Aminoalkylphosphinate inhibitors of D-Ala-D-Ala adding enzyme

Article abstract of DOI:10.1039/a704097k

Pseudo-tri- and -tetra-peptide aminoalkylphosphinic acids of general structure X-LyS-PO₂H-Gly-Ala have been synthesised as transition state analogues for D-Ala-D-Ala adding enzyme. The key synthetic step used to assemble the C-terminal dipeptide unit is a modified Arbusov reaction, coupling bromopropionyl-D-alanine methyl ester to a silylated aminoalkylphosphonite. Kinetic assays with the purified E. coli enzyme reveal that the phosphinate analogues act as reversible competitive inhibitors, with K_i values in the range 200-700 μM. Extended analogues mimicking the peptide chain of the UDPMurNAc-L-Ala-γ-D-Glu-m-DAP substrate show increased binding affinity for the enzyme active site. These are the first reported inhibitors for D-Ala-D-Ala adding enzyme.

Full text of DOI:10.1039/a704097k

View Article Online / Journal Homepage / Table of Contents for this issue
Aminoalkylphosphinate inhibitors of D-Ala-D-Ala adding enzyme
David J. Miller,a Stephen M. Hammond,†,b Daniela Anderluzzib and
Timothy D. H. Bugg*,a
a Department of Chemistry, University of Southampton, Highfield, Southampton,
UK SO17 1BJ
b Department of Microbiology, GlaxoWellcome Research, Via Fleming 4, 37100 Verona, Italy
Pseudo-tri- and -tetra-peptide aminoalkylphosphinic acids of general structure X-Lys-PO2H-Gly-Ala have
been synthesised as transition state analogues for D-Ala-D-Ala adding enzyme. The key synthetic step used
to assemble the C-terminal dipeptide unit is a modified Arbusov reaction, coupling bromopropionyl-D-
alanine methyl ester to a silylated aminoalkylphosphonite. Kinetic assays with the purified E. coli enzyme
reveal that the phosphinate analogues act as reversible competitive inhibitors, with Ki values in the range
200–700 ìM. Extended analogues mimicking the peptide chain of the UDPMurNAc-L-Ala-ã-D-Glu-m-
DAP substrate show increased binding affinity for the enzyme active site. These are the first reported
inhibitors for D-Ala-D-Ala adding enzyme.
the synthesis of tri- or tetra-peptide phosphinate analogues.10
Introduction
We report the development of methodology for the synthesis of
a series of X-Lys-PO2H-Gly--Ala analogues, and we report
the properties of these analogues as inhibitors of E. coli D-Ala-
-Ala adding enzyme.
Enzymes involved in bacterial cell wall peptidoglycan bio-
synthesis represent good targets for development of new anti-
bacterial agents, which are of increasing importance in the light
of increasing bacterial antibiotic resistance.1 While there are
many inhibitors of the later extracellular enzymes (β-lactams,
vancomycin, moenomycin), there are relatively few inhibitors
of intracellular enzymes.1
Results
Retrosynthetic disconnections
The cytoplasmic peptidoglycan precursor UDPMurNAc--
Ala-γ--Glu-m-DAP--Ala--Ala is synthesised from UDP-
MurNAc by a series of ATP-dependent ligases.1 The -Ala-
-Ala dipeptide is synthesised by the ATP-dependent -
alanine:-alanine ligase,2 which has been shown to proceed
mechanistically via an acyl phosphate intermediate.3 Amino-
alkylphosphinate analogues for the ensuing tetrahedral tran-
sition state have been synthesised,4 and were found to act as
potent slow-binding inhibitors for this enzyme, although they
showed only limited antibacterial activity.5 The -Ala--Ala
dipeptide is then condensed with UDPMurNAc--Ala-γ--
Glu-m-DAP 1 by the ATP-dependent -Ala--Ala adding
enzyme.2
-Ala--Ala adding enzyme has previously been purified to
homogeneity from E. coli and found to be a 48 kDa monomeric
enzyme with a turnover number of 784 minϪ1.6 The cloning and
sequencing of the E. coli murF gene encoding -Ala--Ala
adding enzyme has allowed the overexpression of enzyme
activity.7–9 By analogy with -alanine:-alanine ligase and
other amino acid adding enzymes, we anticipated that the
mechanism for this enzymatic reaction would also proceed via
an acyl phosphate intermediate 2 (Scheme 1). Attack on the
acyl phosphate 2 by the amino group of -Ala--Ala proceeds
via tetrahedral transition state 3, which would collapse with loss
of phosphate to give the product UDPMurNAc-pentapeptide
4.
Three retrosynthetic disconnections were attempted for the syn-
thesis of phosphinate target 5, as shown in Scheme 2. Discon-
nection of the C-terminal -alanine unit (disconnection a)
leaves a Lys-PO2H-Ala phosphinate, which could be assembled
either via attack of phosphorus onto an imine (route b), or via
attack of an aminophosphinic acid onto a methacrylate ester
(route c). Alternatively, the C-terminal dipeptide unit could be
coupled directly to phosphorus via route d.
Disconnection b—reaction of an imine with an organo-
phosphorus P᎐H compound. 5-Phthalimidopentanal 6 was pre-
pared by reaction of 5-aminopentan-1-ol with phthalic
anhydride, followed by Swern oxidation,11 in 90% overall yield.
Reaction of 5-phthalimidopentanal with tritylamine in reflux-
ing toluene over 4 Å molecular sieves for 4 days gave the trityl
imine 7 in 100% crude yield. Phosphinic acid 8 was prepared
using the general method of Boyd et al.,12 by heating dry
ammonium hypophosphite and hexamethyldisilazane at
100 ЊC, followed by treatment with methyl methacrylate at 0 ЊC,
in 96% yield. The corresponding methyl ester 9 was prepared
from acid 8 by treatment with diazomethane, and was obtained
as a mixture of diastereoisomers (see Scheme 3).
Activation of either phosphinic acid 8 or methyl ester 9 with
hexamethyldisilazane or trimethylsilyl chloride–triethylamine,
followed by treatment with imine 7, gave no reaction, contrary
to literature reports.13,14 No reaction was observed upon add-
ition of Lewis acids, and no reaction was observed between
ester 9 and imine 7 in the presence of sodium methoxide, which
has also been used for activation of phosphinic esters.4 The lack
of reactivity in this case may be due to steric crowding.
Disconnection c—reaction of a 1-aminoalkylphosphinic acid
with an á,â-unsaturated ester. Following the method of Baylis et
al.,15 the 1-aminoalkyl phosphinic acid 10 was prepared from 5-
phthalimidopentanal 6 and (diphenylmethyl)amine, in 36%
yield. In view of the insolubility of 10 in most organic solvents,
it was converted into the benzyloxycarbonyl (Z) derivative 11
by deprotection in refluxing trifluoroacetic acid, followed by
We wished to examine whether aminoalkylphosphinate ana-
logues of transition state 3 might act as potent and selective
inhibitors for -Ala--Ala adding enzyme. Our initial synthetic
target was the tripeptide phosphinate Ac-Lys-PO2H--Ala--
Ala 5. This enzyme is known to accept lysine in position 3 of
the UDPMurNAc-tripeptide substrate in place of meso-
diaminopimelic acid.8 There are very few literature reports of
† Present address: Scottish Biomedical Research, Clydebank, UK G81
2LG.
J. Chem. Soc., Perkin Trans. 1, 1998
131

View Article Online
H3N
H3N
–
–
CO2
CO2
O
O
2–
O–
OPO3
ATP ADP
UDPMurNAc-L-Ala-γ-D-Glu
N
H
UDPMurNAc-L-Ala-γ-D-Glu
N
H
–
H2N
N
H
CO2
O
O
2
O
1
H3N
–
CO2
O
2–
OPO3
UDPMurNAc-L-Ala-γ-D-Glu
N
H
UDPMurNAc-L-Ala-γ-D-Glu-m-DAP-D-Ala-D-Ala
δ+
–
N
N
CO2
O
H2
H
δ–
4
O
Pi
3
H3N
Aminoalkylphosphinate
transition state analogue 5
O
O
P
–
R
N
H
N
H
CO2
O–
O
Scheme 1 Anticipated catalytic mechanism for -Ala--Ala adding enzyme, showing acyl phosphate intermediate 2, transition state 3, and
transition state analogue 5
R22N
R22N
R22N
a
b
b
O
c
O
d
O
P
a
OR5
R1N
H
P
P
R1N
H
OR5
H
R1N
N
CO2R4
H
H
OR3
OR3
O
O
R3O
O
c
d
R22N
R22N
O
O
P
OR5
R1N
H
P
H
R1N
H
N
CO2R4
H
H
OR3
OR3
O
O
Scheme 2 Retrosynthetic strategies for the synthesis of aminoalkylphosphinate analogues
treatment with benzyl chloroformate, in 85% overall yield (see
Scheme 4).
alkaline conditions, however the opening of the phthalimide
protecting group also occurred rapidly under these conditions.
Following treatment of 12 with aqueous potassium hydroxide,
it was found that the phthalimide could be re-closed by treat-
ment with 1,1Ј-carbonyldiimidazole, to give the diacid 13 in
86% yield.
Having assembled the pseudo-dipeptide phosphinic acid 13,
it was anticipated that the C-terminal -alanine residue could
be coupled via standard methods. However, attempts to couple
acid 13 onto -alanine methyl ester using a range of peptide
Coupling of phosphinic acid 11 with methyl methacrylate
was accomplished using the general method of Yiotakis et al.10
After drying in vacuo over phosphorus pentoxide, acid 11 was
pyrolysed at 110 ЊC with hexamethyldisilazane, and the fused
silyl phosphonite intermediate was treated with methyl meth-
acrylate. This successfully gave the 1-aminoalkyl phosphinic
acid 12 as a mixture of diastereoisomers in 94% yield.
Hydrolysis of the methyl ester 12 could be achieved under
132
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
(a),(b)
O
P
O
P
R′
PhthN
PhthN
O
H2N
OH
DCC, HOBT
R′
CO2H
R
R
6
OH
O
(c)
O
R′ = Me, this work
1) H2N-D-Ala-OBz
2) aqueous work-up
NTr
R′ = H (ref. 10)
7
O
P
O
R′
R′
R′′
O
P
O
P
P
CO2H
(d),(e)
N
CO2R′′′
R
R
(f)
H
NH4HPO2
H
OH
OH
H
O
CO2Me
CO2Me
Scheme 5 Hypothetical intramolecular cyclisation of diacid 13 upon
attempted DCC–HOBT coupling with -alanine methyl ester
HO
MeO
8
9
pared from 12, using diazomethane and diphenyldiazomethane
respectively. However, alkaline hydrolysis rapidly hydrolysed
the phosphinic ester in each case. Accordingly, the benzyl carb-
oxylic ester 16 was prepared by use of benzyl methacrylate in a
Michael addition reaction, followed by methylation using tri-
methylsilyldiazomethane, in 75% yield. The only successful
hydrogenation catalyst found for 16 was palladium black and
cyclohexa-1,4-diene as hydrogen donor, which gave the desired
carboxylic acid 17 in low yield. However, attempts to couple 17
with -alanine methyl ester gave only the demethylated diacid
13! The observed demethylation reaction could be rational-
ised by an intramolecular reaction similar to that shown in
Scheme 5.
Disconnection d—reaction of a 1-aminoalkylphosphinic acid
with an electrophilic dipeptide equivalent. In view of the extra-
ordinary difficulty of coupling the C-terminal -alanine residue
onto a pseudo-dipeptide phosphinate, attempts were made to
couple a dipeptide equivalent directly onto a silylated phos-
phonite, using an Arbusov reaction.16 (2R)-3-Bromo-2-methyl-
propionyl--alanine benzyl ester 18 and 3-bromopropionyl--
alanine methyl ester 19 were both synthesised using standard
coupling methods. Silylation of 1-aminoalkylphosphinic acid
11 was carried out by heating with hexamethyldisilazane at
110 ЊC, at which point the appropriate bromide was added.
