Their idea, stylized in the picture, was to first connect the center of a dumbbell to a ring, then use nucleophilic attack (specifically, aminolysis) by the other half of the dumbbell to form the final rotaxane. Hiratani did something similar to this back in 2002.
The ring is a phenolic crown ether with side groups attached. Why the phenol? Because the first step is an esterification involving the phenolic OH (which points inside the ring) and the half axle, a benzoyl chloride. The resulting "half-rotaxane" is just a benzene-stoppered ester of the crown ether. To form the final rotaxane, an amine with a bulky group attached is added to the half-rotaxane. Nitrogen attacks the carbonyl carbon of the ester group, and voila! The axle becomes an amide and the ring reverts back to a phenolic crown ether. Miraculously, this reaction requires no heat or special treatment of any kind aside from column chromatography. Put the stuff together, stir, wait three days, evaporate the solvent, and voila! Rotaxane city.
