Oxidative cyclisation of N,N-dialkylcatechol amines to heterocyclic betaines via o-quinones: Synthetic, pulse radiolytic and enzyme studies

Article abstract of DOI:10.1039/b006860h

Oxidation of N,N-dialkyldopamines? by either dianisyltellurium oxide or tyrosinase gives 2,3-dihydro-1H-indolium-5-olates which are formed by cyclisation of an intermediate o-quinone. The kinetics of formation and cyclisation of the N,N-dimethyl-o-quinone

Full text of DOI:10.1039/b006860h

View Article Online / Journal Homepage / Table of Contents for this issue
Oxidative cyclisation of N,N-dialkylcatechol amines to hetero-
cyclic betaines via o-quinones: synthetic, pulse radiolytic and
enzyme studies
1
a
b
b
c
John Clews, Christopher J. Cooksey, Peter J. Garratt, Edward J. Land, †
a
d
Christopher A. Ramsden* and Patrick A. Riley
a
School of Chemistry and Physics, Keele University, Keele, Staffordshire, UK ST5 5BG
Department of Chemistry, Christopher Ingold Laboratories, UCL, 20 Gordon Street,
London, UK WC1H 0AJ
CRC Drug Development Group, Paterson Institute for Cancer Research, Christie Hospital
NHS Trust, Manchester, UK M20 4BX
b
c
d
Windeyer Institute, UCL Medical School, 46 Cleveland Street, London, UK W1P 6DB
Received 22nd August 2000, Accepted 27th September 2000
First published as an Advance Article on the web 22nd November 2000
Oxidation of N,N-dialkyldopamines‡ by either dianisyltellurium oxide or tyrosinase gives 2,3-dihydro-1H-indolium-
-olates which are formed by cyclisation of an intermediate o-quinone. The kinetics of formation and cyclisation
5
of the N,N-dimethyl-o-quinone have been studied using pulse radiolysis. The indolium-5-olates do not activate
met-tyrosinase and these results support a mechanism of tyrosinase oxidation of phenols to o-quinones in which the
o-quinone is formed in a single step and not via an intermediate catechol. Similar chemical and enzymatic oxidation
of a higher homologue gives an analogous 1,2,3,4-tetrahydroquinolinium-6-olate. Pulse radiolysis studies show that
this product is formed via a spiro intermediate and not by direct cyclisation to form the six-membered quinolinium
ring. The novel betaines described have been fully characterised and converted to their dimethoxy iodide salts. In
a preliminary investigation of potential anti-cancer pro-drugs, amide derivatives of dopamine do not cyclise when
oxidised to the o-quinone but cyclisation of an N-benzoylmethyl derivative to the corresponding betaine was
observed. This betaine then appears to equilibrate with an N-ylide which, in contrast to the betaine, is a substrate
for tyrosinase.
Introduction
portionation and not by direct enzyme oxidation. This revised
mechanism satisfactorily accounts for the induction or lag
period observed during tyrosinase catalysed oxidation. In
particular, the induction period arises because the enzyme
occurs largely in the inactive met form in which the two copper
Using selected phenolic and catecholic substrates we have
recently demonstrated that the enzyme tyrosinase [EC
1–3
1
.14.18.1] oxidises phenols to o-quinones in one step and not
4
via intermediate catechols as is widely claimed. Thus, in the
tyrosinase mediated oxidation of tyrosine 1 (Scheme 1) the
dopaquinone 2 is formed directly in one step and then cyclises
to cyclodopa 3. A redox reaction subsequently occurs between
the products 2 and 3 giving dopa 4 and dopachrome 5, which
then undergoes further reactions leading to melanin forma-
II
atoms at the active site are in the Cu oxidation state and
6
cannot bind dioxygen. Reduction by a catechol converts the
I
enzyme to the active deoxy form [Cu ] together with formation
of an o-quinone [eqn. (1)]. Deoxytyrosinase then binds dioxy-
gen to form oxytyrosinase [eqn. (2)]. Using phenolic substrates,
oxidation is initially very slow due to the small amount of
enzyme in the active deoxy form and consequently indirect
formation of activating catechol via redox disproportionation
5
tion. An important feature of this mechanism (Scheme 1) is
that the dopa 4 is formed indirectly by non-enzymatic dispro-
(
Scheme 1) is also initially slow. As more catechol is formed
†
‡
E. J. L. is an Honorary Senior Research Fellow at Keele University.
The IUPAC name for dopamine is 4-(2-aminoethyl)pyrocatechol.
by the indirect non-enzymatic route the rate of activation
accelerates accounting for the observed kinetics. It is important
Scheme 1 Reagents: i, O ϩ deoxytyrosinase; ii, met-tyrosinase.
2
4
306
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315
This journal is © The Royal Society of Chemistry 2000
DOI: 10.1039/b006860h

View Article Online
to note that although catechols are not formed directly by
tyrosinase (1 ×→ 4) they are substrates for the enzyme [eqns. (1)
and (3)] and like phenols [eqn. (4)] are effectively oxidised to
o-quinones by the oxy form of the enzyme (e.g. 4→2).
in a similar manner. In fact, during the course of our work,
Jagoe and co-workers in the report of a study of the binding
7
of oxidised catechols to cysteine residues in Src family SH2
domains assumed that o-quinone 7b is unlikely to cyclise to
a 2,3-dihydro-1H-indole derivative although their observation
that compound 8b, like dopamine 8d, was inactive in their
assay is entirely consistent with cyclisation. However, our
search of the literature showed that this type of cyclisation
II
Met[Cu ]tyrosinase ϩ catechol →
I
ϩ
deoxy[Cu ]tyrosinase ϩ o-quinone ϩ 2H
(1)
8
I
II
had previously been encountered by Robinson and Sugasawa,
Deoxy[Cu ]tyrosinase ϩ O → oxy[Cu ]tyrosinase (2)
2
9
and independently by Schöpf and Thierfelder, during studies
II
ϩ
of the biogenesis of morphine alkaloids, but simple systems
were not investigated.
Oxy[Cu ]tyrosinase ϩ catechol ϩ 2H →
II
met[Cu ]tyrosinase ϩ o-quinone ϩ 2H O (3)
2
II
Preparation of indolium-5-olates 10
Oxy[Cu ]tyrosinase ϩ phenol →
I
Synthesis of precursor catechols 8 from 2-(3,4-dimethoxy-
phenyl)ethylamine was achieved using standard methods and
in each case the 3,4-dimethoxyphenyl function was converted
to the catechol using 48% HBr and the resulting tertiary amine
hydrobromide salt converted to the free base using aqueous
sodium bicarbonate. There are many reagents for oxidising
catechols to o-quinones but our interest in hypervalent
deoxy[Cu ]tyrosinase ϩ o-quinone ϩ H O (4)
2
As part of our mechanistic study of tyrosinase we have shown
that N,N-di-n-propyltyramine§ 6b is not oxidised by unactivated
tyrosinase. This provides strong evidence that direct enzymatic
catechol formation does not occur (i.e. 6 ×→ 8). Using pre-
activated tyrosinase, oxidation of the same substrate 6b resulted
in rapid formation of a product that we proposed was the
indolium-5-olate 10b, generated by cyclisation of the inter-
mediate o-quinone 7b and aromatisation of the resulting cyclo-
hexadienone 9b (Scheme 2). This product 10b is not a catechol
1
10
reagents directed our attention to dianisyltellurium oxide
11
(
An Te᎐O)(DAT) which has been shown to be selective for
2
᎐
catechol oxidation in the presence of a wide variety of other
functional groups including amines.
