Die in a Fire
By Sgt York in Science
Wed Jan 10, 2007 at 07:19:05 PM EST
Tags: science, cloning, DNA, bacteria, die in a fire (all tags)
A tale of pyrotechnics, science, manipulation of genetic material, and personal injury, set in the background of a bit of scientific education. Sorry, no actual death unless you include prokaryotes. Of course, if you do, then this story is truly a gruesome tale, involving the death of tens, if not hundreds of millions.
In the popular press, "cloning" is synonymous with Dolly, people that want copies of their pets, or possibly stem cell research. But that's just one specific technique (nuclear transfer) under the umbrella of cloning. It's the most extreme case of cloning, and one that is rarely used, and has very few scientific applications. Today, however, we will discuss one of the more common methods of cloning, plasmid vector cloning. Cloning is one of those hard things to teach; it's hard to know exactly where to begin. So, let's start somewhere you certainly would not expect: Prokaryotic genetics.
When you make a genetic construct, you have to amplify it in order to use it. For small constructs (a few kilobases or so) this is done using plasmids grown up in E. coli. You prepare bugs a specific way, put them in a nearly salt-absent solution, add some plasmid, and put the concoction between two very high voltage electrodes for about a microsecond. Lots of voltage, little amperage, lots of resistance. Hopefully not too much heat, if you do it right. The current causes little holes to open up in the plasma membrane, allowing the DNA access to the inside of the cell. A very small fraction of the bacteria take up the plasmid and call it their own. They will now faithfully replicate that plasmid along with their own genome.
In addition to the main bacterial chromosome1, prokaryotes can carry a number of extragenomic chunks of DNA called plasmids. Like the genomic DNA, plasmids are circular. Picture a long piece of string. Now ball it up. That's a eukaryotic chromosome. Now picture another single long piece of string. Tie the two ends together. Now ball it up. That's a bacterial chromosome. The difference is subtle, but important for a variety of reasons. The only one we'll concern ourselves with here is degradation. At a pH of lower than about 7.5, any free end of DNA will start to break down due to a reaction with water, catalyzed by the acidic nature of the DNA itself. That is to say, that if DNA is in something with a pH of less than 7.5, the DNA will start to destroy itself. Life is inherently unstable. Eukaryotes have complicated end capping systems to prevent this. True to form, bacteria do it the simple way: they make it into a circle.
Back to plasmids. In nature, these are the mechanism by which bugs swap DNA. It's like bacterial spooge. When a plasmid+ bacterium loves a plasmid- bacterium very much, sometimes the plasmid+ bacterium will extend a pilus (a long extension of plasma membrane) to the plasmid- bacterium and inject a plasmid into the plasmid- bacterium. Hot bacterial sex hirez. If the bacteria are not careful, they may unleash the apocalypse. The swapping of these plasmids is how one bacterium will transfer antibiotic resistance to another.
Of course, the bugs get something for their trouble. Namely, ampicillin resistance. That's how we tell which bugs got plasmid and which didn't. The bugs are coddled for about and hour in super-rich media and then spread out on an agar plate laden with ampicillin.
Of course, we irresponsibly hijack this potentially catastrophic process. We engineer plasmids to help us in our mad pursuit of genetic research. We piece together the bare building blocks of a plasmid and then recklessly put into the plasmid an antibiotic resistance gene. The gene encodes a B-lactamase, a
member of a class of enzymes that breaks down the antibiotic ampicillin. Normally, if you plate bacteria on a plate containing ampicillin, they won't grow. However, if you have ampR+ bugs, they grow just fine. That's how we tell who got the plasmid; you plate the whole thing on an ampicillin plate and pick out the colonies that form. If you have the plasmid, it lives. If not, it dies. And don't worry, we don't put in the part that teaches the bug how to make a prokaryotic penis. They are dependent on us for the ability to move plasmid from one bug to another. Besides, we only use ampicillin, and high doses will still kill the bacteria. It's not that reckless.
The technical side of this is as follows: You take a test tube with bacteria in it and add some plasmid. You then take the mix, put it in a cuvette with electrodes in it, load it into the power supply -*ZAP*- and put the bugs in SOC, an exceptionally rich bacterial media, chock-full of glucose. You then let them rest for a while. Then, you plate them by taking out a small volume, placing it on the middle of an agar plate and spreading it around with a small L-shaped glass bar, affectionately known as a hockey stick.
These plasmids are called vectors, in that they are the vehicle by which we work with the DNA of interest. They have a few required features.
