Proc. Natl. Acad. Sci. USA
Vol. 73, No. 6, pp. 1806-1810, June 1976
Biochemistry

Localization of acetylcholine receptors during synaptogenesis

in retina

(a-bungarotoxin/neuron development/cultured neurons)

ZVI VOGEL AND MARSHALL NIRENBERG

Laboratory of Biochemical Genetics, National Heart and Lung Institute, National Institutes of Health, Bethesda, Maryland 20014

Contributed by Marshall Nirenberg, January 9, 1976

ABSTRACT __ Nicotinic acetylcholine receptors are synthe-
sized in chick embryo retina before synapses appear. Most of
the receptors are found associated with neurites in the inner
synaptic layer of the retina; later in development the receptors
appear in the outer synaptic layer. .

Cells dissociated from retina and cultured in vitro form ag-
gregates and also sort out into regions comprised either of
neurites or cell bodies. Most of the nicotinic acetylcholine re-
ceptors of aggregates are found associated with neurite regions.
The receptor distribution of cultured retina cells thus resembles
that of intact retina.

 

125]-Labeled a-bungarotoxin, a specific ligand of the nicotinic
acetylcholine (AcCh) receptor, was used to detect AcCh re-
ceptors in the developing chick embryo retina. This species of
AcCh receptor is a synaptic marker, since the AcCh receptors
of striated muscle cells (1, 2) and certain parasympathetic
neurons (3) are found in abundance only at the site of the sy-
napse. After denervation, the receptors are found over the entire
surface of the cell.

Most of the recent information regarding the nicotinic AcCh
receptor has been obtained with the receptors of the electric
organ of fish, and striated muscle (for reviews see refs. 4-6).
a-Bungarotoxin has been used to detect AcCh receptors in
sympathetic ganglia (7) and brain (8-10), but the nicotinic
AcCh receptors of retina have not been studied in this manner.
The vertebrate retina is a model system for studies on synaptic
communication, since relatively few types of neurons are
present and cells dissociated from retina form synapses in vitro
(11, 12). The retina contains acetylcholine and enzymes that
catalyze the synthesis and the hydrolysis of this compound
(13-17). In addition, electrophysiologic studies show that some
neurons in retina respond to acetylcholine and that both nico-
tinic and muscarinic AcCh receptors are present in the retina
of certain organisms (18-22).

In this report, the concentration of nicotinic AcCh receptors
and their location in chick embryo retina are described as a
function of the developmental age of the retina.

MATERIALS AND METHODS

Assay of Homogenates for !25I-Labeled a-Bungarotoxin
Binding. Neural retinas from white leghorn chick embryos
were dissected in cold Dulbecco’s phosphate-buffered saline
(Gibco) and then homogenized in 0.05 M Tris-HCl, pH 7.4, in
a ground glass Duall homogenizer (Kontes) using 0.25-1.25 ml
of buffer per retina so that the final concentration of protein
was 3-6 mg/ml. Homogenates were frozen immediately and
stored at —191°; binding sites for a-bungarotoxin (aBT) were
stable for at least 1 year. a-Bungarotoxin labeled with 2 atoms
of }25] per molecule of toxin (!>I-aBT) was prepared as de-

 

Abbreviations: aBT, a-bungarotoxin; !251-aBT, a-bungarotoxin labeled
with 2 atoms of !25I per toxin molecule; AcCh, acetylcholine.

1806

scribed previously (23). Stock solutions of !I-aBT contained
0.25-0.5 uM !5]-aBT; 3 mM sodium phosphate buffer, pH 7.5;
150 mM NaCl; and 2 mg of crystalline bovine serum albumin
per ml. Each binding reaction mixture contained the following
components in a final volume of 0.1 ml: 10 nM !*]-aBT (2-4
ul of a stock solution); 0.05 M Tris-HCl, pH 7.4; 0.2 mg of serum
albumin; and 0-180 ug of retina homogenate protein. Reaction
mixtures were incubated for 30 min at 37°, diluted with 3 ml
of solution C (0.05 M Tris-HCl, pH 7.4, and 2 mg of serum al-
bumin per ml) at 8°, and immediately filtered (vacuum)
through a wet cellulose acetate filter (Millipore, EGWP 25 mm
in diameter, 0.2 um pore size). The filter was washed four times
with 3 ml portions of solution C, the filter was dissolved in 10
ml of Instabray (Yorktown Research), and radioactivity was
determined. Zero time values (100-200 cpm, approximately
0.25 fmol of }25J-aBT) were subtracted from the values shown.
The amount of !2]-aBT bound was proportional to protein
concentration in the range studied (0-180 yg of protein). Pro-
tein was determined by modification of the method of Lowry
et al. (24).

