• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. Nov 2007; 177(3): 1655–1665.
PMCID: PMC2147940

Association of Polymorphisms in Odorant-Binding Protein Genes With Variation in Olfactory Response to Benzaldehyde in Drosophila


Adaptive evolution of animals depends on behaviors that are essential for their survival and reproduction. The olfactory system of Drosophila melanogaster has emerged as one of the best characterized olfactory systems, which in addition to a family of odorant receptors, contains an approximately equal number of odorant-binding proteins (OBPs), encoded by a multigene family of 51 genes. Despite their abundant expression, little is known about their role in chemosensation, largely due to the lack of available mutations in these genes. We capitalized on naturally occurring mutations (polymorphisms) to gain insights into their functions. We analyzed the sequences of 13 Obp genes in two chromosomal clusters in a population of wild-derived inbred lines, and asked whether polymorphisms in these genes are associated with variation in olfactory responsiveness. Four polymorphisms in 3 Obp genes exceeded the statistical permutation threshold for association with responsiveness to benzaldehyde, suggesting redundancy and/or combinatorial recognition by these OBPs of this odorant. Model predictions of alternative pre-mRNA secondary structures associated with polymorphic sites suggest that alterations in Obp mRNA structure could contribute to phenotypic variation in olfactory behavior.

INTERACTIONS with the chemical environment provide a sensitive target for forces of natural selection, as evident from the rapid evolution of odorant receptors (Young et al. 2002; Robertson et al. 2003). Drosophila provides an excellent model system for studies of olfaction due to its well-established genetics and the relative simplicity of its olfactory system. Although numerically simpler than the mammalian olfactory system in terms of the number of olfactory sensory neurons, the functional organization of the olfactory system of Drosophila is similar (Hallem et al. 2004). Olfactory sensory neurons in sensilla of the third antennal segments or maxillary palps form convergent projections onto output neurons in ~43 glomeruli in the antennal lobe (Vosshall et al. 2000). Individual neurons express one, or rarely two, odorant receptors from a repertoire of 62 Or genes (Clyne et al. 1999; Vosshall et al. 1999). Uniquely expressed odorant receptors dimerize with the common Or83b receptor, which is essential for transport and insertion of odorant receptors in the chemosensory dendritic membrane (Larsson et al. 2004; Benton et al. 2006). Olfactory sensory neurons that express the same odorant receptor converge on the same antennal lobe glomerulus (Gao et al. 2000; Vosshall et al. 2000; Bhalerao et al. 2003).

Odorants must dissolve in the aqueous perilymph to reach their cognate membrane-associated odorant receptors. Their solubilization and transport is thought to be mediated by odorant-binding proteins (OBPs) that are secreted by support cells. The Drosophila genome encodes 51 OBPs with different spatial patterns of expression (McKenna et al. 1994; Pikielny et al. 1994; Galindo and Smith 2001), which contain a characteristic structural signature of conserved cysteines (Graham and Davies 2002; Hekmat-Scafe et al. 2002).

Altered regulation of expression of different subsets of OBPs has been observed following mating (McGraw et al. 2004), exposure to starvation stress (Harbison et al. 2005), during the development of alcohol tolerance after exposure to alcohol (Morozova et al. 2006), as a correlated response to artificial selection for divergent levels of copulation latency (Mackay et al. 2005) and aggression (Edwards et al. 2006), and as a consequence of pleiotropic effects arising from single P-element-induced mutations that affect olfactory behavior (Anholt et al. 2003).

Whereas the role of OBPs in pheromone recognition has been clearly defined for several insect systems, the precise functions of these abundantly expressed proteins in olfaction remain obscure. Ligand specificities (whether broadly or narrowly tuned), interactions with odorant receptors (for which there remains scant evidence to date), interrelationships among OBPs with overlapping molecular-response profiles and their functional correspondence (if any) with odorant receptors, the significance of altered expression of some OBPs in aggression (Edwards et al. 2006), mating behavior (McGraw et al. 2004; Mackay et al. 2005), and alcohol sensitivity (Morozova et al. 2006), all pose as yet unresolved questions. To date, only one OBP in Drosophila melanogaster, encoded by Lush, has been characterized functionally (Kim et al. 1998; Xu et al. 2005). Flies homozygous for a deletion of the Lush gene do not avoid repellant concentrations of short-chain alcohols (Kim et al. 1998) and do not respond behaviorally or electrophysiologically to the aggregation pheromone 11-cis-vaccenyl acetate (Xu et al. 2005). The Or67d receptor, expressed in a subset of trichoid sensilla, has been identified as the receptor for 11-cis-vaccenyl acetate (Ha and Smith 2006; Kurtovic et al. 2007). Lush appears to be essential for delivering this pheromone to its receptor. Other insights into the functions of OBPs come from a recent study reporting that a polymorphism in Obp57e in D. sechellia determines preference for its host plant, Morinda citrifolia, and that D. melanogaster knock-out flies for Obp57e and Obp57d showed altered behaviors to hexanoic and octanoic acid produced by this plant (Matsuo et al. 2007).

Functional studies on OBPs have been hampered by the lack of Obp mutants, with the exception of Lush. Furthermore, if odorant recognition by OBPs is combinatorial, as is the case for odorant recognition by mammalian odorant receptors (Malnic et al. 1999), functional redundancy may render a laborious “one-gene-at-a-time” approach less than satisfactory, as it would provide only partial insights into the role of any one member of this multigene family in mediating olfactory behavior.

