Glutathione-analogous peptidyl phosphorus esters as mechanism-based inhibitors of γ-glutamyl transpeptidase for probing cysteinyl-glycine binding site

Bioorganic & Medicinal Chemistry 22 (2014) 1176–1194

Contents lists available at ScienceDirect

Bioorganic & Medicinal Chemistry

journal homepage: w ww.elsevier.com/locate/bmc

Glutathione-analogous peptidyl phosphorus esters as

mechanism-based inhibitors of c-glutamyl transpeptidase

for probing cysteinyl-glycine binding site

Mado Nakajima ^a,, Bunta Watanabe ^a, Liyou Han ^a, Bun-ichi Shimizu ^a,à, Kei Wada ^b, Keiichi Fukuyama ^c,§

,

Hideyuki Suzuki ^d, Jun Hiratake ^a,

⇑

^aInstitute for Chemical Research, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan

^bOrganization for Promotion of Tenure Track, University of Miyazaki, 5200 Kihara, Kiyotake, Miyazaki 889-1692, Japan

^cDepartment of Biological Sciences, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan

^dDivision of Applied Biology, Graduate School of Science and Technology, Kyoto Institute of Technology, Goshokaido-cho, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan

a r t i c l e i n f o

a b s t r a c t

Article history:

c

-Glutamyl transpeptidase (GGT) catalyzing the cleavage of

c

-glutamyl bond of glutathione and its

Received 16 October 2013

Revised 6 December 2013

Accepted 7 December 2013

Available online 22 December 2013

S-conjugates is involved in a number of physiological and pathological processes through glutathione

homeostasis. Deﬁning its Cys-Gly binding site is extremely important not only in deﬁning the physiolog-

ical function of GGT, but also in designing speciﬁc and effective inhibitors for pharmaceutical purposes.

Here we report the synthesis and evaluation of a series of glutathione-analogous peptidyl phosphorus

esters as mechanism-based inhibitors of human and Escherichia coli GGTs to probe the structural and ste-

reochemical preferences in the Cys-Gly binding site. Both enzymes were inhibited strongly and irrevers-

ibly by the peptidyl phosphorus esters with a good leaving group (phenoxide). Human GGT was highly

selective for

whereas E. coli GGT signiﬁcantly preferred

analogue at the Cys binding site were necessary for inhibition, suggesting that human GGT was highly

selective for glutathione ( -Glu- -Cys-Gly), whereas E. coli GGT are not selective for glutathione, but still

retained the dipeptide ( -AA-Gly) binding site. The diastereoisomers with respect to the chiral phospho-

rus were separated. Both GGTs were inactivated by only one of the stereoisomers with the same stereo-

chemistry at phosphorus. The strict recognition of phosphorus stereochemistry gave insights into the

stereochemical course of the catalyzed reaction. Ion-spray mass analysis of the inhibited E. coli GGT con-

ﬁrmed the formation of a 1:1 covalent adduct with the catalytic subunit (small subunit) with concomi-

tant loss of phenoxide, leaving the peptidyl moiety that presumably occupies the Cys-Gly binding site.

The peptidyl phosphonate inhibitors are highly useful as a ligand for X-ray structural analysis of GGT

for deﬁning hitherto unidentiﬁed Cys-Gly binding site to design speciﬁc inhibitors.

Keywords:

c

-Glutamyl transpeptidase

Peptidyl phosphorus esters

Transition-state analogues

Glutathione analogues

Mechanism-based inhibitors

Cysteinyl-glycine binding site

Structure–activity relationship

Human GGT

L

-aliphatic amino acid such as

L-2-aminobutyrate (

L

-Cys mimic) at the Cys binding site,

L

-Phe mimic at this site. The C-terminal Gly and a

L-amino acid

c

L

E. coli GGT

1. Introduction

c-glutamyl group to water (hydrolysis) or amino acids and pep-

tides (transpeptidation) by the ping-pong mechanism via a c-glut-

c

-Glutamyl transpeptidase (GGT, EC 2.3.2.2) is a heterodimeric

enzymes found widely in organisms ranging from bacteria, plants

to mammals. GGT catalyzes the cleavage of the -glutamyl amide

bond of glutathione ( -glutamyl- -cysteinylglycine), its S-conju-

-glutamylamides to transfer the

amyl-enzyme ester intermediate (Scheme 1A).^1–4Mammalian GGT

is bound to the external surface of cell membrane and is expressed

in high concentrations in kidney tubules, biliary epithelium and

brain capillaries.^2,5GGT plays a central role in glutathione homeo-

c

-L

L

gates, glutamine and other

c

stasis as the sole enzyme capable of cleaving c-glutamyl bond of

glutathione in the initial step of glutathione metabolism; GGT

hydrolyzes the extracellular glutathione in concert with other

dipeptidases to provide cells with cysteine, the rate-limiting sub-

strate for intracellular de novo synthesis of glutathione.^4–8GGT is

also involved in the metabolism of xenobiotics by cleaving the

⇑

Corresponding author. Tel.: +81 774 38 3231; fax: +81 774 38 3229.

E-mail address: hiratake@scl.kyoto-u.ac.jp (J. Hiratake).

Present address: Shionogi Pharmaceutical Research Center, Shionogi & Co., Ltd,

1-1, Futaba-cho 3-chome, Toyonaka, Osaka 561-0825, Japan.

à

Present address: Graduate School of Life Sciences, Toyo University, 1-1-1-Izumino,

c-glutamyl bond of glutathione-S-conjugates as the initial step

Itakura-machi, Oga-gun, Gumma 374-0193, Japan.

for their degradation and excretion.²Accordingly, the overexpres-

sion of GGT is often observed in human tumors, and its roles in

§

Present address: Department of Applied Chemistry, Graduate School of

Engineering, Osaka University, Suita, Osaka 565-0871, Japan.

http://dx.doi.org/10.1016/j.bmc.2013.12.034

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1177

A

B

C

Scheme 1. Proposed catalytic mechanism of GGT (A), inhibition of GGT by acivicin (B) and inhibition of GGT by the mechanism-based inhibitor 1 (C). Compound 1 is a

mixture of stereoisomers with respect to the ethyl side chain (see also Table 1).

tumor progression,^9,10growth acceleration¹¹and the expression of

such malignant phenotypes of cancer cells as drug resistance^11–13

and metastasis^14,15have been suggested repeatedly. GGT is also

implicated in many physiological disorders such as neurodegener-

ative diseases,^16,17diabetes,¹⁸cardiovascular disease^19–21and pul-

monary disease.²²Hence the serum GGT level has been used

extensively as a diagnostic/prognostic marker of liver dysfunction,

coronary heart diseases, Type 2 diabetes, stroke and atherosclero-

sis.^7,21While GGT is recognized widely as an integral component of

cellular antioxidative defense system,^7,8,13several lines of evidence

have indicated that GGT promotes oxidative stress by producing

Cys-Gly, a highly reactive thiol that generates reactive oxygen spe-

cies (ROS) via metal ion-catalyzed reduction of molecular oxy-

gen.²³Therefore, GGT can also be recognized as a pro-oxidant

enzyme that potentially promotes oxidative stress.^24–27The seem-

ingly contradictory dual aspects of GGT, along with a complicated

glutathione metabolism,²⁸have made the physiological character-

ization of GGT rather complex. However, the implication of GGT in

a number of pathological processes strongly suggests a causative

role of this enzyme^7,23,25,27rather than a simple result of cellular

adaptation to oxidative stress.^7,8Therefore, GGT is an attractive

pharmaceutical target for cancer chemotherapy and a vast array

of physiological disorders involving oxidative stress and glutathi-

one metabolism.^13,15,22

frequent use of acivicin as a GGT inhibitor, irrespective of its toxic-

ity, is primarily due to the irreversible nature of its inhibition; aci-

vicin reacts with the catalytic Thr residue of GGT to make a

covalent adduct to inactivate the enzyme (Scheme 1B).^36,37There-

fore the important criteria for GGT inhibitors of practical use both

in vitro and in vivo are (1) irreversible inhibition with no regain of

enzyme activity until new enzymes are de novo synthesized and (2)

high selectivity for GGT without inhibiting glutamine-hydrolyzing

biosynthetic enzymes. In view of these criteria, we reported a ser-

ies of phosphonate esters as active-site directed and transition-

state analogue inhibitors of GGT.^38,39The phosphonate esters react

in a mechanism-based manner with the N-terminal Thr residue in

the small subunit, the catalytic nucleophile of GGT,^40,41to inacti-

vate the enzyme completely. Among those inhibitors, compound

1 (a mixture of four stereoisomers, see Section 2.5 and Table 1) clo-

sely mimicking the transition state 1 (TS-1) for glutathione hydro-

lysis, exhibited an exceptionally higher inhibitory activity (>180

times) toward human GGT than the phosphonate with a simple

methoxy group in place of the peptidyl chain.³⁹Interestingly, the

E. coli enzyme did not show any extra sensitivity toward the phos-

phonate 1. This observation suggested that the presence of the

peptidyl chain signiﬁcantly improved the afﬁnity of compound 1

to the active site of human GGT, thereby facilitating the formation

of the enzyme–inhibitor covalent complex that is closely analo-

gous to the putative transition state TS-1 (Scheme 1A and C). We

reasoned therefore that human GGT, but not the E. coli enzyme,

held a deﬁned Cys-Gly binding site that could provide an extra,

or in some cases, a major binding energy to enhance the inhibitory

activity of the covalent-type phosphonate inhibitors. The charac-

terization of the Cys-Gly binding site of GGT is also important in

reducing the toxicity of inhibitors by avoiding accidental inhibition

So far a number of GGT inhibitors have been reported,^4,29but

acivicin,

L-(

aS,5S)-

a

-amino-3-chloro-4,5-dihydro-5-isoxazole,

a

classical glutamine antimetabolite antibiotic produced by Strepto-

myces sviceus,³⁰has been used by far the most extensively for both

in vitro and in vivo inhibition of GGT.^31–33Acivicin, however, is not

selective for GGT and inhibits many glutamine-dependent biosyn-

thetic enzymes to show potent cyto- and neurotoxicity.^34,35The

1178

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

of the glutamine-hydrolyzing enzymes; the inhibitors that could

exploit the binding energy effectively from the Cys-Gly binding

site, like compound 1, should be bound tightly to the active site

of GGT, but not to the active site of glutamine-hydrolyzing en-

zymes that are presumably devoid of the Cys-Gly binding site.

Several X-ray crystal structures of bacterial GGTs have been

reported,^36,42–45but none of them were successful in identifying

the Cys-Gly binding site. In contrast, the mammalian enzymes,

whose natural substrate is glutathione and its S-conjugates, are

expected to have a deﬁned binding site for Cys-Gly. Very recently

the X-ray crystal structure of human GGT was solved,⁴⁶present-

ing the ﬁrst tertiary structure of eukaryotic GGT. However, the

Cys-Gly binding site is still undeﬁned by the X-ray crystal struc-

ture of human GGT in complex with glutamate, suggesting that

the structural identiﬁcation of Cys-Gly biding site is very difﬁcult

without a minutely designed ligand that could incorporate a

Cys-Gly moiety.

Here we report the synthesis and evaluation of a series of gluta-

thione-analogous peptidyl phosphorus esters 2a–g, 3 and 5 that

incorporated a Cys-Gly moiety in a transition-state analogous scaf-

fold (Fig. 1). The extent of mechanism-based inhibition depending

on the structural differences in the peptidyl moiety has gained in-

sights into the structural and stereochemical preferences of the

putative Cys-Gly binding site of human and E. coli GGTs. Human

ClAla-Gly, 2c),

alanylglycine (

L

-b-iodoalanylglycine (

-Phe-Gly, 2e), -norleucylglycine (

-Nle-Gly, 2g) and -norleucine methyl ester

-Nle-OMe, 3) (Fig. 1). The -Abu-Gly moiety of 2a is a mimic of

-Cys-Gly in glutathione with the assumption that a methyl is ste-

L

-b-IAla-Gly, 2d),

L-phenyl-

L

-Nle-Gly, 2f),

D-norleucylglycine (

D

L

(L

L

rically analogous to SH. A halogen atom was introduced (chlorine

and iodine for 2c and 2d, respectively) in the hope that the halogen

atom also mimics the SH of the

preferred acceptor substrate of rat kidney GGT,^3,47the norleucyl

series 2f ( -Nle-Gly), 2g ( -Nle-Gly) and 3 ( -Nle-OMe) were

synthesized to mimic -Met-Gly, -Met-Gly and -Met-OMe,

L-Cys residue. In view of L-Met as a

L

D

L

D

L

respectively. The synthesis of a carboxy derivative of 3 was unsuc-

cessful due to hydrolytic instability of the compound probably by

intramolecular nucleophilic catalysis by the adjacent carboxy

group.⁴⁸

The phosphonate 4 carries a dipeptide mimic of L-Abu-Gly

(same as 2a), but with a poor leaving group (OMe) in place of phen-

oxy attached to the phosphorus. Compound 5 is a phosphinate

version of 2b with a Gly-Gly side chain (see Section 2.4).

2.2. Synthesis of phosphonate diesters 2a–g, 3 and 4

The synthesis of the peptidyl moieties is shown in Scheme 2.

The preparation of the peptidyl phosphorus esters required a series

of

- or

-hydroxy acids 7a, e–g were prepared from the corresponding

a

-hydroxy acids with deﬁned stereochemistry that mimic the

GGT was highly selective for a mimic of

no acid such as -2-aminobutyrate (a

cine (a -Met mimic) at the Cys binding site, whereas E. coli GGT

signiﬁcantly preferred an aromatic amino acid ( -Phe) at this site.

Both enzymes required the C-terminal Gly for effective inhibition,

suggesting that E. coli GGT, as well as the human enzyme, still re-

L

-form of an aliphatic ami-

L

D

-amino acid at the Cys moiety of glutathione. The chiral

L

-Cys mimic) and -norleu-

L

a

L

- or D-amino acids (6a, e–g, respectively) with retention of

L

conﬁguration by diazotization followed by hydrolysis.⁴⁹The

optical rotations of 7e–g were in good agreement with the

reported values.⁴⁹The (R)-3-chloro-2-hydroxypropionic acid (7c)

was prepared from commercial (R)-3-chloropropane-1,2-diol (10)

by selective oxidation of the primary hydroxy with nitric acid.^50,51

tained some characters for binding dipeptides such as

L-amino acyl

(L-AA)-Gly. The transition-state mimicry and the phosphorus ste-

reochemistry of these inhibitors successfully allowed access to

the stereochemical course of the nucleophilic attack of the catalytic

Each of the chiral

a-hydroxycarboxylic acids 7a, c, e–g and com-

mercial grade glycolic acid (7b) was condensed with glycine benzyl

ester monotosylate (8) by the conventional peptide coupling meth-

Thr residue on the

reaction.

c-carbonyl of glutathione in the catalyzed

od to give the

a-hydroxycarboxamides 9a–c, e–g, respectively

(Scheme 2). Iodine atom was introduced by treating the chloride

9c with sodium iodide in reﬂuxing acetonitrile to give 9d in 45%

yield. Treatment of 7f with thionyl chloride in methanol gave

methyl (S)-2-hydroxyhexanoate (11) in 84% yield.

2. Results and discussion

2.1. Analogue design

The peptidyl phosphonate diesters 2a–g, 3 and 4 were synthe-

sized as shown in Scheme 3. The hetero-diesters of phosphonic

acid was prepared from the phosphonyl dichloride by stepwise

The synthesized phophonate diesters 2a–g and 3 contain the

peptidyl side chains that mimic

L

-2-aminobutyrylglycine (

L

-Abu-

-b-

Gly, 2a), glycylglycine (Gly-Gly, 2b),

L

-b-chloroalanylglycine (

L

reaction with phenol (a leaving group) and

a-hydroxycarboxa-

mides 9a–c, e–g (a peptidyl side chain) according to our previously

reported procedure.³⁹Thus, the fully protected phosphonic acid 12

(racemic form)³⁹was converted to the phosphonyl dichloride 13.

Selective substitution of a single chlorine atom of 13 was achieved

by pre-mixing 13 and phenol, followed by the addition of Et₃N at

low temperature (ꢀ65 °C) as a catalyst.⁵²The resulting phosphonyl

monochloro monophenyl ester was then allowed to react with the

a-hydroxycarboxamides 9a–g or the a-hydroxyester 11 to give the

phosphonyl hetero-diesters 14a–g and 15, respectively. This proce-

dure, however, usually accompanied the formation of the homo-di-

phenyl and homo-bis-peptidyl esters of 12 regardless of careful

operations, and gave the desired hetero-diesters 14a–g and 15 in

moderate yields after careful chromatographic separation. The

results under several reaction conditions indicated that the initial

reaction with phenol should be conducted at sufﬁciently low tem-

perature by initiating the reaction with NEt₃for substitution of a

single chlorine atom of 13. However, the initial ratio of the inter-

mediate monochloro monophenyl ester (d_P39.6) was not in agree-

ment in several cases with the product ratio of the hetero-diesters,

and the formation of the homo-bis-peptidyl esters was observed

even in the reactions where complete consumption of the

Figure 1. Peptidyl phosphorus esters 2a–g, 3,

inhibitors of GGT.

4 and 5 as mechanism-based

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1179

Scheme 2. Synthesis of a-hydroxycarboxamides 9a–g and a-hydroxyester 11.

dichloride 13 (d_P49.2) was conﬁrmed by ³¹P NMR. This observation

suggested that the intermediate monochloro monophenyl ester

(d_P39.6) reacted with the second peptidyl alcohol by substitution

with phenol, as well as with chlorine. It is noteworthy that the

formation of homo-bis-peptidyl and diphenyl esters was not

suppressed by reversing the order of reaction of 13 with the

2.3. Separation of diastereoisomers with respect to chiral

phosphorus by MPLC

Starting with racemic 2-amino-4-phosphonobutanoic acid,^53,54

the reaction with chiral -hydroxycarboxamides 9a, c–g and

-hydroxyester 11 gave the corresponding peptidyl phosphonate

a

a-hydroxycarboxamides and phenol (data not shown). Reaction

diesters 2a, c–g, 3 and 4 as a mixture of four diastereoisomers with

of the dichloride 13 with methanol and 9a by the same procedure

afforded the peptidyl methyl phosphonate 16. At this stage, the

compounds 14a–g, 15 and 16 were obtained as a mixture of

equal amounts of four diastereoisomers (for 14a, c–g, 15 and 16)

or four stereoisomers (two sets of enantiomers) (for 14b) with

respect to the newly formed chiral phosphorus and the original

chiral a-carbon of the glutamate moiety. In compound 2b, the ab-

sence of a chiral carbon in the peptidyl chain gives a mixture of

four stereoisomers (two sets of enantiomers) with respect to the

same stereogenic centers.

respect to the stereogenic centers at the

a

-carbon of the glutamate

We found in the puriﬁcation step that two sets of diastereoiso-

mers were separated for compounds 2a, c–g, 3 and 4 by reversed

phase medium pressure liquid chromatography (MPLC) (Fig. 2).

We collected each peak separately and designated peak A (slow-

eluting) and peak B (fast-eluting). In contrast, no chromatographic

separation of diastereoisomers was observed with compound 2b.

