Deoxyfluoro-d-trehalose (FDTre) analogues as potential PET probes for imaging mycobacterial infection

Article abstract of DOI:10.1039/c6ob01734g

Mycobacterium tuberculosis, the etiological agent of human tuberculosis, requires the non-mammalian disaccharide trehalose for growth and virulence. Recently, detectable trehalose analogues have gained attention as probes for studying trehalose metabolism and as potential diagnostic imaging agents for mycobacterial infections. Of particular interest are deoxy-[¹⁸F]fluoro-d-trehalose (¹⁸F-FDTre) analogues, which have been suggested as possible positron emission tomography (PET) probes for in vivo imaging of M. tuberculosis infection. Here, we report progress toward this objective, including the synthesis and conformational analysis of four non-radioactive deoxy-[¹⁹F]fluoro-d-trehalose (¹⁹F-FDTre) analogues, as well as evaluation of their uptake by M. smegmatis. The rapid synthesis and purification of several ¹⁹F-FDTre analogues was accomplished in high yield using a one-step chemoenzymatic method. Conformational analysis of the ¹⁹F-FDTre analogues using NMR and molecular modeling methods showed that fluorine substitution had a negligible effect on the conformation of the native disaccharide, suggesting that fluorinated analogues may be successfully recognized and processed by trehalose metabolic machinery in mycobacteria. To test this hypothesis and to evaluate a possible route for delivery of FDTre probes specifically to mycobacteria, we showed that ¹⁹F-FDTre analogues are actively imported into M. smegmatis via the trehalose-specific transporter SugABC-LpqY. Finally, to demonstrate the applicability of these results to the efficient preparation and use of short-lived ¹⁸F-FDTre PET radiotracers, we carried out ¹⁹F-FDTre synthesis, purification, and administration to M. smegmatis in 1 hour.

Full text of DOI:10.1039/c6ob01734g

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Deoxyfluoro-D-trehalose (FDTre) analogues as
potential PET probes for imaging mycobacterial
infection†
Cite this: DOI: 10.1039/c6ob01734g
Sarah R. Rundell,‡a Zachary L. Wagar,‡a Lisa M. Meints,a Claire D. Olson,a
Mara K. O’Neill,a Brent F. Piligian,a Anne W. Poston,a Robin J. Hood,a
Peter J. Woodruffb and Benjamin M. Swarts*a
Mycobacterium tuberculosis, the etiological agent of human tuberculosis, requires the non-mammalian
disaccharide trehalose for growth and virulence. Recently, detectable trehalose analogues have gained
attention as probes for studying trehalose metabolism and as potential diagnostic imaging agents for
mycobacterial infections. Of particular interest are deoxy-[18F]fluoro-D-trehalose (18F-FDTre) analogues,
which have been suggested as possible positron emission tomography (PET) probes for in vivo imaging of
M. tuberculosis infection. Here, we report progress toward this objective, including the synthesis and con-
formational analysis of four non-radioactive deoxy-[19F]fluoro-D-trehalose (19F-FDTre) analogues, as well
as evaluation of their uptake by M. smegmatis. The rapid synthesis and purification of several 19F-FDTre
analogues was accomplished in high yield using a one-step chemoenzymatic method. Conformational
analysis of the 19F-FDTre analogues using NMR and molecular modeling methods showed that fluorine
substitution had a negligible effect on the conformation of the native disaccharide, suggesting that fluori-
nated analogues may be successfully recognized and processed by trehalose metabolic machinery in
mycobacteria. To test this hypothesis and to evaluate a possible route for delivery of FDTre probes specifi-
cally to mycobacteria, we showed that 19F-FDTre analogues are actively imported into M. smegmatis via
the trehalose-specific transporter SugABC-LpqY. Finally, to demonstrate the applicability of these results
to the efficient preparation and use of short-lived 18F-FDTre PET radiotracers, we carried out 19F-FDTre
synthesis, purification, and administration to M. smegmatis in 1 hour.
Received 16th June 2016,
Accepted 15th August 2016
DOI: 10.1039/c6ob01734g
www.rsc.org/obc
and treat.2,3 In 2014, an estimated 5% of all new TB cases—
nearly half a million—were MDR-TB, which had a case fatality
Introduction
Mycobacterium tuberculosis, the organism that causes tubercu- rate approaching 50%.4 To combat drug-susceptible and drug-
losis (TB), infects 2 billion people worldwide, every year resistant TB alike, new drugs and diagnostics—as well as new
causing 10 million active cases of TB and killing 1.5 million tools that support their development—are needed.
people.1 Although drug-susceptible TB is curable, treatment
Nuclear imaging is emerging as a potentially transformative
requires multiple drugs delivered over the course of technology for TB research and clinical applications. As
6–9 months. Non-adherence to treatment has led to the emer- recently discussed by Johnson and co-authors, nuclear
gence of multi- and extensively-drug-resistant TB (MDR- and imaging modalities such as positron emission tomography
XDR-TB), which are extremely difficult and costly to diagnose (PET) may offer the ability to rapidly and noninvasively diag-
nose TB and monitor TB disease progression and response to
treatment, while also allowing correlation of various disease
aDepartment of Chemistry and Biochemistry, Central Michigan University, Mount
characteristics with patient anatomy in real time and in three
dimensions.5 Thus, in principle, nuclear imaging can provide
Pleasant, MI 48859, USA. E-mail: ben.swarts@cmich.edu
bDepartment of Chemistry, University of Southern Maine, Portland, ME, USA
rich information on TB disease state that is complementary to
†Electronic supplementary information (ESI) available: Supplementary figures,
diagnostic methods based on analysis of sputum samples by
microscopy, culture, or nucleic acid detection.
experimental procedures, and NMR spectra. See DOI: 10.1039/c6ob01734g
‡These authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
To date, nuclear imaging approaches for TB have primarily
focused on the use of radiotracers that provide a readout of the
host inflammatory response to infection. In particular,
2-deoxy-2-[18F]fluoro-D-glucose (18F-2-FDG) has been used
extensively for PET imaging of inflammation associated with
TB lesions.5,6 However, 18F-2-FDG is not well-suited to differen-
tiating TB from tumors and other metabolically active tissues
due to non-specific uptake.5,6 Radiolabeled analogues of anti-
mycobacterial compounds, such as isoniazid, have been used
for drug biodistribution studies,7,8 and they may also have
value for TB detection applications.8 Radiotracers that specifi-
cally label bacteria may hold the most promise for in vivo
imaging. Indeed, recent reports have demonstrated the poten-
tial of carbohydrate-based PET probes for imaging various
types of bacterial infection,9–11 but there has been limited pro-
gress for mycobacteria. Jain and co-workers imaged TB with 1-
(2′-deoxy-2′-fluoro-beta-D-arabinofuranosyl)-5-[125I]iodouracil
(
125I-FIAU).12 However, an engineered strain of M. tuberculosis
expressing thymidine kinase was required to permit bacteria-
selective uptake of 125I-FIAU, so this approach is limited to
experimentally infected animals.12 Therefore, as noted by
Johnson and co-authors,5 there is an urgent need for new
mycobacteria-specific radiotracers that enable direct visualiza-
tion of wild-type M. tuberculosis during infection.
Trehalose analogues modified with detectable tags have
recently been explored as mycobacteria-specific imaging
probes, and they are promising candidates for nuclear imaging
applications. Trehalose (1) is a C2-symmetric, non-mammalian
disaccharide that is essential for mycobacterial growth and
virulence due to its central role in cell wall biosynthesis.13,14
Fig. 1 Conserved trehalose metabolic pathways in mycobacteria.
