Sex differences in the acetylcholine receptor kinetics of postnatal and denervated rat muscle

Abstract

1. Single-channel recording from visualised endplates in freshly dissociated muscles from postnatal and denervated rat muscle revealed the presence of a low conductance, fetal type of acetylcholine receptor.

2. Kinetic analysis showed a main component in the burst durations with a mean of 10.8 ± 2.7 ms (n = 29). Receptors from female rats had an additional 27.3 ± 5.5 ms (n = 5) kinetic component which was found in one-third of the 15 female endplates.

3. Recordings from male and female denervated muscles gave more homogeneous kinetics with single time constants of 7.2 ± 1.3 and 7.4 ± 1.3 ms, respectively.

4. It is concluded that the acetylcholine receptor channels present during early development are different from those of denervated muscle.

The acetylcholine (ACh) receptor at the neuromuscular junction is present in two molecularly distinct forms that are developmentally regulated. The adult form, composed of the α₂βεδ subunits, replaces the fetal form, composed of the α₂βγδ subunits, during the first 2 weeks of postnatal development (Mishina et al. 1986).

The fetal form of the receptor has been widely studied in preparations from Xenopus (Kullberg & Owens, 1986), calf (Mishina et al. 1986), mouse (Franke et al. 1992) and rat ‘myoballs’ (Hamill & Sakmann, 1981). The recombinant α₂βγδ receptor forms channels with a conductance of 41 pS and a mean open time of 7.6 ms (Sakmann & Witzemann, 1989). A similar conductance has been measured in native preparations. The kinetics of the receptor, however, have given conflicting results among different preparations. Firstly, kinetic measurements have given variable results in a single preparation. Secondly, there is no agreement between the decay time constant of the synaptic current and the mean open duration of the channel. Miniature endplate currents from young rat muscle have a decay time constant of 4.8 ms (Sakmann & Witzemann, 1989). Fast application of ACh to outside-out patches extracted from young rat muscle (postnatal day (P)5-14), however, had an offset-decay time constant (Colquhoun et al. 1992) that varied between 13 and 3 ms (Villarroel & Sakmann, 1996). A large variability in the opening rate has also been described (Maconochie & Steinbach, 1998).

An attractive hypothesis ascribes this variability to multiple channel types present in the fetal and denervated endplate. Consistent with this idea, a splice variant of the γ-subunit that predicts a peptide lacking part of the extracellular portion has been described (Villarroel, 1999).

Here, we investigate the origin of the kinetic variability using single-channel recording of freshly dissociated muscle fibres. The mean open time has at least two kinetic components: a fast one, partially unresolved, of 0.1 ms presumably arising from activation of a single ACh molecule, and a main component around 7 ms. This main component is highly variable in neonatal muscle, but constant in denervated muscle. In some endplates from female rats we also found a 27 ms component. It is concluded that the embryonic acetylcholine receptor channel has complex kinetics consistent with the presence of several channel types. In contrast, receptors from denervated muscle are more homogeneous. In addition, these results suggest that variation in channel properties may depend on sex.

METHODS

Dissociation of muscle fibres

Neonatal rats (Wistar) were killed with an overdose of pentobarbital. Single muscle fibres were obtained from flexor digitorum brevis muscle by enzymatic dissociation (Bekoff & Betz, 1977; Villarroel & Sakmann, 1996). Briefly, muscles were incubated for 1 h at room temperature in a solution containing 1 mg ml⁻¹ collagenase. Denervated muscles required a concentration of 2 mg ml⁻¹ to dissociate into individual fibres. Dissociated muscle fibres were then maintained in extracellular solution (mm: NaCl 150, KCl 4, CaCl₂ 1.8, MgCl₂ 1, Hepes 10; pH 7.2), with 500 nm tetrodotoxin (Sigma, St Louis, MO, USA). For measurements of end-plate patch current, individual muscle fibres were transferred to an experimental chamber of 0.5 ml volume. All experiments were done at 22-24 °C within 3 h after muscle dissociation.

