Two families of candidate taste receptors in fishes

Mechanisms of Development 122 (2005) 1310–1321 www.elsevier.com/locate/modo

Two families of candidate taste receptors in fishes Yoshiro Ishimarua,*,1, Shinji Okadaa,1, Hiroko Naitoa, Toshitada Nagaia, Akihito Yasuokab, Ichiro Matsumotoa, Keiko Abea,* a

Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan b Laboratory of Reproductive and Developmental Toxicology, National Institute of Environmental Health Sciences, National Institute of Health, Research Triangle Park, NC 27709, USA Received 9 May 2005; received in revised form 24 July 2005; accepted 26 July 2005 Available online 30 August 2005

Abstract Vertebrates receive tastants, such as sugars, amino acids, and nucleotides, via taste bud cells in epithelial tissues. In mammals, two families of G protein-coupled receptors for tastants are expressed in taste bud cells—T1Rs for sweet tastants and umami tastants (L-amino acids) and T2Rs for bitter tastants. Here, we report two families of candidate taste receptors in fish species, fish T1Rs and T2Rs, which show significant identity to mammalian T1Rs and T2Rs, respectively. Fish T1Rs consist of three types: fish T1R1 and T1R3 that show the highest degrees of identity to mammalian T1R1 and T1R3, respectively, and fish T1R2 that shows almost equivalent identity to both mammalian T1R1 and T1R2. Unlike mammalian T1R2, fish T1R2 consists of two or three members in each species. We also identified two fish T2Rs that show low degrees of identity to mammalian T2Rs. In situ hybridization experiments revealed that fish T1R and T2R genes were expressed specifically in taste bud cells, but not in olfactory receptor cells. Fish T1R1 and T1R2 genes were expressed in different subsets of taste bud cells, and fish T1R3 gene was co-expressed with either fish T1R1 or T1R2 gene as in the case of mammals. There were also a significant number of cells expressing fish T1R2 genes only. Fish T2R genes were expressed in different cells from those expressing fish T1R genes. These results suggest that vertebrates commonly have two kinds of taste signaling pathways that are defined by the types of taste receptors expressed in taste receptor cells. q 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Taste receptor; Fish; Taste bud; Phospholipase C-b2

1. Introduction In the vertebrate taste system, environmental chemical compounds are received by taste buds, cell groups of epithelial origin (Stone et al., 1995), and the resulting signals are transmitted to the central nervous system through taste nerves innervating the taste bud cells. In mammals, two families of G protein-coupled receptors (GPCRs), named T1R and T2R families, are expressed in taste bud cells, and are responsible for direct interaction with tastants at the initial step of taste perception (Nelson et al., 2001, 2002; Chandrashekar et al., 2000; Zhao et al., 2003). The * Corresponding authors. Tel.: C81 3 5841 5129; fax: C81 3 5841 8006. E-mail addresses: [email protected] (Y. Ishimaru), [email protected] (K. Abe). 1 These authors contributed equally to this work.

0925-4773/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2005.07.005

mammalian T1Rs consist of three members, T1R1, T1R2, and T1R3, and function as heteromers. The T1R1/T1R3 heteromer is activated by L -amino acids, including glutamate as a representative umami tastant (Nelson et al., 2002; Zhao et al., 2003). The T1R2/T1R3 heteromer is activated by various kinds of sweet compounds, such as sugars, artificial sweeteners, and sweet proteins (Nelson et al., 2001; Zhao et al., 2003). The mammalian T2Rs consist of approximately 30 members, which are more divergent in primary structure than the T1Rs. They were originally identified as GPCR gene clusters in a genetic locus responsible for bitter taste sensation (Adler et al., 2000; Matsunami et al., 2000). T2Rs are expressed in a subset of taste bud cells that are different from those expressing T1Rs (Nelson et al., 2001), and activated by bitter compounds including denatonium benzoate and cycloheximide (Chandrashekar et al., 2000). Based on the

