3590 J. Agric. Food Chem.

1998, 46, 3590−3595

Characterization of the Action of Selected Essential Oil Components

on Gram-Negative Bacteria
Ilkka M. Helander,*,† Hanna-Leena Alakomi,† Kyösti Latva-Kala,† Tiina Mattila-Sandholm,†
Irene Pol,‡ Eddy J. Smid,‡ Leon G. M. Gorris,‡,§ and Atte von Wright†
VTT Biotechnology and Food Research, P.O. Box 1501, FIN-02044 VTT, Finland, and Agrotechnological
Research Institute (ATO-DLO), Bornsesteeg 59, NL-6708 PD Wageningen, The Netherlands

Carvacrol, (+)-carvone, thymol, and trans-cinnamaldehyde were tested for their inhibitory activity
against Escherichia coli O157:H7 and Salmonella typhimurium. In addition, their toxicity to
Photobacterium leiognathi was determined, utilizing a bioluminescence assay. Their effects on the
cell surface were investigated by measuring the uptake of 1-N-phenylnaphthylamine (NPN), by
measuring their sensitization of bacterial suspensions toward detergents and lysozyme, and by
analyzing material released from cells upon treatment by these agents. Carvacrol, thymol, and
trans-cinnamaldehyde inhibited E. coli and S. typhimurium at 1-3 mM, whereas (+)-carvone was
Downloaded from pubs.acs.org by KAROLINSKA INST on 01/23/19. For personal use only.

less inhibitory. trans-Cinnamaldehyde was the most inhibitory component toward P. leiognathi.
Carvacrol and thymol disintegrated the outer membrane and released outer membrane-associated
material from the cells to the external medium; such release by (+)-carvone or trans-cinnamaldehyde
was negligible. Of the tested components, carvacrol and thymol decreased the intracellular ATP
pool of E. coli and also inreased extracellular ATP, indicating disruptive action on the cytoplasmic
membrane.
J. Agric. Food Chem. 1998.46:3590-3595.

Keywords: Essential oils; carvacrol; carvone; cinnamaldehyde; thymol; antimicrobial activity; outer
membrane; lipopolysaccharide

INTRODUCTION are able to pass the OM through abundant porin

proteins providing hydrophilic transmembrane chan-
Plant-derived essential oils have long served as nels, whereas the OM serves as a penetration barrier
flavoring agents in foods and beverages, and due to their toward macromolecules and to hydrophobic compounds,
versatile content of antimicrobial compounds, they and it is for this reason that Gram-negative bacteria
possess potential as natural agents for food preservation are relatively resistant to hydrophobic antibiotics and
(Conner, 1993). The antimicrobial activity of essential toxic drugs (Nikaido and Vaara, 1985; Nikaido, 1996).
oils is assigned to a number of small terpenoid and The OM is, however, not totally impermeable to hydro-
phenolic compounds, which also in pure form have been phobic molecules, some of which can slowly traverse
shown to exhibit antibacterial or antifungal activity through porins (Plésiat and Nikaido, 1992; Nikaido,
(Karapinar and Aktug, 1987; Didry et al., 1993; Conner, 1996). In general, bypassing the OM is a prerequisite
1993; Juven et al., 1994; Oosterhaven et al., 1996; Smid for any solute to exert bactericidal activity toward
et al., 1996). Specific mechanisms of bactericidal action Gram-negative bacteria. It should therefore be worth-
of essential oil constituents such as thymol and carvac- while to evaluate antimicrobial compounds present in
rol remain, however, poorly characterized. The anti- plants for their activity on Gram-negative bacteria in a
bacterial properties of these compounds are evidently number of real food systems. There are indications that
associated with their lipophilic character, leading to also microorganisms, for example, certain lactic acid
accumulation in membranes and to subsequent mem- bacteria, produce small molecular mass antibacterial
brane-associated events such as energy depletion (Con- and antifungal compounds (Helander et al., 1997b) that
ner, 1993; Sikkema et al., 1995). should be very useful in food preservation. It was
Essential oils and their components are known to be therefore of interest to investigate the mode of action
active against a wide variety of microorganisms, includ- of selected essential oil components on Gram-negative
ing Gram-negative bacteria (Deans and Ritchie, 1987; bacteria, focusing on the effects on the OM.
Conner, 1993; Sivropoulou et al., 1996). All Gram-
negative bacteria possess an outer membrane (OM), MATERIALS AND METHODS
which provides the bacterium with a hydrophilic sur-
face, due to the presence of lipopolysaccharide (LPS) Chemicals. Carvacrol, (+)-carvone, trans-cinnamaldehyde,
molecules (Nikaido, 1996). Small hydrophilic solutes and thymol (referred to below as essential oil components)
were obtained from Fluka. Stock solutions (1 mM) of these
were prepared in aqueous 85% ethanol. The maximum
* Author to whom correspondence should be addressed (fax solubility of these compounds in water at 20 °C was deter-
++358 9 455 2028; e-mail Ilkka.Helander_vtt.fi). mined as 10 mM for carvacrol and (+)-carvone, 13 mM for
† VTT.
trans-cinnamaldehyde, and 7 mM for thymol. Chicken egg
‡ ATO-DLO.
white lysozyme (EC 3.2.1.17), n-heptadecanoic acid methyl
§ Present address: Unilever Research Laboratorium, Vlaard-
ester, 1-N-phenylnaphthylamine (NPN), and SDS were pur-
ingen, The Netherlands. chased from Sigma-Aldrich, and Triton X-100 was from BDH.

