Intramolecular Hydrogen-bond Participation in Phosphonylammonium Salt Formation

Article abstract of DOI:10.1021/ol0066330

equation presented A series of phosphonochloridates and phosphonyl dichlorides were prepared, and their reactivity with triethylamine has been investigated using ³¹P NMR spectroscopy. Taken together these studies provide evidence that an intram

Full text of DOI:10.1021/ol0066330

ORGANIC
LETTERS
2000
Vol. 2, No. 24
3887-3890
Intramolecular Hydrogen-Bond
Participation in Phosphonylammonium
Salt Formation
Amos B. Smith, III,* Laurent Ducry, R. Michael Corbett, and Ralph Hirschmann*
Department of Chemistry, UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104
smithab@sas.upenn.edu
Received September 21, 2000
ABSTRACT
A series of phosphonochloridates and phosphonyl dichlorides were prepared, and their reactivity with triethylamine has been investigated
using 31P NMR spectroscopy. Taken together these studies provide evidence that an intramolecular hydrogen-bond is required for
phosphonylammonium salt formation to render the phosphorus more electron-deficient.
Phosphonate esters are privileged surrogates for the unstable,
high energy tetrahedral transition states found in many
chemical reactions, including enzyme-catalyzed reactions.1-3
Effective phosphorylation procedures are, therefore, of
importance in the design and synthesis of, on one hand,
inhibitors of proteolytic enzymes and, on the other, haptens
for the generation of catalytic antibodies.4 These protocols
are also of great interest for the preparation of analogues of
phospho-ester-containing natural products.5
been made to develop more reactive phosphonylating agents
[e.g., bis(benzotriazolyl)-phosphonates].7 During the course
of our work on hapten synthesis, we reported a new method
for the phosphonylation of alcohols.8 This protocol calls for
treatment of phosphonochloridates 2a,b with triethylamine
to afford a new species, the phosphonyl triethylammonium
salts (3a,b; Scheme 1). The reactivity of these phosphony-
Scheme 1
The preparation of mixed phosphonate esters generally
relies on the reaction between phosphonochloridates or
phosphonyl dichlorides with alcohols.6 Efforts have also
* Corresponding authors. E-mail for R. Hirschmann: rfh@sas.upenn.edu.
(1) Lerner, R. A.; Benkovic, S. J.; Schultz, P. G. Science 1991, 252,
659-667.
(2) Benkovic, S. J. Annu. ReV. Biochem. 1992, 61, 29-54.
(3) (a) 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, 234-237. (b) Smithrud, D. B.; Benkovic, P. A.; Benkovic, S.
J.; Taylor, C. M.; Yager, K. M.; Witherington, J.; Phillips, B. W.; Sprengeler,
P. A.; Smith, A. B., III; Hirschmann, R. J. Am. Chem. Soc. 1997, 119,
278-282.
lating agents toward alcohols and amines was found to be
high, often providing superior yields of phosphonate esters
(4) For example, see: (a) Pratt, R. F. Science 1989, 246, 917-919. (b)
Bartlett, P. A.; Giangiordano, M. A. J. Org. Chem. 1996, 61, 3433-3438.
(c) Chen, S.; Lin, C.-H.; Kwon, D. S.; Walsh, C. T.; Coward, J. K. J. Med.
Chem. 1997, 40, 3842-3850.
(6) Zhao, K.; Landry, D. W. Tetrahedron 1993, 49, 363-368 and
references cited therein.
(5) Engel, R. Chem. ReV. 1977, 77, 349-367.
10.1021/ol0066330 CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/28/2000

