[page 6↓]

1.  Introduction

1.1. Kir2 channels

1.1.1. Potassium channels

The diversity of potassium-(K+-)specific channels far exceeds any other group of ion channels. The importance of this superfamily of eucaryotic channels is underlined by the responsibility for many regulatory processes1.

Potassium channels contain alpha subunits and in some cases auxiliary beta subunits. The alpha subunit represents the integral membrane protein responsible for the conduction of potassium ions across the lipid bilayer2. Based on their molecular architecture different channel classes are distinguished (Fig. 1). Most channels are referred to as voltage-gated channels (Kv). They are composed of six transmembrane (TM) helices (S1-S6) and one pore-region (P-region), resulting in the 6TM/1P class. Furthermore, the Ca2+-dependent potassium channels with their additional helix, S0, belong to this class. The simplest structural plan of potassium channels is visible in the 2TM/1P class, named after their typical composition of two TM helices and one P-region and represented by the inward rectifier potassium channels (see next chapter). The class of the so-called ‘background’ or ‘leakage’ channels is made up of two P-regions, hence structurally referred to as 4TM/2P type.

Fig. 1: Alpha-subunits of distinct potassium channel classes. A) voltage-gated (Kv) K+ channels, B) inward rectifier (Kir) potassium channels (M1, M2: transmembrane helices, P: pore-forming region), C) ‘background’ or ‘leakage’ channels


[page 7↓]

All potassium channels require a tetrameric arrangement of four P-regions to make up a K+-selective filter. In contrast to the tetrameric association formed by the assembly of four 1P channel subunits in the plasma membrane3, the 4TM/2P type only needs to dimerize to construct a pore.

The structural key causing potassium selectivity of potassium channels is a highly conserved stretch of eight amino acids (TXXTXGYG)4. This K+ channel signature sequence with a GYG (Gly-Tyr-Gly) motif plays an essential role to selectively allow K+ ions to pass through the pore5: The carbonyl oxygen atoms of the channel pore are in the precise position to substitute for the oxygen of solvating water molecules. Thus, the entering K+ ion can strip its hydration shell at reasonably balanced energy cost. In K+ channels, however, the carbonyl oxygen atoms of the selectivity filter are not narrow enough to allow dehydration of the smaller Na+ion.

1.1.2. Kir – inward rectifier potassium channels

Inwardly rectifying K + (Kir) channels represent the minimal requirement for a potassium channel, however, it is involved centrally in physiological functions. The Kir current was first observed electrophysiologically in frog sceletal muscle by Katz in 19496 and named an anomalous rectifier due to its unique properties, which were in contrast to the known voltage-dependent delayed-rectifier K+ channels. In recent years K+ channel genes were discovered encoding channels whose properties verify his original measurement7. In 1993, the primary sequence of Kir channels was identified by Ho et al. who cloned the ROMK gene transcribed from an mRNA sequence from rat kidney cells.

The Kir superfamily contains at least 15 members in six distinct subfamilies (Kir1, Kir2, Kir3, Kir5, Kir6, Kir7) that respond to many types of regulation. In the usual terminology, members are designated Kirx.y, whereas „x“ determines the subfamily and „y“ names the subtype within the subfamily8. Kir subunit cDNAs encode proteins between 327 and 501


[page 8↓]

Fig. 2: Dendrogram of the Kir family. Subfamilies, subtypes and former names. (modified from Reimann and Ashcroft 1999)

amino acids in length. As shown in the dendrogram (Fig. 2), the channel proteins share about 30-40% homology among the Kir subfamilies and >60% within the subfamilies.

The electrophysiological properties of Kir channels include the conduction of more current in the inward than in the outward direction. This is due to a reduced open probability in a depolarized membrane. Hyperpolarization above the K+ equilibrium potential hardly occurs under physiological conditions at the mammalian plasmamembrane. Consequently, despite their name, Kir channels primarily conduct outward potassium currents.

Due to their peculiar properties in a particular cell, Kir channels have a key role in stabilizing the resting membrane potential. The Kir-mediated outward fluxes prevent action potential firing by small electrical stimuli. In case of strong electrical stimuli, however, these channels close and allow action potentials, thereby preventing massive K+ loss. Furthermore, modulation of these channels can alter the membrane potential, cellular excitability and heart rate, information processing and secretion of neurotransmitters or hormones9.


