Indigo dye production by enzymatic mimicking based on an iron(III)porphyrin

Article abstract of DOI:10.1016/j.jcat.2014.04.012

A novel indigo synthesis is based on a simple and cost-effective model system of the enzymes involved in the natural and biocatalytic productions. The method considers the oxidation of indole by hydrogen peroxide, being catalyzed by an iron(III)porphyrin in ethanol, as solvent, and no other additives. The yields of indigo and of the other oxidized indolinoid derivatives were found to be dependent on the metalloporphyrin system used and on the control of the oxidation conditions. Significant reductions of the environmental impact relatively to the present industrial production and of the costs relatively to the biocatalytic methodologies were obtained. The enhanced indigo production in the presence of the iron(III)porphyrin-ethanol catalytic system relatively to the manganese(III)porphyrin-acetonitrile system can be rationalized by the formation of different active species in the two systems.

Full text of DOI:10.1016/j.jcat.2014.04.012

Journal of Catalysis 315 (2014) 33–40
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Indigo dye production by enzymatic mimicking
based on an iron(III)porphyrin
a
b
b
Susana L.H. Rebelo a, , Margarida Linhares , Mário M.Q. Simões , Artur M.S. Silva ,
⇑
M. Graça P.M.S. Neves b, , José A.S. Cavaleiro b, Cristina Freire a
⇑
a REQUIMTE, Departamento de Química e Bioquímica, Faculdade de Ciências, Universidade do Porto, Rua do Campo Alegre, 4169-007 Porto, Portugal
b QOPNA, Departamento de Química, Universidade de Aveiro, 3810-193 Aveiro, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel indigo synthesis is based on a simple and cost-effective model system of the enzymes involved in
the natural and biocatalytic productions. The method considers the oxidation of indole by hydrogen
peroxide, being catalyzed by an iron(III)porphyrin in ethanol, as solvent, and no other additives. The
yields of indigo and of the other oxidized indolinoid derivatives were found to be dependent on the
metalloporphyrin system used and on the control of the oxidation conditions. Significant reductions of
the environmental impact relatively to the present industrial production and of the costs relatively to
the biocatalytic methodologies were obtained. The enhanced indigo production in the presence of the
iron(III)porphyrin-ethanol catalytic system relatively to the manganese(III)porphyrin-acetonitrile system
can be rationalized by the formation of different active species in the two systems.
Received 10 February 2014
Revised 15 April 2014
Accepted 18 April 2014
Keywords:
Indigo
Indole
Iron(III)porphyrin
Oxidation
Ó 2014 Elsevier Inc. All rights reserved.
Catalysis
Hydrogen peroxide
Mechanism
Ethanol
1. Introduction
result in the production of high amounts of toxic waste products,
and require sophisticated purification processes [7].
Indigo is a historical dye used by almost all the ancient civiliza-
tions and with relevance until our days, mainly due to its incidence
as the primary color of blue jeans [1]. Other indigoid compounds
are also prominent targets in textile industries and for alimentary,
medicinal and cosmetic purposes [2–4].
Until the late 19th century, indigo was obtained from natural
sources and, with the advent of the modern chemical industry,
became one of the first natural molecules to be synthesized [5].
Today, almost all the traded indigo is produced synthetically and,
in general, the branded industrial production is based on variations
of the Pflegers’s method. In the largely used process, the synthesis
involves the reaction of aniline, formaldehyde and hydrogen
cyanide, affording phenylglycinonitrile that is then hydrolyzed to
N-phenylglycine. Subsequently, the N-phenylglycine is treated
with molten mixtures of caustic soda and sodamide at 200 °C to
afford indoxyl, which undergoes further oxidative dimerization
to form indigo [6]. The processes are highly energy demanding,
The high production rates and the amount of hazardous wastes
have been raising significant environmental concern that have
stimulated the reediting of the natural production [8], and since
the work of Ensley et al. in 1983, the microbial productions have
been considered as an important eco-sustainable alternative [9].
These authors reported an Escherichia coli mutant encoding the
enzyme naphthalene dioxygenase (NDO) that was able to synthe-
size indigo from indole. Several other mutant strains have been
tested in order to improve the methodology [10,11], which is still
expensive and has significant drawbacks in obtaining pure indigo
from the culture media (fermentation broth). Enzymatic systems
were also tested, such as cytochrome P450 monooxygenases [12],
non-heme oxygenases such as NDO [13], and engineered mutants
[14,15]. Recently, myoglobin mutants were developed and succeed
to directly activate hydrogen peroxide; the system allowed to
avoid the use of expensive co-factors, NADPH or NADH, required
for the activation of molecular oxygen, thus affording a more prac-
tical and green indigo production [16]. The best indigo yield
obtained was 12%, based on the consumed indole [16]. However,
the cost and purification drawbacks in enzymatic synthesis also
need to be considered and, furthermore, significant quantities of
⇑
Corresponding authors. Fax: +351 220402659.
E-mail addresses: susana.rebelo@fc.up.pt (S.L.H. Rebelo), gneves@ua.pt
(M. Graça P.M.S. Neves).
http://dx.doi.org/10.1016/j.jcat.2014.04.012
0021-9517/Ó 2014 Elsevier Inc. All rights reserved.

34
S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
indigo structural isomers, namely indirubin, are produced
concomitantly.
microwave heating, in acetic acid as the reaction solvent, replacing
the usual hazardous mixture of nitrobenzene and acetic acid. The
complexation of the porphyrin was obtained by the reaction with
iron(II) chloride at reflux in DMF and in the presence of pyridine
[26]. MS FAB+ [FeP], m/z [MꢀCl]+ = 1128.2 (Fig. S1).