(4a,b) or phosphonamides when compared with the reaction
of the corresponding phosphonochloridates. In addition to
improved yields and reaction rates, 3a,b, obtained by the
sequential addition of triethylamine to the phosphonochlo-
ridates, also displayed a greater affinity for alcohols over
amines.8b
Recently, we became interested in applying this method
to phosphonyl dichloride 5. Unexpectedly, no phosphonyl-
ammonium salt was formed when 5 was treated with
triethylamine in deuterated chloroform (Scheme 2). Accord-
ing to 31P NMR spectroscopy, the starting material remained
unchanged.
the 31P NMR resonance (from 44.0 to 54.0), characteristic
for phosphonylammonium salt formation.8
These examples clearly demonstrate that both mono- and
dichlorides can be converted into phosphonylammonium
salts. The difference in reactivity between 2a and 8 (or 11
and 5) must therefore derive from the (N-benzyloxycarbon-
yl)aminomethyl vs phenyl side chain. The presence of an
R-alkyl substituent on compound 2a had already been
demonstrated to be compatible with phosphonylammonium
salt formation,8b indicating that an electronic factor, rather
than steric, most likely accounts for the difference.
Following this reasoning, replacement of the Cbz-carbam-
ate by another electron-withdrawing substituent like a
phthalimide was expected to result in similar reactivity.
Phthalimides 1311 and 1612 were converted, respectively, to
the corresponding monochloride 14 and dichloride 17, and
then treated with triethylamine (Scheme 3). Both 14 and 17
Scheme 2
Scheme 3
Intrigued that this result could arise from the structural
differences between compounds 2 and 5, we examined in
more detail the structural features of phosphonochloridates
that are suitable precursors of phosphonylammonium salts.
For example, monochloride 8 was examined to determine
whether the presence of an alkoxy group, more electro-
negative than chlorine, but having a greater π donor ability
compared to Cl is required. Toward this end, conversion of
dichloride 5 to the corresponding symmetrical diester fol-
lowed by alkaline hydrolysis furnished 7 in 97% yield
(Scheme 2). Monochloride 8,9 obtained by treatment of
compound 7 with oxalyl chloride, however, failed upon
addition of triethylamine (same conditions as for 2) to
generate phosphonylammonium salt 9. The reverse experi-
ment, using phosphonyl dichloride 11 having the same alkyl
side-chain as 2a, was then explored. Conversion of phos-
phonic acid 1010 to the corresponding dichloride 11 and
treatment with triethylamine resulted in a downfield shift of
failed to form the corresponding phosphonylammonium salt
(e.g., 31P NMR), implying that the carbamate group of 2 and
11, absent in 14 and 17, promotes phosphonylammonium
salt formation by some mechanism other than a simple
inductive effect.
This observation suggested that the acidic carbamate
hydrogen might be the distinctive feature. Formation of an
intramolecular hydrogen bond between the carbamate NH
and the phosphonyl oxygen would increase the polarization
of the P-O bond, rendering the phosphorus more electron-
deficient, and therefore more susceptible to attack by the
nucleophilic tertiary amine (Figure 1). The importance of
(7) (a) van der Klein, P. A. M.; Dreef, C. E.; van der Marel, G. A.; van
Boom, J. H. Tetrahedron Lett. 1989, 30, 5473-5476. (b) Dreef, C. E.;
Douwes, M.; Elie, C. J. J.; van der Marel, G. A.; van Boom, J. H. Synthesis
1991, 443-447.
(8) (a) Hirschmann, R.; Yager, K. M.; Taylor, C. M.; Moore, W.;
Sprengeler, P. A.; Witherington, J.; Phillips, B. W.; Smith, A. B., III. J.
Am. Chem. Soc. 1995, 117, 6370-6371. (b) Hirschmann, R.; Yager, K.
M.; Taylor, C. M.; Witherington, J.; Sprengeler, P. A.; Phillips, B. W.;
Smith, A. B., III. J. Am. Chem. Soc. 1997, 119, 8177-8190.
(9) The mono- and dichlorides described in this report were prepared in
situ at 0 °C, concentrated in vacuo in the presence of toluene and used
without further purification.
Figure 1. Electrophilic activation of phosphorus via an intra-
molecular hydrogen-bond.
intramolecular participation of a hydrogen-bond in the base-
catalyzed hydrolysis of phosphonate esters has been re-
3888
Org. Lett., Vol. 2, No. 24, 2000

