New approach to biomimetic transamination using bifunctional [1,3]-proton transfer catalysis in thioxanthenyl dioxide imines

Article abstract of DOI:10.1021/jo0106748

A pyridoxamine equivalent, 9-aminothioxanthene 10,10-dioxide, has been designed that is capable of affording transamination in good to excellent yields of natural as well as artificial amino acids. Amidines and guanidines in catalytic amounts were capable of performing [1,3]-proton transfer in the imines under mild conditions, whereas various simple amines failed. The use of chiral catalysts resulted in modest asymmetric induction (ee ≤ 45%). The electronic dependence in para-substituted phenyl glyoxylate imines, isotope effects, and computational studies support a stepwise, bifunctional mechanism for amidine and guanidine catalysts. Attempts toward an autocatalytic model system are described.

Full text of DOI:10.1021/jo0106748

J . Org. Chem. 2002, 67, 3585-3594
3585
New Ap p r oa ch to Biom im etic Tr a n sa m in a tion Usin g Bifu n ction a l
[1,3]-P r oton Tr a n sfer Ca ta lysis in Th ioxa n th en yl Dioxid e Im in es
Anders Hjelmencrantz and Ulf Berg*
Organic Chemistry 1, Department of Chemistry, Lund University, PO Box 124, S-221 00 Lund, Sweden
ulf.berg@orgk1.lu.se
Received J uly 5, 2001
A pyridoxamine equivalent, 9-aminothioxanthene 10,10-dioxide, has been designed that is capable
of affording transamination in good to excellent yields of natural as well as artificial amino acids.
Amidines and guanidines in catalytic amounts were capable of performing [1,3]-proton transfer in
the imines under mild conditions, whereas various simple amines failed. The use of chiral catalysts
resulted in modest asymmetric induction (ee e 45%). The electronic dependence in para-substituted
phenyl glyoxylate imines, isotope effects, and computational studies support a stepwise, bifunctional
mechanism for amidine and guanidine catalysts. Attempts toward an autocatalytic model system
are described.
Sch em e 1
In tr od u ction
Transamination constitutes an important process in
most biological systems as a pathway of metabolism of
amino acids/R-keto acids.1 The production of amino acids
is one of the most important reactions in native as well
as artificial systems as it is the foundation for enzymes/
proteins (native systems) and chiral building blocks
(artificial systems). It would thus be highly desirable to
construct a simple artificial analogue of the amino
transferase enzyme, making the transamination both
effective and stereocontrolled. Many efforts have been
made to seek efficient and simple systems capable of
achieving transamination. Efforts to mimic transamina-
tion used the cofactor, pyridoxal (vitamin B6, Scheme 1),
which is involved in all the different reaction pathways
performed by the amino transferase enzymes.
Other approaches involve the use of simple pyridine
derivatives to mimic the action of pyridoxal (commonly
referred to as an“electron sink”). The principal idea is to
create an electron-poor center, thereby stabilizing the
intermediate aza-allylic anion. A general problem with
pyridoxamine-based transamination is that ∆G is close
to 0, which may lead to undesirably low yields,2 but has
advantage when a catalytic cycle with high turnover is
sought. Other approaches make use of conjugation in the
product,- and in this way shift the equilibrium toward
the thermodynamically more stable product, in our case
the amino acid imine. The transamination reaction itself
consists of a [1,3]-proton-transfer reaction across an aza-
allyl anionic intermediate (Scheme 1).3 In a native
system, this is performed by the amino group of the lysine
residue in the enzyme active site.4 In artificial systems,
bases such as alkoxides and simple amines have been
used. The first rational investigation of the transamina-
tion reaction was performed by Cram et al.,5 who
investigated the suprafacial/antarafacial character of the
proton-transfer step, as well as the stereoselectivity of
the reaction, using N-(R-methylneopentylidene)-R-phe-
nylethylamine and KOtBu as base. Martell6 has also
made several studies concerning the mechanistic view-
point of transamination, elucidating the different mecha-
nistic details of pyridoxal-catalyzed reactions. Soloshonok
et al.7 have presented a number of papers on the
* To whom correspondence should be addressed. Phone: Int + 46
46 222 81 23. Fax: Int + 46 46 222 41 19.
(5) (a) Guthrie, R. D.; J aeger, D. A.; Meister, W.; Cram, D. J . J . Am.
Chem. Soc. 1971, 93, 5137-5153. (b) J aeger, D. A.; Cram, D. J . J . Am.
Chem. Soc. 1971, 93, 5153-5161. (c) J aeger, D. A.; Broadhurst, M.
D.; Cram, D. J . J . Am. Chem. Soc. 1979, 101, 717-732.
(1) Dugas, H. Bioorganic Chemistry, A Chemical Approach to
Enzyme Action; Springer: New York, 1996.
(2) Lehninger, A. L.; Nelson, D. L.; Cox, M. M. Principles of
Biochemistry, 2nd ed.; Worth Publishers: New York, 1993; pp 511-
514.
(6) (a) Martell, A. E. Acc. Chem. Res. 1989, 22, 115-124. (b) Martell,
A. E. Adv. Enzymol. 1982, 53, 163-199.
(7) (a) Soloshonok, V. A.; Kukhar, V. P. Tetrahedron 1996, 52, 6953-
6964. (b) Soloshonok, V. A.; Soloshonok, I. V.; Kukhar, V. P.; Svedas,
K. V. J . Org. Chem. 1998, 63, 1878-1884. (c) Soloshonok, V. A.; Ono,
T. J . Org. Chem. 1997, 62, 3030-3031. (d) Soloshonok, V. A.; Ono, T.
Tetrahedron 1996, 52, 14701-14712.
(3) (a) Roitenan, J . N.; Cram, D. J . J . Am. Chem. Soc. 1971, 90,
2225-2230. (b) Roitenan, J . N.; Cram, D. J . J . Am. Chem. Soc. 1971,
90, 2231-2241.
(4) Toney, M. D.; Kirsch, J . F. Biochemistry 1993, 32, 1471-1479.
10.1021/jo0106748 CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/01/2002

3586 J . Org. Chem., Vol. 67, No. 11, 2002
Hjelmencrantz and Berg
Sch em e 3
Sch em e 2
transamination of fluoro substituted ketimines, using
mild bases such as triethylamine (TEA) and diazabicyclo-
[3.3.0]undec-7-ene (DBU) and also attempts toward
asymmetric induction using alkoxides of ephedrines to
achieve asymmetric catalysis, resulting in an ee of 36%.
Cainelli et al.8 reported on the transamination of ben-
zhydryl imines of simple aldehydes using 0.1 equiv of
KOtBu as base in THF at room temperature, resulting
in good yields with aldimines. The use of ketimines
resulted in enolization and concomitant condensation of
the substrate. Other efforts to achieve asymmetric induc-
tion have been made by Zwanenburg et al.9 with the aid
of a chiral amino alcoholate catalyst resulting in a
maximum asymmetric induction of 44% ee. Some model
systems have been presented in the field of supramo-
lecular chemistry in efforts to mimic the native system
as far as reaction rates and stereoselectivity are con-
cerned. Breslow et al.10 have presented cyclodextrin-
containing models, where the cavity of the cyclodextrin
acted as the active site of the “enzyme”. This resulted in
an enhanced reaction rate as well as a fair asymmetric
induction. The basic sidearm was placed on the upper
rim of the cyclodextrin cavity, in the form of either an
imidazole unit or an ethylenediamine arm. This model
system also presented a bifunctionally acting base.
Other approaches to mimic the action of the enzyme
lysine residue unit have been made by Ahlberg et al.,11
incorporating an aminoalkyl or azapropellane sidearm
in a pyridoxal-based model system, though a later
investigation showed that the preferred conformation of
the aza propellane was not optimal for a bifunctional
action.12
formation of the corresponding carbamate. This protect-
ing group appeared very attractive, since removal of the
9-proton was very easily accomplished by such mild bases
as pyridine and by even heating in aprotic polar solvents
alone (DMF, DMSO). A small modification of the thiox-
anthene moiety was made, and it emerged as 9-ami-
nothioxanthene 10,10-dioxide, previously prepared by
Catscoulakos.14 DBU and triethylamine were initially
employed for comparison of yields and reaction rate
constants. The substrates tested were all R-keto esters
converted to the corresponding amino acid esters (Scheme
3). It has been reported previously that an equilibrium
mixture of the two possible imines can be afforded by just
synthesizing the model substances, though in our ex-
amples no products were observed other than the desired
product imines together with the starting materials.15
The designed system was thus not prone to thermal
equilibration between the two isomeric imines possible.
Boiling the imines (1-4) in triethylamine resulted only
in recovered nonrearranged starting material. Increasing
the pKa of the corresponding acids to DBU afforded
complete rearrangement in all cases but for 1. This is
probably due to the conjugative effect of the nearby
phenyl group, resulting in an equilibrium mixture be-
tween imines 1a and 1b. Other tested catalysts are given
in Chart 1 together with their pKa values in water. For
the purpose of kinetic studies, the use of DBU proved
futile since the reaction rate (even at -10 °C) was much
too fast to be observed by any non-stopped-flow method.