Alternative processes using chemical methodologies are scarce;
one of these approaches refers to the indigo synthesis with an
organometallic catalyst and an organic hydroperoxide, in an anhy-
drous media [17]; another strategy is the non-catalytic oxidation of
indole by meta-chloroperoxybenzoic acid (m-CPBA) [18]. However,
both processes led to the introduction of high amounts of expen-
sive reactants in the reaction mixture, hindering product isolation,
and the eco-sustainability of the processes, in terms of by-products
release, cost and energy consumption was still compromised.
In the context of cytochrome P450 mimicking by metallopor-
phyrin catalytic systems, several novel and relevant synthetic
routes have been opened, using mild conditions [19–22]. Simulta-
neously, an improved understanding on the reactivity patterns of
the catalytic and biological systems has been achieved, specifically
the nature of the porphyrin nucleus, the central metal, and the
reaction conditions (protic/aprotic solvent and the type of addi-
tives) showed a prominent effect on the reactivity pattern of the
catalytic system [23,24], which was rationalized by the tuning of
the active species formed [25].
The recently reported oxidation of indole in the presence of
different Mn(III)porphyrins, using acetonitrile as solvent and a
co-catalyst, afforded a very poor indigo yield that, after optimiza-
tion, reached only 3% [26]. In the present work, a complimentary
Fe(III)porphyrin catalytic system was developed for indole
oxidation by hydrogen peroxide. In addition, ethanol was consid-
ered as a protic and eco-sustainable solvent [27–29]; furthermore,
ethanol is less expensive relatively to other organic solvents, can
be obtained from biomass, is biodegradable, and can be used as a
fuel.
2.4. Indigo production in the presence of FeP under optimized
conditions
Indole (35.1 mg, 0.3 mmol) and the catalyst (1.0 mg, 0.3%) were
placed on a round-bottom flask and dissolved in 2 mL of ethanol.
The mixture was left under magnetic stirring at room temperature
(ꢁ20 °C) and the hydrogen peroxide was progressively added in a
total amount of 0.6 mmol for 20 min [adequate solutions of H2O2
in ethanol (1:10 v/v) were prepared to allow the accurate addi-
tion]. The reaction was stopped by removing the excess of catalyst
and H2O2 by passing the reaction mixture through a small plug of
silica gel and washing it with hexane/ethyl acetate (1:1). The cata-
lyst, a colored polar metalloporphyrin is not eluted through the
silica column in these conditions and is retained in the top of the
column. The collected mixture was evaporated to dryness under
vacuum and the residue diluted in an exact volume of dimethyl
sulfoxide (DMSO). Since indigo is insoluble in common HPLC
reversed-phase solvents, the amount of indigo was determined
by UV–Vis spectrophotometry in comparison with a calibration
curve obtained with standard solutions of indigo in DMSO.
Otherwise, the precipitation of indigo was obtained through
immersion of the reaction mixture in water, followed by filtration
through filter paper and washing with water and then with ethanol
to afford pure indigo. The resulting water and ethanol filtrate frac-
tions were evaporated to dryness and fractionated by preparative
TLC using a mixture of hexane/ethyl acetate (2:1) as eluent; the
isolated products were characterized by NMR spectroscopy and
mass spectrometry (Supporting Information).
2. Materials and methods
2.1. Chemicals
2.5. Catalytic reactions in the presence of the MnP–CH3CN system
Aqueous hydrogen peroxide 30% (w/w) [H2O2(aq)], urea–H2O2
adduct compound (UHP), and m-chloroperoxybenzoic acid
(m-CPBA, 77%) were purchased from Aldrich. All other reagents
(>95%) and solvents (p.a.) were of high purity and used as received.
Silica gel 60 (0.063–0.200 mm) was used for column chromatogra-
phy, and thin layer chromatography (TLC) was performed using
silica gel 60 with F254 indicator.
Indole (35.1 mg, 0.3 mmol), 1.0 mg of catalyst (0.3%), and 40 mg
of ammonium acetate were placed in a round-bottom flask and
dissolved in 4 mL of acetonitrile. The mixture was left under
magnetic stirring at room temperature (ꢁ20 °C) and the hydrogen
peroxide was progressively added in a total amount of 1.2 mmol
for 20 min (adequate solutions of H2O2 in acetonitrile (1:10 v/v)
were prepared to allow the accurate addition). The reactions were
left to proceed for 10 min. Quantification of indigo was obtained as
described for the FeP–EtOH system.
2.2. Instruments
Electronic spectra were recorded on a range cell Cary 50 BIO
spectrometer. NMR spectroscopy was performed on a Brucker
Avance III spectrometer at a frequency of 400.15 and 100.63 MHz
for 1H and 13C NMR, respectively. The mass spectra were acquired
with electrospray or fast atom bombardment ionization in positive
modes, ES+ or FAB+, respectively, at Unidad de Masas, Universidad
de Santiago de Compostela. GC/MS analyses were carried out on a
Finnigan Trace GC/MS (Thermo Quest instruments) using helium
as the carrier gas (35 cm/s), a DB-5-type fused silica capillary col-
2.6. Non-catalytic oxidation with m-CPBA
Indole (35.1 mg, 0.3 mmol) and K2CO3 (19 mg) were placed in a
round-bottom flask and dissolved in 4 mL of solvent. When ethanol
was used as solvent, the system was denominated m-CPBA–EtOH;
when the solvent was dichloromethane, the system was denomi-
nated m-CPBA–CH2Cl2. The mixture was left under magnetic
stirring at room temperature (ꢁ20 °C) and 103 mg of m-CPBA
(0.6 mmol) were added. The reaction was stopped after 20 min
and the quantification of indigo was obtained using the same pro-
tocol described for the catalytic oxidation in the presence of the
FeP–EtOH system.
umn (30 m, 0.25 mm i.d., 0.25 lm film thickness), and the mass
spectra were acquired with electron impact ionization at 70 eV.
