EPISODE TRANSCRIPT
Protein Folding: Introduction to Proteins (Amit Kessel & Nir Ben-Tal)
Automatically transcribed from the episode audio. Some words and names may be inaccurate.
Listen to the episodeDownload textDownload captions
Imagine for a moment an entire universe of tiny incredibly complex machines operating within your body right now. These aren't you know the cogs and levers you might picture but rather intricate biological origami constantly folding and unfolding performing millions upon millions of tasks every single second. They're involved in everything from say the subtle twitch of your finger to the incredibly complex thought processes humming away in your brain. You quite literally rely on these molecular marvels for every single process that keeps you alive, thinking, and, well, thriving. So let's dig into this. Today, our deep dive takes us into the astonishing world of proteins. Our guiding text, Introduction to Proteins, Structure, Function, and Motion, makes it really clear. These aren't just important molecules. They're absolutely fundamental to life itself. It's no exaggeration to say that protein science sits at the very epicenter of modern biological research. We're talking applications spanning medicine, agriculture, biotechnology.
It's everywhere. Our mission in this deep dive is to really unravel how the precise, intricate, three -dimensional shapes of these proteins dictate their functions, and why understanding what scientists call the protein folding problem is one of the most crucial frontiers in science today. It's truly a puzzle with massive implications. What's fascinating here is that our mission goes beyond just grasping the basics. We aim to give you a really comprehensive understanding of protein folding. And yeah, we'll delve into its technical intricacies, but hopefully in a way that's accessible and engaging for everyone. Then we're going to connect those sort of nitty gritty details to some of the most exciting. real -world applications that are happening right now, like how this knowledge is helping design groundbreaking drugs, discover powerful new antibodies, and even create the next generation of AI tools that could, frankly, revolutionize biology. At its very core, this whole pursuit of understanding structure, dynamics, and function right down at the atomic level, well, that's precisely what we call structural biophysics. It's kind of like deciphering the universe within us, you know, one molecule at a time. The marvel of proteins. Yeah. Structure and function unveiled. Okay, let's start right at the beginning then. The preface of our source material lays it out pretty starkly. Proteins are highly complex molecules that are actively involved in the most basic and important aspects of life. When you really stop and think about that, it's an incredible statement, isn't it?
What exactly are we talking about when we say basic and important aspects of life? Well, metabolism, all chemical reactions, just keeping things running, movement too, right? Everything from tiny cellular transfer mechanisms all the way up to like... Your muscles contracting. Proteins form the very backbone of your body's defense system, fighting off invaders constantly. They act as cellular communicators, sending vital signals, and they're responsible for molecular recognition, knowing exactly what to interact with and what to ignore. It's just an astounding array of duties for, well, a single class of molecule. Yeah, and if we connect this to the bigger picture, it becomes abundantly clear that protein functions aren't just random, right? They are direct, unavoidable results of their very specific, exquisitely crafted three -dimensional structure. And not just the structure, but the subtle dynamic ways those structures move and change. We call these movements structural dynamics. So to truly understand how a protein performs its function, maybe it's catalyzing a reaction or transporting something or recognizing a pathogen. We absolutely have to characterize the physical forces acting on these proteins, and we need to do that right down at the atomic level. This profound insight, this... fundamental realization that structure dictates function.
That's precisely what gave rise to the whole field of structural biophysics. Our source really emphasizes this with what it calls an energy -oriented approach, which basically means we're not just asking what proteins do, but how the intricate dance of molecular interactions and the thermodynamic change as the energy shifts actually provide the power that drives their actions. It's all about understanding their unique energy landscape. You know, the hills and valleys that guide their every single move. OK, so to really navigate that energy landscape, we have to start with the fundamentals, the building blocks. Our book calls them nature's alphabet, which I kind of like. And that alphabet is found in Chapter 2, the 20 natural amino acids. Every single protein, no matter how big or small, is just a chain made up of these 20 unique links strung together like beads on a necklace, basically.
Each amino acid shares a common core structure. a central carbon atom the alpha carbon bonded to a hydrogen an amino group alpha amino a carboxyl group alpha carboxyl and then crucially a unique side chain And it's this side chain, and only this side chain, that gives each amino acid its distinct chemical personality, its role. Right, and these properties are absolutely critical for how proteins eventually fold up and do their jobs. While the central backbone, all the atoms except that side chain, is chemically identical across all amino acids, it's that unique side chain that really differentiates them, makes them special. Now, you mentioned those alpha -carboxyl and alpha -amino groups. At physiological pH, so around 7 in most of your body, the alpha -carboxyl group, it has a low pKa. Which means it tends to give up a proton easily, right? It's negative. Yep, it's protonated to negative. And conversely, the alpha amino group has a high pKa, so it tends to hold on to its protons, stays protonated, positive charge. This charge behavior is super important for individual amino acids just floating around. But when they link up in the protein chain... These backbone charges mostly get neutralized when they form those peptide bonds, except, crucially, right at the very beginning and the very end of the protein sequence, the termini. Exactly. And that side chain diversity you just highlighted, that's precisely what enables such an astonishingly vast range of functional variation. It really does raise an important question, doesn't it? How does this relatively small alphabet, just 20 letters, manage to create millions of distinct, highly specialized proteins, each with its own job? Well, it's like nature's molecular Lego set, isn't it? A brilliantly efficient system. These subtle chemical differences in the side chains act like personalities, allowing proteins to adopt this almost infinite array of shapes and functions, from rigid scaffolding to dynamic enzymes. And these side chains, they're broadly categorized into five basic types based on their chemical properties. non -polar, polar in charge, polar charge, aromatic, and then glycine, which is kind of its own thing. Each type brings unique properties to the table. Take cysteine, for instance. It's unique because its thiol group has sulfur, and it can form disulfide bonds, these strong covalent links. Super important for stabilizing proteins, especially outside the cell. It can also bind metals. Methionine, another sulfur one, is more involved in weaker polar interactions, and it can even act as an antioxidant. Then you have serine and threonine.
