Cheng 2007
Cheng 2007
Cheng 2007
Abstract
The retina of salmonid fishes has two types of cone photoreceptors: single and double cones. At the nuclear level,
these cones are distributed in a square mosaic such that the double cones form the sides of the square and the single
cones occupy positions at the centre and at the corners of the square. Double cones consist of two members, one
having visual pigment protein maximally sensitive to green light ~RH2 opsin!, the other maximally sensitive to red
light ~LWS opsin!. Single cones can have opsins maximally sensitive to ultraviolet ~UV! or blue light ~SWS1 and
SWS2 opsins, respectively!. In Pacific salmonids, all single cones express UV opsin at hatching. Around the time of
yolk sac absorption, single cones start switching opsin expression from UV to blue, in an event that proceeds from
the ventral to the dorsal retina. This transformation is accompanied by a loss of single corner cones such that the
large juvenile shows corner cones and UV opsin expression in the dorsal retina only. Previous research has shown
that adult Pacific salmon have corner cones over large areas of retina suggesting that these cones may be
regenerated and that they may express UV opsin. Here we used in-situ hybridization with cRNA probes and
RT-PCR to show that: ~1! all single cones in non-growth zone areas of the retina express blue opsin and ~2! double
cone opsin expression alternates around the square mosaic unit. Our results indicate that single cone driven UV
sensitivity in adult salmon must emanate from stimulation of growth zone areas.
Keywords: Ultraviolet opsin, Visual pigment, Cone mosaic, Regeneration, Fish retina
Introduction single cones in the ventral retina start switching opsin expression
from SWS1 to SWS2 ~the latter maximally sensitive to blue light!
The retinas of most fishes have cone photoreceptors that are
in a transformation event that expands toward the dorsal retina
distributed in lattice-like formations termed mosaics ~Lyall, 1957a;
~Cheng & Novales Flamarique, 2004; Cheng et al., 2006, 2007!.
Engström, 1963; Ahlbert, 1969!. The disposition and chromatic
The loss of corner cones from the ventral retina in juvenile
organization of these mosaics is crucial to all aspects of vision, and
salmonids has been shown for all species studied ~Lyall, 1957b;
their study can reveal important insights into general mechanisms
Bowmaker & Kunz, 1987; Kunz et al., 1994; Novales Flamarique
of cell pattern formation and function ~Cook & Chalupa, 2000!.
2000, 2001, 2002; Cheng et al., 2006!. Large juveniles ~termed
Like most other fishes, Pacific salmonids have cone photorecep-
smolts because they have undergone a physiological transforma-
tors that, at the nuclear level, are organized in a square mosaic
tion for life in seawater! have corner cones and UV ~SWS1! opsin
pattern ~Lyall, 1957b; Ahlbert, 1976; Bowmaker & Kunz, 1987;
expressing cones restricted to the upper half of the retina ~Cheng
Beaudet et al., 1997; Novales Flamarique, 2002; Cheng et al.,
et al., 2006!.
2006!. At hatching, the mosaic unit consists of four double cones
For many years it was believed that the loss of corner cones and
forming the sides of the square and single cones at the centre and
associated UV sensitivity in salmonid fishes took place at smolt-
at the corners of the square ~Cheng et al., 2006, 2007!. Each
ification, that this loss was complete, and that it was mediated by
member of a double cones expresses an opsin that is maximally
the thyroid hormone precursor, T4 ~Hawryshyn et al., 1989; Brow-
sensitive to red light ~LWS opsin! or green light ~RH2 opsin! and
man & Hawryshyn, 1992!. Because thyroid hormone has also been
the single cones express an opsin ~SWS1! that is maximally
associated with sexual maturation of salmon ~Sower & Schreck,
sensitive to ultraviolet ~UV! light, regardless of if they are centre
1982; Youngson & Webb, 1993!, it was hypothesized that the
or corner cones ~Cheng et al., 2006!. Following hatching and
corner cones present in the retina of adult fish were regenerated
closer to the time of full yolk sac absorption ~the alevin stage!,
and expressed UV opsin ~Beaudet et al., 1997!. It is now known
that corner cones are not completely lost from the juvenile retina
Address correspondence and reprint requests to: Iñigo Novales Flama-
and that this loss and the decrease in UV opsin is variable and
