Differences in sweetener intake among inbred strains of mice are partially determined by allelic variation of the saccharin preference (Sac) locus. Genetic and physical mapping limited a critical genomic interval containing Sac to a 194 kb DNA fragment. Sequencing and annotation of this region identified a gene (Tas1r3) encoding the third member of the T1R family of putative taste receptors, T1R3. Introgression by serial backcrossing of the 194 kb chromosomal fragment containing the Tas1r3 allele from the high-sweetener-preferring C57BL/6ByJ strain onto the genetic background of the low-sweetener-preferring 129P3/J strain rescued its low-sweetener-preference phenotype. Polymorphisms of Tas1r3 that are likely to have functional significance were identified using analysis of genomic sequences and sweetener-preference phenotypes of genealogically distant mouse strains. Tas1r3 has two common haplotypes, consisting of six single nucleotide polymorphisms: one haplotype was found in mouse strains with elevated sweetener preference and the other in strains relatively indifferent to sweeteners. This study provides compelling evidence that Tas1r3 is equivalent to the Sac locus and that the T1R3 receptor responds to sweeteners.
Sweet taste transduction is thought to be initiated by the interaction of a sweetener ligand with a G-protein-coupled taste receptor on apical ends of the taste receptor cells, which evokes a sensation of pleasantness and a consummatory behavioral response (Lindemann, 1996). Many compounds that taste sweet to humans (sweeteners) are palatable to various species of animals, including mice (Beauchamp and Mason, 1991; Bachmanov et al., 2001). Inbred mouse strains display marked differences in their avidity for sweet solutions (Lush, 1989; Capeless and Whitney, 1995; Lush et al., 1995). Much of the difference in sweetener preferences among mouse strains is attributed to allelic variation of the saccharin preference (Sac) locus, on distal chromosome 4 (Phillips et al., 1994; Lush et al., 1995; Bachmanov et al., 1997b; Blizard et al., 1999; Li et al., 2001b). In addition to sweetener preferences, the Sac genotype influences the afferent responses of gustatory nerves to sweeteners (Bachmanov et al., 1997b; Li et al., 2001b), suggesting that the Sac gene is involved in peripheral taste transduction and may encode a sweet-taste receptor.
The T1R family of putative taste receptors consists of three genes expressed in taste receptor cells and located on the distal chromosome 4 (Hoon et al., 1999; Kitagawa et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Sainz et al., 2001), which makes them candidates for the Sac locus. The Tas1r1 (alias Gpr70) gene encoding the T1R1 receptor and the Tas1r2 (Gpr71, T1R2) gene have been excluded as candidates for Sac based on their more proximal chromosomal location (Kitagawa et al., 2001; Li et al., 2001b; Montmayeur et al., 2001). However, the Tas1r3 gene encoding the T1R3 receptor has been mapped to a more distal part of chromosome 4 corresponding to the Sac interval (Kitagawa et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Sainz et al., 2001).
In this study, using high-resolution linkage analysis followed by physical mapping and sequencing, we determined the exact limits of the critical Sac region. Next, we identified genes within the Sac interval and found that Tas1r3 is the most likely candidate for the Sac locus. The low-sweetener-preferring phenotype of 129P3/J strain was rescued by introgressing an allele of Tas1r3 from a high-sweetener-preferring C57BL/6ByJ strain using serial backcrossing during selection of a 129.B6-Sac congenic strain. Finally, sequence variants of Tas1r3 that are likely to have functional significance were identified using analysis of Tas1r3 sequences and sweetener-preference phenotypes in genealogically distant mouse strains. These in-vivo data provide compelling evidence that Tas1r3 is equivalent to the Sac locus and that it encodes a taste receptor responding to sweeteners.
Materials and methods
C57BL/6ByJ (B6) and 129P3/J (formerly 129/J, abbreviated here as 129) mice were purchased from The Jackson Laboratory. Details of breeding F2 hybrids (Bachmanov et al., 1997b) and of marker-assisted selection of a 129.B6-Sac congenic strain (Li et al., 2001b) are described elsewhere.
