Synthesis of phosphonamide and thiophosphonamide dipeptides [3]

Full text of DOI:10.1021/ja015632f

J. Am. Chem. Soc. 2001, 123, 4861-4862
4861
Scheme 1
Synthesis of Phosphonamide and Thiophosphonamide
Dipeptides
Sheila D. Rushing and Robert P. Hammer*
Department of Chemistry, Louisiana State UniVersity
Baton Rouge, Louisiana 70803
ReceiVed February 5, 2001
ReVised Manuscript ReceiVed April 1, 2001
Phosphonopeptides are peptide mimics1 that emulate high-
energy tetrahedral transition states of enzyme-catalyzed peptide
hydrolysis reactions. Thus, they are useful as mechanism-based
inhibitors of aspartyl and metalloproteases.2 Several phosphono-
peptides have also shown powerful antibacterial activity.3 More
recently, these types of compounds have been applied to hapten
design for the generation of catalytic antibodies that possess
peptide ligase activity.4 Herein we report the synthesis of
phosphonamide dipeptides 1a-d (R ) Cbz) (Scheme 1) in
isolated yields up to 30% using our P(III) one-pot activation-
coupling-oxidation procedure.5 Previous attempts to prepare the
hapten precursor 1a from P(V) phosphonochloridate 2 and
D-tryptophanamide under a variety of conditions were unsuccess-
ful, due in part to the steric bulk of the amino acid side chains.6
The difficulty in preparing the phosphonamide 1a exemplifies
the experiences that many laboratories have encountered in
preparing phosphonamidate peptides via P(V) coupling protocols.
The successful preparation of 1 detailed herein using phosphono-
chloridite 3 [P(III)] as the key intermediate and exploration of
parameters for the reaction should provide a road map to others
who wish to prepare phosphonamides not accessible by P(V)
protocols.
Table 1. Phosphonochloridite Generation under Various Solvent
and Base Conditions
entry solvent Ph3PCl2 equiv base (equiv)
31P δ ppm (ratio)c
1
pyridine
2.5b
solvent
200 (0.25), 180 (0.05),
165 (0.04), 158 (0.15),
120 (0.18), 115 (0.05),
39-41 (0.08), 28 (1))
same as above
2
3
4
CH2Cl2
CH2Cl2
CH2Cl2
2.5b
2.0b
1.5b
Et3N (2)
none
193 (0.4), 40d (1)
pyridine (1) 200 (0.3), 193 (0.4),
28 (1)
5
CH2Cl2
1.2b
pyridine (2) 200 (0.85), 28 (1)
a Contains 10-20% Ph3PO. b Number of equivalents needed to
consume 4. c Value in parentheses represents intensity of each peak
relative to Ph3PO. d Ph3PO+HCl-.
The air-stable precursor used for generation of phosphono-
chloridite 3 is racemic, carbamate-protected H-phosphinate amino
acid 4 as a diastereomeric mixture of p-nitrobenzyl (pNb) esters.
The H-phosphinate 4, under certain conditions, was converted to
3 with the commercially available dichlorotriphenylphosphorane
(Ph3PCl2). Because the solvent pyridine and the solvent system
CH2Cl2/Et3N worked well in our model systems,5 they were
initially employed during activation of 4 with Ph3PCl2. However,
only a small amount of 3 (31P ) 200 ppm; Table 1) formed
(admixed with several other as yet unidentified products). This
may be due to the amide functionality in 4 possibly being sensitive
to the combination of Ph3PCl2 with Et3N or pyridine (when used
as solvent). Ph3PCl2 in the presence of Et3N is known to convert
secondary amides to imidoyl chlorides.7 Mechanistically, the
formation of an isocyanate from the carbamate functionality with
Ph3PCl2/Et3N is also possible. We therefore searched for alterna-
tive solvent conditions for the activation step. In standard P(V)
chemistry, successful conversion of a phosphonate monoester to
the corresponding phosphonochloridate with thionyl or oxalyl
chloride is often done in the absence of base.6,8 Therefore, in an
analogous attempt to improve the activation step of our P (III)
protocol,5 treatment of 4 with Ph3PCl2 was performed in CH2Cl2
without any base present, resulting in the complete conversion
of 4 to an activated P(III) species (31P ) 193 ppm). This P(III)
species was later determined, via a study conducted on the
activation step using 4 and GC-MS analysis of the products, to
be lacking the pNb group.9 This study on the activation step was
conducted because of problems with isolation of final products 1
by normal-phase chromatography and because of FAB-MS spectra
of 1 revealing the major products to be ones in which the pNb
group had been apparently cleaved off. This same type of ester
cleavage was also observed in other systems in our laboratory.10
We concluded that in order to prevent pNb cleavage, the presence
of base to scavenge generated HCl during the activation reaction
was essential.
