#4: My first genetic engineering experiment with the Engineer-it Kit
By Rohan M.
OK, I don’t fault you if you’re a micropipette-or-two irked with me at this point. I’ve thrown around genetic-engineering-this and genetic-engineering-that without us actually doing any genetic engineering yet. But that’s about to change right now (and yes, I’m actually still lying…).
We understand that DNA is the code that cells compile and then run to build proteins. We’ve seen it with our own eyes, and held it in our own hands! We understand that by loading tiny cartridges of this code (called plasmids) into cells (for our purposes, E. coli) we can give them the ability to do things they never evolved to do – like colorfully decorate our Petri dishes with bioluminescent tapestries. We understand that to get E. coli cells to execute our non-essential (and only slightly vain) plasmid, we need to ensconce them in a nice, comfy, and well-furnished home full of tasty LB agar, so that they won’t mind going the extra mile.
Among the many, many things we don’t understand, however (sorry, but I’m taking the liberty of assuming that you’re just as dumb as I am!), is how to get our plasmid of choice into our E. coli cells. And – just our luck – cells happen to be very picky eaters. Their bilayer membranes are about as receptive to foreign DNA as a baby is to broccoli. Unfortunately, that’s where this analogy ends, because disguising our genetic payload as a friendly aircraft simply won’t do the trick. To get our microscopic charges to eat their veggies, we’ll need to ignore our parental instincts and melt their mouths open (how funny that you think this is a euphemism…).
Of course, we can’t do any of this until we’ve actually come up with the plasmid in the first place. This is, arguably, the hardest part of any genetic engineering project (check out The Thought Emporium’s recent stream for a taste of what this looks like). We probably won’t cover it in this series, though. Besides being a tad intimidating, there are lots of additional logistical concerns that come with designing, printing, and shipping a piece of genetic code to your front door. Moreover, many biologists tend to remix existing, open source plasmids for common use cases anyway – just like programmers prefer importing a library when possible. We’ll come to understand the basic parts of these genetic libraries, but rewriting them gene-by-gene isn’t a particularly good use of our time. Feynman may have had the luxury of understanding only what he could make, but if he was a biologist, he never would’ve gotten anything done!
With that slight disclaimer out of the way, it’s time to hit the lab. For this experiment, we'll use the Engineer-it Kit. But first things first: before we tell our E. coli cells to “Open up!”, we’ll have to culture them. Remember, right now they are packaged in a stab, where they lounge about in a dormant, sleepy state not unlike hibernation. Our DNA transformation protocol will only work once we’ve woken them up, i.e., prepared them a sumptuous LB feast. Since we should be familiar with how to do this by now, I’ll breeze through this section quite quickly, but you can refer to the previous post if you need a refresher.
We’ll start out by preparing and labeling four LB agar plates – three selective, one non-selective. Pour the non-selective plate before adding the antibiotic, and then pour the three selective plates. Label the non-selective plate N.S. and the selective plates S+, S-, and e, respectively.
Here are my three selective and one non-selective plates. Remember, when letting the molten agar solidify, leave the lid of the Petri dish only partially on!
That’s a lot of different Petri dishes! You might think (as I did) that all we would need to do is grow and transform our bacteria and then “re-plate” them, for a total of just two Petri dishes. And we do, in fact, do exactly that with the N.S. and e plates. So where do S+ and S- come in?
Well, this one’s for the programmers out there. I’m sure you wouldn’t disagree with me that coding time is more than worth its weight in debugging time. Thirty minutes of uncommented code can snowball into a week (or more!) of debugging if you’re not careful. And if that doesn’t scare you yet, it should. Nature has spent millions of years weaving the highly-unreadable, undocumented, and self-regulating code of DNA. So you can imagine just how long it might take to debug even the smallest DNA program if we don’t equip ourselves with the proper tools. Luckily, biologists have recognized what most programmers still contest: the best way to debug is with strategically placed ```print``` statements, and that’s exactly what S+ and S- are.
I’ll explain this in due course, but for now, let’s grow some bacteria! To start, we’ll streak our stab of blank E. coli cells – i.e., completely unmodified K12 E. coli – on the N.S. plate. These are the cells we’re eventually going to engineer, so it’s important that they develop into many well-separated colonies. If the colonies aren’t well-separated, then the bacteria will constantly be competing for resources. When a lot of bacteria begin to congregate around a limited resource, they release chemical signals to discourage other bacteria from trying to crash the party, causing them to “slow down” their growth. In case you were wondering, they do this in the extremely polite and socially-acceptable way of screaming “FAMINE!” in their neighbor’s faces. The recipient bacteria take this enthusiastic warning to heart, and lock down their cell membranes to prevent any nutrients from escaping. If before they were like children stubbornly refusing their veggies, now they’re more akin to underground, steel-reinforced vaults surrounded by armed guards and security cameras. Nothing is getting in or out.
