University of Zurich
Zurich Open Repository and Archive
GABA(A) receptors as in vivo substrate for the anxiolytic action
of valerenic acid, a major constituent of valerian root extracts
Benke, D; Barberis, A; Kopp, S; Altmann, K H; Schubiger, M; Vogt, K E; Rudolph,
U; Möhler, H
Benke, D; Barberis, A; Kopp, S; Altmann, K H; Schubiger, M; Vogt, K E; Rudolph, U; Möhler, H (2009).
GABA(A) receptors as in vivo substrate for the anxiolytic action of valerenic acid, a major constituent of valerian
root extracts. Neuropharmacology, 56(1):174-181.
Postprint available at:
Posted at the Zurich Open Repository and Archive, University of Zurich.
Originally published at:
Neuropharmacology 2009, 56(1):174-181.
GABA(A) receptors as in vivo substrate for the anxiolytic action
of valerenic acid, a major constituent of valerian root extracts
Valerian extracts have been used for centuries to alleviate restlessness and anxiety albeit with unknown
mechanism of action in vivo. We now describe a specific binding site on GABA(A) receptors with nM
affinity for valerenic acid and valerenol, common constituents of valerian. Both agents enhanced the
response to GABA at multiple types of recombinant GABA(A) receptors. A point mutation in the beta2
or beta3 subunit (N265M) of recombinant receptors strongly reduced the drug response. In vivo,
valerenic acid and valerenol exerted anxiolytic activity with high potencies in the elevated plus maze
and the light/dark choice test in wild type mice. In beta3 (N265M) point-mutated mice the anxiolytic
activity of valerenic acid was absent. Thus, neurons expressing beta3 containing GABA(A) receptors
are a major cellular substrate for the anxiolytic action of valerian extracts.
GABAA receptors as in vivo substrate for the anxiolytic action of
valerenic acid, a major constituent of valerian root extracts
Dietmar Benke1, Andrea Barberis1,3, Sascha Kopp2, Karl-Heinz Altmann2, Monika
Schubiger1, Kaspar Vogt1,5, Uwe Rudolph1,4, Hanns Möhler1,2,6
1Institute of Pharmacology and Toxicology, University of Zurich, Winterthurerstr. 190,
CH-8057 Zurich, Switzerland
2Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology (ETH),
Wolfgang-Pauli-Str. 10 CH-8093 Zürich, Switzerland
3Present address: Georgetown University, Washington, USA
4Present address: McLean Hospital, Harvard Medical School, Belmont MA, USA
5Present address: Biocenter, University of Basle, Switzerland
6Collegium Helveticum, Zurich, Switzerland
Running title: Valerenic acid in vivo acts via GABAA receptors
Key words: GABAA receptors, valerenic acid, anxiety, sedation, Valeriana officinalis
Address for correspondence: Uwe Rudolph, MD
Harvard Medical School
Belmont MA, USA
Valerian extracts have been used for centuries to alleviate restlessness and anxiety
albeit with unknown mechanism of action in vivo. We now describe a specific binding
site on GABAA receptors with nM affinity for valerenic acid and valerenol, common
constituents of valerian. Both agents enhanced the response to GABA at multiple
types of recombinant GABAA receptors. A point mutation in the β2 or β3 subunit
(N265M) of recombinant receptors strongly reduced the drug response. In vivo,
valerenic acid and valerenol exerted anxiolytic activity with high potencies in the
elevated plus maze and the light/dark choice test in wild type mice. In β3 (N265M)
point-mutated mice the anxiolytic activity of valerenic acid was absent. Thus, neurons
expressing β3 containing GABAA receptors are a major cellular substrate for the
anxiolytic action of valerian extracts.
Keywords: : GABAA receptors, valerenic acid, anxiety, sedation, Valeriana officinalis
Various extracts from the roots of plants of the genus Valeriana (Valerianaceae) are
used in herbal medicine of many cultures as mild sedatives and tranquilizers.
Valerenic acid is a main constituent of Valeriana officinalis, most widely used in
Europe and USA. In animal experiments, valerenic acid or extracts from valerian
showed tranquilizing and/or sedative activity (Bent et al., 2006; De Feo and Faro,
2003; Hendriks et al., 1985). Activation of adenosine receptors has been implicated
in the action of valerian ingredients (Müller et al., 2002; Schumacher et al., 2002).
