AI-Powered Drug Discovery Identifies Promising MvfR Blocker Against Drug-Resistant Pseudomonas aeruginosa

Drug-resistant infections are becoming one of the biggest health threats worldwide. The United Nations Advisory Committee predicts that these infections could push 24 million people into extreme poverty by 2030 and cause up to 10 million deaths each year by 2050.

Pseudomonas aeruginosa is responsible for approximately one in 10 hospital-acquired infections. The bacterium poses the greatest threat to people who are already vulnerable, including intensive care patients and those with weakened immune systems.

Why Pseudomonas aeruginosa Is Difficult to Treat

Researchers have identified two major reasons why Pseudomonas aeruginosa infections are so difficult to treat. First, the bacterium is naturally resistant to many antibiotics and can develop additional resistance during a single course of treatment.

Second, it can form a slimy protective layer called a biofilm. Biofilms can develop on medical devices such as catheters and ventilators, helping protect the bacteria from antibiotic treatment.

Targeting Bacterial Communication With MvfR

Scientists have also discovered that Pseudomonas aeruginosa communicates through a chemical signaling system called quorum sensing. This system allows the bacteria to sense the number of nearby cells and collectively switch on genes that contribute to infection. In other words, the bacteria can become more harmful when they act together.

A protein called MvfR is a central component of this quorum-sensing system. Previous research has shown that removing MvfR reduces the bacterium’s ability to cause harm, making the protein an attractive target for new drugs.

Using Modern Bacterial Genomes to Find New Drug Candidates

Researchers at the University of New South Wales reported that previous efforts to identify compounds capable of blocking MvfR had two important weaknesses.

First, many studies relied on older DNA and protein data and did not account for the genetic differences found in bacterial isolates collected from patients today. Second, many screening studies focused mainly on predicting how strongly a compound would bind to a target protein, without determining whether the compound had properties expected of an actual drug in the human body.

To address these limitations, the researchers built target proteins using genomes from recent clinical bacterial isolates. They then combined several layers of analysis to go beyond initial binding predictions.

The team downloaded whole-genome sequence data for clinical isolates of Pseudomonas aeruginosa from the NCBI public archive. Using an artificial intelligence structure-prediction tool called AlphaFold2, they constructed a three-dimensional model of the MvfR protein.

Computer Screening of 7,199 Compounds

To identify compounds that could block this version of the MvfR protein, the researchers used a chemical database to create four compound groups, labeled A through D. Together, the groups contained 7,199 compounds associated with four known antibiotics.

The researchers used computer software to dock each compound to the MvfR protein and rank the compounds according to their binding scores. The compound with the highest binding score in each group was labeled A1, B1, C1 or D1, respectively.

Testing Whether the Compounds Could Work as Oral Drugs

The research team next used multiple computer programs to predict how well the leading compounds might perform as oral medicines. The analysis examined questions related to drug safety and stability, including:

• How well would the human intestine absorb each compound?
• Would the compound penetrate the brain, which could indicate a potential safety concern?
• How effectively would the compound be excreted from the body after treatment?
• Would it remain bound to the target protein under abnormal stress levels?
• Could it cause harmful side effects?

Compound A1 Emerges as a Promising Starting Point

The researchers reported that many compounds in groups C and D violated one or more safety criteria, while many compounds in groups A and B did not violate any of the tested criteria.

According to the study, A1 achieved the highest overall scores across the safety and stability assessments. The researchers concluded that it could serve as a starting point for developing MvfR blockers against Pseudomonas aeruginosa.

However, the findings are based on computer modeling and prediction. The researchers said their study was the first to identify this particular compound and that the discovery could provide a foundation for laboratory experiments to test and confirm its performance in living systems, including cell lines and animals.

Views: 64