J. Org. Chem.

1991,56,1373-1381 1373
6 20.6,22.0,30.8,43.9,47.5,70.0,70.9,83.1,85.0, 102.1,113.6, 114.2, 129.2, 129.3, 129.5, 135.7, 136.3, 137.2, 148.4, 152.3, 153.0; mas8
115.2, 126.5, 127.5,127.7,127.8,128.1,128.3,128.4, 128.7, 129.0, spectrum, m / z (relative intensity) 595.2209 (calcd for C a S N 0 8
133.5, 137.2, 138.5, 149.1, 152.6,171.1,200.6; mass spectrum, m/z 595.2206), 580 (121,504 (8), 254 (151,234 (241,160 (791,144(32),
(relative intensity) 579.2635 (M+,calcd for CsH3,NOB579.2621), 105 (100). Cycloadduct 41. Crystallization of 41 in hexane/
382 (l),331 (5), 309 (l), 255 (l),218 (l),190 (l),148 (2), 128 (3), methylene chloride (91) afforded single crystals suitable for X-ray
91 (100). analysis: mp 137-138 "C; IR (CHCl3) 1817 (vs), 1491 (vs), 1378
Vinylidene Carbonate Cycloadducts 40 and 41. Freshly (m); 'H NMR (CDC13) 6 1.76 (8, 3 H), 2.65 (s, 3 H), 4.00 (d, J =
distilled vinylene carbonate (2.71 g, 2.00 mL, 0.03 mmol) was 4.4 Hz, 1H), 4.50 (d, A of AB q, J = 11.5 Hz, 1H), 4.61 (d, J =
transferred to a sealed tube apparatus containing nitrone 8B (692 4.4 Hz, 1 H), 4.71 (d, B of AB q, J = 11.5 Hz, 1 H), 5.03 (AB q,
mg, 1.36 mmol) in 10 mL of anhydrous xylenes. The mixture was J = 9.6 Hz, 2 H), 5.63 (d, J = 5.3 Hz, 1 H), 6.10 (d, J = 5.3 Hz,
heated at 120 OC for 72 h and then concentrated in vacuo. Flash 1 H), 6.76 (dd, J = 8.9, 3.0 Hz, 1 H), 6.85 (d, J = 8.9 Hz, 1 H),
chromatography of the residue (21 hexane/ethyl acetate) provided 7.03 (d, J = 3.0 Hz, 1 H), 7.33 (m, 15 H); 13C NMR (50 MHz) 6
a 2:1 mixture of low Rf isomer 41 to high R, isomer 40 for a 21.9, 68.8, 70.2, 71.2, 80.8, 84.8, 87.2, 102.8, 103.9, 113.1, 113.6,
combined yield of 55%. The diastereomers were separated by 115.1, 126.5, 127.1, 127.5, 127.8,128.0,128.4, 128.5, 128.6, 128.9,
flash chromatography to afford 160 mg (18%) of high Rf isomer 129.2,133.9, 136.1, 137.1,138.2,148.6,152.7,152.9; mass spectrum,
40 and 325 mg (37%) of lower Rf isomer 41. m / z 595.2195 (calcd for C35H33N08595.2206), 505 (29), 504 (89),
High-Pressure Cycloaddition of Nitrone 8B with Vinylene 296 (19), 254 (18).234 (32), 168 (20), 163 (27), 144 (59), 105 (100).
Carbonate. Freshly distilled vinylene carbonate (0.50 mL, 0.678 Anal. Calcd for CSHsNO8: C, 70.58; H, 5.58. Found: C, 70.17;
g, 7.87 mmol) was added to a solution of nitrone 8@in 3 mL of H, 5.57.
anhydrous THF. The solution was placed into a syringe and then
into a high-pressure reactor. The vessel was pressurized to 12 Acknowledgment. We thank the National Institutes
kbar and allowed to react for 72 h. The syringe was removed and of General Medical Sciences (GM 37014) for generous
washed with EtOAc, and the solvent was removed in vacuo. Flash financial support of this program. Dr. Michael Dicken,
Chromatography of the residue (21hexanes/EtOAc) afforded an T h u y Le, Dr. Steve Lander, Jr., and Dr. Grady Lamb are
W20 mixture of diastereomers in 60% yield that were separated acknowledged for helpful discussions during the course of
as above. this investigation. Finally, we acknowledge the support
Cycloadduct 40. Crystallization of 40 in hexane/methylene of Dr. Masood Parvez (X-ray, Pennsylvania State Univ-
chloride (91) afforded single crystals suitable for X-ray analysis: ersity), Dr. Yiu-Fai Lam (NMR), and Caroline Preyer (MS)
mp 200-201 OC; lR (CHClJ 1819 (v8),1497 (m), 991 (m); 'H NMR in obtaining spectral data.
(CDCls) 6 1.73 (s,3 H), 2.64 (9, 3 H), 3.56 (d, J = 2.5 Hz, 1 H),
4.41 (d, J = 2.5 Hz, 3 H), 4.87 (AB q, J = 11.7 Hz, 2 H),5.01 (AB Supplementary Material Available: 'H NMR spectra of
q, J = 9.8 Hz, 2 H), 5.23 (d, J = 5.3 Hz, 1 H), 5.92 (d, J = 5.3 compounds 8a,8@, 11,17, 18,24-30,33,34,38, and 39; ORTEP
Hz, 1 HI, 6.21 (s, 1 H), 6.87 (dd, J = 8.8, 3.0 Hz, 1 H), 6.97 (d, diagrams and X-ray data of compounds a@,28,40, and 41; and
J = 8.8 Hz, 1 H), 7.29 (d, J = 3.0 Hz, 1 H), 7.36 (m, 15 H); 13C tables of fractional coordinates and temperature factors, bond
NMR (50MHz) 6 22.6,47.4,70.4,71.0,71.3,77.2,84.7,90.1, 103.3, distances in angstroms, and bond angles in degrees (44 pages).
105.2, 112.3, 112.8, 114.4, 127.1, 127.5, 127.9, 128.4, 128.5, 129.1, Ordering information is given on any current masthead page.

Anthracenediols as Reactive Dienes in Base-Catalyzed Cycloadditions:

Reductionxycloaddition Reactions of Anthraquinones'
Michael Koerner and Bruce Rickborn*
Department of Chemistry, University of California, Santa Barbara, California 93106
Received August 7, 1990

Anthraquinone is readily reduced to the hydroquinone (9,10-anthracenediol), which under basic conditions
serves as a reactive diene for cyloaddition purposes. Catalytichydrogenation in pyridine solvent provides convenient
access to this species, and efficient reactions occur with dienophiles in situ, provided that they are sufficiently
reactive. Thus N-methylmaleimide (NMM) gives the bicyclic bridgehead diol in near quantitative yield when
the H2/Pd reduction of anthraquinone is carried out in pyridine containing 1 equiv of NMM. Fumaronitrile
and maleonitrile similarly give high yields in stereospecific reactions, with the dienophile geometry retained in
the cycloadduct. Less reactive dienophiles suffer competitive reduction. Dimethyl fumarate in situ givea cycloadduct
(stereospecifically) in only 35-60% yield, with the remainder of the dienophile reduced to dimethyl succinate.
Stepwise reduction followed by addition of dienophile leads to a higher yield in this and related reactions. The
benzologues 5,12-naphthacenedione and 6,13-pentacenedione undergo analogous reactions with NMM, leading
to novel bridgehead diols. The monimine of anthraquinoneexhibits NMR features attributed to syn/anti isomerism.
Under neutral or mildly basic conditions, the aromatic protons on the ring proximal to the NH are clearly
distinguished (500 MHz) from those on the distal ring. The addition of acid causes rapid syn/anti NH exchange
leading to time averaged symmetry. This imine behaves similarly to anthraquinone in the reduction/cycloaddition
sequence. For example, with NMM in situ an essentially quantitative yield of the qovel bridgehead amino alcohol
adduct is obtained. Related benzologue reactions and attempts to extend the sequence to the oxime and methylene
analogues of anthraquinone are described. Base-catalyzed ring opening of the cycloadduct of NMM/anthracenediol
leads to a novel retro-bis-aldol reaction, resulting in the formation of anthraquinone and N-methylsuccinimide.

