Synthesis of P1-Citronellyl-P2-α-D-pyranosyl pyrophosphates as potential substrates for the E. coli undecaprenyl-pyrophosphoryl-N-acetylglucoseaminyl transferase MurG

Article abstract of DOI:10.1016/S0960-894X(01)00653-9

P¹-Citronellyl-P²-α-D-pyranosyl pyrophosphates containing α-D-N-acetylglucoseaminyl, α-D-glucosyl, and α-D-N-acetylmuramyl carbohydrates were synthesized and used in substrate specificity studies of the Escherichia coli MurG enzyme. Oxalyl chloride activation of citronellyl phosphate for coupling to α-D-pyranose-1-phosphates resulted in markedly improved yields over traditional Khorana-Moffatt and diphenyl chlorophosphate activation strategies.

Full text of DOI:10.1016/S0960-894X(01)00653-9

Bioorganic & Medicinal Chemistry Letters 11 (2001) 3107–3110
Synthesis of P1-Citronellyl-P2-ꢀ-D-pyranosyl Pyrophosphates as
Potential Substrates for the E. coli Undecaprenyl-pyrophosphoryl-
N-acetylglucoseaminyl Transferase MurG
Predrag Cudic, Douglas C. Behenna, Michael K. Yu, Ryan G. Kruger,
Lawrence M. Szewczuk and Dewey G. McCafferty*
Johnson Research Foundation and the Department of Biochemistry and Biophysics, The University of Pennsylvania
School of Medicine, Philadelphia, PA 19104-6059, USA
Received 17 August 2001; accepted 11 September 2001
Abstract—P1-Citronellyl-P2-a-d-pyranosyl pyrophosphates containing a-d-N-acetylglucoseaminyl, a-d-glucosyl, and a-d-N-ace-
tylmuramyl carbohydrates were synthesized and used in substrate specificity studies of the Escherichia coli MurG enzyme. Oxalyl
chloride activation of citronellyl phosphate for coupling to a-d-pyranose-1-phosphates resulted in markedly improved yields over
traditional Khorana–Moffatt and diphenyl chlorophosphate activation strategies. # 2001 Elsevier Science Ltd. All rights reserved.
In bacterial cell-wall biosynthesis, the enzyme MurG
catalyzes the final membrane-associated step in the
biosynthesis of the peptidoglycan monomer.1À3 MurG
utilizes the substrates uridyl-diphospho-N-acetyl-
glucosamine (UDP-GlcNAc, 3) and undecaprenyl-pyro-
phosphoryl-N-acetyl-muramyl-pentapeptide (Lipid I, 1)
to produce uridyl-diphosphate (UDP, 4) and undecapre-
nyl - pyrophosphoryl - N - acetyl - muramyl - pentapeptide
(Lipid II, 2) (Scheme 1). Lipid II (2) is subsequently
exported to the outer surface of the bacterial cell where
itbecomes polymerized and crosslinked inot mautre
peptidoglycan by transglycosylases. Because MurG cata-
lyzes an essential cell wall biosynthesis step, it is emerging
as an attractive target for the development of new anti-
microbial agents to help stave off infections due to
multidrug resistant microorganisms.
For years, a limiting factor barring analysis of MurG
has been its difficult purification away from the mem-
brane and the lack of availability of its natural sub-
strate, Lipid I (1), due to its low natural abundance,4
chemical complexity, difficult isolation,5 and propensity
to form detergent-like micelles.4 Recentadvances in the
cloning,6 overexpression,1,7,8 purification,7,8 and crystal
structure9 of Escherichia coli MurG together with the
discovery of water-soluble substrates8,10 and novel assay
Scheme 1. The bacterial peptidoglycan biosynthesis MurG reaction.
