Positioning a euchromatic gene near heterochromatin can influence its expression. To better understand expression-relevant changes in locus positioning, we monitored in vivo movement of centromeres and a euchromatic locus (with and without a nearby insertion of heterochromatin) in developing Drosophila tissue. In most undifferentiated nuclei, the rate of diffusion and step size of the locus is unaffected by the heterochromatic insertion. Interestingly, although the movement observed here is non directional, the heterochromatic insertion allows the flanking euchromatic region to enter and move within the heterochromatic compartment. This study also finds that a constraint on chromatin movement is imposed which is a factor of distance from the centric heterochromatic compartment. This restraint prevents the heterochromatic locus from moving away from the centric heterochromatin compartment. Therefore, because of the constraint, even distinct and non-random nuclear organizations can be attained from random chromatin movements. We also find a general constraint on chromatin movement is imposed during differentiation, which stabilizes changes in nuclear organization in differentiated nuclei.

Introduction

Changes in nuclear organization influence nuclear function. In higher eukaryotes, the positioning of some loci near large blocks of heterochromatin is correlated with gene silencing. Research is now focused on elucidating the mechanisms that help establish and maintain the organization of the interphase nucleus. Live imaging studies following specific loci marked with fluorescence tags have found that most chromatin movement is stochastic but constrained (for reviews, see Misteli, 2005; Spector, 2003). A conundrum in these studies is how a process that is fundamentally random can have a specific, predictable order as its endpoint.

Although the most recent studies leading to an appreciation of the impact of gene positioning on gene expression have used yeast and mammalian cells, a correlation was first established using the bwDominant (bwD) allele of fruit flies. The bwD allele contains a 1.6 Mb insertion of heterochromatin into the distal euchromatin of the right arm of chromosome 2 (2R). Flies heterozygous for bwD exhibit a variegated eye phenotype, demonstrating that the wild-type gene on the homologous chromosome is also inactivated. Fluorescence in situ hybridization (FISH) on cells dissected from third-instar larvae heterozygous for bwD showed that the bw locus associates with centric heterochromatin of the second chromosome (2h). Somatic pairing of the homologous interphase chromosomes (a phenomenon commonly found in almost all nuclei of dipteran insects) results in the bwD allele dragging the wild-type homolog closer to heterochromatin, resulting in silencing of the wild-type gene (Csink and Henikoff, 1996; Dernburg et al., 1996; Harmon and Sedat, 2005).

Studies presented here find that the movement of the bwD locus is random and directionless. Surprisingly, in most bwD nuclei movement of the euchromatic region near the heterochromatic insert was unaffected by the insertion, suggesting that the 1.6 Mb of heterochromatin does not affect the dynamics of locus movement. Further analysis of our data shows that when the stochastic movements of bwD bring it close to the centromere, and its surrounding heterochromatin, a constraint in the movement of the bw region is imposed. This constraint results in an extreme decrease in the step size and radius of confinement of the locus. Interestingly, absence of the heterochromatic insertion in bw+ appears to exclude the wild-type locus from the centric heterochromatic compartment. In addition, we find that the general constraint on chromatin movement that is imposed during differentiation (Thakar and Csink, 2005) also prevents further changes in organization despite an increase in time spent in G0. Our results suggest that association is a stable encounter between the two loci that is maintained in differentiated cells by increasing constraint on chromatin dynamics and mechanisms that prevent the escape of bwD from the heterochromatic compartment.

Results

Development and verification of tools to examine heterochromatic associations in vivo

To track the movement of bw we tagged a flanking locus with LacOperator repeats (lacO) using a P-element transposon (Vazquez et al., 2001). The repeats are inserted 15 kb away in the Dcp1 locus on a wild-type chromosome and as well as in the exact same location and orientation on a chromosome carrying the bwD allele. The transposon reporter in the insertion on the wild-type chromosome can be silenced by a bwD allele on the homolog, showing that this insertion site is a good proxy for the bwD heterochromatic insert. These repeats are visualized by crossing in a second transposon expressing Lac repressor (LacI) fused to a monomeric red fluorescent protein (mRFP) (Campbell et al., 2002; Robinett et al., 1996; Thakar and Csink, 2005). The centromeres were marked using centromere identifier (CID) fused to GFP (CIDGFP) (Henikoff et al., 2000) (Fig. 1A).

Fig. 1.

Distance measurements of bw-centromere signals to determine whether the CIDGFP dot closest to the mRFP dot is the centromere of the second chromosome. (A) Projected image of undifferentiated nuclei anterior to the morphogenetic furrow in the eye imaginal discs. The green dots are CIDGFP and the single purple dot represents the lacO repeats inserted in the bw region bound by mRFP-LacI. The background fluorescence of the unbound mRFP-LacI marks the nucleus. In interphase nuclei, the centromeres of the four paired chromosomes in Drosophila are typically observed as three or four dots. (B) Distance between the tagged locus and the closest CIDGFP dot was computed and divided by the radius of the nucleus. The distribution of distances is displayed as box plots. Box plots are calibrated representations of histograms wherein each horizontal line delimits the 10th, 25th, 50th (median), 75th and 90th percentiles. The numbers within the box plots are the number of nuclei included in the analysis. P-values (Mann-Whitney U-test) are shown above brackets for the respective sets. (C) The distribution is similar to that observed in an earlier study using FISH. Probes specific to the bw locus and the AACAC satellite repeats that make up a small subset of 2Rh were used (Thakar and Csink, 2005). (D) A histogram of mRFP-CIDGFP distances corrected by the radius.

Fig. 1.

Distance measurements of bw-centromere signals to determine whether the CIDGFP dot closest to the mRFP dot is the centromere of the second chromosome. (A) Projected image of undifferentiated nuclei anterior to the morphogenetic furrow in the eye imaginal discs. The green dots are CIDGFP and the single purple dot represents the lacO repeats inserted in the bw region bound by mRFP-LacI. The background fluorescence of the unbound mRFP-LacI marks the nucleus. In interphase nuclei, the centromeres of the four paired chromosomes in Drosophila are typically observed as three or four dots. (B) Distance between the tagged locus and the closest CIDGFP dot was computed and divided by the radius of the nucleus. The distribution of distances is displayed as box plots. Box plots are calibrated representations of histograms wherein each horizontal line delimits the 10th, 25th, 50th (median), 75th and 90th percentiles. The numbers within the box plots are the number of nuclei included in the analysis. P-values (Mann-Whitney U-test) are shown above brackets for the respective sets. (C) The distribution is similar to that observed in an earlier study using FISH. Probes specific to the bw locus and the AACAC satellite repeats that make up a small subset of 2Rh were used (Thakar and Csink, 2005). (D) A histogram of mRFP-CIDGFP distances corrected by the radius.

FISH on fixed diploid nuclei from third-instar larvae has shown that the bwD locus can associate with the centric heterochromatin of the second chromosome (2h) (Csink and Henikoff, 1996; Dernburg et al., 1996). Given the arrangement of chromosomes into territories (Zink et al., 1998), we conjectured that the CIDGFP signal closest to the mRFP signal is the centromere of the second chromosome, and would serve as an in vivo marker for 2h. To test this assumption eye imaginal discs were dissected, primary cultures of the tissue were maintained during microscopy and the distance between lacO repeats near bw and the closest CIDGFP signal was measured (Fig. 1B,D).

We found that the distance between the mRFP signal and the closest CID dot was statistically significantly lower in bwD lacO / bw+ nuclei as compared with bw+ lacO / bw+. The distribution of data was remarkably similar to that obtained in an earlier study that used FISH on eye imaginal discs (Thakar and Csink, 2005), where the probes used were specific to the bw region and the AACAC satellite repeats, which are unique to the heterochromatin of 2R (Fig. 1C). These AACAC repeats, like the centromeres, do not encompass the entirety of 2h, but make up only a small part. The similarity in the distribution of distances between mRFP-CIDGFP to that obtained in the FISH experiments for bw-2Rh confirmed that the closest CID dot is usually the centromere of the second chromosome and validates the use of this system to study nuclear organization and chromatin dynamics of the bwD locus in vivo.

