| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Genetics, Vol. 172, 991-1008, February 2006, Copyright © 2006
doi:10.1534/genetics.105.050567
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Wellcome Trust Centre for Cell Biology, University of Edinburgh, Edinburgh EH9 3JR, United Kingdom
1 Corresponding author: Wellcome Trust Centre for Cell Biology, University of Edinburgh, School of Biological Sciences, Institute of Cell Biology, Michael Swann Bldg., Mayfield Rd., Edinburgh EH9 3JR, United Kingdom.
E-mail: margarete.heck{at}ed.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSLITERATURE CITED |
|---|
In addition to influencing mitotic chromosome structure, condensin proteins are known to play roles during interphase (HAGSTROM and MEYER 2003). A functional link between mitotic chromosome condensation and global regulation of gene expression was initially demonstrated in Caenorhabditis elegans on the basis of the involvement of an SMC2/4 heterodimer in dosage compensation. Transcription from each of the X chromosomes is reduced twofold in hermaphrodites (XX) of this organism to match the level of X-linked gene expression in males (XO). The discovery that a variant SMC4-type protein (DPY-27) is an essential regulator of dosage compensation (through its association with the X chromosome) provided the first clue that SMC proteins might be involved in this process (CHUANG et al. 1994, 1996). Subsequently, MIX-1 was identified as an SMC2-type protein required for both mitosis and dosage compensation, whose restricted localization to the X chromosome was shown to depend on that of DPY-27 (LIEB et al. 1998). The dosage compensation complex also consists of at least two non-SMC subunits, including DPY-26 and DPY-28 (LIEB et al. 1996; HAGSTROM and MEYER 2003). On the other hand, the mitotic function of MIX-1 requires its association with a more conventional SMC4-type protein (HAGSTROM et al. 2002). It has therefore been suggested that MIX-1 may have been enlisted to the dosage compensation complex through the evolution of DPY-27 into a highly specialized SMC protein, altering the higher-order structure of X chromosomes by a mechanism potentially related to that involved in mitotic chromosome condensation.
Strikingly, the same dosage compensation complex is also recruited to the autosomal her-1 gene (a male sex-determination gene) to repress its transcription >20-fold and thereby elicit hermaphrodite differentiation (CHU et al. 2002). This therefore demonstrates a link between the roles of condensin SMC proteins in gene-specific and chromosome-wide transcriptional repression. Various non-SMC condensin subunits also appear to be required for transcriptional silencing in both Saccharomyces cerevisiae (BHALLA et al. 2002) and Drosophila melanogaster (LUPO et al. 2001). However, it remains to be seen if these proteins function during interphase as part of the canonical condensin complex or other alternative SMC-containing complexes. Similarly, Cnd2/CAP-H (a non-SMC condensin subunit) was recently shown to be required for the activity of Cds1 (a G2 checkpoint kinase) in Schizosaccharomyces pombe (AONO et al. 2002). Although all the other condensin components in S. pombe were also required for full Cds1 activity, the condensin SMC proteins (Cut3 and Cut14) differed from Cnd2/CAP-H as they were not required for viability in the presence of hydroxyurea during S phase (AONO et al. 2002). The contrasting interphase requirements for different condensin proteins in S. pombe therefore suggest that these proteins also may not necessarily act in the same complex during interphase as during mitosis.
In this article, we further explore the varied activities of condensin proteins in D. melanogaster by analyzing multiple alleles of gluon, encoding the Drosophila ortholog of SMC4, together with various mutations affecting other condensin components. A large number of gluon alleles were previously generated through imprecise excision of the P element in glu1 (Figure 1A), resulting in additional embryonic lethal lines (glu17C, glu88-37, and glu88-41B) displaying chromosome segregation defects similar to those described in glu88-82 mutant embryos in which chromatin bridges were most frequently observed (STEFFENSEN et al. 2001). We have therefore exploited these different gluon alleles, together with mutations affecting other components of the condensin I complex in Drosophila (SAVVIDOU et al. 2005), to explore possible functions of condensins in interphase as well as in mitosis. These include the embryonic lethal barrL305 allele, which is considered to be a null mutation of the barren gene (BHAT et al. 1996) encoding the Drosophila ortholog of the CAP-H subunit (HIRANO et al. 1997) and the embryonic lethal red sea64 mutant (PHILP 1998), which was subsequently shown by complementation analysis to be allelic to the Drosophila gene encoding CAP-G (referred to hereafter as the dcap-g64 allele). Finally, a larval lethal mutation affecting SMC2, known as smc2jsl2, was discovered when a collection of EMS mutants generated by Jeff Sekelsky (University of North Carolina at Chapel Hill) was screened using an antibody raised against SMC2 in Drosophila (SAVVIDOU et al. 2005). By examining the effect of each of these mutations on position-effect variegation (PEV) at different loci, we demonstrate further effects of different condensin components in modulating interphase gene expression (with the exception of CAP-D2, for which there are currently no mutants available) in addition to regulating the structure and behavior of mitotic chromosomes.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSLITERATURE CITED |
|---|
Selection of homozygous mutant animals:
Embryonic lethal condensin mutations were maintained over CyO balancer chromosomes expressing GFP under control of the Krüppel gene (CASSO et al. 2000) and homozygous mutant embryos were selected on the basis of the absence of fluorescence in the amnioserosa or the Bolwig's organs by sorting in halocarbon oil, as described (STEFFENSEN et al. 2001; SAVVIDOU et al. 2005). As the gluon and prospero genes are on the second and third chromosomes, respectively, the double mutants were maintained over T(2;3)TSTL, a translocation balancer chromosome allowing simultaneous isolation of homozygotes at both loci. The homozygous glu1;pros17 mutant embryos were then identified on the basis of the elevated number of mitotic nuclei in their central nervous system (resulting from the pros17 mutation). As a control, the pros17 mutation was balanced over the TM3-5 balancer chromosome expressing GFP under control of the Krüppel gene (CASSO et al. 2000) and pros17 homozygous embryos were selected on the basis of their lack of GFP expression. The larval lethal smc2jsl2 allele was maintained over a CyO-GFP balancer and homozygous larvae were similarly selected on the basis of the absence of GFP fluorescence.
