated only in males (upregulation phase). In contrast to Dsx1α mRNA, Dsx1β mRNA is deposited as a maternal RNA in both sexes. After consumption of maternal RNAs, zygotic Dsx1β expression begins in both sexes and its male-specific upregulation occurs at 3 h later than the timing of Dsx1α upregulation. This difference of timing for male-specific upregulation of each isoform suggests that maternal sesquiterpenoid signaling first influences the Dsx1α promoter Thereafter, in males, expression of both isoforms is steadily maintained. Dsx1α is further upregulated at the instar 5 when the secondary sex characteristics appear. In females, the Dsx1 locus is persistently silenced except for synthesis of the maternal Dsx1β mRNA (Nong et al., 2017).

To investigate where the Dsx1 activation occurs throughout development, the transgenic D. magna for visualizing the Dsx1 expression was generated (Nong et al., 2017). TALENs that introduce double-strand break at the start codon of Dsx1 in its exon 4 was constructed and used for non-homologous end-joining mediated knock-in of the donor plasmid harboring the mCherry ORF and Dsx1 3′ UTR (Figure 5). Coinjection of the TALEN mRNAs with the donor plasmid led to the establishment of two transgenic lines (Line A and Line B) showing mCherry fluorescence. Both lines have a single copy of the mCherry reporter on one of the Dsx1 alleles. In Line A, another Dsx1 allele is mutated with a 6-bp deletion at the start codon, which suggests that this line does not produce any intact Dsx1 protein (Nong et al., 2020). In contrast, Line B has a wild-type Dsx1 gene on another allele. This genotype enables Line B males to develop sex-specific morphology such as elongation of first antennae and to have the reproductive ability for producing sexual eggs by copulating with females, even though those male-specific traits are slightly feminized (Nong et al., 2017). Localization of mCherry fluorescence was consistent with Dsx1 mRNA localization that was analyzed by in situ hybridization (Kato et al., 2011). Therefore, Line B has been used as a Dsx1 reporter line that recapitulates Dsx1 expression (Mohamad Ishak et al., 2017; Kato et al., 2018; Perez et al., 2021).

Live imaging of mCherry fluorescence in Line B male embryos revealed the dynamic change of spatial expression patterns in early embryos (Nong et al., 2020) (Figure 6). During the initiation phase of Dsx1, no clear mCherry signal was detected probably due to low Dsx1 expression levels (Figure 6, stage i and ii). In the upregulation phase, two mCherry signals were detected at 11 hpo. One was the cell cumulus appearing around the invagination, and the other was around the blastopore (Figure 6, stage iii). The mCherry-expressing cumulus moved to the vicinity of the blastopore, and the signal appeared to weaken gradually (Figure 6, stage iv). In parallel, another mCherry-expressing population migrated from the blastopore to the posterior region (Figure 6, stage iv). When naupliar segmentation occurred, the population of cells showing strong mCherry fluorescence converged in the posterior growth zone although it is not clear whether the two mCherry-expressing cell populations merged (Figure 6, stage v). In other animals, cell populations showing similar migration behavior are known as a primary organizer that induces differentiation in the adjacent area (Bénazéraf and Pourquié, 2013; Akiyama-Oda and Oda, 2003). This suggests that the migrating cell population may induce sexual differentiation in their surrounding cells. The posterior growth zone is known to supply the progenitor cells for the formation of thoracic appendages (Martin and Kimelman, 2009), suggesting that, in male embryos of D. magna, this region distributes masculinized progenitor cells for the development of the male-specific structure of thoracic appendages. Strong mCherry fluorescence was also observed in the buds of the first antennae, and in parallel, strong mCherry fluorescence was detected at the boundary between the head and thoracic segments (Figure 6, stage v). Later, at 18 h, more intense mCherry fluorescence was detected in the first thoracic appendage (Figure 6, stage vi), where hooks for female capture are formed at the later juvenile stage. During these earlier embryonic stages, rather broader mCherry fluorescence patterns were observed, indicating that the embryos are still sexually immature. Localization of mCherry fluorescence of female embryos was similar but its intensity was weaker compared to male embryos until the naupliar segmentation stage. Thereafter, there was no mCherry fluorescence in female embryos (Figure 6).

