The Spatial and Temporal Expression of Outer Segment Proteins during Development of Macaca Monkey Cones

Abstract

purpose. To characterize the spatial and temporal expression of key structural and phototransduction cascade proteins in the monkey cone outer segment (OS).

methods. Retinas from Macaca monkeys from ages fetal day (Fd) 89 through adulthood were double labeled using immunofluorescence for short (S) or long/medium (L/M) wavelength–sensitive cone opsin and either a structural protein (peripherin) or a phototransduction cascade protein (α-transducin [α-T], phosphodiesterase [PDE], or rhodopsin kinase [RK]). The spatial and temporal patterns of expression for each protein at each age were determined and graphed as a percentage of retinal coverage.

results. In both cone types, opsins and phototransduction proteins appear first in the fovea and last at the retinal edge. Peripherin appears concomitantly with opsin in both S and L/M cones, but S cones express peripherin and opsin 1 to 3 weeks before neighboring L/M cones. α-T, PDE, and RK are expressed together in the L/M cone OS shortly after L/M opsin appears. Phototransduction proteins are not expressed in S cones until 1 to 3 weeks after the appearance of S opsin and at the same time that neighboring cones are expressing both L/M opsin and phototransduction proteins.

conclusions. The concomitant appearance of opsin and peripherin strongly suggests roles in promoting the structural integrity of the developing OS. Phototransduction cascade proteins appear in the developing OS at the same time as one another, but after opsin. The significant lag between their expression and that of S cone opsin indicates that phototransduction proteins are not essential for OS formation, nor does opsin expression trigger their expression. The different temporal but similar spatial expression patterns of phototransduction proteins within S and L/M cones suggests that some local signal(s) coordinates their appearance.

Phototransduction is the process by which retinal photoreceptors translate light absorption by photopigments into a neural signal. These photopigments are localized as integral proteins in the infolded membranes forming the outer segment (OS) of the photoreceptor and are composed of 11-cis-retinaldehyde and an opsin protein specific for each photoreceptor type. Each opsin protein, in combination with photoreceptor-specific inner retinal synaptic circuitry, confers specific properties on each photoreceptor system. Rods provide achromatic vision and are able to do so even at very low levels of light. Cones function optimally in bright light and provide high acuity and color vision because of the relative sensitivities of their opsin proteins to long (L), medium (M), or short (S) wavelengths of light. ^{1 2} It has not yet been possible to make an antibody that distinguishes between L and M cone opsins, because of their highly similar amino acid sequences; thus, these cones will be designated L/M cones in this article. The markedly different sequence of S cone opsin has allowed several groups to produce antibodies that distinguish S from L/M cones. ^{3 4 5}  

Although light absorption by photopigments is the first step in phototransduction, a complex cascade involving other OS proteins is essential for phototransduction to be translated into membrane potential changes. ^{2 6 7 8} Light-activated photopigment molecules stimulate the active α-subunit of transducin (α-T), which then enables phosphodiesterase (PDE) to cleave cyclic guanosine 3′,5′-monophosphate (cGMP). This causes cGMP-gated cation channels to close and hyperpolarizes the cell, leading to a reduction in synaptic activity. Re-establishment of the dark current is initiated by the phosphorylation of light-activated opsin by rhodopsin kinase (RK), enhancing the binding and inactivation of opsin by arrestin. Guanylyl cyclase activating protein (GCAP) stimulates guanylyl cyclase to regenerate cGMP, reopening the cation channels. 

Structural proteins are also essential for phototransduction as they maintain intact photoreceptor OS. The proteins peripherin and ROM-1 form protein complexes in the rim regions of the OS that are thought to stabilize these sharp bends in the relatively fluid membrane. ^{9 10 11 12 13 14 15 16} Mutations both of opsin and of peripherin are associated with various photoreceptor degenerative diseases, ^{17 18 19} providing evidence that opsin may also have an important role in maintaining the structural integrity of photoreceptors. 

Information on how and when phototransduction and structural proteins are added to the OS during development is surprisingly scarce in the literature. The only full report in rats shows that the appearance of PDE, peripherin, and a cation channel protein are detectable in rat rod OS on P7d, several days after opsin is found. ²⁰ However, studies of protein expression during development of primate photoreceptors have focused mainly on opsin. ^{21 22 23} Opsin is expressed in a central-to-peripheral manner beginning in and around the fovea between fetal day (Fd)65 for rods and Fd75 for cones. S cones express their opsin before L/M cones across most of the retina, ²³ but all opsins are present at the retinal edge by birth, which occurs around Fd168. In this study we used double-label immunofluorescence to compare cone opsin expression with the spatial and temporal expression patterns of the structural protein peripherin and three phototransduction cascade proteins: α-T, PDE, and RK, throughout fetal Macaca monkey retinal development. 

