Glycoside hydrolases are organized into glycoside hydrolase families (GHFs) and within this larger group, the ␤-galactosidases are members of four families: 1, 2, 35, and 42. Most genes encoding GHF 42 enzymes are from prokaryotes unlikely to encounter lactose, suggesting a different substrate for these enzymes. In search of this substrate, we analyzed genes neighboring GHF 42 genes in databases and detected an arrangement implying that these enzymes might hydrolyze oligosaccharides released by GHF 53 enzymes from arabinogalactan type I, a pectic plant polysaccharide. Because Bacillus subtilis has adjacent GHF 42 and GHF 53 genes, we used it to test the hypothesis that a GHF 42 enzyme (LacA) could act on the oligosaccharides released by a GHF 53 enzyme (GalA) from galactan. We cloned these genes, plus a second GHF 42 gene from B. subtilis, yesZ, into Escherichia coli and demonstrated that cells expressing LacA with GalA gained the ability to use galactan as a carbon source. We constructed B. subtilis mutants and showed that the increased ␤-galactosidase activity generated in response to the addition of galactan was eliminated by inactivating lacA or galA but unaffected by the inactivation of yesZ. As further demonstration, we overexpressed the LacA and GalA proteins in E. coli and demonstrated that these enzymes degrade galactan in vitro as assayed by thin-layer chromatography. Our work provides the first in vivo evidence for a function of some GHF 42 ␤-galactosidases. Similar functions for other ␤-galactosidases in both GHFs 2 and 42 are suggested by genomic data.␤-Galactosidases, such as the LacZ ␤-galactosidase of Escherichia coli, are used as molecular biology tools, industrial enzymes, and models for exploring structure-function relationships. However, these uses do not often yield insight into the physiological roles of these enzymes. Although ␤-galactosidases (EC 3.2.1.23) are also found in glycoside hydrolase families (GHFs) 1, 2, and 35 (groups differing in secondary structure and other attributes [20,21]), we were especially curious about the in vivo function of GHF 42 ␤-galactosidases because they occur in organisms isolated from diverse habitats where lactose would not be present, e.g., hot springs (33, 47, 65), hypersaline environments (27, 54), and soil (Bacillus and Streptomyces spp.).Over a dozen GHF 42 enzymes have been characterized, and most are ␤-galactosidases (as shown by maximal activity with the) with less than 10% relative activity on nongalactosidic chromogens (with two exceptions [27,34]). However, experimental data supporting lactose hydrolysis by GHF 42 ␤-galactosidases are absent (34, 44, 48) or weak. Several prokaryotes possessing a GHF 42 gene are unable to grow on lactose as a sole carbon source (1, 9, 27), and at least two GHF 42 ␤-galactosidases do not cleave lactose in vitro (27,64). The determination of growth on lactose can be complicated by the presence of multiple ␤-galactosidases (7,16,25) because not all of the ␤-galactosidases may be participating equally (or at all) as...

We selected for spore-forming psychrophilic bacteria able to use lactose as a carbon source and one isolate, designated Paenibacillus sp. strain C7, that was phylogenetically related to, but distinct from both Paenibacillus macquariensis and Paenibacillus antarcticus. Some Escherichia coli transformants obtained with genomic DNA from this isolate hydrolyzed X-Gal (5-bromo-4-chloro-3-indoyl-␤-D-galactopyranoside) only below 30°C, an indication of cold-active ␤-galactosidase activity. Sequencing of the cloned insert revealed an open reading frame encoding a 756-amino acid protein that, rather than belonging to a family typically known for ␤-galactosidase activity, belonged to glycoside hydrolase family 3, a family of ␤-glucosidases. Because of this unusual placement, the recombinant enzyme (BglY) was purified and characterized. Consistent with its classification, the enzyme had seven times greater activity with the glucoside substrate ONPGlu (o-nitrophenyl-␤-D-glucopyranoside) than with the galactoside substrate ONPGal (o-nitrophenyl-␤-D-galactopyranoside). In addition, the enzyme had, with ONPGlu, a thermal optimum around 30 to 35°C, activity over a broad pH range (5.5 to 10.9), and an especially low K m (<0.003 mM). Further examination of substrate preference showed that the BglY enzyme also hydrolyzed other aryl-␤-glucosides such as helicin, MUG (4-methylumbelliferyl-␤-D-glucopyranoside), esculin, indoxyl-␤-D-glucoside (a natural indigo precursor), and salicin, but had no activity with glucosidic disaccharides or lactose. These characteristics and substrate preferences make the BglY enzyme unique among the family 3 ␤-glucosidases. The hydrolysis of a variety of aryl-␤-glucosides suggests that the enzyme may allow the organism to use these substrates in the environment and that its low K m on indoxyl-␤-D-glucoside may make it useful for producing indigo.Glycoside hydrolases cleave the bond between two carbohydrates or a carbohydrate moiety and another molecule. The classic Enzyme Classification System (EC) groups glycoside hydrolases by substrate specificity; for instance, the ␤-glucosidases collectively have the designation EC 3.2.1.21. The EC system does not use structural or evolutionary information and is thus less useful for categorizing enzymes when the substrate preferences are not known. Therefore, Henrissat (18) initiated a classification system based on amino acid sequences, hydrophobicity plots, and reaction mechanisms. Under this system, ␤-glucosidases are grouped in two glycoside hydrolase families (GHFs), 1 and 3, and the ␤-galactosidases (EC 3.2.1.23) in four families, 1, 2, 35, and 42. Although this system illustrates enzyme relationships, it does not indicate the natural substrate or function of the enzyme. For example, some GHF 1 ␤-glucosidases possess significant ␤-galactosidase activity (5) and the physiological functions of many of these are unknown.The need for biochemical and physiological studies of microbial glycoside hydrolases is highlighted by the prevalence of open reading frames (OR...