Denervations

Adult Wistar rats were anaesthetised by an intraperitoneal injection of 30-60 mg kg⁻¹ sodium pentobarbital (MTC Pharmaceutical, Cambridge, Ontario, Canada). Deep anaesthesia was assessed by the absence of papillary reflexes. Denervations were performed by sectioning the right sciatic nerve and ligating the proximal segment to prevent re-innervation. Two weeks after denervation the animals were killed with an overdose of pentobarbital, and the muscles dissected for electrical recordings. All surgical procedures were performed according to the guidelines of the University Committee on Laboratory Animals, protocol number 99-030.

Current measurements

Recordings from cell-attached patches were made using standard methods (Hamill et al. 1981). Pipettes had a tip resistance of 5-8 MΩ when filled with rat Ringer solution (mm: NaCl, 135; KCl, 5.4; CaCl₂, 1.8; Hepes, 5; pH 7.2). Ringer solution containing 0.1-1 μm acetylcholine was used as a pipette solution. Endplates were identified using differential interference contrast (DIC) optics as described (Villarroel & Sakmann, 1996).

Data analysis

Patch currents were recorded with an EPC-7 amplifier (List, Darmstadt, Germany), filtered at 3 kHz, digitised using a Cambridge Electronic Design interface (CED micro1401) using the Spike2 program (version 2.29), and stored on a Pentium computer. Single-channel current was recorded at -100 mV from the reversal potential in consecutive segments of 5-20 min duration. Amplitude and duration of the single-channel events were measured using a semi-automatic procedure implemented in IgorPro (WaveMetrics, OR, USA) and run on a Power Macintosh computer. Firstly, a whole-point amplitude histogram was constructed to determine the minimum between the open and closed amplitudes, usually one-half of the maximum amplitude. During the detection, the program skips transitions shorter than 0.277 ms and connects the adjacent events. After each transition had been inspected and accepted, the current mean amplitude was determined and plotted as a function of the corresponding duration (Fig. 1B). In the case of patches with two amplitude types, as occurs late during postnatal development, events of small amplitude were selected. To determine the burst length, open events interrupted by a shut shorter than a time to which we refer as t_crit were concatenated. Values of t_crit were calculated between the last and the previous component of the shut-time distributions to produce an equal number of misclassified events (Colquhoun & Sakmann, 1985). Time constants were determined by fitting of exponential functions to the burst-length distribution, and to bin heights of logarithmic histograms as described in Sigworth & Sine (1987). No correction for missing events was performed. To decide whether two or three time constants were required to fit a distribution, the logarithm of likelihood ratio (LLR) method was used (Horn, 1987). This method rejects three time constant models when the improvement of the fit is not sufficient for the increase in the number of parameters. Data are reported as means ±s.d., unless otherwise specified.

Figure 1. Single-channel conductance of AChR from a male P10 muscle.

A, single-channel recording showing events from fetal and adult forms of the channel. B, separation of the events according to conductance.

RESULTS

Sex differences in the kinetics

Single-channel recordings of endplates from male muscle (P10) showed the characteristic fetal type of channel with a single-channel conductance of 40 pS, and openings that lasted several milliseconds (Fig. 1A). The distribution of the burst lengths revealed two kinetic components, which were more evident when represented in a logarithmic histogram: a slow one of 10 ms and a poorly resolved fast one in the submillisecond range (Fig. 2A). We attribute this fast component either to receptors activated by a single ACh molecule or to fast adult type receptors which have been filtered by the acquisition system (Fig. 1B).

Figure 2. Sex differences in the kinetic components of the fetal ACh receptor.

A, probability density function (PDF) histogram of burst length from a male endplate. The kinetic components are τ₁= 0.2 ms (20.1 %), τ₂= 6.3 ms (79.9 %), 1155 events shown. B, PDF histogram of burst length from a female endplate. The kinetic components are τ₁= 0.4 ms (31.1 %), τ₂= 9.9 ms (39.6 %), and τ₃= 27.1 ms (29.3); 1898 events shown.

In contrast, endplates from female muscles frequently showed very long openings, some of them lasting up to 100 ms. In this case, the LLR method favoured three kinetic components to describe the burst lengths. The fit to the logarithmic histogram revealed an extra component of 27 ms duration (Fig. 2B). This extra component, which was found in 5 out of 15 endplates, accounted for 22 ± 15 % of the total openings (Table 1). In one endplate, we found channel events exclusively of the long type. The time distribution of these channel events revealed, in addition to the submillisecond component, a single component of 18.2 ms (824 events) and no sign of the 10 ms component. The presence of the third kinetic component was significantly associated with sex (χ²= 10.34, P = 1.3 × 10⁻³).