S0021-8561(98)00154-X CCC: $15.00 © 1998 American Chemical Society

Published on Web 08/11/1998
Action of Essential Oil Components on Gram-Negative Bacteria J. Agric. Food Chem., Vol. 46, No. 9, 1998 3591

Proteinase K (EC 3.4.21.64) was from Merck and EDTA from indicate lytic action. Each determination was performed in
Riedel-de-Haen AG. Silicon oils AR200 and AR20 were from quadruplicate. At least three independent lysis tests were
Wacker Chemicals. done.
Bacteria. Escherichia coli ATCC 35150 (O157:H7) and Lipopolysaccharide and Protein Release. The possible
Salmonella typhimurium ATCC 13311 were grown in LB broth release of OM components [see Vaara (1992)] was investigated
(per liter, 10 g of Difco tryptone, 5 g of Difco yeast extract, by analyses of cell-free supernatants after treatment with the
and 5 g of NaCl, pH 7.0) at 37 °C with shaking (200 essential oil components. The bacteria were grown to OD630
revolutions/min). Growth was monitored spectrophotometri- of 0.5, and the cultures were divided in 1-mL portions into
cally at 630 nm. Photobacterium leiognathi ATCC 33469 was microcentrifuge tubes. After centrifugation at room temper-
cultivated at 25 °C with shaking in a medium containing, per ature (12000 revolutions/min, 2 min, Eppendorf microfuge),
liter, 30 g of NaCl, 3.9 g of K2HPO4, 2.1 g of KH2PO4, 5 g of the cells were suspended in 10 mM Tris-HCl buffer, pH 7.2,
NH4Cl, 1 g of MgSO4 heptahydrate, 0.75 g of KCl, 1 g of CaCO3, and centrifuged again as above. The deposited cells were again
5 g of Difco yeast extract, 5 g of Difco tryptone, 3 mL of glycerol, suspended into the Tris buffer, to which the test substance
and 50 mL of 1 M Tris-HCl, pH 7.5. (essential oil components or EDTA) was added, and the
Growth Inhibition Tests. Bacteria were grown to OD630 suspensions were kept at 37 °C for 10 min. The control
of 0.2 and subsequently diluted at 1/100 into LB broth (2 mL) samples were suspended in Tris buffer without additions. After
supplemented with 0.2, 0.6, 2, 6, and 10 µmol of the particular centrifugation, 0.9 mL of the cell-free supernatant was re-
essential oil component (corresponding to end concentrations moved and freeze-dried. For each test substance, two parallel
of 0.1, 0.3, 1, 3, and 10 mM, respectively) and then cultivated vials were prepared so that the total volume of recovered
for 20 h. The OD630 values of the cultures were measured, supernatant was 1.8 mL. The freeze-dried supernatants were
and the lowest concentration that completely inhibited bacte- dissolved in 100 µL of SDS-PAGE sample buffer (Novex) and
rial growth was taken as the minimal inhibitory concentration heated at 100 °C for 10 min. Each sample was then divided
(MIC). in two equal portions, to one of which 10 µL of proteinase K
solution (2.5 mg/mL) was added. Subsequently, the vials were
Photobacter Toxicity Test. This method was a modifica- kept at 60 °C for 1 h. The samples were analyzed by SDS-
tion of the DIN 38412 method originally designed for ecotoxi- PAGE in Novex precast 12% acrylamide gels; 10 µL of lysate
cological examination of water samples, with P. leiognathi as was applied to the gel. The gels with proteinase K-treated
the test organism. From a culture grown for 24 h was made samples were stained with silver (Silver Xpress staining kit,
a 1% (v/v) dilution into 2% (w/v) NaCl. After 15 min at 25 °C, Novex), as were the gels of the untreated samples, which were
the bioluminescence was measured (BioOrbit 1253 Luminom- alternatively stained for protein by the Novex Colloidal Coo-
eter, BioOrbit, Turku, Finland), and the test components massie stain.
appropriately diluted in 2% (w/v) NaCl were added in amounts Lipid (including phospholipid and LPS) released from the
ranging from 0.0125 to 0.5 µmol. The bioluminescence was bacterial cells by the above treatments was assayed on the
again measured after 15 min of incubation. The values basis of fatty acids found in the cell-free supernatants after
obtained were corrected for the decrease in bioluminescence treatment of cultures with essential oil components. The
observed in untreated control tubes, and the results are bacteria were grown to OD630 of 0.6, washed with 10 mM Tris-
expressed as inhibition percentages calculated from duplicate HCl, pH 7.2 (centrifugation at 3000g, 10 min, 25 °C), and
measurements of bioluminescence. resuspended in the same buffer, which was supplemented with