This stencil encodes the zigzag pattern we’ll use when streaking our blank E. coli cells. Each colored, dotted line should be traced out with a new inoculating loop.
Obviously, this doesn’t bode well for our DNA, so we need to make sure this doesn’t happen, and give the colonies we end up engineering all the elbow room in the world. Unlike with the Canvas Kit, where we sort of just spread our bacteria about will-nily, we need to streak our E. coli in a more controlled manner this time around. In particular, we will follow the streaking stencil that comes with the Engineer-It Kit. First dip a fresh inoculating loop in the stab of bacteria and streak along the 1A-1B path. Dispose of the inoculating loop, open up a fresh one, and streak along the 2A-2F path WITHOUT dipping the loop back in the stab. Do the same for the 3A-3F path. By repeatedly stretching out the “tail” of each previous streak, we’ll pull the bacteria further and further apart, until they’re extremely well-separated from about 3E onwards.
Time to incubate? Well, not quite. We still need to consider our S- plate. Like the N.S. plate, we’ll streak it with some of our blank cells, but since we won’t be using these cells in the DNA transformation step, we don’t have to follow a specific pattern. Only after streaking S- can we incubate both of our plates (remembering to stack them upside down in a humidity chamber) for about 12-24 hours (any longer than this, and the cells won’t be the optimal “age” for DNA uptake, resulting in fewer successful transformations).
But why? How exactly will this S- “print statement” help us debug our experiment? Earlier, I called N.S. a non-selective plate and S- a selective plate. That’s because the agar in the N.S. plate was poured before the antibiotic was added, so it isn’t selecting for anything. Meanwhile, the agar in the S- plate does contain antibiotic, so it is selecting for something – namely, bacteria that are resistant against the antibiotic added (chloramphenicol, for me). We added unmodified cells to both plates, so we expect them to grow in N.S. but not in S- (in fact, that’s why we call S- the negative control) and we can exploit this simple fact to tell if something has gone wrong. Imagine that cells do grow on our S- plate – what does this mean? Our selection protocol has failed. Either the antibiotic is defective, or enough of it wasn’t added, or its effect is somehow being inhibited. Whatever the cause, the other selective plates we poured – S+ and e – will have this same issue. If we proceed with the experiment, this lack of selection won’t kill off the bacteria which don’t take in our plasmid. Instead, these untransformed bacteria will continue to hog resources that could’ve gone to the bacteria that do have our plasmid, leaving us with less overall product. Even worse, other microorganisms in the surrounding area, such as molds, might start growing on the plate too, contaminating our results!
Bacteria growing on S- is kind of like a runtime warning. Nothing is wrong with our code per se, but some human error – the biology equivalent of package conflicts – will quite possibly make a mess for us in the long run. So it’s worth taking the time to scrap what we have and try again, wasting an hour tops rather than a whole three days of incubation for lackluster results.
Cells grew on the non-selective plate (left) but not on the selective one (right).
Hopefully your code is bug-free though. If it is, you should see something like the above: lots of nicely separated colonies (little “dots” of bacteria) on the N.S. plate ready for transformation (only incubate them for 12-24 hours for optimal DNA transformation results) and nothing at all on the S- plate. If not, don’t be discouraged. It’s taken me multiple repetitions of each part of this experiment to get everything working, and you’ll likely have to try a few times, too. That said, we can now move on to the real meat of the experiment – actually getting DNA inside our cells.
Dispose of the S- plate and set the Cold Station on the DNA Playground to Ice 4°C. It’s also a good idea to set the Hot Station to Shock 42°C, as we’ll need it right after we finish with the Cold Station. Unscrew the cap on the transformation Buffer and place it in one of the four vial racks on the Cold Station. Then, unwrap a blue inoculating loop (its tip is about 10x smaller than the yellow one) and gently collect anywhere from 10-20 colonies from the incubated N.S. plate. Dip the inoculating loop into the Transformation buffer (keeping it on Cold Station!) and gently rotate it like a blender, dispersing the collected cells throughout the solution without touching the sides.
After collecting cells from the N.S. plate with the blue inoculating loop, gently dispose of them in the Transformation Buffer by rotating. Remember, the Transformation Buffer should be on the Cold Station during the entirety of this process!
Keep mixing the solution until no “clumps” of cells remain, and it appears homogeneous and cloudy. It may take up to 30 seconds of vigorous stirring for this to happen, but don’t dilly-dally (spend no more than 2 minutes!) as the window for DNA transformation is very tight.
After the blank cells have been properly mixed into the Transformation Buffer, it should be a cloudy, homogenous solution, free of any visible clumps. Zero to Genetic Engineering Hero book Fig. 4.9.