Materials and Methods

Tissue Preparation

Fifteen retinas of Macaca monkeys from Fd89 to birth and three from adult monkeys were obtained from either the Regional Primate Research Center at the University of Washington or the Indonesian Primate Center in Bogor, Java. The care and handling of all animals was in conjunction with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and all protocols were approved by the University of Washington Animal Care Committee. Fetuses were delivered by cesarean section under surgical anesthesia and then were given a lethal dose of barbiturate. The eyes were enucleated, the cornea and lens removed, and the posterior globe immersion fixed in 0.1 M phosphate-buffered 4% paraformaldehyde (pH 7.4) for 8 to 16 hours in the refrigerator. The entire horizontal meridian was cryoprotected and serially sectioned at 10 μm onto subbed slides and stored at −20°C in tightly sealed boxes. Every 10th section was stained with cresyl violet to locate retinal landmarks. 

Immunocytochemistry

Frozen sections were thawed for 4 minutes on a 37°C slide-warming plate. Nonspecific binding was blocked by incubating the sections for 60 minutes on a shaker at room temperature with 10% normal goat serum in Tris-buffered saline (TBS) containing 0.1% Triton X-100. The sections were then incubated for 24 hours on a shaker at 4°C in a mixture of primary antibodies (Table 1) diluted in 0.1 M TBS containing 1% normal goat serum (diluent). To provide localization to a single photoreceptor, one of the polyclonal (poly) cone opsin antibodies was mixed with a monoclonal antibody (mAb), or the mAb OS2 to S opsin was mixed with PDE. The sections were washed overnight in TBS-0.1% Triton X-100 and then incubated in biotinylated goat anti-rabbit IgG (1:100 in diluent) for 45 minutes at 37°C, washed in TBS-0.1% Triton X-100 twice for 30 minutes each, incubated in a mixture of avidin Texas red (1:1000 in diluent) and goat anti-mouse IgG fluorescein isothiocyanate (FITC, 1:100 in diluent) for 45 minutes at 37°C, thoroughly washed, and coverslipped (VectaShield; Vector, Burlingame, CA). 

Before starting this study of fetal retina, it was necessary to determine the range or specificity of each antibody (the degree to which an antibody labels only the protein being studied) and the sensitivity (relative ability of an antibody to detect a low level of its antigen). First, we used antibodies (Table 1) that are well characterized in the literature and that have good sensitivity and specificity in detecting OS proteins in mature photoreceptors. Although there has been no previous characterization of some of these antibodies in the developing retina, it is reasonable to assume comparable sensitivity and specificity in fetal retina. Second, we performed a dilution series of these antibodies in Macaca retina to confirm labeling patterns and optimal antibody concentrations. Fetal retinas then were tested with a narrower range around each optimal concentration. The antibody concentrations listed in Table 1 are those that were found to give the highest specific signal-to-background labeling in fetal retina. Third, no attempt was made to determine the quantity of protein present. Although it seems reasonable that increased immunocytochemical labeling intensity over time at the same antibody concentration suggests an increase, other factors make quantitative interpretation difficult. Rather, in a given microscope field, the protein was either present, defined as detection of a minimal level of specific immunofluorescence, or absent. Finally, for each antibody, at least 10 sections per retina were stained and analyzed. This large number decreased individual variability and resulted in highly reproducible data. 

Data Analysis

The presence of proteins within the same OS was determined by rapidly switching between Texas red and FITC filters. Single and double labeling was also documented using double-exposure photography or a filter that allowed Texas red and FITC to be viewed at the same time. 

Because the monkey horizontal meridian doubles in length between Fd89 and Fd168, ²⁸ measurement of the extent of protein expression is difficult to compare across ages. In prior studies from this laboratory, ^{21 23} we have reported developmental data as the percentage of retinal coverage to simplify temporal comparisons. Retinal coverage is determined in sections cut along the horizontal meridian, which includes the fovea and optic disc. All proteins in this study appeared first in or around the fovea and last at the edge of the retina. First, the total number of ×40 microscopic fields is counted from the temporal to nasal edge of each retina. Then, starting at the fovea, the number of ×40 microscopic fields containing any positively labeled cone OSs is counted. The number of fields containing labeled cells is divided by total fields to arrive at percentage of retinal coverage. For example, 10% retinal coverage indicates labeling of only the fovea and nearby surrounding retina, whereas 90% retinal coverage indicates that the protein is found across most of the retina and is approaching its peripheral edges (see graphs in Figs. 2 6 and 8 ). 

