Download Excitatory and Inhibitory Vestibular Pathways to the Extraocular

Excitatory and Inhibitory Vestibular Pathways to the Extraocular Motor
Nuclei in Goldfish
WERNER GRAF, 1 ROBERT SPENCER, 2 HARRIET BAKER, 3 AND ROBERT BAKER 4
1
Laboratoire de Physiologie de la Perception et de l’Action, Centre National de la Recherche Scientifique, 75270 Paris
Cedex 06, France; 2 Department of Anatomy, Medical College of Virginia, Richmond, Virginia 23298; 3 Department of
Neuroscience, Cornell University Medical College at the Burke Medical Research Institute, White Plains, 10605; and
4
Department of Physiology and Neuroscience, New York University Medical Center, New York 10016
structural traits that have been essentially preserved throughout
vertebrate phylogeny by a shared developmental plan.
INTRODUCTION
Early electrophysiological work in the vestibulooculomotor system of mammals persuasively argued that the most
strikingly feature of synaptic organization might be the presence of reciprocal excitatory and inhibitory pathways (Baker
et al. 1973; Highstein 1973; Ito et al. 1973a,b; Magherini et
al. 1974; Precht and Baker 1972). Primary vestibular afferents in each of the three ipsilateral semicircular canal related
circuits were hypothesized to contact inhibitory and excitatory second-order neurons in the ipsilateral vestibular nuclei
that would inhibit ipsilateral, and excite contralateral, motoneurons. For instance, ipsilateral posterior canal afferents
projecting to the superior and medial vestibular complex
would inhibit ipsilateral superior rectus and inferior oblique
motoneurons and excite contralateral trochlear and inferior
rectus motoneurons (Graf et al. 1983). Subsequent work
showed that vestibular signals originating from each of the
six canals formed nearly exclusive contralateral/excitatory
and ipsilateral/inhibitory three neuron arcs terminating in
the appropriate subgroups of extraocular motoneurons (reviewed in Evinger 1988; Highstein and McCrea 1988). To
date, the only exception to this scheme has appeared to be
the presence of ipsilateral excitation and lack of inhibition
to the medial rectus subdivision from the horizontal canal
(Baker and Highstein 1978).
Structure/function work using intracellular horseradish
peroxidase (HRP) as a marker corroborated the electrophysiology and further suggested that second-order vestibular
neuron projections may differ little between closely related
vertebrates, because only minor differences in species-specific arborization patterns were found for individual vertical
and horizontal vestibular neurons, for example cats and rabbits (Graf and Ezure 1986) compared with squirrel monkey
(McCrea et al. 1987). Based on single-cell techniques, reciprocal vestibular innervation also was postulated in the distantly related species of flatfish (Graf and Baker 1983) and
recently, this plan has been suggested in other lower vertebrates by the use of pathway tracers (e.g., lamprey, Pombal
et al. 1994; elasmobranchs, Puzdrowski and Leonard 1994).
In a particularly relevant study, a direct parallel was shown
between the amphibian (Straka and Dieringer 1993) and
0022-3077/97 $5.00 Copyright q 1997 The American Physiological Society
/ 9k11$$my40 J711-6
08-08-97 12:57:37
neupa
LP-Neurophys
2765
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
Graf, Werner, Robert Spencer, Harriet Baker, and Robert
Baker. Excitatory and inhibitory vestibular pathways to the extraocular motor nuclei in goldfish. J. Neurophysiol. 77: 2765–2779,
1997. Electrophysiological, ultrastructural, and immunohistochemical techniques were utilized to describe the excitatory and inhibitory vestibular innervation of extraocular motor nuclei in the goldfish. In antidromically activated oculomotor motoneurons, electrical stimulation of the intact contralateral vestibular nerve produced
short-latency, variable amplitude electrotonic excitatory postsynaptic potentials (EPSPs) at 0.5–0.7 ms followed by chemical EPSPs
at 1.0–1.3 ms. Stimulation of the ipsilateral vestibular nerve produced small amplitude membrane hyperpolarizations at a latency
of 1.3–1.7 ms in which equilibrium potentials were slightly more
negative than resting potentials. The inhibitory postsynaptic potentials (IPSPs) reversed with large amplitudes after the injection of
chloride ions suggesting a proximal soma-dendritic location of
terminals exhibiting high efficacy inhibitory synaptic conductances. In antidromically identified abducens motoneurons and putative internuclear neurons, electrical stimulation of the contralateral vestibular nerve produced large-amplitude, short-latency electrotonic EPSPs at 0.5 ms followed by chemical depolarizations at
1.2–1.3 ms. Stimulation of the ipsilateral vestibular nerve evoked
IPSPs at 1.4 ms that were reversed after injection of current and/
or chloride ions. g-Aminobutyric acid (GABA) antibodies labeled
inhibitory neurons in vestibular subdivisions with axons projecting
into the ipsilateral medial longitudinal fasciculus (MLF). Putative
GABAergic terminals surrounded oculomotor, but not abducens,
motoneurons retrogradely labeled with horseradish peroxidase.
Hence the spatial distribution of GABAergic neurons and terminals
appears highly similar in the vestibuloocular system of goldfish
and mammals. Electron microscopy of motoneurons in the oculomotor and abducens nucleus showed axosomatic and axodendritic
synaptic endings containing spheroidal synaptic vesicles establishing chemical, presumed excitatory, synaptic contacts with asymmetric pre- and/or postsynaptic membrane specializations. The
majority of contacts with spheroidal vesicles displayed gap junctions in which the chemical and electrotonic synapses were either
en face to dissimilar or adjacent to one another on the same soma/
dendritic profiles. Another separate set of axosomatic synaptic endings, presumed to be inhibitory, contained pleiomorphic synaptic
vesicles with symmetric pre- and/or postsynaptic membrane specializations that never included gap junctions. Excitatory and inhibitory synaptic contacts appeared equal in number but were more
sparsely distributed along the soma-dendritic profiles of oculomotor as compared with abducens motoneurons. Collectively these
data provide evidence for both disynaptic vestibular inhibition and
excitation in all subdivisions of the extraocular motor nuclei suggesting the basic vestibulooculomotor blueprint to be conserved
among vertebrates. We propose that unique vestibular neurons,
transmitters, pathways, and synaptic arborizations are homologous
2766
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
/ 9k11$$my40 J711-6
individual oculomotor subgroups, but always more robust
in abducens motoneurons and internuclear neurons. Overall,
our data strongly suggest that the pattern of vestibuloocular
innervation from the three semicircular canal is surprisingly
well conserved between distantly related taxa. Preliminary
results have been published previously in abstract form
(Baker et al. 1986, 1987b).
METHODS
Surgical procedures
All preparative surgery was done under MS222 (1:2,000
wt/vol) anesthesia and immobilization with gallamine triethiodide
(Flaxedil; 2 mg/kg). Twenty-four goldfish were mounted in a
customized fish holder system, and aerated water was continuously
circulated to the gills by an electric pump through a tube fitted to
the mouth. Incisions were infiltrated with local anesthetics (2%
lidocaine with epinephrine). Access to the labyrinths was obtained
by bilateral removal of the bone overlying the hindbrain and cerebellum. Individual branches of the anterior, horizontal, and posterior ampullary nerves along with the utricular, saccular, and lagenar
nerves were exposed from the side of the brain stem. The bone
overlaying the midbrain also was removed so the oculomotor and
trochlear nuclei could be reached either through the intertectal cleft
or by direct visualization after removal of the tectal commissure
and valvula. For gross antidromic identification of all extraocular
motoneurons, bipolar silver ball stimulation electrodes were implanted in both orbits. Orthodromic vestibular activation was attained by bipolar silver wire electrodes placed at the most peripheral location surrounding the three ampullary nerves. In most experiments, one or more of the individual extraocular muscles and
labyrinthine nerve branches was dissected free from overlying tissue, and a set of smaller bipolar electrodes was utilized to obtain
more selective stimulation.
Electrophysiological recordings and data analysis
Glass microelectrodes (0.5–1-mm tip size), containing either 3
M K Acetate, 2 M K Citrate, 2 M LiCl, or 3 M KCl were used
for intracellular recordings from extraocular motoneurons. Oculomotor, trochlear, and abducens motoneurons were identified by
antidromic action potentials following electrical stimulation in the
orbit. Synaptic potentials, either excitatory or inhibitory postsynaptic potentials (EPSPs or IPSPs, respectively), were recorded after
vestibular nerve stimulation. Latency for antidromic invasion
ranged from 0.3 to 0.5 ms and for orthodromic disynaptic activation
varied between 1.0 and 1.7 ms. During recording, the animals were
kept under light pentobarbital sodium (15–20 mg/kg) anesthesia.