[page 9↓]

1.1.2.1.  Kir subfamilies

Very different physiological functions are assigned to Kir channels originating in the diversity of these proteins. Aside from differential cellular expression the diverse subset of Kir channel subtypes determines the pattern of electrical properties.

The Kir1 subtype is a Kir channel with weak inward rectifying properties that is involved in transepithelial membrane transport, predominantly in the kidney, where it is important in homeostasis by conducting large amounts of potassium. Kir3 channels are the predominant mediators of metabotropic inhibition and are activated directly by Gβγ subunits. They influence electrical activity in neuronal, cardial and neurosecretory cells10. Kir4 is known to control K+ homeostasis in glial cells and the inner ear11. ATP-sensitive Kir6 (KATP) channels play an important role in the pancreas; they influence the control of insulin secretion12 in interaction with SUR (sulphonylurea receptor) subunits. Moreover, this channel may contribute to the cell survival during anoxia or hypoglycaemia in neuronal and cardiac structures.

1.1.3. Kir2 subfamily

Recent investigations produced extensive information about the molecular structure, electrophysiology, modulation and localization of the Kir2 inward rectifier potassium channels. They form tetramers of four identical (homomeric) or related (heteromeric) subunits that each contains two putative membrane spanning regions M1 and M2 (Fig. 1B). The transmembrane segments form alpha helices where M1 makes up the outer and M2 the inner helix13. M1 as well as M2 flank a highly conserved pore-forming region called H5 – at this side the Kir2 subfamily shows sequence differences at only one amino acid residue (Fig. 7). In addition to the common molecular motif, a homotetrameric association of subunits is a crucial structural hallmark of Kir2 channels. Kir2.1 is an exception to the rule: it forms heteromultimers if coexpressed with Kir4.114.


[page 10↓]

1.1.3.1.  Electrophysiology

The rectification property of the “classical” inward rectifying potassium channels (Kir2) includes the following components: 1) Inwardly rectifying channels show a linear conductance below the K+ equilibrium potential (Fig. 3). 2) The outward conductance in physiological conditions helps to stabilize the resting potential15. 3) Negligible outward currents during depolarization prevent massive potassium loss and allow prolonged depolarization (plateau of an action potential)16.

Fig. 3: Inwardly rectifying currents recorded from a Xenopus oocyte expressing a Kir channel.
Experimental conditions with symmetrical K+ concentrations. A) Current traces elicited by voltage steps from 0 mV (holding potential) to +80, +40, 0, -40, -80, -120 and –160 mV. B) Current-voltage relationship of the peak currents recorded from the same oocyte.

Therefore, in theory cells with many of these constitutively active Kir channels and no additional regulation should have a relatively constant resting membrane potential, and, in case of depolarization, they may initiate action potentials with a long-lasting plateau17.

1.1.3.2. Basic principles of inward rectification

The ability of the strongly rectifying Kir2 channels to prevent potassium outflow was originally thought to be the effect of an intracellular Mg2+-mediated blockade18. Cations enter the channel pore and impede K+ efflux. Subsequent investigators postulated the intrinsic rectification caused by soluble factors different from Mg2+. Detailed [page 11↓]characterization identified them being polyamines19 (products of L-ornithine metabolism: putrescine, spermidine, spermine). Spermine functions in switching Kir channels off in case of significant depolarization. The Mg2+ block is evidently faster than the polyamine effect, thus decreasing the channel-open probability20.

Experiments focusing on combination of Mg2+-mediated block and intrinsic polyamine gating revealed a competition between these ions. The finding led to the assumption that these substances bind at the same single site of the Kir channel21. Mutagenesis studies have
Fig. 4: Model of channel block by polyamines . A) Spermine and Mg2+ directly block the pore by binding at residue D172. B) E224 and E299 are discussed controversially: spermine binding at residue E224 leads to a high affinity blocking complex (Lee et al. 1999), or intermediate binding at E224+E299 level facilitates the entry of spermine but does not plug the pore (Kubo and Murata 2001).

characterized at least two negatively charged amino acid residues as regulators of inward rectification (Fig. 4, 5). Aspartic acid D172 in the second transmembrane domain (M2) as well as the widely separated aspartic acid D260, glutamate E224 and E299 in the C-terminal region seem to be key positions in the strong inward rectifier Kir222.

