Extreme conservation and rare evolutionary deviations in mammalian NOTCH3 inform the interpretation of CADASIL pathogenicity

Inside the walls of our blood vessels, a molecular gatekeeper stands watch. This gatekeeper is a protein called NOTCH3, a long, complex chain that stretches across the surface of cells, acting as a sensor and a signal transmitter for the body's vascular system. In humans, when the instructions for building this protein go wrong, a severe hereditary disease called CADASIL can develop. This condition damages the tiny blood vessels in the brain, leading to strokes and dementia. For decades, scientists have known that the most dangerous errors in the NOTCH3 instructions involve the loss or alteration of specific chemical building blocks known as cysteines. These cysteines act like tiny staples, folding the protein into a precise, rigid shape. If a staple is missing or moved, the protein can crumple, clump together, and clog the system. The prevailing belief has been that these cysteine staples are so critical that any change to them is almost certainly a disaster for the cell.

However, nature is rarely as simple as a single rule. To understand how rigid these rules really are, a team of researchers looked beyond the human genome. They gathered the genetic blueprints for the NOTCH3 protein from 113 different mammal species, ranging from mice and monkeys to whales and bats. By comparing these blueprints, they were searching for a rare phenomenon: animals that have survived with broken or missing staples in their NOTCH3 proteins. If such animals exist, they offer a natural experiment, showing how a protein can change its shape and still function, or perhaps fail in a way that teaches us something new about human disease. The researchers combined this massive genetic comparison with advanced computer modeling to see how these natural variations affected the protein's structure and behavior.

The study began by confirming just how incredibly similar the NOTCH3 protein is across the mammalian world. In nearly all 113 species examined, the protein's structure was almost identical to the human version, with the cysteine staples appearing in exactly the same places. This extreme conservation suggests that evolution has kept this design unchanged for millions of years because it works so well. Yet, buried within this sea of similarity, the researchers found three striking exceptions where nature had taken a different path. These were not random errors, but specific, stable variations found in jaguars, Brandt's bats, and several other mammals, including some that carry a version of the protein also found in human families with CADASIL.

The first surprise came from the jaguar. In the region of the protein corresponding to the 13th, 14th, and 15th repeating units, the jaguar's NOTCH3 protein is missing eight cysteine staples that are present in every other mammal. In fact, the entire 14th repeating unit is stripped of its stabilizing staples. In a human, such a massive loss of structural support would likely cause the protein to collapse and clump together, triggering disease. But the jaguar is alive and healthy. Computer simulations of the jaguar's protein showed that without these staples, that specific region becomes floppy and disordered, losing its tight, folded shape. Surprisingly, however, this floppy region did not show an increased tendency to clump or aggregate. This finding challenges the simple idea that losing a cysteine staple always leads to immediate clumping. It suggests that the surrounding sequence of the protein matters just as much as the staples themselves, and that in the jaguar, the protein has found a way to remain soluble even while its shape is altered.

A second natural variation was found in Brandt's bat. This species has a deletion in a part of the protein called the negative regulatory region, which normally acts as a safety lock to keep the receptor turned off until it receives a signal. In humans, this lock is very delicate, and if it breaks, the receptor can turn on by itself, causing problems. The bat's deletion removes a small piece of this safety lock. Computer simulations of the bat's protein interacting with the enzyme that cuts the receptor showed that the cut site was more exposed to the enzyme than in humans. The distance between the enzyme and the cut site was shorter in the bat model, suggesting the safety lock is looser. While the researchers could not prove the bat's receptor is constantly active, the simulation indicates that this natural deletion makes it easier for the receptor to be activated. This provides a real-world example of how changing the shape of the safety lock can alter how easily a cell receives a signal, offering a new way to study the fine-tuning of this system.

The third variation involved a version of the protein where a specific section, known as exon 16, is missing. This "X1" deletion removes eight cysteine staples and is found in several mammals, including great apes and elephants. Interestingly, a human family with CADASIL also carries this exact deletion. In the human context, this variant causes the disease, leading to the accumulation of toxic material in blood vessel walls. The researchers modeled this deletion and found that even though the protein might form a new, makeshift connection between two remaining staples to replace the lost ones, the overall structure still became more prone to clumping. This is a crucial insight: even if a protein manages to patch a broken staple with a new one, it does not necessarily return to a healthy state. The loss of the original structural context is enough to cause trouble. The fact that this same deletion exists in healthy mammals but causes disease in humans highlights that the background of the protein—the rest of the sequence surrounding the deletion—plays a massive role in determining whether a change is harmless or deadly.

The study does not claim that these animal variations are perfect models for curing human disease, nor does it suggest that the jaguar or bat versions of the protein are "better" than the human one. Instead, it uses these natural deviations to show that the relationship between a protein's structure and its function is more complex than a simple checklist of missing parts. The researchers demonstrated that a change in the number of cysteine staples does not always lead to the same outcome; sometimes it causes clumping, sometimes it causes disorder, and sometimes it changes how the protein is activated. The context of the surrounding sequence and the specific location of the change are just as important as the change itself.

By looking at how nature has tolerated these changes in other species, the researchers have provided a new framework for understanding human CADASIL. They show that not every cysteine mutation is automatically a catastrophe, and not every structural change leads to disease in the same way. The jaguar's floppy protein, the bat's looser safety lock, and the human family's clumping variant all tell different parts of the same story: the health of a cell depends on a delicate balance of shape, stability, and context. These natural experiments offer scientists new, testable ideas about how to predict which genetic changes will be harmful and which might be tolerated, moving the field beyond a simple count of missing staples toward a deeper understanding of how proteins truly work.