Lafora disease is a progressive myoclonus epilepsy with an early fatal issue. Two genes were identified thus far, the mutations of which cause the disease. The first one, EPM2A, encodes the consensus sequence of a protein tyrosine phosphatase. Its product, laforin, is the object of the present work. We analysed in detail the amino acid sequence of this protein. This suggested, as also observed by others, that it could present two domains, a carbohydrate-binding domain (CBM20, known as a starch-binding domain) and the catalytic domain of a dual-specificity protein phosphatase. We produced the enzyme as two different GST-fused proteins and as an N-terminally His-tagged protein. Differences in solubility were observed between the constructs. Moreover, the N-terminal carbohydrate-binding domain contains a thrombin cleavage site, which is hidden in the simplest GST-fusion protein we produced, but was accessible after introducing a five-residue linker between the engineered cleavage site and the enzyme N-terminus. The two types of constructs hydrolyse pNPP and OMFP with kinetic parameters consistent with those of a dual-specificity phosphatase. We show in addition that the protein not only binds glycogen, but also starch, amylose and cyclodextrin. Neither binding of glycogen nor of beta-cyclodextrin appreciably affects the phosphatase activity. These results suggest that the role of the N-terminal domain is rather that of targeting the protein in the cell, probably to glycogen and the protein complexes attached to it, rather than that of directly modulating the catalytic activity.