random mashup of really really old past posts so sorry for formatting weirdness…
PCR (Polymerase Chain Reaction) is a way to copy specific regions of a DNA template sequence, which we specify by using primers. Primers are short pieces of DNA (oligonucleotides, or “oligos”) which we design to “bookend” our region of interest (amplicon).
PCR is run in cycles of 1️⃣ melt (heat up dsDNA to unzip strands) 2️⃣ anneal (cool down slightly to allow primers to bind and 3️⃣ extend (starting where primers leave off, add nucleotides complementary to template strand until you reach end of template strand). After the 1st cycle (where Pol goes till it runs out of steam or out of time), this end is determined by other strand’s primer because DNA can only *copy* it cannot “compose” so it’ll run off the track corresponding to the position that strand started being copied from in the 1st round. (easier to explain in pics)⠀
You do this over & over 🔁 (30 or so times) to get lots of copies (each time you get 2X as many copies because each new strand becomes another template strand).⠀
Topic #1: Hot Start PCR…
PCR is a fundamental technique in biochemistry and molecular biology, and, like most experiments, there are lots of places where things can go wrong… One “bad outcome” is formation of nonspecific products – if primers bind multiple locations they’ll amplify multiple stretches of DNA, giving you multiple products of different lengths.
The most obvious way to avoid this is to make sure primers have HIGH SPECIFICITY for sequence you want them to bind. You can use free software programs like NCBI BLAST or Primer3 to help you check for specificity & design good primers. More on this in yesterday’s post: http://bit.ly/pcrtrain
But even if your sequence is super specific to a single region of the template, you can still get problems from binding of primer to primer! You can get “wrong pairing” within primers themselves (e.g. hairpins —u—) or between primers (primer dimers). This leads to less primer available to bind template and lower yield (less copies made). Artifact-wise, when you get primer dimers, DNA Pol can end up using primers as a (really short) template, amplifying primer “artifacts” instead of desired amplicon. And the high primer concentrations needed to prevent template-template zipping make such primer pairing more likely because there are more primer fish & fewer template fish in the sea.
When you set up a PCR reaction, you have to mix together ingredients – template, primers, nucleotides, buffer (solution of salts & pH-stabilizers), Mg²⁺, & DNA Pol… We buy a pre-mixed “master mix” of Pol, buffer, nucleotides, & Mg² that makes this a lot simpler, but we still have to add template & primers, set up machine, etc. Because Pol’s part of the mix, it can get a “head start” before you start the cycling. Even if you add the DNA Pol last, right before you stick your (tiny) tubes in thermal cycler, there’s a lag time where, if you’re unlucky, Pol can start working & temp’s low, so you’re at risk of primers settling & giving you nonspecific products.
In hot start PCR, you “hide” Pol until you’re ready to go. The “hider” is often an antibody that binds to Pol & blocks its active site. Antibodies are like primers in the sense that they recognize & bind to specific parts of specific molecules. BUT antibodies are proteins whereas the primers are DNA. And antibodies can recognize & bind to parts of different things (proteins, small molecules etc.) by “mimicking” their surfaces with a combination of well-placed amino acid building blocks (kinda like making a mold of a keyhole) 🗝 Different antibodies recognize specific parts of different molecules, kinda like having a specific key for a specific molecular lock.
The first melt step you separate the template & also separate the antibody, freeing Pol to go to work & since you’re at higher temps, primers have enough energy to seek out their soulmates, so fewer non-specific matches. So you get less nonspecific priming, less nonspecific products you don’t want, & more of the specific product you do want.
In the really olden days (no offense meant to anyone) there weren’t thermal cyclers so ppl had to manually transfer tubes back & forth between heated water baths. Then thermal cyclers came along, freeing arms & minds of grad students in labs around the world! But you still had the nonspecificness problem. The best way to prevent it was to add Pol at the very last minute, but this could be hard when you had a lot of tubes to add to. Then there were some methods introduced using wax to physically separate components in the tube in “layers” until heated. And now there are antibody-based hot start polymerases which are even easier to use.
