acid chain binds to the hydrophobic crevice and is, subsequently, cleaved off; and in
sn-3 orientation, where thesn-3 fatty acid chain binds to the hydrophobic crevice.
The hydrophobic dent is assumed to bind thesn-2 substituent.
In order to evaluate the influence of flexibility, size and polarity of thesn-2 sub-
stituent on stereoselectivity, triacylglycerol was compared to the more flexible ether
derivative and to the more bulky amide and phenyl derivatives (Figure 4). The sug-
gested model explained why ROL hydrolyzes substrates with an ester or ether group
atsn-2 preferably insn-1 position, while the amide and phenyl substrates were hy-
drolyzed insn-3. It also explained why RML, although it is homologous to ROL,
hydrolyzes all substrates insn-1.
The catalytic H257 and L258 in ROL and RML form a binding pocket in the
hydrophobic dent. All substrates bind with thesn-2 substituent close to C 2 of gly-
cerol between the side chains of these two residues, theHis gap.All substrates inter-
acted sterically with the side chain of L258 insn-1 orientation, but less insn-3or-
ientation. It was concluded that this steric interaction governs stereoselectivity: for
rigidand bulky substrates, thesn-1 orientation is less favorable than forflexible
substrates. Insn-3 orientation, flexible and rigid substrates adjusted better to the
His gap, by a conformational change of the glycerol backbone. This adaptation
can be directly evaluated by measuring torsion ØO3–C3around the bond between
the hydrolyzed ester oxygen and glycerol C 3 of the substrate insn-3 orientation.
It was shown that this torsion angle discriminates between rigid and flexible trira-
dylglycerols: it was 160 8 to 170 8 for flexible substrates, but about 120 8 for rigid
substrates (Table 2). For bothMucoraleslipases, the value of this torsion angle cor-
related well with experimentally determined stereoselectivity (Figure 5).
5.4 Modeling and engineering of stereoselectivity 93
Figure 5. Correlation of experimental E values and torsion angle ØO3–C3;forsn-1 preference, data are
directly taken from Table 2; forsn-3 preference, reciprocal values of E values are displayed; horizontal
and vertical bars divided the data points in two regions: E1(sn-1 preference)/(ØO3-C3< 1508 ) and E< 1
(sn-3 preference)/(ØO3-C3< 1508 ).