An improved synthesis of α-phosphonoenamines based on a modified Peterson olefination

Article abstract of DOI:10.1016/j.tetlet.2007.11.066

An efficient, stereoselective method for the synthesis of α-phosphonoenamines based on a modified Peterson olefination is described. The carbanion derived from isolatable intermediate 2 reacts with aromatic or aliphatic aldehydes selectively eliminating in Peterson fashion to deliver functionally rich α-phosphonoenamines 3. The synthetic utility of these enamines is demonstrated by their hydrolysis yielding the homologous carboxylic acids in good yield.

Full text of DOI:10.1016/j.tetlet.2007.11.066

Available online at www.sciencedirect.com
Tetrahedron Letters 49 (2008) 281–285
An improved synthesis of a-phosphonoenamines based on a
modified Peterson olefination
James McNulty,a,* Priyabrata Dasa and Don Gosciniakb
aDepartment of Chemistry, McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada L8S 4M1
bCytec Industries Inc., 7910 Mount Joy Road, Mt Pleasant, TN 38474, USA
Received 27 August 2007; revised 9 November 2007; accepted 12 November 2007
Available online 17 November 2007
Abstract—An efficient, stereoselective method for the synthesis of a-phosphonoenamines based on a modified Peterson olefination is
described. The carbanion derived from isolatable intermediate 2 reacts with aromatic or aliphatic aldehydes selectively eliminating in
Peterson fashion to deliver functionally rich a-phosphonoenamines 3. The synthetic utility of these enamines is demonstrated by
their hydrolysis yielding the homologous carboxylic acids in good yield.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
yield of the desired a-phosphonoenamine intermediate
as (E)/(Z) mixtures, in part due to the sensitivity of
the enamine functionality.
Approaches towards the synthesis of a-amino phospho-
nates 1 (Fig. 1) and derivatives1,2 have been intensely
investigated over the last few decades with applications
as peptide-mimics due to their homology with the corres-
ponding a-amino acids. Not surprisingly, these com-
pounds are inhibitors of a range of proteolytic
enzymes and derivatives have now been shown to exhi-
bit a wide range of biological properties,1 including
antibacterial activity.3 On the other hand, the related
a,b-unsaturated-a-phosphonoenamines such as 3 (Fig. 1)
have received considerably less synthetic attention,4,5
although some have also been shown to possess biolog-
ical activity.6 The syntheses of these latter compounds
have been achieved mainly from a-aminodiphospho-
nates4 using carbanion chemistry, however these proce-
dures provide low (29–47%) to moderate (average 55%)
A reliable, high yielding synthesis of a-phosphonoenam-
ines 3 has hindered application-development of these
derivatives. Gross and Costisella reported the hydrolysis
of these intermediates to give homologated carboxylic
acids,4d work later extended by Degenhardt4e employing
the diphosphonate route. This appears to be a poten-
tially valuable synthetic interconversion (aldehyde to
homologated acid) although it has been rarely applied
due to the difficulty in accessing the necessary phospho-
noenamine intermediate. Finally, a significant advance
was reported by Dufrechou and co-workers wherein
excess base and trimethylsilyl chloride were employed
to trap a-aminophosphonate/aldehyde adducts under-
going in situ Peterson-type elimination.7 The dense
functionality found in a-phosphonoenamines 3 in con-
junction with the commercially available a-N,N-di-
methylaminophosphonate 1 stimulated our interest in
developing the aldol-type reaction shown (Scheme 1)
as a route to phosphonoenamines. In this Letter, we
report that while the direct reaction of enolate derivatives
of 1 with aldehydes is problematic, the trimethylsilyl
derivative of this compound 2 can be prepared and iso-
lated and that it readily enters into the reactions with
both aromatic and aliphatic aldehydes in Peterson fash-
ion to yield a-phosphonoenamines in high yield. These
intermediates undergo ready hydrolysis in aqueous
HBr extending the carbonyl-homologation strategy
O
P
O
P
R3
R4
R3
R4
N
N
R1
R1
O
O
R5
O
O
R6
R2
R2
1
3
Figure 1. General a-aminophosphonate 1 and a-phosphonoenamine 3
structures.
