Pyrroles as antioxidants: Solvent effects and the nature of the attacking radical on antioxidant activities and mechanisms of pyrroles, dipyrrinones, and bile pigments

Article abstract of DOI:10.1021/jo8005073

(Chemical Equation Presented) The absolute rate constants, k_inh, and stoichiometric factors, n, of pyrroles, 2-methyl-3-ethylcarboxy-4,5-di-p- methoxyphenylpyrrole, 6, 2,3,4,5-tetraphenylpyrrole,7, and 2,3,4,5-tetra-p- methoxyphenylpyrrole, 8, compared to the phenolic antioxidant, di-tert-butylhydroxyanisole, DBHA, during inhibited oxidation of cumene initiated by AIBN at 30°C gave the relative antioxidant activities (k _inh) DBHA > 8 > 7 > 6 and n = 2, whereas in styrene, 8 > DBHA. These results are explained by hydrogen atom transfer, HAT, from the N-H of pyrroles to ROO^? radicals. The k_inh values in styrene of dimethyl esters of the bile pigments of bilirubin ester (BRDE), of biliverdin ester (BVDE), and of a model compound (dipyrrinone, 1) gave k _inh in the order pentamethylhydroxychroman (PMHC) BRDE > 1 > BVDE. These antioxidant activities for BVDE and the model compound, 1, and PMHC dropped dramatically in the presence of methanol due to hydrogen bonding at the pyrrolic N-H group. In contrast the k_inh of BRDE increased in methanol. We now show that pyrrolic compounds may react by HAT, proton-coupled electron transfer, PCET, or single electron transfer, SET, depending on their structure, the nature of the solvent, and the attacking radical. Compounds BVDE and 1 react by the HAT or PCET pathway (HAT/PCET) in styrene/chlorobenzene with ROO^? and with the DPPH^? radical in chlorobenzene according to N-H/N-D k_H/k_D of 1.6, whereas the DKIE with BRDE was only 1.2 with ROO^?. The antioxidant properties of polypyrroles of the BVDE class and model compounds (e.g., 1) are controlled by intramolecular H bonding which stabilizes an intermediate pyrrolic radical in HAT/PCET. According to kinetic polar solvent effects on the monopyrrole, 8, and BRDE, which gave increased rates in methanol, some pyrrolic structures are also susceptible to SET reactions. This conclusion is supported by some calculated ionization potentials. The antioxidant mechanism for BRDE with peroxyl radicals is described by the PCET reaction. Experiments using the 2,6-di-tert-butyl-4- (4′-methoxyphenyl)phenoxyl radical (DBMP^?) showed this to be a better radical to monitor HAT activities in stopped-flow kinetics compared to the use of the more popular DPPH^? radical.

Full text of DOI:10.1021/jo8005073

Pyrroles As Antioxidants: Solvent Effects and the Nature of the
Attacking Radical on Antioxidant Activities and Mechanisms of
Pyrroles, Dipyrrinones, and Bile Pigments
Patricia D. MacLean,† Erin E. Chapman,‡ Sarah L. Dobrowolski,† Alison Thompson,‡ and
L. Ross C. Barclay*,†
Department of Chemistry, Mount Allison UniVersity, SackVille, New Brunswick, Canada E4L 1G8, and
Department of Chemistry, Dalhousie UniVersity, Halifax, NoVa Scotia, Canada, B3H 4J3
rbarclay@mta.ca
ReceiVed March 4, 2008
The absolute rate constants, kinh, and stoichiometric factors, n, of pyrroles, 2-methyl-3-ethylcarboxy-4,5-
di-p-methoxyphenylpyrrole, 6, 2,3,4,5-tetraphenylpyrrole,7, and 2,3,4,5-tetra-p-methoxyphenylpyrrole,
8, compared to the phenolic antioxidant, di-tert-butylhydroxyanisole, DBHA, during inhibited oxidation
of cumene initiated by AIBN at 30 °C gave the relative antioxidant activities (kinh) DBHA > 8 > 7 >
6 and n ) 2, whereas in styrene, 8 > DBHA. These results are explained by hydrogen atom transfer,
HAT, from the N-H of pyrroles to ROO• radicals. The kinh values in styrene of dimethyl esters of the
bile pigments of bilirubin ester (BRDE), of biliverdin ester (BVDE), and of a model compound
(dipyrrinone, 1) gave kinh in the order pentamethylhydroxychroman (PMHC) . BRDE > 1 > BVDE.
These antioxidant activities for BVDE and the model compound, 1, and PMHC dropped dramatically in
the presence of methanol due to hydrogen bonding at the pyrrolic N-H group. In contrast the kinh of
BRDE increased in methanol. We now show that pyrrolic compounds may react by HAT, proton-coupled
electron transfer, PCET, or single electron transfer, SET, depending on their structure, the nature of the
solvent, and the attacking radical. Compounds BVDE and 1 react by the HAT or PCET pathway (HAT/
PCET) in styrene/chlorobenzene with ROO• and with the DPPH• radical in chlorobenzene according to
N-H/N-D kH/kD of 1.6, whereas the DKIE with BRDE was only 1.2 with ROO•. The antioxidant
properties of polypyrroles of the BVDE class and model compounds (e.g., 1) are controlled by
intramolecular H bonding which stabilizes an intermediate pyrrolic radical in HAT/PCET. According to
kinetic polar solvent effects on the monopyrrole, 8, and BRDE, which gave increased rates in methanol,
some pyrrolic structures are also susceptible to SET reactions. This conclusion is supported by some
calculated ionization potentials. The antioxidant mechanism for BRDE with peroxyl radicals is described
by the PCET reaction. Experiments using the 2,6-di-tert-butyl-4-(4′-methoxyphenyl)phenoxyl radical
(DBMP•) showed this to be a better radical to monitor HAT activities in stopped-flow kinetics compared
to the use of the more popular DPPH• radical.
Introduction
ROS and protective defense mechanisms, oxidatiVe stress
occurs.1a The potential damage is commonly initiated by
oxygen-centered radicals resulting in free radical chain autoxi-
dation, which is implicated in a number of degenerative diseases
in humans.1b As a result, there is a continuing interest by
scientists and the public on the control of oxidative stress by
antioxidants. Phenolic antioxidants, such as vitamin E in our
diet, deactivate oxygen radicals by reductive processes, and also
Aerobic organisms are continuously exposed to reactive
oxygen species (ROS), and when an imbalance occurs between
* To whom correspondence should be addressed. Tel: (506) 364-2369. Fax:
(506) 364-2313.
† Mount Allison University.
‡ Dalhousie University.
10.1021/jo8005073 CCC: $40.75
Published on Web 07/29/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 6623–6635 6623

MacLean et al.
SCHEME 1
The structural factors controlling the mechanisms and anti-
oxidant activities of phenols are very well understood.6 In
contrast, the antioxidant mechanisms and activities of pyrroles
and pyrrolic derivatives are still open to question, although the
naturally occurring tetrapyrrolic bilirubin was discovered to
possess antioxidant activity 50 years ago.7 More recently,
biliverdin and bilirubin have been studied quite extensively for
their antioxidant properties, and several different antioxidant
mechanisms have been advanced to account for these proper-
ties.8
Previously, we have discussed the antioxidant capabilities of
polypyrroles, bilirubin, biliverdin, and some model compounds
in organic solvents,8f micelles, and lipid bilayers.8e Herein we
report significant advancements in understanding the antioxidant
properties/mechanisms of the bile pigments bilirubin and
biliverdin. Indeed, using pyrroles, dipyrrins, the methyl esters
of bilirubin and biliverdin, and a dipyrrinone (see Figure 1 for
the structures of all molecules used), we have (i) established
the importance of internal hydrogen bonding for antioxidant
ability; (ii) used N-H DKIE measurements to investigate the
role of hydrogen atom transfer (HAT) and proton-coupled
electron transfer mechanism (PCET) transfer; (iii) calculated
ionization potentials (IPs) as an indication of the tendency for
single electron transfer (SET) reactions; (iv) demonstrated the
unreliability of using DPPH• in stopped-flow kinetics typically
used to determine hydrogen atom donating ability of antioxi-
dants; and (v) demonstrated, for the first time, the superiority
of using DPMB• to this effect. Specifically, the experimental
investigations reported herein include (A) determination of the
absolute rate constants, kinh, with peroxyl radicals and the
number of peroxyl radicals trapped, n, per molecule of the pyr-
rolic compounds using inhibition of oxygen uptake kinetics; (B)
kinetic deuterium isotope effects of N-D versus N-H for
reactions of the bile pigments and a model dipyrrinone with
peroxyl radicals and with the nitrogen-centered radical, diphe-
nylpicrylhydrazyl (DPPH•) to elucidate reaction mechanisms;
(C) determination of the effects of variation of the solvent
polarity on these antioxidant properties; and (D) determination
of the rate constants of the antioxidants with DPPH• in a
nonpolar and polar solvent using stopped-flow kinetics, as well
as an evaluation of the use of the synthetic 2,6-di-tert-butyl-4-
(4′-methoxyphenyl)phenoxyl, DBMP•, radical compared to
DPPH• in stopped-flow kinetics. Calculated IPs are reported as
an aid in evaluating the role of HAT or SET reactions as
contributing mechanisms of antioxidant activity.
there is a continuing interest in synthetic antioxidants of high
antioxidant activity.2 In contrast, pyrrole-containing molecules
provide a possible in ViVo protection against free radical damage,
and the possible beneficial effects of the bile pigment bilirubin
in this regard was pointed out in 1990.3a Such natural pyrroles
are of particular interest because they are formed continuously
during the catabolism of hemoglobin, and more recently,
bilirubin was discovered to contribute to the total antioxidant
capacity of human blood plasma.3b We now present results on
the relative antioxidant activities and antioxidant mechanisms
of some synthetic pyrrole-containing molecules and the natural
pigments compared to some conventional phenolic antioxidants.
Phenolic antioxidants deactivate oxygen radicals by a process
that generally involves transfer of hydrogen atoms, resulting in
a stabilized phenolic radical which does not continue the
oxidative chain process (Scheme 1). Pyrroles also contain an
active hydrogen atom (N-H), and by analogy, one might expect
that they would also act as hydrogen atom transfer (HAT) agents
and possess antioxidant activity, especially in compounds
having electron-supplying groups positioned to stabilize
intermediate radicals as is known for active phenolic anti-
oxidants (Scheme 1).
The obvious structural differences between phenols and
pyrroles, where the former have the active phenolic O-H donor
group attached to the aromatic ring in contrast to pyrroles where
the N-H group is imbedded as part of the aromatic system,
are expected to result in very significant differences when
considering their antioxidant properties. In addition, one of the
very surprising observations is the resulting electronic structure
of the pyrrole radical arising from hydrogen atom transfer.4 The
radical has been observed to possess the electronic structure 2
in its lower energy state rather than the “expected” 1, as
indicated by both experimental and theoretical data.5 This is
one of the factors that determines the antioxidant mechanism
and activity of pyrroles.
Results
I. Antioxidant Activities, kinh, and Stoichiometric Fac-
tors, n, of Pyrroles during Inhibited Peroxidation of Cumene
and Styrene. Autoxidation and its inhibition by antioxidants
are reviewed in various articles.6b–d The process commonly
involves peroxyl radicals in initiation, propagation, and termina-
tion steps. For experiments conducted in vitro, this process is
usually initiated by azo-initiators such as AIBN. The mechanism
of uninhibited autoxidation of typical substrates, cumene (a)
and styrene (b), is outlined as follows:
(1) (a) OxidatiVe Stress: Oxidants and Antioxidants; Sies, H., Ed.; Academic
Press: London, 1991. (b) Halliwell, B.; Gutteridge, J. M. C.; Free Radicals in
Biology and Medicine, 3rd ed.; Oxford University Press: New York, 1999.
(2) (a) Foti, M.; Johnson, E. R.; Vinquist, M. R.; Wright, J. S.; Barclay,
L. R. C; Ingold, K. U. J. Org. Chem. 2002, 67, 5190-5196; Foti, M. C. Barclay,
L. R. C.; Ingold, K. U. J. Am. Chem. Soc. 2002, 124, 12881-12888. (b) Pratt,
D. A.; DiLabio, G. A.; Brigati, G.; Pedulli, G. F.; Valgimigli, L. J. Am. Chem.
Soc. 2001, 123, 4625–4626. (c) Barclay, L. R. C.; Vinqvist, M. R.; Mukai, K.;
Itoh, K.; Morimoto, H. J. Org. Chem. 1993, 58, 7416–7420.
(3) (a) McDonagh, A. F. Clinics Perinatol. 1990, 17, 359–369. (b) MacLean,
P. D.; Drake, E. C.; Barclay, L. R. C. Free Radical Biol. Med. 2007, 43, 600–
609.
(4) It would be reasonable to expect that the loss of a hydrogen atom from
pyrrole would result in a nitrogen radical with the unpaired electron in a σ orbital
(structure 1, Scheme 1) leaving six π electrons in the aromatic pyrrole ring.
(5) (a) Bacskay, G. B.; Martoprawiro, M.; MacKie, J. C Chem. Phys. Lett.
1998, 290, 391–398. (b) Gianola, A. J.; Ichino, T.; Hoenigman, R. L.; Kato, S.;
Bierbaum, V. M.; Lineberger, W. C. J. Phys. Chem. A 2004, 108, 10326–10335.
(6) (a) Burton, G. W.; Doba, T.; Gabe, E. J.; Hughes, L.; Lee, F. L.; Prasad,
L.; Ingold, K. U. J. Am. Chem. Soc. 1985, 107, 7053–7065. (b) Barclay, L. R. C.;
Vinqvist, M. R. The Chemistry of Phenols; John Wiley and Sons, Ltd.: Chichester,
England, 2003. (c) Burton, G. W.; Ingold, K. U. Acc. Chem. Res. 1986, 19,
194–201. (d) Ingold, K. U. Acc. Chem. Res. 1969, 2, 1–9. (e) Ingold, K. U.
Chem. ReV. 1961, 61, 563–589.
(7) Bernhard, K.; Ritzel, G.; Steiner, K. U. HelV. Chim. Acta 1954, 3, 306–
313.
6624 J. Org. Chem. Vol. 73, No. 17, 2008

