Fourth Annual DNA Grantees' Workshop

Tuesday, June 24, 2003

AFTERNOON SESSION

Single Nucleotide Polymorphism Detection in Highly Degraded DNA
George F. Sensabaugh, Jr.
Biography

DR. VOSBURGH: Our next speaker is George Sensabaugh. One thing he was telling me is that he has moved mostly into the area of microbiology. He will be talking about SNPs (single nucleotide polymorphisms) and degradation of DNA. In any event, I did notice that our two speakers had a connection through the University of Strath-Clyde in Glasgow.

DR. SENSABAUGH: It's hard to follow Jack, who has done a lot of very nice work. It's work that I am quite envious of because we did some ultraviolet light work together back in the late 1980s. However, we didn't get nearly as far along as he has now, and I now wish that we had. Sensabaugh: Slide 1

Today, I would like to talk about single nucleotide polymorphism (SNP) detection in highly degraded DNA. I want to thank NIJ for funding this project and also for their patience. Last year at this time, we had just got a budget line, and it took another 6 to 8 months to get a visa through INS (Immigration and Naturalization Service), so really we've only been working on this since the end of February. I also want to thank Dr. Vera Simir, my postdoc working on this project. All the work or most of the work that you're going to see is hers.

So here is the problem. Some forensic samples contain DNA that's too highly degraded for routine STR (short tandem repeat) profiling, and we've seen a number of presentations that have mentioned this. Highly degraded DNA often comes from mass disasters, missing persons, wartime missing persons, war crime victims, and old crime samples. Actually, the motivation for the project in the beginning was missing persons and war crime victims. This project was proposed 8 to 10 months prior to 9–11, but it obviously has become more relevant since then. Sensabaugh: Slide 2

In the case of highly degraded DNA, the direct amplification can result in artifacts due to template jumping, although there is some argument as to just how much template jumping actually occurs. The environmental context in which degraded DNA is found is probably more important because it often contains contaminants, such as polymerase inhibitors, that interfere with polymerase chain reaction (PCR).

The objective of the project is to develop an assay that allows genetic typing of highly degraded DNA samples (i.e., DNA fragments smaller than 50 to 100 base pairs in length), avoids amplification artifacts, is insensitive to contaminating interferences, and uses a standard platform for detection—something that is readily available in most laboratories, like a 310 or a 3100. Sensabaugh: Slide 3

Here is the design approach that we took: Sensabaugh: Slide 4

• It's difficult to get too many STRs that fall in the range of 50 to 100 base pairs, but single nucleotide polymorphisms do fit the bill nicely.

• To avoid amplification artifacts, we would like to generate a signal from the target and then amplify that signal.

• To avoid contaminating interferences, we want to use a solid state capture system to facilitate washing away of interferences and other sorts of problems.

• Of course, then, we want to use a standard detection format with size-based markers and fluorescent labels.

In the proposed analytical scheme, we would have a capture probe that is immobilized on a solid support. The capture probe also contains a primer region as well as a region that is complementary to or in sequence with the target DNA. This is the capture probe. Sensabaugh: Slide 5

We then hybridize a strand of DNA to the capture probe and add a detection or reporter probe, which, again, contains a complementary sequence, plus another primer sequence. After that, we ligate. At this point in the scheme, ligation will occur where the reporter probe has a complement to the polymorphic base in the target DNA. Where it is not a complement, no ligation will occur.

Next, we wash away the unligated material and we're left with a full-length piece of ligated DNA that is attached to a solid support. We can then use the primers to amplify that piece of DNA and can detect it according to standard formats. Sensabaugh: Slide 6

A number of variables come into play when developing a model test system, including the length of the probe sequences, both capture and reporter probes, the concentrations of the various elements, the binding and dissociation rate constants under various conditions for the immobilized part, the surface, and the nature of the surface attachment.

There are a couple of design features that we looked at, including the size of the whole hybridization sensor region and issues with sample transport (i.e., how efficiently can we get the sample onto the bound capture probe and how efficiently can we get the reporter probe system into the complex). A lot of it depends on the type of system (e.g., diffusion or convection based) and the distance that the various molecules have to travel in order to make contact for hybridization. Sensabaugh: Slides 7 and 8

We wanted to develop a model test system with these features in mind and to compare the solid state chemistries, investigate the order of steps, and determine the amplification parameters. Then we wanted to investigate the test system for markers that had flanking sequences with different compositions and complexities, determine the effects of reducing target size on either side of the site for sensitivity and specificity, determine whether the SNPs can be detected on both strands simultaneously, and, if we had time, try to expand the detection system using rolling circle amplification for detection rather than using PCR. Rolling circle amplification is an isothermal amplification system and potentially more sensitive than PCR.