Valerenic acid was also shown to modulate or - at high concentrations - activate
GABAA receptors as shown for recombinant receptors expressed in Xenopus oocytes
(Khom et al., 2007) or neonatal brainstem neurons (Yuan et al., 2004). However, the
molecular mechanism of action of any of the valerian ingredients in vivo is yet
unknown. We have now identified a specific binding site for valerenic acid and
valerenol, common constituents of valerian, on GABAA receptors. In addition, it was
demonstrated that valerenic acid mediates anxiolytic activity via GABAA receptors
containing the β3 subunit. Mice containing a single amino acid point mutation in the β3
subunit [β3(N265M)] failed to show anxiolytic activity to valerenic acid, but maintained
an anxiolytic response to diazepam.
2. Materials and Methods
Valerenic acid was purchased from Extrasynthese, Lyon (France). TTX was
purchased from Alomone labs (Israel). All other chemicals were from Sigma-Aldrich
(Switzerland). The α1 subunit-specific antibody is described in Benke et al. (1991).
2.2. Synthesis of valerenic acid derivatives
Valerenol (2), valerenal (3), and O-methyl-valerenol (4) were synthesized according
to Scheme S1.
Scheme S1: i. LiAlH4, THF, rt, 2h, 92%. ii. (COCl)2, DMSO, Et3N, CH2Cl2, -78°C→
RT, 30 min, 89%. iii. CH3I, NaH, ether, -78°C→ RT, 84%.
2.2.1. Valerenol (2)
Valerenic acid (1) (275 mg, 1.17 mmol, 1 eq) was dissolved in THF (10 ml) under Ar.
A 1 M solution of LiAlH4 in THF was then added (1.1 ml, 1.1 mmol, 1 eq) and the
reaction mixture was stirred at room temperature for 2 h. The reaction was quenched
with diluted aq H2SO4 and extracted with diethyl ether. The organic phase was dried
over Na2SO4 and concentrated in vacuo. The resulting crude product was purified by
flash chromatography (FC) in AcOEt/hexane ¼, yielding 238 mg (92%) of Valerenol
(2) as colourless oil. AcOEt = ethyl acetate.
Rf (AcOEt/hexane 1/4): 0.36 (vanilline/H2SO4) 1H-NMR (500 MHz, CDCl3): 5.75 (d, J
= 9.3 Hz, 1H), 4.01 (s, 2H), 3.46 (dd, J1 = 4.9 Hz, J2 = 8.5 Hz, 1H), 2.94-2.92 (m, 1H),
2.20 (t, J = 7.6 Hz, 2H), 1.97-1.96 (m, 1H), 1.89-1.76 (m, 3H), 1.73 (s, 3H), 1.65 (s,
3H), 1.59-1.49 (m, 1H), 1.40-1.27 (m, 2H), 0.77 (d, J = 7.0 Hz, 3H). 13C-NMR (125
MHz, CDCl3): 147.5, 135.1, 133.0, 129.1, 69.2, 47.3, 37.4, 33.4, 33.1, 26.1, 24.5,
28.5, 13.7, 13.5, 12.0.
2.2.2. Valerenal (3)
A solution of oxalyl chloride (9 μl, 0.102 mmol, 1.5 eq) in CH2Cl2 (0.5 ml) under Ar
was cooled to -78° and DMSO (7.5 μl, 0.102 mmol, 1.5 eq) was added. After 15 min
a solution of valerenol (2) (15 mg, 0.068 mmol, 1 eq) in CH2Cl2 (0.1 ml) was added
followed by triethylamine (125 μl, 0.340 mmol, 5 eq.; after 15 additional min). The
cooling bath was removed after 5 min and stirring was continued at room
temperature for 30 min. Then the reaction was quenched with water and the mixture
extracted with diethylether. The organic layer was separated, dried over Na2SO4, and
concentrated in vacuo. The resulting crude product was purified by FC
(ether/pentane 1/10), yielding 13.3 mg (89%) of Valerenal (3) as a colourless oil.
Rf (ether/pentane 1/2): 0.39 (vanillin/H2SO4). 1H-NMR (500 MHz, CDCl3): 9.38 (s,
1H), 6.75 (qd, J1 = 9.8 Hz, J2 = 1.3Hz, 1H), 3.74-3.69 (m, 1H), 2.92-2.98 (m, 1H),
2.24-2.19 (t, J = 7.4 Hz, 2H), 2.06-1.97 (m, 1H), 1.89-1.81 (m, 3H), 1.80 (d, J = 1.3
Hz, 3H), 1.65 (s, 3H), 1.61-1.53 (m, 1H), 1.50-1.42 (m, 2H), 0.80 (d, J = 7.3 Hz, 3H).
13C-NMR(125 MHz, CDCl3): 195.9, 156.1, 139.9, 137.3, 132.4, 47.3, 37.2, 34.5, 32.7,
28.6, 25.0, 24.3, 13.4, 11.7, 9.1.
