CHEMBIOCHEM
COMMUNICATIONS
pose significant synthetic challenges. To address this issue,
a number of chemical methods have been developed for tre-
halose desymmetrization/regioselective hydroxy group manip-
[
10]
[11]
ulation and the formation of 1,1-a,a-glycosidic linkages.
Collectively, these methods are versatile and important for pre-
paring various types of trehalose analogues, particularly those
with complex structures bearing multiple sites of modification,
[
12]
such as SL-1 analogues. However, chemical syntheses of tre-
halose analogues are usually lengthy and low-yielding, which
motivated us to develop a complementary approach. Here, we
describe a chemoenzymatic method for the rapid and efficient
synthesis of trehalose analogues that is both accessible to
non-chemists and well-suited to generating trehalose-based
probes to investigate mycobacteria.
Scheme 2. Synthesis of trehalose from Glc and UDP-Glc by TreT from
T. tenax.
HEPES buffer (pH 7.4) containing 10 mm Glc analogue, 40 mm
UDP-Glc, 20 mm MgCl , and 9.8 mm TreT (reaction volume
2
50 mL). The reactions were incubated at 708C with gentle shak-
ing for 1 h, quenched by addition of cold acetone, and ana-
lyzed by HPLC and high-resolution ESI mass spectrometry (Fig-
ures S1–S18 in the Supporting Information).
We first sought to identify a trehalose biosynthesis pathway
that could be adapted for the chemoenzymatic synthesis of
trehalose analogues. Several natural pathways exist that could,
[
1]
in principle, be exploited for synthetic applications. Two en-
zymes are particularly fitting for this purpose because they
employ simple and readily available glucose (Glc) analogues as
substrates: trehalose phosphorylase (TreP), which catalyzes the
cleavage of trehalose-6-phosphate to release Glc and b-d-Glc-
The Glc analogues that we evaluated were remarkably well
tolerated by TreT (Table 1). In most cases, the corresponding
trehalose analogue products were generated in excellent yield
after only 1 h, which highlighted the efficiency, rapidity, and
generality of the method. Fluoro-, deoxy-, azido-, and stereo-
chemical modifications of the Glc 2-, 3-, and 6-positions were
generally accepted, except for azido substitution at the 2-posi-
tion and inversion of the 3-OH group. Glc analogues bearing
4-position alterations were poor substrates (n.d.–26% yield),
indicating a strict specificity at this position. Nonetheless, 4-po-
sition-modified products were observed in low yield; this, sug-
gests that mutation of the active site of TreT might allow for
improved substrate tolerance. Finally, 5-thio-d-glucose, the
sole 5-position-modified Glc analogue that was tested, was
converted to the novel compound 5-thio-trehalose in quantita-
tive yield. The 5-thio-d-glucose reaction was performed in
10 mm dithiothreitol (DTT), indicating that TreT retains activity
in the presence of DTT, which can facilitate access to other
thio-modified trehalose analogues.
1-phosphate; and trehalose synthase (TreT), which catalyzes
the formation of trehalose directly from Glc and uridine di-
phosphate glucose (UDP-Glc). TreP was recently explored for
its ability to produce trehalose analogues for biopreservation
applications by running the enzyme in the reverse “synthetic”
direction, but substrate specificity tests were limited to natural
sugars, and no yields or product characterization data were
[
13]
reported. Further, the donor in these reactions, b-d-Glc-1-
phosphate, is prohibitively expensive and difficult to synthe-
size. Studies focused on the characterization of TreT from Ther-
motoga maritima and Pyrococcus horikoshii showed that these
enzymes could process a few monosaccharide acceptors other
than Glc—mannose, galactose, and fructose were tested—
[
14]
albeit with low conversion and long reaction times. In addi-
tion, a multistep chemoenzymatic approach to synthesizing 2-
deoxy-2-fluoro-a,a’-trehalose (2-fluoroTre) has been reported,
but it required the use of three enzymes and a Glc-6-phos-
phate analogue intermediate, thus limiting synthetic efficiency
Selected reactions were run on a semi-preparative scale (5–
10 mg) to evaluate the scalability of the method and to con-
firm product structure by NMR spectroscopy. Semi-preparative
reactions were performed as described above (reaction
volume: 1.5–2.0 mL), and the products were readily purified by
silica gel chromatography. Consistent with the small-scale re-
sults, 2-fluoroTre, 2-deoxyTre, 3-fluoroTre, 6-TreAz, and 5-thio-
Tre were obtained in isolated yields of 92–97%, which demon-
[
5a]
and versatility. Overall, the reported results are encouraging,
but the use of enzymes for the practical and efficient synthesis
of trehalose analogues has remained underdeveloped.
We decided to pursue trehalose analogue synthesis using
TreT from the hyperthermophile Thermoproteus tenax
1
strated that scale-up of the TreT reaction is feasible. H and
[
15]
13
(Scheme 2).
In addition to its simple acceptor and donor
C NMR analysis established the product structures, including
substrates and excellent thermostability, TreT from T. tenax
the assignment of 1,1-a,a-stereochemistry for newly formed
glycosidic bonds (Supporting Information). No regio- or stereo-
isomers of the desired trehalose analogue products were de-
tected.
[
15]
cannot degrade trehalose, which distinguishes it from TreT
enzymes in other organisms. We hypothesized that these char-
acteristics would make TreT from T. tenax ideal for synthetic
applications. To initiate the study, we expressed and purified
As discussed above, previously reported efforts to assess the
substrate specificity of trehalose-synthesizing enzymes have
primarily been limited to natural sugars, for example, Glc epi-
mers, for the purpose of developing new compounds for the
[
15]
TreT from E. coli as reported and screened it for reactivity by
using a panel of monofunctionalized Glc analogues. The Glc
analogues that were tested contained fluoro-, deoxy-, azido-,
and stereochemical modifications occurring at all positions of
the sugar ring, which afforded a systematic evaluation of TreT
substrate promiscuity. Reactions were performed in 50 mm
[13]
preservation of biomaterials. In the present work, we consid-
erably expanded this focus by emphasizing the incorporation
of unnatural functionalities that are useful for glycobiology re-
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 5
&
2
&
ÞÞ
These are not the final page numbers!