Molecular Markers for Pyrethrin Autoxidation in Stored Pyrethrum Crop: Analysis and Structure Determination

Article abstract of DOI:10.1021/acs.jafc.6b02959

Pyrethrum is a natural insecticide extracted from Tanacetum cinerariifolium. Six esters, the pyrethrins, are responsible for the extract's insecticidal activity. The oxidative degradation of pyrethrins through contact with aerial oxygen is a potential cause of pyrethrin losses during pyrethrum manufacture. Described here is the first investigation of the autoxidation chemistry of the six pyrethrin esters isolated from pyrethrum. It was found that pyrethrins I and II, the major pyrethrin esters present in pyrethrum, undergo autoxidation more readily than the minor pyrethrin esters, the jasmolins and cinerins. Chromatographic analysis of pyrethrin I and II autoxidation mixtures showed some correlation with a similar analysis performed on extracts from T. cinerariifolium crop, which had been stored for 12 weeks without added antioxidants. Two pyrethrin II autoxidation products were isolated, characterized, and shown to be present in extracts of stored T. cinerariifolium crop, confirming that autoxidation of pyrethrin esters does occur during crop storage.

Full text of DOI:10.1021/acs.jafc.6b02959

Article
pubs.acs.org/JAFC
Molecular Markers for Pyrethrin Autoxidation in Stored Pyrethrum
Crop: Analysis and Structure Determination
Jamie A. Freemont,† Stuart W. Littler,† Oliver E. Hutt,† Stephanie Mauger,†,‡ Adam G. Meyer,†
David A. Winkler,†,§ Maurice G. Kerr,# John H. Ryan,† Helen F. Cole,# and Peter J. Duggan*,†,§
†CSIRO Manufacturing, Bag 10, Clayton South, Victoria 3169, Australia
‡Institut de Recherche de Chimie Paris, CNRS − Chimie ParisTech, 11 rue Pierre et Marie Curie, 75005 Paris, France
#Botanical Resources Australia, 44-46 Industrial Drive, Ulverstone, Tasmania 7315, Australia
§School of Chemical and Physical Sciences, Flinders University, Adelaide, South Australia 5042, Australia
S
* Supporting Information
ABSTRACT: Pyrethrum is a natural insecticide extracted from Tanacetum cinerariifolium. Six esters, the pyrethrins, are
responsible for the extract’s insecticidal activity. The oxidative degradation of pyrethrins through contact with aerial oxygen is a
potential cause of pyrethrin losses during pyrethrum manufacture. Described here is the first investigation of the autoxidation
chemistry of the six pyrethrin esters isolated from pyrethrum. It was found that pyrethrins I and II, the major pyrethrin esters
present in pyrethrum, undergo autoxidation more readily than the minor pyrethrin esters, the jasmolins and cinerins.
Chromatographic analysis of pyrethrin I and II autoxidation mixtures showed some correlation with a similar analysis performed
on extracts from T. cinerariifolium crop, which had been stored for 12 weeks without added antioxidants. Two pyrethrin II
autoxidation products were isolated, characterized, and shown to be present in extracts of stored T. cinerariifolium crop,
confirming that autoxidation of pyrethrin esters does occur during crop storage.
KEYWORDS: Tanacetum cinerariifolium, pyrethrum, pyrethrins, autoxidation, degradation
processing of pyrethrum, but one such process most likely
involves oxidative attack. To date, information about the
oxidative degradation of pyrethrins has been largely derived
from studies related to photoxidation and oxidative mammalian
metabolism. Chen and Casida6 studied the photoxidation of
pyrethrin I and used 14C-labeled pyrethrin I and 2D thin layer
chromatography to show that pyrethrins decompose to give
many products when exposed to a sun lamp in the air. Although
the exact identity of the degradation products was not
determined, many appeared to result from oxidation of the
chrysanthemic acid portion of the pyrethrins. These workers also
indicated that the “alcohol moiety”, presumably the rethrolone
fragment, of pyrethrin I is “most susceptible to photo-
decomposition”.
Extensive studies of the oxidative mammalian metabolism of
pyrethrins have been undertaken by Casida and co-workers,1,7
who identified processes that led to the formation of side-chain
epoxides and diols, as well as oxidative processes that occur on
the chrysanthemic acid portion of the pyrethrins. Preliminary
work by Williams8 led to the suggestion that the diols 7−9
(Figure 1), oxidation products of pyrethrin II and previously
identified by Casida as pyrethrin metabolic products, were
produced upon heating ground pyrethrum seed.
INTRODUCTION
■
“Pyrethrum” is the term used to describe the insecticide that is
extracted and refined from dried pyrethrum daisies (Tanacetum
cinerariifolium). These daisies are grown commercially in
temperate regions or highland areas of the tropics. The potent
insecticidal properties of pyrethrum and its fast action, low
toxicity to plants and mammals, and rapid degradation in the
environment combine to make it ideal for many insecticidal
applications. In particular, it is most suited for “sensitive” uses
such as in food warehouses, homes, pet shampoo, and organic
agriculture and for human health. Pyrethrum contains six
insecticidal esters known collectively as “pyrethrins”.1 These
bioactive compounds occur in two groups (Figure 1), pyrethrins
I (pyrethrin I,1; cinerin I, 2; and jasmolin I, 3) and pyrethrins II
(pyrethrin II, 4; cinerin II, 5; and jasmolin II, 6). Chemically,
these compounds are esters, formed between chrysanthemic acid
or chrysanthemic diacid and three different hydroxylated
cyclopentenones,2 the latter known collectively as the
“rethrolones”. Pyrethrins I and II are the major components of
pyrethrum with the ratio of pyrethrins I to pyrethrins II varying
depending on plant variety, geography, and timing of harvest.
The susceptibility of pyrethrins to degrade upon exposure to
light,1,3,4 oxygen, and heat5 is a desirable attribute from an
environmental perspective. Care, however, must be taken to
minimize pyrethrin losses during manufacture of the pyrethrum
product. To this end, the crop is protected from exposure to light
and high temperatures after harvesting and is amended with the
antioxidant butylated hydroxytoluene (BHT).
