Anal Bioanal Chem (2004) 378 : 753763

DOI 10.1007/s00216-003-2377-0

O R I G I N A L PA P E R

Sami Barrek Olivier Paisse

Marie-Florence Grenier-Loustalot

Analysis of neem oils by LCMS and degradation kinetics

of azadirachtin-A in a controlled environment
Characterization of degradation products by HPLCMSMS
Received: 31 July 2003 / Accepted: 28 October 2003 / Published online: 18 December 2003
Springer-Verlag 2003

Abstract Since it was first isolated, the oil extracted

from seeds of neem (Azadirachtin indica A juss) has been
extensively studied in terms of its efficacy as an insecticide. Several industrial formulations are produced as
emulsifiable solutions containing a stated titer of the active ingredient azadirachtin-A (AZ-A). The work reported
here is the characterization of a formulation of this insecticide marketed under the name of Neem-azal T/S and kinetic studies of the major active ingredient of this formulation. We initially performed liquidliquid extraction to
isolate the neem oil from other ingredients in the commercial mixture. This was followed by a purification using flash chromatography and semi-preparative chromatography, leading to 13C NMR identification of structures such as azadirachtin-A, azadirachtin-B, and azadirachtin-H. The neem extract was also characterized by HPLCMS
using two ionization sources, APCI (atmospheric pressure
chemical ionization) and ESI (electrospray ionization) in
positive and negative ion modes of detection. This led to
the identification of other compounds present in the extract azadirachtin-D, azadirachtin-I, deacetylnimbin, deacetylsalannin, nimbin, and salannin. The comparative
study of data gathered by use of the two ionization sources
is discussed and shows that the ESI source enables the
largest number of structures to be identified. In a second
part, kinetic changes in the main product (AZ-A) were
studied under precise conditions of pH (2, 4, 6, and 8),
temperature (40 to 70 C), and light (UV, dark room and
in daylight). This enabled us to determine the degradation
kinetics of the product (AZ-A) over time. The activation
energy of the molecule (759 kJ mol1) was determined
by examining thermal stability in the range 40 to 70 C.
The degradation products of this compound were identified by use of HPLCMS and HPLCMSMS. The results enabled proposal of a chemical degradation reaction

S. Barrek O. Paisse M.-F. Grenier-Loustalot ()

Service Central dAnalyse, USR 059, CNRS, BP 22,
69390 Vernaison, France
e-mail: mf.grenier-loustalot@sca.cnrs.fr

route for AZ-A under different conditions of pH and temperature. The data show that at room temperature and pH
between 4 and 5 the product degrades into two preferential forms that are hydrolyzed to a single product over time
and as a function of pH change.
Keywords Neem Triterpenoids Azadirachtin-A
Degradation Kinetics HPLCMSMS

Introduction
The oil extracted from seeds of neem (Azadirachtin indica
A juss) has been the subject of considerable research. After isolation of this oil in seminal work by Morgan and
Butterworth in 1968 [1], research on its properties has
consistently shown the efficacy of azadirachtin as an insecticide [2, 3, 4, 5]. When applied to larvae it causes
mortality at different stages of their development and malformations (in particular reduced longevity and fertility of
adults). The 50% lethal dose (LD50) varies with species
from 1 to 4 g azadirachtin per gram of insect weight [2].
The industrial use of neem extracts has now become a reality as a result of these ecotoxicological characteristics,
with the appearance on the market of a variety of commercial formulations for varied applications: protection of
plants and stock [6]. Most products are in the form of emulsifiable solutions, among which are Neem-Azal F, NeemAzal T, and Neem-Azal T/S. All formulations on the market indicate the quantity of the active ingredient only, azadirachtin-A, which is generally found in company of eight
other compounds (Fig. 1) in varying proportions.
Several publications have dealt with different techniques
for the separation and identification of products contained
in neem extracts [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16].
Separation in these studies involved reversed-phase liquid
chromatography (RP-HPLC) [6, 8, 9, 11, 12, 13, 14, 15,
16], using octadecylsilane (C18) columns and detection by
ultraviolet absorption. These coupled methods are effective for separation and detection of compounds present in
the oil extracted from neem but are insufficient to identify

754
Fig. 1 Chemical structures of
the compounds present in neem
extracts

all the compounds in insecticide formulations, because of

their chemical complexity. This work employed liquid
liquid extraction procedures before separation [5, 6, 12,
13]. Other liquid chromatography techniques, e.g. flash
chromatography [13] and high-pressure preparative chromatography [8, 10, 15], respectively, have been used to
separate and purify compounds in extracts.
Two techniques, nuclear magnetic resonance spectroscopy (NMR) and mass spectrometry (MS), have been
used to identify chemical structures. NMR is the most
widely used technique in the literature, especially liquidstate 13C NMR [7, 8, 17, 18, 19].
When using MS, gentle ionization techniques are required because most compounds in neem extracts are

thermolabile. The method most often used is FAB (fast

atom bombardment) with a high-resolution magnetic analyzer [7, 18, 20]. This method requires considerable sample-preparation time, especially for separation and purification of extracts. Using a coupled HPLCMSMS system with APCI as the ionization technique Schaaf et al.
[21] analyzed products in extracts. Only four neem products were detected with this ionization technique.
The properties of azadirachtin make it a very good substitute for synthetic pesticides. Unfortunately, use of azadirachtin is limited by its sensitivity to light and to acid
and alkaline media [5, 22, 23]. The degradation of azadirachtin-A has been investigated as a function of the acidity of the medium [22, 23]. The transformation product of