*
Corresponding author. Tel.: +1 905 525 9140x27393; fax: +1 905 522
2509; e-mail: jmcnult@mcmaster.ca
0040-4039/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.11.066

282
J. McNulty et al. / Tetrahedron Letters 49 (2008) 281–285
O
P
O
O
P
N
R
P
N
N
TMS-Cl
1) sec-BuLi, THF
2) R-CHO
O
O
O
SiMe3
O
O
LDA, THF
O
3a-3j, 70-87%
isolated yield
1
2 (80%)
1) MeLi
2) ArCHO
O
(EtO)2P
N
HO
Ar
Scheme 1.
O
P
(aldehyde to homologated carboxylic acid) to a range of
aromatic and aliphatic carboxylic acids.4d
N
R
Peterson
O
O
O
2. Results and discussion
P
N
3a-3j
O
O
On the one hand, the direct aldolization of various eno-
late derivatives of 1 with aromatic aldehydes or ketones
provided a-phosphonoenamines 3 only in low yield. For
example, reaction of the anion of 1 (formed with MeLi,
THF, ꢀ78 ꢁC) with 4-chlorobenzaldehyde or acetone
gave the corresponding b-hydroxyl-a-aminophospho-
nate adducts in moderate yields of 67% (two diastereo-
mers) and 60%, respectively. Attempts to dehydrate
these intermediates to the a-phosphonoenamines gave
complex mixtures resulting in poor yields of 3. The fail-
ure to selectively dehydrate these intermediates
prompted us to attempt a Peterson variant of the reac-
tion using a functionalized a-silyl carbanion.8 To this
end, we prepared the a-trimethylsilyl derivative 2
(Scheme 1) employing a silylation strategy first described
by Padwa et al. in the silylation of cyanoamines.9 In the
present case, rapid N-silylation of 1 with trimethylsilyl-
chloride gave the trimethylsilyl ammonium salt. Depro-
tonation with LDA promoted the desired N to C
migration of the trimethylsilyl group most likely via a
nitrogen ylide intermediate. The trimethylsilyl interme-
diate 2 proved to be stable to silica gel and could be
stored under argon for several days without decomposi-
tion. The intermediate was however readily hydrolyzed
back to the starting aminophosphonate 1 when treated
with aqueous acid. Formation of the a-silyl carbanion
from the above intermediate 2 using sec-BuLi in THF
(ꢀ78 ꢁC) and the addition of piperonal, as outlined in
Scheme 1, to our delight led to the formation of the cor-
responding phosphonoenamine 3a with 95% conversion
and 87% isolated yield. Although the precise mechanism
of the Peterson reaction is still unclear (so-called betaine
versus oxasiletanide intermediates)8 the present interme-
diate, shown as the betaine (Fig. 2), selected the Peter-
son elimination pathway over the Horner–Emmons
route. Vinylsilanes are also possible products from this
reaction and this ‘challenge’ unambiguously demon-
strates the prominence of the Peterson pathway,7 reac-
tivity also observed with simple a-silylphosphonates.8
SiMe3
Horner-Emmons
-
O
R
Me3Si
N
R
Betaine
vinyl silane
Figure 2. Possible pathways for elimination from the a-silylated
‘betaine-type’ intermediate.
The scope of the reaction was investigated with a range
of aromatic and aliphatic aldehydes the results of which
are summarized in Table 1.
Conversions are high in all cases with isolated yields
being only slightly lower. Under these conditions of
kinetic control, aromatic aldehydes provided a mixture
of (E)/(Z) olefins that rapidly isomerize to give exclusively
the thermodynamically more stable (E)-isomers. No sig-
nificant electronic effect was observed with both electron
rich and electron poor derivatives reacting equally well
(entries 1–6). In the case of aliphatic aldehydes, di-
hydro-cinnamaldehyde (entry 7) provided a single (E)-
adduct, most likely as the kinetic product, while other
aliphatic aldehydes gave (E)/(Z) mixtures (entries 8
and 9). These isomers could be separated and proved
to be configurationally stable. The heteroaryl aldehyde
2-furfural also yielded its corresponding adduct effi-
ciently (entry 10).