Another potential route of oxidative degradation of pyrethrins
is via microbial activity,1 although products formed by this mode
Received: July 3, 2016
Revised: September 4, 2016
Accepted: September 6, 2016
Limited information has been published on the chemical
transformations of pyrethrins that occur during postharvest
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 1. Six insecticidal esters of pyrethrum (1−6) and side-chain oxidation products of pyrethrin II (7−11).
of degradation have not been reported. To date, no evidence has
been found to suggest that microorganisms play a role in the
Termination:
metabolism of pyrethrins under the conditions in which the crop
2L−OO• → [L−OO−OO−L]
is stored.5
(5)
A study of the electrochemical oxidation of synthetic
[L−OO−OO−L] → nonradical products + O2
In−In represents a radical initiator, and LH represents a lipid.
(6)
pyrethroids and pyrethrins I and II was published by Bond and
co-workers.9 They found that pyrethrins could be irreversibly
oxidized under conditions of cyclic voltammetry. No attempt was
made to identify the oxidation products; however, the authors
indicated that a wide range of compounds were formed, and the
overall reaction pathways for oxidation of the pyrethrins are
complex.
To date, the autoxidation chemistry, that is, free radical chain
reactions mediated by molecular oxygen, of pyrethrins has not
been subjected to close investigation. The structural similarity of
the “doubly allylic center” (i.e., methylene centers flanked by two
double bonds) in the rethrolone side chain of the pyrethrins to
unsaturated fatty acids such as linoleate,10 however, enables
parallels to be drawn between the autoxidation of lipids and
potential autoxidation reactions of pyrethrins. The chain
oxidation process that is thought to apply in the autoxidation
of lipids is as follows:11
The major observable product expected from this process is a
hydroperoxide, L−OOH, where L represents the lipid chain.
Hydroperoxides may be unstable and undergo a number of
nonradical decomposition reactions. The rate of formation of L−
OOH (eq 4) is concentration-dependent and also depends on
the C−H bond dissociation energy of L−H. Through
experimental observations and theoretical calculations, it has
been established that the doubly allylic positions are most
susceptible to autoxidation.10
In the study described herein, extracts of stored pyrethrum
crop were analyzed for the presence of pyrethrin degradation
products in the hope of gaining insights into the main causes of
pyrethrin losses during storage and into potential ways of
reducing these losses in the future. Evidence for the presence of
molecular markers of metabolic oxidation and autoxidation of
pyrethrin esters was sought, and the results were rationalized in
terms of the autoxidation processes known to occur with
unsaturated lipids. Included in this work is the first reported
investigation of the autoxidation chemistry of pyrethrins,
jasmolins, and cinerins.
Initiation:
In−In → 2In•
In•+L−H → In−H + L
(1)
•
(2)
MATERIALS AND METHODS
■
Propogation:
Chemicals and Reagents. m-Chloroperbenzoic acid (m-CPBA),
azobis(isobutyronitrile) (AIBN), and BHT were purchased from Sigma-
Aldrich (St. Louis, MO, USA). HPLC solvents were purchased from
Fisher Scientific (Loughborough, UK), and all other solvents were
obtained from Merck (Darmstadt, Germany). Water for HPLC and
•
L +O2 → L−OO•
(3)
•
L−OO•+LH → L−OOH + L
(4)
B
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
UPLC was purified through a Milli-Q system (Millipore, Billerica, MA,
USA).
Purified Pyrethrin Esters. The six individual pyrethrin esters
(Figure 1) were obtained in >99.5% purity from refined pyrethrum
concentrate through a combination of dry column vacuum chromatog-
raphy and careful preparative HPLC using a 900 mm × 100 mm i.d. glass
column packed with 100 μm Chromatorex RP-18 packing material (Fuji
Silysia Chemical Ltd., Kasugai, Japan) using an acetonitrile/water
gradient pumped at 350 mL/min by a micropump. Detection was by
Waters 481 LC spectrophotometer at 223 nm.
General Experimental Procedures. All NMR spectra were
recorded using CDCl3 as the solvent. Proton NMR spectra were
recorded on either an Av400 H spectrometer (Bruker, Billerica, MA,
USA) operating at 400.13 MHz or a BioSpin Av500 spectrometer
(Bruker) operating at 500.13 MHz. Carbon NMR spectra were recorded
either at 100.62 MHz on an Av400 H spectrometer or on a BioSpin
Av200 spectrometer (Bruker) operating at 50.33 MHz. Two-dimen-
sional COSY and HSQC spectra were recorded on an Av400 H
spectrometer, and two-dimensional HMBC were recorded on a BioSpin
Av500 spectrometer. Low- and high-resolution mass spectrometric
experiments (electron impact) were performed on a MAT 95XL mass
spectrometer (ThermoQuest, San Jose, CA, USA). Analysis by LC-MS
used a Q Exactive high-resolution FT mass spectrometer (Thermo
Scientific, Waltham, MA, USA), employing atmospheric pressure
chemical ionization (resolution, 70,000; scan range, m/z 100−1000;
capillary temperature, 320 °C; sheath gas, 25 (arbitrary units); auxiliary
gas, 10 (arbitrary units); discharge voltage, 2.8 kV; discharge current, 5.0
μA; vaporizer temperature, 350 °C). The column used was an Acquity
50 mm × 2.1 mm i.d., 1.7 μm, BEH RP-8 (Waters, Milford, MA, USA).
Solvent A was H2O with 5% CH3CN, and solvent B was CH3CN with
5% H2O. HPLC analysis was programmed as follows: 100% solvent A
for 20 min, followed by a linear gradient from 100% A to 100% B in 10
min, then 100% B for 2 min, then a linear gradient from 100% B to 100%
A in 1 min, and finally a hold at 100% A for 5 min. Flow injection analysis
was performed with the same instrument and settings as described above
with the exception that the vaporizer temperature was 450 °C and the
mass range was m/z 100−1500. The sample was not passed through the
HPLC column; instead, it was introduced with a 0.2 mL/min flow of
methanol, injected directly into the mass spectrometer.