We devised a strategy to overcome these challenges by taking advantage of naturally occurring mutations that have arisen during evolution and that segregate as polymorphic variants in nature. We established isofemale lines from a natural population and inbred them for 20 generations, thus minimizing genetic variation within lines while retaining naturally occurring variation among the lines. We sequenced 13 Obp genes, located in two chromosomal clusters on the second and third chromosome, and used statistical tests for deviations from neutrality to assess patterns of selection. We then assessed whether polymorphisms in these 13 Obp genes were associated with naturally occurring variation in olfactory response to a standard test odorant, benzaldehyde. Finally, we show that single nucleotide polymorphisms (SNPs) in regulatory and coding regions that are associated with variation in chemosensory behavior can impact the predicted structure of pre-mRNA.


Drosophila stocks:

Isofemale lines were created from flies collected from a natural Raleigh (NC) population in 2002 and inbred by 20 generations of full sib mating to create 193 inbred lines. Flies were reared on cornmeal-molasses-agar medium under standard culture conditions of 25°, 70% humidity, and a 12-hr light/dark cycle.

OBP sequences:

Genomic DNA was extracted with the Puregene DNA extraction kit (Gentra, Minneapolis) and PCR primers were designed to amplify overlapping coding regions and 5′- and 3′-untranslated regions of the Obp genes (Obp56a–i and Obp99a–d) for Drosophila melanogaster and for a D. simulans line originally collected in Florida City (FL) by Jerry Coyne. PCR products were purified using Qiaquick columns (QIAGEN, Valencia, CA) and amplified samples were sequenced. Sequences were aligned with the Vector NTI Suite 9.0 program (Informax, Frederick, MD) to identify polymorphic sites. Singletons were excluded from the association and linkage disequilibrium (LD) analyses.

Molecular population genetics:

Neutrality tests were performed using the DnaSP 4.10.3 program (Rozas et al. 2003) (http://www.ub.es/dnasp). D. simulans sequences were compared to the sequences from the D. melanogaster population for the HKA test (Hudson et al. 1987) and the McDonald–Kreitman test (McDonald and Kreitman 1991). Estimates of Tajima's D (Tajima 1993), Fu and Li's D* and F* (Fu and Li 1993), and Fay and Wu's H (Fay and Wu 2000) take into account the calculated population recombination rate (Hudson 1987). Coalescent simulation was used to estimate P-values (two-tailed tests) with 103 coalescent simulations of an infinite site locus conditioned on the sample size; these simulations are implemented for a fixed number of segregating sites. LD between SNPs was analyzed using TASSEL 2.0 software (http://sourceforge.net/projects/tassel). Fisher's exact test was used to determine whether the pairs of sites were in significant LD.

Behavioral assays:

Olfactory behavior was quantified by measuring responses to the standard odorant, benzaldehyde, in a well established “dipstick” assay that we (Anholt et al. 1996; Mackay et al. 1996; Fanara et al. 2002) and others (Devaud 2003; Stockinger et al. 2005) have used previously. Pilot experiments on 5 of the lines over a range of benzaldehyde concentrations established that a concentration of 3.5% (v/v) provided optimal resolution for evaluating variation in olfactory behavior in these lines. We measured olfactory behavior of 4–10 day-old non-virgin flies from 193 wild-derived inbred lines in single-sex groups of five flies/replicate and 10 replicates/sex. All assays were conducted between 2:00 and 4:00 pm in a behavioral room at 25° and 70% humidity under white light. The experimental design was randomized such that measurements on individual lines were collected over several days to average environmental variation. Theoretically a score of 2.5 reflects indifference to the odorant. Note, however, that the precise determination of the boundary between indifference and attraction or avoidance is determined statistically when for example the distribution of scores from mutants is compared to that of a control.

Locomotor reactivity was assessed by Jordan et al. (2007) by subjecting single flies to a mechanical disturbance by tapping the vial twice against a table and recording the amount of time the fly is active in the 45 sec immediately following the disturbance.

Quantitative genetic analysis of olfactory behavior:

We used ANOVA to partition sources of variation in olfactory behavior according to the model Y = μ + L + S + L × S + E, where μ is the overall mean, L is the random effect of line, S is the fixed effect of sex, L × S is the random effect of line by sex, and E is environmental error. The total genotypic variance among lines was estimated as equation M1, where equation M2 is the among-line variance component and equation M3 is the variance for the line-by-sex interaction. The total phenotypic variance was estimated as equation M4, where equation M5 is the environmental variance component. Broad-sense heritability was estimated as equation M6 (Carbone et al. 2006). Narrow-sense heritability was estimated as equation M7, where equation M8 (Falconer and Mackay 1996). The genetic correlation between males and females was calculated as rMF = covMFMσF, where covMF is the covariance of line means for the two sexes and σM and σF are the square roots of the genotypic variances for each sex. Analyses of variance and tests of significance were calculated using the Proc GLM procedure, and variance components were estimated using the Proc VARCOMP procedure in SAS (SAS Institute, Cary, NC).

Genotype–phenotype associations:

Association between polymorphisms and line means for olfactory behavior were analyzed using two way factorial ANOVA with the model Y = μ + S + M + M × S + E, in which μ is the overall mean, marker (M) and sex (S) are fixed effects, and E indicates error variance. We used permutation tests to determine random distributions under the null hypotheses of no association between Obp genotypes and olfactory behavior in response to benzaldehyde (Churchill and Doerge 1994). We performed two different permutation tests. To assess whether we observed more significant associations with olfactory behavior for each gene than expected by chance, we permuted the phenotypes among the markers 1000 times and recorded the number of significant associations at P < 0.05 for each permuted data set. To identify particular polymorphic sites with significant associations with behavior, we similarly permuted the data 1000 times and recorded the lowest P-value of each permuted data set. In both cases we used the 5% significance threshold of the permuted data sets to give an empirical type I error rate that accounts for multiple tests.