They've got hydroxyl groups, and they're absolutely vital for regulation through phosphorylation, adding that phosphate group in glycosylation, which is adding sugar chains. It affects solubility, recognition, lots of stuff. Tyrosine, with its aromatic ring and hydroxyl group, it can also be phosphorylated, sulfated, both key modifications. And its ring helps stabilize this phenylite ion, giving it a peak high around 9 .8. Makes it subtly basic. And what's even cooler, I think, is how the local environment inside a protein can dramatically tweak these p -kyan values. You know, we often think of p -kyan as a fixed property, but inside the cell, or even in different compartments like lysosomes, which are really acidic, pH 4 .7 versus a cytosol at 7 .2. Well, the local pH, or just having other charged residues nearby, can massively shift a residue's p -kyan. It can cause a residue to become ionized, lose its proton when it normally wouldn't. This is a critical molecular trigger for loads of enzymes, like in serine and cysteine protease, there's often a nearby histidine residue.
static interactions, Coulomb interactions, this histidine helps pull the proton off the serine or cysteine, which makes that residue a potent nucleophile, makes it reactive, ready to attack another molecule and start the reaction. Or, you know, in some redox enzymes, nearby positive lysines or arginines can stabilize the negative charge on a deprotonated cysteine, helping it transfer electrons. It's just elegant, really fine -tuned by nature. Yeah. Elegant is a good word for it. And speaking of those aromatic side chains, phenylalanine, tyrosine, tryptophan, histidine too, they have some really special tricks, don't they? Those unique electron clouds they have above and below the ring, they allow for distinct interactions, like pi -pi interactions.
Short -ranged, but powerful, right? Lose strength really fast with distance, like tiny magnets almost. Responsible for forming these aromatic clusters inside proteins, important for binding. positive side chains like lysine or arginine plus maybe surprisingly aromatic rings can even act as hydrogen bond acceptors weaker ones but still it's interesting the book notes that aromatic amino acids showed up relatively late in evolution maybe that explains their sophistication and why they're common in things like ion channel gates where that precise geometry is needed for opening and closing okay so here's where it gets really interesting how do these individual amino acids link up and then fold into this incredible functional machinery. It's a hierarchy, right?
Yeah. Step -by -step thing. From a simple linear chain to a complex working 3D molecule. First up, primary structure. That's just the sequence, the order of the amino acids. Like DNR, CNHL, that sequence is the primary structure determined by your DNA. Then we move to secondary structure. This is about local repeating folds. The common ones are alpha helices and beta sheets. The secret here is the peptide bond linking the amino acids. It's planar, rigid, because of electronic resonance, limits rotation. So rotation is mainly restricted to just two angles along the backbone. Phi. and psi, these constraints are absolutely key. And to visualize these allowed rotations, scientists use this amazing tool called the Romachandran plot. It basically maps out all possible fip -psi combinations, showing which shapes are okay sterically and which ones are just impossible because atoms would crash into each other. The right -handed alpha helix and beta sheets, they show up as these dominant, stable regions on the plot. Glycine, though, is special. Its side chain is just a hydrogen, so it's way more flexible.
It can occupy more areas on the plot, great for turns, alpha helices, they're like molecular spiral staircases. stabilized by hydrogen bonds inside the helix, between residues that are four positions apart in the sequence. And they can be amphipathic, one side water -loving, one side water -fearing, perfect for sitting in membranes or bridging different environments. There are variations too, right? Like the tighter 310 helix, the wider pi helix, the weird polyprolene helix. Beta sheets, on the other hand, are more extended, like a flat zigzag ribbon, stabilized by hydrogen bonds between adjacent strands. and their side chains stick out perpendicular to the sheet they can run parallel or anti -parallel and they often get this characteristic twist gives them rigidity okay moving up another level in this amazing architecture tertiary structure this describes the unique overall 3d fold of a single polypeptide chain formed by combinations of those secondary bits helices and sheets and loops This specific fold is held together by a whole bunch of forces.