Results

Adult Photoreceptors

Immunoreactivity (IR) with all antibodies listed in Table 1 was confined to photoreceptors. Although the phototransduction proteins in this study are thought to have a function concentrated in the OS, staining was also found in other photoreceptor regions. The OS remained stained even at the most dilute antibody concentration used, whereas other cellular labeling was markedly reduced or lost, suggesting that these proteins have the highest concentration in the adult OS but are not confined to it. 

Opsins.

The antibody poly L/M intensely labeled the OS and lightly labeled the inner segment and cell body of most cones (Fig. 1A ). A much smaller cone population was identified by either the polyclonal S opsin JH455 or the mAb OS2. JH455 heavily labeled the cone OS and lightly labeled the entire cell membrane (Fig. 1B) at dilutions up to 1:20,000. OS2 labeled only the OSs of the same population of cones (not shown). 

Structural Proteins.

As visualized by the monoclonal antibodies 3B6 and 5H2, peripherin IR was limited to the OSs of rods and cones (Fig. 1C , arrow). Double labeling with JH455 and poly L/M showed equal staining of S and L/M cone OSs. 

Phototransduction Cascade Proteins.

Labeling by mAb A1.1 to α-T was cone specific and most intense in the OSs but was also heavy in inner segments (ISs), cell bodies, and synaptic terminals (Fig. 1D) . RK, detected by the mAb D11, heavily labeled the OSs and ISs of all cones and lightly stained the cone synaptic terminals (Fig. 1E) . Double labeling with JH455 or poly L/M showed that both mAbs labeled S and L/M cones with the same cellular distribution. PDE 73-87, which detects the γ-subunit of PDE, intensely labeled cone OSs while also lightly labeling the ISs, cell bodies, and synaptic terminals. Rod OSs were also lightly labeled (Fig. 1F) . All cones labeled by PDE 73-87 were also double labeled with mAb A1.1 or OS2, showing that S cones and L/M cones both contained PDE. 

Fetal Photoreceptors

General Developmental Patterns.

All proteins appeared first in photoreceptors within or around the fovea, and then appeared sequentially in more peripheral retina with increasing age. The general pattern of fetal OS and cell body staining was similar to that in the adult. For instance, in the fetal cone,α -T and PDE stained both cell body and developing OS, whereas peripherin labeling was never seen outside the OS at any developmental stage. 

Because opsin was used as the canonical marker, it was important to define its developmental pattern with the retinas and antibodies used in this study. Percentage of retinal coverage for S and L/M opsin (Figs. 2A 2B ) confirms our earlier findings ²³ that S cones express their opsin well before L/M cones at all developmental ages. This large temporal difference in expression cannot be due to a differential sensitivity of opsin antibodies because the same temporal pattern is found for both mAb OS2 and poly JH455. In addition, S opsin mRNA is expressed before L/M opsin mRNA with the same overall spatial pattern but on a slightly earlier temporal scale. ²³  

Structural Proteins.

In older fetuses, there was a large difference in the temporal appearance of S and L/M opsin, with S cones having over 90% coverage by Fd115, whereas L/M cone opsin expression in the same retina had not yet reached 50% coverage. Despite this difference, there was a tight spatial fit between opsin IR and peripherin IR that held true at all ages and in both cone types. Every cone OS that was positively labeled for L/M opsin (Fig. 2A) and S opsin (Figs. 2B 3A 3B 3C) was also positively labeled for peripherin. No cones were seen that labeled for opsin but not for peripherin, and the reverse was also true. The central-to-peripheral change of peripherin expression is illustrated from an Fd108 retina in Figure 3 . Rod, S cone, and L/M cone OS were all positively labeled for peripherin in the central retina (Fig. 3B) , with two S cone OS (Fig. 3A) double-labeled (Fig. 3B , arrows). Note that in fetal cones peripherin is confined to the developing OS, whereas opsin also labels the entire cell membrane. In the far periphery of the same section where only S cones contain opsin (Fig. 3C) , peripherin IR was confined to S cone OS (Fig. 3D , arrow). 

