#lab technique

Hot springs are home to the bacteria Thermus aquaticus, cleverly named "hot water" in Latin, and as a result their replication machinery is uniquely resistant to heat. 

Taq polymerase (as it is shortened) still shares one limitation with all other DNA polymerases, though. It can't start a new strand of DNA. It can only add onto an existing strand of DNA by following a template strand.

That's where primers come in. Primers are very short DNA strands that must be designed by the scientist. They pair to single-stranded DNA to give the Taq polymerase a starting point for its DNA replication:

Image from visionlearning.com - great resource!

Because the polymerase can only replicate DNA when a complementary primer is provided, the PCR process can be very specific. You can use an entire genome as a template, but provide primers for only one gene, and only the DNA for that gene will be replicated.

Okay, so let's pull it all together to understand how the process works. To start with, your sample includes template DNA, primer DNA, Taq polymerase, and all of the DNA building blocks (ATP, GTP, CTP, and TTP). It also includes background molecules that provide an environment favorable for the polymerase, such as magnesium ions.

DNA melting. First, the sample is heated above 90 degrees Celsius to unwind all of the double-stranded template DNA. Primers cannot bind to DNA that is already bound to a fully complementary strand.

Annealing. The sample is cooled partway, to about 60 degrees Celsius, very briefly - say, thirty seconds. The goal is to get the DNA cool enough that the primers can bind to the template DNA, but not hold it there so long that the two template DNA strands have time to bind to each other. 

A primer strand is usually only 20-30 bases long, while a template can be hundreds or thousands; matching up all of those pairs is a much slower process. The sample should include primers for each complementary strand of template DNA.

Replication. The sample is heated up to the optimum working temperature of the polymerase, usually 72 degrees Celsius. The polymerase uses the primers as a starting point and constructs new complementary strands for each of the template strands.

Then the process repeats: the sample is heated back to 90 degrees Celsius for the new double-stranded DNA to melt and the process to start over. This is where the Chain Reaction part of the name comes in; as the procedure continues, the sample contains more and more template DNA, and more can be replicated each cycle. For example, if we started with a single strand of DNA, after the first round we would have two strands, after the second, we would replicate each of those to get four; after the third, eight; and so on.

This process is called exponential amplification, and through it, a large sample of identical DNA sequence can be generated from a very small amount of starting material. I have used this to make plasmid DNA for molecular biology experiments, which I described here, but it's also very important in forensic science: DNA samples collected from crime scenes can be amplified to be tested for matches. Because the amplification happens in a mathematically consistent way, quantitative PCR (qPCR) can be used to estimate the concentration of very small amounts of DNA.

Now imagine: when this technique was invented, scientists had to perform it by hand by moving a tube between different-temperature pools of water every few minutes or seconds. The process takes hours! Modern science students are very grateful for laboratory robots.

Image from BioRad.

Baiting the Hook

A fishing rod doesn't function very well without fishing line: something to pull your hook back in when you're ready. For me, that meant creating a GST fusion protein.

GST is glutathione s-transferase, an enzyme that binds glutathione and, well, transfers it. If you add solid glutathione beads to a solution of GST, the GST binds to it, but it gets stuck if it can't transfer it anywhere. The beads settle to the bottom of liquids. So if you have a solution of GST mixed with other things, and you add beads, shake it around a bit, and then remove the liquid, GST will stay stuck to the beads while everything else gets washed away with the liquid.

How can we use this to extract our actual protein of interest? At this point in the fishing analogy, glutathione beads are the line and GST is the hook. Time to bait the hook with Protein A.

Enter a piece of molecular biology called a plasmid, a relatively small circular piece of synthetic DNA. Note that when I say the DNA is synthetic, I don't mean that the molecule of DNA itself is at all different: I mean its sequence was designed by humans. Plasmids are designed to have sites to insert more DNA into the circle, with important sequences on either side. In this case, I want a plasmid that codes for GST right before the insertion site, where I'll be adding DNA that codes for Protein A. The first gene should lead straight into the second one. Here's a basic illustration of the process. Spoiler alert on what we're planning to do with it.

Image from here.

As the image suggests, we then put the plasmid into bacteria, who aren't smart enough to tell it isn't their own DNA. E. coli is used for this purpose. (The strains we use in lab aren't virulent - they can't make you sick.) When the E. coli make protein out of this DNA, they'll make GST and move right onto Protein A without stopping, keeping the two proteins linked. This is our GST fusion protein.

