Stability studies and determination of carnosic acid and its oxidative degradation products by gas chromatography–mass spectrometry

Article abstract of DOI:10.1016/j.ijms.2016.07.002

The stability of carnosic acid (CA, an important abietane ferruginol-type diterpene, with anti-oxidant, anti-inflammatory, anti-bacterial, and anti-mutagenic properties) exposed to different stress conditions (daylight, high temperatures, different solvents, and humidity) was investigated. Gas chromatography–mass spectrometry (GC–MS) was used to analyse the degraded samples. Structural identification of degradation products was assigned based upon MS fragmentation patterns. The stability experiments were performed on pure CA (96% purity) as well as on commercially-available rosemary extract (containing 50.27% of CA and 5.65% of a diterpene lactone carnosol-C). For the extraction of target compounds three different techniques were compared: ultrasonic extraction (UE), Soxhlet (SE) and microwave extraction (MWE). It was shown that CA always occurs in combination with C, which is produced by oxidation and ring closure from CA, but with UE transformation of CA into C was minimal and constant when compared to SE or MWE. CA dissolved in tetrahydrofuran (THF) remained stable for at least 5?h at room temperature. After a certain time, under all tested conditions, the concentration of CA lowered as a result of the formation of degradation products such as C, rosmanol, 12-O-methyl-CA, 6,7-de-hydro-CA, 7-keto-CA and other unidentified degradants. In comparison to the solid state, the degradation of CA dissolved in organic solvents (THF and ethanol) was significantly faster. In general, protic solvents lead to faster degradation of CA. Only extremely high temperatures (200–300?°C) and long-term exposure to light and humidity (6 months) caused the degradation of CA in powdered samples. The GC–MS method described was also used for determination of CA and C in five different species of the Lamiaceae family (Rosmarinus officinalis L., Salvia officinalis L., Satureja Montana L., Salvia sclarea L., and Salvia glutinosa L.).

Full text of DOI:10.1016/j.ijms.2016.07.002

International Journal of Mass Spectrometry 407 (2016) 29–39
Contents lists available at ScienceDirect
International Journal of Mass Spectrometry
journal homepage: www.elsevier.com/locate/ijms
Stability studies and determination of carnosic acid and its oxidative
degradation products by gas chromatography–mass spectrometry
Masˇa Islamcˇevic´ Razborsˇek∗, Milena Ivanovic´
University of Maribor, Faculty of Chemistry and Chemical Engineering, Smetanova 17, 2000 Maribor, Slovenia
a r t i c l e i n f o
a b s t r a c t
Article history:
The stability of carnosic acid (CA, an important abietane ferruginol-type diterpene, with anti-
oxidant, anti-inflammatory, anti-bacterial, and anti-mutagenic properties) exposed to different stress
conditions (daylight, high temperatures, different solvents, and humidity) was investigated. Gas
chromatography–mass spectrometry (GC–MS) was used to analyse the degraded samples. Structural
identification of degradation products was assigned based upon MS fragmentation patterns. The stability
experiments were performed on pure CA (96% purity) as well as on commercially-available rosemary
extract (containing 50.27% of CA and 5.65% of a diterpene lactone carnosol-C). For the extraction of tar-
get compounds three different techniques were compared: ultrasonic extraction (UE), Soxhlet (SE) and
microwave extraction (MWE). It was shown that CA always occurs in combination with C, which is pro-
duced by oxidation and ring closure from CA, but with UE transformation of CA into C was minimal and
constant when compared to SE or MWE. CA dissolved in tetrahydrofuran (THF) remained stable for at least
5 h at room temperature. After a certain time, under all tested conditions, the concentration of CA lowered
as a result of the formation of degradation products such as C, rosmanol, 12-O-methyl-CA, 6,7-de-hydro-
CA, 7-keto-CA and other unidentified degradants. In comparison to the solid state, the degradation of CA
dissolved in organic solvents (THF and ethanol) was significantly faster. In general, protic solvents lead
to faster degradation of CA. Only extremely high temperatures (200–300 ◦C) and long-term exposure to
light and humidity (6 months) caused the degradation of CA in powdered samples.
Received 22 December 2015
Received in revised form 7 July 2016
Accepted 11 July 2016
Available online 12 July 2016
Keywords:
Carnosic acid
Stability tests
Gas chromatography
Mass spectrometry
The GC–MS method described was also used for determination of CA and C in five different species
of the Lamiaceae family (Rosmarinus officinalis L., Salvia officinalis L., Satureja Montana L., Salvia sclarea L.,
and Salvia glutinosa L.).
