Synthesis and evaluation of l-rhamnose 1C-phosphonates as nucleotidylyltransferase inhibitors

Article abstract of DOI:10.1021/jo401542s

We report the synthesis of a series of phosphonates and ketosephosphonates possessing an l-rhamnose scaffold with varying degrees of fluorination. These compounds were evaluated as potential inhibitors of α-d-glucose 1-phosphate thymidylyltransferase (Cps2L), the first enzyme in Streptococcus pneumoniae l-rhamnose biosynthesis, and a novel antibiotic target. Enzyme-substrate and enzyme-inhibitor binding experiments were performed using water-ligand observed binding via gradient spectroscopy (WaterLOGSY) NMR for known sugar nucleotide substrates and selected phosphonate analogues. IC ₅₀ values were measured and K_i values were calculated for inhibitors. New insights were gained into the binding promiscuity of enzymes within the prokaryotic l-rhamnose biosynthetic pathway (Cps2L, RmlB-D) and into the mechanism of inhibition for the most potent inhibitor in the series, l-rhamnose 1C-phosphonate.

Full text of DOI:10.1021/jo401542s

Article
pubs.acs.org/joc
Synthesis and Evaluation of L‑Rhamnose 1C-Phosphonates as
Nucleotidylyltransferase Inhibitors
Matthew W. Loranger,† Stephanie M. Forget,† Nicole E. McCormick,‡ Raymond T. Syvitski,†,∥
and David L. Jakeman*,†,‡
†Department of Chemistry, Dalhousie University, 6274 Coberg Road, P.O. Box 15,000, Halifax, Nova Scotia B3H 4R2, Canada
‡College of Pharmacy, Dalhousie University, 5968 College Street, P.O. Box 15,000, Halifax, Nova Scotia B3H 4R2, Canada
∥National Research Council of Canada, 1411 Oxford Street, Halifax, Nova Scotia B3H 3Z1, Canada
S
* Supporting Information
ABSTRACT: We report the synthesis of a series of phosphonates
and ketosephosphonates possessing an L-rhamnose scaffold
with varying degrees of fluorination. These compounds were
evaluated as potential inhibitors of α-D-glucose 1-phosphate
thymidylyltransferase (Cps2L), the first enzyme in Streptococcus
pneumoniae L-rhamnose biosynthesis, and a novel antibiotic
target. Enzyme−substrate and enzyme−inhibitor binding experi-
ments were performed using water-ligand observed binding via
gradient spectroscopy (WaterLOGSY) NMR for known sugar nucleotide substrates and selected phosphonate analogues. IC50
values were measured and Ki values were calculated for inhibitors. New insights were gained into the binding promiscuity of
enzymes within the prokaryotic L-rhamnose biosynthetic pathway (Cps2L, RmlB−D) and into the mechanism of inhibition for the
most potent inhibitor in the series, L-rhamnose 1C-phosphonate.
Cps2L (EC 2.7.7.24) is a thymidylyltransferase responsible for
coupling α-D-glucose 1-phosphate (1, Glc-1-P) and deoxythymi-
dine triphosphate (2, dTTP) to form deoxythymidine
diphosphate-α-D-glucose (3, dTDP-Glc) (Figure 1). RmlB (EC
4.2.1.46) is responsible for oxidation of the C4″ hydroxyl unit and
C6″ dehydration. RmlC (EC 5.1.3.13) catalyzes an uncommon
double-epimerization at both C3″ and C5″. Finally, RmlD (EC
1.1.1.133) implements a reduction at C4″ to produce dTDP-β-L-
rhamnose (6) (Figure 2).13 dTDP-β-L-rhamnose has been shown
to act as an allosteric inhibitor of the thymidylyltransferase in the
INTRODUCTION
■
Streptococcus pneumoniae is a prevalent, highly infectious,
Gram-positive pathogen that has shown high clinical resistance
to penicillin and chloramphenicol.1 The mechanisms of
resistance for S. pneumoniae and other more invasive pathogens,
including Mycobacterium tuberculosis, usually entail alterations
to drug target proteins involved in cell wall synthesis.2−5
Bacterial chemo-resistance, including resistance to synergistic
treatments using β-lactams and aminoglycosides, is a growing
barrier to the treatment of invasive pathogens (i.e., S. pneumoniae
and M. tuberculosis).6
Cell wall biosynthesis is a commonly pursued target for a
variety of bacterial enzyme inhibitors as cell wall assembly
is essential for bacterial survival and virulence. L-Rhamnose (6-
deoxy-L-mannose) serves as the linking unit between arabinoga-
lactan and peptidoglycan7,8 and is a necessary constituent of the
bacterial cell wall in many bacterial species.9 The biosynthesis of
L-rhamnose begins with nucleotidylyltransferase enzymes
(Cps2L/RmlA) responsible for generating activated sugars in
the form of glycosylated nucleoside diphosphates (NDPs). These
NDPs are generated from the enzymatic coupling of sugar
1-phosphates with nucleoside triphosphates (NTPs). Once
formed, the NDPs may be further modified by intermediate
enzymes (RmlB−D) before serving as substrates for glycosyl-
transferases.10 Sugar 1-phosphate nucleotidylyltransferases are
vital to most biological glycosylation systems and are often used
for the chemo-enzymatic synthesis of sugar nucleotides.11
Nucleotidylyltransferases are an attractive antimicrobial target,
in that they show broad homology across various species.12
Figure 1. Physiological reaction catalyzed by Cps2L and ordered
Bi−Bi mechanism (inset).
Received: July 17, 2013
Published: September 10, 2013
© 2013 American Chemical Society
9822
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
mimics due to the variety of applicable synthetic methods.17 The
inductive effects of two fluorine atoms results in pKa2 values
between 5.0 and 6.0, which are much lower than the pKa2 values
of physiological phosphates. The preparation of monofluoro-
methylene phosphonates is much less widespread, in part due to
synthetic complexity.24−27 The monofluoromethylene function-
ality (pKa2 5.5−6.527) has been shown to better mimic the
electronic properties of a phosphate oxygen than the methylene
unit of the phosphonate (pKa2 ∼7.0−8.0).17 The variety of studies
probing enzyme systems with substituted phosphonates demon-
strates that there is no universal functionality best suited to mimic
a physiological phosphate.
The design of compounds with structural scaffolds that
enables them to bind multiple targets within a biosynthetic
pathway would be a desirable approach to thwart rising bacterial
drug resistance. Herein, we report the synthesis of a series of
L-rhamnose 1C-phosphonate compounds that utilize phospho-
nate and ketosephosphonate scaffolds, with varying degrees
of fluorination at the C1 position, as isosteric mimics of
β-L-rhamnose 1-phosphate. Select compounds were evaluated
for binding to Cps2L and RmlB−D using water-ligand observed
via gradient spectroscopy (WaterLOGSY) NMR and enzyme
assay techniques
Figure 2. Reactions catalyzed by Cps2L(RmlA) and RmlB−D.
biosynthetic pathway, with the allosteric site presumed to be the
point of regulatory control.9,14,15 A series of thymine-based small
molecule inhibitors was developed by Naismith and co-workers
and was found to exhibit nanomolar inhibition against RmlA
(Pseudomonas aeruginosa).9 To date, these are the most effective
reported inhibitors against nucleotidylyltransferases.
Many phosphonates have biomedical applications.16,17
Fosfomycin, a phosphonate containing natural product isolated
from soil bacteria in 1969, has broad spectrum antibacterial
activity.18 Dichloromethylene diphosphonic acid (chlodronic
acid) has been shown to inhibit RNA polymerase activity in
influenza,19 and both the related mono- and difluoromethylene
diphosphonates have shown activity as bone lysis inhibitors20
and antiviral agents.21,22 Methylene phosphonates have been
devised as nonhydrolyzable mimics of phosphates to probe a
variety of biological systems.16,17 Blackburn first proposed
halogenation of the methylene phosphonate functionality in an
effort to better mimic the properties of the natural phosphate.23
Difluoromethylene phosphonates are often prepared as phosphate
RESULTS AND DISCUSSION
■
Synthesis. The synthesis of all analogues was initiated from
lactone template 7 (Scheme 1), which was produced from
commercially available L-rhamnose following standard carbohy-
drate methylation, benzylation, and demethylation techniques
to generate 2,3,4,-tri-O-benzyl-α/β-L-rhamnopyranose.28,29
a
Scheme 1. Proposed Synthesis of L-Rhamnose Phosphonate and Ketosephosphonate Analogue Compounds
a
(a) CH3PO(OR)2, n-BuLi, THF, −78°C, 1.5 h, R = Me (88%), R = Bn (76%); (b) (i) H2, Pd−C, 18 h, rt, 1 psi; (ii) 6 M HCl(aq), 100°C, 3h; (iii)
NH4OH(aq), pH 8 (quant); (c) (i) H2, Pd−C, 18 h, rt, 1 psi; (ii) NH4OH(aq), pH 8 (quant); (d) MeOXCl, CH2Cl2, py, rt, 1.5 h, R = Me (68%),
R = Bn (60%); (e) (i) H2, Pd−C, 18 h, rt, 1 psi; (ii) 6 M HCl(aq), 100°C, 3h; (iii) NH4OH(aq), pH 8 (quant); (f) (i) H2, Pd−C, 18 h, rt, 1 psi; (ii)
NH4OH(aq), pH 8 (quant); (g) (i) Selectfluor, CH3CN, rt, 18 h; (ii) H2O, 70°C, 2.5 h, R = Me (54%), R = Bn (22%); (h) (i) H2, Pd−C, 18 h, rt,
1 psi; (ii) 6 M HCl(aq), 100°C, 3 h; (iii) NH4OH(aq), pH 8 (quant); (i) (i) H2, Pd−C, 18 h, rt, 1 psi (ii) NH4OH(aq), pH 8 (quant); (j)
F2CHPO(OEt)2, LDA, THF, −78°C, 0.5 h (74%); (k) (i)TMSI, CH2Cl2, rt, 5 h; (ii) NH4OH(aq), pH 8 (95%); (l) MeOXCl, CH2Cl2, py, rt, 1.5 h
(69%); (m) Bu3SnH, AIBN, tol, 105°C, 1 h (85%); (n) TMSI, CH2Cl2, rt, 5 h, (ii) NH4OH(aq), pH 8 (94%).
9823
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
The selectively deprotected sugar was then oxidized using
Albright−Goldman30 conditions, in a dimethyl sulfoxide
(DMSO) and acetic anhydride solvent mixture, to afford the
starting material 7. Lactone 7 was transformed into the dimethyl
or dibenzyl ketosephosphonate species 8a (88% yield) and 8b
(76% yield), using lithiated salts of dimethyl methylphosphonate
and dibenzyl methylphosphonate,31 respectively, in tetrahydro-
furan (THF) at −78 °C. Both compounds were isolated as the
and 10b were performed within 24 h of isolation. Treatment of
10a or 10b with Selectfluor33 electrophilic fluorinating reagent
afforded a diastereotopic mixture of monofluorinated ketose-
phosphonates. The major products 12a and 12b were isolated
in 54% and 22% yields, respectively. It was suspected that the
additional steric bulk of the benzyl protecting groups of 10b
reduced the fluorination efficiency, accounting for the reduced
yield of 12b relative to methyl protected 12a.
