Scientists at the VIB–VUB Center for Structural Biology have uncovered a counterintuitive principle that could reshape how membrane proteins are designed from scratch: Sometimes, making a protein less stable helps it fold correctly. In their study published in the Proceedings of the National Academy of Sciences, the researchers demonstrate that introducing carefully placed “imperfections,” a strategy known as negative design, enables synthetic membrane proteins to fold and assemble efficiently in artificial membranes.

Membrane protein stability

Membrane proteins are essential for life and biotechnology, acting as gateways, sensors, and drug targets. Yet designing them from scratch remains notoriously difficult. Unlike soluble proteins, they must navigate a complex folding process while inserting into lipid membranes and during this step, many designs fail.

Traditional protein design focuses on maximizing the stability of the final folded structure. But the new study shows that, for transmembrane β-barrel proteins, this approach can backfire.

Using a cell-free protein synthesis system combined with synthetic lipid vesicles, the team found that highly optimized designs often misfold and aggregate instead of inserting into membranes.

“Designing for maximum stability alone can actually trap these proteins in the wrong state,” says first author and Ph.D. student Giacomo Pedrelli (VIB-VUB). “They become too eager to fold too early, which leads to aggregation in water before they ever reach the membrane.”

The power of ‘negative design’

To overcome this, the researchers introduced subtle destabilizing features to disrupt premature folding. This negative design strategy reduced aggregation and significantly improved membrane insertion and assembly. Remarkably, these changes did not substantially compromise the final stability of the proteins. Instead, they helped guide the folding pathway, ensuring the protein reached the membrane in a foldable state.

The study also revealed that a protein language model (ESM3), trained on evolutionary data, outperformed traditional physics-based methods in identifying beneficial negative design mutations. While conventional tools predicted these mutations would destabilize the protein, the AI model successfully pinpointed changes that improved assembly in membranes.

The ability to reliably design transmembrane β-barrels opens exciting possibilities. These proteins can form nanopores—tiny channels with applications in biosensing, molecular detection, and next-generation sequencing technologies.

“This work shows that we need to think beyond static structures,” says Prof. Anastassia Vorobieva (VIB-VUB). “By designing not just the final state, but taking into consideration the entire folding journey, we can unlock new possibilities for engineering functional membrane proteins.”

This negative design approach for designing membrane proteins could accelerate the development of synthetic proteins for biotechnology, medicine, and nanotechnology.

A research team has discovered an enhanced CRISPR gene-editing system that could enable targeted delivery inside the human body—a key step toward broader clinical use. Researchers identified a naturally occurring enzyme, Al3Cas12f, that is small enough to fit into adeno-associated virus vectors, a leading targeted delivery method for gene therapies. They then engineered an enhanced version that dramatically improved gene-editing performance in human cells.

The advance addresses a major limitation in CRISPR technology. Commonly used gene-editing proteins are too large for targeted delivery systems, restricting clinical applications to cells modified outside the body, such as blood and bone marrow.

“Smart delivery of gene editing systems is a powerful notion with broad clinical implications, and this basic science finding takes us a significant step toward that future,” said Erica Brown, Ph.D., acting director of NIH’s National Institute of General Medical Sciences (NIGMS).

Using imaging and machine learning tools, researchers at the University of Texas at Austin analyzed the enzyme’s structure. They found it forms a more stable and tightly connected complex than other enzymes of a similar size, allowing it to function more effectively in human cells.

“The expanded interface means the enzyme is much more stable. Compared to the others we looked at, Al3Cas12f basically comes preassembled and ready to go shortly after its pieces are produced,” said David Taylor, Ph.D., a molecular bioscience professor at UT Austin and corresponding author of the paper published in Nature Structural & Molecular Biology.

The team then engineered a variant, known as Al3Cas12f RKK, which significantly improved editing efficiency from less than 10% to more than 80% across tested targets. In a commonly edited region of the genome, efficiency reached 90%.

