Synthesis and Structural Characterization of Trivalent Amino Acid Derived Chiral Phosphorus Compounds

Article abstract of DOI:10.1021/jo035508+

Reaction of the N-toluenesulfonyl derivatives of (S)-alanine, phenylalanine, and valine (4-6) with PhPCl₂ gave in high yield the 4-methyl, benzyl, and isopropyl derivatives (7-9) of 2-phenyl-1-p-toluenesulfonyl-1,3,2-oxazaphospholidin-5-one. Th

Full text of DOI:10.1021/jo035508+

Syn th esis a n d Str u ctu r a l Ch a r a cter iza tion of Tr iva len t Am in o
Acid Der ived Ch ir a l P h osp h or u s Com p ou n d s
,
†
†
†
†
†
William H. Hersh,* Ping Xu, Cheslan K. Simpson, J onathan Grob, Brian Bickford,
†
,§
‡
‡,|
Mohammad Salman Hamdani, Thomas Wood, and Arnold L. Rheingold
Department of Chemistry and Biochemistry,
Queens College and the Graduate Center of the City University of New York,
Flushing, New York 11367-1597, and Department of Chemistry and Biochemistry,
University of Delaware, Newark, Delaware 19716
william_hersh@qc.edu
Received October 13, 2003
Reaction of the N-toluenesulfonyl derivatives of (S)-alanine, phenylalanine, and valine (4-6) with
PhPCl gave in high yield the 4-methyl, benzyl, and isopropyl derivatives (7-9) of 2-phenyl-1-p-
2
toluenesulfonyl-1,3,2-oxazaphospholidin-5-one. The ratios of the (2S,4S)/(2R,4S) diastereomers (cis/
trans isomers) were 1:1, 2:1, and 10:1 for the methyl, benzyl, and isopropyl derivatives 7a ,b, 8a ,b,
and 9a ,b, respectively. For 7a ,b, both isomers could be crystallized, but for the others only the
major isomers were isolable. The X-ray crystal structure of 9a shows that the isopropyl and phenyl
groups are mutually cis and that the tolyl moiety is oriented s-trans to both the isopropyl and
phenyl groups. Reaction of 6 with Cl
trans, and trans/trans diphosphorus heterocycles 11a -c. The major isomer could be crystallized
and isolated free of the other diastereomers. Reaction of 6 with EtPCl gave a 6:1 mixture of cis/
2 2 2 2
PCH CH PCl (10) gave a 56:38:7 mixture of the cis/cis, cis/
2
trans isomers of the ethyl-substituted heterocycles 12a ,b as an inseparable oil but allowed
confirmation of the structure of 11a . Slow epimerization at phosphorus may occur by inversion
but more likely by ring opening/closure, since 7b, 9a , and 11a give rise upon standing in solution
to mixtures containing starting material and 7a , 9b, and 11b, respectively, along with the free
amino acid derivatives 4 and 6. The NMR spectra, and in particular the coupling constants between
the R-hydrogen atom of the amino acid moiety and phosphorus, were used to establish the identities
of the cis and trans isomers. Reaction of 9a with (THF)W(CO)
5
gave the phosphorus-ligated adduct
(
9a )W(CO) (13), and the IR spectrum of this complex shows that 9a is a strongly electron-
5
1
withdrawing ligand. The geometry of the sulfonamide moiety is discussed in detail, as are the H
NMR coupling constants. The data are consistent with the presence of little steric interaction
between the cis isopropyl and phosphorus substituent in 9a , 11a , and 12a and orientation of the
tolyl moiety s-cis to the isopropyl group in 9b, 12b, and 13.
have great utility,1
1-14
they are electron-donating and so
In tr od u ction
are poor choices in the design of chiral Lewis acids.
K u¨ ndig reported one solution to this problem, describing
the use of a relatively electron-deficient fluorinated diaryl
phosphinite ligand in asymmetric Diels-Alder catalysis
Organometallic Lewis acids show promise as catalysts
for reactions such as the Diels-Alder and aldol reac-
tions,1 but the critical combination of high Lewis acidity
and construction of a chiral reaction site is difficult to
achieve. Phosphines are the auxiliary ligands of choice
-9
(4) J aquith, J . B.; Guan, J . Y.; Wang, S. T.; Collins, S. Organome-
for low oxidation state carbonyl complexes, but while
tallics 1995, 14, 1079-1081.
chiral ligands10 such as DiPAMP, BINAP, and DuPHOS
(5) Hollis, T. K.; Bosnich, B. J . Am. Chem. Soc. 1995, 117, 4570-
4
581.
(
6) Evans, D. A.; Murry, J . A.; Kozlowski, M. C. J . Am. Chem. Soc.
†
Queens College.
University of Delaware.
We dedicate this paper to the memory of Mohammad Salman
1996, 118, 5814-5815.
‡
(7) Pignat, K.; Vallotto, J .; Pinna, F.; Strukul, G. Organometallics
2000, 19, 5160-5167.
§
Hamdani. Sal worked in our labs at Queens College as an undergradu-
ate student from 1998 to 2001 and died as part of the rescue effort at
the World Trade Center in New York City, September 11, 2001.
(8) Faller, J . W.; Grimmond, B. J . Organometallics 2001, 20, 2454-
2458.
(9) Huang, Y.; Iwama, T.; Rawal, V. H. J . Am. Chem. Soc. 2002,
124, 5950-5951.
|
Current address: Department of Chemistry and Biochemistry,
University of California, San Diego, 9500 Gilman Drive, La J olla, CA
(10) Pietrusiewicz, K. M.; Zablocka, M. Chem. Rev. 1994, 94, 1375-
9
2093-0358. E-mail: arheingold@ucsd.edu.
1) Bonnesen, P. V.; Puckett, C. L.; Honeychuck, R. V.; Hersh, W.
H. J . Am. Chem. Soc. 1989, 111, 6070-6081.
2) Odenkirk, W.; Rheingold, A. L.; Bosnich, B. J . Am. Chem. Soc.
992, 114, 6392-6398.
3) Hollis, T. K.; Odenkirk, W.; Robinson, N. P.; Whelan, J .; Bosnich,
B. Tetrahedron 1993, 49, 5415-5430.
1411.
(
(11) Knowles, W. S. Acc. Chem. Res. 1983, 16, 106-112.
(12) Noyori, R. Tetrahedron 1994, 50, 4259-4292.
(13) Burk, M. J .; Gross, M. F.; Harper, T. G. P.; Kalberg, C. S.; Lee,
J . R.; Martinez, J . P. Pure Appl. Chem. 1996, 68, 37-44.
(14) Nugent, W. A.; Rajanbabu, T. V.; Burk, M. J . Science 1993, 259,
479-483.
(
1
(
1
0.1021/jo035508+ CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/18/2004
J . Org. Chem. 2004, 69, 2153-2163
2153

Hersh et al.
by chiral iron and ruthenium Lewis acids.15-17 Corey
reported a solution for a non-transition-metal Lewis acid,
in which higher reactivity was imparted by the presence
SCHEME 1
of a cationic nitrogen center adjacent to the boron Lewis
acid site.1
8,19
We recently discovered a class of trivalent
phosphorus compounds in which one or two electron-
withdrawing N-sulfonyl moieties are attached to phos-
phorus.20 As shown in eq 1, for instance, the compound
that we have named TosL is readily prepared from the
corresponding sulfonamide and phosphorus dichloride,
despite the presumed low nucleophilicity of the nitrogen
lone-pairs. Compounds such as TosL can serve as ligands
2
0
that are exceptionally electron-withdrawing, and con-
sistent with this electronic property, these phosphorus
compounds function as effective promoters of rhodium-
catalyzed hydroformylation.2
1,22
Related compounds that have been reported in the
literature suggested an obvious route to the preparation
of a chiral analogue of TosL (Scheme 1). Thus, J ug e´ and
Genet and separately Brown have described the use of
ephedrine to prepare the 1,3,2-oxazaphospholidine 1a.23-25
render the oxygen much more electron withdrawing than
that in ephedrine, and more importantly, the primary
amine is readily sulfonylated to give the electron-
withdrawing sulfonamide moiety. A number of amino
acid-derived heterocycles that contain boron have been
described,28 of which the 1,3,2-oxazaborolidinone 2a is a
In this system, chirality is efficiently and elegantly
transmitted from ephedrine to phosphorus to give a
diastereomerically pure material that contains a stereo-
genic phosphorus atom; presumably the diastereomer in
which the phenyl groups are mutually cis is disfavored
due to the higher steric interaction. The stereochemistry
2
9
well-known Diels-Alder catalyst. Unlike the phospho-
rus center in TosL and 1a , the trivalent boron is an
electron acceptor rather than donor, and its geometry is
planar. Enone complexation to give 2b is proposed to
occur anti to the toluenesulfonyl moiety and includes a
π-acceptor interaction with the electron-rich indole ring.
