Odorant-binding proteins expression patterns in recently diverged species of Anastrepha fruit flies

We studied two species of closely related South American fruit flies, Anastrepha fraterculus and Anastrepha obliqua which, despite being able to interbreed, still show some ecological and reproductive differences. Because part of these differences, such as host and mate preferences, may be related to olfactory perception, we focused our investigation on the differential expression of Odorant-binding protein (OBP) gene family, which participate in initial steps of the olfactory signal transduction cascade. We investigated patterns of expression of eight OBP genes by qPCR in male and female head tissues of both species. The expression patterns of these OBPs suggest that some OBP genes are more likely involved with the location of food resources, while others seem to be associated with mate and pheromone perception. Furthermore, the expression patterns obtained at different reproductive stages indicate that OBP expression levels changed significantly after mating in males and females of both species. All eight OBP genes analyzed here showed significant levels of differential expression between A. fraterculus and A. obliqua, suggesting that they may hold important roles in their olfactory perception differences, and consequently, may potentially be involved in their differentiation.


Results and Discussion
Screening of differentially expressed OBP genes between A. fraterculus and A. obliqua from RNA-seq data. Studies in Drosophila indicate that the detection of differing host odors by sibling species may be reflected in the expression level of the olfactory genes involved 29,30 . Therefore, we expected to find some OBP genes to show differential expression between A. fraterculus and A. obliqua because, even though they may use common hosts, they have different host preferences and also because individuals seem to be able to recognize pheromones of their own species [22][23][24]31 , which are two important ecological and reproductive attributes that may be involved in the species' differentiation. We looked initially for differentially expressed OBP genes between these two species using RNA-seq data of head tissues 27 and found three OBP genes significantly differentially expressed (FDR < 0.001): OBP56a, OBP56d and OBP83cd in virgin females and OBP56d in virgin and post-mating males ( Fig. 1; Table 1). Whereas OBP56d and OBP83cd were up-regulated in A. obliqua, OBP56a was up-regulated in A. fraterculus, at a level of differential expression much higher than the other significant gene comparisons here identified (Table 1), which indicate species differences in patterns of expression.
When we contrast the differences in expression between head tissues of A. fraterculus and A. obliqua with a study that investigated differences in expression of OBP genes in antennae of three Drosophila species which   Table 1. Significance and fold-change values of the RNA-seq differential expression analyses between A. fraterculus and A. obliqua OBP genes. VF = virgin female; VM = virgin male; PM = post-mating male. FDR = false discovery rate.
found 19 OBPs differentially expressed between D. sechellia and D. simulans, 21 OBPs between D. sechellia and D. melanogaster and 20 between D. simulans and D. melanogaster 30 , we notice many fewer differences in gene expression between A. fraterculus and A. obliqua than among the Drosophila species studied, in which 27 of the 52 members of OBP family were expressed in the antenna 32 , even though their divergence time is comparable. It is possible that there are biological aspects that explain this difference in number of differentially expressed OBPs, but we should mention that these studies used very different methodologies, so part of the difference may be due to the smaller bias of microarray analyses and qPCR used for Drosophila when compared to analyses using next-generation transcriptomes, used for Anastrepha, despite the use of two library replicates and a deep sequencing coverage for these libraries.
Since some studies in flies have reported that the most relevant differences in gene expression would occur between reproductive stages (instead of different tissues or life stages) [33][34][35] , we expanded our investigation to include different stages in reproduction. We investigated immature individuals, collected 24 hours after eclosion, to contrast against virgin individuals that were sexually mature, collected 10 days after eclosion, when males and females start showing mating behavior. Furthermore, comparisons between virgin and mated D. melanogaster females showed that the peak of differentially expressed genes was at 1-3 hours post-mating in the adult 34 , 6 hours post-mating in reproductive tissues 36 and 72 hours post-mating in head and brain tissues, with the majority of genes (including OBP99b) showing lower expression at this later time 33 . Since there was no previous information available for Anastrepha species, we considered these results in Drosophila to sample post-mating males and females of Anastrepha at 3, 6, 12, 24 and 48 hours post-mating. We compared differential expression of the different samples to virgin mature individuals, since we were particularly interested in the effect of sexual maturation and mating. We evaluated expression in seven reproductive profiles (immature and mature individuals and five post-mating profiles) of males and females for OBP56a, OBP56d and OBP83cd by qPCR analyses. Besides, we also used qPCR to compare the patterns of expression of five other OBP genes previously described as evolving under positive selection in Anastrepha species: OBP50a, OBP56h-1, OBP56h-2, OBP57c and OBP99c 26, 27 .