A preliminary study of the oxidation of N,N-dimethyl-
dopamine 8a using one equivalent of DAT in CDCl –MeOH
3
1
solution was monitored using H NMR spectroscopy and
the results were extremely promising. All the proton signals
associated with the reactants rapidly disappeared and were
replaced by signals corresponding to quantitative formation
of the betaine 10a and dianisyltellurium. When the reaction
was repeated on a preparative scale the betaine 10a was iso-
lated as a reddish brown solid in 91% yield. Oxidation of
the amines 8b,c similarly gave the betaines 10b,c in high yield.
NMR spectroscopy fully supported the assignment of the 2,3-
dihydro-1H-indolium-5-olate structure 10 to these products.
Typically, the betaine 10b showed only two aromatic protons
(
δ 6.63 and 6.67) consistent with cyclisation together with a
pair of triplets (J 7.5 Hz) at δ 3.07 and 3.99 corresponding to
the dihydroindole ring methylene groups: the methylene group
at particularly low chemical shift is adjacent to the quaternary
nitrogen atom. Equally significant are the chemical shifts and
non-equivalence of the methylene protons of the N-n-propyl
substituents. Thus the non-equivalent methylene protons on
the α-carbon atoms appear at low field as a pair of doublets
of triplets (δ 3.40 and 3.55) and the protons on the β-carbon
atoms appear as a pair of multiplets at higher field (δ 1.40
and 1.65). We assume that in each case the protons at lower
field are those pointing towards the aromatic ring and are de-
shielded by the ring current. A COSY spectrum fully supported
Scheme 2 Reagents: i, O ϩ deoxytyrosinase; ii, met-tyrosinase; iii,
2
An TeO; iv, MeI–K CO .
2
2
3
and we concluded that it cannot lead to redox activation of
1
tyrosinase [eqn. (1)]. Material isolated from the pre-activated
1
enzyme reaction media had a H NMR spectrum consistent
with the proposed betaine structure 10b and we required at
this stage authentic samples of this novel indolium-5-olate
and related products for structure confirmation and studies
of their properties. In this paper we describe the preparation
the proposed proton–proton coupling in the betaine 10b and
13
additional structural confirmation was provided by the
C
2
of examples of novel heterocyclic betaines, including com-
NMR spectrum. The NMR spectra of the betaines 10a,b were
identical with those of the tyrosinase oxidation products pro-
viding confirmation of the structures of the enzyme products.
As might be expected, the UV spectra of the betaines 10
show a small pH dependence. At pH 7.4 in phosphate buffer all
the derivatives 10a–c show an absorption at 290 nm with a
shoulder at ca. 310 nm. At pH 6.5 this shoulder is absent. We
attribute this change at lower pH to complete formation of
salts, which have spectra identical to those of the iodides 11. In
aqueous solution the betaines 10 are clearly in equilibrium with
these salts. We have represented the betaines 10 as the 5-olates
on the understanding that in these tautomers resonance places
pound 10b, by cyclisation of chemically generated o-quinones,
together with a pulse radiolysis study of the mechanisms and
kinetics of these reactions. We also describe the investigation of
three dopamine derivatives designed to explore the possibility
of using tyrosinase mediated betaine formation in vivo as a
method of selectively activating anti-cancer pro-drugs.
Results and discussion
Although intramolecular cyclisations of primary amino groups
onto o-quinones to form 2,3-dihydro-5,6-dihydroxy-1H-indole
5
derivatives are well known (e.g. 2→3, Scheme 1), a priori
negative charge closest to the quaternary centre. This assump-1
it was not obvious that tertiary amines would readily cyclise
tion is supported by AM1 semi-empirical MO calculations.
However, we cannot eliminate the possibility that these tauto-
mers are in equilibrium with the 6-olates as well as their salts.
§
The IUPAC name for tyramine is 2-(4-hydroxyphenyl)ethylamine.
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315
4307

View Article Online
The betaines 10 are hygroscopic and difficult to recrystallise
but further characterisation was achieved by conversion to the
5
,6-dimethoxy iodides 11 using methyl iodide in the presence
of potassium carbonate. In this way analytically pure samples
of the colourless iodides 11a–c were obtained after recrystal-
lisation. Attempts to prepare the monomethyl iodides were
unsuccessful and led to mixtures of mono- and dimethylated
products that were difficult to separate.
The availability of the synthetic betaines 10 has enabled us to
demonstrate experimentally that these betaines do not activate
1
tyrosinase, in accord with our mechanistic conclusions.
We have previously shown that N,N-dimethyltyramine 6a is
oxidised at a negligible rate by tyrosinase unless the enzyme
is activated by a catechol, which is presumptive evidence that
the formation of the corresponding cyclic betaine 10a is unable
1
to abolish the lag-period of the enzyme. However, to demon-
strate unequivocally the inability of the betaines 10 to activate
tyrosinase we have employed an oximetric assay using N,N-
dimethyltyramine 6a as substrate. In these experiments both
pre-incubation with or addition to the incubation mixture of
up to 125 µM of synthetic betaine 10a failed to modify the rate
of oxygen uptake. In contrast, small amounts (10 µM) of the
corresponding catechol 8a produced an immediate increase in
oxygen utilization attributable to accelerated oxidation of the
N,N-dimethyltyramine 6a. Since the activation of tyrosinase is
dependent on recruitment of met-enzyme by reduction of active
site copper atoms, we conclude that the lack of enzyme acti-
vation by the authentic betaine 10a is a consequence of the
significantly different redox properties of the betaine compared
to the corresponding catechol 8a.
Fig. 1 Absorption changes at various times after pulse radiolysis
Ϫ3
of an N O-saturated aqueous solution of 0.58 × 10 M 8a containing
2
Ϫ2
0
.1 M KBr and 10 M phosphate buffer, pH 7.0; dose ~14 Gy. Inset:
time profiles of transmission changes.
5
,6
Mechanism and kinetics of indolium-5-olate formation
1
The intermediate o-quinones 7 could not be detected by H
NMR spectroscopy because they cyclise too rapidly. We have
used pulse radiolysis to identify o-quinone intermediates, e.g.
7
a, and study the mechanism and kinetics of their cyclisations.
For these studies the o-quinones 7 were generated in situ by
one-electron oxidation of the catechol to the semiquinone. The
12,13
semiquinone then disproportionates to give the o-quinone.
One-electron oxidation of the catechol 8a was carried out by
Fig. 2 Absorption changes at various times after pulse radiolysis
pulsed irradiation of an N O-saturated solution in the presence
2
Ϫ3
of an N O-saturated aqueous solution of 0.82 × 10 M 8a containing
2
of KBr buffered to pH 7.0. Fig. 1 shows the changes in absorp-
tion spectrum 45, 315, 950 µs and 2.2 ms after the pulse. The
initial spectrum at 45 µs is due to formation of the semiquinone
Ϫ2
0
.1 M KBr and 10 M phosphate buffer, pH 6.2; dose ~30 Gy. Inset:
time profiles of transmission changes.
1
2 [eqn. (5)].
the above experiment was repeated at pH 6.2. Fig. 2 shows the
changes in absorption spectrum recorded 2.8, 5.9 and 16 ms
after pulse radiolysis of compound 8a and KBr in phosphate
buffer at pH 6.2. The spectrum with wavelength maximum
(
5)
around λ
390 nm is attributed to the o-quinone 7a (ε 1500
max
Ϫ1
3
Ϫ1
dm mol cm ) which decayed unimolecularly, with a rate
constant of 300 s , into a stable species absorbing increasingly
Ϫ1
The spectrum obtained when most of semiquinone 12 had
below 300 nm.
decayed (2.2 ms) showed little evidence for formation of
the expected o-quinone 7a (λmax 400 nm). However, a com-
parison of the decay of the semiquinone 12 at its peak
It was not practicable to study the spectrum of the stable
product below 300 nm using millimolar solutions of compound
8a because of the strong parent absorption in this region.
