Structure-Informed Stress-Testing of the Standard Numbering Scheme for Class B β-Lactamases Using AlphaFold2
This study validates the robustness of the Galleni-Garau standard numbering scheme (BBL) for class B -lactamases across 87 newly modeled families using AlphaFold2, confirming its general applicability while identifying a small subset of variants with structural divergences that warrant experimental follow-up.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Bacteria have evolved a formidable defense against the antibiotics we rely on to treat infections. Among their most effective weapons are enzymes called beta-lactamases, which act like molecular scissors, snipping apart the chemical rings of penicillin and related drugs before they can kill the bacteria. While some of these enzymes use a specific chemical tool to cut, a particularly dangerous group known as metallo-beta-lactamases uses zinc atoms to perform the same task. These zinc-dependent enzymes are a major concern in modern medicine because they can destroy even our most powerful, last-resort antibiotics, leaving doctors with few options for treating severe infections.
To study these enzymes and design new drugs to stop them, scientists need a common language. Because these enzymes come in many different shapes and sizes, researchers created a standard numbering system. This system acts like a universal map, allowing scientists to point to the exact same spot on different enzymes, regardless of how long or short the protein chain is. This map was drawn decades ago, based on a very small number of physical structures that scientists had managed to capture in a laboratory. Today, however, we know of hundreds of these enzymes from bacteria found in soil, water, and hospitals, but we have never seen their physical shapes. The big question is whether the old map, drawn from just a few examples, still works for this vast, unseen world of new enzymes.
A team of researchers set out to test this map using a powerful new tool called AlphaFold, which uses artificial intelligence to predict the 3D shape of proteins based on their genetic code. Instead of waiting years to capture physical structures in a lab, the team used this computer program to build models of 87 different classes of these zinc-using enzymes that had never been seen before. They started by proving their method worked, comparing their computer models of eight known enzymes against their actual, experimentally measured structures. The computer predictions were remarkably accurate, matching the real shapes with a precision so high that the difference was less than half the width of a single atom.
With their method validated, the researchers applied it to the 87 unseen families. They checked if the standard numbering system correctly identified the critical spots where zinc atoms attach to the enzyme. In the vast majority of cases, the old map held up perfectly. The computer models showed that the zinc-attaching spots lined up exactly as the standard system predicted, even for enzymes that had never been seen in a lab. This suggests that the numbering system, created over twenty years ago, is robust enough to guide scientists through the much larger and more diverse family of enzymes discovered since then.
However, the study also found a few interesting exceptions. In eleven of the cases where the map seemed to fail, the researchers discovered a biological reason for the difference, such as the enzyme belonging to a specific subgroup that naturally has a slightly different arrangement. But in two specific families, the computer models revealed something more puzzling. While these enzymes still had the correct sequence of letters to build the zinc-attaching spots, their predicted 3D shapes showed that the active area was twisted or arranged differently than the standard map suggested. These two families, along with one other that showed a similar pattern, represent a small group of enzymes where the physical structure might not match the old expectations.
The researchers conclude that while the standard numbering system is a reliable guide for most of these enzymes, these few outliers are important. They represent a small set of variants where the active site, the very heart of the enzyme's function, may look different in three-dimensional space than we thought. Because the computer models flagged these differences, the study identifies these specific enzymes as high-priority targets for future laboratory work. Scientists will now know to focus their experimental efforts on these few families to confirm their unique shapes, ensuring that our understanding of these bacterial defenses remains accurate as we continue to discover new varieties in the wild.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.