Non -covalent interactions, like van der Waals, forces those tiny fleeting attractions between atoms. Electrostatic interactions, two plus and minus charges attracting. And the big one, the hydrophobic effect, which we'll definitely get into more. And sometimes you get strong covalent disulfide bonds between two cysteines locking things together, especially important for proteins outside the cell. What's really remarkable is how tightly packed the protein's core usually is, almost crystalline. Optimizes those van der Waals forces, makes them strong collectively. But even in that dense core, you can have these small, empty spaces, packing defects, maybe 30 to 100 cubic angstroms. And far from being flaws, they can actually be functional, like creating a path for oxygen and myoglobin to get to its binding site deep inside. Or they can help mediate conformational changes in the whole protein.
It's kind of counterintuitive, isn't it? Shows how what looks like a flaw can actually be a clever design feature, essential for function. These secondary elements, they combine into recurring patterns called motifs, like beta, alpha, beta. And motifs assemble into larger folds. Some are so common and vital, they're called superfolds. like the immunoglobulin fold, the Rossman fold, the P -loop fold, Tim Beryl, you see them over and over again across evolution, often doing really important jobs like binding nucleotides or forming enzyme active sites. And then finally, at the very top of this structural hierarchy, we have quaternary structure. This describes the precise arrangement when multiple polypeptide subunits, multiple folded protein chains, come together to form a larger functional complex.
Think of it like a highly coordinated team. Each member is an individually folded protein, but they work together to achieve something bigger, often more complex. These multi -subunit complexes are mainly stabilized by non -covalent interactions, which is key, actually, because it allows for a dynamic equilibrium. The complex can assemble and disassemble, which is really important for regulation, the driving forces, again largely the hydrophobic effect in electrostatics, often forming these intricate cooperative interaction networks. Arginine, with its positive charge, often acts as a key bridging element, linking subunits through strong hydrogen bonds and salt bridges. Evolutionarily, these complexes often arise through clever tricks like gene duplication.
You copy a gene, the copies diverge a bit, and now you have slightly different subunits that can interact. Or through domain swapping, where bits from two identical proteins actually reach out and grab onto the other copy, stabilizing a larger complex. It's efficient. Subtle mutations in the linker regions can then fine -tune these interactions, making them stable and evolutionarily cheap to produce, leveraging existing folds. Okay, so the 20 standard amino acids are amazing, no doubt. But they alone can't explain the sheer diversity millions of unique proteins out there. That's where post -translational modifications, PTMs, come in, adding entirely new layers of complexity and function after the protein is made. Over 200 different types are known. It dramatically expands the protein alphabet, especially in complex organisms like us. Huge source of diversity in regulation. Phosphorylation is a fantastic example. Such a massive impact from a seemingly small change. It involves adding a bulky, negatively charged phosphate group, usually to serine, threonine, or tyrosine, using enzymes called kinases. This really raises an important question. How does this tiny chemical addition have such a profound effect? The answer lies in the electrostatic field. Adding that big negative charge fundamentally perturbs the local electrostatics. It changes the protein's entire energy landscape, basically pushing it towards new, often lower energy shapes it wouldn't normally adopt. This can trigger huge conformational changes. I mean, in glycogen phosphorylase, adding one phosphate can cause parts of the protein to shift by like... 50 angstroms. That's enormous on a molecular scale. And these shifts then alter the protein's activity, change its binding sites, or tweak the alignment of catalytic residues in an enzyme. It's a molecular switch. A tiny change fundamentally redesigning the protein's behavior. And phosphorylation is just one. The cell has this whole toolbox of modifications, like merestoilation and... palmitoylylation.
These involve sticking on long fatty acid chains, myrcic acid C14, palmitic acid C16. These act like molecular anchors, right? Allowing the protein to embed in or associate tightly with cell membranes. Crucial for getting proteins to the right place in the subcellular localization. Palmitoylylation can also regulate active cysteine residues and enzymes, kind of dynamically controlling their function. And these lipid modifications pop up in disease too. Cancer, diabetes, Alzheimer's, Huntington's shows how important important they are for health. Then there's ubiquitination or isomoyolation, sticking small proteins ubiquitin or isumo onto target proteins. Often this tagging is like a molecular kiss of death, marking the protein for destruction by proteasomes. Proteasomes are like the cell's recycling centers, breaking down old, damaged, or misfolded proteins, so it's a vital... quality control mechanism, ensures only good proteins stick around, also regulates protein levels based on the cell's needs. And we also see glycosylation sticking complex carbohydrate chains, sugars, onto proteins. Very common, especially for proteins on the cell surface or secreted outside the cell. Glycosylation is crucial for making them soluble, protecting them from being chewed up by enzymes, and really importantly for cell recognition, how cells talk to each other. And just like phosphorylation, adding these bulky sugar groups profoundly influences the protein's energy landscape, affects stability, folding, interactions. These diverse PTMs are basically the cell's incredibly clever way of getting maximum functional bang for its genetic buck, creating millions of distinct protein forms from a relatively small number of genes. It's just amazing biological efficiency. The art of folding. A dance of stability and dynamics. Okay, so what does all this intricate detail, amino acids, bonds, PTMs mean for one of the biggest mysteries in biology? The protein folding problem.