Introduction hydrogen-bonded variant is believed to be the intermediate

Novel base-catalyzed Diels-Alder reactions of 9-anthrone responsible for t h e very rapid cycloadditions which are
have recently been described.'Z T h e oxyanion or an amine observed in the presence of base. T h e possibility that
hydroquinones and other substituted anthrone analogues
(1) A preliminary communication has appeared: Koerner, M.; Rick-
born, B. J. Org. Chem. 1989,54, 6. (2) Koerner, M.; Rickborn, B. J. Org. Chem. 1990, 55, 2662.

0022-3263/91/ 1956-1373$02.50/ 0 0 1991 American Chemical Society

1374 J. Org. Chem., Vol. 56, No. 4, 1991 Koerner and Rickborn
might exhibit similar behavior led to the present study. bonded species such as 3a-c ( b N = pyridine) are formed?
Hydroquinone itself has been used in (uncatalyzed) An uoxyanionncomplex of this kind is believed to be the
Diels-Alder reactions by Nakazaki et al.3 and by Jefford reactive intermediate responsible for the reactions de-
et al? to give products such as that shown in eq 1, in which scribed below.
cycloaddition takes place across the 2,5-positions of the
aromatic ring. This regiochemistry may be the result of H --NRa -O..HNFS
thermodynamic control, since cycloaddition across the
1,4-positions would likely be readily reversible for simple
hydroquinones.
I

''H--NR, O'H-.Nb
3a 3b 3c
Addition of N-methylmaleimide (NMM) to the dark
solutions leads to the formation of the cycloadduct 4.
Apparently no other arene 1,4-diols have been used as Interestingly, catalytic reduction of a mixture of an-
cycloaddition partners, although Krohn and Meihesa have thraquinone and NMM in pyridine also gives this prod-
described reactions of substituted anthracenediols which uct, in excellent yield (eq 2).
in principle could have given Diels-Alder adducts although 0
none were found (see later discussion). The parent 9,lO- n 1
MeN S
anthracenediol (1) appears to have escaped examination
as a prospective diene, although this material has been
known for many years.6
pyridine
"
PH 8 4 95%

Successful application of this in situ dienophile proce-

dure requires that both reduction of anthraquinone and
AH bR cycloaddition be faster than hydrogenation of the dieno-
1 2r, R = Me phile. Reduction of the dienophile can certainly occur
2b, R = SiMe3 under these conditions, as shown by the formation of
The dimethyl ether derivative 2a has been studied by N-methylsuccinimide from any excess NMM employed.
Sauer and Wiest? who found it to be somewhat less re- NMM is a very reactive dienophile, as shown in the
active than the parent hydrocarbon in cycloaddition with base-catalyzed cycloaddition to anthrone reported earlierm2
maleic anhydride. More recently Czarnick et al.8 prepared In that reaction the intermediate oxyanion is trapped so
the cycloadduct of ethyl acrylate with the bis(trimethy1- rapidly by NMM that proton transfer from anthrone to
si1oxy)anthracene 2b. These authors did not comment on Et3N becomes rate determining. For the reaction shown
any unusual reactivity features. The bis(trimethoxysiloxy) in eq 2, reduction of anthraquinone is probably the rate-
cycloadduct was subsequently converted to the corre- limiting step. The fact that anthraquinone is reduced more
sponding bridgehead diol, thereby providing an alternative rapidly than NMM can be attributed to preferential cat-
synthetic route to diols similar to those described below. alyst interaction and/or lower reduction potential.
The particular advantage of the in situ dienophile pro-
Results and Discussion cedure stems from the very facile air oxidation of an-
Catalytic reduction of anthraquinone in pyridine solvent thracenediol 1 under basic conditions, which tends to in-
gives a very dark solution. I t is assumed, in analogy with terfere with the stepwise method. However, the latter is
base-catalyzed reactions of anthrone, that hydrogen preferred or needed for less reactive dienophiles, as out-
lined in some examples discussed below.
Fumaronitrile and maleonitrile are also very reactive
(3) Nakazaki, M.; Naemura, K.; Yoshihara, H. Bull. Chem. SOC. Jpn.
1976,48, 3278. dienophiles and work well in the in situ mode with an-
(4) Jefford, C. W.; Wallace, T. W.;Acar, M. J . Org. Chem. 1977, 42, thraquinone. These reactions are stereospecific, leading
1654. exclusively and essentially quantitatively to cycloadducts
(5) (a) Krohn, K.; Miehe, F. Liebigs Ann. Chem. 1986,1329. (b) The 5 and 6, respectively, with no indication of the alternative
reaction of the parent anthracenediol (anion) with MVK was described
somewhat earlier: Dimmel, D. R.; Shepard, D. J . Org. Chem. 1982,47, stereoisomer being formed in either reaction (by NMR, it
22. is estimated that any amount over 1%would have been
(6) Anthracene-9,lO-diol was first generated as the red dianion by detected). These results parallel those reported earlier for
Zn/NaOH reduction of the quinone in 1870 (Boettger,R. J . h k t . Chem.
1870, 2, 133). Subsequently, several reducing agenta including most the base-catalyzed reactions with anthrone and are likewise
commonly Na&O, in NaOH solution (Grandmougin,E. Chem. Ber. 1906, thought to be the result of concerted cycloaddition.
39, 3563) and H2/Pt (Manchot, W.; Gall,H. Chem. Ber. 1925,58,486) Dimethyl fumarate is estimated to be ca. as reactive
have been employed. The diol is rapidly air oxidized to the quinone,
especially under basic conditions. For more recent applications of pre- as NMM. This difference is sufficient to introduce dif-
formed 9JO-anthracenedio1, see ref 6b and Landucci, L. L.; Ralph, J. J . ficulties into the in situ dienophile approach. Under the
Org. Chem. 1982, 47, 3486. The latter papers describe reactions with same conditions (slight excess of dienophile) that led to
p-quinone methides, which have been suggested as models for the cata-
lytic effect of anthraquinone on delignification of wood under reductive quantitative yields of cycloadducts with the more reactive
Conditions. Cycloaddition-cleavagepathways similar to those described dienophiles, dimethyl fumarate gave yields of only 35-6070
in the present work are conceivable mechanistic alternatives to the Mi- of 7 in different runs. The remainder of the dienophile
chael additions depicted in these papers, and indeed this possibility has
previously been suggested by Ralph (J. Ralph, 1989, Proceedings of the was reduced to dimethyl succinate, presumably by com-
Intl. Symposium on Wood Q Pulping Chemistry, N.C. State University, petitive catalytic hydrogenation (see later discussion).
Raleigh, NC, May 22-25,1986, Session 4-2, p 61). We thank Dr. Ralph
for calling this work to our attention.
(7) Sauer, J.; Wiest, H. Angew. Chem., Int. Ed. Engl. 1961, 1, 269. (9) Similar hydrogen-bonded complexes have been postulated by Bahn
(8)Chung, Y.;Duerr, B. F.; McKelvey, T. A,; Nanjappan, P.; Czamik, and Takemura for hydrocarbon solutions of anthrone to which EQN has
A. W. J . Org. Chem. 1989,54, 1018. been added: Baba, H.; Takemura, T. Tetrahedron 1968,4779.
Reduction-Cycloaddition Reactions of Anthraquinones J. Org. Chem., Vol. 56, No. 4, 1991 1375

& 0

e
5 94%

NCN+

(4)

6 98% 8. 8b
Although the yield of cycloadduct 7 is modest the trans of NMM, as shown by the isolation of cycloadduct 9 in
stereochemistry of the fumarate is retained in the product. 42% yield.
0

e 7 35M)%
+ Me02CCH2CH2C02Me (5)