Bromide 18 gave no reaction, however the less hindered brom-
ide 19 was found to react, and after methylation of the crude
product using diazomethane gave the pseudo-tripeptide phos-
phinate 20 in 46% yield (see Scheme 6).
The behaviour of bromides 18 and 19 in this reaction indi-
cates that this Arbusov-type reaction is sensitive to steric crowd-
ing, presumably since the attacking silylated phosphonite is
very bulky. Although the pseudo-tripeptide phosphinate 20 is
lacking the methyl group of the central -alanine residue of the
original target, it is known that -Ala--Ala adding enzyme
will accept the dipeptide Gly--Ala as a substrate in place of -
Ala--Ala, and in general that the substrate specificity for this
enzyme is more strict for the C-terminal amino acid site.17
Therefore it was decided to elaborate phosphinate 20 to give a
series of X-Lys-PO2H-Gly--Ala analogues which could be
tested as enzyme inhibitors.
Elaboration of the amino terminus of pseudo-tripeptide phos-
phinate 20. Removal of the N-terminal Z protecting group by
catalytic transfer hydrogenation using 10% palladium/charcoal
and ammonium formate gave the free amine 21 in 81% yield.
This free amino terminus was then acylated with groups
of varying length, in order to mimic more closely the
UDPMurNAc--Ala-γ--Glu-m-DAP substrate, as shown in
Scheme 7.
Scheme 3 Synthesis of imine 7 and phosphinic ester 9. Reagents and
conditions: (a), phthalic anhydride, toluene, heat, 99%; (b), i, Me2SO,
(COCl)2, Ϫ78 ЊC; ii, NEt3, 91%; (c), Ph3CNH2, toluene, heat, 100%; (d),
HN(SiMe3)2, 110 ЊC; (e), H2C᎐C(CH3)CO2Me, 96% overall; ( f ),
᎐
CH2N2, Et2O, 100%.
PhthN
PhthN
PhthN
(a)
(b),(c)
O
O
O
Ph
Ph
P
P
N
H
Ph
O
N
H
O
H
H
OH
OH
6
11
10
(d)
PhthN
PhthN
O
O
CO2R2
P
ZN
H
P
ZN
H
CO2R
OR1
OH
14 R1 = Me, R2 = Me
12 R = Me
13 R = H
15 R1 = CHPh2, R2 = Me
16 R1 = Me, R2 = CH2Ph
17 R1 = Me, R2 = H
(e)
Scheme 4 Synthesis of pseudo-dipeptide phosphinate 13. Reagents
and conditions: (a), H3PO2, Ph2CHNH2, H2O–EtOH, 80 ЊC, 36%; (b),
CF3CO2H, heat, 85%; (c) PhCH2OCOCl, H2O, pH 9–9.5, <5 ЊC, 100%;
(d), i, HN(SiMe3)2, 110 ЊC; ii, H2C᎐C(CH3)CO2Me, 94%; (e), i, KOH,
᎐
H2O; ii, 1,1Ј-carbonyldiimidazole, CH2Cl2, 86%.
coupling reagents [dicyclohexylcarbodiimide–hydroxybenzotri-
azole (DCC–HOBT); 1,1Ј-carbonyldiimidazole; isobutyl
chloroformate; benzotriazolyloxytris(pyrrolidino)phosphon-
ium hexafluorophosphate (PyBOP)–HOBT; SOCl2] gave only
recovered starting materials. It was observed that an attempted
DCC coupling gave rise to the expected dicyclohexylurea
(DCU) by-product, indicating that acid activation had taken
place. We hypothesise that the failure of this reaction is due to
the formation of a sterically hindered cyclic anhydride inter-
mediate, which hydrolyses upon work-up, as shown in Scheme
5. Yiotakis et al. report the successful coupling of a pseudo-
dipeptide phosphinate lacking an α-substituent (Scheme 5,
RЈ = H),10 thus it appears that the α-methyl group in 13 prevents
attack of an amine nucleophile with such an intermediate.
Selective protection of the phosphinic acid functional group
of 13 also proved problematical. The methyl phosphinate ester
14 and diphenylmethyl phosphinate ester 15 were both pre-
Treatment of amine 21 with acetic anhydride, triethylamine
and dimethylaminopyridine gave the N-acetyl pseudo-tri-
peptide phosphinate 22 in 91% yield. Treatment of 21 with
glutaric anhydride gave the N-glutaryl pseudo-tripeptide phos-
phinate 23 (ESϩ 582 [MHϩ]), in which the N-glutaryl moiety
mimics the α-carboxy group of -glutamic acid. Finally, amine
J. Chem. Soc., Perkin Trans. 1, 1998
133

View Article Online
ment of the crude acid with sodium borohydride in 6:1 propan-
PhthN
2-ol–water, followed by treatment with acetic acid.19 Deprotec-
tion of the phosphinate methyl ester was achieved by alkaline
hydrolysis, and each of the deprotected phosphinic acids was
purified by reversed phase HPLC. The yields of HPLC purified
material obtained for the N-acetyl pseudo-tripeptide phos-
phinic acid 26, N-glutaryl pseudo-tripeptide phosphinic acid 27
and pseudo-tetrapeptide phosphinic acid 28 are 12, 16 and 2.6%
respectively, from the respective fully protected precursors. The
low yield of 28 is thought to be due to inefficient peptide coup-
ling of amine 21. The N-Z-protected pseudo-tripeptide 20 was
also similarly deprotected to give the N-Z-pseudo-tripeptide
phosphinic acid 25 in 26% yield after HPLC purification. The
structures of the final phosphinates (depicted in the protonated
form) are shown in Scheme 8.
OSiMe3
110 o
C
P:
OSiMe3
No reaction
N
H
CO2Bz
ZN
H
Br
O
18
PhthN
OSiMe3
P:
N
H
CO2Me
ZN
H
H2N
Br
O
OSiMe3
19
(a)–(d)
20
O
(1) 110 o
C
O
(2) CH2N2
P
Ph
O
N
H
N
H
CO2H
PhthN
O
OH
25
H2N
O
P
ZN
H
N
H
CO2Me
(a)–(d)
O
O
OMe
22
O
P
20
N
H
N
H
CO2H
Scheme 6 Reactions of silylated phosphonite derivative of phosphinic
acid 11 with bromides 18 and 19. Pseudo-tripeptide phosphinate 20
isolated in 46% overall yield.
O
OH
26
H2N
PhthN
(a)–(d)
23
O
O
P
O
N
H
N
H
CO2H
P
ZN
H
N
CO2Me
H
CO2H
O
OH
O
OMe
27
H2N
20
PhthN
10% Pd/C,
NH4HCO2
(81%)
O
(a)–(d)
O
P
H
24
N
N
H
N
H
CO2H
O
P
O
CO2H
O
OH
RN
H
N
H
CO2Me
28
O
OMe
Structure for 29:
21 R = H
22 R = COCH3
23 R = CO(CH2)3CO2H
24 R = N-Pr-D-Glu(α-OMe)-
H2N
N
CO2H
N
CO2H
H
H
O
O
Scheme 7 Elaboration of pseudo-tripeptide phosphinate 20 to give
protected phosphinate analogues. Conversion of 21 to 22, 23 and 24
described in text and in Experimental section.
D-Ala-D-Ala
Isobutyryl-D-Ala 29
Scheme 8 Deprotection of phosphinate target analogues. Reagents
and conditions: (a), (Bu3Sn)2O, THF, heat; (b), NaBH4, PriOH/H2O
(6:1); (c), AcOH, heat; (d), NaOH, H2O. Overall yields after HPLC
purification: 25, 26%; 26, 12%; 27, 16%; 28, 2.6%.
21 was coupled to N-propionyl--Glu-(α-OMe)-OH to give a
pseudo-tetrapeptide phosphinate 24 (ESϩ689 [MNaϩ]).
Deprotection of the phthaloyl group by hydrazinolysis was
found to form the C-terminal acid hydrazide (RCONHNH2).
Therefore the C-terminal methyl ester of the protected phos-
phinates was first hydrolysed using tributyltin oxide (5 equiv.).18
Removal of the phthaloyl group was then achieved by treat-
Purification of Escherichia coli D-Ala-D-Ala adding enzyme
The murF gene encoding -Ala--Ala adding enzyme in
Escherichia coli has been identified amongst a cluster of genes
134
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
Table 1 Purification of Escherichia coli -Ala--Ala adding enzymea from JM105/pTB3
Volume
(ml)
Proteinb
(mg/ml)
Total protein
(mg)
Activityc
(units/ml)
Sp. activity
(units/mg)
Yield
(%)
Purification
(x-fold)
Step
Crude extract
(NH4)2SO4 ppt.
AcA54 gel filtration
Q Sepharose FPLC
42.0
13.2
60.0
60.0
29.0
45.0
1.87
0.35
1218
594
112
33.0
94.4
14.2
7.2
1.14
2.10
7.6
100
90
61
1.0
1.84
6.7
21.0
20.6
31
18.0
a One unit of enzyme activity is defined as the activity required to convert 1 µmole of substrate per minute. b Protein concentration determined by the
method of Bradford.20 c Enzyme activity determined by Pi release assay.6
Table 2 Kinetic data for -Ala--Ala adding enzyme substrates and
inhibitors. Kinetic assays were carried out as described in the Experi-
mental section. Ki values for phosphinates 26–28 refer to concentrations
of diastereomeric mixture of phosphinate in solution (1:1 mixture of
two diastereoisomers).
involved in peptidoglycan biosynthesis at minute 2 of the E. coli
chromosome.7 The murF gene has been overexpressed in plas-
mid pTB3 under the control of the tac promoter to approxi-
mately 6% of soluble cell protein.9
-Ala--Ala adding enzyme was purified from extracts of
JM105/pTB3 using ammonium sulfate precipitation, gel filtra-
tion chromatography and Q Sepharose FPLC anion exchange
chromatography (see Table 1). The specific activity of the puri-
fied enzyme was 20.6 units mϪ1, and the protein appeared as a
band at 48 kDa by SDS-PAGE.
Substrate/inhibitor
KM/µ
Ki/µ
-Ala--Ala
Isobutyryl--Ala 29
N-Z-Lys-PO2H-Gly--Ala 25
N-Ac-Lys-PO2H-Gly--Ala 26
N-glutaryl-Lys-PO2H-Gly--Ala 27
N-Pr-γ--Glu-Lys-PO2H-Gly--Ala 28
60
—
—
—
—
—
—
16 800
—
700
200
200
Steady state kinetic assays of purified enzyme were carried
out using a phosphate release stopped assay.6 In this assay a
mixture of enzyme, UDPMurNAc-tripeptide, -Ala--Ala,
ATP and MgCl2 in 50 m Tris pH 8.6 is incubated at 20 ЊC. At
selected time intervals aliquots are removed, and the release of
phosphate quantitated using the method of Lanzetta et al.21
Using this assay the amount of phosphate release versus time
was measured over a 10 min assay, and initial rates deduced. In
our hands we were unable to obtain consistent rate measure-
ments using the reported continuous coupled assay involving
pyruvate kinase and lactate dehydrogenase,6,8 although this
coupled assay works well for the preceding enzyme -
alanine:-alanine ligase.5
Using the phosphate release assay, steady state kinetic
parameters were measured for the purified enzyme. The enzyme
was found to obey Michaelis–Menten kinetics, and a KM value
of 60 ( 5) µ was measured for -Ala--Ala under pseudo-first
order conditions (100 µ UDPMurNAc-tripeptide, 1 m
ATP). This value is lower than reported KM values of 220 µ6
and 208 µ8 determined using the continuous coupled assay,
but was verified by several independent analyses. A KM value of
220 ( 10) µ was measured for co-substrate ATP. The sensitiv-
ity of the phosphate release assay (1–10 nmol Pi) was insuffi-
cient to measure the KM for UDPMurNAc-tripeptide, implying
that KM < 20 µ (reported value 11 µ, determined by radio-
chemical assay22).
