In PCR, as in most things, digital is an improvement on analog — it's faster, more efficient, and able to perform more complex tasks. Unlike other digital technologies, digital PCR hasn't replaced analog PCR completely — but it does take over the more complex and intricate tasks.
In 1998, Dan Tawfik and Andrew Griffiths published a paper in Nature Biotechnology on a new technique they called emulsion PCR. Since then, researchers have used the technique for a variety of purposes — refining and adjusting it, developing and publishing their own protocols, and even helping to develop some of the most ubiquitous sequencing equipment found in the lab.
Emulsion PCR, a type of digital PCR, was first developed to study molecular evolution — it amplifies individual molecules by dividing them into bubbles or nano-compartments in a water and oil emulsion, that act as single chambers. "It's like having a billion little test tubes," says Johns Hopkins Sidney Kimmel Comprehensive Cancer Center's Bert Vogelstein.
Turning down the noise
Emulsion PCR was the basis for a technique Vogelstein and his colleagues created called BEAMing, which they developed to look at the sequence of individual molecules of DNA, Vogelstein says. He and his team published the technique in Nature Methods in 2006, and that technology is now a core component of two commercially available next-generation sequencing instruments: Life Sciences' 454 and Applied Biosystems' SOLiD.
Using this technique, researchers have been able to detect rare mutations, as BEAMing is sensitive enough to find one mutation in 10,000 DNA molecules. "Emulsion PCR as it was originally formulated was not a technique in which you could recover individual bubbles," Vogelstein says. "They're basically in the emulsion and you lyse the emulsion. So what we wanted to do is what can be called 'digital PCR,' in that you examine molecules one by one to see if there's a rare variant among them."
Improving the signal-to-noise ratio meant researchers were that much more likely to detect a rare variant instead of overlooking it. The classic way to detect and quantify the number of variants within a population of DNA molecules, or the "analog way" as Vogelstein terms it, calls for taking all the molecules being studied, mixing them together and then looking for abnormalities. "Say all the molecules are red, but there's a rare one that's green. You try to detect that faint green color amongst the mountain of red, and the signal to noise often makes that difficult, or limits the sensitivity," Vogelstein says. "You can imagine if you have a little hill next to a big mountain, it can be hard to see the little hill because the mountain conceals it."
With digital PCR, a researcher can look at each DNA molecule one at a time to determine whether it's a mutant or a wild-type gene. "It either has the variation or not," Vogelstein adds. "It's digital — one or zero — and the signal to noise is, in principle, huge because you have no contaminating molecules to conceal the signal that you'd like to detect." When they first started using digital PCR 10 years ago, with a 384-well plate, Vogelstein and his team were able to amplify molecules and, looking at them in a fluorometric assay, could see 100 or 200 different molecules at once.
But for many applications, like finding rare variants, it's necessary to see one in every 1,000 or 10,000 molecules. "It's at that point that we invented BEAMing," Vogelstein says. "We used the emulsion PCR basic formula that was invented by Tawfik and Griffiths, but in a way that we could recover the DNA from each bubble." In order to do that, Vogelstein included a magnetic bead in each bubble, so the amplification was working on a solid phase rather than in the aqueous nano-compartment. The researcher ends up with one molecular template per 10 bubbles, he says, and can convert those individual template molecules into individual beads — if you start with 10,000 template DNA molecules, you end up with 10,000 beads. "Each bead is now bound to 100,000 to a million identical copies of that initial template molecule, and then you lyse the emulsion, collect the beads on a magnet, or centrifuge them, and then you have all this DNA, which is really so many copies of one molecule," Vogelstein says. "It becomes easy to sequence it, and that's what the SOLiD platform does, and it's also what 454 does, and it's what we do — we use the beads and just put them on a flow cytometer after hybridizing them to detect either the mutant or the wild type."
Variations on a theme
Emulsion PCR may be a standard when it comes to finding mutations in DNA molecules, but in the years since it's been invented, many researchers have developed their own variations on the original in order to facilitate certain types of research or refine the protocol.
