Isolation and characterization of two teleost melanopsin genes and their differential expression within the inner retina and brain.
ABSTRACT Melanopsin is a newly discovered photopigment that is believed to be involved in the regulation of circadian rhythms in tetrapods. Here we describe the characterization of the first two teleost melanopsins (opn4a and opn4b) isolated from Atlantic cod (Gadus morhua). These two teleost genes belong to a subgroup of melanopsins that also include members from Xenopus, chicken, and Takifugu. In situ hybridization revealed that opn4a and opn4b are differentially expressed within the retina and brain. In the larval and adult retina, both melanopsins are expressed in a subset of cells in the inner retina, resembling amacrine and ganglion cells. In addition, opn4a is expressed in the horizontal cells, indicating a separate task for this gene. In the brain, the two melanopsins are separately expressed in two major retinal and extraretinal photosensitive integration centers, namely, the suprachiasmatic nucleus (opn4a) and the habenula (opn4b). The expression of opn4a in the suprachiasmatic nucleus in cod is similar to the melanopsin expression found in Xenopus. This suggests a conserved role for this opsin and an involvement in mediation of nonvisual photoreceptive tasks, such as entraining circadian rhythms and/or hypophysiotrophic systems. The differential expression of opn4b in the habenula suggests that this gene plays a role similar to that of opn4a, in that it is also situated in an area that integrates photic inputs from the pineal as well as other brain regions. Thus, the habenula may be an additional region that mediates photic cues in teleosts.
Isolation and Characterization of Two
Teleost Melanopsin Genes and Their
Differential Expression within the
Inner Retina and Brain
ØYVIND DRIVENES,1ANNE METTE SØVIKNES,1LARS O.E. EBBESSON,2
ANDERS FJOSE,1HEE-CHAN SEO,3
1Department of Molecular Biology, University of Bergen, N-5020 Bergen, Norway
2Department of Fisheries and Marine Biology, University of Bergen,
N-5020 Bergen, Norway
3Sars Centre for Molecular Marine Biology, N-5008 Bergen, Norway
AND JON VIDAR HELVIK1*
Melanopsin is a newly discovered photopigment that is believed to be involved in the
regulation of circadian rhythms in tetrapods. Here we describe the characterization of the
first two teleost melanopsins (opn4a and opn4b) isolated from Atlantic cod (Gadus morhua).
These two teleost genes belong to a subgroup of melanopsins that also include members from
Xenopus, chicken, and Takifugu. In situ hybridization revealed that opn4a and opn4b are
differentially expressed within the retina and brain. In the larval and adult retina, both
melanopsins are expressed in a subset of cells in the inner retina, resembling amacrine and
ganglion cells. In addition, opn4a is expressed in the horizontal cells, indicating a separate
task for this gene. In the brain, the two melanopsins are separately expressed in two major
retinal and extraretinal photosensitive integration centers, namely, the suprachiasmatic
nucleus (opn4a) and the habenula (opn4b). The expression of opn4a in the suprachiasmatic
nucleus in cod is similar to the melanopsin expression found in Xenopus. This suggests a
conserved role for this opsin and an involvement in mediation of nonvisual photoreceptive
tasks, such as entraining circadian rhythms and/or hypophysiotrophic systems. The differ-
ential expression of opn4b in the habenula suggests that this gene plays a role similar to that
of opn4a, in that it is also situated in an area that integrates photic inputs from the pineal
as well as other brain regions. Thus, the habenula may be an additional region that mediates
photic cues in teleosts. J. Comp. Neurol. 456:84–93, 2003.
© 2002 Wiley-Liss, Inc.
Indexing terms: retina; circadian rhythms; suprachiasmatic nucleus; habenula; opsin; Takifugu;
The retina is a sensitive photodetector that captures
light by an 11-cis retinal chromophore bound by the opsin
G-protein-coupled receptors that undergo conformational
change when hit by photons, initiating an intracellular
absorbing proteins are believed to regulate a variety of
physiological processes, of which vision is the best known.
The visual pigments, cone opsins in cone photoreceptors
and rod opsin in rod photoreceptors, are well character-
ized and have been classified by their absorbance spec-
trum and sequence characteristics (Bowmaker and Hunt,
1999). Important findings with rodless/coneless mice sug-
gest that traditional photoreceptors are not required for
the regulation of temporal physiology and circadian pho-
totransduction (Freedman et al., 1999; Lucas et al., 1999).
Instead, several novel nonvision-mediating opsins have
Grant sponsor: The Norwegian Research Council; Grant number:
133875; Grant number: 133881.
*Correspondence to: Jon Vidar Helvik, Department of Molecular Biology,
University of Bergen, P.O. Box 7800, N-5020 Bergen, Norway.
Received 19 July 2002; Revised 30 August 2002; Accepted 14 October
Published online the week of December 9, 2002 in Wiley InterScience
THE JOURNAL OF COMPARATIVE NEUROLOGY 456:84–93 (2003)
© 2002 WILEY-LISS, INC.
been identified in a number of vertebrates that are be-
lieved to be involved in the regulation of diverse physio-
logical processes, such as color camouflage, responses to
seasonal change, and entraining of circadian rhythms
(Shand and Foster, 1999). These include pinopsin in
chicken (Okano et al., 1994), parapinopsin in catfish
(Blackshaw and Snyder, 1997), vertebrate ancient (VA)
opsin in salmon and zebrafish (Soni and Foster, 1997;
Kojima et al., 2000), and melanopsin in Xenopus and
mammals (Provencio et al., 1998, 2000).
