Virtual Screening of the p62/SQSTM1–KEAP1 Interaction Interface in KRAS-Mutant Colorectal Cancer: Pan-Kelch Binding and a p62-Selective Aminobenzothiazole Scaffold

Inside the cells of our bodies, a delicate balance determines whether a cell survives a toxic assault or succumbs to it. One of the key players in this drama is a protein called NRF2, which acts as a master switch for the cell's defense systems. Under normal conditions, another protein named KEAP1 holds NRF2 in check, tagging it for destruction so the cell does not overreact. However, in certain aggressive forms of colorectal cancer, a different protein called p62 hijacks this system. It grabs onto KEAP1 and pulls it away, freeing NRF2 to run unchecked. This runaway defense mechanism helps cancer cells resist a specific type of cell death known as ferroptosis, a process that destroys cells by letting them rust from the inside out. When cancer cells become resistant to ferroptosis, they become much harder to kill with existing treatments.

Scientists have long suspected that if they could find a small molecule—a drug-like chemical—that stops p62 from grabbing KEAP1, they might be able to restore the cell's natural ability to die when under attack. The challenge lies in the architecture of the proteins involved. Both p62 and NRF2 bind to the same general area on KEAP1, a flat, wheel-shaped surface. It is like trying to find a key that fits into a specific lock on a door, only to discover that the lock is so similar to the one next to it that any key opening one will likely open the other. If a drug blocks the wrong spot, it might disrupt the cell's normal defenses rather than helping them. The question researchers faced was whether it is possible to design a molecule that targets only the p62 binding site without disturbing the NRF2 site, or if the two sites are simply too similar to tell apart.

A researcher working independently set out to answer this question by running a massive digital simulation. Instead of testing chemicals in a lab, they used a computer to model how thirty-eight different compounds would interact with the KEAP1 protein. The team focused on the specific region where p62 attaches, known as the KIR site, and compared how well each compound fit there versus how well it fit the nearby NRF2 site. They examined a diverse collection of molecules, ranging from small fragments designed to probe the surface to more complex structures inspired by known drugs. The goal was to see if any single compound could stick tightly to the p62 spot while ignoring the NRF2 spot, effectively acting as a selective blocker.

The results of this digital screening revealed a surprising reality, but with a major caveat. For the vast majority of the compounds tested, the computer could not distinguish between the two sites. When a molecule fit well into the p62 pocket, it almost always fit just as well into the NRF2 pocket. Out of the thirty-eight compounds, thirty-four showed nearly identical binding strengths for both locations. However, the researchers explicitly warned that this result might be an artifact of the simulation method itself. The computer models used two different-sized "search boxes" that overlapped significantly, which can force the software to report similar scores for both sites regardless of whether the drug truly prefers one over the other. Because of this methodological flaw, the conclusion that a "pan-Kelch" inhibitor (a drug that blocks both sites) is the best strategy is currently provisional and unproven. The researchers stated clearly that this reading is the most likely interpretation based on the data, but it must be confirmed by re-running the simulation with perfectly matched search boxes before it can be treated as a fact.

However, the study did identify one standout candidate that broke the pattern. A compound named Arg_Anchor_004, which features a specific chemical structure called an aminobenzothiazole, showed a modest preference for the p62 site over the NRF2 site. It bound more tightly to the target the researchers wanted to block. This molecule passed standard checks for drug safety and quality, making it a promising starting point for future development. The researchers noted that the slight difference in binding strength was small and required further testing to confirm, but the structural reason for this preference was clear: the molecule formed a specific connection with a part of the protein that exists only in the p62 binding area. This finding offers a concrete lead for chemists who might try to build better versions of this molecule in the future.

The study also highlighted another compound, JKE-1674, which appeared to interact with multiple targets involved in cell death, including the KEAP1 protein and other enzymes that manage cellular stress. While the computer simulations suggested this molecule could hit several different spots, the researchers were careful to frame this as a hypothesis rather than a proven fact. They emphasized that these were digital predictions that would need to be verified in a wet lab. The work serves as a map of the chemical landscape, showing that while finding a perfectly selective drug for this specific protein interaction is difficult, there are viable paths forward. The most immediate takeaway is that the two binding sites appear nearly indistinguishable in this specific simulation, but this conclusion is heavily dependent on the imperfect setup of the computer model. Until the simulation is corrected, the idea of a drug blocking both sites remains a hypothesis, while the single compound with a slight preference offers a new direction for those seeking a more targeted solution.