The electrophysiological data were recorded either on magnetic
tape (Neurocorder: Neurodata), or magnetic disks (Nicolet 4096,
Data 6000) and later read out on a laser printer for analysis and
documentation. Some electrophysiological records were photographed directly from a Tektronix oscilloscope (Figs. 1–3) and
others assembled by computer-assisted graphics (Fig. 4).
g-Aminobutyric acid (GABA) immunohistochemistry
The extraocular muscles of four goldfish were injected with 2
ml of 10% HRP. After a survival time of 1 day, animals were
reanesthetized and perfused with fixative solution containing 4.0%
paraformaldehyde in 0.1 M phosphate buffer with 0.002% calcium
chloride (pH 7.4). Vibratome sections (25–50 mm) were prepared
through the abducens, trochlear, and oculomotor nuclei. The sections were first reacted for HRP (Pastor et al. 1991; Spencer and
Baker 1990) to identify extraocular motoneurons, then washed
through several changes of 0.1 M sodium phosphate buffer for up
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
mammalian (Spencer and Baker 1990) horizontal canal system in respect to the putative ipsilateral, inhibitory glycinergic and contralateral, excitatory glutamatergic input to
both abducens internuclear and motoneurons. Nevertheless,
a broad analysis including the anatomy, physiology, and
transmitter histochemistry of canal-specific second-order
projections to each subdivision of extraocular motoneurons
has not been carried out for any taxa evolutionarily distinct
from mammals.
The largest body of nonmammalian oculomotor related
literature has been accumulated in teleostean fish, primarily
cypriniformes (includes goldfish, catfish, and zebrafish),
possibly because this order exhibits robust oculomotor performance with eye movements comparable with those observed in mammals (Easter 1972; Pastor et al. 1992; Schairer
and Bennett 1986). In the goldfish, neurons within the vestibular complex, notably the anterior, descending, and tangential octaval nuclei (McCormick and Braford 1994), have
been found to be distributed consistent with putative separate
inhibitory and excitatory projections to oculomotoneurons
(Allum et al. 1981; Torres et al. 1992). Although initial
intracellular records in goldfish oculomotoneurons did not
report inhibition (Korn and Bennett 1971), structural and
neurophysiological observations in the goldfish horizontal
canal system indicated reciprocal inhibition and excitation
(Baker et al. 1986, 1987a, 1994). Eye velocity and position
signals have been recorded from hindbrain neurons presumed homologous to those in the mammalian vestibular
and prepositus nuclei, suggesting similar central processing
of vestibular and visual signals (Pastor et al. 1994). Collectively, these data provide sufficient rationale to suggest that
it would be useful to directly compare the electrophysiology
and anatomy between teleosts and mammals.
The presence of electrotonic coupling reported in teleost
motor pathways, including extraocular motoneurons, provides
a noteworthy difference between mammals and fish (Korn and
Bennett 1972). Gap junctions were postulated to be largely
localized on soma and proximal dendrites to facilitate initiation
of dendritic spikes thereby synchronizing motoneuron discharge [e.g., during fast phases of nystagmus (Korn and Bennett 1975)]. Because the caudal subgroup of abducens motoneurons was observed to feature more gap junctions than those
in the rostral subdivision (Sterling 1977), it was predicted
to contain more phasic motoneuronal responses (Gestrin and
Sterling 1977). However, the physiological profiles of motoneurons identified in the two subdivisions were demonstrated
to be the same (Pastor et al. 1991) as were the observed
distribution of gap junctions (present study). Data herein also
will show that electrotonic coupling is present in all of the
semicircular canal pathways that produce compensatory, i.e.,
not rapid, eye movements. Hence one goal was to achieve a
consensus with prior work by employing a broad approach to
analyze all of the vestibuloocular circuitry.
In this study we evaluate the evidence for reciprocal excitatory and inhibitory synaptic connections to extraocular motoneurons on the basis of electrophysiological, structural,
and immunohistochemical methods. In comparison with
mammals, postsynaptic inhibition appeared to be of smaller
amplitude and efficacy because of less negative equilibrium
potentials rather than changes in inhibitory synaptic conductance. Electrotonic coupling was highly variable between
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
2767
to 4 h. Subsequently, the sections were processed for the immunohistochemical localization of GABA using an antibody generated
in rabbit against GABA conjugated to bovine serum albumin (Immunonuclear). Sections were immersed for 1 h in 10% normal
goat serum (NGS) in phosphate-buffered saline (PBS; 0.01 M)
containing 0.1% Triton X-100, and then incubated in primary antibody (1:2,000 to 5,000) containing 1% NGS and 0.1% Triton X100 in PBS for 18–24 h at 47C with constant agitation. As a
control, some sections were incubated in the same solution without
either the addition of primary antibody or with antibody preabsorbed with GABA. After the incubation, sections were washed
for 30 min in three changes of PBS, and immersed in biotinylated
anti-rabbit IgG (1:200; Vector) in PBS with 1% NGS for 90 min
at room temperature with constant agitation. After the secondary
antibody incubation, the tissue was washed in PBS and immersed
in an avidin:biotin-HRP complex (1:100; Vector) for 90 min. Sections then were washed through two changes of PBS over 20 min
followed by two changes of 0.1 M phosphate buffer (pH 7.4) for
20 min. This chromogen produced a brown diffuse reaction prod-
/ 9k11$$my40 J711-6
uct. The material was rinsed through several changes of 0.1 M
phosphate buffer, and then placed on pretreated microscope slides,
dehydrated, and mounted with Permount for light microscopic examination using brightfield and Nomarski differential interference
contrast optics.
Electron microscopy
The extraocular muscles of 10 goldfish were injected with 2 ml
of 10% HRP. After a survival time of 1–3 days, the animals were
deeply reanesthetized and perfused with 1% paraformaldehyde and
1.25% glutaraldehyde in 0.1 M phosphate buffer with 0.002 calcium chloride (pH 7.4). After fixation, the brains were removed
and immersed in 0.1 M phosphate buffer with 8% dextrose overnight. Serial 50- to 75-mm coronal sections through the oculomotor
nuclei were cut on a vibratome, and reacted with tetramethylbenzidine or diaminobenzidine to visualize the extraocular motoneuron
pools (Graf and McGurk 1985). Sections containing the oculomotor nuclei were trimmed, postfixed for 2 h in 1% osmium tetroxide
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
FIG . 1. Postsynaptic potentials in rightside oculomotor neurons (rOc) following
vestibular nerve stimulation. A and B: inferior rectus motoneuron recorded with a KCl
electrode. A: antidromic identification. B:
excitatory postsynaptic potential (EPSP)
elicited from the contralateral (left, lVes),
and inhibitory postsynaptic potential
(IPSP) from the ipsilateral (right, rVes),
vestibular nerves. Unitary potentials indicated by arrows. Electrotonic coupling occurred at 0.9 ms (short arrow) and chemical conductances at Ç1.3 ms (long arrow).
C – E: inferior oblique motoneuron recorded with a K Citrate electrode. IPSPs
were evoked by ipsilateral (rVes) stimulation during hyperpolarizing (C and D) and
depolarizing (E) current injection. IPSP
equilibrium potential was found at 04.0 nA
(D). Small arrows in D indicate the axonal
(M) spike and initial segment/soma dendritic (IS-SD) invasion inflections during
antidromic invasion that was blocked in C
during hyperpolarizing current injection.
The long vertical line indicates IPSP onset.
F–J: superior rectus motoneuron recorded
with a K Acetate electrode. F: IPSP elicited
by ipsilateral (rVes) stimulation at twice
threshold intensity (2 1 T) followed by the
antidromic response indicating intrasomatic recording. G: EPSP following contralateral (lVes) stimulation at 4 times
threshold (electrotonic component is indicated by the small arrow). H and I: responses to stimulation of the ipsilateral vestibular nerve (rVes) at 4 times threshold
and 10 times threshold, respectively, indicating no change in IPSP amplitude. J: determination of IPSP latency (small vertical
double arrow) by comparison of intracellular (In) and extracellular (Ex) records.
Amplitude calibrations for A– E are shown
in A and for F–J in F. Time scales are
indicated in the 1st record for each set of
traces.
