The loci of behavioral evolution: Fas2 and tilB underlie differences in pupation site choice behavior between Drosophila melanogaster and D. simulans

The recent boom in genotype-phenotype studies has led to a greater understanding of the genetic architecture of a variety of traits. Among these traits, however, behaviors are still lacking, perhaps because they are complex and environmentally sensitive phenotypes, making them difficult to measure reliably for association studies. Here, we aim to fill this gap in knowledge with the results of a genetic screen for a complex behavioral difference, pupation site choice, between Drosophila melanogaster and D. simulans. In this study, we demonstrate a significant contribution of the X chromosome to the difference in pupation site choice behavior between these species. Using a panel of X-chromosome deletions, we screened the majority of the X chromosome for causal loci, and identified two regions that explain a large proportion of the X-effect. We then used gene disruptions and RNAi to demonstrate the substantial effects of a single gene within each region: Fas2 and tilB. Finally, we show that differences in tilB expression underlie species differences in pupation site choice behavior, and that generally, pupation site choice behavior appears to be correlated with relative expression of this gene. Our results suggest that even complex, environmentally sensitive behaviors may evolve through changes to loci with large phenotypic effects. Author summary Behaviors are complex traits that involve sensory detection, higher level processing, and a coordinated output by the nervous system. This level of processing is highly susceptible to environmentally induced variation. Because of their complexity and sensitivity, behaviors are difficult to study; as a result, we have very little understanding of the genes involved in behavioral variation. In this study, we use common laboratory fruit fly model, Drosophila, to address this gap and dissect the genetic underpinnings of an environmentally sensitive behavior that differs between species. We find that a significant amount of the phenotypic difference between species is explained by a single chromosome. We further show that just two genes on this chromosome account for a large majority of its effect, suggesting that the genetic basis of complex behavioral evolution may be simpler than anticipated. For one of these genes, we show that a species-level difference in gene expression is associated with the difference in behavior. Our results contribute to a growing number of studies identifying the genetic components of behavior. Ultimately, we hope to use these data to better predict the number, types, and effects of genetic mutations necessary for complex behaviors to evolve.

Introduction which we would expect for genes regulating behavior (48). Five of these six genes are well 258 described. At the time of assay, only two of the five characterized genes within this region had 259 non-lethal verified loss-of-function alleles available: Fas2 (Fasciclin2) and mei9 (meiotic 9). We 260 tested knockouts of each for an effect on pupation site choice behavior. We found no significant 261 difference in the pupation index obtained when we crossed the mei-9 A1 mutant allele to the Lhr D. found that Fas2 eb112 hybrids had a significantly higher pupation index than the D. melanogaster 266 knockouts (Wilcoxon test: n= 50-51, c 2 = 6.97, p= 0.0083; Fig 5A), suggesting that Fas2 may be 267 involved in pupation site choice. To ensure this pattern is not unique to these strains, we crossed 268 Fas2 eb112 to additional D. simulans (Mex180) and D. melanogaster (T.1) wild-type strains. We 269 again found the same pattern: the pupation index for knockout hybrids was significantly higher 270 than for D. melanogaster knockouts (Wilcoxon test: n= 52-53, c 2 = 27.2, p<0.0001; Fig 5A). As 271 further verification, we tested a second Fas2 strain: a p-element insertion allele, Fas2 G0293 , and 272 similarly found that the pupation index for Fas2 G0293 hybrids (crossed to Lhr) was significantly 273 higher than that for D. melanogaster knockouts (crossed to T.4; Wilcoxon test: n= 52-53, c 2 = 274 9.73, p=0.0018; Fig 5B). When we used the combined consensus p-value test (49) to look at the 275 overall pattern for Fas2 eb112 , and Fas2 as a whole (i.e. including results from both Fas2 eb112 and 276 Fas2 G0293 ), we found a strongly significant pattern of higher pupation indices for hybrid crosses 277 compared to D. melanogaster crosses (Fas2 eb112 : p= 3.59 x 10 -6 ; Fas2: p= 2.35 x 10 -8 ). 