Other methods hide DNA Pol with aptamers (short oligonucleotides that can kinda fold up & bind things specifically (but not through base-pairing because we’re talking protein-DNA here) that fall off when heated. Or they sequester the DNA Pol in wax beads that melt with heat.
Or you can hide other critical components – some hide the dNTPs or primers with thermolabile (heat-sensitive) chemical groups on the 3’ OH that prevent them from being used until heated. Others hide magnesium in phosphate-magnesium precipitates that won’t dissolve until heated, etc. etc. etc.
Topic #2: Heated Lids
When you run a PCR you have to get it hot enough for the DNA to melt (strands to come apart so that each can be used as a template). Heat is a form of energy & the more heat you add, the more energy the molecules have, so they start wiggling around more & break free of weak attractions to over molecules… The strands separate (but they stay strands because the individual DNA “letters” are linked through a stronger type of bond – covalent bonds that can withstand the heat ).
BUT the DNA you’re trying to separate isn’t alone. Instead it’s in a solution w/all the other things required for the copying reaction (primers, DNA pol, dNTPs, salts, etc.). All this stuff is important & when something goes wrong & you don’t get any product, it’s often because we’ve forgotten to add something (it happens to everyone, don’t feel bad if it happens to you!). But it’s easy to under-appreciate one of the most important ingredients, the thing all that other important stuff is dissolved in, WATER!
When you heat your PCR reaction (over & over & over each cycle… 🔁) you heat ALL the molecules, not just the DNA ones. The water molecules also take some heat & they start moving around more. If they get enough energy, they can escape all of their ties & turn from a liquid to a gas (water vapor) – that is, they evaporate. The evaporated gas is now free to move around & (as a result of random motion) it “tries” to get as far away as possible from the crowded solution because it wants to be as free as possible (have maximum entropy). What’s more free than the open air of the atmosphere?
BUT, the water vapor’s dreams are dashed when it hits a plastic ceiling (the PCR tube’s cap). The temperature of the liquid is controlled by putting the tube in a machine called a THERMAL CYCLER which has a metal block you put your tubes in that hugs the walls of the tube to distribute even heat. BUT if the lid isn’t heated too, the cap of the tube will be cold
Heat flows from hot things to cold things, so the cold lid will steal some of the water vapor’s heat which means the water vapor loses energy. Now, it no longer has enough energy to run away from the neighboring water molecules trying to tag along. So the water molecules stick back together – condense to form liquid water
The water can also no longer avoid attractions to the cap – it adsorbs to the plastic surface, sticking on as a droplet of water. And this frees up more room in the tube’s limited air, so more water can follow suit & evaporate.
The water that evaporates goes solo. It doesn’t take the other ingredients w/it bc they’re nonvolatile (don’t evaporate easily) (think of salt crystals left behind when a puddle of sea water evaporates) This can cause big problems…
You carefully calculate how much of each component to add, you mix them all together, and then some of the water escapes, messing up all those carefully calculated concentrations. Because the volumes are so tiny to begin with (I typically run reactions in 50μL (& a μL is a millionth of a L), a “tiny” loss of water leads to a big change in the concentrations. So you want to prevent water evaporation and/or condensation on the lid.
One way to do this is by using a HEATED LID. You heat up the part of the PCR machine that contacts the cap. This way, the lid won’t steal water vapor’s energy so any vapor that escapes the solution will get stuck in the limited air instead of getting stuck on the lid as a liquid. If too much water tries to escape the air will get crowded too, so it’s no longer as “free” – there’s less benefit to evaporating, so less water will evaporate & evaporated water can get pulled back in.
Before heated lids, scientists would stick a layer of oil over their reactions. Oil doesn’t like water (it’s hydrophobic) so the water won’t evaporate because that would require going through the oil layer. But then you have to go through the oil to get to your product… so heated lids make things much easier.
