Neuromuscular Junction: Principles of Synapse Formation
October 23 , 2008
Structure of the Neuromuscular Junction-
A Classical Synapse
NMJs are accessible for electrophysiological, biochemical, structural and developmental
studies. Much of what we know about the synapses derives from work initiated at NMJs.
For example, Katz and coworkers demonstrated chemical transmission and the quantal
hypothesis at NMJs.
Other tools such as the electric organs of electric fish allowed cloning of the subunits of
acetylcholine receptors (AChRs) and acetylcholinesterase (AChE). Toxins (alpha
bungarotoxin) that bind irreversibly to AChRs have been used to follow synaptic
dynamics. As a result work on NMJ had progressed relatively rapidly. Moreover work in
this area is a paradigm for the studies of synapse formation in the brain since many of the
same questions obtain. For example:
Does the presynaptic cell target a specialized region of the postsynaptic cell?
Does the postsynaptic cell stimulate differentiation of the presynaptic cell?
How is postsynaptic specialization including a high density of neurotransmitter
How stable are synapses?
How is that stability maintained?
How does neural activity affect synapse formation?
Structure of NMJs:
-The matching of pre- and postsynaptic elements of the nerve terminal is very precise.
AChRs are at a density of 10,000X higher at under the nerve terminal than in adjacent
areas away from the terminal.
AChE is concentrated at NMJs in the basement membrane that runs over the entire
muscle fiber but is specialized at NMJs.
A neuromuscular junction, Gold chloride, http://starklab.slu.edu/Physio/Muscle.htm
AChE is located in the basement membrane at the postsynaptic terminal of the
Examples of NBT-AChE stained neuromuscular junctions from post-metamorphic frogs, Letinsky 1980,
Journal of neurocytology
Principles of NMJ formation:
1. Much of axonal guidance and synapse formation is regulated posttranslationally:
The Dynamics of Re-Innervation of Neuromuscular Junction-
In the figure above, consider the muscle to be considerably distant from the motor
neuron cell body that innervates it. In fact, the distance between a cell body and the
muscle innervated by a neuron in a mammal could be in the order of meters. If we were
to section the axon of this nerve cell, we would find that the re-growing nerve would find
its way to the old end plates. Growth cones grow at approximately 50μm per hour, while
a muscle fiber has a diameter varying from 50 to 100 μm per hour. Moreover, the rate of
retrograde and orthograde transport is on the order of millimeters per hour.
It thus seems unlikely that a signal from the cell body (i.e. the genome) could reach the
growth cone fast enough to signal termination of growth and differentiation into a nerve
terminal. This regulation is largely posttranslational.
(Caveat: New evidence for protein synthesis and RNA in growth cones)
2. Cell surface proteins are localized at a very high concentration beneath the nerve
Membrane proteins are typically not icebergs diffusing in a sea of lipids but are often
associated with intracellular cytoskeletal proteins. Aggregation can enhance association
of membrane proteins with cytoskeletal proteins that anchor the assembly in the
membrane. Hence the postsynaptic apparatus will not drift away from the nerve terminal.
3. AChRs are diffusionally and metabolically very stable.
The half life (t1/2) for degradation of AChR prior to innervation is approximately 20
hours. Similar t1/2 for extrajunctional AChR following innervation. That of junction
AChR on the other hand is greater than 10 days.
-Synaptic activity results in a large increase in recycling of AChRs and AChE from the
cell surface and may be important to synaptic plasticity where transmission would be
regulated depending on the amount of use.
4. Arrays of membrane proteins such as cell adhesion molecules may have a greatly
increased avidity for interaction despite the fact that each protein has a low affinity
for its ligand. This is important in cell cell interactions and assembly of the
Increase in avidity NCAM-NCAM interaction
(Kd (Dissociation constant) of 10-6M i.e. low affinity)
We observe a 30-40 fold increase in adhesion between NCAM upon a 2-3 fold
increase of NCAM concentration.
6. Synaptic imprinting: NMJs ave a template in the ECM that permits precise
Cells in the body often die over the course of a lifetime and residual neurons may
take there place. When a nerve cell dies, old synaptic sites vacated by dead innervating
neurons may become reinnervated by neighboring neurons.
Muscles and neurons must contain a template or imprint that is stable in the face
of this turnover and is “instructive” to the newly sprouting growth cone telling it where to
innervate and how to form a nerve terminal.
A- AChR at Neuromuscular Junctions are Metabolically very
B- Mechanisms of Transynaptic modulation of Genome
A- Localization of Membrane Proteins.
From The Molecular Biology of the Cell, Alberts et Al.
7. Nerve terminals compete to innervate their targets:
Synaptogenesis is a slow (2-3 weeks process) that entails sequential appearance of
specializations. At E13 in rat, motor axons grow into developing muscle that still contain
myoblasts and have not separated into individual muscles. As they grow, the nerve’s
growth cones release Ach. At E15, AChRs are found clustered under the nerve ending.
Presynaptically, synaptic vesicles accumulate in the terminal and levels of choline
acetylase (Ach synthesizing enzyme) increase.
At E16, acetyl cholinesterase (AChase) becomes concentrated in the synaptic cleft.
At E18-E20, the metabolic half life of AChR increases form about 24 hours to 10 days.
Subsequently, synaptic folds appear and extrajunctional AChR disappear.
Initially each NMJ is innervated by more than one nerve terminal. These compete
with one another and ultimately one is left. Similar competition occurs in the brain and is
thought to be essential for precise innervation. The rules and molecules governing these
events are at this point unclear but activity is important. For example, blocking neural
activity will allow multiple nerve terminals to persist for a much longer time.
