Comparative functional and evolutionary analysis of essential germline stem cell genes across the genus Drosophila and two outgroup species

27 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 2 In Drosophila melanogaster, bag of marbles (bam) encodes a protein essential for 28 germline stem cell daughter (GSC) differentiation in early gametogenesis. Despite its 29 essential role in D. melanogaster, direct functional evaluation of bam in other closely 30 related Drosophila species reveal this essential function is not necessarily conserved. In 31 D. teissieri, for example, bam is not essential for GSC daughter differentiation. Here, we 32 generated bam null alleles using CRISPR-Cas9 in a species more distantly related to D. 33 melanogaster, D. americana, to interrogate whether bam’s essential GSC differentiation 34 function is novel to the melanogaster species group or a function more basal to the 35 Drosophila genus. To further characterize the extent of the functional flexibility of other 36 GSC regulating genes, we generated a gene ortholog dataset for 366 GSC regulating 37 genes essential in D. melanogaster across 15 additional Drosophila and two outgroup 38 species. We find that bam’s essential GSC function is conserved between D. 39 melanogaster and D. americana and therefore originated prior to the formation of the 40 melanogaster species group. Additionally, we find that ~8% of the 366 GSC genes 41 essential in D. melanogaster are absent in at least one of the 17 species in our ortholog 42 dataset. These results indicate that developmental systems drift (DSD), in which the 43 specific genes regulating a function may change, but the final phenotype is retained, 44 occurs in stem cell regulation and the production of gametes across Drosophila species. 45 46 Article summary 47

from CRISPR induced bam null mutants in D. americana and comparative 48 ortholog analysis of essential GSC regulating genes indicate that the evolutionary origin 49 of bam’s essential GSC differentiation function is likely basal to the Drosophila genus, 50 and there is functional flexibility in at least ~8% of the 366 GSC regulating genes across 51 the 17 included species. 52 53

bam, germline stem cells, comparative functional analysis, CRISPR 54 55 56 57 58

59 Proper production of gametes is critical for reproduction, and in Drosophila it begins with 60 the asymmetric division of germline stem cells (GSCs) to both self-renew, maintaining 61 the germline, and the ultimate differentiation into sperm and eggs (Kahney et al. 2019). 62 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 3 Mis-regulation of this highly sensitive process can quickly lead to sterility, so 63 coordination of these early cellular divisions might be presumed to be highly conserved 64 (Gleason et al. 2018). However, recent results have demonstrated that several GSC 65 regulating genes essential for reproduction in D. melanogaster show signs of positive 66 selection through rapid amino acid diversification, are nonessential for fertility, and/or 67 are completely absent in non-melanogaster Drosophila species (Civetta et al. 2006; 68 Bauer DuMont et al. 2007; Bubnell et al. 2022), DuMont et al. 2021, Choi et al. 2015, 69 Flores et al. 2015. 70 One of these GSC regulating genes is bag-of-marbles (bam). The encoded protein Bam 71 is 442 amino acids with several known functions that are executed in complexes with 72 other protein partners in D. melanogaster. The most well characterized of Bam’s 73 functions is as the switch for GSC daughter differentiation, but Bam also has 74 documented roles in the maintenance of gut integrity, and as a switch for preventing 75 premature differentiation of hematopoietic progenitor cells (McKearin and Spradling 76 1990; Insco et al. 2009; Tokusumi et al. 2011). 77 Bam specifically acts as the switch gene for GSC differentiation for D. melanogaster 78 females and is necessary for terminal differentiation of spermatogonia in males 79 (McKearin and Spradling 1990). In females, bam is repressed in GSCs and bam 80 expression causes differentiation by binding to several protein partners including benign 81 gonial cell neoplasm (Bgcn) in order to repress the production of self-renewal factors 82 Nanos and elF4a. The resultant differentiating cystoblast undergoes several mitotic 83 divisions (Shen et al. 2009, Ohlstein et al. 2000, Li et al. 2013, Li et al. 2009). 84 Simultaneously, Bam concentrates at the fusome, which connects the cysts, and Bam 85 and Bgcn function together to regulate the timing of mitotic divisions between cells. In 86 males, Bam is expressed in GSCs, and as differentiation continues, expression 87 increases (McKearin and Spradling 1990, Sgromo et al. 2018, Pan et al. 2014, Ji et al. 88 2017). Once bam expression reaches a threshold in the early spermatogonia, Bam 89 binds to Bgcn and tumorous testis (tut) (Ting et al. 2013, Insco et al. 2009, Insco et al. 90 2012). Binding represses mei-P26 and ends proliferation, triggering terminal 91 differentiation and beginning meiosis (Shen et al. 2009, Ohlstein et al. 2000, Li et al. 92 2013, Li et al. 2009). Loss of bam function prevents differentiation in both sexes and 93 causes over-proliferation of GSCs in females and spermatogonia in males, leading to 94 tumors and sterility in D. melanogaster (Shivdasani et al. 2003, Ohlstein 1997). 95 Though bam plays an essential role in GSC regulation in D. melanogaster, recent 96