This chromatographic behavior, in view of the distal stereogenic

moiety and the phosphorous atom. No attempts were successful in

separating the diastereoisomers at this stage (see Section 2.3). Re-

moval of the benzyl protecting groups by catalytic hydrogenation

gave the desired phosphonates 2a–c, 2e–g, 3 and 4. Interestingly,

the iodine analogue 14d did not react under the conventional

hydrogenolysis conditions (H₂gas /10% Pd-C). Hence the deprotec-

tion of 14d was achieved under the nucleophilic C–O bond

cleavage conditions using anisole/aluminum chloride to give 2d

in 22% yield.

center at the

a-carbon of the glutamate moiety, suggested that

the separated diastereoisomers (peak A and B) of 2a, 2c–g, 3 and

4 were related to the difference in the stereochemistry at the chiral

⁄

phosphorous (S_Por R_P^⁄) relative to the ﬁxed stereogenic center at

the adjacent peptidyl moiety. It seems that the stereogenic center

(L

a

or

D

a

) at the

a

-carbon of the glutamate moiety is not close en-

ough to the chiral phosphorous and/or the peptidyl -carbon to ef-

a

fect diastereoisomeric differences in the MPLC behavior. Hence

each peak A and B is composed of a mixture of two diastereoiso-

mers with respect to the distal chiral a-carbon (L or D ) of the glu-

a a

tamate moiety. This view was supported by the spectroscopic data

for peak A and B; each gave a single peak or very close two peaks

(equal height) in ³¹P NMR and an apparently single set of peaks in

¹H NMR. We therefore concluded that the separated diastereoiso-

meric pairs (peak A and B) of 2a, 2c–g, 3 and 4 were related to

the chiral phosphorus and that the diastereoisomers belonging to

peak A had the same stereochemistry at the chiral phosphorus

(S_P^⁄, see below), but it was opposite to that of the diastereoisomers

in peak B (R_P^⁄). Note that each peak A and B is still composed of

equal amounts of two diastereoisomers with opposite stereochem-

istry (

L

or D ) at the chiral a-carbon of the glutamyl moiety, but

a a

they behaved like enantiomers due to the distance of this chiral

center. For the same reason, the diastereoisomers of compound

2b behaved like enantiomers and were not separated by MPLC to

give a mixture of four stereoisomers (two sets of enantiomers)

with equal amounts.

For the phosphonate diesters 2a, c–g and 3, the slow-eluting

peak A invariably showed high inhibitory activities, whereas the

fast-eluting peak B exhibited little or no activity (see Sections 2.5

Scheme 3. Synthesis of phosphonate diesters 2a–g, 3 and 4.

1180

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

yield. The crude product was treated as such with N,O-bis(trimeth-

ylsilyl)acetamide (BSA), and the resulting trivalent trimethylsiyl

(TMS) phosphite was allowed to react by conjugate addition with

t-butyl acrylate to give the (S)-phosphinic acid derivative 20

quantitatively. We initially tried to react with ethacrylamide 21

to synthesize directly the corresponding phosphinate derivative

with a ^DL-Abu-Gly analogous peptidyl chain, but the reaction did

not work with less reactive Michael acceptor ethacrylamide 21.

Hence we used the sequential three-step reaction involving the Mi-

chael addition to tert-butyl acrylate to give 20, followed by depro-

tection of the tert-butyl ester and condensation with the glycine

benzyl ester monotosylate (8) to give the full-length phosphinic

acid 23 in a total yield of 50% from 18. Treatment of 23 with oxalyl

chloride followed by reaction with phenol gave the phosphinate

ester 24 in 55% yield. Removal of the benzyl protecting groups by

catalytic hydrogenolysis afforded the desired phosphinate ester 5

with a Gly-Gly mimic peptidyl chain. Compound 5, synthesized

Figure 2. Separation of diastereoisomeric pairs (A and B) of 2f with MPLC. See

Section 4 for conditions of chromatography.

and 2.6). For this reason, we used the inhibitory activities of peak A

(a mixture of two diastereoisomers) in the following discussion.

For compound 2b, the inhibitory activity was evaluated for a mix-

ture of four stereoisomers. The absolute conﬁguration of the chiral

phosphorus was _⁄not dete_⁄rmined experimentally, but was assigned

tentatively as S_Pand R_Pfor the isomers in peak A and peak B,

respectively, from our preliminary results on the X-ray crystal

structural analysis of E. coli GGT in complex with 2a-A (Supple-

mental Fig. S1, see Section 2.5 for details).

from L-vinylglycine derivative 18, was composed of equal amounts

of two diastereoisomers with respect to the chiral phosphorus, but,

as seen for compound 2b, the diastereoisomers of 5 behaved like

enantiomers and were not separable by reversed phase MPLC

due to the absence of a ﬁxed stereogenic center at the

a-carbon

of the adjacent peptidyl moiety. Hence we used compound 5 for

the following enzyme assays as a 1:1 mixture of two diastereoiso-

mers with opposite stereochemistry at the chiral phosphorus.

Compound 5 is the phosphinate version of compound 2b, but it

should be noted that compound 2b, synthesized from racemic 2-

amino-4-phosphonobutanoic acid, was a mixture of four stereoiso-

2.4. Synthesis of phosphinate ester 5

The phosphinate 5 was synthesized as shown in Scheme 4. The

mers (two sets of enantiomers) with respect to the stereogenic

protected L-glutamic acid 17 was converted to the L-vinylglycine

⁄

centers on the phosphorus (R_Por S_P^⁄) and the distal

a

-carbon (

L

derivative 18 by the reported procedure.⁵⁵Phosphorus atom was

introduced by triethylborane-catalyzed radical addition of ammo-

nium hypophosphite⁵⁶to give the hydrophosphinic acid 19 in good

a

or

D ) at the glutamyl moiety, whereas compound 5 was a mixture

a

of two diastereoisomers (L L

, R_P^⁄) and ( , S_P^⁄). For clarity, the

a a

Scheme 4. Synthesis of phosphinate ester 5.

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1181

Table 1

Composition of stereoisomers of 1, 2a–g, 3, 4 and 5 and their stereochemistry

Compound

Composition^a

Stereochemistry^b

⁄

, R_P, D

0

)

1

Four stereoisomers^c

Two diastereoisomers

Four stereoisomers

Two diastereoisomers

(L

a

(L

a

(L

a

(L

a

(L

a

(L

a

(L

a

, S_P

, R_P

,

L

), (

L

, R_P

,

D

), (

)

D

, S_P

,

L

), (D

a

⁄

0

2a, c–f, 3 and 4

Peak A

Peak B

D

, S_P

, R_P

L

a

0

L

a

, S_P^⁄), (

L

), (D

a

⁄

2b

2g

, S_P^⁄), (

L

, R_P^⁄), ( _⁄

D

a

, R_P

)

a

), (

a

⁄

0

Peak A

Peak B

, R_P

,

D

, R_P

,

D

)

a

⁄

0

, S_P

,

D

, S_P

,

D

)

a

⁄

5

, S_P^⁄), (

L

, R_P

)

a

The ratio of stereoisomers _⁄is almost equal (1:1) in each case.

b

or D

), the chiral phosphorus (S_P^⁄or R_P^⁄) and the proximal

a

a-

0

Stereochemistry [ex. (

L

, S_P

,

L

a

)] is indicated in the order of the distal

a

-carbon of the glutamate moiety (

L

a

⁄

0

carbon of the peptidyl moiety (

L

or

D

a

). Inhibitory active stereoisomers [ex. (

L

, S_P

,

L

)] are shown in bold (see Sections 2.5 and 2.6).

a

c

See Section 2.5.

stereoisomeric composition and the conﬁguration of each isomer

for compounds 1, 2a–g, 3, 4 and 5 are summarized in Table 1.

The stereoisomeric composition of compound 1³⁹will be discussed

in Section 2.5. The stereochemistry of peak A and B of compound

2g is worth noting. This compound was synthesized from (R)-a-

hydroxycarboxamide 9g, a mimic of D-Nle-Gly, and its stereochem-

0

istry (

D

) at the

a

-carbon of the peptidyl chain adjacent to the

a

0

phosphorus was opposite to that of compound 2f (

L

a

) derived from

(S)-a-hydroxycarboxamide 9f. Therefore, the chromatographic

⁄

, D _’) isomer of 2g on MPLC should be similar

a

, L _’) isomer of 2f, because, due to the distance

a

behavior of the (R_P

to that of the (S_P

⁄

of the remaining stereogenic center (

L or D ) at the a-carbon of

a a

the glutamate moiety, these two sets of diastereoisomers are pseu-

do-enantiomorphic and hence are expected to behave like enanti-

omers on MPLC. We therefore assigned the stereochemistry at the

chiral phosphorus in peak A (slow-eluting) and peak B (fast-elut-

⁄

ing) as R_Pand S_P^⁄, respectively, for compound 2g. Thus_⁄, the peak

A of 2g was composed of two diastereoisomers (

L

, R_P

,

D

_’) and

a

⁄

(D

, R_P

,

D _’), and peak B was a mixture of two diastereoisomers

a

⁄

D

(L

a

, S_P

,D

a

_’) and (

D

, S_P

,

_’). The diastereoisomers contained in each

a

peak A and B of 2g were pseudo-enantiomorphic and were not sep-

arable by MPLC.

2.5. Inhibitory activities toward human GGT

The inhibitory activities of the phosphorus esters 2a–g, 3, 4 and

5 were evaluated against human and E. coli GGTs according to

our previous method.^38,39Enzyme activity was measured using

7-(c-L-glutamylamino)-4-methylcoumarin (c

-Glu-AMC)⁵⁷as sub-

strate, and release of 7-amino-4-methylcoumarin (AMC) was de-

tected ﬂuorimetrically and continuously as the progress of

enzyme reactions. We employed a hydrolytic condition without

using a

c

-glutamyl acceptor substrate such as Gly-Gly,³because

Figure 3. Typical time-dependent inhibition of GGT by 2a-A at pH 5.5 with c-Glu-

a high concentration of Gly-Gly may compete with the peptidyl

moiety of the phosphorus esters 2a–g, 3, 4 and 5 for the enzyme

Cys-Gly binding site, thereby complicating the inhibition kinetics.

To suppress the possible auto-transpeptidation,^3,4the reaction

was conducted under an acidic condition (pH 5.5), where the

transpeptidation activity is relatively low,⁵⁸and by using lower

AMC as substrate. (A) Progress curves for reactions catalyzed by human GGT with

4.0 -Glu-AMC in the presence of varying concentration of 2a-A: (a) 0, (b) 4, (c)

10, (d) 12, (e) 16 and (f) 20 M. (B) Progress curves for reactions catalyzed by E. coli

GGT with 0.2 -Glu-AMC in the presence of varying concentration of 2a-A: (a) 0,

(b) 10, (c) 20, (d) 30, (e) 40 and (f) 60 M. The formation of product AMC was

lM c

l

lM c

l

monitored continuously by ﬂuorescence (see Section 4.3 Enzyme assay).

concentrations of substrate (4.0 and 0.2

lM of c-Glu-AMC

inhibition curves were ﬁt to the ﬁrst-order rate equation (1) (see

Section 4.3.2) to determine the observed pseudo-ﬁrst-order rate

constant for enzyme inactivation (k_obs) at each inhibitor concentra-

tion. A plot of k_obsagainst inhibitor concentration gave a straight

line without saturation until the highest inhibitor concentration

(data not shown), indicating that the reversible formation of EI

complex was kinetically incompetent under the reaction conditions

employed. Hence the k_onvalues, the second-order rate constants for

enzyme inactivation, were calculated from the slope on the

assumption that the inhibitor was competitive with the substrate

[Eq. (2) in Section 4.3.2]. The results are summarized in Table 2.

for human and E. coli GGT, respectively), which correspond to

1/3 and 1 K_mvalue for human and E. coli GGT, respectively.

The phosphonate diesters 2a–f (peak A, except 2b) with a good

leaving group (phenoxide) served as potent and irreversible inhib-

itors of human GGT. Typical progress curves for the reaction cata-

lyzed by human GGT in the presence of varying concentrations of

compound 2a-A are shown in Figure 3A. The inhibitory activities

were evaluated by calculating the second-order rate constant

(k_on) for enzyme inactivation, assuming a bimolecular reaction of

enzyme and inhibitor to form a covalent enzyme-inhibitor covalent

complex (E-I) (see Section 2.7, Fig. 5). Thus, the time-dependent

1182

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

Table 2

reaction with glutathione (Scheme 1). A methyl as a good mimic of

SH was also shown in the sulfoximine-based transition-state ana-

Inhibitory activities of peptidyl phosphonate diesters 1,³⁹2a–g, 3, 4 and phosphinate

ester 5 toward human and E. coli GGTs^a

logue inhibitors of

c-glutamylcysteine synthetase, the ﬁrst and

b

Compound

Peptidyl moiety mimicking

k_on(M^ꢀ1s^ꢀ1

)

the rate-limiting enzyme in glutathione biosynthesis.⁵⁹We previ-

ously reported that the parent compound 1, a mixture of stereoiso-

Human GGT

E. coli GGT

mers synthesized from racemic

a

-hydroxybutyrate, inactivated

Assuming that the

1

75 3^c

80 2^c

DL-Abu-Gly

human GGT with a k_onvalue of 75 M^ꢀ1s^{ꢀ1 39}

.

2a

145 5.0

109 2.9

L-Abu-Gly

inhibition of human GGT is highly dependent on the stereochemis-

try at the chiral phosphorus, which will be discussed later, the

inhibitory activity is then proportional to the stereoisomeric com-

position. The observed activity of 1 (75 M^ꢀ1s^ꢀ1) was about a half

that of 2a that was a mixture of two diastereoisomers, suggesting

that the previously synthesized 1 was a mixture of four stereoiso-

mers (Table 1) rather than eight, probably due to the isolation of

peak A solely during the chromatographic puriﬁcation by MPLC.

2b

2c

Gly-Gly

44.5 2.5

40.3 0.59

80.6 1.2

188 14

L

-b-ClAla-Gly

-b-IAla-Gly

-Phe-Gly

2d

2e

2f

43.6 2.0

4.44 0.28

63.2 0.76

98.4 2.9

389 24

50.7 0.66

-Nle-Gly

NI^d,e

1.7^e,f

2g

NI^d

1.3^f

D-Nle-Gly

NI^d,e

NI^d

NI^d,e

8.83 0.19

NI^d

⁄

D

3

4

L

-Nle-OMe

-Abu-Gly

Therefore compound 1 was composed of (

L

, S_P

,

L

a

_’), (

L

, R_P

,

_’),

a

⁄

D

L

(D

a

, S_P

,

L

a

_’) and (

D

, R_P

,

_’) as shown in Table 1.

a

5

Gly-Gly

—

1.35 0.11

0.40 0.02^h

NI^g

Removal of the ethyl group from the L-Abu-Gly peptidyl chain

signiﬁcantly decreased the inhibitory activity (2b). It should be

noted that 2b was racemic at both the chiral phosphorus and the

Acivicin

4200 100^h

a

The activities of peak A (a mixture of two diastereoisomers, Table 1) are shown

for 2a, 2c–g, 3 and 4. The inhibitory activities of peak B are also shown for 2f and 2g

(footnote e). The inhibitory activities of peak B were extremely low (see Sections 2.5

and 2.6). For compound 2b, the activities of a mixture of four stereoisomers (two

sets of enantiomers, Table 1) were determined. Compound 5, the phosphinate

version of 2b, was a mixture of two diastereoisomers (Table 1).

a

-carbon at the distal glutamate moiety, and was composed of

equal amounts of four stereoisomers (Table 1). With an

assumption that the only S_P^⁄isomer is active, the observed k_onva-

lue of 2b (a mixture of four) should be doubled when compared to

b

that of 2a (a mixture of two). The corrected value (89.0 M^ꢀ1s^ꢀ1

)

Second-order rate constant for enzyme inactivation (see Section 4.3.2).

Ref. 39.

No inactivation at 1 mM.

Inhibitory activity of Peak B (Table 1)

Error limit was not determined due to duplicate measurements

No inactivation at 1.5 mM.

Ref. 38.

c

was still 61% of the k_onvalue of 2a, indicating a signiﬁcant

contribution of the ethyl group to the enzyme inhibition. Interest-

ingly, the introduction of a chloromethyl (2c) or iodomethyl group

(2d) caused almost no difference in the inhibitory activity as

compared to 2b. We initially anticipated that a chlorine and iodine

atoms might also work as a good mimic of SH, because the van der

Waals volume of SH (14.8 cm³mol^ꢀ1) is similar to that of Cl

d

e

f

g

h

Among the inhibitors synthesized, the phosphonate 2a exhibited

the highest inhibitory activity (k_on= 145 M^ꢀ1s^ꢀ1) toward human

GGT. This value, compared with the inhibitory activity of acivicin

(k_on= 0.40 M^ꢀ1s^ꢀ1) measured under the same reaction conditions,

indicated that 2a was an extremely potent irreversible inhibitor of

human GGT that inactivated the enzyme more than 360 times as

quickly as acivicin. This observation was in accordance with our ini-

(11.62 cm³mol^ꢀ1

) and I

(19.18 cm³mol^ꢀ1), as well as CH₃

(13.67 cm³mol^ꢀ1).⁶⁰However, the introduction of a chloromethyl

(2c) and iodomethyl (2c) had almost no impact on the activity of

the inhibitors, whereas the introduction of an ethyl (2a) (CH₃as

SH mimic) increased the activity by 63% as compared to 2b.

This observation might be due to a little too small and large in

size of chlorine and iodine atom, respectively, for a SH mimic,

with a methyl group in between being a good compromise.

In this sense, a bromine atom with a van der Waals volume of

14.40 cm³mol^ꢀ160might be a reasonably good mimic of SH.

Human GGT thus strictly recognized the size of the substituent

tial expectations that the

L-Abu-Gly peptidyl chain in 2a was highly

analogous to the -Cys-Gly moiety in glutathione to favor its bind-

L

ing to the active site of human GGT, to facilitate covalent bond for-

mation with the catalytic Thr residue and to give the adduct that

was a mimic of the putative transition state (TS-1) for the catalyzed

A

C

B

D

Figure 4. Stereochemistry of the attack of the catalytic nucleophile of human GGT (Thr residue) on the chiral phosphorus of the inhibitor 2f-A (A to B) as a model for

catalyzed reaction of glutathione (C to D).