Competitive reduction of the dienophile can easily be 9

avoided by separating the reduction and cycloaddition By analogy with the corresponding hydrocarbons, one
steps. Application of this sequential approach to the re- might expect the pentacenediol intermediate to be more
action of anthraquinone and dimethyl fumarate (1equiv) reactive (unstable) than anthracenediol. However, this
caused the yield of 7 to rise to 80%. The procedure was reduction state proved to be readily isolable, apparently
the same as for the in situ dienophile reaction, except that because of strong preference for the keto tautomer struc-
the quinone was stirred under H2for 0.5 h, after which the ture 10, a conclusion supported by the NMR spectrum of
atmosphere was replaced by Ar, and then the dienophile this material.
was introduced. OH 0
Direct evidence for base catalysis was obtained by ex-
amining the NMR spectra of a mixture of preformed 1 and
dimethyl fumarate in acetone-d, solvent under various
conditions. Neither the neutral mixture nor one treated
with acid gave any indication of reaction after 2 h. How-
ever, an aliquot treated with Et3N (ca. 5 equiv) gave 30% 10
of cycloadduct in this same time period. This result also Isolated compound 10 serves nicely as a “diene” under
has significance for the related reactions of anthrones, basic conditions, leading to 9 in excellent yield. The en-
reinforcing the conclusion2that the phenol tautomer itself couraging results with NMM in this stepwise procedure
is not the highly reactive species for cycloaddition. suggest that polycyclic quinones may also be employed
Linear benzannulation is known to enhance the re- with less reactive dienophiles, although this point was not
activity of aromatic hydrocarbons toward dienophiles. examined in the present study.
Thus, the reactivity order pentacene > naphthacene > Anthraquinone Monimine: NMR and Reactions.
anthracene is observed for cycloaddition with maleic an- The monimine (12) appears to be the only analogue of
hydride.l0 However, 5-naphthacenone proved to be less anthraquinone with’the same oxidation state and general
reactive than anthrone with NMM under base-catalyzed structure that has been reported. It has been prepared”
cycloaddition conditions.2 We attributed this inversion by LiA1H4 reduction of the oxime 11 as shown in eq 9.
of normal reactivity to diminished concentration of the
reactive oxyanion intermediate. It was not clear from this R
experience what to expect for benzannulated anthra-
quinones, but the experiments described below indicate
that high reactivity toward NMM is retained.
N
Catalytic reduction of a mixture of 5,12- ‘OH
naphthacenedione and NMM in pyridine gave two prod- 11 12 979’0
ucts identified as exo and endo cycloadducts 8a and 8b.
These two very similar materials are formed in a ratio of We employed both this method and an alternative
1.5/1; it is not known which is the major isomer. catalytic reduction procedure described below to obtain
The reaction with 5,13-pentacenedione is even more 12. Very different NMR spectra were exhibited by prod-
striking because this material has only slight solubility in ucts from these methods, although other evidence sug-
pyridine. In spite of this handicap, dissolution, reduction, gested that the materials were the same. These initially
and cycloaddition all occur in competition with reduction puzzling observations are attributed to the large chemical

(11) Costa, A.; Riego, J. M.; Garcia-Raso, A.; Sinisterra, J. V. Liebigs

(10)Biermann, D.; Schmidt, W. J. Am. Chem. SOC.1980,102, 3163. Ann. Chem. 1981, 2085.
1376 J. Org. Chem., Vol. 56,No.4, 1991 Koerner and Rickborn

/---

f b
_.._

Figure 1. Spectra were obtained at 500 MHz in CDC13solvent. A illustrates 12 as isolated from the LiAlH, reduction of oxime 11.
B shows 12 which has been recrystallized from acetone; similar spectra were obtained when 12 was prepared by catalytic reduction
of 11. Spectra C through G show the effects of adding increasing amounts of trifluoroacetic acid (TFA) to the NMR sample used to
obtain B. The amounts of TFA added were: 4 p L (C),8 p L (D), 12 p L (E), 20 p L (F), and a large excess, ca. 100 pL (G),where 1
molar equiv of TFA = ca. 10 pL. Spectrum A exhibited a sharp singlet at 11.0 ppm attributed to the NH proton. In B this peak was
broadened but not measurably shifted. This signal was not seen in the other spectra, presumably due to exchange with the TFA.
shift differences of the aromatic ring protons caused by material is recrystallized (acetone or isopropyl alcohol), or
proximity (syn, anti) of the imine proton (NH). These prepared by catalytic reduction of the oxime, spectrum B
NMR effects are shown in Figure 1. Spectrum A shows or similar results. In both spectra the NH proton appears
the region from ca. 7.6 to 8.8 ppm for material isolated as a singlet at 11.0 ppm, sharp in A and broadened in B.
without any purification step from the LiAlH4 reduction Figure 1C shows the effect of addition of a small amount
procedure. This spectrum was reproducibly observed from (50.5equiv) of trifluoroacetic acid (TFA), and D-G show
various samples of 12 prepared in this manner. When this the changes which result from adding increased amounts
Reduction-Cycloaddition Reactions of Anthraquinones J. Org. Chem., Vol. 56,No. 4,1991 1377
Scheme I is nitrosoanthracenol 19,and in fact we found that 11 can
be prepared by treatment of anthrone with nitrous acid,
following a general procedure for nitrosation of active
methylene compounds.12 An initial attempt to effect
cycloaddition of 11 with NMM was done in DMF solvent,
which had been used previously to effect NMM cyclo-
addition with unsubstituted anthroneS2No reaction was
16 39% \ I 13 100% observed, even when the mixture was heated to 100 OC.
Pyridine solvent was also used, but again no loss of starting
materials was detected. The failure of 11 to undergo
base-catalyzed cycloaddition may be due to very unfa-
vorable equilibria, as outlined in eq 10. Only the oxime
form 11 is observed in NMR spectra of this material taken
in CDC13.
12 0

E.i H- -NO
O
'H
11
15 92% 14 90% OH

of TFA. The smallest amount of TFA caused the 8 aro-

matic proton spin system of A to simplify to the 4 proton
spin system seen in C, as expected for rapid syn/anti imine
interconversion. However, something in addition to 19
syn/anti interconversion must occur, since the new chem-
ical shifts are not simply the averages of the peaks in A The use of 11 under reductive conditions was informa-
or B. Contributions from protonated species are reason- tive even though no new product was obtained. Instead,
able, but this alone does not seem to explain the shift when an equimolar mixture of 11 and NMM in pyridine
phenomena obse~ed,ie., large changes in the triplet shifts was subjected to catalytic hydrogenation, approximately
caused by a small amount of acid, but no appreciable 20% of the cycloadduct 13,identical with product derived
change as more is added. It appears that a change in the from the monoimine 12,was formed, while most of the
state of aggregation of 12 is also needed to rationalize this NMM was reduced to N-methylsuccinimide. Either the
behavior. reduction of 11 or the interconversion of its possible in-
The imine 12 behaves much like anthraquinone when termediate reduction products to 12 occurs more slowly
subjected to catalytic reduction (H2, Pd/C, pyridine) in than reduction of NMM. It was independently established
the presence of reactive dienophiles. Thus NMM, fu- that 11 gives 12 under these hydrogenation conditions.
maronitrile, maleonitrile, and dimethyl fumarate give the The dark red diazo analogue 2016 was also examined
cycloadducts 13,14,15,and 16,respectively. The reactions under the reductive cyclization conditions with NMM in
are stereospecific as shown. The yields are excellent except situ. No bridgehead nitrogen-bearing product was ob-
for the in situ dimethyl fumarate reaction, another parallel served. Instead, the cycloadduct 21,which has been pre-
between anthraquinone and its monimine. Again the yield viously prepared2 by base-catalyzed reaction of anthrone
of this cycloadduct can be improved by the sequential with NMM, was isolated in moderate yield. A control
reduction-cycloaddition approach, showing that the am- experiment without dienophile established that anthrone
inophenol has a reasonable lifetime under these conditions. is formed rapidly from 20 under these reductive conditions,
The monimine (17)of pentacenequinone was prepared and the stepwise procedure applied to 20 gave 21 in near
by reduction of the monoxime, which was obtained by quantitative yield.
treatment of the quinone with hydroxylamine following
the literature method for the anthraquinone analogue. The
in situ dienophile procedure with NMM gave the bridge-
head aminoalcohol derivative 18.
0
21 45%
20

