Cornell University researchers are looking to commercialize a method for selectively sorting single DNA molecules based on their epigenetic features as they move through a nanofluidic device.
Down the road, the team plans to pair the fluorescence-activated sorting system — which is akin to flow cytometry, but operates on a much smaller scale — with high-throughput sequencing for discerning the identity of molecules with or without specific DNA and/or chromatin modifications. The researchers have formed a spinout company called Odyssey Molecular to commercialize the technology.
"With flow cytometry, you want to look at one cell at a time — you want to identify that you have a cell and you want to then identify what surface markers it has, based on fluorescent antibodies," Cornell University genetics and development researcher Paul Soloway told In Sequence. "We're doing the same thing, except with chromatin fragments."
Along with Cornell University applied and engineering researcher Harold Craighead, Soloway is co-corresponding author on a proof-of-principle study on the single-molecule sorting method that appeared online in the Proceedings of the National Academy of Sciences in mid-May.
In the paper, he and colleagues from Cornell explained how they go about selectively tagging specific epigenetic marks — methylated cytosines in double-stranded DNA, in the case of the PNAS study — using fluorescently labeled binding proteins or antibodies. From there, they demonstrated it is possible to sort DNA based on such epigenetic features by automatically tracking fluorescence signals as single molecules move through the inspection volume of a nanofluidic chip.
"We have described the development of a nanofluidic device that can be used as part of an epigenetic analysis workflow," the study's authors explained.
"This device provides nanofluidic confinement to isolate individual molecules by voltage-actuated flow," they elaborated. "These attributes have enabled us to perform real-time fluorescence detection and automated sorting on individual molecules based on their fluorescence signature."
The researchers are hoping to commercialize the system through Odyssey Molecular, which was co-founded in 2010 by Soloway, Craighead, and study co-author Stephen Levy, also at Cornell.
That commercialization effort is "still very much in the development stage," according to Soloway, who said the company has secured some funding and is negotiating with potential industrial partners about the prospect of collaborating.
Nevertheless, the researchers are optimistic that their approach will eventually find favor for researchers in a variety of fields who are interested in selectively studying individual molecules with specific epigenetic marks.
In particular, because it appears possible to fluorescently tag multiple DNA and/or histone-associated marks along the same stretch of sequence, Soloway said the single-molecule sorting method may be useful for directly studying multiple epigenetic features associated with development, cellular differentiation, and a range of other processes.
In addition to the sorting method described in the PNAS study, Soloway noted that the fluorescence-activated nanofluidic technology is also being used as the basis for an analytical system that quantifies the levels of various epigenetic marks on DNA or on the histone proteins associated with it.
A study introducing that potential application of the technology was published in Analytical Chemistry in 2010 and Soloway said the team recently submitted a follow-up study that builds upon and extends its initial epigenetic enumeration efforts.
"There's the enumeration, which allows us to count the abundance of different epigenetic features quantitatively," he explained, "and there's also the sorter that allows us to recover molecules of interest for downstream library preparation and sequencing."
Until now, the ability to detect epigenetic modifications on chromatin has relied primarily on approaches such as chromatin immunoprecipitation, or ChIP, alone or in combination with sequencing, while methods that involve bisulfite conversion coupled with sequencing have been used to look at cytosine methylation within DNA.
But while Soloway called ChIP and bisulfite sequencing "wonderful methods," he argued that they have limitations, particularly when trying to look at multiple epigenetic marks at the same time.
For instance, he explained, Massachusetts General Hospital, Harvard University, and Broad Institute researcher Brad Bernstein and his colleagues defined bivalent domains involving a combination of gene activating and gene silencing chromatin modifications at many developmentally regulated genes in embryonic stem cells.
In a 2006 Cell study, that team relied on a technique known as re-ChIP, which involves taking immunoprecipitate captured with one antibody and re-precipitating it with another.
Though the re-ChIP method proved crucial for demonstrating the presence of H3 lysine 27 methylation and H3 lysine 4 methylation on the same stretch of DNA, the approach is rarely used, owing to technical challenges associated with it, Soloway explained.