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

tastant-responsive profiles and the expression patterns of mammalian taste receptors described above, it was proposed that the discrimination of taste modality, such as sweetness and bitterness, is dependent on the properties of taste receptor cells (Zhao et al., 2003; Mueller et al., 2005). Both T1Rs and T2Rs require G proteins, effector enzymes, and other downstream components to activate intracellular signaling. Two subtypes of G protein a subunit (Ggust and Gi2), phospholipase C (PLC-b2), and ion channel (TRPM5) are expressed in taste bud cells (McLaughlin et al., 1992; Kusakabe et al., 2000; Rossler et al., 1998; Perez et al., 2002; Zhang et al., 2003). Fish species respond to water-soluble chemical compounds, such as amino acids, organic acids, sugars, nucleotides, and bile acids, with higher sensitivity than mammals do (Hara, 1994). In fishes, the taste buds are distributed in the lips, gill rakers, pharynx, oral cavity, and also on the body surface with some variations depending on the species. However, the organizations of taste buds and taste nerves along the anterior–posterior axis are mainly conserved among fish species and are comparable to those of mammals (Puzdrowski, 1987; Kotrschal, 2000). Previously, we reported that PLC-b2 gene is expressed specifically in taste buds of the two fish species, pond loach (Misgurnus anguillicaudatus) and medaka fish (Oryzias latipes), in a manner similar to those of mammals (Asano-Miyoshi et al., 2000; Yasuoka et al., 2004). This observation suggested that at least at the level of effector molecules, vertebrates share a common molecular component in taste signal transduction. In reference to the olfactory system in fishes as the other chemosensory system, two families of odorant receptors, which correspond to mammalian odorant receptors (ORs) or putative pheromone receptors (V2Rs), have been shown to be expressed in the olfactory receptor cells of several fish species (Ngai et al., 1993; Speca et al., 1999). These findings prompted us to search for fish taste receptor families homologous to mammalian T1Rs or T2Rs. We searched for candidate fish taste receptors in genome databases by a homology-based method, and identified two families of candidate taste receptors, fish T1Rs and T2Rs, which showed significant identity to each of the mammalian cognate taste receptor families. In addition, fish T1R and T2R genes were expressed in different subsets of taste bud cells, similarly to mammalian T1R and T2R genes. These results suggested that the two kinds of taste signaling pathway, each of which is dependent on the kind of taste receptors expressed by taste receptor cells, are common to vertebrate species.

2. Results 2.1. Search for fish T1R genes in genome databases To identify genomic sequences encoding proteins homologous to mammalian T1Rs, we first searched in the

1311

genome database of puffer fish (Fugu rubripes), because its genome is smaller in size and annotated better than the other fish species we examined (Venkatesh et al., 2000). Four sequences were identified as genes encoding fugu T1Rs that showed the highest degrees of identity to mammalian T1Rs (Fig. 1(A)). In this process, we also found ff5.24 gene as an ortholog of the goldfish olfactory receptor 5.24 gene (Speca et al., 1999) and ffCaSR genes as orthologs of human calcium sensing receptor genes, both of which are distantly related to mammalian T1R genes (Fig. 1(B)). We next searched for candidate fish T1R genes in the genome databases of two other fish species, zebrafish (Danio rerio) and medaka fish (Oryzias latipes). Using fugu T1R genes as references, four and five T1R genes were identified in the genomes of zebrafish and medaka fish, respectively (Fig. 1(A)). The three zebrafish T1R genes, zfT1R1, zfT1R2a, and zfT1R3, were cloned as cDNAs containing the entire respective coding regions (Fig. 1(A)). As to the genomic structure of fish T1R genes, most of them consisted of six coding exons, while zfT1R1, ffT1R3, and zfT1R3 consisted of seven, eight, and nine coding exons, respectively. Fish T1R1s showed the highest degrees of amino acid identity (39–43%) to mammalian T1R1s. Fish T1R1s showed 56–67% identity among different fish species. Fish T1R3s showed the highest degrees of identity (34–37%) to mammalian T1R3s. Fish T1R3s showed 51– 58% identity among different fish species. On the other hand, fish T1R2s showed almost equal degrees of identity (31–34%) to both mammalian T1R1s and T1R2s. Fish T1R2s showed 42–62% identity among different fish species. Phylogenetic tree drawn using CLUSTAL W classified four groups of vertebrate T1Rs: vertebrate T1R1s (blue circle), vertebrate T1R3s (green circle), fish T1R2s (red circle), and mammalian T1R2s (orange circle) (Fig. 1(B)). Similar branching patterns were observed when the N-terminal extracellular domains and the C-terminal transmembrane domains were analyzed independently (data not shown). We found that fish and mammalian T1R1s branched off from the node distant from that between vertebrate T1R1s and T1R3s, as well as fish and mammalian T1R3s. On the other hand, fish and mammalian T1R2s trichotomously branched off from the node close to that between vertebrate T1R1s and T1R3s. These results suggested that fish T1R1s and T1R3s are orthologous to mammalian T1R1s and T1R3s, respectively, and that fish T1R2s are not orthologous to mammalian T1R2s. Each of the three fish species investigated here has two or three fish T1R2 genes in the genome. They showed 61– 71% identity in each species, and were located tandemly within a region of 10–20 kb in each genome with the exception of mfT1R2c. Some partial genomic sequences encoding proteins that show high degrees of identity (ca. 60%) to fish T1R2s were also found in the zebrafish and medaka fish genome databases (data not shown), although they had incomplete coding regions due to insertions or

1312

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

A zfT1R1

mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 T M 1

T M 2

T M 3

T M 4

zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 T M 5

zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3 zfT1R1 mfT1R1 ffT1R1 hT1R1 zfT1R2a mfT1R2a ffT1R2a hT1R2 zfT1R3 mfT1R3 ffT1R3 hT1R3

