Researchers at the University of California, Berkeley have developed a potential alternative to conventional emulsion PCR, generating uniform droplets for bead-based PCR amplification on a microfabricated droplet maker.
The technology, which was describeddescribed in the May 15 issue of Analytical Chemistry, uses a droplet generator that contains a micropump to produce nanoliter droplets of uniform size. These can be used for sequencing template preparation as well as single-cell genetic analysis.
According to its inventors, the technology promises to increase the yield, amplicon length, and uniformity of emulsion PCR, a process used by several second-generation sequencing systems.
Richard Mathies, the senior author on the study and a professor of chemistry at UC Berkeley, told In Sequence last week that three key challenges in emulsion PCR are the ability to amplify uniformly across multiple DNA targets, to obtain high PCR yields, and to generate long amplicons. “By using these engineered microdroplets, where we have precise control over the volume of the emulsion particles, we can transcend all of these problems,” Mathies said.
He and his team built a droplet generator that contains a cross-injector. At its injector’s junction, two channels are filled with carrier oil, and a third one is filled with a mix containing DNA template, forward primers, bead-coupled reverse primers, and PCR reagents. Driven by an on-chip polydimethylsiloxane membrane pump, about six 2.5-nanoliter droplets of the mix per second form at the junction and are transported through the fourth channel to a tube that collects 6,600 droplets.
According to the study, the system can make droplets ranging from 2 to 5 nanoliters in size.
“The hard part is getting the size you want and making them all the same size,” Mathies said. The PDMS membrane pump allows the researchers to control the size precisely, he explained.
Others also fabricate emulsion droplets for PCR and other applications, notably RainDance Technologies, which won a grant from the National Cancer Institute last year to develop its NanoReactor technology for multiplexed exon amplification (see In Sequence 7/17/2007).
According to its website, RainDance works with picoliter rather than nanoliter droplets, of which it generates “thousands” per second. According to Mathies, RainDance does not employ on-board pumps to control the breakaway of droplets but releases them passively. Picoliter-size droplets “would not be useful for our purposes,” he said. “We need nanoliter-volume droplets in order to produce the kind of products that we want,” such as kilobase-sized PCR products.
“By using these engineered microdroplets, where we have precise control over the volume of the emulsion particles, we can transcend all of these problems.”
In their study, Mathies and his team amplified a single 624-base pair DNA template in a droplet, achieving a yield of between 50 and 100 atomoles. They also successfully analyzed this product by Sanger sequencing, obtaining more than 500 bases of high-quality sequence.
Will Pyrosequencing Profit?
The technology could have applications both for next-generation Sanger sequencing as well as for pyrosequencing, the chemistry used by 454 Life Sciences, according to Mathies. He is working with Microchip Biotechnologies, a Dublin, Calif.-based startup company he co-founded, on commercializing the technology for sequencing applications.
Specifically, MBI is working on an integrated Sanger-based micro DNA sequencer (see In Sequence’s sister publication, GenomeWeb Daily News 4/24/2006) that could incorporate the technology.
In addition, in collaboration with Mostafa Ronaghi at the Stanford University Genome Technology Center, MBI is developing an automated sample-prep system for high-throughput pyrosequencing that makes use of the nanodroplet method (see In Sequence 3/4/2008).
Further, MBI and Ronaghi have a grant application under review at the National Institutes of Health to develop a “superscalar pyrosequencer,” MBI Chief Commercial Officer Barney Saunders told In Sequence this week, and the nanodroplet technology “will be valuable in this work at MBI.”
According to Mathies, the technology could help to increase the fraction of long sequence reads in pyrosequencing, including those of 454 Life Sciences.
Both 454 and Applied Biosystems currently employ emulsion PCR to prepare DNA for their respective second-generation sequencing systems, using what Mathies called a “shake-and-bake” approach to generate the emulsion. This method generates droplets that vary in size, he said, and small ones do not yield sufficient PCR product.
Although some droplets do have the right size to give good results, “you are always fighting the polydispersity,” Mathies said. “With the kinds of droplets that we make, of course, every droplet is a good one,” he added, and the method “can routinely produce kilobase-sized DNA at yields that are orders of magnitude higher than those you can get from the ‘shake-and-bake’ approaches.”
Conventional emulsion PCR technology has also not prevented 454 from extending its read length to more than 400 bases. The company plans to launch these extra-long or XLR reads as part of a technology upgrade later this year.
The upgrade will also involve “significant improvements to the [emulsion] PCR process,” Michael Egholm, 454’s vice president of R&D, told In Sequence this week. “The users will experience a more streamlined process with significant reduction in the hands-on time as well as improved robustness of the process,” he said, but did not provide any details.
“The Mathies lab’s work is elegant and it is certainly possible that controlled droplet formation could improve our [emulsion] PCR process further – we’re not done improving our system yet,” he added.
The XLR reads have a Q20 length of at least 400 bases. Though a large fraction is approximately 500 bases long, due to the fact that 454’s chemistry produces a distribution of reads, the average read length is 400 bases.
It is unclear whether the Berkeley researchers’ nanodroplet method could increase the average read length, but Mathies said his lab will be the first to test this. “As soon as 454 releases [its upgrade], we have got beads ready to go, and we intend to run them on 454, too,” he said.
Applied Biosystems also recently made improvements to its emulsion PCR process. The company said in February that it has replaced its emulsion maker with a smaller device and has optimized its emulsion PCR protocol (see In Sequence 2/19/2008). The new device can generate 10 billion droplets within minutes, which produce 200 to 300 million PCR products. DNA is also more uniformly amplified than before and beads are loaded with more template, according to the company.
Mathies’ lab is currently working on increasing the throughput of its droplet generator by using hundreds of nozzles in parallel instead of just one, enabling his team to produce millions of droplets within hours.
That will make the throughput of high-quality beads containing PCR products “at least similar” to the shake-and-bake approaches, he said. “The question is, what fraction of the beads you produce actually yield good sequence? And in our case, it’s [close to] 100 percent.”