Isoflavones as potentiators of antibacterial activity

Article abstract of DOI:10.1021/jf0302714

Isoflavones isolated from Lupinus argenteus were found to potentiate the antibacterial activity of α-linolenic acid, also found in the same plant. The isoflavones also potentiated the activity of the natural plant antibiotic berberine and the synthetic fluoroquinoline antibiotic norfloxacin. The isoflavones increased the uptake of berberine into Staphylococcus aureus cells, indicating that they may be inhibiting a multidrug resistance pump (MDR). Thus, L. argenteus contains a weak antibacterial and also MDR pump inhibitors, which increase its potency.

Full text of DOI:10.1021/jf0302714

J. Agric. Food Chem. 2003, 51, 5677−5679 5677
Isoflavones As Potentiators of Antibacterial Activity
‡
CEÄ CILE MOREL,† FRANK R. STERMITZ,*,† GEORGE TEGOS,‡ AND KIM LEWIS
Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523-1872, and
Department of Biology, Northeastern University, Boston, Massachusetts 02115
Isoflavones isolated from Lupinus argenteus were found to potentiate the antibacterial activity of
R-linolenic acid, also found in the same plant. The isoflavones also potentiated the activity of the
natural plant antibiotic berberine and the synthetic fluoroquinoline antibiotic norfloxacin. The isoflavones
increased the uptake of berberine into Staphylococcus aureus cells, indicating that they may be
inhibiting a multidrug resistance pump (MDR). Thus, L. argenteus contains a weak antibacterial and
also MDR pump inhibitors, which increase its potency.
KEYWORDS: Lupinus argenteus; Leguminosae; isoflavones; linolenic acid; antibacterial; potentiation;
MDR pump; Staphylococcus aureus
INTRODUCTION
Lupinus argenteus Pursh. subspecies rubricaulis (Greene) Hess
and Dunne had both direct antibiotic activity as well as antibiotic
potentiation properties. This species had previously been studied
only for its alkaloid content (7).
Bacterial resistance to antibiotics is a continuing problem,
and microorganisms seem able to develop resistance to new
drugs as rapidly as they are introduced. One of the major modes
by which pathogens develop resistance is through development
or enhancement of methods for the removal of antibiotics that
have entered cells of the organism. Thus, resistant bacteria
possess efficient systems known generically as multidrug
resistance (MDR) pumps (1). If the action of such pumps can
be inhibited, then the cellular content of antibiotics will be
increased and the antibiotic treatment will again be effective.
We reasoned that plants might develop MDR pump inhibitors
to enhance the activity of their own natural antimicrobial
compounds, and we have been able to show that this is indeed
the case (2-6). Essentially, we have been looking for plant
products that do not themselves possess antibacterial activity
but can potentiate known antibiotics by inhibiting microbial
MDR pumps.
For this search, a bioassay was developed (1) in which a
subinhibitory dose of an antibiotic was combined with plant
extracts or fractions from a chromatography to test for potentia-
tion. Staphylococcus aureus was used as a test organism because
of its known resistance and because an S. aureus mutant, which
lacks the MDR efflux protein (known as NorA), was available
(1). Berberine was chosen as a model antibiotic since its
fluorescent properties when complexed with cellular DNA
allowed for measurement of antibiotic uptake and efflux from
bacterial cells. The value of this method for bioactivity-directed
fractionation was demonstrated by its use in discovery of a
potent flavonolignan MDR pump inhibitor from Berberis species
(2-4) as well as for structure-activity relationship studies on
flavonolignans and flavones as MDR pump inhibitors (5, 6).
Screening of randomly chosen plants revealed that extracts of
MATERIALS AND METHODS
General Experimental Procedures and Chemicals. Instruments
and chromatography procedures used were as previously described (3)
as were detailed methods for bioassays and berberine uptake into cells
(2, 3). Bacterial strains used were S. aureus wild-type (8325-4), NorA
mutant KLE 8, and Berberis megaterium (11561, M. Cannon). Bacteria
were cultured in Mueller-Hinton (MH) broth. Amounts of 105 cells/
mL were inoculated into 1 mL of MH broth and dispensed at 0.2 mL/
well in microtiter plates. Berberine chloride (Aldrich Chemical) was
added at about one-eighth of its minimum inhibitor concentration (MIC),
and MICs of test fractions or compounds were determined by serial
2-fold dilution in the presence and absence of berberine chloride. MIC
was defined as the concentration of an antibacterial agent that
completely prevented cell growth during an 18 h incubation at 37 °C.
Growth was assayed with a microtiter plate reader (Biorad) by reading
absorption at 600 nm.
