Reactions of 1,2-oxaphospholenes. 9. Attempted deprotonation at C5

Full text of DOI:10.1021/jo000751j

1480
J . Org. Chem. 2001, 66, 1480-1483
Notes
Rea ction s of 1,2-Oxa p h osp h olen es. 9.1
Sch em e 1
Attem p ted Dep r oton a tion a t C5
Roger S. Macomber,* Mark Guttadauro,2
Allan R. Pinhas,* and J eanette Krause Bauer
Department of Chemistry, University of Cincinnati,
P.O. Box 210172, Cincinnati, Ohio 45221-0172
allan.pinhas@uc.edu
Received May 16, 2000
We1 previously reported that C5 of 1,2-oxaphosphol-
3-ene 2-oxides 1 (prepared from propargyl alcohols 2 via
allenic phosphoryl derivatives 3,3 Scheme 1) can undergo
facile allylic nucleophilic and free radical substitution
reactions that leave the ring system intact.1b,4 J udging
from 1H and 13C NMR chemical shifts,4 a C5 allylic
hydrogen of 1 should exhibit increased acidity owing to
the proximity of the phosphoryl group. In the present
study, we wished to determine if the C5 position of such
oxaphospholenes can be deprotonated to generate carb-
anion 4 (Scheme 2).5 In selecting an appropriate base to
deprotonate 1, it is necessary to avoid strongly nucleo-
philic oxy bases because of their preferential attack at
phosphorus.6,7 We therefore selected sodium hexameth-
yldisilazide (NaHMDS) for our initial experiments.
The five substrates used in this study were prepared
by previously published methods (Scheme 1). It should
be noted that derivatives of 1 (as well as 3) with R2 * R3
exist as two comparably stable chiral diastereomers. The
diastereomers of 1a and 1b have been assigned by a
combination of 1H NMR spectroscopy and X-ray crystallog-
raphy.1b Interestingly, although 1c was first described
in 1971,8 no mention of the stereochemistry of the
product(s) has heretofore been made. In our hands, 1c is
formed as a 60/40 mixture of two diastereomers; the
major isomer has the Z configuration (with phenyl groups
trans) by X-ray analysis (see Supporting Information).
Addition of 2 equiv of NaHMDS to Z/E-1a at ambient
temperature resulted in a transient burgundy color. The
mixture was quenched with a slight excess of deuterated
trifluoroacetic acid (TFA-d), under conditions to which
all substrates and products are stable. The product
mixture consisted quantitatively of allenic precursor 3a
Sch em e 2
(deuterated at oxygen). It seems likely that the steric
bulk of the C5 tert-butyl group prevented proton abstrac-
tion from C5, directing the base instead to the vinyl
hydrogen at C4, which occupies a perfect geometry for
ring opening via an E2 process.
To block the C4 position, we next examined the Z
stereoisomer of 4-bromo derivative 1b.1b Addition of up
to 7 equiv of NaHMDS (at ambient temperature) gave a
precipitate that, when quenched with TFA-d, afforded a
quantitative yield of the starting material, without
detectable epimerization to the E isomer nor incorpora-
tion of deuterium. We were not able to ascertain the
composition of the precipitate, but lack of epimerization
and deuterium exchange is strong evidence against C5
deprotonation.
(1) Paper 8 in the series: Macomber, R. S.; Rardon, D. E.; Ho, D.
M. Phosphorus, Sulfur Silicon Relat. Elem. 1993, 75, 95. (b) A leading
reference to earlier papers: Macomber, R. S.; Rardon, D. E.; Ho, D.
M. J . Org. Chem. 1992, 57, 3874.
(2) Taken in part from the Ph.D. Dissertation of M.G., University
of Cincinnati, 1999.
(3) Macomber, R. S. J . Org. Chem. 1971, 36, 2713.
(4) Rardon, D.; Macomber, R. S. J . Org. Chem. 1990, 55, 1493.
(5) Calculation on model structures predict that the 3-ene is more
stable than the 4-ene by about 0.5 kcal/mol. When deprotonated, the
3-ene and 4-ene generate the same anion, which has the majority of
charge on C3.
With this additional evidence that the C5 t-Bu group
blocks access to the C5 hydrogen, we next examined 1c
in which the C5 t-Bu group is replaced by an electroni-
cally facilitating and less sterically demanding phenyl
group and the C3 t-Bu by a methyl. When stereoisomeri-
(6) Macomber, R. S.; Constantinides, I.; Garrett, G. J . Org. Chem.
1985, 50, 4711.
(7) Macomber, R. S. J . Am. Chem. Soc. 1983, 105, 4386 and
references therein.
(8) Campbell, I. G. M.; Raza, S. M. J . Chem. Soc. C 1971, 1836.
10.1021/jo000751j CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/25/2001