At the University of New Mexico Health Sciences Center, Jeremy Edwards has been using emulsion PCR for the last 10 years, and says it's definitely made his work much easier. Edwards' lab uses emulsion PCR to do single-molecule assays in a high-throughput manner, and to do many assays simultaneously. His team then uses flow cytometry to interrogate the droplets, but they've also arranged them on slides and used standard microscopy to observe the reactions.
Edwards and his team have also developed their own technique, called dual-primer emulsion PCR, which they published in BioTechniques in May 2010. This protocol is a "-highly sensitive single--molecule clonal amplification method" similar to standard emulsion PCR, with the difference being that both primers are attached to the beads, the team writes in its paper.
"The idea with dual-primer PCR was to develop a simple DNA sequencing strategy where you would have both strands of the DNA present on the bead, so theoretically you could sequence both ends of the DNA," Edwards says. "Traditionally, only one strand would be present on the bead, but getting that other end is difficult. With a dual-primer approach, both strands of DNA attach to the bead, and you would potentially be able to run two separate, distinguished sequencing reactions off of the DNA immobilized on the bead that you used to isolate the DNA out of the drop. You could double your sequencing as well as your information because sequencing both ends of a fragment of DNA is very valuable for reassembly of a genome."
Edwards hoped to overcome the limitations of short reads in next-generation DNA sequencing. "We're using it to do single transcript molecule analysis, to analyze individual transcripts in the full context of the transcript. So rather than knowing the expression level of a gene, you can subdivide a gene up and learn about what exons are being expressed at different levels," he says. Edwards and his team developed this assay with the intent of finding out how exons are spliced together in individual transcripts. They used emulsion PCR to amplify the full transcripts and then interrogated which exons were present.
At the Mount Sinai School of Medicine in New York City, James Wetmur has developed a protocol recently published in Methods in Molecular Biology, which he calls "linking emulsion PCR," and which aids in the measurement of haplotypes. Linking emulsion PCR uses emulsion PCR to isolate single template molecules for "simultaneous PCR of widely spaced markers and uses linking PCR to fuse these amplicons into one short amplicon, which maintains the phase of the markers," Wetmur et al. write. "LE-PCR is illustrated for polymorphisms in human paraoxonase 1 that have been shown to affect transcriptional activity and substrate specificity in the detoxification of organophosphates."
At the Max Planck Institute for Molecular Genetics in Berlin, Jörn Glökler and grad student Tatjana Schütze have been using emulsion PCR in their work for about a year and a half. Doing experiments with aptamers — oligonucleic acids or peptide molecules that bind to target molecules — the researchers start off with a library of 1015 molecules, a "mind-boggling" amount, Glökler says. In order to sift through them for binding, you need to do 10 to 15 selection rounds with an amplification step in between each round of about 15 to 20 PCR cycles. "Stacked up on each other, that means more than 200 PCR cycles. By doing so, often you get a lot of junk, and the junk is more than the signal, so the signal-to-noise ratio is not advantageous," Glökler says. "We thought those with a high efficiency should be kept at bay by using emulsion because you do these emulsions and you can to 10 to the ninth compartments in one milliliter, and everything is kept in one receptacle and can only exploit the energy that is found there."
In order to keep the noise down and do experiments they could not do otherwise, Glökler and Schütze turned to emulsion PCR. "Peptic nucleic acids are like antibodies or antigens, and [emulsion PCR is] a way of using large libraries of different clones and finding those that are specific to a given ligand," Glökler says.
Working along with their colleagues at Max Planck, they have developed a streamlined protocol for emulsion PCR, recently published in Analytical Biochemistry. The subsequent purification of DNA doesn't require any specialized equipment, and can be implemented in a broad range of applications. Starting with an aqueous phase — water with the PCR mix — and another phase that is a combination of three or four components — usually detergents and mineral oil — the two phases are combined to form two layers, and then shaken vigorously in a vortexer to combine them into a turbid emulsion that "looks like milk," Glökler says. After the emulsion is put through PCR, it can be broken up with isobutanol to separate the layers, and a standard PCR clean-up kit can be used to get ahold of the DNA. "We use our new protocol to amplify our selection rounds [of aptamers from the library]," he adds.
