RNA interference provides a way to down-regulate no-longer-wanted mRNA! It’s what I study, so I’m a bit biased, no lie. But I love RNAi and here’s why! Basically, it’s a way for cells to use short pieces of RNA (small RNAs including miRNAs and siRNAs) to selectively “search for” copies of protein recipes that contain specific sequences and then take those recipe copies “out of circulation” so that the “chefs” (protein-making complexes called ribosomes) make less of the corresponding proteins. In nature, RNAi is a fundamental biological process that plays a key role in keeping the tight regulation required to allow cells to grow and develop in a controlled manner (i.e. without becoming cancerous). Additionally, because RNAi is sequence-guided, synthetic small RNAs can be designed to target specific genes; these RNAs can thus be used to investigate the roles of various gene products in diverse biological processes and hold promise for the treatment of diseases for which the genetic cause is known. Small RNAs can have large effects, but can’t go it alone! So bring in the protein Ago and off we’ll go!
In jargon, RNA interference (RNAi) is a post-transcriptional regulatory mechanism that uses small RNAs (including microRNAs (miRNAs) and small interfering RNAs (siRNAs)) as guides to direct a protein called Argonaute (Ago) to bind to and repress target messenger RNAs (mRNAs) containing sequence complementarity to that small RNA. This triggers mRNA degradation and/or translational repression – bottom line: less of the corresponding protein gets made.
Now, let’s de-jargonize that a bit, shall we?
I talk a lot about proteins as cookies and cakes (a nice analogy they do make) but today let’s go a little less bakery-y and a bit more chef-y… When the price of avocados rose, many restaurants took avocado-y items off the menu – the chefs still knew how to make it, they just didn’t because customers couldn’t order it. Just like a restaurant’s menu changes seasonally, our cells produce different proteins at different times but they still have the genetic recipes for making the proteins that aren’t “on the menu” at some point in time. If cells are like molecular restaurants, where the chefs are ribosomes and the dishes they make, how do your cells take proteins off the menu? I spend a lot of my time researching one such way – RNA interference (RNAi). This mechanism sends recipe destruction machinery on the hunt for specific recipes based on “key words” hidden in them. And at stakes is a lot more than a few bad Yelp reviews…
Your body contains billions of cells, each of which has to be able to respond “in real time” to what’s happening around (and inside) you. Responses often involve the use of molecules called proteins, whose original recipes (GENES) are written in DNA and locked up in a membrane-bound compartment called the nucleus. These original copies can’t leave the nucleus (like a reference section of a library), so there’s a lag time between when a cell realizes it needs a protein and when that protein’s ready to use; it has to make a copy of the genetic “recipe” for that protein (transcribe the DNA sequence into messenger RNA (mRNA)) and send that mRNA copy to the “chefs” (ribosomes) that translate it into protein, linking together the protein letters (amino acids) the RNA letters (nucleotides) say to.
One way to help cells respond more quickly is to stockpile the mRNA, but this introduces new complications; you have to tightly regulate its translation to prevent problems like cancer from occurring and you need a way to get rid of mRNAs for proteins you know you won’t need again for a while so that, instead of them taking up space and resources, you can recycle their pieces to make mRNA for proteins you will need.
mRNA is regulated in large part by a fundamental, evolutionarily-conserved process called microRNA (miRNA)-mediated regulation, or RNA interference (RNAi) and it involves types of RNA you don’t usually think of regulating the type of RNA you usually think of (hopefully you think of RNA).
See, RNA can have other functions than as a DNA-protein “go-between” “recipe” role. That’s just one type of RNA – messenger RNA (mRNA). It tends to get most of the attention but there are lots of other types of RNA that are the “end of the line” (i.e. they’re not instructions for some other product, they *are* the product). These are sometimes referred to as “noncoding RNAs” or “functional RNAs”
The type I study most is microRNA (miRNA) – they may be “micro” in size but they’re “macro” in importance! – they work to regulate over half of all our genes. miRNAs have their own recipes in DNA and get written (transcribed) just like the like protein-coding genes do, but they get processed differently. They’re written (transcribed) as long hairpins but they get chopped a couple times and, after processing, they’re ~20 nucleotide (nt, RNA-letter) long & single-stranded. (more on their “birth” (biogenesis) in a minute).⠀
On their own they can’t do much but when bound to a protein called Argonaute (Ago) it’s like an address typed into the GPS of a self-driving car.
Different miRNAs contain different “addresses” for a specific mRNA or set of mRNA “targets” to be silenced (recipes to find & destroy). This address gets entered into a self-driving car (binds a protein called Argonaute (Ago)), and the car takes it to that address and does its thing. (We call this address-loaded car the RNAi-induced-silencing complex (RISC))
The car knows where to “park” because the miRNA’s sequence is partially complementary to a sequence in the mRNA target. Complementary RNA letters can base pair with each other – like matching puzzles pieces – “A” binds “U” and “C” binds “G.” So you can get 2 complementary sequences to bind to each other, and thus a sequence can act as a guide to direct silencing machinery to a complementary sequence.