rique, Department of Biological Sciences, Simon Fraser University, Burnaby, starts before smoltification, at the young alevin stage ~Cheng et al.,
British Columbia V5A 1S6, Canada. E-mail: [email protected] 2006, 2007; see also diminished UV opsin transcript levels in the
269
270 C.L. Cheng and I.Novales Flamarique
ventral vs. the dorsal retina of ; 5 g rainbow trout, termed parr, mosaic. Digital images of representative sections from any given
Veldhoen et al., 2006!. Whether the corner cones present in the sector of the retina were acquired with an E-600 Nikon microscope
adult retina of salmonid fishes express UV opsin and whether they equipped with a DXM-100 digital camera and DIC optics. With the
are regenerated is unknown. In this study, we investigated these aid of a grid system on the computer monitor, the density of double
questions using histological methods that involved cell density cones and single cones ~center cones ⫹ corner cones! in a
counts and in-situ labeling with cRNA probes against the various 32700 mm 2 area was counted and the ratio of double to single
opsin mRNAs in salmon retina. We also amplified opsin transcripts cones ~d0s! computed for each sector of the retina ~d refers to the
using reverse transcriptase polymerase chain reaction ~RT-PCR! to two member double cone pair!. A d0s ratio close to 2 indicates a
compare with the in-situ labeling results. If corner cones expressed total loss of corner cones, whereas a ratio close to 1 indicates full
UV opsin in the adult retina, then these cells would be labeled with corner cone presence ~see Beaudet et al., 1997; Novales Flama-
our UV riboprobe. Alternatively, we expected corner cones to rique, 2001!. The contours of pieces mapped back to the original
express blue opsin, since this is the only other opsin that has been retina were used to trace maps of cone density and double cone to
found in the single cones of salmonid fishes ~Novales Flamarique, single cone ~d0s! ratio. Retinas used to obtain an average map were
2005; Cheng et al., 2006!. of similar size, minimizing differences due to age.
mosaic. Serial cryosections were collected on poly-L-lysine coated corners ~Fig. 1A!. Corner cones faced the partitioning membranes
slides and used for in-situ hybridization with the various opsin of adjacent double cones whereas center cones were located at the
riboprobes as per previous studies ~Cheng & Novales Flamarique, intersection of the partitions, if these were to be imaginary ex-
2004; Cheng et al. 2006, 2007!. The procedure involved rehydrat- tended. The row mosaic had double cones and single cones ar-
ing the sections, permeabilizing them in 10 mg0mL proteinase K ranged primarily in rows, though the partitions from adjacent
~Sigma! for 6 min, followed by exposure to 0.1 M triethanolamine double cones could still trace the incomplete outline of a square
containing 0.25% acetic anhydride dehydration and hybridization ~Fig. 1B!. Although the differences between row mosaic and
overnight at 568C with 1–5 mg of riboprobe in hybridization square mosaic are often blurry at the ellipsoid level ~e.g., both
solution containing 50% formamide and dextran sulfate. Sections mosaics can be viewed as alternating rows of double cones and
were then treated with RNase A ~Sigma! and incubated with single cones!, in general, the partitions of neighboring double
appropriate Fab fragments conjugated to alkaline phosphatase cones, if extended, would intersect at an angle closer to 908 in a
~1:3000; Roche Diagnostics! for 3 h at room temperature. The square versus a row mosaic, and the single cones tend to be
riboprobes were visualized using 5-bromo-4-chloro-3indolyl phos- equidistant in a square mosaic. In the central and centro-temporal
phate with 4-nitroblue tetrazolium chloride ~NBT0BCIP, Roche retina, there were both square and row mosaics complete with
Diagnostics!. In the case of double labeling experiments, sections corner cones ~Figs. 1A, 1B!. More distal dorsal areas showed
were incubated first with anti-DIG Fab fragments conjugated to square to row mosaics with fewer to no corner cones ~Figs. 1C,
alkaline phosphatase ~1:3000! for 2 h, and the DIG-labeled probes 1D!, which was also the pattern observed in sections going from
visualized using NBT-BCIP. The color reaction was stopped by the centro-nasal retina ~Fig. 1E! to the nasal periphery ~Fig. 1F!.