T M 6

T M 7

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

1313

B

hT1R2 mT1R2 zfT1R1 hT1R1 mT1R1

zfT1R2a zfT1R2b

ffT1R1

ffT1R2b

mfT1R1

mfT1R2c mT1R3 100

56

hT1R3

mfT1R2a mfT1R2b

25

ffT1R3 91

42 ffT1R2a

mfT1R3 66 38

zfT1R3

hCaSR zfCaSR ffCaSR

ff5.24 ffCa02.1

mV2R4

gf5.24

zf5.24

10%

mV2R2 Fig. 1 (continued)

3 Fig. 1. Identification of fish T1R genes. (A) Deduced amino acid sequences of zebrafish, medaka fish, fugu, and human T1R genes. The transmembrane domains are indicated by bars above the sequences. The black boxes indicate identical residues among at least seven of the twelve sequences. The accession numbers of fish T1R genes are as follows: zfT1Rs, AB200899 to AB200902; mfT1Rs, AB200905 to AB200909; and ffT1Rs, AB200910 to AB200913. (B) Phylogenetic tree showing fish T1Rs and related receptors. The tree was constructed on the basis of amino acid sequence alignments by the neighbor-joining method. Fish T1Rs were categorized into three groups: a group including mammalian T1R1s (blue circle), a group including mammalian T1R3s (green circle), and a group independent of any vertebrate T1R (red circle). zf5.24 and ff5.24 genes were orthologs of goldfish olfactory receptor 5.24 gene, and zfCaSR and ffCaSR genes were orthologs of human calcium sensing receptor (hCaSR) genes. ffCa02.1 is a fugu receptor distantly related to the mouse vomeronasal receptors, mV2R2 and mV2R4 (Naito et al., 1998). The numbers at nodes are bootstrap support values based on 1000 bootstrap replicates. The bootstrap value at the node where fish T1R2s, mammalian T1R2s, and the other receptors branch cannot be calculated because of its trichotomy. The bootstrap value (25%) indicated near the trichotomous node is for the node where T1R1s and other receptors including vertebrate T1R3s branch. The scale bar indicates 10% amino acid difference.

1314

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

A

T

M

1

zfT2R1a zfT2R1b ffT2R1 T

M

2

T

M

3

zfT2R1a zfT2R1b ffT2R1 T

M

4

zfT2R1a zfT2R1b ffT2R1 T

M

5

zfT2R1a zfT2R1b ffT2R1 T

M

6

T

M

7

zfT2R1a zfT2R1b ffT2R1 zfT2R1a zfT2R1b ffT2R1

orthologous groups

B

mT2R19 hT2R1

mT2R8 hT2R4

ffT2R1

mT2R45

mouse-specific groups

mT2R5 hT2R7

mT2R62 mT2R54

zfT2R1a zfT2R1b

hT2R14 hTAS2R44 hTAS2R43

human-specific group

94

mT2R52 97

mV1RD8 (mV3R)

65 mV1RE9 zfV1R1

mV1RA1

ffV1R1

10%

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

deletions. Their transcripts were not detected in taste bud cells (data not shown) indicating that they are pseudogenes. 2.2. Search for fish T2R genes in genome databases Next, we aimed for the identification of fish genes encoding proteins homologous to mammalian T2Rs. We searched in the fugu genome database using nineteen mammalian T2Rs as references (accession numbers are listed in Section 4.1). The sequences encoding GPCRs with known function were excluded, and the remaining sequences were examined on the basis of BLAST scores and the presence of continuous reading frames (more than 900 bases). As a result, three fugu T2R-like genes, ff157, ff997-1, and ff997-2, were identified (named after scaffold numbering where they were located). In addition, ff767 was obtained by reiterative searches. Using these fugu T2R-like genes as references, we next searched in the genome databases of zebrafish and medaka fish. Several sequences similar to fugu T2R-like genes were found in zebrafish genome, and named after fugu T2Rlike genes as follows: zf157-1 to 2, zf997-1 to 5, and zf767-1 to 3. However, we could not find any T2R-like sequence in the medaka fish genome. Before further analysis, a preliminary in situ hybridization analysis was performed using probes for zebrafish T2R-like sequences. Among them, probes for zf157-1 and zf157-2 gave significant signals in gustatory tissues. Then we designated zf157-1, 2, and ff157 as zfT2R1a, b, and ffT2R1, respectively (Fig. 2(A)). Fish T2R genes had no introns in their coding regions as in the case of mammalian T2R genes. Fish T2Rs showed 13–22% identity to mammalian T2Rs. These values are close to the values, 13–20%, observed between fish T2Rs and vertebrate vomeronasal receptors (Conte et al., 2003; Shi et al., 2003; Pfister and Rodriguez, 2005), which are distantly related to mammalian T2Rs (Fig. 2(B)). These phylogenetic relationships are represented as fish T2R branches isolated from those of mammalian T2Rs and vertebrate V1Rs. Zebrafish T2R1 s showed 33% identity to fugu T2R1. The value between the two zebrafish T2R members was much higher (80%), and these genes were located in tandem within a region of 4 kb of the 9th chromosome in the zebrafish genome.