Uptake of NPN. The uptake of the hydrophobic fluores- essential oil components (final concentration ) 2 mM) or EDTA
cent probe NPN by bacterial cells in buffer suspension was (1 mM). Control samples were suspended in buffer only. After
determined fluorometrically as described recently (Helander the suspensions (10 mL) had been incubated at 37 °C for 10
et al., 1997a). Briefly, cultures were grown to OD630 of 0.5 min, they were centrifuged at 11000 rpm (1 min, Eppendorf
and then washed with and suspended in 5 mM HEPES buffer, microfuge, 25 °C), whereupon clear supernatants (total volume
pH 7.2. Samples of this suspension supplemented with 10 µM ) 9.1 mL) were carefully removed and freeze-dried. After
NPN were monitored with a Shimadzu RF-5000 spectropho- addition of internal fatty acid standard (n-heptadecanoic acid
tometer at 420 nm (excitation wavelength, 350 nm; slit widths, methyl ester, 105 µg), the samples were processed by saponi-
5 nm), and fluorescence levels (in arbitrary units) were fication and methylation as described by Moore et al. (1994).
recorded before and after addition of desired concentrations It is recognized that this procedure leads to an underestima-
of test substances. The stable fluorescence level, typically tion of 3-hydroxytetradecanoic acid, which is partially amide-
reached after 30 s from addition of the test substance, was linked in the lipid A component of the LPS (Zähringer et al.,
taken as the result value. In testing the effect of MgCl2 to 1994) and therefore converted to the corresponding methyl
the carvacrol- or thymol-induced fluoresecence, this salt was ester only by a ∼65% yield (Helander and Haikara, 1995). The
added to the buffer before NPN. Increased fluorescence resulting fatty acid methyl esters were analyzed by capillary
indicates uptake of NPN by the bacterial membrane(s), as the gas chromatography (GC), the instrumentation and conditions
quantum yield of NPN is greatly enhanced in lipid versus of which were described recently (Helander et al., 1997a).
aqueous environment (Loh et al., 1984). Determination of Intra- and Extracellular ATP. To
Bacteriolysis. The effect of essential oil components on study the possible effect of essential oils on the permeability
detergent- or lysozyme-induced bacteriolysis was assayed on of the cytoplasmic membrane for small solutes, internal and
microtiter plates as described recently (Helander et al., 1997a). external ATP pools of E. coli were determined in the presence
The bacteria were first cultivated and washed as above and and absence of carvacrol, thymol, (+)-carvone, and trans-
then suspended in a similar volume of 10 mM HEPES cinnamaldehyde. Cells of E. coli ATCC 35150 were harvested
containing 50 mM NaCl, pH 7.2. The suspension was divided at the exponential growth phase (OD660 of 0.8), washed twice
into two portions, and to one was added the essential oil in 5 mM HEPES buffer, pH 7.2, and concentrated (∼50-fold).
component at the indicated concentration; the other served as Washed cells were subsequently diluted in HEPES buffer to
control. The suspensions were incubated at room temperature a density of ∼0.2 mg of protein/mL at room temperature.
for 10 min, centrifuged, and resuspended in the buffer without Glucose was added at time zero to a final concentration of 0.5%
essential oil components. From such suspensions, aliquots (108 (w/v). After 5 min of incubation, essential oil components
cells in 100 µL) were pipetted into microtiter wells, which (carvone, carvacrol, thymol, or trans-cinnamaldehyde) were
already contained the indicated concentrations of lysozyme, added at a final concentration of 2 mM. In control experi-
Triton X-100, or SDS. Cell lysis was monitored spectropho- ments, no essential oil component was added. The intracel-
tometrically at 405 nm by a Multiskan MCC/340 spectropho- lular and extracellular ATP concentrations were determined
tometer, Labsystems; results are expressed in percentages at regular time intervals by separating the cells from the
based on absorbances of component-pretreated versus un- external medium by silicon oil centrifugation (Ten Brink et
treated cell suspensions at 4 min after addition of detergent al., 1985). Briefly, samples (200 µL) from a cell suspension
or lysozyme. Thus, 100% equals no lysis; lower percentages were transferred to microcentrifugation tubes containing 200
3592 J. Agric. Food Chem., Vol. 46, No. 9, 1998 Helander et al.