The reaction with DBU as catalyst was completed within
seconds after addition. The use of benzamidine (9), on
the other hand, resulted in a reaction rate suitable for
observation by NMR. The reactions were generally
complete within 4 h (Table 1).
Our research has focused on accomplishing mild and
efficient [1,3]-proton-transfer reactions in the model
systems. We have also explored the possibility of bifunc-
tional catalysis. The desired bifunctionality should pref-
erably consist of a species (the catalyst) that is degenerate
between its two tautomers (Scheme 2). Two such classes
of species proved to be symmetric amidines and guani-
dines.
Resu lts a n d Discu ssion
Exp lor a tor y Wor k . In the design of the transamina-
tion system, it was important to find an amine that upon
formation of imines could be easily deprotonated and the
equilibrium shifted toward the desired product (rear-
ranged imine). When scanning the literature, we found
that Carpino et al.13 had reported a thioxanthene system
in a new protecting group for amines/amino acids by
(8) Cainelli, G.; Giacomini, D.; Trere`, A.; Boyl, P. J . Org. Chem. 1996,
61, 5134-5139.
(9) Willems, J . G. H,; de Vries, J . G.; Nolte, R. J . M.; Zwanenburg,
B. Tetrahedron Lett. 1995, 36, 3917-3920.
(10) (a) Fasella, E.; Dong, S. D.; Breslow, R. Bioorg. Med. Chem.
1999, 7, 709-714. (b) Breslow, R.; Canary, J . W.; Varney, M.; Waddell,
S. T.; Yang, D. J . J . Am. Chem. Soc. 1990, 112, 5212.
(11) (a) Wu, Y.-K.; Ahlberg, P. Acta Chem. Scand. 1992, 46, 60. (b)
Wu, Y.-K.; Ahlberg, P. J . Org. Chem. 1992, 57, 6324-6327.
(12) Berg, U.; Bladh, N. Acta Chem. Scand. 1997, 51, 778-784.
(13) Carpino, L. A.; Gao, H.-S.; Ti, G.-S.; Segev, D. J . Org. Chem.
1989, 54, 5887-5897.
(14) Catscoulakos, P. J . Heterocycl. Chem. 1967, 4, 645-646.
(15) Panetta, C. A.; Dixit, A. S. J . Org. Chem. 1980, 45, 4503-4505.

New Approach to Biomimetic Transamination
J . Org. Chem., Vol. 67, No. 11, 2002 3587
Ch a r t 1
Ta ble 1. Yield of Rea r r a n ged Im in es, w ith 5 Mol % of
Ben za m id in e a n d DBU, Resp ectively, in THF a t 20 °Ca
yield (in %) of imine b
F igu r e 2. Reaction-rate profiles of compounds 1-4 measured
as the disappearance of the starting material. Benzamidine
(9) was used as catalyst (10 mol %), and reactions were
performed in THF-d8 at 35 °C. The y-axis gives the logaritms
of the ratios of the integrals of the moving proton in substrate
and product, and the x-axis gives the time in seconds.
compd
R1
R2
benzamidine
DBU
3a
2a
4a
5a
6a
1a
7a
8a
-Me
-Et
98
98
98
42
51
60
68
72
>99
>99
>99
45
-i-Pr
-Et
-CH2CH2CO2Me
-p-NO2C6H4
-p-FC6H4
-Me
-Et
-Me
-Me
-Me
-Me
56
-Ph
68
-p-MeC6H4
-p-OMeC6H4
74
81
a
All yields are recorded after 4 h.
F igu r e 1. NMR spectrum of the 9-methine proton and the
2-methine proton of compounds 4a and 4b, respectively. Added
amount of benzamidine (9) was 10 mol %, and the time interval
between the first measurement (bottom line) to the last
measurement (top line) was 2 h.
F igu r e 3. Solvent dependence on the rate of the proton-
transfer step of 1. 10 mol % of benzamidine (9) was used at a
temperature 35 °C. Same plot variables as in Figure 2.
Kin etics. Kinetics were studied by means of NMR in
THF-d8, and the reaction rate was determined by mea-
suring the integral of the disappearing 9-methine proton
of the starting material (Figure 1). The reaction rate did
indeed vary with the R group, but there seemed to be no
obvious explanation as to the observed differences (Figure
2), other than possibly a steric contribution. For example,
the kinetics of compound 2 turned out to obey a linear
first-order reaction rate, whereas the other compounds
studied were all nonlinear, probably because of approach
to an equilibrium. In all cases, it turned out that a
prolonged reaction time (<18 h) resulted in cleavage of
the product imine (b), as one of the resulting products
after cleavage could clearly be identified as 9-thioxan-
thone 10,10-dioxide. The use of DBU as catalyst resulted
in slightly higher yields in all cases, and as mentioned
earlier, the reactions were very fast. The only reaction
that did not result in a quantitative yield was with
compound 1. Compound 1 was thus studied more exten-
sively as to effect of solvent. We observed the lowest
initial reaction rate with the most polar solvent (aceto-
nitrile, Figure 3). A possible explanation could be that
the rate-determining step in the proton-transfer reaction
is protonation of the aza allylic anion. The use of
acetonitrile, however, increased the yield (79%) and gave
an observed linear first-order behavior in kinetics. The
use of an apolar solvent (benzene) resulted in a lower
yield (55%) as well as nonlinear reaction profile, due to
the equilibration.
Su bstr a te a n d Ca ta lyst Con for m a tion a n d Bi-
fu n ction a l Ca ta lysis. In exploring the concept of bi-
functional catalysis of the base (such as benzamidine),
knowledge of the conformation of the substrate as well
as the base is important. The substrate imines may
assume an E- or Z-configuration. The NMR spectra of
the substrate imines contain signals from two isomers,
most clearly shown for the methine singlet at ca. δ6, with

3588 J . Org. Chem., Vol. 67, No. 11, 2002
Hjelmencrantz and Berg
Ch a r t 2
[1,2-a]pyrimidine (13-1H) can act as bifunctional cata-
lysts. One other candidate, amidine 11, reported by
Buddrus et al.,18 is, however, prevented from having its
free electron pair in a cis conformation to the proton
(Scheme 2). This restriction proved to be crucial since
attempted transamination with 11 yielded only unreacted
starting material. A possible explanation for the incapa-
bility of this amidine to catalyze proton transfer could
be its inability to both abstract and donate a proton
simultaneously and, hence, partially sustain our idea
about bifunctional action via a double-hydrogen-bonded
intermediate (Chart 2). Such an interpretation is con-
tradicted by the action of DBU being an efficient catalyst,
with no possibility for bifunctional catalysis.
Com p u ta tion a l Ap p r oa ch to Su bstr a te a n d Tr a n -
sition Str u ctu r es. We have performed a computational
study of the conformations of the substrates, transition
states, and products by simple semiempirical calculations
(PM3) and by use molecular mechanics for the ground
states. All the substrate imines except 3a are predicted
to prefer an E configuration according to both PM3 and
MMFF calculations. Both imines a and b are saddle-
shaped, b even a little more than a .
J udging from the calculated structure of conformer A,
it appears that approach by a base to abstract the
9-proton is more hindered. We reasoned that optimal
conditions for efficient bifunctional proton-transfer ca-
talysis would be governed by a conformation of starting
imine in which the C(9)-H bond is perpendicular to the
imine plane also for electronic reasons. To gain informa-
tion about the mechanism, we decided to localize transi-
tion structures for proton-transfer reaction by computa-
tions using formamidine as base. It turned out that we
were unable to find any transition state for concerted
bifunctional proton transfer. Instead, formamidine gave
a transition structure for 1 corresponding to the first step
in a stepwise mechanism in which the imine nitrogen
atom acts as a base (Figure 5, TS1). The proton of that
nitrogen is hydrogen-bonded to the syn oxygen of the
sulfone. Furthermore, a proton of the other nitrogen atom
is hydrogen bonded to the carbonyl oxygen of the ester
group. These hydrogen bonds are most likely important
for both reactivity and stereoselectivity of the proton-
transfer reaction. The second hydrogen bond points at
an alternative mechanism for the protonation step in
which the proton is tranfered to the oxygen atom giving
an enol intermediate, which rearranges to the product
imino ester. In such a mechanism, the configuration of
the product would be determined in the last ketonization
step. We were unable to localize a transition state for
proton transfer to the oxygen. Instead, a transition
structure was localized for the protonation at the carbon
atom (TS2) having a hydrogen bond to the carbonyl
oxygen. TS1 is 0.32 kcal/mol higher than TS2. The
calculations refer to gas-phase reaction, and the two
reaction steps have opposite charge developments, which
F igu r e 4. Two different boat-shaped conformations of com-
pound 1. Heats of formation were calculated by the semiem-
pirical PM3 method. The E isomer was lower in energy for
both conformations.
the high-field signal dominating (85-98%). Compound
3a , however, had reversed preference (80:20). An alter-
native interpretation of the NMR spectra would be a
conformation equilibrium of the thioxanthenyl ring sys-
tem as shown in Figure 4. Earlier investigations indicate
a boat-shaped conformation of the central ring.16 The
solid-state structure of 9-cyclohexylthioxanthene 10,10-
dioxide was reported to have the cyclohexyl substituent
in the axial position.17
A way to distinguish between the two alternative
interpretations would be to freeze out the thioxanthenyl
ring flip. To investigate if the two conformers could be
observed by NMR, a sample of 1a in THF-d8 was cooled
to -90 °C, but did not give any evidence for a dynamic
behavior slow enough to be observed on the NMR time
scale. The barrier to inversion between the two conform-
ers could hence be assumed to be small. Another pos-
sibility would be a strongly biased equilibrium or that
we are dealing with a planar three-ring system, though
this seemed highly unlikely, since both our theoretical
studies as well as earlier investigations point at a boat-
shaped conformation of the central ring. A NOESY
spectrum of imine 1a (in THF-d8), showed no cross
coupling peaks between the 9-methine and any other
proton. This behavior would be expected if the proton
were situated in the axial position in agreement with the
chemical shift considerations. Heating a sample of 7a in
2H2-1,1,2,2-tetrachloroethane to 95 °C did not lead to
significant broadening of the signals. Such a high barrier
is expected for E/Z isomerization but not for thioxanthe-
nyl ring inversion.