500 µ), and a Ki value of 200 ( 20) µ deduced using a Dixon
plot. These data are summarised in Table 2.
The measured Ki values are somewhat larger than the KM
value for -Ala--Ala, which contains the α-amino group
absent from the phosphinate inhibitors. In order to assess the
importance of the α-amino group of -Ala--Ala for substrate
binding, the analogue isobutyryl--alanine 29 containing a
methyl group in place of the α-amino group was synthesised
and tested as an inhibitor. Analogue 29 was found to act as a
reversible competitive inhibitor, with a Ki value of 16.8 m.
Thus, the α-amino group does appear to form a significant
binding interaction with the enzyme active site. However,
introduction of the Lys-PO2H-phosphinyl moiety has given
>20-fold tighter binding, relative to 29.
In contrast to the phosphinate inhibitors of -alanine:-
alanine ligase,5 pre-incubation of enzyme and ATP with phos-
phinates 26–28 gave no increase in inhibition. Thus, these
analogues are not time-dependent inhibitors. Attempts to
improve the potency of inhibition by co-addition of 1 m
UDP or UDPGlcNAc, in order to occupy the UDPMurNAc-
tripeptide binding site, also had no effect on enzyme inhibition.
The antibacterial action of phosphinates 26–28 was tested with
cultures of Escherichia coli K12 and Bacillus subtilis W23, and
no retardation of growth was observed at 100 µg mlϪ1 final
concentration.
Inhibition of D-Ala-D-Ala adding enzyme by phosphinates 25–28
A series of preliminary enzyme assays was carried out at 100 µ
-Ala--Ala, in the presence of 1 m concentrations of phos-
phinates 25–28. Under these conditions the N-Z-phosphinate
25 showed no inhibition, however >50% inhibition was
observed for phosphinates 26–28.
Ki values for phosphinates 26 and 27 were determined by
measuring apparent KMЈ values for -Ala--Ala in the presence
of fixed concentrations of inhibitor. Kinetic analysis using
Eadie/Hofstee plots revealed that the observed inhibition was
reversible and competitive with respect to -Ala--Ala. Values
Discussion
The aminoalkylphosphinate analogues described in this paper
represent the first synthetic inhibitors for -Ala--Ala adding
enzyme, a peptidoglycan biosynthetic enzyme found in all bac-
teria. Pseudo-dipeptide phosphinate inhibitors have previously
been synthesised for -alanine:-alanine ligase4 and the -Glu
adding enzyme,24 however there are very few literature reports
of pseudo-tri- or -tetra-peptide phosphinate analogues.10
Our synthetic work has confirmed that the use of silylated
phosphonites for Michael addition or Arbusov-type displace-
ment reactions, after the work of Regan and co-workers12,13,16
and Yiotakis et al.,10 is a powerful method for the assembly of
complex phosphinate skeletons. However, these reactions pro-
ceed only under harsh reaction conditions, and appear to be
sensitive to steric factors.
of Ki were then calculated using the equation Ki = [I]/[(KMЈ
/
KM) Ϫ 1], giving values of 700 ( 50) µ and 200 ( 20) µ
respectively for the N-acyl phosphinate 26 and the N-glutaryl
phosphinate 27. Inhibition by the pseudo-tetrapeptide phos-
phinate 28 was complicated by nanomolar background levels of
inorganic phosphate in the purified inhibitor, leading to
increased scatter at low substrate concentrations. Therefore,
assays were carried out at a fixed substrate concentration (100
µ -Ala--Ala), and variable inhibitor concentrations (0, 200,
The inability to couple a further amino acid onto the carboxy
terminus of a pseudo-dipeptide phosphinic acid 13 represents a
J. Chem. Soc., Perkin Trans. 1, 1998
135

View Article Online
severe limitation in the synthesis of larger phosphinates, and
should be considered by other workers in this area. In this case
we have circumvented this problem by attaching a C-terminal
dipeptide unit as bromopropionyl--alanine methyl ester, using
an Arbusov reaction. However, this approach only allows for
the introduction of a glycine analogue immediately adjacent to
the phosphinate. During the completion of this work Yiotakis
et al. have published a method for the protection of a phos-
phinic acid group using an adamantyl protecting group, which
provides an alternative route for the assembly of higher
phosphinates.23
Kinetic assays of these analogues has revealed that they act
as inhibitors of E. coli D-Ala--Ala adding enzyme, with Ki
values in the range 200–700 µ. Since these Ki values are com-
parable with, but not lower than, the KM value for -Ala--Ala
(60 µ), it seems that these analogues have not fully realised
the target of resembling the transition state for the enzymatic
reaction. However, comparison with the carbon analogue
isobutyryl--Ala (Ki 16.8 m) reveals that the presence of the
phosphinic acid unit has increased the binding efficiency by
>20-fold.
It is also noteworthy that increasing the length of the amino-
terminal peptide chain significantly increases the effectiveness
of inhibition, presumably by mimicking the UDPMurNAc-
tripeptide chain. The N-glutaryl phosphinate 27 and the
pseudo-tetrapeptide 28 have 3-fold lower Ki values than the N-
acetyl phosphinate 26, suggesting that the α-carboxylate of the
-Glu residue forms a favourable binding interaction with the
enzyme, although the presence of the -Ala--Glu amide bond
in 28 gives no further increase in binding affinity. It has been
found recently that the kinetic mechanism of this enzyme pro-
ceeds via initial binding of the UDPMurNAc-tripeptide sub-
strate.8 Therefore, it seems likely that in order to achieve sub-
micromolar levels of inhibition via this approach, a significant
amount of the UDPMurNAc-tripeptide skeleton would need
to be included in the inhibitor design. Similarly, Tanner et al.
have shown that most effective inhibition of the E. coli D-Glu
adding enzyme was achieved using a pseudo-dipeptide phos-
phinate analogue covalently linked to uridine monophos-
phate.24
methylpropionic acid was prepared by the method of Salomon
et al.18 and N-butoxycarbonyl--glutamic acid α-methyl ester
γ-benzyl ester‡ was prepared by the method of Manesis and
Goodman.27
5-Phthalimidopentanal 6 was prepared by reaction of 5-
aminopentan-1-ol with phthalic anhydride,28 to give 5-phthal-
imidopentan-1-ol in 99% yield, followed by oxidation using the
method of Swern et al. in 91% yield11 {νmax(liquid film)/cmϪ1
2724 (w), 1771 (s), 1710 (s); δH(300 MHz, CDCl3) 1.60–1.80
(4H, m), 2.25 (2H, t, J 7), 3.62 (2H, t, J 7), 7.60–7.90 (4H, m,
C6H4), 9.73 (1H, s); δC(75 MHz, CDCl3) 19.09, 27.83, 37.32,
43.07, 123.1, 133.9, 131.9, 168.3, 201.9; m/z (ESϩ) 463 (71%,
[2M ϩ H]ϩ), 485 (100, [2M ϩ Na]ϩ)}. 5-Phthalimido-N-
tritylpentanimine 7 was prepared by reaction of 5-phthal-
imidopentanal with tritylamine in refluxing toluene over 4 Å
molecular sieves for 4 days, in quantitative yield.25
2-Methoxycarbonylpropylphosphinic acid 8 was prepared
from ammonium hypophosphite and methyl methacrylate
using the method of Boyd et al.,12 in 96% yield {νmax(liquid film)/
cmϪ1 3403 (br, m), 2340 (w), 1731 (s); δH(300 MHz, CDCl3) 1.32
(3H, d, J 6.5), 1.83–2.27 (2H, m), 2.85–3.16 (1H, m), 3.70 (3H,
s), 7.19 (1H, d, JH᎐P 560, PH); δC(75 MHz, CDCl3) 19.10, 33.02
(d, JC᎐P 93, PCH2), 33.83, 52.33, 175.7; δP(121 MHz, CDCl3)
34.7 ppm downfield from 85% H3PO4; m/z (ESϪ) 165 (10%,
MϪ)}.
5-Phthalimido-1-[(diphenylmethyl)amino]pentylphosphinic
acid 10 was prepared from 5-phthalimidopentanal 6 by the
method of Baylis et al.15 (mp 134–135 ЊC, lit.,15 134–135 ЊC). N-
Isobutyryl--alanine benzyl ester was prepared by the coupling
of isobutyric acid and -Ala-benzyl ester using hydroxybenzo-
triazole and dicyclohexylcarbodiimide as coupling reagents in
dichloromethane. Debenzylation was carried out by hydrogen-
ation (stirring overnight in ethanol with 10% Pd/charcoal under
a hydrogen atmosphere for 16 h at 25 ЊC). After filtration
through Celite and removal of solvent under reduced pressure,
the deprotected compound was recrystallised from ethyl
acetate–hexane {δH(500 MHz, D2O) 1.06 [6H, d, (CH3)2CH],
1.37 (3H, d, J 8, CH3CH), 2.51 [1H, septet, J 7, (CH3)2CH],
4.28 (1H, q, J 8, NHCH); m/z (FAB) 160 [M ϩ H]; HRMS
(FAB) MHϩ, 160.0974. C7H14NO3 requires 160.0970}.
An alternative explanation for the decreased activity of these
phosphinate analogues compared with the phosphinate inhibi-
tors of -alanine:-alanine ligase4,5 is that the latter enzyme is
strongly inhibited by the product -Ala--Ala,3 and therefore
binds dipeptides tightly. In contrast, -Ala--Ala adding
enzyme is not inhibited by its product UDPMurNAc-
pentapeptide.6,22 The lack of antibacterial activity of phos-
phinates 26–28 could either be due to inefficient transport into
the bacterial cell, or the inability to accumulate inhibitor con-
centrations inside the cell up to the observed values of Ki.
Therefore, the potential of -Ala--Ala adding enzyme as an
antibacterial target remains to be confirmed.
2-Methoxycarbonylpropylphosphinic acid methyl ester 9
To a stirred solution of phosphinic acid 8 (1.02 g, 6.14 mmol)
was added a solution of diazomethane in diethyl ether until
evolution of nitrogen ceased and the yellow colour of dia-
zomethane remained permanent. The solution was stirred
overnight at room temperature to allow excess diazomethane to
evaporate and the solvent was removed under reduced pressure
to give the title compound as a colourless oil (1.10 g, 100%);
νmax(liquid film)/cmϪ1 2338 (w), 1735 (s); δH(300 MHz, CDCl3)
1.29 and 1.30 (3H, 2 d, J 7), 1.75–1.90 and 2.10–2.30 (2H, m,
PCH2), 2.88 (1H, tq, J 7 and 14, CH2CHCH3), 3.68 (3H, s,
CO2CH3), 3.73 and 3.77 (3H, 2 s, POCH3), 7.13 and 7.14 (1H, 2
d, JH᎐P 550, PH); δC(75 MHz, CDCl3) 18.98, 19.20, 32.29 (d,
JC᎐P 93, PCH2), 33.53, 33.89, 52.23, 52.34, 52.93, 53.07, 175.4;
δP(121 MHz, CDCl3) 38.9 and 39.1 (downfield from 85%
H3PO4); m/z (ESϩ) 181 (100%, [M ϩ H]ϩ).
Experimental
General
Nuclear magnetic resonance spectra were recorded on either a
Bruker AM 300 Fourier transform spectrometer (300 MHz) or
a Bruker AM 360 Fourier transform spectrometer. J Values are
given in Hz. Ultraviolet–visible spectra were recorded on a
Cary-1 UV-visible spectrophotometer. Infrared spectra were
recorded on a 1600 series Perkin-Elmer FTIR spectrometer.
