The ocean may seem like an unlikely place to mine uranium, but the valuable metal is actually plentiful there. Scientists estimate that the total quantity of uranium in seawater is roughly 4 billion tons, 1,000 times as great as all terrestrial sources combined. But the metal, present in water as uranyl ion (UO22+), cannot be collected and used to fuel nuclear reactors unless selective extraction methods can be designed to nab the highly dilute element. The challenge is that many other metal ions are present at overwhelmingly greater concentrations. Various methods have been studied, but a front-runner has not yet emerged. University of Chicago chemists Lu Zhou and Chuan He and coworkers have now put a new contender in the race. The team screened thousands of proteins computationally, searching for ones with binding pockets of suitable size, shape, and charge for sequestering UO22+. They customized the best candidate to boost its UO22+ uptake and tested it under various conditions. The group reports that the protein binds UO22+ with unprecedented femtomolar-level affinity, and it exhibits selectivity that is 10,000 times as great for UO22+ as it is for other ions in seawater (Nat. Chem. 2014, DOI: 10.1038/nchem.1856). The protein is thermally stable and can be used repeatedly, the team reports.