Pyrroles As Antioxidants
Initiation:
[RsH] × Ri1⁄2
-d[O2]
dt
kp
(2kt)1⁄2
)
×
(6)
O2
9
[(CH3)2C(CN)N)]2
9
∆8
8 R-OO•(rate ) R1)
(1)
In general, an antioxidant molecule (typically a phenolic
antioxidant, ArOH) “traps” peroxy radicals by H atom donation
with a rate constant of kinh forming a relatively stable radical
(see eqs 7 and 8).
Chain Propagation:
kp
R-OO• + (CH3)2CHPh
9
8 R-OO-H + (CH3)2-C•Ph (2a)
kinh
R′OO• + ArOH 8 R′OOH + ArO•
(7)
(8)
(CH3)2-C•Ph + O2 f (CH3)2-C(OO•)Ph()R′OO•) (3a)
R - OO• + CH2)CHPh f R-OOCH2C•HPh (2b)
R-OOCH2C•HPh + O2 f R-OOCH2CH(OO•)Ph()R′OO•)
R′OO• + ArO• fast8 nonradical products
(3b)
The oxygen uptake is suppressed for a length of time, τ,
R′-OO• + (CH3)2CHPh f R′-OO-H + (CH3)2-C•Ph (4a)
during an induction or inhibition period, and is given by
R′-OO• + CH2)CHPh f R′-OOCH2C•HPh
(4b)
(5)
n × [ArOH]
τ)
(9)
Ri
Chain Termination:
2kt
2R′-OO• 98 nonradical products + O2
where n, the stoichiometric factor, is the number of peroxyl
radicals trapped per molecule of antioxidant. In general, the n
for phenols is 2.0.6
Equation 6, where RsH ) the substrate, represents the kinetic
expression which generally applies to such uninhibited autoxi-
dations.
During the induction period, the rate of oxygen uptake is
given by eq 10:6
FIGURE 1. Pyrroles, phenols,and radicals used in this study.
J. Org. Chem. Vol. 73, No. 17, 2008 6625

MacLean et al.
-d[O2]
dtinh
kp
Ri
)
× [RsH] ×
(10)
kinh
n[ArOH]
Integrating eq 10 gives the incremental oxygen uptake, as shown
in eq 11:
kp[RH]
-∆[O2]t
)
ln(1 - t ⁄ τ)
(11)
kinh
A linear plot of -∆[O2]t versus ln(1 - t/τ) will give a slope of
kp [RH]/kinh from which the absolute rate constant for inhibition,
kinh, is obtained using a known value for the substrate kp.
Alternately, where there is sufficient oxygen uptake at the
beginning of the induction period, the initial, -d[O2]/dtinh can
be measured and kinh calculated using eq 10. In order for a
substance to suppress effectively the oxidation of a substrate,
the kinh must be substantially greater than the rate constant of
propagation, kp of the substrate (eq 10). Application of these
kinetic methods for rate constant calculations requires significant
kinetic chain lengths (e.g., υ > 5) throughout the induction
period to provide measurable oxygen uptake.
In order to quantify the antioxidant capability of the pyrrolic
and phenolic molecules shown in Figure 1, cumene and styrene
were chosen as oxidizable substrates based on the following:
(i) cumene has a low kp of 0.18 M-1 s-1 at 30 °C,9a and
therefore, relatively weak antioxidants give well-defined induc-
tion periods for measurements;9b (ii) styrene has a kp of 41 M-1
s-1at 30 °C,6a allowing kinh values for more efficient antioxidants
to be measured; and (iii) styrene forms a polyperoxide in an
irreversible reaction as it does not contain a readily abstractable
hydrogen atom, and therefore, it does not undergo complicating
side reactions.10 Solutions of cumene or styrene of convenient
concentrations in an inert solvent (e.g., chlorobenzene) were
used to permit measurements of oxygen uptake during inhibition
periods.
FIGURE 2. Oxygen uptake profile of the oxidation of cumene, 1.77
M, in chlorobenzene initiated with AIBN, 21 mM, at 30 °C: (a)
uninhibited; (b) inhibited by 7, 4.00 µM; (c) inhibited by DBHA, 4.06
µM; (d) inhibited by 8, 3.97 µM.
TABLE 1. Antioxidant Activities, kinh, and Stoichiometric Factors,
n, During Peroxidation (a) of Cumene, Inhibited by BHT, 6, 7, and
8; (b) of Styrene Inhibited by DBHA or 8
b
substratea
inhibitor, M × 106 kinh M-1 s-1 × 10-3 nc
(a) cumene, 1.64-1.68 M BHT
5.2d
2*
6 4.99
7 4.00-6.34
8 3.94, 3.97
DBHA
8 2.63-2.67
(0.93)
(1.30)
10.8
2.0
1.8
2.2
2*
(b) styrene, 0.87 M
161d
254
3.2
a Reactions were initiated by AIBN (21.0-22.5 mM) in C6H5Cl at 30
°C. b Rate constants were calculated from plots of -∆[O2]t versus ln(1
- t/τ) using a kp ) 0.18 M-1 s-1 for cumene and 41 M-1 s-1 for
styrene, except for 6, calculated from the initial inhibited oxygen uptake
using eq 10. Results are from at least three determinations with error
limits less than 12%, except for 8 in (a). c Stoichiometric factors were
calculated from n ) Ri × τ/[antioxidant], where Ri ) 2[ArOH]/τ and
ArOH ) DBHA or PMHC, and n* for DBHA and BHT was taken to
be 26. The induction period for 7 was estimated from the intersection of
initial rate and the final rate, not shown in Figure 2. d Values of kinh and
n are taken from ref 8f.
Quantitative kinetic studies of cumene and styrene autoxi-
dation were carried out to determine the antioxidant activity of
our pyrroles. To our knowledge, there are no reported quantita-
tive determinations of antioxidant activities of simple monopy-
rroles, and we found 2,4-dimethylpyrrole not to possess
antioxidant activity.8f However, the stoichiometric factor of
2,3,4,5-tetraphenylpyrrole, 7, in cumene was reported 50 years
ago by Hammond et al.11 Consequently, we started our studies
with a comparison of the antioxidant activities of 7 and the
tetramethoxy derivative, 8, with the commercial antioxidant,
DBHA (11), in cumene. The profiles for the inhibition of oxygen
uptake by pyrroles 7 and 8 show that 8 is a superior antioxidant
compared to 7 but clearly less active than DBHA (see Figure
2). Under these conditions, DBHA reduced the oxygen uptake
almost completely; therefore, its kinh was not calculated, but it
gave reliable determinations of the rate of chain initiation, Ri.
Our kinetic methods now provide an actual antioxidant
activity of 7 along with two other pyrroles, 6 and 8, in cumene
as shown in Table 1(a). The absolute rate constants, kinh, and
stoichiometric factors, n, were compared with those for the
commercial antioxidants BHT. Compound 8 is the most
effective monopyrrole antioxidant studied in cumene, and the
order of antioxidant activity found is 8 > BHT > 7 > 6.
Stoichiometric factors of all three monopyrroles were ap-
proximately 2, indicating that these pyrroles trap two peroxyl
radicals per molecule. The kinh values for 6 and 7, shown in
parentheses in Table 1(a) are overestimated due to the combina-
tion of reactions by peroxyl radicals and by the antioxidant (eqs
5 and 7) because 6 and 7 are relatively weak antioxidants (see
note/reference).12
(8) (a) Neuz˘il, J.; Stocker, R. J. Biol. Chem. 1994, 24, 16712–16719. (b)
Stocker, R.; Glazer, A. N.; Ames, B. N. Proc. Natl. Acad. Sci. U.S.A. 1987, 84,
5918–5922. (c) Stocker, R.; Yamamoto, Y.; McDonagh, A. F.; Glazer, A. N.;
Ames, B. N. Science 1987, 235, 1043–1046. (d) Stocker, R.; Peterhans, E.
Biochim. Biophys. Acta 1989, 1002, 238–244. (e) Hatfield, G. L.; Barclay,
L. R. C. Org. Lett. 2004, 6, 1539–1542. (f) Chepelev, L. L.; Beshara, C. S.;
MacLean, P. D.; Hatfield, G. L.; Rand, A. A.; Thompson, A.; Wright, J. S.;
Barclay, L. R. C. J. Org. Chem. 2006, 71, 22–30. (g) Dudnik, L. B.; Khrapova,
N. G. Membr. Cell. Biol. 1998, 12, 233–240. (h) Adhikari, S. K.; Guha, S. N.;
Gopinathan, C. Int. J. Chem. Kinetics 1994, 26, 903–912. (i) Ihara, H.; Aoki,
Y.; Hashizume, N; Aoki, T.; Yoshida, M.; Osawa, S Clin. Chem. Enzymol.
Commun. 1998, 8, 31–36. (j) Baran˜ano, D. E.; Rao, M.; Ferris, C. D.; Snyder,
S. H. Proc. Natl. Acad. Sci. U.S.A. 2000, 99, 16093–16098. (k) Tomaro, M. L.;
del Batlle, A. M. C. Int. J. Biochem. Cell Biol. 2002, 34, 216–220.
(9) (a) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1965, 43, 2729–2736.
(b) Horswell, E. C.; Howard, J. A.; Ingold, K. U. Can. J. Chem. 1966, 44, 985–
991.
(10) Burton, G. W.; Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 6472–6477.
(11) The stoichiometric value for 7 was reported to be 1.6: Hammond, G. S.;
Boozer, C. E.; Hamilton, C. E.; Sen, J. N. J. Am. Chem. Soc. 1955, 77, 3238-
3244, and also 3.3: Boozer, C. E.; Hammond, G. S.; Hamilton, C. E.; Sen, J. N.
J. Am. Chem. Soc. 1955, 77, 3233-3237. The latter value was determined to be
the same as that of BHT, known to be 2,10 so this value for 7 should also be
taken as 2.
The relative antioxidant profiles in styrene of 8 and DBHA
along with PMHC are shown in Figure 3. The inhibited oxygen
uptake profiles in this substrate show that the relative inhibition
properties of 8 and DBHA are switched compared to those in
cumene (see Figure 2). The antioxidant activities, kinh, values
given in Table 1(b) indicate 8 to be 1.7 times more reactive
than DBHA in styrene. The dependence of the relative inhibition
6626 J. Org. Chem. Vol. 73, No. 17, 2008