Taste preference tests
Consumption of 120 mM sucrose and 17 mM saccharin was measured in individually caged mice using 96 h, two-bottle tests, with water as the second choice, as described in detail elsewhere (Bachmanov et al., 1997b; Li et al., 2001b). Results are presented as sweetener solution intakes expressed per 30 g of body wt (the approximate weight of an adult mouse) per day, or as a preference score (ratio of solution intake to total fluid intake, in %). Indexes of sweetener consumption used for interval mapping were standardized within each gender and experimental group relative to the group mean and standard deviation.
Mouse genotyping and linkage analysis
Mouse genomic DNA was purified from tails by NaOH/Tris (Truett et al., 2000) or phenol/chloroform extraction. Simple sequence length polymorphism (SSLP) markers were tested using a standard protocol (Dietrich et al., 1992), with minor modifications (Bachmanov et al., 1997b; Li et al., 2001b). All other markers were tested using a single-strand conformation polymorphism (SSCP) protocol (Orita et al., 1989) or by sequencing the polymerase chain reaction (PCR) products. Marker and primer details are available upon request. Polymorphic markers were mapped by genotyping F2 and partially congenic mice. Chromosomal positions of non-polymorphic markers were confirmed using radiation hybrid mapping of the T31 mouse–hamster radiation hybrid panel (Research Genetics, Huntsville, AL) according to a standard protocol (http://www.jax.org/resources) and the data were submitted for analysis to the Mouse Genome Database. Markers that mapped in the Sac region were used to screen the RPCI-23 mouse bacterial artificial chromosome (BAC) library (Osoegawa et al., 2000).
Construction of a BAC contig
The RPCI-23 female (C57BL/6J) mouse BAC library (Osoegawa et al., 2000) was screened by hybridization (Church and Gilbert, 1984) with radioactively labeled probes (Feinberg and Vogelstein, 1983). The library was screened twice, first with a probe generated from the yeast artificial chromosome (YAC) clone 178B3 (http://carbon.wi.mit.edu:8000/cgi-bin/mouse/sts_info) and, second, with pooled probes of markers in the Sac region. The YAC 178B3 was chosen for screening because, based on its STS content, it mapped in the Sac region and appeared to be non-chimeric. Positive clones identified by the initial screenings were re-arrayed and hybridized against individual probes. The secondary screening results were confirmed by PCR. BAC insert sizes were determined using pulsed-field gel electrophoresis after digestion with NotI (Lengeling et al., 1999). The BAC contig was assembled using SEGMAP, version 3.48 (Green and Green, 1991).
Analysis of nucleic and amino acid sequences
DNA from BAC 118E21 was isolated using an alkaline lysis protocol followed by CsCl density gradient centrifugation, subcloned and sequenced. Sequence data were assembled using Phred, Phrap and Consed packages of programs. After BAC 118E21 was sequenced, the sequence-tagged site (STS) content of this BAC and overlapping BACs was confirmed by aligning the STS and BAC end sequences with the 118E21 sequence (Sequencher, Gene Codes Corporation, Ann Arbor, MI). Repeat sequences were identified using RepeatMasker (Smit and Green, 2000) and GENSCAN (Burge and Karlin, 1998) was used to predict gene content and exon–intron organization. The predicted proteins were submitted to TBLASTN searches against the nr and the mouse EST databases at NCBI or against Unigene.
To determine the intron–exon boundaries of the Tas1r3 gene, total RNA was extracted using TRIZol Reagent (Life Technologies Inc., Rockville, MD) from enzymatically separated mouse lingual epithelium (Ruiz et al., 1995; Spielman and Brand, 1995), which included fungiform, foliate and circumvallate taste papillae. The RNA was reverse transcribed (Superscript reverse transcriptase, Life Technologies). The cDNA samples were amplified using AmpliTaq DNA Polymerase with GeneAmp (Perkin Elmer Corp., Branchburg, NJ) and intron-spanning primers selected to distinguish genomic and cDNA. Single bands of expected sizes were excised from the gel, purified and sequenced.
Mouse T1R3 protein sequence was predicted from cDNA obtained from tongue epithelium by RT-PCR, and it was analyzed using the computer programs HMMTOP (Tusnády and Simon, 1998) and TOPO (Johns and Speth, 1996) to determine the hydrophobicity and the transmembrane regions. Protein motif prediction was conducted using the Motif search service on GenomeNet (http://www.motif.genome.ad.jp).