1315

2.3. Expression of fish T1R and T2R genes in the gustatory tissues To obtain information about the functions of fish T1Rs and T2Rs, we investigated the tissue distributions of the transcripts by in situ hybridization analysis (Fig. 3). All of the fish T1R and T2R genes in zebrafish were expressed in taste bud cells in lips, gill rakers, and pharynx (panels in Fig. 3). Similar expression patterns were observed in the case of mfT1R genes (data not shown). In these tissues, taste buds are observed as onion-like shaped cell clusters, where PLC-b2 is expressed as a common component of vertebrate taste signal transduction (Asano-Miyoshi et al., 2000; Yasuoka et al., 2004). For zfT1Rs and zfT2Rs, signal frequencies were lower than that for zfPLC-b2 (panels in zfPLC-b2 row in Fig. 3). On the other hand, no signals for zfT1Rs or zfT2Rs were detected in the olfactory epithelium (panels in olfactory epithelium column in Fig. 3), where zf5.24, an ortholog of the goldfish 5.24 gene, is expressed (panels in zf5.24 row in Fig. 3). These results indicated that zfT1R and zfT2R genes are expressed specifically in taste bud cells. Next, to determine which taste bud cells express fish T1R and T2R genes, double-labeling in situ hybridization analysis was carried out using probes for the zfTRs and zfPLC-b2. Both zfT1R3 and zfT2R1a were co-expressed with zfPLC-b2 as demonstrated by the yellow signals in Fig. 4(C) and (F). All the other zfTR members were coexpressed with zfPLC-b2 as well (data not shown). In addition, zfT1R and zfT2R expression were restricted to a subset of zfPLC-b2-positive cells. These results strongly suggested that these candidate fish T1Rs and T2Rs function in taste reception. 2.4. Relationships in expression among fish taste receptor genes In mammals, taste receptor genes are expressed in several patterns, which provide each taste bud cell with a certain tastant responsive profile. Therefore, the expression pattern of fish taste receptor genes can be a clue to elucidating their functions. We investigated the relationships in expression among zfTR genes by triple-labeling in situ hybridization analyses using zfPLC-b2-positive cells as a main population. With regard to the relationships among zfT1Rs, we found that zfT1R1 and zfT1R2s were expressed in a mutual

3 Fig. 2. Identification of fish T2R genes. (A) Deduced amino acid sequences of zebrafish and fugu T2R genes. The transmembrane domains are indicated by bars above the sequences. The black boxes indicate identical residues among at least two of the three sequences. The accession numbers of fish T2R genes are as follows: zfT2R1 s, AB200903 and AB200904; and ffT2R1, AB200914. (B) Phylogenetic tree showing fish T2Rs, mammalian T2Rs, and related receptors. The tree was constructed on the basis of amino acid sequence alignments by the neighbor-joining method. The fish T2Rs showed 13–22% identity to mammalian T2Rs, which are close to 11–20% observed between fish T2Rs and vertebrate vomeronasal receptors. Fish T2Rs (red circles) constitute fish species-specific groups, similar to mouse- or human-specific groups (blue circles). Some of the mammalian T2Rs showed orthologous relationships between human and mouse (green circles). For example, both hT2R4 and mT2R8 are receptors for denatonium benzoate. The numbers at nodes are bootstrap support values based on 1000 bootstrap replicates. The bootstrap value at the node where an orthologous group of mammalian T2Rs (hT2R4 and mT2R8), other mammalian T2Rs, and the others including fish T2Rs branch cannot be calculated because of its trichotomy. The scale bar indicates 10% amino acid difference.

1316

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

lip

gill raker

pharynx

olfactory epithelium

zfT1R1

zfT1R2a

zfT1R2b

zfT1R3

zfT2R1a

zfT2R1b

zfPLC-β2

zf5.24

Fig. 3. Expression of zfT1R and zfT2R genes in gustatory tissues. Tissues containing taste buds were sectioned, and hybridized with digoxigenin-labeled riboprobes for each of the zfT1Rs and zfT2Rs as indicated on the left. The transcripts of zfT1R1, zfT1R2a, zfT1R2b, zfT1R3, zfT2R1a, and zfT2R1b were detected in the taste buds of the lips, gill rakers, and pharynx. The frequencies of zfT1R- and zfT2R-positive cells were lower than that of zfPLC-b2-positive cells. No signals for zfT1Rs or zfT2Rs were detected in the olfactory epithelium, while signals for zf5.24, an ortholog of goldfish olfactory receptor 5.24 gene, were detected in the adjacent section. The black label in olfactory epithelium hybridized with probe for zfT2R1a is a cluster of endogenous pigments. Scale bars at the bottom of each columnZ10 mm.