In cases where more than one polymorphism in a gene was associated with olfactory behavior, we tested for associations between haplotypes of these variants and line means for olfactory behavior by two-way ANOVA with the model Y = μ + S + H + H × S + E, in which haplotype (H) and sex (S) are fixed factors and E indicates error. We conducted post hoc analyses using least-square means to assess the effect of haplotypes and Tukey tests to control the experimentwise error rate.

The additive variance attributable to a marker polymorphism (equation M9) was estimated as equation M10, where a is one-half the difference in mean olfactory behavior between homozygous genotypes for the marker, q is the frequency of the rare marker allele, and p = 1 − q (Falconer and Mackay 1996).

Prediction of RNA secondary structure:

Secondary structures of the full length mRNA and pre-mRNA molecules transcribed from Obp99a, Obp99c, and Obp99d genes were predicted using Mfold and Afold, as described previously (Nackley et al. 2006; Shabalina et al. 2006). Energy minimization was performed by a dynamic programming method that finds the secondary structure with the minimum free energy with sums composed of factors that include stacking and loop length (Nackley et al. 2006; Shabalina et al. 2006). The RNA folding parameters were published by the Turner group (Mathews et al. 1999). Suboptimal stem-loop structures were analyzed by the Hybrid program (Nazipova et al. 1995).


Molecular population genetics of Obp genes:

Odorant receptors evolve rapidly to adapt to changes in the chemical environment (Young et al. 2002; Robertson et al. 2003). Since OBPs are the first components of the insect olfactory system to encounter odorants, they might also be expected to undergo rapid adaptive evolution. The interaction between the chemical environment and an organism, however, is not constant, as different chemical signals become relevant during different developmental stages, under different physiological conditions, and during different social interactions (e.g., courtship and mating). This raises the question whether different members of the Obp gene family follow similar or diverse evolutionary trajectories.

To address this question, we sequenced alleles of 13 Obp genes organized in two representative chromosomal clusters, Obp56a–i on the second chromosome, and Obp99a–d on the third chromosome, in wild-derived inbred lines of D. melanogaster. We initially sequenced 50 alleles of each gene. Preliminary analyses suggested associations with members of the Obp99 gene family and olfactory behavior (see below); therefore, we obtained additional sequences of 143 alleles for the Obp99 genes. We observed 299 SNPs and 18 insertion/deletion (indel) polymorphisms in this sample of Obp genes, with 154 and 163 polymorphic sites in the Obp56 and Obp99 gene clusters, respectively (Table 1). SNP numbers were highly variable, ranging from only a single SNP in Obp56f to 76 SNPs in Obp99c (Table 1).

Population genetic parametersa

Consistent with the large variation in SNP numbers among genes of similar size in the same genomic regions, estimates of nucleotide diversity [the average number of pairwise differences between sequences (π) and the number of segregating sites (θw)] vary over an order of magnitude among the Obp genes (Table 1), suggesting that OBPs have experienced different evolutionary histories. We applied tests for deviation from selective neutrality to members of the Obp56 and Obp99 gene clusters (Hudson et al. 1987; McDonald and Kreitman 1991; Fu and Li 1993; Tajima 1993; Fay and Wu 2000) [corrected for the estimated recombination rate, R, (Hudson 1987)] that are based on detecting reduction in genetic diversity over different evolutionary time scales. Null alleles of Obp56c that contain a premature stop codon at predicted amino acid position 17 segregate in this population with an allele frequency of 0.06. Since we expected these null alleles not to be under selection, we did not include them in tests for deviation from neutrality (our results, however, do not change when the null alleles are included in the analysis, because the software performing the tests treats the missing sequence as missing data in any case). Seven of the 13 Obp genes exhibited signatures of departure from neutrality (Obp56a, Obp56c, Obp56g, Obp56h, Obp99b, Obp99c, and Obp99d; Tables 1 and and2),2), although none of the P-values of these tests survives a strict Bonferroni correction for multiple tests. Tajima's D was significant for Obp99c, Fu and Li's D* for Obp56g and Obp56h, F* for Obp99c and Obp99d, and Fay and Wu's H test statistic indicated a recent selective sweep for Obp56c (Table 1). No significant deviation from neutral expectations was detected by the HKA test (Table 1), but the McDonald–Kreitman test showed deviations from neutrality for Obp56a, Obp99b, and Obp99d (Table 2). For 5 Obp genes (Obp56g, Obp56h, Obp99b, Obp99c, and Obp99d) the departure from neutrality was such that there were more polymorphisms segregating at intermediate frequency than expected under neutral mutation-drift balance, which could be attributable to balancing selection. In contrast, Obp56a and Obp56c show evidence of rapid evolution. Since there is no LD between Obp56a and Obp56c, the signatures of positive selection experienced by these 2 genes are likely not due to hitchhiking (supplemental Table S1 at http://www.genetics.org/supplemental/).

McDonald–Kreitman testsa

Pairwise comparisons of LD between polymorphic sites for the Obp56 and Obp99 genes (except the highly conserved Obp56f) show evidence of extensive historic recombination within each gene cluster. However, there are regions exhibiting LD between genes within each cluster, e.g., Obp99c and Obp99d (Figure 1; supplemental Table S1 at http://www.genetics.org/supplemental/), although long range LD is not common. We did not observe significant LD between polymorphisms in the Obp56 cluster and those in the Obp99 cluster (the Bonferroni corrected significance threshold is P < 2.68 E-6). The diagrams in Figure 1 also illustrate the differences in SNP densities among these genes.