Table 1.

Burst lengths of neonatal and denervated ACh receptors

	Time constant (ms)			Relative area
	τ1	τ2	τ3	A1	A2	A3
Female neonatal	0.3 ± 0.2 (15)	9.7 ± 2.5 (15)	27.3 ± 5.5 (5)	29 ± 13 (15)	61 ± 19 (15)	22 ± 15 (5)
Male neonatal	0.3 ± 0.2 (29)	10.8 ± 2.7 (29)	—	33 ± 16 (29)	64 ± 20 (29)	—
Female denervated	0.5 ± 0.5 (14)	7.4 ± 1.3 (14)	—	48 ± 23 (14)	49 ± 23 (14)	—
Male denervated	0.2 ± 0.1 (8)	7.2 ± 1.3 (8)	—	38 ± 23 (8)	56 ± 31 (8)	—

Time constants and relative areas of burst length distributions. The data are means ± S.D. with number of fibres shown in parentheses.

Developmental changes of the kinetic subtypes

To examine variation of this extra kinetic component during development, we recorded single-channel currents from muscle endplates for several postnatal days. Figure 3 shows the main kinetic component as a function of the age of the animal. Receptors from male muscle showed a single kinetic component that varied between fibres in the range 5.7-16.0 ms (mean 10.8 ± 2.7 ms). Endplates with channels in the 5 ms range were seen in the first, but not in the second week of postnatal development. Channels from female endplates also showed a variable main kinetic component (9.7 ± 2.5 ms), but in addition showed an extra long kinetic component. This component was seen in the first as well as in the second week of postnatal development.

Figure 3. Mean burst length during the early development.

Kinetic components derived from histograms similar to that of Fig. 2 plotted against postnatal days. The mean values during development were τ₂= 10.8 ± 2.7 and τ₂= 9.7 ± 2.5 ms for male muscle (A) and female muscle (B), respectively.

Figure 4. The long component is absent in denervated muscle.

Burst length distribution from a receptor recorded at the synaptic site in a female denervated rat. The time constants were τ₁= 1.0 ms (24.7 %) and τ₂= 7.4 ms (75.3 %); 1733 events shown.

As in the case of the main kinetic component of the female rat muscle, the components of the male muscle are probably heterogeneous. If the time constants of the individual components are close to 10 ms, they are difficult to separate. For unknown reasons, we were more successful in obtaining recordings from male than female muscle endplates.

Comparison to denervated muscle

Since the ACh receptor from female endplates showed an extra long kinetic component, we examined the channel subtypes expressed in female denervated muscle to determined whether this sexual distinction is maintained during denervation. The results, however, indicated that this was not the case (Table 1). ACh receptors from denervated female endplates showed, in addition to the submillisecond component, a main kinetic component of 7.4 ± 1.3 ms (n = 14). The LLR test indicated that two components were sufficient to describe the kinetics in all cases. Thus, the long kinetic component, characteristic of female muscle, seems to appear only during early development.

Endplates from male denervated muscles contained homogeneous channel populations similar to those found in denervated female muscle. Single-channel recording revealed a main kinetic component of 7.2 ± 1.3 ms (n = 8). This value was not significantly different from that of the denervated female muscles. Comparison between neonatal and denervated muscle indicated that the τ₂ kinetic component of the neonatal channels (10.4 ± 2.7 ms, n = 44) was significantly slower than that of the denervated muscle (7.4 ± 1.2 ms, n = 22). The smaller value of the standard deviation indicates that receptors in denervated muscle were more homogeneous than those of neonatal muscle (F = 5.06, P = 9.4 × 10⁻⁵). Receptors from denervated muscles then differed from those of neonatal muscle in two ways: they were more homogeneous, and their kinetics were somewhat faster.

We compared the value of τ₂ among the four endplate types using Student's unpaired t test. Since the comparison involves six possible combinations, we used Bonferroni's adjustment (α= 5 %/6) to correct the confidence interval. Receptors from both female and male denervated muscles were not significantly different (P = 0.72). Receptors from male neonatal muscle were significantly different from those of both female (P = 2.0 × 10⁻⁶) and male (P = 2.4 × 10⁻⁵) denervated muscles. Similarly, the τ₂ component of receptors from female neonatal muscle was significantly different from that of male (P = 0.006), and female denervated muscle (P = 0.006).