It probably seems like I came up with these instructions off the top of my head, but everything we just did actually makes a lot of sense when you think about the chemical properties of a plasmid and cell membranes. In fact, let’s take this opportunity to get a closer look at the molecular makeup of a plasmid (you should already be familiar with cell membranes from our DNA extraction experiment). Even though they typically have only ~3,000 bp (base pairs), we’ll still need to work with a highly simplified mental model to get a handle on how they operate. Most plasmids are split roughly into three core regions: the gene encoding our trait of interest (pigment, in this case), the gene encoding antibiotic resistance, and the ori (origin of replication).
A plasmid is a short, circular strand of DNA, with – for our purposes, at least – three main regions. Zero to Genetic Engineering Hero book Fig. 4.13.
The genes encoding our trait of interest and antibiotic resistance should be relatively straightforward, but what about the ori? In a nutshell, it ensures that our plasmid is copied over when its host bacterium replicates. Each E. coli cell has its own larger plasmid that is in charge of keeping it alive (the chromosomal DNA), but the engineered plasmid we insert ourselves remains distinct from that. In fact, the chromosomal DNA is completely unaware of this extra plasmid, and unless it’s alerted to our plasmid’s existence, it won’t include it when duplicating over all of the bacterium’s core structures during reproduction. So the ori is kind of like a large, obnoxious flag that our plasmid waves to let the rest of the cell know it’s there.OK, but how about the Transformation Buffer? What is it, and why do we need it? Despite its intimidating appellation, the Transformation Buffer is really just a bunch of dissolved Ca2+ ions! Not only is it really difficult to get our plasmid into an E. coli cell, it’s also really difficult to get it near an E. coli cell. That’s because both our plasmid and an E. coli cell’s bilayer membrane have a slight negative charge – the former due to the phosphate groups making up its deoxyribose backbone and the latter due to the hydrophilic heads of its phospholipids. As you might expect, these negative charges are far from thrilled about the idea of spending time with one another, and without intervention they would constantly repel each other. The highly positive calcium ions serve as a middle ground of sorts, which both the plasmid and cell membrane can “attach” to, allowing the plasmid to get close enough to the cell to eventually slip past its defenses. In other words, by ordering the construction of several calcium “bridges” along the cell’s surface, we’ve rendered it capable of taking up DNA – or, in synbio lingo, competent (just imagine how ruthless an adolescent E. coli’s report card is!).
Strongly positive Ca2+ ions help the negatively charged plasmids get close to the negatively charged bilipid membrane of the E. coli cells by playing peacemaker. Zero to Genetic Engineering Hero book Fig. 4.14
Of course, helping the DNA slip inside our E. coli cells is no easy matter. Earlier, I said that we do this by “melting their faces off,” and while that’s certainly not a wrong way of thinking about things, a more accurate – albeit pedestrian (literally!) – analogy lies in potholes. Yes, really! While we do put our cells through the blast furnace, we also cool them down immediately after. Transforming E. coli cells really boils down to a cycle of cold, then hot, then cold: an eruption of microscopic potholes across the plasma membranes of our E. coli cells.
To see this in action, unwrap a blue inoculating loop, dip it into the vial of DNA, and twist it a few times. Then, dip it into the Transformation Buffer (which is still on the Cold Station and contains our competent cells), and swirl for ~10 seconds before disposing of the loop. The DNA is now closing in on our E. coli cells via the added Ca2+ ions. Let the road work begin!
Immediately after being on the Cold Station, the Transformation Buffer must be moved to the Hot Station. That’s why this process is called heat shock.
The Hot Station should now have reached 42°C. Quickly transfer the Transformation Buffer to the Hot Station for no more than 90 seconds! While this sudden, extreme heat is a necessary part of the process, it is still killing some proportion of our cells. Even a few seconds too long on the Hot Station can prove disastrous, so you definitely shouldn’t eyeball it – set a timer! As soon as this minute and a half is up, switch the Transformation Buffer back to the Cold Station and lower the Hot Station down to 37°C. Pour the vial of Recovery Media into the Transformation Buffer, and invert it a few times to make sure everything is well-mixed. Then, move the Transformation Buffer back to the Hot Station once it’s warmed up and let it rest for 12 hours.
Whew! That was a lot of information! But we’ve finally done it: our plasmid is now inside of our E. coli cells! Let’s step through everything we just did to see how that happened. First, we cycled our solution of competent cells and DNA through a cold phase, a hot phase, and then another cold phase. This sudden change in ambient temperature is known as heat shock, and it’s one of the simplest ways to do DNA Transformation. Surprisingly, we don’t know how it works, but for better or worse, that doesn’t stop us from acting like we do (rant oncoming!).