show bam’s sequence and function vary considerably across Drosophila or 97 outgroup species (Bubnell et al. 2022). The total 442 amino acid Bam protein in D. 98 melanogaster differs by 60 fixed amino acid differences (~14%) with its sibling species 99 D. simulans. Bam sequences differing from D. melanogaster by up to 308 (67%) of 100 amino acids in other Drosophila species and up to 87% in outgroup species (Arnce, 101 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 4 Bubnell, Aquadro 2025). Statistical tests of selection for D. melanogaster and D. 102 simulans bam suggest that 94% and 72% of fixed amino acid differences respectively 103 were driven by natural selection between the two current species and their common 104 ancestor (Bubnell et al. 2022). Additional signals of positive selection at bam were 105 detected in other Drosophila lineages across the genus leading to D. yakuba, D. 106 ananassae, and D. rubida for example while other lineages, despite evaluation, showed 107 no evidence of positive selection for protein diversification (Bubnell et al. 2022). This 108 data suggests a striking level of sequence divergence, much of which potentially driven 109 by natural selection, in orthologs of a gene critical for ensuring fertility in a D. 110 melanogaster. 111 Additionally, recent functional studies of bam using complete loss-of-function (null) 112 alleles in Drosophila species (Bubnell et al. 2022) revealed divergent roles across D. 113 simulans, D. melanogaster, D. teissieri, D. yakuba, and D. ananassae. In D. teissieri 114 bam null mutants showed no germline stem cell (GSC) differentiation defects in either 115 sex, suggesting bam lacks its canonical role as a critical differentiation regulator in this 116 species. While in D. ananassae bam null females were sterile, males exhibited normal 117 spermatogenesis. These results demonstrate evolutionary divergence in bam sequence 118 and function, despite its conserved essential role in GSC regulation in other Drosophila 119 lineages. 120 Analyses of DNA polymorphism and divergence among other GSC regulatory genes 121 within the Drosophila melanogaster species group have revealed distinct signatures of 122 adaptive evolution across lineages (Bubnell et al. 2022). Notably, some core 123 components of GSC regulatory networks exhibit dynamic evolutionary trajectories, 124 including lineage-specific gene loss. For example, Yb—a gene expressed in somatic 125 cap cells of the germarium and essential for GSC maintenance and transposable 126 element suppression in D. melanogaster (Szakmary et al. 2009)—shows pronounced 127 amino acid divergence between D. melanogaster and D. simulans (Flores et al. 2015). 128 While Yb null mutants in D. melanogaster result in female sterility (Swan et al. 2001), 129 preliminary evidence from our lab suggested that orthologs of Yb are absent in the D. 130 pseudoobscura, D. persimilis, and D. miranda lineages raising the possibility of 131 functional redundancy or network rewiring. These findings underscore both the 132 evolutionary plasticity of GSC regulation and the potential for critical gene turnover 133 within conserved developmental pathways. 134 A persistent challenge in comparative functional genetics lies in the widespread 135 assumption of ortholog functional conservation, despite limited empirical validation. 136 Most studies focus on single-species models, leaving ortholog activity across taxa 137 largely inferred rather than experimentally confirmed (Tekaia 2016). Systematic 138 assessment of the mechanistic basis and degree of functional conservation is crucial for 139 reconstructing the evolution of gene networks. This limited validation of ortholog 140 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 5 conservation introduces significant limitations when interpreting comparative genomic 141 data: non-conserved ortholog functions imply divergent evolutionary pressures across 142 lineages, thereby confounding hypotheses about ancestral genetic architectures or 143 drivers of molecular evolution. Recent studies in bacteria and Diptera have begun to 144 evaluate conservation of functional genes across closely related species and have 145 identified surprising variability in ortholog functional conservation (Carranza et al. 2018, 146 Bergmiller et al. 2012, Leeuwen et al. 2020, Deng et al. 2024, Hopkins et al. 2024, 147 Zakerzade et al. 2025). Lineage-specific ortholog loss of functional reproductive genes 148 has also been observed between humans and nonhuman primates (Carlisle et al. 149 2024). Altogether, these results suggest a more comprehensive comparative functional 150 analysis of bam and other essential GSC regulating genes across the genus Drosophila 151 could provide insight into the functional evolutionary history and extent of network 152 flexibility of bam and essential GSC genes more broadly. 