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
often occurs in combination with C, and both are commonly found
in species of the Lamiaceae family such as rosemary (Rosmarinus
Carnosic acid (CA) or salvin and carnosol (C) or picros-
alvin belong to the group of abietane ferruginol-type phenolic
diterpenes and are present in many different plants. In plants
the diterpenes are not distributed evenly; they are usually
present within the aerial parts such as leaves, where they play
diverse functional roles [1,2]. It has been confirmed that CA
and C are powerful antioxidants, as their antioxidant activities
are comparable to those of synthetic antioxidant BHT (2,6-
di-tert-butyl-p-hydroxytoluene), BHA (butylated hydroxyanisole),
and TBHQ (2-tert-butyl-hydroquinone) [3,4]. CA and C possess
anti-bacterial, anti-inflammatory, anti-viral, anti-mutagenic, anti-
carcinogenic and many other health-beneficial activities [2]. CA
officinalis L.) or sage (Salvia officinalis L.) [1–8]. The diterpenic con-
tent in plants can differ significantly depending on many factors
such as concentration of CO2 in the air, strong seasonal variations
(exposure to thermal, water and UV-light stress), and fertilisation
and harvest time [2,5,8].
The chemical structure of C was first described by Brieskorn in
1964 [9]. Later it was established that C is a diterpene ı-lactone pro-
duced by oxidation and cyclization from CA and is an intermediate
in the formation of CA’s highly oxygenated derivatives [3,10–12].
C can be converted back to CA by catalytic hydrogenation where
the lactone-ring is opened. The oxidative degradation process of
CA leads to the formation of several different, more or less stable,
degradation products. Most of these, except O-methylated prod-
ucts, indicate antioxidant properties [1,3]. Wenkert has described
oxidative degradation of CA as the “oxidation cascade reaction of
∗
Corresponding author.
E-mail address: masa.islamcevic@um.si (M.I. Razborsˇek).
http://dx.doi.org/10.1016/j.ijms.2016.07.002
1387-3806/© 2016 Elsevier B.V. All rights reserved.

30
M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
CA” [11]. The mechanism of CA oxidative degradation is very similar
to that of its antioxidant activity [12].
2.3. Preparation of plant samples
High-performance liquid chromatography (HPLC) with gra-
dient elution and ultraviolet-visible (UV) detection has been
widely used for the separation and determination of diterpenes
[1,3,4–6,8,10,12–14,16,19]. There are a few cases where in anal-
ysis of diterpenes, the hyphenated technique of LC–MS has been
used [10,14–19]. In the literature, there are no reliable proce-
dures that allow simultaneous separation and determination of CA
and its degradation products using gas chromatography coupled
to mass spectrometry (GC–MS). The reason is in its low thermal,
photo and oxidative stability. Despite their unfavourable physico-
chemical properties, we have been able to develop a reliable
analytical method that allows good separation, rapid, simultane-
ous unambiguous identification and determination of CA, C and
other degradation products in different samples using GC–MS. In
comparison with HPLC, a minor disadvantage is the necessity of a
derivatisation step being required to ensure the thermal stability
and volatility of the compounds. The method described includes
derivatisation with trimethylsilylation (TMS). The spectra of the
silylated investigated compounds are not available in common
mass spectral libraries. This is the first report in which the identify-
ing ions of TMS-derivatives of CA, C, rosmanol, 12-O-methyl-CA,
6,7-de-hydro-CA, and 7-keto-CA are presented and some of the
fragmentation patterns are proposed and discused.
In this experiment, the stabilities of CA and C have been
systematically tested by exposure to different factors including
temperature, daylight, darkness, and moisture. Accelerated and
long-term stability tests were performed and, with respect to
the recorded mass spectra, some of the degradation products
were identified. Additionally, three different extraction techniques
(ultrasonic extraction, Soxhlet extraction, and microwave assisted
extraction) were compared and the extracts were characterised
not only in terms of extraction yields, as described in most articles
[5–7,14,16,20]. but also regarding maximum stability of the CA and
its minimal conversion to C and other degradation products. CA and
C were successfully quantified by GC–MS in five different Lamiaceae
species.
For the extraction of CA and C from plant samples, three different
extraction techniques were compared: ultrasonic extraction (UE),
Soxhlet extraction (SE) and microwave extraction (MWE). For UE,
1 g of an homogenised sample (dry powdered leaves of rosemary)
was weighed into a centrifuge tube and spiked with an appropriate
amount of ISTD-cholesterol. The sample was extracted three times
by sonication with 20 mL of a mixture of THF, EtOH and PYR (v/v/v,
1:1:0.1). After each extraction the sample was centrifuged and the
supernatants were combined, dried over Na2SO4, and concentrated
into dryness by rotary evaporation. The residue was re-dissolved
in 10 mL of the same organic mixture and an aliquot of 200 ␮L was
firstly cleaned by solid phase extraction (SPE) and further by size
exclusion chromatography (SEC).
SPE tubes (Superclean ENVI-Carb 6-mL) were preconditioned
with EtOH (2 × 5 mL), and therafter the sample aliquot was loaded.
The sorbent was never allowed to dry during the conditioning and
sample loading step. The sample was rinsed with 2 mL of EtOH
and the elution was carried out with an additional 30 mL of EtOH.