2.3. Metalloporphyrin synthesis
2.7. 1H NMR of the total reaction mixture
The metalloporphyrin catalyst chloro[5,10,15,20-tetrakis(4-
dimethylamino-2,3,5,6-tetrafluorophenyl)porphyrinate]iron(III),
denominated as FeP (Fig. 1), was prepared using an adapted liter-
ature procedure [30]. The synthesis of the 5,10,15,20-tetrakis(pen-
tafluorophenyl)porphyrin ligand (H2TPFPP) was performed under
In order to determine the yields of the other reaction products,
the reactions were performed following the appropriate procedure,
and after passing through the short silica column, the eluate was
evaporated to dryness and the residue was dissolved in DMSO-d6

S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
35
Ar
Ar
F
F
Cl
Cl
N
M
N
Cl
Ar =
N
Ar =
N
N
F
F
FeP (M = Fe)
MnP (M = Mn)
Ar
Ar
Fig. 1. Metalloporphyrin catalysts.
for the NMR analysis. This analysis considered the non-superim-
posed assigned proton signals of the co-oxidation products.
(I.2), isatin (I.4) and indirubin (I.5), which were observed during
catalysis with heme-enzymes [10,12], whereas the product result-
ing from indole trimerization, 2,2-bis(30-indolyl)-3-oxoindole (I.3)
was detected during the oxidation of indole by laccase [31].
The control of the reactions by TLC showed that higher reaction
times and, consequently, higher amounts of oxidant led to the
bleaching of the formed indigo, as a result of successive oxidation
steps illustrated in Scheme 1. Indole (I.0) is primarily oxidized to 2-
and 3-hydroxyindoles (indoxyls), which are the precursors of the
isolated products [12]. The 3-indoxyl undergoes oxidative dimer-
ization to afford indigo (I.1) and this pigment is further oxidized
with the formation of isatin (I.4) [32,33]. With the present system,
this probably occurs by epoxidation of the median double bond of
indigo, followed by the epoxide ring opening and oxidative cleav-
age. Consequently, the high yields of isatin and the low yields of
indigo suggest that, under the above conditions, indigo is being
bleached.
3. Results and discussion
The oxidation of indole in the presence of the chloro
[5,10,15,20-tetrakis(4-dimethylamino-2,3,5,6-tetrafluorophe-
nyl)porphyrinate]iron(III), further denominated as FeP (Fig. 1), was
carried out in ethanol with no other additives, at room tempera-
ture and with progressive addition of H2O2. Starting from the
previously reported reaction conditions for the catalytic oxidation
of organic compounds in the presence of metalloporphyrins, indigo
could only be obtained and its yield improved, after optimization
of key reaction parameters. Furthermore, the comparison between
the optimized iron(III)porphyrin system with other metallopor-
phyrin catalytic systems, oxidation by m-CPBA and enzymatic
systems, invoking data from earlier work, allowed some consider-
ations on the mechanistic pathways leading to indigo dye and to
other substituted indole derivatives.
3.1. Screening of the reaction conditions
Initially, the oxidation of indole was carried out using literature
reports on catalysis by iron(III)porphyrins [25]. Specifically, in a
reaction of 4 h at a ratio substrate/catalyst (S/C) of 300 and addi-
tion of H2O2 at a rate of 0.6 mmol hꢀ1 (Table 1, entry 1), which cor-
responds to a total molar ratio oxidant/substrate (Ox/S) of 8, the
indigo pigment (I.1, Scheme 1) could not be isolated but was
detected by UV–Vis spectroscopy (k 600 nm in CH2Cl2) [12]. The
same was observed for indirubin (I.5), the pink pigment known
as a contaminant of natural and biocatalytic indigo (k 550 nm in
CH2Cl2). Isatin (I.4, Scheme 1) was the major product isolated with
49% yield and the second most abundant product was 2-oxoindole
(I.2), isolated in 21% yield, followed by the trimer 2,2-bis(30-indo-
lyl)-3-oxoindole (I.3) with 16% yield, and the dimer I.6 with 6%
yield. The fraction identified as ‘‘others’’ was recovered with a total
yield of 7% and included 2-formylaniline and ethyl 2-aminobenzo-
ate. The reaction products were identified and characterized by
NMR and MS studies (Supporting Information).
Expecting to improve indigo yields, key reaction parameters
were screened in order to find the best compromise between a
high conversion of indole and a low degradation of indigo in the
oxidizing media such as: (i) reaction time; (ii) H2O2 addition rate;
(iii) molar ratio substrate/catalyst (S/C); (iv) solvent; (v) reaction
volume; (vi) temperature; and (vii) the form of H2O2, as aqueous
solution or as an urea–H2O2 adduct.
Since indigo is soluble in dimethyl sulfoxide (DMSO) or
dimethylformamide (DMF), but very insoluble in common organic
solvents and water, the control of the reactions by HPLC is not
appropriate. After optimization, indigo could be isolated by precip-
itation from the reaction mixture (see below), but for accurate
comparison during the conditions survey, the quantification of
indigo was obtained by UV–Vis spectrophotometry. The reactions
were stopped at the indicated time passing the mixture through
a small plug of silica to remove the excess of catalyst and oxidant
(see experimental part). The eluted pale green solution contained
mainly 2- and 3-indoxyls, but not indigo, as confirmed by
With the exception of dimer I.6, the other products were also
reported in biological oxidations of indole, namely 2-oxoindole
Table 1
Yields of Indigo and other co-products obtained during indole oxidations with different systems.
(%)a
Indigo I.1 2-Oxo indole I.2 Trimer I.3 Isatin I.4 Indirubin I.5 Dimer I.6 Others Total
Entry Catalytic System (reaction time) Ox/S (nH2O2/nsubstrate
)
g
1
2
3
4
5
FeP–EtOHb (4 h)
8
2
4
2
2
Trace
20
3
9
3
21
31
64
20
13
16
18
5
20
16
49
6
6
10
4
Trace
4
7
13
7
6
16
–
–
–
7
–
–
18
22
99
97
85
90
65
FeP–EtOHb (20 min)
MnP–CH3CNb (30 min) [26]
m-CPBA–EtOH (20 min)
m-CPBA–CH2Cl2; (20 min)
c
c
a
b
c
Yields based on 1H NMR of the total reaction mixture except for entry 1, which are isolated yields.