153 We executed our comparative functional analysis of D. melanogaster-essential GSC 154 regulating genes using a two-pronged approach: generating a bam null allele in an 155 additional divergent Drosophila species and evaluating ortholog presence or absence 156 across 15 diverse Drosophila and two outgroup species (sheep blowfly Lucillia cuprina 157 and house fly Musca domestica. 158 We chose to generate a bam null mutant in D. americana as this species represents a 159 major, more divergent outgroup lineage to the D. melanogaster species group within the 160 Drosophila genus and has previously been successfully edited with CRISPR/Cas9 161 (Vankuren et al. 2018, Lamb et al. 2020). We evaluated cytology and fertility in this null 162 mutant using the strategy from Bubnell et al (2022) to evaluate bam’s function in GSC 163 differentiation. This analysis adds broader evolutionary scope to our knowledge of bam 164 function and provides additional insight into whether bam’s essential role in GSC 165 differentiation is likely basal to all Drosophila species and was lost in specific lineages or 166 whether bam’s critical role was a gained function within the D. melanogaster species 167 group. Defining bam null phenotypes also provide a broader context for understanding 168 the relationship between bam function and positive selection. 169 Null mutants are effective for performing comparative analyses of function for genes, 170 like bam, with orthologs across species of interest. However, generating null mutants is 171 expensive and time consuming in non-model species. Identifying ortholog absences in 172 other species for GSC regulating genes essential in D. melanogaster provides an 173 alternative rapid and cost-effective strategy for identifying potential functional 174 differences in the genes and interaction networks that regulate germline stem cell 175 maintenance, self-renewal and differentiation. To investigate ortholog functional 176 conservation via ortholog presence or absence, we started with a set of 366 GSC 177 regulating genes determined to be essential in D. melanogaster from a functional RNAi 178 screen (Yan et al. 2014) along with two of bam’s close interacting partners that are also 179 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 6 essential for fertility, bgcn and Yb (Szakmary et al. 2009, Li et al. 2009). We then used 180 orthology tools in conjunction with a custom reciprocal blast best hit (RBBH) pipeline to 181 identify ortholog presences/absences across 17 diverse Drosophila and two outgroup 182 species. We also used the custom pipeline to identify physical and genetic GSC 183 interacting gene orthologs and validate ortholog absences using a combination of 184 localized PCR and sequencing when possible. Finally, we integrated functional 185 categories as identified in Yan et al. (2014) for the 366 included GSC regulating genes. 186 This generated gene ortholog dataset enables evaluation of the extent and 187 characteristics of essential GSC regulating gene network flexibility as well as informed 188 predictions regarding their functional evolutionary histories. 189 Here, we report that both female and male D. americana bam null mutants are sterile 190 with GSC regulation defects, indicating bam’s essential GSC regulating function is not 191 novel to the melanogaster species group, is likely basal to the genus Drosophila, and 192 provides additional evidence that species in which bam is not necessary for 193 gametogenesis (e.g., D. teissieri and D. ananassae) represent lineage-specific 194 functional losses. Our comparative ortholog analysis of GSC regulating genes reveals 195 that there is additional functional flexibility beyond bam with ~8% (30 of 366) of the 196 genes absent in one or more of the 17 included species and ~3% absent in one or more 197 of the 15 included Drosophila species. Surprisingly, we find that Yb is from D. 198 pseudoobscura and D. obscura species, although it is necessary for GSC function and 199 development, and therefore fertility, in D. melanogaster. Ortholog conservation does not 200 necessarily indicate conservation of function (e.g., bam), so this represents the 201 minimum functional flexibility in essential GSC regulating genes. These results 202 altogether are consistent with other recent studies showing genes essential in one 203 species for critical functions, like fertility, do not necessarily have the same essential 204 function even among closely related species (Leeuwen et al. 2020). Bam functional 205 variation and the absences of several essential GSC regulating gene across Drosophila 206 are potential examples of developmental systems drift (DSD) or divergence in genetic 207 systems that underpin a conserved phenotype (Weiss and Fullerton 2000, True and 208 Haag 2001). 209 210 211