2768
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
in 0.1 M phosphate buffer with 7% dextrose and 1.5% potassiumferricyanide, stained en bloc for 2 h with 0.5% uranyl acetate in
maleate buffer, dehydrated in methanol and propylene oxide, and
embedded in TAAB 812 resin between vinyl plastic microscope
slides and coverslips.
Selected loci containing oculomotor neurons were analyzed by
serial ultrathin sectioning and electron microscopy. Serial 5-mm
sections through the plastic-embedded vibratome sections were cut,
examined by light microscopy, and remounted on blank BEEM
capsules from which ultrathin sections were cut and collected on
single-slot Formvar-coated grids. Sections were then stained with
uranyl acetate and lead citrate, and examined and photographed
with a Zeiss EM-10CA electron microscope at 11,500, 18,000,
and at sites of synaptic contact zones, at 125,000 (Spencer and
Baker 1990).
RESULTS
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
Localization of oculomotor nuclei
At the outset of each experiment, the trochlear nucleus
was located first by the characteristic antidromic field potential profile initiated from the contralateral orbit (Baker et
al. 1973). Subsequently, superior rectus motoneurons were
identified in the caudal oculomotor complex after stimulation
of the contralateral orbit. Inferior rectus (rostrodorsal) and
inferior oblique (caudoventral) motoneurons were recognized by antidromic activation from the ipsilateral orbit
along with the rostral and caudal subgroups of abducens
motoneurons in the posterior brain stem. Stimulation of the
ipsi- and contralateral vestibular nerves also distinguished
the MLF borders.
Antidromic oculomotor field potentials following electrical stimulation of the orbit were found to extend from 0.75–
1.0 mm lateral from the visually observed midline of the
oculomotor complex. Most of the respective motoneuron
somata are located directly adjacent to the midline (Fig. 5,
D and F) except for the dorsal and lateral extensions to
either side (Graf and McGurk 1985; Pastor et al. 1991).
Intracellular records from somata are illustrated in Figs. 1
and 2, even though the most frequent intracellular recording
site was from dendrites (as in Fig. 3). Motoneuron somata
were identified by antidromic activation characterized by
initial segment/soma dendritic invasion (IS-SD) followed
by a large after depolarization typical of an intrasomatic
penetration (Fig. 1, A, D, and F).
General organization of inhibitory and excitatory
vestibular pathways
Postsynaptic potentials following vestibular nerve stimulation were recorded in inferior rectus, inferior oblique, and
superior rectus motoneurons in Fig. 1, A and B, C–E, and
F–J, respectively. EPSP and IPSP characteristics varied between motoneuronal populations; however, the ipsi- and contralateral spatial organization of inhibition and excitation
was invariant between experiments. At normal resting membrane potential, stimulation of the contralateral vestibular
nerve produced a short-latency depolarization consisting of a
presumed electrotonic and a subsequent chemical component
initiated at Ç0.9 and 1.3 ms, respectively (Fig. 1B, lVes).
Stimulation of the ipsilateral vestibular nerve produced a
small-amplitude 1-mV hyperpolarization, typically at a la-
/ 9k11$$my40 J711-6
FIG . 2. Vestibular evoked synaptic potentials in a right-side trochlear motoneuron after contralateral (lVes, in A and B) and ipsilateral (rVes in C – E)
stimulation before (A and C) and after chloride loading (B and D–F). A and
B: electrotonic (small arrows) and chemical EPSPs after lVes stimulation at 4
times threshold. C – E: IPSPs after 4 times threshold rVes stimulation with the
onset of inhibitory conductance indicated by small arrows in C. D and E:
chloride loading increases the efficacy of the inhibitory synaptic conductance as
compared with C; however, the electrotonic and chemical EPSPs where unaltered
(compare B with A). Large-amplitude unitary inhibitory potentials are shown
during ipsilateral (rVes, E) and without vestibular stimulation (F). Amplitude
calibrations for A– D are shown in A and those for E and F in E. Time scales
are shown in order of record appearance.
tency of 1.3 ms (Fig. 1B, rVes). In general, after 5–10
min of intrasomatic recording with KCl electrodes, smallamplitude, all-or-none depolarizations and hyperpolarization
(arrows in Fig. 1B) were recorded (see also Fig. 2, E and
08-08-97 12:57:37
neupa
LP-Neurophys
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
2769
F). The leakage of chloride ions from the microelectrode
appeared to affect the unitary IPSPs in a nonuniform fashion
(Korn et al. 1992).
Because the major goal of these electrophysiological experiments was to establish the presence of inhibitory synaptic
potentials in extraocular motoneurons, intrasomatic records
were obtained with electrodes filled with either K Citrate
(Fig. 1, C–E) or K Acetate (Fig. 1, F–J). Large hyperpolarizing IPSPs were recorded in the presence of extrinsic membrane depolarization ( /4.0 nA, Fig. 1E). The application
of hyperpolarizing current ( 04.0 nA, Fig. 1D) demonstrated
the inhibitory synaptic effect to be associated with a membrane conductance exhibiting an equilibrium potential
slightly more negative than the resting potential. In fact, the
IPSP equilibrium potential was near the point where the
antidromic invasion of the IS-SD spike was blocked (arrows
in Fig. 1D). Further injection of hyperpolarizing current
( 06.0 nA, Fig. 1C) reversed the inhibitory synaptic conductance to a depolarization, and the antidromic invasion of the
IS-SD component was completely blocked leaving only the
/ 9k11$$my40 J711-6
M spike. Spontaneous and/or evoked unitary synaptic potentials like those illustrated in Figs. 1B and 2, E and F, were
absent from recordings obtained with acetate or citrate filled
electrodes.
Vestibular input to identified superior rectus and trochlear
motoneurons
The presence of contralateral excitation and ipsilateral inhibition was also documented in a superior rectus motoneuron illustrated in Fig. 1, F–J. This series of intracellular
records was obtained with a K Acetate electrode that did
not appear to affect the equilibrium potential for the IPSP
(Fig. 1, H and I). Stimulation of the ipsilateral vestibular
nerve (rVes) at two times threshold (2 1 T) produced a
hyperpolarizing IPSP preceding, but not blocking, antidromic activation of the superior rectus motoneuron (Fig.
1F). Stimulation of the contralateral vestibular nerve (lVes)
at 4 times threshold evoked a short-latency electrotonic depolarization (arrow in Fig. 1G) followed by a more slowly
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
FIG . 3. Dendritic recording from a right-side
oculomotor neuron (rOc, inferior rectus or inferior oblique) with a KCl electrode. A– D: antidromic response at 1 times threshold (A), 2 times
threshold (B), and 4 times threshold ( C and D)
with 03.0 nA (C) and 05.0 nA (D) membrane
hyperpolarization. Arrows in A indicate small-amplitude all-or-none action potentials. E and F:
EPSPs recorded at 2 times threshold (E) and 4
times threshold (F) with different time bases after
contralateral vestibular nerve (lVes) activation
(arrows indicate onset of electrotonic coupling).
G: IPSP (arrow indicates onset) following ipsilateral vestibular nerve (rVes) stimulation presumed
to be reversed after chloride leakage shift of the
equilibrium potential. H: reversed IPSP at 4 times
threshold after 3 min ( 05.0 nA) chloride injection. Amplitude calibrations for A– D are shown
in D, and for E– H in G.
2770
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
Intradendritic recording from oculomotor and trochlear
motoneurons
The majority of intracellular records from extraocular motoneurons were obtained from presumed dendritic, and not
somatic, penetrations (Fig. 3). This assessment was based on
electrophysiological criteria as well as comparison between
medial and lateral MLF penetration sites. Dendritic, as opposed to somatic, records occurred in a ratio of at least 5:1.
The EPSP and IPSP profile, including the response to current
clamp conditions, differed from somatic sites (Fig. 3, G and
H), except for the rapid and sustained response to chloride
injection (H).