Heterochromatin associations in differentiating embryonic nuclei

Earlier research had suggested that in larval central nervous system (CNS) cells bwD heterochromatic associations do not occur until at least 5 hours into G1 (Csink and Henikoff, 1998) and the prevalence of these associations increases with the amount of time since the last mitosis (Thakar and Csink, 2005). Association between bwD-2Rh is not established in the early embryo (Dernburg et al., 1996). To determine whether changes in nuclear organization leading to bwD-2Rh association were possible in any embryonic cells, we examined the nuclei of embryonic neuroblasts, which are capable of differentiation in culture (Furst and Mahowald, 1985). Embryos were collected, aged and disrupted, and the adhered neuroblasts were allowed to develop for 5 or 15 hours before imaging. Under these culture conditions ganglionic clusters as well as neural outgrowths that stained positive for embryonic lethal abnormal visual system (ELAV) are observed only after 15 hours of culturing (Fig. 2A). The fluorescently tagged loci were also easily visible in the cultured neuroblasts (Fig. 2B).

We compared the distance between the mRFP in the bw region and the closest CIDGFP dot in bwD and wild-type nuclei. Neuroblasts cultured for 5 hours do not show association between mRFP and CID in bwD nuclei (Fig. 2C). Interestingly, we found that when allowed to culture for 15 hours, by which time the neuroblasts typically exhibit the morphology of differentiated neurons, large-scale nuclear reorganization of the bwD locus had occurred (Fig. 2D). This demonstrates that the tip of 2R can move across the space of the interphase nucleus to associate with centric heterochromatin even in embryonic cells.

Fig. 2.

bwD association is observed in differentiated embryonic neurons but not undifferentiated neuroblasts. (A) Immunofluorescence with antibody against ELAV, a neuron-specific protein, on embryonic neuroblasts cultured for either 5 hours (left) or 15 hours (right). Although the 5-hour cells do not show much signal for ELAV, those cultured for 15 hours do. The 15-hour cells also tend to form ganglionic clusters and extrude neuronal processes. (B) Single section from a series of z-stack images of embryonic neuroblasts expressing CIDGFP (green) and mRFP-LacI protein bound to the bw-region tagged with lacO repeats (purple). (C) Box plots displaying the 2D distance between the mRFP dot and the closest CIDGFP dot corrected by the nuclear radius from projected images. At least three to four coverslips were imaged with no more than 40 nuclei imaged per coverslip. Change in nuclear organization of the bwD locus is observed with a significant number of cells showing association between the bwD locus and centromere only after 15 hours of culturing. (D) Histograms of data shown in C.

Fig. 2.

bwD association is observed in differentiated embryonic neurons but not undifferentiated neuroblasts. (A) Immunofluorescence with antibody against ELAV, a neuron-specific protein, on embryonic neuroblasts cultured for either 5 hours (left) or 15 hours (right). Although the 5-hour cells do not show much signal for ELAV, those cultured for 15 hours do. The 15-hour cells also tend to form ganglionic clusters and extrude neuronal processes. (B) Single section from a series of z-stack images of embryonic neuroblasts expressing CIDGFP (green) and mRFP-LacI protein bound to the bw-region tagged with lacO repeats (purple). (C) Box plots displaying the 2D distance between the mRFP dot and the closest CIDGFP dot corrected by the nuclear radius from projected images. At least three to four coverslips were imaged with no more than 40 nuclei imaged per coverslip. Change in nuclear organization of the bwD locus is observed with a significant number of cells showing association between the bwD locus and centromere only after 15 hours of culturing. (D) Histograms of data shown in C.

Constrained movement in differentiated cells prevents further nuclear reorganization

There is a marked increase in the constraint on chromatin dynamics upon differentiation (Thakar and Csink, 2005). Here we test whether the constraint in chromatin movement could prevent further reorganization, despite an increase in time since the last mitosis. In the eye imaginal disc of late third-instar larvae the differentiating cells are not only spatially separated from undifferentiated cells, but they are also separated based on a gradient of differentiation in that cells closer to the furrow have more recently undergone differentiation (Fig. 3A). Images of differentiated cells present posterior to the morphogenetic furrow were separated into two categories based on their distance from the furrow (Fig. 3B). The percentage of cells showing association does not change with the gradient of differentiation (Fig. 3C). In light of the constraint on chromatin movement in differentiated nuclei, these results suggest that the confined chromatin movement that accompanies differentiation leads to a lockdown of nuclear organization. Hence, a further increase in association is impossible despite an increase in time spent in the G0 phase.

Fig. 3.

Constraint of chromatin dynamics with differentiation restricts changes in nuclear organization. (A) An eye imaginal disc expressing GFP (pink) in differentiating cells present posterior to the morphogenetic furrow. There is a gradient of differentiation, with cells further away from the furrow having started differentiating earlier than those present closer to the furrow. A series of images of the posterior end of the disc were taken (green box). (B) A projected image representing the area demarcated by the green box is shown. Each image was divided into two halves, categorizing the nuclei into early differentiating and late-differentiating nuclei. (C) Box plots of distances between the mRFP signal from the tagged bwD locus and the closest CIDGFP dot was measured from projected images and corrected by the nuclear radius. No significant difference in the distribution was observed in the late-differentiating cells as compared with the early differentiating cells.

Fig. 3.

Constraint of chromatin dynamics with differentiation restricts changes in nuclear organization. (A) An eye imaginal disc expressing GFP (pink) in differentiating cells present posterior to the morphogenetic furrow. There is a gradient of differentiation, with cells further away from the furrow having started differentiating earlier than those present closer to the furrow. A series of images of the posterior end of the disc were taken (green box). (B) A projected image representing the area demarcated by the green box is shown. Each image was divided into two halves, categorizing the nuclei into early differentiating and late-differentiating nuclei. (C) Box plots of distances between the mRFP signal from the tagged bwD locus and the closest CIDGFP dot was measured from projected images and corrected by the nuclear radius. No significant difference in the distribution was observed in the late-differentiating cells as compared with the early differentiating cells.

Loss of Rabl orientation of chromosomes in nuclei of eye imaginal discs

One of the earliest changes in nuclear organization during interphase is often the breakdown of Rabl orientation, which is the clustering of centromeres at one pole of the nucleus and telomeres at the opposite pole. This arrangement is established during anaphase when the homologs are pulled apart by the spindle apparatus that is attached to the centromere, while the telomeres drag behind (for a review, see Comings, 1980). We were interested in finding out whether Rabl orientation persists in the eye imaginal disc and whether this rearrangement is also influenced by differentiation. Chromosomes were categorized to be in Rabl orientation if the centromeres (CIDGFP signals) were clustered at one pole and lacO repeats near bw (near the end of 2R) at the opposite pole (supplementary material Fig. S1). Similar to the observation in larval CNS (Csink and Henikoff, 1998), a breakdown of Rabl orientation occurs in cells of the eye imaginal disc, this breakdown is more prominent in differentiated than undifferentiated cells (P<0.001 as computed by G-test) (Sokal and Rohlf, 1981) (Table 1).

Table 1.

Percentage of nuclei in the eye imaginal disc in Rabl orientation

Treatment No. of discs Genotype No. of nuclei No. of nuclei in Rabl Rabl orientation (%) Disc end
Only media   3  bw+  175   13   7.42   Anterior  
Only media   3  bw+  966   23   2.38   Posterior  
Only media   3  bwD  274   14   5.10   Anterior  
Only media   3  bwD  894   14   1.56   Posterior  
Microtubule destabilizers   5  bwD  161   13   8.07   Anterior  
Treatment No. of discs Genotype No. of nuclei No. of nuclei in Rabl Rabl orientation (%) Disc end
Only media   3  bw+  175   13   7.42   Anterior  
Only media   3  bw+  966   23   2.38   Posterior  
Only media   3  bwD  274   14   5.10   Anterior  
Only media   3  bwD  894   14   1.56   Posterior  
Microtubule destabilizers   5  bwD  161   13   8.07   Anterior  

The diffusion coefficient and step size of the bwD locus decreases only when it is associated with centric heterochromatin

In other studies, movement of chromatin has been found to be non directional and random (Heun et al., 2001; Marshall et al., 1997; Vazquez et al., 2001). How can such a process have a highly stable chromosomal arrangement as its endpoint? To address this question, we compared the movement of bwD, which has a specific nuclear location, to the locus without the heterochromatic insertion.