Cytoplasmic trehalose (1) is converted to trehalose monomycolate
(TMM), which is then exported and processed by the Ag85 complex to
generate trehalose dimycolate (TDM) and arabinogalactan–mycolate
(AGM). Released trehalose is recycled by SugABC-LpqY. Trehalose-
based probes (red text) administered to mycobacteria can cross the
mycomembrane, likely via a porin(s), and incorporate into the cell via
SugABC-LpqY-mediated uptake and/or via Ag85-mediated incorpor-
ation into TMM, TDM, and/or AGM.
Trehalose metabolic pathways that are conserved in mycobac- image live M. tuberculosis within an infected macrophage,
teria are briefly discussed here to illustrate how they can be showing the potential of trehalose-based probes for myco-
targeted for imaging probe development (Fig. 1). Biosynthesis bacteria-specific imaging within the context of a mammalian
of trehalose occurs from either glucose or polymeric α-glucan host system.29 Subsequently, the Bertozzi group reported a
through metabolic cycles involving the enzymes OtsAB/treha- series of azide-modified trehalose (TreAz) analogues (3–6), all
lase or TreYZ/TreS, respectively.15–17 Intracellular trehalose is of which could be metabolically incorporated into trehalose
converted by Pks13 to the glycolipid trehalose monomycolate glycolipids and subsequently detected by click chemistry
(TMM),18 which, upon translocation across the plasma mem- reaction with alkyne-modified fluorophores.30 Presumably
brane by MmpL3,19–21 is utilized by the Ag85 complex to con- due to the smaller size of the azide, three of the four
struct the mycobacterial outer membrane (also referred to as reported TreAz analogues were taken up by the trehalose
the mycomembrane).22–24 Ag85 has a dual use for TMM, using transporter SugABC-LpqY and incorporated into trehalose
it to generate: (i) arabinogalactan–mycolate (AGM), which is glycolipids via an intracellular route.30 Our group recently deve-
the foundation of the mycomembrane; and (ii) trehalose loped a class of alkyne-modified TMM (AlkTMM) analogues
dimycolate (TDM),
a
mycomembrane-resident glycolipid (7 and 8), which could deliver alkyne-modified mycolic acid
which is involved in pathogenesis.25–27 Free trehalose that is mimics to either TDM or AGM via the action of Ag85, thus
released during these Ag85-mediated processes is recycled tagging the lipid portions of these molecules.31 Together,
back into the cell by the trehalose-specific transporter these studies revealed two routes by which trehalose-based
SugABC-LpqY.28
probes can incorporate detectable tags into mycobacteria: (i)
These pathways, which have only recently been fully eluci- anchoring to the cell wall via periplasmic Ag85-mediated
dated, provide opportunities to rationally design trehalose- mycoloylation; (ii) accumulation in the cytoplasm via uptake
based probes for the detection of mycobacteria (Fig. 2A). The through SugABC-LpqY, with possible downstream incorpor-
Barry and Davis groups reported the first trehalose-based ation into trehalose glycolipids. FITC-Tre, TreAz, and
imaging probe, a fluorescein-modified ketoside trehalose AlkTMM analogues enabled cellular imaging and provided
analogue (FITC-Tre, 2), which was capable of metabolically valuable information about targeting trehalose metabolic
labeling the mycobacterial cell wall via Ag85-mediated in- pathways, but their reliance on optical imaging for detection
corporation into trehalose glycolipids.29 FITC-Tre was used to and, in the case of TreAz and AlkTMM analogues, a second-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 2 Trehalose-based probes for imaging mycobacteria. (A) Previously reported trehalose analogues modified with fluorescein (FITC-Tre) or click
chemistry tags (TreAz and AlkTMM) have been used for optical imaging of intact mycobacterial cells. (B) This report describes the rapid synthesis,
characterization, and mycobacterial uptake of non-radioactive deoxy-[19F]fluoro-D-trehalose (19F-FDTre) analogues 9–12. This study informs the
future development of radioactive deoxy-[18F]fluoro-D-trehalose (18F-FDTre) analogues for PET imaging.
ary labeling step, are not ideally suited to in vivo imaging
applications.
Synthetic approaches to FDTre
analogues
The concept of a trehalose-based nuclear imaging probe
was first introduced by Barry and Davis in 2011.29 The most
attractive candidate for this purpose would be a 18F-modified Trehalose possesses C2-symmetry and a 1,1-α,α-glycosidic bond,
trehalose analogue, which, in principle, could be used for both of which present unique challenges to analogue synthesis.
specific radiolabeling of wild-type M. tuberculosis within an In general, one of two approaches is used to prepare trehalose
infected host to allow in vivo PET imaging of TB. Two major analogues by chemical synthesis. For simpler targets, such as
challenges must be overcome to accomplish this objective. monofunctionalized analogues, trehalose desymmetrization
First, because 18F has a short half-life (t1/2 = 110 min), the syn- and hydroxyl manipulation steps are typically used to modify
thesis and purification of the 18F-labeled trehalose analogue the disaccharide. For more complex structures, for example sul-
must be rapid. Second, one or both of the mycobacteria- folipid analogues, specialized glycosylation methods have been
specific metabolic pathways discussed above (or perhaps as-yet developed to enable 1,1-α,α-stereoselective linkage of two pro-
undiscovered trehalose-related pathways) must be capable of perly functionalized glucose analogues. Although these
incorporating the unnatural fluorine-modified trehalose ana- approaches have been effectively applied to the synthesis of a
logue into live mycobacterial cells, in vitro and during host variety of trehalose analogues, as recently reviewed by Kulkarni
infection (i.e., in a TB animal model or a human patient with and co-workers,32,33 they are often lengthy and low-yielding, and
TB). Herein, we present progress toward both of these goals. synthetic expertise is required for their execution.
We report on the synthesis and conformational analysis of
A brief survey of prior chemical syntheses of FDTre ana-
four non-radioactive deoxy-[19F]fluoro-D-trehalose (19F-FDTre) logues exemplifies these disadvantages. Of the 19F-FDTre ana-
analogues (9–12, Fig. 2B), three of which were prepared using logues reported herein, only 19F-6-FDTre (12) has previously
a rapid one-step chemoenzymatic method that is adaptable to been chemically synthesized. In two independent reports,
the synthesis of radioactive deoxy-[18F]fluoro-D-trehalose 19F-6-FDTre was accessed via trehalose desymmetrization strat-
(18F-FDTre) analogues. We also demonstrate that three of the egies, which required 5 or 8 steps from trehalose and pro-
four synthesized 19F-FDTre analogues were taken up by ceeded in 6 or 13% overall yield, respectively.29,34 Critically,
M. smegmatis, a model organism, via the SugABC-LpqY treha- the final two-step fluorination–deprotection sequences in
lose transporter, thus establishing a potential route for label- these syntheses had long reaction times (32–36 h) and moder-
ing of mycobacteria within a host organism. Together, these ate yields (39–48%).29,34 In the same reports, syntheses with
studies help to lay the groundwork for the future production comparable lengths and yields were used to access an epimer
and use of 18F-FDTre PET probes for in vivo TB imaging and of 19F-4-FDTre containing an axial fluoro group at the 4′-posi-
potentially other applications.
tion.29,34 Methyl ketoside derivatives of trehalose bearing
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
fluoro groups at the 2- and 3-positions have also been pre-
pared29 using established stereoselective glycosylation
methods.35–37 These approaches, while strategically elegant,
are disadvantaged by incomplete 1,1-α,α-stereoselectivity, mul-
tiple protection and deprotection steps, and the use of custom
monosaccharide building blocks. Overall, currently available
methods for the chemical synthesis of trehalose analogues
lack the speed, efficiency, and convenience desired for in-
corporating short-lived 18F into trehalose for PET applications.
