Solid-phase synthesis of protected a-amino phosphonic acid oligomerswzYoshitaka Ishibashi and Masato Kitamura*

Received (in Cambridge, UK) 22nd June 2009, Accepted 30th September 2009

First published as an Advance Article on the web 14th October 2009

DOI: 10.1039/b912231a

By establishing both a highly efficient phosphonamidate

formation and a RuCp-catalyzed cleavage of an allyl linker,

the solid-phase synthesis of Fmoc-(GlyP(OBn))6-OH/DIEA, a

protected form of a new type of unnatural peptide a-amino

phosphonic acid oligomer (APO), has been realized.

a-Amino phosphonic acid oligomers (APOs) can be

considered to be the closest analogues of a-amino carboxylic

acid-based peptides. Incorporating both sp3-hybridized

tetrahedral phosphorus(V) atoms and the dibasic acid

properties of phosphonic acid moieties into the peptide

backbone modifies the size and shape of its three-dimensional

structure, as well as the isoelectronic points of the parent

a-amino acids, which possess smaller, flat, monobasic

carboxylic functions.1,2 In addition, the strong hydrogen

bonding ability of the phosphonamidate might be advanta-

geous for constructing higher-order structures.3 These steric

and electronic differences influence the diverse chemical and

biological activities of APOs through inherent competition

with their carboxylic counterparts for the active sites of

enzymes, antibodies or receptors.1a Intense interest in

these so-called ‘‘phosphorus analogues’’ has increasingly

attracted attention with regard to their sequential synthesis

on a solid phase; however, there have been no reports of

their synthesis since 1975, when Yamauchi reported the

solution-phase synthesis of a protected dimer or trimer using

a phosphonochloridate method,4a and the non-sequential

polymerization of a-amino phosphonic acids or esters.4b This

may be partly due to the lack of efficiency in the direct

dehydrative condensation5 between N-protected a-amino

phosphonic acids and a-amino phosphonic esters, and in

the cleavage of the target molecules from supports under

conditions that do not affect the acid-labile P–N bonds.6 In

pursuit of our final goal of creating APOs, we have realized

here, for the first time, the solid-phase synthesis of

Fmoc-(GlyP(OBn))6-OH/DIEA (DIEA = N,N-diisopropyl-

ethylamine) using an N-Fmoc-protected 1-aminomethylphos-

phonic acid monobenzyl ester, Fmoc-GlyP(OBn)-OH (1),yas the simplest monomer unit.

Fig. 1 illustrates the synthetic cycle, which follows standard

peptide chemistry. The chain is elongated in the N–P direction

on a TentaGel support attached to the monomer via a Jones

a-branched allyl linker.7 One cycle involves Fmoc removal

using piperidine,8 followed by condensation of monomer acid

1 (key point 1). After repetition of this cycle n times, the n-mer

ester is catalytically detached to give Fmoc-(GlyP(OBn))n-OH

(key point 2).

The first key point in Fig. 1 was investigated by the

condensation of 1 with H-GlyP(OBn)-OBn (2), giving

Fmoc-(GlyP(OBn))2-OBn (3) in solution at the outset

(Fig. 2). Successful condensation was attained by the use of

diisopropylcarbodiimide (DIC) and 1-hydroxy-7-azabenzo-

triazole (HOAt), together with DIEA. In 1 h at room

temperature, a modified Carpino’s method quantitatively

converted 1 and 2 in CD3CN to 3 (observed by 31P NMR

analysis). The dimer was isolated in 93% yield. Contrary to

the carboxylic acid case,9 the yield did not exceed 64% in the

absence of DIEA. Removal of Fmoc from 3 was performed

by using piperidine, without any side reactions, to give

H-(GlyP(OBn))2-OBn quantitatively (83% isolated yield).10

31P NMR analysis10 revealed that the reaction proceeds as

shown in Fig. 2. First, the phosphonic acid 1 (d 21.7) was

quantitatively converted into the pyrophosphate anhydride 4

(d 16.7),5,11 10 min after 1 and DIC were mixed in a 1 : 1.5

ratio in CD3CN. The addition of 1.2 equivalents of HOAt to

the solution containing 4 generated the HOAt-activated

Fig. 1 The synthetic cycle of a-amino phosphonic acid oligomers.

