Mitotic chromosome structure: reproducibility of folding and symmetry between sister chromatids

Abstract

Mitotic chromosome structure and pathways of mitotic condensation remain unknown. The limited amount of structural data on mitotic chromosome structure makes it impossible to distinguish between several mutually conflicting models. Here we used a Chinese hamster ovary cell line with three different lac operator-tagged vector insertions distributed over an approximately 1 microm chromosome arm region to determine positioning reproducibility, long-range correlation in large-scale chromatin folding, and sister chromatid symmetry in minimally perturbed, metaphase chromosomes. The three-dimensional positions of these lac operator-tagged spots, stained with lac repressor, were measured in isolated metaphase chromosomes relative to the central chromatid axes labeled with antibodies to topoisomerase II. Longitudinal, but not axial, positioning of spots was reproducible but showed intrinsic variability, up to approximately 300 nm, between sister chromatids. Spot positions on the same chromatid were uncorrelated, and no correlation or symmetry between the positions of corresponding spots on sister chromatids was detectable, showing the absence of highly ordered, long-range chromatin folding over tens of mega-basepairs. Our observations are in agreement with the absence of any regular, reproducible helical, last level of chromosome folding, but remain consistent with any hierarchical folding model in which irregularity in folding exists at one or multiple levels.

Figure 1

Visualization of longitudinal and axial positioning of closely spaced lac operator-containing vector inserts. Isolated mitotic chromosomes from the MC8_I clone show three nearby vector transgene sites (top chromosome arm), with a fourth transgene site on the opposite chromosome arm. Lac repressor and topoisomerase II staining labeled vector transgene sites and chromosome axes, respectively. (A): (a) Lac repressor staining of lac operator-containing vector insert sites; (b) anti-topoisomerase II antibody staining; (c) total DNA stained with DAPI; (d) merged lac repressor (cyan) and topoisomerase II (red) staining; (e) 3× enlargement from (d); (f) merged DAPI (yellow) and lac repressor signal (cyan); (g) 3× enlargement from f. (B): (a–f) Six examples of isolated metaphase chromosomes. Subpanels (left to right): topoisomerase II immunostaining, lac repressor staining, merged (topoisomerase II red, lac repressor cyan), merged, 3× enlargement. (C): Two stereo pairs of metaphase chromosomes reconstructed from optical sections. Total DNA stained with DAPI (blue), anti-topoisomerase II (red), and lac repressor staining (green). (D) 2D model of a fragment of a chromosome with lac op inserts: lac op spots, green; sister chromatid axes, red. Bars: 1 μm.

Figure 2

Quantitative analysis of an imaged CHO metaphase chromosome with lac operator inserts. (A) Anti-topo II immunostaining. (B) Lac op immunostaining. (C) Superimposition of A and B. (D) 3× enlarged region from C. (E) Extraction of geometrical parameters from D: (green and red) centers of extracted 3D coordinates of lac op spots of left and right chromatids, respectively; (blue and cyan) traces of centers of topo II signal for left and right chromatid axes, respectively; (black and pink) 4th degree polynomial and linear fits of axes traces, respectively. (F) Model of metaphase chromosome used for Monte Carlo simulation. The distance between a spot and the chromatid axis was distributed uniformly between 0 and R_max; angles θ and θ′ were distributed uniformly between 0 and 2π. Spot positions in sister chromatids were assigned independently, modeling the absence of any correlation of spot positioning in sister chromatids. Bars: 1 μm.

Figure 3

Longitudinal spot positions are reproducible but show significant variability. (A) Distances between spots 1 and 3 (Fig. 1D) vary significantly between sister chromatids. Scatter plot where each spot corresponds to measured distances on the left and right chromatids for a particular metaphase chromosome. (B) Correlation between chromatid arm length versus longitudinal separation of spots 1 and 3, with plotted values representing average of measurements on sister chromatids. (C) Longitudinal coordinates of individual spots vary within ∼300 nm. Histogram of differences between longitudinal coordinates of “sister” spots, e.g., 1L and 1R. (D) Histogram of distribution of longitudinal distances between spots 1 and 2 normalized by distance between spots 1 and 3.

Figure 4

Lateral positions of individual spots relative to chromatid axis are nonreproducible, uncorrelated between sister chromatids, and independent of the degree of chromosome condensation. (A) Distance histogram for spot 1L to chromatid axis. Positive distances correspond to spots located on the left (exterior) of the chromatid axis, and negative distances correspond to spots on the right (interior) of the axis. (B) Histogram of axial positions for all spots. (C) Scatter plot showing comparison of lateral spot positions for sister chromatids: left spot (x axis) versus right spot (y axis). (D) Axial position versus length of chromosomal arm for the top pair of lac op spots of all chromosomes; (gray circles) right chromatids; (solid circles) left chromatids. (E) Scatter plots from 2D projections of actual (open circles) and simulated (black circles) chromosome measurements demonstrate comparable distributions for axial positions of lac operator spots. Mean values, SDs, and absence of correlation between sister chromatids were similar for experimental and simulated distributions.

Figure 5

Similar distributions of spot positions relative to chromatid axes for real and computer-generated data. Minimal RMS deviations in pixels for spots of sister chromatids were found through an optimization procedure between original chromosomes (abscissa) and chromosomes in which a mirror reflection of the right chromatid was performed (ordinate). The left chromatid was fixed. For the right chromatid, 2D translation and rotation of the projection was allowed for the 2D case (A and B), and 3D translation and rotation around the chromatid axis was allowed for the 3D case (C and D). (A) 2D projections, real data. (B) 2D projections, simulated data. (C) 3D real data. (D) 3D simulated data.