This episode covers a PNAS study reporting the de novo design of miniprotein inhibitors targeting porcine deltacoronavirus (PDCoV). The lead minibinder, MB11, binds the PDCoV RBD with picomolar affinity, broadly neutralizes diverse deltacoronaviruses, and resists multiple biochemical stresses.
0:00Welcome to Base by Base, the paper cast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. So, uh, picture this for a second.
0:09A child in Haiti develops a sudden, severe fever. Right. And the clinical team steps in, they test for the usual suspects, dengue, malaria, you know, the common endemic pathogens, nothing. The results just come back totally negative.
0:24Each is always a terrifying moment for a clinician. Exactly. What they don't realize right away is that the culprit is a virus that up until very recently only existed in birds and pigs, and we have 0 approved vaccines for it.
0:360 therapeutics, our medical cabinet is basically empty. Completely bare. So here's the situation for you, listening right now. We have a virus from an entirely different animal family, actively figuring out how to adapt to human biology.
0:50What really happens when a pathogen like this starts knocking on our cellular doors, and we have absolutely nothing on hand to stop it. It is a massive problem. Right. And more importantly, how could we completely rewrite our global strategy from reacting to pandemics as they unfold to computationally designing and stockpiling custom built, perfectly tailored biological defenses before the wider outbreak even begins?
1:14I mean, that transition moving from reactive medicine to proactive, computationally designed barriers is, uh, arguably the most critical shift happening in modern biotechnology right now. We are moving from hoping we find a natural defense to actually engineering an artificial one from the ground up.
1:30Yeah, and that brings us to the core of our deep dive today. Today we celebrate the work of Nathan G. Avery, David Wiesler, and their collaborative teams at the University of Washington, the University of Utah School of Medicine, and Vera Biotechnology, who have advanced our understanding of pandemic preparedness through the computational design of antiviral mini proteins.
1:49Right. And we are drawing these insights from an incredibly detailed study. It was published in the proceedings of the National Academy of Sciences, or PNAS, on April 29, 2026. This research really maps out a blueprint for handling emerging zoonotic threats before they escalate into global crises.
2:07So let's set the stage here a bit. Most of us hear the word coronavirus, and our minds immediately jump to the beta coronaviruses, you know, SARS, MERS, COVID-19. They have completely dominated the global conversation and, well, our funding priorities for the last few years.
2:23But this research targets a different branch of the family tree entirely. Yeah, so the Coronavirde family actually has 5 genera. This deep dive focuses on the delta coronavirus genus, or decovies. Birds are the natural evolutionary reservoir for decos.
2:38They have circulated in avian populations for a very long time. Okay, so it starts in birds. Right. But the specific pathogen we are dissecting today is Porsian Delta Coronavirus, or PD Cove. It first jumped onto the radar when it was discovered in pigs in Hong Kong back in 2012.
2:53And it didn't just quietly circulate, right? I mean, it caused massive devastating outbreaks in agricultural pig populations globally. It did. And the real problem is the mechanism it uses to break into cells.
3:05PD Covey hijacks a specific cellular receptor called a minopeptidase N, or APN. APN got it. Yeah, the viral spike protein seeks out that APN receptor and uses it as the physical anchor to pull itself into the host cell.
3:19And recently, we have seen multiple documented zoonotic transmissions. Meaning actual spillovers. Exactly. Spillovers of PD CoV into humans in Haiti. So the virus is actively mutating its spike protein to better interface with the human version of that APN receptor.
3:36Wow, which really brings us to the core mission of this research for you listening today. The world desperately needs broad spectrum entry inhibitors. We need a way to jam that interaction completely. So are we basically trying to forge a universal biological keyhole blocker before we even know exactly which door the virus is going to try to unlock next?
3:54That is exactly what they are doing. Instead of relying on traditional antibodies, the researchers here utilize de novo computational protein design. They basically decided to build a completely synthetic defense from scratch.
4:07So not just finding something in nature but building it. Right. Because traditional drug discovery often relies on screening 1000000s of existing natural compounds and just, you know, hoping one of them happens to interfere with the virus.
4:19De Novo design completely flips the script. The researchers use an advanced computational tool called Binecraft. They used it to mathematically calculate the optimal shape of a protein that would bind to the receptor binding loops of the PD Covey spike protein.
4:35So they are essentially looking at the three-dimensional topography of the viral spike. You know, the ridges, the valleys, the precise atomic structure, and they're using generative AI to hallucinate a novel protein that perfectly cups that specific shape.