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

1317

Fig. 4. Co-expression of zfTRs with zfPLC-b2. Sections of pharyngeal region were hybridized with zfTR and zfPLC-b2 probes. The signals of probes were represented as pseudocolors (green for zfTRs and red for zfPLC-b2). The cells positive to zfT1R3 (A) or zfT2R1a (D) were included in the cells positive to zfPLC-b2 (B and E) as indicated by the yellow signals in merged images (C and F). Scale barZ10 mm.

exclusive manner (green and red signals in Fig. 5(A) and (B)). On the other hand, most of zfT1R1 signals co-localized with zfT1R3 (Fig. 5(C)), although there were a small number of cells expressing only zfT1R1 or only zfT1R3 (green and red signals in Fig. 5(C)). The two members of zfT1R2s were partially co-expressed with each other (Fig. 5(D)). They were also partially co-expressed with zfT1R3 (Fig. 5(E) and (F)). However, the number of cells co-expressing zfT1R2s and zfT1R3 was smaller than that of cells co-expressing zfT1R1 and zfT1R3. In addition, there were a significant number of cells expressing only zfT1R2 members. In summary, zfT1R3-positive cells partially co-expressed other zfT1Rs as the largest subsets among zfT1Rs-positive cells (Fig. 5(K)). Next, we examined the relationships between zfT1Rs and zfT2Rs. We chose zfT2R1a as a representative zfT2R, because the zfT2R1b-positive cells were included in zfT2R1a-positive cells (Fig. 5(J)). As shown in Fig. 5(G) and (H), signals for zfT2R1a were clearly separated from those for zfT1Rs. The same result was obtained when the mixed probes for zfT1Rs and zfT2Rs were used (Fig. 5(I)). These results were summarized as Venn diagram (Fig. 5(M)). The zfT1R1-positive cells and zfT1R2s-positive cells belonged to distinct subsets. Most of zfT1R3-positive cells co-expressed other zfT1Rs. There were a small but significant number of cells expressing only zfT1R3 (Fig. 5(K)). These kinds of cells were also observed in mammals. In addition, we found cells expressing two zfT1R2s, zfT1R2a/2b, and three zfT1Rs, zfT1R2a/2b/3 (yellow signals in Fig. 5(L)), which could be a specific

feature of zebrafish taste bud cells. The most notable is the segregation of two kinds of cells, zfT1Rs-positive cells and zfT2Rs-positive cells. This may be a common feature of vertebrate taste bud cells.

3. Discussion In the present study, we identified fish genes encoding candidate taste receptors that are co-expressed with PLCb2, a general component of taste signal transduction in vertebrate taste receptor cells. This is the first evidence indicating that G protein-coupled receptors may be responsible for chemoreception in the fish taste system. 3.1. Molecular evolution of fish T1R genes Fish T1R1s and T1R3s showed the highest degrees of identity to mammalian T1R1s and T1R3s, respectively. This suggests that the divergence of vertebrate T1R1 and T1R3 genes occurred prior to the divergence of fish and mammalian species. The attribution of fish T1R2s was more ambiguous. They trichotomously branched off from the node, which is close to that of separation of vertebrate T1R1 and T1R3 branches (Fig. 1(B)). It is uncertain whether or not the fish and mammalian T1R2 genes have independently diverged from an ancestral T1R gene. Another feature of fish T1R2s was the presence of multiple members in each species. They showed 61–71% identity to each other, and were located within a region of 10–20 kb in

1318

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

the genome. In addition, we found several pseudogenes showing ca. 60% identity to fish T1R2 members, as in the case of mammalian V2R pheromone receptor genes (Matsunami and Buck, 1997). 3.2. Molecular evolution of fish T2R genes We identified one or two fish T2R genes in a fish species. These numbers are much smaller than those of known mammalian T2R genes (25 genes for human, 33 genes for mouse, Adler et al., 2000; Conte et al., 2003; Shi et al., 2003). Recently, the presence of two T2R-like genes in zebrafish genome was reported (DrT2R1-like and DrT2R2like genes, Pfister and Rodriguez, 2005). DrT2R1-like gene is identical to zf997-5 found in our searching process (Section 2.2), and DrT2R2-like gene is identical to zfT2R1b. Supposing that we reckon zf997-5 as a member of zebrafish T2R genes, the total number of zebrafish T2R genes is three. Taking vertebrate V1R genes as examples, Pfister and Rodriguez also reported a smaller number of V1R gene in zebrafish genome (only one) than in mouse genome (137, Rodriguez et al., 2002). The small number of V1R genes was also reported about primate genomes (less than two, Young et al., 2005), suggesting that the number of receptor genes expressed in chemosensory system can vary among the species. Accordingly, it is possible that fish species, when compared to mammals, have a more diminished set of T2R genes. It is also possible that fish has other distantly related T2Rs that we could not identify because of low degree of identity between fish and mammalian T2Rs (Section 2.2) and diversity of fish T2Rs described as follows. Several orthologous relationships were reported between human and mouse T2R genes. For example, human T2R4 shows 65% identity to mouse T2R8, both of which receive the bitter substance, denatonium benzoate (Chandrashekar et al., 2000). On the other hand, there are groups of mammalian T2R genes that do not show orthologous relationships (human- and mouse-specific groups in Fig. 2(B), Shi et al., 2003; Conte et al., 2003). The identity values between human and mouse T2Rs belonging to the latter groups are 32–35%, which are close to the values, 33%, observed between ffT2R1 and zfT2R1 s. Thus, it is possible that fish T2Rs show less orthologous relationships among species. 3.3. The variety of fish taste receptor cells In mammals, there are three kinds of taste receptor cells expressing T1Rs: cells expressing T1R1 and T1R3, those