Genistein (2), biochanin A (4), R-linolenic acid (Z,Z,Z-9,12,15-
octadecatrienoic acid, 1), γ-linolenic acid (Z,Z,Z-6,9,12-octadecatrienoic
acid, 5), and their methyl esters were purchased from Sigma (St. Louis,
MO). INF271 was a gift from Influx Corp. INF is a designation used
for compound numbers by INFlux Corp.
Plant Material. L. argenteus Pursh. subsp. rubricaulis (Fabaceae)
was collected on July 2, 2001 at the same site as previously described
(7) (Colorado State University voucher GH171, identified by D. B.
Dunn, Department of Biology, University of Missouri, Columbia).
Isolation. Dried, ground leaves (164 g) were submerged successively
in solvents (two × 1.0 L each), which upon evaporation yielded as
follows: hexanes extract (1.2 g; 250 µg/mL direct growth inhibitory
activity against S. aureus), CHCl3 (5.4 g; 62.5 µg/mL), EtOAc (0.6 g;
not tested), and MeOH (12.5 g; inactive). Similarly, dried, ground stems
(127 g) were extracted to yield hexanes extract (0.40 g; 125 µg/mL
activity), CHCl3 (0.56 g; inactive), EtOAc (0.26 g; 62.5 µg/mL), and
MeOH (7.2 g; inactive).
* To whom correspondence should be addressed. Tel: 970-491-5158.
Fax: 970-491-5610. E-mail: frslab@lamar.colostate.edu.
† Colorado State University.
Vacuum liquid chromatography (VLC) of each of the nonpolar
extracts followed by bioassay yielded 1 as the directly active component.
Typical procedures were as follows.
‡ Northeastern University.
10.1021/jf0302714 CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/09/2003

5678 J. Agric. Food Chem., Vol. 51, No. 19, 2003
Morel et al.
Table 1. Growth Inhibition of Bacteria by Antimicrobials (Alone) and in Combination with Isoflavone MDR Inhibitors
MIC (µg/mL)
S. aureus
NorA mutant
B. megaterium
alone
4a
31.3
7.81
7.81
0.25
2b
3c
alone
4a
2b
3c
16
5
2.5
0.5
alone
4a
2b
31.3
7.81
7.81
1
3c
berberine
1
5
500
62.5
25
31.3
16
16
31.3
10
5
64
10
5
7.81
2
1.5
0.25
31.3
5
5
250
100
50
31.3
7.81
7.81
0.25
31.3
16
16
norfloxacin
1
0.5
0.25
0.25
0.5
1
0.5
a-c Compounds 2−4 were added at a final concentration of 10 µg/mL.
The dried leaf hexane extract (1.2 g) was separated by VLC (silica
gel, 45 g; hexanes followed by hexanes/EtOAc mixtures; 11 fractions,
10-15 mL each). Fraction 3 (9:1 hexanes/EtOAc,101 mg) was 1.
Similarly, 189 mg of 1 was obtained from the leaf CHCl3 extract.
The leaf EtOAc extract was purifed by VLC (silica gel, 45 g; CHCl3,
followed by CHCl3 /MeOH mixtures; 15 fractions, 40-50 mL each).
Fractions 6 and 7 (9:1 CHCl3/MeOH; 25 µg/mL potentiation activity
with subinhibitory berberine) were combined (74 mg) and purified by
preparative thin-layer chromatography (PTLC; 9:1 CHCl3/MeOH) to
yield 2 (band 2; 7 mg; 25 µg/mL) and orobol (3; band 8; 4 mg; 12.5
µg/mL). The stem EtOAc extract also yielded 2 and 3 in similar
amounts.
The stem CHCl3 extract was purified by VLC (silica gel, 45 g;
gradient of hexanes/EtOAc, EtOAc, MeOH; 15 × 50 mL fractions).
Potentiation active fractions (nos. 10-12, 12.5-25 µg/mL) were
combined (48 mg) and purified by PTLC (9:1 CHCl3/MeOH) to yield
4 (band 1; 4 mg; 6.25 µg/mL) and 2 (band 2; 2 mg).
Although neither of the MeOH extracts was active, they were also
fractionated by VLC. None of the fractions from the leaf extract were
active. The stem extract, however, yielded potentiation active fractions,
which proved to contain small amounts of 2 and 3. An inactive fraction
yielded the alkaloid 2,9-dihydroxyaphyllidine (7).
Identification of 1 was by interpretation of 1H and 13C NMR spectra
and by spectral comparison with a commercial sample. Identifications
1
of 2 and 4 were by H and 13C NMR spectra and MS in comparison
with commercial samples and 3 by analysis of the MS and UV (12)
spectra and the 1H and 13C NMR spectra in comparison with literature
values (13).