In previous studies, direct determination of phosphonate stereo-
chemistry for the D-glucose analogue of 12b was possible using
X-ray diffraction analysis.27 In this instance, however, both
the methyl (12a) and benzyl (12b) phosphonate protected
L-rhamnose analogues were clear viscous liquids, rendering X-ray
crystallographic techniques ineffective for stereochemical determi-
nation about the C1 position. 1D NOESY NMR experiments of
compounds 8a and 12a were hence used to deduce the stereo-
chemistry for compound 12a as R. By irradiating the hydroxy
proton of 8a, the methylene proton H1a produced an NOE
(Figure 4). Irradiation of the second methylene proton H1b
1
single axial hydroxyl isomer as determined using 31P and H
NMR spectroscopy. Olefins 10a and 10b were produced from
the in situ formation, and subsequent elimination, of the oxalyl
esters formed when ketosephosphonates 8a and 8b were
treated with methyl oxalyl chloride (MeOXCl) in a 7:2
CH2Cl2:pyridine solvent mixture. Complete deoxygenation of
the oxalyl esters was not achieved, as the reaction conditions
generally produced an approximate 2:1 ratio of the desired
olefin products to intermediate oxalyl esters. Previously, the use
of freshly distilled pyridine or further refluxing in toluene had
facilitated complete deoxygenation of related gluco-ketose-
phosphonate oxalyl esters.27 Such measures were found to be
ineffective for the L-rhamnose analogues. The intermediate oxalyl
esters were always visible as the less polar species by thin layer
chromatography (TLC) analysis. Olefins 10a and 10b were single
Z stereoisomers as determined by 31P and, 1H NMR spectroscopy,
and by 1D nuclear Overhauser effect spectroscopy (NOESY)
NMR experiments (Figure 3). Irradiation of the olefin proton
Figure 4. 1H NMR and 1D NOESY spectra (300 MHz, CDCl3) for 8a
and 12a. (A) 1H NMR spectrum for 8a. (B) 1D NOESY spectrum for
8a OH. (C) 1D NOESY spectrum for 8a H1b. (D) 1H NMR spectrum
for 12a. (E) 1D NOESY spectrum for 12a OH. Down arrow indicates
site of irradiation. Up arrow indicates a positive NOE. Structure letters
(B, C, and E) correspond with spectra letters. Structure bond lengths
are exaggerated for clarity. Spectra B and E show artifacts resulting from
trace H2O in the samples, which should not be confused with H1b.
Figure 3. 1H NMR and 1D NOESY NMR spectra (300 MHz, CDCl3)
for 10a. (A) 1H NMR spectroscopy for 10a. (B) 1D NOESY spectrum
for 10a H1. Down arrow indicates site of irradiation. Up arrow
indicates a positive NOE.
H1 of compound 10a produced an NOE at the equatorial H3
proton, suggesting a Z orientation about the double bond. If H1
had been in an E orientation, an NOE with the axial H6 and/or
L-rhamnose methyl CH3 moiety would have been expected. Such
effects were not observed. Determination of stereochemistry
about the double bond of olefin 10a has not been reported in its
previous synthesis.32
The L-rhamnose olefins (10a and 10b) were found to be
significantly more prone to CC hydration than the analogous
D-glucose compounds.27 If left open to atmosphere or in CDCl3
overnight, partial reversion to ketosephosphonates 8a and 8b
was observed by TLC analysis and 31P NMR spectroscopy. As a
result, NMR characterization of 10a and 10b was carried out in
CD2Cl2, and subsequent reactions or deprotections using 10a
showed only an NOE with H1a and not the hydroxyl proton
(Figure 4). It was thus reasoned that because only one of the
methylene protons of 8a showed An NOE between the hydroxyl
proton and only one of the methylene protons of 8a indicates
there is no free rotation about the C1−C2 bond of the keto-
sephosphonates was not occurring. The subsequent 1D NOESY
spectrum of 12a, in which the hydroxyl proton was irradiated,
showed that the H1a NOE had been lost, suggesting that the
fluorine was occupying the H1a position at C1 (Figure 4). For
12b, the stereochemistry was assumed to be the same as 12a, as
benzyl protecting groups on the phosphonate would be unlikely to
invert the position of fluorination. Such findings were consistent
with X-ray crystal structures of the analogous D-glucose
9824
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
Scheme 2. Attempted Synthesis of R Monofluoro L-Rhamnose Phosphonate Analogue from Ketosephosphonate 12a
compounds, which also showed the R stereoisomer to be the
major product.27 Nevertheless, we acknowledge that this
assignment of stereochemistry is not unequivocal and necessarily
relies on the equivalent dihedral angle between the PCCO bonds
in compounds 8a, 12a, and 12b.
inhibition studies on the mixture could be performed and qual-
itatively compared to inhibitors 9, 11, and 13. Due to the com-
plexity of the mixture of phosphonates, such additional studies
were not pursued.
For the global deprotection of the methylene and mono-
fluorinated methyl protected analogues 8a, 10a, and 12a, were
globally deprotected via hydrogenolysis using 20 mol % H2−
Pd/C in 1:1 MeOH/EtOAc overnight. This was followed by
treatment with 6 M HCl(aq) at reflux temperatures for 3 h to
cleave the phosphonate methyl esters, and susbsequent titration
to pH 8 with 0.2 M NH4OH(aq) to produce the corresponding
ammonium salts for enzymatic analysis. In the case of the
benzyl protected analogues 8b, 10b, and 12b, hydrogenolysis
was performed directly followed by titration to generate the
final compounds 9, 11, and 13. All hydrogenation and aqueous
acidic deprotection reactions produced the desired compounds
in essentially quantitative yields. In the case of the difluoro
analogues 14 and 17, both benzyl and ethyl protecting groups
were removed using trimethylsilyl iodide (TMSI) in CH2Cl2,
followed by aqueous workup and titration to afford 15 and 18 in
high yields (∼95%) as their ammonium salts. The presence of the
two fluorine atoms at C1 for the difluoro ketosephosphonate 14
seemed to stabilize the anomeric phosphonate linkage under the
harsh TMSI deprotection conditions, given that attempted TMSI
deprotection of the nonfluoro ketosephosphonate 8a resulted in
decomposition of the material. The increasing acidity of 9 < 13 <
15 was anticipated to correlate to a greater preference for the
cyclic (α-pyranose) form of the ketose. However, the ratios be-
tween the α, β, and open chain species were 18:2.5:1 (9), 100 %
α (13), and 11:3:1 (15) as determined by 31P and 19F NMR spec-
troscopy (Figure 5). This remains contrary to previous studies
on analogous gluco-ketosephosphonates in which the difluoro
analogue was found to exist solely in the α-pyranose form.
Herein, the monofluoro analogue was found to be the only
exclusive α-pyranose L-rhamnose ketosephosphonate. In the
case of the difluoro ketosephoshonate 15, 19F NMR spectros-
copy was used to determine the relative ratios of the various
tautomers, due to signal overlap in the 31P NMR spectrum. The
diastereotopic fluorine atoms appeared as AB quartets in the
α and the β pyranose forms of 15, and appeared as a sole 2F
doublet in the open chain form.
Difluoro ketosephosphonate 14 was produced from the treat-
ment of lactone 7 with diethyl difluoromethylphosphonate using
lithium diisopropylamide (LDA) at −78 °C in 74% yield. Com-
pound 14 was converted to the corresponding oxalyl ester 16
using identical reaction conditions to those used to produce
olefins 10a and 10b. The presence of the fluorine atoms at the
C1 position renders the possibility of in situ elimination of the ester
moiety unlikely, given that there are no methylene protons available
for abstraction. The oxalyl ester hydrolysis product (14) was
detected by 31P and 19F NMR spectroscopy in a minor amount
(<3%) relative to the major product (16) after purification, indi-
cating sensitivity of these esters to hydrolysis. Subsequent radical
deoxygenation using azobisisobutyl nitrile and tributyl tin hydride
in toluene, at 105 °C, afforded the difluoro phosphonate 17 in an
excellent yield (85%), with a minor side product (<5%) as detected
by 19F NMR spectroscopy. This side product is likely the axial
analogue of 17 produced from in situ mutarotation of 16 under the
reflux conditions required for radical deoxygenation.
Synthesis of the monofluorinated phosphonate analogue was
attempted from 12a using the oxalyl ester methodology em-
ployed for generation of the difluoro oxalyl ester analogue 16
(Scheme 2). Although complete formation of the oxalyl ester 19
was detected using TLC, standard aqueous workup or direct
removal of the pyridine via concentration of the crude reaction
mixture resulted in complete breakdown of the desired product
into an inseparable mixture of the hydrolysis product 12a and
olefin product 20. An alternative method of ketosephosphonate
deoxygenation27 starting from the monofluoro ketosephospho-
nate 12a and using Et3SiH and trimethylsilyl trifluoromethane-
sulfonate (TMSOTf) was explored but resulted in complete
breakdown of the starting material, as was determined via TLC
analysis. Attempts to perform hydrogenation on the mixture
of olefin products and ketosephosphonates resulted in the pro-
duction of the desired compound 21 (∼40%) as well as signif-
icant amounts of the defluorinated material 22 (∼20%) and
the monofluoro ketosephosphonate 23 (∼40%) as determined
by 31P NMR spectroscopy. Conceivably, demethylation of the re-
sultant mixture of phosphonates (21−23) and subsequent
NMR Binding Studies. Enzyme inhibitor binding experi-
ments were performed (a) to investigate the binding of
9825
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
Figure 5. 31P and 19F (inset) NMR (300 MHz, D2O, pH 7.5) spectra
for 9, 13, and 15 showing ratios of α:β:open chain form (*) of
ketosephosphonates. (A) 9 19:2.5:1. (B) 13 (100% α) verified by D
(19F inset). (C) 15 11:3:1 as determined by E (19F inset) due to 31P
NMR overlap of signals.