Movement of the bw region was followed relative to the closest centromere in undifferentiated cells of the eye imaginal discs (supplementary material Movie 1). To determine the characteristics of chromatin movement, the mean square change in distance (MSD, or <Δd2>), which is the average of Δd2 values over all possible combinations of time points separated by Δt, was computed (Fig. 4A). The slope of the graph is related to the diffusion coefficient based on the relation, <Δd2>=4D <Δt> (Qian et al., 1991). The diffusion coefficient of bwD was comparable to that obtained for the bw+ locus (Table 2). Our results reveal that despite the insertion of approximately 1.6 Mb of heterochromatic repeats into the coding region of bw, the dynamics of chromatin movement of the flanking euchromatin is mostly unchanged.

Table 2.

Diffusion coefficients

Genotype treatment Locus tagged No. of larvae No. of nuclei Diffusion coefficient (μm2/second)
lacO bw+/bw+ bw+-CID   3   40   5.0×10-4 
lacO bwD/bw+ bwD-CID   11   134   4.2×10-4 
lacO bwD/bw+ IU bwD-CID   11   101   5.0×10-4 
lacO bwD/bw+ IA bwD-CID   11   33   8.3×10-5 
lacO bw+/bw+  CID-CID   3   29   8.3×10-5 
lacO bwD/bw+ IU Untreated control for microtubule destabilizers  bwD-CID   9   64   4.2×10-4 
lacO bwD/bw+ IA Untreated control for microtubule destabilizers  bwD-CID   9   32   1.7×10-4 
lacO bwD/bw+ IU Microtubule destabilizers  bwD-CID   9   78   2.5×10-4 
lacO bwD/bw+ IA Microtubule destabilizers  bwD-CID   9   36   1.7×10-4 
lacO bwD/bw Untreated control for microtubule destabilizers   CID-CID   8   88   1.7×10-4 
lacO bwD/bw Microtubule destabilizers   CID-CID   10   109   8.3×10-5 
Genotype treatment Locus tagged No. of larvae No. of nuclei Diffusion coefficient (μm2/second)
lacO bw+/bw+ bw+-CID   3   40   5.0×10-4 
lacO bwD/bw+ bwD-CID   11   134   4.2×10-4 
lacO bwD/bw+ IU bwD-CID   11   101   5.0×10-4 
lacO bwD/bw+ IA bwD-CID   11   33   8.3×10-5 
lacO bw+/bw+  CID-CID   3   29   8.3×10-5 
lacO bwD/bw+ IU Untreated control for microtubule destabilizers  bwD-CID   9   64   4.2×10-4 
lacO bwD/bw+ IA Untreated control for microtubule destabilizers  bwD-CID   9   32   1.7×10-4 
lacO bwD/bw+ IU Microtubule destabilizers  bwD-CID   9   78   2.5×10-4 
lacO bwD/bw+ IA Microtubule destabilizers  bwD-CID   9   36   1.7×10-4 
lacO bwD/bw Untreated control for microtubule destabilizers   CID-CID   8   88   1.7×10-4 
lacO bwD/bw Microtubule destabilizers   CID-CID   10   109   8.3×10-5 

On further examination of these data, we noticed that at the first time point no bw+ nuclei were observed where the distance between bw-CID was less than 1 μm (Fig. 4B). Therefore, we divided our bwD data into two categories based on the relative distance of the bw region and the closest centromere at t=0. Nuclei with initial distance >1.0 μm were categorized as initially unassociated (IU) and nuclei with distances <1.0 μm were categorized as initially associated (IA). We found that for the IA bwD nuclei the diffusion coefficient was much lower than that observed for the IU nuclei, however in the IU bwD loci the chromatin dynamics were similar to bw+ (Fig. 4C; Table 2).

The step size is the change in distance over a specified interval of time. In the yeast nucleus, that has a radius of 1-2 μm, loci tagged with lacO repeats were found to occasionally undergo movement with a step size >0.5 μm. Such large-scale movements have been hypothesized to facilitate changes in nuclear organization (Heun et al., 2001). Similar movement of the bwD locus may occur that allows the locus to come into close proximity to the centric heterochromatic compartment. As observed for the diffusion coefficient, the step size of the IU bwD locus was similar to that observed for the tagged bw+ locus (Fig. 4D). Interestingly, we found that the value of |Δd| in IA bwD nuclei was significantly less (P<0.0001, t-test), when compared with IU nuclei. We also found the absence of |Δd|>0.5 μm in nuclei categorized as IA. The maximum step size in IA nuclei was found to be 0.3 μm as compared with 0.7 μm in IU nuclei (Table 3).

Table 3.

Maximum and mean step size (Δt=15 seconds)

Genotype treatment Locus tagged Number of time points Max. step size |Δd| (μm) Mean step size |Δd| (μm)
lacO bw+/bw+ bw+-CID   381   0.58   0.12  
lacO bwD/bw+ IU bwD-CID   1005   0.71   0.10  
lacO bwD/bw+ IA bwD-CID   327   0.33   0.06  
lacO bw+/bw+  CID-CID   203   0.24   0.06  
lacO bwD/bw+ IU Untreated control for microtubule destabilizers  bwD-CID   448   0.44   0.11  
lacO bwD/bw+ IA Untreated control for microtubule destabilizers  bwD-CID   224   0.39   0.08  
lacO bwD/bw+ IU Microtubule destabilizers  bwD-CID   546   0.70   0.10  
lacO bwD/bw+ IA Microtubule destabilizers  bwD-CID   252   0.37   0.08  
lacO bwD/bw+ Untreated control for microtubule destabilizers   CID-CID   616   0.37   0.079  
lacO bwD/bw+ Microtubule destabilizers   CID-CID   763   0.32   0.055  
Genotype treatment Locus tagged Number of time points Max. step size |Δd| (μm) Mean step size |Δd| (μm)
lacO bw+/bw+ bw+-CID   381   0.58   0.12  
lacO bwD/bw+ IU bwD-CID   1005   0.71   0.10  
lacO bwD/bw+ IA bwD-CID   327   0.33   0.06  
lacO bw+/bw+  CID-CID   203   0.24   0.06  
lacO bwD/bw+ IU Untreated control for microtubule destabilizers  bwD-CID   448   0.44   0.11  
lacO bwD/bw+ IA Untreated control for microtubule destabilizers  bwD-CID   224   0.39   0.08  
lacO bwD/bw+ IU Microtubule destabilizers  bwD-CID   546   0.70   0.10  
lacO bwD/bw+ IA Microtubule destabilizers  bwD-CID   252   0.37   0.08  
lacO bwD/bw+ Untreated control for microtubule destabilizers   CID-CID   616   0.37   0.079  
lacO bwD/bw+ Microtubule destabilizers   CID-CID   763   0.32   0.055  

bwD does not move towards the heterochromatic compartment in a directed manner

The signed step size measurements can be used to determine whether movement is directional. If the movement of the tagged locus was random then the values for step size [Δd=d(tn-1)-d(tn)] would follow a normal distribution, with an equal probability of positive or negative value for the change in step size. However, if the movement were directed towards the closest CID dot, with each time point the distance between bwD-CIDGFP would decrease and the distribution of Δd would shift towards the right. As observed from the histograms (Fig. 4D), the distribution of Δd appears to be normal. The one-sample sign test is a statistical test that determines whether there is a bias towards positive or negative values in a given distribution (Sokal and Rohlf, 1981). The P-value of a one-sample sign test on each of the distributions shown in Fig. 4D was not statistically significant.

Fig. 4.