Ratio substrate/catalyst was 300.
Not determined.

36
S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
O
O
OH
H
N
(a)
(b)
(c)
O
[O]
[O]
N
N
H
N
N
H
H
H
O
I.0
Indole
3-Indoxyl
I.4
Isatin
I.1
Indigo
OH
(a)
3-Indoxyl
[O]
OH
N
H
[O]
OH
N
H
O
O
HN
HN
H
2-Indoxyl
O
NH
NH
O
HO
O
N
H
HN
HN
O
O
N
H
I.5 Indirubin
I.6
I.2
I.3
2,2-bis(3'-indolyl)-3-oxoindole
2-Oxoindole
Scheme 1. Isolated reaction products from indole oxidation by H2O2 in the presence of FeP at room temperature, using ethanol as solvent.
UV–Vis spectrophotometry. This fraction was then evaporated to
dryness under vacuum and the residue turned dark blue, violet
or green, depending on the relative amounts of indigo, indirubin
or isatin present. Then, the residue was completely dissolved in
DMSO and the resulting solutions were analyzed and indigo quan-
tified by UV–Vis spectrophotometry (Fig. 2).
In the first series of experiments (Fig. 2a–c), different reaction
times were tested, using the conditions reported in the first exper-
iment: room temperature, ethanol as the reaction solvent, a S/C
molar ratio of 300 and the progressive addition of aqueous H2O2
at a rate of 0.6 mmol hꢀ1. For the selected reaction times, 1 h, 2 h
and 4 h, the oxidant/substrate molar ratio (Ox/S) was 2, 4 and 8,
respectively. The best indigo yield was obtained after 1 h of reac-
tion time (11%), corresponding to the addition of 2 mol equiv of
oxidant relatively to the substrate. The TLC control of this reaction
also showed nearly total substrate conversion.
In the subsequent experiments (d–f), 2 mol equiv of oxidant rel-
atively to the substrate (Ox/S = 2) was added at different rates,
leading to different reaction times. The results showed that the
decrease in the initial rate of H2O2 addition from 0.6 mmol hꢀ1
(Fig. 2a) to 0.3 mmol hꢀ1 was responsible for a decrease in indigo
yield from 11% to 3%. On the contrary, when the rate of addition
of H2O2 was increased to 1.8 mmol hꢀ1, after 20 min of reaction
the best indigo yield of 20% was obtained. Further increase in the
rate of H2O2 addition to 3.6 mol hꢀ1 resulted again a decrease in
indigo yield to 13%. The TLC control of this last experiment
evidenced a lower substrate conversion and corroborated the
required compromise between the extension of indole oxidation
and indigo over-oxidation.
Based on the best conditions achieved so far (Fig. 2e), additional
reactions were performed with two other catalyst loadings, namely
S/C of 200 and 400 (g–h). The results showed lower indigo yields
Fig. 2. Screening of the reaction conditions for the optimization of indigo yield: (a–c) different reaction times starting from initial conditions; (d–f) different H2O2 addition
rate, and the other conditions are as in (a); (g–m) all the other conditions are as in (e).

S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
37
under these new S/C ratios: the indigo yield decrease from 20% to
(a)
(b)
(c)
5%, when the S/C = 400, probably due to an incomplete conversion
of indole (confirmed by TLC) or to 10% when S/C = 200, due to an
over-oxidation of indigo in the presence of a higher catalyst
amount.
(I.0)
(I.1)
The role of the solvent was also evaluated (i–j) and the oxida-
tion of indole was carried out in the most common solvents used
with iron(III)porphyrin catalysts, the mixture (1:1) of methanol/
dichloromethane as protic solvent, and acetonitrile. In this latter
solvent, the reaction afforded only trace amounts of indigo, which
can be justified by the very low substrate consumption, as was
confirmed by TLC. With the system based on the methanol/dichlo-
romethane mixture, a faster consumption of the substrate was
observed by TLC, when compared with the system in ethanol,
but indigo yield did not exceed 10%. These conditions are probably
leading to the over-oxidation of indigo. Therefore, the results show
a superior catalytic performance in ethanol, a greener and cost-
effective solvent. When the volume of ethanol was reduced to a
half (2 mL instead of 4 mL), the indigo production was maintained
at 20% yield (results not shown).
Since temperature plays an important role on the conversion of
indoxyls into indigo, the effect of a higher temperature was also
verified (k). However, no advantage was obtained by increasing
the reaction temperature to 50 °C, since the indigo yield was only
4%. The metalloporphyrins also catalyze the H2O2 dismutation
reaction [34] and this reaction becomes more important relatively
to the oxygenation pathway at higher temperatures.
(I.5)
(I.3)
(I.4)
(I.2)
(I.6)
Ethanol
Water
Filtrate fraction
t = 0'
20'
Reaction mixture
Fig. 3. Representative draws of TLC controls from the reaction in Table 1, entry 2:
(a) total reaction mixture; (b and c) after washing the indigo precipitate with water
and ethanol, respectively. Non-colored products were observed on the UV lamp at
k = 254 nm.
acetonitrile and the use of ammonium acetate as co-catalyst was
required [26]. This will be denominated as the MnP–CH3CN sys-
tem. On the other side, the non-catalytic oxidation of indole by
m-CPBA was the object of a patent [18], and attending to the
impossibility to compare our results with the reported yields, since
they were determined from the mass of precipitate formed in dif-
ferent conditions and characterized by FTIR and probably included
impure indigo, an adapted procedure was reproduced at room
temperature, and the same quantification methodology used for
FeP experiments was applied. Two different solvents were tested:
dichloromethane as an aprotic solvent, the m-CPBA–CH2Cl2 sys-
tem, and ethanol as a protic solvent, the m-CPBA–EtOH system.