*Corresponding author. Tel.: +1-215-898-7619; fax: +1-215-573-8052; e-mail: deweym@mail.med.upenn.edu
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00653-9

3108
P. Cudic et al. / Bioorg. Med. Chem. Lett. 11 (2001) 3107–3110
methods8,10,11 have paved the way for detailed analysis
of its mechanism and substrate specificity. Van Heije-
noortand co-workers 12 prepared a fluorescentanalogue
of Lipid I, heptaprenyl-pyrophosphoryl-MurNAc-pen-
tapeptide (Ne-Lys-dansylated) from naturally-isolated
UDP-N-acetylmuramyl pentapeptide13 (Park’s nucleo-
tide). Hitchcock and co-workers14 of Lilly Research
laboratories recently developed a convergent total syn-
thetic route to Park’s nucleotide from O-benzyl-N-ace-
tyl-4,6-benzylidinemuramic acid,15 providing the first
route to this class of compounds via total synthesis, thus
facilitating synthetic modification of the peptidoglycan
intermediate architecture. Building upon this precedent,
VanNieuwenhze and colleagues16 (also from Lilly)
recently reported the first total synthesis of native
undecaprenyl–Lipid I. In a related approach, Walker
and co-workers10 synthesized an analogue of Lipid I (5)
with two modifications that improved solubility and
facilitated affinity capture and assay detection. Walker’s
group replaced the C-55 undecaprenol native lipid with
a shorter C-10 chain derived from (R)-(+)-b-citronellol
and also labeled the molecule with d-biotin.
kinetically competent minimal substrate replacements.
We report here a new synthesis of full-length citro-
nellyl–Lipid I and three related P1-citronellyl-P2-a-d-
pyranosyl pyrophosphates and their evaluation as
potential substrates for MurG.
P1-Citronellyl-P2-a-d-glucosyl pyrophosphate (14) and
the related P1-citronellyl-P2-a-d-N-acetylglucoseaminyl
pyrophosphate (15) were prepared in six steps from
hyperacetylated carbohydrate precursors 8 and 9,
respectively (Scheme 2). Selective deacetylation of C-1
with hydrazine acetate17 (69–74% yield) followed by
four step conversion of the anomeric alcohol provided
the glycosyl phosphates 10 and 11 in 42 and 35% yields,
respectively. (R)-(+)-b-citronellyl phosphate (7) was
prepared separately according to the method of Danilov
and Chojancki.18 Attempts at Khorana–Moffatt18 acti-
vation of 7 for pyrophosphate bond formation pro-
duced almost exclusively the guanidine byproduct,
normally an undesirable side reaction associated with
this procedure.19 Similarly, use of diphenyl chloropho-
sphate20 for activation of 7 resulted in erradic yields of
the desired pyrophosphate.
Preliminary substrate specificity studies with soluble
preparations of MurG have shown a distinct preference
for the authentic Lipid I substrate, and a capacity for
glycosylating smaller, truncated Lipid I analogues such
as 6, albeit with less efficiency. This observation has
introduced the possibility of using smaller, synthetically
tractable surrogate substrates for mechanistic evalua-
tion or inhibitor screening. Since the multistep synthesis
of 5 is costly and time consuming,8,10 we set out to both
improve key elements of this synthesis that were not
amenable to preparative scaleup and to also establish if
simpler P1-citronellyl-P2-a-d-pyranosyl pyrophosphates
containing a-d-N-acetylglucoseaminyl, a-d-glucosyl,
and a-d-N-acetylmuramyl carbohydrates might be
The solution to this problem was found by using oxalyl
chloride as an activator to form the highly reactive
dichlorophosphate of 7.21 Treatment of 7 with 20 equiv
of oxalyl chloride and glycosyl phosphates 10 and 11
rapidly generated the desired pyrophosphates 12 and 13
in good yields after hydrolytic workup with acetone/
H2O/triethylamine (88:10:2, v/v/v; 2 h, rt) (75 and 70%
yields, respectively).21 Lastly, to facilitate complete
deprotection of the acetates, 2.2 equiv of sodium meth-
oxide in methanol were required to provide compounds
14 and 15 in 70 and 65% yields, respectively. All inter-
mediates and final products were characterized by NMR
and ESI-MS.22,23
Scheme 2. Reactions and conditions: (a) H2NNH2 acetate, DMF, 15
min, 50 ꢀC; (b) dibenzyl N,N-diethylphosphoramidite, 1,2,4-triazole,
CH2Cl2, rt; (c) 30% H2O2, THF, À70 ꢀC to rt; (d) H2, Pd/C, cyclo-
hexylamine, MeOH, 16 h; (e) (R)-(+)-b-citronellyl-OPOCl2, pyridine,
DMF, 1.5 h, rt; (f) NaOMe 2.2 equiv, MeOH, 15 min, rt.
Figure 1. Biotinylated citronellyl–Lipid I (5) and two related fragments.