Analysis of variation in sweetener preference and Tas1r3 sequences among mouse strains
Sweetener preference data were taken from previous studies for the following mouse strains: 129/Rr, 129/Sv, AKR/J, BALB/cA, BALB/cByJ, C3H/He, C57BL/6ByJ, C57BL/6Ty, C57L/Lac, CBA/Cam, DBA/2Ty, IS/Cam, SEA/GnJ, ST/bJ, SWR/J (Lush, 1989) and CAST/Ei (A. Bachmanov et al., unpublished data). When preferences were available for two substrains, they were averaged and shown as 129, BALB/c and C57BL/6.
A 6.8 kb segment of genomic DNA, including ∼2.6 kb upstream and ∼1.2 kb downstream of Tas1r3, was sequenced in genealogically remote or unrelated (Beck et al., 2000) mouse strains with high (C57BL/6ByJ, SWR/J and CAST/Ei) or low (129P3/J, AKR/J and DBA/2J) sweetener preferences. Comparison of the sequences from these six strains identified a haplotype of six single nucleotide polymorphisms (SNPs) associated with sweetener preference. Next, the regions contributing to this haplotype were sequenced in additional mouse strains (BALB/cByJ, C3H/ HeJ, C57L/J, CBA/J, IS/CamEi, SEA/GnJ and ST/bJ). The C57BL/6J strain sequence obtained from the sequencing of RPCI-23 BAC 118E21 was identical to the C57BL/6ByJ sequence. In some cases, the exact substrain tested for sweetener preference differed from the substrain sequenced, usually because the older substrains were not available for sequencing.
The initial linkage analysis was conducted using an F2 intercross between B6 mice with high sweetener acceptance and 129 mice with the low sweetener acceptance. The F2 mice were phenotyped using 96 h two-bottle tests with sucrose and saccharin, and genotyped with markers polymorphic between the B6 and 129 strains. Interval mapping narrowed the region containing Sac to∼5 cM (Figure 1a,b). This region was further reduced to 0.7 cM (flanked by 280G12-T7 and D4Mon1 markers) during the marker-assisted selection of a 129.B6-Sac segregating partially congenic strain (Figure 1b,c).
A contig of BAC clones representing the Sac-containing region was constructed by screening the RPCI-23 mouse BAC library (Figure 1d). A 246 kb BAC clone, 118E21, including most of the Sac interval, was sequenced. The remaining small proximal part of the Sac region was contained in a sequence from mouse genomic DNA (GenBank accession No. AF185591). The 0.7 cM Sac interval flanked by 280G12-T7 and D4Mon1 had a physical size of 194 kb and contained twelve predicted genes (Figure 1e).
Identification of genes within the Sac interval
Of the twelve predicted genes, four were known (cyclin ania-6b, Dvl, Ubc6p and Tas1r3), two were similar to known human genes (KIAA1716 and SCNN1D) and six were represented as cDNA clones with GenBank accession Nos NM_025338, AK004732, NM_024472, AK010425, AA435261 and NM_026125. Most of the genes identified within the Sac-containing interval are involved or potentially involved in cell division and differentiation [cyclin ania-6b and Dvl (MacLachlan et al., 1995; Lee et al., 1999)], maintenance of intracellular processes [Ubc6p, NM_024472 and genes similar to KIAA1716 and SCNN1D (Waldmann et al., 1995; Gilon et al., 2000; Nagase et al., 2000)], or collagen synthesis (NM_026125). The functions of the predicted genes represented by NM_025338, AK004732, AK010425 and AA435261 are unknown.
Of these twelve genes within the 194 kb Sac interval, only one, Tas1r3 (taste receptor, type 1, member 3), was a G-protein-coupled receptor. A predicted Tas1r3 protein, T1R3, has moderate sequence homology to putative G-protein-coupled taste receptors T1R1 (encoded by Tas1r1, alias Gpr70, 32%; Figure 2a), T1R2 (encoded by Tas1r2, alias Gpr71, 30%) and mGluR4 (encoded by Gpcr1d, 23%) (Hoon et al., 1999; Chaudhari et al., 2000). Sequence comparison of cDNA from mouse lingual epithelium and genomic DNA showed that Tas1r3 contains six coding exons (Figure 2b). It is translated into an 858-amino-acid protein with a predicted secondary structure that includes seven transmembrane domains and a large hydrophilic extracellular N-terminus (Figure 2c). This structure is typical of the G-protein-coupled receptor family 3, which includes the metabotropic glutamate and extracellular calcium-sensing receptors. There is substantial evidence that a G-protein-coupled mechanism is involved in sweet taste transduction (Lindemann, 1996), suggesting that Tas1r3 is the most likely candidate for the Sac locus.
