Research Article

Dosage Compensation Proteins Targeted to X Chromosomes by a Determinant of Hermaphrodite Fate

See allHide authors and affiliations

Science  11 Jun 1999:
Vol. 284, Issue 5421, pp. 1800-1804
DOI: 10.1126/science.284.5421.1800


In many organisms, master control genes coordinately regulate sex-specific aspects of development. SDC-2 was shown to induce hermaphrodite sexual differentiation and activate X chromosome dosage compensation in Caenorhabditis elegans. To control these distinct processes, SDC-2 acts as a strong gene-specific repressor and a weaker chromosome-wide repressor. To initiate hermaphrodite development, SDC-2 associates with the promoter of the male sex-determining gene her-1 to repress its transcription. To activate dosage compensation, SDC-2 triggers assembly of a specialized protein complex exclusively on hermaphrodite X chromosomes to reduce gene expression by half. SDC-2 can localize to X chromosomes without other components of the dosage compensation complex, suggesting that SDC-2 targets dosage compensation machinery to X chromosomes.

In diverse organisms, the choice of sexual fate specifies not only the overt sexual characteristics evident in adults but also the amount of X chromosome gene expression in somatic cells throughout development. X chromosome–wide expression is controlled by the essential process of dosage compensation, which equalizes X chromosome expression between females (XX) and males (XY or XO). In mammals, flies, and nematodes, specialized dosage compensation complexes are targeted exclusively to the X chromosomes of one sex to modulate transcript levels (1, 2). Here we usedCaenorhabditis elegans to understand how the dosage compensation machinery is targeted to X chromosomes of hermaphrodites (XX) and how dosage compensation is coordinately activated with the genetic pathway for sexual development.

Although dosage compensation is a sex-specific process, the nematode dosage compensation complex contains both dosage compensation-specific proteins such as DPY-27 (dumpy) (3–5) and chromatin-associated proteins that are also active in meiosis or mitosis such as DPY-26 (6,7) and MIX-1 (mitosis and X) (8). These three proteins are similar to components of the frog 13S condensin complex that drives mitotic chromosome condensation in vitro (4, 8, 9), implying evolutionary recruitment of ancient mitotic proteins to the regulation of gene expression. All dosage compensation components, including the mitotic and meiotic proteins functional in both sexes, must be directed exclusively to X chromosomes of XX animals by a hermaphrodite-specific factor.

The sex-specific gene that induces hermaphrodite development, including dosage compensation, had not been determined. In males, the xol-1(XO lethal) gene initiates sexual development and represses dosage compensation by inactivating three hermaphrodite-specificsdc (sex and dosage compensation) genes (10, 11). In hermaphrodites, these genes activate dosage compensation and the genetic pathway for hermaphrodite sexual development (12–15). Both SDC-2 and SDC-3 are necessary for localization of dosage compensation proteins to X chromosomes (5,7, 8, 16). XX animals that lack sdc-2 activity develop as males and die from elevated X chromosome expression caused by the failure to dosage compensate (14). In contrast, sdc-1 and sdc-3 XX mutants exhibit less severe phenotypes (13, 15). Here we show that sdc-2 encodes the pivotal sex-specific factor that triggers the hermaphrodite program of development.

Hermaphrodite-Specific Expression of SDC-2

To elucidate the roles of sdc-2 in dosage compensation and sex determination, we cloned sdc-2(17). We used germ line transformation assays to test cosmids (Fig. 1A) and subclones (Fig. 1B) from the sdc-2 region for the ability to rescuesdc-2 XX mutants. The smallest genomic region to confer rescue was a 12.6-kb DNA fragment (Fig. 1B) that corresponds to a single 9.5-kb transcript (Fig. 1D). Detection of a 4.8-kb deletion associated with the sdc-2 allele y74 confirmed the identity of the transcript as sdc-2 (18). The deletion removes the first 606 codons of the transcript and 2.1 kb of upstream regulatory regions, consistent with y74 eliminating gene function. sdc-2 has 19 exons and is predicted to encode a highly charged protein of 2962 amino acids with a coiled-coil motif (Fig. 1C).