Melanopsin was first isolated from light-sensitive pig-
ment skin cells (melanophores) in Xenopus laevis (Prov-
encio et al., 1998) and subsequently from mammals (Prov-
encio et al., 2000). Xenopus melanopsin is expressed in the
retinal pigment epithelium (RPE), iris, and inner nuclear
layer of the retina and in hypothalamic neurons (Proven-
cio et al., 1998). In mammals, expression of melanopsin
seems to be restricted to the ganglion and amacrine cell
layers of the retina (Provencio et al., 2000). Recently,
several studies have shown that melanopsin is expressed
in photosensitive ganglion cells that project into the su-
prachiasmatic nucleus of the mammalian brain, support-
ing the hypothesis that melanopsin may be a key regula-
tor in the photoentrainment of the circadian clock (Gooley
et al., 2001; Berson et al., 2002; Hannibal et al., 2002;
Hatter et al., 2002).
Teleost fish have adapted to almost every niche in the
hydrosphere, ranging from the ocean depths, where no
light penetrates, to the photic zone near the surface. One
key to understanding these remarkable adaptations to
different photic environments is the variation found in
spectral sensitivity and the expression and number of
pigments in the retina and brain of different species. Mo-
lecular studies in freshwater teleosts have given new in-
sight into the molecular adaptation of photopigments in
relation to light (Hunt et al., 1997, 2001; Cowing et al.,
2002). To gain further knowledge of such processes in the
marine environment, we screened marine pelagic fish by
degenerate polymerase chain reaction (PCR) for opsin-
coding transcripts. We identified, in Atlantic cod (Gadus
morhua), two photopigments that were different from
classical visual pigments (cone and rod opsin). These tran-
scripts are the first melanopsin genes isolated from te-
leosts and are shown to belong to a subgroup of melanop-
sins that also includes members from Xenopus, chicken,
and the pufferfish Takifugu. Expression analysis shows
that these photodetectors are located in the retina and
brain in a pattern that is specific for the two different
MATERIALS AND METHODS
Isolation of opsin cDNAs from Atlantic cod
Total RNA was isolated from the retina of adult Atlantic
cod using the Trizol reagent (Life Technologies, Bethesda,
MD). Poly(A)?mRNA was purified from total RNA
with Oligotex resin (Qiagen, Germany). Double-stranded
RACE-ready cDNA and adaptor-ligated RACE-ready
cDNA were made using the Marathon cDNA Amplifica-
tion kit (Clontech, Palo Alto, CA) according to the manu-
A DNA fragment of approximately 750 bp was amplified
by degenerate PCR as previously described (Helvik et al.,
2001). This fragment was subcloned and sequenced. Nu-
cleotide sequences were compared with the GenBank da-
tabase using the BLAST algorithm (Altschul et al., 1990).
In all, five different opsin-like genes were identified, in-
cluding two melanopsins.
RACE-PCR strategies were used to generate full-length
cDNAs of the two melanopsins as described previously
(Seo et al., 1998). 3?-RACE was performed using the gene-
specific primers Me1F (5?-CCA AGT GCG GAA ATC CAC
CAT AAT C, primary PCR) and Me2F (5?-AAG CAG CAG
ATG GAG AGA ACG AAC A, secondary PCR) for opn4a
and Me5F (CCT GGG CGG GGT ACG GGA GCC A, pri-
mary PCR) and Mop2F2 (GTC CAT TCA TGG TCG CCG
TAC, secondary PCR) for opn4b. Similarly, the gene-
specific primers Me4R (5?-GGT GTA GCT CCT GTT GGA
CGC CGT GTA A, primary PCR) and Me6R (5?-AGG GTG
CCG ATG ACG GAG ACG AAG AAA G, secondary PCR)
for opn4a and Me9R (GTC TCG CCA AAA ATC CAC TCT
TTG AA) for opn4b were used for 5?-RACE.
All PCR products were ligated into pGEM-T Easy Vec-
tor (Promega, Madison, WI) and sequenced in both direc-
tions using the ABI Prism Dye Terminator Cycling se-
quencing kit (Perkin Elmer, Norwalk, CT). The nucleotide
sequences were deposited into GenBank with the acces-
sion numbers AF385823 (opn4a) and AY126448 (opn4b).
Searching the Takifugu rubripes database
for melanopsin sequences
By using the Fugu BLAST server (http://fugu.hgmp.
mrc.ac.uk/blast/) with the cod Opn4a protein as the query
sequence, three putative melanopsin genes were identified
(opn4c)]. Further examination by BLAST search (Altschul
et al., 1990) in the Genbank database and phylogenetic
analyses (see below) suggested that these genes probably
also encode melanopsins and they were therefore named
Sequence and phylogenetic analysis
ClustalX (1.64b) was used to align the amino acid se-
quences. Phylogenetic analysis was based on these align-
ments and was performed with the Puzzle 4.0.2 program
(http://www.tree-puzzle.de/) using the maximum-likeli-
hood method, with the substitution model of Strimmler
and von Haeseler (1997). The amino acid sequences in the
long cytoplasmic tail of the melanopsin, which were
greater than the common length of opsins, were removed
prior to alignment. Phylogenetic trees were constructed
with bootstrap confidence values based on 1,000 repli-
Reverse transcription-PCR analysis
Total RNA was isolated from brain, retina, skin, and
liver tissues using Trizol reagent (Gibco-BRL, Grand Is-
land, NY), chloroform/phenol extraction and ethanol pre-
cipitation. To prevent contamination of genomic DNA, the
samples were treated with DNase (Promega). DNase was
inactivated subsequently at 70°C for 5 minutes, and the
samples were again subjected to chloroform/phenol extrac-
tion and ethanol precipitation. Reverse transcription (RT)
was carried out using 2 ?g total RNA, M-MLV reverse
transciptase (Promega), and the gene-specific primers
Mops1RTr (CGC CAA GGT CTT CCT GTA CTT C) for
opn4a and Mops2RTr (CTC GGC TAG TGT GTC TCT
GTA TTT AGA) for opn4b. PCRs were performed using the
85 MELANOPSINS IN TELEOSTS
Advantage 2 PCR enzyme system (Clontech) and 2 ?l of
the cDNA reaction mix. For opn4a, PCR was performed
with the primers Mops1RTf (CGT TGG CTG GAG CTC
CTA CAT C) and Mops1RTr, generating a 466 bp PCR
product. In total 35 cycles were performed at an annealing
temperature of 62°C. For opn4b, PCR was performed with
the primers Mops2RTf (TGG CGA GAC CGG CTG TAG
GA) and Mops2RTr, generating a 673 bp PCR product. In
total 35 cycles were performed at the annealing tempera-
ture of 68°C. Five microliters of each PCR were analyzed
on a 2% agarose-TAE gel. The analysis was repeated three
times with samples from three different fish, and all gave
the same results.