This problem, fundamentally, is about figuring out exactly how a protein's linear amino acid sequence dictates its unique functional 3D structure. It's like, how does this long, seemingly random string of beads automatically tie itself into a perfect, intricate knot that can then do a specific job? It's baffled scientists for ages. Yeah, and what's truly fascinating here is that proteins achieve this specific native state, their working, folded shape, that is only marginally stable. meaning it's just slightly more stable energetically than all the unfolded or partially folded states it could be in. This delicate balance is absolutely crucial for function. If it were too stable, it might be too rigid, right? Couldn't move and interact properly. Too unstable, and it would just fall apart. The stability is determined by its free energy. Lower G means more stable, more energetically preferred. And that free energy change comes from the famous equation Liege -AAA -OS. Liege is about changes from bonding and interactions, and the not -a -query term is all about changes in order or disorder entropy. The dominant forces driving this whole folding process are really remarkable, and they all interplay.
The primary driver is widely considered to be the hydrophobic effect. This is largely entropy -driven. It's about increasing the overall disorder of the water around the protein. See, when nonpolar water -fearing residues are exposed to water, they force the water molecules nearby into these ordered cage -like structures, like tiny icebergs. By burying these nonpolar bits inside the protein during folding, you release those ordered water molecules back into the bulk solvent. They become free, disordered, and that increases the overall entropy of the whole system. Which is favorable. There's also a favorable enthalpy bit, too, as the released water can now make better hydrogen bonds in the bulk. But folding isn't free. There's a big configurational entropy tunnelty. As that flexible chain folds into one specific shape, it loses a ton of freedom of movement, becomes much more ordered. This penalty is huge, maybe 1 .2 to 1 .9 kilocalories per mole per residue, roughly. But it's beautifully balanced by the gain from the hydrophobic effect, leading to that characteristic marginal stability we talked about. Beyond the hydrophobic effect, other interactions are definitely key. Van der Waals interactions, those weak fleeting attractions, they're optimized in the tightly packed core. Individually weak, but add them all up in that dense core and they become collectively very strong. Contribute a lot to stability, though maybe secondary to hydrophobic effect as a driving force. Hydrogen bonds are absolutely critical, especially for forming and stabilizing secondary structures like alpha helices and beta sheets and for overall stability. It costs a lot of energy to bury a polar backbone group in the core if it doesn't form a hydrogen bond. Makes the protein unstable. But when they form those nice, well -ordered H -bonds inside the core, shielded from water, that cost drops dramatically. Makes them highly favorable. And finally, electrostatic interactions plus -minus attractions. Most common on the surface, interacting with water and other polar groups. But they can be important inside the core too, especially if shielded from water by other groups, or even trapped water molecules. And their strength really depends on the local environment, the dielectric constant. How well the surroundings dampen electrical forces. It's also really important to distinguish between thermodynamic stability as the folded state, the lowest energy state, and kinetic stability.
Kinetic stability is more about resistance to unfolding. How long does it take for the protein to fall apart? Kinetically stable proteins often tend to be more rigid, maybe lots of beta sheets, which are very sturdy, or extensive long -range interactions locking things down. These features... physically block solvent molecules from getting to weak spots and starting the unfolding process. Also, things like oligomerization forming complexes and certain PTMs like glycosylation can significantly boost kinetic stability, reinforce the structure, make it last longer. Now, folding inside a cell in vivo, that presents a whole different set of challenges, right? The cytoplasm where lots of proteins are made is incredibly crowded, dense.
packed with molecules. So it's really easy for unfolded or partially folded proteins to just bump into each other nonspecifically. Get tangled up. Aggregate, these aggregates, these clumps, they can actually be energetically more stable than the correctly folded protein sometimes. But they're usually inactive. And often they're outright toxic to the cell. It's a big hurdle. Exactly. And this is precisely where chaperones come in. They are the cell's essential, incredibly sophisticated folding assistants. Chaperonins, like the famous Groel Hsp60 complexes, are these large cage -like protein machines. Their main job is to help refold unfolded or misfolded proteins.
They basically grab these substrate proteins, the ones needing help, and encapsulate them inside this protected cage -like structure. By doing that, they provide this isolated safe space where the protein can try to fold or refold without getting tangled up with all the other sticky molecules in the crowded cytoplasm. Prevents aggregation. And it's not passive. It's an active process. Requires energy from ATP hydrolysis to power its conformational changes. The asymmetrical model for Grohl -Grohes shows this. beautifully. Unfolded protein goes into the GroEL cavity, then ATP binding in the GroES cap. protein trap it inside gives it a chance maybe a second or third chance to fold correctly in isolation now there's this interesting ongoing debate about chaperones do they just prevent aggregation letting the protein fold passively on its own time or do they actually accelerate the folding process itself some studies suggest they can maybe speeding it up by as much as 10 times compared to folding alone even without aggregation the idea is that by confining the protein inside the cage they reduce the entropy penalty of folding favor more compact native -like states over randomly unfolded ones. It's a really cool example of how cells optimize a fundamentally tricky physical process, making sure proteins get to their functional shapes efficiently. Yet it's not just about the final shape, is it? It's the intricate dance it performs. Which brings us to protein dynamics and lostry. We really need to shift from thinking of proteins as rigid statues to seeing them as incredibly dynamic, constantly shifting between different shapes or conformational substates.
