Inhibition of the antibacterial target UDP-(3-O-acyl)-N-acetylglucosamine deacetylase (LpxC): Isoxazoline zinc amidase inhibitors bearing diverse metal binding groups

Article abstract of DOI:10.1021/jm020183v

UDP-3-O-[R-3-hydroxymyristoyl]-GlcNAc deacetylase (LpxC) is a zinc amidase that catalyzes the second step of lipid A biosynthesis in Gram negative bacteria. Known inhibitors of this enzyme are oxazolines incorporating a hydroxamic acid at the 4-position,

Full text of DOI:10.1021/jm020183v

J . Med. Chem. 2002, 45, 4359-4370
4359
In h ibition of th e An tiba cter ia l Ta r get UDP -(3-O-a cyl)-N-a cetylglu cosa m in e
Dea cetyla se (Lp xC): Isoxa zolin e Zin c Am id a se In h ibitor s Bea r in g Diver se Meta l
Bin d in g Gr ou p s
Michael C. Pirrung,*,† L. Nathan Tumey,† Christian R. H. Raetz,‡ J ane E. J ackman,‡ Karnem Snehalatha,‡
Amanda L. McClerren,‡ Carol A. Fierke,§ Stephanie L. Gantt,§ and Kristin M. Rusche§
Department of Chemistry, Levine Science Research Center, Box 90317, Duke University, Durham, North Carolina 27708-0317,
Department of Biochemistry, Box 3711, Duke University Medical Center, Durham, North Carolina 27710, and
Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109-1055
Received April 30, 2002
UDP-3-O-[R-3-hydroxymyristoyl]-GlcNAc deacetylase (LpxC) is a zinc amidase that catalyzes
the second step of lipid A biosynthesis in Gram negative bacteria. Known inhibitors of this
enzyme are oxazolines incorporating a hydroxamic acid at the 4-position, which is believed to
coordinate to the single essential zinc ion. A new structural class of inhibitors was designed to
incorporate a more stable and more synthetically versatile isoxazoline core. The synthetic
versatility of the isoxazoline allowed for a broad study of metal binding groups. Nine of 17
isoxazolines, each incorporating a different potential metal binding functional group, were found
to exhibit enzyme inhibitory activity, including one that is more active than the corresponding
hydroxamic acid. Additionally, a designed affinity label inhibits LpxC in a time-dependent
manner.
In tr od u ction
Strains of three bacteria (Enterococcus faecalis, Myco-
bacterium tuberculosis, and Pseudomonas aeruginosa)
have recently shown resistance to every clinically avail-
able antibioticsincluding vancomycin, often considered
a drug of “last resort”.1 Although many groups have
devised creative ways to get around resistance prob-
lems,2,3 the most effective long-term solution is the
development of drugs that act on unexploited antibacte-
rial targets.4 One such target in Gram negative bacteria
is the biosynthesis of lipid A, a key component of
the outer membrane. The minimal lipid A structure
necessary for bacterial cell growth consists of a diphos-
phorylated tetrasaccharide acylated with six fatty acid
chains on the two glucosamine subunits (Figure 1).5 The
biosynthesis of this minimal lipid A structure has been
extensively studied; all of the enzymes and correspond-
ing genes responsible for its synthesis in Escherichia
coli have been identified.6 Homologous genes have also
been identified from the genomes of most other Gram
negative bacteria.
Over the last two decades, several steps of lipid A
biosynthesis have been investigated as potential anti-
bacterial targets. Early attempts to inhibit the produc-
tion of lipid A primarily focused on the biosynthesis of
2-keto-3-deoxy-D-manno-octulosonic acid (KDO), a key
F igu r e 1. Biosynthesis and structure of lipid A.
sugar residue in the minimal lipid A structure (Figure
1).7,8 Although relatively potent mechanism-based in-
hibitors have been found for two enzymes in this
pathway, none of these inhibitors has proved therapeu-
tically useful.9
Recently, a series of small heterocyclic hydroxamic
acids that inhibit lipid A biosynthesis and are antibac-
terial against E. coli both in vitro and in animal models
have been disclosed.10 These compounds are aryl ox-
azoline hydroxamic acids (Figure 2), and the most potent
compound of the series, L-161,240, has a minimal
inhibitory concentration (MIC) against E. coli compa-
rable to ampicillin and rifampicin (∼2 µg/mL). The
target of these small molecules has been shown to be
* To whom correspondence should be addressed. Tel: 919-681-3482.
Fax: 919-660-1591. E-mail: pirrung@chem.duke.edu.
† Duke University.
‡ Duke University Medical Center.
§ University of Michigan.
10.1021/jm020183v CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/20/2002

4360 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19
Pirrung et al.
F igu r e 3. Retrosynthesis of oxazolines (a) and isoxazolines
(b).
Sch em e 1a
F igu r e 2. Structure of two of Merck’s deacetylase inhibitors.
UDP-3-O-[R-3-hydroxymyristoyl]-GlcNAc deacetylase
(LpxC), a zinc amidase encoded by the lpxC gene.
Unfortunately, the antibacterial spectrum of this series
of compounds is rather limited although a very wide
variety of Gram negative bacteria have been shown
to express homologous LpxCs. Although Enterobacter
cloacae and Klebsiella pneumoniae are sensitive to
L-161,240, the growth of P. aeruginosa and Serratia
marescens is not affected.10 This lack of activity against
other Gram negative bacteria is due, at least in part,
to subtle differences in enzyme structure between the
various organisms.11
a
Reagents: (a) (1) TMBNHODMB, N-methylmorpholine; (2)
Compound 3, Et3N. (b) TFA/Et3SiH/DCM. DMB ) 2,4-Dimethoxy-
benzyl. TMB ) 2,4,6-Trimethoxybenzyl.
Resu lts
LpxC catalyzes the first committed step of lipid A
biosynthesis, the removal of an N-acetyl group from a
3-acyl-N-acetyl-glucosamine moiety (Figure 1). This
enzyme contains a single catalytic zinc ion that is
coordinated by the nitrogen of two histidine residues,
with a third coordinating group that is either a histidine
nitrogen or a carboxylate of Asp or Glu.12 The three-
dimensional structure of this deacetylase is currently
unknown, although efforts to determine it are under-
way.13 No sequence homology with other structurally
characterized zinc amidases has been identified.
One problem in targeting analogues of the original
lead oxazoline hydroxamate bearing various metal
binding groups is the difficulty in obtaining the ap-
propriate serine analogues to serve as precursors (Fig-
ure 3). Although a few of these compounds have been
reported, their synthesis is tedious and therefore would
greatly limit the number of metal binding groups that
could be studied. For instance, diethyl phosphonoserine
is a known compound but is prepared in a six step
synthesis from commercially available starting materi-
als.18 Further, oxazolines are sensitive to nucleophiles
such as thiols that would be of great interest to
incorporate into the inhibitors.19
One requisite feature of these LpxC inhibitors is the
hydroxamic acid moiety. As shown in the earlier studies,
the conversion of the hydroxamic acid to a carboxylic
acid results in the loss of all inhibitory and antibacterial
activity. The well-known affinity of hydroxamic acids
for zinc has led to the proposal that the inhibitors act
by coordinating to the single catalytic zinc ion.14 Thus,
it is thought that the inhibitors function analogously
to matrix metalloproteinase (MMP) inhibitors.
A potential solution to these problems was to change
the heterocyclic scaffold to increase its stability toward
nucleophiles and to allow the facile incorporation of a
wide variety of groups in place of the hydroxamic acid.
One class of heterocycles that fits these criteria is 4,5-
dihydro-isoxazoles (Figure 3). These isoxazolines are
quite stable toward nucleophiles such as thiols and are
easily derived from 1,3-dipolar cycloaddition with an
acrylate or a substituted ethylene.20 Ethylenes substi-
tuted with an electron-withdrawing group generally
react to give a single regioisomer (as shown in Figure
3).21 The corresponding alkenes containing the various
metal binding groups are widely available or easily
synthesized.
A feature of MMP inhibitors that has been studied
in great detail is the zinc coordinating group. Although
hydroxamic acids are frequently employed as metal
binding groups, they often present metabolic and
pharmokinetic problems such as glucoronidation and
sulfation that result in a short in vivo half-life.15,16 Many
hydroxamates are unstable in vivo, being hydrolyzed to
give the toxin hydroxylamine,17 which has fueled the
search for other zinc coordinating functional groups. A
variety of metal binding groups distinct from hydrox-
amic acids have been shown effective in in vitro and in
vivo inhibition of zinc peptidases, including thiols,
â-hydroxy thiols, phosphinic acids, and heterocyclic
amines.17 With these precedents in mind, we undertook
a systematic study of metal binding groups for LpxC
inhibitors, as it applies to both enzymatic and antibac-
terial activity. These data should be of value in future
studies of both lipid A biosynthesis inhibitors and zinc
protease inhibitors.
Before large numbers of isoxazolines were made, we
determined whether the change of heterocycle from an
oxazoline to an isoxazoline would affect activity. A
slightly simplified analogue of the most potent known
inhibitor was therefore designed. Racemic hydroxamate
2 was synthesized according to Scheme 1 beginning with
the previously reported doubly protected hydroxylamine
shown.22 Cycloaddition with nitrile oxide precursor 3
gave isoxazoline 1 as a single regioisomer. Acidic
deprotection gave the expected hydroxamate product,
which exhibited the desired biological activity (Table 1).
The two enantiomers of this compound were made

Isoxazoline Zinc Amidase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19 4361
Ta ble 1. Enzymatic and Antibacterial Activity of the Various Inhibitors
a
The IC50 of this compound could not be determined accurately. See text.
according to Scheme 2 from the previously reported
enantiomeric isoxazolines.23,24
carboxylic acid was converted in an Arndt-Eistert
sequence to the acid chloride and then treated with
diazomethane to give the corresponding R-diazoketone.
Conversion to the R-chloroketone with anhydrous HCl
and displacement with thioacetic acid gave thioester 7.
Removal of the acetate with hydrazine immediately
gave an inseparable 1:1 mixture of the desired thiol 8
and its dimer (Figure 4). Removal of the acetate under
All of other compounds containing the various metal
binding groups were prepared as racemates and com-
pared to the activity of the racemic hydroxamate 2. An
R-thiol-ketone isoxazoline was synthesized from methyl
acrylate according to Scheme 3. Dipolar cycloaddition
followed by saponification gave carboxylic acid 6. The

