Lipid II: Total synthesis of the bacterial cell wall precursor and utilization as a substrate for glycosyltransfer and transpeptidation by penicillin binding protein (PBP) 1b of Eschericia coli

Article abstract of DOI:10.1021/ja0166848

An essential feature in the life cycle of both gram positive and gram negative bacteria is the production of new cell wall. Also known as murein, the cell wall is a two-dimensional polymer, consisting of a linear, repeating N-acetylmuramic acid (MurNAc) a

Full text of DOI:10.1021/ja0166848

11638
J. Am. Chem. Soc. 2001, 123, 11638-11643
Lipid II: Total Synthesis of the Bacterial Cell Wall Precursor and
Utilization as a Substrate for Glycosyltransfer and Transpeptidation
by Penicillin Binding Protein (PBP) 1b of Eschericia coli
Benjamin Schwartz,* Jay A. Markwalder, and Yi Wang
Contribution from the Department of Chemical and Physical Sciences, DuPont Pharmaceuticals Company,
Wilmington, Delaware 19880
ReceiVed July 23, 2001
Abstract: An essential feature in the life cycle of both gram positive and gram negative bacteria is the production
of new cell wall. Also known as murein, the cell wall is a two-dimensional polymer, consisting of a linear,
repeating N-acetylmuramic acid (MurNAc) and N-acetylglucosamine (GlcNAc) motif, cross-linked via peptides
appended to MurNAc. The final steps in the maturation of murein are catalyzed by a single, bifunctional
enzyme, known as a high MW, class A penicillin binding protein (PBP). PBPs catalyze polymerization of the
sugar units (glycosyltransfer), as well as peptide cross-linking (transpeptidation) utilizing Lipid II as substrate.
Detailed enzymology on this enzyme has been limited, due to diffculties in obtaining sufficient amounts of
Lipid II, as well as the availability of a convenient and informative assay. We report the total chemical synthesis
of Lipid II, as well as the development of an appropriate assay system and the observation of both catalytic
transformations.
Introduction
Scheme 1
Bacterial cell walls consist of a repeating, two-dimensional
array of peptidoglycan known as murein.1 Structurally, murein
is composed of linear polysaccharide chains of alternating
N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid
(MurNAc), cross-linked via peptides appended to MurNAc
(Scheme 1). The final steps in the biosynthesis of murein occur
extracellularly, associated with the cell membrane, and are
catalyzed by high molecular weight, class A penicillin-binding
proteins (PBPs), such as Eschericia coli PBP 1b.2,3 This
bifunctional enzyme catalyzes both sugar polymerization (gly-
cosyltransfer) and peptide cross-linking (transpeptidation) utiliz-
ing the cognate monomer, Lipid II, as substrate (Scheme 1).
desirable. Research in this respect has been slow primarily due
to the lack of a prerequisite understanding of the glycosyltransfer
and transpeptidation reactions. In turn, this is due to an inability
to obtain sufficient quantities of substrate, and the lack of an
informative, yet unencumbered, system for studying the enzyme.
Thus, we have developed a total chemical synthesis of Lipid
II, and developed a homogeneous assay format with recombinant
PBP1b that can be used for detailed enzymology of both
glycosyltransfer and transpeptidation.
Inhibition of either activity of PBP1b, and the analogous
enzymes in other bacterial species, is regarded as an excellent
opportunity for antimicrobial intervention. Natural products
which inhibit both reactions in vitro are known; flavomycin
(moenomycin, bambermycin) blocks the glycosyltranser step,
while penicillin (and all â-lactams) blocks transpeptidation.1
Though these compounds are potent antimicrobial agents, they
suffer serious clinical problems, as flavomycin has poor
bioavailability,4 while resistance to â-lactam drugs develops
quickly and is widespread.5 Recent efforts at analoging flavo-
mycin4 and vancomycin6 to produce better inhibitors of the
glycosyltransfer reaction have been encouraging; however, to
drive drug discovery access to novel classes of compounds is
Synthesis of Lipid II
The synthesis of Lipid II began with the commercially
available MurNAc precursor 1, which was coupled with alanine,
deprotected, and selectively monoacetylated to generate the
glycosyl acceptor 2 (Scheme 2). Alanine was introduced at this
step to avoid the unwanted cyclization of compound 2, which
occurred when the lactate was protected as an ester. Compound
2 was coupled with the donor 3, utilizing AgOTF to activate
the glycosyl bromide.7 The phthalimido group in the 2 position
(1) Bugg, T. D. H. Compr. Nat. Prod. Chem. 1999, 3, 241-294.
(2) Nakagawa, J.; Tamaki, S.; Matsuhashi, M. Agric. Biol. Chem. 1979,
43, 1379-1380.
(3) Nakagawa, J.-i.; Tamaki, S.; Tomioka, S.; Matsuhashi, M. J. Biol.
Chem. 1984, 259, 13937-13946.
(4) Goldman, R. C.; Gange, D. Curr. Med. Chem. 2000, 7, 801-820.
(5) Wong, K. K.; Pompliano, D. L. AdV. Exp. Med. Biol. 1998, 456,
197-217.
(6) Ge, M.; Chen, Z.; Onishi, H. R.; Kohler, J.; Silver, L. J.; Kerns, R.;
Fukuzawa, S.; Thompson, C.; Kahne, D. Science 1999, 284, 507-511.
(7) Paulsen, H.; Himpkamp, P.; Peters, T. Liebigs Ann. Chem. 1986, 4,
664-674.
