Applications of Aldolases in Organic Synthesis - Schoevaart Thesis 2000

Applications of aldolases
in organic synthesis
Rob Schoevaart
Front cover:
Sun set in the Mediterranean sea
Er is niets in het bewustzijn dat niet door zijn eigen zelf is.
Phenomenologie des Geistes, Hegel

Applications of aldolases
in organic synthesis
Toepassingen van aldolasen in organische synthese
Proefschrift
ter verkrijging van de graad van doctor
aan de Technische Unversiteit Delft
op gezag van de Rector Magnificus, Prof. ir. K.F. Wakker
ingevolge het besluit van het College voor Promoties
in het openbaar te verdedigen op
dinsdag 23 mei 2000 om 13:30
door
Willem Robert Klaas Schoevaart
geboren op 28 januari 1969 te Coevorden

Dit proefschrift is goedgekeurd door de promotor: Prof. dr. R.A. Sheldon.
Samenstelling promotiecommissie:
Rector Magnificus Technische Universiteit Delft
Prof. dr. A. Bruggink DSM Geleen / KU Nijmegen
Prof.dr.ir. J.J. Heijnen Technische Universiteit Delft
Prof. dr. A.P.G. Kieboom Universiteit Leiden / DSM-Gist
Prof. dr. B. Zwanenburg KU Nijmegen
Dr. W. Fitz ICI
Dr.ir. F. van Rantwijk Technische Universiteit Delft
Dr. ir. F. van Rantwijk heeft als begeleider in belangrijke mate aan de
totstandkoming van dit proefschrift bijgedragen.
ISBN 90-9013780-7
Copyright© 2000 W.R.K. Schoevaart
Printed by Print Partners Ipskamp
This work was carried out as part of the Innovation Oriented research Programme on
Catalysis (IOP Katalyse, no. IKA94010) sponsored by the Netherlands Ministery of
Economic Affairs.
Contents
1 ............................................................................................................. 1
General introduction ......................................................................... 1
IOP Catalysis.......................................................................................................2
Enzymatic aldol reactions....................................................................................3
The aldol reaction................................................................................................4
Aldol reactions with aldolases .............................................................................5
Dihydroxyacetone phosphate dependent aldolases ............................................8
Pyruvate- and phosphoenolpyruvate-dependent aldolases...............................10
Glycine-dependent aldolases ............................................................................11
2-Deoxyribose-5-phosphate aldolase................................................................12
Transketolase....................................................................................................13
The objectives and justification of the thesis .....................................................15
General ..........................................................................................................15
Specific ..........................................................................................................16
References ........................................................................................................16
2 ........................................................................................................... 21
Aldol reactions with DHAP dependent aldolases ......................... 21
Abstract .............................................................................................................21
Introduction........................................................................................................22
Acceptor substrate specificity of FruA, FcuA and RhuA ....................................23
Properties of FruA from different sources..........................................................24
FruA stability under reaction conditions. ........................................................27
A 5 mmol scale reaction.................................................................................27
Conclusion.........................................................................................................28
Experimental .....................................................................................................28
General ..........................................................................................................28
Aldol reaction .................................................................................................28
FruA activity assay.........................................................................................29
DHAP-assay ..................................................................................................29
Enzyme recovery ...........................................................................................30
Synthesis of 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose ............................31
References ........................................................................................................31
3 ........................................................................................................... 35
The stereochemistry of DHAP aldolases....................................... 35
Abstract .............................................................................................................35
Determination of stereoselectivity with chiral GC ..............................................36
Determination of stereoselectivity with an enzymatic assay ..............................39
Comparison of methods.................................................................................41
Stereospecificity of FruA and RhuA ...............................................................42
Experimental .....................................................................................................45
General ..........................................................................................................45
Acetylation of GC-samples ............................................................................46
Preparation of product samples .....................................................................46
Aldol reaction .................................................................................................46
Retro aldol reaction assay .............................................................................47
References ........................................................................................................47
4 ........................................................................................................... 51
In situ generation of DHAP: cascade catalysis ............................. 51
Abstract .............................................................................................................51
Synthesis of dihydroxyacetone phosphate ........................................................52
Integral in situ reaction ......................................................................................55
Sequential one-pot reaction...............................................................................60
Figure 6. ............................................................................................................60
Synthesis of L-glycerol-3-phosphate..............................................................61
Effects of glycerol concentration ....................................................................62
Selectivity.......................................................................................................63
Synthesis of DHAP from L-glycerol-3-phosphate...........................................64
Aldol reaction and dephosphorylation ............................................................64
Chemical phosphorylation of glycerol ............................................................65
Conclusion.........................................................................................................66
Experimental .....................................................................................................66
General ..........................................................................................................66
Assay for DHAP and L-glycerol-3-phosphate ................................................67
DHAP synthesis from glycerol .......................................................................67
Aldol reactions ...............................................................................................68
Chemical phosphorylation of glycerol ............................................................69
References ........................................................................................................69
5 ........................................................................................................... 73
DHAP analogues ............................................................................. 73
Abstract .............................................................................................................73
Hydroxyacetone thiosulfate as DHAP analogue................................................74
Dihydroxyacetone arsenate as DHAP analogue ...............................................76
Results and discussion......................................................................................78
Optimal conditions .........................................................................................79
Use of cosolvents ..........................................................................................80
Acceptor specificity with DHAAs ....................................................................83
Effect of arsenate on the stereoselectivity .....................................................85
Conclusion.........................................................................................................86
Experimental .....................................................................................................86
General ..........................................................................................................86
Preparation of hydroxyacetone thiosulfate (HATS)........................................87
Preparation of 2,3-dihydroxypropane thiosulfate ...........................................88
Assay of aldehydes with alcohol dehydrogenase ..........................................88
Assay of DHA ................................................................................................89
Aldol reaction with dihydroxyacetone arsenate..............................................89
Stereoisomer assay .......................................................................................90
Preparation of 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose .........................91
References ........................................................................................................91
6 ........................................................................................................... 95
Transketolase versus...................................................................... 95
fructose-1,6-bisphosphate aldolase .............................................. 95
Abstract .............................................................................................................95
Introduction........................................................................................................96
α-Hydroxyaldehydes .........................................................................................98
Cross-linked enzyme crystals..........................................................................100
Conclusions.....................................................................................................102
Experimental ...................................................................................................103
General ........................................................................................................103
Transketolase catalysed reactions...............................................................103
FruA catalysed reactions .............................................................................103
Hydroxypyruvate assay................................................................................104
Synthesis of D,L-lactaldehyde .....................................................................104
Synthesis of α-hydroxybutanal.....................................................................105
Synthesis of α-hydroxypentanal...................................................................106
Synthesis of 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose ..........................106
Crystallization and cross-linking..................................................................106
References ......................................................................................................107
7 ......................................................................................................... 111
2-Deoxyribose-5-phosphate aldolase (DERA)............................. 111
Abstract ...........................................................................................................111
Introduction......................................................................................................112
Donor substrates .............................................................................................113
Acceptor substrates.........................................................................................113
Sequential aldol additions ...............................................................................114
Results and discussion....................................................................................116
Stability in cosolvents...................................................................................117
Activity in cosolvents....................................................................................119
Conclusion.......................................................................................................121
Experimental ...................................................................................................121
General ........................................................................................................121
Preparation of DERA ...................................................................................121
Activity assay ...............................................................................................122
Stability test .................................................................................................122
References ......................................................................................................122
Summary........................................................................................ 124
Samenvatting................................................................................. 127
Dankwoord..................................................................................... 130
Publications................................................................................... 133
Curriculum vitae ............................................................................ 134
Index............................................................................................... 135
aan mijn ouders

1
General introduction
The presence of multiple functional groups of similar reactivity in molecules poses
problems for classical organic synthetic methods. Synthesizing and modifying
carbohydrates for example, is accomplished only by applying complex synthetic
schemes with many protection and deprotection steps to ensure regio- and
stereoselectivity. A large proportion of the reactions in the armamentarium of organic
chemists is incompatible with aqueous systems and some modifications are
necessary simply to achieve solubility in non-aqueous solvents. The application of
enzymes1 as catalysts in organic synthesis has succeeded in solving synthetic
problems and developments in enzymology have effectively refuted the perception
that most enzymes are fully dependent on natural substrates, costly, unstable and
require expensive cofactors.
1
Chapter 1
IOP Catalysis
Research supported by the Ministry of Economic Affairs, with the ultimate goal of
creating new economic activity (i.e. to create jobs and prosperity), set itself the
demanding objective of providing an essential creative step on the road from
fundamental science to novel, applicable technology. This implied research based on
a vision of its potential applications in an industrial context2.
Catalytic conversions are involved in the manufacture of more than 80% of the
total volume of chemicals. Hence a solid knowledge base on catalysis is of strategic
interest to the Dutch chemical industry. Catalysis offers the opportunity to steer
chemical conversions into a desired direction. This has resulted in the central
chemical theme of a research programme known as the Innovation Oriented
Research Programme (IOP) Catalysis: precision in chemical conversion. This
precision is required both to save energy and feedstocks and to avoid the formation
of undesired by-products and waste. Catalysis is applied on a large scale in
petroleum refining and in the manufacture of bulk chemicals. This is far less the case
in the manufacture of fine chemicals where classical multi-step chemical conversions
play a much larger role. These classical procedures often involve lower selectivities,
the use of undesirable, toxic, or corrosive reagents, the formation of by-products and
large amounts of waste.
Biocatalysis is of particular interest for the manufacture of the fine chemicals,
such as pharmaceuticals and products of the flavour and fragrance industry. In these
branches specific optically active isomers are often required; biocatalyst have shown
promise in such enantioselective synthesis since they compete with chemical
catalysis on e.g. introduction of chirality, selective modification or resolution of
2
racemates. Beyond that, enzymatic conversion of natural compounds can even yield
“natural” products. For example, “natural” benzaldehyde is distilled from plants, but
retro aldol cleavage of abundant cinnamaldehyde is a “grey–zone chemistry” route to
so-called natural benzaldehyde3. Inexpensive, readily available and renewable
natural precursors, such as fatty or amino acids, can be converted to more highly
valued flavours.
Enzymatic aldol reactions
Catalytic asymmetric aldol reactions remain one of the most interesting and
challenging subjects in synthetic organic chemistry and, without a doubt, in
biocatalysis. This attractiveness and ambition originates from the importance of C-C
bond formation. When comparing simple base catalysed aldol reactions with enzyme
catalysed ones a characteristic of the enzymatic reaction is particularly outstanding:
enantioselectivity. Whereas the simple base catalysed reaction is only
enantioselective with (stoichiometric amounts of) chiral starting materials, application
of enzymes can lead to enantiomerically pure products from completely non-chiral
starting materials. For non-enzymatic reactions regioselectivity is more controlable,
but the use of aldehydes or ketones with free hydroxyl or highly activated groups
might be problematic, making protection and deprotection steps necessary.
Enzymes are usually completely regioselective, reducing protective group chemistry
to a minimum, making biocatalysis an important instrument in C-C bond forming
reactions.
Aldolases have good activity towards hydroxylated compounds which is not
3
Chapter 1
suprising considering their role in the carbohydrate metabolism. This property is
narrowly associated with their common medium: water. Aldol reactions catalysed by
enzymes can consequently be conducted under mild conditions (water, pH 7) and do
not carry the burden of obligatory use of strong bases and organic solvents, thus
obviating again the use of protective groups. The mild conditions and their
biodegradability makes them an excellent environmentally acceptable option as well.
The aldol reaction
In the aldol reaction4 the α-carbon of an aldehyde or ketone adds to the carbonyl
carbon of another. This reaction is called the aldol condensation when the product
(in some cases spontaneously) dehydrates in the course of the reaction, because
the new double bond is conjugated with the C=O bond, forming an α,β-unsaturated
product. The base most often used is OH-, but stronger bases are also applied.
O O OH O
R1 + base R1 R4
H R4
R2 R3 R2 R3
Figure 1. The aldol reaction
In principle, five combinations are possible for the aldol reaction, namely 1) two
identical aldehydes, 2) two identical ketones, 3) two different aldehydes, 4) two
different ketones and 5) an aldehyde and a ketone. Combination 1 has a favorable
equilibrium, in contrast with combination 2. Therefore combination 4 is hard to apply.
4
Combination 3 can give as much as four different products when both aldehydes
have an α-hydrogen. The last combination, also called the Claisen-Schmidt reaction,
is particularly straightforward and affords only one product when the aldehyde bears
no α-hydrogen.
The aldol reaction can create two contiguous stereogenic centers and,
consequently, four stereoisomers. Some control over the stereoselectivity can be
obtained by the use of preformed enolates with e.g. lithium5, Ti6, Zr7, Pd8 or Au9.
Normally, metal Z enolates give the two syn products whereas E enolates are mostly
non-stereoselective. Enantioselectivity can only be obtained by using either chiral
enol derivaties, chiral aldehydes or ketones, or both. Recent studies of catalytic
antibodies10 opened ways to obtain enantiomerically pure aldols via resolution.
Aldol reactions with aldolases
One need only to consider the remarkable complexity and variety of molecules
produced within the biosphere to realize that the enzymes which make these
possible would be of immense synthetic utility. Aldolases are part of a large group of
enzymes called lyases and are present in all organisms. They are mainly involved in
the metabolism of carbohydrates, but also of amino acids and hydroxy acids.
Complex carbohydrates play an important role in various types of biochemical
processes like growth, development, immune response, infection, cell adhesion,
metastasis and numerous signal transduction events. They are synthetically
accessible by application of aldolases. Although their function in vivo is often related
to the degradative cleavage of metabolites, the reactions are reversible and by
5
Chapter 1
choosing the proper conditions synthesis becomes favored. A number of compounds
are accessible11 in this way, particularly analogues of sugars like nitro-, amino-,
deoxy-, thio-, aza-, fluoro-, O-acyl, O-alkyl-sugars or aroma compounds and chiral
intermediates. The more than 30 different aldolases with different donor and
acceptor specificity that currently have been identified, are classified on the basis of
their donor into five groups (Table I).
Table I. Five groups of aldolases, adding either a two or a three carbon unit.
Aldolase Donor Example(s)

O
HO OPO3-2
DHAP-dependent RAMA, FruA from S.
carnosus
O
-
pyruvate-dependent O Sialic acid aldolase
(NeuAc aldolase)
O
-2
OPO3
-
O
phosphoenolpyruvate- NeuAc synthetase
dependent O
O
H2N
glycine-dependent OH L-threonine aldolase from
Candida humicola
O
acetaldehyde-dependent DERA from E. coli
Many of these enzymes are restricted more or less to their natural substrates but this
may not pose a limitation in their use since biologically meaningful compounds are
made by nature in the first place with these aldolases. Hence, applying them in the
synthesis of a derivative will be very advantageous since simple modification of one
6
of the substrates easily leads to modified end products. Selecting the right aldolase
is, therefore, an important stage in designing a synthetic route.
There are two classes12 of aldolases which vary in structure since they
separated early in evolution, although identical reactions are catalysed. Class I
aldolases, which are present in all groups of living organisms, from prokaryotes to
eukaryotes, are characterized by the formation of a Schiff’s base with the donor
substrate (see Figure 2). When the aldolase is treated with sodium borohydride after
the reaction with the donor, enzyme inactivation is observed. Class II aldolases
occur only in prokaryotes and lower eukaryotes such as yeast, algae and fungi. They
require a metal cofactor such as divalent zinc and are inactivated by chelating
compounds such as EDTA.
Remarkably, the binding of an aldehyde via an imine bond was also
successfully applied in a non-enzymatic aldol reaction, using LiN(iso-Pr)2 as the
base13. This is called a directed aldol reaction.
Asp-33
O
Lys-229
C -
O
HO +
NH O
O
-2 P
O3PO -
-
O
OH O H
HO +
+ H2N NH
NH3
NH3+ C
Lys-107 Lys-146 Arg-148
Figure 2. Imine intermediate in the active site of a type I fructose-1,6-bisphosphate
aldolase.
7
Chapter 1
Dihydroxyacetone phosphate dependent aldolases
The best studied aldolase is fructose-1,6-bisphosphate aldolase (FruA) from rabbit
muscle (RAMA). It accepts a wide range of aldehydes in addition to its natural
substrate, glyceraldehyde-3-phosphate, allowing the synthesis of many sugar
derivatives. This tetrameric enzyme is limited in its synthetic application by a low
operational stability14 and is now frequently replaced by the stable monomeric FruA
from Staphylococcus carnosus15,16. L-rhamnulose-1-phosphate aldolase46 (L-RhuA)
and L-fuculose-1-phosphate aldolase46,17 (L-FcuA), both from E. coli, are also
commercially available. Tagatose-1,6-bisphosphate aldolase (TagA) is not
commercially available.
OH O OH O
2- 2-
OPO3 OPO3
R R
OH OH
1 RCHO 2
FruA L-RhuA
+
O
2-
HO OPO3
DHAP
L-FcuA
TagA
OH O
OH O
2-
2- OPO3
OPO3 R
R
OH
OH
3 4
Figure 3. The four complementary DHAP dependent aldolases
8
These four aldolases (Figure 3) catalyse in vivo the reversible asymmetrical aldol
addition of dihydroxyacetone phosphate (DHAP) to D-glyceraldehyde-3-phosphate or
L-lactaldehyde, but they are expected to be as flexible towards the acceptor
substrate as RAMA. Two other DHAP-utilizing aldolases have been described,
namely erythrulose-1-phosphate aldolase18 which yields only one new chiral centre
since formaldehyde is the acceptor substrate and phospho-5-keto-2-deoxygluconate
aldolase19 which has the same stereospecificity as FruA but is not commercially
available.
The interest in the group of DHAP-dependent aldolases is mainly due to their
enantioselectivity. Not only are these DHAP aldolases enantioselective, the
combination of four different DHAP aldolases opens the possibility to preplan the
synthesis of any one of the four possible stereoisomers (Figure 3) normally
generated in this type of reaction, a degree of control unprecedented in organic
synthesis.
The scope of DHAP-dependent aldolases in organic synthesis is restricted by
three factors: namely 1) the acceptor substrate specificity, 2) the stereoselectivity
and 3) the circuitous and laborious synthesis of the donor substrate DHAP. We have
found that a wide variety of aldehydes can be used as the acceptor (see Chapter 2).
The influence of the structure of the acceptor on the stereoselectivity of the aldolases
needs to be determined since it is of vital importance to their utility as fine chemicals
or specialties (Chapter 3). A new preparation of the donor substrate DHAP from very
cheap starting materials was developed to overcome the disadvantages associated
with existing syntheses. The new method was integrated in an unprecedented four
steps, one-pot reaction (Chapter 4). The investigation of DHAP analogues was
mainly limited to the use of arsenate (Chapter 5).
9
Chapter 1
Pyruvate- and phosphoenolpyruvate-dependent aldolases
Pyruvate-dependent aldolases have catabolic functions in vivo, whereas their
counterparts employing phosphoenolpyruvate as the donor substrate are involved in
the biosynthesis of keto acids. However, both classes of enzymes can be used to
prepare similar keto acid products20.
N-acetylneuraminic acid (NeuAc) aldolase or sialic acid aldolase catalyzes the
reversible aldol reaction of N-acetyl-D-mannosamine (ManNAc) and pyruvate (Figure
4) to N-acetyl-D-neuraminic acid (Neu5Ac). The sialic acids are key cell-surface
determinants of mammalian glycoconjugates. They are found at terminal positions of
glycoproteins and glycolipids, which play an important role in biological recognition
processes. Neu5Ac is the best-known member of this special class of amino sugars.
Derivatives of Neu5Ac have been found to act as inhibitors of sialidases21.
HO O OH
NHAc CO2H
O HO
HO CO2H
O OH
HO AcHN
ManNAc OH NeuAc Aldolase OH HO Neu5Ac
Figure 4. N-acetylneuraminic acid (Neu5Ac) aldolase catalysed aldol reaction of N-
acetyl-D-mannosamine (ManNAc) and pyruvate to Neu5Ac.
In vivo its synthesis is accomplished by Neu5Ac synthetase through the irreversible
condensation of phosphoenol pyruvate (PEP) and N-acetylmannosamine. For the
synthesis of oligosaccharides and enzyme inhibitors Neu5Ac has been previously
isolated in small amounts from natural sources, such as cow’s milk, edible birds’
nests or eggs. Successful production of Neu5Ac was achieved by enzymatic
10
synthesis22 from ManNAc and pyruvic acid (space-time yield 470 g / L-1.d-1) with
Neu5Ac aldolase. The retro-aldol reaction is favored and an equilibrium favoring the
aldol reaction product is induced by providing pyruvate in excess (five- to seven-
fold). The excess pyruvate can be decomposed with pyruvate decarboxylase23 in
order to simplify the product isolation. The enzyme, immobilized onto Eupergit-C,
has been used in industrial scale production24 of Neu5Ac.
Glycine-dependent aldolases
Glycine-dependent aldolases catalyze the reversible aldol reaction of glycine with an
aldehyde acceptor to form β-hydroxy-α-amino acids. These compounds are naturally
occurring amino acids (threonine, serine) as well as components of many complex
natural products such as antibiotics and immunosuppressants (e.g. mycestericin D25,
vancomycin, echinocardin D, cyclosporin, katanosin, polyoxin D, empedopeptin etc.).
CHO
+ H2NCH2CO2H
X
NaOH
MeOH
L-aldolase
OH OH
CO2H CO2H
+
NH2 NH2
X X
L-threo D-threo
Figure 5. Selective retro aldol reaction by a L-specific threonine aldolase
11
Chapter 1
Both D- and L-isomers are biologically significant. Threonine aldolases are the best
described glycine-dependent aldolases. Only a few examples are known of its use
in bond-forming reactions26; yields and stereoselectivity are generally low for non-
natural substrates.
Threonine aldolases, with acetaldehyde as their natural acceptor, have been
used extensively in the resolution of racemic β-hydroxy-α-amino acids27 (Figure 5).
2-Deoxyribose-5-phosphate aldolase
The enzyme 2-deoxyribose-5-phosphate aldolase11 (DERA) catalyses in vivo the
reversible aldol reaction of acetaldehyde and D-glyceraldehyde 3-phosphate to form
D-2-deoxyribose-5-phosphate (Figure 6), the sugar moiety of DNA. Consequently
this type I aldolase is widespread in nature. It is the only aldolase that accepts two
aldehydes as substrates.
HO O
HO
DERA
-2
O3PO O + -2
O3PO O
H
D-glyceraldehyde- acetaldehyde OH
3-phosphate
-2
O3PO
O
OH 2-deoxyribose-5-phosphate
OH
Figure 6. Production of D-2-deoxyribose-5-phosphate by DERA.
12
Furthermore, DERA is an item of interest because it needs no phosphorylated
substrates, in contrast to DHAP aldolases. It has been cloned and overexpressed in
E. coli.
Acetaldehyde can be replaced by propanal, acetone and fluoroacetone, making
DERA suitable for production of β-hydroxy aldehydes or ketones. The extent of the
chain elongation is thus not fixed, but either 2- or 3-carbon long. Initial studies28
showed that the acceptor D-glyceraldehyde-3-phosphate can be replaced by other
aldehydes, albeit at a penalty of more than 99% of the reaction rate. Relatively large
amounts of enzyme are consequently required. The stability of the enzyme is one
point of investigation.
Transketolase
Transketolase is not an aldolase but a transferase. It catalyzes the transfer of a
two-carbon ketol moiety from a ketose to an aldose. Transketolase is particularly
attractive as an alternative to FruA, in combination with an appropriate choice of
substrates. When using an α-hydroxyaldehyde as acceptor and hydroxypyruvate as
donor, compounds with structures identical to those produced by FruA are obtained
(Figure 7). This reaction is of key importance in the synthesis of a wide range of
molecules in vitro and the ability to carry out the reaction stereospecifically means
that this biotransformation is of particular commercial interest.
The natural donor substrate, D-xylulose-5-phosphate, can be replaced by β-
hydroxypyruvate. This renders the reaction irreversible, because instead of
glyceraldehyde-3-phosphate carbon dioxide is released. Hydroxypyruvate29 is also
13
Chapter 1
cheaper and more stable than D-xylulose-5-phosphate, which adds to its
attractiveness as donor. The two natural acceptors D-ribose-5-phosphate and
erythrose-4-phosphate can be replaced by a whole range of aldehydes, which will be
discussed in Chapter 6.
OH O OH O
H O OH transketolase
+ OH
R R
-
O
O OH
CO2
O O OH O
1) FDP-aldolase
HO OPO32- OH
R H + R
2) acid phosphatase
OH
Figure 7. Transketolase catalysed reaction of hydroxypyruvate and an α-
hydroxyaldehyde.
Although transketolase requires thiamine pyrophosphate (TPP) and
magnesium(II) ions to function, only catalytic amounts of these cofactors are
required. Moreover, it does not necessarily require phosphorylated compounds as
substrates30 (although the natural substrates are phosphorylated) which is a
significant advantage over FruA. Even though the enzyme is absolute in its
requirement for the (R)-configuration of the hydroxy functionality at C2 of the
aldehyde, there seem to be no other stereochemical restrictions. Of main interest is:
how well does transketolase compete with or complement FruA regarding reaction
rate and conversion and how stereoselective is the enzyme? These questions are
14
addressed in Chapter 6.
The objectives and justification of the thesis
Aldolases doubtlessly constitute a group of catalysts with great synthetic potential. In
order to outgrow the shortcomings in the employment of aldolases in future
commercial processes, generic problems such as the design of non-natural
substrates, the exploration of enzymes from new species, multi-enzyme systems,
enzyme stabilization and application in the synthesis of fine and specialty chemicals
all have to be addressed.
General
The main objective of this thesis is to increase the applicability of enzyme catalyzed
aldol reactions. This implies that the enzymatic reactions have to be adapted for their
use in organic synthesis and economically viable industrial processes. The
development of practical procedures for the synthesis of enantiomerically pure
compounds based on enzyme catalysed aldol reactions will contribute to an
increased atom utilisation factor and, hence, will reduce waste streams and
enantiomeric ballast.
The exploration of enzymatic aldol reactions is justified by the central position
of aldol reactions in organic synthesis. For synthesizing and modifying carbohydrates
for example, the application of aldolases has already demonstrated its potential for
solving synthetic problems.
15
Chapter 1
Specific
The most common aldolases are dependent on dihydroxyacetone phosphate
(DHAP) as their exclusive donor reactant. The use of DHAP as the only donor
substrate restricts the spectrum of products since the wide variety of possible
acceptor substrates is combined with a single donor, whereas it would be preferred
to have a various choice of donor substrates as well. Hence, aldolases that do not
depend on expensive phosphorylated substrates have to be considered as
alternative to DHAP-dependent ones.
The phosphate anion plays a crucial rol in the binding of the reactants and in
the reaction mechanism. The use of DHAP is accompanied by consumption of
expensive organic phosphate that is ultimately transformed into inorganic phosphate,
which is not contributing to an economical proces. The replacement or efficient
preparation of DHAP would increase competitiveness of enyzme catalysed aldol
reactions relative to chemical ones. To investigate the benefits of phosphate free
aldol reactions, 2-deoxyribose-5-phosphate aldolase (DERA) was selected, although
this leads to different products than use of DHAP dependent aldolases. Exploration
of the transferase transketolase as an alternative to DHAP dependent FruA is a very
alternate approach of the DHAP “problem”, however very engaging since identical
products can be made.
References
1
E.J. Toone, E.S. Simon, M.D. Bednarski and G.M. Whitesides, Tetrahedron, 1989
45 (17) 5365-5422
2
Dr. Ir. J.M. Oelderik, Chairman of the IOP Catalysis, April 3, 1998
16
3
U. Krings, R.G. Berger, Appl. Microbiol. Biotechnol., 1998, 49, 1-8
4
J. March, Advanced Organic Chemistry, Wiley-Intersience, New York, 1992, 4th
Ed., 937-945
5
Arnett, Fisher, Nichols, Ribeiro, J. Am. Chem. Soc. 1990, 112, 801
6
Stille, Grubbs, J. Am. Chem. Soc. 1983, 105, 1664
7
D.A. Evans, L.R. McGee , J. Am.Chem.Soc, 1981, 103, 2876
8
Nokami, Mandai, Watanabe, Ohyama, Tsuji, J. Am.Chem.Soc. 1989, 111, 4126
9
Y. Ito, M. Sawamura and T. Hayashi, J. Am. Chem. Soc., 1986, 108, 6405-6406
10
B. List, D. Shabat and C.F. Barbas III, J. Am. Chem. Soc., 1999, 121, 7283-7291
11
For reviews see: Gijsen, H.J.M.; Qiao, L.; Fitz, W.; Wong, C.-H.; Chemical
Reviews 1996, 96(1), 443-469; M.D. Bednarski, E.S. Simon, ACS Sypmposium
Series 1990 466 Chapter 1 and 2; Wong, C.-H, Halcomb, R.L., Ichikawa, Y. and
Kajimoto, T., Angew. Chem. Int. Ed. Engl. 1995, 34, 412-432; Wang, P.G., Fitz,
W.; Wong, ChemTech, 1995, april, 22-32
12
Gefflaut, T., Blonski, C., Perie, J. and Willson, M., Prog. Biophys. Molec. Biol.
1995, Vol. 63, 301-340,
13
Wittig, Frommeld, Suchanek, Angew. Chem. Int. Ed. Engl. 1963, 2, 683
14
Brockamp, H.; Kula M. Tetrahedron Letters, 1990, 31 (49), 7123-7126.
15
Witke, C.; Götz, F. J. Bacteriol., 1993, 175 (22), 7495-7499.
16
Brockamp, H.; Kula, M.; Goetz, F. U.S. 5.162.221, 10 Nov. 1992.
17
Ozaki, A.; Toone, E.J.; Von der Osten, C.H.; Sinskey, A.J.; Whitesides, G.M. J.
Am. Chem. Soc. 1990, 112, 4970-4971
18
G.C. Charalampous, Methods Enzymol. 1962, 5, 283
19
W.A. Anderson, B. Magasanik, J. Biol. Chem. 1971, 246, 5662
20
D.G. Comb and S. Roseman, J.Biol. Chem. 1960, 235, 2529-2537
21
E. Zbiral, H. Brandstetter, R. Christian and R. Schauer, Liebigs Ann. Chem. 1987,
781-786
22
U. Kragl, M. Kittelmann, O. Ghisalba and C. Wandrey, Ann. N.Y. Acad. Sci.1995,
750, 300-305
23
C.-H. Lin, T. Sugai, R.L. Halcomb, Y. Ichikawa, C.-H. Wong, J. Am. Chem. Soc.
1992, 114, 10138
24
M. Mahmoudian, D. Noble, C.S. Drake, R.F. Middleton, D.S. Montgomery, J.E.
17
Chapter 1
Piercey, D. Ramlakhan, M. Todd and M.J. Dawson, Enzyme Microb. Technol.

