Simple enzymatic in situ generation of dihydroxyacetone phosphate and its use in a cascade reaction for the production of carbohydrates: Increased efficiency by phosphate cycling

Article abstract of DOI:10.1021/jo060644a

A new enzymatic method for the generation of dihydroxyacetone phosphate (DHAP) using the acid phosphatase from Shigella flexneri (PhoN-Sf) and the cheap phosphate donor pyrophosphate (PP_i) is described. The utility of this method was demonstrated in an aldolase-catalyzed condensation carried out in one pot in which DHAP was generated and coupled to propionaldehyde to give a yield of 53% of the isolated dephosphorylated end product.

Full text of DOI:10.1021/jo060644a

carbohydrate natural products.7 Several synthetic procedures of
DHAP precursors have been developed based on protected
DHAP dimers,8 DHAP dimethylacetals,9 a cyclic DHAP di-
methyl acetal,10 or benzyl-protected DHAP.11 Many of these
chemical procedures are low yielding and complicated by
multistep purification procedures, unstable intermediates, and/
or expensive toxic reagents. An enzymatic method would
obviate the need for protection and deprotection steps inherent
in chemical phosphorylation. Enzyme-catalyzed reactions are
often compatible with each other, and this makes it, in principle,
possible to combine several enzymes in a one-pot, multistep
reaction sequence. The easiest enzymatic method is the in situ
formation of 2 equiv of DHAP from fructose-1,6-bisphosphate
(FDP) by FDP-aldolase and triose phosphate isomerase,1a,12
but incomplete conversion complicates isolation of products.
Other methods are based on the phosphorylation of dihydroxy-
acetone (DHA) using either glycerol kinase2,13 or recombinant
DHA kinases,14 which need in situ regeneration of ATP.
Oxidation of L-glycerol phosphate by glycerophosphate oxidase
(GPO)15 also results in DHAP. L-Glycerol phosphate can be
generated by phosphorylation of glycerol using glycerol
kinase.13a,16 It is also possible to phosphorylate glycerol by
phytase using cheap PPi as a phosphate donor;17 however,
racemic glycerol phosphate results and D-glycerol phosphate is
not oxidized by GPO.
Simple Enzymatic in situ Generation of
Dihydroxyacetone Phosphate and Its Use in a
Cascade Reaction for the Production of
Carbohydrates: Increased Efficiency by
Phosphate Cycling
Teunie van Herk, Aloysius F. Hartog,
Hans E. Schoemaker, and Ron Wever*
Van’t Hoff Institute for Molecular Sciences, UniVersity of
Amsterdam, Nieuwe Achtergracht 129,
1018 WS Amsterdam, The Netherlands
rweVer@science.uVa.nl
ReceiVed March 24, 2006
A new enzymatic method for the generation of dihydroxy-
acetone phosphate (DHAP) using the acid phosphatase from
Shigella flexneri (PhoN-Sf) and the cheap phosphate donor
pyrophosphate (PPi) is described. The utility of this method
was demonstrated in an aldolase-catalyzed condensation
carried out in one pot in which DHAP was generated and
coupled to propionaldehyde to give a yield of 53% of the
isolated dephosphorylated end product.
Recently, we reported that the acid phosphatase from Shigella
flexneri (PhoN-Sf) is able to phosphorylate a wide variety of
alcoholic compounds regioselectively using cheap PPi as
phosphate donor.18 Here we describe a new and easy method
for the generation of DHAP in which DHA is phosphorylated
by the phosphatase and PPi. The in situ generated DHAP can
be coupled to an aldehyde in an aldolase-catalyzed condensation
using rabbit muscle aldolase (RAMA), resulting in a C-C
As a result of the broad substrate tolerance and predictable
product stereochemistry, dihydroxyacetone phosphate (DHAP)-
dependent aldolases have been frequently utilized in the forma-
tion of carbon-carbon bonds1 of highly functionalized mol-
ecules (13C labeled,2 heterosubstituted,3 deoxy,4 fluoro,5 and high
carbon sugars6) from simple precursors. Additionally, this enzy-
matic approach has been extended to a variety of non-
(7) (a) Matsumoto, K.; Shimagaki, M.; Nakata, T.; Oishi, T. Tetrahedron
Lett. 1993, 34, 4935-4938. (b) Cheˆnevert, R.; Dasser, M. J. Org. Chem.
