Thiophosphate and thiophosphonate analogues of glucose-1-phosphate: Synthesis and enzymatic activity with a thymidylyltransferase

Article abstract of DOI:10.1016/j.carres.2013.06.007

Synthetic methods were investigated for the preparation of O and S-glucosyl thiophosphates and glucosyl 1C-thiophosphonate. Four protected glucosyl thiophosphate compounds were synthesized and characterized as precursors to glucose 1-thiophosphate. The ef

Full text of DOI:10.1016/j.carres.2013.06.007

Accepted Manuscript
Thiophosphate and thiophosphonate analogues of glucose-1-phosphate: syn‐
thesis and enzymatic activity with a thymidylyltransferase
Matthew W. Loranger, Stephen A. Beaton, Katie L. Lines, David L. Jakeman
PII:
S0008-6215(13)00240-1
DOI:
http://dx.doi.org/10.1016/j.carres.2013.06.007
CAR 6503
Reference:
To appear in:
Carbohydrate Research
Received Date:
Revised Date:
Accepted Date:
6 May 2013
10 June 2013
11 June 2013
Please cite this article as: Loranger, M.W., Beaton, S.A., Lines, K.L., Jakeman, D.L., Thiophosphate and
thiophosphonate analogues of glucose-1-phosphate: synthesis and enzymatic activity with a thymidylyltransferase,
Carbohydrate Research (2013), doi: http://dx.doi.org/10.1016/j.carres.2013.06.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Thiophosphate and thiophosphonate analogues of glucose-1-phosphate: synthesis and
enzymatic activity with a thymidylyltransferase
a
a
a
ab
Matthew W. Loranger, Stephen A . Beaton, K atie L. Lines and D avid L. Jakem an*
aD epartm ent of C hem istry, D alhousie U niversity, 6274 C oberg R d., PO Box 15,000, H alifax, N ova Scotia, B3H 4R 2, C anada;
bC ollege of Pharm acy, D alhousie U niversity, 5968 C ollege St., PO Box 15,000, H alifax, N ova Scotia, B3H , 4R2. Em ail:
david.jakeman@dal.ca;
Abstract
Synthetic methods were investigated for the preparation of O and S-glucosyl thiophosphates and
glucosyl 1C-thiophosphonate. Four protected glucosyl thiophosphate compounds were
synthesized and characterized as precursors to glucose 1-thiophosphate. The effect of various
reaction conditions and the nature of the carbohydrate and thiophosphate protecting groups and
how they impact both the yields and ꢀ/ꢁ diastereoselectivity of the glucosyl thiophosphate
products were explored. A novel isomerization from an O-linked to S-linked glucosyl
thiophosphate was observed. ꢀ-D-Glucose-1C-thiophosphonate was synthesized and evaluated as
a substrate for the thymidylyltransferase, Cps2L. Tandem mass spectrometric analysis
determined the position of sulfur in the sugar nucleotide product.
1
. Introduction
Glycosyl phosphates are natural phosphates of importance in biological systems as key metabolic
1
,2
intermediates and as bacterial cell wall constituents. Thiophosphates are analogues of natural
phosphates in which the P—O bond has been replaced with a phosphorus-sulfur (P—S) bond.
The sulfur atom within thiophosphates is capable of mimicking both the binding and
3
,4
nucleophilic properties of the oxygen atom of natural phosphates. Many classes of enzymes
utilize phosphate or pyrophosphate as substrates or recognition elements, and thiophosphates
have provided significant insight into interactions of these enzymes, including determining the
5
stereochemical outcome of phosphoryl transfer at phosphorus. Despite the extensive literature
associated with enzymes transferring phosphates, aspects of the mechanism of enzymatic
1

phosphoryl transfer remain contentious, with metal fluoride complexes providing new insights
6
for those enzymes forming covalent phosphoryl-enzyme intermediates. Phosphonothioates are
non-hydrolysable derivatives of phosphates with the structure RPS(OH) and have been used to
2
probe receptor signaling. Prestwich and coworkers reported the synthesis and evaluation of
7
several phosphonothioate analogues as antagonists of lysophosphatic acid (LPA) receptors.
Sugar phosphorylases are a class of enzyme that use glycosyl phosphates as substrates to furnish
oligosaccharides. These enzymes offer complementary approaches to that of chemical synthesis
in the preparation of oligosaccharides due to the regio- and stereoselectivity of the glycosidic
8
linkage formed.
Glycosyl phosphates including ꢁ- and ꢀ-GlcNAc-1-phosphate, ꢁ- and ꢀ-glucose 1-
phosphate, have various roles in biological pathways, making their thiophosphate or
thiophosphonate analogues attractive synthetic targets with which to probe enzyme function. ꢁ-
D-Glucose 1-phosphate is a substrate for ꢁ-phosphoglucomutase and is suspected to be a
9
,10
precursor in cell wall biosynthesis in Lactococcus lactis.
ꢀ-D-Glucose 1-phosphate is a
substrate for nucleotidylyltransferase enzymes in a number of biosynthetic pathways, including
the incorporation of L-rhamnose into the cell wall of harmful pathogens including Streptococcus
1
1,12,13
pneumoniae, Pseudomonas aeruginosa and Mycobacterium tuberculosis.
Glycosyl thiophosphates have been synthesized in both O and S-glycosylated forms. O-
glycosyl thiophosphates have been previously synthesized as glycosyl donors for the
stereoselective production of glycosides. Hui and coworkers reported the synthesis of acetyl and
1
4
benzyl-protected glycosyl dimethylthiophosphates in 51% and 86% yields, respectively. They
generated several diastereotopic mixtures of fully protected glycosyl thiophosphates with varying
ꢀ/ꢁ selectivity, depending on the nature of the glycosyl protecting groups. More recently,
Piekutowska and Pakulski generated a series of benzoyl protected mannose, galactose and
glucose S-glycosyl thiophosphates by reacting anomeric thiocyanates with O-alkyl or O-
1
5
trimethylsilyl phosphites. However, the global deprotection of these glycosyl thiophosphates
was not reported.
The only example of a free sugar glycosyl thiophosphate was reported by Wang and
1
6
coworkers in 2011 using enzymatic approaches. Using NahK (EC 2.7.1.162), an N-
acetylhexosamine 1-kinase, they were able to produce GlcNAc-1-thiophosphate on a milligram
scale from GlcNAc and commercially available ꢂ(S)ATP. The GlcNAc-1-P uridyltransferase
2