Fixation of embryos and larval brains:
A fast formaldehyde fix was used to preserve cortical cytoskeletal structures such as microtubules on the basis of protocols described elsewhere (THERKAUF and HECK 1999; ROTHWELL and SULLIVAN 2000). Dechorionated embryos were promptly rinsed in filter-sterilized EBR (Ephrussi Beadle Ringer's solution; 130 mM NaCl, 4.7 mM KCl, 1.9 mM CaCl2, 10 mM Hepes, pH 6.9) and flushed into a small glass vial using heptane. The embryos were then fixed in equal volumes of heptane and 40% EM-grade formaldehyde (Electron Microscopy Sciences, Fort Washington, PA) with extremely gentle rocking for 3 min, ensuring that the embryos were spread in a monolayer at the interface between the two phases. The lower formaldehyde phase was then removed carefully using a long-tipped glass Pasteur pipette and replaced with an equal volume of Merck ARISTAR methanol to remove the vitelline membrane. The vial was capped again, shaken vigorously, vortexed for 15 sec, and then left to stand upright for 1 min. The devitellinized embryos that had sedimented to the bottom of the vial were removed using a long-tipped glass Pasteur pipette and transferred to a 1.5-ml microcentrifuge tube. The methanol was replaced with fresh Merck ARISTAR methanol four times to remove any remaining formaldehyde or heptane and mixed after each replacement. Fixed embryos were stored in methanol at –20° before processing for immunofluorescent detection as described (THERKAUF and HECK 1999; ROTHWELL and SULLIVAN 2000). Developmental staging of embryos was assessed according to the numbering scheme introduced by Mary Bownes (BOWNES 1975, 1982) and later refined by Eric Wieschaus and Christiane Nüsslein-Volhard (WIESCHAUS and NÜSSLEIN-VOLHARD 1998). Brains from third instar larvae were fixed and processed for immunofluorescence as described (BONACCORSI et al. 2000; MCHUGH et al. 2004).
Antibodies:
Primary antibodies were used for immunofluorescence or immunoblotting as follows: 
-tubulin (mouse monoclonal antibody B5-1-2 used at 1:1000; Sigma, St. Louis), phosphohistone H3 (P-H3; rabbit polyclonal used at 1:500; Upstate Biotechnology, Lake Placid, NY), P-H3 (mouse monoclonal used at 1:500; New England Biolabs, Beverly, MA), lamin (rabbit polyclonal used at 1:500; provided by Paul Fisher, State University of New York, Stony Brook, NY), Barren rabbit polyclonal used at 1:500 (BHAT et al. 1996), SMC2 rabbit polyclonal used at 1:500 (SAVVIDOU et al. 2005), SMC4 rabbit polyclonal used at 1:500 (STEFFENSEN et al. 2001), SCC1 rabbit polyclonal used at 1:500 (WARREN et al. 2000), and CAP-D2 rabbit polyclonal used at 1:10,000 for immunoblotting (SAVVIDOU et al. 2005). All fluorescent or horseradish-peroxidase (HRP)-conjugated secondary antibodies (Molecular Probes, Eugene, OR; Jackson ImmunoResearch Laboratories, West Grove, PA; and Amersham Biosciences, Buckinghamshire, UK) were used according to the manufacturer's instructions.