Once parked, the car can either directly silence it or recruit other helper proteins to repress translation of the mRNA (temporarily keep it from being made into protein) and/or degrade the mRNA completely so they can’t be used to make more protein (don’t worry – these are just copies of the “original” DNA instructions that are still held safely in the nucleus).
Which way it goes depends in part on how well the sequences match. There’s a critical “seed sequence” of ~6-8 letters in the beginning that serves as the “code word” – this “has” to match, but the whole thing doesn’t have to match in order for Ago to bind and recruit mRNA degradation machinery. What happens next, in the “effector phase” depends largely on the amount of matching.
If the whole sequence *does* match, and it’s in Ago2 (the only one of the 4 human versions of Ago that can slice under normal conditions) Ago can cut the target sequence. This type of fully-complementary guide usually comes from a different type of small RNA (sRNA) called small interfering RNA (siRNA). siRNA typically comes from double-stranded RNA (dsRNA) that usually comes from exogenous (outside sources). This long dsRNA gets chopped into ~22 nt sRNAs by a protein called Dicer.
One source of dsRNA is viruses, so plants, insects, and some other critters use siRNA for antiviral defense and stuff (e.g. using pieces of an invading virus’ RNA to target that virus). But we have a more complex immune system to take care of those, so our RNAi system has shifted to focusing on miRNA-mediated repression. But our cells still have the capacity to deal with siRNA, so we can hijack the system to direct Ago to specific genes. Often if cell biologists want to see what a specific protein does in cells, they’ll use RNAi – they put double-stranded RNA that gets chopped into siRNA containing the protein’s mRNA address into those cells to selectively take that item off the menu. This is often called “genetic knockdown.”
But most of the time, our cells are dealing with miRNA & its imperfect matches, so Ago calls for backup, recruiting (with the help of a scaffolding protein called GW182) deadenylation (poly(A) tail removing) complexes & decapping complexes which remove the mRNA’s protective ends so they can be chewed up.
Your body uses miRNA-mediated mRNA regulation as a sort of thermostat to regulate levels of proteins being made inside each of your billions of cellular homes. It’s always working (or it better be!) and it works on tons of targets using miRNA it makes.
But what miRNA to make? Depends on what protein recipes you want to take off the menu. And there are LOTS of options! If you had to have a totally unique one for each recipe that’d be way too complicated. So how to make things easier? Combine and conquer!
The “code words” that match the miRNA are usually in the 3’ untranslated region (3’UTR) of mRNA – this 3’UTR is like the “backmatter” of a recipe book (the index & glossary & stuff that’s there to help but doesn’t contain actual info about what to add when). mRNAs have many target sites (aka mRNA response elements or MREs) in their 3’ UTRs, so they can be targeted by multiple miRNAs. And the same miRNA address can take Ago to many different recipes because different mRNAs can have some of the same sites. This allows many proteins to be targeted by the same miRNA (or copies of it). And this is great for regulating functionally related proteins that are needed (and not needed) at the same time. Even if you want them at the same time, you want to be able to regulate them individually in case you overestimate demand for one of them, etc. So in addition to having some shared sites, you also have some different sites. So you can target different groups of proteins at the same time (think Venn-diagrammy).
To help visualize the significance of this, let’s go back to our avocado adventure… If there’s an avocado shortage, it doesn’t just affect avocado toast – it also affects guacamole. So a restaurant might have to stop making both. If both of their recipes have the same “code word” the same miRNA can “erase” them both -> Often cells want to regulate the levels of multiple related proteins at the same time, so they often contain sites for the same miRNA.
But it’s not just an avocado shortage that might make you want to take guacamole off the menu. What if the restaurant finds that no one’s ordering it at breakfast time? – they can take it off the breakfast menu – but they don’t want to take off the avacado toast at that time because people want it – this is why recipes have different combos of miRNA sites, allowing you to control related genes separately. And that’s where combos come in.
So, let’s all “tip” RNAi for keeping each of our cells running like a 5-star restaurant!
Now, as promised, a bit more detail about miRNA biogenesis, starting with “why do our cells use miRNA instead of siRNA?”