4362 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19
Pirrung et al.
Sch em e 2a
Sch em e 4a
a
Reagents: (a) EDC, Y-NH2.
a
Reagents: (a) J ones oxidation. (b) (1) EDC, NH2OTr; (2) TFA/
Sch em e 5a
DCM.
Sch em e 3a
a
a
Reagents: (a) Compound 3, Et3N. (b) Me3SiBr.
Reagents: (a) (1) Compound 3, N-methylmorpholine; (2)
NaOH. (b) (1) SOCl2; (2) CH2N2; (3) HCl; (4) CH3COSH, Et3N. (c)
N2H4.
Sch em e 6a
a
Reagents: (a) (1) 2-Chloroethanol, pyridine; (2) 160 °C. (b)
F igu r e 4. Compound 8 was isolated as an inseparable
mixture of these two compounds.
Et3N, 120 °C. (c) Compound 3, Et3N. (d) Me3SiBr.
Sch em e 7a
other conditions gave little or no product. In retrospect,
the formation of the dimer is not surprising given that
the carbonyl is quite electrophilic (being doubly substi-
tuted with R-electron-withdrawing groups) and dimer-
ization results in the formation of a stable six-membered
ring.
Carboxylic acid 6 could easily be converted to deriva-
tives 9-12 via a carbodiimide-mediated coupling reac-
tion as shown in Scheme 4. A dipolar cycloaddition
reaction with readily available alkenes gave derivatives
13-15 (Scheme 5). The cycloaddition reactions are quite
clean and give a single regioisomer in each example.
Phosphonic acid 17 was synthesized in two steps begin-
ning with commercially available diethyl vinyl phos-
phonate according to Scheme 5. Again, the cycloaddition
reaction gives a single regioisomer.
Several unsuccessful attempts were made to synthe-
size methyl vinyl phosphinate esters according to lit-
erature procedures. Eventually, the three step synthesis
shown in Scheme 6 was adopted. Methyl dichlorophos-
phine was treated with 2-chloroethanol in the presence
of pyridine to form the corresponding phosphine, which
was concentrated and heated to 160 °C (neat). The
resultant oil was heated (neat) with Et3N to form alkene
19, which was purified by Kugelrohr distillation. This
represents a very efficient new synthesis of methyl vinyl
a
Reagents: (a) Compound 3, Et3N. (b) NaOH.
phosphinate esters, a very useful class of compounds
in natural product chemistry.25 Dipolar cycloaddition
followed by deprotection gave phosphinic acid 21.
Thiol 22 (Scheme 7) was synthesized from the com-
mercially available allyl thiopropionate. As may be
expected, examination of the crude NMR spectrum of
the cycloaddition reaction showed that a significant
amount of the undesired regioisomer formed. However,
the two regioisomers could be easily separated by flash
chromatography. Finally, compounds 25 and 26 (Scheme
8) were synthesized from the known diastereomeric
epoxides 24, which, in turn, were derived from butadi-
ene monoepoxide.26 Epoxides 24 were converted to

Isoxazoline Zinc Amidase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19 4363
Sch em e 8a
F igu r e 5. Pro-drug approaches for improving the antibacte-
rial activity of compound 17 (RdCH2CH2OH).
F igu r e 6. Structural superposition of energy-minimized
structures of oxazoline and isoxazoline LpxC inhibitors.
a
Reagents: (a) Compound 3, Et3N. (b) Thiourea. (c) (Me3Si)2S,
TBAF.
enzyme. Even though in this case the inhibition is weak,
the formyl hydrazine group present in 10 represents a
previously unreported metal binding functional group.
More interestingly, thiols 8 and 22 were found to have
low micromolar inhibition (27 and 22 µM) of LpxC,
nearly comparable to hydroxamate 2. Their potency is
predicted to be even greater at physiological pH because
the thiol will be more ionized. These assays were
performed at pH 6.0. It is interesting that thiol 22 is
nearly as active as the hydroxamate even though it can
function only as a monodentate zinc chelator.31 This is
commonly seen; it has been suggested that this observa-
tion is due to easier ionization of the thiol as compared
to hydroxamates and a lower desolvation penalty.32
Given the activity of thiol 22, it is interesting that
R-hydroxy thiols 26a ,b do not inhibit LpxC. R-Hydroxy
thiols have made excellent zinc binding groups in
various MMP inhibitors, often exhibiting equivalent or
improved activity as compared to the corresponding
hydroxamate.33 Excitingly, phosphonic acid 17 has an
IC50 of 4 µM, a considerable improvement over the
corresponding hydroxamic acid. This is highly un-
usual: phosphonates typically inhibit zinc amidases
10-100-fold less effectively than hydroxamates. In
earlier work,11 a carbohydrate substrate analogue with
a hydroxamate replacing the acetamide was about 3-fold
more potent vs the E. coli enzyme than an analogous
carbohydrate phosphinate. This compound is much more
potent vs the Aquifex aeolicus enzyme.
Thioacetate 7 unexpectedly proved to have significant
enzymatic activity, although this activity was difficult
to characterize. The IC50 of this compound was ap-
proximately 12 µM, comparable to the corresponding
hydroxamate. However, it is questionable how mean-
ingful this value is. The activity of the enzyme dropped
by 85% upon addition of 36 µM inhibitor. However, 15%
residual activity remained upon addition of up to 350
µM inhibitor. Moreover, the compound appeared to show
time-dependent inactivation of the enzyme. One expla-
nation for the peculiar inhibitory activity of this com-
pound is that it may be a substrate for the enzyme.
thiiranes 25 with inversion of stereochemistry via a
known transformation utilizing thiourea.27 Epoxides 24
could also be opened with a hydrogen sulfide deriva-
tive28 to give diastereomeric thiol-alcohols 26.
Table 1 gives a complete summary of the inhibitory
and antibacterial properties of the compounds prepared
above. Racemic hydroxamate 2 was found to have an
IC50 of 13 µM. Although the compound had no effect on
the growth of wild type E. coli, it did significantly inhibit
the growth of EnvA1 E. coli, a strain of bacteria with a
mutation in lipid A biosynthesis.10 As expected, the
active enantiomer of 2 was found to have the (S)
configuration. This corresponds to the same spatial
relationship of groups as the biologically active (R)-
oxazoline L-159,692. Interestingly, the active enanti-
omer of 2 has an IC50 very similar to that of L-159,692.
Despite the equivalent inhibitory activity, the (S) isomer
of 2 has reduced antibacterial activity as compared to
that of its oxazoline counterpart. Though (S)-2 has no
effect on the growth of wild-type E. coli, the oxazoline
(L-159,692) and isoxazoline (2) have a nearly equivalent
inhibitory effect on the growth of EnvA1 E. coli. The
MIC of (S)-2 against EnvA1 E. coli is 3.9 µg/mL, which
is comparable to the MIC of the original lead compound
(3.6 µg/mL) against the same strain of bacteria.29
The nearly equivalent inhibitory activity of the two
heterocycles (2 and L-159,692) supports the hypothesis
that the heterocycle acts as a scaffold that holds the
metal binding group and the hydrophobic aryl group in
a particular spatial orientation. The two heterocycles
have different patterns of hydrogen bond acceptors; such
properties therefore seem to be less important factors
in inhibition. Overlaying the energy-minimized confor-
mations of these two compounds shows that the posi-
tions of the hydroxamate carbons differ by only 1.1 Å
(Figure 6).30
Several isoxazolines with metal coordinating groups
other than hydroxamic acid show significant inhibition
of LpxC enzymatic activity. Phosphinic acid 21 and
formyl hydrazine 10 are weakly active against the

4364 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19
Pirrung et al.
demonstrated that 25a is also a time-dependent inhibi-
tor with a KI of 0.8 ( 0.2 mM and kinact of 4 ( 2 M-1
min-1. While these data clearly demonstrate time-
dependent inhibition, they do not demonstrate active
site modification. Further studies are underway to
clarify the mechanism of this inhibition. Given the
dearth of structural and mechanistic knowledge of
LpxC, these inhibitors may provide important clues
about the active site amino acids.
Despite their similar or even improved enzymatic
potency as compared to hydroxamate 2, the compounds
with the various metal binding groups described above
displayed no antibacterial activity against wild-type or
hypersensitive EnvA1 E. coli (Table 1). This was not
entirely unexpected given the difficulty in finding
compounds that will cross the two membranes of Gram
negative bacteria. Phosphonate 17, in particular, would
be very unlikely to cross the hydrophobic barriers
presented by the bacteria. Various prodrug approaches
(Figure 5) were attempted in order to “mask” the polar
phosphonate group as a more hydrophobic phosphonate
ester that could be cleaved in vivo to form compound
17, but these attempts were unsuccessful.36 Phosphate
prodrugs are widely utilized to deliver phosphates
across mammalian cell membranes,37 but there has been
no reported prodrug system for delivering such com-
pounds across bacterial cell membranes. Compound 7,
along with its good enzymatic activity, has the polarity
and solubility profile necessary to cross the double
membrane. Presumably, its lack of antibacterial activity
is due to either active extrusion or fast metabolism of
this compound.
F igu r e 7. Time-dependent inhibition of LpxC. LpxC (100 µM)
was incubated with 0.5 mM 25b at 30 °C in 100 mM bis-Tris,
pH 7, 12% DMF. At the indicated times, an aliquot was
removed and diluted 10-fold into assay buffer. The initial
velocity was assayed as described in the Experimental Section.
The data are well-described by a single-exponential decay with
an apparent first-order rate constant of 0.020 ( 0.001 min-1
.
Inset: The observed rate constants for inhibition show a
hyperbolic dependence on the inhibitor concentration. The data
are well-described by eq 2 with a second-order rate constant
of 90 ( 30 M-1 min-1 and KI of 0.3 ( 0.1 mM.
Control experiments establish that 7 is stable in assay
buffer alone over several hours. If the enzyme can
hydrolyze 7, the active inhibitor may be thiol 8. How-
ever, this cannot explain the time-dependent inactiva-
tion of the enzyme. Thioesters have been used to acylate
amines under physiological conditions.34 Therefore, it
may be possible that this functional group is acetylating
an enzyme amine or thiol and thereby rendering the
enzyme inactive. It is also conceivable that it is a slow,
tight binding inhibitor, though we consider this pos-
sibility less likely. Although there have been reports of
similar functional groups being used as thiol “prodrugs”
in vivo, this work represents the first report of this novel
zinc binding group directly having enzymatic inhibitory
activity.
In addition to the competitive zinc binding groups
described above, two potential affinity labels were also
investigated. These inhibitors incorporate thiiranes and
epoxides, functional groups that have been previously
reported to covalently modify the active sites of other
zinc amidases, rendering them inactive.27,35 They are
believed to act through electrophilic activation of the
highly strained heterocycle by the active site zinc ion.
Subsequent attack by an active site nucleophile (be-
lieved to be Glu-404 in MMPs) gives a covalently
modified enzyme that is incapable of turnover.
The time-dependent inhibition of 25a ,b was investi-
gated by incubating LpxC with excess inhibitor. At
various times, an aliquot was removed and diluted 10-
fold into assay buffer to measure the remaining enzyme
activity. Compound 25b clearly inhibits LpxC in a time-
dependent manner (Figure 7). The apparent first-order
rate constant for inactivation is dependent on the
concentration of inhibitor, yielding a second-order rate
constant (kinact) of 90 ( 30 M-1 min-1 and a KI of 0.3 (
0.1 mM. This inactivation extrapolates to 100% inhibi-
tion at long incubation times. Furthermore, the inhibi-
tion is essentially irreversible; when the inactivated
enzyme is diluted 20-fold into buffer, no increase in
activity is observed over 45 h. Similar experiments
Discu ssion
Using an isoxazoline heterocycle as a scaffold for
groups that could potentially coordinate the active site
zinc ion in LpxC, a number of novel inhibitors have been
prepared. Our inhibition data in Table 1 suggest the
following preference for zinc coordination: hydroxamate
≈ phosphonate ≈ thioacetoxyacetyl > thiolacetyl >
phosphinate > carboxylate, sulfonamide, and hydrazide.
This ranking is similar to that determined in a study
of MMP-1 (fibroblast collagenase) inhibitors38 where the
preference for the zinc coordinating group was observed
as follows: hydroxamate . formyl hydroxylamine >
thiol > phosphinate > R-amino acrylate > carboxylate.
This study is the only reported systematic study of
changes of the metal binding group in zinc proteinase
inhibitors while holding the rest of the inhibitor con-
stant. Although the rankings are similar, our studies
identified two additional metal binding groups with
inhibitory activity comparable to hydroxamates.
The specific ranking of the zinc binding groups
presented here may not be general for this enzyme.
Slight changes in the metal binding geometry are likely
to translate into slight changes in the conformation of
the heterocyclic scaffold or slightly alter hydrogen bond
interactions with the heterocycle. This, in turn, may
lead to slight changes in the way the aromatic group
fits into a hydrophobic pocket of the enzyme. Therefore,
a true quantitative measure of the effectiveness of the
metal binding groups would require an optimization of
the aromatic group and/or heterocycle for each zinc
binding group tested.17 Even if the quantitative results