Chemoenzymatic synthesis is an attractive alternative
approach to preparing trehalose analogues, particularly if sen-
sitive functionalities, such as radionuclides, must be quickly
incorporated into the product. In principle, naturally occurring
trehalose-synthesizing enzymes offer the ability to stereoselec-
tively forge a 1,1-α,α-glycosidic bond between two unprotected
glucose derivatives, thus bypassing the slow and inefficient
chemical desymmetrization and glycosylation methods dis-
cussed above.38 Ideally, an enzyme or series of enzymes would
be capable of rapidly converting deoxy-[18F]fluoro-D-glucose
(18F-FDG) analogues into 18F-FDTre analogues for PET imaging
of mycobacterial infections. For example, 18F-2-FDG, which is
commercially available and widely used in the clinic for PET
imaging of tumors,39 could be enzymatically transformed into
18F-2-FDTre for TB imaging applications.
In 2011, Barry and Davis reported the first chemoenzymatic
synthesis of the non-radioactive version of 18F-2-FDTre, 19F-2-
FDTre (9), using a three-enzyme system modeled after the
OtsAB trehalose biosynthetic pathway in Escherichia coli.29 The
E. coli OtsAB pathway uses trehalose-6-phosphate synthase
(OtsA) to catalyze a glycosylation reaction between glucose-6-
phosphate and uridine diphosphate (UDP)-glucose, generating
trehalose-6-phosphate, which is subsequently dephosphory-
lated by trehalose-6-phosphate phosphatase (OtsB) to give tre-
halose.40,41 In adapting this pathway to the laboratory
synthesis of 19F-2-FDTre, the authors used yeast hexokinase to
first transform 19F-2-FDG into 19F-2-FDG-6-phosphate, which
was then processed sequentially by OtsA and alkaline phos-
phatase (in lieu of OtsB) to give 19F-2-FDTre (9).29 This work
established the concept of using enzymes to access a FDTre
analogue from the corresponding FDG analogue and should
be adaptable to the synthesis of 18F-2-FDTre from 18F-2-FDG.
However, in addition to being applied to only a single reaction
example, this approach is limited by the need for three enzy-
matic steps, the inability to use 6-position-modified glucose
substrates, and the use of overnight reactions.
In 2014, we reported a one-step chemoenzymatic method
for the synthesis of trehalose analogues that addressed
many of the limitations discussed above.42 Our method
exploited the unidirectional trehalose biosynthetic pathway in
Thermoproteus tenax, which is a hyperthermophilic bacterium
that utilizes a trehalose synthase (TreT) enzyme to catalyze the
direct formation of trehalose from glucose and UDP-glucose
(reaction scheme shown in Fig. 3A).43 We showed that re-
combinant T. tenax TreT, which was expressed and purified from
Fig. 3 Optimization of TreT activity for FDTre synthesis. (A) Scheme for
TreT-catalyzed one-step synthesis of 19F-FDTre analogues from 19F-FDG
analogues. A luminescence-based UDP detection assay was used to
assess enzyme activity during reaction optimization and substrate
testing. Relative enzyme activity was assessed for varying (B) tempera-
ture, (C) pH, (D) salt concentration, and (E) substrate. Reactions were
performed with 1 µg TreT at 23 °C in 50 mM Tris-HCl buffer, 10 mM
acceptor substrate, 0.4 mM UDP-glucose, 20 mM MgCl2, and 200 mM
NaCl, unless noted otherwise. For pH dependence experiments, the fol-
lowing buffers were used: pH 2, sodium citrate; pH 4 sodium acetate,
pH 7.4, sodium phosphate; pH, 8 Tris-HCl; pH 10, sodium carbonate.
generating the corresponding trehalose analogues quickly
(60 min) and in high yield (up to >99%, as determined by
HPLC) in a single step.42 Because this reaction exhibited prom-
ising preliminary results on 19F-FDG substrates and possessed
high efficiency, substrate flexibility, and speed, we reasoned
that it would be an ideal platform for the development of
18F-FDTre PET imaging probes.
Results and discussion
Optimization of TreT activity for 19F-FDTre synthesis
E. coli, could tolerate a variety of different glucose analogues In our previous report,42 four commercial 19F-FDG analogues
(e.g., azido, deoxy, fluoro, and stereochemical modifications), were screened as substrates for T. tenax TreT. Microscale
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
enzymatic reactions (50 µL) were performed in 50 mM HEPES activity of 135 14%. This result is of significance because, as
buffer (pH 7.4) using 10 µM TreT, 10 mM 19F-FDG, 40 mM mentioned above, 18F-2-FDG is a widely used and commer-
UDP-glucose, 300 mM NaCl, and 20 mM MgCl2, with incu- cially available PET imaging agent that could be readily con-
bation and shaking at 70 °C for 60 min. After quenching with verted by TreT to 18F-2-FDTre for PET imaging applications.
cold acetone, an HPLC assay employing an aminopropyl 19F-3- and 19F-6-FDG also showed high activities of 86 9%
analytical column and refractive index detection was used to and 89 3%, respectively. Although 18F-3- and 18F-6-FDG are
determine reaction yields. Complete conversion of 19F-2-, not commercially available, their radiosyntheses have been
19F-3-, and 19F-6-FDG analogues to the corresponding reported,44,45 so it would be feasible to converge these
19F-FDTre analogues was observed, whereas the 19F-4-FDG methods with the TreT reaction to access radioactive 18F-3- and
reaction did not proceed appreciably under the conditions 18F-6-FDTre chemoenzymatically. 19F-4-FDG showed no activity
tested (nor did other glucose analogues with modifications at in this assay, confirming that chemoenzymatic synthesis of
the 4-position, likely due to a strict requirement for an equa- 19F-4-FDTre using wild-type T. tenax TreT is not achievable.
torial hydroxyl group at this site). Based on these promising
Synthesis and purification of 19F-FDTre analogues
preliminary results, here we sought to optimize the TreT reac-
With TreT’s substrate tolerance for 19F-FDG analogues estab-
tion for the production of FDTre analogues. To facilitate these
efforts, we replaced the slow, low-sensitivity HPLC assay with a
rapid and sensitive luminescence-based glycosytransferase
assay. This assay, which detects UDP released from the UDP-
lished and with some reaction parameters optimized, we pro-
ceeded to the synthesis of 19F-FDTre analogues (Fig. 4). 19F-2-,
19F-3-, and 19F-6-FDTre (9, 10, and 12) were prepared by enzy-
sugar donor upon reaction, was used to optimize TreT reaction
matic synthesis from the corresponding commercially avail-
conditions and assess relative enzyme activities on 19F-FDG
substrates.
able 19F-FDG analogues (14, 15, and 17) (Fig. 1A). The
enzymatic reactions were performed on a larger scale (4 mL) in
50 mM Tris-HCl buffer (pH 8.0) using 10 µM TreT, 300 mM
Using the natural substrates glucose (13) and UDP-glucose,
the luminescence assay was used to evaluate the dependence
of TreT activity on temperature, pH, buffer, and NaCl concen-
tration (Fig. 3A–D). The growth temperature for T. tenax in its
native environment is 86 °C and TreT is heat-stable,43 so it was
not surprising that an elevated reaction temperature of 70 °C
was optimal for TreT activity (Fig. 3B). High enzyme activity
was also observed at 90 °C, which allows for the use of TreT at
higher temperatures to decrease reaction times if needed.
Different buffers spanning a range of pH values (2–10.4) were
also tested (Fig. 3C). With the highest activity occurring at pH
7–8, pH 8.0 was chosen since it appeared to provide better
protein stability during storage. At pH 7–8, the enzyme had
comparable activity in various buffers, including HEPES,
sodium phosphate, and Tris-HCl. We used 50 mM Tris-HCl
due to its easy removal during product purification. According
to the luminescence assay, TreT activity decreased as NaCl con-
centration increased above 100 mM (Fig. 3D). However, we did
not observe significantly lower reaction conversions in high
salt conditions by HPLC or TLC,§ and low salt conditions led
to protein precipitation. Therefore, NaCl was maintained at
300 mM to achieve a balance between enzyme activity and
stability.