Autoradiography of '%5J-aBT Bound to Intact Retina.
Neural retina was dissected in cold solution A [the Dulbecco-
Vogt modification of Eagle’s medium (Gibco Catalog no. H-21)
with 20 mM Hepes (N-2-hydroxyethylpiperazine-N’-2-eth-
anesulfonic acid), pH 7.4, instead of NaHCOs, and 2 mg of
serum albumin per ml of medium]. Pieces of retina were in-
cubated for 60 min at 37° in 0.5-1.0 ml of solution A with 10
nM !25]-aBT and then were washed five times with 5 ml of cold
solution A and twice with solution A without serum albumin.
The tissue was fixed with 10% (vol/vol) neutral formalin or 2.5%
(wt/vol) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4)
and embedded in paraffin. Serial sections, 8 um thick, were
mounted on glass slides, and the paraffin was removed; the
slides were coated with Kodak NTB-2 emulsion diluted 1:1 with
water. Slides were cooled to gel the emulsion and stored at 4°
in the dark. Autoradiographs were developed for 4 min at 20°
with Kodak D-19 developer diluted 1:1 with water and fixed
for 5 min with Kodak F5. The slides were rinsed with water;
some were stained with 0.05% toluidine blue in 0.66 M
borax.

Retina Cell Cultures. Cells were dissociated from retinas of
8-day-old chick embryos and aggregated in vitro by a modi-
fication of the method described by Sheffield and Moscona (11).
Single cells, dissociated from retina with 0.25% trypsin (three
times crystallized, Worthington) and 50 ug/ml of DNase I
(Worthington), were cultured in rotating flasks (80 rpm) in a
37° incubator in a humidified atmosphere of 5% CO2-95% air.
Each flask contained 9 X 108 cells and 1.5 ml of medium (80%
Eagle's Basal Medium and 20% fetal bovine serum). The me-
dium was changed every other day. After 7 days 10 nM !?*I-
aBT was added to the aggregates in the rotating flasks. One
hour later the aggregates were collected and washed four times
 

Biochemistry: Vogel and Nirenberg

 

 

 

 

 

Q

=

5

5

a

Z 25+

Pa

2

© 20F

x

5

9 15}

2

a

& 10

°

Bs o

Oo bee 87]
2 0 6% 30 ~~«90
= MINUTES

Fic. 1. The rate of !25I-aBT binding. Portions of a 2-week-old
chick retina homogenate (62.2 ug of protein per reaction mixture) were
incubated with 1 nM !25]-aBT with or without 2 »M unlabeled aBT
at the temperatures and times specified.

with growth medium by centrifugation, and three times with
growth medium without serum. The aggregates were fixed with
glutaraldehyde, embedded in paraffin, sectioned, and subjected
to autoradiography as described above.

RESULTS AND DISCUSSION

125].oBT Binding. The rate of binding of !*I-aBT to re-
ceptors in a homogenate of chick retina is shown in Fig. 1.
Maximum !5]-aBT binding was obtained after 15 min of in-
cubation at 37°; rates of binding were lower at 19° and 0°.
Nonspecific binding of !?5-aBT in the presence of a large ex-
cess of unlabeled aBT (2 uM) was relatively low.

