Calcium channels involved in neurotransmitter release at adult, neonatal and P/Q-type deficient neuromuscular junctions (Review)

Calcium channels involved in neurotransmitter release at adult, neonatal and P/Q-type deficient neurFull Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=imbc20
Molecular Membrane Biology
Calcium channels involved in neurotransmitter release at adult, neonatal and P/Q-type deficient neuromuscular junctions (Review)
Francisco J. Urbano, Marcelo D. Rosato-Siri & Osvaldo D. Uchitel
To cite this article: Francisco J. Urbano, Marcelo D. Rosato-Siri & Osvaldo D. Uchitel (2002) Calcium channels involved in neurotransmitter release at adult, neonatal and P/Q-type deficient neuromuscular junctions (Review), Molecular Membrane Biology, 19:4, 293-300, DOI: 10.1080/0968768021000035087
To link to this article: https://doi.org/10.1080/0968768021000035087
Published online: 09 Jul 2009.
Submit your article to this journal
Article views: 343
View related articles
Summary
Different types of voltage-dependent calcium channels (VDCCs) have been recognized based on their molecular structure as well as their pharmacological and biophysical properties. One of these, the P/Q type, is the main channel involved in nerve evoked neurotransmitter release at neuromuscular junctions (NMJs) and many central nervous system synapses. However, under parti- cular experimental or biological conditions, other channels can be involved. L-type VDCC presence at the NMJ has been demonstrated by the contribution to the perineural calcium currents (ICa) at adult mice Bapta-loaded NMJs. This is probably a result of a reduction in Ca
2+ inactivation. The L-type current was
not coupled to neurotransmitter release, but became coupled, as demonstrated by the release of acetylcholine, after the inhibition of serine/threonine protein phosphatases with okadaic acid (OA). Thus, under these conditions, L-type channels were unmasked at Bapta- but not at Egta-loaded NMJs. This suggests that the speed, not the capacity, of the calcium chelator was decisive in preventing Ca
2+ -inactivation and facilitating the
contribution to neurotransmitter release. At neonatal rat NMJs, N-type VDCCs were involved early during development whereas P/Q-type VDCCs play a main role at all stages of development. Furthermore, P/Q-type VDCCs were more efficiently coupled to neurotransmitter release than N-type VDCCs. This difference could be accounted for by a differential location of these channels at the release site. Neuromuscular transmission in P/ Q-type calcium channel knock out ataxic mice jointly depends on both N-type and R-type channels and shows several altered properties including low quantal content. Thus, calcium chan- nels may be recruited to mediate neurotransmitter release with a functional hierarchy where the P/Q channel seems to be the channel most suited to mediate exocytosis at NMJs.
Keywords: Neuromuscular junction, development, Bapta, Egta, calcium channels.
Abbreviations: AChRs, Acetylcholine recepotors; EPP, endplate potentials; MEPP, miniature endplate potentials, o-AgaIVA, o- Agatoxin IVA; o-ConoGVIA, o-Conotoxin GVIA; Bapta-AM, 1,2- bis(2-aminophenoxy)ethane-N,N,N,N-tetraaceticacid acetoxymethyl ester; Egta-AM, Ethylene glycol-bis(b-aminoethyl ether)-N,N,N’,N’- tetraacetic acid acetoxymethyl ester; OA, okadaic acid; DMSO, dimethylsuphoxide; [Ca
2+ ], calcium concentration; VDCCs, voltage-
dependent calcium channels; dTC, d-tubocurarine; TEA, tetraethy- lammonium; 3,4-DAP, 3,4-diaminopyridine.
Introduction
2+ ions to enter the cell from
extracellular space. Neuronal VDCCs have been sub-divided on the bases of their electrophysiological and pharmacolo- gical properties into low voltage activated or T-type channels and high voltage activated (HVA) channels. HVA channels are a class that include the L-, N-, P/Q- and R-types. The HVA channels have mainly been characterized by their different sensitivity to pharmacological modulators and inhibitory toxins (Uchitel 1997). High voltage-activated Ca
2+
channels consist of an a1 sub-unit that forms the core of the channel, in addition to b, a2 ± d, and possibly g sub-units that modulate the functional properties of the a1 sub-unit. Molecular cloning experiments have identified 10 different genes encoding the a1 sub-unit (Cav 1.1 ± Cav 3.3, for review, Ertel et al. 2000). Moreover, molecular biology studies have identified many of the structural determinants that govern the functional pharmacological properties of the a1 sub-unit, as well as sites of interaction with modulatory proteins (Birnbaumer et al. 1998).
