Topographic Arrangement of S-cone Photoreceptors in the Retina of the Diurnal Nile Grass Rat (Arvicanthis niloticus)

Abstract

Purpose.: The retina of Arvicanthis niloticus, a diurnal murine rodent closely related to Rattus (rats) and Mus (mice), contains ≈30% to 35% cones and has several cone-driven functional characteristics found in humans. In this study the organization of these cone photoreceptors was examined, with emphasis on those expressing the S-opsin photopigment (S-cones).

Methods.: Cones were labeled with antibodies against M- and S-opsins. Their topographic arrangement was examined on images of retinal flatmounts using density measures, nearest-neighbor distance, and Voronoi domain analysis. Partial sequencing of the S-opsin DNA was also performed to determine whether this visual pigment was blue/violet or UV sensitive.

Results.: Cone photoreceptors (estimated total population ≈1.450 million) came in two distinct types that express either M/L- or S-opsin. Both types were present across the retinal surface. S-cones (≈7–8% of the total cone population) achieved a higher density in a discrete temporodorsal sector of the retina. The S-cone mosaic was irregular. Finally, S-cones were likely to be UV sensitive, according to genetic analysis.

Conclusions.: The topographic arrangement of cone photoreceptors in the retina of the diurnal Nile grass rat A. niloticus represents a highly pertinent model to improve understanding of the pathologic course of and related therapy for retinal disease involving cones.

Cone photoreceptors in human retinas achieve their highest densities in the fovea. They mediate high-spatial-resolution daylight vision and color discrimination. Their progressive loss in a variety of retinal degenerations induces a central visual field scotoma, which almost inevitably leads to legal blindness (visual acuity of 20/200 or less). Ideally, such diseases would be investigated in animal species with dense cone populations. Although genetically engineered rodless mice have recently become available, ^1,2 ground squirrels, guinea pigs, or tree shrews remain popular models for investigating structure, function, and pathologic features of cone-rich mammalian retinas. Members of the Arvicanthis genus (a genus closely related to Rattus and Mus, but with a diurnal lifestyle) may be a useful alternative. ³ The first immunohistochemical investigations dedicated to the visual system of two geographically distant Arvicanthis species, the Sudanian (A. ansorgei) and the Nile (A. niloticus) grass rats, pointed out the efficiency of preexisting antibodies in these species and established that their retinas contain ≈30% to 35% cones. ^3–5 Additional experiments ^5,6 in A. niloticus showed that its retina displays postreceptoral neural features commonly observed in diurnal mammals and that it has several cone-driven functional characteristics (as assessed with the electroretinogram) found in human retinas (large photopic a-wave amplitudes, photopic hill effect, and critical flicker fusion beyond 60 Hz). In the present study, we determined the distribution of the cone population (more specifically those expressing the S-opsin photopigment) in A. niloticus with the rationale of acquiring normative data against which to compare the effects of experimental manipulations and/or aging. 

Material and Methods

Animals

This study was performed on young adult (2–6 months of age) Nile grass rats (A. niloticus) of both sexes derived from a breeding colony established at the University of Alberta. The animals were raised on a 12:12 light–dark cycle (lights on at 5 AM; ambient temperature 21 ± 1°C; relative humidity ≈50%) and supplied ad libitum with water and standard rodent diet (formula 5001 LabDiet; Nutrition International, Richmond, IN). Experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research and with the guidelines laid down by the NIH (National Institutes of Health) in the United States regarding the care and use of animals for experimental procedures. The University of Alberta Animal Care and Use Committee approved the present work. 

Primary Antibodies

Cone photoreceptors were screened using anti-M/L- and anti-S-opsin polyclonal antibodies (AB5405 and AB5407, working dilution 1:500; both from Chemicon, Temecula, CA;) raised in rabbit against the last 42 and 38 amino acids respectively, at the C-terminus of recombinant human red/green- and blue opsins. ⁷ These antibodies have been reported to label the outer segments (OS) and cell membranes of specific types of cones in human, mouse, and ground squirrel retinas. ^8,9 Their specificity was confirmed in Western blot analyses from Nile grass rat retinal tissue (described later). For double-labeling experiments, the S-opsin AB5407 rabbit polyclonal antibody was replaced with an affinity-purified goat polyclonal antibody raised against a 20-amino-acid synthetic peptide mapping within amino acids 1 to 50 of human blue-sensitive opsin (EFYLFKNISSVGPWDGPQYH; sc-14363; Santa Cruz Biotechnology Inc., Santa Cruz, CA; working dilution 1:200). This antibody has been used to identify S-opsin expressing cones in a range of species. ^10–14 Competition controls where sc-14363 was preincubated with the antigenic peptide ^11,13 yielded no labeling. 