Conformer B is favorable to approach by a base to
abstract the 9-methine proton for both steric and elec-
tronic reasons. The resulting intermediate would be a
more planar resonance stabilized anion before reproto-
nation to the rearranged imine. Conformer A, on the
other hand, has to undergo a larger conformational
change to achieve such stabilization. We consider it
reasonable to assume that B is the preferred conforma-
tion for all compounds except maybe for 3a , as suggested
by the NOESY spectrum.
The second stereochemical issue is the conformation
of the catalyst. To achieve concerted bifunctional cataly-
sis, the catalyst must be able to both donate and abstract
a proton. Two such classes of basic compounds are
amidines and guanidines. DBU, however, cannot meet
up to the requirements. Benzamidine 9 and the com-
mercially available 1,3,4,6,7,8-hexahydro-2H-pyrimido-
(16) Ternay, A. L.; Evans, S. A. J . Org. Chem. 1974, 39, 2941-2946.
(17) Chu, S. S. C.; Mangion, M. Acta Crystallogr. 1975, B31, 2134-
2136.
(18) Dauwe, C.; Buddrus, J . Synthesis 1995, 2, 171-172.

New Approach to Biomimetic Transamination
J . Org. Chem., Vol. 67, No. 11, 2002 3589
Ta ble 2. Yield of Rea r r a r a n ged Im in es Accor d in g to
Sch em e 3, w ith 5 Mol % of Gu a n id in e (12) a s Ca ta lyst in
THF a t 23 °Ca
compd
R1
R2
yield of imine b (%) ee (%)
1
2
3
4
-Ph
-Me
-Et
-Et
63
95
98
98
43
45
24
0
-i-Pr
-Me
-CH2CH2COOMe -Me
a
All yields are recorded after 4 h.
of the guanidine. The enantiomeric excess of this base
varied between 0 and 45% (Table 2). Imine 4 resulted in
a racemic mixture of 4b. Somewhat higher asymmetric
induction was achieved for 1-3. The lack of asymmetric
induction of 4 may be a result of the high reaction rate
observed in THF (Figure 2). Imine 2 resulted in the
highest enantiomeric excess (45%), followed by imine 1
(43%). Both these imines contain rather sterically de-
manding substituents, which may be a reason for the
increased asymmetric induction. Chiral amines, such as
dihydroquinidine, did not give rise to any proton transfer.
Step w ise ver su s Con cer ted Ca ta lysis. Even though
several studies indicate that amidines do act bifunction-
ally,22 Ahlberg et al. claim that in bifunctional catalysis
the concerted mechanism is higher in energy than a step-
wise mechanism.20 One observation substantiating this
idea appears on addition of base to all imines-a colored
solution immediately emerges. A possible explanation
would be that we have a charge-transfer complex (be-
tween base and substrate) leading to coloration or an
intermediate ion formation. To eliminate the possibility
of charge-transfer complexation, a simple test was done
with KOtBu as base. This base is incapable of forming a
charge-transfer complex, but a colored solution still re-
sulted upon addition of the base, hence eliminating the
possibility of a charge-transfer complex as a cause for
the observed coloration. In final essence, we are left with
the option of an intermediate anion. We envisage, how-
ever, that a two-point interaction as in bifunctional cat-
alysis increases the possibility of enantioselective proto-
nation (Chart 1).
F igu r e 5. Two transition structures for the stepwise proton
transfer of 1 by formamidine calculated by the semiempirical
PM3 method.
certainly affects the quantitative properties of the transi-
tion structures and energies. We are currently working
on higher level computations in order to gain information
of the mechanism and to guide us in the search for better
enantioselectivity.
Th e Electr on ic In flu en ce. The effect on the proton-
transfer reaction was examined by substitution of the
phenyl ring in compound 1. Substituents were p-OMe,
p-Me, p-F, and p-NO2 (Chart 2). A moderate influence
on the reaction rate was observed, with the rate constant
varying within 1 order of magnitude depending on the
substituent (Figure 6).
With the use of LFER para-substituent parameters
(σp),23 a Hammett plot (Figure 7) displayed a poorly linear
relationship, with a F ) -0.58, which is a low value for
a rate-determining protonation step but could be ex-
plained by a hydrogen bond to the carbonyl oxygen as in
TS1 or to the acceptor carbon in the deprotonation tran-
sition state. It would have been further enlightening to
study the reaction rate behavior of compounds 1 and 5-8,
but since these reactions proceeded at a very high reac-
tion rate under identical conditions, any comparison was
impossible. Little has been studied on the electronic effect
of para substitution of phenyl derivatives of substrates
in transamination reactions. The only report (to our
Isotop e Effect. To further elucidate the nature of the
reaction, a deuterated base was synthesized (13-2H) with
a degree of deuteration of >98% according to NMR.19 The
base was used in excess (1.1 equiv). This resulted in 66
( 2% deuteration of the rearranged imine (2b). Using
an iterative method, simulation of the course of bifunc-
tional reaction, gave 66% deuteration assuming a kinetic
isotope effect of 1.0, and it was possible to determine a
maximum value for the isotope effect of 1.2 (giving 64%
deuterated product). This is identical to the value found
by Ahlberg et al. for amidine-catalyzed 1,3-proton-
transfer reactions in indene derivatives.20 Such a small
isotope effect is hardly compatible with a rate-determin-
ing protonation.
Asym m etr ic Ca ta lysis. These results were further
evaluated in our attempts toward asymmetric catalysis,
by the synthesis and use of a phenyl glycine based guani-
dine 12 first synthesized by Corey et al.21 The pseudo-C2
symmetry of this base is a prerequisite for efficient
asymmetric induction due to the tautomeric properties
(19) Kaempfen, U.; Eschenmoser, A. Helv. Chim. Acta 1989, 72,
185-195.
(22) (a) J anne´, K.; Ahlberg, P. J . Chem. Soc., Chem. Commun. 1976,
1040-1041. (b) Ek, M.; Ahlberg, P. Acta Chem. Scand. 1984, B38, 211-
217.
(23) Isaacs, N. Physical Organic Chemistry, 2nd ed.; Longman
Scientific & Technical: Essex, 1995; pp 152-153.
(20) Ahlberg, P.; J anne´, K.; Lo¨fås, S.; Nettelblad, F.; Swahn, L. J .
Phys. Org. Chem. 1989, 2, 429-447.
(21) (a) Corey, E. J .; Grogan. M. J . Org. Lett. 1999, 1, 157-160.

3590 J . Org. Chem., Vol. 67, No. 11, 2002
Hjelmencrantz and Berg
base in an autocatalytic manner.25 The product imine has
a proton sponge like structure26 (compare, e.g., 1,8-(bis-
dimethylamino)naphthalene, 14). Since the nitrogen-
carbon bond in the imine moiety is planar, a similar
crowding of the free electron pairs in the product imine
is obtained resulting in a stronger base than the starting
imine (25) and, hence, should be able to act as a base to
accomplish [1,3]-proton transfer. In addition, some cal-
culations (molecular mechanics and semiempirical PM3)
were made that pointed toward a negative ∆H for the
desired reaction. However, the only imines that could be
formed with amine 24 were with highly reactive alde-
hydes such a crotonaldehyde. Other (enolizable) alde-
hydes were simply enolized by the diamine, resulting in
condensation. Efforts to rearrange this crotyl imine failed.
Only unrearranged starting material was recovered when
subjected to several bases such as KOtBu, DBU, or
benzamidine. This could very well be the result of a
sterically demanding environment of the 9-proton caused
by the dimethyl amino group in the 1-position or an
energetically unfavorable product.
F igu r e 6. Varying reaction rates of compounds 1 and 5-8.
All performed in CD3CN/THF-d8 (5:2), with 10 mol % benza-
midine (9) at 303 K, and measured as the disappearance of
the imine (a ).