Mass spectra were recorded on a VG-70-250 mass spectrometer
in Electron Impact (EI), Chemical Ionisation (CI) or Fast
Atom Bombardment (FAB) mode or a VG Platform Quad-
rupole Electrospray Ionisation mass spectrometer (ES).
All chemicals were supplied by Sigma-Aldrich Chemical Co.
Tritylamine (triphenylmethylamine) was prepared by the
method of Soroka and Sygmunt.25 -Alanine methyl ester was
prepared by the method of Miller et al.,26 (2R)-3-Bromo-2-
5-Phthalimido-1-aminopentylphosphinic acid
A stirred solution of 5-phthalimido-1-(diphenylmethyl)amino-
pentylphosphinic acid15 10 (1.38 g, 2.98 mmol) in trifluoroace-
tic acid (15 ml) was heated under reflux under an atmosphere of
dry nitrogen for 1 h. After cooling to room temperature, tri-
fluoroacetic acid was removed under reduced pressure and the
‡ In this name α-methyl ester corresponds to esterification of the carb-
oxy group attached to the α-carbon, and γ-benzyl ester corresponds
to esterification of the carboxy group attached to the γ-carbon. The
IUPAC name for this compound is N-butoxycarbonyl--glutamic acid
1-methyl ester 5-benzyl ester.
136
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
residue partitioned between ethyl acetate (50 ml) and water (50
ml). The separated aqueous layer was further extracted with
ethyl acetate (2 × 50 ml) and then evaporated in vacuo. The title
compound was obtained as a white crystalline solid after recrys-
tallisation from water–acetone (0.752 g, 85%), mp 208–210 ЊC;
νmax(KBr disc)/cmϪ1 3300 (s, br), 2346 (m), 1773 (s), 1710 (s);
δH(300 MHz, D2O) 1.28–1.90 (6H, m), 3.04 (1H, m, H2NCHP),
3.56 (2H, t, J 7, PhthNCH2), 6.91 (1H, d, JH᎐P 530, PH), 7.69
(4H, br s, C6H4); δC(75 MHz, D2O) 25.06 (d, JC᎐P 8.5,
PCHCH2), 28.25, 29.81, 39.70, 52.71 (d, JC᎐P 92.5, H2NCHP),
125.6, 137.0, 133.6, 173.0; δP(121 MHz, D2O) 20.4 (downfield
from 85% H3PO4); m/z (FAB) 297 (40%, [M ϩ H]ϩ), 319 (22,
[M ϩ Na]ϩ).
52.06, 67.04, 123.1, 127.9, 128.1, 128.4, 132.0, 133.8, 136.2,
156.3, 168.3, 175.8; δP(121 MHz, CDCl3) 53.67 (downfield from
85% H3PO4); m/z (NH3, CIϩ) 531 (8%, [M ϩ H]ϩ, 513 (12,
[M ϩ H Ϫ H2O]ϩ).
Methyl 3-[(1-benzyloxycarbonylamino-5-phthalimidopentyl)-
(methoxy)phosphonyl]-2-methylpropionate 14
To a stirred solution of pseudo-dipeptide phosphinic acid 12
(516 mg, 0.974 mmol) in chloroform (20 ml) was added drop-
wise a solution of diazomethane in diethyl ether until the evolu-
tion of nitrogen ceased and the yellow colour of diazomethane
was no longer discharged. The solution was then left to stand
for 16 h to allow excess diazomethane to evaporate. The solvent
was removed under reduced pressure to give the title compound
as a colourless oil (530 mg, 100%); νmax(CHCl3 solution)/cmϪ1
3227 (m), 1771 (s), 1711 (s, br), 1529 (m); δH(300 MHz, CDCl3)
1.20 and 1.25 (3H, 2d, J 8, CH3CH), 1.36–2.35 (8H, m), 2.87
(1H, ddd, J 7.5, 13.5, 24.5, CH3CHCH2), 3.66–4.03 (9H, m),
4.97–5.53 (3H, m), 7.32 (5H, br s, C6H5CH2), 7.65–7.81 (4H, m,
C6H4); δC(75 MHz, CDCl3) 18.60, 18.71, 18.91, 19.03, 27.73 (d,
JC᎐P 12, PCHCH2), 27.20, 27.74, 29.00 (d, JC᎐P 88.5, PCH2CH),
33.60, 37.06, 49.51, 49.75 (d, JC᎐P 106, NHCHP), 51.95, 123.0,
127.8, 127.9, 128.0, 128.1, 128.4, 131.9, 133.8, 136.1, 156.2,
168.3, 175.7; δP(121 MHz, CDCl3) 52.65, 52.93, 52.98, 53.73
(downfield from 85% H3PO4).
5-Phthalimido-1-(benzyloxycarbonylamino)pentylphosphinic
acid 11
To a stirred solution of 5-phthalimido-1-aminopentylphos-
phinic acid (3.87 g, 13.1 mmol) in water (50 ml) at 0 ЊC was
added 4  sodium hydroxide solution to pH 9.5. A solution of
benzyl chloroformate (2.46 g, 14.4 mmol) in dioxane (50 ml)
was then added dropwise over 10 min whilst maintaining the
solution in the range pH 9–9.5. The solution was then stirred at
0 ЊC for a further 4 h, maintaining the solution in the range pH
9–9.5. After warming to room temperature, the solution was
extracted with diethyl ether (3 × 100 ml) and the separated
aqueous layer was acidified to pH 1 by dropwise addition of 2 
hydrochloric acid. The resulting white precipitate was extracted
with ethyl acetate (3 × 100 ml) and the combined organic
extracts were washed with brine (300 ml). The organic solution
was then dried (MgSO4), filtered and evaporated to give the title
compound as a white foam (5.68 g, 100%); νmax(CH2Cl2
solution)/cmϪ1 2354 (w), 1771 (w), 1712 (s); δH(300 MHz,
CDCl3) 1.31–1.91 (6H, m), 3.63 (2H, t, J 7, PhthNCH2), 3.85–
3.95 (1H, m, NHCHP), 5.04 (2H, s, C6H5CH2), 5.62 (1H, d, J 9,
NHCHP), 6.96 (1H, d, JH᎐P 567, PH), 7.29 (5H, br s, C6H5CH2),
7.63–7.78 (4H, m, C6H4); δC(75 MHz, CDCl3) 22.70, 26.15,
27.85, 37.22, 49.97 (d, JC᎐P 100, NHCHP), 67.03, 123.1, 133.9,
125.7, 128.0, 128.4, 134.0, 136.0, 156.7, 168.4; δP(121 MHz,
CDCl3) 32.6 (downfield from 85% H3PO4); m/z (ESϩ) 431
(100%, [M ϩ H]ϩ), 453 (34, [M ϩ Na]ϩ); HRMS (FAB) MHϩ,
431.1373. C21H24N2O6P requires 431.1372.
Benzyl 3-[(1-benzyloxycarbonylamino-5-phthalimidopentyl)-
(methoxy)phosphonyl]-2-methylpropionate 16
A stirred mixture of phosphinic acid 11 (5 g, 11.6 mmol) and
hexamethyldisilazane (3.70 ml, 17.4 mmol) was heated under an
atmosphere of dry nitrogen at 110 ЊC for 40 min. After cooling
the fused mixture to 85 ЊC, benzyl methacrylate (2.36 ml, 14.0
mmol) was added dropwise over 10 min and the resulting mix-
ture stirred for 1 h at 85 ЊC. After cooling to 70 ЊC, ethanol (60
ml) was added and the resulting solution stirred whilst slowly
cooling to room temperature. The solvent was removed under
reduced pressure and the residue redissolved in 8% aqueous
sodium hydrogen carbonate (80 ml). This solution was washed
with diethyl ether (3 × 80 ml) and then acidified to pH 1 by
dropwise addition of 10% aqueous hydrochloric acid. The
resulting white precipitate was extracted into ethyl acetate
(3 × 100 ml) and the pooled organic extracts were washed with
brine (300 ml). The organic solution was then dried (Na2SO4),
filtered and evaporated to give a white foam. This was redis-
solved in 1:1 (v/v) acetonitrile–methanol (100 ml) and to this
stirred solution under a dry nitrogen atmosphere was added
trimethylsilyldiazomethane (2 , solution in hexanes) until evo-
lution of nitrogen ceased and the yellow colour of trimethyl-
silyldiazomethane was permanent. The solution was then
stirred for 2 h at room temperature. The solvent was removed
under reduced pressure and the residue purified by flash
chromatography on silica gel, eluting with a solvent gradient
[from 100% dichloromethane to 24:1 (v/v) dichloromethane–
methanol] to give the title compound as a white foam (5.37 g,
75%); νmax(KBr disc)/cmϪ1 3400 (w), 1772 (m), 1712 (s, br), 1508
(m); δH(500 MHz, CDCl3) 1.27 and 1.32 (3H, 2d, J 7.5,
CH3CH), 1.40–2.00 (7H, m), 2.32 (1H, m, CH3CHCH2), 2.85
(1H, m), 3.60–4.12 (6H, m), 4.80–5.20 (5H, m), 7.32–7.36
(10H, m), 7.69–7.83 (4H, m); δC(75 MHz, CDCl3) 18.86, 18.97,
19.24, 19.36, 22.98 (d, JC᎐P 11.5, PCHCH2), 27.50, 28.02, 28.65,
29.82, 34.06, 37.33, 50.05, 50.27 (d, JC᎐P 105, NHCHP), 51.95,
52.19, 66.81, 67.23, 123.3, 128.1, 128.2, 128.3, 128.4, 128.7,
132.2, 134.0, 136.0, 136.4, 156.5, 168.6, 175.3, 175.4; δP(200
MHz, CDCl3) 48.76, 48.86, 50.04 and 50.73 (downfield from
85% H3PO4); m/z (FABϩ) 621 (30%, [M ϩ H]ϩ), 643 (10,
[M ϩ Na]ϩ).
Methyl 3-[(1-benzyloxycarbonylamino-5-phthalimidopentyl)-
(hydroxy)phosphonyl]-2-methylpropionate 12
A stirred mixture of phosphinic acid 11 (480 mg, 1.12 mmol)
and hexamethyldisilazane (0.35 ml, 1.68 mmol) was heated at
110 ЊC under an atmosphere of dry nitrogen for 40 min. After
cooling the fused mixture to 90 ЊC, methyl methacrylate (0.15
ml, 1.34 mmol) was added dropwise over 5 min and the mixture
stirred at 98 ЊC for 25 min. The mixture was then allowed to
cool to 70 ЊC and ethanol (10 ml) was added. The resulting
solution was stirred at this temperature for 2 min and then
allowed to slowly cool to room temperature. The solvent was
removed under reduced pressure and the residue redissolved in
10% aqueous sodium hydrogen carbonate (15 ml). This solution
was washed with diethyl ether (3 × 15 ml) and the separated
aqueous layer was acidified to pH 1 with 2  aqueous hydro-
chloric acid. The resulting white precipitate was extracted with
ethyl acetate (4 × 40 ml) and the combined organic extracts
were washed with brine (200 ml). The organic solution was then
dried (MgSO4), filtered and evaporated to give the title com-
pound as a white foam (560 mg, 94%); νmax(KBr disc)/cmϪ1
3335 (m, br), 1772 (s), 1707 (s, br), 1527 (s); δH(300 MHz,
CDCl3) 1.24 (3H, d, J 6.5, CH3CH), 1.40–2.18 (8H, m), 2.83–
2.93 (1H, m, CH3CH), 3.60–3.65 (2H, m, PhthNCH2), 3.65
(3H, s, CO2CH3), 3.87–3.98 (1H, m, NHCHP), 5.06 (2H, ABq,
J 12.5, C6H5CH2), 5.48 (1H, d, J 9.5, NHCHP), 7.31 (5H, br s,
C6H5CH2), 7.63–7.78 (4H, m, C6H4); δC(75 MHz, CDCl3) 18.70,
18.82, 22.77 (d, JC᎐P 11.5, PCHCH2), 26.89, 27.77, 29.87 (d,
JC᎐P 91.5, PCH2), 33.45, 37.21, 49.98 (d, JC᎐P 105, NHCHP),
Methyl 3-[(1-benzyloxycarbonylamino-5-phthalimidopentyl)-
(benzyloxy)phosphonyl]-2-methylpropionate 15
To a stirred mixture of benzophenone hydrazone (393 mg, 2.00
J. Chem. Soc., Perkin Trans. 1, 1998
137

View Article Online
mmol) and anhydrous magnesium sulfate (200 mg) in dichlo-
romethane (10 ml) was added manganese dioxide (869 mg, 10.0
mmol) and the mixture stirred for 1.5 h at room temperature.