10.1021/ja0166848 CCC: $20.00 © 2001 American Chemical Society
Published on Web 11/06/2001

Lipid II: Total Synthesis of the Bacterial Cell Wall Precursor
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11639
Scheme 2
a HAlaOTMSE, DIEA, HOBt, PyBOP, 1:1 THF:CH2Cl2, room
temperature, 2 h, 81%. b Catalyst TsOH, MeOH, 75 °C, 30 min, 89%.
c Pyridine, 1 equiv of AcCl, -30 °C, 15 min, then warm to room
temperature, 80%. d 3 equiv of AgOTF, 2 equiv of compound 3, 4 Å
sieves, CH2Cl2, -40 °C, drybox, 3 h, then O.N., 60%. e Amino-modified
resin, 4 Å sieves, butanol, 85 °C, 24 h, then pyridine, Ac2O, room
temperature, O.N., 53%. f 10%Pd/C, H2, MeOH, 1 h, 93%. g Tetrazole,
(BnO)2PN(i-Pr)2, CH2Cl2, -30 °C to room temperature, 1 h, then
m-CPBA, -40 °C to room temperature, 1 h, 55%. h TBAF, THF, room
temperature, 45 min, 87%. i H-γ-D-Glu(OTMSE)-Lys(TEOC)-D-Ala-
D-AlaOTMSE, DIEA, HOBt, PyBOP, 1:1 THF:CH2Cl2, room temper-
Figure 1. Measurement of the glycosyltransferase reaction. (Top) Lipid
II (2 µM) was incubated at 30 °C with PBP1b (60 nM) in 50 mM
HEPES buffer (pH 7.5) containing 0.085% decyl PEG, 10% DMSO,
and 10 mM MgCl2. Aliquots were removed, labeled, and injected on
an anion-exchange HPLC column. Labeled Lipid II elutes at 16-17
min. Samples are at t ) 0, 10, 20, 30, and 40 min (time increasing
front to back). (Bottom) Lipid II was incubated with enzyme as in the
top panel, with the addition of 10 µg/mL of flavomycin. Samples are
at the same reaction times.
j
ature, 2 h, 61%. 10%Pd/C, H2, MeOH, 1 h, 100%. k CDI, dry DMF,
room temperature, 4 h, then MeOH, then undecaprenyl phosphate, room
l
temperature, 48 h, 39%. TBAF, DMF, room temperature, 24 h, then
specificity preferences with respect to both the sugar12 and lipid
moieties,9 it is expected that the synthetic route will not be
limited in this manner. We will pursue generation of these
analogues in the future, for both mechanistic and screening
purposes.
3% NaOMe, MeOH, 0 °C, 1 h, 35%.
of compound 3 directs attack predominantly on the â face, to
give the desired product 4. Following coupling, the phthalimido
group was deprotected utilizing an ethylenediamine-derivatized
resin,8 and the amino group re-acetylated with acetic anhydride.
This compound was then converted to the desired phosphate
by successive anomeric deprotection and phosphorylation to
yield compound 5. The phosphorylation is done in a one-pot,
two-step procedure by initial formation of the phosphite,
followed by in situ oxidation with m-CPBA.9,10 The carboxyl
protecting group of alanine was removed, and the resulting acid
was coupled to protected tetrapeptide to yield the desired
glycopeptide (6). The phosphoryl group was freed, activated
with CDI, and coupled to undecaprenyl phosphate. Finally,
removal of the peptidyl and hydroxyl protecting groups provided
Lipid II in 12 steps and 0.7% overall yield. This report
constitutes the first published total chemical synthesis of Lipid
II;11 the only previously published synthesis of Lipid II is by
Walker and co-workers, using the enzyme MurG to convert the
monosaccharide (and biogenic) precursor Lipid I to the desired
material.9
Observation of Glycosyltransfer Reaction
For almost 20 years, studies of the glycosyltransfer reaction
have required cumbersome and lengthy paper chromatographic
analysis of radiolabled substrates, incubated with either ether-
permeabilized whole bacterial cells or recombinant PBP1b as
a source of enzyme. Though some elegant work has been done
by Suzuki,13 Ghuysen,14 and van Heijenoort15 to study the
glycosyltransfer reaction using this methodology, it is unsuitable
for detailed enzymology, as throughput is low and product
analysis is unavailable.
Using the ready supply of Lipid II afforded by the chemical
synthesis, we constructed an HPLC-based assay to interrogate
the glycosyltransfer reaction. Lipid II contains only a single
primary amino group, making it suitable for derivatization with
fluorescamine. Importantly, this can be done postreaction,
removing any concerns about the effects of labels on the reaction
progress. Incubation of Lipid II with reconstituted PBP1b in a
simple detergent system led to the observation of a time-
dependent depletion of substrate (Figure 1, top). It is worth
pointing out that the system is homogeneous, with both enzyme
and substrate presumably partitioning into detergent micelles.
DMSO (10%) was included in the assays, as it was found to
accelerate the enzymatic reaction, as reported previously.14
It may be useful to compare the synthetic and chemoenzy-
matic routes to Lipid II. The total yield of material appears to
be comparable for each method (milligram quantities), and it
is expected that both should scale without difficulty. One
difference in these approaches may be in their abilities to
generate Lipid II analogues. While MurG is noted to have
(8) Stangier, P.; Hindsgaul, O. Synlett 1996, 2, 179-181.
(9) Ye, X.-Y.; Lo, M.-C.; Brunner, L.; Walker, D.; Kahne, D.; Walker,
S. J. Am. Chem. Soc. 2001, 123, 3155-3156.
(10) Sim, M. M.; Kondo, H.; Wong, C.-H. J. Am. Chem. Soc. 1993,
115, 2260-2267.
(12) Ha, S.; Chang, E.; Lo, M.-C.; Men, H.; Park, P.; Ge, M.; Walker,
S. W. J. Am. Chem. Soc. 1999, 121, 8415-8426.
(11) Scientists at Eli Lilly and Co. have recently reported a synthesis of
Lipid II, but details have not yet been published. VanNieuwenhze, M. S.;
Mauldin, S. C.; Zia-Ebrahimi, M.; Hornback, W. J.; Saha, S. L.; Blaszczak,
L. C. Abstracts of the April 4, 2001, National Meeting of the American
Chemical Society, San Diego, California; American Chemical Society:
Washington, DC, 2001; Abstract No. 470
(13) Hara, H.; Suzuki, S. FEBS Lett. 1984, 168, 155-160.
(14) Terrak, M.; Ghosh, T. K.; van Heijenoort, J.; Van Beeumen, J.;
Lampilas, M.; Aszodi, J.; Ayala, J. A.; Ghuysen, J.-M.; Nguyen-Disteche,
M. Mol. Microbiol. 1999, 34, 350-364.
(15) Van Heijenoort, Y.; Gomez, M.; Derrien, M.; Ayala, J.; van
Heijenoort, J. J. Bacteriol. 1992, 174, 3549-3557.