GlmU (EC 2.3.1.157) was then utilized to produce UDP(ꢁS)-GlcNAc in a respectable yield
Scheme 1).
No synthetic route currently exists to access fully deprotected O-glycosyl and S-glycosyl
(
sugar 1-thiophosphates or glucosyl 1C-thiophosphonate. We report the synthesis of a series of
differentially protected O-glucosyl and S-glucosyl thiophosphates. We looked at the influence of
reaction conditions and protecting groups on the diastereotopic ratio of products and explored
common deprotection strategies towards the production of O- and S- glucose 1-thiophosphate.
The thiophosphonate analogue, Glucose 1C-thiophosphonate, was synthesized and enzymatically
transformed into a sugar nucleotide analogue of deoxythymidine diphosphate-ꢀ-D-glucose
1
7
(
dTDP-Glc) using thymidylyltransferase Cps2L. This compound represents the first example of
a thiophosphonate containing sugar nucleotide analogue.
1
6
Scheme 1 Enzymatic synthesis of GlcNAc-1TP and UDP(ꢁS)-GlcNAc
3

Scheme 2 Reactions catalyzed by Cps2L.
2
. Results and Discussion
The formation of glucosylated thiophosphates was explored using acetyl (1) or benzyl (2)
protected glucopyranose as the starting sugar together with diethylchlorothiophosphate (3) using
1
4
the conditions described by Hui or diphenylchlorothiophosphate (4) using the conditions
1
8
described for preparation of glycosyl phosphates (Scheme 3). Table 1 summarizes the various
reaction conditions and isolated yields for ꢀ/ꢁ-D-glucopyranosyl thiophosphate synthesis.
Scheme 3 The synthesis of protected ꢀ/ꢁ-D-glucopyranosyl thiophosphates (5–7)
Table 1 Reaction conditions for ꢀ/ꢁ-D-glucopyranosyl thiophosphate synthesis
c
Entry Reactants
Conditions
T (°C)
Product(s)
ꢀ/ꢁ
Yield
1
2
3
4
5
1 + 3
2 + 3
2 + 3
2 + 4
2 + 4
THF/n-BuLi
THF/n-BuLi
-78
-78
0
5
1:3
2:1
57%
67%
0%
6
CH
CH
Cl
Cl
/DMAP
/DMAP
N/A
7a/7b
7a/7b
N/A
2
2
a
b
0
1 , 4:1
24%
16%
2
2
a
b
THF/n-BuLi
-78
6:1 , 4:1
a
b
c
Initial ratio for O-linked product 7a. Final ratio for S-linked product 7b. Yields are reported as
overall isolated yields for all anomers and isomers.
The reaction solvent and base were investigated in an attempt to enhance yield and selectivity.
When diethylchlorothiophosphate (3) was used as the electrophile (entries 1-3), both n-
4

butyllithium (n-BuLi) and lithium diisopropylamide (LDA) were comparably effective bases in
tetrahydrofuran (THF), however 4-dimethylaminopyridine (DMAP) in dichloromethane
(
CH Cl ) was ineffective in promoting glycosyl coupling. The yields for 5 and 6 (entry 1-2) were
2 2
comparable will analogous methyl protected glycosyl thiophosphates reported by Hui (51-
1
4
8
6%). For ease of preparation, n-BuLi was chosen as the base of choice for the production of 5
and 6. When diphenyl chlorothiophosphate (4) was used as the electrophile (entries 4-5), either
n-BuLi/THF or DMAP/CH Cl conditions proved to be sufficient for the production of 7a. The
2
2
complete structures of 7a and 7b can be found in Scheme 4. The reactions conditions used
affected both the diastereotopic ratio of ꢀ/ꢁ products as well as the overall yield. When n-BuLi
/
THF conditions were used the resultant ꢀ/ꢁ ratio of 7a was found to be ~6:1 with a 16% overall
yield (entry 5). When DMAP/CH Cl conditions were used, only the ꢀ isomer was obtained with
2
2
a 24% overall yield (entry 4). The yields of 7a/7b (entry 4) were significantly less than
analogous glycosyl phosphates synthesized using DMAP/CH Cl conditions reported previously
2
2
1
8
(
55-82%). While reduced coupling efficiency can be partially attributed to the steric bulk of the
phenyl protecting groups of 4 relative to the ethyl protected 3, the driving factor in the reduced
yields is believed to be the excellent leaving group capacity of the diphenylthiphosphate moiety,
leading to decomposition during purification. The DMAP/CH Cl method was preferred in the
2
2
synthesis of 7a due to its stereospecificity and improved yield relative to the n-BuLi/THF
coupling method.
The diastereotopic ratios of the starting sugars 1 (ꢁ only) and 2 (ꢀ/ꢁ ~3:1) were not
retained in the production of any of the glucosyl thiophosphates (5–7). Varying temperature
conditions from -78°C to -40°C did not affect the diastereotopic ratio of products for compounds
5
and 6. Using the benzyl-protected sugar (2) moderate improvement in overall yield was
observed relative to the acetyl-protected sugar (1), which can be attributed to the acetyl groups of
deactivating the anomeric hydroxyl group’s ability to act as a nucleophile in the coupling
1
reaction with the chlorothiophosphate. Reactions involving the chlorothiophosphate 3 produced
an ꢀ/ꢁ ratio of 1:3 (entry 1) with respect to coupling with the acetyl-protected sugar (1) and an
ꢀ/ꢁ ratio of 2:1 (entry 2) with respect to coupling with the benzyl-protected sugar (2). In the
absence of neighboring group participation, the anomeric effect drives the ratio of axial ꢀ
products for the benzyl-protected sugar and accounts for the formation of the dominant ꢀ product
in 6. The C2 acetate group of 1 effectively increases the proportion of the equatorial (ꢁ) hydroxyl
5