Next, the relative activities of TreT on 19F-FDG analogues
(14–17) relative to the natural acceptor glucose (13) were deter-
mined using the luminescence assay (Fig. 3E). Consistent with
the reaction yields obtained from our earlier HPLC assay, the
luminescence assay showed that TreT had high activity on
19F-2-, 19F-3-, and 19F-6-FDG. Surprisingly, 19F-2-FDG appeared
to be a better substrate for TreT than glucose, with a relative
Fig. 4 (A) TreT-catalyzed chemoenzymatic synthesis and purification of
19F-2-, 19F-3-, and 19F-6-FDTre (9, 10, and 12) on a semi-preparative
scale. Isolated yields following spin dialysis and ion exchange purifi-
cation steps are shown. (B) Chemical synthesis of 19F-4-FDTre (11).
Representative (C) 1H and (D) 13C NMR spectra for 19F-2-FDTre (see ESI†
for all spectra).
§The observed decrease in luminescence signal at higher NaCl concentrations is
likely not due to decreased TreT activity, but instead due to effects on the
coupled luciferase reaction. The assay manufacturer recommends maintaining
NaCl at ≤200 mM.
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
NaCl, and 20 mM MgCl2. During reaction optimization, we native trehalose within the environment of the mycobacter-
also noted that a 1 : 2 ratio of 19F-FDG : UDP-glucose gave ium. Therefore, minimal distortion of the trehalose structure
yields similar to the previously used 1 : 4 ratio. Therefore, reac- upon deoxyfluorination is desired for this application. While
tions were carried out using 20 mM 19F-FDG and 40 mM UDP- substitution of a hydroxyl group with a smaller fluorine atom
glucose, with incubation and shaking at 70 °C for 60 min. TLC does not have a notable steric impact, the higher electro-
analysis of the reactions showed complete conversion of each negativity of fluorine versus oxygen, as well as its inability to
19F-FDG analogue to the corresponding 19F-FDTre analogue.
act as a hydrogen bond donor, have the potential to modulate
Bearing in mind the speed needed for future production of sugar structure and function. To provide insight into the effect
radioactive 18F-FDTre, we developed a rapid and straight- of monodeoxyfluorination on the structure of trehalose, we
forward method to isolate 19F-FDTre analogues from crude used NMR and molecular modeling to define and compare the
enzymatic mixture. First, the TreT enzyme was removed via solution conformations of unmodified trehalose and the four
spin dialysis employing a 10 kDa molecular weight cutoff synthesized 19F-FDTre analogues.
(MWCO) centrifugal filter unit. Next, the aqueous filtrate was
The structure of native trehalose has been investigated exten-
treated with mixed-bed ion exchange resin to remove all ionic sively due to its biological importance and applications in bio-
species, leaving only the neutral 19F-FDTre product in aqueous preservation. Experimental studies have primarily focused on
solution after filtration of the resin beads. Using this protocol, establishing trehalose’s solution conformation using NMR-
19F-2-, 19F-3-, and 19F-6-FDTre were isolated in yields of 79%, derived intra- and inter-residue dihedral angles, most impor-
74%, and 74%, respectively, each on a ∼20 mg scale. Given tantly those around the glycosidic bond, φH (H1–C1–O1–C1′)
that complete reaction conversion was observed by TLC, it is and ψH (C1–O1–C1′–H1′) (see Fig. 5 for φH and ψH definitions).
likely that the loss of product was due to some binding of When interpreted through Karplus curves, vicinal 3JHH coupling
19F-FDTre product to the resin beads. NMR and ESI-MS were constants provide intra-residue dihedral angles, whereas long-
1
3
used to characterize the 19F-FDTre products (representative H range heteronuclear JCOCH coupling constants provide the
and 13C NMR data are shown for 19F-2-FDTre in Fig. 4C and D). glycosidic dihedral angles φH and ψH. An early NMR study of treha-
3
3
The anomeric JHH coupling constants of all three enzy- lose provided JHH coupling constants that clearly established
matically synthesized 19F-FDTre analogues ranged from 4C1 chair conformations for the identical glucopyranoside
3.5–4.0 Hz, which established 1,1-α,α stereochemistry of the rings.47 Before suitable NMR techniques were available, experi-
newly formed glycosidic bonds. Additionally, H and 19F NMR mental values for φH and ψH were obtained from X-ray crystallo-
1
spectra exhibited the expected peak characteristics for the graphic analysis of trehalose dihydrate (φH = –41°, ψH = –58°)48
fluorinated position of each analogue. Together, the one-step and anhydrous trehalose (φH = –60°, ψH = –59°),49 as well as
synthesis and ion exchange purification procedures are con- from optical rotation studies of trehalose in solution (φH = ψH
=
venient and fast, occurring on a timescale that is sufficient for –60°).50 Later on, several groups used NMR techniques to
3
18F-FDTre synthesis (demonstrated in the section “Protocol for extract trehalose’s glycosidic JCOCH coupling constants, which
the rapid synthesis, purification, and administration of 19F-2- were converted to φH and ψH values using a Karplus-type corre-
FDTre to M. smegmatis as a model for future 18F-FDTre PET lation curve developed by Tvaroška and co-workers (eqn (1)).51
imaging applications”).
3JCOCH ¼ 5:7 cos2ðθÞ ꢀ 0:6 cosðθÞ þ 0:5
ð1Þ
Although 19F-4-FDTre was not accessible through chemo-
enzymatic synthesis with wild-type T. tenax TreT, we decided to
synthesize it chemically so that all four monofluorinated treha-
lose analogues could be systematically evaluated in confor-
mational studies and bacterial uptake experiments (Fig. 4B).
To initiate the synthesis, known trehalose heptabenzoate
derivative 18,46 bearing a free axial 4-position OH group, was
prepared via a reported route requiring 3 steps from trehalose.
Bis(2-methoxyethyl)aminosulfur trifluoride (BAST) was used to
fluorinate the free 4′-position of 18 with inversion of stereo-
chemical configuration, which was followed by NaOMe-mediated
debenzoylation to give 19F-4-FDTre (11) in 73% yield from 18
Using this approach, the groups of Batta (3JCOCH = 3.3 Hz;
φH = ψH = −41°),52 Jiménez-Barbero (3JCOCH = 3.0 Hz; φH = ψH
=
1
over two steps. H NMR spectroscopic analysis of 19F-4-FDTre
showed H4′ chemical shift (δ 4.36 ppm), splitting pattern
(doublet of triplets), and coupling constants (3JHH = 9.5 Hz,
2JHF = 51.5 Hz) that were consistent with the presence of an
equatorial fluorine atom at the 4′-position.