278 Next, we used RNAi with the elav-Gal4 driver to reduce expression of Fas2 throughout 279 the nervous system in D. melanogaster. We compared the proportion of experimental flies that 280 pupated on the food for the RNAi cross (UAS-Fas2 x elav-Gal4) to that of two controls (both 281 crossed to elav-Gal4): the background stock in which the RNAi lines were created (y v; attP2, 282 y+) and the Gal4-1 stock, which has a hairpin targeting Gal4 in VALIUM20 to control for Gal4 283 effects. We found that a significantly higher proportion of RNAi flies pupated on the food 284 compared to the control flies from either the background (Wilcoxon test: n= 42-48, p<0.0001 285 after sequential Bonferroni correction) or Gal4-1 cross (Wilcoxon test: n= 42-43, p<0.0001 after 286 sequential Bonferroni correction; Fig 5C). These results are unchanged when we control for 287 density effects (Fig S5A), providing further evidence for Fas2's role in pupation site choice. 288 289

Gene knockouts and RNAi knockdown suggest that tilB is involved in divergent pupation 290
behavior 291 The second region of interest identified by our deficiency screen was the region deleted by 292 DF(1)Exel6255 (X:21,519,203 -22,517,665; Fig 4D). Within this region are 28 genes, of which 293 22 are protein coding (Table S3B). Of the 22 protein coding genes, 14 are expressed in larvae -294 13 of which have some expression in the larval nervous system (48). Of these, 7 are described. 295 We obtained knockout strains for both of the characterized genes expressed in the larval nervous 296 system that had verified loss-of-function alleles available at the time: tilB (touch insensitive larva 297 B) and wap (wings apart). We found no significant difference in the pupation index obtained 298 when we crossed the wap 2 mutant allele to the Lhr D. simulans strain and the T.4 D. 299 melanogaster strain (Wilcoxon test: n= 51-55, c 2 =0.32, p=0.57; Fig S4A), indicating that wap is 300 unlikely to be involved in pupation site choice. In contrast, when we crossed the tilB 1 and tilB 2 301 mutant alleles to D. simulans (Lhr) and D. melanogaster (T.4), we found that the tilB knockout 302 hybrids had significantly higher pupation indices than the D. melanogaster knockouts for both 303 alleles (Wilcoxon tests; tilB 1 : n= 56, c 2 = 6.61, p= 0.0101; Fig 6A; tilB 2 : n= 57, c 2 = 6.61, p= 304 0.0101; Fig 6B). To test whether this difference is consistent in other backgrounds, we crossed 305 both the tilB 1 and tilB 2 mutant alleles to additional D. simulans (Mex180) and D. melanogaster 306 (T.1) wild-type strains. We had a difficult time crossing our tilB strains to the Mex180 strain, so 307 our sample sizes for these crosses are smaller, but there is a nonsignificant trend towards a higher 308 pupation index for the knockout hybrids compared to the D. melanogaster hybrids for both 309 alleles (Wilcoxon tests; tilB 1 : n=36-37, c 2 = 2.80, p= 0.0943, Fig 6A; tilB 2 : n=12-18, c 2 = 3.44, p= 310 0.0638, Fig 6B). We used the combined consensus p-value test (49) to look at the overall pattern 311 for tilB 1 , tilB 2 , and tilB as a whole (i.e. including results from both tilB 1 and tilB 2 ), and found a 312 strongly significant pattern of higher pupation indices for hybrid crosses compared to D. 313 melanogaster crosses (tilB 1 : p= 0.0022; tilB 2 : p= 0.0037; tilB: p= 2.47 x 10 -5 ). 314 As for Fas2 above, we then used RNAi with the elav-Gal4 driver to reduce expression of 315 tilB throughout the nervous system in D. melanogaster. We found that a significantly higher 316 proportion of RNAi flies pupated on the food compared to the control flies from either the 317 background (Wilcoxon test: n= 46-47, p<0.01 after sequential Bonferroni correction) or Gal4-1 318 cross (Wilcoxon test: n= 41-46, p<0.01 after sequential Bonferroni correction; Fig 6C), These 319 findings are consistent when we control for density effects (Fig S5B), providing additional 320 support for tilB's role in pupation site choice. 