P(CH3NCH2CH2)3N as a dehydrobromination reagent: Synthesis of vitamin A derivatives revisited

Article abstract of DOI:10.1021/jo010648+

Although P(CH₃NCH₂CH₂)₃N (1) was found to be less effective than 1,5-diazabicyclo[4.3.0]non-5-ene (DBN) or 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in the removal of hydrogen bromide from vitamin A intermediates 13-cis-10-bromo-9,10-dihydroretinyl acetates (6) and 14-bromo-9,14-dihydroretinyl acetate (11) when the reaction was carried out in refluxing benzene, in acetonitrile at room temperature it was superior to DBN and DBU. A ³¹P NMR study of this reaction suggests that the carbanion generated from acetonitrile-d₃ in the presence of 1 is the basic species that initiates the elimination step. Diastereoselectivity of the nucleophilic addition of (Z)-HC≡ C(CH3)=CHCH20H to the carbonyl group of (E)-2-methyl-4-(2′,6′,6′-trimethyl-1′-cyclohexen- 1′-yl)-3-butenal (2) was only moderate (20%), and (9R*, 10S*)-13-cis-11,12 -didehydro-9,10-dihydro-10-hydroxyretinol (3b) predominated. The LiAlH4 reduction of the C≡C bond in the diastereoisomeric diols 3 afforded 13-cis-9,10-dihydro-10-hydroxyretinols 4a and 4b as major products together with 11-cis-13-cis-isomers and the deoxygenated compound (3EZ,5EZ,8E)-3,7-dimethyl-9-(2,6,6-trimethyl-1-cyclohexen-1-yl)- 1,3,5,8-nonatetraene (9). Reaction of 15-acetates of the pure diastereoisomeric allylic alcohols 4a and 4b with PBr3 occurred with significant but not identical retention of configuration, and with concomitant formation of the rearranged bromide 11.

Full text of DOI:10.1021/jo010648+

420
J . Org. Chem. 2002, 67, 420-425
P (CH3NCH2CH2)3N a s a Deh yd r obr om in a tion Rea gen t: Syn th esis
of Vita m in A Der iva tives Revisited
Andrzej E. Wro´blewski† and J ohn G. Verkade*
Department of Chemistry, Iowa State University, Ames, Iowa 50011-3111
jverkade@iastate.edu
Received J une 25, 2001
Although P(CH3NCH2CH2)3N (1) was found to be less effective than 1,5-diazabicyclo[4.3.0]non-5-
ene (DBN) or 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in the removal of hydrogen bromide from
vitamin A intermediates 13-cis-10-bromo-9,10-dihydroretinyl acetates (6) and 14-bromo-9,14-
dihydroretinyl acetate (11) when the reaction was carried out in refluxing benzene, in acetonitrile
at room temperature it was superior to DBN and DBU. A 31P NMR study of this reaction suggests
that the carbanion generated from acetonitrile-d3 in the presence of 1 is the basic species that
initiates the elimination step. Diastereoselectivity of the nucleophilic addition of (Z)-HCt
C(CH3)dCHCH2OH to the carbonyl group of (E)-2-methyl-4-(2′,6′,6′-trimethyl-1′-cyclohexen-1′-yl)-
3-butenal (2) was only moderate (20%), and (9R*,10S*)-13-cis-11,12-didehydro-9,10-dihydro-10-
hydroxyretinol (3b) predominated. The LiAlH4 reduction of the CtC bond in the diastereoisomeric
diols 3 afforded 13-cis-9,10-dihydro-10-hydroxyretinols 4a and 4b as major products together with
11-cis-13-cis-isomers and the deoxygenated compound (3EZ,5EZ,8E)-3,7-dimethyl-9-(2,6,6-trimethyl-
1-cyclohexen-1-yl)-1,3,5,8-nonatetraene (9). Reaction of 15-acetates of the pure diastereoisomeric
allylic alcohols 4a and 4b with PBr3 occurred with significant but not identical retention of
configuration, and with concomitant formation of the rearranged bromide 11.
Sch em e 1a
In tr od u ction
The synthesis of compound 1 in this laboratory1 and
the discovery of its very strongly basic nature2 prompted
us to evaluate the reactivity of 1 in the elimination of
H-X species (X ) Cl, Br, I, OSO3R, etc) from organic
substrates.3,4 As a part of this project we focused our
attention on the application of this approach to the
synthesis of all-trans Vitamin A and its various isomers
by introducing the fifth CdC bond into the 9,10 position
of the retinol skeleton.5 Although the synthesis of all-
trans Vitamin A has been successfully accomplished
using DBN,6 an improvement of this synthesis by the use
of 1 could be envisioned due to higher basicity of the
latter compound.2 Here we report on such an improve-
ment and propose a mechanism for the mode of action
of 1.
a
Reagents: (i) HCtC(CH3)CdCHCH2OH, 2BuLi, THF, then
water; (ii) LiAlH4 in ether; (iii) Ac2O in 2,4,6-collidine; (iv) PBr3
in ether, -20° to rt; (v) 1, DBN, or DBU.