group in the free sugar by way of blocking the axial site via neighboring group participation,
which consequently drives the reaction to the preferential ꢁ form in 5. Complete ꢀ selectivity
was achieved with the coupling of 4 and benzyl-protected sugar (2) using the DMAP/CH Cl
2
2
conditions (entry 4), although the yields were significantly reduced relative to the n-BuLi/THF
coupling conditions with 4 (entry 2). The enhanced selectivity can be attributed to the phenyl
protecting groups on 4, which make the thiophosphate moiety more electron rich, driving it to
the ꢀ position via the anomeric effect.
The S-linked isomer 7b was isolated in an 8% yield following purification relative to the
O-linked 7a (16%) using the preferred DMAP/CH Cl coupling conditions (combined yield =
2
2
2
4%) (entry 4). For the purification of 7a and 7b, triethylamine neutralization of the silica using
a 95:5 hexane:triethylamine slurry was required. In the absence of triethylamine, the material
was found to break down completely during purification and only the hydrolysis products were
recovered. The O-linked species was found to breakdown in the presence of CDCl to the free
3
benzyl sugar and phenyl thiophosphoric acid within 24 hours, as was observed using TLC
analyses, however, the compounds were found to be more stable in CD Cl . In CD Cl an
2
2
2
2
interesting isomerization (Scheme 4) with sugar-thiophosphate 7a was observed, in which the
exclusively ꢀ 7a compound gradually interconverted to the S-linked species 7b with a resulting
ꢀ/ꢁ ratio of 4:1 within 48 hours, after isolation. This transformation was observed by the large
31
up-field shift in the P NMR spectrum when the phosphorous-sulfur bond shifts from a double
bond, as in the case of the O-linked species, to a single bond as in the case of the S-linked
3
1
species. The P NMR spectra overlay of the various forms of the benzyl glucosyl thiophosphates
1
(
6–7) is shown in Figure 1. Compound 7b was confirmed to be the S-linked isomer by H NMR
3
spectroscopy, which showed the H-P coupling of the ꢀ O-linked species ( J1 = 9.5 Hz, 6.20
, P
3
ppm) had changed and shifted slightly up field ( J1 = 11.1 Hz, 6.30 ppm). An overlay of the H1
, P
splitting patterns for compounds (6, 7a and 7b) is shown in Figure 2. The S-glycosyl
thiophosphate appeared to be the more stable isomeric form and was not as vulnerable to
glycosidic cleavage by CDCl relative to its O-glycosyl counterpart.
3
6

Scheme 4 Proposed isomerization of glucopyranosyl-thiophosphate (7a)
3
1
Figure 1 P NMR spectra for glucosyl thiophosphates 6 (A) in CDCl , 7a (B) and 7b (C) in
3
CD Cl .
2
2
7

1
Figure 2 H NMR spectra for glycosyl thiophosphates 6 (A) in CDCl ; 7a (B) and 7b (C) in
3
CD Cl displaying ꢀ and ꢁ splitting patterns. The (*) indicates the start of isomerization from 7a
2
2
to 7b ~6 hours after isolation.
The significantly higher yields of the diethylthiophosphates relative to the diphenyl
thiophosphate prompted an attempt at separating the ꢀ and ꢁ diastereomers of 6. Many possible
separation conditions were explored using various elutants (EtOAc/hexane, Water/MeOH), as
well as normal phase silica, silver impregnated silica, reversed-phased C18 silica, alumina and
basic alumina media conditions. Under all circumstances the product was either recovered as a
diastereotopic mixture of anomers or hydrolyzed during the separation.
Despite unsuccessful separation attempts, the ꢀ/ꢁ mixtures of glycosyl thiophosphates
may not be of crucial concern. Given the stereospecificity of nucleotidylyltransferases and
glycosyltransferases, they can potentially select and drive the ꢀ/ꢁ equilibrium toward the
preferred form (ꢀ or ꢁ) of the glycosyl thiophosphate. After attempted separation of the anomers,
deprotection of the glycosyl thiophosphates was explored. The attempted literature deprotection
1
9
conditions employed with benzyl sugar product are summarized in Table 2. Basic LiOH
2
0
21
conditions were found to be unsuccessful and the acidic conditions of TMSI, TMSBr and
2
2
BCl3 were all found to be too harsh, resulting in anomeric cleavage of the product.
Hydrogenolysis of the products was explored next in the hope that it would remove the benzyl
protecting groups on the sugar moiety and/or the phenyl protecting groups on the thiophosphate
8

moiety. While compounds containing sulfur have been well known to poison hydrogenation
2
3,24
catalysts,
the incorporation of ammonium formate into the reaction media has been shown to
2
5
circumvent this problem in certain circumstances involving sulfate hydrogenolysis. In our
hands none of these hydrogenolysis conditions enabled deprotection of the glucosyl
thiophosphates.
Table 2 Attempted deprotection conditions for ꢀ/ꢁ-D-glucopyranosyl diethylthiophosphates
Compound
Deprotection Conditions
Result
6
6
6
LiOH(sat), dioxane/water, 50°C
No reaction
TMSI, CH
TMSBr, CH
BCl , CH Cl
Pd/C, MeOH/EtOAc
PtO , MeOH/EtOAc
, NH CO (1-15 eq), MeOH/EtOAc
Cl
Cl
, -78°C
, 0°C
Anomeric cleavage
Anomeric cleavage
Anomeric cleavage
No reaction
2
2
, 0°C
2
2
7
7
7
7
a, 7b
a, 7b
a, 7b
a, 7b
3
2
2
H
2
H
2
No reaction
2
H
2
Pd/C or H
PtO
No reaction
2
2
4
2
We next turned our synthetic efforts towards the preparation of a non-scissile, isosteric
analogue of ꢀ-D-glucose 1-phosphate containing sulfur. The synthesis of glucosyl
thiophosphonate (12) was accomplished from globally protected glucosyl phosphonate (11),
2
6
which was prepared as described previously. Lawesson’s Reagent (10) was used in the
transformation of 9 to 11. The use of 10 for the replacement of the oxo group of phosphorous
2
7
(
P=O) with the thio (P=S) group is relatively common. While many reports require refluxing
2
7,28
toluene conditions to affect transformation (P=O to P=S),
the optimal conditions were found
to be heating at 55°C for 12 hours, with dry toluene. Compound 11 was isolated in 72% yield
after purification (Scheme 5). The robust C—P thiophosphonate anomeric linkage of 11,
compared to the protected glucosyl thiophosphates O—P linkage of 5-7, facilitated easy global
deprotection. This was accomplished using TMSI, followed by purification via cellulose
chromatography to afford 12 in a 42% yield.
9

