Highly efficient aldol additions of DHA and DHAP to N-Cbz-amino aldehydes catalyzed by l-rhamnulose-1-phosphate and l-fuculose-1-phosphate aldolases in aqueous borate buffer

Article abstract of DOI:10.1039/c1ob06263h

Aldol addition reactions of dihydroxyacetone (DHA) to N-Cbz-amino aldehydes catalyzed by l-rhamnulose-1-phosphate aldolase (RhuA) in the presence of borate buffer are reported. High yields of aldol adduct (e.g. 70-90%) were achieved with excellent (>98:2 syn/anti) stereoselectivity for most S or R configured acceptors, which compares favorably to the reactions performed with DHAP. The stereochemical outcome was different and depended on the N-Cbz-amino aldehyde enantiomer: the S acceptors gave the syn (3R,4S) aldol adduct whereas the R ones gave the anti (3R,4R) diastereomer. Moreover, the tactical use of Cbz protecting group allows simple and efficient elimination of borate and excess of DHA by reverse phase column chromatography or even by simple extraction. This, in addition to the use of unphosphorylated donor nucleophile, makes a useful and expedient methodology for the synthesis of structurally diverse iminocyclitols. The performance of aldol additions of dihydroxyacetone phosphate (DHAP) to N-Cbz-amino aldehydes using RhuA and l-fuculose-1-phosphate aldolase (FucA) catalyst in borate buffer was also evaluated. For FucA catalysts, including FucA F131A, the initial velocity of the aldol addition reactions using DHAP were between 2 and 10 times faster and the yields between 1.5 and 4 times higher than those in triethanolamine buffer. In this case, the retroaldol velocities measured for some aldol adducts were lower than those without borate buffer indicating some trapping effect that could explain the improvement of yields. The Royal Society of Chemistry 2011.

Full text of DOI:10.1039/c1ob06263h

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 8430
www.rsc.org/obc
PAPER
Highly efficient aldol additions of DHA and DHAP to N-Cbz-amino
aldehydes catalyzed by L-rhamnulose-1-phosphate and L-fuculose-1-phosphate
aldolases in aqueous borate buffer†
Xavier Garrabou,a Jordi Calveras,b Jesu´s Joglar,a Teodor Parella,c Jordi Bujonsa and Pere Clape´s*a
Received 27th July 2011, Accepted 13th September 2011
DOI: 10.1039/c1ob06263h
Aldol addition reactions of dihydroxyacetone (DHA) to N-Cbz-amino aldehydes catalyzed by
L-rhamnulose-1-phosphate aldolase (RhuA) in the presence of borate buffer are reported. High yields
of aldol adduct (e.g. 70–90%) were achieved with excellent (>98 : 2 syn/anti) stereoselectivity for most S
or R configured acceptors, which compares favorably to the reactions performed with DHAP. The
stereochemical outcome was different and depended on the N-Cbz-amino aldehyde enantiomer: the S
acceptors gave the syn (3R,4S) aldol adduct whereas the R ones gave the anti (3R,4R) diastereomer.
Moreover, the tactical use of Cbz protecting group allows simple and efficient elimination of borate and
excess of DHA by reverse phase column chromatography or even by simple extraction. This, in addition
to the use of unphosphorylated donor nucleophile, makes a useful and expedient methodology for the
synthesis of structurally diverse iminocyclitols. The performance of aldol additions of dihydroxyacetone
phosphate (DHAP) to N-Cbz-amino aldehydes using RhuA and L-fuculose-1-phosphate aldolase
(FucA) catalyst in borate buffer was also evaluated. For FucA catalysts, including FucA F131A, the
initial velocity of the aldol addition reactions using DHAP were between 2 and 10 times faster and the
yields between 1.5 and 4 times higher than those in triethanolamine buffer. In this case, the retroaldol
velocities measured for some aldol adducts were lower than those without borate buffer indicating some
trapping effect that could explain the improvement of yields.
preparation of these compounds.10,12,19–23 Therefore, improving
Introduction
the efficiency of this step by increasing the catalytic activity of
The catalytic asymmetric carbon–carbon bond formation by the
the aldolases, the yield of the process and the atom economy
aldol addition reaction is one of the simplest and most powerful
by using non-phosphorylated donor substrates is of paramount
strategies for the synthesis of polyoxygenated molecules with
importance from scientific and industrial viewpoints. Although
many stereogenic centers, such as carbohydrates and analogues.1–4
the preparation and synthetic applications of DHAP and other
Dihydroxyacetone (DHA) constitutes an important C3 building
phosphorylated analogues have reached a high degree of sophisti-
cation and efficiency24 the preferred choice for the preparation of
unphosphorylated targets is DHA.