The relation between !*I-aBT concentration and the amount
of labeled toxin bound is shown in Fig. 2. The amount of
125].¢BT bound increased sharply as the concentration of
1235]_aBT was elevated from 1 to 7.5 nM. Higher concentrations
of 125]-@BT (10-40 nM) increased binding of labeled aBT only
slightly. Little nonspecific binding of !*I-aBT was observed
in the presence of 2 uM unlabeled aBT. The number of specific
aBT binding sites in chick retina, 2 weeks after hatching, esti-

 

 

 

 

 

 

oO 02 03

 

 

 

0

fmol DIIODO-a-BUNGAROTOXIN BOUND

10 20 30 40
DIIODO-e-BUNGAROTOXIN (nM)

Fic. 2. Relation between 125{-diiodo-aBT concentration and the
amount of labeled toxin bound. Portions of a 2-week-old chick retina
homogenate (62.2 ug of protein per reaction mixture) were incubated
for 5 min at 20°, in the absence (O), or presence (@), of 2 uM unlabeled
aBT. !25]-Labeled a-bungarotoxin then was added at the concen-
trations specified and the reaction mixtures were incubated and as-
sayed for bound radioactivity as described in Materials and Methods.
A Scatchard plot of specific !25I-aBT binding; i-e., binding obtained
after subtracting the radioactivity bound in the presence of 2 uM
unlabeled toxin, is shown in the insert; the ordinate refers to the
bound/free }25I-aBT ratio; the abscissa corresponds to nM toxin
bound per 62.2 ug of protein.

Proc. Natl. Acad. Sci. USA 73 (1976) 1807

 

     

x
1004
Zz
x<
© got
O
ow OOF
Zz
Z 40
& ?.
Py oq
20+

 

 

 

0-9 6-8 7 -6 -§ -4 -3
LOG MOLARITY

Fic. 3. Inhibitors of }25I-aBT binding. Portions of a 2-week-old
chick retina homogenate (50 wg of protein per 80 ul of reaction mix-
ture) were incubated for 5 min at 20° in the presence of the com-
pounds indicated. }75I-aBT (10 nM) then was added (the final volume
was 100 ul) and tubes were incubated for an additional 10 min at 20°.
The final concentration of each compound is indicated on the abscissa.
Tubes with acetylcholine also contained 3 uM eserine sulfate. One
hundred percent on the ordinate corresponds to 6.25 fmol of !I-oBT
specifically bound in the absence of inhibitor. This value was 35% of
that obtained after 30 min of incubation at 37°. Abbreviations rep-
resent the following: a-BT, unlabeled a-bungarotoxin; NIC, nicotine;
d-TC, d-tubocurarine; ACH, acetylcholine plus 3 1M eserine sulfate;
CARB, carbamylcholine; ATR, atropine; ESE, eserine sulfate; PILO,
pilocarpine.

mated from a Scatchard plot (Fig. 2, insert), is 525 fmol/mg of
protein. The (!251-aBT-receptor) complex dissociates in a bi-
phasic manner in the presence of 2 uM unlabeled aBT (50%
dissociation in 30 min at 37°). Slower rates of dissociation were
observed at 20° and 0° (data not shown).

Receptor Specificity. Nicotinic and muscarinic AcCh re-
ceptors can be distinguished by differences in affinity for li-
gands. In Fig. 3, the effects of various ligands upon !*I-aBT
binding are shown. Binding was inhibited 50% by 1 nM unla-
beled aBT, 0.3 uM nicotine, and 1 »M d-tubocurarine, each
compound known to have a high affinity for nicotinic AcCh
receptors. Fifty percent inhibition of toxin binding was obtained
with 6 uM acetylcholine (3 uM eserine sulfate also was present,
a concentration without effect on aBT binding), and with 100
uM carbamylcholine. Compounds that interact preferentially
with muscarinic AcCh receptors, such as pilocarpine and at-
ropine, either did not affect }25I-aBT binding or inhibited
binding only at relatively high concentrations. Putative neu-
rotransmitters such as L-glutamic acid, glycine, y-aminobutyric
acid, dopamine, and [-norepinephrine did not affect «BT
binding at 10-1000 uM; however, serotonin stimulated aBT
binding 15%. Choline (100 4M) inhibited !>I-aBT binding
(data not shown).