Expression of VDCC and their role in neurotransmitter
release at the neuromuscular junctions
Different types of VDCCs support neurotransmitter release at synapses (Reuter 1996, Catterall 1998, 2000). At mature mammalian NMJs, evoked neurotransmitter release is blocked completely by the P/Q-type channel blocker o- Agatoxin IVA (Uchitel et al. 1992, Protti and Uchitel 1993, Katz et al. 1997, Uchitel 1997).
However, other VDCCs may be coupled to neurotrans- mitter release. Various novel types of VDCCs mediating neurotransmitter release at NMJs have been reported. This has been demonstrated during reinnervation (Katz et al.
1996), during functional recovery from Botulinum toxin type- A poisoning (Santafe et al. 2000), as well as during different developmental stages (Gray et al. 1992, Fu and Huang 1994, Sugiura and Ko 1997, Rosato-Siri and Uchitel 1999, Santafe et al. 2001, 2002).
L-type VDCCs at adult neuromuscular junctions
The presence of L-type VDCCs at both the neuronal cell bodies and proximal dendrites of many Central Nervous System neurons has been shown (Day et al. 1997, Plant et al.
1998, Westenbroeket al. 1998, Iwasaki et al. 2000). Previous immunocytochemical studies failed to demonstrate the binding of antibodies to the a sub-units of the L-type channels at mammalian NMJs (Day et al. 1997, Westenbroek et al.
1998). However, L-type VDCCs in adult NMJs specifically stain with antibodies against the a 1D sub-unit and co-localize
*To whom correspondence should be addressed. e-mail: [email protected] Present addresses: {Department of Physiology and Neuroscience, NYU School of Medicine, 550 First Avenue, New York, NY 10016. {Biophysics Sector S.I.S.S.A./I.S.A.S; Via Beirut 2-4; 34014 Trieste, Italy.
Molecular Membrane Biology, 2002, 19, 293±300
Molecular Membrane Biology ISSN 0968-7688 print/ISSN 1464-5203 online # 2002 Taylor & Francis Ltd
http://www.tandf.co.uk/journals DOI: 10.1080/0968768021000035087
with labelled post-synaptic AChR’s (R. Pagani and O. D. Uchitel, unpublished work). Nevertheless, further studies are currently taking place in the laboratory to clarify whether these AChR are located pre- or post-synaptically on the NMJs.
Electrophysiologicalexperiments searching for an effect of L-type channel blockers on neuromuscular transmission at the adult NMJs were negative. However, considering that L- type channels are strongly modulated by both intracellular [Ca
2+ ] and phosphorylation (Catterall 1998, 2000), experi-
ments where the nerve terminal was loaded with cell-
permeant Ca 2+
buffers were performed. Levator auris nerve-musclepreparationswere incubated in
the presence of either Bapta-AM or Egta-AM cell permeant calcium chelators (Urbano and Uchitel 1999, Urbano et al.
2001a). It is well known that these AM forms, once into the cell and after the action of cytoplasmic esterases, become active intracellular calcium buffers (Tsien 1981). In these experi- ments, cell-permeant Ca
2+ buffers were successfully loaded
into the terminals, as suggested by the drastical decrement observed in the K
+ -evoked MEPP frequency at both Bapta
and Egta loaded NMJs (figure 1(a)). Similar results have been described (Tanabe and Kijima 1992; Losavio and Muchnik 1997, Protti and Uchitel 1997, Rosato-Siri et al. 2002).
Pre-synaptic L-type perineural currents at Bapta-AM-
incubated NMJs
Perineural pre-synaptic currents were recorded on the levator auris longus muscle as previously described (Urbano and Uchitel 1999, Urbano et al. 2001a). Perineural calcium currents (ICa) are voltage drops, across the resistance of the perineurium, which are proportional to the conductance changes that take place at synaptic terminals (Mallart 1985a,b). In the presence of dTC and the K
+ channel
blockers TEA and 3,4-DAP, a recording electrode placed into the perineural bundles of NMJs showed the capacitance artefact followed by downward and upward deflections. Here, the downward component reflects the Na entry into the last heminodes while the upward deflection corresponds to the Ca
2 influx (ICa) from the terminal (Mallart 1985a,b). An L-type
component of perineural ICa was observed after the Bapta loading of adult mice NMJs (figure 1(b)). Furthermore, this component remained after P/Q-type VDCCs were fully blocked (figure 1(c)), revealing the existence of a nitrendi- pine-sensitive, L-type component of pre-synaptic perineural ICa. Surprisingly, nitrendipine did not affect neurotransmitter release under the same conditions (Urbano and Uchitel 1999). Next, to determine whether L-type VDCCs may become coupled to neurotransmitter release when serine/ threonine phosphatases are inhibited (Arenson and Gill 1996), the effect of okadaic acid was studied (i.e. a potent serine/threonine protein phosphatases inhibitor) on neuro- transmitter release at Bapta-loaded NMJs.