has been assigned on the basis of the X-ray structures of
the oxide26 and borane27 derivatives 1b,c. We are not
aware of this ligand having been tested for its electronic
properties in transition-metal adducts, but a prediction
of its electronic character would be difficult since the
3
0
The iridium-containing 1,3,2-oxazirolidinone 3a is a
rare transition-metal heterocycle, and like the boron
center in borolidinone 2a the iridium center is an electron
pair acceptor and is coplanar with the attached nitrogen,
oxygen, and the centroid of the Cp* ligand. However,
attack by phosphine ligands to give 3b occurs so that the
dialkylamino group is electron-donating while the oxygen
is electron-withdrawing.20
Amino acids possess the same â-amino hydroxy func-
tionality as ephedrine, but the carboxy group should
(
15) K u¨ ndig, E. P.; Bourdin, B.; Bernardinelli, G. Angew. Chem.,
Int. Ed. Engl. 1994, 33, 1856-1858.
16) Kundig, E. P.; Saudan, C. M.; Bernardinelli, G. Angew. Chem.
999, 38, 1220.
17) Kundig, E. P.; Saudan, C. M.; Alezra, V.; Viton, F.; Bernar-
dinelli, G. Angew. Chem. 2001, 40, 4481-4485.
18) Hayashi, Y.; Rohde, J . J .; Corey, E. J . J . Am. Chem. Soc. 1996,
18, 5502-5503.
19) Ryu, D. H.; Lee, T. W.; Corey, E. J . J . Am. Chem. Soc. 2002,
24, 9992-9993.
20) Hersh, W. H.; Xu, P.; Wang, B.; Yom, J . W.; Simpson, C. K.
Inorg. Chem. 1996, 35, 5453-5459.
21) Magee, M. P.; Luo, W.; Hersh, W. H. Organometallics 2002,
1, 362-372.
22) Magee, M. P.; Li, H.-Q.; Morgan, O.; Hersh, W. H. Dalton Trans.
3
large PR and phenyl groups are mutually trans on the
(
heterocycle as seen for the two phenyl groups in the
ephedrine case (1a ), as well as trans to the toluenesulfo-
nyl group as seen in the boron case (2a ).
In this paper, we describe (1) the synthesis of phos-
phorus analogues of 2a and 3a , (2) an example of the
high diastereoselectivity at phosphorus like that seen for
1
(
(
1
1
(
(
1
a , (3) extension to the synthesis of a chiral diphosphorus
analogue, (4) a rare example of an X-ray structure of a
chiral trivalent phosphorus compound, and (5) the high
electron-withdrawing ability of this new moiety as judged
(
2
(
2
003, 387-394.
(
(
5
by the IR spectrum of the W(CO) adduct, like that seen
23) J ug e´ , S.; Genet, J . P. Tetrahedron Lett. 1989, 30, 2783-2786.
24) J ug e´ , S.; Stephan, M.; Laffitte, J . A.; Genet, J . P. Tetrahedron
for TosL. Part of this work has been communicated
Lett. 1990, 31, 6357-6360.
25) Brown, J . M.; Carey, J . V.; Russell, M. J . H. Tetrahedron 1990,
6, 4877-4886.
26) Schwalbe, C. H.; Chopra, G.; Freeman, S.; Brown, J . M.; Carey,
J . V. J . Chem. Soc., Perkin Trans. 2 1991, 2081-2090.
27) J ug e´ , S.; Stephan, M.; Genet, J . P.; Halut-Desportes, S.;
31
previously.
(
4
(
(28) Deloux, L.; Srebnik, M. Chem. Rev. 1993, 93, 763-784.
(29) Corey, E. J .; Loh, T. P. J . Am. Chem. Soc. 1991, 113, 8966-
(
8967.
J eannin, S. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1990,
C46, 1869-1872.
(30) Grotjahn, D. B.; Groy, T. L. Organometallics 1995, 14, 3669-
3682.
2
154 J . Org. Chem., Vol. 69, No. 6, 2004

Amino Acid Derived Phosphorus Compounds
SCHEME 2
2 2
reproducibility in CH Cl is puzzling since all reactions
were carried out with strict exclusion of air and water
but we presume the cause is the presence of what might
+
-
be variable concentrations of the weak acid Et
which is more soluble in CH Cl than in ether and
toluene. In ether and toluene, by contrast, both the
3
NH Cl ,
2
2
+
-
3
starting amino acid carboxylate and the Et NH Cl are
fairly insoluble. One run was carried out in THF at room
temperature, and it exhibited solubility characteristics
similar to CH
2 2
Cl and gave a relatively low 6.2:1 dias-
tereomer ratio.
Resu lts
Comparison of diastereomer ratios for the three reac-
tions was most reproducibly made in toluene at room
temperature, with the alanine derivative giving a 1:1
ratio, phenylalanine a 2:1 ratio, and valine a 10:1 ratio.
A single crystallization of the mixture of diastereomers
of 9 from ether at -35 °C reproducibly gave a 50% yield
of the major isomer 9a free of any detectable minor
isomer 9b. Subsequent recrystallizations gave the major
isomer contaminated by small amounts of the minor
isomer, but absent the need for large amounts of the pure
diastereomer, no attempt to optimize the yield has been
made. To identify the minor product 9b, fractional
crystallization was attempted, since the compounds are
too polar to move on TLC. No crystallization solvent could
be found that would selectively precipitate 9b and so
numerous crystallizations from ether were carried out
to remove 9a . Since this crystallization also resulted in
increasing amounts of precipitation of 9b as its concen-
tration in the residue increased, the best purification we
could achieve was a 1:1 ratio of 9a /9b. This was sufficient
to identify all peaks due to 9b in the H and C NMR
spectra, and all are simply slightly shifted from those due
to 9a . Elemental analysis of the mixture was consistent
with 9b being an isomer of 9a , and the optical rotation
of the mixture showed that 9b as expected has a non-
zero rotation. There is no reason to suppose that this
compound is anything other than the stereoisomer of 9a ,
in which the isopropyl and phenyl groups are trans rather
than cis to each other (vide infra).
Syn th esis. Following a procedure similar to that used
2
0
for the synthesis of TosL, triethylamine was first added
to a solution of the N-toluenesulfonyl derivatives of
S-alanine (4), phenylalanine (5), and valine (6) in toluene,
each giving a white suspension of what is presumably
+
the Et
3
NH salt of the corresponding sulfonamido car-
2
boxylate. A solution of PhPCl was then added to the
stirred suspensions, resulting in a visible change to a
merely cloudy white suspension, presumably due to the
+
-
3
less voluminous two equivalents of Et NH Cl . Stirring
was continued for up to 3 h while the reactions were
monitored by 31P NMR spectroscopy, after which filtra-
tion of the ammonium salt followed by solvent removal
gave the products as pale yellow solids in greater than
9
0% yield (Scheme 2). Analysis of the reaction mixtures
3
1
by P NMR spectroscopy indicated the presence of two
new phosphorus-containing compounds with peaks near
1
30 ppm, subsequently identified as the desired diaster-
eomeric cis and trans-1,3,2-oxazaphospholidinones 7-9,
along with any excess PhPCl at 161.5 ppm. The ratio of
1
13
2
the new compounds was similar in the reaction solutions
to that observed after workup, indicating that no signifi-
cant isomerization occurred over a short time.
The highest diastereomer ratio was observed for valine
derivative 9, but was somewhat variable in the solvent
initially used, ether, ranging from a low of 6.2:1 to a high
of 9.8:1 for room temperature reactions, although typi-
cally the ratio was 7 to 9:1. The product ratios determined
by 31P NMR spectroscopy were confirmed by analysis of
For the other mixtures, the relatively low diastereomer
ratios resulted in more difficulty in selective crystalliza-
tion, but in the case of the alanine derivatives 7a and
7b each isomer could be crystallized free of the other,
although not in high yield. No prolonged attempt was
made to isolate the minor isomer of 8.
