proc. Natl. Acad. Sci. USA
Vol. 78, No. 4, pp. 2145-2149, April 1981.
Biochemistry

A topographic gradient of molecules in retina can be used to

identify neuron position
(cell membrane/antigen/embryo/synapses/hybrid cells)

CG. Davip TRISLER, MICHAEL D. SCHNEIDER, AND MARSHALL NIRENBERG

Laboratory of Biochemical Genetics, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20205

Contributed by Marshall Warren Nirenberg, December 24, 1980

ABSTRACT A monoclonal antibody was obtained that binds
to cell membrane molecules distributed in a topographic gradient
in avian retina. Thirty-five-fold more antigen was detected in dor-
soposterior retina than in ventroanterior retina. Most of the an-
tigen was associated with the synaptic layers of the retina. Less
antigen was detected in cerebrum, thalamus, cerebellum, and
optic tectum, but little or none was found in non-neural tissues
tested. The antigen was found on most or all cell types in retina,
and the concentration of antigen found is a function of the square
of the circumferential distance from the ventroanterior pole of the
gradient toward the dorsoposterior pole. Thus, the antigen can be
used as a marker of cell position along the ventroanter-
ior-dorsoposterior axis of the retina.

 

Topographic relationships between retina ganglion neurons are
conserved, forming point-to-point representations of the retina,
when ganglion neurons synapse in tectum or certain other re-
gions of brain. Thus, the retina is a favorable model system for
studying the formation of synaptic circuits. However, the mo-
lecular basis for spatial order has not been defined. Sperry (1)
hypothesized that two orthogonal gradients of molecules on ret-
ina ganglion neurons and corresponding gradients of compli-
mentary molecules in the optic tectum might determine the
specificity of synaptic connections between retina and tectum
neurons. Other mechanisms proposed include adhesive inter-
actions between migrating neurites, myelination of bundles of
axons, and formation of extracellular channels by glia to guide
axons in appropriate directions (2).

Antibodies provide a means of identifying surface molecules
and reactions required for neural function (3-7). Rabbit anti-
serum to clonal retina hybrid cells was used to detect an antigen
with a restricted domain in retina (8).

Our objective was to detect cell surface molecules with topo-
graphic specificity in the retina, as candidates for neuronal rec-
ognition molecules. Monoclonal antibodies were obtained by
fusing P3X63 Ag8 mouse myeloma cells (9) with spleen cells
from mice immunized with small portions of dorsoposterior or
ventral chicken embryo retina. A hybridoma antibody was ob-
tained that binds to cell membrane molecules that are distrib-
uted in a dorsoposterior > ventroanterior gradient in retina.

RESULTS

The rationale for the experiments is shown in Fig. 1. Spleen
cells from mice immunized with small portions of dorsoposterior
or ventral neural retina from the left eyes of 14-day chicken
embryos were fused with P3X63 Ag8 mouse myeloma cells, and
the binding of hybridoma antibody to cultured cells from the
sector of the retina used for immunization was compared with

 

The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked “advertise-
ment” in accordance with 18 U. §. C. §1734 solely to indicate this fact.

2145

 

 

MOUSE IMMUNIZATION WITH CHICK RETINA SEGMENTS
, —— ra N, .
Ai P Ai iP
Ye lA
Vv v
HYBRIDOMA ANTIBODY SPECIFICITY
DD - D
f f a x
At iP t 1 A P
4 i ‘ i
Vv NUMBER v NUMBER
OF OF
HYBRIDOMA HYBRIDOMA
COLONIES COLONIES
tee t + 1 + + 5
+ + 13 + + 3
— — 54 — ~ 101

 

 

 

 