Resu lts a n d Discu ssion
in Scheme 1 and parallel those described in the litera-
ture.6 However, each step of the sequence was optimized,
and the major byproducts of all the transformations were
separated and identified. The required aldehyde 2 was
obtained from â-ionone by introducing a carbon atom
from trimethylsulfonium methyl sulfate under phase-
transfer catalysis (PTC) conditions,7 followed by trans-
formation of the intermediate epoxide into 2 according
to a literature procedure8 in 98% overall yield. Attempts
Syn th esis. The major steps in the synthesis of vitamin
A via the elimination of hydrogen bromide are depicted
* To whom correspondence should be addressed. Fax: 515-294-0105.
† On leave of absence from the Institute of Chemistry, Medical
University, Lo´dz, Poland.
(1) Schmidt, H.; Lensink, C.; Xi, S. K.; Verkade, J . G. Z. Anorg. Allg.
Chem. 1989, 578, 75.
(2) (a) Verkade, J . G. Acc. Chem. Res. 1993, 26, 483. (b) Kisanga,
P.; Verkade, J . G.; Schwesinger, R. J . Org. Chem. 2000, 65, 5431.
(3) Arumugan, S.; Verkade, J . G. J . Org. Chem. 1997, 62, 4827.
(4) Liu, X.; Verkade, J . G. J . Org. Chem. 1999, 64, 4840.
(5) Verkade, J . G.; Mohan, T.; Wroblewski, A. E. U.S. Pat. 5698737;
Chem. Abstr. 1998, 128, 75536.
(6) Oediger, H.; Kabbe, H.-J .; Mo¨ller, F.; Eiter, K. Chem. Ber. 1966,
99, 2012. Oediger, H.; Eiter, K. Ger. Pat. 1157606; Chem. Abstr. 1964,
60, 5569a.
10.1021/jo010648+ CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/21/2001

P(CH3NCH2CH2)3N as a Dehydrobromination Reagent
J . Org. Chem., Vol. 67, No. 2, 2002 421
to synthesize pure 2 from â-ionone via the Darzens
reaction9 were unsuccessful, leading to mixtures contain-
ing 2 contaminated with its R,â isomer and other impuri-
ties which are possibly of a polymeric nature.
5b and diacetate 10b was accomplished when 2 equiv of
the anhydride and an excess of the base were employed.
Using chloroform or methylene chloride as a solvent and
pyridine as a base led only to incomplete conversion of
the starting diols 4 and excessive formation of diacetates.
An even more complex mixture was formed when acetyl
chloride was used as an acetylation reagent.
The synthesis of the retinol skeleton through a C14
+
C6 coupling has usually been accomplished by the addi-
tion of dimagnesium salts of 2-methyl-2-penten-4-yl-1-
ols to the aldehyde 2 or preferentially to its R,â isomer.6,10
Although the reaction is sluggish with 2, it could be
accelerated using lithium salts.11 Thus, the dilithium salt
of cis-2-methyl-2-penten-4-yl-1-ol (obtained by the addi-
tion of 2 equiv of butyllithium to a tetrahydrofuran (THF)
solution of the alcohol) was treated with 2 to give a 2:3
mixture of the 13-cis-11,12-didehydro-9,10-dihydro-10-
hydroxyretinols (9R*,10R*)-3a and (9R*,10S*)-3b, re-
spectively, in 85% yield of crude product, which was found
to be sufficiently pure, however, for the subsequent
transformation. The ratio of diastereoisomers, established
from 1H NMR spectral data, was independent of the
reaction conditions, as well as of the cation (i.e., Li+ or
Mg2+).
Synthesis of diastereoisomeric bromides 6 from the
secondary allylic alcohols 5 was most satisfactorily ac-
complished with phosphorus tribromide.6 Other bromi-
nating reagents, namely, PPh3/NBS14 and PPh3/CBr4,15
led only to traces of 6 in complex reaction mixtures.
Transformation of the hydroxy group to a bromide with
PBr3 also provided interesting regio- and stereochemical
results. Independent of the ratio of (9R*,10R*)-5a to
(9R*,10S*)-5b, approximately 10-30% of the rearranged
bromide 11 was formed. Furthermore, when pure 5a was
treated with PBr3, nearly complete conversion to a
mixture of (9R*,10R*)-6a and (9R*,10S*)-6b (3:1) con-
taining ca. 10% of 11 was observed. Under the same
conditions, pure 5b gave a mixture of 6a and 6b (3:7)
and 23% of 11.