ؒ
(
λmax 310 nm) with the transmission-against-time curve
However, since N3 tends to react at least an order of magnitude
Ϫ 14
observed at 400 nm over the same time scale (see inset to Fig. 1)
provides evidence of a delayed growth at 400 nm over ~0.25 ms
after the pulse. This is probably due to a short-lived o-quinone
faster with catechols than Br2
,
it was still possible with
᝽
Ϫ4
1 × 10 M solutions of compound 8a to obtain practically
complete scavenging of the alternative one-electron oxidant
ؒ
i.e. 7a with a lifetime of a few milliseconds. The increase in the
N3 and make spectroscopic observations inside the region of
Ϫ
rate of decay of Br2 observed at λmax 360 nm on addition of
parent 8a absorption down to 250 nm. Accordingly, Fig. 3
shows the change in absorption spectrum 1.3 and 14 ms after
pulse radiolysis of an N O-saturated solution of compound 8a
᝽
the catecholamine 8a led to measurement of a rate constant
8
Ϫ1 Ϫ1
of 1.3 × 10 M
s
for the formation of the semiquinone 12
2
(
eqn. (5)). The semiquinone absorption with a maximum at
in the presence of NaN , buffered to pH 6.2 with phosphate.
3
3
Ϫ1
Ϫ1
λmax 310 nm (ε 11 900 dm mol cm , based on thiocyanate
dosimetry) decayed by disproportionation with a rate constant
(
Care was taken to keep the azide concentration as low as
possible in order to minimise the likelihood of nucleophilic
8
Ϫ1 Ϫ1
Ϫ
12
2k) of 7.5 × 10 M s .
addition of N3 to the o-quinone formed. The spectrum
measured at 1.3 ms is assigned largely to the o-quinone 7a,
together with a small amount of residual semiquinone 12, and
Since o-quinones derived from catecholamines can become
12
longer-lived on changing from neutral to acid conditions,
4
308
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315

View Article Online
product generated pulse radiolytically (filled circles) matches
rather well the spectrum of the isolated and fully characterised
indolium-5-olate 10a (full line).
Preparation of quinolinium-6-olate 15
The clean preparation of the indolium-5-olates 10 prompted
us to investigate the preparation of quinolinium-6-olates,
e.g. 15, via an analogous oxidative cyclisation of the higher
homologues. In this case we recognised that an alternative 5-
exo-trig cyclisation of the intermediate o-quinone 14 to give a
spirobetaine 16 might be favoured compared to six-membered
ring formation (14→15)(Scheme 3). However, the cyclisation
Fig. 3 Absorption changes 1.3 and 14 ms after pulse radiolysis of
Ϫ4
an N O-saturated aqueous solution of 1.0 × 10 M 8a containing
2
Ϫ3
Ϫ2
3
× 10 M NaN and 10 M phosphate buffer, pH 6.2; dose ~33 Gy.
3
Inset: time profiles of transmission changes.
Scheme 3 Reagents: i, An TeO; ii, MeI–K CO .
2
2
3
product 16 does not enjoy aromatic stabilisation and can be
expected to equilibrate with the o-quinone precursor 14 or
rearrange (16→17).
N,N-Diethyl-3-(3,4-dihydroxyphenyl)propylamine 13 upon
treatment with one equivalent of DAT gave the 7-hydroxy-
1
,2,3,4-tetrahydroquinolinium-6-olate 15 quantitatively as
1
monitored by H NMR. There was no NMR evidence of the
formation of the spirobetaine 16 during the reaction. On a
preparative scale the betaine 15 was isolated in 84% yield as
a reddish brown solid which was readily converted into the
crystalline dimethoxy iodide 18, mp 230–231 ЊC, in 90% yield.
The structures 15 and 18 were fully supported by their spectro-
scopic properties: significantly only two aromatic protons are
Fig. 4 Absolute absorption spectrum, measured at pH 7.4, of the final
product of oxidation of the catechol 8a: 2,3-dihydro-1,1-dimethyl-6-
hydroxy-1H-indolium-5-olate 10a. (The filled circles are from the pulse
radiolysis experiments and the full line from measurements on the
1
observed in the H NMR spectra.
Mechanism and kinetics of quinolinium-6-olate formation
betaine prepared by An Te᎐O oxidation).
2
᎐
When the formation and cyclisation of the o-quinone 14 was
studied using pulse radiolysis an interesting difference to the
cyclisation of the dopamine derivatives 8 was observed. In
particular, an additional intermediate was detected during
the formation of the betaine 15 and we believe that this inter-
mediate is the spirobetaine 16.
that obtained at 14 ms is assigned to the final product. From the
rate of formation of semiquinone at λ
310 nm, a rate con-
max
9
Ϫ1 Ϫ1
stant of 5.8 × 10 M
s
was obtained for the reaction shown
in eqn. (6).
(
6)
One-electron oxidation of the catechol 13 was carried out by
pulsed irradiation of an N O-saturated solution in the presence
2
of NaN , buffered to pH 6.2 with phosphate. The initial
3
product, which had completely formed 40 µs after the pulse,
showed absorption maxima at λ 310 and 350 nm (Fig. 5). This
species is almost certainly the semiquinone radical 19 formed
via the reaction shown in eqn. (8).
It is suggested that the final product is the indolium-5-olate
0a, resulting from the intramolecular cyclisation of the o-
quinone 7a, formed via the disproportionation reaction shown
in eqn. (7), followed by reaction 7a→[9a]→10a (Scheme 2).
1
(
8)
(
7)
Based on thiocyanate dosimetry, and taking into account that
From the pseudo first-order build-up of semiquinone
9 Ϫ1 Ϫ1
the yield of the final product is half that of the initial yield
absorption at 310 nm, a rate constant of 4.1 × 10 M
s
ؒ
of N3 radicals, the absolute spectrum of the radiolytically
was obtained for the reaction shown in eqn. (8). Based on
produced final product, corrected for parent 8a depletion, was
calculated and is presented in Fig. 4, together with the spec-
trum of an authentic sample of the betaine 10a prepared by
thiocyanate dosimetry, a molar absorption coefficient of
12 100 M cm was obtained for the semiquinone 19 at 310
Ϫ1
Ϫ1
nm which decayed bimolecularly with a rate constant (2k) of
7
Ϫ1 Ϫ1
An Te᎐O oxidation. It can be seen that the spectrum of the
4.1 × 10 M s .
2
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315
4309

View Article Online
Fig. 7 Absorption changes 24 and 912 ms after pulse radiolysis of
Ϫ4
an N O-saturated aqueous solution of 1.0 × 10 M 13 containing
2
Ϫ3
Ϫ2
3
× 10 M NaN and 10 M phosphate buffer pH 6.2; dose ~30 Gy.
3
Inset: time profiles of transmission changes.
Fig. 5 Absorption changes at various times after pulse radiolysis of
Ϫ4
an N O-saturated aqueous solution of 0.96 × 10 M 13 containing
2
Ϫ3
Ϫ2
3
× 10 M NaN and 10 M phosphate buffer, pH 6.2; dose ~30 Gy.
Ϫ1
Ϫ
3
(k = 230 s ) using Br2 as oxidant is illustrated in Fig. 6. There
᝽
Inset: time profile of transmission changes.
is a window in the absorption of the catechol amine 13 in the
λ 250 nm region allowing the use of the necessary higher
concentrations (millimolar) [k(Br2 ϩ 13) = 2.4 × 10 M s ]
to ensure efficient Br2 scavenging.