Of the many variants the team produced, Al3Cas12f RKK stood above the rest. The team introduced instructions for RKK directly into a line of human cells originally isolated from a patient with leukemia. Mutations in several of the genes they aimed to edit were associated with diseases such as cancer, atherosclerosis, and amyotrophic lateral sclerosis (ALS).

The authors expect to build on their encouraging results. They next plan to conduct tests of the nuclease’s performance when packaged into AAV vectors, which, if successful, could bring gene editing therapy for many diseases much closer to reality.

New findings reveal that obesity significantly impaired the quality and longevity of antibody responses to a Pseudomonas aeruginosa vaccine in a mouse model. The impaired antibody production was due to defects in germinal centers, a transient part of the immune system where specialized immune cells, called B cells, produce antibodies and build memory against pathogens.

Researchers say the findings, which are published in The Journal of Immunology, provide an important reason for why traditional vaccines, which rely on high antibody production, tend to underperform in people with obesity.

“We hope these findings shift the focus of vaccine design and lead to more effective, tailored vaccines for the millions of people living with obesity who are at higher risk for severe respiratory infections,” said Wendy L. Picking, Ph.D., Professor in the Department of Pathobiology and Integrative Biomedical Sciences at the University of Missouri and lead author of the study.

Unexpected strength of lung T cells

Though the antibody response was decreased, the vaccine did generate a strong response from lung tissue-resident memory T cells. These specialized cells live permanently in the lungs and do not circulate through the bloodstream.

In response to the P. aeruginosa vaccine, resident memory T cells provided early, critical protection against infection that was not observed in mice fed a normal or low-fat diet. This suggests that the tissue-resident memory T cells could be compensating for antibody deficiencies.

“Instead of just trying to boost blood antibody levels, we should intentionally design vaccines that prioritize tissue-resident immunity, ensuring protection directly where pathogens like Pseudomonas enter the body,” shared Dr. Picking.

Growing urgency for better vaccines

P. aeruginosa is a leading cause of severe pneumonia for people with obesity, and emerging antibiotic resistance increasingly makes the infection difficult to treat, highlighting the need for effective vaccines.

To date, no other studies have examined the effectiveness of vaccines targeting gram-negative bacterial pathogens, like P. aeruginosa, in people with obesity. Understanding the relationship between obesity and the immune system addresses a significant gap in current vaccine research.

Next steps in vaccine development

The researchers plan to build on these findings by identifying the specific molecular signals that enable the lung tissue-resident memory T cells to become activated despite the chronic inflammation associated with obesity. This could allow researchers to optimize vaccine formulations to further boost these resident memory cells.

Ultimately, the researchers seek to create a vaccine that ensures robust protection for all individuals, regardless of their metabolic health.

In many parts of East Africa, small pools of water that form after heavy rainfall are ideal breeding sites for the Anopheles mosquitoes that transmit malaria. Researchers at the Karlsruhe Institute of Technology (KIT) have analyzed how such environmental conditions affect the effectiveness of mosquito nets. They combined high-resolution climate and hydrology models with malaria data from Kenya to enable better assessments of when and where the nets are especially effective at preventing infections. Their results have been published in Scientific Reports.

More than 600,000 people die of malaria in sub-Saharan Africa every year. How widespread the disease becomes depends not only on medical care and preventive measures but also on environmental factors such as rainfall, temperature, and especially the formation of temporary bodies of water.

“Such pools of water determine where Anopheles mosquitoes breed and increase the risk of infections,” said Professor Harald Kunstmann of the Institute of Meteorology and Climate Research Atmospheric Environmental Research (IMKIFU) at KIT’s Campus Alpin in Garmisch-Partenkirchen. “Thanks to today’s high-resolution environmental models, we know exactly when and where that occurs.”

With his team, Kunstmann has investigated whether and how such data can be used to maximize the effectiveness of countermeasures. “One of the simplest tools for fighting malaria is mosquito nets that protect people from mosquito bites at night,” said Dr. Diarra Dieng of the IMKIFU, who was a major contributor to the project. “We wanted to find out how much they actually reduce transmission, and where their use has the greatest impact.”