Figure 6. WaterLOGSY NMR spectra (700 MHz, 10:90 D2O/H2O)
showing binding of 11 and 9 to Cps2L. Asterisk indicates spectrum
showing binding (peaks above baseline). Sample compositions (see
1
General Methods). (A) H NMR spectrum of 11 and benzoic acid
phosphonate analogues (9 and 11) to enzymes (Cps2L, RmlB−D)
and to gain insight into the possible mechanism of inhibition
for these compounds, (b) to investigate the capacity for sugar-
nucleotides dTDP-D-glucose (3) and dTDP-L-rhamnose (6)
(isolated using previously established literature protocols)34 to
bind to multiple enzymes (Cps2L, RmlB−D) within the bac-
terial biosynthetic pathway, and (c) to substantiate the ordered
Bi−Bi mechanism of 3 and the allosteric binding model of 6 to
Cps2L. WaterLOGSY NMR was used to accomplish these
goals. WaterLOGSY NMR is a 1D NOE experiment in which
the irradiation of bulk water effectively facilitates magnetization
transfer from active-site bound water molecules to bound
ligands via the enzyme−ligand complex.35 Binding compounds
will maintain a perturbed magnetic state when released back
into solution from the enzyme−ligand complex and produce an
opposite sign NOE relative to nonbinding compounds. It is
noted that with WaterLOGSY, a strong (sub nanomolar)
binding interaction could result in a spectrum that appears as
a nonbinding event.35 This situation is not anticipated as our
inhibitors are expected anticipated to be in the nanomolar to
micromolar range, this situation was not anticipated. In
addition, a weak or nonspecific binding event could result in
the vanishing of peaks due to signal cancellation resulting from
the effects of bound and unbound states.36 Due to significant
signal overlap within the carbohydrate region (3−4 ppm), binding
events were most easily observed by the isolated L-rhamnose
and/or thymidine methyl signals between 1−2 ppm.
The WaterLOGSY experiments of inhibitors 9 and 11 both
showed binding to Cps2L (Figure 6). Binding was most clearly
determined by the distinct methyl peak (∼1.1 ppm) and meth-
ylene protons (1.5−2.5 ppm) phasing in the opposite direction
to the nonbinding benzoic acid signals (7.0−8.0 ppm). Com-
pound 11 also bound epimerase RmlC but did not appear to
bind to RmlB or D (Figure 7). Competitive binding experi-
ments were performed to see whether 11 would bind in the
presence of the natural substrates Glc-1-P (1) or dTTP (2). As
expected, Glc-1-P (1) did not bind to Cps2L in the absence of
dTTP. The ordered Bi−Bi mechanism (Figure 1) of Cps2L re-
quires binding of dTTP (2) to induce a conformational change
(control nonbinder). (B) 11 with Cps2L. (C) 1H NMR spectrum of 9
and benzoic acid. (D) 9 with Cps2L.
Figure 7. WaterLOGSY NMR spectra (700 MHz, 10:90 D2O/H2O) for
11. Asterisk indicates spectrum showing binding. Sample compositions
(see General Methods). (A) 1H NMR spectrum of 11 and benzoic acid.
(B) 11 with RmlB. (C) 11 with RmlC. (C) 11 with RmlD.
necessary to facilitate the binding of 1 (Figure 8). However,
analogue 11 was found to bind in the presence of substrate 1,
yet not in the presence of substrate 2. This suggests that 11
may, in fact, be binding to the same site as substrate 1, and that
the conformational change induced by 2 might hinder the ability
of 11 to bind Cps2L.
The WaterLOGSY spectra indicate binding of dTDP-Glc
(3) to Cps2L (Figure 9A). In the time it took to prepare the
9826
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
RmlD was observed (Figure 9D). The WaterLOGSY spectra of
dTDP-Rha (6) with Cps2L and RmlB-D is shown in Figure 10.
Figure 8. WaterLOGSY NMR spectra (700 MHz, 10:90 D2O/H2O)
of 11 and substrate 1 or 2 with Cps2L. Sample compositions (see
General Methods). (A) 1H NMR spectra of 1 and 11. (B) 11 binding
to Cps2L, 1 nonbinding. (C) 1H NMR spectrum of 2, 11, and benzoic
acid with Cps2L. (D) 2 binding to Cps2L, 11 nonbinding.
Figure 10. WaterLOGSY NMR spectra (700 MHz, 10:90 D2O/H2O)
showing dTDP-Rha 6 binding (above baseline) or nonbinding (below
baseline). Sample compositions (see General Methods). Asterisk
indicates spectra showing binding. (A) 6 with Cps2L. (B) 6 with
RmlB. (C) 6 with RmlC. (D) 6 with RmlD.
The product of RmlD is dTDP-Rha (6). dTDP-Rha (6) bound
Cps2L as anticipated by crystallography and kinetic studies for
homologous enzymes.14 Unequivocal binding was also observed
for 6 to RmlC. For RmlB and RmlD, we observed significant
reduction in intensity of the signals for 6 relative to the benzoic
acid control, and the signals for 6 were also difficult to phase and
remained dispersive. As indicated above, this is likely due to weak
binding of the enzymes.
Enzyme Inhibition Studies. Compounds 9 and 11 were
first evaluated as Glc-1-P (1) analogue substrates for Cps2L
given that the enzyme had previously demonstrated a broad
substrate specificity,34,37,38 including poor conversion (3−10%)
observed for β-L-fucosyl phosphate.34 Negligible turnover
(<1%) to sugar nucleotide products were observed by HPLC
over 24 h using 2−10 EU of Cps2L with a 8:1 ratio of 9 or 11
relative to dTTP substrate (2) (Supporting Information). This
suggested that the L-rhamnose phosphonates did not act as
substrates for Cps2L. Compounds 9, 11, 13, 15, and 18 were
each evaluated as inhibitors of Cps2L at various concentrations
between 0 and 100 mM against 10 mM concentrations of
substrates 1 and 2. The percentage conversion was calculated as
the concentration of product formed (3) relative to the remaining
starting material (2) for each concentration of inhibitor tested
relative to a control reaction in which no inhibitor was present.
IC50 curves were generated for compounds 9, 11, and 13
(Supporting Information). Although compounds 15 and 18 did
exhibit minor inhibition (<10%) at concentrations approaching
100 mM, IC50 curves could not be generated within the con-
centration range sampled (0−100 mM). The IC50 and Ki values
(as calculated using the Cheng−Prusoff equation39) are presented
in Table 1.
Figure 9. WaterLOGSY NMR spectra (700 MHz, 10:90 D2O/H2O)
showing dTDP-Glc 3 binding or nonbinding. Sample compositions
(see General Methods). Asterisk indicates spectra showing binding.
(A) 3 with Cps2L. (B) 3 with RmlB. (C) 3 with RmlC. (D) 3 with
RmlD. (a) Residual imidazole binding the His6-Tag on Cps2L. (b)
Benzoic acid control nonbinder.
sample, RmlB turned 3 over to 4, despite no additional NAD+,
as demonstrated by the new peak at ∼1 ppm for the methyl
substituent (Figure 9B), and loss of the C4″ and C6″ proton
signals (independently observed). These results are indicative
that both the RmlB catalyzed oxidation and dehydration oc-
curred, suggesting the enzyme contains endogenous NAD+.
Binding of 3 to RmlC was clearly observed (Figure 9C), whereas,
on the basis of the cancellation of signals, a weak binding to
It was found the L-rhamnose 1C−phosphonate analogue 11
exhibited the most potent inhibition (IC50 = 5.7 mM) relative
9827
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
pH 7.5) was prepared and used in the inhibition assays.
Concentrations between 0 and 25 mM of 24 were tested in an
attempt to determine an IC50 value for 24 against Cps2L. All
control samples were run in the absence of inhibitor 24 with
equal volume proportions of DMSO to ensure any inhibition
observed could not be attributed to solvent effects. For compound
24, no inhibition was observed up to 20 mM. At 25 mM, minor
inhibition (<10%) was observed. Due to the reduced solubility of
24 in Tris−HCl (pH 7.5) and reduced efficiency of Cps2L in the
presence of increasing concentrations of DMSO (Supporting
Information), concentrations beyond 25 mM of 24 were not
investigated. Although compound 24 may be a poor inhibitor for
Cps2L, its IC50 is significantly higher than that of phosphonate and
ketosephosphonate analogues 9, 11, and 13. We were initially
surprised that compound 24 failed to significantly inhibit Cps2L
based on the high homology between Cps2L and RmlA. Com-
pound 24 binds to the allosteric site in RmlA by interacting with
seven key residues. These are Gly115, Phe118, His119, Lys249,
Ala251, Cys252, and Glu255.9 In Cps2L, only three of these
residues are conserved (Gly115, Ala251, and Glu255), which could
result in reduced binding and, consequently, poorer inhibition of
Cps2L relative to RmlA in the presence of compound 24.
The most potent phosphonate inhibitor (11) was subjected
to a coupled spectrophotometric inhibition assay to determine
the mode of inhibition with respect to substrates 1 and 2. The
Cps2L reaction was coupled with inorganic pyrophosphatase
(IPP) to produce phosphate, which was used as a substrate
for human purine nucleoside phosphorylase (PNP) along with
7-methyl-6-thioguanosine (MESG) (25) (λmax = 330 nm) to
produce 7-methyl-6-thioguanine (26) (λmax = 355 nm) and
α-D-ribose 1-phosphate (27). A coupled assay was necessary
because Cps2L substrate 2 and product 3 have the same λmax
value (254 nm). The coupled assay is shown in Scheme 3.40
The kinetic parameters for Cps2L are displayed in Table 2. The
Km value for 1 matched closely with previously reported studies
using HPLC based assays.37 The Km value for 2 with Cps2L
has not been reported previously. Inhibitor 11 was found to
be competitive with respect to 1 (Ki = 536 μM) and uncom-
petitive with respect to 2 (Ki = 497 μM) using the coupled
assay method. The Michaelis−Menten and Lineweaver−Burk
plots can be found in the Supporting Information. Both values
were significantly higher than predicted from the Cheng−
Prusoff equation and IC50 values determined by the HPLC
method, using only Cps2L.
Table 1. Inhibition of Cps2L by L-Rhamnose Phosphonates
a
c
b
c
d
compound
IC50 (mM)
Ki (μM)
Ki (μM)
pKa2
9
23.9
5.7
295
70
358
85
6.5
7.0
5.7
5.0
5.0
11
13
15
18
24
18.3
225
274
e
>100
e
>100
e
>25
a
b
c
With respect to Glc-1-P. With respect to dTTP. Calculated using
the Cheng−Prusoff equation39 assuming competitive inhibition.
d
Estimated from previous pKa2 studies of analogous fluoro and
nonfluoro gluco-phosphonate and ketosephosphonate analogues.27
e
IC50 value beyond limit of inhibitor concentrations tested.
to ketosephosphonate 9 (IC50 = 23.9 mM) and monofluoro
ketosephosphonate 13 (IC50 = 18.3 mM). This is likely due to
the lack of the C2 axial hydroxyl moiety in the phosphonate
scaffold, which may reduce inhibitor-enzyme binding efficiency
for the ketosephosphonates. The presence and quantity of
fluorine atoms (0−2) seems to have a significant impact on
the potency of inhibition. The incorporation of fluorine into
phosphonate analogues of natural phosphates has been shown
to reduce the pKa2 of phosphonates by ∼0.5−1 units per
fluorine atom.27 The assays were conducted at pH 7.5 to ensure
complete ionization of phosphonate functionality. Monofluoro
analogue 13 produced noticeably improved inhibition relative
to its nonfluorinated analogue 9. On the basis of previous
studies of fluorinated ketosephosphonates,27 the pKa2 values
for these compounds were estimated to be ∼5.7 and ∼6.5 for
13 and 9, respectively. The ability of ketosephosphonate 9
1
to undergo mutarotation, as determined by 31P and H NMR
spectroscopy, relative to the exclusively α-pyranose analogue
13, is also a probable factor in explaining the reduced inhibition
observed for 9. The fact that difluoro analogues 15 and 18
(pKa2 ∼5.0) showed only minor inhibitory activity (<10%)
relative to the other analogues (9, 11, and 13) suggests that the
presence of an extra fluorine atom and additional pKa2 reduc-
tion is detrimental to inhibition.