1
the more complex H NMR spectra. Approximately 20
experiments were conducted in which the reaction sol-
vent, time (up to 3 days), temperature, and reactant
concentrations were varied, but no correlations with
diastereoselectivity emerged. For instance, while a 2.7:1
ratio was observed in refluxing ether, lowering the
reaction temperature to -35 °C gave a 6.3:1 ratio, lower
X-r a y Str u ctu r e. Crystals of 9a suitable for X-ray
diffraction were obtained on the first recrystallization
from ether. The absolute configuration was assumed to
be (S) from the naturally occurring amino acid carbon
than most room temperature runs. Use of methylene
+
3
chloride, a solvent in which the initially formed Et NH
3
1
salt of the sulfonamido carboxylate was completely
atom, and the refinement was consistent with this. The
structure is shown in Figure 1, and selected bond lengths,
angles, and torsional angles are given in the Supporting
Information for comparison to related compounds. The
key results are stereochemical: the isopropyl and phenyl
groups are cis to each other, and the tolyl moiety is in
the s-trans conformation relative to these 1,3-substitu-
ents on the heterocycle.
+
-
3
soluble and the final product Et NH Cl was partially
soluble, gave results that were even less reproducible.
Room-temperature reactions gave ratios from 3.3 to 10.6:
1
, a -78 °C reaction gave a 6.1:1 ratio, and a reaction
conducted at reflux gave a 4.3:1 ratio. Most of the room-
temperature reactions in CH Cl gave ratios greater than
:1, but the greater reproducibility of the ether and
2
2
6
toluene reactions made these the solvents of choice, with
toluene giving the most reproducible ratios but ether
giving the easiest synthesis in reliable yield. The lack of
Syn th esis of a Ch ir a l Dip h osp h in e. Reaction of 6
with tetrachlorodiphosphine 10 in THF gave a mixture
that after 5 min of reaction exhibited four peaks in the
3
1
P NMR spectrum: a singlet at 156.7 ppm, a 1:1 pair of
3
doublets at 159.2 and 154.5 ppm with J PP ) 6.8 Hz, and
(31) Hersh, W. H.; Xu, P.; Simpson, C. K.; Wood, T.; Rheingold, A.
L. Inorg. Chem. 1998, 37, 384-385.
a singlet at 159.3 ppm, in a 56:38:7 molar ratio of
J . Org. Chem, Vol. 69, No. 6, 2004 2155

Hersh et al.
there was no immediate change (a sample withdrawn
after 0.5 h contained e0.2% of the trans isomer 9b), by
the end of 1 week, the diastereomer ratio of 9a :9b was
8
9:11, and it remained constant as continued decomposi-
tion to give increasing amounts of 6 occurred. Since
samples enriched in 9b might isomerize back to this 89:
1
1 mixture, we attempted to optimize the yield of 9a by
allowing the mixtures to stand at room temperature in
solution after crystallizing out the 9a . In fact, a 65:35
mixture of 9a /9b did isomerize to an 86:14 mixture after
one week in toluene, but decomposition to 6 was competi-
tive with the isomerization reaction, so no additional 9a
could be crystallized out. Finally, the solid 89:11 mixture
was allowed to stand at ambient temperature over a
F IGURE 1. ORTEP drawing of 9a .
3
-week period in order to see if it might be converted to
SCHEME 3
a single isomer as has been seen during the elegant self-
resolution process described by Vedejs,3 but in our case
no isomerization occurred.
2,33
A pure sample of 7b exhibited similar behavior. After
1
day at 20 °C in toluene, the ratio of 7a /7b was 4:96,
and after 1 week, 23:77, but like 9 accompanied by
analogous decomposition to 4. The diphosphorus com-
pound 11a seemed to isomerize faster than this albeit in
3
CDCl : after 1 day at room temperature, a 90.2:9.3:0.5
mixture of 11a /11b/11c was observed, and after 3 days
an 88:12 mixture of 11a /11b with about an equal amount
of 6 was observed, but with no detectable 11c.
NMR Sp ectr a . Peak assignments for 7-9, 11, and 12
were made by a combination of 2D COSY and HETCOR
experiments and for the R-ring hydrogen by selective
homonuclear decoupling experiments. In addition, selec-
tive heteronuclear decoupling of the ortho phenyl hydro-
gens resulted in collapse of the broad phosphorus signal
(in the undecoupled spectrum) to a doublet, allowing
confirmation of coupling of the phosphorus and ring
hydrogen. While the determination of the cis/trans ster-
eochemistry was made on the basis of the X-ray structure
of 9a , the NMR coupling constant data allowed all of the
other stereochemistries to be determined with reference
to this one compound. The key observation is that all of
diphosphines. When the reaction was carried out in
toluene, the three compounds were observed to form in
a 49:42:9 molar ratio (Scheme 3). The major product is
proposed to be cis/cis diastereomer 11a while the minor
product is proposed to be trans/trans diastereomer 11c,
each of which would give a singlet in the 31P NMR
spectrum. The last product, cis/trans diastereomer 11b,
would be expected to exhibit a pair of doublets due to
the cis isomers exhibit 3
J
) 3.4-3.7 Hz while all of
PH
PH
3
the trans isomers exhibit J ) 1.4-1.8 Hz for the ring
hydrogen. In addition, the trans isomers for valine
derivatives 9b, 11b,c, and 12b each exhibit, compared
to the cis isomers, both an upfield shift of one of the
isopropyl methyl groups that suggests anisotropic shield-
ing by the toluenesulfonyl moiety, and a concomitant
3
-bond phosphorus-phosphorus coupling as seen. Since
the 2-carbon bridge linking the two phosphorus atoms
is different from the phenyl group on phosphorus of 7-9,
3
6
was combined with EtPCl
2
as well (Scheme 3), to better
reduction in J HH from ∼7 Hz to 3 Hz, suggesting a
model the alkane linker, to give the cis and trans isomers
2a ,b. The initial reaction mixture exhibited two peaks
conformational rotation of the isopropyl group. The
W(CO)5 adduct of 9a , that is, 13, exhibited similar
changes in the isopropyl group compared to 9a . These
and other data characteristic of the cis/trans isomers are
collected in tabular form in the Supporting Information.
Both the decoupled and undecoupled NMR spectra of
the diphosphine 11a were complicated by the presence
1
3
1
in the P NMR spectrum at 163.7 and 169.9 ppm in a
:1 ratio due to the cis and trans isomers. In contrast to
6
the other reactions, neither isomer could be crystallized
and the mixture that was spectroscopically characterized
was a clear oil.
Sta bility. The air, water, and thermal stability of 9a
is surprisingly high. Crystalline 9a exhibited no decom-
3
of virtual coupling, and the critical coupling constant J PH
could not be measured directly. The undecoupled 1H
1
position upon exposure to air overnight, as judged by H
signal was an apparent doublet of triplets, and simulation
of this multiplet as an ABXX′ system (where the phos-
phorus atoms are the XX′ pair) gave an excellent fit using
and 31P NMR spectroscopy, but roughly half decomposed
in CDCl
product analysis has been carried out. Addition of water
solution in the air gave only
3
solution upon overnight exposure to air; no
3
3
the observed values J HH ) 5.8 Hz and J PH ) 3.6 Hz
(1-15 equiv) to a CDCl
3
traces of decomposition within 1 h. A sample of 9a was
(
32) Vedejs, E.; Donde, Y. J . Am. Chem. Soc. 1997, 119, 9293-9294.
stored at 18 °C in toluene in a glovebox for 1 week. While
(33) Vedejs, E.; Donde, Y. J . Org. Chem. 2000, 65, 2337-2343.