321 322 6. tilB is expressed more highly in D. melanogaster strains 323 We performed RT-qPCR to quantify relative tilB transcript abundance for two strains of D. 324 simulans (Per005 and Geo288) and two strains of D. melanogaster (CA1 and T.4). For the two D. 325 simulans strains and the Geo288 D. melanogaster strain, we collected larvae from two stages of larval development (96 and 120 hours following oviposition). For the other D. melanogaster 327 strain (T.4), we were only able to obtain enough larval tissue at 96 hours following oviposition, 328 due to low fecundity. Because the relative transcript abundance data had a skewed distribution, 329 we used the reciprocal root transformation to normalize the species and overall distributions 330 (Shapiro-Wilk tests for normality: all p >0.34). We then analyzed the relative transcript 331 abundance of tilB using a nested ANOVA with the following factors: species, strain nested 332 within species, larval age (96 and 120 hours), and the interaction between species and larval age. 333 The interaction term between species and larval age was not significant (p=0.41), so we removed 334 it from the model. We found that larvae from the 120-hour sampling period had significantly 335 lower tilB expression than 96-hour larvae (F1,14= 6.16, p= 0.026). While we did not detect any 336 significant differences between the 2 strains from the same species (F2,14=2.39, p=0.13), we 337 found a significantly higher average relative amount of tilB transcript in D. melanogaster larvae 338 compared to D. simulans larvae (F1,14= 9.74, p= 0.0075; Fig 7). 339 Because we performed rt-PCR on an extreme strain and a strain closer to the average for 340 each species, these four lines represent a continuum of pupation site choice behavior, with T.4 (D. 341 melanogaster) having the lowest proportion of pupae on the food, followed by CA1 (D. 342 melanogaster), then Per005 (D. simulans) and Geo288 (D. simulans) having the highest 343 proportion of pupae on the food (Fig 1). These four strains follow an identical pattern for tilB 344 gene expression, with T.4 having the highest relative transcript abundance, and Geo288 having 345 the lowest (Fig 7). Although it is not possible to detect a significant effect with a sample size of 4, 346 this suggests that tilB gene expression may be negatively correlated with the proportion of pupae 347 on the food (Spearman's rank correlation: rs = -1, p< 0.10). 348

Effect sizes 350
We used the results above to estimate how much of the difference in pupation site preference 351 between D. melanogaster and D. simulans can be attributed to: i) the X chromosome, ii) our 352 deficiencies of interest (Df(1)BSC869/DF(1)ED6720 and Df(1)Exel6255), and iii) Fas2 and tilB. 353 We first calculated the "species difference ratio" (using the data from our parental/hybrid screen 354 in Fig 3) by dividing the median proportion of males on the food for D. simulans by the median 355 proportion of males on the food for D. melanogaster males (species difference ratio = 5.27). To 356 determine how much of this species difference ratio can be attribute to the X chromosome, we 357 calculated an "X effect ratio" by dividing the median proportion of simX males on the food by 358 the median proportion of melX males on the food. By then dividing the "X effect ratio" by the 359 "species difference ratio", we found that the X chromosome accounts for approximately 55.6% 360 (95% CI= 31.4%-80.2%) of the difference in pupation site preference between D. melanogaster 361 and D. simulans. It is important to note, however, that this may be an overestimate, as calculating 362 effect sizes using only reciprocal hybrids does not account for potential transgressive autosomal 363 effects (50). 364 We then estimated the effect size of our deficiencies by calculating "deficiency effect 365 ratios" (the median pupation index when the deficiency was crossed to the D. simulans Lhr 366 strain/the median pupation index when the deficiency was crossed to the D. melanogaster T.4 367 strain). Because Df(1)BSC869 and DF(1)ED6720 overlap, we used their mean deficiency effect 368 ratio to estimate the effect size of the overlapping region, and found that it explains 369 approximately 44.1% (95% CI= 33.5%-74/5%) of the X chromosome effect. Similarly, we found 370 that Df(1)Exel6255 explains approximately 52.3% (95% CI= 33.7%-85.9%) of the X 371 chromosome effect. 