Serine/threonine phosphorylation couples L-type VDCCs to
neurotransmitter release
Protein phosphorylation is a fast process that plays a cardinal role in neuronal excitability, regulation, plasticity and excitotoxicity (Hunter 1995, Tokuda and Hatase 1998,
Turner et al. 1999). The balance between protein kinases and protein phosphatases regulates phosphorylation. In fact, when protein phosphatases are inhibited, numerous proteins of the cell become `trapped’ in their phosphorylated forms.
In particular, L-type VDCC function is regulated by serine/ threonine phosphorylation (Artalejo et al. 1992, Frace and Hartzell 1993, Haby et al. 1994, Schuhmann et al. 1997, Catterall 1998, 2000, Chik et al. 1999, Sena et al. 1999, 2000). Also, many other proteins involved in neurotransmitter release are regulated by phosphorylation (Tokuda and Hatase 1998, Turner et al. 1999).
The effects of okadaic acid (OA, 1mM) were studied on evoked neurotransmitter release mediated solely by L-type VDCCs (i.e., after blocking P/Q-type VDCCs) at Bapta-loaded NMJs. figure 2(a) shows that the bath application of OA significantly reduced the number of failures (i.e. incremented
Figure 1. L-type VDCCs in adult Bapta-loaded neuromuscular junctions. (a) Bar diagram showing that 10 mM K
+ -evoked MEPP
frequency at DMSO-, Bapta-AM and Egta-AM incubated NMJs. *p50.001 compared with DMSO-incubated NMJs. Each bar represents the mean+SEM of at least 15 fibres per NMJ (n=12 muscles). (b) Effect of 10 mM nitrendipine on perineural ICa of Bapta- AM-incubated NMJs. (c) Effect of nitrendipine on perineural ICa,
recorded in the presence of 120 nM o-Aga IVA from a Bapta-AM- incubated NMJ. Note how in both (b) and (c), nitrendipine effects were reversible. Perineural ICa were recorded in the presence of dTC (30 mM), TEA (10 mM) and DAP (250 mM), and correspond to the average of three currents elicited by nerve stimulation every 30 s. Statistical significance (p-values) was evaluated by the two-tailed Welch’s t-test (for unpaired values and not assuming equal variances). Data reproduced with permission from Urbano and Uchitel (1999).
294 F. J. Urbano et al.
quantal content). Furthermore, an increment in the sponta- neous MEPP frequencywas also observed in the presence of OA (Urbano et al. 2001a). OA-dependent increments of both quantal content and MEPP frequency were abolished by the L-type VDCC blockers nitrendipine and calciseptine (Urbano et al. 2001a), indicating that these channels were finally coupled to neurotransmitter release under these conditions.
Next, the latency and rise time was studied of EPPs mediated by either P/Q- or L-type VDCCs in Bapta-loaded NMJs. Figure 2(b) shows representativeEPP latencies when only P/Q-type or L-type VDCCs were available to participate in neurotransmitter release. Interestingly, delay histograms corresponding to L-type VDCC mediated EPP were more scattered than those corresponding to the P/Q-type VDCC mediated EPP. Differences in the EPP rise time were also observed, showing P/Q-type VDCC mediated EPP to be faster than those mediated by L-type VDCCs (Urbano et al. 2001a). These results suggest that L-type VDCCs are located further from the active vesicle release zones than P/Q-type VDCCs. However, voltage-clamp experiments are necessary to study this point more carefully.
