1
Are the pheromones of female and male Heliothis
virescens genetically linked? – a QTL analysis
Written by Thomas Rietbergen, 11307595 Date of submission: 24-07-2020
2 Summary
When trying to find a mate organisms often use some form of sexual communication. Lepidoptera use chemical signals (pheromones) in order to attract mates. The female pheromone is well studied and is involved in the long-range communication while the male pheromone is much less studied and thought to play a role at the short range. While the pheromones themselves have been well studied the way in which a sexual communication system such as this could evolve is poorly understood. The hypothesis is that sexual communication is under stabilizing selection, this is supported by the fact that the main purpose of sexual communication seems to be species recognition. However, the variation in pheromone blend that is observed within different species can not be explained by stabilizing selection. This hypothesis can also not explain speciation. In order to develop a new hypothesis it is necessary to better understand the genetics underlying sexual communication. This study aims to find out whether the male and female pheromones in Heliothis virescens are regulated by the same genomic regions, this is done using quantitative trait locus (QTL) analysis. The results show that the female and male pheromone seem to be regulated by different genomic regions.
Introduction
Sexual communication in moths
When looking for a suitable mate, animals use some form of sexual signals. In insects, these sexual signals are mainly chemical. These chemical signals are called (sex) pheromones. Lepidoptera, and more specifically moths, are insects that have been well studied for their pheromone communication. Female moths produce species-specific pheromones and emit them from the pheromone gland to attract males over long distances (Ando et al., 2004). Much research has been performed regarding the female pheromone signal and the male response to this signal. This has led to the discovery that these pheromones contain a blend of different fatty acids of various lengths (Roelofs et al., 1974; Tumlinson et al., 1975; Vetter & Baker, 1983). Detection of these pheromones by males causes the males to fly upwind to try and find the female (Vickers & Baker, 1994). Male moths also produce pheromones and release it from elaborate structures, so-called hairpencils, in proximity to the female. Thus, the male pheromone is believed to play a role in short-range communication (Baker et al., 1981).
Male pheromones in Lepidoptera
The effects of the male pheromone of Lepidoptera has not been studied as extensively as the female pheromones, however some discoveries have been made. In the noctuid moth Heliothis virescens the detection of male pheromones by females led to a decrease in the release of female pheromones (Hendricks & Shaver, 1975). These male pheromones also caused females to mate less than females who had not been in contact with male pheromones (Hosseini et al., 2016). These results suggest that the male sex pheromone might play a role in mate guarding, which could serve as a form of competition between males.
Hillier and Vickers (2004) found that the hair pencils of male heliothine moths are important for the mating success of the males, as males who had their hair pencils removed had lower mating success compared to mock treated males (Hillier & Vickers, 2004). This effect was reversed by releasing male
3 pheromone extract from filter paper, but only if the extract consisted of pheromones produced by the same species of moth. Introducing pheromones from a different species (in this case from Heliothis subflexa) did not restore the mating success of male H. virescens. This suggests that the male sex pheromone may play a role in species recognition and/or mate choice in heliothine moths.
Evolution of sexual communication
Female sex pheromone signals in moths are generally thought to be under stabilizing selection (Groot et al., 2016). This hypothesis is supported by studies that show that the males prefer the most common female pheromone signal (Zhu et al., 1997). The fact that the pheromones play a role in species recognition (Hillier &Vickers, 2004) further supports this hypothesis. The problem with this stabilizing selection is that it cannot explain the observed diversity in pheromone blends within different moth species. Additionally, it does not explain how the sexual communication systems could evolve and lead to speciation. Since both males and females use pheromones to signal and to potentially discriminate between mates, the genetic architecture and thus the biosynthetic pathways may be overlapping. An overlap of the pathways could explain some of the observed variation in pheromone blend in one sex as being a by-product of a change in pheromone blend in the other sex. Under stabilizing selection one would expect the observed variation in the female pheromone to be small, however this is not what is observed. There seems to be something more going on that we do not yet know of. If there were to be genetic overlap between the male and female pheromone of H. virescens this could help to explain the amount of variation present in the female pheromone. If we assume the female pheromone to be under stabilizing selection we would expect the genetic variation to be small for the involved genes, but if these same genes were to also be involved in the male pheromone the variation could increase again. As the male pheromone could be under stabilizing selection in a different direction, or perhaps under some other type of selection.
