YMCNE-02928; No of Pages 11 Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Molecular and Cellular Neuroscience journal homepage: www.elsevier.com/locate/ymcne

3Q1

Alexey V. Rossokhin ⁎, Irina N. Sharonova, Julia V. Bukanova, Sergey N. Kolbaev, Vladimir G. Skrebitsky

4

Research Center of Neurology, Russian Academy of Medical Sciences, 105064 Moscow, Russia

5

a r t i c l e

6 7 8 9 10

Article history: Received 5 June 2014 Revised 29 September 2014 Accepted 7 October 2014 Available online xxxx

11 12 13 14 15 16 17

Keywords: GABAA receptor Penicillin Isolated neurons Patch clamp Molecular modeling Monte-Carlo energy minimization

R O

i n f o

O

F

2

Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism

a b s t r a c t

T

E

D

P

GABAA receptors (GABAAR) mainly mediate fast inhibitory neurotransmission in the central nervous system. Different classes of modulators target GABAAR properties. Penicillin G (PNG) belongs to the class of noncompetitive antagonists blocking the open GABAAR and is a prototype of β-lactam antibiotics. In this study, we combined electrophysiological and modeling approaches to investigate the peculiarities of PNG blockade of GABA-activated currents recorded from isolated rat Purkinje cells and to predict the PNG binding site. Whole-cell patch-сlamp recording and fast application system was used in the electrophysiological experiments. PNG block developed after channel activation and increased with membrane depolarization suggesting that the ligand binds within the open channel pore. PNG blocked stationary component of GABA-activated currents in a concentrationdependent manner with IC50 value of 1.12 mM at −70 mV. The termination of GABA and PNG co-application was followed by a transient tail current. Protection of the tail current from bicuculline block and dependence of its kinetic parameters on agonist affinity suggest that PNG acts as a sequential open channel blocker that prevents agonist dissociation while the channel remains blocked. We built the GABAAR models based on nAChR and GLIC structures and performed an unbiased systematic search of the PNG binding site. Monte-Carlo energy minimization was used to find the lowest energy binding modes. We have shown that PNG binds close to the intracellular vestibule. In both models the maximum contribution to the energy of ligand–receptor interactions revealed residues located on the level of 2′, 6′ and 9′ rings formed by a bundle of M2 transmembrane segments, indicating that these residues most likely participate in PNG binding. The predicted structural models support the described mechanism of PNG block. © 2014 Published by Elsevier Inc.

E

C

1

R

36 40 38 37

R

39

1. Introduction

42

The γ-aminobutyric acid type A receptor (GABAAR) is a member of the Cys-loop receptor family of pentameric ligand-gated ion channels, which also comprises excitatory nicotinic acetylcholine receptors (nAChR) and 5-hydroxytryptamine receptors, as well as the inhibitory glycine receptors (Betz, 1990; Connolly and Wafford, 2004). Cys-loop receptors consist of an extracellular domain (ECD) including agonist binding pocket and a transmembrane domain (TMD) forming an ion channel. The TMD of each subunit contains four transmembrane helixes (M1–M4) and a large intracellular loop M3–M4. Five M2 segments form the receptor pore. GABAARs are the major inhibitory receptors in the central nervous system (CNS), formed by combination of α1–6, β1–3, γ1–3, ρ1–3, ε, π, δ or θ subunits with the predominant receptor being α1β2γ2 and with a subunit stoichiometry of 2:2:1 (Hevers and Luddens, 1998; Sieghart, 2006). Alpha and beta subunits are necessary

47 48 49 50 51 52 53 54 55

U

45 46

N C O

41

43 44

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

⁎ Corresponding author at: Research Center of Neurology RAMS, 105064 Moscow, by-str. Obukha 5, Russia. E-mail address: [email protected] (A.V. Rossokhin).

for receptor activation by GABA and the γ subunit defines sensitivity to benzodiazepines. The GABAAR is a target for many pharmacological compounds of different classes including benzodiazepines, barbiturates, steroids, and noncompetitive antagonists (Cascio, 2006; Sieghart, 2006). The last ones include β-lactam antibiotics, t-butylbicyclophosphorothionate (TBPS), insecticides and picrotoxin (PTX). Despite large structural diversity, these noncompetitive antagonists are believed to bind in different positions within a common “convulsant” binding pocket in the chloride channel lumen (Chen et al., 2006; Olsen, 2006). The amino acid composition of the M2 helixes determines the ion selectivity and conductance properties of the channel (Galzi et al., 1992; Keramidas et al., 2004). A classic GABAAR blocker is the convulsant PTX that inhibits GABAergic transmission by channel blockade (Dillon et al., 1995; Newland and Cull-Candy, 1992). Accumulating evidences indicate that a PTX binding site is located within the channel pore near the cytoplasmatic end (Buhr et al., 2001; Erkkila et al., 2008; FfrenchConstant et al., 1993). Another well-known open channel blocker of GABAAR is penicillin (Chow and Mathers, 1986; Twyman et al., 1992). While extensive efforts have been made to determine the molecular mechanism of PTX inhibition of GABAAR channels, paradoxically little

http://dx.doi.org/10.1016/j.mcn.2014.10.001 1044-7431/© 2014 Published by Elsevier Inc.

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76

121

133

All recorded Purkinje cells (n = 46) displayed responses to GABA which were modulated by PNG. Fig. 3A illustrates the inhibitory effects of penicillin on currents activated by 5 μM GABA. Penicillin (0.1– 10 mM) applied together with GABA, at − 70 mV suppressed the responses to GABA, measured at steady state, in a concentrationdependent manner. The effect of PNG developed quickly, and was easily reversible. By itself, penicillin (up to 10 mM) had no action on resting currents (data not shown). Application of PNG in the absence of GABA did not change the peak response of subsequent GABA applications (not shown), suggesting that the compounds can only access the binding site in the open state. When GABA receptors were activated by 5 μM GABA the IC50 value for PNG inhibition of GABA-induced currents was 1.1 ± 0.12 mM and Hill coefficient was 0.84 ± 0.10 (n = 6) (Fig. 3B).

134 135

2.2. Penicillin-induced inhibition is more potent at high vs low GABA concentrations

136

The degree of block produced by a fixed concentration of an open channel blocker is expected to increase with agonist concentration

102 103 104 105 106 107 108 109 110 111 112 113 114 115 116

122 123 124 125 126 127 128 129 130 131 132

137

C

100 101

E

98 99

R

96 97

R

94 95

O

92 93

C

90 91

N

88 89

U

86 87

Fig. 1. Sequence alignment of M2 segment of nACh, GABAA and GLIC receptors. Sequences taken from the Protein Data Bank with PDB codes: nAChR 2BG9 and GLIC 2XQ3 and from the UniProt database with ascension numbers GABAA_α1 P14867, GABAA_β2 P47870, GABAA_β1 P18505, and GABAA_γ2 P18507. Sequences aligned relative to the highly conserved Arg 0′ and Lys 0′ (GABAA and nAChR) or Pro 23′ (GABAA, nAChR and GLIC) residues highlighted in bold. Last residue number is shown. Dashed and solid line rectangles underline the difference in M2 amino acid composition in the GABAA_nAChR and GABAA_GLIC models.