Conformational analysis of 19F-FDTre analogues
The success of 18F-FDTre analogues as mycobacteria-specific
Fig. 5 Definition of the glycosidic dihedral angles φH and ψH for treha-
PET imaging probes will depend on their ability to mimic lose and 19F-FDTre analogues.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 1 Experimentala and theoretical glycosidic dihedral angles for trehalose and 19F-FDTre analogues
Experimental valuesa
Theoretical values
φH b; 3JCOCH
ψH b; 3JCOCH
φH; 3JCOCH
ψH; 3JCOCH
Ref.
b
b
Compound
Trehalose (1)
−41°; 3.3 Hz
−44°; 3.0 Hz
−42°; 3.2 Hz
−41°; 3.3 Hz
−44°; 3.0 Hz
−42°; 3.2 Hz
49
50
51
44
52
53
−50°; 2.5 Hz
−38°; 3.6 Hz
−50°; 2.5 Hz
−45°; 2.9 Hz
−54°; 2.1 Hz
−45°; 2.9 Hz
−52°; 2.3 Hz
−52°; 2.3 Hz
−52°; 2.3 Hz
−51°; 2.4 Hz
−52°; 2.3 Hz
−50°; 2.5 Hz
−38°; 3.6 Hz
−50°; 2.5 Hz
−45°; 2.9 Hz
−54°; 2.1 Hz
−45°; 2.9 Hz
−52°; 2.3 Hz
−52°; 2.3 Hz
−52°; 2.3 Hz
−51°; 2.4 Hz
−52°; 2.3 Hz
54
c
−44°; 3.0 Hz
−41°; 3.3 Hz
−43°; 3.1 Hz
−44°; 3.0 Hz
−43°; 3.1 Hz
−44°; 3.0 Hz
−40°; 3.4 Hz
−41°; 3.3 Hz
−44°; 3.0 Hz
−42°; 3.2 Hz
19F-2-FDTre (9)
19F-3-FDTre (10)
19F-4-FDTre (11)
19F-6-FDTre (12)
c
c
c
c
a Only NMR-determined values are shown. b Calculated using the Karplus equation of Tvaroška (eqn (1)).51 c This study.
−44°),53 and Cheetham (3JCOCH = 3.2 Hz; φH = ψH = −42°)54 well within the φH and ψH ranges for unmodified trehalose
obtained results that were in close agreement (Table 1). In that were established earlier by others and confirmed by our
addition, theoretical values for φH and ψH of trehalose, calcu- group in this study. Applying the experimentally determined
lated using various computational methods, ranged from −38° φH and ψH dihedral angles as constraints, the lowest energy
to −54° (Table 1).47,52,53,55–57 Altogether, the experimental and conformers of trehalose and the 19F-FDTre analogues were re-
theoretical studies were generally in good accord with one calculated using MM3* and presented as a superposition in
another, and indicated that trehalose has a relatively rigid Fig. 6B. Taken together, the NMR and modeling data pre-
structure in aqueous solution.
sented here demonstrate that monodeoxyfluorination has little
Based on the relatively inflexible nature of trehalose, impact on the solution conformation of trehalose, suggesting
coupled with the negligible effect deoxygenation has on its that components of mycobacterial trehalose metabolism may
structure,58,59 we predicted that monodeoxyfluorination would recognize and process FDTre analogues similarly to trehalose.
not have a major impact on its conformation. This hypothesis
Uptake of 19F-FDTre analogues by M. smegmatis
was supported by a simple conformational search employing
the MM3* force field,60 which predicted tightly grouped φH After confirming that FDTre analogues faithfully mimic the
and ψH values (−51° to −52°) for the lowest energy conformers solution structure of trehalose, we tested their uptake in
of trehalose and the four FDTre analogues (Table 1). A super- M. smegmatis, a frequently used avirulent model organism for
position of the five MM3*-derived theoretical structures, M. tuberculosis. As discussed in the introduction, mycobacteria
shown in Fig. 6A, depicts their similarity.
possess at least two conserved, mycobacteria-specific pathways
NMR spectroscopy was used to experimentally determine through which FDTre analogues could accumulate in the cell:
the analogues’ solution conformations. 1H, 13C, COSY, and (i) extracellular Ag85-mediated mycoloylation to anchor FDTre
HSQC NMR spectra were acquired and analyzed to make reson- to the cell wall; (ii) SugABC-LpqY-mediated uptake into the
3
ance assignments for each compound. Analysis of vicinal JHH cytoplasm, possibly followed by incorporation into cell-wall tre-
values confirmed that the 4C1 chair conformations of the halose glycolipids. While no cellular uptake experiments have
hexose rings were not distorted by fluorine substitution. To been reported for FDTre analogues, it was previously shown
determine the long-range 3JCOCH coupling constants across the that 19F-2-FDTre (9) reacted with a TDM analogue in vitro in
glycosidic bond, we performed a 2D excitation-sculptured the presence of recombinant Ag85.29 However, Ag85 operates
indirect-detection NMR experiment (EXSIDE)61 with anomeric in the opposite direction in vivo, converting two molecules of
proton(s) selected (H1 for trehalose; H1 and H1′ for the TMM to one molecule each of TDM and trehalose, the latter of
19F-FDTre analogues). The dihedral angles φH and ψH were which is recycled back into the cell (see Fig. 1).28 Therefore, it
then calculated using Tvaroška’s Karplus-type equation shown is possible that Ag85-catalyzed mycoloylation of FDTre ana-
above (eqn (1)).51 Representative EXSIDE spectra for trehalose logues would not occur appreciably in living cells, unless a
and 19F-2-FDTre are shown in Fig. 6C and D, respectively (see high concentration of the analogue were available in the peri-
Fig. S1 in the ESI† for all EXSIDE spectra).
plasm to force the reverse reaction.
Previous cellular labeling experiments with TreAz analogues
suggest that the trehalose recycling route may provide efficient
For native trehalose, the 3JCOCH constant and corresponding
glycosidic bond dihedral angles (3JCOCH = 3.0 Hz; φH = ψH
=
−44°) were in excellent agreement with previously reported uptake of FDTre analogues by mycobacteria. SugABC-LpqY
3
data (Table 1). For the 19F-FDTre analogues, the JCOCH con- is constitutively expressed and required for virulence of
stants spanned 3.0–3.4 Hz, which corresponded to φH and ψH M. tuberculosis, and its function in vivo is to import free
values ranging from −40 to −45° (Table 1). These values fall trehalose that is released by Ag85-catalyzed mycoloylation
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
M. smegmatis and M. bovis BCG.30 On the other hand, 3-TreAz (4),
which also labeled the cell surface, was not taken up by
SugABC-LpqY in either species.30 Instead, it was presumably
anchored to the cell wall via periplasmic Ag85-mediated myco-
loylation. A much higher concentration of 3-TreAz (500 µM)
was required to achieve labeling, and even then it was only
about 10–20% of the surface labeling given by a much lower
concentration of 2- and 6-TreAz (25 µM).30 4-TreAz (5) was also
taken up by SugABC-LpqY in M. smegmatis, though not in
M. bovis BCG.30 These studies indicated that SugABC-LpqY is
tolerant of some trehalose analogues bearing small hydroxyl
replacement modifications, and that it may incorporate such
analogues into mycobacterial cells more efficiently than Ag85.¶
Based on these data, we hypothesized that subtly modified
FDTre analogues would undergo SugABC-LpqY-mediated
transport into mycobacteria.
We evaluated the four 19F-FDTre analogues synthesized in
this study for SugABC-LpqY-mediated uptake into whole cells
of M. smegmatis. To determine whether uptake was passive or
dependent on the trehalose transporter SugABC-LpqY, three
bacterial strains were tested in parallel: (i) M. smegmatis wild
type; (ii) M. smegmatis ΔsugC mutant, which lacked a func-
tional SugABC-LpqY transporter; and (iii) M. smegmatis ΔsugC::
sugC complement, which had the transporter restored.42 Each
strain was cultured in the presence of 25 µM 19F-FDTre or left
untreated, then cells were washed and lysed. Cytosolic extracts
containing water-soluble species were dried and subjected to
trimethylsilyl (TMS) derivatization using N-methyl-N-(tri-
methylsilyl) trifluoroacetamide (MSTFA). TMS-derivatized
extracts were analyzed by GC-MS to provide a qualitative
assessment of 19F-FDTre uptake (Fig. 7 shows representative
GC-MS data for 19F-2-FDTre; see Fig. S3 in the ESI† for data on
other analogues).