Thymidylyltransferases show high sequence homology
(>60%) across various bacterial species (see sequence align-
ment, Supporting Information), with Cps2L (S. pneumoniae)
and RmlA (P. aerugenosa) sharing 68% identity. Commercially
available thymine analogue 24 (Figure 11) was shown to be a
On the basis of the coupled inhibition assay results, it is
expected that inhibitor 11 binds to the same site as 1 both prior
to (competitive model) and after (uncompetitive model) the
conformational change induced by the binding of 2. Although
binding in the presence of 2 was not observed by WaterLOGSY
NMR spectroscopy, it could be rationalized by Cps2L having
higher affinity for 2 (Km,app = 152 μM) relative to 11 (Ki =
497 μM). Past15 and recent9 studies of nucleotidylyltransferase
inhibition have all focused on the study of allosteric inhibitor
compounds with a common trend being that each of the allosteric
binders contained a thymine derived moiety necessary for binding.
The scaffold of inhibitor 11 is more similar to substrate 1, with no
thymine component. To our knowledge, it is the first reported
sugar−phosphonate active site binding competitive inhibitor of
Cps2L. The synthesis of the sugar nucleotide analogues of 9, 11,
and 13 would produce analogues of dTDP-Rha (6). These ana-
logues would be predicted to be more potent inhibitors than their
phosphonate analogues and, with the introduction of the thymidine
Figure 11. Structure of RmlA (P. aeruginosa) inhibitor (24).
potent inhibitor of RmlA (P. aeruginosa). Naismith and co-
workers report an IC50 value of 0.32 μM for allosteric com-
petitive inhibitor compound 24.9 We chose this compound as a
benchmark for our synthesized compounds.
We determined, using the Cheng−Prusoff equation, a Ki
value of ∼0.2 μM for both Glc-1-P and dTTP relative to RmlA
for compound 24. With RmlA and Cps2L being 68% identical,
it was expected that comparable inhibition might be observed
for Cps2L. Using the predicted Ki for RmlA, an IC50 value be-
tween 10−20 μM was predicted for our HPLC based Cps2L in-
hibition assay. A solution of 24 (50 mM in 1:1 DMSO/Tris−HCl,
9828
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
Scheme 3. Cps2L, IPP, and Human PNP Coupled Assay
(1.1%) followed by charring. Either 300 or 500 MHz NMR spec-
trometers were used for all structural NMR data. Peak assignments
Table 2. Apparent Substrate Kinetics and Inhibition
Parameters for Cps2L in the Presence of 11
were aided by H−1H correlated spectroscopy (COSY) and H−13C
heteronuclear single quantum coherence (HSQC) experiments when
necessary. 1D NOESY was used to determine stereochemistry for
select compounds. A 700 MHz spectrometer equipped with a 1.7 mm
CryoProbe was utilized for all WaterLOGSY binding experiments.
HPLC analysis for all enzymatic inhibition assays was performed as
previously described.34 Compounds containing a nucleotide base chro-
mophore were monitored at 254 nm absorbance. HPLC runs were
performed for 15 min at a flow rate 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 to 10 min and a linear decline to 90:10 A/B from 10 to
11 min followed by isocratic 90:10 A/B was used for all assays. Buffer
A was aqueous 12 mM Bu4NBr, 10 mM KH2PO4, and 5% HPLC grade
CH3CN. Buffer B was HPLC grade CH3CN. Low resolution mass spectra
were obtained with a hybrid triple quadrupole linear ion trap ionization
(ESI) source. Samples were run using a flow rate of 120 μL/min and an
isocratic 70:30 CH3CN/2 mM aqueous ammonium acetate (pH 5.5)
buffer. The capillary voltage was set to 4500 kV, with a declustering
potential of 60 V and a curtain gas of 20 (arbitrary units) for positive and
negative mode, respectively. Analyst version 1.4.1 (Applied Biosystem)
software was used for analysis.
WaterLOGSY NMR Analysis. WaterLOGSY NMR samples were
composed of binding substrates Glc-1-P (5 mM) and/or dTTP
(5 mM) with MgCl2 (5 mM) co-factor or sugar nucleotides (5 mM)
with MgCl2 (5 mM) and enzymes Cps2L, RmlB, RmlC, or RmlD
(0.05 mM). Benzoic acid (5 mM) was used as a control nonbinder for
all experiments. For experiments using phosphonate inhibitors, 5 mM
concentrations were used. All samples were prepared in Tris−HCl
(100 mM, pH 7.4) buffer with 10% D2O to a total volume of 50 μL.
Benzoic acid was used as a control nonbinder with the exception of the
competitive model of Glc-1-P (1) and Rha-1C-P (11) (Figure 8). In
each spectrum, proton signals for the Tris−HCl buffer (∼3.5 ppm)
appear as the major signal in the 1H NMR spectra. In spectra containing
Cps2L, residual signals of imidazole left over from purification (∼7.1
and 7.8 ppm) could not be removed and should not be confused with
thymidine or benzoic acid aromatic proton signals.
HPLC Enzyme Assay and Inhibition Conditions. Cps2L,
RmlB-D were overexpressed and purified as previously described.34,41
The concentration of Cps2L was measured spectrophotometrically at
280 nm using a calculated extinction coefficient of 29.8 mM−1 cm−1.
All stock solutions were prepared in Tris−HCl buffer (100 mM, pH 7.4).
Inhibition assays were conducted using phosphonate analogues
(0−100 mM), or compound 24 (0−25 mM) with Glc-1-P (10 mM),
dTTP (10 mM), MgCl2 (20 mM), and Cps2L (8 EU) in a final volume
of 50 μL of Tris−HCl buffer. Samples were incubated for 17 min at
37 °C, at which point 25 μL aliquots were quenched in an equal volume
of HPLC grade MeOH. Conversions were approximately 30% after
17 min. In the case of compound 24, a concentrated stock solution
(50 mM in 1:1 DMSO/Tris−HCl, pH 7.5) was used. IC50 curves and
values were generated based on HPLC conversion using Grafit 5 Eri-
thacus software. Ki values were approximated relative to Glc-1-P and
dTTP substrates using the Cheng−Prusoff equation, which assumes a
competitive mode of inhibition. Due to the time intensive requirements
of performing HPLC-based inhibition assays, we did not carry out
experiments in duplicate.
1
1
compound
Km,app (μM)
Ki (μM)
kcat. (s‑1)
kcat./Km (s‑1 μM‑1)
a,
a,
b,
c
c
c
1
125
139
152
60
0.475
d
c
c
c
2
54
0.352
a,e
11
536
497
b,f
a
b
With respect to variable Glc-1-P. With respect to variable dTTP.
c
d
Calculated from coupled spectrophotometric assay. Previously
reported.37 Competitive inhibition. Uncompetitive inhibition.
e
f
component into their scaffolds, would potentially function as
allosteric competitive inhibitors.
CONCLUSION
■
We have reported the synthesis and enzymatic evaluation of a
series of L-rhamnose 1C-phosphonate and ketosephosphonate
analogues and investigated the impact of fluorination on the
inhibition of nucleotidylyltransferase Cps2L. WaterLOGSY
studies with compounds 9 and 11 demonstrated the potential
of synthetic substrate analogues to bind to multiple enzymes
within a single biosynthetic pathway. WaterLOGSY studies of
dTDP-Glc (3) and dTDP-Rha (6) also highlighted the capacity
for sugar nucleotides to bind to, and potentially inhibit, mul-
tiple enzymes within a single bacterial biosynthetic pathway.
The synthesized compounds bound multiple enzymes in the
L-rhamnose pathway and were competitive inhibitors for the
first enzyme, Cps2L. L-Rhamnose-1C-phosphonate (11) was
found to be a better inhibitor than all of the ketosephospho-
nates, presumably due to reduced binding efficiency from the
presence of the C2 hydroxyl group on the ketosephosphonate
scaffold. For the series of ketosephosphonates, the presence of
one fluorine atom at the C1 position improved inhibition by
∼25% relative to the nonfluorinated phosphonate whereas the
presence of two fluorine atoms was detrimental to inhibition.
We expect the sugar nucleotide analogues of these L-rhamnose
phosphonate and ketosephosphonate analogues will signifi-
cantly magnify their capacity to inhibit Cps2L and potentially
other enzymes within the prokaryote L-rhamnose biosynthetic
pathway. Such studies are currently in progress and results will
be reported in due course.
EXPERIMENTAL SECTION
■
General Methods. All reagents and solvents were purchased and
used without further purification. Synthetic reactions were performed
under N2 atmosphere unless otherwise indicated. Dibenzyl methyl-
phosphonate was prepared and characterized as previously described.31
Reaction progress was generally monitored by thin layer chromatog-
raphy. Plates were visualized using an ethanol dipping solution
containing p-anisaldehyde (3.4%), sulfuric acid (2.2%), and acetic acid
9829
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
Spectrophotometric Enzyme Coupled Kinetic and Inhibition
Assay Conditions. Recombinant human PNP was overexpressed
and purified using the same procedure as described for Cps2L and
RmlB-D, with the following modifications; overnight induction at
37 °C was initiated with 0.1% w/v lactose.42 The concentration of
PNP was measured spectrophotometrically at 280 nm using an extinc-
tion coefficient of 29.8 mM−1 cm−1. Inorganic pyrophosphatase (IPP)
from Escherichia coli was purchased from a supplier. Stock solutions
of 0.1 EU/μL IPP were prepared in Millipore water and stored at
−30 °C. Thawed aliquots were kept in a fridge and were used for up to
1 month after thawing. MESG was purchased from a supplier. MESG
stock solutions were prepared in Millipore water no more than 2
weeks prior to use and aliquots were stored at −30 °C. MESG so-
lutions were used as soon as possible after thawing. All other stock
solutions were prepared in Tris−HCl buffer (50 mM, pH 7.5).
