In Drosophila melanogaster, X chromosome dosage compensation is achieved by doubling the transcription of most X-linked genes. The male-specific lethal (MSL) complex is required for this process and binds to hundreds of sites on the male X chromosome. The MSL1 protein is essential for X chromosome binding and serves as a central scaffold for MSL complex assembly. The amino-terminal region of MSL1 binds to hundreds of sites on the X chromosome in normal males but only to approximately 30 high-affinity sites in the absence of endogenous MSL1. Binding to the high-affinity sites requires a basic motif at the amino terminus that is conserved among Drosophila species. X chromosome binding also requires a conserved leucine zipper-like motif that binds to MSL2. A glycine-rich motif between the basic and leucine-zipper-like motifs mediates MSL1 self-association in vitro and binding of the amino-terminal region of MSL1 to the MSL complex assembled on the male X chromosome. It is proposed that the basic region may mediate DNA binding and that the glycine-rich region may promote the association of MSL complexes to closely adjacent sites on the X chromosome (Li, 2005).

Significant progress has been made in understanding the regulation of transcription of individual genes in eukaryotes. It has also become apparent that the transcription of many genes within a particular region of a chromosome can be cocoordinately regulated by mechanisms that involve changes in the local chromosome structure. The most dramatic example of this is X chromosome dosage compensation, the equalization of X-linked gene transcription between XY males and XX females. In mammals, this is achieved by compacting one X chromosome in females into an inactive heterochromatic structure. In Drosophila melanogaster, the male X chromosome is modified to a more open structure that somehow leads to a precise doubling of transcription of nearly all the predicted 2,240 X-linked genes (Li, 2005 and references therein).

Dosage compensation in Drosophila requires the ribonucleoprotein male-specific lethal (MSL) complex. The complex binds to hundreds of sites along the male X chromosome. The core protein components are MSL1, MSL2, MSL3, MLE, and MOF. Loss-of-function mutations in any of the genes encoding these proteins lead to male-specific lethality, due to a failure in dosage compensatation. A sixth protein, JIL1, preferentially associates with the male X chromosome and has been shown to coimmunoprecipitate with components of the MSL complex. Loss-of-function jil1 mutations are, however, lethal to both sexes, indicating a vital role for JIL1 in addition to X chromosome dosage compensation. The two noncoding RNA components of the complex, roX1 and roX2, share little sequence similarity but are genetically redundant and appear to be functionally interchangeable. The MSL complex does not assemble in females since one protein component, MSL2, is absent (Li, 2005).

MSL1 plays a central role in assembly of the MSL complex (Scott, 2000). The amino-terminal domain of MSL1 binds to MSL2 (Copps, 1998, Scott, 2000). It has been suggested that the interaction between MSL1 and MSL2 was via predicted amphipathic coiled-coil α-helical regions that are found within the interacting domains (Scott, 2000). In addition, the carboxyl-terminal domain of MSL1 binds to both MSL3 and MOF (Scott, 2000). Subsequent studies have shown that MOF and MSL3 bind to adjacent regions in the MSL1 carboxyl-terminal domain (Morales, 2004). Further, formation of the MSL1/MSL3/MOF complex leads to a significant increase in the histone acetylase activity of MOF (Morales, 2004), which preferentially acetylates histone H4 at lysine 16. Both MSL3 and MOF have been shown to bind RNA nonspecifically in vitro and thus may have a role in incorporation of roX RNAs into the complex (Akhtar, 2000). Incorporation of MLE into the complex is presumably via interaction with the roX RNA, since MLE contains an RNA binding domain but does not appear to interact (Copps, 1998) with any of the other protein components of the complex (Li, 2005).

While progress has been made in understanding MSL complex assembly, how the complex specifically binds to hundreds of sites on the male X chromosome and then upregulates transcription so precisely remains poorly understood. One model for X chromosome binding is that the first step involves recognition of approximately 30 'high-affinity' or 'chromatin entry' sites on the X chromosome by the MSL1/MSL2 dimeric complex (Kageyama, 2001). Additional binding to the high-affinity sites within the roX1 and roX2 genes requires MLE. Subsequently, the other components bind, and the complex then spreads along the chromosome to hundreds of other sites. One of the key observations that support this model is that the MSL1/MSL2 complex binds to the high-affinity sites in the absence of MSL3, MLE, or MOF (Gu, 1998, Lyman, 1997). MSL1 and MSL2, however, do not contain any of the well-characterized DNA binding domains. Previously it was found that a deletion mutant of MSL1 that lacks the first 84 amino acids (aa) binds to MSL2, MSL3, and MOF but fails to bind to the X chromosome (Scott, 2000). This result suggested an important role for the amino-terminal region in X chromosome binding. This study shows that a conserved basic segment at the amino terminus of MSL1 is essential for binding to the high-affinity sites on the X chromosome in the absence of endogenous MSL1. It was also found that the adjacent region of MSL1 mediates MSL1 self-association. Lastly, the importance of the predicted coiled-coil region of MSL1 in binding to MSL2 was confirmed and that this interaction was shown to be essential for binding of the amino-terminal region of MSL1 to the X chromosome (Li, 2005).

