Rat Neuromuscular Junction Research Articles

The effects of exercise-induced fatigue on acetylcholinesterase expression and activity at rat neuromuscular junctions.

Acetylcholinesterase is the enzyme that terminates neurotransmission by hydrolyzing the acetylcholine released by the motoneurons at the neuromuscular junctions. Although acetylcholinesterase has been studied for almost a century, the underlying relationship between exercise-induced fatigue and acetylcholinesterase activity at the synaptic cleft is not clear. The purpose of this study was to assess the effects of exercise-induced fatigue on the expression and activity of acetylcholinesterase at the neuromuscular junctions. The expression and activity of acetylcholinesterase at the gastrocnemius neuromuscular junctions was decreased transiently by exercise-induced fatigue and then gradually increased over 24 hr. The expression of acetylcholinesterase in the 24 hr recovery group returned to the level of the control (non-exercised) group, but the activity of acetylcholinesterase remained significantly lower. These data suggest that the decrease of acetylcholinesterase expression and activity may be involved in the production and/or maintenance of exercise-induced fatigue.

Onset dynamics of type A botulinum neurotoxin-induced paralysis

Experimental studies have demonstrated that botulinum neurotoxin serotype A (BoNT/A) causes flaccid paralysis by a multi-step mechanism. Following its binding to specific receptors at peripheral cholinergic nerve endings, BoNT/A is internalized by receptor-mediated endocytosis. Subsequently its zinc-dependent catalytic domain translocates into the neuroplasm where it cleaves a vesicle-docking protein, SNAP-25, to block neurally evoked cholinergic neurotransmission. We tested the hypothesis that mathematical models having a minimal number of reactions and reactants can simulate published data concerning the onset of paralysis of skeletal muscles induced by BoNT/A at the isolated rat neuromuscular junction (NMJ) and in other systems. Experimental data from several laboratories were simulated with two different models that were represented by sets of coupled, first-order differential equations. In this study, the 3-step sequential model developed by Simpson (J Pharmacol Exp Ther 212:16–21,1980) was used to estimate upper limits of the times during which anti-toxins and other impermeable inhibitors of BoNT/A can exert an effect. The experimentally determined binding reaction rate was verified to be consistent with published estimates for the rate constants for BoNT/A binding to and dissociating from its receptors. Because this 3-step model was not designed to reproduce temporal changes in paralysis with different toxin concentrations, a new BoNT/A species and rate (kS) were added at the beginning of the reaction sequence to create a 4-step scheme. This unbound initial species is transformed at a rate determined by kS to a free species that is capable of binding. By systematically adjusting the values of kS, the 4-step model simulated the rapid decline in NMJ function (kS ≥0.01), the less rapid onset of paralysis in mice following i.m. injections (kS = 0.001), and the slow onset of the therapeutic effects of BoNT/A (kS < 0.001) in man. This minimal modeling approach was not only verified by simulating experimental results, it helped to quantitatively define the time available for an inhibitor to have some effect (tinhib) and the relation between this time and the rate of paralysis onset. The 4-step model predicted that as the rate of paralysis becomes slower, the estimated upper limits of (tinhib) for impermeable inhibitors become longer. More generally, this modeling approach may be useful in studying the kinetics of other toxins or viruses that invade host cells by similar mechanisms, e.g., receptor-mediated endocytosis.Electronic supplementary materialThe online version of this article (doi:10.1007/s10928-008-9087-3) contains supplementary material, which is available to authorized users.

Verapamil-fentanyl interaction on rat neuromuscular junction

Verapamil-fentanyl interaction on rat neuromuscular junction

Muscarinic M1 acetylcholine receptors regulate the non‐quantal release of acetylcholine in the rat neuromuscular junction via NO‐dependent mechanism