The reaction of acetylenic bonds with LiAlH4 usually
leads to pure trans products, while the exclusive forma-
tion of cis isomers is expected when hydrogenation is
performed over deactivated palladium catalysts.12 In our
hands, the LiAlH4 reduction of 3 afforded a mixture of
the required diastereoisomeric 11,12-trans diols 4 and
11,12-cis diols 8 in a ratio of approximately 4:1 in overall
65% yield. This mixture was easily separated by silica
gel chromatography. In addition, a nonpolar fraction was
also obtained (5%), which was later identified as a liquid
mixture of four isomeric hydrocarbons, 9. Partial separa-
tion of the diastereoisomeric diols 4 was achieved by
column chromatography after which the more polar diol
4b (eluting second from the column) crystallized. The
ratio of (9R*,10R*)-4a to (9R*,10S*)-4b in the crude
reduction product reflected a 2:3 ratio of diastereoiso-
meric compounds 3. For NMR characterization, the
diastereoisomeric cis-diols 8 were separated on silica gel
after several chromatographic runs.
Initially we studied the conversion of mixtures of 6 and
11 into the isomeric retinyl acetates 9-cis-13-cis-7a and
9-trans-13-cis-7b in refluxing benzene (Table 1).6 It was
found that in the presence of equimolar amounts of 1,5-
diazabicyclo[4.3.0]non-5-ene (DBN) or 1,8-diazabicyclo-
[5.4.0]undec-7-ene (DBU), mixtures of diastereoisomers
6a and 6b were almost quantitatively transformed after
15 min into mixtures of 7a and 7b, while the rearranged
bromide 11 underwent elimination only partially. The
same transformation performed with DBN in refluxing
toluene left only a trace of unreacted 11 after 15 min.
Selective protection of the primary hydroxy groups in
4a and 4b as the corresponding acetates was achieved
with acceptable selectivity using an excess of acetic
anhydride in 2,4,6-collidine.13 Quantitative conversion of
pure 4b into a 9:1 mixture of monoacetate (9R*,10S*)-
Progress of the elimination of hydrogen bromide from
a mixture of 6a , 6b, and 11 with 1, DBN, or DBU in
acetonitrile at room temperature was monitored by H
(7) Bosansky, M.; Goemoery, J .; Kacina, R.; Styrakova, L. Czech Pat.
254032; Chem. Abstr. 1989, 111, P57525. Babler, H. J .; Schlidt, S. A.
Tetrahedron Lett. 1992, 33, 7697.
(8) Rosenberger, M.; J ackson, W.; Saucy, G. Helv. Chim. Acta 1980,
63, 1665.
(9) Ramamurthy, V.; Tustin, G.; Yau, C. C.; Liu, R. S. H. Tetrahe-
dron 1975, 31, 193.
(10) Isler, O.; Huber, W.; Ronco, A.; Kofler, M. Helv. Chim. Acta
1947, 30, 1911.
(11) Takahashi, J .; Oritani, T.; Yamashita, K. Agric. Biol. Chem.
1989, 53, 2711.
(12) Hudlicky, M. Reductions in Organic Chemistry; Ellis Horwood
Ltd.: Chichester, England, 1984. Seyden-Penne, J . Reductions by the
Alumino- and Borohydrides in Organic Synthesis; Wiley-VCH: New
York, 1997.
1
NMR spectroscopy. It was found that 1 is more reactive
than DBN or DBU, leading to disappearance of the
signals of 6 and 11 in 30 min, while complete conversion
with DBN was observed only after 1 h. Using DBU under
these conditions, traces of unreacted 6 and 11 still
remained after 1 h.
Purification of the products of the dehydrohalogenation
reactions included a water workup and filtration through
(13) Ishihara, K.; Kurihara, H.; Yamamoto, H. J . Org. Chem. 1993,
58, 3791.
(14) Bose, A. K.; Lal, B. Tetrahedron Lett. 1973, 3937.
(15) Calzada, J . G.; Hooz, J . Org. Synth. 1974, 54, 63.

422 J . Org. Chem., Vol. 67, No. 2, 2002
Wro´blewski and Verkade
products
Ta ble 1. Yield s a n d Ster eoch em ica l Resu lts of HBr Elim in a tion fr om th e Br om id es 6 a n d 11
starting materials
6a :6ba
11:(6a + 6b)b
reagent
solvent
time, min
7a :7bc
12:7d
yield, %
75:25
40:60
31:69
38:62
31:69
31:60
35:65
31:69
9:91
30:70
23:77
10:90
23:77
23:77
20:80
23:77
DBN
DBU
DBN
DBN
DBN
DBU
1
benzene
benzene
benzene
toluene
acetonitrile
acetonitrile
acetonitrile
acetonitrile
15
15
30
15
60
60
60
60
66:34
40:60
36:64
42:58
36:64
37:63
37:63
36:64
15:85
20:80
34:66
23:77
24:76
30:70
18:82
22:78
e
e
61f
50f
61f
63f
49f
52f
1
Based on relative intensities of H12 signals. Based on integrals of H9 signals. c Based on integrals of H3C18 signals. Based on
a
b
d
integrals of H14 signals. e Reaction proceeded to more than 95% completion. f For a mixture containing only 7a , 7b, and 12.