2000, 65, 4529-4531.
(8) (a) Effenberger, E.; Straub, A. Tetrahedron Lett. 1987, 28, 1641-
1644. (b) Pederson, R. L.; Esker, J.; Wong, C.-H. Tetrahedron 1991, 47,
2643-2648. (c) Jung, S.-H.; Jeong, J.-H.; Miller, P.; Wong, C.-H. J. Org.
Chem. 1994, 59, 7182-7184.
(9) (a) Valentin, M.-L.; Bolte, J. J. Bull. Soc. Chim. Fr. 1995, 132, 1167-
1171. (b) Charmantray, F.; El Blidi, L.; Gefflaut, F.; Hecquet, L.; Bolte, J.;
Lemaire, M. J. Org. Chem. 2004, 69, 9310-9312. (c) Gefflaut, T.; Lemaire,
M.; Valentin, J.-L.; Bolte, J. J. Org. Chem. 1997, 62, 5920-5922.
(10) Ferroni, E. L.; DiTella, V.; Ghanayem, N.; Jeske, R.; Jodlowski,
C.; O’Connell, M.; Styrsky, J.; Svoboda, R.; Venkataraman, A.; Winkler,
B. M. J. Org. Chem. 1999, 64, 4943-4945.
* To whom correspondence should be addressed. Phone: +31 20 5255110.
Fax: +31 20 5255670.
(1) (a) Bednarski, M. D.; Simon, E. S.; Bischofberger, N.; Fessner, W.-
D.; Kim, M. J.; Lees, W.; Saito, T.; Waldmann, H.; Whitesides, G. M. J.
Am. Chem. Soc. 1989, 111, 627-635. (b) Gijsen, H. J. M.; Qiao, L.; Fitz,
W.; Wong, C.-H. Chem. ReV. 1996, 96, 443-473. (c) Machajewski, T. D.;
Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39, 1352-1374. (d) Fessner,
W.-D.; Helaine, V. Curr. Opin. Biotechnol. 2001, 12, 547-586.
(2) Wong, C.-H.; Whitesides, G. M. J. Org. Chem. 1983, 48, 3199-
3205.
(11) Meyer, O.; Rohmer, M.; Grosdemange-Billiard, C. Tetrahedron Lett.
2004, 45, 7921-7923.
(12) Fessner, W.-D.; Walter, C. Angew. Chem., Int. Ed. Engl. 1992, 31,
614-616.
(13) (a) Crans, D. C.; Whitesides, G. M. J. Am. Chem. Soc. 1985, 107,
7019-7027. (b) Simon, E. S.; Grabowski, S.; Whitesides, G. M. J. Am.
Chem. Soc. 1989, 111, 8920-8921.
(3) (a) Von der Osten, C. H.; Sinskey, A. J.; Barbas, C. F.; Pederson, R.
L.; Wang, Y. F.; Wong, C.-H. J. Am. Chem. Soc. 1989, 111, 3924-3927.
(b) Straub, J. A.; Effenberger, F.; Fischer, P. J. Org. Chem. 1990, 55, 3926-
3932.
(14) Sa´nchez-Moreno, I.; Garc´ıa-Garc´ıa, J. F.; Bastida, A.; Garc´ıa-
Junceda, E. Chem. Commun. 2004, 1634-1635.
(15) Fessner, W.-D.; Sinerius, G. Angew. Chem., Int. Ed. Engl. 1994,
33, 209-212.
(4) (a) Wong, C.-H.; Mazenod, F. P.; Whitesides, G. M. J. Org. Chem.