Research Center for Materials Science and the Department ofChemistry, Nagoya University, Chikusa, Nagoya 464-8602, Japan.E-mail: kitamura@os.rcms.nagoya-u.ac.jp; Fax: +81 52-789-2958w This article is part of a ChemComm ‘Catalysis in Organic Synthesis’web-theme issue showcasing high quality research in organicchemistry. Please see our website (http://www.rsc.org/chemcomm/organicwebtheme2009) to access the other papers in this issue.z Electronic Supplementary Information (ESI) available: Experimentaldetails for the solution-phase and solid-phase synthesis. See DOI:10.1039/b912231a

This journal is �c The Royal Society of Chemistry 2009 Chem. Commun., 2009, 6985–6987 | 6985

COMMUNICATION www.rsc.org/chemcomm | ChemComm

Publ

ishe

d on

14

Oct

ober

200

9. D

ownl

oade

d by

UN

IVE

RSI

TY

OF

AL

AB

AM

A A

T B

IRM

ING

HA

M o

n 22

/10/

2014

05:

35:0

9.

View Article Online / Journal Homepage / Table of Contents for this issue

phosphonic acid 5 (d 30.7). The signal completely disappeared

30 min after the addition of 1 equivalent of amine 2 (d 29.3)

and 1 equivalent of DIEA, giving a new set of signals at d 26.2

and 28.0, which could be assigned to the phosphorous atoms

of the phosphonic diester and the phosphonamidate in 3,

respectively. The coexistence of DIEA at the stage of forming

5 from 4 gave, however, oxazaphospholine 6 (d 13.9)5 in

ca. 80% yield, which was not further converted to 3 when 2

was added to 6. On the other hand, the reaction of 5 with 2 was

not completed without DIEA, instead giving a mixture of 3, 5

(d 30.7) and another species (d 24), probably the 2/HOAt salt.

The use of N-hydroxybenzotriazole (HOBt) instead of HOAt

stopped the reaction at the stage of anhydride 4 formation.

With benzotriazole-1-yl-oxy-trispyrrolidinophosphonium

hexafluorophosphate (PyBOP), the HOBt-activated ester was

obtained, whereas no reaction with 2 occurred, even with the

assistance of DIEA.

Evidence from a series of experiments clearly indicated

the importance of the N(7) atom in HOAt for facilitating

phosphonamidate P–N bond formation, as has been suggested

in the case of carboxamide.9 The simultaneous addition of 2

and DIEA to 5 would generate a stable DIEA/HOAt

(pKa 3.47)9c complex, moving the equilibrium predominantly

to the side of 3. The generation of the very reactive phosphonyl-

trialkylammonium species, as reported by Smith et al.,5 might

be also involved, although the 31P NMR spectrum of a 1 : 1

mixture of 5 and DIEA showed no signals in region of d 445

assignable to the species. Other methods used for phosphon-

amidate formation between a-amino carboxylic esters and

N-protected a-amino phosphonic monoesters were not

efficient.10 In situ-activation of 1 with diphenylphosphoryl

azide (DPPA)6b or isolation of the phosphonochloridate

(SOCl2, 2 h, rt), followed by the addition of 2 and triethyl-

amine, gave 3, at best, in 80% yield (by 31P NMR analysis),5

which is not high enough for oligomer synthesis.12

The second issue, detachment of the APO from Jones’ allyl

linker, was investigated by the use of our deallylation catalyst,

[RuCp(P(C6H5)3)(CH3CN)2]PF6,13 and model compound 7a,

prepared from 8a via the condensation of 1, followed by Fmoc

removal and the recondensation of 1.10 The standard

conditions reported ([substrate] = 50 mM, [catalyst] = 0.5 mM,

CH3OH, 25 1C, 1 h) detached the dipeptide Fmoc-

(GlyP(OBn))2-OH (7b) (31P NMR: d 24 and 29.5), together

with compounds with resonances at d 22.6 and 12.4,

which could be assigned to Fmoc-GlyP(OBn)-OCH3 and

H-GlyP(OBn)-OH, produced by self-acid-catalyzed P–N bond

cleavage. This problem was solved by the addition of DIEA

(50 mM) to the catalytic system to neutralize the acid, giving

7c (d 18.9 and 29.9) in quantitative yield after 3 h, together

with the corresponding branched allyl methyl ether 8b

(branch/normal ratio ca. 3 : 1). Compound 7c was isolated

in >90% yield. As this detachment reaction proceeded

without the need for any additional nucleophiles, such as

potassium carboxylate, alkoxide, morpholine, pyrrolidine,

formate, tributyltin hydride, borohydride or silane, which

are generally required for conventional Pd methods,13 the

isolation of the product will be simplified when applied to its

solid-phase synthesis.