Driven by the thermal energy just buzzing around, these motions are diverse too. Hinge motions, like parts swinging open and closed, rotations of whole domains, translations, parts sliding, even bits of secondary structure folding and unfolding locally, like myosin, the motor protein. Hinge motions are literally how it converts ATP energy into mechanical force for muscles. Many enzymes use hinge motions too, opening and closing access to their active sites, regulating catalysis. There are even databases, like MoleMoveDB, dedicated to visualizing these amazing movements. Shows proteins are truly living machines. And what's truly fascinating here is how these built -in dynamics lead to allostery. This is a key regulatory thing where binding a ligand, any molecule that binds at one spot, the allosteric site, influences what happens at a distant spot, often the active site. Our understanding here has evolved. From early, more rigid models, like monads, we've moved toward the conformational selection model.
This model suggests proteins aren't just static waiting for a ligand to bind and force a change. Instead, they're constantly, spontaneously sampling different shapes. like a dancer trying out poses. The ligand then preferentially binds to and stabilizes one of those existing shapes, shifting the energy landscape to favor that state. Koshland's induced fit model still plays a role too, describing how sometimes slower adjustments happen to optimize binding, maybe involving latch residues that click into place. Hemoglobin is the classic example, right? Carrying oxygen in your blood. Oxygen binding makes it switch from a low affinity T state tense to a high affinity R state. Relax.
When oxygen binds the iron in the heme group, the iron moves slightly. This pulls on a key amino acid, the proximal histidine, which then pulls on a whole helix segment, the F -helix. This chain reaction changes the whole central cavity, making it easier for more oxygen to bind. Cooperative binding. And there's another histidine, the distal one, that helps stabilize the bound oxygen and actually makes it harder for poisonous carbon monoxide to bind as tightly. Clever design. These allosteric changes have huge functional effects everywhere in biology. They can open or close active sites, like we said, or subtly reshape them to change binding affinity. For instance, in the enzyme DHFR, loops that aren't even part of the active site transmit these small coordinated motions across about eight angstroms to the active site. This tiny motion actually stabilizes the transition state, making the enzyme way more efficient. Another amazing example, the SH2 domain of a kinase. A tiny flip, a cis -trans isomerization of just one peptide bond completely redesigns the protein surface, changes which ligands it prefers to bind. It just hammers home that function isn't just about the static shape. It's deeply tied to dynamic flexibility and responsiveness. And adding yet another layer of... frankly mind -bending complexity we have intrinsically unordered proteins iups or intrinsically disordered regions idrs these are proteins or bits of proteins that just don't have a stable 3d structure on their own in isolation instead they exist as these dynamic ensembles of rapidly interconverting shapes like a constantly shifting cloud This disorder comes from a really flat energy land state. No single shape is strongly preferred. What's super intriguing is that these IUPs often do fold into a stable structure, but only when they bind to a specific partner molecule.
Now, this binding might have an entropy cost, maybe leading to lower affinity than structured interactions, but it also allows for incredibly rapid binding and release. This fast on -off rate is crucial for complex processes like signaling cascades or regulatory networks where things need to happen quickly and transiently. Medically, IUPs are hugely relevant. They're involved in loads of diseases, cancer, neurodegenerative diseases like Alzheimer's and Parkinson's, cardiovascular disease, diabetes, makes them important but also really challenging drug targets. Their flexibility presents unique problems and opportunities for drug design. But sometimes this whole intricate dance of folding just goes terribly wrong. with devastating results. And here's where it gets really interesting in a rather grim way, because the consequences of protein misfolding can be severe, leads to these debilitating, often fatal diseases. We group these under the term amyloidosis.
Over 40 human diseases characterized by proteins clumping together, aggregating into these toxic, fine fibers called amyloids. Yeah, think about really devastating conditions like Alzheimer's disease that involves beta amyloid and tau proteins misfolding and clumping in the brain, or Parkinson's disease characterized by aggregates of alpha -synuclein. Huntington's disease, another neurodegenerative one, involves the Huntington protein. The general idea is that misfolded proteins often expose more hydrophobic, sticky patches on their surface.
This makes them prone to clumping together, aggregating, and precipitating out as these highly ordered toxic amyloid fibrils. Exactly how they're toxic isn't always fully clear. It probably varies. But it's generally thought of as a harmful gain of function the aggregates themselves do damage rather than just losing the original protein's function. But this raises a really provocative question. Are amyloid proteins always bad? There's this fascinating hypothesis from Nobel laureate Eric Kandel about a protein called CPEB. In sea slugs, it seems to be involved in long -term memory, and it aggregates in a controlled way, forming a functional amyloid structure that seems crucial for maintaining the synaptic connections that store memories. So... It highlights a potential evolutionary advantage for controlled aggregation in specific contexts. Maybe it's not always pathological.