1983, 48, 3493-3497. (b) Liu, K. K.-C.; Kajimoto, T.; Chen, L.; Zhong,
Z.; Ichikawa, Y.; Wong, C.-H. J. Org. Chem. 1991, 56, 6280-6289.
(5) Wong, C.-H. In ACS Symposium Series 456: SelectiVe Fluorination
in Organic and Bioorganic Chemistry; Welch, J. T., Ed.; American Chemical
Society: Washington, DC, 1991; pp 156-162.
(16) Hettwer, J.; Oldenburg, H.; Flaschel, E. J. Mol. Catal. B 2002, 19,
215-222.
(17) Schoevaart, R.; van Rantwijk, F.; Sheldon, R. A. J. Org. Chem.
2000, 65, 6940-6943.
(18) (a) Tanaka, N.; Hasan, Z.; Hartog, A. F.; van Herk, T.; Wever, R.
Org. Biomol. Chem. 2003, 1, 2833-2839. (b) van Herk, T.; Hartog, A. F.;
van der Burg, A. M.; Wever, R. AdV. Synth. Catal. 2005, 347, 1155-1162.
(6) Bednarski, M. D.; Waldmann, H. J.; Whitesides, G. M. Tetrahedron
Lett. 1986, 27, 5807-5810.
10.1021/jo060644a CCC: $33.50 © 2006 American Chemical Society
Published on Web 07/08/2006
6244
J. Org. Chem. 2006, 71, 6244-6247

FIGURE 1. Optimization of DHAP formation. Standard reaction mixture contains 100 mM PPi, 100 mM DHA in 100 mM acetate buffer at 30
°C. The reaction was started by addition of 2 µM (54 µg/mL) PhoN-Sf. DHAP concentration was determined by the coupled assay as described
in the Supporting Information. (A) PhoN-Sf concentration dependency at pH 5.3. (B) pH dependency. (C) DHA concentration dependency at pH
4.5. (D) PPi concentration dependency at pH 4.5.
SCHEME 1. One-Pot Reaction: Formation of Dephosphorylated Aldol Adduct from DHA, PPi, and Propionaldehyde
coupled product with two new stereocenters with high ste-
reospecificity (Scheme 1).
DHAP with initial DHA concentrations ranging from 0.1 to 2.5
M, respectively. From the initial rate of the reaction versus the
DHA concentration (not shown), it was possible to obtain the
kinetic constants. The Km value was 3.6 M, and the kcat was 40
s-1. The fourth optimization factor is the effect of the PPi
concentration (panel D). Increasing the PPi concentration causes
a further increase in the DHAP formation. In a reaction mixture
containing 100 mM DHA, 100 mM PPi is converted to 5 mM
DHAP and a maximal concentration of 19 mM DHAP is
reached at a concentration of 400 mM PPi.
Combining the optimization factors increases the DHAP
concentration considerably. A reaction mixture containing 500
mM DHA and 240 mM PPi at pH 4.5 and 2 µM (54 µg/mL)
PhoN-Sf resulted in 52 mM of DHAP. Addition of a second
portion of 240 mM PPi after 100 min pushed the yield to 104
mM of DHAP (Figure 2).
To demonstrate the usefulness of this DHAP generating
system in the synthesis of carbohydrates and other compounds,
an aldolase (RAMA) was added in the presence of 100 mM
propionaldehyde. The formation of DHAP under optimal
conditions, as determined above, was started by the addition of
PhoN-Sf and followed by 31P NMR (Figure 3). After 30 min,
RAMA was added, resulting in the formation of the phospho-
rylated aldol adduct to a maximal concentration of 33 mM.