ported.13 We propose here that in similar fashion, a five-
membered-ring intramolecular hydrogen-bond activates the
phosphorus atom in 2a and 2b and is thereby responsible
for formation of the phosphonylammonium salts (Figure 1).
This result was confirmed by carrying out the same
experiment on a 1:1 mixture of monochlorides 2a and 22a.
Addition of triethylamine (3-fold excess) resulted in the
complete conversion of 2a into the corresponding phosphon-
ylammonium salt 3a, whereas the N-methylated derivative
21a remained unmodified (Figure 2). The 31P NMR spectra
1
Infrared and H NMR spectroscopy indicate that com-
pound 2a indeed forms an intramolecular hydrogen-bond.
The IR spectra of compound 2a in chloroform displayed two
NH stretches; a broad peak at 3340 cm-1 derives from
hydrogen-bonded conformers, whereas the sharp peak at
3445 cm-1 arises from non-hydrogen-bonded conformers.14,15
The spectra are essentially identical over the concentration
range studied (5 × 10-2 to 1 × 10-3 M), indicating that the
hydrogen-bond is intra- and not intermolecular. Likewise,
1
variable concentration H NMR spectroscopy between 5 ×
10-2 to 1 × 10-3 M revealed only a small concentration
dependence of the NH chemical shift (0.25 ppm), consistent
with the presence of an intramolecular hydrogen-bond.14
Additional support for electrophilic phosphorus activation
arises from the N-methylated analogues of compounds 2a,b
(Scheme 4). Treatment of diethyl phosphonate 198 with
Scheme 4
Figure 2. 31P NMR spectra: (a) monochlorides 2a and 22a (1:1
mixture); (b) phosphonylammonium salt 3a and monochloride 22a,
derived by reaction of sample a with 3 equiv of triethylamine.
of compounds 20, 21a, and 22a displayed two resonances16
which were assigned to carbamate rotamers on the basis of
temperature-dependent 31P NMR studies with compounds 20
and 22a (coalescence observed at 325 K).
To explore further the scope of this reaction, we synthe-
sized an analogue of monochloride 8 containing an ortho-
substituted N-acetyl group which was anticipated to form
via a six-membered ring, an intramolecular hydrogen bond
with the phosphonyl (see 25, Scheme 5). This experiment
was anticipated to determine whether aromatic phosphono-
chloridates react with tertiary amines, providing the phos-
phonyl oxygen is hydrogen-bonded. To this end, hydrogena-
tion of 2417 followed in turn by acetylation and hydrolysis
of the derived phosphonate 25 furnished monoester 26.
Unfortunately, conversion to the corresponding monochloride
proved impossible. Realizing that this difficulty was likely
due to nucleophilic addition of the acetyl group to the
phosphonochloridate, we targeted cyclic carbamate 31.
Toward this end, reduction of 2718 with sodium borohy-
dride and zinc chloride,19 followed by formation of the cyclic
carbamate and cross-coupling of the resulting aryl iodide with
diethyl phosphite, afforded phosphonate 28.20 Treatment of
methyl iodide in the presence of sodium hydride furnished
the N-methylated phosphonate 20; basic hydrolysis then gave
21a in 69% yield. The corresponding Fmoc analogue 21b
was obtained by hydrogenation of 21a, followed by repro-
tection of the amine with FmocCl. Conversion of 21a and
21b to the corresponding monochlorides 22a and 22b,
followed by treatment with triethylamine, did not result in
phosphonylammonium salt formation, strongly supporting
the hydrogen-bond hypothesis.
(10) Rahil, J.; Pratt, R. F. J. Chem. Soc., Perkin Trans. 2 1991, 947-
950.
(11) Yamauchi, K.; Kinoshita, M.; Imoto, M. Bull. Chem. Soc. Jpn. 1972,
45, 2531-2534.
(12) Chen, S.; Lin, C.-H.; Kwon, D. S.; Walsh, C. T.; Coward, J. K. J.
Med. Chem. 1997, 40, 3842-3850.
(13) Naylor, R. A.; Williams, A. J. Chem. Soc., Perkin Trans. 2 1976,
1908-1913.
(14) Gellman, S. H.; Dado, G. P.; Liang, G.-B.; Adams, B. R. J. Am.
Chem. Soc. 1991, 113, 1164-1173.
(15) (a) Mizushima, S.-I.; Simanouti, T.; Nagakura, S.; Kuratani, K.;
Tsuboi, M.; Baba, H.; Fujioka, O. J. Am. Chem. Soc. 1950, 72, 3490-
3494. (b) Boussard, G.; Marraud, M.; Aubry, A. Biopolymers 1979, 18,
1297-1331.
(16) Similar rotamers are observed by 13C NMR.
(17) Cadogan, J. I. G.; Sears, D. J.; Smith, D. M. J. Chem. Soc. 1969,
1314-1318.
Org. Lett., Vol. 2, No. 24, 2000
3889