Genetics
In order to test this overlap and to then better understand how these sexual communication systems could evolve it is necessary to unravel the genetic architecture and find the genes that underlie the variation seen within a species. Looking at the genetic architecture is important because linkages between sets of genes involved in different systems can help to form a better idea of the evolution of these systems. For example, no linkages have been found between female pheromone genes and male response genes (Groot et al., 2016). Linkages between male pheromone genes and other genes have not been tested yet because the genes that underlie variation in the male sex pheromone are currently unknown. This study aims to identify these genes/genomic locations so that further linkage testing between sets of genes can take place. Also, if the genes involved in male and female pheromone production overlap this can be seen as evidence of some level of genetic entanglement.
Pheromone biosynthesis
As stated before the female pheromone is made up of a blend of different fatty acids of different lengths, but the male pheromone contains many of the same components or components that share a precursor. The difference between the pheromones is in the ratios of different components. In H. virescens the major critical component of the female pheromone is (Z)-11-hexadecenal (Z11–16:Ald)
4 (Roelofs et al., 1974) while the major component of the male pheromone is hexadecanyl acetate (16:Ac) (Teal & Tumlinson, 1989). Because the pheromones share many of the same components it is likely that the biosynthetic pathways are also (partly) shared. Groot et al. (2014) looked at variation in female pheromone ratios and found a major effect QTL corresponding with delta-11-desaturase. Because this delta-11-desaturase is responsible for an early step in the biosynthetic pathway of H. virescens (Groot et al., 2016) it can be argued that the delta-11-desaturase could also explain variation in the male pheromone ratios. Since the major components of the male and female pheromones are produced from the same precursor. This study aims to identify whether the male pheromone ratio in H. virescens is also regulated by the delta-11-desaturase gene.
Materials/Methods Phenotype
QTL analysis was performed to identify the genomic regions of H. virescens that underlie variation in the male pheromone. To perform this analysis it is necessary to have a well defined phenotype and adequate knowledge of the genotype. For the phenotype a number of different male pheromone component ratios and amounts of single compounds were used, a list of the chosen phenotypes can be found in figure 1. These phenotype values were obtained from H. virescens individuals that were a backcross from individuals of the High and Low lines as described in Groot et al. (2019). The High and Low lines differed from each other in the ratio of hexadecenal/(Z)-11-hexadecenal (16:Ald/Z11-16:Ald) in females, with High line females having a high ratio and Low line females having a low ratio of these components.
The means of the male components/ratios between the High and Low line were compared and I tested if they were significantly different. Based on the normality of the data I performed either a two sample T-test (for normally distributed data) or a Mann-Whitney U test (for non-normally distributed data) (see table 1 for a full list of tested phenotypes). The phenotype values that showed a significant difference between the male backcross individuals of the High and Low lines were used to perform QTL analysis. Since these selection lines were used by Groot et al. (2014) and Groot et al. (2019) to look at variation in the female H. virescens pheromone I also performed QTL analysis on the female ratio as a re-evaluation of the genetic architecture of the female sat/unsat ratio using genome-wide single nucleotide polymorphisms (SNPs) instead of AFLPs.
Genotype
The individuals used in this experiment had their genome sequenced using Restriction-site
Associated DNA sequencing (see Groot et al. (2019) for more details). Bioinformatics were used to align these sequences to the H. virescens reference genome and to call SNPs. SNPs were selected on the basis of completeness (hose with < 80% genotype completeness were removed), segregation distortion (highly distorted SNPs were removed) and quality (in case SNPs show identical segregation patterns, the SNP with the highest quality was kept). These SNPs where used to produce a genetic map using the R/qtl package (Broman et al., 2003). The SNPs were grouped using a LOD threshold of 3.5, followed by ripple using maximum likelihood in 5 marker windows with the following
5 With the map completed QTL analysis was performed on the female and male pheromone
phenotype values that showed significant differences between High and Low line individuals. Analysis was again performed using the R/qtl package (Broman et al., 2003). The function scanone() was used perform single QTL analysis, depending on the normality of the phenotype I used either a normal or non-parametric model and Haley-Knott regression was the method used for the analysis. Co-localization of the male and female phenotype values was tested by checking whether male and female QTL mapped to the same linkage group. The location of the delta-11-desaturase gene HvirLAPQ was identified using tblastn against the H virescens reference genome followed by checking the mapping of scaffolds on the linkage map. The looking for possible candidate genes underlying male pheromone variation was done using blastx against the database of known proteins on the NCBI website. Effect sizes of the detected QTL of the male phenotypes were calculated using the effectplot() function.
