A mutant D-fructose-6-phosphate aldolase (Ala129Ser) with improved affinity towards dihydroxyacetone for the synthesis of polyhydroxylated compounds

Article abstract of DOI:10.1002/adsc.200900772

A mutant of D-fructose-6-phosphate aldo-lase (FSA) of Escherichia coli, FSA A129S, with im-proved catalytic efficiency towards dihydroxyacetone (DHA), the donor substrate in aldol addition reac-tions, was explored for synthetic applications. The K_cat/K_M value for DHA was 17-fold higher with FSA A129S than that with FSA wild type (FSA wt). On the other hand, for hydroxyacetone as donor sub-strate FSA A129S was found to be 3.5-fold less effi-cient than FSA wt. Furthermore, FSA A129S also ac-cepted glycolaldehyde (GA) as donor substrate with 3.3-fold lower affinity than FSA wt. This differential selectivity of both FSA wt and FSA A129S for GA makes them complementary biocatalysts allowing a control over donor and acceptor roles, which is par-ticularly useful in carboligation multi-step cascade synthesis of polyhydroxylated complex compounds. Production of the mutant protein was also improved for its convenient use in synthesis. Several carbohy-drates and nitrocyclitols were efficiently prepared, demonstrating the versatile potential of FSA A129S as biocatalyst in organic synthesis.

Full text of DOI:10.1002/adsc.200900772

FULL PAPERS
DOI: 10.1002/adsc.200900772
A Mutant d-Fructose-6-Phosphate Aldolase (Ala129Ser) with
Improved Affinity towards Dihydroxyacetone for the Synthesis
of Polyhydroxylated Compounds
Josꢀ A. Castillo,a,b Christine Guꢀrard-Hꢀlaine,a,b Mariana Gutiꢀrrez,c
Xavier Garrabou,c Martine Sancelme,a,b Melanie Schꢁrmann,e Tomoyuki Inoue,d
Virgil Hꢀlaine,a,b Franck Charmantray,a,b Thierry Gefflaut,a,b Laurence Hecquet,a,b
Jesffls Joglar,c Pere Clapꢀs,c Georg A. Sprenger,d,e,* and Marielle Lemairea,b,*
a
Laboratoire SEESIB, Clermont Universitꢀ, Universitꢀ Blaise Pascal, 24 avenue des Landais, 63177 Aubi›re cedex, France
Fax : (+33)-(0)4-7340-7717; phone: (+33)-(0)4-7340-7584; e-mail: Marielle.Lemaire@univ-bpclermont.fr
CNRS, UMR 6504, 24 avenue des Landais, 63177 Aubi›re, France
Biotransformation and Bioactive Molecules Group, Instituto de Química Avanzada de CataluÇa – CSIC, Jordi Girona
b
c
18-26, 08034 Barcelona, Spain
Present address: Institute of Microbiology, Universität Stuttgart, Allmandring 31, 70550 Stuttgart, Germany
d
Fax : (+49)-(0)711-685-65725; phone: (+49)-(0)711-685-65488; e-mail: g.sprenger@imb.uni-stuttgart.de
Institute for Biotechnology 1, Forschungszentrum Jꢁlich, 52405 Jꢁlich, Germany
e
Received: November 6, 2009; Revised: March 12, 2010; Published online: April 7, 2010
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200900772.
Abstract: A mutant of d-fructose-6-phosphate aldo- makes them complementary biocatalysts allowing a
lase (FSA) of Escherichia coli, FSA A129S, with im- control over donor and acceptor roles, which is par-
proved catalytic efficiency towards dihydroxyacetone ticularly useful in carboligation multi-step cascade
(DHA), the donor substrate in aldol addition reac- synthesis of polyhydroxylated complex compounds.
tions, was explored for synthetic applications. The Production of the mutant protein was also improved
kcat/KM value for DHA was 17-fold higher with FSA for its convenient use in synthesis. Several carbohy-
A129S than that with FSA wild type (FSA wt). On drates and nitrocyclitols were efficiently prepared,
the other hand, for hydroxyacetone as donor sub- demonstrating the versatile potential of FSA A129S
strate FSA A129S was found to be 3.5-fold less effi- as biocatalyst in organic synthesis.
cient than FSA wt. Furthermore, FSA A129S also ac-
cepted glycolaldehyde (GA) as donor substrate with Keywords: Ala129Ser mutant; carbohydrates; d-fruc-
3.3-fold lower affinity than FSA wt. This differential tose-6-phosphate aldolase; dihydroxyacetone; nitro-
selectivity of both FSA wt and FSA A129S for GA cyclitols
Introduction
donor to furnish aldose-type products, opening new
strategies for biocatalytic aldol addition reactions.[7]
Wild-type d-fructose-6-phosphate aldolase (FSA wt, Therefore it can be considered both as a ketose- and
gene fsaA) from Escherichia coli, first described in an aldose-aldolase. Mechanistically, it belongs to the
2001,[1,2] has attracted considerable attention due to class I aldolase family (i.e., no metal cofactor re-
its remarkable catalytic potential in the asymmetric quired), in which an enamine (i.e., nucleophile) is
synthesis of polyhydroxylated compounds.[3–7] This formed by reaction of the donor with a lysine residue
enzyme was found to catalyze the 3S,4R (syn) stereo- at the active site.[8] When compared with the well-
selective reversible aldol addition of dihydroxyace- known DHAP-dependent aldolases, the hallmark of
tone (DHA) on d-glyceraldehyde 3-phosphate (d- FSA is its ability to catalyze direct aldol additions of
G3P) to afford d-fructose 6-phosphate (d-F6P).[1] Fur- unphosphorylated DHA skipping a tedious, time-con-
thermore, it was later discovered by some of us that suming and expensive dihydroxyacetone phosphate
FSA wt is able to accept glycolaldehyde (GA) as (DHAP) synthesis. Moreover, it is the first uncovered
Adv. Synth. Catal. 2010, 352, 1039 – 1046
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1039

Josꢀ A. Castillo et al.