a pad of deactivated alumina to give mixtures of isomeric
retinyl acetates consisting of 7a and 7b as major products
(see Table 1). Elimination of hydrogen bromide from 6a
and 6b leads exclusively to the 13-cis-isomers 7a and 7b.
On the other hand, from the rearranged bromide 11,
formation of all-trans-retinyl acetate (12a ) and its 9-cis-
In the present work, the generation of -CH2CN is
isomer 12b was found as expected rather than the more
sterically strained 13-cis-isomers 7a and 7b. No improve-
supported by a 31P NMR study of the elimination of
hydrogen bromide from 6 and 11 with 1 in acetonitrile-
ment in the yield of vitamin A isomers was noticed when
d3. Before addition of the bromides, the 31P NMR spec-
the dehydrobromination of the bromides 6 and 11 with
trum of an acetonitrile-d3 solution of 1 showed signals of
1 in acetonitrile was carried out at room temperature
1 (singlet at 120.8 ppm) and of minute quantities of its
1
over 30 min. Comparison of the H NMR spectra of the
deuterated cation (three lines of equal intensity at -10.0
crude products from the DBU, DBN, and 1-induced
dehydrobrominations suggests that 1 in acetonitrile may
induce some decomposition of highly conjugated products.
When the mixture of the bromides 6 and 11 was treated
with 1 in acetonitrile at 0 °C for 2 1/4 h, incomplete
conversion was observed.
ppm).21 When the dehydrobromination21 reaction was
complete, a small excess of 1 was still present, while more
than 80% of 1 was converted to its deuterated form and
less than 20% was found as the protonated species.
Evidence for the 9-cis-configuration for 7a and 9-trans
for 7b is provided by their 1H NMR spectra. Although
Str u ctu r al, Ster eoch em ical, an d Mech an istic Con -
no differences between analogous coupling constants
sid er a t ion s. Initially, we were disappointed with the
among the H7-H8 and H10-H11-H12 sets of protons
lower dehydrobromination conversion rates using 1
in both isomers were found, it was noticed that H8 in 7a
compared with DBN or DBU in refluxing benzene. In
is deshielded by 0.5 ppm relative to H8 in 7b. The same
hindsight, this observation can be rationalized on the
phenomenon was found when the chemical shifts of H8
greater steric bulk around the basic phosphorus center
in all-trans and 13-cis-retinal and methyl retinoate were
in 1 compared with DBN and DBU, especially as positive
compared with the 9-cis- and 9-cis-13-cis-isomers, re-
charge is built up on 1 as it becomes protonated and
spectively.22 The presence of all-trans-12a and possibly
transannulated to the trigonal bipyramidal cation 13. In
also 9-cis-12b was concluded from the observation of H14
acetonitrile, however, the superiority of 1 over DBN and
triplets at δ 5.58 ppm22 and a long-range correlation of
DBU in effecting HBr elimination from 6 and 11 at room
this proton to H3C2 at 1.86 ppm22 in its 1H-1H COSY
temperature stems from
spectrum. Further evidence that 12a and 12b had been
the greater basicity of 1 (pKa of 13 = 33 in acetonitrile2b)
compared with the basicities of DBU (pKa = 23.9)16,17 and
DBN (pKa = 23.4).17 Although the pKa of 13 is the same
as a reported value of pKauto ) 33 of acetonitrile,18
evidence has been put forth for a value of 44 for this
autoprotolysis constant.16 In any case we have found that
nonionic bases such as 1 are capable of deprotonating
acetonitrile to -CH2CN with sufficient efficiency to allow
it to act as a sterically small but potent base in affecting
the dehydrohalogenation of primary, secondary and
tertiary alkyl halides19 and as a nucleophile in the
synthesis of 3-hydroxy nitriles from its reactions with
aldehydes and ketones.20
formed was the appearance of all of their 13C signals23
(except for some in the aliphatic region owing to overlaps)
in the 13C NMR spectra of the dehydrohalogenation
reaction mixtures.