Figure 1.
LD in the Obp56 and Obp99 clusters. Boxes below the diagonal reflect R2 values for all possible marker combinations and boxes above the diagonal indicate the corresponding P-values. Obp gene structures are denoted at the top by the horizontal line with ...

The pattern of polymorphisms in Obp99d contains an unusually large number of nonsynonymous substitutions (23 of 28), many of which occur at intermediate frequencies at adjacent nucleotide positions, generating extensive amino acid diversity (supplemental Table S2 at http://www.genetics.org/supplemental/). This unusual pattern of SNPs is not caused by a tandem polymorphic duplication of Obp99d (data not shown), and is consistent with the inference of balancing selection acting on this gene (Tables 1 and and22).

Quantitative genetics of olfactory behavior:

We quantified naturally occurring variation in olfactory behavior among the 193 Raleigh inbred lines and observed broad variation in responsiveness to benzaldehyde, which appeared normally distributed and ranged from attractant (<<2.5) to repellant responses (>>2.5) (Figure 2a). To test whether the low responses were specific to olfactory behavior or due to a general deficit in locomotion, we assessed the correlation between locomotor behavior in response to a mechanical stimulus, which exhibits considerable inbreeding depression (Jordan et al. 2007) and olfactory behavior in these lines. The correlation between locomotion scores and olfactory response scores was not significantly different from zero (Figure 2b). Thus, the variation in olfactory behavior cannot be explained by inbreeding depression for general locomotor impairment.

Figure 2.
Distribution of mean olfactory response scores for male and female flies of 193 wild-derived inbred lines (a) and lack of correlation between olfactory response scores and locomotion scores (b). Olfactory responses were measured at an odorant concentration ...

ANOVA shows significant variation in olfactory behavior between males and females, among lines, and for the sex-by-line interaction (Table 3). The environmental error variance component (equation M11) is high (0.621), demonstrating the sensitivity of olfactory behavior to uncontrollable environmental variance. The line variance component (equation M12) is also high (0.435), reflecting substantial genetic variation among lines. The estimate of broad-sense heritability (H2) is H2 = 0.441. Assuming strict additivity, the estimate of narrow-sense heritability (h2) is h2 = 0.283. This value is larger than previously estimated in a different population (Mackay et al. 1996; h2 = 0.084 and 0.134 for chromosome 1 and chromosome 3 substitution lines, respectively). The substantial genetic component to variation in olfactory behavior provides a favorable scenario for association analysis. The significant sex-by-line interaction term indicates that there is variation in sexual dimorphism in the response to benzaldehyde, although the cross-sex genetic correlation (rMF = 0.893; Table 3) is higher than observed in a previous study of chromosome substitution lines (Mackay et al. 1996).

Variance components and quantitative genetic parameters from ANOVA of olfactory behavior in response to benzaldehyde

Association of polymorphisms in Obp genes with olfactory behavior:

We used a two-step strategy to identify polymorphisms in Obp genes that are associated with olfactory responsiveness to benzaldehyde. First, we conducted association tests with 50 lines, and used permutation analysis to estimate the number of associations one would observe by chance for each gene. We then compared the expected number of random associations with the observed number of associated polymorphic markers to identify Obp genes for further analysis. Obp99a and Obp99d showed a higher number of associations than expected by chance (supplemental Table S3 at http://www.genetics.org/supplemental/). Therefore, we focused our efforts on the Obp99 cluster, and obtained DNA sequences for an additional 143 alleles from the same population.

Association analyses with 193 alleles for each of the four Obp99 genes revealed individual genotype–phenotype associations with olfactory responsiveness that exceeded the permutation threshold for multiple testing in Obp99a, Obp99c, and Obp99d (Table 4, Figure 3). Since we found a significant line-by-sex interaction when we analyzed phenotypic variation in olfactory behavior in this population, we included a marker-by-sex interaction term in the association model. However, none of the significant SNPs had significant marker-by-sex interactions. The effects of the SNPs on olfactory behavior are quite large, ranging from 0.40 to 1.33 genetic standard deviation units and 0.26–0.88 phenotypic standard deviation units in the population of inbred lines. However, if we assume strict additivity, the individual polymorphisms explain only 3–6% of the total additive variance in olfactory behavior (Table 4).

Figure 3.
Associations between polymorphisms in the Obp99 cluster with variation in behavioral responses to benzaldehyde. The Obp99a, b, c, and d gene structures are schematically represented at the top of each graph with blue boxes representing exons, orange boxes ...
Associations of Obp99 genotypes with olfactory behavior

Polymorphic markers C75G in Obp99a and C141G in Obp99c are located in introns, whereas markers G67A and T78G in Obp99d are in the coding region. The G67A SNP results in a glutamine to lysine substitution, whereas the T78G SNP is synonymous. These SNPs are in LD (P < 0.0001). In addition, the C29A SNP in Obp99d is near the permutation threshold for statistical significance (Figure 3). This nonsynonymous substitution changes a cysteine into tyrosine, thereby eliminating a cysteine that could contribute to the formation of a structurally important disulfide bond. This SNP is in substantial LD (P < 0.01) with T78G. Haplotype analysis of the G67A and T78G polymorphisms in Obp99d shows a significant difference in phenotypic values between haplotypes GT (71% frequency) and AG (5% frequency) in Obp99d (Figure 3). The observed differences in phenotypic values between these haplotypes were the same for both sexes.