Pyrroles As Antioxidants
Two other compounds, 4 and 5, were synthesized and tested
to further establish the relevance of the dipyrrinone model, 1.
The dipyrrin, 4, lacks a second N-H group but possesses an
electron-donating methoxy group in the pyrrolic ring. It did not
exhibit antioxidant activity. While compound 5 contains two
N-H groups, it was not an antioxidant under the conditions
used. The lack of HAT activities for compounds 2-5 is not
unexpected, as these results support the earlier conclusion
regarding the importance of internal hydrogen bonding to
stabilize the pyrrolic radical in order for a dipyrrinone to possess
significant antioxidant activity8f in a HAT reaction (see Discus-
sion).
(b) Deuterium Kinetic Isotope Effects on Dipyrrinone,
1, and Dimethyl Esters of Bilrubin and Biliverdin with
Peroxyl Radicals and DPPH•. The study of antioxidant
mechanisms of phenols using deuterium kinetic isotope effects
(DKIE) was begun in the 1960s by Howard and Ingold,14 and
their experiments provided the first clear evidence for the HAT
antioxidant mechanism for phenolic antioxidants. In our study
reported herein, we employed DKIE with peroxyl and DPPH•
radicals in an attempt to provide insight into the antioxidant
mechanism of mono- and polypyrrolic molecules.
FIGURE 3. Oxygen uptake profile of the oxidation of styrene, 0.87
M, in chlorobenzene initiated with AIBN, 21 mM, at 30 °C: (a)
uninhibited; (b) inhibited by DBHA, 2.69 µM; (c) inhibited by 8, 2.64
µM; (d) inhibited by PMHC, 2.52 µM.
efficiencies of phenolic antioxidants on the substrate used was
pointed out decades ago,13a and the results herein indicate again
the importance of the substrate when determining antioxidant
properties. In addition, the relative reactivity of a pyrrole may
also depend on the structure of the peroxyl radical. Tertiary
peroxyl radicals (e.g., from cumene, eq 3a) are generally known
to exhibit lower reactivity compared to that of secondary peroxyl
radicals13b (e.g., from styrene, eq 3b), and so relative antioxidant
properties must be reported with reference to a common
substrate. The remaining experiments with peroxyl radicals were
carried out using styrene as the substrate. These results show
that an activated pyrrole (e.g., 8) possesses HAT activity of the
N-H group at least as great as that of a typical phenolic
antioxidant like DBHA. Such results provide a useful compari-
son with the antioxidant properties of the dipyrrinones and the
bile pigments.
II. Antioxidant Activities and Mechanisms of Dipyrrino-
nes and Bile Pigments (a) Applications Using Synthetic
Model Compounds. Model compounds of the bile pigments,
such as the dipyrrinones 1-3 (Figure 1), contain important
structural features present in both biliverdin and bilirubin and
are expected to mimic some of the important physical-organic
properties of these bile pigments without exhibiting their marked
sensitivity to laboratory light. Earlier we found that 1 has
antioxidant properties comparable to those of BVDE and
proposed that free N-H groups are required in both rings of a
dipyrrinone for significant antioxidant activity since a dipyr-
rinone with both nitrogen atoms protected with methyl groups
lacked such activity.8f To test this more specifically, we have
now synthesized two other dipyrrinones containing single
N-CH3 groups: 2, in which the pyrrolic ring contains the
N-CH3, and 3, where the lactam ring contains a N-CH3 group.
We now find that neither 2 nor 3 exhibits significant antioxidant
activity under conditions used for 1.8f
A. Deuterium Kinetic Isotope Effect with Peroxyl Radi-
cals. Kinetic studies of H atom, kH, and D atom, kD, abstractions
by peroxyl radicals under aerobic conditions were determined
using values for the inhibited oxidation of styrene initiated by
AIBN in chlorobenzene. The rate constants, kH and kD, and the
DKIE (kH/kD) of 1, BVDE, and BRDE are reported in Table 2.
Both compounds 1 and BVDE exhibited a DKIE of 1.6. The
DKIE for BRDE was only 1.2.
TABLE 2. Rate Constantsa,kH and kD (M-1 s-1 × 10-4) (N-H/
N-D of Pyrrolic Compounds), for Hydrogen Atom and Deuterium
Atom Abstraction during the Inhibited Peroxidation of Styreneb
compoundc
kH
kD
kH/kD
1
13
16
26.6
8
10
22.5
1.6
1.6
1.2
BVDE
BRDE
a Rate constants are an average of at least three determinations, and
error limits were less than 25%. b Oxidation of styrene (0.87 M) in
C6H5Cl was initiated by AIBN (21.4-22.4 mM) at 30 °C in C6H5Cl
containing trace amounts of H2O or D2O to ensure N-D groups do not
undergo back exchange. c Compound concentrations ranged from 4.21 to
11.5 µM.
B. Deuterium Kinetic Isotope Effect with DPPH• Radicals.
Kinetic studies of H atom and D atom abstractions by DPPH•
in chlorobenzene under anaerobic conditions were monitored
by following the pseudo-first-order decay of the 526 nm
absorption of DPPH• radicals, kobs. Using pseudo-first-order
kinetics, the concentration of the antioxidant [A-H] was between
8 and 2200 times larger than that of DPPH•. At least five
different concentrations of antioxidant were employed, and at
least three separate measurements of kobs were made at each
concentration. The second-order rate constants, k2, were deter-
mined by plots of the kobs versus [A-H] from the following
equation:
(12) Methods of estimating the kinh of weak antioxidants have analyzed the
data by relating the ratio of initial rates of oxidation of a substrate both in the
absence and the presence of antioxidant in terms of the composite rate constants
for inhibition, kinh, and termination by peroxyl radicals by a relationship, nkinh
[AH]0 /(Ri2kt)1/2, assuming n ) 2. (See, for example: Amorati, R.; Pedulli, G. F.;
Valgimigli, L.; Attanasi, O. A.; Filippone, C.; Fiorucci, C.; Saladino, R. J. Chem.
Soc., Perkin Trans. 2 2001, 2142–2146. Using this relationship and a reported
2kt of 0.035 × 106 M-1 s-1 in cumene at 30 °C (Howard, J. A. AdV. Free Radical
Chem. 1972, 4, 49–73.), we find that the kinh of 6 given in Table 1 is overestimated
by about 30% and that of 7 is overestimated by 6%. Since these estimates depend
on the 2kt used in the calculations, the values used in Table 1 or calculated
values are not claimed to be reliable but the relative order of antioxidant activities
shown in the text are correct.
kobs ) k0 + k2[A-H]
(12)
Least-squares fitting of the plots gave R2 values >0.98 for all
compounds except for DBHA, which gave R2 ) 0.90.
Rate constants of H atom, kH, and D atom, kD, and DKIE
(kH/kD) of model compound 1, DBHA, and BRDE in chlo-
(13) (a) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1964, 42, 2324–2333.
(b) Korcek, S.; Chenier, J. H. B.; Howard, J. A.; Ingold, K. U. Can. J. Chem.
1972, 50, 2285–2297.
J. Org. Chem. Vol. 73, No. 17, 2008 6627