A large-amplitude, stable resting membrane potential of
065 to 072 mV was encountered immediately after successful penetration of a dendrite. Frequently, orbital stimulation
at a low-intensity straddling antidromic threshold produced
small-amplitude all-or-none action potentials exhibiting a
fast rise time and two separate exponential decays (Fig. 3,
A and B). Stimulation at a slightly higher intensity (2 times
threshold) produced a large-amplitude all-or-none action potential with a uniform, fast rising phase exhibiting either two
(Fig. 3B) or three (Fig. 3C) component, exponential-like
falling phases. Generally, slight membrane hyperpolarization
(Fig. 3D, 05.0 nA) could block the large-amplitude invasion leaving a small-amplitude, all-or-none (5–20 mV) de-
/ 9k11$$my40 J711-6
polarization. We interpret this type of recording as a sequential invasion of the antidromic action potential of the initial
segment (Fig. 3D) and the soma-dendritic compartment
(Fig. 3, B and C). This action-potential profile was largely
only seen during antidromic activation when the action potential progresses from the axon toward the dendrites. By
contrast, contralateral vestibular nerve stimulation (lVes)
produced electrotonic (arrows in Fig. 3, E and F) and chemical EPSPs eliciting action potentials with fast rising phases
and a single exponential falling phase. These observations
suggest that in this case the regenerative current was being
initiated distal in the dendritic tree and propagated toward
the soma.
In representative intradendritic records, irrespective of microelectrode type, electrical stimulation of the ipsilateral (inhibitory) vestibular nerve (rVes) produced a short-latency
(1.3 ms), small-amplitude membrane depolarization (Fig.
3, G and H). After recording with a chloride microelectrode,
for only 1 min, the reversed IPSPs also initiated action potentials (Fig. 3G). An active chloride injection (3 min 05.0
nA) altered the IPSP equilibrium potential such that rVes
stimulation at 4 times threshold produced a high-frequency
discharge of the proximal soma-dendritic compartment in
the presence of a 02.0-nA holding current (Fig. 3H). During
chloride loading, lVes stimulation at 4 times threshold elicited EPSPs with either single or multiple dendritic action
potentials (Fig. 3F).
In summary, disynaptic inhibition and excitation including
electrotonic coupling could be documented in all vertical
recti [exclusive of the medial rectus (cf., however, Suwa
and Baker 1996)] and oblique motoneurons. Although the
equilibrium potential for the IPSP, around 065 mV, was
never as negative as that observed in mammalian oculomotor
motoneurons, 080 mV (Llinas and Baker 1972), the effect
of chloride loading in either the soma or dendrite was comparable. In all oculomotoneurons, IPSPs were quickly reversed
and the membrane depolarization could initiate action potentials spreading throughout the soma-dendritic compartment.
Intracellular records from abducens motoneurons and
internuclear neurons
Abducens motoneurons were antidromically identified in
the rostral and caudal subgroups of the abducens nucleus in
the posterior brain stem (Pastor et al. 1991). Intracellular
records were also obtained from nonantidromically activated
putative abducens internuclear neurons located adjacent to,
and somewhat dorsolateral to the caudal subgroup of abducens motoneurons. In both types of neurons, electrical stimulation of the contralateral vestibular nerve (rVes) evoked
robust, short-latency electrotonic EPSPs with a latency of
0.5 ms followed by chemical depolarizations at 1.1–1.3 ms
(Fig. 4A). Stimulation of the ipsilateral vestibular nerve
revealed IPSPs with a latency of 1.4 ms that initially were
always hyperpolarizing in direction. Soon after penetration
with chloride containing electrodes, IPSPs were quickly reversed. As illustrated in Fig. 4 B, injection of hyperpolarizing
current from either K Acetate or K Citrate electrodes shifted
the membrane resting potential toward the IPSP equilibrium
potential, and the hyperpolarization could be entirely reversed from an intrasomatic electrode location. Like in the
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
rising EPSP that initiated action potentials. Stimulation of
the ipsilateral vestibular nerve at 4 times threshold and 10
times threshold (H and I) did not result in larger amplitude
potentials indicating that the IPSP was already near maximum at 2 times threshold. The 1.3-ms latency for this IPSP
was determined by comparing the adjacent intra- (In) and
extracellular (Ex) responses (Fig. 1J).
Trochlear motoneurons were the easiest to identify because peripheral current spread to the IIIrd nerve was minimal (Fig. 2, A–F). These motoneurons always exhibited
large-amplitude inhibitory (reversed IPSPs) as well as excitatory potentials following ipsilateral and contralateral vestibular stimulation, respectively (Baker et al. 1973; Llinas and
Baker 1972; Precht and Baker 1972). A rather distinct electrotonic component was always associated with excitatory
potentials (Fig. 2, A and B, see also Fig. 1B). Trochlear
motoneurons were frequently recorded after chloride loading
as had been previously done in mammalian trochlear neurons
(Llinas and Baker 1972). Seemingly much smaller amounts
of chloride ( õ1 min injection) reversed the equilibrium potential for the IPSP to values close to that presumed for the
EPSP; however, little change was noted in the electrotonic
and/or chemically evoked EPSP (compare Fig. 2, A and B).
After intrasomatic chloride loading of trochlear motoneurons, electrical stimulation of the ipsilateral vestibular nerve
always produced large-amplitude IPSPs that even could sustain action potentials in the soma (Fig. 2, D and E; also see
Fig. 3, E–H). These action potentials were superimposed
on large-amplitude (30–40 mV) depolarizations including
evoked unitary IPSPs ranging from 5 to 15 mV in amplitude
(Fig. 2E). As had been previously demonstrated in mammalian trochlear motoneurons (Llinas and Baker 1972), inhibitory synaptic transmitter (GABA) appeared to be continuously released from presynaptic terminals (Fig. 2F).
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
2771
GABA immunohistochemistry of vestibulooculomotor
pathways
FIG . 4. Intracellular recording from a right-side abducens motoneuron and
abducens internuclear neuron after electrical stimulation of the horizontal canal
nerves. A: EPSPs after gradual increase of contralateral vestibular nerve (lVes)
stimulation ranging from 1.5 to 10 times threshold recorded with a K Acetate/
LiCl electrode. Electrotonic and chemical depolarizations are indicated by the
1st and 2nd arrow, respectively. B: IPSPs elicited at constant 4 times threshold
ipsilateral vestibular nerve (rVes) stimulation. Latency indicated by vertical
arrow. Current clamp was varied from /4.0 to 08.0 nA to illustrate the discrete
changes in IPSP amplitude from negative to positive with the equilibrium potential occurring at about 05.0 nA (middle trace). The hyperpolarizing IPSP profile
recorded at resting membrane potential is indicated by the oblique arrow. C:
EPSPs and IPSPs in an abducens internuclear neuron following 4 times threshold
contralateral (lVes) and ipsilateral (rVes) vestibular nerve stimulation, respectively. Oblique arrows indicate the onset of electrotonic and chemical depolarization, respectively. Amplitude and times scales are indicated for each set of
records.
/ 9k11$$my40 J711-6
In five goldfish, GABA antibodies were used in conjunction with retrograde labeling of oculomotor and abducens
motoneurons to localize putative inhibitory neurons in the
vestibular nuclei, identify their axonal trajectories in the
MLF, and establish which subdivisions of the extraocular
motor nuclei contained terminal arborizations. Figure 5A
shows a low magnification coronal section containing the
entire hindbrain and cerebellum. GABA immunostaining is
notably robust in the Purkinje cell and external granular
cell layer of the cerebellum (CC and EG) along with the
octavolateral column including the vestibular and mechanosensory medial nuclei (VN and M) (McCormick and Braford 1994). The brain stem consisted of largely unlabeled
fiber tracts and few GABAergic cells in the medulla, but
terminal arborizations were present in most mid- and hindbrain nuclei.
A higher magnification photomicrograph through the caudal end of the anterior octaval nucleus, located just dorsal
to the descending trigeminal (V) and secondary gustatory
(trg) tract, always showed vestibular neurons with intense
reaction product (inset in Fig. 5B). In general, GABAergic
neurons appeared to be more dense in the anterior octaval
complex and relatively sparser in the caudal descending octaval subdivisions. Axons of inhibitory vestibular neurons
could be followed through the vestibular commissure into
the MLF (Fig. 5C). The rostral MLF axonal trajectory
tended to show a preferential location to the ventral MLF,
especially closer to the oculomotor complex (Fig. 5C). Distinctly fewer GABAergic axons were found in the descending MLF (not illustrated).
In two experiments, oculomotor (Fig. 5, D, F, and H) and
abducens (Fig. 5, E and I) motoneurons were retrogradely
labeled with HRP and followed by GABA localization. Dendrites of oculomotoneurons extend laterally into the MLF
(Fig. 5D) and the axons leave via the ipsilateral oculomotor
nerve. The coronal view at the mid level of the oculomotor
complex in Fig. 5D is shown at higher power in Fig. 5F to
indicate reaction product in the axons of the MLF and terminals in the area containing presumed inferior rectus, inferior
oblique, and superior rectus motoneurons. GABAergic termination on a HRP retrogradely labeled motoneuron (Fig.