Western Blot Analysis

Freshly dissected retinas (n = 3; 2 months of age) were homogenized in SDS buffer (4% [wt/vol] sodium dodecyl sulfate, 0.13 M Tris, 2% [vol/vol] 2-mercaptoethanol, 20% [vol/vol] glycerol; pH 6.8) with a protease inhibitor cocktail (Complete; Roche Applied Science, Mannheim, Germany). Samples (25 μg protein) were resolved by SDS-PAGE on 8% to 10% acrylamide gels. Proteins were transferred to polyvinylidene fluoride (PVDF) membranes, blocked for 1 hour with 5% nonfat milk diluted in TBS-T (20 mM Tris, 137 mM NaCl, and 0.1% Tween-20; pH 7.6), incubated overnight with anti-M/L-opsin (AB5405) or anti-S-opsin (AB5407) primary antibodies (dilution 1:500 in the blocking solution), washed three times for 10 minutes each in TBS-T and reacted for ≈1 hour with a donkey anti-rabbit IgG, HRP-conjugated enhanced chemiluminescence (ECL) antibody (1:5000 in the blocking solution; NA934; GE Health Care, Little Chalfont, UK). After washing three times for 10 minutes each in TBS-T, the protein bands were visualized using ECL reagent (NEL 103; Perkin Elmer, Wellesley, MA) and an imaging station (Eastman Kodak, Rochester, NY). These antibodies detected major products at ≈39 and ≈37 kDa, respectively corresponding to M/L- and S-opsin (Fig. 1A). 

Immunohistochemistry

Cryosections were collected from 4% paraformaldehyde-fixed retinas (n = 2; 4 months of age) cut serially at 20 μm, parallel to the nasotemporal axis, and mounted on glass slides (Superfrost/Plus; Fisher Scientific, Pittsburgh, PA). After they were extensively washed in PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄-7H₂O, 1.4 mM KH₂PO₄; pH 7.3), the sections were blocked for 2 hours in a medium containing PBS +0.1% Triton X-100+5% nonfat milk, and reacted overnight with anti-M/L- and anti-S-opsin antibodies diluted appropriately in a 1:10 solution of the previous blocking medium. On the following day, the sections were washed with PBS, blocked again for 1 hour and exposed for 2 hours to goat anti-rabbit Alexa488-tagged secondary antibody (Molecular Probes Inc., Eugene, OR) diluted to 1:500 in a 1:10 solution of the blocking medium. In double-staining experiments, donkey anti-rabbit-Alexa594 and donkey anti-goat-Alexa488 (1:500) were used as secondary antibodies. Sections were washed extensively in PBS, covered with an antifade solution (ProLong gold antifade reagent, P36939; Molecular Probes), and coverslipped. Control labeling without primary antibody remained negative. All reactions were run at room temperature. 

Additional retinas (2.5–6 months of age) were prepared as flatmounts. After euthanatization (Euthanyl; Bimeda-MTC Animal Health Inc., Cambridge, ON, Canada), the eyes were enucleated (after a small incision was made in the dorsal margin of the ora serrata for orientation). The cornea and lens were removed. The retinas were then carefully dissected from the eye cup, cut into four distinct quadrants, fixed overnight at 4°C in 4% paraformaldehyde and processed with anti-M/L-opsin (n = 2), anti-S-opsin (n = 2), a mixture of anti-M/L+S-opsin (n = 2) antibodies, as well as (n = 2) with a rabbit anti-γ-transducin antibody ⁵ (dilution 1:1000, PAB-00801G; CytoSignal, Irvine, CA) as described earlier, with the exception that incubations with the primary and secondary antibodies were performed at 4°C and lasted up to 3 days. 

Quantitative Investigations on Flatmounted Retinas

Representative samples were imaged with a laser confocal microscope (LSM 510 Axiovert 100M; Carl Zeiss Meditec, Inc.) or a fluorescence microscope (DM6000 Leica; Deerfield, IL) equipped with a computer controlled motorized stage. Raw images were then processed with the use of several computer programs (ImageJ ver. 1.40g; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD, and available at http://rsb.info.nih.gov/ij/, or Image-Pro ver. 4.0; Media Cybernetics, Silver Spring, MD, for quantitative measurements, and with Photoshop 6.0, Adobe, San Jose, CA, for illustrations). 

The density of M/L- and S-opsin expressing cones was determined from manual counts applied to raw images (area = 0.092 mm²) taken with a 40× oil immersion objective in the center (from 0 to 1250 μm from the optic nerve head [ONH]), in the mid periphery (from 1250 to 2500 μm away from the ONH) and in the far periphery (>2500 μm from the ONH) of each retinal quadrant of the flatmounted retinas (Figs. 2A, 2B). Every labeled cone OS in these frames, which did not touch the borders, was scored and tagged at its base for additional analysis (nearest-neighbor distance and Voronoi domain ^15,16 ) by using the appropriate ImageJ plug-ins. 