expressing T1R2 and T1R3, and those expressing only T1R3. We found that zebrafish as well has these three kinds of taste receptor cells. Intriguingly, fish T1R1s and T1R3s were phylogenetically related to mammalian T1R1s and T1R3s, respectively. This suggests that the functions of fish T1R1s and T1R3s may be conserved among these species. On the other hand, fish T1R2s were not closely related to mammalian T1R2s. This means that each fish T1R2 is species-specific. We found a significant population of cells expressing only fish T1R2 members. It is possible that fish T1R2s function independently of fish T1R3s. The presence of multiple members of fish T1R2s may increase the variety of fish taste receptor cells. These members were co-expressed with fish T1R3s in all possible combinations. Especially, the presence of cells co-expressing all of these fish T1Rs suggests that more broadly tuned responsiveness exists in fish taste receptor cells than in mammals. 3.4. Discrimination of taste modality in vertebrates Mammalian T1Rs and T2Rs are activated by chemical compounds that cause preferable and aversive behaviors, respectively. This taste recognition takes place in the central nervous system, but is dependent on which kind of taste receptor is expressed in peripheral taste receptor cells (Zhao et al., 2003; Mueller et al., 2005). Several types of taste receptor cells have been identified in fishes as well as in mammals. Notably, the cells expressing fish T1Rs are segregated from those expressing fish T2Rs. This suggests that the discrimination of taste modality in fishes also depends on the properties of peripheral sensory cells, raising the hypothesis that vertebrates could show common mechanisms of taste information processing. Identification of ligands for fish T1Rs and T2Rs and behavioral analyses will provide important insight to allow verification of this hypothesis.

4. Experimental procedures 4.1. Database search The TBLASTN program was used to search for genomic sequences showing significant identity to mammalian T1Rs and T2Rs in the public genome databases of fugu (Fugu v.2.0 assembly, http://www.ensembl.org/Fugu_rubripes/), zebrafish (Zv4 assembly, http://www.ensembl.org/Danio_rerio/), and medaka fish (200406 revision, http://biol1.bio. nagoya-u.ac.jp:8000/). In these database releases, 90% of "

Fig. 5. Relationships in expression among zfTRs. Sections of pharyngeal region were hybridized with three kinds of probes in the combinations indicated by colored letters at the bottom of the panels: zfTRs and zfPLC-b2 (A–K) and three zfTRs (L). The signals of probes were represented as pseudocolors (green, red and blue). A relative population of cells expressing each receptor gene to zfPLC-b2-expressing cells is represented by bar under the panels, and percentage of co-expression between zfTRs is illustrated with colors. The blue bar under the panel A represents 10% population of zfPLC-b2-positive cells. Scale barZ 10 mm. (M) Venn diagram illustrating the relationships in zfTR expression.