Finally, to look for the serine/threonine protein kinases that mediated the OA-dependent effects shown above, H-7 was used, a potent inhibitor of both cAMP-dependent protein kinase and protein kinase C (Hidaka et al. 1984). Indeed, H-7 was able to significantly reduce both of the OA-dependent effects (Urbano et al. 2001a), which is consistent with the hypothesis that OA exerted its effects by enhancing the cellular serine/threonine phosphorylation.However, the exact
type of protein kinase involved and its role on neurotransmit- ter release is currently under study in the laboratory.
Possible role of the L-type VDCC at adult NMJs
According to the results, Bapta was required to unmask pre- synaptic L-type VDCCs. Moreover, the coupling of these channels to neurotransmitter release was possible after the inhibition of serine/threonine phosphatases. These findings, combined with the well-known involvement of L-type VDCCs on neurotransmitter release during both development and processes of recovery after injury (Fu and Huang 1994, Katz et al. 1996, Sugiura and Ko 1997, Rosato-Siri and Uchitel 1999, Santafe et al. 2001, 2002), suggest that L-type VDCCs could be mediating a wide range of cellular functions. Indeed, some examples supporting this idea have been described; L- type VDCC function seems to mediate the neuroprotective effect of vitamin D (Brewer et al. 2001) and gene expression (Murphy et al. 1991). Also, in hippocampal neurons, Ca
2+
influx through post-synaptically located L-type VDCCs causes the translocation of calmoduline from the cytoplasm to the nucleus, leading to the activation of transcription factors (Deisseroth et al. 1998).
Voltage-dependent calcium channels at the neonatal rat neuromuscular junction
Many physiological and morphological changes have been described during the course of neuromuscular junction
Figure 2. L-type VDCCs are coupled to neurotransmitter release after serine/threonine phosphorylation in the presence of Bapta. (a) Superimposed consecutive traces during nerve stimulation in the presence of o-Aga IVA, 30 min after bath application of 1 mM OA (o-Aga IVA+OA) and washout (OA-washout) from the same Bapta-loaded NMJ. Membrane resting potential, Vm, was 770 mV. Note the decrement in failures (i.e. increment on the EPP number) when OA was present. (b) EPP latency distribution histograms corresponding to both P/Q-(&) and L- mediated EPP (&). Illustrative EPPs at both conditions are shown as an inset. The histograms were fitted to a single Gaussian distribution. Mean latencies+SD: P/Q-mediated, 1.90+0.24 (r
240.9); L-mediated, 2.20+0.16 (r 240.9). Both distributions were significantly different (p50.001).
Histograms of latency of the EPP of P/Q- and L-type VDCCs were fitted to a single Gaussian (r 2 40.9). A non-parametric Kruskal-Wallis ANOVA
was used to evaluate statistical significance between both histograms. Data reproduced with permission from Urbano et al. (2001a).
295Differential expression of calcium channels at NMJ
maturation (Sanes and Lichtman 1999, for review). For instance, muscle fibres were multiply innervated, displayed EPP with different peaks and lower mean quantal content and underwent changes in the pharmacology of VDCCs coupled to neurotransmitter release.
Developmental change of VDCCs
At neonatal rat neuromuscular junctions, both Ca 2+
channel blockers o-Aga IVA and o-Cono GVIA (P/Q- and N-type VDCC blockers, respectively) were capable of inhibiting neurotransmitter release, with o-Aga IVA being more effective (Rosato Siri and Uchitel 1999). This pharmacologi- cal profile, however, was clearly dependent on the develop- mental stage of the animal. VDCCs of the N-type were involved in synaptic transmission early in development (embryos and 0 ± 4 day-old rats) while P/Q-type VDCCs played a major role at all stages of development (figure 3). The temporary participation of the N-type channels in synaptic transmission could reflect a role of these channels in events related to newly formed NMJs and synapse elimination.
Differential coupling of VDCCs
The involvement of two types of Ca 2+
channels, P/Q and N, rather than P/Q alone, provided an interesting model to test whether the different pathways for Ca
2+ entry were coupled
to neurotransmitter release with equal effectiveness and also whether a particular arrangement of these VDCC sub-types and the space-time overlapping of calcium microdomains contributed to influence short-term plasticity processes (Zucker 1996). It is well known that fast neurotransmitter release depends on the extracellular concentration of Ca
2+
(at low concentrations) in a non-linear relationship where the amplitude of the end-plate potentials varies as the 4
th power
of Ca 2+
concentration (Dodge and Rahamimoff 1967). This has been interpreted as a cooperative action of four Ca
2+
ions in the release of each neurotransmitter vesicle. In mammals, this relationship has also been described as non- linear, with a slope close to 3.0 (Cull-Candy et al. 1980).
