Prodrug forms of N-[(4-deoxy-4-amino-10-methyl)pteroyl]glutamate-γ- [ψP(O)(OH)]-glutarate, a potent inhibitor of folylpoly-γ-glutamate synthetase: Synthesis and hydrolytic stability

770

J. Med. Chem. 2006, 49, 770-788

Prodrug Forms of N-[(4-Deoxy-4-amino-10-methyl)pteroyl]glutamate-γ-[ψP(O)(OH)]-glutarate,

a Potent Inhibitor of Folylpoly-γ-glutamate Synthetase: Synthesis and Hydrolytic Stability

Yan Feng and James K. Coward*

Departments of Medicinal Chemistry and Chemistry, UniVersity of Michigan, Ann Arbor, Michigan 48109-1055

ReceiVed September 2, 2005

Ester prodrugs of the phosphinate pseudopeptide N-[(4-deoxy-4-amino-10-methyl)pteroyl]glutamate-γ-[ψP-

(O)(OH)]-glutarate (1a) were synthesized. H-phosphinic acids derived from N-Cbz vinyl glycine esters were

converted to the desired pseudopeptides by Michael addition to R-methyleneglutarate esters. Pivaloyloxym-

ethyl (POM) ester moieties were incorporated in both the N-terminal and C-terminal fragments prior to

formation of either C-P bond. N-Alkylation of the corresponding amides derived from p-(N-methyl)-

aminobenzoic acid with 2,4-diamino-6-(bromomethyl)pteridine gave the target compounds. POM esters of

methotrexate and the corresponding γ-glutamyl conjugate were also synthesized using the same strategy.

All prodrugs were evaluated in Chinese hamster ovary cells. Although the pseudopeptide prodrugs were

ineffective, prodrugs of methotrexate and the corresponding γ-glutamyl conjugate were equipotent with the

parent compounds. Stability of the prodrugs was investigated in both phosphate buffer and cell line medium

to provide a rationale for the observed biological data.

Introduction

Folic acid (pteroyl-L-glutamate, PteGlu) is an essential vitamin

which, when reduced to (6S)-5,6,7,8-tetrahydrofolic acid (H₄-

PteGlu), leads to one-carbon donors involved in the biosynthesis

of serine, glycine, methionine, and thymidylate and also in de

novo synthesis of purines.¹Due to the critical functions of folic

acid in human metabolism, numerous antifolates have been

designed and synthesized as drugs for chemotherapy of malig-

nant diseases, such as leukemia²and solid tumors,^3-5as well

as for nonmalignant diseases such as rheumatoid arthritis and

psoriasis.⁶In most cells, both folates and antifolates occur

exclusively as their polyglutamate conjugates,⁷which are formed

in a reaction catalyzed by folylpoly-γ-glutamate synthetase

(FPGS, EC 6.3.2.17).⁸FPGS catalyzes an ATP-dependent

ligation of glutamic acid to reduced folates, including H₄PteGlu,

as well as anticancer drugs, such as methotrexate (MTX,

AMPteGlu, 2a), 5,10-dideazatetrahydrofolate, and related de-

rivatives. There are at least two contributions of FPGS-catalyzed

biosynthesis of polyglutamate derivatives to cell viability. First,

because these conjugates cannot exit the cell,⁹the intracellular

concentration of folates and antifolates (∼10 µM) are much

higher than that found in the extracellular compartments (∼10

nM).⁷Second, the polyglutamate derivatives of folates or

antifolates are usually better substrates or inhibitors than the

parental monoglutamates in a variety of folate-dependent

reactions.⁸A Chinese hamster ovary (CHO) cell line that

transports monoglutamate folates normally, but cannot synthe-

Figure 1. Free acid and prodrug forms of AMPte-Glu-γ-{ψP(O)(OH)}-

size folylpolyglutamates, is unable to survive even in the

Glu (1a-d), AMPte-Glu (2a, 2b), and AMPte-Glu-γ-Glu (2c, 2d).

presence of supraphysiological levels of reduced monoglutamate

folates.¹⁰Accordingly, specific inhibition of FPGS, resulting

have developed resistance to antifolates due to the reduced

activity of FPGS.^4,11

in decreased levels of the polyglutamate conjugates, should lead

to a block of folate-dependent, one-carbon biosynthetic reactions

and a cessation of DNA replication in the tumor cells.^4,11In

addition, resistance to several antifolates has been attributed to

reduced FPGS activity in several tumor-derived cell lines.¹²

Thus, FPGS is a new target for cancer chemotherapy to

introduce selective folate deficiency and/or attack tumors that

On the basis of the mechanism of FPGS-catalyzed ligation

of glutamic acid to reduced folates and antifolates,¹³we

previously designed and synthesized a phosphorus-containing

pseudopeptide analogue of AMPteGlu-γ-Glu (2c), N-[(4-amino-

4-dexoxy-10-methyl)pteroyl]-L-glutamyl-γ-[Ψ{P(O)(OH)CH₂}]-

glutarate (AMPteGlu-γ-[Ψ{P(O)(OH)CH₂}]-glutarate, 1a, Fig-

ure 1),^14-16to mimic the tetrahedral intermediate formed during

the ATP-dependent reaction. This compound is the most potent

in vitro inhibitor of FPGS described to date based upon the

* Corresponding author. Phone: 734-936-2843. FAX: 734-647-4865.

E-mail: jkcoward@umich.edu.

Published on Web 12/15/2005

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 771

Table 1. Conditions for the Preparation of N-Cbz-glutamic Acid

heterocyclic pharmacophore 2,4-diaminopteridine. Unfortu-

nately, when used as an inhibitor of cell growth in human cancer

cells (CCRF-CEM), the phosphinic MTX analogue 1a was

completely ineffective,¹⁵presumably due to its highly hydro-

philic structure and inability to act as a substrate for either of

the folate transporters reduced folate carrier (RFC) or the folate

binding protein (FBP) to penetrate the cell membrane.¹⁷To

overcome this transport barrier, a prodrug strategy was adapted

and investigated as an alternative approach for intracellular

FPGS inhibition, which could ultimately be lethal to tumor-

derived cell lines.

Di-POM Ester^a

X

base

yield

entry (equiv)

(equiv)

additive (equiv)

solvent temp (%)^b

Prodrugs are derivatives of drug molecules that require a

chemical or enzymatic transformation in order to release the

pharmacologically active drug within cells. To date, prodrug

strategies have proved to be an important method in pharmacol-

ogy for the design and modification of drug candidates, because

they can result in better transport, uptake, and metabolism, all

of which are very important factors, together with the bioactivity

of the parent compound, in determining ultimate drug efficacy.

In preclinical and clinical studies, increasing the membrane

permeability of polar and hydrophilic drug candidates with good

inhibitory potency has been widely investigated.^18-24For those

candidates with acid functionalities, which have low cell

permeability and could therefore cause drug delivery problems,

masking the hydrophilic moieties results in a significant increase

of the lipophilicity, which can make them good substrates for

membrane transport.²⁵Numerous reports in the literature,

including clinical studies leading to commercial drugs, have

shown the value of using esters as prodrugs, among which

pivaloyloxymethyl (POM) esters are particularly attractive,

because they are very hydrophobic and are hydrolyzed (enzy-

matic or nonenzymatic) to the parent acid in an irreversible

reaction.^26-31Since the membrane transport barrier is presum-

ably a major reason the antifolate analogue 1a is ineffective as

an inhibitor of cell growth,¹⁵converting the phosphinic

pseudopeptide 1a to prodrug esters may be an effective solution

to overcome the transport barrier problem.³²Herein, we report

the syntheses of POM ester (1b and 1c) and methyl ester (1d)

prodrugs of the phosphinic pseudopeptide 1a (Figure 1) and an

investigation of their chemical stabilities in phosphate buffer

and cell line medium. Because FPGS inhibitor 1a is based on

the MTX heterocycle and the pharmacology of MTX has been

investigated extensively,^2,6,33two reference compounds, 2b and

2d, prodrug esters of MTX (AMPte-Glu, 2a) and its γ-glutamyl

conjugate (AMPte-Glu-γ-Glu, 2c), were also synthesized.

1

2

3

4

5

6

Cl (2.0) NaOH (2.2) AgNO₃(1.0)

Cl (2.2) K₂CO₃(2.5) AgNO₃(2.2)

Cl (2.2) K₂CO₃(2.5) NaI (2.5)

DMF

THF

acetone reflux 45

dioxane reflux 45

rt

13

10

27

Cl (3.0) DBU (3.0)

DMAP (0.2)

reflux 28

Cl (4.0) K₂CO₃(2.5) NaI (4.0)

Cl (4.0) DBU (2.5)

DMAP (0.2)

NaI (4.0)

7

8

9

I (2.2) K₂CO₃(2.5)

I (5.0) DBU (3.0)

I (5.0) Ag₂CO₃(5.0)

-

DMF

dioxane reflux 24

dioxane reflux 15

rt

24

DMAP (0.3)

-

10 I (5.0) Ag₂CO₃(3.0)

11 I (5.0) Ag₂CO₃(3.0) (MeO)₂P(O)Me (5) DMF

12 I (20) Ag₂CO₃(5.0) (MeO)₂P(O)Me (20) DMF

DMF

rt

71

80

92

^aAll of the reactions proceeded overnight. ^bIsolated yield.

studied, AMPteGlu-γ-Glu (2c) is not a good RFC ligand and is

not effectively taken up by cells^34,35in the absence of hydrolysis

by secreted γ-glutamyl hydrolase.^36,37The POM esters and their

parent acids, 1 and 2, were evaluated in wild-type CHO cells

as well as a MTX-resistant mutant³⁸and a “rescued” CHO cell

line arising from transfecting the MTX-resistant mutant with

DNA coding for human RFC.³⁹

Results and Discussion

In our previous work, 1a was synthesized successfully by

coupling of the free phosphinic acid 4a and acyl azide 3 (eq 1)

which, in turn, was obtained from a pteroic acid analogue

derived from the coupling of 2,4-diamino-6-(bromomethyl)-

pteridine and N-methyl-p-aminobenzoic acid.^14,16,40The syn-

thesis of 6(S)- and 6(R)-5,10-dideaza-5,6,7,8-tetrahydrofolic acid

and its poly-γ-glutamate derivatives have recently been reported

using similar methods.⁴¹Accordingly, based on the same

strategy, target compounds 1b-d and 2d should be accessible

as well by reaction of azide 3 with POM esters 4b-d and 4f in

the presence of an appropriate base. POM ester 2b, a derivative

of MTX should also be available via coupling of 3 with

glutamate di-POM ester (7).

The main transport protein involved in the cellular uptake of

MTX and related compounds is the reduced folate carrier (RFC).

On the basis of the ideas discussed above, the prodrug esters

should not require functional RFC for cell uptake and would

cross the cell membrane by passive diffusion. However, MTX

(2a) requires RFC for uptake, and although not as extensively

As a key step, preparation of POM ester of N-Cbz-glutamic

acid using a POM halide was then investigated (Table 1).^32,42

When using POM chloride and N-Cbz glutamic acid as the

substrates, the yields were consistently low under several

conditions (Table 1, entries 1-4), even in the presence of

772 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

Scheme 1

With di-POM N-Cbz-glutamate 5 in hand, the synthesis of

L-glutamic acid bis-POM ester, 7, was investigated (Scheme

1). Unfortunately, efforts to remove the Cbz group from the

amino group of the POM ester 5 were unsuccessful. No desired

product was obtained. Instead, pyroglutamate derivatives 9a and

9b were isolated and identified. That suggested that although

the desired free amine di-POM ester 7 was formed in the

hydrogenolysis, it was not stable. Immediately following its

formation, intramolecular reaction of the amine and POM ester

gave the more stable cyclized product, lactam 9a, part of which

reacted further with the formaldehyde generated in situ and was

then reduced to yield the N-methylated derivative 9b. The use

of Boc₂O to trap the free amine as formed during hydrogenolysis

was investigated. Unfortunately, the pyroglutamate derivatives

9a and 9b were still the major products, with only 19% of the

desired N-Boc-protected di-POM glutamate obtained. Regardless

of the low yield and the byproducts, the small amount of N-Boc-

protected POM ester was converted to the amine TFA salt

(7‚TFA) and allowed to react with the azide 3 in the presence

of triethylamine. However, this attempted synthesis of POM

ester 2b was unsuccessful and none of the desired product was

observed. The POM ester of the γ-glutamyl dipeptide 6 was

also prepared. Unfortunately, similar to the N-Cbz-glutamic

POM ester 5, removal of the Cbz from the amino group was

also unsuccessful. After hydrogenolysis, multiple spots were

observed on TLC analysis, indicating probable intermolecular

reaction of the amine with the POM ester moiety. This approach

was abandoned due to the lability of free amines 7 and 8.

catalysts such as silver nitrate and DMAP. Only when sodium

iodide was included to effect in situ generation of POM iodide,

together with an increase in the reaction temperature, were

modest yields (45%) of the desired product obtained (Table 1,

entries 5, 6). When POM iodide (POMI)⁴³was employed

directly, an excellent yield was achieved when the iodide and

the additive dimethyl methylphosphonate³²were both present

in large excess (Table 1, entry 12). When less iodide and

phosphonate were used, the yield was lower but still good (Table

1, entry 11). If the additive was not utilized, the yield of the

reaction decreased even further (Table 1, entry 10). The use of

POMI under other conditions led to unsatisfactory yields (Table

1, entries 7-9).

Another approach to the prodrugs was then pursued, and the

retrosynthetic route is illustrated in Scheme 2. In this approach,

2,4-diamino-6-(bromomethyl)pteridine 11b is used instead of

pteroyl azide 3 to introduce the heterocycle. This strategy avoids

Scheme 2

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 773

Scheme 3

the generation of free amino acid POM esters that would react

as described above (Scheme 1). Although the secondary amine

of N-Me-p-ABA might also be problematic in terms of possible

reaction with the POM esters, the low nucleophilicity of the

aromatic amine as well as steric hindrance of the adjacent phenyl

and methyl groups should prevent inter- or intramolecular

reaction of the amine and POM. The N-methyl p-aminobenzoyl-

protected POM ester of the peptide or pseudopeptide 10 could

be prepared in several ways. In the case of AMPte-Glu-(R-

POM)-γ-Glu-(R,γ-di-POM) (2d), synthesis of the N-acylated

γ-glutamyl peptide would be effected through standard solution-

phase peptide coupling reactions,⁴⁴followed by conversion to

the corresponding tri-POM ester. In the case of 1b,c, the prodrug

esters of AMPteGlu-γ-[Ψ{P(O)(OH)CH₂}]-glutarate (1a), two

routes are available. One is through an Arbuzov reaction of the

2-amino-4-bromobutyrate 16 with phosphinic acid 15 (route

A).^45,46Phosphinic acid 15 could be obtained via a Michael

addition of (TMSO)₂PH to R-methyleneglutarate 17.^47,48Bro-

mide 16 could be prepared from 2-amino-4-hydroxybutyrate 19,

which is derived from 2-aminobutyrolactone 20. The other route

is through a Michael addition of the phosphinic acid 18 to

R-methyleneglutarate 17 (route B). Phosphinic acid 18 could

be synthesized from vinylglycine 21 by a radical reaction with

ammonium hypophosphite under Montchamp’s condition.^16,49

N-Protected vinylglycine 21 could be derived from aminobu-

tyrolactone 20 following a procedure developed by our group.^16,50

Accordingly, the syntheses of prodrug esters 1 and 2 were

then pursued. POM ester 2b was first synthesized starting from

acid 22, as shown in Scheme 3. By using the same conditions

used previously to synthesize pteroyl azide 3, acid 22 was

converted in situ to the corresponding azide 23,^14,51which was

then allowed to react with di-tert-butyl L-glutamate (24a) to

give compound 25a. By treatment with TFA, the two tert-butyl

groups on 25a were removed to afford the corresponding free

acid, which was converted to the POM ester 27 was then

deprotected by hydrogenolysis to give 29. 2,4-Diamino-6-

(bromomethyl)pteridine 11b was synthesized in situ by reaction

of 2,4-diamino-6-(hydroxymethyl)pteridine hydrochloride 11a

with dibromotriphenylphosphine.¹⁴Upon addition of 29 to 11b

in the presence of Hu¨nig’s base, POM ester 2b was obtained in

good yield.

Encouraged by the successful synthesis of 2b, synthesis of

our second target compound, 2d, the tris-POM ester of AMPte-

Glu-γ-Glu (2c), was then investigated. Acid 22 was converted

to the corresponding azide 23 and then reacted with the R-tert-

butyl glutamate 24b in the presence of tetramethylguanidine

(TMG) to give acid 25b, which was then converted to the

corresponding acyl azide by reaction with diphenylphosphoryl

azide. Reaction of the resulting azide and di-tert-butyl glutamate

(24a) afforded tri-tert-butyl ester 26 in excellent yield. Treatment

of 26 with TFA gave the triacid, which was esterified to the

corresponding POM ester 28 under the standard conditions. The

Cbz group of tris-POM ester 28 was removed by hydrogenolysis

to give the coupling partner 30. The resulting secondary amine

was allowed to react with 11b, generated in situ from 11a, to

give the AMPteGlu-γ-Glu-tri-POM ester 2d in good yield.

With the successful synthesis of prodrug esters 2b and 2d

completed, a similar approach to synthesize the prodrug esters

of AMPteGlu-γ-[Ψ{P(O)(OH)CH₂}]-glutarate was then inves-

tigated, initially via using the Arbuzov reaction (Scheme 2, route

A). Accordingly, efforts were made to synthesize the phosphinic

acid 35 (Scheme 4). The dimer of methyl acrylate 32a was first

prepared following a literature procedure in which hydroquinone

was used as an additive.⁵²However, the yield of the desired

dimer was quite low (ca. 30%). Since hydroquinone usually is

used to prevent polymerization reactions, we carried out the

774 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

Scheme 4

Scheme 5

dimerization in its absence with encouraging results. Neat methyl

acrylate was treated with tri-n-butyl phosphine at low to ambient

temperature to yield the dimerized methyl ester 32a in an

acceptable yield, the best yield reported to date. Under the same

conditions, di-tert-butyl R-methyleneglutarate 32b was also

synthesized in good yield. Dimer 32a was then hydrolyzed to

give the free acid 33 in nearly quantitative yield.⁵²Upon

treatment of acid 33 with the POMI under the standard

esterification conditions, di-POM R-methyleneglutarate 34 was

obtained in modest yield. Subsequently, reaction of the POM

ester 34 with the P^IIIspecies generated in situ under Regan’s

conditions⁴⁸gave the desired phosphinic acid 35 in excellent

yield.