As expected, all cytosolic extracts contained a prominent
peak for native trehalose, which is endogenously biosynthe-
sized in the cytoplasm as described in the introduction. For
samples that were treated with 19F-2-, 19F-3-, and 19F-6-FDTre
(9, 10, and 12), the corresponding FDTre peaks were observed
in wild-type M. smegmatis but not in the untreated controls,
indicating that these analogues entered the cell. Critically, the
FDTre peaks were absent in the M. smegmatis ΔsugC mutant
and restored in the M. smegmatis ΔsugC::sugC complement,
Fig. 6 Conformational analysis of trehalose and 19F-FDTre analogues.
Superpositions of the lowest energy conformers of trehalose and
19F-FDTre analogues (1, 9–12) determined by conformational searches
using MM3* (A) without constraints and (B) with NMR-determined φH which confirmed that SugABC-LpqY was solely responsible for
and ψH values as constraints. The fluorine atom of each 19F-FDTre struc-
the observed analogue uptake. By contrast, 19F-4-FDTre (11)
was not taken up by M. smegmatis.
ture is colored red. Representative EXSIDE spectra are shown for (C) tre-
3
halose (1) and (D) 19F-2-FDTre (9). JCOCH values were measured from
Between the TreAz and FDTre uptake experiments, it can be
EXSIDE spectra using a scaling factor of 25, as shown. See Fig. S1 and S2
concluded that small 2- and 6-position monosubstitutions are
fairly well-tolerated by SugABC-LpqY in M. smegmatis. Because
4-TreAz and 19F-3-FDTre were taken up by SugABC-LpqY but
3-TreAz and 19F-4-FDTre were not, there appears to be mixed
in the ESI† for additional EXSIDE spectra and structural images for all
compounds.
processes.28 Although SugABC-LpqY is known to be highly
specific for α,α-trehalose versus other disaccharides (e.g.,
α,β-trehalose and maltose),28 the Bertozzi group demonstrated
that 2- and 6-TreAz (3 and 6) were actively imported into the cell
by SugABC-LpqY and subsequently incorporated into trehalose
¶Ag85 may be involved in FDTre labeling at lower probe concentrations, most
likely through the following pathway: SugABC-LpqY-mediated uptake followed by
Pks13-catalyzed mycoloylation to form fluorine-modified TMM, which would
then undergo MmpL3-mediated export and further mycoloylation by Ag85 to
glycolipids at relatively low concentrations (25 µM) in both give fluorine-modified TDM (refer to Fig. 1).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
ively convert 19F-2-FDG to 19F-2-FDTre in only 15 min at 70 °C.
After the reaction was completed (15 min), the mixture was
subjected to spin dialysis (10 min) followed by treatment with
mixed-bed ion exchange resin (25 min). After passage through
a 0.2 µm sterilization filter, an aqueous solution of pure 19F-2-
FDTre of known concentration (approximately 5 mM∥) was in
hand, and the entire process took only 1 hour. TLC and 1H
NMR analysis of the 19F-2-FDTre product established its iden-
tity and purity (see Fig. S4A and S4B in the ESI†). Immediately
after the filtration step, an appropriate volume of the 19F-2-
FDTre solution was added to growing M. smegmatis strains
(wild type, ΔsugC mutant, ΔsugC::sugC complement) to achieve
a final concentration of approximately 25 µM. The cultures
were incubated for 1 hour, then processed and analyzed by
GC-MS as described above. The uptake results from this experi-
ment were identical to those shown in Fig. 7 (see Fig. S4C in
the ESI†).
These results demonstrate that the TreT method can be
used to synthesize and purify 18F-FDTre analogues on a time-
scale that would still allow time to administer them to an
animal or patient before the extinction of the radionuclide
signal. This is especially significant considering that many
hospitals already have the infrastructure both to synthesize
18F-2-FDG and image patients using PET, as it is often used to
detect and monitor the treatment of cancer. Thus, our method
could tap into existing medical infrastructure and expand its
functionality to include imaging of TB patients.
Fig. 7 Representative GC-MS data demonstrating SugABC-LpqY-
dependent uptake of 19F-2-FDTre (9) by M. smegmatis. Standards or
cytosolic extracts from cells were dried, TMS-derivatized, and analyzed
by GC-MS. (i) Trehalose standard; (ii) 19F-2-FDTre standard; (iii) untreated
M. smegmatis wild type; (iv) 19F-2-FDTre-treated M. smegmatis wild
type; (v) 19F-2-FDTre-treated M. smegmatis ΔsugC mutant; (vi) 19F-2-
FDTre-treated M. smegmatis ΔsugC::sugC complement. See Fig. S3 in
the ESI† for GC-MS data on other 19F-FDTre analogues.
tolerance at these positions, perhaps due to the disruption of
a critical interaction (e.g., hydrogen bond) at this region of the
substrate. It has not yet been studied whether the homologous
SugABC-LpqY transporters in pathogenic mycobacteria, such
as M. tuberculosis, process FDTre analogues similarly to the
M. smegmatis transporter. However, the consistency in TreAz
uptake observed between M. smegmatis and M. bovis BCG30—a
close genetic relative of M. tuberculosis—suggests that there
may be overlap for at least a subset of the FDTre analogues. In
sum, the considerable substrate tolerance of SugABC-LpqY,
revealed by uptake experiments on the TreAz and FDTre series,
provides justification for targeting this pathway with 18F-FDTre
analogues for TB PET imaging applications.
Conclusions
Most significantly, this study established an uncomplicated yet
robust synthetic process for rapidly accessing FDTre ana-
logues. The heat-stable glycosyltransferase TreT from T. tenax
exhibited high activity on three out of four FDG substrates,
which allowed their complete conversion to FDTre products in
a matter of minutes—a favorable timescale for 18F-FDTre radio-
synthesis. The expeditious purification protocol reported here
is also fitting for radiosynthesis. Fortuitously, the FDTre
product is the only neutral species in the TreT reaction
mixture (as long as substrate conversion is quantitative), so
purification was accomplished by a simple ion exchange step
following enzyme removal. Also of benefit, the TreT synthesis
process has a minimal environmental impact since it uses a
single biocatalytic step and an entirely aqueous purification.
At present, the scale of the TreT reaction is limited by the cost
of UDP-glucose (∼$300 g−1), which we are currently working to
address. Regardless, the reaction scale used in this study is
more than sufficient for radiosynthesis applications. Because
radiosynthesis is usually performed with small amounts of
radioactive starting material, a future challenge will be to
Protocol for the rapid synthesis, purification, and
administration of 19F-2-FDTre to M. smegmatis as a model for
future 18F-FDTre PET imaging applications
The data presented above indicate that TreT catalysis is ideally
suited to the radiosynthesis of 18F-FDTre analogues. The ready
availability of 18F-2-FDG at medical imaging facilities, coupled
with the exceptionally high activity of TreT on this substrate, is
of particular importance. We envision that TreT catalysis could
be used in a radiosynthesis facility to convert freshly prepared
18F-2-FDG to 18F-2-FDTre, followed by product purification and
administration to an M. tuberculosis-infected animal or human
patient. To demonstrate the feasibility of this idea, we develo-
ped a protocol for 19F-2-FDTre production and utilization that
models this process and can be executed in 1 hour.
∥The 5 mM concentration was estimated based on quantitative conversion of
the substrate to product, which in this instance was confirmed by TLC and 1H
NMR. Wash steps resulted in a four-fold dilution from the starting substrate con-
In an effort to minimize FDTre preparation time, we shor-
tened the reaction time and found that TreT could quantitat- centration, which was 20 mM.
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
adapt the TreT reaction scale and stoichiometry to this setting. accumulates in nervous tissue at concentrations high enough
Also of note, 4-FDTre was not accessible via TreT catalysis and to counteract protein aggregation. Thus, 18F-FDTre analogues
was instead chemically synthesized in this study. However, if have the potential to impact various areas of biomedicine,
4-FDTre shows high promise for PET imaging applications, its and the methods established herein provide a platform for
radiosynthesis could potentially be accomplished by subjecting accessing this promising new class of nuclear imaging
intermediate 18 to sequential trifluoromethanesulfonylation, probes.