washing the sections in glycine-HCl ~0.1M, pH 2.2!. To visualize The ventral retina had few to no corner cones, the mosaic was
the fluorescein-labeled probes, the sections were then incubated square ~Fig. 1G!, and cone density increased dramatically towards
with anti-fluorescein Fab fragments conjugated to alkaline phos- the periphery ~Fig. 1H!. These patterns are summarized in Fig. 2
phatase ~1:3000! for 2 h, and stained with Fast red ~Roche! for 2 h. where the d0s ratio is shown to be lowest ~range: 1.0–1.3! in the
Sense probes were used as negative controls and did not hybridize centro-dorso-temporal retina, and the highest cone densities are
in any of the retinas. Sections from retinas of younger fish were found in the ventro-temporal and dorso-nasal peripheries. All
used as positive controls. The same microscopy set-up used to salmon species studied showed the same patterns, which were also
collect images from EPON embedded sections was used to photo- consistent with those reported by Beaudet et al. ~1997!.
graph cryosections.
Identification of cone spectral classes
Reverse transcriptase polymerase
Our in-situ hybridization experiments resulted in similar findings
chain reaction (RT-PCR)
for all salmonid species. Regardless of whether they were center or
Two retinas from freshly killed coho salmon were extracted from corner cones, single cones labeled with the blue opsin riboprobe
the eyecups, immersed in RNA-later solution ~RNAqueous-Midi, ~Figs. 3A, 3B! and none of these cones labeled with the UV
Ambion!, and stored at 48C. The following day, each retina was cut riboprobe ~Fig. 3C!. Each member of a double cone labeled with
into small pieces corresponding to sectors of the main retina used either the green or red riboprobe, such that each double cone was
for histological analyses. These were immersed in new RNA-later a green-red pair ~Figs. 3D, 3E!. In tangential sections progressing
solution and stored at 48C. Ultraviolet opsin partial cDNA was from the pigment epithelium towards the ganglion cell layer,
generated by RT-PCR amplification of DNAse-treated total RNA labeling by the green riboprobe appeared first indicating that the
isolated from each homogenized piece of retina, as described red cone is displaced slightly vitreal with respect to its green
previously. Primers used for RT-PCR were designed from the UV partner ~Fig. 3D!. The expression of either of these two opsins
opsin sequence to span across introns to distinguish amplicons alternated along the unit square mosaic so that each corner cone
from DNA contamination. These primers were: UV opsin forward was flanked by two green and two red cones ~Fig. 3E!. These
5 ' -GGG CTT TGT GTT CTT TGC TG–3 ', reverse 5 ' -GGT ACT results depict the chromatic organization sketched in Fig. 4. The
CCT CGT TGT TTG TG–3 ', b-actin was used as a control, b-actin only UV labeling ever detected occurred in the dorsal periphery
forward 5 ' -CCC ATG GAG CAC GGT ATC ATC AC-3 ', reverse ~Fig. 3F!. These cells stained heavily and were often found in
5 ' -GCG TGG GGC AGA GCG TAA CCT TC-3 '. Reaction prod- areas where the retina was not fully differentiated ~growth zones!.
ucts were analyzed by electrophoresis on 1% agarose gels in 1 ⫻ By comparison, the central retina of the young alevin ~used as a
Tris Boreate EDTA ~TBE! containing 0.5 mg0mL ethidium bro- control! showed labeling of all single cones with the UV riboprobe
mide, and photographed. These experiments were conducted to ~Fig. 3G!, as has been reported previously ~Cheng et al., 2006!.
identify any UV opsin transcripts in the main retina of adult coho Labeling with the rod riboprobe was observed throughout the
salmon that may had been missed by in-situ hybridization, because retina. This labeling was restricted to the small inner segments of
the latter technique is not as sensitive. rod photoreceptors that interdigitated between cones in the light-
adapted retina ~e.g., Fig. 3A!. In accordance with the lack of UV
labeling observed in the main retina by in-situ hybridization, we
Results
did not amplify any UV opsin transcripts by RT-PCR analysis of
isolated retinas. That UV transcripts were not found by RT-PCR in
Distributions of morphologically-defined
peripheral regions may reflect variations in the amount of periph-
cone photoreceptor types
eral retina extracted between fish or differential expression be-
The cone photoreceptors were arranged in mosaics that varied tween individuals.