Enzyme kinetic assays were performed for each substrate, dTTP and
Glc-1-P. Reactions containing 1 mM Glc-1-P or dTTP, 0−300 μM
dTTP or Glc-1-P, 8 μM Human PNP, 0.05 EU IPP, 5.6 mM MgCl2,
and 400 μM MESG in 50 mM Tris−HCl (pH 7.5) were initiated by
the addition of 4.9 nM Cps2L. Assays were made to a final reaction
volume of 80 μL and performed in a 384 well plate. Initial velocities
were monitored continuously by UV spectrometry at λ = 360 nm
using Softmax pro 4.8. Inhibition assays were performed using the
same reaction conditions with variable concentrations (250−1000 μM)
of Rha-1C−P. Initial velocities were converted from mAu/min to
μM/min using a 0−100 μM Pi standard curve. Michaelis−Menten and
inhibition equations were fit using Grafit 5 Erithacus Software.
Rf = 0.22 (20:80 EtOAc/hexane); δH 7.25−7.45 (m, 25H, 5 × C6H5),
5.80 (s, 1H, OH), 4.60−5.24 (m, 10H, 5 × CH2Ph), 4.23 (dd, 1H, H4,
3J4,5 = 9.3 Hz, 3J4,3 = 2.6 Hz), 4.10 (m, 1H, H6), 3.76 (d, 1H, H3 3J3,4
=
3
3
2.6 Hz), 3.67 (t, 1H, H5, J5,4 = J5,6 = 9.5 Hz), 2.70 (dd, 1H, H1a,
2
2
2JH,P = 18.8 Hz, JH,H = 15.4 Hz), 1.78 (dd, 1H, H1b, JH,P = 18.8 Hz,
2JH,H = 15.7 Hz), 1.37 (d, 3H, H7, 3J7,6 = 6.3 Hz); δC 135.8−138.8 (5 ×
2
C, C6H5), 127.5−129.2 (25 × CH, C6H5), 97.3 (d, C, C2, JC,P
=
=
3
7.72 Hz), 81.4 (CH, C4), 80.4 (CH, C5), 78.5 (d, CH, C3, JC,P
12.77 Hz), 75.4 (CH2, CH2Ph), 74.9 (CH2, CH2Ph), 73.1 (CH2,
CH2Ph), 69.0 (CH, C6), 68.5 (d, CH2, POCH2Ph, 2JC,P = 5.5 Hz), 66.9
(d, CH2, POCH2Ph, 2JC,P = 5.5 Hz), 34.3 (d, CH2, C1, 1JC,P = 136.40 Hz),
18.0 (CH3, C7); δP 30.49 (s, 1P). HRMS (ESI−): found [M − H]−
707.2762. C42H44O8P1 requires [M − H]− 707.2779.
Dimethyl (2,6-Anhydro-3,4,5-tri-O-benzyl-1-deoxy-L-rhamno-
hept-1-enopyranosyl)phosphonate (10a). Dimethyl (tri-O-benzyl-
1-deoxy-L-rhamno-heptulopyranosyl)phosphonate (8a) (0.660 g, 1.18
mmol) was dissolved in a mixture of 7:2 CH2Cl2:pyridine (9 mL), and
methyl oxalyl chloride (MeOXCl) (1.13 mL, 10.58 mmol) was added
to the mixture dropwise with stirring under nitrogen. After 90 min,
TLC showed complete consumption of 8a so the reaction was
quenched with EtOH (1 mL) and was allowed to stir for 10 min
before CH2Cl2 (13 mL) and NaHCO3 (5 mL, sat.) were added. The
reaction was stirred for an additional 5 min after which the mixture
was poured into NaHCO3 (10 mL) and extracted with CH2Cl2 (3 ×
15 mL) before being dried (Na2SO4) and concentrated. The resulting
material was purified using column chromatography (70:30−100:0
EtOAc/hexane) to afford the product (10a) as a clear viscous liquid
(0.433 g, 68% yield): Rf = 0.46 (80:20 EtOAc/hexane); δH (CD2Cl2)
Synthetic Procedures and Characterizations. Dimethyl (3,4,5-
Tri-O-benzyl-1-deoxy-α-L-rhamno-heptulopyranosyl)phosphonate
(8a). Dimethyl methylphosphonate (0.93 mL, 8.31 mmol) was
dissolved in anhydrous THF (5 mL) under nitrogen, and the mixture
was cooled to −78 °C using a dry ice/acetone slurry. n-Butyllithium
(2.5 M in hexane) (0.35 mL, 8.31 mmol) was added, and the solution
was stirred for 30 min before 2,3,4,-tri-O-benzyl-L-rhamno-1,5-lactone
(7) (1.197 g, 2.8 mmol) was added dropwise in a solution of THF
(5 mL). After 75 min, TLC showed complete consumption of 7. The
reaction was quenched with 10% NH4Cl (10 mL), and the THF layer
was extracted from the mixture. The aqueous layer was then extracted
from the mixture with CH2Cl2 (3 × 10 mL). The organic extracts were
combined, dried (Na2SO4), and concentrated. The result material was
purified using column chromatography (40:60 EtOAc/hexane) to
afford the product (8a) as a clear liquid (1.378 g, 88% yield): Rf = 0.25
(40:60 EtOAc/hexane); δH (CDCl3) 7.20−7.50 (m, 15H, 3 × C6H5),
5.74 (s, 1H, OH), 4.50−5.10 (m, 6H, 3 × CH2Ph), 4.12 (dd, 1H, H4
2
7.20−7.50 (m, 15H, 3 × C6H5), 5.02 (d, 1H, H1, JH,P = 12.4 Hz),
4.40−4.80 (m, 6H, 3 × CH2Ph), 4.19 (d, 1H, H3 3J3,4 = 2.8 Hz), 3.89
3
3
(m, 1H, H6), 3.80 (dd, 1H, H4, J4,5 = 5.6 Hz, J4,3 = 2.8 Hz), 3.75
3
5
(dd, 6H, 2 × OCH3, JH,P = 11.4 Hz, JH,H = 1.1 Hz), 3.70 (dd, 1H,
H5, 3J5,6 = 8.7 Hz, 3J5,4 = 5.6 Hz), 1.46 (d, 3H, H7, 3J7,6 = 6.2 Hz); δC
165.75 (C, C2), 137.3−138.0 (3 × C, C6H5), 127.5−129.0 (15 × CH,
C6H5), 92.8 (d, CH, C1, 1JC,P = 191.4 Hz), 80.5 (CH, C5), 78.1 (CH,
3
C4), 75.9 (CH, C6), 75.2 (d, CH, C3, JC,P = 14.9 Hz), 73.8, (CH2,
CH2Ph), 72.4 (CH2, CH2Ph), 71.5 (CH2, CH2Ph), 52.7 (d, CH3,
2
2
OCH3, JC,P = 5.5 Hz), 52.3 (d, CH3, OCH3, JC,P = 5.5 Hz), 19.1
(CH3, C7); δP 20.17 (s, 1P). HRMS (ESI+): found [M + Na]+
561.2009. C30H35O7P1 requires [M + Na]+ 561.2013.
Dibenzyl (2,6-Anhydro-3,4,5-tri-O-benzyl-1-deoxy-L-rhamno-
hept-1-enopyranosyl)phosphonate (10b). Dibenzyl (tri-O-benzyl-1-
deoxy-L-rhamno-heptulopyranosyl)phosphonate (8b) (0.318 g, 0.45
mmol) was dissolved in a mixture of 7:2 CH2Cl2:pyridine (7 mL), and
methyl oxalyl chloride (MeOXCl) (0.4 mL, 4.18 mmol) was added to
the mixture dropwise with stirring under nitrogen. After 90 min, TLC
analysis showed complete consumption of the starting material so the
reaction was quenched with EtOH (0.5 mL) and was allowed to stir
for 10 min before CH2Cl2 (13 mL) and NaHCO3 (5 mL, sat.) were
added. The reaction mixture was stirred for an additional 5 min after
which it was poured into NaHCO3 (10 mL, sat.) and extracted with
CH2Cl2 (3 × 15 mL) before being dried (Na2SO4) and concentrated.
The result material was purified using column chromatography (35:65
EtOAc/hexane) to afford the product as a clear viscous liquid (10b)
(0.188 g, 60% yield): Rf = 0.44 (50:50 EtOAc/hexane); δH (CD2Cl2)
3
3J4,5 = 9.4 Hz, J4,3 = 2.7 Hz), 4.03 (m, 1H, H6), 3.83 (d, 3H, OCH3,
3
3JH,P = 11.1 Hz), 3.76 (d, 1H, H3 J3,4 = 2.7 Hz), 3.69 (d, 3H, OCH3,
3
3JH,P = 11.1 Hz), 3.64 (t, 1H, H5, 3J5,4 = J5,6 = 9.4 Hz), 2.60 (dd, 1H,
H1a, 2JH,P = 17.9 Hz, 2JH,H = 16.0 Hz), 1.69 (dd, 1H, H1b, 2JH,P = 17.9
Hz, 2JH,H = 16.0 Hz) 1.32 (d, 3H, H7, 3J7,6 = 6.4 Hz); δC 138.3−138.8
(3 × C, C6H5), 127.5−129.0 (15 × CH, C6H5), 97.1 (d, C, C2, 2JC,P
=
=
3
7.6 Hz), 81.5 (CH, C4), 80.3 (CH, C5), 78.2 (d, CH, C3, JC,P
12.5 Hz), 75.4 (CH2, CH2Ph), 74.8 (CH2, CH2Ph), 73.2 (CH2,
CH2Ph), 69.0 (CH, C6), 53.9 (d, CH3, OCH3, 2JC,P = 5.8 Hz), 51.8 (d,
CH3, OCH3, 2JC,P = 5.8 Hz), 33.2 (d, CH2, C1, 1JC,P = 135.12 Hz), 18.0
(CH3, C7); δP 32.22 (s, 1P). HRMS (ESI−): found [M − H]−
555.2164. C30H36O8P1 requires [M − H]− 555.2153.
2
7.00−7.20 (m, 25H, 5 × C6H5), 4.86 (d, 1H, H1, JH,P = 12.7 Hz),
Dibenzyl (3,4,5-Tri-O-benzyl-1-deoxy-α-L -rhamno-
heptulopyranosyl)phosphonate (8b). Dibenzyl methylphosphonate
(0.374 g, 1.35 mmol) was dissolved in anhydrous THF (5 mL) under
nitrogen, and the mixture was cooled to −78 °C using a dry ice/acetone
slurry. n-Butyllithium (0.5 mL, 1.25 mmol) was added, and the solution
was stirred for 30 min before 2,3,4,-tri-O-benzyl-L-rhamno-1,5-lactone
(7) (0.220 g, 0.5 mmol) was added dropwise in a solution of THF
(1 mL). After 75 min, TLC showed complete consumption of (7). The
reaction was quenched with 10% NH4Cl (10 mL), and the THF layer
was extracted from the mixture. The aqueous layer was then extracted
from the mixture with CH2Cl2 (3 × 10 mL). The organic extracts were
combined, dried (Na2SO4), and concentrated. The resulting material
was purified using column chromatography (25:75 EtOAc/hexane) to
afford the product as a clear viscous liquid (8b) (0.270 g, 76% yield):
4.20−4.80 (m, 10H, 5 × CH2Ph), 4.01 (d, 1H, H3 3J3,4 = 1.8 Hz), 3.58
(m, 2H, H4, H6), 3.37 (dd, 1H, H5, 3J5,6 = 8.5 Hz, 3J5,4 = 4.6 Hz), 1.11
3
(d, 3H, H7, J7,6 = 6.4 Hz); δC 165.6 (C, C2), 137.5−138.7 (5 × C,
1
C6H5), 128.0−129.2 (25 × CH, C6H5), 92.2 (d, CH, C1, JC,P
=
=
3
190.9 Hz), 81.2 (CH, C5), 78.2 (CH, C4), 78.0 (d, CH, C3, JC,P
14.5 Hz), 75.8 (CH, C6), 73.7 (CH2, CH2Ph), 72.8 (CH2, CH2Ph),
72.3 (CH2, CH2Ph), 67.6 (d, CH2, POCH2Ph, 2JC,P = 5.0 Hz), 67.3 (d,
2
CH2, POCH2Ph, JC,P = 5.0 Hz), 19.2 (CH3, C7); δP 17.78 (s, 1P).