ride 31 was accomplished using oxalyl chloride with catalytic
DMF. The presence of an intramolecular hydrogen-bond was
verified via variable concentration IR spectroscopy14,15 in
CHCl3 (concentration range 0.05 to 0.0025 M), as well as
Scheme 5
1
variable concentration H NMR spectroscopy (0.0025 to
0.00010 M). In the IR experiment, the only observable NH
stretch was at 3285 cm-1, indicating the hydrogen-bonding
1
was intramolecular. Likewise, during the H NMR experi-
ment, the NH resonance at 9.42 ppm did not vary across the
concentration range studied, again suggesting the hydrogen-
bonding is intramolecular. This observation is in marked
contrast to 33, obtained through a similar route from 5-iodo-
2-nitrobenzoic acid.21 The X-ray crystal structure of phos-
phonate 33 illustrates that no internal hydrogen-bond is
possible, but the IR spectra showed a non-hydrogen-bonded
NH stretch at 3422 cm-1 (5 × 10-2 to 2.5 × 10-3 M in
CHCl3), and at high concentration (above 10-2 M), a weak
intermolecularly hydrogen-bonded NH stretch at 3240 cm-1.
Treating a solution of 31 in CDCl3 with triethylamine did
not, however, result in the formation of phosphonylammo-
nium salt 32 (as observed by 31P NMR spectroscopy).
Presumably, the aromatic ring deactivates the phosphorus
atom by resonance so that, despite the presence of a
hydrogen-bond, no reaction with triethylamine occurs.
In summary, we have obtained evidence that phosphorus
activation via the participation of an intramolecular hydrogen-
bond accounts for the formation of phosphonylammonium
salts 3a,b and 12. Accordingly, phosphono mono- and
dichloridates 5, 8, 14, 17, and 22a,b that lack the ability to
form an intramolecular hydrogen-bond remain unmodified
upon triethylamine addition. We have also demonstrated that
aromatic phosphonochloridates, regardless of the presence
of an intramolecular hydrogen-bond (see 5, 8, and 31), are
not sufficiently reactive to form phosphonylammonium salts.
28 with NaOH in water to hydrolyze the phosphonate
unfortunately resulted in the cleavage of the cyclic carbamate.
This problem was circumvented by preparing ethyl benzyl
phosphonate 29 using ethyl benzyl phosphite.21 Hydro-
genolysis removed the benzyl group in 29 to furnish 30 in
97% yield. Conversion of 30 to the corresponding monochlo-
Acknowledgment. Financial support was provided by the
National Institutes of Health (National Institute of General
Medical Sciences) through Grant GM-41821. L.D. and
R.M.C. gratefully acknowledge the Swiss National Science
Foundation and the National Insitiutues of Health, respec-
tively, for postdoctoral fellowships.
(18) (a) Newman, M. S.; Logue, M. W. J. Org. Chem. 1971, 36, 1398-
1401. (b) Reissenweber, G.; Mangold, D. Angew. Chem., Int. Ed. Engl.
1981, 20, 882-883.
(19) Yamakawa, T.; Masaki, M.; Nohira, H. Bull. Chem. Soc. Jpn. 1991,
64, 2730-2734.
(20) Jaffre`s, P.-A.; Bar, N.; Villemin, D. J. Chem. Soc., Perkin Trans. 1
1998, 2083-2089.
Supporting Information Available: Spectroscopic and
analytical data for 7, 20, 21a-b, 25, 26, 28, 29, 30, and 33
as well as experimental procedures. This material is available
free of charge via the Internet at http://pubs.acs.org.
(21) Ethylammonium phosphite, prepared as in Hammond, P. R. J. Chem.
Soc. 1962, 2521-2522, was converted to ethylphenyl phosphite by treatment
with benzyl alcohol and pivaloyl chloride, similarly to Hill, J. M.; Lowe,
G. J. Chem. Soc., Perkin Trans. 1 1995, 2001-2007.
(22) Friedla¨nder, P.; Bruckner, S.; Deutsch, G. Liebigs Ann. Chem. 1912,
388, 23-49; we have only observed formation of 5-iodo-2-nitrobenzoic
acid, and no 3-iodo-2-nitrobenzoic acid.
OL0066330
3890
Org. Lett., Vol. 2, No. 24, 2000