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

1319

1320

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321

fugu genome (ca. 400 Mb), 92% of zebrafish genome (ca. 1700 Mb), and 99% of medaka fish genome (ca. 800 Mb) were covered. These information are from Ensembl office and medaka fish genome web site. For fugu T2Rs search, nineteen mammalian T2Rs sequences: hT2R1 (BC095521), hT2R3 (BC095523), hT2R13 (BC095518), hT2R16 (BC095524), mT2R2 (BK001080), mT2R7 (BK001072), mT2R8 (AF227148), mT2R10 (AF412304), mT2R14 (BK001075), mT2R19 (AF227149), mT2R21 (BK001077), rT2R1 (AF227140), rT2R2 (AF240765), rT2R3 (AF227141), rT2R4 (AF227142), rT2R5 (AF227143), rT2R7 (AF227144), rT2R8 (AF227145), and rT2R9 (AF227146) obtained from the GenBank, were used as references. The identified sequences were analyzed using CLUSTAL W program (Thompson et al., 1994) available in the DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp/search/clustalw-j.html) and TMPRED program (Hofmann and Stoffel, 1993) available in the EMBnet.ch (http://www.ch.embnet.org/ software/TMPRED_form.html). The phylogenetic trees were constructed on the basis of amino acid sequence alignments by the neighbor-joining method as implemented in CLUSTAL W program. The stabilities of trees were estimated by bootstrap analysis for 1000 replications using the same program. Tree figures were generated by TreeView program (Page, 1996). 4.2. cDNA clones Total RNA was isolated from the lips and gill rakers of adult zebrafish, and reverse transcribed into cDNA using oligo dT primer and First-strand cDNA synthesis kit (Amersham Biosciences Inc., Piscataway, NJ). The entire coding regions of zfT1R1, zfT1R2a, zfT1R3, zfT2R1a, and zfT2R1b, and partial coding regions of zfT1R2b and zfPLCb2 were amplified from the above cDNA using specific primers synthesized according to the genomic sequences and pyrobest DNA polymerase (Takara Bio Inc., Otsu, Japan). 4.3. Fish strains Zebrafish (Wako-Riken strain) were kindly provided by Dr Hitoshi Okamoto (RIKEN Brain Science Institute, Wako, Japan) and medaka fish (Cab strain) were a kind gift from Dr Makoto Furutani-Seiki (ERATO, Kondoh Differentiation Signaling Project, Kyoto, Japan). The fishes were maintained under a 14 h/10 h day/night cycle in a closed circulating system at 25 8C (AQUA, Tokyo, Japan). 4.4. In situ hybridization Fresh frozen sections of zebrafish and medaka fish were prepared, and hybridized following the procedure described previously (Kusakabe et al., 2000). For doubleand triple-labeling fluorescent in situ hybridization analysis, horizontal or sagittal sections of the pharyngeal

region were prepared to be 7 mm in thickness, and the tyramide signal amplification (TSA) method was used for the detection of signals. Sections were hybridized simultaneously with two or three kinds of probes labeled with fluorescein, digoxigenin, or biotin. Each labeled probe was sequentially detected by incubation with peroxidase-conjugated anti-fluorescein antibody (Roche), peroxidase-conjugated anti-digoxigenin antibody (Roche), and peroxidase-conjugated streptavidin (NEN Life Science Products, Boston, MA), and by subsequent reaction with fluophor-labeled tyramides, TSA-AlexaFluor 488, TSA-AlexaFluor 555, and TSA-AlexaFluor 647 (Molecular Probes, Eugene, OR) as a substrate for peroxidase. After each TSA reaction, peroxidase was inactivated by incubation with 1% hydrogen peroxide for 30 min. Fluorescent images were obtained with a laser scanning confocal microscope LSM510 (Carl Zeiss Japan, Tokyo, Japan) and transformed into pseudocolor images using LSM5 Image Browser (Carl Zeiss Japan, Tokyo, Japan). For triple-labeling fluorescent in situ hybridization analyses, more than four hundred of cells positive to zfPLC-b2 were examined for the co-existence of signals in each combination of zfTR probes.

Acknowledgements This study was supported by Grant-in-Aid 16108004 (to K.A.), Grant-in-Aid 16688006 (to I.M.) from the Ministry of Education, Culture, Sports, Science and Technology, and by Grant-in-Aid for Creative Scientific Research 13GS0015 (to K.A.) from the Japan Society for the Promotion of Science.

References Adler, E., Hoon, M.A., Mueller, K.L., Chandrashekar, J., Ryba, N.J., Zuker, C.S., 2000. A novel family of mammalian taste receptors. Cell 100, 693–702. Asano-Miyoshi, M., Abe, K., Emori, Y., 2000. Co-expression of calcium signaling components in vertebrate taste bud cells. Neurosci. Lett. 283, 61–64. Chandrashekar, J., Mueller, K.L., Hoon, M.A., Adler, E., Feng, L., Guo, W., et al., 2000. T2Rs function as bitter taste receptors. Cell 100, 703–711. Conte, C., Ebeling, M., Marcuz, A., Nef, P., Andres-Barquin, P.J., 2003. Evolutionary relationships of the Tas2r receptor gene families in mouse and human. Physiol. Genomics 14, 73–82. Hara, T.J., 1994. Olfaction and gustation in fish: an overview. Acta Physiol. Scand. 152, 207–217. Hofmann, K., Stoffel, W., 1993. TMbase—a database of membrane spanning proteins segments. Biol. Chem. Hoppe-Seyler 374, 166. Kotrschal, K., 2000. Taste(s) and olfaction(s) in fish: a review of specialized sub-systems and central integration. Pflugers Arch. 439, R178–R180.