HRMS (ESI+): found [M + Na]+ 713.2606. C42H43O7P1 requires [M +
Na]+ 713.2639.
Dimethyl (3,4,5-Tri-O-benzyl-1-deoxy-1-R-fluoro-α-L-rhamno-
heptulopyranosyl)phosphonate (12a). Dimethyl (2,6-anhydro-
tri-O-benzyl-1-deoxy-L-rhamno-hept-1-enopyranosyl)phosphonate
(10a) (0.250 g, 0.464 mmol) was dissolved in acetonitrile (5 mL), and
9830
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
3
Selectfluor (0.350 g, 0.99 mmol) was added to the mixture all at once
with stirring overnight at room temperature under nitrogen. Water
(2 mL) was then added, and the solution was heated at 70 °C for 2.5 h
before being poured into water (10 mL) and extracted with CH2Cl2
(2 × 10 mL). The organic layer was dried (Na2SO4), concentrated,
and purified using column chromatography (25:75 EtOAc/hexane) to
afford the single stereoisomer (12a) (0.141 g, 54% yield): Rf = 0.24
(30:70 EtOAc/hexane). By NOE analysis, the CHFP stereocenter was
determined to be R configuration. δH: (CDCl3) 7.20−7.50 (m, 15H, 3 ×
C6H5), 5.57 (s, 1H, OH), 4.50−5.10 (m, 7H, 3 × CH2Ph, H1), 4.15
(dd, 1H, H4 3J4,5 = 9.1 Hz, 3J4,3 = 2.7 Hz), 4.09 (m, 2H, H3, H6), 3.91
(d, 3H, OCH3, 3JH,P = 11.1 Hz), 3.84 (d, 3H, OCH3, 3JH,P = 11.1 Hz),
3.68 (t, 1H, H5, 3J5,4 = 3J5,6 = 9.1 Hz), 1.31 (d, 3H, H7, 3J7,6 = 6.2 Hz).
δC: 138.4−138.8 (3 × C, C6H5), 127.5−129.0 (15 × CH, C6H5), 97.9
(CH, C5), 75.6 (d, CH, C3, JC,P = 5.5 Hz), 75.4 (CH2, CH2Ph), 74.9
(CH2, CH2Ph), 72.5 (CH2, CH2Ph), 69.4 (CH, C6), 65.6 (d, CH2,
OCH2, JC,P = 6.2 Hz), 17.7 (CH3, C7), 16.5 (d, CH2CH3, JC,P = 6.1
2
3
2
Hz); δP 8.17 (t, 1P, JP, F = 96.9 Hz); δF −117.36 to −119.00 (dd, 1F,
2JF, F = 307.0 Hz, 2JF, P = 96.9 Hz), −119.12 to −120.70 (dd, 1F, 2JF, F
=
307.0 Hz, 2JF, P = 96.9 Hz). HRMS (ESI−): found [M − H]− 619.2296.
C32H38F2O8P1 requires [M − H]− 619.2278.
Diethyl (3,4,5-Tri-O-benzyl-1-deoxy-1,1-difluoro-2-O-methyloxal-
yl-α-L-rhamno-heptulopyranosyl)phosphonate (16). Diethyl (tri-O-
benzyl-1-deoxy-L-rhamno-heptulopyranosyl)-1-difluorophosphonate
(14) (0.149 g, 0.24 mmol) was dissolved in a mixture of 7:2
CH2Cl2:pyridine (5 mL), and methyl oxalyl chloride (MeOXCl) (0.2
mL, 2.1 mmol) was added to the mixture dropwise with stirring under
nitrogen. After 90 min, TLC analysis showed complete con-
sumption of 14 so the reaction was quenched with EtOH (0.5 mL)
and was allowed to stir for 10 min before CH2Cl2 (13 mL) and
NaHCO3 (5 mL, sat.) were added. The mixture was stirred for an
additional 5 min after which it was poured into NaHCO3 (10 mL, sat.)
and extracted with CH2Cl2 (3 × 15 mL) before being dried (Na2SO4)
and concentrated. The result material was purified using column
chromatography (30:70 EtOAc/hexane) to afford the product as a clear
viscous liquid (16) (0.101 g, 69% yield): Rf = 0.64 (50:50 EtOAc/
hexane); δH (CDCl3) 7.28−7.48 (m, 15H, 3 × C6H5), 5.37 (br s, OH,
H3), 4.62−5.00 (m, 6H, 3 × CH2Ph), 4.21−4.43 (m, 4H, 2 ×
OCH2CH3), 4.07 (m, 1H, H6), 3.94 (s, 3H, OCH3), 3.86 (t, 1H, H5,
2
(d, C, C2, JC,F = 24.7 Hz), 85.8 (m, CH, C1), 81.2 (CH, C4), 80.1
3
3
(CH, C5), 75.4 (dd, CH, C3, JC,F = 19.5 Hz, JC,P = 13.7 Hz), 72.8
(CH2, CH2Ph), 69.4 (CH, C6), 54.4 (d, CH3, OCH3, 2JC,P = 5.9 Hz),
54.1 (d, CH3, OCH3, 2JC,P = 5.9 Hz), 17.9 (CH3, C7). δP: 20.60 (d, 1P,
2
2JP,F = 74.8 Hz). δF: −216.8 (d, 1F, JF,P = 75.1 Hz). HRMS (ESI−):
found [M − H]− 573.2045. C30H35F1O8P1 requires [M − 1]−
573.2059.
Dibenzyl (3,4,5-Tri-O-benzyl-1-deoxy-1-R-fluoro-α-L-rhamno-
heptulopyranosyl)phosphonate (12b). Dibenzyl (2,6-anhydro-tri-O-
benzyl-1-deoxy-L-rhamno-hept-1-enopyranosyl)phosphonate (10b)
(0.142 g, 0.206 mmol) was dissolved with acetonitrile (3 mL), and
Selectfluor (0.157 g, 0.44 mmol) was added to the mixture all at once
with stirring overnight at room temperature under nitrogen. Water
(1 mL) was then added, and the solution was heated at 70 °C for 2.5 h
before being poured into water (10 mL) and extracted with CH2Cl2
(2 × 10 mL). The organic layer was dried (Na2SO4), concentrated,
and purified using column chromatography (16:84 EtOAc/hexane) to
afford the R (12b) as a clear viscous liquid (0.033 g, 22% yield): Rf =
0.45 (20:80 EtOAc/hexane); δH (CDCl3) 7.29−7.45 (m, 25H, 5 ×
C6H5), 5.66 (s, 1H, OH), 4.63−5.25 (m, 11H, 5 × CH2Ph, H1), 4.17
3J5,4 = J5,6 = 9.4 Hz), 3.72 (dd, 1H, H4 J4,5 = 9.4 Hz, J4,3 = 2.4 Hz),
1.43 (d, 3H, H7, 3J7,6 = 6.2 Hz), 1.47 (m, 6H, 2 × OCH2CH3); δC 157.1
(CO, C(O)C(O)OCH3), 154.5 (CO, C(O)C(O)OCH3), 137.6−138.6
(3 × C, C6H5), 127.5−128.8 (15 × CH, C6H5), 106.3 (m, CF2, C1)
80.0 (CH, C4), 78.9 (CH, C5), 75.8 (CH2, CH2Ph), 75.3 (CH2,
CH2Ph), 73.0 (CH, C3), 72.3 (CH2, CH2Ph), 72.1 (CH, H6), 65.5 (d,
CH2, OCH2, 2JC,P = 6.3 Hz), 65.2 (d, CH2, OCH2, 2JC,P = 6.3 Hz), 54.0
(CH3, OCH3), 17.9 (CH3, C7), 16.5 (m, CH2CH3); δP 3.82 (dd, 1P,
3
3
3
2JP,F = 104.9 Hz, 2JP,F = 91.3 Hz); δF −109.6 to −111.22 (dd, 1F, 2JF,F
=
3
3
320.6 Hz, 2JF,P = 91.3 Hz), −113.60 to −115.25 (dd, 1F, 2JF,F = 320.6 Hz,
2JF,P = 104.9 Hz). HRMS (ESI+): found [M + Na]+ 729.2258.
(dd, 1H, H4 J4,5 = 9.6 Hz, J4,3 = 2.7 Hz), 4.08−4.15 (m, 2H, H3,
3
3
3
H6), 3.69 (t, 1H, H5, J5,4 = J5,6 = 9.6 Hz), 1.30 (d, 3H, H7, J7,6
=
C35H41F2O11P1 requires [M + Na]+ 729.2247.