The four systems, FeP–EtOH, MnP–CH3CN, m-CPBA–CH2Cl2 and
m-CPBA–EtOH were compared in terms of indigo yield determined
by UV–Vis spectrophotometry (Fig. 4) and in terms of the yields of
the other reaction products, determined by 1H NMR of the total
reaction mixtures (Table 1, entries 2–5). The yield of indigo by
UV–Vis confirmed the yields obtained by NMR, indicating that no
losses by precipitation or other phenomena have occurred during
Based on conditions (e), the possibility of adding the oxidant in
an anhydrous form, as a urea–H2O2 adduct (UHP), was also consid-
ered (l). In this experiment, indigo was obtained in the same 20%
yield, confirming that there is no advantage on the use of the more
expensive adduct. It also merits to be mentioned that indigo yield
was <0.2% when a reaction was performed under conditions (e),
but in the absence of a catalyst.
During the experiments, it was observed that once precipitated
[33], indigo behaves as a quite stable compound toward oxidation
and indigo could be isolated by precipitation from the reaction
mixture by immersion of the reaction mixture into water. The dark
blue solid obtained was filtered and washed with water and then
with ethanol, and dried. Moreover, the yield of 19% obtained for
the isolated indigo matches the yield obtained by UV–Vis in the
best conditions (e). After precipitation, filtration and washing
procedures, indigo was obtained with high purity (Fig. S2). The
two filtrate fractions collected, one in water and the other in etha-
nol, were fractionated by preparative TLC after evaporation of the
corresponding solvents. As shown in the cartoon of Fig. 3, the
remaining reaction products were distributed between the two
filtrate fractions; the water fraction contains four components:
2-oxoindole (I.2), isatin (I.4), indirubin (I.5), and dimer (I.6), and
the ethanol fraction contains only the trimer (I.3). The yields of
the reaction products collected in Table 1, entry 2, were obtained
by the 1H NMR analysis of the total reaction mixture.
The efficiency of the process described was maintained even
when scale-up experiments were performed using amounts of
indole 10 and 100 times higher. The 100 times scale-up experi-
ment was repeated in 5 batches in order to test the reproducibility
of the process and g% of indigo was found in the range 19 5%.
3.2. Comparison with MnP and m-CPBA
The efficiency of catalytic indigo production in the presence of
FeP, denominated as system FeP–EtOH, was compared with two
other chemical systems, the catalytic oxidation with MnP (Fig. 1)
and the oxidation by m-CPBA. The MnP demonstrated to be the
best catalyst on a series of Mn(III)porphyrins tested in the
biomimetic oxidation of indole. The reactions were performed in
Fig. 4. UV–Vis spectra of the total reaction mixtures in DMSO for the reactions with
the FeP–EtOH and MnP–CH3CN catalytic systems and oxidations by m-CPBA in
protic and aprotic solvent (m-CPBA–EtOH and m-CPBA–CH2Cl2), in comparison with
the UV–Vis of pure solutions of indigo and indirubin (in the inset).

38
S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
the procedure. The reactions were performed in duplicate. One of
the assays was used to quantify the amount of indigo in DMSO
by UV–Vis spectrophotometry, and the other was used to obtain
the 1H NMR spectra of the total reaction mixtures in DMSO-d6.
As can be observed, the FeP–EtOH system shows a significantly
higher indigo yield (20%) and a relatively lower yield of indirubin if
compared with the other systems. In the presence of MnP–CH3CN,
indigo was obtained in a 3% yield, and under the non-catalytic
systems, m-CPBA–EtOH and m-CPBA–CH2Cl2, the indigo dye was
obtained with yields of 9% and 3%, respectively. Concerning the
non-catalytic oxidations by m-CPBA, a notable difference was
observed on indigo yield in the two solvents tested, 9% versus 3%,
for ethanol and dichloromethane, respectively.
Regarding the distribution of the other co-products, a different
tendency in the reactivity of the catalytic system based on FeP
(entry 2) is observed in comparison with the MnP catalytic system
(entry 3), while the m-CPBA oxidations (entries 4 and 5) approach
the reactivity pattern of FeP system but probably led to more
extensive over-oxidation of indigo. In the presence of MnP, the
major product was 2-oxoindole (I.2) with 64% yield and all the
other products were obtained with yields 67%. This fraction is
not so important for the other oxidative systems and relatively
higher yields of trimer (I.3), indigo (I.1) and other compounds were
observed. Low amounts of isatin (I.4) were obtained with FeP–
EtOH system under the optimized conditions (6%, entry 2).
It is worth to mention that these co-products have biological
and economical interest, since isatin and 2-oxoindole are heterocy-
clic building blocks finding important application in the synthesis
of pigments and bioactive compounds [35,36], and this synthetic
methodology can be applied to obtain derivatives. Recently, indiru-
bin is raising significant interest due to its anti-tumor and
anti-viral properties [37,38].
Further support on this reactivity can be obtained considering
the mechanism proposed for the formation of these active species
(Eqs. (1) and (2)) [23,29]; the formation of a high valent oxo-
species MnV@O species could be favored in the MnP–CH3CN
system due to the presence of: (i) less electron-withdrawing por-
phyrin nucleus and manganese as central metal that facilitate the
formation of high oxidation states, when compared to FeP; (ii) a
buffering substance such as ammonium acetate as co-catalyst,
which could successively assist the deprotonation of H2O2 (Eq.
(1)) followed by the protonation of the hydroperoxy-species to
achieve its dehydration and formation of the high valent MnV@O
species (Eq. (2)).