11640 J. Am. Chem. Soc., Vol. 123, No. 47, 2001
Schwartz et al.
Scheme 3
Figure 2. Effect of inhibitors on the glycosyltransferase reaction.
Enzymatic reactions were carried out as indicated in Figure 1, with
either no inhibitor (0), 100 µM penicillin G (O), or 10 µg/mL of
flavomycin (×)
Flavomycin (or Moenomycin) is a known glycosyltransferase
inhibitor of PBP1b,16 and has been used in past studies to verify
the proposed source of observed activity in assays. The addition
of flavomycin at 10 µg/mL to our reaction led to complete
inhibition of substrate depletion (Figure 1, bottom).
Polymeric Nature of Glycosyltransferase Reaction
To verify that the reaction observed between Lipid II and
PBP1b was the result of glycosyltransfer, attempts were made
to analyze the enzymatic product. Bacterial muramidases are
enzymes capable of hydrolyzing un-cross-linked murein at the
â-1,4 linkage between MurNAc and GlcNAc, producing mono-
meric material17 (Scheme 3). Thus, the addition of muramidase
to our in vitro reaction with PBP1b was expected to yield a
new chemical species, amenable to the same labeling reaction
as Lipid II.
To produce a homogeneous product, and thus simplify the
analysis, reactions were conducted in the presence of penicillin
G, which has been shown to specifically inhibit cross-linking
of murein. Important to this analysis, the addition of penicillin
G was found not to affect the glycosyltransfer reaction (Figure
2).
When reactions were carried out in the presence of murami-
dase, the disappearance of Lipid II was accompanied by the
appearance of a new peak (Figure 3). Consistent with the
possibility that this species was labeled monomer, the new peak
was equal in intensity to Lipid II (monomer and Lipid II are
both labeled through lysine), and eluted more rapidly off of the
anion-exchange HPLC column (monomer has fewer negative
charges than Lipid II as the lipid diphosphate moiety is lost
during the enzymatic reaction). Importantly, this new peak was
not generated by the action of muramidase itself on Lipid II, as
control reactions carried out in the absence of PBP1b showed
no turnover (data not shown).
To obtain the identity of the muramidase-dependent species,
unlabeled product from a reaction containing both PBP1b and
muramidase was isolated and subjected to mass spectral analysis.
Material was obtained by first carrying out a complete reaction
with Lipid II and PBP1b, followed by removal of all small
(<10 000 MW) species by dialysis. Muramidase was then
Figure 3. Glycosyltransferase reaction in the presence of muramidase.
The enzymatic reaction was carried out under similar conditions to
Figure 1, with the addition of 0.02 mg/mL of muramidase. Aliquots
were removed, labeled, and injected on an anion-exchange HPLC
column. Labeled monomer elutes at 14.5 min and labeled Lipid II elutes
at 16 min. Samples are at t ) 0, 2, 6, 10, 15, 30, and 60 min (time
increasing front to back).
added, followed by a second dialysis, in which digested polymer
was collected. This material was found to have a MW of 966.4,
as predicted for the un-cross-linked monomer, and when labeled
was shown to correspond with the new, muramidase-dependent
HPLC peak. Thus, the in vitro enzymatic glycosyltransfer
reaction produces polymeric peptidoglycan.
Observation of Transpeptidation Reaction
As PBP1b is a bifunctional enzyme, reactions with Lipid II
should yield not only polymerized sugars, but also D-Ala as
the result of interstrand peptide cross-linking. It has not been
possible to quantitatively study this reaction, as short peptides
are not utilized as substrates in transpeptidation.18 Qualitatively,
it has been possible to establish the transpeptidation reaction
by digesting mature peptidoglycan with lysozyme, and analyzing
the resultant products.3 For kinetic measurements of transpep-
tidation, the hydrolysis of thioester peptide analogues has been
used to emulate the natural reaction.11 Though thioester hy-
drolysis is likely to be a useful approximation of the transpep-
tidation reaction, as penicillin irreversibly inhibits both reactions,
questions about the substrate specificity and processivity of
PBP1b cannot be answered with this methodology.
(16) Huber, G. In Mechanism of Action of Antibacterial Agents; Springer,
New York, 1979; Vol. 5, pp 135-153.
(17) Barrett, J. F.; Schramm, V. L.; Shockman, G. D. J. Bacteriol. 1984,
159, 520-526.
(18) Jamin, M.; Damblon, C.; Miller, S.; Hakenbeck, R.; Frere, J.-M.
Biochem. J. 1993, 292, 735-741.

Lipid II: Total Synthesis of the Bacterial Cell Wall Precursor
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11641
glycosyltransferases19 and class B, high MW PBPs18), so there
exists in vivo precedence for the results obtained with PBP1b.
Attempts were made to analyze the muramidase-digested
products of this reaction, as with un-cross-linked material.
However, an identifiable species could not be found, possibly
due to the heterogeneous nature of the products.
Conclusions
Using a synthetically prepared Lipid II, a recombinant form
of E. coli PBP1b, and a simple HPLC analytical system, we
have observed both glycosyltransferase and transpeptidase
activities in a simple in vitro system. Products of the observed
glycosyltransfer reaction were shown to be un-cross-linked
murein by the isolation of polymeric product, and subsequent
digestion by muramidase to yield the expected monomeric
material.
Figure 4. Measurement of the transpeptidase reaction. Lipid II (2 µM)
was incubated at 30 °C with PBP1b (60 nM) in 50 mM HEPES buffer
(pH 7.5) containing 0.085% decyl PEG, 10% DMSO, and 10 mM
MgCl2. Aliquots were removed, labeled, and injected on a reverse-
phase C18 HPLC column. Labeled D-Ala elutes at 21.8 min. Samples
are at t ) 0, 2, 6, 10, 20, and 40 min (time increasing front to back).
The largest peak corresponds to 0.28 µM D-Ala.
There are many crucial, yet currently unexplored mechanistic
aspects involved in the polymerizations carried out by PBP1b,
and the analogous enzymes in other bacterial species. These
include the possible processivity of these enzymes, as well as
the effects of detergents, metals, organic cosolvents, and
inhibitors. Using the system described here, we will begin to
study these questions and support future drug discovery efforts.