372 Finally, we estimated the effect size of our two identified candidate genes, Fas2 and tilB. 373 For each allele we tested, we calculated a "knockout effect ratio" (the median pupation index 374 when the knockout was crossed to the D. simulans Lhr strain/the median pupation index when 375 the knockout was crossed to the D. melanogaster T.4 strain) and then we used the average 376 knockout effect ratio of the two alleles for each gene (i.e. Fas2: average of Fas2 eb112 and 377 Fas2 G0293 ; tilB: average of tilB 1 and tilB 2 ) to estimate the effect size. We found that However, we used 11 D. melanogaster and 12 D. simulans strains sourced from around the globe 396 to demonstrate this difference. While the species difference holds when comparing the grand 397 mean of all strains for each species, there is substantial variation within species. This variation is 398 so significant that the species' distributions overlap, with some D. simulans strains, like Mex180 399 and Cal006, more closely resembling D. melanogaster strains in pupation behavior (Fig 1). To 400 our knowledge, our study is the first to extensively record this variation, which presents a useful 401 tool for better understanding the evolution of pupation site choice behavior. These documented 402 differences in pupation behavior among Drosophila species, in combination with our 403 understanding of the environmental variables that affect this behavior within species 404 (24,25,47,51), will be useful in identifying the selection pressures (if any) that affect the 405 evolution of this trait. Differences in pupation site choice behavior may be a form a niche 406 partitioning where species co-occur, as has been previously suggested (34). Alternatively, 407 pupation site choice may be an adaptive response to parasite or parasitoid presence (52). A 408 globally sourced panel of lines with significant variation, such as our own, provides an inroad for 409 studies comparing pupation behavior to differences in the ecology of each collection site, such 410 that we can better understand the ultimate causes of this behavioral evolution. 411

412
The genetic architecture of pupation site choice behavioral evolution 413 Using hybrid crosses in the same background that controlled for maternal inheritance, we were 414 able to estimate the effect of the X chromosome on pupation behavior. We found a significant 415 effect of the X, in that reciprocal hybrid males pupate in similar locations as the X-donating 416 parent. We calculate that this chromosome explains 55.6% (95% CI= 31.4%-80.2%) of the total 417 phenotypic difference between parent strains. Although the exact contribution of the X 418 chromosome reported here may be an overestimate due to transgressive autosomal effects that 419 cannot be detected in a hybrid background (50), a similar X-effect has been detected for pupation 420 behavior when comparing D. simulans and D. sechellia (37). 421 We also found that hybrid females, which inherit one X chromosome from each parent, Further, we show that this is not simply an effect of hemizygosity within this region in two ways. 447 First, we demonstrate that this effect is lost when we crossed these knockout strains to D. Here we show that a substantial amount of the difference in pupation site choice behavior 490 between D. melanogaster and D. simulans is attributed to the X chromosome. This difference is 491 not, however, entirely explained by the X, and further study of the autosomal genome using the 492 same approach may reveal additional genes with substantial effects. 493 Using engineered chromosomal deletions, we identify two regions on the X chromosome 494 associated with pupation site choice behavior. We then use gene knockouts and RNAi to show 495 that the large majority (if not entire) effect of each of these regions is explained by a single gene. 496 Overall, this indicates that individual genes can have substantial effects on behavioral differences.