FULL PAPERS
the donor substrate that provided the best reaction
rates and conversions to the aldol adducts.[3,5,6] There-
fore, aldehydes that are tolerated by FSA wt using
HA as donor may not react using DHA.[5–7] This low
affinity towards DHA is also the reason why the self-
aldol product of GA (i.e., d-threose, 5) was the only
product formed when DHA and GA were mixed in
the presence of FSA wt.[7] Since this peculiar reactivi-
ty map often limits some synthetic applicability, it was
regarded as interesting and significant to explore
novel FSA variants with distinct substrate specificity,
which complement FSA wt. In the course of investiga-
tions on the FSA function and mechanism, Schꢁr-
mann and Sprenger envisaged a redesign of the FSA
active site with the aim to change the native FSA al-
Scheme 1. Examples of polyhydroxylated compounds
chemo-enzymatically prepared with FSA wt.[4–7]
aldolase that shows significant activity towards several dolase activity into transaldolase activity.[15,16] Accord-
donor substrates, for example, hydroxyacetone (HA), ing to the crystallographic structure of FSA, the resi-
1-hydroxy-2-butanone (HB) and GA.[9] Recently, dues Leu107 and Ala129 were identified as putatively
some of us reported on a transaldolase B variant involved in the binding site of the C-1 hydroxy group
(TalB F178Y)[10] that is also able to use DHA as of the donor substrate.[2] Hence, three mutants of
donor in aldol reactions and a double mutant (TalB FSA wt were constructed by site-directed mutagene-
F178Y/R181E)[11] which exhibited improved affinity sis: two single A129S and L107N and a double
towards non-phosphorylated acceptor aldehydes such mutant L107N/A129S exchange.[15,16] Interestingly, al-
as d- and l-glyceraldehyde and GA. DHAP-depen- though no transaldolase activity was observed on any
dent rhamnulose aldolase (RhuA) has been described of the mutants, the single A129S one gave a strikingly
to accept DHA in the presence of borate buffer.[12] In improved kcat/KM towards DHA, d-G3P and d-F6P,
addition, to alter RhuA donor substrate specificity and potential complementary synthetic abilities to
from DHAP to DHA, Sugiyama et al. reported the those of FSA wt. It is likely that the serine residue
development of an in vivo selection system for the di- allows hydrogen bonding that stabilizes the DHA
rected evolution of this aldolase.[13]
Due to its broad substrate tolerance, FSA wt has bilization of the Schiff-base intermediate.[2,15,16]
been efficiently used to synthesize different poly- Regarding the interesting catalytic properties of
donor substrate and also might be involved in the sta-
hydroxylated compounds such as iminocyclitols and FSA A129S, a clear picture of its synthetic capabilities
carbohydrates (Scheme 1). Thus, the syntheses of d- has to emerge. To shed light on this point, in this
fagomine (1), deoxynojirimycin (DNJ, 2), 1,4,5-tri- work we were prompted to study this FSA A129S
deoxy-1,4-imino-d-arabinitol (5-DDAB, 3) and ana- mutant for the synthesis of polyhydroxylated com-
logues were described by Clapꢀs and co-workers.[4,6] pounds. Furthermore, kinetic parameters with differ-
Other illustrative iminocyclitols such as compound 4 ent substrates were measured to define precisely the
were synthesized by Wong and co-workers.[5] In addi- substrate spectrum of this biocatalyst.
tion, the use of GA as donor provided the first biocat-
alytic strategy of self- and cross-aldol additions of GA
to prepare d-threose (5) and 5-deoxy-l-xylose (6).[7]
All along, these synthetic studies on FSA wt have
Results and Discussion
revealed different attractive features. Recombinant Recently, we reported the kinetic parameters of FSA
FSA wt can be produced in high amount since it can wt for the retroaldol reaction of d-F6P and for aldol
be efficiently expressed in E. coli. Moreover, the al- reactions using DHA and HA as donors and GA as
dolase possesses an unusual thermostability, likely donor and acceptor.[7] To exploit the catalytic proper-
due to its tight packing in a homodecamer form.[1–2,14] ties of FSA A129S, the kinetic parameters for the
Hence, FSA wt can be used with acceptable purity above-mentioned reactions, particularly for DHA and
after a simple heat treatment of a crude extract[1,4] or HA as donors and GA as donor and acceptor, were
even in whole cells.[5] Besides, the enzyme tolerates also determined. In addition, we have evaluated the
organic co-solvents such as 10–20% dimethylforma- kinetic parameters for formaldehyde and 4-nitrobuta-
mide and reactions can be performed at room temper- nal, as examples of acceptor substrates to evaluate
ature or at lower temperature.[4]
As judged by the Vmax/KM values, FSA wt functions affinity (Table 1).
the potential effect of the mutation on the acceptor
~40-fold and ~20-fold better with HA and GA, re-
As mentioned in the introduction for FSA wt, HA
spectively, than with DHA.[7] This explains why HA is and GA (Table 1, entries 2 and 3) were shown to be
1040
asc.wiley-vch.de
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1039 – 1046

A Mutant d-Fructose-6-Phosphate Aldolase (Ala129Ser) with Improved Affinity towards Dihydroxyacetone
Table 1. Steady-state kinetic parameters of aldol and retroaldol reactions catalyzed by FSA wt and FSA A129S.
FSA wt
FSA A129S
kcat[a,c] [sÀ1]
Entry
Substrate
KM [mM]
kcat[a,b] [sÀ1]
kcat/KM [sÀ1 mMÀ1]
KM [mM]
kcat/KM [sÀ1 mMÀ1]
1
2
3
4
5
6
7
DHA
HA
32Æ2[d]
17.4Æ0.5[d]
0.2Æ0.05[d]
62.8Æ5.5[d]
254Æ29
ND[f]
116[d]
2527[d]
16.5[d]
16.5[d]
118
4[d]
11Æ1
22Æ3
0.18Æ0.02
12Æ5
149Æ63
77Æ6
0.9
760
899
4.4
69
41
25
0.4
1
145[d]
83[d]
0.3[d]
0.5
GA donor
GA acceptor
4-Nitrobutanal
Formaldehyde
G3P[e]
4.4
163
27.7
380
ND[f]
173
ND[f]
216
0.4
422
0.8
[a]
kcat =(Vmax/Vmax D-F6P)”kcat D-F6P; Vmax D-F6P is the value measured for cleavage of d-fructose-6-phosphate, see more details
in Supporting Information.