Intraspecific differential expression analysis. We used qPCR to investigate how the eight selected OBP genes were expressed along different reproductive stages in each species, to address two distinct questions: 1) whether immature and mature virgin individuals had significant differential expression and 2) whether expression significantly changes after mating, comparing individually five post-mating profiles against mature virgin individuals. We conducted and reported our species-specific analyses separately for males and females because studies in A. obliqua, B. dorsalis and D. melanogaster showed sex-dependent differences in levels of expression of chemosensory genes, which indicates that they experience, interact with, and adapt to their chemical environments differently 35,37,38 . Even though the reference genes here used in real time PCR analyses were originally described for different biological conditions, and only for A. obliqua 37 , they were also effective for studies on circadian clock genes of A. fraterculus 39 . Our tests indicate that these reference genes are also adequate for the conditions and tissues here tested, since we failed to observe significant differential expression levels among the profiles and species analyzed (data not shown). Since a reliable set of reference genes is critical for real time PCR analyses, these results expand once again the conditions, species, and tissues for which these reference genes are suitable.
We will refer to immature virgin males and females as immature males and females, mature virgin males and females as virgin males and females and post-mating mature males and females as post-mating males and females, to simplify reading. A. fraterculus showed significant differential expression for one OBP between immature and virgin females (OBP50a, up-regulated in virgin females), and for five OBPs between immature and virgin males (OBP56d, OBP99c and OBP83cd, up-regulated in immature males; OBP56h-2 and OBP57c, up-regulated in virgin males) (Fig. 2). We also found three OBPs up-regulated in post-mating females, in comparison with virgin females: OBP56a and OBP99c three hours after mating (PF3) and OBP83cd three and 12 hours after mating (PF3 and PF12), while the other five genes were down-regulated in some of the post-mating profiles. In A. fraterculus post-mating males, two OBPs were up-regulated: OBP56d (PM6, PM12 and PM48) and OBP57c (PM6), while OBP56h-1 was down-regulated (PM48) in post-mating males (Fig. 2). We highlight the OBP56d that showed opposite patterns between sexes, being down-regulated in post-mating females and up-regulated in post-mating males. A. obliqua showed significant differential expression for five OBPs between immature and virgin females (OBP56a and OBP83cd, up-regulated in immature females; OBP56d, OBP56h-2 and OBP57c, up-regulated in virgin females), and for all OBP genes between immature and virgin males (OBP56a and OBP83cd, up-regulated in immature, and the other genes up-regulated in virgins) (Fig. 3). These results indicated a similar expression pattern in these profiles between A. obliqua males and females, which was not found for A. fraterculus. Furthermore, in A. obliqua females, six genes were up-regulated in post-mating females, in comparison with virgins: OBP56d (PF24 and PF48), OBP56h-1 (PF12 and PF24), OBP56h-2 (PF12), OBP99c (PF24), OBP50a (PF12, PF24 and PF48) and OBP83cd (PF6, PF24 and PF48), whereas in males most OBPs that showed differential expression among the profiles tested were down-regulated in post mating, with the exception of OBP83cd (PM6, PM12 and PM48), which was up-regulated ( Fig. 3).
