SEARCH

SEARCH BY CITATION

Keywords:

  • flavonoids;
  • flavonoid glycosides;
  • G protein-coupled inwardly rectifying potassium channels;
  • G protein-coupled receptor;
  • naringin;
  • tertiapin-Q

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

BACKGROUND G protein-coupled inwardly rectifying potassium (KIR3) channels are important proteins that regulate numerous physiological processes including excitatory responses in the CNS and the control of heart rate. Flavonoids have been shown to have significant health benefits and are a diverse source of compounds for identifying agents with novel mechanisms of action.

EXPERIMENTAL APPROACH The flavonoid glycoside, naringin, was evaluated on recombinant human KIR3.1–3.4 and KIR3.1–3.2 expressed in Xenopus oocytes using two-electrode voltage clamp methods. In addition, we evaluated the activity of naringin alone and in the presence of the KIR3 channel blocker tertiapin-Q (0.5 nM, 1 nM and 3 nM) at recombinant KIR3.1–3.4 channels. Site-directed mutagenesis was used to identify amino acids within the M1–M2 loop of the KIR3.1F137S mutant channel important for naringin's activity.

KEY RESULTS Naringin (100 µM) had minimal effect on uninjected oocytes but activated KIR3.1–3.4 and KIR3.1–3.2 channels. The activation by naringin of KIR3.1–3.4 channels was inhibited by tertiapin-Q in a competitive manner. An alanine-scan performed on the KIR3.1F137S mutant channel, replacing one by one aromatic amino acids within the M1–M2 loop, identified tyrosines 148 and 150 to be significantly contributing to the affinity of naringin as these mutations reduced the activity of naringin by 20- and 40-fold respectively.

CONCLUSIONS AND IMPLICATIONS These results show that naringin is a direct activator of KIR3 channels and that tertiapin-Q shares an overlapping binding site on the KIR3.1–3.4. This is the first example of a ligand that activates KIR3 channels by binding to the extracellular M1–M2 linker of the channel.


Abbreviations
CGP36742 or SGS742

3-aminopropyl-n-butylphosphinic acid

GPCRs

G protein-coupled receptors

KIR3

G protein-coupled inwardly rectifying potassium channels

LPA

lysophosphatidic acid

PIP2

phosphatidylinositol 4,5-bisphosphate

r KIR1.1

rat renal outer medullary potassium

TPN-Q

tertiapin-Q

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

G protein-coupled inwardly rectifying potassium channels (KIR3/GIRK) are members of a family of inwardly rectifying potassium channels. These channels are activated by G protein-coupled receptors (GPCRs) such as opioid, adenosine, muscarinic, GABAB and dopamine receptors (Ikeda et al., 1995; 1996; 1997; Luscher et al., 1997). Four subunits have been identified termed KIR3.1–3.4 (Lesage et al., 1994). In general, heterotetrameric channels are formed by KIR3.1 coupling to either KIR3.2 or KIR3.4 subunits resulting in the formation of KIR3.1–3.2 channels (present mainly in brain regions) (Kobayashi et al., 1995; Lesage et al., 1995) and KIR3.1–3.4 channels (present mainly in the heart and endocrine regions) (Corey and Clapham, 1998; Gregerson et al., 2001). Such channels play an important role in regulating neuronal excitability and heart rate (Signorini et al., 1997; Wickman et al., 1998).

The activation of KIR3 channels involves many intrinsic factors including Mg2+ and Na+, pH, phosphatidylinositol 4,5-bisphosphate (PIP2) and intracellular proteins such as Gi/o proteins (Nichols and Lopatin, 1997; Mark and Herlitze, 2000). Hormones, transmitters and peptides indirectly activate KIR3 channels by preferentially stimulating Pertussis toxin-sensitive GPCRs (Sadja et al., 2003) causing the Gα subunit to replace its bound GDP with GTP. This causes the Gβγ subunit to dissociate from the Gα subunit, which in turn directly binds and activates the GIRK channel at several intracellular sites on the N- and C-termini (Lesage et al., 1994; Nishida and MacKinnon, 2002). The activation of KIR3 channels by the Gβγ subunit in turn stabilizes PIP2 interactions (Sui et al., 1998; Logothetis and Zhang, 1999) and is accelerated by some members of the regulators of G protein signalling family of proteins (Dascal, 1997; Mark and Herlitze, 2000; Sadja et al., 2003). Finally the activation signal is terminated when Gα-GTP is hydrolysed back to Gα-GDP, which enables the Gβγ subunit to associate and reform the Gi/o trimer. Some members of the regulators of G protein signalling family of proteins also contribute to this process, paradoxically also accelerating the deactivation kinetics (Benians et al., 2005).

A number of direct inhibitors of KIR3 channels have been reported including some opioids (Ulens et al., 1999), antipsychotics and antidepressants (Kobayashi et al., 2000; 2003; 2004), and the bee toxin, tertiapin-Q (Jin et al., 1999). In contrast, only ethanol (Kobayashi et al., 1999; Lewohl et al., 1999) and halothane (Weigl and Schreibmayer, 2001; Yamakura et al., 2001) have been reported to directly activate the channels.

Flavonoids are natural polyphenolic agents found in all plants (Mattila et al., 2000). They are secondary metabolites, which are consumed in significant amounts from beverages, fruits and vegetables. Flavonoids have been shown to have significant health benefits (Birt et al., 2001) being involved in a variety of biological processes. They have been shown to reduce heart disease (Renaud and de Lorgeril, 1992), be protective against cancers (Kandaswami et al., 1991; Hertog et al., 1995) and neurodegenerative diseases such as Alzheimer's disease (Mandel and Youdim, 2004). Thus, flavonoids are a diverse source of compounds for identifying agents with novel mechanisms of action.

Naringin and its aglycosylated analogue (±)-naringenin are bioflavanoids found in grapefruit. Both agents have a chiral centre at the C2 position of the middle ring (Figure 1) resulting in two stereoisomers. Both enantiomers are found naturally in the fruit and contribute to the fruits' bitter flavour. In this study, we evaluated the effects of naringin (racaemic mixture) and (±)-naringenin on recombinant wild-type and mutant KIR3 channels expressed in Xenopus oocytes. The data show that naringin but not (±)-naringenin activates KIR3 channels. The activation is inhibited by low nanomolar concentrations of tertiapin-Q in a competitive manner indicating a common or overlapping binding site on the KIR3 channel. Site-directed mutagenesis studies replacing aromatic amino acids [phenylalanine (Phe) and tyrosine (Tyr)] located on the extracellular linker that connects the intracellular transmembrane domains 1 and 2 (M1–M2 linker) of the homotetrameric KIR3.1F137S mutant channel to alanine (Ala) identified Tyr148 and Tyr150 as important for naringin's affinity. This is the first report of a molecule that binds to the extracellular vestibule of the KIR3 channel and triggers channel opening.

image

Figure 1. Structures of naringin and naringenin.

Download figure to PowerPoint

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

Materials

Naringin (±)-naringenin, barium chloride (BaCl2), adenosine, theophylline and LPA were obtained from Sigma, Australia. Neohesperidose was obtained from ABCR GmbH& Co. KG, Karlsruhe, Germany. 3-Aminopropyl-n-butylphosphinic acid (CGP36742 or SGS742) was a gift from Dr Wolfgang Froestl (formerly Novartis, Switzerland).

Rat KIR3.4, human GABAB(1b) and GABAB2 subcloned in pcDNA3.1(−), and rat KIR3.1 subcloned in pBluescript were gifts from Drs Fiona Marshall and Andrew Green (Glaxo Wellcome, UK). Human KIR3.1 and KIR3.2 subcloned into pCMV6-XL5 and pCMV6-XL4 were obtained from Origene Technologies, Inc., MD, USA.

Molecular biology

Human GABAB(1b) and GABAB2 cDNAs were linearized using EcoRI. Human KIR3.1, rat KIR3.1 and KIR3.4 cDNAs were linearized with XbaI. Human KIR3.2 cDNA was linearized with SmaI and SacI and the DNA extracted via QIAquick gel extraction kit (QIAGEN Pty Ltd, Australia). mRNAs were transcribed in vitro using T7 mMessage mMachine transcription kit (Ambion Inc., Austin, TX, USA) for all linearized cDNAs.

