Oscillating Voltage in Interneuron I of the Melibe swimming CPG
John F.Wesseling
for: Stuart Thompson
Hopkins Spring Class 1990
February 8, 1991
abstract
Interneuron 1 of the Melibe swimming CPG was studied with
both intracellular and extracellular electrophysiological
techniques. By varying the concentrations of Calcium and
Magnesium in the extracellular saline, an oscillating voltage
that happenned in the absence of action potentials was observed
in Interneuron 1. These results were taken to indicate that
Interneuron 1 is driven at least in part by a wave of
extracellular excitation.
Introduction
Central Pattern Generators have been studied in a number of
different species (see Getting 1983). These rhythmically active
neural circuits often drive coordinated behaviors that are essential to
the survival of the organisms in which they occur. Some CPGs, such
as the ones that control the swimming behaviors of Tritonia and
Melibe, are thought to be simple enough to be studied
reductionistically; that is, it may be possible to account for the
electrical output of these circuits by appealing to the intrinsic
properties and interconnectivities of specific component cells.
such a reductionistic analysis turns out to be possible, these CPGS
may someday give us insight into the functioning of a set of much
more complex neural pattern generators. Perhaps some of the
fundamental properties of simple CPGs will be found to be elemental
in the workings of the neural networks that underlie many human,
motor and cognitive, behaviors.
In this paper, we will describe a continuation of the work done
by Stuart Thompson (Personal Communication) on the CPG that
drives the swimming behavior of the nudibranch mollusk Melibe
leonina. We consider this network to be uniquely interesting for at
least two reasons. First, unlike many simple CPGs (see Getting 1983),
it generates a seemingly regular output that does not fatigue in the
absence of proprioceptive feed back. In fact, it takes a specific
stimulus to halt this CPG (Thompson, personal communication).
Second, this CPG apparently consists of a very few cells, each one
contributing uniquely to the generator, and each one being essential
to the behavior of the network. (Thompson, Personal Communication)
In contrast, the output of many of the more studied CPGs is due to an
integration of distinct groups of cells, each group pooling the similar
outputs of its various members to produce a united effect on the
function of the network. In these networks, not every cell is crucial
to the overall behavior; often ablation of a single cell does not affect
the network much at all. (Getting 1983). Because of the uniqueness
of the Melibe CPG, it is possible that the mechanisms that are
responsible for its behavior are also distinct from the general
mechanisms that have been shown to play key roles in the
functioning of other CPGs.
Thompson has identified four of the component cells of this
Melibe CPG (figure 1, Personal Communication). The current theory
is that the four are arranged bilaterally in the central ganglion, two
per side, one each in the Pleural and Pedal ganglia (fig 1). These cells
have been labeled the right and left Interneuron 1s (Pleural ganglia),
and right and left Interneuron 2s (Pedal). The Int.1, and Int.2 of
each side are thought to normally fire bursts of about 40 spikes
synergistically with each other, and antagonistically (and thus in
inverse phase) with the interneurons on the opposing ganglia. When
the interneurons of one side fire, it is thought that those on the
opposing side are silent. On either side, during a burst of action
potentials, Int.1 begins before Int.2, and Int.2 is thought to continue
firing after Int.l becomes quiescent. Since there is an electrical
coupling between Int.1l and Int.2 (Thompson, personal
communication) it is easy to suppose that Int.1 triggers bursting in
Int.2. Less apparent, though, is the mechanism that triggers Int.1.
This paper describes an attempt to sort out the sort of mechanism
that controls bursting in Int.1.
Materials and Methods.
The nudibranch mollusk Melibe leonina was collected in the
kelp beds near Delmonte Beach in Monterey Bay, California, at a
depth of between 10 and 20 feet. The experimental animals were
between 9 and 18 cm long; these large animals swim at a known rate
which aids identification of Int.1. The animals were held in tanks
kept fresh and at sea temperature (-13-15C) by a constant trickle of
sea water. Pieces of kelp (Macrosystis) provided a substratum. The
Melibe were fed both mysids (collected in the kelp beds off of
Hopkins Marine Station, Pacific Grove, California) and commercially
grown brine shrimp. Whenever possible, the Melibe were fed the
more nutritious mysids (Freya Sommer, personal communication).