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
mammalian abducens nucleus (Spencer et al. 1989), the
efficacy of inhibition could be demonstrated by the blockade
of the antidromic field potential following electrical stimulation of the ipsilateral horizontal canal nerve (not illustrated).
After contralateral vestibular nerve stimulation (lVes)
EPSPs with electrotonic and chemical components were also
recorded from abducens internuclear neurons in both the
rostral and caudal subgroups (Baker et al. 1994). Large
amplitude hyperpolarizing IPSPs were recorded after ipsilateral vestibular nerve (rVes) activation (Fig. 4C). In summary, these electrophysiological data document a pattern of
reciprocal canal-related disynaptic excitation and inhibition
in all extraocular subdivisions including the abducens internuclear neurons. As noted in the DISCUSSION, the synaptic
profiles in each subgroup were also found to be remarkably
similar to their mammalian homologues.
2772
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
/ 9k11$$my40 J711-6
08-08-97 12:57:37
neupa
LP-Neurophys
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
5H) is compared with an unlabeled motoneuron (Fig. 5G)
to illustrate the density of terminal boutons surrounding the
soma. By contrast, low (Fig. 5E) and high (Fig. 5I) power
photomicrographs through either the rostral or caudal abducens nuclei showed the absence of any axonal and terminal
arborizations containing GABA reaction product. In summary these observations suggest a presumed inhibitory role
for GABA throughout the oculomotor complex; however,
not in the abducens nucleus where presumably glycine is
thought to be the inhibitory transmitter (Spencer et al. 1989).
Hence the pattern of inhibition as well as the putative transmitter profile appears to have been conserved in the teleostmammalian vestibulooculomotor plan (Evinger 1988).
Electron microscopy of oculomotor nuclei
ings, in general, showed axosomatic and axodendritic synaptic endings (as in Fig. 6, B and C) and were either of pure
chemical nature (Fig. 7, A and B) or exhibited chemical
contacts and gap junctions consistent with electrotonic coupling (Fig. 7, E and F). Chemical synapses with spheroidal
vesicles were characterized by asymmetric pre/postsynaptic
membrane specializations (Fig. 7A, inset). Their ultrastructure was also characterized by the presence of pale matrix
mitochondria (Fig. 7, A and B). These types of axosomatic
endings that contained large numbers of synaptic vesicles
were never observed to establish gap junction contacts.
Mixed chemical and electrotonic axosomatic synaptic endings contained few spheroidal vesicles. The chemical synaptic
contacts were also characterized by asymmetric pre- and/or
postsynaptic membrane specializations with an Ç20-nm intersynaptic space (Fig. 7E, top left inset). Gap junction contacts
displayed an Ç2-nm separation between the pre- and postsynaptic membranes (Fig. 7E, bottom right inset). Both types
of contacts were rarely observed in relationship to the same
postsynaptic profile (arrows in Fig. 7F). In this case, as in
other synaptic profiles, the two contacts were found to be
spatially separated. The presumed inhibitory axosomatic synaptic endings (Fig. 7, C and D) were characterized by symmetric pre-/postsynaptic membrane specializations (Fig. 7D, inset). They contained pleiomorphic vesicles and established
only chemical synaptic contacts on motoneurons, i.e., these
endings never exhibited gap junctions.
Abducens nucleus motoneurons (Fig. 8A) also showed
typical cytoplasmic organelles similar to those observed in
cats (Spencer and Baker 1983) including cisternal arrays
of granular endoplasmic reticulum. In most instances, the
nucleus also exhibited invaginations. As a rule, abducens
motoneurons displayed a more densely packed arrangement
of axosomatic synaptic endings than found in oculomotor
motoneurons (compare Figs. 6B and 8B). In addition, synaptic endings were differentially distributed along soma,
proximal dendrites, and distal dendrites in abducens motoneurons in respect to the content of spheroidal and pleiomorphic vesicles (Fig. 8C). In general, a mixture of synaptic
contacts containing pleomorphic and spheroidal vesicles was
found at the soma and at the proximal dendrites (Fig. 8 B),
whereas predominantly spheroidal vesicle containing synapses were present at distal dendrites (Fig. 8C). Comparatively more synaptic contacts contained pleiomorphic vesicles except for those located on the dendritic areas.
Abducens motoneurons also possessed synaptic contacts
of presumed inhibitory and excitatory nature containing
pleiomorphic vesicles and spheroidal vesicles, respectively
FIG . 5. Immunohistochemical localization of g-aminobutyric acid (GABA). A: coronal section through the brain stem
and cerebellum at the level of anterior octaval (AO) vestibular subgroup. B: high magnification of the AO and M subdivisions
of the octavolateral complex (at arrow in A) with the inset showing cells containing immunoreaction product. C: coronal
section showing presumed GABAergic axons in the medial longitudinal fasciculus (MLF) with the higher magnification
inset indicated by the arrow. D–I: double-labeled sections showing horseradish peroxidase (HRP) retrograde identification
of extraocular motoneurons with GABA staining. D: coronal section through the midbrain of the goldfish at the mid level
of the oculomotor nucleus. E: coronal section through the rostral subgroup of abducens motoneurons. F: high magnification
view of the bilateral oculomotor complex showing GABAergic reaction product in MLF axons and terminals surrounding
HRP and unlabeled motoneuronal somata. G and H: GABA terminal patterns on 2 motoneurons indicated by arrows in F.
I: high-magnification view of abducens motoneuronal somata as indicated by the arrow in E. Abd, abducens nucleus; AO,
anterior octaval nucleus; CC, corpus cerebellum; EG, ementiae granularis; IV, 4th ventricle; LL, lateral lemniscus; mV,
trigeminal nucleus; M, medial nucleus; MLF, medial longitudinal fasciculus; nIII, oculomotor nerve; Oc, oculomotor nucleus;
RF, reticular formation; trg, secondary gustatory tract; VN, vestibular nuclei, VIIIn, vestibular nerve; VIIn, facial nerve
(abbreviations, after McCormick and Braford 1994).
/ 9k11$$my40 J711-6
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
To provide evidence for the presence of electrotonic excitation and chemical inhibition, the pattern and mode of terminal arborization was studied in both the oculomotor and
abducens motor nuclei of six goldfish retrogradely labeled
with HRP (Figs. 6A and 8A, insets, respectively). Oculomotor neurons were identified first in thick plastic sections (e.g.,
Fig. 6A, bottom inset) that were subsequently thin sectioned
for electron microscopy (Spencer and Baker 1983). Cytoplasmic organelles typical of other vertebrates, e.g., Golgi
apparatus and mitochondria, were observed (Fig. 6A). Cisternal arrays of granular endoplasmic reticulum (ger) were
well developed, and in most instances the nucleus also exhibited the type of invaginations seen in mammalian oculomotor
neurons (Spencer and Baker 1983). A moderately high density of axosomatic synaptic endings was found in all oculomotor subdivisions and trochlear motoneurons (e.g., arrows
in top of Fig. 6A). The density of axosomatic contacts varied
among motoneurons, but the number of axodendritic synaptic endings always appeared lower, despite the relatively
large size and length of the dendrites (arrows in Fig. 6 B). In
general, the large-sized dendrites exhibited few axodendritic
synapses (Fig. 6C). Most synaptic contacts were located on
medium and small caliber dendrites in the dendritic neuropile
lateral to the oculomotor nucleus. All such synapses contained spheroidal synaptic vesicles and possessed chemical
contact zones characterized by asymmetric pre-/postsynaptic membrane specializations (Figs. 6C and 7, A and B). Gap
junctions were rarely associated with axodendritic synaptic
endings, particularly those on distal dendrites.
Two kinds of presumed excitatory synapses and one type
of a presumed inhibitory synapse could be discerned on
oculomotor motoneurons. Putative excitatory synaptic end-
2773
2774
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
(Fig. 9A and C). Excitatory synapses existed in the chemical
only (Fig. 9, A–C), as well as the mixed chemical/electrotonic configuration (Fig. 9D). All chemical synaptic contact
zones in the abducens nucleus were characterized by asymmetric pre-/postsynaptic membrane specializations (Fig.
9C). Gap junction contacts were separated by a 2-nm cleft
between pre- and postsynaptic membranes (compare with
Fig. 7E). In conclusion, the ultrastructure of terminal arborizations differed little between oculomotor and abducens motoneurons.