The methylene derivative 22 was of interest both as a

quinone analogue and via reduction as a substituted an-
HN
, throne (23). The usual catalytic hydrogenation of a mix-
17
ture of 22 and NMM in pyridine gave no cycloadduct.
Instead, both starting materials were reduced to give
N-methylsuccinimide and 10-methylanthrone (23), re-
spectively.
The substituted anthrone 23 was subjected to conditions
18 87% (NMM, THF, Et3N, 25 "C) which give very rapid cyclo-
Other Quinone Analogues. The oxime 11 is an in- adduct formation with anthrone itself? but no reaction was
teresting candidate for similar reactions, both directly and
under reductive conditions. One tautomeric form of 1 1 (12) Touster, 0.Org. React. 1953, 7, 327.
1378 J. Org. Chem., Vol. 56,No. 4, 1991 Koerner and Rickborn
a a with methyl vinyl ketone (MVK) under basic conditions.
Although no cycloadducts were observed by Krohn and
Miehe, we suggest that cycloadduct 26 may be an inter-
mediate in this process, and note that the factors which
C"2 n r;n make the conversion of dithranol (1,8-dihydroxy-9-
22 23
anthrone) cycloadducts to Michael adducts2 more facile
than the corresponding anthrone reactions may contribute
observed. The methyl substituent thus prevents reaction, to this process. It is interesting to note that the minor
reminiscent of the failure of ethyl crotonate to enter into product (13%) formed from the reagents shown in eq 14
cycloaddition with anthrone under these conditions.2 Of arises from the opposite regioisomer, i.e. the formal Mi-
the 10-position substituents examined, only hydroxyl and chael adduct is bonded at C-9 rather than C-10; this
amino groups appear to be well-tolerated in the cyclo- product may arise from the alternative cycloadduct re-
addition process under the conditions which we have ex- gioisomer. Formation of both materials may simply reflect
plored to date. modest regioselectivity in the Diels-Alder step.
Double Retro-Aldol Reaction. The cycloadducts of 0
anthrone with several different dienophiles are efficiently
converted to the ring-opened Michael adducts by either OH 0 OH 1.Na&04