Considering the set of OBP genes we tested here, A. obliqua showed more OBP genes differentially expressed between immature and virgins than A. fraterculus and, in both species, males had a higher number of OBPs differentially expressed upon maturation (Afra_F = 1; Afra_M = 5; Aobl_F = 5; Aobl_M = 8), maybe due to an earlier reproductive maturation 40 . Before maturation, females tend to search for protein sources and plant hosts for foraging and eventual oviposition, shifting from a post-eclosion developmental phase during which they reach reproductive maturity to one in which they prepare for maximum reproduction 36 . Females seek carbohydrates and proteins for their own development and for that of their eggs. Likewise, immature males spend a large proportion of their time feeding to accumulate energy, which will define their reproductive success 41 . Based on these needs, we hypothesize that genes up-regulated in immature flies are more probably involved with the location of food resources during foraging activity, instead of location of mate partners. Besides, sexual maturation in insects is accompanied by the need to locate a mate and, subsequently, in the case of females, to switch from mate searching to oviposition behavior 7, 42 . In C. capitata, for instance, studies revealed that sexually mature virgin females are attracted to the odor of male-produced pheromone, while mated females are more interested in finding suitable oviposition sites and are therefore more attracted to the odor of host fruits [42][43][44][45] . Considering these findings, we suggest that OBP genes up-regulated in virgins, in comparison with immature flies, may be involved in pheromone perception.
Our comparisons between virgin and post-mating profiles revealed that OBP expression levels of the genes we investigated changed significantly after mating. Whereas in A. fraterculus females most OBPs analyzed were down-regulated post-mating, in A. obliqua females we found the opposite, the majority of OBPs were up-regulated post-mating, which is the same pattern observed in D. melanogaster females 38 . After mating, females become less attracted to males, probably due to a lower expression of pheromones stimulated by seminal fluid proteins transferred from the males, tending to become unreceptive to further mates and look for oviposition sites 33,[46][47][48] . Thus, OBPs up-regulated in post-mating females could be involved in finding oviposition sites, whereas down-regulated OBPs, in pheromone perception. On the other hand, mature males are actively involved in lek formation, in which the recognition of male pheromones is extremely important for male aggregation and for female attraction 43,49 , which would explain a few up-regulated OBPs in post-mating males. A transcriptome analysis in C. capitata showed that sexual maturation induces profound expression changes in females and modest variations in males, whereas post-mating changes were modest in females, and consistent both during maturation and post-mating in males 46 , a pattern that was not repeated for the OBP genes we investigated in Anastrepha. In both species, we observed different patterns of expression between males and females, as was previously described for Drosophila OBPs 38 . Only OBP56a showed the same pattern of expression between males and females of A. obliqua. Sex-specific gene expression differences found in OBPs were still more likely between virgin and post-mating profiles, since each sex has different behavioral and physiological attributes to maximize their reproductive potentials.
Since we used data from whole head tissues, we must consider that the OBPs here studied may have been expressed in other parts of the head, such as the brain or the taste organs 33,50 and not in the olfactory tissues, and therefore would not be involved in olfactory processes. However, in D. melanogaster, homologues of all eight of these OBPs were expressed in the antennae, especially OBP56d, OBP57c and OBP99c, expressed at high levels in both males and females 51,52 . Some of the OBP genes described here, OBP56a, OBP56d, OBP56h, OBP57c and OBP99c, were also found differentially expressed at different reproductive stages of D. melanogaster 38 (pairwise amino acid similarities between Anastrepha and D. melanogaster's OBPs are shown in the Supplementary  Table S1). Although the patterns of expression are not similar, the recurrence of the same OBP genes differentially expressed in both species may indicate a similar importance in these flies' reproductive process. Furthermore, OBP56a, OBP56d and OBP56h genes from D. melanogaster belong to the same gene cluster 53 , which may indicate they have the same promoters. Nonetheless, as it was observed for D. melanogaster, the expression patterns of OBP genes from the same gene cluster may differ in A. fraterculus and A. obliqua, suggesting that even OBP genes in the same gene cluster may be regulated independently. This independent regulation is supposed to be a necessary requirement for subfunctionalization or neofunctionalization during evolution, when daughter genes of duplication events either allow refinement and/or expansion in perception of the chemical environment, or yet the acquisition of specialized chemosensory functions 38 .