755

azadirachtin-A under the action of UV was identified by

Shaun et al. [16] and Johnson et al. [22], but the degradation products in aqueous media remain unknown and are
the subject of this work.
In this work we used HPLCMS coupled with two ionization techniques to achieve rapid identification of neem
products and comparison of APCI and ESI. To characterize the chemical structures of the species present in the
extract, however, several isolation and purification steps
were applied to each product to determine kinetic behavior in a reaction mixture in a controlled environment. In a
second part, we describe a study of the degradation kinetics of azadirachtin-A as a function of temperature, pH,
and light in order to determine the number of breakdown
products. We then propose chemical structures of these
products, identified by use of HPLCMS and HPLCMS
MS.

Material and methods

Solvents and reagents
The products used were all HPLC-grade HPLC water (Millipore,
France), acetonitrile (Merck, France), dichloromethane, n-hexane
(SDS, France), and methanol (Prolabo, France). All these HPLCgrade solvents were used for both liquid chromatography and the
different liquid extraction steps. The azadirachtin-A standard was
purchased from SigmaAldrich, France, and Neem-Azal T/S (1%
azadirachtin-A) is marketed by Trifolio-M, Germany.
Azadirachtin-A purification procedure
Liquidliquid extraction
The extraction method used was that of Sundaram [12]. The commercial formulation (Neem AzalT/S; 50 mL) was mixed with
methanolwater (90:10 (v/v), 1 L). The alcoholic mixture was first
subjected to two liquidliquid extractions with a mixture of 125 mL
hexane and 125 mL water. Apolar or slightly polar compounds that
limit the extraction of azadirachtin-A in subsequent steps are eliminated in the hexane phase.
The active ingredient in the alcoholic phase is extracted twice
with 100 mL dichloromethane. After separation, the dichloromethane phase is recovered and evaporated at 50 C under nitrogen,
yielding a yellowish liquid.
Chromatographic purification and preparation
Chromatographic purification was carried out in two steps, flash
chromatography followed by semi-preparative chromatography, to
isolate the quantity of product required for our work.
Flash chromatography purification. The extraction procedure used
was that of Yamasaki et al. [13] with several improvements. Silica
gel (300 g) was mixed with approximately 500 mL mobile phase in
a glass column. The yellowish liquid obtained by extraction of the
commercial product was deposited on the stationary phase at the
head of the column. The product was then eluted under slight nitrogen pressure with ethermethanol, 49:1 (v/v) as mobile phase.
Sixty fractions were collected.
Analytical determination was by thin-layer chromatography
and visualization was with sulfuric acid. Fractions were pooled according to the RF values of their components and concentrated by
evaporation of the solvent. The dry residue was then dissolved in
2 mL acetonitrile and purified by semi-preparative chromatography.