5,7,4′-Trimethoxyisoflavone (7). A mixture of 4 (92 mg, 0.32 mmol)
and 5 mL of a 2.0M solution of trimethylsilyldiazomethane (TMSD)
in MeOH was stirred at room temperature overnight. The excess TMSD
was decomposed with AcOH, and the mixture was evaporated to
dryness. VLC purification (Si gel; 6:4 hexanes/EtOAc) yielded 81 mg
(0.26 mmol, 80% yield) of 7 (14).
Figure 1. Structures of antibacterial compounds and potentiators.
RESULTS AND DISCUSSION
respects as was 7, obtained by methylation of 4. The linolenic
acids had weak direct activity against the Gram-positive
organisms S. aureus and B. megaterium, while the isoflavones
were inactive at the highest dose tested (100 µg/mL). Compound
4 was, however, a potentiator of berberine and norfloxacin
antibiotic activity, as well as in combination with 1 and 5,
against wild-type S. aureus and against B. megaterium. For
example, a combination of 6.25 µg/mL of 4 and subinhibitory
(30 µg/mL) berberine or subinhibitory 1 (30 µg/mL) completely
inhibited S. aureus and B. megaterium growth. In Table 1, data
are given for experiments in which each of the isoflavones was
added at 10 µg/mL to berberine, 1, 5, and the fluoroquinoline
antibiotic norfloxacin and tested against wild-type S. aureus,
B. megaterium, and a mutant S. aureus that lacked the NorA
efflux pump protein. Potentiation of activity was observed in
all cases. Note that the direct activities of the four antibiotics
against the NorA mutant S. aureus had dropped significantly
in comparison with the wild-type. Thus, the linolenic acids (at
5 and 10 µg/mL direct activity against the mutant) can be added
to the large group of plant compounds whose inherent activity
is hidden by the presence of bacterial efflux pumps (8). To
Dried, ground leaves and stems were extracted successively
with hexanes, CHCl3, EtOAc, and MeOH, and the extracts were
tested for direct growth inhibition against S. aureus and also
for potentiation of subinhibitory berberine. The hexanes and
CHCl3 extracts yielded directly active fractions after VLC, and
these were shown to contain 1 (Figure 1). Acid 1 had weak
activity (MIC, 62.5 µg/mL). The leaf EtOAc extract, when
fractionated, yielded additional 1 and also fractions that were
not directly active but that potentiated berberine. PTLC of these
fractions yielded 2 and 3 as the active potentiators. The nonpolar
stem extracts yielded 1 as well as 2 and 4. None of the fractions
from the leaf MeOH extract were active, but the stem MeOH
extract yielded additional small amounts of 2 and 3 as well as
the quinolizidine alkaloid 2,9-dihydroxyaphyllidine (7). This
alkaloid was inactive, either directly or as a potentiator, as was
aphyllidine (7).
To measure potentiation activities in more detail and to test
for structure-activity relationships, we used commercial samples
of 1-5. Commercial samples of the methyl esters of 1 and 2
and 6,7,4′-trimethoxyisoflavone (6) proved to be inactive in all

Isoflavones Potentiate Antibiotics
J. Agric. Food Chem., Vol. 51, No. 19, 2003 5679
(2) Stermitz, F. R.; Lorenz, P.; Tawara, J. N.; Zenewicz, L.; Lewis,
K. Synergy in a medicinal plant: antimicrobial action of
berberine potentiated by 5′-methoxyhydnocarpin, a multidrug
pump inhibitor. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1433-
1437.
(3) Stermitz, F. R.; Tawara-Matsuda, J.; Lorenz, P.; Mueller, P.;
Zenewicz, L.; Lewis, K. 5′-Methoxyhydnocarpin-D and pheophor-
bide a: Berberis species components which potentiate berberine
growth inhibition of resistant Staphylococcus aureus. J. Nat.
Prod. 2000, 63, 1146-1149.
(4) Stermitz, F. R.; Beeson, T. D.; Mueller, P.; Hsiang, J.-F.; Lewis,
K. Staphylococcus aureus MDR efflux inhibitors from a Berberis
and a Mahonia (sensu strictu) species. Biochem. Syst. Ecol. 2001,
29, 793-798.
(5) Guz, N. R.; Stermitz, F. R.; Johnson, J. B.; Beeson, T. D.; Willen,
S.; Hsiang, J.-H.; Lewis, K. Flavonolignan and flavone inhibitors
of a Staphylococcus aureus multidrug resistance pump: Structure-
activity relationships. J. Med. Chem. 2001, 44, 261-268.
(6) Stermitz, F. R.; Scriven, L.; Tegos, G.; Lewis, K. Two flavonols
from Artemisia annua which potentiate the activity of berberine
and norfloxacin against a resistant strain of Staphylococcus
aureus. Planta Med. 2002, 68, 1140-1141.