A complete density map of the S-cone population (Fig. 2D) was further achieved in a computer-assisted manner by using the Analyze Particles tool of ImageJ software. Threshold and object size values were adjusted for the S-cone population in any previous frames that could be detected accurately, and the process was applied to a mask containing 101 nonoverlapping fields of interest (squares 350 × 350-μm wide; area, 0.1225 mm²; spacing, 200 μm) superimposed on a digital photographic montage of the flatmounted retina obtained by the automated confocal microscope system (single focal plane; 10× objective). Counting variability between manual detection and computer-aided processing, as assessed a posteriori in an additional 10 randomly selected fields of interest, was no more than 5%, which confirms the reliability of the quantitative method used. ¹⁷ Counts were not corrected for possible tissue shrinkage. 

Estimates of cone populations were calculated for an average retinal area of 43 mm² (mean value from flatmounts: 42.85 ± 2.5 mm²; Statview software; SAS Institute Inc, Cary, NC). Significance was set at P = 0.05. All values are given as the mean ± SD. 

Partial Sequencing of the S-opsin Gene

Genomic DNA was prepared from ear notch biopsies of 3 Nile grass rats (all at 3 weeks of age) and one mouse (for control) (DNeasy Blood and Tissue Kit; cat. no. 69504; Qiagen Sciences, Germantown, MD). PCR was performed by using the following primers designed within exon 1 of rat Opn1sw (GenBank accession number NM_031015; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD): 5′-CTGGGATGGGCCTCAGTAC-3′ (forward) and 5′-AGGCCTCCAGAGCACAAAC-3′ (reverse). PCR conditions were as follows: initial denaturation at 94°C for 5 minutes; denaturation at 94°C for 30 seconds, annealing at 60°C for 30 seconds, extension at 72°C for 30 seconds for 30 cycles; and final extension at 72°C for 10 minutes The resulting 275-bp product was visualized on a 2% TAE gel, the band was excised, and the DNA was extracted with a gel extraction kit (Qiaex II, cat. No. 20021; Qiagen Sciences). DNA was eluted in ultrapure water and sequenced in the forward and reverse directions using the primers just shown. Sequencing was performed with genetic analyzer (3130xl; Applied Biosystems, Inc. [ABI], Foster City, CA). Amino acids at positions corresponding to 46 through 114 in human S-opsin (41 through 109 in rat) were examined. 

Results

Cone Types in the Nile Grass Rat Retina

In cryosections of Nile grass rat retinas, antibodies AB5405 (against M/L-opsin) and AB5407 (against S-opsin) stain outer segments (OS) from cone photoreceptors (negative to the anti-rhodopsin monoclonal antibody Rho4D2; data not shown). The double labeling procedure shows further that: (1) OS stained with AB5405 do not label with sc-14363 (Figs. 1B, 1D); and (2) OS stained with AB5407 are also labeled with sc-14363. Complete colocalization of the two labels demonstrates that both antisera recognize the same S-opsin (Figs. 1E–G). Antibodies AB5405 and sc14363 also faintly stain a few cones (≈50 per section) from OS to synaptic pedicles (Figs. 1C, 1D). The reason for such whole cell labeling (which is absent in mouse and rat) is unclear, but likely related to the specificity of these antibodies in A. niloticus. Of note, the commonly used anti-S-opsin antibody JH455 (mapped at the C-terminal of the human S-opsin; Jeremy Nathans; Johns Hopkins University School of Medicine, Baltimore, MD) also stains S-cones entirely in several species. ^13,18–20 The cell bodies of cones are located in the two first rows (outer side) of the outer nuclear layer and have comparable sizes (average vertical diameter: 9.0 ± 1.2 μm for the S-cones, n = 76; 8.4 ± 0.8 μm for the M/L-cones, n = 41). M/L- and S-opsin expressing OS have very different distributions (approximate M/L:S ratio in Fig. 1B is 13–14:1), which suggests that they reflect two distinct cone populations. Close inspection of two flatmounted retinas, processed with a mixture of anti-M/L- and anti-S-opsin antibodies, indeed confirmed that the retina of A. niloticus possesses no dual-pigment cones. This is illustrated in samples taken at the naso-dorsal periphery (Figs. 1H, 1I) and in the ventral quadrant which contains most of the dual-pigment–expressing cones in the mouse (Figs. 1J, 1K). 

Figure 1.