Phototransduction Proteins.

mAb A1.1 to α-T and mAb D11 to RK were double-labeled against poly L/M (Figs. 4A 4B 4C 4D) . PDE expression in L/M cones was studied by staining an adjacent section with PDE 73-87. In central Fd108 retina, most cones were double labeled for L/M opsin (Figs. 4A 4C) andα -T (Fig. 4B) or RK (Fig. 4D) . In central Fd113 fovea, all cone OSs were labeled for PDE (Fig. 5B ); one S cone was double labeled by mAb OS2 (Fig. 5A 5s) and PDE 73-87 (Fig. 5B 5s) . PDE was the first phototransduction protein to appear after L/M opsin and was present at Fd89 only in central cones (Fig. 6A 6B ). By Fd101, α-T, RK, and PDE also were expressed within central L/M cones (Fig. 6A) , and by Fd108 to Fd113 most central L/M cones contained phototransduction proteins (Figs. 4A 4B 4C 4D 5A 5B 6A) . With increasing fetal age α-T, RK, and PDE appeared together in more peripheral cones until all three reached the retinal edge by Fd155 (Fig. 6A) . Although there were slight differences between RK, PDE, andα -T at each age after Fd101, no consistent pattern was found for one of these proteins leading the others. 

Strikingly different spatiotemporal expression patterns emerged when the appearance of S opsin was compared with the expression of phototransduction proteins. mAbs A1.1 and D11 were double labeled against polyclonal JH455, and mAb OS2 was double labeled against polyclonal PDE 73-87. As shown graphically in Figures 2B and 6B and photographically in Figures 5A 5C 7A and 7C 7S opsin was detected in cones across much of the retina by Fd115. At Fd108 to Fd113 all the central and midperipheral S cone OSs labeled for PDE (Figs. 5A 5B) ,α -T (Figs. 5C 5D) , and RK (Figs. 7A 7B ). However, the far peripheral S cones in the same sections did not double-label for any of the phototransduction proteins (Fig. 6B) . For instance, in midperipheral Fd108 retina, all L/M cone and three S cone OSs contained both opsin and RK (Figs. 7A 7B arrows), but in the far periphery, although S cones contained opsin and had short OSs (Fig. 7C) , RK could not be detected (Fig. 7D) . By Fd155, RK, PDE, and α-T were present in all cones to the retinal edge (Fig. 6A 6B) , more than a month after S opsin was first detected there. 

To clarify these different expression patterns, the retinal coverage of all three phototransduction proteins in each cone subtype was averaged at the different fetal ages and then compared with S and L/M opsin expression (Fig. 8) . Although S cones expressed opsin long before L/M cones, both cone types first expressed phototransduction proteins in their OSs at the same retinal eccentricity. In other words, S cone expression of phototransduction proteins did not occur until L/M opsin and/or α-T, PDE, or RK were present in neighboring L/M cones. 

Discussion

Postive Evidence for Important Roles of Opsin and Peripherin in the Structural Integrity of the Developing Photoreceptor OS

Opsin is the photoreceptor subtype-specific apoprotein that associates with 11-cis-retinaldehyde to form the visual pigment of the photoreceptor OS. Evidence for a structural role for opsin in the OS membrane is supported by its high concentration in the OSs of rods and cones. ²⁹ Studies of patients with retinitis pigmentosa, many of whom have known mutations of rod opsin, have shown that the OSs undergo disorganization and gradual shortening, presumably due to OS membrane instability ^{19 30 31 32 33 34 35 36} Our current results and prior studies showing the presence of opsin in the initial connecting cilia of photoreceptor OSs ^{21 23} also suggest an important role for opsin in maintaining the structural integrity of the earliest OS membranes. 

Peripherin is the product of the wild-type retinal degeneration slow (rds) gene localized to chromosome 6p. ^{17 18} It associates with ROM-1 to form a multisubunit protein complex in the rim region of the OS, stabilizing the sharp bends in OS disks. ^{11 12 13 14 15 16} Although peripherin^−/− mice fail to form OS, ROM-1^−/− mice form functional OS, ¹⁶ supporting a major role for peripherin in OS disc membrane stability. Retinal degeneration in the rds mouse and some forms of autosomal dominant retinitis pigmentosa are characterized by dysfunctional or absent peripherin protein. ^{17 18 37 38 39} This is the first study to show that peripherin is coexpressed with opsin in the initial stages of OS development. This finding adds strong evidence that peripherin has a necessary role in maintaining OS structural integrity throughout life. 

