Studies on the short range spreading of the male specific lethal (MSL) complex on the X chromosome in Drosophila

The male specific lethal (MSL) complex consists of at least six proteins (MSL-1–3, MOF, MLE, JIL1) and two non-coding RNAs (Kelley and Kuroda, 1995; Meller et al., 1997). In males the complex binds to many sites on the X chromosome (Baker et al., 1994; Kelley et al., 1999; Demakova et al., 2003). Its presence on the male X chromosome has been hypothesized to bring about dosage compensation (Kuroda et al., 1991), but other evidence suggests a role in modifying the impact of chromosomal imbalance on genomic gene expression (Hiebert and Birchler, 1994; Bhadra et al., 1999; Pal Bhadra et al., 2005). In females the MSL complex is not formed due to the translational repression of the msl-2 mRNA by the Sex-lethal protein (SXL) (Bashaw and Baker, 1997; Kelley et al., 1997). Ectopic expression of MSL-2 in females results in binding to the two X chromosomes (Kelley et al., 1995). MSL-1 and MSL-2 are core components of the complex, and both are required for the binding of the complex to the X (Lyman et al., 1997). The expression levels of the MSL complex are critical for its correct targeting to the X chromosome (Demakova et al., 2003). The concept that the MSL complex could spread from nucleation sites was initially proposed from two types of evidence (Bhadra et al., 1999; Kelley et al., 1999). Low level of MSL-2 in transgenic females results in a reduced number of binding sites (Kelley et al., 1997), while over-expression of MSL-2 in males extends the MSL complex spreading from roX transgenes into flanking autosomal chromatin (Park et al., 2002). The msl mutations are also known to eliminate chromosomal binding of the complex to the X chromosome (Kelley et al., 1997; Lyman et al., 1997).

The mechanism of how the MSL complex recognizes its targets is unknown. The ‘chromatin entry sites’ model suggests that the MSL complex initially assembles at ∼35 chromatin entry sites and subsequently spreads bi-directionally along the chromosome (Kelley et al., 1999). This model was proposed based on the observation that a limited number of binding sites (∼35) were detected in male mutants for mle, msl-3, or mof (Kelley et al., 1997; Lyman et al., 1997). Among those high affinity sites, two have been identified and correspond to the roX1 and roX2 genes. Autosomal roX transgenes recruit the MSL complex from which it then spreads into flanking regions (Oh et al., 2003). Further studies revealed that the consensus binding region in both roX1 and roX2 is within male-specific DNaseI hypersensitive sites (DHS) and is represented by the two core sequences GAGAG and CTCTC (Kageyama et al., 2001; Park et al., 2003). The DHS were later found not to be required for initiation of cis-spreading of the MSL complex, and the roX RNA plays an important role in spreading (Bai et al., 2004). Further studies using ChIP-chip and ChIP-seq have extended the number of chromatin entry sites to potentially 150 or more (Alekseyenko et al., 2008). The ‘entry site’ model was deemed an oversimplification by further studies on chromosome transpositions, which showed that any substantial portion of the X chromosome displays a normal binding when inserted into an autosome, regardless of the presence of a previously defined entry site (Fagegaltier and Baker, 2004). This model implies that binding activity of the MSL complex to the X is dependent on the affinity of the complex to target sites and there is no spreading of the complex. This model was further tested in studies using chromatin immunoprecipitation (ChIP) (Dahlsveen et al., 2006), and the common elements for complex recruitment were analyzed.

In the present study, MSL spreading was evaluated using transposed fragments on both the X chromosome and autosomes by monitoring MSL complex recruitment and spreading. A series of autosomal insertions of X segments derived from around the white locus were examined for their ability to bind the MSL complex. Immunostaining followed by fluorescence in situ hybridization (FISH) revealed that not all X to A (autosome) transposition fragments accumulate the MSL complex at insertion sites but larger sections of the X will attract the complex that extends over sequences in the smaller insertions that alone have no MSL complex. In addition, by examining MSL binding in Tp(3:1)O5 flies, a transposition strain containing an autosomal segment present in the X, a limited spreading of the MSL complex at the junction into the autosomal sequences was found. Chromatin immunoprecipitation was used to confirm the nature of MSL complex spreading to the autosome segment. The results suggest that spreading of the complex does indeed occur but is limited under these circumstances.