prolonged exposure to the same conditions used for cy-

cloadduct formation, or by the use of a stronger base in
protic solvent.2 Similar reaction of the anthracenediol
derivatives described above would lead to C-10 hydroxy 0 OH
analogues which might be susceptible to further cleavage. 25 26
Indeed, when cycloadduct 4 was subjected to conditions 9" 9 OH
(MeOH/THF, isopropylamine) which gave facile conver-
sion of the anthrone adduct 21 to its the ring-opened
Michael adduct isomer (analogous to 24, but lacking the
(2-10 hydroxyl), the only products observed were anthra-
quinone and N-methylsuccinimide. This reaction almost 27 (39%)
certainly occurs in stepwise fashion, with protonation of
the intermediate carbanion (imide enolate) formed by in- Although the different conditions employed by Krohn
itial bond cleavage occurring prior 'to scission of the second and Miehe preclude more direct comparisons, we have
carbon-carbon bond. The overall reaction from 4 is for- been able to isolate cycloadducts from 25 with both NMM
mally a double retro-aldol reaction. and methyl acrylate. We would expect that doubly acti-
vated dienophiles such as those employed in the current
study would lead to cycloaddition followed by the double
retro-aldol process, an overall formal redox dispropor-
tionation, i.e., reduction of the dienophile with concomitant
oxidation of the anthracenediol to the quinone. The major
difference between singly and doubly activated dienophiles
is the presence of the adjacent electron-withdrawinggroup
in the (intermediate) Michael adduct from vicinal doubly
activated substrate, which allows cleavage of the second
I C-C bond to occur via formation of a relatively stable ester
Me
24 enolate or similar intermediate.
Conclusions. Cycloadducts are formed between an-
0 thracenediols (and related aminoanthracenols)and reactive
dienophiles under mild base conditions. Stronger base/
protic medium can lead to a formal redox disproportion-
ation, which is believed to occur mechanistically as a
double retro-aldol reaction. The fundamental chemistry
The net reaction obtained by combining eqs 2 and 13 appears to be analogous to that described previously for
is the reduction of NMM to N-methylsuccinimide, al- anthrones? in which base-catalyzed Diels-Alder reactions
though this occurs by a circuitous route with the hydrogen were shown to occur more rapidly than Michael additions.
introduced from a proton source. This feature was dem- Cycloadditions may be carried out by either a one-step (in
onstrated by carrying out the reaction in MeOD, which led situ dienophile) or a two-step experimental procedure, with
to the formation of N-methy1succinimide-2,3-d2. Since advantages to both. As in,the anthrone reactions, the
related "reduction" of the dienophile double bond might ultimate product can depend strongly on apparently minor
account for the reduced yields of cycloadduct in e.g. the changes in reaction conditions.
dimethyl fumarate reaction (eq 5), a control experiment
was carried out which demonstrated that the cycloadduct Experimental Section
7 is stable under the conditions used for its formation. Reductions were carried out under an atmosphere of H2
However, it also can be cleaved in analogy to eq 13 under maintained by a balloon, which was attached by a syringe tip
the more forcing conditions of isopropylamine in methanol through a septum. The catalyst was 10% Pd/C from two different
solvent. lots. Unless otherwise noted, 'H spectra were recorded a t 500
MHz, and 13C NMR spectra a t 75 MHz, both with CDC13solu-
Krohn and Mieheh have reported that anthraquinone tions. MS, MS (CI) (chemically induced, methane), and MS (DCI)
and analogues such as 1,8-dihydroxyanthracene-9,1O-dione (desorption chemical ionization) data were obtained by Dr. Hugh
(25) will react with olefins activated by a single electron- Webb. Melting points were taken in open capillary tubes and
withdrawing group to give Michael adducts. For example, are uncorrected. Combustion analyses were performed by Desert
27 is formed as the major product (39%) when 25 is treated Analytics, Tucson, AZ, and by Galbraith Laboratories, Knoxville,
Reduction-Cycloaddition Reactions of Anthraquinones J . Org. Chem., Vol. 56, No. 4, 1991 1379
TN. The pyridine was Fisher Scientific reagent grade, used as through solution, and this was taken as an indication of com-
received. Starting materials were either commercial reagents or pletion. Suction filtration and evaporation gave crude product
prepared as described earlier.* which exhibited peaks attributed to 8a,b in a ratio of ca.65/35.
4,9[ l'~]-Benzeno-3a,4,9,9a-tetrahydro-4,9-dihydroxy-2- Recrystallizationfrom methanol gave 77 mg (65%) of the major
met hy 1-1H-benz[flisoindole- 1f (2H)-dione (4). A suspension isomer in essentially pure form as a colorless solid, mp 237-240
of 612 mg (2.94 mmol) of anthraquinone, 330 mg (2.97 mmol) of "C dec; 'H NMR 6 2.44 (s,3 H), 3.32 (8, 2 H), 4.44 (br s, 2 H),
NMM, and 25 mg of 10% Pd/C in 25 mL of pyridine was vig- 7.30 (m, 2 H), 7.47 (m, 2 H), 7.69 (m, 2 H), 7.82 (m, 2 H),and
orously stirred while Hz gas was introduced from a balloon (ca. 7.87 ppm (s,2 H); 13CNMR (acetone-ds) 6 52.8,120.6, 120.9,1!27.0,
0.5 L) via a syringe tip with the needle introduced through a 127.4, 128.8, 133.0, and 176.8 ppm; IR (KBr) 3400 (br, s), 3022
septum and extending below the surface of the liquid. Flow was (w), 1655 (br, s), 1430, (m), 1275 (m), 1125 (m), 872 cm-'; MS 371
maintained by allowing gas to exit via a narrow-gauge needle. (P, 21,261 (19), 260 (loo), 231 (141,202 (28); MS (CI) calcd for
After 0.5 h, the mixture wagsuction filtered and the filtrate was CSHl8NO4 (P + H) 372.1257, found 372.1246. Identification of
evaporated under vacuum. The residue was washed with cold the minor isomer rests on the NMR of the crude mixture, which
methanol to give 893 mg (95%) of a colorless solid mp 258-260.5 exhibited peaks that closely paralleled those of the major product.
"C;'H NMR 6 2.51 (s, 3 H), 3.24 (8, 2 H), 4.35 (8, 2 H), 7.25 (m, 6,13-Dihydro-6,13-dihydroxy-15H,19H-6,13
4 H), 7.45 (m, 2 H), and 7.67 (m, 2 H) ppm; 13CNMR 6 24.4, 51.7, pentacenel6,18(17H)-dione(9). (a) In Situ NMM. A mixture
120.2,120.5,126.8,127.3, 138.4,140.0,and 177.4 ppm; IR (KBr) of 101 mg (0.33 m o l ) of 6,13-pentacenedione,44 mg (0.40 "01)
3420 (br, s), 3080 (w), 3000 (w), 2945 (w), 1680 (vs), 1430 (m), 1375 of NMM, and 15 mg of 10% Pd/C in 15 mL of pyridine was
(m), 1200 (m), 1130 (m), 1025 (m) cm-'; MS (CI) 322 (P + H, 5), hydrogenated in the usual manner for 0.5 h and then stirred
211 (60),210 (58), 195 (261,112 (100); calcd for C19H16N04 (P + overnight at room temperature. Examinationof the crude residue
H) 322.1039, found 322.1059. after filtration and evaporation of volatiles showed that 9 was the
trans -1 1,12-Dicyano-9,10-ethano-9,lO-dihydro-9,lO-di- major product, along with N-methylsuccinimide. Further vacuum
hydroxyanthracene (5). Similar hydrogenationwas carried out treatment (1Torr) removed the latter material to give 58 mg
over a period of 2 h of a suspension of anthraquinone (280 mg, (42%) of essentially pure 9, as a colorless solid, mp 269-270 "C.
1.35 mmol), fumaronitrile (106 mg, 1.36 mmol), and 15 mg of 10% The NMR of this material was identical with that described below.
Pd/C in 5.0 mL of pyridine. Suction filtration followed by (b) Stepwise Procedure. Reduction was carried out by
evaporation of volatiles gave a residue which was recrystallized bubbling Hz through a mixture of 9,13-pentacenedione (90 mg,
from CHC13 to give 361 mg (94%) of tan cubic crystals: mp 0.29 mmol) and 10 mg of Pd/C in 12 mL of pyridine for 0.75 h.
180-184 "C dec; 'H NMR (acetone-d,) 6 3.47 (s, 2 H), 7.37 (m, Methanol (2 mL) was added, and the mixture was suction filtered
4 H), 7.68 (m, 2 H), and 7.75 (m, 2 H) ppm; '% NMR (acetone-ds) to remove the catalyst. Vacuum evaporation of the volatiles gave
6 43.7, 75.6, 118.4, 121.0, 122.3, 128.1, 140.0, and 141.2 ppm; IR 87 mg (9670) of a pale yellow solid, mp 330 OC, identified as
(KBr) 3350 (br, s), 3060 (w), 2924 (w), 2250 (m), 1445 (w), 1360 13-hydroxypentacene-9(13H)-one(10) on the basis of its spectral
(vs), 1280 (m), 1255 (m), 1021 (s) cm-'; MS (DCI) 210 (P - 78, properties: 'H NMR 6 2.45 (d, 1 H, J = 8.5 Hz), 6.15 (d, 1 H,
loo), 181 (291,152 (1%81 (49). Anal. Calcd for C1&12N202:C, J = 8.5 Hz), 7.58 (t, 2 H, J = 8 Hz), 7.65 (t, 2 H, J = 8 Hz), 7.97
74.99; H, 4.20. Found: C, 74.72; H, 4.18. (d, 2 H, J = 8 Hz), 8.06 (d, 2 H, J = 8 Hz), 8.36 (s,2 H), and 8.90
cis - 11,12-Dicyano-9,10-ethano-9,lO-dihydro-9,10-di- (s,2 H); IR (KBr) 3420,(br,m), 3060 (w), 1652 (m), 1605 (s), 1590
hydroxyanthracene (6). Analogous treatment of a mixture of (s), 1445 (s), 1390 (m), 1370 (s), 1285 (vs), and 1191 (vs) cm-'.
220 mg (1.06 mmol) of anthraquinone, 81 mg (1.04 mmol) of A solution of 72 mg (0.114 m o l ) of 10 and 13 mg (0.117 mmol)
maleonitrile, and 10 mg of 10% Pd/C in 4.0 mL of pyridine for of NMM in 5 mL of THF was treated with 4 drops of triethyl-
2 h gave after workup and recrystallization from CHC13294 mg amine and allowed to stand overnight. The solvent was vacuum
(98%) of a colorless solid: mp 214-217 "C dec; 'H NMR (ace- evaporated to give 41 mg (84%) of 9 as a colorless solid: mp
toned6) 6 3.77 ( 8 , 2 H), 7.32 (m, 2 H), 7.37 (m, 2 H), 7.65 (m, 2 269-271 "C; 'H NMR 6 2.49 (s,3 H), 3.42 ( ~ , H), 2 4.58 (s,2 H),
H), and 7.76 (m, 2 H) ppm; 13C NMR (acetone-d6)6 42.8,75.5, 7.47 (m, 4 H), 7.83 (m, 2 H), 7.88 (m, 2 H), 7.92 (s,2 H), and 8.11
117.5, 120.8, 122.1,127.8,128.0, 139.9, and 141.8 ppm; IR (KBr) ppm (s, 2 H); 13C NMR (THF-d8) 6 30.76, 52.86, 76.05, 119.82,
3400 (br, vs), 3075 (w), 2920 (m), 2240 (m), 1455 (m), 1230 (m) 119.87,120.69,126.64,128.73,133.36,133.43,138.13,140.78,176.57;
cm-'; MS (DCI) 210 (P - 78, 461, 180 (15), 152 ( E ) , 79 (100). IR (KBr) 3400 (br, 4,2070 (w), 2810 (w), 1695 (vs), 1445 (s), 1390
Several attempts to obtain combustion analyses resulted in low (s), 1285 (s), 1135 (s) cm-'; MS 421 (P, l), 311 (23), 310 (P- 111,
carbon percentages (ca. 2%) for reasons that are not understood; loo), 308 (38), 252 (22), 155 (12), 126 (14); MS (CI) calcd for
all MS efforts failed to give a parent ion. CZ7HzoNO4 (P + H) 422.1430, found 422.1412.
Dimethyl 9,lO-Et hano-9,1O-dihydro-g,lO-dihydroxy- Anthraquinone Monoxime (11). This material was prepared
anthracene-1 1,12-trans-dicarboxylate (7). A mixture of 585 by two different methods in essentially quantitative yield. The
mg (2.81 m o l ) of anthraquinone, 410 mg (2.85 mmol) of dimethy literature p r ~ e d u r e utilizes
'~ anthraquinone and hydroxylamine.
fumarate, and 20 mg of 10% Pd/C in 10 mL of pyridine was The alternative described here involves nitrosation of anthrone,
hydrogenated for 3 h. The usual workup gave a residue which following a general procedure outlined by Touster for active
by 'H NMR exhibited singlets at 3.70 and 2.64 ppm due to di- methylene compounds.'* To a solution of 1.00 g (5.15 mmol) of
methyl succinate in addition to absorptions for the product de- 9-anthrone in 50 mL of DMF was added 10 mL of water and 20
scribed below. Recrystallization from methanol gave 597 mg mL of concentrated HCl. The ice bath cooled, and the stirred
(60%) of 7 as a colorless solid: mp 140-141.5 "C; 'H NMR 6 3.59 mixture was treated in several portions with 1.11 g (13 mmol) of
(s,6 H),5.4 (br s, 2 H), 7.22 (m,4 H),7.47 (m, 2 H),and 7.65 ppm NaNOZ,and stirring was continued with the bath removed for
(m, 2 H); 13CNMR B 52.39,55.03, 76.35, 120.1, 121.5, 126.8,141.4, 12 h. The colorless precipitate was suction filtered and recrys-
143.4, and 172.6 ppm; IR (KBr) 3360 (br, vs), 3080 (m), 3020 (m), tallized from ethanol to give 1.11g (98%) of 11: mp 224.5-225
1740 ( 4 , 1 4 8 2 (s), 1345 (s), 1255 (vs)cm-'; MS (CI) 210 (P - 144, "C (lit." mp 224 "C); 'H NMR 6 7.58 (t, 1 H, J - 8 Hz), 7.66 (t,
61), 194 (30), 180 (291,152 (27), 145 (761,113 (loo), 85 (37). Anal. 2 H, J = 8 Hz), 7.74 (t, 1 H, J = 8 Hz), 8.20 (d, 1 H, J = 8 Hz),
Calcd for C&I1806: C, 67.79; H, 5.12. Found: C, 67.54; H, 5.04. 8.28 (d, 1H, J = 8 Hz), 8.42 (d, 1 H, J = 8 Hz), 8.90 (br s, 1H),
This product (7) was also prepared, in 80% yield, by stepwise and 9.02 ppm (d, 1H, J = 8 Hz); 13C NMR (a~et0ne-d~) 6 125.3,
reduction (1h), replacement of the Hz atmosphere by Ar, followed 127.1, 127.9, 128.0, 130.1, 131.2, 131.9, 132.2,134.0, 134.1, 137.2,
by addition of dimethyl fumarate. 143.2, and 183.5 ppm; MS 223 (P, 100), 207 (25), 193 (28), 179
endo- and exo-3a,4,9,9a-Tetrahydro-4,9[2',3']-benzeno- (23), 165 (26), 121 (19),76 (26); IR (KBr) 3280 (br, s), 3075 (w),
4,9-dihydroxy-2-methylnaphthaleno-lH-benz[ flisoindole- 2980 (w), 1650 (vs), 1600 (vs), 1444 (s), 1360 (vs), and 1030 (s)
1,3(2H)-dione (8a,b). A suspension of 90 mg (0.35 mmol) of cm-'.
naphthacenequinone, 60 mg (0.54 mmol) of NMM, and 15 mg Anthraquinone Monimine (12). The reduction of 11 was
of 10% Pd/C in 15 mL of pyridine was hydrogenated in the usual carried out as described by Costa et al." A solution of 11 (2.76
way. The heterogeneous yellow mixture developed a dark bur-
gundy red color at the start of the reaction, and the color faded (13) Haq, M. A.; Ray, J. N.; Tuffail-Malkana,M. J . Chem. Soc. 1934,
to pale orange if the reduction was interrupted. After 0.5 h the 1326.
red color no longer appeared when additional H, was passed (14) Schunck, E.; Marchlewski, L. Chem. Ber. 1894,27, 2125.
1380 J. Org. Chem., Vol. 56, No. 4, 1991 Koerner and Rickborn