1997, Vol. 20, april, 393-400
25
K. Shibata, K. Shingy, V.P. Vassilev, K. Nishide, T. Fumita, M. Node, T. Kajimoto,
C.-H. Wong, Tetrahedron Letters 1996, 37 (16), 2791-2794.
26
T. Kimura, V.P. Vassilev, F.-J. Shen and C.-H. Wong, J. Am. Chem. Soc. 1997,
119, 11734-11742
27
Celgene Corporation, US patent No. 5346828, 1994
28
C.F. Barbas III, Y.-F. Wang and C.-H. Wong, J. Am. Chem. Soc. 1990, 12, 2013-
2014
29
Large scale synthesis: K.G. Morris, M.E.B. Smith, N.J. Turner, M.D. Lilly, R.K.
Mitra and J.M. Woodley, Tetrahedron: Asymmetry, 1996, 7 (8), 2185-2188
30
U. Schörken and G.A. Sprenger, Biochim. et Biophys. Acta, 1998 1385, 229-243
18
19
20
2
Aldol reactions with DHAP dependent aldolases
“Kinetiek is de ethiek van de moderne tijd”
Peter Sloterdijk
Abstract
The synthetic application of enzymatic aldol reactions using the commercially
available DHAP-aldolases, D-fructose-1,6-bisphosphate (FruA), L-fuculose-1-
phosphate (FcuA) and L-rhamnulose-1-phosphate (RhuA) aldolase, was
investigated. All three aldolases were shown to have a broad specificity towards the
aldehyde acceptor. The kinetic properties of FruA from two bacterial and one
mammalian source were compared and appeared to be rather similar, but the
bacterial aldolases, were much more stable. The addition of dihydroxyacetone
phosphate to butanal catalyzed by FruA from Staphylococcus carnosus was
performed on a 5 mmol scale, demonstrating the applicability of aldolase catalysed
reactions.
21
Chapter 2
Introduction
The dihydroxyacetone monophosphate (DHAP) dependent aldolases (Figure 3, page
8), which in their natural role perform a retro-aldol reaction, have mainly been used
to synthesize a wide variety of modified carbohydrates.31,32,33,34 With their natural
substrates they exhibit essentially absolute control over the two newly created
stereogenic centers,35 each aldolase being specific for one of the four possible
stereoisomers. However, in order to be a useful synthetic tool aldolases must not
only function with a wide variety of aldehyde substrates but should also exhibit the
same degree of stereospecificity as is observed with their natural substrates. In this
chapter we have investigated the substrate spectrum of commercially available class
I DHAP aldolases.
Fructose-1,6-bisphosphate aldolase (FruA) from Staphylococcus carnosus36,37
and Staphylococcus aureus38 are known to be very heat and pH stable.39 The former
organism is a safer production host than the pathogenic S. aureus and, hence, its
aldolase has been studied more intensively. L-Rhamnulose-1-phosphate aldolase40
(RhuA) and L-fuculose-1-phosphate aldolase40,41 (FcuA), both from E. coli, were also
available. Tagatose-1,6-bisphosphate aldolase (TagA) is not commercially available
and was not investigated.
The best studied DHAP-depending aldolase is FruA from rabbit muscle (RAMA.
It accepts a number of unnatural aldehydes42 as substrate, but its practical
application is hampered by a low operational stability43. The substrate spectrum of
FruA from S. carnosus suggests that it has a similarly relaxed acceptor
specificity,43,44 but its kinetic properties have not been studied in any depth.
22
Acceptor substrate specificity of FruA, FcuA and RhuA
The three commercially available DHAP aldolases FruA∗, FcuA and RhuA were
compared in their ability to accept a number of non-natural aldehyde substrates.
Conversions (see Table I) show that FruA and RhuA have very similar properties.
With the exception of chloroacetaldehyde, methylglyoxal, phenylacetaldehyde and
pivaldehyde (10 to 20% difference), almost equal values were found. With FcuA
consistently lower conversions were found. Surprisingly glycolaldehyde, which
resembles the natural substrate (L-lactaldehyde) most, showed only a conversion of
22%.
Table I. Conversions of aldolases (determined by DHAP consumption)
acceptor FruA FcuA RhuA
Formaldehyde 81 26
Acetaldehyde 84 57 88
Propanal 67 60 76
Butanal 78 61 81
i-Butyraldehyde 78 69 81
Pivaldehyde 85 68 56
Glycolaldehyde 96 22 99
Glyoxal 98 43 98
Chloroacetaldehyde 76 68 92
Methylglyoxal 86 29 99
Phenylacetaldehyde 66 31 79
∗
From Staphylococcus carnosus.
23
Chapter 2
Similarities in kinetic properties are expected since all the class I DHAP
aldolases bind the same donor substrate by involving equivalent amino acid
residues. Consequently, the tolerance towards unnatural acceptor substrates is
expected to vary only within certain limits.
Properties of FruA from different sources
The FruAs from rabbit muscle (RAMA), S. carnosus and S. aureus were compared
for activity in the aldol reaction as well as the retro-aldol reaction. The cleavage of
the natural reactant –D-fructose-1,6-bisphosphate- was used to assay the retro-aldol
activity. The catalytic activity in the synthetic direction was measured with DHAP as
donor; D-glyceraldehyde-3-phosphate (GAP) and propanal were used as acceptors.
The reactions were monitored by a coupled enzymatic assay45 of DHAP. In this
assay DHAP is reduced to glycerol-3-phosphate with NADH using glycerol-3-
phosphate dehydrogenase. The decrease in absorption at 340 nm is proportional to
the amount of DHAP present.
The results of the assays are tabulated in Table II. All three aldolases
converted GAP and propanal at similar rates. Under the reaction conditions used,
the specific synthesis activity of the bacterial aldolases was approximately 50% of
their cleavage activity, whereas RAMA had the same activity in both reactions.
24
Table II. Specific activities of FruA from different sources.
splitting synthesis synthesis

aldolase Fruc-1,6-bisphosphate propanal + DHAP GAP + DHAP
U/mg enzyme U/mg enzyme U/mg enzyme
S. carnosus 26.7 12.2 12.2

S. aureus 16.8 7.2 8.5
RAMA 22 24 22.4
The acceptor specificity of the three aldolases was investigated by incubating DHAP
and a 100% excess of acceptor aldehyde with 1 PAU* of aldolase. Propanal was
chosen instead of the conventionally used glyceraldehyde-3-phosphate because of
better reproducibility, comparable kinetic properties and ease of use.
The relative initial rates and conversions of the reaction of DHAP with the
different aldehydes are displayed in Table III. The kinetic parameters of RAMA are
comparable with those of the FruA from S. carnosus or S. aureus, as might be
expected on the basis of the similarity in structure of the active sites. Conjugated
aldehydes such as benzaldehyde and acrolein (Vini = 0) are not substrates, with the
exception of pyridinecarboxaldehydes43. Longer alkyl chains result in a lower
(relative) Vini and conversion. Aldehydes bearing an electron-withdrawing group such
as chloroacetaldehyde have an activated carbonyl group, which increases the initial
rate. Bulky and aromatic groups result in lower rates and conversions. Hence, we
conclude that the generalizations made for the substrate spectrum of RAMA42 also
apply to bacterial type I FruA.
*
1 PAU (propanal aldolase unit) converts propanal at an initial rate of 1 µmol per minute.
25
Chapter 2
Table III. Relative initial reaction rates and conversions of aldol reactions with DHAP.
Aldehyde Relative Vini (%) Conversion (% (t ))

carnosus aureus RAMA carnosus aureus RAMA
formaldehyde 133 57 43 81 (2h) 16 (2h) 31 (3h)

acetaldehyde 108 141 103 84 (2.5h) 64 (3h) 87 (2h)
propionaldehyde 100 100 100 67 (4h) 68 (2h) 67 (3h)
butanal 37 72 23 78 (4h) 53 (4h) 69 (3h)
pentanal 36 27 20 64 (3h) 50 (4h) 67 (4h)
hexanal 32 22 22 62 (5h) 56 (4h) 63 (3h)
heptanal 8 11 9 a
octanal 2 <1 <1
isobutyraldehyde 8 16 10
isovaleraldehyde 7 9 8
glyoxylic acid 17 13 14 94 (4h) 96 (6h) 97 (5h)
glyoxal 65 21 16 98 (5h) 83 (6h) 89 (6h)
methylglyoxal 35 32 9 86 (6h) 64 (5h) 69 (6h)
glycolaldehyde 253 354 145 96 (0.5h) 95 (3h) 94 (0.7h)
chloroacetaldehyde 308 282 275 76 (1.5h) 85 (4h) 87 (0.5h)
phenylacetaldehyde 33 93 28 66 (2h) 37 (1h) 22 (1h)
2-pyridine 6 9 2
carboxyaldehyde
3-pyridine 7 2 3
carboxyaldehyde
4-pyridine 6 3 <1
carboxyaldehyde
[a] Undisplayed conversions could not be determined because reaction rates were slower then the
hydrolysis of DHAP
26
FruA stability under reaction conditions.
The reaction of DHAP with butanal was used to compare the stability of RAMA
(tetramer, 160 kD) and S. carnosus FruA (monomer, 40 kD). After the maximum
conversion was reached, the biocatalysts were concentrated and washed with buffer
in Amicon centripreps (cut-off 10 kD). The aldolase from S. carnosus proved to be
the most stable enzyme, exhibiting 45% of its initial activity compared with 0.1% for
RAMA. With glyoxylic acid as acceptor aldehyde, 100% of the initial activity was
observed with the bacterial FruA and 91% with RAMA. This yield of RAMA is
consistent with its reported44 rate of deactivation under these conditions of 2.47% per
h, which corresponds to 89% residual activity after 4.5 h. The washed and
concentrated aldolase from S. carnosus could readily be reused.
A 5 mmol scale reaction
In order to test the validity of the figures from Table I and III for scale-up of reactions,
the reaction of butanal and DHAP was conducted on a 5 mmol scale. The
concentration of DHAP was increased from 20 to 50 mM. The conversion of DHAP
was 78%. Treatment with wheat germ acid phosphatase (WGAP) followed by
acetylation gave 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose in 53% overall yield.
Hence we conclude that the values form a good basis for performing the reactions
on a larger scale.
27
Chapter 2
Conclusion
All the aldolases tested accepted a broad palette of aldehydes as substrate
analogues. Comparison of FruA, FcuA and RhuA showed that FcuA functions less
well when non-natural substrates are used, while FruA and RhuA both performed
well and with similar profiles. More than twenty aldehydes were found to be
substrates for the aldolases from rabbit muscle, S. carnosus and S. aureus. The
bacterial aldolases were more stable than RAMA, especially with more apolar
aldehydes. Reaction rates and chemical yields were comparable. Isolation of
enzyme from the reaction mixture is feasible with high recovery of activity for S.
carnosus aldolase and makes it more efficient than RAMA.
Experimental
General
1 13
H and C spectra were recorded in CDCl3 at 300 Mhz on an Oxford NMR 300. UV
spectroscopy was performed with a Varian Cary 3 Bio equipped with a Cary
temperature controller. DHAP was prepared from its ethyl hemiacetal dimer barium
salt (Fluka).
Aldol reaction
Aldol reactions were performed in 1 ml 50 mM TRIS buffer pH 7.6 containing 20 mM
28
DHAP and 40 mM aldehyde, reactions were initiated by adding 1.0 U aldolase. At
certain intervals 50 µl aliquots were taken and quenched with 15 µl 7% perchloric
acid. After 30 min 10 µl 1M NaOH was added followed by 175 µl of 50 mM TRIS pH
7.6. This neutralized mixture was then assayed for DHAP. The reaction was followed
until only background hydrolysis of DHAP was measured (0.0047 mM/min).
FruA activity assay
The reaction products, DHAP and GAP were assayed with a coupled enzyme
system. To 1.95 ml 50 mM Tris-buffer pH 7.6 containing 0.16 mM NADH in a 2 ml
cuvette, were added 20 µl 190 mM fructose-1,6-bisphosphate, 20 µl of a mixture
containing 1.25 unit D-glyceraldehyde-3-phosphate dehydrogenase and 12.5 unit
triose-1-phosphate isomerase. Then 50 µl diluted aldolase was added and
absorption monitored at 25 °C. 1 unit (U) aldolase converts 1 µmol fructose-1,6-
bisphosphate per minute.
DHAP-assay
DHAP was assayed with a coupled enzyme system: reduction of DHAP with NADH-
consuming glycerol-3-phosphate dehydrogenase14. From a diluted DHAP solution,
50 µl was added in a quartz cuvette containing 1.95 ml 50 mM TRIS pH 7.6, 0.16
mM NADH, 1.25 U D-glycerol-3-phosphate dehydrogenase and 12.5 U triose-1-
phosphate isomerase. The absorption was monitored at 340 nm at 20 °C. The molar
29
Chapter 2
adsorption coefficient taken was 6.22 l.mmol-1cm-1. Reaction rates were calculated
with the method of least squares. Relative rates were calculated by deviding Vpropanal
by Valdehyde and multiplying by 100.
At intervals of 0.5 - 5 minutes during 3 - 30 minutes 50 µl aliquots were taken
and quenched with 15 µl 7% perchloric acid. After 30 min 10 µl 1M NaOH was added
followed by 175 µl of 50 mM TRIS pH 7.6. This neutralized mixture was then
assayed for DHAP.
Enzyme recovery
In a 25 ml roundbottomed flask equipped with a magnetic stirrer which contained 10
ml 50 mM Tris-buffer pH 7.6, 50 mM DHAP and 100 mM aldehyde, 10 U FruA
(lyophilized RAMA and S. carnosus FruA) were added and stirred for 3 h at RT. After
this the reaction mixture was cooled to 4 °C, poured into (10 or 30 kDa) Amicon
centripreps, centrifuged, washed with 2 x 10 ml Tris-buffer and concentrated to 0.5
ml. The residual activity was determined by adding 10 µl concentrated recovered
enzyme to 1.95 ml 50 mM Tris-buffer pH 7.6 and 0.16 mM NADH in a 2 ml cuvet.
Subsequent addition of 20 µl mixture of 1.25 U glyceraldehyde-3-phosphate and
12.5 U triosephosphate isomerase exposed any residual DHAP or product if present.
Then the reaction was started by addition of 20 µl 190 mM fructose-1,6-bisphosphate
and absorption was monitored at 25 °C.
30
Synthesis of 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose
A 50 ml solution containing 50 mM Tris pH 7.6, 50 mM DHAP, 100 mM butanal and
48 U S. carnosus aldolase was stirred at room temperature. Conversion was 78%
(based on consumption of DHAP) after 4 h. Treatment with wheat germ acid
phosphatase (WGAP) overnight at pH 5.5 gave after extraction with ethylacetate
214 mg (c.y. 53% yield) product. Acetylation with 20 ml pyridine and 10 ml acetic
anhydride and subsequent column chromatography (EtOAc/Hexane 1:3) yielded 250
mg (67%) of colorless oil. [α]D25 = 25.0° (c 1.05, CHCl3).
1
H-NMR (300 Mhz, CDCl3)
δ=4.75 (d, H-1a), 4.88 (d, H-1b); 5.31 (d, H-3); 5.37 (m, H-4); 1.55 (m, H-5a), 1.33
(m, H-5b); 1.55 (m, 2H-6); 0.94 (t, 3H-7); 2.22 (s, 3H-1’); 2.16 (s, 3H-3’); 2.078 (s,
3H-4’).
13
C-NMR (300 Mhz, CDCl3)
δ=66.79 (C-1), 198.41 (C-2), 77.15 (C-3), 71.52 (C-4), 32.38 (C-5), 18.51 (C-6),
13.71 (C-7), 170.07 (CO-1’), 20.76 (CH3-1’), 170.19 (CO-3’), 20.42 (CH3-3’), 169.81
(CO-4’), 20.36 (CH3-4’)
References
31
J. Peters, H. Brockamp, R. Minuth, M. Grothus, A. Steigel, M. Kula and L.
Elling, Tetrahedron: Asymmetry 1993, 4 (6), 1173-1182

32
C.-H Wong and G.M. Whitesides, J. Org. Chem. 1983, 3199-3205.
31
Chapter 2
33
J.R. Durrwachter, D.G. Drueckhammer, K.. Nozaki, H.M. Sweers and C.-H. Wong,
J. Am. Chem. Soc. 1986, 108, 7812-7818.