TUNEL labeling of staged embryos:
Combined antibody staining and terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) of staged embryos was performed using the Appligene Oncor Apotag kit, with modifications detailed below. Fixed embryos were permeabilized for 1 hr in PBS with 1% Triton X-100 and washed for 2 min in PBSTx (PBS with 0.05% Triton X-100) until all samples were ready for the Apotag reaction. The PBSTx was removed by careful aspiration and at least 100 µl equilibration buffer per sample was applied for 2 min. The buffer was then removed and at least 100 µl of working terminal deoxynucleotidyl transferase (TdT) enzyme solution (150 µl TdT enzyme with 350 µl reaction buffer) was immediately applied before the embryos were incubated for 1 hr at 37° with rotation on a mixer. The embryos were then washed in stop/wash solution with gentle agitation for 15 sec and incubated for 10–15 min. The embryos were washed three times for 10 min in PBSTx and blocked in PBS + 3% bovine serum albumin (Sigma) for 30 min at room temperature. They were next washed for 5 min in PBSTx and incubated overnight at 4° in mouse antidigoxigenin antibody (Boehringer, Indianapolis) diluted 1:100 in PBSTx. The embryos were then washed six times for 5 min each time before performing any additional antibody incubations as described earlier or staining of DNA with 0.05 µg/ml DAPI. The mouse antidigoxigenin antibody was detected using an Alexa Fluor 594 conjugated anti-mouse antibody (Molecular Probes).
Microscopy:
All preparations were examined with an Olympus Provis microscope, equipped with epifluorescence optics. Images were captured with an Orca II CCD camera (Hamamatsu, Bridgewater, NJ) and Smart Capture 2 software (Digital Scientific). Fixed embryos were mounted in 90% glycerol/10% PBS. Image analysis was performed using IPLab Spectrum version 3.1 and Adobe Photoshop version 4.0.
Preparation of larval brain samples for SDS-PAGE:
Ten to 20 brains were suspended in 50 µl lysis EBR (EBR with 10 mM EDTA, 10 mM dithiothreitol, 1:100 dilutions of each of the protease inhibitors PMSF, CLAP, and Trasylol). The material was homogenized using a Kontes pestle and mini-motor immediately before adding 25 µl of warm (65°) 3x SDS-PAGE sample buffer with DTT (6% SDS, 150 mM trizma base, 30% glycerol, 0.03% bromophenol blue, 6 mM EDTA, 100 mM dithiothreitol). A hole was then pierced in the top of tube and it was heated at 100° for 10 min. Finally, the tube was spun briefly and the sample was loaded onto a polyacrylamide gel or stored at –20°.
Immunoblotting:
Protein extracts were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Schleicher & Schuell, Keene, NH) in a trans-blot apparatus. Membranes were blocked in PBS + 0.1% Tween 20 (PBSTw) and 5% semiskimmed milk for 1 hr at room temperature (RT) and then incubated for 1–1.5 hr with the primary antibody in PBSTw. After washing three times for 5, 15, and 10 min with PBSTw, the membranes were incubated in an HRP-linked secondary antibody for 1 hr in PBSTw at RT. Finally, the membranes were washed as above in PBSTw and immunocomplexes were detected by enhanced chemiluminescence (Amersham Biosciences). The nitrocellulose was then exposed to Kodak XAR-5 film for autoradiography and films were developed using a Konica SRX-101A developer.
PhosphorImager analysis:
Samples were processed as described for immunoblotting but instead of using an HRP-conjugated secondary antibody, a Cy5-conjugated antibody (Jackson ImmunoResearch Laboratories) was used at a dilution of 1:200. The intensity of the signal was measured using a STORM 860 scanning phosphorImager and calculated using the ImageQuant program (Amersham). The intensity was finally normalized using -tubulin as a loading control. Values were corrected by subtracting the background of the corresponding lane from each band.
Analysis of position-effect variegation:
Male flies from each mutant line (balanced over CyO) were crossed against wm4h virgin females at 25° and the resultant straight-winged male progeny were selected and allowed to age for 4 days after eclosion. The flies were then frozen at –80° and the heads were decapitated by vortexing. The fly heads were isolated by repeated sieving through nylon meshes and 100 fly heads were counted and placed in a 1.5-ml microfuge tube. The drosopterin pigments were extracted and analyzed by spectrophotometry as described (ASHBURNER 1989), using the heads of similarly treated w1118 flies as a blank. Crosses between gluon virgin females and male flies carrying SUPor-P transposon insertions on the Y chromosome (YAN et al. 2002) or between condensin mutant males and virgin females carrying hsp70-w+ transgenes on the fourth chromosome (CRYDERMAN et al. 1999) were similarly performed as described above. Crosses with w1118 flies were used as controls for all reporters of PEV to generate hemizygotes in an otherwise wild-type background, thereby avoiding potential complications associated with background modifiers of PEV on balancer chromosomes (WALLRATH and ELGIN 1995).
Frequency distributions of variegated eye color in different mutant backgrounds were generated by photographing both eyes of individual 4-day-old male progeny against a uniform white background with an Olympus C-2020 Z digital camera. The level of pigmentation was then quantified using the histogram feature of Adobe Photoshop 4.0 to measure the mean pixel intensity of eye color in the images. To correct for any possible variations in ambient illumination during image capture, the luminosity of the traced eyes was subtracted from that of the white background. The resulting values were then rounded and a frequency distribution was plotted, using the ratio of the mean absorbance to the mean pixel intensity in each genetic background to calibrate the respective frequencies in terms of absorbance.