UPLC analyses were performed on an Acquity UPLC system
(Waters) using a 50 mm × 2.1 mm i.d., 1.7 μm, Acquity BEH Shield RP-
18 column (Waters), with a flow rate of 0.4 mL/min and detection by
photodiode array (PDA) scanning in the range from 190 to 400 nm. For
the analysis of the autoxidation of individual pyrethrins, solvent A was
H2O with 50% CH3CN and solvent B was CH3CN with 5% H2O. UPLC
analysis was programmed as follows: 100% solvent A at 0 min followed
by a linear gradient from 100% A to 99.9% B in 4 min, then a linear
gradient from 99.9% B to 100% A in 0.5 min, and finally a hold at 100% A
for 2 min. For the analysis of crop extracts, the same solvent regimen and
program as used for LC-MS analysis, described above, was employed.
Data acquisition and processing were performed using Empower
software (Waters), and representative chromatograms were extracted at
223 nm for all analyses except those for the analytical scale autoxidation
experiments. In that case, the following wavelengths were used; 1, 223
nm; 2, 228 nm; 3, 221 nm; 4, 229 nm; 5, 221 nm; 6, 229 nm.
Semipreparative HPLC separations were performed on a HPLC
system (Waters) using a 250 mm × 22 mm i.d., 5 μm, Altima RP-18
column with detection by a 2487 dual-wavelength absorbance detector
(Waters) set at 223 nm. Preparative HPLC was performed on a HPLC
system (Waters) using a 300 mm × 40 mm i.d., 15 μm, Deltaprep RP-18
column, a mobile phase of 40% acetonitrile/water, and a flow rate of 80
mL/min and with detection by a 490E programmable multiwavelength
detector (Waters) set at 223 nm. Radial chromatography was performed
with silica gel (60 PF254) coated (4 mm) glass rotor plates using a
Chromatotron 7924T (Harrison Research, Palo Alto, CA, USA).
Crop Material. Samples of crop extracted included T. cinerariifolium
crop harvested from a pyrethrum site, which had not been stored (i.e.,
“unstored crop”), and crop from the same site that had been stored for
12 weeks at 22 °C (i.e., “stored crop”). These samples did not contain
BHT.
Preparation of 8′,9′ and 10′,11′ Epoxides and Diols from
Pyrethrin II. Synthesis and Isolation of Pyrethrin II 8′,9′ and 10′,11′
Epoxides. To a solution of pyrethrin II (195 mg, 0.52 mmol) in
dichloromethane (3 mL) was added m-CPBA (0.79 mmol) portionwise
over 5 min. The reaction was monitored by TLC developed with
EtOAc/petroleum spirits (1:4, v/v) until determined to be complete
(1.5 h). The solids were removed by filtration through glass fiber paper,
and the filtrate was concentrated to give a white sticky solid. The solid
was purified by radial chromatography (elution with 1:4 to 3:7 EtOAc/
petroleum spirits) providing two bands. Further purification of each
band by semipreparative HPLC (eluent: 50% CH3CN/H2O, 8 mL/
min) afforded the 8′,9′-epoxide 1 (39 mg, 19%) and the 10′,11′-epoxide
2 (14 mg, 7%) as clear colorless oils (both compounds were isolated as a
∼1:1 mixture of diastereomers).
8′,9′-Epoxypyrethrin II 12:12 δH (400 MHz, CDCl3) 6.46 (1H, m),
5.83 (1H, m), 5.68 (1H, m), 5.50 (1H, m), 5.42 (1H, m), 3.73 (3H, s),
3.44 (1H, m), 3.24 (1H, m), 2.91 (1H, ddd, J = 2.2, 6.4, 18.7 Hz), 2.54
(1H, dt, J = 4.8, 14.1 Hz), 2.40 (1H, dd, J = 7.3, 14.3 Hz), 2.30−2.20
(2H, m), 2.07 (3H, s), 1.95 (3H, d, J = 1.4 Hz), 1.74 (1H, t, J = 5.4 Hz),
1
1.31 (3H, s), 1.24 (3H, d, J = 2.3 Hz) (all diastereomeric H signals
overlap); δC (100 MHz, CDCl3) 204.0, 203.9, 171.3, 171.3, 168.2, 168.2,
167.0, 166.9, 140.3, 140.3, 139.1, 132.1, 132.0, 129.9, 120.8, 73.6, 73.5,
57.4, 57.3, 56.8, 56.8, 52.0, 42.2, 42.1, 35.9, 35.9, 33.1, 33.0, 30.7, 22.5,
22.4, 22.3, 20.6, 14.5, 14.5, 13.0 (some diastereomeric 13C signals
overlap); m/z (EI, 70 eV) found M•+ 388.1875, C22H28O6 requires
388.1880.
10′,11′-Epoxypyrethrin II 13:12 δH (400 MHz, CDCl3) 6.46 (1H, m),
5.69−5.58 (2H, m), 5.09 (1H, m), 3.76 (1H, m), 3.73 (3H, s), 3.25−
3.11 (2H, m), 3.02 (1H, m), 2.89 (1H, dd, J = 6.5, 18.5 Hz), 2.66 (1H,
m), 2.27−2.19 (2H, m), 2.07 (3H, s), 1.94 (3H, d, J = 1.5 Hz), 1.74 (1H,
d, J = 5.2 Hz), 1.30 (3H, s), 1.23 (3H, s) (all diastereomeric 1H signals
overlap); δC (100 MHz, CDCl3) 203.5, 203.4, 171.2, 168.1, 165.2, 165.1,
141.8, 138.9, 131.2, 131.1, 129.8, 128.9, 128.8, 73.4, 73.4, 51.8, 48.6,
47.8, 42.0, 42.0, 35.8, 33.0, 30.6, 29.7, 22.3, 21.9, 21.8, 20.4, 14.4, 12.9
(some diastereomeric 13C signals overlap); m/z (EI, 70 eV) found M+•
388.1854, C22H28O6 requires 388.1880.