block for the construction of these compounds through aldol
addition.5–12 In this regard, a great deal of progress has been
achieved in biocatalytic aldol additions of dihydroxyacetone
phosphate (DHAP) or DHA to aldehydes.13–18
During our ongoing project on the chemo-enzymatic synthe-
sis of iminocyclitols, the enzymatic aldol addition of DHAP
or DHA to N-Cbz-amino aldehydes is the key step for the
Recently, it was uncovered that L-rhamnulose-1-phosphate
aldolase (RhuA) can accept DHA as donor when the reactions
were carried out in the presence of borate,25 similarly to what
was suggested in pioneering studies with arsenate salts.26,27 In
this context, we reported that RhuA can catalyze the aldol
addition of DHA to N-Cbz-amino aldehydes without borate.23
Interestingly, when sodium borate was added the rates of aldol
formation improved between 35- and 100-fold.23 The tolerance
of unphosphorylated DHA by a Class II DHAP-dependent
aldolase appears to be an exclusive property of RhuA, since
the stereocomplementary L-fuculose-1-phosphate aldolase from
E. coli (FucA) had no detectable activity with DHA either with
or without borate added.23 This is a remarkable difference with
the arsenate esters, which were accepted by both RhuA and FucA
aldolases.26
aDept Biological Chemistry and Molecular Modeling. Instituto de Qu´ımica
Avanzada de Catalun˜a, IQAC-CSIC, Jordi Girona 18-26, 08034, Barcelona,
Spain. E-mail: pere.clapes@iqac.csic.es; Fax: +34 932045904; Tel: +34
934006112
bPresent Address: Department of Medicinal Chemistry, College of Phar-
macy, University of Texas, Austin, USA
cServei de Ressona`ncia Magne`tica Nuclear, Departament de Qu´ımica,
Universitat Auto`noma de Barcelona, Bellaterra, Spain
† Electronic supplementary information (ESI) available: See DOI:
10.1039/c1ob06263h
8430 | Org. Biomol. Chem., 2011, 9, 8430–8436
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 1 L-Rhamnulose-1-phosphate aldolase catalyzed aldol addition
reactions of DHA-borate to N-Cbz-amino aldehydes (1)a
b
Entry
Aldehyde
%
syn : antic
1
2
3
4
5
6
7
8
1a
91
99
95
94
99
94
91
62
64
93
93
89
42
52
99
94
95
nr
94 : 6
(S)-1b
(R)-1b
(S)-1c
(R)-1c
(S)-1d
(R)-1d
(S)-1e
(R)-1e
(S)-1f
(R)-1f
(S)-1g
(S)-1h
(R)-1h
1i
>98 : 2
45 : 55
>98 : 2
76 : 24
>98 : 2
>2 : 98
>98 : 2
>2 : 98
>98 : 2
>2 : 98
>98 : 2
>98 : 2
>2 : 98
>98 : 2
81 : 19
>98 : 2
nr
9
10
11
12
13
14
15
16
17
18
1j
(S)-1k
(R)-1k
a Reaction conditions: [Aldehyde] = 65 mM; [DHA] = 100 mM; RhuA =
0.6 U; volume of the mixture 300 mL; 200 mM sodium borate pH 7.5.
Percentage of aldol adduct measured after 24 h. b Percentage of aldol
adduct 2 formed after 24 h respect to the limiting substrate, i.e. acceptor
aldehyde. The amount of product was measured by HPLC from the peak
integration using an external standard method. c The stereochemistry of
the aldol adducts of the aldol additions of DHA to the aldehydes 1a, (S)-1b,
(S)- (R)-1c, (S)- (R)-1d, (S)- (R)-1e, (R)-1f, (R)-1h, 1i, and 1j, was inferred
from the corresponding iminocyclitols obtained after reductive amination
of the corresponding aldol adduct, after elimination of the excess of DHA
and borate (see experimental procedures and ESI†). The stereochemical
outcome of the aldol additions of DHA to aldehydes (R)-1b, (S)-1f, (S)-1g,
(S)-1h and (S)-1k, were determined by HPLC (see experimental procedures
and ESI†); syn and anti adducts refer to those with (3R,4S) and (3R,4R)
configuration, respectively. nr: no reaction.
Besides the intrinsic tolerance of RhuA for DHA, the measured
retroaldol rates for some aldol adducts in the presence of borate
were low or negligible as compared with the synthetic ones,
making the process virtually irreversible.25 Therefore, it was further
suggested that the aldol adduct may be trapped by formation of
borate complexes which would be less active substrates for the
aldolase.25
In this paper, we investigated the exploitation of the aldol
additions of DHA-borate to a structural variety of N-Cbz-
amino aldehydes catalyzed by RhuA for the two-step synthesis
of iminosugars. Since the acceptance of DHA seems to be an
exclusive property of RhuA, we also investigated whether the
addition of sodium borate may improve the performance of
enzymatic aldol addition reactions of DHAP to N-Cbz-amino
aldehydes catalyzed by RhuA and FucA aldolases.
Scheme 1 RhuA- and FucA-catalyzed aldol addition reactions of DHA
and DHAP to selected N-Cbz-amino aldehydes (1). i) when X = PO32-, two
steps: a) dephosphorylation by acid phosphatase, b) reductive amination,
H2/Pd; ii) when X = H, one step: a) reductive amination, H2/Pd.
Stereocenter control (compounds 3 and 4, respectively): 4* (3) and 1*
(4) by the aldolase, 2* (3) and 3* (4) by reductive amination; 5* (3) and 7a*
(4) are selected from the starting N-Cbz-amino aldehyde. The absolute
configuration of 3* (3) and 2* (4) depended on the aldolase and it is
conserved upon the reaction with electrophiles.
The full equivalence of the stereochemical outcome observed
for the DHA and DHAP nucleophiles, indicates the unbiased
orientation of both donors in the active site of RhuA catalyst.19,21,29
Differences in the syn/anti ratio between the S and R enantiomers
of N-Cbz-amino aldehydes were also reported using DHAP
donor.21 Notably, the additions of DHA-borate to (R)-1d, (R)-
1e, (R)-1f and (R)-1h (Table 1, entries 7, 9, 11 and 14) furnished
exclusively the anti adducts, whereas the (S)-aldehydes yielded
always the syn ones. This high stereoselectivity towards the R
enantiomers of N-Cbz-amino aldehydes at 25 ◦C using DHA
donor with borate contrasted with the aldol addition reactions
Results and discussion
Enzymatic aldol addition catalyzed by L-rhamnulose-1-phosphate
aldolase
Aldol additions of DHA to Cbz-N-amino aldehydes 1 in the
presence of 200 mM sodium borate pH 7.5 (Scheme 1)28 at 25 ◦C
furnished high conversions (Table 1), which are comparable to
those achieved under different optimized conditions using DHAP
as donor (see Table S1, ESI†).19,21,29
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 8430–8436 | 8431
©

View Article Online
with DHAP catalyzed by RhuA where different mixtures of
syn/anti aldol adduct were obtained.21
The possibility that the borate may lead to the formation of
borate-ester complexes with the aldol adduct was also investigated.