The solution was filtered to give a solution of diphenyl-
diazomethane. This solution was added dropwise to a stirred
solution of pseudo-dipeptide phosphinic acid 12 (530 mg, 1.00
mmol) in dichloromethane (5 ml). The resulting solution was
then heated under reflux for 2 h. The solvent was removed
under reduced pressure and the residue was purified by flash
chromatography on silica gel [eluting with 1:1 (v/v) ethyl
acetate–light petroleum] to give the title compound as a colour-
less oil (335 mg, 48%); δH(300 MHz, CDCl3) 1.01–2.20 (10H,
m), 2.23–2.39 (1H, m), 2.56–2.80 (1H, m, CH3CH), 3.49–3.61
(5H, m), 3.86–4.06 (1H, m, NHCHP), 4.92–5.11 and 5.28–5.29
(3H, m, C6H5CH2 and NHCHP), 6.53 and 6.57 [1H, 2s,
(C6H5)2CH], 7.32–7.40 (15H, m, 3 × C6H5), 7.67–7.83 (4H, m,
C6H4); δC(75 MHz, CDCl3) 18.96, 19.07, 19.35, 19.47, 23.00 (d,
JC᎐P 12, PCHCH2), 27.38, 27.97, 30.61 (d, JC᎐P 86.5, PCH2CH),
33.84, 37.45, 50.50, 50.86 (2d, JC᎐P 105, NHCHP), 52.15, 67.23,
76.12, 78.01, 123.33, 126.7, 127.0, 127.3, 127.5, 127.7, 128.1,
128.2, 128.3, 128.4, 128.5, 128.7, 128.8, 129.0, 132.3, 134.0,
136.5, 140.8, 144.4, 156.5, 168.5, 175.7, 175.9; δP(121 MHz,
CDCl3) 53.29, 53.45, 53.55, 53.68 (downfield from 85% H3PO4).
ethyl acetate (10 ml). The separated organic layer was washed
with saturated aqueous sodium hydrogen carbonate (10 ml), 5%
hydrochloric acid (10 ml) and brine (10 ml). The organic layer
was then dried (MgSO4), filtered and evaporated. The residue
was purified by flash chromatography on silica gel [eluting with
2:1 (v/v) light petroleum–ethyl acetate] to give the title com-
pound as a white crystalline solid which was recrystallised from
ethyl acetate–hexane (123 mg, 70%), mp 73–74 ЊC; νmax(Nujol
mull)/cmϪ1 3325 (s), 1726 (s), 1655, 1531; δH(500 MHz, CDCl3)
1.27 (3H, d, J 5), 1.47 (3H, d, J 7.5), 2.65 (1H, m), 3.38 (1H, dd,
J 5.5, 10), 3.61 (1H, dd, J 8.5, 10), 4.69 (1H, quintet, J 7.5), 5.20
(2H, ABq, J 12), 6.14 (1H, d, J 6.5), 7.38 (5H, m); m/z (FAB)
329 (100%, [M ϩ H]ϩ, 248 (24, [M Ϫ Br]ϩ).
3-Bromopropionyl-D-alanine methyl ester 19
To a stirred solution of 3-bromopropionic acid (3.30 g, 21.6
mmol) in dry tetrahydrofuran (30 ml) under an atmosphere of
dry nitrogen was added N-methylmorpholine (2.36 ml, 21.6
mmol) and the solution cooled to Ϫ15 ЊC. Isobutyl chlorofor-
mate (2.80 ml, 21.6 mmol) was then added and the solution
stirred for 15 min. To this solution was added a solution of (R)-
alanine methyl ester hydrochloride (3.00 g, 21.6 mmol) and N-
methylmorpholine (2.36 ml, 21.6 mmol) in dry N,N-
dimethylformamide (30 ml) plus washings via a cannula. The
solution was stirred for a further 15 min at Ϫ15 ЊC and then
slowly allowed to warm to room temperature. The solution was
filtered and the solvent removed in vacuo. The residue was par-
titioned between ethyl acetate (50 ml) and water (50 ml). The
separated organic layer was washed with saturated aqueous
sodium hydrogen carbonate (50 ml), 2  hydrochloric acid (50
ml) and brine (50 ml). The organic layer was then dried
(MgSO4), filtered and evaporated to give a yellow oil that was
purified by flash chromatography on silica gel [eluting with 1:1
(v/v) ethyl acetate–petroleum ether] to give the title compound
as a white crystalline solid that was recrystallised from ethyl
acetate–hexane (2.29 g, 45%), mp 70–72 ЊC (lit.,29 54–56 ЊC);
νmax(CH2Cl2 solution)/cmϪ1 3417 (w), 1740 (s), 1674 (s), 1515
(m); δH(300 MHz, CDCl3) 1.43 (3H, d, J 7), 2.72–2.88 (2H, m),
3.57–3.69 (2H, m), 3.76 (3H, s), 4.63 (1H, quintet, J 7), 6.28
(1H, br); δC(75 MHz, CDCl3) 18.65, 27.21, 39.60, 48.31, 52.73,
169.3, 173.5; m/z (ESϩ) 238/240 (12%, [M ϩ H]ϩ), 260/262 (70,
[M ϩ Na]ϩ).
3-[(1-Benzyloxycarbonylamino-5-phthalimidopentyl)(hydroxy)-
phosphonyl]-2-methylpropionic acid 13
A solution of pseudo-dipeptide phosphinic acid methyl ester 12
(2.50 g, 4.72 mmol) in 1  aqueous potassium hydroxide (45 ml)
was stirred for 1 h at room temperature. The solution was acid-
ified to pH 1 by dropwise addition of 2  hydrochloric acid and
the resulting white precipitate was extracted with ethyl acetate
(3 × 50 ml). The combined organic extracts were dried
(MgSO4), filtered and evaporated to give the ring-opened
phthalamic acid as a white foam (2.36 g, 94%).
To a stirred solution of the phthalamic acid (1.90 g, 3.56
mmol) in dry tetrahydrofuran (20 ml) was added 1,1Ј-
carbonyldiimidazole (1.73 g, 10.7 mmol) and the solution
stirred at room temperature for 1 h. The solvent was removed
under reduced pressure and the residue redissolved in 10%
aqueous sodium hydrogen carbonate (20 ml). This solution was
washed with diethyl ether (3 × 20 ml) and then acidified to pH 1
by dropwise addition of 2  hydrochloric acid. The resulting
white precipitate was extracted with ethyl acetate (3 × 40 ml)
and the combined organic extracts were washed with brine (120
ml). The organic solution was then dried (MgSO4), filtered and
evaporated to give the title compound as a white foam (1.58 g,
86%); νmax/cmϪ1 3322 (s, br), 1770 (s), 1713 (s, br), 1530; δH(300
MHz, CDCl3) 1.18 (3H, d, J 6, CH3CH), 1.33–2.32 (8H, m),
2.73–2.95 (1H, m, CH3CH), 3.61 (2H, t, J 7, PhthNCH2), 3.85
(1H, m, NHCHP), 4.97–5.13 (2H, m, C6H5CH2), 5.80–5.98
(1H, br, NHCHP), 7.23–7.31 (5H, br s, C6H5), 7.61–7.76 (4H,
m, C6H4); δC(75 MHz, CDCl3) 19.23, 23.11, 26.93, 27.99, 30.00
(d, JC᎐P 88, PCHCH2), 34.12, 37.55, 50.30 (d, JC᎐P 105,
NHCHP), 67.26, 123.3, 128.0, 128.1, 128.6, 132.2, 134.0, 136.4,
156.9, 168.6, 178.7; δP(121 MHz, CDCl3) 52.01 (downfield from
85% H3PO4); m/z (ESϩ) 517 (100%, [M ϩ H]ϩ).
Methyl (2R)-2-{3-[(1-benzyloxycarbonylamino-5-phthalimido-
pentyl)(methoxy)phosphonyl]propionylamino}propionate 20
A stirred mixture of phosphinic acid 11 (430 mg, 1.00 mmol)
and hexamethyldisilazane (316 µl, 1.50 mmol) was heated at
110 ЊC for 40 min under an atmosphere of dry nitrogen. After
cooling to 90 ЊC, 3-bromopropionyl--alanine methyl ester 19
(238 mg, 1.00 mmol) was added and the resulting mixture
stirred at 90 ЊC under a dry nitrogen atmosphere for 1 h. The
mixture was cooled to 70 ЊC and ethanol (10 ml) was added.
The solution was allowed to cool to room temperature and the
solvent removed under reduced pressure. The residue was redis-
solved in 10% aqueous sodium hydrogen carbonate and washed
with diethyl ether (3 × 10 ml). The separated aqueous solution
was acidified to pH 1 by dropwise addition of 2  hydrochloric
acid and the resulting white precipitate extracted into ethyl
acetate (3 × 20 ml). The pooled organic extracts were washed
with brine (60 ml) then dried (MgSO4), filtered and evaporated
under reduced pressure to give a white foam. This compound
was dissolved in dichloromethane (40 ml) and treated with a
solution of diazomethane in diethyl ether until the evolution of
nitrogen ceased and the yellow colour of diazomethane
remained permanent. The solution was allowed to stand for 16
h, treated dropwise with acetic acid, and the solvent removed
under reduced pressure. The crude product was purified by
flash chromatography on silica gel [eluting with 49:1 (v/v)
chloroform–methanol] to give the title compound as a white
foam (274 mg, 46%); νmax(Nujol mull)/cmϪ1 3225 (m), 1772 (m),
(2R)-2-(3-Bromo-2-methylpropionylamino)propionic acid benzyl
ester 18
To a stirred solution of (2R)-3-bromo-2-methylpropionic acid
(91 mg, 0.545 mmol) and N-methylmorpholine (60 ml, 0.545
mmol) in dry tetrahydrofuran (3 ml) at Ϫ15 ЊC under a dry
nitrogen atmosphere was added isobutyl chloroformate (71 ml,
0.545 mmol) and the solution stirred for 15 min. To the result-
ing solution was added a solution of (R)-alanine benzyl ester N-
toluene-p-sulfonate (192 mg, 0.545 mmol) and N-methyl-
morpholine (60 ml, 0.545 mmol) in dry N,N-dimethyl-
formamide (2 ml). This solution was then stirred whilst slowly
warming to room temperature. The solvent was removed in
vacuo and the residue partitioned between water (5 ml) and
138
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
1708 (s), 1743 (s), 1655 (s), 1537 (s); δH(300 MHz, CDCl3) 1.33–
1.37 (3H, m, CH3CH), 1.38–1.96 [6H, m, NCH2(CH2)3], 2.04–
2.19 (2H, m, PCH2), 2.43–2.62 (2H, m, PCH2CH2), 3.62–3.71
(8H, m), 3.99 (1H, q, J 10, NHCHP), 4.52 (1H, quintet, J 7,
CHCH3), 4.97–5.08 (2H, m, PhCH2), 5.42 and 5.97 (1H, 2d, J 7,
NHCHP), 6.95–7.08 (1H, m, NHCHCH3), 7.30 (5H, br s,
C6H5), 7.64–7.80 (4H, m, C6H4); δC(75 MHz, CDCl3) 18.01,
19.95, 20.19, 21.14, 21.36, 21.57 (m, PCH2), 22.93, 23.10, 23.26,
27.35, 27.75, 27.93, 37.29, 37.46, 48.18, 48.77, 49.50, 50.11
(multiplet, NHCHP), 51.91, 51.99, 52.11, 52.20, 52.40, 67.06,
67.11, 67.17, 67.23, 123.26, 132.19, 127.96, 128.00, 128.10,
128.16, 128.27, 128.59, 134.02, 136.25, 156.66, 168.55, 171.09,
173.59; δP(121 MHz, CDCl3) 54.85, 55.03, 55.31 and 55.37
(downfield from 85% H3PO4); m/z (ESϩ) 602 (9%, [M ϩ H]ϩ),
624 (100, [M ϩ Na]ϩ); HRMS (FAB) MHϩ, 602.2304.