The eluate was collected under gravity flow and evaporated to dry-
ness. The solvent exchange to THF was performed by adding two
0.5 mL portions of THF and vortexing. For SEC, Bio-Beads S-X3 gel
(200–400 mesh, Bio-Rad Laboratories, Richmond) was allowed to
swell in THF for 24 h before use. The extract was quantitatively
transferred into a glass column (15 mm I.D. and 30 cm long) filled
with slurried beads. After the sample sank into the column bed
it was rinsed with 2 mL of THF and additional 50 mL of THF was
added as the mobile phase. Different fractions (0–10 mL, 10–20 mL,
20–30, 30–50 mL) were collected into glass-stoppered graduated
test tubes. The third fraction (20–30 mL), which contained CA and
C, was concentrated to dryness and re-dissolved in 1 mL of THF. An
aliquot of 50 ␮L was evaporated by a gentle stream of dry nitrogen
and the extract was derivatised by adding 100 ␮L of MSTFA, 50 ␮L
of pyridine and by heating for 1.5 h at 70 ◦C. Prior to analysis, 200 ␮L
of InjSTD was added and the solution was diluted with THF to the
final volume of 1 mL. 1 ␮L was injected into the GC–MS.
For comparison, 1 g of homogenised sample was spiked with
ISTD, 50 mL of an organic mixture was added (THF, EtOH and PYR,
v/v/v, 1:1:0.1) and the samples were submitted to:
2. Materials and methods
- MWE, using MW reactor (Pro Labo MICROWAVE 3.6, France), for
30 min at 30 ◦C or 80 ◦C,
2.1. Chemicals and reagents
- SE, refluxed for 3 h at 80 ◦C.
All reagents and solvents used were of analytical grade.
Methanol (MeOH) and ethanol (EtOH) were purchased from Riedel-
de Haën, tetrahydrofuran (THF), and pyridine (PYR) from Merck
(Germany). Anhydrous sodium sulphate (Na2SO4) was obtained
from J.T. Baker (Netherlands), N-methyl-N-trimethylsilyl trifluo-
roacetamide (MSTFA) from Fluka Chemie (Switzerland), carnosic
acid (CA, 96.4%) and carnosol (C, 98.2%) were from the Alexis cor-
poration (Switzerland), cholesterol (99%) and cholesteryl acetate
(95%) were supplied by Sigma-Aldrich (Germany).
After each extraction the samples were further prepared as
described above by UE.
2.4. Instrumentation and GC–MS conditions
The analyses were performed on two different GC–MS systems:
- HP GC 6890 coupled to an HP MS 5973 (Waldbronn, Germany).
1 ␮L of the sample was injected in the split mode (split ratio 10:1).
Chromatographic separation was performed on a DB-5MS capil-
lary column (J&W Scientific, Folsom, CA, USA; 30 m × 0.25 mm i.d.,
0.25 ␮m thick) where the temperature program was as follows:
initial 105 ◦C (0.8 min), 12 ◦C min−1 to 200 ◦C (0.1 min), 7 ◦C min−1
to 290 ◦C (6 min), 25 ◦C min−1 to 320 ◦C (10 min). Helium was used
as the carrier gas (flow 1.5 mL min−1). MS was operated in the EI+
mode, with electron energy at 70 eV, and MS data were collected
in full scan mode (m/z 50–750 amu).
2.2. Plant samples
Air-dried leaves of Rosmarinus officinalis L. (rosemary), Salvia
officinalis L. (common sage), Satureja Montana L. (winter savory),
Salvia sclarea L. (clary sage) and Salvia glutinosa L. (sticky sage),
were used for the analyses. The dried plant material was ground,
homogenised, and stored in darkness at room temperature.
Commercially-available powdered rosemary extract containing
carnosic acid (50.27%) and carnosol (5.65%) was also used in the
analyses.
- A Polaris Q ion trap mass spectrometer coupled to a Trace GC
ultra gas-chromatograph (Thermo electron corporation, USA)
was used as well. Except for the oven temperature program, all
other GC–MS conditions were the same as mentioned before;

M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
31
Table 1
Storage conditions for photo or thermal stability studies of CA and C.
