By Kevin Davies
February 11, 2010 | EXCLUSIVE -- In a wide-ranging presentation at CHI’s Molecular Medicine Tri-Conference (MMTC) earlier this month, Stanford professor Ron Davis offered the first peek at the basics of an ingenious electrical sensing system that will apparently constitute at least part of the third-generation sequencing platform to be unveiled shortly by Ion Torrent Systems.
Ion Torrent Systems, which has offices in Guilford, CT, and San Francisco, has been in stealth mode since the irrepressible Jonathan Rothberg founded the company in 2007, the latest in a string of successful biotech launches including CuraGen, 454 Life Sciences, and RainDance Technologies. The impermeable secrecy cloak around Ion Torrent – something of a Rothberg trademark – finally looks set to be lifted later this month as Rothberg delivers the coveted closing talk at this year’s Advances in Genome Biology and Technology (AGBT) conference on February 27.
But Davis, who is on the scientific advisory board of Ion Torrent, presented some tantalizing snippets of information on the technology at CHI’s flagship molecular medicine conference in San Francisco. The Stanford chemist reviewed some published work from his laboratory, led by Nader Pourmand, now a faculty member at University of California Santa Cruz.
“We’ve licensed this to Ion Torrent, and they’ve done a fabulous job of CMOS implementation and re-engineering the technology. We just did it in a very basic, simple way. They’ve turned it into a much more robust device,” said Davis. “I think it might be quite disruptive.”
One of the underlying themes of Davis’s research is the need for speed in DNA sequencing, rather than what he views as an unhealthy obsession with maximizing capacity. “There’s a lot of need [for speed], especially clinical, where that paradigm might not fit very well,” said Davis. He touched upon a variety of approaches, many from his own lab, including variations on pyrosequencing using $10 CMOS (complementary metal oxide semiconductor) chips or thermal detection rather than light, as well as exciting technologies from other companies such as Halcyon Molecular, which is directly reading DNA sequence using heavy metal labeling and electron microscopy. “If it works, it could be very inexpensive. You won’t need huge data collections because you can read these long drafts,” said Davis.
But of particular interest was technology developed a few years ago by Pourmand. Born in Iran, Pourmand’s family moved to Sweden after the revolution, where he went to university. He later moved to Stanford as a postdoc, where he said his goal has been “pushing the envelope and making DNA sequencing cheaper and faster,” preferably using electronics. (Pourmand agreed to discuss his academic work with Bio-IT World, but stressed he was under a non-disclosure agreement and could not discuss Ion Torrent.)
“I don’t like optics!” he continued. “Bulky, expensive. Any time you deal with optics, you have to label molecules, so you don’t detect first-order effects. If you don’t label and have a way to detect, you have a way to detect first-order effect.”
Initially, Pourmand worked with Mostafa Ronaghi (now at Illumina) on pyrosequencing, trying to find ways to detect the negatively charged pyrophosphate that is released as a by-product during DNA synthesis. In 2000, Pourmand even launched a company called Xagros, but after it floundered in the economic crash of 2001, he returned to Stanford, where he focused on electrical engineering approaches.
Bring Back the Spark
“Using optical systems or luminescence is not ideal for biology,” said Pourmand. “If we bridge biology with electrical engineering, we might achieve goals we can’t do right now.” His first patent for a sensor to detect nucleotide incorporation into a DNA strand was filed that year and issued in 2006. At that time, Pourmand didn’t fully understand the reason behind the negative charge, because of the presence of Magnesium ions shielding the pyrophosphate. His research was funded by an NIH grant to develop a portable CMOS-based platform for pathogen detection.
In early 2004, Pourmand and Davis realized the basis of the net negative charge when a nucleotide is incorporated into a DNA strand -- a single hydrogen ion (proton) is released upon each incorporation event, which rapidly diffuses away from the sensor. Pourmand filed a patent on this sensor in 2005, which was published in 2006.
In April 2006, Pourmand, Davis and colleagues published a paper in the Proceedings of the National Academy of Sciences, entitled “Direct electrical detection of DNA synthesis.” The PNAS paper was a proof-of-principle, but as one of numerous ingenious ideas for novel DNA sequencing methodologies emanating from the Davis lab, did not attract undue attention at the time. Davis described the idea in his MMTC talk as follows:
“You take a gold plate and tether a DNA sequence to it. If you put a base in, what happens is you fix a negative charge … on the DNA. It releases a proton in the process…. What we think is being detected here is really a transient electrical disturbance. You feed this into an amplifier, and you get a transient signal. This is actually very fast, because you’re seeing the base… It’s only determined by the rate of triphosphate binding and coupling… This can also be done on a CMOS chip.”