212 Fly stocks and rearing 213 We raised fly stocks on standard cornmeal-molasses food at room temperature, and we 214 used yeast-glucose food for fertility assays. We acquired lines with sequenced genomes 215 for 15 Drosophila and two outgroup species: Drosophila simulans (strain: w501), 216 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 7 Drosophila sechellia (strain: sech25), Drosophila teissieri (strain: GT53w), Drosophila 217 yakuba (strain: Tai18E2), Drosophila takahashii (strain: IR98-3 E-12201), Drosophila 218 elegans (strain: 14027-0461.03), Drosophila serrata (strain: Fors4), Drosophila 219 ananassae (strain: 14024-0371.14) , Drosophila pseudoobscura (strain: MV2-25), 220 Drosophila obscura (strain: BZ-5 IFL), Drosophila willistoni (strain: 14030-0811.24), 221 Drosophila mojavensis (strain: 15081-1352.22), Drosophila virilis (strain: 15010-222 1051.87), Drosophila grimshawi (strain: 15287-2541.00), Drosophila rubida (strain: PH 223 161), Musca domestica (strain: aabys), Lucilia cuprina (strain: Lc7/37) (S1 Table A). We 224 also acquired D. simulans, sechellia, yakuba, serrata, willistoni, mojavensis, virilis,and 225 grimshawi from the National Drosophila Species Stock Center (NDSSC) 226 (http://blogs.cornell.edu/drosophila/), D. takahashii from Kyorin-fly Drosophila species 227 stock center (https://shigen.nig.ac.jp/fly/kyorin/), D. teissieri, D. pseudoobscura, the 228 house fly Musca domestica, and the sheep blowfly Lucillia cuprina as gifts from Daniel 229 Matute, Andy Clark, Jeffrey Scott, and Max Scott, respectively. D. elegans and D. 230 ananassae were gifts from Artyom Kopp. D. obscura and D. rubida were gifts from 231 Dmitri Petrov. We acquired the D. americana white eye mutant line (Lamb et al. 2020) 232 for null generation as a gift from Trisha Wittkopp (S1 Table A). 233 234 Strategy for generating a bam null phenotype in D. americana 235 We generated the bam null disruption by targeting the first exon of bam and introducing 236 an early stop codon in the coding sequence using CRISPR/Cas9 gene editing. Because 237 null homozygotes are sterile and thus cannot be maintained, we developed two bam 238 disruption lines (one marked by 3x3P-Dsred and the other by 3xP3-YFP) which can be 239 maintained as heterozygous lines (S1 Tables B-F). In order to phenotype the null 240 homozygote, we crossed bam3xP3-Dsred/bamwt and bam3xP3-YFP/bamwt flies to create a 241 bam disruption null homozygotes (bam3xP3-Dsred/bamYFP) which we then identified via 242 fluorescent eye screen. 243 244 Bam null construct cloning 245 Yasir Ahmed gifted us bam nucleotide sequences for D. americana (now on NCBI as 246 G96 accession: PRJNA475270), and we performed cloning design in Geneious (S1 247 Table D). We generated PCR products using the NEB Q5 High Fidelity 2x master mix, 248 then gel extracted and purified using the NEB Monarch DNA gel extraction kit. For PCR, 249 sequencing, and cloning, we used IDT primers. We also generated donor plasmids for 250 both the 3xP3-YFP and 3xP3-Dsred bam disruption lines using the strategy outlined in 251 Bubnell et al. (2022). We prepared and purified plasmids for embryo injections with the 252 Qiagen plasmid plus midi-prep kit followed by phenol-chloroform extraction for further 253 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 8 RNase removal and then sequenced plasmids with whole plasmid sequencing 254 (Plasmidsaurus). 255 256 CRISPR/Cas9 and gRNA selection 257 We used Geneious to select gRNAs with no predicted off-targets in the reference 258 genomes for D. americana (S3). We generated these as synthetic gRNAs (sgRNAs) 259 from Synthego and used up to two gRNAs per injection to improve the likelihood of 260 successful CRISPR events (S1 Table F). 261 262 Embryo injections 263 Genetivision performed CRISPR/Cas9 injections including the appropriate plasmid 264 donor, sgRNAs, and Cas9 protein (Synthego) into the D. americana line. We screened 265 bam disruption lines for eye color to identify CRISPR/Cas9 mutant flies in-house using a 266 Nightsea fluorescence system with YFP (cyan) and DsRed (green) filters. We 267 backcrossed positive flies to generate lines which were maintained as heterozygous 268 stocks. We confirmed CRISPR insertions by linear sequencing (Plasmidsaurus). 