Isoxazoline Zinc Amidase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19 4365
presented here do not represent a general trend, they
clearly represent a starting point for the design of new
LpxC inhibitors with different structure/activity profiles
and potentially increased or modified antibacterial
activity. The four new zinc binding groups that were
identified will hopefully serve as new lead structures
from which more effective inhibitors can be developed.
Moreover, two previously unknown zinc binding groups
were identified, which may be useful in the design of
inhibitors for the myriad of other therapeutically im-
portant metalloamidases.
groups apart from hydroxamic acids that are effective
in the inhibition of a variety of zinc enzymes, though it
is unlikely that the ranking of zinc binding groups
presented here will apply directly to other zinc ami-
dases.
Exp er im en ta l Section
En zym e Assa ys a n d An tiba cter ia l Testin g. Bu ffer s
a n d Rea gen ts. [R-32P]UTP was purchased from NEN Dupont.
PEI-Cellulose thin-layer chromatography (TLC) plates were
obtained from E. Merck, Darmstadt, Germany. Bis-tris buffer
(Ultrapure reagent) and bovine serum albumin (BSA) were
purchased from Sigma. Dimethyl sulfoxide (DMSO) was
purchased from Mallinckrodt.
Both oxazolines and isoxazolines tend to exist in an
“envelope” conformation in which one carbon is slightly
out of plane from the four remaining atoms (Figure 6).
MM2 calculations and literature reports show that the
C-5 position is “puckered” in both heterocycles by
approximately 15°.39,40 Importantly, the metal binding
group is bound to C-5 of the isoxazoline (the puckered
carbon) but to C-4 of the oxazoline (a planar carbon).
Therefore, it appears that the oxazoline provides a
slightly more rigid scaffold for the metal binding group
than does the isoxazoline. Though this does not seem
to significantly affect the IC50 of the isoxazoline 2, this
may have the implication of changing the structure-
activity relationship (SAR) of the aromatic ring as
compared to the known inhibitors when more potent
compounds are tested. In fact, preliminary studies of
the SAR of the isoxazoline hydroxamates as compared
to the oxazoline hydroxamates indicate moderate di-
vergences in activity when compounds in the mid-
nanomolar range are tested.41 The relative flexibility of
C-5 of the isoxazoline may contribute to the wide range
of metal binding groups that are effective at this
position.
Lp xC Activity Assa y. The E. coli LpxC substrate, [R-32P]-
UDP-3-O-(R-3-hydroxymyristoyl)GlcNAc was prepared and
purified as described previously by acylation of [R-32P]-UDP-
GlcNAc using purified E. coli LpxA (provided by T. J . O.
Wyckoff, Duke University). These assays were performed at
30 °C and contained 3 µM UDP-3-O-(R-3-hydroxymyristoyl)-
GlcNAc, 1 mg/mL BSA in 40 mM bis-tris, pH 6.0. The activity
assays were performed in plastic microcentrifuge tubes in a
reaction volume of 20 µL. At each time point (chosen so that
the total conversion to product was less than 10%), 5 µL
portions of each reaction mixture were removed and added to
1 µL of 1.25 M NaOH to stop the reaction. The alkaline
samples were incubated for an additional 10 min at 30 °C to
ensure complete hydrolysis of the ester-linked acyl chains from
the LpxC substrate and product and then were neutralized
by the addition of 1 µL of 1.25 M acetic acid and 1 µL of 5%
trichloroacetic acid. The neutralized samples were incubated
on ice for 5 min and centrifuged for 2 min in a microcentrifuge.
Portions of the supernatants (2 µL) were spotted onto PEI-
cellulose TLC plates for separation of the remaining substrate
(detected as [R-32P]UDP-GlcNAc) from the product (detected
as [R-32P]UDP-GlcN). After they were air-dried, the plates were
soaked for 10 min in methanol to improve resolution before
chromatography. The plates were developed with 0.2 M
aqueous guanidine-HCl as the solvent system. The radioactive
spots on the plates were analyzed using a PhosphorImager
equipped with ImageQuant software (Molecular Dynamics,
Inc.) to determine the yields of product produced in each
reaction mixture.
In h ibition of Lp xC Activity. Purified E. coli LpxC was
used to characterize inhibitors. Stock solutions (10 mg/mL) of
each inhibitor were made in 100% DMSO, and any further
dilutions of these compounds were also made in 100% DMSO.
E. coli LpxC (2.5 nM) was incubated with varied concentra-
tions (0.1-500 µg/mL) of each inhibitor in the presence of 1
mg/mL BSA for 30 min on ice prior to starting the assays. The
final assay is 20% in DMSO. The LpxC/inhibitor stocks were
then diluted 2.5-fold into assay mixtures containing 3 µM
UDP-3-O-acyl-GlcNAc in 40 mM bis-tris, pH 6.0, at 30 °C, to
give a final enzyme concentration of 1 nM E. coli LpxC. Assay
tubes were supplemented with each of the inhibitors and
additional DMSO, if necessary, to maintain the same concen-
tration of the inhibitor and DMSO in the assay as was present
in the preincubation with enzyme. The initial velocities were
plotted as a function of inhibitor concentration for each
compound. The resulting data were fit to eq 1, where vi ) the
initial velocity at a given concentration of inhibitor and vc )
the initial velocity of a control reaction containing no inhibitor,
to yield the IC50 for inhibition at 30 °C by each compound.
In summary, a comprehensive study of the utility of
various zinc binding groups in LpxC inhibitors has been
performed. A new class of scaffolds, the isoxazolines,
was found to be effective in inhibition of this enzyme.
Though the antibacterial activity of isoxazoline hydrox-
amate 2 was surprisingly low, additional SAR studies
being performed in our labs have indicated that quite
potent antibacterial compounds can be developed based
on the isoxazoline scaffold given the correct choice of
the aryl group at the 3-position of the ring.42 Because
we have shown that the oxazoline scaffold is not strictly
required for activity, it remains to be seen the extent
to which the scaffold can be varied. To this end, a
systematic study of scaffolds should be undertaken.
Several zinc binding groups attached to the isoxazo-
line were found to have greater or nearly equal enzy-
matic activity as compared to the corresponding hy-
droxamic acid. Although the specific ranking of these
inhibitors is likely of little importance, this study has
identified several new lead structures from which potent
inhibitors can likely be developed. Additionally, a mech-
anism-based inhibitor that shows time-dependent ir-
reversible enzyme inhibition has been identified. This
is the first systematic study of metal binding groups for
LpxC deacetylase inhibition. This work represents the
most comprehensive direct study of the zinc binding
group in any family of zinc amidase inhibitors to date43
and complements the study of Colletti et al.35 on zinc
proteases. The rankings provided should be a useful
starting point for attempts to identify zinc binding
vi/vc ) IC50/([I] + IC50
)
(1)
Tim e-Dep en d en t In h ibition of Lp xC. Because high
protein concentrations stabilize LpxC activity, assays were
performed at concentrations >0.3 mg/mL that enabled a slow
alternative substrate, UDP-GlcNAc, to be used. Stock solutions
of 25a ,b were prepared in 100% dimethyl formamide (DMF).
E. coli LpxC (100 µM) was incubated at 30 °C with varied
concentrations (0-0.8 mM) of each inhibitor in 100 mM bis-