(1) (a) Kaplan, A. P.; Bartlett, P. A. Biochemistry 1991, 30, 8165. (b)
Bertenshaw, S. R.; Rogers, R. S.; Stern, M. K.; Norman, B. H. J. Med. Chem.
1993, 36, 173.
(2) Aspartyl: (a) Bartlett, P. A.; Hanson, J. E.; Giannousis, P. P. J. Org.
Chem. 1990, 55, 6268. (b) Ikeda, S.; Ashley, J. A.; Wirsching, P.; Janda, K.
D. J. Am. Chem. Soc. 1992, 114, 76. Metallo: (c) Giannousis, P. P.; Bartlett,
P. A. J. Med. Chem. 1987, 26, 1603. (d) Barelli, H.; Dive, V.; Yiotakis, A.;
Vincent, J. P.; Checler, F. Biochem. J. 1992, 287, 621.
(3) (a) Allen, J. G.; Atherton, F. R.; Hall, M. J.; Hassall, C. H.; Holmes,
S. W.; Lambert, R. W.; Nisbet, L. J.; Ringrose, P. S. Nature 1978, 272, 56.
(b) Lejczak, B.; Kafarski, P.; Sztajer, H.; Mastalerz, P. J. Med. Chem. 1986,
29, 2212.
(4) Hirschmann, R.; Smith, A. B., III; Taylor, C. M.; Benkovic, P. A.;
Taylor, S. D.; Yager, K. M.; Sprengeler, P. A.; Benkovic, S. J. Science 1994,
265, 234.
(5) Fernandez, M. d. F.; Vlaar, C. P.; Fan, H.; Liu, Y.-H.; Fronczek, F. R.;
Hammer, R. P. J. Org. Chem. 1995, 60, 7390.
The best solvent/base combination for activation of 4 with
Ph3PCl2 was finally determined to be 2 equiv of pyridine (∼3%)
in CH2Cl2. With 1 equiv of pyridine nearly equal amounts of 3
(31P ) 200 ppm) and the P(III) species (31P ) 193 ppm)9 that
resulted from pNb cleavage formed. Furthermore, as the number
(6) (a) Smith, A. B., III; Taylor, C. M.; Benkovic, P. A.; Taylor, S. D.;
Hirschmann, R. Tetrahedron Lett. 1994, 35, 6853. (b) Hirschmann, R.; Yager,
K. M.; Taylor, C. M.; Witherington, J.; Sprengeler, P. A.; Phillips, B. W.;
Moore, W.; Smith, A. B., III J. Am. Chem. Soc. 1997, 119, 8177. (c)
Hirschmann, R.; Yager, K. M.; Taylor, C.; Moore, W.; Sprengeler, P. A.;
Witherington, J.; Philips, B. W.; Smith, A. B., III J. Am. Chem. Soc. 1995,
117, 6370.
(8) (a) Musiol, Hans J.; Grams, F.; Rudolph-Bohner, S.; Moroder, L. J.
Org. Chem. 1994, 59, 6144. (b) Malachowski, W. P.; Coward, J. K. J. Org.
Chem. 1994, 59, 7616. (c) Maffre-Lafon, D.; Escale, R.; Girard, J. P.
Tetrahedron Lett. 1994, 35, 4097.
(9) We propose that activation in the absence of base forms 3 but quickly
converts, after pNb cleavage, to the respective phosphonodichloridite, with
concomitant p-nitrobenzyl chloride formation.
(7) Relles, H. M.; Schluenz, R. W. J. Am. Chem. Soc. 1974, 96, 6469.
(10) Fan, H.; Rushing, S.; Hammer, R. P. Unpublished results.
10.1021/ja015632f CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/01/2001

4862 J. Am. Chem. Soc., Vol. 123, No. 20, 2001
Communications to the Editor
Table 2. Synthesis of Phosphonamide and Thiophosphonamide
Dipeptidesa
entry
Y
X
productb
% yieldc
31P
1
2
3
4
NH2
NH2
OCH3
OCH3
O
S
O
S
1a
1b
1c
1d
18
14
15
30
28.8, 29.4
85.3, 86.6
30.5, 30.8
87.3, 88.4
a 1a-d were prepared by first generating the proposed phosphity-
lating agent 7 from 3 and 2 equiv of DIEA. Then 5 was coupled to 7
to form the phosphonamidite 6a or 6b followed by sulfurization or
oxidation with elemental S8 or t-BuOOH, respectively. b 1a-d were
1
isolated using HPLC and characterized by 31P NMR, H NMR, and
Figure 1. 31P spectra: (A) species 3 (200 ppm), Ph3PO (27.5 ppm), and
unreacted 4 (36, 37 ppm) and Ph3PCl2 (64 ppm); (B) proposed
oxazaphospholine 7 (180.3 ppm); (C) species 6a (123.6, 124.9 ppm);
and (D) target 1b (87.2 and 88.3 ppm).