OK, rewind a little bit more. New story, which will ultimately tie in with the rest. Bacteria have their own viruses, called bacteriophage, or just "phage". They were first discovered as little clear patches on otherwise uniform lawns of bacteria grown on agar plates, like bacterial colonies in reverse. After a while, the culprit was identified: a tiny chunk of nucleic acid coated in protein that would invade the bacterium, hijack it, and make millions of copies of itself. The protein attaches itself to the bacterium, shifts around a bit and injects the nucleic acid payload into the bug. The nucleic acid then takes over the bacterium's molecular machinery, forcing it to make more nucleic acid and more protein chunks, which are reassembled into fully functional phage. The simplest of these will make more and more phage until the physical forces are too much for the bacterial membrane, which ruptures, allowing the phage to go on and infect nearby bugs.
- ori a locus of origin. This is the site at which the bacterial replication machinery will sit down and start copying the plasmid.
- ampR Ampicillin resistance gene. This encodes a specific B-lactamase, which will break up ampicillin.
- mcs The multiple cloning site. More on this to follow.
Cloning reactions are finicky, and you typically have to do several different conditions to see which one gives you the best results. So, in the interest of time, you do them all at once. And to prevent cross-contamination of each sample, you have to clean the hockey stick between platings. Cleaning involves dunking the hockey stick in a petri dish of ethanol and then passing it through a bunsen burner flame. This is referred to as "flaming the stick." Feel free to use it as a homosexual euphemism. After flaming it, you set it aside, handle pointed down into a holder, L-shaped side up, to cool. Don't want to cook the bugs until their time has come.
But the process is not 100% efficient. Certain phages would exhibit restricted growth in certain bacteria, and that confused people for a while. It suggested that the bugs had some kind of resistance to the phage, that they could fight it. But how could such a simple organism have an immune system? Enter restriction enzymes. Bacteria express enzymes (called restriction enzymes, or RE's) in their cytoplasm that recognize specific short (6-10 base pair2) sequences of DNA and cut them. They protect their own DNA by methylating it with enzymes that seek out the same sequence the RE recognizes and add a methyl (-CH3) group to it. When the virus gets in, its DNA is not methylated, and the restriction enzymes go to work, chewing it up into very specific chunks. Restriction enzymes from different bugs recognize different sequences, and of course different phages have different genomes, so a bacterium will have an increased resistance to phages that have a lot of sequences its RE's recognize. If the phage doesn't have a sequence recognized by an RE in the bacterium, it has free reign.
Have you ever seen ethanol burn? Probably not; it's pretty hard to see, the flame of pure ethanol is almost transparent. When you flame the stick, you see a yellowish, popping flame for a moment as the bugs and media burn off, then nothing except heat shimmer. But that shimmer is really burning ethanol. Hanging on a glass rod. Suspended in the air. Somewhere near a dish full of more invisibly flammable liquid. Cue music.
We hijacked those, too. There are hundreds of categorized REs from different bacteria, and each one recognizes a specific sequence. And we use them to cut DNA in specific ways. The MCS in the plasmid is a short sequence of DNA that happens to have an overlapping series of restriction sites. We use this to put in the plasmid the chunks of DNA we are interested in.
Restriction sites are palindromes, and are cut symmetrically, but not always right in the middle. You have three classes of cutters: 3' overhangs, 5' overhangs, and blunt. A few examples to start out with:
This cuts at the |. Blunts are the simplest cutters. If you cut a linear piece of DNA, you end up with two pieces with blunt ends; neither strand is longer than the other.
Again, this cuts at the |. If you cut a linear piece of DNA, you end up with a little extra 5' sequence hanging off.
GAGCT|C If you start with a linear piece of DNA, you end up with two chunks that have a little extra sequence hanging over on the 3' end.
Once upon a time yours truly was doing just that. I've always had a habit of placing a blue "diaper" (an absorbent pad) under whatever I was doing; it makes cleanup easier and decreases cross contamination between experiments. The plates, tubes, and dish of ethanol were on the pad. The hockey stick, rack and bunsen burner were off to the side, away from the flammable pad and ethanol. Apparently, not far enough. Mistake #1.
OK, file that one away for a minute. The basic idea behind cloning is that you take your plasmid, and cut it in a specific place to yield a linear piece of DNA. You then take your gene of interest and cut it out of whatever it's in. You then stick the insert in the gap you've made, affixing it to both ends and making it a plasmid again, only slightly (or a lot) larger. OK, minute's up.
Obviously, you can't actually pick up the insert and spot weld it into place. You have to rely on enzymes, called ligases, that will join a free 3' end to a free 5' end. However, these adhere to the rules of biochemistry; they just join whatever two pieces happen to line up and happen to be joinable. It's random. If you use a blunt cutter, this means that the ends have to happen to be together at the same time as the ligase is bound to one of them. That could result in what you want, an insert stuck into a circularized plasmid. Or, it could result in two inserts being connected, two plasmids being connected, a plasmid being simply recircularized without an insert, or an insert being circularized with no plasmid. Due to the spatial relationships involved, the latter two occur with much greater frequency than the former three. The benefit is that any blunt cut end can be joined with any other blunt cut end, but it's really not ideal. It's only done if it's your only option.
