Scanning tunneling microscopy of deoxyribonucleic acid during replication

Previous scanning tunneling microscopy (STM) experiments have produced several images of deoxyribonucleic acid (DNA) molecules, but never that of DNA in the act of replicating. We present such an image, obtained on an uncoated and unlabeled molecule extracted during replication from the cell. The results were highly reproducible, and generally consistent with previously reported cases of STM imaging of coated or uncoated DNA. I 8 Our micrographs show the parent molecule in the act of splitting; the Y -shape characteristic of the replication process is clearly imaged. Scanning tunneling microscopy is emerging as a very promising technique for studying biological specimens with atomic-scale resolution. The technique is particularly suitable for small-size molecular structures, since the small size reduces the problems due to electrical insulation. Among those, DNA is by far the most interesting. Other authors have reported STM images of uncoated DNA molecules, either in water or in air. H Most of such studies were performed on unlabeled DNA. In the present case, images were obtained on uncoated and unlabelled DNA in air, thereby reproducing the excellent experimental conditions of Ref. 3. Highly polymerized DNA extracted from salmon (Sigma 0-1626), suspended in a 4% phosphate buffered saline solution (PBSS), was deposited on atomically flat substrates obtained by cleaving highly oriented pyrolitic graphite (HOPG). After deposition, the DNA-containing solution drops were mechanically spread on the substrate and then dried in an oven at 50°C for 2 h. This approach produces DNA specimens suitable for STM imaging without metal coating or labeling. The preparation technique and the STM methodology were tested by performing experiments on a number of specimens. We confirmed many of the empirical findings of Ref. 3. For example, we did find that the bias can produce a significant interaction between tip and specimen, thereby altering the observed molecules between subsequent scans. We also found that periodic structures can be simulated by the graphite surface after deposition of DNA-free saline solution. Such structures, however, have periodicities one order of magnitude larger than DNA. In order to rule out spurious substrate effects in the images discussed here, we examined approximately 500 images from DNA-free regions, never finding structure that could be confused with DNA. We agree, therefore, with Ref. 3, and find that graphite substrates are much better than polished metal substrates in producing high-quality images. We empirically found that the thermal treatment of the specimen during dehydration is extremely important in producing stable DNA molecules on graphite substrates. The replicating DNA molecule is shown by the STM images of Figs. 1 and 2. These images were produced at atmospheric pressure in the topographic mode, using a modified McAllister scanning tunneling microscope with a gold tip. Prior to taking micrographs of the DNA molecule, the STMs performance was checked by imaging the graphite substrate with atomic-scale resolution. Figure 1 (a) shows a 438 X 436 A 2 area, with simulated colors; Fig. 1 (b) presents the same image with a three-dimensional plot. Figure 2(a) presents a grey-scale image of the small area in Fig. 1 (a), 221 X 177 A 2, and Fig. 2 (b) a three-dimensional plot of a slightly larger area, 221 X 221 A • The images are part of a set including a total of27 micrographs; each part of the same molecule was imaged several times over a period of hours, demonstrating the excellent reproducibility of the experiment and the stability of the specimen. In particular, we found no evidence of the tip-specimen interaction effects that were previously mentioned (Ref. 3). From Figs. 1 and 2, one observes the typical periodicity related to the DNAs double helix. This periodicity is emphasized by the line height profile of Fig. 2 (c), derived from the data of Fig. 2 (a) for the dashed line following the axis of the DNA branch on the left-hand side. We estimate a period of approximately 44 A. This is well within the range of periodicity values reported by other authors, and in particular by