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1183

attached to the b-carbon of the side chain of the Cys moiety and

that a SH group or its equivalent in size seems preferable at this

site. This is highlighted by a signiﬁcantly low activity of 2e with

Since the phosphonate inhibitors 2a, c–f are a good mimic of the

putative transition state (TS-1) for the catalyzed reaction of gluta-

thione (Scheme 1), the stereochemistry at phosphorus is closely re-

lated to the stereochemical outcome of the catalyzed reaction with

glutathione. This is illustrated in Figure 4. In the active site of hu-

man GGT, the isomer(s) 2f-A with S_P^⁄conﬁguration at phosphorus

a side chain mimicking L-Phe-Gly. This compound with a bulky

phenyl group attached to the b-carbon was found to be an

extremely weak inhibitor of human GGT (only 3% of compound

2a). This result is consistent with the acceptor substrate speciﬁcity

is accommodated and reacted with the catalytic Thr residue (cOH)

of rat kidney GGT, where

L

-Phe-Gly was a signiﬁcantly poor

that approaches the phosphorus opposite to the phenoxy leaving

group and then substitutes the phenoxy in line with inversion of

conﬁguration at phosphorus, thus making an enzyme-inhibitor

covalent complex with S_P^⁄conﬁguration (panel A to B). In contrast,

dipeptide substrate.³It is noteworthy, however, that elongation

of the side chain of 2a by two methylene units did not

severely affect the inhibitory activity; compound 2f with

a

⁄

L

-Nle-Gly-analogous side chain inhibited human GGT with a k_on

the isomer(s) 2f-B with R_Pconﬁguration at phosphorus is not

value of 63.2 M^ꢀ1s^ꢀ1that was 44% of that of 2a. This result, com-

reactive probably because of the unfavorable conﬁguration for

accommodation in the enzyme active site and/or for in-line substi-

tution by the catalytic Thr residue. If this stereochemistry holds for

bined with the moderate activities of 2c and 2d, and remarkably

low activity of 2e, indicated that the c-position of the side chain

that is usually occupied by the SH of Cys was strictly recognized

by the enzyme, whereas the group beyond this position might

be forced out of the active site, thereby escaping from the steric

constraint and recognition by the enzyme. Furthermore, the ste-

the catalyzed reaction, then the nucleophilic attack of the Thr

cOH

occurs on the Si-face (above the plane) of the -carbonyl of gluta-

c

thione to make a tetrahedral oxyanion with S conﬁguration (panel

C to D). This stereochemical course seems quite reasonable from a

reochemistry at the

of the C-terminal Gly had a signiﬁcant impact on the inhibition

activity: the inversion of stereochemistry at the -carbon of the

Cys moiety (2g) or removal of the C-terminal Gly (3) resulted in

a complete loss of inhibitory activity (Table 2). Thus, human GGT

was inactivated strongly only by the phosphonate with a peptidyl

side chain that is structurally and stereochemically equivalent to

a

-carbon of the Cys moiety and the presence

steric point of view: the catalytic Thr

cO approaches the phospho-

rus (for 2a-A) or the -carbonyl (for glutathione) from the less-hin-

c

a

dered side that is opposite to the L-Cys side chain and makes a

tetrahedral (S_P^⁄)-adduct (for 2a-A) or (S)-oxyanion (for glutathi-

one) in which the O-Thr is located above the plane, while the side

chain of the L-Cys moiety occupies below (Fig. 4 B and D). Since the

phosphorus stereochemistry of the active inhibitor is assigned

L

-Cys-Gly. This is consistent with the observation that GGT accepts

tentatively as S_P^⁄, we cannot rule out the possibility that all the

only -amino acids or peptides as

L

c

-glutamyl acceptor substrate³

stereochemistry shown in Figure

4

would be reverse_⁄d: the

and is in good agreement with the putative substrate recognition

and inhibition mechanisms proposed by us that human GGT

strongly recognizes the negative charge at the C-terminal Gly of

glutathione.³⁹

active-site Thr cOH attacks from the bottom side on the R_Pphos-

phorus of 2a-A to make a (R_P^⁄)-adduct, which corresponds to the

attack of Thr OH on the Re-face (below) of the -carbonyl of

c

glutathione to make a (R)-tetrahedral intermediate. This mode of

We found that peak A of compounds 2a, c–f was invariably

active, whereas peak B was completely inactive towards human

GGT. These results indicated that human GGT strictly recognized

the stereochemistry at phosphorus; the isomers A (tentatively

assigned as S_P^⁄) was exclusively accommodated and/or reacted

with the active-site Thr residue, leading to enzyme inactivation.

The assignment of the conﬁguration at phosphorus was derived

from our preliminary results on the X-ray crystallographic anal-

ysis of E. coli GGT in complex with 2a-A, in which S_Pconﬁgura-

tion was observed at the chiral phosphorus that was covalently

reaction, however, appears unfavorable from a stereochemical

point of view, because Thr

cOH attacks the phosphorus or the

c-carbonyl from the same side of the Cys side chain. The deﬁned

stereochemical assignments have to await the X-ray structural

analysis of human GGT in complex with 2a-A.

In this model, we also proposed the mode of recognition of the

Cys-Gly moiety; human GGT has a speciﬁc binding site for the Cys

0

moiety to recognize strictly its stereochemistry (

L ) and the size of

a

its side chain up until the sulfur atom probably by a narrow cavity

around this part, but no recognition is inferred for any group be-

yond that position. This mode of recognition of the Cys residue is

reasonable for GGT that accepts a variety of glutathione-S-conju-

gates as substrates. This is illustrated by the reaction of GGT with

bound to the O

c atom of Thr391 (Supplemental Fig. S1). Since

both human and E. coli GGTs were inactivated by the same iso-

mers (peak A), the stereochemical preference of both enzymes

for the chiral phosphorus is the same (see Section 2.6). We

therefore tentatively assigned the original stereochemistry at

2f-A (Fig. 4A and B) where the

L-Nle residue is bound or recognized

by the enzyme up until the -carbon. No stereoisomers with

c

D

-con-

⁄

phosphorus of 2a-A (peak A) as S_Pby assuming the in-line sub-

ﬁguration at the Cys moiety are permitted for binding and reaction

with the enzyme active site. In addition, the C-terminal carboxy

group of the Gly moiety is strictly recognized by an electrostatic

interaction with a positively charged active-site residue.³⁹The ac-

tive-site residue that recognizes the negative charge of the accep-

tor substrate in the transpeptidation reaction was recently

reported to be Lys562.⁶¹Since the peptidyl chain of the present

phosphonate inhibitors is equivalent to the acceptor substrate

and the Cys-Gly moiety of glutathione, this residue is probably

responsible for the recognition of the C-terminal negative charge

in the present peptidyl phosphonate inhibitors, as well as that of

the C-terminal Gly of the substrate glutathione.

stitution at phosphorus with inversion of conﬁguration [Note

that the phosphorus stereochemical assignments for the phenyl

phosphonate (before reaction) and the enzyme-inhibitor complex

(after reaction) are the same (S_P^⁄) due to the sequence rule for

the ligands around the phosphorus, irrespective of the inversion

of conﬁguration].

Table 2 also highlights the strict stereochemical recognition by

human GGT. Neither peak A nor peak B of the phosphonate 2g was

inhibitory toward human GGT, indicating that none of the four

stereoisomers of 2g (Table 1) was acceptable by the enzyme irre-

spective of the conﬁguration at phosphorus. This was due to the

D

-conﬁguration at the

a

-carbon of the Cys moiety of 2g. Con-

A good leaving group such as phenoxy attached to the phospho-

rus was absolutely necessary for enzyme inactivation. The methyl

versely, peak B of compound 2f was totally inactive irrespective

of the

L

-conﬁguration at the

a

-carbon of the Cys moiety, because

ester 4 was totally inactive toward the enzyme, irrespective of the

⁄

the isomers in peak B had the R_Pconﬁguration at phosphorus.

presence of S_Pphosphorus and

L-Abu-Gly peptidyl chain. This is

Therefore human GGT is highly selective for S_P^⁄-phosphorus and

naturally understood by the inactivation mechanism; the electro-

philic phosphorus with a good leaving group reacted with the ac-

L-

a-carbon at the Cys moiety in the peptidyl chain, and does not ac-

cept any other stereoisomers.

tive-site Thr

cOH to phosphonylate and inactivate the enzyme

1184

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

(see Section 2.7 and Fig. 5). The structure of 4 is highly analogous

to glutathione, but the enzyme activity was not affected apprecia-

bly even in the presence of a high concentration (1 mM) of com-

pound 4 (Supplementary, Fig. S2). This result indicated that,

contrary to our initial expectations, the afﬁnity of the phosphonate

4 with the enzyme active site was substantially low. This is prob-

ably due to the sp³-hybridized tetrahedral phosphorus carrying a

methoxy group that is substantially different in size and shape

chiral phosphorus, the E. coli enzyme was not highly selective,

either; E. coli GGT was inactivated slowly by 2f-B and 2g-A both

with R_P^⁄phosphorus, whereas these compounds were totally inac-

tive toward human GGT. It appears therefore that E. coli GGT does

not particularly prefer

ing site, and that the enzyme does not strictly recognize the stereo-

chemistry at chiral phosphorus and the -carbon of the Cys

L-Cys or its equivalent at the Cys-Gly bind-

a

moiety. The properties of E. coli GGT suggest that the natural and

preferred substrates of E. coli GGT are not glutathione or its S-con-

from the sp²-hybridized

c-carboxamide of glutathione and from

the tetrahedral oxyanion. If this is the case, the initial binding of

2a is expected to be unfavorable, although the enzyme was inacti-

vated by the phosphonate rather fast. In fact, the K_ivalue for a

related phosphonate inhibitor was 0.17 mM.³⁹

jugates, but other exogenous

cessed by this enzyme as a nutrition source.⁶⁵According to the

substrate preference of the E. coli enzyme, such -glutamyl pep-

tides presumably contain an aromatic moiety ( -Phe, Tyr, Trp and

so on) or its equivalent at the peptide moiety. The identiﬁcation

of those -glutamyl peptides is intriguing in terms of natural

occurrence of novel -glutamyl compounds.

c-glutamyl amides that can be pro-

c

L

The phosphinate inhibitor 5, a closely related analogue of

the phosphonate 2b, showed

a

substantially low activity

c

(k_on= 1.4 M^ꢀ1s^ꢀ1) toward human GGT. The phosphinate 5 was a

mixture of two diastereoisomers, while the phosphonate 2b was

four. Hence the net inhibitory activity of 5 was even worse and

no more than 1.6% of that of 2b. Despite its low activity, compound

5 caused a time-dependent inhibition of human GGT, giving a ser-

ies of concave upward progress curves in the inhibition assay (Sup-

plementary Fig. S3). This is indicative of covalent modiﬁcation of

the enzyme active site by the phosphinate 5, but the reaction

was substantially slow. The signiﬁcantly lower activity of phosph-

inate 5 as compared to the phosphonate counterpart 2b was totally

unexpected from our initial expectations that the phosphinate es-

ter 5 would be more reactive than the phosphonate ester 2b due to

c

The relatively non-selective nature of the E. coli enzyme is sup-

ported by the inhibitory activity of 3 that is devoid of the C-termi-

nal Gly moiety. This compound was totally inactive toward human

GGT, but an appreciable activity (k_on= 8.8 M^ꢀ1s^ꢀ1) was observed

with E. coli GGT. Compound 3, however, was substantially less ac-

tive (ca. 17%) than 2f, implying that the C-terminal Gly still con-

tributed noticeably to the afﬁnity of the inhibitor in the active

site. This result is indicative of a putative Cys-Gly binding site still

conserved in E. coli GGT. Several X-ray crystal structural analyses

have been reported for bacterial GGTs,^42–45,66but none of them

was successful in identifying the Cys-Gly binding site. In the crystal

structures of E. coli GGT, the detection of intact glutathione failed

even after the soaking of the enzyme crystals into a glutathione

solution for a short period followed by rapid freezing.⁴²An attempt

to detect the structure of intact glutathione in complex with

Helicobactor pylori GGT was made by using a catalytically inactive

mutant enzyme (Thr380Ala), but the Cys-Gly portion of the en-

zyme-bound glutathione was invisible by X-ray structural analysis

the absence of the ground-state resonance stabilization by p –d

p

conjugation with the oxygen and phosphorus.⁶²We therefore

speculated that, in the reaction of the active-site Thr OH and

c

the phosphorus inhibitors, the inductive (-I) effect by an electro-

negative oxygen had a larger impact on the reactivity than the

p –d conjugation resonance effect. We reported previously that

p

the reaction of GGT with phosphonate-diester based inhibitors

proceeded via a dissociative transition state in which a substantial

P–O bond cleavage is inferred.³⁹In this transition state, the posi-

tively charged phosphorus atom might be stabilized effectively

by an electron-donating resonance effect of the oxygen to facilitate

the reaction.

due to considerable mobility beyond the c

-glutamyl group.⁴⁴These

results suggested that no deﬁned Cys-Gly binding site was present

in bacterial GGTs. However, the properties of E. coli GGT as probed

by the present phosphonate inhibitors cogently suggested the

presence of a putative or remaining characters of the binding site

for the Cys-Gly moiety that accommodates the amino acyl

2.6. Inhibitory activities toward E. coli GGT

(AA)-Gly moiety of c-glutamylamide substrates or an acceptor

substrate AA-Gly. In this regard, we could exploit the differences

in the structural and stereochemical preferences of enzymes in

the Cys-Gly binding site for generating selective inhibitors; a

Under the same reaction conditions, the activities of 2a–f, 3, 4

and 5 (peak A, except 2b) were evaluated for the inhibition of

E. coli GGT. The peak A of the phosphonate diesters 2a, c–f and 3

invariably caused time-dependent inhibition of E. coli GGT

(Fig. 3B), whereas peak B was uniformly inactive.

phosphonate diester, for example, with a D-AA-Gly or L-AA side

chain that incorporates an appropriate aromatic substituent at

the AA moiety would be selective for E. coli GGT without inhibiting

the human enzyme.

In sharp contrast to human GGT, the E. coli enzyme was strongly

inactivated by 2e with a

was more than 3.5 times as active as 2a. This is consistent with

the fact that aromatic -amino acids such as -Phe, -His and

L

-Phe-Gly analogous peptidyl chain that

We previously reported the inhibitory activities of a series of

simple phosphonate diesters with a substituted phenol as a leaving

group.³⁹The E. coli enzyme was uniformly two orders of magnitude

more sensitive to those inhibitors than human GGT. It is worth not-

ing that in the present phosphonate inhibitors 2a–f with a peptidyl

side chain, the inhibitory activities (k_onvalues) toward E. coli GGT

were not so much different from those toward the human enzyme

(Table 2). These observations can be explained by the difference in

the extent of recognition of the Cys-Gly moiety by the enzymes;

human GGT is highly selective for the Cys-Gly moiety of inhibitors

and is strongly inhibited only by the inhibitors with an appropriate

peptidyl chain or its equivalent that ﬁts the deﬁned Cys-Gly

binding site, whereas the E. coli enzyme is relatively unselective

for the Cys-Gly moiety and is inhibited generally by simple phos-

phonates without the Cys-Gly peptidyl moiety. Hence the phos-

phonate diesters without an appropriate Cys-Gly mimic peptidyl

chain are rather weak inhibitors of human GGT, but can be highly

inhibitory toward the E. coli enzyme. The present peptidyl

L

-DOPA serve as a good acceptor substrate of E. coli GGT in trans-

peptidation.⁶³Inhibitors 2a–d and 2f inactivated E. coli GGT with

k_onvalues ranging from 50.7 to 188, but no clear relationship

was observed between their inhibition potency and the structure

of the side chains at the Cys moiety.

nearly twice as good as

-Phe as an acceptor substrate,⁶³but the

inhibitor 2f with -Nle-Gly mimetic peptidyl chain (k_on=

L-Met was reported to be

L

a

L

50.7 M^ꢀ1s^ꢀ1) was only13% as active as 2e. A hydrogen bond

involving a sulfur atom⁶⁴might be playing a role in the recognition

of Met as noted for capD, a GGT family enzyme from Bacillus

anthracis.⁶¹

In contrast to human GGT, the E. coli enzyme was inactivated

slowly by 2g with a

E. coli enzyme was not highly selective for the

the -carbon of the Cys-Gly peptidyl chain. With respect to the

D-Nle-Gly peptidyl chain, suggesting that the

L-conﬁguration at

a

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1185

phosphonates 2a–f have the Cys-Gly mimicking peptidyl chain and

hence are potent inhibitors of human GGT as well as the E. coli

enzyme, if other requirements such as stereochemistry at the

phosphorus and the side chain of the peptidyl moiety are met. This

view is consistent with the extremely weak inhibition of human

GGT by acivicin (k_on= 0.4 M^ꢀ1s^ꢀ1) that lacks the Cys-Gly moiety,

whereas the E. coli enzyme is highly sensitive to this compound

(k_on= 4200 M^ꢀ1s^ꢀ1) (Table 2).³⁹In this regard, the recently re-

ported inhibitor OU749 and its analogues by Hanigan et al. bound

to the acceptor substrate binding site can be an attractive option

for selective inhibition of mammalian GGTs,^67–69but their revers-

ible nature with micro-molar K_ivalues might be insufﬁcient for

complete elimination of GGT activities for a certain period of time

in vitro and in vivo as was observed with acivicin.

The phosphonate 4 with a poor leaving group and the phosph-

inate 5 did not inhibit E. coli GGT at all. Since the phosphinate 5

was weakly inhibitory toward human GGT, it stands out as the sole

compound that inhibited human GGT selectively among the phos-

phorus ester-based inhibitors. The structure of 5 might be a good

template for designing novel inhibitors selective for human GGT.

enzyme-inhibitor adduct that mimics the putative transition state

(TS-1) for glutathione degradation (Scheme 1). For this purpose,

the Cys-Gly mimicking peptidyl chain has to remain bound to

the phosphorus after the reaction with the enzyme active site. This

was conﬁrmed by an ion-spray mass spectroscopy of E. coli GGT

inactivated by compound 2a-A. We used E. coli GGT rather than

the human enzyme, because highly and heterologously glycosyl-

ated human GGT would make the mass analysis very complex.

E. coli GGT was treated with 2a-A under varying conditions, and

the small subunit from each experiment was separated by HPLC.⁷⁰

The isolated small subunit (GGTS, ca. 20,000 Da) was analyzed by

electrospray ionization mass spectroscopy (ESI-MS) according to

our previous method for detecting the catalytic residue (Thr391

in GGTS) of E. coli GGT.⁴⁰The untreated E. coli GGT was also

processed in the same manner as reference. The results are shown

in Figure 5. The GGTS from the untreated enzyme gave a peak at

m/z 20,014 (Fig. 5A) that corresponds to the expected molecular

weight (20,010) of the unmodiﬁed small subunit calculated from

its amino acid sequence.⁷⁰The small subunit obtained from the en-

zyme treated under a weak reaction condition (1.7 mM 2a-A,

20 min) gave two peaks at m/z 20,015.5 and 20,323 (Fig. 5B). Final-

ly, the small subunit from the enzyme with a stronger treatment

(4 mM 2a-A, 3 h) gave a single peak at m/z 20,323 (Fig. 5C). The

mass difference (307.5) between the two peaks corresponded well

to the calculated mass (308) of the inhibitor that is bound cova-

2.7. Mass spectroscopic analysis of E. coli GGT in complex with

inhibitor

The structural deﬁning of the Cys-Gly binding site of human

GGT is one of the most challenging, but the most rewarding sub-

jects in the studies on GGT. Directing towards this goal, the pepti-

dyl phosphonate inhibitors 2a–f was designed also for use as a

ligand for X-ray structural analysis of human GGT in view of the

lently to the catalytic Thr Oc through a P-O bond with concomitant

loss of the phenoxy group as shown in Scheme 1C. These results

clearly indicated that the inhibitor 2a-A reacted with the active site

of the enzyme to leave the Cys-Gly mimicking peptidyl moiety

A

B

C

Figure 5. Mass spectra of E. coli GGT small subunit (GGTS) incubated with 2a-A. (A) 0 mM, (B) 1.7 mM, 20 min, (C) 4.0 mM, 3 h.

1186

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

bound to the active site of GGT. Therefore the peptidyl

phosphonate inhibitors 2a–f are highly promising as a ligand for

X-ray structural analysis of GGT for deﬁning the Cys-Gly binding

site.

Racemic benzyl 2-(N-benzyloxycarbonylamino)-4-phosphonobut-

anoate (12) was synthesized as previously described.³⁹Dry CH₂Cl₂

and dry triethylamine (Et₃N) were prepared by distillation from

CaH₂and stored over Molecular Sieves 4 Å. E. coli

peptidase was puriﬁed from the periplasmic fraction of a recombi-

nant strain of E. coli K-12 (SH642).⁶³Human

-glutamyl

c-glutamyl trans-

c

3. Conclusions

transpeptidase HC-GTP (T-72) was a generous gift from Asahi Kasei

Corporation (Osaka, Japan). The enzyme preparation contained a

signiﬁcant amount of bovine serum albumin (BSA) as a stabilizer

(GGT content <1%), but the enzyme was catalytically homogeneous

and was used without further puriﬁcation.