4366 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19
Pirrung et al.
Tris, pH 7, 12% DMF. At various times, aliquots were diluted
10-fold into 27 mM UDP-GlcNAc in 100 mM bis-Tris, pH 7, at
30 °C in order to measure remaining enzyme activity. The
initial velocity (<10% product formation) was determined from
the measurement of UDP-GlcN by the following method. An
aliquot (25 µL) of the assay mixture was diluted with 75 µL of
1 M borate, pH 9, and subsequently mixed with 30 µL of 10
mM fluorescamine (in acetonitrile). Fluorescamine reacts with
the free amine of UDP-GlcN to form a fluorescent adduct (λex
) 395 nm, λem ) 485 nm). The concentration of the product
was determined from the fluorescence by comparison to a
UDP-GlcN standard curve.44 In the absence of inhibitor,
enzyme activity was constant for >3 h under these conditions.
The initial velocities were plotted as a function of incubation
time, and the data fit to an exponential decay to determine
the pseudo-first-order rate constant for inactivation. These
observed rate constants were then plotted as a function of
inhibitor concentration, and eq 2 was fit to these data. To
examine whether this inhibition was reversible, LpxC (0.2 mM)
was incubated with 0.5 mM 25b at 30 °C for 4 h in 100 mM
bis-Tris, pH 7, 12% DMF. The inactivated enzyme was then
diluted 20-fold into 100 mM bis-Tris, pH 7, which decreases
the inhibitor concentration to <25 µM, 10-fold below the KI.
Aliquots were removed at various times and assayed as
described above. The initial velocity of both LpxC and inacti-
vated LpxC altered by <10% over a 45 h incubation.
N-(2,4,6-Tr im eth oxyben zyl)-O-(2,4-d im eth oxyben zyl)-
3-(4-m et h oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r b oxyl-
ic Acid Hyd r oxya m id e (24) a n d N-(2,4,6-Tr im eth oxyben -
zyl)-O-(2,4-d im eth oxyben zyl)a cr ylic Hyd r oxa m a te (1).
NH(TMB)ODMB22 (0.2 g, 0.55 mmol) and N-methylmorpholine
(0.24 mL, 2.2 mmol) were stirred in 5 mL of CH2Cl2. Acryloyl
chloride (0.179 mL, 2.20 mmol) was added, and the solution
was stirred overnight. The reaction was quenched (carefully)
with 10 mL of saturated NaHCO3 and stirred for 30 min. CH2-
Cl2 (10 mL) was added, and the organic layer was removed
and dried over MgSO4. The mixture was filtered and evapo-
rated to give 0.24 g (100%) of the acrylic hydroxamate as a
yellow solid. A portion of solid (0.24 mmol, 0.1 g) and chloro-
oxime 3 (0.36 mmol, 67 mg) was stirred in 5 mL of EtOAc at
room temperature. N-Methylmorpholine (0.36 mmol, 40 µL)
was added dropwise, and the solution was stirred for 3 h. TLC
indicated that the reaction was complete. The product was
purified by prep TLC (3:2 EtOAc:Hex, Rf ) 0.30) to give 87
mg (65%) of a clear oil. IR (thin film): 2955, 2843, 1665, 1611,
1
1513, 1206, 1151,1038 cm-1. H NMR (CDCl3): δ 7.57 (2H, d,
J ) 8.7 Hz), 7.03 (1H, d, J ) 8.7 Hz), 6.88 (2H, d, J ) 8.7 Hz),
6.40-6.42 (2H, m), 6.12 (2H, s), 5.54 (1H, t, J ) 9.3 Hz), 5.08
(1H, d, J ) 14.1 Hz), 4.94 (1H, d, J ) 14.1 Hz), 4.82 (1H, d, J
) 9.9 Hz), 4.79 (1H, d, J ) 9.9 Hz), 3.83 (3H, s), 3.82 (3H, s),
3.81 (3H, s), 3.77 (3H, s), 3.74 (6H, s), 3.33 (1H, d, J ) 11.1
Hz), 3.32 (1H, d, J ) 7.5 Hz). 13C NMR: δ 169.59, 161.63,
161.04, 160.70, 159.85, 159.22, 155.12, 132.67, 128.15, 121.77,
115.07, 113.84, 104.13, 103.60, 98.29, 90.16, 76.68, 71.32, 55.61
(2C), 55.31 (3C), 39.01, 37.66. FAB MS (M + H): 567. HRMS
m/z calcd for C30H34N2O9, 567.2342; found, 567.2329.
kobs ) kinact[I]/(1 + [I]/KI)
(2)
Disk Diffu sion Test for An tiba cter ia l Activity of Lp xC
In h ibitor s. A 5 mL culture of E. coli strain R477 (wild type)
was grown to stationary phase at 37 °C in LB broth containing
30 µg/mL streptomycin. This culture was diluted with LB broth
to A600 ) 0.2, and then, a sterile cotton swab was used to
spread an even lawn of the diluted culture onto an LB agar
plate. Filter paper disks (6 mm diameter) were saturated with
50 µg of the compound (in 100% DMSO) to be tested for
antibacterial activity. The disks were placed on the lawn of
freshly plated cells, and the plates were then incubated at 37
°C. Inhibition of bacterial growth was detected as a zone of
clearing around each disk. The diameter of the zone of growth
inhibition was measured for each compound after overnight
incubation. A diameter of 6 mm indicated that there was no
visible inhibition of growth beyond the edges of the disk.
MIC. Antibacterial activity of potent LpxC inhibitors (2 and
L-161,240) was determined using two strains of E. coli, the
wild-type strain (R477), and G17S LpxC mutant strain (EnvA).
Overnight cultures of R477 and G17S were grown at 37 °C
and then diluted to OD660 ) 0.1. The diluted culture was
diluted again 1:100 into 50 µL of LB containing varied
concentrations of inhibitor or DMSO as a control. A 96 well
microtiter plate was used such that each well contained a
different concentration of inhibitor (0.001-500 µg/mL). The
cultures were allowed to grow for 7 h at 37 °C with shaking.
The MIC was defined as the lowest inhibitor concentration that
inhibited growth as measured by no increase in A660 during
the time of the assay.
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid Hyd r oxya m id e (2). Isoxazoline 2 (75 mg, 0.13
mmol) and 0.11 mL of triethylsilane (0.65 mmol) were stirred
in 3 mL of 50% trifluoroacetic acid (TFA) in CH2Cl2. After 3
h, the solution was evaporated and was triturated with ether
to give 22 mg (73%) of a pink solid; mp 152-153 °C (dec). IR
(thin film): 3166, 2950, 2832, 1631, 1611, 1517, 1254, 828
cm-1. 1H NMR (DMSO-d6): δ 10.99 (1H, bs), 9.04 (1H, bs), 7.62
(2H, d, J ) 8.4 Hz), 7.00 (2H, d, J ) 8.4 Hz), 4.94 (1H, dd, J
) 7.5, 11.1 Hz), 3.79 (3H, s), 3.59 (1H, dd, J ) 11.1, 17.1 Hz),
3.52 (1H, dd, J ) 7.2, 16.8 Hz). 13C NMR: δ 165.64, 160.56,
155.60, 128.22, 120.93, 114.14, 77.34, 55.28, 38.33. FAB MS
(M + H): 237. Anal. (C11H12N2O4) C, H, N.
(R)- an d (S)-3-(4-Meth oxyph en yl)-4,5-dih ydr o-isoxazole-
5-ca r boxylic Acid Hyd r oxya m id e (2). (R)- or (S)-Alcohol
422 (0.1 g, 0.48 mmol) was dissolved in 100 mL of acetone.
J ones reagent (1.2 mmol, 2.5 equiv) was added, and the
solution was stirred for 3 h. The reaction was quenched with
1 mL of MeOH and concentrated. The residue was partitioned
between water and CH2Cl2. The organic layer was extracted
with aqueous NaOH. The aqueous layer was acidified and
extracted twice with CH2Cl2. The organic layer was dried over
MgSO4, filtered, and evaporated to give 44 mg (42%) of acid 5
(see racemic 6 for characterization). Using EDC, a small
sample of the acid was coupled to (R)-R-phenethylamine. The
de could be measured by NMR by observing the doublets at δ
1.52 and δ 1.47. Acid 5 (0.12 g, 0.54 mmol) was dissolved in
20 mL of CH2Cl2. O-Trityl hydroxylamine (0.15 g, 0.55 mmol)
was added followed by EDC (0.14 g, 0.75 mmol). The solution
was stirred for 48 h after which time the solution was
concentrated and partitioned between EtOAc and water. The
organic layer was washed with 5% citric acid(aq), saturated
aqueous NaHCO3, and brine. The solution was dried over
MgSO4, filtered, and concentrated. The residue was treated
with 30% TFA in CH2Cl2 (15 mL) for 20 min. MeOH (1 mL)
was added, and the solution was concentrated. The residue
was triturated with ether to give the desired hydroxamate 2
as a white solid (62 mg, 49% from 5).
4-Meth oxyben zoh yd r oxim in oyl Ch lor id e (3).45 Hydrox-
ylamine hydrochloride (5.60 g, 80.9 mmol) was added to 2.6
mL of water. Triethylamine (11.3 mL, 80.9 mmol) was added
followed by 40 mL of tetrahydrofuran (THF). The suspension
was sonicated briefly. Trimethyl orthoformate (28.0 mL, 257
mmol) was added followed by p-anisaldehyde (8.90 mL, 73.5
mmol) and 40 mL of THF. The solution was stirred for 4 h.
The solution was evaporated and partitioned between CH2Cl2
and water. The organic layer was dried over MgSO4 and
evaporated to give a white solid (11.2 g, 101%). To the white
solid was added 100 mL of DMF. N-Chlorosuccinimide (9.80
g, 73.5 mmol) was added, and HCl gas from the headspace of
a concentrated HCl bottle was bubbled into the solution until
the reaction became warm. After it was stirred for 1 h at room
temperature, the solution was poured over 150 mL of water
and 150 mL of ether. The ether was washed 6× with water,
dried over MgSO4, filtered, and evaporated to give 11.1 g (81%)
of a yellow oil that solidified overnight.
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid (6). p-Anisaldehyde oxime (0.36 g, 2.37 mmol) and
N-chlorosuccinimide (0.32 g, 2.37 mmol) were stirred in 2 mL
of DMF. A small amount of HCl gas from the headspace of a
concentrated HCl bottle was bubbled into the solution until
the reaction became warm. The solution was stirred for 1 h at