Fic. 1. (Upper) Hybridoma cell lines were derived from mice im-
munized with cells mechanically dissociated from dorsoposterior (Left)
or ventral (Right) retina from the left eyes of 14-day White Leghorn
chicken embryos (Gallus gallus) (sections labeled 4 and 1, respectively,
in Fig. 2). Symbols, A, D, P, and V correspond to anterior, dorsal, pos-
terior, and ventral, respectively. The choroid fissure, through which
axons exit or enter the retina, shown extending from the ventroanteri-
or margin of the retina, was used as a landmark for dissection. Female
BALB/c mice were injected intraperitoneally at 0,7, and 14 days with
8 x 10° retina cells in 0.4 ml of Dulbecco’s phosphate-buffered saline
and intravenously with 2 x 10° cells in 0.1 ml of the saline. On day
17, 108 dissociated spleen cells from each mouse were fused (10) with
2 x 107 P3X63 Ag8 mouse myeloma cells. After fusion, the cells were
suspended with 5 x 10° spleen cells from a mouse that had not been
immunized in 50 ml of medium A. [Medium A is selective medium
D20SHAT of Berzofsky et ai. (11) except that it contains 10% fetal bo-
vine serum, 1 «M aminopterin, 2 milliunits of insulin per ml, and each
nonessential amino acid at 0.1 mM.] Cells were added to 96-well plates
(Falcon) (186,000 cells per 74 j:1 of medium A per well), incubated at
37°C in a humidified atmosphere of 10% CO,/90% air, and fed addi-
tional medium A (0.1 ml per well) on the 5th and 10th days of incu-
bation. (Lower) To identify hybridomas synthesizing antibodies to re-
gional antigens in retina, cells from the portion of 12-day chicken
embryo retina (left eye) used for mouse immunization, or the rest of
the retina, were dissociated with trypsin and collagenase, and transfer
plates (12) were inoculated with 1.2 x 10° of the cells per well. Cells
were cultured for 2 days in 90% Eagle’s minimal essential medium/
10% fetal bovine serum; the medium was replaced with 50 yl of me-
dium conditioned by hybridoma cells, and plates were incubated for 30
min at 37°C. Each well was washed three times with 150 j] of solution
B (1 mg of pigskin gelatin per ml of phosphate-buffered saline) at 4°C.
Each cell monolayer was incubated for 30 min at 37°C with 50 yl of
solution B containing 1.6 nM rabbit '*°I-labeled F(ab’), antibody frag-
ment [?25]-F(ab’))] directed against mouse IgG heavy and light chains
(9799 cpm) and 500 xg of bovine serum albumin. Cells were washed
as above, and bound radioactivity was determined. F(ab’), (Cappel)
was purified on a mouse IgG-Sepharose 4B column and iodinated with
mono-!“] Bolton—-Hunter reagent (Amersham) (13).

binding to cells from the remainder of the retina. Cultured ret-
ina cells were used to increase the probability of detecting an-
tibodies directed against cell surface molecules. Sixty-eight of
2146 — Biochemistry: Trisler et al.

672 wells inoculated contained hybridoma colonies derived
from a mouse immunized with dorsoposterior retina cells. One
cell line (14H3) synthesized antibody that bound more to dor-
soposterior retina cells than to cells from the rest of the retina,
whereas 13 hybridoma antibodies bound equally to cells from
both portions of retina. Antibodies to retina were not detected
with 54 additional cell lines. Hybridoma colonies derived from
a mouse immunized with ventral retina cells were detected in
109 of the 672 wells plated. Five antibodies of the 87 tested
bound somewhat more to cells from ventral retina than to cells
from the rest of the retina, and 3 antibodies bound equally.
Thus, 14% of the 155 hybridoma cell lines examined synthesize
antibodies to retina; however, only one antigen was detected,
termed TOP for toponymic (i.e., a marker of position), with an
asymmetric distribution, more abundant in dorsoposterior ret-
ina than in the remainder of the retina.

Distribution of TOP Molecules in Retina. Retinas from right
or left eyes of chicken embryos were cut into eight sections as
shown in Fig. 2, and cells from each section were assayed for
{!*1-F(ab’),‘anti-TOP antibody‘TOP antigen] complexes. Bi-
laterally symmetrical gradients of TOP were found in retina
from right and left eyes. The highest concentrations of antigen
detected were in dorsoposterior or dorsal retina and the lowest,
in ventroanterior or ventral retina. Little }°I-F(ab’), bound to

 

VW toot tot
TOPOGRAPHIC GRADIENT OF ANTIGEN

     
 

DORSAL DORSAL

 

VENTRAL
RIGHT RETINA

VENTRAL
LEFT RETINA

pMOL 1251-F(ab‘lg BOUND/mg PROTEIN

 

 

 