from square to row at the ellipsoid level ~Fig. 1!. The square In general we found that opsin content, as judged by intensity
mosaic consisted of double cones forming the sides of the unit of riboprobe labeling, was much lower in adult fish than in
square and single cones in the center and, when present, at the juveniles ~Cheng et al., 2006! suggesting an overall downregula-
272 C.L. Cheng and I.Novales Flamarique
Fig. 1. Cone mosaic formations in the light adapted retina of adult Pacific salmon. A. square mosaic with corner cones from the centro
dorsal retina. An asterix ~*! indicates a corner cone while a black arrow head indicates a centre cone; the partitioning membrane
separating double cone members is indicated by a white arrow head. B. Row mosaic with corner cones from the centro-temporal
retina; t indicates a triple cone. C–F. Row mosaics devoid of corner cones from the proximal ~C, E! and distal ~D, F! dorsal and nasal
retina, respectively. G, H. Square mosaics without corner cones from the proximal ~G! and distal ~H! ventral retina. Cones are smaller
and more closely packed towards the peripheral retina. Magnification bar ~in a! ⫽ 25 mm holds for all panels.
Opsin expression in adult Pacific salmon 273
Fig. 2. Retinal maps of cone distributions in the retina of Pacific salmon ~n ⫽ 3 per map!. Shown for each location are the average
cone density and the associated d0s ratio ~in parenthesis!. Cone densities are expressed in thousands per square millimeter. A larger
circle indicates that the average cone density for that location was at least 1 standard deviation above the mean from all locations pooled
together. A smaller circle indicates the opposite. In each map, the full line polygon delineates the minimum area in which high corner
cone densities were found ~d0s ratio ⬍1.4!. Also embedded within each map is the approximate size of the corresponding smolt retina
~reproduced from Cheng et al., 2006! and its associated area of high corner cone density ~dashed polygon!. In salmonid fishes, the
embryonic fissure ~ef ! runs from the ventral to the central retina ~approximate location of the optic nerve head! pointing toward the
temporal retina. D, dorsal, N, nasal. Magnification bar ⫽ 0.55 cm ~coho!, 0.80 cm ~chum!, 0.61 cm ~chinook!, 0.46 cm ~chinook jack!,
and 0.54 cm ~pink!.
274 C.L. Cheng and I.Novales Flamarique
Fig. 3. Micrographs of tangential cryosections from the light adapted retina of adult Pacific salmon. A–C. Sections from the central
retina show a full square mosaic in which all the single cones label with the blue riboprobe ~dark blue stain, A, B! and none label with
the UV riboprobe ~C!. The rod riboprobe labels rod inner segments ~red stain, black arrows!, which occupy the space between cones.
D, E. The green and red riboprobes label alternating double cone members in the square mosaic; in this case, the blue and red stains
indicate labeling by the green and red opsin riboprobes, respectively. F. Labeled cones ~white arrows! expressing UV opsin mRNA in
the peripheral dorsal retina. G. All single cones in the central retina of the alevin label with the UV opsin riboprobe ~blue stain!; the
rhodopsin riboprobe labels rod inner segments in the juvenile retina ~red stain!. Other symbols as in Fig. 1. Magnification bar ~in A! ⫽
15 mm holds for ~D–G!, and ⫽ 25 mm for ~B, C!.
tion of opsins with age. For instance, in chinook salmon, the Discussion
precocial jacks showed much stronger labeling than normal spawn-
ers ~which are older by 1–3 years!. Similar conclusions were This study demonstrates that UV opsin is not expressed in the main
reached by comparing absorbance of visual pigments obtained by ~non-peripheral! retina of the sexually mature Pacific salmon
microspectrophotometry ~results not shown!. species studied. The corner cones present in the dorso-temporal
Opsin expression in adult Pacific salmon 275
Cheng, C.L., Novales Flamarique, I., Hárosi, F.I., Rickers- Lyall, A.H. ~1957a!. Cone arrangements in teleost retinae. Quarterly
Haunerland, J. & Haunerland N.H. ~2006!. Photoreceptor layer of Journal of Microscopy Science 98, 189–201.
salmonid fishes: Transformation and loss of single cones in juvenile Lyall, A.H. ~1957b!. The growth of the trout retina. Quarterly Journal of
fish. Journal of Comparative Neurology 495, 213–235. Microscopy Science 98, 101–110.