2
156 J . Org. Chem., Vol. 69, No. 6, 2004

Amino Acid Derived Phosphorus Compounds
assumed to be double the observed “triplet” coupling),
(
SCHEME 4
3
3
for values of J PP′ g 15 Hz. In fact the value J PP′ ) 20.4
Hz was indirectly determined by simulation of the 13
ABX signal for the bridging methylene carbon, where the
C
1
3
single C renders the phosphorus atoms chemically
inequivalent (the AB pair) and gives a 6-line pattern.34
Further confirmation of the stereochemical assignment
came from the isopropyl methyl chemical shifts: both are
“normal” in 11a (1.22, 1.15 ppm), but the mixture of all
three stereoisomers exhibited doublets at 0.39 and 0.375
ppm for 11c and 11b, respectively, characteristic of the
high-field signals of the trans isomers 9b and 12b.
Coor d in a tion to W(CO)
properties of 9a , as we have done before20 it was
combined with (THF)W(CO) to give the adduct 13 (eq
) in 63% yield.
5
. To assess the electronic
5
2
Coordination of 9a to tungsten via phosphorus was
evident from the change in the 31P NMR chemical shift
from 133.0 ppm in 9a to 146.9 ppm in 13, as well as from
the presence of the characteristic tungsten-phosphorus
1
83
satellites due to the 14% natural abundance of
W,
1
1
giving J PW ) 346 Hz. The H NMR exhibited features of
both 9a and 9b, with a phosphorus-hydrogen coupling
constant for the ring methine characteristic of 9a ( J PH
3
)
3.4 Hz) but also a hydrogen-isopropyl hydrogen cou-
pling constant for the ring methine characteristic of 9b
3
(
J HH ) 3.4 Hz). Also like 9b, one of the isopropyl methyl
13
doublets was shifted upfield to 0.32 ppm. In the C NMR
spectrum, the tungsten carbonyl ligands exhibited the
2
2
characteristic trans ( J PC ) 38.4 Hz) and cis ( J PC ) 8.5
Hz) phosphorus-carbon coupling constants of phosphine
tungsten pentacarbonyl complexes.20 The infrared spec-
ranes such as 14b have been prepared by base-free
elimination of TMSCl from the phosphorus chloride and
trum exhibited unusually high-frequency carbonyl stretch-
ing frequencies indicative of an electron-withdrawing
ligand, along with the anhydride-like carbonyl stretch for
N,O-bis(trimethylsilyl)-R-amino acids,3
7-40
and while again
diastereomeric purity of the products has not been
discussed, one spectrum was published and showed the
ratio to be close to unity.40 Synthesis of trivalent phos-
phorus heterocycles by substitution of glycine derivatives,
again in a manner similar to that seen in Scheme 2, has
also been described; the 31P NMR chemical shift of the
glycine-derived 2-phenyl substituted 1,3,2-oxazaphos-
-
1
the P-OCdO moiety at 1811 cm
.
Discu ssion
Su r vey of 1,3,2-Oxa za p h osp h olid in on e Syn th esis,
Ster eoch em istr y, a n d Sta bility. Only a limited num-
ber of syntheses of 1,3,2-oxazaphospholidinones, formally
derived from R-amino acids, as well as one diastereose-
lective synthesis of a related 1,3,2-oxazaphospholidin-4-
one, have been reported (Scheme 4). The first to be
reported were prepared from amino acids via substitution
reactions at phosphorus similar to those in Scheme 2, to
4
1
pholidinone 15 is very similar to those of 7-9. Hetero-
cycle 15 has also been prepared by a room-temperature
4
2
exchange reaction of germanium heterocycle 16. Both
a trivalent and pentavalent phosphorus heterocycle have
been prepared by elimination reactions from acyclic
precursors. While 17 forms at 150 °C and so demon-
give a number of related spirophosphoranes such as
4a ,3
5,36
(37) Fu, H.; Tu, G.-Z.; Li, Z.-L.; Zhao, Y.-F.; Zhang, R.-Q. J . Chem.
Soc., Perkin Trans. 1 1997, 2021-2022.
1
and while these first examples included chiral
amino acids, diastereomeric purity of the products was
not discussed nor was data reported that could allow the
de to be evaluated. More recently, related spirophospho-
(
38) Zhang, N. J .; Lu, H. Y.; Chen, X.; Zhao, Y. F. Synthesis 1998,
376-378.
(
39) Fu, H.; Tu, G.-Z.; Li, Z.-L.; Zhao, Y.-F. Synthesis 1998, 855-
8
58.
(
40) Lu, K.; Tu, G.-Z.; Guo, X.-F.; Sun, X.-B.; Liu, Y.; Feng, Y.-P.;
(
(
34) Hersh, W. H. J . Chem. Educ. 1997, 74, 1485-1489.
Zhao, Y.-F. J . Mol. Struct. 2002, 610, 65-72.
(41) Lamande, L.; Munoz, A. Phosphorus, Sulfur Silicon 1993, 75,
241-244.
(42) Lavayssiere, H.; Dousse, G.; Satge, J . Phosphorus, Sulfur
Silicon 1990, 53, 411-422.
35) Garrigues, B.; Munoz, A.; Koenig, M.; Sanchez, M.; Wolf, R.
Tetrahedron 1977, 33, 635-643.
36) Munoz, A.; Garrigues, B.; Wolf, R. Phosphorus Sulfur 1978, 4,
7-52.
(
4
J . Org. Chem, Vol. 69, No. 6, 2004 2157

Hersh et al.
4
3
strates the potentially high stability of the P-O linkage,
temperature is most reasonably attributed to the pres-
ence of unpredictable amounts of acid in solution at this
epimerization stage. Once the heterocycle forms from the
rapidly inverting chlorophosphine, however, we propose
that the ring closure is essentially irreversible, and
monitoring of reactions at early reaction time has never
given any indication of different stereoselectivities.
4
4
1
8 only forms in low yields. Elimination of 2,3-pro-
panediol to give cyclic glycine derivatives such as 19
occurs under mild conditions.45 The final reaction in
Scheme 4 was reported to give cis-20 as the major product
at -5 °C (>2:1 cis/trans), but upon refluxing, trans-20
was obtained as the major product (1:5 cis/trans).46
As this survey makes clear, transfer of chirality from
the organic fragment to phosphorus is not observed
An alternative explanation for thermodynamic product
formation is intramolecular isomerization by inversion
at phosphorus; while there is no evidence for it in any of
these cases, it is not precluded. For instance, measured
barriers to phosphorus inversion are relatively low in a
(
14a ,b) or has not been sought (15-19) in most cases.
Only for 20 where moderate selectivity was observed and
of course for ephedrine (1a , Scheme 1) has diastereose-
lectivity been reported previously. The synthesis of
ephedrine-derived 1b,c has been reported to give a 1:1
mixture of cis and trans diastereomers initially, followed
by isomerization to give exclusive formation of the trans
49
50
1
,3,2-dioxaphospholane as well as in acylphosphines
where the electron-withdrawing acyl group allows anal-
ogy to the carboxylate and sulfonamide groups here.
Nevertheless, epimerization of 7 and 9 is always ac-
companied by decomposition with formation of the free
N-toluenesulfonyl amino acid, so the isomerization here
might most reasonably arise via P-O or P-N cleavage
followed by ring-closure. Regardless of mechanism, the
observed isomerization of 7b to 7a , and of the 65:35
mixture of 9a /9b back toward the 89:11 mixture, suggests
the cis diastereomer is indeed thermodynamically favored
for these N-sulfonyl-1,3,2-oxazaphospholidinones.
diastereomer 1a before oxidation or reaction with BH
3
as shown in Scheme 1. Both in this case and that of 20,
formation of the trans product is evidently favored
thermodynamically, in apparent contrast to the cis-
selectivity for N-toluenesulfonyl heterocycles 8, 9, 11, and
1
2. With the exception of this study, no attempts to
correlate steric effects with diastereoselectivity at phos-
phorus have been reported. Clearly, the correlation of cis/
trans ratio with steric bulk here suggests a mechanism
by which the cis product is favored: the toluenesulfonyl
moiety adopts a conformation that is predominantly
s-trans to the R-amino acid substituent, so that if P-O
bond formation occurs first due to the higher acidity of
the carboxylic acid, then subsequent nucleophilic attack
of nitrogen on phosphorus occurs so as to minimize the
steric interaction of the toluenesulfonyl and the phenyl-
phosphorus groups, leaving the larger R-amino acid
substituent and phenyl groups mutually cis. For the
isopropyl moiety, not only was a 10:1 cis:trans ratio
observed for the phenyl moiety of 9, but a 6:1 cis:trans
ratio was observed for the smaller ethyl moiety of 12.