Dissection
All experiments were performed on the excised central ganglia
of Melibe leonina. Taking care to keep the ganglia submerged in sea
water, ganglia were exposed by making a single slit in the membrane
directly above the ganglia. First the nerve roots, then surrounding
connective tissue and the esophagus directly caudal, and rostral to
the ganglia were cut, leaving the ganglia and a bit of the esophagus
detached from the rest of the animal. Care was taken not to damage
the two largest circumesophageal connectives.
Within three minutes of the first cut, the ganglia and a bit of
the esophagus were excised from the animal, and placed in à
chamber filled with ASW (-11C) and chilled by Peltier chip. The
remaining esophagus was then cut free from the circum esophageals,
and the ganglia were pinned to the bottom of the chamber by the
attached connective tissue. The ganglia were soaked in EB saline for
35 minutes. EB was then replaced by HiOz saline and the ganglia
were pinned down more firmly, exposing one of the Int.Is, the region
of one of the Pleural ganglia just rostral to the tentacular lobe. That
area was then desheathed with a pair of fine forceps. Finally, the
HiOz saline in the chamber was replaced by ASW (via perfusion-see
below).
Perfusion
Saline changes were done by flushing an amount of saline that
equaled 10 times the volume of the chamber over the ganglia, taking
care to keep the ganglia beneath water level at all times.
Electrophysiology
A glass pipette with a tip diameter of
Extracellular.
approximately 20-50 um was attached to the cell body of Int.l by
suction. Somatic currents were amplified with a Grass differential
amplifier. Data was recorded on a chart recorder (Gould/Brush 220).
The cell body of Int.1 was impaled with a 2-8
Intracellular.
Mohm glass microelectrode backfilled with 3 M KCl. The
transmembrane voltage was amplified with either a Getting micro
electrode amplifier (Model 5). Voltages were recorded on a chart
recorder (Gould/Brush 220).
Serotonin (5HT) In some preparations, the neural circuit is not
functioning (Jones, personal communication). It is believed that
washing the ganglia with a 5HT saline can prompt the circuit to start
functioning. (Jones, personal communication) In preparations whère
we had trouble locating Int.1, we did this, hoping that it would
initiate the bursting behavior of the cell, and enable us to identify
the cell.
Physiological Salines (all Phs at 7.8)
(all concentrations given in mM)
ASW (artificial sea water)
470 Nacl, 10 KCl, 10 CaCl2, 50 MgC12,
10 HEPES(C8HI7N204SNa)
Hi Oz
470 mM NaCl, 10 KCl, 10 CaCi2, 50 MgC12,
10 Hepes, 500 Sucrose
EB
Collagenase, and Dispase in solution of ASW.
10 mg Dispase, 2 mg Collagenase / 1 ml ASW
5-HT (serotonin
10-5 5HT in solution with ASW
HiDi
470 Nacl, 10 KCI, 5 CaCl2, 100 MgCI2, 10 Hepes
O Ca++
470 Nacl, 10 KCl, 60 MgCI2, 10 Hepes
Salts and drugs were obtained from Stuart Thompson's laboratory,
Hopkins Marine Station, Pacific Grove, CA.
Results:
Int.1 is a small cell located in the saddle-shaped groove just
rostral to the tentacular lobe on the dorsal side of the Pleural
ganglion (fig la). In functioning networks, this cell was identified by
its characteristic bursting firing pattern. Under normal conditions,
the cell spikes up to 40 times per burst and its interburst interval is
at least as long as the bursting phase: the cell bursts with a
maximum frequency of 15 per minute in large animals. (Thompson,
personal communication) Though it was not possible to find this cell
in every preparation, of the 20 preparations in which we did record
from a cell that satisfied the specifications, we never observed à
second cell in the area that also had the characteristic bursting
pattern. This confirms previous observations by Stuart Thompson
and Brad Jones (personal communications). Although we completed
the majority of our experiments within 5 hours of the original
dissection, in one experiment we were able to hold the regularly
firing Int.1 at 7.5 hours after the dissection, and even then, we
noticed no physiological abnormalities.