Table 1. Minimal Inhibitory Concentrations (MIC) of

Essential Oil Components for E. coli O157:H7 and S.
typhimurium
MIC (mM)
component E. coli S. typhimurium
carvacrol 3 1
(+)-carvone 10 10
trans-cinnamaldehyde 3 3
thymol 3 1

µL of a 2:1 mixture of silicon oil AR200 (F ) 1.05 g/mL) and Figure 1. Inhibition of bioluminescence of P. leiognathi by
silicon oil AR20 (F ) 0.96 g/mL) on top of a layer of 100 µL of essential oil components.
10% (w/v) trichloroacetic acid with 2 mM EDTA. The cells
were spun through the silicon oil (5 min, 12000g), and samples Table 2. Values for NPN Uptake Induced by Essential
(5 µL) of both aqueous layers were taken to determine the ATP Oil Components in Suspensions of E. coli O157:H7 and S.
content using the firefly luciferase assay (Lundin and Thore, typhimurium
1975). Luminescence was recorded using a BioOrbit 1250 fluoresecence increasea
luminometer (BioOrbit, Turku, Finland). (rel fluorescence units ( SD)
concn MgCl2
component (mM) (mM) E. coli S. typhimurium
RESULTS thymol 0.5 - 22 ( 3.5 24 ( 11
1 - 44 ( 2.5 36 ( 11
Effect of Essential Oil Components on Bacterial 2 - 90 ( 5.3 44 ( 18
Growth. Each of the four essential oil components 2 2 91 ( 3.5 57 ( 13
inhibited the growth of E. coli and S. typhimurium. The carvacrol 0.5 - 15 ( 2.1 23 ( 17
minimal inhibitory concentrations listed in Table 1 1 - 38 ( 2.1 30 ( 24
2 - 83 ( 4 45 ( 26
reveal that E. coli was equally inhibited by thymol, 2 2 88 ( 3.2 42 ( 9
carvacrol, and trans-cinnamaldehyde, whereas (+)- (+)-carvone 2 - 3(3 3(3
carvone was considerably less inhibitory. A similar 10 - 0 0
pattern was obtained for S. typhimurium, which was trans-cinnamaldehyde 2 - 0 0
slightly more sensitive than E. coli toward carvacrol and a Total fluorescence values in the presence of the test component

thymol. subtracted from the background value obtained after NPN addi-
Toxicity of Essential Oil Components to P. tion. Each determination was done in triplicate.
leiognathi. The toxicity test based on the inhibition
of bioluminescence of the Gram-negative Photobacte- marked sensitization by thymol toward SDS, whereas
rium revealed inhibitory activity for each of the com- S. typhimurium was sensitized by thymol to a small
ponents, as shown in Figure 1. trans-Cinnamaldehyde extent toward Triton X-100. (+)-Carvone had a small
was most inhibitory, followed (in the order of toxicity) sensitizing effect toward SDS in E. coli, whereas trans-
by carvacrol, thymol, and (+)-carvone. cinnamaldehyde exhibited no sensitization for any
Effect of Essential Oil Components on the Up- combination or species.
take of NPN. Thymol and carvacrol brought about Table 5 shows that for a sensitizing effect of thymol
increased NPN uptake for both E. coli and S. typhimu- toward SDS and Triton X-100 in S. typhimurium, a
rium, as opposed to (+)-carvone and trans-cinnamalde- thymol concentration of at least 2 mM was required and
hyde, which exhibited no effect on NPN uptake for that the effect was slightly but certainly not totally
either species. None of the components yielded any reversed by MgCl2.
fluorescence in tests without bacterial cells. As shown
in Table 2, there was a large variation in the NPN Release of LPS and Protein. Shown in Figure 2
uptake values with S. typhimurium. This strain tended is a silver-stained gel in which proteinase K-treated
to take relatively high amounts of NPN also in the samples of cell-free supernatants from suspensions
absence of permeabilizing components, indicating a less treated with the essential oil components or with EDTA
stable OM than that of the E. coli strain. Magnesium were electrophoresed. In contrast to the supernatant
ions (supplied as MgCl2) in concentrations equimolar of untreated cells, the EDTA supernatant revealed a
to those of carvacrol or thymol did not affect the NPN prominent ladder pattern typical of smooth-type LPS.
uptake induced by the latter substances (Table 2). Supernatants of carvacrol- and thymol-treated suspen-
Sensitization of Bacteria to Lytic Effects. Since sions revealed ladder patterns that stained with inten-
the increased uptake of NPN observed for carvacrol and sity similar to that of the EDTA supernatant material.
thymol indicated cell envelope-permeating activity, On the basis of the staining intensity, (+)-carvone and
further experiments were performed to measure sensi- trans-cinnamaldehyde supernatants contained much
tization of the bacteria to detergent- or lysozyme- less LPS. The observed LPS patterns were all quali-
induced bacteriolysis. The test species were subjected tatively similar. After silver staining, the samples that
to treatments with the essential oil components and, were not treated with proteinase K revealed a number
subsequently, to the detergents SDS or Triton X-100 or of protein bands in patterns that did not differ between
to lysozyme. The lytic effects for each of the test essential oil or EDTA supernatants; again, thymol and
bacteria are shown in Tables 3 and 4. SDS alone had a carvacrol yielded strongest staining (not shown). In the
slight lytic effect on E. coli but a strong effect on S. gel stained with Coomassie blue, there was only one
typhimurium. Tested at a concentration of 1%, Triton prominent band present in supernatant samples of cells
X-100 alone slightly lysed both E. coli and S. typhimu- treated with EDTA, thymol, and carvacrol. The esti-
rium. The exposure to carvacrol sensitized E. coli to mated molecular mass of this protein was 37 kDa, and
SDS and S. typhimurium to Triton X-100, whereas no thus it is most likely to be one of the main porin proteins
sensitization to lysozyme was observed. There was of the E. coli OM (Nikaido and Vaara, 1985).
Action of Essential Oil Components on Gram-Negative Bacteria J. Agric. Food Chem., Vol. 46, No. 9, 1998 3593