34
K.K.-C. Liu, T.Kajimoto, L. Chen, Z. Zhong, Y. Ichikawa and C.-H. Wong, J. Org.
Chem. 1991, 56, 6280-6289.

35
H.J.M. Gijsen, L. Qiao, W. Fitz and C.-H. Wong, Chemical Reviews 1996, 96(1),
443-469
36
C. Witke and F. Götz, J. Bacteriol. 1993, 175 (22), 7495-7499.
37
H. Brockamp, M. Kula and F. Goetz, U.S. 5.162.221, 10 nov. 1992.
38
F. Götz, S. Fischer and K.-H.Schleifer, Eur. J. Biochem. 1980, 108, 295-301.
39
C. De Montigny and J. Sygusch, Eur. J. Biochem. 1996, 241, 243-248.
40
W.-D. Fessner, G. Sinerius, A. Schneider, M. Dreyer, G.E. Schulz, J. Badia and J.
Aguilar, Angew. Chem. 1991, 103, 596-599

41
A. Ozaki, E.J. Toone, C.H. Von der Osten, A.J. Sinskey and G.M. Whitesides, J.
Am. Chem. Soc. 1990, 112, 4970-4971

42
M.D. Bednarski, E.S. Simon, N. Bischofberger, W.-D. Fessner, M.-J. Kim, W.
Lees, T. Saito, H. Waldmann and G.M. Whitesides, J. Am. Chem. Soc. 1989, 111,
627-635.
43
H. Brockamp and M. Kula, Tetrahedron Letters, 1990, 31 (49), 7123-7126.
44
H. Brockamp and M. Kula, Appl. Microbiol. Biotechnol., 1990, 34, 287-291.
45
H.U. Bergmeyer, Methods of enzymatic analysis, Verlag Chemie, Mannheim 1984;
Vol. IV p. 342-350
32
33
34
3
The stereochemistry of DHAP aldolases
Abstract
The four DHAP dependent aldolases – D-fructose-1,6-bisphosphate (FruA), L-
fuculose-1-phosphate (FcuA), L -rhamnulose-1-phosphate (RhuA) and D-tagatose-
1,6-bisphosphate (TagA) aldolase - constitute a complementary toolkit for
predesigning chirality in carbohydrates. Since the stereochemical outcome may differ
when substrates in the aldol reaction are modified, it is necessary to find
unambiguous evidence for the configuration and optical purity of the products. We
used chiral GC and developed a new enzymatic assay to determine the
stereoselectivity of the aldolases towards a number of acceptor substrates.
Determination of isomer formation revealed that D-fructose-1,6-bisphosphate
aldolase from S. carnosus and L-rhamnulose-1-phosphate aldolase from E. coli are
highly stereospecific.
35
Chapter 3
Determination of stereoselectivity with chiral GC
The stereochemical preferences of aldolases have never been investigated in detail.
Kinetic studies (Chapter 2) have demonstrated that almost any aldehyde is accepted
with a preference for α-hydroxyaldehydes or activated ones such as
chloroacetaldehyde. The selectivity of the aldolase for each of the four possible
stereoisomers remains an unanswered question, however, although it is of crucial
importance in the context of synthetic application. The reaction of butanal with DHAP
catalysed by FruA from S. carnosus was chosen to investigate the steric course of
the aldol reaction in detail. The reaction products were examined, after
dephosphorylation and acetylation, by means of chiral GC analysis. Previously46,
diastereomers were identified by NMR, but chiral GC has the advantage that
enantiomers can also be discriminated.
OAc O OAc O
OAc OAc
O +
HO OPi 1. FruA S. carnosus OAc OAc
(78% yield)
1 (3S,4R) (90%) 2 (3R,4S) (< 0.5%)
+
2. WGAP (53% yield)
CHO 3. Ac2O/pyridine
(67% yield)
OAc O OAc O
OAc OAc
+
OAc OAc
3 (3S,4S) (6%) 4 (3R,4R) (4%)
Figure 1. Aldol reaction of butanal and DHAP catalysed by S. carnosus FruA.
Selectivities are in parentheses.
36
The reaction of butanal and DHAP, catalyzed by FruA afforded a major (90%
selectivity) product to which we assigned the (3S,4R) configuration 1 (Figure 1) on
the basis of the known steric preferences of the catalyst.47 Two minor, presumably
stereoisomeric, products were formed with 6% and 4% selectivity.
In order to elucidate their structures we synthesized two of the remaining three
possible stereoisomers of the aldol adduct independently. Reaction of butanal and
DHAP catalyzed by RhuA46 afforded the (3R,4S) product 2; FcuA 46,48 afforded the
(3R,4R) stereoisomer 4 (Figure 2). The (3S,4S) stereoisomer 3 could not be
synthesized directly, because no D-tagatose-1,6-bisphosphate aldolase was
available. However, epimerisation of 2 with sodium acetate in methanol at room
temperature afforded, after 4 h, an equilibrium mixture of 45% 2 and 55% of its 3-
epimer 3. Epimerisation presumably proceeds via enolisation and is also observed
under acidic conditions (p-TosOH was added). Hence, all four aldol adducts of
butanal and DHAP could be resolved by chiral GC.
To reconfirm the structural assignments, 1 was similarly subjected to
epimerisation; after 2 h at room temperature an equilibrium mixture of 60% 1 and
40% of its 3-epimer 4 was obtained. On the basis of chiral GC the minor products of
FruA from S. carnosus were identified as the anti stereoisomers 3 (6% selectivity)
and 4 (4 % selectivity). Because the stereogenic centre at C-4 is not readily
susceptible to epimerisation, the formation of 3 must reflect a lack of complete
stereoselectivity at C-4, In contrast, epimerisation of 1, either during the reaction of
the work-up, could contribute to the observed formation of 4. NMR analysis of the
(phosphorylated) reaction products suggests that S. carnosus aldolase forms 8.5%
of the anti stereoisomers 3 and 4. Hence, the remaining 1.5% presumably results
from epimerisation during the work-up.
37
Chapter 3
OAc O OAc O
OAc OAc
OAc OAc
1 2
1) FruA CHO 1) L-RhuA
2) WGAP
2) WGAP
3) Ac2O
O 3) Ac2O/pyridine
/pyridine
2-
NaOAc HO OPO3
MeOH NaOAc
MeOH
1) L-FcuA
2) WGAP
3) Ac2O/pyridine
OAc O OAc O
OAc OAc
OAc OAc
4 3
Figure 2. Structural assignments of the four possible aldol adducts.
The use of GC requires derivatisation of reaction products. Unfortunately, this
method turned out to be rather inflexible and labor intensive. For every single
aldehyde the derivatisation method must be altered, because of too short or too long
retention times resulting in diminished peak separation. Also, new product samples
with specific configuration must be synthesized each time. However, in the case of
butanal, for the first time, a complete resolution of DHAP aldolase products was
established.
38
Determination of stereoselectivity with an enzymatic assay
To bypass the limitations encountered with chiral GC, a new coupled enzymatic
method, based on the reversibility of the aldol reaction, was developed to detect the
stereoisomers formed by DHAP aldolases. This simple coupled enzymatic assay
makes use of the known stereopreference46,47 of the commercially available
aldolases in retro-aldol reactions (Figure 3). Three of the four stereoisomers could be
detected directly, the fourth one was calculated. The method does not require
derivatisation or complete work-up of the product samples, only removal or
inactivation of the biocatalyst.
OH O OH O
2- 2-
OPO3 OPO3
R R
OH OH
1 RCHO 2
FruA L-RhuA
+
O
OH
2-
HO OPO3 2-
NADH HO OPO3
DHAP GDH
L-FcuA
TagA
OH O OH O
2- 2-
OPO3 OPO3
R R
OH OH
3
4
Figure 3. Retro aldol reactions and reduction of DHAP
The assay is as elementary as the aldol reaction itself and incorporates three steps
(Figure 3): 1) removal of aldolase, 2) reduction of any residual DHAP and 3) isomer
39
Chapter 3
detection. Steps 2) and 3) are repeated for each stereoisomer. First, after the initial
synthetic aldol reaction has reached its maximum conversion the aldolase activity is
quenched by addition of acid or membrane filtration, which was, in our hands, the
fastest way to remove the aldolase activity. Secondly, a sample of the aldolase-free
mixture containing the aldol adducts is added to a suprasil cuvette containing NADH.
Addition of glycerol-3-phosphate dehydrogenase (GDH) mediates reduction of any
residual DHAP49 present in the mixture (∆A1, Figure 4A). Third, one of the aldolases
is added to selectively cleave one of the four possible products to a corresponding
amount of DHAP, which is reduced in situ by NADH in the presence of GDH (∆A2,
Figure 4A). The reduction step renders the retro-aldol reaction irreversible, hence,
the amount of reduced DHAP is equal to the amount of NADH consumed, which is
monitored by UV.
1.3 GDH added

1.4 GDH added
1.3 1.2
1.2 1.1
∆A1
absorption
absorption
1.1 ∆A1
1 FcuA added
1 FruA added
0.9 0.9
0.8 0.8 ∆A2
∆A2
0.7 0.7
0.6
0.5 0.6
0 10 20 30 40 0 10 20 30
t (min)
t (min)
A B
Figure 4. Example of detection of the butanal-DHAP adduct made with RhuA. (A)
The absorption of NADH is measured at 340 nm. Addition of GDH yields ∆A1 which
is correlated to the amount of residual DHAP. Subsequent addition of aldolase yields
∆A2 which is correlated to the amount of stereoisomer with the fructose
configuration. (B) Distinction between two isomers can be made because of different
retro aldol reaction rates. ∆A2 is now extrapolated from the graph.
40
It may seem paradoxical to distinguish between stereoisomers using the
same enzyme which produced them, but since the minor isomers are cleaved
sluggishly (see Figure 4B), there is a clear distinction in the assay between the four
aldol adducts. Since absorption is measured versus time, the concentration is easily
extrapolated from an absorption-time plot.
Since TagA is not commercially available, only three of the four possible
stereoisomers could be detected via this retro aldol reaction. With FruA, FcuA and
RhuA the corresponding isomers are detected; the missing part, based on consumed
DHAP from the synthetic aldol reaction, is equal to the product with tagatose
configuration. Otherwise, this information can be obtained by allowing the initial
aldolase to convert all the four isomers it produced, thus determining the sum of
concentrations.
In practice this method of analyzing stereoisomers is very convenient, since
no work-up of the reaction product or derivatisation is necessary and small samples
can be taken during the synthetic reaction and analyzed in minutes. In fact, even
when only one aldolase from the set of four is available, this method would prove
itself useful since – besides the significant importance of product formation
monitoring – it is still possible to determine the fraction of the main isomer produced
by this enzyme, based on the total DHAP consumption.
Comparison of methods
The results of the enzymatic assay were compared (see Table I) with those
obtained by stereoisomer detection with gas chromatography and NMR.46
41
Chapter 3
Examination of the distribution of stereoisomers found for the reaction of butanal and
DHAP with GC on the one hand and the enzymatic assay on the other revealed only
small differences. These are probably due to some epimerisation taking place during
the derivatisation necessary for GC analysis and underline the strength of the
enzymatic assay. According to our retro-aldol assay the reaction of acetaldehyde
with DHAP catalyzed by RhuA yields about 2% less of the syn isomers than found by
NMR analysis. For butanal this discrepancy is 1%. We note that NMR analysis also
requires work-up of the samples which might lead to a certain deviation in values.
Table I. Enzymatic assay versus GC and NMR analysis. Configurations can be
found in the Figures 1, 2 and 3
Acceptor Analysis Aldolase Configuration (%) Conversion
1 2 3 4
butanal GC FruA 90 0 4 6 78%

enzymatic FruA 89.7 0 2.6 7.7 80%
acetaldehyde NMR RhuA 69 31 84%
enzymatic RhuA 2 65 21 12 88%
butanal NMR RhuA 83 17 82%
enzymatic RhuA 14 68 0 18 81%
Stereospecificity of FruA and RhuA
The steric course of the FruA mediated aldol reaction of butanal has been elucidated
by gas chromatography. The steric preferences of RhuA and FcuA were studied with
a number of acceptors46, but since NMR was used only proportions of
42
diastereoisomers could be measured. With our new method it was possible to
quickly scan a range of substrates, delineating a more general picture of the
properties of aldolases regarding stereospecificity.
Table II. Stereochemical preferences of DHAP aldolases. Bold figures

refer to the configurations in Figures 1, 2 and 3.
RCHO FruA RhuA

R 1 2 3 4 1 2 3 4
CH3 88 0 3.5 8.5 1.5 64 21.5 13.0

CH2CH3 96 0 0 4.0 5.0 81 13 1
CH2CH2CH3 90 0 2.5 7.5 14 68 18 0
CH(CH3)2 92.5 0 0 7.5 5.0 68.5 10 16.5
C(CH3)3 96 0 0 4.0 9.5 90.51 0.0 1
CH2OH 96 0 0 4.0 0.3 99 0.2 0.5

CHO 991 0 1.0 1
0.5 99.51 0.0 1
COCH3 95.51 0 4.5 1

3.0 89.5 0.0 7.5
CH2Cl 94 0 4 2.0 0.0 96.5 1.5 2.0
CH2Ph 97.5 0 1.5 1.0 3.0 75.0 20.5 1.5
1
The retro-aldol reaction was too slow for the separate determination of the Tag-stereoisomer
The steric course of the FruA mediated reaction of DHAP and ten different
aldehydes, was investigated using our enzymatic assay (Table II). We found that the
FruA from S. carnosus is highly, but far from absolutely, stereospecific for products
with the fructose configuration, which were formed for > 95% in the majority of the
cases. The specificity for the fructose configuration was less pronounced for the
aldol reactions of acetaldehyde and butanal. However, a high selectivity for the
43
Chapter 3
fructose configuration was observed with related non-polar aldehydes, such as
trimethylacetaldehyde (pivaldehyde) and phenylacetaldehyde.
The amounts of the major, “fructose” stereoisomers in the aldol adducts of
glyoxal and methylglyoxal could not be determined directly owing to the extremely
low rate of the corresponding retro-aldol reactions. The cleavage of the adduct of
glyoxal, for example, was 250 times slower than that of the phenylacetaldehyde
adduct. The sluggish retro-aldol reactions prevented a reliable observation of the
transition from cleavage of the major stereoisomers to that of the minor ones.
However, the stereoisomers with L-rhamnulose and L-fuculose configuration could be
measured and from these the sum of the products with fructose and tagatose
configuration was calculated.
The minor products from the FruA catalyzed reactions mainly had the tagatose
configuration, in particular with acetaldehyde, butanal and dimethylacetaldehyde.
The optical antipodes of the main product, the stereoisomers with the L-rhamnulose
configuration, were not detected at all.
The RhuA catalyzed aldol reactions of the same range of aldehydes were in
general only moderately stereoselective. A high selectivity for the “rhamnulose”
product was only observed with glycolaldehyde and glyoxal. We note that in the
latter case, as well as 2,2-dimethylpropanal, the “rhamnulose” stereoisomer could
not be determined directly because the retro-aldol reaction was too slow.
Consequently, only the sum of the products with L-rhamnulose and tagatose
configuration could be calculated for these reaction products. Contrary to FruA, the
RhuA catalysed reaction also formed detectable amounts (up to 9.5%) of the optical
antipode of the major stereoisomer (with fructose stereochemistry).
44
Conclusion
GC analysis showed that the reaction of butanal with DHAP catalysed by FruA from
S. carnosus afforded the (3S,4R) stereoisomer 1 as the major product (90%)
together with 6% of the (3S,4S) isomer 3 and 4% of the (3R,4R) isomer 2. In contrast
to this method the enzymatic assay is a direct and flexible procedure for the
quantitative determination of stereoisomers formed in aldol reactions catalysed by
DHAP dependent aldolases. It obviates the need for work-up and/or derivatisation; is
much less labor-intensive and does not entail any risk of product epimerisation. The
procedure is universal for DHAP aldol adducts, consequently a whole range of
substrates can be investigated. This proved unambiguously, for the first time, that
the steric preference of FruA and RhuA for the newly created stereogenic center at
C-(3) is far from absolute. The method is also useful for fast monitoring of aldolase
catalysed syntheses. Its main limitation is when the retro-aldol reaction is too slow to
discriminate reliably between cleavage of the major and minor products.
Experimental
General
1
H and 13C spectra were recorded in CDCl3 at 300 Mhz on an Oxford NMR 300.
UV spectroscopy was performed with a Varian Cary 3 Bio equipped with a Cary
temperature controller. Chiral gaschromatography was performed with a
diacetyltertbutylsilyl-β-cyclodextrine (50 m x 0.25 mm, df 0.25 µm) column. DHAP
was prepared from its ethyl hemiacetal dimer barium salt (Fluka).
45
Chapter 3
Acetylation of GC-samples
Acetylation prior to GC-injection was conducted by solving 2 to 3 mg of analyte in 0.1
ml pyridine in a 2 ml sample bottle. After addition of 50 µl acetic anhydride the
mixture was shaken for 10 sec at room temperature. Immediately 1 ml diethylether
and 0.3 ml 4N HCl was added to neutralize the mixture and extract the acetylated
compound. The water layer was removed and the organic layer was dried with
sodium sulfate. This solution was then ready for injection. Under these conditions no
racemisation occurred, only after subjecting the analyte for more than 5 min to the
acetylation mixture epimerisation was observed.
Preparation of product samples
Butanal and DHAP were reacted following the procedure of the preparation of 1,3,4-
tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose with 50 U L-fuculose-1-P aldolase or 50 U L-
rhamnulose-1-P aldolase. After extraction with ethylacetate the products were
acetylated according to the acetylation of GC-samples.
Aldol reaction
Aldol reactions were performed in 0.5 ml 50 mM TRIS buffer pH 7.6 containing 20
mM DHAP and 50 mM aldehyde and were initiated by adding 5 U aldolase. After
complete conversion the reaction mixture was centrifuged in an eppendorf
46
centrifuged and transferred into Microcon Centrifugal filter (cutoff 10 kD) and again
centrifuged for 10 minutes. The filtrate contained no aldolase activity. Alternatively,
the complete mixture can be quenched with 20 µl 70% perchloric acid and allowed to
stand for 30 minutes. Precipitate was centrifuged off. Neutralization was not
necessary.
Retro aldol reaction assay
This assay is a modification of the DHAP coupled enzyme assay. DHAP was
assayed by reduction with NADH-consuming glycerol-3-phosphate dehydrogenase.
From the deproteinated reaction mixture 10 to 20 µl was added in a quartz cuvette
containing 1.97 ml 50 mM TRIS pH 7.6, 0.16 mM NADH, 1.25 U glyceraldehyde-3-
phosphate and 12.5 U triose-1-phosphate isomerase. The absorption was monitored
at 340 nm at 20 °C. The molar adsorption coefficient taken was 6.22 l.mmol-1cm-1.
Reaction rates were calculated with the method of least squares. After complete
conversion of residual DHAP, retro aldol reactions were initiated by adding 10 µl
aldolase (5 units). This was repeated for each of the three aldolases.
References
46
W.-D. Fessner, G. Sinerius, A. Schneider, M. Dreyer, G.E. Schulz, J. Badia and J.
Aguilar, Angew. Chem. 1991, 103, 596-599

47
H. Brockamp and M. Kula, Appl. Microbiol. Biotechnol. 1990, 34, 287-291.
47
Chapter 3
48
A. Ozaki, E.J. Toone, C.H. Von der Osten, A.J. Sinskey and G.M. Whitesides,
Am. Chem. Soc. 1990, 112, 4970-4971

