Fresh views on DNA structure

Abstract

BioTechniquesVol. 53, No. 5 Tech NewsOpen AccessFresh views on DNA structureJeffrey M. PerkelJeffrey M. PerkelSearch for more papers by this authorPublished Online:3 Apr 2018https://doi.org/10.2144/000113949AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinkedInReddit Even though nearly 60 years have passed since James D. Watson and Francis Crick solved the structure of DNA, we still have much to learn about the iconic double helix.“These are complex molecules, and only now are [we] developing the techniques, the technology that allows us to study them systematically with high precision,” says Carlos Bustamante, professor of molecular and cell biology at the University of California, Berkeley.Bustamante should know: In 1992, his lab pioneered single-molecule biophysical studies of DNA when they attached a magnetic bead to the double helix and measured elasticity as force was applied along the molecule's helical axis. Yet 13 years later, using the same technique, they were surprised to learn that the textbook understanding of how DNA responds to such force — an understanding that they, in part, developed — was incorrect.The conventional wisdom, explains Bustamante, was that pulling the DNA molecule taut would cause it to unwind and then stretch. And indeed, in 1996, his team determined that by yanking on a piece of DNA with progressively greater force that the molecule can exist in any of three elasticity “regimes.” But, in 2005, his team discovered that during the first regime, as the molecule is being extrended, it actually overwinds instead of unwinding. It is only at a certain force level that the molecule begins to unwind, and finally, to physically stretch. The idea that a helix would overwind upon being extended is clearly counterintuitive. And yet, this is precisely what was observed. How a helix responds (unwinding or overwinding) to extension is determined by its so-called torsional-extension coupling constant. That constant was believed to hover around +200 pN nm. But, in fact, Bustamante's team measured it at -90 pN nm.“We discovered that not only did we have the wrong value for the torsional-extension coupling constant of DNA, but we actually had the wrong sign for it,” says Bustamante. “So, there are still a lot of things that need to be understood about a molecule like DNA.”Call it a technology tipping point in DNA structure research, but today scientists are applying a battery of novel microscopy techniques to understand DNA at length scales from the atomic to the genomic. That's important, says Bustamante, because each approach reveals a different aspect of the molecule, as in the classic story of three blind men describing an elephant's appearance. “It is only through the concerted and coordinated view of the molecule from different aspects that we are going to ultimately get a much better appreciation of what makes the DNA such a special molecule.”David Bazett-Jones harnessed the power of electron microscopy to debunk the 30-nm chromatin fiber.Courtesy of David Bazett-JonesTaekjip Ha, a physicist at the University of Illnois, Urbana-Champaign, uses single-molecule FRET to study DNA bending.Courtesy of Brian StaffersSingle-molecule candy canesBart Hoogenboom takes his view of DNA from atomic force microscopy (AFM).In AFM, a sharp microscopic tip at the end of a flexible diving board-like cantilever scans over the sample line by line like a Gramophone needle, mapping the topography of the sample beneath it. Typically, that mapping is achieved by measuring the deflection of the cantilever as the stylus is dragged across the sample, but that scanning mode can both drag and distort the molecules being imaged. As a result, DNA in an AFM typically lacks the resolution of crystalline samples, looking more like a squashed rope with its helical substructure obscured.Yet AFM, explains Hoogenboom, a University College London physicist, offers multiple advantages for imaging DNA. Unlike optical microscopy, AFM's resolution is not limited by the diffraction limit of light (200 nm or so); it offers single-digit nanometer resolution. And unlike electron microscopy and X-ray diffraction, AFM can be applied to individual molecules in aqueous solution.Hoogenboom and his team recently optimized the method by implementing a suite of changes, including shrinking the cantilever about 10 times and measuring topography by changes in the cantilever resonance frequency as it passes over the sample. They call the method “constant-amplitude phase-modulation AFM.”Imagine scaling up the atoms in the sample to the size of a ping pong ball, explains Hoogenboom. The cantilever tip would be like an inverted mountain, several miles high, hovering above it.