With the phosphinic acid 35 in hand, synthesis of bromide

39, the other component for the Arbuzov reaction, was

investigated (Scheme 5). Azide 23 was generated in situ from

acid 22 as described above (Scheme 3) followed by reaction

with R-aminobutyrolactone 20 to give lactone 36 in excellent

yield. The lactone was then hydrolyzed by treatment with

sodium hydroxide to afford the γ-hydroxybutyric acid 37, a

proposed precursor of POM ester 39. Unfortunately, after 37

was treated with POMI under the standard conditions, no POM

ester 38 was isolated. Instead, reversion to lactone 36 was

observed. A possible explanation may be that acid 37 was indeed

esterified to the corresponding POM ester 38. However, under

basic conditions, the hydroxy group at the γ-position attacked

the POM ester moiety to yield the lactone 36. This facile

intramolecular reaction is reminiscent of what was observed

during attempted removal of the Cbz group from N-Cbz-

glutamic acid bis-POM ester 5 (Scheme 1).

In addition to pursuing the synthesis of γ-bromo-POM ester

39, we also investigated conditions to realize the Arbuzov

reaction between a P^IIIspecies and a related γ-bromo ester 41

(Scheme 6). Ester 41 was first synthesized by coupling acid 22

with the methyl ester of 2-amino-4-bromobutyric acid. However,

reaction of the phosphinic acid 35 with the methyl ester 41 under

Regan’s conditions⁴⁸failed to provide the desired pseudopeptide.

A model reaction was then investigated using a smaller P^III

species, 44 (prepared in situ by the reaction of ammonium

hypophosphite and hexamethyldisilazine), which was combined

with alkyl bromide 41. Unfortunately, under several different

conditions, the desired Arbuzov substitution reaction was not

observed (Scheme 6). These failures are not surprising on the

basis of several examples in the literature that document

intramolecular attack of the amide carbonyl group at a distal

carbon bearing a good leaving group.^46,53,54In the present case,

intramolecular cyclization presumably occurred in which the

amide carbonyl of 39 or 41 attacked the γ-bromide. On the basis

of these results, this approach was abandoned.

Therefore, an alternate approach (Scheme 2, route B) to the

target phosphinate esters via Michael addition of a phosphinic

acid to an R-methyleneglutarate was explored starting from

N-Cbz-vinylglycine esters 46 (Scheme 7).¹⁶To establish the

viability of this approach in the synthesis of ester derivatives

of pseudopeptide 1a, synthesis of methyl ester 1d was explored.

Formation of the first phosphorus-carbon bond was effected

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 775

Scheme 6

Scheme 7

by reaction of the vinylglycine methyl ester 46a with ammonium

hypophosphite in the presence of triethylborane open to the air

for several hours.¹⁶Addition of the resulting phosphinic acid

47a to the Michael acceptor, dimethyl R-methyleneglutarate 32a,

using BSA as the P^IIIinitiator afforded the N-Cbz-protected

phosphorus-containing pseudopeptide, which was methylated

by trimethylsilyl diazomethane (TMSCHN₂)⁵⁵to give the

tetramethyl ester 48a.¹⁶Removal of the Cbz group on the amino

group of 48a by hydrogenolysis yielded the free amine 49a,

which was allowed to couple with acid 22 to afford the key

intermediate 50a for the prodrug ester 1d, and a possible

precursor to prodrug esters 1b and 1c.

the secondary amine precursor 51, which was allowed to react

with pteridinylmethyl bromide 11b generated in situ from 11a.

Interestingly, the isolated product was not the expected tetra-

methyl ester but the diisopropylethylamonium salt of the

phosphinic acid (P-OH), in which the methyl ester (P-OMe)

was removed. The salt was converted to the desired phosphinic

acid 1d by treatment with Dowex-H.

Demethylation of the phosphinic methyl ester may be due to

the presence of bromide ion formed in the reaction, since

phosphonic acid alkyl esters can be dealkylated by halide anion

(Figure 2, path A).^56,57However, the use of AgOTf as a bromide

ion scavenger failed to intercept the demethylation reaction.

Because the amount of the dibromotriphenylphosphine used in

the in situ formation of 11b from 11a was in excess, it is

Removal of the Cbz group of the tetramethyl ester of the

phosphinate pseudopeptide 50a by hydrogenolysis in THF gave

776 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

Figure 2. Possible mechanisms of demethylation of phosphinic acid P-O-methyl esters.

Scheme 8

possible that the reaction proceeded via an alternate mechanism.

(Figure 2, path B). Which of the two pathways is operative

remains to be determined.

of a methyl ester was chosen since the tert-butyl groups can be

easily removed with an acidic reagent, such as TFA. Following

the procedure employed to synthesize tetramethyl pseudopeptide

50a, the corresponding tris-tert-butyl ester 50b was synthesized

(Scheme 7) starting with N-Cbz-vinylglycine tert-butyl ester 46b

(see Supporting Information).^60,61Treatment of 46b with am-

monium hypophosphite and triethylborane in methanol open to

air afforded the phosphinic acid 47b. Compound 47b was then

allowed to react with BSA to form the corresponding nucleo-

philic P^IIIspecies followed by reaction with di-tert-butyl

R-methylene glutarate 32b to afford the corresponding tri-tert-

butyl phosphinic acid, which was reacted with TMSCHN₂

to give the desired pseudopeptide 48b. The Cbz group was

then exchanged for N-Me-p-ABA simply by hydrogenation and

coupling with acid 22 in the presence of EDC to provide the

desired key intermediate 50b. Surprisingly, treatment of the

pseudopeptide 50b with TFA for 2 h gave a 5:1 mixture of two

products, one of which (minor) was the expected tri-

acid containing a P-OMe ester, while the other was the fully

hydrolyzed tetraacid 52. Reaction of the mixture with TFA

overnight led to complete hydrolysis of compound 50b to

give 52 (Scheme 8). This result indicates that the phos-

phinic acid methyl ester is as sensitive to acidic hydrolysis

as are the carboxylic acid tert-butyl esters. Tetraacid 52 was

then used in an attempt to prepare the corresponding tetra-

POM ester 53 using POMI under the standard conditions.

However, the yield of 53 was very low (ca 10%), in addition to

a mixture of partially esterified mono-, di-, and tri-POM

esters.⁴²

Encouraged by the successful synthesis of the methyl ester

1d, research on the synthesis of POM ester 1b was initiated.

With the tetramethyl phosphinic pseudopeptide 50a in hand,

hydrolysis of the methyl esters to give the corresponding

tetraacid, for conversion to the desired POM ester 53 was

investigated (Scheme 8). However, hydrolysis of the tetramethyl

ester to its free acid proved to be surprisingly slow. Several

attempts using lithium hydroxide or sodium hydroxide in a

mixture of methanol and water failed, either at room temperature

or at reflux. ¹H NMR spectra of the isolated products indicated

incomplete hydrolysis of the methyl esters. Initially, it was

thought that the phosphinic acid methyl ester was the most

difficult to hydrolyze. However, after the methyl ester of the

phosphinic group was hydrolyzed using thiophenol and triethyl-

amine,^58,59further hydrolysis of the resulting trimethyl ester still

could not be completed. Because the completely hydrolyzed

and incompletely hydrolyzed products were difficult to separate

and the nonhydrolyzed sites were difficult to determine, use of

the crude hydrolysis product to prepare the corresponding POM

ester is untenable.

Considering the fact that the carboxylic methyl esters cannot

be hydrolyzed completely under standard basic conditions, but

that the phosphinic acid methyl ester can be removed during

the subsequent coupling reaction with 6-(bromomethyl)pteridine

11b (Figure 2), another approach to the synthesis of POM ester

53 was pursued. The use of a tert-butyl ester precursor instead

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 777

Scheme 9

Given the inability to convert free tetraacid 52 to the desired

POM ester precursor, 53, a strategy was adopted to install the

POM ester on the precursors prior to the formation of either

C-P bond of the phosphinate pseudopeptide. It was found that

POMI is stable to overnight exposure (room temperature) to

aqueous Na₂S₂O₃solution (Y. Feng, unpublished results). This

suggested that perhaps the POM ester moiety of a phosphinate

pseudopeptide would survive exposure to the numerous aqueous-

organic extractions in a multistep synthesis if all P-C bonds

were synthesized under acidic or neutral conditions after

installation of the POM esters on the N-terminal and C-terminal

fragments. This approach was then investigated (Scheme 9).

Initially, methyl 4-bromobutyrate 41 was converted to phe-

nylselenide 54, which, following hydrolysis to the free acid 55,

was esterified with POMI to give POM ester 56 in excellent

overall yield. It should be noted that the successful use of alkyl

bromide 41 as the electrophilic partner in the synthesis of

selenide 54 is in marked contrast to the failure of 41 to undergo

reaction with two nucleophilic P^IIIspecies, derived from 35 and

42, in the attempted synthesis of 1b (Scheme 5) and 45 (Scheme

6), respectively. Evidence for the attenuated nucleophilicity of

P^IIIspecies in reactions with unactivated alkyl halides has been

reported previously.¹⁶

Table 2. Initial Cytotoxicity Data for Prodrug Forms of FPGS

Substrates and Inhibitors (Chinese hamster ovary cells;74 IC₅₀, nM)

compd

Pro-3

R2

43-10

compd Pro-3

R2

43-10

1a

1b

1c

1d

>1000 >1000 >1000

2a

2b

2c

2d

5

8

20

6

350

75

5

9

3

5

>1000 >1000

700 350

500 >1000

800

250

300

formation of the desired tri-POM ester 59 in 46% yield over

three steps. Removal of the N-Cbz group of tris-POM pseudopep-

tide 59 afforded the secondary amine 60. Reaction of 60 with

11b (generated in situ from 11a) led to a mixture of POM ester

1b and the corresponding free phosphinic acid, 1c, the latter

presumably arising from halide-mediated demethylation (Figure

2). Methylation of the mixture with CH₂N₂in EtOAc led to

POM ester 1b, in modest yield. Since HPLC data showed that

the substitution reaction between 60 and 11b, followed by

treatment with CH₂N₂, proceeded with complete consumption

of 60 to yield a mixture of 1b and triphenylphosphine oxide,

the modest yield of the last step may be due to the lability of

the POM ester to nucleophiles such as H₂O and MeOH under

conditions used for purification by preparative HPLC or

preparative TLC.

Neutral conditions were chosen to carry out the oxidative

elimination of the phenylselenide moiety to afford the corre-

sponding olefin.⁶²The POM ester 56 was heated at reflux

temperature in a mixture of H₂O₂and THF. As expected, the

oxidative elimination reaction proceeded very smoothly and the

desired N-Cbz-vinylglycine POM ester 57 was obtained in

excellent yield. This result showed that under these conditions,

the POM ester is stable. Encouraged by this result, 57 was

subsequently converted to the H-phosphinic acid 58 using

Montchamp’s phosphorus radical reaction.^16,49Michael addition

of 58 to di-POM-R-methyleneglutarate (34) was accomplished

by BSA-mediated, in situ generation of the nucleophilic bis-

TMS derivative of 58 and monitoring of the reaction by ³¹P

NMR.¹⁶The use of TMSCHN₂in a mixture of toluene and

methanol to trap the resulting phosphinic acid as its correspond-

ing methyl ester 59 resulted in complete methanolysis of the

POM ester. However, use of CH₂N₂in EtOAc resulted in

Biological Data and Hydrolytic Stability. Cytotoxicity

assays were carried out in collaboration with Prof. Larry

Matherly, Karamanos Cancer Center and Department of Phar-

macology, Wayne State University, Detroit, MI. Five prodrug

esters (1b-1d, 2b, 2d) of the phosphinic acid-containing

pseudopeptide 1a, methotrexate 2a, and its γ-glutamyl “con-

jugate” 2c, were evaluated in three cell lines derived from

Chinese hamster ovaries (CHO). These included wild-type cells

Pro3, with a fully functional reduced folate carrier (RFC);³⁸

a

mutant cell line, R2, which is ca. 70-fold less sensitive to

antifolates due to a marked decreased expression of RFC;³⁸and

the cell line 43-10, in which sensitivity to antifolates has been

restored as a result of transfection of RFC cDNA to R2 cells,

thereby restoring high sensitivity to MTX.³⁹As shown in Table

2, pseudopeptides 1a-d were uniformly ineffective in all cell

lines. However, the prodrug forms of MTX (AMPteGlu) and

its γ-glutamyl conjugate, AMPteGlu-γGlu, were effective and

778 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

Table 3. Half-Lives of the Prodrugs in Phosphate Buffer and Cell

Medium^a

t

_1/2(h)

phosphate buffer

(20 mM, pH 8.0)

cell medium^b

(pH 8.0)

compd

1b

1c

1d

2b

2d

3

8

1

3

-

2

c

>100

3.5

0.8

0.5

^aDetermined by HPLC. ^bRPMI-1640 medium. ^cNot determined.

in several cases equipotent with the parent compound, i.e., 2b

vs 2a, 2d vs 2c. As noted in the Introduction, the POM esters

of MTX and its γ-glutamyl conjugate were synthesized for use

as controls. It was anticipated that the prodrug forms 2b and

2d would cross the cell membrane by diffusion, after which

the esters would be hydrolyzed to provide MTX or AMPteGlu-

γ-Glu, respectively.

These data suggest that hydrolysis of the prodrug esters is

occurring, either via an enzyme-catalyzed or nonenzymic route,

prior to and/or after entering the cell. If true, the resulting

γ-glutamyl peptide derived from 2d would be a substrate for

γ-glutamyl hydrolase (GH, EC 3.4.19.9), a secreted cysteine

peptidase,³⁷leading to free MTX (2a). However, the free acid-

containing phosphinic acid pseudopeptide 1a, the desired

hydrolysis product of 1b, 1c, and 1d, does not penetrate the

cell membrane¹⁵and cannot be a GH substrate. Thus, extra-

cellular hydrolysis of all esters moieties of 1b-1d, 2b, and 2d

would lead, respectively, to a highly polar phosphapeptide 1a,

which is unable to cross the cell membrane or to free MTX

(2a), the latter either directly from 2b or via GH-catalyzed

hydrolysis of the isopeptide bond of 2c following ester hy-

drolysis (2d f 2c f 2a).⁶³The proposed extracellular hydroly-

sis would explain the cytotoxicity data, i.e., the small effect of

the POM ester moieties on cell penetration and cytotoxicity of

phosphinic acid pseudopeptides 1b-d vs toxicity of POM esters

2b and 2d, which suggest ultimate hydrolysis to MTX (2a).

Similar results were obtained in vivo with 2b vs 2a and have

been reported in the patent literature.⁶⁴It should be noted that

the free phosphinic acids, 1c and 1d, have considerably

improved solubility when compared to the P-OMe ester 1b.

These solubility differences may explain the marginal improve-

ment in cytotoxicity noted for 1c and 1d vs 1b.

A search of the literature revealed a paucity of kinetics data

on the hydrolysis of POM esters in aqueous buffer or cell

medium. Therefore, we undertook a study of the stability of

the four prodrug esters in phosphate buffer (20 mM, pH 8.0)

and also in cell medium (RPMI-1640, pH 8.0). A comparison

of prodrug half-lives determined in phosphate buffer and cell

culture medium is given in Table 3. As expected on the basis

of literature precedent,⁶⁵POM esters 1b, 1c, 2b, and 2d were

more labile than the methyl ester 1d. The more rapid reaction

of prodrug esters in cell medium vs phosphate buffer is

presumably due to the presence in medium of numerous oxygen,

nitrogen, and sulfur nucleophiles (e.g., amino acids, glutathione,

vitamins, etc.) in addition to water.

Figure 3. Hydrolysis of AMPteGluPOM₂(2b) in phosphate buffer

(20 mM, pH 8.0, 5% DMSO). Data for the hydrolysis of 2b (b) are fit

to a pseudo-first-order kinetics equation. Formation of the mono-POM

derivatives (4, 0) is fit to a rectangular hyperbola. Following a lag

period of ca. 3 h, formation of MTX (2a) (O) occurs with a linear

dependence on time.

second mono-POM derivative, which in turn was formed slightly

faster than the completely hydrolyzed product, MTX (2a). The

overall hydrolysis scheme can be summarized as follows:

AMPteGlu(R,γ-OPOM₂) (2b) f AMPteGlu(R- or γ-OPOM)

f AMPteGlu(R- or γ-OPOM) f AMPteGlu (2a). On the basis

of steric considerations, we hypothesize that the γ-POM ester

of 2b is more labile than the R-POM ester, thus leading to initial

rapid hydrolysis of the γ-POM ester and slower hydrolysis of

the R-POM ester. Formation of the free acid MTX (2a) is

observed, albeit at a rate indicating that hydrolysis of the initially

formed major product is slow. After prolonged reaction (115

h), the observed products were a mono-POM ester (ca. 60%),

presumed to be AMPteGlu(R-OPOM-γ-OH), and 2a (ca. 40%),

with none of the initially formed minor product, presumed to

be AMPteGlu(R-OH-γ-OPOM), remaining.

Hydrolysis of AMPteGlu(R-OPOM)-γ-Glu(R,γ-OPOM₂) (2d)

proceeded much more rapidly (t_1/2) ca. 0.5 h) than was

observed with 2b (data not shown). However, it was soon

determined that the product of this reaction was not the

corresponding free acid but rather a pyroglutamic acid derivative

61, resulting from intramolecular reaction of the peptide nitrogen

on the γ-POM ester moiety (Scheme 10). Subsequent hydrolysis

of each of the remaining two R-POM esters occurred with t_1/2

) ca. 4 h in phosphate buffer, thus providing an estimate for

Scheme 10

The wide variety of nucleophilic reactions other than hy-

drolysis that are possible in cell medium led us to focus a more

detailed analysis of the hydrolytic stability of prodrug esters

1b-d, 2b, and 2d in phosphate buffer, pH 8.0, as described

below. The bis-POM ester of MTX, i.e., AMPteGlu(R,γ-

OPOM₂), 2b, was the first compound studied in detail (Figure

3). Product analysis (LC-MS) revealed that a product contain-

ing only one POM ester was formed much more rapidly than a

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 779

Figure 5. Hydrolysis of AMPteGlu[P(O)(OH)]GluPOM₃(1c) in

phosphate buffer (20 mM, pH 8.0, 5% DMSO). Data for the hydrolysis

of 1c (b) are fit to pseudo-first-order kinetics. Formation of the bis-

POM derivative(s) (0) is fit to a rectangular hyperbola. Following a

lag period of ca. 6 h, formation of the mono-POM derivatives (])

occurs with linear dependences on time. At any time point, the sum of

peak percentage data is not 100% due to the formation of several

unidentified minor products during these experiments as monitored by

HPLC.