General fly maintenance
Unless otherwise stated, we maintained all fly strains for these experiments in 20 mm diameter 511 vials containing standard cornmeal-molasses-yeast medium at 25°C under a 12h:12h light/dark 512 cycle at 50% relative humidity. Under these conditions, we established non-overlapping two-513 week lifecycles as follows. For all stocks, except LHM and Lhr (see below), we transferred male 514 and female adult flies into fresh vials containing food media supplemented with live yeast on the 515 surface for 1-3 days, at which point the flies were discarded. 14 days later (after all progeny had 516 eclosed), we again transferred adult flies into fresh vials for 1-3 days to begin the next generation. 517 We maintained LHM and Lhr identically, except we additionally regulated density by transferring 518 only 10 males and 10 females to begin the next generation. 519 520

Characterizing pupation behavior for D. melanogaster and D. simulans 521
We measured pupation behavior for 11 D. melanogaster and 12 D. simulans strains collected 522 from various locations throughout the world (Table S1) To measure pupation behavior, we placed 10 males and 10 females from a specific line 529 (both 3-5 days old) into half pint bottles and allowed females to oviposit overnight on a 10 mm 530 diameter petri dish filled with food medium that was placed in the opening of the bottle. In total, 531 we set up 5 bottles for each line. The following morning, we transferred 100 eggs from the petri 532 dishes into vials containing food medium that were lined with an acetate sleeve on which the 533 larvae could pupate. In total, we set up 5-8 vials per line. Vials were held at 25°C for 8 days, at 534 which time the liner was removed and the locations of the pupae were recorded (8 days was long 535 enough for almost all larvae to pupate without any flies eclosing). A pupa was considered "on" 536 the food if it was within 1 cm of the food surface, while all pupae that were further than 1 cm 537 from the food surface were considered "off" the food. supplemented with an ad lib amount of live yeast on the surface. We then pushed a long foam 549 plug down into the vial, leaving approximately 1 cm of space above the food surface. We held 550 flies under these conditions for 3 days, at which time they were transferred from these "cross 551 vials" into "pupation vials" that contained food medium with no added yeast, and were lined 552 with an acetate sleeve on which the larvae could pupate. We always set up crosses using D. 553 melanogaster females and D. simulans males, because crosses in the opposite direction were 554 never successful. 555

Measuring pupation behavior in F1 hybrids 557
To create F1 hybrid males and females, we used D. simulans males from the Lethal hybrid 558 rescue (Lhr) strain (36). The Lhr mutation restores viability in F1 hybrid males, which are 559 usually lethal (69). To create F1 hybrid females and F1 males with a D. melanogaster X 560 chromosome ("melX" males), we crossed wild-type females from our LHM strain (provided by 561 Dr. William Rice) to Lhr D. simulans males (Fig 2A). Because we were unable to successfully 562 cross D. simulans females to D. melanogaster males, we created F1 hybrid males with the D. 563 simulans X chromosome ("simX" males) by crossing D. melanogaster LHM females that carry a 564 compound X chromosome (C(1)DX y f) (68) to D. simulans Lhr males. The compound X in these 565 females ensured that the X chromosome was transmitted from D. simulans fathers to their F1 566 hybrid sons (Fig 2B). Thus, for each direction of the cross, we combined D. melanogaster 567 females with D. simulans males, as described above. This crossing scheme ensures that all 568 maternal inheritance (cytoplasmic and mitochondrial) in the reciprocal male hybrid crosses 569 originates from the D. melanogaster parent. Thus, any differences we observe between melX and 570 simX males are directly attributable to their different X chromosomes. 