g, 12.4 mmol) in 130 mL of T H F was cooled in an ice bath, and 122.4,127.7,127.8,140.5, and 142.0ppm; IR (KBr) 3340 (m), 3280
2.54 g (67.1mmol) of LiAlH, was added. The mixture was then (m), 3150 (br, s), 2940 (m), 2250 (m), 1600 (m), 1460 (s), 1225 (s),
refluxed for 5 h, cooled, and quenched by careful addition of water +
1005 ( 8 ) cm-'; MS (CI) 288 (P H, O.l),209 (loo),193 (12),180
(11mL). After stirring overnight, the mixture was filtered and (26),130 (9),76 (67);calcd for C18H14N30(P + H) 288.1177,found
the filtrate with THF washes was evaporated in vacuo to give 2.49 288.1157.
g (97%) of tan solid, mp 225.5-226 O C (lki6 mp 224-225 "C). Dimethyl 9-Amino-9,lO-ethano-9,10-dihydro-10-hydroxy-
Material formed in this way reproducibly exhibited NMR spectra anthracene-traas-ll,l2-dicarboxylate (16). A mixture of the
expected for 12 not undergoing rapid (NMR time frame) syn/anti monimine 12 (322mg, 1.55 mmol), dimethyl fumarate (225mg,
interconversion (see Figure 1A): 'H NMR S 7.71 (m, 2 H), 7.79 1.56 mmol), and 20 mg of 10% Pd/C in 10 mL of pyridine was
(m, 2 H), 7.97 (d, 1 H, J = 8 Hz), 8.33 (d, 1 H, J = 8 Hz), 8.40 vigorously stirred while H2gas was bubbled through the solution.
(d, 1 H, J = 8 Hz), 8.64 (d, 1 H, J = 8 Hz), and 11.0 ppm (s, 1 After 1.5 h, the mixture was fdtered and then vacuum evaporated.
H, exchanges with D,O);13C NMR 6 123.3,126.0,126.9,127.9, Analysis of the crude product by 'HSJMR showed the presence
131.6,131.7,133.5,133.7,163.1,and 183.3 ppm; MS(C1): 236 (P of cycloadduct 16 (30%), starting material 12 (70%),and dimethyl
+
+ 29,ll),209 (23),208 (P H, 100),207 (33),179 (9);IR (KBr) succinate (70%).
3055 (w), 1660 (m), 1582 (s), 1361 (s), 1310 (vs), 1184 (s) cm-'. The yield was raised to 60% by repeating the procedure with
The effect of added trifluoroacetic acid on the 'H NMR is shown 5 equiv of dimethyl fumarate, and 60% was also obtained by the
in Figure 1. stepwise reduction-cycloaddition as described for the analogous
Recrystallization of this product from acetone gave material quinone. Compound 16 was isolated by silica gel chromatography
which exhibited (broadened) spectra such as that shown in Figure (90/10CH2C12/acetone)as a colorless solid mp 116118 OC dec;
lB, as did material prepared by catalytic reduction. Thus hy- 'H NMR 6 2.60 (br s, 2 H), 3.26 (d,1 H, J = 5 Hz), 3.37 (d, 1 H,
drogenation in the usual manner of a solution of 115 mg (0.52 J = 5 Hz), 3.57(8, 3 H), 3.58 (s, 3 H), 7.23 (m, 4 H), 7.38 (m, 1
mmol) of 11 and 10 mg of 10% Pd/C in 7 mL of pyridine for 1 H), 7.49 (m, 1 H), 7.53 (m, 1 H), and 7.67 ppm (m, 1 H); '9 NMR
h gave, after vacuum evaporation, 96 mg (90%)of an off-white 6 119.5,120.0,120.5,126.4,126.5,140.2,141.4,141.7,142.1,171.7,
solid: mp 223-224 OC; 'H NMR 6 7.71 (t, 2 H, J = 8 Hz), 7.78 and 172.9 ppm; IR (KBr) 3480 (br), 3390 (m), 3210 (m), 3080 (m),
(t, 2 H, J = 8 Hz), 8.0 (very broad s, 1 H), 8.36 (broadened d, 2 2970 (m), 1735 (s), 1595 (m), 1270 (m) cm-'; MS (CI) 354 (P +
H, J = 8 Hz), 8.5 (very broad s, 1 H), and 11.0 ppm (broad s, 1 H, 2),210 (58),209 (loo),208 (44),180 (14),145 (57),113 (33),
H); '9NMR 6 127.3,127.4,131.6,131.7,163.0,and 183.3 ppm. 85 (13);calcd for CmHzoNO5(P + H) 354.1355,found 354.1348.
The MS (CI) and IR spectra were essentially identical with those Pentacenequinone Monoimine (17). The monoxime of
described above. Monimine 12 from both procedures gave the 9,13-pentacenedione was prepared by appropriate modification
same cycloadduct with NMM. of the literature method for the anthraquinone analogue." A
4-Amino-4,9[ lf,2']-benzeno-3a,4,9,9a-tetrahydro-9- mixture of the quinonei(260gm,0.844 mmol) and 1.7 g (30mmol)
hydroxy-2-methyl- 1H-benz[f]isoindole-l,3(2H)-dione (13). of KOH in 170 mL of ethanol was treated with 610 mg (8,77"01)
A mixture of 70 mg (0.314mmol) of 12, 36 mg (0.324mmol) of of hydroxylaminehydrochloride dissolved in 0.5 mL of water. The
NMM, and 10 mg of 10% Pd/C in 4.0 mL of pyridine was vig- yellow suspension was refluxed for 20 h, at which time it was dark
orously stirred while the contents of a H2-filled balloon was red and appeared to be homogeneous. After cooling to room
bubbled into the solution by means of a syringe needle tip im- temperature, the mixture was treated with acetic acid, and the
mersed below the surface. A slow flow was maintained by allowing resulting yellow suspension was reduced to about half volume on
excess gas pressure to exit via a narrow gauge needle. After 2 h a hot plate, after which water was added. Filtration followed by
the mixture was filtered, and the filtrate was evaporated to give recrystallization from ethanol/water gave 222 mg (82%) of the
107 mg (100%)of an off-white solid mp 107-200 OC dec; 'H NMR oxime as a yellow solid mp 269-271 OC; 'H NMR (acetone-de)
6 2.50 (8, 3 H),2.72 (s, 2 H), 3.10 (d, 1 H, J = 9 Hz), 3.22 (d, 1 6 7.70(m, 4 H), 8.10 (d, 1 H, J = 8 Hz), 8.17 (d, I H, J = 8 Hz),
H, J = 9 Hz), 4.50 (s, 1 H), 7.23 (m, 1 H), 7.24(m, 2 H), 7.29 (m, 8.20 (d, 1H, J = 8 Hz), 8.24 (d, 1 H, J = 8 Hz), 8.79(8, 1 H), 8.88
1 H), 7.39 (m, 1 H), 7.48 (m, 1 H), 7.55 (m, 1 H), and 7.70ppm (5, 1 H), 8.98 (s, 1 H), 9.82 (s, 1 H), and 12.0 ppm (br s, 1 H); 13C
(m, 1 H); 13CNMR (75MHz) 6 25.4,52.0,52.4,59.5,64.4,120.3, NMR (THF-de) 6 124.9,127.8,128.8,129.0,129.3,129.5,130.0,
120.5,120.7,126.9,127.1,127.2,139.4,139.5, 141.0,and 141.3 ppm; 130.3,130.4,133.1,133.8,133.9,136.3,136.6,and 143.8 ppm; IR
IR (KBr) 3320 (m), 3295 (m), 3200 (br, s), 3060 (m), 3005 (m), (KBr) 3290 (br), 1640 (m), 1360 (m), 1295 (m), 1195 (m) cm-';
1680 (vs), 1440 (s), 1290 (s), 1140 (s), 1010 cm-'; MS (CI) 321 (P MS (CI) 309 (31),308 (P - 15,100),307 (85), 295 (29),294 (31),
+ H, 6),231 (9),208 (loo),180 (18),112 (59);calcd for C19H1,N203 293 (13).
(P + H) 321.1223,found 321.1231. Reduction of the oxime (35mg, 0.108 mmol) was carried out
9-Amino-trans - 11,12-dicyano-9,10-ethano-9,10-dihydro- in T H F by adding LiAlH, (38mg, 1.0 mmol) at 0 OC, followed
10-hydroxyanthracene (14). Similar treatment of a mixture by 5 h at reflux. Water (5 drops) was added to quench the excess
of 12 (87mg, 0.39 mmol), fumaronitrile (30mg, 0.39mmol), and hydride, and the mixture was filtered to remove inorganic material.
10 mg of 10% Pd/C in 5.0 mL of pyridine for 40 min gave, after Evaporation and recrystallization of the residue from ethanol gave
recrystallization from CHC13,105 mg (90%)of an off-white solid 33 mg (99%) of 17 as pale yellow needles: mp 322 OC dec; IR
mp 176-178.5 OC dec; 'H NMR (acetone-d,) 6 3.07 (br s, 3 H), (KBr) 3060 (w), 1661 (s), 1640 (m), 1570 (s), 1445 (s), 1390 (s),
3.44 (d, 1 H, J = 5 Hz), 3.47 (d, 1 H, J = 5 Hz), 7.35 (m, 4 H), 1308 (vs), 1190 (s), 1175 (s), and 968 (s) cm-'; MS 308 (26),307
7.69 (m, 3 H), and 7.75 ppm (m, 1 H); 13C NMR (acetone&) 6 (P, loo),277 (14),252 (lo),126 (19),125 (15);calcd for CZH13NO
43.9,45.1,59.9, 75.7,118.9,121.0,121.7,122.1,122.5,127.7,127.8, 307.0986,found 307.0992. Like the anthraqu:inone monoimine
127.9,128.0,140.5,140.7,141.6,and 142.2 ppm; IR (KBr) 3500 discussed in greater detail in the text, 17 exhibited NMR char-
(m), 3400 (m), 3250 (br, s), 3082 (m), 2945 (m), 2250 (m), 1455 acteristics which depended upon the sample history, again at-
(m), 1300 (s), 750 (4cm-'; MS (CI) calcd for C18H14N30(P + H) tributed to syn/anti interconversion rates: 'H NMR (slow in-
288.1117,found 288.1127. terconversion) 6 7.67 (m, 4 H),8.09 (m, 4 H), 8.47,(8, 1 H),8.96
9-Amino-cis -11,12-dicyano-9,10-ethano-9,lO-dihydro-lO- (s, 1 H), 9.00 (8, 1 H), 9.20 (s, 1 H), and 10.95 ppm (br s, 1 H);
hydroxyanthracene (15). A solution of 100 mg (0.480mmol) (rapid interconversion) 6 7.67 (m, 4 H), 8.09 (m, 4 H), 8.85 (br
of 12,40 mg of maleonitrile (0.51mmol), and 10 mg of 10% Pd/C s, 2 H), 8.98 (s, 2 H).
in 5.0 mL of pyridine was hydrogenated for 40 min. The solvent Cycloadduct of 17 + NMM (18). A mixture of 17 (24mg,
was vacuum evaporated, and the residue was recrystallized from 0.078mmol), NMM (20mg, 0.18 mmol), and 5 mg of Pd/C in
CHCl, to give 127 mg (92%) of tan solid: mp 164-166 "C dec; 6 mL of pyridine was subjected to H2bubbling for 0.5 h and then
'H NMR (acetone-d,) 6 3.05 (br s, 2 H), 3.64 (d, 1 H, J = 10.5 allowed to stand overnight. The crude residue from the initial
Hz), 3.75 (d, 1 H, J = 10.5 Hz), 6.50 (s, 1 H), 7.31 (m, 2 H), 7.37 filtration and evaporation was taken up in T H F and filtered
(m, 2 H), 7.64 (m, 2 H), 7.71 (m, 1 H), and 7.77ppm (m, 1 H); through a short plug of neutral alumina (1g), and then volatiles
'3c NMR (acetone-ds)6 43.1,44.5,117.8, 118.0,120.8,121.6,122.2, including N-methylsuccinimide were removed by full pump
vacuum with mild heating. Essentially pure 18 was obtained (29
mg, 87%) as a colorless solid: mp 259.5-260 "C; 'H NMR 6 2.47
(15) Rigaudy, J.; Cauquis, C.; Izoret, G.; Baranne-Lafont, J. Bull. Soc. (s, 3 H), 3.00 (br s, 2 H), 3.26 (d, 1 H, J = 9 Hz), 3.37 (d, 1 H,
Chim. Fr. 1961, 1842. J = 9 Hz),4.73 (8, 1 H), 7.47 (m, 4 H), 7.84 (m, 5 H), 7.95 (s, 1
 J . Org. Chem. 1991,56,1381-1386 1381