MacLean et al.
TABLE 3. Rate Constantsa, kH and kD(M-1 s-1) (N-H/N-D), for
Hydrogen and Deuterium Atom Abstraction by DPPH• (1.4-1.5 ×
10-4 M) in C6H5Clb
SCHEME 2
compoundc
kH
2.6
153
3.63e
kD
kH/kD
1
1.6
141
0.41
1.6
(1.1)d
9
activities of available active antioxidants used in this investiga-
tion, with the object of determining if HBA solvents affect the
antioxidant activities and mechanisms of these compounds
especially in view of the fact that the two esters, BVDE and
BRDE, of the bile pigments gave quite different results using
kinetic isotope effects (Table 2).
BRDE
DBHA
a Rate constants at 30 °C are an average of at least two experimental
runs, excluding DBHA (one experiment), using pseudo-first-order
stopped-flow methods with an R2 value greater than or equal to 0.98.
b C6H5Cl contained trace amounts of H2O or D2O (see Table 2 footnote
b). c Compound concentrations ranged from 0.31 to 0.0011 M and
DPPH• 1.4-1.5 × 10-4 M giving a [DPPH•] to [A-H] ratio of 1:8 to
1:2200. d This value is uncertain due to the typical experimental varia-
A. Kinetic Solvent Effects (KSE) with Peroxyl Radicals.
Quantitative kinetic studies of styrene autoxidation with and without
the presence of methanol were carried out to obtain the KSE of a
HBA solvent on the antioxidant activity of monopyrrole 8,
dipyrrinone 1, BVDE, and BRDE compared to that of PMHC
(Table 4). In the presence of a HBA solvent, the kinh of PMHC
dropped 7-fold, a common trend with phenolic antioxidants in HBA
solvents.17 In contrast, the effect of a HBA/ionizing solvent on
the antioxidant activity of the mono- and polypyrroles varied
considerably. While the antioxidant activity of 8 was only slightly
suppressed by methanol from 25.4 × 104 to 17.5 × 104 M-1
s-1(Table 4), the antioxidant activities of both 1 and BVDE were
completely suppressed in the presence of methanol and, surpris-
ingly, the antioxidant activity of BRDE activity increased from
22.5 × 104 to 37.9 × 104 M-1 s-1 (see ref 18).
tions in rate constants. e In hexane, we obtained the value kH ) 22.4
17e
M-1 s-1, in agreement with the reported value, 22.6 M-1 s-1
.
robenzene are reported in Table 3. Compound 1, a dipyrrinone
which mimics the terminal pyrroles of biliverdin and bilirubin,
exhibited a DKIE of 1.6. The hindered phenol, DBHA, showed
a high DKIE of 9 as previously reported for other phenols.14–16
While the plots according to eq 12 for deuterated (N-D) BRDE
were distinguished from those of the N-H compound (see
Supporting Information), the calculated DKIE (Table 3) was
not significantly different from experimental variations. The
reaction of BVDE with DPPH• by stopped-flow studies could
not be quantified using pseudo-first-order kinetics as the UV
spectrum of BVDE interfered with the absorbance of DPPH•
used in the measurements.
The DKIE of the selected model dipyrrinone, 1, was
remarkably consistent, 1.6, with that of BVDE, showing that 1
is an appropriate model for these free radical reactions and
indicates participation of the pyrrolic N-H group in the
mechanism (see Discussion). The lower DKIE result with
BRDE (1.1) makes an interpretation uncertain and, while the
value with peroxyl radicals (1.2) appears to distinguish BRDE
from 1 and BVDE, the latter two giving DKIE of 1.6, there is
no obvious explanation for such a difference in DKIE. Conse-
quently, we decided to study solvent effects on these compounds
to determine if their kinetic behavior and mechanisms could be
differentiated. Compound 8 was included in some of these
experiments as a representative monopyrrole.
III. Kinetic Solvent Effects on 8, Dipyrrinone, 1, Bilru-
bin and Biliverdin Dimethyl Esters with Peroxyl Radicals
and with DPPH•. Pyrrolic molecules, such as the bile pigments,
have important antioxidant properties in polar aqueous
systems.3b Consequently, it was important to determine the
effects of a polar solvent on representative pyrroles under
controlled conditions. The importance of hydrogen bonding
between the hydroxyl group of phenols and protic solvents on
the antioxidant activities of phenols was recognized decades
ago.17a Such solvent effects have been extensively investigated
more recently and arranged on a systematic basis for phenolic
antioxidants by Ingold et al.17b–j based on the assumption that
hydrogen bond accepting solvents (HBAs) prevent HAT from
the bonded OH group so that HAT is limited to the “free” (non-
H-bonded) OH group (see Scheme 2).
TABLE 4. Antioxidant Activities, kinh, and Stoichiometric Factors,
n, of PMHC, BRDE, BVDE, 1, and 8 during the Inhibited
Peroxidation in C6H5Cl, of (a) Styrene, (b) Styrene with MeOH,
Initiated by AIBN
a
substrate
inhibitor, M × 106 kinh M-1 s-1 × 10-4
nb
(a) styrene, 0.87 M
PMHCc
1c
380
2*
12.4
10.2
22.5
25.4
57.3
1.8
2.7
2.0
BVDEc
BRDEc
8
(b) styrene, 1.74 M, PMHC 2.23-9.24
2*
MeOH, 11.1 M
8 9.06-9.86
1
17.5
2.1
-
not measurable
not measurable
37.9
BVDE
BRDE 9.14- 13.0
-
1.8
a Rate constants for reactions with peroxyl radicals initiated by AIBN
(21.6-22.5 mM) at 30 °C calculated from plots of eq 11, using a kp
)
41 M-1 s-1 for styrene. Results are from at least three measurements
with error limits less than 10% for PMHC and BRDE, and less than
20% for 8. The kinh for 8 in styrene is taken from Table 1. b The
stoichiometric factor of PMHC from ref 6. The stoichiometric factors were
calculated from n ) Ri × τ/[antioxidant], where Ri ) 2[ArOH]/τ and
ArOH) PMHC. c Values of kinh and
n were previously repo-
rted.6a,8e,f
(17) (a) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1964, 42, 1044–1056.
(b) Litwinienko, G.; Ingold, K. U. Acc. Chem. Res. 2007, 40, 222–230. (c)
Litwinienko, G.; Ingold, K. U. J. Org. Chem. 2005, 70, 8982–8991. (d)
Litwinienko, G.; Ingold, K. U. J. Org. Chem. 2004, 69, 5888–5897. (e)
Litwinienko, G.; Ingold, K. U. J. Org. Chem. 2003, 68, 3433–3438. (f) Snelgrove,
D. W.; Lusztyk, J.; Banks, J. T.; Mulder, P.; Ingold, K. U. J. Am. Chem. Soc.
2001, 123, 469–477. (g) Valgimigli, L.; Banks, J. T.; Lusztyk, J.; Ingold, K. U.
J. Org. Chem. 1999, 64, 3381–3383. (h) Valgimigli, L.; Ingold, K. U.; Lusztyk,
J. J. Org. Chem. 1996, 61, 7947–7950. (i) MacFaul, P. A.; Ingold, K. U.; Lusztyk,
J. J. Org. Chem. 1996, 61, 1316–1321. (j) Valgimigli, L.; Banks, J. T.; Lusztyk,
J.; Ingold, K. U. J. Am. Chem. Soc. 1995, 117, 6226–6231. (k) Musialik, M.;
Litwinienko, G. Org. Lett. 2005, 7, 4951–4955. (l) Foti, M. C.; Daquino, C.;
Geraci, C. J. Org. Chem. 2003, 69, 2309–2314. (m) Foti, M. C.; Johnson, E. R.;
Vinqvist, M. R.; Wright, J. S.; Barclay, L. R. C.; Ingold, K. U. J. Org. Chem.
2002, 67, 5190–5196. (n) Barclay, L. R. C.; Edwards, C. E.; Vinqvist, M. R.
J. Am. Chem. Soc. 1999, 121, 6226–6231.
Such systematic quantitative studies of free radical reactions
on pyrrolic molecules are lacking. However, it was desirable
to investigate the effect of a polar solvent on the antioxidant
(14) Howard, J. A.; Ingold, K. U. Can. J.Chem. 1962, 40, 1851–1864.
(15) de Heer, M. Doctoral Dissertation, de Universiteit Leiden, May 2002.
(16) Mayer, J. M. Annu. ReV. Phys. Chem. 2004, 55, 363–390.
6628 J. Org. Chem. Vol. 73, No. 17, 2008