C29H37N3O9P requires 602.2267.
(2R)-2-{3-[(1-Benzyloxycarbonylamino-5-phthalimidopentyl)
(hydroxy)phosphonyl]propionylamino}propionic acid 25
To a stirred solution of Z-protected pseudo-tripeptide phos-
phinate 20 (200 mg, 0.333 mmol) in dry tetrahydrofuran (10 ml)
was added bis(tributyltin) oxide (706 µl, 1.33 mmol) and the
solution heated under reflux for 48 h. After cooling to room
temperature, the solvent was removed under reduced pressure
and the residue redissolved in ethyl acetate (10 ml). This solu-
tion was extracted with 10% aqueous sodium hydrogen carbon-
ate solution (3 × 5 ml) and the pooled aqueous extracts were
washed with n-pentane (3 × 10 ml). The separated aqueous
layer was acidified to pH 1 by dropwise addition of 2  hydro-
chloric acid and the resulting white precipitate extracted with
ethyl acetate (3 × 15 ml). The pooled organic extracts were
washed with brine (45 ml) then dried (MgSO4), filtered and
evaporated. The residue was redissolved in 6:1 (v/v) propan-2-
ol–water (2 ml) and to this stirred solution was added sodium
borohydride (31 mg, 0.810 mmol). The solution was stirred for
16 h and then acetic acid (0.2 ml) was added. After effer-
vescence had subsided, the flask was stoppered and then heated
to 80 ЊC for 2 h. After cooling to room temperature the solution
was diluted to ~5 ml by addition of water and washed with
diethyl ether (3 × 5 ml). The water was removed in vacuo and
the residue redissolved in 1  potassium hydroxide solution (2
ml). This solution was stirred at room temperature for 4 h. The
solution was neutralised by dropwise addition of 2  hydro-
chloric acid and the water removed by freeze drying. The crude
product was purified by reversed phase HPLC (25 cm × 16 mm
Zorbax ODS column, eluting with a gradient from 0.1% tri-
fluoroacetic acid–water to 0.1% trifluoroacetic acid–acetonitrile
over 40 min, flow rate 4 ml minϪ1, detecting at 254 nm) to give
the title compound as a white fluffy solid (39 mg, 26%); HPLC
retention time tR (conditions as above) 18 min, mp 116–120 ЊC;
νmax(KBr disc)/cmϪ1 3404 (m, br), 1691 (s), 1539 (m); δH(300
MHz, D2O) 1.37 and 1.39 (3H, 2d, J 7.5, CH3CH), 1.44–1.92
[8H, m, H2NCH2(CH2)3 and PCH2CH2], 2.33–2.51 (2H, m,
PCH2CH2), 2.83–2.98 [2H, m, H2NCH2(CH2)3], 3.71–3.79 (1H,
m, CH3CH), 4.29 (1H, dt, J 7.5, 13, NHCHP), 5.11–5.12 (2H,
m, PhCH2), 7.39 (5H, br s, C6H5); δC(75 MHz, D2O), 18.68,
24.26, 24.73, 24.88, 25.46, 28.65, 29.01, 30.25 (d, PCH2), 41.81,
51.34, 52.02 (d, J 102, NHCHP), 69.84, 130.0, 130.15, 131.0,
131.4, 139.1, 160.7, 177.1, 179.2; δP(121 MHz, D2O) 47.30
(downfield from 85% H3PO4); m/z (ESϩ) 444 (100%, [M ϩ H]ϩ),
466 (9, [M ϩ Na]ϩ); HRMS (FAB) MHϩ, 444.1912.
C19H30N3O7P requires 444.1900.
Methyl (2R)-2-{3-[(1-amino-5-phthalimidopentyl)(methoxy)-
phosphonyl]propionylamino}propionate 21
To a stirred solution of Z-protected pseudo-tripeptide phos-
phinate 20 (234 mg, 0.389 mmol) in 5% ammonium formate
solution in methanol (10 ml) was added 10% palladium on acti-
vated charcoal (200 mg, 0.188 mmol Pd) and the mixture stirred
at room temperature for 2 h. The catalyst was removed by filtra-
tion through a short pad of Celite and the solvent removed in
vacuo. The residue was purified by flash chromatography on
silica gel [eluting with 19:1 (v/v) chloroform–methanol] to give
the title compound as a colourless oil (147 mg, 81%); νmax(liquid
film)/cmϪ1 3260 (m, br), 1766 (m), 1708 (s), 1743 (s), 1666 (s),
1549 (m); δH(300 MHz, CDCl3) 1.37 (3H, d, J 7, CH3CH), 1.38–
1.88 [6H, m, NCH2(CH2)3], 1.92 (2H, br, NH2), 1.94–2.25 (2H,
m, PCH2), 2.40–2.67 (2H, m, PCH2CH2), 2.80–2.95 (1H, m,
H2NCHP), 3.65–3.72 (8H, m), 4.51 (1H, quintet, J 7, CH3CH),
7.08–7.23 (1H, m, CH3CHNH), 7.68–7.83 (4H, m, C6H4); δC(75
MHz, CDCl3) 17.97, 19.39, 19.79 (2d, JC᎐P 86, PCH2), 22.25,
23.32, 23.49, 28.11, 29.91, 37.43, 48.06, 49.99 (d, JC᎐P 100,
H2NCHP), 51.67, 52.29, 123.1, 132.0, 133.9, 168.3, 171.1,
173.4; δP(121 MHz, CDCl3) 57.96, 58.02, 58.23, 58.32 (down-
field from 85% H3PO4); m/z (ESϩ) 468 (12%, [M ϩ H]ϩ), 490
(30, [M ϩ Na]ϩ).
Methyl (2R)-2-{3-[(1-acetylamino-5-phthalimidopentyl)-
(methoxy)phosphonyl]propionylamino}propionate 22
To a stirred solution of α-amino pseudo-tripeptide phosphinate
21 (371 mg, 0.794 mmol) in dichloromethane (5 ml) at 0 ЊC was
added triethylamine (188 µl, 1.35 mmol), followed by acetic
anhydride (105 µl, 1.11 mmol) and 4-dimethylaminopyridine
(small crystal) and the solution stirred for 5 min. The reaction
was quenched by the addition of a drop of acetic acid and the
solvent removed under reduced pressure. The residue was redis-
solved in water (10 ml) and extracted with dichloromethane
(3 × 10 ml). The pooled organic extracts were washed with
brine (30 ml) then dried (MgSO4), filtered and evaporated. The
residue was purified by flash chromatography on silica gel [elut-
ing with 19:1 (v/v) chloroform–methanol] to give the title com-
pound as a white foam (367 mg, 91%); νmax(CH2Cl2 solution)/
cmϪ1 3050 (m), 1768 (w), 1707 (s), 1740 (m), 1674 (m), 1543 (m),
1515 (m); δH(300 MHz, CDCl3), 1.38 (3H, 2d, J 7, CH3CH),
2.02 (3H, 3s, CH3CONH), 1.59–2.70 (10H, m), 3.64–3.75 (8H,
m), 4.24–4.39 (1H, m, NHCHP), 4.47–4.58 (1H, m, CH3CH),
6.33–6.36 (1H, m, NH), 6.96–7.22 (1H, m, NH), 7.68–7.84
(4H, m, C6H4); δC(75 MHz, CDCl3) 17.92, 18.09, 20.99 (d,
JC᎐P 87, PCH2), 22.96, 23.10, 26.66, 26.87, 27.39, 27.86, 27.99,
28.61, 37.42, 37.09, 46.46, 48.24, 50.63, 52.30, 123.3, 132.2,
134.1, 168.5, 168.6, 171.2, 173.6; δP(121 MHz, CDCl3) 50.68,
50.76, 55.17, 55.29, 55.85, 55.90 (downfield from 85%
H3PO4); m/z (ESϩ) 510 (16% [M ϩ H]ϩ), 532 (100, [M ϩ Na]ϩ);
HRMS (FAB) MHϩ, 510.2015. C23H33N3O8P requires
510.2005.
(2R)-2-{3-[(1-Acetylamino-5-aminopentyl)(hydroxy)-
phosphonyl]propionylamino}propionic acid 26
To a stirred solution of N-acetyl pseudo-tripeptide phosphinate
methyl ester 22 (276 mg, 0.542 mmol) in dry tetrahydrofuran
(20 ml) was added bis(tributyltin) oxide (1.11 ml, 2.17 mmol)
and the solution heated under reflux for 16 h. After cooling to
room temperature the solvent was removed under reduced pres-
sure and the residue redissolved in ethyl acetate (10 ml). This
solution was extracted with 10% aqueous sodium hydrogen
carbonate (3 × 5 ml) and the pooled aqueous extracts were
washed with n-pentane (3 × 10 ml). The aqueous solution was
acidified to pH 1 by dropwise addition of 2  hydrochloric acid
and the water removed in vacuo. The residue was redissolved in
6:1 (v/v) propan-2-ol–water (2 ml) and sodium borohydride (48
mg, 1.26 mmol) was added. This solution was stirred for 16 h at
room temperature and then acetic acid (0.2 ml) was added.
When effervescence had subsided, the flask was stoppered and
heated at 80 ЊC for 2 h. After cooling to room temperature the
solvent was removed in vacuo and the residue redissolved in
water (10 ml). This solution was washed with diethyl ether
(3 × 10 ml) and the solvent removed by freeze drying. The crude
product was purified by reversed phase HPLC (25 cm × 16 mm
Zorbax ODS column, eluting with a gradient from 0.1% tri-
fluoroacetic acid in water to 0.1% trifluoroacetic acid in
J. Chem. Soc., Perkin Trans. 1, 1998
139

View Article Online
acetonitrile over 40 min, flow rate 4 ml minϪ1, detecting at 220
nm) to give the title compound as a white fluffy solid (31 mg,
12%); HPLC tR (conditions as above) 9 min, mp 148–152 ЊC
(decomp.); νmax(KBr disc)/cmϪ1 3421 (s, br), 1640 (s), 1547 (m);
δH(360 MHz, D2O) 1.08 and 1.39 (3H, 2d, J 7, CH3CH), 1.43–
1.87 (8H, m), 2.03 (3H, s, CH3CO), 2.36–2.53 (2H, m,
PCH2CH2), 2.96 [2H, t, J 7.5, H2NCH2(CH2)3], 3.96–4.04 (1H,
m, NHCHP), 4.31 (1H, q, J 7.5, CH3CH); δC(90 MHz, D2O)
18.32, 18.69, 24.52, 25.12 (d, JC᎐P 16, PCHCH2), 26.32, 28.11,
29.07, 31.04 (d, JC᎐P 35, PCHCH2), 41.81, 50.97 (d, JC᎐P 77,
NHCHP), 51.54, 176.3, 177.8, 180.0; δP(121 MHz, D2O) 42.16
(downfield from 85% H3PO4); m/z (ESϩ) 352 (8% [M ϩ H]ϩ).
mediate -glutamic acid γ-benzyl ester α-methyl ester trifluoro-
acetate salt as a white solid (1.65 g, 78%).