H), 7.99 (a, 1 H), and 8.13 ppm (8, 1 H); 13C NMR 24.51, 51.60, A mixture of 22 (212 mg, 1.02 mmol), NMM (126 mg, 1.14
52.12,59.10, 119.26, 119.33, 119.62, 120.00,120.17,120.33, 126.42, mmol), and 10 mg of 10% Pd/C in 10 mL of pyridine was hy-
128.04, 132.26, 136.83, 138.34, 138.58, 176.43, 177.62; IR (KBr) drogenated for 40 min, stirred an additional 30 min, and then
3400 (br, s), 3380 (m), 3210 (m), 3047 (m), 2942 (w), 1680 (vs), worked up in the usual manner. Analysis of the crude residue
1440 (vs), 1285 (vs), 1140 (s) cm-'; MS (CI) 421 (P, 11, 323 (51, by NMR showed no absorptions anticipated for cycloadduct;
295 (16), 294 (131, 114 (121,112 (71), 111 (12), 55 (12), '43 (100); instead, major amounts of N-methylsuccinimide and 10-
MS (CI) calcd for CnHz1N2O3(P+ H) 421.1530, found 421.1541. methyl-9-anthrone (23) were formed. The latter was identical
Reaction of 20. The procedure of Regitz16was used to prepare with 23 that had been independently prepared by methylation
10-diaza-9-anthrone (20) from 9-anthrone and p-toluenesulfonyl of anthrone and also by reduction of 2 2 10-methyl-9-anthrone
azide" in 90% yield. Bright red 20 (mp >300 OC) has 'H NMR (23) has mp 64-66 OC (lit.19 mp 64.5-66.5 "C); 'H NMR 6 1.59
6 7.33 (d, 2 H, J = 8 Hz), 7.41 (t, 2 H, J = 8 Hz), 7.70 (t, 2 H, (d, 3 H, J = 7.5 Hz), 4.30 (4, 1 H, J = 7.5 Hz), 7.44 (t, 2 H, J =
J = 8 Hz), and 8.54 ppm (d, 2 H, J = 8 Hz). 7.5 Hz), 7.50 (d, 2 H, J = 7.5 Hz), 7.62 (t, 2 H, J = 7.5 Hz), and
A solution of 20 (126 mg, 0.594 mmol) and NMM (68 mg, 0.613 8.31 ppm (d, 2 H, J = 7.5 Hz).
mmol) in 7 mL of pyridine was reduced in the usual manner with Double Retro-Aldol Reaction of 4 in MeOD. A solution of
15 mg of 10% Pd/C catalyst. After 1h of vigorous stirring with 321 mg (1.0 mmol) of cycloadduct 4 in a solvent consisting of 7
Hz, and 5 h of additional stirring, the mixture was worked up in mL each of MeOD and THF was treated with 1 mL of iso-
the usual manner. The deazotized cycloadduct 21 was estimated propylamine. After 24 h at room temperature, the solution was
by NMR to have been formed in 45% yield, along with N- separated from the yellow needles of anthrquinone which had
methylsuccinimide and 9-anthrone. An essentially quantitative precipitated, and the solvent was removed under vacuum, with
yield of 21 was formed when this procedure was carried out by care to retain the relatively volatile N-methylsuccinimide. The
stepwise reduction-cycloaddition. The product 21 was identical 'H NMR spectrum of the residue showed a singlet at 3.00 ppm
with material prepared by direct reaction of anthrone with NMM, (N-Me; relative area = 3) and a broadened triplet-like absorption
as recently described.2 at 2.72 ppm (relative area = 2.1, compared to 4.0 for the un-
Reaction of 22. The procedure of Clar was used to prepare deuterated N-methylsuccinimidemethylene absorption),signifying
10-methylene-9-anthrone1* (22) from 9-anthrone and aqueous formation of the dideuterio compound with high efficiency.
formaldehyde. Compound 22 was obtained in 91% yield as
colorless crystals: mp 146.5-147 OC (lit.19 mp 145-147.5 "C); 'H Acknowledgment. Support of this work by the donors
NMR 6 6.34 (s, 2 H), 7.54 (t, 2 H, J = 7 Hz), 7.65 (t, 2 H, J = of t h e Petroleum Research Fund, administered by the
8 Hz), 8.01 (d, 2 H, J = 8 Hz), and 8.35 ppm (d, 2 H, J = 7 Hz). American Chemical Society, is gratefully acknowledged.