 

Cone distribution in cross sections and flatmounted retinas. (A) Western blots of A. niloticus retinal tissue. AB5405 and AB5407 antibodies detect major products at ≈39 and ≈37 kDa (lanes M and S, respectively) as expected for M/L- and S-opsins. A long exposure time (>4 minutes) may explain the faint bands (at ≈80 and ≈130 kDa) observed with the M/L-opsin antibody. (B) Near-central retinal cross section treated with M/L (AB5405; red)- and S (sc-14363; green)-opsin antibodies. These antibodies stain distinct OS. Pseudo co-localization of both opsins in the leftmost cone is due to accidental superimposition of two OS (the thickness of the sections is 20 μm). (C) M/L-cone entirely labeled with antibody AB5405. (D) S-cone as revealed with antibody sc-14363. (E) Cross section treated with anti-S-opsin antibody AB5407. (F) Same cross section treated with anti-S-opsin antibody sc-14363. (G) Merged pictures show that these two antibodies detected the same S-cone OS. (H) Example of M/L- (red) and S- (green) opsin expressing cones as viewed in a flatmounted retina (nasodorsal periphery). (I) Enlargement of the inset in (H) illustrates that expression of M/L- and S-opsins was mutually exclusive in the cones of A. niloticus. (J, K) Absence of opsin coexpression in an additional double-labeled flatmount retina. Pictures taken in the midperiphery of the ventral quadrant where dual-pigment cones are most evident in the mouse. OS, outer segment; IS, inner segment; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; RPE, retinal pigment epithelium.

Cone Densities in the Nile Grass Rat Retina

M/L-cones.

M/L-cones distribute across the retina of A. niloticus. Their density increased smoothly (1.6-fold) but significantly (P < 0.001, Mann-Whitney U test), from the far periphery to the ONH (Table 1). Apart from the central retina, high cone density was notable in the dorsal part of the temporal quadrant (white rectangles, Fig. 2A; and darkest area, Fig. 2C). Variations in M/L-cone densities between quadrants and between the main retinal axis never reached statistical significance. Given an average packing density of 31,000 ± 5,700/mm² (n = 76), the total number of M/L-cones in A. niloticus retina was estimated at ≈1.350 million. 

Table 1.

 

Cone Densities in the Retina of A. niloticus

	Far Periphery (n/mm2)	Mid Periphery (n/mm2)	Center (n/mm2)	Estimated Total (n)§
Single labeling
M-cones*	23,580 ± 2,010 (n = 14)	35,130 ± 1,470 (n = 8)	37,700 ± 1,620 (n = 8)	≈1,350,000
S-cones*	1,660 ± 240 (n = 10)	2,680 ± 750 (n = 6)	3,200 ± 560 (n = 5)	100,000
S-cones†	1,720 ± 540 (n = 41)	2,600 ± 540 (n = 31)	3,100 ± 590 (n = 20)	100,450
Total S- and M-cones*	25,240	37,810	40,900	≈1,450,000
Ratio S-cones over total cones (%)	6.6	7.1	7.6	6.8
Total Cones (γ-transducin labeling)‡	28,350 ± 1,950 (n = 8)	36,400 ± 2,500 (n = 4)	40,300 ± 2,600 (n = 8)	≈1,495,000
Double labeling
M-cones*	22,820 ± 3,280 (n = 4)		35,810 ± 2,660 (n = 4)	≈1,260,000
S-cones* †	1,760 ± 230 (n = 17)	2,350 ± 240 (n = 15)	3,100 ± 360 (n = 12)	≈104,670
Ratio S-cones over total cones (%)	7.1		8.0	7.7

 (n) indicates the number of sampling frames examined per retinal locations.

 *Manual procedure.

 †Computer-aided procedure.

 ‡Recalculated from Gaillard et al., Figure 1D. ⁵

 §For a 43-mm² retina.

Figure 2.

 

Analysis of the cone photoreceptor populations from flatmounted retinas (same Nile grass rat specimen). (A) M/L-cones; (B) S-cones. Manual counts were performed at ×40 magnification. The rectangular regions are of the size matching the counting frame of 0.092 mm². White rectangles: areas of higher cone density (>36,000/mm² for the M/L-cones and >3,600/mm² for the S-cones, i.e., respectively, 28% and 68% above the average density in the rest of the retina). Cone densities at specific retinal locations are indicated. (C) Isodensity curves for the M/L-cone population (as determined from manual counts over a total of 76 frames). (D) Isodensity curves (modeled from 101 fields counted in ImageJ) optimally illustrate the higher S-cone density in the temporodorsal sector of an A. niloticus retina. A similar map (modeled from 44 fields counted in ImageJ) was obtained with the retina double-labeled for M/L- and S-cones.

S-cones.

As was the case for the M/L-cones, S-cones populated the entire retina and displayed an approximately twofold increase in cell density from periphery (1660 ± 240/mm²) to center (3200 ± 560 /mm²), as deduced in a manual counting approach (Fig. 2B, Table 1). The collected values are considerably lower (12–14-fold) than those obtained for the M/L-cones at similar retinal locations (compare Fig. 2A with 2B). Given an average packing density of 2320 ± 820/mm² (n = 21), the total number of S-cones in the A. niloticus retina was estimated at ≈100,000. 