For 11, the observed 56:38:7 ratio would be generated
by a 3:1 cis/trans ratio for each of the two ring-forming
steps pitting the isopropyl group against the two-carbon
bridge, a ratio which seems low given the size of this
group, but we have no explanation for this result.
One might suppose that the P-O-C(dO)R linkage,
which is rigorously a mixed anhydride formed from the
corresponding phosphorus acid chloride and carboxylic
acid would be only somewhat less reactive than a PCl
linkage, albeit with the potential for added instability due
to Arbuzov-type rearrangement due to the presence of
51
51-56
the P-O bond. In fact, acyclic analogues are known
and, consistent with the range of syntheses and reaction
conditions in Scheme 4, show that this can be a surpris-
ingly robust linkage. Interesting examples of both stable
and unstable mixed anhydrides are known in biological
55,56
systems. The former include pyrophosphate analogues
and aminoacyl-adenylate, the high-energy intermediate
used at the outset of protein biosynthesis to charge
57
tRNA. The latter unstable mixed anhydrides include,
perhaps surprisingly, amino acid-derived heterocycles
proposed to be present only as transient intermediates
58,59
in solvolysis reactions of acyclic phosphonamides (21),
This explanation for thermodynamic product formation
requires that the intermediate chlorophosphine be al-
lowed to epimerize, if (as seems likely) the initial attack
of carboxylate on phosphorus is not diastereoselective.
and pentavalent phosphorus intermediates proposed to
be present in a laboratory peptide oligomerization se-
quence (22) and perhaps in prebiotic polypeptide syn-
60
thesis and/or protein biosynthesis (23).
We have noticed that reaction of 6 with PCl
3
gives a
heterocycle that undergoes rapid epimerization in the
(
49) Hommer, H.; Gordillo, B. Phosphorus, Sulfur Silicon Relat.
+
-
3
presence of Et NH Cl on the NMR time scale, but gives
Elem. 2002, 177, 465-470.
4
7
sharp peaks in the NMR when it is crystalline. A recent
paper similarly shows that HCl can catalyze chlorophos-
phine epimerization,48 and the epimerizations of both 1a
(50) Egan, W.; Mislow, K. J . Am. Chem. Soc. 1971, 93, 1805-1806.
(
51) Peterson, L. K.; Burg, A. J . Am. Chem. Soc. 1964, 86, 2587-
2
591.
52) Bilevich, K. A.; Evdakov, V. P. Zh. Obsch. Khim. 1965, 35, 365-
368.
(
and 20 occur in the presence of trialkylammonium
salts.2
5,46
(53) Nifant’ev, E. E.; Fursenko, I. V. Usp. Khim. 1970, 39, 2187-
216.
(
The source of the variability in diastereoselec-
2
1
tivity for the synthesis of 9 as a function of solvent and
54) Hargis, J . H.; Mattson, G. A. J . Org. Chem. 1981, 46, 1597-
602.
(
43) Nesmerova, L. I.; Gololobov, Y. G. Zh. Obshch. Khim. 1981, 51,
663-1664 (Engl. 1417-1418).
44) Lai, P.; Wang, H.; Chen, R.; Liu, A.; Xu, J . Gaodeng Xuexiao
Huaxue Xuebao 1992, 13, 191-194.
45) Devedjiev, I.; Petrova, K.; Glavchev, I. Synth. Commun. 2000,
0, 4411-4415.
46) Totschnig, K.; Ellmerer-M u¨ ller, E. P.; Peringer, P. Phosphorus,
(55) Ahlmark, M.; Veps a¨ l a¨ inen, J .; Taipale, H.; Niemi, R.; J a¨ rvinen,
1
T. J . Med. Chem. 1999, 42, 1473-1476.
(56) Ahlmark, M. J .; Vapsalalnen, J . J . Tetrahedron 2000, 56, 5213.
(57) Voet, D.; Voet, J . G.; Pratt, C. W. Fundamentals of Biochemistry;
Wiley: New York, 2002.
(
(
3
(58) J acobsen, N. E.; Bartlett, P. A. J . Am. Chem. Soc. 1983, 105,
1613-1619.
(
Sulfur Silicon 1996, 113, 173-177.
(
(
(59) J acobsen, N. E.; Bartlett, P. A. J . Am. Chem. Soc. 1983, 105,
1619-1626.
47) Hersh, W. H. Unpublished results.
48) Humbel, S.; Bertrand, C.; Darcel, C.; Bauduin, C.; J uge, S.
(60) Fu, H.; Li, Z.-L.; Zhao, Y.-F.; Tu, G.-Z. J . Am. Chem. Soc. 1999,
121, 291-295.
Inorg. Chem. 2003, 42, 420-427.
2
158 J . Org. Chem., Vol. 69, No. 6, 2004

Amino Acid Derived Phosphorus Compounds
ring alkane carbon bonded to the nitrogen ring atom
being the envelope carbon atom. In most of the known
structures, it is this same carbon atom that is the furthest
removed from the mean plane.2
6,61,63,66,69,70
The nitrogen
itself is slightly pyramidalized in 9a , since the sum of
2
the angles about nitrogen (ΣN) is 350°; planar sp -
3
hybridized nitrogen would yield ΣN ) 360°, while for sp -
hybridized nitrogen ΣΝ ≈ 320°. For instance, pyrami-
dalization in 1d and 25 is small, while in 26, the nitrogen
with the CF
3
2
moiety anti to the P-NEt moiety, compa-
rable to the anti orientation of the tolyl and phenyl
groups in 9a , is fairly planar, while the other nitrogen
with the syn CF group is almost as pyramidalized as
3
that in 9a . The effect of the pyramidalization is to
increase the distance between the syn groups. Further
details of the sulfonamide geometry in 9a and 26 will be
described below (vide infra).
Heter ocycle X-r a y Str u ctu r e. The structure deter-
mination of 9a is unique - to the best of our knowledge
it is the only 1,3,2-oxazaphospholidinone and the only
acetoxyphosphine and just the second trivalent 1,3,2-
oxazaphospholidine structure.61 Most 1,3,2-oxazaphos-
pholidine structures both are based on the amino alcohol
The most striking feature of these heterocycles is the
geometry at phosphorus. The O-P-N angles are ex-
tremely acute in 9a (90.9°) and 24 (91.9°) as is the
corresponding N-P-N angle in 26 (85.5°), while in the
six-membered ring compound 25 this angle is 101° and
in the tetrahedral phosphine oxide 1d , 95.6°. If this
bending were due to the constraints of the ring (it is not
simply a geometric consequence of the longer phosphorus
bonds), one might expect the third group attached to
trivalent phosphorus to be held at a much less acute
angle, but with average O-P-R and N-P-R angles
ephedrine and are tetrahedral phosphine oxides, sulfides,
or boranes.2
6,27,62-66
X-ray structures of trivalent phos-
phines such as 9a include one of a 1,3,2-oxazaphospho-
6
1
lidine that is itself part of a larger macrocycle (24) and
one of a 1,3,2-oxazaphosphorinane (25),67 a six-membered
ring heterocycle that like 9a has a phenyl on phosphorus
as well. Comparison of data for these compounds, a
(
9
where R is the nonring substituent) of 100.5-101.5° for
a , 24, and 25, this does not seem to be the case although
6
8
recently reported sulfonamide analogue of TosL (26),
for 26 with an average N-P-R angle of 104.6° it might
and a representative tetrahedral phosphine oxide (1d )26
chosen since the ring substituents are mutually cis,
shows that neither the presence of the carbonyl nor the
sulfonyl moiety has much effect on the structure of the
heterocycle. The P-O and P-N bonds in 9a are slightly
be. The normal angle about phosphorus in acyclic triva-
71
lent alkyl and aryl phosphines is 102-103°, so in fact
the observed angles for the nonring substituent are close
to those expected for unstrained bond angles, as is that
for the O-P-N ring angle in the six-membered ring
compound 25. Hence neither the deviation from a tetra-
hedral angle for the nonring substituent nor the acute
ring angle, which is normal for phospholidines, is a
consequence of the heteroatoms or sulfonamide moiety.