To examine the dominance relationships between mutant alleles identified as strong enhancers or suppressors of variegation, the In(1)wm4h chromosome was crossed into each of the mutant lines using wm4h;L/CyO virgin females. Reciprocal crosses between strong enhancers of variegation and strong suppressors of variegation were then performed and the absorbance of 100 heads from 4-day-old straight-winged male progeny was measured as described above.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSLITERATURE CITED |
|---|
|
As with glu88-82 homozygotes, no easily discernible effect was observed before metaphase in glu1;pros17 mutant embryos, although 11% of metaphase figures appeared abnormal, containing more than the usual chromosome complement (Figure 3A). These aneuploid or polyploid metaphases presumably resulted from segregation defects in a previous mitotic cycle and are consistent with the giant aneuploid nuclei observed when SMC4 is depleted by RNA interference (RNAi) in C. elegans (HAGSTROM et al. 2002). Furthermore, chromatin bridges were observed in glu1;pros17 mutant embryos at anaphase and telophase (Figure 3C) with similar frequencies to those previously observed in glu88-82 homozygotes (STEFFENSEN et al. 2001). Therefore, the high frequency of segregation defects observed in glu88-82 mutant embryos can be reproduced by increasing the mitotic activity of glu1 mutants and presumably depleting wild-type SMC4 levels more quickly than in glu1 mutants alone. Previously, we described how the proportion of mitotic cells in metaphase was significantly greater in embryos homozygous for the glu88-82 allele (STEFFENSEN et al. 2001). Although somewhat obscured by the greater proportion of cells in prophase/prometaphase when all mitotic stages are represented (Figure 3B), we were similarly able to detect a twofold increase in the proportion of mitotic cells in metaphase relative to anaphase in glu1;pros17 mutant embryos and a more than fourfold increase in the proportion of metaphase figures in dcap-g64 embryos (Figure 3C), in contrast to wild-type or pros17 controls in which the proportion of metaphase and anaphase figures is roughly equal.
|
Primers flanking the original P-element insertion site in the glu1 allele were used to amplify the remaining P element from homozygous mutant embryos. These PCR products were then sequenced and the nature of the lesion in each of these mutant alleles was determined (Figure 1). Like the original glu1 allele, each of these excision alleles contained stop codons in all reading frames at the 5'-end of the P element, so the mutants are predicted to be nulls. In each of these mutant alleles, a 122-residue truncated protein containing only the first 119 amino acids of SMC4 could be produced, lacking even the Walker A motif (Figure 1A). The sequence for the molecular lesions explains why the red eye color resulting from the mini-white transgene is absent in each of these excision alleles, as the entire coding sequence is deleted in most alleles (with the exception of glu88-41B, in which only the 3' half of the mini-white transgene is present). Furthermore, the sequence data also suggest that the glu88-82 allele might be a neomorph in which the Hsp70 promoter could possibly drive expression of a truncated SMC4 protein containing amino acids 195–1409. However, we have been unable to verify the presence of such a truncated protein as previously generated antibodies were raised against an N-terminal fragment, corresponding to amino acids 80–303 of wild-type SMC4 (STEFFENSEN et al. 2001). Sequencing of the molecular lesion in glu17C revealed a large deletion of 
31 kb, resulting in a knockout for SMC4. Although a number of other genes were also removed by the deletion, the glu17C allele was found to affect PEV in the same way as most other gluon alleles (as described below).
Sequencing of the dcap-g64 allele confirmed that the gene encoding CAP-G was indeed affected in these mutants, revealing a G954T transversion that results in a D254Y substitution. This mutation is significant as it affects a residue that is invariant among CAP-G sequences from different organisms and is found in a predicted HEAT repeat in the most conserved region of the protein. The discovery of HEAT repeats in CAP-G and CAP-D2 orthologs from S. pombe, S. cerevisiae, and Xenopus was previously thought to signify possible regions of interaction between condensin proteins (NEUWALD and HIRANO 2000), as these tandemly repeated bihelical structures are thought to play roles in protein-protein interactions (ANDRADE and BORK 1995). Indeed, such interactions were subsequently confirmed by co-immunoprecipitation experiments in which the HEAT-repeat-containing N-terminal moiety of CAP-D2 was shown to be required for interactions with other condensin proteins while the CAP-D2 C-terminal region interacted specifically with CAP-G (BALL et al. 2002). We therefore speculate that the replacement of an invariant aspartate residue with a tyrosine in a highly conserved region of CAP-G may disrupt a salt bridge required for formation of the bihelical HEAT repeat structure, thereby rendering the protein unable to interact normally with other condensin subunits.