Experimental Section
General Procedures. 1H NMR spectra were recorded on dilute
solutions in CDCl3, CD3OD, or DMSO-d6 at 300 MHz on Varian Unity
instruments. Chemical shifts are reported in parts per million (δ)
downfield from tetramethylsilane. Low-resolution mass spectral analyses
were performed on an Micromass Platform II (ESI) instrument. High-
resolution mass spectra were obtained on a Finnigan MAT95S (ESI)
instrument. Reactions were performed under an atmosphere of dry
nitrogen or argon in flame-dried glassware and were monitored for
completeness by thin-layer chromatography (TLC) by using silica gel
60 F-254 (0.25 mm) plates. Visualization of TLC plates was ac-
complished by I2 vapor, phosphomolybdic acid in ethanol, ceric
ammonium molybdate in aqueous methanol, or UV light absorption at
254 nm. Flash column chromatography was performed by the method
of Still, using 230-400 mesh silica gel (E Merck). Tetrahydrofuran
was distilled from potassium/benzophenone ketyl immediately prior
to use. Other solvents and reagents were purchased from commercial
sources and were used without further purification.
Cloning, Expression ,and Purification of PBP1b. The gene
encoding PBP1b was amplified from the Streptococcus pneumonia
genomic DNA library (American Type Culture Collection) by PCR. A
Nde I site was included at the 5′ end of the sense primer, and a Xho I
site was included at the 5′ end of the antisense primer. The PCR product
was subsequently inserted into a pet-21a expression vector (Novagen)
between the Nde I and BamH I sites. As a result, a hexahistidine tag
was introduced to the c-terminus of the protein. E. coli cells (BL21-
DE3) transformed with the vectors showed moderate level expression
of PBP1b (5-7 mg/L in LB medium). After harvesting, cells were
suspended in loading buffer (100 mM NaCl, 50 mM NaPi, pH 8.0)
and disrupted in a French Press apparatus. After brief centrifugation,
PBP1b protein was found in pellets. The pellets were dissolved in
loading buffer containing 1% CHAPS, and loaded onto a Ni-agarose
column. The column was washed with 5× and 3× bed volumes of
load buffer and loading buffer containing 25 mM imidazole, respec-
tively. The protein was then eluted by applying loading buffer
containing 500 mM imidazole to the column (1% CHAPS was included
in all column steps). The purity of the protein was >95%, as judged
by SDS-page analysis.
Figure 5. Relative rates of glycosyltransfer (0) and transpeptidation
(O) reactions. Conditions for analyses of the two reactions are the same
as in Figure 1 and Figure 4.
An advantage of the current assay system is that D-Ala can
also be visualized with the same labeling reaction, as fluores-
camine will react with the amino group of amino acids. As
shown in Figure 4, a time-dependent formation of D-Ala is
observed upon incubation of PBP1b with Lipid II. As mentioned
previously, penicillin G is known to inhibit the transpeptidation
reaction, by chemical modification of a key active site serine
residue.11 The addition of penicillin G to the in vitro reaction
catalyzed by PBP1b resulted in a lack of detectable D-Ala
formation (data not shown), confirming the enzymatic source
of this product. To our knowledge, this is the first direct
observation of the transpeptidation reaction, and will allow
future study of this activity.
It is interesting to compare the timecourses for the glycosyl-
transfer and transpeptidation reactions (Figure 5). While the
glycosyltransfer reaction is virtually complete in 30 min, the
transpeptidation reaction occurs on a longer time scale. Mecha-
nistically, this indicates that in vitro, the active sites of PBP1b
do not act in concert, as transpeptidation is still occurring after
completion of the glycosyltransfer reaction. Based on the
inability of penicillin G to inhibit this reaction (Figure 2), it
appears that both active sites can function independently. The
relevance of these results in vivo is not known; however, it
should be noted that bacteria possess enzymes capable of
catalyzing each of these reactions singly (monofunctional
Assay of Transglycosylase Activity. Assays consisted of 270 µL
of buffer (0.085% C10E6 PEG, 50 mM HEPES, pH 7.5), 10% DMSO,
and 2 µM Lipid II, and were conducted at 30 °C. Reactions were
(19) Wang, Q. M.; Peery, R. B.; Johnson, R. B.; Alborn, W. E.; Yeh,
W.-K.; Skatrud, P. L. J. Bacteriol. 2001, 183, 4779-4785.

11642 J. Am. Chem. Soc., Vol. 123, No. 47, 2001
Schwartz et al.
initiated by the addition of recombinant PBP1b (60nM final concentra-
tion). Aliquots (25 µL) were removed at various times and analyzed
by adjusting the pH to >9, followed by addition of fluorescamine (1
mM) and HPLC separation on an anion-exchange column (SAX1,
Supelco Co.). A linear gradient of 20 mM to 1 M ammonium acetate
(in methanol) was used as eluant. N-Acetyllysine was used as a standard
to quantitate the fluorescent signal observed. Reactions involving the
observance of monomer were carried out identically, with the exception
of added muramidase from Streptomyces globisporus (Calbiochem) at
0.02 mg/mL.
Assay of Transpeptidase Activity. Assay conditions were identical
with those of the transglycosylase assays. Analysis, however, was
carried out by using a Reverse-phase C-18 HPLC column. A linear
gradient of water (+0.1% TFA) and acetonitrile (+0.1% TFA) was
used as eluant. Authentic material was used as a standard for
identification and quanitation of product.
Isolation of Cell Wall Monomer for Mass Spectral Analysis.
Enzyme (60 nm) was preincubated in reaction buffer for 30 min with
penicillin G (100 µM). Lipid II (2 µM) was added, and after 1 h the
reaction was placed in a Slide-A-Lyzer (Pierce) with a 10 000 MW
cutoff and dialyzed against water for 4 h (3 × 200 mL). To the resultant
polymeric material was added 0.02 mg of muramidase from S.
globisporus (Calbiochem) and the mixture was allowed to react for 30
min at 30 °C. The crude reaction was spun through a microconcentrator
(Amicon) with a 10 000 MW cutoff, and the flowthrough was
lyopholized for analysis. ESI-MS for C39H66N8O20: calcd [M + H+]
967.4, found 967.4
in the addition funnel. The acid chloride solution was added dropwise
to the reaction over 15 min. The reaction was allowed to warm slowly
to 0 °C, after which it was diluted with 200 mL of CH2Cl2 and
transferred to a separatory funnel. The organic layer was washed 2
times with 150 mL of 0.01 M HCl and once with 150 mL of brine.