AcCh Receptors During Embryonic Development. The
amount of bound !25]-q@BT is shown in Fig. 4 as a function of
the developmental age of chick retina. Under the conditions
used (10 nM !25]-@BT, 30 min incubation at 37°) approximately
80% of the receptors bind aBT. Specific binding sites for aBT
are present in low concentration (7.5 fmol/mg of protein) in
chick retina on the 6th embryonic day. The concentration of
receptors increases 10-fold between the 6th and the 13th em-
bryonic day; however, >80% of the receptors are synthesized
between the 13th and last embryonic day during the period of
retina synapse formation. These results show that the receptors
1808 Biochemistry: Vogel and Nirenberg

 

—

 

 

 

 

 

06+ 3
? |
zit g
é 5
oO 4
& od 2
~ z
5 1s
5 | 2
o z
a Z
Z 0
<x O2t ae
o 3
e Ee
E a
a a 4
0 - 250°

IN OVO DAYS POST HATCHING

Fic. 4. Amount of !251-aBT bound as a function of the develop-
mental age of chick retina. Filled symbols (@) represent pmol of toxin
bound specifically per mg of protein; open symbols (©) represent pmol
of toxin bound specifically per retina. Each point is the mean of 18
to 60 values (3 to 10 retina homogenates; six concentrations of protein
were assayed from each homogenate). Specific toxin binding is shown;
nonspecific binding has been subtracted.

are synthesized before synapses appear and suggest that ces-
sation of both receptor accumulation and synapse formation
may be coupled. The maximum receptor concentration is at-
tained at the time of hatching (400 fmol of }25I-aBT bound per
mg of protein). Almost the same receptor concentration is found
in adult retina; however, the number of specific binding sites
for aBT per adult retina (2700 fmol per retina) is twice that of
the newly hatched chick (1350 fmol per retina). Thus, adult
retina either has more synapses, more receptors per synapse,
or more extrajunctional receptors than the retina of the newly
hatched chick. Adult rabbit retina bound 107 fmol of }5]-aBT
per mg of protein under the same conditions.

Receptor Distribution in Retina. The distribution of nico-
tinic AcCh receptors in retina was determined by autoradi-
ography. Adult retina was incubated with '**I-aBT, fixed,
sectioned, and either stained with toluidine blue or subjected
to autoradiography as shown in Fig. 5A and B, respectively. Cell
bodies and neurites rich in synaptic connections are found in
separate layers (Fig. 5A). The upper layer consists of photore-
ceptor cells; next is the outer plexiform layer which consists of
neurites and synapses; next is a layer of cell bodies, the inner
nuclear layer; and then there is a wide layer of synaptic con-
nections and neurites, the inner plexiform layer. The lowest
layer of cell bodies is composed of ganglion neurons; the last
layer consists of ganglion neuron axons.

The autoradiographs show that most of the binding sites for
125]_qBT are located in the inner plexiform layer, which consists
primarily of synaptic connections between processes of bipolar,
amacrine, and ganglion neurons. Four horizontal bands with
relatively high concentrations of silver grains can be distin-
guished within the inner plexiform layer. Silver grains also are
found in the outer plexiform layer, which contains synapses
between photoreceptor, horizontal, and bipolar neurons. Rel-
atively few grains are associated with cell body layers, or with

Proc. Natl. Acad. Sci. USA 73 (1976)

 

Fic. 5. Section of an 8-month-old chicken retina. (A) Tolui-
dine-blue-stained section, not subjected to autoradiography. (B) The
corresponding field in an unstained serial section, subjected to au-
toradiography for 110 days. The initial specific activity '*]-aBT was
335 Ci/mmol of toxin. Bright field views are shown in both panels; the
bar represents 25 um. Abbreviations of retina layers are as follows:
NF, nerve fiber layer, GC, ganglion cell layer, IP, inner plexiform layer
(i.e., inner synaptic layer); IN, inner nuclear layer, OP, outer plexiform
layer (i.e., outer synaptic layer); IS and OS, inner and outer segments
of photoreceptor cell layer, respectively.

s
axons of ganglion neurons. }25]-«-Bungarotoxin does not bind
to pigment cells, choroid, pecten, or sclera.