F

2.1. Inhibition of GABA-induced currents by penicillin

84 85

O

120

83

because this allows the blocker increased access to its binding site (Ascher et al., 1979). The influence of agonist concentration on the extent of blockade was determined by measuring the inhibition (by 1 mM PNG) of currents evoked by increasing concentration of GABA (Fig. 4A). The amplitudes of currents were measured at the end of drug co-application. PNG inhibition was tested at GABA concentrations from 3 to 300 μM. The inhibition was GABA concentration-dependent, being larger at higher concentrations of GABA (Fig. 4B). Penicillin inhibited the steady-state component of current induced by 3 μM GABA to 75 ± 6.1% of control (n = 4), to 49 ± 2.7% for current induced by 10 μM GABA (n = 4), and to 36 ± 6.2% (n = 4) at GABA concentrations of 100 μM (Fig. 4C). The comparison of concentration–response curve for GABA in control and during co-application with 1 mM PNG shows that the blocker both inhibited the maximal GABA current and shifted dose–response curve to the left (Fig. 4B). The concentration– response curve for GABA experienced a leftward shift: from 15.5 μM in control conditions to 6.0 μM in the presence of 1 mM PNG (n = 4). The shift in GABA dose–response curves fits well to open channel blocking action of PNG.

138

2.3. Penicillin block at different holding potentials

157

To determine the voltage dependence of PNG effect, PNG was co-applied with GABA at several membrane potentials. Fig. 5A illustrates the effect of 1 mM PNG on the current elicited by 5 μM GABA at −110, −90, −70, −50, −30, −10, +10, and +30 mV. The GABA current–voltage relationship was roughly linear in the absence of PNG, but showed significant inward rectification in the presence of PNG (Fig. 5B). Thus, the block of GABA-induced currents was greater at more positive holding potentials. At 1 mM, PNG depressed stationary GABA current by 49 ± 6% at −110 mV and by 87 ± 1% at +30 mV. A voltage-dependent effect suggests that action of a drug is located within the channel pore of the receptor. Therefore, we analyzed the results according to a simple one-site blockade model (Woodhull, 1973). The voltage dependence of the block is shown in Fig. 5C. The proportion of blocked current increased with membrane depolarization and reached the maximum of ~ 90% block at + 30 mV. Using Woodhull model, we estimated the electrical depth of the PNG binding site in GABAAR δ = 0.39 ± 0.03 and respective KD(0) = 200 ± 20 μM (see Experimental methods).

158

2.4. Penicillin block prevents the channel closure and agonist dissociation

176

The termination of GABA and PNG co-application was followed by transient increase in the inward current (“rebound” or “tail” current) (Fig. 3A). The onset of this current coincided with the initiation of perfusion with normal extracellular solution. The amplitude and kinetics

177

R O

2. Results

81 82

P

119

79 80

T

117 118

work has been done on the mechanisms of penicillin interaction with GABAA channel. Penicillin G (PNG) is a prototype of β-lactam antibiotics which are the antibacterial agents. However, administration of these drugs can cause toxic side effect on CNS. Epileptogenic properties of PNG were documented since the 1940s (Chow et al., 2005) and since then adverse effects of PNG were thoroughly investigated (Curtis et al., 1972). The widely accepted theory of the pathogenesis of convulsions induced by penicillin and related β-lactam compounds suggests that these drugs suppress GABAergic transmission in the CNS (Chow and Mathers, 1986). At millimolar concentrations, PNG causes a voltage-dependent open channel block of GABAAR (Fujimoto et al., 1995; Pickles and Simmonds, 1980; Twyman et al., 1992). These data suggest that PNG enters the open pore of GABAAR and then occludes it. This theory is supported by competition between PNG and PTX for suppression of GABAA receptors (Bali and Akabas, 2007). It has been demonstrated that PTX inhibits GABAAR by a noncompetitive, open channel block mechanism (Newland and Cull-Candy, 1992; Olsen, 2006; Yoon et al., 1993). Single channel recordings have shown that PNG reduced mean life-time of the GABA-activated channels in the open state rather than changed the channel conductance (Chow and Mathers, 1986; Twyman et al., 1992) indicating the drug influence on the receptor gating. However, the exact mechanism of PNG interaction with GABAAR and the site of PNG binding in GABAAR pore still remain unknown. In this study, we combined electrophysiological and modeling approaches to investigate the peculiarities of PNG blockade of GABAactivated currents in the isolated cerebellar Purkinje cells and to build the structural model of PNG binding in the pore of GABAAR. We confirm that PNG is an open channel blocker and extend the previous findings by demonstrating that PNG acts as a “sequential blocker”, which prevents the channel from closing while blocked. Such mechanism of block is also known as a “foot-in-the door” effect and was described previously for acetylcholine (Adams, 1976; Neher and Steinbach, 1978), NMDA (Benveniste and Mayer, 1995; Vorobjev and Sharonova, 1994) and GABAA receptors (Kolbaev et al., 2002). We have designed a structural model of PNG binding in the GABAAR pore. The molecular modeling revealed that the maximum contribution to the energy of ligand– receptor interactions is provided by the residues located on the level of 2′, 6′ and 9′ rings formed by a bundle of M2 transmembrane segments, indicating that these residues most likely participate in PNG binding. The proposed model indicates that the binding sites of PNG and PTX are overlapping.

D

77 78

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

E

2

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156

159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175

178 179 180

3

R O

O

F

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

192 193 194 195 196 197 198

D

antagonist, only during generation of this current. We have observed that application of bicuculline did not change neither the amplitude, nor the time course of the current (Fig. 6A, right panel), while PNG (5 mM) completely suppressed it (Fig. 6A, middle panel). Protection of tail current from bicuculline block is consistent with the hypotheses that agonists remain bound in the open-blocked state. However, these experiments do not exclude the possibility that the decay of tail current reflects the closing kinetics of an open, but unliganded receptor, and therefore binding of competitive antagonists has no effect on this process. According to the model of sequential open channel block (Benveniste and Mayer, 1995), the time course of the tail current will depend on the dissociation rate constant for both PNG and the agonists. Thus, a clear relationship between tail current decay kinetics and agonist affinity should be observed, similar to the agonist dependence of NMDA receptor deactivation (Lester and Jahr, 1992) and NMDA receptor blockade by 9-aminoacridine (Benveniste and Mayer, 1995). We tested this hypothesis using three GABAAR agonists with different affinities for

E

T

C

190 191

E

188 189

R

186 187

R

184 185

of this current were dependent on blocker concentration and holding potential (Figs. 3A, 5A). The maximal relative amplitude of the tail current measured from the level of the steady-state current was observed when the degree of block was maximal — at PNG concentration of 10 mM and holding potential of +30 mV. The tail current was not observed when PNG was not applied (Fig. 6A, left panel) or was continuously present in the solution (Fig. 6A, middle panel). These properties of tail current are consistent with the sequential model of open channel block (Adams, 1976; Antonov and Johnson, 1996; Neher and Steinbach, 1978). Sequential blockers prevent the channel from closing while blocked. In this model channel closure takes place after dissociation of the blocker, which occurs via the open state. The appearance of tail current is due to transition from open-blocked to the open state of the receptor (Benveniste and Mayer, 1995; Vorobjev and Sharonova, 1994). This model also implies that a sequential open channel blocker prevents dissociation of the agonist while the channel is blocked. To test the hypothesis that GABA is trapped on the receptors during the tail current development, we applied bicuculline, a competitive GABAA receptor

N C O

182 183

U

181

P

Fig. 2. PNG structure and the method of systematic unbiased search of the minimum energy ligand–receptor complexes. A. PNG structural formulae. B. PNG minimal energy conformation. C. PNG pulling and rotation in the pore of GABAA receptor. Three of five subunits are shown for clarity (α1 — cyan, β2 — rose, γ2 — magenta). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Block by PNG of GABA-induced currents in isolated cerebellar Purkinje cells. A. Changes in the amplitude of currents induced by 5 μM GABA during coapplication with increasing concentrations of PNG. A current trace from a whole-cell voltage clamp recording is shown. PNG concentrations are indicated on the right. Increasing concentrations of PNG led to a decrease of steady-state currents and to an increase of the amplitudes of tail currents. The onsets of traces are slightly shifted relative to each other for the best presentation of the tail currents. Co-application of GABA and PNG is indicated by a horizontal line above the current traces. The cell was held at −70 mV. B. Concentration–response analysis for PNG induced block of currents activated by 5 μM GABA. The data points are the mean steady-state currents normalized to the control current, and the error bars are standard deviations. The line is unweighted least-squares fit of the data to a Hill equation (Eq. (2)).

