3.2 Constructing the NADPH cycle with the modification ofPdPDH
There is no NADP+-dependent galactitol dehydrogenase
in the ENZYME database, and it is difficult to mine the
NADP+-dependent galactitol dehydrogenase in nature.
Therefore, we must develop a rational design for
NADP+-dependent galactitol dehydrogenase. We minedPd PDH and Rl GDH from the enzyme bank, and their
dehydrogenase activities were compared (Figure S2A). The activity ofPd PDH with respect to galactitol dehydrogenation was 17.29 times
greater than that of Rl GDH (25.42 and 1.47 U
mg−1, respectively) (Figure S2B). Therefore,Pd PDH was selected as the template for the rational design.
We docked Pd PDH (PDBID: 7E6O) with the cofactor
NAD+ (PubChem CID: 5892) using the Schrodinger
software, and the docking results are shown using LigPlot software
(Figure S3A). We selected 22 amino acids within 5 Å of
NAD+ that interact with the cofactor based on hydrogen
bond or hydrophobic interaction. The Pd PDH-NAD+combination model was predicted using the CSR-SALAD website (Table S3
and Table S4). According to the predicted results, we designed 100
mutants of Pd PDH (Table S5) based on 13 amino acid residues
(Figure S3B). Subsequently, a point mutation of Pd PDH was
simulated using Schrodinger software, and the mutants were docked with
the cofactor NADP+. Then, the docking binding energies
of the mutants were counted (Table S5). A total of 14 mutants were
selected according to their lower binding energies, indicating that they
had a higher affinity for NADP+ than the other
mutants. Furthermore, their specific activities for
NAD+ and NADP+ were measured. As
shown in Table 1, nine mutants
completely reversed the cofactor preference
and exhibited improved specific
activity toward NADP+. Six mutants had
NADP+ activities of more than 15 U
mg−1; particularly, the NADP+activity of the D36A/I37R mutant reached 19.4 U mg−1.
However, the triple mutants
(A14T/D36A/I37R and A14S/D36A/I37R) did not
improve their activities further
than those of the double mutants. The single mutants (D36A, D36G, and
I37R) exhibited no activity against NADP+.
The screening accuracy was
increased sevenfold the computer-aided screening (9/14) compared with
the original screening (9/100).
The kinetic parameters of Pd PDH mutants and the wild-type are
shown in Table 2. TheKm value of NADP+ with the
D36A/I37R mutant was 0.942 mM and was lower than that of the wild-type
enzyme (infinite), proving that there was a substantial increase in the
affinity of the D36A/I37R mutant for NADP+. Moreover,
the specificity constant
(kcat/Km ) of the D36A/I37R mutant
toward NADP+ reached 0.322
mM−1s−1, whereas that of the
wild-type was 0 mM−1s−1. Conversely,
the specificity constant
(kcat/Km ) of the wild-type toward
NAD+ was 0.4872
mM−1s−1, whereas that of the
D36A/I37R mutant was 0 mM−1s−1.
Furthermore, we identified the binding free energy of each cofactor with
the wild-type and its mutant D36A/I37R (Table S6).
The binding free energy of the
D36A/I37R mutant toward NADP+ decreased by 54% (from
−8.011 to −12.335 kcal mol−1), whereas that toward
NAD+ increased by 46% (from −13.614 to −7.325 kcal
mol−1). These results suggested that the binding
affinity of the D36A/I37R mutant toward NADP+ was
higher than that toward NAD+. Therefore, we obtained
an NADP+-dependentPd PDHD36A/I37R for performing the NADPH cycle
during tagatose production.