(16) Regitz, M. Chem. Ber. 1964,97, 2742.

Supplementary Material Available: 'H NMR spectra of
(17) Doering, W. v. E.; DePuy, C. H. J. Am. Chem. SOC.1953,75,5955. +
4,6,8a 8b, 10, and 13-18 and 13C NMR spectra of 4,6,8a +
(18) Clar, E. Chem. Ber. 1936,69, 1687. 8b, 13-16, and 18 (67 pages). Ordering information is given on
(19) Heymann, H.; Trowbridge, L. J . Am. Chem. SOC.1950, 72, 84. any current masthead page.

Pentaalkylstiboranes. 1. Synthesis of Homobenzylic Alcohols, Homoallylic

Alcohols, Ethyl 5-Aryl-5-hydroxypent-2-enoates, and B-Hydroxypropionic
Acid Derivatives via Pentaalkylstiboranest
Yao-Zeng Huang* and Yi Liao
Laboratories of Organometallic Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of
Sciences, 345 Lingling Lu, Shanghai 200032, China
Received March 28, 1990

Although pentarilkylstiboranes have long been known, their applications in organic synthesis have not been
exploited. It has been found that quaternary stibonium salts (n-Bu3SbCHzE]+X-(E = Ph, CH=CH2, CH-
CHCO,Et, COZEt, CN; X = Br, I, BPh4) on treatment with RLi (R = n-Bu, t-Bu, Ph) afford pentaalkylstiboranes,
n-Bu3Sb(R)CHzE,which react with aromatic aldehydes to give, after subsequent hydrolysis, homobenzylic alcohols,
homoallylic alcohols, ethyl 5-aryl-5-hydroxypent-2-enoates, ethyl @-aryl-0-hydroxypropionates,and @-aryl-@-
hydroxypropionitriles, respectively, in good to excellent yields. The reaction is chemoselective for aldehydes.

Introduction stiborane, methyltetraphenylstiborane, and unreacted

Few reports' concerning t h e application of organo- benzophenone. On the other hand, Wittig and Laib' re-
antimony compounds in organic synthesis have appeared ported that Me2Sb(CH2Ph)2Brreacted with PhLi to yield
in the literature. Henry and Wittig2 claimed that tri- Me2SbCH(CH2Ph)Ph,a product hypothesized t o result
phenylstibonium methylide, prepared from methyltri- from the rearrangement of an antimony ylide. The only
phenylstibonium iodide and phenyllithium, reacted with successful Wittig-type process reported for an antimony
benzophenone to form acetaldehyde. However, Doleshall ylide was t h e reaction of triphenylstibonium tetra-
et aL3later reported a quite different result: that reaction phenylcyclopentadienylide,formed from triphenylstibine
of methyltriphenylstibonium iodide or tetraphenylborate
with phenyllithium followed by introduction of benzo- (1) Freedman, L. D.; Doak, G. 0. J. Orgonomet. Chem. 1988,351,26
phenone into the reaction mixtures gave a pentaorganyl- and the references therein.
( 2 ) Henry, M. C.; Wittig, G. J. Am. Chem. SOC.1960, 82, 563.
(3) Doleshall, G.; Nesmeyanov, N. A,; Reutov, 0. A. J. Orgonomet.
'This paper is the 85th report of the synthetic application of Chem. 1971, 30, 369.
elementoorganic compounds of group 15 and 16. (4) Wittig, G.; Laib, H. Liebigs Ann. Chem. 1953, 580, 57.

0022-3263/91/1956-1381$02.50/00 1991 American Chemical Society