Mitotic defects in smc2 mutant larvae:
We discovered a larval lethal mutation affecting SMC2, known as smc2jsl2, among a collection of EMS mutants after immunoblotting larval protein extracts with an antibody generated against the N-terminal region (amino acids 24–309) of SMC2 (SAVVIDOU et al. 2005) to detect lines in which the protein showed reduced expression or altered mobility. Unlike the previously characterized condensin mutant alleles, we did not detect mitotic phenotypes in smc2jsl2 embryos and the observed hatching frequencies were not consistent with embryonic lethality. However, the brains in smc2jsl2 homozygous third instar larvae were much smaller than either wild-type or heterozygous brains (Figure 4B), suggesting a proliferation defect. When third instar larval brains from heterozygous (control) and homozygous larvae were stained for phosphorylated histone H3, -tubulin, and DNA, the percentage of mitotic cells in smc2jsl2 homozygous mutant brains (0.3%) was found to be significantly lower than in heterozygous or wild-type controls (1.6%). Similarly, the percentage of mitotic cells in younger second instar smc2 jsl2 homozygous mutant brains was only 0.4%, compared to 1.3% in controls. Analysis of the few mitotic cells that were present in homozygous brains showed several defects in chromosome morphology. Chromosomes often appeared fuzzy with almost no sister-chromatid resolution while cells in telophase showed extensive chromatin bridges (Figure 4A). However, the stage of many mitotic cells (up to 56% in the case of third instar mutant larval brains) could not be easily determined as the chromosome structure was so abnormal, whereas cells from control heterozygous brains showed normal chromosome morphology and segregation (Figure 4A). All telophase- or anaphase-like cells in smc2jsl2 homozygotes had chromosome bridges. Furthermore, 2.7% of interphase cells in smc2jsl2 homozygotes were also joined by chromosome bridges (forming apparently binucleate or trinucleate cells), compared to 0.1% in control brains, presumably reflecting defective chromosome segregation and failed cytokinesis in previous mitotic cycles. Nevertheless, no obvious delays in mitotic progression were apparent in smc2jsl2 mutant brains, although quantitation was limited by the significantly reduced mitotic index and not easily classifiable mitotic figures.
|
-tubulin loading control.
The effect of condensins on gene expression:
Although the various subunits of the condensin complex are clearly essential for proper resolution of chromatids during mitosis (SAKA et al. 1994; BHAT et al. 1996; LIEB et al. 1998; SUTANI et al. 1999; STEFFENSEN et al. 2001; HAGSTROM et al. 2002; STEAR and ROTH 2002; HUDSON et al. 2003; SAVVIDOU et al. 2005), it seemed plausible that they might also play a role in chromatin behavior during interphase as condensin SMC proteins have long been known to play a role in dosage compensation in C. elegans (CHUANG et al. 1994; LIEB et al. 1998). Moreover, chromatin-bridging phenotypes similar to those of gluon mutants had been observed in Su(var)2055 embryos (KELLUM and ALBERTS 1995), which are null mutants for the gene-encoding heterochromatin protein 1 (HP1) with dosage-dependent effects on PEV (EISSENBERG et al. 1992; DORN et al. 1993). Therefore, we hypothesized that various components of the condensin complex might also be active during interphase, possibly influencing chromatin structure in a manner similar to their effects on mitotic chromosome structure. To explore the possible interphase role of condensin components, the various mutant alleles shown to display a consistent mitotic phenotype were crossed against white mottled 4 (wm4h) flies to examine their effect on PEV, together with reporters of PEV at other loci (Figure 5). The In(1)wm4h inversion places the white gene (conferring red eye color) near a region of pericentric heterochromatin on the X chromosome (TARTOF et al. 1984) and suppression of PEV by expansion of euchromatin typically results in increased expression of the white gene and thus in a higher proportion of red ommatidia. On the other hand, enhancement of PEV by expansion of heterochromatin leads to decreased expression and more white-eyed flies (SASS and HENIKOFF 1998). In addition to performing crosses with condensin mutants, established suppressors and enhancers of variegation were also examined to contextually evaluate the strength of effects on PEV seen with the different condensin mutants (aside from the original glu1 allele, as an intact white reporter gene in the P element precluded its use in these analyses).
|
|
Finally, several alleles of different condensin genes identified as particularly strong enhancers or suppressors of variegation were crossed against each other in a wm4h background to test for possible epistatic effects. Visual inspection of the predominantly red fly eyes in the resulting male progeny initially suggested that suppressors of variegation such as Su(var)2055 and glu88-82 might be partially dominant over enhancers of variegation such as glu17C, dcap-g64, and barrL305. If Su(var)2055 proved to be dominant over glu17C, dcap-g64, or barrL305, then this might have suggested that the interphase function of condensins acts upstream of HP1 to modulate transcription. Conversely, if these mutant alleles proved to be dominant over Su(var)2055, this might imply that condensin function may be downstream of HP1 activity, possibly reflecting HP1-regulated transcription of condensins as with other cell-cycle regulators (DE LUCIA et al. 2005). However, spectrophotometric measurements revealed intermediate eye pigment levels in these male fly heads (supplemental Figure 1 at http://www.genetics.org/supplemental/), indicating instead that the effect of the different mutant alleles on PEV was approximately dose dependent, with their independent influences combining additively in the assay. Similarly, intermediate eye pigment levels were observed in crosses between either glu88-82 and dcap-g64 or glu88-82 and barrL305 in a wm4h background, suggesting that any suppressor effects potentially due to expression of a neomorphic SMC4 C-terminal truncation in glu88-82 were not dominant over the enhancer activity of other condensin mutants, consistent with the observation of both predominantly red- and white-eyed glu88-82 heterozygotes (Figure 6, A and D).