Results
Male phenotype
In order to test for a co-evolutionary response in the male pheromone following selection on the female ratio of 16:Ald to Z11:16-Ald, the means of the male components/ratios between the High and Low line were compared and I tested if they were significantly different. Based on the normality of the data I performed either a two sample T-test (for normally distributed data) or a
Mann-Whitney U test (for non-normally distributed data) (table 1) to see if the High and the Low line differed significantly regarding the specific variable. The phenotypes showing a significant difference between High and Low line males where then used for the QTL analysis. A graphical comparison of the selection lines/crosses for the most important phenotype values can be seen in figure 1.
Table 1: Results of two sample T-tests/Mann-Whitney U tests. Significant P-values (p < 0.05) are shown in bold.
Male phenotype Test type P value
Ratio saturated/unsaturated Two sample T-test 5.4e-10 Ratio 16:OAc/Z11-16:OAc Mann-Whitney U test 7.4e-11
Z11-16:OAc Mann-Whitney U test 1.8e-8
Z7-16:OAc Mann-Whitney U test 0.014
18:OH Mann-Whitney U test 0.779
16:OH Mann-Whitney U test 0.365
16:OAc Mann-Whitney U test 0.029
16:Ald Mann-Whitney U test 0.025
14:OH Mann-Whitney U test 0.392
14:OAc Mann-Whitney U test 0.061
Decalactone Mann-Whitney U test 0.963
MeSA Mann-Whitney U test 0.785
6
Figure 1: Male phenotype values for traits showing a significant difference between High and Low lines. The average phenotype values for the different lines and crosses was calculated and either a two sample T-test or Mann-Whitney U test was performed using the High and Low values. A) Ratio between saturated and
unsaturated pheromone components (two sample T-test, p < 0.001). B) Ratio between 16:OAc and Z11-16:OAc Whitney, p < 0.001). C) Amount of Z7-16:OAc Whitney, p = 0.014). D) Amount of 16:Ald (Mann-Whitney, p = 0.025).
Female pheromone QTL scans
A QTL scan of the female ratio saturated/unsaturated was performed as a re-evaluation of the genetic architecture of the female sat/unsat ratio in Groot et al. (2014) and Groot et al. (2019) using genome-wide SNPs instead of AFLPs. The QTL scan shows 2 almost significant (p < 0.20) QTL and 1 suggestive (p < 0.67) QTL (figure 2). The QTL are situated on linkage group 5 and 14, while the suggestive QTL is on linkage group 2. A QTL scan of this ratio was previously performed in Groot et al. (2019) using the same selection lines, where they found a QTL explaining most of the observed variation between the High and Low lines. Since they found a single locus that mapped to the delta-11-desaturase gene explaining nearly all of the variation between the lines, I used tblastn to check whether one of the linkage groups with a QTL on contained the 11-desaturase gene. The delta-11-desaturase gene was found on linkage group 5.
A second QTL scan was done on just the amount of 16:Ald in the female pheromone, the results of this scan can be seen in figure 3. While the QTL on linkage groups 5 and 14 are present when looking at just the 16:Ald the suggestive QTL on linkage group 2 is no longer present.
7
Figure 2: QTL plot of the female ratio saturated/unsaturated. The LOD scores for all markers along the 33 different linkage groups are shown using Haley-Knott regression. The black, blue and red lines represent the 0.67, 0.20 and 0.05 significance thresholds as determined by 1000 permutations.
Figure 3: QTL plot of the female 16:Ald amount. The LOD scores for all markers along the 33 different linkage groups are shown using Haley-Knott regression. The black, blue and red lines represent the 0.67, 0.20 and 0.05 significance thresholds as determined by 1000 permutations.
8 QTL scans were performed for the male pheromone phenotypes that showed a significant difference between the High and Low lines. These phenotypes include the ratio saturated/unsaturated, the ratio 16:OAc/Z11-16:OAc, the amount of Z7-16:OAc and the amount of 16:Ald. The aim was to test whether the same or different genomic regions are contributing to the indirect selection response in male phenotypes compared to the direct selection response in the females.
The QTL scans of the ratios sat/unsat and 16:OAc/Z11-16OAc identified the same peak on linkage group 7 at p < 0.67 and p < 0.20, respectively (Fig 4, 5). This peak did not co-localize with any of the QTL identified for the sat/unsat ratio in the females. In order to check for potential candidate genes affecting these ratios in the male pheromone I performed a blastx, however no easily recognized candidate genes were found on linkage group 7 (see supplemental figures). The QTL on linkage group 7 was found to explain around 45% (45.6% for the ratio sat/unsat and 45.0% for the ratio 16:OAc/Z11-16:OAc) of the difference between High and Low line (table 2).
Table 2: Effect sizes of the male QTL peaks.