49
H.U. Bergmeyer, Methods of enzymatic analysis, Verlag Chemie, Mannheim 1984;
Vol. IV p. 342-350
48
49
50
4
In situ generation of DHAP: cascade catalysis
Abstract
A total of four enzymatic steps were combined, in a one-pot reaction, to synthesize
carbohydrates starting from glycerol. This approach avoided restraints normally
associated with the preparation of DHAP. First, phosphorylation of glycerol by
reaction with pyrophosphate in the presence of phytase at pH 4.0 in 95% glycerol
afforded racemic glycerol-3-phosphate in 100% yield. The L-enantiomer of the latter
underwent selective aerobic oxidation to dihydroxyacetone phosphate (DHAP) at pH
7.5 in the presence of glycerolphosphate oxidase and catalase. Subsequently, D-
fructose-1,6-bisphosphate aldolase catalyzed the aldol reaction of DHAP with
butanal. Finally, dephosphorylation of the aldol adduct was mediated by phytase at
pH 4 affording 5-deoxy-5-ethyl-D-xylulose in 57% yield from L-glycerol-3-phosphate.
Switching off the activity of the phosphatase by a pH change during oxidation and
aldol reaction resulted in better results than integral in situ phosphorylation.
51
Chapter 4
Synthesis of dihydroxyacetone phosphate
Enzyme catalyzed aldol addition is a powerful methodology for building carbohydrate
derivatives.50,51 Aldolases, which in their natural function perform a retro-aldol
reaction, are capable of catalyzing a large variety of carbon-carbon bond constituting
reactions52. The group of aldolases which depend on dihydroxyacetone phosphate53
(DHAP) as the donor substrate is particularly attractive as total control over the
stereochemical outcome of the aldol reaction is obtained (see Figure 1, page 8).
They are hampered, nevertheless, in their practical application by the lack of cheap
and efficient access to DHAP.
DHAP was first prepared53 by enzymatic dismutation of D-fructose-1,6-
bisphosphate (Figure 1, below). D-Fructose-1,6-bisphosphate aldolase (FruA)
mediates its retro aldol reaction to equal amounts of DHAP and glyceraldehyde-3-
phosphate. A second aldehyde can then be coupled to DHAP by the same aldolase.
The overall equilibrium may however not be favorable for synthesis and the workup
is complicated by residual substrate54.
It is possible to phosphorylate dihydroxyacetone directly54 with POCl3 but this
results in complex reacton mixtures. A nine step synthesis was first developed from
3-chloro-1,2-propanediol55, with an overall yield of only 15%. The synthesis was
improved by switching to dihydroxyacetone as starting material (4 steps). The
present-day preparation of DHAP proceeds via the dimer of dihydroxyacetone, the
form in which it excists in non-aqueous media, through ketalisation with
triethylorthoformate to 2,5-diethoxy-p-dioxane-2,5-dimethanol56. The phosphorylation
of this intermediate has been extensively investigated and can be performed in
different ways but, because deprotection of 2,5-diethoxy-p-dioxane-2,5-diphosphate
52
to DHAP is at best 66%, overall yields remain low. By improving the work-up
conditions for 2,5-diethoxy-p-dioxane-2,5-dimethanol higher yields of DHAP were
obtained (Colbran57 27%, Effenberger58 35%, Pederson59 55%, Jung60 61%). An
efficient final hydrolysis step is possible by using 3-acetoxy-2,2-dimethoxypropyl
phosphate61 (95%), but this compound requires a five-step synthesis with an overall
yield of only 54%. The hydrolysis of 3-bromo-2,2-dimethoxy-1-propyl dibenzyl
phosphate62 yields only 56% DHAP. Summarizing, the chemical synthesis of DHAP
O OPO3-2
OEt
EtO
-2
O3PO O
2,5-diethoxy-p-dioxane-2,5-diphosphate
O
H3CO OCH3
HO OPO3-2 HO OH
+
H
+ ATP
H
glycerol kinase
HO OPO3-2
DHAP
GPO FDP aldolase
TPI
catalase pyrophosphate
OH phytase
GPO/catalase
-2
O3PO OPO3-2
HO OPO3-2 O
OH
OH
OH
HO OH OH
involves lengthy procedures and modest over-all yields.
Figure 1. Two chemical and four enzymatic routes to DHAP.
The enzymatic preparation63 from dihydroxyacetone was accomplished for the
53
Chapter 4
first time by phosphorylation with ATP54 catalysed by glycerol kinase, in combination
with a regeneration system composed of acetyl phosphate and acetate kinase in
83% yield. A method that is very efficient as regards cost, number of steps and
chemical yield proceeds via the oxidation of L-glycerol-3-phosphate mediated by

64
glycerol phosphate oxidase (GPO) . Accummulation of hydrogen peroxide is
suppressed by adding catalase. Racemic glycerol-3-phosphate can be used,
notwithstanding that GPO exclusively oxidizes the L-enantiomer, making the process
more cost-effective than with pure L-glycerol-3-phosphate. Although conversions run
up to 95%, inhibition of GPO by DHAP is even in the best case substantial (20%
residual activity at [DHAP] = 100mM). However, under these conditions, enzymatic
aldol reactions can be performed in situ, thus avoiding inhibition of GPO and making
yields of up to 96% feasible, the record for DHAP synthesis. The enzymatic
phosphorylation of glycerol cannot be coupled to this system, because of the low
oxygen tolerance of glycerol kinase.
Phosphorylation with a phospholipase was also reported65. The alcohol
function in dihydroxyacetone can replace the polar choline head group of the natural
reactant in a phospholipase D mediated phosphorylation. Subsequent hydrolysis by
phospholipase C then yields DHAP and the corresponding 1,2-diacyl glycerol. This
phosphorylation has an efficiency of only 72% and thus cannot compete with the
other enzymatic methods.
A simple and economical route to DHAP – and therefore to carbohydrates –
should involve a limited number of steps from inexpensive starting materials. The
reaction(s) should preferably be enzymatic as this would obviate the need for
protection and deprotection steps inherent in chemical phosphorylation.66 Moreover,
54
chemical phosphorylation is likely to be incompatible with subsequent enzymatic
steps. Owing to the high energy content of the phosphate ester bond, DHAP
synthesis generally requires an organic phosphate. Alternatively, DHAP can be
prepared by enzymatic oxidation of L-glycerol-3-phosphate which is, in vivo
produced by glycerol kinase catalysed reaction of glycerol with ATP which would
again call for a source of organic phosphate for regeneration. These obstacles can,
in principle, be circumvented by using a phosphatase as the phosphorylation catalyst
and pyrophosphate as an inexpensive source of phosphate.67,68
Combination of such a phosphorylation of glycerol with in situ oxidation to
DHAP and subsequent enzymatic aldol condensation and dephosphorylation would
provide a one-pot procedure from glycerol to carbohydrates (Figure 2).
Integral in situ reaction
Initially, we examined the production of DHAP in a mixture of pyrophosphate,
glycerol / water, phosphatase and GPO / catalase / oxygen. This combination can be
coupled to an aldolase and an acceptor aldehyde. To make sure that only
dephosphorylation of the aldol adduct takes place, oxidation (and subsequent aldol
reaction) must be fast compared to the dephosphorylation rate of DHAP. This implies
slow phosphorylation as well. To avoid base catalysed background hydrolysis of
glycerol phosphate and DHAP, acid phosphatases were the obvious choice.
Oxidation and aldol reaction are optimally performed at neutral pH, but the enzymes
display activity under mildly acidic conditions as well.
55
Chapter 4
OH
HO OH
pyrophosphate
3-
Phosphatase PO4
3-
PO4 Phosphatase
2-
OPO3 OH OH
2- 2-
HO OH HO OPO3 HO OPO3
+ +
1/2
O2 H2O
GPO
Catalase
H2O2
O O
Phosphatase 2-
HO OH HO OPO3 (DHAP)
3-
PO4
O
FruA R H
OH O OH O
Phosphatase 2-
OH OPO3
OH PO4
3- OH
Figure 2. One pot preparation of carbohydrates from glycerol.
The acid phosphatases used were phytase69 from Aspergillus ficuum, which is a
cheap and readily available industrial enzyme and bovine intestinal mucosa
phosphatase (BIMP). The effect of the pH on the phosphorylation of glycerol
mediated by phytase and BIMP was examined (see Figure 3). With both enzymes
the productivity increased with the glycerol concentration. The synthesis of DHAP
was performed at up to 55% glycerol. The pH optimum of the system was 5.3 with
56
phytase and 6.8 with BIMP. More DHAP was produced when BIMP was used. The
pH optima of phytase and GPO are more distant than those of BIMP and GPO,
which may account for the lower production of DHAP by the system with phytase.
10
mM DHAP
0
4,8 5,8 6,8 7,8
pH
Figure 3. DHAP synthesis with pyrophophate and phosphatase and GPO/catalase
after overnight reaction. s phytase, l BIMP, ...... 0.5 M glycerol, ___ 55% glycerol.
At lower glycerol concentrations pyrophosphate as well as DHAP are hydrolysed but
it was not clear how much DHAP was hydrolysed to dihydroxyacetone. Thin layer
chromatography showed in each case large amounts of this hydrolysis product,
increasing with decreasing glycerol concentration. DHAP could not be detected with
the coupled enzymatic assay for DHAP by adding DHA to a solution containing
pyrophosphate, BIMP phosphatase, NADH, GDH/TPI, hence, phosphorylation of
DHA back to DHAP does not occur. Similarly, no phosphorylation could be detected
in 30% glycerol. Increasing the amount of glycerol is, thus, the only effective way to
suppress undesired hydrolysis. Unfortunately, the concentration of glycerol should
not exceed 55%, since the activity of GPO decreases dramaticly above this value
(Figure 4). Only swift consumption of DHAP by the aldolase can in principle prevent
57
Chapter 4
untimely dephosphorylation.
100
80
GPO Vrel (%) 60
40
20
0
0 20 40 60 80 100
glycerol (%)
Figure 4. Oxidation of L-glycerol-3-phosphate to DHAP by GPO in varying
concentrations of glycerol at pH 7.5. Conversion was measured after 1h. Reaction
rates were expressed relative to the highest rate obtained at 55% glycerol.
Preliminary to the integration of the aldolase in the reaction system, the rate of
DHAP hydrolysis by BIMP and aldol reaction were examined with equal initial
concentration of DHAP (10 mM). Both reactions depend on the glycerol
concentration (Figure 5). At 55% glycerol hydrolysis decreased to zero, while the
aldol reaction still took place (rates expressed in mM consumption per hour). Under
this condition the aldol reaction can be conducted without unwanted DHAP
hydrolysis. The activities are no more than guidelines, since they depend heavily on
the substrate concentration, which will be different in the one-pot reaction. With the
same amount of enzymes combined in the one-pot reaction starting from glycerol,
pyrophosphate and an aldehyde, predictably low concentrations of DHAP were
found, as a result of the direct consumption of in situ produced DHAP by the
aldolase.
58
aldol reaction
15 hydrolysis
one pot
mM DHAP / h
10
0
0 10 20 30 40 50 60
% glycerol
Figure 5. Aldol reaction and hydrolysis of DHAP (10 mM initially) with 1 mg bovine
intestinal mucosa phosphatase (BIMP) per ml. The “one pot” shows DHAP
concentration after overnight incubation.
Finally, the mixture must be dilluted to initiate the dephosphorylation of the
adduct. The dephosphorylation rate depends on the acceptor (Table I) used; in all
the cases the phosphatase displayed good activity.
Table I. Hydrolysis of the aldol adduct (initial concentration 10mM) by BIMP.
acceptor mM/h
acetaldehyde 3.13
propionaldehyde 2.00
butanal 2.29
Unfortunately, after combining all the components of the multi-enzyme reaction,
only dihydroxyacetone was found in the mixture as the final product. No conditions
were found where the aldol reaction prevailed over the dephosphorylation of the
aldol adduct. This outcome reveals, that in this dynamic system of four enzymes, the
59
Chapter 4
equilibrium is unfavorable for the dephosphorylated aldol adduct. Since the aldol
reaction is reversible, the phosphatase can perpetually hydrolyse DHAP released in
the retro aldol reaction. From experimental observation it is obvious that DHAP is a
better substrate than the aldol adduct. The main product from this multienzyme
reaction is therefore dihydroxyacetone.
Sequential one-pot reaction
The oxidation and aldol reactions are optimally performed at pH 7.5; hence, to avoid
interference by the phosphatase its hydrolytic activity should be zero at neutral pH.
This restricts the choice of phosphatase to the acid phosphatases, because the use
of an alkaline phosphatase would introduce the problem of base catalysed
background hydrolysis of glycerol phosphate and DHAP. The phosphatase of choice
was phytase from Aspergillus ficuum, which is a cheap and readily available
industrial enzyme used in animal feed. It has two pH optima for the hydrolysis of its
natural reactant, phytic acid: pH 2.2 and 5.
100 0% Figure 6.
L-G-3-P hydrolysis (%)
50%
80
Hydrolysis (2h) of L-glycerol-3-
60
phosphate (50mM) catalysed
40
by phytase (1 mg / ml) in 0 and
20
50% glycerol at different pH.
0
2 3 4 5 6 7 8
pH
60
The pH profile of the hydrolysis of L-glycerol-3-phosphate (see Figure 6) shows an
optimum at pH 3 - 4. At pH 7.5 phytase is stable, but unreactive towards either L-
glycerol-3-phosphate or DHAP, which makes it possible to combine the
phosphorylation, oxidation and aldol addition steps into a one-pot procedure by pH
control. Phytase exhibited the same hydrolytic activity towards DHAP as to L-
glycerol-3-phosphate.
Synthesis of L-glycerol-3-phosphate
The production of L-glycerol-3-phosphate by phytase-mediated reaction with
pyrophosphate was pH dependent with a very broad optimum (see Figure 7). At pH
2 phosphorylation was substantial and it decreased above pH 4 to become zero at
pH 7. The pH optimum shifted with increasing glycerol concentration from pH 3 to 4,
but as long as the pH was kept between these values good results were obtained.
No change in pH was observed during the reaction.
30 85% Figure 7. Production of L-

50%
glycerol-3-phosphate catalysd
L-G-3-P (mM)
10%
20
by phytase (1 mg / ml) in 10, 50
10 and 85% glycerol at different
pH after 2.5 h incubation with

0
150 mM pyrophosphate.
2 3 4 5 6 7 8
pH
61
Chapter 4
Since only L-glycerol-3-phosphate can be oxidized to DHAP, the D-isomer (which is
presumably produced in equal amounts) will remain unused in solution. In the final
dephosphorylation step it is converted back to glycerol and phosphate. This also
implies that the yield of the reaction, based on pyrophosphate, cannot surpass 50%.
Effects of glycerol concentration
The time to reach maximum conversion, based on the amount of
pyrophosphate, was hardly affected by glycerol concentrations between 10% and
85% (v/v), but at 95% the rate had dropped by a factor of ten (data not shown). At
still lower water concentrations the enzyme activity decreased to zero in pure
glycerol, although some activity still could be observed when only 0.25% water was
present.
100
4.5h
phosphorylation (%)
80 24h
60
40
20
0
0 20 40 60 80 100
glycerol (%)
Figure 8. Phosphorylation of glycerol with pyrophosphate (150 mM) by phytase (1
mg / ml) at pH 4.0 in glycerol / water mixtures. Racemic glycerol-3-phosphate is
obtained in 100% yield in 95% glycerol after 24h.
62
The yield of D,L-glycerol-3-phosphate increased with increasing glycerol
concentration (see Figure 8), presumably because the competing hydrolysis of
pyrophosphate and glycerol-3-phosphate is suppressed at low water concentrations.
At 95% glycerol concentration a quantitative conversion of pyrophosphate into
glycerol-3-phosphate was obtained after 24h reaction time.
Selectivity
Phytase is believed to be non-stereoselective in the phosphorylation of
glycerol.67 Since only L-glycerol-3-phosphate is a substrate for GPO, the production
of D-glycerol-3-phosphate cannot be monitored by this enzymatic assay. The
consumption of pyrophosphate and production of phosphate and glycerol phosphate