For KIR3.1–3.4 and KIR3.1–3.2 channel expression, the mRNA ratio injected was 1:1 with a total RNA concentration of 100 and 80 ng per oocyte respectively. For GABAB-KIR3 receptor expression, the mRNA ratios used for GABAB(1b) : GABAB2 : KIR3.1 : KIR3.4 were 1:2:1:1 with a total RNA concentration of 80 ng per oocyte. For GABAB-KIR3 mutant expression, the mRNA ratio was 1:2:1 [GABAB(1b) : GABAB2 : mutant] and the RNA concentration ranged between 40 and 130 ng per oocyte. For expressing KIR3 mutations alone, the final concentration used was 25 ng per oocyte.

Electrophysiological recording

Animal ethics approval was granted by The University of Sydney Animal Ethics Committee and followed the NH&MRC guidelines on the Australian code of practice for the care and use of animals for scientific research (L24/2–2006/3/4267). In brief, female Xenopus laevis were anaesthetized with tricaine (850 mg·500 mL−1). Several ovarian lobes were surgically removed by a small incision on the abdomen of the X. laevis. The lobes were cut into small pieces and were rinsed thoroughly with oocyte releasing buffer 2 [OR2; 82.5 mM NaCl, 2 mM KCl, 1 mM MgCl2, 5 mM HEPES (hemi-Na)]. The lobes were digested with collagenase A (2 mg·mL−1 in OR2; Boehringer Manheim, Germany) at room temperature. The oocytes were further washed with OR2 and stored in ND96 wash solution [96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 5 mM HEPES (hemi-Na)] supplemented with 2.5 mM sodium pyruvate and 0.5 mM theophylline until ready for injection. Stage V–VI oocytes were selected and microinjected with 50.6 nL of mRNA. After injection, the oocytes were maintained at 18°C in the presence of ND96 wash solution augmented with 2.5 mM sodium pyruvate, 0.5 mM theophylline and gentamicin at 50 µg·mL−1.

Naringin was stored at 500 mM stock in DMSO and (±)-naringenin was stored as 100 mM in DMSO at −20°C. Prior to recording, the compounds were diluted with either: (i) 45 mM K+ buffer: 45 mM NaCl, 45 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 5 mM HEPES (hemi-Na salt); or (ii) 90 mM K+ buffer: 90 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 5 mM HEPES (hemi-Na salt) to the desired concentration with a final concentration of DMSO equal to 0.8%.

Whole-cell currents were measured using a two-electrode voltage clamp set-up composed of a Digidata 1200, Geneclamp 500B amplifier and pClamp 8 (Axon Instruments Inc., Foster City, CA, USA), together with a Powerlab/200 (AD Instruments, Sydney, Australia) and Chart version 5.5 program for PC. Voltage was maintained at −60 mV. For current–voltage analysis, the holding potential was −30 mV and currents were measured from −80 mV to 60 mV in 10 mV increments in response to 100 ms voltage steps. This measurement was performed after a 2 min perfusion of each of the following buffers: ND96, 45 mM K+, 90 mM K+, naringin (100 µM) in 45 and 90 mM K+ buffers for GIRK1/4 and ND96, 45 mM K+, naringin (100 µM) in 45 mM K+ buffer for KIR3.1–3.2 and mutant channels. Traces recorded in ND96 were subtracted offline from traces recorded under the various buffers in order to correct for leak and endogenous oocyte currents.

The recording microelectrodes were filled with 3 M KCl [or 3 M KCl and 11 mM ethylene glycol tetraacetic acid (EGTA)] and the resistance was between 0.2 and 1.0 mΩ. Two to five days post injection, oocytes were used for recording. Oocytes were initially superfused with ND96 until a stable base current was achieved at which point the buffer was switched to a high K+ buffer (45 or 90 mM K+). Compounds were only applied once a stable base current was reached. A 6 min wash period was applied between drug applications.

Evaluating the effect of naringin at intracellular sites of GIRK1/4 channels

Oocytes expressing KIR3.1–3.4 and responding to extracellularly applied naringin (100 µM) or ethanol (100 mM) were further injected with 50.6 nL of water (control), ethanol (2 M) (positive control) or naringin (2 mM) (note: as the volume of a single oocyte is ∼1 µL, the final intracellular concentration of ethanol and naringin were ∼100 mM and ∼100 µM respectively). The effects on the whole-cell currents were monitored for up to 40 min.

Binding studies

Naringin (10 µM) was tested on a series of primary binding assays performed by the National Institute of Mental Health's Psychoactive Drug Screening Program, Contract # NO1MH32004 (NIMH PDSP) directed by Bryan L. Roth MD, PhD at the University of North Carolina at Chapel Hill and Project Officer Jamie Driscol at NIMH, Bethesda MD, USA. For experimental details please refer to the PDSP web site http://pdsp.med.unc.edu/.

Site-directed mutagenesis studies

Sense and antisense oligonucleotide primers (Table S1) were designed to introduce the point mutations within the KIR3.1 and KIR3.4 subunits. Phenylalanine (Phe; F) 137 of the KIR3.1 subunit was mutated to serine (Ser; S) while Ser143 of the KIR3.4 subunit was mutated to threonine (Thr; T) in order to form the homotetrameric channels KIR3.1F137S and KIR3.4S143T (Chan et al., 1996). Using the KIR3.1F137S subunit, a series of secondary mutations were performed replacing either tyrosine (Tyr; Y) or Phe in the M1–M2 linker to alanine (Ala; A) using the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). DNA plasmids containing the singly or doubly mutated KIR3.1F137S subunit were transformed into XL1-Blue supercompetent cells (Stratagene, CA, USA) and cultured in LB broth with ampicillin (5 µg·mL−1). Successful mutants were identified by restriction enzyme analysis and verified by DNA sequencing. The mRNA of the single and double mutations were expressed in oocytes alone or in the presence of the GABAB(1b,2) receptor before evaluating with naringin, adenosine or GABA.

Statistical analysis

Data are represented as the mean (±SEM) from a specified number of independent experiments or as mean (95% CI). For the concentration–response curves, data points were fitted using GraphPad Prism 5. The current was normalized to the maximum concentration of agonist in the following ratio (I/Imax) unless otherwise stated. The concentration–response curves were plotted using current ratios (Y-axis) and plotted against log of the concentration (X-axis) and fitted to the following formula:

  • image

Where, I= current response, Imax= maximum current, nH= Hill slope and [A]= agonist concentration.

The binding constant for antagonist (Kb) was estimated using the Schild equation Kb= ({A}/{A*}−1) −[Ant], where {A} is the EC50 of agonist in the presence of antagonist, {A*} is the EC50 of agonist in the absence of antagonist, [Ant] is the concentration of antagonist, and simple competitive antagonism (m = 1) is shown by plotting the log (dose ratio − 1) versus −log [Ant].

Statistical analyses were performed using a one-way anova followed by Tukey's multiple comparison post hoc test when comparing multiple groups (unless otherwise stated) or with a paired Student's t-test when comparing two groups. Statistical probability (P) are expressed as *P < 0.05, **P < 0.01 and ***P < 0.001.