Scheme 5 Synthesis of ꢀ-D-glucose 1C-thiophosphonate
-D-glucose 1C-thiphosphonate (12) was evaluated as
ꢀ
a
substrate for
thymidylyltransferase Cps2L, and compared to ꢀ-D-glucose 1-phosphate (14) and its
phosphonate analogue ꢀ-D-glucose 1C-phosphonate (13) (Scheme 2). Compounds 12-14 were
coupled with deoxythymidine triphosphate (15, dTTP) to produce sugar nucleotides 16-18.
Conversions were monitored at 30 minutes and 24 hours by HPLC analysis (Table 3). At varying
concentrations of Cps2L (2 EU, 7 EU, 12 EU), the maximum conversion of glucosyl
thiophosphonate 12 to sugar nucleotide 16 was 15% (entry 1). The HPLC analysis did not
differentiate between whether oxygen or sulfur had acted as the nucleophile, giving rise to two
potential products (X, Y = O, Z = S) or (X, Z = O, Y = S). Analysis of the ESI-MS/MS
-
fragmentation pattern for the [M-H] product peak injected from an aliquot of the enzyme assay
revealed that oxygen operated as the nucleophile, rather than the sulfur atom, confirming the
structure of the sugar nucleotide analogue as compound 16 (Figure 3). If the sulfur atom were
-
acting as the nucleophile, an [NMP] fragment of 377, as opposed to 321, would have been
observed. It is conceivable that the ꢁ-thiophosphonate in 16 may exist as two diastereomers at
the ꢁ-phosphorus. Our HPLC and mass spectrometric analyses were unable to differentiate
between these two forms, however, the observation of only one peak by HPLC suggests
tentatively that the compound exists as only one diastereomer at the ꢁP position. The GlmU
catalyzed synthesis of UDP(ꢁS)-GlcNAc did not report whether two diastereomers were formed,
31
16
nevertheless, inspection of their P NMR suggests a single diastereomeric product. If this
discovery is corroborated, it demonstrates that both GlmU and Cps2L, and potentially this class
of enzyme, are able to distinguish between the pro-R and pro-S thiophosphonate oxygen atoms.
We have recently demonstrated that Cps2L has approximately a 10-fold difference in reactivity
1
7
between the monofluoromethylene analogues of glucose ketosephosphonates, demonstrating
that the Cps2L active site is able to discriminate between isosteric non-physiological substrates
1
0

possessing chirality where no chirality is present in the physiological substrate. Glucosyl
phosphonate 13 showed high turnover to its sugar nucleotide product 17, as has been previously
2
6
reported (entry 2), and was a more active substrate than compound 12.
Table 3 Comparison of conversions to sugar nucleotides for ꢀ-D-glucose 1-phosphate analogues
Entry
Compound
Product
Conversion 30 min
Conversion 24 h
1
2
3
12
13
14
16
17
18
2%
15%
96%
95%
95%
100%
Figure 3 Tandem mass spectrometry (ESI-MS/MS EPI) scan of Cps2L product 16.
3
. Conclusion
1
1

Several methods for the synthesis of glucosyl-thiophosphates were explored and evaluated. It
was found that the coupling method as well the protecting groups of the carbohydrates and
thiophosphates had a significant impact on both the stereoselectivity and yield of the reactions.
The diastereoselectivity did not appear to be effected by varying the temperature conditions.
Separation of ꢀ/ꢁ diastereomers was not achieved for the protected thiophosphates. The first
example of isomerization from an O-linked glucosyl thiophosphate to an S-linked species was
reported. Deprotection of the various glucosyl thiophosphates was unsuccessful due to the
fragility of the glycosidic bond or poisoning of the hydrogenolysis catalyst by sulfur. This
highlights the need for alternative deprotection strategies or modified thiophosphorylating
reagents. By contrast, ꢀ-D-Glucose 1-thiophosphonate was successfully synthesized and
converted to a thiophosphonate-containing sugar nucleotide analogue utilizing the
thymidylyltransferase Cps2L. The first example of a glycosyl thiophosphonate containing sugar
nucleotide analogue was reported. The enzymatic formation of compounds which incorporate
both phosphonate and thio- functionalities into their scaffold introduces a greater diversity of
non-scissile, isosteric, sugar nucleotide analogues. These types of compounds may have
biological applications as probes for glycosyltransferases or other related enzymes.
4
4
. Experimental
.1. General Methods
All reagents and solvents were purchased and used without further purification. Synthetic
reactions were performed under N atmosphere unless otherwise indicated. Reaction progress
2
was generally monitored by thin-layer chromatography using silica gel 60 F254 plates (Silicycle).
Plates were visualized using a an ethanol dipping solution containing p-anisaldehyde (3.4%),
sulfuric acid (2.2%) and acetic acid (1.1%) followed by charring or ultraviolet light (ꢃ = 254
nm). Dry ice/acetone and dry ice/acetonitrile slurries were used to generate reaction temperature
conditions of -78 and -40°C respectively. For attempted diastereomeric separation, the silver
impregnated silica was generated by stirring normal silica in a 1 M solution of AgNO , followed
3
by decantation of the solvent. The silica was then dried in the oven over night at 60°C and
column loaded in a 10/90 EtOAc/hexane slurry. Hydrogenation reactions were generally
1
2