Sequence variants of Tas1r3
As a candidate for Sac, Tas1r3 should have sequence variants corresponding to phenotypical Sac alleles. To assess this correspondence, sequences of Tas1r3 and surrounding genomic DNA were analysed in mouse strains with known sweetener preferences (see details in Materials and methods). Two haplotypes consisting of six SNPs (Figure 2b) distinguished strains with high and low sweetener preferences (Figure 3). Two of these SNPs resulted in amino acid substitutions of threonine (found in all high-preferring strains) for alanine (found in all low-preferring strains) at position 55 (Thr55Ala) and isoleucine for threonine at position 60 (Ile60Thr; Figure 2a), both within the predicted extracellular N-terminal domain of T1R3 (Figure 2c). Two other variants occurred in the upstream regions, the fifth polymorphism was intronic and the sixth variant was a silent mutation (Ser) in exon 1.
Using a positional cloning approach, we narrowed the Sac-containing region to a 194 kb interval. One gene within this interval, Tas1r3, encodes a G-protein-coupled receptor (T1R3) that is expressed in taste receptor cells (Kitagawa et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Sainz et al., 2001). Based on the effects of the Sac genotype on peripheral sweet-taste responsiveness (Bachmanov et al., 1997a,b; Li et al., 2001b) and on the known mechanism of sweet-taste transduction (Lindemann, 1996), Tas1r3 is the most likely candidate for Sac.
If Tas1r3 is identical to Sac, substitution of Tas1r3 alleles must result in phenotypical changes attributed to the Sac locus. This can be tested through transgenic (‘phenotype rescue’ or ‘knock-out’) experiments. We obtained equivalent data using the 129.B6-Sac congenic mice. Introgression of the 194 kb chromosomal fragment containing the Tas1r3 allele from the high-sweetener-preferring B6 strain onto the genetic background of the 129 strain fully rescued its low sweetener preference phenotype: sweetener intake of the congenic mice was as high as that of mice from the donor B6 strain. This introgression was essentially equivalent to production of BAC-transgenic mice (Antoch et al., 1997; Probst et al., 1998) because the size of the donor chromosomal fragment was comparable to the size of a typical BAC. These data demonstrate that substitution of Tas1r3 alleles results in behavioral changes attributed to the Sac locus and therefore provide the strongest evidence to date that Tas1r3 is identical to Sac and that the T1R3 receptor responds to sweeteners.
Further evidence that Tas1r3 is equivalent to Sac was obtained from correspondence of Tas1r3 sequence variants to sweetener-preference phenotypes in inbred mouse strains, which are largely attributed to allelic variation of the Sac locus. Tas1r3 has multiple SNPs among the inbred strains we examined. We reasoned that variants that sort mouse strains into high- and low-sweetener preferring categories are more likely to be responsible for the Sac phenotype. To minimize the chances for sharing of chromosomal fragments due to identity-by-descent, we analysed mouse strains with unrelated or distant genealogies. Using this approach, we identified a haplotype of six SNPs concordant with sweetener preferences. Two of these SNPs resulted in amino acid substitutions at positions 55 (Thr55Ala) and 60 (Ile60Thr) within the predicted extracellular N-terminus domain of T1R3. The same Thr55Ala and Iso60Thr polymorphisms have been identified by others (Kitagawa et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Sainz et al., 2001) and they could affect ligand binding (Rana et al., 2001) or receptor dimerization (Max et al., 2001). The SNPs in non-coding upstream and intronic regions may affect expression of the gene (Phillips, 1999; Rana et al., 2001).
What are the ligands of T1R3? The B6 and 129 strains differ in responses to several different classes of sweeteners, and to ethanol, which has a sweet-taste component (Bachmanov et al., 1996, 2001). Our genetic analysis of the B6 × 129 F2 hybrids demonstrated that the Sac genotype affects behavioral and neural responses to representatives of several different sweet-tasting chemicals (sugars, non-caloric sweeteners, sweet amino acids) and also ethanol consumption (Bachmanov et al., 1997a; Tordoff et al., 1998). The effects of Sac genotype on taste responses to a variety of compounds suggest that they evoke sweet-taste sensation via a common receptor, T1R3.