MIIIP þ H2O2—Hþ ! MIIIP—OOH
MIIIP—OOH þ Hþ ! MVP@O þ H2O
MIIIP—OOH ! MIVP@O þ HOꢀ
ð1Þ
ð2Þ
ð3Þ
On the other hand, in the FeP–EtOH or FeP–MeOH systems, the
presence of (i) porphyrin nucleus with more electron-
a
a
withdrawing character and iron metal center, with less electron
density, and (ii) the absence of a co-catalyst could avoid the dehy-
dration step (Eq. (2)) and the active species is possible to be the
hydroperoxy-complex FeIII–OOH; this species would be signifi-
cantly activated by hydrogen bridging in the presence of a protic
solvent, making the distal oxygen of the peroxide bridge more elec-
tron deficient and more reactive toward an electrophilic attack.
Further evidence of the two mechanisms was obtained by UV–
Vis spectrophotometry of the two systems using cis-cyclooctene as
substrate (Fig. S3, S.I) [23]; this analysis during the oxidation of
indole is not possible due to the colored nature of the reaction
products that overlap the Soret bands. In the presence of MnP–CH3-
CN, after addition of H2O2, a decrease in the Soret band at 478 nm
is observed, with a concomitant increase in a band at 429 nm, indi-
cating the formation of a second manganese complex, probably in a
different oxidation state. This is in accordance with the multiple
oxidation states during the oxo-species mechanism. On the other
hand, in the presence of FeP–MeOH [23], a second band is not
observed, indicating a different mechanism relatively to the previ-
ous one, with no changes in the oxidation state of the porphyrin
complex. This is in accordance with the proposed hydroperoxy-
species mechanism [23].
Some evidences allow inferring that a free radical mechanism is
not the main pathway in the process, such as the observed
epoxidation capabilities of both systems, as well as, the different
selectivity toward naphthalene oxidation observed for Mn and
Feporphyrins and Fenton systems [19,23,42]. Nevertheless, a radi-
cal-based mechanism has been hypothesized in hydroxylation
reactions, where the homolytic cleavage of the metal-hydroperoxy
species can generate HOꢀ radicals at low temperature (Eq. (3)),
which can initiate a radical oxidation mechanism. In order to infer
about the action of a free radical process, the use of radical trappers
was considered, namely BHT (bis-tert-butylhydroxytoluene) and
iodine, but due to their concomitant transformations in the pres-
ence of these systems, the results were not conclusive (S.I. 1). As
a result, although less probable, a pathway based on a free radical
mechanism could not be completely excluded.
3.3. Considerations on the biomimetic mechanism
The mechanistic elucidation on metalloporphyrin catalysis, as
well as, on heme enzymatic pathways has been a matter of exten-
sive discussion in the literature [23,39]. The actual consensus
about the putative active species involved in these processes
was, in general, based on a reactivity pattern [23–25,39] or consid-
ering the parallelism with non-porphyrinic (non-heme) systems
[40]. Although some reactive intermediates have been spectro-
scopically trapped and used to confirm the obtained evidences
[23], their unequivocal assignment has been controversial [23,40].
The reactivity pattern observed in previous works [23–25] and
the obtained evidences in this work suggest that the active species
formed are dependent on the catalytic system used.
In particular, the system MnP–CH3CN was efficient toward
alkane hydroxylation, performing the activation of inert saturated
C–H bonds at room temperature, while the system FeP–MeOH was
inactive toward those substrates [25]. In addition, although both
MnP–CH3CN and FeP–EtOH are able to perform the epoxidation
of alkenes [24,41], the former is more reactive toward less acti-
vated double bonds and often afforded some allylic oxidation,
while with the latter system, allylic oxidation was not observed
[23]. The two systems have also different reactivity toward aro-
matic rings, since MnP–CH3CN system afforded aromatic epoxida-
tion [19,20], while systems FeP–EtOH or FeP–MeOH typically
afforded hydroxylated compounds and derived quinones [23]. This
type of reactivity has been explained with the formation of a highly
reactive oxo-species [MV@O], when the reactions are performed
with the MnP–CH3CN system, that it is able to hydroxylate
saturated bonds, or a hydroperoxy-species [MIII–OOH], when the
reactions are performed with the FeP–EtOH system; this last spe-
cies is able to epoxidize double bonds and matches the reactivity
of m-CPBA [23,24].
In the present work, the different selectivity for indigo (I.1) or
2-oxoindole (I.2) observed for the two catalytic systems also
corroborates the action of two diverse active species. A rationaliza-
tion for the observed reactivity on indole oxidation is proposed in
Scheme 2, considering the formation of different active species,
namely
a hydroperoxy-species (Scheme 2a), an oxo-species
(Scheme 2b) or the involvement of radicals (Scheme 2c). The reac-
tion of a hydroperoxy-species with aromatic systems has been

S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
39
Scheme 2. Possible mechanisms acting during indole oxidation by H2O2 in the presence of catalytic systems FeP–EtOH and MnP–CH3CN, considering the formation of
different active species: (a) a hydroperoxy-species; (b) an oxo-species and (c) a radical species.
rationalized to occur by aromatic hydroxylation through an elec-
trophilic substitution mechanism [40,43], while the oxo-species
showed inedited efficiency for aromatic epoxidation [19,20].
Attending to the previous, when indole reacts with the hydroper-
oxy-species the 3-hydroxyl should be preferentially formed, since
the positive charge of the intermediate is stabilized at position 2
due to the proximity of nitrogen (Scheme 2a). On the other hand,
when indole reacts with the oxo-species through epoxidation,
followed by the opening of the epoxide ring by internal rearrange-
ment and restoring of aromaticity, the 2-indoxyl should be prefer-
entially formed, attending to the higher acidity of the proton at the
2-position of the pyrrolidine ring relatively to H-3 (Scheme 2b).
The formation of a cation-radical via radical reaction will also favor
the formation of 3-indoxyl (Scheme 2c).
relatively to the wild enzyme [14] can be explained by the change
in specific aminoacids in the active site frame. Aminoacids carrying
hydroxyl groups, as threonine 252 or tyrosine 96 (Fig. 5) [44], have
been replaced by aminoacids with hydrophobic side chains,
namely alanine [45] or methionine [14], respectively, it was
The 3-indoxyl can undergo oxidative coupling to form indigo
(I.1) or trimer (I.3), while the 2-indoxyl readily isomerizes to the
particularly stable keto form, the 2-oxoindole (I.2) (Scheme 2), thus
explaining the observed different reactivity/selectivity observed
with the two systems.