Trait LG LOD AA mean (SD) AB mean (SD) % parental
difference explained Males sat/unsat 7 1.82 68.54 (12.09) 125.02 (13.35) 45.6 16:OAc/Z11-16:OAc 7 2.08 87.03 (34.02) 217.52 (37.41) 45.0 16:Ald 1 1.80 1.43 (0.13) 0.91 (0.11) 91.2 2 2.42 1.32 (0.10) 0.84 (0.14) 83.8 Z7-16:OAc 1 1.82 0.38 (0.03) 0.55 (0.05) 125.0 10 2.18 0.36 (0.03) 0.53 (0.04) 125.0 17 2.18 0.37 (0.03) 0.53 (0.04) 117.6 29 3.66 0.52 (0.04) 0.32 (0.03) 149.9
9
Figure 4: QTL plot of the male ratio 16:OAc/Z11-16:OAc. The LOD scores for all markers along the 33 different linkage groups are shown using non-parametric interval mapping. The black and blue lines represent the 0.67 and 0.20 significance thresholds as determined by 1000 permutations.
Figure 5: QTL plot of the male ratio saturated/unsaturated. The LOD scores for all markers along the 33 different linkage groups are shown using non-parametric interval mapping. The black and blue lines represent the 0.67 and 0.20 significance thresholds as determined by 1000 permutations.
When the relative amount of 16:Ald in the male pheromone is considered, a significant (p < 0.05) QTL on linkage group 2 and a suggestive (P < 0.67) QTL on linkage group 1 were found (figure 6). The
10 QTL on linkage groups 1 and 2 explain 91.2% and 83.8% of the difference between High and Low line individuals respectively (table 2).
Figure 6: QTL plot of the male 16:Ald amount. The LOD scores for all markers along the 33 different linkage groups are shown using non-parametric interval mapping. The black, blue and red lines represent the 0.67, 0.20 and 0.05 significance thresholds as determined by 1000 permutations.
Lastly, a QTL scan was performed on the amount of Z7-16:OAc in the male pheromone. This scan showed 3 suggestive QTL, on linkage groups 1, 10 and 17. There is also a significant QTL found on linkage group 29. All four of these QTL explained over 100% (125.0% for the QTL on linkage group 1 and 10, 117.6% for linkage group 17, and 149.9% for linkage group 29) of the difference between the High and Low line (table 2). The QTL on linkage group 29 worked in the opposite direction of those on the other linkage groups.
11
Figure 7: QTL plot of the male Z7-16:OAc amount. The LOD scores for all markers along the 33 different linkage groups are shown using Haley-Knott regression. The black, blue and red lines represent the 0.67, 0.20 and 0.05 significance thresholds as determined by 1000 permutations.
Discussion
The aim of this study was to determine which genes/genomic regions underlie the variation in the male pheromone of H. virescens. Further, I wanted to find out whether these regions overlap with the delta-11-desaturase gene, as this is the main gene underlying variation of the female
pheromone. QTL analysis of the female ratio showed several peaks with the peak on linkage group 5 corresponding to delta-11-desaturase. The QTL analysis of the different male pheromone
components showed several peaks, however no peak was found on linkage group 5. This suggests that the variation in the male pheromone cannot be explained by the same genes as the variation in the female pheromone. The QTL, and therefore the genetics, showed no overlap between the male and female pheromone. However, it should be noted that I found significant differences in the phenotype values of the High and Low line males. This is despite the fact that the selection lines were created to get high and low values for the female ratio. Hence, this suggests that there has been an indirect selection response in the males.
Both the female ratio and the male 16:Ald amount showed a QTL peak on linkage group 2, this would suggest overlap between the male and female pheromones. However, after performing a separate QTL scan on only the female 16:Ald amount in which the peak on linkage group 2 was no longer present I conclude that the peaks on linkage group 2 coincidental.
The QTL scans for the male ratios both showed a peak on linkage group 7, indicating that this linkage group is important for the variation in the male pheromone. Because this was the only peak found I decided to look for candidate genes on this linkage group using blastx. I was unable to find any
12 probable candidate genes on this linkage group. This could be because the genes on this linkage group have secondary effects I am currently unaware of or because there are undiscovered genes situated on linkage group 7.