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

2.5. Molecular modeling

224

In parallel with electrophysiological investigation we used a molecular modeling approach to predict the binding site of PNG in the GABAAR pore. We performed a systematic unbiased search of the binding site pulling PNG through the pore and rotating this molecule around its long axis (see Experimental methods). In that way we found MEC of the ligand–receptor complex at each level of the pore. Fig. 7 presents the energy profile of PNG in the pore of GABAA_nAChR and GABAA_GLIC models. The minimums on the energy curve correspond to the probable sites of PNG binding. In the case of the GABAA_nAChR model one can extract two strongly pronounced minimums in the upper (9 Å) and lower (32 Å) parts of the pore. Only one minimum in the lower part of the pore exists at the depth of 26 Å in the case of the GABAA_GLIC model. Below we will refer to them as upper and lower binding sites, respectively. Fig. 7 shows that the lower binding sites do not coincide in the GABAA_nAChR and GABAA_GLIC models. In both models the energy of Van der Waals and electrostatic interactions changes in opposite directions at PNG moving toward the intracellular space, increases (in absolute values) in the first case and decreases in the second one (data not shown). The growth of the energy of Van der Waals interactions correlates with a decrease of the pore lumen. When PNG is pulled deeper in the lower binding site the Van der Waals interactions attenuate and the total energy of the ligand– receptor interactions decreases (Fig. 7). The difference of the energy profiles in the upper part of the pore is mainly caused by electrostatic component of the energy of ligand– receptor interactions. At physiological pH PNG carboxyl group is negatively charged. In the GABAA_nAChR model six positively charged lysine residues (γ2 20′ and α1, β2, γ2 24′) in the C-terminal part of M2 segments face the pore and strongly interact with PNG. In the GABAA_GLIC model five lysine residues 24′ are moved to the M2–M3 loop that attenuates electrostatic interactions in the upper part of the pore. The upper binding site is found only in the GABAA_nAChR model. At this site Van der Waals and electrostatic interactions approximately equally contribute to the ligand–receptor interaction energy (data not shown). The predicted structural model of PNG binding at the upper site is shown in Fig. 8 (A, B). PNG occludes the pore (Fig. 8A) and accepts two hydrogen bonds from Lys289 (γ2 24′). Along with Lys289, residues Lys279 (24′ α1), Lys285 (20′, γ2), Ile282 (17′ γ2) and His267 (17′ β2) provide strong contribution to the ligand–receptor interaction energy in the upper binding site (Fig. 8A, B). In our models we did not consider the interaction of the ionized residues with counter-ions because the location of the last ones is unknown. Therefore it is reasonable to suggest that some of the positively charged lysine residues facing the pore can be screened by the oppositely charged ions. To estimate the influence of the screening effect on formation of the energy minimum corresponding to the upper binding site we performed three additional series of numerical experiments. We built three modifications of the GABAA_nAChR model with: (1) all six lysine residues (20′ and 24′), (2) two α1 Lys279 and one γ2 Lys289 at 24′ level, and (3) two β2 Lys274 and one γ2 Lys289 at 24′ level generated in the neutral form. We found that in the models with completely or partially neutralized charges on the lysine residues the energy minimum corresponding to the upper binding site disappears (data not shown). Lower binding site was found in both models. Predicted structural models of PNG binding in the lower site in the GABAA_nAChR and GABAA_GLIC models are shown in Figs. 8 (C, D) and 9, respectively. At this site Van der Waals interactions prevail than the electrostatic ones in the ligand stabilization (− 20.1 vs −5.7 kcal/mol in GABAA_nAChR and −29.3 vs −0.4 kcal/mol in GABAA_GLIC). In both models PNG occludes the pore binding in the extended conformation. However, in these models PNG carboxyl group is oppositely oriented in the pore (Figs. 8D, 9B). In the GABAA_GLIC model PNG accepts H-bonds from β2 Thr256 (6′) and γ2 Thr271 (6′) residues.

225 226

O

R

R

E

C

T

E

D

P

R O

O

F

4

U

N

C

Fig. 4. The degree of penicillin effects depends on GABA concentration. A. Representative current traces induced by GABA at various concentrations (3, 10 and 30 μM) in control (thin line) and during coapplication with 1 mM PNG (thick line). The lines above each panel indicate the drug application periods. B. PNG modifies EC50 value of GABA. Concentration–response curves for GABA were obtained in control (squares) and in the presence of 1 mM PNG (triangles). Data points represent average from four cells. Data are normalized with respect to the response to 300 μM GABA in the absence of PNG. The EC50 value for GABA was 15.5 ± 0.2 μM in the absence of PNG and 6.0 ± 0.21 μM in the presence of 1 mM PNG (n = 4). C. Comparison of the percent of inhibition induced 1 mM PNG of sustained GABA currents evoked by different GABA concentrations. Mean values of GABA current recorded during co-application with 1 mM PNG normalized to that evoked by GABA alone (control); n = 4.

217 218 219 220 221 222 223

this receptor — β-alanine, GABA and muscimol, for which it was demonstrated that the time course of current decay (deactivation) depended on agonist affinity (Jones et al., 1998). The time course of decay of PNG-induced tail currents is much faster when β-alanine was applied with PNG (Fig. 6B, left panel) than when GABA was the agonist (Fig. 6B, middle panel) and much slower when muscimol was used as receptor agonist (Fig. 6B, right panel).

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 Q2 285 286 287 288

5

T

E

D

P

R O

O

F

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

289

E

C

Fig. 5. Voltage-dependence of the effect of penicillin on GABA currents. A. Representative current traces induced by 5 μM GABA at different membrane potentials in control (left) and during co-application with 1 mM PNG (right). The lines above panels indicate the drug application periods. B. The current–voltage relationship for the control (5 μM GABA) currents (squares) and the currents blocked by 1 mM PNG (triangles). Responses obtained at different membrane potentials are normalized to the response at −70 mV in the absence of PNG. C. Woodhull analysis of PNG block. Normalized block is plotted versus the holding potential. The data are fitted using Eq. (3) after normalizing. The fraction of the membrane field sensing by PNG δ is 0.39 ± 0.03. The respective KD(0) is 200 ± 20 μM.

300

3. Discussion

301 302

Open channel block is a process by which the ligand binds to the site inside of a channel pore and blocks the flow of ions through that channel. Pore blocker molecules constitute a helpful tool for understanding the general architecture of the permeation pathway and the gating properties of ion channels. Our goal in this study was to examine the PNG-mediated inhibition of the GABAAR and to correlate electrophysiological data with the structural model of PNG binding predicted by molecular modeling approach. There are two main mechanisms by which blockers can interact with ion channel gating, namely, “trapping” and “foot-in-the door”. In the first case, channel closure and agonist dissociation can occur while blocker is bound, trapping the blocker within the receptor's channel until subsequent agonist binding and channel opening permit blocker dissociation. In the second case, bound blocker does not permit the agonist dissociation and the channel transition to the closed conformation.

