Neurons immunoreactive to choline acetyltransferase in the turtle retina

Vision Res. Vol. 32, No. 11, pp. 202%2030, 1992 Printed in Great Britain. All rights reserved

0042-6989/92 $5.00 + 0.00 Copyright 0 1992 Pergamon Press Ltd

Neurons Immunoreactive to Choline Acetyltransferase in the Turtle Retina GLORIA D. GUILOFF,* HELGA KOLB*

Received 9 July 1991; in revised form 7 April 1992

Light microscopic immunocytochemistry using anti-choline acetykransferase (ChAT) was performed to stain putative cholinergic amacrine cells in turtle retina. C&AT-immunoreactive somata lie in the inner nuclear (INL) and ganglion cell (GCL) layers. Three types of amacrine cells were found according to the location of their somata and their dendritic stratifkation pattern in the inner plexiform layer (IPL). Type I amacrines lie in the row of cells closest to the INL/IPL limits and they branch along the sl/s2 border of the IPL. Type II amacrines are displaced to the GCL and they ramify along the s3/s4 border of the IPL. Type III amacrines lie in the middle of the INL, 2-3 rows away from the IPL limits and their dendrites appear to be bi- or t&strati&d in sl and s3-4 of the IPL. The turtle ChAT-IR amacrines are thus similar to the types described in chicken retina. A regubu, non-random mosaic formed by stained type II amacrine cells was ohserved in the GCL. Their density in mid-central retina was 750 cells/mm*, tapering off to 393 cells/mm’ in peripheral retina. Our study indicates that a pair of cholinergic amacrine cell types in turtle retina is arranged in mirror-image symmetry contributing to sublamina “a” and subhunina “b” of the IPL, l&e in other vertebrate retinas.

Turtle retina ChAT-immunoreactive neurons Amacrine cells

INTRODUCTION

In order to eventually understand neural processing in the retina, it has become increasingly important to find out which neurotransmitters are involved in specific circuits and where their actions are occurring. The traditional neurotransmitter acetylcholine, for example, has been implicated as the driving excitatory input to “complex” retinal ganglion cells in general and to direction-selective ganglion cells in particular (Daw & Ariel, 1981; Ariel & Daw, 1982b; Famiglietti, 1987).

Choline@ amacrine cells have been found in several species and are expected to be present in the turtle retina as well, particularly as this retina is known to have a large complement of “complex” cells. Certainly, acetylcholine has been implicated in the generation of directional selectivity in turtle retinal ganglion cells (Ariel & Adolph, 1985). A brief report concerning choline acetyltransferase (ChAT) immunostaining in cone pedicles of the turtle retina (Criswell & Brandon, 1987) is the only attempt, as far as we know, to investigate cholinergic neurons in this species. We think it is important to determine which subpopulations of amacrine cells are choline@ and to identify their branching patterns, distribution and stratification in the inner plexiform layer (IPL). Thus, the purpose of this study was to observe and identify putative cholinergic amacrine cells

*Physiology Department, School of Medicine, University of Utah, 410 Chipeta Way, Room 156, Research Park, Salt Lake City, UT 84108, U.S.A.

using light microscopic immunocytochemistry in the turtle retina. In order to do this, we have used an antibody against ChAT, the rate limiting enzyme and last step in the biosynthetic pathway that forms acetylcholine. This enzyme, therefore, is a good indicator of putative cholinergic neurons (Eckenstein & Thoenen, 1982). A preliminary report of this work has been published elsewhere (Guiloff & Kolb, 1991).

METHODS

Small adult turtles (Psetcdemys scriptu elegans, under 6 in. carapace length) gave the best results for immunocytochemistry in our hands. Data from three animals, i.e. six retinas, were used for this study, although immunostaining was tried unsuccessfully on eight ad- ditional animals. Turtles were kept in a departmental animal facility with food pellets and water ad libitum on a 12 hr light-dark cycle. They were dark-adapted l-2 hr or overnight, then anesthetized with an i.m. injection of vetalar or ketamine (100 mg/kg wt) for about 30 min, decapitated and pithed. The eyeballs were quickly dissected out, cut in half and the posterior halves placed in 0.1 M phosphate buffer-saline solution (PBS), pH 7.4. The retinas were dissected under a microscope; as much vitreous as possible was carefully dissected out and the rest of the vitreous was drained with tissue paper wedges when necessary. The pigment epithelium was peeled or brushed off. Four radial razor-blade cuts were made in order to flatten the retina and then it was placed, with