31
could be monitored using P NMR, however. A reaction mixture with 95% glycerol
31
and an initial pyrophosphate concentration of 150 mM was analyzed for P before
100% conversion was reached. This revealed that the phosphate concentration was
100 mM whereas the pyrophosphate concentration was 50 mM, a decrease of 100
mM (67% conversion). This observation demonstrates that, under these conditions,
pyrophosphate was converted to one equivalent of glycerol phosphate and one
equivalent of phosphate, without any hydrolysis of pyrophosphate or glycerol-3-
phosphate. The concentration of D,L-glycerol-3-phosphate therefore was 100 mM.
Enzymatic detection of L-glycerol-3-phosphate with GPO and GDH gave a
concentration of 50 mM, exactly matching the NMR results and confirming that
phytase is not stereoselective. But, since glycerol-2-phosphate was absent in the
reaction mixture phytase is completely regiospecific.
63
Chapter 4
Synthesis of DHAP from L-glycerol-3-phosphate
After complete conversion of the pyrophosphate, the pH was raised to 7.5 to
suppress the hydrolysis of L-glycerol-3-phosphate (and DHAP) and maximize GPO’s
activity. The concentration of glycerol must also be adjusted since GPO activity is
low at 95 %. To achieve maximum space-time yields a balance must be found
between high L-glycerol-3-phosphate concentration and compatibility with GPO
activity. At pH 7.5 the oxidation rate of L-glycerol-phosphate was increased fiftyfold
by lowering glycerol concentration (Figure 4) from 95% to 55%, which convincingly
shows that dilution is highly advantageous.
A 1.5 mmol (pyrophosphate) reaction was carried out in 95 % glycerol. The
time to reach full conversion at this concentration was one day. After phosphorylation
was completed the mixture was diluted with water to 55% glycerol and the pH was
adjusted to 7.5. Addition of GPO/catalase and entrainment with oxygen for three
hours afforded DHAP in a concentration of 43 mM corresponding to 50% yield based
on pyrophosphate.
Aldol reaction and dephosphorylation
Aldol reactions can be performed with the in situ prepared DHAP in glycerol. To
this end we used the D-fructose-1,6-bisphosphate aldolase (FruA) from
Staphylococcus carnosus to prepare the aldol adduct of DHAP and butanal.70
Butanal and aldolase were added to the mixture containing DHAP, phytase,
glycerolphosphate oxidase, catalase and 55% glycerol. In fact, it is not necessary to
64
do this step separately; the oxidation and aldol reaction can be carried out
simultaneously64, thus saving hours of reaction time. The consumption of DHAP was
used to monitor the course of the reaction. Under these conditions the aldolase
displayed normal activity and modification of the mixture to enhance its performance
was therefore unnecessary. We expect that the combination of the relaxed acceptor
specificity of DHAP dependent aldolases with this new method opens the way to the
synthesis of a wide spectrum of possible structures.
The addition of extra phosphatase for removal of the phosphate group is not
necessary since phytase is still present and active. After oxidation and aldol reaction,
phytase-mediated dephosphorylation of the butanal-DHAP adduct was carried out by
simply lowering the pH to 4.0. Apparently phytase exhibits a wide substrate
specificity in the hydrolysis reaction and, hence, is expected to be generally
applicable for a broad range of substrates in this final step. Extraction afforded 5-
deoxy-5-ethyl-D-xylulose in 57% yield from from DHAP (or 29% from
pyrophosphate).
Chemical phosphorylation of glycerol
Alternatively, glycerolphosphate can be made chemically from glycerol and
phosphoric acid by simply heating glycerol with phosphoric acid. Under these
conditions 1.2 M of D,L-glycerolphosphate can be obtained. Problematic however is
the high salt concentration, which requires substantial dilution. This lowers the final
DHAP concentration to about 20 mM.
65
Chapter 4
Conclusion
Phytase-catalyzed phosphorylation of glycerol with pyrophosphate afforded D,L-
glycerol-3-phosphate in quantitative yield. L-glycerol-3-phosphate underwent GPO-
catalyzed oxidation to DHAP after adjustment of the pH to 7.5. The D-isomer was not
converted but its presence had no effect on the subsequent steps. Under the
conditions used for oxidation and aldol reaction (pH 7.5) phytase did not hydrolyze L-
glycerol-3-phosphate or DHAP. Manipulation of pH thus gives total control over
phytase’s activity; it can be “switched off” by increasing pH from 4 to 7.5 and
“switched on” again by lowering back to pH 4.0.
This alliance of four different enzymes in a one-pot cascade of four enzymatic
transformations provides an attractive procedure for performing aldol reactions with
DHAP aldolases starting from the cheap, readily available glycerol and
pyrophosphate. Combined with the broad substrate specificity of DHAP aldolases
towards acceptor substrates it may constitute a simple procedure for the synthesis of
a wide variety of carbohydrates from glycerol.
Experimental
General
31
P spectra were recorded at 300 Mhz on an Oxford NMR 300 with phosphoric
acid as external standard. UV spectroscopy was performed with a Varian Cary 3 Bio
equipped with a Cary temperature controller. The aldolase and oxidase were
66
obtained from Roche Diagnostics. All other enzymes and chemicals were purchased
from Sigma.
Assay for DHAP and L-glycerol-3-phosphate
DHAP was assayed with a coupled enzyme system: reduction with NADH-
consuming glycerol-3-phosphate dehydrogenase enables determination of DHAP by
measuring NADH concentration with UV-spectroscopy. From a diluted DHAP
solution, 50 µl was added in a quartz cuvette containing 1.95 ml 50 mM TRIS pH 7.6,
0.16 mM NADH, 1.25 U glyceraldehyde-3-phosphate and 12.5 U triose-1-phosphate
isomerase. The absorption was monitored at 340 nm at 20 °C. Blank DHAP: 0.0047
mM / min. The molar adsorption coefficient taken was 6.22 l.mmol-1cm-1.
L-glycerol-3-phosphate was detected by oxidizing it to DHAP with GPO and
subsequent assay for DHAP. Samples of 125 µl containing L-glycerol-3-phosphate
were diluted four-fold with 200 mM Tris buffer, pH 8.0. This buffer neutralizes acidic
samples thus preventing hydrolysis by phytase. In a 8 ml vial with gastight Teflon
cap 5 µl GPO/catalase mixture (25 units / 250 units) was added and oxygen was
applied for 30 seconds while the vial was shaken vigorously. After one-hour
incubation at room temperature the mixture was assayed for DHAP.
DHAP synthesis from glycerol
To a 10 ml solution containing 95% (v/v) glycerol and 150 mM pyrophosphate pH 4.0
67
Chapter 4
(adjusted with 2 N HCl) freeze-dried phytase was added (10 mg ). The mixture was
incubated for 24h at 37 °C in a 30 ml flask with Teflon cap and shaken gently. After
cooling down to room temperature the pH was raised to 7.5 by addition of 2.0 N
sodium hydroxide. Water was added (7.4 ml) to obtain a final glycerol concentration
of 55%. Then 100 µl GPO/catalase mixture (50 units / 500 units) was added and
oxygen was applied for three minutes. This was repeated after 30 and 60 minutes.
The flask was shaken at room temperature. Oxidation was stopped after 3h. The
yield of DHAP (from pyrophosphate) was 50% (0.75 mmol). Since phytase produces
equal amounts of the enantiomers the yield of D,L-glycerol-3-phosphate is 100%.
Aldol reactions
To a 10 ml mixture containing 43 mM DHAP 88 µl butanal (100mM) and 12.5 units of
fructose-1,6-bisphosphate aldolase from Staphylococcus carnosus were added.
Conversion was 78% after four hours. Dephosphorylation by phytase still present in
the mixture started by lowering the pH to 4 with 2N HCl and stirring overnight. The
glycerol/water mixture was extracted with ethylacetate, the ethylacetate layer with
water and dried with sodium sulfate. After evaporation of the solvent 39 mg 5-deoxy-
1
5-ethyl-D-xylulose was isolated (analysed using H NMR spectra of authentic
samples70). The yield was 56% from L-glycerol-3-phosphate.
68
Chemical phosphorylation of glycerol
Synthesis of L-glycerol-3-phosphate from glycerol and phosphoric acid (85%)
was achieved by heating an equimolar mixture (17.8 ml) overnight at 110 °C and
subsequently adding 1 equivalent water followed by stirring for 5h at 80 °C. This
resulted in an final concentration of 0.6 M L-glycerol-3-phosphate (8.4% yield). When
equal amounts of α- and ß-glycerolphosphate are formed, phosphate concentration
is about 4.7 M.
References
50
H.J.M. Gijsen, L. Qiao, W. Fitz and C.-H. Wong, Chem. Rev. 1996, 91 (1), 443-
469
51
W.-D. Fessner and C. Walter, Angew. Chem. Int. Ed. Engl. 1992, 31 (5), 614-616
52
P.G. Wang, W. Fitz and C.-H. Wong, Chemtech 1995, april, 22-23
53
O. Meyerhof and K. Lohmann, Biochem. Z. 1934, 271, 89
54
C.-H. Wong and C.M. Whitesides, J. Org. Chem. 1983, 48, 3493-3497
55
C.E. Ballou and H.O.L. Fischer, J. Am. Chem. Soc. 1956, 78, 1659
56
H.O.L. Fischer and H. Mildbrand, Ber. Dtsch. Chem. Ges. 1924, 57, 710
57
R.L. Colbran, J.K.N. Jones, N.K. Matheson and I. Rozema, Carbohyd. Res. 1967,
4, 355-358
58
F. Effenberger and A. Straub, Tet. Lett, 198728 (15), 1641-1644
59
R.L. Pederson, J. Esker and C.-H. Wong, Tetrahedron 1991, 47 (14/15), 2643-
2648
60
S.-H Jung, J.-H. Jeong, P. Miller and C.-H. Wong, J. Org. Chem. 1994, 59, 7182-
7184
61
M.-L Valentin and J. Bolte, Bull. Soc. Chim. Fr. 1995, 132, 1167-1171
62
T. Gefflaut, M. Lemaire, M.-L Valentin and J. Bolte, J. Org. Chem. 1997, 62, 5920-
69
Chapter 4
5922
63
D.C. Crans and G.M. Whitesides, J. Am. Chem. Soc. 1985, 107, 7019-7027
64
W.-D. Fessner and G. Sinerius, Angew. Chem. Int. Ed. Engl. 1994, 33 (2), 209-
212
65
P. D’Arrigo, V. Piergianni, G. Pedrocchi-Fantoni and S. Servi, J. Chem. Soc.
Chem. Commun. 1995, 2505-2506
66
a) R.L. Colbran, J.K.N. Jones, N.K. Matheson and I. Rozema, Carbohyd. Res.
1967, 4, 355-358, b) F. Effenberger and A. Straub, Tet. Lett. 1987, 28 (15), 1641-
1644, c) R.L. Pederson, J. Esker and C.-H. Wong, Tetrahedron 1991, 47 (14/15),
2643-2648, d) S.-H Jung, J.-H. Jeong, P. Miller and C.-H. Wong, J. Org. Chem.,
1994, 59, 7182-7184
67
A. Pradines, A. Klaébé, J. Périé, F. Paul and P. Monsan, Tetrahedron 1988, 44
(20), 6386-6386
68
A. Pradines, A. Klaébé, J. Périé, F. Paul and P. Monsan, Enzyme Microb. Technol.
1991, 13, 19-23
69
J. Dvorakova, Folia Microbiol. 1998, 43 (4), 323-338
70
R. Schoevaart, F. van Rantwijk and R.A. Sheldon, Tetrahedron: Asymmetry 1999,
10 (4), 705-711
70
71
72
5
DHAP analogues
Abstract
Hydroxyactone thiosulfate was prepared as a substrate analogue for DHAP, but it
was not converted by DHAP utilizing enzymes. The scope of aldol reactions with in
situ formed dihydroxyacetone arsenate with different aldehydes catalysed by
fructose-1,6-bisphosphate aldolase was examined. The use of inorganic arsenate
and dihydroxyacetone afforded higher yields than dihydroxyacetone phosphate and
precluded the normal, elaborate preparation of phosphorylated dihydroxyacetone.
Cosolvents which were applied to increase the solubility of hydrophobic aldehydes
also increased their reaction rates and stabilised the enzyme as well. The
stereospecifity is somewhat higher with the natural donor substrate.
73
Chapter 5
Hydroxyacetone thiosulfate as DHAP analogue
The possible replacement of dihydroxyacetone phosphate (DHAP) as donor in aldol
reactions catalysed by DHAP depending aldolases is of considerable synthetic
significance. The first DHAP analogue (1, Figure 1) reported71 was acetol phosphate,
with a relative activity of only 1%. 3-Chloro-, 3-bromo- and 3-iodoacetol phosphate72
(2) were synthesised from 3-chloro-1,2-propanediol, to investigate their capability to
bind irreversibly in the active-site of aldolases, triose phosphate isomerases and
glycerolphosphate dehydrogenase. The iodo analogue of DHAP inactivated FruA
rapidly, while the others had no effect. It is unclear if the chloro and bromo
compounds were substrates, however, it is known that 3-fluoroacetol phosphate73 is
not. They did act as effective reversible inhibitors, which demonstrates their ability to
bind in the active site.
The first active substrate analogue was 4-hydroxy-3-oxobutylphosphonic
acid74,75 (3), which was synthesized in six steps from acrylic acid. Its reaction with
FruA was 10% of that of DHAP. 1,3-Dihydroxy-2-butanone 3-phosphate76 (4),
prepared in seven steps from butyl-2-hydroxyacetate, had the same activity but the
very similar phosphonomethyl glycolate (5) was not active at all. The phosphoramide
(6) and phosphorothioate (7) analogues of DHAP prepared from bromoacetyl
chloride or 2-chloro-2-propen-1-ol77 both showed a relative activity of 10%. These
two compounds hydrolyse readily under the reaction conditions. For example, 6 was
hydrolysed fifty times as fast as DHAP. Consequently, also considering their reduced
activity, the synthetic utility of 6 and 7 is quite limited, even though they are relatively
stable at pH 9. The synthesis of 3, 6 and 7 could also be achieved by the GPO
mediated oxidation of the corresponding glycerol-3-phosphate analogues64.
74
DHAP analogues
O O O
-2 -2 -2
OPO3 4 HO OPO3 7 HO SPO3
1
O
O O
-2 5 HO -2 -
2 X OPO3 O PO3 8 HO OSO3
X = Cl, Br, I
O
O
3 HO -2 -2
PO3 6 HO NHPO3
Figure 1. DHAP analogues
The crucial role of the phosphate group was demonstrated by dihydroxyacetone
monosulfate76 (8), which was completely inactive. This shows the narrow limits
between which the aldolase tolerates changes in the donor substrate, in contrast
with the quite relaxed acceptor specificity.
Summarising, no effective replacement for DHAP has yet been found. In view of the
lack of a theoretical framework for selecting promising reactant candidates, such
replacements can only be identified by experiment. We decided to explore the
capabilities of the thio analogue of 8, dihydroxyacetone thiosulfate (HATS), which we
synthesised via the haloacetol phosphate72 route (Figure 2). In the final step the
chlorine moiety of hydroxyacetone chloride is exchanged for thiosulfate. Afterwards a
shorter route was devised, namely from dihydroxyacetone which had a slightly
higher overall recovery, although only two in stead of six steps were employed.
75
Chapter 5
OH OH Cl
5 steps OH O Cl
8% overall
Na2SSO3
OH O SSO3Na
PhCOCl O Na2SSO3
OH O OH (HATS)
Ph C O O OH
12% overall
Figure 2. Synthesis of the DHAP analogue hydroxyaceton thiosulfate (HATS)
In an aldolase catalysed reaction with propanal no product could be discovered. The
reduction of HATS with NADH mediated by glycerophosphate dehydrogenase did
not proceed either, although this test is very sensitive. The use of triosephosphate
isomerase in combination with hydrazine did not give an indication that any aldehyde
was formed. Finally, when crude 2,3-dihydroxypropane thiosulfate prepared from
thiosulfate and 3-chloro-1,2-propanediol was exposed to GPO and oxygen no HATS
was found. Sofar, it is has become clear that HATS is not a substrate for any DHAP
utilizing enzyme, which reconfirms the importance of the phosphate moiety.
Dihydroxyacetone arsenate as DHAP analogue
In many phosphate-depending enzymatic reactions arsenate esters can mimic
phosphate esters78. This approach obviates the use of organic phosphates that
require kinases and ATP, as well as ATP regeneration systems for their synthesis.
Because arsenate esters are formed spontaneously and reversible in aqueous
76
DHAP analogues
solutions - in contrast with phosphate esters - experimental procedures are
simplified. The synthetic application of the in situ reversibly formed dihydroxyacetone
arsenate (DHAAs) has been demonstrated in the preparation of uncommon sugars79.
In that case DHAAs replaced dihydroxyacetone phosphate (DHAP) in reactions
catalysed by DHAP-dependent aldolases (see figure 3). This group consists of four
enzymes which all have a unique specificity for one of the four possible
stereoisomers. They are named according to specificity D-fructose-1,6-bisphosphate
(FruA), L-fuculose-1-phosphate (FcuA), L-rhamnulose-1-phosphate (RhuA) and D-
tagatose-1,6-bisphosphate (TagA) aldolase.80
O OH O
2- aldolase 2-
HO OAsO3 OAsO3
RCHO R
OH
HOAsO32-
O OH O
HO OH OH
R
OH
Figure 3. Arsenate catalysed aldol reaction with DHAP-aldolases
With FruA from Staphylococcus carnosus as aldol reaction catalyst we explored the
use of many unnatural acceptor substrates, cosolvents and the effects of arsenate
on the stereoselectivity of the enzyme.
77
Chapter 5
Results and discussion
Monitoring the progress of the aldol reaction requires the detection of either
dihydroxyacetone (DHA) or the aldehyde. Enzymatic detection of DHA would require
its phosphorylation to DHAP with ATP, mediated by a kinase81, preliminary to
measuring DHAP by monitoring its reduction with NADH mediated by
glycerolphosphate dehydrogenase (GDH). The phosphorylation step would delay the
analysis by hours and is not always complete. The NADH / GDH system can also be
used to detect DHAAs, however, since its Km is hundred fold higher78 than the one of
DHAP, DHAAs reacts very slowly rendering the system unsuitable for routine
analysis. To circumvent this problem DHA can be reduced with borohydride to
glycerol, which in its turn is detected by phosphorylation with ATP coupled to an ADP
detection system. As this approach requires four steps, we opted for a more
straightforward approach by means of detecting the aldehyde through reduction with
NADH mediated by yeast alcohol dehydrogenase (ADH). This method is fast, but
cannot detect all aldehydes, since the substrate spectrum of alcohol dehydrogenase
is more limited than the one of FruA. For example formaldehyde, i-butyraldehyde,
pivaldehyde, phenylacetaldehyde, methylglyoxal, glyoxal and glyceraldehyde were
found to be substrates for FruA but not for ADH. These aldehydes can alternatively
be detected by reaction with hydrazine, which is monitored at 240 nm, although this
would require all the individual molar extinction coefficients to be the determined
first.
In order to study the scope of DHAAs as a substrate analogue, propanal was
chosen as an example acceptor substrate, since it dissolves readily in water and is a
good substrate for both FruA and ADH.
78
DHAP analogues
Optimal conditions
Two factors which are considered to have a major influence on the enzymatic aldol
reaction with DHAAs were first examined: the pH and the arsenate concentration.
When the pH was varied between pH 6 and 9 the enzymatic aldol reaction took
place, with an optimum at pH 7.6. At values above pH 9 or below pH 6 no aldol
reaction was observed. The base catalysed reaction – at a pH above neutral – was
never detected.
Although arsenate is catalytically79 involved in the reaction, a large excess is
needed to compensate for the reduced binding of the substrate by the enzyme: Km is
fivefold higher and Vmax eightfold lower78 for DHAAs compared to DHAP when FruA
is used as a catalyst. With an arsenate concentration being as low as a few mol
percent, aldol reactions could still be observed. However, with 20 mM DHA present,
the reaction rate reached a maximum at about 120 mM arsenate, a sixfold excess
(see figure 4). Practically speaking this means that arsenate can be used to buffer
the reaction medium. Higher arsenate concentrations gradually slowed down the
aldolase, presumably by the high salt concentration. The reduced activity could, in
principle, be compensated by increasing the amount of enzyme. For example, the
use of DHA and arsenate in concentrations of up to 0.5 M has been reported79, but
this required a fifty fold increase of the amount of aldolase. Because FruA is
inactivated by propanal at concentrations above 200 mM, which effect increases with
less hydrophilic aldehydes, the aldehyde concentration should be kept as low as
practicable. As a compromise, we performed the reaction at 20 mM DHA, 120 mM
arsenate and 50 mM aldehyde.
79
Chapter 5
100
80
% conversion
60
40
20
0
0 50 100 150 200 250
mM arsenate
Figure 4. Optimal arsenate concentration with 20 mM dihydroxyacetone, 50 mM
propanal and 1 unit fructose-1,6-bisphosphate aldolase, conversion after 24h.
Use of cosolvents
The more hydrophobic aldehydes that we used as acceptor, were sparingly soluble
in water. Because the resulting inhomogeneous reaction system provoked
deactivation of the biocatalyst and hampered representative sampling, the use of
organic cosolvents became an attractive proposition.
The effect of a number of cosolvents on the activity of FruA from S. carnosus
was assessed by monitoring the aldol reaction of DHAAs and propanal (Table I). The
addition of 25 % cosolvent in some cases, such as DMSO, DMF and tert. butyl
alcohol, slightly activated the enzyme compared with reaction in aqueous medium. In
other cases a reduction in activity of up to 50% was observed. The lowered activity in
ethanol is most likely caused by ester formation of arsenate and the cosolvent, thus
lowering the effective arsenate concentration. In the case of tert. butyl alcohol the
80
DHAP analogues
effect of unwanted ester formation presumably is reduced by steric hindrance.
Addition of cosolvent lead in many cases to increased enzyme activity, most likely
provoked by withdrawal of water from the protein causing structural changes82. The
effect of 25 % cosolvent (v/v) on the 24-h stability of FruA from S. carnosus as well
as those from Staphylococcus aureus and rabbit muscle (RAMA) was measured via
a standard FruA assay (Table I). It is apparent that the bacterial aldolases from S.
carnosus and S. aureus are superior in stability compared with the traditionally used
fructose-1,6-bisphosphate aldolase from rabbit muscle (RAMA). Some general
trends could be observed: in all cases dimethylformamide (DMF) increased stability
whereas acetonitrile and tert. butyl alcohol had the opposite effect. FruA from S.
carnosus is remarkably stable in DMSO. When it was added to pure DMSO, shaken
for 1 hour and centrifuged, the pellet contained 85% of the original activity while the
remaining 15% was found in the supernatant.
Table I. Activity and stability of aldolases from different sources in cosolvents.
Vrel1 stability2
cosolvent (25%) S. carnosus S. carnosus S. aureus RAMA
DMSO 104 151 146 66

1,2-Dimethoxyethane 107 134 125 65
Ethanol 50 116 99 92
Dioxane 87 114 123 60
DMF 118 112 139 106
Acetonitrile 92 98 91 0
t-BuOH 121 81 86 18
Water 100 100 100 96
1
Based on conversion of propanal after 2 hours relative to the activity in water.
81
Chapter 5
2
Cleavage of fructose-1,6-bisphosphate after 1 day incubation at room temperature
relative to the rate in water at t = 0.
Because FruA is activated as well as stabilized in the presence of 25% 1,2-
dimethoxyethane or DMF, the time to reach maximum conversion is reduced,
making these solvents an attractive additive, even with accptor aldehydes that
readily dissolve in water. The preferred cosolvent will depend on the best stability /
activity combination as well as on other restrains, such as work-up procedures. For
example, activity in DMSO is only slightly increased, but stability is excellent. With
tert. butyl alcohol the activity is much higher, but the stability is lowered. Ethanol has
only half the activity, but it has a good stability and is of course attractive because it
is easy to separate from the reaction mixture.
60
6h
Conversion (%)
40
20
0
0 10 20 30 40 50 60 70 80
% DMSO (v/v)
Figure 5. Use of DMSO as a cosolvent for arsenate catalysed aldol reaction of
DHA with propanal.
The effect of DMSO was studied in more detail. Propanal was, for practical
82
DHAP analogues
reasons, used as the acceptor. Strictly spoken this aldehyde would not require a
cosolvent, but its use allowed us to study the effect of the cosolvent without
interference by heterogeneity issues. The conversion of propanal and DHAAs in the
presence of FruA from S. carnosus slightly increased up to 50% (v/v) of DMSO and
dropped sharply above 60%.
Acceptor specificity with DHAAs
Aldol reactions of DHAAs with a number of aldehydes were performed. Since our
method of analysis restricted the use of aldehydes, the conversion of aldehydes that
could not be beasured by the ADH assay: formaldehyde, i-butyraldehyde,
pivaldehyde, phenylacetaldehyde, glyoxal and methylglyoxal were monitored by
assaying DHAP assay after (partial) phosphorylation of the unconverted DHA. The
hydrophobic acceptors hexanal, heptanal, i-butyraldehyde, pivaldehyde, 3-
phenylpropionaldehyde required the addition of 25% (v/v) DMSO as cosolvent for
reliable sampling.
All the aldehydes employed were found to be acceptor substrates (Table II).
The enzyme specificity found when DHAAs was used as donor is in agreement with
expectations based on aldol reactions with DHAP83,84, in which aldehydes with an α-
hydroxyl group as well as activated aldehydes are normally good substrates.
However, DHAAs gave higher products yield with hydrophobic aldehydes than with
hydrophilic ones, which runs against the general trend observed with DHAP. This
manifestation is a great advantage of the usage of arsenate, although it entails a
penalty as regards reaction rate, leading to an increase in reaction time in all cases.
83
Chapter 5
Under optimized conditions and with the same amount of enzyme, non-polar
aldehydes reacted approximately six times slower, and apolar ones twenty times
slower, with DHAAs than with DHAP.
Table II. Conversion of DHAAs and DHAP by FruA
RCHO Donor
R DHAAs DHAP
H 67 (48h) a 81 (2h)
CH3 78 (48h) 84 (2.5h)
CH2CH3 79 (24h) 67 (4h)
(CH2)2CH3 86 (24h) 78 (4h)
(CH2)3CH3 95 (24h) 64 (3h)
(CH2)4CH3b 76 (24h) 62 (5h)
(CH2)5CH3b 54 (24h) 30 (16h)b
CH(CH3)2 b 83 (48h) a 78 (16h)b
C(CH3)3 b 69 (48h) a 85 (16h)b
(CH2)2Phb 83 (24h) 65 (16h) b
CH2Ph 77 (48h) a 66 (2h)
CHO 83 (48h) a 98 (5h)
COCH3 73 (48h) a 86 (6h)
CH2Cl 63 (24h) 76 (1.5h)
CH2OH 83 (24h) 96 (0.5h)
a) Detected by phosphorylation of DHA (error ±10% ).
b) 25% DMSO was added
84
DHAP analogues
Effect of arsenate on the stereoselectivity
To analyse the stereochemistry of the products, phosphorylation is necessary to
perform a stereoisomer assay85. This assay is based on DHAP detection in the retro
aldol reaction for each of the four possible stereoisomers, catalysed by the
appropriate DHAP dependent aldolase. The concentration of each single
stereoisomer can be detected in this way, giving full insight in the stereoselectivity of
the aldol reaction. The phosphorylation of the aldol adducts, preliminary to the retro-
aldol assay, was performed using a glycerol kinase and ATP. It could not be carried
out to complete conversion, since aldol adducts are even worse substrates for
glycerol kinases than is DHA. But, since glycerol kinase did not seem to discriminate
between the different stereoisomers – which was demonstrated by using RhuA
instead of FruA as the aldol catalyst - this method is nevertheless suited for
determining the ratios of the stereoisomers.
Table III. Stereochemistry of aldol adducts.
RCHO DHAP DHAAs

