Full Papers
the complete elimination of H O avoids enzyme inactivation
2
2
Table 2. Efficiency of the different immobilized biocatalysts to eliminate
the H O concomitantly formed from the enzymatic reduction of NAD .
2 2
+
by chemical oxidation and, importantly, avoids the unspecific
oxidation of DHA that reduces the product yield of the bio-
Biocatalyst
[H
2
O
2
] in bulk
Efficiency of H
[%]
2
O
2
elimination
[24]
transformation dramatically.
The action of both NOX and
[
mm]
CAT recycles and eliminates the redox cofactor and the perox-
ide in situ, respectively, within the same porous microenviron-
ment, which increases both the kinetics and yield of the DHA
biosynthesis.
Ag-AG-NOX
140
19
2
–
86
99
[
a]
Biocat 3-3
Biocat 3-1
[
b]
[
a] Biocat 3-3: NOX and CAT are immobilized separately on two different
carriers. [b] Biocat 3-1: NOX and CAT are co-immobilized on the same car-
rier. [c] If we consider 100% of H is the concentration of H pro-
This new heterogeneous multienzyme system resulted in
a four times higher glycerol conversion to DHA than the multi-
enzyme system formed by GlyDH from Cellulomonas sp co-im-
2
O
2
2 2
O
duced and detected in the enzymatic reaction performed by NOX immo-
bilized on Ag-AG and further crosslinking with dexCHO. See Experimental
Section.
[23]
mobilized with Xylitol reductase on silica nanoparticles. Fur-
thermore, this trienzyme immobilized biocatalyst reaches even
higher DHA yields than those found for other 2-hydroxyke-
tones, such as 4-hydroxy-2-butanone and 2-hydroxycyclohexa-
[25]
cause some of the H O2 was decomposed spontaneously.
noene. Therefore, the rational integration of protein engi-
neering and immobilization techniques has resulted in immo-
bilized multienzyme systems with better properties to yield
DHA by the selective oxidation of glycerol.
2
Therefore, accumulation values might be underestimated be-
cause of the indirect method we used to titer H O (see Experi-
2
2
mental Section). However, NOX and CAT immobilized separate-
ly on two different carriers oxidized NADH quantitatively to ac-
cumulate 19 mm H O ; 7.3% of its theoretical yield. This means
2
2
Conclusions
that CAT partially eliminated the H O . Nevertheless, we only
2
2
detected 2 mm of H O in the quantitative NADH oxidation cat-
2
2
A supported cell-free platform based on a trienzyme system,
glycerol dehydrogenase (GlyDH), NADH oxidase, and the cata-
lase from bovine liver, was designed and tested for the selec-
tive oxidation of glycerol to 1,3-dihydroxyacetone with both
in situ redox-cofactor recycling and H O elimination. Firstly, we
alyzed by NOX and CAT co-immobilized on the same carrier.
Such a residual H O concentration is 0–1% of the theoretical
2
2
yield. Hence, the colocalization of NOX and CAT accumulate
negligible H O in the bulk solution as the levels of H O de-
2
2
2
2
2
2
tected are quite close to the detection limit of the colorimetric
assay. These results indicate that immobilized CAT is able to
eliminate H O in situ and it does this more efficiently if it is
engineered GlyDH to minimize its product inhibition that led
to poor yields. The product inhibition of this enzyme was im-
proved by at least five times by combining protein engineer-
ing and immobilization techniques. The engineered GlyDH was
co-immobilized with the other two enzymes, which confined
them into the same micrometric and porous environment. As
result of such confinement, the heterogeneous biocatalyst pro-
duced up to 9.5 mm 1,3-dihydroacetone, which is 18- and 6-
fold higher than glycerol dehydrogenase itself and the soluble
multienzyme system, respectively. Moreover the co-immobi-
lized multienzyme system presented a 4.5 times higher pro-
2
2
co-immobilized with NOX on the same carrier porous surface.
The success of this co-immobilized and co-crosslinked engi-
neered biocatalyst is driven by the synergy between the im-
mobilization and post-immobilization chemistry that reduces
the product inhibition of GlyDH and co-immobilization of the
three enzymes that enable their spatial colocalization within
the porous carriers. Such a spatial localization of the immobi-
lized multienzyme system improves both the yield and kinetics
of the biotransformation. This improvement is mainly because
+
ductivity and eliminated the H O formed during the NAD re-
2
2
+
both NAD recycling and H O2 elimination are much more
2
cycling quantitatively. Therefore, this work demonstrates, once
more, that the interdisciplinary engineering of biocatalysts re-
sults in integral solutions to overcome process bottlenecks.
This biocatalyst opens new opportunities for process engineer-
ing to scale up the process under the best reaction conditions.
rapid and efficient in the porous microenvironment. The better
performance of these two processes causes a higher effective
+
concentration of NAD available for the glycerol oxidation and
an undetectable H O2 concentration accumulated inside the
2
+
pores. As expected, the higher effective NAD concentration
improved the total turnover number (TTN) of the system up to
2
0. The TTN values were not as high as those reported for Experimental Section
other multienzyme cascades that involve orthogonal redox-co-
Chemicals
[
36]
factor recycling. The TTN numbers are low probably because
in this multienzyme cascade the limiting step is the product in-
+
NAD and NADH were purchased from GERBU Biotechnik GmbH
+
(Wieblingen, Germany). Glycidol, glycerol, triethylamine (TEA), poly-
ethylene glycol (PEG), sodium borohydride, sodium periodate, per-
oxidase from horseradish (HRP), and SIGMAFAST DAB (3,3’-diamino-
benzidine tetrahydrochloride) with Metal Enhancer Tablet Sets
hibition of GlyDH rather than the NAD recycling. The ob-
served TTN values for the co-immobilized system are compara-
ble to those reported for DHA production catalyzed by GlyDH
from Cellulomonas sp and Xylose reductase from Pichia stipitis
(DAB) were supplied by Sigma–Aldrich (St. Louis, IL). DHA was sup-
[
23]
immobilized on silica nanoparticles and GlyDH from Citro-
bacter braakii and NOX from Thermus thermophius co-immobi-
plied by Acros Organics (Geel, Belgium). Crosslinked agarose beads
(4%) were from Agarose Beads Technology (Madrid, Spain). The
Coomassie (Bradford) protein assay kit was purchased from Pierce
[
10]
lized on agarose beads under similar conditions. Moreover,
&
ChemCatChem 0000, 00, 0 – 0
6
ꢂ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!