2023

2024 GLORIA D. GUILOFF and HELGA KOLB

the ganglion cell layer (GCL) side upward, on #50 Whatman filter paper, making sure that the retina was properly stuck to the filter paper. Excess vitreous which might still be clinging to the retina was brushed off under the microscope at this stage. This filter paper with the retina attached was then placed in fresh cold fixative solution (4% formaldehyde in PBS) overnight, at 4°C. Frozen, 14 pm-thick, radial sections were cut from some retinas and mounted on subbed slides for subsequent immunostaining.

Following a PBS wash, 0.3% H,O,, in PBS was used for 15 min (sections) or 30 min (wholemounts), and then another PBS wash was done. An optional preblock step with 10% normal goat serum (Vector Lab. Inc., Calif.) in PBS + 0.5% Triton X-100 for 1.5 hr (~tio~) or 4 hr (wholemounts) followed by a brief PBS wash may be done. In our hands, this optional step could be omitted without apparently altering the results. For the incubation in the primary antibody, ChAT, polyclonal anti- human placental ChAT raised in rabbit, obtained as a gift from Drs Mariani and Hersch or ~mmer~ally from Chemicon International Inc. (Calif.), was used. Three dilutions were used, namely 1: 500, 1: 1000 and 1: 2000, made up with PBS + 0.5% Triton X-100. The incubation was performed overnight for sections and for 4 days for wholemounts, at 4°C. All incubations were done in a humid-tight box or tightly-capped small flasks, with agitation. Some sections were incubated omitting the primary antibody as controls. After a PBS rinse, the tissue was incubated in a biotinylated secondary antibody, affinity purified goat anti-rabbit IgG (H + L, Vector Lab. Inc., Calif.), diluted 1: 50 with PBS + 0.5% Triton X-100, for 1 hr (sections), or for 2 days (wholemounts), at room temperature, then rinsed with PBS. The incubation in the tertiary solution, ABC (Vectas- tain, Vector Lab. Inc., Calif.), for 1 hr (sections) or for 2 days (wholemounts) followed, at room temperature. After another PBS wash, the retinas were then preincu- bated with diaminobenzidine (DAB) for 10 min (sections) or for 20min (wholemounts). The DAB reaction (10ml fresh DAB solution + 6.6~1 H,O,) was done under visual control and usually took from 3 to 10 min. This reaction was stopped by immersing the tissue in ice-cold PBS, followed by a PBS wash. Some prep arations were subsequently osmicated with 0.2% 0~0, in PBS (under visual control) to enhance the immunostain. Finally, the sections were air-dried before dehydration and embedding in permount and the wholemounts were dehydrated in a graded series of ethanol, cleared in xylene and mounted in permount.

ChAT-immunoreactive (ChAT-IR) cells were exam- ined by light microscopy, drawn with the aid of a camera lucida and photographed. One well-stained retina was analyzed for nearest-neighbor relations (Wissle & Rieman, 1978). Stained cell bodies from two areas: mid-central retina, about 2 mm ventral to the linear visual streak, and peripheral retina, 5 mm dorsal to the linear visual streak, were mapped onto a drawing of the retina and the distances (in pm) between cells were measured by reference to a scale in the eyepiece graticule.

Shrinkage in the retinal wholemounts was about 8% with the technique we used; the data were not corrected for shrinkage.

Finally, some comments seem appropriate about the difficulties encountered when trying out the few available ChAT antibodies. We report here successful staining in the turtle retina with one batch of the polyclonal anti- human placental antibody raised in rabbit provided by Dr Mariani, with another batch provided by Dr Hersch, and with one batch from Chemicon International Inc. (Calif.). However, we also tried, unsuccessfully, antibodies from Boehringer-Mannheim Biochemicals (Indi- anapolis, Ind.) and a second batch provided by Dr Hersch. It appears to be a hit-or-miss situation to be able to stain the molecule in a species like turtle, because different species apparently seem to have diverse epi- topes on their ChAT molecules. Thus, several authors have reported their inability to obtain a reaction with various commercial antibodies against ChAT. Clearly, if more sources (different species) for obtaining the ChAT molecule to make the antisera were used in the future, more investigations in a larger variety of species could be carried out successfully.