571 After 3 days in the cross vial, we transferred males and females into pupation vials for 24 572 hours, at which time the flies were removed. While screening hybrid pupation behavior, we also 573 concurrently screened pupation behavior for the parental D. melanogaster strain (LHM) and the 574 parental D. simulans strain (Lhr) for comparison. Parental strain cross vials contained only a 575 moderate amount of yeast, were set up with only 5 males and 5 females (pure species crosses 576 produce more offspring), and did not have a plug pushed down into the vial, but were otherwise 577 treated identically to the hybrid crosses. In total, we set up 30-33 vials per treatment. 578 All pupation vials were held at 25°C for 8 days, at which time the liner was removed. We 579 removed any remaining larvae, and cut the liner at a point 1 cm above the food surface. The 580 portion of the liner that contained pupae within 1 cm of the food surface was returned to the 581 original vial (the "on vial"), while the portion of the liner with pupae further off of the food 582 surface was placed in another vial containing food medium (the "off vial"). The flies that eclosed 583 were sexed and counted 7 days later (15 days post-egg); all flies that eclosed in the "on vial" 584 were considered flies that pupated on the food surface, while all flies that eclosed in the "off 585 vial" were considered flies that pupated off the food surface. We then calculated the proportion 586 of individuals that pupated on the food for each type of individual (genotype and sex). 587 588

Mapping hybrid pupation behavior using the Bloomington Deficiency Kit 589
Because we found a significant effect of the X chromosome on the difference in pupation site 590 choice behavior between D. simulans and D. melanogaster, we devised a crossing scheme using 591 molecularly engineered chromosomal deficiencies to screen the X chromosome for loci 592 contributing to this difference. These deficiencies are part of the Bloomington Deficiency Kit 593 (38), available from the Bloomington Drosophila Stock Center (BDSC). We assayed a total of 90 594 deficiency strains covering 87% of the X chromosome (Table S2). We restricted our deficiency 595 screen to lines from the BSC, Exelixis, and DrosDel sets to control for strain background effects 596 while also maximizing chromosome coverage. 597 To set up crosses, we collected deficiency females as young virgins (2-3 hours after 598 eclosing) and crossed them to D. simulans males from the Lhr strain. After 3 days in the cross 599 vial, we transferred males and females into pupation vials for 24-48 h, at which time the flies were removed. We then divided the pupations vials into "on" and "off" vials as we did for our F1 601 hybrids (above). 602 These crosses produced two types of hybrid female that were heterozygous for D. 603 melanogaster/D. simulans at each autosome (Fig 2C). food and the proportion of balancer hybrid females that pupated on the food. We then used these 614 measures to calculate a "pupation index" as the proportion of deficiency hybrids pupating on the 615 food divided by the proportion of balancer females pupating on the food. To increase the 616 accuracy of our estimates, we only included pupation vials in our analysis that yielded at least 10 617 of each type of female. For each deficiency hybrid strain we measured, we report the median 618 pupation index of all replicates, because there were often high-scoring outliers that significantly 619 skewed the mean pupation index. These outliers almost always had abnormally high pupation 620 indices, so focusing on median values makes our findings more conservative. 621 Any deficiency hybrid cross with a median pupation index greater than 1 indicates that 622 more deficiency females pupated on the food compared to balancer females, potentially because 623 the deficiency includes D. melanogaster genetic variation that is involved in pupation site choice 624 behavior. Alternatively, simply creating flies that are hemizygous at a locus on the X 625 chromosome may result in a variety of pleiotropic effects that make larvae less likely to climb up 626 the vial. To test for this, when a deficiency hybrid cross showed a pupation index significantly 627 greater than 1 (Table S2) P{lacW}fas2 G0293 /FM7c), and the mei-9 A1 mutant allele (w 1 mei-9 A1 /FM7h; BDSC stock #6792).