R Fru Rhu Fcu “Tag” Fru Rhu Fcu “Tag”
CH3 88 0 3.5 8.5 90a 10 0 a
CH2CH3 96 0 0 4.0 93a 5 2 a
CH2CH2CH3 90 0 2.5 7.5 95a 0 5 a
CH(CH3)2 92.5 0 0 7.5 86a 0 14 a
a
Since phosphorylation was incomplete, the proportion of the product with the
tagatose configuration could not be calculated
85
Chapter 5
When comparing the steric course of the FruA mediated aldol reactions of DHAAs
and DHAP, the high stereospecificity of the aldolase is evident in both cases (Table
III). The fraction of products with D-fructose (“Fru”) and D-tagatose (“Tag”)
configuration is somewhat lower for DHAAs, however. Moreover, the products with
the L-rhamnulose (“Rhu”) configuration, which were never detected in aldol adducts
of DHAP84, were found with acetaldehyde and to a lower extend also with propanal
as the acceptor substrate. Hence, the use of arsenate instead of phosphate clearly
has some effect on the stereoselectivity of the aldol reaction.
Conclusion
In situ formed dihydroxyacetone arsenate (DHAAs) can advantageously replace the
traditionally used DHAP in D-fructose-1,6-bisphosphate aldolase catalysed aldol
reactions. The procedure obviates the elaborate separate preparation of the donor
and gives improved yields with hydrophobic acceptor aldehydes. The use of DHAAs
somewhat reduces the stereoselectivity of the aldol reaction. Cosolvents can be
used with a triple advantage: increased solubility of hydrophobic aldehydes, faster
reaction and stabilisation of the enzyme.
Experimental
General
86
DHAP analogues
temperature controller. Bacterial aldolases were obtained from Roche Diagnostics as
a gift. DHAP was prepared from its ethyl hemiacetal dimer barium salt (Fluka). All
other enzymes and chemicals were purchased from Sigma. Thin-layer
chromatography (TLC) on pre-coated aluminium foil of silica gel (Merck) was
performed with butanol / acetic acid / water (3:2:5). Detection was effected by
dipping in a solution containing 24 g ammonium molybdate, 1 g of cerium(III) sulfate
and 10% (v/v) H2SO4 per 500 ml and subsequent heating.
Preparation of hydroxyacetone thiosulfate (HATS)
Method 1. 3-Chloro-1-hydroxyacetone was synthesised according to Hartman72. A
stirred solution of this product (110 mg, 1.01 mmol) in water (1 ml) and ethanol (1.5
ml), was treated with sodium thiosulfate pentahydrate (320 mg, 2eq), and then
allowed to stir for 3 hours. The resultant mixture was concentrated in vacuo and the
residual material was solved in dry ethanol, filtrated and again concentrated in
vacuo giving the titel compound as a white powder (25 mg, 12%).
Method 2. In a 250 ml flask 2 g dihydroxyacetone (22.2 mmol) was solved in 80 ml
of dry pyridine containing 100 mg 4-dimethylaminopyridine (0.8 mmol). At –5 °C 2.31
ml (1 eq) of benzoylchloride was added dropwise. After the addition was complete,
the mixture was stirred for one-half hour at room temperature. Subsequently, sodium
thiosulfate pentahydrate (11 g, 2eq) was added and the mixture was stirred for two
hours. The pyridine was evaporated (with toluene), water was added and the
solution was filtered. Then the water was evaporated (with ethanol) and 660 mg
87
Chapter 5
(14% yield) of solid material was isolated. TLC indicated that both methods yielded
the same product, which was used without additional purification in enzymatic
assays.
Preparation of 2,3-dihydroxypropane thiosulfate
A stirred solution of 3-chloro-1,2-propanediol (1g, 9 mmol) in water (8ml) was treated
with sodium thiosulfate pentahydrate (1g, 0.75 eq) and the mixture was stirred
overnight at room temperature and then heated at 85 °C for 5 hours. Ethanol was
added and the crude product was concentrated in vacuo. This yielded a very viscous
residue (1g, 53 %) which was used without further purification in enzymatic assays.
Assay of aldehydes with alcohol dehydrogenase
The aldehydes were assayed with alcohol dehydrogenase (ADH) from yeast. To
1.97 ml of a 50 mM Tris-buffer pH 7.6 containing 0.16 mM NADH in a 2 ml cuvette,
10 µl solution containing 5 units ADH was added. Then a 20 µl sample of tenfold
diluted reaction mixture was added and the absorption was monitored at 25 °C. The
molar absorption coefficient taken was 6.22 l.mmol-1cm-1.
88
DHAP analogues
Assay of DHA
After maximum conversion the reaction mixture was centrifuged in an eppendorf
centrifuge and transferred into a Microcon Centrifugal filter (cutoff 10 kD) and again
centrifuged for 10 minutes. The filtrate contained no aldolase activity. To 1 ml of this
aldolase free mixture containing a maximum amount of 4 mM dihydroxyacetone, 5.5
mg ATP (10 mM) and 10 units glycerol kinase from Pseudomonas species were
added. After shaking for 6 hours at pH 7.5 the mixture was assayed for DHAP.
DHAP was assayed with a coupled enzyme system. To 1.95 ml 50 mM Tris-
buffer pH 7.6 containing 0.16 mM NADH in a 2 ml cuvette, 10 µl of a mixture
containing 1.25 unit D-glycerol-3-phosphate dehydrogenase and 12.5 unit triose-1-
phosphate isomerase were added. Then 40 µl of the phosphorylation mixture was
added and the absorption monitored at 25 °C.
Aldol reaction with dihydroxyacetone arsenate
In a gastight bottle dihydroxyacetone (1,8 mg, 20 mM) and aldehyde (50 mM) were
added to 1 ml of 120mM arsenate buffer pH 7.6. Then 0.5 ml was taken out to
monitor the background reaction. To the other half, 1 unit of fructose-1,6-
bisphosphate aldolase from Staphylococcus carnosus was added. After shaking for
24 hours the mixtures were analyzed.
89
Chapter 5
FruA activity assay
Two aldol reactions - one with 25 % cosolvent (v/v) - were performed as described
above. After incubating at room temperature for two hours, the amount of propanal
converted in both samples was assayed with ADH. The conversion in water was set
at 100%.
FruA stability assay
To 1.95 ml of a 50 mM Tris-buffer pH 7.6 containing 0.16 mM NADH in a 2 ml
cuvette, were added 20 µl of a 190 mM fructose-1,6-bisphosphate as well as 20 µl of
a mixture containing 1.25 unit D-glycerol-3-phosphate dehydrogenase and 12.5 unit
triose-1-phosphate isomerase. Then 50 µl diluted incubation mixture containing the
aldolase was added. The reduction of the reaction products DHAP and GAP by
NADH was monitored at 25 °C. 1 unit aldolase converts 1 µmol fructose-1,6-
bisphosphate per minute.
Stereoisomer assay
The reaction mixture containing the aldol adducts and DHA was deproteinated and
partially phosphorylated as described above. 20 to 40 µl of this reaction mixture was
then added to a quartz cuvette containing 1.97 ml of a 50 mM TRIS buffer pH 7.6,
0.16 mM NADH, 1.25 unit GDH and 12.5 units triose-1-phosphate isomerase. The
90
DHAP analogues
absorption was monitored at 340 nm at 25 °C. After complete conversion of residual
DHAP, the retro-aldol reaction was initiated by adding 10 µl aldolase (5 units). The
amount of DHAP released from this reaction is equal to the amount of adduct with
the configuration corresponding with the used aldolase. For each of the three
aldolases a new sample was taken from the deproteinated solution.
Preparation of 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose
In a flask dihydroxyacetone (36 mg, 20 mM) and 88 µl butanal (50 mM) were added
to 20 ml of 120mM arsenate buffer pH 7.6 containing 25% DMF. To initiate the
reaction, 40 units of fructose-1,6-bisphosphate aldolase from Staphylococcus
carnosus was added. After stirring for 24 hours conversion was 86% (assay) the
mixture was extracted with diethylether. This afforded after drying and evaporation of
the solvent, 100 mg of product and unreacted aldehyde. Acetylation in 2 ml pyridine
and 1 ml acetic anhydride yielded 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose (65%
yield) (analysed using 1H NMR spectra of authentic samples).
References
71
I.A. Rose and E.L. O’Connell, J. Biol. Chem, 1969, 244, 126
72
F. C. Hartman, Biochemistry, 1970, 9 (8), 1776-1782
73
R. Ducan, Thesis, Stanford University, 1995
74
D. Stribling, Biochem. J. 1974, 114, 725
75
H.-L. Arth and W.-D. Fessner, Carbohydrate Res. 1998, 305, 313-321
91
Chapter 5
76
N. Bischofberger, H. Waldmann, T. Saito, E.S. Simon, W. Lees, M.D. Bednarski
and G.M. Whitesides, J. Org. Chem. 1988, 53, 3457-3465
77
R. Duncan and D.G. Drueckhammer, Tet. Lett., 1993, 34 (11), 1733-1736
78
J.R. Durrwachter, D.G. Druekhammer, K. Nozaki, H. Sweers and C.-H. Wong, J.
Am.Chem.Soc. 1986, 108, 7812
79
D.G. Drueckhammer, J.R. Durrwachter, R.L. Pederson, D.C. Crans, L. Daniels
and C.-H. Wong, J. Org. Chem. 1989, 54, 70-77
80
H.J.M. Gijsen, L. Qiao, W. Fitz and C.-H. Wong, Chem. Rev. 1996, 91 (1), 443-
469
81
C.-H. Wong and C.M. Whitesides, J. Org. Chem 1983, 48, 3493-3497
82
Ö. Almarson and A.M. Klibanov, Biotechnology and Bioengineering 1996, 49, 87-
92
83
M.D. Bednarski, E.S. Simon, N. Bischofberger, W.-D Fessner, M.-J Kim, W. Lees,
T. Saito, H. Waldmann and G.M. Whitesides, J. Am. Chem. Soc. 1989, 111, 627-
635
84
R. Schoevaart, F. van Rantwijk, R.A. Sheldon, Tetrahedron: Asymmetry 1999, 10
(4), 705-711
85
R. Schoevaart, F. van Rantwijk, R.A. Sheldon, submitted
92
93
94
6
Transketolase versus
fructose-1,6-bisphosphate aldolase
Abstract
Two approaches were compared for the synthesis of optically pure polyhydroxy
ketones via asymmetric enzymatic C-C bond formation. D-Fructose-1,6-
bisphosphate aldolase (FruA) catalyzes the aldol reaction of an aldehyde with
dihydroxyacetone phosphate (DHAP). Transketolase, on the other hand, catalyzes
the formation of the same product, together with CO2, by irreversible reaction of an
α-hydroxyaldehyde with β-hydroxypyruvate. When the α-hydroxyaldehyde is racemic
the reaction is a kinetic resolution, i.e. both enantio- and diastereoselective.
Transketolase catalyzed reactions proved to be excellent alternatives to FruA-based
synthesis, with the added advantages that transketolase is readily available from a
recombinant E. coli and it does not require a phosphorylated substrate. Cross-linked
enzyme crystals of tranketolase were more stable but substantially less active than
the native enzyme.
95
Chapter 6
Introduction
Transketolase86 catalyzes the reversible transfer of a ketol group between D-
xylulose-5-phosphate and D-ribose-5-phosphate (Figure 1) or D-xylulose-5-
phosphate and erythrose-4-phosphate in the pentose phosphate pathway. Together
with transaldolase it creates a reversible link between two main metabolic pathways,
the pentose phosphate pathway and glycolysis.
CH2OH
O H
CH2OH O
C
O H
O HO
OH C
HO transketolase OH
OH OH
OH -2 OH
OH OPO3
-2
OPO3 -2 OH
OPO3
-2
OPO3
Xylulose Ribose Glyceraldehyde Sedoheptulose

5-phosphate 5-phosphate 3-phosphate 7-phosphate
Figure 1. Natural substrates of transketolase
Transketolase requires vitamin B1 derived thiamine pyrophosphate (TPP)87 and
magnesium(II) ions to function, but in only catalytic amounts. The enzyme is a dimer
with the cofactor and the active site located at the interface between the two identical
subunits88. A variety of sources for transketolase have been described; however,
commercially available samples are generally derived from yeast89 or spinach
leaves.90 More recently, enzyme obtained from an over-expressed E. coli
transformant91,92 has attracted considerable interest. This source is more suitable for
96
Transketolase versus fructose-1,6-bisphosphate aldolase
large scale production of transketolase. The properties93 of all three are very similar,
but the specific activity of the E. coli transketolase is about ten fold higher.
When β-hydroxypyruvate94,95,96 is used instead of D-xylulose-5-phosphate as
the ketol donor substrate the reaction becomes irreversible. Moreover, with the
appropriate choice of α-hydroxyaldehydes (Figure 2) products identical to the ones
made by D-fructose-1,6-bisphosphate aldolase (FruA) mediated aldol reactions are
obtained.
OH O OH O
H O OH transketolase
+ OH
R R
-
O
O OH
CO2
O O OH O
1) FDP-aldolase
HO OPO32- OH
R H +
2) acid phosphatase R
OH
Figure 2. Transketolase and FruA can make identical products.
Hydroxypyruvate decomposes slowly in solution, but it is stable enough in Tris
and glycylglycine buffers97. The use of a pH autotitrator gives the best results since
the progress of the reaction is accompanied by a pH increase caused by the
consumption of hydroxypyruvic acid, which can simply be fed on demand98 to control
the pH. Using this system, transketolase can be applied on a large scale99 with high
productivity (space-time yield of 45 gL-1d-1 for L-erythrulose100,101,102).
Over sixty aldehydes103 have been identified as acceptor substrates.
97
Chapter 6
Transketolase is rather flexible towards the acceptor substrate, although its
requirement for the (R) configuration of any hydroxy functionality at C-2 is absolute.
This feature has been used to prepare optically pure L-2-hydroxyaldehydes by
means of kinetic resolution of racemic α-hydroxyaldehydes104. Aldehydes without a
hydroxyl group at C-2 are also converted. A variety of products is accessible via the
transketolase route, e.g. N-phenylglycolhydroxamate105, natural ketoses and
analogues106, labeled xylulose107, the pheromone (+)-exo-Brevicomin108 and the
glycosidase inhibitor Fagomin109, 1,4-dideoxy-1,4-imino-D-arabinitol110, aromatics
from D-glucose111 and D-sedoheptulose112. The half-life of the enzyme is strongly
influenced by the specific aldehyde. Especially non-polar aldehydes such as
acetaldehyde and aromatic aldehydes are detrimental to the enzyme stability and
cause a decrease in the half-life from 24 hours to under one hour.
Transketolase also provides access to compounds that cannot be obtained
using DHAP aldolases. For example α,β-unsaturated aldehydes are not substrates
for FruA, but α-hydroxy-β,γ

-unsaturated aldehydes are substrates for transketolase.
Transketolase can even accommodate α,β-unsaturated aldehydes as substrates,
once again demonstrating its versatility. Mutants of transketolase have been
prepared to study its active side structure and the mechanism of the reaction113,114,
but also to broaden its substrate spectrum115.
α-Hydroxyaldehydes
Because the use of α-hydroxyaldehydes in transketolase catalyzed reactions affords
products identical with those made by FruA, we were particularly interested in this
98
group of acceptors. The synthesis of many α-hydroxyaldehydes has been
described98,116. A few strategies can be distinguished: ozonolysis of allylic alcohols
with reductive workup, addition reaction of glycidaldehyde dimethyl acetal,
chlorination at the α-position of an aldehyde followed by hydrolysis117 and synthesis
via silyl protected 2-hydroxyesters. Chiral α-hydroxyaldehydes can be
enantiomerically pure or racemic. Because transketolase is absolute in its
requirement for the (R) configuration of the hydroxy functionality at C-2, racemic
aldehydes also give enantiomerically pure products via a kinetic resolution. Acetals
are commonly used for stable storage of aldehydes and their hydrolysis is
accomplished under acidic aqueous conditions. Storage of aldehydes as aqueous
solutions is likewise convenient.
We chose a small group of substrates for transketolase for comparison with the
synthesis of the same products using FruA (Table I). Two commercially available α-
hydroxyaldehydes were glycolaldehyde and D,L-glyceraldehyde. Lactaldehyde was
synthesized via the reduction of methylglyoxal 1,1-dimethylacetal118. Chlorination at
the α-position of butanal and pentanal followed by hydrolysis was used to prepareα-
hydroxybutanal and α-hydroxypentanal. The corresponding aldehydes for the FruA-
catalyzed syntheses were formaldehyde, glycolaldehyde, acetaldehyde, propanal
and butanal, which are all commercially available.
With the exception of the last entry (Table I) the overall reaction times were
quite similar, even when initial reaction rates differed. However, the conversions of
the transketolase catalysed reactions, with the exception of glyceraldehyde, were
almost quantitative, which is significantly higher than those of FruA. α-
Hydroxyaldehydes are also particularly good substrates for FruA, but only reactions
generating an identical product are considered here. Hence, the use of transketolase
99
Chapter 6
was in most cases superior on account of the high yield. The exception was the
synthesis of D-xylulose mediated by FruA which gave a higher reaction rate and
conversion than the corresponding reaction of transketolase.
In practice, however, the complete synthetic route must be taken into account.
Transketolase requires other aldehydes than FruA and some might have to be
synthesized while others are commercially available. This has to be considered for
each different end product.
Table I. Comparison of transketolase versus FruA synthesizing identical products.
FruA acceptor = TK acceptor =

RCHO R(CHOH)CHO
R Vini conversion Vini conversion
H 100 81 (2h) 100 100 (0.5h)

CH3 81 84 (2.5h) 21 99 (2h)
CH2CH3 75 67 (4h) 58 97 (2h)
(CH2)2CH3 28 70 (4h) 23 96 (4h)
CH2OH 190 96 (0.5h) 12 75 (6h)
Reaction conditions: 25 mM donor, 50 mM (100 mM racemic) aldehyde and 1
unit enzyme.
Cross-linked enzyme crystals
In order to improve the stability of transketolase, the application of cross-linked
transketolase crystals was examined. Stabilization with lipids119,120 was not
successful and immobilization on Eupergit-C had already been accomplished121.
100
Crystallization to collect X-ray crystallographic data from E. coli transketolase had
opened the possibility to make cross-linked enzyme crystals122. Cross-linking
crystals of FruA from rabbit muscle had also been accomplished123,124, but the
stability was not better than bacterial FruA, which happened to be very stable. Cross-
linked crystals generally have increased stability, especially when cosolvents are
applied.
Crystals of transketolase were prepared in 35% activity yield and subsequently
cross-linked with glutaraldehyde. with a yield of 13%. Upon cross-linking more than
62% activity was lost, resulting in an overall activity yield of only 13%.
50
40
TK activity (%)
30
20
10
0
0 1 2 3 4 5
Glutaraldehyde (mM)
Figure 3. Activity of transketolase crystals after cross-linking with glutaraldehyde.
The most active cross-linked transketolase crystals were obtained at exactly 1.6 mM
glutaraldehyde (Figure 3). Any deviation from this concentration resulted in a
dramatic reduction in activity. It is not clear why the reduction in activity is so abrupt.
The stability of cross-linked transketolase crystals was tested in glycylglycine
101
Chapter 6
buffer and compared with the native enzyme (Figure 4). It was found that the crystals
were more stable, but since 87% activity was lost during their preparation, this
increased stability is not beneficial when normal buffers are used. Addition of more
native enzyme would be far more efficient.
120 native TK
TK-CLEC
100
TK activity (%)
80
60
40
20
0
0 5 10 15 20 25
t (days)
Figure 4. Stability of native and cross-linked TK.
Conclusions
Transketolase is capable of competing with FruA as far as reaction rate and
conversion of the corresponding acceptor substrates is concerned. In order to
assess the best way to synthesize a target molecule both the accessibility of the
different acceptor aldehydes and their activity as a substrate have to be considered.
In cases where it is not possible to use certain aldehydes, e.g. α,β-unsaturated
aldehydes with FruA, transketolase is the automatic choice. In other examples the
use of transketolase could lead to higher yields, making it competitive with FruA. In
102
the preparation of polyhydroxy ketones, use of either one should always be
considered.
Experimental
General
temperature controller.
Transketolase catalysed reactions
In a gastight bottle hydroxypyruvate (HPA, 6.6 mg Li-salt, 25 mM) and an α-
hydroxyaldehyde (50 mM for glycolaldehyde and 100 mM for racemic ones) were
added to 1 ml of TRIS buffer pH 7.6. Then 0.1 ml was withdrawn to monitor the
background reaction. The reaction was initiated by adding transketolase (1 unit). At
7.6. The neutralized mixture was then assayed for HPA by lactate dehydrogenase
mediated reduction with NADH.
FruA catalysed reactions
Aldol reactions were performed in 1 ml 50 mM TRIS buffer pH 7.6 containing 25 mM
DHAP and 50 mM aldehyde, reactions were initiated by adding aldolase (1 unit). At
103
Chapter 6
7.6. This neutralized mixture was then assayed for DHAP. The reaction was followed
until only background hydrolysis of DHAP was measured (0.0047 mM/min).
Hydroxypyruvate assay
50 µl aliquots were taken from the reaction mixture at intervals and quenched with 15
µl 7% perchloric acid. After 30 minutes 10 µl 1M NaOH was added and 175 µl of 50
mM TRIS pH 7.6. From this neutralized and diluted mixture a 40 µl sample was
added to a quartz cuvette containing 1.95 ml 50 mM TRIS pH 7.6 and 0.16 mM
NADH. After the addition of 1.7 units lactate dehydrogenase (10 µl of 1 mg in 1 ml
sat. A2S) the absorption was monitored at 340 nm at 25 °C. The molar absorption
coefficient taken was 6.22 l.mmol-1cm-1.
Synthesis of D,L-lactaldehyde
To a three-necked flask containing 7.2 g LiAlH4 (190 mmol) in 150 ml Et2O 20.4 g
(172 mmol) methylglyoxal 1,1-dimethylacetal was added dropwise under a nitrogen
atmosphere. After the addition of 75 ml Et2O the mixture was refluxed for 30 minutes
and then quenched with saturated sodium chloride solution. After drying with sodium
sulfate the solvent was evaporated and the resulting oil distilled (bp 62-67 °C 30 mm
Hg). Deprotection was effected by heating a 0.7 M solution pH 1.5 at 60 °C for 1
hour. This solution was ready for use after neutralization and could be stored at 5 °C
for months. The exact concentration of the aldehyde was determined by reaction
with hydrazine monitored at 240 nm. The molar absorption coefficient taken was
104
2.73 l.mmol-1cm-1.
Synthesis of α-hydroxybutanal
In a flask equipped with a stirrer and reflux condenser were placed 67 ml (0.75 mole)
of butyraldehyde and 30 ml of methylene chloride. After the solution was cooled to
10 °C in an icebath, a mixture of 72 ml (0.75 mole) of sulfuryl chloride (SO2Cl2) and
10 ml of methylene chloride was slowly added maintaining the temperature between
15 and 40 °C. After completion of the addition, the reaction was stirred for one-half
hour. After addition of 20 ml of methylene chloride the mixture was refluxed for one
hour. All the volatile material was then distilled rapidly at reduced pressure.
Distillation through a 20 cm Vigreux column afforded 24.9 g product (25%) (b.p. 104-
110 °C).
To avoid polymerisation, acetalyzation should be conducted immediately. To 77 ml
of dry methanol was added slowly 5.2 g of sodium, the temperature was maintained
below 20 °C and 24.9 g of α-chlorobutyraldehyde was added. The ice-bath was
removed and the reaction mixture was stirred for one hour. To neutralise any
unreacted sodium water was carefully added. Then anhydrous sodium sulfate and
ethyl acetate was added. After filtration the solvents were rapidly distilled off at
reduced pressure. Distillation at reduced pressure (1 mm Hg) 13.9 g (46%) of α-
hydroxybutyraldehyde dimethyl acetal (b.p. 39-42 °C).
Deprotection was accomplished in 2 hours at pH 1.5 and 60 °C.
105
Chapter 6
Synthesis of α-hydroxypentanal
Using the same procedure as for the preparations of a-chlorobutyraldehyde, the α-
chloropentanal (b.p. 103°C) could be prepared affording 25 g (30% yield) α-
hydroxypentanal dimethyl acetal (b.p. 95 °C, 25 mm Hg).
Deprotection was accomplished in 3 hours at pH 1.5 and 70 °C.
Synthesis of 1,3,4-tri-O-acetyl-5-deoxy-5-ethyl-D-xylulose
A 171 ml solution containing 50 mM Tris pH 7.8, 25 mM β-hydroxypyruvate (HPA, 50
mM 2-hydroxypentanal and 171 units transketolase was stirred at room temperature.
Conversion was 99.4% (based on consumption of HPA) after 16 h. Extraction with
ethylacetate afforded 1.412 g (c.y. 90% yield) product and L-2-hydroxypentanal.
Acetylation with pyridine and acetic anhydride yielded 1,3,4-tri-O-acetyl-5-deoxy-5-
ethyl-D-xylulose (analysed using 1H NMR spectra of authentic samples.
Crystallization and cross-linking
Crystallization of transketolase A from recombinant Escherichia coli K12 was
performed in an eppendorf reaction vessel, sealed in a jar containing CaCl2 and kept
at 4 oC. In each vessel 90 ml of transketolase solution (17 mg / ml) was placed in
120 ml 50 mM glycylglycine buffer pH 8.0 containing 25% PEG 1450. To initiate
crystallization 15 ml of sonificated transketolase crystals were added. After 4 days
crystals were collected by centrifugation and suspended in 560 ml 500 mM
triethanolamine-buffer pH 7.6 also containing 25% PEG and divided into smaller
portions. Varying amounts of 50 mM glutaraldehyde were added for cross-linking
and the solutions were shaken for one hour at 4 oC. The insoluble TK-crystals were
106
washed with water and assayed for activity.
References
86
P. Srere, J.R. Cooper, M. Tabachnik and E. Racker, Archiv. of Biochem. and
Biophysics, 1958, 74, 295-305
87
G. Schneider and Y. Lindqvist, Bioorganic Chem. 1993, 21, 109-117
88
Y. Lindqvist, G. Schneider, U. Ermler and M. Sundström, The EMBO Journal,
1992, 11 (7), 2373-2379
89
G. De La Haba, I. Leder and E. Racker, J. Biol. Chem. 1955, 214, 409
90
J. Villafranca and G. Axelrod, J. Biol. Chem., 1971, 246, 3126-3130
91
G.R. Hobbs, M.D. Lilly, N.J. Turner, J.M. Ward, A.J. Willets and J.M. Woodley, J.
Chem. Soc. Perkin Trans. 1, 1993, 165-166
92
M. Gyamerah and A.J. Willetts, Enzyme Microb. Technol., 1997, 20, 127-134
93
G.A. Sprenger, U. Schörken, G. Sprenger and H. Sahm, Eur. J. Biochem, 1995,
230, 525-532
94
Large scale synthesis: K.G. Morris, M.E.B. Smith, N.J. Turner, M.D. Lilly, R.K.
Mitra and J.M. Woodley, Tetrahedron: Asymmetry, 1996, 7 (8), 2185-2188
95
HPLC analysis: R.K. Mitra and J.M. Woodley, Biotechnology Techniques, 1996, 10
(3), 167-172
96
Enzymatic assay: A.W. Holldorf, Methods of enzymatic analysis, 1966, 3th Ed.,
Vol. 6, 614-618
97
V. Dalmas and C. Demuynck, Tetrahedron: Asymmetry, 1993, 4 (11), 2383-2388
98
Y. Kobori, D.C. Myles and G.M. Whitesides, J. Org. Chem., 1992, 57, 5899-5907
99
J.M. Woodley, R.K. Mitra and M.D. Lilly, Ann. N. Y. Acad. Sci, 1996, 799, 434-445
100
R.P. Chauhan, J.M. Woodley and L.W. Powell, Ann. N. Y. Acad. Sci, 1996, 799,
545-554
101
R.P. Chauhan, L.W. Powell, J.M. Woodley, Biotech. Bioeng., 1997, 56 (3) 345-
351
102
J. Bongs, D. Hahn, U. Schörken, G.A. Sprenger, U. Kragl and C. Wandrey,
Biotechnolgy Lett., 1997, 19 (3), 213-215
107
Chapter 6
103
U. Schörken and G.A. Sprenger, Biochimica et Biophysica Acta, 1998, 1385, 229-
243
104
F. Effenberger, V. Null and T. Ziegler, Tetrahedron Lett., 1992, 33 (36), 5157-
5160
105
M.D. Corbett and B.R. Chipko, Biochem. J., 1977, 165, 263-267
106
J. Bolte, C.Demuynck and H. Samaki, Tetrahedron Lett., 1987, 28, 5525-5528
107
C.Demuynck, J. Bolte, L. Hecquet and H. Samaki, Carbohydrate Research, 1990,
206, 79
108
D.C. Myles, P.J. Andrulis III and G.M. Whitesides, Tetrahedron Lett. 1991, 32 (37)
4835-4838
109
F. Effenberger and V. Null, Liebigs Ann. Chem. 1992, 1211-1212
110
L. Hecquet, M. Lemaire, J. Bolte and C. Demuynck, Tetrahedron Lett., 1994, 35
(47), 8791-8794
111
K.M. Draths, D.L. Pompliano, D.L. Conley, J.W. Frost, A. Berry, G.L. Disbrow,
R.J. Staversky and J.C. Lievense, J. Am. Chem. Soc. 1992, 114, 3956-3962
112
V. Dalmas and C. Demuynck, Tetrahedron: Asymmetry, 1993, 4 (6), 1169-1172
113
U. Nilsson, L. Meshalkina, Y. Lindqvist and G. Schneider, J. Biol. Chem., 1997,
272 (3) 1864-1869
114
G. Schneider and Y. Lindqvist, Biochimica et Biophysica Acta, 1998, 1385, 387-
398
115
C. French and J.M. Ward, Ann. N. Y. Acad. Sci, 1996, 799, 11-18
116
C. André, J. Bolte and C. Demuynck, Tetrahedron: Asymmetry, 1998, 9, 1359-
1367
117
C.L. Stevens, E. Farkas and B. Gillis, J. Am. Chem. Soc., 1954, 2695-2698
118
J.R. Durrwachter, D.G. Drueckhammer, J. Am. Chem. Soc., 1986, 108, 7812-
119
Y. Okahata, Y. Fujimoto and K. Ijiro, J. Org. Chem., 1995, 60, 2244-2250
120
Y. Okahata and T. Mori, Tibtech, 1997, 15, 50-54
121
S.P. Brocklebank, R.K. Mitra, J.M. Woodley and M.D Lilly, Ann. N. Y. Acad. Sci,
1996, 799, 729-736
122
J. Littlechild, N. Turner, Acta Cryst. 1995, D51, 1074-1076
123
J. Sygusch and D. Beaudry, Journal of Biological Chemistry, 1984, 259 (16)
10222-10227
108
124
S.B. Sobolov, A. Bartoszko-Malik, T.R. Oeschger and M.M. Montelbano,
Tetrahedron Letters, 1994, 35 (42), 7751-7754
109
110
7
2-Deoxyribose-5-phosphate aldolase (DERA)
Abstract
D-2-Deoxyribose-5-phosphate aldolase (DERA) is the only aldolase which accepts
two aldehydes as substrates. The natural donor acetaldehyde can be replaced by
e.g. acetone, making DERA suitable for synthesizing both β-hydroxyaldehydes and
β-hydroxy ketones by chain elongation with 2- or 3-carbons, respectively. Low
reaction rates for non-natural acceptor substrates however, call for long reaction
times and large amounts of enzyme making DERA a rather inefficient catalyst. The
stability could be improved by addition of cosolvent.
111
Chapter 7
Introduction
Whereas other aldolases use ketones as aldol donors, D-2-deoxyribose-5-phosphate
aldolase (DERA) accepts two aldehydes as substrates, namely acetaldehyde and D-
glyceraldehyde-3-phosphate. These two aldehydes undergo reversible C-C bond
formation to give 2-deoxyribose-5-phosphate (Figure 1). The newly generated chiral
center always has the (S) configuration.
HO O
HO
DERA
-2
O3PO O + -2
O3PO O
H
D-glyceraldehyde- acetaldehyde OH
3-phosphate
-2
O3PO
O
OH 2-deoxyribose-5-phosphate
OH
Figure 1. Synthesis of 2-deoxyribose-5-phosphate by 2-deoxyribose-5-phosphate
aldolase (DERA) from D-glyceraldehyde-3-phosphate and acetaldehyde.
DERA plays a key role in the biosynthetic pathway125 of the sugar moiety of DNA.
Obviously, this enzyme can provide a route to a wide range of potentially biologically
active compounds, in particular the synthesis of deoxysugars such as deoxyriboses,
2-deoxyfucose analogs, dideoxyhexoses, trideoxyhexoses, deoxythiosugars and