Preparation by semi-preparative chromatography. The commercial product was purified by semi-preparative chromatography using an HP 1050 instrument (HewlettPackard) equipped with a
quaternary pumping system and manual Rheodyne 7525 injection
valve with 100-L loop and coupled to a Kratos Spectroflow 757
diode-array UV detector and an HP 3396A integrator-recorder.
Separation was achieved on a 250 mm10 mm Interchrom Uptisphere ODSB column (Interchim). The mobile phase was water
and acetonitrile with an elution gradient of 16 to 100% acetonitrile
in 40 min at a flow-rate of 4.4 mL min1. Detection was at the
wavelength =215 nm. After several repeated injections of the
same fraction, azadirachtin-A was recovered and evaporated under
a stream of nitrogen. To eliminate solvent traces for further NMR
analysis azadirachtin-A was then treated with a vacuum pump (at
15 mm Hg) for several hours.
Characterization of chemical species isolated
Liquid chromatography (LCUVMS)
Analyses were carried out with an HP1100-MSD system. The
chromatograph was an HP 1100 equipped with a 250 mm3 mm
Interchim Uptisphere ODSB column maintained at 40 C and fitted
with a 10 mm3 mm pre-column. UV detection was performed at
=215 nm. The mobile phase was water and acetonitrile with an elution gradient of 16% to 100% acetonitrile in 40 min at a flow-rate
of 0.4 mL min1. Samples were diluted in acetonitrile and filtered
through a 0.45 m pore diameter Teflon membrane before analysis.
Mass detection involved two ionization modes at available atmospheric pressure APCI and ESI using two polarities (positive or
negative). The acquisition conditions for mass spectra were thoroughly optimized. The flushing gas was ultrapure nitrogen at a pressure of 55 psig, a temperature of 350 C and a flow-rate of 12 L min1.
The mass-range detected was between 200 and 800 m/z, with fragmentor voltages of +80 V and +100 V for ESI and APCI, respectively, depending on the polarity of the ions detected. In APCI
mode vaporizer temperature was 480 C and the discharge current
at the corona needle was set at 35 A under negative APCI conditions and at 8 A in positive mode. Ion-extraction voltages applied
to the transfer capillary in the two ionization sources with negative
and positive polarity were 2500 V and 3500 V.
To enhance ionization a T-connector was placed between UV
detector output and the MSD source input. When using negative
polarity a mixture of water, ammonium formate, and aqueous ammonia solution at pH 10.5 was added before the sources (APCI and
ESI) of the mass spectrometer at flow-rates of 0.6 and 0.2 mL min1,
respectively. In positive polarity, a mixture of water and formic
acid, pH 3.7, was added under the same conditions as above.
Nuclear magnetic resonance (NMR)
Liquid state 13C NMR spectra (50 MHz) were recorded with a Bruker
Avance 200 spectrometer equipped with a 5-mm 1H/13C probe.
Samples were dissolved in deuterated chloroform [13C(CDCl3)=
77 ppm] and analyses were conducted at room temperature. Acquisition conditions were: pulse angle 30, spectral width 12,000 Hz,
acquisition time 1.41 s, and repetition interval 2 s. The number of
accumulations was set at 16,000 to obtain a good signal/noise ratio.
Experimental design and monitoring of degradation kinetics
The degradation of azadirachtin-A was studied under different
conditions of pH, light, and temperature.
Preparation of solutions
Several solutions of azadirachtin-A were prepared at different pH.
Depending on the target pH they were buffered and adjusted to the
final value with phosphoric acid (pH 20.1), acetic acid (pH 40.1
or 60.1), or sodium carbonate (pH 80.1)

756
Kinetic changes during temperature variation
This study was performed in a dark room to limit photodegradation. Degradation of azadirachtin-A was followed at 40, 50, and
70 C.
The disappearance of azadirachtin-A was followed by means of
liquid chromatography with determination of the concentration of
residual active ingredient using the external standard method.

trile. HPLCMSMS analyses were carried out using a Waters

2690 separation module chromatographic system coupled to a
Waters 996 photodiode-array UV detector and a Micromass Quattro II triple-quadrupole mass spectrometer. The chromatographic
conditions for elution were the same as those described above.
We used ESI with positive and negative polarity under the same
conditions as described for HPLCMS. Collision energy used in
tandem MS was 20 eV and the collision gas was argon (Ar).

Degradation kinetics under UV radiation

Results and discussion

Irradiation device. The device used comprised a UV-exposure
chamber containing a medium-pressure Hg lamp (Heraeus Noblelight) with the characteristics: emission lamp TQ 150, power 150 W,
light flux 6.2 W for UV-C, 3.6 W for UV-B and 4.5 W for UV-A.
Irradiation. Freshly prepared samples in 10-mL sealed quartz tubes
were placed in the exposure chamber 17 cm from the emission
source at room temperature. To prevent excessive increase of the
temperature of the sample as a result of irradiation it was continuously cooled with a stream of air. The disappearance of the product was followed at intervals of 70 min by liquid chromatography
until 90% of the initial concentration was reached.
Identification of breakdown products
The different products formed during these treatments were identified by liquid chromatography coupled to mass spectrometry
(HPLCMS and HPLCMSMS). The chromatographic conditions were as described above, except that the mobile phase was
water (adjusted to pH 6.3 with ammonium formate) and acetoni-

Fig. 2 Chromatographic profile obtained from the separation method developed

Identification of nine products

isolated from neem extracts by HPLCUVMS
The yellowish liquid solution obtained after liquidliquid
extraction was diluted with acetonitrile and injected (1 L)
on to the chromatographic column; a profile typical of
neem oils (Fig. 2) was obtained. To attribute the structure
corresponding to each chromatographic peak (Fig. 1) we
performed HPLCUVMS with ESI (positive and negative) and APCI (positive and negative).
Table 1 shows the ions detected for each molecule by
use of the two ionization techniques, and their retention
times. Also presented in this table are the attributions for
the base peaks and the most important ions detected. The
results show that deacetylslannin, deacetylnimbin, slannin
and nimbin were not detected in negative ESI and APCI.