Pure standards of CA (96.4%) and C (98.2%) and rosemary extract containing CA (50.27%) and C (5.65%) dissolved in EtOH or THF
Stability conditions
Short-term
Storage conditions
Temperature
Time of exposure
Room temperature
Refrigerator
daylight
darkness
25 ◦
4 ◦
C
C
C
2 ◦
C
C
5 h, 24 h, 4 days
5 h, 24 h, 4 days
C
2 ◦
Long-term
Room temperature
Refrigerator
daylight
darkness
25 ◦
4 ◦
2 ◦
C
C
1 month
1 month
C
2 ◦
Accelerated
Thermostatically-controlled oven
darkness
70 ◦
2 h
Rosemary extract containing CA (50.27%) and C (5.65%) in powdered/solid state
Stability conditions
Long-term
Storage conditions
Temperature
Time of exposure
Room temperature
Freezer
Test chamber
Test chamber
daylight
darkness
relative humidity 75%
relative humidity 75%
25 ◦
C
2 ◦
2 ◦
C
C
C
3 months
3 months
3 months
6 months
−16 ◦
C
C
40 ◦
40 ◦
C
C
2 ◦
2 ◦
Accelerated
Thermostatically-controlled oven
darkness
100 ◦C, 200 ◦C, 300 ◦
C
1 min
the oven temperature program was: initial 105 ◦C (0.8 min),
12 ◦C min−1 to 200 ◦C (0.1 min), 7 ◦C min−1 to 290 ◦C (1 min),
15 ◦C min−1 to 310 ◦C (4 min).
GC–MS immediately after preparation. The initial contents of CA
and C were determined from the corresponding calibration curves.
The results were used to control the decreasing contents of both
compounds and for monitoring the progress of the degradation pro-
cess during exposure to the stress conditions. Then several equal
parts from each solution were transferred into graduated glass-
stoppered test tubes and the prepared samples were exposed to
different storage conditions for different periods of time (Table 1).
After exposure to these conditions, an aliquot of 200 ␮L from each
solution was combined with 100 ␮L of ISTD (1000 mg L−1), evapo-
rated, derivatised, spiked with InjSTD, diluted to 1 mL with THF and
measured by GC–MS.
The identification of CA and C from the samples was established
by comparing their retention times and mass spectras with the
derivatised standard compounds, whilst other compounds were
identified by interpretations of their fragmentation patterns, by
comparing their spectral properties with those reported in the lit-
erature [21] or in the Willey and NIST mass spectra libraries.
2.5. Preparation of calibration solutions and calibration curves
The rosemary powdered extract (500 mg in a solid state) was
also exposed to different storage conditions (Table 1). After expo-
sure, the powder was dissolved in THF (10 mg/10 mL), and a 200 ␮L
aliquote was taken for further analysis.
According to the results of the stability tests, THF was used
as a solvent for the preparation of standard solutions. Standard
stock solutions of CA, C, cholesterol (ISTD) and cholesteryl acetate
(InjSTD), were prepared by weighing 10 mg of each into four 10-mL
volumetric flasks, then dissolving in THF (␥ = 1000 mg L−1). Calibra-
tion solutions were prepared in conical glass flasks by combining
different volumes (50–500 ␮L) of CA and C stock solutions with
100 ␮L of ISTD stock solution.
3. Results and discussion
3.1. Plant sample extraction
The solvent was evaporated to dryness and trimethylsilyl (TMS)
derivatives were created by adding 100 ␮L of MSTFA, 50 ␮L of PYR
and by heating for 1.5 h at 70 ◦C. Prior to analysis, 100 ␮L of InjSTD
(␥ = 1000 mg L−1) was added and the solutions were diluted to 1 mL
with THF. 1 ␮L of these solutions were injected into the GC–MS
system in triplicate. In addition, three replicate analyses of the cal-
ibration solutions were performed within 2 weeks. Curves were
constructed by linear regression of the peak-area ratio of CA or C to
In this study, the plant material was extracted using three differ-
ent extraction techniques (UE, SE and MWE) and the results were
compared. The aim was to choose the optimal extraction proce-
dure which would ensure maximum stability of the compounds
and minimal formation of degradation products. CA appears to
be a thermo-sensitive compound. The concentrations of CA and C
differed significantly in the final extracts depending on the extrac-
tion method used. Th−e1concentration of CA was approximately six
times higher (33 mg g ) in the samples extracted by UE than when
extracted by SE (5.7 mg g−1) or MWE (6.8 mg g−1) (Fig. 1). Using UE
the concentration of C was minimal (1.8 mg g−1) and the area ratios
between CA/C, CA/ISTD, and C/ISTD were constant. SE and MWE
resulted in an increase in the C concentration because of CA degra-
dation. The higher temperatures used in the SE (approx. 80 ◦C) and
long-term extraction (3 h) promoted CA degradation and its con-
version to C and some other degradants, which are explained in
Section 3.2. Similar results were obtained in the extractions using
microwaves. Based on the results, for further investigation the UE
at room temperature was adopted as the optimal one.
the ISTD (y), versus the concentrations (x) in mg L−1
.
2.6. Stability tests
The stability tests were performed on pure CA (96% purity),
as well as on commercially-available rosemary extract (contain-
ing 50.27% of CA and 5.65% of C). Because of their more affordable
prices, most of the stability tests were performed on commercially-
available rosemary extracts, and a small number were carried out
on pure standards. The stabilities of the compounds were studied
in the solid state as well as after dissolving in THF or EtOH.