performed on 10–30 mg of glycosyl-thiophosphate with platinium(IV) oxide (80 mol%) or
palladium on carbon (20 mol%) in 10 mL of 1:1 EtOAc:EtOH or 1:1 EtOAc:MeOH respectively.
The mixtures were shaken under H in a Parr apparatus at 54 PSI for 5–18 h. The mixtures were
2
3
1
then filtered and TLC analysis, P NMR spectroscopy or Low Resolution Mass Spectrometry
was used as an indication of reaction progress. Bruker AV-300 or AV-500 MHz NMR
spectrometers were utilized for all NMR spectra via the Nuclear Magnetic Resonance Research
3
1
Resource (NMR ). For CDCl and CD Cl , H NMR chemical shifts are reported as d in units of
3
2
2
3
1
parts per million (ppm) downfield from tetramethylsilane (ꢀ 0.0). P NMR chemical shifts are
reported as ꢀ in units of parts per million (ppm) relative to 85% H PO . Low resolution mass
3
4
spectra were obtained with an Applied Biosystems hybrid triple quadrupole linear ion trap
ionization (ESI) source. The mass spectrometer was coupled with an Agilent 1100 HPLC
instrument with a Phenomenex Kinetex 2.6 u (150 x 2.10 mm) column. Samples were run using
a flowrate of 120 ꢄL/min and an isocratic 70/30 CH CN/2 mM aqueous ammonium acetate (pH
3
5
.5) buffer. The capillary voltage was set to ±4500 kV, with a declustering potential of ±60V and
a curtain gas of 20 (arbitrary units) for positive and negative mode, respectively. Analyst version
.4.1 (Applied Biosystem) software was used for analysis. HPLC analysis for all enzymatic
1
2
9
assays was performed as previously described using a Hewlett Packard Series 1050 instrument
with an Agilent Zorbax 5 ꢄm Rx-C18 column (150 x 4.6 mm). Compounds containing a
nucleotide base chromophore were monitored at 254 nm absorbance. HPLC runs were performed
for 15 min at a flowrate of 1 mL/min. A linear gradient from 90/10 A/B to 40/60 A/B over 8 min
followed by a 40/60 A/B plateau from 8-10 min and a linear decline to 90/10 A/B from 10-11
min followed by isocratic 90/10 A/B was used for all assays. Buffer A was aqueous 12 mM
Bu NBr, 10 mM KH PO AND 5% HPLC grade CH CN. Buffer B was HPLC grade CH CN.
4
2
4
3
3
4
.1 HPLC enzyme assay conditions:
2
9
Cps2L was over-expressed, purified and assayed as previously described. The concentration of
Cps2L was measured spectrophotometrically at 280 nm using calculated extinction coefficient
-1
-1
2
9.8 mM cm . Enzyme reactions containing dTTP (1.0 mM) ꢀ-D-glucose 1-phosphate
analogues 12-14 (2.0 mM), MgCl (2.2 mM) and inorganic pyrophosphatase (0.5 EU) were
2
initiated by addition of Cps2L (2 EU) in Tris-HCl buffer (20 mM final concentration, 50 µL final
volume). The reactions were incubated for 30 minutes or 24 hours at 37°C, quenching with
1
3

methanol (50 µL), and centrifuged (5 min at 12,000 x g) to precipitate the denatured enzymes
prior to HPLC analysis. Percent conversions were calculated as the ratio of sugar nucleotide
product (16-18) formed relative to remaining dTTP. In the absence of Cps2L, no turnover to
sugar nucleotide product was observed. Retention times can be found in the supporting
information.
4
4
.2. Synthesis
.2.1. Diphenylchlorothiophosphate (4)
Diphenylchlorothiophosphate was synthesized using a previously established literature
3
0
procedure. Phenol (1.88 g, 20.0 mmol) was dissolved in CH Cl (5.2 mL) and triethylamine
2
2
(2.8 mL (20.0 mmol). Half of the solution was added dropwise over 15 minutes to a solution of
PSCl (1 mL, 10.0 mmol) in CH Cl (3 mL) at 0°C. After 2.5 hours, TLC showed full
3
2
2
consumption of phenol so the remaining half of the solution was added to the reaction mixture
dropwise and the solution was allowed to stir overnight. TLC analysis after 18 hours showed one
major product by TLC (Rf = 0.63 10/90 EtOAc/hexane). The material was isolated directly using
normal phase chromatography (100% hexane) and collected as a white crystalline solid (1.400 g,
5
0%). ꢅ (CDCl ) 58.8 (s, 1P, P1).
P 3
4
.2.2. 2,3,4,6-Tetra-O-acetyl-ꢀ/ꢁ-D-glucopyranosyl diethylthiophosphate (5)
2
,3,4,6-Tetra-O-acetyl-ꢁ-D-glucopyranose (1) (0.35 mmol) was dissolved in anhydrous THF (2
mL) and cooled to -78°C using a dry ice/acetone slurry. n-Butyllithium (0.17 mL, 0.43 mmol)
was added and the solution was stirred for one hour before diethylchlorothiophosphate (3) (0.063
mL, 0.4 mmol) was added dropwise. The solution turned a red-brown color upon warming to
room temperature. The reaction was stirred overnight before being diluted with ice cold water
(10 mL) and extracted with dichloromethane (10 mL). The organic extract was washed with ice
cold aqueous HCl (0.5 M, 15 mL) and saturated sodium bicarbonate solution (15 mL). After
drying with anhydrous sodium sulfate, the mixture was concentrated to a brown liquid. The
material was purified using normal phase chromatography (isocratic 30/70 EtOAc/hexane) to
afford the pure product 5 as a clear liquid, 1:3 ꢀ/ꢁ diastereotopic mixture (0.204g, 57% yield), RF
3
=
0.58 (40/60 EtOAc/hexane). ꢅ (CDCl ) 5.96 (dd, 0.25H, J = 3.4 Hz, J1, P = 9.8 Hz, ꢀH1),
H
3
1, 2
3
3
5
5
.48 (t, 0.25H, J = J4, 5 = 9.8 Hz, ꢀH4), 5.34 (dd, 0.75H, J = 7.9 Hz, J1,P = 10.6 Hz, ꢁH1),
3, 4 1, 2
3
3
.25 (t, 0.75H, J = J = 9.8 Hz, ꢁH4) 5.06-5.15 (m, 2H, H2, H3), 5.00 (m, 0.25H, ꢀH5)
3
, 4
4, 5
1
4