4:49Perfectly, yes. They are designing what they call mini binders. mini bongers. Right. These are ultra-compact synthetic proteins. They range from just 54 to 190 amino acids in length. Oh, wow, that is small.
5:00It is. For context, a standard human antibody is a massive complex molecule made of over a 1000 amino acids. These mini binders are tiny, stripped down homing missiles. That is wild. So the software generated 101 different digital blueprints that pass the initial thermodynamic scoring.
5:21But having a digital blueprint is only step one. Oh, absolutely. I have to ask, it's one thing to design a protein on a computer, but how often does a software prediction like Alpha Fold 3 actually translate perfectly to a physical working protein in the real world?
5:36Well, that is the $10000 question, isn't it? So to test that, they deployed alpha fold three. Alpha Fold 3 is a deep learning model that predicts how a linear string of amino acids will fold into a three-dimensional structure based on the fundamental laws of physics and chemistry.
5:50The researchers fed their 101 hallucinated sequences into alpha fold 3 to predict not just their individual shapes, but exactly how they would physically dock with the viral spike protein. Like a digital simulation of the docking process.
6:02Exactly. And based on those structural predictions, they narrowed the field down from 101 to just 16 of the most promising candidates. All right, so Alpha Fold 3 says this digital blueprint works. But you can't just hit print on a 3D printer for a protein.
6:17You need to hijack a biological factory, right? Like using yeast or bacteria to actually grow it. Yeah, so you introduce the genetic instructions for these 16 mini binders into E. coli bacteria. The bacteria, read the synthetic DNA and manufacture the proteins for you.
6:32Just brewing them? Essentially, yes. And once the researchers harvested these physical mini proteins, they face their 1st major experimental hurdle. They had to prove the proteins actually stick to the viral spike in reality, not just in a computer simulation.
6:47Right, which they tested using bio layer and therometry, or BOI for short. Yes. And the way this works is really cool. They attach the synthetic mini binder to the end of a microscopic glass tip, right?
6:58And they submerge that tip into a solution filled with the viral spike protein and then pass a light beam through the tip. It is highly precise. Yeah, because as the viral proteins stick to the mini binder, It actually changes the thickness of the biological layer on the tip, which alters how the light reflects back.
7:15It gives you a real time, highly precise measurement of binding affinity. It proves the physical attraction is there, but, you know, sticking to a glass tip in a controlled buffer solution is late years away from stopping a dynamic live virus from infecting a complex living cell.
7:30Oh for sure. So the researchers had to prove functional neutralization. But without accidentally causing a lab leak of a dangerous pathogen, I assume. Right, safety first. To test live infections safely, they engineered a surrogate.
7:43They took vesicular stomatitis virus or VSV. It is a relatively harmless virus, often used in labs, and they genetically modified it, so that its outer shell displayed the dangerous PD Covey spike protein.
7:55Oh, wow. So they gave a harmless virus, the exact biochemical exterior, the pathogen they wanted to stop. Exactly. So if the mini binders can neutralize this surrogate from entering human cells in a patriot dish, it practically proves they can block the real deal.
8:09So they ran the gauntlet. Out of those initial 101 digital dreams, one absolute champion emerged in the physical tests, a synthetic mini protein they named MB 11. Let's look at the actual performance data for MB 11, because this is where the computational predictions meet biological reality.
8:28The research is reported that MB11 binds to the PD CoV receptor, with a KD of 155 pick a molar. Yeah. And it functionally neutralized the live surrogate virus with an IC 50 of 216 Pekamolar. Those numbers are incredible.
8:43Right. But for those of us who aren't in the lab every day, can you translate pico molar affinity? Just how sticky is this mini protein compared to standard antibodies? Sure. So a standard high quality natural antibody produced by the human immune system might bind to its target in the anomolar range.
8:58MB 11 is operating in the Picamolar range. Okay, and a Pekamol is... Pekamol is 1000th of an animal. So when we look at a KD, which is an equilibrium dissociation consonant of 155 pico molar, we are looking at an affinity that is orders of magnitude tighter than natural antibodies.
9:13It literally acts like molecular super glue. Molecular super glue. I love that. Wait, you mentioned KD and IC 50. You explain that KD is basically how tightly it acts like Superglue. But what is IC 50 actually measuring in this context?
9:27Right. So the IC 50 is the half maximal inhibitory concentration. It measures the functional stopping power. Okay. If you have a population of cells exposed to the virus, the IC 50 tells you exactly how much of your MB 11 mini protein you need to add to cut the viral infection rate exactly in half.