Nomenclature

The nomenclature of all molecular targets (receptors, ion channels, enzymes, etc.) cited in this work conforms to the British Journal of Pharmacology's Guide to Receptors and Channels (Alexander et al., 2009).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

High concentrations of naringin activate endogenous K+ channels and calcium-activated chloride channels in Xenopus oocytes

In the presence of high K+ buffers (45 or 90 mM K+ buffers) to increase the K+ driving force at the holding potential (−30 mV), small basal currents were observed in uninjected oocytes (10–30 nA), indicating the presence of endogenous potassium channels. When naringin (≥100 µM) was added to uninjected cells, a small response (≤10–20 nA) was detected in the presence of 45 mM K+ buffer. As naringin is reported to weakly activate Ca2+-activated K+ channels (BKCa) (Saponara et al., 2006) and Xenopus oocytes are known to express an endogenous homologue of these channels, it was thought that naringin may be activating these channels (Kanjhan et al., 2005). In addition, naringin (≥100 µM) elicited a sharp current indicative of activating the endogenous Ca2+-activated Cl- channel in both ND96 and 45 mM K+ buffer (Figure S1) and the levels of each varied with individual cells. Therefore for subsequent experiments, intracellular Ca2+ was buffered with 11 mM EGTA in the recording microelectrode and was allowed to equilibrate in the oocyte for 15 min prior to recording. In addition, oocytes injected with KIR3 channels were stored at 4°C for 24–48 h prior to use as this treatment helped reduce the expression of the unwanted Ca2+-activated Cl- channels [modified from (Varecka and Peterajova, 1990)]. Any oocytes that responded with a sharp inward current typical of Ca2+-activated Cl- channel activity were discarded.

Naringin but not (±)-naringenin activates KIR3 channels

In oocytes expressing GABAB receptors and KIR3.1–3.4 channels, both GABA (100 µM and 3 µM) and naringin (100 µM) activated the channel (Figure S2A). The effect of GABA (3 µM) but not naringin (100 µM) was blocked by the competitive GABAB/C receptor antagonist, CGP36742 100 µM (Figure S2B), indicating the activation of the KIR3 channel by naringin is not via the GABAB receptor. Furthermore, naringin (100 µM) but not (±)-naringenin (100 µM) or the C7 sugar moiety of naringin, neohesperidose (100 µM) activated oocytes expressing KIR3.1–3.4 and KIR3.1–3.2 channels (Figure S2B & S2C).

Naringin activates KIR3 channels independently of the extracellular K+ concentration

In oocytes expressing only KIR3.1–3.4 channels, upon switching buffer solutions from ND96 to 45 mM K+, basal currents of the order 313 ± 37 nA (n= 71) were generated. Naringin (100 µM) further activated the channel by 240 ± 22 nA (n= 71). In the presence of 90 mM K+ buffer, basal currents increased compared with 45 mM K+ buffer and were of the order 1406 ± 340 nA (n= 17). Naringin (100 µM) further increased the current by 949 ± 120 nA (n= 17). An example of the current produced by naringin (100 µM) in the presence of 45 and 90 mM K+ buffers at KIR3.1–3.4 is shown in Figure 2. The current–voltage relationship in the presence and absence of naringin (100 µM) using 45 and 90 mM K+ buffers at KIR3.1–3.4 is also shown in Figure 2B. Under these conditions, the response of KIR3.1–3.4 to 45 and 90 mM K+ is strongly rectified. Naringin (100 µM) further activates the channel and this effect is voltage-independent.

image

Figure 2. (A) Effect of naringin (100 µM; duration indicated by open bar) on KIR3.1–3.4 channels expressed in Xenopus oocytes in the presence of 45 mM K+ buffer (duration indicated by hatched bar) and 90 mM K+ buffer (duration indicated by filled bar), measured at −60 mV. (B) Current–voltage relationship for naringin-sensitive currents generated from wild-type KIR3.1–3.4 expressed in Xenopus oocytes and measured at 45 and 90 mM K+ buffers. Current responses were measured at a holding potential of −30 mV. Currents were measured in a 10 mV increment from −80 to 60 mV in response to 100 ms voltage steps in presence of ND96 buffer, 45 mM K+ buffer, naringin (100 µM) in the presence of 45 mM K+ buffer, 90 mM K+ buffer and naringin (100 µM) in the presence of 90 mM K+ buffer. Current–voltage relationships in ND96 buffer were subtracted offline from traces recorded in 45 mM K+ buffer, naringin (100 µM) in the presence of 45 mM K+ buffer, 90 mM K+ buffer and naringin (100 µM) in the presence of 90 mM K+ buffer to correct for leak and endogenous oocyte currents. Each voltage point is shown as mean ± SEM of current (µA) from three to six oocytes. (C) Example of an oocyte expressing KIR3.1–3.4. In the presence of 45 mM K+ naringin generated an outward current when clamped at −60 mV in a concentration-dependent manner. (D) Example of an oocyte expressing KIR3.1–3.4. In the presence of 45 mM K+ naringin generated outward currents in a concentration-dependent manner when clamped at −60 mV.

Download figure to PowerPoint

In oocytes expressing only KIR3.1–3.2, the basal currents produced in the presence of 45 and 90 mM K+ buffers were 583 ± 71 nA (n= 50) and 823 ± 127 nA (n= 23) respectively. Naringin (100 µM) further increased the currents by 207 ± 25 nA (n= 38) and 505 ± 194 nA (n= 23) respectively.

Table 1 summarizes the activity of naringin at KIR3.1–3.4 and KIR3.1–3.2 under various conditions. Figure 2C and D shows an example of the responses in an oocyte expressing KIR3.1–3.4 and KIR3.1–3.2 respectively. Naringin dose-dependently activated these channels. The EC50 for naringin at KIR3.1–3.4 and KIR3.1–3.2 channels in the presence of 45 mM K+ were not significantly different (P > 0.05; one way anova, followed by Tukey's multiple comparison test; Table 1). The EC50 values for naringin at KIR3.1–3.4 and KIR3.1–3.2 channels in the presence of 90 mM K+ were also not significantly different (P > 0.05; one way anova, followed by Tukey's multiple comparison test; Table 1). This indicates that the activity of naringin at KIR3.1–3.4 and KIR3.1–3.2 channels is independent of the extracellular K+ concentration. In all cases, the Hill coefficients (nH) did not significantly vary between KIR3.1–3.4 and KIR3.1–3.2 (P > 0.05; Table 1).

Table 1.  Effect of naringin on wild-type KIR3 channels expressed in Xenopus oocytes using two-electrode voltage clamp methods
Channel I −60 mV, 45 K (nA) (n) Naringin EC50[95% CI] (µM) (logEC50± SEM) n H ±SEM n
  • a

    Activity determined using 90 mM K+ buffer.

  • b

    Coinjected with mRNA of GABAB(1b), GABAB2 subunits.

  • Unless otherwise stated, 45 mM K+ was used for recording.

KIR3.1–3.4313 ± 37 (71)120.9 [69.7–209.8] (2.08 ± 0.01)0.93 ± 0.036
1406 ± 340 (17)a71.0a[55.4–91.0] (1.85 ± 0.08)0.84 ± 0.085
91.8b[40.2–209.5] (1.96 ± 0.09)0.86 ± 0.184
KIR3.1F137S551 ± 104 (7)214.7 [81.5–565.7] (2.33 ± 0.01)0.86 ± 0.016
KIR3.4S143T350 ± 124 (11)104 [71–152] (2.02 ± 0.08)1.0 ± 0.26
KIR3.1–3.2583 ± 71 (50)111.0 [50.8–242.3] (2.05 ± 0.08)0.93 ± 0.114
823 ± 127 (23)a58.6a[33.0–105.2] (1.77 ± 0.04)1.03 ± 0.075

The EC50 and nH values for naringin on KIR3.1–3.4 in the presence and absence of GABAB(1b,2) receptors were not significantly different (P > 0.05; one way anova, followed by Tukey's multiple comparison test; Table 1) further indicating that the activation by naringin is not via the GABAB receptor. Thus, naringin activates KIR3 channels independently of the extracellular K+ concentration and with similar potency at KIR3.1–3.4 and KIR3.1–3.2 channels.