In situ hybridization
Eyes and brains from 1-year-old juvenile Atlantic cod
and the whole head from larval Atlantic cod were sampled
after killing the fish by decapitation. All fish were cultured
at the Department of Fisheries and Marine Biology, Uni-
versity of Bergen. Eyes from adult fish were removed, and
the dorsal region of the head was opened prior to fixation
in 4% paraformaldehyde in phosphate-buffered saline
(PBS; pH 7.2) for 48 hours at 4°C. Eyes and brain were
then dissected from the head prior to overnight treatment
in 25% sucrose PBS (pH 7.2) solution with 30% OCT
(Tissue Tek Sakura Finetek Europe). Whole eyes and
brains were placed in a mold containing 100% OCT and
lined up in an anterior-posterior orientation perpendicu-
lar to (for transversal sections) or parallel to (for sagittal
sections) the bottom of the mold before rapid freezing on
an iron block precooled in liquid nitrogen. Flat mounts of
the whole retina were made according to Helvik et al.
Digoxigenin (DIG)-labelled RNA probes from the Atlan-
tic cod melanopsins were prepared following manufactur-
er’s instruction (Roche, Basel, Switzerland), and the probe
concentrations were determined by spot tests. In situ hy-
bridization was carried out according to Barthel and Ray-
mond (1993), with some modifications (Helvik et al.,
2001). Labelling was visualized with NBT and BCIP.
Sense probes were used to control for nonspecific staining
on parallel sections. Sections from five different fish were
used for each tissue and stage shown. Sections were ana-
lyzed with a Leica DMLB microscope (Wetzlar, Germany),
and images were captured using an integrated image sys-
tem (CoolSNAP-Pro Digital Camera Kit; Media Cybernet-
ics, Silver Spring, MD) with Nomarski optics. Adobe Pho-
toshop was used for arranging images, labelling, and
printing. The appropriate rules and procedures were fol-
lowed (The Norwegian Department of Agriculture, 1996,
sections 21 and 22, concerning animal treatment) in all
cases involving fish.
Key features of the two Atlantic cod
melanopsin amino acid sequences
The full lengths for both cod melanopsin cDNAs (opn4a
and opn4b) were isolated by PCR. The opn4a cDNA is
2,242 bp long and has a 1,686 bp open reading frame
(ORF), with the first putative translation initiation site at
nucleotide (nt) 264, predicting a protein of 561 amino acids
in length (Fig. 1). A second possible start codon is located
at nt 285, yielding a protein of 554 amino acids. However,
the first putative start codon is in an environment identi-
cal to the Kozak consensus sequence (ccATGg) for trans-
lation initiation (Kozak, 1987). This is a more favorable
environment than the second (cgATGg) initiation site. The
2,215 bp opn4b cDNA has a 1,845 bp ORF encoding a
615-amino-acid-long protein (Fig. 1) with a single start
codon at nt 202 (acATGgg).
embnet.org/software/TMPRED_form.html) of the Opn4a
and Opn4b proteins revealed typical opsin characteristics,
with seven transmembrane domains and an extracellular
N-terminal and a cytoplasmic C-terminal tail (Fig. 1). The
two cod melanopsin proteins are 75% identical when the
tail sequences are excluded (Table 1). However, Opn4a
and Opn4b show the highest sequence similarity to Opn4a
and Opn4b from Takifugu (83%) respectively, whereas
weaker similarity is shared with the mouse and human
homologues (?53% and ?56%). It has previously been
shown that melanopsins from Xenopus and mammals
show a high degree of similarity to invertebrate opsins
(Provencio et al., 1998, 2000). This is also true for the cod
melanopsins, which are ?35% identical to octopus rhodop-
sin and only ?25% identical to vertebrate rhodopsins (not
shown). Another common characteristic of melanopsin
proteins is the increase in length of the third cytoplasmic
loop that might be involved in the interaction with down-
stream components of the phototransduction cascade
(Terakita et al., 2002). As observed for other melanopsins
and many invertebrate opsins, the two cod proteins have
an aromatic residue [Tyr-106 (Opn4a), Tyr-89 (Opn4b);
Fig. 1] instead of an acidic residue in the third transmem-
brane helix, a position that is proposed to stabilize the
Schiff base (Gartner and Towner, 1995; Provencio et al.,
1998, 2000). The cytoplasmic and extracellular tails are
far less well conserved with respect to both sequence and
length for the different melanopsins. In addition, all mela-
nopsin proteins identified so far contain a high proportion
of potential phosphorylation sites, suggesting that the
activity of melanopsins might be regulated by kinases
(Provencio et al., 1998, 2000).
Phylogenetic relationship of melanopsin
and other opsins
Maximum-likelihood phylogenetic analysis was per-
formed using the inferred amino acid sequences from the
two cod melanopsin cDNAs and various sequences from
different opsin families. Scallop opsin was used as the out
group (Kojima et al., 1997). The resulting phylogenetic
tree positions melanopsin and invertebrate opsins to-
gether on one branch, whereas VA opsins and the different
classes of visual pigments are separated on the other (Fig.