In the 2006 paper, Pourmand attached a DNA oligonucleotide to a gold electrode (other metals will also work). A single nucleotide is then incorporated by DNA polymerase into the complementary DNA strand. As the base is incorporated, it introduces a single negative charge into the DNA duplex, with the concomitant release of a proton from the 3’ hydroxyl group of the DNA primer, which rapidly diffuses away. The negative charge, if sufficiently close to the electrode, induces a compensating positive surface charge on the highly polarizable electrode. This in turn results in a small electrical pulse, which summed across all of the templates results in a transient current detected by a voltage-clamp amplifier.
Although no sequence data using the CMOS chip were reported then or since, Pourmand et al. suggested that the induced charge effect held steady for a ‘detection zone’ up to 30 microns from the electrode – a space that, in principle, could accommodate DNA strands of many thousand bases. In that zone, Pourmand explained, “the sensor can pick up the signal. Immediately it will disappear, because it is a transient signal.”
By washing away the nucleotide solution and introducing another nucleotide, much like pyrosequencing, one can repeat the cycle many times over. “If the nucleotide incorporates, you see the current – the sIgnal is proportional to number of nucleotides getting incorporated,” says Pourmand. However, unlike pyrosequencing, the CMOS system does not require any modified nucleotides or enzymatic cascades. “You’re using as natural as possible dNTPs and only a single enzyme,” says Pourmand. This also improves accuracy when measuring repetitive sequences.
Pourmand’s own research has focused on building a “GeneMeter,” with a CMOS chip and a small but growing number of array features. In 2008, his group described a multiplexed detection system using a 6x6 grid of electrodes. That work has since progressed to a 10x10 array, or up to 100 different sequences on a chip. The potential advantage of using CMOS is large-scale manufacturing and fabrication, which can be done very cheaply in mass production. “With more users, it will become cheaper, exactly like a CD player,” says Pourmand. “It’s absolutely the simplest platform you can think of.”
Speed and Size
Asked about the potential speed of sequencing using the CMOS approach, Pourmand says, “The incorporation rate for nucleotides using polymerase [in nature] is 300 bases/second, or about 3-5 ms/nucleotide. If we can implement a system that can incorporate fluidics, that’s the speed [we can obtain]. We don’t need to wait seconds, it can go really fast. It cannot be as fast as the potential of a nanopore, but not far behind that. This is the fastest sequencing technology you can think of – the fastest and basically the simplest. Any other platform you have to do ligation, enzymes, image, dephosphorylation. Each base takes two hours. Here we’re talking subseconds.”
Pourmand did not divulge any details of his lab’s sequencing progress, but said, “I can tell you it is really promising, really great results … This has potential to get to high school students and middle school students and to do biology really inexpensively.”
How much of the Pourmand/Davis technology has been incorporated into the Ion Torrent system is unclear for the time being, as are other scientific and business specifications. (Pourmand, it should be noted, is not currently listed as an advisor to Ion Torrent, but that is likely to change in the coming weeks.)
In 2009, Rothberg and long-time research associate Wolfgang Hinz were awarded a patent (USPTO #20090026082) for measuring analytes using electronic CMOS sensors, in which DNA sequencing was but one application. “The invention contemplates sequencing of nucleic acids based on changes in the [chemical sensor] current,” Rothberg and Hinz wrote, adding that the claimed methods were “not dependent upon the mechanism by which the current change is effected.”
According to the patent, “The invention provides a method for sequencing … a new nucleic acid strand by incorporating one or more known nucleotide triphosphates sequentially at the 3′ end of the sequencing primer, detecting the incorporation of the one or more known nucleotide triphosphates by a change in current.” That incorporation could be measured in many ways, including changes in the concentration of nucleotides or pyrophosphate or pH.
Regardless of the actual specifications, Stanford’s Davis said the development of such a portable, affordable device – a goal of many other third-generation sequencing companies -- could usher in an era of “distributed sequencing… Right now, we have very expensive sequencing. But sometimes you don’t need that capacity. You don’t want to run a bunch of things together, so you wait until you have a collection… I think scaling back the size of the sequencer, having low reagent costs, might be quite disruptive.”
The comparison Davis offered is to the computer industry. If the genomes centers are equivalent to main frames and next-gen sequencers (such as Illumina, 454 and Life Technologies) the mini computer, then the emergence of small sequencing devices, such as Ion Torrent’s perhaps, would be akin to the personal computer.
Davis summed up his ultimate goal: “I want to build a $10 sequencer that can sequence the genome, and do it for maybe $10. That’s what we need. That’s biology.”