Although the reactivity of a-phosphonoenamines
remains largely been unexplored (vide supra) the obvious
synthetic potential of the homologation strategy
prompted our further interest. The use of boiling hydro-
chloric acid was reported to give carboxylic acids
through initial enamine hydrolysis and cleavage of the
derived a-ketophosphonate intermediate.4d Three exam-
ples of this homologation were reported by Gross and

J. McNulty et al. / Tetrahedron Letters 49 (2008) 281–285
283
Table 1. Peterson reaction of a-trimethylsilyl-a-dimethylamino-
Table 2. Hydrolysis of a-phosphonoenamines to provide the corre-
phosphonate with aldehydes
sponding homologous carboxylic acids
O
P
N
O
P
°
1) sec-BuLi, -78 C, THF
O
O
O
O
O
O
P
48% HBr
°
°
C
O
2) RCHO, -78 Cto 25
Si
N
R
OH
100 oC, 10 min.
R
N
O
2
3
R
3
4
Entry RCHO
Conversion Isolated yield E:Zc
(%)a
3
of 3b (%)
Entry
1
Phosphonoenamine
4 Isolated yield (%)
O
O
O
CO H
2
N
P
1
>95 3a
87
100:0
O
O
O
H
O
(2a)
(3a)
(4a) 90
H
O
O
O
O
O
O
O
CO H
2
2
3
4
(2b) >90 3b
(2c) >94 3c
(2d) >95 3d
67
80
85
100:0
100:0
100:0
N
P
H
2
3
(3b)
(3c)
(4b) 93
(4c) 95
H
O
H
O
P
COH
O
2
N
Cl
O
O
H
Cl
H
Cl
MeO
O
P
O
COH
O
O
2
N
H
5
6
(2e) >95 3e
(2f) >95 3f
82
80
100:0
100:0
4
(3d)
(4d) 98
H
O2N
O2N
O2N
O2N
O2N
O2N
O
O
P
CO H
O
H
2
N
5
6
(3e)
(3f)
(4e) 97
(4f) 90
O
H
O
7
8
9
(2g) >90 3g
(2h) >90 3h
(2i) >90 3i
75
70
70
100:0
60:40
65:35
H
O
O
N
P
COH
2
O
O
H
H
O
1). Longer heating time gives very poor yield of the
homologated acid, presumably due to rapid decarboxyl-
ation, whereas shorter heating time gives only a-keto-
phosphonate. The use of HBr under these reaction
conditions proved to be general for both electron rich
and electron deficient aryl derivatives (Table 2, entries
2–5) with high yields of carboxylic acid being obtained
in all cases. Lastly, the hexanal derived a-phosphono-
enamine 3f was readily hydrolyzed to heptanoic acid
(Table 2, entry 6) extending the scope to aliphatic deriv-
atives also.
H
O
O
10
(2j) >95 3j
75
100:0
H
a Conversion is based on the recovered aldehyde partner.
b Isolated yield of the phosphonoenamine 3 after chromatographic
purification.
c The (Z)-isomer derived of aromatic aldehydes fully converts to the
(E)-isomer, see text. In the case of aliphatic aldehydes, both config-
urational isomers were separated.
While several well-known carbonyl homologation strat-
egies are employed in synthetic chemistry, most involve
several steps to extend an aldehyde or ketone to the
homologated aldehyde. The present two-step method
appears to be potentially useful as in the overall
process, an aldehyde is converted into the homologous
carboxylic acid 4 via the intermediacy of 3. Few other
methods have been reported to effect this precise trans-
formation (aldehyde to homologated acid).10 Further-
more, the method allows efficient conversion of
common aldehydes to phenylacetic acid and substituted
derivatives which are valuable industrial intermediates
in the production of pharmaceuticals, cosmetics and
fragrances.11
Costisella4d and the route appears to have been applied
only once.4e The protocol is perhaps underutilized due
to the difficulties in accessing intermediates 3 and the
need for optimization of the conditions for enamine
hydrolysis, which produces ketophosphonate intermedi-
ates and also results in decarboxylation with b-aryl
acids.4e Investigation of the hydrolysis of adduct 3a
under various conditions revealed that high yields of
the hydrolysis product could be obtained under a tightly
held set of conditions. Treatment of the piperonal
adduct 3a with 48% HBr and gentle warming to 100 ꢁC
with a heat-gun led to rapid hydrolysis and isolation
of the homologous acid in 90% yield (Table 2, entry

284
J. McNulty et al. / Tetrahedron Letters 49 (2008) 281–285
3. Conclusion
was weighed compound 2 (74.0 mg, 0.278 mmol) under
argon and dry THF (556 lL) added to make a 0.5 M
solution. The flask was stirred for 15 min. at ꢀ78 ꢁC
whereupon sec-BuLi (239 lL, 0.334 mmol, 1.4 M stock,
C6H12) was added slowly. After 40 min, a 0.5 M solution
(in THF) of piperonal (50 mg, 0.333 mmol) was added
slowly to the reaction flask at ꢀ78 ꢁC. The flask was
kept at ꢀ78 ꢁC for 2 h and then slowly warmed to room
temperature where it was stirred for further 2 h. The
resulting mixture was concentrated to remove solvent.
