Would you like a phosphate cherry on top or on the side of your protein? Kinases are like “presentation chefs” that embellish the protein products your “line cooks” (ribosomes) make – but they do *charge* – the “cherries” are bulky, negatively-charged phosphate groups, which can change not just what the protein looks like, but also how the protein acts which is great when you want that to change! But if kinases go too cherry-adding-happy, proteins can start acting up. And when there’s a problem in the system, like a rogue presentation chef, problems like cancer can occur, so kinases are sometimes targets of anticancer drugs like Gleevec, which targets a rogue chef called Bcr-Abl.
note: this post is updated from September 2019
I had kinases on the mind today because today I used a kinase to radiolabel some RNA – yup, some kinases phosphorylate things other than proteins, but today I’m going to focus on the protein ones.
You can think of your genome as a collection of recipes for all the things your cells will ever need to make. Some of these recipes are for making protein & RNA “chefs” that help “bake” (translate) other proteins or RNAs from other recipes.
There’s only so much you can tell by reading a recipe, which is 1 of many reasons why simply sequencing the DNA of everything won’t give us all the answers, which is one of the reasons why structural biology (where we look at the shape of molecules like proteins and how that shape affects how they function and how they interact with one another) is so important
The “line cooks” initially putting together the protein are protein/RNA complexes called ribosomes and they’re pretty restricted to following the recipe word for word. But even if you have that recipe in front of you, it’s hard to “visualize” the final product until it’s complete
The recipe tells you what amino acids (protein building blocks) to add in what order. These amino acids will be linked together through their generic backbone by strong, covalent, peptide bonds so we know that amino acids next to each other in the recipe will be next to each other in the products (like linked chains in a charm bracelet). But what we don’t know is what else they’ll be near because proteins get folded up into (beautiful) 3D structures
Different amino acid “ingredients” or “letters” have different properties because in addition to the generic backbone that makes them easy to link together any 2, they have unique “charms” called “side groups” sticking out. Some charms are big and bulky, others are small and flexible, some are hydrophobic (water-avoided), some are hydrophilic (water-loving), some are charged (+ or -) others are neutral.
These properties influence how the protein gets folded (e.g. hide the hydrophobic parts in the center, put the hydrophilic ones on the outside).
You know how some people fold their pizza in half (not judging, just analogizing…) – parts of the pizza that were once on opposite sides are now close together. Similarly, just because 2 amino acids are far apart in the recipe doesn’t mean that they’ll be far apart in the final structure
BUT if 2 amino acids are far apart (or close together) in the “final” structure it doesn’t mean they *always* are because there’s often not just “1” final structure. Proteins can change shape subtly or dramatically & such “conformational changes” can be caused by post-translational modifications (changes to the protein after they’re made)
Unlike the line cooks, “presentation chefs” can add embellishments (parsley sprig anyone?) that can change the protein’s shape and appearance. One common type of display cook are KINASES. They modify proteins after they’re made (post-translationally) by transferring a phosphoryl group (-PO₃²⁻) group to one of a protein’s charms -> phosphorylation
Usually the phosphoryl group comes from the end of a nucleotide, most commonly from Adenosine TriPhopshatte (ATP). Nucleotides? Like the ones that make up RNA & DNA? Yep – the very same! ATP gives you the adenine “A” of RNA (and if you remove an oxygen from ATP’s sugar to get deoxy adenosine triphosphate (dATP) you get the precursor to DNA’s deoxyadenine “A”)
Although it’s technically a phosphorylation group being added (the 4th oxygen comes from the protein), we normally just say that kinases add phosphates. Phosphate (PO₄³⁻) has a central phosphorous(P) atom connected to 4 oxygen(O) atoms. It has “extra” electrons (e⁻) so it’s ➖ charged. Putting so many negative charges right next to each other gives you an “unstable” high-energy molecule because the negative charges repel and it’s hard for the phosphate bonds to hold on. Enter the “escorts” – positively charged metals like Magnesium (Mg²⁺) complex with the phosphates and chaperone them as they move from place to place. And proteins like kinases that need to bind them have metals in their binding pocket as well as charged amino acids sticking into the binding pocket that can do a “changing of the guards.”
Kinases are a type of enzyme – they catalyze (speed up) reactions without getting used up (they can add a phosphate to 1 protein then turn around & add 1 to another protein). Kinases catalyze the phosphoryl group transfer by helping hold together the reactants (ATP & protein substrate) in the right position for the transfer and stabilizing unstable intermediates. It can do this because the amino acid charms sticking out into its “active site” make a perfect environment for the hand-off. But how do they know where to add?