A schematic representation of neuromuscular junction development in amphibians.
From Letinsky 1980, Journal of neurocytology.
Cell culture has contributed importantly to our understanding of how NMJs form
but reductionists beware!
Evidence for neural control of synapse formation-
Focus on Ach as first event in synapse formation and mediator of muscle activity
that is important in regulating subsequent events. Prior to innervation, AChR are
distributed diffusely over the myotube surface. Following innervation and maturation,
they are all found in the subsynaptic membrane. Increases in AChR results from the
synthesis of new receptors at extrajunctional regions and is thus not due to the spread of
AChR away from the synapse. Reinnervation of muscle causes a decrease in AChR
synthesis at extrajunctional region. Following denervation of a muscle, we can observe
that direct muscle stimulation blocks increases in AChR synthesis. Thus, muscle activity
per se and not synaptic activity regulates AChR synthesis. Following denervation, muscle
stimulation inhibits the increase in extrajunctional AChRs.
As noted above, muscle activity also regulates multiple innervation of muscle
fibers. Blockage of muscle activity in embryonic muscle results in maintenance of
multiply innervated fibers. Direct muscle stimulation inhibits multiple innervation and
hence is not directly synaptically dependant (i.e. is not necessary to activate AChR.).
Is there a specialized region on the cell surface that is targeted by the ingrowing
AChR aggregation in Culture (Anderson and Cohen)
AChR labeled with
alpha Bungartoxin can be flourescenated . It binds irreversibly to AChRs and can be used
to track them.
Muscle cells in culture without neurons form aggregates which similar to those at
developing NMJs in vivo.
However the nerve does NOT preferentially innervate these densities rather is grows over
the muscle cell and the density of AChRs disperses and collects around the nerve…..So
much for Accam’s razor.
Activity dependent release of ACh is not important in this process since the AChRs are
not functional once bound with bungarotoxin.
Receptors become localized by Diffusion-Trapping:
Region of the muscle cell surface is altered by the incoming nerve so that AChRs diffuse
rapidly into this region but diffuse out at a much lower rate.
What constitutes the trap and how is it established?
Rapsyn, agrin, MuSk (muscle specific kinase), dystroglycan
Rapsyn is an intracellular peripheral membrane protein that is associated with
AChRs (it co-purifies with AChRs). It is localized to synaptic regions in muscle and to
AChR aggregates which form in culture. The removal of rapsyn with “Chaotropic”
agents disperses AChR aggregates in culture.
Expression of rapsyn alone in frog oocytes leads to self aggregation.
Co-expression of AChR with rapsyn will cause the aggregation of AChR and co-
localization with rapsyn.
Mice null in the gene for rapsyn have synapses where the AChR are more broadly
and less densely associated that at normal neuromuscular junctions. Presumably, rapsyn
is sufficient for the clustering of AChRs.
Rapsyn binds to Actin and associates with another complex of membrane
proteins, including alpha and beta dystroglycan, that are linked to the extracellular matrix
as well as to the membrane cytoskeletal protein utrophin.
Mice null for rapsyn have NMJs with a diffuse distribution of AChRs.
Agrin can be isolated from Torpedo muscle cell. Experimental findings demonstrate that
crude basement membrane (Basal Lamina) extracts increases AChR aggregation in
Use of cell culture to assay for aggregating activity led to biochemical identification and
molecular cloning of agrin. It was viewed as a molecule that stimulated aggregation of
Mice null for agrin have severely disrupted NMJs
Muscle Specific Receptor kinase
Localized to neuromuscular junction
Binds agrin (De Chiara et al, Cell 1966)
Phosphorylated in response to Agrin (Activated by Agrin)
Mice null in the MuSK gene have no neuromuscular junctions
Muscles had no AChR clustering and nerve terminals grew all over the cell.
The cartoon below describes a plausible model for the function of these 3 molecules in
Simple Model of Neuromuscular Junction Formation
Growth cone grows into
muscle mass containing
myofibers with a diffuse
distribution of AChRs.
Agrin in incorporated into the B.M.
and forms part of the synaptic imprint.
Growth cones release agrin ARIA is also incorporated into the
which activates MuSK basement membrane.
MuSK in turn activates unknown intracellular intermediates
which stimulate rapsyn aggregation. Alpha ( ) and Beta (ß)
dystroglycan also aggregate along with rapsyn. The AChR
may then diffuse into this region and become trapped.
Growth cones also release Ach which cause synaptic activity
over the entire muscle fiber. This activity inhibits transcription
of AChRs at all nuclei of the myofiber.
A second fiber, Glial growth fiber (or acetylcholine receptor inducing
activity) activates a distinct receptor tyrosine kinase (Erbs) which are
dimer receptor kinases that increase the synthesis of AChRs.
Erbs become concentrated at neuromuscular junction through MuSK
setting up a compartment in which just the subsynaptic nuclei continue to
Here is the BEWARE PART for all you reductionists.
More recent work in vivo has found that muscle contains a compartment of AChRs that is
that attracts the incoming nerve. That is the muscle is instructive for the nerve and not the
reverse as in culture where the nerve does NOT recognize the AChRs.
This compartment of AChRs is MuSk dependent….MuSk null mice do have these
The nerve will disperse these clusters of AChRs (see paper for class by Misgeld et al) but
it releases agrin that stabilizes (not induces) them and results in NMJ formation.