[b]
[c]
[d]
[e]
[f]
k
k
cat D-F6P =27 sÀ1.[16]
cat D-F6P =84 sÀ1.[16]
Values from Garrabou et al.[7]
Kinetic parameters for d,l-glyceraldehyde 3-phosphate were reported by Sprenger et al.[16]
ND: not done.
largely better donors than DHA (entry 1). FSA wt Synthetic Examples using FSA A129S as Biocatalyst:
has a KM value for DHA of 32 mM; the A129S var- Nitrocyclitols
iant has an improved KM value of 11 mM. Therefore,
as judged by kcat/KM, the mutation induced a decrease 4-Nitrobutanal has been widely used in our laboratory
in the catalytic selectivity towards HA and GA to synthesize nitrocyclitols by a chemo-enzymatic
donors and an improvement towards DHA. In terms methodology using DHAP as donor substrate.[17–19]
of kcat/KM, it is noticeable that the activity towards The one-pot cascade process for nitrocyclitols prepa-
HA and GA was reduced by factors of 3.5 and 3.3, re- ration involved the formation of two carbon-carbon
spectively (entries 2 and 3). Interestingly, FSA A129S bonds. First, the aldol reaction was catalyzed by d-
functions 17-fold better than FSA wt with DHA. On fructose 1,6-bisphosphate aldolase (FruA from rabbit
the other hand, the mutation induced no significant muscle, RAMA) in the presence of DHAP and this
effect on the acceptors since the catalytic activity was was followed by an intramolecular Henry reaction.
found to be in the same range with 4-nitrobutanal as This strategy furnished nitrocyclitols in a highly ste-
well as with GA (entries 4 and 5). For both enzymes, reoselective way. For example, nitrocyclitol (7) was
GA, 4-nitrobutanal and formaldehyde (entries 4, 5 obtained (see Table 2, entry 1) as a major diastereo-
and 6) were poor acceptors compared to d,l-G3P. Re- isomer in 60% isolated yield.[17]
markably, d,l-G3P is clearly the best acceptor for
For the sake of comparison, in the present work we
both aldolases. In spite of the low kcat/KM good con- performed the two-step cascade carbon-carbon form-
versions and isolated yields were achieved with both ing process by the aldol addition of DHA to 4-nitro-
4-nitrobutanal and formaldehyde (vide infra).
butanal catalyzed by FSA A129S. Interestingly, in
These results point out that the mutant FSA A129S spite of the fact that 4-nitrobutanal was not a good
is an improved biocatalyst for the aldol additions of substrate for FSA A129S (Table 1, entry 5), the nitro-
DHA to aldehydes. This differential reactivity of FSA cyclitol 7 was isolated in 73% isolated yield also as a
A129S as compared with FSA wt is of paramount im- major diastereoisomer (Table 2, entry 1). 4-Nitrobuta-
portance since it allows one to modulate donor and nal was also submitted to aldol reaction with HA as
acceptor roles in cascade self- and cross-aldol addition donor (Table 2, entry 2). In this case, the enzymatic
reactions, particularly when using GA.
reaction was fully stereoselective, while the intramo-
The production of enzymes for the development of lecular Henry reaction was not. This lower diastereo-
biosynthetic methodologies is of utmost importance. selectivity was probably due to the absence of the C-1
In this connection, the production of FSA A129S was hydroxy group, which might direct the approach of
improved using a classical plasmid vector carrying the the nucleophilic nitro a-carbon to the carbonyl group.
fsaA129S gene with a kanamycin resistance cassette The whole process gave a mixture of diastereoisomers
to have a continuous supply of the biocatalyst in large in 87% isolated yield, which were separated by classi-
quantities and at a reasonable cost (for details see cal chromatography furnishing nitrocyclitols 8, 9, 10
Supporting Information).
and 11 in 6%, 4%, 55% and 22% isolated yields, re-
spectively. The stereochemistries of these new nitrocy-
clitols were unequivocally assigned, based on the 3S-
Adv. Synth. Catal. 2010, 352, 1039 – 1046
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1041

Josꢀ A. Castillo et al.
FULL PAPERS
Table 2. Aldol additions catalyzed by FSA A129S.
Entry Acceptor
Donor
Product
Isolated yield[a] and/or conversion[b]
1
DHA (7)
73%[a]
2
3
HA
(8)–(11)
(8) 6%;[a] (9) 4%;[a] (10) 55%;[a] (11) 22%[a]
DHA (12)
52%[a]
4
5
6
7
8
9
DHA (13)
DHA (14)
68%[a]
80%[b]; 42%[a]
HA
(15)
70%[b]
DHA (16)
81%[b]; 67%[a]
86%[b]; 68%[a]
38%[b,c]
HA
GA
(17)
(18)
[a]
Isolated yield.
[b]
Conversion of the aldehyde to product was determined by HPLC (for details see Supporting Information).
d-Threose formation (20% by HPLC) was observed arising from self-aldol addition of GA.
[c]
stereoselectivity of the aldolase and on high-field
NMR studies (coupling constants and NOESY experi- A129S gave generally higher isolated yields than
ments). those with FruA and the synthetic process were con-
Remarkably, the aldol additions catalyzed by FSA
An enantiomerically pure 2-hydroxy-4-nitrobuta- siderably simplified since a number of steps required
nal[19] was submitted to aldolization with DHA for DHAP preparation and phosphate group removal
(Table 2, entry 3). In this case, the two-step cascade of the resulting adduct were avoided.
chemo-enzymatic process was fully stereoselective
and nitrocyclitol 12 was isolated in 52% yield.