Pyrroles As Antioxidants
TABLE 5. Rate Constants,a k2 (M-1 s-1), for the Reaction of
B. Kinetic Solvent Effects (KSE) with DPPH• Radicals.
Initial KSE studies on pyrroles 7 and 8 with the DPPH• radical
were carried out in tert-butylbenzene, methanol, and methanol
containing acetic acid. tert-Butylbenzene was used as a nonpolar
solvent, and the initial rates method was used17l,19 because of
limited solubility of these compounds in alkanes.
b
DPPH• with BRDE and 1 in tert-Butylbenzene, Methanol, and
Methanol Containing Acetic Acid
solvent
compoundc
t-butyl C6H5
MeOH
MeOH + Acetic Acidd
BRDEe
1
77
5.1
1100
3.3
4075
18
a Rate constants at 30 °C calculated from an average of at least three
determinations using second-order stopped-flow methods. Individual
rates were determined from plots of the initial decay of DPPH•
b
absorption and k2 calculated from eq 13. DPPH• concentrations ranged
from 137 to154 µM. c Compound concentrations ranged from 96 to 174
µM. d The acetic acid concentration was 100 mM. e Results from at least
three determinations with error limits less than 10% for BRDE and less
than 30% for 1.
IR ) 2a × t0 + b
(14)
Therefore, since the Abs0, [A-H]0, and t0 are known, the second-
order rate constant can be calculated. Rate constants for DBHA,
BRDE, and 1 are given in Table 5. Interestingly, the dipyrrinone,
1, and the tetrapyrrole, BRDE, exhibited quite different
behaviors in a polar solvent. The k2 of BRDE increased by 14-
fold, whereas the k2 of 1 slightly decreased in methanol.
Surprisingly, both second-order rate constants of BRDE and 1
increased with the addition of acetic acid to methanol. The other
dipyrrinones 2-4, and dipyrrin, 5 (Figure 1), did not have
measurable rate constants with DPPH•.
IV. Ionization Potentials (IPs) of Monopyrroles, Deriva-
tives and Model: Compounds. It was thought that consideration
of the ionization potentials of the two pyrroles, 7 and 8,
compared to those of pyrrole, dimethylpyrrole, and some model
compounds representative of the bile pigments would help
clarify the possible reaction mechanisms involved in the
inhibition experiments. Calculated IPs for pyrrole and 2,4-
dimethylpyrrole from the literature are compared with our
calculations for tetraphenyl- and tetra-4-methoxyphenylpyrrole
in Table 6. As would be expected, electron-supplying groups
FIGURE 4. Absorption decay by time for the reaction of DPPH•, with
8 in (a) tert-butylbenzene, DPPH•, 99.7 µM; 8, 108 µM; (b) MeOH,
DPPH•, 105 µM; 8, 89.9 µM; and (c) MeOH containing 100 mM acetic
acid, DPPH•, 105 µM; 8, 89.9 µM. The concentrations are before
mixing in the stopped-flow wherein the concentrations will be one-
half.
Qualitatively, 7 was not an active hydrogen atom donor in
the three solvents, and similarly, 8 exhibited no activity in the
nonpolar solvent, tert-butylbenzene (see Figure 4a). However,
in methanol, 8 was surprisingly active, and its activity increased
in methanol in the presence of acetic acid (see Figure 4b,c).
The decay of DPPH• in the presence of acetic acid was too
fast to measure for calculations of the initial rates by this method.
Quantitative studies on BRDE and I were carried out by
monitoring the second-order decay of the absorption of the
DPPH• radicals, and the second-order rate constants, k2, were
calculated from the initial rates (IR) of the DPPH• decay as
outlined below:
k2 ) IR ⁄ (Abs0) × [A-H]0)
(13)
TABLE 6. Ionization Potentials (IPs) of Pyrroles and Derivatives
where (Abs0) is the initial absorbance and [A-H]0 is the initial
concentration of the antioxidant. The IR of the decay was
calculated by fitting the initial decay to the parabola equation y
) ax2 + bx + c. R2 values for the parabola equation were >0.98.
The IR at t0 is given by the following equation:
(18) A reviewer questions the calculations where we used the kp ) 41 M-1
s-1 for styrene containing methanol (Table 4b). Based on a reported kp of 0.3
M-1 s-1 for the reaction of peroxyl radicals with methanol (Denisov, E. T.;
Denisova, T. G. Handbook of Antioxidants, 2nd ed.; CRC Press: Boca Raton,
FL, 2000),this reviewer suggests that the use of the kp for styrene is “certainly
wrong”. However, some control experiments showed that this is definitely not
the case. For example, experiments with the concentrations of AIBN and
methanol shown in Table 4b indicated no oxygen uptake until styrene was added.
In addition, in the styrene/methanol mixture used, the rate of oxidation depended
only on the styrene concentration. The rate at a styrene concentration of 0.87 M
was 0.154 × 10-6 M s-1 and at 1.74 M styrene, the rate ) 0.326 × 10-6 M s-1
,
a 2-fold increase, as expected. The Ri was determined for these experiments
and the oxidizability, (kp/(2kt)1/2, eq 6) of styrene determined to be 4.49 × 10-2
M-1/2 s-1/2. This value is consistent with literature values (3.4-4.6 × 10-2 M-
1/2 s-1/2 for the oxidizability of styrene in polar solvents (Howard, J. A.; Ingold,
K. U. Can. J. Chem. 1964, 42, 1044–1056). Quantitative kinetic studies between
peroxyl redicals and alkenes have been carried out by others in methanol solvent,
even with very reactive perhaloperoxyl radicals without interfering reactions by
methanol ( Shoute, L. C.; Alfassi, Z. R.; Neta, P.; Huie, R. E. J. Phys. Chem.
1994, 98, 5701–5704. ). The lack of interference by methanol was attributed to
the lower reactivity of the HOCH2O2• radical. Further examination of this question
of the reactivity of peroxyl radicals with methanol is beyond the scope of this
work.
a Values are calculated. b This compares with an experimental value
of 189 kcal /mol.20
lower the IP values as the resulting cation would be more
stabilized. Methyl groups at positions 2 and 4 lower the IP,
and the presence of aryl groups, especially those bearing para-
methoxy groups, provided marked lowering of the IP values.
Such substituent effects are probably most significant on the
(19) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2nd ed.;
McGraw-Hill, Inc.: New York, 1995.
J. Org. Chem. Vol. 73, No. 17, 2008 6629

MacLean et al.
kinetics in polar solvents and could account for the observed
variations of antioxidant activities and even mechanism shifts,
for example, from hydrogen atom abstraction to single electron
transfer (see Discussion).
Discussion
I. Antioxidant Activities and Mechanisms of Pyrroles
in a Nonpolar Solvent. As proposed in the Introduction,
pyrroles might be expected to act as hydrogen atom donors from
the N-H group to provide antioxidant activity, much like the
well-known HAT mechanism from the O-H group of phenolic
antioxidants. Assuming that the resulting pyrrolyl radical adopts
the 5π-electron system, the effect of strong electron-supplying
or -attracting groups present in 6, 7, and especially in 8 should
control the relative stabilities of the incipient radicals in an
analogous fashion as recognized6 for ortho and para substituents
in phenoxyl radicals. Therefore, the observed relative HAT
antioxidant activities of 8 > 7 > 6 during the inhibition of
peroxyl radical reactions on a substrate in a nonpolar solvent
(Table 1) is as expected since electron-supplying aryl groups,
especially in 7 and 8, accelerate the rate-determining antioxidant
step (eq 15).
The IP of the dipyrrinone, 1, is the highest among the
calculated values, and this is probably due to the strong electron-
attracting carbonyl group conjugated with the system. This
contrasts with the bilirubin model compound where the IP drops
by nearly 70 kcal/mol compared to that of 1 (Table 6). It is
thought that this “theoretical” structure would mimic the
behavior of the two central pyrrole rings of derivatives of
bilirubin (e.g., BRDE, Figure 1) because, in the latter, the two
carbonyl groups of the dipyrrinones should give equal and
opposite electron-attracting effects.
V. Reactions of Phenols with DPPH• and an Oxygen-
Centered Radical. While DPPH• has been the most popular
radical with which to monitor hydrogen atom transfer activities
(HAT) of phenols, other stable oxygen-centered radicals have
been used in stopped-flow kinetics, such as galvinoxyl (G•)21
and the 2,6-di-tert-butyl-4-(4′-methoxyphenyl)phenoxyl, DB-
MP•, radical which has been shown to follow the same relative
trend in rate constants as peroxyl radicals upon reaction with
vitamin E type phenols.22 We carried out a few exploratory
experiments for the purpose of showing the trend of solvent
effects with DPPH• compared to the DBMP• radical using the
pseudo-first-order methods. The results for HAT reactions with
the nitrogen-centered radical, DPPH•, in heptane (Table 7) agree
with literature values.17e With acetic acid present, the rate
decreased substantially, as expected.17e
The superior antioxidant activity of 8 is attributed to the
electron-supplying para-methoxyphenyl groups. The four aryl
groups in 7 and 8 cause steric crowding especially at positions
3 and 4, and as a result, these aryl groups are almost orthogonal
to the plane of the pyrrole ring in 8, while those at positions 2
and 5 were less twisted from this plane in the minimum energy
conformation. (See Supporting Information for a minimized
structure of the parent, 8, the derived radical, and calculated
twist angles.) The twist angles at positions 2 and 5 decrease in
the radical by 12-16° compared to the parent, while those at
positions 3 and 4 change by only 2-3°. Consequently, most of
the stabilization of the radical is expected to come from the
groups at positions 2 and 5, where the stabilizing effects of the
methoxy groups can occur through the aromatic systems by
direct resonance (see Figure 5).
TABLE 7. Trends in Rate Constants, k (M-1 s-1) of
2,4,6-Trimethyl- (13) and 2,6-Di-tert-butylphenol (12) with DPPH•
(1-2 × 10-4 M) and DBMP• (6.0-8.5 × 10-5 M) in Heptane,
Methanol, and Methanol-Acetic Acid (AA) (10 mM)
compounda
solvent
DPPH•
DBMP•
13
heptane
MeOH
MeOH
+AA
heptane
MeOH
MeOH
+AA
47.7
179b
96.4
13.3
0.948
0.132
17.1b
12
22.2
1.02
NAc
a Compound concentrations ranged from 0.41 to 0.37 mM. b The
values in methanol are higher than reported, probably due to partial
ionization of the phenols and incursion of the SPLET reaction.17e c The
rate was too low to measure (<0.03 M-1 s-1).
FIGURE 5. Resonance structures of compound 8 radical.
Compound 7 lacks the methoxy groups of 8, so it is less active
in the HAT reaction, and compound 6, which bears a strong
electron-attracting ester carbonyl group at position 3, exhibits
very low antioxidant activity as a result. 2,4-Dimethylpyrrole
did not exhibit antioxidant activity in styrene.8f The inductive
effects of two methyl groups do not provide sufficient stability
of the pyrrolyl radical, and in this regard, the result is similar
to that found for dimethylphenol which possesses only very
weak antioxidant activity in styrene.6a
II. Antioxidant Mechanisms of Dipyrrinones, the Deriva-
tives BVDE and BRDE of Bile Pigments: Deuterium Kinetic
Isotope Effects (DKIE) N-H/N-D. The rationalization of
antioxidant activity of the dipyrrinones, 1-4, and the dipyrrin,
5, and application to biliverdin and bilirubin are more complex
issues. Earlier results using the dipyrrinones showed that only
systems bearing free N-H groups in both rings (e.g., compound
1) exhibited significant antioxidant activity. This was accounted
for by calculations showing strong hydrogen bonding in the
However, with the oxygen-centered radical, DBMP•, a
different trend was observed. With the latter as the monitoring
radical, the HAT activities of trimethylphenol, 13, and that of
the hindered phenol, 12, both dropped significantly in methanol
compared to that in heptane. As expected, methanol suppresses
the HAT of the phenols by hydrogen bonding at the phenolic
OH group, so that the DBMP• radical reacts with the “free”
ArOH (Scheme 2). This implies that this radical is superior to
DPPH• to monitor the HAT activities of H atom donors in a
solvent that supports ionization.
(20) von Niessen, W.; Cederbaum, L. S.; Diercksen, G. H. F. J. Am. Chem.
Soc. 1976, 98, 2066–2073.
(21) Shi, H.; Noguchi, N.; Niki, E. Methods Enzymol. 2001, 335, 157–166.
(22) Mukai, K.; Fukyda, K.; Ishizu, K. Chem. Phys. Lipids 1988, 46, 31–36.
6630 J. Org. Chem. Vol. 73, No. 17, 2008