The organic layer was then dried over MgSO4, filtered, and concentrated
in vacuo. The crude material was purified by silica gel chromatography,
with EtOAc as eluant (Rf ) 0.47). Upon removal of EtOAc, a foamy
1
solid (compound 2) was obtained (2.8 g, 80%). H NMR (300 MHz,
CD3OD) δ 8.22 (d, 1H, J ) 8.4 Hz), 7.20-7.36 (m, 5H), 4.82 (d, 1H,
J ) 3.6 Hz), 4.56 (ABq, 2H, JAB ) 12.1 Hz, ∆ν ) 61 Hz), 4.25-4.33
(m, 3H), 4.14-4.21 (m, 3H), 3.87-3.95 (m, 1H), 3.75-3.91 (m, 1H),
3.60 (dd, 1H, J ) 10.6, 8.8 Hz), 3.46 (t, 1H, J ) 9.3 Hz), 2.04 (s, 3H),
1.85 (s, 3H), 1.35 (d, 3H, J ) 7.4 Hz), 1.33 (d, 3H, J ) 6.9 Hz),
0.94-1.00 (m, 2H), 0.00 (s, 9H). 13C NMR (75.4 MHz, CDCl3) δ
172.93, 172.75, 171.67, 170.39, 136.99, 128.57, 128.15, 128.08, 97.18,
79.94, 77.67, 70.47, 69.92, 69.77, 63.78, 63.12, 52.55, 48.08, 23.26,
20.85, 19.17, 17.73, 17.27, -1.51.
Reaction D. In a drybox, compound 2 (2.8 g, 4.7 mmol) was placed
in 25 mL of dry CH2Cl2 and stirred for 20 min with 4 Å sieves. The
solution was cooled to -40 °C, and silver triflate (3.6 g, 14.1 mmol)
was added. 1-Bromo-2-deoxy-2-N-phthalimido-3,4,6-tri-O-acetyl-D-
glucopyranoside was dissolved in 30 mL of dry CH2Cl2 and added
dropwise over 3 h via addition funnel. After addition, the reaction was
allowed to continue overnight at -40 °C, followed by slow warming
to room temperature. The crude reaction was diluted with 200 mL of
CH2Cl2 and washed 2 times with 50 mL of NaHCO3 and once with 50
mL of brine. The organic layer was dried, filtered, and concentrated in
vacuo to yield a crude yellow oil. The oil was purified by silica gel
chromatography in 3:1 toluene:acetone (Rf ) 0.17) to yield the desired
Reaction A. To a solution of protected muramic acid 1 (3.88 g,
8.23 mmol) in 60 mL of 1:1 THF:CH2Cl2 was added HAlaOTMSE
(1.5 g, 7.9 mmol). To this mixture was added DIEA (2.84 mL, 16.4
mmol), HOBt (1.14 g, 8.46 mmol), and PyBOP (4.4 g, 8.46 mmol).
The reaction was stirred under N2 for 2 h, after which solvent was
removed. The residue was redissolved in 400 mL of EtOAc and washed
3 times with 150 mL of 0.01 M HCl and once with 150 mL of brine.
The organic layer was dried over MgSO4, filtered, and concentrated in
vacuo. The crude reaction was purified by silica gel chromatography,
with 2:1 EtOAc:Hex as eluant (Rf ) 0.41), to yield a white powder
1
coupled disaccharide as a white solid (2.85 g, 60%). H NMR (300
MHz, CDCl3) δ 7.86-7.91 (m, 2H), 7.75-7.79 (m, 2H), 7.15-7.27
(m, 5H), 6.94 (d, 1H, 7.7 Hz), 6.76 (d, 1H, J ) 7.3 Hz), 5.82 (dd, 1H,
J ) 10.6, 9.1 Hz), 5.38 (d, 1H, J ) 8.5 Hz), 5.20 (t, 1H, J ) 9.7 Hz),
4.99 (d, 1H, J ) 3.7 Hz), 4.46 (ABq, 2H, JAB ) 12.4 Hz, ∆ν ) 43
Hz), 4.39-4.49 (m, 2H), 4.10-4.31 (m, 7H), 4.01-4.08 (m, 1H), 3.94
(t, 1H, J ) 9.6 Hz), 3.82-3.88 (m, 1H), 3.45-3.64 (m, 3H), 2.13 (s,
3H), 2.07 (s, 3H), 2.02 (s, 3H), 1.92 (s, 3H), 1.85 (s, 3H), 1.49 (d, 3H,
J ) 6.9 Hz), 1.45 (d, 3H, J ) 7.4 Hz), 0.97-1.02 (m, 2H), 0.03 (s,
9H). 13C NMR (100.52 MHz, DMSO-d6) δ 173.52, 172.10, 169.81,
169.60, 169.55, 169.33, 169.20, 137.37, 134.96, 128.02, 127.29, 127.08,
123.64, 96.81, 95.59, 76.29, 75.99, 75.07, 70.53, 70.05, 68.45, 68.35,
68.12, 62.61, 62.11, 61.35, 54.66, 53.54, 47.47, 22.50, 20.30, 20.26,
19.92, 18.57, 17.00, 16.52, -1.65. HRMS (FAB) for C48H63N3O19Si:
calcd [M + H+] 1014.390, found 1014.389.