2.2.3. O-Methyl valerenol (4)
To a solution of valerenol (2) (19 mg, 0.09 mmol, 1 eq) in diethylether (1 ml) was
added sodium hydride (2.3 mg, 0.10 mmol,1.1 eq) at -78°C under Ar. After 15 min
methyl iodide (6 µl, 0.10 mmol, 1.1 eq) was added and the mixture stirred at room
temperature for 1 h. The reaction was then quenched with water and extracted with
diethylether. The organic layer was separated, dried over Na2SO4, and concentrated
in vacuo. The crude product was purified by FC (AcOEt/hexane 1/10), yielding 16 mg
(84%) of O-methylvalerenol (4) as a colourless oil. Rf (AcOEt/hexane 1/10): 0.76
(vanillin/H2SO4). 1H-NMR (500 MHz, CDCl3): 5.73 (dq, J1 = 9.1 Hz, J2 = 1.3 Hz, 1H),
3.83-3.76 (m, 2H), 3.48-3.43 (m, 1H), 3.28 (s, 3H), 2.96-2.89 (m, 1H), 2.19 (t, J = 7.6
Hz, 2H), 2.00-1.92 (m, 1H), 1.88-1.77 (m, 2H), 1.75-1.67 (m, 1H), 1.70 (d, J = 7.6 Hz,
3H), 1.65-1.63 (m, 3H),1.57-1.50 (m, 1H), 1.40-1.34 (m, 2H), 0.77 (d, J = 6.9 Hz, 3H).
13C-NMR (125 MHz, CDCl3): 135.1, 130.2, 129.8, 129.0, 78.9, 57.2, 47.3, 37.4, 33.3,
33.2, 28.6, 26.2, 24.5, 13.7, 13.2, 12.0.
2.3. Preparation of crude neuronal membranes
Crude membranes were prepared from adult rat brain tissue. Following
homogenization in 10 volumes of 10 mM Tris-HCl, pH 7.4, 0.32 M sucrose, 5 mM
EDTA, 0.1 mM phenylmethylsulfonyl fluoride and centrifugation at 1000 g for 10 min
the resulting supernatant was centrifuged for 20 min at 12,000 g to obtain the crude
membrane pellet. The crude membranes were resuspended in buffer and stored at -
80°C until used. For [3H]valerenic, [3H]flunitrazepam and [3H]flumazenil binding,
crude membrane were thawed and washed three times in 50 mM Tris pH 7.4. For
[3H]muscimol binding, crude membranes were resuspended in 20-40 volumes of 5
mM Tris-HCl 7.4, frozen in liquid nitrogen, thawed and centrifuged for 20 min at
45,000 g. This procedure to eliminate endogenous GABA was repeated two times.
Radioligand binding using [3H]muscimol, [3H]flunitrazepam and [3H]flumazenil was
performed as described previously (Benke et al., 1991).
2.4. [3H]Valerenic binding assay
To establish a radioligand binding assay, valerenic acid was custom-labeled by
tritium using random catalytic exchange of hydrogen by tritium (6.5 Ci/mmol; RC
Tritec, Teufen, Switzerland). For [3H]valerenic acid binding, crude rat brain
membranes were washed three times in 50 mM Tris-HCl pH 7.4 and incubated (200
μg protein) with 300 nM [3H]valerenic acid in a total volume of 0.2 ml for 45 min at
room temperature. Subsequently, the reactions were filtered though glass microfiber
filters and washed with ice cold buffer (50 mM Tris-HCl pH 7.4). Filters were then
processed for liquid scintillation counting using a Tricarb 2500 liquid scintillation
analyzer. Non-specific binding was assessed in parallel by coincubation with 100 μM
unlabelled valerenic acid. Binding of [3H]valerenic acid (300 nM) to well-washed
membranes increased linearly with the protein concentration up to 1.5 mg/ml.
Equilibrium of binding was reached after 20 min at room temperature (measured at
50 and 300 nM [3H]valerenic acid and 200 μg protein). Saturation binding
experiments were performed with 3 - 600 nM [3H]valerenic acid and 200 μg protein
per assay. Binding data were analyzed using the programs Kell (Biosoft, UK) and
GraphPad Prism 4.0 (GraphPad Software, USA).