As you might recall, siRNA comes from Dicer chopping up dsRNA. Although our cells don’t typically use dsRNA to elicit specific siRNA-mediated antiviral attack, dsRNA does play a role in our immune responses, albeit a more generic one. Our cells have proteins that bind to dsRNA, serving as sensors for invaders -> the presence of dsRNA can set off innate immune system alarm bells in our cells. That wouldn’t be good if the dsRNA was something we were making for regulatory means…
Further disqualifying it, dsRNA is “hard” to make because you have to copy 2 strands and coordinating the copying of a single strand is hard enough!⠀
So, instead of coming from long dsRNA, miRNAs are made from long single-stranded RNA that folds up on itself to form a hairpin. So you still get a double-stranded region to feed Dicer, but you avoid those other problems. ⠀
miRNA birth starts with transcription, which is where RNA polymerase makes an RNA copy of DNA. Most of the time when people talk about transcription they’re referring to making mRNAs (our protein recipe copies). To make mRNAs, genes are transcribed into pre-mRNAs, which get edited through a process called splicing that removes regulatory introns and stitches back together the protein-coding exons. They also get a 5’ (starting end) 7mG cap (a modified backwards nucleotide) and a 3’ (ending end) poly(A) tail (lots of the RNA letter A) added “generically.” The cap & tail help tell the cell that those are mRNAs. It helps them get trafficked out of the nucleus and into the cytoplasm, where protein-making machinery recognizes and binds them and gets to work. ⠀
Since miRNAs are NOT mRNAs, they don’t need (or want) those mRNA identifiers. So, even though they do get a cap when they’re born (cuz that’s coupled with transcription) and some get a tail, those are removed during processing, where the miRNA gets cut out of the middle of the original transcript.⠀
The cap & tail also provide protection for mRNA’s ends because there are RNA exonucleases (end chewers) that degrade “raw ends.” Since miRNAs don’t have this protection, they wouldn’t last long without protection from proteins, so they’re passed off from one protein to the next.⠀
miRNAs often have their own genes (though multiple miRNAs can be transcribed from the same gene and then separated) and their sequences are such that they fold back upon themselves into long hairpins (aka stem-loops)(you might be familiar with “base pairing” between strands of DNA or RNA? well, base pairing can also occur within a strand, and RNA often tends to do this).⠀
These long hairpins are called primary miRNA (pri-miRNA) are bound in the nucleus by a complex called Microprocessor that contains a pair of RNA scissors (RNA endonuclease) that chops of the non-loop end of the hairpin to form a shorter hairpin (~60nt) called pre-miRNA. ⠀⠀
Microprocessor is heterotrimeric – it as 3 (tri) parts and the parts aren’t all the same (hetero). The scissors-part (endonuclease) is a protein called Drosha and the other 2 units are copies of its helper, DGCR8 (aka Pasha in flies and worms) which helps recognize it as a pri-miRNA in need of processing and holds it in place for Drosha to cut. Drosha has 2 pairs of scissors (2 RNase III) domains. One pair cuts on one side of the hairpin & the other pair cuts across from it, with a 2nt offset so it leaves a little overhang. That overhang is important for helping it get recognized and processed by Dicer instead of binding to one of the immune sensor proteins which binds blunt ended dsRNAs. ⠀
This (still premature) miRNA then gets transported into the cytoplasm. Since it’s not an mRNA and it doesn’t have a cap complete with binding partners to help shuttle it out the mRNA way, it uses a different helper, Exportin 5 (Xpo-5). Xpo-5 clamps onto the base of the hairpin, protecting its ends. And it also binds another protein called Ran, which binds GTP to provide energy to power the process. Xpo-5 helps shuttle the pre-miRNA out, then Ran GTPase-activating proteins in the cytoplasm convince Ran to “spend” the GTP (hydrolyze it to GDP), causing it to change shape and release the pre-miRNA and the exporters get recycled.⠀
In the cytoplasm, instead of being bound by protein-making workers, pre-miRNA is bound by another pair of scissors, Dicer. Dicer lops off the loop, turning the pre-mRNA hairpin into an miRNA duplex. Like Drosha, Dicer cuts both strands at an offset, so the duplex has ~2nt 3’ overhangs on each side (each side being ~22nt long)⠀
This duplex gets loaded into another protein called Argonaute (Ago), to form a pre-RISC (RNA Induced Silencing Complex). Pre??? Yup – you’re still not done!⠀
As we discussed above, the targeting power of miRNA comes from it containing a sequence that’s complementary to (can base pair with) a sequence in the target mRNA (usually in the target’s 3’ UTR (untranslated region) (the sequence past the end of the protein instructions but before the generic tail).⠀
But in the duplex form, the guide strand miRNA’s sequence is being hidden by the “passenger strand” (the other half of the duplex that was across from it in the hairpin). So Ago ejects the passenger strand to form a mature RISC and positions the guide strand to go searching for targets. When it finds one, it shuts down protein production. note: which strand it ejects is determined in part by which end of the duplex is less thermodynamically stable (and thus easier for Ago to grab tightly onto the end of a single strand). ⠀
Since RNAi is my favorite topic, I’ve done several more in-depth posts on various aspects of all this stuff, and you can find links to them (and tons of other topics) here 👉 http://bit.ly/2OllAB0
And here’s a link to an RNAi overview figure I shared on WikiMedia Commons for download & use: https://commons.wikimedia.org/wiki/File:RNAi_overview.png