Pyrroles As Antioxidants
change indicated by eq 16 is -6.7 kcal/mol and for eq 17 only
-0.6. kcal/mol. It is improbable that the thermochemistry alone
accounts for the low DKIE for these reactions.
1(N-H)+ROO• f (N•) + ROO-H
1(N-H)+DPPH• f (N•) + DPPH-H
(16)
(17)
FIGURE 6. Hydrogen bonding in the intermediate radical of a
dipyrrinone.
The low but consistent kH/kD of 1.6 observed for compound
1 and BVDE with ROO• radicals and 1 with DPPH• as well as
BRDE with these radicals (Tables 2 and 3) is indicative of a
transition state involving both transfer of the N-H hydrogen
and electron. Bearing in mind the tendency for a pyrrole to
transfer a π-electron from the ring, the proton-coupled electron
transfer (PCET) mechanism17b,27 should prevail. Nevertheless,
the different observed kinetic solvent effects exhibited by these
compounds showed unexpected trends for such a common
mechanism and require separate interpretations especially in
view of the expectation that the HAT and PCET mechanisms
are expected to show similar kinetic solvent effects from analogy
with such effects on phenols.17b
III. Kinetic Solvent Effects on 8, Dipyrrinone, 1, BRDE,
and BVDE on Reactions with Peroxyl Radicals and with
DPPH•. The antioxidant activities and mechanisms of these
pyrrolic compounds are controlled by several factors: (i) the
molecular structures, (ii) the nature of the attacking radical, and
(iii) the solvent used. A change in any one of these factors may
switch the mechanism or change the activity. In some cases,
even competing reactions may occur. The effects of a hydrogen
bond accepting (HBA) or ionizing solvent for reactions with
peroxyl radicals will be reviewed followed by those with
DPPH•, and finally, the two processes will be compared by a
quantitative treatment. A HBA solvent, such as methanol, is
expected to reduce HAT antioxidant activity by hydrogen
bonding of the solvent (N-H-OHCH3). In contrast, if the
antioxidant activity is dependent on a SET mechanism, the
ionizing solvent will stabilize the polar transition state involved
and the antioxidant activity will increase. To simplify rather
complex solvent effects on these structures, the effects of
methanol could depend on the predominance of HAT or SET
within the HAT/PCET pathways and on the electron deficiency
of the attacking radical.
intermediate radical (Figure 6).8f The results herein on the
isomeric dipyrrinones, 2 and 3 (Figure 1), provide important
support for this proposal. An N-methyl group on either ring
would block such H bonding in the intermediate pyrrolyl radical.
In these compounds, 2 lacks the pyrrolic N-H for HAT while
3 does not possess a donor H bond to stabilize the radical, and
both of these compounds are inactive as HAT antioxidants. Even
a strong electron-supplying methoxy group (compound 4) is not
sufficient to impart significant HAT activity, without internal
H bonding. In addition, in 4, the N-H of the pyrrolic ring is
probably H bonded to the N of the other ring to contribute to
the low reactivity of 4. The dipyrrin 5 has NH groups in both
rings but also proved to be inactive in a HAT reaction, even in
cumene. In solution 5 exists as a delocalized cation, as indicated
1
by the H NMR spectrum, where the N-H signals and alkyl
groups in the two pyrrolic rings exhibited identical resonances.23
The protonated dipyrrin cation need not exclusively maintain
the natural Z-configuration at the connecting CH. As this
configuration was shown to be necessary for stabilization of
the intermediate pyrrolyl radical by internal H bonding,8f HAT
activity is lost by 5. There is some precedent for the role that
internal H bonding can play in controlling the antioxidant
activity of other polyfunctional compounds, such as catechols
and 1,8-naphthalene diols.2a Such intramolecular interaction
could also be a factor in controlling the antioxidant activity of
the more complex bile pigments. However, it does not dif-
ferentiate between the various possible pathways for the reaction.
Some of these proposed earlier include HAT from the connect-
ing CH2 group in the case of bilirubin,8c radical addition to the
pyrrolic rings,8g HAT from free N-H groups,8f and SET in polar
media.8e
To the best of our knowledge, there are no reports regarding
recent applications of DKIE for radical reactions relevant to
the antioxidant ability of pyrrolic systems. We recently reported
some DKIE on H atom transfer from other heterocyclic
compounds, including a dihydropyridine derivative and dihy-
droacridane. The kH/kD values ranged from 1.3 to 2.1 for
reactions with peroxyl radicals and 1.8-2.4 with DPPH•.24 Our
observed DKIE ratios of about 1.6 for compounds 1 and BVDE
(Tables 2 and 3) suggest a relatively unsymmetrical transition
state for these reactions. It was thought that an estimate of
reaction enthalpies for the HAT reactions with ROO• and with
DPPH• with the N-H group of 1 (eqs 16 and 17) might support
this explanation for a low kH/kD. Using the calculated BDE,
78.3 kcal/mol, of the N-H bond of the pyrrole ring A of 18f
(see Figure 6) and the experimental O-H BDE, 85.0 kcal/mol,
of tert-butylhydroperoxide,25 and 78.9 kcal/mol recently deter-
mined for the BDE of the DPPH-H bond,26 the enthalpy
The small effect of methanol on the reaction of the pyrrole,
8, with peroxyl radicals (a decrease of only 32% in kinh, Tables
1 and 4) contrasts with the typical large effects observed on
the kinh of methanol for phenolic antioxidants and for this solvent
effect on 1 and on BVDE (see below). The different solvent
effects of alcohols on 8 compared to those on phenols can be
accounted for by two factors: (i) pyrrole 8 involves a
N-H-OHCH3 interaction, whereas phenols have
a
OH-OHCH3 interaction, and these presumably have different
H bond strengths,28 and (ii) the net effect will depend on the
predominance of HAT or SET within the HAT/PCET pathways.
As noted in our previous article,8f the antioxidant activities
of dipyrrinone, 1, and BVDE are due to the very strong
(27) (a) Hodgkiss, M. J.; Damrauer, N. H.; Presse´, S.; Rosenthal, J.; Nocera,
D. G. J. Phys. Chem. B 2006, 110, 18853–18858. (b) Presse´, S.; Silby, R.
J. Chem. Phys. 2006, 124, 164504(1-7). (c) Hammes-Schieffer, S. Acc. Chem.
Res. 2001, 34, 273–281. (d) Decornez, H.; Hammes-Schieffer, S. J. Phys. Chem.
A 2000, 104, 9370–9384. (e) Cukier, R. I. J. Phys. Chem. 1996, 100, 15428–
15443. (f) Reece, S. Y.; Hodgkiss, J. M.; Stubbe, J. A.; Nocera, D. G. Philos.
Trans. R. Soc. London, Ser. B 2006, 361, 1351–1364.
(28) The fact that the hydrogen bond donating ability of R-tocopherol (and
presumably PMHC) as measured by the RH2 of 0.37017f is less than that of pyrrole,
0.408,32a indicates that different H-bonding effects are NOT the factor causing
the greater decrease in antioxidant activity of PMHC compared to 8.
(23) Wood, T. E.; Berno, B.; Beshara, C. S.; Thompson, A. J. Org. Chem.
2006, 71, 2964–2971.
(24) Mulder, P.; Litwinienko, G.; Lin, S.; MacLean, P. D.; Barclay, L. R. C.;
Ingold, K. U. Chem. Res. Toxicol. 2006, 19, 79–85.
(25) Clifford, E. P.; Wenthold, P. G.; Garejev, R.; Lineberger, W. C.; DePuy,
C. H.; Bierbaum, V. M.; Ellison, G. B. J. Chem. Phys. 1998, 109, 10293–10310.
(26) Foti, M. C.; Daquino, C. Chem.Commun. 2006, 3252–3254.
J. Org. Chem. Vol. 73, No. 17, 2008 6631