Our results implicate Obp99a, Obp99c, and Obp99d in the recognition of benzaldehyde and suggest that these OBPs are either redundant for the recognition of this odorant, or more likely, that OBPs recognize odorants in a combinatorial manner similar to odorant recognition by odorant receptors (Malnic et al. 1999). Thus, a given odorant would interact with multiple OBPs with different affinities in a concentration-dependent manner and a given OBP would recognize multiple odorants on the basis of its molecular response profile.

Effects of polymorphisms associated with olfactory behavior on predicted secondary pre-mRNA structure:

To gain insights into the mechanism by which synonymous or noncoding SNPs could affect phenotypic variation in olfactory behavior, we predicted secondary structures for pre-mRNA molecules transcribed from alternative SNP genotypes. The C75G polymorphism in Obp99a is located in an intron. Frequent local secondary structures predicted for sequences with C and G in position 75 are very similar except for the pairing in position 75, where C is paired and G is not paired (Figure 4). Position 75 in Obp99a pre-mRNA is close to the splice site (35 bp upstream from the exon/intron boundary), and the different predicted RNA structures could potentially affect splicing (Solnick 1985).

Figure 4.
Predicted local stem-loop structures associated with polymorphic markers C75G in Obp99a, C141G in Obp99c, and G67/T78, G67/G78, and A67/G78 in Obp99d. The local stem-loop structures in Obp99d modulate base pairing of their neighbors in an ~100-nt ...

The C141G polymorphism in Obp99c is also located in an intron. Different stable structures are predicted for Obp99c sequences with nucleotides C and G in position 141, where C is paired with a higher probability (P = 0.56) than G (P = 0.18) (Figure 4; supplemental Table S4 at http://www.genetics.org/supplemental/). It is interesting that G in position 141 is frequently unpaired in suboptimal secondary structures, although G has the highest potential for base pairing among all four nucleotides.

The three common haplotypes formed by SNPs at positions 67 and 78 of Obp99d produce different optimal local secondary structures (Figure 4), although the free energy for the local secondary structures in this region does not differ dramatically (from −133.9 to −126.5 kcal/mol, Figure 4). Structure predictions for Obp99d mRNAs showed that nucleotides in positions 67 (G67A, a nonsynonymous polymorphism) and 78 (T78G, a synonymous polymorphism) are frequently involved in the neighboring stem-loop structures and modulate base pairing of neighboring nucleotides (Figure 4). Nucleotides G and A in position 67 have different potentials for base pairing for the different haplotypes (P = 0.7 for G67/T78, P = 0.61 for G67/G78, and P = 0.05 for A67/78G, respectively, supplemental Table S4 at http://www.genetics.org/supplemental/). The rarest haplotype (A67/T78) has a dramatically different potential for base pairing in position 78 (P = 0.2 for T78, supplemental Table S4) compared to the three more common haplotypes. Since these polymorphisms are in strong LD, we cannot infer whether the effect on olfactory behavior is due to a structural change in the protein or the mRNA.


We have used a population genetics approach in conjunction with behavioral measurements to identify OBPs that recognize benzaldehyde, while at the same time gaining insights in the history of natural selection, mutation, and recombination of members of the OBP multigene family. Sequence analyses showed that not all OBPs share the same evolutionary history. While patterns of polymorphism in six OBPs do not depart from those expected under neutral mutation–random drift balance, statistical tests for deviations from neutrality identify signatures of positive selection or balancing selection for seven OBPs. These diverse evolutionary trajectories may result from the diversity of biological functions influenced by OBPs. Differential expression of OBPs has been observed in lines artificially selected for aggression (Edwards et al. 2006), alcohol sensitivity (Morozova et al. 2006), copulation latency (Mackay et al. 2005), and starvation stress resistance (Harbison et al. 2005).

Sequence analyses showed that OBPs have different evolutionary histories and statistical tests for deviations from neutrality identify different signatures of selection for eight OBPs. The neutrality tests used detect signatures of selection over different evolutionary time scales (Sabeti et al. 2006); this is most likely the reason that inferences regarding selection were not consistent for the different tests. Tajima's D (Tajima 1993) and Fu and Li's D* (Fu and Li 1993) tests are based on different sensitivities of summary measures of nucleotide diversity within a species to a selective sweep and have different powers to detect departure from neutrality. Fay and Wu's H test (Fay and Wu 2000) analyzes high-frequency-derived alleles as a signature of a recent selective sweep and, thus, extends over a recent evolutionary period as high-frequency-derived alleles rapidly reach fixation (Sabeti et al. 2006). The results of these tests may be confounded by demographic history. Changes of population size can affect Tajima's D (Tajima 1993) and Fu and Li's D* (Fu and Li 1993) tests. Populations that have undergone a recent bottleneck in population size have a similar compressed genealogy to populations that are under positive selection; Tajima's D (Tajima 1993) and Fu and Li's D* (Fu and Li 1993) statistics are expected to be positive in both cases. Negative Tajima's D (Tajima 1993) and Fu and Li's D* (Fu and Li 1993) statistics are expected for populations undergoing recent expansion of population size, which increases in θw. Population subdivision can confound Fay and Wu's test by generating a nonsignificant H statistic. In contrast, the McDonald–Kreitman test (Mcdonald and Kreitman 1991) compares the ratios of nonsynonymous and synonymous substitutions within and among species and detects deviations from neutrality that persist over a long evolutionary time. Perhaps the best indicators of the diverse patterns of evolution of Obp genes are the different levels of nucleotide diversity in closely linked genes that presumably experience similar mutation and recombination rates (Figure 1).