6.3 Hz); δC 135.0−139.0 (5 × C, C6H5), 127.6−129.1 (25 × CH,
C6H5), 98.1 (d, C, C2, 2JC,F = 25.2 Hz), 85.1 (m, CH, C1), 81.3 (CH,
Diethyl (2,6-Anhydro-3,4,5-tri-O-benzyl-1-deoxy-1,1-difluoro-β-L-
rhamno-heptulopyranosyl)phosphonate (17). Diethyl (tri-O-benzyl-
2-deoxy-methyloxalyl-L-rhamno-heptulopyranosyl)-1-difluorophospho-
nate (16) (0.100 g, 0.140 mmol) was dissolved in toluene (7 mL) and
transferred to a three-neck round-bottom flask that had been degassed
for 10 min. Tributyltin hydride (Bu3SnH) (87 μL, 0.33 mmol) was
added followed by dropwise addition of azobisisobutylnitrile (AIBN)
(20 mol %, 0.14 mL, 0.028 mmol) over 1 min. The reaction was heated
to 105 °C and stirred for 1 h. The reaction mixture was then cooled to
room temperature and diluted with EtOAc (7 mL) followed by
concentration. The resulting material was purified using column
chromatography (27:73 EtOAc/hexane) to afford the product as a
clear viscous liquid (27) (0.076 g, 85% yield): Rf = 0.32 (40:60 EtOAc/
hexane); δH (CDCl3) 7.25−7.50 (m, 15H, 3 × C6H5), 4.60−5.00 (m,
6H, 3 × CH2Ph), 4.20−4.40 (m, 5H, H3, 2 × OCH2CH3), 3.91 (m, 1H,
H2), 3.75 (t, 1H, H5, 3J5,4 = 3J5,6 = 9.4 Hz), 3.54 (dd, 1H, H4 3J4,5 = 9.5
3
3
C4), 80.2 (CH, C5), 75.4 (dd, CH, C3, JC,F = 25.8 Hz, JC,P = 12.8
Hz), 72.8 (CH2, CH2Ph), 69.4 (CH, C6), 69.2 (d, CH2, POCH2Ph,
2
2JC,P = 6.4 Hz), 69.0 (d, CH2, POCH2Ph, JC,P = 6.4 Hz), 18.0 (CH3,
C7); δP 18.92 (d, 1P, 2JP,F = 76.5 Hz); δF −215.70 (d, 1F, 2JF,P = 76.2
Hz). HRMS (ESI−): found [M − 1]− 725.2668. C42H43F1O8P1
requires [M − 1]− 725.2685.
Diethyl (3,4,5-Tri-O-benzyl-1-deoxy-1,1-difluoro-α-L-rhamno-
heptulopyranosyl)phosphonate (14). Diisopropylamine (0.49 mL,
2.9 mmol) was dissolved in anhydrous THF (3 mL), and the mixture
was cooled to −78 °C using a dry ice/acetone slurry.
n-Butyllithium (1.2 mL, 2.9 mmol) was added, and the solution was
stirred for 5 min before being transferred to an ice bath and allowed to
warm to 0 °C over the course of 30 min. Diethyl difluoromethyl-
phosphonate (0.5 mL, 3.2 mmol) was cooled to −78 °C in THF
(2 mL) and added dropwise to the solution of LDA at −78 °C. After
stirring for 15 min, 2,3,4,-tri-O-benzyl-L-rhamno-1,5-lactone (7) (0.40 g,
0.93 mmol) in THF (3 mL) was added dropwise to the solution. TLC
analysis showed complete consumption of the lactone after 30 min. The
reaction was quenched with 10% NH4Cl (15 mL) followed by addition
of diethyl ether. The aqueous layer was removed and extracted with
diethyl ether (3 × 15 mL). The organic extracts were combined, dried
(Na2SO4), and concentrated. The resulting material was purified using
column chromatography (24:76 EtOAc/hexane) to afford the product
as a clear viscous liquid (14) (0.423 g, 74% yield): Rf = 0.37 (25:75
EtOAc/hexane); δH (CDCl3) 7.20−7.60 (m, 15H, 3 × C6H5), 5.91 (s,
1H, OH), 4.70−5.05 (m, 6H, 3 × CH2Ph), 4.33−4.45 (m, 4H, 2 ×
3
Hz, J4,3 = 2.7 Hz), 3.50 (m, 1H, H6), 1.30−1.45 (m, 9H, H7, 2 ×
OCH2CH3); δC 138.2−138.6 (3 × C, C6H5), 127.6−128.8
(15 × CH, C6H5), 84.0 (CH, C4), 80.0 (CH, C5), 77.2 (CH, C6),
75.7 (CH, CH2, C2, CH2Ph), 74.6 (CH2, CH2Ph), 72.7 (CH, C3, 3JC,P
=
2
5.9 Hz), 72.4 (CH2, CH2Ph), 64.9 (t, CH2, OCH2, JC,P = 7.4 Hz),
3
17.7 (CH3, C7), 16.6 (d, CH2CH3, JC,P = 4.9 Hz); δP 6.20 (t, 1P,
2JP,F = 108.0 Hz); δF −115.27 to −117.01 (dd, 1F, JF, F = 314.7 Hz,
2
2JF,P = 102.3 Hz), −123.62 to −125.34 (dd, 1F, 2JF,F = 314.7 Hz, 2JF,P
=
102.3 Hz). HRMS (ESI+): found [M + Na]+ 627.2320. C32H39F2Na1O7P1
requires [M + Na]+ 627.2294.
3
3
OCH2CH3), 4.27 (d, 1H, H3, J3,4 = 2.8 Hz), 4.16 (dd, 1H, H4 J4,5
=
=
Representative Procedure for the Deprotection of Methyl
Protected Analogues. A solution of methyl protected compound 8a,
10a, or 12a (50 mg), 10% palladium on carbon (20 mg, 0.02 mmol),
ethyl acetate (3 mL), and methanol (3 mL) was degassed under
vacuum and saturated with hydrogen gas (1 atm balloon). The reaction
was stirred overnight before the catalyst was filtered, and the solution
9.6 Hz, 3J4,3 = 2.8 Hz), 4.10 (m, 1H, H6), 3.79 (t, 1H, H5, 3J5,4 = 3J5,6
9.6 Hz), 1.47 (td, 3H, OCH2CH3, 3JH,H = 7.0 Hz, 4JH,P = 0.6 Hz), 1.43
3
4
(td, 6H, 2 × OCH2CH3, JH,H = 7.0 Hz, JH,P = 0.6 Hz), 1.39 (d, 3H,
H7, 3J7,6 = 6.3 Hz); δC 138.5−138.7 (3 × C, C6H5), 127.4−128.8 (15 ×
CH, C6H5), 96.4 (m, C, C2), 89.8 (m, CF2, C1), 81.1 (CH, C4), 79.7
9831
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
was concentrated. The material was directly dissolved in 6 M aqueous
HCl (5 mL) and refluxed for 3 h before being concentrated with
methanol (3 × 5 mL) and titrated to pH 8 with aqueous NH4OH
(0.2 M). Lyophilization provided compounds 9, 11, and 13 were ob-
tained in quantitative yields as their diammonium salts.
Representative Procedure for the Deprotection of Benzyl
Protected Analogues. A solution of benzyl protected compound 8b,
10b, or 12b (50 mg), 10% palladium on carbon (20 mg, 0.02 mmol),
ethyl acetate (3 mL), and methanol (3 mL) was degassed under
vacuum and saturated with hydrogen gas (1 atm balloon). The reaction
was stirred overnight before the catalyst was filtered, and the solution
was concentrated to a colorless oil. The material was dissolved in water
(2 mL) and titrated to pH 8 with aqueous NH4OH (0.2 M). Upon
lyophilization, compounds 9, 11, and 13 were obtained in quantitative
yields as their diammonium salts.
Representative Procedure for the Deprotection of Ethyl
Protected Analogues. A solution of ethyl protected compound 14
or 17 (50 mg) was dissolved in dichloromethane (2 mL) and cooled
to 0 °C in an ice bath before TMSI (55 equiv) was added dropwise
with stirring. The ice bath was removed, and the reaction was allowed
to stir at room temperature for 5 h before being quenched with
methanol (4 mL) and concentrated. The crude mixture was dissolved
in water (10 mL) and extracted with diethyl ether (8 × 10 mL). The
aqueous layer was concentrated with methanol (3 × 10 mL) before
being taken up in water (2 mL) and passed though an Amberlite
cation exchange column (H+ form). Acid fractions were pooled,
concentrated, and titrated to pH 8 with aqueous NH4OH (0.2 M).
Lyophilization provided compounds 15 and 18 were obtained in high
yields (94−95%) as their diammonium salts.
H3, 3J1,2 = 2.4 Hz), 3.80 (m, 1H, H2), 3.54 (dd, 1H, H4, 3J4,5 = 9.3 Hz,
3J4,3 = 3.4 Hz), 3.33 (m, 2H, H5, H6), 1.23 (d, 3H, H7, 3J7,6 = 5.7 Hz).
δC: 77.96 (s, CH, C2), 76.70 (s, CH, C6), 73.12 (s, CH, C4), 72.19 (s,
CH, C5), 68.24 (s, CH, C3), 16.72 (s, CH3, C7). δP: 3.08 (t, 1P, 2JP,F
=
=
=
2
2
89.2 Hz). δF: −110.09 to −111.60 (dd, 1F, JF,F = 297.1 Hz, JF,P
2
2
80.5 Hz), −119.76 to −121.22 (dd, 1F, JF,F = 296.00 Hz, JF,P
78.3 Hz). HRMS (ESI−): found [M − H]− 277.0289. C7H12F2O7P1
requires [M − H]− 277.0294.
ASSOCIATED CONTENT
* Supporting Information
■
S
NMR spectra, HPLC traces, kinetic parameter calculations,
amino acid sequence alignments, Cps2L tolerance to DMSO,
Michaelis−Menten and Lineweaver−Burk plots, and IC50
curves. This material is available free of charge via the Internet
at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*D. L. Jakeman. E-mail: david.jakeman@dal.ca.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
RmlB-D (Aneurinibacillus thermoaerophilis) plasmids were
provided by Dr. Joseph Lam, Department of Molecular and
Cellular Biology, University of Guelph, Canada.41 Human PNP
plasmid was provided by Dr. Vern L. Schramm, Department of
Biochemistry, Albert Einstein College of Medicine, USA.42 This
work was supported in part by the Natural and Engineering
Sciences Council of Canada and the Canadian Institutes of
Health Research.
Ammonium-(1-deoxy-α-L-rhamno-heptulopyranosyl)-
phosphonate (α-L-Rhamnose-1C-ketosephosphonate) (9). δH:
(D2O) 3.86 (m, 1H, H4), 3.60−3.83 (m, 2H, H3, H6), 3.24 (m,
1H, H5), 1.70−2.00 (m, 2H, H1), 1.15 (d, 3H, H7, 3J7,6 = 6.2 Hz). δC:
93.93 (s, C, C2), 68.22−72.87 (4s, 4 × CH, C3−6), 48.80 (s, CH2,
C1) 16.80 (s, CH3, C7). δP: 15.41 (s, 1P). HRMS (ESI−): found [M −
H]− 257.0417. C7H14O8P1 requires [M − H]− 257.0432.
Ammonium-(2,6-anhydro-1-deoxy-β-L-rhamno-heptulopyranosyl)-
phosphonate (β-L-Rhamnose-1C-phosphonate) (11). δH: (D2O) 3.80
REFERENCES
■
3
(1) Pallares, R.; Viladrich, P. F.; Linares, J.; Cabellos, C.; Gudiol, F.
Microb. Drug Resist. (New Rochelle, NY, U. S.) 1998, 4, 339−347.