The effect of condensin mutations on PEV at other loci:
The effect of the different condensin mutations on PEV was also examined at several other loci, including the bwD (brown-dominant) allele (SLATIS 1955; HENIKOFF et al. 1995), the SbV (stubble-variegated) allele (HAYASHI et al. 1990), and the Fab-7 repressed mini-white reporter gene in 5F24(25,2) flies (ZINK and PARO 1995). However, none of the crosses showed any effect of condensin mutations on PEV. This contrasts with previously published results in which the barrL305 allele was shown to suppress 5F24(25,2) variegation in adult female progeny (LUPO et al. 2001) but is nonetheless consistent with more recent findings concerning the lack of any discernible effect of dcap-g mutations on regulation of gene expression by the Fab-7 Polycomb-group response element (DEJ et al. 2004). However, as the wm4h inversion places the white gene near the bobbed multicopy rDNA array within pericentric heterochromatin, it is possible that the PEV assay performed with wm4h flies may be more sensitive to mutations affecting condensin components if such proteins are enriched at rDNA loci as in other species (FREEMAN et al. 2000; CABELLO et al. 2001; BHALLA et al. 2002; UZBEKOV et al. 2003; D'AMOURS et al. 2004; WANG et al. 2004) and since various condensin proteins in Drosophila have been shown to be enriched at mitotic centromeres (STEFFENSEN et al. 2001; SAVVIDOU et al. 2005). To evaluate whether the effects of condensin mutations on PEV were specific to particular chromosomal regions, additional crosses were performed between selected alleles and flies carrying transgenic white reporter genes at different chromosomal loci.
The potential specificity of effects at rDNA-proximal loci was further explored by performing crosses between selected gluon alleles and flies carrying insertions of the SUPor-P transposon at various intervals on the Y chromosome (YAN et al. 2002), which carries a white reporter gene flanked by su(Hw)-binding regions (ROSEMAN et al. 1995). Interestingly, no effect on PEV was seen with KV168 (J448) or KV81 (C882) lines, in which the SUPor-P transposon, respectively, maps to the h10 and h11–13 Y chromosomal "h" bands, although each of the different gluon alleles tested was shown to affect PEV in the KV128 (D285) and KV113 (B947) lines, in which the SUPor-P transposon, respectively, maps to the h20 and h22–24 Y chromosomal "h" bands and thus relatively close to the Y-bobbed rDNA locus (h20–21). In the case of crosses performed with the KV128 insertion (which maps closest to the rDNA locus), the glu17C and glu88-37 alleles clearly suppressed variegation of the white reporter gene in the SUPor-P transposon, although suppression of variegation was evidently weaker with the glu88-82 allele (Figure 7, A and B). By contrast, each of the different gluon alleles was found to enhance variegation in crosses with KV113 flies (Figure 7C), with no pigmented ommatidia observed in heterozygous mutant progeny. Together, these results further suggest that the rDNA proximal loci may be more sensitive to the effect of condensin mutations.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSLITERATURE CITED |
|---|
It is also worth noting that mitotic spindles appear normal in both gluon and dcap-g64 mutant embryos, in contrast to the spindle defects observed with different condensin mutations in S. cerevisiae (LAVOIE et al. 2000, 2002; OUSPENSKI et al. 2000). Normal mitotic spindles are also observed in the Drosophila barrL305 mutant (BHAT et al. 1996), chicken DT40 cells depleted of SMC2 (HUDSON et al. 2003), and SMC4 depleted by RNAi in C. elegans (HAGSTROM et al. 2002), so it appears that the condensin proteins do not affect spindle assembly in metazoans. Therefore, the segregation defects in gluon mutants do not appear to be due to defective spindle formation but are more consistent with a role for condensins in preventing chromosome entanglement. By contrast, abnormal mitotic spindles were frequently observed in the smc2jsl2 allele, although this might be partly explained by the difficulty in preparing neuroblast squashes from significantly smaller larval brains.