In order to investigate the stabilities of CA and C in organic
solvents, pure compounds as well as commercially-available
rosemary extract were dissolved in EtOH and THF, separately
(␥ = 1000 mg L−1). A 200 ␮L aliquot from each solution was com-
bined with 100 ␮L of ISTD (1000 mg L−1), evaporated, derivatised,
spiked with InjSTD, diluted to 1 mL with THF and measured by
The mixture of organic solvents THF-EtOH-PYR proved to be an
effective solvent for the extraction of CA from different aromatic
plants. Cleaning and separation methods using a combination of
SPE and SEC were suitable for removing interferences and for iso-
lating the investigated compounds. After SEC, different fractions
were collected and derivatised; the major compounds of the third

32
M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
Fig. 1. Rosemary leaf extracts obtained after (a) SE, for 3 h at 80 ◦C, (b) MWE for 30 min at 80 ◦C, (c) MWE for 30 min at 30 ◦C, and (d) UE at room temperature.
fraction (20–30 mL) were CA and C. Pentacyclic triterpenoids, such
as betuline, betulinic acid, oleanolic acid, and ursolic acid were also
eluted in this fraction. These were identified by comparison with
standard compounds [22], but they were not quantitatively evalu-
ated (Fig. 2a). The CA and C contents from commercially-available
rosemary extract were determined from the corresponding

M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
33
Fig. 2. Chromatograms of silylated compounds present in Rosmarinus officinalis L. leaf extract obtained after SPE and SEC by collecting different fractions, (a) fraction 20–30 mL;
(b) a typical chromatogram of a commercially available rosemary extract with declared contents of CA and C (50.27% and 5.65%).

34
M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
Fig. 3. Chromatograms of silylated compounds investigated present in a solution of CA in THF: (a) freshly prepared (b) after being stored in the refrigerator at 4 ◦C for 4 days
(c) after being stored in the refrigerator at 4 ◦C for 1 month.

M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
35
Fig. 4. Typical EI mass spectra of (a) CA-3TMS derivative (b) C-2TMS derivative.
calibration curves, and compared with the declared values specified
by the manufacturer. After ten independent analyses it was con-
firmed that the extract contained 50.3 1.27% of CA and 5.5 0.42%
of C (Fig. 2b).
DMSO–acetonitrile, or ethyl acetate–acetonitrile). Interestingly,
both compounds present in rosmemary extract were stable up to
21 h when the extract was dissolved in DMSO [13]. Cuvelier showed
that pure CA dissolved in MeOH and exposed to daylight and air-
oxygen completely decomposed in 15 days and the more stable
compound C was formed [10]. Our results confirmed that pure CA
(96.4%) was stable for at least 5 h (at room temperature) when it
was dissolved in the aprotic solvent THF. During this time the con-
centrations of CA and C were constant and the peak areas of CA, C,
and ISTD were constant. In the freshly prepared CA (96.4%) in THF,
besides CA, the minor contents of 12-O-methyl-CA and rosmanol
were also identified (Fig. 3a). Prolonged storage time resulted in CA
degradation. After being dissolved in THF and stored in the refriger-
ator for 4 days the CA partially degraded (by app. 10%) and at least
three new compounds were formed, of which only 6,7-dehydro-
CA was identified (Fig. 3b). Longer storage time (1 month in the
refrigerator) resulted in further degradation of CA (by 90%) which
3.2. Stabilities of CA and C in organic solvents
It is well known that diterpenic compounds are very unsta-
ble and sensitive to many different factors and that they can
be rapidly decomposed or transformed into other compounds
[2,3,5,8–14,18–20]. The stabilities of pure CA and C dissolved in
organic solvents have been described previously, and it has been
confirmed that CA is more stable in non-polar organic solvents
such as DMSO and less so in polar ones such as MeOH [12,13].
Thorsen and Hildebrandt reported that pure CA dissolved in DMSO
is stable for several days, whilst C showed significant degrada-
tion within a few hours in all solvents tested (MeOH, DMSO,

36
M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
produced at least ten new compounds where 6,7-dehydro-CA and
7-keto-CA could be identified (Fig. 3c). Degradation was faster at
room temperature and daylight than in a refrigerator. Degradation
took place to a greater extent in a protic solvent (EtOH) than in an
aprotic one (THF). Degradation products were the same. Identifica-
tion of degradation products is very difficult as the pure standards
are not commercially available, mass spectra could not be found
in the literature, and only few of them are present in the spectral
libraries. Therefore, in the future, additional efforts will be needed
in order to identify all of the degradation products of CA.
CA was observed. After 4 days in a refrigerator, C slightly degraded
(by 8%) into two new compounds, which were not identified. After
one month in a refrigerator, large-scale degradation of C occured
(by 87%) and led to the formation of at least five new compounds,
which could not be identified. Degradation was faster at room tem-
perature and daylight than in a refrigerator. Degradation was faster
and greater in EtOH than in THF. From the results we concluded that
under the same conditions, C degraded slightly more slowly than
CA and formed a significantly lower number of new compounds.