2). The melanopsin family, however, is also separated into
two different subgroups. Opn4a and Opn4b from cod and
Takifugu cosegregate with melanopsins from Xenopus and
chicken, whereas the human, rat, and mouse are grouped
together with a third melanopsin (Opn4c) from Takifugu
on a discrete branch. It seems, therefore, that the melan-
opsins and invertebrate opsin classes have been separated
from the vertebrate opsins and have subsequently di-
verged into two subgroups. This is supported by the iden-
tification of three melanopsins in Takifugu, where mem-
bers are situated in each group. Moreover, it is noteworthy
that the proteins in subgroup I have a considerably longer
(30–35 aa) N-terminal tail and a shorter C-terminal com-
86Ø. DRIVENES ET AL.
logues. The following amino acid sequences were deduced from cDNA
nucleotide sequences and were compared with the cod and Takifugu
melanopsins: Xenopus melanopsin (Provencio et al., 1998), chicken
melanopsin (Accession number AAK59988), mouse melanopsin, and
Sequence comparison of cod melanopsins and their homo-
human melanopsin (Provencio et al., 2000). The seven predicted
transmembrane regions are overscored and numbered. Critical resi-
dues of chromophore linkage (asterisk) and disulfide bridge formation
(solid circles) are indicated. The solid square indicates the residue
corresponding to the acidic counterpart of vertebrate visual pigments.
ponent compared with the melanopsins in subgroup II
To investigate the expression of opn4a and opn4b in
different tissues, we performed RT-PCR analysis on sam-
ples from the skin, retina, brain, and liver (Fig. 3). Strong
expression could be detected for both genes in the brain
and retina samples, whereas only weak expression was
found in the skin sample. No expression was observed in
the liver, which served as a negative control.
Distribution in the retina
In situ hybridization using DIG-RNA probes recogniz-
ing both melanopsin transcripts demonstrates that both
types of melanopsins are expressed in the retina from an
early stage onward (Fig. 4A–E). opn4a and opn4b are
expressed in a specific manner with distinct characteris-
tics. opn4a expression is dominant in an array of cells just
below the photoreceptors, corresponding to the layer of
horizontal cells, during the entirety of development (Fig.
4F,H,J). Staining is also observed in what appears to be
bipolar cells in the inner nuclear layer at the larval stage
(Fig. 4F). This is not conclusive, however, because of the
extensive staining in this layer. The sense probe, however,
generated little or no staining (Fig. 4C). This suggests that
the diffuse staining surrounding the bipolar cells may
come from axons projecting from opn4a-positive horizon-
tal cells (Fig. 4B,F,H,J). A few single cells close to the
inner plexiform layer, resembling amacrine cells, and a
few cells in the ganglion cell layer are also stained (Fig.
4F). Tangential section of the inner retina at the level of
the horizontal cells shows that the opn4a-positive cells
have a mosaic distribution in which stained single cells
are surrounded by several unstained cells (Fig. 4I). This
indicates that only a subpopulation of horizontal cells
expresses opn4a. Although opn4a expression is dominant
in the horizontal cell layer, the opn4b-positive cells are
located on both sides of the inner plexiform layer resem-
bling amacrine and ganglion cells (Fig. 4D,E,G,K). At the
early larval stage, opn4b-positive cells are also found in
the horizontal cell layer, but this labelling disappears
completely during development, suggesting that opn4b is
developmentally regulated (Fig. 4E,G).
tree is constructed through the maximum-likelihood method. Node
values represent an analysis of 1,000 bootstrap trials. Scale bar indi-
cates genetic distance. Chicken (Gallus gallus; GALL) melanopsin
(MEL; AAK59988); fruit fly (Drosophila melanogaster; DR) RH4
(P08255); human (H) BLUE (NP001699), GREEN (NP000504), RED
(P04000; Nathans et al., 1986), rod opsin (ROD; P08100; Nathans and
Hogness, 1984), MEL AAF24978; Provencio et al., 2000); mouse (Mus
musculus; MUS) MEL (AAF24979; Provencio et al., 2000), giant oc-
topus (Octopus dofleini; OCTOPUS) RHO (CAA30644; Ovchinnikov et
al., 1988), Takifugu rubripes (FUGU) GREEN (AF226989), ROD
(AF201471), Atlantic salmon (Salmo salar; SALM), vertebrate an-
cient (VA; AAC60124; Soni and Foster, 1997), yesso scallop (Patin-
opecten yessoensis; SCALL) OP2 (O15974; Kojima et al., 1997), Xeno-
pus laevis (XENO) MEL (AF014797; Provencio et al., 1998), ROD
(P29403; Saha and Grainger, 1993), RED (U90895; Chang and Harris,
1998), UV (P51473; Starace and Knox, 1998), zebrafish (ZF) UV
(AF109373), BLUE (AF109372), GREEN1 (AF109369), GREEN2
(AAD24753), RED (AAD24754), ROD (AF109368; Vihtelic et al.,
1999), VA (BAA94289; Kojima et al., 2000), and Atlantic cod (Gadus
morhua; COD) Opn4a (AF385823), Opn4b (AY126448).
Phylogenetic analysis of various opsins. The phylogenetic
brain, and skin. RT-PCR analysis of total RNA purified from skin
(lanes 2, 6), retina (lanes 3, 7), brain (lanes 4, 8), and liver (lanes 5, 9).
The amplification reaction was performed after incubation with
(?RTase) or without (–RTase) reverse transcriptase. Gene-specific
primers amplified a 466 bp opn4a product (A) or a 673 bp opn4b
product (B). The arrowheads in lane 2 point to specific products.
Lanes 1 and 10 shows a 100 bp ladder, and numerical values indicate
the number of base pairs.