Fic. 2. Gradients of TOP molecules in retina from right (e) and
left (a) eyes of 14-day chicken embryos. e———-@ and a——a, anti-TOP
antibody; @----@ and a----4, P3X63 Ag8 antibody; o——o and
A——A, buffer B without antibody to TOP. Values shown within
appropriate sections of retina are pmol of '”*I-F(ab’), specifically bound
per mg protein (i.e., pmol of !7°I-F(ab’), bound in the presence of anti-
TOP antibody minus pmol of }”°I-F(ab’), bound in the presence of
P3X63 Ag8 antibody). The assay conditions for TOP antigen used in
this and subsequent experiments, except where stated, were as follows.
Each retina was cut into eight sections as shown. Retina cells were
mechanically dissociated. in phosphate-buffered saline at 4°C by trit-
uration (10 times) with a 200 yz] micropipette tip (Medical Laboratory
Automation), and 100 j] of a cell suspension (usually 160 yg of protein;
range, 50-250 yg) was added to each well of polyvinyl chloride 96-well
V-bottom plates (Dynatech) treated with solution B. Cells were cen-
trifuged at 1300 x g for 5 min at 4°C and supernatant solutions were
decanted. Each pellet was washed three times in solution B at 4°C (150
pl each wash), suspended in 50 yl of an antibody solution containing
1.0 yl of ascites fluid in solution B (except where specified), and in-
cubated for 30 min at 4°C. Retina cells were washed three times as
above, suspended in 50 yl of solution B containing 440 nM rabbit }751-
F(ab’), (5-10 x 10* cpm) and 500 yg of bovine serum albumin, and
incubated at 4°C for 30 min, unless otherwise specified. Pellets were

. washed three times as described above, wells were separated, and ra-
dioactivity was determined. Each value shown is the mean of two or
three determinations. Protein was determined by a modification of the
method of Lowry et al. (14).

Proc. Natl. Acad. Sci. USA 78 (1981)

 

    
 
   
   
   
 

    
   
  
 

 

  

 

A. PROTEIN = [|B MONOCLONAL ANTI- 2 [0.5251 Ftab’i2 ANTI. |
o CONCENTRATION & gi BODY CONCEN- i | MOUSE IgG CONCEN-
26 TOTAL, | @ | TRATION 230; TRATIO |
g : >, / Re
a a a
£ 5 hit ; |
: 4 a
3° Z 2%
7 434 3 |
wm m | a
= 2 7 2 NONSPECIFIC
3 f ‘ol won. ¢ VENTRAL ig! HIE
g Lz RETINA | 8 Bi j
a NONSPECIFIC | = At =
-O-6 18 a
0 200 400 6007 00.0010.010.1 127, 4
ug RETINA PROTEIN $ wi ASCITES FLUID 3 uM 1251-Flab‘)2
a a

Fic. 3. (A) Effect of retina protein concentration on ?”°I-F(ab’),
binding in the presence of anti-TOP antibody (A) or P3X63 Ag8 anti-
body (0). @, }”°I-F(ab’}, bound specifically. (B) Effect on hybridoma
antibody (ascites fluid) concentration on )7°I-F(ab’), binding. Solid
symbols, specific binding of ‘”5I-F(ab’)) due to anti-TOP antibody; open
symbols, nonspecific binding with P3X63 Ag8 antibody. @, 0, Dorsal
retina, sections 4 and 5; a, 4, ventral retina, sections 1 and 8. Reaction
mixtures contained 220 nM '*]-F(ab’), (3.68 x 107° uCi/pmol). (C)
Effect of concentration of rabbit '7*]-F(ab’), anti-mouse IgG. Solid
symbols, specific '**I-F(ab’), binding; open symbols, nonspecific bind-
ing to cells from dorsal retina (@, C) or ventral retina (m, 0, a, A).

retina cells when antibody to TOP was omitted or was replaced
with P3X63 Ag8 antibody. Other hybridoma antibodies includ-
ing A2B5 (6) bound equally to cells from different regions of
retina (not shown).