Figure 1

Molecular analysis ofsdc-2. (A) Corresponding genetic and physical maps of the X chromosome near sdc-2. Nearby genes (egl-15 and lin-14), DNA polymorphisms (nP8 and nP3), and overlapping cosmids (EEG4, HHG9, and C03B2) are shown. sdc-2 resides on C03B2; m.u. indicates map unit. (B) DNA transformation rescue experiments to define sdc-2. A partial restriction map of C03B2 is shown. C03B2 subclones were tested for their ability to rescue sdc-2(y46) partial loss-of-function mutants (‡),sdc-2(y74) null mutants (§), or both (†). +, rescue; −, failure to rescue. The number of rescuing lines relative to total lines is indicated in parentheses. The minimal rescuing region (12.6 kb) is bounded by dashed vertical lines. B, Bgl II; X, Xba I; W, BsiW I; E, BstE II; T, Tth111 I. (C)sdc-2 gene structure. Horizontal line, promoter; black boxes, exons; spaces, introns; open box, 3' untranslated region. Approximate location of the y74 deletion is shown. (D) Northern blot of poly(A)+ RNA from wild-type XX embryos probed with sdc-2 sequences. sdc-2encodes a highly charged 344-kD protein with a coiled-coil motif from amino acids 1226 to 1256.

Immunofluorescence experiments with SDC-2-specific antibodies (Fig. 2) (19) showed the SDC-2 expression pattern to be distinct from that of all other dosage compensation proteins. Dosage compensation proteins such as DPY-27 are diffusely distributed throughout nuclei of very young embryos (<30 cells) (Fig. 2A) in both sexes and only later become specifically localized to X chromosomes of XX (Fig. 2, D and F) but not XO animals (4, 7, 8). SDC-2 differs in three important ways. First, SDC-2 is not expressed in very young embryos (Fig. 2B). Its initial expression occurs around the 40-cell stage, a time that corresponds to the assembly of the dosage compensation machinery on X chromosomes (Fig. 2D) (4). Second, SDC-2 localizes to hermaphrodite X chromosomes from the onset of its expression (Fig. 2, E and F). Third, SDC-2 is not expressed in wild-type XO embryos (Fig. 3, A to C), which indicates that SDC-2 is sex-specifically regulated. sdc-2 is repressed in males by the XO-specific gene xol-1: in xol-1 XO mutants, SDC-2 is expressed and appears to be localized to X (Fig. 3, D to F). The observations that SDC-2 is expressed exclusively in XX embryos, that its initial expression coincides with its own X localization, and that sdc-2 is required for X localization of other dosage compensation proteins (Fig. 2, G to I) (5, 7, 8) suggest that SDC-2 is the hermaphrodite-specific protein that triggers assembly of the dosage compensation machinery on X chromosomes.

Figure 2

SDC-2 localizes to X chromosomes of XX embryos. False color confocal immunofluorescence images of wild-type (A to F) and sdc-2mutant (G to I) XX embryos costained with anti-DPY-27 (green), anti-SDC-2 (red), and the DNA-intercalating dye 4',6-diamidino-2-phe- nylindole (DAPI) (blue) are shown. Merged image in (F) superimposes images from the first two columns with DAPI. Yellow color indicates overlap between DPY-27 and SDC-2. (A to C) SDC-2 is not expressed in young XX embryos (<30 cells) (B), unlike the diffusely distributed DPY-27 protein (A). (D to F) SDC-2 is first expressed in 40- to 50-cell embryos, exhibiting a punctate pattern (E) coincident (F) with that of the X chromosome-localized DPY-27 (D). (G to I) Specificity of anti-SDC-2 is shown by the lack of SDC-2 staining in ansdc-2 null embryo (>100 cells) (H). DPY-27 cannot localize to X in sdc-2 mutants (G).