For the amount of Z7-16:OAc a significant QTL was found on linkage group 29, this is interesting since Z7-16:OAc is not produced using delta-11-desaturase but rather by delta-9-desaturase. No genetic overlap was found between the female and male pheromones in this study, so this does not support my hypothesis that the higher than expected variation in the female pheromone could be the result of conflicting selections for the genetically linked male and female pheromones. However the fact that the High and Low lines males differed significantly for multiple phenotype values shows that there is an indirect selection response in the males while the selection lines where selecting for the female pheromone. This suggests that there is something going on that I did not find during analysis. If we combine this with the fact that the sequence data contained a larger amount of sequencing mistakes than expected it is fair to assume that the data used for the QTL analysis was not of high quality. Because of this I would say that no matter what the outcome of the study would have been it has to be taken with a grain of salt. In order to find out more about this indirect selection response in the males it is necessary to perform similar research on a dataset that contains fewer sequencing mistakes.
References
• Ando, T., Inomata, S., Yamamoto, M. (2004). Lepidopteran Sex Pheromones. In: Schulz S. (eds) The Chemistry of Pheromones and Other Semiochemicals I. Topics in Current Chemistry, vol 239. Springer, Berlin, Heidelberg. (https://doi.org/10.1007/b95449). • Baker, T. C., Nishida, R., Roelofs, W. L. (1981). Close-Range Attraction of Female Oriental
Fruit Moths to Herbal Scent of Male Hairpencils. Science. 214 (4527), 1359-1361. (doi:10.1126/science.214.4527.1359).
• Broman et al. (2003) R/qtl: QTL mapping in experimental crosses. Bioinformatics 19:889-890 doi:10.1093/bioinformatics/btg112.
• Groot, A. T., Schofl, G., Inglis, O., Donnerhacke, S., Classen, A., Schmalz, A., et al. (2014). Within-population variability in a moth sex pheromone blend: genetic basis and behavioural consequences. Proc. R. Soc. B 281: 20133054. (https://doi.org/10.1098/rspb.2013.3054) • Groot, A. T., Dekker, T., Heckel, D. G. (2016). The Genetic Basis of Pheromone Evolution in
Moths. Annu. Rev. Entomol. 61, 99-117. (doi:10.1146/annurev-ento-010715-023638) • Groot, A. T., van Wijk, M., Villacis-Perez, E., Kuperus, P., Schofl, G., van Veldhuizen, D.,
Heckel, D. G. (2019). Within-population variability in a moth sex pheromone blend, part 2: selection towards fixation. R. Soc. open sci. 6:182050.
• Hendricks, D. E., & Shaver, T. N. (1975). Tobacco Budworm: Male Pheromone Suppressed Emission of Sex Pheromone by the Female. Environmental Entomology. 4 (4), 555-558. (https://doi.org/10.1093/ee/4.4.555).
• Hillier, N. K., & Vickers, N. J. (2004). The Role of Heliothine Hairpencil Compounds in Female Heliothis virescens (Lepidoptera: Noctuidae) Behavior and Mate Acceptance. Chemical Senses. 29 (6), 499–511. (doi:10.1093/chemse/bjh052).
13 • Hosseini, S. A., van Wijk, M., Ke, G., Goldansaz, S. H., Schal, C., Groot, A. T. (2016).
Experimental evidence for chemical mate guarding in a moth. Sci. Rep. 6, 38567. (doi:10.1038/srep38567).
• Roelofs, W. L., Hill, A. S., Cardé, R. T., Baker, T. C. (1974). Two sex pheromone components of tobacco budworm moth, Heliothis virescens. Life Sci. 14, 1555 – 1562. (doi:10. 1016/0024-3205(74)90166-0).
• Teal, P. E. A., & Tumlinson, J. H. (1989). Isolation, identification, and biosynthesis of compounds produced by male hairpencil glands of Heliothis virescens (F.) (Lepidoptera: Noctuidae). Journal of Chemical Ecology. 15 (1), 413-427.
• Tumlinson, J. H., Hendricks, D. E., Mitchell, E. R., Doolittle, R. E., Brennan, M. M. (1975). Isolation, identification and synthesis of the sex pheromone of the tobacco budworm. J. Chem. Ecol. 1, 203– 214. (doi:10.1007/ BF00987869).
• Vetter, R. S., & Baker, T. C. (1983). Behavioral responses of male Heliothis virescens in a sustained flight tunnel to combinations of seven compounds identified from female sex pheromone glands. J. Chem. Ecol. 9, 747– 759. (doi:10.1007/BF00988780).
• Vickers, N. J., & Baker, T. C. (1994). Reiterative responses to single strands of odor promote sustained upwind flight and odor source location by moths. Proc. Natl. Acad. Sci. USA. 91, 5756-5760.
• Zhu, J., Chastain, B. B., Spohn, B. G., Haynes, K. F. (1997). Assortative Mating in Two
Pheromone Strains of the Cabbage Looper Moth, Trichoplusia ni. Journal of Insect Behavior. 10 (6), 805-817.