Consistent with our previous findings (STEFFENSEN et al. 2001), quantitation of the mitotic phenotypes in condensin mutant embryos suggests that most cells are delayed in prophase/prometaphase and metaphase, whereas the majority of cells reaching anaphase display chromatin bridges. The observation that embryonic lethal gluon and dcap-g alleles accumulate in metaphase while larval alleles do not is possibly because the embryonic lethal alleles are more severe (STEFFENSEN et al. 2001). On the other hand, the absence of any similar mitotic delay in cultured cells depleted of condensin components (COELHO et al. 2003; HUDSON et al. 2003) might reflect checkpoint differences in cultured cells from various species, as a metaphase delay together with chromatin bridges has also been observed in C. elegans mix-1 mutant embryos, in which the ortholog of SMC2 is affected (LIEB et al. 1998). A roughly threefold greater number of prometaphase figures was previously reported for dcap-g mutant embryos compared to wild type (DEJ et al. 2004), although different criteria were used to classify prometaphase figures in this study, with only the morphology of the chromosomes (but not the mitotic spindle or nuclear lamina) scored for the purposes of quantitation. Nevertheless, our analyses based on consistent classification of both mitotic spindle and chromosome behavior confirm the accumulation of mitotic cells in prophase/prometaphase in both embryonic lethal gluon and dcap-g alleles. On the basis of our observations and previous findings that mitotic chromosome condensation is delayed but not abolished in condensin mutants (STEFFENSEN et al. 2001; HUDSON et al. 2003), we speculate that condensation checkpoints may exist in intact cells, similar to previously reported topoisomerase II and DNA damage checkpoints (DOWNES et al. 1994; RIEDER and Cole 1998; GIMÉNEZ-ABIÁN et al. 2002; MIKHAILOV et al. 2004) or possible monitoring of kinetochore tension through the spindle checkpoint (MUSACCHIO and HARDWICK 2002). If so, the presence of such a checkpoint may prevent the most extreme condensation defects from occurring, so segregation defects are the main phenotypes usually observed in whole organisms with reduced condensin activity (LIEB et al. 1998; STEFFENSEN et al. 2001; HAGSTROM et al. 2002; STEAR and ROTH 2002). By contrast, more extreme chromosome structural defects or compromised kinetochore–microtubule interactions may be apparent only in cultured cells from different species (COELHO et al. 2003; HUDSON et al. 2003; HIROTA et al. 2004; ONO et al. 2004; SAVVIDOU et al. 2005).
Unfolding the role of condensins during interphase:
In contrast to previous reports consistently describing condensins primarily as suppressors of variegation (LUPO et al. 2001; DEJ et al. 2004), we have found that the same mutant alleles can promote either suppression or enhancement of variegation at different loci on the same or different chromosomes (Table 1). The observation that most mutations affecting condensin proteins resulted in strong enhancement of variegation at the wm4h locus was initially surprising, as one might expect that proteins that compact DNA would encourage expansion of heterochromatin, resulting in suppression of variegation as previously reported (BHAT et al. 1996; DEJ et al. 2004). Nevertheless, one should also bear in mind the possibility that different condensin proteins may not function in the same way or as members of the same complex during interphase, consistent with the contrasting requirements for different condensin proteins during interphase in S. pombe (AONO et al. 2002) or during silencing at mating-type loci in S. cerevisiae (BHALLA et al. 2002). Indeed, different proteins are already known to be associated with the SMC2/4 heterodimer in C. elegans that participates in dosage compensation (CHUANG et al. 1994; LIEB et al. 1996, 1998). We recently verified the existence of a canonical condensin I complex in Drosophila embryos (SAVVIDOU et al. 2005) but it is still unknown what components an interphase condensin complex might contain in flies. Although the various condensin proteins are enriched in actively dividing tissues such as larval brains (STEFFENSEN et al. 2001; SAVVIDOU et al. 2005), we were unable to detect these proteins in equivalent amounts of extracts prepared from nondividing tissues such as third instar larval salivary glands (supplemental Figure 4 at http://www.genetics.org/supplemental/). However, it is possible that the interphase activity of condensin proteins may be specific to particular tissues or developmental stages and may require comparatively low protein levels to exert significant epigenetic effects.