Figure 3

SDC-2 protein is absent from wild-type XO embryos but assembles dosage compensationcomponents on X in SDC-2–expressing XO embryos. (A to F) Confocal images of wild-type XO and XX (A to C) or xol-1(y155) mutant XO (D to F) embryos stained with anti-SDC-2 (red). All embryos carried the integrated Pxol-1::gfpreporter transgene (yIs34), in which the male-specific xol-1 promoter drives expression of thegfp gene, which encodes the green fluorescence protein (GFP) (29). (A) Only XO animals produce nuclearly restricted GFP. (B and C) The wild-type XO embryo does not express SDC-2, but the XX embryo shows punctate staining that persists throughout development. (E and F) In the xol-1(y155) XO mutant, SDC-2 is expressed and localized to X. (G to I) Confocal images of an SDC-2–expressing XO embryo costained with anti–β-galactosidase (red) and anti-SDC-3 (green). The embryo carries an integrated X chromosome–linked array (yIs29) with Pdpy-30::sdc-2transgenes and an integrated array (yIs2) with the male-specific Pxol-1::lacZ reporter gene, which is similar to yIs34 (11). SDC-3 is absent from wild-type XO embryos older than 100 cells (16), but in SDC-2–expressing XO embryos, SDC-3 accumulates and becomes localized to X.

Dosage Compensation Activated by SDC-2

If sdc-2 is the sex-specific switch that is both necessary and sufficient to activate dosage compensation, ectopic expression of SDC-2 in males should initiate dosage compensation, causing XO-specific lethality from underexpression of X chromosome–linked genes. SDC-2 was expressed in males from a chromosomally integrated transgene (yIs30) in whichsdc-2 transcription was controlled by the constitutively active dpy-30 promoter (20, 21). Extensive XO-specific lethality resulted from the ectopic expression ofsdc-2: 81% of yIs30/+ XO progeny were dead (Table 1). The male survivors were small, slow growing, and mating defective, further indications of inappropriate X chromosome expression. All mutant phenotypes, including lethality, were suppressed by mutations in hermaphrodite dosage compensation genes, either sdc-3 or dpy-27, which indicates that ectopic expression of sdc-2 activated the XX mode of dosage compensation in XO animals (Table 2).

Table 1

Ectopic expression of sdc-2 kills XO animals.

View this table:
Table 2

Ectopic expression of sdc-2 initiates hermaphrodite development in XO animals.

View this table:

The incomplete male lethality caused by SDC-2 suggested that another dosage compensation protein was limiting in males. SDC-3 was a likely candidate, because it associates with hermaphrodite X chromosomes and is required for X localization of other dosage compensation proteins. Moreover, SDC-3 is only weakly expressed in XO embryos (16). Whereas overexpression of SDC-3 [fromyIs3 (16)] caused only 2% male lethality, and expression of SDC-2 caused 83% lethality, overexpression of both sdc-2 and sdc-3 caused 99% male lethality (Table 1), which suggests that in XX animals SDC-3 assists SDC-2 in activating dosage compensation.

Direct demonstration that SDC-2 is sufficient to trigger assembly of the dosage compensation complex on X chromosomes was achieved by comparing the staining pattern of SDC-3 in both wild-type and SDC-2-expressing XO embryos. SDC-3 by itself does not associate with the male X chromosome: it is diffusely distributed in the nuclei of wild-type XO embryos (<100 cells) and in XO embryos engineered to overexpress SDC-3 (16). In contrast, SDC-3 appears to be specifically localized to X chromosomes in SDC-2-expressing XO embryos (Fig. 3, G to I). Thus, SDC-2 is the hermaphrodite-specific factor that activates dosage compensation.

Dosage Compensation Machinery Targeted to X Chromosomes by SDC-2

Does SDC-2 require other dosage compensation proteins for its association with X chromosomes? Evidence that SDC-2 can associate with X independently of other dosage compensation components would distinguish SDC-2 from all known dosage compensation proteins and implicate it in the recognition of X. The dosage compensation proteins DPY-26, DPY-27, MIX-1, and SDC-3 fail to associate with X chromosomes in certain dosage compensation mutants (sdc-2,sdc-3, and dpy-30) and are not stably expressed in other mutants (dpy-26, dpy-27, anddpy-28) (5, 7, 8, 16). In contrast, SDC-2 accumulates to significant quantities by midembryogenesis indpy-26, dpy-27, dpy-28,dpy-30, and sdc-3 mutants and exhibits a distinctly punctate nuclear pattern that is indistinguishable from the wild-type, X-localized pattern (Fig. 4) (22). Thus, despite a reduction in the amount of SDC-2 in dosage compensation mutants, SDC-2 appears to associate with X chromosomes without other components of the dosage compensation complex. This result implies that SDC-2 plays a central role in X chromosome recognition and confers chromosome specificity to dosage compensation.