Addition of both the phosphatase and aldolase at the same time
minimizes the dephosphorylation of the formed DHAP by
To find the optimal conditions for DHAP formation, several
parameters were varied as depicted in Figure 1. Time-dependent
optima are observed since, after the consumption of the PPi,
DHAP will be dephosphorylated by the phosphatase. Panel A
shows the effect of the phosphatase concentration. Varying the
PhoN-Sf concentration between 1 and 5 µM (27-135 µg/mL)
yielded maximum DHAP concentrations around 4.3 mM
between 110 and 20 min, respectively. Considering reaction
time, a concentration of 2 µM (54 µg/mL) PhoN-Sf was used
in further studies. The phosphorylating activity of PhoN-Sf is
highly dependent on pH, as is shown in panel B. The highest
DHAP concentration (5.5 mM) was reached at pH 4 after 100
min of incubation. Increasing the pH increased the initial rate
of DHAP formation, but the decomposition of DHAP due to
hydrolysis was also accelerated. At pH 6, only 2 mM of DHAP
was formed. At pH 3.5, the transphosphorylation reaction was
considered too slow to be practically viable, also because this
pH deviates from the aldolase pH optimum. This pH dependency
has also been observed in the phosphorylation of inosine to 5′-
inosine monophosphate by PhoN-Sf.18a Combining yield and
reaction time, pH 4.5 was considered as starting point for further
optimization. As shown in panel C, the DHAP yield increases
with higher initial DHA concentrations. In a reaction mixture
containing 100 mM PPi, the yield increases from 5 to 67 mM
J. Org. Chem, Vol. 71, No. 16, 2006 6245

in the aldol coupling. Scheme 2 illustrates this enzymatic phos-
phoryl transfer. During phosphorylation and formation of the
aldol adduct, several phosphate transfer reactions are thought
to occur at the same time.18a (1) The phosphatase reacts with
PPi to produce a binary PPi-enzyme complex, which (2) disso-
ciates to yield an activated phosphorylated enzyme intermediate
(E1‚P). (3) Reaction of this complex with water would result in
hydrolysis of the complex, but (4) the E1‚P intermediate may
also transfer the phosphate group to an acceptor (in this case
DHA), which (5) dissociates to form E1 and DHAP. In the next
step, the aldol condensation occurs, and DHAP is coupled to
an aldehyde by RAMA to form the phosphorylated aldol adduct
Z-P. As the reaction proceeds in time, PPi becomes exhausted
and dephosphorylation of Z-P occurs via (re)formation of the
phosphorylated enzyme intermediate which reacts with water
(hydrolysis). Our experiments show that the phosphate group
from the aldol adduct (Z-P) hydrolysis is used to phosphorylate
another DHA molecule when present in excess to generate more
DHAP, which is used in aldol coupling. As described above,
hydrolysis proceeds via the E1‚P intermediate, which is the
central intermediate used in the phosphorylation. In other words,
the phosphate group is transferred more than one time, and
higher yields can be obtained.
FIGURE 2. Combination of the optimization factors. Reaction
mixtures contain 0.5 M DHA (closed symbols) or 1 M DHA (open
symbols) and varying concentrations of PPi in 100 mM acetate buffer
(pH 4.5) at 30 °C. The reaction was started by addition of 2 µM (54
µg/mL) PhoN-Sf. The second PPi portion (240 mM) was added after
100 min (0.5 M DHA, 240 mM PPi).
A 10 mL 100 mM propionaldehyde reaction was carried out.
Three portions of 240 mM PPi were sufficient to convert >95%
of the aldehyde to aldol adduct. Dephosphorylation was finished
overnight. The product was isolated in 53% yield and was
identical to 5,6-dideoxy-D-threo-2-hexulose.