Flies were cultured on cornmeal dextrose medium at 25°C. All the transposition stocks were obtained from the Bloomington Drosophila Stock Center (Indiana University). Their genotypes are: y w– rst–/y+Y; TE89/TE89; z[1] w[11E4]/Dp(1:Y)y[+]; Tp(1:2) TE34C, kuz[3]/CyO; Tp(1:3)w[zh], sc[1] z[1] w[zh]; In(2LR)DTD86, Dp(1:2)w^70h ho2/In(2LR)O, Cy dp[lvl] pr cn2; Df(1)N-71h/C(1)M3, y[2]; Dp(1:2)^51b/+; and Tp(3:1)O5; D¹/+ and C(1)DX, y¹ f¹; +/+. The breakpoints of the X–autosome transpositions are listed in Fig. 1. To produce females with ectopic expression of MSL-2, a transgene [(w⁺)H83M2–6I] of a P-element msl-2 construct with a mini-white reporter gene (Kelleyet al., 1995) was introduced by crossing heterozygous males of w–/Y; (w⁺)H83M2–6I/TM3 GFP with females of the transposition line. For ISWI mutant analysis, genetic crosses for generating the ISWI mutant with a TE89 insertion are diagrammed in supplementary figure 1 (online suppl. fig. 1; for suppl. material see www.karger.com?doi=10.1159/000207524). Non-GFP third instar male larvae were selected for immunostaining.

Polytene chromosomes from the third instar larvae were dissected, fixed, and processed for antibody staining according to the protocols as described (Kuroda et al., 1991; Bhadra et al., 1999). In brief, salivary glands were dissected in 0.7% NaCl and fixed in phosphate-buffered saline (PBS) (136 mM NaCl, 1.1 mM K₂HPO₄, 2.7 mM KCl, 8.0 mM Na₂HPO₄, pH 7.3) solution containing 0.1% Triton X-100 and 3.7% formaldehyde on a siliconized coverslip for 1 min and then in 50% acetic acid, 3.7% formaldehyde for 2–5 min. The coverslip was picked up with the slide and inverted. The glands were squashed and the slides were placed in liquid nitrogen or on dry ice for the separation of the coverslip from the slide. After the removal of all coverslips, the slides were washed in PBS and blocked with PBT (PBS, 1% BSA, 0.2% Triton X-100, 0.02% azide) for 30 min. The primary antibody binding was performed at an appropriate dilution in PBT at 4°C overnight. The secondary antibody binding conjugated with fluorophore (diluted 1:100 to 1:200 in PBT) was performed at room temperature for 30 min to 3 h. The slides were mounted with Vectashield mounting medium containing 4′,6-diamidino-2-phenyl- indole (DAPI) (Vector Laboratory, Inc. Burlingame, CA) and examined with a Zeiss fluorescence microscope (Carl Zeiss, Inc, Oberkochen, Germany). The images were prepared using Adobe Photoshop 7.0 software.

Probe preparation. All probes were prepared from PCR products fluorochrome-labeled by nick translation with TexasRed CTP (red) (Perkin Elmer) or AlexaFluor488-dUTP (green) (Invitrogen). Nick translation was performed according to a protocol modified from Wiegant et al. (1996). In brief, a 50 µl mixture of DNA (5 µg) with non-labeled dNTPs (0.2 mM), TexasRed CTP or AlexaFluor488-dUTP (0.05 mM), DNA polymerase I (4 U/µl), and DNase (2 mU/µl) was incubated for 2 h at 15°C. After nick translation, the probes were purified with 1× TE saturated Bio-Gel P-60 (Bio-Rad Laboratories) and ethanol-precipitated using 50 µg autoclaved salmon sperm DNA as carrier. The final pellet was dissolved in 25 µl of 1× TE, 2× SSC buffer (pH 7.0) for FISH.