1
after extensive drying (4.15 g, 81%). H NMR (300 MHz, CDCl3) δ
7.40-7.45 (m, 2H), 7.22-7.36 (m, 8H), 6.88 (d, 1H, J ) 7.3 Hz),
6.24 (d, 1H, J ) 8.8 Hz), 5.54 (s, 1H), 4.95 (d, 1H, J ) 4.0 Hz), 4.57
(ABq, 2H, JAB ) 11.7 Hz, ∆ν ) 67 Hz), 4.41 (t, 1H, J ) 7.4 Hz),
4.15-4.30 (m, 4H), 4.10 (q, 1H, J ) 6.7 Hz), 3.80-3.87 (m, 1H),
3.60-3.77 (m, 3H), 1.91 (s, 3H), 1.38 (d, 3H, J ) 6.9 Hz), 1.36 (d,
3H, J ) 6.6 Hz), 0.93-0.99 (m, 2H), 0.00 (s, 9H). 13C NMR (75.4
MHz, CDCl3) δ 172.67, 172.66, 170.51, 137.13, 136.87, 129.01, 128.63,
128.28, 128.24, 128.12, 125.94, 101.40, 97.50, 81.73, 78.20, 77.45,
70.05, 68.86, 63.77, 63.17, 53.16, 48.08, 23.34, 19.45, 17.93, 17.28,
-1.48. HRMS (FAB) for C33H46N2O9Si: calcd [M + H+] 643.3051,
found 643.3064.
Reaction E. Disaccharide (3.29 g, 3.24 mmol) was placed in a 250
mL round-bottom flask with 16 g of diaminoethylene-derivatized
Merrifield resin (ref), 4 Å sieves, and 150 mL of dry butanol. The
mixture was heated at 85 °C under dry atmosphere for 24 h, at which
time the majority of starting material was gone as judged by TLC.
After being cooled to room temperature, the reaction was filtered and
concentrated to yield a syrupy oil. The oil was redissolved in 12 mL
of pyridine, to which 12 mL of acetic anhydride was added. The reaction
was stirred at room temperature overnight, after which it was
concentrated and purified by silica gel chromatography in 4:3 toluene
acetone (Rf ) 0.26) to yield the peracetylated sugar as a white solid
Reaction B. To a 250 mL RB flask containing 4.15 g of Ala-coupled
MurNAc (6.45 mmol) was added 100 mL of methanol and 120 mg of
TsOH. The reaction was warmed to 75 °C, and after 30 min the reaction
was judged complete by the disappearance of s.m. by TLC (2:1 EtOAc:
Hex). The solvent was removed from the crude reaction, and the
resulting oil was purified by silica gel chromatography, with 5:1 EtOAc:
MeOH as eluant (Rf ) 0.62). After removal of solvent, a foamy solid
1
1
(1.58 g, 53%). H NMR (300 MHz, CDCl3) δ 7.28-7.38 (m, 52H),
was obtained (3.2 g, 89%). H NMR (300 MHz, CD3OD) δ 7.20-
7.21 (d, 1H, J ) 7.3 Hz), 6.99 (d, 1H, J ) 6.9 Hz), 6.15 (d, 1H, J )
9.5 Hz), 5.05-5.16 (m, 3H), 4.59 (ABq, 2H, JAB ) 12.3 Hz, Dn ) 42
Hz), 4.48 (t, 1H, J ) 7.3 Hz), 4.39-4.43 (m, 2H), 4.29-4.37 (m, 2H),
3.98-4.22 (m, 6H), 3.77 (d, 2H, J ) 5.2 Hz), 3.56-3.66 (m, 2H),
2.15 (s, 3H), 2.04 (s, 3H), 2.03 (s, 3H), 2.03 (s, 3H), 1.96 (s, 3H), 1.94
(s, 3H), 1.42 (d, 3H, J ) 6.3 Hz), 1.39 (d, 3H, J ) 6.6 Hz), 0.96-1.02
(m, 2H), 0.03 (s, 9H). ESI-MS for C42H63N3O18Si: calcd [M + H+]:
926.4, found 926.8.
Reaction F. To a round-bottom flask containing benzyl-protected
sugar (1.58 g, 1.71 mmol) was added 300 mg of 10% Pd/C and 50 mL
of methanol. The reaction was stirred vigorously under an atmosphere
of hydrogen until reaction was judged complete by TLC. The reaction
was then filtered through Celite and concentrated in vacuo to yield the
desired compound as a white solid (1.32 g, 93%). 1H NMR (300 MHz,
7.37 (m, 5H), 4.83 (d, 1H, J ) 3.7 Hz), 4.58 (ABq, 2H, JAB ) 12.3
Hz, ∆ν ) 72 Hz), 4.28 (q, 1H, J ) 7.3 Hz), 4.27 (q, 1H, J ) 6.7 Hz),
4.14-4.20 (m, 2H), 3.91 (dd, 1H, J ) 10.4, 3.5 Hz), 3.78 (d, 1H, J )
9.9 Hz), 3.57-3.69 (m, 3H), 3.45 (t, 1H, J ) 9.2 Hz), 1.85 (s, 3H),
1.35 (d, 3H, J ) 7.3 Hz), 1.33 (d, 3H, J ) 6.3 Hz), 0.94-1.00 (m,
2H), 0.00 (s, 9H). 13C NMR (75.4 MHz, CDCl3) δ 173.53, 173.06,
170.82, 137.11, 128.54, 128.07, 127.97, 97.00, 79.75, 77.22, 72.04,
69.95, 69.73, 63.89, 61.73, 52.66, 48.13, 23.19, 19.39, 17.57, 17.24,
-1.51. HRMS (FAB) for C26H42N2O9Si: calcd [M + H+] 555.2738,
found 555.2757.
Reaction C. Diol (3.2 g, 5.8 mmol) was dissolved in 100 mL of
dry CH2Cl2, and cooled to -30 °C in a three-neck flask with an addition
funnel. Pyridine (1.0 mL, 12.4 mmol) was added to the reaction. Acetyl
chloride (0.464 mL) was dissolved in 30 mL of dry CH2Cl2 and placed

Lipid II: Total Synthesis of the Bacterial Cell Wall Precursor
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11643
CDCl3) δ 7.43-7.46 (m, 2H), 6.27 (d, 1H, J ) 5.50 (br s, 1H)), 5.13
(t, 1H, J ) 9.4 Hz), 5.08 (t, 1H, J ) 8.8 Hz), 4.59 (q, 1H, J ) 6.6 Hz),
4.41-4.50 (m, 2H), 4.34 (dd, 1H, J ) 12.5, 4.0 Hz), 4.07-4.28 (m,
6H), 3.96-4.00 (m, 1H), 3.81-3.88 (m, 2H), 3.69 (dd, 1H, J ) 10.8,
8.7 Hz), 3.60-3.66 (m, 1H), 2.14 (s, 3H), 2.06 (s, 3H), 2.04 (s, 3H),
2.03 (s, 3H), 2.01 (s, 3H), 1.99 (s, 3H), 1.46 (d, 3H, J ) 7.3 Hz), 1.39
(d, 3H, J ) 6.6 Hz), 1.00-1.06 (m, 2H), 0.05 (s, 9H). ESI-MS for
C35H57N3O18Si: calcd [M + Na+] 858.3, found 858.6.