Isoxazoline Zinc Amidase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19 4367
room temperature. CH2Cl2 (10 mL) was added followed by 0.45
mL (5 mmol) of methyl acrylate. N-Methyl morpholine (0.275
mL, 2.5 mmol) was added dropwise. After 1 h, the solution
was concentrated and partitioned between H2O and EtOAc.
The organic layer was washed 4× with water and evaporated
onto silica gel. The product was purified by flash chromatog-
raphy (3:1 Hex:EtOAc, Rf ) 0.33) to give 0.47 g (84%) of a white
solid, which was dissolved in 20 mL of 1:1 THF:MeOH. NaOH
(1 M, 2.5 mmol, 2.5 mL) was added, and the solution was
stirred overnight. The solvent was evaporated, and water (10
mL) was added. The aqueous solution was washed with CH2-
Cl2 and then acidified with 3.5 mL of 1 M HCl. The mixture
was extracted 2× with 15% 2-propanol in CH2Cl2. The solvent
was dried over MgSO4, filtered, and evaporated to give 0.52 g
(84%) of a white solid; mp 168-170 °C. IR (thin film): 2955,
g, 0.27 mmol) was stirred in 3 mL of CHCl3. EDC (61 mg, 0.32
mmol) was added followed by 18 mg (0.3 mmol) of formyl
hydrazine. The solution was sonicated briefly and then allowed
to stand at room temperature overnight, and a white precipi-
tate formed. The precipitate was filtered to give 37 mg (52%)
of a white solid; mp 195-200 °C (dec). IR (Nujol): 3185, 1684,
1620, 1518, 1463, 1253, 831 cm-1. 1H NMR (DMSO-d6, 95 °C):
δ 10.05 (1H, bs), 9.75 (1H, bs), 7.99 (1H, bs), 7.61 (2H, d, J )
9 Hz), 6.99 (2H, d, J ) 9 Hz), 5.13 (1H, dd, J ) 7.5, 11.4 Hz),
3.81 (3H, s), 3.66 (1H, dd, J ) 11.1, 17.1 Hz), 3.55 (1H, dd, J
) 7.2, 17.1 Hz). 13C NMR: δ 168.84, 161.47, 159.86, 156.44,
129.13, 121.59, 115.00, 78.37, 56.12, 39.59. FAB MS (M + H):
264. HRMS m/z calcd for C12H14N3O4, 264.0984; found, 264.0974.
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid Th ia zol-2-yla m id e (11). Acid 6 (0.27 mmol, 60 mg),
2-aminothiazole (0.27 mmol, 27 mg), and EDC (0.35 mmol, 68
mg) were stirred in 5 mL of CH2Cl2 for 4 h. A white precipitate
formed. The precipitate was filtered and washed with CH2Cl2
to give 32 mg (39%) of 11 as a white powder; mp 210 °C (dec).
IR (Nujol): 2855, 2710, 1690, 1608, 1583, 1459, 1361, 1263
1713, 1727, 1610, 1226, 901 cm-1 1H NMR (CDCl3): δ 7.61
.
(2H, d, J ) 9 Hz), 6.93 (2H, d, J ) 9 Hz), 5.19 (1H, dd, J )
7.2, 9.3 Hz), 3.85 (3H, s), 3.71 (1H, d, 9.9 Hz), 3.71 (1H, d, 7.5
Hz). 13C NMR: δ 175.82, 162.68, 157.26, 129.34, 122.56,
115.13, 80.06, 55.84, 40.56. FAB MS (M + H): 221. Anal.
(C11H11NO4) C, H, N.
1
cm-1. H NMR (CDCl3): δ 12.5 (1H, bs), 7.63 (1H, d, J ) 8.7
2-Th ioa cetyl-1-[3-(4-m eth oxyp h en yl)-4,5-d ih yd r o-isox-
a zol-5-yl]eth a n on e (7). Acid 6 (1.0 g, 4.5 mmol) was stirred
in 50 mL of CH2Cl2. SOCl2 (15 mL) was added, and the solution
was gently heated until it became clear. After it was stirred
for 2 h, the solvent was removed and the acid chloride was
dissolved in ether (25 mL). This was slowly added to a freshly
prepared solution of diazomethane that was made from 4.0 g
(18 mmol) of Diazald. After the mixture was allowed to stand
at room temperature for 1 h, air was blown over the top to
remove the solvent. The crude product was recrystallized from
boiling ethanol to give 0.8 g of the R-chloroketone. A portion
of this (0.4 mmol, 0.1 g) was dissolved in 7 mL of THF.
Thioacetic acid (36 µL, 0.5 mmol) was added, and the solution
became cloudy immediately. After 1 h, the solution was
evaporated and partitioned between CH2Cl2 and water. The
organic layer was dried over MgSO4, filtered, and evaporated
to give 0.12 g (69% from 6) of the protected thiol; mp 81 °C.
Hz), 7.50 (1H, d, J ) 3.6 Hz), 7.27 (1H, d, J ) 3.6 Hz), 7.01
(2H, d, J ) 8.7 Hz), 5.33 (1H, t, J ) 9 Hz), 3.80 (3H, s), 3.69
(2H, d, 9 Hz). 13C NMR: δ 167.66, 160.63, 160.60, 155.78,
137.68, 128.32, 120.75, 114.15, 114.10, 78.06, 55.28, 38.23. FAB
MS (M + H): 304. Acceptable combustion microanalytical data
could not be obtained for this compound. HRMS m/z calcd for
C
14H14N3O3S, 304.0756; found, 304.0759.
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid P yr id in -2-yla m id e (12). Acid 6 (60 mg, 0.27 mmol),
2-aminopyridine (25 mg, 0.27 mmol), and EDC (68 mg, 0.35
mmol) were stirred in CH2Cl2 for 4 h. The solution was
concentrated and purified by prep TLC (3:2 EtOAc:Hex, Rf )
0.60) to give 60 mg (75%) of compound 12; mp 145 °C. IR
(Nujol): 3184, 2855, 1702, 1610, 1583, 1435, 1299, 1257, 1177
cm-1 1H NMR (CDCl3): δ 9.12 (1H, bs), 8.30 (1H, ddd, J )
.
0.6, 1.8, 4.8 Hz), 8.18 (1H, dt, J ) 1.2, 8.4 Hz), 7.74-7.66 (1H,
m), 7.62 (2H, d, J ) 8.7 Hz), 7.06 (1H, ddd, 0.9, J ) 4.8, 7.2
Hz), 6.92 (2H, d, J ) 9 Hz), 5.22 (1H, dd, J ) 6.9, 10.2 Hz),
3.84 (3H, s), 3.75 (1H, d, J ) 6.9 Hz), 3.75 (1H, d, J ) 10.2
Hz). 13C NMR: δ 169.78, 161.41, 156.53, 150.16, 148.00,
138.10, 128.55, 120.52, 120.26, 114.15, 113.85, 78.61, 55.37,
39.95. FAB MS (M + H): 298. Anal. (C16H15N3O3 + H2O) C,
H, N.
IR (Nujol): 1230, 1692, 1608, 1463, 1254 cm-1 1H NMR
.
(CDCl3): δ 7.63 (2H, d, J ) 8.8 Hz), 6.93 (2H, d, J ) 9.2 Hz),
5.22 (1H, dd, J ) 6, 12 Hz), 4.08 (1H, d, J ) 17.2 Hz), 3.99
(1H, d, J ) 17.2 Hz), 3.84 (3H, s), 3.75 (1H, dd, J ) 6, 17.2
Hz), 3.55 (1H, dd, J ) 12, 17.2 Hz), 2.37 (3H, s). 13C NMR: δ
202.86, 194.44, 161.65, 156.68, 128.80, 121.08, 114.49, 83.93,
55.72, 38.58, 36.95, 30.44. FAB MS (M + H): 294. Anal.
(C14H15NO4S) C, H, N.
2-Mer ca p to-1-[3-(4-m eth oxyp h en yl)-4,5-d ih yd r o-isox-
a zol-5-yl]eth a n on e (8). Hydrazine (1 M, 0.18 mmol, 0.18 mL)
was slowly added to protected thiol 7 (68 mg, 0.23 mmol) that
was being stirred in 3 mL of THF. A new spot appeared by
TLC (2.5% EtOH in CH2Cl2, Rf ) 0.33) and was purified by
prep TLC to give a mixture of the desired product and its
dimer. 1H NMR (CDCl3): δ 7.60 (2H, d, J ) 8.7 Hz), 6.92 (2H,
d, J ) 9 Hz), 5.25 (1H, dd, J ) 5.7, 11.7 Hz), 3.83 (3H, s),
3.76-3.64 (3H, m), 3.55 (1H, dd, J ) 11.7, 16.8 Hz). FAB MS
(M + H and 2M + H): 252 and 503.
3-(4-Met h oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-su lfon -
ic Acid Am id e (13). Chloro-oxime 3 was generated in situ
from 0.33 mmol (50 mg) of p-anisaldehyde oxime and N-
chlorosuccinimide (44 mg, 0.33 mmol) as described in the
synthesis of 6. Vinyl sulfonamide46 (43 mg, 0.4× mmol) was
added as a solution in CH2Cl2 (10 mL). N-Methyl morpholine
(44 µL, 0.33 mmol) was added, and the solution was stirred
for 2 h. The solution was concentrated and partitioned between
EtOAc and water. The organic layer was washed 3× with
water, dried over MgSO4, and filtered, and the solvent was
evaporated. CH2Cl2 was added (3 mL), and a white precipitate
formed. The precipitate was filtered to give 45 mg (63%) of a
white solid; mp 175-176 °C. IR (Nujol): 3346, 3227, 1613,
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid Hyd r a zid e (9). Carboxylic acid 6 (0.06 g, 0.27
mmol) was stirred in 5 mL of CH2Cl2. EDC (0.104 g, 0.54
mmol) was added followed by 85 µL (2.7 mmol) of anhydrous
hydrazine. The solution was stirred overnight and then
evaporated. Water (10 mL) was added to the residue, and the
suspension was sonicated briefly. The resultant white solid
was filtered to give 41 mg of product (65%); mp 173-175 °C.
1521, 1455,1366, 1338, 1158 cm-1 1H NMR (acetone-d6): δ
.
7.66 (2H, d, J ) 9 Hz), 7.00 (2H, d, J ) 9 Hz), 6.43 (2H, bs),
5.58 (1H, dd, J ) 5.1, 10.8 Hz), 3.93 (1H, dd, J ) 10.8, 8.3
Hz), 3.84 (3H, s), 3.82 (1H, dd, J ) 5.4, 18.6 Hz). 13C NMR: δ
162.39, 157.27, 129.44, 121.17, 115.04, 92.40, 56.04, 38.94. FAB
MS (M + H): 257. Anal. (C10H9N2O4S) C, H, N.
5-Meth a n esu lfon yl-3-(4-m eth oxyp h en yl)-4,5-d ih yd r o-
isoxa zole (14). Chloro-oxime 3 (2.74 mmol) was generated
from the corresponding oxime as was described in the syn-
thesis of 6. CH2Cl2 was added followed by methyl vinyl sulfone
(0.24 mL, 2.74 mmol). Triethylamine (0.38 mL, 2.74 mmol) was
slowly added, and the solution was stirred for 1 h. The solution
was concentrated and triturated with water. The white solid
was filtered, giving 0.5 g (71%) of product; mp 167-168 °C.
IR (Nujol): 3015, 1609, 1601, 1521, 1460, 1303, 1131, 1014
1
IR (Nujol): 3297, 3245, 1689, 1609, 1460, 828 cm-1. H NMR
(DMSO-d6): δ 9.43 (1H, bs), 7.61 (2H, d, J ) 9 Hz), 7.00 (2H,
d, J ) 9 Hz), 4.98 (1H, dd, J ) 7.5, 11.1 Hz), 4.36 (1H, bs),
3.79 (3H, s), 3.60 (1H, dd, J ) 11.1, 16.8 Hz), 3.50 (1H, dd, J
) 7.5, 17.1 Hz). 13C NMR: δ 168.79, 161.39, 156.41, 129.04,
121.80, 114.97, 78.69, 56.11, 39.30. FAB MS (M + H): 236.
Acceptable combustion microanalytical data could not be
obtained for this compound.
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid (N′-for m yl-h yd r a zid e) (10). Carboxylic acid 6 (0.6
cm-1
.
1H NMR (DMSO-d6): δ 7.68 (2H, d, J ) 8.4 Hz), 7.02
(2H, d, J ) 8.4 Hz), 5.88 (1H, dd, J ) 5.1, 10.8 Hz), 3.95 (1H,