Autoradiographs of sections of rat and rabbit retina are shown
in Fig. 6. Again, most of the silver grains are associated with the
inner plexiform layer. Some putative ganglion and amacrine
neurons are labeled in rat retina (Fig. 6A); however, the outer
plexiform layer of the rat is essentially unlabeled. In rabbit
retina (Fig. 6B and C) silver grains are found in both the inner
and outer plexiform layer. Approximately 10-20% of the rabbit
ganglion cell bodies and some cell bodies in the lower portion
of the inner nuclear layer, presumably amacrine neurons, are

 

Fic. 6. Autoradiography of sections of adult rat and rabbit retina
labeled with !25I-aBT. (A) Bright field view of a stained section of
Fisher rat retina subjected to autoradiography for 23 days. The initial
specific activity of !251-aBT was 260 Ci/mmol of toxin. (B) and (C)
Phase contrast views of stained sections of New Zealand white rabbit
retina subjected to autoradiography for 27 days; the initial specific
activity of 1251-aBT was 220 Ci/mmol of toxin. The bar represents 25
pm.
Biochemistry: Vogel and Nirenberg

 

Fic. 7. Autoradiography of sections of 6- to 17-day-old chick
embryo retina, labeled with }75I-aBT. Retina sections were subjected
to autoradiography for 29 days. The initial specific activity of 51-aBT
was 485 Ci/mmol of toxin. The same magnification was used for each
photomicrograph; the bars represent 25 um. (A) Phase contrast view
of a stained autoradiograph of 6-day-old embryo retina. (B) Bright
field view of the area shown in panel A before staining. (C) Phase
contrast view of a stained autoradiograph of 9-day-old embryo retina.
(D) Bright field view of the same area shown in panel C before stain-
ing. (E) Bright field view of a stained section of 13-day-old embryo
retina not subjected to autoradiography. (F) Bright field view of the
corresponding area of an unstained autoradiograph of an adjacent
section. (G) and (H) 17-Day-old embryo retina sections, treated as
in panels FE and F, respectively.

labeled. Some cells in the outer synaptic layer, thought to be
horizontal neurons, also are labeled. The concentration of
ganglion neurons is considerably higher in the center of the eye
compared to the periphery; however, the concentration of silver
grains in the inner plexiform layer does not change markedly
(<2-fold).

Three types of control experiments were performed. First,
incubation of chicken retina with 200 uM d-tubocurarine or
1 uM unlabeled aBT reduced the amount of !25]-aBT bound
to retina by 85% and 95%, respectively, and the few silver grains
found after autoradiography were randomly distributed. Sec-
ond, autoradiography of 0.5 xm thick, Epon-embedded sections
of chick retina revealed the same silver grain distribution as that
of 8 um paraffin sections described above. Since Epon sections
have an essentially flat surface coated with the autoradiograph
emulsion, the grain distribution does not result from surface
irregularities of sections after paraffin has been removed. Third,
the concentration of specific !1-aBT binding sites in intact
retina is approximately the same as that of homogenates pre-
pared from sister retinas, which suggests that most binding sites
in intact retina are accessible in the labeled toxin.

Embryonic chick retinas at various stages of development
were labeled with !25I-aBT and sections were subjected to
autoradiography. Pairs of photomicrographs, one emphasizing
retina structure, the other silver grain distribution, are shown
in Fig. 7 for the 6th, 9th, 13th, and 17th embryonic days. The
first neurons to appear are ganglion neurons which, by the 6th
embryonic day, are found ina layer next to the margin of the
retina (25, 26). Autoradiographs of retina on the 6th embryonic
day (Figs. 7A and B) show that !*5]-aBT binds to ganglion
neurons but not to most neuroblasts. Coulombre (25) has re-
ported that the inner and outer plexiform layers of chick retina
appear on the 8th and 9th embryonic days, respectively.
125}_«-Bungarotoxin binds predominantly to the inner plexi-
form layer of 9-, 13-, and 17-day-old chick embryo retina. The
receptors were not found in appreciable numbers in the outer
synaptic layer until the 17th embryonic day.