Assay Conditions. The effects of varying retina protein, hy-
bridoma antibody, or }*1-F(ab’), anti-mouse IgG are shown in
Fig. 3. Specific binding of !I-F(ab’), was proportional to retina
protein in the range 65-300 yg of protein (Fig. 3A). Half-max-
imal and maximal specific binding of '!*I-F(ab'), to dorsopos-
terior retina cells were obtained with 1:1350 and 1:100 dilutions
of ascites antibody to TOP, respectively; higher concentrations
of antibody to TOP reduced specific !I-F(ab'), binding (Fig.
3B). Specific !I-F(ab’), binding to ventral retina was low; how-
ever, the concentrations of anti-TOP antibody required for half-
maximal and maximal specific binding did not differ greatly from
those found with dorsal retina. Thus, no obvious difference was
detected in the affinity of the antibody for antigen in dorsal and
ventral retina. Nonspecific binding of !*I-F(ab’), to retina cells
was low at all concentrations of P3X63 Ag8 antibody tested. The
antibody to TOP was identified as an IgG] with « light chains
(not shown). Half-maximal and maximal specific binding of }°1-
F(ab’), were obtained with approximately 0.5 and 2 uM )I-
F(ab’),, respectively, both to dorsal and ventral retina cells (Fig.
3C). Nonspecific ‘*I-F(ab’), binding in the presence of P3X63
Ag8 ascites antibody was proportional to ‘I-F(ab’), concen-
tration and was not saturating at 3.52 4M, the highest '”I-
F(ab’), concentration tested.

Geometry of the Gradient. The retina grows by accretion of
concentric rings of neurons at the periphery; thus, central retina
is the oldest portion of the retina and peripheral retina is the
youngest. To determine whether the antigen gradient is a polar
gradient that rotates around the center of the retina with uni-
form antigen concentration along any arc from center to pe-
riphery or is a circumferential gradient extending from dorso-
posterior to ventroanterior retina, left retinas of 14-day chicken
embryos were cut into eight central and eight peripheral sec-
tions (Fig. 4A) which were assayed for TOP. A 35-fold gradient
of antigen was found extending from dorsoposterior to ventroan-
terior margins of the retina aligned parallel to the long axis of
the choroid fissure.

As shown in Fig. 4B, strips of retina extending from the dor-
soposterior to ventroanterior margins, or perpendicular to this
axis from anterior to posterior margins, were removed and each
Biochemistry: Trisler e¢ al.

 

      

 

ia
Zz
Z © QUTER i
of 8
& &
1) a
é E
a
8 z
3 3
8 ! o
a =
7
e 7
270 J
Q1r27345 678 g 5 10 50
= RETINA SECTION = ” PERCENT OF MAXIMAL
DISTANCE

Fic. 4. Geometry of the TOP gradient in 14-day chicken embryo
retina. Specifically bound !”51-F(ab’), (pmol/mg of protein) is shown
on the ordinate in A and B and within the appropriate segment of ret-
ina tested in A. (A) Each retina (left eye) was cut into eight 45° sections
(7.25 mm in length from the center to periphery ofretina) and each was
divided into central (4.9 mm) and outer (2.35 mm) segments. (B) Dem-
onstration that TOP concentration detected depends on the square of
distance from the ventroanterior margin of the retina. Percentage of
maximal circumferential distance is shown on the abscissa; 100% cor-
responds to 14.5 mm. 4, Strips of retina from ventroanterior (0%) to
dorsoposterior (100%) retina margins, 14.5 x 2.5 mm, were removed
from eight retinas (left eyes), and each was cut into nine segments (1.6
x 2.5 mm) as shown; each segment was assayed for TOP antigen. 5,
Strips of retina from dorsoanterior (0%) to posterior (100%) margins
of the retina perpendicular to the choroid fissure were prepared and
assayed as above. ©, Data from A. The length of the arc from the ventral
pole of the gradient to the center of each segment was calculated by
assuming the retina to be a hemisphere and using equations for spher-
ical triangles.

was cut into nine pieces which were assayed for antigen. The
concentration of TOP molecules detected varied continuously
and logarithmically with the logarithm of distance along the
circumference of the retina from the ventroanterior pole of the
gradient to the dorsoposterior pole, with a slope of 2. In con-
trast, little or no change was detected in retina cells along a
perpendicular axis from anterior to posterior margins of the ret-
ina. The data are described somewhat better by a power func-
tion than by a logarithmic function, but a logarithmic function
has not been ruled out.

 

Fic. 5. Autoradiographs of 14-day chicken embryo retina. (A and B) D
silver grains over cell soma in the inner nuclear layer appear dim due to s

Proc. Natl. Acad. Sci. USA 78 (1981) 2147

The concentration of [!25-F(ab’),‘anti-TOP antibody-TOP
antigen] complex detected is described by the relationship:

[TOP complex] = distance’.

