June 15, 2012 (Vol. 32, No. 12)
Benjamin Bronfin, M.D. Cambridge Biomedical
Rohit Kumar Mahajan, Ph.D. Vical
Chris Meda Response Genetics
Taka Murakami University of British Columbia
Vyacheslav Palchevskiy, Ph.D. UCLA
An Exclusive Q&A with Our Expert Panel
Once the concept of PCR popped into Kary Mullis’ head in 1983 it wasn’t long before scientists realized they had a revolutionary technology in their hands. Because now they could take the smallest slice of DNA and amplify it an infinite number of times for analysis and further research. Indeed, PCR so transformed the life science field that Mullis was awarded the Nobel Prize in Chemistry in 1993.
Many researchers regard PCR as a workhorse in molecular biology. Novel applications, new reporter chemistries, and advancing microfluidics are already leading to higher throughput and more sensitive PCR techniques.
Mikael Kubista, Ph.D., one of GEN’s editorial advisory board members, helped pioneer the development of real-time qPCR. In the May 1 issue of GEN Dr. Kubista noted that qPCR is developing rapidly, although most current efforts are on pre-analytic and data-mining processes rather than on the qPCR technique itself.
In this special GEN Tech Tips PCR feature, we asked some of the world’s other leading practitioners of PCR methods to talk about ways to improve and get more out of their work with various PCR techniques. The people interviewed for this special feature included Benjamin Bronfin, M.D., director of R&D at Cambridge Biomedical; Rohit Kumar Mahajan, Ph.D., associate director, assay development at Vical; Chris Meda, consultant and former president at Response Genetics; Taka Murakami, manager of the genotyping facility at the University of British Columbia; Vyacheslav Palchevskiy, Ph.D., postdoctoral fellow at UCLA, pulmonary and critical care division; and Kevin Papenfuss, scientific consultant specializing in PCR, formerly of Life Technologies. You will find that their insights also have applications to your research.
What kinds of PCR techniques are you most familiar with?
Dr. Bronfin: We perform real-time PCR (qPCR) and reverse transcription (RT)-qPCR for clinical diagnostic and clinical trials. Our lab performs both CAP/CLIA and GLP types of tests. Sample types include serum, whole blood, stool, saliva, urine, and cell lysate. Average number of PCR runs per month is about 10 with 3–4 for clinical diagnostics and 5–7 to support clinical trials.
Dr. Mahajan: As a plasmid DNA (pDNA)-based company, we use PCR-related techniques for product-specific applications such as monitoring cleaning of our manufacturing equipment, performing preclinical safety studies to quantify the risk of product integration into the genome of injected tissue, and monitoring product expression potency for lot release and stability.
We use quantitative PCR to target either product-specific sequences or pDNA backbone sequences. Quantitation of the target sequence is achieved through the use of standard curves made from 6–7 log dilutions of product pDNA. For drug product potency assays, we have developed a one-step RT-PCR assay that measures levels of mRNA expressed in transfected tissue culture cells.
In this assay, pDNA product lots are transfected into cells in parallel with a specific product reference; product potency is measured relative to the reference lot. All our routine assays are real-time PCR assays using TaqMan® probes.
Ms. Meda: Response Genetics works mostly with paraffin-embedded tumor tissue that we microdissected and subjected to real-time PCR, which has pretty much become the platform of choice due to its speed and specificity.
Mr. Murakami: Our PCR-based genotyping assays involve endpoint and melting curve analysis, copy-number assay, allelic discrimination, SNP analysis, and PCR-based pathogen detection. We work with all types of tissues and cells, including from animal disease models. Our associates average 4,500 PCR reactions each per month.
Dr. Palchevskiy: Our laboratory performs real-time PCR and genotyping. Since we are primarily working with animal models, genotyping of mouse tails or ears is crucial to identify knockout strains. For mRNA gene expression we perform qPCR, using primarily Applied Biosystems’ gene-expression assays with the assumption that they perform at 100% efficiency.
The type of samples that we are using are mouse tracheas that were heterotopically transplanted subcutaneously to the back of a different strain recipient (usually a knockout strain). We’re also interested in bronchial and axial lymph nodes in mice.