The species causing the difference absorption maximum at
λ 250 nm decayed in turn, over several hundred milliseconds,
Ϫ
8
Ϫ1 Ϫ1
᝽
Ϫ
᝽
Ϫ1
unimolecularly, with a rate constant (k) of 7.1 s at pH 6.2
ؒ
(
using N3 initiation), to give a stable product with a difference
maximum at λ 290 nm (Fig. 7). This resembles the final product
obtained on oxidation of compound 8a and identified as the
indolium-5-olate 10a. Using the same assumptions as described
earlier for the radiolytic formation of 10a, the absolute spec-
trum of the final product obtained on oxidation of compound
1
3 was calculated, and is presented in Fig. 8a, together with the
spectrum of an authentic sample of the quinolinium-6-olate 15
at the same pH. Tyrosinase catalysed oxidation of compound
1
1
3 at physiological pH (7.4) also gave rise to a final product
with the same UV spectrum as that produced by pulse radio-
lytic oxidation (Fig. 8b).
The transient spectrum with a difference maximum at 250 nm
(
Fig. 7) is assigned to the spirobetaine 16. This maximum is
similar to those of several more stable compounds with similar
15–17
chromophores.
The lack of detection of an analogous
spirobetaine intermediate in the oxidation of the lower homo-
logue 8a is understandable, since 4-membered ring formation
is much less favoured. The species causing the difference
maximum at λ 250 nm is unlikely to be due to the intermediate
1
7 as the analogous species 9 were not detected on oxidation of
the dopamine derivatives 8. Again, on the basis of thiocyanate
dosimetry and that the spirobetaine yield is half that of the
ؒ
initial yield of N , the absolute absorption spectrum of the
3
spiro intermediate 16 at pH 6.2 was obtained and is illustrated
in Fig. 9. The weak absorption of intermediate 16, tailing into
the visible, possibly due to a forbidden n–π* transition, is con-
sistent with the presence of a cyclohexadienone component in
the assigned structure.
Fig. 6 Time profiles of transmission changes showing matching 1st
order decay at 400 nm, and build-up at 250 nm, following pulse radio-
Ϫ3
lysis of an N O-saturated aqueous solution of 1.0 × 10 M 13 con-
2
Ϫ2
taining 0.1 M KBr and 10 M phosphate buffer, pH 6.2; dose ~26 Gy.
The rapid first-order decay of the spirobetaine 16 leading
to the final product 15 suggests that this occurs via 1,2-
rearrangement to intermediate 17 followed by rapid aromatis-
ation. 1,2-Rearrangement of spirointermediates are well known:
an example is the rearrangement of carbocyclic spiro-species
formed by oxidative radical coupling during the biosynthesis of
some alkaloids. The novel 1,2-nitrogen rearrangement pro-
posed in Scheme 3 is therefore fully in accord with expectation.
As with the amine 8a, there was evidence for an o-quinone
ε 1300 M cm at 400 nm) with a half-life of a few milli-
Ϫ1
Ϫ1
(
seconds growing in as the semiquinone 19 decayed (see inset to
Fig. 5). The decay of the o-quinone in this case, however,
instead of leading directly to a stable betaine, gave rise instead
to a new transient absorption with a peak at 250 nm. The
matching first order decay at 400 nm and build-up at 250 nm
4
310
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315

View Article Online
limitation of its distribution essentially to invasive tissue pro-
vides in principle a mechanism for selectively activating anti-
cancer pro-drugs within melanoma cells by employing in vivo
oxidation of a phenol. In particular we envisage that if an anti-
cancer drug were incorporated into a dopamine derivative of
the general type 24 (or the corresponding tyramine derivative)
then tyrosinase mediated betaine formation (24→26) might
lead to rapid hydrolysis of the reactive quaternary amide 26
Fig.
8
Absolute absorption spectrum of 1,1-diethyl-7-hydroxy-
,2,3,4-tetrahydroquinolinium-6-olate 15, which is the final product of
1
oxidation of the catechol 13, measured at: (a) pH 6.2 (the filled circles
being from the pulse radiolysis experiments and the full line from meas-
18
with selective release of the drug within the cancer cell.
urements on the betaine prepared by An Te᎐O oxidation); and, (b)
pH 7.4 (the filled circles being from the pulse radiolysis experiments and
the dotted line from the tyrosinase-catalysed oxidation of catechol 13).
Although we recognise that amide nitrogens are poor nucleo-
philes we believed that the reactivity of the o-quinone inter-
mediates towards intramolecular cyclisation merited an initial
investigation of simple amides and related species. Accordingly
we prepared the novel amides 24a,b and the aminoketone
derivative 25c. These compounds were prepared from N-methyl-
2
2
-(3,4-dimethoxyphenyl)ethylamine using standard methods
and were fully characterised.
Fig. 9 Absolute spectrum, measured at pH 6.2, of the unstable
spirobetaine 16.
Consideration of the canonical forms 20 and 21 suggests that
this reaction is symmetry allowed. Furthermore, the transition
state 22 can be expected to be stabilised by a favourable inter-
action between the quinone LUMO and amine HOMO
The catechols 24a,b were rapidly oxidised by tyrosinase with
a concurrent rise in the absorbance at 400 nm, consistent
with formation of the corresponding o-quinone. Radiolytic
oxidation of catechols 24a,b also showed growth of a weak
absorbance at 400 nm which is evidence of o-quinone form-
ation as the corresponding semiquinones decay. This absorbance
was stable for ӷ 10 seconds. Chemical oxidation (DAT or
DDQ) similarly led to o-quinone formation without any
(
23). However, we cannot exclude the possibility that the final
product 15 is formed indirectly by ring opening of the
spirobetaine followed by recyclisation (i.e. 16→14→17→15).
1
evidence of cyclisation when monitored by H NMR. We
conclude that under all these conditions cyclisation to form
the N-acyl betaines 26 does not occur.
Investigation of potential substrates for tyrosinase
activated hydrolysis
Tyrosinase has limited occurrence in healthy human beings
but significant amounts occur in malignant melanoma. This
When the oxidation of the aminoketone 25c by tyrosinase
was investigated there was an initial rapid uptake of oxygen.
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315
4311

View Article Online
Scheme 4 Reagents: i, O ϩ deoxytyrosinase; ii, met-tyrosinase.
2
oxygen stoichiometry of this phase was also approximately
0
.5 moles of oxygen per mole of substrate, which is consistent
with the proposed transformation (Scheme 4).
Pulse radiolysis of the aminoketone 25c under oxidative con-
ditions at pHs 6 and 7 led to formation of the semiquinone
(
λmax 310 nm) which decayed via a second-order process. A
subsequent change in decay kinetics to first-order, detected
at 400 nm, is consistent with the presence of the very short-lived
o-quinone intermediate 28 (τ < 1 ms), in agreement with the
1
oximetry results. Furthermore, in a H NMR study, chemical
oxidation of the catechol 27 (R = OH) by one equivalent of
DDQ resulted in disappearance of the N-Me singlet (δ 2.4)
and appearance of a new singlet (δ 3.2). Significantly, after
oxidation there was a complete absence of aliphatic protons
1
in the region δ 2–3. Comparison with the H NMR spectra of
the authentic betaines 10 suggests that this observation is also
entirely consistent with the formation of the betaine 29 via the
short-lived o-quinone 28.
Fig. 10 Combined oximetry and spectrophotometric data of the
tyrosinase-catalysed oxidation of the catechol 25c showing (a) data for
the initial phase oxidation up to 180 s and (b) the secondary phase up
to 30 min.
The structure of the betaine 29 and its proposed oxidation
by tyrosinase merits further comment. We have previously
concluded that the indolium-5-olates 10 are not oxidised by
tyrosinase and do not take part in redox exchange reactions
with the o-quinone products of oxytyrosinase-catalysed oxid-
ation of monohydric phenols to generate catechol substrates
for met-tyrosinase. Since the mechanism of catechol oxidation by
oxytyrosinase involves met-tyrosinase in an intermediate stage
This was associated with a shift in the peak absorbance from
2
66 to 278 nm, and, although not associated with any peak
in the spectrum, a rise in absorbance at 320 nm (Fig. 10a).