Association of bwD with the centromere greatly decreases the movement and radius of confinement of the locus. Data obtained from short time-lapse movies was used to calculate the diffusion coefficient and step size and to test for directionality of movement in bwD and bw+ nuclei. (A) Movement of the bw locus was followed relative to the closest centromere marked by CIDGFP in undifferentiated cells of the eye imaginal discs. The average mean square change in distance <Δd2> over all possible Δt values (±1 s.e.m.) was plotted for each genotype. The 1.6 Mb insertion of heterochromatic repeats does not alter the dynamics of movement of the bw locus. The diffusion coefficients in Table 2 were calculated by taking the mean value for <Δd2> at Δt= 3 and 6 seconds from the data set shown here. (B) Histogram of the distribution of distances between mRFP-CIDGFP at t=0, from the data shown in A. Note that no bw+ nuclei were observed where the distance between mRFP-CIDGFP signal was less than 1.0 μm. (C) bwD nuclei were further categorized into IU (distance between bwD and the closest CIDGFP dot was >1.0 μm at t=0) and IA (distance between bwD and the closest CIDGFP dot was <1.0 μm at t=0). The graph shows the mean square change in distance of the IA and IU bwD nuclei. (D) Histograms of step size (Δd) at Δt=15 seconds. The mean step size of bwD (IA) was found to be statistically significantly lower with a P-value of <0.001 (t-test), than either bw+ or bwD (IU). (E) To calculate the true radius of confinement for bw+ and IU bwD, chromatin movement was followed for a longer time (1-2 hours) with images taken every 3 minutes to obtain graphs where the plots for the average mean square change in distance <Δd2> over all possible Δt values reached a plateau. Image analysis was done in 3D. The plot of <Δd2> (±1 s.e.m.) in the longer movies for both bw+ and IU bwD tend to plateau at approximately Δt=21 minutes. As observed in A, the plateau height of the graph for IA bwD was significantly lower than that observed for bw+ and IU bwD, indicating that its movement is highly confined.

Fig. 4.

Association of bwD with the centromere greatly decreases the movement and radius of confinement of the locus. Data obtained from short time-lapse movies was used to calculate the diffusion coefficient and step size and to test for directionality of movement in bwD and bw+ nuclei. (A) Movement of the bw locus was followed relative to the closest centromere marked by CIDGFP in undifferentiated cells of the eye imaginal discs. The average mean square change in distance <Δd2> over all possible Δt values (±1 s.e.m.) was plotted for each genotype. The 1.6 Mb insertion of heterochromatic repeats does not alter the dynamics of movement of the bw locus. The diffusion coefficients in Table 2 were calculated by taking the mean value for <Δd2> at Δt= 3 and 6 seconds from the data set shown here. (B) Histogram of the distribution of distances between mRFP-CIDGFP at t=0, from the data shown in A. Note that no bw+ nuclei were observed where the distance between mRFP-CIDGFP signal was less than 1.0 μm. (C) bwD nuclei were further categorized into IU (distance between bwD and the closest CIDGFP dot was >1.0 μm at t=0) and IA (distance between bwD and the closest CIDGFP dot was <1.0 μm at t=0). The graph shows the mean square change in distance of the IA and IU bwD nuclei. (D) Histograms of step size (Δd) at Δt=15 seconds. The mean step size of bwD (IA) was found to be statistically significantly lower with a P-value of <0.001 (t-test), than either bw+ or bwD (IU). (E) To calculate the true radius of confinement for bw+ and IU bwD, chromatin movement was followed for a longer time (1-2 hours) with images taken every 3 minutes to obtain graphs where the plots for the average mean square change in distance <Δd2> over all possible Δt values reached a plateau. Image analysis was done in 3D. The plot of <Δd2> (±1 s.e.m.) in the longer movies for both bw+ and IU bwD tend to plateau at approximately Δt=21 minutes. As observed in A, the plateau height of the graph for IA bwD was significantly lower than that observed for bw+ and IU bwD, indicating that its movement is highly confined.

Anchoring of heterochromatin and constrained chromatin dynamics: calculation of radius of confinement

Loci have never been found to be free to move through the totality of the nuclear space. At the least, movement of a locus is limited to its chromosomal territory (Vazquez et al., 2001; Zink et al., 1998). Stable changes in nuclear organization may arise if an additional constraint in the movement confines a locus to, or excludes it from, a subregion of the chromosomal territory. The graphs in Fig. 4A,C increase linearly, but do not reach a plateau during the duration of the movie. Therefore, to compute the radius of confinement, movement of the tagged loci was followed for a longer time. Images were taken every 30 seconds for a total of 300 seconds (supplementary material Movie 2).

The plateau height on the MSD graph can be used to estimate the radius of confinement, which indicates the size of the volume through which the locus is free to diffuse. Although the graph for IA nuclei reached a plateau, the graphs for the bw+ nuclei as well as the bwD IU nuclei continued their linear trend (supplementary material Fig. S2). Therefore, movies spanning 1-2 hours were made with images taken every 3 minutes (supplementary material Movies 3-5). As opposed to the linearly increasing trend of the plot for MSD for bwD IU and the bw+ locus in the shorter movies (Δt=3 seconds and Δt=30 seconds), the plot in the longer movies (Δt=3 minutes) reached a plateau approximately 21 minutes from the start of the movie (Fig. 4E). We calculated the radius of confinement of these loci from the average MSD at Δt=30 minutes (Table 4).

Table 4.

Radius of confinement

Genotype Locus tagged No. of nuclei Radius of confinement (2D) (μm) Radius of confinement (3D) (μm)
lacO bw+/bw+ bw+-CID   12   1.29   1.13  
lacO bwD/bw+ IU bwD-CID   22   1.19   1.06  
lacO bwD/bw+ IA bwD-CID   7   0.40   0.40  
lacO bw+/bw+  CID-CID   6   0.35   0.35  
Genotype Locus tagged No. of nuclei Radius of confinement (2D) (μm) Radius of confinement (3D) (μm)
lacO bw+/bw+ bw+-CID   12   1.29   1.13  
lacO bwD/bw+ IU bwD-CID   22   1.19   1.06  
lacO bwD/bw+ IA bwD-CID   7   0.40   0.40  
lacO bw+/bw+  CID-CID   6   0.35   0.35  

Exclusion of bw+ from the heterochromatic compartment

An examination of the distributions of the 3D distances between the bw region and the centromere from both the 30-second and 3-minute movies reveal a marked absence of smaller distances in the bw+ data set (Fig. 5A). Indeed, only 1% of the distances were less than 1 μm in bw+, whereas 36% of the distances were less than 1 μm in bwD. If bw+ were free to move through the full volume of the chromosomal domain, it would be expected that by chance the bw locus would get closer to the centromere. However, a clear truncation of the distribution is seen (Fig. 5A) and the traces never continue below 1 μm (Fig. 5B). This suggests that, not only is bw+ not associating with the heterochromatic compartment, but it is actually being excluded from interacting with the chromatin surrounding the centromere. Perhaps the insertion of heterochromatin allows the bwD locus to come into close proximity to centric heterochromatin, giving it a license to enter the compartment.

Decrease in locus dynamics when getting close to the centromere is seen only in bwD nuclei

Similar to our analysis for the shorter movies, we observed a significant decrease in the step size of the bwD locus in IA nuclei when compared with the IU bwD nuclei in the Δt=3-minute movies. Interestingly, within the IU bwD category a significant decrease in the step size of the bwD locus was observed in nuclei where the bw-CID distance was between 1 and 2 μm as compared with those with distances >2.0 μm. From the distance measurements at t=0 we thus separated the nuclei into three categories. The first category included nuclei where the 3D distance between CIDGFP and mRFP at t=0 seconds was >2.0 μm, and the second category included nuclei where the distance was between 1.0 and 2.0 μm. The third category included nuclei where the initial distance was <1.0 μm (Fig. 5C). The step size in the second and third categories are statistically significantly smaller as compared with the step size in the first category (P-value first category vs second category and first category vs third category <0.0001, t-test) (Table 5). Although there is a decrease in the mean step size from the second to the third category, this difference is not statistically significant (P=0.28).

Table 5.

Mean step size (Δt=3 minutes)

Locus tagged Distance from CIDGFP (μm) Mean 3D step size (μm) Standard deviation No. of time points
bw+>2.0   0.381   0.289   196  
bwD>2.0   0.421   0.399   214  
bwD 1.0-2.0   0.236   0.233   331  
bwD<1.0   0.212   0.204   147  
CID-CID   0.16   0.157   105  
Locus tagged Distance from CIDGFP (μm) Mean 3D step size (μm) Standard deviation No. of time points
bw+>2.0   0.381   0.289   196  
bwD>2.0   0.421   0.399   214  
bwD 1.0-2.0   0.236   0.233   331  
bwD<1.0   0.212   0.204   147  
CID-CID   0.16   0.157   105  

We also found that, when the distance between the centric heterochromatic compartment and the bwD locus is >2.0 μm, the fluctuation in distances between the two loci is more apparent (Fig. 5C, left panel) and is similar to that observed for the bw+ nuclei (Fig. 5B). In contrast to bwD, cases in which the first category bw+ gets closer than 2 μm from the centromere are not accompanied by a decrease in step size. In addition, we failed to observe any bwD loci in the third category that `escaped' from close association with the centromere.