Crude membranes prepared from rat brain tissue were washed once with 10 mM
Tris-HCl pH 8, 150 mM NaCl, 1 mM EDTA, 200 mg/l bacitracin, 0.1 mM
phenylmethylsulfonyl fluoride, 2.3 mg/l aprotinin, 1 mM benzamidine and
resuspended in the same buffer to a protein concentration of 5 mg/ml followed by
addition of sodium deoxycholate to a final concentration of 0.5%. After incubation for
30 min at 4°C, insoluble material was removed by centrifugation for 30 min at
100,000 g. For immunoprecipitation of GABAA receptors, aliquots (0.5 ml) of the
deoxycholate extract were incubated with α1-subunit-selective antiserum overnight at
4°C (Benke et al. 1991). Receptor-antibody complexes were precipitated by
incubation with 50 μl of protein A-agarose for 60 min. After extensive washing, the
precipitates were resuspended in 10 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM
EDTA, 200 mg/l bacitracin, 0.1 mM phenylmethylsulfonyl fluoride, 2.3 mg/l aprotinin,
1 mM benzamidine, 0.2% Triton X-100 and subjected to radioligand binding
experiments using 300 nM [3H]valerenic acid (see above), 6 nM [3H]flumazenil and
10 nM [3H]Ro 15-4513 as described previously (Benke et al., 1991).
2.6. cDNAs and site-directed mutagenesis
Rat α1, α2, α3, α4, α5, β2, β3 and γ2s GABAA receptor subunit cDNAs were used and
transferred in human embryonic kidney cells (HEK) as described previously (Siegwart
et al. 2002). For the β3 subunit, site-directed mutagenesis substituting an asparagine
residue at position 265 with a methionine residue, β2(N265M) and β3(N265M) was
perfomed as described (Siegwart et al., 2002; 2003).
Current recordings from recombinant GABAA receptors in HEK 293 cells were
performed using the whole cell patch-clamp technique as described previously
(Benson et al., 1998). Primary cultures of hippocampal neurons were prepared from
wild type E18 rat brain (van Rijnsoever et al., 2005) and whole cell voltage-clamp
recordings were made using standard techniques (Hamill et al., 1981). Briefly:
Cultures were continuously superfused with ACSF containing (in mM: 125 NaCl, 26
NaHCO3, 1.25 NaH2PO4, 2.5 KCl, 1 MgCl2, 2.5 CaCl2, and 11 glucose, oxygenated
with 95% O2-5% CO2), TTX was added to block Na channels and APV (100 uM) and
NBQX (20 uM) were added to isolate miniature inhibitory postsynaptic currents
(mIPSCs). Whole cell voltage-clamp recordings of spontaneous mIPSCs from
pyramidal cells were obtained at room temperature with a holding potential of –60
mV and a high chloride containing internal solution (in mM: 100 CsCl, 2 MgCl2, 1
EGTA, 2 ATP, 0.3 GTP, and 40 HEPES, pH 7,2, 300 mOsm).
2.8. Mutant mice
The generation and breeding of homozygous mutant mice containing a (N265M)
mutation in the GABAA receptor β3 subunit was performed as described previously
(Jurd et al., 2003).
2.9. Behavioral tests
The anxiolytic drug action was tested in the light/dark choice test and in the elevated
plus maze test as described (Löw et al., 2000; Rudolph et al., 1999). The animals
were maintained on a 12:12 h reversed light–dark cycle, with ad libitum food and
water. All manipulations described here had been approved by the Cantonal
Veterinary Office of Zurich; they conformed to the ethical standards required by the
Swiss Act and Ordinance on Animal Protection and the European Council Directive
Data are expressed as mean ± SEM or mean ± SD as indicated in the legend to the
figure. If required, data were analyzed for statistical significance using the test
indicated in the respective figure legend. P values less than 0.05 were considered
3.1. Allosteric interaction of valerenic acid with the GABA and benzodiazepine
Using radioligand binding, we aimed at identifying the binding site for valerenic acid.
In a first set of experiments the potential interaction of valerenic acid with the
benzodiazepine and GABA binding site was analyzed. [3H]Flumazenil,
[3H]flunitrazepam and [3H]muscimol binding to crude rat brain membranes was dose-
dependently increased by valerenic acid to a maximum of 200-500% with EC50
values of 23 ± 1 μM, 7 ± 1 μM and 42 ± 6 μM, respectively (Fig. 1). Valerenol also
potentiated benzodiazepine binding to GABAA receptors with an EC50 of 16 ± 7 μM
as tested in the [3H]flunitrazepam binding assay (not shown). These results indicate
that valerenic acid and valerenol allosterically interact with the benzodiazepine and
GABA binding sites of GABAA receptors.
