NEW YORK (GenomeWeb) – Caerus Molecular Diagnostics is developing a new single-molecule sequencing technology that promises long reads with low error rates.
The Mountain View, California-based startup, which currently has two employees, is working on proof of concept for its so-called activator sequencing method, which relies on the release of an activator from dNTP that switches on an enzyme.
"We don't have any sequencing data yet, although we hope to be able to get sequencing data in the coming months," said Javier Farinas, the company's founder, CEO, and CTO.
Farinas, a former director of assay development at Caliper Life Sciences, founded Caerus in 2008 and has been funding it through government grants. In 2010, for example, the company won a $500,000 grant from the National Human Genome Research Institute to explore a label-free sequencing method that it has since abandoned.
This was followed by other grants, including one through the Qualifying Therapeutic Discovery Project program to develop a diagnostic platform based on the label-free method, and another from NHGRI to develop a microfluidic mRNA integrity assay. And a year ago, Caerus received a three-year, $701,000 grant under NHGRI's Advanced DNA Sequencing Technology program to focus on the current activator sequencing technology.
Initially, Caerus worked on a label-free sequencing method dubbed Millikan sequencing, named after the famous 1909 oil drop experiment by Robert Millikan to measure the charge of an electron. That method relies on recording changes in charge that occur when a polymerase adds a nucleotide to DNA. "Coming from a biophysics background, it was appealing to use this 100-year-old method that was extremely sensitive," Farinas explained. "You can get sensitivities down to a single electron using this technique, with a really rudimentary apparatus."
However, it turned out to be tricky to apply the Millikan approach to sequencing, which requires the system to be submerged in water, and after working on it for a few years, the researchers concluded that it was not going to be possible to use this method for single-molecule sequencing.
For amplified DNA, they did show proof-of-principle of sequencing, which they published last month in Analytical Biochemistry. However, it did not make sense to develop that approach commercially. "It was pretty evident that we were not being able to catch up to the rest of the field using that method," Farinas recalled, noting that Ion Torrent came out with the PGM around the same time, which offered many of the same benefits, including label-free sequencing, that Millikan sequencing promised. "Basically, they had a commercial product and we had some preliminary data," he said.
Thus, Caerus switched gears and has been working on the activator sequencing method, for which it has filed several patents, for the last two years or so. "I wanted to come up with a method that was allowing both long read lengths but would have high base calling accuracy," Farinas said.
The approach is conceptually based on a sequencing chemistry developed by Sunney Xie at Harvard and published in 2011 that uses terminally labeled dNTPs where the label only becomes fluorescent in the presence of phosphatase. "That works nicely but it doesn't necessarily give you a great benefit over what's commercially available," Farinas said.
For the activator sequencing approach, he attached not a fluorophore but an enzyme activator to the terminal phosphate of the dNTP. When a polymerase incorporates the nucleotide into DNA, the activator is set free and turns on an engineered enzyme, which converts a substrate into a detectable product, for example, a fluorophore. The enzyme acts as a molecular amplifier of the signal because a single activator leads to a large number of enzyme products. This amplification, in turn, reduces the sequencing error rate.
A key step was to identify an enzyme that is mostly inactive in the absence of the activator and shows high activity in its presence. That enzyme was eventually provided by Marc Ostermeier, a professor at Johns Hopkins University who had been studying protein switches. His lab had developed a fusion protein between maltose binding protein and beta-lactamase that uses maltose as an activator to switch on the enzyme. Lactamase activity can be measured in various ways, including fluorescence, luminescence, or pH changes. "It was the perfect enzyme for this approach, and I was fortunate to collaborate with him and get the ability to use the enzyme," Farinas said.
His lab is now testing the fusion protein for sequencing, using dNTPs that release maltose when they are incorporated into DNA. "The preliminary data to date suggest that the enzyme does in fact have the properties necessary to work in such a system. We're just in the midst of extending those results and trying to show sequencing," he said. "The key is, will we be able to get sequencing data at the single-molecule level with low enough error rates to be competitive with everything else that's out there, or is going to be out there in two to three years?"
While the switch protein is not perfect, he said, the parameters of the system can hopefully be tweaked to provide a good enough signal for sequencing. After that, it will be "just normal development issues" of getting the system to work fast enough and finding a detection system that is sufficiently sensitive.
One open question is how the sequencing method will deal with long stretches of homopolymers. "Initial calculations suggest that it should do well but that's probably going to be the dominant error rate that's going to limit the usefulness of the technique," Farinas said. "We have to explore that in the future."
The initial goal is to generate read lengths on the order of several thousand bases, although read lengths could go into the tens of thousands of bases in the future. The key benefit of the technology would be its low error rate, with a target on the order of 0.1 percent, compared to the high error rates of current single-molecule sequencing methods from Pacific Biosciences or Oxford Nanopore Technologies.
In addition, unlike other single-molecule sequencing methods, the activator sequencing would be highly scalable, Farinas said, because it would be easy to parallelize. It could use a number of detection methods, including non-optical ones, but Caerus currently relies on fluorescence as its readout method because it is convenient.
The company hopes to be able to generate sequence data within the next six to 12 months and plans to raise private funding at that point in order to develop the technology further.