Department of Neurobiology, Duke University Medical Center, Durham, North Carolina 27710
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Pettit, Diana L., Matthew C. Helms, Psyche Lee, George J. Augustine, and William C. Hall. Local excitatory circuits in the intermediate gray layer of the superior colliculus. We have used photostimulation and whole cell patch-clamp recording techniques to examine local synaptic interactions in slices from the superior colliculus of the tree shrew. Uncaging glutamate 10-75 µm from the somata of neurons in the intermediate gray layer elicited a long-lasting inward current, due to direct activation of glutamate receptors on these neurons, and brief inward currents caused by activation of presynaptic neurons. The synaptic responses occurred as individual currents or as clusters that lasted up to several hundred milliseconds. Excitatory synaptic responses, which reversed at membrane potentials near 0 mV, could be evoked by uncaging glutamate anywhere within 75 µm of an intermediate layer neuron. Our results indicate the presence of extensive local excitatory circuits in the intermediate layer of the superior colliculus and support the hypothesis that such intrinsic circuitry contributes to the development of presaccadic command bursts.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The intermediate gray layer of the superior
colliculus commands saccadic eye movements by generating high-frequency
bursts of action potentials in spatially restricted populations of
cells. The spatial and temporal properties of these bursts determine the time of onset, amplitude, and direction of an impending saccade (Munoz and Wurtz 1995a,b
; Sparks 1978
).
This causal relationship between the pattern of activity in a
population of nerve cells and a defined behavior has aroused interest
in understanding the neural processing that underlies the
spatiotemporal properties of these bursts. Most models of presaccadic
bursting include a wide ranging network of inhibitory connections
within the intermediate layer that help shape the distribution of
electrical activity. However, these models differ in the predicted role
of local excitatory connections. Some models postulate that local
recurrent excitatory circuitry produces an increase in excitation among
neighboring cells. This increase in local excitation, in concert with
the inhibitory connections, determines which cells produce the
high-frequency command bursts (Arai et al. 1994
;
Van Opstal and Van Gisbergen 1989
). In contrast, a
second type of model does not require intrinsic excitatory connections
and suggests that extrinsic synaptic inputs, such as those arising from
the frontal eye fields, impose the pattern of excitation
(Schlag-Rey et al. 1992
). One way to evaluate these
models is to examine intrinsic synaptic circuits; the presence of
robust local excitation would provide a physiological substrate for
intrinsic excitatory circuitry postulated to contribute to the
production of presaccadic bursts.
Here we examine short-range synaptic interactions among intermediate
layer neurons in collicular slices. With traditional methods, it has
been difficult to determine the extent to which the response properties
of these cells are determined by extrinsic inputs or by intrinsic
circuitry. In particular, local synaptic interactions cannot be
selectively elicited by electrical stimulation, which will also
activate synaptic inputs arising from fibers of passage. We have
avoided this problem by using localized laser photolysis of caged
glutamate (Pettit et al. 1997) to exclusively activate
local synaptic circuits by "photostimulation" (Callaway and
Katz 1993
; Katz and Dalva 1994
). This approach
reveals extensive intrinsic excitatory connections among neurons in the
intermediate layer, consistent with the hypothesis that such
interactions contribute to the development of presaccadic bursts.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Collicular slices
Coronal or parasagittal slices (300 µm thick) were prepared
from the superior colliculus of 13- to 20-day-old tree shrews. The
well-defined laminar organization of the colliculus in this species
made it possible to visually select and record from single neurons in
identified layers. Slices were superfused at room temperature with
oxygenated physiological saline (in mM: 119 NaCl, 2.5 KCl, 1.3 MgCl2, 2.5 CaCl2, 1 NaH2PO4, 26.2 NaHCO3, and 11 glucose) containing 100-150 µM carboxy-nitrobenzyl-caged glutamate
(Molecular Probes, Eugene, OR). In some experiments, 1 µM
tetrodotoxin was added to the saline to block synaptic transmission.
Whole cell patch-clamp recordings were made from 27 intermediate and
deep layer neurons, as described in Lee et al. (1997).