13
C-substituted D-2-deoxyribose-5-phosphate. It also affords a route to a variety of
chiral aldehydes. Synthesis of N-heterocycles which are used as building blocks for
112
antihypertensive drugs is possible via azide substituted acceptors.
This type I aldolase126 has been isolated from different sources, but is
commonly derived from the E. coli gene, overexpressed in the same microbe. DERA
from E. coli is a dimer of identical subunits. It is not commercially available, but the
organism containing the gene, which is suitable for preparing the enzyme on a large
scale127 can be obtained from the ATCC.
Donor substrates
The natural nucleophilic reactant, acetaldehyde, can be replaced by propanal,
acetone or fluoroacetone128. Introducing a methyl group at C-2 makes it possible to
synthesize chiral β-hydroxyketones, besides β-hydroxyaldehydes. In this case two
new stereogenic centers are formed, having the 2(R) and 3(S) configuration. The
use of fluoroacetone as a donor results in the regioselective addition at the non-
fluorinated carbon. Acetol, chloroacetaldehyde, pyruvonitrile and glycolaldehyde are
not donor substrates for DERA, presumably due to steric hindrance in the active site.
Hence, at C-2 of this compound, hydrogen can only be substituted by a methyl
group.
Acceptor substrates
A wide variety of aldehydes, including aldose sugars128, can be used as acceptor for
DERA, but the reaction rates with unnatural acceptor substrates are relatively low
113
Chapter 7
(Table 1) and a relatively large amount of enzyme is required. The adducts of 2-
chloroaldehydes can be converted to epoxy aldehydes and epoxy ketones. N- and
S-heterocycles129 are obtained by the use of 3-azido- or 3-thiolaldehydes. The need
for a phosphate group at the 3-hydroxy position is reflected in the reaction rate; use
of D- or L-glyceraldehyde results, in both cases, in a 250-fold reduction in reaction
rate compared with D-glyceraldehyde-3-phosphate.
Table 1. Relative rates for DERA substrates (references 128 and 129).
Acceptor Vrel
D-glyceraldehyde-3-P 100
L-glyceraldehyde-3-P 5
D-glyceraldehyde 0.4
L-glyceraldehyde 0.4
acetaldehyde 0.03
chloroacetaldehyde 0.3
propionaldehyde 0.07
butanal 0.03
Sequential aldol additions
A problem associated with DERA is its competing self aldol reaction of acetaldehyde.
In principle each acceptor will react with the donor acetaldehyde, but when the
reaction rate is near or lower than the rate (Table I) with acetaldehyde as acceptor,
selfcondensation130,131 of acetaldehyde will predominate. This limits the number of
useful acceptor substrates. However, it has been used to advantageously synthesize
114
2,4,6-trideoxyhexoses, which are precursors for cholesterol lowering agents132. This
compound is created by a three-substrate aldol reaction (Figure 2).
O O O
DERA R O OH
+
R H + H H
R=Me, 22%
R= (CH2)2COOH, 80% OH
Figure 2. Sequential three-substrate aldol additions with DERA.
The yields vary widely (3-80%) and long reaction times (2 weeks with 2500 units
DERA / mmol) are necessary. With DERA having a specific activity of 58.1 units per
mg, 4.3 g of pure enzyme per liter would be needed. Having an activity less than
10% makes it almost impossible to perform these reactions on a reasonable scale.
Even low yields of reaction products from a four-substrate aldol reaction131 can be
detected.
O O O R
DERA,
2- OH
RAMA
+ + HO OPO3 O
R H H
OH
OH OH
Figure 3. Sequential one-pot aldol reaction catalysed by DERA and RAMA.
The combination of DERA with D-fructose-1,6-bisphosphate aldolase from rabbit
muscle using a 2-substituted acetaldehydes (see Figure 3) allowed the synthesis of
mixtures of 5-deoxy ketoses133 in low yields (10-50%) after long reaction times (6
115
Chapter 7
days). Furthermore the combination of DERA with N-acetylneuraminic acid aldolase
(NeuAc aldolase) is possible, leading also to 2,4,6-trideoxyhexoses.
Summarizing, DERA shows flexibility towards both donor and acceptor substrate
analogues, making it possible to synthesize a variety of β-hydroxyaldehydes or
ketones. The very low reaction rates for non-natural acceptor substrates requires
sometimes as much as 250.000 units of the enzyme per liter, while the reaction time
is in the order of weeks. DERA is therefore a rather inefficient enzyme.
Results and discussion
A possible solution to the problem of low activity towards acceptor substrates that is
associated with the use of DERA could be the use of cross-linked enzyme crystals to
increase activity and/or long term stability during the reaction. Crystallization
procedures have been developed for acquisition of x-ray crystallographic data of
DERA crystals134. However, presumably because of impure enzyme preparations, no
DERA crystals could be obtained in our laboratory leaving the possible benefits of
cross-linking unexplored.
In a second approach DERA was covalently immobilized on Eupergit C 250L.
with 45% activity retention of activity. Immobilization provided a simple procedure to
separate the enzyme from the reaction mixture, but no improvement in the
operational stability compared with the native enzyme was observed.
116
Stability in cosolvents
To conclude our investigation of DERA the use of cosolvents was investigated. Since
the acceptor substrates in DERA catalyzed reactions are often quite hydrophobic,
solubility can pose a significant obstacle. Incomplete mixing of substrates with the
aqueous medium causes enzyme deactivation as well as a reduced rate of reaction.
To increase the solubility of hydrophobic substrates the use of cosolvents was
examined. Shorter reaction times would be possible in view of the higher
concentration of the acceptor substrate by complete mixing with the reaction
medium. These cosolvents can furthermore increase enzyme stability because no
liquid-liquid interface is present, thereby reducing deactivation of the enzyme during
the reaction. The stability of DERA can be defined as the residual activity after
incubation in water / solvent mixtures. In Table 2 the residual activities of DERA after
one day incubation in different water / solvent combinations (75:25, v/v) are
tabulated. It is apparent that DERA can be combined rather well with cosolvents. For
example DMSO and DMF increased its activity and the long term stability of the
aldolase under these conditions is remarkable as well: after one week of incubation it
was still a completely active. Acetonitrile is detrimental for the stability since it
rendered the enzyme completely inactive after one day already. The behavior of
DERA in ethanol is perplexing: with 25% solvent the residual activity is zero after one
week, whereas with 20% solvent an initial reduction of activity is seen after which it
increases to become 130% of the original activity after one week.
117
Chapter 7
Table 2. Stability (expressed as residual activity) of DERA in 25% cosolventa.
Residual activity (%)

Cosolvent (25%) 1 day 1 week
Ethanol 73 0
DMSO 150 115
Dioxane 85 0
Acetone 55 55
DMF 128 128
t-BuOH 63 0
Acetonitrile 0 0
a
Residual activity in the cleavage of D-2-deoxyribose-5-phosphate
To examine the effect of cosolvent under reaction conditions, that is with the donor
aldehyde present, the stability of DERA was determined in 100 mM acetaldehyde
(Table 3) and 10% ethanol or DMSO. This cosolvent concentration represents a
compromise between enzyme activity (see next paragraph) and stability.
Acetaldehyde clearly has a negative effect on the stability of DERA during
incubation. After one week the enzyme retained only 21% activity, whereas in the
presence of ethanol and DMSO 33 and 36% was found, respectively. Since large
amounts of catalyst are normally required, even modest improvements in stability are
significant as many units of enzyme are preserved. Hence, the use of these
cosolvents in DERA-catalyzed aldol reactions clearly has practical benefits.
118
Table 3. Stability of DERA in 100 mM acetaldehyde and 10% cosolvent.
acetaldehyde % activity (1 day) 7d
- 100 99
+ 38 21
+/10%ethanol 71 33
+/10% DMSO 58 36
Activity in cosolvents
Besides the need for stability of the catalyst during the reaction its activity under the
specified conditions is also important. Determination of the activity of DERA in 25%
cosolvent can be performed in two ways: analysis of the synthetic reaction or
analysis of the cleavage (retro-aldol reaction) of D-2-deoxyribose-5-phosphate, which
is the standard DERA activity test. To follow the progress of the synthetic reaction
sampling is required at time intervals. With the standard activity test a single
measurement is sufficient. We chose this last approach for a quick assessment of
the activity. For the standard DERA activity assay the enzymes triose-3-phosphate
isomerase (TPI) and glycerol-3-phosphate dehydrogenase (GDH) are required. TPI
isomerises D-glyceraldehyde-3-phosphate liberated by DERA from 2-deoxyribose-5-
phosphate to dihydroxyacetone phosphate which in its turn is reduced by NADH
catalysed by GDH. These two enzymes were found to be stable and active in up to
30% ethanol or DMSO (determined with D-glyceraldehyde-3-phosphate).
119
Chapter 7
140 Figure 4.
120 Activity and stability (1

% activity
100
week incubation) of DERA
80
60 in 25% ethanol.
activity
40
stability
20
0
0 5 10 15 20 25
% EtOH
175
150 Figure 5.
125 Activity and stability (1 week