Mr. Papenfuss: I have developed applications and instruments for use with TaqMan and SYBR Green qPCR, SNP genotyping, and TaqMan Digital PCR. Most of my experience is in ultra-low volume, mid-density platforms. Being in a product-development environment, we generally worked with a small number of high-quality cDNA or genomic DNA samples.
Do you have any special ideas or techniques for improving PCR reproducibility?
Dr. Bronfin: One of the keys for PCR reproducibility is DNA/RNA extraction. We usually use column-based extraction kits that show very good efficiency and repeatability. The hot-start approach can also improve specificity and reproducibility for some PCR reactions.
It’s important to have a correct concentration and ratio for your primers and probe, because low reproducibility may be a sign of low primer concentration. Some of the smaller steps like centrifuging the plate before putting it into PCR machine may have a very positive effect as well.
Dr. Mahajan: We have also developed qPCR applications to quantify intact pDNA or large fragments of pDNA in mixed-population samples. These assays amplify target regions that are much larger than those used in typical real-time applications. This requires additional optimization of buffer components and cycling parameters to ensure efficient PCR amplification.
Ms. Meda: The expression “garbage in, garbage out” is highly applicable to PCR work. In our case much of our success resulted from skilled sample preparation. If you don’t have a good sample it doesn’t matter how exquisite your instrument or technique may be. So we focused on helping pathologists improve their sampling technique, to get tumor material into paraffin quickly, and to fix the sample properly.
Mr. Murakami: The quantity of DNA template is one of the most significant factors for PCR reproducibility. Some ideas for avoiding reproducibility errors are to use color-changeable master mix after DNA addition. This lets us know if DNA was added or missed.
We also employ the PCR instrument as an additional check to determine whether DNA was added or not. We do this by measuring the anion concentration of template by UV.
Dr. Palchevskiy: Reproducibility, especially for qPCR, depends on many factors but most important are quality of RNA, efficiency of RT reaction, and the efficiency in primer amplification. Therefore, stable and consistent protocol for acquiring and process of samples is instrumental to produce good quality RNA that can be used for downstream applications.
Quality of RNA can be checked by bioanalyzer or agarose gel. I prefer the former since it requires less sample, but it can be more expensive with large numbers of samples. Also significant are the quality of primer pairs and probes. Running samples in triplicate provides a good control for any pipetting errors. Also, it is a good idea to have a positive control(s), previously amplified sample(s) that you know is going to work and should provide you with some point of reference.
Mr. Papenfuss: Consistency and quality of the sample preparation are of huge importance. While it may be cheaper to use a home-brew sample-prep method, the extra expense of using a commercial kit is absolutely worth it. Also, depending on the PCR platform being used, the default cycling protocol may not be optimal for your samples, potentially leading to inconsistent results. If the protocol is adjustable, it’s possible to improve reproducibility in that manner.
The actual setup of the PCR reaction should not be overlooked either. Having well thought out, established lab practices and scientists who are comfortable implementing them logically leads to reactions that are set up consistently from experiment to experiment, which should help improve reproducibility.
Describe ways to improve PCR doubling efficiency.
Dr. Bronfin: I think doubling efficiency depends a lot on primer/probe specificity. Ideally melting temperature of reverse and forward primers should be within 5°C of each other. I know that some people suggest that adding additional MgCl2 may increase robustness, but I personally never saw any benefit from it.
Increasing extension time to 2–3 minutes may also help. Very often diluting your DNA prior to reaction in RNase/DNase free water can yield much higher efficiency, especially if your sample type is stool or cell lysate.
Dr. Mahajan: All our assays are optimized during the development phase to ensure reproducibility and maximum efficiency of PCR (i.e., doubling efficiency). We typically evaluate critical parameters such as primer and probe concentrations and reaction buffer conditions.
We use a standard optimization approach of running matrices of varying concentrations of reagents to find the optimal ratio that delivers maximal normalized fluorescence and minimal cycle threshold. We then test our optimized conditions on multiple standard curve preparations to ensure reproducibility.