Recordings were accepted only if the holding current was <100 pA when
the membrane potential was voltage clamped at
60 mV. The patch
pipette solution contained (in mM) 100 gluconic acid, 2-10 EGTA, 5 MgCl2, 2 ATP, 0.3 GTP, and 40 HEPES; pH to 7.2 with CsOH.
During the experiment, a fluorescent dye, Oregon Green 488 BAPTA-1 (200 µM; Molecular Probes, Eugene, OR), was dialyzed from the patch
pipette into the neuron to visualize the soma and dendrites of
individual cells with a confocal microscope (Noran Odyssey) during the
experiment. To characterize the cell morphology in more detail,
biocytin (10%) was included in the pipette solution for 16 of these
neurons. The slices were fixed and a diamino-benzidine reaction was
performed after the experiment to reveal the biocytin.
Photostimulation
The instrument used for photostimulation is described in
Wang and Augustine (1995). In brief, ultraviolet light
from an argon ion laser was delivered through a microscope objective
and focused on the slice. With this arrangement, glutamate was
photolyzed consistently over an area 10 µm diam at the focal plane
(Pettit et al. 1997
; Pettit and Augustine
1998
). We assume a constant light density and constant
concentration of "caged" glutamate for each photostimulation event.
An electronic shutter was used to vary the duration of the light pulse
(5-25 ms) and thus the total amount of photostimulation. The laser
beam was aligned in the center of the objective and remained stationary
in the field of view. The distance between the uncaging spot and the
fluorescently labeled cell was varied by translating the microscope
stage, and this distance was determined directly from the confocal image.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Electrical recordings from the cell bodies of intermediate and
deep gray layer neurons were used to detect responses evoked by
photostimulation (n = 27). For all of the cells,
uncaging glutamate at distances 10-75 µm from the soma elicited
inward currents with two distinctive components (Fig.
1A). The first component
consisted of a long-lasting inward current with a duration of several
hundred milliseconds (Fig. 1A, top trace). Such responses
were insensitive to 1 µM tetrodotoxin, which blocks voltage-gated
sodium channels and, therefore, the action potentials that require
these channels (Fig. 1B). This component closely resembled
responses mediated by glutamate receptors (Pettit et al.
1997) and probably arose from direct activation of glutamate
receptors on the neurons from which recordings were made. Consistent
with this interpretation, and the range of action of uncaged glutamate
(Katz and Dalva 1994
; Pettit et al.
1997
), these responses were always observed at locations within
20 µm of the somata or dendrites of the collicular neurons, but were
often absent from more distant sites (Fig. 1A, bottom trace).
|
The second component consisted of brief currents with a half-maximal
duration of 4.7 ± 0.2 ms (mean ± SE, n = 42 responses). These brief currents could be evoked even at locations that
evoked no direct responses (Fig. 1A, bottom trace) and were
blocked by application of tetrodotoxin (1 µM; Fig. 1B).
These properties indicate that the brief currents arose from excitation
of neurons innervating the patch-clamped cells. The synaptic currents
occurred with a wide range of latencies and occurred as isolated events or in clusters with durations up to hundreds of milliseconds (Fig. 1A, bottom trace). The frequency of these events varied
within a range of 20-40 Hz. Typically the greatest number of responses occurred within the first 50 ms after the light flash, presumably reflecting the time course of the glutamate-induced depolarization of
the stimulated presynaptic cells. There was a mean of 8.8 ± 0.5 synaptic events/stimulus, although this may be an underestimate due to
superimposition of synaptic responses. This is a very high density of
evoked responses in comparison with other brain areas where
photostimulation has been attempted, such as visual cortex (Callaway and Katz 1993; Dalva and Katz
1994
; Sawatari and Callaway 1996
).
At a holding potential of 60 mV, the evoked synaptic currents had a
mean amplitude of 36 ± 1.9 pA (n = 56 responses).
We conclude that these currents were excitatory because they reversed their polarity when the holding potential was depolarized to positive potentials (Fig. 1C). Further evidence of their excitatory
nature is their decay time constant of 4.1 ± 0.3 ms, which is
consistent with the decay of glutamate-mediated synaptic currents
(Hestrin 1993
), whereas GABA-mediated currents typically
decay much more slowly (Jones and Westbrook 1995
). In
addition, any chloride-mediated inhibitory currents would likely escape
detection because of the small difference between the chloride reversal
potential (
66 mV) and the holding potential (
60 mV).