RESULTS

We describe here the distribution, density, nearest- neighbor relations (WHssle & Riemann, 1978) and stratification in the IPL, of ChAT-immunoreactive (ChAT-IR) neurons in the turtle retina, as observed in wholemounted retinas and in frozen, radial sections. Control sections, i.e. incubated without the primary antibody, showed no ChAT immunostaining.

Wholemounts

We have observed ChAT-immunostained cell bodies and their dendrites in wholemount preparations of turtle retina forming a regmar mosaic [Fig. l(a)]. ChAT-IR cell bodies occurred in the inner nuclear layer (INL) and in the GCL. In the IPL, the stained amacrine cell dendrites could be seen forming an obvious plexus. This plexus was evident at two levels in the IPL: in strata sl and sl/s2 of sublamina ‘ra” [arrowheads, Fig. l(c)] and in stratum s3/s4 of sublamina “b” fFig. l(b, d)]. In this plexus, dendrites appeared to run together in thick bundles or fascicles leaving relatively stain-free spaces [Fig. l(b)].

Figure 2 shows camera-lucida drawings of some of the individual amacrine cells that we observed in wholemounted retinas; these are not complete ceIb, however, because we could not distinguish confidently where the dendrites of one cell ended and those of a neighboring cell started. The ChAT-IR amacrines had small (5-1Opm dia), round cell bodies, often giving rise to a single descending or ascending apical dendrite that then branched profusely in the strata of the IPL where it contributed to the plexus, i.e. stratum s3 for GCL-occurring cells, stratum sl/s2 for INL-occurring cells. The branching pattern was tufted and profuse, with greater numbers of dendrites arising progressively further from the cell body. ChAT-IR amacrines could be considered

ChAT-IMM~OR~A~IVE NEURONS IN TURTLE RETINA

FIGURE 1. (a) Low-power micrograph showing the regularity of the mosaic made by ChAT-IR cell bodies in mid-central, flat-mounted, turtle retina, seen with interference-contrast light microscopy. The focus is in the ganglion cell layer, arrows point to stained somata. Magnification bar = 50 pm. (b) Micrograph illustrating the plexus along the s3/s4 border of the IPL seen in flat-mount view. Arrowheads point to the bundles of dendrites forming this plexus. A type. II ChAT-IR amacrine cell soma (arrow) lying in the ganglion cell layer of the turtle retina is also in focus. Magnification bar = 20 pm. (c) Micrograph of a vertical (radial) 14 pm-thick frozen section through turtle retina showing a type I C&AT-IR amacrine cell (arrow) b~nc~ng in the sl/s2 border of the IPL (arrowhead). Madison bar = 20 pm. (d) Micrograph of a vertical (radial) 14 pm-thick frozen section through turtle retina showing a type II C&AT-IR amacrine cell (arrow) branching in the s3/s4 border of the IPL

(arrowhead). Magnification same as in (c).


5 P q

FIGURE 2. Camera lucida drawings of flat-mounted views of ChAT- IR neurons in the turtle retina. The cells illustrated are probably not complete because of difficulty in distinguishing where the dendrites of one cell ended and those of a neighboring cell started. Magnification

bar = 25 pm.

to have the rudiments of the “starburst” amacrine cell morphology so clear in other species.

Cone pedicies in the outer retina also appeared to be ChAT-IR (not shown), as has been reported previously (Criswell & Brandon, 1987).

Vertical sections

On viewing frozen, radial 14 pm-thick sections of the turtle retina we could clearly see immunostained cell bodies lying in the inner nuclear layer and in the ganglion cell layer. We were able to distinguish three types of ChAT-IR amacrine cells in our frozen sections, according to the position of the somata in the INL or GCL and their stratification patterns in the IPL (Figs l(c,d) and 3(a)]. ChAT staining in the IPL occurred in four bands: two relatively narrow bands were strongly stained in strata sl/s2 and s3/s4 [Fig. l(c,d)], and two weaker bands were sometimes discernible in strata sl and s3-s4 [arrowheads, Fig. 3(a)]. In some preparations the fasciculation of dendrites forming a plexus described from the wholemount view could also be observed.