The computer-assisted counting procedure allowed a more detailed analysis of the S-cone distribution in this retina. Significant differences in S-cone numbers were observed between retinal quadrants (P = 0.0001, ANOVA, F = 7.55, df = 3). S-cones appeared more numerous in the temporal quadrant than in any other quadrant (P ≤ 0.003, Fisher PLSD post hoc test). Therefore, S-cones distributed preferentially along the temporonasal compared with the dorsoventral axis (P = 0.009, one-tailed t-test). Finally, S-cones were denser (3760 ± 345/mm²; n = 12) in a restricted sector of the temporal retina superior to the ONH (Fig. 2D), than elsewhere in the retina (≈2160 ± 625/mm²; n = 89; P = 0.0001, one-tailed t-test). In this sector, S-cones represented ≈10% of the total cone population. 

Estimated Cone Densities in the Double-Labeled Retina

Taking advantage of the fact that the Nile rat retina has no dual-pigment cone, individual M/L- and S-cone distributions were further assessed in one of the double-labeled flatmounted retinas. Counts in this retina (Table 1) gave an average packing density of 29,320 ± 7,400/mm² (leading to a total of ≈1.260 million), in the M/L-cones, and an average packing density of 2,430 ± 890/mm² (leading to a total of 104,500) in the S-cones. The S-cones were also more numerous along the temporonasal axis, peaking ≈2 mm temporally and ≈1 mm superior to the ONH (3910 ± 940/mm²; n = 8 compared with 2130 ± 450/mm²; n = 36 in the rest of the retina). 

Cone Mosaics in the Nile Grass Rat Retina

Collectively, counts of M/L- and S-cones yield ≈1.350 to 1.450 million total cones in A. niloticus, which is in accord with the ≈1.495 million obtained previously ⁵ after retinas were processed with an anti-γ-transducin antibody as a general cone marker. ²¹ Of the total cone population, ≈92% to 93% express M/L-opsin and ≈7% to 8% express S-opsin (Table 1). In addition to these measurements, we examined the spatial organization of the respective M/L- and S-cone populations by computing for each cone the distance with its nearest neighbor (NND) and its Voronoi domain (VD; defining points in the plane that were closer to that cone than to any other cone in the mosaic). VDs intersecting the sampling frame and NNDs of cells closer to the frame borders than to any cells in the sample were discarded from the analysis. 

S-cones.

The regularity of S-cone distribution in the retina of A. niloticus was examined in 15 of the sampling frames initially used for pilot density measurements (Fig. 2B), one of which was located within the temporodorsal cone-rich sector. Depending on their location, these sampling frames (0.092 mm²) contain between 134 and 368 S-cones. On average, the NND was 11.21 μm ± 1.75 (n = 15), the regularity index (RI; ratio of the mean NND to the SD) was 1.94 ± 0.14, and the NND index (Rn; ratio of observed NND to the mean random distance) was 1.04 ± 0.046. The NND in a given sampling frame appeared to be closely related to the cone density in that frame (Figs. 3A, 3B); and the regularity indexes tended to be lower (P = 0.042, one-tailed U test) in the periphery than in the center of the retina. Finally, distributions of the NNDs were similar in each frame (Fig. 3A), and all deviated significantly from a normal distribution (P < 0.01, χ² test of normality). As suggested with the NNDs, Voronoi domains were larger (P = 0.01, one-tailed U test) in the retinal periphery than in the center (Fig. 3C). Corresponding RIs (ratio of the mean Voronoi area to the SD) were low, and did not vary with eccentricity (Table 2). Investigations in the additional, double-labeled retina (Figs. 1J, 1K) provided similar results: both NND and VD regularity indexes were low (respectively, 2.20 ± 0.10, n = 8; and 2.02 ± 0.3, n = 8) and remained stable with eccentricity. Together these findings indicate that the S-cone mosaic of A. niloticus is not regular. 

Figure 3.

 

Cone mosaics in the retina of A. niloticus. (A) Distributions of the NNDs for S-cones taken from fields T3 (periphery; temporal quadrant) and D6 (center; dorsal quadrant). Counting fields (0.092 mm²) contained respectively 179 and 329 S-cones. (B) Distributions of the NNDs for M/L-cones. Sampling areas (0.01 mm²) were taken from fields T2 (periphery; temporal quadrant) and D7 (center; dorsal quadrant) and contained, respectively, 244 and 397 M/L-cones. The Bell-shaped curve in each histogram corresponds to the Gaussian fit. (C) Two examples of Voronoi domains for S-cones from fields T3 and D6. Field positions are given in Figure 2A for the M/L-cones and Figure 2B for the S-cones.

Table 2.