Atlantic cod opn4a and opn4b are expressed in retina,
TABLE 1. Amino Acid Identity (%) of Opn4a and Opn4b From Atlantic Cod With Melanopsins From Different Species
88Ø. DRIVENES ET AL.
various developmental stages. Retina from 1-day-old cod larvae (A–E)
contains cells positive for both opn4a and opn4b. Transversal sections
of the whole larvae (A) show opn4a-positive cells in the inner nuclear
layer, especially in horizontal cells (B), but staining in a few cells
resembling amacrine and ganglion cells is also observed. No staining
is observed with the opn4a sense probe (C). opn4b shows less expres-
sion in the inner nuclear layer than opn4a, but more ganglion cells are
stained (D,E). In 40-day-old cod larvae (F), clear staining of opn4a in
horizontal cells (arrow) and more diffuse staining around bipolar cells
are observed, although some bipolar cell bodies are stained (white
arrowhead). Some amacrine cells and ganglion cells are also express-
ing opn4a (black arrowheads). In 40-day-old larvae, no expression of
Melanopsin-expressing cells in the retina of Atlantic cod at
opn4b is found in the horizontal cell layer (G). opn4b Transcripts are
found in cells on both side of the inner plaxiform layer resembling
amacrine and ganglion cells (arrowheads and arrows, respectively). In
1-year-old juveniles (H–K), the array of opn4a-positive cells in the
region of horizontal cells is prominent (H,J), in addition to diffuse
staining surrounding bipolar cells. The mosaic organization of opn4a-
positive horizontal cells is apparent at this developmental stage (I).
opn4b expression is not found in horizontal cells (K) but in subpopu-
lations of cells in the ganglion and amacrine cell layers. Retina layers:
pigment epithelium with black melanin granula (PE), outer nuclear
layer (ONL), inner nuclear layer (INL), inner plexiform layer (IPL),
and ganglion cell layer (GCL). Scale bars ? 50 ?m in A,C, 20 ?m in
E,F,H,J, 10 ?m in B,D,G,I.
89MELANOPSINS IN TELEOSTS
The distribution of melanopsin-positive cells seems to
be uniform over the entire retina. The distance between
the opn4a-expressing cells in the horizontal cell layer in-
creases during development, however, a feature that be-
comes even more apparent when comparing juvenile and
adult Atlantic cod (data not shown).
No staining was observed in the pigment epithelium
surrounding the photoreceptor cells or in the iris pigment
(data not shown), where Xenopus melanopsin is known to
be expressed (Provencio et al., 1998). One cannot be too
positive about the lack melanopsin expression in the pig-
ment epithelium, however, insofar as the black melanin
granula could mask a weak color staining. Melanopsin
expression was not detected in the photoreceptor layer
containing cones and rods.
Distribution in the brain
The two melanopsins are expressed separately in the
supraoptic/suprachiasmatic nucleus (SOC; opn4a) and the
habenula (opn4b). These regions receive retinal and/or
pineal afferents and are nonvisual photic integration ar-
eas (Fig. 5; Holmqvist et al., 1994; Yanez and Anadon,
1996). opn4a expression is restricted to a small cell popu-
lation in the most lateral and rostral part of the SOC, as
described for the Atlantic salmon, Salmo salar, brain
(Holmqvist et al., 1994) close to the periventricular preop-
tic nucleus (NPP). A transverse section shows the
melanopsin-expressing cells along the lateral boundary of
the SOC, bordering the optic nerve (Fig. 5A). Parasagittal
sections show that the cells form a column extending
along the rostral-caudal axis (Fig. 5B) and more laterally
form a small cell cluster (Fig. 5C). opn4b expression is
localized to a discrete cell population extending ventrolat-
erally within the dorsal habenula (Fig. 5D). The sense
probes used as controls on adjacent sections showed no
staining in these cells (not shown).
The present study is the first description of melanopsin
genes in teleosts, providing new evolutionary information
on this gene family and the mRNA distribution within the
inner retina and brain. We used a degenerate PCR ap-
proach to isolate two melanopsin transcripts. The inferred
cod melanopsin proteins have several features in common
with homologues in the pufferfish Takifugu, Xenopus,
chicken, and mammals. One of these features is the strong
similarity to invertebrate opsins in both sequence identity
and critical amino acid substitutions. Hence, melanopsin
is grouped together with invertebrate opsins in the phy-
logenetic tree, suggesting a common ancestor for both
groups of genes. Although melanopsins are similar to in-
vertebrate opsins, for example, octopus rhodopsin, they
also have some distinct features. One of these is a long
cytoplasmic C-terminal tail, with a high number of phos-
phorylation sites, suggesting regulation by kinases (Prov-
encio et al., 1998, 2000).
Interestingly, phylogenetic analysis shows that melan-
opsins can be separated into two subgroups. Subgroup I
contains melanopsins from mammals and one melanopsin
from Takifugu, whereas subgroup II contains all the other
melanopsins, including the two cod genes reported here.
This grouping is supported by expression data. Tran-
scripts from subgroup I have been detected only in the
retina, whereas members in subgroup II are expressed in
retina and brain tissues, suggesting functional differences
as well. Hence it is likely that an ancient melanopsin has
been duplicated into two paralogous genes, which subse-
quently have acquired different characteristics. One of
these duplicates has been lost in the mammalian lineage,
whereas both genes have been kept in the fish lineage as
demonstrated by the presence of members from each sub-
group in Takifugu. Thus, it is tempting to speculate that
cod could have at least one melanopsin in subgroup I. The
existence of extra melanopsin members in cod and
Takifugu has probably been generated by a genome dupli-
cation event occurring in the lineage leading to modern
ray-finned fishes (Amores et al., 1998; Postlethwait et al.,
Melanopsin expression in the retina
Lack of expression in the cod photoreceptor cells of the
two melanopsins verifies the nonvisual nature of these
opsins in teleosts. This has been observed in other verte-
brates. Both the location and the cell morphology indicate
that opn4a is expressed in horizontal cells. Cone-
connecting types of horizontal cells have long axonal pro-
jections between bipolar and amacrine cells (Wagner,
1990). The diffuse staining of opn4a in the inner nuclear
layer may suggest that melanopsin mRNA is transported
out into these axons. Expression in horizontal cells has
also been described for Xenopus (Provencio et al., 1998).