Thus, cell position along a ventroanterior—dorsoposterior axis
of retina can be identified by the concentration of TOP detected:

in which [TOP complex] is the fraction of maximal pmol of !”I-
F(ab’), specifically bound per mg of protein, F,/F yy, and dis-
tance is the fraction of maximal circumferential distance from
the ventroanterior to the dorsoposterior poles of the gradient,
D,/Dmax: Under the conditions used, Finx is 20 pmol of 125]
F(ab’), bound specifically per mg of protein and D,,,, is 14.5
mm. Thus, the calculated mean position in retina of cells that
bind 5 pmol !”I-F(ab’), specifically per mg protein is 7.25 mm
from the ventroanterior pole of the gradient, which agrees well
with the experimental values.

Autoradiography and Immunofluorescence. Autoradiogra-
phy revealed most antigen in dorsoposterior chicken embryo
retina in the inner and outer synaptic layers of the retina (Fig.
5 A and B). The antigen was detected in lesser amounts on the
soma of most, or all, cell types in dorsoposterior retina. Little
antigen was detected in ventroanterior retina (Fig. 5 C and D).
Similar results were obtained by immunofluorescence (not
shown). All cells mechanically dissociated from 8-day chicken
embryo dorsoposterior retina exhibited punctate ring fluores-
cence. All cells from middle retina also were fluorescent, but
less intensely than dorsoposterior retina cells. At each location,
no obvious heterogeneity in cell population was seen. No flu-
orescent cells were detected in ventroanterior retina, although
low levels of TOP were detected in ventroanterior retina by'”I-
F(ab’), binding.

Expression of Antigen During Development. The concen-
tration of TOP antigen detected was higher in dorsoposterior
retina than in ventroanterior retina at every age tested from the
4-day embryo through the adult (Fig. 6A Inset), and the axis and
polarity of the gradient did not change during development
(Fig. 6A). The concentration of antigen detected in the dorsal
half of the retina increased 3-fold between the 4th and 12th days
of embryonic development and then decreased slightly the
adult. Antigen concentration detected in ventral retina did not
vary with development. The amount of protein per retina and

 

ark-field (A) and phase-contrast (B) views of dorsal retina. In A, some
taining of cells by toluidine blue. (C and D) Dark-field (C) and phase-

contrast (D) views of ventral retina. R, photoreceptor layer; OS, outer synaptic layer; IN, inner nuclear layer; IS, inner synaptic layer; G, ganglion
cell layer; A, ganglion cell axon layer. (x 630.) Outer halves of 14-day chicken embryo dorsal or ventral retina were immersed in liquid Freon and

then in liquid nitrogen and were cut in sections 16 wm thick. Each se
fluid diluted 1:50 with solution B) for 45 min at 4°C, washed six times
B containing 20 yg of fluorescein conjugate of rabbit IgG anti-mouse I
each section was incubated with 50 yl of solution B containing 440 mM

ction was incubated with antibody to TOP or P3X63 Ag8 antibody (ascites
(15 min) with solution B (50 yl per wash), incubated with 50 yl of solution
gG (Cappel) for 45 min at 4°C, and washed as above. For autoradiography,
125]. F(ab’), (L x 10° cpm) and 500 yg of bovine serum albumin for 30 min

at 4°C, washed as above, and coated with NTB-2 nuclear track emulsion (Kodak). Slides were exposed in the dark (22 days) at 4°C in the presence

of a dessicant and stained with 0.02% toluidine blue.
 

    
 
   

  

 

 

 

2148 Biochemistry: Trisler et al.

z ,

2 | rd T T T T

Q A. TOPOGRAPHIC GRADIENT | 2 B. ANTIGEN AND

x VS. DEVELOPMENTAL AGE ti PROTEIN/RETINA

2 20b 16 = 100 100
E g 1251.Ftab’

Bo s [-Fiab’}o

z fo}

3 2 <
© 4 a 10 2
>

z 3 «
= G PROTEIN i
5 10 2 1.0 WF
f S &
a Ba a
a) 2 z
5 z

: g oO 0.1
8 4

2 $

2 a 0.01 L l \ iy10.01
z 4 8 12 16 20 ADULT

DAYS AFTER FERTILIZATION

Fic. 6. TOP antigen in chick retina as a function of developmental
age. (A) Each retina (left eye) was cut into eight sections as shown in
Fig. 2. Symbols and days in ovo are: 0, 8; A, 10;5, 12; , 14; v, 16;
@, 18; and m, adult. (Inset) }°1-F(ab’), bound specifically is shown on
the ordinate; days in ovo and adult (AD) are shown on the abscissa.
©, dorsal half of retina (sections 3-6); a, ventral half of retina (sections
1, 2, 7, and 8). Data for retina from 4- and 6-day embryos are shown
only in Inset. (B) 0, '*I-F(ab’), bound specifically per retina; 4, protein
per retina; @, ordinate represents pmol **I-F(ab’), bound specifically
per mg of protein.