DISCUSSION
Because all vertebrate species studied to date represent a
widely divergent taxonomy exhibiting a wide repertoire of
eye movements, we believe these new data support the hypothesis that the ancestral, hence primitive, vestibulooculomotor plan was based on the utilization of semicircular canal-specific disynaptic excitation and inhibition. The corroborative chemical, structural, and physiological evidence
/ 9k11$$my40 J711-6
argues that the neurons, transmitters, pathways, and synaptic
arborizations on extraocular motoneurons have been highly
conserved and as a result are likely homologous traits that
are embryologically linked throughout vertebrate phylogeny.
Electrophysiology
In antidromically identified oculomotor motoneurons innervating inferior/superior oblique and inferior/superior
recti muscles, electrical stimulation of the contralateral vestibular nerve produced short-latency disynaptic electrotonic
EPSPs (0.5–0.7 ms) followed by disynaptic chemical depolarizations with latencies of 1.0–1.3 ms (Figs. 1 and 2).
Stimulation of the ipsilateral vestibular nerve produced
small-amplitude hyperpolarizations beginning at a latency
of 1.3–1.7 ms that gradually increased in amplitude to peak
at 4.0–7.0 ms. Manipulation of membrane potential by current injection through the microelectrode demonstrated the
short-latency ipsilateral hyperpolarization to be associated
with a membrane conductance exhibiting an equilibrium po-
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
FIG . 6. Light and electron microscopic visualization of oculomotor motoneurons illustrating the distribution of synaptic
endings. A, top inset: darkfield photomicrograph of a 50-mm vibratome section through the mid level of the oculomotor
nucleus containing motoneurons labeled by retrograde transport of HRP [tetramethylbenzidine (TMB) reaction]. Bottom
inset: semithin (1 mm) section through 3 motoneurons [diaminobenzidine (DAB) reaction, toluidine blue counterstain]. The
oculomotoneuron in the electron micrograph is indicated by the asterisk and the high density of axosomatic synaptic endings
by arrows. ger, cisternal arrays of granular endoplasmic reticulum; G, Golgi apparatus; m, mitochondria; Nuc, nucleus. B:
low-magnification electron micrograph of a primary dendrite (d) emanating from the soma (S) of a labeled oculomotor
neuron with few synaptic contacts (arrows). C: low-magnification electron micrograph of the dendritic neuropile (d) lateral
to the oculomotor nucleus with numerous axon contacts in the central core (arrows). Calibrations in A are 100 mm (top
inset), 20 mm (bottom inset), and 5 mm (electron micrograph); B, 5 mm; and C, 1 mm.
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
2775
tential slightly more negative than resting potential. Reversal
of this IPSP by injected current, as opposed to Cl injection,
was highly dependent on intrasomatic (Fig. 1) as opposed
to intradendritic (Fig. 3) recording sites.
Injection of chloride ions significantly displaced the equilibrium potential in the depolarizing direction to the extent
that IPSP rereversal was closer to the equilibrium potential
for the contralateral EPSP. Following chloride loading, efficacy of the inhibitory synaptic conductance was demonstrated by the large difference in amplitude of the two evoked
depolarizations. The ability to change the equilibrium potential with chloride ions is similar to that utilized to study
recurrent and vestibular commissural inhibition in the goldfish Mauthner cell (Faber and Korn 1975). This model system has been immensely informative concerning inhibitory
postsynaptic conductances in vertebrates, especially that related to glycinergic transmission (e.g., the abducens nucleus;
/ 9k11$$my40 J711-6
Fig. 4) (Faber and Korn 1987, 1988). Of particular interest
is the extrapolation of current- and voltage-clamp data showing that the chloride channels exhibit a steep voltage dependency such that the lower equilibrium potential in goldfish
extraocular motoneurons as compared with mammals (e.g.,
065 vs. 090 mV) would not be expected to alter canalrelated inhibition. Because both the magnitude and duration
of inhibitory responses would in fact be noticeably increased
in the presence of larger excitation, we conclude that the
reciprocal semicircular canal inhibitory and excitatory pathways play an equally important role in coordinated eye
movement.
In the horizontal canal system, intracellular records from
antidromically identified abducens motoneurons revealed robust short-latency, disynaptic electrotonic EPSPs at 0.5-ms
latency followed by disynaptic chemical depolarizations at
1.3 ms after electrical stimulation of the contralateral vestib-
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
FIG . 7. Organization of chemical and
electrical synaptic contacts on somatic and
dendritic profiles in the oculomotor nucleus. A and B: electron micrographs of
axosomatic synaptic endings establishing
chemical synaptic contacts (large arrow)
characterized by asymmetric pre- and/or
postsynaptic membrane specializations (A,
inset). C and D: electron micrographs of
axosomatic synaptic endings establishing
chemical synaptic contacts (large arrow)
characterized by symmetric pre- and/or
postsynaptic membrane specializations (D,
inset). E and F: electron micrographs of
synaptic endings that establish both chemical (large arrow) and electrical gap junction contacts (small arrows). Insets in E
illustrate chemical (top left) and gap junction (bottom right) contact zones. d, dendrite; s, soma. Calibrations: A, 0.25 mm (A,
inset, 0.1 mm); B, 0.5 mm; C and D, 0.5
mm (D, inset, 0.1 mm); E, 0.25 mm (E,
inset, 0.05 mm); and F, 0.5 mm.
2776
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
ular nerve. As expected, stimulation of the ipsilateral vestibular nerve revealed disynaptic IPSPs at 1.4 ms that were
also reversed after injection of current and/or chloride ions.
Because disynaptic inhibition and excitation was demonstrated for adjacent rostral and caudal subgroups of abducens
internuclear neurons, the reciprocal horizontal canal organization appears to be similar in goldfish and mammals. The
/ 9k11$$my40 J711-6
larger IPSP amplitude and more negative equilibrium potential for goldfish abducens versus oculomotor neurons suggests that either the density of inhibitory boutons and/or
the amplitude of the channel conductances are larger for
glycinergic versus GABAergic terminal arborizations, respectively. This difference in inhibitory efficacy between
comparable motoneuron groups was not detected in mam-
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
FIG . 8. Light and electron microscopic
visualization of abducens motoneurons indicating the distribution of synaptic endings. A, bottom inset: darkfield photomicrograph of motoneurons located in the
caudal subgroup labeled by retrograde
transport of HRP (TMB reaction). Top inset: brightfield photomicrograph of a semithin (1 mm) section (DAB reaction), and
the motoneuron in the electron micrograph
is indicated by an asterisk. Axosomatic
synaptic contacts are shown by arrows. B:
low-power electron micrograph of a large
caliber primary dendrite showing little variation in the distribution of spheroidal and
pleiomorphic vesicles. C: low-power electron micrograph of distal dendritic neuropile showing chemical synaptic contact
zones with predominately pleiomorphic
vesicles at the proximal site (left) and almost exclusively spheroidal vesicles at the
distal end (right). d, dendrite; p, pleiomorphic vesicle; s, spheroidal vesicle. Calibrations in A are as follows: bottom inset, 100
mm; top inset, 20 mm; and electron micrograph, 5 mm; B and C, 2 mm.
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
2777
mals (Spencer et al. 1992). By contrast, in all species, GABAergic inhibition has been found to be the most robust in
trochlear motoneurons.
Immunochemistry
Double label experiments using GABA antibodies and
HRP injections localized inhibitory neurons in the anterior
octaval subdivision of the vestibular nucleus (Fig. 5)
(McCormick and Braford 1994). Axonal trajectories were
identified toward, and into, the ipsilateral MLF along with
observing terminal arborizations distributed largely on the
somata and proximal dendritic trees of trochlear and oculomotor, but not abducens, motoneurons. Thus the presence,
spatial distribution, and magnitude of GABAergic inhibition
in oculomotor neurons produced by presumed GABAergic
neurons of the vertical vestibular system also appears to be
analogous between fish and mammals (Spencer et al. 1992).
/ 9k11$$my40 J711-6
Immunohistochemical observations in the abducens nucleus,
especially the absence of GABA axons and terminals, was
consistent with findings in mammals demonstrating glycine
to be the inhibitory neurotransmitter of second-order vestibular neurons related to the horizontal vestibuloocular reflex
(Spencer et al. 1989, 1992).