Immuno-FISH. FISH was performed immediately after the second washing of slides labeled with the secondary antibody forimmunostaining. Slides were fixed in 10% formaldehyde for 10 min and sequentially washed in PBS and 100% ethanol. After slides were dried, 7 µl of 1× TE, 2× SSC with 140 ng/ml of autoclaved salmon sperm DNA (Sigma) were applied to the slides at the center of the chromosome spreads. After the application of a mineral oil-coated plastic coverslip, the slides, together with the probes, were heated in a metal tray at 100°C for 5 min for denaturation. After cooling on ice, 7 µl of the probes (diluted to 20% of original) were applied to the slides for hybridization at 55°C overnight in a humidity chamber. After hybridization, slides were washed in 2× SSC for 20 min at 55°C and then mounted with Vectashield mounting medium containing DAPI as a counterstain.

Chromatin samples were prepared from third instar larvae of either males or females as described (www.igh.cnrs.fr/equip/cavalli/link.labgoodies.html; Chua et al., 2001). Briefly, 150–200 mg of third instar larvae (sufficient for four independent immunoprecipitations) was homogenized in 1.8% formaldehyde of B1 buffer (60 mM KCl, 15 mM NaCl, 4 mM MgCl₂, 15 mM HEPES (pH 7.6), 0.5% Triton X-100, 0.5 mM DTT, 10 mM sodium butyrate, EDTA-free protease inhibitor cocktail (Roche)) for 15 min (total time starting from beginning of homogenization) at room temperature (RT). Glycine was added to a final concentration of 225 mM and the homogenate was mixed and incubated for 5 min at RT, followed by centrifugation at 4000 g at 4°C with discard of the supernatant. The pellet was washed three times in B1 buffer, and once in lysis buffer (140 mM NaCl, 15 mM HEPES pH 7.6, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.5 mM DTT, 0.1% sodium deoxycholate, 0.05% SDS, 10 mM sodium butyrate, with protease inhibitors). The same conditions applied for pelleting following each wash. The pellet was resuspended in 0.5 ml lysis buffer with addition of SDS and N-lauroylsarcosine to 0.5% each. The resuspended cross-linked material was incubated for 10 min at 4°C in a rotating wheel and then subjected to sonication. Each sample was sonicated 5–6 times for 30 s each, using a Fisher Scientific Sonicator (Model w375, Heat Systems-ultrasonic, Inc.) at output 3 and duty cycle 30%. The sonicated chromatin solution was centrifuged 5 min at RT at maximum speed to remove the debris. The supernatant was purified and concentrated with a Centricon YM-100 column (4212), pre-blocked by BSA (1 mg/ml), and centrifuged at 1000 g three times for 40 min while adding ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 0.6 mM EGTA, 16.7 mM Tris-HCl, pH 8.1, 167 mM NaCl, 10 mM sodium butyrate, 0.5 mM DTT, with protease inhibitors). 100 µl protein G agarose/salmon sperm DNA beads (Upstate Scientific Technology) equilibrated with ChIP buffer was added to samples for pre-clearance at 4°C on a rotation wheel for 4 h. Immunoprecipitation (IP) was then performed on a sample of the supernatant collected from pre-incubation using 3 µl/IP of MSL-2 polyclonal antibody (Santa Cruz) incubated overnight at 4°C, using no antibody as a control. After antibody binding, 60 µl of protein G agarose beads were added and incubated on a rotating wheel at 4°C for 4 h. Beads were then subsequently washed with a low salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.1, 150 mM NaCl), a high salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.1, 500 mM NaCl), a lithium chloride wash buffer (0.25 M LiCl, 1% IGEPAL CA630, 1% deoxycholic acid (sodium salt), 1 mM EDTA, 0.5 mM EGTA, 10 mM Tris, pH 8.1, 10 mM sodium butyrate, 0.5 mM DTT), and then two washes of TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). The immunoprecipitated protein/DNA was eluted in 500 µl elution buffer (1% SDS, 0.1 M NaHCO₃) at 65°C for 15 min. Beads were removed and then the samples were reverse cross-linked, with addition of 25 µl of 4 M NaCl, and 5 µl of Proteinase K (10 mg/ml), and incubated at 65° for 5 h. Following reversal of cross-links, the precipitated DNA was recovered using phenol/chloroform extraction and ethanol precipitation, and then re-suspended in 20 µl of nuclease-free H₂O. 1 µl of this DNA was used as a template for real-time PCR.