FAB-MS. c Isolated yield.
amidite diastereomers preferentially formed, which, after sulfuri-
zation or oxidation, resulted in compound 1 (Table 2). Leucine
methyl ester and glycine ethyl ester were also coupled to 7
followed by sulfurization, resulting in isolated yields of 28% and
26%, respectively (see Supporting Information). As NMR spectra
indicate almost complete conversion of starting H-phosphinates
to final phosphonamides, the moderate yields obtained here are
likely indicative of losses during purification due to the hydrolytic
instability of phosphonamides.13
Scheme 2
We also explored an alternative structure for the phosphitylating
agent. We speculated that the observed reactive species could be
analogous to the phosphonyltrialkylammonium salts observed by
Hirschmann and co-workers in phosphonate activation.6b,c To test
that hypothesis, we generated the phosphitylating agent (31P, ≈180
ppm) by adding the phosphonochloridite 3 to PS-DIEA resin
(DIEA attached to a polystyrene resin). We reasoned that if a
reactive phosphinyltrialkylammonium salt forms, then it would
remain attached to the resin. However, after the reaction mixture
was filtered from the resin, the newly formed phosphitylating
agent was found only in solution, and no 31P signal was detected
on the solid support by gel 31P NMR. Therefore, in our case, we
concluded that the species that resonated around 180 ppm was
not a phosphinyltrialkylammonium salt but the oxazaphospholine
7.
In conclusion, the P(III) approach to phosphonopeptides
described here provides a route to phosphonamide 1a (and 1c)
that was previously unattainable by P(V) methods. Additionally
the use of reduced phosphorus intermediates allowed preparation
of thiophosphonamide peptides 1b and 1d. Further improvements
in this method using nonparticipating amino protection, milder
activation reagents, and solid-supported approaches are currently
under investigation.
of equivalents of pyridine increased from two to large excess (e.g.,
pyridine as solvent), the number and amount of side-products
increased (Table 1, entry 1 vs entries 4 and 5). With the optimized
solvent/base conditions, very few side products were evident from
31P NMR (Table 1, entry 5; Figure 1A).
The next step of the one-pot procedure was the coupling of 3
to D-tryptophanamide 5a or D-tryptophan methyl ester 5b to form
the phosphonamidite 6 (Scheme 2). The major problem initially
encountered with the amine coupling of 5 to 3 was the formation
of an unknown substance (31P, ≈180 ppm) that competed with
the formation of 6. If 5 was added in one portion to the
phosphonochloridite 3, only the peak ≈180 ppm would be
observed. However, if 5 was added over 15-20 min, the intensity
of the peak ≈180 ppm and the expected four diastereomeric peaks
(31P, 123, 124, 126, 128 ppm) would be nearly equal. We therefore
directed our investigation toward elucidating the identity or
properties of that unknown species (31P, ≈180 ppm.).
In the presence of tertiary amines urethane-protected amino
acids and peptides are known to cyclize to form oxazolones,11
which are themselves reactive acylating agents. We reasoned that
similar cyclization could be occurring with our urethane-protected
phosphonochloridite 3 to form an oxazaphospholine12 (31P, ≈180
ppm). We generated the proposed oxazaphospholine 7 by adding
1.5 equiv of the tertiary base, diisopropylethylamine (DIEA), to
3. Next amine nucleophile 5a or 5b and DIEA (2 equiv each)
were added in one portion. After about 15 min, two phosphon-
Acknowledgment. We thank the Louisiana Board of Regents
(LEQSF) and NSF (CHE-9500992) for financial support. We thank the
NSF for funding to purchase NMR (CHE-9809187) and mass spectrom-
etry (CHE-9809178) instrumentation used in this work. We thank LSU
and the National NOBCChE Chapter for a Huel D. Perkins Graduate
Fellowship and a Procter and Gamble Fellowship, respectively, to S.D.R.
Supporting Information Available: Experimental details for the
synthesis and characterization of phosphonopeptides (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
(11) Benoiton, N. L. In The Peptides; Undenfriend, S., Meienhofer, J., Eds.;
Academic Press: San Diego, CA, 1983; Vol. 5, Chapter 4.
JA015632F
(12) For an example of a P(III) oxazaphospholine formed from Cl2POEt
and RCONHCH2CO2Et with Et3N see: Malenko, D.; Nesterova, L.; Luky-
anenko, S.; Sinitsa, A. Zh. Obshch. Khim. 1989, 59, 2626. Oxazaphospholine
formation has also been proposed as an intermediate in phosphonate [P(V)]
activation and coupling, but was discounted in favor of phosphonyltrialkyl-
ammonium salts (see refs 6b,c).
(13) (a) Rahil, J.; Haake, P. J. Am. Chem. Soc. 1981, 103, 1723. (b)
Yamauchi, K.; Ohtsuki, S.; Kinoshia, M. J. Org. Chem. 1984, 49, 1158. (c)
Elliot, R. L.; Marks, N.; Berg, M. J.; Portoghese, P. S. J. Med. Chem. 1985,
28, 1208.