Interspecific differential expression analysis. Differences in OBP gene expression between species could result in different olfactory perception, since it may indicate that species experience, interact with, and adapt to their chemical environments differently 54 . To investigate potential differences in gene expression between A. fraterculus and A. obliqua, we investigated expression patterns across several reproductive stages, focusing on the differential expression of individual OBPs. Once again, we restricted our interspecific comparisons of gene expression to the same sex, because OBPs have shown sex-biased gene expression in flies 34,37,38 . The comparison between A. fraterculus and A. obliqua considering reproductive stages of immature and virgin females (Fig. 4, IF and VF profiles) revealed that most OBPs that displayed differential expression pattern were significantly more expressed in A. fraterculus, with the exception of OBP56a and OBP83cd in immature and OBP56d in virgins, more expressed in A. obliqua. This pattern was different in the comparison of males (Fig. 4, IM and VM profiles), since some OBPs (OBP56a, OBP56h-2 and OBP57c) were more expressed in A. fraterculus, others (OBP56h-1, OBP50a and OBP83cd) were more expressed in A. obliqua, and yet two of them (OBP56d and OBP99c) showed an antagonistic pattern, more expressed in immature males of one species but in virgins of the other species. The interspecific comparison among post-mating profiles revealed that OBP56d, OBP56h-2 and OBP57c showed more profiles significantly differentially expressed between A. fraterculus and A. obliqua females (Fig. 4, PF profiles), and OBP56d, OBP56h-1, OBP57c and OBP99c between males (Fig. 4, PM profiles). The putative orthologs of these genes are very similar between A. fraterculus and A. obliqua (See Supplementary Table S2 for pairwise amino acid identities), OBP56d being the most divergent, showing 88% similarity across species. We highlight the common occurrence of OBP56d and OBP57c as differentially expressed in male and female contrasts, and the fact that OBP56d had the highest number of differentially expressed profiles between these species, seven profiles for males and six for females. OBP57c gene expression was shown to increase in females of D. melanogaster exposed to odors of other females, which could suggest a function related to female pheromone reception 38 . To the best of our knowledge, no putative functions were associated to OBP50a, OBP56d and OBP56h. Seven OBPs were differentially expressed between males of A. fraterculus and A. obliqua at least one day after mating (24 and 48 hours after mating). Interestingly, at the same period, females showed the lowest numbers of OBPs differentially expressed between species (Fig. 5).
Although the qPCR results do not exactly corroborate the RNA-seq expression data, which is not surprising since the profiles are not exactly the same, the comparisons of each OBP expression between A. fraterculus and A. obliqua were concordant in both analyses. In summary, all eight OBP genes investigated here by qPCR showed significant differential expression between A. fraterculus and A. obliqua, but OBP56a, OBP56d, OBP57c and the two copies of OBP56h showed the most divergent patterns between A. fraterculus and A. obliqua. The two copies originally described for OBP56d for both species were suggested to represent intraspecific variation 26 , so they were considered as a unique gene to conduct the in silico gene expression analysis. Although two sequences were also associated with OBP56d for B. dorsalis (81% of pairwise identity between the two copies) 55 , the large pairwise identity between the two copies (94% and 95% for A. fraterculus and A. obliqua, respectively) and the single product produced in the qPCR analysis suggests that both sequences are different alleles of a single gene in the Anastrepha species studied. The intra-and interspecific variability previously reported, associated with the significant differential expression found between the two species, make OBP56d an interesting candidate gene to investigate differentiation among Anastrepha species. On the other hand, the two copies of OBP56h in Anastrepha may have arisen via gene duplication that preceded the divergence of A. fraterculus and A. obliqua. These two copies have a pairwise identity of 29% in A. fraterculus and 33% in A. obliqua 26 . The presence of two paralogous copies could have favored functional diversification, and promoted different expression patterns 56 . We clearly observed this differentiation in the interspecific contrast between males, in which OBP56h-1 and OBP56h-2 showed very distinct expression patterns.