It is worth to mention that some associations can be withdrawn
by comparing the present results on MnP–CH3CN and FeP–EtOH
systems, with the biocatalytic oxidations of indole by different
enzymes and respective mutants reported in literature. As an
example, the indole oxidation by the wild-type P450 enzymes have
shown undetectable or very low yields of indigo and high yields of
2-oxoindole [12], and this match the reactivity of the system
MnP–CH3CN and, consequently, can be related to the action of an
oxo-species as the final oxidant species, in accordance with the
proposed for wild-type P450 enzymes [24]. On the other hand,
considering P450 mutants, the improved yields of indigo obtained
Fig. 5. Adapted representation of the active site environment of enzyme P450cam
showing the effect of proton donating aminoacids on the evolution of the
hydroperoxy to oxo-species in the catalytic cycle [39,41].
,

40
S.L.H. Rebelo et al. / Journal of Catalysis 315 (2014) 33–40
proposed that these modifications decrease the polarity and proton
donation characteristics of the residues and result in the more dif-
ficult protonation of the hydroperoxy-species avoiding its conver-
sion into oxo-species (Fig. 5). In these cases, the higher indigo
yields obtained with this type of mutants was ascribed to the
action of a Fe–OOH species as the active species [45,46] and can
be related with the FeP–EtOH system.
Furthermore, the relative high yields of indigo observed with
naphthalene dioxygenase (NDO) and related non-heme enzymes
are in accordance with the observed binding of indole to the iron
co-factor through a hydroperoxy intermediate [13,15]. In the over-
all, the results point out to the higher performance for indigo yield
when the catalytic systems act through a hydroperoxy-species.
These evidences can be seen as a key to understand and eventually
anticipate the mutants’ activity.
References
[1] E.S.B. Ferreira, A.N. Hulme, H. McNab, A. Quye, Chem. Soc. Rev. 33 (2004) 329.
[2] G. Karapetyan, K. Chakrabarty, M. Hein, P. Langer, ChemMedChem 6 (2011) 25.
[3] T.C. Barden, Top. Heterocycl. Chem. 26 (2011) 31.
[4] K. McClay, B. Wan, Y. Wang, S. Cho, J. Yu, B. Santarsiero, S. Mehboob, M.
Johnson, S. Franzblau, R. Steffan, Appl. Microbiol. Biotechnol. 97 (2013) 7151.
[5] A. de Meijere, Angew. Chem. Int. Ed. 44 (2005) 7836.
[6] H. Zollinger, Color Chemistry, Synthesis, Properties and Applications of Organic
Dyes and Pigments, third ed., Wiley-VCH, Zürich, 2003. p. 1–15.
[7] U. Bergmann, M. Gang, R. Kolhaupt, Purification of indigo, U.S. Patent 5536842,
1996.
[8] T. Maugard, E. Enaud, P. Choisy, M.D. Legoy, Phytochemistry 58 (2001) 897.
[9] B.D. Ensley, B.J. Ratzkin, T.D. Osslund, M.J. Simon, L.P. Wackett, D.T. Gibson,
Science 222 (1983) 167.
[10] T. Mouri, N. Kamiya, M. Goto, Biochem. Eng. J. 53 (2011) 229.
[11] W. Weyler, T.C. Dodge, J.J. Lauff, D.J. Wendt, Microbial production of indigo,
U.S. Patent 6303354 B1, 2001.
[12] E.M.J. Gillam, L.M. Notley, H. Cai, J. De Voss, F.P. Guengerich, Biochemistry 39
(2000) 13817.
[13] E. Carredano, A. Karlsson, B. Kauppi, D. Choudhury, R.E. Parales, J.V. Parales, K.
Lee, D.T. Gibson, H. Eklund, S. Ramaswamy, J. Mol. Biol. 296 (2000) 701.
[14] A. Çelik, R.E. Speight, N.J. Turner, Chem. Commun. (2005) 3652.
[15] K. McClay, C. Boss, I. Keresztes, R.J. Steffan, Appl. Environ. Microbiol. 71 (2005)
5476.
[16] J. Xu, O. Shoji, T. Fujishiro, T. Ohki, T. Ueno, Y. Watanabe, Catal. Sci. Technol. 2
(2012) 739.
[17] Y. Yamamoto, Y. Inoue, U. Takaki, H. Suzuki, Bull. Chem. Soc. Jpn. 84 (2011) 82.
[18] Y. Yamamoto, U. Takaki, S. Aoki, I. Hara, Process for the preparation of indigo
compounds, U.S. Patent 5112987, 1992.
[19] S.L.H. Rebelo, M.M.Q. Simoes, M.G.P.M.S. Neves, A.M.S. Silva, J.A.S. Cavaleiro,
Chem. Commun. (2004) 608.
[20] P. Costa, M. Linhares, S.L.H. Rebelo, M.G.P.M.S. Neves, C. Freire, RSC Adv. 3
(2013) 5350.
[21] M.M.Q. Simões, R. De Paula, M.G.P.M.S. Neves, J.A.S. Cavaleiro, J. Porphyrin.
Phthalocyan. 13 (2009) 589.
[22] S.M.G. Pires, S.L.H. Rebelo, M.M.Q. Simões, I.C.M.S. Santos, M.M. Pereira,
M.G.P.M.S. Neves, J.A.S. Cavaleiro, Appl. Catal. A: Gen. 449–440 (2012) 51.
[23] S.L.H. Rebelo, M.M. Pereira, M.M.Q. Simões, M.G.P.M.S. Neves, J.A.S. Cavaleiro, J.
Catal. 234 (2005) 76.