From rain to infection: A modeling chain

The researchers combined various model types for their study, with climate models providing temperature and precipitation data and hydrological simulations showing where water can accumulate to form potential breeding sites. Based on these data, an epidemiological model predicts the resulting spread of malaria. The analysis was based in part on malaria data from Kenya.

“Our approach is the first to consider the entire chain, from atmospheric processes to the formation of breeding sites to disease transmission, enabling us to make the first experimental determination of how effective mosquito nets really are at reducing infections,” Dieng said.

The researchers quantified the extent of changes in malaria transmission and incidence under different environmental conditions with and without mosquito nets. They were able to show that systematic use of mosquito nets significantly reduced the number of infectious insect bites, causing malaria incidence to decrease by around 40% on average, and in some regions by over 50%.

They also showed the extent to which trends were influenced by local environmental factors. Temperature, precipitation, and the availability of temporary breeding sites determine when and where mosquitoes can breed most successfully, which in turn determines the effectiveness of preventive measures.

Planning targeted preventive measures

The study shows how climate data can be used for practical health care decisions. High-resolution environmental data make it possible to assess malaria risk with much greater geographical precision and to estimate the expected benefits of preventive measures. Health programs could use this information to identify regions where targeted intervention would be especially effective and where additional measures might be needed.

“For the first time, we have data that show what really helps,” Dieng said. “If we understand the relationships between environmental conditions and preventive measures, we can put limited resources to better use.”

Research and innovation in Texas A&M University’s biomedical engineering department often centers around clinical impact on patients. Beyond the lab, however, some faculty are finding breakthroughs in the classroom.

Dr. Charles Patrick, professor of practice, published findings in the journal Biomedical Engineering Education. The study details his success implementing a scaffolded inquiry-based learning model in the classroom. Patrick found learning outcomes improved through an approach that allowed students to practice design throughout the semester before a summative final design project.

“Students were highly engaged and it’s been well published that the more engaged a student is, the more they learn,” Patrick said. “They worked in small groups, which helped develop their teamwork and communication skills. We also measured their imagination competency at the beginning of the semester using validated surveys. This increased when measured at the end of the semester.”

Patrick tested the approach in a course titled “Biomimicry, Biomimetics, and Bioinspired Approaches to Medical Device and Technology Design.” The aim is for students to use nature as a model to solve engineering problems.

“This class is focused on teaching students how to not necessarily start with a blank sheet of paper, but to look at nature and see how it has optimized or influenced some aspect of medical design,” he said. “We can make an exact copy of nature, or emulate it, or just be inspired by it.”

Biomimicry is responsible for some of the world’s most recognizable inventions. For example, bur seeds inspired Velcro and whale fins inspired wind turbines.

Texas A&M biomedical engineering researchers Drs. Taylor Ware and Abhisek Jain each found inspiration in nature. Ware developed self-assembling polymers inspired by fire ants, while Jain created vessel-chips by mimicking human microvasculature.

“Nature has already optimized the energy and mechanics of processes while we’re still struggling to make medical devices efficient,” Patrick said. “We are looking to nature to teach us how to optimize engineering.”

The course is one of three design frameworks the department offers students to build competency and gain experience before their final capstone class, where they work with companies to refine or design a real biomedical device.

“This course is a nature-inspired framework,” Patrick said. “Another is with NASA—a space engineering design framework—and the third is the department’s core biodesign framework. Students can choose to learn all three different types of design frameworks.”

To build familiarity with design principles in biomimicry, Patrick used three lessons to increase competency before a final project. The first assignment used LEGO sets as a low-risk way for students to familiarize themselves with the process through a method they already understood. He took inspiration from other universities as well as his own children at home using LEGO Serious Play as a means of learning.

For the assignment, the students are tasked with grasping a yellow sphere. Patrick gives them a curated list of “critters” to choose from—like a hawk—to ideate how they might grasp the object. The students then decide how they’ll emulate or copy that function.

“I wanted something they’re familiar with that still stoked creativity,” Patrick said. “There are enough different parts that they could mimic what they see in nature. It lowers activation energy due to their familiarity. They get to be creative, but still learn new concepts at the same time.”