TOP antigen detected per retina increased 470- and 620-fold,
respectively, between the 4-day embryo and the adult; the
amount of antigen detected per mg of protein remained rela-
tively constant (Fig. 6B). These results show that a gradient of
TOP molecules is formed early in retina development, during
the period of active neuroblast proliferation and neuron genesis,
and that the gradient is maintained after neuron genesis ceases.

Tissue Specificity. Highest concentrations of TOP antigen
detected were in regions of the nervous system derived from
prosencephalon (forebrain): retina > cerebrum > thalamus
(Table 1). Low levels of antigen were found in dorsal and ventral
retina pigment epithelium, optic nerve, optic tectum, and cer-
ebellum; little or no antigen was detected in heart, liver, kid-
ney, or cells from blood.

Chicken Embryo with an Ectopic Eye. During the course
of these studies, a chicken embryo with three eyes was found
(Fig. 7). The third eye was situated in the middle of the fore-
head, facing in a dorsoanterior direction. Retinas from right,
middle, and left eyes contained gradients of TOP molecules
with normal polarity and alignment with the choroid fissure.

Table 1. Distribution of TOP antigen in chicken tissues

'25].F(ab’), specifically
bound, pmol/mg protein

 

14-day
Tissue embryo Adult

Dorsal neural retina 12.0 7.70
Ventral neural retina 1.08 2.70
Cerebrum 3.30 3.12
Thalamus 2.13 _
Optic nerve _— 0.58
Optic tectum 0.15 _
Cerebellum 0.23 0.46
Dorsal retina pigment epithelium 0.26 _
Ventral retina pigment epithelium 0.25 _
Heart, liver, kidney, or blood cells 0.002-0.055

 

Proc. Natl. Acad. Sci. USA 78 (1981)

 

3 ToT
A.

  

 

BEAKS

 

#MOL 1251-F(ab'}2 BOUND/RETINA SECTION

 

Poppy
12345 67 8
RETINA SECTION

Fic. 7. (A) TOP antigen gradients in retinas from the right (0),
middle (2), and left (a) eyes of a 14-day chicken embryo with three eyes,
Total ‘*°1-F(ab'), bound per retina section is shown. Reaction mixtures
contained 2.38 nM '°1-F(ab’)s. (B) Frontal view of head of embryo. The
third eye is situated on the forehead facing in a dorsoanterior direction.
The embryo had two pairs of beaks, two brains in one head, and one
body.

Thus, a gradient of TOP antigen was generated with normal
orientation in the supernumerary eye despite the abnormal
orientation of the eye in the embryo.

Species Specificity. Gradients of TOP molecules with similar
opientation and symmetry were detected in turkey, quail, and
duck embryo retina, 17, 15, and 16 days after fertilization, re-
spectively (Fig. 8). '°I-F(ab’), concentrations in reaction mix-
tures were low (0.074-0.26 nM); thus, bound !*]-F(ab’), also
was low. With 440 nM ’”°1-F(ab’),, 7.5 pmol of !*I-F(ab’),
bound specifically per mg of dorsoposterior quail retina protein,
comparable to chicken retina. The antigen was not detected in
retina of goldfish, Xenopus laevis, Rana pipiens, or Fisher rats
(not shown).

Properties of TOP Antigen. A dorsal—ventral gradient of
GM, ganglioside molecules (II’NeuAc-G,Ose,Cer) in chicken
embryo retina has been postulated but not detected (15). Bovine
brain gangliosides (10-10,000 4M) did not inhibit the binding
of anti-TOP antibody to retina (not shown). Rabbit antisera to
mono-, di-, and trisialogangliosides (gift of C. Alving) bound to
retina; however, a ganglioside gradient was not detected.