A recent X-ray crystallography analysis reveals many amino acids lining the pore yet to be altered by mutagenesis23. Their further conclusions display the voltage-dependent rectification in a simple fashion: The lining of the cytoplasmic pore contains polar, negatively charged amino acid side chains interspersed with hydrophobic amino acid side chains. These amino acids exactly match with polyamines, which are long chains of positive amino groups separated by hydrophobic spacers.


[page 12↓]

1.1.3.3.  Modulation

Fig. 5: Synopsis of Kir2 channel modulations. For detailed explanation of (a) to (r) see table 1.

Despite their rather simple structure, members of the Kir2 subfamily play a central role in many physiological processes. They depend prominently on multiple regulators24 such as pH, G-proteins, kinases, polyamines, nucleotides and phospholipids, but also on pharmaceutical products like the immunosuppressant cyclosporin A25. Due to the complexity of modulators, a synopsis is given (Fig. 5) summarizing the most important factors with their site of interaction. Moreover, it underlines the extraordinary significance of this information with respect to normal and pathophysiological conditions.


[page 13↓]

Table 1: Kir2 channel modulators.

modulator

channel effect

Kir2.1

Kir2.2

Kir2.3

Kir2.4

reference

 

Mg2+, spermine

inward rectification

+ (r)

   

Yang 1995, Lee 1999

Lee 1999

Kubo 2001

pH extracellular

inhibition

  

+ (r)

+ (h)

Hughes 2000

pH intracellular

inhibition

– (r)

 

+ (r)

 

Qu 1999, Qu 2000

ATP

inhibition

 

+

 

Collins 1996

PIP2

activation

+ (r)

   

Huang 1998

G-protein

suppression

 

+

 

Cohen 1996b

M1 ACh receptor

inhibition

 

+

 

Chuang 1997

5HT1A

inhibition

+

   

Cohen 1996a

Substance P

suppression

    

Stanfield 1985

PSD-95 binding

cytosceletal binding

+

 

+

 

Cohen 1996a

PKC phosphorylation

decreased activity

+

 

+ (r)

+ (h)

Kir2.1: Doupnik 1995, Kir2.3: Zhu 1999b, Kir2.4: Hughes 2000

Asn glycosylation

glycosylation

   

+ (h)

Hughes 2000

tyrosine kinase

phosphorylation

   

+ (h)

Hughes 2000

cAMP/cGMP dependent

phosphorylation

  

+ (r)

+ (h)

Kir2.3: Cohen 1996a

Kir2.4: Hughes 2000

endoplasmic reticulum

ER-export consensus sequence FxYENEV

+ (r)

   

Stockklausner 2001

+ = effect, – = no effect, (r) = rat, (h) = human

1.1.3.4. Distribution of Kir2 channels in the brain

Kir2 channels are expressed by diverse cells including neuronal cells, myocytes, blood cells, sceletal muscle fibers, macrophages, osteoclasts, endothelial and placental cells26. Most information about distribution of Kir channels is based on in situ hybridization experiments27. In general, Kir2 mRNAs are widely distributed throughout the rat brain. Kir2.1, Kir2.2 and Kir2.328 are expressed in dentate gyrus, olfactory bulb, caudate putamen, and piriform cortex (see discussion). Kir2.1 and Kir2.3 are less prominent in thalamus, cerebellum and brainstem, where Kir2.2 displays a prominent expression pattern. The Kir2.4 subunit29 seems to be strongly expressed in several nuclei of medulla and brainstem.


[page 14↓]

So far, only few studies investigated the distribution of Kir proteins in the mammalian brain by using specific antibodies30. Up to date, available data are largely restricted to brain nuclei at the cellular level. Consequently, a more detailed localization at the subcellular and regional level in cells linked to physiological or clinical questions are to the fore.

1.2. Basal ganglia

The basal ganglia are a major neural system of subcortical telencephalic brain nuclei. They are involved in movement disorders, motivated behavior and drug addiction; they are strongly interconnected with each other as well as linked to most elements of the mammalian brain. The main structures are the striatum (STR, caudate nucleus and putamen), pallidum (GP), substantia nigra (pars compacta (SNc) and pars reticulata (SNr)), ventral tegmental area (VTA) and subthalamic nucleus (STN). The pallidum is further divided into an external pallidum (GPe, corresponding to the globus pallidus in rat) and an internal segment (GPi, corresponding to the entopeduncular nucleus in rat).