Table 3. Lytic Activity of Lysozyme, Triton X-100, and SDS against E. coli O157:H7 with (+) or without (-)
Pretreatment with Essential Oil Components (2 mM)
percentage (( SD) remaining turbidity after 4 min in the lytic substance

test carvacrol thymol (+)-carvone trans-cinnamaldehyde

substance concn - + - + - + - +
lysozyme 10 µg/mL 101 ( 4 98 ( 1 101 ( 3 101 ( 3 99 ( 2 100 ( 1 100 ( 4 100 ( 2
Triton X-100 0.1% 104 ( 3 101 ( 1 101 ( 3 103 ( 2 100 ( 2 102 ( 1 100 ( 3 101 ( 2
Triton X-100 1% 95 ( 4 92 ( 1 94 ( 3 93 ( 3 93 ( 3 94 ( 2 93 ( 4 94 ( 4
SDS 0.1% 101 ( 1 56 ( 17 102 ( 2 42 ( 6 99 ( 1 94 ( 4 100 ( 3 99 ( 3
SDS 1% 96 ( 4 47 ( 16 93 ( 2 34 ( 6 93 ( 3 85 ( 4 94 ( 4 93 ( 4

Table 4. Lytic Activity of Lysozyme, Triton X-100, and SDS against S. typhimurium with (+) or without (-)
Pretreatment with Essential Oil Components (2 mM)
percentage (( SD) remaining turbidity after 4 min in the lytic substance

test carvacrol thymol (+)-carvone trans-cinnamaldehyde

substance concn - + - + - + - +
lysozyme 10 µg/mL 99 ( 1 97 ( 1 98 ( 1 97 ( 2 98 ( 5 97 ( 4 98 ( 4 96 ( 1
Triton X-100 0.1% 101 ( 1 88 ( 5 101 ( 2 87 ( 5 102 ( 2 100 ( 2 102 ( 3 98 ( 2
Triton X-100 1% 92 ( 1 80 ( 5 92 ( 1 77 ( 6 95 ( 3 91 ( 2 94 ( 3 91 ( 1
SDS 0.1% 68 ( 14 37 ( 20 62 ( 14 28 ( 18 57 ( 15 74 ( 13 57 ( 15 70 ( 12
SDS 1% 33 ( 9 27 ( 16 29 ( 7 19 ( 11 26 ( 1 42 ( 1 26 ( 2 36 ( 2

Table 5. Effect of Thymol Concentration and of MgCl2 on Thymol-Induced Sensitization of S. typhimurium

percentage (( SD) remaining turbidity after 4 min in the lytic substance

test 0.1 mM thymol 1 mM thymol 2 mM thymol 2 mM thymol + 2 mM MgCl2

substance concn - + - + - + - +
lysozyme 10 µg/mL 98 ( 1 100 ( 5 100 ( 3 99 ( 3 98 ( 1 97 ( 2 99 ( 2 99 ( 2
Triton X-100 0.1% 101 ( 1 101 ( 2 102 ( 1 97 ( 5 101 ( 2 87 ( 5 101 ( 3 98 ( 1
Triton X-100 1% 92 ( 1 92 ( 3 93 ( 3 88 ( 4 92 ( 1 77 ( 6 92 ( 2 88 ( 3
SDS 0.1% 65 ( 7 76 ( 6 62 ( 1 54 ( 6 62 ( 14 28 ( 18 58 ( 2 38 ( 16
SDS 1% 29 ( 7 36 ( 12 28 ( 6 28 ( 6 29 ( 7 19 ( 11 22 ( 2 27 ( 11