Here is the system that we have started to work with. Just to point out a couple of features: The capture site is 40 nucleotides long. Similarly, the reporter site is 40 nucleotides long. The G at the end of the reporter site or the A at the end of the alternative reporter site are the polymorphic nucleotides that we're trying to detect in our SNP. We're using an M13 primer site on both sides because those are nice commercially available primers and we don't have to concern ourselves with them. Sensabaugh: Slide 9

The two reporters are differentiated by a 10-base insert, such that after ligation, we have the ligation products and one PCR product off of the ligation product will be 116 base pairs and the other one will be 126. So that is something that's easily detectable using standard size-based detection systems. Sensabaugh: Slide 10

So we started out with a homologous liquid phase system to investigate some of the initial parameters, and this is the result of an early experiment. This is now in a homologous liquid phase system, as we're not using solid phase for this. It's throwing the reagents into the test solution and seeing what happens. Sensabaugh: Slide 11

Note that where we have used the G reporter, we're getting a band in the right size range. Where we use the A reporter, we're getting a band that's about 10 nucleotides longer or 10 base pairs longer. That's all right. The only problem is that we have a specific result here; that is, we have an A target in an A reporter or a G target in a G reporter, but we're also getting results when we have the non-specific reporter present. When we have no target, we get nothing. Sensabaugh: Slide 12

So this could be accounted for by non-specific ligation or it could be accounted for by something independent of the ligation reaction; the latter is in fact the case. If we leave out the ligase step, we get the same results as we get here. That tells us that we have something going on that's independent of ligation.

In homologous liquid phase system, the reporter probes themselves serve as primers. Once they initiate a reaction, then they begin to amplify, so we get an amplification product. This brought up two questions: First, how could we get the reporter probes or the target out of the system, because those are causing the problem, and secondly, how could we verify that we're really getting good ligation? Sensabaugh: Slide 13

Therefore, we redesigned the system so that we could test ligation reactions directly. Here are the illustrated results: In fact we are getting specific ligation. We have the G target, G reporter, and a product; the G target, A reporter, and no product; the A target, G reporter, and no product; and the A target, A reporter, and a product.

The ligation, then, was okay, because what we were seeing really was independent of the ligation and not a nonspecific ligation reaction. Instead we were essentially getting a tail PCR reaction with the reporters as the primers in the first stage that subsequently switched over to a straight PCR amplification.

Knowing that, we wanted to find out which ligases worked. There are several commercially available ligases, most notably T4, Taq ligase, and an ampligase, which is used for the ligase chain reaction. Sensabaugh: Slide 14

They have different temperature activity ranges. We tested these to get an assessment of how much ligation product is produced as a function of ligase and temperature. Sensabaugh: Slide 15

You can see a little blip here for the T4 at 25°C, and then as the temperature goes up, the peak increases in size. By the time you get up to 70 or 75°C, the peak is good-sized, and at 80°C, the peak disappears. Sensabaugh: Slide 16

This figure summarizes the results of the preceding experiment and some related experiments. In essence the amount of ligation product increases as a function of the amount of ligase used and of the temperature at which the ligation is done, with a maximum ligation occurring at elevated temperatures between 65 and 75°C.

Subsequent experiments have shown some of the variables that affect DNA ligation. Efficiency increases with increasing temperature. The amount of ligation product increases with the amount of target that is added to the system. There is also a hybridization probe concentration effect, whereas the ligation product can be saturated if the target is in a concentration in excess to the hybridization probe concentration. Sensabaugh: Slide 17

Once you go above a particular level, adding more ligase has no effect, and we're still working on maximizing some of the ligation conditions.

Since working with the solid phase was an integral part of the detection scheme, we also explored solid phase chemistries. There are a number of chemistries for attaching DNA and single-stranded DNA onto solid supports. Sensabaugh: Slide 18

Of these, the DNA–BIND from Corning is the one that we have done experiments with, but we want to go the direction of polyacrylamide. By attaching an acrylic group to the 3' end of a probe, we can incorporate the DNA strand into a polyacrylamide polymer, and those are supposed to be very stable and amenable to PCR. Sensabaugh: Slide 19

Some of the other chemistries are not compatible with PCR, at least in their current configuration, largely because they involve attachment to polystyrene, which has a tendency to melt at higher temperatures, and rather than replace the temperature block in our thermocycler every time we do a run, we think it would be better to use the thermally stable plastic for the attachment.

This is a solid phase test. We've used the Corning DNA–BIND PCR plate, capture, ligation, amplification, and here's what we get. We've demonstrated that we can get the product that we are looking for.

There is a shoulder on the side of each one of those peaks, which may be a 1 base extension product, as is often seen with PCR. There also is a peak here [indicating] that's about 50 nucleotides larger, which shows up sometimes. We haven't done the sequencing on it to find out what it is but that's on the to-do docket.

But at least we've got to the point of knowing that the detection scheme works and the concept works. Now, it's a matter of manipulating the conditions so that we can get it to work with the efficiency that would be useful for use on the kinds of samples in which we are interested.

I think that we'll finish here. Again, I thank NIJ and all of you for your support on this project.