Next, both the condensation method and the detachment

process established in solution were applied to the solid-

phase synthesis of Fmoc-(GlyP(OBn))6-OH/DIEA. Thus,

initial monomer 1 was connected to a TentaGel-attached

Jones’ branched allyl linker 8c (A (m = 0) in Fig. 1)

(0.26 mmol OH g�1, swelled in CH3CN) by use of the

DIC/HOAt/DIEA method (rt, 10 h). Washing of the

1-connected resin with DMF, CH3OH and CH2Cl2, followed by

drying at 0.1 mmHg, gave B (n = 1), which showed 31P NMR

signals at d 22.6, 23.4 and 24.6 (4 : 85 : 11). A standard

piperidine treatment of B (n = 1) (CH2Cl2, rt, 1 h) released

free amine A (m = 1) (31P NMR signals at d 28.0 and 28.3

(1 : 1)). Repetition of the (1) swelling, (2) condensation,

(3) washing and (4) drying processes afforded B (n = 2)

(31P NMR: d 25 and 29), which was again converted to A

(m = 2) by the (5) piperidine, (6) washing and (7) drying

processes. Three repetitions of the (1)–(7) processes,

followed by one cycle of the (1)–(4) processes afforded

TentaGel-attached B (n = 6). Fig. 3 shows the 31P NMR

spectra obtained in each cycle. The signals labelled J corres-

pond to the phosphorus atom of the diester moiety, the signals

labelled n to those of internal phosphonamidates and the

signals labelled & to those of N-terminal examples. As the

chain is elongated, the intensity of the internal phosphorus

signals increase, although all signals tend to be broadened.

This broadening can be ascribed to the increase in the number

of diastereomers originating from the stereogenic centers of

the phosphorus and carbon atoms. In each step, the 31P NMR

Fig. 2 The supposed reaction pathways of DIC/HOAt/DIEA-

promoted condensation.

6986 | Chem. Commun., 2009, 6985–6987 This journal is �c The Royal Society of Chemistry 2009

Publ

ishe

d on

14

Oct

ober

200

9. D

ownl

oade

d by

UN

IVE

RSI

TY

OF

AL

AB

AM

A A

T B

IRM

ING

HA

M o

n 22

/10/

2014

05:

35:0

9.

View Article Online

signals of the terminal phosphonamidates appear at around

d 29, and the chemical shifts are lowered to about d 33 after

Fmoc removal. This phenomenon can be understood by

assuming that the Fmoc moiety has a magnetic shielding

effect, an idea that is consistent with our observations in the

conversion of 3 (d 25.8, 27.0 and 28.0 (50 : 4 : 46))

to H-(GlyP(OBn))2-OBn (d 25.6 and 32.3) in CDCl3.10

Compound B (n = 6) was subjected to a cationic

RuCp/P(C6H5)3-catalyzed detachment13 (B= 80 mg (0.1 mmol),

S/C = 100, [RuCp catalyst] = 1 mM, [DIEA] = 100 mM,

CH3OH, rt, 24 h, >95% conversion) to give Fmoc-

(GlyP(OBn))6-OH/DIEA (31P NMR: d 19, 29.5 and 31.5;

MALDI-TOF MS: 1360.5 (Fmoc-(GlyP(OBn))6-ONa); tR in

HPLC 3.3 min (column Develosil ODS-UG 4.6 mm � 25 cm,

eluent 8 : 2–7 : 3 CH3OH–H2O, flow rate 1.0 mL min�1)).

In summary, to the best of our knowledge, a protected

a-amino phosphonic acid hexamer, Fmoc-(GlyP(OBn))6-OH/

DIEA, has been synthesized for the first time by solving

two problems: namely, establishing a DIC/HOAt/DIEA

dehydrative-condensation method and achieving a RuCp/

P(C6H5)3-catalyzed detachment from Jones’ allyl linker. Until

now, it has not been possible to regard APO as a synthetic

target because of the instability of the P–N bond. Although

there is no a-substituent and the final product is still protected,

our new method should inspire further research into the design

and preparation of various phosphorus analogues of

higher peptides, which have high potential in peptide and

nucleotide sciences, and future biochemical technologies.

The characteristics of APOs should make our new type of

unnatural peptide an important complement to existing

unnatural biooligomers.14 The synthesis of unprotected

H-(AlaP(OH))6-OH/DIEA is an on-going project in our group.

Notes and references

y a-Amino phosphonic acid abbreviation follows the IUPAC ruledetermined for usual a-amino acids by the addition of ‘‘P’’ at theend of the three character code. e.g. GlyP for a-aminomethyl phosphonicacid, where H-GlyP(OH)-OH represents H2NCH2P(O)(OH)2.