The stability of CA (50.27%) was studied in the rosemary
extract dissolved in THF and EtOH exposed to different conditions
(described in Section 2.6). It was found that the CA quickly (in less
Furthermore, the stability of pure C (98.2%) was studied. In a
freshly prepared solution of C in THF, the presence of 6,7-de-hydro
Table 2
Structural formulas and names of identified TMS-diterpenes, and the molecular weights of their TMS derivatives with characteristic fragmentation ions and characteristic
fragmentation ions of all unidentified TMS-degradation compounds.
Identified diterpenic compounds
Molecular weight of TMS derivative (g/mol)
Characteristic fragmentation ions m/z (relative intensity%)
OH
CH3
HO
COOH
CH3
H
HC3C arCnHo3 sic acid (CA)
548
335 (100), 431 (90), 548 (15)
Carnosic acid (CA)
OH
CH3
HO
O
CH3
O
H3C
CH3
Carnosol (C)
474
348 (80), 430 (100), 474 (20)
Carnosol (C)
H3C
O
CH3
HO
CH3
COOH
H3C
CH3
12-O-methyl-carnosic acid
490
277 (60), 373 (100), 490 (30)
12-O-methyl-carnosic acid
OH
CH3
HO
CH3
O
OH
H3C
CH3
6,7-dehydro-carnosic acid
546
387 (60), 429 (100), 546 (30)
6,7-dehydro-carnosic acid
OH
CH3
HO
CH3
O
OH
O
H3C
CH3
7-keto-carnosic acid
562
359 (99), 444 (100), 562 (30)

M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
37
Table 2
(Continued)
Identified diterpenic compounds
7-keto-carnosic acid
Molecular weight of TMS derivative (g/mol)
Characteristic fragmentation ions m/z (relative intensity%)
OH
CH3
HO
O
CH3
O
C6
OH
H3C
CH3
Rosmanol (epi-, iso-)
Rosmanol (epi-, iso-)
562
433 (90), 518 (100), 562 (40)
Unidentified compounds
Characteristic fragmentation ions m/z
»unknown 1«
»unknown 2«
»unknown 3«
»unknown 4«
»unknown 5«
»unknown 6«
»unknown 7«
»unknown 8«
341, 387, 415, 430
284, 372, 428
341, 357, 369, 474
284, 372, 428, 460, 555
301, 371, 489, 504
301, 371, 489, 504
474, 517, 634
429, 447, 555, 673
than 5 h) transformed into C and other degradants under all con-
ditions tested. Again, degradation took place to a greater extent in
EtOH than in THF. Degradation products were the same as those
formed after the one-month aging of pure CA (96.4%) in THF. 12-
O-Methyl-CA, 6,7-dehydro-C, 7-keto-CA, C, and rosmanol could be
identified. After accelerated heating at 70 ◦C for 2 h, the content of
CA in EtOH decreased by approximately 50%, whilst in THF degra-
dation was even more rigorous, as CA was completely degraded
and quantitatively converted to C.
From the results we can conclude that temperature is the major
factor affecting the degradation of CA and C in solutions, and light
exposure further accelerates the degradation of both compounds.
In general, more protic solvents lead to faster degradation.
and intense fragment ion can be observed at m/z 431 (elimination
of TMSCOOH group) and is characteristic of diterpene compounds
of ferruginol type [21]. The characteristic ion at m/z 335 is formed
by the cleavage of 3 bonds (between C4-C5, C5-C10, and C9-C10
atoms) within the CA molecule.
EI+ mass spectra of a C-2TMS derivative is shown in Fig. 4b. From
the mass spectra in Fig. 4b the molecular ion of C-2TMS derivative at
m/z 474 can be clearly seen. The main and most intense ion occurs
at m/z 430, as the result of COO-group elimination and is specific
for diterpenic lactones [21]. The secondary fragment at m/z 348 is
formed from the ion at m/z 430 as a result of the cleavage of three
C
C bonds between C1 and C10, C10 and C5 and C5 and C6 atoms,
and a transfer of the CH3 group within the molecule.
Structural formulas and the names of all identified TMS-
diterpenes, including the molecular weights of their TMS
derivatives, with characteristic fragmentation ions, are presented
in Table 2. MS data of all unidentified compounds are also reported
in Table 2.
3.3. Stabilities of CA and C in the solid state
In the next step, the stability of rosemary extract (containing
50.27% of CA) in the solid state was tested. It was confirmed that
powder extract was stable for up to 3 months under all conditions
tested and during this time its composition did not change. After 6
months of exposure to +40 ◦C and relative humidity of 75%, signifi-
cant changes occurred: the content of CA decreased by 28%, whilst
the C content increased by 97%. A change in the colour and odour of
the extract was also observed. Powder was also heated for 1 min at
100, 200 and 300 ◦C separately, and analysed. The purpose of this
study was to examine the possibility of controlled thermal trans-
formation of CA to C and other compound(s). No changes appeared
after heating to 100 ◦C; as usual, besides CA, minor traces of 12-O-
methyl-CA were also present in the extract. When heated to 200 ◦C,
degradation of CA appeared and at least three new unidentified
degradation products were formed. At 300 ◦C, CA was completely
degraded and at least four new unidentified degradation products
were formed.