When we compare this new one-pot method of in situ genera-
tion of DHAP in combination with an aldolase-coupling reaction
with the method using phytase to phosphorylate glycerol and
subsequent oxidation to DHAP by GPO,17 a number of advant-
ages can be seen. (1) Only two enzymes are necessary compared
to four used in the phytase method. (2) In this new method, pH
switching is not necessary. The phytase phosphorylation is
carried out at pH 4.0, whereas the oxidation has to be performed
at pH 7.5. Further, the dephosphorylation of the product needs
a second pH switch back to pH 4.0. (3) No dilution of the
reaction mixture to lower the glycerol concentration to obtain
sufficient GPO activity is necessary in our method. (4) The
glycerol phosphate formed is racemic in the phytase method,
and since D-glycerol phosphate is not converted by GPO, only
50% of the phosphorylated product is used. (5) In our method,
increased yields can be obtained due to cycling of the E-P
intermediate when sufficient aldehyde and DHA are present.
FIGURE 3. 31P NMR analysis of a one-pot phosphorylation, aldol
coupling, and dephosphorylation reaction. Reaction mixture contains
240 mM PPi, 500 mM DHA, and 100 mM propionaldehyde in 100
mM acetate buffer (pH 4.5). The reaction was started by the addition
of 2 µM (54 µg/mL) PhoN-Sf. After 30 min, 10 units of RAMA were
added.
converting it directly into the aldol adduct and results in an
increased yield. This can be explained by comparison of the
Km values for DHAP of both enzymes. RAMA has a Km value
of 49 µM,19 whereas PhoN-Sf has a Km value of 0.5 mM (data
not shown), demonstrating a higher affinity of the aldolase for
DHAP compared to the phosphatase.
To further optimize the reaction, an experiment was carried
out in which RAMA was allowed to react with propionaldehyde
and DHAP (commercially obtained). After reaction in which
>95% of the phosphate from DHAP was transferred to the aldol
adduct, the reaction mixture was split into three parts for further
incubations. To the three incubations was added phosphatase,
to incubations two and three was added additional aldehyde,
and to incubation three was added DHA. The concentrations
of the dephosphorylated aldol adduct were measured in all three
incubations. To our surprise, the concentration of aldol adduct
in the third incubation was significantly higher (2.3 times) com-
pared to incubations one and two. Similar results were also
obtained using butyraldehyde (1.6 times). This shows that the
phosphate group from the aldol adduct can be transferred to
DHA, which becomes DHAP and is subsequently used again
We conclude that this cascade reaction provides an attractive
enzymatic procedure in which DHAP is generated in situ from
DHA and cheap PPi in an easy way and which can be coupled
to aldol reactions with DHAP-dependent aldolases. Advantage
is taken of the two-way action of the phosphatase. First it
catalyzes the simple phosphorylation of DHA, avoiding the
problems with chemical phosphorylation, and second it dephos-
phorylates the aldol adduct, avoiding nonenzymatic dephos-
phorylation which, in general, may cause decomposition of the
products. Although we used only propionaldehyde and butyral-
dehyde as substrates for RAMA, the enzyme accepts many other
natural and unnatural aldehydes.1a When other aldolases are
used, such as L-fuculose 1-phosphate aldolase and L-rhamnulose
1-phosphate aldolase, other stereoisomers can be generated. Thus
our new method may lead to the synthesis of a wide array of
compounds that up to now could only be synthesized via a
number of steps. A drawback to our method may be the large
(19) Callens, M.; Kuntz, D. A.; Opperdoes, F. R. Mol. Biochem.
Parasitol. 1991, 47, 1-10.
6246 J. Org. Chem., Vol. 71, No. 16, 2006

SCHEME 2. Enzymatic Phosphoryl Transfer Reactions in the Formation of Aldol Adductsa
a E1 ) PhoN; E2 ) RAMA/FruA; Pi ) inorganic phosphate; R ) alkyl, aryl, etc.; Z ) aldol adduct.
to 1 mL, and 4 units of RAMA were added. After 5 h, again, 300
amounts of DHA and Pi that are present during purification.