Notably, the proportion of LPS-specific fatty acids was

nearly 2 times higher in the thymol and carvacrol
supernatants than in the control sample, and this
proportion was also higher in the EDTA supernatant.
Effect of Essential Oil Components on ATP
Pools. Addition of glucose to washed cells of E. coli did
not change the size of the internal ATP pool (Figure 3A),
showing that harvesting and washing had no significant
effect on the bioenergetic status of the cells. Addition
of carvacrol resulted in a gradual decrease of the
internal ATP pool from 3.1 to <0.3 nmol/mg of protein
in 10 min (Figure 3A). In the same time period, a small
but significant increase of external ATP was observed
to a level corresponding to 0.75 nmol of ATP/mg of
protein (Figure 3B). Also, thymol decreased the intra-
cellular ATP pool with the appearance of extracellular
ATP (Figure 3C,D), although the changes were less
Figure 2. SDS-PAGE of proteinase K-treated cell-free prominent than those observed for carvacrol. The
supernatants of E. coli O157:H7 exposed to essential oil amount of ATP appearing outside cannot entirely be
components (2 mM) or EDTA (1 mM). An equal volume (10 explained by the decrease of the intracellular pool,
µL) of each sample was electrophoresed, whereupon the gel indicating that also increased ATP turnover occurs,
was stained with silver. Lanes: 1, untreated (control); 2, EDTA
supernatant; 3, thymol supernatant; 4, carvacrol aupernatant;
either inside or outside the cells.
5, (+)-carvone supernatant; 6, trans-cinnamaldehyde super- Addition of (+)-carvone (Figure 3A) or trans-cinna-
natant. maldehyde (Figure 3C) did not have a notable effect on
the size of the internal ATP pool during the incubation
Analysis of Released Lipid. Analysis of fatty acids period. Furthermore, exposure of the cells to (+)-
of the cell-free supernatants of E. coli treated with the carvone (Figure 3B) or trans-cinnamaldehyde (Figure
essential oil components revealed typical cellular fatty 3D) did not induce the appearance of extracellular ATP.
acids (Table 6). There was considerable liberation of These results show that of all tested oil components
lipid by 1 mM EDTA, as indicated by the nearly 5-fold carvacrol and thymol increased the permeability of the
higher amount of fatty acid in the EDTA supernatant membrane for ATP.
as compared to the control sample. Also, thymol and
carvacrol supernatants contained much lipid, whereas DISCUSSION
(+)-carvone had liberated slightly more lipid than found
in the control sample. The lipid content of the trans- With respect to bactericidal activity and membrane-
cinnamaldehyde supernatant was similar to the control. disintegrating properties, the essential oil components
3594 J. Agric. Food Chem., Vol. 46, No. 9, 1998 Helander et al.

Table 6. Fatty Acid Analysis of Cell-Free Supernatants of E. coli Cultures Treated with EDTA or with the Indicated
Essential Oil Components for 10 min at 37 °C in Tris Buffer
µg of fatty acid in 9.1 mL of culture supernatant treated with
nothing EDTA thymol carvacrol (+)-carvone trans-cinnamaldehyde
fatty acid (control) (1 mM) (2 mM) (2 mM) (2 mM) (2 mM)
12:0 0.5 2.2 1.9 1.9 0.6 0.4
14:0 0.5 4.2 2.5 2.7 1.0 0.4
14:0(3-OH) 0.7 7.0 4.9 4.9 1.0 0.5
16:1 1.2 4.6 2.6 2.3 1.6 0.6
16:0 2.0 6.5 3.7 3.3 2.8 1.1
18:1 0.9 3.6 1.8 1.4 1.3 0.5
sum of all fatty acids 5.8 28.1 17.4 16.5 8.3 3.5
sum LPS-specifica fatty acids 1.7 13.4 9.3 9.5 2.6 1.3
percentage proportion of LPS-specific 29 48 53 58 31 26
fatty acids
a LPS-specific fatty acids: 12:0; 14:0; 14:0(3-OH).