Behavioral responses to benzaldehyde showed a greater range of phenotypic variation than observed previously with standard inbred laboratory stocks, and were elicited at a higher concentration of benzaldehyde. Olfactory responses were not correlated with locomotor reactivity scores, indicating that low olfactory-response scores did not result from locomotor impairments due to inbreeding depression. ANOVA showed significant variation in sexual dimorphism but a high genetic correlation between the sexes (rMF = 0.893; Table 3). Variation in sexual dimorphism in avoidance responses to benzaldehyde has also been observed previously in chromosome substitution lines (Mackay et al. 1996) and among co-isogenic P-element insertion lines that affect olfactory behavior (Anholt et al. 1996; Sambandan et al. 2006).

The extensive phenotypic variation in our wild-derived inbred population provided a favorable scenario for association analyses. We detected associations with SNPs in Obp99a, Obp99c, and Obp99d which implicate these OBPs in recognition of benzaldehyde. However, we examined only ~25% of the Obp gene family and it is likely that additional OBPs would contribute to phenotypic variation in the response to this odorant. It should be noted that OBPs that interact with odorants, such as benzaldehyde, but for which there is no segregating variation in the population under study, would go undetected by our approach. Furthermore, the detection power for associations depends on the sample size, and additional associations that make a smaller contribution to the observed phenotypic variation might be detected if the population size were expanded. Similarly, our analysis to date has focused on only the commonly used test odorant benzaldehyde. Expanding this analysis to include the entire family of Obp genes with a battery of odorants would enable a comprehensive characterization of ligand specificities of the OBP family. However, our experiments with a single odorant and a limited number of Obp genes already show that the recognition of benzaldehyde by OBPs is redundant and likely combinatorial, reminiscent of odorant recognition by odorant receptors (Malnic et al. 1999). Functional redundancy may allow the persistence of segregating null alleles in the population, observed by us and others (Takahashi and Takano-Shimizu 2005).

Previous association studies in Drosophila have implicated SNPs in noncoding regions of Catsup in phenotypic variation in sternopleural bristle number, environmental plasticity of abdominal bristle number, and starvation resistance (Carbone et al. 2006). SNPs associated with variation in longevity, locomotor behavior, starvation resistance, and bristle number have been identified also in functional regions of the protein (Carbone et al. 2006). Tests for association of SNPs in Obp genes and responsiveness to benzaldehyde revealed four polymorphisms implicating three OBPs, all within the Obp99 cluster, in the recognition of this odorant. Whereas nonsynonymous SNPs in coding regions of Obp genes can affect ligand binding by introducing variation in protein structure, the most parsimonious explanation for the phenotypic effects of synonymous SNPs and SNPs in regulatory regions would be alterations in mRNA structure. Indeed, structure predictions of mRNAs encoded by alternative haplotypes of Obp99d show that a single base substitution can have a profound effect on secondary mRNA structure (Figure 4), which could affect its transport, splicing, ribosome binding, or translation efficiency (Kimchi-Sarfaty et al. 2007). Our theoretical predictions of causal effects of altered mRNA structures on the behavioral phenotype, however, will need to be supported experimentally in the future.

Elegant electrophysiological studies have generated molecular response profiles of odorant receptors in the Drosophila antennae and maxillary palps and shown that a single odorant can activate multiple odorant receptors (De Bruyne et al. 1999; Hallem et al. 2004; Carlson and Hallem 2006). Thus far, the function of OBPs in odorant recognition has remained enigmatic, as there is no clear correlation between expression patterns of OBPs and odorant receptors. The population genetics approach described here represents a first step toward defining molecular recognition profiles of the OBP family. Such information together with the expression patterns of odorant receptors of known response profiles will, ultimately, clarify how these olfactory gene families interact in enabling the fly to sense its chemical environment.


We thank Philip Awadalla for helpful discussions, Theodore J. Morgan for assistance with the permutation tests, Katherine Jordan for sharing locomotion data prior to publication, and Aleksey Y. Ogurtsov for assistance with the secondary structure predictions. This work was supported by National Institutes of Health (NIH) grant GM-059469 (to R.R.H.A.) and by the Intramural Research Program of the NIH, National Library of Medicine (S.A.S.).


Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession nos. EU088428EU089648 for Drosophila melanogaster sequences and EU089649EU089661 for D. simulans sequences.