Real-time PCR was performed on the resuspended DNA from ChIP described above using the Power SYBR Green PCR Master Mix kit (ABI). Three sets of primers were chosen. Two pairs (5′-88B and 92C-3′) were designed for cytologically visualized binding regions adjacent to each breakpoint, and one pair (88B-150K) was designed for a non-binding region as an internal control. The primer sequences were as follows: for 5′-88B: 5′-ACT TGT TGT GAA TTG TGC CGT GGG-3′ (forward) and 5′-TCA AGC CAT CAT GTT GAA GTG GCG-3′ (reverse), for 92C-3′: 5′-TCC ACC CTC AAT CTT GGA GGA ACA-3′ (forward) and 5′-TTT GGG ACA TGG GCT TGG GTT T-3′ (reverse), and for 88B-150K: 5′-GGC TGG GTG CAG CGA AAT GAA TAA-3′ (forward) and 5′-TGC CTC GCC AAT TGA GTT TAT GCC-3′ (reverse). PCR amplification and fluorescence detection were performed in a 25 µl final volume containing 100 ng of each primer and 1 µl of the immunoprecipitated DNA using the Applied Biosystems 7300 Real-Time PCR System with a thermocycler profile of 50°C for 2 min, 95°C for 10 min, and followed by 40 cycles of 95°C for 15 s, 55°C for 1 min. Each sample was processed three times. Primers for β-tubulin (forward: 5′-AGC TCA GCA CCC TCT GTG TAA T-3′ and reverse: 5′-AGC TGG AGC GCA TCA ATG TGT A-3′) were used as the internal reference for normalizing the amount of immunoprecipitated DNA on the basis of the different Ct value. The relative quantification (RQ) for each primer was determined according to the ΔCt analysis based on the manufacturer’s instructions (Applied Biosystems 7300 Real-Time PCR system, sequence detection software Version1. 3.1.).

Previous studies had suggested that very small X chromosome derived fragments could not attract the MSL complex at autosomal positions, which were, however, bound by the complex if present in a larger context on the X chromosome or in a larger X–autosomal transposition (Bhadra et al., 1999). This finding led to the proposal that the MSL complex could spread from nucleation sites (Bhadra et al., 1999). Indeed, extensive spreading was subsequently found from transgenes carrying the roX genes (Kelley et al., 1999), confirming the concept of spreading. To examine this issue further, the MSL complex binding was first examined in various autosomal insertions of X fragments. The fragments tested were all derived from around the w gene and were of differing lengths, including the transposing elements (TEs) and transpositions depicted in Fig. 1. The w locus is located in the 3C2 region of the salivary gland polytene map and is about 5.9 kb in length. The TEs tested in this study included TE89 and Tp(1:2)TE34C, both having the smallest piece of X chromosome examined in this study (3C2-3C5). An independent small insertion examined was Tp(1:3)w^zh (3C2-3C3). Another insertion, Dp(1:2)^51b, was reported to extend more proximally on the X than the TEs, with breakpoints of 3C2-3D6 (www.flybase.org). Another insertion, Dp(1:2)w^70h (3A7-3C2-3), has a proximal breakpoint within the TE limits but extends much more distally. From immunostaining and FISH studies in normal males, the TEs are not labeled at all with anti-MSL-2 antibodies, which indicates that no MSL complex is recruited to them (Fig. 2). Similarly, there is no complex binding observed for the Tp(1:3)w^zh insertion (Fig. 2C). Two larger insertions, Dp(1:2)w^70h and Dp(1:2)^51b, were found to attract the MSL complex (Fig. 2D, E), and a similar recruitment was also found in females with ectopic expression of MSL-2, generated by crossing the msl-2 transgene (w+)H83M2–6I into these transposition lines (Fig. 2F, G). This result indicates that if the X–autosome transposition can recruit the MSL complex in normal males, the inserted piece of X chromosome is also sufficient to recruit the MSL complex in females with ectopic expression of MSL-2.

