It has been shown [1,2,3,4] that the following glycosyl hydrolases can be, on the basis of sequence similarities, classified into a single family:

• β-glucosidases (EC 3.2.1.21) from various bacteria such as Agrobacterium strain ATCC 21400, Bacillus polymyxa, and Caldocellum saccharolyticum.
• Two plants (clover) β-glucosidases (EC 3.2.1.21).
• Two different β-galactosidases (EC 3.2.1.23) from the archaebacteria Sulfolobus solfataricus (genes bgaS and lacS).
• 6-phospho-β-galactosidases (EC 3.2.1.85) from various bacteria such as Lactobacillus casei, Lactococcus lactis, and Staphylococcus aureus.
• 6-phospho-β-glucosidases (EC 3.2.1.86) from Escherichia coli (genes bglB and ascB) and from Erwinia chrysanthemi (gene arbB).
• Plants myrosinases (EC 3.2.1.147) (sinigrinase) (thioglucosidase).
• Mammalian lactase-phlorizin hydrolase (LPH) (EC 3.2.1.108 / EC 3.2.1.62). LPH, an integral membrane glycoprotein, is the enzyme that splits lactose in the small intestine. LPH is a large protein of about 1900 residues which contains four tandem repeats of a domain of about 450 residues which is evolutionary related to the above glycosyl hydrolases.

One of the conserved regions in these enzymes is centered on a conserved glutamic acid residue which has been shown [5], in the β-glucosidase from Agrobacterium, to be directly involved in glycosidic bond cleavage by acting as a nucleophile. We have used this region as a signature pattern. As a second signature pattern we selected a conserved region, found in the N-terminal extremity of these enzymes, this region also contains a glutamic acid residue.