T. Pons et al., Prediction of a common beta-propeller catalytic domain for fructosyltransferases of different origin and substrate specificity, PROTEIN SCI, 9(11), 2000, pp. 2285-2291
The three-dimensional (3D) structure of fructan biosynthetic enzymes is sti
ll unknown. Here, we have explored folding similarities between reported mi
crobial and plant enzymes that catalyze transfructosylation reactions. A se
quence-structure compatibility search using TOPITS, SDP, 3D-PSSM, and SAM-T
98 programs identified a beta -propeller fold with scores above the confide
nce threshold that indicate a structurally conserved catalytic domain in fr
uctosyltransferases (FTFs) of diverse origin and substrate specificity. The
predicted fold appeared related to that of neuraminidase and sialidase, of
glycoside hydrolase families 33 and 34, respectively. The most reliable st
ructural model was obtained using the crystal structure of neuraminidase (P
rotein Data Bank file: 5nn9) as template, and it is consistent with the loc
ation of previously identified functional residues of bacterial levansucras
es (Batista et al., 1999; Song & Jacques, 1999). The sequence-sequence anal
ysis presented here reinforces the recent inclusion of fungal and plant FTF
s into glycoside hydrolase family 32, and suggests a modified sequence patt
ern {H-x(2)-[PTV]-x(4)-[LIVMA]-[NSCAYG]-[DE]-P-[NDSC]-[GA]} for this family
.