It is almost 14 years since the first report that a component of the MSL complex selectively binds to hundreds of sites on the male X chromosome (Kuroda, 1991). Although neither MSL1 nor MSL2 has a readily identifiable DNA binding domain, both are essential for binding to the high-affinity sites on the X chromosome (Lyman, 1997). This study shows that a conserved basic motif at the amino terminus of MSL1 is required for binding to high-affinity sites. Short stretches of basic amino acids are involved in DNA recognition by basic leucine zipper (bZIP) and basic helix-loop-helix proteins. By analogy, a possible role for the basic region of MSL1 is to recognize a DNA sequence within the ~30 high-affinity sites on the male X chromosome. Consistent with this possibility, it was found that replacement of three of the conserved basic amino acids at positions 3, 4, and 6 by alanine eliminated binding of the amino-terminal region of MSL1 to all but five of the high-affinity X chromosome binding sites. Two of the five sites mapped to the location of the roX genes (roX1 at 3F and roX2 at 10C). It is possible that binding to these sites could be via association with the RNA components of the complex rather than via DNA recognition. Two conserved aromatic amino acids in the basic region appear to be important for binding specificity, since alanine substitution leads to increased binding to the autosomes. Aromatic and nonpolar amino acids in the basic domain of bZIP protein C/EBPα are important for DNA recognition and binding specificity, respectively. Investigating these possibilities will require in vitro binding studies with DNA sequences from the three high-affinity sites that have been identified. It is also possible that the amino terminus of MSL1 could bind RNA, since several proteins bind to RNA via basic-rich motifs. If so, the MSL1/MSL2 complex would associate with the nascent RNA of genes transcribed within the high-affinity sites. However, binding of MSL1 to the X chromosome is not disrupted by RNaseA treatment (Buscaino, 2003). This suggests that it is more likely that the MSL1/MSL2 heteromeric complex recognizes DNA sequences within the high-affinity sites (Li, 2005).

bZIP and basic helix-loop-helix proteins bind to DNA as dimers with bZIP dimers, forming coiled coil structures. Coiled coil domains contain a heptad repeat of the form (a-b-c-d-e-f-g)_n, where positions a and d are commonly occupied by apolar residues. Oligomerization then occurs through the formation of a multistranded, α-helical coiled coil in which a and d residues become internalized and hence shielded from the aqueous environment. It has been proposed that the short heptad substructures observed in the sequences of both MSL1 and MSL2 (residues 128 to 143 and 25 to 40, respectively) could provide a simple means by which chain dimerization could be effected in vivo (Scott, 2000). This study identified highly conserved apolar regions that lay immediately N terminal to the heptad motif in both chains (residues 113 to 121 and 5 to 14, respectively). Alanine substitution of four amino acids in the MSL1 apolar region eliminated binding to MSL2 in vitro and in vivo. A possible explanation for the critical importance of the MSL1 apolar region is that this acts as a trigger motif that facilitates coiled-coil formation. In the case of long heptad-containing regions, trigger motifs are sometimes used in the sequence to provide a short length of highly stable coiled coil that acts as a nucleating point for subsequent coiled-coil formation. For short lengths of coiled coil, however, other features may play an important role in either stabilizing or facilitating the formation of coiled-coil structure. If the apolar region of MSL1 does serve as a trigger for dimerization, then the first turn of the α-helix would be expected to be important in zipping together the two proteins. Consistent with this suggestion, it was found alanine substitution of the first two apolar amino acids in the a position and of the charged amino acids in the e and g positions of the first heptad decreased binding to MSL2 in vitro. It should also be noted that the RING finger domain of MSL2, which immediately follows the short heptad motif, is also important for binding (Copps, 1998) to MSL1 (Li, 2005).

Remarkably, it was found that the amino-terminal domain of MSL1 lacking the basic motif (Δ26HA) binds to all sites on the male X chromosome. This appears to be because Δ26HA binds to full-length MSL1 incorporated into the complete MSL complex. It was found that a glycine-rich region between the basic and coiled-coil motifs facilitated MSL1 self-association in vitro and binding to the MSL complex in vivo. The glycine-rich and leucine zipper-like motifs appear to function independently, as MSL1 self-association does not require MSL2 and deletion of the glycine-rich motif (e.g., Δ84HA) does not disrupt binding to MSL2. However, binding to the MSL complex in vivo does require interaction with MSL2; a mutation (mut_apo) that disrupts binding of MSL1 to MSL2 in vitro also eliminates binding to the male X chromosome, despite containing a complete glycine-rich motif. An explanation for these observations is that the MSL1NHA:MSL2 heterodimer binds to sites on the X chromosome immediately adjacent to sites occupied by the endogenous MSL complex and that the binding is stabilized by association of MSL1NHA with MSL1 in the complex. The 18D10 high-affinity site appears to consist of a cluster of sites of intermediate or weak affinity for the MSL complex. It is likely that stable binding to the X chromosome involves some cooperativity between MSL complexes bound to adjacent sites of differing affinity. MSL1 self-association may then be important in cooperative binding of MSL complexes to the male X chromosome, but testing this proposal will require evaluation of a series of alanine substitution mutations within the glycine-rich region. Alternatively, the results do not preclude the possibility that MSL1NHA is recruited to the male X chromosome by interaction with both MSL2 and MSL1 in prebound MSL complex. Interaction with MSL1 in the complex must also be necessary, since Δ84HA does not bind to MSL1 or male X chromosome, yet binds to MSL2 (Li, 2005).