Reaction J. A solution of the phosphotriester from reaction I (252
mg, 0.14 mmol) in 10 mL of absolute ethanol was treated with 10%
Pd/C (100 mg) and placed under an atmosphere of H2. The mixture
was stirred 1 h, treated with anhydrous MgSO4, and filtered. The solid
was rinsed with ethanol, and the combined filtrates were treated with
triethylamine (100 µL), concentrated under reduced pressure, and
azeotroped twice with 5 mL of dry toluene to afford 240 mg (100%)
1
of the triethylammonium phosphate salt as a glassy foam. H NMR
Reaction G. To a 500 mL round-bottom flask containing the product
of reaction E (1.32 g, 1.58 mmol) was added 338 mg of tetrazole (4.725
mmol) and 50 mL of dry CH2Cl2. After the mixture was cooled to
-30 °C, dibenzyl(N,N-diisopropyl)phosphoramidite (1.07 mL, 3.1
mmol) was added dropwise via syringe, and the reaction was slowly
warmed to room temperature, followed by stirring for 1 h. The reaction
was then cooled to -40 °C, and 2 g of mCPBA was added. The reaction
was then allowed to continue for 30 min at 0 °C, and 30 min with
slow warming to room temperature. The reaction was then diluted with
200 mL of CH2Cl2, and washed successively 2 times with 50 mL of
10% sodium thiosulfate, 2 times with 50 mL of concentrated Sodium
bicarbonate, and 2 times with 50 mL of water. The organic layer was
then dried, filtered, and concentrated in vacuo. The crude material was
then purified by silica gel chromatography in 3:2 toluene:acetone
(300 MHz, DMSO-d6) δ 8.10-8.55 (m, 4H), 8.03 (d, 1H, J ) 8.4
Hz), 7.80-7.89 (m, 1H), 6.88-6.95 (m, 2H), 5.30-5.35 (m, 1H), 5.20
(t, 1H, J ) 10.5 Hz), 4.83 (t, 1H, J ) 10.6 Hz), 4.66 (d, 1H, J ) 8.4
Hz), 3.90-4.48 (m, 15H), 3.83-3.87 (m, 1H), 3.54-3.78 (m, 4H),
3.80-3.94 (m, 8H), 2.05 (s, 3H), 1.93 (s, 3H), 1.91 (s, 3H), 1.89 (s,
3H), 1.78 (s, 3H), 1.73 (s, 3H), 1.70-1.95 (m, 4H), 1.05-1.60 (m,
27H), 0.82-0.96 (m, 6H), -0.010 (s, 18H), -0.023 (s, 9H). HRMS
(FAB) for C63H111N8O29PSi3 + Na+: calcd [M + Na+] 1581.640, found
1581.641.
Reaction K. Under dry argon, a solution of the phosphate salt from
reaction J (75 mg, 0.045 mmol) in 0.4 mL of DMF was treated with
carbonyldiimidazole (38 mg, 0.23 mmol). The solution was stirred 4 h
at ambient temperature, treated with anhydrous MeOH (17 µL, 0.42
mmol), and stirred an additional 0.5 h. To this solution was added a
solution of undecaprenyl phosphate triethylammonium salt (38 mg,
0.036 mmol) in 1.5 mL of CH2Cl2. The volume of the solution was
reduced to 1 mL under a stream of dry argon, and the reaction was
stirred for 48 h at ambient temperature. The mixture was concentrated
under reduced pressure and purified by anion exchange chromatography
(Bio-Rad High Q, elution with NH4OAc in MeOH). The appropriate
fractions were concentrated under reduced pressure and lyophilized
twice from water-MeOH to afford 34 mg (39%) of the pyrophosphate
1
(Rf ) 0.18) to yield compound 5 as a foamy solid (0.95 g, 55%). H
NMR (300 MHz, CDCl3) δ 7.42 (d, 1H, J ) 5.9 Hz), 7.33-7.36 (m,
10H), 7.19 (d, 1H, J ) 7.3 Hz), 5.95-5.99 (m, 2H), 5.02-5.17 (m,
6H), 4.46-4.55 (m, 2H), 4.40 (d, 1H J ) 8.5 Hz), 4.30 (td, 1H, J )
13.3, 3.8 Hz), 4.04-4.22 (m, 6H), 3.82-3.99 (m, 3H), 3.51-3.61 (m,
2H), 2.07 (s, 3H), 2.06 (s, 3H), 2.04 (s, 3H), 2.02 (s, 3H), 1.97 (s, 3H),
1.83 (s, 3H), 1.43 (d, 3H, J ) 7.4 Hz), 1.40 (d, 3H, J ) 7.0 Hz),
0.84-0.95 (m, 4H), 0.01 (s, 9H). 13C NMR (150.8 MHz, CDCl3) δ
173.87, 172.54, 171.06, 170.99, 170.75, 170.48, 169.24, 135.65, 128.64,
128.59, 127.92, 127.88, 100.04, 95.85, 95.81, 76.87, 75.05, 74.27, 72.45,
72.13, 71.13, 69.59, 69.55, 69.45, 69.42, 68.12, 63.77, 61.89, 61.60,
54.53, 53.66, 53.61, 48.17, 23.21, 22.83, 20.82, 20.61, 20.57, 20.54,
18.77, 17.88, 17.19, -1.56. HRMS (FAB) for C49H70N3O21PSi: calcd
[M + Na+] 1118.390, found 1118.388.