Figure l(c) shows that type I amacrine cell bodies lie in the row of cells closest to the INL/IPL limits and their dendrites branch in the IPL along the stratum sl/s2 border, while Fig. l(d) shows that type II amacrines are displaced to the GCL and they ramify in the IPL along the stratum s3/s4 border. Type III amacrine cell somata could be seen to lie in the middle of the INL, 2-3 rows of cell bodies away from the INL/IPL limits, and their weaker-staining dendrites are probably bi- or tri- stratified in the IPL contributing to faint staining in strata sl and s3-s4 [Fig. 3(a)].

Amacrine cells or ganglion cells?

The majority of the ChAT-IR neurons lying in the GCL appeared to be amacrine cells, as judged by cell body size. Their typical dimension was S-10 pm in diameter, which is the common size range for amacrine cell somata in the turtle retina (Kolb, 1982). A few larger ChAT-IR cell bodies in the GCL may correspond to

ganglion cells. For example, the cell illustrated in Fig. 3(b) (large arrow) has a cell body diameter of over 19 pm. For comparison, the ChAT-IR type I, II and III amacrine cells are also pointed out [small arrows, Fig. 3(b)]. A few ChAT-IR fibers were occasionally seen in the nerve fiber layer [arrowheads, Fig. 3(b)]. These large putative ChAT-IR ganglion cell bodies were not entered in the quantitative analysis below.

Nearest -neighbor analysis

In order to determine whether the stained ChAT-IR cells were randomly distributed or not, we did a nearest- neighbor analysis (Wtissle & Rieman, 1978) with data from two regions containing ChAT-IR cell bodies in the GCL of a single wholemounted turtle retina. Histograms depicting the results are shown in Fig. 4. Distances (in pm) to nearest neighbors of stained cells (bars, Fig. 4) are plotted vs the frequencies at which they occurred in mid-central (2 mm ventral to the linear visual streak) and peripheral (5 mm dorsal to the linear visual streak) regions of the turtle retina. The distribution of the cells was then compared to a normal distribution curve (Gaussian, closed circles in Fig. 4) and to a random distribution curve (Poisson, open circles in Fig. 4). The ChAT-IR cell bodies in both samples, mid-central retina [Fig. 4(A)] and peripheral retina [Fig. 4(B)] were distributed closer to the Gaussian curve than to the Poisson curve, and are thus considered to be distributed non- randomly. The density of ChAT-IR cell bodies in central retina, close to the linear visual streak, was higher at 750 cells/mm’ than in peripheral retina where it was 393 cells/mm2.

DISCUSSION

Our study shows that three types of ChAT-IR amacrine cells are present in the turtle retina with cell bodies occurring both in the INL and displaced to the GCL. According to their location and stratification pattern in the IPL, ChAT-IR amacrine cells in the turtle retina appear to correspond to the three types described previously in the chicken retina (Millar, Ishimoto, Chubb, Epstein, Johnson & Morgan, 1987). We found that type I ChAT-IR amacrine cell bodies were located in the row of cells closest to the border of the INL/IPL layers and their dendrites branched in the IPL in an intense band along the strata sl/s2 border. These cells are probably equivalent to the “starburst a” cholinergic amacrine cell type of the rabbit retina (Famiglietti, 1983). Type II ChAT-IR amacrine cell bodies in the turtle retina were displaced to the GCL and they ramified in the IPL in a strongly stained band along the strata s3/s4 border. These cells are probably the turtle counterpart of the “starburst b” cholinergic amacrine cell type of the rabbit retina (Famiglietti, 1983). Cholin- ergic amacrine cell types I and II in turtle retina are thus arranged in a reciprocal array in the IPL, as mirror-image, symmetrical pairs for sublamina “a” and sublamina “b”, just like has been seen in other vertebrate retinas. In the GCL, Type II ChAT-IR amacrines formed a