1.2.1. Principle neuron and interneurons

Simple neuroanatomical techniques (cresyl violet) characterize the homogeneous structure of the striatum, predominantly composed of medium-sized spiny neurons and perforated by numerous fiber bundles. Using axon tracing and immunocytochemistry, the neostriatum seems to be a complex mosaic of connections, cell morphologies and neurochemical markers. Aside from spiny projection neurons, at least four classes of interneurons contribute to the neuronal framework.

The principal medium-sized spiny neuron represents over 90% of neurons in the rodent neostriatum31. The cell bodies range from 12-20 µm in size and these projection cells show the highest spine density in the brain and have highly collateralized axons. Since the spiny [page 15↓]neurons are the main output of the striatum, the principal neuron is the primary circuit element. In addition, most of the incoming axons target their spines.

After identification as projection cells in the 1970s, the projections of spiny neurons were examined. Approximately one half of these cells forms the direct pathway (see chapter 1.2.3.), while the remaining neurons project via the indirect pathway 32. Both subpopulations form GABAergic synapses, but these two classes of spiny neurons are different in their cytochemical profile containing either enkephalin or substance P33. No preferential distribution is present in the striatum.

1.2.1.1. Kir current in striatal spiny neurons

Medium spiny neurons move between two membrane states, referred to as the down-state and the up-state34. During the down-state, the membrane potential remains at relatively hyperpolarized potentials (-85 mV) and the spiny cells do not generate action potentials. In contrast, during the episodes of depolarization (up-state) the membrane potential resides just below threshold (-60 mV) and action potential discharge may occur35. The transition to the depolarized state depends on temporally coherent activity.

The membrane conductance in the hyperpolarized down-state is mainly determined by inwardly rectifying Kir2 potassium channels36. These channels help to hold the neuron close to the K+ equilibrium potential. Thus, the Kir current reduces the effect of temporally or spatially isolated excitatory synaptic input and acts like a barrier for asynchronous excitatory potentials. In up-states, the influence of Kir channels is minimal due to their inward rectification properties.

1.2.1.2. Cholinergic interneurons

Up to 60 µm in length, cholinergic interneurons are the main source of the high acetylcholine (ACh) levels in the striatum and neurotransmission of these cells is of central importance for the final striatal output. The giant interneurons account for less than 2% of [page 16↓]cells in the striatum. They can be identified by staining for choline acetyltransferase (ChAT)37, the enzyme for synthesis of the transmitter acetylcholine from acetyl-CoA and choline. Communication between the striatal compartments (see chapter 1.2.2.) is thought to be performed by very extensive axonal fields and widespread dendritic trees of cholinergic interneurons. Due to their tendency to preferentially arborize in the matrix, the cholinergic cells seem to transmit information from the patch to the spiny projection neurons in the matrix38 that represent the main synaptic target of cholinergic interneurons. Aside from some dopaminergic input, cholinergic interneurons are mainly innervated by cortical and thalamic projections39.

The capability of firing irregularly in a spontaneous tonic pattern named these cells tonically active neurons (TANs). In contrast to spiny neurons, the cholinergic cells integrate the effects of only a small numbers of afferent axons40. TANs exhibit a pause in firing, which is thought to represent the physiological correlate of motor learning and rewarded movements41. Therefore, cholinergic neurons play a pivotal role for striatal output and learning in normal and pathophysiological conditions.

1.2.2. Patch and matrix

Both in the primate and rodent neostriatum, two compartments can be distinguished. The so-called cell islands in primates correspond to striosomes = patches in the rat, where they occupy approximately 10% of striatal space42. The areas surrounding the patches are called matrix. The compartmental organization can be distinguished using biochemical markers such as transmitters, related enzymes or metabolites, and transmitter binding and uptake sites. The patchy organization of µ-opiate receptors as first reported by Pert et al. in 1976 provided an excellent and constant marker of the patch compartment in the embryonic, young and adult rat, and throughout both the dorsal and ventral striatum43. Another marker is acetylcholinesterase (AChE) that is more intensively detectable in the matrix. The AChE-[page 17↓]poor striosomes form three-dimensional labyrinths and correspond to the µ-opiate receptor patches44.

With respect to the neurochemical differences of patch and matrix, it seems plausible to assume functional heterogeneity. Separate sets of dopamine midbrain neurons target the two compartments, respectively. Dopaminergic afferents from part of the SN project to striosomes, whereas much of cell group A8 (retrorubral area) projects to the matrix45. It suggests a differential regulation between striosome-related versus matrix-related basal ganglia circuits.