Figure 4

SDC-2 exhibits a punctate staining pattern in dosage compensation mutants. Confocal images of wild-type (A), sdc-3 (C), and dpy-26(E) mutant embryos stained with anti-SDC-2 (red). (B, D, and F) Enlarged sections of mutant embryos costained with anti–SDC-2 (red) and DAPI (blue). Regions of overlap are fuchsia. SDC-2 protein amounts are reduced in the mutants but SDC-2 appears to be X-localized. A similar SDC-2 pattern occurs in dpy-27, dpy-28, anddpy-30 mutants, and a wild-type pattern occurs insdc-1 and dpy-21 mutants, which exhibit mild dosage compensation defects (3, 12,22). dpy-30 participates in dosage compensation by activating sdc-3 (16). These results implicate SDC-2 in X chromosome recognition. Because SDC-2 staining was faint in the mutants, the brightness was enhanced during imaging to show the punctate pattern.

Hermaphrodite Sexual Development Induced by SDC-2

In addition to its pivotal role in dosage compensation,sdc-2 plays a separate role in sex determination, promoting hermaphrodite sexual development in concert with sdc-1 andsdc-3 (13–15). Is sdc-2 the sex-specific trigger for hermaphrodite development as it is for dosage compensation? If so, XO animals that express sdc-2 should develop as hermaphrodites. Because such XO animals are dead, dosage compensation mutations were used to suppress the lethality and permit assessment of sexual fate. About 31% of yIs30 XO animals rescued by a dpy-27 null mutation were fertile hermaphrodites and 5% were intersexual (Table 2), which shows thatsdc-2 can trigger hermaphrodite development in XO animals. In contrast, all yIs30 XO animals rescued by ansdc-3 null mutation were male, consistent with the role ofsdc-3 in sex determination (Table 2). Becausesdc-2 feminized only 36% of XO animals, complete sexual transformation might require overexpression of sdc-2 and another sdc gene, just as overexpression of sdc-2and sdc-3 are needed to fully activate dosage compensation (Table 1).

Hermaphrodite sexual development requires transcriptional repression of the male autosomal gene her-1 (hermaphrodite). Mutations in sdc-1, sdc-2, or the sex determination domain of sdc-3 derepress her-1transcription in XX animals and cause masculinization, which suggests that these genes collaborate to turn off her-1(15, 23). To assess whether SDC-2 is a direct molecular repressor of her-1, we asked whether endogenous SDC-2 protein can associate with her-1 regulatory sequences in vivo. We created transgenic strains in which extrachromosomal DNA arrays included multiple copies of either her-1 regulatory regions (24) or control DNA. Arrays also includedlac operator repeats (lacO) (25) and a transgene encoding a LacI::GFP fusion protein (26). LacI::GFP binds to lacOsequences, which allows arrays to be detected by GFP autofluorescence (26, 27). Of 400 embryos from two strains with independent her-1 arrays, >90% showed colocalization of SDC-2 and GFP (Fig. 5, A to F). In contrast, of 200 embryos with control arrays, none showed any colocalization (Fig. 5, G to L). SDC-2 also localized to X chromosomes in all experimental and control embryos (Fig. 5, A to L). Thus, SDC-2 associates with her-1 promoters in vivo.

Figure 5

SDC-2 associates withher-1 promoter sequences. Confocal images of wild-type XX embryos carrying arrays (green), with (A to C) or without (G toI) her-1 promoters, stained with anti-SDC-2 (red). All arrays also contain lacO sequences and express a transgene encoding a lacrepressor::GFP fusion protein (LacI::GFP). LacI::GFP binds to lacOsequences within the DNA arrays, which allows them to be detected by green autofluorescence (26, 27). Colocalization (yellow) of GFP and SDC-2 indicates that SDC-2 binds to her-1-containing arrays but not to control arrays. Enlargement of single nuclei from wild-type embryos bearing her-1 arrays (D toF) or control arrays (J to L). Arrows indicate arrays tagged by LacI::GFP, and arrowheads indicate X chromosomes. (M to O) Confocal images of an sdc-3(Tra) XX embryo bearing her-1promoter arrays that bound SDC-2 in wild-type embryos.sdc-3(Tra) prevents SDC-2 from associating withher-1 but not X. (P to R) Enlargement of a single nucleus from the sdc-3(Tra) strain. (S) Model. sdc-2 encodes the sex-specific factor that induces hermaphrodite development in XX animals by repressing the male sex-determining gene her-1 and by triggering assembly of dosage compensation machinery on X, including proteins (DPY-26 and MIX-1) active in meiosis or mitosis. In XO embryossdc-2 is repressed by the male-specific XOL-1 protein.xol-1 is the direct molecular target of the X chromosome counting mechanism that determines sex (27). Repression is indicated by —∣.