 

Cone Mosaics

	Location	Fields (n)	Cells in Fields (n)	NND (Mean μm)	RI	VD (Mean μm2)	RI
S-cones	Center	5	250–368	9.15 ± 0.67	2.03 ± 0.10	287.00 ± 59.30	1.82 ± 0.08
	Periphery	8	134–208	12.41 ± 0.9	1.84 ± 0.14	528.65 ± 62.20	1.82 ± 0.18
M/L-cones	Center	5	330–392	3.54 ± 0.20	4.67 ± 0.33	20.50 ± 1.90	3.84 ± 0.42
	Periphery	7	224–280	4.15 ± 0.25	4.02 ± 0.18	31.00 ± 4.80	3.58 ± 0.28
All cones	Center	5	350–395	3.55 ± 0.25	4.25 ± 0.18	21.00 ± 2.18	3.88 ± 0.20
	Periphery	8	245–320	4.12 ± 0.22	4.18 ± 0.30	28.40 ± 2.85	3.57 ± 0.28

 Average NND, VD, and RI found for counting fields located in the far periphery (<350 μm from the ora serrata) and in the center (within 1000 μm from the ONH) of A. niloticus retina. Field areas were 0.01 mm² except for the S-cones (0.092 mm²).

M/L-cones.

The M/L-cone mosaic was studied on the retina that was initially used for density measurements (Fig. 2A). On average, NND and VD values were found to be respectively 3 and 14 to 17 times smaller than those observed in the S-cone population (Table 2). Regardless of the retinal position, the distributions of NND and VD values (Fig 3B) were symmetric around the mean (i.e., fitting a Gaussian distribution; P > 0.1 in all cases, χ² test of normality). Corresponding regularity indexes were equally high across the retina. Together these findings indicate that the M/L-cone mosaic is more regular than the S-cone mosaic is. 

Total Cone Mosaic.

The total cone mosaic was ascertained from two retinas processed with the anti-γ-transducin antibody, as previously described. ⁵ Results were essentially similar to those obtained for the M/L-cones (Table 2). 

UV Sensitivity of S-cones

According to genetic analysis, the presence of Phe86 is exclusive to all UV-sensitive S-opsins, whereas this site is occupied by a different amino acid in the blue/violet-sensitive S pigments. ^22–24 Sequencing the first exon of the S-opsin gene in three different specimens of A. niloticus returned the following translated amino-acid sequence: 46-FVFFVGTPLNATVL VATLHYKKLRQPLNYILVNVSLGGFLFCIFSVFTVFIASCHGYFLF-GRHVCALEA-114. This sequence contains Phe46, Phe49, Thr52, Phe86, Thr93, and Ala114, indicating that the S-opsin in A. niloticus is likely sensitive to UV by comparison with known mammalian sequences (Table 3). 

Table 3.

 

Amino Acids at Key Positions in the S-opsin Pigment of Some Mammalian Species

Species	λmax	Amino Acid Positions
		46	49	52	86	93	114	118
(1) Homo sapiens	424	Thr	Leu	Phe	Leu	Pro	Gly	Thr
(2) Bos taurus	435	Phe	Phe	Thr	Tyr	Ile	Ala	Cys
(3) Cavia porcellus	430	Ile	Cys	Thr	Val	Ala	Gly	Ser
(4) Sciurus carolinensis	440	Phe	Phe	Thr	Tyr	Val	Ala	Ser
(5) Talpa europaea	UV	—	Phe	Thr	Phe	Thr	Ala	Ser
(6) Mus musculus	359	Phe	Phe	Thr	Phe	Thr	Ala	Ser
(7) Rattus norvegicus	358	Phe	Phe	Thr	Phe	Thr	Ala	Ser
(8) Arvicanthis niloticus	—	Phe	Phe	Thr	Phe	Thr	Ala	—

 GenBank Accession numbers: (1) NM_001708; (2) NM_174567; (3) AY_552608; (4) DQ_302163; (5) EU_130563; (6) NM_007538; (7) NM_031015.

Discussion

Previous immunohistochemical investigation with an antibody recognizing the cone-specific Gγ8 transducin subunit showed that cone photoreceptors (≈30%–35% of the total photoreceptor population) in the diurnal muridae A. niloticus are distributed evenly across the retina and that their packing density (≈36,000/mm² on average) increases only 1.8-fold from the far periphery to center. ⁵ The present study provides new insight into the topographic arrangement of cone photoreceptors in this rodent. 

Cone photoreceptors in A. niloticus segregate in two complementary populations: a major one expressing the M/L-opsin pigment and a minor one expressing the S-opsin pigment. Their respective proportions (13–14:1) are conserved in most mammalian species with rod-dominated retinas, including humans (outside the fovea ^25,26 ). Of note, the M/L:S ratio in A. niloticus is very similar to that in the rat retina ²⁷ ; further proof that this ratio is independent of the proportion of cone over total photoreceptor population. Dual opsin expression in a single cone was not detected by immunohistochemistry. Opsin coexpression in rodents examined so far is primarily a species-specific feature: it has been observed in pocket gopher, siberian hamster, common house mouse (Mus) and guinea pig; it is absent in the wood mouse (Apodemus sylvaticus), squirrel, and degu (Table 4 and related references). A recent report indicates that opsin coexpression is also absent in agoutis. ⁴⁰ In some instances (rat and gerbil), however, pigment coexpression is present postnatally, but is downregulated by 1 month of age. ³¹ Further studies are needed to explore whether this is also the case in the Nile grass rat. 