 

   

12 T T T T T T T T
TURKEY

3

@

Q2

rar

Be

Se

a

ge

get

gs

- >

7 Oo 4

oo

x

=

a

oO

0 Ly
12345 67 8

 

 

RETINA SECTION

Fic. 8. TOP gradients in retina from Japanese quail (0), Coturnix
coturnix japonica (15-day embryo); White Pekin duck (a), Anas pla-
tyrhynchos (16-day embryo); and turkey (9), Meleagris gallopavo (17-
day embryo). The eggs hatch 17, 28, and 28 days after fertilization,
respectively. }”°I-F(ab’), concentrations and ,.Ci/pmol were as follows:
quail, 0.26 nM, 0.90 wCi/pmol; duck, 0.074 nM, 2.92 uCi/pmol; and
turkey, 0.078 nM, 2.38 uCi/pmol.
Biochemistry: Trisler et al.

Table 2. Effect of trypsin or heat on TOP antigenicity

 

 

 

1251 -F(ab‘)o
bound
specifically
to retina
Treatment of retina cells cells
Exp. 0-30 min 30—40 min cpm %
1 Control + Trypsin inhibitor 1700 100
Trypsin + Trypsin inhibitor 93 «6
Trypsin + trypsin
inhibitor — 1803 106
2 4°C, 30 min 1607 100
100°C, 30 min 1328

 

Inactivation of TOP retina molecules in 14-day chicken embryo dor-
gal retina by trypsin or heat. In Exp. 1, retina cells were incubated for
30 min at 37°C, in phosphate-buffered saline alone or with 11 »M tryp-
sin (crystallized three times, Worthington) or with 11 »M trypsin in-
activated with 12 uM soybean trypsin inhibitor (Worthington), and
then for 10 min in soybean trypsin inhibitor. TOP ascites fluid was
diluted 1:1000; 1.86 nM !”5]-F(ab’), (90.5 nCi/pmol) was used. In Exp.
2, TOP and P3X63 Ag8 antibodies were diluted 1:100; 1.82 nM 1257.
F(ab’), (96.9 nCi/pmol) was used.

TOP antigenicity was upon incubation at 100°C or by incu-
bation with trypsin (Table 2). All TOP antigenicity in retina cell
homogenates was recovered from the 100,000 x g particulate
fraction; soluble antigen was not detected (not shown).

DISCUSSION

Fusion of spleen cells from a mouse immunized with dorso-
posterior chicken embryo retina with P3X63 Ag8 mouse my-
eloma cells yielded a line of hybridoma cells that synthesizes
antibody to molecules that are distributed in a topographic gra-
dient in the retina. At least 35-fold more antigen was detected
in dorsoposterior than in ventroanterior retina. The antigen was
found by autoradiography in highest concentration in 14-day
chicken embryo retina on neurites in the inner and outer syn-
aptic layers. The antigen was found by immunofluorescence on
most, or all, cells from dorsoposterior and middle portions of
8- and 14-day chicken embryo retina. These results suggest that
antigen molecules are distributed in the retina on the basis of
cell position, rather than cell type.

Neurons in dorsal and ventral retina differ in several ways.
For example, Xenopus retina is composed ofat least three clonal
domains (16). One cell on each side of the 16-cell embryo gives
rise to dorsal retina, another cell gives rise to middle retina, and
a third cell gives rise to cells that migrate across the midline of
the embryo and form ventral retina on the opposite side. In
addition, dorsal retina ganglion neurons synapse in ventral tec-
tum, whereas ventral retina ganglion neurons synapse in dorsal
tectum (1, 17, 18), establishing a continuous, point-to-point re-
tino-tectal map.

Neurons are generated in chicken embryo retina between the
Qnd and 12th days after fertilization; central retina is the oldest
portion of the retina, and peripheral retina is the youngest. The
concentration of TOP molecules detected in dorsoposterior ret-
ina increased as the diameter of the retina increased, and it
varied with cell position at every stage tested, from the 4-day
embryo through the adult, suggesting that the gradient of TOP
is established as neurons and glia (and possibly their precursors)
are generated in retina. Proliferation of clonal populations of
cells with different genotypes in mosaic mice results in radial
patterns of cells in neural retina and retina pigment epithelium
(19); whether the gradient of TOP molecules is due to clonal
inheritance remains to be determined.