The oxygen stoichiometry of this reaction was found to be
ca. 0.5 per mole of substrate. This stoichiometry is consistent
with the rapid oxidation of the catechol 27 (R = OH) to the
corresponding quinone 28 (Scheme 4). However, the absence
of a quinone spectrum implies rapid conversion to another
species with a peak absorbance in the 280 nm region, account-
ing for the shift in absorbance maximum from 266 nm.
Comparison with the absorption spectrum at pH 7.4 of the
related indolium-5-olate oxidation product 10b of the catechol
6
(eqn. (3)), which requires reactivation (eqn. (1)), the enzymatic
oxidation of the betaine 29 was unexpected. An important
feature relevant to the oxidation of the betaine 29 is the
presence of a particularly acidic proton on the carbon atom
α to the carbonyl and ammonium groups. The betaine 29, in
contrast to the betaines 10, can, therefore, be expected to be
in equilibrium with the N-ylide tautomer 30, which is a catechol
derivative. It may be that it is this tautomer 30 that is further
oxidised to the quinone 31, which is itself a stabilised N-ylide.
Based on oximetry studies of the phenol 27 (R = H),
we believe that the N-ylide tautomer 30, in contrast to the
betaine 29, may be able to activate met-tyrosinase. Thus the
phenol 27 (R = H) was oxidised without pre-activation and with
almost no lag period, showing similar kinetics to those of the
analogous catechol 25c. The corresponding methyl ether was
not oxidised. The ability of the betaine 29, in contrast to the
8
b, which exhibits a peak at 290 nm with a marked shoulder
1
between 310 and 320 nm, suggests that the observed spectral
changes in the initial phase of this reaction are due to formation
of the betaine 29. Subsequently, in a second experiment of
longer duration a further phase of very slow oxygen utilisation,
associated with a concurrent rise in an absorbance at 418 nm,
was observed (Fig. 10b). This secondary oxidation is ascribed
to the formation of the relatively stable o-quinone 31. The
4
312
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315

View Article Online
betaines 10, to equilibrate with an N-ylide tautomer 30, which is
a neutral catechol derivative, is therefore of special significance
and fully accounts for the observed properties.
Although of some mechanistic interest, these preliminary
studies of model systems that might be prototype pro-drugs
activated in situ by tyrosinase were unsuccessful. In the systems
that we have studied, i.e. 24 and 25, either the nitrogen function
was insufficiently nucleophilic for cyclisation to occur or N-
ylide formation appears to stabilise the cyclic product and pre-
vent hydrolysis. Nevertheless, these results are fully consistent
with our conclusions on the mechanism of tyrosinase oxidation
of phenols which are described in the previous section. We
believe that it should be possible to design other derivatives that
avoid these limitations and thus lead to useful pro-drugs and
alternative strategies are under consideration.
In conclusion, our detailed studies of the formation of novel
heterocyclic betaines by cyclisation of dopamine derivatives via
intermediate o-quinones provide firm evidence that tyrosinase
oxidises phenols to o-quinones in one step and not via an
intermediate catechol. Catechols are formed by an indirect
non-enzymic mechanism and this accounts for the charac-
teristic lag period associated with tyrosinase oxidations. Where-
as dopamine derivatives cyclise to the betaines in one step, the
higher homologue forms the corresponding betaine derivative
in a two step mechanism involving a spiro intermediate.
To monitor tyrosinase-catalysed oxidation simultaneous
oximetric and spectrophotometric measurements were made
1
using the apparatus and methods previously described.
Preparation of indolium-5-olates 10
Unless specific details are given, catecholamines were prepared
as hydrobromide salts using the literature method cited in
each individual preparation. In each case the free amine was
prepared by treating the salt with excess saturated aqueous
sodium bicarbonate and extraction into CHCl3.
2
,3-Dihydro-1,1-dimethyl-6-hydroxy-1H-indolium-5-olate
23
1
0
0a. To a stirred solution of N,N-dimethyldopamine 8a (0.5 g,
.003 mol) in CHCl –MeOH (9:1)(50 cm ) under a nitrogen
3
3
atmosphere was added dropwise (15 min) dianisyltellurium
3
oxide (0.98 g, 0.003 mol) in CHCl –MeOH (9:1)(20 cm ). The
3
resulting red solution was then stirred at ambient temperature
(
30 min). The reaction mixture was partitioned with water
3
Experimental
(50 cm ) and the organic phase was separated and washed with
water (2 × 30 cm ). The combined aqueous layers were then
3
Melting points were determined using a Reichert Kofler Block
apparatus and are uncorrected. Infrared spectra were recorded
on a Perkin-Elmer 881 spectrophotometer with only major
absorbances being quoted. Unless otherwise stated IR spectra
washed with CHCl and the water was removed under reduced
3
pressure to yield compound 10a (0.45 g, 91%) reddish brown
Ϫ1
solid, mp 208 ЊC with some decomp. at 118 ЊC; νmax/cm 1622,
1
492, 1305, 1266, 1152, 1041, 840; λmax (0.1 M phosphate
1
were measured as KBr discs. H NMR spectra were recorded
buffer): pH 7.4, 290 (ε 3360) and 310 (sh) (ε 1500) nm; pH 6.5,
at ambient temperatures using a JEOL GSX270 FT-NMR
spectrometer with tetramethylsilane (TMS) as an internal
reference, and were run in deuterated chloroform solution
unless otherwise stated. Chemical shifts are quoted in parts per
million and the following abbreviations are used: s = singlet;
d = doublet; t = triplet; q = quartet; m = multiplet; br = broad.
Elemental analyses were determined using a Perkin-Elmer 240
CHN Elemental Analyser. Low resolution mass spectra were
recorded on an AEI MS12 Mass Spectrometer at 70 eV electron
impact ionisation. Separations by column chromatography
were carried out using aluminium oxide (150 mesh, Aldrich)
deactivated with water to Brockmann grade IV unless other-
wise stated. Flash chromatography was performed using silica
gel (Janssen Chimica) 0.035–0.07 mm. Preparative radial
2
3
2
86 (ε 3700) nm; δH (D O) 2.98 (t, J 7.0 Hz, 2H), 3.17 (s, 6H),
2
.85 (t, J 7.0 Hz, 2H), 6.51 (s, 1H), 6.64 (s, 1H); δC (D O)
2
6.1 (t), 54.1 (q), 68.5 (t), 101.9 (d), 110.1 (d), 120.3 (s), 136.6
ϩ
(
s), 151.3 (s), 151.7 (s); m/z 179 (48%)(M ). The product
᝽
was further characterised by conversion to the 5,6-dimethoxy
iodide 11a.
In a similar manner the following betaines were prepared
23
24
from the amines 8b and 8c and subsequently converted to
their iodides.
2
,3-Dihydro-1,1-di-n-propyl-6-hydroxy-1H-indolium-5-olate
1
0b (0.45 g, 90%) reddish brown solid, mp 115–120 ЊC;
Ϫ1
νmax/cm (KBr) 3422, 2971, 1491, 1385, 1297, 1267, 1152,
1
2
037, 960, 862, 843; λmax (0.1 M phosphate buffer): pH 7.4,
90 (ε 4120) and 312 (sh) (ε 1450) nm; pH 6.5, 290 (ε 3920)
(
chromatotron) chromatography was carried out on a Harrison
nm; δH (D O) 0.85 (t, J 7.3 Hz, 6H), 1.40 (m, 2H), 1.65
2
Research Ltd Chromatotron 7924 using a 2 mm plate with silica
gel 60 PF254 containing gypsum (Merck). All solvents were pre-
distilled and dried appropriately prior to use. Concentration
and evaporation refer to the removal of volatile materials under
reduced pressure on a Büchi Rotovapor. Substances stated to be
identical were so with respect to mps, mixed mps and IR spectra.