757
Table 1 Retention times, peaks in ESI and APCI mass spectra, and attributions for Neem compounds
Compounds

Azadirachtin I

Retention
time
(min)

ESI mode

APCI mode

m/z (ESI+) Attribution

m/z(ESI) Attribution

m/z (APCI+) m/z (APCI)

17.1

641a

[M+Na]+

617a

[MH]

ndb

617a

501

[M+H(H2O+tigOH)]+

599
557
517

[M(H+H2O)]
[M(H+AcOH)]
[M(H+tigOH)]

Azadirachtin H

17.9

685a
645
627
545

[M+Na]+
[M+HH2O]+
[M+H2H2O]+
[M+H(H2O+tigOH)]+

661a
643
601
561

[MH]
[M(H+H2O)]
[M(H+AcOH)]
[M(H+tigOH)]

nd

661a

Azadirachtin D

20.3

699a
677
541

[M+Na]+
[M+H]+
[M+H(2H2O+tigOH)]+

675a
643
615

[MH]
[M(H+MeOH)]
[M(H+AcOH)]

nd

675a

Azadirachtin A

21.3

743a
721
685
585

[M+Na]+
[M+H]+
[M+H2H2O]+
[M+H(2H2O+tigOH)]+

719a
687
659

[MH]
[M(H+MeOH)]
[M(H+AcOH)]

703a

719a

Azadirachtin B

21.9

685a
645
627
545

[M+Na]+
[M+HH2O]+
[M+H2H2O]+
[M+H(H2O+tigOH)]+

661a
643
599

[MH]
[M(H+H2O)]
[M(H+H2O+CO2)]

nd

nd

Deacetylnimbin

28.8

521a
467
449

[M+Na]+
[M+HMeOH]+
[M+H(MeOH+H2O)]+

nd

499a

nd

Deacetylslannin

29.3

577a
555
523

[M+Na]+
[M+H]+
[M+HMeOH]+

nd

555a

nd

Nimbin

31.0

563a
509
481

[M+Na]+
[M+HMeOH]+
[M+HAcOH]+

nd

541a

nd

Slannin

32.1

619a
597
459

[M+Na]+
nd
[M+H]+
[M+Na(AcOH+tigOH)]+

597a

nd

aBase
bNot

peak
detected

This could be explained by a lack of hydroxide groups in

the corresponding structures (only one OH group occurs
in the structures of deacetylslannin and deacetylnimbin).
Also, azadirachtin I, azadirachtin H, azadirachtin D, and
azadirachtin B are not detected in positive APCI. The
most suitable ionization mode for analysis of the neem extract is positive ESI, because all the base peaks for each
compound are detected.
All the mass spectra are available on-line in the supplementary material.
NMR characterization of neem compounds
Because of the large number of protons and the complexity of the compounds structures, 1H NMR spectra were
complex and the lack of resolution made proton attribution difficult. 13C NMR spectra on the other hand were

useful [7, 8, 17, 18, 19] as a result of the greater spectral

range of this atom (200 ppm).
Among the nine neem compounds that could be isolated by the long purification procedure, only three could
be characterized by NMR. The very small quantities of
compounds recovered (several mg) did not furnish a sufficient signal-to-noise ratio for all the samples.
The three compounds that could be studied were AZ-A,
AZ-B and AZ-H. Table 2 lists the complete attributions of
the chemical shifts of all the carbons. These attributions
were obtained by use of DEPT pulse sequences and theoretical calculation rules for different chemical shifts. Carbons in position 2 of each compound can be used as a molecular probe when following the disappearance of these
compounds as the reaction mixture changes (degradation,
transformation) under chemical restrictions in a controlled
environment.

758
Table 2 13C NMR chemical shifts (, ppm) of the three compounds isolated (solvent CDCl3, reference TMS)
Number of
carbon atoma

Azadirachtin-A Azadirachtin-B Azadirachtin-H

2
3
4
5
6
8
9
10
11
12
13
14
15
17
19
20
22
23
24
25
27
28
29
30
32
34
36
38
41
44
47
48
49
50
51

104.2
44.9
45.6
74.5
73.9
73.0
53.1
37.1
50.4
69.1
67.1
29.8
70.7
173.3
52.7
70.1
68.6
48.7
83.6
108.8
76.5
25.1
107.5
147.0
21.4
171.8
52.6
169.5
166.2
18.4
20.8
128.8
137.5
14.2
11.9

aSee

79.5
43.9
44.1
73.6
74.3
73.3
53.4
35.2
51.3
69.5
67.7
32.1
71.4
174.1
52.7
69.4
66.6
48.9
83.6
109.1
76.1
25.1
107.5
146.7
21.2
173.4
52.4
b
166.9
18.1
b
128.6
138.5
14.2
11.9

101.0
48.5
43.3
76.2
74.0
70.0
52.3
37.1
47.7
73.0
67.1
30.8
72.2
b
b
73.0
69.8
48.2
83.5
108.6
77.2
25.3
107.5
147.2
20.6
173.5
52.7
169.8
166.7
18.6
20.6
128.7
138.2
14.4
11.9