and Aktug, 1987; Didry et al., 1993). Although the

enhanced NPN uptake was already observed at some-
what lower concentrations than the MICs, thymol or
carvacrol did not directly act as OM-permeabilizing
agents such as EDTA or polyethylenimine, which dis-
integrate the OM at concentrations that do not affect
bacterial survival or growth (Vaara, 1992; Helander et
al., 1997a). Furthermore, the observation that MgCl2
had no major effect on the OM-disintegrating ability of
carvacrol or thymol suggests mechanisms of action other
than chelation of divalent cations from the OM (EDTA)
or intercalation into the OM with replacement of
stabilizing cations and concomitant disturbances of
interactions between OM constituents (polyethylen-
imine). Of the compounds studied, trans-cinnamalde-
hyde is exceptional because, without exerting disinte-
grativeeffectsontheOM,itstronglyinhibitedenterobacterial
growth and the bioluminescence of P. leiognathi. The
conclusion can thus be drawn that trans-cinnamalde-
hyde and partly also thymol and carvacrol gain access
to the periplasm and to the deeper parts of the cell;
indeed, the OM-traversing porin proteins have been
shown to allow the penetration of lipophilic probes at
Figure 3. Effect of the essential oil components (each at 2
mM) on intracellular (A, C) and extracellular (B, D) ATP pools
significant rates (Plésiat and Nikaido, 1992; Nikaido,
of E. coli ATCC 35150. Panels A and B show ATP levels after 1996). It will be of interest to measure the antibacterial
addition of carvacrol (solid circles) or (+)-carvone (solid action of these components on a porin-deficient Gram-
squares) at 5 min (indicated by arrow). Panels C and D show negative bacterial mutant; such an experiment would
ATP levels after addition of thymol (solid circles) and trans- be especially helpful in attempts to reveal the target of
cinnamaldehyde (solid squares). As a control, the ATP pools action of trans-cinnamaldehyde. Why exposure to trans-
were monitored in the absence of essential oil components
(open circles). The data come from one set of experiments with cinnamaldehyde does not result in the disintegration
duplicate measurements ((0.07 nmol). of OM, whereas carvacrol and thymol cause significant
liberation of OM components, might be related to the
investigated in this study fall into different categories. phenolic character of carvacrol and thymol; phenols are
Thymol and carvacrol both have prominent OM- known for their membrane-disturbing activities (Keweloh
disintegrating properties, as indicated by their enhanc- et al., 1990; Sikkema et al., 1995). This property of
ing effect on NPN uptake and LPS release, as well as phenol is actually widely utilized in extraction of LPS
sensitization to detergents. These compounds also from bacteria (Helander et al., 1985). A plausible reason
inhibited bacterial growth at concentrations similar to underlying the inability of (+)-carvone to inhibit growth
those required for OM disintegration and increased the or to affect the OM could be that it is pumped out from
permeability of the cytoplasmic membrane to ATP. the periplasm at a rate exceeding its penetration rate
trans-Cinnamaldehyde was inhibitory at a similar (Nikaido, 1994). (+)-Carvone, on the other hand, pos-
concentration as thymol and carvacrol for the growth sesses a strong antifungal activity (Oosterhaven et al.,
of the enteric bacteria and exhibited high toxicity in the 1996; Smid et al., 1996). The same is true for trans-
photobacter test, yet it exhibited neither OM-disinte- cinnamaldehyde. In a study with Saccharomyces cer-
grating activity nor depletion of intracellular ATP. (+)- evisiae it was demonstrated that trans-cinnamaldehyde
Carvone was ineffective on the OM and exhibited low (at 5 mM) caused a (partial) collapse of the integrity of
activity both in growth inhibition and in photobacter the cytoplasmic membrane, leading to excessive leakage
tests and did not affect the cellular ATP pool. of metabolites and enzymes from the cell and finally loss
The MIC values presented here for carvacrol, thymol, of viability (Smid et al., 1996).
and trans-cinnamaldehyde agree with previously re- Beside shedding light on the differential toxicity of
ported values for E. coli and S. typhimurium (Karapinar essential oil components on Gram-negative bacteria, our
Action of Essential Oil Components on Gram-Negative Bacteria J. Agric. Food Chem., Vol. 46, No. 9, 1998 3595