Elimination of hydrogen bromide with strong bases
usually proceeds via an E2 mechanism with an anti
orientation of departing substituents24 (Scheme 2). Com-
parison of the ratios of starting bromides 6a and 6b with
the ratios of the products 7a and 7b (Table 1) as well as
the independence of the isomeric ratio of 7a and 7b from
polarity changes of the solvents provides support for the
E2 mechanism in our case. Thus, (9R*,10R*)-6a is
expected to give 9-cis-13-cis-7a , while 9-trans-13-cis-7b
should be formed from (9R*,10S*)-6b (Scheme 2). More-
over, the ratio of the bromides 6 and 11 are in reasonable
correlation with the ratios of the 13-cis-7 to 13-trans-12
(16) Kaljurand, I.; Rodima, T.; Leito, I.; Koppel, I. A.; Schwesinger,
R. J . Org. Chem. 2000, 65, 6202, and references therein.
(17) Leffek, K.; Pruszynski, P.; Thanapaalasingham, K. Can. J .
Chem. 1989, 67, 590.
(18) Kolthoff, I. M.; Chantooni, M. K., J r. J . Phys. Chem. 1968, 72,
2270.
(19) Mohan, T.; Arumugam, S.; Wang, T.; J acobson, R. A.; Verkade,
J . G. Heteroatom Chem. 1996, 7, 455.
(20) Kisanga, P. B.; McLeod, D.; D’Sa, B. A.; Verkade, J . G. J . Org.
Chem. 1999, 64, 3090.
(21) D’Sa, B. A.; Kisanga, P.; Verkade, J . G. J . Org. Chem. 1998,
63, 3961.
(22) Liu, R. S. H.; Asato, A. E. Tetrahedron 1984, 40, 1931.
(23) Englert, G. Helv. Chim. Acta 1975, 58, 2367.
(24) March, J . Advanced Organic Chemistry; J ohn Wiley & Sons:
New York, 1992; pp 982-1050.

P(CH3NCH2CH2)3N as a Dehydrobromination Reagent
J . Org. Chem., Vol. 67, No. 2, 2002 423
Sch em e 2
Sch em e 4a
a
Reagents: (i) PCC; (ii) NaBH4, MeOH/CH2Cl2, 1:1, 0°.
Sch em e 5
Sch em e 3
when unstable 14 was reduced with NaBH4 at 0 °C,34
a
70:30 mixture of (9R*,10R*)-5a and (9R*,10S*)-5b was
formed. Reports of precedents in the literature rationalize
the stereochemistry of hydride reductions of structurally
related R-methyl-â,γ-unsaturated ketones by invoking
the Felkin-Anh model26 in conjunction with π*CdO-π*CdC
attractive interactions that stabilize the transition state
of the nucleophilic addition.35-37 Application of this model
(Scheme 5) to the reduction of 14 rationalizes the
formation of (9R*,10R*)-5a as a major product. It has
been reported that the ketone 13-trans-14 can be reduced
isomers, supporting the conclusion that 11 is the pre-
dominant, if not the sole, source of the latter compounds.
Transformation of the OH in 5a and 5b to a bromide
in 6a and 6b (Scheme 2) undoubtedly involves an ionic
intermediate,25 because the formation of significant
amounts of the rearranged bromide 11 was observed and
because partial epimerization at C10 in 5a and 5b
occurred. The close resemblance of the 1H NMR patterns
of 5a and 6a , as well as those of 5b and 6b, suggests
that 6a is the major product of bromination of pure 5a ,
while 6b is predominantly formed from pure 5b. Thus,
in our case, phosphorus tribromide converts the second-
ary allylic alcohols 5 into bromides 6 with predominant
retention of configuration. To support this conclusion, the
relative configurations in the diastereoisomeric alcohols
3a and 3b and/or 4a and 4b were established. At first
we resorted to considerations of asymmetric induction in
the addition of (Z)-HCtC(CH3)dCHCH2OH to the car-
bonyl group of 2. When nucleophilic attack occurs from
the less hindered face in the conformation predicted from
the Felkin-Anh model,26 wherein the 2-(2,6,6-trimeth-
ylcyclohexene-1-yl)ethenyl group is viewed as the L
substituent, then (9R*,10S*)-3b predominates (Scheme
3). Although the diastereoisomeric excess (de) in this
addition is only moderate (20%), in reactions of closely
structurally related 2-substituted propanals with lithium
or magnesium salts of diverse acetylenes,27-32 de’s value
never exceeded 33%, and the configuration of the major
product was found to be in full agreement with that
predicted from the Felkin-Anh model.26
with NaBH4 and with AlH(OtBu)3,39 but diastereose-
38
lectivity of the reactions was not disclosed.