1 Book and review: (a) Aminophosphonic and Aminophosphinic Acids,ed. P. V. Kukhar and R. H. Hudson, Wiley-VCH,Weinheim, 1991,pp. 1–634; (b) P. Kafarski and B. Lejczak, Phosphorus, SulfurSilicon Relat. Elem., 1991, 63, 193–215; (c) The first report ona-amino phosphonic acids: M. Horiguchi and M. Kandatsu,Nature, 1959, 184, 901–902.

2 M. Kitamura, M. Yoshimura, M. Tsukamoto and R. Noyori,Enantiomer, 1996, 1, 281–303.

3 B. P. Morgan, J. M. Scholtz, M. D. Ballinger, I. D. Zipkin andP. A. Bartlett, J. Am. Chem. Soc., 1991, 113, 297–307.

4 (a) K. Yamauchi, Y. Mitsuda and M. Kinoshita, Bull. Chem. Soc.Jpn., 1975, 48, 3285–3286; (b) K. Yamauchi, J. Synth. Org. Chem.,Jpn., 1988, 46, 654–666.

5 R. Hirschmann, K. M. Yager, C. M. Tayor, J. Witherington,P. A. Sprengeler, B. W. Phillips, W. Moore and A. B. Smith III,J. Am. Chem. Soc., 1997, 119, 8177–8190.

6 (a) N. S. Sampson and P. A. Bartlett, J. Org. Chem., 1988, 53,4500–4503; (b) K. Yamauchi, S. Ohtsuki andM. Kinoshita, J. Org.Chem., 1984, 49, 1158–1163; (c) M. Hariharan, R. J. Motekaitisand A. E. Martell, J. Org. Chem., 1975, 40, 470–473; (d) For areport on relatively stable phosphonamidates, see: D. A. McLeod,R. I. Brinkworth, J. A. Ashley, K. D. Janda and P. Wirsching,Bioorg. Med. Chem. Lett., 1991, 1, 653–658.

7 X. Zhang and R. A. Jones, Tetrahedron Lett., 1996, 37, 3789–3790.8 L. A. Carpino and G. Y. Han, J. Org. Chem., 1972, 37, 3404–3409.9 (a) L. A. Carpino, J. Am. Chem. Soc., 1993, 115, 4397–4398;(b) L. A. Carpino and A. El-Faham, Tetrahedron, 1999, 55,6813–6830; (c) L. A. Carpino, H. Imazumi, B. M. Foxman,M. J. Vela, P. Henklein, A. El-Faham, J. Klose and M. Bienert,Org. Lett., 2000, 2, 2253–2256.

10 For details, see the ESIz.11 (a) R. Greenhalgh, R. M. Heggie and M. A. Weinberger, Can. J.

Chem., 1970, 48, 1351–1357; (b) N. S. Corby, G. W. Kenner andA. R. Todd, J. Chem. Soc., 1952, 1234–1242.

12 Y. Hayakawa, S. Wakabayashi, H. Kato and R. Noyori, J. Am.Chem. Soc., 1990, 112, 1691–1696.

13 M. Kitamura, S. Tanaka and M. Yoshimura, J. Org. Chem., 2002,67, 4975–4977.

14 (a) Z. Zhang and E. Fan, J. Org. Chem., 2005, 70, 8801–8810;(b) D. Seebach, A. K. Beck and D. J. Bierbaum, Chem. Biodiversity,2004, 1, 1111–1239; (c) K. Oh, K.-S. Jeong and J. S. Moore,Nature, 2001, 414, 889–893; (d) K. Kirshenbaum, R. N.Zuckermann and K. A. Dill, Curr. Opin. Struct. Biol., 1999, 9,530–535; (e) S. E. Schneider and E. V. Anslyn, in Advances inSupramolecular Chemistry, ed. G. W. Gokel, JAI Press,Greenwich, 1999, vol. 5, pp. 55–120; (f) M. J. Soth andJ. S. Nowick, Curr. Opin. Chem. Biol., 1997, 1, 120–129;(g) R. P. Cheng, S. H. Gellman and W. F. DeGrado, Chem.Rev., 2001, 101, 3219–3232.

Fig. 3 The 31Pt{1H} NMR spectrum change during APO elongation

on TentaGel (CDCl3, 25 1C). a: Dimer, b: trimer, c: tetramer,

d: pentamer, e: hexamer. The right-hand spectra show the Fmoc-

deprotected oligomers. K: CH(CHQCH2)(CH2)7CONH-TentaGel.

This journal is �c The Royal Society of Chemistry 2009 Chem. Commun., 2009, 6985–6987 | 6987

Publ

ishe

d on

14

Oct

ober

200

9. D

ownl

oade

d by

UN

IVE

RSI

TY

OF

AL

AB

AM

A A

T B

IRM

ING

HA

M o

n 22

/10/

2014

05:

35:0

9.

View Article Online