Fig. 5.

bwD can get closer to the centromere than bw+. (A) Histograms of all the 3D measurements from all the movies in this study with Δt= 30 seconds and 3 minutes suggest that the heterochromatic insertion in bwD allows it to come into close proximity (<1.0 μm) to centric heterochromatin. (B,C) Traces showing the distances between mRFP-CID at each time point in individual movies (each movie a different color trace). Movement of the tagged bw region in undifferentiated nuclei of the eye imaginal discs was followed for 1-2 hours. Images were taken every 3 minutes. Image analysis was done in 3D. Nuclei were divided into the three categories. The first category contains nuclei where the distance between CIDGFP and mRFP at t=0 was >2.0 μm, the second category contains nuclei where the distance was between 1.0 and 2.0 μm, and the third category contains nuclei where the distance was <1.0 μm. When the distance between the centromere and the bwD locus is >2.0 μm the fluctuation in distances between the two loci is more apparent (left panel), as compared with the change in distance when the relative distance between bwD and the closest centromere is <2.0 μm. (B) Traces from movies of wild-type nuclei. (C) Traces from movies of bwD nuclei. (D) Graph of the distance between the two closest CIDGFP dots in six nuclei. These centromeres do not experience any large-scale movements.

Fig. 5.

bwD can get closer to the centromere than bw+. (A) Histograms of all the 3D measurements from all the movies in this study with Δt= 30 seconds and 3 minutes suggest that the heterochromatic insertion in bwD allows it to come into close proximity (<1.0 μm) to centric heterochromatin. (B,C) Traces showing the distances between mRFP-CID at each time point in individual movies (each movie a different color trace). Movement of the tagged bw region in undifferentiated nuclei of the eye imaginal discs was followed for 1-2 hours. Images were taken every 3 minutes. Image analysis was done in 3D. Nuclei were divided into the three categories. The first category contains nuclei where the distance between CIDGFP and mRFP at t=0 was >2.0 μm, the second category contains nuclei where the distance was between 1.0 and 2.0 μm, and the third category contains nuclei where the distance was <1.0 μm. When the distance between the centromere and the bwD locus is >2.0 μm the fluctuation in distances between the two loci is more apparent (left panel), as compared with the change in distance when the relative distance between bwD and the closest centromere is <2.0 μm. (B) Traces from movies of wild-type nuclei. (C) Traces from movies of bwD nuclei. (D) Graph of the distance between the two closest CIDGFP dots in six nuclei. These centromeres do not experience any large-scale movements.

Movement of centromeres

In S. cerevisiae the movement of chromatin near centromeres has been previously described as being highly constrained (Heun et al., 2001; Marshall et al., 1997). To learn about the dynamics of movement of centromeres in higher eukaryotes, we calculated the diffusion coefficient (Table 2) and step size at Δt=15 seconds (Table 3) from movies with a Δt=3 seconds. A distinct lack of large-scale movements was observed (Fig. 5D), with the largest step at Δt=15 seconds being 0.37 μm. The radius of confinement (Table 4) and step size at Δt=3minutes (Table 5) were calculated from data obtained from longer movies with Δt=3minutes. We found that movement of the centromeres is highly constrained, with a very low diffusion coefficient of 8.3×10–5 μm2/seconds, which is very similar to that observed for the bwD locus in IA nuclei (supplementary material Fig. S3).

Effect of inhibitors of microtubule depolymerization on chromatin dynamics

Treatment of S. cerevisiae with drugs that destabilize microtubules has been reported to result in an increase in the radius of confinement of centromeres (Marshall et al., 1997). Although a second study found that such treatment caused the centromeres to acquire a more central position in the S phase, the authors did not observe a change in chromatin dynamics or any change in organization for other tagged loci (Heun et al., 2001). To find out whether treatment with chemicals that inhibit microtubule polymerization would perturb nuclear organization or chromatin dynamics in higher eukaryotes, we incubated eye imaginal discs in nocodazole and colchicine. Upon treatment, we did not observe a change in association of the bwD locus with centric heterochromatin (Fig. 6A), major changes in bw region chromatin dynamics (Fig. 6B), significant change in the diffusion coefficient of the centromeres (Fig. 6C) or the percentage of cells in Rabl orientation (Table 1). Interestingly, we did observe a significant decrease in the step size of centromeres upon incubation with the microtubule-depolymerizing agents (Table 3), with what appears to be a decrease in the number of larger movements (Fig. 6D). In addition, treatment with microtubule-destabilizing agents caused an apparent decrease in the radius of confinement of centromeres (Fig. 6C), in contrast to the increase in the confinement radius observed in yeast (Marshall et al., 1997).

Discussion

As a cell undergoes differentiation, the nucleus undergoes changes in volume, shape, global chromatin condensation and distribution, and qualitative and quantitative changes in nuclear protein content. Such changes in morphology are bound to impact the transcriptional profile of the nucleus. The bwD system used in this study has served as an important model to examine the correlation between the state of gene expression and its positioning in the nucleus. Previous studies have found that there was an increase in the percentage of cells where the bwD locus was associated with centric heterochromatin in differentiating cells of the eye disc (Thakar and Csink, 2005). To test the generality of this observation, we examined primary cultures of embryonic neuroblasts and found no bwD heterochromatic association in nuclei cultured for only 5 hours. Based on the number of cells (16-20) typically present in the ganglionic clusters after 15 hours of culturing, we hypothesize that the single neuroblast (observed during 5 hours of culture) would have to undergo rapid cell divisions, making it impossible to form associations in the small time period. This is similar to the lack of association observed in nuclei derived from embryos at nuclear stage 13 that are undergoing rapid divisions, and only cycle between the S- and M-phase (Dernburg et al., 1996).

Fig. 6.

Microtubule destabilizers decrease the step size of centromeres, but do not perturb nuclear organization or chromatin dynamics of bwD. (A) Box plots showing the distribution of distances between bwD locus and the nearest CIDGFP signal corrected by the radius. No change was observed in nuclear organization upon treatment with nocodazole and colchicine, two chemicals that destabilize microtubules, as compared with the untreated control. (B) MSD plot (±1 s.e.m.) comparing the movement of bwD-CIDGFP in movies where images were taken at a single plane of focus every 3 seconds. A large change in chromatin dynamics was not observed for the bwD locus upon treatment. (C) MSD (±1 s.e.m.) plots comparing the movement of centromeres in the presence of microtubule-destabilizing agents. Although the diffusion coefficient for the movement of centromeres was comparable, a significant decrease in the radius of confinement was observed upon treatment. (D) Histograms of the distribution of step sizes of centromeres calculated at Δt=15 seconds. A significant decrease in step size of centromeres was observed in nuclei treated with microtubule destabilizers (P<0.0001).

Fig. 6.

Microtubule destabilizers decrease the step size of centromeres, but do not perturb nuclear organization or chromatin dynamics of bwD. (A) Box plots showing the distribution of distances between bwD locus and the nearest CIDGFP signal corrected by the radius. No change was observed in nuclear organization upon treatment with nocodazole and colchicine, two chemicals that destabilize microtubules, as compared with the untreated control. (B) MSD plot (±1 s.e.m.) comparing the movement of bwD-CIDGFP in movies where images were taken at a single plane of focus every 3 seconds. A large change in chromatin dynamics was not observed for the bwD locus upon treatment. (C) MSD (±1 s.e.m.) plots comparing the movement of centromeres in the presence of microtubule-destabilizing agents. Although the diffusion coefficient for the movement of centromeres was comparable, a significant decrease in the radius of confinement was observed upon treatment. (D) Histograms of the distribution of step sizes of centromeres calculated at Δt=15 seconds. A significant decrease in step size of centromeres was observed in nuclei treated with microtubule destabilizers (P<0.0001).