Considering that we selected a set of OBP genes for this study that were the most differentiated between A. fraterculus and A. obliqua, either because they showed significant differential expression in Illumina transcriptomes, or because they showed significant evidence of positive selection, it is not surprising that all genes here analyzed were differentially expressed between A. fraterculus and A. obliqua. Since OBPs are involved in species-specific behaviors related to host and mate location, differences in OBP expression between A. fraterculus and A. obliqua could help explain their preference for different host fruits and mate choice. Although we cannot assume that the OBPs here analyzed are in fact responsible for such ecological differences, or would be the sole contributors to these differences, we found significant differences in their expression between A. fraterculus and A. obliqua, indicating that these genes are differentially regulated in these species, which may result in different olfactory responses and consequently, lead to important ecological and behavioral consequences. The differences in OBP expression here described may make these genes relevant to investigate diversification and speciation in these and in other species in the fraterculus group, as well as be considered targets for species-specific control of these pest species.

Conclusions
Intraspecific expression analyses at different reproductive stages of eight OBP genes showed that before mating, some OBP genes were more expressed in immature flies (OBP56a and OBP83cd for A. obliqua; OBP56d, OBP99c and OBP83cd for A. fraterculus), whereas others were more expressed in mature virgin flies (OBP56d, OBP56h-1, OBP56h-2 and OBP57c for A. obliqua; OBP50a, OBP56h-2 and OBP57c for A. fraterculus). Considering when these genes are expressed, it is possible that the former genes may be more involved in the search for food, whereas the latter genes may be more associated with mating. Furthermore, we found that expression levels of the OBP genes here studied changed significantly after mating in all sexes for both species. Interspecific comparisons showed high levels of differential expression for OBP genes between A. fraterculus and A. obliqua, especially OBP56d, OBP57c and the two copies of OBP56h, which suggests that these genes may have played important roles in the evolution and differentiation of these species.

RNA-seq expression analyses.
A differential expression analysis was performed between A. fraterculus and A. obliqua using transcriptomes of adult heads (accession numbers SRX2030818-SRX2030827 and SRX2030830-SRX2030833) 27 , where the flies' olfactory organs (antennae and maxillary palps) are located. The data included four profiles per species at different reproductive stages: mature virgin and mature post-mating for both sexes. The cDNA libraries were prepared in duplicates and each consisted of ten pooled individuals. Mature virgin individuals were collected 10 days after eclosion. Post-mating individuals were at least 10 days old and collected 15-20 hours after the first successful mate. More details about the cDNA libraries' construction and analysis are described elsewhere 27 . We used MAFFT v7 (available online and using default parameters) 57 to align 22 pairs of orthologous OBPs of A. fraterculus and A. obliqua previously identified 26,37 (Supplementary Table S2). Based on these alignments, we trimmed each ortholog OBP to obtain sequences with the same length. The set of sequences from each species was used as reference which was used to align that species' reads and estimate OBP expression. In so doing, we avoided problems in the RNA abundance quantification step arisen by interspecific polymorphism.