1042
asc.wiley-vch.de
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1039 – 1046

A Mutant d-Fructose-6-Phosphate Aldolase (Ala129Ser) with Improved Affinity towards Dihydroxyacetone
Carbohydrates
of complex sugars and analogues such as d-xylulose
and d-fructose, not accessible via FSA wt catalysis,
As a first example of carbohydrates synthesis using was catalyzed by FSA A129S. Both aldolases will
FSA A129S, l-erythrulose (13) was synthesized from expand the scope of carbohydrates and analogues
DHA and formaldehyde as starting materials. This that can be prepared enzymatically and will open
aldol reaction afforded l-erythrulose (13) in 68% iso- novel synthetic opportunities for the carboligation
lated yield as a unique enantiomer.
cascade synthesis of polyhydroxylated compounds.
Interestingly, the aldol addition of DHA to GA Most importantly, the enzyme has shown a high dia-
proceeded in 80% aldehyde conversion to d-xylulose stereoselectivity affording always the syn (3S,4R) con-
(14) (42% isolated yield not optimized, Table 2 figuration. The use of an expression vector with an
entry 5). This was the most striking result since, using added kanamycin resistance cassette led to a remark-
FSA wt, d-threose was exclusively formed, arising able improvement of the enzyme yields in comparison
from the self-aldol addition reaction of GA.[6,7] This to the system with ampicillin resistance alone. Thus,
differential reactivity was a consequence of the modi- FSA A129S production was improved for its use in
fied catalytic properties of FSA A129S (Table 1, en- organic synthesis at multi-gram scale.
tries 1 and 3). Moreover, the aldol addition of DHA
to d-glyceraldehyde (d-GAL) furnished d-fructose
(16) in 67% isolated yield (Table 2, entry 7), while Experimental Section
FSA wt was nearly inactive to catalyze this aldol reac-
tion.[3] This might be due to better catalytic affinity of For Kinetic Measurements
FSA A129S for DHA which could afford a protection
of the donor active site against aldehyde inhibition.
One unit of FSA cleaves 1 mmol of d-fructose 6-phosphate
per min to afford d-glyceraldehyde 3-phosphate and dihy-
droxyacetone at pH 7.5 (TEA 50 mM buffer) and 258C.
This phenomenon has been previously demonstrated
for class I aldolases.[20]
Despite the lower catalytic activity of FSA A129S
towards HA as compared with FSA wt, the condensa-
For 4-Nitrobutanal as Acceptor with FSA wt
Aldol addition reaction of DHA to 4-nitrobutanal: Measure-
ment of DHA was carried out with glycerol dehydrogenase
(GDH) from Cellulomonas sp. To a solution containing
DHA (100 mM) and FSA wt powder (1.5 mg to 3 mg, 1.2 U
to 2.4 U) in 100 mM tris-ethanolamine (TEA) buffer pH 7.5
at 258C, 4-nitrobutanal (in concentration varying from 40 to
200 mM) was added. The final volume was 1 mL. Then, ali-
quots (20 mL) were withdrawn at different times and diluted
in water (380 mL). Samples (20 mL) were then dissolved in a
solution (180 mL) composed of GDH (20 U) and NADH
(0.7 mM) in TEA buffer. The consumption of NADH was
monitored spectrophotometrically at 340 nm and 258C. One
mmol of NADH oxidized was equivalent to 1 mmol of
DHA.
tion of HA to GA and d-GAL proceeded with good
conversion to the corresponding 1-deoxy-d-xylulose
(15) and 1-deoxy-d-fructose (17) (Table 2, entries 6
and 8).[6,7] In aldol additions involving GA (Table 2,
entries 5, 6 and 9), the side conversion to self-aldol
addition product, d-threose, was moderate (20%) as a
consequence of the reduced affinity of FSA A129S to
GA as donor and acceptor. The enzymatic condensa-
tion of DHA on d-threose was attempted, unfortu-
nately in a one-pot/two-step process where d-threose
was accumulated and then followed by addition of
DHA, the corresponding d-ido-hept-2-ulose was
never observed.
Finally, as ascertained by NMR spectroscopy and
within the limits of detection, FSA A129S was fully
stereoselective yielding the corresponding syn (3S,4R)
aldol adducts in all cases. As illustrated, the known l-
erythrulose, d-xylulose and d-fructose were exclusive-
ly formed.
Km and Vmax for 4-Nitrobutanal and Formaldehyde as
Acceptor with FSA A129S
Aldol addition reaction of DHA to 4-nitrobutanal or formal-
dehyde: Measurement of DHA was carried out with glycerol
dehydrogenase (GDH) from Cellulomonas sp. To a solution
containing DHA (30 mM) and FSA A129S powder (0.2 mg,
0.72 U for nitrobutanal and 0.4 mg to 1.3 mg, 1.4 U to 4.6 U
for formaldehyde) in 100 mM tris-ethanolamine (TEA)
buffer pH 7.5 at 258C, 4-nitrobutanal (in concentration vary-
ing from 50 to 120 mM) or formaldehyde (in concentration
varying from 20 to 200 mM) was added. The final volume
was 1 mL. Then, aliquots (20 mL) were withdrawn at differ-
ent times and diluted in water (80 mL). Samples (20 mL)
were then dissolved in a solution (980 mL) composed of
GDH (20 U) and NADH (0.2 mM) in TEA buffer. The con-
sumption of NADH was monitored spectrophotometrically
at 340 nm and 258C. One mmol of NADH oxidized was
Conclusions
From the kinetic data obtained with different sub-
strates, the mutant FSA A129S was found up to 17-
fold more efficient to condense DHA than FSA wt.
FSA A129S also accepted HA and GA as donor sub-
strates with 3.5- and 3.3-fold lower affinity, respective-
ly, than FSA wt. This differential selectivity of both
enzymes makes them complementary biocatalysts.
Hence, we have also demonstrated that the synthesis equivalent to 1 mmol of DHA.
Adv. Synth. Catal. 2010, 352, 1039 – 1046
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1043

Josꢀ A. Castillo et al.
71.1ACHTGNURTENNUGN
(C-3), 28.2 (C-4), 25.0 (C-5), 22.0 (C-7); HR-MS (ESÀ):
FULL PAPERS
Km and Vmax of donor substrates were measured as previ-
ously described.[7]
m/z=190.0711, calcd. for [C7H12NO5]: 190.0715.