% activity
100
incubation) of DERA in 25%
75
DMSO.
50
activity
25 stability
0
0 10 20 30
% DMSO
The activity of DERA in water / ethanol mixtures immediately drops to become
zero at 25 % ethanol (Figure 4). The stability was excellent at concentrations of up to
20%. Its activity in DMSO was much better (Figure 5); between 0 and 20% cosolvent
the activity of DERA rose to about 150%. At 30% cosolvent three quarters of the
activity remained. The enzyme is very stabile in this cosolvent. These findings are in
favor of the use of DMSO as an additive in DERA catalyzed aldol reactions.
120
Conclusion
The ability of DERA to catalyze the synthesis of either β-hydroxyaldehydes or β-
ketones would broaden the scope of aldolases in organic synthesis. However, DERA
is a rather inefficient enzyme. The very low reaction rates for non-natural acceptor
substrates requires, in some cases, 250.000 units of the enzyme per liter often with
reaction times of weeks and only moderate yields. Improvement of the stability by
addition of cosolvents compensated to some extent the poor performance of this
aldolase but its synthetic utility remains very limited.
Experimental
General
temperature controller
Preparation of DERA
The E. coli strain DH5α (ATCC 86963), with the deo C system containing plasmid p
VH17, was grown in LB broth containing ampiciline at 37 °C on a 0.7 liter scale.
DERA was purified by centrifugation, cell disruption and ammonium sulfate
fractionation. This provided 400 units of DERA (should be 4000 units according to
literature).
121
Chapter 7
Activity assay
The reaction product, GAP was assayed using a coupled enzyme system. To 1.95
ml 50 mM Tris-buffer pH 7.6 containing 0.16 mM NADH in a 2 ml cuvette, were
added 20 µl 190 mM 2-deoxyribose-5-phosphate, 20 µl of a mixture containing 1.25
unit D-glycerol-3-phosphate dehydrogenase and 12.5 unit triose-1-phosphate
isomerase. Then 50 µl diluted aldolase was added and absorption monitored at 25
°C. 1 unit (U) aldolase converts 1 µmol 2-deoxyribose-5-phosphate per minute.
Stability test
Conditions: 1 unit enzyme per 1 ml assay containing the appropriate amount of
cosolvent (v/v), the remainder is 50 mM TRIS-buffer pH 7.6. After incubation a
sample was taken (typically 50 µl which diluted when activity is high) and the activity
is measured with the assay described above.
References
125
E. Racker, J. Biol. Chem. 1952, 196 (1), 347-365
126
B.L. Horecker, S. Pontremoli, C. Ricci and T. Cheng, Proc. Natl. Acad. Sci. U.S.A.
1961, 47, 1942
127
C.-H. Wong, E. Garcia-Junceda, L. Chen, O. Blanco, H.J.M. Gijsen and D.H.
Steensma, J. Am. Chem. Soc. 1995, 117, 3333-3339
128
C.F. Barbas III, Y.-F. Wang and C.-H. Wong, J. Am. Chem. Soc. 1990, 112,
2013-2014
122
129
L. Chen, D.P. Dumas and C.-H. Wong, J. Am. Chem. Soc. 1992, 114, 741-748
130
H.J.M. Gijsen and C.-H. Wong, J. Am. Chem. Soc. 1994, 116, 8422-8423
131
H. Gijsen and C.-H. Wong, J. Am. Chem. Soc. 1995, 117, 7585-7591
132
R. Rosen and C.H. Heathcock, Tetrahedron 1986, 42, 4909
133
H.J.M. Gijsen and C.-H. Wong, J. Am. Chem. Soc. 1995, 117, 2947-2948
134
E.A. Stura, S. Ghosh, E. Garcia-Junceda, L. Chen, C.-H. Wong and I.A. Wilson,
Proteins: structure, function and genentics 1995, 22, 67-72
123
Applications of aldolases in organic synthesis
Summary
Biocatalysts have shown promise in enantioselective synthesis since they compete
with chemical catalysis on e.g. introduction of chirality, selective modification or
resolution of racemates. Aldolases constitute a group of catalysts with many
potential applications in carbon-carbon bond forming reactions. In order to overcome
the shortcomings concerning their employment in future commercial processes,
generic problems like design of non-natural substrates, exploration of enzymes from
new species, enzyme stabilization and multi-enzyme systems need to be
investigated.
In Chapter 2 the acceptor substrate specificity of the commercially available
dihydroxyacetone phosphate (DHAP) dependent aldolases, D-fructose-1,6-
bisphosphate (FruA), L-fuculose-1-phosphate (FcuA) and L-rhamnulose-1-
phosphate (RhuA) aldolase, was investigated. Enzymatic detection of the donor
DHAP was found to be the fastest and most accurate way of determining the
progress of the aldol reactions and, hence, the kinetic properties of the aldolases.
Comparison of FruA, FcuA and RhuA showed that FcuA functioned less well
when non-natural substrates were used, whereas FruA and RhuA both performed
well and with similar profiles. More than twenty aldehydes were found to be
substrates for the aldolases from rabbit muscle, S. carnosus and S. aureus. The
bacterial aldolases were much more stable than the mammalian one.
Since the stereochemical outcome may differ when substrates in the aldol
reaction are modified, it is necessary to find unambiguous evidence for the
configuration and optical purity of the products. In Chapter 3 chiral
gaschromatography and a new retro aldol reaction assay were used to determine the
124
stereoselectivity of the aldolases towards a number of acceptor substrates. This
proved unambiguously, for the first time, that the steric preference of FruA and RhuA
for the newly created stereogenic center at C-(3) was far from absolute. The
enzymatic method is also useful for fast monitoring of aldolase catalysed syntheses.
In Chapter 4 a one-pot reaction of four enzymatic steps is described. Phytase-
catalyzed phosphorylation of glycerol with pyrophosphate was a key step which
afforded D,L-glycerol-3-phosphate in quantitative yield. Switching off the activity of
the phosphatase by a pH change during oxidation and aldol reaction resulted in
better results than integral in situ phosphorylation. Dephosphorylation was effected
by lowering the pH back to 4. Combined with the broad substrate specificity of DHAP
aldolases towards acceptor substrates it constitutes a simple procedure for the
synthesis of a wide variety of non-natural carbohydrates from glycerol. At the
moment this is the most economical preparation of DHAP.
Chapter 5: Hydroxyacetone thiosulfate was prepared as a substrate analogue
for DHAP, but it was not converted by DHAP utilizing enzymes. In situ formed
dihydroxyacetone arsenate was investigated in depth as analogue for DHAP. Use of
arsenate allows uncomplicated performance of aldol reactions, since synthesis of
phosphorylated donor is avoided. Analysis was afforded by enzymatic detection of
the acceptor substrates. It was found that cosolvents could be used with a triple
advantage: increased solubility of hydrophobic aldehydes, faster reaction and
stabilisation of the enzyme. The stereospecifity is somewhat higher with the natural
donor substrate.
Two approaches for the synthesis of optically pure polyhydroxy ketones (with
transketolase and FruA) were compared in Chapter 6. In order to assess the best
way to synthesize a target molecule both the accessibility of the different acceptor
125
aldehydes and their activity as a substrate were considered. Transketolase was
capable of competing with FruA as regards reaction rate and conversion of the
corresponding acceptor substrates.
In Chapter 7 the use of 2-deoxyribose-5-phosphate aldolase (DERA) is
examined. It is the only aldolase which accepts two aldehydes as substrates. The
ability of DERA to catalyze the synthesis of either β-hydroxyaldehydes or β-ketones
broadens the scope of aldolases in organic synthesis, however, DERA is a rather
inefficient enzyme. Improvement of the stability by addition of cosolvents
compensated to some extent the poor performance of this aldolase but its synthetic
utility remains very limited.
Summarizing, this work has resulted in a new synthesis of DHAP which makes this
important substrate cheap and readily available in large amounts. The
stereoselectivity of the aldol reaction was clarified by developing a “stereoisomer
assay”. The use of stable and active bacterial aldolases was promoted as an
alternative for aldolase from rabbit muscle. Dihydroxyacetone arsenate was shown
to be attractive for facile labscale experimentation. Furthermore, investigation of
other aldolases demonstrated these enzymes to be usefull catalysts.
Rob Schoevaart
126
Samenvatting
Biokatalysatoren zijn veelbelovend voor het gebruik in enantioselectieve syntheses,
omdat ze kunnen concurreren met chemische katalysatoren met betrekking tot de
introductie van chiraliteit, selectieve modificatie of resolutie van racematen.
Aldolasen vormen een groep katalysatoren met vele potentiële toepassingen in
koolstof-koolstof band vormende reacties. Om hun toepassingsmogelijkheden in
commerciële processen te vergroten, moeten algemene problemen zoals het
ontwerpen van niet-natuurlijke substraten, enzym stabiliteit, multi-enzym systemen
onderzocht worden. Tevens moeten enzymen uit nieuw gevonden organismen
uitgeprobeerd worden.
In Hoofdstuk 2 is de acceptor substraat specificiteit van de commercieel
verkrijgbare dihydroxyaceton fosfaat (DHAP) afhankelijke aldolasen, D-fructose-1,6-
bisfosfaat (FruA), L-fuculose-1-fosfaat (FcuA) and L-rhamnulose-1-fosfaat (RhuA)
aldolase beschreven. Enzymatische detectie van de donor DHAP bleek de snelste
en meest nauwkeurige manier om de voortgang van de aldol reacties en dus de
kinetisch eigenschappen van aldolasen te onderzoeken.
Een vergelijking van FruA, FcuA en RhuA toonde aan dat FcuA minder goed
functioneert als niet-natuurlijke substraten worden gebruikt, terwijl FruA en RhuA
beiden goed presteren met gelijkwaardige eigenschappen. Meer dan twintig
aldehyden werden als substraat voor de aldolasen van konijnespierweefsel,
Staphylococcus carnosus en Staphylococcus aureus gevonden. De bacterieële
aldolasen zijn veel stabieler dan de dierlijke.
Omdat de stereochemische uitkomst van de reactie anders kan zijn als niet
natuurlijke substraten in de aldol reactie gebruikt worden, is het nodig om
127
onomstotelijk bewijs te vinden voor de configuratie en de optische zuiverheid van de
producten. In Hoofdstuk 3 is chirale gaschromatografie en een nieuwe retro aldol
reactie assay gebruikt om de stereoselectiviteit van de aldolasen met een aantal
acceptor substraten te bepalen. Hieruit bleek voor het eerst dat de sterische
voorkeuren van FruA en RhuA voor het nieuwe chirale centrum op C-(3) niet
absoluut is. De enzymatische methode is ook handig voor een snelle bepaling van
de productvorming in aldolase gekatalyseerde reacties.
In Hoofdstuk 4 is een éénpotsreactie beschreven die bestaat uit vier
verschillende enzymatische stappen. Fytase gekatalyseerde fosforylering van
glycerol met pyrofosfaat was de sleutel tot de kwantitatieve bereiding van D,L-
glycerol-3-fosfaat. Het uitschakelen van de aktiviteit van de fosfatase door een pH
verandering (van pH 4 naar pH 7) tijdens de oxidatie en de aldol reactie leverde
betere resultaten op dan een integrale in situ fosforylering. Door de pH weer te
verlagen naar 4 kwam de defosforylering op gang. In combinatie met de brede
substraat specificiteit van DHAP aldolasen voor acceptor substraten vormt de
eenpotsreaktie een eenvoudige procedure voor de synthese van een keur aan
onnatuurlijke koolhydraten uit glycerol. Op dit moment is het ook de meest
economische bereiding van DHAP.
Hoofdstuk 5: Hyroxyaceton thiosulfaat was gesynthetiseerd als substraat
analoog voor DHAP, maar het werd niet omgezet door DHAP-afhankelijke enzymen.
In situ gevormde dihydroxyaceton arsenaat is uitgebreid onderzocht als analoog
voor DHAP. Het gebruik van arsenaat staat het ongecompliceerd uitvoeren van aldol
reacties toe, omdat de synthese van gefosforyleerde donor vermeden wordt.
Analyse van de reactie werd deze keer uitgevoerd door middel van enzymatische
detectie van de acceptor substraten. Cosolvents konden met een drievoudig
128
voordeel gebruikt worden: betere oplosbaarheid van hydrofobe aldehyden, snellere
reactie en stabilisatie van het enzyme. De stereospecificiteit is voor het natuurlijke
donorsubstraat iets hoger.
Twee benaderingen voor de synthese van optisch zuivere polyhydroxy ketonen
(met transketolase en FruA) werden vergeleken in Hoofdstuk 6. Om de beste
synthese route te vinden voor het doel molecuul is het nodig de toegankelijkheid van
de verschillende acceptor aldehydes en hun activiteit als substraat te overwegen.
Transketolase was in staat te concurreren met FruA m.b.t. de reactiesnelheid en
omzetting van de overeenkomstige acceptor substraten.
In Hoofdstuk 7 is 2-deoxyribose-5-phosphate aldolase (DERA) onderzocht. Het
is de enige aldolase die twee aldehyden als substraat accepteert. De mogelijkheid
van DERA om de synthese van β-hydroxyaldehyden or β-ketonen te katalyseren
verbreedt de scope van aldolasen in organische syntheses. Helaas is DERA een
zeer inefficient enzym. Verbetering van de stabiliteit door de toevoeging van
cosolvents compenseerde de slechte prestaties een beetje, maar de synthetische
toepasbaarheid blijft beperkt.
Samengevat heeft dit onderzoek tot een nieuwe synthese van DHAP geleid
waardoor dit belangrijke substraat goedkoop en gemakkelijk in grote hoeveelheden
gemaakt kan worden. De stereoselectiviteit van de aldol reaktie is opgehelderd door
hiervoor een speciale enzymatische assay te ontwikkelen. Voordelen van het
gebruik van stabiele en actieve bacteriële aldolasen boven het gebruik van de
dierlijke uit konijnespierweefsel zijn benadrukt. Dihydroxyaceton arsenaat bleek
aantrekkelijk te zijn voor eenvoudige experimenten op labschaal. Het onderzoek aan
andere aldolasen heeft de bruikbaarheid van deze enzymen aangetoond.
Rob Schoevaart
129
Dankwoord
Dit werk is mede tot stand gekomen dankzij de hulp van vele mensen, waarvan ik
een aantal persoonlijk wil bedanken.
Als eerste natuurlijk Roger Sheldon, die mij de mogelijkheid heeft geboden het
promotieonderzoek te doen. Ik heb de afgelopen vier jaar met plezier gewerkt
onder jouw begeleiding. Wat ik zeker geleerd heb is kritisch naar m’n eigen werk
te kijken alsof ik nooit gezien had en het ter beoordeling voor me lag. Zoals je
vaak tijdens werkbesprekingen zei: “The proof of the pudding is in the eating”.
Op zoek naar nieuwe ideeën hoopte ik altijd “the best thing since sliced bread”
te vinden.
Fred van Rantwijk, verantwoordelijk voor de dagelijkse begeleiding, heeft een
grote invloed gehad op eigenlijk alle aspecten van het onderzoek. Niet alleen
stond je altijd klaar voor vragen (waarbij je altijd instant wist waarover het
ging, wat een geheugen!), ook kwam je altijd weer met nieuwe ideeën op de
proppen. Verder denk ik graag terug aan alle buitenlandse avonturen.
De heren van HSL, ook ik moet toegeven dat pret omgekeerd evenredig is met
de hoeveelheid daglicht. De sporadische behoefte aan zonlicht heeft zich bij
jullie vooral vertaald in bijzondere LAT relaties. Luuk “The Force”, schei-kundig,
ook in de scheikunde, Delft zou zonder jou niet hetzelfde geweest zijn. Ik
herinner me vele kilometers in de Starlet, de verhuizing in 37 minuten, de
ontbijtkoffie in de “casa Parallelia“ en Gezellig Samen Mobiel. “Hacking”, wat
entropie in de thermodynamica is, ben jij op het lab en heel ver daarbuiten. De
nachtelijke biertjes op het strand van La Grande Motte, alle bezwaarschriften
130
die de centen weer in de juiste portemonnee deden belanden, het afblaffen van
Luuk’s ex-huisbaas en je experimenteerlust kleurden het leven in Delft.
Michel Verhoef, wat kan ik zeggen? Het liefst natuurlijk: dinsdag, klokslag
negen uur, de Wijnhaven. Hadden we maar meteen aandelen in die tent gekocht.
Als collega was je toch vooral collega-brouwer.
Rute, ondanks dat HSL LSL is geworden, is er niets veranderd. Ik hoop dat
ik nog eens net zo snel Portugees kan leren als jij Nederlands geleerd hebt.
Esther Geenen heeft als HLO-stagiaire een aantal enzym substraten
gesynthetiseerd. Dit was vaak vrij lastig, maar je inzet was groot. Goede blonde
herinneringen heb ik nog steeds aan het concert van K’s Choice.
Paloma, Gerd-Jan (“louder”), Göran (we moeten weer eens brouwen), Mike,
Gerard, Michiel Van Vliet, Margreth, Arné, Martin “Sjaak” (wil je me straffen?),
Annemieke, Emrin, Fred van de Velde en nog vele anderen wil ik bedanken voor
alle goede contacten die ik altijd met ze heb.
Verder wil ik bedanken Adrie Knol voor de massaspectrometrie, Anton van
Estrik, Anton Sinnema en Joop Peters voor de altijd snelle en kundige acquisitie
van NMR spectra en de hulp bij de vragen die dit opriep; Bert van der Hulst en
Chris van Drongelen voor het bestellen van chemicaliën en andere
laboratoriumbenodigdheden en de gezellige koffiepauzes tijdens de practica;
Ernst Wurz voor z’n twee rechterhanden, Leen Maat, Herman van Bekkum; en
Jan Baas voor alle zorg over de UV (vooral toen we nog OS-2 WARP hadden!).
Mieke van der Kooij bedankt voor de administratieve ondersteuning en vele
andere zaken.
Many thanks to Dr. Tischer and Peter Rasor from Roche Diagnostics,
Penzberg, Germany for the generous gifts of enzymes. I think it’s obvious this
thesis would not have been the same without good old S. carnosus aldolase. Zum
wohl!
Via het IOP ben ik in aanraking gekomen met vele mensen. De leden van de
131
begeleidingscommissie ben ik zeer dankbaar voor het bijwonen van de
halfjaarlijkse vergaderingen, tijdens welke het onderzoek op de rails werd
gehouden en geen station overgeslagen werd. De commissieleden waren: Tom
Kieboom (DSM-Gist), Wolfgang Fitz (ICI), dhr. Van Scharrenburg (Solvay
Duphar), Gerard Kramer (Unilever Research Vlaardingen), dhr. Ter Burg
(Tastemaker ). dhr. J. Frankena (Solvay Pharmaceuticals), dhr. Weenen (ICI),
Prof. dr. H.E. Schoemaker (DSM Research), Erik De Vroom (DSM-Gist), dhr.
Winkel (ICI) en dhr. Dijkstra (Solvay Pharmaceuticals).
Niels Ouwerkerk, bedankt voor jouw DERA kweek en wie had ooit gedacht
dat we nog collega’s zouden worden!
Georg Sprenger and Ulrich Schörken, Institut für Biotechnologie 1,
Forschungszentrum Jülich GmbH, Germany; thanks for the crude extract of
transketolase and the assistance with the purification and the initial
crystallizations. Furthermore, thanks for the pleasant company, also during the
beers.
Simon de Vries van Enzymologie ben ik dankbaar voor de hulp bij het
kweken en zuiveren van “DERA”.
Pa, jammer dat je mijn promotie niet meer hebt mogen meemaken, maar ik weet
zeker dat je trots op me was.
Karel en Nel bedankt voor het bewaren van de vrede als ik thuis zat te werken
en niet voor de witte haren op m’n zwarte truien en de vlekken in de
vloerbedekking.
En natuurlijk als laatste, Martine “Tien” bedankt, voor meer dan teveel om op te
noemen en je Shift+F7 tip.
132
Publications
1. R. Schoevaart, F. van Rantwijk and R.A. Sheldon, “Class I fructose-1,6-
bisphosphate aldolases as catalysts for asymmetric aldol reactions”,
Tetrahedron Asymmetry 1999, 10 (4), 705-711
2. R. Schoevaart, F. van Rantwijk and R.A. Sheldon, “Carbohydrates from
glycerol: an enzymatic four-step one-pot synthesis”,
Chemical Communications, 1999, 24, 2465-2466
3. R. Schoevaart, F. van Rantwijk and R.A. Sheldon, “A four-step enzymatic
cascade for the one-pot synthesis of non-natural carbohydrates from glycerol”,
submitted.
4. R. Schoevaart, F. van Rantwijk and R.A. Sheldon, “The stereochemistry of non-
natural aldol reactions catalysed by DHAP aldolases”, submitted.
5. R. Schoevaart, F. van Rantwijk and R.A. Sheldon, “Facile enzymatic aldol
reactions with dihydroxyacetone arsenate”, submitted.
133
Curriculum vitae
Rob Schoevaart werd op 28 januari 1969 geboren te Coevorden. In 1988 behaalde
hij zijn VWO diploma aan de Rijksscholengemeenschap in Coevorden. In hetzelfde
jaar begon hij met de studie scheikunde aan de Katholieke Universiteit Nijmegen.
Daar werd Organische Chemie onder leiding van Prof. dr. B. Zwanenburg en dr.
G.J.F. Chittenden als hoofdvak gekozen en Microbiologie onder leiding van Prof. dr.
ir. G.D. Vogels als bijvak. In 1994 werd het doctoraal diploma behaald en in 1995
werd begonnen met het promotieonderzoek bij de vakgroep Organische Chemie en
Katalyse aan de Technische Universiteit Delft onder begeleiding van Prof. Dr. R.A.
Sheldon. Momenteel is de auteur werkzaam als post-doc bij het Leids Instituut voor
Chemisch onderzoek onder leiding van Prof. Dr. A. Kieboom.
134
Index
α-hydroxyaldehydes .............................97, 113
α-hydroxybutanal........................................ 105 B
β-hydroxyketones ....................................... 113
α-hydroxypentanal...................................... 106 biocatalysis..................................................... 2
β-hydroxypyruvate ........................................ 97
C
2
carbon dioxide
2-deoxyribose-5-phosphate..........................12 release of ................................................. 13
cascade catalysis ......................................... 51
3 catalase ........................................................ 54
catalytic antibodies......................................... 5
3R,4R............................................................37 chemical phosphorylation............................. 65
3R,4S ........................................................8, 37 chiral GC....................................................... 36
3S,4R ............................................................37 Claisen-Schmidt reaction ............................... 5
3S,4S ............................................................37 Class I............................................................. 7
contents .......................................................... 5
A cosolvents..................................................... 80
curriculum vitae .......................................... 134
acceptor substrate specificity........................23
alcohol dehydrogenase D
assay of aldehydes................................... 88
aldehydes dankwoord .................................................. 130
hydrophobic.............................................. 80 dephosphorylation........................................ 64
aldol adduct DERA.................................................... 12, 111
hydrolysis of ............................................. 59 acceptor substrates................................ 113
aldol reaction activity in cosolvents .............................. 119
general........................................................4 stability in cosolvents ............................. 117
aldol reactions D-glyceraldehyde-3-phosphate
enzymatic ................................................... 3 dehydrogenase ........................................ 24
aldolase...........................................................5 DHA
2-deoxyribose-5-phosphate aldolase12, 111 assay of.................................................... 89
Class I.........................................................7 DHAAs
Class II........................................................7 acceptor specificity................................... 83
dihydroxyacetone phosphate dependent conversion by FruA .................................. 84
aldolases ................................................ 8 DHAP.................................... 22, 25, 52, 54, 65
from rabbit muscle...................................... 8 analogues................................................. 73
from Staphylococcus aureus .................... 22 assay of.................................................... 29
from Staphylococcus carnosus .................. 8 In situ generation of.................................. 51
fructose-1,6-bisphosphate aldolase ......... 21 dihydroxyacetone arsenate.......................... 76
glycine-dependent.................................... 11 dihydroxyacetone monophosphate .............. 22
L-fuculose-1-phosphate aldolase............... 8 dihydroxyacetone phosphate
L-rhamnulose-1-phosphate aldolase.......... 8 synthesis of .............................................. 52
phosphoenolpyruvate-dependent............. 10 DNA synthesis.............................................. 12
pyruvate-dependent ................................. 10 donor .............................................................. 6
RAMA .........................................................8 D-ribose-5-phosphate................................... 96
specific activities....................................... 25 D-xylulose-5-phosphate ......................... 13, 96
stability................................................ 27, 81
tagatose-1,6-bisphosphate aldolase .......... 8 E
threonine aldolases .................................. 12
alsolase enolates .......................................................... 5
FruA activity assay ................................... 29 entropie....................................................... 130
arsenate enzyme recovery .......................................... 30
esters........................................................76 epimerisation ................................................ 37
erythrose-4-phosphate ................................. 96
Eupergit-C .................................... 11, 100, 116
135
F One pot preparation of carbohydrates ......... 56
one-pot ......................................................... 55
Force...........................................................130 one-pot reaction
integral in situ ........................................... 55
G sequential................................................. 60
general introduction ........................................ 1 P

glutaraldehyde ............................................ 101
glyceraldehyde-3-phosphate .................. 8, 112 pentose phosphate pathway......................... 96
glycerol phosphate oxidase..........................54 phosphatase
glycolaldehyde.............................................. 23 from Aspergillus ficuum............................ 56
glycolysis ...................................................... 96 from bovine intestinal mucosa ................. 56
phospholipase C ...................................... 54
H phytase..................................................... 56
wheat germ acid phosphatase................. 27
hydroxyacetone thiosulfate...........................74 phosphoenolpyruvate................................... 10
Hydroxypyruvate........................................... 13 POCl3 ............................................................ 52
propanal.................................................. 23, 24
I publications................................................. 133
pyrophosphate.............................................. 55
Imine intermediate .......................................... 7
in situ oxidation ............................................. 55 R
inhibitors .......................................................74
initial rates..................................................... 25 reaction rates................................................ 26
IOP Catalysis .................................................. 2 retro aldol reactions...................................... 39
retro-aldol reaction.................................. 22, 24
K
S
Kinetiek .........................................................21 samenvatting .............................................. 127
Schiff’s base ................................................... 7
L self aldol reaction........................................ 114
sialic acid aldolase........................................ 10
L-FcuA ...................................................... 8, 22
L-lactaldehyde .............................................. 23 sialic acids .................................................... 10
L-RhuA...................................................... 8, 22 stereoselectivity
lyases..............................................................5 effect of arsenate ..................................... 85
enzymatic assay....................................... 39
M NMR analysis ........................................... 42
structural assignments ................................. 37
ManNAc ........................................................10 summary..................................................... 124
N T
N-acetyl-D-mannosamine.............................10 thiamine pyrophosphate......................... 14, 96

N-acetylmannosamine.................................. 10 transaldolase ................................................ 96
N-acetylneuraminic acid ...............................10 transferase.................................................... 13
NADH....................................................24, 104 transketolase
Neu5Ac synthetase....................................... 10 from E. coli ............................................... 97
NMR from spinach leaves ................................. 96
31P NMR .................................................. 63 transketolase .......................................... 13, 95
O cross-linked enzyme crystals ................. 100
from yeast ................................................ 96
objectives and justification............................15 versus FruA ............................................ 100
136

Applications of Aldolases in Organic Synthesis - Schoevaart Thesis 2000

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Applications of Aldolases in Organic Synthesis - Schoevaart Thesis 2000

Enviado por

Direitos autorais:

Formatos disponíveis

Applications of aldolases

Sun set in the Mediterranean sea

Phenomenologie des Geistes, Hegel

Toepassingen van aldolasen in organische synthese

ter verkrijging van de graad van doctor

aan de Technische Unversiteit Delft

op gezag van de Rector Magnificus, Prof. ir. K.F. Wakker

ingevolge het besluit van het College voor Promoties

in het openbaar te verdedigen op

dinsdag 23 mei 2000 om 13:30

Willem Robert Klaas Schoevaart

geboren op 28 januari 1969 te Coevorden

Rector Magnificus Technische Universiteit Delft

Prof. dr. A. Bruggink DSM Geleen / KU Nijmegen

Prof.dr.ir. J.J. Heijnen Technische Universiteit Delft

Prof. dr. A.P.G. Kieboom Universiteit Leiden / DSM-Gist

Prof. dr. B. Zwanenburg KU Nijmegen

Dr. W. Fitz ICI

Dr.ir. F. van Rantwijk Technische Universiteit Delft

totstandkoming van dit proefschrift bijgedragen.

Copyright© 2000 W.R.K. Schoevaart

Printed by Print Partners Ipskamp

Catalysis (IOP Katalyse, no. IKA94010) sponsored by the Netherlands Ministery of

aan mijn ouders

The presence of multiple functional groups of similar reactivity in molecules poses

problems for classical organic synthetic methods. Synthesizing and modifying

carbohydrates for example, is accomplished only by applying complex synthetic

stereoselectivity. A large proportion of the reactions in the armamentarium of organic

chemists is incompatible with aqueous systems and some modifications are

necessary simply to achieve solubility in non-aqueous solvents. The application of

enzymes1 as catalysts in organic synthesis has succeeded in solving synthetic

problems and developments in enzymology have effectively refuted the perception

require expensive cofactors.

demanding objective of providing an essential creative step on the road from

fundamental science to novel, applicable technology. This implied research based on

a vision of its potential applications in an industrial context2.

total volume of chemicals. Hence a solid knowledge base on catalysis is of strategic

chemical theme of a research programme known as the Innovation Oriented

Research Programme (IOP) Catalysis: precision in chemical conversion. This

of undesired by-products and waste. Catalysis is applied on a large scale in

in the manufacture of fine chemicals where classical multi-step chemical conversions

large amounts of waste.

Biocatalysis is of particular interest for the manufacture of the fine chemicals,

promise in such enantioselective synthesis since they compete with chemical

catalysis on e.g. introduction of chirality, selective modification or resolution of

retro aldol cleavage of abundant cinnamaldehyde is a “grey–zone chemistry” route to

so-called natural benzaldehyde3. Inexpensive, readily available and renewable

Enzymatic aldol reactions

challenging subjects in synthetic organic chemistry and, without a doubt, in

catalysed ones a characteristic of the enzymatic reaction is particularly outstanding:

enantioselectivity. Whereas the simple base catalysed reaction is only

enantioselective with (stoichiometric amounts of) chiral starting materials, application

of enzymes can lead to enantiomerically pure products from completely non-chiral

starting materials. For non-enzymatic reactions regioselectivity is more controlable,

might be problematic, making protection and deprotection steps necessary.

Enzymes are usually completely regioselective, reducing protective group chemistry

to a minimum, making biocatalysis an important instrument in C-C bond forming

Aldolases have good activity towards hydroxylated compounds which is not

suprising considering their role in the carbohydrate metabolism. This property is

enzymes can consequently be conducted under mild conditions (water, pH 7) and do

biodegradability makes them an excellent environmentally acceptable option as well.

The aldol reaction

Figure 1. The aldol reaction