Syn th esis. The key amine in the study was synthe-
sized according to the route outlined in Scheme 4,
starting from commercially available thioxanthone, which
was oxidized by NaIO4/RuCl3.27 9-Thioxanthone 10,10-
dioxide was converted to its oxime (15) by reaction with
HONH2‚HCl in pyridine and ethanol and thereafter
acetylated by acetic anhydride/pyridine. The acetylated
oxime 16 was reduced to the key amine 17 by reduction
of BH3‚THF complex. The earlier reported yield of this
amine was not reproducible, but instead turned out be
lower. The workup of this product proved to be an ordeal,
since the use of the borane reagent produced a thick
white precipitate, complicating isolation of the product.
Attempts to produce the simple imine by reacting 9-thiox-
anthone 10,10-dioxide with ammonia (in the presence of
TiCl4 as catalyst) failed, since the product of this reaction
was much too unstable to allow isolation and concomitant
reduction. Nonaromatic imines were synthesized by
reacting the key amine 17 with the desired R-keto ester
in CH2Cl2, in the presence of dry MgSO4, for 16 h at room
temperature. In the case of aromatic imines, the reaction
was performed in dry benzene in the presence of molec-
ular sieves (4 Å), with BF3‚OEt2 as the activating
catalyst, and heated to reflux for 12 h. Glyoxylates with
different para-substitution patterns were synthesized
according to the method by Babudri et al. using the LiCl/
CuBr-catalyzed Grignard reaction with methyl chlo-
rooxoacetate.28 Synthesis of amidine 11 and guanidine
12 were performed according to known literature proce-
dures. Deuteration of compound 13-1H to produce 13-2H
was accomplished by repeated solvation and evaporation
of the guanidine three times from CD3OD.
F igu r e 7. Hammett plot for the observed reaction rates as
depicted in Figure 6, yielding a F ) -0.58.
knowledge) investigated the rearrangement of a series
of para-substituted phenylalanine derivatives in the
context of transamination.24 That investigation pointed
at an accelerated reaction rate when electron-donating
substituents in the para position were present, such as
p-aminophenylalanine, whereas p-nitrophenylalanine was
found to be inactive.
Au toca ta lytic Mod el. Since DBU could catalyze
transamination in our model systems, we decided to
incorporate a basic moiety in the thioxanthene ring
system. The base was incorporated into the model system
as a dimethyl amino group in the 1-position (24). The
Synthesis of the 1-dimethylaminothioxanthone dioxide
structure (24, Scheme 5) was performed by an earlier
reported procedure by Snieckus et al.29 with the starting
2-N,N-(dimethylamino)-N,N-diethylbenzamide (20) pre-
pared according to the method reported by Murphy et
al.30 The resulting key ketone 22 was converted to the
(25) Soai, K.; Shibata, T.; Sato, I. Acc. Chem. Res. 2000, 33, 382.
(26) Hibbert, F.; Hunte, P. P. J . Chem. Soc., Perkin Trans. 2 1983,
1895-1899.
idea was to use the rearranged imine to act as its own
(27) Su, W. Tetrahedron Lett. 1994, 35, 4955-4958.
(28) Babudri, F.; Fiandanese, V.; Marchese, G.; Punzi, A. Tetrahe-
dron 1996, 52, 13513-13520.
(24) Taylor, D. C.; Wightman, F. Phytochemistry 1987, 26, 2167-
2179.
(29) Beaulieu, F.; Snieckus, V. J . Org. Chem. 1994, 59, 6508-6509.

New Approach to Biomimetic Transamination
J . Org. Chem., Vol. 67, No. 11, 2002 3591
Sch em e 4
Sch em e 5
amine by formation of imine 23 in NH3 with TiCl4 as
catalyst. The imine was converted into its amine 24 by
reduction in MeOH/NaBH4. The final product (25) was
produced by reaction of 24 in CH2Cl2 with croton alde-
hyde in the presence of MgSO4 at room temperature.
with a hydrogen bond to the exocyclic carbon or carbonyl
oxygen in the transition state rather than a rate-deter-
mining protonation step. The electronic dependence in
para-substituted phenyl glyoxylate imines, isotope effects,
and conformational studies support a stepwise, bifunc-
tional mechanism for amidine and guanidine catalysts.
This is further substantiated by our observation of color-
ation of the reaction solution upon addition of the base.
The computations support a stepwise mechanism with
two essentially equi-energetic transition states, and the
amidine transition state gives valuable information of
how to improve the stereoselectivity of the reaction. An
attempt to construct an autocatalytic system failed be-
cause of difficulties to synthesize the required imines.
Experiments aiming at improved enantioselective reac-
tions are in progress.
Con clu sion s
We have described a new mild and efficient biomimetic
transamination method for R-amino acid esters. Amines,
such as triethylamine and dihydroquinidine, failed to
catalyze the proton transfer, whereas the more basic
benzamidine and DBU proved to give high yields, DBU,
in addition, gave high reaction rates. The use of chiral
catalysts resulted in an asymmetric induction with ee’s
up to 45%. Introduction of substituents in the para
position of the phenylglycine precursor (1) resulted in a
small variation in the observed first-order reaction con-
stants (varying only within 1 order of magnitude). The
low negative value of F indicates that we might be dealing
Exp er im en ta l Section
NMR (1H, 13C) was performed at 400 and 100 MHz, respec-
tively. Melting points are reported herein as uncorrected
melting points. All commercially available starting materials
were used as received unless otherwise noted. Thin-layer
(30) Murphy, J . A.; Roome, S. J . J . Chem. Soc., Perkin Trans. 1 1995,
1349-1358.

3592 J . Org. Chem., Vol. 67, No. 11, 2002
Hjelmencrantz and Berg
chromatography was performed on silica gel plates with
aluminum supports. Purification of all imines was performed
on a chromatotron using premade plates with neutral alumi-
num oxide. Investigations of enantioselectivity was performed
by quenching the sample with gaseous HCl and a small
amount of water. The sample was dried and thereafter treated
with trifluoroacetic anhydride or CBz-Cl together with excess
pyridine and allowed to stand at room temperature for 30 min
to produce the corresponding N-trifluoroacetyl- or Cbz-
protected amino acid esters. The sample was thereafter passed
through a short silica column to remove any basic remains. A
small part of the sample was taken out, diluted with aceto-
nitrile, and injected onto a column (column dimensions 30 ×
0.5 cm) with 2-10% 2-propanol in hexane as the mobile phase
at a flow rate of 0.5 mL/min.
was dissolved in methylene chloride and filtered to afford a
clear solution. The crude material was purified by chroma-
tography (Rf ) 0.35, 10% ether/CH2Cl2) and recrystallized from
heptane/toluene to afford white crystals: yield 75%; mp 158-
1
160 °C; H NMR (CDCl3) δ 8.16-8.13 (m, 2H), 8.02-7.99 (m,
2H), 7.61-7.52 (m, 7H), 7.47-7.43 (m, 2H), 5.93 (s, 1H), 3.90
(s, 3H); 13C NMR (CDCl3) δ 165.33, 165.20, 140.20, 136.67,
133.39, 132.67, 132.62, 129.24, 128.22, 128.20, 126.70, 124.14,
60.98, 52.78; HRMS (FAB+) calcd for C22H18NO4S (M + 1)
392.0957, found 392.0962; FT-IR (KBr pellet cm-1) 1740, 1637,
1298, 1164.
N-(2-Meth ylben zoylfor m yl)-9-th ioxa n th en yl 10,10-d i-
oxid e im in e (1b): mp 155-156 °C; Rf ) 0.27 (10% ether/CH2-
Cl2); 1H NMR (CDCl3) δ 8.25 (ddd, J ) 6.98,1.69, 0.67 Hz, 1H),
8.16 (ddd, J ) 7.62, 1.69, 0.48 Hz, 1H), 8.06 (ddd, J ) 7.41,
1.90, 0.42 Hz, 1H), 7.72-7.62 (m, 6H), 7.51-7.37 (m, 4H), 7.87
(s, 1H), 3.76 (s, 3H); 13C NMR (CDCl3) δ 171.12, 157.22, 139.94,
139.26, 138.51, 137.78, 133.26, 132.26, 131.00, 130.25, 129.20,
128.62, 128.52, 128.18 (br), 127.56, 125.33 (br), 123.18 (br),
70.10, 53.01; HRMS (FAB+) calcd for C22H18NO4S (M + 1)
392.0957, found 392.0963; FT-IR (KBr pellet cm-1) 1735, 1603,
1299, 1171.
The isotope effect experiments were analyzed in Excel using
a iterative method based upon the rate equation
d[PH(D)
dt
]
) kH(D)[NH(D)][S]
Gen er a l P r oced u r e for Syn t h esis of Im in es 2a -4a .
N-(9-Th ioxan th en yl 10,10-dioxide)eth yl Isopr opylfor m yl-
im in e (2a ). 9-Aminothioxanthene 10,10-dioxide (0.408 mmol,
100 mg) and ethyl 3-methyl-2-oxobutyrate (0.408 mmol, 0.60
µL) were mixed together in a dry round-bottomed flask. Dry
toluene was added (3.0 mL). The resulting mixture was lightly
heated to dissolve all components in the reaction. The reaction
solution was allowed to cool to ambient temperature, where-
upon molecular sieves (4 Å, activated) were added. The
reaction flask was heated to reflux for 17 h. The molecular
sieves were separated and rinsed well with several portions
of CH2Cl2. The combined organic layers were evaporated in
vacuo to afford a yellow oil. The crude material was purified
by chromatography (5% ether in CH2Cl2, Rf ) 0.68) to afford
a pale yellow oil, which crystallized upon standing: yield 75%;
where P, N, and S refer to product, catalyst, and substrate
and H(D) refers protonated and deuterated species, respec-
tively. The reverse reaction was neglected in the calculations.