In previous works, we have observed that the aldol adducts formed
from the aldol addition of DHAP, DHA or glycolaldehyde to
N-Cbz-aminoaldehydes, consist of a mixture of an acyclic and
cyclic hemiaminal compounds in equilibrium.12,19,20,30,31 As an
1
example, the H NMR spectra at 263 K of pure aldol adduct
(3R,4S,5S)-2b (Fig. 1), which was obtained from the addition
of DHA to (S)-2b (Table 1, entry 2), confirmed a major acyclic
compound and two hemiaminal conformers (I) and (II) (Fig. 1)
(i.e. rotamers from the Cbz group) with the syn configuration
(3R,4S), as ascertained by NMR (see ESI†).32
Fig. 2 Observed 11B NMR of a sample of (3R,4S,5S)-2b (a) without
borate added and at (3R,4S,5S)-2b : sodium borate millimolar ratios of (b)
114 : 42, (c) 106 : 136, (d) 100 : 209, and (e) 89 : 357.
at different molar ratios of (3R,4S,5S)-2b : borate. Thus, when
[(3R,4S,5S)-2b] ≥ [borate] the 11B NMR spectrum showed the
appearance of a signal (Fig. 2b,c) at d ~ -10 ppm, while at higher
borate concentrations (Fig. 2d,e) signals at d ~ -14 ppm and ~ -19
ppm also appeared. Taking into consideration the structure of
the (3R,4S,5S)-2b hemiaminal and by comparison with literature
data on borate-sugars complexes,33–35 it could be reasoned that
the signal at ~ -10 ppm could arise from the formation of an
a,b-bidentate BL2 complex, those at ~ -14 ppm from a,b-bidentate
BL complexes and that at ~ -19 ppm from a,g-bidentate BL or
BL2 complexes.33,36
Although carrying exhaustive modelization studies on the
potential borate esters that could arise from the (3R,4S,5S)-2b
hemiaminal was beyond the scope of this work, DFT calculations
at the B3LYP/6-31G** level (see ESI†) allowed to propose the
complexes shown in Scheme 2 as plausible structures responsible
for our observations.
The formation of borate ester complexes with the aldol adduct
appears to be plausible, although we cannot conclude that this
could be the only explanation for the high yields obtained.
Indeed, among the examples shown in Table 1, the behavior of
prolinal derivatives is remarkable (Table 1, entries 17 and 18):
the aldol addition of DHA-borate to the (S)-1k furnished almost
quantitative conversions to (3R,4S,5S)-2k, whereas no reaction
was detected with the enantiomer (R)-1k. Both acceptors were
well tolerated by RhuA with DHAP donor.22,29 Moreover, it can
be anticipated that the protected secondary amine of (3R,4S,5S)-
2k and (3R,4S,5R)-2k aldol adducts precludes the formation of a
cyclic hemiaminal.
As a proof of concept of this methodology, the synthesis of the
potent glycosidase inhibitor 1,4-dideoxy-1,4-imino-L-arabinitol
(LAB, 3) was conducted (Scheme 3).37–40
The aldol adduct intermediate 2a was obtained in a 73%
isolated yield and, after reductive amination, 600 mg of LAB
(see experimental and ESI†) were obtained. This result compares
much more favorably to that obtained using the corresponding
Fig. 1 1H NMR spectra of pure (3R,4S,5S)-2b at 263 K (a) and records
at different millimolar ratios of (3R,4S,5S)-2b/sodium borate: b) 114 : 42;
c) 106 : 136; d) 100 : 209; e) 89 : 357.
On increasing the concentration of borate the 1H NMR signals
corresponding to the acyclic species decreased while the ones
assigned to the cyclic hemiaminals became broad (see signals cor-
responding to the methyl group 7, and those of the hydroxymethyl
moiety 1 of the acyclic compound, Fig. 1). This result suggests that
different borate complexes are formed which interconvert in a slow
equilibrium (Fig. 1). Unfortunately, we were unable to resolve the
NMR signals, even by decreasing the temperature below 263 K
and, therefore, none of the possible borate-hemiaminal structures
could be elucidated. Analysis of the same samples by 11B NMR
(Fig. 2) revealed the formation of different borate complexes
8432 | Org. Biomol. Chem., 2011, 9, 8430–8436
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Scheme 2 Formation of potential hemiaminal-borate complexes derived from aldol adduct (3R,4S,5S)-2b.