Table 4.

 

Characteristics of the S-cone Population in A. niloticus Compared with Other Rodents

Animal	Lifestyle	% Cones	Ratio M/L:S	S-cones
				Approx. n	L max	Retinal Distribution	Peak Location	Peak Density (×1000/mm2)
Pocket gopher (Thomomys bottae) 12	S	26.4	1:2 (DPC)	650,000	UV, 367	Even	32–37	Centroventral
Chilean cururo (Spalacopus cyanus) 11	S	10	5–14:1	?	UV, 365	DVG	6.5	Ventral half
Siberian Hamster (Phodopus sungorus) 28,29	N	1–2	1:1 (DPC)	220,000	UV, 360	Even	8–12	Ventronasal
Syrian Hamster (Mesocricetus auratus) 28,30	N	1–2	N/A	Absent	N/A	N/A	N/A	N/A
House mouse (Mus musculus) 10,31–34	N	3	2:1 (DPC)	95,000	UV, 359	DVG	22–24	Mid-ventral
Wood mouse (Apodemus sylvaticus) 31	N	?	25–18:1	7–8,000	UV, ?	Even	0.5	Central, ONH
Rat (Rattus norvegicus) 27,35	N	1	14:1	20,000	UV, 359	Even	0.85	Central, ONH
African giant rat (Cricetomys gambianus) 36	N	<0.5	N/A	Absent	N/A	N/A	N/A	N/A
Mongolian gerbil (Meriones unguiculatus) 37	D	12–14	20:1	130,000	UV, 360	Even	?	?
Chilean degu (Octodon degus) 38	D	32	13:1	221.000	UV, 362	Even	5.4–7.0	Centroventral
Guinea pig (Cavia porcellus) 39	C	8–17	(DPC)	?	BV, 430	DVG	?	Ventral half
Agouti (Dasyprocta aguti) 40	D	17–19	13:1	260,000	N/A	Even	1.5	Temporocentral
Ground squirrel (Spermophilus beecheyi) 41,42	D	86	14:1	530,000	BV, 436	Even	20	Dorsonasal rim
Nile grass rat (Arvicanthis niloticus)	D	33	13–14:1	100,000	UV, ?	Even	3.3–4.1	Temporodorsal

 S, subterranean; N, nocturnal; C, crepuscular; D, diurnal; BV, bluish-violet sensitive; DPC, dual-pigment cones; DVG, dorsoventral gradient; ?, data not available; N/A, not applicable.

S-cone topography is extremely variable among mammalian species, ranging from a complete absence to an even distribution across the retina. ^26,43 In rodents, notable interspecies differences have been reported in the mouse, hamster, and squirrel. ^25,28,31,44 S-cones in A. niloticus are distributed throughout the retina as in Rattus and Apodemus (Table 4). Their density (≈2320 cells/mm² on average) increases (1.8-fold) from periphery to center, and is maximum in the temporodorsal quadrant. A pioneer investigation ³ in another Arvicanthis species, A. ansorgei, showed a similar distribution but with a higher centripetal increase (threefold), a significantly lower packing density (≈1480 cells/mm²), and no topographic preference. The minor discrepancies in S-cone densities may be related either to sampling bias or to species differences. There is evidence, for instance, that densities of retinal cells, including photoreceptors, differ tremendously between mice species or strains. ^{25,31,45–47}  

There is an S-cone-rich region in the temporodorsal quadrant of the Nile grass rat retina. For a still unknown reason, the location of this S-cone-rich region is closer to that found in ferret, lynx, and cheetah ^43,48 than that in most other rodents where the S-cone rich region is localized in the ventral retina (Table 4). It should be noted, however, that the retinal location of such S-cone-rich regions is extremely variable between mammalian species and that the visual function of these regions is not well understood. ^26,43 In A. niloticus, the S-cone-rich region approximately matches the area where M/L-cones are the densest. Both cone types might thus form a broad, horizontally oriented, centrotemporally elongated elliptical streak. Such retinal specialization may provide a better scanning of the natural environment in the frontal visual field. Further experiments are needed to examine whether there is also a corresponding higher density of rods and ganglion cells in this retinal sector that might subserve the higher visual acuity observed in A. niloticus compared with mice and rats. ⁵  

Within each nuclear layer of the retina, neurons of a given type are commonly spaced in an orderly manner, forming planar arrays known as mosaics ¹⁵ that are established during development through local interactions between immediate neighboring cells of the same type (for a review, see Ref. 49). Topographic analysis of mosaic patterns may be useful to pathologists: A recent study in humans ⁵⁰ showed for instance that progressive tritan color vision deficiency is accompanied by a progressive disorganization of the S-cone mosaic, which ultimately disrupts the whole cone mosaic. Knowledge of the normal pattern of retinal cell organization in A. niloticus is therefore essential for comparative and developmental studies as well as for experimental manipulations. ⁵¹  