(1R,2S,3R,6R)-1-Methyl-6-nitrocyclohexane-1,2,3-triol
(10): isolated yield: 898 mg (55%); Rf =0.34 (CH2Cl2/ace-
tone, 6:4); mp 162–1648C; [a]2D0: +21 (c 1.0, MeOH);
(1S,2S,3R,6R)-1-(Hydroxymethyl)-6-nitrocyclo-
hexane-1,2,3-triol (7)
1H NMR (400 MHz, CD3OD): d=4.52 (dd,
J
6,5ax =12.7,
6,5eq =3.8 Hz, 1H, H-6), 3.69 (ddd, J3,4ax =11.6, J3,2 =9.3,
3,4eq =5.2 Hz, 1H, H-3), 3.02 (d, J2,3 =9.3 Hz, 1H, H-2), 2.41
J
J
4-Nitrobutanal (500 mg, 4.3 mmol) was dissolved in H2O
(50 mL). Then, dihydroxyacetone (320 mg, 3.6 mmol) and
FSA A129S aldolase powder (125 mg, 338 U) were added.
The reaction mixture was placed on a reciprocal shaker for
17 h at 258C (175 rpm). MeOH (50 mL) was added and the
mixture was centrifuged (20000 rpm, 40 min, 58C). The su-
pernatant was concentrated under reduced pressure, afford-
ing a yellow oil. The crude material was purified by silica
gel chromatography using CH2Cl2/MeOH (90:10) as eluent
to afford the desired nitrocyclitol 7; isolated yield: 538 mg
(m, 1H, H-5ax), 2.03–1.95 (m, 1H, H-4eq), 1.92 (ddd, Jgem
=
12.7, J3,4ax =7.3, J5eq,6 3.8 Hz, 1H, H-5eq), 1.36 (s, 3H, CH3),
1.40–1.28 (m, 1H, H-4ax); 13C NMR (100 MHz, CD3OD):
d=91.6 (C-6), 80.0 (C-2), 74.7 (C-1), 70.9 (C-3), 29.6 (C-4),
25.0 (C-5), 23.0 (C-7); IR: n=3505, 3428, 2982, 2914, 1539,
1452, 1369, 1330, 1298, 1260, 1220, 1190, 1128, 1109, 1074,
1038, 899, 876, 858, 779, 723, 694, 642 cmÀ1; elemental analy.
calcd. (%) for C7H13NO5 (Mw 191.18): C 43.98, H 6.85, N
7.33, O 41.84; found: C 44.00, H 6.94, N 7.46, O 41.60.
(1S,2S,3R,6R)-1-Methyl-6-nitrocyclohexane-1,2,3-triol
1
(73%). H NMR and 13C spectra and [a]2D0 values were iden-
tical to those obtained previously in our group.[17] 1H NMR
(11): isolated yield: 358 mg (22%); Rf =0.24 (CH2Cl2/ace-
tone, 6:4); mp 125–1278C; [a]2D0: À58 (c 1.0, MeOH);
1H NMR (400 MHz, CD3OD): d=4.55 (dd, J6,5ax =9.7, J6,5eq
(400 MHz, CD3OD): d=4.79 (dd, 1H, H-6, J6ax,5ax =12.9,
J
J
6ax,5eq =4.0 Hz), 3.82 (d, Jgem =11.0 Hz, 1H, H-7), 3.73 (ddd,
3,4ax =11.6, J3,2 =9.4, J3,4eq =4.7 Hz, 1H, H-3), 3.39 (d, J2,3
=
7.1 Hz, 1H, H-6), 3.45 (ddd, J3,4ax =11.3, J3,2 =9.6, J3,4eq
=
9.4 Hz, H-2), 3.35 (d, Jgem,=11 Hz, 1H, H-7’), 2.49–2.36 (m,
1H, H-5eq) 2.03–1.99 (m, 1H, H-5eq), 1.99–1.95 (m, 1H, H-
4eq), 1.38–1.26 (m, 1H, H-4ax); 13C NMR (100 MHz,
CD3OD): d=86.4 (C-6), 77.1 (C-1), 75.2 (C-2), 71.0 (C-3),
4.9 Hz, 1H, H-3), 3.23 (d, J2,3 =9.6 Hz, 1H, H-2), 2.11–1.98
(m, 3H, 2H-5 and H-4eq), 1.47–1.34 (m, 1H, H-4ax), 1.18 (s,
3H, CH3); 13C NMR (100 MHz, CD3OD): d=92.7 (C-6),
81.6 (C-2), 76.1 (C-1), 71.3 (C-3), 29.8 (C-4), 26.0 (C-5), 15.2
(C-7); IR: n=3389, 2955, 2912, 1547, 1429, 1373, 1344, 1292,
1196, 1128, 1099, 1084, 1041, 1015, 984, 966, 947, 878, 781,
714 cmÀ1; elemental anal. calcd. (%) for C7H13NO5 (Mw
191.18): C 43.98, H 6.85, N 7.33, O 41.84; found: C 44.13, H
6.89, N 7.36, O 41.62.
61.9 (C-7), 29.6 (C-4), 24.7 (C-5); [a]2D0:
+ 22 (c 1.0,
MeOH).
1-Methyl-6-nitrocyclohexane-1,2,3-triols (8)–(11)
Hydroxyacetone (630 mg, 8.5 mmol) and FSA A129S aldo-
lase powder (250 mg, 675 U) were dissolved in water
(100 mL). 4-Nitrobutanal (1.1 g, 9.4 mmol) was added to the
reaction mixture. The reaction mixture was placed on a re-
ciprocal shaker for 18 h at 258C (175 rpm). The pH was
maintained between 7.5 and 8.5. MeOH (50 mL) was added
and the mixture was centrifuged (20000 rpm, 40 min, 58C).
The supernatant was concentrated under reduced pressure
at 258C. The crude material was purified by silica gel chro-
matography using CH2Cl2/acetone (6:4) as eluent to afford
the following nitrocyclitols 8–11; global yield: 87%.