Nitric oxide (NO), previously demonstrated to participate in the regulation of the resting membrane potential in skeletal muscles via muscarinic receptors, also regulates non-quantal acetylcholine (ACh) secretion from rat motor nerve endings. Non-quantal ACh release was estimated by the amplitude of endplate hyperpolarization (H-effect) following a blockade of skeletal muscle post-synaptic nicotinic receptors by (+)-tubocurarine. The muscarinic agonists oxotremorine and muscarine lowered the H-effect and the M1 antagonist pirenzepine prevented this effect occurring at all. Another muscarinic agonist arecaidine but-2-ynyl ester tosylate (ABET), which is more selective for M2 receptors than for M1 receptors and 1,1-dimethyl-4-diphenylacetoxypiperidinium (DAMP), a specific antagonist of M3 cholinergic receptors had no significant effect on the H-effect. The oxotremorine-induced decrease in the H-effect was calcium and calmodulin-dependent. The decrease was negated when either NO synthase was inhibited by N(G)-nitro-L-arginine methyl ester or soluble guanylyl cyclase was inhibited by 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one. The target of muscle-derived NO is apparently nerve terminal guanylyl cyclase, because exogenous hemoglobin, acting as an NO scavenger, prevented the oxotremorine-induced drop in the H-effect. These results suggest that oxotremorine (and probably also non-quantal ACh) selectively inhibit the non-quantal secretion of ACh from motor nerve terminals acting on post-synaptic M1 receptors coupled to Ca(2+) channels in the sarcolemma to induce sarcoplasmic Ca(2+)-dependent synthesis and the release of NO. It seems that a substantial part of the H-effect can be physiologically regulated by this negative feedback loop, i.e., by NO from muscle fiber; there is apparently also Ca(2+)- and calmodulin-dependent regulation of ACh non-quantal release in the nerve terminal itself, as calmidazolium inhibition of the calmodulin led to a doubling of the resting H-effect.

DOES NITRIC OXIDE MODULATE TRANSMITTER RELEASE AT THE MAMMALIAN NEUROMUSCULAR JUNCTION?

1. Application of the nitric oxide (NO) donor, sodium nitrite and the NO synthase substrate l-arginine had no effect on nerve-evoked transmitter release in the rat isolated phrenic nerve/hemidiaphragm preparation; however, when adenosine A(1) receptors were blocked with the adenosine A(1) receptor antagonist 1,3-dipropyl-8-cyclopentylxanthine (DPCPX) prior to application of sodium nitrate or l-arginine, a significant increase in transmitter release was observed. In addition, the NO donor s-nitroso-N-acetylpenicillamine (SNAP) significantly increased transmitter release in the presence of DPCPX. In the present study, we have made the assumption that these NO donors elevate the level of NO in the tissue. Future studies should test other NO-donating compounds and also monitor the NO concentrations in the tissue to ensure that these effects are, in fact, NO induced. 2. Elevation of cGMP in this preparation with the guanylyl cyclase activator 3-(5'-hydroxymethyl-2'-furyl)-1-benzyl indazole (YC-1) significantly enhanced transmitter release. In the presence of DPCPX and the selective guanylyl cyclase inhibitor 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), which blocks the production of cGMP, the excitatory effects of sodium nitrite and l-arginine were abolished. 3. These results suggest that NO serves to enhance transmitter release at the rat neuromuscular junction (NMJ) via a cGMP pathway and this facilitation of transmitter release can be blocked with adenosine. Previously, we demonstrated that adenosine inhibits N-type calcium channels. Because NO only affects transmitter release when adenosine A(1) receptors are blocked, we suggest that NO enhances transmitter release by enhancing calcium influx via N-type calcium channels. Further studies are needed to confirm that NO alters transmitter release via cGMP and that this action involves the N-type calcium channel. 4. The results of the present study are consistent with a model of NO neuromodulation that has been proposed for the mammalian vagal-atrial junction. This model suggests that NO acts on NO-sensitive guanylyl cyclase to increase the intracellular levels of cGMP. In turn, cGMP inhibits phosphodiesterase-3, increasing levels of cAMP, which then acts on the N-type calcium channels to enhance calcium influx, leading to an increase in transmitter release. Our only modification to this model for the NMJ is that adenosine serves to block the modulation of transmitter release by NO.

Glycerotoxin stimulates neurotransmitter release from N‐type Ca2+ channel expressing neurons

Glycerotoxin (GLTx) is capable of stimulating neurotransmitter release at the frog neuromuscular junction by directly interacting with N-type Ca2+ (Cav2.2) channels. Here we have utilized GLTx as a tool to investigate the functionality of Cav2.2 channels in various mammalian neuronal preparations. We first adapted a fluorescent-based high-throughput assay to monitor glutamate release from rat cortical synaptosomes. GLTx potently stimulates glutamate secretion and Ca2+ influx in synaptosomes with an EC50 of 50 pm. Both these effects were prevented using selective Cav2.2 channel blockers suggesting the functional involvement of Cav2.2 channels in mediating glutamate release in this system. We further show that both Cav2.1 (P/Q-type) and Cav2.2 channels contribute equally to depolarization-induced glutamate release. We then investigated the functionality of Cav2.2 channels at the neonatal rat neuromuscular junction. GLTx enhances both spontaneous and evoked neurotransmitter release causing a significant increase in the frequency of postsynaptic action potentials. These effects were blocked by specific Cav2.2 channel blockers demonstrating that either GLTx or its derivatives could be used to selectively enhance the neurotransmitter release from Cav2.2-expressing mammalian neurons.