Fig. 1 for carbon-atom numbering

not present in the structure

bCarbon

Kinetic and thermodynamic studies

Table 3 Half-life (min) as a function of light and pH

pH

t1/2(min)

2
4
6
8

Darkness

Sunlight

UV

14032
141133
188278
75357

14097
142413
63360
62053

340
1274
1410
1050

Table 4 Rate constants (k1), half-life times (min), and activation

energies (Ea) as a function of temperature and pH
k1(s1)

Ea
(kJ mol1)

1160
385
116

1.00105
3.00105
1.00104

66.9

40
50
70

23104
11552
1650

5.00107
1.00106
7.00106

79.9

40
50
70

23104
2888
1155

5.00107
4.00106
1.00105

83.4

40
50
70

38508
11552
1444

3.00107
1.00106
8.00106

77.7

pH

Temperature (C)

40
50
70

t1/2(min)

Effect of temperature
The effect of temperature on degradation kinetics was
studied at the same pH. We noticed that the rate of disappearance of azadirachtin-A increased with temperature in
dependently of pH. For a given temperature half-life values
tended to increase as pH increased, except for the combination (50 C, pH 6) (Table 4). These results show that the
disappearance kinetics of azadirachtin-A depend on media pH and on temperature.
Activation energy calculations
The curves of the disappearance of azadirachtin-A vs.
time in all the experiments were exponential. Based on

Influence of light
Three experiments were carried out to determine the influence of light on the degradation kinetics of azadirachtin-A. The first was at room temperature and in daylight,
the second with UV irradiation, and the third in darkness.
The influence of these conditions was also examined at
pH 2, 4, 6, and 8. The half-lives measured (Table 3) show
that kinetics of disappearance are higher under UV illumination than in daylight or darkness at the four pH values
used. When the influence of daylight and darkness was
examined, no notable variations were observed at pH 2, 4
and 8; at pH 6, however, disappearance in daylight was
three times faster than in the dark.

Fig. 3 Variation of ln (A) as a function of pH

759

these experimental data, it can be deduced that the disappearance of azadirachtin-A is first order (apparent
order).
The experimental data enable the apparent rate constant (k1) to be determined graphically by plotting the
variation of ln (Cr) vs. time. By applying Eq. (1) to the experimental data, we could determine the resulting rate
constants and activation energies (Table 4):
(1)
. = $H (D 57
where Ea is the activation energy (kJ), R the ideal gas constant (J), and T the temperature (K).
Fig. 4 Structures of azadirachtin-A and its degradation products

The variation of the activation energy as a function of

pH was studied. The resulting curve is bell-shaped with a
maximum determined at pH 5.7 and a minimum at pH 2.
In addition, the activation energy calculated at pH 8 is
smaller than that at pH 4. On this basis, one can expect the
half-lives to evolve in the sequence t1/2(pH 2)<t1/2(pH 8)
<t1/2(pH 4)<t1/2(pH 6), because the energy barrier is more
important at pH 6 and 4 than at pH 8 and 2. Surprisingly,
t1/2(pH 8) was higher than t1/2(pH 6) for the temperature
values studied. To explain these results, we examined the
variation of ln(A) as a function of pH (Fig. 3). The resulting curves have a similar evolution as the curve obtained for

760
Fig. 5 Chromatograms obtained from degradation products: (A) pH 2, (B) pH 4 and 6,
(C) pH 8

the variation of the activation energy vs. pH (bell-shaped).

This result shows that A depends on pH and on temperature, as indicated by the work of Laidler [22]. Thus, we are
investigating a complex phenomenon where both A and Ea
vary as a function of the pH of the media.
For azadirachtin-A, the activation energy was 75
9 kJ mol1.
Analysis of degradation products
We detected several degradation products in the reaction
mixture (Fig. 4). Using separation conditions already developed, we identified these products by LCMS. Figure 5
shows the chromatograms obtained at the different pH
values studied.
Identification of breakdown products
The chemical structures of the products could be unambiguously characterized by use of their mass spectra. The
different positional isomers possible could not, however,

be identified by simple HPLCMS analysis. Using the

data obtained with HPLCMSMS, on the other hand, we
could identify the precise structure of each isomer in the
reaction mixture. This study was made possible by fine
analysis of each of the spectra.
All the mass spectra are available on-line in the supplementary material.
Analysis of product P1. The mass spectrum of product P1
in negative ESI mode has a base peak at m/z 723 corresponding to the [MH] ion. The ion at m/z 691 is formed
by the elimination of a molecule of methanol from the ion
at m/z 723 [M(H+MeOH)]. The peak at m/z 663 results
from the loss of a molecule of acetic acid [M(H+
AcOH)], and loss of a molecule of carbon dioxide yields
the [M(H+AcOH+CO2)] ion at m/z 619.
The mass spectrum of product P1 in positive ESI contains two low-intensity pseudomolecular peaks at m/z 725
and 747 that are attributed to the [M+H]+ and [M+Na]+
ions. The base peak at m/z 689 corresponds to loss of two
water molecules from the [M+H]+ ion. The peak detected
at m/z 707 corresponds to the elimination of one water
molecule [M+HH2O]+.