SN2 reaction with a [18F]fluoride nucleophile, and deprotection.
FDTre analogues must emulate native trehalose to serve as
probes for trehalose-related metabolic processes, which
prompted us to investigate the effect of deoxyfluorination on
Acknowledgements
This work was funded by a grant from the National Institutes
of Health (R15 AI117670) to B. M. S. and P. J. W., as well as
Cottrell College Scholar Awards from the Research Corporation
to B. M. S. (22525) and P. J. W. (20185). L. M. M. was supported
by a Provost’s Fellowship from CMU. R. Kalscheuer is thanked
for providing bacterial strains and L. Zhao is thanked for a
critical reading of the manuscript.
the conformation of trehalose. Molecular mechanics predic-
tions that deoxyfluorination would not distort trehalose’s
structure were corroborated by NMR analyses, which showed
virtually no variability in intra- and inter-residue dihedral
angles between trehalose and the FDTre analogues.
Accordingly, FDTre analogues accurately mimic the confor-
mation of trehalose, suggesting that they can potentially be
used to probe trehalose metabolism in mycobacteria and other
organisms.
An important potential application of 18F-FDTre analogues
is in vivo PET imaging of M. tuberculosis infection. In support
of this objective, it will be critical to thoroughly characterize
the metabolism of FDTre analogues by mycobacteria. Here, we
provided evidence that FDTre analogues were successfully
recognized and taken up through the trehalose transporter
SugABC-LpqY in whole M. smegmatis cells. Thus, the trehalose
recycling pathway can serve as a biochemical sink through
which FDTre analogues can accumulate in the cell. Further
studies on FDTre uptake are warranted, including quantitation
of SugABC-LpqY-mediated FDTre uptake in M. smegmatis and
pathogenic mycobacteria, as well as uptake in other types of
bacteria and mammalian cells to evaluate FDTre specificity for
mycobacteria. In addition, the preliminary uptake experiments
reported herein were conducted at micromolar concentrations
to probe the transporter’s substrate tolerance. Future studies
should assess uptake efficiencies at FDTre concentrations that
accurately reflect in vivo radiotracer concentrations. The possi-
bility of Pks13, MmpL3, and Ag85 (and potentially other treha-
lose-processing proteins) being involved in the uptake of
FDTre analogues should also be investigated.¶ The non-radio-
active 19F-FDTre analogues reported herein will facilitate these
experiments. Toward the long-term goal of this project, we are
currently translating our results to the synthesis and evalu-
ation of 18F-FDTre probes for PET imaging of mycobacterial
infection in an animal model.
Notes and references
1 C. Dye, Lancet, 2006, 367, 938–940.
2 A. Matteelli, A. Roggi and A. C. Carvalho, Clin. Epidemiol.,
2014, 6, 111–118.
3 L. Norbis, P. Miotto, R. Alagna and D. M. Cirillo,
New Microbiol., 2013, 36, 111–120.
4 World Health Organization, Global tuberculosis report
2015, 2015.
5 D. H. Johnson, L. E. Via, P. Kim, D. Laddy, C.-Y. Lau,
E. A. Weinstein and S. Jain, Nucl. Med. Biol., 2014, 41, 777–
784.
6 A. O. Ankrah, T. S. van der Werf, E. F. J. de Vries,
R. A. J. O. Dierckx, M. M. Sathekge and
A. W. J. M. Glaudemans, Clin. Transl. Imaging, 2016, 4, 131–
144.
7 L. Liu, Y. Xu, C. Shea, J. S. Fowler, J. M. Hooker and
P. J. Tonge, J. Med. Chem., 2010, 53, 2882–2891.
8 E. A. Weinstein, L. Liu, A. A. Ordonez, H. Wang,
J. M. Hooker, P. J. Tonge and S. K. Jain, Antimicrob. Agents
Chemother., 2012, 56, 6284–6290.
9 X. Ning, W. Seo, S. Lee, K. Takemiya, M. Rafi, X. Feng,
D. Weiss, X. Wang, L. Williams, V. M. Camp, M. Eugene,
W. R. Taylor, M. Goodman and N. Murthy, Angew. Chem.,
Int. Ed., 2014, 53, 14096–14101.
10 G. Gowrishankar, M. Namavari, E. B. Jouannot, A. Hoehne,
R. Reeves, J. Hardy and S. S. Gambhir, PLoS One, 2014, 9,
e107951.
Finally, 18F-FDTre analogues may have applications
beyond imaging mycobacterial infection. Many members of
the gut microbiota use trehalose as a carbon source and 11 E. A. Weinstein, A. A. Ordonez, V. P. DeMarco,
osmoprotectant62,63 and may metabolize 18F-FDTre, which
could yield information on the dynamic changes that these
bacteria undergo in response to diet, stress, and antibiotic
A. M. Murawski, S. Pokkali, E. M. MacDonald, M. Klunk,
R. C. Mease, M. G. Pomper and S. K. Jain, Sci. Transl. Med.,
2014, 6, 259ra146.
treatment. It has also been proposed that trehalose may have 12 S. L. Davis, N. A. Be, G. Lamichhane, S. Nimmagadda,
value in treating neurodegenerative diseases due to its ability
to prevent protein aggregation and induce autophagy.64
M. G. Pomper, W. R. Bishai and S. K. Jain, PLoS One, 2009,
4, e6297.
18F-FDTre analogues could be used to study the metabolic 13 A. D. Elbein, Y. T. Pan, I. Pastuszak and D. Carroll,
fate of trehalose in animals and confirm whether it
Glycobiology, 2003, 13, 17R–27R.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
14 H. Tournu, A. Fiori and P. Van Dijck, PLoS Pathog., 2013, 9, 30 B. M. Swarts, C. M. Holsclaw, J. C. Jewett, M. Alber,
e1003447.
D. M. Fox, M. S. Siegrist, J. A. Leary, R. Kalscheuer and
C. R. Bertozzi, J. Am. Chem. Soc., 2012, 134, 16123–16126.
31 H. N. Foley, J. A. Stewart, H. W. Kavunja, S. R. Rundell and
B. M. Swarts, Angew. Chem., Int. Ed., 2016, 55, 2053–2057.
15 K. A. L. De Smet, A. Weston, I. N. Brown, D. B. Young and
B. D. Robertson, Microbiology, 2000, 146, 199–208.
16 P. J. Woodruff, B. L. Carlson, B. Siridechadilok, M. R. Pratt,
R. H. Senaratne, J. D. Mougous, L. W. Riley, S. J. Williams 32 V. A. Sarpe and S. S. Kulkarni, Trends Carbohydr. Res., 2013,
and C. R. Bertozzi, J. Biol. Chem., 2004, 279, 28835–28843. 5, 8–33.
17 H. N. Murphy, G. R. Stewart, V. V. Mischenko, A. S. Apt, 33 M. A. Chaube and S. S. Kulkarni, Trends Carbohydr. Res.,
R. Harris, M. S. B. McAlister, P. C. Driscoll, D. B. Young 2013, 4, 1–19.
and B. D. Robertson, J. Biol. Chem., 2005, 280, 14524– 34 A. F. Hadfield, L. Hough and A. C. Richardson, Carbohydr.
14529. Res., 1979, 71, 95–102.
18 S. Gavalda, F. Bardou, F. Laval, C. Bon, W. Malaga, 35 X. Li, H. Ohtake, H. Takahashi and S. Ikegami, Tetrahedron,
C. Chalut, C. Guilhot, L. Mourey, M. Daffé and
A. Quémard, Chem. Biol., 2014, 21, 1660–1669.
19 A. E. Grzegorzewicz, H. Pham, V. A. K. B. Gundi,
2001, 57, 4297–4309.