Junctional and extrajunctional acetylcholinesterase in skeletal muscle fibers

The asymmetric A 12 acetylcholinesterase (AChE) molecular form, consisting of three tetrameric catalytic oligomers and three non-catalytic subunits of collagen Q (ColQ), is the functional AChE form in the neuromuscular junction. Its extremely high concentration and sharp localization in the junction is mostly due to the binding of this AChE form to perlecan in the synaptic basal lamina. In the rat neuromuscular junctions, about two-thirds of AChE molecules appear to be bound by ionic interactions involving calcium and the rest is probably bound covalently. In immature rat muscles, the A 12 AChE forms are expressed also extrajunctionally. During the early post-natal period, this expression is completely suppressed in rat fast muscles, whereas it extends into adulthood in the slow soleus muscles because of the differences in the extrajunctional expression of ColQ between fast and slow muscles. The level of the A 12 molecular forms of AChE in the extrajunctional muscle regions is regulated by the motor nerve, probably via the pattern of muscle fibre activations triggered by the nerve. The pattern of muscle activations also regulates the extrajunctional expression of the catalytic subunits of AChE: phasic, infrequent, high frequency activations enhance expression, whereas prolonged tonic low-frequency activations tend to decrease it. Calcineurin signalling pathway seems to be involved.

Localization of neuronal calcium sensor-1 at the adult and developing rat neuromuscular junction

Neuronal calcium sensor (NCS-1; frequenin) is a calcium-binding protein involved in the regulation of neurotransmission in the central and peripheral nervous systems from insects to vertebrates. This study reports the localization of NCS-1 immunoreactivity, by Western blotting and immunohistochemistry, at the adult and developing postnatal rat neuromuscular junction. Our confocal immunofluorescence results on the whole-mount muscle and on semithin cross-sections are indicative of the localization of NCS-1 to motor axon terminals. There is no evidence of immunoreactivity in the postsynaptic side of the neuromuscular junctions or teloglial Schwann cells. These results suggest that NCS-1 is involved in the formation and function of presynaptic nerve terminal part of the neuromuscular junction during synaptogenesis and in adult mammals.

Ultrastructural localization of high-affinity choline transporter in the rat neuromuscular junction: enrichment on synaptic vesicles.

In cholinergic neurons, Na(+)- and Cl(-)-dependent, hemicholinium-3-sensitive, high-affinity choline uptake system is thought to be the rate-limiting step in acetylcholine (ACh) synthesis. The system is highly regulated by neuronal activity; the choline uptake is increased by a condition in which ACh release is favored. Here we analyzed the ultrastructural localization of the high-affinity choline transporter (CHT) in the rat neuromuscular junctions with two separate antibodies. The majority (>90%) of immunogold labeling of CHT was observed on synaptic vesicles rather than the presynaptic plasma membrane. Less than 5% of the gold-silver particles were associated with the plasma membrane, and more than 70% of such particles were localized within or in close vicinity to presynaptic active zones. Our morphological data support the recent hypothesis that trafficking of CHT from synaptic vesicles to the plasma membrane couples neuronal activity and choline uptake.

Muscarinic autoreceptors related with calcium channels in the strong and weak inputs at polyinnervated developing rat neuromuscular junctions