“Think of moving this whole mountain up and down by, to scale, about a few meters,” he explains. “We measure how the resonance frequency of this mountain changes because it interacts with the ping pong ball. And all this we do without crushing the ping pong balls.”Using this method, Hoogenboom and research associate Carl Leung were able to reproducibly pick out candy cane-like striations corresponding to the periodic major and minor groove spacing of DNA. Unlike standard amplitude-modulation AFM, the molecule was not squished in the process, producing a diameter of about 1.8 nm for double-stranded DNA (the accepted diameter is 2 nm). Even more remarkably, they not only detected the standard right-handed B-form DNA but also left-handed conformations in the same molecule, although the biological significance of that observation is unclear.Now, Hoogenboom hopes to apply his technique to study DNA replication in real time. “It opens the possibility of doing experiments which help us to understand the replication process of genetic information at a scale which have been hardly accessible before, except by looking at just large ensembles of molecules,” he says.Regular or al dente?Taekjip Ha also explores DNA at the single-molecule level. But rather than structure, Ha, a physicist at the University of Illinois, Urbana-Champaign, investigates the molecule's biophysical properties using a technique he has helped pioneer called single-molecule fluorescence-resonance energy transfer (FRET).Ha's research interest concerns DNA flexibility. Over long distances, DNA can be thought of as a highly flexible rod — like cooked spaghetti. But there is controversy over the molecule's behavior over short distances, say 100 bases or so. One camp believes that DNA at this length scale behaves more like a rigid rod — dry spaghetti, perhaps. The other camp suspects the molecule at this length scale is more akin to a wet noodle. So, the question is, how does nature take its DNA: Regular or al dente?This distinction is not merely academic. If DNA is rigid, then transcription factors must be able to recognize a linear molecule and bend it, a situation that requires energy. The alternative is to capture the transiently bent DNA form and stabilize it. “These are two extreme possibilities,” explains Ha.To figure it out once and for all, Ha and his Ph.D. student Reza Vafabakhsh used single-molecule FRET and TIRF microscopy to measure the ability of molecules between 67 and 106 base pairs to form a “loop,” or circularize, joining two complementary single-stranded overhangs coupled to a FRET donor and acceptor pair. The molecules were tethered at their centers to a glass slide and monitored over time as salt was added to facilitate DNA hybridization.The data showed that even molecules as short as 67-nucleotides can loop, albeit less efficiently than longer molecules, suggesting that transcription factors actually stabilize bent conformations rather than bend rigid ones. (Unpublished data push that limit down to 50 bases, says Ha.) Furthermore, they found that some sequences bend more efficiently than others, with equivalent-length molecules looping at different rates based on the number of consecutive adenine residues they contain.According to Ha, these findings potentially explain nucleosome distribution across the human genome. “Sequence-dependent flexibility of the DNA may influence where the nucleosomes are formed on the DNA, and where it can also regulate the gene expression process,” he says.Ha hopes to apply his technique to determine how modified bases, such methylcytosine, or mismatches influence flexibility. In the meantime, when it comes to the question of DNA flexibility, Ha says he's satisfied that question has been answered. “I think if someone wants to debate it further they'll have to come up with an even better experiment.”Hopping supercoilsCees Dekker is applying extension of the technique Bustamante pioneered, magnetic tweezers, to explore DNA biophysics at longer length scales. Dekker is interested in the dynamics of DNA supercoils, hyper-wrapped topological structures resembling a tangled telephone cord that form, for instance, as molecular motors (such as RNA polymerases) travel along DNA templates.According to Dekker, the management of supercoiling is a critical biological problem. “This coiled stuff is in your nucleus, and to express the genes there, to read off the genetic information, you have to locally uncoil [the] DNA molecule.”