Proc. Natl. Acad. Sci. USA 78 (1981) 2149

Dorsal and ventral retina cells also differ in adhesive speci-
ficity—i.e., cells from dorsal retina adhere preferentially to cells
from ventral retina or tectum and vice versa (20-22). However,
the properties of TOP molecules differ from those reported for
preferential adhesion of dorsal retina cells and for retina adhe-
sion factors such as cognin (4), CAM (5), and ligand and agglu-
tinin (23).

The dorsoposterior portion of chicken or pigeon retina con-
tains a fovea, 3- to 5-fold more amacrine synapses than ventral
retina (24), and a relatively high concentration of red droplets
in photoreceptor cells and functions as a binocular visual field
for pecking (25). Thus, dorsoposterior retina differs from other
portions of retina in embryologic development, migration of
cells and axons across the midline of the embryo, synapse spec-
ificity, adhesive specificity, and function.

The results suggest that a gradient of TOP molecules is
formed by a gradient of cells which have different numbers of
antigenic TOP molecules depending on cell position in the ret-
ina. No evidence was found for topographically distributed dif-
ferences in antigen affinity for antibody or in the proportion of
cells expressing antigen. Variation in antigen accessibility has
not been excluded. The mechanism of generating and main-
taining the gradient of TOP, highly ordered with respect to the
axis of the retina, remains to be determined.

The function of TOP molecules has not been determined.
However, since TOP antigen concentration detected is contin-
uously graded and distributed on the basis of cell position not
cell type, our working hypothesis is that TOP molecules play
a role in the coding of positional information in the retina.

Sperry, R. W. (1963) Proc. Natl. Acad. Sci. USA 50, 703-710.

Silver, J. & Sidman, R. L. (1980) J. Comp. Neurol. 189, 101-111.

Goldschneider, I. & Moscona, A. A. (1972) J. Cell Biol. 53,

435-449.

4. Hausman, R. E. & Moscona, A. A. (1979) Exp. Cell Res. 119,
191-204.

5. Thiery, J.-P., Brackenbury, R., Rutishauser, U. & Edelman, G.
M. (1977) J. Biol. Chem. 252, 6841-6845.

6. Eisenbarth, G. S., Walsh, F. S. & Nirenberg, M. (1979) Proc.
Natl. Acad. Sci. USA 76, 4913-4917.

7. Barnstable, C. J. (1980) Nature (London) 286, 231-235.

8. Trisler, G. D., Donlon, M. A., Shain, W. G. & Coon, H. G.
(1979) Fed. Proc. Fed. Am. Soc. Exp. Biol. 38, 2368-2373.

9. Kohler, G. & Milstein, C. (1975) Nature (London) 256, 495-497.

10. Galfre, G., Howe, S. C., Milstein, C., Butcher, G. W. & How-
ard, J. C. (1977) Nature (London) 266, 550-352.

ll. Berzofsky, J. A., Hicks, G., Fedorko, J. & Minna, J. (1980) J.
Biol. Chem. 255, 11188-11191.

12. Schneider, M. D., & Eisenbarth, G. S. (1979) J. Immunol. Meth-
ods 29, 331-342.

13. Bolton, A. E. & Hunter, W. M. (1973) Biochem. J. 133, 529-539.

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

15. Marchase, R. B. (1977) J. Cell Biol. 75, 237-257.

16. Hirose, G. & Jacobson, M. (1979) Dev. Biol. 71, 191-202.

17. Hamdi, F. A. & Whitteridge, D. (1954) Quant. J. Exp. Physiol.
39, 111-119.

18. McGill, J. L, Powell, T. P. S. & Cowan, W. M. (1966) J. Anat.

100, 5-33

orpr

-19. Sanyal, S. & Zeilmaker, G. M. (1977) Nature (London) 265,

731-733.

20. Barbera, A. J., Marchase, R. B. & Roth, S. (1973) Proc. Natl.
Acad. Sci. USA 70, 2482-2486.

21. Gottlieb, D. L., Rock, K. & Glaser, L. (1976) Proc. Natl. Acad.
Sci. USA 73, 410-414.

92. Cafferata, R., Panosian, J. & Bordley, G. (1979) Dev. Biol. 69,
108-117.

93. Rutz, R. & Lilien, J. (1979) J. Cell Sci. 36, 323-342.

_ Yazulla, §. (1974) J. Comp. Neurol. 153, 309-324.

95. Galifret, Y. (1968) Zeitschrift fur Zellforschung 86, 535-545.

Lo]
ns