The spatial arrangement of functional synaptic connections was examined
for seven neurons whose anatomy (confirmed with biocytin histology) was
consistent with their identity as premotor neurons (Hall and Lee
1997). That is, they were located in the intermediate layer and
had large (>20 µm diam) somata and multipolar dendritic arbors. In
other studies, neurons with these features have been shown to generate
command signals for saccades (Moschovakis et al. 1988
).
Photostimulating at sites along several axes within the field of the
microscope objective revealed an isotropic excitatory surround of
synaptic transmission (Fig. 2). This
excitatory surround was characteristic of all of the recordings from
intermediate layer neurons (Fig. 3). When
the light beam was moved laterally within the intermediate layer,
synaptic events were evoked at every location. In fact, over a range of
50 µm, the number of events evoked by a light flash was remarkably
constant (Fig. 3A), suggesting a constant density of local
excitatory inputs converging on the premotor cells. Likewise, when the
light beam was placed at sites along a superficial axis perpendicular
to the surface of the colliculus, responses could be detected from each
location within 75 µm of the soma (Fig. 3B). Stimulating
along axes with other orientations also evoked very similar synaptic
responses. Combining responses from all axes reveals that there was no
significant spatial gradient in the distribution of synaptic inputs
within 75 µm of a given intermediate layer neuron (Fig.
3C). Similarly, the amplitude of the synaptic responses did
not vary systematically with distance, either between or within axes
(Fig. 3D). Responses evoked near cell bodies were somewhat
smaller than those observed for regions distant from the cell body, but
this may be a consequence of the large direct responses that were
generated when the light spot was close to the target neuron. Our
results indicate that a high and relatively constant density of strong
synaptic excitation surrounds each intermediate layer neuron.
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Previous studies that used electrical stimulation suggested a rich
plexus of intrinsic synaptic circuitry in the intermediate layer. For
example, electrical stimulation of the intermediate layer produced
transynaptic excitation of premotor neurons (McIlwain 1982) and lateral inhibitory interactions (Meredith and
Ramoa 1998
; Munoz and Istvan 1998
). However, the
likely stimulation of fibers of passage made it difficult for such
experiments to distinguish between extrinsic and intrinsic sources of
synaptic input to the cells of this layer. Because the uncaged
glutamate does not activate axons (Callaway and Katz
1993
) and because slicing the colliculus eliminated most
extrinsic sources of synaptic input, we were able to selectively
activate local inputs. We find that there are indeed extensive
excitatory synaptic interactions among the neurons of the intermediate
gray layer. These local excitatory circuits could provide the substrate
for positive feedback that sustains and intensifies the low-frequency
activity that precedes the command burst for a saccade (Glimcher
and Sparks 1992
; Munoz and Wurtz 1995a
). They
may also contribute to the prolonged bursts of excitatory postsynaptic
currents seen in intermediate layer neurons in response to electrical
stimulation of sensory inputs from the superficial layer (Lee et
al. 1997
). Premotor cells in the intermediate layer give rise
to local axonal arbors that may underlie these interactions
(Hall and Lee 1997
; Moschovakis et al.
1988
).
Our results support the hypothesis that intrinsic excitatory interactions contribute to the development of presaccadic command bursts in the superior colliculus. Future work will be needed to address the contributions of intrinsic inhibitory connections and extrinsic inputs. It is probable that all of these sources of synaptic input work in concert to shape the spatiotemporal profile of electrical activity in the intermediate layer, and the present results demonstrate that photostimulation will be useful for analyzing their relative contributions.
![]() |
ACKNOWLEDGMENTS |
---|
We thank Drs. N. Cant, D. Fitzpatrick, E. Keller, G. Ozen, and D. Sparks for helpful comments on this manuscript. D. L. Pettit and M. C. Helms contributed equally to this paper.
This work was supported by National Institutes of Health Grants EY-08233 and NS-34045.
![]() |
FOOTNOTES |
---|
Address for reprint requests: W. C. Hall, Dept. of Neurobiology, Box 3209, Duke University Medical Center, Durham, NC 27710.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 8 September 1998; accepted in final form 3 November 1998.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|