3.5. Plant extracts analysis
CA and C were successfully quantified by GC–MS in five dif-
ferent Lamiaceae species. Fig. 5 presents the GC chromatogram of
the silylated extract of Salvia officinalis L. From CA and C, different
biologically-active components in plants, such as rosmarinic acid,
caffeic acid, betuline, betulinic acid, oleanolic acid, ursolic acid, and
different monosaccharides were separated and identified, but they
were not quantified. The average contents of CA and C were deter-
mined in mg g−1 per dry weight (Table 3). The content of CA ranged
from traces (<0.1 mg g−1) up to 33 mg g−1. The lowest concentration
of CA was determined in Salvia glutinosa and the highest in Ros-
marinus officinalis, whilst in other plants it was undetected. C was
detected only in rosemary and sage in concentrations from 0.80 to
1.78 mg g−1
.
3.4. Confirmation of CA and C
4. Conclusions
EI+ mass spectra of a CA-3TMS derivative is shown in Fig. 4a.
In all chromatograms the diterpenic compounds investigated were
well distinguished by order of elution during gas chromatography
(CA is eluted before C) and by intensities of the fragment ion sig-
nals in their mass spectra (Table 2). The molecular ion of CA-3TMS
derivative is easily visible at m/z 548 (Fig. 4a). The most important
The GC–MS method described is sensitive and suitable for simul-
taneous quantification of structurally-related phenolic diterpenes
such as carnosic acid (CA) and carnosol (C) and their degradation
products in different plant samples. Before GC–MS analysis, the
method requires derivatisation of the compounds. In this way the

38
M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
Fig. 5. Chromatogram of silylated compounds present in Salvia officinalis L. extract.
Table 3
Contents of carnosic acid (CA) and carnosol (C) in different plant extracts (mg g−1 per dry weight); each content value is the mean of two replicated analyses.
Plant species
Compound
Rosmarinus officinalis
Salvia officinalis
Salvia sclarea
Salvia glutinosa
Satureja montana
*
*
*
*
Carnosic acid (CA)
Carnosol (C)
33.02
1.78
3.51
0.80
<0.1
*
*
Not identified.
compounds are converted to more stable and more volatile TMS
derivatives. Although UV–VIS provides useful information for the
identification of antioxidant phytochemicals such as diterpenes,
the use of conventional approaches based on spectra is often lim-
ited when samples contain similar compounds whose absorption
spectra are almost identical. Unambiguous identification can be
done using other detection techniques like mass spectrometry (MS)
which provides additional structural information for easier identi-
fication and quantification of the compounds.
Extraction of diterpenic compounds from plant samples using
ultrasound at room temperature is more suitable than Soxhlet
extraction or extraction using microwaves. With ultrasonic extrac-
tion, conversion of CA into C is minimal and constant. The stability
of CA and C were studied in the extracts as well as in the pure
standards. From the results it is clear that the degradation pro-
cess of CA is solvent, temperature, light and time-dependent. At all
tested conditions, the concentration of CA was lowered as a result
of the formation of C, rosmanol, 12-O-methyl-carnosic acid, 6,7-de-
hydro-CA, 7-keto-CA and other unidentified degradation products.
Finally, it has to be emphasised that the maximum stability of
CA can be achieved only by careful laboratory work which includes
the selection of a suitable extraction method, solvent, using an inert
atmosphere, and avoiding humidity and higher temperatures.
[4] M.E. Cuvelier, C. Berset, H. Richard, Antioxidant constituents in sage (Salvia
officinalis), J. Agric. Food Chem. 42 (1994) 665–669.
[5] K. Schwarz, W. Terens, E. Schmauderer, Antioxidative constituents of
Rosmarinus officinalis and Salvia officinalis. III. Stability of phenolic
diterpenes of rosemary extracts under thermal stress as required for
technological processes, Z. Lebensm. Unters. Forsch. 195 (1992) 104–107.
[6] M. Jacotet-Navarro, N. Rombaut, A.-S. Fabiano-Tixier, M. Danguien, A. Bily, F.
Chemat, Ultrasound versus microwave as green processes for extraction of
rosmarinic, carnosic and ursolic acids from rosemary, Ultrason. Sonochem. 27
(2015) 102–109.
[7] M.B. Jemia, R. Tundis, A. Maggio, S. Rosselli, F. Senatore, F. Menichini, M.