(1S,2S,3R,4R,6R)-1-(Hydroxymethyl)-6-nitrocyclo-
hexane-1,2,3,4-tetraol (12)
To a solution of (R)-1,1-dimethoxy-4-nitrobutan-2-ol (30 mg,
0.17 mmol) in H2O (1 mL) was added a cation exchange
resin (Dowex 50”8, H+ form, 0.5 g). The suspension was
stirred at 458C for 1 h affording (R)-2-hydroxy-4-nitrobuta-
nal (quantitative by TLC). The resin was filtered off using
H2O (2 mL) as rinse. Then, the pH was adjusted to 7.5 with
1M NaOH. To this solution were added DHA (15 mg,
0.17 mmol) followed by FSA A129S (40 mg lyophilized
powder, 98 U). The reaction mixture was placed on a recip-
rocal shaker for 17 h at 258C (175 rpm). MeOH (50 mL)
was added and the mixture was centrifuged (20000 rpm,
40 min, 58C). The supernatant was concentrated under re-
duced pressure. The residue was purified by silica gel chro-
matography using CH2Cl2/MeOH (8:2) as eluent, affording
the desired nitrocyclitol 12; yield: 19 mg (52%). The NMR
data were identical to those previously published.[19]
(1S,2S,3R,6S)-1-Methyl-6-nitrocyclohexane-1,2,3-triol (8):
isolated yield: 100 mg (6%); Rf =0.53 (CH2Cl2/acetone, 6:4);
mp 143–1458C; [a]2D0: À17.4 (c 1.0, MeOH); IR: n=3435,
3300, 2949, 2909, 2548, 2450, 1541, 1452, 1373, 1316, 1292,
1221, 1136, 1063, 1049, 1013, 970, 920, 862, 812, 750 cmÀ1
;
1H NMR (400 MHz, CD3OD): d=4.71 (dd,
6,5eq =3.8 Hz, 1H, H-6), 3.79 (dd, J3,4ax =8.4, J3,2 =4.5 Hz,
J6,5ax =10.4,
J
1H, H-3), 3.69 (d, J2,3 =4.5 Hz, 1H, H-2), 2.52–2.41 (m, 1H,
H-5ax), 1.98–1.89 (m, 1H, H-4eq), 1.87–183 (m, 2H, H-4ax
and H-5eq), 1.28 (s, 3H, CH3); 13C NMR (100 MHz,
CD3OD): d=90.2 (C-6), 75.9 (C-2), 75.0 (C-1), 71.6 (C-3),
27.3 (C-4), 23.7 (C-7), 22.3 (C-5); HR-MS (ES+): m/z=
214.0706, calcd. for [C7H13NO5 +Na+]: 214.0691.
(1R,2S,3R,6S)-1-Methyl-6-nitrocyclohexane-1,2,3-triol (9):
isolated yield: 62 mg (4%); Rf =0.37 (CH2Cl2/acetone, 6:4);
[a]2D0: +43 (c 1.0, MeOH); 1H NMR (400 MHz, CD3OD):
d=4.68 (t, J6,5ax =5.4, J6,5eq =5.4, Hz, 1H, H-6), 3.82 (ddd,
1H NMR (400 MHz, CD3OD): d=5.11 (dd,
6,5eq =4.0 Hz, 1H, H-6), 4.08 (dd, J=6.2, 3.3 Hz, 1H, H-4),
J6,5ax =13.0,
J
3.85 (d, J2,3 =9.8 Hz, 1H, H-2), 3.82 (d, Jgem =11.1 Hz, 1H,
H-7), 3.70 (dd, J3,2 =9.8, J3,4 =3.3 Hz, 1H, H-3), 3.36 (d, J=
11.1 Hz, 1H, H-7’), 2.57 (dt, J5ax,5eq =13.1, J5ax,6 =13.1, J5ax,4
2.4 Hz, 1H, H-5ax), 2.13 (td, J5eq,5ax =13.2, J5eq,6 =4.0, J5eq,4
=
=
J
3,4ax =7.9, J3,2 =7.0, J3,4eq =4.4 Hz, 1H, H-3), 3.66 (d, J2,3 =
3.9 Hz, 1H, H-5eq); 13C NMR (100 MHz, CD3OD): d=82.9
(C-6), 77.2 (C-1), 73.0 (C-2), 70.2 (C-3), 68.6 (C-4), 61.9 (C-
7), 32.1 (C-5); [a]2D0: + 9 (c 0.7, MeOH)
7.0 Hz, 1H, H-2), 2.14–2.09 (m, 2H, H-5), 1.94–1.85 (m, 1H,
H-4), 1.85–1.74 (m, 1H, H-4’), 1.32 (s, 1H, CH3); 13C NMR
(100 MHz, CD3OD): d=92.0 (C-6), 77.6 (C-2), 73.7ACTHNUGTRNEUNG(C-1),
1044
asc.wiley-vch.de
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1039 – 1046

A Mutant d-Fructose-6-Phosphate Aldolase (Ala129Ser) with Improved Affinity towards Dihydroxyacetone
1.6 Hz, 2H), 3.79 (dd, J=10.1, 3.5 Hz, 1H), 3.69 (m, 2H),
l-Erythrulose (13)
3.58 (m, 4H), 3.45 (m, 2H); 13C NMR (101 MHz, D2O): d=
104.6, 101.6, 98.2, 82.1, 81.5, 80.8, 76.2, 75.5, 74.6, 69.8, 69.4,
67.7, 64.0, 63.5, 62.8, 62.5.