Synthesis and Isolation of Diols from 8′,9′-Epoxypyrethrin II. 8′,9′-
Epoxypyrethrin II 12 (39 mg, 0.10 mmol) was dissolved in methanol/
water (2:1, 600 μL), and one drop of 0.5 M H2SO4 was added. The
reaction was stirred for 1 h at room temperature and analyzed by TLC,
developed with EtOAc/petroleum spirits (2:3, v/v), which showed no
starting material present. A 5% NaHCO3 solution was added to
neutralize the reaction, and then the mixture was concentrated to
remove the methanol. The aqueous phase was extracted with EtOAc
(four times), and the combined organic extracts were dried (MgSO4),
filtered, and concentrated. Purification by semipreparative HPLC
(eluent: 45% CH3CN/H2O, 10 mL/min) afforded 8′,9′-dihydroxy
pyrethrin II 711 (20 mg, 49%) as a clear colorless oil (mixture of
diastereomers). δH (400 MHz, CDCl3) 6.45 (1H, m), 5.88 (1H, ddd, J =
6.0, 10.5, 16.9 Hz), 5.68 (1H, brt, J = 5.1 Hz), 5.37 (1H, m), 5.27 (1H,
m), 3.91 (1H, m), 3.73 (3H, s), 3.62 (1H, m), 3.22 (1H, brs), 2.93 (1H,
dd, J = 6.3, 18.9 Hz), 2.81 (1H, brs), 2.58−2.40 (2H, m), 2.28 (1H, ddd
(obscured), J = 2.0, 3.9, 19.0 Hz), 2.23 (1H, dd (obscured), J = 5.2, 9.7
Hz), 2.06 (3H, s), 1.94 (3H, d, J = 1.3 Hz), 1.74 (1H, dd, J = 2.2, 5.2 Hz),
1
1.30 (3H, s), 1.24 (3H, s) (all diastereomeric H signals overlap); δC
Extraction of Crop Material. A known amount of sample (∼500
mg) was weighed into a 50 mL centrifuge tube. Ethyl acetate (25 mL)
was added, and the sample was mixed at room temperature for 2 h on a
rotisserie shaker. The mixture was centrifuged (4 °C, 2000 rpm, 10 min)
and the liquor decanted and concentrated under vacuum. The residue
was taken up in acetonitrile (2 mL), and a portion (200 μL) was taken
and further diluted up to 1.5 mL with acetonitrile. The diluted sample
was analyzed by UPLC (3 μL injection).
(100 MHz, CDCl3) 206.5, 171.4, 171.4, 168.3, 168.3, 168.3, 141.0,
139.1, 137.4, 137.4, 130.1, 117.8, 117.8, 75.4, 75.4, 73.7, 73.6, 72.9, 72.8,
52.0, 42.2, 42.1, 35.9, 35.9, 33.2, 33.2, 30.9, 30.9, 28.1, 28.1, 22.5, 20.6,
20.6, 14.6, 14.5, 13.1 (some diastereomeric 13C signals overlap); m/z
(EI, 70 eV) found M•+ 406.1981, C22H30O7 requires 406.1986.
Synthesis and Isolation of Diols from 10′,11′-Epoxypyrethrin II.
10′,11′-Epoxypyrethrin II 13 (10 mg, 0.025 mmol) was dissolved in
C
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 2. Synthesis of compounds 7−9 identified as pyrethrin II mammalian metabolic products by Casida.2
methanol/water (2:1, 900 μL), and one drop of 0.5 M H2SO4 was
added. The reaction was stirred for 1 h at room temperature and
analyzed by TLC, developed with EtOAc/petroleum spirits (2:3, v/v),
which showed no starting material present. A 5% NaHCO3 solution was
added to neutralize the reaction, and the mixture was concentrated to
remove the methanol. The aqueous phase was extracted with EtOAc
(four times), and the combined organic extracts were dried (MgSO4),
filtered, and concentrated. Purification by semipreparative HPLC
(eluent: 40% CH3CN/H2O, 3 mL/min) afforded 10′,11′-dihydroxy
pyrethrin II 8 (6 mg, 59%) and 8′,11′-dihydroxy pyrethrin II 9 (3 mg,
30%), respectively, as clear colorless oils (both diols were obtained as
mixtures of diastereomers).
under nitrogen at −8 °C until analysis by UPLC. Immediately prior to
analysis, the samples were diluted with acetonitrile. The extent of the
disappearance of pyrethrin I over time was measured relative to the BHT
peak. This procedure was reproduced with pyrethrin II, cinerins I and II,
and jasmolins I and II. The results are shown graphically in Figure 4.
Curves were fit until convergence using OriginPro 9.1 software
(OriginLab Corp., Northampton, MA, USA) employing the Rational
Nelder Model and the Levenberg−Marquardt iteration algorithm and
according to the following equation:
y = (x + a)/(b0 + b1*(x + a) + b2*(x + a) ∧ 2)
(7)
Preparative Autoxidation Experiment Performed with Pyr-
ethrins II. A freshly prepared solution of pyrethrins II, consisting of 80%
pyrethrin II, 12.9% cinerin II, and 2.7% jasmolin II (1.08 g, 0.1 M), and
AIBN (0.01 M) in benzene (29 mL) was placed in a 100 mL round-
bottom flask fitted with an air condenser and heated at 64 °C in an oil
bath for 24 h. The solvent was then removed under reduced pressure.
The pyrethrins II autoxidation mixture (∼1 g) was processed using
preparative HPLC, and peaks of interest were collected and analyzed by
UPLC and LC-MS. The compound eluting at 26.25 min from the UPLC
column, compound A, was obtained but was contaminated with another
component, compound B, eluting at 24.75 min in the UPLC.
Compound A was passed through the HPLC two more times; however,
it remained contaminated, whereas compound B could be isolated in
pure form. Compound A (13 mg) was obtained as a yellow oil, which
darkened over time. Compound B (14 mg) was a colorless oil.
Compound A was identified as 7′,8′-epoxypyrethrin II (10) and
compound B as 7′,8′-dihyroxypyrethrin II (11).