4368 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19
Pirrung et al.
dd, J ) 10.5, 18.6 Hz), 3.87 (1H, dd, J ) 4.8, 18.3 Hz), 3.80
(3H, s), 3.04 (3H, s). 13C NMR: δ 161.09, 156.54, 128.75,
119.59, 114.24, 91.16, 55.36, 36.28, 35.71. FAB MS (M + H):
256. Anal. (C11H13NO4S) C, H, N.
3 (0.335, 1.99 mmol) was added to a solution of vinyl phos-
phinate 19 (0.335 g, 1.99 mmol) in 25 mL of CH2Cl2. Et3N (0.28
mL, 2 mmol) was added to the solution, and it was stirred at
room temperature for 1 d. The solution was evaporated onto
silica gel and purified by flash chromatography (2.5% EtOH
in CH2Cl2, Rf ) 0.1) to give 0.37 g of 20. This was dissolved in
15 mL of CH2Cl2 and treated with 0.6× mL (4.7 mmol) of
bromotrimethylsilane. The solution was stirred for 3 d at room
temperature. Water was added (1 mL), and the solution was
stirred for 3 h. The solution was evaporated, and the crude
residue was recrystallized from a few milliliters of EtOH to
give 0.21 g (41%) of 21; mp 182-183 °C. IR (Nujol): 1609,
3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zole-5-ca r box-
ylic Acid Hyd r oxym eth yl-a m id e (15). Chloro-oxime 23
(0.38 mmol, 70 mg) was stirred in 5 mL of THF. N-Hydroxy-
methyl acrylamide (48% solution in water, 74 µL, 0.38 mmol)
was added followed by Et3N (0.38 mmol, 53 µL). A white
precipitate immediately formed. After 20 min, the filtrate was
evaporated, MeOH (5 mL) was added, and the mixture was
sonicated. The resultant solid was filtered giving 25 mg (26%)
of nearly pure product. An analytical sample could be obtained
by chromatography (3:1 EtOAc:Hex, Rf ) 0.2); mp 152 °C (dec).
IR (Nujol): 3353, 3300, 1663, 1611, 1518, 1462, 1254, 1050
1
1516, 1464, 1306, 1253, 1179, 985 cm-1. H NMR (DMSO-d6):
δ 7.61 (2H, d, J ) 8.4 Hz), 6.99 (2H, d, J ) 8.7 Hz), 4.68 (1H,
ddd, J ) 4.8, 9.9, 15 Hz), 3.78 (3H, s), 3.75-3.39 (2H, m), 1.35
(3H, d, J ) 14.4 Hz). 31P NMR: δ 41.29. 13C NMR: δ 160.55,
155.56 (d, J ) 7.5 Hz), 128.25, 120.92, 114.11, 77.19 (d, J )
109 Hz), 55.27, 36.05, 12.18 (d, J ) 93 Hz). FAB MS (M + H):
256. Anal. (C11H14NO4P) C, H, N.
1
cm-1. H NMR (acetone-d6): δ 8.08 (1H, bs), 7.66 (2H, d, J )
9 Hz), 6.99 (2H, d, J ) 9 Hz), 5.05 (1H, dd, J ) 6.3, 11.7 Hz),
4.78-4.70 (2H, m), 3.84 (3H, s), 3.68 (1H, dd, J ) 11.4, 16.8
Hz), 3.59 (1H, dd, J ) 6.3, 16.8 Hz). 13C NMR: δ 169.95,
160.61, 155.69, 128.25, 120.83, 114.14, 78.55, 62.29, 55.27,
37.95. FAB MS (M + H): 251. HRMS m/z calcd for C12H14N2O4,
250.0954; found, 250.0943.
Th iop r op ion ic Acid S-[3-(4-Meth oxyp h en yl)-4,5-d ih y-
d r o-isoxa zol-5-ylm eth yl)ester (22). Chloro-oxime 3 (2.74
mmol) was generated in situ as described in the synthesis of
6. CH2Cl2 (10 mL) was added followed by allyl thiopropionate
(2.7 mmol, 0.37 mL). After it was stirred overnight, the
solution was evaporated onto silica gel and purified by flash
chromatography (3:1 Hex:EtOAc, Rf ) 0.52) to give 0.487 g
(64%) of a white solid; mp 62 °C. IR (thin film): 2990, 2979,
[3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zol-5-yl]p h os-
p h on ic Acid Dieth yl Ester (16). Chloro-oxime 3 (2.74 mmol)
was generated in situ as described in the synthesis of 6.
Diethyl vinylphosphonate (2.74 mmol, 0.42 mL) was added
followed by slow addition of Et3N (2.74 mmol, 0.38 mL). The
solution was stirred for 1 h and then concentrated. The residue
was dissolved in EtOAc (50 mL), and the mixture was washed
3× with water and once with brine. The solution was dried
over MgSO4, filtered, and evaporated onto silica gel. The
product was purified by flash chromatography (EtOAc, Rf )
0.3) to give 0.69 g (80%) of 16 as an oil. IR (thin film): 2983,
1
2832, 1695, 1608, 1518, 1254, 831 cm-1. H NMR (CDCl3): δ
7.59 (2H, d, J ) 9 Hz), 6.91 (2H, d, J ) 9 Hz), 4.91-4.81 (1H,
m), 3.83 (3H, s), 3.40 (1H, dd, J ) 10.2, 16.8 Hz), 3.23 (1H,
dd, J ) 6, 14.1 Hz), 3.19 (1H, dd, J ) 6.3, 14.1 Hz), 3.06 (1H,
dd, J ) 6.9, 16.5 Hz), 2.61 (q, 2H, J ) 7.8 Hz), 1.18 (3H, t, J
) 7.8 Hz). 13C NMR: δ 199.42, 160.97, 155.83, 128.16, 121.75,
114.05, 79.17, 55.38, 39.65, 37.49, 32.63, 9.77. FAB MS (M +
H): 280. Anal. (C14H17NO3S) C, H, N.
1
1609, 1517, 1306, 1255, 1022, 972 cm-1. H NMR (CDCl3): δ
7.60 (2H, d, J ) 9 Hz), 6.91 (2H, d, J ) 9 Hz), 4.83 (1H, ddd,
J ) 2.1, 10.2, 11.7), 4.28-4.18 (4H, m), 3.83 (3H, s), 3.63 (1H,
dd, J ) 10.2, 23.7 Hz), 3.63 (1H, dd, J ) 11.7, 22.2 Hz), 1.36
(3H, t, J ) 7.2 Hz), 1.32 (3H, t, J ) 7.2 Hz). 13C NMR: δ 161.19,
155.68 (d, J ) 6 Hz), 128.38, 120.90, 114.11, 74.86 (d, J ) 168
Hz), 63.57 (d, J ) 6.9 Hz), 63.16 (d, J ) 6.6 Hz), 55.37, 37.99,
16.59, 16.52. 31P NMR: δ 19.64. FAB MS (M + H): 314. Anal.
(C14H20NO5P) C, H, N.
[3-(4-Met h oxyp h en yl)-4,5-d ih yd r oisoxa ol-5-yl]m et h -
a n eth iol (23). Thioester 22 (75 mg, 0.27 mmol) was stirred
in 4 mL of methanol. NaOH (1M in H2O, 0.27 mL, 0.27 mmol)
was added followed by 1 mL of water. The white precipitate
that was formed after it was stirred overnight was filtered
giving 19 mg (32%) of the thiol; mp 185 °C (dec). IR (Nujol):
1
1610, 1518, 1461, 1376, 1253, 829 cm-1. H NMR (DMSO-d6):
δ 7.59 (2H, d, J ) 9 Hz), 6.98 (2H, d, J ) 9 Hz), 4.96-4.85
(1H, m), 3.78 (3H, s), 3.55 (1H, dd, J ) 10.5, 17.1 Hz), 3.22
(1H, dd, J ) 6.9, 17.1 Hz), 3.14-2.99 (2H, m). 13C NMR: δ
160.40, 155.80, 128.01, 114.05, 78.74, 55.24, 42.40. FAB MS
(M + H): 224. Anal. (C11H13NO2S) C, H, N.
[3-(4-Meth oxyp h en yl)-4,5-d ih yd r o-isoxa zol-5-yl]p h os-
p h on ic Acid (17). Phosphonate ester 16 (0.25 g, 0.80 mmol)
was dissolved in 15 mL of CH2Cl2. Bromotrimethylsilane (0.34
mL, 6.4 mmol) was added, and the solution was stirred for 3
d at room temperature. The reaction was quenched by addition
of 1 mL of water, and the mixture was stirred for 20 min. The
solution was concentrated and made basic by the addition of
saturated NaHCO3. The solution was evaporated and dissolved
in a minimal amount of water. Concentrated HCl was added
to acidify the solution. The white precipitate that formed was
filtered and dried to give 121 mg (59%) of 17; mp 238-240 °C.
IR (Nujol): 2775, 2356, 1610, 1516, 1459, 1376, 1180, 1040
cm-1. 1H NMR (DMSO-d6): δ 8.4 (2H, bs), 7.59 (2H, d, J ) 8.4
Hz), 6.98 (2H, d, J ) 8.7 Hz), 4.62 (1H, t, J ) 11.4 Hz), 3.78
(3H, s), 3.74-3.57 (1H, m), 3.47-3.30 (1H, m). 13C NMR: δ
160.49, 155.42, 128.17, 121.17, 114.15, 75.6 (d, J ) 162 Hz),
55.31, 37.31. FAB MS (M - H): 256. HRMS m/z calcd for
(R*,R*)-3-(4-Meth oxy-p h en yl)-5-th iir a n yl-4,5-d ih yd r o-
isoxa zole (25a ). (R*,R*)-3-(4-Methoxy-phenyl)-5-oxiranyl-4,5-
dihydro-isoxazole (24a , 0.05 g, 0.23 mmol) was dissolved in
1 mL of MeOH. Thiourea (17 mg, 0.23 mmol) was added, and
the suspension was heated to 80 °C for 5 h in a pressure tube.
After it was cooled, the solution was partitioned between 10
mL each of DCM and H2O. The organic layer was dried over
MgSO4 and evaporated to give 57 mg of a white solid (100%);
mp 125-128 °C. IR (Nujol): 1610, 1597, 1518, 1464, 1253,
1
1180 cm-1. H NMR (CDCl3): δ 7.60 (2H, d, J ) 9.3 Hz), 6.92
(2H, d, J ) 9.3 Hz), 4.88 (1H, ddd, J ) 5.1, 6.6, 10.5 Hz), 3.84
(3H, s), 3.40 (1H, dd, J ) 10.5, 16.8 Hz), 3.19 (1H, dt, J ) 5.7,
6.3 Hz), 3.13 (1H, dd, J ) 6.6, 16.8 Hz), 2.47 (1H, dd, J ) 1.5,
6.3 Hz), 2.30 (1H, dd, J ) 1.5, 5.4 Hz). 13C NMR δ. FAB MS
(M + H): 236. Anal. (C12H14NO2S) C, H, N.
C
10H11NO5P, 256.0375; found, 256.0381. Anal. (C10H12NO5P)
C, H, N.
Met h yl-vin yl-p h osp h in ic Acid 2-Ch lor o-et h yl E st er
(19). Methyl dichlorophosphine (0.5× g, 4.3 mmol) was stirred
in 25 mL of dry ether under N2. To this solution was added
0.58 mL (8.6 mmol) of 2-chloroethanol followed by 1.7 mL (22
mmol) of pyridine. The solution was stirred overnight, filtered,
and evaporated to give a clear oil. The oil was heated under
N2 in a pressure tube to 160 °C for 3 h. The reaction was
allowed to cool, and 0.65 mL (4.7 mmol) of Et3N was added.
The vessel was sealed and heated to 120 °C for 2 h. The residue
was purified by Kugelrohr distillation (160 °C, 2 Torr) to give
0.335 g (47%) of an oil identical to the known compound.47
(R*,S*)-3-(4-Meth oxy-p h en yl)-5-th iir a n yl-4,5-d ih yd r o-
isoxa zole (25b). (R*,S*)-3-(4-Methoxy-phenyl)-5-oxiranyl-4,5-
dihydro-isoxazole (24b, 82 mg, 0.37 mmol) was treated exactly
as above (25a ) to give 78 mg (100%) of a while solid; mp 127-
130 °C. IR (Nujol): 1611, 1597, 1519, 1457, 1255, 1179 cm-1
.
1H NMR (CDCl3): δ 7.61 (2H, d, J ) 8.7 Hz), 6.92 (2H, d, J )
8.7 Hz), 4.32 (1H, ddd, J ) 6.9, 8.1, 10.5 Hz), 3.84 (3H, s),
3.49 (1H, dd, J ) 10.2, 16.5 Hz), 3.33 (1H, dd, J ) 6.9, 16.5
Hz), 3.13 (1H, ddd, J ) 5.1, 6.0, 8.1 Hz), 2.62 (1H, dd, J ) 1.8,
6.0 Hz), 2.37 (1H, dd, J ) 1.8, 5.4 Hz). 13C NMR: δ 161.32,
156.08, 128.49, 121.91, 114.38, 85.37, 55.69, 40.63, 36.24,
25.06. FAB MS (M + H): 236. Anal. (C12H13NO2S) C, H, N.
Meth yl-[3-(4-m eth oxyp h en yl-4,5-d ih yd r o-isoa zol-5-yl)-
p h osp h in ic Acid 2-Ch lor o-eth yl Ester (20). Chloro-oxime

