Protein enables better lanthanide separations

Lanthanides are an essential part of modern life, used in permanent magnets, electronics, and catalysts. But these 15 rare earth elements are chemically very similar, making them difficult to separate from one another after they’ve been extracted from ores.

Researchers led by Joseph A. Cotruvo Jr. at The Pennsylvania State University have now taken a crucial step in their quest to purify lanthanides using a natural protein called lanmodulin (LanM), which could offer a more efficient and environmentally friendly approach to isolating the metals. They have discovered a new type of LanM that is better at differentiating between lighter and heavier lanthanides than previous forms of the protein and unraveled the intricate mechanism behind its ability, paving the way for further improvements (Nature 2023, DOI: 10.1038/s41586-023-05945-5).

“The research is beautiful,” says Scott Banta, a protein engineer at Columbia University, who was not involved in the work. “This really charts a path towards designing biomolecules to discriminate between the very, very small differences in the lanthanide ions.”

Lanthanides typically form +3 ions, and the lighter ions such as neodymium are only slightly larger in size than the heavier ions such as dysprosium. Conventional separation methods use toxic phosphonate ligands and organic solvents to extract the elements from aqueous solution, potentially requiring hundreds of wash cycles.

LanM proteins can perform these separations more efficiently because they carry loops of amino acids called EF hands, which are rich in carboxylate groups that coordinate to the ions. A LanM produced by the bacterium Methylorubrum extorquens AM1 grabs light lanthanides about 5 times better than heavy lanthanides, thanks to tiny differences in how it coordinates the ions.

In previous experiments, Cotruvo’s team loaded this protein, Mex-LanM, into a separation column and used it to tease apart lanthanide mixtures dissolved in dilute aqueous acid. The researchers achieved high purities after just a couple of passes through the system.

The researchers have now trawled through a genome database to discover almost 700 sequences that encode other LanM proteins. One of these comes from Hansschlegelia quercus, a bacterium found in English oak tree buds. Although this protein, Hans-LanM, doesn’t bind lanthanides quite as well as Mex-LanM, it is much better at discriminating between them, with an almost 40-fold higher affinity for lighter lanthanides than heavier lanthanides.

X-ray crystallography revealed how this works. When Hans-LanM binds heavier lanthanides, the ions coordinate to 9 of the protein’s oxygen atoms, which are part of a wider network of hydrogen bonds. In contrast, lighter lanthanide ions coordinate to 10 of the protein’s oxygens. This strengthens the binding and also shifts one of the protein’s carboxylates, a movement that is amplified through the hydrogen-bonding network. Those adjustments ultimately enable two entire protein molecules to embrace and form a dimer, locking the ions firmly in place and thus providing an additional boost to the protein’s preference for light lanthanides.

The researchers tested a mutant version of Hans-LanM in a separation column, giving it a mixture of neodymium and dysprosium that mimicked the composition of electronic waste containing rare-earth magnets. The mutant separated more than 99% of each element from the simulated waste at over 98% purity after a single pass through the column—a big improvement on the Mex-LanM-based separation.

“Being able to separate them completely, with high yield and recovery, is remarkable,” says Banta, who also works on protein-based lanthanide separation.

It’s unlikely that the Hans-LanM proteins stuck to the column are actually able to form dimers, Cotruvo says, so the team is now tethering pairs of Hans-LanM together in a way that would help them to get the additional selectivity afforded by dimerization. They also plan to mine their trove for other LanMs and tweak them to produce variants that offer even better separation.

Whether this will lead to a cost-effective alternative to conventional lanthanide separation methods remains to be seen: “It comes down to how many times you can run the process” through the same column, Cotruvo says. But he and his team have already used a separation column containing the previous Mex-LanM almost 100 times with no loss in performance, and “we are working towards scaling this up,” he says.