Outputs are as distinct as inputs: neurons from the matrix project to the SNr via direct and indirect pathway. In contrast, patches preferentially project to dopaminergic nigrostriatal neurons in the SNc (Fig. 6). They may either inhibit dopaminergic neurons in SNc, or disinhibit them through inhibition of GABAergic interneurons.

The local axon collaterals of spiny neurons as well as their dendrites are restricted to the compartment of their parent cell body46. In contrast, dendrites of cholinergic interneurons cross the compartments. Their axon collaterals are denser in the matrix, whereas cell bodies and dendrites are approximately equally distributed47.

1.2.3. Understanding the basal ganglia circuitry

Observations of humans afflicted with degenerative diseases of basal ganglia were essential for the understanding of the basic circuitry. Diseases related to basal ganglia include Parkinson’s disease (PD), Tourette’s syndrome, schizophrenia, Huntington’s chorea and hemiballism.

Parkinson’s disease is the most frequent disorder of the basal ganglia caused by loss of dopaminergic neurons and resulting depigmentation in the substantia nigra. Typical features are extreme underactivity, poverty of movement (hypokinesia), infrequency of swallowing, abnormal postural reflexes, absence of arm swing in walking and reduced velocity of movement (bradykinesia) up to inability to walk forwards (freezing). Also firm and tense [page 18↓]muscles (rigidity) and a low-frequency resting tremor are seen in many patients, often beginning in one limb and spreading to one side48.

Both the features of the disease and the success of available therapies can be understood to a remarkable degree, when the neuronal circuits are considered, on which movements are based. Cerebral cortex and basal ganglia are the main structures of movement planning and program selection (Fig. 6).

Neuronal transmission through the basal ganglia is performed in circuits. Beginning at the cerebral cortex most information enters the basal ganglia at the level of the striatal projection neurons. The axons form asymmetrical excitatory glutamatergic synapses primarily on the head of dendritic spines. Afterwards, the circuit continues to the GABAergic neurons located in GPi (entopeduncular nucleus in rats) and SNr. This so-called direct pathway facilitates a certain movement. Another half of spiny projection neurons make use of an alternative route also including GPe and STN (indirect pathway). It suppresses a movement in contrast to the direct pathway. Glutamatergic projections from the subthamalic nucleus to GPi/SNr are thought to excite the GPi, consequently increasing pallidal inhibition of the thalamus49. Both GPe and STN influence all striatal structures and are affected by cortical and thalamic projections50.

From GPi/SNr the circuit proceeds to the output nuclei of basal ganglia, i.e. ventral anterior, ventromedial and mediodorsal thalamic nuclei and superior colliculus. These structures are under tonic GABAergic inhibition by pallidal and nigral spontaneous firing. Finally, the cortex receives excitatory input from this complex circuit.

This model of circuitry contains multiple parallel basal ganglia-thalamocortical circuits (motor, oculomotor, dorsolateral prefrontal, lateral orbitofrontal and anterior cingulate). Depending on the cortical input, three striatal territories were distinguished (motor, associative and limbic or ventral striatum)51. Neither the striatal subdivision nor the specific loops are supposed to be considered as strictly bordered or without neuronal interactions. Recent studies52 suggest an information flow through the basal ganglia not by segregated parallel circuits, but by an ascending spiral interconnecting different functional regions.


[page 19↓]

Information flow proceeds from limbic to associative, further to motor domains of the striatum via the midbrain dopamine cells. Each component disinhibits the next step in the spiral and inhibits itself by a feedback loop.

Fig. 6: Schematic representation of the neuronal circuitry underlying basal ganglia function. For details see text.

1.2.4. Striatal regulation of movements

Basal ganglia seem to play a selecting rather than initiating role in motor activity. Neuronal activity in the striatum is thought to be a “movement template” whose occurrence could become facilitated or suppressed. To perform a certain movement, inhibition of GPi/SNr is essential (Fig. 6). Therefore, activation of the direct pathway as well as suppression of the indirect pathway is required to interrupt the tonic pallidal/nigral inhibition. Consequently, [page 20↓]this disinhibition of thalamic neurons would result in excitation of the cortex and the appearance of motor action.

The efficacy of the direct and indirect pathway is regulated by two striatal key structures: the patch compartment and the cholinergic interneuron.