On the basis of the X-ray structure of 9a , the 1,3-
interaction between the isopropyl and phenyl groups is
evidently small. The distance between the isopropyl
hydrogen atom and the phenyl ring is 2.94 Å, while the
longer (∼0.05 Å) than those in 24 and 25, which might
be ascribed both to a weaker anomeric effect in 9a ,26 and
the presence of the less basic carboxylate and sulfona-
mide moieties. These bonds are also ∼0.1 Å longer than
those in 1d , but this can be a result of the switch from
trivalent phosphine to pentavalent phosphine oxide.
Other bond lengths are comparable for similarly hybrid-
ized atoms. The ring itself is close to planar, with the
72
sum of the van der Waals radii is ∼2.97 Å. As a first
guess it is reasonable to propose that the isopropyl forces
the tolyl ring to be in the anti conformation, which then
forces the phenyl to be anti to it, as previously discussed
from a kinetic point of view. The diastereomer 9b
requires that the sulfonyl tolyl be s-cis to either the
isopropyl or phenyl substituents, but the use of molecular
modeling software suggests that the steric contacts are
(
61) Bonningue, C.; Houalla, D.; Wolf, R.; J aud, J . J . Chem. Soc.,
Perkin Trans. 2 1983, 773-776.
62) Bartczak, T. J .; Galdecki, Z.; Antipin, M. Y.; Struchkov, Y. T.
Phosphorus, Sulfur Silicon 1984, 19, 11-16.
63) Setzer, W. N.; Black, B. G.; Hubbard, J . L. Phosphorus, Sulfur
Silicon 1990, 47, 207-214.
64) Katagiri, N.; Yamamoto, M.; Iwaoka, T.; Kaneko, C. J . Chem.
Soc., Chem. Commun. 1991, 1429-1430.
65) Duthu, B.; El Abed, K.; Houalla, D.; Wolf, R.; J aud, J . Can. J .
Chem. 1992, 70, 809-816.
66) Kruszynski, R.; Bartczak, T. J .; Majewski, P. Heteroatom Chem.
001, 12, 146-150.
67) Huang, Y.; Arif, A. M.; Bentrude, W. G. J . Org. Chem. 1993,
8, 6235-6246.
68) Konya, D.; Philouze, C.; Gimbert, Y.; Greene, A. E. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 2002, C58, o108-o111.
2
minor due to the intervening SO , which extends the
distance from the tolyl moiety to the other ring substit-
uents. We therefore next consider data directly relevant
to the sulfonamide geometry.
NMR Det er m in a t ion of R in g St er eoch em ist r y.
The solution NMR data that are relevant to the confor-
(
(
(
(
(69) Bartczak, T. J .; Galdecki, Z. Acta Crystallogr. 1983, C39, 219-
222.
(
2
(70) Bartczak, T. J .; Galdecki, Z.; Rutkowska, M. Acta Crystallogr.
1983, C39, 222-224.
(
5
(71) Bruckmann, J .; Kruger, C.; Lutz, F. Z. Naturforsch., B: Chem.
Sci. 1995, 50, 351-360.
(
(72) Bondi, A. J . Phys. Chem. 1964, 68, 441-451.
J . Org. Chem, Vol. 69, No. 6, 2004 2159

Hersh et al.
SCHEME 5
F IGURE 2. View of 9a (partial structure) along the N-P-
O-C plane and Newman projection of of 9a , viewed down the
S-N bond.
angles near 90° and is greater than 10 Hz for dihedral
angles near 180° in tetrahedral 2-oxo-1,3,2-oxazaphos-
pholidines but ∼3-5 Hz in trivalent 1,3,2-oxazaphos-
phorinanes and 1,3,2-diazaphosphorinanes. The presence
or absence of a phosphorus lone pair may influence the
magnitude of this coupling constant, and the value of 3.7
Hz in 9a for the P-N-C-H dihedral angle of 139° taken
from the X-ray structure suggests that the six-membered
ring phosphorinanes provide the better comparison since
like 9a and 9b they are trivalent. The 1.4 Hz value in
mations of 9a and 9b are the methine and ring hydrogen
coupling constants, as well as the isopropyl methyl
chemical shifts. The X-ray structure of 9a exhibits a
3
H-C-C-H dihedral angle of 168°, and J HH ) 7.1 Hz is
3
consistent with this. In 9b on the other hand, J HH ) 3.0
9
b suggests that the dihedral angle is closer to 90° and
Hz, which would be the result of a 52° dihedral angle
using the Karplus equation.73 Such an angle would be
so the ring hydrogen is increasingly axial compared to
that in 9a . This result is sensible in terms of avoiding
steric interactions between the isopropyl and tolyl moiety,
by making the isopropyl increasingly equatorial. The
coupling constant in 13, by this argument, cannot be
directly compared, since now the phosphorus lone pair
is tied up by coordination to tungsten. That is, the 3.4
Hz value in 13 might be considered to be low, if the
proper comparison is to the 2-oxo-1,3,2-oxazaphospho-
lidines which would suggest a dihedral angle closer to
consistent with rotation of the isopropyl moiety to
another staggered conformation, as might occur to relieve
tolyl-methyl interactions if the tolyl were s-cis to the
isopropyl group, as drawn in Scheme 5. Consistent with
this, the chemical shift of one of the isopropyl methyl
signals is unusually far upfield, due to its position above
the face of the arene ring. Furthermore, upon coordina-
3
tion of 9a to tungsten, J HH ) 3.4 Hz in 13, and one
5
interpretation of this is that the bulky W(CO) moiety
9
5
0°, and so the bulky W(CO) moiety not only results in
forces the tolyl to be anti to it rather than to the phenyl,
again as drawn in Scheme 5, so that as in 9b rotation of
the isopropyl moiety occurs. Since the same patterns in
isopropyl moiety chemical shift and coupling constant
occur in 12b and in 11b,c, the toluenesulfonyl moiety is
evidently syn to the isopropyl in these cases as well.
Interestingly, the parent toluenesulfonylamino acid 6 also
the tolyl and isopropyl moieties adopting an s-cis con-
formation, but in forcing the isopropyl to be more
equatorial, just as in 9b.
Th eor etica l a n d Str u ctu r a l Su r vey of Su lfon a -
m id e Geom etr y. To understand the steric effect of the
toluenesulfonyl moiety, it is necessary to know if the
observed geometry is the norm. As shown in Figure 2,
the sulfonamide moiety lies in a staggered conformation,
with the tolyl roughly anti to the putative nitrogen lone
pair (vide infra). There is no indication of any π-overlap
between either of the sulfonyl oxygen double bonds and
the nitrogen lone pair, and the pyramidalization of the
nitrogen provides further evidence of the absence of
π-delocalization of the lone pair, and also allows the steric
interaction of the cis phenyl and isopropyl groups to be
minimized. Stereoelectronic effects have been investi-
3
exhibits a coupling constant ( J HH ) 4.7 Hz) suggestive
of a 60° H-C-C-H dihedral angle; a simple PM3
calculation shows this to be a reasonable outcome of a
hydrogen-bonded seven-membered ring structure com-
parable to those seen for t-Boc-R-amino acids;74 a similar
geometry is in fact observed in two X-ray structures
incorporating valine-analogues in sulfonamidopeptide
mimetics.75
The other available coupling constant, 3J PH, has been
considered by several groups in both five and six-member
ring heterocycles, with the conclusion that P-N-C-H
coupling constants are too variable for evaluation of the
gated in the six-membered ring cases of which 25 is
representative,6
7,78
including the anomeric effect which
would be the stabilizing interaction of both the nitrogen
dihedral angle.2
6,67,76,77
Generally J PH ≈ 0 Hz for dihedral
3
and oxygen lone pairs with the P-Ph σ* orbital when
the phenyl is axial,6
7,78
and nitrogen-phosphorus lone
(73) Bovey, F. A.; J elinski, L.; Mirau, P. A. Nuclear Magnetic
pair repulsion, which would complement the anomeric
effect.79 On this basis both the anomeric n-σ* and the
nitrogen-phosphorus lone pair-lone pair repulsion in-
teractions would be operative since the electron-with-
Resonance Spectroscopy, 2nd ed.; Academic Press: San Diego, 1988.
(
(
74) Branik, M.; Kessler, H. Chem. Ber. 1975, 108, 2176-2188.