In this study, we designed and synthesized an array of novel

peptidyl phosphorus esters as potent mechanism-based inhibitors

of human and E. coli GGTs. The most active compound 2a with a

L-Abu-Gly analogous peptidyl chain was more than 700 times as

Melting points (mp) were recorded using a Mettler FP62 melt-

ing point apparatus and are corrected. ¹H and ³¹P NMR spectra

were recorded on a JEOL JNM-AL300 (300 MHz for ¹H) and JOEL

JNM-AL400 (400 MHz for ¹H) spectrophotometers. ¹H chemical

shifts (d_H) are reported downﬁeld of tetramethylsilane as an inter-

nal reference (d_H0.0), except for measurements in D₂O for which

sodium 3-(trimethylsilyl)propanesulfonate was used as an internal

standard (d_H0.0). Splitting patterns are abbreviated as follows: s,

singlet; d, doublet; t, triplet, q; quartet; m, multiplet. ³¹P chemical

shifts (d_P) are reported downﬁeld of 85% H₃PO₄as an external stan-

dard (d_P0.0). Infrared spectra were recorded on a HORIBA FT-720

Fourier Transform infrared spectrophotometer. Mass spectra were

obtained with a JEOL JMS-700 spectrometer. Mass spectra of the

enzyme were recorded on a Sciex API-3000 mass spectrometer

(PE Sciex) interfaced with an ion-spray ion source. Elemental anal-

yses were performed on a Yanaco MT-5. Reactions were monitored

by analytical thin layer chromatography (TLC) on silica gel 60 F₂₅₄

plates (Merck, #5715, 0.25 mm). Compounds were puriﬁed by

ﬂash column chromatography on silica gel 60 N (Kanto Kagaku,

spherical and neutral, 40–50 mm, No. 37563-79) or by medium-

pressure reversed-phase column chromatography using a Yamazen

YFLC System with ODS-SM-50B column (Yamazen Co., Osaka, Ja-

pan). The spectral data (¹H NMR and/or IR) for known compounds

7a, c, e–g, 9b, 11, 13 and 18 are described in Supplementary data.

The spectral data supported the structure or are in good agreement

with those reported (references cited in Supplemental data).

active as acivicin toward human GGT, if the stereoisomeric compo-

sition is considered. As a close mimic of the transition state (TS-1)

for glutathione degradation, the peptidyl phosphorus esters served

as a good chemical probe to deﬁne hitherto unidentiﬁed Cys-Gly

binding site of mammalian and bacterial GGTs. So far the substrate

speciﬁcity of GGT for the Cys-Gly binding site have been estimated

by using

the acceptor substrates AA-Gly bind to the same site as the Cys-Gly

moiety of the -glutamyl donor substrates (glutathione). However,

c

-glutamyl acceptor substrates³with the assumption that

c

the approximation of substrate speciﬁcity by acceptor substrates

might be misleading, because the apparent activity is signiﬁcantly

dependent on the extent of deprotonation or the pK_aof the amino

group of the dipeptidyl acceptor substrates (AA-Gly).⁵⁸In this re-

gard, the phosphonate esters with a peptidyl side chain in the pres-

ent study precisely reﬂect the properties of the Cys-Gly binding

site; human GGT has a speciﬁc binding site for Cys-Gly to recognize

strictly its stereochemistry (

the sulfur atom, but no recognition is expected for any group be-

yond the sulfur ( -position). This mode of substrate recognition

L) and the size of its side chain up until

c

seems quite reasonable from the proposed physiological function

of mammalian GGT; the enzyme accepts glutathione (unsubstitut-

ed

to accommodates structurally diverse glutathione S-conjugates

(S-substituted -Cys) for catabolism of xenobiotics. These appar-

ently contradictory demands can be met nicely by strict recogni-

tion around the , b and -position (S atom) of the Cys residue,

while any groups beyond the -position remain unbound, probably

L-Cys) as a primary natural substrate, but at the same time has

L

a

c

by sticking out of the active site. The stereochemical preference of

GGT for the chiral phosphorus (tentatively assigned as S_P^⁄) shown

in this study gives insights into the stereochemical course of the

catalyzed reaction with glutathione. In this context, the peptidyl

phosphorus diesters such as 2a-A is an ideal ligand for X-ray struc-

tural analysis of GGT to deﬁne the putative Cys-Gly binding site.

This line of work is in progress, and the results will be published

in due course.

Finally, the irreversible nature of the present phosphonate

inhibitors is highly advantageous over reversible inhibitors for

needs to eliminate the GGT activities completely until new en-

zymes are de novo synthesized. This is exactly the major reason

why acivicin, a rather weak and nonselective inhibitor of mamma-

lian GGT, has been and is still being used widely, irrespective of its

potent cyto- and neurotoxicity. Therefore the properties of the Cys-

Gly binding site as probed by the present peptidyl phosphorus es-

ters should give a basis for future molecular design for potent and

irreversible inhibitors speciﬁc for human GGT, as well as for bacte-

rial enzymes.

4.2. Synthesis

4.2.1. (S)-2-Hydroxybutanoic acid (7a)

(S)-2-Aminobutanoic acid (6a) (4.00 g, 38.8 mmol) was dis-

solved in 0.5 M H₂SO₄(155 mL, 77.6 mmol). This solution was

cooled to 0 °C and a solution of NaNO₂(16.0 g, 233 mmol) in H₂O

(40 mL) was added slowly with stirring. After stirring at this tem-

perature for 3 h, the reaction was allowed to warm to room tem-

perature and was stirred for further 2 h. The reaction mixture

was extracted 7 times with ethyl acetate (7 ꢁ 150 mL). The com-

bined organic extracts were washed with sat. NaCl and dried over

anhydrous Na₂SO₄. The solvent was removed in vacuo to afford the

hydroxy acid 7a (3.98 g) as an oil. This product was used for the

preparation of compound 9a without puriﬁcation.

4.2.2. (R)-3-Chloro-2-hydroxypropanoic acid (7c)

A

solution of (R)-(+)-3-chloropropane-1,2-diol (10) (5.0 g,

45.5 mmol) in water (5 mL) was placed in a 100 mL test tube. To

the bottom of this test tube, 60% (13 N) nitric acid (6.99 mL,

90.9 mmol) was injected slowly by a pipette, keeping the layers

separated. The mixture was allowed to stand at room temperature

without mixing for 3 days until the two layers became a clear

homogeneous solution. Nitrogen oxide gas (Caution! toxic) was re-

moved from the reaction mixture by argon stream in a fuming

hood, and the solvent was removed in vacuo to afford crude

(R)-3-chloro-2-hydroxypropanoic acid (7c) (6.83 g) as a colorless

4. Experimental procedures

4.1. General methods

All chemicals were obtained commercially and used without

further puriﬁcation unless otherwise sated. 7-(

c-L-Glutamylami-

no)-4-methylcoumarin ( -Glu-AMC) was purchased from Sigma.

c

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1187

oil. The product was used for the preparation of compound 9c

without puriﬁcation.

4.2.8. (R)-Benzyl 2-(3-chloro-2-hydroxypropanamido)acetate

(9c)

Compound 9c was prepared from crude (R)-3-chloro-2-

hydroxypropanoic acid (7c) (45.2 mmol) and glycine derivative 8

(18.4 g, 54.6 mmol) by the same procedure as described for the

synthesis of 9a and was puriﬁed by ﬂash column chromatography

(50% ethyl acetate in hexane), followed by recrystallization from

ethyl acetate and hexane (1:12) to afford 9c as a colorless powder

4.2.3. (S)-2-Hydroxy-3-phenylpropanoic acid (7e)

Compound 7e was prepared from

L-phenylalanine (10.0 g,

60.5 mmol) by the same procedure as described for the synthesis

of 7a and was puriﬁed by recrystallization from a mixture o f

diethyl ether and hexane (1:10) to yield pure 7e as a colorless

power (6.18 g, 61%). Mp 123.8–124.6 °C. ½a _D²⁷

ꢂ

ꢀ21.3 (c 1.04, H₂O)

(6.85 g, 56%). Mp 87.7–88.8 °C; ½a _D²⁷

ꢂ

ꢀ35.4 (c 1.08, EtOH). ¹H NMR

[lit.⁴⁹

½

a ²_D⁰

ꢂ

ꢀ21.3 (c 2.35, H₂O)]. Anal. Calcd for C₉H₁₀O₃: C, 65.05;

(300 MHz, CDCl₃): d_H7.39–7.31 (m, 5H, Ph), 7.24 (br, 1H, NH), 5.20

(s, 2H, PhCH₂O), 4.41 [dd, J = 6.0 and 3.6 Hz, 1H, CH(OH)CONH],

H, 6.07. Found: C, 64.98; H, 6.04.

4.16 (dd, J = 18.3 and 5.7 Hz, 1H, Glya-H_a), 4.09 (dd, J = 18.3 and

4.2.4. (S)-2-Hydroxyhexanoic acid (7f)

Compound 7f was prepared from

5.7 Hz, 1H, Gly -H_b), 3.91 (dd, J = 11.4 and 3.6 Hz, 1H, ClCH_aH_b),

a

L

-norleucine (8.00 g,

3.84 (dd, J = 11.4 and 6.0 Hz, 1H, ClCH_aH_b). IR (KBr) 3384 (br),

1736, 1655, 1541, 1221, 1088, 953, 762, 704 cm^ꢀ1. Anal. Calcd

for C₁₂H₁₄ClNO₄: C, 53.05; H, 5.19; N, 5.16. Found: C, 52.78; H,

5.12; N, 5.33.

61.0 mmol) by the same procedure as described for the synthesis

of 7a and was puriﬁed by recrystallization from a mixture of

diethyl ether and hexane (1:12) to give pure 7f as a highly hygro-

scopic colorless crystalline mass (3.18 g, 39%). ½a _D³⁰

ꢀ16.2 (c 1.02,

ꢂ

1 M NaOH) [lit.⁴⁹

C₆H₁₂O₃: C, 54.53; H, 9.15. Found: C, 54.33; H, 9.12.

½

a ²_D⁰

ꢂ

ꢀ16.3 (c 3.92, 1 M NaOH)]. Anal. Calcd for

4.2.9. (R)-Benzyl 2-(2-hydroxy-3-iodopropanamido)acetate (9d)

The chloride 9c (4.00 g, 14.8 mmol) and sodium iodide (22.0 g,

148 mmol) were dissolved in acetonitrile (200 mL). The solution

was stirred at 80 °C for 3 days. Insoluble salt (NaBr) was removed

by ﬁltration, and the ﬁltrate was evaporated. The residue was dis-

solved in ethyl acetate (300 mL), washed with sat. NaCl

(2 ꢁ 200 mL) and was dried over anhydrous Na₂SO₄. After removal

of the solvent in vacuo, the residue was puriﬁed by ﬂash column

chromatography (60% ethyl acetate in hexane), followed by recrys-

tallization from a mixture of ethyl acetate and hexane (1:20) to af-

4.2.5. (R)-2-Hydroxyhexanoic acid (7g)

Compound 7g was prepared from

D-norleucine (8.00 g,

61.0 mmol) by the same procedure as described for the synthesis

of 7a and was puriﬁed by recrystallization from a mixture of diethyl

ether and hexane (1:12) to yield pure 7g as highly hygroscopic

colorless crystalline mass (4.19 g, 52%). ½a _D³⁰

ꢂ

+16. 8 (c 0.99, 1 M

NaOH) [lit.⁴⁹

½

a ²_D⁰

ꢂ

ꢀ16.3 (c 3.92, 1 M NaOH) for the S enantiomer].

Anal. Calcd for C₆H₁₂O₃: C, 54.53; H, 9.15. Found: C, 54.11; H, 9.11.

HRMS (FAB, glycerol): Calcd for C₆H₁₃O₃(MH⁺) 133.0865; found

133.0865.

ford 9d as a colorless powder (2.41 g, 45%). Mp 84.0–85.1 °C. ½a _D²⁷

ꢂ

ꢀ36.2 (c 1.01, EtOH). ¹H NMR (300 MHz, CDCl₃): d_H7.39–7.33

(m, 5H, Ph), 7.21 (br, 1H, NH), 5.20 (s, 2H, PhCH₂O), 4.21 [ddd,

J = 6.3, 5.7 and 3.9 Hz, 1H, CH(OH)CONH], 4.16 (dd, J = 18.3 and

4.2.6. (S)-Benzyl 2-(2-hydroxybutanamido)acetate (9a)

6.0 Hz, 1H, Glya-H_a), 4.11 (dd, J = 18.3 and 4.5 Hz, 1H, Glya-H_b),

Glycine benzyl ester monotosylate (8) (19.3 g, 57.3 mmol), 1-

hydroxybenzotriazole monohydrate (HOBtꢃH₂O) (7.75 g, 57.3 mmol)

and N-methylmorpholine (NMM) (6.59 g, 65.1 mmol) were

dissolved in CH₂Cl₂(150 mL). This solution was cooled to 0 °C. To

this solution were added the hydroxy acid 7a (3.98 g, 38.3 mmol)

and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride

(EDC) (11.0 g, 57.3 mmol), successively. After stirring at room

temperature for 18 h, the reaction mixture was concentrated in

vacuo to remove CH₂Cl₂. Ethyl acetate (250 mL) was added to the

residue, and the resultant mixture was washed successively with

1 N HCl, sat. NaHCO₃and sat. NaCl, and was dried over anhydrous

Na₂SO₄. After removal of solvent in vacuo, the residue was puriﬁed

by ﬂash column chromatography (40–60% ethyl acetate in hexane)

3.60 (dd, J = 10.2 and 3.9 Hz, 1H, ICH_aH_b), 3.53 (dd, J = 10.2 and

6.3 Hz, 1H, ICH_aH_b), 3.18 (d, J = 5.7 Hz, 1H, OH). IR (KBr) 3352

(br), 1741, 1662, 1535, 1392, 1242, 1201, 1078, 957, 715 cm^ꢀ1

.

Anal. Calcd for C₁₂H₁₄INO₄: C, 39.69; H, 3.89; N, 3.86. Found: C,

39.87; H, 3.97; N 3.84.

4.2.10. (S)-Benzyl 2-(2-hydroxy-3-phenylpropanamido)acetate

(9e)

Compound 9e was prepared from 7e (6.13 g, 36.9 mmol) and

glycine derivative 8 (14.9 g, 44.2 mmol) by the same procedure

as described for the synthesis of 9a and was puriﬁed by recrystal-

lization from ethyl acetate (80 mL) to give 9e as a colorless powder

(8.57 g, 74%). Mp 106.7–107.9 °C. ½a _D²⁷

ꢂ

ꢀ66.7 (c 1.00, EtOH). ¹H

to afford 9a as a pale yellow oil (4.93 g, 51%). ½a _D²⁷

ꢂ

ꢀ26.2 (c 0.80,

NMR (300 MHz, CDCl₃): d_H7.77 (d, J = 9.3 Hz) and 7.44–7.22 (m)

(10H, 2 ꢁ Ph), 7.28 (br t, 1H, NH), 5.19 (s, 2H, PhCH₂O), 4.37 [dd,

J = 8.7 and 3.6 Hz, 1H, CH(OH)CONH], 4.13 (dd, J = 18.6 and

EtOH). ¹H NMR (300 MHz, CDCl₃): d_H7.39–7.30 (m, 5H, Ph), 7.08

(br, 1H, NH), 5.18 (s, 2H, PhCH₂O), 4.18–4.02 [m, 3H, CH(OH)CONH

and Glya-H], 3.08 (br, 1H, OH), 1.88 (dqd, J = 14.4, 7.5 and 4.2 Hz,

5.7 Hz) and 4.06 (dd, J = 18.6 and 5.7 Hz) (2H, Glya-H), 3.27 (dd,

1H, CH₃CH_aH_b),1.69 (dqd, J = 14.4, 7.5 and 7.2 Hz, 1H, CH₃CH_aH_b),

0.98 (t, J = 7.5 Hz, 3H, CH₃). IR (NaCl) 3396 (br), 1747, 1655, 1533,

1194, 1128, 1057, 1032, 984, 741, 698 cm^ꢀ1. Anal. Calcd for C₁₃H_17-

NO₄ꢃ0.2H₂O: C, 61.26; H, 6.88; N, 5.50. Found: C, 61.43; H, 7.07; N,

5.57. HRMS (FAB, glycerol): Calcd for C₁₃H₁₈NO₄(MH⁺) 252.1236;

found 252.1238.

J = 13.8 and 3.6 Hz, 1H, PhCH_aH_b), 2.87 (dd, J = 13.8 and 8.7 Hz,

1H, PhCH_aH_b), 2.61 (br, 1H, OH). IR (KBr) 3309 (br), 1759, 1660,

1541, 1456, 1385, 1362, 1186, 1092, 1036, 964, 746, 704,

588 cm^ꢀ1. Anal. Calcd for C₁₈H₁₉NO₄: C, 68.99; H, 6.11; N, 4.47.

Found: C, 68.75; H, 6.15; N, 4.35.

4.2.11. (S)-Benzyl 2-(2-hydroxyhexanamido)acetate (9f)

4.2.7. Benzyl 2-(2-hydroxyacetamido)acetate (9b)

Compound 9f was prepared from 7f (2.03 g, 15.4 mmol) and

glycine derivative 8 (6.22 g, 18.4 mmol) by the same procedure

as described for the synthesis of 9a and was puriﬁed by ﬂash col-

umn chromatography (30 to 50% ethyl acetate in hexane) to give

Compound 9b was prepared from glycolic acid (7b) (2.28 g,

30.0 mmol) and glycine derivative 8 (4.87 g, 14.5 mmol) by the same

procedure as described for the synthesis of 9a, and was puriﬁed by

ﬂash column chromatography (50% acetone in hexane) to give 9b as

a pale yellow oil (2.36 g, 73%). Anal. Calcd for C₁₁H₁₃NO₄ꢃ0.2H₂O:

C, 58.25; H, 5.95; N, 6.17. Found: C, 58.26; H, 5.98; N, 6.05. HRMS

9f as a colorless oil (4.17 g, 97%). ½a _D³¹

ꢂ

ꢀ30.3 (c 1.03, EtOH). ¹H

NMR (300 MHz, CDCl₃): d_H7.40–7.28 (m, 5H, Ph), 7.17 (br t, 1H,

NH), 5.17 (s, 2H, PhCH₂O), 4.14 [m, 1H, CH(OH)CONH], 4.13 (dd,

(FAB, glycerol): Calcd for

224.0924.

C

₁₁H₁₄NO₄(MH⁺) 224.0923; found

J = 18.0 and 5.7 Hz, 1H, Glya-H_a), 4.03 (dd, J = 18.0 and 5.7 Hz,

1H, Gly -H_b), 3.43 (d, J = 5.7 Hz, 1H, OH), 1.89–1.77 [m, 1H,

a

1188

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

CH₃(CH₂)₂CH_aH_b], 1.70–1.57 [m, 1H, CH₃(CH₂)₂CH_aH_b], 1.47–1.26

(m, 4H, CH₃CH₂CH₂), 0.90 (t, J = 7.1 Hz, 3H, CH₃). IR (NaCl) 3398

(br), 2956, 1747, 1657, 1533, 1456, 1387, 1358, 1194, 1132, 962,

908 cm^ꢀ1. HRMS (FAB, glycerol): Calcd for C₁₅H₂₂NO₄(MH⁺)

280.1549; found 280.1550. Anal. Calcd for C₁₅H₂₁NO₄: C, 64.50;

H, 7.58; N 5.01. Found: C, 63.94; H, 7.58; N 4.93.

idue was added ethyl acetate (40 mL), and the insoluble salt was

removed by ﬁltration. The crude product contained the desired

product 14a (49%, d_P28.8, 28.6, 28.4 and 28.2) along with the

corresponding diphenyl phosphonate (13%, d_P24.5) and dialkyl

phosphonate (38%, d_P39.5) as byproducts (³¹P NMR). The ﬁltrate

was concentrated, and the crude product was puriﬁed by ﬂash col-

umn chromatography (60% ethyl acetate in toluene) to afford the

phosphonate diester 14a as a mixture of four diastereoisomers

4.2.12. (R)-Benzyl 2-(2-hydroxyhexanamido)acetate (9g)

Compound 9g was prepared from 7g (2.20 g, 16.7 mmol) and

glycine derivative 8 (6.76 g, 20.0 mmol) by the same procedure

as described for the synthesis of 9a and was puriﬁed by ﬂash col-

umn chromatography (30% ethyl acetate in hexane) to give 9g as

(ca. 2:2:1:1) with respect to the chiral

a-carbon and phosphorus

(1.99 g, 53%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H

7.37–7.10 (m, 20H, 4 ꢁ Ph), 7.20–6.78 (4 ꢁ br t, 1H, amide NH),

5.79 (br d, J = ca. 7 Hz) and 5.59 (d, J = 7.6 Hz) (1H, carbamate

NH), 5.23–5.04 (m, 6H, 3 ꢁ PhCH₂O), 4.85–4.75 [m, 1H, CH(OP)-

CONH], 4.55–4.37 [m, 1H, NH(CO)CH(CH₂)₂P], 4.15–3.95 and

a clear oil (4.50 g, 97%). ½a _D²⁷

ꢂ

+30.7 (c 0.94, EtOH). ¹H NMR

(300 MHz, CDCl₃): d_H7.40–7.28 (m, 5H, Ph), 7.00 (br t, 1H, NH),

5.19 (s, 2H, PhCH₂O), 4.17 [m, 1H, CH(OH)CONH], 4.15 (dd,

3.80–3.65 (m, 2H, Glya-H), 2.40–1.70 (m, 6H, CH₂CH₂P and CH_3-

J = 14.4 and 5.4 Hz, 1H, Gly

a

-H_a), 4.08 (dd, J = 14.4 and 5.4 Hz,

CH₂), 0.93 (2 ꢁ t, J = 7.6 Hz) and 0.82 (2 ꢁ t, J = 7.6 Hz) (3H, CH₃).