Using intracellular recording, we studied how several muscarinic antagonists affected the evoked endplate potentials in singly and dually innervated endplates of the levator auris longus muscle from 3 to 6-day-old rats. In dually innervated fibers, a second endplate potential (EPP) may appear after the first one when we increase the stimulation intensity. The lowest and highest EPP amplitudes are designated “small-EPP” and “large-EPP,” respectively. In singly innervated endplates and large-EPP, we found an inhibition of acetylcholine release by M1-receptor antagonists pirenzepine and MT-7 (more than 30%) and M2-receptor antagonists methoctramine and AF-DX 116 (more than 40%). The small-EPP was also inhibited by both M2-receptor antagonists methoctramine (approximately 70%) and AF-DX 116 (approximately 40%). However, the small-EPP was enhanced by M1-receptor antagonists pirenzepine (approximately 90%) and MT-7 (approximately 50%). The M4-receptor selective antagonists tropicamide and MT-3 can also increase the small-EPP amplitude (75% and 120%, respectively). We observed a graded change from a multichannel involvement (P/Q- N- and L-type voltage-dependent calcium channels) of all muscarinic responses (M1-, M2- and M4-mediated) in the small-EPP to the single channel (P/Q-type) involvement of the M1 and M2 responses in the singly innervated endplates. This indicates the existence of a progressive calcium channels shutoff in parallel with the specialization of the adult type P/Q channel. In conclusion, muscarinic autoreceptors can directly modulate large-EPP generating ending potentiation, and small-EPP generating ending depression through their association with the calcium channels during development.

Ca2+ channels and synaptic transmission at the adult, neonatal, and P/Q-type deficient neuromuscular junction.

Different types of voltage-activated Ca(2+) channels have been established based on their molecular structure and pharmacological and biophysical properties. One of them, the P/Q-type, is the main channel involved in nerve-evoked neurotransmitter release at neuromuscular junctions and the immunological target in Eaton-Lambert Syndrome. At adult neuromuscular junctions, L- and N-type Ca(2+) channels become involved in transmitter release only under certain experimental or pathological conditions. In contrast, at neonatal rat neuromuscular junctions, nerve-evoked synaptic transmission depends jointly on both N- and P/Q-type channels. Synaptic transmission at neuromuscular junctions of the ataxic P/Q-type Ca(2+) channel knockout mice is also dependent on two different types of channels, N- and R-type. At both neonatal and P/Q knockout junctions, the K(+)-evoked increase in miniature endplate potential frequency was not affected by N-type channel blockers, but strongly reduced by both P/Q- and R-type channel blockers. These differences could be accounted for by a differential location of the channels at the release site, being either P/Q- or R-type Ca(2+) channels located closer to the release site than N-type Ca(2+) channels. Thus, Ca(2+) channels may be recruited to mediate neurotransmitter release where P/Q-type channels seem to be the most suited type of Ca(2+) channel to mediate exocytosis at neuromuscular junctions.

Nifedipine-mediated mobilization of intracellular calcium stores increases spontaneous neurotransmitter release at neonatal rat motor nerve terminals.

The modulation of spontaneous release of acetylcholine by specific Ca2+ channel blockers was studied at neonatal rat neuromuscular junction. During early postnatal periods (0-4 days), blockers of N- and P/Q-type Ca2+ channels did not affect miniature endplate potential (MEPP) frequency. Unexpectedly, treatment with the L-type Ca2+ channel antagonist nifedipine, although not when treated with isradipine, nitrendipine, or calciseptine, resulted in strong increase in MEPP frequency. The potentiation effect of nifedipine was dose-dependent with a 56-fold maximum effect with 15 microM. The effect decreased during the first two postnatal weeks and disappeared by the third. The effect of nifedipine was not dependent on extracellular Ca2+ and was not altered by the presence of other Ca2+ channel blockers. In contrast, it was abolished by depleting intracellular Ca2+ stores with 2 microM thapsigargin and was partially inhibited by 10 microM ryanodine. In conclusion, we report a new ryanodine receptor-mediated effect of nifedipine on neonatal neuromuscular junction that may indicate the developmental expression of a specific receptor channel that interacts with intracellular Ca2+ stores. This effect of nifedipine should also be considered when using this drug as either a therapeutic or a research tool.

Glutamate regulation of non-quantal release of acetylcholine in the rat neuromuscular junction.