Pouring the Recovery Media into our Transformation Buffer and then heating everything up to a comfy 37°C helps the cells inside recover from heat shock. . Zero to Genetic Engineering Hero book Fig. 4.15
So yes, the pothole explanation I offered a few moments ago might not be right. But it is what we think is happening. The sudden switch from cold to hot is believed to make a cell’s plasma membrane more fluid and induce the formation of pores, just like how steam or warm water at a spa causes the pores on your skin to open up. Now, our cell is no longer the picky eater it usually is – in fact, its mouth has become barely defined, loose, and fluid. Of course, this is actually a bad thing, and a few of our cells will be unable to prevent their innards from spilling out or proteins from denaturing. Which, in the cellular world, means death. But most of them will survive, and a significant fraction of the survivors will take up some of our plasmid through their pores. But even these cells are in critical condition, straddling the line between life and death.
To make sure they survive, we need to take immediate action. First, we transfer the cells to a much cooler environment, so their membranes start to seal back up (which also ensures the plasmid is trapped inside) and their proteins can work at top efficiency. But even this isn’t enough. That’s why we douse our cells in some Recovery Media, which is just a liquid version of LB agar. Recovering from sudden temperature extremes is very energy intensive, and the Recovery Media provides our cells with all the various nutrients they need to get back in shape. And that’s pretty much it! We rupture cell membranes by expanding them with heat, close them back up with cold, and give our cells a few extra meals so that they make it through the whole ordeal. A microscopic pothole, created and resealed on demand, with almost surgical precision.
Of course, this explanation might also be complete nonsense, and isn’t that so annoying? Unfortunately, no one seems to care very much about figuring out if it is or isn’t, so long as it works. I’ll spare you my rant, but you may want to check out this interesting paper which suggests some initial directions for future progress (over a decade ago…).
For the final step of the experiment, plate both the S+ and e plates, with our positive cells and newly transformed cells, respectively. Zero to Genetic Engineering Hero book Fig. 4.16, 4.17.
Once we’ve let our transformed cells rest for a while, we can plate them on the e plate. We can also plate the vial of “+ cells” (which we haven’t touched so far) on the S+ plate. Let both plates dry so that no liquid pools on their surface. This may take as long as 30 minutes, so be patient! Finally, we can incubate both plates for 72 hours, setting the temperature down from 37°C to 30°C after the first 24 hours. After three days, you should see something like this:
Both plates should have cells on them, and in different colors, too (since one color comes from the plasmid of pre-engineered cells, and the other comes from the plasmid we engineered yourself). Apologies that my own e plate isn’t there – I forgot to photograph it before cleaning everything up!
But this experiment is not as straightforward as our previous one. First of all, the E. coli cells on our two plates are expressing different colors, and it may not be clear what each one represents. Let’s start with the S+ plate. Just like S-, S+ is a control plate – a ```print``` statement that we use to debug the possibility of human error. But if we already do this with S- then why do we have to do it again? The difference between these two plates is subtle, but crucial.
S- tests against the possibility of some environmental error in the way we prepared our LB agar plates. S+ on the other hand – which is positive control – helps us ensure that our transformation process actually works. Usually, synthetic biologists want to put genes that do much more complex things that fluoresce inside organisms. These more complicated genes are larger, more expensive, and more finicky than those that only code for a single pigment, and so they really don’t want to waste them due to some silent error in their transformation process. As a result, they always make sure to test out the transformation process with a much simpler gene to make sure any discrepancy in their results is due to something else. In this case, Amino has already given us a pre-transformed strain of bacteria (with a slightly different plasmid, which is why the positive control is a different color). While a positive control isn’t actually that useful given that we’re inserting a pigment-inducing gene in both cases, it will be extremely helpful to know how to prepare and interpret these controls as we move onto more complex experiments.
All that’s to say that if we see cells have grown on our S+ plate, it means that the transformation protocol has worked! So if there are no cells on our e plate – which contains the cells we transformed ourselves as opposed to the ones Amino pre-transformed for us – then it’s due to something other than the transformation process. In other words, seeing colorful cells on both plates is exactly what we want! It means the experiment has been a success!
So are we done? Not quite. We’ve gotten everything working, but we’re still missing out on a whole boatload of theory. Namely, how exactly did our E. coli cells use our plasmid to express such colorful proteins? There’s a bunch of beautiful biology involved in answering that question, all of which we’ll be using in future experiments. But this post is already getting quite long. So head over to this “in-between” post to check out how transcription – one of the two main processes that connects plasmid to protein – really works. Hope to see you there, and in our next experiment, where we’ll exchange the painter’s brush for the mechanic’s wrench, and start thinking of our cells less like living tapestries and more like microscopic factories!
Stay tuned for blog post #4.5, in which we look more into what's happening inside the cells during genetic engineering!