Table 2 Initial reaction rates (vo, mmol aldol adduct formed h-1 mg-1) of
the aldol addition reaction of DHAP to (R)-1h catalyzed by L-fuculose-1-
phosphate aldolase at 25 ◦
C
vo (mmol h-1 mg-1)a
Entry Aldehyde acceptor FucA catalyst pH Boratea
TEAa
1
2
3
4
(R)-1h
(R)-1h
(R)-1h
(R)-1h
F131A
Wild-type
F131A
7.0 66
7.0 14
6.6 42
7.4 82
22(15)b
1.5
12
Scheme
3
Chemo-enzymatic synthesis of 1,4-dideoxy-1,4-imino-
F131A
38
L-arabinitol (LAB, 3) mediated by L-rhamnulose-1-phosphate aldolase.
a Reactions were conducted at 25 ◦C using 200 mM sodium borate, 50 mM
TEA adjusted to pH 6.6, 7.0 or 7.4. [DHAP] = 50 mM; [(R)-1h] = 65 mM,
volume = 360 mL, FucA wild-type = 2.3 U (8 U mL-1), FucA F131A =
0.024 U (0.08 U mL-1). Values are the mean of three determinations with
a standard error between 11–13%. b The DHAP solution was adjusted to
pH 6.9–7.0 and the reaction was conducted without any buffer solution
added.
azido aldehyde as acceptor for the synthesis of LAB (4% isolated
yield).25 The Cbz group was tactically essential to effectively purify
the aldol adducts intermediates by simple column chromatography
onto a hydrophobic Amberliteꢀ XAD16. Once the reaction
crude was loaded onto the column, the excess of DHA, borate
and enzyme were eliminated by washing with aqueous HCl
(10 mM) and plain water. The aldol adduct was eluted with an
ethanol : water mixture.
R
vo values were significantly different when using 200 mM sodium
borate, 50 mM TEA at pH 7 or the DHAP solution adjusted to
pH 7 without buffer added (Table 2).
Effect of borate buffer on the yield of aldol addition reactions of
DHAP to N-Cbz-amino aldehydes catalyzed by RhuA and FucA
aldolases
Indeed, when borate was used, the vo for FucA F131A was
between 2.1- to 4-fold higher than that with TEA or when no
buffer solution was used (Table 2, entry 1) and up to 10-fold
higher for FucA wild type (Table 2, entry 2). The pH also
affected significantly the vo of the aldol additions catalyzed by
FucA catalysts: values >7 were the most favorable (Table 2,
entries 3 and 4), albeit higher pH must be avoided since the
rate of DHAP degradation increases and aldehydes may undergo
side reactions.42,43 A similar behavior was found when using the
(S)-1h as acceptor and either using FucA as an ammonium
sulphate suspension or in solution (see ESI† for the preparation
of FucA in solution and the results).
For preparative purposes, we found that the aldol additions
of DHAP to N-Cbz-amino aldehydes at 4 ◦C provided the best
yields in FucA catalysis.21,22,43 Therefore, to better compare with
previously published data22 and conduct the reactions under
optimum conditions to avoid DHAP degradation, the screening
of the reactions using 200 mM sodium borate was performed
at 4 ◦C.
As mentioned in the introduction, the presence of borate decreased
the retroaldol activity of RhuA catalyst for some aldol adducts,
while the aldol addition of DHAto aldehydes in borate is efficiently
catalyzed by RhuA.25 In this way, the process might be virtually
irreversible. Moreover, it was demonstrated that cyclic aldol
adducts form stable complexes with borate and that might help to
shift the reaction towards the synthetic direction. Therefore, it was
considered both interesting and significant to investigate if those
borate effects can also be found in the aldol addition reactions
of DHAP to N-Cbz-amino aldehydes catalyzed by L-fuculose-1-
phosphate aldolase (FucA) wild type and FucA F131A mutant,
which were the most active ones towards sterically demanding
a-substituted aldehydes and conformationally restricted prolinal
derivatives.22 For the sake of comparison, RhuA catalyst was
included in the set of experiments.
First, the influence of the buffer composition, i.e., borate or
triethanolamine, on the initial rate, vo, of the aldol addition of
DHAP to (R)-1h, as model reaction, was investigated. Working
with RhuA, variations on the buffer type and strength (up to
200 mM) did not affect the initial reaction rate.41 On the other
hand, using FucA catalysts under the same reaction conditions the
Working with RhuA catalyst, the aldol adduct yields using
DHA in the presence of borate were similar to those with DHAP
(see Table S2 in ESI†) and therefore there is no general positive
effect of borate in this case. On the other hand, the presence of
borate has a positive effect on the aldol adduct yield for FucA
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 8430–8436 | 8433
©

View Article Online
Table 3 L-Fuculose-1-phosphate aldolase catalyzed aldol addition reac-
tions of DHAP to N-Cbz-amino aldehyde derivatives in the presence of
borate
2h, which forms cyclic hemiaminals, is more strongly inhibited
than that of (3R,4R,5R)-2k, whose tertiary amine prevents the
formation of that cyclic species.
a
%
Conclusions
Entry
Aldehyde
FucA wild-typeb
FucA F131Ab
The aldol additions of DHA to a variety of N-Cbz-amino
aldehydes catalyzed by RhuA in the presence of borate provide
high conversions with excellent stereochemistry, which compares
favorably to the reactions performed with DHAP. In addition
to that, the tactical use of Cbz protecting group allows simple
and efficient elimination of borate and excess of DHA by reverse
phase column chromatography or even by simple extraction.
Moreover, the use of DHA as donor for the preparation of
unphosphorylated targets simplifies the synthetic strategy avoid-
ing the manipulation of the phosphate moiety and fostering the
methodology to industrial application. To sum up, this is a useful
and expedient methodology for the preparation of structurally
diverse iminocyclitols.
The effect of borate on the performance of the aldol additions
of DHAP to N-Cbz-amino aldehydes was different for RhuA and
FucA aldolases. In general, it was positive for the yield of aldol
adduct but depended on each particular acceptor. Interestingly, for
FucA catalysis, the rate of the aldol addition reaction of DHAP
to N-Cbz-amino aldehydes in borate were between 2 to 10 times
faster than that in TEA buffer. Moreover, the yields of aldol adduct
improved between 1.5 to 4-fold for both FucA wild-type and the
F131A mutant. In this case, it appears that the borate inhibits the
retroaldol reaction favoring the aldol adduct formation.