Bruno, M.E. Kchouk, M.R. Loizzo, NMR-based quantification of rosmarinic and
carnosic acids, GC–MS profile and bioactivity relevant to neurodegenerative
disorders of Rosmarinus officinalis L. extracts, J. Funct. Foods 5 (2013)
1873–1882.
[8] S. Munné-Bosch, L. Alegre, K. Schwarz, The formation of phenolic diterpenes
in Rosmarinus officinalis L. under Mediterranean climate, Eur. Food Res.
Technol. 210 (2000) 263–267.
[9] C.H. Brieskorn, A. Fuchs, Uber die inhaltsstoffe von Salvia off L. XIII. Die
Struktur des Pikrosalvins, eines diterpen-o-diphenol-lactons aus dem
Salbeiblatt, Chem. Ber. 95 (1962) 3034–3041.
[10] M.E. Cuvelier, H. Richard, C. Berset, Characterization of two new carnosic acid
derivatives by spectrophotometry, mass spectrometry and electrochemical
detection, 2nd International Electronic Conference on Synthetic Organic
Chemistry (ECSOC-2) (1998) 1–30.
[11] E. Wenkert, A. Fuchs, J.D. McChesney, Chemical artifacts from the family
Labiatae, J. Org. Chem. 30 (1965) 2931–2934.
[12] T. Masuda, Y. Inaba, Y. Takeda, Antioxidant mechanism of carnosic acid:
structural identification of two oxidation products, J. Agric. Food Chem. 49
(2001) 5560–5565.
[13] M.A. Thorsen, K.S. Hildebrandt, Quantitative determination of phenolic
diterpenes in rosemary extracts. Aspects of accurate quantification, J.
Chromatogr. A 995 (2003) 119–125.
References
[14] E. Ibán˜ez, A. Kubátová, F.J. Sen˜oráns, S. Cavero, G. Reglero, S.B. Hawthorne,
Subcritical water extraction of antioxidant compounds from rosemary plants,
J. Agric. Food Chem. 51 (2003) 375–382.
[15] K. Bentayeb, C. Rubio, R. Batlle, C. Nerin, Direct determination of carnosic acid
in a new active packaging based on natural extract of rosemary, Anal. Bioanal.
Chem. 389 (2007) 1989–1996.
[1] S. Munne-Bosch, L. Alegre, Subcellular compartmentation of the diterpene
carnosic acid and its derivatives in the leaves of rosemary, Plant Physiol. 125
(2001) 1094–1102.
[2] S. Birtic´, P. Dussort, F.X. Pierre, A.C. Bily, M. Roller, Carnosic acid,
Pytochemistry 115 (2015) 9–19.
[3] S.L. Richheimer, M.W. Bernart, G.A. King, M.C. Kent, D.T. Bailey, Antioxidant
activity of lipid-soluble phenolic diterpens from rosemary, JAOCS 73 (1996)
507–514.

M.I. Razborsˇek, M. Ivanovi´c / International Journal of Mass Spectrometry 407 (2016) 29–39
39
[16] G. Vicente, M.R. García-Risco, T. Fornari, G. Reglero, Isolation of carsonic acid
[19] Y. Zhang, J.P. Smuts, E. Dodbiba, R. Rangarajan, J.C. Lang, D.W. Armstrong,
Degradation study of carnosic acid carnosol, and rosemary extract
(Rosmarinus officinalis L.), J. Agric. Food Chem. 60 (2012) 9305–9314.
from rosemary extracts using semi-preparative supercritical fluid
chromatography, J. Chromatogr. A 1286 (2013) j208–215.
[17] M. Herrero, M. Plaza, A. Cifuentes, E. Ibán˜ez, Green processes for the
extraction of bioactives from rosemary: chemical and functional
characterization via ultra-performance liquid chromatography–tandem mass
spectrometry and in-vitro assays, J. Chromatogr. A 1217 (2010) 2512–2520.
[18] Y. Song, H. Yan, J. Chen, Y. Wang, Y. Jiang, P. Tu, Characterization of in vitro
and in vivo metabolites of carnosic acid a natural antioxidant, by high
performance liquid chromatography coupled with tandem mass
spectrometry, J. Pharm. Biomed. Anal. 89 (2014) 183–196.
[20] S. Bas¸ kan, N. Öztekin, F.B. Erim, Determination of carnosic acid and rosmarinic
acid in sage by capillary electrophoresis, Food Chem. 101 (2007) 1748–1752.
[21] C.R. Enzell, I. Wahlberg, Terpenes and terpenoids, in: G.R. Waller, O.C. Dermer
(Eds.), Biochemical Applications of Mass Spectrometry, First Supplementary
Volume, Wiley-Interscience, New York, 1980, pp. 311–406.
[22] M.I. Razborsˇek, D. Brodnjak-Voncˇina, V. Dolecˇek, E. Voncˇina, Determination of
oleanolic, betulinic and ursolic acid in lamiaceae and mass spectral
fragmentation of their trimethylsilylated derivatives, Chromatographia 67
(2008) 433–440.