Isoxazoline Zinc Amidase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19 4369
negative bacteria. Inhibition of diverse UDP-3-O-(r-3-hydroxy-
myristoyl)-N-acetylglucosamine deacetylases by substrate ana-
logues containing zinc binding motifs. J . Biol. Chem. 2000, 275,
11002-11009.
(R*,S*)-2-Mer ca p to-1-[3-(4-m eth oxy-p h en yl)-4,5-d ih y-
d r o-isoxa zol-5-yl]eth a n ol (26a ). (R*,R*)-3-(4-Methoxy-phen-
yl)-5-oxiranyl-4,5-dihydro-isoxazole (24a , 50 mg, 0.23 mmol)
was dissolved in 4 mL of THF and 1 mL of H2O. (TMS)2S (95
µL, 0.45 mmol, 2 equiv) was added followed quickly by TBAF
(1 M in THF, 0.45 mmol, 2 equiv). After it was stirred at 0 °C
for 1 h, the solution was evaporated. The residue was treated
with 5 mL of boiling MeOH (the solid did not completely
dissolve). After the solution was allowed to cool, the white solid
was filtered to give pure product (25 mg, 43%); mp 168-170
(12) J ackman, J . E.; Raetz, C. R. H.; Fierke, C. A. Site-directed
mutagenesis of the bacterial metalloamidase UDP-(3-O-acyl)-
N-acetylglucosamine deacetylase (LpxC). Identification of the
zinc binding site. Biochemistry 2001, 40, 514-523.
(13) Christianson, D.; Whittington, D. Personal communication.
(14) J ackman, J . E.; Raetz, C. R. H.; Fierke, C. A. UDP-3-O-(R-3-
hydroxymyristoyl)-N-acetylglucosamine deacetylase of Escheri-
chia coli is a zinc metalloenzyme. Biochemistry 1999, 38, 1902.
(15) Mulder, G. J .; Meerman, J . H. N. Sulfation and glucuronidation
as competing pathways in the metabolism of hydroxamic acids:
the role of N,O-sulfonation in chemical carcinogenesis of aro-
matic amines. Environ. Health Persp. 1983, 49, 27.
(16) Vassiliou, S.; Mucha, A.; Cuniasse, P.; Georgiadis, D.; Lucet-
Levannier, K.; Beau, F.; Kannan, R.; Murphy, G.; Knauper, V.;
Rio, M.; Basset, P. Yiotakis, A.; Dive, V. Phosphinic pseudo-
tripeptides as potent inhibitors of matrix metalloproteinases: a
structure-activity study. J . Med. Chem. 1999, 42, 2610-2620.
(17) Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing, A. J . H. Design
and therapeutic application of matrix metalloproteinase inhibi-
tors. Chem. Rev. 1999, 99, 2735-2776.
(18) Smith, A. B.; Yager, K. M.; Taylor, C. M. Enantioselective
synthesis of diverse a-amino phosphonate diesters. J . Am. Chem.
Soc. 1995, 117, 10879-10888.
(19) Oxazolines unsubstituted at the 5-position rearrange or
polymerize in the presence of alkanethiols. L. N. Tumey.
Unpublished results.
(20) Padwa, A., Ed. 1,3-Dipolar Cycloaddition Chemistry; Wiley: New
York, 1984.
(21) Weidner-Wells, M. A.; Fraga-Spano, S. A.; Turchi, I. J . Unusual
regioselectivity of the dipolar cycloaddition reactions of nitrile
oxides and tertiary cinnamides and crotonamides. J . Org. Chem.
1998, 63, 6319-6328.
°C. IR (Nujol): 3481, 1602, 1597, 1530, 1477, 1261 cm-1 1H
.
NMR (DMSO-d6): δ 7.57 (2H, d, J ) 9.0 Hz), 6.98 (2H, d, J )
9.0 Hz), 4.65-4.55 (1H, m), 4.02 (1H, bs), 3.80-3.74 (1H, m),
3.78 (3H, s), 3.50-3.29 (2H, m), 2.97 (1H, dt, J ) 3.6, 13.5
Hz), 2.78 (1H, ddd, J ) 8.4, 9.6, 13.2 Hz). 13C NMR: δ 160.54,
155.84, 128.08, 121.89, 114.20, 82.61, 69.45, 55.29, 42.87, 35.7.
FAB MS (M + H): 254. Anal. (C12H15NO3S) C, H, N.
(R*,R*)-2-Mer ca p to-1-[3-(4-m eth oxy-p h en yl)-4,5-d ih y-
d r o-isoxa zol-5-yl]eth a n ol (26b). (R*,S*)-3-(4-Methoxy-phen-
yl)-5-oxiranyl-4,5-dihydro-isoxazole (24b, 50 mg, 0.23 mmol)
was dissolved in 5 mL of THF. (TMS)2S (95 µL, 0.45 mmol, 2
equiv) was added followed immediately by TBAF (1 M in THF,
0.45 mmol, 2 equiv). After it was stirred for 1 h, a new spot
was observed by TLC (10% EtOH in DCM, Rf ) 0.65). This
product was purified by flash chromatography and then
recrystallized from MeOH to give 18 mg (31%) of a white solid;
mp 170-174 °C. IR (Nujol): 3416, 1610, 1518, 1461, 1255 cm-1
.
1H NMR (DMSO-d6): δ 7.57 (2H, d, J ) 8.7 Hz), 6.97 (2H, d,
J ) 8.7 Hz), 5.18 (1H, dd, J ) 3.0, 6.3 Hz), 4.69 (1H, ddd, J )
3.6, 8.4, 12.0 Hz), 3.77 (3H, s), 3.66-3.58 (1H, m), 3.37 (1H,
dd, J ) 10.2, 16.8 Hz), 3.24 (1H, dd, J ) 8.4, 16.8 Hz), 2.75-
2.61 (2H, m). 13C NMR: δ 160.23, 155.62, 127.83, 121.84,
114.02, 81.76, 71.24, 55.23, 36.32, 35.28. FAB MS (M + H):
254. Anal. (C12H15NO3S + H2O) C, H, N.
(22) Barlaam, B.; Hamon, A.; Maudet, M. New hydroxylamines for
the synthesis of hydroxamic acids. Tetrahedron Lett. 1998, 39,
7865.
(23) Ukaji, Y.; Sada, K.; Katsuhiko, I. Enantioselective synthesis of
2-isoxazolines via asymmetric 1,3-dipolar cycloaddition of nitrile
oxides to achiral allyl alcohols. Chem. Lett. 1993, 1847-1850.
(24) The ee was determined by converting compound 5 to its R-phen-
ethyl amide derivative. The de of the (R) isomer was found to
be 95% while the de of the (S) isomer was found to be 87%.
(25) (a) Minowa, N.; Hirayama, M.; Fukatsu, S. Asymmetric synthesis
of (+)-phosphinothricin and (+)-2-amino-4-phosphonobutyric
acid. Tetrahedron Lett. 1984, 25, 1147-1150. (b) Gross, V. H.;
Ack n ow led gm en t. Financial support was provided
by NIH AI-42151 (M.C.P.), NIH GM-40602 (C.A.F.),
NIH GM GM08858 (Biological Chemistry traineeship
to L.N.T.), and an American Chemical Society Division
of Medicinal Chemistry-Bristol Myers Squibb Fellow-
ship (L.N.T.). The assistance of L. LaBean in adminis-
trative support of this work is greatly appreciated.
Gnauk, T. R-Amino acids and derivatives. VIII.
A simple
synthesis of phosphinothricin. J . Prakt. Chem. 1976, 318, 157-
160.
(26) Balsamo, A.; Breschi, M. C.; Chiellini, G.; Lucacchini, A.;
Macchia, M.; Martinelli, A.; Martini, C.; Nardini, C.; Orlandini,
E.; Romagnoli, F.; Rossello, A. Conformationally restrained
â-blocking oxime ethers. 2. Synthesis and â-adrenergic properties
of diastereoisomeric anti and syn 2-(5′-(3′-aryl-substituted)-
isoxazolidinyl)-N-alkylethanolamines. Eur. J . Med. Chem. 1994,
29, 855-867.
Refer en ces
(1) Levy, S. B. The challenge of antibiotic resistance. Sci. Am. 1998,
46-53.
(2) Wright, G. D. Resisting resistance: new chemical strategies for
battling superbugs. Chem. Biol. 2000, 7, R127-R132.
(3) Haddad, J .; Vakulenko, S.; Mobashery, S. An antibiotic cloaked
by its own resistance enzyme. J . Am. Chem. Soc. 1999, 121,
11922-11923.
(27) Brown, S.; Bernardo, M. M.; Li, Z.; Kotra, L. P.; Tanaka, Y.;
Fridman, R.; Mobashery, S. Potent and selective mechanism-
based inhibition of gelatinases. J . Am. Chem. Soc. 2000, 122,
6799-6800.
(4) Travis, J . Reviving the antibiotic miracle? Science 1994, 264,
360-362.
(5) Raetz, C. R. H. Bacterial endotoxins: extraordinary lipids that
activate eukaryotic signal transduction. J . Bacteriol. 1993, 175,
5745-5753.
(28) (a) Hu, J .; Fox, M. A.
A convenient trimethylsilylthioxy-
dehalogenation raction for the preparation of functionalized
thiols. J . Org. Chem. 1999, 64, 4959-4961. (b) J usto de Pomar,
J . C.; Soderquist, J . A. Regio- and stereospecific synthesis of (O-
TIPS)-protected 2-hydroxyalkylmercaptans from epoxides and
triisopropylsilanethiol. Tetrahedron Lett. 1998, 39, 4409-4412.
(29) Chen, M.; Steiner, M. G.; Laszlo, S. E.; Patchett, A. A.; Anderson,
M. S.; Hyland, S. A.; Onishi, H. R.; Silver, L. L.; Raetz, C. R. H.
Carbohydroxamido-oxazolidines: antibacterial agents that
target lipid A biosynthesis. Bioorg. Med. Chem. Lett. 1999, 9,
313-318.
(30) Standard MMP2 calculations were performed using Chem-3D
Pro version 4.0.
(31) Phosphinates are also frequently found to be monodentate zinc
binders.
(32) Babine, R. E.; Bender, S. L. Molecular recognition of protein-
ligand complexes: applications to drug design. Chem. Rev. 1997,
97, 1359-1472.
(33) Levin, J . I.; DiJ oseph, J . F.; Killar, L. M.; Sharr, M. A.; Skotnicki,
J . S.; Patel, D. V.; Xiao, X.; Shi, L.; Vavre, M.; Campbell, D. A.
The asymmetric synthesis and in vitro characterization of
succinyl mercapto alcohol and mercaptoketone inhibitors of
matrix metalloproteinases. Bioorg. Med. Chem. Lett. 1998, 8,
1163-1168.
(6) Wyckoff, T. J . O.; Raetz, C. R. H.; J ackman, J . E. Antibacterial
and antiinflammatory agents that target endotoxin. Trends
Microbiol. 1998, 6, 154-159.
(7) Coutrot, P.; Dumacrcay, S.; Finance, C.; Tabyaoui, M.; Tabyaoui,
B.; Grison, C. Investigation of new potent KDO-8-phosphate
synthetase inhibitors. Bioorg. Med. Chem. Lett. 1999, 9, 949-
952.
(8) Goldman, R.; Kohlbrenner, W.; Lartey, P.; Pernet, A. Antibacte-
rial agents specifically inhibiting lipopolysaccharide synthesis.
Nature 1987, 329, 162-164.
(9) See the following for a recently described KDO synthase inhibitor
having weak antibacterial activity. Birck, R. M.; Holler, T. P.;
Woodard, R. W. Identification of a slow tight-binding inhibitor
of 3-deoxy-D-manno-octulosonic acid 8-phosphate synthase. J .
Am. Chem. Soc. 2000, 122, 9334-9335.
(10) Onishi, R. H.; Pelak, B. A.; Gerckens, L. S.; Silver, L. L.; Kahan,
F. M.; Chen, M.; Patchett, A. A.; Galloway, S. M.; Hyland, S. A.;
Anderson, M. S.; Raetz, C. R. H. Antibacterial agents that inhibit
lipid A biosynthesis. Science 1996, 274, 980-982.
(11) J ackman, J . E.; Fierke, C. A.; Tumey, L. N.; Pirrung, M. C.;
Uchiyama, T.; Tahir, S. H.; Hindsgaul, O.; Raetz, C. R. H.
Antibacterial agents that target lipid A biosynthesis in gram-