³¹P NMR (121 MHz, CDCl₃): d_P28.76, 28.57, 28.40, 28.14. IR (NaCl)

3292 (br), 1743, 1684, 1529, 1313, 1201, 933, 738 cm^ꢀ1. Anal.

Calcd for C₃₈H₄₁N₂O₁₀P: C, 63.68; H, 5.77; N, 3.91. Found: C,

63.54; H, 5.78; N 3.74.

1H, Gly -H_b), 2.68 (d, J = 5.1 Hz, 1H, OH), 1.91–1.79 [m, 1H,

a

CH₃(CH₂)₂CH_aH_b], 1.71–1.58 [m, 1H, CH₃(CH₂)₂CH_aH_b], 1.47–1.26

(m, 4H, CH₃CH₂CH₂), 0.91 (t, J = 6.9 Hz, 3H, CH₃). IR (NaCl) 3404

(br), 2956, 1749, 1652, 1533, 1456, 1387, 1358, 1192, 1132, 960,

908 cm^ꢀ1. HRMS (FAB, glycerol): Calcd for C₁₅H₂₂NO₄(MH⁺)

280.1549; found 280.1548. Anal. Calcd for C₁₅H₂₁NO₄ꢃ0.2H₂O: C,

63.68; H, 7.62; N, 4.95. Found: C, 63.77; H, 7.67; N 4.91.

4.2.16. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{[N-

(benzyloxycarbonylmethyl)carbamoyl]methyl(phenyl)-(R_P,S_P)-

phosphono}butanoate (14b)

4.2.13. (S)-Methyl 2-hydroxyhexanoate (11)

Compound 14b was synthesized from the phosphonic dichlo-

ride 13 (2.18 g, 4.91 mmol) and the alcohol 9b (1.10 g, 4.91 mmol)

by the same procedure as described for the synthesis of 14a and

was puriﬁed by ﬂash column chromatography (45% acetone in hex-

ane) to yield pure 14b as a mixture of two diastereoisomers (1.31 g,

39%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H7.37–7.13 (m,

20H, 4 ꢁ Ph), 7.04 (br, 1H, amide NH), 5.66 (br, 1H, carbamate NH),

5.18, 5.15 and 5.08 (3 ꢁ s, 2H, PhCH₂O), 4.66–4.46 [m, 1H,

Thionyl chloride (2.00 g, 16.8 mmol) was added dropwise to

methanol (20 mL) at 0 °C, and the mixture was allowed to warm

to room temperature. After stirring at room temperature for

30 min, the hydroxy acid 7f (1.39 g, 10.5 mmol) was added to the

solution. After stirring at room temperature for 1.5 h, the mixture

was concentrated in vacuo. The resulting residue was puriﬁed by

ﬂash column chromatography (40% ethyl acetate in hexane) to af-

ford the methyl ester 11 as a colorless oil (1.30 g, 84%). ½a _D²⁷

ꢂ

ꢀ5.50

NH(CO)CH(CH₂)₂P] and 4.46 [dd, J_HP= 15.2 Hz, J = 8.0 Hz, 3H,

3

(c 0.92, EtOH). Anal. Calcd for C₇H₁₄O₃: C, 57.51; H, 9.65. Found: C,

CH₂(OP)CONH], 4.12–3.92 (m, 2H, Glya-H), 2.43–2.2 (m) and

56.78; H 9.47.

2.15–1.80 (m) (4H, CH₂CH₂P). ³¹P NMR (121 MHz, CDCl₃): d_P

29.12, 29.04. IR (NaCl) 3298 (br), 1745, 1685, 1535, 1255, 1200,

1065, 937, 741, 696 cm^ꢀ1. Anal. Calcd for C₃₆H₃₇N₂O₁₀Pꢃ0.5H₂O:

C, 61.98; H, 5.49; N, 4.02. Found: C, 62.11; H, 5.44; N 4.02.

4.2.14. (R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-

(dichlorophosphoryl)butanoate (13)

Oxalyl chloride (1.41 g, 11.1 mmol) was added to a solution of

racemic benzyl 2-(benzyloxycarbonylamino)-4-phosphonobutano-

ate (12) (2.01 g, 4.93 mmol) in dry CH₂Cl₂(10 mL) in the presence

of a catalytic amount (1 drop) of DMF. Evolution of gas (CO and

CO₂) was observed immediately. The reaction was complete in

2 h at 30 °C. Volatiles were removed by vacuum evaporation, fol-

lowed by removal of trace of HCl by ﬂushing Ar ﬂow to give the

phosphonic dichloride 13 as an yellow oil. Compound 13 was used

fresh for the next reaction without puriﬁcation.

4.2.17. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(R)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]-2-

chloroethyl(phenyl)-(R_P,S_P)-phosphono}butanoate (14c)

Compound 14c was synthesized from the phosphonic dichlo-

ride 13 (3.26 g, 7.37 mmol) and the alcohol 9c (1.33 g, 4.91 mmol)

by the same procedure as described for the synthesis of 14a and

was puriﬁed by ﬂash column chromatography (25% ethyl acetate

in toluene) to afford pure 14c as a mixture of four diastereoisomers

(1.45 g, 40%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H7.37–

7.16 (m, 20H, 4 ꢁ Ph), 7.02 and 6.84 (2 ꢁ br t, 1H, amide NH), 5.81

(br d, J = ca. 6 Hz) and 5.56 (d, J = 7.6 Hz) (1H, carbamate NH), 5.21–

5.00 [m, 7H, 3 ꢁ PhCH₂O and CH(OP)CONH], 4.68–4.34 [m, 1H,

4.2.15. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(S)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]propyl(phenyl)-

(R_P,S_P)-phosphoryl}butanoate (14a)

3

Phenol (748 mg, 7.94 mmol) in dry CH₂Cl₂(5 mL) was added to

a solution of the phosphonic dichloride 13 (3.52 g, 7.94 mmol) in

dry CH₂Cl₂(20 mL), and the solution was cooled to ꢀ65 °C. Dry

Et₃N (804 mg, 7.94 mmol) was added dropwise to the cold solution

over 15 min, and the mixture was stirred at ꢀ65 °C for 30 min. The

mixture was then allowed to warm to room temperature. After

stirring at room temperature for 2 h, the consumption of 13 (d_P

49.2) and the formation of monochloro monophenyl ester (d_P

39.6) were conﬁrmed with ³¹P NMR. A solution of the alcohol 9a

(1.33 g, 5.30 mmol) in dry CH₂Cl₂(5 mL) was then added to the

mixture at 0 °C, followed by the addition of Et₃N (536 mg,

5.30 mmol) to initiate the reaction. The mixture was stirred at

0 °C for 30 min and was allowed to warm to room temperature.

After stirring at room temperature for 4 h, the reaction mixture

was concentrated in vacuo to remove CH₂Cl₂. To the resulting res-

NH(CO)CH(CH₂)₂P], 4.16 (dd, J_HP= 18.3 Hz and J = 6.3 Hz) and

3

3.96 (dd, J_HP= 18.3 Hz and J = 4.8 Hz) (2H, Gly

a-H), 4.10–3.73

and 3.56–3.48 (2 ꢁ m, 2H, and ClCH₂), 2.50–2.25 and 2.23–1.96

(2 ꢁ m, 4H, CH₂CH₂P). ³¹P NMR (121 MHz, CDCl₃): d_P29.86,

29.75, 29.51. IR (NaCl) 3292 (br), 1747, 1684, 1533, 1255, 1200,

1061, 939, 741, 696 cm^ꢀ1. Anal. Calcd for C₃₇H₃₈ClN₂O₁₀P: C,

60.29; H, 5.20; N, 3.80. Found: C, 60.22; H, 5.33; N, 3.68.

4.2.18. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(R)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]-2-

iodoethyl(phenyl)-(R_P,S_P)-phosphono}butanoate (14d)

Compound 14d was prepared from the phosphonic dichloride

13 (2.38 g, 5.37 mmol) and the alcohol 9d (1.30 g, 3.58 mmol) by

the same procedure as described for the synthesis of 14a and

was puriﬁed by ﬂash column chromatography (25% ethyl acetate

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1189

in toluene) to afford pure 14d as a mixture of the diastereoisomers

(538 mg, 18%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H

7.37–6.95 (m, 21H, 4 ꢁ Ph and amide NH), 5.81 (br, 1H, carbamate

NH), 5.22–5.03 (m, 6H, 3 ꢁ PhCH₂O), 4.97–4.88 [m, 1H,

CH(OP)CONH], 4.60–4.44 [m, 1H, NH(CO)CH(CH₂)₂P], 4.14 (dd,

d, J = ca. 7 Hz) and 5.58 (d, J = 7.5 Hz) (1H, carbamate NH), 5.21–

5.04 (m, 6H, 3 ꢁ PhCH₂O), 4.98–4.85 [m, 1H, CH(OP)CONH], 4.59–

3

4.44 [m, 1H, NH(CO)CH(CH₂)₂P], 4.10 (dd, J_HP= 18.0 Hz and

3

J = 6.0 Hz), 3.97 (dd, J_HP= 18.0 Hz and J = 5.4 Hz) and 4.15–3.93

(m) (2H, Glya-H), 2.42–2.23 (m), 2.20–1.94 (m) and 1.93–1.63

3

³J_HP= 18.3 Hz and J = 6.0 Hz) and 3.97 (dd, J_HP= 18.3 Hz and

(m) [6H, CH₂CH₂P and CH₃(CH₂)₂CH₂], 1.31 and 1.17 (2 ꢁ m, 4H,

CH₃CH₂CH₂), 0.86 and 0.78 (2 ꢁ m, 3H, CH₃). ³¹P NMR (121 MHz,

CDCl₃): d_P28.80, 28.59, 28.37, 28.11. IR (NaCl) 3290 (br), 2956,

1745, 1682, 1589, 1529, 1491, 1454, 1254, 1201, 1047, 1005,

935 cm^ꢀ1. Anal. Calcd for C₄₀H₄₅N₂O₁₀Pꢃ0.5H₂O: C, 63.74; H, 6.15;

N, 3.72. Found: C, 63.62; H, 6.16; N, 3.59. HRMS (FAB, glycerol):

Calcd for C₄₀H₄₆N₂O₁₀P (MH⁺) 745.2890; found 745.2894.

J = 4.5 Hz) (2H, Gly

a

-H), 3.64–3.50 and 3.21–3.19 (2 ꢁ m, 2H,

ICH₂), 2.47–1.96 (m, 4H, CH₂CH₂P). ³¹P NMR (121 MHz, CDCl₃):

d_P29.59, 29.47, 29.32. IR (NaCl) 3288 (br), 1743, 1682, 1489,

1456, 1255, 1200, 1053, 939, 754, 696 cm^ꢀ1. Anal. Calcd for C₃₇H_38-

IN₂O₁₀P: C, 53.63; H, 4.62; N, 3.38. Found: C, 54.04; H, 4.72; N, 3.33.

HRMS (FAB, glycerol): Calcd for C₃₇H₃₉IN₂O₁₀P (MH⁺) 829.1387;

found 829.1375.

4.2.22. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(S)-1-

methoxycarbonylpentyl(phenyl)-(R_P,S_P)-phosphono}butanoate

(15)

4.2.19. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(S)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]-2-

phenylethyl(phenyl)-(R_P,S_P)-phosphono}butanoate (14e)

Compound 14e was prepared from the phosphonic dichloride

13 (2.18 g, 4.91 mmol) and the alcohol 9e (1.54 g, 4.91 mmol) by

the same procedure as described for the synthesis of 14a and

was puriﬁed by ﬂash column chromatography (30% ethyl acetate

in toluene) to afford pure 14e as a mixture of four diastereoisomers

(804 mg, 21%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H

7.39–6.97 (m, 25H, 5 ꢁ Ph), 6.57 (br t, 1H, amide NH), 5.57 (br d,

1H, carbamate NH), 5.28–5.04 [m, 7H, 3 ꢁ PhCH₂O and CH(OP)-

CONH], 4.41 and 4.28 [2 ꢁ m, 1H, NH(CO)CH(CH₂)₂P], 4.02–3.83

Compound 15 was prepared from the phosphonic dichloride 13

(2.24 g, 5.07 mmol) and the alcohol 11 (494 mg, 3.38 mmol) by the

same procedure as described for the synthesis of 14a and was puri-

ﬁed by ﬂash column chromatography (40% ethyl acetate in hexane)

to afford pure 15 as a mixture of four diastereoisomers (763 mg,

37%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H7.38–7.12

(m, 15H, 3 ꢁ Ph), 5.61 (br) and 5.50 (br d, J = ca. 7 Hz) (1H, carba-

mate NH), 5.23–5.07 (m, 6H, 3 ꢁ PhCH₂O), 4.96–4.82 [m, 1H,

CH(OP)CO₂Me], 4.57–4.38 [m, 1H, NH(CO)CH(CH₂)₂P], 3.68 and

3.60 (2 ꢁ s, 3H, CH₃O), 2.44–2.25 (m) and 2.25–1.90 (m) (4H, CH_2-

CH₂P), 1.90–1.75 (m) and 1.74–1.60 (m) [2H, CH₃(CH₂)₂CH₂], 1.32

and 1.14 (2 ꢁ m, 4H, CH₃CH₂CH₂), 0.88 (t, J = 6.3 Hz) and 0.77 (t,

J = 6.6 Hz) (3H, CH₃). ³¹P NMR (121 MHz, CDCl₃): d_P29.61, 29.51,

28.05, 27.94. IR (NaCl) 3265 (br), 2956, 1753, 1724, 1591, 1493,

1454, 1344, 1255, 1205, 1053, 1026, 931, 756, 696 cm^ꢀ1. Anal.

Calcd for C₃₂H₃₈NO₉P: C, 62.84; H, 6.26; N, 2.29. Found: C, 62.71;

H, 6.26; N, 2.26.

(m) and 3.62 (dd, J = 18.3 and 4.8 Hz) (2H, Glya-H), 3.33 (d,

J = 12.6 Hz) and 3.25–2.89 (m) (2H, PhCH₂), 2.29–1.76 (m), 1.75–

1.50 (m), 1.49–1.27 (m) (4H, CH₂CH₂P). ³¹P NMR (121 MHz, CDCl₃):

d_P29.98, 29.85, 28.99, 28.78. IR (NaCl) 3294 (br), 1745, 1687, 1529,

1456, 1387, 1346, 1254, 1200, 935, 752, 698, 580 cm^ꢀ1. Anal. Calcd

for C₄₃H₄₃N₂O₁₀P: C, 66.32; H, 5.57; N, 3.60. Found: C, 66.21; H,

5.55; N 3.62.

4.2.20. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(S)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]pentyl(phenyl)-

(R_P,S_P)-phosphono}butanoate (14f)

4.2.23. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(S)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]propyl(methyl)-

(R_P,S_P)-phosphono}butanoate (16)

Compound 14f was prepared from the phosphonic dichloride

13 (3.26 g, 7.37 mmol) and the alcohol 9f (1.37 g, 4.91 mmol) by

the same procedure as described for the synthesis of 14a and

was puriﬁed by ﬂash column chromatography (35% ethyl acetate

in toluene) to afford pure 14f as a mixture of four diastereoisomers

(1.96 g, 54%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H7.36–

7.15 (m, 20H, 4 ꢁ Ph), 7.05–6.76 (4 ꢁ br t, 1H, amide NH), 5.83 (br

d, J = 5.1 Hz) and 5.60 (d, J = 7.8 Hz) (1H, carbamate NH), 5.23–5.04

(m, 6H, 3 ꢁ PhCH₂O), 4.97–4.87 [m, 1H, CH(OP)CONH], 4.57–4.43

[m, 1H, NH(CO)CH(CH₂)₂P], 4.10 (dd, ³J_HP= 18.3 Hz and

Compound 16 was prepared from the phosphonic dichloride 13

(3.18 g, 7.17 mmol), dry methanol (230 mg, 7.17 mmol) and the

alcohol 9a (1.20 g, 4.78 mmol) by the same procedure as described

for the synthesis of 14a and was puriﬁed by ﬂash column chroma-

tography (40% ethyl acetate in diethyl ether) to afford pure 16 as a

mixture of four diastereoisomers (624 mg, 20%, a pale yellow oil).

¹H NMR (300 MHz, CDCl₃): d_H7.40–7.19 (m, 15H, 3 ꢁ Ph), 7.11

(br, 1H, amide NH), 5.79 (br t, J = ca. 8 Hz, 1H, carbamate NH),

5.20–5.06 (m, 6H, 3 ꢁ PhCH₂O), 4.87–4.75 [m, 1H, CH(OP)CONH],

4.55–4.35 [m, 1H, NH(CO)CH(CH₂)₂P], 4.15–3.94 and 3.82–3.63

3

J = 6.0 Hz), 3.96 (dd, J_HP= 18.3 Hz and J = 4.8 Hz) and 4.15–3.77

(2 ꢁ m, 2H, Gly

a

-H), 3.71 and 3.59 (2 ꢁ d, J_HP= 11.1 Hz, 3H,

(m) (2H, Gly

a

-H), 2.42–2.25 (m), 2.25–1.94 (m) and 1.93–1.63

CH₃O), 2.33–2.10 (m) and 2.10–1.71 (m) (6H, CH₂CH₂P and CH_3-

CH₂), 0.95 (t, J = 7.5 Hz, 3H, CH₃). ³¹P NMR (121 MHz, CDCl₃): d_P

33.24, 33.05, 32.79. IR (NaCl) 3275 (br), 2952, 1743, 1724, 1682,

1531, 1456, 1244, 1215, 1196, 1043, 995, 739, 698 cm^ꢀ1. HRMS

(FAB, glycerol): Calcd for C₃₃H₄₀O₁₀N₂P (MH⁺) 655.2421; found

655.2422.

(m) [6H, CH₂CH₂P and CH₃(CH₂)₂CH₂], 1.31 and 1.17 (2 ꢁ m, 4H,

CH₃CH₂CH₂), 0.86 (m) and 0.78 (t, J = 7.1 Hz) (3H, CH₃). ³¹P NMR

(121 MHz, CDCl₃): d_P28.80, 28.59, 28.37, 28.11. IR (NaCl) 3290

(br), 2956, 1747, 1684, 1591, 1533, 1489, 1456, 1248, 1201,

1047, 1005, 935 cm^ꢀ1. Anal. Calcd for C₄₀H₄₅N₂O₁₀P: C, 64.51; H,

6.09; N, 3.76. Found: C, 64.57; H, 6.08; N 3.70.