At neonatal NMJs, Ca 2+
dependence of nerve-evoked neurotransmitter release produced a supralinear increment of quantal content together with a higher cooperative degree when release was mediated by P/Q- rather than N-type channels (figure 4). This difference has been interpreted to be due to a difference in pre-synaptic distribution of the channels, with the P/Q-type VDCCs located more closely to the Ca
2+ sensor than the N-type (Rosato Siri et al. 2002). If
this hypothesis is correct, Ca 2+
entering via N-type Ca 2+
channels would diffuse along a greater distance before reaching the Ca
2+ sensor, increasing the probability of Ca
2+
being bound by a calcium bufferÐparticularly by a fast one like Bapta (Adler et al. 1991). Indeed, at the adult and late neonatal mouse neuromuscular junction, loading the nerve terminal with Bapta did not affect P/Q-type VDCCs mediated neurotransmitter release (Urbano and Uchitel 1999). On the other hand, at early stages of development (0 ± 4 post-natal days), loading the nerve terminals with Bapta strongly reduces synaptic transmission. This effect was related to
the capacity for buffering Ca 2+
influx specifically through N- type VDCCs (Rosato Siri et al. 2002, figure 4).
Hypothetical model
A diminished density of VDCCs at the active site per se may result in less efficient neurotransmitter release and low co- operativity. However, this is unlikely since reducing the number of functional channels in the adult wild-type NMJ did
Figure 3. Voltage-dependent calcium channels at neonatal rat neuromuscular junctions. (a) Plots show the correlation between the percentage of inhibition (of quantal content) and the time of development (age of the animals) in the presence of either o-CgTx GVIA (1 mM; closed circles) or o-Aga IVA (100 nM; open circles). Negative numbers of age represent embrionary days. Each value represents the mean+SE mean of the data pooled from a nerve- muscle preparation (at least 12 NMJs per muscle). The lines represent the best linear fit of the data for o-CgTx GVIA (r
2 =0.8,
p50.01) and o-Aga IVA (r 2 =0.37, p=0.27). (b) Bar diagrams show
the effect of the drugs on quantal content, expressed as a percentage of the control value. The toxins o-Aga IVA (100 nM) and o-CgTx GVIA (1 ± 5 mM) significantly reduced the evoked response in embryonic and 0 ± 4 day-old rats. On the other hand, in 5 ± 11 day-old rats, o-Aga IVA (100 nM) maintained its strong effect on the evoked response shown at early stages while o-CgTx GVIA (1 ± 5 mM) lacked any effect. Control fibres were assayed in a low Ca
2+ -high Mg
2+ (0.7 ± 1 mM/5± 8 mM) saline solution. Treated
fibres were assayed in the same muscles after 1-h incubation with the respective drug. Each column represents the mean+SE mean of data pooled from 2 ± 5 nerve± muscle preparations (at least 12 NMJs per muscle). m was calculated by the failure method. Stimulation frequency was 0.5 Hz. **p50.0001; *p50.05 compared with the values obtained in the same muscles before addition of the calcium channel blockers. Data from Rosato-Siri and Uchitel (1999).
296 F. J. Urbano et al.
not mimic the alterations in co-operativity observed when neurotransmitter release was mediated by N-type channels (Urbano et al. 2001b).
One simple way to explain the differences in the co- operativity is to assume the existence of different Ca
2+
sensors (of SNARE complexes), each one with its own distinct Ca
2+ co-operativity, interacting with each type of
channel. However, considering that both types of calcium channels are present in the same active site, this possibility is unlikely.
The difference in the calcium dependence of neurotrans- mitter release between P/Q- and N-type calcium channels may result from the different interaction and affinity of VDCCs to synaptotagmin and the SNARE complex (i.e.
syntaxin, SNAP-25, VAMP-synaptobrevin) (Catterall 1998, 2000), leading to a different spatial relationship between the calcium sensor of the SNARE complex and the VDCCs (e.g. being the P/Q- closer to the sensor than the N-type; figure 5).
In this model, during development, P/Q-type VDCCs are able to `displace’ N-type ones far from the SNARE complex, thus reducing the calcium sensitivity of N-type mediated
release. The binding at the synaptic protein interaction sites (synprint) of P/Q and N-type VDCCs to SNARE proteins is thought to be…