4. Conclusions
The present catalytic system embodies an important alternative
to the high-consuming harsh chemicals and high-dependence tem-
peratures of the current industrial indigo production, leading to a
significant reduction of its environmental impact. It is also a
preferable methodology comparing to natural and biocatalytic pro-
cesses, allowing a reduction of costs and the minimization of the
drawbacks related to enzyme stability and products isolation.
A careful selection on the reaction parameters, such as the
amount of oxidant, and its rate of addition to the reaction mixture
were found to be essential to improve the yield of indigo. The best
yield obtained, 20%, represents an important upgrading relatively
to the 12% reported for biocatalytic processes [16]. The superior
[24] S.L.H. Rebelo, M.M.Q. Simões, M.G.P.M.S. Neves, A.M.S. Silva, J.A.S. Cavaleiro,
A.F. Peixoto, M.M. Pereira, M.R. Silva, J.A. Paixão, A.M. Beja, Eur. J. Org. Chem.
(2004) 4778.
[25] S.L.H. Rebelo, M.M. Pereira, P.V. Monsanto, H.D. Burrows, J. Mol. Catal. A:
Chem. 297 (2009) 35.
[26] M. Linhares, S.L.H. Rebelo, M.M.Q. Simões, A.M.S. Silva, M.G.P.M.S. Neves, J.A.S.
Cavaleiro, C. Freire, Appl. Catal. A: Chem. 470 (2014) 427.
[27] R.A. Sheldon, Chem. Soc. Rev. 41 (2012) 1431.
catalytic performance of this system when compared to
a
Mn(III)porphyrin system was rationalized by different mechanistic
pathways, acting through different active species that can find con-
nections with enzymatic pathways. The Fe(III) porphyrin system is
proposed to act through a hydroperoxy Fe–OOH intermediate,
instead of an oxo-species, also confirmed by a similar reactivity
pattern with the non-catalytic oxidation in the presence of the
stoichiometric m-CPBA, although this showed a lower indigo yield.
The present work provides important elements for the design of
catalytic and eco-sustainable systems for the oxidation of indole
and production of indigo and related indolinoid compounds. Heter-
ogeneous catalytic systems based on the present system can bring
improved proficiency to the process and will be the focus of future
developments.
[28] P.G. Jessop, Green Chem. 13 (2011) 1391.
[29] M. Rocha, S.L.H. Rebelo, C. Freire, Appl. Catal. A: Gen. 460–461 (2013) 116.
[30] R. De Paula, M.A.F. Faustino, D.C.G.A. Pinto, M.G.P.M.S. Neves, J.A.S. Cavaleiro, J.
Heterocycl. Chem. 45 (2008) 453.
[31] C. Ganachaud, V. Garfagnoli, T. Tron, G. Lacazio, Tetrahedron Lett. 49 (2008)
2476.
[32] A.J. Kettle, B.M. Clark, C.C. Winterbourn, J. Biol. Chem. 279 (2004) 18521.
[33] G.A. Russell, G. Kaupp, J. Am. Chem. Soc. 91 (1969) 3851.
[34] J. Rosenthal, L.L. Chng, S.D. Fried, D.G. Nocera, Chem. Commun. (2007) 2642.
[35] P. Limpachayaporn, M. Schäfers, O. Schober, K. Kopka, G. Haufe, Bioorg. Med.
Chem. 21 (2013) 2025.
[36] A.H. Abadi, S.M. Abou-Seri, D.E. Abdel-Rahman, C. Klein, O. Lozach, L. Meijer,
Eur. J. Med. Chem. 41 (2006) 296.
[37] V. Myrianthopoulos, M. Kritsanida, N. Gaboriaud-Kolar, P. Magiatis, Y.
Ferandin, E. Durieu, O. Lozach, D. Cappel, M. Soundararajan, P.
Filippakopoulos, W. Sherman, S. Knapp, L. Meijer, E. Mikros, A. Skaltsounis,
ACS Med. Chem. Lett. 4 (2013) 22.
[38] E.C. Davis, A. Heredia, R. Redfield, Use of indirubin and its derivatives in the
treatments of hiv infection and heart failure, WO2007033208A3, 2009.
[39] M. Oszajca, A. Drzewiecka-Matuszek, A. Franke, D. Rutkowska-Zbik, M.
Brindell, M. Witko, Graz´yna Stochel, R. van Eldik, Chem. Eur. J. 20 (2014) 2328.
[40] A. Thibon, V. Jollet, C. Ribal, K. Sénéchal-David, L. Billon, A.B. Sorokin, F. Banse,
Chem. Eur. J. 18 (2012) 2715.
Acknowledgments
The authors thank the Fundação para a Ciência e a Tecnologia
(FCT, Portugal), the European Union, QREN, FEDER, COMPETE,
for funding REQUIMTE (PEst-C/EQB/LA0006/2011 and NORTE-
07-0124-FEDER-000067-nanochemistry) and QOPNA (project
PEst-C/QUI/UI0062/2013; FCOMP-01-0124-FEDER-037296), and
the Portuguese National NMR Network, also supported by funds
from FCT.
´
[41] M.E. Lipinska, S.L.H. Rebelo, C. Freire, J. Mater. Sci. 49 (2014) 1494.
[42] E. Boyland, P. Sims, J. Chem. Soc. (1953) 2966.
[43] Y. Watanabe, in: K.M. Kadish, K.M. Smith, R. Guilard (Eds.), The Porphyrin
Handbook, vol. 4, Academic Press, San Diego, 2000, p. 97.
[44] T.L. Poulos, B.C. Finzel, A.J. Howard, Biochem. 25 (1986) 5314.
[45] K. Makamura, M.V. Martin, F.P. Guengerich, Arch. Biochem. Biophys. 395
(2001) 25.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2014.04.012.
[46] K.P. Vatsis, M.J. Coon, Arch. Biochem. Biophys. 397 (2002) 119.