Figure 4. Hydrolysis of AMPteGlu[P(O)(OMe)]GluPOM₃(1b) in

Phosphate buffer (20 mM, pH 8.0, 5% DMSO). Data for the hydrolysis

of 1b (b) are fit to pseudo-first-order kinetics. Formation of the bis-

POM derivative(s) (0) is fit to a rectangular hyperbola. Following a

lag period of ca. 3 h, formation of the mono-POM derivatives (])

occurs with linear dependence on time. At any time point, the sum of

peak percentage data is not 100% due to the formation of several

unidentified minor products during these experiments as monitored by

HPLC.

Figure 5), and >100 h (1d, data not shown). These results are

consistent with the literature reports that the POM esters are

more susceptible to hydrolysis than the corresponding methyl

esters.⁶⁵LC-MS data indicate that the tris-POM derivatives,

the rate of hydrolysis of R-POM esters in related compounds

of interest in this research.

Hydrolysis of phosphapeptides 1b-1d each proceeded at a

very different rate with half-lives of 3 h (1b, Figure 4), 8 h (1c,

Scheme 11

780 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

1b and 1c, react initially to form a mixture of bis-POM

derivatives, followed by a much slower loss of POM esters to

produce the free acid, 1a, via a series of mono-POM derivatives.

The difference in hydrolysis rates of bis-POM species derived

from the P-OMe (1b) and P-OH (1c) derivatives presumably

reflects the fact that the R-esters are closer to the central

phosphinic ester or acid, respectively, than the initially hydro-

lyzed γ-ester. The phosphinic acid 1c is ionized at pH 8.0 and

the more proximal R-esters may be less prone to attack by

negatively charged nucleophiles such as hydroxide ion. After a

more extended reaction time (24 h), the distribution of the bis-

POM:mono-POM derivatives of both 1b and 1c was determined

to be ca. 2:1. These are thought to be the γ-COOH, bis-R-POM

ester derivative and a mixture of two γ-CO₂H, mono-R-POM

esters, derived from precursors containing free γ-CO₂H and free

R-CO₂H, respectively, assuming that hydrolysis of the γ-ester

is faster than the R-ester as hypothesized above.

The postulated hydrolysis pathways are shown in Scheme

11, in which hydrolysis at the γ-ester, indicated in bold, leads

ultimately to one bis-POM ester and two mono- POM esters.

Only small amounts of the final product, a triacid (R¹) Me)

from 1b or tetraacid (1a, R¹) H) from 1c, were observed under

these conditions. All routes to these final products involve a

combination of hydrolytic reactions at one γ-ester and two

R-esters, with the latter being rate-limiting and, in the event,

preventing significant formation of 1a or its P-OMe ester.

The research described in this paper (Figures 4 and 5, Scheme

11) provides the first data on the hydrolytic lability of isopeptide

prodrug esters in phosphate buffer and cell medium, pH 8.0.

The results obtained with 1b-d indicate that spontaneous

hydrolysis of the prodrug esters does not lead to the parent

compounds, 1a. In addition, the high lipophilicity of POM esters

1b (log P ) 3.50), 1c (log P ) 2.21), 2b (log P ) 2.54), and

2d (log P ) 2.90), results in solubility problems that have been

noted by others.^42,66Thus, hydrolytic stability combined with

poor aqueous solubility leads to poor bioavailability of this initial

series of pseudopeptide prodrugs. Studies on the lability of these

esters in cell lysates and with cloned esterases are underway,

and the results will be reported in due course. Future research

will focus on the synthesis and evaluation of compounds with

alternate prodrug esters and/or containing a free γ-carboxyl

group, both of which should result in compounds with better

solubility properties as well as improved bioavailability.

1b of phosphinate pseudopeptide 1a allows for the installation

of varied ester moieties on different carboxylic acids, to give

additional prodrug candidates for structure-activity relationship

(SAR) investigations.

All newly synthesized pseudopeptide (1) and peptide (2)

prodrugs were evaluated in three cell lines derived from Chinese

hamster ovaries (CHO), including wild-type cells with a fully

functional reduced folate carrier (RFC), a mutant cell line with

decreased expression of RFC, and a cell line in which RFC

function has been restored by transfection of RFC-deficient CHO

cells with human RFC cDNA. Although the pseudopeptide

prodrugs 1b-d were ineffective in all cell lines, prodrugs 2b

and 2d were effective and in several cases equipotent with their

parent compounds, MTX and its γ-glutamyl conjugate, 2a and

2c, respectively. The cytotoxicity data suggested that hydrolysis

of the prodrugs esters may have occurred prior to entering the

cell. Thus, the hydrolysis rate of the prodrugs was investigated

in both phosphate buffer (pH 8.0) and cell line medium (pH

8.0). In phosphate buffer, all of the prodrugs, except the tris-

methyl ester 1d, could be hydrolyzed with the loss of a single

POM moiety within a few hours. In cell line medium, the initial

hydrolysis rates of the prodrugs were even faster. The results

indicate that POM ester prodrugs of substituted glutamic acid

(e.g., MTX and its γ-glutamyl conjugates, or phosphinate

pseudopeptides) contain one hydrolytically sensitive POM ester

moiety, presumably at the more accessible C-terminal γ-car-

boxyl group. Hydrolysis of the remaining POM esters, presum-

ably at the R-carboxyl group(s), is much slower. The latter esters

may be substrates for cellular esterases, a hypothesis that is

currently being tested experimentally. Since the C-terminal

γ-POM ester is assumed to be the most labile, its removal in

future prodrug candidates may be appropriate. The desired

lipophilicity of prodrug esters results in poor aqueous solubility,

a characteristic that could be partially ameliorated by the

synthesis of prodrug esters of γ-glutamyl peptides or pseudopep-

tides containing a free C-terminal γ-carboxylic acid. Research

on the development of this and other second-generation prodrugs

of 1a is currently underway in our laboratory.

Experimental Section

General Procedures. All reactions involving moisture-sensitive

reagents were carried out under an inert atmosphere of nitrogen or

argon in oven-dried glassware. All solvents used in moisture-

sensitive reactions were dried as follows. Tetrahydrofuran and

toluene were freshly distilled over sodium/benzophenone. Dichlo-

romethane was freshly distilled over CaH₂. Triethylamine was

distilled from KOH and stored over 4 Å molecular sieves. Thin-

layer chromatography was performed with aluminum-backed silica

gel 60-F254 plates. Column chromatography was performed with

silica gel 60 (230-400 mesh) according to the protocol of Still et

al.⁶⁷When applicable, column chromatography was carried out by

using the CombiFlash system (ISCO, Inc.) for better separation.

Melting points were taken on a Thomas-Hoover Mel-Temp ap-

paratus and are uncorrected. All NMR spectra were recorded on a

Bruker AVANCE DRX300 300-MHz or AVANCE DRX500 500-

MHz spectrometer. ¹H NMR spectra were recorded and are reported

as follows: chemical shifts in ppm downfield from internal

tetramethylsilane with multiplicity, integrated intensity, coupling

constant in hertz. ¹³C NMR spectra were obtained at 75 or 126

Conclusion

Several routes to synthesize prodrug esters of methotrexate,

its γ-glutamyl conjugate, and a phosphinate pseudopeptide were

investigated. The most effective method for synthesis of a

pseudopeptide POM ester prodrug uses N-Cbz vinylglycine

POM ester as the key intermediate to incorporate the N-terminal

C-P bond and (R,γ-bis-POM)-R-methyleneglutarate to set the

C-terminal P-C bond. Several other approaches proved to be

impractical for a variety of reasons. POM esters were found to

be quite stable under neutral or mild acidic conditions but more

labile under basic conditions. However, hydrolysis of the

phosphonic pseudopeptide tetramethyl ester was ineffective

under common basic conditions, whereas while the phosphinic

acid P-OMe moiety was sensitive to acid, such as TFA, and

also to halide ion. Conditions required to effect phosphorus-

carbon bond formation were also investigated and optimized.

Prodrug esters 1b, 1d, and 2b and 2d were synthesized

successfully by coupling of 2,4-diamino-6-(bromomethyl)-

pteridine with the N-p-ABA derivatives of peptide or pseudopep-

tide esters. Moreover, the synthetic route to the tris-POM ester

1

MHz and referenced to tetramethylsilane. Assignment of the H

and ¹³C peaks was aided by acquisition of DEPT and 2D spectra

(COSY, HETCOR) of the compounds. Unless otherwise mentioned,

the reagents used were purchased from Aldrich or Fisher. Am-

monium hypophosphite was purchased from Fluka and dried over

P₂O₅under vacuum for 2 days at room temperature prior to use.

Analytical HPLC method A: flow rate, 1 mL/min; eluant A,

phosphate buffer (20 mM, pH 7.0); eluant B, acetonitrile; gradient,

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 781

0 min, 2% B, 35 min, 50% B, 90 min, 50% B; column, Chrompack

Kromasil 100 C18, 250 × 4.6 mm. Analytical HPLC method B:

flow rate, 1 mL/min; eluant A, ammonium acetate buffer (20 mM,

pH 6.5); eluant B, acetonitrile; gradient, 0 min, 2% B, 30 min,

80% B, 40 min, 80% B; column, Chrompack Kromasil 100 C18,

250 × 4.6 mm). Detection was by UV/vis, λ ) 254 and 310 nm.

Pivaloyloxymethyl iodide (POMI) was prepared according to the

literature,⁴³decolorized with saturated aqueous NaS₂O₃solution,

and purified by distillation (bp 64-66 °C/10-11 mmHg, lit.⁴³bp

35 °C/0.1 mmHg). N-(N-Cbz-L-glutamyl-γ)-L-glutamic acid,⁴⁴as

well as compounds 11a,^51,68,6922,⁷⁰24a,^71,7224b,^71,7340,^16,50and

46a,¹⁶were prepared according to the literature procedure. Experi-

mental details on the synthesis of 46b from 40, via oxidative

elimination of a selenide, are given in the Supporting Information,

since such details and spectral data are not available in the

literature.^60,61

171.42, 161.23, 80.11, 49.70, 38.81, 27.23, 21.32, 21.14. Further

elution (EtOAc:hexanes ) 3:1, R_f) 0.2) to give 21 mg (40%) of

9b as a colorless oil. ¹H NMR (CDCl₃): 5.85 (d, J ) 5.5 Hz, 1H),

5.79 (d, J ) 5.5 Hz, 1H), 4.10-4.17 (m, 1H), 2.85 (s, 3H), 2.34-

2.44 (m, 3H), 2.05-2.06 (m, 1H), 1.21 (s, 9H). ¹³C NMR

(CDCl₃): δ 177.42, 171.45, 161.34, 80.23, 49.82, 38.81, 30.25,

27.23, 21.45, 21.01.

Di-tert-butyl N-[4-(N-benzyloxycarbonyl-N-methyl)aminoben-

zoyl]-L-glutamate (25a). Diphenylphosphoryl azide (1.375 g, 5

mmol) and triethylamine (505 mg, 5 mmol) were added into a

solution of acid 22 (1.14 g, 4 mmol) in 5 mL of DMF. The solution

was stirred at room temperature for 2.5 h. Di-tert-butyl L-glutamate

24a (1.554 g, 6 mmol) and another portion of triethylamine (606

mg, 6 mmol) were added. The resulting reaction mixture was then

stirred at room temperature overnight. The resulting turbid mixture

was poured into 200 mL of EtOAc and washed with 1 N HCl,

NaHCO₃, and brine. After being dried over Na₂SO₄, the solvent

was removed on a rotary evaporator to give a yellow oil as the

crude product, which was purified by flash chromatography (EtOAc:

hexanes ) 1:1, R_f) 0.6) to give 2.1 g (90% over two steps) of

Bis(pivaloyloxymethyl) N-benzyloxycarbonyl-L-glutamate (5).

To a solution of N-Cbz-L-glutamic acid (281 mg, 1 mmol) in 5

mL of DMF was added silver carbonate (828 mg, 3 mmol), dimethyl

methylphosphonate (620 mg, 5 mmol), and POMI (1.21 g, 5 mmol).

The color of the resulting mixture changed from gray greenish to

yellowish slowly. The resulting mixture was kept at room temper-

ature overnight and the precipitate was moved by filtration. The

residue was washed with EtOAc. Some precipitate was formed in

the filtrate again. The filtering operation was repeated twice until

the filtrate was clear. The filtrate was then combined, washed with

saturated NaHCO₃and brine, and dried over NaSO₄. The solvent

was removed to give a yellow oil as the crude product, which was

purified by flash chromatography (EtOAc:hexanes ) 1:5, R_f) 0.4)

1

25a as an oil. H NMR (CDCl₃): δ 7.81 (d, J ) 8.52 Hz, 2H),

7.33 (m, 5H), 7.08 (d, J ) 7.34 Hz, 1H), 5.18 (s, 2H), 4.66-4.68

(m, 1H), 3.35 (s, 3H), 2.20-2.44 (m, 3H), 2.04-2.09 (m, 1H),

1.41-1.50 (m, 18H). ¹³C NMR (CDCl₃): 173.04, 171.69, 166.69,

155.45, 146.57, 136.65, 131.36, 128.92, 128.51, 128.32, 128.17,

125.40, 82.82, 81.28, 68.02, 53.30, 37.80, 32.05, 28.45, 28.42,

27.81. ESI-HRMS (m/z): calcd for C₂₉H₃₈N₂O₇Na [M + Na]⁺

549.2577, found 549.2592.

r-tert-butyl N-[4-(N-benzyloxycarbonyl-N-methyl)aminoben-

zoyl]-L-glutamate (25b). Diphenylphosphoryl azide (206 mg, 0.75

mmol) and triethylamine (101 mg, 1 mmol) were added into a

solution of acid 22 (75 mg, 0.5 mmol) in 2 mL of DMF. The

solution was stirred for 2.5 h. R-tert-butyl glutamate 24b (102 mg,

0.5 mmol) and tetramethylguanidine (138 mg, 1.2 mmol) were

added and dissolved in 5 min. Two days later, the solvent was

removed. The residue was acidified with 5% citric acid and

extracted with EtOAc. The extracts were washed with H₂O and

brine and dried over Na₂SO₄. Purification of the crude product with

silica gel column chromatography (methanol:CH₂Cl₂) 1:15, R_f)

0.3) gave 146 mg (86%) of the desired product as a yellow oil. ¹H

NMR (CDCl₃): δ 10.30 (s, br, 1H), 7.80 (d, J ) 8.61 Hz, 2H),

7.28-7.38 (m, 7H), 7.11 (d, J ) 7.50 Hz, 1H)), 5.30 (s, 2H), 4.68-

4.75 (m, 1H), 3.33 (s, 3H), 2.45-2.53 (m, 2H), 2.27-2,32 (m,

1H), 2.03-2.10 (m, 1H), 2.17 (s, 9H). ¹³C NMR (CDCl₃): δ 177.75,

171.59, 167.66, 155.52, 146.69, 136.55, 131.00, 128.95, 128.57,

128.36, 128.24, 125.41, 83.32, 68.12, 53.05, 37.77, 30.69, 28.40,

28.03. ESI-HRMS (m/z): calcd for C₂₅H₃₀N₂O₇Na [M + Na]⁺

493.1951, found 492.1954.

1

to yield 409 mg (80%) of 5 as a colorless oil. H NMR (CDCl₃):

δ 7.32-7.42 (m, 5H), 5.74-5.91 (m, 4H), 5.45 (s, br, 1H), 5.12-

5.21 (m, 2H), 4.47 (s, br, 1H), 2.49-2.54 (m, 2H), 2.25-2.27 (m,

1H), 2.01-2.06 (m, 1H), 1.20-1.32 (m, 18H). ¹³C NMR (CDCl₃):

177.46, 177.26, 171.62, 171.01, 156.35, 136.46, 129.10, 128.92,

128.78, 128.62, 128.51, 80.32, 80.04, 67.53, 53.48, 39.01, 38.80,

30.17, 27.43, 27.28, 27.21, 27.18. ESI-HRMS (m/z): calcd for

C₂₅H₃₅NO₁₀Na [M + Na]⁺532.2159, found 532.2161.

Bis(pivaloyloxymethyl) N-[N-benzyloxycarbonyl-r-pivaloyl-

oxymethyl-L-glutamyl-γ]-L-glutamate (6). To a solution of the

N-(N-Cbz-L-glutamyl-γ)-L-glutamic acid (100 mg, 0.36 mmol) in

5 mL of DMF was added dimethyl methylphosphonate (446 mg,

3.6 mmol) and Ag₂CO₃(607 mg, 2.2 mmol) followed by the

addition of POMI (871 mg, 3.6 mmol). Some gas was released.

The color of the suspension changed from gray to yellow. The

mixture was then stirred under room temperature for 18 h. The

precipitate was filtered off, and the filtrates were poured into 200

mL of EtOAc. The whole mixture was washed with saturated

NaHCO₃and brine and dried over Na₂SO₄. Removal of the solvent

gave a yellow oil as the crude product, which was purified with

flash chromatography (EtOAc:hexanes ) 1:2, R_f) 0.55) to give

Di-tert-butyl N-{r-tert-butyl-N-[4-(N-benzyloxycarbonyl-N-

methylamino)benzoyl]-L-glutamyl-γ}-L-glutamate (26) Diphen-

ylphosphoryl azide (330 mg, 1.2 mmol) and triethylamine (151 mg,

1.5 mmol) were added into a solution of acid 25b (470 mg, 1 mmol)

in 3 mL of DMF. The solution was stirred for 2.5 h. Di-tert-butyl

L-glutamate (102 mg, 0.5 mmol) and tetramethylguanidine (173

mg, 1.5 mmol) were added. The turbid resulting mixture was then

kept at room temperature overnight. The resulting mixture was

poured into 200 mL of EtOAc and washed with 2 N HCl (40 mL

× 3), NaHCO₃(40 mL ×3), and brine (40 mL x 3). After being

dried over Na₂SO₄, the solvent was removed on a rotary evaporator

to give a colorless oil as the crude product, which was purified by

flash chromatography (EtOAc:hexanes ) 1:1, R_f) 0.35) to yield

1

126 mg (70%) of 6 as a colorless oil. H NMR (CDCl₃): δ 7.32

(s, 5H), 6.60 (d, J ) 7.34 Hz, 1H), 5.59-5.88 (m, 6H), 5.03 (s,

2H), 4.54-4.61 (m, 1H), 4.33-4.35 (m, 1H), 2.39-2.43 (m, 2H),

2.26-2.31 (m, 2H), 2.15-2.16 (m, 2H), 1.95-2.00 (m, 2H), 1.18

(s, 27H). ¹³C NMR (CDCl₃): δ 177.55, 177.48, 177.29, 172.29,

171.73, 171.18, 170.83, 156.55, 136.54, 128.90, 128.57, 128.49,

80.35, 80.25, 80.11, 67.46, 53.76, 51.88, 46.42, 39.12, 32.13, 30.26,

27.44. ESI-HRMS (m/z): calcd for C₃₆H₅₂N₂O₁₅Na [M + Na]⁺

775.3265, found 775.3278.