The responses to an increment in time were added to previous
values of P, N, and S until 100% reaction.
The molecular modeling calculations were performed on a
Silicon Graphics O2 workstation using the SPARTAN 5.0.3
program.31 The geometries were evaluated with full optimiza-
tion employing the semiempirical PM3 Hamiltonian or mo-
lecular mechanics MMFF force field. The conformational
search was performed employing the systematic Monte Carlo
method. Transition states were localized, and vibrational
frequencies were calculated for transition states verifying the
existence of a single imaginary frequency for the proper
reaction coordinate. Default convergence criteria for SCF
convergence were used, whereas transition structure optimiza-
tion required several thousand cycles and small steps in each
cycle (0.1 au).
Gen er a l Exp er im en ta l P r oced u r es. Transaminations
were performed at ca. 0.02 mmol (∼10 mg) scale of imine
dissolved in THF or THF/CH3CN (5:2) with a concentration
of 0.035 M substrate. A stock solution of benzamidine (0.05
M, in THF) was added, and the solution was stirred at room
temperature for 4 h. The solvent was evaporated, and the
crude material was purified by chromatography on a chroma-
totron. Imines 2a ,b-4a ,b were simply passed through a short
column of neutral Al2O3 (attempts to purify on silica gel
resulted in hydrolysis). Compounds 1a ,b and 6a ,b-8a ,b were
separated on Al2O3 (neutral) with 10% ether/CH2Cl2. Due to
problems in separation of 5a and 5b, no further attempts were
made to purify the individual compounds.
1
mp 70-72 °C; H NMR (CDCl3) δ 8.08 (dd, J ) 7.5, 1.5 Hz,
2H), 7.51 (dd, J ) 7.50, 1.67 Hz, 2H) 7.50-7.44 (m, 2H), 7.41-
7.38 (m, 2H), 5.94 (s, 1H), 4.23 (q, J ) 7.1 Hz, 2H), 3.12 (sept,
J ) 6.9 Hz, 1H), 1.41 (d, J ) 6.9 Hz, 6H), 1.15 (t, J ) 7.1 Hz,
3H); 13C NMR (CDCl3) δ_ 174.06, 164.81, 140.56, 136.57,
132.44, 127.85, 126.02, 123.93, 61.90, 59.61, 35.80, 19.81,
14.19; HRMS (EI+) calcd for C20H21NO4S (M) 371.1192, found
371.1200; FT-IR (KBr pellet cm-1) 1730, 1662, 1257, 1166.
N-(2-Isop r op yleth ylfor m yl)-9-th ioxa n th en yl 10,10-d i-
oxid e im in e (2b): Rf ) 0.64 (10% ether in CH2Cl2); mp 68-
1
70 °C; H NMR (CDCl3) δ 8.21-8.19 (m, 1H), 8.02 (ddd, J )
7.5, 1.3, 0.7 Hz, 1H), 7.94 (ddd, J ) 8.6, 1.6, 0.4 Hz, 1H), 7.72-
7.69 (m, 1H), 7.66-7.63 (m, 2H), 7.61 (dd, J ) 7.6, 1.5 Hz,
1H), 7.58 (dd, J ) 7.5, 1.5 Hz, 1H), 4.44 (d, J ) 7.0 Hz, 1H),
4.35 (q, J ) 7.1 Hz, 2H), 2.43 (octet, J ) 6.8 Hz, 2H), 1.38 (t,
J ) 7.1 Hz, 3H), 0.90 (d, J ) 6.8 Hz, 3H), 0.86 (d, J ) 6.7 Hz,
3H); 13C NMR (CDCl3) δ 171.58, 157.82, 139.88, 137.86, 133.44,
133.19, 132.21, 131.52, 130.62, 129.94, 129.19, 127.91, 125.11,
122.91, 73.14, 61.51, 33.28, 19.43, 18.56, 14.52; HRMS (FAB+)
calcd for C20H22NO4S (M + 1) 372.1270 found 372.1269; FT-
IR (KBr pellet cm-1) 1732, 1635, 1319, 1167.
Kinetics were run in THF-d8 for imines 2a -4a and CD3-
CN/THF-d8 (5:2) for imines 1a and 5a -8a , in a total volume
of 0.7 mL, at a concentration of 0.035 M. Solvents were used
as received, and runs were made at 303 K. Benzamidine was
added as a stock solution in THF-d8 (0.05 M).
N-(9-Th ioxa n th en yl 10,10-d ioxid e)eth yl m eth ylfor m yl
im in e (3a ): yield 73%; Rf ) 0.60 (10% ether/CH2Cl2); mp 45-
Gen er a l P r oced u r e for Syn th esis of Com p ou n d s 1a
a n d 5a -8a . N-(9-Th ioxa n th en yl 10,10-d ioxid e)m eth yl
Ben zoylfor m yl Im in e (1a ). Methyl benzoylformate (2.04
mmol, 0.29 mL) was dissolved in dry benzene (15 mL) in a
dry round-bottomed flask. Dry benzene was added (20 mL).
BF3‚OEt2 (2.04 mmol, 0.26 mL) was added, and the solution
was stirred at room temperature for 5 min. 9-Aminothioxan-
thene 10,10-dioxide (2.04 mmol, 500 mg) was added, and the
heterogeneous mixture was heated to reflux for 15 min and
thereafter allowed to cool to room temperature. Activated
molecular sieves (4 Å) were added, and the resulting reaction
mixture was heated to reflux for 12 h. The solvent was
removed by evaporation in vacuo. The resulting crude solid
1
50 °C; H NMR (CDCl3) δ 8.14-8.11 (m, 2H), 7.55-7.50 (m,
6H), 6.19 (s, 1H), 4.48 (q, J ) 7.1 Hz, 2H), 2.29 (s, 3H), 1.47 (t,
J ) 7.1 Hz, 3H); 13C NMR (CDCl3) δ 166.71, 164.96, 139.57,
136.62, 132.74, 128.15, 125.60, 124.27, 62.77, 58.01, 16.22,
14.38; HRMS (FAB+) calcd for C18H18NO4S (M + 1) 344.0957,
found 344.0952; FT-IR (KBr pellet cm-1) 1741, 1632, 1311,
1151.
N-(2-Eth ylm eth ylfor m yl)-9-th ioxa n th en yl 10,10-d iox-
id e im in e (3b): Rf ) 0.57 (10% ether/CH2Cl2); mp 52-54 °C;
1H NMR (CDCl3) δ 8.23-8.20 (m, 1H), 8.03 (ddd, J ) 7.53,
1.88, 0.43 Hz, 1H), 7.95 (ddd, J ) 7.53, 1.88, 0.45 Hz, 1H),
7.75-7.72 (m, 1H), 7.65 (dd, J ) 6.58, 0.9 Hz, 1H), 7.59 (dd, J
) 6.48, 0.84 Hz, 1H), 7.58-7.56 (m, 2H), 4.91 (q, J ) 6.51 Hz,
1H), 4.30 (q, J ) 7.11 Hz, 2H), 1.56 (d, J ) 6.51 Hz, 3H), 1.34
(t, J ) 7.15 Hz, 3H); 13C NMR (CDCl3) δ 172.04, 157.02, 133.31,
(31) Hehre, W. J .; Burke, L. D.; Shusterman A. J .; Pietro, W. J .
SPARTAN, version 5.0.3; Wavefunction, Inc., 18401 Von Karman, Suite
370, Irvine, CA, 92715.

New Approach to Biomimetic Transamination
J . Org. Chem., Vol. 67, No. 11, 2002 3593
132.35, 130.77, 130.02, 128.71, 127.76, 125.21, 123.07, 61.79,
61.46, 20.00, 14.42; HRMS (FAB+) calcd for C18H18NO4S (M
+ 1) 344.0957, found 344.0943; FT-IR (KBr pellet cm-1) 1745,
1641, 1315, 1160.
N-(Met h yl(p -m et h ylben zoyl)for m yl)-9-t h ioxa n t h en yl
10,10-d ioxid e im in e (7b): Rf ) 0.15 (10% ether/CH2Cl2); mp
1
162-164 °C; H NMR (CDCl3) δ 8.23 (dd, J ) 7.32, 1.52 Hz,
1H), 8.13 (dd, J ) 7.40, 1.54 Hz, 1H), 8.04 (dd, J ) 7.40, 1.52
Hz, 1H), 7.67 (tdd, J ) 7.28, 2.14, 1.56 Hz, 2H), 7.60 (tdd, J )
7.60, 2.14, 1.44 Hz, 2H), 7.49 (d, J ) 8.36 Hz, 2H), 7.46 (dd, J
) 6.85, 0.58 Hz, 1H), 7.23 (d, J ) 7.9 Hz, 2H), 5.79 (s, 1H),
3.73 (s, 3H), 2.37 (s, 3H); 13C NMR (CDCl3) δ 171.35, 156.99,
139.97, 139.33, 138.50, 137.79, 135.66, 133.29, 132.24, 131.07,
130.99, 130.24, 129.96, 128.61, 128.16, 127.45, 125.30, 123.07,
69.94, 53.03, 21.42; HRMS (FAB+) calcd for C23H20NO4S (M
+ 1) 406.1114, found 406.1107; FT-IR (KBr pellet cm-1) 1731,
1539, 1321, 1154.