9:44The lower is better. much better. An IC 50 of 216 pickle molar means you need an astonishingly minuscule amount of this protein to completely crash the virus's ability to infect cells. It isn't just sticking to the virus.
9:57It is functionally incapacitating it. That is amazing. But, you know, if you're listening to this and thinking, well, wait, if this virus mutated to jump from pigs to humans, when it just mutated spike protein to escape this MB 11 protein, you are exactly where the researcher's heads were at.
10:13Exactly. We need something that covers a broad spectrum of mutations, not just the one specific strain circulating today. Right. And the researchers knew that. So they threw distantly related Delta coronaviruses at MV 11.
10:23They tested it against Sparacove and Municove just to map its breadth. And I looked at the sequence data for Miniko before we started. It shares only 59% of its spike sequence identity with PD Kofi, just 59%.
10:36Very low. That is like recognizing the melody of a song. Even when it is being played on a completely different instrument in a totally different time signature. The core shape of the threat is the same, and MB 11 still recognizes it.
10:49It does. It neutralized them all. In head-to-head laboratory tests against a state-of-the-art neutralizing natural antibody called PD 33, the synthetic MB 11 outperformed it, it outperformed it in both absolute potency and cross-strain breadth.
11:03Wow. It jammed the cellular entry mechanism for viruses that had diverged significantly on the evolutionary tree. So to understand exactly how MB 11 achieves this broad neutralization, the researchers use cryolelectron.
11:15microscopy or cryo-EM. Yeah, an incredible tool. They flash froze the viral spike, physically bound to MB 11 and bulb barted it with electrons to capture a 2.8 Angstrom resolution map of the interaction.
11:28They wanted to see the molecular straitjacket in action, basically. And the CryoEM map revealed the physical mechanics of the block. MB 11 perfectly wedges itself onto the exact receptor binding loops that the virus relies on to connect with the human APN receptor.
11:43It sits right in the way. Exactly. It physically occupies the space, creating what we call hysteric clash. The viral key cannot enter the cellular lock because MB 11 is welded over the keys teeth. The structural data also provided the ultimate validation for the computational design, didn't it?
11:58When they align the physical structure captured by the cryo EM microscope, with the original alpha fold 3 digital prediction, They found an incredibly tight .6 Angstrom variant. Yeah, that is mind blowing.
12:10Right. A .6 angst from variants is almost unfathomable. I mean, for context, a single hydrogen atom is about one angstrom wide. Yeah. So the computer predicted this shape and how it would mechanically lock onto a complex viral spike down to the subatomic level.
12:24It is stunning accuracy. And that brings us back to your listener's theoretical question about viral mutation. To rigorously test the virus's ability to evolve a workaround, the team employed deep mutational scanning.
12:39They took the genetic code for the viral spike and systematically forced it to undergo every single possible amino acid mutation at the binding interface. They literally mapped 10s of 1000s of microscopic evolutionary moves.
12:53So they preemptively simulated the future evolution of the virus, just to see if any possible mutation could shake off the NBA 11's straightjacket. Yes, and they found that NBA 11 imposes an exceptionally high barrier to viral escape.
13:04The mechanism relies on evolutionary biology. MB 11 targets what we call the conserved core of the viral spike. Conserved core. Right. These are the critical engine parts that the virus absolutely must maintain to interface with the human APN receptor.
13:18Oh, I see. So the virus is trapped in a biological checkmate. If it mutates those specific amino acids to make MB 11 fall off, it simultaneously destroys its own ability to grab onto the human cell. Exactly.
13:29It can't escape the drug without breaking its own mechanism for infection. It's brilliant. But uh, the final major hurdle for any therapeutic is durability. Right? If we are talking about stockpiling these as a global defense, they cannot be fragile.
13:44No, they can't. And massive natural antibodies often require strict cold chains storage, and they degrade easily under stress. So the researchers subjected MB 11 to a brutal biochemical stress test. They heated it to 70 degrees Celsius.
13:58That is 158 degrees Fahrenheit for a full hour. And it still retained its binding affinity. Which is fantastic. But also, because PD Covey is an enteric virus, meaning it primarily targets the gastrointestinal tract, any effective treatment really needs to survive the human gut.
14:13The gut is a harsh environment. Very harsh. They soaked MB 11 in highly acidic conditions, dropping the ph to 2.2 for 2 hours to simulate stomach acid. And then they exposed it to aggressive digestive enzymes like trypsin, pyma trypsin, and elastase.