FIG1 am cell 1

Mw sectic

ChAT-IMMUNOREACMVE NEURONS IN TURTLE RETINA 2027

JRE 3. (a) Micrograph of a vertical (radial) 14pm-thick frozen section through mid-centml turtle retina showing the uance of ChAT-immunomactivity. Amamine cell bodies in the inner nuclear layer (types I and Iii) and in the ganglion ayer @CL, type II) are stained (arrows), as well as four bands in the inner pkziform layer QPL, arrowheads). lification bar = 20 pm. @) Interf enmcecontrast light microscopy micrograph of a vertical (radial) 14pm-thick frozen III through turtle retina. Large arrow points to a large ChAT-IR cell (GC) in the ganglion cell layer. Small arrows indicate smaller ChAT-IR somata in the inner nuclear layer and ganglion cell layer for comparison. Arrowheads in the nerve

fiber layer point to putative ChAT-IR fibers. Nagni6cation same as in (a).


ChAT, did-central retina {n&4)

35 F r 3o

e 25 9 u 20 e n l5

c 10 Y

5

IData

I

* Normal curve

-0- Poisson curve

0 10 20 30 40 50 60 70 60 90 100 110 120

Distance to nearest neighbor (pm)

ChAT, peripheral retina (n=28)

Data

-a- Normal curve

*- Poigsan curve

a 10 PO 30 40 50 60 70 80 60 100 110 120

I3iiMnce to near@ neighbor (w)

FIGURE 4. Histqmms ~~~t-~~~r analysis of two regions of GMT-IR somata in the CCL of wh~e-mounts turtle retina. The distant PO nag@ nei6hbors (in gm, sevwal dkances per cell) are pietWl vs tie frequtncieg, @umber of cells) at which they occur, and compared to a normal @&ibution (Gaussian curve) and to a ra&om distribution (Poisson

curve). The cells am disttibuted non-randomly in both samples: mid-central retina (A) and peripheral retina (B).

regular mosaic which was found by nearest-neighbor analysis to correspond to a non-rartdom distribution.

Type III ChAT-IR ama@it&e cell somata lie in the middle of the INL, 2 or 3 rows CfT c&s away from the INL/IPL border and their do&rites, al&o* very weakly stained, appeared to zle hi+ .W tri-stratiBed in the IPL, in strata sl and s3-&. T&ES&~+@ corrwnd well with the type III cl~l&r& am&&s ~~~1~ described for chicken retina (Millar et al., 1987). Interest- ingly, type III ChAT-IR amacrine cells have not been observed in species other than chicken, and now turtle, so far.

The ChAT-IR bands which we observed in the IPL may originate from processes of amacrine and ganglion cell types which are mono-, bi- or TV-stratified in stata sl, sl/s2, s3, s3/s4 and/or s4 of the IPL. Many mospho- logical neuronal types have been described in the turtle

retina, with di@&rent stratification patterns in the IPL (Kolb, 1982; Kolb, Perltnan L Normann, 1988). Of thee the smaIl&Id amacrine ceil rno~holo~~ type A8 branches in sl of the IPL, thus making it a good candidate for our type I ChAT-IR amacrine cell. In the same way, the ~maEi&ekl amacrine cell type A9, ramify- ing in s3/s4 of&e WL, wo&d q&ify to be the type II ChAT-IR ama&ne cell described in this study. For the type III ChAT-IR amacrine cell, the hi-stratified, medium-field A14, which branches in sl and s3 of the IPL, and/or the bi-stratified, medic-field A15, which ramifies in sl/s2 and s4fs5 of the IPL, are likely candi- dates. Further studies are needed to clarify which of these morphological amacrine cell types are the cholinergic ones, and also to ascertain their p~~~pa~on in driving the direction-selective ganglion cells in the turtle retina (see below). The ChAT-IR banding pattern in the

IPL also indicates where we might expect ganglion cell there is quantitative ultrastructural evidence from our types that are known to be driven by ACh to branch, for laboratory that the majority of the inputs to ganglion example, the “complex” DS ganglion cell types G4, G7, cells in turtle retina are indeed from amacrine cells G14, G15, G17, G19, G20, G21 and G22. (Guiloff, Jones & Kolb, 1988).