N-(9-Th ioxa n th en yl 10,10-d ioxid e)d im eth ylglu ta r a te-
2-im in e (4a ): yield 86%; mp 58-60 °C; Rf ) 0.21 (10% eter/
1
CH2Cl2); H NMR (CDCl3, 400 MHz) δ 8.09 (dd, J ) 7.5, 1.7
Hz, 2H), 7.58-7.46 (m, 4H), 7.32 (d, J ) 7.7 Hz, 2H), 6.15 (s,
1H), 3.78 (s, 3H), 3.65 (s, 3H), 3.20 (t, J ) 6.6 Hz, 2H), 2.90 (t,
J ) 6.2 Hz, 2H). 13C NMR (CDCl3, 100 MHz) δ 173.31, 166.22,
163.91, 140.37, 136.52, 132.52, 127.98, 126.19, 124.07, 59.97,
52.80, 52.04, 31.86, 29.25; HRMS (FAB+) calcd for C20H20NO6S
(M + 1) 402.1012 found 402.1021; FT-IR (KBr pellet cm-1
1748, 1736, 1650, 1299, 1162.
)
N-(9-Th ioxa n th en yl 10,10-d ioxid e)m eth yl(p-m eth oxy-
ben zoyl)for m yl im in e (8a ): Yield 59%; Rf ) 0.29 (20% ether/
N-(2-Dim eth ylglu ta r a te)-9-th ioxa n th en yl 10,10-d iox-
id e im in e (4b): mp 61-63 °C; Rf ) 0.10 (10% ether/CH2Cl2);
1H NMR (CDCl3) δ 8.18-8.15 (m, 1H), 7.98 (dd, J ) 7.3, 2.2
Hz, 1H), 7.90 (dd, J ) 7.5, 1.8 Hz, 1H), 7.85-7.83 (m, 1H),
7.73-7.69 (m, 2H), 7.66-7.60 (m, 2H), 4.99-4.96 (m, 1H), 3.79
(s, 3H), 3.44 (s, 3H), 2.31-2.07 (m, 4H); 13C NMR (THF-d8) δ
173.42, 171.65, 159.46, 141.40, 140.65, 139.73, 133.64, 133.08,
132.49, 131.64, 130.84, 129.73, 128.45, 125.50, 123.39, 65.54,
1
CH2Cl2); yield 58%; mp 168-170 °C; H NMR (CDCl3) δ 8.13
(dd, J ) 7.25, 2.1 Hz, 2H), 7.95 (d, J ) 8.90 Hz, 2H), 7.54-
7.51 (m, 4H), 7.47-7.46 (m, 2H), 7.03 (d, J ) 9.0 Hz, 2H), 5.88
(s, 1H), 3.91 (s, 3H), 3.88 (s, 3H); 13C NMR (CDCl3) δ 165.61,
164.49, 163.23, 140.54, 136.59, 132.65, 130.04, 128.11, 126.75,
126.04, 124.06, 114.61, 60.73, 55.78, 52.70; HRMS (FAB+)
calcd for C23H20NO5S (M + 1) 422.1063, found 422.1058; FT-
IR (KBr pellet cm-1) 1755, 1683, 1312, 1166.
52.623, 51.34, 30.28, 29.95; HRMS (FAB+) calcd for C20H20
-
N-(Met h yl(p -m et h oxyb en zoyl)for m yl)-9-t h ioxa n t h e-
n yl 10,10-d ioxid e im in e (8b): Rf ) 0.18 (20% ether/CH2Cl2);
NO6S (M + 1) 402.1012, found 402.1018; FT-IR (KBr pellet
cm-1) 1752, 1739, 1661, 1305, 1181.
1
mp 166-167 °C; H NMR (CDCl3) δ 8.22 (dd, J ) 7.87, 1.43
N-(9-Th ioxa n t h en yl 10,10-d ioxid e)et h yl(p -n it r ob en -
zoyl)for m ylim in e (5a ): yield 79%; Rf ) 0.26 (30% eter/CH2-
Hz, 1H), 8.12 (dd, J ) 7.63, 1.43 Hz, 1H), 8.04 (dd, J ) 7.64,
1.40 Hz, 1H), 7.67 (tdd, J ) 8.20, 1.64, 0.90 Hz, 2H), 7.61 (tdd,
J ) 7.50, 1.65, 1.20 Hz, 2H), 7.52 (d, J ) 8.40 Hz, 2H), 7.47
(dd, J ) 8.21, 0.90 Hz, 1H), 6.95 (d, J ) 8.40 Hz, 2H), 5.78 (s,
1H), 3.82 (s, 3H), 3.74 (s, 3H); 13C NMR (CDCl3) δ 171.46,
159.83, 156.92, 139.98, 139.35, 137.80, 133.30, 132.26, 131.11,
131.00, 130.71, 130.25, 128.74, 128.61, 128.15, 125.32, 123.09,
1
Cl2); yield 69%; H NMR (CDCl3) δ 8.24 (d, J ) 8.8 Hz, 2H),
8.09 (dd, J ) 7.50, 1.0 Hz, 1H), 8.03 (dd, J ) 7.50, 1.1 Hz,
1H), 7.77 (d, J ) 8.9 Hz, 2H), 7.72-7.64 (m, 4H), 7.56-7.54
(m, 2H), 5.92 (s, 1H), 4.27 (m, 2H), 1.27 (t, J ) 7.2 Hz, 3H);
13C NMR (CDCl3) δ 169.41, 158.58, 147.92, 145.21, 139.94,
139.04, 137.76, 133.35, 132.47, 131.31, 130.89, 130.52, 128.77,
128.30, 127.98, 125.46, 124.08, 123.10, 69.39, 62.61, 32.01,
22.82, 14.17; HRMS (FAB+) calcd for C23H19N2O6S (M + 1)
451.0965, found 451.0959; FT-IR (KBr pellet cm-1) 1732, 1620,
1521, 1348, 1313, 1165.
114.66, 69.56, 55.55, 53.03; HRMS (FAB+) calcd for C23H20
-
NO5S (M + 1) 422.1063, found 422.1066; FT-IR (KBr pellet
cm-1) 1723, 1627, 1302, 1138.
(+)-2-P h en yl-(3a r ,7a f)-3a ,4,5,6,7a -h exa h yd r o-1H-ben z-
im id a zole (11): yield 86%; mp 152-153 °C; [R]21 ) 191 (c
D
N-(Meth yl(p-flu or oben zoyl)for m yl)-9-th ioxan th en yl 10,-
10-d ioxid e im in e (5b) could not be separated from 5a and
was characterized in the mixture.
) 3.98, pyridine); 1H NMR (CDCl3) δ 7.76-7.73 (m, 2H), 7.35-
7.25 (m, 3H), 5.32 (s, 1H), 3.09 (m, 2H), 2.31 (m, 2H), 1.82 (m,
2H), 1.54 (m, 2H), 1.38 (m, 2H).
N-(9-Th ioxa n th en yl 10,10-d ioxid e)m eth yl(p-flu or oben -
zoyl)for m ylim in e (6a ): yield 76%; Rf ) 0.22 (10% ether/CH2-
(3R,7R)-3,7-Dip h en yl-1,4,6-t r ia za b icyclo[3.3.0]oct -4-
1
en e (12): yield 33%; mp 155-157 °C; [R]21 ) 20.1 (c ) 0.55,
D
Cl2); yield 65%; mp 156-158 °C; H NMR (CDCl3) δ 8.23 (dd,
1
CHCl3); H NMR (CDCl3) δ 7.35 (d, J ) 7.2 Hz, 4H), 7.30 (m,
4H), 7.21 (m, 2H), 5.15 (t, J ) 6.3 Hz, 2H), 3.48 (t, J ) 7.5 Hz,
2H), 3.00 (dd, J ) 6.1 Hz, J ) 7.6 Hz, 2H).