10′,11′-Dihydroxypyrethrin II 8:12 δH (400 MHz, CDCl3) 6.46 (1H,
m), 5.67 (1H, d, J = 5.9 Hz), 5.55 (1H, m), 5.39 (1H, m), 4.73 (1H, m),
3.74 (3H, s), 3.71−3.49 (3H, m), 3.24 (1H, m), 2.99 (1H, m), 2.90 (1H,
dd, J = 6.3, 18.8 Hz), 2.28−2.18 (3H, m), 2.09 (3H, s), 1.95 (3H, d, J =
1.4 Hz), 1.74 (1H, dd, J = 1.5, 5.2 Hz), 1.31 (3H, s), 1.24 (3H, s) (all
diastereomeric 1H signals overlap); δC (100 MHz, CDCl3) 205.3, 205.2,
171.4, 168.3, 166.4, 141.4, 139.1, 130.8, 130.7, 130.0, 128.6, 128.5, 67.9,
67.8, 66.3, 52.1, 42.3, 35.9, 33.2, 30.9, 22.7, 22.5, 20.6, 14.4, 13.1 (some
diastereomeric 13C signals overlap).
8′,11′-Dihydroxypyrethrin II 9:12 δH (400 MHz, CDCl3) 6.46 (1H,
m), 5.92−5.83 (1H, m), 5.78−5.65 (2H, m), 4.33 (1H, m), 4.15 (2H, t, J
= 4.9 Hz), 3.74 (3H, s), 2.92 (1H, dd, J = 6.3, 18.8 Hz), 2.67 (1H, dd, J =
4.5, 11.4 Hz), 2.55 (1H, dt, J = 3.9, 14.0 Hz), 2.47 (1H, dd, J = 7.8, 14.1
Hz), 2.32−2.21 (2H, m), 2.06 (3H, s), 1.95 (3H, d, J = 1.2 Hz), 1.75
(1H, dd, J = 2.3, 5.2 Hz), 1.38 (1H, t, J = 5.5 Hz), 1.31 (3H, s), 1.25 (3H,
s) (all diastereomeric 1H signals overlap); δC (100 MHz, CDCl3) 206.0,
206.0, 171.4, 171.4, 168.4, 168.1, 140.9, 139.1, 133.4, 130.1, 130.0, 73.8,
73.7, 70.8, 70.7, 63.1, 63.1, 52.1, 42.2, 35.9, 33.2, 33.2, 32.0, 30.9, 30.9,
22.5, 20.6, 14.6, 13.1 (some diastereomeric 13C signals overlap).
Radical Stability Predictions. Radical stabilities were calculated
using the MOPAC semiempirical molecular orbital program (Stewart
Computational Chemistry, Colorado Springs, CO, USA) using the AM1
parametrization. The radical stabilities were calculated by removing the
relevant hydrogen atom from the geometry-optimized structure of
pyrethrin I and using the UHF method to calculate the stability of the
resulting radical. The geometry optimization was carried out using the
Tripos force field13 and Gasteiger−Huckel charges in the Sybylx1.1
modeling package (Certera, Princeton, NJ, USA).
7′,8′-Epoxypyrethrin II 10: δH (400 MHz, CDCl3) 6.44 (1H, d, J =
9.5 Hz), 5.66 (1H, m), 5.57−5.39 (2H, m), 5.23 (1H, dd, J = 1.1, 9.7 Hz)
3.72 (3H, s), 3.15−3.10 (1H, m), 2.98−2.84 (3H, m), 2.67−2.58 (1H,
m), 2.43 (1H, dd, J = 6.0, 14.2 Hz), 2.30 (3H, m), 2.05 (3H, brs), 1.93
(3H, brs), 1.76−1.70 (1H, m), 1.30 (3H, s), 1.22 (3H, s) (all
diastereomeric 1H signals overlap); δC (50 MHz, CDCl3) 204.0, 203.9,
171.4, 168.3, 167.7, 167.6, 139.6, 139.5, 139.1, 135.1, 130.0, 119.8, 73.6,
73.5, 58.8, 58.3, 52.0, 42.2, 42.1, 36.0, 33.2, 33.1, 30.8, 26.1, 22.5, 20.6,
14.6, 13.1 (some diastereomeric 13C signals overlap); IR (cm−1) 2952
(br), 1711 (st), 1655, 1435, 1385, 1262, 1222, 1175, 1149, 1112, 994,
926, 831, 761; HRMS (APCI) m/z calcd for C22H29O6 [M + H]+
389.1959, found 389.1934.
7′,8′-Dihyroxypyrethrin II 11: δH (400 MHz, CDCl3) 6.45 (1H, d, J =
9.5 Hz), 5.98−5.86 (1H, m), 5.70−5.63 (1H, m), 5.35 (1H, dd, J = 1.3,
17.3 Hz), 5.27 (1H, d, J = 10.6 Hz), 4.08−4.00 (1H, m), 3.72 (4H, brs),
3.10 (1H, brdd, 3.5, 17.8 Hz), 2.91 (1H, dt, J = 6.6, 18.9 Hz), 2.63 (1H,
brd, J = 37.4 Hz), 2.54−2.38 (2H, m), 2.32−2.19 (2H, m), 2.06 (3H, s),
1.92 (3H, s), 1.73 (1H, d, J = 5.2 Hz), 1.29 (3H, s), 1.23 (3H, s) (all
diastereomeric 1H signals overlap); δC (50 MHz, CDCl3) 206.9, 171.4,
168.7, 168.5, 168.3, 141.4, 141.3, 139.1, 136.7, 130.1, 117.5, 75.5, 75.5,
73.8, 73.7, 73.4, 73.2, 52.1, 42.1, 42.2, 35.9, 33.2, 30.9, 26.5, 26.4, 22.5,
20.6, 14.6, 14.6, 13.1 (some diastereomeric 13C signals overlap); IR
(cm−1) 3435 (br), 2952 (br), 1708 (st), 1645, 1435, 1386, 1266, 1222,
1176, 1150, 1112, 994, 923, 830, 761, 734; HRMS (APCI) m/z calcd for
C22H29O6 [M − H2O + H]+ 389.1959, found 389.1959; calcd for
C22H31O7 [M + H]+ 407.2064, found 407.2063; calcd for C22H30O7Na
[M + Na]+ 429.1884, found 429.1882.