297 298

303 304 305 306 307 308 309 310 311 312 313 314 315

R

295 296

N C O

293 294

U

291 292

R

299

In the GABAA_GLIC model the ligand phenyl ring deeply penetrates into α1–β2 intersubunit interface and strongly interacts with α1 Val260 (5′), Val257 (2′), Thr256 (1′) and β2 Ala252 (2′) residues (Fig. 9). In the GABAA_nAChR model the PNG phenyl ring faces the α1–β2 interface and residues α1 Thr265 (10′), Leu264 (9′), Thr261 (6′) and β2 Thr256 (6′) form binding pocket for this moiety (Fig. 8D). In both models residues α1 Leu264 (9′), Thr261 (6′) and β2 Thr256 (6′), Ala252 (2′) strongly contribute to the ligand–receptor interactions. Apart from the above, residues γ2 Ser267 (2′) in GABAA_nAChR and γ2 Leu274 (9′), α1 Val257 (2′) in GABAA_GLIC make a significant contribution to the ligand–receptor energy.

290

Simultaneous wash out of the agonist and blocker entails an appearance of the tail currents. It is generally accepted that PNG as well as PTX acts as an open channel blocker in GABAAR. Both drugs require agonist binding and channel opening to access their binding site (Dillon et al., 1995; Newland and Cull-Candy, 1992). Our results are in agreement with these data. However, we found an important difference between PNG and PTX blocking effects. It was shown that channel closure and agonist dissociation are permitted while PTX is bound in the GABAA-receptor-mediated channel, resulting in trapping of the drug in the channel (Bali and Akabas, 2007). This characteristic of the action of PTX, which is termed “trapping channel block”, distinguishes it from “sequential block,” which prevents the channel from closing while blocked (Adams, 1976; Antonov and Johnson, 1996; Neher and Steinbach, 1978). We report here that for PNG, the mechanism of block is clearly different from PTX. Our results are in agreement with the observation that PNG is an open channel blocker and extend these studies by showing that PNG acts as a sequential open channel blocker which prevents agonist dissociation from the open-blocked state. The appearance of tail currents indicates that the blocker has to exit the channel before it can be closed, as has been described for NMDA receptor blocker such as tacrine, 9-aminoacridine, tetrapentylammonium and adamantane derivatives IEM-1857 (Antonov and Johnson, 1996; Benveniste and Mayer, 1995; Sobolevsky, 2000; Sobolevsky et al., 1999; Vorobjev and Sharonova, 1994). In the present experiments the tail currents were evoked following termination of co-application of GABA and PNG (Fig. 3A). These tail currents were resistant to block by competitive GABAA receptor antagonist bicuculline and their decay kinetics were dependent on agonist affinity

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

E

D

P

R O

O

F

6

353 354 355

R

R

O

351 352

C

349 350

N

347 348

(Fig. 6). These results are in accord with a sequential model for block of GABAA receptor by PNG, in which the agonist cannot dissociate from the open-blocked state of the receptor. We have observed that at low GABA concentrations (1–3 μM) 1 mM PNG practically does not change the amplitude of steady-state component of GABA-induced current and only prolongs it (Fig. 4). This finding is consistent with the properties of the sequential open channel block observed with single channel recording. It was shown that during the block of nicotinic acetylcholine receptors by local anesthetic QX-222 the burst length increases, but the cumulative open time within a burst remains constant in the presence of an ion channel blocker (Neher, 1983).

Thus, PNG blocks GABAAR by “foot-in-the door” mechanism. In contrast, tail currents do not appear after termination of PTX block. Bali and Akabas (2007) have shown that PTX molecule is trapped in the closed GABAAR. Even though PTX also blocks the channel, it does not interfere with the gating mechanism, suggesting that PTX molecule can reside in its blocking site even when the channel is closed and that the channel pore is large enough to accommodate this molecule. Along with electrophysiological studies we used a molecular modeling approach to predict the PNG binding site in the GABAAR pore. However, the reliability of predictions, which can be made with the help of homology models, is highly sensitive to the choice of the templates. In

U

345 346

E

C

T

Fig. 6. Properties of the tail current induced by co-application of GABA agonists and PNG. A. Effects of modulators of GABAA receptors on the tail current triggered after termination of GABA and PNG co-application. GABAA receptor competitive antagonist bicuculline does not influence the decay of the response neither in control (left panel), nor after co-application of GABA with PNG (right panel), while PNG itself completely blocks the tail current (middle panel). For drug applications two different pipettes were used. Control responses obtained in the absence of modulator during current deactivation (black curves) are superimposed with the responses obtained in the presence of PNG or bicuculline during washout (gray curves). The bars above panels indicate the drug application periods. All records were obtained from one neuron. B. Amplitude and kinetics of tail currents depend on agonist affinity. Representative traces induced by β-alanine (ALA) (left panel), GABA (middle panel) and muscimol (MUSC) (right panel) in control (thin lines) and during co-application with 1 mM PNG (thick lines). The time constants of the tail current deactivation (single exponential fitting) after agonist + PNG co-application were 13 ms for β-alanine, 54 ms for GABA and 142 ms for muscimol (indicated on the right). The horizontal lines above panels indicate the drug application periods. All records were obtained from one neuron.

Fig. 7. Energy profiles of PNG in the pore of GABAA_nAChR and GABAA_GLIC models.

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

356 357 358 359 360 361 362 363 364 365 366

7

E

D

P

R O

O

F

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

373 374 375 376

E

R

372

R

370 371

this study, we have built homology models of GABAAR using nAChR and GLIC structures as templates. Electron microscopy structure of nAChR obtained with 4 Å resolution (Unwin, 2005) is used traditionally for modeling of GABAAR. However, in this template conformation of TMD corresponds to the closed state of the channel (Miyazawa et al., 2003) and some additional procedures are necessary to model the conducting conformation of GABAAR (O'Mara et al., 2005). Recently, X-ray structure of bacterial pentameric ligand-gated receptor GLIC became available (Bocquet et al., 2009). Apparent advantages of GLIC template are an open conformation of the channel and its higher resolution (2.9 Å).

N C O

368 369

One of the aims of our study was to check the possibility of usage of GLIC structure as a template for homology modeling of TMD of GABAAR. We performed an unbiased systematic search of the PNG binding sites in the pore of the GABAA_nAChR and GABAA_GLIC models. The upper binding site was found only in the GABAA_nAChR model (Fig. 7). Negatively charged PNG is very attractive for positively charged residues lining the pore at the extracellular vestibule. However, residues composing M2 segment are different in the homology models based on nAChR and GLIC structures (Fig. 1). Six lysine residues facing the pore exist in the 20′ and 24′ rings in the GABAA_nAChR model while only

U

367

C

T

Fig. 8. Predicted structural model of PNG binding in GABAA_nAChR model. View from extracellular space (A, C) and side view (B, D) on the upper (A, B) and lower (C, D) binding sites. In A–D parts of the M2 helixes are shown. PNG and side chains of residues which strongly contributed to the ligand–receptor energy are represented by sticks. In B and D frontal subunit is not shown for clarity. H-bonds are depicted by dotted lines. Subunit α1 is in cyan, β2 is in rose, and γ2 is in magenta. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 9. Predicted structural model of PNG binding in GABAA_GLIC model. View from extracellular space (A) and side view (B) on the lower binding site. In A and B parts of the M2 helixes are shown. PNG and side chains of residues which strongly contributed to the ligand–receptor energy are represented by sticks. In B frontal subunit is not shown for clarity. H-bonds are depicted by dotted lines. Subunit α1 is in cyan, β2 is in rose, and γ2 is in magenta. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