4.2.24. (2R,S)-2-Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]

propyl(phenyl)-(S_P^⁄)-phosphono}-butanoic acid (2a-A).

4.2.21. (2R,S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{(R)-1-

[N-(benzyloxycarbonylmethyl)carbamoyl]pentyl(phenyl)-

(R_P,S_P)-phosphono}butanoate (14g)

(2R,S)-2-Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]propyl

(phenyl)-(R_P^⁄)-phosphono}-butanoic acid (2a–B)

Compound 14g was prepared from the phosphonic dichloride

13 (3.67 g, 8.28 mmol) and the alcohol 9g (2.08 g, 7.45 mmol) by

the same procedure as described for the synthesis of 14a and

was puriﬁed by ﬂash column chromatography (40% ethyl acetate

in toluene) to afford pure 14g as a mixture of four diastereoisomers

(1.57 g, 28%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H7.37–

7.12 (m, 20H, 4 ꢁ Ph), 7.05–6.76 (4 ꢁ br t, 1H, amide NH), 5.81 (br

The phosphonate diester 14a (227 mg, 0.317 mmol, a mixture of

four diastereoisomers) was dissolved in a mixture of methanol

(20 mL) and H₂O (4 mL)containingacetic acid(95.0 mg, 1.58 mmol).

Hydrogen gas was passed through the solution in the presence of 5%

Pd/C for 2 h at room temperature. The consumption of 14a was mon-

itored by TLC (ethyl acetate/hexane, 2:1, phosphomolybdate,

R_f= 0.5), and the formation of product was conﬁrmed by TLC

1190

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

⁄

(butan-1-ol/acetic acid/H₂O, 5:2:2, ninhydrin, R_f= 0.55). Pd/C was

removedbyﬁltrationthroughCelite, and theﬁltratewasevaporated.

The residue was puriﬁed by medium-pressure reversed-phase

column chromatography on ODS-S-50B (Yamazen Co., Osaka, Ja-

pan). The column was eluted (6 mL/min) with a linear gradient of

40 to 60% methanol in H₂O. The UV-positive (254 nm) fractions con-

taining the product (TLC, butan-1-ol/acetic acid/H₂O, 5:2:2, ninhy-

drin) were collected and lyophilized to afford the S_P^⁄isomer (2a-A)

(41.3 mg, 33%, slow-eluted isomer, eluted at 50% methanol) and

slow-eluted isomer, eluted at 50% methanol) and the R_Pisomer

(2c-B) (28.9 mg, 13%, fast-eluted isomer, eluted at 35% methanol)

each as a colorless solid.

Compound 2c-A (a mixture of two diastereoisomers). ¹H NMR

(400 MHz, D₂O): d_H7.47 (t, J = 8.0 Hz), 7.34 (t, J = 8.0 Hz), 7.29 (d,

J = 8.0 Hz), 6.98 (t, J = 7.6 Hz) and 6.91 (d, J = 8.4 Hz) (5H, Ph),

5.29 and 4.96 [2 ꢁ m, 1H, CH(OP)CONH], 4.10–3.89 (m, 4H, ClCH₂

and Glya-H), 3.84–3.78 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 2.53–2.11

(m, 4H, CH₂CH₂P). ³¹P NMR (121 MHz, D₂O): d_P31.42. IR (KBr)

3427 (br), 2944, 1674, 1491, 1228, 1203, 1070, 947, 768,

690 cm^ꢀ1. Anal. Calcd for C₁₅H₂₀ClN₂O₈Pꢃ0.43H₂O: C, 41.85; H,

4.88; N, 6.51. Found: C, 42.12; H, 4.97; N, 6.21. HRMS (FAB, glyc-

⁄

the R_Pisomer (2a-B) (17.0 mg, 13%, fast-eluted isomer, eluted at

40% methanol) each as a colorless solid.

Compound 2a-A (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.47 (t, J = 8.0 and 7.8 Hz, 2H), 7.33 (t,

J = 7.8 Hz, 1H) and 7.26 (d, J = 7.8 Hz, 2H) (Ph), 4.98–4.86 [m, 1H,

erol): Calcd for

C

₁₅H₂₁³⁵ClN₂O₈P (MH⁺) 423.0724; found

423.0722. Calcd for C₁₅H₂₁³⁷ClN₂O₈P (MH⁺) 425.0695; found

425.0698.

CH(OP)CONH], 4.00 (d, J = 18.0 Hz, 1H, Glya-H_a), 3.93 (d,

J = 18.0 Hz, 1H, Gly -H_b), 3.92–3.85 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P],

a

Compound 2c-B (a mixture of two diastereoisomers). ¹H NMR

(400 MHz, D₂O): d_H7.45 (t, J = 8.4 Hz), 7.32 (t, J = 8.4 Hz), 7.24 (d,

J = 8.4 Hz), 6.98 (t, J = 7.6 Hz) and 6.91 (d, J = 8.4 Hz) (5H, Ph),

5.24–5.18 (m) and 4.96 (m) [1H, CH(OP)CONH], 4.04–3.79 [m,

2.48–2.19 (m, 4H, CH₂CH₂P), 1.90–1.76 (m, 2H, CH₃CH₂), 0.85 (t,

J = 7.5 Hz, 3H, CH₃). ³¹P NMR (121 MHz, D₂O): d_P30.91. IR (KBr)

3435 (br), 2976, 1668, 1489, 1205, 1051, 1007, 943, 769,

690 cm^ꢀ1. Anal. Calcd for C₁₆H₂₃N₂O₈Pꢃ1.25H₂O: C, 45.23; H,

6.05; N, 6.59. Found: C, 45.28; H, 5.90; N, 6.74. HRMS (FAB, glyc-

erol): Calcd for C₁₆H₂₄N₂O₈P (MH⁺) 403.1270; found 403.1263.

Compound 2a-B (a mixture of two diastereoisomers). ¹H NMR

(400 MHz, D₂O): d_H7.44 (t, J = 8.1 Hz), 7.32 (t, J = 8.1 Hz), 7.21 (d,

J = 8.4 Hz), 6.98 (t, J = 7.6 Hz) and 6.91 (d, J = 8.0 Hz, 1H) (5H, Ph),

5H, ClCH₂, Glya-H and NH₂(CO₂H)CH(CH₂)₂P], 2.54–2.09 (m, 4H,

CH₂CH₂P). ³¹P NMR (121 MHz, D₂O): d_P31.31; IR (KBr) 3427 (br),

2945, 1674, 1491, 1228, 1203, 1070, 947, 768, 690 cm^ꢀ1. Anal.

Calcd for C₁₅H₂₀ClN₂O₈PꢃH₂O: C, 40.87; H, 5.03; N, 6.36. Found: C,

40.73; H, 4.89; N, 6.35. HRMS (FAB, glycerol): Calcd for

C

₁₅H₂₁³⁵ClN₂O₈P (MH⁺) 423.0724; found 423.0732. Calcd for

₁₅H₂₁³⁷ClN₂O₈P (MH⁺) 425.0695; found 425.0686.

3

4.92–4.85 (m) and 4.61 (ddd, J_HP= 9.4 Hz, J = 4.9 and 4.9 Hz)

[1H, CH(OP)CONH], 3.99 (s, 1H, Glya-H_a), 3.78 (s, 1H, Glya-H_b),

3.88 (t, J = 5.4 Hz) and 3.82 (dd, J = 11.6 and 5.4 Hz) [1H, NH₂(CO_2-

H)CH(CH₂)₂P], 2.47–2.31 (m), 2.35–2.20 (m) and 2.20–2.05 (m)

(4H, CH₂CH₂P), 1.95–1.76 (m, 2H, CH₃CH₂), 0.96 and 0.93 (2 ꢁ t,

J = 7.4 Hz, 3H, CH₃). ³¹P NMR (121 MHz, D₂O): d_P30.71. IR (KBr)

3429 (br), 2978, 1668, 1489, 1202, 1051, 1009, 941, 769,

690 cm^ꢀ1. HRMS (FAB, glycerol): Calcd for C₁₆H₂₄N₂O₈P (MH⁺)

403.1270; found 403.1268.

4.2.27. (2R,S)-2-Amino-4-{(R)-1-[N-

(carboxymethyl)carbamoyl]-2-iodoethyl(phenyl)-(S_P^⁄)-

phosphono}butanoic acid (2d-A)

To

a solution of the phosphonate diester 14d (257 mg,

0.310 mmol) in dry nitromethane (4 mL) was added anisole

(403 mg, 3.72 mmol) and aluminum chloride (248 mg, 1.86 mmol)

at room temperature. After stirring at room temperature for 1 h,

the reaction mixture was diluted with 1 N HCl (20 mL) and washed

with diethyl ether (3 ꢁ 30 mL) to remove anisole and nitrometh-

ane. The aqueous layer was concentrated in vacuo to ca. 10 mL

(do not dry up) and was directly applied to the medium-pressure

reversed-phase column chromatography on ODS-S-50B. The col-

umn was eluted (6 mL/min) with a linear gradient of 30–50%

methanol in H₂O. The UV-positive (254 nm) fractions containing

the product (TLC, butan-1-ol/acetic acid/H₂O, 5_⁄:2:2, ninhydrin)

were collected and lyophilized to afford the S_Pisomer (2d-A)

(35.0 mg, 22%, slow-eluted isomer, eluted at 50% methanol) as col-

orless solid and the R_P^⁄isomer (2s-B) (trace, detected by TLC, fast-

eluted isomer, eluted at 40% methanol).

4.2.25. (2R,S)-2-Amino-4-{[N-(carboxymethyl)carbamoyl]

methyl(phenyl)-(R_P^⁄,S_P^⁄)-phosphono}butanoic acid (2b)

Compound 2b was prepared from 14b (500 mg, 0.727 mmol, a

mixture of two diastereoisomers) by the same procedure as de-

scribed for the synthesis of 2a and was puriﬁed by medium-pres-

sure reversed-phase column chromatography (ODS-S-50B) with a

linear gradient of 5–35% methanol in H₂O. The UV-positive

(254 nm) fractions containing the product (TLC, butan-1-ol/acetic

acid/H₂O, 5:2:2, ninhydrin) eluted at 30% methanol were collected

and lyophilized to give compound 2b as a colorless solid (79.2 mg,

29%) (a mixture of four stereoisomers). ¹H NMR (300 MHz, D₂O): d_H

7.47 (dd, J = 7.8 and 7.4 Hz, 2H), 7.33 (t, J = 7.4 Hz, 1H) and 7.26 (d,

J = 7.8 Hz, 2H) (Ph), 4.80–4.68 [m, 2H, CH₂(OP)CONH], 3.97 (s, 2H,

Compound 2c-A (a mixture of two diastereoisomers). ¹H NMR

(400 MHz, D₂O): d_H7.48 (t, J = 7.8 Hz), 7.34 (t, J = 7.8 Hz), 7.31 (d,

J = 7.8 Hz), 6.98 (t, J = 7.6 Hz) and 6.91 (d, J = 7.6 Hz, 1H) (5H, Ph),

Glya-H), 3.91 [br t, 1H, NH₂(CO₂H)CH(CH₂)₂P], 2.48–2.22 (m, 4H,

CH₂CH₂P). ³¹P NMR (121 MHz, D₂O): d_P31.73, 31.70. IR (KBr)

3419 (br), 2943, 1670, 1489, 1228, 1201, 1070, 1024, 941, 769,

690 cm^ꢀ1. Anal. Calcd for C₁₄H₁₉N₂O₈Pꢃ1.5H₂O: C, 41.90; H, 5.53;

N, 6.98. Found: C, 41.97; H, 5.50; N 7.10. HRMS (FAB, glycerol):

Calcd for C₁₄H₂₀N₂O₈P (MH⁺) 375.0957; found 375.0959.

5.07 (ddd, J_HP= 8.0 Hz, J = 4.0 and 4.0 Hz) and 4.73–4.67 (m) [1H,

3

CH(OP)CONH], 4.12–3.91 [m, 3H, NH(CO)CH(CH₂)₂P and Glya-H],

3.66–3.38 (m, 2H, ICH₂), 2.52–2.11 (m, 4H, CH₂CH₂P). ³¹P NMR

(121 MHz, D₂O): d_P31.19. IR (KBr) 3435 (br), 2945, 1668, 1633,

1489, 1228, 1201, 1053, 497, 768, 690 cm^ꢀ1. Anal. Calcd for C₁₅H_20-

IN₂O₈PꢃH₂O: C, 33.85; H, 4.17; N, 5.26. Found: C, 33.89; H, 3.97; N

4.2.26. (2R,S)-2-Amino-4-{(R)-1-[N-(carboxymethyl)carbam

oyl]-2-chloroethyl(phenyl)-(S_P^⁄)-phosphono}butanoic acid (2c-

A). (2R,S)-2-Amino-4-{(R)-1-[N-(carboxymethyl)carbamoyl]-2-

chloroethyl(phenyl)-(R_P^⁄)-phosphono}butanoic acid (2c-B)

Compound 2c was prepared from 14c (379 mg, 0.515 mmol) by

the same procedure as described for the synthesis of 2a and was

puriﬁed by medium-pressure reversed-phase column chromatog-

raphy (ODS-S-50B) with a linear gradient of 30–50% methanol in

H₂O. The UV-positive (254 nm) fractions containing the product

(TLC, butan-1-ol/acetic acid/H₂O, 5:2:2, ninhydrin) were collected

5.25. HRMS (FAB, glycerol): Calcd for

515.0080; found 515.0068.

C

₁₅H₂₁IN₂O₈P (MH⁺)

4.2.28. (2R,S)-2-Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]-

2-phenylethyl(phenyl)-(S_P^⁄)-phosphono}butanoic acid (2e-A).

(2R,S)-2-Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]-2-

phenylethyl(phenyl)-(R_P^⁄)-phosphono}butanoic acid (2e-B)

Compound 2e was prepared from 14e (253 mg, 0.325 mmol))

by the same procedure as described for the synthesis of 2a and

was puriﬁed by medium-pressure reversed-phase column chroma-

tography (ODS-S-50B) with a linear gradient of 40–60% methanol

⁄

and lyophilized to afford the S_Pisomer (2c-A) (72.9 mg, 34%,

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1191

in H₂O. The UV-positive (254 nm) fractions containing the product

(TLC, butan-1-ol/acetic acid/H₂O, 5:2:2, ninhydrin) were collected

N, 6.12. Found: C, 47.35; H, 6.35; N, 6.17. HRMS (FAB, glycerol):

Calcd for C₁₈H₂₈N₂O₈P (MH⁺) 431.1583; found 431.1583.

⁄

and lyophilized to afford the S_Pisomer (2e-A) (59.3 mg, 39%,

⁄

slow-eluted isomer, eluted at 60% methanol) and the R_Pisomer

4.2.30. (2R,S)-2-Amino-4-{(R)-1-[N-(carboxymethyl)

(2e-B) (33.8 mg, 22%, fast-eluted isomer, eluted at 55% methanol)

each as a colorless solid.

carbamoyl]pentyl(phenyl)-(R_P^⁄)-phosphono}-butanoic acid

(2g-A). (2R,S)-2-Amino-4-{(R)-1-[N-(carboxymethyl)carbamoyl]

pentyl(phenyl)-(S_P^⁄)-phosphono}-butanoic acid (2g-B)

Compound 2e-A (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.40 (t, J = 7.8 Hz, 2H), 7.36–7.25 (m, 4H),

7.23–7.17 (m, 2H) and 7.03 (d, J = 8.1 Hz, 2H) (2 ꢁ Ph), 5.12 [m,

1H, CH(OP)CONH], 3.88 (d, J = 17.4 Hz) and 3.80 (d, J = 17.4 Hz)

Compound 2g was prepared from 14g (300 mg, 0.403 mmol) by

the same procedure as described for the synthesis of 2a and was

puriﬁed by medium-pressure reversed-phase column chromatog-

raphy (ODS-S-50B) with a linear gradient of 40 to 60% methanol

in H₂O. The UV-positive (254 nm) fractions containing the product

(TLC, butan-1-ol/acetic acid/H₂O, 5:2:2, ninhydrin) were collected

(2H, Glya-H), 3.80–3.75 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.18

(dd, J = 13.8 and 4.2 Hz) and 3.10 (dd, J = 13.8 and 8.1 Hz) (2H,

PhCH₂), 2.28–2.07 (m, 4H, CH₂CH₂P). ³¹P NMR (121 MHz, D₂O):

d_P30.79. IR (KBr) 3446 (br), 3064, 1668, 1491, 1252, 1211,

1053, 1007, 947, 758, 690 cm^ꢀ1. Anal. Calcd for C₂₁H₂₅N₂O_8-

Pꢃ0.5H₂O: C, 53.28; H, 5.54; N, 5.92. Found: C, 53.45; H, 5.44; N,

5.99. HRMS (FAB, glycerol): Calcd for C₂₁H₂₆N₂O₈P (MH⁺)

465.1427; found 465.1428.

⁄

and lyophilized to afford the R_Pisomer (2g-A) (97.3 mg, 52%,

⁄

slow-eluted isomer, eluted at 60% methanol) and the S_Pisomer

(2g-B) (46.9 mg, 25%, fast-eluted isomer, eluted at 60% methanol)

each as a colorless solid.

Compound 2g-A (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.47 (dd, J = 8.1 and 7.5 Hz, 2H), 7.33 (t,

J = 7.5 Hz, 1H) and 7.26 (d, J = 8.1 Hz, 2H) (Ph), 4.97–4.87 [m, 1H,

Compound 2e-B (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.45–7.24 (m, 8H) and 7.14 (d, J = 8.4 Hz, 2H)

(2 ꢁ Ph), 5.12 [m, 1H, CH(OP)CONH], 3.71 (s, 2H, Gly

a

-H), 3.74–

CH(OP)CONH], 4.01 (d, J = 17.7 Hz, 1H, Gly

a-H_a), 3.94 (d,

3.64 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.25 (dd, J = 14.4 and 3.6 Hz)

and 3.11 (dd, J = 14.4 and 8.4 Hz) (2H, PhCH₂), 2.18–1.81 (m, 4H,

CH₂CH₂P). ³¹P NMR (121 MHz, D₂O): d_P30.64. IR (KBr) 3419 (br),

3064, 1668, 1491, 1456, 1404, 1340, 1250, 1203, 1055, 1009,

964, 941, 764, 700 cm^ꢀ1. Anal. Calcd for C₂₁H₂₅N₂O₈Pꢃ1.5H₂O: C,

51.32; H, 5.74; N, 5.70. Found: C, 51.26; H, 5.86; N, 5.66. HRMS

(FAB, glycerol): Calcd for C₂₁H₂₆N₂O₈P (MH⁺) 465.1427; found

465.1436.

J = 17.7 Hz, 1H, Gly -H_b), 3.93–3.85 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P],

a

2.47–2.18 (m, 4H, CH₂CH₂P), 1.85–1.66 [m, 2H, CH₃(CH₂)₂CH₂],

1.24–1.11 (m, 4H, CH₃CH₂CH₂), 0.78 (t, J = 6.0 Hz, 3H, CH₃). ³¹P

NMR (121 MHz, D₂O): d_P30.85. IR (KBr) 3419 (br), 2956, 1668,

1491, 1207, 1039, 1007, 941, 767, 690 cm^ꢀ1. Anal. Calcd for C₁₈H_27-

N₂O₈Pꢃ0.5H₂O: C, 49.20; H, 6.42; N, 6.38. Found: C, 49.12; H, 6.33;

N, 6.42. HRMS (FAB, glycerol): Calcd for C₁₈H₂₈N₂O₈P (MH⁺)

431.1583; found 431.1593.

