Barren, the Drosophila homolog of XCAP-H, is one of three non-SMC subunits of condensin, a conserved 13S multiprotein complex required for chromosome condensation. Mutations in barren (barr) were originally shown to affect sister-chromatid separation during mitosis 16 of the Drosophila embryo, whereas condensation defects were not detected. In contrast, mutations in yeast homologs of barren result in defective mitotic chromosome condensation as well as irregular chromatid separation. We have used double-stranded RNA-mediated interference (RNAi) to deplete Barren in Drosophila S2 cells. Our analyses indicate that inactivation of barr leads to extensive chromosome condensation and disrupts chromatid segregation.
CHROMOSOME condensation in eukaryotic cells requires condensin, a 13S multiprotein complex that is conserved in all organisms studied to date (Hiranoet al. 1997; Sutaniet al. 1999; Freemanet al. 2000; Kimuraet al. 2001). This complex, identified initially in Xenopus egg extracts, consists of five subunits termed XCAPs (Xenopus chromosome-associated proteins): two of these belong to the structural maintenance of chromosome (SMC) family and contain coiled-coils with globular terminal domains that interact with ATP and DNA (Kimuraet al. 1999); the remaining three are defined as non-SMC proteins and play essential, although not well understood, roles in condensation (Kimuraet al. 2001). Mitotic kinases control both localization of the condensin complex to chromosomes and its supercoiling and condensation activities (Kimura et al. 1998, 2001; Sutaniet al. 1999; Giet and Glover 2001).
Two sets of evidence indicate 13S condensin is required for chromosome condensation. Immunodepletion of affinity-purified condensin fractions in Xenopus egg extracts converted compact sperm chromatin into a round-shaped interphase-like chromatin sphere, whereas adding condensin back restored the activity of chromosome assembly (Hiranoet al. 1997). In yeast, mutations in the condensin subunits resulted in increased distance between two labeled probes on a mitotic chromosome and disorganization of rDNA array (Strunnikovet al. 1995; Ouspenskiet al. 2000; Lavoieet al. 2002). This approach, coupled with biochemical analyses, has been exploited to define roles played by the SMC subunits in chromosome condensation and segregation, including the yeast homologs of the SMC-type subunit XCAP-C (Guacciet al. 1993; Sakaet al. 1994; Strunnikovet al. 1995). More recently, mutants in the Drosophila SMC4 ortholog have been shown to be defective in proper chromosome condensation and segregation throughout different developmental stages (Steffensenet al. 2001).
Studies of the non-SMC subunits of condensin have been carried out mainly on Barren, the homolog of the XCAP-H subunit. barren (barr) was originally isolated as a Drosophila mutant affecting embryonic neuronal cell divisions (Bhatet al. 1996). The analysis of dividing cells in mutant barr embryos revealed that chromosomes fail to separate their chromatids during anaphase, giving rise to extensive chromatin bridging (Bhatet al. 1996). This phenotype is similar to that observed in DNA topoisomerase II mutants in Saccharomyces cerevisiae (Holmet al. 1989). barr chromosomes, however, appear to condense regularly and the missegregation phenotype does not appear to be a consequence of defects in chromosome condensation. Thus, the involvement of barr in chromosome condensation has been inferred solely from the observation that Barren and topoisomerase II coimmunoprecipitate and interact in vitro and in vivo (Bhatet al. 1996). In addition, mutants in brn1 and cnd2 genes, which encode the Barren homolog in budding and fission yeast, respectively, exhibit chromosome condensation defects (Aonoet al. 2002; Lavoieet al. 2002). This has led us to ask why Drosophila barr mutants do not exhibit the same class of condensation defects revealed by their yeast counterparts.
We have used double-stranded RNA-mediated interference (RNAi) to deplete barr function in Drosophila S2 cultured cells. The addition of 15 μg of barr double-stranded RNA (dsRNA) into S2 culture cells completely ablated the endogenous mRNA after 72 hr as shown by RT-PCR (Figure 1A). Cytological analysis of treated metaphases revealed that chromosome condensation was highly defective upon depletion of the barr gene product (Figure 2). In 98% of metaphases scored (n = 75), condensing chromatin appeared fuzzy and loose. Sister-chromatid morphology was largely disorganized and chromosomes were not distinguishable from one another. This chromatin undercondensation pattern was also detectable during prophase and by using different DNA dyes (i.e., Giemsa, Figure 2F). To assess whether heterochromatin was also influenced by this global decondensation effect, we performed fluorescence in situ hybridization (FISH) on metaphase spreads using the dodecasatellite probe. The dodecasatellite probe specifically hybridizes to pericentromeric heterochromatin of chromosome 2 (Carmenaet al. 1993). In all barr metaphases analyzed (n = 20) the fluorescent signals appeared larger and more diffuse with respect to untreated control metaphases (Figure 3). We measured the area of each fluorescent dot and found that it was, on average, 1.8 times larger in barr cells than in controls. This suggests that some degree of chromatin decondensation may also occur at pericentromeric regions. This observation is consistent with Barren localization in the centromeric region (Steffensenet al. 2001) where it is thought to be required for the proper function of centromeres. In addition, studies in yeast have shown that brn-1 may be necessary for the formation of functional mitotic kinetochores (Ouspenskiet al. 2000).