  • Anholt, R. R. H., R. F. Lyman and T. F. C. Mackay, 1996. Effects of single P-element insertions on olfactory behavior in Drosophila melanogaster. Genetics 143: 293–301. [PMC free article] [PubMed]
  • Anholt, R. R. H., C. L. Dilda, S. Chang, J. J. Fanara, N. H. Kulkarni et al., 2003. The genetic architecture of odor-guided behavior in Drosophila: epistasis and the transcriptome. Nat. Genet. 35: 180–184. [PubMed]
  • Benton, R., S. Sachse, S. W. Michnick and L. B. Vosshall, 2006. Atypical membrane topology and heteromeric function of Drosophila odorant receptors in vivo. PLoS Biol. 4: e20. [PMC free article] [PubMed]
  • Bhalerao, S., A. Sen, R. Stocker and V. Rodrigues, 2003. Olfactory neurons expressing identified receptor genes project to subsets of glomeruli within the antennal lobe of Drosophila melanogaster. J. Neurobiol. 54: 577–592. [PubMed]
  • Carbone, M. A., K. W. Jordan, R. F. Lyman, S. T. Harbison, J. Leips et al., 2006. Phenotypic variation and natural selection at Catsup, a pleiotropic quantitative trait gene in Drosophila. Curr. Biol. 16: 912–919. [PubMed]
  • Carlson, J. R., and E. A. Hallem, 2006. Coding of odors by a receptor repertoire. Cell 125: 143–160. [PubMed]
  • Churchill, G. A., and R. W. Doerge, 1994. Empirical threshold values for quantitative trait mapping. Genetics 138: 963–971. [PMC free article] [PubMed]
  • Clyne, P. J., C. G. Warr, M. R. Freeman, D. Lessing, J. Kim et al., 1999. A novel family of divergent seven-transmembrane proteins: candidate odorant receptors in Drosophila. Neuron 22: 327–338. [PubMed]
  • de Bruyne, M., P. J. Clyne and J. R. Carlson, 1999. Odor coding in a model olfactory organ: the Drosophila maxillary palp. J. Neurosci. 19: 4520–4532. [PubMed]
  • Devaud, J. M., 2003. Experimental studies of adult Drosophila chemosensory behaviour. Behav. Processes 64: 177–196. [PubMed]
  • Edwards, A. C., S. M. Rollmann, T. J. Morgan and T. F. C. Mackay, 2006. Quantitative genomics of aggressive behavior in Drosophila melanogaster. PLoS Genet. 2: e154. [PMC free article] [PubMed]
  • Falconer, D. S., and T. F. C. Mackay, 1996. Introduction to Quantitative Genetics, Ed. 4. Prentice Hall, Englewood Cliffs, NJ.
  • Fanara, J. J., K.O. Robinson, S.M. Rollmann, R. R. H. Anholt and T. F. C. Mackay, 2002. Vanaso is a candidate quantitative trait gene for olfactory behavior in Drosophila melanogaster. Genetics 162: 1321–1328. [PMC free article] [PubMed]
  • Fay, J. C., and C. I. Wu, 2000. Hitchhiking under positive Darwinian selection. Genetics 155: 1405–1413. [PMC free article] [PubMed]
  • Fu, Y. X., and W. H. Li, 1993. Statistical tests of neutrality of mutations. Genetics 133: 693–709. [PMC free article] [PubMed]
  • Galindo, K., and D. P. Smith, 2001. A large family of divergent Drosophila odorant-binding proteins expressed in gustatory and olfactory sensilla. Genetics 159: 1059–1072. [PMC free article] [PubMed]
  • Gao, Q., B. Yuan and A. Chess, 2000. Convergent projections of Drosophila olfactory neurons to specific glomeruli in the antennal lobe. Nat. Neurosci. 3: 780–785. [PubMed]
  • Graham, L. A., and P. L. Davies, 2002. The odorant-binding proteins of Drosophila melanogaster: annotation and characterization of a divergent gene family. Gene 292: 43–55. [PubMed]
  • Ha, T. S., and D. P. Smith, 2006. A pheromone receptor mediates 11-cis-vaccenyl acetate-induced responses in Drosophila. J. Neurosci. 26: 8727–8733. [PubMed]
  • Hallem, E. A., M. G. Ho and J. R. Carlson, 2004. The molecular basis of odor coding in the Drosophila antenna. Cell 117: 965–979. [PubMed]
  • Hekmat-Scafe, D. S., C. R. Scafe, A. J. McKinneyand and M. A. Tanouye, 2002. Genome-wide analysis of the odorant-binding protein gene family in Drosophila melanogaster. Genome Res. 12: 1357–1369. [PMC free article] [PubMed]
  • Harbison, S.T., S. Chang, K. P. Kamdar and T. F. C. Mackay, 2005. Quantitative genomics of starvation stress resistance in Drosophila. Genome Biol. 6: R36. [PMC free article] [PubMed]
  • Hudson, R. R., 1987. Estimating the recombination parameter of a finite population model without selection. Genet. Res. 50: 245–250. [PubMed]
  • Hudson, R. R., M. Kreitman and M. Aguadé, 1987. A test of neutral molecular evolution based on nucleotide data. Genetics 116: 153–159. [PMC free article] [PubMed]
  • Jordan, K. W., M. A. Carbone, A. Yamamoto, T. J. Morgan and T. F. C. Mackay, 2007. Quantitative genomics of locomotor behavior in Drosophila melanogaster. Genome Biol. 8: R172. [PMC free article] [PubMed]
  • Kim, M. S., A. Repp and D. P. Smith, 1998. LUSH odorant-binding protein mediates chemosensory responses to alcohols in Drosophila melanogaster. Genetics 150: 711–721. [PMC free article] [PubMed]
  • Kimchi-Sarfaty, C., J. M. Oh, I. W. Kim, Z. E. Sauna, A. M. Calcagno et al., 2007. A “silent” polymorphism in the MDR1 gene changes substrate specificity. Science 315: 525–528. [PubMed]