Recently both GABA and glutamic acid decarboxylase
were localized in the extraocular motor nuclei of monkey
and cat (Spencer et al. 1992). The densest GABA-related
termination patterns were found in the trochlear nucleus, and
the lightest, in the abducens nucleus. Comparison of single
versus multiple postsynaptic profiles on respective somadendritic profiles suggested two separate inhibitory GABAergic synaptic inputs onto oculomotor and trochlear motoneurons. For example, in goldfish and mammals, vertical
saccade-related premotor neurons in the rostral interstitial
nucleus of the MLF and anterior octaval vestibular neurons
are the most likely candidates. Hence it is well established
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
FIG . 9. Organization of chemical and
electrical synaptic contacts on somatic and
dendritic profiles of abducens motoneurons. A: electron micrograph of axosomatic
synaptic endings on motoneurons containing either spheroidal or pleiomorphic
synaptic vesicles. B: electron micrograph
of a nodal synapse associated with a motoneuron soma. C: electron micrograph illustrating vesicle and pre- and/or postsynaptic
membrane morphology of presumed inhibitory and excitatory chemical axosomatic
endings. D: electron micrograph of axosomatic synaptic vesicles establishing both a
chemical synaptic contact with intermediate dense line and postsynaptic density including an adjacent electrotonic synaptic
contact (gap junction; small arrows). Inset
in D shows the 2-nm separation of the gap
junction. a, axon; GJ, gap junction; idl, intermediate dense line; psd, postsynaptic
density; p, pleiomorphic vesicle; s, spheroidal vesicle. Calibrations: A, 2 mm; B, 1 mm;
C, 0.1 mm; and D, 0.1 mm (D, inset, 0.05
mm).
2778
W. GRAF, R. SPENCER, H. BAKER, AND R. BAKER
that GABA is the major inhibitory neurotransmitter utilized
by premotor neurons involved in vertical eye movements.
In contrast, glycine is the inhibitory neurotransmitter of most
premotor neurons that are related to horizontal eye movements. On the basis of rationale related to the embryological
origin of extraocular motoneurons and muscles, the differences in neurotransmitters utilized in the vertical and horizontal eye movement systems have been suggested to reflect
the evolutionary history correlated to each extraocular subdivision (Baker 1991).
Ultrastructure
/ 9k11$$my40 J711-6
This work was supported by National Institutes of Health Grants DC00239, EY-04613, EY-02007, and NS-13742.
Address for reprint requests: R. Baker, Dept. of Physiology and Neuroscience, New York University Medical Center, 550 First Ave., New York,
NY 10016.
Received 5 September 1996; accepted in final form 21 January 1997.
REFERENCES
ALLUM, J.H.J., GREEFF, N. G., AND TOKUNGA, A. Projections to the rostral
and caudal abducens nuclei in the goldfish. In: Progress in Oculomotor
Research, edited by A. F. Fuchs and W. Becker. New York: Elsevier,
1981, p. 253–262.
BAKER, R. A contemporary view of the phylogenetic history of eye muscles
and motoneurons. In: Vestibular and Brainstem Control of Eye, Head,
and Body Movements, edited by H. Shimazu and Y. Shinoda. Tokyo:
Karger: Japan Scientific Society Press, 1991, p. 3–19.
BAKER, R., DELGADO-GARCIA, J. M., GRAF, W., AND SPENCER, R. F. Vestibular ocular organization in vertebrates is a largely conserved phylogenetic
plan. Soc. Neurosci. Abstr. 13: 172, 1987a.
BAKER, R., GRAF, W., AND BAKER, H. Inhibition and excitation in the
oculomotor system of fishes. Soc. Neurosci. Abstr. 12: 460, 1986.
BAKER, R., GRAF, W., AND SPENCER, R. F. Morphological and physiological
correlates of vestibular excitatory and inhibitory pathways to extraocular
motor nuclei in the goldfish. Neuroscience 22: S384, 1987b.
BAKER, R. AND HIGHSTEIN, S. M. Vestibular projections to medial rectus
subdivision of oculomotor nucleus. J. Neurophysiol. 41: 1629–1646,
1978.
BAKER, R., MARSH, E., AND SPENCER, R. F. Transneuronal transport of
biocytin through abducens motoneurons outlines the internuclear and
vestibular pathways in teleost fish. Soc. Neurosci. Abstr. 20: 1400, 1994.
BAKER, R., PRECHT, W., AND BERTHOZ, A. Synaptic connections to trochlear
motoneurons as determined by individual vestibular nerve branch stimulation in the cat. Brain Res. 64: 402–406, 1973.
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
The electron microscopic visualization of goldfish oculomotor nucleus neurons revealed axosomatic and axodendritic
synaptic endings containing spheroidal synaptic vesicles that
established chemical, presumed excitatory, synaptic contacts
characterized by asymmetric pre- and/or postsynaptic membrane specializations (Figs. 6 and 7). In addition, many
axosomatic synapses also showed gap junction contacts consistent with electrotonic coupling. Other axosomatic synaptic
endings contained pleiomorphic synaptic vesicles establishing chemical synaptic contacts on motoneurons. These presumed inhibitory endings were never observed to establish
gap junctions.
Electron microscopy of HRP-labeled abducens motoneurons in the goldfish showed numerous axo–somatic/dendritic synaptic endings containing spheroidal synaptic vesicles establishing both chemical (asymmetric pre- and/or
postsynaptic specialization) and extensive electrotonic (opposed membranes exhibiting gap junctions) synaptic contacts (Figs. 8 and 9). Other axosomatic synaptic endings
containing pleiomorphic vesicles established only chemical
synaptic specializations (Spencer and Baker 1983; Spencer
et al. 1989).
Comparison of the flatfish oculomotor and trochlear nuclear complex with that in the goldfish showed a difference
in the relative placement of chemical versus electrotonic
contacts of excitatory second-order vestibular neurons
(Spencer et al. 1986). In the goldfish, chemical contacts
were located at some distance from gap junctions, whereas
in the flatfish, gap and chemical junctions were adjacent to
each other (compare Fig. 7, E and D, unpublished observations). By contrast to the goldfish abducens nucleus, the
spatial relationship of chemical and electrotonic contacts was
closer and like the arrangement in flatfish. The functional
significance of this difference in chemical versus electrical
junctional placement is as yet undetermined. Furthermore,
the larger question concerns the significance of electrotonic
coupling per se in the vestibulooculomotor system of fish.
A prior interpretation argued that electrotonic transmission
played an important role in synchronizing motoneuronal discharge during either fast phases or saccades (Korn and Bennett 1971) and that the site of coupling was more restricted
to the soma rather than the dendrites (Korn and Bennett
1972). However, the ultrastructural data suggest that electrotonic coupling is restricted to excitatory boutons largely located on the proximal dendritic tree and they originate from
second-order vestibular neurons that produce the slow, not
fast, phase. Furthermore, electronic coupling is likely not
necessary for appropriate vestibuloocular reflex function be-
cause, mammals, for instance, do not have gap junctions
anywhere in the oculomotor system. The electrophysiological role of gap junctions in oculomotor function remains to
be determined.
The collective experimental evidence in goldfish and
flatfish (unpublished observations) now provide electrophysiological, immunohistochemical, and ultrastructural evidence for second-order vestibular inhibitory pathways in two
evolutionarily separate teleostean designs. These results
point to a common organizational pattern of vestibulooculomotor connectivity across vertebrate species. First, excitatory synapses appear to be located predominantly on dendrites and inhibitory synapses on somata and proximal dendrites. Second, although detailed comparisons reveal
quantitative differences in density of presumed inhibitory
and excitatory contacts (including gap junctions between
extraocular motoneurons), it is the overall similarity of the
vertebrate vestibulooculomotor blueprint that is most striking. Third, on a phylogenetic scale, these observations on
reciprocal inhibition and excitation are noteworthy, because
each of the many species studied are quite divergent from
lampreys which presumably exhibit the most primitive vestibuloocular plan (Pombal et al. 1994). Fourth, we conclude
that the conceptual view of reciprocal excitatory-inhibitory
control of motoneurons may be consistent from inception
throughout vertebrate phylogeny. If so, these relatively few
and easily identifiable neurons in fish offer significant potential for pursuing long-standing questions concerning the visual and vestibular regulation of oculomotor function and
the sites underlying adaptive plasticity.
VESTIBULAR INPUT TO EXTRAOCULAR MOTOR NUCLEI
/ 9k11$$my40 J711-6
MCCORMICK, C. M. AND BRAFORD, M. R. Organization of inner ear endorgan projections in the goldfish, Carassius auratus. Brain Behav. Evol.
