It’s very hard taking a selfie of yourself kicking a punching bag – but it kinda works great for today’s topic – GENETIC KNOCK-DOWN & KNOCK-OUT – and how you can use RNAi as a tool to take specific protein-making down – but not out – you’ll need something like CRISPR to bring that about! Today I want to tell you more about how scientists can use RNA interference (RNAi) as a tool to “knock down” a specific gene and see what effects are seen!

The reason I’ve been purifying all these proteins is because I’m trying to study this RNAi process – you can learn a lot more of it here: 

But basically, if you think of your cells as “bakeries” where the original recipes are genes and the “cookies and cakes” are proteins, RNAi is a way to take specific items off the menu temporarily by using small RNAs (~20 RNA letters (nucleotides) long) to guide a protein called Argonaute (Ago) to bind to specific sequences in RNA copies of the recipe to block those copies of the recipes from being used by the “chefs” (protein/RNA complexes called ribosomes) for protein-making (translation).

This is possible because, since the original recipes are kept safely locked up in DNA form in the nucleus (at least in eukaryotic cells which, unlike bacteria, have membrane-bound compartments like nuclei) but the chefs are in the cytoplasm (general cell area outside of such compartments) – so cells have to make copies of the recipe in the form of messenger RNA (mRNA) and then send that mRNA out of the nucleus where the chefs basically get flooded with recipes so they pick up the ones that are most abundant and/or have help getting noticed.

Prokaryotes like bacteria make copies too even though they don’t have nuclei because in addition to keeping the DNA from accidentally getting hurt, it allows for multiple chefs to make it at the same time and it provides opportunities for regulation. Your cells have lots of recipes and make lots of things, so competition’s fierce – and the fewer copies of a certain recipe there are, the less likely it is to be made (note: this is an oversimplification, as there are regulatory ways to “cheat” and get expressed more. 

But, in general, if there are FEWER copies of a recipe, it’s less likely to be made – this is the case with genetic KNOCK-DOWN – we can achieve this with RNAi which intercepts the recipe *copies* – and if there are NO copies, then it CAN’T be made – this requires removing or damaging the *original* recipe (as in that DNA version your cells try so hard to protect) – this can be done with tools like CRISPR and it’s more permanent.

To help understand the difference – think back to that restaurant – if you walk into a restaurant at 6am you might get a breakfast menu whereas at 6pm you’d get a dinner one. It’s not that the restaurant doesn’t magically stop knowing how to make a pancake when the clock strikes noon (and if you ask nicely they might still make you one). But they assume people won’t want pancakes at night.

Similarly, your cells produce different proteins at different times because they only make copies of certain recipes to give to the chefs and they can regulate what copies they make and when – but they still have the genetic recipes for making the proteins that aren’t “on the menu” at some point in time  To go “on the menu” they have to be transcribed – copied into messenger RNA (mRNA) form and taken into the cytoplasm (general cell area) where the chefs are – it’s this mRNA that RNAi destroys. As a result, knocking down a gene (such as with RNAi) is like taking something off the menu whereas knocking out a gene (such as with CRISPR/Cas9 genome editing) is destroying the original recipe so the restaurant doesn’t know how to make it anymore. No more pancakes. Ever.

Even within the realm of RNAi there are different “extremes” – largely based on whether or not the target mRNA gets destroyed by getting sliced and chewed up or just quieted down by hiding it away (sequestration) and/or preventing ribosomes and/or their helpers from latching on and protein-making. 

The term RNAi can refer to a few different things which are often now split into their own categories – RNAi and miRNA-mediated-gene-regulation. They involve a lot of the same players – a small RNA that acts as a guide that gets carried by the protein Ago to a target mRNA that has a complementary sequence (kinda like sticking an address into a self-driving car – they just differ in how close of a guide/target match there is (you need fully complementarity to get slicing) and how the silencing is carried out (does the target actually get cut (“sliced”)?) 

There’s a critical “seed sequence” of ~6-8 letters in the beginning of the small RNA “guide” that has to match, but the whole thing doesn’t have to match in order for Ago to bind and recruit mRNA degradation machinery. But if the whole sequence *does* match, and it’s in Ago2 (the only one of the 4 human versions of Ago that can slice) Ago can cut the target sequence. This type of fully-complementary guide often comes from double-stranded RNA (dsRNA) and is known as small interfering RNA (siRNA) and it usually comes from exogenous (outside sources) – in insects and plants and stuff these sources can be things like viruses because they use siRNA as an immune tool. But we have shifted away from this use of RNAi – and instead, the exogenous source of siRNA for mammals is most likely to be scientists!

Humans and other mammals have kinda evolutionarily moved-away from the slicing and towards less-complementary, more “versatile” guides – your cells naturally use a form of RNAi that uses microRNAs (miRNAs) which are less specific, but can target multiple related genes (kinda like being able to take all the breakfast-only items off the menu when it’s lunch time).