The dopaminergic neurons of the substantia nigra pars compacta (SNc) target spines of striatal projection neurons and are thought to control the cortical inputs synapsing on the same spine head53. The SNc neurons facilitate the direct pathway via D1 receptors and inhibit the indirect pathway via D2 receptors, thus synergistically favoring movements. The dopaminergic SNc neurons are controlled by the patch compartment, emphasizing the striking functional differences between the two striatal areas.

In contrast, giant cholinergic interneurons facilitate the indirect pathway via muscarinic M1 receptors and inhibit the direct pathway via M4 receptors. In animal models of Parkinson’s disease, persistent oscillatory activity of cholinergic interneurons can be seen instead of normal tonic firing, underlining the functional and clinical relevance of these neurons54.

1.2.5. The therapeutic dilemma

Effective therapies to treat the motor impediments in Parkinson´s disease described above have been available for a long time. Presently, therapeutic intervention in younger patients starts with dopamine agonists, which often improve the motor impediments. With progression of the disease addition of L-dopa (L-dihydroxy-phenylalanine) is necessary and very effective even in advanced states of the disease. The traditional choice of cholinergic antagonists is less important today and restricted to the relief of tremor. It remains essential, however, in order to achieve a marked improvement in dyskinesia during neuroleptic therapy. However, even an excellent therapeutic regime implies a medical dilemma: many side effects are the patient’s price for temporary motor improvement. Dryness of the mouth, blurring of vision and constipation are the first troubles in many patients making use of anticholinergics, and nausea as well as sickness are the primary problems under L-dopa therapy.


[page 21↓]

L-dopa application also causes more troublesome complications. Induction of involuntary movements such as grimacing, lingual-labial dyskinesia and restlessness, limits the use. Consequently, therapeutic skill is required to maintain the fine balance between a mobile patient and some degree of dyskinesia. Hallucinations, confusion and impairment of memory limit the advantages of cholinergic antagonists. Moreover, long-term treatment cannot prevent the advance of Parkinson’s disease, and finally, L-dopa therapy causes severe dose-response fluctuations (on-off-phenomenon) as well as uncontrollable dyskinesia, thus terminating the pharmacological management.

1.2.6. Intention of the research

Current therapies of Parkinson’s disease as described above rely on manipulation of transmission at these striatal key positions. Other targets have to be found to develop new therapeutic strategies for treating motor deficits. In principle, every step of the signal transduction cascade seems to be a possible approach for new pharmacological regimes. Special attention must be paid to those biological functions, which are performed by a variety of effector proteins in different cell types. This diversity may allow addressing individual neurons and individual neuronal circuits separately.

With respect to their enormous diversity (see chapter 1.1.3.3.), potassium channels may represent suitable targets for therapeutic intervention. The success of this type of strategy is well documented in cases like the treatment of cardiac arrhythmias and diabetes (see discussion). For the treatment of Parkinson’s disease or related disorders, however, promising effects may be expected only if the channels are selectively or at least predominantly expressed in one of the key motor structures. Only in this case, selective channel activation or inhibition may specifically influence the outcome of movements.

In a first step to throw light on the possible involvement of Kir2 channels in striatal pathways, the present study aimed to:

(1) analyze the protein distribution of all members of the Kir2 channel subfamily in the rat brain.

(2) study whether members of the Kir2 channel protein family, if present in the rat basal ganglia, display a similar or distinct basal ganglia localization.


[page 22↓]

(3) learn whether any member of the Kir2 subfamily is predominantly found at the striatal key motor structures.

It may be expected that the results of this study increase the understanding of specific interactions that take place in the striatum and ultimately determine the output of the basal ganglia. The knowledge of the detailed pattern of Kir2 channel distribution is necessary to understand electrophysiological influences at the subcellular level as well as the processing within the striatum. Consequently, the study may be helpful in further examinations to clarify how a disturbed striatal circuitry causes the symptoms of diseases of the basal ganglia.