Autoxidation Experiments Performed with Pyrethrin I and II,
Cinerin I and II and Jasmolin I and II. Stock solutions of the six
pyrethrins 1−6 (>99.5% purity, 1.0 M) and a stock solution of AIBN
(0.1 M) were made up in cyclohexane. A stock solution of BHT (1.0 M)
was made up in petroleum ether.
An autoxidation experiment was first performed with pyrethrin I:
Stock solutions of pyrethrin I (10 μL) and AIBN (10 μL) were added to
a series of 1.9 mL UPLC vials and then diluted with cyclohexane (80 μL)
to give final pyrethrin I and AIBN concentrations of 0.1 and 0.01 M,
respectively. The vials were sealed and placed in an oven set at 64 °C.
Vials were removed periodically and quenched with BHT solution (50
μL). For the zero time experiment, BHT (50 μL) was first added to the
stock solution of pyrethrin I (10 μL) and cyclohexane (80 μL), followed
by AIBN (10 μL). After each vial was removed from the oven, the
solvent was removed under a stream of nitrogen and the vial was stored
D
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
RESULTS AND DISCUSSION
■
Analysis of Crop Material for Molecular Markers of
Oxidative Metabolism. Pyrethrum crop was examined for the
presence of diols 7−9 (Figure 1), identified by Casida7 as
resulting from biological metabolism of pyrethrin esters and
proposed by Williams8 to be present in ground pyrethrum seed
after heating at 100 °C for 10 h. These compounds were
synthesized using methods slightly modified from those
previously reported,12 as outlined in Figure 2, and used as
analytical standards for the analysis of extracts prepared from
crop.
Crop was extracted with a range of organic solvents of various
polarities, and the extracts were carefully analyzed by reverse
phase UPLC. Not surprisingly, the extracts contained a complex
mixture of compounds. Careful analysis revealed no evidence for
the presence of the three diols 7−9 in any of the extracts studied.
Autoxidation Studies of Pyrethrin Esters. Oxidative
metabolism that occurs in higher living systems typically employs
P450 enzymes.14 Whereas free radicals have been implicated as
intermediates in these processes, the involvement of oxo-heme
species distinguishes oxidative metabolism from autoxidation. It
follows, therefore, that oxidative metabolism of pyrethrins is
likely to produce a different product distribution from that
produced from the autoxidation of pyrethrins in the air.
As indicated above, the pentadienyl substituent of the
pyrethrins is somewhat reminiscent of the structures found in
unsaturated lipids, the autoxidation chemistry of which has been
extensively investigated. Figure 3 shows that a key step in the
Figure 4. Plots of the percent decrease of pyrethrin esters versus time in
AIBN-promoted autoxidation reactions. PI, 1; CI, 2; JI, 3; PII, 4; CII, 5;
JII, 6.
pyrethrins I and II but less for the cinerins and jasmolins (Figure
4). It was found that at least 80% of pyrethrins I and II had
degraded after 24 h and very little of either of these pyrethrins
remained after 96 h. The control reactions in which AIBN was
not included led to only 10−11% degradation of these
compounds. The fact that the cinerins and jasmolins are much
less susceptible to autoxidation than pyrethrins I and II is
consistent with the above arguments about the ease of hydrogen
abstraction from the 7′-position, because abstraction of a
hydrogen from this position in the cinerins and jasmolins
would give a less highly conjugated and hence less stable free
radical.
Identification of Autoxidation Products in Pyrethrum
Crop Extract. UPLC traces of the reaction mixtures from the
AIBN-initiated autoxidation experiments showed a significant
number of new peaks, which mainly eluted before the pyrethrins
and increased in intensity with time. An attempt was made to find
evidence for the autoxidation of pyrethrins in the T. cinerar-
iifolium crop by correlating UPLC peaks that appeared in the
AIBN-initiated autoxidation reaction mixtures, which also
increased in intensity as the reaction proceeded, with peaks
found in the crop extracts. Between 10 and 20 peaks showed such
a correlation, based on retention time (Figure 5).
The strongest correlation between the number of peaks in
crop extracts coincident with apparent autoxidation products,
from the AIBN-initiated reactions, occurred with the extracts of
stored crop samples, which had been stored in the absence of
antioxidants. The three most intense peaks that grew over time in
each of the pyrethrin I and II autoxidation reaction mixtures, and
were found to be coincident with peaks in the stored crop
extracts, were identified and their UV profiles extracted and
compared (Figure 5, peaks 2−7). UPLC peaks 2 and 7 in the
autoxidation mixture were both found to have UV profiles quite
different from those of the coincident peaks in stored crop
extract. Conversely, the UV profiles of UPLC peaks 3−6 in the
autoxidation mixtures appeared to be similar to the UV profiles of
coincident peaks in stored crop extract. The latter results provide
strong circumstantial evidence that pyrethrin autoxidation occurs
Figure 3. Calculated stabilities (kJ/mol) of radicals produced by
hydrogen abstraction from various positions on the pyrethrin I
molecule.
autoxidation chain is the abstraction of hydrogen from the
substrate to generate a substrate radical (eq 2). The investigation
of the autoxidation chemistry of pyrethrins described here began
with semiempirical calculations of the relative stabilities of the
possible free radicals generated by hydrogen abstraction from
pyrethrin I. The MOPAC semiempirical molecular orbital
program using the AM1 parametrization was employed, and
relative radical stabilities were calculated via the UHF method.
The results are displayed in Figure 4 and predict that hydrogen
abstraction from the 7′-position of pyrethrin I would produce the
most stable free radical. This is not unexpected, given that such a
radical would be highly delocalized.