Simultaneous Expression of α-T, PDE, and RK

The synchronous functioning of the phototransduction cascade requires the presence of all its members ^{2 6 7 8} ; for instance, α-T could do little to facilitate phototransduction if PDE were not present. It thus seems logical that all the cascade proteins would appear in the OS together. Our results showed that α-T, PDE, and RK were expressed in both S and L/M cone OS at approximately the same time as one another, but after opsin. A similar developmental pattern has recently been reported for rat rod OS ²⁰ in which PDE and the cation channel protein both appear on P7d. These similar results for four different cascade proteins in monkey and rat retina strongly suggest that the photoreceptor initiates synthesis of all phototransduction proteins simultaneously. The appearance of opsin in the OS before these phototransduction proteins is probably explained by its additional role as a structural protein. Localization of α-T, PDE, and RK throughout the rest of the photoreceptor, both in the developing and the mature retina, suggests that these proteins have additional functions in other cell activities. 

The Expression Delay between Opsin and Phototransduction Proteins

Because OS function is intimately related to phototransduction, we originally hypothesized that phototransduction proteins would be expressed shortly after the OS began to form. Results from this study and others ²³ that primate S cones express opsin 1 to 3 weeks before L/M cones therefore predict that S cones should express phototransduction proteins well ahead of L/M cones. Our results clearly showed that this was not the case; rather, S cones had a morphologic OS containing opsin and peripherin for 1 to 3 weeks before expression of phototransduction proteins. This strongly indicates that cascade protein expression is not directly tied to opsin expression. 

Because the initiation of phototransduction protein expression in S cones overlaps the onset of both opsin and these proteins in neighboring L/M cones, this suggests the presence of some local coordinating signal(s). This signal could be released from the pigment epithelium, which has been shown to increase opsin content and stimulate normal OS developmental differentiation in Xenopus photoreceptors. ⁴⁰ Another possible outer retinal signal could be the maturation of interphotoreceptor matrix proteins that carry the various forms of vitamin A across the pigment epithelium and interphotoreceptor space and into the photoreceptor. ⁴¹ Alternatively, the signal may be triggered within the L/M cones by the expression of opsin and spread to the S cones through cell-to-cell communication or local diffusion. ^{22 23} Another possibility is that the signal arises from the inner retina through cone synapses onto bipolar cells. ⁴² The onset of synaptic activity in the L/M cone synaptic circuit could initiate a retrograde signal that in turn triggers outer cone development. Whatever form the local signal takes, it induces the simultaneous expression of critical phototransduction proteins in all cone types in a given retinal region. It is likely that this region of retina would then have a functional cone dark current that may enhance correlated firing within the entire cone synaptic chain. Identification of these signals and delineation of the other events in primate retinal development that coordinate the final stages of cone functional capability await further studies. 

 Supported by National Institutes of Health Grant EY-04536, Research to Prevent Blindness (AH), and the Lions’ Sight Conservation Foundation of the Northwest (SS).

 Submitted for publication March 1, 1999; revised September 1, and October 26, 1999; accepted November 8, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Antibodies Used in the Study

Antibody	Antigen	Host	Dilution	Donor	Reference
Poly L/M	L/M cone opsin	Rabbit	1:2,000	J. Saari	4
JH455	S cone opsin	Rabbit	1:10,000	J. Nathans	5
OS2	S cone opsin	Mouse	1:25,000	A. Szel	3
3B6	Peripherin	Mouse	1:20	R. Molday	11 12
5H2	Peripherin	Mouse	1:20	R. Molday	11 12
A1.1	α-T	Mouse	1:50	J. Hurley	4
PDE 73-87	PDE	Rabbit	1:1,000	B. Fung	26
D11	RK	Mouse	1:500	K. Palczewski	24 25

Figure 1.

 

Immunofluorescence staining of adult Macaca monkey retina demonstrating the localization of the antibodies used in this study. All sections are aligned so that the pigment epithelium (PE) and photoreceptor outer segments (OS), inner segments (IS), cell bodies (CB), and synaptic terminals (S) are at the same level. (A) Poly L/M antibody at a dilution of 1:2000 heavily labeled most L/M cone OS. (B) A minority of cone OSs were heavily labeled for S opsin by poly antibody JH455 (1:10,000), and light labeling was seen in the IS and CB. (C) 3B6 (1:20) was one of two mAbs used to label the structural protein peripherin. Labeling was confined to rod and cone (arrow) OSs. (D) mAbs A1.1 to α-T heavily labeled all cone OS, IS, CB, and S at a dilution of 1:50. (E) mAbs D11 (1:500) heavily labeled RK in cone OS and IS and was lighter in the CB and S. (F) PDE was detected by poly PDE 73-87, which labeled rod OSs and all cones at dilutions up to 1:10,000. The cone OS labeled most intensely, but staining was also present in the IS and S.