As embryonic development progressed toward the first instar stage that corresponded to 72 hpo, mCherry signal was localized to organs characteristic of males, first antennae, first thoracic leg, external genitalia, and testis (Figure 7, instar 1). The expression of Dsx1 in these male-specific traits stably continued until the fourth instar stage (Figure 7, instar 4). After males showed secondary sex characteristics at the instar 5, mCherry fluorescence expanded to the other several male-specific structures, including the carapace edge below the helmet, the tip of the penis, and the skeletal muscle (Figure 7, instar 7) (Nong et al., 2020). These imaging results suggest that Dsx1 is tightly regulated by time and place for controlling male development.

A previous study reported that knockdown of Dsx1 led to feminization in males (Kato et al., 2011). Long dsRNA targeting Dsx1 mRNA was injected into eggs that were destined to be developed into males. By Dsx1 RNAi, the first antennae of males became short like those of females (Figure 8A) and a copulatory hook for capturing females was lost from the first thoracic leg (Figure 8B). In addition to the feminization of somatic tissues, the gonads also showed sex reversal from testes to ovaries (Figure 8C). When the chemically modified Dsx1-targeting siRNA was injected, the effects of Dsx1 silencing in adults became severe enough to allow the knockdown males for laying eggs (Kato et al., 2018), showing that Dsx1 is the key sex determination factor in D. magna.

Phenotyping of the Dsx1 mutants Line A and Line B aforementioned also showed Dsx1 dependency of male trait development (Figure 4) (Nong et al., 2017; Nong et al., 2020). Compared to Line B males, Line A males showed more severe feminization throughout development, which were obvious in the first antennae and carapace (Figure 9A). The first antennae of Line A were as short as those of wild-type females in younger juveniles in contrast to Line B males with rather elongated first antennae. At instar 5 when the secondary sex characteristics appeared, the difference in the degree of feminization between the two mutants became clearer. Line B male showed the transformation of first antennae and carapace into the mature adult-like structures at this instar whereas Line A gradually transformed morphology of those two male traits across the following instar stages. Feminization of Line A males was also observed throughout the body such as larger body size (Figures 9B,C), the modest opening of the ventral carapace (Figure 9C), bulky head (Figure 9D), female-like genital and anus (Figure 9E). In addition, Line A male developed ovary-like darkened gonads but did not produce any eggs (Figures 9B,C). Male-specific traits are developed exclusively dependent on Dsx1 activity in D. magna.

On the D. magna genome, upstream of the Dsx1α transcription start site, there is an element that resembles the consensus sequence of the binding site for the bZip transcription factor Vrille. This element overlaps a potential Dsx binding site (Figure 3) (Mohamad Ishak et al., 2017). A similar element harboring both the bZip transcription factor and Dsx binding sites had been characterized well as an enhancer of yolk protein gene in Drosophila (An and Wensink, 1995). Importantly, in D. magna, male-specific expression of Vrille occurred at 6 hpo before upregulation of Dsx1 begins (Figure 10A) even though Vrille did not show any sexually dimorphic expression in the later stages of embryos (Figure 10B). Vrille knockdown reduced Dsx1 expression in male embryos (Figures 10C,D) whereas its overexpression in females activated Dsx1 expression and in turn led to the development of male trait, elongation of first antennae (Mohamad Ishak et al., 2017). Disruption of the Vrille binding site on the Dsx1α promoter region inhibited Dsx1 activation in male embryos, demonstrating that Vrille is required for Dsx1 activation during the upregulation phase. Interestingly, there was a potential binding site of the sesquiterpenoid receptor Methoprene-tolerant (Met) in the Vrille promoter, suggesting the possibility that the Met may directly regulate Vrille expression (Mohamad Ishak et al., 2017). In Drosophila, Vrille functions as a sex regulatory factor downstream of Dsx (Lebo et al., 2009). Vrille might be repeatedly co-opted in the sex-determining pathways with the different hierarchical position in animals as reported in the other sex-determining genes (Herpin and Schartl, 2015).