To a stirred solution of -glutamic acid γ-benzyl ester α-
methyl ester trifluoroacetate salt (808 mg, 2.21 mmol) in dichlo-
romethane (10 ml) at 0 ЊC was added triethylamine (617 µl, 4.43
mmol) followed by propionic anhydride (426 µl, 3.32 mmol).
This solution was stirred for 5 min and then the reaction was
quenched by the addition of a drop of acetic acid. The solvent
was removed under reduced pressure and the residue redis-
solved in ethyl acetate (50 ml). This solution was washed with
10% aqueous citric acid (50 ml), water (50 ml) and brine (50
ml). The separated organic solution was then dried (MgSO4),
filtered and evaporated. The residue was purified by flash chro-
matography on silica gel [eluting with 1:1 (v/v) light petroleum–
ethyl acetate] to give the title compound as a colourless oil
which solidified upon standing (586 mg, 86%), mp 55–56 ЊC;
νmax(CH2Cl2 solution)/cmϪ1 3407 (w), 1740 (s), 1680 (s), 1661
(s), 1514 (m); δH(300 MHz, CDCl3) 1.14 (3H, t, J 7.5, CH3CH2),
1.96–2.55 (4H, m), 2.22 (2H, q, J 7.5, CH3CH2), 3.73 (3H, s,
CO2CH3), 4.64 (1H, dt, J 5, 7.5, NHCHCH2), 5.12 (2H, s,
PhCH2), 6.21 (1H, d, J 7.5, NH), 7.36 (5H, br s, C6H5); δC(75
MHz, CDCl3) 9.52, 27.23, 29.36, 30.28, 51.48, 52.47, 66.74,
128.2, 128.3, 128.5, 135.6, 172.5, 172.7, 173.7; m/z (ESϩ) 308
(100%, [M ϩ H]ϩ), 330 (28, [M ϩ Na]ϩ).
2-(3-{[1-(4-Carboxybutanoylamino)-5-aminopentyl](hydroxy)-
phosphoryl}propanoylamino)propanoic acid 27
To a stirred solution of α-amino pseudo-tripeptide phosphinate
methyl ester 21 (68 mg, 0.146 mmol) in dichloromethane (2 ml)
at 0 ЊC was added triethylamine (40 µl, 0.293 mmol) followed by
glutaric anhydride (25 mg, 0.220 mol) and 4-dimethylamino-
pyridine (small crystal) and the solution stirred for 5 min. The
reaction was quenched by the addition of a drop of acetic acid
and the solvent removed under reduced pressure. The residue
was redissolved in 10% aqueous sodium hydrogen carbonate (5
ml) and washed with diethyl ether (3 × 5 ml). The aqueous solu-
tion was then acidified to pH 1 by dropwise addition of 2 
hydrochloric acid and extracted with ethyl acetate (3 × 10 ml).
The pooled organic extracts were washed with brine (30 ml)
then dried (MgSO4), filtered and evaporated, to give the N-
glutaryl pseudo-tripeptide phosphinate 23 (ESϩ 582 [MHϩ]).
The crude solid 23 was redissolved in dry tetrahydrofuran (5
ml) and bis(tributyltin) oxide (175 ml, 0.344 mmol) was added.
This stirred solution was heated under reflux for 48 h and after
cooling to room temperature the solvent was removed under
reduced pressure. The residue was triturated with n-pentane
(3 × 5 ml) and then redissolved in 6:1 (v/v) propan-2-ol–water
(2 ml). To this solution was added sodium borohydride (16 mg,
0.430 mmol) and the solution stirred for 16 h at room temper-
ature. Acetic acid (0.2 ml) was added and after effervescence had
subsided, the flask was stoppered and heated at 80 ЊC for 2 h.
After cooling to room temperature the solvent was removed in
vacuo and the residue redissolved in 1  potassium hydroxide
solution (1 ml). This solution was stirred for 1 h at room tem-
perature then neutralised by the dropwise addition of 2 
hydrochloride acid. After isolation by freeze drying, the crude
product was purified by reversed phase HPLC (25 cm × 16 mm
Zorbax ODS column, eluting with a gradient from 0.1% trif-
luoroacetic acid in water to 0.1% trifluoroacetic acid in
acetonitrile over 40 min, flow rate 4 ml minϪ1, detecting at 220
nm) to give the title compound as a white fluffy solid (10 mg,
16%); HPLC tR (conditions as above) 10.5 min, mp 132–140 ЊC
(decomp.); νmax(KBr disc)/cmϪ1 3429 (s, br), 1649 (s), 1539 (m);
δH(300 MHz, D2O) 1.38 (3H, d, J 7.5, CH3CH), 1.43–1.92
(10H, m), 2.33–2.51 (6H, m), 2.96 [2H, t, J 8, H2NCH2(CH2)3],
3.99 (1H, ddd, J 3, 9, 12, NHCHCH2), 4.30 (1H, q, J 7.5,
NHCHCH3); δC(67.5 MHz, D2O) 16.79, 20.29, 21.38, 23.10,
24.47, 26.86, 28.81, 33.65, 35.45, 39.97, 48.00, 49.52, 175.8,
176.3, 178.5; δP(121 MHz, D2O) 42.85 (downfield from 85%
H3PO4); m/z (ESϩ) 424 (100%, [M ϩ H]ϩ), 446 (15, [M ϩ Na]ϩ);
HRMS (FAB) MHϩ, 424.1797. C16H31N3O8P requires
424.1849.
N-Propionyl-D-glutamic acid á-methyl ester
To a stirred solution of N-propionyl--glutamic acid δ-benzyl
ester α-methyl ester (502 mg, 1.64 mmol) in ethanol (20 ml) was
added 10% palladium on activated charcoal (87 mg, 81.8 mmol
Pd). This mixture was stirred under an atmosphere of hydrogen
for 1.5 h at room temperature. The catalyst was removed by
filtration through a short pad of Celite and then the solvent was
removed under reduced pressure. The product was recrystal-
lised from ethyl acetate–hexane to give the title compound as a
white crystalline solid (360 mg, 100%), mp 72–73 ЊC; νmax(KBr
disc)/cmϪ1 3288 (s, br), 1818 (s), 1738 (s), 1650 (s), 1538 (s);
δH(300 MHz, CDCl3) 1.15 (3H, t, J 7.5, CH3CH2), 1.91–2.04
(1H, m), 2.15–2.22 (1H, m), 2.37–2.52 (2H, m), 2.28 (2H, q,
J 7.5, CH3CH2), 3.75 (3H, s, CO2CH3), 4.66 (1H, dt, J 5, 8,
NHCH), 6.50 (1H, d, J 8, NH); δC(75 MHz, CDCl3) 9.57, 27.23,
29.37, 30.00, 51.45, 52.60, 172.6, 174.5, 176.8; m/z (ESϩ) 218
(12%, [M ϩ H]ϩ).
(2R)-2-(3-{[1-(4-Carboxy-4-propionylaminobutanoylamino)-5-
aminopentyl](hydroxy)phosphonyl}propionylamino)propionic
acid 28
To a stirred solution of N-propionyl--glutamic acid α-methyl
ester (340 mg, 1.57 mmol) in dichloromethane (10 ml) at 0 ЊC
was added N-hydroxysuccinimide (180 mg, 1.57 mmol) fol-
lowed by dicyclohexylcarbodiimide (323 mg, 1.57 mmol) and
the solution stirred at 0 ЊC for 1 h. The resulting precipitate was
removed by filtration and the solvent removed under reduced
pressure. The residue was redissolved in diethyl ether (10 ml)
and filtered again to remove residual dicyclohexylurea. The
solvent was then removed under reduced pressure and the crude
activated ester redissolved in dichloromethane (2 ml).
To a stirred solution of α-amino pseudo-tripeptide phosphi-
nate 21 (202 mg, 0.433 mmol) in dichloromethane (5 ml) at 0 ЊC
was added triethylamine (218 ml, 1.57 mmol) followed by the
solution of crude activated ester prepared earlier, plus wash-
ings. This solution was then stirred for 1 h at 0 ЊC and for 16 h at
room temperature. The solvent was removed under reduced
pressure and the residue redissolved in ethyl acetate (5 ml). This
solution was washed with 10% aqueous citric acid (5 ml), satur-
ated aqueous sodium hydrogen carbonate (5 ml) and brine (5
ml). The separated organic layer was then dried (MgSO4), fil-
tered and evaporated to give the protected pseudo-tetrapeptide
phosphinate 24 (ESϩ 689 [M ϩ Naϩ]).
N-Propionyl-D-glutamic acid ã-benzyl ester á-methyl ester
A solution of N-butoxycarbonyl--glutamic acid γ-benzyl ester
α-methyl ester27 (2.04 g, 5.82 mmol) in 25% trifluoroacetic acid
in dichloromethane (30 ml) was stirred at room temperature for
1 h. The solvent was removed under reduced pressure and the
residue redissolved in water (20 ml). This solution was extracted
with diethyl ether (3 × 20 ml) and the water removed from the
separated aqueous layer by freeze drying. This gave the inter-
The residue was redissolved in dry tetrahydrofuran (5 ml)
140
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
and bis(tributyltin) oxide (594 µl, 1.17 mmol) and the stirred
solution heated under reflux for 48 h. After cooling to room
temperature the solvent was removed under reduced pressure
and the residue triturated with n-pentane (3 × 5 ml). The resi-
due was then redissolved in 6:1 (v/v) propan-2-ol–water (2 ml)
and to this solution was added sodium borohydride (28 mg,
0.730 mmol). This solution was then stirred for 16 h at room
temperature and acetic acid (0.2 ml) was added. After effer-
vescence had subsided the flask was stoppered and heated to
80 ЊC for 2 h. The solution was allowed to cool to room tem-
perature and the solvent removed in vacuo. The residue was
redissolved in 1  potassium hydroxide solution (1 ml) and this
solution stirred for 1 h at room temperature. The solution was
neutralised by dropwise addition of 2  hydrochloric acid and
the water removed by freeze drying. The crude product was
purified by reversed phase HPLC (25 cm × 16 mm Zorbax ODS
column, eluting with a gradient ranging from 0.1% trifluoro-
acetic acid in water to 0.1% trifluoroacetic acid in acetonitrile
over 40 min, flow rate 4 ml minϪ1, detecting at 220 nm) to give
the title compound as a white powder (5.5 mg, 2.6%); HPLC tR
(conditions as above) 11 min; δH(360 MHz, D2O, COSY analy-
sis) 1.14 (3H, t, J 7.5, CH3CH2), 1.40 (3H, d, J 7.5, CH3CH),
1.59–2.24 [8H, m, H2NCH2(CH2)3 and HO2CCHCH2CH2],
2.29–2.48 (8H, m, PCH2CH2, HO2CCHCH2CH2 and
CH3CH2), 3.02 [2H, m, H2NCH2(CH2)3], 3.64 (1H, t, J 6.5,
HO2CCHCH2), 4.19–4.34 (2H, m, NHCHP and CH3CH);
δP(121 MHz, D2O) 40.91 (downfield from 85% H3PO4); m/z
(FAB) 495 (MHϩ); HRMS (FAB) MHϩ, 495.2353.
C19H36N4O9P requires 495.2220.
and freeze-dried to give the title compound as a white powder
(12.8 mg); HPLC tR (conditions as above) 12.8 min; m/z (ESϪ)
1050 (15%, [M Ϫ H]Ϫ).