761

Based on these data, we can deduce that the molar mass

of product P1 is 724 g mol1. This mass is higher than that
of azadirachtin-A (720 g mol1) by 4 amu. At pH 2 this
structure can be attributed on the basis of the hydrolysis of
one of the two ester groups in azadirachtin-A in positions
2 and 9, and by addition of one molecule of water to the
double bond at position 4849.
This structural hypothesis was confirmed by use of
HPLCMSMS coupling. The mass spectrum of the azadirachtin-A standard shows loss of 100 Da, characteristic
of tiglic acid (tigOH). The mass spectrum of the 707 Da
daughter ions reveals loss of 118 Da from the 707, 689
and 671 Da ions. Assuming this loss is simultaneous elimination of one water and one tiglic acid molecule, we recorded
the mass spectrum of the neutral loss of 118 Da. It showed
the elimination of a neutral fragment of 118 Da from the
725 [M+H]+, 707 [M+HH2O]+, 689 [M+H2H2O]+, and
671 [M+H3H2O]+ ions, confirming addition of a water
molecule to the double bond of the tiglic group at position
4849.
The HPLCMS coupling study of azadirachtin-A and
its homologues (azadirachtin-B, azadirachtin-H and azadirachtin-D) showed that elimination of a molecule of methanol in negative ESI is possible only between the ester
function and the hydroxyl group on carbon 2. The mass
spectrum of 723 Da [MH] daughter ions recorded in
negative ESI shows losses of methanol from the 723 Da,
619 Da and 601 Da ions. To confirm this, we recorded the
neutral loss of a 32 Da mass from product P1 and thus obtained confirmation of our hypotheses. These results confirm that the hydrolysis reaction involved the ester function on carbon 9.
Analysis of product P2. Degradation product P2 also occurs at pH 2. The mass spectrum in negative ESI mode has
a base peak at m/z 687, corresponding to the [MH] ion.
The spectrum also contains a peak at m/z 583, characteristic of simultaneous loss of molecules of CO2 and acetic acid
[M(H+CO2+AcOH)]. The presence of a peak at m/z 551
is because of the elimination of a molecule of methanol
from the 583 ion [M(H+CO2+AcOH+MeOH)]. This
elimination characterizes the presence of the ester function and of the hydroxyl group on carbon 2.
The positive ESI mass spectrum contains a low-intensity peak at m/z 689 [M+H]+. Its intensity was amplified by
reducing the fragmentation energy applied in the collision
chamber of the mass spectrometer. The base peak of this
spectrum is at m/z 671, corresponding to the [M+HH2O]+
ion. Under these conditions we also see two peaks at
m/z 653 and m/z 571, corresponding to simultaneous loss
of one water molecule and one tiglic acid molecule, respectively, from the 671 ion. The pseudomolecular peak
at m/z 706 is the ammonium adduct [M+NH4]+. This phenomenon results from use of ammonium salts in the mobile phase.
On the basis of these data obtained with positive and
negative ESI, it can be concluded that the molar mass of
product P2 is 688 g mol1, equivalent to a loss of a mass of
32 g mol1 from the starting product azadirachtin-A. Un-