Pivaloyloxymethyl L-pyroglutamate (9a) and Pivaloyloxy-

methyl N-methyl-L-pyroglutamate (9b). To a solution of di-POM

N-Cbz-glutamate (6) (102 mg, 0.2 mmol) in 10 mL of EtOH was

added 20 mg of Pd/C. The resulting mixture was then stirred under

H₂at room temperature overnight. Removal of the Pd/C and solvent

gave 61 mg of the crude product, which was purified by flash

chromatography (EtOAc:hexanes ) 3:1, R_f) 0.4) to give 22 mg

1

628 mg (86%) of the title compound was obtained as an oil. H

NMR (CDCl₃): δ 7.83 (d, J ) 8.55 Hz, 2H), 7.54 (d, J ) 6.81

Hz, 1H), 7.28-7.34 (m, 7H), 6.65 (d, J ) 7.32 Hz, 1H). 5.17 (s,

2H), 4.57-4.62 (m, 1H), 4.45-4.48 (m, 1H), 3.33 (s, 3H), 2.25-

2.36 (m, 4H), 2.03-2.10 (m, 4H), 0.90-1.51 (m, 27H). ¹³C NMR

(CDCl₃): δ 172.65, 172.61, 171.57, 171.38, 166.91, 155.44, 146.50,

136.64, 131.27, 128.91, 128.48, 128.29, 125.38, 82.72, 82.66, 81.10,

67.98, 60.78, 53.50, 52.76, 37.80, 32.83, 31.88, 28.44, 28.39, 28.34,

21.44, 14.59. ESI-HRMS (m/z): calcd for C₃₈H₅₃N₃O₁₀Na [M +

Na]⁺734.3629, found 734.3649.

1

(45%) of 9a as a colorless oil. H NMR (CDCl₃): δ 6.63 (s, 1H),

δ 5.84 (d, J ) 5.5 Hz, 1H), 5.80 (d, J ) 5.5 Hz, 1H), 4.30 (dd, J₁

) 5.0 Hz, J₂) 8.7 Hz, 1H), 2.41-2.51 (m, 1H), 2.34-2.39 (m,

2H), 2.20-2.24 (m, 1H), 1.22 (s, 9H). ¹³C NMR (CDCl₃): δ 177.48,

782 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

Bis(pivaloyloxymethyl) N-[4-(N-benzyloxycarbonyl-N-methyl)-

aminobenzoyl]-L-glutamate (27). A solution of the di-tert-butyl

ester 25a (2.1 g, 4 mmol) in 5 mL of dichloromethane was cooled

to 0 °C and TFA (5 mL) was added dropwise. The resulting mixture

was then warmed to room temperature and stirred overnight. The

solvent was removed on a rotary evaporator to give a sticky oil,

which was put on the pump overnight. The oil was then dissolved

in 20 mL of DMF. To the solution was added dimethyl meth-

ylphosphonate (7.44 g, 30 mmol), silver carbonate (8.28 g, 30

mmol), and POMI (14.5 g, 30 mmol). The whole was then stirred

at room temperature overnight. The solid was filtered off. The

filtrate was collected and poured into 400 mL of EtOAc. The

combined organic solution was washed with NaHCO₃and brine

and dried over Na₂SO₄. The solvent was removed on a rotary

evaporator to give a black yellow oil, which was purified by silica

gel column chromatography (EtOAc:hexanes ) 1:2, R_f) 0.3) to

without further purification. Evaporation of solvent gave 480 mg

(95%) of 29. ¹H NMR (CDCl₃): δ 7.64 (d, J ) 8.61 Hz, 2H), 6.77

(d, J ) 7.47 Hz, 1H), 6.54 (d, J ) 8.64 Hz, 2H), 5.86 (d, J ) 5.43,

1H), 5.67-5.74 (m, 3H), 4.76-4.83 (m, 1H), 4.33 (s, br, 1H), 2.84

(s, 3H), 2.46-2.58 (m, 2H), 2.25-2.31 (m, 1H), 2.02-2.14 (m,

1H), 1.16-1.20 (m, 18H). ¹³C NMR (CDCl₃): δ 177.52, 177.36,

172.24, 171.42, 167.54, 152.70, 129.28, 121.32, 111.69, 80.30,

80.00, 52.24, 39.13, 39.10, 30.56, 30.50, 27.21. ESI-HRMS (m/z):

calcd for C₂₅H₃₆N₂O₉Na [M + Na]⁺531.2319, found 531.2328.

Bis(pivaloyloxymethyl) N-[(4-deoxy-2,4-diamino-10-methyl)-

pteroyl]-L-glutamate (2b). To a suspension of dibromotriphenyl-

phosphine (253 mg, 0.6 mmol) in 1 mL of dimethylacetamide was

added (2,4-diamino-6-pteridinyl)methanol hydrochloride (11a) (45

mg, 0.2 mmol). The resulting mixture turned to clear and red in a

few minutes with emission of gas. The reddish solution was stirred

at room temperature for 18 h before diethylisopropylamine (103

mg, 0.8 mmol) and a solution of the POM ester 29 (152 mg, 0.3

mmol) in dimethylacetamide (1 mL) was added. The resulting

reaction mixture was then kept at 60 °C for 2 days. The solvent

was removed and the residue was extracted with 200 mL of EtOAc.

The extracts were combined and washed with NaHCO₃and brine

and dried over sodium sulfate. Removal of the solvent gave 302

mg of the crude product, which was purified by silica gel column

chromatography (MeOH:CH₂Cl₂) 1:20, R_f) 0.3) to yield 113

1

yield 1.5 g (60%) of the desired product as a pale yellow oil. H

NMR (CDCl₃): δ 7.81 (d, J ) 8.70 Hz, 2H) 7.26-7.40 (m, 7H),

6.95 (d, J ) 7.45, 1H), 5.91 (d, J ) 5.49 Hz, 1H), 5.71-5.80 (m,

3H), 5.20 (s, 2H), 4.82-4.83 (m, 1H), 3.38 (s, 3H), 2.52-2.58(m,

2H), 2.16 (m, 1H), 2.14-2.16 (m, 1H), 1.20-1.23 (m, 18H). ¹³C

NMR (CDCl₃): 177.48, 177.33, 172.26, 171.01, 166.90, 155.43,

146.86, 136.61, 130.71, 128.94, 128.55, 128.34, 128.22, 125.43,

80.45, 80.10, 68.08, 52.52, 39.16, 39.12, 37.75, 30.49, 27.23, 26.96.

ESI-HRMS (m/z): calcd for C₃₃H₄₂N₂O₁₁Na [M + Na]⁺665.2686,

found 665.2695.

Bis(pivaloyloxymethyl) N-{r-pivaloyloxymethyl-N-[4-(N-Cbz-

N-methyl)aminobenzoyl]-L-glutamyl-γ}-L-glutamate (28). To a

solution of the tert-butyl ester 26 (711 mg, 1 mmol) in 5 mL of

CH₂Cl₂at 0 °C was added 5 mL of TFA dropwise. The resulting

solution was warmed to room temperature and stirred for another

4 h. The solvent was removed in vacuo. The TFA in the residue

was removed by multiple coevaporations with CH₂Cl₂and hexanes.

The resulting residue was then put on the high-vacuum pump

overnight to afford 560 mg (quantitative) of 27 as a white oily

solid. ¹H NMR (MeOH-d₄): δ 7.88 (d, J ) 8.60 Hz, 2H), 7.40 (d,

J ) 8.60 Hz, 2H), 7.29-7.35 (m, 5H), 5.16 (s, 2H), 4.57-4.60

(m, 1H), 4.43-4.46 (m, 1H), 3.34 (s, 3H), 2.44-2.47 (m, 2H),

2.38-2.41 (m, 2H), 2.30-2.34 (m, 1H), 2.10-2.18 (m, 2H), 1.90-

1.93 (m, 1H). ¹³C NMR (MeOH-d₄): δ 175.3, 174.2, 174.0, 173.9,

168.6, 155.9, 146.6, 136.7, 131.6, 128.6, 128.3, 128.2, 128.0, 125.5,

67.8, 53.0, 52.1, 36.8, 32.1, 30.2, 27.0, 26.9, 26.7.

1

mg (83%) of 2b as a yellow oil. H NMR (methanol-d₄): δ 8.49

(s, 1H), 7.68 (d, J ) 8.80 Hz, 2H), 6.76 (d, J ) 8.86 Hz, 2H),

5.68-5.82 (m, 4H), 4.75 (s, 2H), 4.58 (dd, J₁)9.14 Hz, J₂)5.46

Hz, 1H), 3.28-3.29 (m, 1H), 3.17 (s, 3H), 2.50 (t, J ) 7.28, 2H),

2.03-2.24 (m, 2H), 1.11 (s, 18H). ¹³C NMR (methanol-d₄): δ

177.32, 177.24, 171.90, 171.18, 169.26, 163.64, 162.99, 154.45,

152.15, 149.21, 148.16, 129.29, 122.33, 121.00, 111.58, 80.07,

79.81, 55.54, 52.47, 38.71, 30.13, 26.24, 25.87; IR (KBr, cm^-1):

3332, 3161, 2975, 1755, 1632, 1607, 1557, 1509, 1480, 1450, 1366,

1332, 1282, 1201, 1157, 1114, 1032, 986, 827, 767, 736. ESI-MS

(m/z): [M + Na]⁺: 705.3. HRMS (m/z): calcd for C₃₂H₄₂N₈O₉

[M + Na]⁺705.2972, found 705.2967. UV λ_max: (0.1 N NaOH)

255, 306, 378, 460 nm; (0.1 N HCl) 245, 308. Analytical HPLC

method A: t_R) 47.6 min. Analytical HPLC method B: 27.4. Anal.

(C₃₂H₄₂N₈O₉·0.5H₂O): C, H, N.

Bis(pivaloyloxymethyl) N-{r-pivaloyloxymethyl-N-[4-(N-meth-

yl)aminobenzoyl]-L-glutamyl-γ}-glutamate (30). To a solution of

the POM ester 28 (204 mg, 0.23 mmol) in 10 mL of EtOH was

added 40 mg of Pd/C, and the mixture stirred under H₂at room

temperature overnight, after which the catalyst was removed by

filtration. Evaporation of the filtrate gave 180 mg of the crude

product, which was purified by silica gel column chromatography

(EtOAc:hexanes ) 1:1, R_f) 0.25) to yield 165 mg (96%) of 30 as

To a solution of acid 27 prepared above (163 mg, 0.3 mmol) in

15 mL of DMF were added dimethyl methylphosphonate (1.12 g,

9 mmol), Ag₂CO₃(745 mg, 2.7 mmol), and POMI (2.18 g, 9 mmol).

The resulting mixture was allowed to stir at room temperature

overnight. The solution was then concentrated and the residue was

extracted with EtOAc. The extracts were combined and washed

with saturated NaHCO₃and brine and dried over Na₂SO₄. The

solvent was removed to give 346 mg of a red oil as the crude

product, which was purified by flash chromatography (EtOAc:

hexanes ) 1:1, R_f) 0.3) to yield 227 mg (86%) of 28 as a colorless

1

a colorless oil. H NMR (CDCl₃): δ 7.68 (d, J ) 8.5 Hz, 2H),

7.23 (d, J ) 6.8 Hz, 1H), 6.81 (d, J ) 7.5 Hz, 1H), 6.59 (d, J )

8.5 Hz, 2H), 5.88 (d, J ) 5.5 Hz, 1H), 5.78 (d, J ) 5.5 Hz, 1H),

5.74 (d, J ) 3.4 Hz, 1H), 5.73 (d, J ) 3.4 Hz, 1H), 5.70 (d, J )

5.5 Hz, 1H), 5.67 (d, J ) 5.5 Hz, 1H), 4.69-4.73 (m, 1H), 4.58-

4.63 (m, 1H), 2.35-2.52 (m, 4H), 2.17-2.28 (m, 2H), 2.06-2.14

(m, 1H), 1.98-2.05 (m, 1H), 1.21 (s, 9H), 1.20 (s, 9H), 1.19 (s,

9H). ¹³C NMR (CDCl₃): δ 177.6, 177.5, 177.3, 173.0, 171.7, 171.4,

170.7, 167.7, 152.4, 129.4, 121.7, 112.0, 80.4, 80.3, 80.1, 52.8,

52.0, 39.2, 39.1, 32.6, 30.8, 30.3, 27.7, 27.6, 27.2, 26.9. ESI-HRMS

(m/z): calcd for C₃₆H₅₃N₃O₁₄Na [M + Na]⁺774.3425, found

774.3445.

1

oil. H NMR (CDCl₃): δ 7.83 (d, J ) 8.62 Hz, 2H), 7.71 (d, J )

6.66 Hz, 1H), 7.29-7.36 (m, 7H), 6.65 (d, J ) 7.60 Hz, 1H), 5.90

(d, J ) 5.47 Hz, 1H), 5.78 (d, J ) 5.47 Hz, 1H), 5.75 (d, J ) 2.05

Hz, 1H), 5.74 (d, J ) 2.05 Hz, 1H), 5.70 (d, J ) 5.49 Hz, 1H),

5.66 (d, J ) 5.49 Hz, 1H), 4.67-4.71 (m, 1H), 4.60-4.64 (m,

1H), 2.39-2.51 (m, 4H), 2.24-2.30 (m, 1H), 2.13-2.23 (m, 2H),

2.00-2.06 (m, 1H), 1.21 (s, 9H), 1.20 (s, 9H), 1.19 (s, 9H). ¹³C

NMR (CDCl₃): δ 177.6, 177.5, 177.3, 173.0, 171.8, 171.0, 170.6,

167.0, 155.4, 146.8, 136.7, 128.9, 128.5, 128.3, 125.4, 80.5, 80.4,

80.2, 68.0, 53.1, 52.1, 39.2, 39.1, 37.8, 32.4, 30.3, 27.5, 27.3, 27.2,

27.0, 26.9, 21.4. ESI-HRMS (m/z): calcd for C₄₄H₅₉N₃O₁₆Na [M

+ Na]⁺908.3793, found 908.3816.

Bis(pivaloyloxymethyl) N-{N-[(4-deoxy-2,4-diamino-10-meth-

yl)pteroyl]-r-pivaloyloxymethyl-L-glutamyl-γ}-glutamate (2d).

To the suspension of dibromotriphenylphosphine (152 mg, 0.36

mmol) in 1 mL of dimethylacetamide (DMA) was added (2,4-

diamino-6-pteridinyl)methanol hydrochloride (11a) (28 mg, 0.12

mmol). The resulting mixture turned to clear and red in a few

minutes with emission of gas. The reddish solution was stirred at

room temperature for 20 h before diisopropylethylamine (62 mg,

0.48 mmol) and a solution of the POM ester 30 (135 mg, 0.18

mmol) in dimethylacetamide (1 mL) was added. The resulting

mixture was then kept at 60 °C for 30 h and monitored by analytical

HPLC. The solvent was removed and the residue was extracted

Bis(pivaloyloxymethyl) N-[4-(N-methylamino)benzoyl]-L-

glutamate (29). The di-POM ester 27 (642 mg, 1 mmol) was

dissolved in 10 mL of ethanol. To the solution was added 130 mg

(20%) of 5% palladium on activated carbon. The mixture was

saturated with hydrogen and stirred overnight under H₂, after which

the catalyst was removed by filtration with a short silica pad. TLC

showed the product was sufficiently pure for use in the next step

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 783

with 200 mL of EtOAc. The extracts were combined and washed

with NaHCO₃and brine and dried over Na₂SO₄. Removal of the

solvent gave 310 mg of the crude product, which was purified by

silica gel column chromatography (MeOH:CH₂Cl₂) 1:20, R_f)

0.25) to yield 83 mg (75%) of 2d as a yellow oil. ¹H NMR

(CDCl₃): δ 8.56 (s, 1H), 7.65 (d, J ) 8.4 Hz, 2H), 7.51 (d, J )

5.9 Hz, 1H), 7.43 (d, J ) 7.2 Hz, 1H), 6.89 (s, br, 2H), 6.67 (d, J

) 8.4 Hz, 2H), 6.25 (s, br, 2H), 5.86 (d, J ) 5.5 Hz, 1H), 5.77 (d,

J ) 5.5 Hz, 1H), 5.72 (d, J ) 4.8 Hz, 1H), 5.71 (d, J ) 4.8 Hz,

1H), 5.68 (d, J ) 5.4 Hz, 1H), 5.64 (d, J ) 5.4 Hz, 1H), 4.60-

4.66 (m, 4H), 3.11 (s, 3H), 2.40-2.49 (m, 4H), 2.20-2.24 (m,

1H), 2.13-2.19 (m, 2H), 1.99-2.05 (m, 1H), 1.19 (s, 18H), 1.17

(s, 9H). ¹³C NMR (CDCl₃): δ 177.5, 177.4, 177.3, 173.2, 171.7,

171.6, 171.0, 167.7, 163.4, 163.0, 155.4, 151.9, 149.9, 147.4, 129.4,

122.4, 121.5, 111.8, 80.4, 80.3, 80.1, 56.1, 52.9, 52.0, 39.5, 39.1,

39.1, 32.5, 30.3, 27.5, 27.2, 27.1, 26.8; IR (KBr, cm^-1):3322, 3155,

2974, 2935, 2874, 1755, 1607, 1558, 1535, 1508, 1481, 1449, 1368,

1332, 1281, 1200, 1157, 1112, 1029, 985, 942, 854, 827, 766. ESI-

MS (m/z): [M + Na]⁺: 948.4. HRMS (m/z): calcd for C₄₃H₅₉N₉O₁₄

[M + Na]⁺948.4079, found 948.4086. UV λ_max(phosphate buffer,

pH 7): 225, 360, 310, 380 nm^-1. Analytical HPLC method A: t_R

) 58.2 min. Analytical HPLC method B: t_R) 29.9 min. Anal.