Although the changed nuclear organization is not directly regulated by differentiation-specific signals, differentiation does impose a constraint on chromatin movement (Thakar and Csink, 2005). These observations allow us to model the interplay between cell cycle, nuclear organization and differentiation. In undifferentiated cells of the eye disc the nuclei are cycling and the chromatin is more dynamic. This dynamic property allows both the bw and bwD loci to move around the chromosomal territory in a random-walk manner. During this time a portion of the bwD nuclei will establish associations with the centric heterochromatin within its chromosomal territory. At a later point in G1 the process of differentiation begins imposing constraint on chromatin movement and the organization of the nucleus becomes less flexible. The fact that we do not observe a further increase in the percentage of cells showing association posterior to the morphogenetic furrow despite an increase in time spent in G0 (Fig. 3C), supports the idea that during differentiation constrained chromatin dynamics causes a lockdown of nuclear organization.

The arrangement of chromosomes in the Rabl orientation and its consequent breakdown is another prominent change in nuclear organization during development. Nuclei of early Drosophila embryos are arranged in the Rabl orientation because of the very rapid divisions that consist only of S- and M-phases (Foe and Alberts, 1985). In addition, polytene nuclei of the larva are arranged in Rabl orientation (Hochstrasser et al., 1986). It has been assumed that this arrangement would be the standard for most diploid nuclei in flies. However, our ability to visualize all the centromeres and a chromosome tip concurrently in vivo allows us to dismiss this idea. We find that less than 10% of even the dividing nuclei have this arrangement. Indeed, the percentage of nuclei in the Rabl orientation in the differentiated cells of the eye disc is approximately 2%, which suggests that most of the differentiated nuclei in adults would not be in Rabl orientation. The fact that the Rabl orientation is almost completely lost suggests that breakdown of the Rabl orientation happens early in G1, as has been observed in cells of the larval CNS where a breakdown occurs within the first 2 hours of G1 (Csink and Henikoff, 1998).

A common theme in early studies of locus dynamics is that the movement of chromatin is directionless, but constrained (Gartenberg et al., 2004; Heun et al., 2001; Marshall et al., 1997; Thakar and Csink, 2005; Vazquez et al., 2001). Occasional large-scale movements were observed and these were speculated to help in nuclear reorganization. These studies do not directly address how a stochastic process can achieve the characteristic organization of various nuclear components seen in a wide range of cell types. It can be argued that this is because the tagged loci being observed had either already attained the desired nuclear organization (centromeres and telomeres were positioned near the nuclear periphery) or the specific tagged loci did not exhibit any preferential organization during interphase. More recently, directional movement of a synthetic locus towards the center of the nucleus upon transcriptional induction has been found (Chuang et al., 2006), but it is not yet known how this applies to the natural induction or silencing of genes.

In this study, we tagged the bwD locus with lacO repeats to study the dynamics of chromatin movement that gives rise to the changed nuclear organization. We found no evidence that the journey of bwD towards the centric heterochromatic compartment is directional. It is possible that there is occasional, rare directionality of bwD towards the centric heterochromatin that we were unable to detect. However, a simpler explanation is that bwD moves around in the confines of the chromosomal territory where it occasionally comes into contact with centric heterochromatin, allowing the stable capture of the locus by the compartment.

We also find that the insertion of heterochromatin in bwD allows it to come into close proximity to the centromere. This close proximity to the centromere is never observed for the wild-type locus (Fig. 5A). One interpretation of this observation is that the bwD insertion has a license to gain proximity to the centromere by entering the centric heterochromatic compartment. When, during the random walk of the bwD locus, it approaches centric heterochromatin, a constraint on movement can be imposed. This constraint prevents bwD from moving away from the centric heterochromatic compartment, although it appears to be able to move within the compartment, albeit with slower dynamics. This capture of bwD results in a stable, changed nuclear organization. The change in nuclear organization is thus a result of an active mechanism that prevents the bwD locus from moving away from the centric heterochromatic compartment.

What factors could be responsible for the capture and confinement of bwD? Possible players include proteins involved in the nuclear matrix and the various nuclear lamins. In addition, there are several distinguishing features of heterochromatin based on histone modification, nucleosome density and heterochromatin-specific non-histone proteins. Some of these heterochromatic markers have been found on the insertion in bwD (Delattre et al., 2000; Platero et al., 1998). In general, heterochromatin is hypoacetylated and is thought to be a prominent target for histone deacetylases (HDAC). Several compounds can be used to specifically inhibit HDACs, however we found no effect of two of these compounds on bwD association or any chromatin dynamics (Rajika Thakar, Nuclear organization and the dynamics of chromatin movement in Drosophila, PhD thesis, Carnegie Mellon University, 2005). These results indicate that constraint is not mediated by factors immediately directly sensitive to the histone acetylation state. Similar observations were made by Gilchrist et al. (Gilchrist et al., 2004), where treatment with an HDAC inhibitor did not alter the nuclear organization of centromeres. However, sensitivity to HDAC inhibitors may be influenced by the proliferation state of the cell and the length of treatment with the inhibitor (Taddei et al., 2005). Unfortunately, treatment with higher concentrations of the inhibitors or for longer time induced cell death in our cultures of eye imaginal discs. In addition, it has been found by ourselves and others (Pile et al., 2001) that feeding larvae HDAC inhibitors slows larval developmental progression. Such developmental delay would itself have an effect on nuclear organization that would make interpretation of such experiments difficult.

Another potential player in regulating chromatin dynamics was suggested by earlier work done on yeast, where inhibition of microtubules was seen to partially release some loci from confinement (Marshall et al., 1997). We treated primary cultures of Drosophila cells with microtubule destabilizers, but found no perturbation of nuclear organization or chromatin dynamics of bwD-CID. Surprisingly, this treatment did result in a significant decrease in the step size and radius of confinement of centromeres. We believe that the change in the dynamics of centromeres (CID-CID) did not give rise to a discernable change in the movement of bwD-CID in the IU category because of a difference in magnitude in their movements, such that the small change in the movement of centromeres did not sufficiently change the relatively larger movements of bwD-CID in IU nuclei. Interestingly, we did not observe a change in the movement of bwD-CID in IA nuclei either (the dynamics of whose movements are similar to that of centromeres), suggesting that the movement of the bwD locus is truly independent of microtubules, in that it continues to track along with the altered centromeric movement.

There are several possible ways that microtubules could influence centromere movement. As microtubules and tubulin are usually concentrated in the cytosol, their influence on centromere dynamics may be an indirect influence on the nuclear envelope. Such an impact on the nuclear envelope could be caused by change in kinetics of polymerization of tubulin outside the nucleus. Since the centromeres are peripherally located, the movement of these may be altered, whereas more centrally positioned loci would be insensitive. Alternatively, movement of centromeres may be influenced directly by nuclear tubulin that has been observed in some interphase nuclei (Walss et al., 1999) and associated with the nucleolus (Andersen et al., 2002).

Fig. 7.

Models of bwD behavior near the centromere. A constraint in movement is imposed when the distance between bwD-CIDGFP is between 1 and 2 μm. The distribution of the traces in Fig. 6 could be interpreted in two ways. (A) Constraint may be imposed wherever bwD contacts the surface of the heterochromatic compartment and the seemingly large variation in distance from the centromere could be because of the variable position of the centromere within the mass of heterochromatin. (B) bwD moves within the centric heterochromatin, varying its distance to the centromere. Possibly the endpoint of this random walk is a stable, close association with the centromere.

Fig. 7.

Models of bwD behavior near the centromere. A constraint in movement is imposed when the distance between bwD-CIDGFP is between 1 and 2 μm. The distribution of the traces in Fig. 6 could be interpreted in two ways. (A) Constraint may be imposed wherever bwD contacts the surface of the heterochromatic compartment and the seemingly large variation in distance from the centromere could be because of the variable position of the centromere within the mass of heterochromatin. (B) bwD moves within the centric heterochromatin, varying its distance to the centromere. Possibly the endpoint of this random walk is a stable, close association with the centromere.

Previous work of ours found that despite the constraint in chromatin dynamics imposed at differentiation, the speed of movement was unchanged (Thakar and Csink, 2005). By contrast, here we find that the constraint on movement of the bwD locus not only causes a decrease in the radius of confinement, but it is also accompanied by a decreased diffusion coefficient. A similar decrease has been observed in Drosophila spermatocytes (Vazquez et al., 2001). However, it is possible that the decreases in the diffusion coefficients are artifacts of the microscopy system. It could be that the movement is so constrained that it reaches a plateau within the short time period of imaging, such that the diffusion coefficient is not being calculated from the linear portion of the MSD graph.