Formaldehyde (676 mg of aqueous solution 37% w/w,
8.3 mmol) and dihydroxyacetone (750 mg, 8.3 mmol) were
dissolved in water (85 mL). Then FSA A129S powder
(300 mg, 810 U) was added to the solution. The pH was ad-
justed at 7.5. The mixture was shaked for 21 h at 150 rpm
and 258C. MeOH (85 mL) was added and the mixture was
centrifuged (20000 rpm, 108C, 40 min). The supernatant was
concentrated under reduced pressure at 308C. The crude
material was purified by silica gel chromatography using
CH2Cl2/MeOH (9:1) as eluent to afford l-erythrulose 14 as
colourless oil; yield: 676 mg (68%). The NMR spectra were
identical to those of a commercial l-erythrulose sample:
1H NMR (400 MHz, D2O): d=4.62 (d, J=19.4 Hz, 1H, H-
1), 4.54 (d, J=19.4 Hz, 1H, H-1’), 4.47 (dd, J=4.3, 3.9 Hz,
1H, H-3), 3.89 (dd, J=12.2, 4.4 Hz, 1H, H-4), 3.87 (dd, J=
12.2, 3.9 Hz, 1H, H-4’); 13C NMR (100 MHz, D2O): d=
212.3 (C-2), 75.9 (C-3), 65.9 (C-1), 62.9 (C-4); [a]2D0: + 12 (c
1.0, H2O) [lit.21 [a]D20: + 13 (c 2.44, H2O)].
1-Deoxy-d-fructose (17)
Hydroxyacetone (79 mg, 1.1 mmol) and d-glyceraldehyde
(76 mg, 0.9 mmol) were dissolved in NaHCO3, 50 mM,
pH 7.5 buffer (11 mL). To this solution FSA A129S powder
(17 mg, 47 U) was added. Then the procedure and work up
were identical to that described for compound 14. The pu-
rification was similar to that described for d-fructose in the
following conditions: first, the column was eluted with
(ethyl acetate:CHCl3 1:1):MeOH 9:1 to remove the non-
converted starting substrates. Then, the product was eluted
with a mixture of (ethyl acetate:CHCl3 1:1):MeOH:H2O
from 80:16:4 to 70:20:10. After concentration under
vacuum, the residue was lyophilized to afford 1-deoxy-d-
fructose as a sticky white solid; isolated yield: 89 mg (68%);
1
[a]2D0: À87 (c 0.9, H2O) [lit.23 [a]D25: À80 (c 2.7, H2O)]. H and
13C NMR matched those reported by Jones et al.[23] 1H NMR
(500 MHz, D2O): d=4.06 (m, 1H), 4.01 (m, 1H), 3.99 (m,
1H), 3.80 (m, 2H), 3.65 (dd, J=18.2, 8.5 Hz, 2H), 3.62 (m,
2H), 3.59 (m, 2H), 3.57 (m, 2H), 2.28 (m, 1H, keto form),
1.47 (m, 1H, b-pyranose), 1.41 (m, 1H, b-furanose), 1.36 (m,
1H, a-pyranose); 13C NMR (101 MHz, D2O): d=213.7,
105.3, 101.7, 98.3, 82.5, 81.4, 80.6, 80.2, 77.0, 76.5, 74.8, 73.0,
72.1, 71.1, 70.9, 70.7, 69.6, 69.4, 63.4, 63.0, 62.8, 61.3, 25.8,
24.7, 23.9, 21.9.
d-Xylulose (14)
Dihydroxyacetone (111 mg, 1.2 mmol) and glycoladehyde
(59 mg, 1.0 mmol) were dissolved in NaHCO3 50 mM,
pH 7.5 buffer (12 mL). To this solution FSA A129S powder
(19 mg, 52 U) was added and the reaction mixture placed on
a reciprocal shaker (200 rpm) at 258C. When the limiting
substrate was consumed (~48 h) the reaction was filtered
through active charcoal to remove the enzyme. The filtrate
was adjusted to pH 5.0 with dilute HCl and freeze-dried.
The residue was purified by flash column chromatography
HPLC Measurements of Conversion of the Enzyme-
Catalyzed Reaction
on silica gel with [ethyl acetate:CHCl3ACTHNUTRGNEUNG(1:1):MeOH (9:1)] as
eluent. The oil obtained was then dissolved in water and
lyophilized to afford d-xylulose as a pale yellow oil; isolated
yield: 51 mg (42%); [a]2D0: À28 (c 1.1, H2O) [lit.22 [a]2D5: À32
For d-xylulose (14), 1-deoxy-d-xylulose (15), d-fructose
(16), 1-deoxy-d-fructose (17), d-arabinose (18), d-threose
(19), see Supporting Information.
1
(c 2.7, H2O)]; H NMR (500 MHz, D2O) (mixture of cyclic
and acyclic forms): d=4.54 (d, J=19.4 Hz, 1H), 4.43 (d, J=
19.4 Hz, 1H), 4.35 (s, 1H), 4.28 (m, 2H), 4.17 (m, 1H), 4.13
(dd, J=9.5, 6.2 Hz, 1H), 4.08 (dd, J=9.5, 6.5 Hz, 2H), 3.97
(m, 1H), 3.94 (d, J=5.6 Hz, 2H), 3.79 (dd, J=9.6, 4.2 Hz,
1H), 3.60 (m, 1H), 3.55 (m, 2H), 3.51 (d, J=11.1 Hz, 2H),
3.46 (d, J=12.1 Hz, 3H); 13C NMR (101 MHz, D2O) (mix-
ture of cyclic and acyclic forms): d=213.1, 105.9, 103.1, 80.7,
76.4, 76.0, 75.5, 75.0, 72.1, 72.1, 70.0, 66.2, 63.2, 62.6, 62.0.
Acknowledgements
JAC acknowledges the French foundation Vaincre Les Mal-
adies Lysosomales for a postdoctoral fellowship. MG and
XG acknowledge the I3P-CSIC predoctoral scholarship pro-
gram. This work was supported by the Spanish MCINN
CTQ2009-07359/BQU, La Maratꢀ de TV3 foundation (Ref:
050931), Generalitat de Catalunya DURSI 2005-SGR-00698
and ESF project COST CM0701. GS acknowledges the Deut-
sche Forschungsgemeinschaft which supported work of Mela-
nie S. and T.I. through SFB 380/B21. We thank Dr. Natalie
Trachtmann and Liliya Malikova for help with cloning of
vector pJFfsaA129S-kan.
d-Fructose (16)
Dihydroxyacetone (108 mg, 1.2 mmol) and d-glyceraldehyde
(86 mg, 0.96 mmol) were dissolved in NaHCO3 50 mM,
pH 7.5 buffer (12 mL). To this solution FSA A129S powder
(21 mg, 56 U, 68 mg protein) was added and the reaction
mixture placed on a reciprocal shaker (200 rpm) at 258C.