P. Cudic et al. / Bioorg. Med. Chem. Lett. 11 (2001) 3107–3110
3109
P1-Citronellyl-P2-a-d-N-acetylmuramyl pyrophosphate
(19) was synthesized from O-benzyl 4,6-benzylidine-
muramic acid15 (16). First, benzylidine 16 was converted
to dibenzylphosphate 17 in 30% yield over six steps
according to the method of Hitchcock and colleagues.14
Hydrogenolysis of 17 followed by oxalyl chloride-medi-
ated coupling to 7 produced pyrophosphate 18 in 74%
yield over two steps. Lastly, global deacetylation with
NaOMe/MeOH produced the desired compound 19 in
78% yield. Compound 19 and intermediates were char-
acterized by elemental analysis, NMR and ESI-MS
(Scheme 3).24
during preparative scale up of this synthesis (>25
mmol), we routinely observed a marked reduction in
yields for this step (<35% yield). To address this issue,
a preparative synthesis of citronellyl-Lipid I was devised
whereby oxalyl chloride activation was successfully
employed as a replacementfor hte diphenyl chloro-
phosphate method (Scheme 4).
On a 25 mmol scale, the phenylsulfonylethyl ester of
protected phosphate 1714 was removed by treatment
with 1,8-diazabicyclo[5.4.0]undec-7-ene(1,5-5) (DBU)
and the resulting acid was coupled to H-l-Ala-d-iso-
Glu(OTMSE)-l-Lys(TEOC)-d-Ala-d-Ala-OMe25 using
PyBOP, HOBt, and NMM to yield compound 20 in
15% overall yield over 10 steps.26 Peptide 20 was quan-
titatively deprotected by catalytic hydrogenation and
subsequently coupled to 7 using oxalyl chloride activa-
tion (74% yield after hydrolytic workup with acetone/
H2O/triethylamine solvent mixture (88:10:2 v/v/v) for
2 h at rt). Global deprotection was accomplished by
treatment with NaOMe/MeOH followed by tetrabutyl-
ammonium fluoride (TBAF). Preparative reversed-
phase HPLC purification over octadecyl silica (Vydac
C18, 50 mM ammonium acetate, pH 4.5) afforded pure
21 in 5% overall yield from 16. NMR and ESI-MS
spectra for 21 exactly matched that of reported values.10
Walker’s total synthesis of citronellyl–Lipid I (5) on a
ꢁ1.0 mmol scale8 employed diphenyl chlorophosphate
activation of 7 for assembly of the key pyrophosphate
bond. Although this reaction occurred in 68% yield,
To evaluate the activity of compounds 14, 15, 19, and
21, a new activity assay for MurG was devised. Instead
of monitoring the production of Lipid II, this dis-
continuous assay monitored the conversion of the sec-
ond MurG substrate, UDP-GlcNAc to UDP using thin-
layer chromatography. [a-32P]-UDP-GlcNAc was pre-
pared chemo-enzymatically from glucoseamine-1-phos-
phate and [a-32P]-UTP according to the method of
Anderson and co-workers.27 Soluble recombinant E. coli
MurG was prepared according to the method of
Walker.8 Conversion of [a-32P]-UDP-GlcNAc to
[a-32P]-UDP was monitored by TLC on PEI-cellulose
plates that were eluted with 0.3 M guanidine hydro-
chloride in water and visualized by phosphorimagery
(Molecular Dynamics). Compounds 21, 14, 15, or 19
(100 mM) and [a-32P]-UDP-GlcNAc (100 mM) were
incubated with MurG (0.01 mg/mL) for 20 min at
25 ꢀC in 20 mM HEPES pH 8, 2 mM MgCl2. Upon
Scheme 3. Reactions and conditions: (a) 1-(3-dimethylaminopropyl)-
3-ethylcarbodiimide
hydrochloride,
2-(phenylsulphonyl)ethanol,
DMAP, CH2Cl2, 0 ꢀC to rt, 2 h, 96%; (b) AcOH, H2O, reflux, 45 min;
(c) Ac2O, pyridine, 0 ꢀC to rt, 85%; (d) H2, Pd/C, AcOH, 16 h, 90%;
(e) dibenzyl N,N-diethylphosphoramidite, 1,2,4-triazole, CH2Cl2, rt ;
(f) 30% H2O2, THF, À70 ꢀC to rt, 32%; (g) H2, Pd/C, cyclohexyl-
amine, MeOH, 16 h, quantitative; (h) (R)-(+)-b-citronellyl-OPOCl2,
pyridine, DMF, 1.5 h, rt, 74%; (i) NaOMe 2.2 equiv, MeOH, 15 min,
rt, 78%.