1
2
3
4
5
6
7
8
1a
63c(52d)
65c(68d)
85c(72d)
48c(16d)
65c(16d)
60c(22d)
62c(29d)
75c(27d)
47c(20d)
7c(2d)
50(59d)
64c(64d)
83c(76d)
60c(38d)
76c(56d)
60c(36d)
65c(68d)
78c(50d)
71c(39d)
70c(60d)
(S)-1b
(R)-1b
(S)-1f
(R)-1f
(S)-1g
(S)-1h
(R)-1h
(S)-1k
(R)-1k
9
10
a Percentage of aldol adduct 2 formed respect to the limiting substrate,
DHAP. The amount of aldol adduct formed was determined by HPLC
using an external standard method. b The stereochemical outcome of
the aldol addition of DHAP to N-Cbz-amino aldehydes catalyzed by
FucA wild-type and FucA F131A in borate buffer was identical to the
ones previously published for the same reactions conducted with DHAP
adjusted to pH 6.9–7.0 without any buffer solution added (see ref. 22).
c Reactions were conducted at 4 ◦C in 200 mM sodium borate pH 7.0.
[DHAP] = 40 mM; [aldehyde] = 68 mM, volume = 360 mL, FucA wild-
type = 2.3 U (8 U mL-1), FucA F131A = 0.024 U (0.08 U mL-1); conversions
measured after 24 h. d Experiments conducted with DHAP adjusted to pH
6.9–7.0 without any buffer solution added. Reactions were conducted at
4 ◦C. The rest of the conditions were identical to the experiments performed
in borate. Conversions measured after 24 h. Data from ref. 22
wild-type and FucA F131A mutant catalysis (Table 3, entries 4–
6 and 8–9). Interestingly, this was more significant for aldehydes
1f, 1g and 1h (Table 3 entries 4–8), which had previously been
determined to be poor acceptor substrates particularly with FucA
wild-type,22 and limited for good substrates such as 1a and 1b
(Table 3, entries 1–3). It is noteworthy that, under these conditions
and with the exception of the prolinal derivatives, the conversions
achieved with FucA wild-type for the most sterically demanding
a-substituted aldehydes were comparable to those achieved by
FucA F131A mutant, the most active biocatalyst found towards
those substrates.22
Measurements of the initial retroaldol velocities (voretro) for the
adducts (3R,4R,5R)-2h and (3R,4R,5R)-2k (Fig. 3) catalyzed by
FucA F131A mutant at 25 ◦C in borate, showed a significant
33- and 2-fold reduction, respectively, as compared to those with
TEA buffer: from 20 to 0.6 mmol h-1 mg-1 for (3R,4R,5R)-2h
and from 571 to 260 mmol h-1 mg-1 for (3R,4R,5R)-2k, without
and with 200 mM sodium borate, respectively. This indicates
that some phosphorylated aldol adducts may be weak substrates
for the aldolase in borate buffer, thus favoring the synthetic
reaction as observed with some examples with RhuA catalysts
(vide supra). Interestingly, the retroaldolization of (3R,4R,5R)-
Experimental
Materials
Dihydroxyacetone phosphate (DHAP) was obtained from the
cyclic dimer precursor 2,5-diethoxy-p-dioxane-2,5-dimethanol-O-
21-O-51-bisphosphate, which was synthesized in our lab using a
procedure described by Jung et al.44 with slight modifications.
DHAP is liberated from the precursor by mild acidic hydrolysis
at 65 ◦C. N-Cbz-amino aldehydes used in these studies were
synthesized in our lab using procedures published in previous
works.22 L-Rhamnulose-1-phosphate aldolase [CoII] (3.8 U mg-1
protein) (1 U catalyzes the cle◦avage of 1 mmol of L-rhamnulose-
1-phosphate per minute at 25 C and pH 7.5 (100 mM Tris·HCl
+ 150 mM KCl));23,45 L-fuculose-1-phosphate aldolase wild-type
9.8 U mg-1 protein and FucA F131A mutant 0.1 U mg-1 were
obtained as previously describe◦d (1 U cleaves 1 mmol of L-fuculose-
1-phosphate per minute at 25 C and pH 7.5 (100 mM Tris·HCl
+ 150 mM KCl)).22,46 Aqueous sodium borate solutions were
prepared by adjusting the desired pH of a solution of boric acid
with NaOH (2 M). Deionized water was used for preparative
HPLC and Milli-Q-grade water for analytical HPLC. All other
solvents used were of analytical grade.
HPLC analyses
R
HPLC analyses were performed on a RP-HPLC XBridgeꢀ C18,
5 mm, 4.6 ¥ 250 mm column (Waters). The solvent system used
was: solvent (A): 0.1% (v/v) trifluoroacetic acid (TFA) in H2O
and solvent (B): 0.095% (v/v) TFA in CH3CN/H2O 4 : 1; gradient
Fig. 3 Aldol adducts used for the initial retroaldol velocities.
8434 | Org. Biomol. Chem., 2011, 9, 8430–8436
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
elution from 10% to 70% B over 30 min to monitor reactions
with aldehydes 1a, 1b, 1c, 1i, 1j; or 30% to 90% B over 30 min to
monitor those with 1d, 1e, 1f, 1g, 1h, 1k, flow rate was 1 mL min-1,
detection was at 215 nm and column temperature at 30 C. The
amount of aldol adduct produced was quantified from the peak
areas using an external standard methodology.
the analytical reactions. Samples were withdrawn at 5, 10, 20, 30
and 45 min and analyzed by HPLC as described above. Linear
correlations with the reaction time were found for aldol adduct
production lower than 15% with respect to the limiting reactant,
DHAP. The estimated standard error for three determinations was
between 11–13%.