J ) 7.38, 1.9 Hz, 1H), 8.10 (dd, J ) 7.61, 1.7 Hz, 1H), 8.05
(dd, J ) 7.60, 1.8 Hz, 1H), 7.68 (tdd, J ) 6.73, 2.59, 1.4 Hz,
2H), 7.62-7.52 (m, 4H), 7.48 (dd, J ) 7.25, 1.8 Hz, 1H), 7.24
(m, 2H), 5.92 (s, 1H), 2.93 (s, 3H); 13C NMR (CDCl3) δ 165.12,
163,72, 140.03, 136.66, 132.67, 130.52, 130.42, 128.26, 126.65,
124.17, 116.55, 116.33, 61.05, 52.89; HRMS (FAB+) calcd for
9-Th ioxa n th on oxim e 10,10-d ioxid e (15): Rf ) 0.36 (10%
ether in CH2Cl2); yield 85%; 1H NMR (DMSO-d6) δ 8.47 (dd, J
) 7.8, 1.01 Hz, 1H), 8.11 (dd, J ) 6.5, 1.3 Hz, 1H), 8.05 (dd, J
) 7.03, 1.29 Hz, 1H), 7.91 (dd, J ) 7.17, 1.12 Hz, 1H), 7.75
(m, 4H); 13C NMR (DMSO-d6) δ 142.91, 137.48, 136.63, 135.02,
134.34, 133.63, 131.81, 131.56, 130.72, 129.01, 127.68, 123.98,
123.57.
C
22H17FNO4S (M + 1) 410.0863, found 410.0869; FT-IR (KBr
pellet cm-1) 1781, 1655, 1328, 1148.
N-(Meth yl-(p-flu or oben zoyl)for m yl)-9-th ioxa n th en yl-
10,10-d ioxid e im in e (6b): Rf ) 0.12 (10% ether/CH2Cl2); mp
154-155 °C; 1H NMR (CDCl3) δ 8.23 (dd, J ) 7.77, 1.4 Hz
1H), 8.11 (dd, J ) 7.52, 1.0 Hz, 1H), 8.05 (dd, J ) 7.52, 1.0
Hz, 1H), 7.68 (tdd, J ) 7.58, 2.81, 1.4 Hz, 2H), 7.69-7.61 (m,
2H), 7.59-7.55 (m, 2H), 7.48 (dd, J ) 7.58, 0.9 Hz, 1H), 7.12-
7.06 (m, 2H), 5.82 (s, 1H), 3.76 (s, 3H). 13C NMR (CDCl3) δ
171.00, 164.06, 161.60, 157.52, 139.98, 139.22, 137.80, 134.20,
134.17, 133.31, 132.33, 131.12, 131.02, 130.36, 129.39, 129.31,
128.44, 128.05, 125.38, 123.10, 116.21, 116.00, 69.31, 53.15;
HRMS (FAB+) calcd for C22H17FNO4S (M + 1) 410.0863, found
410.0886; FT-IR (KBr pellet cm-1) 1763, 1650, 1310, 1123.
9-Th ioxa n th on oxim e O-a ceta te 10,10-d ioxid e (16):
1
yield quantitative. Rf ) 0.58 (10% ether in CH2Cl2); H NMR
(CDCl3) δ 8.22 (m, 2H), 8.12 (m, 1H), 8.05 (m, 1H), 7.70 (m,
4H), 2.32 (s, 3H); 13C NMR (CDCl3) δ 167.94, 150.16, 138.56,
137.60, 133.40, 133.12, 132.51, 131.63, 131.16, 131.07, 128.54,
128.10, 124.52, 123.89, 19.90.
9-Am in oth ioxa n th en e 10,10-d ioxid e (17): yield 62%; mp
1
165-168 °C; H NMR (CDCl3) δ 8.10 (dd, J ) 7.69, 1.34 Hz,
2H), 7.75 (d, J ) 7.71 Hz, 2H), 7.59 (m, 2H), 7.50 (td, 2H),
5.29 (s, 1H), 2.13 (s, 2H, -NH2); 13C NMR (CDCl3) δ 143.44,
137.73, 133.28, 128.28, 127.14, 124.48, 52.70; FT-IR (KBr pellet
cm-1) 3409, 3327, 1278, 1139.
N-(9-Th ioxa n t h en yl 10,10-d ioxid e)m et h yl(p -m et h yl-
ben zoyl)for m ylim in e (7a ): yield 64%; Rf ) 0.28 (10% ether/
CH2Cl2); yield 60%; mp 165-167 °C; Rf ) 0.2 (10% ether in
2-Nitr o-N,N-d ieth ylben za m id e (18): yield 83%; 1H NMR
(CDCl3) δ 8.19 (dd, J ) 8.3, 1.2 Hz, 1H), 7.71 (m, 1H), 7.57
(m, 1H), 7.40 (dd, J ) 7.6, 1.1 Hz, 1H), 3.60 (br s, 2H), 3.13 (q,
J ) 7.14 Hz, 2H), 1.30 (t, J ) 7.13 Hz, 3H), 1.07 (t, J ) 7.14
Hz, 3H); 13C NMR (CDCl3) δ 167.07, 144.85, 134.30, 133.51,
129.45, 127.88, 124.64, 42.64, 38.82, 13.47, 11.88.
2-Am in o-N,N-d ieth ylben za m id e (19): yield quantitative;
1H NMR (CDCl3) δ 7.14 (m, 1H), 7.06 (dd, J ) 8.0, 1.5 Hz,
1H), 6.71 (m, 2H), 4.13 (s, 2H, -NH2), 3.42 (br. s, 4H), 1.23
1
CH2Cl2); H NMR (CDCl3) δ 8.14 (dd, J ) 7.0, 2.57 Hz, 2H),
7.88 (dd, J ) 7.0, 0.2 Hz, 2H), 7.54-7.51 (m, 4H), 7.47-7.44
(m, 2H), 7.35 (d, J ) 7.9 Hz, 2H), 5.91 (s, 1H), 3.88 (s, 3H),
2.47 (s, 3H); 13C NMR (CDCl3) δ 165.48, 165.15, 143.36, 140.37,
136.61, 132.65, 130.74, 129.94, 128.21, 128.14, 126.72, 124.08,
60.82, 52.72, 21.84; HRMS (FAB+) calcd for C23H20NO4S (M
+ 1) 406.1114, found 406.1111; FT-IR (KBr pellet cm-1) 1726,
1541, 1309, 1161.

3594 J . Org. Chem., Vol. 67, No. 11, 2002
Hjelmencrantz and Berg
(m, 6H); 13C NMR (CDCl3) δ 170.56, 144.68, 129.99, 126.79,
121.57, 117.46, 116.48, 68.86, 46.31, 24.13, 23.30.
then-9-one 10,10-dioxide (1.19 g, 4.14 mmol) in dry CH2Cl2 (140
mL), cooled on an ice bath, was added TiCl4 (0.897 mL, 8.28
mmol, 2.0 equiv). The mixture was stirred at room tempera-
ture for 10 min, before the dark purple heterogeneous mixture
was bubbled through with gaseous NH3 for 45 min, after which
time CH2Cl2 (25 mL) was added. The mixture was filtered and
evaporated to dryness. The raw material was purified on a
chromatotron (Rf ) 0.27, 10% ether/CH2Cl2). Approximately
30% of the raw material was isolated as the starting ketone:
2-(N,N-Dim eth yla m in o)-N,N-d ieth ylben za m id e (20). To
a solution of 2-amino-N,N-diethylbenzamide (0.032 mol, 6.09
g) in dry THF (50 mL) was added NaH (0.080 mol, 3.2 g, 60%
dispersion in mineral oil). MeI (0.080 mol, 4.98 mL) was added,
and the reaction mixture was refluxed for 17 h. The reaction
mixture was then allowed to cool to room temperature and
thereafter carefully poured into a saturated NH4Cl solution
(ca. 150 mL). The organic phase was separated, and the
aqueous phase was extracted with EtOAc (3 × 50 mL). The
combined organic layers were dried over Na2SO4 and evapo-
rated in vacuo to afford a brown oil. The raw material was
purified by vacuum distillation and retrieved as a pale yellow
1
yield 71%; mp 150-152 °C deg; H NMR (CDCl3) δ 11.78 (br
s, 1H, dNH), 8.15 (br. d, J ) 6.7 Hz, 1H), 8.10 (m, 1H), 7.77
(dd, J ) 7.7, 1.0 Hz, 1H), 7.70 (m, 2H), 7.53 (t, J ) 7.76 Hz,
1H), 7.32 (dd, J ) 7.0, 1.0 Hz, 1H), 2.78 (s, 6H); 13C NMR
(CDCl3) δ 162.84, 152.20, 139.54, 137.40, 136.66, 133.17,
131.46, 130.84, 127.49, 123.04, 122.91, 122.84, 116.37, 43.70;
HRMS (FAB+) calcd for C15H15N2O2S (M + 1) 287.0855, found
287.0852; FT-IR (KBr pellet cm-1) 3224, 1602, 1296, 1163.