Compound 2g-B (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.44 (dd, J = 7.8 and 7.4 Hz, 2H), 7.31 (t,

J = 7.4 Hz, 1H) and 7.21 (d, J = 7.8 Hz, 2H) (Ph), 4.97–4.87 [m, 1H,

CH(OP)CONH], 3.89 [t, J = 5 Hz, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.82

4.2.29. (2R,S)-2-Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]

pentyl(phenyl)-(S_P^⁄)-phosphono}-butanoic acid (2f-A). (2R,S)-2-

Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]pentyl(phenyl)-

(R_P^⁄)-phosphono}-butanoic acid (2f-B)

(s, 2H, Glya-H), 2.44–2.21 (m, 4H, CH₂CH₂P), 1.93–1.83 [m, 2H,

CH₃(CH₂)₂CH₂], 1.42–1.25 (m, 4H, CH₃CH₂CH₂), 0.87 (t, J = 6.9 Hz,

3H, CH₃). ³¹P NMR (121 MHz, D₂O): d_P30.56 and 30.53. IR (KBr)

3429 (br), 2958, 1668, 1491, 1203, 1045, 1009, 939, 766,

690 cm^ꢀ1. HRMS (FAB, glycerol): Calcd for C₁₈H₂₈N₂O₈P (MH⁺)

431.1583; found 431.1583.

Compound 2f was prepared from 14f (447 mg, 0.601 mmol) by

the same procedure as described for the synthesis of 2a and was

puriﬁed by medium-pressure reversed-phase column chromatog-

raphy (ODS-S-50B) with a linear gradient of 20 to 60% methanol

in H₂O. The UV-positive (254 nm) fractions containing the product

(TLC, butan-1-ol/acetic acid/H₂O, 5:2:2, ninhydrin) were collected

4.2.31. (2R,S)-2-Amino-4-{(S)-1-methoxycarbonylpentyl

(phenyl)-(S_P^⁄)-phosphono}butanoic acid (3-A). (2R,S)-2-Amino-

4-{(S)-1-methoxycarbonylpentyl(phenyl)-(R_P^⁄)-phosphono}

butanoic acid (3-B)

Compound 3 was prepared from 15 (479 mg, 0.784 mmol) by

the same procedure as described for the synthesis of 2a and was

puriﬁed by medium-pressure reversed-phase column chromatog-

raphy (ODS-S-50B) with a linear gradient of 40 to 60% methanol

in H₂O. The UV-positive (254 nm) fractions containing the product

(TLC, butan-1-ol/acetic acid/H₂O, 5_⁄:2:2, ninhydrin) were collected

and lyophilized to afford the S_Pisomer (3-A) (140 mg, 46%,

⁄

and lyophilized to afford the S_Pisomer (2f-A) (157 mg, 56%,

⁄

slow-eluted isomer, eluted at 60% methanol) and the R_Pisomer

(2f-B) (71.0 mg, 25%, fast-eluted isomer, eluted at 60% methanol)

each as a colorless solid.

Compound 2f-A (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.47 (t, J = 7.8 Hz, 2H), 7.33 (t, J = 7.8 Hz, 1H)

and 7.26 (d, J = 7.8 Hz, 2H) (Ph), 4.97–4.87 [m, 1H, CH(OP)CONH],

4.02 (d, J = 18.0 Hz, 1H, Glya-H_a), 3.95 (d, J = 18.0 Hz, 1H, Glya-

H_b), 3.93–3.86 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 2.45–2.18 (m, 4H,

CH₂CH₂P), 1.86–1.65 [m, 2H, CH₃(CH₂)₂CH₂], 1.24–1.11 (m, 4H,

CH₃CH₂CH₂), 0.78 (t, J = 6.3 Hz, 3H, CH₃). ³¹P NMR (121 MHz,

D₂O): d_P30.90. IR (KBr) 3435 (br), 1664, 1491, 1207, 1041,

1007, 941, 766, 690 cm^ꢀ1. Anal. Calcd for C₁₈H₂₇N₂O₈PꢃH₂O: C,

48.21; H, 6.52; N, 6.25. Found: C, 48.39; H, 6.30; N, 6.27. HRMS

(FAB, glycerol): Calcd for C₁₈H₂₈N₂O₈P (MH⁺) 431.1583; found

431.1571.

⁄

slow-eluted isomer, eluted at 60% methanol) and the R_Pisomer

(3-B) (67.3 mg, 22%, fast-eluted isomer, eluted at 60% methanol)

each as a colorless solid.

Compound 3-A (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.46 (dd, J = 7.5 and 7.5 Hz, 2H), 7.32 (t,

J = 7.5 Hz, 1H) and 7.24 (d, J = 7.5 Hz, 2H) (Ph), 5.00 [m, 1H,

CH(OP)CO₂Me], 3.85 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.79 (s, 3H,

CH₃O), 2.45–2.17 (m, 4H, CH₂CH₂P), 1.84–1.68 [m, 2H, CH₃(CH₂)_2-

CH₂], 1.24–1.07 (m, 4H, CH₃CH₂CH₂), 0.78 (t, J = 6.6 Hz, 3H, CH₃).

³¹P NMR (121 MHz, D₂O): d_P31.36. IR (KBr) 3467 (br), 2958,

1747, 1631, 1592, 1491, 1406, 1348, 1288, 1207, 1080, 1024,

935, 769, 690, 542 cm^ꢀ1. Anal. Calcd for C₁₇H₂₆NO₇P 0.5H₂O: C,

51.51; H, 6.87; N, 3.53. Found: C, 51.40; H, 6.88; N. 3.64. HRMS

(FAB, glycerol): Calcd for C₁₇H₂₇NO₇P (MH⁺) 388.1525; found

388.1519.

Compound 2f-B (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H7.44 (dd, J = 7.8 and J = 7.4 Hz, 2H), 7.31 (t,

J = 7.4 Hz, 1H) and 7.21 (d, J = 7.8 Hz, 2H) (Ph), 4.97–4.87 [m, 1H,

CH(OP)CONH], 3.88 [t, J = 5 Hz, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.80

(s, 2H, Glya-H), 2.44–2.21 (m, 4H, CH₂CH₂P), 1.93–1.83 [m, 2H,

CH₃(CH₂)₂CH₂], 1.42–1.26 (m, 4H, CH₃CH₂CH₂), 0.87 (t, J = 6.9 Hz,

3H, CH₃). ³¹P NMR (121 MHz, D₂O): d_P30.60 and 30.56. IR (KBr)

3431 (br), 2958, 1664, 1491, 1205, 1047, 1018, 937, 764,

690 cm^ꢀ1. Anal. Calcd for C₁₈H₂₇N₂O₈Pꢃ1.5H₂O: C, 47.26; H, 6.61;

1192

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

Compound 3-B (a mixture of two diastereoisomers). ¹H NMR

11.5 (br s, 1H, P-OH), 7.41–7.19 (m, 10H, 2 ꢁ Ph), 6.95 (d,

¹J_HP= 533 Hz, 1H, P-H), 5.14 and 5.07 (2 ꢁ s, 4H, 2 ꢁ PhCH₂O),

4.42 (br t, 1H, NH), 2.05–1.53 (m, 4H, CH₂CH₂P). ³¹P NMR

(121 MHz, CDCl₃): d_P36.39. The crude phosphinic acid 19 was used

for the next reaction without further puriﬁcation.

Compound 19 was dissolved in dry CH₂Cl₂(15 mL) under argon.

N,O-Bis (trimethylsilyl)acetamide (2.50 g, 12.0 mmol) and tert-bu-

tyl acrylate (564 mg, 4.40 mmol) was added successively to the

solution, and the mixture was stirred at room temperature for

22 h. The reaction mixture was cooled to 0 °C and quenched with

ethanol (3 mL) plus two drops of triﬂuoroacetic acid. The mixture

was allowed to warm to room temperature. After stirring at room

temperature for 30 min, the solvent was removed in vacuo. Ethyl

acetate (70 mL) was added to the residue, and the resulting solu-

tion was washed successively with 10% aqueous KHSO₄and sat.

NaCl, and was dried over anhydrous Na₂SO₄. The solvent was re-

moved in vacuo to afford the crude dialkylphosphinic acid 20

(300 MHz, D₂O): d_H7.46 (dd, J = 7.8 and 7.8 Hz, 2H), 7.32 (t,

J = 7.8 Hz, 1H) and 7.23 (d, J = 7.8 Hz, 2H) (Ph), 5.06 [m, 1H,

CH(OP)CO₂Me], 3.83 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.69 and

3.68 (2 ꢁ s, 3H, CH₃O), 2.45–2.14 (m, 4H, CH₂CH₂P), 1.96–1.83

[m, 2H, CH₃(CH₂)₂CH₂], 1.45–1.24 (m, 4H, CH₃CH₂CH₂), 0.87 (t,

J = 6.9 Hz, 3H, CH₃). ³¹P NMR (121 MHz, D₂O): d_P30.84. IR (KBr)

3467 (br), 2954, 1749, 1616, 1592, 1492, 1410, 1348, 1261,1211,

1085, 1055, 1028, 933, 766, 692, 555 cm^ꢀ1. Anal. Calcd for C₁₇H_26-

NO₇P 0.5H₂O: C, 51.51; H, 6.87; N, 3.53. Found: C, 51.31; H, 6.66; N

3.59. HRMS (FAB, glycerol): Calcd for C₁₇H₂₇NO₇P (MH⁺) 388.1525;

found 388.1530.

4.2.32. (2R,S)-2-Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]

propyl(methyl)-(S_P^⁄)-phosphono}-butanoic acid (4-A). (2R,S)-2-

Amino-4-{(S)-1-[N-(carboxymethyl)carbamoyl]propyl(methyl)-

(R_P^⁄)-phosphono}-butanoic acid (4-B)

Compound 4 was prepared from 16 (324 mg, 0.495 mmol) by

the same procedure as described for the synthesis of 2a and was

puriﬁed by medium-pressure reversed-phase column chromatog-

raphy (ODS-S-50B) with a linear gradient of 5 to 55% methanol

in H₂O. The ninhydrin-positive fractions (TLC, butan-1-ol/acetic

(2.43 g) as a colorless oil. TLC (CH₂Cl₂/methanol/acetic acid,

80:20:1, cerium sulfate) R_f= 0.6 (ca. 80% purity). ¹H NMR

(300 MHz, CDCl₃): d_H8.22 (br, 1H, P-OH), 7.38–7.24 (m, 10H,

2 ꢁ Ph), 5.15 and 5.08 (2 ꢁ s, 4H, 2 ꢁ PhCH₂O), 4.40 [m, 1H,

3

NH(CO)CH(CH₂)₂P], 2.45 (dt, J_HP= 11.4 Hz, J = 8.4 Hz, 2H, PCH_2-

⁄

t

acid/H₂O, 5:2:2) were collected and lyophilized to afford the S_P

CH₂CO₂Bu), 2.14 and 1.92 (2 ꢁ m, 2H, CHCH₂CH₂P), 1.90 (m, 2H,

CH₂CO₂Bu), 1.66 (m, 2H, CHCH₂CH₂P), 1.42 (s, 9H, Bu). ³¹P NMR

(121 MHz, CDCl₃): d_P57.72. IR (NaCl) 3356 (br), 2978, 1728,

1533, 1456, 1367, 1250, 1215, 1155, 1049, 976, 847, 739,

t

isomer (4-A) (68.1 mg, 41%, slow-eluted isomer, eluted at 30%

⁄

methanol) and the R_Pisomer (4-B) (21.5 mg, 13%, fast-eluted iso-

mer, eluted at 10% methanol) each as a colorless solid.

Compound 4-A (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H4.92–4.79 [m, 1H, CH(OP)CONH], 4.05 (d,

698 cm^ꢀ1HRMS (FAB, NBA): Calcd for C₂₆H₃₅NO₈P (MH⁺)

.

520.2100; found 520.2106. The crude 20 was used for the next

J = 18.0 Hz, 1H, Gly

a

-H_a), 3.98 (d, J = 18.0 Hz, 1H, Gly

a

-H_b), 3.90–

reaction without further puriﬁcation.

3.82 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.82 (d, ³J_HP= 12.0 Hz, 3H, CH_3-

OP), 2.27–2.01 (m, 4H, CH₂CH₂P), 1.93 (dq, J = 7.2 and 6.6 Hz, 2H,

CH₃CH₂), 0.98 (t, J = 7.2 Hz, 3H, CH₃). ³¹P NMR (121 MHz, D₂O): d_P

35.09. IR (KBr) 3419 (br), 2976, 1732, 1664, 1539, 1448, 1406,

1346, 1219, 1043, 1001, 893, 827, 775 cm^ꢀ1. Anal. Calcd for C₁₁H_21-

N₂O₈P 0.8H₂O: C, 37.25; H, 6.42; N, 7.90. Found: C, 37.21; H, 6.55;

N, 7.98. HRMS (FAB, glycerol): Calcd for C₁₁H₂₂N₂O₈P (MH⁺)

341.1114; found 341.1114.

4.2.35. (S)-Benzyl 2 -(N-benzyloxycarbonylamino)-4-{2-[N-

(benzyloxycarbonylmethyl)carbamoyl]ethyl(hydrogen)phos

phino}butanoate (23)

Crude compound 20 (2.43 g, ca. 4.0 mmol) was dissolved in a

mixture of triﬂuoroacetic acid (10 mL) and CH₂Cl₂(10 mL) at

0 °C. The solution was stirred at room temperature for 7 h. Triﬂu-

oroacetic acid was removed in vacuo, followed by two cycles of

azeotropic evaporation with toluene. The resulting residue was

dissolved in ethyl acetate (80 mL), washed successively with

1 N HCl and sat. NaCl, and died over anhydrous Na₂SO₄. The sol-

vent was removed in vacuo to afford the carboxylic acid 22

(1.88 g, 102%) as a pale yellow oil. To a solution of crude 22 in

CH₂Cl₂(50 mL), was added glycine benzyl ester monotosylate

(8) (1.48 g, 4.40 mmol), 1-hydroxybenzotriazole monohydrate

(HOBt H₂O) (594 mg, 4.40 mmol) and N,N-diisopropylethylamine

(2.07 g, 16.0 mmol) at 0 °C with stirring. Reaction was initiated by

adding EDC (3.07 g, 16.0 mmol). After stirring at room tempera-

ture for 5 h, the reaction mixture was concentrated in vacuo to re-

move CH₂Cl₂. Ethyl acetate (100 mL) was added to the residue,

and the resultant solution was washed successively with 1 N

HCl, sat. NaHCO₃and sat. NaCl, and was dried over anhydrous

Na₂SO₄. After removal of the solvent in vacuo, the residue was

puriﬁed by ﬂash column chromatography (4% methanol and

0.5% acetic acid in CH₂Cl₂) to afford 23 (1.23 g, 50% from 18) as

a pale yellow oil. TLC (CH₂Cl₂/methanol/acetic acid, 80:20:1, cer-

ium sulfate, R_f= 0.5). ¹H NMR (300 MHz, CDCl₃): d_H10.23 (br, 1H,

OH), 7.38–7.00 (m, 16H, 3 ꢁ Ph and NH), 5.14 and 5.11 (2 ꢁ s, 4H,

2 ꢁ PhCH₂O), 5.08 and 5.03 (2 ꢁ d, J = 12 Hz, 2H, PhCH_aH_b-

Compound 4-B (a mixture of two diastereoisomers). ¹H NMR

(300 MHz, D₂O): d_H4.90–4.78 [m, 1H, CH(OP)CONH], 4.04 (d,

J = 17.7 Hz, 1H, Glya-H_a), 3.98 (d, J = 17.7 Hz, 1H, Glya-H_b), 3.89–

3

3.83 [m, 1H, NH₂(CO₂H)CH(CH₂)₂P], 3.79 (d, J_HP= 11.1 Hz, 3H,

CH₃OP), 2.26–2.02 (m, 4H, CH₂CH₂P), 1.91 (dq, J = 7.5 and 6.6 Hz,

2H, CH₃CH₂), 0.97 (t, J = 7.5 Hz, 3H, CH₃). ³¹P NMR (121 MHz,

D₂O): d_P35.04. IR (KBr) 3428 (br), 2978, 1728, 1658, 1540, 1448,

1406, 1348, 1223, 1040, 1003, 895, 827, 773 cm^ꢀ1. HRMS (FAB,

glycerol): Calcd for C₁₁H₂₂N₂O₈P (MH⁺) 341.1114; found 341.1116.

4.2.33. (S)-Benzyl 2-(benzyloxycarbonyl)but-3-enoate (18)⁵⁵

Detailed synthetic procedure and the spectral data are written

in Supplementary data.

4.2.34. (S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-[2-(tert-

butoxycarbonyl)ethylhydroxyphosphinyl]butanoate (20)

To a solution of the (S)-vinylglycine derivative 18 (1.30 g,

4.00 mmol) and ammonium hypophosphite (NH^þH₂PO^ꢀ, 830 mg,

4

2

10.0 mmol) in methanol (9.3 mL) was added triethylborane

(4.00 mL of 1 M solution in THF, 4.00 mmol) at room temperature.

The mixture was stirred vigorously open to air. After stirring for

5 h, the reaction was concentrated in vacuo to remove the solvent.

The residual oil was partitioned between 10% aqueous KHSO₄

(30 mL) and ethyl acetate (40 mL). The aqueous layer was ex-

tracted with ethyl acetate (2 ꢁ 40 mL), and the combined the or-

ganic layers were dried over anhydrous Na₂SO₄. The solvent was

removed in vacuo to afford the phosphinic acid 19 (1.51 g, 97%)

as a colorless oil. TLC (CH₂Cl₂/methanol/acetic acid, 80:20:1, cer-

ium sulfate) R_f= 0.4, single spot. ¹H NMR (300 MHz, CDCl₃): d_H

OCONH), 4.39 [m, 1H, NH(CO)CH(CH₂)₂P], 3.99 and 3.97 (2 ꢁ s,

3

2H, Gly

a

-H_aH_b), 2.53 (dt, J_HP= 14.7 Hz and J = 7.8 Hz, 2H, PCH_2-

CH₂CONH), 2.14 and 1.95 (2 ꢁ m, 2H, CHCH₂CH₂P), 1.95 (m, 2H,

PCH₂CH₂CONH), 1.72 (m, 2H, CHCH₂CH₂P). ³¹P NMR (121 MHz,

CDCl₃): d_P54.60. IR (NaCl) 3321 (br), 2949, 1743, 1718, 1533,

1456, 1394, 1340, 1254, 1192, 1047, 970, 825, 739, 698 cm^ꢀ1

.

Anal. Calcd for C₃₁H₃₅N₂O₉P: C, 60.98; H, 5.78; N, 4.59. Found:

C, 60.69; H, 5.83; N, 4.57.

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

1193

4.2.36. (S)-Benzyl 2-(N-benzyloxycarbonylamino)-4-{2-[N-

AMC stock solution (40

lM in water) at 25 °C (the ﬁnal

c-Glu-

(benzyloxycarbonylmethyl)carbamoyl]ethyl(phenyl)-(R_P,S_P)-

phosphino}butanoate (24)

AMC concentration of 4.0 lM). The release of AMC was monitored

continuously for 10 min with a Hitachi F-2000 spectrophotometer

(350 nm excitation, 440 nm emission). AMC concentrations were

calculated using a standard calibration curve of ﬂuorescence inten-

Oxalyl chloride (699 mg, 5.51 mmol) was added to a solution of

23 (840 mg, 1.38 mmol) and one drop of DMF in dry CH₂Cl₂

(10 mL) at 0 °C. After stirring at room temperature for 1 h, toluene

was added to the reaction, and the volatiles were evaporated. The

resulting residue was dissolved in dry CH₂Cl₂(10 mL), and phenol

(259 mg, 2.75 mmol) in dry CH₂Cl₂(2 mL) was added to the solu-

tion. The solution was cooled to 0 °C, and dry Et₃N (209 mg,

2.06 mmol) was added to initiate the reaction. The consumption

of 23 was monitored by TLC (CH₂Cl₂/methanol/acetic acid,

80:20:1, cerium sulfate, R_f= 0.5), and the formation of product

was conﬁrmed by TLC (ethyl acetate/hexane, 5:1, cerium sulfate,

R_f= 0.4). After stirring at room temperature for 2 h, the reaction

mixture was concentrated in vacuo. The residue was dissolved in

ethyl acetate (100 mL) and was washed successively with 1 N

HCl and sat. NaCl, and was dried over anhydrous Na₂SO₄. After re-

moval of the solvent in vacuo, the residue was puriﬁed by ﬂash col-

umn chromatography (3% methanol in CH₂Cl₂) to afford the

sity (F) versus AMC concentration (C):

ﬂuorescence intensity was proportional to the concentration of

AMC up to 2.0 M. The Michaelis constant (K_m) for -Glu-AMC of

human GGT was determined as 12.6 M under these conditions.