That's where the overhang cutters come in. Say your insert has the following sequence (not to scale):
(other DNA)---SacI---gene---EcoRV---(more other DNA) And you have a plasmid with a sequence like
You cut the insert out with SacI and EcoRV, giving you a 3' overhang on one end and a blunt on the other. You then cut your plasmid with SacI and HindIII, giving you a 3' overhang on one end and a 5' overhang on the other. Put that in your ligation reaction, and the cut SacI sites line up Recall your base pairs: A:T, G:C. (vector in bold):
The ends can be ligated together in only one way, it's the only way to get the right free ends lined up. This has an added benefit in that the two pieces of DNA will have an affinity for each other due to the base pairing. It's a lot more likely that you'll find these two pieces of DNA in the same space, close enough for ligase to do its thing. But you have some trouble at the other end....(again, vector in bold)
They don't line up. But not to fear....you can fill in that extra sequence with a klenow reaction. Klenow is a little chunk of a bacterial DNA polymerase enzyme. If you give it a template, nucleotides (building blocks of DNA) and a free 3' end, it will copy the template opposite that end. So on the vector side, it will add (in italics):
And now the two can join as a blunt end:
Put that in your ligation reaction, and you have your construct. But it's just bare, circularized DNA....what do you do with it? We have to get it into bacteria, but we've taken away their natural ability to take up this DNA. How do you get it in? You electrocute the motherfuckers.
I had just flamed my stick and set it in the rack; I could feel the heat coming off it and see the "heat shimmer". I turned to get the next plate and tube ready when I see motion out of the corner of my eye...the stick had shifted in the holder. I assumed (mistake #2) that the flame was already out, and didn't take heed of the fact that the base of the L on the hockey stick was now hanging over the petri dish of ethanol. Mistake #3.
You put the bugs in a solution with almost no salt (i.e., very high resistance, as in Ohms) and give it a little voltage. And pray that you did it right so it doesn't go *pow* from the heat buildup (this happens if you don't clean up your reactions correctly). The current causes small holes to open in the bacterial membrane, which allows passage of the plasmid into the cytoplasm, where it will grow happily as the bug grows in the rich media in which it will soon find itself.
A drop formed on the end of the stick and fell, directly into the petri dish, which immediately burst into flame. At first I only felt it as heat, but then I saw it as the plastic of the dish caught fire. I quickly grabbed the corners of the diaper and dumped the whole thing in the sink before things got worse. Mistake #4. Square pads have four corners. People have two hands. Pick up a sheet of paper by two corners. Now predict where liquid would flow. Now predict where burning liquid would flow. Invisibly burning liquid.
Now you're set. Bacteria love to grow, and as they grow they will duplicate the plasmid faithfully. Thanks to some special engineering at that ori site I mentioned earlier, there will even be more plasmid than genomic DNA. It's easy to separate out, as the plasmid is significantly smaller than the genomic DNA. All that's left now is QC; you have to make sure everything worked how it was supposed to work. The ratio of plasmid:vector has to be right, or you'll get multiple copies of your insert in the plasmid (concatamerization), which can cause trouble down the road. Things in the reactions are dominated by chaos, so whatever can go wrong, will go wrong. You compensate for this by doing several reactions in parallel and checking your work through creative use of restriction enzymes; you can cut the DNA up and run it on a size separation gel; each construct will give a distinct pattern of band sizes as a result of certain single or multiple restriction enzyme digests.
When you get a 3rd degree burn, it doesn't hurt. Pain is the manifestation of chemical signals sent from damaged tissue. When a cell is broken, it generates the proteins endothelin and bradykinin through the breakdown of cellular components. These diffuse away and contact receptors on nearby nociceptive nerves. But flame destroys those proteins just as efficiently as it destroys the collagen and elastin of the matrix. You feel nothing. But you do smell something. Plenty of things; hair, flesh, clothing.... My arm was on fire.
It was surreal. At first, I thought it was just the lab coat, and I tore it off and threw it into the sink as well. But the liquid had flowed right up the sleeve from the wrist, and a pool of the burning ethanol had clung to the inside of my right forearm. In a moment of panic I stepped towards the sink again, to douse the flame in water. Fortunately, intelligence beat instinct and I grabbed another lab coat instead, wrapping my arm and smothering the flame.
It was about a day before it started to hurt like a bitch. The scar has faded and shrunk, it's about 3cm across today, about a third of the size of the original wound. But my pride has yet to recover.
The moral of the story: Don't fuck with mother nature. Messing around with genes will only lead to heartache and pain.
1Although bacteria don't really have chromosomes, I will refer to them as such in this story for ease of communication.
2 There are others as well, called "infrequent cutters" that recognize longer sequences. As you can imagine, any short sequence of DNA will occur more frequently than a longer one. Here, however, we are only concerned with the frequent cutters. We want restriction sites that are likely to be found naturally in any given gene of interest.