The disadvantage to the protocol is that it requires an extra step and it also increases the volume of the aqueous phase. "Let's say you have an aqueous phase of 15 microliters and you add 300 microliters of oil phase to it, then you have a larger volume, and PCR machines typically hold 100 microliters at the max," Glökler says. The sample must also be purified before it can be analyzed.
As with all sequencing methods, there are advantages and dis-advantages to emulsion PCR and its variations. Whether you use beads, dual primers, or BEAMing, each method has its pros and cons.
Emulsion PCR has the advantage of keeping copy errors at bay when the DNA is copied. "With normal PCR you have the bulk solution volume of about 100 microliters. If you perform a PCR of many cycles, you get copies on copies on copies, and the more you do that, the more errors you get. And these errors stack on top of each other," Glökler says. "The problem is if you have efficiency in the PCR of a factor of two, that's fine, but quite often, you have differences. Some templates have a different efficiency so that means if you have an efficiency of one plate of 1.9 and efficiency of another plate with 1.99, this doesn't seem to be very much, but the more cycles you do, the more the population with the higher efficiency will supersede the other one. If you keep them separate, so that right from the beginning you have separate receptacles or bubbles, then you would confine them, and they would all reach the same endpoint, so that way you can keep them at bay from the artifacts you can get by PCR."
The choice of whether to use emulsion PCR or one of the variations depends on what kind of application you want to use it for, Schütze says. You wouldn't want to use it for anything beyond 15 or 20 cycles of normal PCR, because if you have a single target instead of multiple complex targets, then it's not necessary to go through the trouble of using the technique. Glökler adds that "once you handle libraries with different lengths, there is automatically a competition between these and you get a shift in balance of the distribution." In this way, emulsion PCR is not a replacement for other PCR techniques, but a complement to them.
Hopkins' Vogelstein says that other technologies like qPCR or mass spectrometry can be used to quantify variants, and they work reasonably well when the fraction of variants that you're trying to detect is greater than 5 percent or 10 percent. "But if you go down lower than that, to 1 percent, or 0.1 percent, or 0.01 percent, which is what you often need to do for clinical applications, then they're not sensitive enough," he says. "They are analog-based techniques rather than digital, and the advantage of digital is simply the signal-to-noise ratio." While other techniques take less time to prepare, if you're looking for something uncommon or rare, analog techniques will have a hard time detecting those rare variants, and wouldn't be very useful, he adds.
There are a billion nano-compartments in a test tube and 100 billion nano-compartments in a 96-well plate, and the sequencing of all those molecules in done simultaneously in emulsion PCR, so it only takes an hour or two to look at all of them. "When people talk about massively parallel sequencing, that's what they're talking about because you're doing all these amplifications, not sequentially, but in a parallel fashion in a test tube," Vogelstein says. If you're using beads, he adds, the advantage is that you can recover individual beads for downstream applications, and you can get many more copies of DNA on a bead than on a flat substrate.
"When people put beads in there, the advantage is that you can isolate the bead from the emulsion and the beads can all be analyzed in a high throughput way, like DNA sequencing, for example," New Mexico's Edwards says. "The people that are not using that approach to get the beads out have the disadvantage that they have to build all the approaches for assaying these molecules into the droplets because they're hard to recover."
When using single molecule assays in emulsion, the challenge is in recovering material for the assay, Edwards adds. "Even if you're using beads, it can be quite laborious to go get all those beads out of the emulsion because they're covered in all kinds of oil and getting them clean and usable is not a trivial process," he says.
Whatever the intended use or protocol, most researchers who use emulsion PCR agree that it is the digital PCR standard, and when looking for rare variants, it is the right method for the job.