(2) Smith, A. M.; Klugman, K. P. Antimicrob. Agents Chemother. 1998,
42, 1329−1333.
(br d, 1H, H3, J3,4 = 3.3 Hz), 3.74 (m, 1H, H2), 3.50 (dd, 1H, H4,
3J4,5 = 9.1 Hz, 3J4,3 = 3.3 Hz), 3.30−3.15 (m, 2H, H5, H6), 1.95 (ddd,
2
3
2
2H, H1a, H1b, JH,P = 18.2 Hz, J1,2 = 6.7 Hz, JH,H = 1.42 Hz), 1.15
3
(d, 3H, H7, J7,6 = 6.0 Hz). δC: 76.09−72.24 (3 × CH, C4−6), 73.76
(d, CH, C2, JC,P = 11.1 Hz), 71.07 (d, CH, C3, JC,P = 9.3 Hz),
2
3
(3) Grebe, T.; Hakenback, R. Antimicrob. Agents Chemother. 1996, 40,
829−834.
1
29.75−28.68 (d, CH2, C1, JC,P = 135.8 Hz), 16.98 (s, CH3, C7). δP:
24.81 (s, 1P). HRMS (ESI−): found [M − H]− 241.0478. C7H14O7P1
(4) Smith, A. M.; Klugman, K. P. Antimicrob. Agents Chemother. 1995,
39, 859−867.
requires [M − H]− 241.0483.
(5) Markiewicz, Z.; Tomasz, A. J. Clin. Microbiol. 1989, 27, 405−410.
(6) Strateva, T.; Yordanov, D. J. Med. Microbiol. 2009, 58, 1133−
1148.
Ammonium-(1-deoxy-R-1-fluoro-α-L-rhamno-heptulopyranosyl)-
phosphonate (α-L-Rhamnose-1CF-ketosephosphonate) (L-Rham-
nose-1CF-ketosephosphonate) (13). δH: (D2O) 4.43 (dd, 1H, H1,
2JH,F = 45.4 Hz, 2JH,P = 4.2 Hz), 3.86 (m, 1H, H4), 3.70−3.78 (m, 2H,
(7) Qu, H.; Xin, Y.; Dong, X.; Ma, Y. FEMS Microbiol. Lett. 2007,
275, 237−243.
3
3
H3, H6), 3.27 (t, 1H, H5, J5,4 = J5,6 = 9.8 Hz), 1.16 (d, 3H, H7,
3J7, 6 = 6.1 Hz). δC: 86.28−87.64 (m, CH, C1), 72.61 (s, CH, C5),
(8) Ma, Y.; Stern, R. J.; Scherman, M. S.; Vissa, V. D.; Yan, W.; Jones,
V. C.; Zhang, F.; Franzblau, S. G.; Lewis, W. H.; McNeil, M. R.
Antimicrob. Agents Chemother. 2001, 45, 1407−1416.
(9) Alphey, M. S.; Pirrie, L.; Torrie, L. S.; Boulkeroua, W. A.;
Gardiner, M.; Sarkar, A.; Maringer, M.; Oehlmann, W.; Brenk, R.;
Scherman, M. S.; McNeil, M.; Rejzek, M.; Field, R. A.; Singh, M.;
Gray, D.; Westwood, N. J.; Naismith, J. H. ACS Chem. Biol. 2013, 8,
387−396.
70.56 (s, CH, C3), 70.09 (s, CH, C4), 68.31 (s, CH, C6), 16.93 (CH3,
C7). δP: 11.00 (d, 1P, 2JP,F = 67.9 Hz). δF: −213.80 (d, 1F, 2JF,P = 68.7 Hz).
HRMS (ESI−): found [M − H]− 275.0338. C7H13F1O8P1 requires
[M − H]− 275.0338.
Ammonium-(2,6-anhydro-1-deoxy-1,1-difluoro-β-L-rhamno-
heptulopyranosyl)phosphonate (β-L-Rhamnose-1CF2-phospho-
nate) (L-Rhamnose-1CF2-ketosephosphonate) (15). δH: (D2O)
3
3
4.09 (d, 1H, H3, J3,4 = 3.6 Hz), 3.79 (dd, 1H, H4, J4,5 = 10.2 Hz,
(10) Barton, W. A.; Lesniak, J.; Biggins, J. B.; Jeffrey, P. D.; Jiang, J.
Q.; Rajashankar, K. R.; Thorson, J. S.; Nikolov, D. B. Nat. Struct. Biol.
2001, 8, 545−551.
3J4,3 = 3.6 Hz), 3.75 (m, 1H, H6) 3.31 (t, 1H, H5, J5,4 = J5,6
=
3
3
9.7 Hz), 1.20 (d, 3H, H7, 3J7,6 = 6.3 Hz). δC: 72.05 (s, CH, C5), 70.29
(s, CH, C3), 68.90−70.20 (2s, 2 × CH, C4, C6), 16.76 (s, CH3, C7).
(11) Moretti, R.; Thorson, J. S. J. Biol. Chem. 2007, 282, 16942−
2
16947.
δP: 3.92 (t, 1P, JP,F = 77.2 Hz). δF: −116.84 to −118.43 (dd, 1F,
2JF,F = 295.4 Hz, 2JF,P = 77.2 Hz), −122.08 to −123.62 (dd, 1F, 2JF,F
=
(12) Zuccotti, S.; Zanardi, D.; Rosano, C.; Sturla, L.; Tonetti, M.;
Bolognesi, M. J. Mol. Biol. 2001, 313, 831−843.
(13) Kornfeld, S.; Glaser, L. J. Biol. Chem. 1961, 236, 1791−1794.
(14) Blankenfeldt, W.; Asuncion, M.; Lam, J. S.; Naismith, J. H.
EMBO J. 2000, 19, 6652−6663.
295.4 Hz, 2JF,P = 77.2 Hz). HRMS (ESI−): found [M − H]− 293.0240.
C7H12F2O8P1 requires [M − H]− 293.0243.
Ammonium-(2,6-anhydro-1-deoxy-1,1-difluoro-β-L-rhamno-
heptulopyranosyl)phosphonate (β-L-rhamnose-1CF2-phosphonate)
(L-rhamnose-1CF2-phosphonate) (18). δH: (D2O) 4.41 (br d, 1H,
(15) Melo, A.; Glaser, L. J. Biol. Chem. 1965, 240, 398−405.
9832
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833

The Journal of Organic Chemistry
Article
(16) Engel, R. Chem. Rev. 1977, 77, 349−367.
(17) Romanenko, V. D.; Kukhar, V. P. Chem. Rev. 2006, 106, 3868−
3935.
(18) Chang, W.; Dey, M.; Liu, P.; Mansoorabadi, S. O.; Moon, S.;
Zhao, Z. K.; Drennan, C. L.; Liu, H. Nature 2013, 496, 114−118.
(19) Cload, P. A.; Hutchinson, D. W. Nucleic Acids Res. 1983, 11,
5621−5628.
(20) Fonong, T.; Burton, D. J.; Pietrzyk, D. J. Anal. Chem. 1983, 55,
1089−1094.
(21) Chen, W.; Flavin, M. T.; Filler, R.; Xu, Z. J. Chem. Soc., Perkin
Trans. 1998, 3979−3988.
(22) Hutchinson, D. W. Antiviral Res. 1985, 5, 193−205.
(23) Blackburn, G. M. Chem. Ind. 1981, 134−138.
(24) Wnuk, S.; Robins, M. J. Am. Chem. Soc. 1996, 118, 2519−2520.
(25) Nieschalk, J.; Batsanov, A.; OHagan, D.; Howard, J. Tetrahedron.
1996, 52, 165−176.
(26) Berkowitz, D. B.; Bose, M.; Pfannenstiel, T. J.; Doukov, T. J.
Org. Chem. 2000, 65, 4498−4508.
(27) Forget, S. M.; Bhattasali, D.; Hart, V. C.; Cameron, T. S.;
Syvitski, R. T.; Jakeman, D. L. Chem. Sci. 2012, 3, 1866−1878.
(28) Fernan
Hernandez-Vaz
M. L.; Sanchez-San
J. Eur. J. Med. Chem. 2010, 45, 4827−4837.
́
dez-Herrera, M. A.; Mohan, S.; Lop
quez, J. M. V.; Per
chez, L.; Regla, I.; Pinto, B. M.; Sandoval-Ramírez,
́
ez-Munoz, H.;
ez-Cervantes, E.; Escobar-Sanchez,
́ ́
̃
́
́
́
́
(29) van Well, R. M.; Marinelli, L.; Erkelens, K.; van der Marel, G. A.;
Lavecchia, A.; Overkleeft, H. S.; van Boom, J. H.; Kessler, H.;
Overhand, M. Eur. J. Med. Chem. 2003, 2303−2313.
(30) Albright, J.; Goldman, L. J. Am. Chem. Soc. 1967, 89, 2416−
2423.
(31) Meyer, O.; Grosdemange-Billiard, C.; Tritsch, D.; Rohmer, M.
Org. Biomol. Chem. 2003, 1, 4367−4372.
(32) Norris, A. J.; Toyokuni, T. J. Carbohydr. Chem. 1999, 18, 1097−
1105.
(33) Nyffeler, P. T.; Duron, S. G.; Burkart, M. D.; Vincent, S. P.;
Wong, C. H. Angew. Chem., Int. Ed. 2005, 44, 192−212.
(34) Timmons, S. C.; Mosher, R. H.; Knowles, S. A.; Jakeman, D. L.
Org. Lett. 2007, 9, 857−860.
(35) Peng, J. W.; Moore, J.; Abdul-Manan, N. Prog. Nucl. Magn.
Reson. Spectrosc. 2004, 44, 225−256.
(36) Skinner, A. L.; Laurence, J. S. J. Pharm. Sci. 2008, 97, 4670−
4695.
(37) Beaton, S. A.; Huestis, M. P.; Sadeghi-Khomami, A.; Thomas, N.
R.; Jakeman, D. L. Chem. Commun. 2009, 238−240.
(38) Timmons, S. C.; Hui, J. P.; Pearson, J. L.; Peltier, P.; Daniellou,
R.; Nugier-Chauvin, C.; Soo, E. C.; Syvitski, R. T.; Ferrieres, V.;
Jakeman, D. L. Org. Lett. 2008, 10, 161−163.
(39) Cheng, Y. C.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22,
3099−3108.
(40) Webb, M. R. Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 4884−4887.
(41) Allard, S. T.; Giraud, M. F.; Whitfield, C.; Graninger, M.;
Messner, P.; Naismith, J. H. J. Mol. Biol. 2001, 307, 283−295.
(42) Lewandowicz, A.; Schramm, V. L. Biochemistry 2004, 43, 1458−
1468.
9833
dx.doi.org/10.1021/jo401542s | J. Org. Chem. 2013, 78, 9822−9833