der these pH conditions we can imagine transformation by

hydrolysis and loss of one water molecule from position
2425, as shown on the reaction path.
As indicated for product P1, the hydrolysis site is determined by elimination of one methanol molecule in the
negative ESI mass spectrum and confirmed by HPLCMS
MS. To determine the elimination position of the water
molecule we examined the mass spectrum of 671 Da daughter ions in positive ESI mode. This revealed the presence
of a peak at m/z 493, loss of 178 Da from the parent ion.
Because this is consistent with dehydration on carbons 24
and 25, we recorded the mass spectrum of the neutral loss
of 178 Da in parallel, confirming our hypothesis. The
spectrum obtained shows that the 178 Da loss is possible
from the 671 and 653 Da ions.
Analysis of product P3. The positive ESI mass spectrum
of product P3 has a base peak at m/z 742, corresponding
to the pseudomolecular ion [M+NH4]+, the formation of
which results from use of ammonium salts in the mobile
phase. Similarly, [M+Na]+ and [M+K]+ pseudo-molecular
ions are observed at m/z 747 and 763. Characteristic fragments seen at m/z 707, 703, and 689 can be attributed to
the [M+HH2O]+, [M+NaCO2]+, and [M+HH2O]+ ions,
respectively. These hypotheses were confirmed by the
mass spectra recorded in MSMS mode.
The negative ESI mass spectrum has a base peak at m/z
723, from the [MH] ion, and two other ions at m/z 705
and 679, from the [M(H+H2O)] and [M(H+CO2)]
fragment ions.
These positive and negative ESI results lead to the conclusion that the mass of degradation product P3 is 724 g mol1
and that product P3 is a P1 positional isomer. MSMS
data indicate the absence of a peak resulting from loss of
a methanol molecule in negative ESI and confirm loss of
a 118 Da fragment for the structure shown in Fig. 4.
Analysis of product P4. The mass spectrum of degradation
product P4 in negative ESI mode has a base peak at m/z
687, ion [MH], and low intensity peaks of fragment
ions, in particular the [MH196] ion, providing information about the structure of this product.
The mass spectrum recorded in positive ESI indicates the
presence of a base peak at m/z 671, the [M+HH2O]+ fragment ion, and three pseudomolecular peaks at m/z 706, 711,
and 727, corresponding to the [M+NH4]+, [M+Na]+, and
[M+K]+ ions. Formation of these adducts results from the
use of ammonium salts in the mobile phase. We also note
the presence of peaks from the loss of tiglic acid at m/z
589 [M+HtigOH]+ and m/z 571 [M+H(H2O+tigOH)]+.
It can be concluded that degradation product P4 is a
positional isomer of product P2, because examining the
neutral loss at 32 Da gave a negative result. This confirms
that the hydrolysis reaction involves the ester function on
carbon 2. The presence of the [MH196] fragment ion
in the negative ESI mass spectrum shows that a water
molecule was eliminated between the hydroxyl group on
carbon 5 and the hydrogen on carbon 6. These results
were confirmed by MSMS detection of a peak at m/z 515

762

in the mass spectra of the 711 Da daughter ions in positive

ESI and the 687 Da daughter ions in negative ESI. These
data enable attribution of the structure given in Fig. 4.
Analysis of product P5. The positive ESI mass spectrum of
degradation product P5 contains a base peak at m/z 701,
the sodium adduct [M+Na]+. The peak at m/z 717 can be
attributed to the potassium adduct [M+K]+. This spectrum
also contains the peak of the [M+HH2O]+ fragment ion
at m/z 661.
On the basis of the data obtained in positive ESI mode
it can be concluded that the molar mass of the latter is
679 g mol1, equivalent to loss of 41 Da from the starting
product azadirachtin-A. This result can be interpreted
as the substitution of the acetate group by hydroxyl via
deacetylation (Fig. 4).
Analysis of product P6. Analysis of degradation product
P6 by negative ESI affords a simple mass spectrum with
only one peak at m/z 705, from the [MH] ion. The positive
ESI mass spectrum, on the other hand, contains a pseudomolecular peak at m/z 729 of the [M+Na]+ ion. This spectrum also contains evidence of successive eliminations of
water molecules from the [M+H ]+ ion at m/z 689, 671, and
653, corresponding to the [M+HH2O]+, [M+H2H2O]+
and [M+H3H2O]+ ions, respectively The ions observed
at m/z 571 and 553 result from losses of tiglic acid:
[M+H(2H2O+tigOH)]+ and [M+H(3H2O+tigOH)]+.
On the basis of these data it can be concluded that the
molar mass of product P6 is 706 g mol1, corresponding to
loss of 14 g mol1 from the starting product. This product
is obtained at pH 4, 6 and 8 and its kinetics of appearance
are accelerated in alkaline medium. These observations
show that one of its ester functions is hydrolyzed. To determine the site of hydrolysis we examined the daughter
ions of [MH] in negative ESI. This spectrum shows the
absence of loss of a methanol molecule that characterizes
the ester function on carbon 2, as shown in Fig. 4.
Analysis of product P7. The negative ESI mass spectrum
of degradation product P7 has a base peak at m/z 691 from
the [MH] ion and a peak at m/z 673 resulting from loss
of a water molecule [M(H+H2O)]. Analysis of the positive ESI mass spectrum shows a pseudomolecular peak at
m/z 715 from the [M+Na]+ ion and peaks of fragment ions
[M+HH2O]+, [M+H2H2O]+, and [M+H3H2O]+ at m/z
675, 657 and 639, respectively. Elimination of tiglic acid
from the last two ions yields the peaks at m/z 557
[M+H(2H2O+tigOH)]+ and 539 [M+H(3H2O+tigOH)]+.
These results enable us to deduce that the molar mass
of this product is 692 g mol1, corresponding to loss of
28 g mol1 from the starting product. This can be interpreted in terms of hydrolysis of two ester functions in the
molecule, results consistent with the chemical structure
shown in Fig. 4.
Analysis of product P8. Degradation product P8 was detected only in negative ESI. Its mass spectrum contained a
peak at m/z 673, attributable to the ion [MH]. This prod-

uct results from transformation of product P7 by the loss

of a water molecule from its precursor. The mass spectra
of [MH] daughter ions reveals the loss of 196 Da, yielding a peak at m/z 477. This result proves that dehydration
occurs between the hydroxyl group on carbon 5 and the
hydrogen on carbon 6. These results enable us to propose
the chemical structure shown in Fig. 4.