The intense labelling surrounding the bipolar cells, par-
ticularly during early stages of development, makes it
hard to conclude that these cells express opn4a. In mam-
mals, melanopsin expression is found in a subpopulation
of cells in the ganglion and inner nuclear cell layers (Prov-
encio et al., 2000; Hatter et al., 2002), which is in agree-
ment with our findings in Atlantic cod. In particular, the
distribution of opn4b-expressing cells in the cod retina at
the late larval stage is highly consistent with that ob-
served in mammals. However, it is known that dislocated
amacrine and ganglion cells can be found on both sides of
the inner plexiform layer (Ota et al., 1999). Consequently,
we cannot be too definite about the classification of
melanopsin-expressing cells observed in the inner retina.
Even though we find opn4a-expressing cells in different
cell classes of the inner retina, expression in horizontal
cells is dominant. This is similar to the pattern found in
Xenopus. opn4b expression on the other hand is found only
in the inner nuclear layers, in cells resembling amacrine
cells, and in the ganglion cell layer of mature retina. Such
a pattern is also found in mammals. It is reasonable to
assume, based on these observations, that the two mela-
nopsins in teleosts represent two different lineages/types
of melanopsin, one horizontal cell layer dominated and the
other ganglion/amacrine cell layer dominated. This subdi-
vision is also reflected in higher vertebrates. The phyloge-
netic analysis confirms this subdivision into two groups.
Expression analysis of the three melanopsins from
Takifugu and analysis of the missing subgroup I type
melanopsin in cod are needed to verify further the com-
plexity in melanopsin family structure and expression in
the retina of teleosts.
Although functional photoreception of melanopsins has
not yet been demonstrated, the occurrence of an opsin-
type transcript in ganglion cell and inner nuclear layers
indicates that cell types other than conventional photore-
ceptors are photosensitive. In agreement with this, Hatter
et al. (2002) showed that melanopsin-expressing ganglion
90Ø. DRIVENES ET AL.
venile cod brain. Parasagittal tracing of the cod brain (top) illustrates
the level of the transverse section of opn4a in the SOC (A) and opn4b
in the habenula (D). The transverse images in A show in situ hybrid-
ization of the DIG-labelled melanopsin RNA probe (left) and adjacent
Nissl stained section (right) within the SOC. opn4a expression in the
SOC is also shown in parasagittal sections (boxed) through the SOC
Differential distributions of melanopsin mRNA in the ju-
(B) and more laterally (C). A transverse section demonstrates opn4b
expression in the dorsal habenula (D). BS, brainstem; Cb, cerebellum;
Hb, habenula; NPP, periventricular preoptic nucleus; OC, optic chi-
asm; ON, optic nerve; OT, optic tectum; Pit, pituitary; SOC,
supraoptic/suprachiasmatic nucleus; SV, saccus vasculosus; T, telen-
cephalon, V, ventricle. Scale bars ? 20 ?m in A, 10 ?m in B,C, 50 ?m
cells in rat are photosensitive. However, ganglion cells
also express two blue-light photoreceptors, cryptochromes
1 and 2 (Miyamoto and Sancar, 1998), which might be
involved in this photoreception. Hence, several light-
detecting molecules may play important roles in the crit-
ical regulation of circadian rhythms in both fish and mam-
What function melanopsin may have in the visual pro-
cess through its expression in the horizontal cells is cur-
rently not known. It has been shown, however, that am-
bient light conditions affect the morphology of synaptic
elements within the cone pedicle and modulate the spatial
properties of the horizontal cells and that dopamine and
retinoic acid are important neurochemical signals in these
processes (Weiler et al., 1998, 2000). One could imagine
that these mechanisms integrate light detection in the
horizontal cells directly by melanopsin photosensitivity.
In addition, the larval cod retina consists only of cone
cells, with rod cells appearing later in development. We
find that melanopsin is expressed in horizontal cells in the
pure cone retina, which indicates a function related to
cone vision physiology rather than balancing rod-cone in-
puts to the horizontal cells.
The pattern of opn4a expression in the inner nuclear
and ganglion cell layers in the cod retina matches the
expression pattern of VA opsins found in the retina of
salmon and zebrafish (Soni et al., 1998; Kojima et al.,
2000). This raises an interesting question: Do retinal cells
in teleosts express two types of opsins in the same cell or
in different cell subpopulations? Further studies on the
function of such a photomechanism in the inner nuclear
layer of the retina may give new perspectives on visual
processes and light adaptation. The term nonvisual pig-
ment could be inappropriate for melanopsins if it turns out
that melanopsins contribute to image processing down-
stream of conventional photoreceptors.
Distribution in the brain
This is the first description of melanopsin-expressing
cells in the teleost brain. We demonstrate that two mela-
nopsins, opn4a and opn4b, are expressed in the SOC and
the habenula, respectively. These regions have not been
identified previously in teleosts as opsin-like molecule-
containing areas. In salmon, the SOC is known to receive
retinal and pineal projections and contains hypophysiotro-
phic dopaminergic neurons (Holmqvist et al., 1994), sug-
gesting a role in circadian rhythms and hypophysiotrophic
regulation. In Xenopus, cells expressing melanopsin have
been found in a similar structure, the suprachiasmatic
nucleus (Provencio et al., 1998). In recent years, other
opsins (cone and VA) and phototransduction molecules
(?-transducin) have been identified in deep brain struc-
tures in the Atlantic salmon (Philp et al., 2000). The
?-transducin-immunoreactive cells in the salmon caudal
and ventral telencephalon are located similarly but more
rostrally than the opn4a-expressing cells we show here in
the cod brain. However, these cells in the salmon have not
been shown to contain opsin molecules (Philp et al., 2000),
suggesting that another opsin, yet to be identified, may
reside in these cells. Melanopsin may be that opsin, but
further characterization of melanopsin and the cells that
express it is necessary before a possible hypophysiotrophic
and/or circadian rhythm function can be attributed to
The expression of opn4b in the habenula suggests a role
in the mediation of responses to photic cues within an
integration zone, which receives and projects to numerous
brain regions (Yanez and Anadon, 1996). Further studies,
however, are needed to assign a specific role to melanopsin
in the habenula.