4370 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 19
Pirrung et al.
(34) Hackeng, T. T.; Griffin, J . H.; Dawson, P. E. Protein synthesis
by native chemical ligation: expanded scope by using straight-
forward methodology. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
10068-10073.
(35) Colletti, S. L.; Myers, R. W.; Darkin-Rattray, S. J .; Gurnett, A.
M.; Dulski, P. M.; Galuska, S.; Allocco, J . J .; Ayer, M. B.; Li, C.;
Lim, J .; Crumlet, T. M.; Cannova, C.; Schmatz, D. M.; Wyvratt,
M. J .; Fisher, M. H.; Meinke, P. T. Broad spectrum antiprotozoal
agents that inhibit histone deacetylase: structure-activity
relationships of apicidin. Part 1. Bioorg. Med. Chem. Lett. 2001,
11, 107-111.
(36) See the following for a similar approach that was used success-
fully to deliver thiol inhibitors across bacterial cell membranes.
Huntington, K. M.; Yi, T.; Wei, Y.; Pei, D. Synthesis and
antibacterial activity of peptide deformylase inhibitors. Bio-
chemistry 2000, 39, 4543. Bacterial intracellular environments
are commonly net-reducing, making disulfide bonds unstable.
Bessette, P. H.; Aslund, F.; Beckwith, J .; Georgiou, G. Efficient
folding of proteins with multiple disulfide bonds in the E. coli
cytoplasm. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 13703-13708.
(37) Krise, J . P.; Stella, V. J . Prodrugs of phosphates, phosphonates,
and phosphinates. Adv. Drug Delivery Rev. 1996, 19, 287-310.
(38) Castelhano, A. L.; Billedeau, R.; Dewdney, N.; Donnelly, S.;
Horne, S.; Kurz, L. J .; Liak, T. J ., Martin, R.; Uppington, R.;
Yuan, Z.; Krantz, A. Novel indolactam-based inhibitors of matrix
metalloproteinases. Bioorg. Med. Chem. Lett. 1995, 5, 1415.
(39) Grunanger, P.; Vita-Finzi, P. Isoxazoles: Part One; J ohn Wiley
& Sons: New York, 1991; p 453.
(40) Hou, X.; Wu, X.; Dai, L.; Cao, B.; Sun, J . Novel N,S- and
N,Se-planar chiral [2,2]-paracyclophane ligands: synthesis and
application in Pd-catalyzed allylic alkylation. J . Chem. Soc.,
Chem. Commun. 2000, 1195-1196.
(41) Zhu, J .; Tumey, L. N.; Pirrung, M. C. Unpublished results.
(42) Zhu, J . Unpublished results.
(43) A possible exception is a recently published study of the SAR of
a
class of histone deacetylase (HDAC) inhibitors. However,
HDAC inhibitors are not a general model for other zinc amidase
inhibitors because it is frequently observed that HDAC inhibitors
do not require a metal binding group for activity. See ref 35.
(44) Wang, W.; Maniar, M.; J ain, R.; J acobs, J .; Trias, J .; Yuan, Z.
A fluorescence-based homogeneous assay for measuring activity
of UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine de-
acetylase. Anal. Biochem. 2001, 290, 338-346.
(45) Bianchi, G.; Gamba-Invernizzi, A.; Gandolfi, R. Isoxazoline
derivatives. VII. Behavior of 5-acyl-2-isoxazolines with bases.
Ring-chain tautomerism of 5-hydroxy-2-pyrrolin-4-ones. J . Chem.
Soc., Perkin Trans. 1 1974, 1757-1763.
(46) Matlack, A. S. Ethylenesulfonamide. J . Org. Chem. 1958, 23,
729.
(47) Sakakura, T.; Huang, X.; Tanaka, M. Hydroformylation-
amidocarbonylation of methylvinylphosphinate. Application to
synthesis of glufosinate. Bull. Chem. Soc. J pn. 1991, 64,
1707-1709.
J M020183V