After 48 h the reaction was worked up and the crude isolat-
ed as described for compound 14. The flash column chroma-
tography over silica gel was eluted with (ethyl acet-
ACHTUNGTRENNUNGate:CHCl3 1:1):MeOH 9:1 to remove the non-converted
starting substrates. Then, the product was eluted with a mix- References
ture of (ethyl acetate:CHCl3, 1:1):MeOH:H2O from 86:10:4
to 75:20:5. After concentration under vacuum, the residue
[1] M. Schꢁrmann, G. A. Sprenger, J. Biol. Chem. 2001,
was lyophilized to afford d-fructose as a white solid; isolated
yield: 140 mg (81%); H NMR (500 MHz, D2O): d=4.00 (d,
J=3.5 Hz, 1H), 3.93 (s, 1H), 3.90 (s, 1H), 3.88 (d, J=
276, 11055–11061.
[2] S. Thorell, M. Schꢁrmann, G. A. Sprenger, G. Schneid-
er, J. Mol. Biol. 2002, 319, 161–171.
1
Adv. Synth. Catal. 2010, 352, 1039 – 1046
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1045

Josꢀ A. Castillo et al.
FULL PAPERS
[3] M. Schꢁrmann, M. Schꢁrmann, G. A. Sprenger, J. Mol.
Catal. B: Enzym. 2002, 19–20, 247–252.
[4] J. A. Castillo, J. Calveras, J. Casas, M. Mitjans, M. P. Vi-
nardell, T. Parella, T. Inoue, G. A. Sprenger, J. Joglar,
P. Clapꢀs, Org. Lett. 2006, 8, 6067–6070.
[5] M. Sugiyama, Z. Hong, P.-H. Liang, S. M. Dean, L. J.
Whalen, W. A. Greenberg, C.-H. Wong, J. Am. Chem.
Soc. 2007, 129, 14811–14817.
[6] A. L. Concia, C. Lozano, J. A. Castillo, T. Parella, J.
Joglar, P. Clapꢀs, Chem. Eur. J. 2009, 15, 3808–3816.
[7] X. Garrabou, J. A. Castillo, C. Guꢀrard-Hꢀlaine, T. Par-
ella, J. Joglar, M. Lemaire, P. Clapꢀs. Angew. Chem.
2009, 121, 5629–5633; Angew. Chem. Int. Ed. 2009, 48,
5521–5525.
[13] M. Sugiyama, Z. Y. Hong, W. A. Greenberg, C. H.
Wong, Bioorg. Med. Chem. 2007, 15, 5905–5911.
[14] For another example of thermostable aldolase: H. J.
Lamble, S. F. Royer, a D. W. Hough, M. J. Danson,
G. L. Taylor, S. D. Bull, Adv Synth Catal. 2007, 349,
817–821.
[15] G. A. Sprenger, M. Schꢁrmann, M. Schꢁrmann, S.
Johnen, G. A. Sprenger, H. Sahm, T. Inoue, U. Schçrk-
en, in: Asymmetric Synthesis with Chemical and Biolog-
ical Methods, 1st edn., (Eds.: D. Enders, K.-E. Jaeger),
Wiley-VCH, Weinheim, 2007, pp 312–326.
[16] M. Schꢁrmann, Dissertation, University of Dꢁsseldorf,
2001.
[17] L. El Blidi, D. Crestia, E. Gallienne, C. Demuynck, J.
Bolte, M. Lemaire, Tetrahedron: Asymmetry 2004, 15,
2951–2954.
[18] L. El Blidi, M. Ahbala, J. Bolte, M. Lemaire, Tetrahe-
dron: Asymmetry 2006, 17, 2684–2688.
[19] L. El Blidi, Z. Assaf, F. Camps, H. Veschambre, V.
Thꢀry, J. Bolte, M. Lemaire, ChemCatChem 2009, 1,
463–471.
[20] T. Gefflaut, C. Blonski, J. Periꢀ, M. Willson, Prog. Bio-
phys. Mol. Biol. 1995, 63, 301–340.
[8] A. K. Samland, G. A. Sprenger, Appl. Microbiol. Bio-
technol. 2006, 71, 253–264.
[9] For recent reviews about aldolases: a) P. Clapꢀs, W.-D.
Fessner, G. A. Sprenger, A. K. Samland, Curr. Opin.
Chem. Biol. 2010, 14, in press; b) S. M. Dean, W. A.
Greendberg, C.-H. Wong, Adv. Synth. Catal. 2007, 349,
1308–1320.
[10] S. Schenider, T. Sandalova, G. Schneider, G. A. Spreng-
er, A. K. Samland, J. Biol. Chem. 2008, 283, 30064–
30072.
[11] S. Schneider, M. Gutiꢀrrez, T. Sandalova, G. Schneider,
P. Clapꢀs, G. A. Sprenger, A. K. Samland, ChemBio-
Chem 2010, in press.
[12] M. Sugiyama, Z. Y. Hong, L. J. Whalen, W. A. Green-
berg, C. H. Wong, Adv. Synth. Catal. 2006, 348, 2555–
2559.
[21] H. De Wilde, P. De Clercq, M. Vandewalle, H. Rçper,
Tetrahedron Lett. 1987, 28, 4757–4758.
[22] R. M. Hochster, R. W. Watson, Arch. Biochem. Bio-
phys. 1954, 48, 120–129.
[23] N. A. Jones, S. F. Jenkinson, R. Soengas, M. Fanefjord,
M. R. Wormald, R. A. Dwek, G. P. Kiran, R. Devendar,
G. Takata, K. Morimoto, K. Izumori, G. W. J. Fleet, Tet-
rahedron: Asymmetry 2007, 18, 774–786.
1046
asc.wiley-vch.de
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2010, 352, 1039 – 1046