Expression of cod melanopsins in skin
Melanopsin was initially isolated from light-sensitive
pigment cells (melanophores) in Xenopus (Provencio et al.,
1998). Here we show that both opn4a and opn4b are
expressed in the epidermis/dermis of Atlantic cod, though
at low levels compared with retina and brain (Fig. 3). In
situ hybridization analysis on skin samples did not reveal
any clear expression above the background level for opn4a
and opn4b (not shown). This suggests a low level of mela-
nopsin expression in the fish skin (Fig. 3). Earlier studies
have shown that light may have a primary effect directly
on pigment cells in teleosts that may indicate that photo-
receptors do exist in these cells (Oshima and Yokozeki,
1999; Fujii, 2000; Oshima, 2001). Our analysis suggests
that melanopsin might be involved in the regulation of
these responses, but further analysis is needed to identify
the cellular location of such light detectors.
In conclusion, the data presented here add to the com-
plexity of the light-detection systems in teleosts. The dis-
tribution of melanopsin in the same region as VA opsin in
the retina and in different regions in the brain may indi-
cate both shared and specialized functions for these opsins
in the regulation of photoentrainment in fish. However,
further genetic and physiological analysis is needed to
elucidate the true roles of these putative light detectors in
The Takifugu rubripes data have been provided freely
by the Fugu Genome Consortium for use in this article
only. We thank Dr. Frank Nilsen for helpful suggestions
in making the phylogenetic tree and Hilde Killingstad for
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local
alignment search tool. J Mol Biol 215:403–410.
Amores A, Force A, Yan YL, Joly L, Amemiya C, Fritz A, Ho RK, Langeland
J, Prince V, Wang YL, Westerfield M, Ekker M, Postlethwait JH. 1998.
Zebrafish hox clusters and vertebrate genome evolution. Science 282:
Barthel LK, Raymond PA. 1993. Subcellular localization of alpha-tubulin
and opsin mRNA in the goldfish retina using digoxigenin-labeled cRNA
probes detected by alkaline phosphatase and HRP histochemistry.
J Neurosci Methods 50:145–152.
Berson DM, Dunn FA, Takao M. 2002. Phototransduction by retinal gan-
glion cells that set the circadian clock. Science 295:1070–1073.
Blackshaw S, Snyder SH. 1997. Parapinopsin, a novel catfish opsin local-
ized to the parapineal organ, defines a new gene family. J Neurosci
Bowmaker JK, Hunt DM. 1999. Molecular biology of photoreceptor spectral
sensitivity. In: Archer SN, Djamgoz MBA, Loew ER, Partridge JC,
Vallerga S, editors. Adaptive mechanisms in the ecology of vision.
Dordrecht: Kluwer Academic Publishers. p 439–462.
Chang WS, Harris WA. 1998. Sequential genesis and determination of cone
and rod photoreceptors in Xenopus. J Neurobiol 35:227–244.
Cowing JA, Poopalasundaram S, Wilkie SE, Bowmaker JK, Hunt DM.
2002. Spectral tuning and evolution of short wave-sensitive cone pig-
ments in cottoid fish from Lake Baikal. Biochemistry 41:6019–6025.
92Ø. DRIVENES ET AL.
Freedman MS, Lucas RJ, Soni B, von Schantz M, Munoz M, David-Gray Z,
Foster R. 1999. Regulation of mammalian circadian behavior by non-
rod, non-cone, ocular photoreceptors. Science 284:502–504.
Fujii R. 2000. The regulation of motile activity in fish chromatophores.
Pigment Cell Res 13:300–319.
Gartner W, Towner P. 1995. Invertebrate visual pigments. Photochem
Gooley JJ, Lu J, Chou TC, Scammell TE, Saper CB. 2001. Melanopsin in
cells of origin of the retinohypothalamic tract. Nat Neurosci 4:1165.
Hannibal J, Hindersson P, Knudsen SM, Georg B, Fahrenkrug J. 2002.
The photopigment melanopsin is exclusively present in pituitary ade-
nylate cyclase-activating polypeptide-containing retinal ganglion cells
of the retinohypothalamic tract. J Neurosci 22:RC191.
Hatter S, Liao HW, Takao M, Berson DM, Yau KW. 2002. Melanopsin-
containing retinal ganglion cells: architecture, projections, and intrin-
sic photosensitivity. Science 295:1065–1070.
Helvik JV, Drivenes O˚, Næss T, Fjose A, Seo HC. 2001. Molecular cloning
and characterization of five opsin genes from the marine flatfish At-
lantic halibut (Hippoglossus hippoglossus). Vis Neurosci 18:767–780.
Holmqvist BI, O¨stholm T, Ekstro ¨m P. 1994. Neuroanatomical analysis of
the visual and hypophysiotrophic systems in Atlantic salmon (Salmo
salar) with emphasis on possible mediators of photoperiodic cues dur-
ing parr–smolt transformation. Aquaculture 121:1–12.
Hunt DM, Fitzgibbon J, Slobodyanyuk SJ, Bowmaker JK, Dulai KS. 1997.
Molecular evolution of the cottoid fish endemic to Lake Baikal deduced
from nuclear DNA evidence. Mol Phylogenet Evol 8:415–422.
Hunt DM, Dulai KS, Partridge JC, Cottrill P, Bowmaker JK. 2001. The
molecular basis for spectral tuning of rod visual pigments in deep-sea
fish. J Exp Biol 204:3333–3344.
Kawamura S. 1995. Photransduction, excitation and adaptation. In:
Djamgoz MBA, Archer SN, Vallerga S, editors. Neurobiology and clin-
ical aspects of the outer retina. London: Chapman and Hall. p 105–131.
Kojima D, Terakita A, Ishikawa T, Tsukahara Y, Maeda A, Shichida Y.
1997. A novel Go-mediated phototransduction cascade in scallop visual
cells. J Biol Chem 272:22979–22982.