Autoxidation experiments were carried out with the six
pyrethrin esters (Figure 1) following a procedure that was
modified from that reported for the autoxidation of fatty acids.10
Cyclohexane solutions of the isolated pyrethrin esters (0.1 M)
containing AIBN (0.01 M) were heated under air at 64 °C in
sealed vials. Vials were allowed to cool and quenched with BHT
at various time points and their contents analyzed by UPLC. The
extent of the disappearance of the pyrethrin esters over time was
measured relative to the BHT peak. The analysis of the AIBN-
initiated reaction mixtures showed substantial degradation of
E
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 5. UPLC trace of the extract of T. cinerariifolium crop that had been stored for 12 weeks without BHT (B), compared with UPLC traces of
pyrethrin I (A) and II (C) autoxidation mixtures.
when crop material is stored for extended periods in the absence
of antioxidants.
peak due to residual cinerin II was observed in the pyrethrin II
mixture as the pyrethrin II used had some cinerin II
contamination. As found with the autoxidation experiments
with pure pyrethrins, pyrethrin II was found to degrade more
quickly than cinerin II in this latter autoxidation experiment.
Stored crop was freshly extracted, and all three samples, the two
resulting from autoxidation reactions and the crop extract, were
analyzed by LC-MS (Figure 6).
Of the components of the pyrethrin I autoxidation mixture,
identified as potentially important on the basis of the arguments
given above, peaks 5 and 6, the compound responsible for peak 6
could not be readily identified in the LC-MS trace of stored crop
The contention that autoxidation is a contributory cause of
pyrethrin losses during crop storage would be further supported
if some of the autoxidation products generated in laboratory
experiments could be isolated, identified, and shown unambig-
uously to be present in stored crop material. The autoxidation
reactions were therefore repeated with pyrethrin I and II on a
gram scale, using a 24 h reaction time with the hope of generating
enough of the autoxidation products to allow isolation and
characterization. These samples had UPLC profiles very similar
to those produced on a smaller scale, although a more significant
F
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 6. Relevant sections of LC-MS trace of the extract of T. cinerariifolium crop that had been stored for 12 weeks without BHT (A), compared with
LC-MS traces of pyrethrin I (B) and II (C) autoxidation mixtures and a CH3CN blank (D). Peak detection was via ion current in the mass range m/z
250−600.
extract as this peak was obscured by large peaks due to pyrethrin
II and cinerin II in the crop extract. Peak 5 in the LC-MS trace of
stored crop extract appeared to represent two separate
compounds, one of which matched a compound in the pyrethrin
I autoxidation mixture, whereas the other matched a compound
in the pyrethrin II autoxidation mixture.
The chosen components of the pyrethrin II autoxidation
mixture, peaks 3 and 4, could be observed as relatively small
peaks in the LC-MS trace of stored crop extract. It was found that
the larger of the two peaks, peak 4, had the same dominant
molecular ion (m/z 389.1959/389.1960, corresponding to a
molecular formula of C22H29O6 for this M + H+ peak) in both
mixtures. The smaller peak, peak 3, was found to have at least two
matching components (m/z 389.1958 and 411.1775) as well as
matching components from the pyrethrin I autoxidation mixture
(m/z 309.1695 and 343.1903).
As the LC-MS results suggested that peak 4 resulted from a
single component, it was chosen for purification and character-
ization. The whole of the remaining pyrethrin II autoxidation
reaction mixture (∼1 g) was passed through a preparative scale
C18-reversed phase HPLC column using aqueous acetonitrile as
the eluent, and the compound of interest was obtained as a
colorless oil (compound A). LC-MS analysis of this compound
was consistent with that previously obtained in crop and
autoxidation mixtures; however, UPLC analysis showed it to be
impure. HPLC purification was repeated two more times, but the
compound remained contaminated. It became apparent that
compound A was unstable in the aqueous acetonitrile eluent and
consistently degraded to a component that eluted much earlier,
with a UPLC retention time of 24.8 min (peak 1). This latter
compound, compound B, was obtained in high purity, and mass
spectral analysis showed it to have molecular ions at m/z
407.2063, corresponding to a molecular formula of C22H31O7 for
the M + H+ peak, and m/z 429.1882, corresponding to a
molecular formula of C22H30O7Na for the M + Na+ peak. The
strongest peak in the spectrum had m/z 389.1959, corresponding
to (M − H2O) + H+ and had a mass identical to that of
compound A, from which it is apparently formed. This mass
spectrometric evidence strongly suggested that there was a
component, compound A, formed in the autoxidation reaction of
pyrethrin II that was also present in stored crop extract, and that
this compound was hydrolytically unstable, hydrolyzing to
compound B.
Single-ion LC-MS analysis of stored crop extract was used to
further confirm that compound A was present in stored crop. A
similar analysis was used to show that compound B was also
present in stored crop. The UPLC retention time for this
compound was also coincident with a peak in stored crop extract.
A small peak at the same retention time was observed in
pyrethrin II autoxidation mixtures. It is logical that compound B
was formed in only small amounts in the autoxidation
experiments as these were performed in dried solvent.
Compound B showed a strong infrared band at 3435 cm−1,
consistent with the presence of an alcohol functionality.
Conversely, compound A did not show a strong IR band in the
same region. Compound A has a molecular formula of C22H28O6,
which is related to the molecular formula of pyrethrin II by the
addition of a single oxygen. At this point, suspicions arose that
compound A could be an epoxide and compound B a 1,2-diol,
formed by hydrolytic opening of the epoxide. This is consistent
with previous observations of the behavior of the pyrethrin
epoxides shown in Figure 2.
Compounds A and B were extensively investigated using one-
and two-dimensional NMR spectroscopy. It was soon estab-
lished that compounds A and B were structurally related to
pyrethrin II, with both possessing the entire chrysanthemic
diacid and cyclopentenone part of pyrethrin II. The unsaturated
side chains of compounds A and B, however, were different from
G
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
pyrethrin II. Two-dimensional COSY, HSQC, and HMBC
spectra were used to establish that the structures of compounds A
and B are the diastereomeric epoxide 10 and the diastereomeric
diol 11 (Figure 1). These compounds are very closely related to
the structures of the 8′,9′-epoxide 12 and 8′,9′-diol 7 derived
from pyrethrin II. In fact, the NMR spectra of compounds A and
B are very similar to those of the previously prepared
compounds, although they are sufficiently different to be certain
that they are not the same compounds. In addition, the UPLC
retention times for compounds A and B are slightly different
from those of 7 and 12. It has been shown that epoxides are
formed during autoxidation reactions of lipids.11 The mechanism
for the formation of compound A (6) has not been determined
but is very likely derived from the free radical formed by
hydrogen abstraction from the 7′-position of pyrethrin II.