The 11,12-cis-configuration in diastereoisomeric 8a and
8b was deduced from the lower values of H11-H12
coupling constants (11.2 Hz) compared with those in the
respective 11,12-trans-isomers 4a and 4b (15.6 Hz). The
structures of the isomeric hydrocarbons 9 in the mixture
were deduced from 1H and 13C NMR spectral data,
including 1H-1H COSY experiments. Thus, the 13C NMR
spectrum clearly shows four signals for the terminal
methylene carbons in the 111.8-114.5 ppm region, and
it lacks signals for carbons attached to oxygen. Two-
1
1
dimensional H-1H correlation allowed us to assign H
resonances to two major constituents of the mixture 9a
and 9b having the 10,11-trans-configuration (J H10-H11
)
15.0 and 15.4 Hz, respectively). The cis arrangement of
the substituents around the C12-C13 bond in 9a can be
assigned on the basis of the 0.5 ppm downfield shift of
H14 relative to that in 9b. This conclusion is expected
from the structural resemblance of the C9-C15 and
C13-C7 subunits in the pairs of compounds 9a and 9b,
and 7a and 7b, respectively. The formulations of 9 were
supported by MS data. The CI-MS spectrum showed a
prominent peak for the molecular ion and two others
resulting from scissions at the two allylic positions.
The mechanism of formation of the isomeric hydrocar-
bons 9 has not been investigated, but neither 8 nor 4
afforded detectable amounts of 9 when treated with
An additional argument for the (9R*,10S*) configura-
tion in 3b, as well as, in 4b and 5b was gained from the
stereochemistry of the reduction of ketone 14 prepared
by PCC oxidation33 of the acetate 5b (Scheme 4). Thus
(25) March, J . Advanced Organic Chemistry; J ohn Wiley & Sons:
New York, 1992; pp 293-500.
(26) (a) Cherest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968,
2199. (b) Anh, N. T. Top. Curr. Chem. 1980, 88, 146.
(27) Oka, T.; Murai, A. Tetrahedron 1998, 54, 1.
(28) Yamamoto, Y.; Nishii, S.; Maruyama, K. J . Chem. Soc., Chem.
Commun. 1986, 102.
(34) Ward, D. E.; Rhee, C. K. Can. J . Chem. 1989, 67, 1206.
(35) Sato, F.; Takeda, Y.; Uchiyama, H.; Kobayashi, Y. J . Chem. Soc.,
Chem. Commun. 1984, 1132.
(29) Hirano Y.; Djerassi, C. J . Org. Chem. 1982, 47, 2420.
(30) Anastasia, M.; Ciuffreda, P.; Del Puppo, M.; Feicchi, A. J . Chem.
Soc., Perkin Trans. 1 1983, 383.
(31) Fujimoto, Y.; Kimura, M.; Teresawa, T.; Khalifa, F. A. M.;
Ikekawa, N. Tetrahedron Lett. 1984, 25, 1805.
(32) Craig, A. S.; Norman, A. W.; Okamura, W. H. J . Org. Chem.
1992, 57, 4374.
(36) Shimazaki, M.; Hara, H.; Suzuki, K. Tetrahedron Lett. 1989,
30, 5447.
(37) Suzuki, K.; Katayama, E.; Tsuchihashi, G. Tetrahedron Lett.
1984, 25, 2479.
(38) Baraldi, P. G.; Barco, A.; Benetti, S.; Guarneri, M.; Manfredini,
S.; Pollini, G. P.; Simoni, D. Tetrahedron Lett. 1988, 29, 1307.
(39) Baraldi, P. G.; Bigoni, A.; Guarneri, M.; Manfredini, S.; Pollini,
G. P.; Simoni, D. Farmaco 1993, 48, 1515.
(33) Corey, E. J .; Suggs, J . W. Tetrahedron Lett. 1975, 2647.

424 J . Org. Chem., Vol. 67, No. 2, 2002
Wro´blewski and Verkade
organic phase was washed with brine until neutral and was
dried over MgSO4. Solvents were evaporated and unreacted
alcohol was distilled off in vacuo (0.2 Torr, bath 95 °C). The
yellow oil left in the flask (7.046 g, 85%) was identified by NMR
spectroscopy as a mixture of 3a and 3b of sufficient purity
(TLC) to be used in the next step. A sample was purified on a
silica gel column with hexanes-ethyl acetate (10:1 to 3:1, v/v)
to give pure 3a and 3b (a 2:3 mixture) as a colorless oil.
lithium aluminum hydride for prolonged periods of time.
On the other hand, all four isomers of 9 were found in
the crude product when the 15-tetrahydropyranyl deriva-
tive of 3 was reduced with LiAlH4.
In conclusion, it was found that (9R*,10S*)-13-cis-11,-
12-didehydro-9,10-dihydro-10-hydroxyretinol (3b) was
the major product of the addition of (Z)-LiCtC(CH3)d
CHCH2OLi to the carbonyl group of (E)-2-methyl-4-
(2′,6′,6′-trimethyl-1-cyclohexen-1′-yl)-3-butenal (2). In the
presence of PBr3 the 15-acetates of diols (9R*,10R*)-4a
and (9R*,10S*)-4b were transformed into 13-cis-10-
bromo-9,10-dihydroretinyl acetates (6) with significant
retention of configuration. Proazaphosphatrane 1 in
acetonitrile eliminates HBr from vitamin A precursors
6 and 11 faster than DBU and DBN.