269 270 Fertility assays 271 We executed following the strategy from Flores et al. (2015) for all female fertility 272 assays. We collected and aged virgin females to sexual maturity (3-4 days for D. 273 americana) (Markow and O’Grady 2006). We collected all generated genotypes from 274 each bottle to control for bottle effects. Wildtype virgin males for each species were also 275 aged until sexual maturity and distributed from different bottles across female 276 genotypes. We crossed single females with two males, allowed to mate for nine days, 277 then flipped onto new vials for nine more days, and then finally cleared from the vials 278 while the offspring develop. We counted progeny daily and cleared to get total adult 279 progeny per female. We also conducted male fertility assays with wildtype females and 280 males of all generated genotypes. We executed fertility assays for D. americana on 281 yeast-glucose food and all fertility experiments were kept at room temperature 282 (approximately 21 degrees C). 283 Fertility assay statistics 284 We used estimation statistics to assess fertility assay mean difference (effect size) in 285 number of adult progeny between the wildtype bam genotype and the bam null 286 homozygote and heterozygote genotypes. We generated estimation statistics and 287 shared control Cumming plots using www.estimationstats.com (Ho 2019) 288 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 9 (github.com/lukearnce/bam_null_ortholog). We used estimation statistics to enable 289 determination of the size of the impact of the bam genotype on fertility via a non-290 parametric methodology. We reported significance as an effect size outside the 95% 291 confidence interval (github.com/lukearnce/bam_null_ortholog). 292 293 Immunostaining 294 We used the following primary antibodies: anti-Hts-1B1 (mouse, 295 AB_528070, Developmental Studies Hybridoma Bank, concentrate 1:40) and anti-vasa 296 (rat, AB_760351, DSHB, concentrate 1:20. We also used the following secondary 297 antibodies: Alexaflour goat anti-rat 488 and goat anti-mouse 568 (Invitrogen) at 1:500. 298 We performed immunostaining as described in Bubnell et al. (2022). In short, we 299 digested ovaries and testes in cold 1x PBS and pipetted up and down to improve 300 antibody permeability, fixed tissues in 4% paraformaldehyde, washed in PBST (1X PBS, 301 0.2% Triton-X 100), blocked in PBTA (1X PBS, 0.2% Triton-X 100, 3% BSA) (Alfa 302 Aesar), and next, incubated in the appropriate primary antibody in PBTA overnight. We 303 washed (PBST), blocked (PBTA), and incubated tissues in the appropriate secondary 304 antibody for two hours. Tissue was then washed again (PBST) and finally mounted in 305 mounting media with DAPI (Prolong glass antifade with NucBlue, Invitrogen) for 306 imaging. 307 308 Microscopy 309 We imaged ovaries and testes on a Zeiss i880 confocal microscope with 405, 488, and 310 568 nm laser lines at 40X (Plan-Apochromat 1.4 NA, oil) (Cornell BRC Imaging Core 311 Facility). We analyzed and edited images using Fiji (ImageJ). 312 313 McDonald-Kreitman test of selective neutrality for the D. americana lineage. 314 We tested D. americana bam for significant departures from neutrality via McDonald-315 Kreitman test (MKT) (McDonald and Kreitman 1991) using polymorphism data gifted by 316 Yasir Ahmed and the bam sequence from D. virilis as an outgroup. We implemented the 317 strategy for MKT analysis of D. americana bam from Bubnell et al. (2022). In brief, we 318 aligned bam sequences using PRANK (version) with the -codon and -F parameters 319 using the PRANK tree guide. We used the codeml package from PAML (version 4.9) 320 (Yang 1997, 2007) to generate the predicted common ancestor sequences for 321 calculating lineage-specific divergence for bam with the MKT. We used PRANK 322 alignments and trees as inputs to codeml with control file parameters (noisy=9, 323 verbose=2, runmode=0, seqtype=1, CodonFreq=2, clock=0, aaDist=0, model=0, 324 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 10 NSsites=0, icode=0, getSE=0, RateAncestor=1, Small_diff=.5e-6, cleandata=0, 325 method=1) (github.com/lukearnce/bam_null_ortholog). We conducted an MKT 326 comparing nonsynonymous and synonymous changes (Egea et al. 2008) using 327 the http://mkt.uab.cat/mkt/mkt.asp webtool. We excluded polymorphic sites at less than 328 12% frequency, classified as slightly deleterious alleles not yet removed by purifying 329 selection (Charlesworth and Eyre-Walker 2008). We used the predicted common 330 ancestral species sequence to calculate lineage-specific divergence. We recorded 331 values for the contingency table, the P-value of the χ2, alpha, and proportion of fixations 332 predicted to be due to positive selection (Eyre-Walker 2006) 333 (github.com/lukearnce/bam_null_ortholog). 