However, according to the Merck Index,20 several more polar
and charged carbohydrates are soluble in methanol allowing
further purification. Currently, we are performing directed
evolution on the phosphatase to increase the affinity for DHA,
so that, in principle, much lower concentrations of DHA and
PPi are sufficient to keep the phosphate transfer cycle going
and purification of the aldol adduct may be easier.
mM of aldehyde was added to portions 2 and 3. Reactions were
allowed to react overnight and were analyzed by HPLC for amounts
of dephosphorylated end product. A RAMA background reaction
(500 mM DHA, 2 times 300 mM aldehyde, and 8 units of RAMA)
accounted for only 3% of the total conversion.
Isolation and Characterization of Aldol Product 5,6-Dideoxy-
D-threo-2-hexulose. To a 10 mL solution containing 500 mM DHA,
240 mM PPi, 100 mM propionaldehyde, and 36 units of RAMA in
100 mM acetate buffer (pH 5.0) at 30 °C was added 2 µM (54
µg/mL) PhoN-Sf. After 2 and 4 h, additional PPi was supplemented
(0.53 g, 240 mM), and the pH was adjusted to 5.0 with 5 M NaOH.
Dephosphorylation to the end product was complete after 24 h.
The solution was freeze-dried, and the residue was extracted with
methanol (overnight). Remaining phosphate salts were filtered off,
and the filtrate was concentrated in vacuo. Purification was
performed by silica gel column chromatography (ethyl acetate:
methanol ) 19:1). Product was obtained as a slightly yellow oil
(79 mg; 53% based on propionaldehyde, which was shown to be
>95% pure by HPLC). The isolated product was characterized1a
by 1H and 13C NMR. Mass spectrometry was carried out on a JEOL
JMS-SX/SX 102 A Tandem Mass Spectrometer using Fast Atom
Bombardment (FAB+), (m/z): [M + Na]+ calcd for C6H12O4Na,
171.146; found, 171.05.
Experimental Section
31P NMR. Phosphorylated products, PPi, and free phosphate (Pi)
were quantified by phosphor nuclear magnetic resonance at 202
MHz using a 10 mm 15N-31P probe. Chemical shifts (δ) are
expressed in parts per million relative to 85% phosphoric acid. At
zero time, a spectrum was taken from the reaction mixture
containing 240 mM PPi, 500 mM DHA, 100 mM propionaldehyde,
and 100 mM sodium acetate buffer (pH 4.5) in a 10 mm NMR
tube at 30 °C in a total volume of 2.75 mL. The reaction was
initiated by addition of 2 µM (54 µg/mL) PhoN-Sf. After 30 min,
10 units of RAMA were added. Concentrations of product and
reactants were determined using 50 mM dimethylmethylphospho-
nate (DMMP) in deuterated water as an external standard, which
was coaxially inserted in the NMR tube.
Phosphate Recycling Experiment. Sixteen units of RAMA were
added to a 2 mL solution containing 300 mM DHAP and 300 mM
aldehyde (either propion- or butyraldehyde) in 200 mM acetate
buffer (pH 6) at 30 °C. After 5 h, another portion of 300 mM
aldehyde was allowed to react overnight. Over 95% of DHAP was
converted to the phosphorylated aldol adduct at this time. The
mixture was split in three parts of 0.5 mL. The first portion was
completed with 2 µM (54 µg/mL) of PhoN-Sf. To the second
portion were added 2 µM of PhoN-Sf and 300 mM of aldehyde.
The third reaction was completed with 2 µM of PhoN-Sf, 300
mM of aldehyde, and 500 mM of DHA. The volumes were adjusted
Acknowledgment. We gratefully acknowledge Prof. A. J.
Lange, University of Minnesota, for providing the plasmid
pET3a PhoN-Sf. The assistance with NMR experiments from
Lidy van der Burg is appreciated. This work was supported by
the Dutch National Research School Combination (NRSC-
Catalysis).
Supporting Information Available: General experimental
1
details, 31P and H NMR spectral data, and HPLC traces. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(20) The Merck Index, 12th ed.; Budavari, S., Ed.; Merck Publishing
Group: Rahway, NJ, 1996; pp 757-758.
JO060644A
J. Org. Chem, Vol. 71, No. 16, 2006 6247