"That was one of the motivations for us developing this [molecule sorting] method," he said. "What we were hoping to do was to provide more richness of information to existing epigenetic methods than was currently possible."
The sorting method itself hinges around nanofluidic devices that have been custom-fabricated from silica at Cornell.
Within these chips, molecules that have been labeled with fluorescently tagged probes targeting specific modifications move through channels with an inspection volume that's around 160 attoliters — small enough to be able to see individual molecules, while large enough to allow for reasonable throughput, Soloway explained.
"At that small confinement volume, we can look at solutions of one nanomolar of chromatin fragments and, based on Poisson statistics, be quite confident that we're looking at a single molecule at a time," he said.
From there, each molecule reaches a fork where it is automatically siphoned based on its fluorescence signature: molecules bearing fluorescent signals associated with the epigenetic mark of interest are sent one way, while those without are sent another.
Again, the principle resembles fluorescence-activated cell sorting, but because researchers are looking at signals stemming from chromatin fragments rather than whole cells, the nanofluidic sorting system requires a much more sensitive detector.
In the research setting, the team has been reusing the same nanofluidic devices whenever possible, though Soloway explained that they do eventually become clogged and need to be replaced.
The hardware components of the system — which currently include a confocal microscope, lasers mounted on a light table to excite the fluorescent molecules, optical components, and avalanche photodiodes for detection — remain unchanged from one experiment to the next.
In the PNAS study, researchers used Life Technologies' red dye TOTO-3 to label both methylated and unmethylated DNA plasmids that had been linearized with restriction enzymes.
To distinguish between double-stranded DNA with or without methylated cytosines, they also used a version of the DNA methyl binding protein MBD1 that had been tagged with green fluorescent protein. The researchers reported that they were able to sort molecules with 98 percent accuracy in this system.
And though they verified the identity of output molecules using quantitative PCR in the current study, researchers explained that it should also be possible to test these epigenetically sorted molecules with microarrays or high-throughput sequencing.
Soloway noted that sequencing platforms with relatively straightforward library preparation processes and low input volume requirements, such as the Helicos system, might be preferable, given the small amounts of sorted genetic material coming out of the nanofluidic system.
Even so, he emphasized that the researchers have not made any decisions about which platforms they will pair with the fluorescence-activated nanofluidic device.
The team is currently looking at the library preparation steps associated with various instruments and hopes to take a crack at sequencing some of the material sorted using the nanofluidic system over the next 12 months or so.
"There are additional changes in library preparation that we may need to implement because we're dealing with sorted molecules — and low amounts of them," Soloway said. "We're in the process of optimizing some library preparation methods that could be used with different platforms."
In addition to the detection of individual epigenetic marks, the team is also keen to use its nanofluidic approach for gauging the presence of two or more epigenetic features on the same chromatin fragment or DNA molecule — something that appears to be possible by using multiple, independent antibodies or binding proteins linked to distinct fluorescent tags.
"If we wanted to take a look at more than two, we could conceivably do that as well using our single molecule method," Soloway added. "But that is almost certainly out of the question with existing chromatin IP methods."
While Soloway conceded that there are concerns over whether interactions between different methylation-binding proteins and/or antibodies might hinder the ability of each to bind at authentic epigenetic marks, he and his colleagues have not run into such problems so far.
In unpublished experiments, they found that they could successfully sort molecules based on the presence or absence of both the histone trimethylation mark H3K9 and DNA methylation during their first attempt to directly detect two epigenetic marks at once.
"For that combination of epigenetic features, there's no interference that we can discern," Soloway said.
"Hopefully we'll be able to expand upon this as we push this forward," he added, noting that more research will be needed to determine whether there are any epigenetic features for which detection with the nanofluidic system is mutually exclusive.
The researchers are also working on ways to increase the throughput of their system — both by adding more nanofluidic channels and tweaking the preparation protocol for the input sample — as they move toward commercialization of their epigenetic sorting and enumeration devices.
"One of the challenges we have now is that our research device is only a single-channel device, which limits our throughput," Soloway said. "Some of our commercial efforts are focused on trying to improve throughput by orders of magnitude."