334 335 Genomes used in analysis 336 We downloaded genomes from NCBI for 15 Drosophila and two outgroup species: 337 Drosophila simulans (assembly: Prin_Dsim_3.1), Drosophila sechellia (assembly: 338 ASM438219v2), Drosophila teissieri (assembly: Prin_Dtei_1.1), Drosophila yakuba 339 (assembly: Prin_Dyak_Tai18E2_2.1), Drosophila takahashii (assembly: ASM1815269), 340 Drosophila elegans (assembly: ASM1815250), Drosophila serrata (assembly: Dser1.1), 341 Drosophila ananassae (assembly: ASM1763931v2) , Drosophila pseudoobscura 342 (assembly: UCI_Dpse_MV25), Drosophila obscura (assembly: ASM1815110v1), 343 Drosophila willistoni (assembly: UCI_dwil_1.1), Drosophila mojavensis (assembly: 344 ASM1815372v1), Drosophila virilis (assembly: Dvir_AGI_RSII-ME), Drosophila 345 grimshawi (assembly: ASM1815329v1), Drosophila rubida (assembly: ASM3504616v1), 346 Musca domestica (assembly: Musca_domestica.polishedcontigs.V.1.1), Lucilia cuprina 347 (assembly: ASM2204524v1) (S1 Table A). 348 349 Ensembl ortholog analysis 350 We used the Ensembl Compara online ortholog tool (Kersey et al. 2010 and Accessed 351 date: June 2022) to collect ortholog predictions for 366 GSC regulating genes for the 10 352 Drosophila species that were available in Ensemble Compara (D. melanogaster, D. 353 simulans, D. sechellia, D. yakuba, D. ananassae, D. pseudoobscura, D. willistoni, D. 354 mojavensis, D. virilis, D. grimshawi) and two outgroup species (L. cuprina and M. 355 domestica) (S2 Tables A-B). This tool catalogs relevant information about each ortholog 356 including sequence alignment, target percent ID (percentage of orthologous sequence 357 matching the Drosophila melanogaster sequence), query percent ID (percentage of 358 Drosophila melanogaster sequence matching the orthologous sequence), gene order 359 conservation score (evaluating synteny), and high or low ortholog confidence 360 (calculated using results from other categories) (Kersey et al. 2010). Predicted orthologs 361 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 11 were included as confident predictions if sequence alignment is equal to or greater than 362 25% identity (github.com/lukearnce/bam_null_ortholog). 363 364 Reciprocal Best Blast Hit (RBBH) ortholog pipeline 365 To execute our RBBH ortholog analysis, implemented a multi-stage pipeline: 366 1. Initial RBBH 367 We first downloaded highly contiguous, long-read full genome sequences from NCBI for 368 all included species (15 Drosophila species plus the two outgroup species). D. teissieri, 369 D. takahashii, D. elegans, D. serrata, D. obscura, D. rubida, M. domestica , and L. 370 cuprina were added to the initial set of species from Ensembl Compara. Then, we used 371 custom scripts to perform forward and reverse BLASTp searches to identify potential 372 orthologs and filter ortholog hits for GSC regulating genes as well as GSC gene 373 interaction network genes (S2, github.com/lukearnce/bam_null_ortholog). 374 2. RBBH + syntenic evaluation 375 For genes with predicted absences, we conducted forward and reverse BLASTn 376 searches as well as reciprocal BLASTp searches for the gene predicted absent and the 377 syntenic genes (three on each side) flanking the GSC gene in D. melanogaster. We 378 executed this to search for genes that are actually present but may have been predicted 379 absent due to location at the end of contigs (S2, S3, 380 github.com/lukearnce/bam_null_ortholog). 381 3. Direct validation of predicted gene ortholog absences 382 We evaluated predicted GSC gene ortholog absences with retained syntenic blocks 383 directly via sequencing. We developed primers (IDT) for PCR amplification the 384 sequence between retained syntenic genes. We then gel extracted, purified, and 385 sequenced PCR products were to directly verify gene absence (S2, S3, 386 github.com/lukearnce/bam_null_ortholog). 387 388 Interaction networks and functional information 389 We cataloged physical interactors and genetic interactors from Flybase datasets for 390 GSC regulating genes with predicted absences and evaluated interaction network 391 genes for orthologs across included species using the same RBBH pipeline (Choi JY 392 and Aquadro CF 2015) as well as ortholog predictions from Li et al. (2021). Annotated 393 molecular functions (Choi and Aquadro 2015), functional categories identified by 394 complex-enrichment analysis of the 366 GSC genes, and defect type, defined by the 395 observed phenotypic effect of RNAi knockdown, were also incorporated from Yan et. al 396 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 12 (2014). Defect types include GSC loss (cell viability), GSC loss (agametic), 397 differentiation defect, and oocyte-specific phenotypes/late oogenesis (S3). 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414

