Researchers in Japan have used scanning tunneling microscopy to visualize the positions of guanine bases in long, single-stranded DNA — an achievement that they said paves the way to direct DNA sequencing.
According to the scientists, who published their results online in Nature Nanotechnology last week, the results are "a step towards the realization of electronic-based single-molecule DNA sequencing." Although they only identified one type of the four bases precisely in their study, they said that with variations to their method, "it should be possible to identify all of the base molecules."
"They have accomplished something that has been a dream of scanning probe microscopists for over 20 years," said William Andregg, CEO and CSO of Halcyon Molecular, a company that has been working on sequencing DNA by transmission electron microscopy, a different imaging method.
In STM, a scanning tip passes over the surface of a metallic or semiconducting sample, and measures tunneling currents, which are a function of the so-called density of states of the sample, and can be translated into an image.
Using tunneling microscopy obtained by lock-in detection, a way to obtain a weak signal from a noisy background, on short, single-stranded synthetic DNA oligomers, the authors — Hiroyuki Tanaka and Tomoji Kawai from the Institute of Scientific and Industrial Research at Osaka University — found that guanine bases differ from other bases as well as the underlying substrate in having a characteristic peak in their electronic state near an applied voltage of -1.6 volts.
"To the best of our knowledge, there are no previous STM reports where the characteristic state density of [DNA] base molecules, adsorbed onto a metal substrate, has been detected in this way," they wrote.
Using the same method, they went on to study long single-stranded DNA. In the past, they wrote, studies of DNA by STM have been hampered by "the unsuitability of the sample preparation methods," which usually require molecules to be deposited onto a surface from the gas phase. To overcome this limitation, they had previously developed a pulse-injection technique for depositing biological molecules — such as DNA — that cannot be vaporized. However, that method "suffers from the fact that [single-stranded] DNA can easily form secondary structures," they write, so there are no long stretches of DNA available for sequencing.
As a solution, the researchers have now developed a method for stretching out and fixing single-stranded M13mp18 phage DNA, which is about 7.2 kilobases long. The method relies on flow effects resulting from the oblique injection of a DNA solution onto a copper substrate using the pulse-injection technique.
After fixing the DNA, they were able to obtain an image of about 140 bases in which bright spots that correspond to guanine bases "match almost perfectly" with the known positions of Gs, "illustrating that STM can be used to sequence the guanine of real DNA."
At the moment, though, the approach is far from practical, and the researchers acknowledged that in order to "make further advances towards the use of STM as a practical tool for sequencing," they need to overcome a number of hurdles, such as being able to recognize all four types of bases, increase the speed, and improve the precision of their method.
Measurements are currently slow, requiring between a second and a minute for each point in a spectrum. According to Andregg, the imaging speed presented in the paper would translate to more than 10,000 years for sequencing a human genome.
One way to increase the speed, according to the authors, would be new STM software and control mechanisms that can find chain-shaped molecules, such as DNA, so no time is wasted on scanning regions where no sample is present.
Further time and cost could be saved by sequencing "from the topographic image alone without using a lock-in amplifier," they wrote, "but it would still be necessary to use a method for identifying contamination."
Another way to speed up the process would be to distinguish guanine by comparing two images obtained at a certain voltage, or comparing the "STM bias dependence." If that turns out to be feasible, "it should be possible to perform sequencing at high speed without the need for spectroscopy," they wrote.
Another problem will be to visualize all four types of bases. "If vibrational spectroscopy is performed using inelastic tunneling spectroscopy, it should be possible to identify all of the bases molecules," the authors noted.
According to Halcyon's Andregg, the method's speed could probably be improved by four orders of magnitude, but that might still not be sufficient for most sequencing applications. "I have not seen a clear path for any probe-based imaging technique to give the kind of throughputs that are needed to enable personalized medicine," he told In Sequence by e-mail.
[ pagebreak ]
He also cautioned that the specialized conditions needed to inject the DNA into a vacuum "are not scalable."
Andregg's startup company, Halcyon Molecular, has been working on a different way to image single molecules of DNA directly, using electron microscopy (see In Sequence 11/11/2008), but is not yet ready to talk about its results in public.
At a meeting in San Diego in April organized by the National Human Genome Research Institute, George Church, a professor at Harvard Medical School and an advisor to Halcyon, said the company uses transmission electron microscopy and labels the DNA bases.
William Glover, CEO of ZS Genetics, agreed that speed would be a major challenge for the Japanese researchers' STM approach. "Scanning microscopes tend to be a good deal slower than other types, so they would need to be careful about what their applications were," he said. "For it to be a general-purpose instrument, the speed would be a big engineering challenge. But they might not need it to be a general-purpose instrument."
For example, Glover said, the STM method — because of its potentially long reads — could be useful for sequencing through stretches of DNA that are otherwise difficult to sequence or assemble, even if it only produced a high-resolution map of the guanine bases. "For hard-to-sequence areas ... the problem is reassembly," he said. "Something like this could completely solve that problem."
His own company, which has also been working on a transmission electron microscopy-based approach to sequencing DNA (see In Sequence 5/27/2008), is currently improving its methods "while being protective of our capital," he said. ZS Genetics raised an undisclosed amount of funding over the last year but is waiting to scale up its operations until capital market conditions have improved, he added.