4
3
ꢁ
.00-4.30 (m, 6H, 2OCH , H6a, H6b), 3.82 (m, 0.75H, ꢁH5), 2.07, 2.04, 2.03, 2.01 (4 x s, 4 x
2
H, 4 x OAc), 1.28-1.39 (m, 6H, 2 x CH ), ꢅ 169.3-171.0 (4 x C, C(O)OCH ), 96.5 (d, CH,
3
C
3
2
2
C1, J = 3.6 Hz), 94.0 (d, CH, JC, P = 4.1 Hz, ꢀC1), 60.3-72.7 (4 x CH, C2-C5, 3 x CH , C6,
C, P
2
OCH CH ), 14.3-32.0 (2 x CH , OCH CH , 4 x CH , C(O)OCH ), ꢅ 67.09 (s, 0.25P ꢀP=S),
2
3
3
2
3
3
3
P
+
+
6
6.94 (s, 0.75P, ꢁP=S), HRMS (ESI ): found [M+Na] 523.0999. C H Na O P S requires
18 29 1 12 1 1
+
[
M+Na] 523.1010.
4
.2.3. 2,3,4,6-Tetra-O-benzyl-ꢀ/ꢁ-D-glucopyranosyl diethylthiophosphate (6)
,3,4,6-Tetra-O-benzyl-ꢀ/ꢁ-D-glucopyranose (2) (0.270 g, 0.50 mmol) was dissolved in
2
anhydrous THF (3 mL) under nitrogen and cooled to -78°C using a dry ice/acetone slurry. n-
Butyllithium (0.22 mL, 0.55 mmol) was added and the solution was stirred for one hour before
diethylchlorothiophosphate (3) (0.1 mL, 0.64 mmol) was added dropwise. The reaction was
allowed to come to room temperature overnight. The solution was poured into a suspension of
diethyl ether (10 mL) and water (10 mL) and extracted with ethylacetate (7 mL) before being
washed with saturated sodium bicarbonate (2 x 10 mL) and dried with anhydrous sodium sulfate.
The mixture was concentrated in vacuo and collected as a clear liquid. The material was then
purified using normal phase chromatography (isocratic 9/91 EtOAc/hexane) to afford the product
(
6) as a clear liquid, 2:1 ꢀ/ꢁ diastereotopic mixture (0.241 g, 67% yield), R = 0.32 (10/90
F
3
3
EtOAc/hexane). ꢅ (CDCl ) 7.20-7.50 (m, 20H, 4 x C H ), 6.13 (dd, 0.67H, J = 3.3 Hz, J
H
3
6
5
1, 2
1, P
3
3
=
9.8 Hz, ꢀH1), 5.4 (dd, 0.33H, J
= 7.9 Hz, J
= 10 Hz, ꢁH1), 4.50-5.10 (m, 8H, 4 x
1, P
1
, 2
PhCH ), 4.20 (m, 4H, 2OCH ) 4.00-4.10 (m, 2H, H3, H4), 3.60-3.86 (m, H2, H5, H6a, H6b)
2
2
1
.28-1.39 (m, 6H, 2 x CH ), ꢅ 137.7-138.7 (4 x C, C H ), 127.7-128.6 (20 x CH, C H ), 99.2 (d,
3 C 6 5 6 5
2
2
CH, J = 5.7 Hz, C1ꢁ), 95.8 (d, CH, J = 5.7 Hz, C1ꢀ), 68.0-84.8 (5 x CH , CH Ph, C6, 4 x
C, P
C, P
2
2
CH, C2-C5), 64.3-64.7 (m, OCH CH ), 15.8-16.0 (m, OCH CH ), ꢅ 67.80 (s, 0.33P, ꢁP=S),
2
3
2
3
P
+
+
6
7.38 (s, 0.67P, ꢀP=S), HRMS (ESI ): found [M+Na] 715.2448. C H Na O P S requires
38
45
1
8 1 1
+
[
M+Na] 715.2465.
4
.2.4. 2,3,4,6-Tetra-O-benzyl-ꢀ/ꢁ-D-glucopyranosyl diphenylthiophosphate (7)
,3,4,6-Tetra-O-benzyl-ꢀ/ꢁ-D-glucopyranose (2) (0.270 g, 0.5 mmol) and 4-
2
dimethylaminopyridine (1.293 g, 1.23 mmol) were dissolved in dichloromethane (10 mL) and
o
stirred under N at 0 C for 15 minutes. A solution of diphenylchlorothiophosphate (4) (0.321 g,
2
o
2
.28 mmol) in CH Cl (5 mL) was added dropwise and stirred under N at 0 C for 3 hours The
2
2
2
.
1
5