Mr. Murakami: We have found that choice of master mix for the first cycling and including the correct Taq polymerase for the particular platform improves efficiency and productivity. To get earlier Ct value depends on the master mix. Multiple PCR is one of several ways to improve productivity. High-throughput instruments are another way to increase efficiency and throughput.
Dr. Palchevskiy: Quality RNA will provide good cDNA and consistent efficient amplification. Properly diluted sample is necessary as you do not want to exhaust your reagents or primers that are present in excess, except for house-keeping genes for multiplex qPCR reactions. A good primer/probe set that is highly sequence-specific to the gene (exon or exon-to-exon junction) of interest is helpful.
Another technique we employ is to re-purify the sample, if necessary, and perform a DNase digestion to prevent any source of DNA contamination. For SYBR, specificity of primers is absolutely essential, as well as checking the melting curve at the end of each run.
Mr. Papenfuss: Problems with PCR reaction efficiency can be attacked from several angles, and some or all of these may need to be explored. The quality of the primers and probes (if using a probe-based method) will affect everything. Be sure to use an established primer/probe design software package or website. Depending on the target, you may not be able to change your primers/probes much, if at all, but use the best ones you can.
Next, the master mix can affect reaction efficiency. If your PCR platform requires a specific master mix, by all means use it. If not, experiment with different mixes as they will each interact with your samples and reaction components differently. It is also possible to use any of a myriad of additives in the master mix, especially when dealing with difficult samples.
Lastly, the PCR cycling protocol directly affects reaction efficiency. Most systems are set up for the average sample, but if you have nonstandard samples, the average PCR cycling protocol may not be appropriate. Some systems are not adjustable, but if you are using a system that is, making changes to the temperature or duration of the various reaction steps can improve efficiency.
Given PCR’s sensitivity, how do you deal with potential contamination issues?
Dr. Bronfin: We perform all our work in biosafety cabinets and we have an SOP in place for cleaning them. We have separate suits for handling samples, preparing master mix, and making controls and standards. As a rule, during extraction and setting the plate only one tube is open at any given time. We always prepare samples in the following order: negative control and/or blank, samples, positive control. Gloves are changed after each step and if contamination is suspected.
Dr. Mahajan: Since we share the building with our pDNA manufacturing facility, potential contamination issues are a particular challenge. This is especially true with our limits assays such as integration or cleaning verification where the expected result is a low or negative PCR result.
In addition to the standard negative control samples in each run, our primary method of eliminating potential contamination issues is through strict adherence to procedural controls (i.e., dedicated and isolated work spaces, rigorous cleaning of work spaces before and after reaction set up, barrier tip pipette tips, etc.). These practices have served us well so far, and we have not had any significant contamination issues.
Ms. Meda: PCR has changed a lot in 20 years. You no longer need three separate rooms to carry out PCR, and reagents have improved tremendously. Even today, you can’t expect PCR to work well in a contaminated environment. However, reagents have improved tremendously, and the method has become more robust. The advent of Roche AmpliPrep, and many of the TaqMan real-time PCR reagents have improved the technology tremendously as well.
Mr. Murakami: There are two possible sources of contamination in our work: the DNA template or reagents. Carefully collected tissue samples should not contain genetic contamination, while premixed reagents containing buffer, MgCl2, and Taq polymerase minimize contamination. It becomes essential to improve handling and processing through less pipetting, particularly for the master mix preparation. Automated dispensing is just one option we’re using for sample preparation.
Dr. Palchevskiy: Since many things can inhibit qPCR reaction, it is necessary to dilute your cDNA several fold and use it as a standard curve to check efficiency. If any contaminants are present, the curve and the Ct vs. concentration would not be what you expect it to be, and it will point to the potential source and solutions. Diluting out the cDNA sample is one way to minimize any inhibition effects that might be introduced by RT reaction.
Mr. Papenfuss: Contamination should rarely be an issue if you follow good laboratory practices. Often, filtered pipette tips and “good lab hands” are sufficient, but that’s obviously not the case in all labs with all operators. Filtered tips are an absolute must. A good first practice is to use a commercially available surface decontaminant, designed to remove DNA, on surfaces and pipettors before use.