To determine whether the phenotype associated with RNAi of barr reflected a general chromatin undercondensation phenomenon or whether it could be attributed directly to depletion of barr, we treated S2 cells with gluon dsRNA. gluon encodes the Drosophila SMC4 homolog and has been shown to be required for chromosome condensation and sister-chromatid resolution (Steffensenet al. 2001). We found that the endogenous gluon mRNA was completely depleted after 72 hr treatment (Figure 1A) and that its depletion resulted in disruption of chromatin condensation (Figure 1B). Interestingly, the chromatin decondensation pattern in gluon dsRNA-treated cells is distinct from that observed in barr (Figure 2, B-D), but very similar to that described for gluon mutant neuroblast chromosomes (Steffensenet al. 2001). Chromosomes in glu (RNAi) cells appear swollen and the chromatin downy. In contrast, barr chromatin seems unrolled and diffuse and appears to branch off the main chromosome axis, clearly visible in the Giemsa-stained spreads. Moreover, using version 1.6 of the National Institutes of Health (NIH) image tool for MacIntosh (http://rsb.info.nih.gov/nih-image), we measured both chromosome length and width in untreated and treated cells to ask whether axial condensation or completion of the chromatin loops, or both, was affected. For this purpose we focused our analysis on metacentric chromosomes, where both longitudinal and transversal axes were identifiable despite the irregular condensation phenotype. We observed that in barr cells the average chromosome length and width did not significantly differ from that of the control (35.40 vs. 37.61 pixels and 14.98 vs. 12.56 pixels, respectively), whereas gluon chromosomes were longer (51.89 pixels) and wider (18.75 pixels) than the control. Collectively, these data suggest that inactivation of these two different subunits (one SMC and one non-SMC) of the condensin complex has diverse effects on chromosome condensation and chromatin loop organization. Furthermore, these observations support the view that components of the condensin complex have distinct, specialized functions (Aonoet al. 2002).
Chromatin from cells treated with barr dsRNA appeared decondensed throughout anaphase and telophase. An extremely high proportion of these figures exhibited chromatin bridges and laggards (98%, n = 55, Figure 4, F-H), very likely a consequence of disorganized chromatin fibers. We speculate that the extensive chromosome decondensation may give rise to chromosomes disentangling and failure of chromatid resolution.
We next asked whether these chromatid segregation abnormalities were accompanied by spindle defects. We immunostained treated cells with the anti-α-tubulin antibody and observed that spindle components were not affected by depletion of barr and appeared normal during all stages of mitosis (Figure 4). Moreover, mitotic progression of treated S2 cells was not influenced, as the percentage of cells at different stages did not differ significantly from the control (Figure 1C). Furthermore, cytokinesis progressed regularly, as no binucleated cells were observed, indicating that, despite massive anaphase bridging, cell membranes were successfully pinched toward the chromatin bridge. This was also confirmed by fluorescence-activated cell sorter (FACS) profiling of treated cells; no increase of the 8C peak at the expense of the 2C and 4C peaks was observed, as normally occurs when a high proportion of cells become polyploid (Sommaet al. 2002). The FACS analysis, however, revealed a general reduction of the 2C peak and an increase of the area between the 2C and 4C peaks (Figure 1B). We believe that the expanded area between 2C and 4C peaks represents 2C cells that, despite chromosome condensation defects, complete mitosis and enter the subsequent cell cycle with an aberrant DNA content.
To further characterize the phenotype of barr mutant cells, we analyzed whether the localization of DNA topoisomerase II, which has previously been shown to interact with Barr (Bhatet al. 1996), was affected in Barr-depleted cells. By immunostaining both mutant and control cells with an antitopoisomerase II antibody, we observed the localization of this enzyme throughout the different stages of mitosis. We found that topoisomerase II was associated with chromatin during prophase, metaphase, anaphase, and telophase in both barr and control cells (Figure 4), suggesting that its localization is independent of barr. Moreover, this analysis indicates that chromosome decondensation is a direct effect of Barr depletion and not a consequence of DNA topoisomerase II misbehavior.
To date, we do not understand why defects in chromosome condensation are present in Barr-depleted S2 cells and not in mutant embryos. Chromosome compaction was found to be normal in a null mutation of barren, so this discrepancy cannot be explained by allele-specific defects. It is possible that barr mutant embryos still retain maternal Barr product, which enables chromosomes to condense properly. Alternatively, it is conceivable that the organization of chromatin in embryonic nuclei differs from that of S2 cells, as the former support frequent and rapid mitotic divisions. In this highly dynamic scenario, Barren may not be strictly required for chromatin compaction, or its function may be redundant. One can also argue that the chromatin disorganization observed in S2 cells is a peculiarity of this cell type. We believe this is unlikely, however, as we show that RNAi of gluon causes failure in chromosome condensation, a phenotype similar to that elicited by gluon mutant neuroblast cells, thus indicating that S2-treated cells do mimic Drosophila mutations. In addition, the observation that dsRNAs of different condensin subunits have different effects on chromosome condensation makes RNA interference a useful tool to molecularly dissect this important aspect of chromosome dynamics.
We thank E. Cundari for FACS analysis and Donna Arndt-Jovin for the antitopoisomerase II antibody. We are also grateful to L. Ciapponi and M. Gatti for critical reading of the manuscript. This work has been supported by a grant from the Consiglio Nazionale delle Ricerche.
Communicating editor: K. Golic
- Received February 21, 2003.
- Accepted June 17, 2003.
- Copyright © 2003 by the Genetics Society of America