1
as a white powder. H NMR (300 MHz, CD3OD) δ 5.31-5.43 (m,
3H), 5.27 (t, 1H, J ) 9.9 Hz), 4.99-5.15 (m, 13H), 4.68-4.90 (m,
2H), 4.41-4.61 (m, 3H), 4.23-4.39 (m, 5H), 4.01-4.19 (m, 8H), 3.63-
3.90 (m, 4H), 3.04 (t, 2H, J ) 6.5 Hz), 2.27 (t, 2H, J ) 5.7 Hz), 1.79-
2.22 (m, 60H), 1.50-1.76 (m, 42H), 1.30-1.45 (m, 12H), 0.87-1.02
(m, 6H), 0.00 (s, 18H), -0.01 (s, 9H). ESI-MS for C118H200N8O32P2-
Si3: calcd [M - 2H2-] 1192.6, found 1193.4.
Reaction H. In a 100 mL round-bottom flask was added compound
5 (770 mg, 0.70 mmol) in 20 mL of THF. Tetrabutylammonium fluoride
solution (2.6 mL, 1 M in THF) was added, and the reaction was
followed by TLC. After 45 min, THF was removed in vacuo, and the
crude reaction was redissolved in 200 mL of EtOAc. After extracting
the EtOAc solution 2 times with 20 mL of 1 M HCl, the organic layer
was dried, filtered, and concentrated to yield a clear oil. The oil was
azeotroped 3 times with 2 mL of toluene to yield the desired deprotected
compound as a white solid (610 mg, 87%). 1H NMR (300 MHz, CDCl3)
δ 7.85 (d, 1H, J ) 7.4 Hz), 7.80 (d, 1H, J ) 3.6 Hz), 7.25-7.39 (m,
10H), 6.18 (d, 1H, J ) 9.9 Hz), 5.99 (dd, 1H, J ) 5.9, 2.5 Hz), 4.96-
5.17 (m, 6H), 4.75 (q, 1H, J ) 6.6 Hz), 4.43-4.50 (m, 2H), 4.06-
4.35 (m, 5H), 3.96 (dd, 1H, J ) 9.9, 8.4 Hz), 3.86-3.90 (m, 1H),
3.60-3.79 (m, 3H), 2.07 (s, 3H), 2.05 (s, 3H), 2.04 (s, 3H), 2.02 (s,
3H), 1.99 (s, 3H), 1.84 (s, 3H), 1.52 (d, 3H, J ) 7.4 Hz), 1.36 (d, 3H,
J ) 6.6 Hz). ESI-MS for C44H58N3O21P: calcd [M + Na+] 1018.3,
found 1018.5.
Reaction L. A solution of the pyrophosphate from reaction K (21
mg, 0.0088 mmol) in 2 mL of DMF was treated with a 1 M solution
of tetra(n-butyl)ammonium fluoride in THF (1 mL, 1.0 mmol). The
solution was stirred 24 h, concentrated under reduced pressure, and
purified by anion exchange chromatography (Bio-Rad High Q, elution
with NH4OAc in MeOH). The appropriate fractions were concentrated
under reduced pressure and lophilized twice from water-MeOH to
afford a mixture of tetra- and triacetates. The mixture was dissolved in
1 mL of CH2Cl2 and treated with 0.5 M NaOMe in MeOH (0.16 mL,
0.08 mmol). The solution was stirred 1 h, treated with 0.16 mL more
NaOMe in MeOH, and stirred 1 h longer. The mixture was concentrated
under reduced pressure and purified by anion exchange chromatography
(Bio-Rad High Q, elution with NH4OAc in MeOH). The appropriate
fractions were concentrated under reduced pressure and lyophilized
twice from water-MeOH to afford 6 mg (35%) of Lipid II tris-
1
Reaction I. In a 25 mL round-bottom flask was added the
phosphodisaccharide from reaction H (70 mg, 0.092 mmol) in 5 mL
of a 1:1 THF:CH2Cl2 solution. Protected tetrapeptide (139 mg, 0.14
mmol) (purchased from New England Peptide, Inc., Fitchburg, MA)
was added, followed by diisopropylethylamine (48.4 µL, 0.28 mmol),
HOBt (20.3 mg, 0.15 mmol), and finally PyBOP (78 mg, 0.15 mmol).
The reaction was stirred for 3 h, after which solvent was removed and
the crude residue was redissolved in 30 mL of CH2Cl2 and transferred
to a separatory funnel. The organic layer was washed 3 times with 5
mL of dilute HCl, dried over magnesium sulfate, filtered, and
concentrated in vacuo. The resulting oil was purified by silica gel
chromatography in 9:1 EtOAc:MeOH (Rf ) 0.43) to yield compound
(ammonium) salt as a white powder. H NMR (300 MHz, CD3OD) δ
5.48-5.56 (m, 1H), 5.42 (t, 1H, J ) 7.5 Hz), 5.03-5.19 (m, 13H),
4.43-4.56 (m, 3H), 4.16-4.42 (m, 7H), 4.04-4.12 (m, 1H), 3.79-
3.94 (m, 3H), 3.59-3.75 (m, 3H), 3.48 (dd, 1H, J ) 10.2, 8.1 Hz),
2.93 (t, 2H, J ) 6.0 Hz), 1.91-2.38 (m, 48H), 1.48-1.89 (m, 44H),
1.33-1.45 (m, 12H). HRMS (FAB) for C94H156N8O26P2: calcd [M -
2H2-] 936.5225, found 936.5216.
Acknowledgment. The authors wish to thank Dr. Ross Stein
for advice on enzymology and Dr. Subramanain Sabesan, Dr.
Steven Seitz, and Professor Dale Drueckhammer for advice on
carbohydrate chemistry. We also wish to thank Dr. Nilsa
Graciani and Chun Qi Zhu for preliminary preparations of
peptide and Dr. Maria Barbosa and Dr. Mike Kurilla for the
gift of flavomycin.
1
6 as a white solid (98 mg, 61%). H NMR: (300 MHz, CD3OD) δ
7.28-7.35 (m, 10H), 5.76 (dd, 1H, J ) 6.1, 3.1 Hz), 5.21 (t, 1H, J )
9.7 Hz), 3.45-4.67 (m, 23H), 2.98-3.07 (m, 2H), 2.16-2.32 (m, 2H),
1.57-2.06 (m, 24H), 1.23-1.47 (m, 16H), 0.78-1.01 (m, 6H), 0.00
(s, 9H), -0.01 (s, 9H), -0.03 (s, 9H). HRMS (FAB) for C77H123N8O29-
PSi3: calcd [M + H+] 1739.752, found 1739.750.
JA0166848