reaction was quenched with distilled water (15 mL), extracted with dichloromethane (2 x 15 mL)
and dried with anhydrous sodium sulfate. The mixture was concentrated in vacuo and collected
as a clear viscous liquid. The material was purified using normal phase chromatography
(isocratic 20/80 EtOAc/hexane) using basic silica (5% triethylamine, 95% hexanes) to afford two
products (0.095 g, combined yield 24%). The oxygen linked thiophosphate 7a R = 0.74, and the
F
sulfur linked thiophosphate 7b R = 0.52 (25% EtOAc/75% hexane).
F
1
-O-Diphenylthiophosphate-2,3,4,6-tetra-O-benzyl-ꢀ/ꢁ-D-glucopyranose (7a)
3
3
ꢅ (CD Cl ) 7.15-7.45 (m, 30H, 6 x C H ), 6.20 (dd, 1H, J = 3.3 Hz, J = 9.5 Hz, H1),
H 1, P
2
2
6
5
1, 2
4
1
.48-5.03 (m, 8H, 4 x PhCH ), 3.97-4.16 (2H, H3, H4), 3.52-3.83 (m, 4H, H2, H5, H6a, H6b), ꢅC
2
37.5-138.7 (6 x C, C H ), 121.0-129.0 (30 x CH, C H ), 93.7 (CH, C1), 67.9-81.6 (5 x CH ,
6
5
6
5
2
+
+
CH Ph, C6, 4 x CH, C2-C5), ꢅ 57.86 (s, 1P, P1), HRMS (ESI ): found [M+Na] 811.2451
2
P
+
C H Na O P S requires [M+Na] 811.2465.
46
45
1
8 1 1
1
-S-Diphenylthiophosphate-2,3,4,6-tetra-O-benzyl-ꢀ/ꢁ-D-glucopyranose (7b)
3
3
ꢅ (CD Cl ) 7.10-7.40 (m, 30H, 6 x C H ), 6.30 (dd, 0.80 H, J = 5.0 Hz, J = 11.1 Hz,
H 1, P
2
2
6
5
1, 2
3
3
ꢀ
3
1
H1), 5.07 (dd, 0.20 H, J = 9.7 Hz, J
= 14.2 Hz, ꢁH1), 4.96-4.39 (m, 8H, 4 x PhCH2),
1, P
1, 2
2
3
.51-3.91(m, 5H, H2, H3, H4, H5, H6a), 3.20 (dd, 1H, J6
= 1.76, J
= 10.91, H6b), ꢅC
6b, 5
b, 6a
35.4-139.0 (6 x C, C H ), 120.4-130.4 (30 x CH, C H ), 67.9-82.7 (5 x CH , CH Ph, C6, 4 x
6
5
6
5
2
2
+
+
CH, C2-C5), ꢅ 18.77 (s, 0.80P, ꢀP=O), 16.69 (s, 0.20P, ꢁP=O), HRMS (ESI ): found [M+Na]
P
+
8
11.2462 C H Na O P S requires [M+Na] 811.2465.
46
45
1
8 1 1
Diethyl (tetra-O-benzyl-ꢀ-D-heptulopyranosyl)-1-thiophosphonate (11)
To a suspension of 9 (0.099 g, 0.146 mmol) in toluene (8 mL) was added Lawesson's reagent
10) (0.059 g, 0.146 mmol) in one portion. The reaction was warmed to 55°C for 12 h and then
(
cooled to room temperature. The yellow reaction mixture was filtered, concentrated in vacuo and
purified using normal phase chromatography (isocratic 15/85 EtOAc/hexane) to afford the
product (9) as a white solid (0.072 g, 71% yield), R = 0.61 (80/20 EtOAc/hexane). ꢅ (CDCl )
F
H
3
3
7
.17-7.39 (m, 20H, 4 x C H ), 4.45-4.87 (m, 9H, 4 x PhCH , H2), 4.16 (q, 2H, J
= 7.0 Hz,
H, H
6
5
2
3
OCH CH ), 4.12 (q, 2H, J
= 7.0 Hz, OCH CH ), 3.66-3.82 (m, 6H, H3, H4, H5, H6, H7a,
2 3
2
3
H, H
1
6

3
3
H7b), 2.47 (m, 2H, H1a, H1b), 1.33 (t, 3H, J
= 7.0 Hz, OCH CH ), 1.28 (t, 3H, J
= 7.0
H, H
H, H
2
3
Hz, OCH CH ), ꢅ 127.7-138.8 (4 x C, C H , 20 x CH, C H ), 72.1-82.2 (4 x CH , CH Ph, 3 x
2
3
C
6
5
6
5
2
2
3
2
CH, C4-C6), 79.3 (d, CH, J = 14.1 Hz, C3), 70.2 (d, CH, J = 3.2 Hz, C2), 68.7 (CH , C7),
3
, P
2, P
2
2
2
6
2.7 (d, CH , J = 6.9 Hz, OCH CH ), 62.6 (d, CH , J = 6.9 Hz, OCH CH ,), 30.5 (d, CH ,
2 C, P 2 3 2 C, P 2 3 2
1
+
+
J1, P = 116.0 Hz, C1), 16.4 (m, OCH CH ), ꢅ 96.81 (s, 1P, P1), HRMS (ESI ): found [M+Na]
2
3
P
+
7
13.2494 C H Na O P S requires [M+Na] 713.2672.
39 47 1 7 1 1
Ammonium-(1-deoxy-ꢀ-D-heptulopyranosyl)thiophosphonate (12)
Under N at 0ºC, a stirring solution of 11 (0.072 g, 0.105 mmol) in anhydrous CH Cl (0.5 mL)
2
2
2
was treated with iodotrimethylsilane (740 ꢄL, 5.25 mmol) and allowed to warm to room
temperature over 2 h. The reaction was monitored by TLC analysis using 6:3:1 n-propanol-NH3-
H O (Rf: 0.70), and quenched by methanol. The mixture was concentrated in vacuo, dissolved in
2
H O (10 mL) and washed with diethyl ether (8 x 10 mL). The resulting acidic aqueous layer was
2
immediately adjusted to pH 8 with aqueous NH OH (0.2 M), concentrated to 5 mL and purified
4
by cellulose column (1.5 x 10 cm, 2.5 g) (n-propanol-NH -H O, 80/10/10). The resulting fraction
3
2
was lyophilized to afford the target phosphonate 10 as a white film (0.013 g, 42%). ꢅ (D O)
H
2
2
4
.43 (m, 1H, H2), 3.68 (d, 1H, J
= 11.2 Hz, H7a), 3.44-3.56 (m, 4H, H3, H4, H6, H7b), 3.37
a, 7b
7
2
(
m, 1H, H5), 2.41 (m, 2H, H1a, H1b), ꢅ 71.4-73.5 (2 x s, 2 x CH, C4, C6), 72.1 (d, CH, J
=
C
2, P
1
3
4
.1 Hz, C2), 70.3 (d, CH, J = 13.7 Hz, C3), 69.5 (CH, C5), 60.4 (CH , C7), 28.7 (d, CH , J
3
, P
2
2
1, P
-
-
=
112.0 Hz, C1), ꢅ 87.90 (s, 1P, P1), HRMS (ESI ): found [M-1] 257.0463 C H O P S
P
7
14
7
1
1
-
requires [M-1] 257.0245.
Supplementary Data
Supplementary data (NMR spectra for compounds 5-7, 11-12, and HPLC retention times
for sugar nucleotides and related compounds).
References
1
2
. Pale, P.; Whitesides, G. M. J. Org. Chem. 1991, 56, 4547-4549.
. Fujita, S.; Oka, N.; Matsumura, F.; Wada, T. J. Org. Chem. 2011, 76, 2648–2659.
1
7

