Nucleic acid amplification-based molecular methods, such as PCR, RT-PCR, loop-mediated isothermal amplification (LAMP), helicase dependant amplification (HDA), and other similar techniques, derive their utility from their combination of high specificity for what they amplify (targets must anneal to amplification primers of some form in each of these methods) and their impressive amplification capacity. Recall that an idealized PCR reaction can make 1012 copies in only 40 thermal cycles from a single target molecule in the starting reaction. It is this very power which can also be the Achilles’ heel of the techniques, in the form of amplicon contamination.
Containing contamination in the early days
If we think back to the early days of laboratory molecular diagnostics, when methods were mostly limited to classical endpoint PCR reactions analyzed through gel electrophoresis, it’s not hard to see how the problem arises. Your nice, shiny new molecular laboratory was a thing of beauty and cleanliness—until you took your first set of PCR reactions out of the thermocycler, and opened the caps. From that point on, the potential to aerosolize and disperse microscopic droplets, each loaded with large numbers of reaction products, just increased with every step. Loading your sample into the agarose gel? Thanks to loading buffer, you could actually see small traces of the sample diffuse out of the well and into the surrounding large buffer volume. Since for every assay you must at least have run a positive control reaction, the risk was inescapable. Just a single amplicon from a prior reaction, somehow introduced into a new sample, would amplify and make the sample look “positive” regardless of its true status.
Contamination control in this setting was primarily limited to a large number of physical separation measures. Molecular laboratories were divided into two physically separate areas (pre-amplification and amplification processes, or “clean”) and post-amplification processes (“dirty”); or even better, three areas: pre-amplification for reaction mix preparation (“clean”), sample preparation and addition to mix (“clean,” although note at this stage, amplifiable template from samples and controls is present and of some contamination risk, albeit at much lower concentration than post-amplification), and amplification/post-amplification (“dirty”).
Separation of these areas began with simple steps such as dedicated pipettes for each task, and could go so far as to have each area on separate floors of a building, or in separate buildings; having antechambers for rooms, with room-specific gowns; maintaining unidirectional workflows (no movement of material from post-amplification back to pre-amplification areas); and air pressure control (maintaining post-amplification areas at negative pressure, to reduce aerosol transmission outward). The use of biological safety cabinets (BSCs) or their simpler cousins, PCR hoods, was taken as a given for all pre-amplification steps. Further steps to combat contamination often included room-bathing shortwave UV lights set to run every night, damaging any residual nucleic acid; aerosol blocking pipette tips, to avoid contamination of pipettes which can then act to dispense amplicon with every use; “sticky mats” at room entrances, to clean shoes; and various personal items beyond common lab protective equipment (shoe covers, hair nets, face masks). The infrastructural requirements of all of this, and the fact that with all of this contamination events still sometimes occurred, was a major impediment to wider adoption of molecular techniques in the clinical lab in early days of molecular medicine.
The introduction of real-time PCR techniques, where the reaction process could be analyzed from the outside optically and there was no longer a requirement to open the amplicon-laden tube, was hailed as a major advance in this regard, and in truth it does a great deal to reduce these risks. However, sometimes even real-time PCR reactions get opened post-reaction either intentionally or accidentally. For this reason, all of the physical control measures outlined above are still valuable and should be considered for implementation wherever feasible. While they are costly and inconvenient, that pales in comparison to the disruption and inconvenience caused by an actual contamination, which will generally entail shutting down molecular lab services for a period of time, discarding all stocks of potentially contaminated reagents and materials, extensive cleaning, often in multiple cycles, and rounds of testing negative control reactions until “amplicon cleanliness” is restored.
Better living through chemistry?
In addition to all of these physical contamination control measures, there are also some chemistry-based approaches to contamination control which are available to today’s molecular laboratorian. With respect to positive control material, space limits us from going into detail, but suffice it to say that it may be possible to genetically engineer in small sequence variations between control material and real targets—most commonly, a few base changes to create a unique “restriction endonuclease” recognition site in controls. This forms a site within the amplicon where a specific enzyme will cleave the DNA and render the product non-amplifiable. This approach (which I used to frequently employ in my LDTs) allows for reduced risk of contamination from positive control material; since these are used in every run, they can be statistically the biggest risk of contamination when the assay is for uncommon targets, as in some infectious agents. At best, though, this is a partial fix, since at least some of your samples must be truly positive.
There is another chemistry-based contamination control approach which is available in most amplification methods and has completely general utility; it allows for the uniform destruction of any and all carryover, contaminating amplicon in a reaction before it can cause a problem, and it does so in a simple, elegant manner. This approach, known as uracil-N-glycosylase or “ung” treatment, is the focus of the rest of this month’s column.
The ung enzyme
First, some basic nucleotide biology. Under regular physiological conditions, the base cytosine (C) can undergo spontaneous deamination, or in other words, loss of the ring’s projecting -NH2 groups and replacement with an =O. This converts the cytosine to a uracil. U, with an important consequence; while C pairs to G, U pairs to A. Spontaneous deamination thus creates a point mutation wherever it occurs, with one template strand being the original, correct sequence and the other strand, now with the uracil, misdirecting a replicating polymerase to insert an A for what should be a G—a transition, in genetic parlance. This is neither an insignificant nor a rare event; estimates are that in every human cell over 24 hours something on the order of 100 cytosine deamination events occur. Not surprisingly for such a common and potentially serious mutation, life has devised a way to detect and repair this specific damage event. Known as base excision repair, it begins with the uracil-N-glycosylase enzyme, which acts to search through dsDNA for any uracil residues. The ung enzyme acts to cut out any uracil bases it finds, leaving the sugar-phosphate DNA backbone intact but a hole in the series of bases. This hole or gap in the series of bases acts to block progress of a DNA polymerase using the strand as a template.
In order to put this into practice as a contamination control measure, we need to do two things. First, we need to make our DNA amplification products contain at least some percentage of uracil. Fortunately, most DNA polymerases, when presented with dUTP as a nucleotide, will accept it in place of dTTP, albeit with sometimes slightly lower efficiency. Since only a single uracil residue per amplicon strand is all that is needed to mark it for ung cleavage, this lower incorporation efficiency is balanced by using a mixture of mostly dTTP with some dUTP in the reaction mix. Most of the time the polymerase will incorporate a T against an A template location, at full efficiency; only occasionally will it pause to incorporate a U. Overall good replication efficiency is maintained, but every amplicon will have at least some uracil residues.
The second half of this approach lies in adding ung enzyme to every PCR reaction as it is set up. This can either already be in the master mix preparation, or be added in as part of the reaction setup. Ung isn’t a thermostable enzyme, and we want it to act before any PCR (or alternate amplification method) occurs, so the approach is to pre-incubate the reactions containing ung for a few minutes prior to an initial high temperature strand denaturation step. During this incubation time, the ung will act to cleave any carried-over, uracil-containing contaminant template material. The subsequent high temperature allows the start of amplification by denaturing target strands for primer access (and in most cases, activating a “hot start” polymerase), and also acts to permanently inactivate the ung so it can’t interfere with the production of new, uracil-containing amplicons.
As formulated into most premade reaction kits, application of ung contamination control therefore consists of nothing more than adding the ung to dUTP containing master mix (if not already present), and adding a few-minute PCR reaction preincubation step at moderate temperatures (usually 37°-45°C).
While other chemically based methods for contamination control do exist, none are as easily introduced as the uracil-N-glycosylase based approaches. In an ideal modern molecular lab, both extensive physical contamination control measures and chemistry-based approaches should be combined and together can greatly reduce risk of false positive results from amplicon contamination.