377 378 379 380 381 382 383 384 385 386

410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452

485

4.1. Preparation of Purkinje cells

486

All experiments were conducted in accordance with the requirements of the Ministry of Public Health of the Russian Federation and were consistent with EU directive for Use of Experimental Animals of the European Community. Maximal efforts were made to minimize both the number and sufferings of used animal. Dissociated neurons were prepared from 14–18-day-old Wistar rats using the method described previously (Vorobjev et al., 1996). Briefly, sagittal slices of the cerebellum were incubated at room temperature for 1–6 h on a mesh near the bottom of a beaker. The incubation solution had the following composition (in mM): NaCl 124, KCl 3, CaCl2 2.4, MgCl2 1.5, NaH2PO4 1.3, NaHCO3 26, glucose 10, phenol red 0.01%, continuously bubbled with carbogene (5% CO2 + 95% O2). One at a time, slices were transferred to the recording chamber and neurons were isolated by vertical vibration of a glass sphere, 0.7 mm in diameter, placed close to the surface of the slice (Vorobjev, 1991). Manipulation and cell identification were performed using an inverted microscope. Isolated Purkinje cells were distinguished from other cerebellar cells based on their large cell bodies (approximately 20 μm) and characteristic pear shape attributable to the stump of the apical dendrite. The solution for dissociation and recording had the following composition: NaCl 150, KCl 5, CaCl2 2.7, MgCl2 2.0, HEPES 10, pH adjusted to 7.4 with NaOH.

487

4.2. Whole-cell recording

509

Voltage-clamp recording was obtained using the whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). Glass recording patch pipettes were prepared from filament-containing borosilicate tubes using a two-stage puller. The electrodes, having resistance of 2–2.5 MΩ,

510

O

F

4. Experimental methods

R O

408 409

P

406 407

453 454

D

404 405

subunit dramatically reduced the PNG blocking potency. Bali and Akabas (2007) established the competitive character of PNG and PTX interactions at the GABAAR block. Our experimental data and structural model allow us to suppose that PTX and PNG binding sites overlap but not completely coincide. Purkinje cells are thought to express a homogeneous population of GABAA receptors composed of α1-, β2/3-, and γ2-subunits (Laurie et al., 1992; McKernan and Whiting, 1996). However, Kelley et al. (2013) have observed that the β1 subunit in Purkinje cells is expressed at the similar level as β2 one. Thus, it is quite probable that both subunit combinations (α1β2γ2 and α1β1γ2) of GABAAR are presented in Purkinje cells. Sequence alignment of β1 and β2 subunits has shown that M2 segments of these subunits are almost identical (Fig. 1). Only one amino acid in 15′ position is different in M2 segment of these subunits, namely Asn in the β2 subunit is replaced by Ser in β1 one. However, 15′ residue doesn't participate in PNG binding in both predicted binding modes. Therefore, modeling results obtained for GABAAR of α1β2γ2 composition are also valid for α1β1γ2 one. In conclusion, in this work we performed electrophysiological and modeling investigation of PNG blockade of GABAAR. We have shown experimentally that PNG acts as a sequential blocker, which requires a channel opening and after binding prevents a channel transition to the closed state. Thus, PNG block corresponds to a “foot-in-the door” mechanism. We used nAChR and GLIC structures as templates for modeling GABAAR and proposed two models of PNG binding in GABAAR pore. In both cases residues located on the level of 2′, 6′ and 9′ rings revealed the maximum contribution to the energy of ligand–receptor interactions. The predicted structural models support the described mechanism of PNG block. The results also imply that GLIC can be used for homology modeling of the pore region of GABAA R. This work can help in mapping of GABAA R for such medically important drugs as β-lactam antibiotics.

T

402 403

C

400 401

E

398 399

R

396 397

R

394 395

O

393

C

391 392

N

389 390

one 20′ lysine remains in the GABAA_GLIC model. Residues belonging to the 21′–24′ rings move to the M2–M3 loop in the last model. This explains the lack of the upper binding site in the GABAA_GLIC model. We suggested that the upper binding site should not be detected even in the GABAA_nAChR model, since the charges at the ionizable residues should be compensated by counter-ions which are vastly presented at the pore extracellular vestibule. We performed additional numerical experiments to confirm this suggestion. Transformation of all or part of the positively charged residues 20′ and 24′ in the unionized state resulted in the disappearance of the energy minimum corresponding to the upper binding site. Lower binding site was found in both models. These sites are located at the different pore depths 26 and 32 Å (defined according to z coordinate of PNG root atom) in the GABAA_GLIC and GABAA_nAChR models, respectively (Fig. 7). Interestingly, nearly identical sets of residues at the 2′–9′ levels contribute to the interaction with PNG at these sites. The residues mainly involved in the interactions with PNG in the GABAA_GLIC model are located on one helix turn higher than in the GABAA_nAChR model according to the shift revealed by the structural alignment of M2 segments of nAChR and GLIC (Fig. 1). According to the structural model of PNG binding in the GABAA_GLIC pore, the ligand forms two H-bonds with γ2 and β2 6′ threonine residues (Fig. 9B). Previously it was shown (Chen et al., 2006) that some of the insecticides (α-endosulfan, lindane, fipronil) and other noncompetitive antagonists (EBOB, TBPS, PTX and BIDN) for the GABAA receptor with widely diverse structure fit the 2′ to 9′ pore region forming hydrogen bonds with Thr 6′ and hydrophobic interactions with Ala 2′, Thr 6′ and Leu 9′, thereby blocking the channel. We found two PNG binding modes with similar energy of the ligand– receptor interactions, −27.7 vs −25.8 kcal/mol for the GABAA_GLIC and GABAA_nAChR models, respectively (Figs. 8C, D and 9A, B). In both modes PNG binds in the extended conformation but orientation of the molecule is opposite (Figs. 8D and 9B). In the GABAA_GLIC model PNG carboxyl group is oriented toward the extracellular space and located approximately at the depth of 23 Å that is 0.46 of the whole membrane depth (taking into account that the height of the TMD is about 50 Å). This value rather closely matches the value of δ (0.39) obtained according to the Woodhull model from our experimental data. In this model δ defines the fraction of the membrane potential acting at the binding site assuming a uniform electric field distribution across the membrane. In contrast, in the GABAA_nAChR model PNG charged moiety is oriented toward the intracellular space. In the GABAA_nAChR model the lower binding site is located deeper in the pore than in the GABAA_GLIC model and the proximity of the ring of positively charged Arg residues at the 0′ level can influence the orientation of the carboxyl group. It was established experimentally that PNG influences the receptor gating reducing the mean life-time of the GABA-activated channels in the open state (Chow and Mathers, 1986; Twyman et al., 1992). Our model predicts that PNG protrudes its phenyl ring deep into the intersubunit interface (Fig. 9A), thus preventing transition of the receptor to the closed state. We can speculate that this mode of PNG molecule interaction with ion pore keeps the channel in the open-blocked state. It was suggested that the channel gate in Cys-loop receptors is located at the 9′–14′ levels (Miller and Smart, 2010). The participation of residues at the 9′ level in the ligand stabilization at the binding site found in both models should also complicate the channel transition to the closed state. Thus, our model explains the absence of “trapping” in experiments with PNG. PTX blocks GABAAR according to “trapping” mechanism (Dillon et al., 1995; Newland and Cull-Candy, 1992). Therefore, PTX binding site must be below the channel activation gate (Bali and Akabas, 2007). A large number of experimental studies indicate that residues of the 2′ (Buhr et al., 2001; Ffrench-Constant et al., 1993; Zhang et al., 1994) and 6′ (Erkkila et al., 2008; Sedelnikova et al., 2006) levels strongly contribute to PTX stabilization at the binding site. Sugimoto et al. (2002) have shown that mutation of the residue of 6′ level in the β2

