Featured Articles

Glucosidases: Defining domain function

Image courtesy of Takane Katayama

Carbohydrate-processing enzymes such as β-glucosidases display a wide range of substrate specificities to control their interactions with and—in the case of glucosidases—hydrolyze their sugar targets. The PA14 domain has been identified in a large number of glucosidases, and has a putative role as a carbohydrate binding domain, but there is no direct evidence for this function and thus its presence remains a mystery. Now Takane Katayama and colleagues, reporting in Biochemical Journal, provide structural and biochemical evidence for its function in KmBglI, a β-glucosidase from Kluyveromyces marxianus.

The PA14 domain was first identified in the protective antigen of anthrax toxin, and has been identified in a variety of proteins including proteases, adhesins, and signaling molecules. Of the 820 known occurrences, 230 are found in glycoside hydrolases, suggesting a particularly important role within this class of enzymes.

KmBglI is a member of the glycoside hydrolase family 3 (GH3), of which only a few enzymes have been structurally characterized. Of these, ExoI contains an N-terminal barrel domain and a C-terminal sandwich domain, whereas Bgl3B contains an extension in its C-terminus, creating a third domain.

Katayama and colleagues began their characterization of KmBglI by assaying several potential substrates as paranitrophenol-labeled monosaccharides, and discovered that glucose is preferred over xylose, with little or no activity against six other sugars. Extension to larger substrates suggested that laminaribiose is a preferred disaccharide, with 180-fold lower activity against the corresponding trisaccharide, laminaritriose.

To understand the basis of this specificity, the authors solved the X-ray crystal structure of KmBglI in the presence and absence of glucose. KmBglI is similar in structure to ExoI and Bgl3B, with a two helix deletion in the N-terminal domain that mirrors the structure of Bgl3B. However, KmBglI diverges from these proteins by the insertion of the PA14 domain into the C-terminal domain. The loops of the PA14 domain create a lid over the enzyme active site. Biochemical analysis of several PA14 domain mutants confirmed that these loops limit the size of the +1 subsite, thus making the enzyme specific for mono- and disaccharides.

The variable nature of the loop sequences in putative PA14 domains indicates that, whereas the domain is likely to play a general role in binding carbohydrates, the specific effect on each enzyme will vary. Thus further work is needed to understand the biochemical roles of other PA14 domains. The study by Katayama and colleagues also indicates that the PA14 domain might be more ubiquitous than previously thought. In particular, the KmBglI PA14 domain maps closely to the original domain found in anthrax protective antigen, but the identity of these sequences is quite low (<25%). The structure also maps closely with the N-terminal domains of several GH2 family members which are not annotated as PA14 domains. Additional work to establish structural motifs will improve our understanding of this domain throughout biology.

Author: Catherine Goodman