Naringin activates KIR3 channels at an extracellular site

In order to determine whether naringin activated KIR3 channels by an extracellular or intracellular site, oocytes expressing KIR3.1–3.4 channels were first activated by extracellular application of naringin (100 µM). Then naringin (50.6 nL of a 2 mM stock) or water (50.6 nL) was injected into the oocyte. Neither water nor naringin generated a current under these conditions even after a 40 min recording period (Figure 3A and C). In contrast, oocytes expressing KIR3.1–3.4 were activated by extracellular and intracellular application of ethanol (100 mM; Figure 3B). Figure 3B shows a sustained activation by ethanol. Taken together the data indicate that naringin activates KIR3.1–3.4 channels only at an extracellular site.

image

Figure 3. Effects of naringin (100 µM; filled bar) and ethanol (100 mM) when applied extracellularly to Xenopus oocytes expressing KIR3.1–3.4 channels in the presence of 45 mM K+ buffer (open bar). When either (A) water (50.6 nL; filled bar), (B) ethanol (100 mM–50.6 nL of 2 mM stock solution; filled bar) or (C) naringin (50.6 nL of 2 mM stock solution; filled bar) is injected into the cytoplasm of the oocyte, only ethanol produced a sustained response. No response was observed to water or naringin, indicating that naringin does not act at an intracellular site.

Download figure to PowerPoint

The effect of naringin is inhibited by barium

Ba2+, a non-specific KIR channel blocker, significantly inhibited the effect of naringin (100 µM) on KIR3.1–3.4 channels in a concentration-dependent manner (Figure 4). Figure 4A shows an example of a trace where Ba2+ (100 µM and 3 mM) blocked 80% and 100% the effect of naringin respectively (P= 0.0001; Student's t-test; Figure 4B).

image

Figure 4. (A) Naringin (100 µM; open bar) activates oocytes expressing KIR3.1–3.4 in the presence of 45 mM K+ buffer (open bar). BaCl2 (100 µM and 3 mM; filled bars) inhibited the effect of naringin (100 µM; open bar). (B) Histogram indicating the inhibition by 100 µM and 3 mM BaCl2, respectively, in the presence of naringin (100 µM) alone. Each column represents the mean (±SEM) of four oocytes from at least two harvests. Statistical significance is as indicated: ***P < 0.001.

Download figure to PowerPoint

The activation of KIR3 channels by naringin is not via an endogenous GPCR

To determine whether naringin stimulated an endogenous receptor, we evaluated the binding of naringin to a range of known GPCR/ion channels. Naringin did not bind to any GPCR/ion channel tested to date (Table S2).

In addition to the binding studies, naringin was evaluated on endogenous adenosine receptors known to express in the Xenopus oocyte (Kobayashi et al., 2002). In oocytes expressing KIR3.1–3.4 channels, adenosine produced a concentration-dependent response by activating KIR3.1–3.4 channels via the endogenous receptor. The EC50 value obtained for adenosine was 2.7 µM (95% CI = 2.36–3.0) and the nH was 1.57 ± 0.07 (n= 3; Table 4). In order to determine whether naringin activated this receptor, theophylline (100 µM), a non-selective adenosine receptor antagonist was used. Theophylline (100 µM) blocked the effect of adenosine (1 µM) but did not block the effect of a low concentration of naringin (10 µM; n= 3; Figure S3), indicating that the effects of naringin on recombinant KIR3.1–3.4 channels are not via the endogenous adenosine receptor.

Table 4.  Effect of stimulating wild-type and mutant KIR3 channels expressed in Xenopus oocytes by adenosine and LPA through their endogenously expressed receptor
Channel Adenosine LPA
EC50[95% CI] (µM) (logEC50± SEM) n H ±SEM EC50[95% CI] (nM) (logEC50± SEM) n H ±SEM
KIR3.1–3.42.66 [2.36–3.0] (0.42 ± 0.01)1.57 ± 0.01416 [360–480] (−6.38 ± 0.03)2.15 ± 0.27
KIR3.1F137SY148A3.60 [2.7–6.3] (0.56 ± 0.02)1.09 ± 0.05329 [175–618] (−6.48 ± 0.13)1.20 ± 0.4
KIR3.1F137SY150A3.10 [1.8–5.4] (0.49 ± 0.06)1.00 ± 0.12

Tertiapin-Q competitively inhibited naringin indicating an overlapping binding site

Tertiapin-Q is a potent KIR1.1 and KIR3 channel blocker (Jin and Lu, 1998; Kanjhan et al., 2005), which binds to the external vestibule of the K+-conduction pore that is formed by the linker between M1 and M2 segments (Ramu et al., 2004) to inhibit the basal current. To determine whether tertiapin-Q inhibits naringin in a competitive or non-competitive manner, we evaluated the effect of naringin alone and in the presence of three concentrations of tertiapin-Q (0.5 nM, 1 nM and 3 nM). We chose low concentrations of tertiapin-Q because we wanted to have maximal inhibition of naringin-stimulated response and minimal basal inhibition. We also compared the effect of GABA alone and in the presence of the maximum concentration of tertiapin-Q (3 nM). Figure 5A and B shows the concentration–response curves for naringin and GABA respectively. Tertiapin-Q shifted the concentration–response curves for naringin to the right (Figure 5A). In addition the Schild slope did not significantly deviate from 1, indicating that tertiapin-Q competitively blocks naringin at KIR3.1–3.4 channels. The Kb (binding constant) for tertiapin-Q was found to be 184 ± 23 pM, 20- to 25-fold more potent than previously reported (Jin and Lu, 1998; Kanjhan et al., 2005). These data infer that tertiapin-Q and naringin share a common or overlapping binding site.

image

Figure 5. (A) Concentration–response curves for naringin alone, and in the presence of 0.5 nM, 1 nM and 3 nM tertiapin-Q at KIR3.1–3.4 channels expressed in oocytes. Tertiapin-Q shifted the EC50 of naringin in a parallel manner. The inset shows the curve of log [dose ratio (DR) − 1] versus −log [Ant]. The slope of −1.32 ± 0.25 was not significantly different from 1, indicating competitive antagonism. (B) Concentration–response curve for GABA alone and in the presence of 3 nM tertiapin-Q at GABAB(1b,2) coexpressed with KIR3.1–3.4 channels in oocytes. Tertiapin-Q had no effect on the EC50 of GABA but started to inhibit the maximal response, indicating non-competitive inhibition. (C) Example trace of an oocyte expressing GABAB(1b,2) and KIR3.1–3.4 channels. High K+ buffer (45 mM K+) stimulates a basal current, which, in the presence of GABA (100 µM), is further activated. Tertiapin-Q (1 µM) inhibits the basal current by 80%. When tertiapin-Q (1 µM) is applied with GABA (100 µM), the response to GABA is reduced by 90%. Data points are expressed as mean ± SEM (n= 4 oocytes from at least two harvests). (D) Effect of GABA (100 µM) in the presence of various concentrations of tertiapin-Q. Tertiapin-Q reduces the GABA (100 µM) response in a concentration-dependent manner. The concentrations of tertiapin-Q required to inhibit GABA (100 µM) responses were at least 10-fold higher than those inhibiting naringin response. Each column represents the mean (±SEM) of four oocytes from at least two harvests. Statistical significance is as indicated: **P < 0.001.

Download figure to PowerPoint

Tertiapin-Q (3 nM) did not shift the concentration–response curve for GABA (Figure 5B). Instead tertiapin-Q inhibited the maximum response to GABA 100 µM (Figure 5C and D). Figure 5C shows a trace of GABA stimulating recombinant GABAB(1b,2) and KIR3.1–3.4 channels. When tertiapin-Q (1 µM) was added, the basal response was dramatically reduced. In the presence of tertiapin-Q (1 µM), GABA (100 µM) exerted only 10% of the original response. Figure 5D shows a concentration-dependent inhibition of the GABA (100 µM) response by tertiapin-Q (0, 3, 10 and 100 nM). The data indicate that tertiapin-Q inhibits GABA in a non-competitive manner.

Effects of naringin and tertiapin-Q on homotetrameric KIR3.4S143T and KIR3.1F137S mutant channels

If naringin and tertiapin-Q share a common or overlapping binding site, then naringin must bind to one or both KIR3 subunit subtypes that make up the KIR3.1–3.4 tetramer. In order to determine which subunit(s) naringin binds to, we evaluated the effects of naringin on the homotetrameric KIR3.4S143T and KIR3.1F137S mutant channels. Phe137 of KIR3.1 is a site for synergy interactions with KIR3.2–3.4. Mutating this position to serine, the corresponding amino acid in KIR3.4 enables translocation of the protein to the cell surface and facilitates homomeric assembly of the channel (Chan et al., 1996). This feature simplifies the study by allowing us to evaluate the effect of naringin on one subunit without the effect of the other. In contrast, KIR3.4 channels can form homotetramers but the currents are much smaller and the mean open time is too fast. Thus by mutating serine 143 to threonine, the resulting mutant channel, KIR3.4S143T, produces more robust and sustained currents for the study (Chan et al., 1996).