Red u ction of 3 w ith LiAlH4. To a suspension of LiAlH4
(1.736 g, 45.75 mmol) in ether (75 mL) cooled under argon to
2 °C, a solution of 3 (7.05 g, 23.3 mmol) in ether (25 mL) was
added dropwise at a rate just slow enough to keep the
temperature of the reaction mixture below 6 °C. Then the
mixture was stirred at room temperature for 16 h. After cooling
to -10 °C, water was slowly added until the vigorous reaction
ceased. The reaction mixture was acidified with dilute H2SO4
at 0 °C, the ether layer was separated, and the water phase
was extracted with ether (2 × 50 mL). The combined organic
phases were washed with water, aqueous NaHCO3, then water
again until neutral, dried over MgSO4, and evaporated. The
residue was kept in vacuo (0.1 Torr) at room temperature for
48 h to give 6.779 g of a very viscous yellow oil which was
subjected to chromatography on silica gel (50 g) with hexanes-
ethyl acetate mixtures to give a mixture of isomers 9 (oil, 0.315
g, 5%), a mixture of diastereoisomers 8 (0.898 mg, 12.6%), a
partially separated mixture of diastereoisomers 4a and 4b
(2.224 g, 31.3%), and crystalline 4b (1.503 g, 21.2%) after
solvent evaporation. Diastereoisomers 8 were further chro-
matographed on silica gel with 10:1 hexanes-ethyl acetate
(v/v) to afford pure samples of 8a and the more polar 8b.
Syn th esis of Mon oa ceta tes 5a a n d 5b. A mixture of 4b
(1.397 g, 4.588 mmol), 2,4,6-collidine (3.03 mL) and acetic
anhydride (0.87 mL, 9.17 mmol) was left at room temperature
for 48 h. After dilution with ether (50 mL) the mixture was
washed with cold diluted H2SO4, water, aqueous NaHCO3, and
water again, and then it was dried over MgSO4. Chromatog-
raphy on silica gel with 10:1 and 5:1 hexanes-ethyl acetate
(v/v) afforded diacetate 10b (174 mg, 9.5%, oil) and 5b (1.440
g, 89.5%, oil). In an analogous fashion, a mixture of 4a and
4b was esterified. Chromatography on silica gel with hexanes-
ethyl acetate mixtures (100:5 to 2:1, v/v) allowed partial
separation of 5a and 5b.
Exp er im en ta l Section
Unless otherwise noted, materials were obtained from
commercial suppliers and were used without purification.
Solvents were reagent grade, predried over molecular sieves,
and when necessary, distilled from sodium-benzophenone
ketyl prior to use (THF, ether). Removal of solvents from oily
reaction products was carried out by applying vacuum and
magnetically stirring at room temperature until 20 mTorr was
achieved. However, residual hexane was still present, espe-
cially for viscous compounds as judged from their 1H NMR
spectra. Efforts to remove hexane by heating under vacuum
resulted in decomposition.
1H NMR spectra were measured on Nicolet NT-300 and
Varian VXR-300 NMR spectrometers in chloroform-d, while
13C and 31P NMR spectra were recorded on a Varian VXR-300
NMR spectrometer in chloroform-d and acetonitrile-d3, respec-
tively. Chemical shifts are reported in parts per million
downfield from tetramethylsilane using chloroform (1H, 7.23
ppm) and chloroform-d (13C, 77.07 ppm) resonances as second-
ary standards. 1H and 13C chemical shift assignments were
supported by 2D 1H-1H correlations performed on 2, 4b, 8a ,b,
6, and 11 as a mixture, 9, and mixtures of 7a and 7b, and by
1
2D H-13C correlations recorded for 4b, 8b and a mixture of
Rea ction of P h osp h or u s Tr ibr om id e w ith 5 (Gen er a l
P r oced u r e). To a solution of 5 (10 mmol) in ether (10 mL)
PBr3 (12 mmol) was injected at -20 °C under argon. The
reaction mixture was stirred for 1 h while allowing it to reach
room temperature during that time. It was then diluted with
ether (40 mL), washed with cold brine, aqueous NaHCO3, and
then brine until neutral and dried over MgSO4. After evapora-
tion of the ether, the crude product was left in vacuo (0.02
Torr) to give a mixture of 6a , 6b, and 11 as a yellow oil in
85-95% yield. This material slowly turned brown when left
at room temperature, and for this reason it was used im-
mediately in the next step.