Glutamate, previously demonstrated to participate in regulation of the resting membrane potential in skeletal muscles, also regulates non-quantal acetylcholine (ACh) secretion from rat motor nerve endings. Non-quantal ACh secretion was estimated by the amplitude of endplate hyperpolarization (H-effect) following blockade of skeletal muscle post-synaptic nicotinic receptors by (+)-tubocurarine and cholinesterase by armin (diethoxy-p-nitrophenyl phosphate). Glutamate was shown to inhibit non-quantal release but not spontaneous and evoked quantal secretion of ACh. Glutamate-induced decrease of the H-effect was enhanced by glycine. Glycine alone also lowered the H-effect, probably due to potentiation of the effect of endogenous glutamate present in the synaptic cleft. Inhibition of N-methyl-d-aspartate (NMDA) receptors with (+)-5-methyl-10,11-dihydro-5H-dibenzocyclohepten-5,10-imine (MK801), dl-2-amino-5-phosphopentanoic acid (AP5) and 7-chlorokynurenic acid or the elimination of Ca2+ from the bathing solution prevented the glutamate-induced decrease of the H-effect with or without glycine. Inhibition of muscle nitric oxide synthase by NG-nitro-l-arginine methyl ester (l-NAME), soluble guanylyl cyclase by 1H[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) and binding and inactivation of extracellular nitric oxide (NO) by haemoglobin removed the action of glutamate and glycine on the H-effect. The results suggest that glutamate, acting on post-synaptic NMDA receptors to induce sarcoplasmic synthesis and release of NO, selectively inhibits non-quantal secretion of ACh from motor nerve terminals. Non-quantal ACh is known to modulate the resting membrane potential of muscle membrane via control of activity of chloride transport and a decrease in secretion of non-quantal transmitter following muscle denervation triggers the early post-denervation depolarization of muscle fibres.

Guanylyl Cyclase Participates in the Mechanism of Glutamatergic Modulation of Acetylcholine Non-Quantum Release in the Rat Neuromuscular Junction

It was shown that glutamate (10 μ M to 1 μM) suppresses in a dose-dependent manner the non-quantum release of acetylcholine from rat motor nerve endings; the release intensity was estimated by the H effect. The action of glutamate was completely eliminated by the blockade of guanylyl cyclase by 1 μM ODQ. An increase in the intracellular cGMP concentration by 1 μ M dibutyryl-cGMP reduced the H effect in a similar manner as glutamate did.

Presynaptic localization of the small conductance calcium-activated potassium channel SK3 at the neuromuscular junction.

Small conductance, calcium-activated potassium channels (SK channels) are present in most neurons, in denervated muscles and in several non-excitable cell types. In excitable cells SK channels play a fundamental role in the generation of the afterhyperpolarization which follows an action potential, thereby modulating neuronal firing and regulating excitability. To date, three channel subunits (SK1-3) have been cloned from mammalian brain. Since SK3 only has been shown to be expressed in muscles upon denervation, this channel may be involved in hyperexcitability and afterhyperpolarization observed in muscle cells in the absence of the nerve. Using confocal microscopy and SK3 specific antibodies, we demonstrate that SK3 immunoreactivity is present at the rat neuromuscular junction in denervated but also in innervated muscles. In denervated muscle fibers, SK3 is localized in the extrajunctional as well as the junctional plasma membrane, where it appears to be less abundant in the acetylcholine receptor-rich domains, corresponding to the crests of the postsynaptic folds. In innervated muscles, SK3 is not detectable in the muscle fiber but is present at the neuromuscular junction and seems to be localized presynaptically in the motor nerve terminals. Axonal accumulation of SK3 immunoreactivity occurs above and below a ligature of rat sciatic nerve, indicating that the SK3 protein is transported in both directions along the axons of the motor neurons. During rat development SK3 immunoreactivity is not found at the neuromuscular junction until day 35 of postnatal development when SK3 first appears in the motor neuron terminals. These results indicate that SK3 channels are components of the presynaptic compartment in the mature neuromuscular junction, where they may play an important regulatory role in synaptic transmission.

N-Benzylpiperidine derivatives of 1,2,4-thiadiazolidinone as new acetylcholinesterase inhibitors

A new family of 1,2,4-thiadiazolidinone derivatives containing the N-benzylpiperidine fragment has been synthesised. The acetylcholinesterase (AChE) inhibitory activity of all compounds was measured using Ellman's method and some of them turned out to be as potent as tacrine. Furthermore, compound 13 was as active as tacrine in reversing the blockade induced by tubocurarine at rat neuromuscular junction. Additionally, receptor binding studies provided new lead compounds for further development of α 2-adrenergic and sigma-receptor antagonists. Molecular dynamic simulation using X-ray crystal structure of AChE from Torpedo californica was used to explain the possible binding mode of these new compounds.

