PCR is commonly used for forensic DNA typing, but forensic samples contain a number of substances that can confound results if they are co-extracted with DNA. In an effort to gain a better understanding of how these compounds inhibit PCR, a team of researchers led by Bruce McCord of Florida International University recently turned to real-time PCR to gain some insight into the mechanisms of PCR inhibition.
In a paper published in the December issue of the Journal of Forensic Sciences, McCord and colleagues describe how they studied the effect of amplicon length, melting temperature, and sequence on PCR inhibition. They designed primers with three different amplicon lengths and three different melting temperatures to target a single homozygous allele in the HUMTH01 locus, and then measured the effect on amplification efficiency for each primer pair by adding different concentrations of various PCR inhibitors — calcium, humic acid, collagen, melanin, hematin, indigo, and tannic acid.
They found that a surprisingly broad range of inhibition mechanisms can occur during the PCR process depending on the type of inhibitor. These mechanisms include inhibition of the Taq polymerase, DNA template binding, and effects on reaction efficiency.
Some inhibitors were found to affect the reaction in several different ways, the authors noted in the paper. For example, calcium appears to be a Taq inhibitor, while humic acid binds to the DNA template, and collagen exhibits both effects.
The study suggests that analysts can use the melt curve data from real-time PCR to help identify potential inhibitors in forensic samples, which can then help them determine the best method to help remove the substance as well as the optimal analytical approach to take in order to improve their results.
PCR Insider spoke to McCord this week to discuss the study and its findings. An edited transcript of the discussion follows.
What led you to look into these inhibition mechanisms?
As you can probably guess, samples for forensic analysis aren't coming from pristine environments. There can be situations where you're trying to extract DNA from a piece of bone, from some type of fabric sample, or maybe even mixed in with soil. If you're recovering a body that's been in the ground or in a stream, things like humic acid can leach in. There are components in the tissue like collagen and calcium that could co-extract with the DNA.
In the forensic context, we are looking at amplification of a multiplex set of short tandem repeats — it could be as many as 15 or 16 different loci that we're looking to amplify simultaneously. In those situations, sometimes you only get what's called a partial profile: not all of the loci amplify. Sometimes you'll see complete dropout of one or both alleles from a particular chromosome. This is often ascribed to inhibition, but … because it’s a forensic sample, you don't have good information about what was present in the environment where the sample was. It can be very difficult to determine what is causing the lack of amplification when you're doing these big multiplexes.
The forensic community is trying to develop various ways to clean up and purify samples during the extraction process. What we wanted to do is take a look at what kinds of things might be present in an extracted DNA sample and how they would affect the polymerase chain reaction.
So we wanted to look at the mechanism. We figured that instead of trying to develop methods for removal of contaminants, it might be a good idea to try and figure out why certain types of contaminants affect the PCR a priori, and then you'd have a better idea of what types of cleanup might allow you to recover more of these alleles.
How did you set up the experiment described in the paper?
Usually what you find when a sample is inhibited is that the larger-sized alleles tend to drop out, but with some kinds of inhibition you also see that individual loci will drop out in a sort of random fashion, not necessarily based on the size of the allele, or the size of the amplicon.
[ pagebreak ]
So we wanted to explore the effects of sequence, primer melting, and different concentrations of various inhibitors, and we picked inhibitors that others in previous work had reported would cause a problem, for example, collagen, humic acid, tannic acid, calcium, and some other inhibitors.
We then took one particular location on a short tandem repeat and we changed the primer location so that we could track differences in melting temperature, differences in the size of the amplified allele, and, of course, sequence differences — all focused on one small location surrounding an amplified fragment that's commonly used in forensic science, the HUMTH01 locus.
Then we needed a technique to explore, and we used real-time PCR with SYBR Green. We examined the melt curves of the amplified product and we examined the amplification efficiency and the quantity of the amplified product that was produced as we changed the concentration of different inhibitors.
What we found was kind of fascinating. We were able to show that for an inhibitor like calcium, which probably interferes with magnesium, a cofactor for the polymerase enzyme, we saw effects on the efficiency of the amplification and a loss of concentration of product, but no effect on the melt curve of the amplified product.
However, when we looked at things like humic acid, we saw something very different. The [cycle threshold, or Ct] for the amplification, which is directly proportional to the quantity of the template, didn't change when we added calcium, but when we added humic acid we saw a shift in the Ct indicating that less DNA template is available. And if you look at the melt curve for the humic acid and for some of the other things that we looked at, melanin, et cetera, you see very big changes. You see a drop in the melt temperature and you see changes in the melt curve itself, and this indicated to us that some of the inhibitors that we looked at appear to affect the Taq polymerase, while other inhibitors affect, presumably, binding DNA. In other words, they affect the ability of the polymerase to amplify that template and effectively reduce the amount of available template for the polymerase.