DISCUSSION

We have shown that the fetal acetylcholine receptor has heterogeneous kinetics at the single-channel level. This partly explains the variability observed in the offset kinetics of the receptor (Villarroel & Sakmann, 1996). In addition to the intrinsic variability in τ₂, an extra kinetic component found in female muscle could also contribute to the variability in the offset kinetics. An ACh receptor population was characterised by a 27 ms kinetic component, which seems to be exclusive to female muscle, as we were unable to detect it in muscle derived from male rats. Besides the presence of the 27 ms component in female muscle, the τ₂ component was highly variable. These two unrelated factors are likely to contribute to the kinetic variability of neonatal channels. In contrast, receptors from denervated muscle showed more homogeneous kinetics with more consistent values for the time constants. The single-channel conductance showed no apparent variability in ACh receptors derived from male or female muscle.

A simple explanation for these findings is to assume that fetal muscle expresses more than one receptor subtype. Messenger RNA encoding the γ-subunit from rat muscle can undergo alternative splicing to generate several subunit forms (Villarroel, 1999). One of these mRNA forms, γ3, predicted a complete subunit lacking part of the N-terminus. It remains to be determined whether γ3, or another form yet to be found, could generate the kinetic components described here. A splice variant of the AChR that generates long channels has been described in mice (Mileo et al. 1995). A similar subunit, if present in rat, may generate the long openings found in female muscle.

We cannot eliminate the possibility that the receptor itself might have different gating modes in male and female muscle, due to some unknown mechanism. Alternatively, heterogeneity in kinetic behaviour might result from different orders of subunit arrangement. Since the channel conductance is independent of the exact arrangement of residues around the selectivity filter (Villarroel et al. 1992), a receptor with different subunit arrangement would generate channels with similar conductance. Whether the agonist binding site in the ACh receptor is located in the interface between α- and γ- or δ-subunits (Sine et al. 1995), or is located in the α-subunit (Miyazawa et al. 1999), the gating properties depend on interactions among subunits. The channel might open differently whether the γ- or the δ-subunit is placed between the two α-subunits. This seems unlikely here, since it is difficult to imagine a mechanism by which subunits assemble differently in neonatal and denervated muscle. Another reason for heterogeneity that needs to be considered is the formation of receptors lacking one subunit. In the case of the rat receptor, this possibility seems unlikely since recombinant receptors from rat were unable to form channels lacking one subunit (Herlitze et al. 1996). This, however, was not the case for ACh receptors from mouse, where recombinant receptors do form γ-less as well as δ-less channels (Charnet et al. 1992).

The ACh receptors from denervated muscle reported here had a burst duration similar to that of receptors from rat myotubes (Jaramillo & Schuetze, 1988). Myotubes from embryonic receptors had a burst-duration distribution with two components of 7.4 ms (47 %) and 0.3 ms (53 %) consistent with the main kinetic component of 7.4 ± 1.3 and 7.2 ± 1.3 ms for denervated female and male muscle, respectively. The present results indicate that receptors from myotubes are representative of denervated muscle rather than embryonic type.

One possible significance of the kinetic differences reported here could be to make the channel faster in denervated muscle. In neonatal fibres, the long duration of mEPSCs is essential for the appearance of spontaneous contractions (Jaramillo et al. 1988). The absence of long-lasting depolarisation prevents spontaneous contraction in the adult fibre. The presence of such long-lasting channels in an adult fibre could have a deleterious effect. A mutation causing the long-channel syndrome produces myopathic changes in the endplate. Since the ACh receptor has large Ca²⁺ permeability at physiological conditions (Villarroel & Sakmann, 1996), a long lasting EPSC is likely to produce a massive Ca²⁺ influx into the fibre. The presence of faster channels in the denervated fibre may constitute a mechanism to prevent this massive Ca²⁺ influx. If neonatal channels were to be present in denervated muscle, some endplates may have synaptic currents with very slow decay. Male channels can be as slow as 16 ms, and female channels as slow as 27 ms. This observation is interesting because if by any reason fetal channels were to appear in the endplate of denervated muscle, the deleterious effect would be more severe in female muscle.