(C₄₃H₅₉N₉O₁₄·0.5H₂O): C, H, N.

Bis(pivaloyloxymethyl) 2-[(hydroxyphosphinoyl)methyl]glu-

tarate (35). The suspension of ammonium hypophosphite (249 mg,

3 mmol) and 1,1,1,3,3,3-hexmethyldisilazane (497 mg, 3.1 mmol)

was heated at ∼110 °C under argon. Some gas was released and

some sublimate was formed on the wall of the flask. Ninety minutes

later, the mixture turned homogeneous and then cooled to room

temperature, followed by the addition of the solution of the POM

ester 34 (186 mg, 0.5 mmol) in 2 mL of freshly distilled CH₂Cl₂.

The resulting mixture was stirred at room temperature overnight.

HCl (5 mL, 2 N) was used to quench the reaction. The organic

layer was separated and the aqueous layer was extracted with CH₂-

Cl₂(30 mL × 3). The organic extracts were combined and dried

over Na₂SO₄. Removal of the solvent gave 205 mg (94%) of 34 as

1

a colorless oil. H NMR (CDCl₃): δ 9.99 (s, 1 H), 7.15 (d, J )

567.4 Hz, 1H), 5.69-5.84 (m, 4H), 2.90 (m, 1H), 2.37-2.43 (m,

2H), 2.17∼2.20 (m, 1H), 1.89-2.03 (m, 3H), 1.15-1.22 (m, 18H).

¹³C NMR (CDCl₃): δ 177.44, 177.32, 172.73, 172.64, 171.40,

80.20, 79.96, 39.10, 38.17, 31.77, 31.16, 30.51, 27.40, 27.21. ³¹P

NMR (CDCl₃): δ 35.02. ESI-HRMS (m/z): calcd for C₁₈H₃₁O₁₀-

PNa [M + Na]⁺461.1553, found 461.1554.

(2S)-2-N-[4-(N-carbobenzyloxy-N-methyl)aminobenzoyl]ami-

nobutyrolactone (36). To a solution of acid 22 (1.14 g, 4 mmol)

were added diphenylphosphoryl azide (1.1 g, 4 mmol) and

triethylamine (505 mg, 5 mmol). The resulting solution was stirred

at room temperature for 2.5 h, followed by the addition of (S)-2-

aminobutyrolactone hydrobromide 20 (728 mg, 4 mmol). The

resulting mixture was allowed to react at room temperature

overnight and then poured into 200 mL of EtOAc, washed with

saturated NaHCO₃and brine and dried over Na₂SO₄. Removal of

the solvent and purifying the crude product with silica gel column

chromatography gave 1.4 g (95%) of 36 as a white solid. ¹H NMR

(CDCl₃): δ 7.76 (d, J ) 8.60 Hz 2H), 7.23-7.30 (m, 7H), 7.21

(d, J ) 6.40 Hz, 1H) 5.15 (s, 2H), 4.74 (q, J ) 8.74 Hz, 1H), 4.37

(t, J ) 8.79 Hz, 1H), 4.19 (dd, J₁)8.90 Hz, J₂)6.51 Hz, 1H), 3.28

(s, 3H), 2.59 (m, 1H), 2.31 (m, 1H). ¹³C NMR (CDCl₃): δ 176.40,

176.24, 167.34, 167.27, 155.48, 146.69, 136.57, 130.39, 128.94,

128.55, 128.38, 128.25, 125.24, 68.04, 66.53, 60.82, 49.70, 37.67,

29.57, 21.45, 14.59. ESI-HRMS (m/z): calcd for C₂₀H₂₀N₂O₅Na

[M + Na]⁺391.1270, found 391.1273.

(2S)-2-N-[4-(N-carbobenzyloxy-N-methyl)aminobenzoyl]amino-

4-hydroxybutyric Acid (37). To a solution of the lactone 36 (736

mg, 2 mmol) in 3 mL of methanol was added 3 mL of 1 N NaOH.

The mixture was stirred at room temperature for 2 h. The methanol

was removed on a rotary evaporator. The resulting aqueous mixture

was acidified with 2 N HCl and extracted with EtOAc. The extracts

were put together and dried over dried over Na₂SO₄. Removal of

the solvent gave 710 mg (92%) of 37 as a colorless oil, which was

used without further purification. ¹H NMR (CD₃OD): δ 7.85 (d, J

) 7.76 Hz, 2H), 7.36 (d, J ) 8.18 Hz, 2H), 7.30 (m, 5H), 5.14 (s,

2H), 4.80 (t, J ) 9.72 Hz, 1H), 4.44 (t, J ) 8.22 Hz, 1H), 4.29 (q,

J ) 7.28 Hz, 1H), 3.28 (s, 3H), 2.53 (m, 1H), 2.41 (t, J ) 10.4 Hz,

1H). ¹³C NMR (CDCl₃): δ 176.61, 168.02, 155.72, 146.76, 136.73,

130.83, 128.67, 128.30, 128.24, 125.40, 67.84, 66.40, 49.53, 36.90,

28.47, 20.00. ESI-HRMS (m/z): calcd for C₂₀H₂₂N₂O₆Na [M +

Na]⁺409.1376, found 409.1362.

(2S)-Methyl 4-bromo-2-N-[4-(N-carbobenzyloxy-N-methyl)-

aminobenzoyl]aminobutyrate (41). To a solution of the free acid

40 (395 mg, 1.5 mmol) in 5 mL of anhydrous MeOH at 0 °C was

added SOCl₂(0.22 mL, 3 mmol) dropwise. The resulting solution

was then warmed to room temperature and stirred overnight.

Removal of the solvent in a vacuum gave 783 mg of a yellow solid

as the desired product, which was used in the next step without

purification. To the mixture of the product obtained above and p-(N-

carbobenzoyloxy-N-methyl-amino)benzoic acid (456 mg, 1.6 mmol)

in 10 mL of freshly distilled CH₂Cl₂at 0 °C were added Et₃N (162

mg, 1.6 mmol) and EDC (307 mg, 1.6 mmol). The solution was

then warmed to room temperature and stirred for 2 h. HCl (2 mL,

2 N) was added to quench to reaction. The resulting mixture was

diluted with 100 mL of EtOAc, washed with NaHCO₃and brine,

and dried over Na₂SO₄. Removal of the solvent gave 603 mg

of the crude product, which was purified by flash chroma-

Dimethyl r-methyleneglutarate (32a). Methyl acrylate (31a)

(45 mL, 43 g, 500 mmol) was cooled to -10 °C, followed by the

addition of tri-n-butylphosphine (10.1 g, 50 mmol). (Caution: This

reaction is exothermic). The reaction was allowed to warm to room

temperature, and after stirring for 1.5 h, the solution was concen-

trated on a rotary evaporator and the residue distilled under reduced

pressure to give 27.9 g (65%) of 32a as a colorless oil (bp 95-98

1

°C/10-11 Torr, lit.⁵²bp 74-77 °C/9.5 Torr). H NMR (CDCl₃):

δ 6.20 (s, 1H), 5.61 (s, 1H), 3.76 (s, 3H), 3.64 (s, 3H), 2.64 (t, J

) 6.57 Hz, 2H), 2.52 (t, J ) 7.86 Hz). ¹³C NMR (CDCl₃): δ

173.52, 167.51, 139.17, 126.38, 52.31, 52.01, 33.30, 27.73.

Di-tert-butyl r-methyleneglutarate (32b). To neat tert-butyl

acrylate (31b) (4.38 g, 34 mmol) was added tri-n-butylphosphine

(808 mg, 4 mmol) at room temperature, and the resulting mixture

was stirred for 2 h. The desired product was then isolated through

distillation (bp 95-98 °C/2 mmHg, lit.⁷⁴bp 67-70 °C/0.3 mmHg)

to afford 3.27 g (75%) of 32b as a colorless oil. ¹H NMR

(CDCl₃): δ 6.08 (s, 1H), 5.49 (d, J ) 1.27 Hz, 1H), 2.55 (t, J )

7.39 Hz, 2H), 2.55 (t, J ) 7.5 Hz, 2H), 2.41 (t, J ) 7.5 Hz, 2H),

1.54 (s, 9H), 1.44 (s, 9H). ¹³C NMR (CDCl₃): δ 172.59, 166.43,

141.12, 124.78, 81.06, 80.73, 34.74, 28.50, 28.46, 27.96.

r-Methyleneglutaric acid (33). A mixture of the dimethyl ester

32a (3.44 g, 20 mmol), NaOH (2.4 g, 60 mmol), 10 mL of

methanol, and 10 mL of H₂O was heated to reflux for 2 h. The

solution was acidified with 6 N HCl to pH < 1. The methanol was

removed on a rotary evaporator. The aqueous mixture was then

extracted with EtOAc (100 mL × 5). The organic extracts were

combined and dried over Na₂SO₄. Removal of the solvent afforded

2.76 g (96%) of a white solid as the desired product. Mp: 130-

1

132 °C (Lit.¹⁶mp: 130-136 °C). H NMR (DMSO-d₆): δ 12.33

(s, 1H), 6.04 (s, 1H), 5.59 (s, 1H), 2.34∼2.50 (m, 4H). ¹³C NMR

(DMSO- d₆): δ 174.55, 168.63, 140.50, 125.67, 33.34, 27.61.

Bis(pivaloyloxymethyl) r-methyleneglutarate (34). To the

solution of the acid 33 (710 mg, 5 mmol) in 10 mL of DMF were

added Ag₂CO₃(4.14 g, 15 mmol), dimethyl methylphosphonate

(3.72 g, 30 mmol), and POMI (7.26 g, 30 mmol). The resulting

mixture was stirred overnight. The solid was filtered off. The residue

was washed with 200 mL of EtOAc. The filtrate was collected and

washed with NaHCO₃and brine. After being dried over Na₂SO₄,

the solvent was removed. Purification of the residue by silica gel

column chromatography (EtOAc:hexanes ) 1:8, R_f) 0.55) gave

1

780 mg (42%) of 34 as a colorless oil. H NMR (CDCl₃): δ 6.27

(s, 1H), 5.82 (s, 2H), 5.73 (s, 2H), 5.70 (s, 1H), 2.64 (t, J ) 6.87

Hz, 2H), 2.55 (t, J ) 6.16 Hz, 2H), 1.19 (s, 18H). ¹³C NMR

(CDCl₃): δ 177.51, 171.57, 165.39, 137.93, 128.40, 80.12, 79.81,

39.17, 39.11, 32.96, 27.22. ESI-HRMS (m/z): calcd for C₁₈H₂₈O₈-

Na [M + Na]⁺395.1682, found 395.1683.

784 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

tography (hexanes:EtOAc ) 1:1, R_f) 0.5) to give 528 mg (76%

over two steps) of 41 as a colorless oil. ¹H NMR (CDCl₃): δ 7.77

(d, J ) 8.54 Hz, 2H), 7.25-7.29 (m, 7H), 5.14, (s, 2H), 3.79 (s,

3H), 3.58 (t, J ) 6.56 Hz, 1H), 3.42 (t, J ) 6.82 Hz, 1H), 3.28 (s,

3H), 2.29-2.49 (m, 2H). ¹³C NMR (CDCl₃): δ 172.60, 167.20,

155.44, 146.64, 136.58, 130.86, 128.92, 128.54, 128.38, 128.25,

125.27, 68.03, 52.89, 52.16, 41.34, 37.70, 35.35, 29.17. ESI-HRMS

(m/z): calcd for C₂₁H₂₃BrN₂O₅Na [M + Na]⁺485.0688, found

485.0683.

in the next step without further purification. ¹H NMR (CDCl₃): δ

11.18 (s, br, 1H), 7.29-7.34 (m, 5H), 7.07 (d, J ) 553 Hz, 1H),

5.09 (s, 2H), 4.29 (m, 1H), 2.03-2.11 (m, 1H), 1.78-1.90 (m,

3H), 1.45 (s, 9H). ¹³C NMR (CDCl₃): δ 170.98, 156.58, 136.55,

128.94, 128.62, 128.54, 83.26, 67.52, 54.81, 30.09, 28.34, 24.39.

³¹P NMR (CDCl₃): δ 37.61. MS (m/z, M + Na⁺): 380.

To the solution of the H-phosphinic acid obtained above (290

mg, 0.8 mmol) was added the solution of di-tert-butyl R-methyl-

eneglutarate (614 mg, 2.4 mmol) in 2 mL of freshly distilled CH₂-

Cl₂dried over MgSO₄through a syringe with a filter under argon,

followed by addition of BSA (812 mg, 4 mmol). The resulting

solution was heated at reflux temperature overnight and then

quenched with 2 N HCl (5 mL). The mixture was stirred vigorously

for 3 h. The organic layer was separated. The aqueous layer was

extracted with EtOAc (10 mL × 3). The organic solutions were

then combined and dried over Na₂SO₄. Removal of the solvent gave

970 mg of the crude product as a colorless oil, which was used in

the next step without further purification.

The product obtained above was dissolved in 5 mL of MeOH,

followed by addition of trimethylsilyldiazomethane (2.0 M in Et₂O,

1 mL, 2 mmol) until the color of the mixture remained green yellow.

The resulting mixture was then stirred at room temperature for

another hour. A drop of AcOH was added to quench the reaction.

Removal of the solvent gave 986 mg of the crude product, which

was purified by flash chromatography (EtOAc:hexanes ) 3:2, R_f

) 0.3) to yield 209 mg (42% over three steps) of 48b as a colorless

oil.¹H NMR (CDCl₃): δ 7.29-7.35 (m, 5H), 5.60-5.67 (m, 1H),

5.09 (s, 2H), 4.25-4.26 (m, 1H), 3.62-3.73 (m, 3H), 2.67 (m,

1H), 2.22-2.26 (m, 2H), 2.16-2.19 (m, 2H), 1.86-1.89 (m, 3H),

1.71-1.74 (m, 3H), 1.45 (s, 9H), 1.44 (s, 9H), 1.43 (s, 9H). ¹³C

NMR (CDCl₃): δ 173.77, 172.30, 172.26, 172.23, 171.01, 156.40,

136.67, 128.89, 128.53, 83.00, 82.96, 82.94, 81.80, 81.74, 81.01,

80.96, 67.34, 54.94, 51.62, 51.42, 39.79, 39.59, 33.02, 32.96, 30.54,

30.07, 29.81, 29.72, 29.36, 28.46, 28.38, 28.36, 25.51, 25.36, 25.11,

24.74, 24.48, 24.38, 24.01. ³¹P NMR (CDCl₃): δ 57.23, 57.12,

56.68, 56.62. ESI-HRMS (m/z): calcd for C₃₁H₅₀NO₁₀PNa [M +

Na]⁺650.3070, found 650.3075.

Dimethyl N-[(N-carbobenzyloxy-r-methyl)-L-glutamyl-γ]-[Ψ-

{P(O)(OCH₃)CH₂}]-glutarate (48a). To a solution of the methyl

vinylglycine 46a (360 mg, 1.44 mmol) in 10 mL of methanol were

added ammonium hypophosphite (300 mg, 3.6 mmol) and trieth-

ylborane (1 M in hexanes, 1.5 mL, 1.5 mmol). The resulting mixture

was stirred open to the air at room temperature for 2 h. The solvent

was then removed on a rotary evaporator while the temperature of

the water bath was kept lower that 30 °C. KHSO₄(20 mL, 1 M)

was added to the residue. The resulting aqueous mixture was then

extracted with EtOAc (50 mL × 3). The organic extracts were

combined and dried over Na₂SO₄. Removal of the solvent gave

430 mg of the desired phosphinic acid 47a, which was used in the

next step without purification. ¹H NMR (methanol-d₄): δ 6.69 (d,

J ) 546 Hz, 1H), 7.00 (m, 5H), 4.76 (s, 2H), 3.94-3.96 (m, 1H),

3.38 (s, 3H), 1.68-1.77 (m, 1H), 1.46-1.58 (m, 3H). ¹³C NMR

(methanol-d₄): δ 170.91, 155.85, 135.38, 126.84, 126.41, 126.18,

65.09, 52.91, 52.68, 50.29, 21.62. ³¹P NMR (CD₃OD): δ 33.36.

A solution of the phosphonic acid 47a (505 mg, 1.6 mmol)

obtained above and dimethyl R-methyleneglutarate (550 mg, 3.2

mmol) in 10 mL of fresh-distilled CH₂Cl₂was dried over MgSO₄

and then injected into a round-bottom flask through a syringe filter

under argon. BSA (1.02 g, 5 mmol) was added. The resulting

solution was stirred at room temperature overnight. The resulting

mixture was allowed to open to the air, followed by addition of 5

mL of 2 N HCl. The mixture was stirred vigorously for 2 h. The

organic layer was separated. The aqueous layer was extracted with

EtOAc (10 mL × 3). The organic solutions were then combined

and dried over Na₂SO₄. Removal of the solvent gave 1.2 g of a

colorless oil as the crude product, which was used in the next step

without further purification.

The product obtained above was dissolved in 15 mL of MeOH,

followed by addition of trimethylsilyldiazomethane (2.0 M in Et₂O,

2 mL, 4 mmol) until the color of the mixture remained green yellow.

The resulting mixture was then stirred at room temperature for

another hour. A drop of AcOH was added, resulting in loss of the

color. Removal of the solvent gave 1.3 mg of a yellow oil as the

crude product, which was purified with CombiFlash (MeOH:EtOAc

) 1:50, R_f) 0.2) to yield 600 mg (75% over 3 steps) of 48a as a

colorless oil. ¹H NMR (CDCl₃): δ 7.89-7.92 (m, 2H), [7.78 (d, J

) 6.8 Hz, 0.5H), 7.73 (d, J ) 6.9 Hz, 0.2H), 7.69 (d, J ) 7.0 Hz,

0.3H), amide N-H, 3 rotamers], 7.29-7.36 (m, 7H), 4.76-4.82

(m, 1H), 3.78-3.79 (m, 3H), 3.61-3.73 (m, 9H), 2.80-2.86 (m,

1H), 2.33-2.37 (m, 2H), 2.27-2.32 (m, 2H), 2.24-2.27 (m, 1H),

1.93-2.00 (m, 3H), 1.75-1.87 (m, 2H). ¹³C NMR (CDCl₃): δ

174.9, 174.8, 173.2, 173.1, 172.5, 172.4, 167.0, 155.4, 146.7, 136.7,

130.9, 128.9, 128.5, 128.4, 128.3, 125.4, 68.0, 53.04, 53.02, 52.62,

52.60, 52.2, 52.0, 51.92, 51.87, 51.7, 39.1, 38.9, 37.8, 31.54, 31.45,

29.2, 24.5. ³¹P NMR (CDCl₃): δ 57.42, 57.36, 56.8. ESI-HRMS

(m/z): calcd for C₂₂H₃₂NO₁₀PNa [M + Na]⁺524.1662, found

524.1666.