For the second region, we obtained knockouts for two genes: tilB and wap. We screened two tilB 647 mutant alleles, tilB 1 and tilB 2 (y w tilB 1/2 /FM4; Kernan et al., 1994; provided by Daniel Eberl), 648 and the wap 2 mutant allele (wap 2 /FM6; DSC stock # 8133). Like the deficiency strains, each of 649 our gene disruptions is held over a balancer chromosome with a visible marker. To measure the 650 pupation behavior of hybrids containing knockout copies of these D. melanogaster genes, we 651 crossed each D. melanogaster knockout strain to Lhr using the previously described methods, 652 and calculated the pupation index as the proportion of knockout females on food / the proportion 653 of balancer females on the food. As for our deficiency screen, any hybrid knockouts with a 654 median PI greater than 1 for the hybrid cross suggest that the knockout gene may be involved in 655 pupation site choice. We also crossed each D. melanogaster knockout strain to a wild-type D. 656 melanogaster (T.4) strain to ensure that hybrid knockouts with a pupation index greater than 1 657 are not simply an artifact of being hemizygous for this particular gene. We crossed knockout 658 strains that displayed the pattern we expect for a gene involved in pupation site choice (i.e. a 659 pupation index significantly greater than 1 when crossed to D. simulans, which is also 660 significantly greater than the pupation index when crossed to D. melanogaster) to an additional 661 D. simulans (Mex180) and D. melanogaster (T.1) wild-type strain for verification. 662 Our knockout screen identified two genes that appear to be involved in pupation site 663 choice: tilB and Fas2. We further tested the effects of these genes on pupation behavior using 664 we detect between these controls and our RNAi crosses must therefore be due to the expression 677 of the gene-specific (tilB or Fas2) hairpin RNA. We set up pupation vials using the methods 678 described above, and for each cross, we calculated the proportion of RNAi (or control) flies on 679 the food (removing any data points with fewer than 20 experimental flies). If more RNAi flies 680 pupate on the food in the experimental cross (in which the expression of the gene is driven down) 681 compared to the control crosses (in which gene expression is unaffected), this provides further 682 support for that gene's involvement in pupation site choice. 683 684 Testing candidate genes for species-specific differences in larval transcript expression 685 We selected two each of our 11 D. melanogaster and 12 D. simulans strains to test for larval 686 stage-specific expression differences of candidate genes using real time RT-PCR -one extreme 687 and one average. For D. simulans, we selected Geo288 and Per005 (Table S1), because Geo288 688 has the highest pupation index of any D. simulans strain and Per005 is closest to the species 689 mean (Fig 1). For D. melanogaster, we selected CA1 and T.4 (Table S1) because T.4 has the 690 lowest pupation index of any D. melanogaster strain, and CA1 is closer to the species mean (Fig  691   1). 692 To harvest larvae from these strains, we allowed adult females to oviposit in standard 693 vials containing food media between the hours of 8 AM and 12 PM over two consecutive days. 694 120 hours after the final oviposition day, we floated larvae out of the food media using a 20% 695 sucrose in water solution, sucked them up using a transfer pipet, briefly rinsed them with DI 696 water on cheesecloth, and snap froze them using liquid nitrogen (72). In this way, we collected 697 20-30 mg of larvae from two developmental time points: 96, and 120 hours following 698 oviposition. These time points approximate early wandering and late wandering larval stages. 699 We chose these time points because they are presumably when pupation site choice occurs, and 700 because larvae are large enough for many to be harvested at once using the above methods. We 701 extracted mRNA using the Qiagen RNeasy Plus Mini Kit, and prepared cDNA using the 702 Promega Verso kit. 703 To quantify transcript abundance, we designed primers that span a single intron near the 704 3' end of tilB (73). Additionally, we used primers for the gene RpL32 as an internal control 705 (74,75). Fas2 is a complex gene with multiple splice forms, so we were unsuccessful in 706 designing general primers that would amplify all transcripts in both species. For this reason, we 707 did not include Fas2 in these experiments. A full list of primers and transcript lengths can be 708 found in Table S4A. For each stage and strain, we prepared two to three biological replicates, 709 which we then amplified in two technical replicates for 40 rounds of qPCR. Using RpL32 710 transcript number as an internal control, we calculated relative transcript abundance while 711 correcting for species differences in primer efficiency. We estimated primer efficiency 712 differences by serially diluting gDNA from each of our D. melanogaster and D. simulans strains, 713 performing qPCR, and using a standard curve to calculate adjusted amplification factors (Table  714 S4B). To ensure that we were amplifying cDNA made from RNA, and not gDNA contamination,