Notes
J . Org. Chem., Vol. 66, No. 4, 2001 1481
Sch em e 3
cally pure Z-1c was treated with 10 equiv of NaHMDS
at ambient temperature, a transient burgundy color
appeared and faded immediately. Following a H2O quench,
an epimerized 75:25 Z/ E mixture of 1c was recovered
in 30% yield. In addition, ring-opened keto-phospinate
5c9 was isolated in 70% yield (Scheme 3). Compound 5c
can be seen as arising from facile10 hydrolysis of 1,2-
oxaphosphole-4-ene 6c, a slightly less stable isomer of
the 3-ene.5,11 Both the epimerization of 1c and its
isomerization to 5c can be accommodated by a mecha-
nism involving intermediate 4c.
At this point, our attention shifted to potassium
hydride (KH), a more powerful but less hindered base,
which is able to deprotonate both hindered substrates12
and the γ-position of R,â-unsaturated ketones.13 When
stereoisomerically pure Z-1a was allowed to react with
3.5 equiv of KH at ambient temperature, followed by a
2-propanol quench, allenic precursor 3a was formed in
22% yield. Reisolated substrate 1a (29% yield) proved to
be partially epimerized (5:1 Z:E). The major product was
ring-opened phosphinate 5a , an analogue of the product
formed from 1c. Thus, KH seems able to promote
deprotonation at C5 leading to formation of 6a , rather
than simply deprotonating at C4 as did NaHMDS. It is
interesting to note that treatment of 3a with KH,
followed by quench with TFA-d, gives starting material
with the allenic proton (as well as the hydroxyl proton)
deuterated.
elimination (of bromide) reaction at C4;14 however, the
formation of 7 rather than 5a suggests that 1a must not
be involved as an intermediate in the formation of 7.
Alternatively, the formation of 3a could be attributed to
ring-opening elimination of bromide from 1b by hydride.
Treatment of Z-1c with 4 equiv of KH at room
temperature led to a dark red solution that was stable
up to a week. TFA workup led to a mixture of epimerized
1c (9%) and 5c (53%), similar to the results with
NaHMDS (Scheme 3). Attempts to trap presumed inter-
mediate 4c with other electrophiles (e.g., methyl iodide,
TMSCl) were unsuccessful.
In an attempt to determine the mechanism of forma-
tion of 7 from 1b, we next studied the reaction of 1d ,
which lacks an extractable hydrogen at C4 and C5. While
1d was inert toward NaHMDS, it reacted readily with
KH (4 equiv) to yield allenic precursor 3d and starting
material 1d in a 2:5 ratio. When 3d itself was treated
with 5 equiv of KH, propargyl phosphinate 8 was formed
in 50% yield and 20% of the starting material 3d could
be recovered.
Compound 1e, the debrominated analogue of 1d , also
was inert toward NaHMDS. However, when 1e was
allowed to react with 3 equiv of KH, the only product
was phosphinic acid 9 (representing net reduction) plus
Though C4-brominated derivative Z-1b was inert to
NaHMDS, it did react with KH to give a product mixture
consisting of allene 3a (20%) and unexpected enol 7
(80%). No debrominated 1a was detected in the product
unreacted starting material. Significantly, 9 was previ-
ously identified as the major product of the electron-
transfer reactions of 1e with sodium naphthalenide or
lithium dimethylcuprate.6 These results strongly suggests
that, in the absence of an exchangeable hydrogen, KH
also is reacting by an electron-transfer pathway.
mixture. The formation of 3a might suggest the inter-
mediacy of 1a , formed by an addition (of hydride)-
(9) The structure of 5c was established by comparison to a simulated
spectrum generated by RACCOON Version 2.0 by P. F. Schatz,
Department of Chemistry, University of Wisconsin-Madison, Madison,
WI 53706 (see Supporting Information for calculated values).
(10) Bergesen, K. Acta Chem. Scand. 1965, 19, 1784; 1969, 23, 2556.
(11) Known compound 6d 8 undergoes ready acid-catalyzed isomer-
ization to 1d .2,8
(13) Brown, C. A. J . Org. Chem. 1974, 39, 3913.
(14) A similar addition-elimination has been noted in the reaction
(12) Brown, C. A. Synthesis 1974, 427.
of the C4,C5-dibrominated derivative of 1b with KH.1b