◦
Reactions with DHA as donor in borate buffer
Initial retroaldol velocity (vo) with phosphorylated aldol adducts
Analytical scale reactions (300 mL total volume) were conducted
in 2 mL test tubes with screw cap stirred with a vortex mixer
(VIBRAX VXR basic, Ika) at 1000 rpm and 25 ◦C. Aldehydes
1 (19.5 mmol) were dissolved in dimethylformamide (60 mL)
and these solutions were mixed with a DHA solution (90 mL,
30.0 mmol) and buffer (60 mL of 1.0 M sodium borate, pH 7.5).
RhuA (90 mL, 0.16 mg of protein, 0.6 U, corresponding to 2 U
mL-1) was then added to start the aldol reaction. Conversions
were measured at 24 h reaction time. Reaction monitoring was as
follows: samples (25 mL) were withdrawn at several reaction times,
diluted with methanol (975 mL) and analyzed by HPLC under the
conditions described above.
Retroaldol assays with aldol adducts were carried out as follows: A
solution of NADH (0.9 mL, 2 mM) in Tris-HCl (110 mM, pH 7.5)
or in sodium borate (220 mM, pH 7.5), containing KCl (160 mM),
the aldolase (variable amount depending on the activity) and
glycerol-3-phosphate dehydrogenase (1.7 U mL-1) was incubated
at 25 ◦C for 5 min. To this solution the aldol adduct (3R,4R,5R)-
2h or (3R,4R,5R)-2k (100 mL, 100 mM) was added up to a total
final volume of 1 mL. The decrease of absorbance of NADH at
340 nm was measured. The estimated standard error for three
determinations was between 5–10%.
Addition of DHA to aldehydes in borate buffer. Laboratory scale
reactions (70 lmol scale). General procedure
Enzymatic aldol reactions with DHAP as donor
To determine the stereochemistry of the aldol adducts obtained
with RhuA using DHA and aldehydes 1a, (S)-1b, (S)-1c, (R)-
1c, (S)-1d, (R)-1d, (S)-1e, (R)-1e, (R)-1f, (R)-1h, 1i, and 1j in
the presence of sodium borate, reactions (~70 mmol scale) were
conducted in 25 mL screw capped test tubes (2.5 mL, total
reaction volume). Each aldehyde (0.07 mmol) was dissolved in
dimethylformamide (0.5 mL) and this solution was mixed with a
DHA solution (1.9 mL, 0.09 mmol) in 263 mM sodium borate,
pH 7.5. Stock solution of RhuA (0.1 mL, 5 U) in Tris-HCl 100
mM, pH 7.5 was added to start the reaction. The stock solution
of RhuA was prepared as follows: the ammonium sulphate
suspension of RhuA (1.76 mL, 113 U mL-1) was centrifuged and
the pellet resuspended in Tris-HCl 100 mM pH 7.5 (400 mL). After
48 h, MeOH (5 mL) was added to the reaction mixture to stop
the enzymatic reaction and precipitate the enzyme. MeOH was
removed in a rotary evaporator. Then, the mixture was filtered
through a 0.45 mm cellulose membrane filter. The filtrate was
adjusted to pH 3.0 and loaded onto a semi-preparative column
Analytical scale reactions (360 mL total volume) were conducted
in 2 mL test tubes with screw cap stirred with a vortex mixer
(VIBRAX VXR basic, Ika) at 1000 rpm at 4 ◦C. Aldehydes 1
(24.5 mmol) were dissolved in dimethylformamide (72 mL). Each
of the aldehydes were mixed with water (47.2 mL) or sodium borate
(47.2 mL of 1.52 M, pH 7.0 solution), a freshly neutralized (pH
6.9–7.0) DHAP solution (141 mL, 14.4 mmol), and the aldolase
(for RhuA catalyzed reactions: 100 mL, 0.19 mg of protein, 0.72 U
corresponding to 2 U·mL-1 reaction; for reactions with FucA wild-
type or FucA mutants: 100 mL, 0.29 mg of protein, (corresponding
to 2.9 U FucA wild-type (8 U mL-1), and 0.029 U FucA F131A
(0.08 U mL-1)) was added to start the aldol reaction. Conversions
were measured at 24 h of reaction time. Reaction monitoring was
as follows: samples (25 mL) were withdrawn at several reaction
times, diluted with methanol (975 mL) and analyzed by HPLC
under the conditions described above. In parallel, samples (25 mL)
were mixed with a solution (25 mL) of acid phosphatase (5.3 U
mL-1 in 400 mM sodium citrate pH 4.5) and incubated for 24 h.
After dilution with methanol (950 mL), samples were analyzed by
HPLC.
R
X-Terraꢀ (19 ¥ 250 mm). Products were eluted using CH3CN
gradient from 0% to 52% CH3CN in 30 min. The flow rate was
10 mL min-1 and the products were detected at 215 nm. Analysis
of the fractions was accomplished under gradient conditions (10
to 70% of solvent B in 30 min) on the analytical HPLC. Pure
fractions were pooled and lyophilized. The solid thus obtained
was dissolved in H2O/MeOH 1 : 1 and treated with H2 (50 psi)
in the presence of 10% Pd/C (50 mg) at room temperature
during 24 h. After removal of the catalyst by filtration through
a 0.45 mm nylon membrane filter, the filtrate was adjusted to pH
6.4 with formic acid and lyophilized obtaining a solid material.