Low Calcium Manipulation
Ganglia were bathed in nominal 0 Ca saline in order to disrupt
the network by altering the time course of the various cellular
mechanisms which depend upon free Ca. In four out of five
experiments in which healthy ganglia were bathed in 0 Ca saline, the
regular firing pattern of the swimming CPG was disrupted, but did
not stop firing altogether. (figure 3b) Reintroduction of normal saline
had the immediate effect of halting all firing. The disruption was
reversed in all four instances by leaving the ganglia to soak in the
normal saline for longer than twenty minutes.
In an extracellular recording from a regularly bursting RInt.I,
the 0 CA manipulation disrupted regular bursting. Bursts became
longer, and happened much less frequently. This erratic pattern
continued for 50 minutes, at which time normal ASW was
reintroduced. Within 2 minutes, all spiking activity had halted. By
17 minutes after perfusion, regular bursting had resumed.
We sought to study this phenomenon by recording from Int.I
intracellularly. After at least fifteen minutes of regular bursting, we
introduced zero nominal free calcium saline (0 Ca) to the ganglia. We
recorded the transmembrane voltage fluctuation of Int.l while
soaking in 0 Ca for 20 minutes. Again, the presence of 0 Ca disrupted
the normal bursting action of Int.1, but bursting did not halt until
ASW was reintroduced to the preparation (fig 2). At this point, an
oscillating voltage was recorded in the absence of action potentials
(fig 2d).
The oscillations appear to be following the same rhythmic
pattern as the bursts of action potentials that had just been silenced.
The furthest hyperpolarized Int.1 becomes during normal bursting is
approximately -55my. Under normal circumstances, the cell begins
firing when the voltage reaches -40mv. That is, an increase of 15 mV
is almost always enough to cause the cell to fire. The oscillation
range seen here is about 12 my, 4/5 of the normal sub-threshold
zone. The oscillations are not smooth fluctuations in voltage, but
rather are marked by sudden quantal jumps in voltage which occuf
in concurrence with the depolarizing phase of the oscillations.
Hi Divalent Saline Manipulation
To monitor the effects on int.1 of decreasing the efficacy of
multisynaptic connections in the network, we bathed the ganglia in
HiDi saline. Such a saline is believed to decrease the strength of each
monosynaptic connection, and thus by a multiplicative effect,
substantially lower the probability of making consistent
multisynaptic connections, without significantly influencing the
spike initiating threshold voltage of individual cells (Getting and
Willows 1974).
We bathed the ganglia in HiDi while recording intracellularly
from the regularly bursting Int.1 once. Within 7 minutes of
perfusion, RInt.1 went from regularly bursting, to quiescent. In
contrast to the 0 CA experiments, reintroduction of ASW was not
necessary in order to halt all spiking activity. The individual bursts
progressively shortened until all spiking halted altogether. As each
burst began comprising only a very few spikes, an underlying wave
of depolarization upon which the spikes had been riding became
apparent. In the complete absence of spiking, the voltage oscillations
continued (fig 3). Although the depolarized phase of these
oscillations had quantal jumps in voltage associated with them, the
jags" on the data trace were less obvious than the ones present after
low calcium perfusions. This effect was reversed by reintroduction
of ASW.