But we naturally use non-fully-complementary (and thus non-sliced miRNAs) guides of a different endogenous source (our cells make it and use it). In fact our cells use microRNA (miRNA) to regulate over half of all our genes. miRNAs are made from special genes (not all genes make proteins – and miRNA can also target recipes for non-proteins-including other miRNA!). But they get processed differently than protein-making genes

Their primary miRNA (pri-miRNA) transcript folds up on itself to form a long hairpin (stem-loop) that gets cut into a shorter hairpin by Microprocessor while it’s in the nucleus, then the top of the hairpin gets chopped off by Dicer in the cytoplasm to form a pre-miRNA duplex that gets loaded into Ago, which ejects one strand to reveal the sequence that can bind to target sequences in mRNAs. 

The specific but not *too specific*-ness of miRNA makes it great for targeting related things at the same time, so it’s great for its natural purpose but not great if you want to use RNAi as a tool to artificially knock down specific genes (take certain items temporarily off the men). In that case you want something more specific., you can artificially introduce siRNA that are fully complementary to the recipe you want to remove – thus they’re more specific and specifically-effective. 

You get higher specificity by introducing guides that are fully complementary so less likely to match multiple things. Unlike miRNA, which can have many targets, since you need a lot more matching letters for siRNA, it’s a lot more specific (think of trying to randomly guess a password of 6 vs 20 letters). 

The stricter requirements for slicing are an advantage for us – and make sense evolution-wise too. If you’re gonna destroy something, you want to be super sure that it’s actually the thing you want to destroy – even when you’re just dealing with copies, not the original. Because molecules don’t really have “hit lists” and they’re not “motivated” to do anything – they just move around randomly – happen to collide with something – and if they like that something they might stick around. The more they like the thing the more likely they are to stick. And the more of the thing there is, the more likely they are to meet.

If there’s a lot of something you might bump into it a lot and, even if it’s not your “fave” you still have a sticking chance. And the more sticking chances there are the more slicing chances there are – unless that is, the target can’t be sliced – remember you need more complementary targets because slicing requires more coordination (the protein has to adopt a slightly different shape to accommodate it) and stickiness so that the guide/target duplex will “hold still” while it gets cut so it gets cut in the right place. 

So it requires extensive complementarity – especially around the site that’s gonna get sliced – the slicing always occurs in the same spot (between the target letters across from the 10th & 11th guide letters (g10 & g11) – because Ago holds the guide really tightly at its 5’ end in a little pocket that has a shape and + charge that’s perfect for gripping that end’s phosphate. 

It’s actually really cool – the way it holds the RNA guide it puts “on display” the “seed sequence” – the first 6 or so guide letters to serve as an initial scout. Even when the guide sequences are totally different, Ago holds this display pose so well that you can see the RNA shape in crystal structures (further along the guide it gets more disordered so you can’t make out some of the bases)

And the scissor location comes from the way the protein folds up – and since it folds the same and holds guides the same, the slicing “active site” will always be aligned with that 10-11 spot. So, whether cutting occurs depends on 1) if the Ago has slicer activity (humans have 4 versions of Ago (homologs) and only 1 of them (Ago2) can slice) – and 2) if the guide and target have complementarity there – if they do, the target will be right there in the scissors’ path – but if they’re not complementary there the target will be “too far away” to get cut. 

 You can actually “see” Ago cleave RNA in a tube. Last week I was doing a bunch of “slicer assays” where I tested the cut-a-bility of some target RNAs. Basically I mix Ago, guide, and a fully-complementary radiolabeled target. If the target gets cut, the radiolabeled piece is smaller & runs faster when you separate the RNA pieces by size on a urea-PAGE gel. 

I was just mixing purified stuff and seeing how well it got cut. But you have more of a challenge if you want to do it in actual cells. Because you have to get the siRNA in there. There are a couple ways to do this.

Only 1 strand (the guide strand) is needed, but you can’t stick a single strand of RNA into cells or it’ll get chewed up right away. You need to protect it. You can transfect cells (transfect means stick in foreign stuff, often by disrupting the cells’ membrane and sneaking in) with double-stranded RNA (dsRNA) for a “temporary effect”

This dsRNA bypasses the nuclear processing miRNA has to undergo (no Microprocessing required) and joins in at the Dicing step. This is temporary because the cell can’t make more of it, it can only use what you give it.

Alternatively, for a more permanent effect, you can stick in a gene vector and tell the cells to make short hairpin RNA (shRNA). It’s kinda like just adding a task to your cell’s normal repertoire – it lives in the nucleus & the cells transcribe it and process it like they would miRNA.

Even in this case, you still aren’t messing with the original DNA gene like you would be with “genetic engineering” tools. You’re just affecting the mRNA copies.

In addition to knocking down specific genes you know you want to knock down, you can use do an “RNAi screen” where you use a “library” of sequences to knockdown lots of different genes in different cells – screen the cells for some effect, then look to see what sequence was in that cell.

sorry I didn’t have a chance to finish formatting this – I got my experiments done & I’m trying to finally get an early-ish night!

more on topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉

Leave a Reply

Your email address will not be published.