3.2. [3H]Valerenic acid binding
To directly assess the valerenic acid binding site of GABAA receptors, valerenic acid
was radiolabeled by random catalytic exchange of hydrogen by tritium. [3H]Valerenic
acid binding to brain membranes revealed both a high affinity binding site (KD = 25 ±
20 nM) and a low affinity site (KD = 16 ± 10 μM). [3H]Valerenic acid was displaced not
only by valerenic acid but also by valerenol with even higher potency (IC50 = 3 ± 2
nM, Tab. 1). A 1000 fold lower displacing potency was found for hydroxy-valerenic
acid whereas O-methyl valerenol, valerenal and valeric acid (each at 100 μM) were
completely inactive (Tab. 1).
To verify that GABAA receptors indeed harbour the binding site for [3H]valerenic acid,
GABAA receptors were immunoprecipitated from deoxycholate extracts of rat brain
membranes using an antibody directed against the α1 subunit. Radioligand binding
to the well-washed immunoprecipitates demonstrated that the valerenic acid binding
site was coimmunoprecipitated with the classical benzodiazepine sites as
demonstrated by the specific [3H]flumazenil, [3H]Ro 15-4513 and [3H]valerenic acid
binding to the precipitate (Fig. 2). Thus, high affinity [3H]valerenic acid binding sites
are located on GABAA receptors.
To test whether the [3H]valerenic acid binding site corresponds to any of the well
established modulatory binding sites of the GABAA receptor, competition experiments
were performed. Ligands for the GABA (THIP, SR95531), benzodiazepine
(flunitrazepam), barbiturate (phenobarbital), channel blocker (picrotoxinin) and the
loreclezole site (each up to 100 μM) did not affect [3H]valerenic acid binding (not
shown). However, [3H]valerenic acid binding was enhanced in the presence of
propofol (680 ± 126%), etomidate (190 ± 21%) and alphaxolon (200 ± 19%),
suggesting an allosteric interaction with sites for these anaesthetics (Tab. 2). The
anti-inflammatory agent mefenamic acid, which shares structural similarities with
loreclezole and etomidate (Halliwell et al., 1999), was found to inhibit [3H]valerenic
acid binding with an IC50 of 7 ± 3 μM. These results suggest that valerenic acid may
bind to a previously unrecognized site on GABAA receptors that is allosterically linked
to sites for anaesthetics and mefenamic acid.
3.3. Effect of valerenic acid of recombinant receptors expressed in HEK 293 cells
Valerenic acid was recently shown to enhance the GABA-induced chloride currents
of recombinant GABAA receptors expressed in Xenopus oocytes (Khom et al. 2007).
Here we used a different expression system (HEK 293 cells) to confirm a positive
allosteric potentiation of valerenic acid on recombinant GABAA receptors. On GABAA
receptors expressed in HEK 293 cells, valerenic acid strongly enhanced a
submaximal GABA response (EC10 at 3 μM) as demonstrated for GABAA receptors
comprising the subunits α1β2γ2 (EC50 = 5.6 ± 0.6 μM), α2β2γ2 (11.8 ± 0.9 μM), α3β2γ2
(8.1 ± 1.3 μM) and α5β2γ2 (8.6 ± 1.8 μM) with similar Hill coefficients (Hill 1.7 ± 0.2;
1.6 ± 0.2; 1.7 ± 0.4; 1.7 ± 0.5 respectively) (Fig. 3A). The GABA response at α2β3γ2
was likewise enhanced by valerenic acid as tested at 10 μM (Fig. 3D). Valerenic acid
(10 μM) enhanced the GABA response also at α4β2γ2 receptors (350 ± 50 % of
control) while diazepam but not bretazenil (Benson et al. 1998) was inactive (Fig.
In the absence of GABA, there was no direct activation of GABAA receptors by 10 μM
valerenic acid as tested at α1β2γ2 receptors (not shown). However, at high
concentrations of valerenic acid (≥ 30 μM ) a direct activation of α1β2γ2 receptors was
previously observed (Khom et al., 2007).
At α1β2 GABAA receptors, valerenic acid (up to 50 μM) failed to enhance the GABA
response, while alphaxolone and propofol (each at 10 μM) retained their potentiating
activity (Fig. 3C).
Valerenol, a congener of valerenic acid, likewise enhanced the GABA response as
tested at α2β3γ2 receptors (Fig. 3D). However, structurally different compounds such
as valeric acid, valeric anhydride, valeronitrile, δ-valerolactame and γ- or δ-
valerolactone (each at 10 μM) failed to alter the GABA response as measured at
α1β2γ2 receptors (not shown).
Asparagine 265 in the β2 subunit was suggested to play an important role for the
potentiating effect of valerenic acid based on its mutation to serine (Khom et al.,
2007). We used β2 and β3 subunit cDNAs in which the asparagine residue 265 was
mutated to methionine (Siegwart et al., 2002). When recombinant GABAA receptors