75) Gennari, C.; Salom, B.; Potenza, D.; Longari, C.; Fioravanzo,
E.; Carugo, O.; Sardone, N. Chem.sEur. J . 1996, 2, 644-655.
76) Hutchins, R. O.; Maryanoff, B. E.; Albrand, J . P.; Cogne, A.;
Gagnaire, D.; Robert, J . B. J . Am. Chem. Soc. 1972, 94, 9151-9158.
77) Setzer, W. N.; Black, B. G.; Hovanes, B. A.; Hubbard, J . L. J .
Org. Chem. 1989, 54, 1709-1713.
(
(
(78) Huang, Y. D.; Yu, J . H.; Bentrude, W. G. J . Org. Chem. 1995,
60, 4767-4773.
2
160 J . Org. Chem., Vol. 69, No. 6, 2004

Amino Acid Derived Phosphorus Compounds
TABLE 1. Nitr ogen P yr a m id a liza tion a n d
Con for m a tion a bou t th e S-N Bon d in
N-Su lfon yl-1,3,2-oxa za h eter olid in es
minimum in which the groups are staggered and the
nitrogen lone pair is now anti to the carbon attached to
sulfur, (4) these conformations interconvert by nitrogen
inversion with a lower energy barrier than they would
by rotation, and (5) X-ray structures reveal a range of
conformations83 with a preference for the secondary
minimum energy conformation over the global minimum
conformation.82
The geometry observed in 9a as shown in Figure 2 is
in fact the secondary minimum energy conformation.
Inversion at nitrogen would clearly have the effect of
pushing the axial phenyl and isopropyl groups closer
together, while 180° rotation about the S-N bond would
place all three ring substituents on the same side of the
ring. Thus, there are good steric reasons for supposing
that the observed sulfonamide conformation is preferred.
However, inspection of models does not reveal any
compelling reasons for preference of this diastereomer
over the phosphorus epimer, particularly given the
possibility of inversion at nitrogen.
A number of X-ray structures of N-sulfonamido-1,3,2-
oxazametallacycles formed from R-amino acids have been
carried out.3
0,84-87
The geometry of the toluenesulfonyl
moiety appears not to have been discussed previously.
Data from our analysis of a few recent examples (3a , 27-
3
2) together with those of 9a and 26 is collected in Table
1
. The position of the putative nitrogen lone pair was
calculated as being equiangular from the three observed
nitrogen substituents, and this provides a more intuitive
measure of pyramidalization at nitrogen. Half of the
amino acid examples are essentially planar at nitrogen,
and interestingly, each of these exhibits a nearly eclipsed
conformation, with the arene and lone pair syn peripla-
nar,88 which formally corresponds to the minimum energy
eclipsed conformation. The more pyramidalized examples
exhibit staggered conformations, and the tolyl and lone
pair are anti periplanar in three of the four amino acid
examples, which corresponds to the calculated secondary
minimum energy conformation. Only the triflate example
a
Sum of the three angles at nitrogen. b Calculated angle for lone
pair to be equiangular to each atom attached to nitrogen.
Counterclockwise angle starting from the lone pair on nitrogen.
c
(
26) is different, giving as noted above sulfonamide
drawing character of the toluenesulfonyl group is evi-
dently an inductive effect. In the absence of π-effects, it
is stereoelectronically comparable to the alkyl or aromatic
groups seen in compounds such as 24 and 25.
nitrogen atoms that differ in pyramidalization, and that
exhibit the minimum energy eclipsed conformation, so
it would be interesting to examine amino acid analogues,
for which no X-ray structures are available. Given this
modest range of amino acid analogues that exhibit
similar conformations it is possible that the observed
Theoretical study of sulfonamide geometry is an active
area of research.8
0-82
Some relevant conclusions are that
(1) conjugation of the type present in planar amides is
absent, (2) the nitrogen is pyramidal but the total energy
is relatively insensitive to nitrogen hybridization and so
X-ray structures may be somewhat misleading, (3) there
is a 2-fold torsional barrier with a minimum at the
conformation where the nitrogen lone-pair roughly eclipses
the carbon attached to sulfur and the nitrogen substit-
uents eclipse the sulfonyl oxygen atoms, and a secondary
(79) Gray, G. A.; Nelson, J . H. Org. Magn. Reson. 1980, 14, 8-13.
(
80) Elguero, J .; Goya, P.; Rozas, I.; Catal a´ n, J .; De Paz, J . L. G.
THEOCHEM 1989, 184, 115-129.
81) Bindal, R. D.; Golab, J . T.; Katzenellenbogen, J . A. J . Am. Chem.
Soc. 1990, 112, 7861-7868.
82) Nicholas, J . B.; Vance, R.; Martin, E.; Burke, B. J .; Hopfinger,
A. J . J . Phys. Chem. 1991, 95, 9803-9811.
83) K a´ lm a´ n, A.; M a´ ty a´ s, C.; Argay, G. Acta Crystallogr. 1981, B37,
868-877.
(
(
(
J . Org. Chem, Vol. 69, No. 6, 2004 2161

Hersh et al.
geometry is preferred not only for steric reasons but for
electronic reasons as well.
substituent of valine; hence the cis stereochemistry is
apparently mediated by steric interactions with the
intervening N-sulfonyl moiety, which evidently does not
interact with the smaller methyl substituent of alanine.
One unexpected feature of this work is the highly
electron-withdrawing nature of 9a as a transition-metal
ligand. The sulfonamide moiety was expected to be
electron-withdrawing, but the carboxylate seems to be
even more so. This work extends the number of sulfona-
mide-substituted heterocycles that exhibit interesting
and high stereoselectivity, so exploration of the steric
space of this substituent appears to be in order. Future
research will be directed toward extensions of this
substituted amino acid methodology for the synthesis of
new chiral phosphorus compounds.
We began this account by noting the similarities of
ephedrine-derived heterocycles to the boron and iridium
heterocycles 2 and 3 (Scheme 1). Addition of a ligand to
boron yields 2b in which the 1,3-substituents are mutu-
ally cis as well as mutually anti to the sulfonyl tolyl
moiety, exactly as seen in 9a albeit in contrast to 1a . The
stereochemistry in the boron case was considered to be
controlled by a π-stacking interaction between the enone
and the indole ring, but our results suggest the steric
interaction with the intervening toluenesulfonyl moiety
must be important in giving the 1,3-cis substitution. The
absence of this bulky group allows the trans stereochem-
istry to predominate in 1a . The iridium case is better
compared to 13, where addition of the bulky W(CO)
5
Exp er im en ta l Section
moiety to 9a forces the toluenesulfonyl moiety to rotate
placing all three substituents on the same side of the ring,
much like 3b.
Electr on -With d r a w in g Ch a r a cter of 9a . A byprod-
uct of the design of this new type of ligand is the
surprising discovery that it is apparently even more
electron-withdrawing than the prototype sulfonylphos-
phoramide TosL (eq 1), on the basis of the high CO
Compounds 7a ,b, 8a , 9a , 11a , and 12a ,b were made by
analogous procedures, although details of the purifications
differ. The procedure for 9a and the partial purification of 9b
is representative.
(
2S,4S)- a n d (2R,4S)-4-Isop r op yl-5-oxo-2-p h en yl-3-p-
tolu en esu lfon yl-1,3,2-oxa za p h osp h olid in e (9a ,b). In an
inert atmosphere glovebox under recirculating nitrogen was
added a solution of Et N (1.31 g, 13.0 mmol) in 3 mL of toluene
3
to a suspension of 6 (1.00 g, 3.69 mmol) in 20 mL of toluene to
give at first a nearly clear solution after which a thick
suspension precipitated out. Toluene was added (10 mL), and
5
vibrational frequencies of the W(CO) adducts. Compara-
tive data for TosL and a variety of phosphorus ligands
have been published,20 and all three CO bands for 13,
then a solution of PhPCl
was added dropwise. The mixture was allowed to stir for 3 h,
and 31P NMR analysis of a few drops of the solution in CDCl
2
(0.93 g, 5.17 mmol) in 5 mL of toluene
the W(CO)
the W(CO)
5
adduct of 9a , are higher in frequency than
adducts of both P(OMe) and TosL. As far as
5
3
3
we are aware the electronic properties of the acetoxy
functional group on phosphorus basicity and acidity have
not been assessed, any more than that of the sulfonamido
group before our report.20 Last, when P NMR data was
examined for the presence of correlation with ligating
at both 15 min and 3 h exhibited a 91:9 ratio of 9a /9b. The
mixture was then filtered through Celite, the solid was washed
successively with ether and benzene, and then the solvent was
removed in vacuo. The yellow crude product (1.53 g) was
crystallized from 40 mL of ether at -35 °C to give 0.25 g of
white crystals. Three more crystallizations from (successively)
20, 10, and 5 mL of ether at -35 °C gave a total of 0.79 g
31
2
0
properties, only the tungsten-phosphorus coupling
constant 1
highest for P(OMe)
J
PW had any utility. In this case, it is still
(386 Hz) but now is second-highest
(57% yield) of 9a contaminated by 0.5% of 9b.