U

387 388

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

E

8

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484

488 489 490 491 492 493 494 495 496 497 498 499 Q3 500 501 502 503 504 505 506 507 508

511 512 513

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

4.4. Data analysis

537 538

Whole-cell records were analyzed off-line being exported as text files to Prism (GraphPad Software, San Diego, CA) for further analysis. Agonist concentration–response curve was fit by the least-square method to

533 534

539 540

nH

543 544 545 546

where I is the peak current evoked by agonist concentration [A], Imax is the peak current evoked by a maximal agonist concentration (300 μM GABA), EC50 is the concentration giving half the maximal current, and nH is the Hill coefficient. To quantify the inhibitory effects of PNG upon the current induced by a constant GABA concentration the following equation was used:   nH Þ; I=I cont ¼ 1= 1 þ ð½PNG=IC50 Þ

549 550 551 552 553

U

I=I cont ¼ 1=ð1 þ ð½PNG=KD ð0ÞÞ exp ðzδFVM =RTÞ; 555

ð2Þ

where I is the GABA current at a given concentration of blocker, Icont is the peak current evoked by GABA application without blocker, IC50 is the concentration of the ligand producing a half-maximal block of GABA-mediated responses, and nH is the Hill coefficient. Voltage dependence of the blocking effect was estimated in the paradigm of “Woodhull model” (Woodhull, 1973) using following relationship:

N C O

548

ð1Þ

E

542

  nH ; = 1 þ ð½A=EC50 =Þ

C

I=I max ¼ ð½A=EC50 Þ

T

531 532

R

529 530

R

527 528

ð3Þ

560 561

where I and Icont are the GABA currents at a given membrane potential VM in the presence of blocker and control conditions respectively. KD(0) is the dissociation constant at VM = 0 mV, δ is the electrical depth of the binding site, z is the electrical charge of PNG (−1), F, R is the Faraday and universal gas constant respectively and T is the ambient temperature (F / RT = 0.04 mV−1). The curve fit was performed using the program Prism (GraphPad Software). Data values are presented as mean ± S.E.M.

562

4.5. Homology modeling

563

In this study, we built the model of open GABAAR with α1β2γ2 subunit combination. Since PNG binds within the pore region we restricted

556 557 558 559

564

4.6. Model building

600

The starting backbone geometry of the GABAA_nAChR and GABAA_GLIC models was taken from nAChR and GLIC structures, respectively. The side chain torsions in those residues, which were identical in a template and corresponding model, were assigned starting values as in the template. The all-trans conformations were used as starting approximations for side chains of residues that mismatched between GABAAR and template. Monte-Carlo (MC) energy minimization method (Li and Scheraga, 1987) was applied to optimize the GABAAR models and the ligand–receptor complexes. Energy was minimized in the space of generalized coordinates using ZMM program (http:// www.zmmsoft.com). Nonbonded interactions were calculated using AMBER force field (Weiner et al., 1984) with a cutoff distance of 9 Å. Electrostatic interactions were calculated with solvent exposure- and distance-dependent dielectric function (Garden and Zhorov, 2010) without cutoff. MC-minimization (MCM) was performed in two stages. At the first stage, the energy was MC-minimized with constraints. Cα atoms in the model were constrained to corresponding positions in the template with the help of pins. A pin is a flat-bottom penalty function (Brooks et al., 1985) that allows penalty-free deviations of the respective atom up to 1 Å from the template and imposes an energy penalty for larger deviations. After the constrained MCM trajectory converged, all constraints were removed and the model was refined by the unconstrained MCM procedure. The difference between structures found in the constrained and unconstrained trajectories indicated whether the constrained search yielded a stable structure. MCM trajectory was terminated when 10,000 consecutive energy minimizations did not improve the energy of the apparent global minimum.

601

F

536

525 526

O

535

A modification of the fast perfusion system was used for solution exchange (Vorobjev et al., 1996). Isolated Purkinje cells were first patch clamped and then lifted into the application system, where they were continuously perfused with control bath solution. The substances were applied through а glass capillary ~ 0.2 mm in diameter located within 0.2 mm from the cell. Solution exchange time was measured at the open electrode tip by switching between solutions of different osmolarities. The time constant of a rise of this current was 4.5 ± 0.1 ms (n = 20). For activation of GABAAR, in most experiments GABA was applied for periods of 0.5–1 s, at 40–60 s intervals. Stock solutions of 10 mM GABA and 1 M PNG were prepared in water. Stock serial dilutions provided the final concentrations given below. All reagents were obtained from Sigma.

R O

523 524

520

P

4.3. Drug solutions and drug application

518 519

565 566

D

522

516 517

the model of GABAAR to its TMD. As a prototype for GABAAR building, we used the electron microscopy (EM) structure of nicotinic acetylcholine receptor from Torpedo marmorata (nAChR) (Miyazawa et al., 2003) and X-ray structure of prokaryotic pentameric ligand-gated receptor from Gloeobacter violaceus (GLIC) (Bocquet et al., 2009). In GLIC there are no Cys residues forming SS-bond in the Cys-loop of ECD, but the overall architecture of GLIC is similar to that of Cys-loop receptors (Corringer et al., 2010). Furthermore GLIC was crystallized in a putatively open state and at a higher resolution of 2.9 Å against 4 Å for nAChR (Bocquet et al., 2009). Thus, GLIC appears to be a suitable candidate to build the GABA AR model and for probing pore block events in detail. Since TMD conformation of nAChR was obtained for the closed state of the channel (Miyazawa et al., 2003), for study the open channel block we used the GABAAR model proposed by O'Mara et al. (2005) in which the pore was gently enlarged to conducting state using molecular dynamics simulations. Further in the text we will refer to the models built on the base of nAChR and GLIC structures as GABAA_nAChR and GABAA_GLIC, respectively. Alignment of M2 segment of nACh, GABAA and GLIC receptors is shown in Fig. 1. To facilitate the comparison of different subunits, a common numbering system is used for M2 amino acid positions, starting with 0′, a highly conserved positively charged Arg residue in anionic channels or Lys residues in cationic channels located near the M2 intracellular end. In GLIC cationic amino acids are absent at the N-terminal part of M2, therefore, we aligned its M2 residues according to conserved Pro 23′. This alignment predicts that Asn223 is located at 0′ position in GLIC (Fig. 1). Start of M2 helix in nAChR falls on residue 1′. Moving outwards through the pore, positions 2′, 6′, 9′, 13′, 17′, 20′ and 24′ face the pore, according to the EM structure of nAChR (Miyazawa et al., 2003). However, X-ray structure of GLIC reveals that the N-terminal residue of M2 helix falls on residue −3′ (Bocquet et al., 2009). In GLIC the residues in positions −2′, 2′, 6′, 9′, 13′, 17′ and 20′ face the pore, while residue 24′ enters the M2–M3 loop. Thus, the amino acid composition of M2 segments in the GABAA_nAChR and GABAA_GLIC models differs.

E

521

were filled with recording solution of the following composition (in mM): CsCl 100, CsF 40, CaCl2 0.2, MgCl2 4, HEPES 10, EGTA 5, ATP-Na 4 (pH adjusted to 7.2 with CsOH). Recordings were carried out at room temperature (20–23 °C) using an EPC 7 patch-clamp amplifier. Currents were filtered at 3 kHz, sampled at 10 kHz, and stored on a computer disk. Cells were held at a membrane potential of − 70 mV, and I–V relationships were generated with test potentials from − 110 mV to +30 mV by 20 mV intervals.