Kojima D, Mano H, Fukada, Y. 2000. Vertebrate ancient-long opsin: a
green-sensitive photoreceptive molecule present in zebrafish deep
brain and retinal horizontal cells. J Neurosci 20:2845–2851.
Kozak M. 1987. At least six nucleotides preceding the AUG initiator codon
enhance translation in mammalian cells. J Mol Biol 196:947–950.
Lucas RJ, Freedman MS, Munoz M, Garcia-Fernandez JM, Foster RG.
1999. Regulation of the mammalian pineal by non-rod, non-cone, ocular
photoreceptors. Science 284:505–507.
Miyamoto Y, Sancar A. 1998. Vitamin B2-based blue-light photoreceptors
in the retinohypothalamic tract as the photoactive pigments for setting
the circadian clock in mammals. Proc Natl Acad Sci USA 95:6097–
Nathans J, Hogness DS. 1984. Isolation and nucleotide sequence of the
gene encoding human rhodopsin. Proc Natl Acad Sci USA 81:4851–
Nathans J, Piantanida TP, Eddy RL, Shows TB, Hogness DS. 1986. Mo-
lecular genetics of inherited variation in human color vision. Science
Okano T, Yoshizawa T, Fukada Y. 1994. Pinopsin is a chicken pineal
photoreceptive molecule. Nature 372:94–97.
Oshima N. 2001. Direct reception of light by chromatophores of lower
vertebrates. Pigment Cell Res 14:312–319.
Oshima N, Yokozeki A. 1999. Direct control of pigment aggregation and
dispersion in Tilapia erythrophores by light. Zool Sci 16:51–54.
Ota D, Downing JEG, Cook JE. 1999. Neuronal and glial cell types re-
vealed by NADPH-diaphorase histochemistry in the retina of a teleost
fish, the grass goby (Zosterisessor ophiocephalus, Perciformes, Gobi-
idae). Anat Embryol 200:487–494.
Ovchinnikov YA, Abdulaev NG, Zolotarev AS, Artamonov ID, Bespalov IA,
Dergachev AE, Tsuda M. 1988. Octopus rhodopsin. Amino acid se-
quence deduced from cDNA. FEBS Lett 232:69–72.
Philp AR, Garcia-Fernandez JM, Soni BG, Lucas RJ, Bellingham J, Foster
RG. 2000. Vertebrate ancient (VA) opsin and extraretinal photorecep-
tion in the Atlantic salmon (Salmo salar). J Exp Biol 203:1925–1936.
Postlethwait JH, Yan YL, Gates MA, Horne S, Amores A, Brownlie A,
Donovan A, Egan ES, Force A, Gong Z, Goutel C, Fritz A, Kelsh R,
Knapik E, Liao E, Paw B, Ransom D, Singer A, Thomson M, Abdul-
jabbar TS, Yelick P, Beier D, Joly JS, Larhammar D, Rosa F, et al.
1998. Vertebrate genome evolution and the zebrafish gene map. Nat
Provencio I, Jiang G, De-Grip W, Hayes W, Rollag M. 1998. Melanopsin: an
opsin in melanophores, brain, and eye. Proc Natl Acad Sci USA 95:
Provencio I, Rodriguez IR, Jiang GS, Hayes WP, Moreira EF, Rollag MD.
2000. A novel human opsin in the inner retina. J Neurosci 20:600–605.
Saha MS, Grainger RM. 1993. Early opsin expression in Xenopus embryos
precedes photoreceptor differentiation. Brain Res Mol Brain Res 17:
Seo HC, Drivenes Ø, Ellingsen S, Fjose A. 1998. Expression of two ze-
brafish homologues of the murine Six3 gene demarcates the initial eye
primordia. Mech Dev 73:45–57.
Shand J, Foster RG. 1999. The extraretinal photoreceptors of non-
mammalian vertebrates. In: Archer SN, Djamgoz MBA, Loew ER,
Partridge JC, Vallerga S, editors. Adaptive mechanisms in the ecology
of vision. Dordrecht: Kluwer Academic Publisher. p 197–222.
Soni BG, Foster RG. 1997. A novel and ancient vertebrate opsin. FEBS
Soni BG, Philp AR, Foster RG, Knox BE. 1998. Novel retinal photorecep-
tors. Nature 394:27–28.
Starace DM, Knox BE. 1998. Cloning and expression of a Xenopus short
wavelength cone pigment. Exp Eye Res 67:209–220.
Strimmer K, von Haeseler A. 1997. Likelihood-mapping: a simple method
to visualize phylogenetic content of a sequence alignment. Proc Natl
Acad Sci USA 94:6815–6819.
Terakita A, Yamashita T, Nimbari N, Kojima D, Shichida Y. 2002. Func-
tional interaction between bovine rhodopsin and G protein transducin.
J Biol Chem 277:40–46.
Vihtelic TS, Doro CJ, Hyde DR. 1999. Cloning and characterization of six
zebrafish photoreceptor opsin cDNAs and immunolocalization of their
corresponding proteins. Vis Neurosci 16:571–585.
Wagner HJ. 1990. Retinal structure of fishes. In: Douglas RH, Djamgoz
MBA, editors. The visual system in fish. London: Chapman and Hall. p
Weiler R, Schultz K, Pottek M, Tieding S, Janssen-Bienhold U. 1998.
Retinoic acid has light-adaptive effects on horizontal cells in the retina.
Proc Natl Acad Sci USA 95:7139–7144.
Weiler R, Pottek M, He SG, Vaney DI. 2000. Modulation of coupling
between retinal horizontal cells by retinoic acid and endogenous dopa-
mine. Brain Res Rev 32:121–129.
Yanez J, Anadon R. 1996. Afferent and efferent connections of the habe-
nula in the rainbow trout (Oncorhynchus mykiss): an indocarbocyanine
dye (DiI) study. J Comp Neurol 372:529–543.
93MELANOPSINS IN TELEOSTS