Figure 2.

 

Graphs of percentage of retinal coverage comparing the spatiotemporal expression of L/M opsin (A) and S opsin (B) with that of peripherin in the fetal monkey retina. The fovea is 0%, and the retinal peripheral edge is 100%. Note that S opsin was consistently present at a more peripheral retinal eccentricity than L/M opsin at all ages up to Fd155. Note also that peripherin retinal coverage exactly matched that of both cone opsins at all ages studied. Both proteins were present in all cones at the retinal edge before birth, which occurs around Fd168.

Figure 3.

 

Double-label immunofluorescence of S opsin (left column) and peripherin (right column) at Fd108. (A, B) In central retina, S opsin labeled the entire cone (A), whereas peripherin (B, arrows) was limited to the outer segment (OS) of the same cells. Peripherin was present in all rod and cone OSs at this age in central retina. (C, D) In the far periphery, an immature S cone (C) had a tiny OS that stained for peripherin (D, arrow). At this eccentricity, both opsin expression and peripherin labeling were limited to S cones.

Figure 4.

 

Double-label immunofluorescence for L/M opsin (A, C), α-T (B), and RK (D) in Fd108 central retina. (A, B) All L/M cones expressed both opsin (A) andα -T (B). Both opsin and α-T were located in the outer segment and throughout the membrane of the entire cone. A single S cone heavily labeled for α-T (B, s) was negative for L/M opsin (A, s). (C, D) In central retina all the L/M cones labeled for opsin (C) also contained RK (D) which was limited mostly to the outer and distal inner segments.

Figure 5.

 

Double-label immunofluorescence of S opsin (A, C), PDE (B), and α-T (D). (A, B) In central retina, one S cone labeled by the mAb OS2 (A, s) had a heavily labeled outer segment, but the cell membrane was only lightly labeled. This outer segment also labeled for PDE (B, s). Surrounding L/M cone outer segments contained PDE, but both L/M and S cone cell bodies were only lightly labeled. (C, D) In Fd113 midperipheral retina, the entire cell membrane and outer segment of one S cone was labeled by JH455 for opsin (C, s) and α-T (D, s). Surrounding L/M cones also contained α-T with the same localization.

Figure 6.

 

Percentage of retinal coverage comparing the expression of L/M opsin (A) and S opsin (B) with the appearance of phototransduction cascade proteins α-T, RK, and PDE in the fetal monkey retina. The fovea is 0%, and the peripheral retinal edge is 100%. The ages are the average of two to three fetuses for each age group except Fd89, which had one fetus. All phototransduction proteins appeared slightly later than L/M opsin and much later than S opsin. PDE was the first protein to label either cone subtype, but generally all three proteins appeared in close sequence across the retina.

Figure 7.

 

Double label immunofluorescence of S opsin (A, C) and RK (B, D) at two eccentricities of the same Fd108 retina section. (A, B) In central retina, three S cones labeled for opsin (A) and had outer segments that also contained RK (B, arrows). Neighboring L/M cones also labeled for RK. (C, D) In peripheral retina, two S cones had heavily labeled tiny outer segments (C), but RK (D) was not detectable at this retinal eccentricity.

Figure 8.

 

Summary of percentage of retinal coverage from fetal monkey retina comparing S and L/M opsin expression with the appearance of phototransduction proteins in each cone subtype. The fovea is 0% and the peripheral retinal edge is 100%. The retinal coverage of phototransduction proteins was obtained by averaging the percentage of coverage of α-T, PDE, and RK at each age. Note that S opsin was expressed well in advance of L/M opsin, but phototransduction protein expression in both cone subtypes occurred at the same retinal eccentricity, shortly after L/M opsin appeared.

The authors thank the Regional Primate Research Center at the University of Washington and the Indonesian Primate Center at Bogor, Java, for their cooperation in providing the retinas used in this study; and Shelly Gollard and Dan Possin for technical assistance in producing the illustrations.