14:28The actual molecular machinery our bodies use to violently tear proteins apart for digestion. Yeah, and MB 11 just shrug them off. Its ultra compact, tightly folded synthetic structure. makes it incredibly resilient to heat, acid, and enzymatic degradation.
14:43So when we cool back and look at the broader clinical and policy implications here. MB 11 sort of shatters a fundamental long-standing rule in virology. It really does. Because historically, there has always been a strict trade-off between potency and breath.
14:58If an antibody is a highly potent neutralizer. It is usually hyper specific to one viral strain. But if an antibody is broad enough to catch multiple variants, its overall stopping power is usually significantly weaker.
15:11MB 11 achieves extreme ultropotency while maintaining broad coverage across an entire viral genus. And the practical applications for global health are immense. Because MB 11 is so resilient to gastrointestinal stress.
15:24It opens the door to oral or mucosal delivery systems. You know, you don't necessarily need a sterile intravenous drip in a hospital setting. You could potentially deliver it as a lyophilized powder. Just freeze dried into a pill.
15:36Or package it in lipid nanoparticles. And the manufacturing economics are completely transformed, aren't they? Oh, absolutely. Traditional monoclonal antibotors require massive, expensive bioreactors filled with mammalian cells, and those are incredibly delicate and slow to grow.
15:52But synthetic mini proteins can be grown in standard E. coli fermentation tanks. It is essentially the same technology we use to brew beer, but engineered to yield high purity pharmaceuticals. Exactly.
16:04You can scale production to 1000000s of doses rapidly and cost effectively. It completely circumvents the massive supply chain bottlenecks we see with traditional biologic drugs. Of course, we do need to outline the current limitations and the necessary next steps for this technology.
16:20The study operates in vitro and in controlled surrogate models. This specific mini binder hasn't been subjected to human clinical trials yet. No, it hasn't. The paramount question for future studies is human immunogenicity, because MB 11 is an entirely synthetic molecule, a shape that does not exist in nature, we have to evaluate whether the human immune system will eventually recognize it as a foreign invader and mount an immune response to clear it from the bloodstream.
16:45Right. Though it is worth noting that previous research on similarly designed mini proteins suggests their ultra-compact size gives them a very low immunogenic profile. They are often too small and stable to trigger a massive immune alarm.
16:59That's the hope. And looking forward, the research team proposes several ways to increase the barrier to escape even further. They are exploring homo oligumeric designs. Okay, what does that mean? This involves chemically linking multiple MB 11 units together to create a multi-pronged molecule that binds to multiple points on the virus simultaneously.
17:19It leverages avidity to lock it down even tighter. Oh, that makes sense. They also suggest creating a therapeutic cocktail, because MB 11 wasn't the only successful mini binder they found, right? They also isolated variants like MB5 and MB7 from that initial digital screen.
17:33Right. So if you deploy a cocktail of 3 distinct synthetic mini proteins, each locking down a slightly different mechanism on the concert core, the mathematical probability of a virus mutating to escape all 3 simultaneously drops effectively to zero.
17:47It's incredible. We now possess the capability to map a viral family. computationally predict its vulnerabilities, hallucinate perfectly matched inhibitors, and validate them structurally down to the atomic level.
17:59This means we have crossed a threshold where we can computationally print and stockpile cheap, resilient therapeutics for entire families of viruses before they even fully adapt to humans. It is a complete paradigm shift.
18:12So to distill the core insight for you today. MBLM is an ultra potent computationally designed mini protein that neutralizes a broad spectrum of delta coronaviruses by perfectly blocking their cellular entry mechanism.
18:23Its extreme synthetic resilience to heat, acid, and digestive enzymes makes it an ideal, highly scalable candidate for proactive stockpiling against futures and not ex spillovers. Beautifully summarized.
18:35Thank you. But I want to leave you with a thought-provoking question. What does this mean for the future of human health? When our 1st line of defense against an emerging pandemic, isn't a scramble for a vaccine, but a library of predesign proteins waiting on a shelf.
18:49And more importantly, if artificial intelligence can computationally design a perfect subatomic biological defense in a matter of weeks. How close are we to a reality, where bad actors can use those exact same accessible tools to design the perfect, entirely synthetic offense?
19:05This episode was based on an open access article under the CCBY 4.0 license. You can find a direct link to the paper and the license in our episode description. If you enjoy this, follow or subscribe in your podcast app and leave a 5 star rating.
19:20If you'd like to support our work, use the donation link in the description. Now stay with us for an original track created, especially for this episode. And inspired by the article you just heard about.
19:29Thanks for listening and join us next time as we explore more science based by base.