In summary, the oxidation chemistry of the pyrethrins was
investigated in an attempt to identify the causes of pyrethrin
losses during crop storage. Compounds previously identified as
metabolic oxidation products of pyrethrin II were prepared and
used as analytical standards to investigate pyrethrum crop
extracts; however, they were not detected in these extracts. It is
thought, therefore, that biological degradation processes are
unlikely to be the cause of pyrethrin losses during crop storage.
Autoxidation experiments were performed with isolated
pyrethrins for the first time. It was found that the pyrethrins
bearing the butadiene side chain, pyrethrin I and II, degrade
significantly more quickly than the jasmolins and cinerins. The
autoxidation of pyrethrin I and II was studied in more detail and
found to yield a complex mixture of products, a number of which
appeared to be also present in crop extracts. In particular, the
7′,8′-epoxide and 7′,8′-diol of pyrethrin II were isolated from
autoxidation mixtures and confirmed to be present in extracts of
crop material that had been stored for an extended period
without BHT. These results provide strong evidence that
pyrethrin autoxidation occurs in pyrethrum crop on standing
and hence may contribute to pyrethrin losses during crop
storage. It is expected that the pyrethrin autoxidation products
identified here will be suitable analytical standards useful in
gauging the effectiveness of amelioration measures aimed at
minimizing pyrethrin losses during postharvest processing.
Manufacturing Program, including the funding of Stephanie
Mauger’s internship.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Jo Cosgriff and Carl Braybrook, CSIRO, for LC-MS
analyses and Roger Mulder, CSIRO, for assistance with the
interpretation of NMR spectra. We also thank Pree Johnston,
CSIRO, for performing the curve fitting for Figure 4.
REFERENCES
■
(1) Pyrethrum Flowers: Production, Chemistry, Toxicology, and Uses;
Casida, J. E., Quistad, G. B., Eds.; Oxford University Press: New York,
1995.
(2) Kikuta, Y.; Yamada, G.; Mitsumori, T.; Takeuchi, T.; Nakayama,
K.; Katsuda, Y.; Hatanaka, A.; Matsuda, K. Requirement of catalytic-triad
and related amino acids for the acyltransferase activity of Tanacetum
cinerariifolium GDSL lipase/esterase TcGLIP for ester-bond formation
in pyrethrin biosynthesis. Biosci., Biotechnol., Biochem. 2013, 77, 1822−
1825.
(3) Angioni, A.; Dedola, F.; Minelli, E. V.; Barra, A.; Cabras, P.; Caboni,
P. Residues and half-life times of pyrethrins on peaches after field
treatments. J. Agric. Food Chem. 2005, 53, 4059−4063.
(4) Minello, E. V.; Lai, F.; Zonchello, M. T.; Melis, M.; Russo, M.;
Cabras, P. Effect of sunscreen and antioxidant on the stability of
pyrethrin formulations. J. Agric. Food Chem. 2005, 53, 8302−8305.
(5) Atkinson, B. L.; Blackman, A. J.; Faber, H. The degradation of the
natural pyrethrins in crop storage. J. Agric. Food Chem. 2004, 52, 280−
287.
(6) Chen, Y.-L.; Casida, J. E. Photodecomposition of pyrethrin I,
allethrin, phthalthrin and dimethrin. Modifications in the acid moiety. J.
Agric. Food Chem. 1969, 17, 208−215.
(7) Casida, J. E.; Kimmel, E. C.; Elliott, M.; Janes, N. F. Oxidative
metabolism of pyrethrins in mammals. Nature 1971, 230, 326−327.
(8) Williams, A. M. Degradation of Natural Pyrethrins in Crop Storage.
Bachelor of Science Honours Thesis, University of Tasmania, Hobart,
Australia, 2003.
(9) Coomber, D. C.; Tucker, D. J.; Bond, A. M. Electrochemical
oxidation of pyrethroid insecticides at glassy carbon electrodes in
acetonitrile. Electroanalysis 1998, 10, 163−172.
(10) Tallman, K. A.; Roschek, B.; Porter, N. A. Factors influencing the
autoxidation of fatty acids: effect of olefin geometry of the non-
conjugated diene. J. Am. Chem. Soc. 2004, 126, 9240−9247.
(11) Yin, H.; Xu, L.; Porter, N. A. Free radical lipid peroxidation:
mechanisms and analysis. Chem. Rev. 2011, 111, 5944−5972.
(12) Ando, T.; Toia, R. F.; Casida, J. E. Epoxy and hydroxyl derivatives
of (S)-bioallethrin and pyrethrins I and II: synthesis and metabolism. J.
Agric. Food Chem. 1991, 39, 606−611.
(13) Clark, M.; Cramer, R. D.; Vanopdenbosch, N. Validation of the
general-purpose Tripos 5.2 force-field. J. Comput. Chem. 1989, 10, 982−
1012.
(14) Ortiz de Montellano, P. R. Hydrocarbon hydroxylation by
cytochrome P450 enzymes. Chem. Rev. 2010, 110, 932−948.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jafc.6b02959.
Representative UPLC traces of pyrethrin autoxidation
mixtures; data from AIBN-promoted autoxidation experi-
ments; LC-MS analysis of individual peaks in autoxidation
mixtures and crop extracts; figures showing UPLC, LC-
MS, IR, and NMR characterization of compounds A and B
(PDF).
AUTHOR INFORMATION
■
Corresponding Author
*(P.J.D.) Phone: +61-39545 2560. E-mail: peter.duggan@csiro.
au.
Funding
The research described here was funded by Horticulture
Australia Limited with co-investment from Botanical Resources
Australia and funds from the Australian Government. Co-
investment was also provided by the CSIRO Biomedical
H
DOI: 10.1021/acs.jafc.6b02959
J. Agric. Food Chem. XXXX, XXX, XXX−XXX