It's a powerful reminder that even the dark side of protein folding might have unexpected evolutionary roots and even serve vital functions sometimes. Designing the future. Protein folding's impact on technology and health. Okay, so we've journeyed through this incredible complexity amino acids. folding dynamics even misfolding now let's shift gears let's look at the impact of all this knowledge how is this fundamental understanding of protein folding and dynamics letting us design the future especially when it comes to making better more targeted drugs 3 .1 drug design and action tailoring molecules to targets right so at the heart of pretty much all drug action is protein ligand interaction that intricate dance often called lock and key but we know it's way more dynamic than that right ligands small molecules peptides even bigger macromolecules they bind to proteins to kick off or regulate countless processes catalysis regulation transport signaling you name it and the driving forces they largely echo what we talked about for folding itself The hydrophobic effect is huge, pushing water out of the binding site when the ligand fits in. That's favorable entropy, plus favorable enthalpy from the released water making better bonds elsewhere. Van der Waals interactions contribute a lot to the tight packing, the snug fit, ensuring good affinity.
And electrostatic interactions, plus minus attractions, they're absolutely crucial for specificity. Making sure the drug hits the right target. Highly geometry dependent. When we talk drug design, two things are paramount. Affinity and specificity. Affinity is how strongly the ligand binds. Kuntz and colleagues estimated an upper limit for typical affinity, maybe around 13 kilomole. Enzymes, though, can achieve way stronger binding, but specifically to the transition state of the reaction they catalyze. A brilliant idea from Linus Pauling. By binding super tightly to that high -energy transition state, they speed up the reaction immensely. Then the products release quickly.
Specificity, on the other hand, is all about hitting the right target and only the right target. It comes down to the number of precise, geometry -dependent interactions. But what's truly fascinating here, and actually a massive challenge for drug design, is this thing called binding promiscuity. You know, we often think of proteins as being super specific, but the reality is many proteins can actually bind multiple different ligands. This can happen in a few ways. Different ligands might use the same binding site residues, just with different strengths. Or they might bind to totally different residues nearby. Or maybe use a mix of overlapping and distinct sites.
And it can even involve different dynamic shapes, different conformations of the binding site itself. Maybe certain residues are accessible only in one shape, or their charge changes. This promiscuity. It's tricky. On one hand, it makes designing hyperspecific drugs hard, could lead to off -target effects. But on the other hand, it suggests proteins might be quite tunable. Maybe we can design novel molecules to bind them therapeutically. And maybe some drugs work by hitting multiple targets weekly. It complicates the simple picture, but it's probably closer to reality in a complex cell. Yeah, that complexity is key. And this deep understanding isn't just for drugs targeting existing proteins.
It's also revolutionizing protein engineering for industrial enzymes. Enzymes are vital industrial workhorses, right? Used in food production cheese, beer textiles, making pharmaceuticals precisely. Basically tiny, efficient biological factories, super precise, super fast. When we want to tweak these enzymes for industry, there are two main ways. Rational design or directed evolution. Rational design uses our knowledge of structure. and function plus computer modeling to make specific targeted amino acid changes, like protecting sibilucin, an enzyme, and detergents from oxidation by swapping a key methionine for something like serine or alanine makes it last longer. Directed evolution is more like accelerated natural selection in a test tube. You make random mutations in the gene and screen like crazy for the improved trait you want. Let evolution do the heavy lifting.
And when it comes to engineering enzymes, especially for industry, a really common goal is improving thermostability. making them tougher, able to withstand heat. Based on everything we know about folding, several clever strategies have emerged. One is to regenify the unfolded state. Make it harder to unravel in the first place. You can do that by putting in rigid prolines, getting rid of flexible glycines, adding bulky aromatics for better packing, or even engineering new disulfide bonds to staple things together. Another way is to strengthen the hydrophobic core. Optimize the packing of nonpolar residues, make it more resistant to water getting in. You could also boost electrostatic interactions, add more salt bridges, those plus -minus attractions inside the protein. Promoting alpha helices, which are intrinsically stable structures, can help too.
And finally, optimizing interactions on the surface with the solvent can prevent irreversible aggregation, which is a big problem under stress. These are all rational strategies derived directly from understanding the forces that hold proteins together. Our deep grasp of protein -ligand interactions also underpins the whole drug discovery pipeline, which is, you know, a long, tough road from idea to medicine. It involves several key stages. First, identifying targets proteins involved in disease. GPCRs, G -protein coupled receptors, are huge targets here because they do so much signaling. Then finding hits molecules showing some initial activity. Then leads more promising compounds and tons of lead optimization. Computational drug design is absolutely indispensable now.
One key approach is the pharmacopher hypothesis that's ligand -based. You look at known drugs for a target, identify common features, H -bond donor susceptors, aromatic rings, charges, and use that pattern to design new molecules. Then there's docking, which is receptor -based. You computationally try to fit potential drug molecules into the protein's binding site, predict how they bind, how strongly. Docking is tough, though. You have to sample all the possible shapes of the flexible ligand in the protein. For millions of ligands. That's trillions and trillions of possibilities. Huge computational cost. Once you have leads, lead optimization refines them using more accurate free energy calculations, things like alchemical methods or MMPB, GBSA, trying to boost affinity, improve selectivity, and crucially, reduce toxicity by minimizing off -target binding. Now, how do drugs actually interact with enzymes, often as inhibitors, stopping them working? Competitive inhibitors bind right in the active site, blocking the natural substrate.