  • Kurtovic, A., A. Widmer and B. J. Dickson, 2007. A single class of olfactory neurons mediates behavioural responses to a Drosophila sex pheromone. Nature 446: 542–546. [PubMed]
  • Larsson, M. C., A. I. Domingos, W. D. Jones, M. E. Chiappe, H. Amrein et al., 2004. Or83b encodes a broadly expressed odorant receptor essential for Drosophila olfaction. Neuron 43: 703–714. [PubMed]
  • Mackay, T. F. C., J. B. Hackett, R. F. Lyman, M. L. Wayne and R. R. H. Anholt, 1996. Quantitative genetic variation of odor-guided behavior in a natural population of Drosophila melanogaster. Genetics 144: 727–735. [PMC free article] [PubMed]
  • Mackay, T. F. C., S. L. Heinsohn, R. F. Lyman, A. J. Moehring, T. J. Morgan et al., 2005. Genetics and genomics of Drosophila mating behavior. Proc. Natl. Acad. Sci. USA 102: 6622–6629. [PMC free article] [PubMed]
  • Malnic, B., J. Hirono, T. Sato and L. B. Buck, 1999. Combinatorial receptor codes for odors. Cell 96: 713–723. [PubMed]
  • Mathews, D. H., J. Sabina, M. Zuker and D. H. Turner, 1999. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure. J. Mol. Biol. 288: 911–940. [PubMed]
  • Matsuo, T., S. Sugaya, J. Yasukawa, T. Aigaki and Y. Fuyama, 2007. Odorant-binding proteins OBP57d and OBP57e affect taste perception and host-plant preference in Drosophila sechellia. PLoS Biol. 5: e118. [PMC free article] [PubMed]
  • McDonald, J. H., and M. Kreitman, 1991. Adaptive protein evolution at the Adh locus in Drosophila. Nature 351: 652–654. [PubMed]
  • McGraw, L. A., G. Gibson, A. G. Clark and M. F. Wolfner, 2004. Genes regulated by mating, sperm, or seminal proteins in mated female Drosophila melanogaster. Curr. Biol. 14: 1509–1514. [PubMed]
  • McKenna, M.P., D. S. Hekmat-Scafe, P. Gaines and J. R. Carlson, 1994. Putative Drosophila pheromone-binding proteins expressed in a subregion of the olfactory system. J. Biol. Chem. 269: 16340–16347. [PubMed]
  • Morozova, T. V., R. R. H. Anholt and T. F. C. Mackay, 2006. Transcriptional response to alcohol exposure in Drosophila melanogaster. Genome Biol. 7: R95. [PMC free article] [PubMed]
  • Nackley, A. G., S. A. Shabalina, I. E. Tchivileva, K. Satterfield, O. Korchynskyi et al., 2006. Human catechol-O-methyltransferase haplotypes modulate protein expression by altering mRNA secondary structure. Science 314: 1930–1933. [PubMed]
  • Nazipova, N. N., S. A. Shabalina, A. Y. Ogurtsov, A. S. Kondrashov, M. A. Roytberg et al., 1995. SAMSON: a software package for the biopolymer primary structure analysis. Comput. Appl. Biosci. 11: 423–426. [PubMed]
  • Pikielny, C. W., G. Hasan, F. Rouyer and M. Rosbash, 1994. Members of a family of Drosophila putative odorant-binding proteins are expressed in different subsets of olfactory hairs. Neuron 12: 35–49. [PubMed]
  • Robertson, H. M., C. G. Warr and J. R. Carlson, 2003. Molecular evolution of the insect chemoreceptor gene superfamily in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 100: 14537–14542. [PMC free article] [PubMed]
  • Rozas, J., J. C. Sánchez-DelBarrio, X. Messeguerand and R. Rozas, 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19: 2496–2497. [PubMed]
  • Sabeti, P. C, S. F. Schaffner, B. Fry, J. Lohmueller, P. Varilly et al., 2006. Positive natural selection in the human lineage. Science 312: 1614–1620. [PubMed]
  • Sambandan, D., A. Yamamoto, J. J. Fanara, T. F. C. Mackay and R. R. H. Anholt, 2006. Dynamic genetic interactions determine odor-guided behavior in Drosophila melanogaster. Genetics 174: 1349–1363. [PMC free article] [PubMed]
  • Shabalina, S. A., A. Y. Ogurtsov and N. A. Spiridonov, 2006. A periodic pattern of mRNA secondary structure created by the genetic code. Nucleic Acids Res. 34: 2428–2437. [PMC free article] [PubMed]
  • Solnick, D., 1985. Alternative splicing caused by RNA secondary structure. Cell 43: 667–676. [PubMed]
  • Stockinger, P., D. Kvitsiani, S. Rotkopf, L. Tirian and B. J. Dickson, 2005. Neural circuitry that governs Drosophila male courtship behavior. Cell 121: 795–807. [PubMed]
  • Tajima, F., 1993. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585–595. [PMC free article] [PubMed]
  • Takahashi, A., and T. Takano-Shimizu, 2005. A high-frequency null mutant of an odorant-binding protein gene, Obp57e, in Drosophila melanogaster. Genetics 170: 709–718. [PMC free article] [PubMed]
  • Vosshall, L., A. Wong and R. Axel, 2000. An olfactory sensory map in the fly brain. Cell 102: 147–159. [PubMed]
  • Vosshall, L., H. Amrein, P. Morozov, A. Rzhetsky and R. Axel, 1999. A spatial map of olfactory receptor expression in the Drosophila antenna. Cell 96: 725–736. [PubMed]
  • Xu, P., R. Atkinson, D. N. Jonesand and D. P. Smith, 2005. Drosophila OBP LUSH is required for activity of pheromone-sensitive neurons. Neuron 45: 193–200. [PubMed]
  • Young, J. M., C. Friedman, E. M. Williams, J. A. Ross, L. Tonnes-Priddy et al., 2002. Different evolutionary processes shaped the mouse and human olfactory receptor gene families. Hum. Mol. Genet. 11: 535–546. [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...