Proc. Natl. Acad. Sci. USA 73 (1976) 1809

 

FIG. 8. Sections of an aggregate formed from dissociated 8-
day-old chick embryo retina cells, cultured in vitro for 7 days. (A)
Toluidine-blue-stained section, not subjected to autoradiography,
phase contrast view. (B) Bright field view of a stained autoradiograph
exposed for 36 days. The initial specific activity of !*I-aBT was 300
Ci/mmol of toxin. The bar represents 50 um.

The AcCh receptors are distributed fairly uniformly
throughout the inner synaptic layer on the 9th and 11th em-
bryonic days; by the 17th embryonic day 2 horizontal bands of
receptors were observed in some autoradiographs, and in the
adult, four diffuse bands of receptors can be seen within the
inner synaptic layer. Acetylcholinesterase activity first appears
in the 4-day-old chick embryo retina associated with ganglion
neurons and later with amacrine neurons (13, 17, 18); after
hatching four horizontal bands of acetylcholinesterase activity
also are present in the inner synaptic layer of chick retina (17,
18).

On the 13th embryonic day retina cells with abundant AcCh
receptors were found which resemble amacrine neurons.
Synaptic connections first appear in chick retina on the 13th
embryonic day (27-29). Thus genes for nicotinic AcCh recep-
tors are expressed in some retina cells at least 7 days before sy-
napses appear. Striated muscle cells also synthesize nicotinic
AcCh receptors in the absence of neurons (28, 30, 31) and after
denervation (2, 32, 33).

Both rod and cone photoreceptor cells of the turtle synthesize
acetylcholine (16). In chick retina, nicotinic AcCh receptors are
associated with dendrites of ganglion neurons and cells which
resemble amacrine neurons. Our working hypothesis is that
acetylcholine is a transmitter of bipolar neurons, and that some
amacrine and ganglion neuron responses to bipolar neurons are
mediated by nicotinic AcCh receptors. Further work also is
needed to determine whether nicotinic AcCh receptors in the
outer plexiform layer of chick and rabbit, but not rat, retina are
synthesized by certain horizontal and/or bipolar neurons.

The relatively high concentrations of nicotinic AcCh re-
ceptors in the synaptic layers of chick retina and the low con-
centrations of receptors associated with cell bodies and axons
of ganglion neurons suggest that nicotinic AcCh receptors of
retina may be localized at certain synaptic sites much as AcCh
receptors are localized at the neuromuscular synapse (1, 2) and
at certain synapses of parasympathetic neurons (3). Preliminary
results with an immunochemical assay for detecting nicotinic
AcCh receptors at the electron microscope level (34) suggest
that most of the receptors are not randomly distributed in the
inner synaptic layer but instead are found in small areas which
contain relatively high concentrations of receptors.

Retina Cell Cultures. Neural retina tissue from 8-day-old
chick embryos was dissociated into single cells and the cells then
were cultured in rotating flasks under conditions that favor
aggregation as described by Sheffield and Moscona (11). A
section of a cell aggregate cultured for 10 days is shown in Fig.
8A. Cell bodies and processes were found to sort out into discrete
regions which resemble the layers of cell bodies and processes
of the intact retina. A similar section of an aggregate which was
incubated with !251-aBT and subjected to autoradiography is
shown in Fig. 8B. Labeled toxin bound primarily to the neu-
1810

Biochemistry: Vogel and Nirenberg

rite-rich regions; considerably less toxin was associated with cell
bodies, These results show that dissociated retina cells form
aggregates, synthesize nicotinic AcCh receptors, and extend

neurites which together with AcCh receptors sort out from cell
bodies.

Cultured retina cells can be used as a model system for studies

related to the process of synapse formation. At least three types
of synapses form in vitro (11, 12), and the concentration of
synaptic connections synthesized in vitro almost equals that
found in the intact retina (35).

eonr

10.

11.
12.

13.

Miledi, R. (1960) J. Physiol. 151, 24-30.

Axelsson, J. & Thesleff, S. (1959) J. Physiol. 149, 179-198.
Kuffler, S. W., Dennis, M. J. & Harris, A. J. (1971) Proc. R. Soc.
London Ser. B. 177, 555-568.