MacLean et al.
intramolecular hydrogen bond (9.0 kcal/mol) from the N-H of
the adjacent lactam ring as calculated for the intermediate
radical. This strong intramolecular H bond formed in the
nitrogen radicals derived from 1 and BVDE could dominate
their antioxidant properties and may determine the different
effect of methanol on their HAT activities compared to that on
BRDE, which forms strong intermolecular H bonds giving
dimers in solution.29 Therefore, the loss in activities of 1 and
BVDE observed with peroxyl radicals in the presence of
methanol (Table 4) can be attributed to hydrogen bonding in
two ways: (i) intermolecular hydrogen bonding between the
solvent and the antioxidant active pyrrole ring prevents HAT
by the radical and/or (ii) an intermolecular hydrogen bond
formed between the adjacent pyrrole ring and solvent blocks
the intramolecular hydrogen bond (see Figure 6). Surprisingly,
the antioxidant activity of BRDE increased somewhat (by 69%,
Table 4) in the presence of a polar solvent, which suggests SET
may be predominant in the HAT/PCET antioxidant mechanism.
The kinetic solvent effect of radical reactions of DPPH• is
an important issue because this radical is very frequently used
to monitor the “so-called antioxidant activities” by stopped-
flow kinetics, and because of the convenience of this method,
it has had many applications to natural products.30 This radical
is highly electron-deficient, especially at the reaction center, due
to contributing polar resonance structures, such as B and C in
Figure 7.
FIGURE 8. The effect of acetic acid on DPPH•/ pyrrole reactions.
The effect of acetic acid on the DPPH• reactions with these
pyrroles may be accounted for by one or both of two factors:
(i) the addition of acid to the solvent system may cause
protonation of DPPH•, resulting in increased reactivity of the
radical (see Figure 8), and (ii) the presence of acid may displace
the equilibrium toward the DPPH-H product. This latter effect
is quite clear in Figure 4 since the reaction does not proceed to
completion in the presence of methanol alone; the addition of
acetic acid is required. Consequently, the latter effect is the more
probable explanation for the accelerating effect of acetic acid.
Overall, these results demonstrate that SET is an important
pathway in the reactions of 8 and BRDE, especially with DPPH•
in an ionizing solvent.
Snelgrove et al. developed a relationship to predict rate
constants in a polar hydrogen bond accepting solvent from the
rate constant in a nonpolar solvent and a knowledge of the linear
free energy relationships of Abraham et al.,32a where RH2 is the
hydrogen bond donating ability of the substrate (e.g., antioxi-
dant), and ꢀH is the hydrogen bond accepting ability of the
2
FIGURE 7. Resonance structures of DPPH•.
solvent. Their derived equation made it possible to calculate
rate constants in a hydrogen bond accepting solvent from the
measured rate constant in a nonpolar solvent, as shown below
(eq 18).17f
The dipyrrinone, 1, and BRDE exhibit remarkably different
effects in the presence of methanol. The rate constant of 1 with
DPPH• in methanol is only slightly lower than that in a
hydrocarbon solvent, whereas BRDE exhibits a large increase
(by 14-fold, Table 5) on changing from the hydrocarbon solvent
to methanol. Such contrasting kinetic solvent effects on
compounds with similar structures as 1 and BRDE can be
accounted for by a different predominance in the HAT/PCET
reactions. In methanol, electron transfer may predominate on a
PCET reaction on BRDE. The slight effect of methanol on the
rate constant of 1 with DPPH• can be accounted for by
competing SET and HAT reactions. That is, methanol reduces
the classical HAT reaction due to interfering H bonding at the
pyrrolic N-H group and simultaneously increases the rate
through a SET reaction. It is interesting to note that, under
similar conditions, phenols were reported to react with DPPH•
by a combination of HAT and SPLET mechanisms.16 The HAT/
PCET reaction with 1 occurs with DPPH• but apparently not
with peroxyl radicals and illustrates the major effect that the
electron-deficient DPPH• has on both the kinetics and the
mechanism of the reactions of antioxidants.
H
H
log(ks⁄M-1 s-1) ) log(k0⁄M-1 s-1) - 8.3R2 ꢀ2
(18)
It is of interest to determine if such a quantitative treatment
can be applied to the solvent effect results of our pyrrolic
compounds, in particular, to compare the data with the peroxyl
and DPPH• radicals. We can predict the HAT rate constants
(k0) with peroxyl radicals of our compounds in methanol with
H
the following assumptions: (i) we can use R2 ) 0.370 from
R-tocopherol for PMHC;17f (ii) that the k0 in styrene can be
represented as that in a saturated hydrocarbon; (iii) a correction
for the concentration of methanol (ꢀ2H ) 0.4132a) can be made
for those experiments in styrene/methanol since the theory is
H
for a neat solvent; and (iv) we can use R2 ) 0.408 from
pyrrole32a for our pyrrolic compounds. First, it is useful to test
the calculation in styrene for a known antioxidant, PMHC. The
results with peroxyl radicals in styrene with PMHC (Table 8(a))
show that the experimental rate in methanol 57 × 104 M-1 s-1
is very similar to the calculated HAT value, 47 × 104 M-1 s-1
.
Calculations for compound 1 and BVDE give very low values
of 1.1 and 0.93 for ks but are in general agreement with our
experimental results (Table 4(b)) where the rates were too small
The acceleration effects of acetic acid on the reactions of 8,
BRDE, and 1 (Figure 4 and Table 5) were completely
unexpected. The resulting enhanced rate constants cannot be
attributed to the SPLET reaction, where acetic acid decreases
the rates, at least for phenols by suppressing their ionization.16,31
(31) The very low acidity of pyrrole, (pKHA) 23; Bordwell, F. G.; Zhang,
X.; Chen, J.-P. J. Org. Chem. 1991, 56, 3216–3219.) rules out the ionization of
the N-H group of pyrroles so that the SPLET mechanism observed for phenols
in ionizing solvents17b would not occur for these pyrroles.
(32) (a) Abraham, M. H.; Grellier, P. L.; Prior, D. V.; Duce, P. P.; Morris,
J. J.; Taylor, P. J. J. Chem. Soc., Perkin Trans. 2 1989, 699–711. (b) Abraham,
M. H.; Grellier, P. L.; Prior, D. V.; Morris, J. J.; Taylor, P. J. J. Chem. Soc.,
Perkin Trans. 2 1990, 521–529.
(29) Lightner, D. A.; Trull, F. R. Spectrosc. Lett. 1983, 16, 785–803.
(30) A recent search of SciFinder Scholar using the words ‘DPPH radical
and antioxidants’ responded with at least 3500 examples. Many of these employed
DPPH• to measure the antioxidant properties of plant extracts from aqueous
mixtures.
6632 J. Org. Chem. Vol. 73, No. 17, 2008

Pyrroles As Antioxidants
TABLE 8. Experimental Rate Constants in a Nonpolar Solvent,
decreases in IPs such as those of the tetraarylpyrroles 7 and 8
compared to pyrrole might be indicative of a propensity toward
single electron transfer (SET) rather than hydrogen atom transfer
(HAT).37
k0, in Methanol, kexp, and Calculated in Methanol, ks
a
b
a
k0 (M-1 s-1 × ks (M-1 s-1 × kexp (M-1 s-1
×
radical compound
10-4
)
10-4
)
10-4
)
(a) ROO•
PMHC
8
1
BVDE
BRDE
1
380
25
12
10
22
47
2.3
1.1
0.93
2
0.21
3.2
57
18
Actual comparisons to our kinetic data become complicated
by the fact that competitions between HAT or SET pathways
are also expected to be affected by the ionization capacity of
the solvent used and the nature of the attacking radical, ROO•
or DPPH•. Nevertheless, some clear comparisons between the
kinetic data and IPs can be made. For example, in the data for
reactions of BVDE and its model compound 1 with ROO•
(Table 4), the effects of methanol indicate an ordinary solvent
effect that retards HAT-type reactions by hydrogen bonding;
the high IP of 1 supports this conclusion, and the SET reaction
is retarded. In contrast the effect of methanol on the rate
constants of 8 and BRDE with peroxyl radicals in styrene
(Tables 1 and 4), a small decrease and a small increase,
respectively, appear ambiguous. However, it would appear that
either HAT or SET reactions could predominate within the
HAT/PCET reactions wherein the relatively very low IP of 8
indicates susceptibility toward a SET partway. So the SET
pathway must predominate in the faster reactions of 8 and
BRDE with the more electron-deficient DPPH• radical in
methanol (Figure 4 and Table 5). Finally, the relatively low IP
of the bilirubin model compound (Table 7) indicates that the
central pyrrole rings of BRDE could be susceptible to facile
electron transfer reactions promoting the PCET process.
V. Reactions of Phenols with DPPH• and the 2,6-Di-tert-
butyl-4-(4′-methoxy)phenoxyl, DBMP•, Radical. The use of
DPPH• as the monitoring radical under anaerobic conditions is
very convenient30 but susceptible to complex kinetic solvent
effects as indicated in the Results. We carried out a few
experiments to compare results with this radical and the oxygen-
centered radical, DBMP•. The results in Table 7 illustrate the
complexity of results using DPPH• where a polar solvent results
in a marked acceleration in reaction rate most likely due to initial
partial ionization of the phenol, followed by rapid electron
transfer (SPLET).17b Consequently incorrect conclusions on
“antioxidant activities” will result when using this radical in
common alcoholic solvents. Indeed, such results should NOT
be referred to as “antioxidant” experiments. We suggest that
the use of an oxygen-centered radical, such as DBMP•, of much
lower electron affinity than DPPH•, is much superior for such
monitoring of HAT activities by antioxidants using stopped-
flow kinetics because it does exhibit “normal” solvent effects.38
The phenolic precursor for the radical is available from a
convenient one-pot aryl coupling synthesis.39
38
3.3
1100
(b) DPPH•
5.1
77
BRDE
a Results taken from Tables 1, 4, and 5. b Calculated using eq 18.
to measure. In contrast, our experimental results for 8 and
BRDE are not consistent with the predicted rates calculated
from eq 18. In this case, the experimental values are 8 and 19
times larger, respectively, than the calculated values. It is thus
concluded that 8 and BRDE, unlike PMHC, do not react via a
simple HAT mechanism, but may undergo a HAT/PCET
reaction in a polar solvent or possibly a combination of reaction
pathways where SET becomes more prominent.
Using the same method, HAT rate constants, ks, of com-
pounds 1 and BRDE with DPPH• in methanol are calculated.
Experimental rates in methanol with compound 1 and BRDE
were 15 and nearly 350 times higher, respectively, than those
calculated assuming the HAT mechanism, supporting the
importance of a SET mechanism in the reaction of DPPH• with
these compounds. The calculated ks values herein for the pyrrolic
compounds cannot be claimed to be highly reliable because of
the assumptions involved. However, they do indicate that great
care must be shown when determining antioxidant activities in
polar solvents, especially when DPPH• is used as the monitoring
radical when reported “antioxidant activities” are of questionable
significance.
IV. Ionization Potentials (IPs) and Electron Transfer
Reactions of Pyrroles. A few selected examples of chemical
and physical evidence from the literature indicate that substituted
pyrroles are susceptible to both HAT and SET reactions. The
oxidation of the pyrrole 7 is known to form the tetraphenylpyr-
rolyl radical which is stable in solution.33 2,5-Di-tert-butyl- and
2,5-diphenylpyrrole form long-lasting radical cations by electron
transfer in ESR observations.34 Gas phase ionization of mono-
substituted pyrroles showed that their IPs gave good linear
correlations with σ+ constants,35 where the F ) -18.2 was
p
indicative of high sensitivity of the ionization to substituent
effects. The self-initiated autoxidation of 1,2,3-trimethylpyrrole
in chlorobenzene at 131 °C was interpreted as a SET reaction
to oxygen.36
We are well aware of the limitations when comparing gas
phase IP calculations to experimental results in polar solvents.
However, as indicated in our Results, ionization potentials of
the substituted pyrroles compared to pyrrole, and of a bilirubin
model compound compared to a typical dipyrrinone (see Table
6), might provide some insight into the antioxidant activities
and mechanisms of these compounds. In particular, large
(37) It has been suggested that, in the case of phenolic antioxidants, a lower
IP of about 45 kcal/mol compared to phenol indicates that the SET antioxidant
mechanism predominates. Wright, J. S.; Johnson, E. R.; DiLabio, G. A J. Am.
Chem. Soc. 2001, 123, 1173–1183. However, such an estimate would be subject
to a large uncertainty by solvent effects.
(38) In hindsight, there is some evidence that the oxygen-centered radical,
DBMP•, is a useful monitor to determine HAT activities in different solvents.
Earlier we reported the rate constants of several phenolic antioxidants using
DBMP• in several solvents.17n For example, from the value of PMHC in hexane
(33) (a) Kuhn, R.; Kainer, H. Biochim. Biophys. Acta 1953, 12, 325–328.
(b) Blinder, S. M.; Peller, M. L.; Lord, N. W.; Aamodt, L. C.; Ivanchukov,
N. S. J. Chem. Phys. 1962, 36, 540–544.
(34) Avila, D. A.; Davies, A. G. J. Chem. Soc., Perkin Trans. 2 1991, 1111–
1118.
(35) Linda, P.; Marino, G.; Pignataro, S. J. Chem. Soc. (B) 1971, 1585–
1587.
(36) Beaver, B.; Teng, Y.; Guiree, P.; Hapiot, P.; Neta, P J. Phys. Chem. A
1998, 102, 6121–6128.
of 93.9 × 103 M-1 s-1 as k0 and the RH ) 0.370 (found for R-tocopherol),17c
2
the calculated ks value in 1-propanol (ꢀH2 ) 0.45)32a of 3.90 × 103 M-1 s-1 and
in acetone (ꢀH ) 0.50)32a of 2.74 × 103 M-1 s-1, employing eq 18, are in
2
reasonable agreement with the reported values17n of 4.40 × 103 M-1 s-1 (1-
propanol) and 3.28 × 103 M-1 s-1 (acetone).
(39) Kawamura, Y.; Satoh, T.; Miura, M.; Nomura, M. Chem. Lett. 1998, 9,
931–932.
J. Org. Chem. Vol. 73, No. 17, 2008 6633