The pulse radiolysis experiments were performed with a 9–12
MeV Vickers linear accelerator, using 50–200 ns pulses with
doses up to ≈33 Gy, and quartz capillary cells of optical path
(
m, 2H), 3.07 (t, J 7.5 Hz, 2H), 3.40 (dt, J 7.9 and 4.4 Hz,
H), 3.55 (dt, J 7.9 and 4.4 Hz, 2H), 3.99 (t, J 7.5 Hz, 2H),
2
6
6
(
2
.63 (s, 1H), 6.67 (s, 1H); δ (CDCl ) 9.8 (q), 16.0 (t), 27.5 (t),
C
3
1.8 (t), 67.1 (t), 103.5 (d), 109.9 (d), 123.1 (s), 132.0 (s), 150.5
ϩ
s), 151.5 (s); m/z (FAB) 236.1639 (MH ). C H NO requires
14 22 2
36.1650.
,3-Dihydro-6-hydroxyspiro[indole-1,1Ј-piperidin]-1-ylium-
-olate 10c (0.40 g, 80%), red brown solid, mp 143 ЊC with some
2
5
Ϫ1
decomp. at 120 ЊC; νmax/cm (KBr) 3412, 1624, 1490, 1456,
300, 1269, 1203, 1156, 1079, 1032, 864; λmax (0.1 M phosphate
buffer): pH 7.4, 288 (ε 3810) and 310 (sh) (ε 1800) nm; pH 6.5,
86 (ε 3800) nm; δ (D O) 1.50 (m, 1H), 1.87 (m, 5H), 3.00 (t,
19,20
2
.5 cm. Ϫ Absorbed doses were determined from the transient
1
Ϫ2
(
SCN)2 formation from air-saturated 10 M KSCN solu-
᝽
21
tions, as described by Adams et al., but using the recently
updated Gε value of 2.59 × 10 m J obtained by Buxton and
Stuart, G being the radiation chemical yield of (SCN) , and
ε its molar absorption coefficient (units of dm mol cm )
2
Ϫ4
2
Ϫ1
H
2
J 7.2 Hz, 2H), 3.22 (d, J 12.4 Hz, 2H), 3.47 (m, 2H), 3.91 (t,
J 7.2 Hz, 2H), 6.59 (s, 1H), 6.74 (s, 1H); δ (D O) 15.0 (t), 16.2
22
Ϫ
2
᝽
3
Ϫ1
Ϫ1
C
2
(
t), 20.7 (t), 55.1 (t), 57.5 (t), 97.4 (d), 105.0 (d), 115.7 (s), 131.5
at 475 nm. Saturation of such solutions with N O results in a
ϩ
2
(s), 146.1 (s), 147.4 (s); m/z 219 (2%)(M᝽
).
Ϫ
doubling of the (SCN)2 yield. Generation of the one-electron
oxidising species Br2 or N3 was achieved by irradiating N O-
᝽
Ϫ
ؒ
᝽
2
Preparation of 5,6-dimethoxyindolium iodides 11
Ϫ3
Ϫ3
saturated aqueous solutions of 100 × 10 M KBr or 3 × 10
Ϫ
ؒ
M NaN . Under such conditions Br2 or N3 radicals are
2,3-Dihydro-5,6-dimethoxy-1,1-dimethyl-1H-indolium iodide
3
᝽
3
formed within ~0.1 µs after the radiation pulse, via the following
11a. To compound 10a (250 mg) and acetone (30 cm ) was
reactions:
added sufficient water to ensure solution (1 or 2 drops). Methyl
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315
4313

View Article Online
3
iodide (2 cm ) and potassium carbonate (300 mg) were then
dimethoxyphenyl)ethylamine (1.5 g). Following an exothermic
3
added and the resulting mixture was heated under reflux (24 h).
After cooling the mixture was filtered and the solvent removed
under reduced pressure to give a yellow solid that was recrystal-
lised from ethyl acetate–methanol and identified as compound
reaction, dichloromethane (5 cm ) and K CO (3.0 g) were
2 3
added and the mixture was stirred (8 h). The resulting suspen-
3
sion was diluted with CHCl (15 cm ), filtered and the residue
3
3
washed with a further portion of CHCl (15 cm ). Evaporation
3
1
1a (390 mg, 83%), colourless crystals, mp 200–201 ЊC (Found:
and drying under high vacuum gave a colourless liquid (1.8 g)
that was identified as the desired amide and used without
further purification. The amide was dissolved in CH Cl
C, 42.82; H, 5.67; N, 4.03. C H INO requires: C, 42.98;
H, 5.41; N, 4.18%); νmax/cm 1600, 1512, 1464, 1433, 1354,
1
3
12
18
2
Ϫ1
2
2
3
310, 1268, 1244, 1225, 1166, 1064, 990, 968, 854, 840; δ (D O)
(15 cm ) and, with cooling at Ϫ80 ЊC, 1 M BBr in CH Cl solu-
3 2 2
3
H
2
.24 (t, J 7.1 Hz, 2H), 3.38 (s, 6H), 3.77–3.79 (2 × s, 6H), 4.09
tion (15 cm ) was added dropwise. After standing overnight,
3
(
t, J 7.1 Hz, 2H), 6.99 (s, 1H), 7.16 (s, 1H).
water (3 cm ) was added and the mixture evaporated under high
In a similar manner the following iodides were prepared from
vacuum to give a viscous oil. Flash chromatography (silica gel:
the corresponding indolium-5-olate.
,3-Dihydro-5,6-dimethoxy-1,1-di-n-propyl-1H-indolium
ethyl acetate–cyclohexane–MeOH (8:2:1) followed by ethyl
2
acetate–cyclohexane–MeOH–MeCO H (40:10:1:1) as eluent)
2
iodide 11b (95%), colourless crystals, mp 172–173 ЊC (Found:
gave a colourless oil which solidified on grinding with Et O
2
C, 49.11; H, 6.69; N, 3.57. C H INO requires C, 49.10; H,
and was identified as N-methyl-N-[2-(3,4-dihydroxyphenyl)-
ethyl]acetamide 24a (400 mg, 25%), tiny crystals, mp 134–
16
26
2
Ϫ1
6
1
.65; N, 3.58%); νmax/cm 1602, 1509, 1464, 1442, 1418, 1358,
304, 1271, 1243, 1222, 1074, 996; λmax (0.1 M phosphate
Ϫ1
136 ЊC; νmax/cm 3254, 1590, 1528, 1445, 1412, 1293, 1236,
buffer): pH 7.4, 284 (ε 4115); unchanged at pH 6.5; δ (CDCl )
1118, 1040, 1018, 871 and 824; δH (d -DMSO) (both amide
H
3
6
0
.95 (t, J 7.0 Hz, 6H), 1.31 (m, 2H), 1.65 (m, 2H), 3.19 (t,
J 7.0 Hz, 2H), 3.82 (s, 3H), 4.0 (s, 3H), 4.12 (t, J 7.0 Hz, 4H),
.23 (t, J 7.0 Hz), 6.70 (s, 1H), 7.61 (s, 1H). 2,3-Dihydro-5,6-
diastereoisomers observed) 1.76 and 1.94 (2s, 3H), 2.56 (dt,
2H), 2.77 and 2.88 (2s, 3H), 3.35 (m, 2H), 6.43 (m, 1H), 6.57 (m,
ϩ
4
1H) and 6.62 (m, 1H); HRMS (EI) Found: MH , 210.1127;
᝽
dimethoxyspiro[indole-1,1Ј-piperidin]-1-ylium iodide 11c (78%),
colourless crystals, mp 228–229 ЊC (Found: C, 47.88; H, 5.76,
N, 3.73. C H INO requires C, 48.00; H, 5.91; N, 3.73%);
Calc. for C H NO : 210.1130.