415 Bam shows a lineage-specific signature of positive selection in D. americana 416 We detected a significant departure from neutrality suggesting positive selection 417 favoring accelerated amino acid substitutions via the McDonald-Kreitman (1981) test for 418 bam in the D. americana lineage. Contingency table values (Neutral Polymorphism: 42, 419 Neutral Divergence: 0, Non-neutral Polymorphism: 27, and Non-neutral Divergence: 3) 420 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 13 show the ratio of nonsynonymous to synonymous variation between species is greater 421 than the same ratio within species, consistent with expectations for positive selection (χ2 422 = 4.390 P-value = 0.036) and 100% of fixations estimated to be due to positive selection 423 (alpha = 1.00) (github.com/lukearnce/bam_null_ortholog). This raised the possibility that 424 there may have been a functional change at D. americana bam. 425 426 Bam is necessary for female and male fertility and germ cell differentiation in D. 427 americana 428 We used CRISPR/Cas9 gene editing to generate bam null alleles in D. americana, a 429 representative of a lineage divergent from D. melanogaster in the genus Drosophila. 430 Wildtype (+/+) and bam null heterozygotes, (bam3xP3-Dsred/bam+) and (bam3xP3-YFP/bam+), 431 are all fertile in both males and females (Fig. 1). However, bam null homozygotes, 432 (bam3xP3-Dsred/bam3xP3-YFP), were completely sterile in males and females (P < 0.0001, 433 permutation test, Fig. 1a&d) (github.com/lukearnce/bam_null_ortholog). As in D. 434 melanogaster, D. simulans, D. yakuba, and D. ananassae (Bubnell, et al. 2022), one 435 copy of wildtype bam is sufficient to rescue the bam null sterility phenotype in D. 436 americana. 437 438 439 440 441 442 443 444 445 446 447 Figure 1. 448 Fertility and cytological analyses of bam function in adult D. americana. Fertility is 449 presented as adult D. americana progeny per fly for each bam genotype and 450 presented separately for females (a) and males (d). The raw data as progeny per 451 fly is plotted on the upper axes with the mean difference for the three genotype 452 comparisons against the shared control wildtype illustrated in the Cumming 453 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 14 estimation plots on the lower axes. Mean differences are plotted as bootstrap 454 resampling distributions. Each mean difference is depicted as a dot, and 95% 455 confidence intervals are indicated by the vertical black bars. Immunostaining of 456 ovaries (b and c) and testes (e and f) of wildtype (b and e) and null bam 457 genotypes (c and f). Composite Z-projections for ovaries and testes show 458 staining for the germline (vasa), fusome (1B1), and nuclei (DAPI) with separate 459 single channels for each image illustrated in the side panels (i. vasa, ii. DAPI, iii. 460 1B1). Wildtype tissue phenotypes are indicated with arrows and mutant tissue 461 phenotypes are indicated with arrowheads. 462 463 To confirm the bam null sterile fertility phenotype was due to defects in GSC function as 464 expected if bam function is conserved between D. melanogaster and D. americana, we 465 evaluated the cytology of D. americana bam null ovaries and testes (Fig. 1). We imaged 466 3-5 day-old ovaries from D. americana bam wildtype and bam null females that we 467 immunostained with antibodies to vasa and 1B1 and mounted with DAPI. (Lavoie et al. 468 1999). Homozygous bam null cytology recapitulated the classic bag of marbles 469 phenotype, with over-proliferation of small, undifferentiated GSC-like cells in the ovaries 470 (Fig. 1c) and testes (Fig. 1f) in contrast to bam wildtype ovaries (Fig 1b) and testes (Fig 471 e) which consist of cysts made of larger differentiating germline cells. Our cytological 472 data reveal that bam is necessary for early germ cell differentiation in D. americana, 473 consistent with fertility assay results (Fig. 1a & d). 474 475 Essential GSC gene ortholog absences across species 476 While functional genetic analyses for bam across diverse lineages in the genus 477 Drosophila revealed some striking variation in bam’s role in fertility and germ cell 478 differentiation, the financial and time costs required to generate null mutants for other 479 GSC genes across diverse species are prohibitive at this point. Therefore, we next 480 chose an alternative, albeit less sensitive, approach to evaluate the functional 481 consistency of the roles of GSC regulating genes that are essential in D. melanogaster. 482 Focusing on the experimentally defined set of GSC regulating genes determined to be 483 essential in D. melanogaster (Yan et al. 2014), we defined a functional difference in 484 GSC regulation pathways between species as the absence of an essential ortholog in 485 D. melanogaster in any other Drosophila and/or outgroup species. 