Models to explain the apparent variation in the size of the area of constraint suggest insights into the organization of the centric heterochromatin and the dynamics of movement within the heterochromatic compartment. The centromere encompasses only a small subset of the area occupied by the centric heterochromatic compartment of the second chromosome (Fig. 7). Constraint might occur as soon as bwD encounters the centric heterochromatic compartment. This model would assume that the centromere itself can be near the edge of the heterochromatic compartment (Fig. 7A). However, this model is contradicted by the fact that we never see the bw+ locus in very close association with the centromere, as would be expected if it were closer to the bulk of euchromatin. Alternatively, one could interpret our data (Fig. 5C) as indicating two levels of constraint, the first one imposed when bwD encounters and then enters the centric heterochromatic compartment (distance between bwD-CID decreases to 1-2 μm), and a secondary constraint might then be imposed when, within the compartment, the bwD locus encounters the centromere, resulting in stable association (Fig. 7B). The nuclear components that mediate this confinement and exclude the euchromatic loci from close association with the centromere will be the focus of future work.

Materials and Methods

Fly lines

All fly lines were maintained at 25°C on cornmeal molasses medium, except when reared for egg laying, which was done on grape-agar plates (Ashburner, 1989). The movement of the bw locus was followed by tagging it with lacO repeats. To mark the wild-type locus we had previously mobilized P{wmc, lacO} on the X chromosome (Vazquez et al., 2001), to replace P[ry+t7.2=PZ] present in the Dcp-1 locus (15 kb distal from bw) (Thakar and Csink, 2005). In a new screen we mobilized P{wmc, lacO} on X to replace bwDP{hsp-w-hsp26-pt-T}Dcp-1 present in the Dcp-1 locus on the same homolog as the bwD insertion. The position and the length (10.1 kb) of lacO repeats were checked by polytene FISH and Southern blot analysis. Fly line w1118, P{wmc, Ubq-mRFP-LacI-NLS} is as previously described (Thakar and Csink, 2005). Fly line w1118; P{wmc, CIDGFP} was obtained from S. Henikoff (Fred Hutchinson Cancer Research Center, Seattle, WA).

Measurement and imaging of chromatin movement

To study the dynamics of chromosomal movement w1118; P{wmc, CIDGFP} bwD P{wmc, lacO} Dcp-1/CyO or w1118; P{wmc, CIDGFP} P{wmc, lacO} Dcp-1/CyO females were crossed to w1118, P{wmc, Ubq-mRFP-LacI-NLS}; P{wmc, CIDGFP} males to obtain w1118, P{wmc, Ubq-mRFP-LacI-NLS}/w1118; P{wmc, CIDGFP} bwD P{wmc, lacO} Dcp-1/P{wmc, CIDGFP} larvae for imaging. We found that for optimal visualization of the centromeres, two copies of the CIDGFP-containing transgene were required. To visualize the lacO-tagged loci, only one copy of the transgene bearing the mRFP-LacI sequence was required, because with two copies we could not distinguish the signal for the fluorescent protein bound to the repeats over the background fluorescence from the unbound protein. Eye-antennal discs were dissected from crawling third-instar female larvae obtained from the cross described above. In each experiment data were collected from three to 11 discs for each genotype and/or treatment.

Primary cultures of discs were maintained as previously described (Thakar and Csink, 2005). For imaging, each disc was always placed on the coverslip in the same orientation. As eye discs have a natural curvature, the disc was placed such that the bottom part of the curvature was in contact with the coverslip, and the flaps on the side, which curled up, were pressed down flat on the coverslip. The layer of undifferentiated cells imaged was the first layer of cells encountered after the peripodial membrane. Differentiated cells present posterior to the morphogenetic furrow were chosen with the criterion that they were part of the ommatidial clusters and could be easily discerned based on their morphology. These are not a part of the layer of cells arrested in G2.

All microscopy on nuclei was performed using a Deltavision system (Applied Precision) built around an Olympus IX70 microscope. The images were gathered with a cooled CCD camera (Micromax 350, Photometrics) using a 100×/1.35 NA objective (Olympus). Time-lapse movies were made using the Attofluor cell chamber (Molecular Probes) or the FCS2 chamber (Bioptech). The chamber type did not influence the viability of the tissues, as indicated by the uptake of calcein AM (Molecular Probes) and the lack of nuclei staining positive for ethidium bromide (EtBr) (data not shown). All images were acquired and deconvolved using Softworx (Applied Precision).

To assess the degree of association between bwD and CIDGFP dots, 3D stacks of images were taken, and image analysis was done on the 2D projected image. Distance between the mRFP signal and closest GFP signal was measured using Softworx (Applied Precision) and corrected by the nuclear radius computed by drawing polygons around the nucleus demarcated by the background fluorescence of the unbound mRFP-LacI fusion protein.

For the shortest time-lapse movies, images were taken at a single plane of focus every 3 seconds for a total of 45 seconds. For the longer movies, 3D optical stacks 0.5 μm apart were taken every 30 seconds for 300 seconds or every 3 minutes for 1-2 hours. For the movies lasting >1 hour, refocusing to optimal z-section was typically required because of drift during microscopy. The mean square change in distance <Δd2> was calculated for all possible values of Δt for an individual nucleus. The average value of <Δd2> obtained from at least six nuclei or as reported in the Tables, was plotted on the graph. The error bars in all the graphs are one standard error. The diffusion coefficient was calculated by taking the mean value for <Δd2> at Δt=3 and 6 seconds.

All analysis for the longer movies was done in 3D for the following reason. During live imaging of our primary cultures with the longer time lapse we also observed the nuclei undergoing what is generally referred to as `nuclear rotation' (supplementary material Movies 3, 6) (Paddock and Albrecht-Buehler, 1988). Similar rotations are not observed in S. cerevisiae or some mammalian cells. Although the significance of these rotations remains a mystery, it is very likely to be highly cell-type specific, as it was most pronounced in the cultured neuroblasts. In our study, we found that nuclear rotation gave a false sense of movement, because upon analysis in 2D the magnitude of movement (scale for the y-axis), radius of confinement and step size calculated for bw+ and the IU bwD nuclei from the MSD plots was larger than that obtained from measurements in 3D (supplementary material Fig. S4; Table 4).

Image analysis was done using a code written for Matlab 7.0. In brief, the tagged loci were tracked by the following criteria: the tracker should be positioned on local peaks of fluorescence in the new image; the new position of the tracker should not be far from its position in the previous image; and no two tracked points should have exactly the same XYZ coordinates. A detailed description of the algorithm as well as the source code for the program can be viewed and downloaded at http://www.cs.cmu.edu/~ggordon/JCSDotTrack/DotTrack.zip. All statistical procedures are as described in Sokal and Rohlf (Sokal and Rohlf, 1981).

Treatment with microtubule destabilizers

Eye imaginal discs were cultured in media supplemented with microtubule destabilizers for 1 hour. All imaging was done in the presence of the inhibitors. The final concentration of inhibitors added was 100 μM nocodazole (Sigma) and 100 μM colchicine (Sigma) as modified from Ashburner (Ashburner, 1989). Stock solution of nocodazole (10 mM) was prepared in dimethyl sulfoxide (DMSO), and colchicine (1 mg/ml) was prepared in water. As a control, media with equivalent volume of DMSO and water was added. After all imaging, viability of the tissue was checked by incubating the tissue in EtBr for 1 minute; only data obtained from discs that did not take up any EtBr was included in the analysis. Concentrations of nocodazole above 100 μM typically caused some cell death, as observed by the uptake of EtBr.