1482 J . Org. Chem., Vol. 66, No. 4, 2001
Notes
Sch em e 4
It is our judgment that all our results can be rational-
ized most economically by a finely balanced competition
between proton transfer and electron-transfer processes
(Scheme 4). The ring opening of 1 (E ) H) to allene 3 is
an example of base-promoted E2 process, the reverse of
the electrophilic cyclization of 3 to 1. The formation of
ring-opened ketone 5, accompanied by epimerization of
the starting material, we interpret as resulting from
removal of a C5 proton by base, formation of intermediate
carbanion 4, protonation to give 6, followed by subsequent
hydrolysis to 5. Since this pathway is followed when there
is also a C4 hydrogen available (1a and 1c), the C5
hydrogen must be at least comparably reactive (in the
case of 1a ) or more reactive (in the case of 1c). In the
case of 1b (E ) Br), the hydrolysis of 6b leads to enol 7
rather than ring opened ketone 5b.
carbanionic intermediate 10 must not survive long
enough to protonate and regenerate 1 (E ) H).
The rearrangement of 3 to 8 also is likely to be a simple
base-promoted prototropy. However, the formation of 9
from allene 3 or from 1 (E ) Br) must involve an electron-
transfer process, as previously was shown using sodium
naphthalenide.6
Exp er im en ta l Section
Compounds 1a -e were prepared by methods already reported
in the literature. The reference numbers are given in Scheme
1. Spectral data for compounds 1 and 3 may be found in these
references. All information about the X-ray structure of Z-1c
may be found in the Supporting Information. Also in the
Supporting Information are full experimental details for all
reactions. Below are two typical procedures.
Rea ction of Z-1c w ith Na HMDS. A solution of 19.9 mg (0.07
mmol) of Z-1c in 2.5 mL of THF was placed under argon and
allowed to mix as 774 µL (0.774 mmol) of a 1.0 M solution of
NaHMDS in THF was added dropwise over 2 min. Each drop
produced a transient burgundy color, which faded immediately
but persisted longer and longer with each drop, eventually
leaving a bronze colored solution. Addition of 2.5 mL of D2O
caused the solution to fade to a pale yellow, from which the THF
was removed in vacuo. This solution was extracted three times
with CH2Cl2 and dried over MgSO4, and the solvent was removed
in vacuo to give 4.3 mg (0.02 mmol) of epimerized 75:25 E/ Z-
1c, identified by 1H NMR spectroscopy. The remaining aqueous
layer was hydrolyzed with TFA-d1, diminishing the milky color
of the solution. The aqueous solution was extracted two times
The ring opening of 1 (E ) H) to allene 3, as discussed
above, can be regarded as a base-promoted E2 process.
Because reaction of KH with C4-brominated oxaphos-
pholenes 1b and 1d gave neither the corresponding
debrominated oxaphospholenes (1a and 1e, respectively)
nor the products from 1a or 1e, we do not consider the
addition (of hydride at C4)-elimination (of bromide from
C4) mechanism to be a viable route for the reactions of
1b and 1d . Therefore, the corresponding reaction of 1
(E ) Br) to allene 3 most likely proceeds by an electron-
transfer process (i.e., halogen-metal exchange). Proposed

Notes
J . Org. Chem., Vol. 66, No. 4, 2001 1483
with CH2Cl2 and dried over MgSO4, and the solvent was removed
in vacuo to give 14.4 mg (0.05 mmol, 68.5% yield) of 5c.
concentrated HCl to pH 3 (pH paper), producing a precipitate
that was isolated by filtration and dissolved in acetone to remove
inorganic salts. The solvent was removed in vacuo to give 636.4
mg of a tan solid with a melting point of 90-95 °C. The solvent
was removed in vacuo from the CH2Cl2 layer from the extrac-
tions to yield 87.4 mg (0.32 mmol) E/ Z-1c, which represents a
recovery of 8.7% of epimerized starting material. The solid
recovered from the bicarbonate layer was further dried at 0.025
mm pressure for 5 h to give 568.4 mg (1.97 mmol) of a pale yellow
solid identified as 5c, in a yield of 53.3%.
Rea ction of Z-1c w ith P ota ssiu m Hyd r id e. Glassware for
this experiment was dried at 100 °C for 72 h. A suspension of
35% potassium hydride in mineral oil was dispersed initially
with a screwdriver and then agitated for 4 h. Into a flask was
added 1489.7 mg (13.0 mmol) of potassium hydride dispersion,
which was washed three times with hexane under argon, and
the remaining solvent as evaporated in an argon stream. The
potassium hydride was suspended in 10 mL of THF into which
a solution of 1000.1 mg (3.70 mmol) Z-1c in 15 mL of THF was
added dropwise over 2 min, producing an dark red color after 1
h. The suspension was allowed to react for 48 h, producing a
clear, deep red solution after the unreacted potassium hydride
was allowed to settle. The supernatant was decanted from the
reaction flask and acidified with TFA to pH 6 (pH paper), causing
the solution to turn a clear, yellow. Analysis by TLC (50/50 CH2-
Cl2/ethyl acetate) indicated only very polar products. 1H NMR
spectroscopy of the product mixture showed epimerized E/ Z-
1c as the minor product and compound 5c as the major product.
The solvent was removed in vacuo to give an oil, which was
dissolved in CH2Cl2, leaving potassium trifluoroacetate behind
as a precipitate. The solution was extracted with aqueous sodium
bicarbonate, and the bicarbonate layer was acidified with
Ack n ow led gm en t. The authors would like to thank
the reviewers for many very helpful suggestions.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra for
new compounds 5a , 5c, 7, and 9, full experimental details for
each reaction, results of the calculations mentioned in ref 5,
and the chemical shifts for ref 11; ORTEP diagram and tables
of crystal data, bond lengths, bond angles, atomic coordinates,
and anisotropic thermal parameters for compond Z-1c. This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O000751J