|
The apparent specificity of condensin-mediated epigenetic phenomena to the fourth chromosome is intriguing. The fourth chromosome of Drosophila is unlike other autosomes but similar to the X or Y chromosomes insofar as flies haploid for these chromosomes are viable, albeit sterile in the case of XO males and usually sterile in the case of haplo-4 flies (LINDSLEY and GRELL 1968). The fourth chromosome is also unusual among the autosomes as it is predominantly heterochromatic, although it also contains interspersed euchromatic and heterochromatic domains (SUN et al. 2000, 2004), whereas domains of constitutive heterochromatin are normally limited to pericentric and telomeric DNA in most metazoans. Similarly, the Y chromosome contains over three times more pericentric heterochromatin than either the second or the third chromosome in Drosophila (PEACOCK et al. 1978; ADAMS et al. 2000). The behavior of telomeric gene silencing on the fourth chromosome differs from that of other autosomes and appears to more closely resemble the general pattern of pericentric gene silencing dependent on HP1 (CRYDERMAN et al. 1999; SUN et al. 2004). Other similarities between the fourth chromosome and the X chromosome seem to be revealed by the chromosomal specificity of painting of fourth (POF) binding in different Drosophila species. For example, the POF protein binds only the fourth chromosome in D. melanogaster males but decorates the entire X chromosome exclusively in D. busckii males (LARSSON et al. 2004). Moreover, POF binding to the male X chromosome (as well as to the fourth chromosome) in species such as D. ananassae and D. malerkotliana has been shown to colocalize with MSL3, suggesting a relationship between sex-linked dosage compensation and a protein associated with the fourth chromosome (LARSSON et al. 2004). Finally, it appears that the fourth chromosome may pair with the X in meiosis as persistent associations between pericentric heterochromatin of the X chromosome and the fourth chromosome have been described in prophase I oocyte nuclei (DERNBURG et al. 1996), while triplo-4 causes an increased frequency of X chromosome nondisjunction (SANDLER and NOVITSKI 1956). Similarly, duplications of the X chromosome can interfere with segregation of either the fourth chromosome or the nonexchange X chromosomes (HAWLEY et al. 1992). On the basis of these various similarities, we envisage that the apparent chromosomal specificity of interphase condensin activities might reflect recognition of a particular chromatin structure or sequence elements common to heterochromatin of the fourth chromosome and rDNA-proximal loci on the sex chromosomes. Although such specific chromosomal targets remain to be identified, it is possible that these might include the Hoppel element/1360 transposon, as this appears to be associated with heterochromatin formation on the fourth chromosome (CRYDERMAN et al. 1999; SUN et al. 2004) and also appears to be enriched in X chromosomal pericentric heterochromatin (BARTOLOMÉ et al. 2002).
Strikingly, we have observed both strong suppression and enhancement of variegation at the white gene in a wm4h background in the case of the glu88-82 allele. By contrast, the other gluon mutant alleles and dcap-g64 behaved only as strong enhancers of variegation at this locus, even though each of these other mutations yielded similar mitotic phenotypes to glu88-82. If the glu88-82 allele proves to be a neomorph as suggested by sequence data, then the suppression of variegation caused by this allele might be explained in terms of the activity of a C-terminal truncated protein. Nevertheless, the effect of glu88-82 on PEV at other loci shows enhancer or suppressor behavior similar to that of other gluon alleles. We therefore postulate that condensins might affect gene silencing by regulating the spread of heterochromatin, possibly by exerting effects on chromatin boundary elements or insulators (LUPO et al. 2001; CUVIER et al. 2002). In the absence of full condensin function, gene expression might be either enhanced or repressed as developmentally regulated patterns of gene expression become locked (PIRROTTA 1997; FRANCIS and KINGSTON 2001; RINGROSE and PARO 2004). It is also possible that the contrasting effects of the glu88-82 allele on PEV at a given locus might reflect the relative ability of a truncated SMC4 protein to interact with distinct subunits of the condensin I or condensin II complex, as non-SMC subunits of these different complexes in vertebrates display distinct localization patterns and appear to contribute differently to mitotic chromosome architecture (ONO et al. 2003, 2004). Similarly, differential behavior of distinct SMC4-containing condensin complexes at different loci might explain the contrasting effects of both glu88-82 and other gluon alleles at different loci on the Y chromosome. However, relatively little is currently known about the behavior of the condensin II complex in Drosophila, aside from a possible role for the CAP-D3 subunit in male meiosis (SAVVIDOU et al. 2005). Consequently, further analysis of the interplay between different condensin proteins and the transcriptional machinery will be required to elucidate mechanistic details of their potential involvement in epigenetic phenomena. Intriguingly, condensin mutants in S. cerevisiae were shown to relocalize telomeric Sir2p to the centromere proximal rDNA and displayed histone hyperacetylation at telomeres (MACHÍN et al. 2004), while deletion of SIR2 increased instability of rDNA by impairing association of the Scc1 cohesin subunit (KOBAYASHI et al. 2004). As the sir2 gene appears to be similarly nonessential for viability in Drosophila yet has minor effects on position-effect variegation (ÅSTRÖM et al. 2003), it is tempting to speculate that the effects of condensins on gene expression might be also mediated at least in part by SIR2 in this species.
In conclusion, it remains to be seen exactly how condensins might behave differently to facilitate interconversion between different states of chromatin compaction in interphase and to compact chromatin in mitosis. The current observations, however, can be unified if one postulates that condensin organizes a particular type of chromatin, which is then acted upon by other components to either bring about chromosome condensation in mitosis or regulate gene expression at particular loci. It is therefore clear that there is still much to discover about how these proteins might operate at various stages of the cell cycle and in different developmental contexts.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODS |
|---|
= 17.5) and 3, respectively, in control progeny but the number of pigmented ommatidia was consistently zero for each of the gluon mutant heterozygotes.