36 R. Namme, T. Mitsugi, H. Takahashi and S. Ikegami,
Eur. J. Org. Chem., 2007, 3758–3764.
M. S. Scherman, E. J. North, T. Hess, V. Jones, V. Gruppo, 37 M. A. Rodriguez, O. Boutureira, M. I. Matheu, Y. Diaz,
S. E. M. Born, J. Kordulakova, S. S. Chavadi, C. Morisseau,
A. J. Lenaerts, R. E. Lee, M. R. McNeil, M. Jackson,
S. Castillon and P. H. Seeberger, J. Org. Chem., 2007, 72,
8998–9001.
J. Korduláková, S. S. Chavadi, C. Morisseau, A. J. Lenaerts, 38 M. Walmagh, R. Zhao and T. Desmet, Int. J. Mol. Sci., 2015,
R. E. Lee, M. R. McNeil and M. Jackson, Nat. Chem. Biol.,
2012, 8, 334–341.
20 S. P. S. Rao, S. B. Lakshminarayana, R. R. Kondreddi,
M. Herve, L. R. Camacho, P. Bifani, S. K. Kalapala,
J. Jiricek, N. L. Ma, B. H. Tan, S. H. Ng, M. Nanjundappa,
16, 13729–13745.
39 P. Som, H. L. Atkins, D. Bandoypadhyay, J. S. Fowler,
R. R. MacGregor, K. Matsui, Z. H. Oster, D. F. Sacker,
C. Y. Shiue, H. Turner, C.-N. Wan, A. P. Wolf and
S. V. Zabinski, J. Nucl. Med., 1980, 21, 670–675.
S. Ravindran, P. G. Seah, P. Thayalan, S. H. Lim, B. H. Lee, 40 H. M. Giaever, O. B. Styrvold, I. Kaasen and A. R. Strøm,
A. Goh, W. S. Barnes, Z. Chen, K. Gagaring, J. Bacteriol., 1988, 170, 2841–2849.
A. K. Chatterjee, K. Pethe, K. Kuhen, J. Walker, G. Feng, 41 A. R. Strøm and I. Kaasen, Mol. Microbiol., 1993, 8, 205–210.
S. Babu, L. Zhang, F. Blasco, D. Beer, M. Weaver, V. Dartois, 42 B. L. Urbanek, D. C. Wing, K. S. Haislop, C. J. Hamel,
R. Glynne, T. Dick, P. W. Smith, T. T. Diagana and
U. H. Manjunatha, Sci. Transl. Med., 2013, 5, 214ra168.
R. Kalscheuer, P. J. Woodruff and B. M. Swarts,
ChemBioChem, 2014, 15, 2066–2070.
21 K. Tahlan, R. Wilson, D. B. Kastrinsky, K. Arora, V. Nair, 43 T. Kouril, M. Zaparty, J. Marrero, H. Brinkmann and
E. Fischer, S. W. Barnes, J. R. Walker, D. Alland, C. E. Barry B. Siebers, Arch. Microbiol., 2008, 190, 355–369.
3rd and H. I. Boshoff, Antimicrob. Agents Chemother., 2012, 44 T. J. Tewson, M. J. Welch and M. E. Raichle, J. Nucl. Med.,
56, 1797–1809. 1978, 19, 1339–1345.
22 N. Sathyamoorthy and K. Takayama, J. Biol. Chem., 1987, 45 T. R. Neal, W. C. Schumann, M. S. Berridge and B. R. Landau,
262, 13417–13423. J. Labelled Compd. Radiopharm., 2005, 48, 845–854.
23 J. T. Belisle, V. D. Vissa, T. Sievert, K. Takayama, 46 R. W. Bassily, R. I. El-Sokkary, B. A. Silwanis,
P. J. Brennan and G. S. Besra, Science, 1997, 276, 1420–
1422.
A. S. Nematalla and M. A. Nashed, Carbohydr. Res., 1993,
239, 197–207.
24 K. M. Backus, M. A. Dolan, C. S. Barry, M. Joe, P. McPhie, 47 K. Bock, J. Defaye, H. Driguez and E. Bar-Guilloux,
H. I. Boshoff, T. L. Lowary, B. G. Davis and C. E. Barry 3rd,
J. Biol. Chem., 2014, 289, 25041–25053.
25 R. Ryll, Y. Kumazawa and I. Yano, Microbiol. Immunol.,
2001, 45, 801–811.
Eur. J. Biochem., 1983, 131, 595–600.
48 G. M. Brown, D. C. Rohrer, B. Berking, C. a. Beevers,
R. O. Gould and R. Simpson, Acta Crystallogr., Sect. B:
Struct. Crystallogr. Cryst. Chem., 1972, 28, 3145–3158.
26 J. Indrigo, R. L. Hunter and J. K. Actor, Microbiology, 2003, 49 G. A. Jeffrey and R. Nanni, Carbohydr. Res., 1985, 137,
149, 2049–2059. 21–30.
27 E. Ishikawa, T. Ishikawa, Y. S. Morita, K. Toyonaga, 50 A. Moyano, C. A. Duda and E. S. Stevens, J. Am. Chem. Soc.,
H. Yamada, O. Takeuchi, T. Kinoshita, S. Akira, Y. Yoshikai
and S. Yamasaki, J. Exp. Med., 2009, 206, 2879–2888.
28 R. Kalscheuer, B. Weinrick, U. Veeraraghavan, G. S. Besra
1990, 112, 7406–7407.
51 I. Tvaroška, M. Hricovíni and E. Petráková, Carbohydr. Res.,
1989, 189, 359–362.
and W. R. Jacobs, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 52 G. Batta, K. E. Kövér, J. Gervay, M. Hornyák and
21761–21766. G. M. Roberts, J. Am. Chem. Soc., 1997, 119, 1336–1345.
29 K. M. Backus, H. I. Boshoff, C. S. Barry, O. Boutureira, 53 A. Poveda, C. Vicent, S. Penades and J. Jiménez-Barbero,
M. K. Patel, F. D’Hooge, S. S. Lee, L. E. Via, K. Tahlan, Carbohydr. Res., 1997, 301, 5–10.
C. E. Barry 3rd and B. G. Davis, Nat. Chem. Biol., 2011, 7, 54 N. W. H. Cheetham, P. Dasgupta and G. E. Ball, Carbohydr.
228–235.
Res., 2003, 338, 955–962.
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
55 M. K. Dowd, P. J. Reilly and A. D. French, J. Comput. Chem., 60 N. L. Allinger, Y. H. Yuh and J. H. Lii, J. Am. Chem. Soc.,
1992, 13, 102–114. 1989, 111, 8551–8566.
56 Q. Liu, R. K. Schmidt, B. Teo, P. A. Karplus and J. W. Brady, 61 V. V. Krishnamurthy, J. Magn. Reson., Ser. A, 1996, 121,
J. Am. Chem. Soc., 1997, 119, 7851–7862. 33–41.
57 S. C. C. Nunesa, A. J. L. Jesus, M. J. Moreno and 62 F. Peris-Bondia, A. Latorre, A. Artacho, A. Moya and
M. E. S. Eusébio, Carbohydr. Res., 2010, 345, 2048–2059. G. D’Auria, PLoS One, 2011, 6, e22448.
58 F. L. Lin, H. van Halbeek and C. R. Bertozzi, Carbohydr. 63 T. Duong, R. Barrangou, W. M. Russell and
Res., 2007, 342, 2014–2030.
59 L. C. Lin, Ph.D. dissertation, University of California,
Berkeley, 2006.
T. R. Klaenhammer, Appl. Environ. Microbiol., 2006, 72,
1218–1225.
64 E. Emanuele, Curr. Drug Targets, 2014, 15, 551–557.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016