So we can use these results, and then correlate what we saw from the amplifications to give us a much better idea about how different types of inhibitors affect the ability to amplify DNA.
It sounds like this variation in different types of inhibition was surprising.
It was. We could see certain inhibitors that would appear to bind DNA. Other ones appeared to affect Taq, and some did both. And we could categorize the types of inhibitors we saw based on the responses we saw in the real-time PCR.
Why do you think no one explored this issue previously if these inhibitors are known to be problematic?
There are a few papers here and there, but nothing as broad as what we looked at. Some people have looked at different types of inhibitors. We have been looking at a way to improve the efficiency of forensic genotyping by reducing the size of the amplicons, and had figured that inhibition should also be improved if you have a smaller amplicon because the amplification should occur a little quicker because it's a shorter read length. But instead we were seeing some of these sequence-specific differences, and so that's kind of why we looked into it.
But most people who are looking at this just want to solve the problem of inhibition. You read paper after paper where people are mainly interested in saying, 'Hey, I've extracted a sample of bone. It had some inhibitors and I did these steps to try and clean it up,' but nobody has really questioned why: Why was it inhibited, and what was the process of inhibition? We knew that maybe it's the Taq, maybe it's binding DNA, but never knew specifically what was going on.
So we just got curious because we needed to know why the shorter amplicons, which we call mini-STRs, worked better and if, in fact, they would affect inhibition. We had developed these shorter amplicons for badly degraded DNA, but it appeared to work for inhibited DNA as well.
So do these findings mean that the shorter amplicons won't work for all inhibitors?
We find that the shorter amplicons do work better, but there are situations where you would have DNA binding where they may work less well. Another thing that we found is that the effects of DNA binding are far more apparent when the amplicon size is larger. So there are two different effects: there's the generic effect of length, and then there's a more sequence-specific effect, but that also may be length dependent because of the fact that the sequence has to be there that the DNA would bind to.
[ pagebreak ]
What does this mean for your work and for others who are doing PCR-based forensics?
There might be situations where you're trying to identify a perpetrator, or at least give some statistical probability about the likelihood that this particular profile matches that of an individual, and you're testifying to this in court. The question can come up of why you are only getting a partial profile. And this is something that needs to be explained. Is this a partial profile because the DNA was badly degraded, or is it a partial profile because the DNA is inhibited?
Generally speaking, if you have less than a full profile, you may want to explain to the jury and provide a reasonable explanation of why a particular set of amplified products did not show up. And this [real-time PCR approach] helps because we can now say, 'There's a particular locus and we see this locus disappearing more often with this particular type of inhibitor, and we know this particular inhibitor, say humic acid, tends to bind DNA.' And we can come up with a reasonable scientific explanation for why.
And furthermore, if the person is amplifying DNA and they see that a particular type of inhibitor is binding, or if they can use the real-time PCR to monitor the melt curve effect, then they might be able to tell what kind of inhibitor it is. They can correlate the effects of the melt curve and the effects of the allele dropout with a particular type of inhibitor and know they have to treat it in a fashion that's going to remove this kind of inhibitor. They can then develop toolkits based on the effect that they see — giving a better idea of what kind of inhibitor is present and how it should be treated.
So you would want to look at the melt curves for the real-time PCR to try to give you an idea of how the inhibitor is affecting the product. It's a two-step process in forensics: First you do the real-time PCR to quantify the amount of DNA, and the second step is to do an additional amplification to look at the different loci, the short tandem repeat loci. The real-time PCR is used for two purposes. First to determine the amount of DNA and second to make sure the DNA that you're amplifying is human.
So then in parallel with that, you can also use the real-time step to gain some insight into potential inhibitors.
Yes, you can glean some information from the melt curve and the shape of the PCR amplification curves.
What's next? Do you plan to further develop the short amplicon approach you mentioned?
We're currently working with different real-time systems to further examine these kinds of effects. We also want to look at other types of amplified products. There are other types of real-time PCR systems that we're starting to look at. Not just SYBR Green analysis, but some of the newer techniques — [Promega's] Plexor, for example, which doesn't use probes. Of course you have to be able to develop a melt curve when you're doing this kind of analysis. We also want to expand the types of inhibitors that we've looked at so far.