Conclusions
An HPLC method has been developed for separation of the
major products of neem. Use of this method with preparative or the equivalent semi-preparative columns enabled
the pure products to be isolated. These were identified by
use HPLCMS in conjunction with two modes of ionization, ESI and APCI. As shown in this work, the best technique for characterizing all the compounds in the neem
extract is positive ESI. This ionization method renders all
molecules detectable and identifiable. It is also the most
sensitive in terms of the total ion-current formed. The results obtained with positive APCI are comparable with
those published by Schaaf et al. [21] and show the fundamental importance of the choice of the ionization source
and polarity for identification and analysis of neem products. Some of the major compounds in extracts could be
analyzed by liquid state 13C NMR. The data obtained agreed
with results obtained by HPLCMS and confirm that neem
extracts contain several compounds whose activities can be
amplified to different extents in complex mixtures. It was
concluded from kinetic studies carried out in this work
that the process of degradation of azadirachtin-A is accelerated by ultraviolet radiation, irrespective of the pH of
the medium (Table 3). This process is more rapid in basic
medium (pH 8) and highly acidic medium (pH 2), for which
the lowest half-reaction times were recorded. In addition,
the effect of daylight is low in comparison with darkness,
except at pH 6. The study of pH variations showed that
the kinetics of disappearance of azadirachtin-A are slowest at pH 6. This was confirmed by the activation energies
calculated as a function of pH (Table 4).
The identification of breakdown products by HPLCMS
and HPLCMSMS in different controlled media showed
that the disappearance of the starting product azadirachtin-A varies with pH. At pH 4, 6, and 8, its degradation is
initiated by hydrolysis of ester functions and the formation of products P6 and P7, confirming the hypothesis put
forward by Sundaram [11]. This hydrolysis is followed by
dehydration to yield product P8. At pH 2, we identified
five breakdown products, four of which are two-by-two
positional isomers ((P1 and P3), (P2 and P4)). These results
enable us to propose different degradation reactions for decomposition of this product in controlled environments.

References
1. Butterworth JH, Morgan ED (1968) J Chem Soc Chem Commun 2324

763
2. Mordue A, Blackwell J (1993) J Insect Physiol 39:903924
3. Ivbijaro M (1990) Insect Sci Appl 11:149152
4. Naqvi SNH, Ahmed SO, Mohammad FA (1991) Pak J Pharm
Sci 4:7176
5. Sundaram KMS (1996) J Environ Sci Health B 31:913948
6. Schiffers BC, Dieye A, Ntema P, Dieye B, Ekukole G (1997)
Med Fac Landbouww Univ Gent 62:225233
7. Govindachari TR, Sandhya G, Ganeshraj SP (1992) Indian J
Chem 31B:295298
8. Shaun J, David ME (1997) J Chromatogr A 761:5363
9. Thejavathi R, Shirish RY, Ravindranath B (1995) J Chromatogr A 705:374379
10. Govindachari TR, Suresh G, Gopalakrishnan G (1995) J Liq
Chromatogr 18:34653471
11. Sundaram KMS, Curry J (1993) J Liq Chromatogr 16:3275
3290
12. Sundaram KMS, Sloane L, Curry J (1993) J Environ Sci Health
B 28:221241
13. Yamasaki BR, Klocke JA, Lee MS, Stone GA, Darlington MV
(1986) J Chromatogr 356:220226

14. Hull CJ Jr, Dutton WR, Switzer BS (1993) J Chromatogr

633:300304
15. Govindachari TR, Sandhya G, Ganeshraj SP (1990) J Chromatogr 513:389391
16. Shaun J, David EM, Wilson ID, Spraul M, Hofmann M (1994)
J Chem Soc Perkin Trans 2 1:14991502
17. Govindachari TR, Sandhya Ganeshraj GSP (1992) J Nat Prod
55:596601
18. Ley SV, Anderson JC, Blaney WM, Jones PS, Lidert Z, Morgan ED, Robinson NG, Santafianos D, Simmonds MS, Toogood PL (1989) Tetrahedron 45:51755192
19. Bokel M, Cramer R, Gutzeit H, Reeb S, Kraus W (1990) Tetrahedron 46:775782
20. Kraus W, Bokel M (1981) Chem Ber 114:267275
21. Schaaf O, Jarvis AP, Andrew van der Esch S, Giagnacovo G,
Oldham NJ (2000) J Chromatogr A 886:8997
22. Laidler KL (1984) J Chem Educ 61:494498
23. Sundaram KMS, Sloane L, Curry J (1995) J Liq Chromatogr
18:363376