Water was added (5 mL) to the residue, and the result-
ing mixture was extracted with dichloromethane
(3 · 15 mL). The combined organic layers were dried
(MgSO4), filtered and concentrated. Purification using
silica gel column chromatography (EtOAc) yielded 3a,
79.1 mg, (87%) as a yellow semisolid. Rf (EtOAc):
In conclusion, we describe the synthesis and isolation of
an a-silyldimethylaminophosphonate 2 and reaction of
its anion with aldehydes yielding a-phosphonoenamines.
The scope of the Peterson route to a-phosphonoenam-
ines 3 described herein is wide and an expanded range
of derivatives are now available using this methodology
in higher yields than previous related methods and with
high (E)-selectivity.6,7 While certain of these derivatives
have been prepared using the Horner–Emmons olefin-
ation reaction of diphosphonates,4 the greater effi-
ciency of the Peterson olefination likely derives from
the higher oxygenophilicity of silicon versus phospho-
rus,8 enabling milder reaction conditions and minimiz-
ing the decomposition of labile a-phosphonoenamines
3. An optimized protocol is described for the hydrolysis
of these intermediates in warm hydrobromic acid which
appears to be a general method for the overall conver-
sion of both aromatic and aliphatic aldehydes into the
homologous carboxylic acid derivatives. Work in our
laboratories continues with the exploration of the reac-
tivity of a-phosphonoenamines with a view to expand-
ing their chemical utility and will be reported in due
course.
0.36. 1H NMR (200 MHz, CDCl3):
d 1.35 (t,
JHH = 7.0 Hz, 6H); 2.65 (d, JPH = 2.0 Hz, 6H); 4.13
(m, 4H); 5.94 (s, 2H); 6.67 (d, JPH = 14.86 Hz, 1H);
6.75 (d, JHH = 7.8 Hz, 1H); 7.45 (s, 1H); 13C NMR
(50 MHz, CDCl3): d 16.4 (d, J = 6.8 Hz); 43.1 (d,
J = 2.3 Hz); 61.6 (d, J = 5.6 Hz); 101.1; 107.9; 109.3;
125.0; 129.4; 130.9 (d, J = 31.8 Hz); 137.8 (d,
J = 182.1 Hz); 147.3; 147.5; 31P NMR (80 MHz,
CDCl3): d 18.0; HRMS (M)+ calcd. for C15H22NO5P:
327.1237; found: 327.1236.
General procedure for Table 2: Synthesis of 4a. Into a
flame dried flask was weighed phosphonoenamine 3a
(50 mg, 0.158 mmol) and 3.0 mL of 48% HBr was
added to the flask. The flask was heated with a heat
gun for 10 min. The flask was cooled immediately.
Deposits were seen in the reaction flask, which was
extracted with diethyl ether (3 · 15 mL). The com-
bined organic layers were dried over (MgSO4),
filtered and concentrated to yield 4a, 26.8 mg, (95%);
off-white solid, mp 125–127 ꢁC; CAS registry number
[2861-28-1].
4. Experimental
Synthesis of 2. Into a 20 mL flame-dried round bottom
flask, containing a magnetic stirring bar, was added
diethyl
N0,N-dimethylaminophosphonate
(500 lL,
4.128 mmol) under argon. To this was added dry THF
(8.26 mL). The contents were cooled to ꢀ78 ꢁC and
stirred for 15 min. whereupon TMS–Cl (550 lL,
4.349 mmol) was added to the reaction flask over
5 min. Upon stirring for 30 min., a solution of LDA
(2.5 mL, 5.0 mmol, 2 M, THF) was added slowly to
the reaction flask maintained at ꢀ78 ꢁC. The flask was
kept at ꢀ78 ꢁC for 2 h and then slowly warmed to room
temperature where it was stirred for a further 6 h. The
resulting mixture was concentrated to remove solvent.