Discussion
The principal finding of these studies is that the patterned
rhythmic electrical activity in the Melibe central pattern generator
that is responsible for the character of the Melibe swimming
behavior can occur in the absence of spiking activity in at least one
Interneuron one. We can see from fig 2, and fig 3 that rhythmic
voltage oscillations continue in the soma of Int.l in the absence of
voltage spikes. Since neither the temporary introduction of HiDi, noi
of the 0 Ca salines permanently destroys the normal propagating
action of spikes along excitable membrane, we can conclude that
these voltage oscillations are not merely due to passive current
spread from another, active part of the cell (such as the axon hillock)
and that the oscillations really do happen in the absence of spikes in
this cell.
The mechanisms that are responsible for these voltage
oscillations have a large influence on the bursting behavior of the cell
under normal conditions. First of all, we can tell that the
mechanisms are operating in the normally bursting cell by observing
the cell while the whole ganglia is being bathed in a Hi Di saline.
During that experiment, the voltage oscillation is incrementally
revealed as the number of spikes per burst decreases; the current
flow that causes the oscillations is present in the bursting cell.
Secondly, the size of the oscillations indicates that they play a
significant role in any spiking activity of the cell since the amplitude
of the voltage swings is at least 80 per cent of the sub (spike
initiating) threshold zone. By understanding the mechanisms
responsible for these observed voltage fluctuations, we will
understand something about the mechanism that controls the
patterned firing activity of Int.1.
There are three possible mechansims, any or all of which could
contribute to the observed voltage oscillations in Int.1.
1» Int.1 could be an endogenous burster; some property of the
cell itself could be sufficient to cause the voltage to vary periodically.
Such cells are common in molluscan ganglia (Smith and Thompson
1987), and though their pattern is usually dependent upon the
occurrence of action potentials, that an endogenous voltage oscillator
could function without the aid of action potentials is conceivable
(Strumwasser 1968).
2» The cell could be driven by an endogenous Post Inhibitory
Rebound mechanism (Jones 1986); a mechanism by which cells fire
when they are released from inhibition. It could be that the
hyperpolarized phases of the voltage oscillations that we saw are
caused by inhibitory input from the antagonistic Int.l and Int.2.
this case, the depolarized phases could be accounted for easily by à
simple absence of inhibition.
3» The cell could be driven by excitatory input from another
cell in the network. In this case, the depolarized phase would be
explained by excitatory (positive inward current) input, and the
hyperpolarized phase would be explained as the absence of
excitation.
The voltage oscillation is probably not due to either the first or
second mechanisms. The voltage increase is marked with a series of
fast jumps in the voltage, yielding an irregular recording. These
sudden jags in the data trace can easily be explained as direct
consequences of the opening of receptor channels. Since these
quantal jumps in voltage are in a depolarizing direction, and since
several of them occur near the normal spike initiating threshold of
this cell, they are most easily explained as excitatory input. In
contrast, the hyperpolarized phase of the oscillation is relatively
smooth, indicating that the ipsps, if existent, are quite a bit smaller,
and thus have less of an effect, than the excitatory input.
The question that remains is, what sort of mechanism excités
Int.1 in this periodic fashion? Thompson (personal communication)
tested Int.1 for excitatory input from the antagonistic Int.1, and
Int.2, and found none. However, he did observe an electrical
connection between Int.2 and Int.1. It could be that the depolarized
phase of the voltage that we see in Int.l is due to passive current
spread from a spiking Int.2. If that is the case, then, contrary to the
current model of this network, Int.2 can fire rhythmic bursts of
action potentials without the input of Int.1. However, in this case, it
would be doubtful that Int.2 habitually drives Int.l because
Thompson (personal communication) has made several simultaneous
recordings from Int.1 and Int.2, and has observed that Int.l
generally begins bursting before Int.2. It could instead, be that Int.l
is being driven by other, as of yet unidentified, cells in this network.