11
16
3
In a similar manner compound 24b was prepared from tri-
fluoroacetic anhydride and N-methyl-2-(3,4-dimethoxyphenyl)-
ethylamine.
15
22
2
δ (D O) 1.50 (m, 1H), 1.8–2.0 (m, 5H), 3.18 (t, J 7.2 Hz, 2H),
H
2
3
.39 (d, J 12.0 Hz, 2H), 3.65 (dt, J 4.6 and 12.3 Hz, 2H), 3.77
N-Methyl-N-[2-(3,4-dihydroxyphenyl)ethyl]trifluoro-
acetamide 24b (469 mg, 29%), micro-crystals, mp 113–115 ЊC
(Found: C, 50.32; H, 4.59; N, 5.09. C H F NO requires C,
(
s, 3H), 3.79 (s, 3H), 4.08 (t, J 7.2 Hz, 2H), 7.00 (s, 1H), 7.16
(
s, 1H).
11
12
3
3
Ϫ1
5
1
0.19; H, 4.56; N, 5.32%); νmax/cm 3368, 1684, 1618, 1533,
450, 1374, 1196, 1151, 1084 and 952; δH (d -DMSO) (both
1
,1-Diethyl-7-hydroxy-1,2,3,4-tetrahydroquinolinium-6-olate
6
amide diastereoisomers observed) 2.7 (2t, 2H), 2.98 and 3.00
2q, JHF 0.7 and 1.7 Hz, 3H), 3.50 (2t, 2H), 6.45 (m, 1H), 6.59
1
5. To a stirred solution of N,N-diethyl-3-(3,4-dihydroxy-
(
phenyl)propylamine 13 (0.5 g, 0.002 mol), prepared from 3-
and 6.60 (2d, J 2.2 Hz, 1H), 6.63 and 6.65 (2d, J 7.8 Hz, 1H);
(
3,4-dimethoxyphenyl)propionic acid by the method of Ginos
ϩ
).
24
3
m/z 263 (6%)(M᝽
et al., in CHCl –MeOH (9:1)(50 cm ) under a nitrogen
atmosphere was added dropwise (15 min) dianisyltellurium
oxide (0.8 g, 0.002 mol) in CHCl –MeOH (9:1)(20 cm ). The
resulting red solution was then stirred at ambient temperature
3
3
N-(4-Bromobenzoylmethyl)-N-[2-(3,4-dihydroxyphenyl)-
3
ethyl]-N-methylammoniumbromide 25c.
A mixture of N-
methyl-2-(3,4-dimethoxyphenyl)ethylamine (2.0 g) in toluene
(
(
30 min). The reaction mixture was partitioned with water
50 cm ) and the organic phase was separated and washed with
3
3
(
50 cm ) and NaHCO (2.0 g) and Na SO ؒ7H O (1.0 g)
3 2 3 2
3
3
in water (20 cm ) was stirred under a nitrogen atmosphere. 4-
Bromophenacyl bromide (2.8 g) was added and stirring was
water (2 × 30 cm ). The combined aqueous layers were then
washed with CHCl and the water was removed under reduced
pressure to yield compound 15 (0.42 g, 84%) reddish brown
solid, mp 95–100 ЊC; νmax/cm (KBr) 3405, 1653, 1501, 1399,
3
maintained (4 h). The organic phase was washed with saturated
3
Ϫ1
Na HPO solution (50 cm ) and then shaken with aqueous 1 M
2
4
3
HBr (20 cm ) to precipitate a viscous yellow oil, which solidified
1
2
273, 1251, 1197, 1008; λmax (0.1 M phosphate buffer): pH 7.4,
86 (ε 2830) and 308 (sh) (ε 1000) nm; pH 6.5, 284 (ε 2800) nm;
on standing (2 days). This material was dissolved in warm
3
3
CHCl (10 cm ) and hexane (20 cm ) was added to give a yellow
δ (D O) 1.09 (t, J 7.0 Hz, 6H), 2.0 (m, 2H), 2.57 (t, J 7.0 Hz,
3
H
2
solid (2.70 g) that was identified as compound 25a and used
2
3
1
H), 3.4–3.8 (m, 6H), 6.42 (s, 1H) and 6.63 (s, 1H); δ (D O)
.0 (q), 12.4 (t), 19.5 (t), 50.0 (t), 56.8 (t), 102.2 (d), 109.9 (d),
16.2 (s), 122.5 (s), 144.7 (s), 146.0 (s); m/z 221 (26%)(M ).
c 2
without further purification.
3
ϩ
The product 25a was suspended in CH Cl (40 cm ) and 1 M
2
2
᝽
3
BBr in CH Cl (11 cm ) was added dropwise (5 min) at room
The product was further characterised by conversion to the
3
2
2
temperature. The resulting dark red–brown solution was
6
,7-dimethoxy iodide 18.
allowed to stand at room temperature (2 days) after which a
3
colourless solid had precipitated. Aqueous 1 M HBr (1 cm )
6
,7-Dimethoxy-1,1-diethyl-1,2,3,4-tetrahydroquinolinium
was added and the solvent evaporated at room temperature.
iodide 18. In the manner described for the preparation of
compound 11a, the betaine 15 (250 mg) was treated with MeI–
K CO in acetone to give a yellow solid that was recrystallised
3
The residue was stirred with water (10 cm ) and filtration gave
a pale yellow solid (2.51 g). This material was recrystallised
2
3
3
from water (200 cm ) and identified as compound 25c (1.46 g,
from ethyl acetate–MeOH and identified as compound 18
380 mg, 90%), colourless crystals, mp 230–231 ЊC (Found: C,
7.54; H, 6.64; N, 3.51. C H INO requires C, 47.73; H, 6.41;
6
3%), tiny pale yellow crystals, mp 234 ЊC (Found: C, 45.58;
(
4
H, 4.18; N, 2.86. C H Br NO requires C, 46.05; H, 4.32;
17
19
2
3
15
24
2
Ϫ1
Ϫ1
N, 3.16%); νmax/cm 3441, 3175, 2963, 1687, 1588, 1517, 1400,
347, 1283, 1237, 1193, 1125, 1074, 1009, 961, 840, 824 and 790;
δ (d -DMSO) 2.88 (t, 2H), 2.92 (s, 3H), 3.16 (br m, 2H), 5.06
N, 3.71%); νmax/cm (KBr) 3425, 2973, 2940, 2362, 1613, 1514,
443, 1259, 1233, 1222, 1181, 1056, 1019, 940, 904, 860, 813;
δ (D O) 1.14 (t, J 7.1 Hz, 6H), 2.09 (m, 2H), 2.79 (t, J 6.4 Hz,
1
1
H
6
H
2
(
br m, 2H), 6.5–6.7 (m, 3H), 7.85–7.95 (q, 4H), 9.81 (br s, 1H);
2
6
H), 3.6–3.7 (m, 4H), 3.78 (s, 3H), 3.79 (s, 3H), 3.8–3.9 (m, 2H),
.85 (s, 1H), 6.95 (s, 1H).
ϩ
m/z 367 (18%)(MH᝽
).
Preparation of N-(4-bromobenzoylmethyl)-N-[2-(4-hydroxy-
Preparation of catechols 24 and 25
phenyl)ethyl]-N-methylammonium bromide 27 (R ؍ H).