Very large, weakly stained ChAT-IR cell bodies were observed in the GCL and they may correspond to true REFERENCES ganglion cells, because we also saw ChAT-IR fibers in the nerve fiber layer. These putative ganglion cells had

Ariel, M. & Adolph, A. R. (1985). Neurotransmitter inputs to direc-

weakly stained dendrites which appeared to ramify at the tionally sensitive turtle retinal ganglion cells. Journal of Neurophys- iology, 54, 1123-I 143.

same level in the IPL as the dominant plexus of ChAT- Ariel, M. & Daw, N. W. (1982a). Effects of choline@ drugs on

IR type I and type II amacrine cell dendrites. A possible receptive field properties of rabbit retinal ganglion cells. Journal of

correspondence between these ChAT-IR ganglion cells Physiology, London, 324, 135-160.

and previously described morphological types of Ariel, M. & Daw, N. W. (1982b). Pharmacological analysis of direc-

ganglion cells (Kolb, 1982; Kolb et al., 1988), may be tionally sensitive rabbit retinal ganglion cells. Journal of Physiology, London, 324, 165-185.

made with the following types: the medium-sized, bi- Barlow, H. B. & Levick, W. R. (1965). The mechanism of directionally

stratified G14 (sl/s2, s3), G15 (sl, s4) and G16 (sl/s2, s4) selective units in rabbit retina. Journal of Physiology, London, 178,

and the large, u-i-stratified G17 (sl/s2, s2/s3, s3/s4). Most 477-504.

of these ganglion cells project to the optic tectum in the Criswell, M. H. & Brandon, C. J. (1987). Immunocytochemical

turtle (Guiloff, Amermtiller & Kolb, 1989; Guiloff & evidence that turtle cones are cholinergic. Investigative Ophthal- mology and Visual Science (Suppl.), 28, 278.

Kolb, 1992), and furthermore, some of them are believed Daw, N. W. & Ariel, M. (1981). Effect of synaptic transmitter drugs

to correspond to direction-selective physiological types on receptive fields of rabbit retinal ganglion cells. Vision Research,

(Jensen & DeVoe, 1983). 21, 164331648.

Excitatory input from cholinergic amacrine cells is Eckenstein, F. & Thoenen, H. (1982). Production of specific antisera

thought to make a strong contribution to the receptive and monoclonal antibodies to choline acetyltransferase: Character- ization and use for identification of choline@ neurons. EMBO

field properties of the direction-selective ganglion cells in Journal, 1, 363-368.

turtle retina, as in other vertebrates (Barlow 8z Levick, Ehinger, B. & Dowling, J. E. (1987). Retinal neurocircuitry and

1965; Oyster, 1968; Wyatt dz Daw, 1976; Ariel & Daw, transmission. In Bjorklund, T., Hokfelt, T. & Swanson, W. (Eds),

1982a, b; Glickman, Adolph & Dowling, 1982; Famigli- Handbook of chemical neuroanatomy, Vol. 5, Integrated systems of

etti, 1985). In fact, the turtle retina has become one of the CNS, Part 1 (pp. 389-446). Amsterdam: Elsevier.

Famiglietti, E. V. Jr (1983). “Starburst” amacrine cells and cholinergic the preferred preparations to study complex visual neurons: Mirror-symmetric ON and OFF amacrine cells of rabbit

phenomena such as color, orientation-, motion-, and retina. Brain Research, 261, 138-144.

direction-selectivity, because it offers several advantages Famiglietti, E. V. Jr (1985). Synaptic organization of on-off direction-

over its mammalian counterparts in terms of experimen- ally selective ganglion cells in the rabbit retina. Society for Neuro-

tal ease in handling. Thus, electrophysiological studies science Abstracts, 11, 337.