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.Taekjip Ha measured DNA flexibility using this single-molecule FRET assay. DNA looping, stabilized by complementary single-stranded ends, brings a FRET donor into close contact with a FRET acceptor, producing a measurable fluorescent signal.Courtesy of Taekjip HaAtomic force microscopy of DNA revealing the periodic spacing of a stretch of right-handed DNA.Courtesy of Carl Leung and Bart HoogenboomTo 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.Supercoil “hopping.”(A) A “kymograph” of a 21-kb DNA molecule at in the absence of supercoiling shows a homogeneous fluorescent intensity along the molecule. (B) The same DNA molecule after applying 61 positive turns is shorter and contains bright fluorescent spots (i.e., supercoiling). At 2.9 s the molecule nicks due to photodamage, and immediately uncoils. Image series showing the nicking event in detail; the time interval between the images shown is 20 ms. Source: Science Express, 13 Sept 2012, doi: 10.1126/SCIENCE.1225810“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 pictureIn 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.Where's the 30-nm fiber?(A) A phosphorus map obtained by electron spectroscopic imaging (ESI) of a region of a mouse embryo fibroblast nucleus. A large block of compact chromatin is present in the center of the field, and represents a constitutive heterochromatin domain called a chromocenter. The domain is transcriptionally repressed and is composed largely of repetitive satellite DNA. According to the prevailing models of chromatin organization, this chromatin should be structured as 30-nm and higher order fibers. Since the physical section contains many overlapping chromatin fibers in the chromocenter, tomography is required to resolve the overlapping fibers in 3D. (B) An ESI/tomogram obtained from phosphorus maps of the field shown in A. (C-D) Regions of the tomogram in B shown at higher magnification. 30-nm fibers are not observed in this cell nucleus. Only 10-nm chromatin fibers (arrowheads) are observed both within the chromocenter itself (D) and on its periphery (C,E). Scale bar represents 500 nm in A, B and 150 nm in C-E. Image courtesy of David Bazett-Jones, adapted from EMBO Reports, Sept. 18, 2012, doi: 10.1038/embor.2012.139.“So, it basically says the genomic position is perfectly related to where it's located in the cell,” says Dekker — an observation that appears to have little impact on gene expression, but does influence chromosome segregation and replication timing. Dekker is now working to determine if similar behavior occurs in eukaryotic cells.David Bazett-Jones of the Hospital for Sick Children in Toronto is also interested in chromatin structure, specifically the superstructure of eukaryotic chromosomes and how these are folded within cell nuclei.The textbook definition is that DNA organizes first into 10-nm fiber “beads-on-a-string” configuration. That fiber then folds into a 30-nm fiber, and it is this structure that helps compact two meters of DNA into a micron-scale nucleus, and that comprises most of the transcriptionally silent “closed” chromatin in cells — or so researchers thought.According to Bazett-Jones, the-30 nm fiber is easy to identify in cell-free conditions. “You just sort of glare at a 10-nm fiber in solution, in cell-free conditions, and it will form a 30-nm fiber.”The question was, does it exist in vivo? To find out, Bazett-Jones combined two electron microscopy techniques, electron spectroscopic imaging (ESI), that is used in materials science to study elemental composition, and electron tomography, which enables 3D reconstructions of samples. Bazett-Jones' team applied this approach to interphase mouse nuclei, where, using nitrogen-to-phosphorous ratios they could clearly detect the 10-nm fibers. But, there was nary a 30-nm fiber to be found.Those data appeared in EMBO Reports, and the implication, says Bazett-Jones, is that the 10-nm fiber suffices to pack DNA into nuclei. Nevertheless, says Bazett-Jones, the old guard still needs convincing. “There is a population of molecular cell biologists who really want to hang on to the 30-nm fiber. They grew up with it, and it's going to be a little while to get that out of their system.”Nearly 60 years after Watson and Crick, it could be time to update the textbooks again.FiguresReferencesRelatedDetails Vol. 53, No. 5 Follow us on social media for the latest updates Metrics History Published online 3 April 2018 Published in print November 2012 Information© 2012 Author(s)PDF download