After the LEGO project, students used virtual reality tools to redesign surgical instruments. Finally, they took a trip to The Gardens at Texas A&M University to draw inspiration for their final project.

“Whether it’s a bee, a certain flower or plant, or an animal they see, that’s their inspiration for their final design project,” Patrick said. “Every time I’ve done it, the students say this is the first time they’ve ever actually gotten outside of the classroom to do an assignment. They enjoy being able to think differently and look at things from another perspective.”

Patrick hopes the course leaves students with a newfound confidence in using their imagination in engineering.

“Something our education system does well in K-12 is making sure that students use their imagination and creativity,” Patrick said. “When students get to college, for some reason, we often stop that. The greatest thing is a blank sheet of paper and a pencil, unlimited imagination and brainstorming. This class helps them do that.”

Engineered cells are a high-value genetic asset that is key to many fields, including biotechnology, medicine, aging, and stem cell research, with the global market projected to reach $8.0 trillion USD by 2035. Yet the only ways to keep the cells safe are strong locks and watchful guards.

In Science Advances, a team of U.S. researchers present a new approach to genetically securing precious biological material. They created a genetic combination lock in which the locking or encryption process scrambled the DNA of a cell so that its important instructions were non-functional and couldn’t be easily read or used.

The unlocking, or decryption, process involves adding a series of chemicals in a precise order over time—like entering a password—to activate recombinases, which then unscramble the DNA to their original, functional form.

The researchers conducted an ethical hacking exercise on the test lock and found that random guessing yielded a 0.2% success rate, remarkably close to the theoretical target of 0.1%.

Turning the assets into locks
The U.S.’s Centers for Disease Control and Prevention (CDC) and Department of Homeland Security have reported an uptick in the theft and smuggling of high-value biological materials, including specially engineered cells. In recent years, there has also been a record rise in unauthorized shipments and attempts at industrial espionage. In the wrong hands, these materials could be misused to create bioweapons or deliberately harm the environment.

Currently, valuable cells are primarily protected by physical measures such as locks, cameras, and guards. Once these barriers are breached, there is little left to prevent the cells from being stolen and misused.

In this study, researchers used a cybersecurity-inspired approach to protect cells at the DNA level by using the cell’s own biological security system. They developed a scenario-based simulation using a designing group (blue team) and a decrypting group (red team).

First, the development (blue team) scrambled the DNA by rearranging and flipping genetic instructions so the cell could no longer read them correctly. They started with a functional genetic unit that includes a promoter (the ON switch) and the gene of interest. They then broke this unit into separate parts, arranged them in the wrong order, and flipped some segments backwards.

To make sure these can be unscrambled later, they placed special DNA sequences called recombinase attachment sites around them.

For decryption, the team used a precise sequence of chemicals to trigger the cell’s machinery to physically rearrange the scrambled DNA and restore it to its functional state. They created a biological keypad with nine distinct chemicals, each acting as a one-digit input.

By using the same chemicals in pairs to form two-digit inputs, where two chemicals must be present simultaneously to activate a sensor, they expanded the keypad to 45 possible chemical inputs without introducing any new chemicals. They also added safety penalties—if someone tampers with the system, toxins are released—making it extremely unlikely for an unauthorized person to access the cells.

Then the red team, who kept out of the encryption process development, stepped in as ethical hackers, tasked with trying to break into the system and access the hidden genetic information. In their first attempt, they uncovered 10 different chemical combinations that partially unlocked the cells, revealing weak spots in the design.

After the developers patched these flaws, the hackers tried again. This time, only the exact passcode worked, showing that the odds of an unauthorized person guessing it had dropped to just two in 990, or 0.2%.

The researchers note that the strong performance of this biological lock signals a shift in biological security, in which genetic material is protected by safety algorithms built into the DNA itself, making the assets their own protectors.

This study designed the system around engineered E. coli cells, but further research is needed to determine whether it can be applied to other organisms and scaled to protect multiple genes or assets within a single cell.