7a and 7b. Two-dimensional spectra were obtained on a Varian
VXR-300 spectrometer using standard COSY and HETCOR
experiments. The CI-MS spectrum of 9 was determined on a
Finnegan 4023 mass spectrometer.
For preparative chromatographic separations, silica gel (60-
200 mesh, EM Science) was used, except for vitamin A isomers,
for which only deactivated alumina (neutral, Baker) was found
useful. TLC analyses were performed using plates precoated
with silica gel (IB-2 and IB-F) or alumina (IB-F) (both Baker-
flex from Baker). The following solvent systems were em-
ployed: 3:1 hexanes-ethyl acetate (v/v, for silica gel plates)
and 10:1 hexanes-ethyl acetate (v/v, for alumina plates).
Elim in a tion of HBr fr om a Mixtu r e of 6a , 6b, a n d 11
w ith 1, DBN, or DBU (Gen er a l P r oced u r e). A mixture of
6a , 6b, and 11 obtained in a previous experiment was dissolved
in benzene or toluene (1 mmol in 1 mL) and was refluxed with
1.2 equiv of 1, DBN, or DBU. Alternatively, the mixture of
bromides was dissolved in acetonitrile (1 mmol in 1 mL) and
was stirred at room temperature with 1.2 equiv of 1, DBN, or
DBU. The reaction mixture was diluted with ether and washed
with brine until neutral. The crude products were filtered
through deactivated alumina (12 g for 5 mmol) to give mixtures
of 7a and 7b (major) and 12a and 12b (minor).
Syn th esis of th e Keton e 14. The acetate 5b (0.193 g, 0.55
mmol) was dissolved in CH2Cl2 (2 mL) and then PCC (0.178
g, 0.83 mmol) was added portionwise at 0 °C. After 3 h at 0 °C
hexane (10 mL) was added, the organic solution was with-
drawn by pipet and concentrated. The yellow residue was
chromatographed on silica gel (5 g) with hexanes-ethyl acetate
(20:1, v/v) to give 4 (60 mg, 31%) as a yellowish oil. On the
basis of the NMR spectra (see Supporting Information) the
material was judged to be ca. 90% pure.
Trimethylsulfonium methyl sulfate was prepared (88%
yield; lit.,40 70-75% yield), 2-methyl-2[2′-(2′′,6′′,6′′-trimethyl-
1′′-cyclohexen-1′′-yl)ethenyl]oxirane (98.3% yield, lit.8 94%
yield), and (RS)-(E)-2-methyl-4,2′,6′,6′-trimethyl-1′-yl)-3-bu-
tenal (2) (quantitative; lit.8 85% yield) were prepared according
to methods in the cited literature.
13-cis-11,12-Did eh yd r o-9,10-d ih yd r o-10-h yd r oxyr etin -
ols (3). The dilithium salt of cis-2-methyl-2-penten-4-yl-1-ol
was prepared from the alcohol (3.715 g, 38.6 mmol) and
n-butyllithium (30.9 mL, 2.5 M in hexane) in tetrahydrofuran
(100 mL) below -15 °C. After stirring for 1 h below -15 °C, a
room-temperature solution of 2 (5.63 g, 27.3 mmol) in THF
(20 mL) was added while the reaction flask temperature was
maintained below -15 °C. The reaction mixture was stirred
for 1 h at this temperature, and then it was allowed to warm
to room temperature. After addition of cold water (100 mL)
the product was extracted with ether (2 × 200 mL). The
(40) Mosset, P.; Gree, R. Synth. Commun. 1985, 15, 749.

P(CH3NCH2CH2)3N as a Dehydrobromination Reagent
J . Org. Chem., Vol. 67, No. 2, 2002 425
Red u ction of th e Keton e 14. To a solution of 14 (77 mg,
0.22 mmol) in a 1:1 mixture of CH2Cl2 and methanol, NaBH4
(8.4 mg, 0.22 mmol) was added portionwise at 0 °C. After 30
min, hexane (20 mL) was added, followed by aqueous H2SO4
(6%) (0.1 mL). The ice-water bath was then removed and the
reaction mixture was dried with MgSO4. The crude product
was chromatographed on silica gel (5 g) with hexanes-ethyl
acetate (10:1, v/v) to give a 7:3 mixture of 5a and 5b,
respectively. 1H NMR integrals of H3C16,17, H3C19, H3C18,
H3C20, HC10, and HC12 in 5a and 5b were compared and
averaged (see Discussion).
Ack n ow led gm en t. We thank the National Science
Foundation and the donors of the Petroleum Research
Fund, administered by the American Chemical Society,
for grant support.
Su p p or tin g In for m a tion Ava ila ble: 1D and 2D NMR
spectra and NMR and mass spectral assignments. This mate-
rial is available free of charge via the Internet at http://pubs.org.
J O010648+