Wait, how does it get into bacteria again? It's called a transformation, and one of the simplest methods is by heat shock. If you suddenly heat E. coli up almost hot enough to kill it, pores will appear in its cell membranes, allowing some of the plasmid DNA to float inside. Then we give the bacteria a little treat for its trouble by growing it in extra nice food for a while, and boom: your DNA is inside the bacteria. You can make E. coli make any protein you want (theoretically), due to that cool thing I keep talking about where all life on Earth speaks the same genetic language.

This is the same way that insulin is mass-produced for diabetics!

But we're not done yet: some of that bacteria has the plasmid DNA and some doesn't. At this time, it turns out that antibiotic resistance can actually be useful to scientists. Somewhere on the other side of that circular plasmid, there is another gene - not a fusion protein, just another gene - that codes for resistance to a certain antibiotic. Our E. coli isn't resistant to the antibiotic, but if it picks up the plasmid DNA, it will start making the resistance protein and become resistant.

After we've let our E. coli recover a little from its near-death heat shock experience, we just dump it in a big vat of antibiotic. Good thing it doesn't have feelings. The bacteria that didn't pick up our gene die, and the ones that did survive.

The next part is simple, but it's called cloning, which sounds cool, so I'm going to tell you about it anyway. After you allow your bacteria to grow and multiply a little - a few hours usually - you take a little of the liquid they're growing in and spread it on a petri dish. After you've left it to grow in a warm place overnight, it looks like this:

Image from here.

Each of those little spots is a colony of bacteria with our plasmid, and each colony is descended from one single cell - a colony of clones. Now all we have to do is pick one, grow it for a while, and then purify our fusion protein from it, which can be done in a variety of ways. 

We kill a lot of bacteria in science.

Gone Fishing

Okay, so now we have our baited hook, our GST linked to Protein A. We throw that fusion protein into cell guts and wait to see if it has Protein B attached to it when it comes back out. We pull it out with glutathione beads as I already described. This part is easy!

I should note here that going fishing in cell guts gives you a better idea of what would actually happen in a cell, but it doesn’t prove that Proteins A and B directly bind each other: Protein A could bind Protein X could bind Protein Y could bind Protein B. If you want to find out for sure whether A and B touch each other directly, you have to go fishing in pure Protein B. Usually, you do both types of assay!

...Looking For a Fish In The Dark?

Okay, so all science metaphors break down eventually. 

We pull Protein A out of the cell guts by its GST leash, bringing along anything bound to it. We remove it from the glutathione beads, either by a special wash or by cutting the fusion protein with an enzyme. Then we put our recovered sample through a gel that separates it by size, so we can see all the components clearly. Now we've arrived at the actual Western Blot. It's also called an immunoblot, because here come the antibodies.

Antibodies bind infectious agents in your body so you can recognize and kill them. You can also grow antibodies to bind almost any protein.

From our gel, we transfer our protein to a membrane where it will be immobilized, but still available for other stuff to bind to it. How do we accomplish this transfer? Through the magic of sticking the gel and membrane together and running electricity through them, basically.

So now we have a membrane. It's full of Protein A, and we want to see if Protein B is there too, but of course they're both invisible. We cover the membrane in an antibody that binds to Protein B (if it's there). But the antibody is also invisible; this doesn't solve our problem. So we use another antibody that binds the first antibody AND has a glowing marker called luciferase.

(Why not make one antibody that binds the target AND has a glowing marker on it? Because it's cheaper, basically. Antibodies that bind specific things are only useful for that specific thing, whereas antibodies that bind other antibodies more generally can be shared between experiments. We put the expensive luciferase marker on the broad-use antibody.)

Luciferase isn't related to Lucifer, but it does have the same root word meaning light. You STILL can't see anything on the membrane with the luciferase marker. The light it emits is too weak. But you CAN take it into an old-timey darkroom with old-fashioned film, develop it, and THEN. FINALLY. You can see whether you have managed to collect Protein B with Protein A.

Congratulations; now you know whether two molecules touch each other.

This procedure is not universal; every lab likes to do each step a little differently. It takes a few days to perform, mostly in the growing of the bacteria, but also depending on how long you want to perform each incubation step. It's something that molecular biologists tend to take for granted, but I like looking back at it to recognize how much biological knowledge is on display. It's crazy. And I didn't even tell you the part where you have to douse your experiment in milk!

(Okay, I'll tell you. The western blot membrane is very good at immobilizing proteins; if you add the antibodies directly to it, they'll just get stuck in it, rather than bind their targets. By letting the membrane sit in milk for a while, the complex milk proteins "fill up" the membrane so the antibodies can bind correctly. What works, works!)