Dimethyl (r-methyl-L-glutamyl-γ)-[Ψ{P(O)(OMe)CH₂}]-glu-

tarate (49a). To a solution of the tetramethyl phosphinic pseudopep-

tide 48a (501 mg, 1 mmol) in 20 mL of freshly distilled THF was

added 100 mg of 5% Pd/C, and the mixture was saturated with H₂.

The resulting mixture was then stirred at room temperature under

H₂overnight. The Pd/C was filtered off. The filtrate was collected

and concentrated to give 370 mg of an oil as the crude product,

which was purified by flash chromatography (MeOH:EtOAc ) 1:4,

R_f) 0.4) to yield 348 mg (95%) of 49a as a colorless oil. ¹H NMR

(CDCl₃): δ 3.74 (s, 3H), 3.72 (s, 3H), 3.68 (s, 3H), 3.64-3.71 (m,

3H), 3.48-3.51 (m, 1H), 2.82-2.88 (m, 1H), 2.33-2.38 (m, 2H),

2.22-2.26 (m, 1H), 1.97-2.03 (m, 3H), 1.77-1.88 (m, 4H), 1.68

(s, br, 2H). ¹³C NMR (CDCl₃): δ 175.93, 175.04, 175.00, 173.32,

173.28, 54.89, 54.77, 52.58, 52.55, 52.15, 51.67, 51.61, 51.56,

51.51, 39.02, 38.83, 38.80, 31.66, 31.56, 30.71, 30.43, 29.99, 29.72,

29.23, 29.14, 27.36, 25.48, 25.14, 24.83, 24.75, 24.40. ³¹P NMR

(CDCl₃): δ 56.97, 56.56. ESI-HRMS (m/z): calcd for C₁₄H₂₆NO₈-

PNa [M + Na]⁺390.1294, found 390.1289.

Di-tert-butyl (r-tert-butyl-L-glutamyl-γ)-[Ψ{P(O)(OMe)CH₂}]-

glutarate (49b). This compound was obtained in 95% yield from

1

48b using the same procedure described above for 49a. H NMR

(CDCl₃): δ 4.05 (m, 1H), 3.68-3.77 (m, 3H), 2.66-2.72 (m, 1H),

2.37 (m, 2H), 2.18-2.28 (m, 4H), 2.11-2.13 (m, 1H), 1.84-1.91

(m, 3H), 1.51 (s, 9H), 1.46 (s, 9H), 1.44 (s, 9H). ¹³C NMR

(CDCl₃): δ 173.74, 173.68, 173.63, 172.40, 172.37, 172.35, 168.19,

84.58, 84.55, 81.76, 81.71, 81.63, 80.94, 80.92, 53.87, 52.50, 52.45,

52.34, 52.29, 52.20, 52.14, 52.09, 39.78, 39.50, 33.08, 33.04, 32.99,

30.53, 30.35, 30.09, 29.71, 29.60, 28.49, 28.42, 28.38, 28.20, 24.84,

24.38, 24.13, 23.69, 23.52. ³¹P NMR (CDCl₃): δ 57.74, 57.35. ESI-

HRMS (m/z): calcd for C₃₇H₅₇N₂O₁₅PNa [M + Na]⁺494.2883,

found 494.2888.

Di-tert-butyl (r-tert-butyl-N-carbobenzyloxy-L-glutamyl-γ)-

[Ψ{P(O)(OMe)CH₂}]-glutarate (48b). To a solution of the tert-

butyl vinylglycine 46b (122 mg, 0.4 mmol) in 2 mL of methanol

were added ammonium hypophosphite (91 mg, 1.1 mmol) and

triethylborane (1 M in hexanes, 0.5 mL, 0.5 mmol). The resulting

mixture was stirred open to the air at room temperature for 3 h.

The solvent was then removed on a rotary evaporator while the

temperature of the water bath was kept lower than 30 °C. To the

residue was added 10 mL of 1 M KHSO₄. The resulting aqueous

mixture was then extracted with EtOAc (20 mL × 3). The organic

extracts were combined and dried over Na₂SO₄. Removal of the

solvent gave 150 mg of the desired product 47b, which was used

Dimethyl {N-[p-(N-carbobenzyloxy-N-methylamino)benzoyl]-

r-methyl-L-glutamyl-γ}-[Ψ{P(O)(OMe)CH₂}]-glutarate (50a).

A solution of the free amine 49a (300 mg, 0.82 mmol) and p-(N-

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 785

carbobenzyloxy-N-methylamino)benzoic acid (285 mg, 1 mmol)

in 10 mL of fresh-distilled CH₂Cl₂was stirred at 0 °C under N₂

for 5 min, followed by addition of EDCl (192 mg, 1 mmol) in one

portion. The resulting mixture was then warmed to room temper-

ature and stirred for 2 h. The reaction mixture was put in an ice-

water bath and 2 mL of 2 N HCl was added to quench the reaction.

The resulting mixture was poured into 200 mL of EtOAc, washed

with NaHCO₃and brine, and dried over Na₂SO₄. Removal of the

solvent gave 610 mg of the crude product, which was purified by

silica gel column chromatography (2% MeOH in EtOAc, R_f) 0.4)

to yield 478 mg (92%) of 50a as a colorless oil. ¹H NMR

(CDCl₃): δ 7.89-7.92 (m, 2H), 7.78 [(d, J ) 6.84 Hz, 0.5 H),

7.73 (d, J ) 6.89 Hz, 0.2 H), 7.69 (d, J ) 7.00 Hz, 0.3 H), amide

N-H, three rotamers], 7.28-7.36 (m, 7H), 3.78-3.79 (m, 3H),

3.61-3.75 (m, 9H), 2.80-2.86 (m, 1H), 2.33-2.37 (m, 2H), 2.24-

2.32 (m, 2H), 2.10-2.22 (m, 1H), 1.93-2.00 (m, 3H), 1.75-1.87

(m, 2H). ¹³C NMR (CDCl₃): δ 174.88, 174.84, 173.22, 173.18,

172.48, 172.43, 166.97, 155.45, 146.69, 136.66, 130.90, 128.92,

128.51, 128.44, 128.32, 125.40, 68.02, 53.04, 53.02, 52.62, 52.57,

52.16, 51.92, 51.87, 51.74, 39.12, 38.88, 37.78, 31.54, 31.45, 29.29,

24.54. ³¹P NMR (CDCl₃): δ 57.42, 57.36, 56.84. ESI-HRMS (m/

z): calcd for C₃₀H₃₉N₂O₁₁PNa [M + Na]⁺657.2189, found

657.2204.

0.2) to afford 37 mg of a yellow solid. The yellow solid was then

stirred for 3 h with 40 mg of Dowex-H in MeOH. Removal of the

resin and solvent gave 25 mg (42% over three steps) of 1d as a

1

yellow solid. H NMR (MeOH-d₄): δ 8.58 (s, 1H), 7.75 (d, J )

14.7 Hz, 2H), 6.78 (d, J ) 14.7 Hz, 2H), 4.85 (s, 2H), 4.53-4.54

(m, 1H), 3.72 (s, 3H), 3.60-3.69 (m, 6H), 3.25 (s, 3H), 2.80 (m,

1H), 2.32-2.33 (m, 2H), 2.21 (m, 1H), 2.08 (m, 2H), 1.97-1.98

(m, 1H), 1.88-1.93 (m, 1H), 1.70 (m, 3H). ¹³C NMR (MeOH-d₄):

δ 174.95, 172.48, 171.81, 167.75, 162.16, 161.10, 152.37, 150.62,

147.70, 146.97, 127.87, 120.78, 119.70, 110.08, 54.03, 53.21, 52.68,

50.13, 49.83, 49.57, 41.18, 38.52, 37.17, 31.64, 29.80, 27.63, 27.55,

26.10, 25.38, 23.11. ³¹P NMR (MeOH-d₄): δ 42.48. ESI-HRMS

(m/z): calcd for C₂₈H₃₇N₈O₉PNa [M + Na]⁺683.2319, found

683.2328. UV λ_max: (phosphate buffer, pH 7.0) 225, 260, 310, 375

nm.AnalyticalHPLCmethodB: t_R)23.6min.Anal.(C₂₈H₃₇N₈O₉P‚3CH₃OH)

C, H, N.

N-{[4-(N-benzyloxycarbonyl-N-methyl)amino]benzoyl}-L-

glutamyl-γ}-[Ψ{P(O)(OH)CH₂}]-glutaric Acid (52). To a solu-

tion of tri-tert-butyl phosphinate 50b (22 mg, 0.03 mmol) in 0.5

mL of freshly distilled CH₂Cl₂was added TFA (0.1 mL, 1.2 mmol)

at 0 °C. The resulting solution was warmed to room temperature

and stirred overnight. Removal of the solvent in a vacuum gave

1

the free tetraacid as an oily solid (yield 17 mg, 100%). H NMR

(MeOH-d₄): δ 7.90 (d, J ) 8.4 Hz, 2H), 7.42 (d, J ) 8.4 Hz, 2H),

7.31-7.37 (m, 5H), 7.04 (d, J ) 8.1 Hz, 1H), 5.17 (s, 2H), 4.67

(dd, J₁) 8.4 Hz, J₁) 4.3 Hz, 1H), 3.35 (s, 3H), 2.82 (m, 1H),

2.38-2.43 (m, 2H), 2.28-2.36 (m, 2H), 2.13-2.14 (m, 1H), 1.90-

2.10 (m, 5H). ¹³C NMR (MeOH-d₄): δ 176.67, 175.46, 173.96,

173.80, 173.62, 168.89, 168.74, 155.86, 146.66, 136.72, 131.48,

129.56, 129.23, 128.58, 128.38, 128.22, 127.96, 125.46, 67.81,

53.82, 53.73, 53.61, 38.93, 36.89, 32.06, 31.18, 31.11, 30.30, 29.75,

28.93, 26.38, 25.66, 23.95. ³¹P NMR (MeOH-d₄): δ 53.39. MS

(m/z, M + Na⁺): 601.

Di-tert-butyl {N-{p-[(N-carbobenzyloxy-N-methyl)amino]ben-

zoyl}-r-tert-butyl-L-glutamyl-γ}-[Ψ{P(O)(OMe)CH₂}]-glut-

arate (50b). This compound was obtained in 92% yield from 49b

using the same procedure described above for 50a. ¹H NMR

(CDCl₃): δ 7.88 (d, J ) 7.87 Hz), 7.46-7.53 (m, 1H), 7.29-7.35

(m, 7H), 5.18 (s, 2H), 4.64-4.71 (m, 1H), 3.60-3.71 (m, 3H),

3.34 (s, 3H), 2.66 (m, 1H), 2.18-2.24 (m, 4H), 2.09-2.14 (m,

1H), 1.80-1.99 (m, 3H), 1.69-1.78 (m, 2H), 1.49 (s, 9H), 1.42

(s, 9H), 1.41-1.46 (m, 9H). ¹³C NMR (CDCl₃): δ 173.73, 172.24,

172.22, 172.19, 172.17, 171.17, 171.13, 166.91, 166.83, 155.44,

146.57, 136.66, 131.22, 128.91, 128.48, 128.32, 128.29, 125.40,

82.98, 82.96, 82.94, 81.80, 81.77, 81.75, 80.95, 67.99, 53.88, 53.84,

53.78, 53.74, 53.68, 53.59, 51.74, 51.69, 51.55, 51.50, 39.89, 39.82,

39.68, 39.65, 37.79, 33.00, 32.96, 32.91, 30.69, 30.30, 30.23, 29.96,

29.80, 29.70, 29.58, 29.51, 28.46, 28.41, 28.37, 25.21, 25.01, 24.94,

24.78, 24.49, 24.11, 24.05. ³¹P NMR (CDCl₃): δ 57.84, 57.72,

57.35, 57.26. ESI-HRMS (m/z): calcd for C₃₉H₅₇N₂O₁₁PNa [M +

Na]⁺783.3598, found 783.3617.

Dimethyl {N-[p-(N-methyl)aminobenzoyl]-r-methyl-L-glutamyl-

γ-[Ψ{P(O)(OCH₃)CH₂}]-glutarate (51). To a solution of the

tetramethyl phosphinic pseudopeptide 50a (60 mg, 0.09 mmol) in

5 mL of fresh-distilled THF was added 12 mg of 5% Pd/C, and

the mixture was saturated with H₂. The resulting mixture was then

stirred at room temperature under H₂overnight. The Pd/C was

filtered off. The filtrate was collected and concentrated to give 45

mg (100%) of 51 as an oil, which was used without further

purification. ¹H NMR (CDCl₃): δ 7.75 (d, J ) 12.2 Hz, 2H), 7.25-

7.34 (m, 1H), 6.57 (d, J ) 12.2 Hz, 2H), 4.79 (m, 1H), 4.36 (m,

1H), 3.77 (s, 3H), 3.61-3.72 (m, 9H), 2.86 (s, 3H), 2.31-2.33

(m, 2H), 2.27-2.29 (m, 2H), 2.23-2.25 (m, 1H), 1.93-1.98 (m,

3H), 1.82-1.87 (m, 2H). ¹³C NMR (CDCl₃): δ 174.91, 173.25,

172.85, 167.65, 152.61, 129.41, 121.59, 111.63, 52.93, 52.54, 52.13,

51.85, 39.00, 38.83, 31.50, 30.55, 29.73, 29.21, 29.05, 24.93. ³¹P

NMR (CDCl₃): δ 57.20, 56.71. ESI-HRMS (m/z): calcd for

C₂₂H₃₃N₂O₉PNa [M + Na]⁺523.1821, found 523.1824.

Methyl (2S)-2-N-[p-(N-carbobenzyloxy-N-methyl)aminoben-

zoyl]amino-4-phenylselenylbutyrate (54). To the suspension of

diphenyl diselenide (463 mg, 1 mmol) in 15 mL of EtOH was added

sodium borohydride (95 mg, 2.5 mmol) in portions under argon,

until the orange color of the mixture disappeared. The solution of

bromide 41 (1.19 g, 3.2 mmol) in 20 mL of EtOH was then added.

After 30 min, the resulting mixture was poured into 200 mL of

Et₂O. The resulting mixture was washed with NaHCO₃and brine

and dried over Na₂SO₄. Removal of the solvent gave 620 mg of

the crude product, which was purified by flash chromatography

(hexanes:EtOAc ) 1:9, R_f) 0.45) to give 483 mg (90%) of 54 as

1

a colorless oil. H NMR (CDCl₃): δ 7.79 (d, J ) 8.6 Hz, 2H),

7.49-7.50 (m, 2H), 7.31-7.38 (m, 7H), 7.24-7.27 (m, 3H), 7.07

(d, J ) 7.8 Hz, 1H), 5.20 (s, 2H), 4.91-4.95 (m, 1H), 3.75 (s,

3H), 3.35 (s, 3H), 2.94-3.01 (m, 2H), 2.35-2.40 (m, 1H), 2.18-

2.22 (m, 1H). ¹³C NMR (CDCl₃): δ 172.88, 166.90, 155.45, 146.70,

136.64, 133.22, 131.07, 130.04, 129.61, 128.97, 128.58, 128.35,

128.27, 127.59, 125.38, 68.09, 53.23, 53.03, 37.77, 33.40, 23.68.

MS (m/z, M + Na⁺): 563. Anal. (C₂₇H₂₈N₂O₅): C, H, N.

(2S)-2-N-[p-(N-carbobenzyloxy-N-methyl)aminobenzoyl]amino-

4-phenylselenylbutyric Acid (55). To a solution of the methyl ester

54 (390 mg, 0.72 mmol) in 4 mL of MeOH and 1 mL of CH₂Cl₂

at 0 °C was added 1 mL of 1 N LiOH. The resulting mixture was

stirred at room temperature for 2 h. Dowex-H was added to

neutralize the reaction mixture and was then removed by filtration.

The filtrate was collected and concentrated in vacuo to yield 380

mg (100%) of 55, which was used without further purification. ¹H

NMR (CDCl₃): δ 7.69 (d, J ) 7.0 Hz, 2H), 7.29-7.34 (m, 7H),

7.10-7.16 (m, 5H), 5.15 (s, 2H), 4.64 (m, 1H), 3.22 (s, 3H), 2.89

(m, 2H), 2.27 (m, 1H), 2.13 (m, 1H). ¹³C NMR (CDCl₃): δ 168.50,

155.53, 146.47, 136.56, 132.54, 130.82, 129.48, 128.95, 128.58,

128.52, 128.35, 127.11, 125.14, 68.13, 53.40, 37.57, 32.74, 4.07.

ESI-MS (m/z, [M-H⁺+ Na⁺]): 548.9.