data further validate the importance of low molecular Lundin, A.; Thore, A. Analytical information obtainable by
mass lipophilic compounds. Despite their limited solu- evaluation of the time course of firefly bioluminescence in
bility in water, these compounds are able to penetrate the assay of ATP. Anal. Biochem. 1975, 66, 47-63.
Gram-negative bacteria and may thus be able to influ- Moore, L. V. H.; Bourne, D. M.; Moore, W. E. C. Comparative
ence the proliferation of certain pathogenic or spoilage distribution and taxonomic value of cellular fatty acids in
bacteria in food-related environments. thirty-three genera of anaerobic Gram-negative bacilli. Int.
J. Syst. Bacteriol. 1994, 44, 338-347.
Nikaido, H. Prevention of drug access to bacterial targets:
ACKNOWLEDGMENT
permeability barriers and active efflux. Science 1994, 264,
We thank Päivi Lepistö, Anna-Liisa Ruskeepää, and 382-388.
Marja-Liisa Jalovaara for excellent technical assistance. Nikaido, H. Outer membrane. In Escherichia coli and Salmo-
nella: Cellular and Molecular Biology; Neidhardt, F. C., Ed.;
LITERATURE CITED ASM Press: Washington, DC, 1996; Vol. I., pp 29-47.
Nikaido, H.; Vaara, M. Molecular basis of bacterial outer
Conner, D. E. Naturally occurring compounds. In Antimicrobi- membrane permeability. Microbiol. Rev. 1985, 49, 1-32.
als in Foods; Davidson, P. M., Branen, A. L., Eds.; Dekker:
Oosterhaven, K.; Chambel Leitao, A.; Gorris, L. G. M.; Smid,
New York, 1993; pp 441-468.
E. J. Comparative study on the action of S-(+)-carvone, in
Deans, S. G.; Ritchie, G. Antibacterial properties of plant
situ, on the potato storage fungi Fusarium solani var.
essential oils. Int. J. Food Microbiol. 1987, 5, 165-180.
coeruleum and F. sulphureum. J. Appl. Bacteriol. 1996, 80,
Didry, N.; Dubreuil, L.; Pinkas, M. Activité antibactérienne
535-539.
du thymol, du carvacrol et de l’aldéhyde cinnamique seuls
ou associes (Antibacterial activity of thymol, carvacrol, and Plésiat, P.; Nikaido, H. Outer membranes of gram-negative
cinnamaldehyde singly or in combinations). Pharmazie bacteria are permeable to steroid probes. Mol. Microbiol.
1993, 48, 301-304. 1992, 6, 1323-1333.
Helander, I. M. Isolation and electrophoretic analysis of Sikkema, J.; de Bont, J. A. M.; Poolman, B. Mechanism of
bacterial lipopolysaccharides. In Enterobacterial Surface membrane toxicity of hydrocarbons. Microbiol. Rev. 1995,
Antigens: Methods for Molecular Characterization; Kor- 59, 201-222.
honen, T. K., Mäkelä, P. H., Dawes, E. A., Eds.; Elsevier: Sivropoulou, A.; Papanikolaou, E.; Nikolaou, C.; Kokkini, S.;
Amsterdam, 1985; pp 263-274. Lanaras, T.; Arsenakis, M. Antimicrobial and cytotoxic
Helander, I. M.; Haikara, A. Cellular fatty acids and alkenyl activities of Origanum essential oils. J. Agric. Food Chem.
residues in Megasphaera and Pectinatus species: contrast- 1996, 44, 1202-1205.
ing profiles and detection of beer spoilage. Microbiology Smid, E. J.; Koeken, J. P. G.; Gorris, L. G. M. Fungicidal and
1995, 141, 1131-1137. fungistatic action of the secondary plant metabolites cin-
Helander, I. M.; Alakomi, H.-L.; Latva-Kala, K.; Koski, P. namaldehyde and carvone. In Modern Fungicides and
Polyethyleneimine is an effective permeabilizer of Gram- Antimicrobial Compounds; Lyr, H., Russell, P. E., Sisler,
negative bacteria. Microbiology 1997a, 143, 3193-3199. H. D., Eds.; Intercept: Andover, U.K., 1996; pp 173-180.
Helander, I. M.; von Wright, A.; Mattila-Sandholm, T.-M.
Potential of lactic acid bacteria and novel antimicrobials Ten Brink, B.; Otto, R.; Hansen, U. P.; Konings, W. N. Energy
against Gram-negative bacteria. Trends Food Sci. Technol. recycling by lactate efflux in growing and non-growing cells
1997b, 8, 146-150. of Streptococcus cremoris. J. Bacteriol. 1985, 162, 383-390.
Juven, B. J.; Kanner, J.; Schved, F.; Weisslowicz, H. Factors Vaara, M. Agents that increase the permeability of the outer
that interact with the antibacterial action of thyme essential membrane. Microbiol. Rev. 1992, 56, 395-411.
oil and its active constituents. J. Appl. Bacteriol. 1994, 76, Zähringer, U.; Lindner, B.; Rietschel, E. Th. Molecular struc-
626-631. ture of lipid A, the endotoxic center of bacterial lipopolysac-
Karapinar, M.; Aktug, S. E. Inhibition of foodborne pathogens charides. Adv. Carbohydr. Chem. Biochem. 1994, 50, 211-
by thymol, eugenol, menthol and anethole. Int. J. Food 276.
Microbiol. 1987, 4, 161-166.
Keweloh, H.; Weyrauch, G.; Rehm, H.-J. Phenol-induced
membrane changes in free and immobilized Escherichia coli. Received for review February 18, 1998. Revised manuscript
Appl. Microbiol. Biotechnol. 1990, 33, 66-71. received June 1, 1998. Accepted June 30, 1998. This work was
Loh, B.; Grant, C.; Hancock, R. E. W. Use of the fluorescent financially supported by the European Commission through
probe 1-N-phenylnaphthylamine to study the interactions the Projects NISINPLUS (FAIR-CT96-1148) and Green Chemi-
of aminoglycoside antibiotics with the outer membrane of cals for Crop Protection (FAIR-CT95-0722).
Pseudomonas aeruginosa. Antimicrob. Agents Chemother.
1984, 26, 546-551. JF980154M