These experiments were intended to determine the stereochemical
outcome of the reactions, therefore the purification procedures
were not optimized and quantities around 5–10 mg of each
iminocyclitol were obtained. The 1H and 13C NMR matched those
previously described and, from the corresponding 1H spectra, the
diastereomeric excess for each reaction was inferred similarly as
previously reported (see ESI†).19–22 The stereochemical outcome
Enzymatic aldol reactions with DHAP as donor: initial velocities
(vo) of aldol adduct formation with the acceptor (R)-1h
The initial velocities of aldol formation (vo) were determined by
measuring the amount of aldol adduct produced by HPLC at
the initial reaction times. The amount of aldol adduct produced
was quantified by an external standard method. Reactions were
conducted as described above for analytical reactions using 360 mL
reaction volume: 47.2 mL of 380 mM or 1.52 M triethanolamine-
HCl (TEA) pH 6.6, 7.0 or 7.4, or 47.2 mL of 1.52 M sodium
borate pH 6.6, 7.0 or 7.4, final concentration 50–200 mM TEA and
200 mM sodium borate. RhuA = 0.72 U (2 U mL-1), FucA wild-
type = 2.9 U (8 U mL-1), FucA F131A = 0.029 U (0.08 U mL-1). The
amount of aldehyde, dimethylformamide, and freshly neutralized
DHAP solution used for these reactions was identical to that for
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 8430–8436 | 8435
©

View Article Online
of the aldol additions of DHA to aldehydes (R)-1b, (S)-1f, (S)-1g,
(S)-1h and (S)-1k, were determined by HPLC. The corresponding
diastereomeric mixtures of unphosphorylated aldol adducts were
separated under the HPLC conditions and compared to those
obtained in a previous work (see ESI†).22
13 P. Clape´s, W.-D. Fessner, G. A. Sprenger and A. K. Samland, Curr.
Opin. Chem. Biol., 2010, 14, 154–167.
14 W.-D. Fessner and S. Jennewein, in Biocatalysis in the Pharmaceutical
Biotechnological Industries., ed. R. N. Patel and M. Dekker, New York,
2007, pp. 363–400.
15 M. Sugiyama, W. Greenberg and C.-H. Wong, J. Synth. Org. Chem.
Jpn., 2008, 66, 605–615.
16 A. K. Samland and G. A. Sprenger, Appl. Microbiol. Biotechnol., 2006,
71, 253–264.
17 W.-D. Fessner, in Asymmetric Organic Synthesis with Enzymes, ed. V.
Gotor, I. Alfonso and E. Garcia-Urdiales, Wiley-VCH Verlag GmbH
& Co. KGaA, Weinheim, 2008, pp. 275–318.
Preparative synthesis of 4-dideoxy-1,4-imino-L-arabinitol (LAB)
As a proof of synthetic utility of these methodology, the aldol
addition of DHA with aldehyde 1a in the presence of borate was
scaled up to 200 mL total volume at 12 mmol scale: 1a (2.3 g,
12 mmol) was dissolved in dimethylformamide (40 mL). To this
solution, DHA (1.4 g, 15 mmol) in water (120 mL) and 1.0 M
sodium borate, pH 7.5 (40 mL, 200 mM final concentration) was
added. RhuA (400 U, 32 mL of an ammonium sulphate suspension
of 12.5 U mL-1 were centrifuged and the pellet dissolved in 32 mL
of 200 mM sodium borate, pH 7.5). After 48 h the reaction mixture
was loaded onto a chromatography column (5 ¥ 15 cm, 295 mL)
filled with Amberlite XAD 16. First, the main impurities were
removed by washing with aq HCl 10 mM and water (4 column
volume, 1178 mL), then the product was eluted with ethanol : water
3 : 2 (2 column volume, 589 mL). Pure fractions were collected
and evaporated to dryness (2.5 g, 73% yield). Following the
procedure described above, treatment of the aldol adduct with
H2 in the presence of Pd/C, furnished the 1,4-dideoxy-1,4-imino-
L-arabinitol (LAB) (600 mg), which was not further purified (see
ESI†). Physical and spectral data matched those obtained in a
previous work.19
18 R. Wischnat, R. Martin, S. Takayama and C.-H. Wong, Bioorg. Med.
Chem. Lett., 1998, 8, 3353–3358.
19 L. Espelt, T. Parella, J. Bujons, C. Solans, J. Joglar, A. Delgado and P.
Clape´s, Chem.–Eur. J., 2003, 9, 4887–4899.
20 L. Espelt, J. Bujons, T. Parella, J. Calveras, J. Joglar, A. Delgado and P.
Clape´s, Chem.–Eur. J., 2005, 11, 1392–1401.
21 J. Calveras, M. Egido-Gaba´s, L. Go´mez, J. Casas, T. Parella, J. Joglar,
J. Bujons and P. Clape´s, Chem.–Eur. J., 2009, 15, 7310–7328.
22 X. Garrabou, L. Gomez, J. Joglar, S. Gil, T. Parella, J. Bujons and P.
Clape´s, Chem.–Eur. J., 2010, 16, 10691–10706.
23 X. Garrabou, J. Joglar, T. Parella, J. Bujons and P. Clape´s, Adv. Synth.
Catal., 2011, 353, 89–99.
24 M. Schu¨mperli, R. Pellaux and S. Panke, Appl. Microbiol. Biotechnol.,
2007, 75, 33–45.
25 M. Sugiyama, Z. Y. Hong, L. J. Whalen, W. A. Greenberg and C.-H.
Wong, Adv. Synth. Catal., 2006, 348, 2555–2559.
26 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.
27 R. Schoevaart, F. vanRantwijk and R. A. Sheldon, J. Org. Chem., 2001,
66, 4559–4562.
28 According to the work of Sugiyama et al. (see ref. 25) the model aldol
reaction reached maximum yield at about 200 mM of sodium borate.
Therefore, this concentration was used through this work.