3
For more convenient large-scale synthesis, in a similar
manner 4.15 mL (29.8 mmol, 3.5 equiv) of triethylamine, 2.31
g (8.51 mmol, 1 equiv) of 6, and 2.13 g (11.9 mmol, 1.4 equiv)
for 9a (346 Hz) with TosL yet lower (329 Hz), but it is
unclear if the value reflects higher electron-withdrawing
character for 9a relative to TosL or is merely due to the
oxygen atom bound to phosphorus.
of PPhCl
2
were combined in 80 mL of ether, and the mixture
was stirred at room temperature for 1 h. The suspended
+
-
Et
3
NH Cl was removed by filtration and washed with 3 × 5
Con clu sion s
mL of ether. The solvent was removed from the combined
ethereal solution in vacuo giving a pale yellow solid in a clear
yellow oil. This residue was heated under vacuum at 50 °C
Reaction of the inexpensive and readily available
toluenesulfonyl derivative of valine with PhPCl gives a
2
2
for 1.5 h to remove unreacted PPhCl , resulting in formation
near-quantitative yield of a new phosphorus heterocycle
in high diastereomeric purity. A single crystallization
gives one diastereomer in overall 50% yield and so
constitutes one of the simplest syntheses of an enantio-
merically pure trivalent phosphorus compound; the
of a pale yellow powder (2.91 g, 91% yield) consisting of a
spectroscopically pure mixture of 9a and 9b in a 9.2:1 ratio.
Fractional crystallization then proceeded as follows. Inside the
glovebox, the 2.91 g of the mixture was dissolved in 45 mL of
ether and the solution was allowed to stand at -35 °C for
several days to give 1.62 g (50% yield) of small pale yellow
crystals. The filtrate was stripped to dryness and the residue
2
method has been extended to give a C -symmetric chiral
diphosphine as well. In contrast to the well-known
ephedrine-derived 1,3,2-oxazaphospholidine, the ring
substituents are cis to each other. High diastereoselec-
tivity is only achieved with the relatively large isopropyl
(1.30 g) was recrystallized from 20 mL of ether at -35 °C to
give 0.33 g (10% yield) of a second crop of small white crystals.
Recrystallization (0.89 g residue from 10 mL ether) gave a
third crop of crystals (0.18 g, 6% yield). The first crop contained
no detectable 9b, the second crop contained 1% 9b, and the
third crop 9% of 9b. The final residue (0.70 g) consisted of a
(84) Antolini, L.; Battaglia, L. P.; Gavioli, G. B.; Corradi, A. B.;
1
[R]
.04:1 mixture of 9a /9b. Major isomer 9a : mp 129-131 °C;
Grandi, G.; Marcotrigiano, G.; Menabue, L.; Pellacani, G. C. J . Am.
Chem. Soc. 1983, 105, 4327-4332.
27
D
) +84.83 (c ) 2.426, C
6
H
6
); IR (CHCl
3
) 3023, 1791, 1360,
-
1 1
(85) Antolini, L.; Battaglia, L. P.; Bonamartini Corradi, A.; Mar-
1164 cm ; H NMR (CDCl
3
) δ 7.84 (d, J ) 8.2 Hz, 2H), 7.60
cotriagiano, G.; Menabue, L.; Pellacani, G. C. J . Am. Chem. Soc. 1985,
3
(
m, 2H), 7.46 (m, 3H); 7.38 (d, J ) 8.2 Hz, 2H), 3.53 (dd, J HH
1
07, 1369-1375.
3
)
7.1 Hz, J PH ) 3.7 Hz, 1H; assignment of the 7.1 Hz coupling
(
86) Corradi, A. B. Coord. Chem. Rev. 1992, 117, 45-98.
3
constant to J HH was made by homonuclear decoupling of the
(87) Battistuzzi, G.; Borsari, M.; Menabue, L.; Saladini, M.; Sola,
3
.53 ppm signal by irradiation at δ 1.71), 2.47 (s, 3H); 1.71
M. Inorg. Chem. 1996, 35, 4239-4247.
(88) Kane, S.; Hersh, W. H. J . Chem. Educ. 2000, 77, 1366.
(octet, J ) 6.9 Hz, 1H), 0.94 (d, J ) 6.9 Hz, 3H), 0.74 (d, J )
2
162 J . Org. Chem., Vol. 69, No. 6, 2004

Amino Acid Derived Phosphorus Compounds
6
.8 Hz, 3H). 31P NMR (CDCl
3
) 133.0 ppm; 13C NMR (CDCl
3
)
noncrystalline 57:43 mixture of 9a /9b used above for the
1
1
1
71.3 (d, J PC)12.4 Hz), 145.2, 139.6 (d, J PC ) 38.6 Hz), 134.7,
31.3, 130.3, 128.9 (d, J PC ) 4.4 Hz), 128.3 (d, J PC ) 18.8 Hz),
27.7 (d, J PC ) 4.8 Hz), 61.6, 31.5, 21.7, 19.7, 18.2 ppm. Min or
optical rotation.
Ack n ow led gm en t. Financial support for this work
from the National Institutes of Health (GM59596-01),
the City University of New York PSC-CUNY Research
Award Program, the CUNY Alliance for Minority Par-
ticipation program funded by the National Science
Foundation (C.K.S.), and a Pfizer Undergraduate PRE-
PARE Fellowship (J .G.) is gratefully acknowledged.
2
7
27
isom er 9b: [R]
D
) +93.8 (calculated from [R]
D
) +88.65
(c ) 2.14, C
6
H
6
) for a sample consisting of a 57.4:42.6 mixture
) δ 7.60 (m, 3H), 7.46 (m, 4H), 7.21
1
of 9a /9b); H NMR (CDCl
d, J ) 8.2 Hz, 2H), 4.06 (dd, J HH)3.0 Hz, J PH)1.4 Hz, 1H;
3
3
3
(
3
assignment of the 3.0 Hz coupling constant to J HH was made
both by simulation of the observed isopropyl CH multiplet and
by homonuclear decoupling of the 4.06 ppm signal by irradia-
tion at δ 2.50), 2.50 (d of septets, J ≈ 3.1, 6.8 Hz, 1H), 2.41 (s,
Su p p or tin g In for m a tion Ava ila ble: Preparation of com-
3
1
3
H), 1.16 (d, J ) 7.1 Hz, 3H), 0.64 (d, J ) 6.7 Hz, 3H);
P
1
13
pounds 6, 7a ,b, 8a , 11a , 12a ,b, and 13; IR and H, C, and
1
3
NMR (CDCl
3
) 136.7 ppm; C NMR (CDCl
3
) 171.0 (d, J PC
)
31
P NMR spectra for compounds 4-9 and 11-13; decoupling
1
(
1
3.0 Hz), 144.5, 139.4 (d, J PC ) 50.4 Hz), 136.8, 132.6, 130.5
d, 25.8 Hz) 129.7, 128.9 (d, 5.3 Hz)
27.1, 60.9, 29.2, 21.6, 18.0, 14.9 ppm. Anal. Calcd for
SP: C, 57.29; H, 5.34; N, 3.71. Found: C, 57.15; H,
.24; N, 3.64 for 9a ; C, 56.47; H, 5.23; N, 3.58 for the
results for 9a ; the X-ray structure determination of 9a ; and
tables of comparative NMR and X-ray data. This material is
available free of charge via the Internet at http://pubs.acs.org.
J
PC
)
J
PC
)
18 4
C H20NO
5
J O035508+
J . Org. Chem, Vol. 69, No. 6, 2004 2163