It is controversial as to whether single or multiple sweet-taste receptors and transduction pathways exist (Bartoshuk, 1987; Bernhardt et al., 1996; Lindemann, 1996; DuBois, 1997). Our data are consistent with the existence of a single sweet-taste receptor. But it is also possible that the involvement of T1R3 in responses to diverse compounds is attributable to its dimerization with other members of the T1R family (Max et al., 2001; Montmayeur et al., 2001) or to the existence of T1R3 splice variants (Max et al., 2001) (X. Li and D. Reed, unpublished data), which may create multiple forms of sweet-taste receptors. However, the most abundant naturally occurring sweeteners, sugars, are chemically similar and may not require multiple receptors for their recognition. This may explain why the family of putative sweet (T1R) receptors is much smaller than the family of putative bitter (T2R) receptors (Adler et al., 2000).
Striking species differences in behavioral and physiological responses to sweeteners imply that the structure of sweet receptors varies among species (Beauchamp and Mason, 1991). A taste receptor for the sugar trehalose (Ishimoto et al., 2000), identified in Drosophila melanogaster, has no homology with T1R3, suggesting the independent origin of insect and mammalian sugar taste receptors. Among mammals, cats, which are strict carnivores, apparently do not prefer carbohydrate sweeteners, but do prefer certain amino acids humans describe as sweet (Beauchamp et al., 1977). Even primate species differ in the ability to recognize certain high-potency sweeteners as sweet (Glaser et al., 1995). TAS1R3, the human ortholog of mouse Tas1r3, exists in a region of conserved synteny, 1p36. Preliminary sequencing of TAS1R3 detected common SNPs (D. Reed et al., unpublished data), which may be involved in genetic variation of the sensory and behavioral responses to sugars in humans (Reed et al., 1997).
Finally, an understanding of sweet receptors may have practical applications. Most high-potency, low-calorie sweeteners were discovered by the accidental tasting of chemicals synthesized for other purposes (Walters, 1991). Models guided by the Tas1r3 taste receptor structure should allow the design and development of new potent ligands (DuBois, 1997; Nofre, 2001).
To summarize, we have now defined the limits of the Sac-containing critical genomic interval, restricted to the 194 kb DNA fragment, and have identified 12 genes within this region. Although all these genes can be considered candidates for Sac, based on gene expression and predicted protein structures, Tas1r3 is the most likely candidate. We have further demonstrated that experimental manipulation of the genomic interval containing Tas1r3 affects the taste responses to sweeteners attributed to the Sac locus, which provides novel in-vivo evidence for the identity of Tas1r3 and Sac. Finally, we have conducted the first quantitative analysis of the relationship between the Tas1r3 sequence variants and sweetener-preference phenotypes in genealogically diverse mouse strains. In addition to Tas1r3 missense polymorphisms reported by others, we have identified three other polymorphisms in non-coding regions that may affect Tas1r3 function. Together, these data provide strong evidence that Tas1r3 is identical to Sac and that it is associated with the taste quality of sweetness.
Note added in proof
While this paper was in press, a new study [
The data described in this paper were presented at the XXIII Annual Meeting of the Association for Chemoreception Sciences, Sarasota, FL, USA, 25–29 April 2001 (Li et al., 2001a). Sequence data reported in this paper have been deposited in GenBank and assigned the accession No. AF311386. We thank Maja Bućan for guidance with physical mapping and comments on the manuscript, Tim Wiltshire for help with BAC library screening, R. Arlen Price for initial genotyping of the F2 hybrids and Linda M. Bartoshuk, Taufiqul Huque, John Teeter, Charles Wysocki, Nancy Rawson and Joseph Brand for providing advice and discussion. We also thank Lianchun Chen, Dulce San Juan, Joe Catanese and Al Cairo for help with physical mapping. The excellent technical support of Sarah Tobias and the Pfizer Global Research and Development–Alameda Genotyping and Sequencing Core is acknowledged. This work was supported by National Institutes of Health grants R01DC00882 (G.K.B.), R03DC03509, R01DC04188 and R01-DK55853 (D.R.R.), R01AA11028 (M.G.T.) and R03DC03853 (A.A.B.), and by a special grant from the Ambrose Monell Foundation.