514 515

9

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 Q4 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx

670 671 672

This work was supported by the Russian Foundation for Basic Research (12-04-00304) and Grant 517.2014.4 from the Foundation for Support of Russian Scientific Schools.

673

References

674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691

Adams, P.R., 1976. Drug blockade of open end-plate channels. J. Physiol. 260, 531–552. Antonov, S.M., Johnson, J.W., 1996. Voltage-dependent interaction of open-channel blocking molecules with gating of NMDA receptors in rat cortical neurons. J. Physiol. 493 (Pt 2), 425–445. Ascher, P., Large, W.A., Rang, H.P., 1979. Studies on the mechanism of action of acetylcholine antagonists on rat parasympathetic ganglion cells. J. Physiol. 295, 139–170. Bali, M., Akabas, M.H., 2007. The location of a closed channel gate in the GABAA receptor channel. J. Gen. Physiol. 129, 145–159. Benveniste, M., Mayer, M.L., 1995. Trapping of glutamate and glycine during open channel block of rat hippocampal neuron NMDA receptors by 9-aminoacridine. J. Physiol. 483 (Pt 2), 367–384. Betz, H., 1990. Ligand-gated ion channels in the brain: thee amino acid receptor superfamily. Neuron 5, 383–392. Bocquet, N., Nury, H., Baaden, M., Le Poupon, C., Changeux, J.P., Delarue, M., Corringer, P.J., 2009. X-ray structure of a pentameric ligand-gated ion channel in an apparently open conformation. Nature 457, 111–114. Brooks, C.L., Pettitt, B.M., Karplus, M., 1985. Structural and energetic effects of truncating long ranged interactions in ionic polar fluids. J. Chem. Phys. 83, 5897–5908.

650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666

C

648 649

E

646 647

R

644 645

R

642 643

O

640 641

C

638 639

N

636 637

U

634 635

F

Acknowledgments

633

O

669

631 632

R O

667 668

ZMM program was used to build the model of the PNG molecule. Geometry of this compound was optimized using an HGRID (Hot GRID) procedure that submits a large number of MCM trajectories from randomly generated starting points and collects low-energy structures found in each trajectory. Bond angles at all heavy atoms of the ligand were allowed to vary during energy minimization. The atomic charges of the ligand were calculated by the AM1 method (Dewar et al., 1985) using MOPAC program (http://openmopac.net). Since PNG is an open channel blocker we performed the docking of the ligand within the pore region of GABAAR. MCM protocol allows finding energetically optimal ligand–receptor complexes, by varying the position and orientation of the ligand, as well as the torsion angles of the ligand and the surrounding amino acid side chains. However, this search is carried within a certain area around the starting position of the ligand. Due to the lack of experimental data on the location of the PNG binding site we performed an unbiased systematic search of the lowest energy binding modes. We pulled PNG along the pore axis, which coincides with the Z axis of the Cartesian coordinate system, with a step of 1 Å. A position of the drug along the pore was determined by z coordinate of its root atom — nitrogen in the β-lactam group (Fig. 2A). Z axis zero corresponds to the pore entrance from the extracellular space. Minimal energy conformation (MEC) of PNG (Fig. 2B) was used as a starting one in our numerical experiments. At each step, z coordinate of PNG root atom was constrained thus fixing the drug position along the z-axis. However, the drug was able to rotate around the root atom and to move normally to the pore axis so that an MCM trajectory would yield an optimal position and orientation of the ligand at the given level of the pore. Taking into account the fact that an asymmetric drug is placed in an asymmetric environment we also rotated the PNG molecule around its long axis (Fig. 2C) with a step of 30○ in order to find the minimum energy of the ligand–receptor complex at each level. The method of the ligand rotation in the pore of an ion channel was described previously in Rossokhin et al. (2006). In this way the grid of starting points covering all inner surfaces of the receptor pore was formed. In each point of the grid the ligand– receptor complex was MC-minimized and the energy of the ligand– receptor interaction was calculated. The ligand–receptor energy extracted from the energetically best structure found at each level was plotted against the pore depth forming the PNG energy profile in the GABAAR pore.