Purification of D-Ala-D-Ala adding enzyme from JM105/pTB3
(see Table 1)
All steps were performed at 4 ЊC unless otherwise specified.
Enzyme activity was followed by the previously reported phos-
phate release assay.6 Protein concentration was determined by
the method of Bradford.20 Buffer A consisted of 50 m HEPES
(pH 7.2), 5 m MgCl2 and 1 m EDTA.
JM105/pTB3 was grown at 37 ЊC in 5 dm3 Luria Broth con-
taining 100 µg mlϪ1 ampicillin to an A595 of 0.6, whereupon
IPTG was added to a final concentration of 1 m, to induce the
tac promoter. Cells were then grown for a further 5 h at 37 ЊC
and harvested by centrifugation at 6000 g for 10 min. The cell
pellet (13.9 g) was stored overnight on ice, and re-suspended in
40 ml of 100 m HEPES (pH 7.2) containing 300 m NaCl, 5
m MgCl2, 1 m EDTA and 5 m DTT, and was passed twice
through a French press at 1000 psi. Cell debris was removed by
centrifugation at 25 000 g for 30 min.
Powdered ammonium sulfate was gradually added to the
supernatant to a a final concentration of 25% saturation, and
the solution was stirred for 1 h. The solution was cleared by
centrifugation at 12 000 g for 30 min, and ammonium sulfate
was added to the supernatant to a concentration of 50% satur-
ation. After the solution was stirred for a further 1 h, the pre-
cipitate was collected by centrifugation at 12 000 g for 30 min.
The precipitate was re-suspended in 10 ml of buffer A, and was
loaded onto an Ultrogel AcA54 gel filtration column (2.5 × 108
cm) and eluted at 60 ml hϪ1 with buffer A. Fractions containing
-Ala--Ala adding enzyme activity were pooled, and applied
directly to a Q Sepharose 16/10 anion exchange column (Phar-
macia). The enzyme was eluted at 4.0 ml minϪ1 using a gradient
of 100–300 m KCl in buffer A over 300 ml, activity eluting at
200 m KCl. The purified enzyme appeared as a 48 kDa protein
by SDS-PAGE, and was judged to be >95% homogeneous. Its
Uridinediphospho-N-acetylmuramyl-L-alanyl-ã-D-glutamyl-m-
diaminopimelic acid (UDPMurNAc-tripeptide)
Cell Wall Synthesis Medium (CWSM) contains, per dm3:
Na2HPO4 0.26 g, NH4Cl 2.0 g, KCl 4.0 g, MgCl2 4.0 g, Na2SO4
0.15 g, FeSO4 0.1 g, glucose 2.0 g, uracil 40 mg, -glutamic acid
120 mg, -lysine 500 mg, ,-diaminopimelic acid 120 mg, -
alanine 50 mg, chloramphenicol 50 mg, ampicillin 50 mg and -
cycloserine 150 mg. CWSM was sterilised by autoclaving the
mixture before addition of FeSO4 and antibiotics. Solutions
containing antibiotics and FeSO4 were added through a 0.22 µm
sterile filter immediately before inoculation.
specific activity was 20.6 u mgϪ1
.
Kinetic assays of D-Ala-D-Ala adding enzyme
KM values for -Ala--Ala and ATP, and Ki values for -Ala--
Ala analogues were determined using the Pi release assay previ-
ously described.6 Assays (100 µl) contained 100 m Tris (pH
8.6), 10 m MgCl2, 100 µ UDPMurNAc-tripeptide (10 nmol),
1 m ATP, 5–200 µ -Ala--Ala and enzyme (6 × 10Ϫ4 units),
and were incubated for a fixed time period (5 or 10 min) at 37 ЊC
prior to addition of colour reagent (800 µl) and sodium citrate
(100 µl) and measurement of A660. Assays were carried out in
duplicate at varying concentrations of -Ala--Ala (or ATP),
with a reaction containing no -Ala--Ala as a control, which
showed no background ATPase activity. Pi release was quanti-
tated using 1–10 nmol potassium phosphate standards, and KM
values were calculated by plots of 1/v vs. 1/[S]. Ki values were
measured by determination of the apparent KM (KMЈ) for -
Ala--Ala in the presence of a fixed concentration of inhibitor,
and Ki calculated using KMЈ = KM(1 ϩ [I]/Ki).
Assays using the HPLC-purified pseudo-tetrapeptide phos-
phinate 28 were complicated by background levels of Pi in the
inhibitor which precluded assays at low concentrations of -
Ala--Ala. Therefore, assays were carried out at fixed substrate
concentrations (100 µ -Ala--Ala) and variable inhibitor
concentrations (0–500 µ). Ki was determined by a Dixon plot
(1/v versus [I]),30 giving a linear plot and a value of 200 ( 20) µ
for Ki.
Uridinediphospho-N-acetylmuramyl--Ala-γ--Glu-m-DAP
was isolated from cultures of Bacillus subtilis W23 (NCIMB
11 824). B. subtilis was grown to stationary phase in 50 ml of
PYP medium (bacteriological peptone 20 g dmϪ3, yeast extract
1.5 g dmϪ3, KH2PO4 5.5 g dmϪ3) in a baffled flask at 37 ЊC with
shaking at 240 rpm. This was used to inoculate at 2% 4 × 500
ml PYP cultures in 2 l baffled flasks and grown to OD660 1.5.
The cultures were chilled in an ice bath and cells were resus-
pended in 50 ml of CWSM. This was added (25 ml each) to
2 × 225 ml prewarmed CWSM cultures in 2 l baffled flasks and
incubated at 37 ЊC with shaking at 240 rpm for 45 min before
chilling in an ice bath. Cells were collected by centrifugation at
5000 g (10 min) and stored at Ϫ20 ЊC (5.8 g cells, wet weight).
Thawed cells were re-suspended in ice cold 5% trichloroacetic
acid (30 ml, 5 ml gϪ1 cells) and left on ice for 30 min. Precipi-
tated material was removed by centrifugation at 5000 g (10
min). The pellet was extracted with ice-cold 5% trichloroacetic
acid (2 × 15 ml). The supernatants were pooled and extracted
with diethyl ether (3 × 60 ml). The aqueous phase was neutral-
ised with 3  NaOH and concentrated to ~1.5 ml in vacuo. This
solution was clarified by microcentrifugation and applied to a
Biogel P2 column (1.5 × 30 cm column, flow rate 0.9 ml minϪ1
)
taking 2 ml fractions. Each fraction was tested for UV absorb-
ance at 262 nm. Active fractions were analysed by reversed
phase HPLC [4.6 mm × 25 cm Zorbax ODS column, eluting
with 0.1  NH4H2PO4 (pH 4.6), flow rate 1 ml minϪ1, detecting
at 262 nm] using an authentic standard of the title compound
to identify fractions containing UDPMurNAc-tripeptide. Frac-
tions containing UDPMurNAc-tripeptide only were pooled
Note added in proof
A recent paper published since the submission of this article
provides several experimental observations which strongly
support the hypothetical intramolecular mechanism proposed
in Scheme 5 for reactions of phosphinic acid 13 (L. A. Reiter
and B. P. Jones, J. Org. Chem., 1997, 62, 2808).
J. Chem. Soc., Perkin Trans. 1, 1998
141

View Article Online
13 E. A. Boyd, M. Corless, K. James and A. C. Regan, Tetrahedron
Acknowledgements
Lett., 1990, 31, 2933.
14 K. Afarinkia, C. W. Rees and J. I. G. Cadogan, Tetrahedron, 1990,
46, 7175.
15 E. K. Baylis, C. D. Campbell and J. G. Dingwall, J. Chem. Soc.,
Perkin Trans. 1, 1984, 2845.
The authors wish to thank Professor C. Walsh (Harvard Med-
ical School) and Dr J. van Heijenoort (CNRS, Paris) for assist-
ance with overexpression of the murF gene, Lesley Robinson
for preliminary synthetic work, Philip Brandish for authentic
samples of UDPMurNAc-tripeptide, and Sylvie Gehanne,
Chiara Savoia and Enrico Domenici for helpful discussions.
We would like to acknowledge the award of a collaborative
studentship (to D. J. M.) by the Medical Research Council, and
financial support from GlaxoWellcome.
16 E. A. Boyd, A. C. Regan and K. James, Tetrahedron Lett., 1994, 35,
4223.
17 F. C. Neuhaus and W. G. Struve, Biochemistry, 1965, 4, 120.
18 C. J. Salamon, E. G. Mata and O. A. Mascaretti, J. Org. Chem.,
1994, 59, 7259.
19 J. O. Osby, M. G. Martin and B. Ganem, Tetrahedron Lett., 1984,
25, 2093.
20 M. M. Bradford, Anal. Biochem., 1976, 72, 248.
21 P. A. Lanzetta, L. J. Alvarez, P. S. Reinach and O. Candia, Anal.
Biochem., 1979, 100, 95.
22 C. Michaud, D. Blanot, B. Fluoret and J. van Heijenoort, Eur. J.
Biochem., 1987, 166, 631.
23 A. Yiotakis, S. Vassilion, J. Jiracek and V. Dive, J. Org. Chem., 1996,
61, 6601.
24 M. E. Tanner, S. Vaganay, J. van Heijenoort and D. Blanot, J. Org.
Chem., 1996, 61, 1756.
25 M. Soroka and J. Sygmunt, Synthesis, 1988, 370.
26 D. D. Miller, F. L. Hsu, R. R. Ruffolo Jr. and P. N. Patil, J. Med.
Chem., 1976, 19, 1382.
27 N. J. Manesis and M. Goodman, J. Org. Chem., 1987, 52, 5331.
28 T. W. Greene and P. G. M. Wuts, in Protective Groups in Organic
Synthesis, Wiley Interscience, New York, 2nd edn., 1991.
29 H. Takahata, Y. Ohnishi, H. Takehara and K. Tsuritani, Chem.
Pharm. Bull., 1981, 29, 1063.
References
1 T. D. H. Bugg and C. T. Walsh, Nat. Prod. Rep., 1992, 9, 199.
2 C. T. Walsh, J. Biol. Chem., 1989, 264, 2393.
3 L. S. Mullins, L. E. Zawadzke, C. T. Walsh and F. M. Raushel,
J. Biol. Chem., 1990, 265, 8993.
4 W. H. Parsons, A. A. Patchett, H. G. Bull, W. R. Schoen, D. Taub, J.
Davidson, P. L. Combs, J. P. Springer, H. Gadebusch, B.
Wiessberger, M. E. Valiant, T. N. Mellin and R. D. Busch, J. Med.
Chem., 1988, 31, 1772.
5 K. Duncan and C. T. Walsh, Biochemistry, 1988, 27, 3709.
6 K. Duncan, J. van Heijenoort and C. T. Walsh, Biochemistry, 1990,
29, 2379.
7 C. Parquet, B. Flouret, D. Mengin-Lecreulx and J. van Heijenoort,
Nucleic. Acids Res., 1989, 17, 5379.
8 M. S. Anderson, S. S. Eveland, H. R. Onishi and D. L. Pompliano,
Biochemistry, 1996, 35, 16 264.
9 T. D. H. Bugg, J. van Heijenoort and C. T. Walsh, unpublished
results.
30 I. H. Segel, in Enzyme Kinetics, Wiley Interscience, New York,
1993, p. 109.
10 A. Yiotakis, A. Lecoq, A. Nicolaou, J. Labadie and V. Dive,
Biochem. J., 1994, 303, 323.
11 A. J. Mancuso, S. L. Huang and D. Swern, J. Org. Chem., 1978, 43,
2480.
Paper 7/04097K
Received 11th June 1997
Accepted 5th August 1997
12 E. A. Boyd, A. C. Regan and K. James, Tetrahedron Lett., 1992, 33,
813.
142
J. Chem. Soc., Perkin Trans. 1, 1998