Selected OBP genes and qPCR primers design. We chose a set of OBP genes that showed differential expression in the RNA-seq data analysis, as well as OBP genes that were previously described under positive selection in A. fraterculus and A. obliqua 26,27 to perform quantitative PCR (qPCR) analyses. Eight OBP genes were considered, for which two primer pairs were designed, using the software Primer 3 version 4.0.0 (available at http://bioinfo. ut.ee/primer3/). We aligned Anastrepha OBP sequences with their putative D. melanogaster's orthologs, to identify putative intron positions using MAFFT 57 . Whenever possible, primer pairs were designed across introns to control for potential amplification of genomic DNA. We designed the primers in regions that were not variable between A. fraterculus and A. obliqua sequences, which allowed us to use the same primers for both species, although variations still could occur in the amplified fragment. We tested primer concentrations of 0.3, 0.5 and 0.6 μM in a final reaction volume of 10 μl, and used dissociation curve analyses to evaluate primer-specific amplifications for all OBP primer pairs. Once the best concentrations were determined, primer efficiency was tested using eight cDNA concentrations from a serial dilution starting at 5 ng/μl. We used a pool consisting of five heads of A. fraterculus females, with three technical replicates on these tests. All primers were tested, considering efficiency (E) between 95-105% and standard curve correlation coefficient (R 2 ) higher than 0.95 64 . When both primer pairs were efficient, we chose the one with the closest E value to 100% and R 2 value to 1. Primer pairs selected for qPCR analyses and their results for efficiency tests are shown in the Supplementary Table S3. Profiles analyzed by qPCR. Our experimental design for the qPCR analysis consisted of fourteen profiles equally divided between females and males, sampled with three biological replicates, with each replicate composed of a pool of five heads. The profiles analyzed represented three reproductive stages of the adult life, immatures (collected 24 hours after pupae eclosion), mature virgins, referred here simply as virgin individuals (collected 10 days after pupae eclosion) and post-mating (10-days-old individuals after the first successful mating). We collected post-mating samples by mating 10-days-old virgin males and females, maintained at separate cages since eclosion in the same controlled environment room (25 °C, 60-90% humidity and natural photoperiod), in a proportion of 1:1. When mating started, the couple was gently removed to another cage. To avoid incomplete matings, we considered only matings that lasted longer than 40 minutes and captured the very first successful matings for both sexes. After the males had dismounted, females and males were separated and collected at five different times: 3, 6, 12, 24 and 48 hours post-mating. Post-mating individuals were maintained under the same controlled environment until sacrifice, to reduce interferences on their normal gene expression levels.
Isolation of total RNA and cDNA synthesis. Total RNA was extracted using the TRIzol/chloroform protocol 65 . RNA quality was visualized in agarose gel electrophoresis for integrity and absorbance was measured in NanoVue ™ Plus Spectrophotometer (GE Helthcare). Samples were quantified with Qubit ® 2.0 Fluorometer, using the RNA BR assay kit (Invitrogen ™ ). Before transcription, total RNA was treated with DNase I Amplification Grade (Invitrogen ™ ), according to the manufacturer's protocol, to remove residual genomic DNA. One μg of treated RNA was converted into cDNA using iScript ™ cDNA Synthesis Kit (Bio-Rad) and cDNA samples were diluted to 5 ng/μl for the qPCR assays. qPCR assays. We performed the qPCRs in a CFX96 Touch ™ Real-Time PCR Detection Systems (Bio-Rad), using SsoFast ™ EvaGreen ® Supermix (Bio-Rad). Reactions were made with a final primer concentration of 0.6 μM in a final volume of 10 μl, which was the best primer concentration in all primer concentrations tests. Cycling parameters for all primers were 30 seconds at 95 °C, followed by 40 cycles of dissociation at 95 °C for 10 seconds and annealing and extension at 60 °C for 1 minute. A fluorescence reading was made at the end of each extension step. For melt curve analysis we used a protocol with temperatures that varied from 65 °C to 95 °C with increments of 0.5 °C for 5 seconds and continuous fluorescent measurements. No template controls (NTC) and inter-run calibrators were included in all qPCR plates. To check for reproducibility, three technical replicates were carried out for each sample. Relative quantification was calculated for each OBP gene with the Bio-Rad CFX Manager ™ software (Bio-Rad), using the comparative ∆∆ 2 C t method 66 in contrast with three Anastrepha reference genes: rpS17 (Ribosomal proteinS17), rpL18 (Ribosomal proteinL18) and ef1a (Elongation factor-1α) 37 . In these analyses, we contrasted both, the within species differential expression of OBP genes among different reproductive stages, and the between species differential expression of each OBP gene. Data were statistically analyzed by ANOVA followed by Tukey's test using Prism 5.01 software (GraphPad Software, San Diego, CA, USA). A value of p < 0.01 was considered statistically significant.