Dimethyl {N-[(4-deoxy-2,4-diamino-10-methyl)pteroyl]-r-

methyl-L-glutamyl-γ}-[Ψ{P(O)(OH)CH₂}]-glutarate (1d). Di-

bromotriphenylphosphine (127 mg, 0.3 mmol) was suspended in 1

mL of dry dimethylacetamide, followed by addition of (2,4-diamino-

6-pteridinyl)methanol hydrochloride (11a) (23 mg, 0.1 mmol). The

resulting mixture turned clear and red in few minutes with emission

of gas. The reddish solution was stirred at room temperature for

20 h before diethylisopropylamine (52 mg, 0.4 mmol) and a solution

of the tetramethyl ester 51 (45 mg, 0.09 mmol) in dimethylaceta-

mide (1 mL) was added. The resulting mixture was then kept at 60

°C for 40 h and monitored by HPLC. The solvent was removed in

a vacuum to give 240 mg of the crude product, which was purified

by silica gel column chromatography (MeOH:CH₂Cl₂) 1:1, R_f)

Pivaloyloxymethyl (2S)-2-N-[p-(N-carbobenzyloxy-N-methyl)-

aminobenzoyl]amino-4-phenylselenylbutyrate (56). To a solution

of the acid 55 (273 mg, 0.52 mmol) in 5 mL of dry DMF was

added POMI (630 mg, 2.6 mmol), dimethyl methylphosphonate

(322 mg, 2.6 mmol), and Ag₂CO₃(360 mg, 1.3 mmol). The

786 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

resulting mixture was allowed to stir at room temperature overnight

and quenched with 2 mL of 2 N HCl. The precipitate was removed

by filtration. The filtrate was collected and concentrated to give

483 mg of the crude product, which was purified by flash

chromatography (hexanes:EtOAc ) 2:1, R_f) 0.45) to yield 276

mg (83%) of 56 as a colorless oil. ¹H NMR (CDCl₃): δ 7.78 (d, J

) 8.6 Hz, 2H), 7.50-7.52 (m, 2H), 7.32-7.39 (m, 7H), 7.26-

7.29 (m, 3H), 6.78 (d, J ) 7.7 Hz, 1H), 5.89 (d, J ) 5.5 Hz, 1H),

5.78 (d, J ) 5.5 Hz, 1H), 5.21 (s, 2H), 4.95-4.99 (m, 1H), 3.38

(s, 3H), 2.97 (m, 2H), 2.34-2.41 (m, 1H), 2.16-2.24 (m, 1H),

1.22 (s, 9H). ¹³C NMR (CDCl₃): δ 177.30, 171.19, 166.82, 155.44,

146.89, 136.61, 133.49, 130.89, 129.64, 128.96, 128.58, 128.37,

128.19, 127.76, 125.46, 80.43, 68.11, 53.12, 39.16, 37.76, 33.21,

27.24, 23.46. MS (m/z, M + Na⁺): 663. Anal. (C₃₂H₃₆N₂O₇Se):

C, H, N.

172.89, 171.34, 171.30, 170.73, 170.67, 167.12, 167.05, 155.42,

146.72, 136.66, 130.73, 128.90, 128.48, 128.28, 125.34, 80.49,

80.31, 80.19, 79.98, 67.99, 53.38, 53.13, 53.05, 52.93, 52.03, 51.82,

39.23, 39.11, 38.79, 38.66, 37.74, 31.20, 31.15, 31.11, 30.48, 30.11,

29.73, 29.40, 28.59, 28.51, 28.07, 27.61, 27.22, 25.31, 24.87, 24.58,

24.28, 24.14, 23.95, 23.77. ³¹P NMR (CDCl₃): δ 57.18, 57.04,

56.61. ESI-HRMS (m/z): calcd for C₄₅H₆₃N₂O₁₇PNa [M + Na]⁺

957.3762, found. 957.3786.

Bis(pivaloyloxymethyl) {N-[p-(N-methyl)aminobenzoyl]-r-

pivaloyloxymethyl-L-glutamyl-γ}-[Ψ{P(O)(OCH₃)CH₂}]-glu-

tarate (60). To a solution of the tri-POM ester 59 (56 mg, 0.06

mmol) in 10 mL of freshly distilled THF was added 12 mg of Pd/

C. The resulting mixture was then stirred overnight under H₂at

room temperature, after which the Pd/C was removed by filtration.

The filtrate was collected and concentrated to give the 40 mg of

the crude product, which was purified by the preparative TLC

(EtOAc:hexanes ) 4:1, R_f) 0.6) to yield 35 mg (73%) of 60 as a

colorless oil. ¹H NMR (CDCl₃): 7.75 (d, J ) 8.4 Hz, 2H), 7.37-

7.42 (m, 1H), 6.58 (d, J ) 8.4 Hz, 2H), 5.81-5.88 (m, 3H), 5.68-

5.80 (m, 5H), 4.77-4.81 (m, 1H), 3.63-3.74 (m, 3H), 2.87 (s,

3H), 2.87 (m, 1H), 2.35-2.42 (m, 2H), 2.22-2.27 (m, 2H), 2.08-

2.14 (m, 1H), 1.90-2.00 (m, 3H), 1.77-1.88 (m, 2H), 1.19-1.21

(m, 27H). ¹³C NMR (CDCl₃): δ 177.48, 177.41, 177.37, 173.09,

173.04, 173.01, 172.96, 171.39, 171.34, 171.12, 171.10, 167.79,

167.76, 167.71, 152.66, 152.62, 129.50, 121.57, 121.54, 121.49,

111.73, 80.54, 80.48, 80.33, 80.30, 80.25, 79.99, 53.13, 52.98,

52.93, 52.84, 52.01, 51.96, 51.83, 51.79, 51.73, 39.12, 38.77, 38.71,

38.61, 31.23, 31.17, 31.14, 30.59, 30.43, 30.07, 29.71, 29.28, 28.60,

Pivaloyloxymethyl (2S)-N-[p-(N-carbobenzyloxy-N-methyl)-

aminobenzoyl]-2-vinylglycine (57). To a solution of the POM ester

56 (276 mg, 0.43 mmol) in 25 mL of freshly distilled THF was

added 0.5 mL of H₂O₂(30% in H₂O). The resulting mixture was

heated at reflux temperature for 1 h, diluted with 200 mL of EtOAc,

washed with brine, and then dried over Na₂SO₄. Removal of the

solvent gave 210 mg of the crude product, which was purified by

flash chromatography (hexanes:EtOAc ) 2:1, R_f) 0.4) to yield

1

187 mg (90%)of 57 as a colorless oil. H NMR (CDCl₃): δ 7.82

(d, J ) 8.6 Hz, 2H), 7.29-7.37 (m, 7H), 6.92 (d, J ) 7.4 Hz, 1H),

5.95-6.01 (m, 1H), 5.91 (d, J ) 5.5 Hz, 1H), 5.80 (d, J ) 5.5 Hz,

1H), 5.44 (dd, J₁)17.1 Hz, J₂)1.4 Hz, 1H), 5.36-5.38 (m, 1H),

5.34 (dd, J₁)10.4 Hz, J₂)1.4 Hz, 1H), 5.19 (s, 2H); 3.36 (s, 3H),

1.22 (s, 9H). ¹³C NMR (CDCl₃): δ 177.28, 169.89, 166.57, 155.45,

146.85, 136.60, 131.80, 130.84, 128.94, 128.56, 128.34, 128.27,

125.44, 119.14, 80.33, 68.08, 55.26, 39.18, 37.74, 27.23. ESI-

HRMS (m/z): calcd for C₂₆H₃₀N₂O₇Na [M + Na]⁺505.1951, found

505.1950.

28.51, 27.61, 27.23, 24.57, 24.47, 24.28.

³¹P NMR (CDCl₃): δ

57.00, 56.89, 56.47. ESI-HRMS (m/z): calcd for C₃₇H₅₇N₂O₁₅PNa

[M + Na]⁺823.3394, found 823.3395.

Bis(pivaloyloxymethyl) {N-[(4-deoxy-2,4-diamino-10-methyl)-

pteroyl]-r-pivaloyloxymethyl-L-glutamyl-γ}-[Ψ{P(O)(OCH₃)-

CH₂}]-glutarate (1b). To the suspension of dibromotriphenylphos-

phine (63 mg, 0.15 mmol) in 0.5 mL of N,N-dimethylacetamide

was added (2,4-diamino-6-pteridinyl)methanol hydrochloride (11a)

(12 mg, 0.05 mmol). The resulting mixture turned to a red

homogeneous solution in a few minutes with emission of gas. The

reddish solution was stirred at room temperature for 18 h before

diisopropylethylamine (26 mg, 0.2 mmol) was added. To the

resulting mixture was added a solution of the POM ester 60 (40

mg, 0.05 mmol) in dimethylacetamide (1 mL). The resulting mixture

was then kept at 60 °C for 2 days and reaction progress was

monitored by analytical HPLC. After the reaction was completed,

the solvent was removed in a vacuum. The residue was dissolved

in 100 mL of EtOAc and washed with 1 N HCl (5 mL × 3) and

brine (5 mL × 3) and then dried over Na₂SO₄. Removal of the

solvent gave 116 mg of a brown oily solid as the crude product,

which was dissolved in 5 mL of EtOAc and treated with CH₂N₂

(∼1 mmol). Concentration of the reaction mixture gave 120 mg of

the crude product, which was purified by semipreparative HPLC

to yield 20 mg (41%) of 1b as a yellow oil. ¹H NMR (DMSO-d₆):

δ 8.56 (s, 1H), 8.47, (d, J ) 7.0 Hz), 7.72 (d, J ) 8.5 Hz, 2H),

7.72 (bs, 1H), 7.48 (bs, 1H), 6.82 (d, J ) 8.5 Hz, 2H), 6.67 (bs,

2H), 5.65-5.76 (m, 6H), 4.78 (s, 2H), 4.37-4.38 (m, 1H), 3.52

(d, J ) 10.5 Hz, 3H), 3.21 (s, 3H), 2.75 (m, 1H), 2.34-2.40 (m,

2H), 2.04-2.10 (m, 1H), 1.85-1.91 (m, 4H), 1.78-1.85 (m, 3H),

1.17 (s, 9H), 1.12 (s, 9H), 1.10 (s, 9H). ³¹P NMR (DMSO-d₆): δ

55.80, 55.76, 55.63, 55.56. ESI-HRMS (m/z): calcd for C₄₄H₆₃N₈O₁₅-

PNa [M + Na]⁺997.4048, found 997.4053. UV λ_max: (0.1 N

NaOH) 255, 306, 370 nm; (0.1 N HCl) 245, 308. Analytical HPLC

method B: t_R) 25.8 min.

Bis(pivaloyloxymethyl) {N-{p-[(N-carbobenzyloxy-N-methyl)-

amino]benzoyl}-r-pivaloyloxymethyl-L-glutamyl-γ}-[Ψ{P(O)-

(OCH₃)CH₂}]-glutarate (59). To a solution of the POM ester 57

(73 mg, 0.15 mmol) and ammonium hypophosphite (38 mg, 0.45

mmol) in 1 mL of MeOH was added triethylborane (1 M in hexanes,

0.18 mL, 0.18 mmol). The resulting mixture was stirred open to

the air at room temperature for 3 h. The solvent was then removed

on a rotary evaporator while the temperature of the water bath was

kept lower than 30 °C. KHSO₄(2 mL, 1 M) was added to the

residue. The resulting aqueous mixture was extracted with EtOAc

(10 mL × 3). The organic extracts were combined and dried over

Na₂SO₄. Removal of the solvent gave 82 mg of the desired product

58, which was used in the next step without further purification.

To the solution of the phosphinic 58 obtained above in 1 mL of

freshly distilled CH₂Cl₂was added the solution of bis(pivaloyl-

oxymethyl) R-methyleneglutarate 34 (140 mg, 0.38 mmol) in 1 mL

of freshly distilled CH₂Cl₂dried over MgSO₄through a syringe

with a filter under argon, followed by addition of BSA (153 mg,

0.75 mmol). The resulting solution was heated at reflux temperature

overnight and then quenched with 2 N HCl (2 mL). The mixture

was stirred vigorously for 3 h and diluted with 50 mL of EtOAc

(10 mL × 3). The resulting mixture was washed with brine and

dried over Na₂SO₄. Removal of the solvent gave 61 mg of the crude

product as a light yellow oil, which was used in the next step

without further purification.

The product obtained above was dissolved in 5 mL of EtOAc,

followed by addition of the solution of CH₂N₂in Et₂O until the

color of the mixture remained green yellow. The resulting mixture

was then stirred at room temperature for another hour. A drop of

AcOH was added to quench the reaction. Removal of the solvent

gave 70 mg of the crude product, which was purified by flash

chromatography (EtOAc:hexanes ) 2:1, R_f) 0.4) to give 65 mg

Bis(pivaloyloxymethyl) {N-[(4-deoxy-2,4-diamino-10-methyl)-

pteroyl]-r-pivaloyloxymethyl-L-glutamyl-γ}-[Ψ{P(O)(OH)CH₂}]-

glutarate (1c). To a suspension of dibromotriphenylphosphine (76

mg, 0.18 mmol) in 0.5 mL of N,N-dimethylacetamide was added

(2,4-diamino-6-pteridinyl)methanol hydrochloride (11a) (14 mg,

0.06 mmol). The resulting mixture turned to a red homogeneous

solution in a few minutes with emission of gas. The reddish solution

was stirred at room temperature for 18 h before diisopropylethyl-

amine (30 mg, 0.23 mmol) was added. To the resulting mixture

1

(46% over 3 steps) of 59 as a colorless oil. H NMR (CDCl₃): δ

7.93-8.00 (m, 1H), 7.91 (d, J ) 8.4 Hz, 2H), 7.29-7.36 (m, 7H),

5.61-5.89 (m, 6H), 5.17 (s, 2H), 4.75-4.79 (m, 1H), 3.62-3.73

(m, 3H), 3.34 (s, 3H), 2.86-2.87 (m, 1H), 2.35-2.40 (m, 2H),

2.22-2.27 (m, 2H), 2.18-2.21 (m, 1H), 1.97-2.03 (m, 3H), 1.82-

1.88 (m, 2H). ¹³C NMR (CDCl₃): δ 177.42, 177.36, 177.33, 172.94,

Potent Inhibitor of Folylpoly-γ-glutamate Synthetase

Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2 787

(8) McGuire, J. J.; Coward, J. K. Pteroylpolyglutamates: biosynthesis,

degradation and function. In Folates and Pterins Biochemistry and

Chemistry of Folates; Blakley, R. L., Benkovic, S. J., Eds.; Wiley:

New York, 1984; pp 135-190.

(9) Zeng, H.; Chen, Z. S.; Belinsky, M. G.; Rea, P. A.; Kruh, G. D.

Transport of methotrexate (MTX) and folates by multidrug resistance

protein (MRP) 3 and MRP1: Effect of polyglutamylation on MTX

transport. Cancer Res. 2001, 61, 7225-7232.

(10) McBurney, M. W.; Whitmore, G. F. Isolation and biochemical

characterization of folate deficient mutants of Chinese-hamster cells.

Cell 1974, 2, 173-182.

(11) McGuire, J. J.; Coward, J. K. Folylpolyglutamate synthetase as a

target for therapeutic intervention. Drugs Future 2003, 28, 967-

974.

(12) Liani, E.; Rothem, L.; Bunni, M. A.; Smith, C. A.; Jansen, G.;

Assaraf, Y. G. Loss of folylpoly-γ-glutamate synthetase activity is a

dominant mechanism of resistance to polyglutamylation-dependent

novel antifolates in multiple human leukemia sublines. Int. J. Cancer

2003, 103, 587-599.

(13) Banerjee, R. V.; Shane, B.; McGuire, J. J.; Coward, J. K. Dihydro-

folate synthetase and folylpolyglutamate synthetasesdirect evidence

for intervention of acyl phosphate intermediates. Biochemistry 1988,

27, 9062-9070.

(14) Valiaeva, N.; Bartley, D.; Konno, T.; Coward, J. K. Phosphinic acid

pseudopeptides analogous to glutamyl-γ-glutamate: synthesis and

coupling to pteroyl azides leads to potent inhibitors of folylpoly-γ-

glutamate synthetase. J. Org. Chem. 2001, 66, 5146-5154.

(15) McGuire, J. J.; Haile, W. H.; Valiaeva, N.; Bartley, D.; Guo, J. X.;

Coward, J. K. Potent inhibition of human folylpolyglutamate syn-

thetase by a phosphinic acid mimic of the tetrahedral reaction

intermediate. Biochem. Pharmacol. 2003, 65, 315-318.

(16) Bartley, D. M.; Coward, J. K. A stereoselective synthesis of

phosphinic acid phosphapeptides corresponding to glutamyl-γ-

glutamate and incorporation into potent inhibitors of folylpoly-γ-

glutamyl synthetase. J. Org. Chem. 2005, 6757-6774.

(17) Matherly, L. H.; Goldman, D. Membrane transport of folates. Vitam.

Horm. 2003, 66, 403-456.

(18) Stella, V. J.; Himmelstein, K. J. Prodrugs and site-specific drug

delivery. J. Med. Chem. 1980, 23, 1276-1282.

(19) Sakamoto, F.; Ikeda, S.; Tsukamoto, G. Studies on prodrugs. 1.

Preparation and characterization of acyloxyallylester of ampicillin.

Chem. Pharm. Bull. 1983, 31, 2698-2707.

(20) Sakamoto, F.; Ikeda, S.; Tsukamoto, G. Studies on prodrugs. 2.

Preparation and characterization of (5-substituted 2-oxo-1,3-dioxo-

len-4-yl)methyl esters of ampicillin. Chem. Pharm. Bull. 1984, 32,

2241-2248.

(21) Ikeda, S.; Sakamoto, F.; Kondo, H.; Moriyama, M.; Tsukamoto, G.

Studies on prodrugs. 3. A convenient and practical preparation of

ampicillin prodrugs. Chem. Pharm. Bull. 1984, 32, 4316-4322.

(22) Jarvinen, T.; Jarvinen, K. Prodrugs for improved ocular drug delivery.

AdV. Drug DeliV. ReV. 1996, 19, 203-224.

(23) Mori, N.; Kodama, T.; Sakai, A.; Suzuki, T.; Sugihara, T.; Yamagu-

chi, S.; Nishijima, T.; Aoki, A.; Toriya, M.; Kasai, M.; Hatano, S.;

Kitagawa, M.; Yoshimi, A.; Nishimura, K. AS-924, a novel, orally

active, bifunctional prodrug of ceftizoxime: Physicochemical proper-

ties, oral absorption in animals, and antibacterial activity. Int. J.

Antimicrob. Ag. 2001, 18, 451-461.

(24) Testa, B.; Mayer, J. M. Hydrolysis in drug and prodrug metabolism:

chemistry, biochemistry, and enzymology; Wiley-VCH: Weinheim,

2003.

(25) Beaumont, K.; Webster, R.; Gardner, I.; Dack, K. Design of ester

prodrugs to enhance oral absorption of poorly permeable com-

pounds: challenges to the discovery scientist. Curr. Drug. Metab.

2003, 4, 461-485.

was added a solution of the POM ester 60 (32 mg, 0.04 mmol) in

dimethylacetamide (1 mL). The resulting mixture was then kept at

60 °C for 2 days and reaction progress was monitored by analytical

HPLC. After the reaction was completed, the solvent was removed

in a vacuum. The residue was dissolved in 100 mL of EtOAc and

washed with 1 N HCl (5 mL × 3) and brine (5 mL × 3) and then

dried over Na₂SO₄. Removal of the solvent gave 134 mg of a brown

oily solid as the crude product, which was purified by preparative

HPLC (method B) to yield 12 mg (31%) of 1c as a yellow solid.