The ISWI mutation causes a bloated chromosome phenotype that is much stronger for the X chromosome in the presence of the MSL complex (Corona et al., 2002). The related mutation, nurf310, causes a similar effect that can be visualized for a single transgene (Badenhorst et al., 2002; Bai et al., 2007). To test further if the TE segment associates with the MSL complex, binding was examined in the ISWI mutant together with TE89. The TE89 was crossed into ISWI mutant lines, and the polytene chromosomes were probed with anti-MSL-2 antibodies. As indicated in Fig. 3, the TE insertion, in the ISWI mutant background, does not change in morphology to produce a comparable bloated site in the autosome relative to the location on the normal X (Fig. 3).

It is interesting that the TEs and Tp(1:3)w^zh do not recruit the MSL complex to their autosomal insertion sites, while the two larger duplications show the MSL complex colocalizing with the white FISH signal. Comparing one of the duplications Dp(1:2)^51b to the Tp(1:3)w^zh insertion, it can be hypothesized that a nucleation site for MSL binding exists between the proximal breakpoints of the Tp(1:3)w^zh and the Dp(1:2)^51b duplication (Fig. 1). Similarly, the sequence between the distal breakpoints of the TEs and Dp(1:2)w^70h (Fig. 1) might contain another nucleation site.

Cytologically, the genes roughest(rst) and verticals(vt) are located proximal to the w gene. In situ hybridization to polytene chromosomes of wild type shows that those genes are close together in the 3C region (Fig. 4A–D). The transposed insertions TE89, Tp(1:3)w^zh and Dp(1:2)^51b, overlapping at their distal ends (see Fig. 1), were examined for the presence of the rst (3C3-3C4) and/or vt(3C5+) genes. The results show that both rst and vt were observed in Dp(1:2)^51b (Fig. 4I–L), but not in the Tp(1:3)w^zh insertion (Fig. 4E–H). The rst gene was also contained in the TE89 insertion (Fig. 4M–P). Thus, the Tp(1:3)w^zh insertion contains less X chromosomal material than the TEs. Because the MSL complex covers the extent of the larger insert, which contains the sequences of the smaller ones, this result suggests that a site in the larger construct nucleates short range spreading, as previously suggested (Bhadra et al., 1999).

Kelley et al. (1999) have found that autosomal insertion of the roX genes will recruit the MSL complex and that the MSL complex spreads into the adjacent autosomal regions. Another region of 18D10 on the X chromosome was also reported to have a similar binding pattern of MSL binding, when inserted into an autosome (Oh et al., 2004). However, the results from Fagegaltier and Baker (2004) and Oh et al. (2004) indicate that when an autosomal fragment was inserted into the X chromosome, there was no MSL complex binding observed on the insertion, suggesting no spreading of the MSL complex in normal males. On a large scale view, this is obvious (Fig. 5). However, to test the possibility of short range spreading, the binding of the MSL complex to Tp(3:1)O5 was examined. Immunostaining for MSL-2 in Tp(3:1)O5 male flies revealed the presence of limited spreading of the MSL complex at either side of this insert as revealed by the overlap of signals from the MSL complex and autosomal sequences labeled by FISH probes (Fig. 5A–C, E, F). The FISH probes were generated from different primers designed near each breakpoint on the autosomal fragment. The primer sequences, resulting product sizes, and binding regions on the autosome are summarized in Table 1. This result suggests that the MSL complex on the X chromosome can spread onto autosomal sequences, roughly limited to 4–5 kb in distance on each side of the insertion (Fig. 6).

The accuracy of colocalization might be obscured by cytological methods and photographic exposure time. In order to confirm the spreading of the MSL complex from the X into the autosomal regions, chromatin immunoprecipitation (ChIP) was performed using anti-MSL-2, followed by real-time PCR for amplification of DNA fragments immunoprecipitated from binding regions. Three pairs of primers were used to measure the enrichment of binding sequences (Fig. 7). Two pairs were designed within autosomal binding regions observed from in vivo immunostaining, one pair from upstream of the 88B binding region (Fig. 5C), and the other pair from downstream of the 92C binding region (Fig. 5F), designated 88B-5′ and 92C-3′, respectively. The third pair (called 88B-150K) located more internally at 88B and proximally about 150 kb beyond the binding region (Fig. 5D), was used as a control sequence (no recruitment of MSL complex on polytene chromosomes). Rather than using the entire binding sequence, the size was reduced to ∼220-bp fragments for immunoprecipitation detection. Females were also selected as a negative control compared to males for each pair of primers to test the enrichment of DNA sequence from immunoprecipitation, as there is no MSL-2 on the X or autosomes in females. Relative quantification of binding sequence was measured by real-time PCR. From this study, compared to the control IP, the 5′ end of 88B and the 3′ end of 92C are both enriched in the MSL-2-immumoprecipitation in males but not in females, and the internal region (88B-150K) also did not accumulate MSL-2 in chromatin immunoprecipitation in either males or females (Fig. 7). The ChIP results are in concordance with the polytene immunolabeling and observations (Fig. 5), suggesting that the MSL complex on the X chromosome can spread into autosomal sequences over a short range.