The basal current of KIR3.1F137S in 45 mM K+ was 551 ± 104 nA (n= 7) (Table 1) and naringin further evoked a current of 128 ± 25 nA (n= 6). The EC50 and nH for naringin at KIR3.1F137S (Table 1) were not significantly different from wild-type KIR3.1–3.4 channels (P > 0.05). The basal current of KIR3.4S143T in 45 mM K+ was 350 ± 124 nA (n= 6) and naringin further evoked currents of 400 ± 120 nA (n= 6) (Table 1). The EC50 and nH for naringin at KIR3.4S143T were not significantly different from wild-type KIR3.1–3.4 channels (P > 0.05; Table 1). Collectively, the data indicate that naringin activates both KIR3.1F137S and KIR3.4S143T subunits with similar potencies compared with wild-type KIR3.1–3.4 channels (Figure 6A; Table 1).

image

Figure 6. (A) Concentration–response curves for naringin on KIR3.1–3.4, KIR3.4S143T and KIR3.1F137S channels expressed in oocytes. No statistical differences were detected in the EC50 or nH for naringin in these channels. (B) A representative trace showing the basal current generated by 45 mM K+ buffer on oocytes expressing KIR3.1F137S. Ba2+ (1 mM) blocked 90% of the basal current while tertiapin-Q could not block the currents unless the concentration was >1 µM. (C) Concentration–response curves for naringin alone and in the presence of 0.5 nM and 3 nM tertiapin-Q on oocytes expressing KIR3.1F137S channels. Tertiapin-Q could not shift the concentration–response curves of naringin to the right nor could it inhibit the maximal currents indicating naringin is not blocked by tertiapin-Q at KIR3.1F137S channels.

Download figure to PowerPoint

In contrast, the effect of tertiapin-Q on KIR3.1F137S was dramatically reduced compared with wild-type KIR3.1–3.4 channels. The concentration of tertiapin-Q required to significantly inhibit basal currents generated by the KIR3.1F137S mutant channel was ≥3 µM. Figure 6B is an example of a trace where 45 mM K+ buffer generated a basal response. Ba2+ (1 mM) completely inhibited the basal response but tertiapin-Q (3 nM–3 µM) had little effect. Furthermore, tertiapin-Q (0.5 nM and 3 nM) did not shift the concentration–response curve for naringin on the KIR3.1F137S mutant (Figure 6C). These data indicate that tertiapin-Q does not potently bind to KIR3.1F137S channels and most likely binds with high potency to the KIR3.4 subunit. Although it is possible that the F137S mutation alters the binding site of tertiapin-Q but Ramu and colleagues (Ramu et al., 2004) showed that the whole P-loop of KIR3.1 is insensitive to tertiapin-Q.

Effects of naringin, GABA, adenosine and LPA on functional KIR3.1F137S double mutants

Nine additional mutations were generated from the KIR3.1F137S mutation: KIR3.1F137SF107A, KIR3.1F137SY120A, KIR3.1F137SY128A, KIR3.1F137SY130A, KIR3.1F137SF134A, KIR3.1F137SF136A, KIR3.1F137SY146A, KIR3.1F137SY148A and KIR3.1F137SY150A. These were evaluated in order to identify amino acids important for naringin activity but not GABA activity. By identifying amino acids important for naringin activity on the KIR3 channel will further support our hypothesis that naringin directly activates KIR3 channels. Tables 2 and 3 summarize the data.

Table 2.  Effect of naringin on KIR3.1F137S mutants expressed in Xenopus oocytes
Channel I −60 mV, 45 K (nA) (n) Naringin (100 µM) (n) EC50[95% CI] (µM) (logEC50± SEM) n H ±SEM
  • a

    Coinjected with GABAB(1b), GABAB2 subunits.

  • Concentration–response curve was determined by extrapolating all the available data using Prism 5.0.

  • ***

    P < 0.001,

  • **

    P < 0.01: comparison with KIR3.1–3.4 logEC50 value (one way anova followed with Tukey's multiple comparison test).

  • ###

    P < 0.001,

  • ##

    P < 0.01: comparison with KIR3.1–3.4 nH value (one way anova followed with Tukey's multiple comparison test).

  • n.i., no ionic currents detected by naringin (100 µM).

KIR3.1F137SY107A263 ± 7 (12)115.5 ± 18 (12)148.4 [23.8–924.5] (2.17 ± 0.15)0.58 ± 0.08
KIR3.1F137SY120A161 ± 17 (19)155.5 ± 25 (19)84.8 [53.3–134.9] (1.93 ± 0.008)0.90 ± 0.01
KIR3.1F137SY128A123 ± 9 (17)64 ± 8 (17)107.8 [46.6–249.4] (2.03 ± 0.03)1.02 ± 0.063
KIR3.1F137SF130A36 ± 4 (4)n.i. (4)
KIR3.1F137SF134A39 ± 3 (7)n.i. (6)
KIR3.1F137SF136A56 ± 10 (9)n.i. (9)
KIR3.1F137SY146A25.0 ± 3.5 (5)n.i. (4)
KIR3.1F137SY148A181 ± 30 (19)66 ± 12 (17)4665 (3.67 ± 0.31)***0.35 ± 0.03###
185 ± 16 (15)a142 ± 34 (9)4280a† (3.63 ± 0.56)***0.37 ± 0.06###
KIR3.1F137SY150A62 ± 5 (16)27.5 ± 2.6 (16)2849 (3.46 ± 0.21)**0.46 ± 0.04###
45 ± 4 (20)a35.6 ± 10.9 (13)2833a† (3.45 ± 0.10)**0.51 ± 0.02##
Table 3.  Effect of GABA on GABAB(1b,2) receptors coexpressed with wild-type and mutant KIR3 channels in Xenopus oocytes
Channel I −60 mV, 45 K (nA) (n) GABA (100 µM) (n) EC50[95% CI] (µM) (logEC50± SEM) (n) n H ±SEM
  1. n.i., no ionic current was detected by GABA (100 µM).

KIR3.1–3.4252 ± 38.5 (7)207 ± 28 (7)3.1 [1.1–8.8] (0.49 ± 0.22) (3–5)0.93 ± 0.33
KIR3.1F137S777 ± 160 (8)349 ± 175 (8)1.43 [0.44–4.7] (0.16 ± 0.13)0.57 ± 0.11
KIR3.1F137SY107A262 ± 30 (12)130 ± 421.37 [0.5–3.7] (0.14 ± 0.09)0.61 ± 0.09
KIR3.1F137SY120A155 ± 18 (13)133 ± 31 (13)6.71 [3.4–13.1] (0.83 ± 0.05)0.67 ± 0.06
KIR3.1F137SY128A129 ± 18 (9)26 ± 9 (9)2.76 [1.5–5.0] (0.44 ± 0.05)0.80 ± 0.08
KIR3.1F137SF130A34.8 ± 4 (5)n.i. (5)
KIR3.1F137SF134A66.7 ± 4.4 (3)n.i. (3)
KIR3.1F137SF136A28.5 ± 3.7 (11)n.i. (11)
KIR3.1F137SY146A53 ± 8 (6)n.i. (4)
KIR3.1F137SY148A185 ± 16 (15)153 ± 20 (13)2.42 [2.0–2.9] (0.38 ± 0.02)1.23 ± 0.05
KIR3.1F137SY150A45.2 ± 3.9 (20)33 ± 4 (19)1.22 [0.76–2.0] (0.087 ± 0.05)0.94 ± 0.10

The basal currents observed for the double mutants KIR3.1F137SF130A, KIR3.1F137SF134A, KIR3.1F137SF136A and KIR3.1F137SY146A did not significantly differ from the basal currents produced by uninjected oocytes in the presence of 45 mM K+, and ranged from 25 to 56 nA (Table 2). Naringin (100 µM) did not further activate oocytes expressing these mutants. In addition Ba2+ (3 mM) had no effect at these mutants (data not shown).