They increase the apparent Km, but Vmax stays the same if you flood the system with substrate. Non -competitive or allosteric inhibitors bind somewhere else. They change the enzyme's shape or activity, reducing Vmax or affinity, and substrate concentration doesn't really overcome it. Let's look at examples. Ibuprofen, the NSAID. It inhibits the cyclooxygenic COX by binding its active site, mimicking arachidonic acid, stops inflammatory signals. ACE inhibitors like Captopril, Lisinopril, crucial for blood pressure. They inhibit angiotensin -converting enzyme by mimicking the transition state of its substrate. Blocks the enzyme, lowers blood pressure. Really clever mimicry. And speaking of prime targets, G -protein -coupled receptors, GPCRs, are absolutely central. These amazing membrane proteins, like 20 -30 % of our genome, involved in so much. Our senses, smell, taste.
Internal signaling via hormones, neurotransmitters. Their structure is classic. Seven alpha helices weaving through the membrane. And the mechanism is fascinating. An agonist ligand binds outside, causes a shape change inside. Exactly. And this complex dynamic mechanism makes them incredibly hot targets for drugs. Orthosteric drugs bind where the natural ligand binds, like adrenaline binding to its receptor. Aripiprazole, the antipsychotic, is a partial agonist at the D2 dopamine receptor's orthosteric site. But the really exciting area now, arguably, is allosteric drugs. These bind somewhere else on the receptor, not the main site. and they modulate activity, maybe make the natural ligand bind better or worse, or stabilize an active or inactive state. This raises an important question, though. Why bother with allosteric drugs if we have orthosteric ones? Well, several reasons. They offer an alternative, potentially allowing for more subtle fine -tuned control.
And crucially, allosteric sites tend to be less conserved across different receptor subtypes than the main orthosteric sites. This means allosteric drugs can often be more subtype -selective. Hit only the receptor involved in the disease, hopefully fewer side effects. Maraviroc, the HIV drug, is a great example. It targets CCR5, a coreceptor HIV uses to get into cells. It binds allosterically and stabilizes an inactive shape of CCR5. It doesn't block the main binding site, just locks the receptor so HIV can't use it. blocks a key transmission switch movement. Powerful example of manipulating protein dynamics therapeutically, locking the door on the virus. Wow. Okay. This deep dive into folding also connects directly to our immune system, doesn't it? Especially in antibody discovery and engineering. Antibodies themselves are amazing proteins. About 10 % of our genome is involved in making them and related immune molecules. They, along with T -cell receptors, are the immune system's precision targeting systems, recognizing specific invaders. T -cells recognize specific peptide fragments presented by MHC molecules like a molecular wanted poster. And B -cells produce these incredibly specific antibodies to... It really is. And what's truly fascinating here is this kind of dual nature of antibodies.
They're famous for being highly specific to their target antigen. But their binding sites also show functional versatility. They can sometimes bind multiple related things. Promiscuity again. So how can they be both specific and promiscuous? It's a great question. Research suggests certain residues like bulky tyrosines contribute a lot to the overall binding energy of the affinity. Makes the interaction strong. But it's often the shorter chain polar residues, serines, threonase, aspartates, asparagens, that provide the exquisite specificity. Through precise, short -range, electrostatic interactions and hydrogen bonds that really depend on the exact shape of the antigen fitting perfectly.
So it's this interplay. Strong general binding driven by some residues, fine -tuned specificity driven by others. And understanding this is critical for antibody maturation and engineering. Your immune system naturally matures antibodies over time through somatic mutations, tiny changes in the antibody genes that improve binding affinity. But interestingly, most of these beneficial mutations don't directly touch the antigen. Instead, they cause subtle global conformational changes within the antibody structure itself. These shifts slightly reposition other residues, effectively increasing the number of productive interaction sites with the antigen, making the binding tighter overall.
Like, in one antibody studied, 48G7, maturation involved residues R96 and TR33, forming new interactions with the antigen after these conformational shifts, even though they didn't interact initially. Boosted affinity hugely. This dynamic adaptability is incredible. It means we can now potentially custom -design therapeutic antibodies with enhanced specificity and affinity for all sorts of targets. Imagine engineering antibodies for specific cancer cells or neutralizing specific viruses or toxins with pinpoint accuracy. It's a massive frontier in precision medicine. Engineering our own defenders.