Cheng, C.L., Gan, K. & Novales Flamarique, I. ~2007!. The ultraviolet Martens, G.D. ~2000!. Topographical changes to the cone photoreceptor
opsin is the first opsin expressed during retinal development of sal- mosaic in thyroxin challenged rainbow trout ~Oncorhynchus mykiss!.
monid fishes. Investigative Ophthalmology and Visual Science 48, University of Victoria Master Thesis, 134 pp.
866–873. Novales Flamarique, I. ~2005!. Temporal shifts in visual pigment ab-
Cook, J.E. & Chalupa, L.M. ~2000!. Retinal mosaics: New insights into sorbance in the retina of Pacific salmon. Journal of Comparative
an old concept. Trends in Neuroscience 23, 26–34. Physiology A 191, 37– 49.
Deutschlander, M.E., Greaves, D.K., Haimberger, T.J. & Hawry- Novales Flamarique, I. ~2002!. Partial re-incorporation of corner cones
shyn, C.W. ~2001!. Functional mapping of ultraviolet photosensitivity in the retina of the Atlantic salmon ~Salmo salar!. Vision Research 42,
during metamorphic transitions in a salmonid fish, Oncorhynchus 2737–2745.
mykiss. Journal of Experimental Biology 204, 2401–2413. Novales Flamarique, I. ~2001!. Gradual and partial loss of corner
Engström, K. ~1963!. Cone types and cone arrangements in teleost cone-occupied area in the retina of rainbow trout. Vision Research 41,
retinae. Acta Zoologica (Stockholm) 44, 179–243. 3073–3082.
Forsell, J., Ekström, P., Novales Flamarique, I., Holmqvist, B. Novales Flamarique, I. ~2000!. The ontogeny of ultraviolet sensitivity,
~2001!. Expression pattern of pineal UV- and green-like opsins in cone disappearance and regeneration in the sockeye salmon Oncorhyn-
teleosts. Journal of Experimental Biology 204, 2517–2525. chus nerka. Journal of Experimental Biology 203, 1161–1172.
Hawryshyn, C.W., Martens, G., Allison, T.W. & Anholt, B. ~2003!. Novales Flamarique, I. & Hawryshyn, C.W. ~1996!. Retinal develop-
Regeneration of ultraviolet-sensitive cones in the retinal cone mosaic of ment and visual sensitivity of young Pacific sockeye salmon ~Onco-
thyroxin-challenged post-juvenile rainbow trout ~Oncorhynchus mykiss!. rhynchus nerka!. Journal of Experimental Biology 199, 869–882.
Journal of Experimental Biology 206, 2665–2673. Sower, S.A. & Schreck, C.B. ~1982!. Steroid and thyroid hormones
Hawryshyn, C.W., Arnold, M.G., Chiasson, D. & Martin, P.C. ~1989!. during sexual maturation of coho salmon ~Oncorhynchus kisutch! in
The ontogeny of ultraviolet photosensitivity in rainbow trout ~Salmo seawater or freshwater. General Comparative Endocrinology 47, 42–53.
gardneri !. Visual Neuroscience 2, 247–254. Veldhoen, K., Allison, T.W., Veldhoen, N., Anholt, B.R., Helbing,
Hoar, W.S. ~1988!. The physiology of smolting salmonids. In Fish Phys- C.C. & Hawryshyn, C.W. ~2006!. Spatio-temporal characterization of
iology, ed. Hoar, W.S. & Randall, D.J., pp. 275–343. New York: retinal opsin gene expression during thyroid hormone-induced and
Academic Press, Inc. natural development of rainbow trout. Visual Neuroscience 23, 169–179.
Kunz, Y.W., Wildenburg, G., Goodrich, L. & Callaghan, E. ~1994!. Youngson, A.F. & Webb, J.H. ~1993!. Thyroid hormone levels in Atlantic
The fate of ultraviolet receptors in the retina of the Atlantic salmon salmon ~Salmo salar! during the return migration from the ocean to
~Salmo salar!. Vision Research 34, 1375–1383. spawn. Journal of Fish Biology 42, 293–300.