Y. Ishimaru et al. / Mechanisms of Development 122 (2005) 1310–1321 Kusakabe, Y., Yasuoka, A., Asano-Miyoshi, M., Iwabuchi, K., Matsumoto, I., Arai, S., et al., 2000. Comprehensive study on G protein alphasubunits in taste bud cells, with special reference to the occurrence of Galphai2 as a major Galpha species. Chem. Senses 25, 525–531. Matsunami, H., Buck, L.B., 1997. A multigene family encoding a diverse array of putative pheromone receptors in mammals. Cell 90, 775–784. Matsunami, H., Montmayeur, J.P., Buck, L.B., 2000. A family of candidate taste receptors in human and mouse. Nature 404, 601–604. McLaughlin, S.K., McKinnon, P.J., Margolskee, R.F., 1992. Gustducin is a taste-cell-specific G protein closely related to the transducins. Nature 357, 563–569. Mueller, K.L., Hoon, M.A., Erlenbach, I., Chandrashekar, J., Zuker, C.S., Ryba, N.J., 2005. The receptors and coding logic for bitter taste. Nature 434, 225–229. Naito, T., Saito, Y., Yamamoto, J., Nozaki, Y., Tomura, K., Hazama, M., et al., 1998. Putative pheromone receptors related to the Ca2Csensing receptor in Fugu. Proc. Natl Acad. Sci. USA 95, 5178–5181. Nelson, G., Hoon, M.A., Chandrashekar, J., Zhang, Y., Ryba, N.J., Zuker, C.S., 2001. Mammalian sweet taste receptors. Cell 106, 381–390. Nelson, G., Chandrashekar, J., Hoon, M.A., Feng, L., Zhao, G., Ryba, N.J., et al., 2002. An amino-acid taste receptor. Nature 416, 199–202. Ngai, J., Chess, A., Dowling, M.M., Necles, N., Macagno, E.R., Axel, R., 1993. Coding of olfactory information: topography of odorant receptor expression in the catfish olfactory epithelium. Cell 72, 667–680. Page, R.D., 1996. TreeView: an application to display phylogenetic trees on personal computers. Comput. Appl. Biosci. 12, 357–358. Perez, C.A., Huang, L., Rong, M., Kozak, J.A., Preuss, A.K., Zhang, H., et al., 2002. A transient receptor potential channel expressed in taste receptor cells. Nat. Neurosci. 5, 1169–1176. Pfister, P., Rodriguez, I., 2005. Olfactory expression of a single and highly variable V1r pheromone receptor-like gene in fish species. Proc. Natl Acad. Sci. USA 102, 5489–5494. Puzdrowski, R.L., 1987. The peripheral distribution and central projections of the sensory rami of the facial nerve in goldfish. Carassius auratus. J. Comp. Neurol. 259, 382–392.

1321

Rodriguez, I., Del Punta, K., Rothman, A., Ishii, T., Mombaerts, P., 2002. Multiple new and isolated families within the mouse superfamily of V1r vomeronasal receptors. Nature Neurosci. 5, 134–140. Rossler, P., Kroner, C., Freitag, J., Noe, J., Breer, H., 1998. Identification of a phospholipase C beta subtype in rat taste cells. Eur. J. Cell Biol. 77, 253–261. Shi, P., Zhang, J., Yang, H., Zhang, Y.P., 2003. Adaptive diversification of bitter taste receptor genes in Mammalian evolution. Mol. Biol. Evol. 20, 805–814. Speca, D.J., Lin, D.M., Sorensen, P.W., Isacoff, E.Y., Ngai, J., Dittman, A.H., 1999. Functional identification of a goldfish odorant receptor. Neuron 23, 487–498. Stone, L.M., Finger, T.E., Tam, P.P., Tan, S.S., 1995. Taste receptor cells arise from local epithelium, not neurogenic ectoderm. Proc. Natl Acad. Sci. USA 92, 1916–1920. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. Venkatesh, B., Gilligan, P., Brenner, S., 2000. Fugu: a compact vertebrate reference genome. FEBS Lett. 476, 3–7. Yasuoka, A., Aihara, Y., Matsumoto, I., Abe, K., 2004. Phospholipase C-beta 2 as a mammalian taste signaling marker is expressed in the multiple gustatory tissues of medaka fish. Oryzias latipes. Mech. Dev. 121, 985–989. Young, J.M., Kambere, M., Trask, B.J., Lane, R.P., 2005. Divergent V1R repertoires in five species: Amplification in rodents, decimation in primates, and a surprisingly small repertoire in dogs. Genome Res. 15, 231–240. Zhang, Y., Hoon, M.A., Chandrashekar, J., Mueller, K.L., Cook, B., Wu, D., et al., 2003. Coding of sweet, bitter, and umami tastes: different receptor cells sharing similar signaling pathways. Cell 112, 293–301. Zhao, G.Q., Zhang, Y., Hoon, M.A., Chandrashekar, J., Erlenbach, I., Ryba, N.J., et al., 2003. The receptors for mammalian sweet and umami taste. Cell 115, 255–266.