1-(N,N-Dim et h yla m in o)t h ioxa n t h en -9-a m in e 10,10-
Dioxid e (24). NaBH4 (7.17 mmol, 277 mg, 4 equiv) was
dissolved in MeOH (7 mL). 1-(N,N-Dimethylamino)thioxan-
then-9-imine 10,10-dioxide (1.79 mmol, 513 mg, 1 equiv)
dissolved in CH2Cl2 (7 mL) was added to the NaBH4/MeOH
solution. The reaction mixture was stirred at room tempera-
ture for 17 h. The solvent was evaporated in vacuo, the residue
was dissolved in CH2Cl2, and the inorganic remains were
filtered off. The solvent was removed, and the crude product
was purified on a chromatotron: Rf ) 0.12 (10% ether in CH2-
Cl2); yield 50%; mp 153-155 °C; 1H NMR (CDCl3) δ 8.06 (d, J
) 7.82 Hz, 1H), 7.80 (dd, J ) 7.96, 1.1 Hz, 1H), 7.55 (m, 2H),
7.46 (m, 2H), 7.35 (dd, J ) 8.0, 1.2 Hz, 1H), 5.52 (s, 1H, 9-H),
2.89 (s, 6H), 2.27 (br. s, 2H, -NH2); 13C NMR (CDCl3) δ 153.13,
143.84, 139.27, 137.61, 136.10, 133.06, 128.71, 128.62, 127.88,
124.50, 124.17, 118.88, 48.55, 45.67; HRMS (FAB+) calcd for
oil: yield quantitative; n21 ) 1.5394; bp 110 °C (0.05 mbar);
D
1H NMR (CDCl3) δ 7.27 (m, 1H), 7.18 (m, 1H), 6.93 (m, 2H),
3.80 (m, 1H), 3.33 (m, 1H), 3.20 (m, 1H), 3.09 (m, 1H), 2.80 (s,
6H), 1.24 (t, J ) 7.13 Hz, 3H), 1.10 (t, J ) 7.13 Hz, 3H); 13C
NMR (CDCl3) δ 171.75, 149.70, 130.17, 129.88, 128.81, 121.08,
117.37, 43.92, 43.09, 39.16, 14.12, 12.98. HRMS (FAB+) calcd
for C13H21N2O (M + 1) 221.1655, found 221.1652; FT-IR (NaCl
tablet cm-1) 1621, 1456, 1287.
2-(N,N-Dim eth yla m in o)-6-(p h en ylsu lfon yl)-N,N-d ieth -
ylben za m id e (21). Dry and distilled TMEDA (21.78 mmol,
3.13 mL) was dissolved in dry THF (140 mL), and the mixture
was cooled to -78 °C. s-BuLi (21.78 mmol, 1.2 M, 18.15 mL)
was added, and the solution was stirred for 10 min. 2-(N,N-
Dimethylamino)-N,N-diethylbenzamide (18.15 mmol, 4.0 g,
4.16 mL) dissolved in dry THF (40 mL) was added droppwise.
After completion of addition, the reaction mixture was stirred
for 30 min at -78 °C. Benzenesulfonyl fluoride (24.94 mmol,
3.0 mL) dissolved in dry THF (40 mL) was added droppwise
to the lithiated diethyl amide. After completion of addition,
the cooling bath was removed and the reaction mixture was
stirred at room temperature for 2 h, after which time it was
quenched with saturated NH4Cl solution. The organic phase
was separated, and the aqueous phase was extracted twice
with CH2Cl2. The combined organic phases were dried over
Na2SO4 and filtered, and the solvent was removed in vacuo.
The crude product was purified via flash chromatography Rf
) 0.36 (10% ether in CH2Cl2) and retrieved as a viscous
C
15H17N2O2S (M + 1) 289.1012, found 289.1024; FT-IR (KBr
pellet cm-1) 3388, 3352, 1287, 1154.
N-(9-(1-N,N-Dim eth yla m in o)th ioxa n th en e 10,10-d iox-
id e)cr otylim in e (25). To a solution of 1-(N,N-dimethylami-
no)thioxanthen-9-amine 10,10-dioxide (0.163 mmol, 47 mg) in
dry CH2Cl2 (2 mL) was added crotonaldehyde (0.163 mmol,
13.4 µL). Dried MgSO4 (500 mg) was added, and the mixture
was stirred at room temperature for 14 h. The drying agent
was filtered off and the solvent evaporated in vacuo. The
semisolid residue was purified by chromatography (Rf ) 0.54,
20% ether in CH2Cl2). The pure product was retrieved as pale
yellow crystals: yield 81%; mp 181-183 °C; 1H NMR (CDCl3)
δ 8.10 (dd, J ) 7.4, 0.8 Hz, 1H), 7.89 (dd, J ) 7.3, 1.4 Hz, 1H),
7.75 (dd, J ) 8.0, 1.4 Hz, 1H), 7.58-7.47 (m, 4H), 7.42 (dd, J
) 8.1, 1.3 Hz, 1H), 6.30 (s, 1H), 6.27-6.09 (m, 2H), 2.76 (s,
6H), 1.80 (d, J ) 6.1 Hz, 3H); 13C NMR (CDCl3) δ 163.63,
153.97, 142.23, 139.70, 139.09, 138.28, 133.61, 132.94, 132.35,
130.43, 129.50, 128.74, 125.42, 124.59, 120.09, 62.94, 46.09,
18.62. HRMS (FAB+) calcd for C19H21N2O2S (M + 1) 341.1325,
found 341.1320; FT-IR (KBr pellet cm-1) 1684, 1310, 1158.
Meth yl p-flu or op h en ylp yr u va te (26): yield 67%; 1H
NMR (CDCl3) δ 8.09 (dd, J ) 9.1, 5.3 Hz, 2H), 7.19 (dd, J )
8.0, 6.1 Hz, 2H), 3.97 (s, 3H).
greenish liquid: yield 92%; n21 ) 1.5760; 1H NMR (CDCl3) δ
D
8.08 (m, 2H), 7.74 (dd, J ) 8.0, 1.07 Hz, 1H), 7.47 (m, 3H),
7.39 (t, J ) 8.02 Hz, 1H), 7.21 (dd, J ) 8.0, 1.0 Hz, 1H), 3.69
(m, 1H), 3.51 (m, 1H), 3.07 (m, 2H), 2.73 (s, 6H), 1.32 (t, J )
7.15 Hz, 3H), 1.10 (t, J ) 7.18 Hz, 3H); 13C NMR (CDCl3) δ
167.41, 151.83, 142.28, 139.61, 133.43, 132.03, 129.83, 129.17,
128.85, 124.59, 123.93, 44.88, 43.61, 39.37, 13.41, 12.73; HRMS
(FAB+) calcd for
C19H25N2O3S (M + 1) 361.1588 found
361.1597; FT-IR (NaCl tablet cm-1) 1625, 1323, 1163, 1451,
1278.
1-(N,N-Dim eth yla m in o)th ioxa n th en -9-on e 10,10-Diox-
id e (22). To a cooled solution (0 °C) of LDA, prepared in situ
from diisopropylamine (46 mmol, 6.46 mL, dry and distilled)
and BuLi (33.29 mmol) in 140 mL of dry THF, was added
dropwise a solution of 2-(N,N-dimethylamino)-6-(phenylsulfo-
nyl)-N,N-diethylbenzamide (16.64 mmol, 6.0 g) in 100 mL of
dry THF. After completion of the addition, the reaction mixture
was stirred at 0 °C for 30 min. The reaction was thereafter
quenched with saturated NH4Cl (150 mL), and the organic
phase was separated. The aqueous phase was extracted with
ether (2 × 100 mL). The combined organic layers were washed
with saturated NaCl (100 mL) and dried over Na2SO4, and
the solvent was evaporated in vacuo. Purification was made
by passing the raw material through a silica column using
CHCl3 as eluent: Rf ) 0.48 (10% ether in CH2Cl2); yield 84%;
mp 170-172 °C; 1H NMR (CD2Cl2) δ 8.14 (m, 1H), 8.06 (m,
1H), 7.77 (m, 2H), 7.58 (m, 1H), 7.51 (dd, J ) 7.5, 1.2 Hz, 1H),
7.20 (dd, J ) 9.00, 1.16 Hz, 1H), 2.99 (s, 6H); 13C NMR (CDCl3)
δ 177.82, 153.90, 143.32, 139.10, 134.37, 133.65, 133.58,
133.28, 129.03, 123.26, 121.65, 116.92, 113.55, 44.67; HRMS
(EI+) calcd for C15H13NO3S (M) 287.0613, found 287.0614; FT-
IR (KBr pellet cm-1) 1643, 1296, 1268, 1154.
Meth yl p-tolu ylpyr u vate (27): yield 72%; 1H NMR (CDCl3)
δ 7.91 (d, J ) 8.2 Hz, 2H), 7.31 (d, J ) 9.2 Hz, 2H), 3.97 (s,
3H), 2.44 (s, 3H).
1
Meth yl p-m eth oxyp h en ylp yr u va te (28): yield 75%; H
NMR (CDCl3) δ 8.04-7.99 (m, 2H), 7.00-6.95 (m, 2H), 3.96
(s, 3H), 3.90 (s, 3H).
Ack n ow led gm en t. We are grateful to Dr. Per-Ola
Norrby, DTU, Lyngby, Denmark, for help with analysis
of the isotope effects and to the Swedish Natural Science
Research Council and the Royal Physiographic Society
for financial support.
Su p p or tin g In for m a tion Ava ila ble: Copies of 1H and 13C
NMR spectra for new compounds and Figure S1 showing
component evolution in the isotope effect experiment. This
material is available free of charge via the Internet at
http://pubs.acs.org.
1-(N,N-Dim eth yla m in o)th ioxa n th en -9-im in e 10,10-Di-
oxid e (23). To a solution of 1-(N,N-dimethylamino)thioxan-
J O0106748