A comprehensive epigenomic analysis of Daphnia pulex revealed that, on the Dsx1 locus, there were sex differences of active and repressive histone marks, histone H3 trimethylation at lysine 4 (H3K4me3), and histone H3 trimethylation at lysine 27 (H3K27me3) in addition DNA methylation status (Kvist et al., 2020). The H3K4me3 level was around 300-fold higher but the H3K27me3 level was around 5,000-fold lower in males compared to those in females. CpG methylation level was higher in females. These epigenetic marks suggest opened and closed chromatin states in males and females respectively. Although epigenetic marks on D. magna Dsx1 remain unknown, a previous study found that Dsx1 enhancer element-targeting Crispr/Cas9 could not induce any mutation on the Dsx1 enhancer element in female embryos but resulted in indel formation in males (Mohamad Ishak et al., 2017). CRISPR/Cas9 is known to have weaker activity against the closed chromatin (Chen et al., 2016; Daer et al., 2016; Jensen et al., 2017), supporting the sex difference of the epigenetic state on the Dsx1 locus.

The RNA binding protein Alan shepard (Shep) ortholog was proved as a post-transcriptional regulator for Dsx1α mRNA in D. magna (Perez et al., 2021). The relationship between Shep and Dsx1α mRNA was found unexpectedly based on phenotypic analysis of female embryos injected with DsRed2 reporter RNA harboring the Dsx1α or Dsx1β 5′ UTR (Kato et al., 2018). Despite lacking the Dsx1 ORF, the chimeric DsRed2 RNA including the Dsx1α 5′ UTR induced development of the male trait and activation of Dsx1 in females. The RNA that coded for only Dsx1α 5′ UTR also led to Dsx1 activation. To investigate the molecular mechanism underlying the unexpected function of Dsx1α 5′ UTR for Dsx1 regulation, the proteins interacting with Dsx1α 5′ UTR were examined by the MS analysis with D. magna lysates and one of the identified proteins was Shep. D. magna Shep gene is expressed both in males and females during embryogenesis. Silencing of Shep in the Dsx1 reporter line increased mCherry expression in male and female embryos (Figures 11A,C). In contrast, overexpression of Shep reduced the mCherry expression in the Dsx1 reporter line (Figures 11B,F). However, both loss- and gain-of-function did not change Dsx1 transcript levels (Figures 11D–F), suggesting that Shep represses Dsx1 expression at a post-transcriptional level. Dsx1α 5′ UTR has a sequence like “tra-2 and GLI element (TGE)” (Figure 3), which has been identified as a binding site of Sup-26, an ortholog of Shep in C. elegans. In vitro reporter assay demonstrated that Shep represses translation of Dsx1α mRNA via TGE (Figure 13A) (Perez et al., 2021).

Dsx1 gene transactivation function of the Dsx1α 5′ UTR is attributed to the inclusion of this 5′ UTR sequence within long noncoding RNA DAPALR (Dsx1 alpha promoter associated long RNA) (Figure 12A). This lncRNA is transcribed from upstream of the transcription start site of Dsx1α mRNA in sense orientation and overlaps Dsx1α 5′ UTR (Figure 3) (Kato et al., 2018). DAPALR is a capped and non-polyadenylated RNA. This lncRNA expression is male-specific as well as Dsx1 but its expression level is around 10 times lower than that of Dsx1. DAPALR is activated by Vrille, which may suggest co-regulation of DAPALR with Dsx1. RNAi-mediated knockdown of DAPALR in males led to their feminization both in somatic tissues and germline, which resulted in the production of offspring (Figure 12B). This feminized phenotype was caused by a reduction of the Dsx1 transcript level (Figure 12C). In contrast, overexpression of DAPALR in females induced upregulation of the Dsx1 gene, resulting in the development of the male trait, elongation of the first antennae (Figures 12D–F). DAPALR activates Dsx1 in trans and Dsx1α 5′ UTR functions as a transactivation element of DAPALR (Kato et al., 2018).

In in vitro translation system including the reporter mRNA harboring Dsx1α 5′ UTR, DAPALR addition canceled the Shep-dependent translational suppression (Perez et al., 2021). Similarly, Dsx1α 5′ UTR had the ability for canceling the Shep function (Figure 13A). Thus, Dsx1α 5′ UTR was also designated as a “DAPALR core element” (Figure 3). Importantly, DAPALR and its core element inhibited the function of Shep as a translational repressor in a dose-dependent manner (Figure 13B). DAPALR functions as a decoy of Shep through its core element that corresponds to Dsx1α 5′ UTR, which in turn increases Dsx1 expression.