29 J. Calveras, J. Casas, T. Parella, J. Joglar and P. Clape´s, Adv. Synth.
Catal., 2007, 349, 1661–1666.
30 The cyclic hemiaminal arises from a reversible spontaneous intramolec-
ular addition of the amino group of the carbamate to the ketose group
of the corresponding polyhydroxy N-Cbz-aminoaldol adduct.
Acknowledgements
This work was supported by the Spanish MICINN projects
CTQ2009-07359 and CTQ2009-08328, Generalitat de Catalunya
(2009 SGR 00281), and ESF project COST CM0701. X. Garrabou
acknowledges the CSIC for an I3P predoctoral scholarship.
31 X. Garrabou, J. A. Castillo, C. Gu
e´rard-He´laine, T. Parella, J. Joglar,
M. Lemaire and P. Clape´s, Angew. Chem., Int. Ed., 2009, 48, 5521–
5525.
32 Broad signals were observed when recording the spectra at 298 K
indicating the interchange between the cyclic and acyclic species..
33 R. van den Berg, J. A. Peters and H. van Bekkum, Carbohydr. Res.,
1994, 253, 1–12.
34 S. Chapelle and J.-F. Verchere, Tetrahedron, 1988, 44, 4469–4482.
35 S. Chapelle and J.-F. Verchere, Carbohydr. Res., 1989, 191, 63–70.
36 Although several a,b + a,g-bidentate B2L complexes of monosaccha-
rides have been reported to show 11B NMR signals at ~-19 ppm, these
complexes were observed only at high pH values (>9).
37 A. B. Walls, H. M. Sickmann, A. Brown, S. D. Bouman, B. Ransom,
A. Schousboe and H. S. Waagepetersen, J. Neurochem., 2008, 10, 1–9.
38 N. G. Oikonomakos, C. Tiraidis, D. D. Leonidas, S. E. Zographos, M.
Kristiansen, C. U. Jessen, L. Norskov-Lauritsen and L. Agius, J. Med.
Chem., 2006, 49, 5687–5701.
39 K. Fosgerau, N. Westergaard, B. Quistorff, N. Grunnet, M. Kristiansen
and K. Lundgren, Arch. Biochem. Biophys., 2000, 380, 274–284.
40 T. B. Mercer, S. F. Jenkinson, B. Bartholomew, R. J. Nash, S. Miyauchi,
A. Kato and G. W. J. Fleet, Tetrahedron: Asymmetry, 2009, 20, 2368–
2373.
41 Experiments conducted at 50 mM and 200 mM of triethanolamine
(TEA) or 200 mM of sodium borate at pH 7 or 7.4 at 25 ◦C gave
similar vo values: 71 and 77 mmol of aldol adduct formed h-1 mg-1,
respectively.
42 W.-D. Fessner and C. Walter, Top. Curr. Chem., 1997, 184, 97–194.
43 T. Suau, G. Alvaro, M. D. Benaiges and J. Lopez-Santin, Biotechnol.
Bioeng., 2006, 93, 48–55.
References
1 H. H. Wasserman, S. F. Martin and Y. Yamamoto, Stereoselective
Carbon–Carbon Bond Forming Reactions, Elsevier, Oxford, U.K., 1999.
2 R. Mahrwald, Modern Aldol Reactions, Vol. 1: Enolates, Organocataly-
sis, Biocatalysis and Natural Product Synthesis, Wiley-VCH, Weinheim,
2004.
3 R. Mahrwald, Modern Aldol Reactions, Vol. 2: Metal Catalysis, Wiley-
VCH, Weinheim, 2004.
4 B. Schetter and R. Mahrwald, Angew. Chem., Int. Ed., 2006, 45, 7506–
7525.
5 D. Enders, M. Voith and A. Lenzen, Angew. Chem., Int. Ed., 2005, 44,
1304–1325.
6 M. Markert and R. Mahrwald, Chem.–Eur. J., 2008, 14, 40–48.
7 B. Erni, C. Siebold, S. Christen, A. Srinivas, A. Oberholzer and U.
Baumann, Cell. Mol. Life Sci., 2006, 63, 890–900.
8 S. S. V. Ramasastry, K. Albertshofer, N. Utsumi, F. Tanaka and C. F.
Barbas III, Angew. Chem., Int. Ed., 2007, 46, 5572–5575.
9 S. S. V. Ramasastry, K. Albertshofer, N. Utsumi and C. F. Barbas, Org.
Lett., 2008, 10, 1621–1624.
10 J. A. Castillo, J. Calveras, J. Casas, M. Mitjans, M. P. Vinardell, T.
Parella, T. Inoue, G. A. Sprenger, J. Joglar and P. Clape´s, Org. Lett.,
2006, 8, 6067–6070.
11 M. Sugiyama, Z. Hong, P. H. Liang, S. M. Dean, L. J. Whalen, W.
A. Greenberg and C.-H. Wong, J. Am. Chem. Soc., 2007, 129, 14811–
14817.
12 A. L. Concia, C. Lozano, J. A. Castillo, T. Parella, J. Joglar and P.
Clape´s, Chem.–Eur. J., 2009, 15, 3808–3816.
44 S.-H. Jung, J.-H. Jeong, P. Miller and C.-H. Wong, J. Org. Chem., 1994,
59, 7182–7184.
45 L. Vidal, O. Durany, T. Suau, P. Ferrer, M. D. Benaiges and G. Caminal,
J. Chem. Technol. Biotechnol., 2003, 78, 1171–1179.
46 O. Durany, G. Caminal, C. de Mas and J. Lopez-Santin, Process
Biochem., 2004, 39, 1677–1684.
8436 | Org. Biomol. Chem., 2011, 9, 8430–8436
This journal is
The Royal Society of Chemistry 2011
©