Future Directions
Future experiments should determine the cell (or cells)
responsible for this patterned excitation of Int.1. By recording
simultaneously from Int.2 and Int.1, while doing the Hi Di, and OCa
manipulations, one could test the hypothesis that the excitatory input
to Int.1 comes directly from Int.2. Simultaneous recordings from
every pair of the identified cells in this network during regular firing
could yield important information about the order, and
synchronization of the four cells. This information could be used to
support or weaken the hypothesis that neither of the two
antagonistic neurons can be directly responsible for the excitatory
input witnessed in Int.1. In fact, this information would be essential
to any computational model that might be developed to test the
general validity of hypotheses about the various mechanisms that
make up this network.
Though Int.1 is clearly getting patterned excitatory input, it is
not necessarily true that that input is responsible for all, or even
most of the current flow that we observe as oscillating voltage in
Int.1. It could be that the input is riding upon a wave of voltage
fluctuation that is due to some endogenous property of the cell itself.
To test such a hypothesis, one could watch for voltage oscillations in
Int.1 while attempting to block all of the spiking activity in the
whole ganglia (TTX), and hence block all extracellular input into
Int.1.
Acknowledgements
First and foremost, I thank Stuart Thompson. Not only has
he aided me greatly with this project, but, put bluntly, were it
not for him, I would be planning to do something other with my
life than science. No amount of gratitude will ever be close to
enough.
I thank Sam wang
from whom I have learned the bulk of my
science knowledge.
am grateful to William Gilly for lots of
technical assistance,
Freya Sommer who helped with the animals,
Alan Baldridge and Susan Baxter, Anthony Morielli, David Epel and
Chuck Baxter both of whom supplied encouragement, Mary Lucero,
and Jennifer Levitt.
Thanks especially to my peers in the spring class, Dianne,
Shannon, and Simone for emotional support of both me and the
slugs. Finally, I thank the whole Hopkins community that was
exceptionally pleasant and generally helpful.
Literature Cited
"Modification of
(1974)
Getting, P.A. and A.O. Dennis Willows.
Neuron Properties by Electrotonic Synapses. II. Burst
J. Neurophysiol. vol
Formation by Electrotonic Synapses.
XXXVII, No. 5. 858-868.
"Neural Control of Swimming in Tritonia."
Getting, P.A. (1983)
in Neural Origin of Rhythmmic Movement. pp 129-158. SEB
Symposium XXXVII: Cambridge U. Press: Cambridge.
Jones, B.R. (1986) "Slow Ionic Currents Underlying Postinhibitory
Rebound in Aplysia Neurons."
Doctoral Dissertation Hopkins
Marine Station, Pacific Grove, CA.
"Slow Membrane Currents
(1987
Smith, S.J. and Stuart Thompson.
in Bursting Pace-Maker Neurones of Tritonia." J. Physiol.
382: 425-448.
Strumwasser F. (1986) "Membrane and Intracellular Mechanisms
Governing Endogenous Activity in Neurons." in Physiological
and Biochemical Aspects of Nervous Integration. pp. 329-
341. Prentice Hall, Englewood Cliffs, N.J.
figure legends.
figure 1. (a) diagram of M. leonina central ganglia from dorsal view
with Interneurons! and 2. (b) schematic of current CPG model.
figure 2. Low Ca experiment--Interneuron 1 (a) normal firing of
Interneuron 1 in active CPG. (b) recording of the transmembrane
voltage of Interneuron 1 soma as the ganglia is bathed in 0 Ca saline.
(c) voltage trace after the reintroduction of normal saline. (d)
voltage trace after reintroduction of normal saline-expanded gain.
figure 3. Hi Di experiment —Interneuron 1 (a) voltage recording of
Interneuron 1 approximately 4 minutes after introduction of Hi Di
saline. (b) recording 7 minutes after introduction of Hi Di saline
figure 1
(a)
pedal
gan.
(b)
Syn.
at
caudal
pleural
gang.
tendkula

(int.)

Lnt 2)
int.
Rint.1
Rint.2

m.
a.
2
figure 2

(a)
L

(b)
l


(c)

OS
dnakenn
(d)
105
d
P
figure 3
40n
(a)
A




Le
40my
(b.
1OS
nnn