Facilitation of spontaneous acetylcholine release induced by activation of cAMP in rat neuromuscular junctions

Regulation of neurotransmitter release is thought to involve modulation of the release probability by protein phosphorylation. Activation of the cAMP-protein kinase A (PKA) pathway has been shown to facilitate synaptic transmission in mammalian neuromuscular synapses, although the relevant phosphorylation targets are mostly unknown. We found that the inhibitor of the phosphodiesterase aminophylline (1 mM AMIN), the membrane-permeable analog of cAMP, 8-Br-cAMP (5 mM) and, the direct adenylate cyclase activator, forskolin (20 μM), induced an increase of miniature end-plate potentials (MEPPs) frequency in rat neuromuscular junctions. We investigated the possible involvement of the voltage-dependent calcium channels (VDCC), since these proteins are known to be phosphorylated by PKA. But this possibility was ruled out, since the increase in MEPPs frequency was not attenuated by the VDCC blocker Cd 2+ (100 μM) and it was observed when AMIN was studied on hyperosmotic response, which is independent of [Ca 2+] o and of Ca 2+ influx through the VDCC. The lack of action of AMIN on MEPPs frequency when [Ca 2+] i was diminished by exposing the preparations to zero Ca 2+-EGTA solution (isotonic condition) or when nerve terminals were loaded with a permeant Ca 2+ chelator (BAPTA-AM) (hypertonic condition), indicate that cAMP-mediated presynaptic facilitation is a function of nerve terminal Ca 2+ concentration. We also found that AMIN exerted a comparable increase in MEPPs frequency in control and high K + (10 and 15 mM), suggesting a single mechanism of action for spontaneous and K +-induced secretion.

Hebbian Mechanisms Revealed by Electrical Stimulation at Developing Rat Neuromuscular Junctions

Synapse competition and elimination are widespread developmental processes, first demonstrated at neonatal neuromuscular junctions. Action potential activity was long shown to exert a powerful influence, but mechanisms and contribution relative to other factors are still not well understood. Here we show that replacement of natural motoneuronal discharge with synchronous activity suppresses elimination of polyneuronal innervation of myofibers. This requires the simultaneous chronic conduction block (tetrodotoxin) and distal electrical stimulation of motor axons during ectopic synaptogenesis in denervated adult soleus muscle. If in fact chronic stimulation is applied without central block of motor axons, the time course of synapse elimination is as fast as in control muscles undergoing natural activity. Our findings follow the prediction of Hebb's postulate and imply that asynchronous activity drives developmental synapse elimination in muscle. They further suggest that motoneurons could become transiently synchronized during development and regeneration, helping to establish the initial polyneuronal innervation.

Do ryanodine receptors regulate transmitter release at the neuromuscular junction of the rat?

It has been suggested that calcium that is stored in nerve terminals can be released via activation of ryanodine receptors and this source of calcium could serve to modulate evoked transmitter release. Calcium influx via voltage dependent calcium channels could lead to calcium induced calcium release via ryanodine receptors in neuronal tissue. This additional source of calcium could contribute to the total calcium that is available for transmitter release or it could result in having a negative feedback action on calcium influx and transmitter release. We examined the effect of blocking and activating the ryanodine receptors on quantal transmitter release at the rat neuromuscular junction. Intracellular recording techniques were used to monitor end-plate potentials and miniature end-plate potentials. The data supports the view that intracellular calcium released via ryanodine receptors suppresses calcium influx leading to depressed quantal release.

Heme oxygenase-2 expression at rat neuromuscular junctions

The neuromuscular junction is specialized for rapid transmission of electrical signals. Nitric oxide synthase (NOS) is concentrated at the junction, and NO modulates transmission and could influence signaling pathways. Increasing evidence suggests that carbon monoxide (CO) serves as a neurotransmitter, and heme oxygenase (HO), the enzyme that catalyzes the formation of CO, is often colocalized with NOS. Immunoreactivity for HO-2 was present at rat neuromuscular junctions of leg muscles and persisted in denervated muscle indicating the localization of the enzyme to the postsynaptic surface. In contrast, HO-2 immunoreactivity was absent from the en grappe and orbital en plaque endplates of extraocular muscle (EOM), while only the global en plaque endplates possessed HO-2 immunoreactivity. The difference between EOM and leg endplates may arise from EOM's unique physiology. The presence of HO-2 at neuromuscular junctions suggests CO could serve as a pre- and post-synaptic messenger.