Scientists formerly employed by electron microscopy sequencing firm Halcyon Molecular have published a method they developed at the company for stretching single DNA strands and placing them on a substrate.
The technique, called molecular threading, was developed in collaboration with researchers at Harvard University, and allows users to deposit DNA ranging in size from nanometers to tens of micrometers in a very straight and uniform manner, facilitating the analysis of their base sequence from EM images.
Halcyon stopped working on DNA sequencing last year (Daily Scan 8/22/2012), and Palo Alto-based Aeon Biowares, which employs several former Halcyon researchers, has taken over commercialization of the firm's sequencing-related technologies.
The principle of molecular threading, which was invented by Halcyon co-founders William and Michael Andregg, is simple: a small glass needle coated with a hydrophobic polymer is inserted into a solution of DNA, where the DNA is believed to sit coiled up in small balls near the surface. When the needle is pulled out of the drop, it picks up a single or multiple strands of DNA, depending on the pH and other environmental conditions, and stretches the thread out in the air. The needle can then deposit the linear strand onto a surface, for example a substrate for imaging by transmission electron microscopy.
In their paper, published this summer in PLOS One, the researchers showed that they can place single-stranded and double-stranded DNA of more than 100 kilobases in length within 100 milliseconds or less per strand. By attaching the needle to a piezo-positioner, they were able to create arrays of parallel DNA strands on a substrate.
According to Christopher Own, a corresponding author of the paper, molecular threading offers a number of advantages over molecular combing, a widely used method for depositing DNA onto substrates for imaging.
For example, while molecular combing works well for applications that use light microscopy, the DNA is not straight enough for accurate EM-based sequencing. "Molecular threading, as we report in the paper, is at least a factor of 10 straighter and less wavy over long distances than molecular combing," he said.
In addition, DNA is more easily controlled by molecular threading, so it can be placed into arrays of very high density. By comparison, Own and his colleagues found that when they used molecular combing, the spacing between strands was much greater, and the start point of each strand was unpredictable.
Own, who used to be principle research physicist at Halcyon and oversaw the firm's EM and imaging work, believes that molecular threading is "an enabling step" for sequencing DNA by transmission EM. "Other groups have tried to sequence using combed DNA, and we tried to do it as well," he said. "The big issue with doing it without threading is that the stretching consistency over long distances is indeterminate. If you don't know what the stretching is in a particular portion of the DNA, then reading off the base position becomes very difficult."
Further, the speed of sequencing increases with straighter DNA. "If the DNA is wiggly or wavy, you have to chase it over two dimensions," Own explained. "With molecular threading, it's almost like a railroad track, you just follow the DNA in a single direction, and typically, you don't have to move so much laterally when you're imaging. That results in fewer wasted pixels, and the throughput is much faster if you have stretched and well-linearized DNA."
"The technique appears to provide excellent consistency, which is necessary for any commercial applications," said William Glover, founder and CEO of ZS Genetics, another company that has been developing transmission EM-based sequencing (IS 7/16/2013).
While more work needs to be done to eliminate bundling of several DNA strands, he said, "overall, it looks promising."
However, depositing linear strands of DNA on a substrate is just one of several steps required for EM-based sequencing, and Halcyon focused its efforts on other areas as well, including functionalized tags to label the DNA bases, imaging, and data analysis.
Last month at the Microscopy and Microanalysis annual meeting, Own presented Halcyon data showing that his team was able to label a single type of base in single-stranded DNA, using a multi-heavy-atom label attached by a non-enzymatic method, and determine the distance between the bases by EM imaging. In areas where the base density is low, the accuracy is good, he said, and the technique is sensitive to copy number variations over long distances.
Going below 10 nanometers, the sensitivity is not yet sufficient to determine the sequence accurately, though, likely because the labels interact with each other. "One way to improve that is to decrease the length of the linkers that bind the base to the backbone," Own said.
Once that problem is solved, reconstructing the entire DNA sequence from four superimposed images would be "quite straightforward," he said. "Getting accurate reads at the nucleotide level for monochromatic DNA is a prerequisite, and it's a major step."
Halcyon stopped working on DNA sequencing sometime last year, apparently for economic reasons. Aeon Biowares, which employs several former Halcyon researchers, has taken over the commercialization of the firm's technologies, and has licensed the intellectual property around molecular threading from Halcyon.
"On the business side, Aeon Biowares is the successor for Halcyon Molecular’s DNA sequencing-by-electron microscopy efforts," Chris Melville, Aeon's CEO, told In Sequence in an e-mail message. Besides molecular threading, this includes "enabling technologies for the other components of a complete sequencing platform," such as microscopy labels and purification and analysis methods for labeled products. These could also be applied in areas other than sequencing. "We are the decision-making party for any potential licensing or co-development inquiries related to these technologies," he said.
While "the currently demonstrated killer app for molecular threading is sequencing," Melville said, "any fairly linear polymer that is sufficiently long and soluble is a candidate for threading," and the method could possibly be applied to proteins or carbon nanotubes as well.
He said that Aeon has not yet licensed the molecular threading technology to other parties.