Absence of cell body–specific labels for each cone subtype led us to rely on markers expressed in the OS to map cone mosaics in A. niloticus. This approach revealed contrasting spatial arrangements between cone subtypes: the S-cone mosaic is distinctly irregular (with nearest-neighbor RI and Rn indexes just above the theoretical upper limit for randomness ⁵² ), whereas that of the M/L-cones is more ordered. Among rodents, such dissimilar patterning has been described recently in the diurnal agouti, Dasyprocta aguti. ⁴⁰ This pattern seems to be the rule in rod-dominated mammalian retinas. ⁴³ The present results further suggest that the S-cone mosaic (with significantly large NND and VD values) is not integrated into the dominant cone mosaic in A. niloticus and does not interfere with the regularity of the total cone mosaic (which furthermore remains stable across the retina as observed in the horseshoe bat ⁵³ ). Such interference ⁴³ occurs probably only between two largely regular and independent mosaics, as is the case in primates. S-cones in the Nile rat retina may be too sparse to disturb the general cone pattern to any large extent. 

The presence of UV photopigments in retinal cones is usually considered to be the ancestral template from which emerged the classic mammalian S-type pigments after multiple mutations. ^24,54 According to an extensive literature review, all UV-sensitive S pigments in mammals have Phe86. ^22–24 Substitution of Tyr86Phe indeed is sufficient to shift the λ_max of the bovine blue/violet-sensitive opsin from 430 to 360 nm. Of the rodents studied so far, only diurnal squirrels, prairie dogs (Sciurus carolinensis, Spermophilus beecheyi, and Cynomys ludovicianus; all from the Sciuridae family), and crepuscular guinea pigs (Cavia porcellus; Cavidae) have been found to possess blue/violet-sensitive S-cones. ^44,55,56 In general, muridae have UV-sensitive S-cones (or no S-cones at all; Tables 3, 4). Although diurnal, Arvicanthis retains this characteristic, like its nocturnal relatives Mus, Rattus, and Apodemus, the S-opsin of A. niloticus has the exclusive Phe86 suggesting that S-cones are UV-sensitive. To determine whether these cones respond efficiently to UV light, the spectral transmission properties of the the lens and cornea must be assessed. Of note, however, in all the rodents studied so far having UV-sensitive S-cones, the eye optics transmit in the near-UV range. ^{11,13,37,38,57}  

The function of UV vision in diurnal rodents with UV-sensitive S-cones (gerbil and degu) is a matter of debate. Contribution to color discrimination, improvement of visual perception at dawn and dusk (when UV light levels are overrepresented in the solar spectrum ³⁷ ), as well as involvement in social behavior (detection of urine and fur reflectance) have all been postulated. ^38,58,59 These postulates apply well to the Nile grass rat which has a typical dual-cone retina, is mostly diurnal with peaks of activity at dawn and dusk, and spends long daylight periods interacting with its juvenile offspring outside densely interconnected burrows. ⁶⁰ Whether and how A. niloticus relies on UV-sensitive cones in its natural environment nevertheless remains an open question. 

In conclusion, the Nile grass rat A. niloticus has a dual cone retina with a majority of M/L-cones and a minority of S-cones. The S-cones are UV-sensitive, distribute throughout the retinal surface, and peak in the temporodorsal quadrant. Omitting the latter, these characteristics compare well with those found in Mongolian gerbil ⁵⁷ and Chilean degu, ^38,59 two distantly related diurnal rodents, but with similar lifestyles. Further investigations will examine how S-cones in A. niloticus are integrated into the retinal circuitry and whether these photoreceptors may participate in some form of color perception. 

Footnotes

 Supported by Canadian Institutes of Health Research (CIHR) Grant 151145; an Alberta Heritage Foundation for Medical Research (AHFMR) equipment grant; the Canadian National Institute for the Blind; the Olive Young Foundation; and The Lena McLaughlin Foundation (Mona & Rod McLennan). FG was supported by the International Society for Clinical Electrophysiology of Vision (ISCEV; short lab visit grant 2008). YS is an AHFMR Senior Scholar.

Footnotes

 Disclosure: F. Gaillard, None; S. Kuny, None; Y. Sauvé, None

Footnotes

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

The authors thank Laura Smale for generously providing breeders from her Nile grass rat colony at Michigan State University (supported by a grant from the National Institute for Mental Health, R01 MH53433) to start the colony that presently exists at the University of Alberta. The authors also thank Paul Freund for performing the Western Blot analyses, Marc-Andre Filion for support in density measurements, William Ted Allison for his advice on the characterization of the S-opsin gene, Sue Kenney, and The Applied Genomics Centre (TAGC), Department of Medical Genetics, and The Division of Gastroenterology (CEGIIR), University of Alberta, for technical support in sequencing.