Water was added (10 mL) to the residue, and the result-
ing mixture was extracted with dichloromethane
(3 · 15 mL). The combined organic layers were dried
(MgSO4), filtered and concentrated. The product,
Rf = 0.32 (EtOAc, pink-red to ninhydrin), was purified
by silica gel column chromatography (EtOAc). The sil-
ica gel was neutralized by adding five drops of triethyl-
amine into the initial silica gel slurry. The title
compound 2, 882.2 mg (80%) was isolated as a brown
Acknowledgements
We thank NSERC, Cytec Canada Inc. and McMaster
University for financial support of this work.
References and notes
1. For some recent examples, see: (a) Wu, J.; Sun, X.; Xia,
H.-G. Green Chem. 2006, 8, 365–367; (b) Palacios, F.;
Ochoa de Retana, A. M.; Alonso, J. M. J. Org. Chem.
2005, 70, 8895–8901; (c) Azizi, N.; Saidi, M. R. Tetra-
hedron 2003, 59, 5329–5332; (d) Steere, J. A.; Sampson, P.
B.; Honek, J. F. Bioorg. Med. Chem. Lett. 2001, 12, 457–
460; (e) Kafarski, P.; Lejczak, B. Curr. Med. Chem. Anti
Cancer Agents 2001, 1, 301–312.
2. Li, S.; Whitehead, J. K.; Hammer, R. P. J. Org. Chem.
2007, 72, 3116–3118.
3. Pratt, R. F. Science 1989, 246, 917–919.
4. (a) Qian, D. Q.; Shi, D. S.; Cao, R. Z.; Liu, L. Z.
Heteroatom Chem. 1999, 10, 271–276; (b) Costisella, B.;
Keitel, I.; Gross, H. Tetrahedron 1981, 37, 1227–1232; (c)
Ahlbrecht, H.; Farnung, W. Synthesis 1977, 336–338; (d)
Gross, H.; Costisella, B. Angew. Chem., Int. Ed. Engl.
1
oil. H NMR (200 MHz, CDCl3): d 0.14 (s, 9H); 1.33
(t, JHH = 6.4 Hz, 6H); 2.38 (d, JPH = 22.2 Hz, 1H);
2.47 (s, 6H); 4.07 (m, 4H); 13C NMR (50 MHz, CDCl3):
d ꢀ0.6 (d, J = 2.5 Hz); 16.5 (dd, J = 6.30, 2.5 Hz); 45.5
(d, J = 3.8 Hz); 55.3 (d, J = 112.5 Hz); 60.7 (t,
J = 7.8 Hz); 31P NMR (80 MHz, CDCl3): d 30.75;
HRMS (M)+ calcd. for C10H26N1O3SiP: 267.1421,
found: 267.1420.
General procedure for Table 1: Synthesis of 3a. Into a
flame-dried flask, containing a magnetic stirring bar,

J. McNulty et al. / Tetrahedron Letters 49 (2008) 281–285
285
1968, 7, 391–392; (e) Degenhardt, C. R. Synth. Comm
1982, 12, 415–421.
5. Schindler, N.; Ploger, W. Chem. Ber. 1971, 104, 2021–
2022.
6. Zumstein, F.; Klingseisen, F. Ger. Offen. 1990, DE
4029444 A1.
8. van Staden, L. F.; Gravestock, D.; Ager, D. J. Chem. Soc.
Rev. 2002, 31, 195–200; Ager, D. J. Org. React. 1990, 38, 1.
9. Padwa, A.; Eisenbarth, P.; Venkatramanan, M. K.; Wong,
G. S. K. J. Org. Chem. 1987, 52, 2427–2432.
10. Zhou, G. B.; Zhang, P. F.; Pan, Y. J. Tetrahedron 2005,
61, 5671–5677.
7. Dufrechou, S.; Combret, J. C.; Malhiac, C.; Collignon, N.
Phosphorus, Sulphur Silicon 1997, 127, 1–14.
11. Kohlpaintner, C. W.; Beller, M. J. Mol. Catal. A: Chem.
1997, 116, 259–267.