Scheme 4. Reactions and conditions: (a) DBU, CH2Cl2, rt; (b) PyBOP, NMM, l-Ala-g-d-Glu(OTmse)-l-Lys(Teoc)-d-Ala-d-Ala(OMe), DMF, rt;
65% over two steps; (c) H2, Pd/C, cyclohexylamine, MeOH, 16 h, quant; (d) (R)-(+)-b-citronellyl-OPOCl2, pyridine, DMF, 1.5 h, rt, 74%; (e)
NaOMe 2.2 equiv, MeOH, 15 min, rt, 46%.

3110
P. Cudic et al. / Bioorg. Med. Chem. Lett. 11 (2001) 3107–3110
completion of the reaction, 5 mL of the reaction mixture
was spotted onto the TLC plates and the plates were
eluted with 0.3 M guanidine hydrochloride ([a-32P]-
UDP-GlcNAc Rf 0.77; [a-32P]-UDP Rf 0.11). Though
full-length citronellyl-Lipid I 21 was processed by the
enzyme, the truncated pyrophosphates were not recog-
nized as substrates by soluble MurG to the sensitivity
limits afforded by this assay (<0.01% productconver-
sion). Increasing the concentration of compounds 14,
15, and 19 (100-fold), the concentration of MurG (100-
fold), and the time of the reaction (120 min) did not
improve turnover of these compounds. Interestingly, in
trans addition of free l-Ala-d-isoGlu-l-Lys-d-Ala-d-
Ala pentapeptide (Bachem) also had no stimulatory
effect on the MurG reaction with compounds 14, 15,
and 19. Neither compounds 14, 15, or 19 possessed
antimicrobial activity against a Bacillus subtilis Gram-
positive bacterial strain.
5. Anderson, J. S.; Matsuhashi, M.; Haskin, M. A.; Stro-
minger, J. L. J. Biol. Chem. 1967, 242, 3180.
6. Mengin-Lecreulx, D.; Texier, L.; van Heijenoort, J. J.
Nucleic Acids Res. 1990, 18, 2810.
7. Crouvoisier, M.; Mengin-Lecreulx, D.; van Heijenoort, J.
FEBS Lett. 1999, 449, 289.
8. Ha, S.; Chang, E.; Lo, M. C.; Men, H.; Park, P.; Ge, M.;
Walker, S. J. Am. Chem. Soc. 1999, 121, 8415.
9. Ha, S.; Walker, D.; Shi, Y.; Walker, S. Protein Sci. 2000, 9,
1045.
10. Men, H.; Park, P.; Ge, M.; Walker, S. J. Amer. Chem.
Soc. 1998, 120, 2484.
11. Branstrom, A. A.; Midha, S.; Longley, C. B.; Han, K.;
Baizman, E. R.; Axelrod, H. R. Anal. Biochem. 2000, 280, 315.
12. Auger, G.; Crouvoisier, M.; Caroff, M.; van Heijenoort,
J.; Blanot, D. Lett. Peptide Sci. 1997, 4, 371.
13. Park, J. T. J. Biol. Chem. 1952, 194, 877.
14. Hitchcock, S. A.; Eid, C. N.; Aikins, J. A.; Zia-Ebrahimi,
M.; Blaszczak, L. J. Am. Chem. Soc. 1997, 120, 1916.
15. Jeanloz, R. W.; Walker, E.; Sinay, P. Carbohydr. Res.
1968, 6, 184.
This data, taken together with prior substrate specificity
studies conducted with soluble and membrane prepara-
tions of the enzyme8,28 suggests that MurG requires
P1-lipid-P2-N-acetylmuramyl pyrophosphate substrates
containing an intact amide between the muramyl lactyl
ether sidechain and the pentapeptide. The length and
composition of this peptide is important, for truncation
of this pentapeptide to diamide 6 reduced activity by
500-fold as compared to 5,8 and elimination of the lac-
tate amide substituent in the related compound 19
eliminated activity altogether. In an effort to elucidate
the molecular basis of MurG specificity, further struc-
ture–activity studies with synthetic P1-lipid-P2-a-d-pyr-
anosyl pyrophosphate peptides are underway and will
be reported in due course.