Okay, let's unpack this even further then. Beyond therapeutics, the real cutting edge seems to be where protein science meets AI. artificial intelligence especially in finally cracking that protein folding puzzle how are these new ai tools and prediction methods changing the game for understanding and even designing proteins 3 .3 new ai tools and prediction solving the folding puzzle yeah historically figuring out protein structures experimentally it's been incredibly hard laborious expensive think about membrane proteins vital but super tricky to work with only about 3 .5 percent of known structures back in 2017 were membrane proteins because they're so hard to crystallize for x -ray methods this bottleneck really spurred the development of computational prediction methods ab initio approaches try to predict structure just from the amino acid sequence alone Note templates. Traditionally very tough, computationally heavy, but useful for refining experimental data or for dynamic simulations. Comparative modeling, or homology modeling, has usually been the go -to. Find a similar protein with a known structure, a template, and use that as a starting point. Accuracy really depends on how similar the sequences are. Over 50 % identity. Usually gives a pretty good model.
But it still struggles with flexible loops, often important for function. and getting the side chain positions exactly right. Fold recognition or threading is slightly different. Tries to match a sequence to known folds in the database based on shared properties, secondary structure propensities, solvent exposure, etc. Sees if the sequence fits a known fold. Right. And if you connect this to the bigger picture, the field is rapidly moving towards these incredibly powerful integrated experimental computational approaches. It's a really exciting trend. You take lower resolution experimental data, maybe from cryo -EMs, access NMR, CD, free T stuff, that doesn't give you the full atomic picture on its own.
And you combine it with computational modeling. The experimental data provides crucial geometric constraints, narrows down the search space dramatically. This massively improves prediction accuracy, especially for big floppy complex things like the yeast nuclear pore complex. And cryo -EM itself has had a revolution. Single particle cryo -EM can now get near atomic resolution, like 4 -5 angstroms sometimes. Unimaginable just a decade ago. Molecular dynamics, MD simulations are also playing a huge role. Simulating atoms moving over time. Gets amazing insights into dynamics, folding pathways, function. Early MD studies, like on the small protein BPTI, had to use super simplified models, like two spheres per amino acid. Just not enough computer power.
But now, distributed computing like folding at home, specialized hardware like the Anton supercomputer. We can simulate millisecond long folding events for small proteins. Extract real thermodynamic data. MD relies on these force fields parameter sets describing how atoms interact. Developing accurate force fields is ongoing work. Especially tricky for non -protein molecules, like drugs, and there's always the trade -off with solvent. Model every single water molecule explicitly. More accurate but way slower. Use implicit solvent models. Faster but less precise. Scientists constantly juggle these choices. Which brings us right up to the current revolution, AI's role in smashing through these old barriers. Artificial intelligence, especially deep learning, is just dramatically accelerating structure prediction, hitting accuracy levels that honestly seem like science fiction not long ago. Breakthroughs came from realizing you can predict structure from sequence variation across species. By looking at which residues co -evolve change together, you can infer which ones must be close together in the 3D structure.
provides powerful constraints. And the CASP competition critical assessment of protein structure prediction running since 94 has been a huge driver, blind tests pushing the algorithms forward. The progress in the last few CASP rounds, particularly with things like AlphaFold from DeepMind, has been absolutely staggering. It's fundamentally changed structural biology. Yeah, it really has. So this raises an important question. What does this future of increasingly accurate, increasingly fast protein structure prediction, powered by AI, actually mean for us? The implications are, well, they're vast, transformative, really exciting.
In drug design, it means much faster, much more accurate identification of targets and lead compounds. We might be able to design drugs against targets previously considered undruggable because their structures were unknown or too complex. Opens up whole new therapeutic avenues. For antibody engineering, imagine custom -designing antibodies with unprecedented precision for specific pathogens or specific cancer markers, accelerating immunotherapy and vaccine development hugely. In synthetic biology and protein engineering, we could start creating novel proteins with functions that don't even exist in nature, optimizing enzymes for biofuels, creating new biomaterials. The possibilities are wild. And fundamentally, it means getting much deeper insights into misfolding diseases.
Understand exactly how proteins go wrong, paving the way for targeted therapies that fix the root cause, not just symptoms. being able to predict structure accurately and quickly, unlocking biological design at its most fundamental level. So we've really gone on a journey, haven't we? From the basic building blocks, the amino acids, through that intricate dynamic dance of protein folding, and now to how this fundamental biology is being harnessed to reshape medicine, biotech, even AI itself. It's just a testament to the incredible power of understanding things at the molecular level and, well, human ingenuity. Indeed. What we've unpacked today, understanding this precise language of protein structure and dynamics right down at the atomic level, it isn't just academic curiosity, is it? It genuinely unlocks incredible potential, potential to tackle some of the biggest challenges we face in health, maybe environment, technology. It gives us the tools to not just observe life, but potentially to engineer it with a level of control we've never had before.
So what does this all mean for you listening right now? If every living cell really is this complex symphony orchestrated by millions of perfectly folded proteins all working in harmony, What new symphonies might we start to compose now that we're finally beginning to properly read nature's molecular sheet music? Think about the mind -boggling possibilities. Designing life -saving drugs that seemed impossible just years ago. Creating novel biological systems straight out of science fiction. All by mastering this intricate, beautiful art of molecular origami.
It's a future that's unfolding incredibly rapidly right before our eyes. And it's a future you're now, hopefully, a little better equipped to understand.