The activity of GABA on GABAB(1b,2) and KIR3.1F137S channel expressed in oocytes was not significantly different from wild-type KIR3.1–3.4 channels (P > 0.05; Table 3). When GABA (100 µM) was tested against the double mutants KIR3.1F137SF130A, KIR3.1F137SF134A, KIR3.1F137SF136A and KIR3.1F137SY146A, it did not further activate oocytes expressing these mutants. The absence of ionic currents regardless of using either GABA or naringin indicates that these amino acids are important for cell surface expression and or channel gating.

Detectable currents in the presence of naringin and GABA were obtained from the double mutants KIR3.1F137SY107A, KIR3.1F137SY120A, KIR3.1F137SY128A, KIR3.1F137SY148A and KIR3.1F137SY150A. The basal currents evoked by the double mutants in the presence of 45 mM K+ are summarized in Table 2 and ranged from 62 to 263 nA. In general, these mutants had lower basal currents compared with wild-type KIR3.1–3.4 channels. Naringin (100 µM) further evoked currents ranging from 27 to 155 nA (Table 2).

Figure 7 shows the current–voltage (I–V) relationship for naringin from −80 to +80 mV using 45 mM K+ in the absence and presence of naringin (100 µM) mutant GIRK1 channels. The curves indicate all mutant channels are strongly rectified and the effect of naringin on these channels is independent of the voltage. Example traces for naringin (100 µM)-induced responses at wild-type mutant channels are shown as an inset in Figure 7.

image

Figure 7. Current–voltage relationship for naringin-sensitive currents. The insets are representative traces of naringin (100 µM)-evoked currents at (A) KIR3.1F137S, (B) KIR3.1F137SY107A, (C) KIR3.1F137SY120A, (D) KIR3.1F137SY128A, (E) KIR3.1F137SY148A and (F) KIR3.1F137SY150A mutant channels expressed in Xenopus oocytes and measured at 45 mM K+ buffer. For I–V curves (A) KIR3.1F137S, (B) KIR3.1F137SY107A, (C) KIR3.1F137SY120A, (D) KIR3.1F137SY128A, (E) KIR3.1F137SY148A and (F) KIR3.1F137SY150A current responses were measured at a holding potential of −30 mV. Currents were measured in a 10 mV increment from −80 to 60 mV in response to 100 ms voltage steps in the presence of ND96 buffer, 45 mM K+ buffer (open symbols), and naringin (100 µM) in the presence of 45 mM K+ buffer (solid symbols). Current–voltage relationships in ND96 buffer were subtracted offline from traces recorded in 45 mM K+ buffer, and naringin (100 µM) in the presence of 45 mM K+ buffer to correct for leak and endogenous oocyte currents. Each voltage point is shown as mean ± SEM of current (µA) from three to six oocytes.

Download figure to PowerPoint

Concentration–response curves for naringin were determined for all functional mutants. There were no significant differences in the EC50 or nH for naringin at the double mutants KIR3.1F137SY107A, KIR3.1F137SY120A and KIR3.1F137SY128A (P > 0.05; Table 2; Figure 8A) when compared with either KIR3.1F137S or wild-type KIR3.1–3.4 channels. In contrast, there were significant differences in the nH and EC50 of naringin at KIR3.1F137SY148A and KIR3.1F137SY150A double mutants (Table 2; Figure 8A and B) compared with GIRK1F137S or wild-type GIRK1/4 channels, indicating that Tyr148 and Tyr150 may play a role in either the binding and/or gating of naringin at these channels. The low nH for naringin at KIR3.1F137SY148A and KIR3.1F137SY150A may also reflect the incomplete concentration–response curve, as the high EC50 values may be distorting these figures. Both the EC50 and nH values for naringin on the double mutants KIR3.1F137SY148A and KIR3.1F137SY150A could not be experimentally determined because of solubility problems encountered by naringin at the high concentrations required to complete the concentration–response curves. Instead, EC50 and nH values were calculated by extrapolating the available data using GraphPad Prism v5.0 software. From the extrapolated curves, EC50 values of 4665 µM and 2861 µM were obtained for naringin at KIR3.1F137SY148A and KIR3.1F137SY150A respectively. Thus the effect of naringin was substantially reduced by approximately 40-fold at KIR3.1F137SY148A when compared with wild-type KIR3.1–3.4 channels (P < 0.001) and approximately 20-fold when compared with KIR3.1F137S (P < 0.01). Furthermore, naringin's effect on KIR3.1F137SY150A was reduced by 20-fold in comparison with wild-type KIR3.1–3.4 channels (P < 0.01).

image

Figure 8. (A) Concentration–response curves for naringin on oocytes expressing KIR3.1–3.4 (inline image), KIR3.1F137S (▴), KIR3.1F137SY107A (▵), KIR3.1F137SY120A (○), KIR3.1F137SY128A (□), KIR3.1F137SY148A (▾) and KIR3.1F137SY150A (●). Data points are expressed as mean ± SEM (n= 3–18 oocytes per point from at least two harvests). There is a statistically significant difference in the EC50 and Hill coefficient (nH) values for naringin at KIR3.1F137SY148A and KIR3.1F137SY150A compared with wild-type KIR3.1–3.4 and KIR3.1F137S channels (Table 2), indicating that Tyr148 and Tyr150 play a role in the binding and/or gating of naringin at KIR3 channels. (B) Concentration–response curves for naringin on oocytes expressing KIR3.1F137S, KIR3.1F137SY148A, KIR3.1F137SY150A and GABAB(1b,2) with KIR3.1F137SY148A (◆) or KIR3.1F137SY150A (◊). Data points are expressed as mean ± SEM (n= 3–11 oocytes from at least two harvests). There were no significant differences between the EC50 or Hill coefficient values for naringin at KIR3.1F137SY148A or KIR3.1F137SY150A in the presence or absence of GABAB receptors (P > 0.05; one-way anova followed by Tukey's post hoc test). (C) Concentration–response curves for GABA on oocytes expressing GABAB(1b,2) coupled with either KIR3.1–3.4, KIR3.1F137S, KIR3.1F137SY107A, KIR3.1F137SY120A, KIR3.1F137SY128A, KIR3.1F137SY148A and KIR3.1F137SY150A. Key to symbols used as in (A) and (B). Data points are expressed as mean ± SEM (n= 3–11 oocytes from at least two harvests). There are no significant differences between the EC50 or Hill coefficient values for GABA at wild-type KIR3.1–3.4 and mutant channels (P > 0.05; one-way anova followed by Tukey's post hoc test). (D) Concentration–response curves for adenosine on oocytes expressing wild-type KIR3.1–3.4, KIR3.1F137SY148A and KIR3.1F137SY150A mutant channels. Data points are expressed as mean ± SEM (n= 4 oocytes per point from at least two harvests). (E) Concentration–response curves for LPA on oocytes expressing wild-type KIR3.1–3.4 and KIR3.1F137SY148A mutant channel. Data points are expressed as mean ± SEM (n= 3 oocytes per point from at least two harvests).

Download figure to PowerPoint

There was no significant difference in the activity of naringin at KIR3.1F137SY148A and KIR3.1F137SY150A in the presence or absence of GABAB receptors [P > 0.05; Student's t-test (comparing logEC50 value for each mutant in the presence or absence of GABAB receptors); Table 2; Figure 8B].

GABAB(1b,2) receptors were also expressed with the KIR3.1 double mutants, KIR3.1F137SY107A, KIR3.1F137SY120A, KIR3.1F137SY128A, KIR3.1F137SY148A and KIR3.1F137SY150A. Tables 2 and 3 summarize the basal currents generated in the presence of 45 mM K+ (ranging from 35 to 262 nA) and the additional currents evoked by GABA (100 µM) (ranging from 26 to 153 nA).