¹H NMR (DMSO-d₆): δ 9.28 (s, 1H), 9.08 (s, 1H), 8.72 (s, 1H),

8.64, (d, J ) 6.5 Hz), 7.75 (d, J ) 9.0 Hz, 2H), 7.56 (bs, 1H), 7.36

(bs, 1H), 6.82 (d, J ) 9.0 Hz, 2H), 5.63-5.76 (m, 6H), 4.88 (s,

2H), 4.38 (m, 1H), 3.25 (s, 3H), 2.75 (m, 1H), 2.36 (m, 2H), 1.78-

1.93 (m, 5H), 1.63-1.72 (m, 3H), 1.19 (s, 9H), 1.16 (s, 9H), 1.13

(s, 9H). ³¹P NMR (DMSO-d₆): δ 47.11. ESI-HRMS (m/z): calcd

for C₄₃H₆₁N₈O₁₅PNa [M + Na]⁺983.3892, found 983.3054. UV

λ_max: (0.1 N NaOH) 255, 306, 370 nm; (0.1 N HCl) 245, 308.

Analytical HPLC method B: t_R) 23.5 min.

Cytotoxicity and Hydrolysis Studies. Compounds 1a-d and

2a-d were dissolved in DMSO to provide 1 mM stock solutions.

Serial dilutions were done either in DMSO (cell culture cytotoxicity)

or phosphate (20 mM, pH 8.0) (hydrolysis studies). Cells were

grown in RPMI 1640 cell culture medium (Sigma) as previously

described³⁹and exposed continuously to compounds for a period

of 72 h. Cytotoxicity was determined by use of a Cell Titer-Blue

Cell Viability kit (Promega), and IC₅₀data (Table 2) were obtained

from dose-response curves obtained over a concentration range

from 1 nM to 1 µM. For hydrolysis studies (Figures 3-5), the t )

0 points indicate the time a sample was applied to the HPLC column

by the autosampler following dilution of the stock solution with

phosphate buffer or cell line medium and filtration of the sample.

Acknowledgment. This research was supported by a grant

from the National Cancer Institute (CA28097). We thank Prof.

Larry H. Matherly and Ms. Sarah Stapels, Karamanos Cancer

Center and Department of Pharmacology, Wayne State Uni-

versity, Detroit, MI, for their efforts in carrying out the cell

cytotoxicity studies of 1a-d and 2a-d. We also thank Dr.

David Bartley for his helpful discussions of synthetic organo-

phosphorus chemistry, Dr. John Tomsho for his helpful dis-

cussions and advice on HPLC analysis, and Dr. Kristin Meagher

for calculation of log P values for the prodrugs investigated in

this research.

Supporting Information Available: Experimental procedures

and characterization of compound 46b; combustion analysis results

1

for compounds 1d, 2b, and 2d; H NMR spectral data for 1b-d,

2b, and 2d; ¹³C NMR spectral data for 1d, 2b, and 2d; and ³¹P

NMR spectral data for 1b-d. This material is available free of

charge via the Internet at http://pubs.acs.org.

References

(1) Shane, B. Folylpolyglutamate synthesis and role in the regulation of

one-carbon metabolism. Vitam. Horm. 1989, 45, 263-335.

(2) Longo-Sorbello, G. S. A.; Bertino, J. R. Current understanding of

methotrexate pharmacology and efficacy in acute leukemias. Use of

newer antifolates in clinical trials. Haematologica 2001, 86, 121-

127.

(3) Rosowsky, A. Progress in medicinal chemistry; Elsevier Sciences

Publisher: Amsterdam, 1989; pp 1-252.

(4) McGuire, J. J. Anticancer antifolates: Current status and future

directions. Curr. Pharm. Design 2003, 9, 2593-2613.

(5) Kompis, I. M.; Islam, K.; Then, R. L. DNA and RNA synthesis:

Antifolates. Chem. ReV. 2005, 105, 593-620.

(6) Markham, A.; Faulds, D. Methotrexatesa review of its pharmaco-

dynamic and pharmacokinetic properties, and therapeutic efficacy

in rheumatoid-arthritis and other immunoregulatory disorders. Clin.

Immunother. 1994, 1, 217-244.

(7) Cossins, E. A. Folates in biological materials. In Folates and Pterins

Biochemistry and Chemistry of Folates; Blakley, R. L., Benkovic,

S. J., Eds.; Wiley: New York, 1984; pp 1-59.

(26) Jusko, W. J.; Lewis, G. P. Comparison of ampicillin and hetacillin

pharmacokinetics in man. J. Pharm. Sci. 1973, 62, 69-76.

(27) Ehrnebo, M.; Nilsson, S. O.; Boreus, L. O. Pharmacokinetics of

ampicillin and its prodrugs bacampicillin and pivampicillin in man.

J. Pharmacokinet. Biopharm. 1979, 7, 429-451.

(28) Josefsson, K.; Bergan, T.; Magni, L.; Pring, B. G.; Westerlund, D.

Pharmacokinetics of bacmecillinam and pivmecillinam in volunteers.

Eur. J. Clin. Pharmacol. 1982, 23, 249-252.

(29) Klepser, M. E.; Marangos, M. N.; Patel, K. B.; Nicolau, D. P.;

Quintiliani, R.; Nightingale, C. H. Clinical pharmacokinetics of newer

cephalosporins. Clin. Pharmacokinet. 1995, 28, 361-384.

(30) Barditch-Crovo, P.; Toole, J.; Hendrix, C. W.; Cundy, K. C.; Ebeling,

D.; Jaffe, H. S.; Lietman, P. S. Anti-human immunodeficiency virus

(HIV) activity, safety, and pharmacokinetics of adefovir dipivoxil

(9-2-(bis-pivaloyloxymethyl)-phosphonylmethoxyethyl adenine) in

HIV-infected patients. J. Infect. Dis. 1997, 176, 406-413.

(31) Cundy, K. C. Clinical pharmacokinetics of the antiviral nucleotide

analogues cidofovir and adefovir. Clin. Pharmacokinet. 1999, 36,

127-143.

788 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 2

Feng and Coward

(32) Patel, D. V.; Gordon, E. M.; Schmidt, R. J.; Weller, H. N.; Young,

M. G.; Zahler, R.; Barbacid, M.; Carboni, J. M.; Gullobrown, J. L.;

Hunihan, L.; Ricca, C.; Robinson, S.; Seizinger, B. R.; Tuomari, A.

V.; Manne, V. Phosphinyl acid-based bisubstrate analogue inhibitors

of ras farnesyl-protein transferase. J. Med. Chem. 1995, 38, 435-

442.

(33) Estlin, E. J. Continuing therapy for childhood acute lymphoblastic

leukaemia: Clinical and cellular pharmacology of methotrexate,

6-mercaptopurine and 6-thioguanine. Cancer Treat. ReV. 2001, 27,

351-363.

(34) Galivan, J.; Balinska, M. Studies of formation and efflux of

methotrexate polygkutamates with cultured hepatic cells. In Folyl

and Antifolyl Polyglutamates, Goldman, I. D.; Chabner, B. A. Bertino,

J. R., Eds; Plenum Press: New York, 1983; pp 235-246.

(35) Hoffbran, Av.; Tripp, E.; Houlihan, C. M.; Scott, J. M. Studies on

uptake of synthetic conjugated folates by human marrow cells. Blood

1973, 42, 141-146.

(36) O’Connor, B. M.; Rotundo, R. F.; Nimec, Z.; McGuire, J. J.; Galivan,

J. Secretion of γ-glutamyl hydrolase in vitro. Cancer Res. 1991, 51,

3874-3881.

(37) Galivan, J.; Ryan, T. J.; Chave, K.; Rhee, M.; Yao, R.; Yin, D. Z.

Glutamyl hydrolase: Pharmacological role and enzymatic charac-

terization. Pharmacol. Ther. 2000, 85, 207-215.

(38) Flintoff, W. F.; Nagainis, C. R. Transport of methotrexate in Chinese-

hamster ovary cellssa mutant defective in methotrexate uptake and

cell binding. Arch. Biochem. Biophys. 1983, 223, 433-440.

(39) Wong, S. C.; Proefke, S. A.; Bhushan, A.; Matherly, L. H. Isolation

of human cDNAs that restore methotrexate sensitivity and reduced

folate carrier activity in methotrexate transport-defective Chinese-

hamster ovary cells. J. Biol. Chem. 1995, 270, 17468-17475.

(40) Chaykovsky, M.; Brown, B. L.; Modest, E. J. Methotrexate analogues.

6. Replacement of glutamic-acid by various amino-acid esters and

amines. J. Med. Chem. 1975, 18, 909-912.

(41) Tomsho, J. W.; McGuire, J. J.; Coward, J. K. Synthesis of (6R)-

and (6S)-5,10-dideazatetrahydrofolate oligo-γ-glutamate: kinetics of

multiple glutamate ligations catalyzed by folylpoly-γ-glutamate

synthetase. Org. Biomol. Chem. 2005, 3, 3388-3398.

(42) Kuo, C. H.; Plevyak, S. P.; Biftu, T.; Parsons, W. H.; Berger, G. D.

The synthesis of C3-methyl, C3-decarboxy-zaragozic acid-AsA

potent squalene synthase inhibitor. Tetrahedron Lett. 1993, 34, 6863-

6866.

(43) Knochel, P.; Chou, T. S.; Jubert, C.; Rajagopal, D. Preparation and

reactivity of highly functionalized organometallics at the alpha psition

of oxygen or nitrogen. J. Org. Chem. 1993, 58, 588-599.

(44) Coward, J. K.; Parameswaran, K. N.; Cashmore, A. R.; Bertino, J.

R. 7,8-Dihydropteroyl oligo-γ-^L-glutamatesssynthesis and kinetic

studies with purified dihydrofolate-reductase from mammalian

sources. Biochemistry 1974, 13, 3899-3903.

(45) Thottathil, J. K.; Przybyla, C. A.; Moniot, J. L. Mild arbuzov reactions

of phosphonous acids. Tetrahedron Lett. 1984, 25, 4737-4740.

(46) Malachowski, W. P.; Coward, J. K. The Chemistry of Phospha-

peptidessinvestigations on the synthesis of phosphonamidate, phos-

phonate, and phosphinate analogues of glutamyl-γ-glutamate. J. Org.

Chem. 1994, 59, 7625-7634.

(47) Thottathil, J. K.; Ryono, D. E.; Przybyla, C. A.; Moniot, J. L.;

Neubeck, R. Preparation of phosphinic acidssmichael additions of

phosphonous acids esters to conjugated systems. Tetrahedron Lett.

1984, 25, 4741-4744.

(48) Boyd, E. A.; Corless, M.; James, K.; Regan, A. C. A versatile route

to substituted phosphinic acids. Tetrahedron Lett. 1990, 31, 2933-

2936.

(54) Berlicki, L.; Obojska, A.; Forlani, G.; Kafarski, P. Design, synthesis,

and activity of analogues of phosphinothricin as inhibitors of

glutamine synthetase. J. Med. Chem. 2005, 48, 6340-6349.

(55) Hashimoto, N.; Aoyama, T.; Shioiri, T. New methods and reagents

in organic-synthesis. 14. A simple efficient preparation of methyl-

esters with trimethylsilyldiazomethane (TMSCH₂N₂) and its applica-

tion to gas-chromatographic analysis of fatty-acids. Chem. Pharm.

Bull. 1981, 29, 1475-1478.

(56) Krawczyk, H. A convenient route for monodealkylation of diethyl

phosphonates. Synthetic Commun. 1997, 27, 3151-3161.

(57) Lherbet, C.; Castonguay, R.; Keillor, J. W. Transesterification of

trialkyl phosphates from alkyl bromides. Tetrahedron Lett. 2005, 46,

3565-3567.

(58) Muller, B.; Martin, T. J.; Schaub, C.; Schmidt, R. R. Synthesis of

phosphonate analogues of CMP-Neu5Ac determination of R(2-6)-

sialyltransferase inhibition. Tetrahedron Lett. 1998, 39, 509-512.

(59) Tsukamoto, T.; Haile, W. H.; McGuire, J. J.; Coward, J. K.

Mechanism-based inhibition of human folylpolyglutamate syn-

thetase: Design, synthesis, and biochemical characterization of a

phosphapeptide mimic of the tetrahedral intermediate. Arch. Biochem.

Biophys. 1998, 355, 109-118.

(60) Decicco, C. P.; Nelson, D. J.; Luo, Y.; Shen, L.; Horiuchi, K. Y.;

Amsler, K. M.; Foster, L. A.; Spitz, S. M.; Merrill, J. J.; Sizemore,

C. F.; Rogers, K. C.; Copeland, R. A.; Harpel, M. R. Glutamyl-γ-

boronate inhibitors of bacterial Glu-tRNAGln amidotransferase.

Bioorg. Med. Chem. Lett. 2001, 11, 2561-2564.

(61) Chowdhury, A. R.; Boons, G.-J. The synthesis of diaminopimelic

acid containing peptidoglycan fragments using metathesis cross

coupling. Tetrahedron Lett. 2005, 46, 1675-1678.

(62) Williams, R. M.; Sabol, M. R.; Kim, H. D.; Kwast, A. Novel ring

contractions via 2,3 Wittig type rearrangementsssynthesis of 2-deoxy-

2-methylenebicyclomycin. J. Am. Chem. Soc. 1991, 113, 6621-6633.

(63) The fact that the tris-POM ester, 2d, is slightly more cytotoxic in a

transport-deficient (R2) cell line than the free acid, 2c, may indicate

that the prodrug ester enters the cell more effectively than the parent

isopeptide. Subsequent intracellular hydrolysis would then occur as

described.

(64) Saito, T.; Iwazawa, H.; Nakajima, S.; Fujita, H. Antitumor metho-

trexate diesters for oral administration JP 63227588; Chem. Abstr.

1989, 110, 154322.

(65) Jansen, A. B. A.; Russell, T. J. Some novel penicillin derivatives. J.

Chem. Soc. 1965, 2127-2132.

(66) Manne, V.; Yan, N.; Carboni, J. M.; Tuomari, A. V.; Ricca, C. S.;

Brown, J. G.; Andahazy, M. L.; Schmidt, R. J.; Patel, D.; Zahler,

R.; Weinmann, R.; Der, C. J.; Cox, A. D.; Hunt, J. T.; Gordon, E.

M.; Barbacid, M.; Seizinger, B. R. Bisubstrate inhibitors of

farnesyltransferasesa novel class of specific inhibitors of ras

transformed-cells. Oncogene 1995, 10, 1763-1779.

(67) Still, W. C.; Kahn, M.; Mitra, A. Rapid chromatographic technique

for preparative separations with moderate resolution. J. Org. Chem.

1978, 43, 2923-2925.

(68) Baugh, C. M.; Shaw, E. Synthesis of 6-hydroxymethylypteridines.

J. Org. Chem. 1964, 29, 3610-3612.

(69) Boyle, P. H.; Pfleiderer, W. Pteridines. 69. Synthesis and reactivity

of 2,4-diamino-6-(hydroxymethyl)pteridine. Chem. Ber. 1980, 113,

1514-1523.

(70) Fu, S.-C. J.; Reiner, M. A new synthesis of p-methylaminobenzoyl-

L-glutamic acid. J. Org. Chem. 1965, 30, 1277-1278.

(71) Roeske, R. Preparation of tert-butyl esters of free amino acids. J.

Org. Chem. 1963, 28, 1251-1253.

(49) Deprele, S.; Montchamp, J. L. Triethylborane-initiated room-tem-

perature radical addition of hypophosphites to olefins: synthesis of

monosubstituted phosphinic acids and esters. J. Org. Chem. 2001,

66, 6745-6755.

(72) Kolasa, T.; Miller, M. J. 1-Hydroxy-3-amino-2-piperidone (delta-

N-hydroxycycloornithine) derivativesskey intermediates for the

synthesis of hydroxamate-based siderophores. J. Org. Chem. 1990,

55, 1711-1721.

(73) Altschul, R. The reversible esterification of carboxylic acids with

isobutene and trimethylethylenesquantitative studies and synthetic

applications. J. Am. Chem. Soc. 1946, 68, 2605-2609.

(74) Majer, P.; Jackson, P. F.; Delahanty, G.; Grella, B. S.; Ko, Y. S.; Li,

W. X.; Liu, Q.; Maclin, K. M.; Polakova, J.; Shaffer, K. A.; Stoermer,

D.; Vitharana, D.; Wang, E. Y.; Zakrzewski, A.; Rojas, C.; Slusher,

B. S.; Wozniak, K. M.; Burak, E.; Limsakun, T.; Tsukamoto, T.

Synthesis and biological evaluation of thiol-based inhibitors of

glutamate carboxypeptidase II: Discovery of an orally active GCP

II inhibitor. J. Med. Chem. 2003, 46, 1989-1996.

(50) Koch, T.; Buchardt, O. Synthesis of ^L-(+)-selenomethionine. Syn-

thesis 1993, 1065.

(51) Rosowsky, A.; Freisheim, J. H.; Bader, H.; Forsch, R. A.; Susten, S.

S.; Cucchi, C. A.; Frei, E. Methotrexate analogues. 25. Chemical

and biological studies on the γ-tert-butyl esters of methotrexate and

aminopterin. J. Med. Chem. 1985, 28, 660-667.

(52) Kagan, J.; Tolentino, L.; Ettlinger, M. G. Solution of a classical

problemstautomerism and isomerism in R-methylglutaconic acid

series. J. Org. Chem. 1975, 40, 3085-3093.

(53) Coward, J. K.; Lok, R. Cleavage of N-carbobenzyloxy group in

neutral and basic mediasneighboring-group participation of carbam-

ate moiety. J. Org. Chem. 1973, 38, 2546-2548.

JM050871P

Article Doi

Prodrug forms of N-[(4-deoxy-4-amino-10-methyl)pteroyl]glutamate-γ- [ψP(O)(OH)]-glutarate, a potent inhibitor of folylpoly-γ-glutamate synthetase: Synthesis and hydrolytic stability

DOI: 10.1021/jm050871p

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm050871p

Full text of DOI:10.1021/jm050871p

Products guided by the article

R&D Labs maybe for 875668-08-9

Relevant to this article

Hot Product