A
N-Methyl-N-[2-(3,4-dihydroxyphenyl)ethyl]acetamide
Acetic anhydride (3.0 g) was added to N-methyl-2-(3,4-
24a.
solution of NaHCO (2.8 g) and Na SO ؒ7H O (1.0 g) in water
(20 cm ) was magnetically stirred under an atmosphere of
3
2
3
2
3
4
314 J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315

View Article Online
3
nitrogen. Toluene (50 cm ) was added followed by N-methyl-4-
3 C. J. Cooksey, P. J. Garratt, E. J. Land, C. A. Ramsden and
25
P. A. Riley, Biochem. J., 1998, 333, 685.
methoxyphenylethylamine hydrochloride (3.0 g) and then 4-
bromophenacyl bromide (2.9 g). After further stirring (4 h) the
organic phase was separated, washed with saturated Na HPO
solution (50 cm ) and then stirred with aqueous HBr (3 cm of
9% HBr made to 15 cm ). After standing overnight the colour-
less precipitate was collected and washed with H O (5 cm ) and
then ether (2 × 5 cm ) to give the desired tertiary amine (2.2 g,
0%); δH (d -DMSO) 2.96 (s, 3H), 3.02 (t, J 5.7 Hz, 2H), 3.35
br s, 2H), 3.72 (s, 3H), 5.13 (s, 2H), 6.89 (d, J 8.5 Hz, 2H), 7.22
d, J 8.5 Hz, 2H), 7.84 (d, J 8.7 Hz, 2H), 7.95 (d, J 8.7 Hz, 2H),
4
5
A. Sánchez-Ferrer, J. N. Rodríguez-López, F. García-Cánovas and
F. García-Carmona, Biochim. Biophys. Acta, 1995, 1247, 1.
H. S. Raper, Physiol. Rev., 1928, 8, 245; W. C. Evans and
H. S. Raper, Biochem. J., 1937, 31, 2162; J. M. Nelson and C. R.
Dawson, Adv. Enzymol., 1944, 4, 99; A. B. Lerner, T. B. Fitzpatrick,
E. Calkins and W. H. Summerson, J. Biol. Chem., 1950, 187, 793;
H. S. Mason, W. L. Fowlks and E. Peterson, J. Am. Chem. Soc.,
2
4
3
3
3
4
3
2
3
1
1
955, 77, 2914; H. Wyler and J. Chiovini, Helv. Chim. Acta, 1968, 51,
5
6
476; M. R. Chedekel, E. J. Land, A. Thompson and T. G. Truscott,
(
(
9
J. Chem. Soc., Chem. Commun., 1984, 1170; S. Naish-Byfield and
P. A. Riley, Pigm. Cell Res., 1998, 11, 127.
6 H. S. Mason, Nature, 1956, 177, 79; K. Lerch, in Metal Ions in
Biological Systems, ed. H. Sigel, Marcel Dekker, Inc., New York,
Vol. 13, pp. 143–186; E. I. Solomon and M. D. Lowery, Science,
.91 (br s, 1H). This material was demethylated without further
purification using the following procedure.
The methyl ether (1.6 g) was suspended in CH Cl and with
magnetic stirring under an atmosphere of nitrogen, BBr3
2
2
1
993, 259, 1575.
C. T. Jagoe, S. E. Kreifels and J. Li, Bioorg. Med. Chem. Lett., 1997,
, 113.
R. Robinson and S. Sugasawa, J. Chem. Soc., 1932, 789.
7
8
3
(
4 cm , 1M in CH Cl ) was added. The solid dissolved to give a
2 2
7
gold coloured solution which after standing overnight was pale
brown and a brown gum had precipitated. Hydrolysis by stirring
with 1 M HBr gave a pale cream powder that was identified as
the phenol 27 (R = H) (HBr salt) (1.13 g, 74%), mp 120–122 ЊC;
νmax/cm (KBr) 3509, 3435, 3221, 2969, 1688, 1586, 1514, 1471,
450, 1406, 1339, 1271, 1243, 1069, 1008, 964, 834, 814;
δ (d -DMSO) 2.95 (s, 3H), 3.35 (m, 2H), 5.07 (br d, 1H), 5.19
br d, 1H), 6.73 (d, J 8.3 Hz, 2H), 7.09 (d, J 8.3 Hz, 2H), 7.86
d, J 8.3 Hz, 2H), 7.95 (d, J 8.3 Hz, 2H), 9.32 (v br s, 1H),
.89 (br s, 1H); m/z (Electrospray) 348.0605 (MH ; Br).
C H NO Br requires 348.0599.
9 C. Schöpf and K. Thierfelder, Liebigs Ann. Chem., 1932, 497, 22.
10 C. A. Ramsden, Chem. Soc. Rev., 1994, 23, 111; C. A. Ramsden,
J. Heterocyl. Chem., 1999, 36, 1573.
Ϫ1
11 S. V. Ley, C. A. Meerholz and D. H. R. Barton, Tetrahedron, 1981,
3
7, 213.
1
1
2 C. Lambert, T. G. Truscott, E. J. Land and P. A. Riley, J. Chem.
Soc., Faraday Trans., 1991, 87, 2939.
H
6
(
13 E. J. Land, J. Chem. Soc., Faraday Trans., 1993, 89, 803.
14 C. J. Cooksey, E. J. Land, F. A. P. Rushton, C. A. Ramsden and
P. A. Riley, Quant. Struct.-Act. Relat., 1996, 15, 498.
5 J. E. Gervay, F. McCapra, T. Money, G. M. Sharma and A. I. Scott,
J. Chem. Soc., Chem. Commun., 1966, 142.
6 A. I. Scott, P. A. Dodson, F. McCapra and M. B. Meyers, J. Am.
Chem. Soc., 1963, 85, 3702.
7 E. Wenkert, N. F. Golob and R. A. J. Smith, J. Org. Chem., 1973, 38,
4068.
8 A. M. Jordan, T. H. Khan, H. M. I. Osborn, A. Photiou and
P. A. Riley, Bioorg. Med. Chem, 1999, 7, 1775.
9 J. P. Keene, J. Sci. Instrum., 1964, 41, 493.
0 J. Butler, B. W. Hodgson, B. M. Hoey, E. J. Land, J. S. Lea,
E. J. Lindley, F. A. P. Rushton and A. J. Swallow, Radiat. Phys.
Chem., 1989, 34, 633.
1 G. E. Adams, J. W. Boag, J. Currant and B. D. Michael, in
Pulse Radiolysis, ed. M. Ebert, J. P. Keene, A. J. Swallow and
J. H. Baxendale, Academic Press, London, 1965, p. 117.
2 G. V. Buxton and C. R. Stuart, J. Chem. Soc., Faraday Trans., 1995,
(
9
ϩ
79
1
1
1
1
1
7
19
2
Acknowledgements
The pulse radiolysis experiments were performed at the
Paterson Institute for Cancer Research Free Radical Research
Facility, Christie Hospital NHS Trust, Manchester. The
facility is supported by the European Commission T.M.R.
Programme—Access to Large-Scale Facilities. E. J. L. thanks the
Cancer Research Campaign for financial support. We thank
the EPSRC National Mass Spectrometry Service Centre for
high-resolution mass spectra.
1
2
2
2
2
9
1, 279.
References
3 J. G. Cannon, F.-L. Hau, J. P. Long, J. R. Flynn, B. Costall and
1
C. J. Cooksey, P. J. Garratt, E. J. Land, S. Pavel, C. A. Ramsden,
P. A. Riley and N. P. M. Smit, J. Biol. Chem., 1997, 272,
R. J. Naylor, J. Med. Chem., 1978, 21, 248.
24 J. Z. Ginos, G. C. Cotzias, E. Tolosa, L. C. Tang and A. LoMonte,
J. Med. Chem., 1975, 18, 1194.
25 T. Katsu, H. Ono, K. Tasaka and Y. Fujita, Chem. Pharm. Bull.,
1984, 32, 4185.
2
6226.
2
J. Clews, C. J. Cooksey, P. J. Garratt, E. J. Land, C. A. Ramsden
and P. A. Riley, Chem. Commun., 1998, 77.
J. Chem. Soc., Perkin Trans. 1, 2000, 4306–4315
4315