P

629 630

Buhr, A., Wagner, C.A., Fuchs, K., Sieghart, W., Sigel, E., 2001. Two novel residues in M2 of the g-aminobutyric acid type A receptor affecting gating by GABA and picrotoxin affinity. J. Biol. Chem. 276, 7775–7781. Cascio, M., 2006. Modulating inhibitory ligand-gated ion channels. AAPS J. 8, E353–E361. Chen, L., Durkin, K.A., Casida, J.E., 2006. Structural model for gamma-aminobutyric acid receptor noncompetitive antagonist binding: widely diverse structures fit the same site. Proc. Natl. Acad. Sci. U. S. A. 103, 5185–5190. Chow, P., Mathers, D., 1986. Convulsant doses of penicillin shorten the lifetime of GABAinduced channels in cultured central neurones. Br. J. Pharmacol. 88, 541–547. Chow, K.M., Hui, A.C., Szeto, C.C., 2005. Neurotoxicity induced by beta-lactam antibiotics: from bench to bedside. European J. Clin. Microbiol. Infect. Dis.Off. Public. Eur. Soc. Clin. Microbiol. 24, 649–653. Connolly, C.N., Wafford, K.A., 2004. The Cys-loop superfamily of ligand-gated ion channels: the impact of receptor structure on function. Biochem. Soc. Trans. 32, 529–534. Corringer, P.J., Baaden, M., Bocquet, N., Delarue, M., Dufresne, V., Nury, H., Prevost, M., Van Renterghem, C., 2010. Atomic structure and dynamics of pentameric ligand-gated ion channels: new insight from bacterial homologues. J. Physiol. 588, 565–572. Curtis, D.R., Game, C.J., Johnston, G.A., McCulloch, R.M., MacLachlan, R.M., 1972. Convulsive action of penicillin. Brain Res. 43, 242–245. Dewar, M.J.S., Zoebisch, E.G., Healy, E.F., Stewart, J.J.P., 1985. AM1: a new general purpose quantum mechanical molecular model. J. Am. Chem. Soc. 107, 3902–3909. Dillon, G.H., Im, W.B., Carter, D.B., McKinley, D.D., 1995. Enhancement by GABA of the association rate of picrotoxin and tert-butylbicyclophosphorothionate to the rat cloned α1β2γ2 GABAA receptor subtype. Br. J. Pharmacol. 115, 539–545. Erkkila, B.E., Sedelnikova, A.V., Weiss, D.S., 2008. Stoichiometric pore mutations of the GABAAR reveal a pattern of hydrogen bonding with picrotoxin. Biophys. J. 94, 4299–4306. Ffrench-Constant, R.H., Rocheleau, T.A., Steichen, J.C., Chalmers, A.E., 1993. A point mutation in a Drosophila GABA receptor confers insecticide resistance. Nature 363, 449–451. Fujimoto, M., Munakata, M., Akaike, N., 1995. Dual mechanisms of GABAA response inhibition by fl-lactam antibiotics in the pyramidal neurones of the rat cerebral cortex. Br. J. Phamacol. 116, 3014–3020. Galzi, J.L., Devillers-Thiery, A., Hussy, N., Bertrand, S., Changeux, J.P., Bertrand, D., 1992. Mutations in the channel domain of a neuronal nicotinic receptor convert ion selectivity from cationic to anionic. Nature 359, 500–505. Garden, D.P., Zhorov, B.S., 2010. Docking flexible ligands in proteins with a solvent exposure- and distance-dependent dielectric function. J. Comput. Aided Mol. Des. 24, 91–105. Hamill, O.P., Marty, A., Neher, E., Sakmann, B., Sigworth, F.J., 1981. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch. Eur. J. Physiol. 391, 85–100. Hevers, W., Luddens, H., 1998. The diversity of GABAA receptors. Pharmacological and electrophysiological properties of GABAA channel subtypes. Mol. Neurobiol. 18, 35–86. Jones, M.V., Sahara, Y., Dzubay, J.A., Westbrook, G.L., 1998. Defining affinity with the GABAA receptor. J. Neurosci. Off. J. Soc. Neurosci. 18, 8590–8604. Kelley, M.H., Ortiz, J., Shimizu, K., Grewal, H., Quillinan, N., Herson, P.S., 2013. Alterations in Purkinje cell GABAA receptor pharmacology following oxygen and glucose deprivation and cerebral ischemia reveal novel contribution of beta1-subunit-containing receptors. Eur. J. Neurosci. 37, 555–563. Keramidas, A., Moorhouse, A.J., Schofield, P.R., Barry, P.H., 2004. Ligand-gated ion channels: mechanisms underlying ion selectivity. Prog. Biophys. Mol. Biol. 86, 161–204. Kolbaev, S.N., Sharonova, I.N., Vorobjev, V.S., Skrebitsky, V.G., 2002. Mechanisms of GABA(A) receptor blockade by millimolar concentrations of furosemide in isolated rat Purkinje cells. Neuropharmacology 42, 913–921. Laurie, D.J., Wisden, W., Seeburg, P.H., 1992. The distribution of thirteen GABAA receptor subunit mRNAs in the rat brain. III. Embryonic and postnatal development. J. Neurosci. Off. J. Soc. Neurosci. 12, 4151–4172. Lester, R.A., Jahr, C.E., 1992. NMDA channel behavior depends on agonist affinity. J. Neurosci. Off. J. Soc. Neurosci. 12, 635–643. Li, Z., Scheraga, H.A., 1987. Monte Carlo-minimization approach to the multiple-minima problem in protein folding. Proc. Natl. Acad. Sci. U. S. A. 84, 6611–6615. McKernan, R.M., Whiting, P.J., 1996. Which GABAA-receptor subtypes really occur in the brain? Trends Neurosci. 19, 139–143. Miller, P.S., Smart, T.G., 2010. Binding, activation and modulation of Cys-loop receptors. Trends Pharmacol. Sci. 31, 161–174. Miyazawa, A., Fujiyoshi, Y., Unwin, N., 2003. Structure and gating mechanism of the acetylcholine receptor pore. Nature 423, 949–955. Neher, E., 1983. The charge carried by single-channel currents of rat cultured muscle cells in the presence of local anaesthetics. J. Physiol. 339, 663–678. Neher, E., Steinbach, J.H., 1978. Local anaesthetics transiently block currents through single acetylcholine-receptor channels. J. Physiol. 277, 153–176. Newland, C.F., Cull-Candy, S.G., 1992. On the mechanism of action of picrotoxin on GABA receptor channels in dissociated sympathetic neurones of the rat. J. Physiol. 447, 191–213. Olsen, R.W., 2006. Picrotoxin-like channel blockers of GABAA receptors. Proc. Natl. Acad. Sci. U. S. A. 103, 6081–6082. O'Mara, M., Cromer, B., Parker, M., Chung, S.H., 2005. Homology model of the GABAA receptor examined using Brownian dynamics. Biophys. J. 88, 3286–3299. Pickles, H.G., Simmonds, M.A., 1980. Antagonism by penicillin of gamma-aminobutyric acid depolarizations at presynaptic sites in rat olfactory cortex and cuneate nucleus in vitro. Neuropharmacology 19, 35–38. Rossokhin, A., Teodorescu, G., Grissmer, S., Zhorov, B.S., 2006. Interaction of d-tubocurarine with potassium channels: molecular modeling and ligand binding. Mol. Pharmacol. 69, 1356–1365.

D

4.7. Ligand docking

T

628

E

10

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777

A.V. Rossokhin et al. / Molecular and Cellular Neuroscience xxx (2014) xxx–xxx Sedelnikova, A., Erkkila, B.E., Harris, H., Zakharkin, S.O., Weiss, D.S., 2006. Stoichiometry of a pore mutation that abolishes picrotoxin-mediated antagonism of the GABAA receptor. J. Physiol. 577, 569–577. Sieghart, W., 2006. Structure, pharmacology, and function of GABAA receptor subtypes. Adv. Pharmacol. 54, 231–263. Sobolevsky, A.I., 2000. Quantitative analysis of tetrapentylammonium-induced blockade of open N-methyl-D-aspartate channels. Biophys. J. 79, 1324–1335. Sobolevsky, A.I., Koshelev, S.G., Khodorov, B.I., 1999. Probing of NMDA channels with fast blockers. J. Neurosci. Off. J. Soc. Neurosci. 19, 10611–10626. Sugimoto, M., Fukami, S., Kayakiri, H., Yamazaki, S., Matsuoka, N., Uchida, I., Mashimo, T., 2002. The beta-lactam antibiotics, penicillin-G and cefoselis have different mechanisms and sites of action at GABA(A) receptors. Br. J. Pharmacol. 135, 427–432. Twyman, R.E., Green, R.M., MacDonald, R.L., 1992. Kinetics of open channel block by penicillin of single GABAA receptor channels from mouse spinal cord neurones in culture. J. Physiol. 445, 97–127. Unwin, N., 2005. Refined structure of the nicotinic acetylcholine receptor at 4 A resolution. J. Mol. Biol. 346, 967–989.

Vorobjev, V.S., 1991. Vibrodissociation of sliced mammalian nervous tissue. J. Neurosci. Methods 38, 145–150. Vorobjev, V.S., Sharonova, I.N., 1994. Tetrahydroaminoacridine blocks and prolongs NMDA receptor-mediated responses in a voltage-dependent manner. Eur. J. Pharmacol. 253, 1–8. Vorobjev, V.S., Sharonova, I.N., Haas, H.L., 1996. A simple perfusion system for patchclamp studies. J. Neurosci. Methods 68, 303–307. Weiner, J., Kollman, P.A., Case, D.A., Singh, U.C., Chio, C., Alagona, G., Profeta, S., Weiner, P.K., 1984. A new force field for molecular mechanical simulation of nucleic acids and proteins. J. Am. Chem. Soc. 106, 765–784. Woodhull, A.M., 1973. Ionic blockage of sodium channels in nerve. J. Gen. Physiol. 61, 687–708. Yoon, K.W., Covey, D.F., Rothman, S.M., 1993. Multiple mechanisms of picrotoxin block of GABA-induced currents in rat hippocampal neurons. J. Physiol. 464, 423–439. Zhang, H.G., ffrench-Constant, R.H., Jackson, M.B., 1994. A unique amino acid of the Drosophila GABA receptor with influence on drug sensitivity by two mechanisms. J. Physiol. 479 (Pt 1), 65–75.

F

778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794

11

U

N C O

R

R

E

C

T

E

D

P

R O

O

812

Please cite this article as: Rossokhin, A.V., et al., Block of GABAA receptor ion channel by penicillin: Electrophysiological and modeling insights toward the mechanism..., Mol. Cell. Neurosci. (2014), http://dx.doi.org/10.1016/j.mcn.2014.10.001

795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811

Block of GABA(A) receptor ion channel by penicillin: electrophysiological and modeling insights toward the mechanism.

GABA(A) receptors (GABA(A)R) mainly mediate fast inhibitory neurotransmission in the central nervous system. Different classes of modulators target GA...
2MB Sizes 0 Downloads 6 Views