The cell accomplishes this coiling and uncoiling dance through the concerted effort of gyrases and topoisomerases. Yet no one had ever been able to study the basic behavior of supercoils, even on naked DNA.
To attack the problem, Dekker's team attached a 21-kb segment of fluorescently labeled DNA to a glass capillary, while on the other end he coupled a magnetic bead. Then, he pulled that DNA straight up (perpendicular to the slide) and applied a rotational magnetic force to induce a supercoil. Finally, he turned the molecule on its side, so that the DNA helical axis was parallel to the slide, and watched what happened.
In this configuration, DNA supercoils (or plectonemes) appear as bright points on a dimmer fluorescent line. Initially, though, the team could make no traction because DNA supercoiling requires, well, supercoiling. “If you have a single nick in your 20,000 base pair-long molecule…that broken backbone will now basically swivel around the unbroken one and all the torsion is released,” explains Dekker. Thus, the team had to optimize labeling and imaging conditions to stave off photodamage as long as possible, for instance by distancing the fluorophore from the DNA backbone and adding an oxygen-scavenging system. In the end, they were able to record molecules for all of about two seconds before they unwound — but it was enough.
When Dekker's team did record the behavior of the plectoneme spots, they observed two distinct forms of motion. The first was expected, a kind of slow diffusion as the supercoils translocate left to right and back again. The other motion, though, was completely unanticipated: Supercoils that could essentially “hop,” almost instantaneously, from one location to another, sometimes kilobases away.
“This ‘writhe,’ as we call it, can sort of transfer,” he explains. “It can coil out in this position, and coil in this [other] position…. So, the global arrangement of this whole supercoiled structure can suddenly rearrange.”
Whether the same behavior occurs in living cells remains to be determined, as Dekker's experiment involved naked DNA in the absence of protein. Still, the experimental conditions did use near-physiological salt concentrations and forces. “The parameters where we explore these dynamics are very close to what would be relevant in your cells,” he says.The big picture
In 2011, Job Dekker (no relation to Cees), of the University of Massachusetts Medical School, published a paper examining the overall structure of the 4-megabase circular chromosome of the bacteria, Caulobacter.
Dekker's primary interest pertained to organization, “whether the genome…is randomly organized inside the cell or does it have some kind of preferred position so that particular DNA sequence elements reside in a particular location in the cell.”
To address this question, Dekker's team first used a long-range interaction mapping technique called 5C to identify segments of the chromosome located in close spatial proximity. From those data, a low-resolution model of the Caulobacter chromosome was developed that suggested the molecule exists as a “slightly twisted ellipsoid structure.” At one pole of that ellipsoid, as it turns out, is a sequence associated with chromosome segregation called parS.
The team suspected parS could define the positioning of the ellipsoid in the cell. When they moved parS to different positions along the circle, the chromosome maintained its ellipsoid form, but rotated to keep the parS sequence at one pole, like a taut, twisted rubber band wrapped around two posts. “That was really, I think, one of the first examples as far as I know where a chromosome structure was first solved to inform or to help identify cis-acting sequence elements that contribute to building that structure,” notes Dekker.
The question was, how does that ellipsoid reside in the cell?
To determine that, the team used in vivo imaging of cells in which an array of lac operator sites was positioned at various locations along the genome. The parS sequence was visualized using a yellow fluorescent protein; lacO was cyan. The data showed that parS was always located at one pole of the rod-shaped cell, while the lacO array moved along the cell's long axis according to its genomic position. From the “zero” position up to two megabases away, lacO moved progressively further away from parS; farther than that, it began to work its way back toward parS.