Primary cultures of embryonic neuroblasts

Primary cultures of embryonic neuroblasts were established by modifying the protocol described by Furst and Mahowald (Furst and Mahowald, 1985). Approximately 100-150 embryos were washed in sterile dH2O, 70% ethanol and dechorianated in 2.5% bleach made up in 50% ethanol for 2 minutes. This surface sterilizes the embryos and also helps remove the excess yeast that may have carried over from the grape plates. After dechoriation the embryos were washed with 70% ethanol, followed by three washes in Chan and Gehring's modified media (Ashburner, 1989). The embryos were transferred to a microfuge tube containing 200 μl Shield and Sang's media (Sigma) supplemented with 2.5% fetal bovine serum (FBS) (Gibco) and Penicillin-Streptomycin (Gibco). They were homogenized with disposable blue pestles (Kimble Kontes). The homogenate was passed through a 10 μm PP column (Lida), collected in a sialized microfuge tube and spun at 1240 g for 1 minute at room temperature. The pellet was suspended in 600 μl media, and passed through a pulled sialized pasteur pipette at least five times. The 100 μl suspension of cells was plated on glass-bottom dishes (MatTek Cultureware). The cells were allowed to adhere to the glass coverslip for 2 hours, the media was changed and the adhering neuroblasts were imaged 5 hours or 15 hours from the time of seeding. Immunofluorescence on cultures was performed on cells fixed with 3.7% paraformaldehyde for 20 minutes. The anti-ELAV antibody was kindly provided by J. Pollock (Duquesne University, Pittsburgh, PA) and used at a dilution of 1:600.

Acknowledgements

We thank Andrew Belmont of the University of Illinois and Steven Henikoff of the Fred Hutchinson Cancer Research Center for fly lines. Andrew Gove provided advice concerning the calculations and software used in this paper. This work was supported by an American Cancer Society grant (RSG-00-073-04-DDC) to A. Csink.

References

Andersen, J. S., Lyon, C. E., Fox, A. H., Leung, A. K., Lam, Y. W., Steen, H., Mann, M. and Lamond, A. I. (
2002
). Directed proteomic analysis of the human nucleolus.
Curr. Biol.
12
,
1
-11.
Ashburner, M. (
1989
).
Drosophila: A Laboratory Handbook
. Cold Spring Harbor: Cold Spring Harbor Laboratory Press.
Campbell, R., Tour, O., Palmer, A., Steinbach, P., Baird, G., Zacharias, D. and Tsien, R. (
2002
). A monomeric red fluorescent protein.
Proc. Natl. Acad. Sci. USA
99
,
7877
-7882.
Chuang, C. H., Carpenter, A. E., Fuchsova, B., Johnson, T., de Lanerolle, P. and Belmont, A. S. (
2006
). Long-range directional movement of an interphase chromosome site.
Curr. Biol.
16
,
825
-831.
Comings, D. E. (
1980
). Arrangement of chromatin in the nucleus.
Hum. Genet.
53
,
131
-143.
Csink, A. K. and Henikoff, S. (
1996
). Genetic modification of heterochromatic association and nuclear organization in Drosophila.
Nature
381
,
529
-531.
Csink, A. K. and Henikoff, S. (
1998
). Large-scale chromosomal movements during interphase progression in Drosophila.
J. Cell Biol.
143
,
13
-22.
Delattre, M., Spierer, A., Tonka, C. H. and Spierer, P. (
2000
). The genomic silencing of position-effect variegation in Drosophila melanogaster: interaction between the heterochromatin-associated proteins Su(var)3-7 and HP1.
J. Cell Sci.
113
,
4253
-4261.
Dernburg, A. F., Broman, K. W., Fung, J. C., Marshall, W. F., Philips, J., Agard, D. A. and Sedat, J. W. (
1996
). Perturbation of nuclear architecture by long-distance chromosome interactions.
Cell
85
,
745
-759.
Foe, V. E. and Alberts, B. M. (
1985
). Reversible chromosome condensation induced in Drosophila embryos by anoxia: visualization of interphase nuclear organization.
J. Cell Biol.
100
,
1623
-1636.
Furst, A. and Mahowald, A. P. (
1985
). Differentiation of primary embryonic neuroblasts in purified neural cell cultures from Drosophila.
Dev. Biol.
109
,
184
-192.
Gartenberg, M. R., Neumann, F. R., Laroche, T., Blaszczyk, M. and Gasser, S. M. (
2004
). Sir-mediated repression can occur independently of chromosomal and subnuclear contexts.
Cell
119
,
955
-967.
Gilchrist, S., Gilbert, N., Perry, P. and Bickmore, W. A. (
2004
). Nuclear organization of centromeric domains is not perturbed by inhibition of histone deacetylases.
Chromosome Res.
12
,
505
-516.
Harmon, B. and Sedat, J. (
2005
). Cell-by-cell dissection of gene expression and chromosomal interactions reveals consequences of nuclear reorganization.
PLoS Biol.
3
,
e67
.
Henikoff, S., Ahmad, K., Platero, J. S. and van Steensel, B. (
2000
). Heterochromatic deposition of centromeric histone H3-like proteins.
Proc. Natl. Acad. Sci. USA
97
,
716
-721.
Heun, P., Laroche, T., Shimada, K., Furrer, P. and Gasser, S. M. (
2001
). Chromosome dynamics in the yeast interphase nucleus.
Science
294
,
2181
-2186.
Hochstrasser, M., Mathog, D., Gruenbaum, H., Saumweber, H. and Sedat, J. W. (
1986
). Spatial organization of chromosomes in the salivary gland nuclei of Drosophila melanogaster.
J. Cell Biol.
102
,
112
-123.
Marshall, W. F., Straight, A., Marko, J. F., Swedlow, J., Dernburg, A., Belmont, A., Murray, A. W., Agard, D. A. and Sedat, J. W. (
1997
). Interphase chromosomes undergo constrained diffusional motion in living cells.
Curr. Biol.
7
,
930
-939.
Misteli, T. (
2005
). Concepts in nuclear architecture.
BioEssays
27
,
477
-487.
Paddock, S. W. and Albrecht-Buehler, G. (
1988
). Rigidity of the nucleus during nuclear rotation in 3T3 cells.
Exp. Cell Res.
175
,
409
-413.
Pile, L. A., Lee, F. W. and Wassarman, D. A. (
2001
). The histone deacetylase inhibitor trichostatin A influences the development of Drosophila melanogaster.
Cell. Mol. Life Sci.
58
,
1715
-1718.
Platero, J., Csink, A., Quintanilla, A. and Henikoff, S. (
1998
). Changes in chromosomal localization of heterochromatin binding proteins during the cell cycle in Drosophila.
J. Cell Biol.
140
,
1297
-1306.
Qian, H., Sheetz, M. P. and Elson, E. L. (
1991
). Single particle tracking. Analysis of diffusion and flow in two-dimensional systems.
Biophys. J.
60
,
910
-921.
Robinett, C. C., Straight, A., Li, G., Willhelm, C., Sudlow, G., Murray, A. and Belmont, A. S. (
1996
). In vivo localization of DNA sequences and visualization of large-scale chromatin organization using lac operator/repressor recognition.
J. Cell Biol.
135
,
1685
-1700.
Sokal, R. R. and Rohlf, F. J. (
1981
).
Biometry
. New York: W. H. Freeman.
Spector, D. L. (
2003
). The dynamics of chromosome organization and gene regulation.
Annu. Rev. Biochem.
72
,
573
-608.
Taddei, A., Roche, D., Bickmore, W. A. and Almouzni, G. (
2005
). The effects of histone deacetylase inhibitors on heterochromatin: implications for anticancer therapy?
EMBO Rep.
6
,
520
-524.
Thakar, R. and Csink, A. K. (
2005
). Changing chromatin dynamics and nuclear organization during differentiation in Drosophila larval tissue.
J. Cell Sci.
118
,
951
-960.
Vazquez, J., Belmont, A. S. and Sedat, J. W. (
2001
). Multiple regimes of constrained chromosome motion are regulated in the interphase Drosophila nucleus.
Curr. Biol.
11
,
1227
-1239.
Walss, C., Kreisberg, J. I. and Luduena, R. F. (
1999
). Presence of the betaII isotype of tubulin in the nuclei of cultured mesangial cells from rat kidney.
Cell Motil. Cytoskeleton
42
,
274
-284.
Zink, D., Cremer, T., Saffrich, R., Fischer, R., Trendelenburg, M. F., Ansorge, W. and Stelzer, E. H. (
1998
). Structure and dynamics of human interphase chromosome territories in vivo.
Hum. Genet.
102
,
241
-251.

Supplementary information