Famiglietti, E. V. Jr (1987). Starburst amacrine cells in cat retina are combined with the application of pharmacological associated with bistratified, presumed directionally selective,

agents in the turtle retina have shown that ACh produces ganglion cells. Brain Research, 413, 404-408.

an enhancement of ganglion cell spike activity while the Glickman, R. D., Adolph, A. R. & Dowling, J. E. (1982). Inner

ACh potentiator physostigmine reduces or eliminates plexiform circuits in the carp retina: Effects of cholinergic agonists,

directional selectivity (Ariel & Adolph, 1985). The same GABA and substance P on the ganglion cells. Brain Research, 234, 81-99.

study revealed that another traditional neurotransmitter, Guiloff, G. D. & Kolb, H. (1991). Choline acetyl-transferase immuno-

gamma aminobutyric acid (GABA) inhibits ganglion cell reactive (CbAT-IR) neurons in the turtle retina. Znvestigatiue

spike activity and that the GABA antagonist picrotoxin Ophthalmology and Visual Science (Suppl.), 32, 1262.

reduces or eliminates ganglion cell directional selectivity. Guiloff, G. D. & Kolb, H. (1992). Ganglion cells types of the turtle

Therefore, the evidence suggests that the neurotransmit- retina that project to the optic tectum: Intracellular HRP injections of retrogradely, rhodamine-marked cell bodies. Visual Neuroscience,

ter likely to excite turtle direction-selective ganglion cells 8, 295-3 13.

is ACh, while lateral inhibition in the null direction may Guiloff, G. D., Ammermiiller, J. & Kolb, H. (1989). Morphology of

be mediated by GABA-containing amacrine cells (Hurd ganglion cells that project to the optic tectum in the retina of the

& Eldred, 1989). Ariel and Adolph (1985) suggest that turtle. Znvesfigarive Ophthalmology and Visual Science (Suppl.), 30, 348.

the inhibitory neurotransmitter glycine might also be Guiloff, G. D., Jones, J. & Kolb, H. (1988). Organization of the inner involved, regulating presynaptic excitatory pathways plexiform layer of the turtle retina: An electron microscopic study.

leading to the direction-selective ganglion cell (pre- Journal of Comparative Neurology, 272, 28&292.

sumably via glycinergic amacrine cells). This situation is Hurd, L. B. & Eldred, W. D. (1989). Localization of GABA- and

similar to the one reported for rabbit direction-selective GAD-like immunoreactivity in the turtle retina. Visual Neuroscience,

ganglion cells (Barlow & Levick, 1965; Oyster, 1968; 3, 9-20.

Jensen, R. J. & De Voe, R. D. (1983). Comparisons of directionally Wyatt & Daw, 1976; Ariel & Daw, 1982a, b; Famiglietti, selective with other ganglion cells of the turtle retina: Intracellular

1985). Interestingly, the choline@ amacrine cells de- recording and staining. Journal of Comparative Neurology, 217,

scribed in this study stratify in the IPL in levels that 27 l-287.

would permit them to make excitatory synaptic inputs Kolb, H. (1982). The morphology of the bipolar cells, amacrine cells

onto the different direction-selective ganglion cell types and ganglion cells in the retina of the turtle Pseudemys scripta elegans. Philosophical Transactions of the Royal Society of London

ChAT-IMMUNOREACTIVE NEURONS IN TURTLE RETINA 2029

of the turtle retina (Guiloff & Kolb, 1992). Furthermore, Series B. 298. 355-393.


Kolb, H., Perlman, I. & Nonnann, R. A. (1988). Neural organization of the retina of the turtle Clemys caspica: A light microscopic and Golgi study. Visual Neuroscience, I, 47-72.

Millar, T. J., Ishimoto, I., Chubb, I. W., Epstein, M. L., Johnson, C. D. & Morgan, I. G. (1987). Cholinergic amacrine cells of the chicken retina: A light and electron microscope immunocytochemical study. Neuroscience, 21, 725-143.

Oyster, C. W. (1968). The analysis of image motion by the rabbit retina. Journal of Physiology, London, 199, 613435.

Wiissle, H. & Riemann, H. J. (1978). The mosaic of nerve cells in the mammalian retina. Proceedings of the Royal Society of London Series B, 200, 441-461.

Wyatt, H. J. & Daw, N. W. (1976). Specific effects of neurotransmitter antagonists on ganglion cells in rabbit retina. Science, 192, 204-205.

Acknowledgements-We thank Drs Louis Hersch and Andrew Mariani for sending us batches of human placental antiChAT. The funds to generate the antisera to ChAT were provided by the Alzheimer’s Disease and Related Disorders Assoc. Inc. This work was supported by NIH Grant EY 04855 to HK.

Documents

Neurons immunoreactive to choline acetyltransferase in the turtle retina