MacLean et al.
Conclusions
Thus it appears that the observation of a DKIE is important to
distinguish experimentally between PCET and sequential elec-
tron transfer followed by proton transfer. There is an interesting
analogy between this mechanism and PCET observed in a N-H
to oxygen H bonded amidium carboxylate interface27 in which
a N-H/N-D DKIE of 1.2 at 27 °C was observed.27a,bThis
complex has received considerable theoretical interest,27c–e and this
model system was reviewed as a model for radical processes in
biological systems.27f Recent NMR evidence for such a complex
between a hydrogen bond donor and peroxyl radicals indicates a
significance bond strength for a complex44 and provides support
for the role of such H bonded complexes in PCET reactions.
Our results show that some previous antioxidant mechanisms
for reactions of bile pigments with peroxyl radicals, such as
HAT from the connecting -CH2- group of bilirubin8b or
addition of peroxyl radicals to the pyrrolic rings8g are untenable.
The popular DPPH• radical for monitoring HAT activities
of phenols gives anomalous kinetic solvent effects which makes
it unreliable for determination of HAT activities in protic
solvents. In contrast, an oxygen-centered radical, 2,6-di-tert-
butyl-4-(4′-methoxyphenyl)phenoxyl, provides reliable HAT
kinetic data that agree with theoretical predictions and is
recommended for such studies using stopped-flow kinetics.
The polypyrroles in bile pigments exhibit antioxidant proper-
ties in various heterogeneous aqueous/lipid dispersions, includ-
ing blood plasma.3b This work provides new evidence toward
the antioxidant actiVities and mechanisms of pyrrolic compounds
in solution. A simple monopyrrole requires strong electron-
donating groups (e.g., p-methoxyphenyl) as in tetra-(4′-meth-
oxyphenyl)pyrrole, 8, for effective HAT antioxidant ability as
large as those of polyalkyl phenolic antioxidants with peroxyl
radicals in chlorobenzene.
The antioxidant mechanisms and activities of the dipyrrinones
and related bile pigment derivatives, BVDE and BRDE, are
strongly influenced by various factors including the structure
of the pyrrolic moiety, the nature of the attacking radical, and
the ionizing capacity of the solvent such that a single compre-
hensive antioxidant reaction mechanism cannot be applicable
for these pyrroles. Only dipyrrinones that possess two “free”
N-H groups are active antioxidants, confirming the proposal
that strong internal H bonding from an adjacent ring (e.g., as
in 1) is required to provide HAT activity in the HAT/PCET
reactions. The HAT activity for 1 is also supported by N-D/
N-H DKIE results, the retarding effect of a HBA solvent,
methanol, on the kinh due to external H bonding, and the high
ionization potential of 1. The bile pigment of biliverdin (BVDE)
showed similar DKIE and solvent effects to 1 on reaction with
peroxyl radicals and therefore reacted by a HAT/PCET path.
Pyrroles readily undergo electron transfer reactions, probably
due to the facile formation of a 5π-electron ring system. In
methanol containing DPPH•, this is accompanied by proton
transfer. In the case of the readily ionizable 8, the proton transfer
might follow SET in methanol (Figure 8). Compound BRDE
may follow this route on reaction with DPPH• in the presence
of both methanol and acetic acid. This is in sharp contrast to
the reactions of phenols where claims for this mechanism are
judged to be “without merit”.17b
Experimental Section
Materials: Bilirubin dimethyl ester and biliverdin dimethyl
ester were obtained from a commercial supplier and used as
received. The commercial inhibitors, 2,2′,5,7,8-pentamethyl-6-
hydroxychroman (PMHC), 2,6-di-tert-butyl-4-methoxyphenol
(DBHA), 2,6-di-tert-butylphenol (DTBP), 2,4,6-trimethylphenol
(TMP), and 2,6-di-tert-butyl-4-methylphenol (BHT) were re-
crystallized from methanol before use. AIBN was recrystallized
from methanol from a solution prepared at room temperature
and cooled. Commercial styrene was separated from added
stabilizer by bulb-to-bulb vacuum distillation at room temper-
ature. The distilled styrene was passed “neat” through chro-
matographic silica gel immediately before use. Cumene, purest
grade, was passed through silica gel before use. Solvents used
were of HPLC purity. Diphenylpicrylhydrazyl radical (98%) was
used as received. The concentration of DPPH• was determined
from its molar absoptivity value, ꢀ, of 11 500 M-1 cm-1 (λmax
) 516 nm).17l 2,6-Di-tert-butyl-4-(4′-methoxyphenyl)phenol, the
phenolic precursor of DBMP•, was synthesized according to the
procedure outlined by Kawamura et al.39 by the aryl coupling
reaction from 2,6-di-tert-butylphenol and 4-bromoanisole in the
presence of tri-(4-methoxy)phenylphosphine catalyzed by cesium
carbonate and palladium acetate. The product recrystallized from
hexane was identical to that reported earlier.17n The reported
procedure17n was followed to prepare the DBMP• radical. The
concentration of DBMP• was determined from its molar absorp-
tivity value, ꢀ, of 24 000 M-1 cm-1 (λmax ) 370 nm).17n
Syntheses and characterization of pyrrolic compounds 2 and 3
are given in the Supporting Information. The synthesis of 4 was
reported earlier;8f however, an alternative method was used in
this work, which is also reported in the Supporting Information.
The synthesis of 5 was reported earlier.23 The syntheses of the
monopyrroles 6, 7, and 8 were reported previously; however,
the details are given herein along with complete characterization.
FIGURE 9. PCET mechanism for reaction of a pyrrolic molecule with
a peroxyl radical.
The antioxidant mechanism of BRDE with peroxyl radicals
in chlorobenzene is best described by the PCET mechanism. A
scheme for the mechanism is proposed in Figure 9, where a
N-H group of a pyrrolic compound like BRDE is H bonded
to a peroxyl radical, known to be highly polarized with the
terminal oxygen bearing higher negative charge.40,41 In this
associated complex, proton vibrational motion is coupled to
electron transfer so a kinetic isotope effect would be expected.27
(40) Barclay, L. R. C. The Structure of Organic Peroxyl Radicals In Peroxyl
Radicals; Alfassi, Z., Ed; Wiley: Chichester, England, 1997; Chapter 3, pp 27-
48.
(41) The polarity of peroxyl radicals has been proposed to account for some
significant observations such as their diffusion from nonpolar regions of lipid
membranes40 or micelles,42 and hydrogen bonding by water could account for
their reduced reactivity resulting in lower termination rate constants in aqueous
lipid systems.43 Recently, direct evidence was provided for hydrogen bonding
of alkyl peroxyl radicals with an alcohol of high H-atom donating ability by
showing a large increase in the esr lifetime of sec-alkyl peroxyls and a enthalpy
change of-3.4 kcal/mol in the H-bonded complex.44 These results add support
for our proposed scheme since it is most likely that similar H-bonding occurs
between pyrrolic N-H groups and peroxyl radicals as shown in Figure 9.
(42) Yazu, K.; Yamamoto, Y.; Niki, N.; Miki, K.; Ukegawa, K. Nihon
Yukagakkaishi 2000, 49, 611–615.
(43) Barclay, L. R. C.; Baskin, K. A.; Locke, S L.; Vinqvist, M. R. Can.
Chem. 1987, 67, 1366–1369.
(44) Mugnaini, V.; Lucarini, M. Org. Lett. 2007, 9, 2725–2728.
(45) Wayner, D. D. M.; Burton, G. W. Handbook of Free Radicals and
Antioxidants in Biomedicine; CRC Press: Boca Raton, FL, 1989; Vol. 3, pp 223-
232.
6634 J. Org. Chem. Vol. 73, No. 17, 2008

Pyrroles As Antioxidants
Deuteration of Inhibitors: Inhibitors were dissolved in hexadeu-
terated acetone and D2O until all O-H or N-H hydrogen atoms were
deuterated. The acetone was then distilled, and the sample was
dissolved in chlorobenzene with traces of D2O. NMR spectra confirmed
deuterated O-H/N-H groups in both hexadeuterated acetone/D2O and
chlorobenzene solutions. The N-H’s for BVDE are not distinguishable
using NMR spectroscopy, and therefore, the N-H groups were
assumed to be deuterated by using the above procedure.
Autoxidation/Inhibition Procedures: Autoxidations were car-
ried out at 30 °C under 760 Torr of air in a dual-channel oxygen
uptake apparatus equipped with a sensitive pressure transducer
described previously.45 The procedures for conditioning the ap-
paratus and conducting an inhibition experiment have been
described previously.2c
mixed using a stopped-flow assembly and monitored using a
UV-vis spectrometer.
Acknowledgment. A.T. and L.R.C.B. acknowledge Discov-
ery Grants from the Natural Sciences and Engineering Research
Council of Canada for financial support of this research.
Supporting Information Available: Syntheses and charac-
terization of pyrrolic compounds. Calculations of ionization
potentials and the twist angles for 8 and the derived radical.
The structures of 8 (parent) and the derived radical with the
related twist angles. Detailed tables for oxygen uptake kinetics
and stopped flow kinetics. This material is available free of
charge via the Internet at http://pubs.acs.org.
Stopped-Flow UV-Vis Experiments: Experiments were con-
ducted at 30 °C under an argon atmosphere. Samples were rapidly
JO8005073
J. Org. Chem. Vol. 73, No. 17, 2008 6635