Footnotes and Endnotes

1 Sanguinetti, M.C. 1997

2 Biggin, P.C. 2000; Minor, D.L. 1999

3 Glowatzki, E. 1995

4 Isomoto, S. 1997; Minor, D.L. 1999; MacKinnon, R. 1998

5 Reimann, F. 1999; Doyle, D.A. 1998; Taglialatela, M. 1994

6 Katz, B. 1949; Hodgkin, A.L., Huxley, A.F., Katz, B. 1952

7 Doupnik, C.A. 1995; Isomoto, S. 1997

8 Minor, D.L. 1999; Chandy and Gutman 1995; Jan, L.Y. 1997

9 Chuang, H. 1997; Isomoto, S. 1997, Reimann, F. 1999

10 Doupnik, C.A. 1995; Krapivinsky, G. 1995

11 Ishii, M. 1997; Hibino, H. 1997, Ruppersberg, J.P. 2000

12 Sakura, H. 1995

13 Yang, J. 1995; Minor, D.L. 1999

14 Raab-Graham, K.F. 1998; Fakler, B. 1996

15 Reimann, F. 1999; Raab-Graham, K.F. 1998

16 Isomoto, S. 1997; Kubo, Y. 1993

17 Isomoto, S. 1997

18 Kubo, Y. 1993

19 Lopatin, A. 1994; Fakler, B. 1995

20 Doupnik, C.A. 1995; Ruppersberg, J.P. 2000

21 Yamashita, T. 1996

22 Yang, J. 1995; Lu, Z. and MacKinnon, R. 1994; Stanfield, P.R. 1994, Lee, J.-K. 1999, Kubo, Y. and Murata, Y. 2001

23 Nishida, M. and MacKinnon, R. 2002

24 Isomoto, S. 1997; Reimann, F. 1999; Qu, Z. 1999/2000; Cohen, N.A. 1996; Zhu, G. 1999; Huang, C.-L. 1998; Coulter, K.L. 1995; Namba, N. 1996; Wischmeyer, E. 1998

25 Chen, H. 1998

26 Doupnik, C.A. 1995; Perillan, P.R. 2000; Raab-Graham, K.F. 1998; Falk, T. 1995; Karschin, C. 1999; Koyama, H. 1994

27 Morishige, K.-I. 1993; Horio, Y. 1996; Karschin, C. 96; Bredt, D.S. 95; Töpert, C. 98; Falk, T. 95

28 Karschin, C. 1999; Karschin, C. 1996; Horio, Y. 1996; Falk, T. 1995

29 Töpert, C. 1998; Töpert, C. 2000

30 Raab-Graham, K.F. 1998; Cohen, N.A. 1996; Perillan, P.R. 2000

31 Kawaguchi, Y. 1995, Wilson, C. 1998

32 Parent, A. 2000; Gerfen, C. 1992; Aizman, O. 2000

33 Penny, G.R. 1986; Kita, H. 1988; Groenewegen, H.J. 1996; Kawaguchi, Y. 1990

34 Wilson and Kawaguchi 1996; Nicola 2000

35 Nisenbaum, E.S. 1995; Wilson, C. 1993

36 Nicola, S.M. 2000; Mermelstein, P.G. 1998

37 Bolam, J.P. 1984

38 van Vulpen, E.H.S. 1998; Kawaguchi, Y. 1995

39 Calabresi, P. 2000; Wilson, C.J. 1998

40 Wilson, C. 1998; Kawaguchi, Y. 1995

41 Bennett, B.D. 2000; Watanabe, K. 1998; Kawaguchi, Y. 1995; Raz, A. 1996

42 Trytek, E.S. 1996; Wilson, C. 1998; Bolam, J.P. 1988

43 Pert, C.B. 1976; Herkenham, M. and Pert, C.B. 1981; Desban, M. 1993; Caboche, J. 1991

44 Graybiel, A. 1990

45 Dahlstrom, A. and Fuxe, K. 1964; Graybiel, A. 1990; Eblen, F. 1995; Surmeier, D.J. 1996

46 Gerfen, C. 1992; Wilson, C. 1998

47 Kawaguchi, Y. 1995; Graybiel, A. 1990

48 Bergmann, H. 1998

49 Groenewegen, H.J. 1996; Kawaguchi, Y. 1990

50 Parent, A. 2000; Parent, A. 1990; Groenewegen, H.J. 1996

51 Alexander, G.E. 1986, 1990; Baev, K.V. 1995; Joel, D. and Weiner, I. 2000; Flaherty, A.W. 1991

52 Haber, S.N. 2000

53 Cepeda, C. 1998; Graybiel, A.M. 1990; Bergmann, H. 1998; Gerfen, C.R. 1989; Groves, P.M. 1995

54 Raz, A. 1996



© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 3.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
15.06.2004