GABA concentration–response curves were also determined for all functional mutants (Figure 8C). The EC50 values for GABA (Table 3) at these mutants were not significantly different from either the single KIR3.1F137S mutant or wild-type KIR3.1–3.4 channels (P > 0.05) (Figure 8C). Furthermore, the nH for GABA at KIR3.1–3.4 was not statistically different from those obtained from the mutant channels (P > 0.05).

To further provide evidence that the KIR3.1F137SY148A and KIR3.1F137SY150A mutants do not affect the affinity of ligands stimulating GPCRs, we evaluated the effects of adenosine and LPA on their prospective receptors expressed endogenously in oocytes. Adenosine concentration–response curves were constructed for KIR3.1F137SY148A and KIR3.1F137SY150A (Figure 8D) and compared with wild-type KIR3.1–3.4. The EC50 and nH values for adenosine at either KIR3.1F137SY148A or KIR3.1F137SY150A were not significantly different from wild-type KIR3.1–3.4 (P > 0.05; Table 4), indicating that the mutations did not affect adenosine activity.

The effect of LPA was also evaluated at KIR3.1F137S and KIR3.1F137SY148A channels. We chose to stimulate the LPA receptor because at low LPA concentrations, GIRK channels are activated (Itzhaki Van-Ham et al., 2004) while at high LPA concentrations, Ca2+-activated Cl- channels are stimulated (Liliom et al., 1996) – reflecting what was observed with naringin in these studies. In oocytes expressing KIR3.1–3.4 channels, LPA produced a concentration-dependent response by activating KIR3.1–3.4 channels via the endogenous receptor. The EC50 and the nH values for LPA on wild-type KIR3.1–3.4 and KIR3.1F137SY148A mutant channels were not significantly different (P > 0.05; Figure 8E; Table 4).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

KIR3 channels play an important role in regulating postsynaptic potentials in the central and peripheral nervous systems (Dascal, 1997). Such channels have been implicated in pain perception and in the pathophysiology of several diseases such as epilepsy and ataxia (Luscher and Slesinger, 2010). Agents that can regulate neuronal excitability via KIR3 channels through activation, modulation or inhibition have the potential to alleviate pain, reduce increased heart rate associated with arrhythmias and treat epilepsy (Luscher and Slesinger, 2010).

Flavonoids possess many biological actions and exert significant peripheral and central actions such as antihepatotoxic, anti-allergic, anti-inflammatory, anti-osteoporotic, antitumour and neurological activities. Their exact mechanisms of action remain unclear. This study evaluated the effects of the flavonoids, naringin and its aglycosylated analogue (±)-naringenin on recombinant wild-type and mutant KIR3 channels expressed in Xenopus oocytes. The data presented show that naringin but not (±)-naringenin activates KIR3 channels (KIR3.1–3.2 and KIR3.1–3.4) indicating that the sugar moiety found on the C7 position is important for the activation. Although flavonoids are known to affect many channels including GABAA receptors (Campbell et al., 2004; Hall et al., 2005; Johnston et al., 2006) and BKCa channels (Nardi et al., 2003), with few exceptions (Viswanathan et al., 1984; Fernandez et al., 2006; Meotti et al., 2006a,b; 2007; Loscalzo et al., 2008;) most reported studies show that the aglycosylated flavonoids are more effective than their glycosylated counterparts. Of particular significance is the study by Saponara and colleagues who report that (±)-naringenin was more potent than naringin at activating the related BKCa channels (Saponara et al., 2006).

The activation of KIR3 channels generally involves the downstream signalling effectors of Pertussis toxin-sensitive GPCRs. Thus, upon GPCR activation by specific neurotransmitters, hormones or peptides, the inactive Gαi/o-GDP is converted to Gαi/o-GTP and releases Gβγ subunits from the Gαi/oβγ trimeric complexes. Consequently, the dissociated Gβγ interacts with the GIRK channel C- and N-termini regions leading to channel activation (Dascal, 1997; Yamada et al., 1998; Sadja et al., 2003).

Few substances have been reported to activate KIR3 channels via a mechanism that is independent of GPCR stimulation. Only ethanol (Kobayashi et al., 1999) and halothane (Weigl and Schreibmayer, 2001) directly activate KIR3 channels. Ethanol activates the channels independently of the Gαi/o protein (Kobayashi et al., 1999; Lewohl et al., 1999) and is proposed to act at the C-terminal domain (Lewohl et al., 1999; Hara et al., 2001; Zhou et al., 2001) by either enhancing PIP2 interactions or by increasing the activity of phosphatidylinositol transfer protein, which could increase the levels of PIP2 (Zhou et al., 2001). Recently structure-based mutagenesis was used to probe a putative alcohol-binding pocket located in the cytoplasmic domains of KIR3 channels (Aryal et al., 2009). The study showed that ethanol could have two binding sites: an activating site and an inhibitory site.

In contrast, halothane is a direct partial inverse agonist on KIR3 channels. At low doses (µM range) it inhibits KIR3 channels while at high doses (mM range) it activates the channel via allosteric promotion of Gβγ dissociation from the Gαi/o protein (Weigl and Schreibmayer, 2001; Milovic et al., 2004). The mechanism by which halothane activates KIR3 channels to promote Gβγ dissociation from the Gαi/o protein is not known.

In this study, we showed that naringin also activates KIR3 channels via a GPCR-independent mechanism and on KIR3.1–3.4, the activation was blocked by tertiapin-Q. Tertiapin-Q is a potent blocker of certain KIR including KIR3.1–3.2, KIR3.1–3.4, the large conductance K+ channel and G protein-insensitive inwardly rectifying rat KIR1.1 channels exhibiting low nanomolar affinity. Mutagenesis studies identified several amino acids important for tertiapin-Q's affinity for KIR1.1 activity. These amino acids were located on the external vestibule of the K+-conduction pore, specifically the M1–M2 linker (Jin et al., 1999). Further evidence to indicate binding of tertiapin-Q occurs at the M1–M2 linker came from studies using the inwardly rectifying KIR1.1 channel. KIR1.1 is insensitive to tertiapin-Q but when the M1–M2 linker of the KIR3.4 but not the KIR3.1 subunit was substituted into the KIR1.1 subunit, high affinity for the toxin was conferred (Ramu et al., 2004). The affinities reported in these studies for tertiapin-Q at KIR3, KIR1.1GIRK4 chimeras and KIR1.1 channels were determined electrophysiologically measuring channel function. Thus the amount of basal block at any given concentration is assumed to relate to its affinity.

From our studies, we observed that the activation of wild-type KIR3 channels by naringin was inhibited by tertiapin-Q in a competitive manner indicating that naringin and tertiapin-Q share a common or overlapping binding site. When measured in this manner, the affinity of tertiapin-Q was found to be 20-fold more potent than previously reported. Thus tertiapin-Q can bind to KIR3 channels with high affinity and not affect the basal current, but as the concentration increases the probability of the closed channel state also increases.

Interestingly tertiapin-Q was over 1000-fold weaker at KIR3.1F137S mutant receptors compared with wild-type KIR3.1–3.4 receptors, supporting the work by Ramu et al. (2004). However as there is no commercially available radioligand, the study by Ramu and colleagues could not differentiate between a gating or binding mechanism. In our studies we evaluated the effect of naringin in the presence and absence of tertiapin-Q on the KIR3.1F137S mutant and found that tertiapin-Q could not inhibit the effect of naringin despite a significant effect on wild-type receptors by the same tertiapin-Q concentration. This indicates that naringin and tertiapin-Q do not share a common binding site on the KIR3.1 subunit. Instead, we hypothesize that naringin and tertiapin-Q share an overlapping binding site on the KIR3.4 subunit, as naringin does not differentiate between the KIR3.1 and KIR3.4 subunits, activating both KIR3.1 and KIR3.4 homomeric channels (KIR3.1F137S and KIR3.4S143T) with similar activity.