(7) Arslanian, R. L.; Harris, G. H.; Stermitz, F. R. New quinolizidine
alkaloids from Lupinus argenteus and its hosted root parasite
Castilleja sulphurea. Stereochemistry and conformation of some
naturally occurring cyclic carbinolamides. J. Org. Chem. 1990,
55, 1204-1210.
(8) Tegos, G.; Stermitz, F. R.; Lomovskaya, O.; Lewis, K. Multidrug
pump inhibitors uncover remarkable activity of plant antimicro-
bials. Antimicrob. Agents Chemother. 2002, 46, 3133-3141.
(9) Markham, P. N.; Westhaus, E.; Klyacho, K.; Johnson, M. E.;
Neyfakh, A. A. Multiple novel inhibitors of the NorA multidrug
transporter of Staphylococcus aureus. Antimicrob. Agents Chemo-
ther. 1999, 43, 2404-2408.
(10) Lee, J.-Y.; Kim, Y.-S.; Shin, D.-H. Antimicrobial synergistic
effect of linolenic acid and monoglyceride against Bacillus cereus
and Staphylococcus aureus. J. Agric. Food Chem. 2002, 50,
2193-2199.
(11) Kaatz, G. W.; Seo, S. M.; O’Brien, L.; Wahiduzzaman, M.;
Foster, T. J. Evidence for the existence of a multidrug efflux
transporter distinct from NorA in Staphylococcus aureus. Anti-
microb. Agents Chemother. 2000, 44, 1404-1406.
(12) Mabry, T. J.; Markham, K. R.; Thomas, M. B. The Systematic
Identification of FlaVonoids; Springer-Verlag: New York, 1970;
p 204.
(13) Asres, K.; Mascagni, P.; O’Neill, M. J.; Phillipson, J. D.
Isoflavonoids from Bolusanthus speciosus (Bolus) Harms
Leguminosae. Z. Naturforsch. 1985, 40c, 617-620.
(14) Tahara, S.; Hashidoko, Y.; Ingham, J. L.; Mizutani, J. New 5-O-
methylisoflavones in the roots of yellow lupin (Lupinus luteus
L. cv. Barpine). J. Agric. Biol. Chem. 1986, 50, 1809-1819.
Figure 2. Accumulation of berberine in cells of S. aureus alone or with
added 2−4 or INF271. Berberine uptake was measured by the increase in
fluorescence following binding to DNA and expressed as RFU (relative
fluorescence units).
provide further evidence that the potentiation of antibiotic
activity was due to inhibition of an MDR efflux pump, we
examined transport of berberine into cells of S. aureus (Figure
1). The effects of the isoflavones were compared with those of
the known S. aureus NorA pump inhibitor INF271 (9). Although
INF271 was somewhat more potent than the isoflavones by 30
min, all three isoflavones had increased the content of berberine
in the S. aureus cells, with 4 being the most effective. If NorA
inhibition had been complete or if it was the only efflux pump
protein in S. aureus, then there should have been no difference
in cell berberine content in the NorA mutant in the control as
compared to the isoflavone treated case. There is at least one
more MDR in S. aureus (11), and this might explain the
presence of residual inhibition of berberine accumulation in the
mutant.
There are both similarities and differences among the
structure-activity relationships of these isoflavones with some
similar flavones previously studied (5). Luteolin, the flavone
analogue of 3, was inactive as an MDR pump inhibitor while 3
(present work) was quite active. On the other hand, both tetra-
O-methylluteolin and its isoflavone analogue 7 were inactive.
In the flavone series, monomethoxy B-ring derivatives were
generally more active than the disubstituted B-ring derivatives.
Activity was previously reported in the 20-50 µg/mL range
for 1 against S. aureus and Berberis cereus, and a combination
of 10 µg/mL 1 with 10 µg/mL of some inactive monoglycerides
was more effective than 1 alone (10). Linolenic acids are
widespread in various seed oils, and isoflavones are common
in foods such as soy products. Isoflavone synergists present in
the human diet may enhance the effectiveness of standard
antibiotic treatments for some bacterial diseases.
Received for review April 9, 2003. Revised manuscript received July
14, 2003. Accepted July 16, 2003. This work was supported by the
National Institutes of Health Grant RO1 GM59903 (K.L.) and Colorado
State University Agricultural Experiment Station Project 271 (F.R.S.).
Mass spectra were obtained on instruments supported by the National
Institutes of Health shared instrumentation Grant GM49631.
LITERATURE CITED
(1) Hsieh, P.-C.; Siegel, S. A.; Rogers, B.; Davis, D.; Lewis, K.
Bacteria lacking a multidrug pump: A sensitive tool for drug
discovery. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 6602-6606.
JF0302714