Cohen, J. B. & Changeaux, J. P. (1975) Annu. Rev. Pharmacol.
15, 83-103.

de Robertis, E. & Schacht, J., eds. (1974) Neurochemistry of
Cholinergic Receptors (Raven Press, New York).

Bennett, M. V. L., ed. (1974) “Synaptic transmission and neuronal
interaction,” Society of General Physiologists Series (Raven
Press, New York.), Vol. 28.

Greene, L. A., Sytkowski, A. J., Vogel, Z. & Nirenberg, M. W.
(1973) Nature 243, 163-166.

Moore, W. J. & Loy, N. J. (1972) Biochem. Biophys. Res. Com-
mun. 46, 2093-2099.

Eterovic, V. A. & Bennett, E. L. (1974) Biochim. Biophys. Acta
362, 346-355.

Polz-Tejera, G., Schmidt, J. & Karten, H. J. (1975) Nature 258,
349-351.

Sheffield, J. B. & Moscona, A. A. (1970) Dev. Biol. 23, 36-61.
Stefanelli, A., Zacchei, A. M., Caravita, S., Catoldi, A. & leradi,
L. A. (1967) Experientia 23, 199-200.

Stell, W. K. (1972) in Handbook of Sensory Physiology, ed.

14.

15.
17.
18.

19.
20.

21.
22.

24.

26.

28.

29.
30.
31.

Proc. Natl. Acad. Sci. USA 73 (1976)

Fourtes, M. G. F. (Springer-Verlag, Berlin, Heidelberg and New
York), Vol. 7, part 2, pp. 111-213.

Gouras, P. (1976) in Pharmacology and Toxicology of the Eye,
ed. Dikstein, 5. (Charles C Thomas, Springfield, Illinois), in press.
Lindeman, V. F. (1947) Am. J. Physiol. 148, 40-44.

Lam, D. M. (1972) Proce. Nail. Acad. Sci. USA 69, 1987-1991.
Shen, S. C., Greenberg, P. & Boel, E. J. (1956) J. Comp. Neurol.
106, 433-461.

Nichols, C. W. & Koelle, G. B. (1969) J. Comp. Neurol. 133, 1-16.
Noell, W. K. & Lansansky, A. (1959) Fed. Proc. 18, 115.
Masland, R. H., Ames, A., II & Livingstone, C. J. (1975) Neurosct.
Abst. 5th Meeting of the Soc. for Neuroscience, New York, Vol.
1, page 110, Abstract 173.

Ames, A. & Pollen, D. A. (1969) J. Neurophysiol. 32, 424-442.
Straschill, M. & Perwein, J. (1973) Pfligers Arch. 339, 289-298.
Vogel, Z., Sytkowski, A. J. & Nirenberg, M. W. (1972) Proc. Natl.
Acad. Sci. USA 69, 3180-3184.

Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.
(1951) J. Biol. Chem. 193, 265-275.

Coulombre, A. J. (1965) Am. J. Anat. 96, 153-189.

Kahn, A. J. (1973) Brain Res. 63, 285-290.

Meller, K. (1968) in Veroffentlichungen aus der Morpholog-
ischen Pathologie, eds. Buchner, F. & Giese, W. (Gustav
Fischer-Verlag, Stuttgart), Vol. 77, pp. 1-77.

Sheffield, J. B. & Fischman, D. A. (1970) Z. Zellforsch. M ikrosk.
Anat. 104, 405-418.

Hughes, W. F. & LaVelle, A. (1974) Anat. Rec. 179, 297-302.
Fambrough, D. & Rash, J. E. (1971) Dev. Biol. 26, 55-68.
Sytkowski, A. J., Vogel, Z. & Nirenberg, M. W. (1973) Proc. Nail.
Acad. Sci. USA 70, 270-274.

Miledi, R. & Potter, L. T. (1971) Nature 233, 599-603.
Hartzell, H. C. & Fambrough, D. M. (1972) J. Gen. Physiol. 60,
248-262.

Daniels, M. & Vogel, Z. (1975) Nature 253, 339-341.

Vogel, Z., Daniels, M. & Nirenberg, M. W. (1974) Fed. Proc. 33,
1476.