If her-1 is a functional target of SDC-2 in vivo, many copies of her-1 regulatory regions on arrays might titrate some SDC-2 from X, thereby impairing dosage compensation. Indeed, two different genetic assays, both sensitive indicators of X chromosome expression, showed that dosage compensation was compromised in animals with her-1 arrays. First, arrays containing her-1promoter sequences enhanced the mutant phenotypes of XX animals with reduced sdc-3 activity. Without her-1 arrays,sdc-3(y126)/+ XX animals are wild type, and sdc-3/sdc-3 XX animals (fromsdc-3/+ mothers) are fully viable but exhibit weak dosage compensation phenotypes (3, 15). However, withher-1-containing arrays, 29% ofsdc-3/+ XX animals showed dosage compensation-specific defects and 40% ofsdc-3/sdc-3 XX animals were dead. Second,xol-1 XO animals, normally dead from inappropriately activated dosage compensation (10), were rescued byher-1 arrays. Thus, association of SDC-2 with multipleher-1 regulatory regions has functional consequences for dosage compensation. These genetic assays, together with theher-1 array assays, indicate that SDC-2 acts directly to repress her-1 transcription, thus initiating hermaphrodite sexual development.

sdc-3 is also required to repress her-1, but it does not trigger hermaphrodite development: XO animals engineered to overexpress SDC-3 develop as males (Table 1). To characterize the interaction between SDC-2 and SDC-3 in repressing her-1, we assessed the effect on SDC-2 of an sdc-3(Tra) mutation, which disrupts sex determination by derepressing her-1. This mutation has no effect on dosage compensation (15). SDC-2 failed to associate with her-1 arrays insdc-3(Tra) mutants but did associate with X chromosomes (Fig. 5, M to R). Thus, sdc-3(Tra) mutations derepressher-1 by preventing the association of SDC-2 with her-1regulatory regions, which demonstrates the interdependence of SDC-2 and SDC-3 and validates use of the assay to identify repressors ofher-1. These results show that SDC-2 has different requirements for its association with her-1 and X, and they reveal the basis for the separation in sex determination and dosage compensation functions of sdc-3 (15, 16,28).

SDC-2 is the pivotal sex-specific factor that initiates the hermaphrodite program of sexual development and activates dosage compensation (Fig. 5S). It participates directly in both processes through an association with chromatin, acting in one case as a strong gene-specific repressor and in the other as a weaker chromosome-wide repressor. The distinct modes of SDC-2 repression are consistent with SDC-2 being a transcriptional repressor that resembles no known transcription factors. SDC-2 triggers sexual development by inactivating the male sex-determining gene her-1, which is repressed at least 20-fold (23). In contrast, SDC-2 achieves dosage compensation by reducing X chromosome expression twofold. The extent of repression conferred by SDC-2 is likely specified by interactions with its protein partners. Repression of her-1requires interplay between SDC-2 and the sex determination domain of SDC-3. In contrast, modulation of X expression requires SDC-2 to collaborate with the dosage compensation machinery and the dosage compensation-specific domains of SDC-3, which includes a pair of zinc fingers (16).

How can a robust transcriptional repressor also trigger assembly of the dosage compensation complex on X chromosomes? SDC-2 may activate dosage compensation by first associating with X chromosomes, perhaps with SDC-3, and then recruiting other dosage compensation components to X, including chromosome segregation proteins. Alternatively, SDC-2 may coordinate the assembly of dosage compensation complexes off DNA; complete complexes would then recognize and associate with X chromosomes. In both cases, SDC-2 could confer chromosome specificity to dosage compensation by recognizing X chromosomes, as implied by the apparent association of SDC-2 with X in the absence of intact dosage compensation complexes.

  • * To whom correspondence should be addressed. E-mail: bjmeyer{at}


Stay Connected to Science

Navigate This Article