DF/D

C = 0.11 nM^ꢀ1. The

l

c

l

The hydrolytic activity of E. coli GGT was measured under the same

conditions, except that the ﬁnal substrate concentration was

0.2

lM. The K_mvalue for

c-Glu-AMC of E. coli GGT was determined

as 0.2

l

M under the same conditions.

4.3.2. Time-dependent inhibition assay

The inhibition of GGT was measured under pseudo-ﬁrst order

rate conditions. Reaction was initiated by adding enzyme to a pre-

incubated mixture of varying concentrations of the inhibitor and

the substrate (the ﬁnal concentrations of 4.0 and 0.2

lM of

c-Glu-AMC for human and E. coli GGT, respectively) in 100 mM so-

phosphinate ester 24 as

a

mixture of two diastereoisomers

dium succinate buffer (pH 5.5) at 25 °C. Time-dependent inhibition

of the enzyme was followed by monitoring the release of AMC con-

tinuously for 10 min. The resulting concave progress curves were

analyzed by ﬁtting the data to the ﬁrst-order rate equation (1) to

calculate the observed pseudo-ﬁrst-order rate constants for en-

zyme inactivation (k_obs) using the KaleidaGraph v.3.5 program

package (Synergy Software).

(524 mg, 55%, a pale yellow oil). ¹H NMR (300 MHz, CDCl₃): d_H

7.37–7.08 (m, 20H, 4 ꢁ Ph), 6.61 (br t, 1H, amide NH), 5.92 (br d,

J = 7.8 Hz) and 5.84 (br d, J = 7.5 Hz) (1H, carbamate NH), 5.18–

5.06 (m, 6H, 3 ꢁ PhCH₂O), 4.40 [m, 1H, NH(CO)CH(CH₂)₂P], 4.07–

3.90 (m, 2H, Glya-H), 2.54 [m, 2H, PCH₂CH₂CONH], 2.28–1.76 [m,

6H, PCH₂CH₂CONH and CHCH₂CH₂P]. ³¹P NMR (121 MHz, CDCl₃):

d_P56.68 and 56.48. IR (NaCl) 3276 (br), 2949, 1747, 1718, 1541,

1491, 1456, 1254, 1196, 1045, 924, 825, 741, 696 cm^ꢀ1. Anal. Calcd

for C₃₇H₃₉N₂O₉P: C, 64.72; H, 5.72; N, 4.08. Found: C, 64.44; H,

5.49; N 4.05.

½Pꢂ ¼ ½Pꢂ₁½1 ꢀ expðꢀk_obstÞꢂ

ð1Þ

where [P] and [P]₁are the concentrations of AMC formed at time t

and at time approaching inﬁnity, respectively. Since the plot of k_obs

vs. inhibitor concentration ([I])

4.2.37. (S)-2-Amino-4-{2-[N-(carboxymethyl)carbamoyl]

ethyl(phenyl)-(R_P^⁄,S_P^⁄)-phosphino}butanoic acid (5)

The phosphinate ester 24 (211 mg, 0.307 mmol) was dissolved

in a mixture of methanol (20 mL) and H₂O (4 mL) containing acetic

acid (92.3 mg, 1.54 mmol). Hydrogen gas was passed through the

solution in the presence of 5% Pd/C for 2.5 h at room temperature.

The consumption of 24 was monitored by TLC (ethyl acetate/hex-

ane, 5:1, cerium sulfate, R_f= 0.3), and the formation of product

was conﬁrmed by TLC (butan-1-ol/acetic acid/H₂O, 5:2:2, ninhy-

drin, R_f= 0.35). The reaction mixture was ﬁltered through Celite,

and the ﬁltrate was evaporated. The residue was puriﬁed by med-

ium-pressure reversed-phase column chromatography on ODS-S-

50B. The column was eluted (6 mL/min) with a linear gradient of

5 to 55% methanol in H₂O. The UV-positive (254 nm) fractions con-

taining the product (TLC, butan-1-ol/acetic acid/H₂O, 5:2:2, ninhy-

drin) were collected and lyophilized to afford 5 as a mixture of two

exhibited no saturation under standard inhibitor concentrations,

the second-order rate constant for enzyme inactivation (k_on) was

calculated according to the Eq. (2) derived from the kinetic mecha-

nism described below:

k_obs¼ k_on½Iꢂ=ð1 þ ½Sꢂ=K_m

Þ

ð2Þ

where S is the substrate

c-Glu-AMC; [S] and K_mare 0.2 and 0.2

lM

(E. coli GGT), and 4.0 and 12.6 M (human GGT).

l

diastereoisomers (22.8 mg, 20%,

a

colorless solid). ¹H NMR

4.4. Ion-spray mass spectroscopy

(300 MHz, D₂O): d_H7.46 (dd, J = 7.5 and 7.5 Hz, 2H), 7.31 (t,

J = 7.5 Hz, 1H) and 7.22 (d, J = 7.5 Hz, 2H) (Ph) 3.80 [t, J = 5.1 Hz,

The mass spectra were recorded on a Sciex API-3000 mass spec-

trometer (PE Sciex) interfaced with an ion-spray ion source. The

1H, NH(CO)CH(CH₂)₂P], 3.73 (s, 2H, Glya-H), 2.64 [m, 2H, PCH₂CH_2-

2

E. coli GGT (600

(1.7 mM) in 50 mM Tris–HCl buffer (pH 8.0) (total volume

100 L) at 37 °C for 20 min or with 2a-A (4 mM) for 3 h. The activ-

ity of the partially or completely inactivated enzyme was mea-

sured by hydrolysis of -Glu-AMC. The small subunit of the

lg) was inactivated by incubation with 2a-A

CONH], 2.39 [dt, J_HP= 13.5 Hz, J = 8.1 Hz, 2H, PCH₂CH₂CONH],

2.32–2.03 (m, 4H, CHCH₂CH₂P). ³¹P NMR (121 MHz, D₂O): d_P

62.53 and 62.48. IR (KBr) 3431 (br), 3068, 1647, 1491, 1406,

1200, 1072, 1024, 928, 833, 769, 692 cm^ꢀ1. HRMS (FAB, glycerol):

Calcd for C₁₅H₂₂N₂O₇P (MH⁺) 373.1165; found 373.1168.

l

c

inactivated enzyme was isolated by reversed-phase HPLC (COSMO-

SIL 5C18-AR-300, 4.6 ꢁ 150 mm) and was lyophilized as previously

described.⁷⁰The lyophilized small subunit was dissolved in a mix-

4.3. Enzyme assay

ture of 100

was injected directly into mass spectrometer using a syringe

(1.46 mm inside diameter) at a ﬂow rate of 2.5 L/min. The quad-

rupole was scanned over a range of 500–2000 atomic mass units

(amu) with a step size of 0.1 amu and a dwell time of 0.2 ms.

Ion-spray voltage was set at 5 kV, and oriﬁce potential was 60 V.

lL of 0.1% formic acid and 100 lL of acetonitrile, and

4.3.1. General

The hydrolytic activity of human GGT was measured by using

l

4.0

as substrate at 25 °C and pH 5.5.³⁹Assays were initiated by adding

10 L enzyme stock solution to 100 mM succinate-NaOH buffer

(pH 5.5) in a total volume of 1 mL containing 100 L of -Glu-

lM 7-(N-c-glutamylamino)-4-methylcoumarin (c-Glu-AMC)

l

c

1194

M. Nakajima et al. / Bioorg. Med. Chem. 22 (2014) 1176–1194

27. Dominici, S.; Paolicchi, A.; Corti, A.; Maellaro, E.; Pompella, A. Methods Enzymol.

2005 , 29 , 484.

Data analysis were performed using the BioMultiView program (PE

Sciex).

28. Meister, A.; Anderson, M. E. Ann. Rev. Biochem. 1983 , 52 , 711.

29. Joyce-Brady, M.; Hiratake, J. Curr. Enzyme Inhib. 2011 , 7 , 71.

30. Hanka, L. J.; Martin, D. G.; Neil, G. L. Cancer Chemother. Rep. 1973 , 57 , 141.

31. Stole, E.; Seddon, A. P.; Wellner, D.; Meister, A. Proc. Natl. Acad. Sci. U.S.A. 1990 ,

87 , 1706.

Acknowledgments

We thank Asahi Kasei Corporation for the generous gift of hu-

man GGT [HC-GTP (T-72), Product Code 635-22771, 200 U]. This

study was supported in part by Grant-in-Aid for Scientiﬁc Research

from Japan Society for the Promotion of Science, the Ministry of

Education, Culture, Sports, Science and Technology in Japan to

J.H. (contract no. 80199075).

32. Stole, E.; Smith, T. K.; Manning, J. M.; Meister, A. J. Biol. Chem. 1994 , 269 , 21435.

33. Smith, T. K.; Ikeda, Y.; Fujii, J.; Taniguchi, N.; Meister, A. Proc. Natl. Acad. Sci.

U.S.A. 1995 , 92 , 2360.

34. Neil, G. L.; Berger, A. E.; McPartland, R. P.; Grindey, G. B.; Bloch, A. Cancer Res.

1979 , 39 , 852.

35. O’Dwyer, P. J.; Alonso, M. T.; Leyland-Jones, B. J. Clin. Oncol. 1984 , 2 , 1064.

36. Wada, K.; Hiratake, J.; Irie, M.; Okada, T.; Yamada, C.; Kumagai, H., et al J. Mol.

Biol. 2008 , 380 , 361.

37. Williams, K.; Cullati, S.; Sand, A.; Biterova, E. I.; Barycki, J. J. Biochemistry 2009 ,

78 , 2459.

38. Han, L.; Hiratake, J.; Tachi, N.; Suzuki, H.; Kumagai, H.; Sakata, K. Bioorg. Med.

Chem. 2006 , 14 , 6043.

39. Han, L.; Hiratake, J.; Kamiyama, A.; Sakata, K. Biochemistry 2007 , 46 , 1432.

40. Inoue, M.; Hiratake, J.; Suzuki, H.; Kumagai, H.; Sakata, K. Biochemistry 2000 , 39 ,

7764.

Supplementary data

Supplementary data associated with this article can be found, in

the online version, at http://dx.doi.org/10.1016/j.bmc.2013.12.034.

41. Castonguay, R.; Halim, D.; Morin, M.; Furtos, A.; Lherbet, C.; Bonneil, E., et al

Biochemistry 2007 , 46 , 12253.

References and notes

42. Okada, T.; Suzuki, H.; Wada, K.; Kumagai, H.; Fukuyama, K. Proc. Natl. Acad. Sci.

U.S.A. 2006 , 103 , 6471.

43. Okada, T.; Suzuki, H.; Wada, K.; Kumagai, H.; Fukuyama, K. J. Biol. Chem. 2007 ,

282 , 2433.

1. Keillor, J. W.; Castonguay, R.; Lherbet, C. Methods Enzymol. 2005 , 401 , 449.

2. Taniguchi, N.; Ikeda, Y. Adv. Enzymol. Relat. Areas Mol. Biol. 1998 , 72 , 239.

3. Allison, R. D. Methods Enzymol. 1985 , 113 , 419.

4. Hiratake, J.; Suzuki, H.; Fukuyama, K.; Wada, K.; Kumagai, H. In Handbook of

Proteolytic Enzymes ; Rawlings, N. D., Salvesen, G. S., Eds., 3rd ed.; Academic

Press: Oxford, 2013; p 3712.

5. Ikeda, Y.; Taniguchi, N. Methods Enzymol. 2005 , 401 , 408.

6. Hanigan, M. H.; Ricketts, W. A. Biochemistry 1993 , 32 , 6302.

7. Whitﬁeld, J. B. Crit. Rev. Clin. Lab. Sci. 2001 , 38 , 263.

8. Zhang, H.; Forman, H. J.; Choi, J. Methods Enzymol. 2005 , 401 , 468.

9. Hanigan, M. H. Carcinogenesis 1995 , 16 , 181.

44. Morrow, A. L.; Williams, K.; Sand, A.; Boanca, G.; Barycki, J. J. Biochemistry 2007 ,

46 , 13407.

45. Wada, K.; Irie, M.; Suzuki, H.; Fukuyama, K. FEBS J. 2010 , 277 , 1000.

46. West, M. B.; Chen, Y.; Wickham, S.; Heroux, A.; Cahill, K.; Hanigan, M. H., et al J.

Biol. Chem. 2013 , 288 , 31902.

47. Tate, S. S.; Meister, A. J. Biol. Chem. 1974 , 249 , 7593.

48. Steffens, J. J.; Sampson, E. J.; Siewers, I. J.; Benkovic, S. J. J. Am. Chem. Soc. 1973 ,

95 , 936.

10. Corti, A.; Franzini, M.; Paolicchi, A.; Pompella, A. Anticancer Res. 2010 , 30 ,

1169.

49. Bauer, T.; Gajewiak, J. Tetrahedron 2004 , 60 , 9163.

50. Koelsch, C. F. J. Am. Chem. Soc. 1930 , 52 , 1105.

11. Hanigan, M. H.; Gallagher, B. C.; Townsend, D. M.; Gabarra, V. Carcinogenesis

1999 , 20 , 553.

12. Godwin, A. K.; Meister, A.; O’Dwyer, P. J.; Huang, C. S.; Hamilton, T. C.;

Anderson, M. E. Proc. Natl. Acad. Sci. U.S.A. 1992 , 89 , 3070.

13. Pompella, A.; Tata, V. D.; Paolicchi, A.; Zunino, F. Biochem. Pharmacol. 2006 , 71 ,

231. and references cited therein.

14. OBrador, E.; Carretero, J.; Ortega, A.; Medina, I.; Rodilla, V.; Pellicer, J. A.;

Estrela, J. M. Hepatology 2002 , 35 , 74.

15. Benlloch, M.; Ortega, A.; Ferrer, P.; Segarra, R.; Obrador, E.; Asensi, M., et al J.

Biol. Chem. 2005 , 280 , 6950.

51. Wisniewski, K. Org. Prep. Proced. Int., New J. Org. Syn. 1999 , 31 , 211.

52. Hirschmann, R.; Yager, K. M.; Taylor, C. M.; Witherington, J.; Sprengeler, P. A.;

Phillips, B. W.; Moore, P.; Smith, A. B., III J. Am. Chem. Soc. 1997 , 119 , 8177.

53. Kosolapoff, G. M. J. Am. Chem. Soc. 1948 , 70 , 1971.

54. Chambers, J. R.; Isbell, A. F. J. Org. Chem. 1964 , 29 , 832.

55. Krol, W. J.; Mao, S.-S.; Steele, D. L.; Townsend, C. A. J. Org. Chem. 1991 , 56 , 728.

56. Depèle, S.; Montchamp, J.-L. J. Org. Chem. 2001 , 66 , 6745.

57. Smith, G. D.; Ding, J. L.; Peters, T. J. Anal. Biochem. 1979 , 100 , 136.

58. Cook, N. D.; Peters, T. J. Biochim. Biophys. Acta 1985 , 832 , 142.

59. Hiratake, J.; Irie, T.; Tokutake, N.; Oda, J. Biosci. Biotechnol. Biochem. 2002 , 66 ,

1500.

16. Sian, J.; Dexter, D. T.; Lees, A. J.; Daniel, S.; Jenner, P.; Marsden, C. D. Ann. Neurol.

1994 , 36 , 356.

60. Bondi, A. J. Phys. Chem. 1964 , 68 , 441.

17. Owen, A. D.; Schapira, A. H.; Jenner, P.; Marsden, C. D. Ann. N. Y. Acad. Sci. 1996 ,

786 , 217.

18. Meisinger, C.; Löwel, H.; Heier, M.; Schneider, A.; Thorand, B. J. Int. Med. 2005 ,

258 , 527.

19. Ruttmann, E.; Brant, L. J.; Concin, H.; Diem, G.; Rapp, K.; Ulmer, H. Circulation

2005 , 112 , 2130.

61. Hu, X.; Legler, P. M.; Khavrutskii, I.; Scorpio, A.; Compton, J. R.; Robertson, K. L.;

Friedlander, A. M.; Wallqvist, A. Biochemistry 2012 , 51 , 1199.

62. Hudson, R. F.; Keay, L. J. Chem. Soc. 1960 , 1859.

63. Suzuki, H.; Kumagai, H.; Tochikura, T. J. Bacteriol. 1986 , 168 , 1325.

64. Gregoret, K. M.; Rader, S. D.; Fletterick, R. J.; Cohen, F. E. Proteins Struct. Funct.

Genet. 1991 , 9 , 99.

20. Kengne, A. P.; Czernichow, S.; Stamatakis, E.; Hamer, M.; Batty, G. D. J. Hepatol.

2012 , 1083 , 57.

65. Suzuki, H.; Hashimoto, W.; Kumagai, H. J. Bacteriol. 1993 , 175 , 6038.

66. Boanca, G.; Sand, A.; Okada, T.; Suzuki, H.; Kumagai, H.; Fukkuyama, K., et al J.

Biol. Chem. 2007 , 282 , 534.

67. King, J. B.; West, M. B.; Cook, P. F.; Hanigan, M. H. J. Biol. Chem. 2009 , 284 , 9059.

68. Wickham, S.; Regan, N.; West, M. B.; Kumar, V. P.; Thai, J.; Li, P.-K.; Cook, P. F.;

Hanigan, M. H. J. Enzyme Inhib. Med. Chem. 2012 , 27 , 476.

69. Wickham, S.; Regan, N.; West, M. B.; Thai, J.; Cook, P. F.; Terzyan, S. S., et al

Biochem. J. 2013 , 450 , 547.

21. Mason, J. E.; Starke, R. D.; Van Kirk, J. E. Prev. Cardiol. 2010 , 13 , 36.

22. Lowry, M. H.; McAllister, B. P.; Jean, J.-C.; Brown, L. A. S.; Hughey, R. P.;

Cruikshank, W. W., et al Am. J. Respir. Cell Mol. Biol. 2008 , 38 , 509.

23. Stark, A.-A.; Zeiger, E.; Pagano, D. A. Carcinogenesis 1993 , 14 , 183.

24. Glass, G. A.; Stark, A.-A. Env. Mol. Mutagen. 1997 , 29 , 73.

25. Paolicchi, A.; Minott, G.; Tonarelli, P.; De Cesare, D.; Mezzetti, A.; Dominici, S.,

et al J. Invest. Med. 1999 , 47 , 151.

70. Hashimoto, W.; Suzuki, H.; Nohara, S.; Tachi, H.; Yamamoto, K.; Kumagai, H. J.

Biochem. 1995 , 118 , 1216.

26. Dominici, S.; Paolicchi, A.; Lorenzini, E.; Maellaro, E.; Comporti, M.; Pieri, L.,

et al BioFactors 2003 , 17 , 187.

Article Doi

Glutathione-analogous peptidyl phosphorus esters as mechanism-based inhibitors of γ-glutamyl transpeptidase for probing cysteinyl-glycine binding site

DOI: 10.1016/j.bmc.2013.12.034

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.bmc.2013.12.034

Full text of DOI:10.1016/j.bmc.2013.12.034

Products guided by the article

R&D Labs maybe for 172723-38-5

Relevant to this article

Hot Product