3
4
5
6
. Eckstein, F. J. Amer. Chem. Soc. 1970, 92, 4718-4723.
. Goody, R. S.; Eckstein, F. J. Amer. Chem. Soc. 1971, 93, 6252-6257.
. Cullis, P. M.; Maxwell, A., Weiner, D. P. Biochemistry 1997, 36, 6059-6068.
. Baxter, N. J.; Olguin, L. F.; Golicnik, M.; Feng, G.; Hounslow, A. M.; Bermel, W.;
Blackburn, G. M.; Hollfelder, F.; Waltho, J. P.; Williams, N. H. Proc. Natl. Acad. Sci.
2
006, 103, 14732-14737.
. Xu, Y.; Jiang, G.; Tsukahara, R.; Fujiwara, Y.; Tigyi, G.; Prestwich, G. D. J. Med. Chem.
006, 49, 5309-5315.
. Nakai, H.; Kitaoka, M.; Svensson, B.; Ohtsubo, K. Curr. Opin. Chem. Biol. 2013, 17
01-309.
. Levander F.; Andersson, U.; Radstrom, P. Appl. Environ. Microbiol. 2001, 67, 4546–
553.
7
8
9
2
3
4
1
1
0. Sjoberg, A.; Hahn-Hagerdal, B.; Appl. Environ. Microbiol. 1989, 55, 1549–1554.
1. Alphey, M. S.; Pirrie, L.; Torrie, L. S.; Boulkeroua, W. A.; Gardiner, M.; Sarkar, A.;
Maringer, M.; Oehlmann, W.; Brenk, R.; Sherman, M. S.; McNeil, M.; Rejzek, M.; Field,
R. A.; Singh, H. M.; Gray, D.; Westwood, N. J.; Naismith, J. H. ACS Chem. Biol. 2013,
8
, 387−396.
1
1
1
1
1
2. Qu, H.; Xin, Y.; Dong, X.; Ma, Y. FEMS Microbiol. Lett. 2007, 275, 2437-243.
3. Melo, A.; Glaser, L. J. Biol. Chem. 1965, 240, 398-405.
4. Zhang, G.; Yu, B.; Deng, S.; Hui, Y. J. Carbohydr. Chem. 1998, 17, 547-556.
5. Piekutowska, M.; Pakulski, Z. Carbohydr. Res. 2008, 343, 785-792.
6. Cai, L.; Ban, L.; Guan, W.; Mrksich, M.; Wang, P. W. Carbohydr. Res. 2011, 346, 1576-
1
580.
1
1
1
7. Forget, S. M.; Bhattasali, D.; Hart, V. C.; Cameron, T. S.; Syvitski, R. T.; Jakeman, D. L.
Chem. Sci. 2012, 3, 1866-1878.
8. Huestis, M.; Aish, G.; Hui, J.; Soo, E.; Jakeman, D. L. Org. Biomol. Chem., 2009, 6, 477-
4
84.
9. Coppola, C.; Saggiomo, V.; Di Fabio, G.; De Napoli, L.; Montesarchio, D. J. Org. Chem,
007, 72, 9679-9689.
2
1
8

2
0. Heissigerova, H.; Kocalka, P.; Hlavackova, M.; Imberty, A.; Breton, C.; Chazalet, V.;
Moravcova, J. Collect. Czech. Chem. Commun. 2006, 71, 1659-1672.
2
2
1. McKenna, C. E.; Schmidhauser, J. J. Chem. Soc., Chem. Commun., 1979, 0, 739.
2. Elgibo-Gabas, M.; Canals, D.; Casas, J.; Llebaria, A.; Delgado, A. ChemMedChem 2007,
2
, 992-994.
2
2
2
2
3. Albers, P.; Pietsch, J.; Parker, S. F. J. Mol. Catal. A: Chem. 2001, 173, 275-286.
4. Vassilakis, D.; Barbouth, N.; Oudar, J. Catal. Lett. 1990, 5, 321-330.
5. Ali, A. M.; Taylor, S. D. J. Pept. Sci. 2010, 16, 190-199.
6. Beaton, S. A.; Huestis, M. P.; Sadeghi-Khomami, A.; Thomas, N. R.; Jakeman, D. L.
Chem. Commun. 2009, 238-240.
2
2
2
7. Ozturk, T.; Ertas, E.; Mert, O. Chem. Rev. 2007, 107, 5210-5278.
8. Karp, G. M. J. Org. Chem. 1999, 64, 8156-8160.
9. Timmons, S. C.; Mosher, R. H.; Knowles, S. A.;; Jakeman, D. L. Org. Lett. 2007, 9, 857-
8
60.
3
0. Hoque, M. E. U.; Shuchismita, D.; Guha, A. K.; Kim, C. K.; Lee, B. S.; Lee, H. W. J.
Org. Chem. 2007, 72, 5493-5499.
1
9

Highlights
Preparation of protected forms of glucose 1-thiophosphate
Isomerization from an O-linked to S-linked glucosyl thiophosphate
Preparation of glucose 1-thiophosphonate
Enzymatic conversion of glucose 1-thiophosphonate by thymidylyltransferase