16. VanNieuwenhze, M. S.; Maudlin, S. C.; Zia-Ebrahimi,
M.; Aikins, J. A.; Blszczak, L. C. J. Am. Chem. Soc. 2001, 123,
6983.
17. Excoffier, G.; Gagnaire, D.; Utille, J.-P. Carbohydr. Res.
1975, 39, 368.
18. Danilov, L. L.; Chojancki, T. FEBS Lett. 1981, 131, 310.
19. Moffatt, J. G.; Khorana, H. G. J. Am. Chem. Soc. 1961,
83, 649.
20. Warren, C. D.; Jeanloz, R. W. Meth. Enzymol. 1978, 122.
21. Imperiali, B.; Zimmerman, J. W. Tetrahedron Lett. 1990,
31, 6485.
22. Analytical data for 14: MS (ESI) m/z 477.5 (MÀH); 1H
NMR (D2O, 250 MHz) d 0.87 (d, J=7.3 Hz, 3H), 1.17–1.51
(m, 7H), 1.60 (s, 3H), 1.66 (s, 3H), 2.01 (m, 2H) 3.31–4.6 (m,
6H), 5.00 (s, 1H), 5.22 (s, 1H).
1
23. Analytical data for 15: MS (ESI) m/z 520.5 (M+H); H
NMR (D2O, 250 MHz) d 0.89 (s, 3H), 1.05–1.35 (m, 7H), 1.61
(s, 3H), 1.67 (s, 3H), 1.81 (s, 3H), 2.06 (m, 3H), 3.66–3.97 (m,
6H), 5.22 (s, 1H), 5.48 (s, 1H).
Acknowledgements
1
24. Analytical data for 19: MS (ESI) m/z 748.6 (M+H), H
NMR (CD3OD, 250 MHz) d 0.95 (d, J=6 Hz, 3H), 1.38 (d,
J=6.7 Hz, 3H), 1.42 (m, 1H), 1.59 (s, 3H), 1.66 (s, 3H), 1.80
(s, 3H), 1.88 (s 1H), 2.08 (m, 2H), 3.20–3.90 (m, 7H), 4.60 (m,
4H), 5.78 (m, 2H); 13C NMR (DMSO-d6, 500 MHz) 19.93,
19.97, 22.00, 22.51, 54.00, 55.66, 60.29, 69.55, 70.71, 77.57,
77.92, 78.91, 80.69, 94.55, 127.40, 171.72, 174.97.
The authors gratefully acknowledge Dr. Stephen Hitch-
cock of Lilly Research Laboratories for the gift of ben-
zyl-N-acetyl-4,6-benzylidinemuramic acid. Funding
from the ACS (RPG CCE-98797), the McCabe Fund,
and the NIH (RO1 AI46611) supported this work.
25. The protected pentapeptide was prepared from Cbz-d-
Ala-d-Ala-OMe in 65% overall yield using standard solution-
phase methods (PyBOP, HOBt, NMM) and Cbz amino acids.
26. Abbreviations: PyBOP, (benzotriazol-1-yloxy)tris-pyrroli-
dinophosphonium hexafluorophosphate; HOBt, 1-hydro-
xybenzotriazole; NMM, N-methylmorpholine.
References and Notes
1. Mengin-Lecreulx, D.; Texier, L.; Rousseau, M.; van Heije-
noort, J. J. Bacteriol. 1991, 173, 4625.
2. Bugg, T. D. H.; Walsh, C. T. Nat. Prod. Rep. 1992, 192.
3. Holtje, J.-V. Microbiol. Mol. Biol. Rev. 1998, 62, 181.
4. van Heijenoort, Y.; Gomez, M.; Derrien, M.; Ayala, J.; van
Heijenoort, J. J. Bacteriol. 1992, 174, 3459.
27. Kelly, T. M.; Stachula, S. A.; Raetz, C. R. H.; Anderson,
M. S. J. Biol. Chem. 1993, 268, 19866.
28. Silva, D. J.; Bowe, C. L.; Branstrom, A. A.; Baizman,
E. R.; Sofia, M. J. Bioorg. Med. Chem. Lett. 2000, 10, 2811.