Different kinases recognize different target consensus sequences – like their “dating profile” describing their “ideal match” sequence around the phosphorylation site – but there’s some wiggle room (and some are more promiscuous than others)(have higher specificity) – the ideality is due to that sequence making the substrate complement the shape, charge, etc. of their binding pocket (like maybe their binding pocket is negatively charged, so they want something with positively charged amino acids near the phosphorylation site)(again, something you wouldn’t know from the linear sequence alone). But what’s the real “phosphorylation site?” which amino acid “charm” is actually getting changed?
Only a few of the charms can be modified like this, the major ones being -> Tyrosine (Tyr or Y), Serine (Ser or S), Threonine (Thr or T) – what do they have in common? A hydroxyl (-OH) group that can ditch the hydrogen and act as a nucleophile and take that phosphoryl group to make a phosphate PO₄³⁻ group. note: Histidine (His or H) can also (less commonly) be phosphorylated, but on a Nitrogen (N)
But under cellular conditions, there are plenty of H⁺ around (the cellular pH is above their pKa), so these charms don’t really want to give up a H⁺ because that would make them negatively charged (-O⁻) -> enter the kinase -> when the protein substrate binds, there’s a basic amino acid just waiting to pull it off. And when it does, you get a strong nucleophile (the substrate’s -O⁻ really doesn’t like being negative and seeks out something that wants more electrons to share their extra with-> so it attacks the gamma phosphate conveniently sticking out right next to it).
Hold on – didn’t you say phosphate has “extra” electrons and is – charged? Why does it want more? The phosphate group is ➖ overall BUT the central P is actually partly ➕ because the electronegative (electron-hogging) Os pull electrons (e⁻) away from it & metal or + charged charms in the binding site can pull on this charge too -> makes the P electrophilic(electron-loving)
The exposed -O⁻ can therefore attack it -> this forms a temporary pentacovalent intermediate (P bound to 5 Os) and then the bond between Β & γ phosphates is severed -> you end up with ADP (Adenosine DiPhosphate) and a phosphorylated protein
Now these products have fewer favorable contacts with the protein so they’re less tightly bound & can “fall out” so the kinase can do it all again but the enzyme has to “open up” to allow this (another conformational change) & this is often the slowest, “rate-limiting” step
Many kinases can phosphorylate both Serine AND Threonine (which are both pretty small) but Tyr & His are bulky, so they require “more space” in the active site so they need different kinases. Thankfully, there are lots of different kinases (~2% of the recipes in the cookbook). Thats a lot of kinases & and they do a lot of work. ~1/3 of all proteins in your cells likely have at least 1 phosphorylation & proteins can be phosphorylated at multiple sites, by one or multiple kinases. Phosphorylation can be hierarchical -> (e.g. site 2 doesn’t get phosphorylated until site 1 does – sometimes because phosphorylation of site 1 makes it a “match” for a second kinase to phosphorylate site 2) But “what’s the point”?
When proteins are phosphorylated, they now have a highly concentrated, bulky, charge -> this can cause the protein to change its shape (conformation) (e.g. maybe the negative charge is repulsive so the atoms around it try to “move away” – but because all the protein’s amino acids are linked together, changes in one can have a “ripple effect” with parts of the protein that have a more defined shape (structural domains) often moving as a group and more flexible “linker” regions serving as hinges. You can also have the opposite – if there are + charged amino acids around, they’ll be glad to hang out around the phosphate and can “clamp down” around it & oxygens in the phosphoryl group can hydrogen-bond (H-bond) with group(s) on the protein for a similar “clamping” effect
These conformational changes can activate or inactivate the proteins (e.g. they might bring the catalytic amino acids together to form an active site that wasn’t there before – Gilmore Girls flashback – a lap is an illusion…)
The added phosphate also changes a protein’s “appearance” to other proteins and molecules, potentially opening up new binding opportunities. Lots of proteins involved in cell signaling have specific domains that seek out & bind to phosphorylated residues.
Sometimes it’s another kinase that seeks it out -> kinases can phosphorylate other kinases, activating those kinases which then phosphorylate other kinases activating those kinases that then… leads to a phosphorylation cascade that can quickly transmit signals -> and it amplifies the signal along the way because each kinase can activate multiple copies of the kinase it activates.