Important amino acids that confer naringin activity were also identified in this study by performing an alanine-scan of all aromatic amino acids within the M1–M2 linker of the KIR3.1 subunit. We chose to mutate aromatic amino acids because the core structure of naringin has two aromatic ring systems, which are likely to interact with aromatic amino acids within the M1–M2 linker region either via π–π and/or hydrophobic interactions. Of the 10 mutations studied, four mutations were not functional: KIR3.1F137SF130A, KIR3.1F137SF134A, KIR31F137SF136A and KIR3.1F137SY146A. The lack of functionality was not altogether surprising as Phe130, 134, 136 and Tyr146 are conserved in the KIR3 family. Of particular interest was KIR3.1F137SY146A. This mutation is located in the K+ selectivity filter TIGYG/TIGFG (Heginbotham and MacKinnon, 1992; Heginbotham et al., 1992; Doyle et al., 1998), which is highly conserved within the K+ channel family. Mutating this position in the Shaker K+ channel also knocks out agonist-induced ionic currents (Heginbotham et al., 1994). Thus, this particular mutation was expected to be involved in conductance and indeed this mutation lacked any conductance when stimulated by GABA or naringin.

Jin et al. (1999) also found a lack of conductance when they mutated Phe at positions 132 and 134 of the related rat KIR1.1 channel to Ala. Phe132 and 134 correspond to Tyr134 and 136 of the KIR3.1 channel. As GABA and naringin could not stimulate KIR3.1F137SF134A and KIR3.1F137SF136A channels, the Phe134 and 136 are most likely involved in channel conductance as such positions are conserved throughout the channel family.

In contrast, KIR3.1F137SY107A, KIR3.1F137SY120A, KIR3.1F137SY128A, KIR3.1F137SY148A and KIR3.1F137SY150A formed functional channels. These mutations had, in general, reduced currents: a property that could be attributed to (i) the endogenous Gα and Gβγ subunit concentrations, which in this study were not controlled; (ii) KIR3 mutant expression levels; and/or (iii) reduced conductivity levels indicating possible changes in channel conformations that lead to the reduced conductivity (Colquhoun, 1998). Reduced conductance or expression levels may interfere with the pharmacodynamic properties of a compound leading to shifts in the EC50 values for ligands and possibly observing them to be lower than expected. However, this was not the case as all mutants responded similarly to GABA with no significant changes in GABA, adenosine or LPA activity. Thus, GPCR expression levels (Henry et al., 1995) and the various levels of endogenous Gα subunits (Shea et al., 2000; Zhang et al., 2002) within oocyte batches did not contribute to GPCR activity when coupled to the KIR3.1 mutants and therefore the mutations had no influence on the way GABA, adenosine or LPA activated the channels. These data are consistent with the hypothesis that naringin binds to the M1–M2 linker of the KIR3 channel, a site distinct from direct G protein interactions and that Tyr148 and 150 may play a role in either the binding and or gating of naringin at KIR3 channels.

Interestingly in the study by Jin and colleagues, a number of amino acids within the M1–M2 linker of KIR1.1 that affected the affinity of tertiapin-Q were identified (Jin et al., 1999), including Phe146 and 148. Phe146 and 148 of rat KIR1.1 form part of the external vestibule of the ion conduction pore and when the rat KIR1.1 sequence is aligned against the KIR3.1 subunit, the phenylalanines correspond to Tyr148 and 150 of the KIR3.1 channel. The fact that both Phe146 and 148 of rat KIR1.1 and Tyr148 and 150 of KIR3.1 subunits affect the affinities of tertiapin-Q and naringin, respectively, indicates the importance of these positions for possible ligand–channel interactions.

Despite the vast knowledge on the structure of the KIR3 channel, the pharmacology of this channel family remains largely undeveloped. This study highlights naringin as a lead molecule to further develop into direct activators of the channel. However, naringin cannot act as a drug per se: flavonoid glycosides such as naringin are readily metabolized to the aglycone by microflora in the intestine (Manach et al., 2004). Furthermore, naringin exhibits low uptake by intestinal Caco-2 cells (Tourniaire et al., 2005). This indicates that only trace amounts of naringin (approximately 0.5% of ingested naringin) are available in plasma when taken orally (Ishii et al., 2000) and may not be able to reach target sites such as the heart and brain where KIR3 channels are expressed. However, when given i.p., naringin exerts sedative (Fernandez et al., 2006) and anxiolytic (Fernandez et al., 2009) effects in mice. This does not accord with the results from microdialysis studies that show no detectable levels of naringin in rat brain after administration of 30 mg·kg−1, i.p. (Tsai, 2002). However, in the same study it was shown that in the presence of a P-glycoprotein blocker (Tsai, 2002) naringin concentrations are increased in the brain. However, naringin is an example of a generic activator of KIR3 channels being equipotent on KIR3.1–3.2 and KIR3.1–3.4. Only compounds acting on specific KIR3 isoforms may be of therapeutic value. The full elucidation of the naringin binding site will aid in the development of such agents.

In conclusion, our study shows that naringin activates KIR3 channels by binding to the M1–M2 linker of the KIR3.1 and KIR3.4 subunits. Mutagenesis studies on the KIR3.1F137S showed that Tyr148 and Tyr150 may be involved in naringin's effects suggesting that channel–ligand interactions may occur via π–π interactions. As tertiapin-Q can completely inhibit naringin at KIR3.1–3.4 channels in a competitive manner, they must share a common or overlapping binding site, hypothesized to be at a site on the KIR3.4 subunit.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

We are very grateful to the Department of Pharmacology, The University of Sydney, for managing and maintaining the X. laevis colony. T.T.Y. acknowledges the financial support of the Australian Postgraduate Award and the John Lamberton Scholarship.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest
  9. References
  10. Supporting Information

Supporting Information: Teaching Materials; Figs 1–8 as PowerPoint slide.

Figure S1 Ca2+-activated Cl- channel activity elicited by naringin in (A) ND96 and (B) 45 mM K+ buffer. The distinctive sharp inward currents were removed by 11 mM EGTA in the intracellular pipettes. Current responses that contained the distinctive sharp inward current were discarded from further analysis. (C) Current voltage relationship for GIRK 1/4 wild-type channels in the presence of 45 mM K+ buffer with (○) and without (□) 11 mM EGTA in the pipette solution, and in the presence of 100 μM naringin in 45 mM K+ buffer with (●) and without (■) 11 mM EGTA in the pipette solution. The residual current at ND96 is subtracted from the measured current, and each cell is then normalised to the current at −100 mV.

Figure S2 (A) Effect of a maximum concentration of GABA (100 μM; filled bar), submaximal concentration of GABA (3 μM; open bar) and naringin (100 μM; backward hatched bar) on GABAB(1b,2) receptors coupled to GIRK1/4 channels expressed in Xenopus oocytes in the presence of 45 mM K+ buffer (open bar). CGP36742 (100 μM; forward hatched bar), a competitive GABAB receptor antagonist inhibited the response of GABA (3 μM; open bar) but not naringin (100 μM; backward hatched bar). (B) Effect of naringin (100 μM; duration indicated by filled bar) activating the channel whereas (±)-naringenin (100 μM; duration indicated by filled bar) and the sugar moiety, neohesperidose (100 μM; duration indicated by solid bar) had no effect on GIRK1/4 channels expressed in oocytes in the presence of 45 mM K+ buffer (duration indicated by open bar). (C) Effect of naringin (100 μM; duration indicated by filled bar) activating the channel whereas (±)-naringenin (100 μM; duration indicated by filled bar) and the sugar moiety, neohesperidose (100 μM; duration indicated by filled bar) had no effect on GIRK1/2 channels expressed in oocytes in the presence of 45 mM K+ buffer (duration indicated by open bar).

Figure S3 An example of a trace showing naringin (100 μM; duration indicated by black bar), naringin (10 μM; duration indicated by white bar) and adenosine (1 μM; duration indicated by white bar) evoked currents. Theophylline (100 μM; duration indicated by black bar) blocks the response by adenosine but not naringin at 10 μM.

Table S1 Sense oligonucleotide primers (from 5′ to 3′)

Table S2 Summary of the binding results for naringin (10 μM)

FilenameFormatSizeDescription
BPH_1315_sm_Figs1-8.pptx608KSupporting info item
BPH_1315_sm_supp_info.doc731KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.