Uncovering the functional constraints underlying the genomic organization of the odorant-binding protein genes

Abstract

Animal olfactory systems have a critical role for the survival and reproduction of individuals. In insects, the odorant-binding proteins (OBPs) are encoded by a moderately sized gene family, and mediate the first steps of the olfactory processing. Most OBPs are organized in clusters of a few paralogs, which are conserved over time. Currently, the biological mechanism explaining the close physical proximity among OBPs is not yet established. Here, we conducted a comprehensive study aiming to gain insights into the mechanisms underlying the OBP genomic organization. We found that the OBP clusters are embedded within large conserved arrangements. These organizations also include other non-OBP genes, which often encode proteins integral to plasma membrane. Moreover, the conservation degree of such large clusters is related to the following: 1) the promoter architecture of the confined genes, 2) a characteristic transcriptional environment, and 3) the chromatin conformation of the chromosomal region. Our results suggest that chromatin domains may restrict the location of OBP genes to regions having the appropriate transcriptional environment, leading to the OBP cluster structure. However, the appropriate transcriptional environment for OBP and the other neighbor genes is not dominated by reduced levels of expression noise. Indeed, the stochastic fluctuations in the OBP transcript abundance may have a critical role in the combinatorial nature of the olfactory coding process.

Keywords: chemosensory system; chromatin domain; expression noise; gene cluster constraint; olfactory reception.

Fig. 1.—

Frequency distribution of the 3,434 Drosophila clusters. Frequency distribution of the 3,434 Drosophila clusters, which is conditioned on the cluster size (i.e., number of genes per cluster) and the BLS value (total time of cluster conservation in million years ago). The 58 significant clusters after correcting for multiple testing are depicted in red.

Fig. 2.—

Transcriptional environment in clusters that include OBP genes. Path analysis model for the causal relationships among cluster constraint probability (pBLS), the minimum age of a gene in the cluster (GA), the EB, the EI, and the EN. The GA is the exogenous variable. The numbers on the lines indicate the path coefficients. Solid and dashed arrows represent significant and nonsignificant relationships.

Fig. 3.—

Genomic features of OBP genes. Relationship between pBLS and the SSE value using (A) all OBP genes and (B) after removing the recent OBP duplicates (red points).

Fig. 4.—

Chromatin features of the clusters that include OBP genes. Relationships between the cluster constraint probability (pBLS) and (A) the proportion of nucleotides annotated as TE and (B) JIL-1 binding intensity in coding regions. The ρ values are the correlation coefficients of these associations. Distribution of the correlation coefficients between pBLS values and (C) the proportion of TE and (D) JIL-1 binding intensities in Drosophila clusters obtained by computer simulations (10,000 replicates of 31 clusters). The arrow indicates the correlation coefficients observed for clusters including OBP genes (P < 1e−5 and P = 0.010, for the TE proportion and JIL-1 binding intensity, respectively). The shaded area in the right tail represents the 5% of the total distribution area.

Fig. 5.—

The cluster including the lush (Obp76a) gene. The cluster (pBLS value of 0.999983) including lush (Obp76a) and other 19 non-OBP genes (blue boxes). The coordinates (from 19,570 k to 19,680 k) correspond to the 3L chromosome of Drosophila melanogaster. The intensity peaks below the genes indicate the EI values across 30 developmental stages (in different colors).