That can make way too strong a signal! You need activation regulation! Some kinases have regulatory regions “built in” as “regulatory domains” attached to the “catalytic domain” where the transfer takes place. Others have a separate regulatory subunit (a whole nother protein as opposed to 2 parts of the same protein)
And the signal can be “turned off” by other proteins called phosphatases that remove the phosphates
If the regulation is impaired, kinases can go wild modifying proteins -> tons of proteins end up always on or always off; signal’s on all the time, etc. which can clearly cause problems – in fact, it can cause cancer. which is why kinases are a frequent drug target.
An example of this is Gleevec™️ (imatinib) – a targeted cancer treatment that revolutionized the treatment of a form of leukemia called chronic myeloid leukemia (CML) which causes excess immature white blood cells. Gleevec is a tyrosine kinase inhibitor (TKI) – specifically it inhibits a mutant kinase called Bcr-Abl
Bcr (breakpoint cluster region) and Abl (Ableson tyrosine kinase) are usually 2 separate proteins, with their genetic recipes on 2 separate chromosomes. But in some patients with certain cancers, the chromosomes get broken and stitched back together incorrectly so that the tips of the 2 chromosomes swas placep. And they do so smack dab in the middle of the genes for those 2 proteins, Bcr and Abl, creating a fusion protein.
When they’re stitched together, Abl loses some of its regulatory parts like an autoinhibitory region and gains parts that allow it to interact with new partners and form dimers and multimers. When they form these multimers, they start phosphorylating each other (autophosphorylation) which helps activate each other. And provides binding sites for other proteins. So those bind, and they get phosphorylated too. So Abl becomes more active than usual and it starts phosphorylating things it normally wouldn’t, so it can set off the”wrong” signaling pathways
ATP binds to Bcr-Abl in a conveniently-placed ATP binding pocket. Well, ATP can bind *unless* somethings blocking the ATP binding pocket. Something like Gleevec (imatinib)… When Gleevec binds there, ATP can’t – they compete for the same site. And if no ATP binds there’s no phosphate to transfer.
But there’s more to the story. Gleevec can’t always bind there – the pocket is only the right shape for it when the kinase is in an “inactive” conformation – one that can’t bind a substrate. In order to bind the substrate, an “activation loop” has to get out of the way of where the substrate needs to bind, taking it from an inactive conformation to an active conformation. Gleevec forms bonds with this loop and if it’s not there Gleevec can’t compete. So Gleevec serves 2 functions by binding in the ATP pocket – it competes with ATP for binding and if no ATP binds there’s nothing to transfer, and it locks the kinase into an inactive state where substrate can’t bind either.
The binding of Gleevec is tight and specific, which is good because your cells have lots of kinases and you don’t want to target them all. In fact, this is largely why kinases were long considered “undruggable”
Gleevec gets this exclusivity because it forms extensive specific interactions with the amino acid (protein building block) parts sticking out into the binding pocket.
Change any of those amino acids and you change the Gleevec-binding activity. And this is precisely what happens in some patients, causing them to develop resistance to the drug. Many patients who develop resistance, however, do so because of mutations in other areas of the protein that cause the protein to remain in an activated conformation (open to bind both ATP & substrate). And this activated form can’t bind Gleevec because Gleevec is “optimized” for binding the inactive form. But later drugs have been designed to treat resistance Bcr-Abl and there are now multiple Bcr-Abl inhibitors on the market. here’s a recent review article of where things stand: https://journals.lww.com/hemasphere/Fulltext/2020/10000/Chronic_Myeloid_Leukemia_in_2020.22.aspx
If you want to learn more, the story of imatinib – which was invented by Nicholas Lyndon at Ciba-Geigy in the late 1990s and developed as a CML treatment by Brian Druker who began clinical trials started in 1998, pushing through to FDA-approval in May 2001 – is detailed in a great book, The Philadelphia Chromosome: A Genetic Mystery, a Lethal Cancer, and the Improbable Invention of a Lifesaving Treatment by Jessica Wapner
I got to hear Brian Druker speak at the ASBMB conference a couple years ago (which was AWESOME!) http://bit.ly/2KgFKL5
link to the paper for the figures: http://cancerres.aacrjournals.org/content/62/15/4236.long
more about enzymes: http://bit.ly/enzymecatalysis
more about enzyme inhibitors: http://bit.ly/enzymeinhibition