Using immunostaining and FISH, the recruitment of the MSL complex in different transposition insertions derived from around the w gene have been examined. Applying FISH for white in this study established the exact site for assaying MSL association. The results show that transpositions of sufficient length are able to recruit the MSL complex. Examination of TE89, the larger of the nonbinding autosomal inserts, in an ISWI background confirmed that it does not attract the MSL complex. ISWI mutants show a ‘bloated’ chromosome phenotype, and the effect is greatly amplified by the MSL complex (Corona et al., 2002; Pal Bhadra et al., 2005). If the TEs can recruit the MSL complex to its insertion site in an autosome, it should show this chromosomal phenotypic effect to some extent similarly to the nurf310 mutants (Bai et al., 2007).

The transposition Tp(1:3)w^zh males do not show binding at its insertion site of 61D (Fig. 2C). This result differs from the claims of Fagegaltier and Baker (2004), who reported the recruitment of the complex on this insertion site. However, our results were confirmed using a FISH probe of the w gene, a sequence contained in each insertion. By using this probe, another site of MSL binding on 3L was excluded (Fig. 2), which might account for this discrepancy. A combination of both TE89 and Tp(1:3)w^zh was also examined for the binding of the MSL complex; the labeling of the MSL complex was not seen at either insertion site as verified by w gene FISH (not shown). Unlike TEs and Tp(1:3)w^zh, the two larger insertions Dp(1:2)w^70h and Dp(1:2)^51b show binding to the MSL complex in males (Fig. 2D, E). This binding ability also occurs in females with ectopic expression of MSL-2 (Fig. 2F, G), showing that the level of ectopic MSL-2 is sufficient to recruit the MSL complex. These data suggest that MSL target determinants are present in these insertion sequences and are located outside of the limits of TE or Tp(1:3)w^zh (Fig. 1). These regions might contain a common consensus DNA motif comparable to other DNA binding fragments (DBFs) (Dahlsveen et al., 2006; Alekseyenko et al., 2008). In this study, the single genes rst and vt were probed, and we found that vt is present in Dp(1:2)^51b (Fig. 4I–L) but not in Tp(1:3)w^zh (Fig. 4E–H). The rst gene is excluded as containing a binding target because it is present in the TE insertion (Fig. 4M–P).

Most strikingly, using immunolabeling and FISH analyses, we visualized the spreading of the MSL complex into autosomal sequences in Tp(3;1)O5 (Fig. 5). The spreading is limited to a few kilobases (see Table 1), which is not as extensive as observed with roX transgenes that condition a long range spreading up to 1 Mb (Kageyama et al., 2001). ChIP analysis was performed with an MSL-2 antibody followed by real-time PCR using autosomal primers to confirm the local spreading. Female samples and the 80B-150kb primers (located near the genomic position 150 kb of 88B) were both used as negative controls. Autosomal sequences adjacent to the X breakpoints were precipitated at much greater amounts in males indicating that there is complex spreading from the X regions. Oh et al. (2004) examined Tp(3:1)O5 using only chromosomal immunostaining and concluded that there was no spreading into autosomal sequences, which is true in a global sense. Here, the combined immunolabeling and FISH, together with the more sensitive ChIP procedure, reveals limited spreading, which would likely not be detected in the aforementioned study. The results do indicate an affinity of X fragments for the MSL complex and constraints on extensive spreading into autosomal sequences, which is consistent with the general conclusions of Fagegaltier and Baker (2004), Oh et al. (2004), Dahlsveen et al. (2006), and Alekseyenko et al. (2008) concerning MSL spreading.

This work was supported by NIH grant R01 JM068042.