43: 189–205, 1994.
MCCREA, R. A., STRASSMAN, A., AND HIGHSTEIN, S. M. Anatomical and
physiological characteristics of vestibular neurons mediating the vertical
vestibulo-ocular reflex of the squirrel monkey. J. Comp. Neurol. 264:
571–594, 1987.
PASTOR, A. M., DE LA CRUZ, R. R., AND BAKER, R. Characterization and
adaptive modification of the goldfish vestibuloocular reflex by sinusoidal
and velocity step vestibular stimulation. J. Neurophysiol. 68: 2003–2015,
1992.
PASTOR, A. M., DE LA CRUZ, R. R., AND BAKER, R. Eye position and
velocity integrators reside in separate brainstem nuclei in the goldfish
hindbrain. Proc. Natl. Acad. Sci. USA 91: 807–811, 1994.
PASTOR, A. M., TORRES, B., DELGADO-GARCIA, J. M., AND BAKER, R. Discharge characteristics of medial rectus and abducens motoneurons in the
goldfish. J. Neurophysiol. 66: 2125–2140, 1991.
POMBAL, M. A., RODICIO, M. C., AND ANADÓN, R. The anatomy of the
vestibulo-ocular system in lampreys. In: Information Processing Underlying Gaze Control, edited by J. M. Delgado-Garcı́a, E. Godaux, and
P.-P. Vidal. Tokyo: Elsevier, 1994, p. 1–11.
PRECHT, W. AND BAKER, R. Synaptic organization of the vestibulo-trochlear
pathway. Exp. Brain Res. 14: 158–184, 1972.
PUZDROWSKI, R. L. AND LEONARD, R. B. Vestibulo-oculomotor connections
in an elasmobranch fish, the Atlantic stingray, Dasyatis sabina. J. Comp.
Neurol. 339: 587–597, 1994.
SCHAIRER, J. O. AND BENNETT, M.V.L. Changes in gain of the vestibuloocular reflex induced by combined visual and vestibular stimulation in
goldfish. Brain Res. 373: 164–176, 1986.
SPENCER, R., WANG, S. F., AND BAKER, R. The pathways and functions of
GABA in the oculomotor system. In: GABA in the Retina and Central
Visual System, edited by R. R. Mize, R. Marc, and A. M. Sillito. Amsterdam: Elsevier, 1992, p. 307–334.
SPENCER, R. F. AND BAKER, R. Immunocytochemical localization of GABA
and parvalbumin in the extraocular motor nuclei of the cat and rhesus
monkey. Invest. Ophthalmol. Visual Sci. 31: 287, 1990.
SPENCER, R. F. AND BAKER, R. Morphology and synaptic connections of
physiologically-identified second-order vestibular axonal arborizations
related to cat oculomotor and trochlear motoneurons. Soc. Neurosci.
Abstr. 9: 1088, 1983.
SPENCER, R. F., WEISER, M., GRAF, W., AND BAKER, R. Vestibulo-oculomotor connections in goldfish and winter flounder. Biol. Bull. 171: 498,
1986.
SPENCER, R. F., WENTHOLD, R. J., AND BAKER, R. Evidence for glycine
as the putative inhibitory neurotransmitter of vestibular, reticular and
prepositus hypoglossi neurons that project to the cat abducens nucleus.
J. Neurosci. 9: 2718–2736, 1989.
STERLING, P. Anatomy and physiology of goldfish oculomotor system. I.
Structure of abducens nucleus. J. Neurophysiol. 40: 557–572, 1977.
STRAK A, H. AND DIERINGER, N. Electrophysiological and pharmacological
characterization of vestibular inputs to identified frog abducens motoneurons and internuclear neurons in vitro. Eur. J. Neurosci. 5: 251–260,
1993.
SUWA, H. AND BAKER, R. Anatomy and physiology of vestibular and internuclear neurons responsible for nasally-directed eye movement in teleosts. Soc. Neurosci. Abstr. 22: 1833, 1996.
TORRES, B., PASTOR, A. M., CABRERA, B., SALAS, C., AND DELGADO-GARCIA, J. M. Afferents to the oculomotor nucleus in the goldfish (Carassius
auratus) as revealed by retrograde labeling with horseradish peroxidase.
J. Comp. Neurol. 324: 449–461, 1992.
08-08-97 12:57:37
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 18, 2017
EASTER, S. S. Pursuit eye movements in goldfish (Carassius auratus). Vision Res. 12: 673–688, 1972.
EVINGER, C. Extraocular motor nuclei: location, morphology and afferents.
In: Neuroanatomy of the Oculomotor System, edited by J. A. ButtnerEnnever. Amsterdam: Elsevier/North Holland, 1988, p. 81–117.
FABER, D. S. AND KORN, H. Inputs from the posterior lateral line nerves
upon the goldfish Mauthner cell. II. Evidence that the inhibitory components are mediated by interneurons of the recurrent collateral network.
Brain Res. 96: 349–356, 1975.
FABER, D. S. AND KORN, H. Voltage-dependence of glycine-activated chloride channels: A potentiometer for inhibition. J. Neurosci. 7: 807–811,
1987.
FABER, D. S. AND KORN, H. Unitary conductance changes at teleost
Mauthner cell glycinergic synapses: a voltage-clamp and pharmacologic
analysis. J. Neurophysiol. 60: 1982–1999, 1988.
GESTRIN, P. AND STERLING, P. Anatomy and physiology of goldfish oculomotor system. II. Firing patterns of neurons in abducens nucleus and
surrounding medulla and their relation to eye movements. J. Neurophysiol. 40: 573–588, 1977.
GRAF, W. AND BAKER, R. Adaptive changes of the vestibulo-ocular reflex
in flatfish are achieved by reorganization of central nervous pathway.
Science Wash. DC 221: 777–779, 1983.
GRAF, W. AND EZURE, K. Morphology of vertical canal related second order
vestibular neurons in the cat. Exp. Brain Res. 52: 35–48, 1986.
GRAF, W., MCCREA, R. A., AND BAKER, R. Morphology of posterior canal
related secondary vestibular neurons in rabbit and cat. Exp. Brain Res.
52: 125–138, 1983.
GRAF, W. AND MC GURK, J. Peripheral and central oculomotor organization
in the goldfish, Carassius auratus. J. Comp. Neurol. 227: 391–401, 1985.
HIGHSTEIN, S. M. Synaptic linkage in the vestibulo-ocular and trochlear
reflex pathways in the rabbit. Exp. Brain Res. 17: 301–314, 1973.
HIGHSTEIN, S. M. AND MCCREA, R. A. The anatomy of the vestibular nuclei.
In: Neuroanatomy of the Oculomotor System, edited by J. A. ButtnerEnnever. Amsterdam: Elsevier/North Holland, 1988, p. 177–202.
ITO, M., NISIMARU, N., AND YAMAMOTO, M. The neural pathways mediating
reflex contraction of extraocular muscles during semicircular canal stimulation in rabbits. Brain Res. 55: 183–188, 1973a.
ITO, M., NISIMARU, N., AND YAMAMOTO, M. The neural pathways relaying
reflex inhibition from semicircular canals to extraocular muscles of rabbits. Brain Res. 55: 189–193, 1973b.
KORN, H. AND BENNETT, M.V.L. Dendritic and somatic impulse initiation
in fish oculomotor neurons during vestibular nystagmus. Brain Res. 27:
169–175, 1971.
KORN, H. AND BENNETT, M.V.L. Electrotonic coupling between teleost
oculomotor neurons: restriction to somatic regions and relation to function
of somatic and dendritic sites of impulse initiation. Brain Res. 38: 433–
439, 1972.
KORN, H. AND BENNETT, M.V.L. Vestibular nystagmus and teleost oculomotor neurons: functions of electrotonic coupling and dendritic impulse
initiation. J. Neurophysiol. 38: 430–451, 1975.
KORN, H., ODA, Y., AND FABER, D. S. Long-term potentiation of inhibitory
circuits and synapses in the central nervous system. Proc. Natl. Acad.
Sci. USA 89: 440–443, 1992.
LLINAS, R. AND BAKER, R. A chloride dependent inhibitory postsynaptic
potential in cat trochlear motoneurons. J. Neurophysiol. 35: 364–385,
1972.
MAGHERINI, P. C., PRECHT, W., AND SCHWINDT, P. C. Functional organization of the vestibular input to ocular motoneurons of the frog. Pfluegers
Arch. 349: 149–158, 1974.
2779