Topic #3: Polymerase choices
DNA Pols have that general mechanism in common, but different organisms have slightly different DNA Pols that vary in efficiency, processivity, fidelity, & thermosensitivity⠀
- Efficiency: how fast can it go?⠀
- Processivity: how many nucleotides can it add before it falls off template?⠀
- Fidelity: how many typos does it make?⠀
- Thermosensitivity: how much heat can it take?⠀
You can get higher fidelity (fewer typos) if your Pol has a “proofreading” 3′→ 5′ exonuclease (DNA end-chewing) domain that can sense errors, “backspace” to remove them, & then put in the correct letter. This proofreeding is important because errors will get copied… & copied… & copied… BUT it slows down process so you get lower efficiency.⠀
Other parts of DNA Pol proteins can also help out. A way to increase efficiency is by increasing processivity – keep Pol on the template. Constantly falling off & hopping back on surely slows you down! Processivity-enhancing domains (a “domain” is just a protein “section”) or separate processivity-enhancing “subunits” bind dsDNA to help latch Pol on. But importantly they don’t bind “too tightly” & they can bind any sequence – this allows Pol to stay on but slide along⠀
BUT before you can copy strands you have to unzip them & this too is energetically expensive (like peeling apart 2 pieces of stuck together tape). You have to put in energy to give the DNA molecules more energy so they wiggle around more & the strands come apart.⠀
In your cells, enzyme helpers called helicases help unzip them using chemical energy from ATP, but our “bare-bones” PCR version doesn’t have these helpers. Instead, we get the needed energy from heat. In the melt step, we physically heat up the dsDNA so the strands come apart. And we have to get it REALLY hot! (~95°C or 200°F). Human DNA Pol would be pretty useless at this temp bc same heat that causes strands of DNA to come apart (yay!) can also cause proteins to unfold (eek!) (just like chains remain chains when you melt DNA (you don’t break up strong covalent bonds), heat denaturation of proteins leaves you w/chains of amino acids)⠀
Thankfully there are organisms called thermophiles that have evolved to live in super hot environments (like near thermal vents in the ocean). They have super-strong proteins that can withstand high temps needed PCR. The “classic” PCR Pol is Taq, which was discovered in 1976 and gets its name because it comes from the thermophilic bacterium Thermus aquaticus. Taq really made PCR possible – before that, scientists trying to copy DNA in the lab were using a DNA Pol from e. coli bacteria. That DNA Pol couldn’t take the heat, so after each heat step, they’d have to add more! In fact, the first thermal cycler, named “Mr. Cycle” was designed with an open system so you could keep adding more. https://bit.ly/2FoUFkD ⠀
So Taq was a major, crucial, discovery. But it’s definitely not “perfect” – Taq tends to make a lot of typos (low fidelity). Pfu DNA polymerase (from Pyrococcus furiosus) makes fewer errors (so higher fidelity). BUT it has relatively low efficiency. Could we do better?⠀
It’s hard to get all 4, but that hasn’t stopped scientists from trying! Scientists can stitch together parts they like from different Pols to get Pol “chimeras” w/enhanced functions. Our lab uses a chimera called Phusion Polymerase (not a paid endorsement, just what we use!) It’s based off of a Pfu-like DNA Pol (w/proofreading capability for increased fidelity) fused to a small dsDNA-binding protein called Sso7 (from Sulfolobus sulfactaricus) which serves as a processivity-enhancing domain. It can add 1000 nucleotides (1 kb) in only 15 seconds w/few errors! So we time out the extension step accordingly (e.g. if we want to copy a 4kb segment, we’ll set the extension step for 4×15=60s⠀
So, DNA Pol was one choice we need to make, but there are others too.
more about all sorts of things: #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0 or search blog: https://thebumblingbiochemist.com