Wear gloves—not to protect yourself from the reaction, but to protect the reaction from the operator. Obviously, if any solutions containing DNA get on the gloves, they should immediately be changed before spreading any contamination. While setting up the reactions, be aware of where you are moving the pipettors. For example, a pipette tip full of DNA should never be passed directly over a negative reaction or other unknowns.
In some circumstances, it makes sense to have separate areas in the lab for addition of unknown samples and positive controls, lowering the risk of contaminating the area where clean master mixes and negative reactions are prepared. Once finished, decontaminate again.
Suggest ways to obtain acceptable results from samples of suspect quality.
Dr. Bronfin: That’s a tricky one. There are some steps you may take to make it better, but at some point if your sample is poor quality nothing will work. If you start with RNA, extraction is important to keep your samples at 4°C during all the process and use a temperature-controlled centrifuge. Nested PCR protocol and increasing the number of cycles may also help.
Dr. Mahajan: Every now and again we receive samples for testing that are either in an uncharacterized matrix or in a matrix known to affect the efficiency of PCR. Our primary solution has been to dilute the sample to a level where we have diluted out any inhibitors while retaining sufficient target sequences to measure in the quantitative range of the assay.
We verify the acceptability of the result by running an aliquot of the diluted test sample spiked with a positive control pDNA. We compare the result of the spiked sample with the result of a sample of water spiked with the same positive control pDNA.
Ms. Meda: With solid tissue samples, the best way to assure results is by training those who take and prepare samples. This is also true for blood or other types of samples. No matter how wonderful your detection system, a lot of things can go wrong during sample collection and preparation in order to obtain a satisfactory answer.
Mr. Murakami: Sequencing of PCR products from gel electrophoresis helps, as does melting curve analysis with SYBR Green with qPCR.
Dr. Palchevskiy: It is difficult to trust any samples containing low-quality RNA. It is easier for DNA samples since for standard PCR you are looking preferably for the presence or absence of the gene or DNA region. But for RNA, because it is message-level expression, low quality might degrade and compromise measuring the exact expression levels.
If suspect quality means inhibition products from either purification or RT reaction, it is always preferable to re-purify the sample either by column (with DNase digestion for RNA) or precipitation. If suspect quality means degradation (more concern for RNA than for DNA samples) then I would say change the protocol either of sample collection or storage.
I spoke with Core facilities at UCLA and they mentioned that there are specific kits for microarray based on degraded samples. However, you need to analyze the data with extreme caution.
Mr. Papenfuss: The easiest, most straightforward approach is a commercially available PCR cleanup column. It may be worth trying a few to see which works best with your sample types and PCR system. If cleaning up your sample is not possible, perhaps structure the experiment so a yes or no answer is sufficient instead of going after actual Ct values.
Do you employ concentration or other preparative methods for dilute samples?
Dr. Bronfin: During the DNA/RNA extraction step, for some low-concentration samples, we use less extraction volume compared to the original sample volume. It’s important to remember that although lower extraction volume may give you a higher concentration, it usually yields less total DNA. We don’t use any concentration column or other similar methods.
Dr. Palchevskiy: Our laboratory prefers not to concentrate samples as any such procedure results in more manipulation of RNA and even greater risk of degradation. However, if necessary and if there are enough samples, we might use lithium chloride or ethanol precipitation (mostly for DNA).
Another alternative if the sample is dilute and not pure, we might employ re-purification of the sample and re-suspension in the smaller volume. It is our practice to re-suspend samples in the smaller (minimum) volume (usually 20 µL) since it is easier to dilute than to concentrate sample.
Mr. Papenfuss: The most basic concentration method, simply evaporating off some of the water in the sample, works for very clean samples but may lead to problems with some real-world samples, due to the fact that it also concentrates any PCR inhibitors/contaminants present.
A PCR cleanup column is one of the best solutions. The sample gets cleaned up, which is a bonus, and then you elute the sample into water or buffer, concentrating it as much as needed. Another possibility, depending on experimental goals, is to look into digital PCR, which allows for excellent quantification of fairly dilute samples.