486 Our initial Ensembl ortholog analysis predicted absences for 311 of 366 GSC regulating 487 genes for nine Drosophila and two outgroup species (S2 Tables A-B, 488 github.com/lukearnce/bam_null_ortholog). Number of absences per species ranged 489 from 27 (7.38%) in D. simulans to 222 (60.66%) in the outgroup species Lucillia cuprina 490 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 15 (sheep blowfly). Next, we used our Reciprocal Best Blast Hit (RBBH) ortholog 491 assessment that revealed predicted absences for 79 of 366 GSC regulating genes 492 across 15 Drosophila and two outgroup species. Number of absences per species 493 ranged from two (0.5%) in D. simulans to 39 (10.7%) in D. rubida. Finally, with our most 494 stringent approach combining RBBH and syntenic evaluation of orthologs, we predicted 495 absences for 30 of 366 GSC regulating genes across 15 Drosophila and two outgroup 496 species (Fig. 3) (S3, github.com/lukearnce/bam_null_ortholog). Our most stringently 497 assessed number of absences per species ranged from one (0.27%) in D. simulans to 498 13 (3.6%) in the outgroup species Musca domestica (house fly). 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 16 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 Figure 2. 536 Predicted absences of 366 GSC regulating genes essential in D. melanogaster 537 (predicted by Yan et al. 2014) by species plus bgcn and Yb using different 538 ortholog detection strategies. 539 a. Percentage of GSC genes predicted absent by species for three ortholog 540 identification strategies. Light grey bars represent ortholog predictions using 541 Ensembl, dark grey represent predictions with RBBH alone, and black represent 542 RBBH plus syntenic evaluation. The bar below the species indicates divergence 543 time from D. melanogaster (MY). 544 b. Presence and absence of GSC genes across 15 Drosophila and two outgroup 545 species that are predicted absent in at least one species after RBBH and 546 syntenic evaluation. Blue indicates gene presence and black indicates gene 547 absence. 548 549 Verification of gene absences with retained syntenic blocks 550 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 17 To experimentally confirm the predicted GSC regulating gene ortholog absences we 551 found with RBBH + syntenic evaluation with retained syntenic blocks, we PCR amplified 552 and sequenced the syntenic block spanning the expected gene absence. The 553 sequencing results verified all predicted absences with retained synteny therefore our 554 RBBH + syntenic computational predictions represented real absences in the amplified 555 regions (github.com/lukearnce/bam_null_ortholog). Using our final ortholog presence 556 and absence results, we then sought to identify potential patterns in gene presence or 557 absence across the included species. Primarily, we evaluated whether and in what ways 558 essential gene conservation varied across gene functional categories and interaction 559 network size (S3). 560 561 Variable GSC gene conservation across species and functional categories 562 We found that GSC regulating genes with predicted absences are unequally 563 represented across functional categories identified by complex enrichment. At the 564 extremes, there are zero predicted absences in the proteasome functional category (15 565 genes) and two of six genes (33%) in the Kinetochore and spindle functional category 566 (Fig. 3). Absences categorized by defect types show less variation ranging from 7.14% 567 for GSC loss (cell viability) (12 of 168) to 11.11% for genes with oocyte-specific 568 phenotypes/late oogenesis (five of 45). 569 570 571 572 573 574 575 576 577 578 579 580 581 582 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 18 583 584 585 586 587 588 589 Figure 3. 590 Percentage of GSC genes absent by functional category. Categories are pulled 591 from the gene-interaction network generated in Yan et al. (2014). The “No 592 subcategory (105)” group includes genes represented in the interaction network 593 map without clear categorical associations and the “No functional category (57)” 594 includes GSC genes that do not appear in the interaction network map. 595 596 Variability in absent GSC gene interaction network size and absences 597 Of the 30 GSC regulating genes (from the 366 in Yan et al. (2014)) with predicted 598 absences across included species, 23 genes have physical and/or genetic interactions 599 (Fig. 4, S3). 13 of these 23 genes also have predicted absences in their interaction 600 networks, and most of these interaction network absences are in the same species as 601 their related absent GSC regulating gene (Fig. 5, S3). 602 603 604 605 606 607 608 609 610 611 612 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 19 613 614 615 616 Figure 4. 617 Interaction network size and absences in absent GSC genes. Genes with at least 618 one absence in the included species are listed with white bars representing the 619 size of their interaction networks including genetic and physical interactors. The 620 number of interaction network genes absent are represented with black bars. 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 20 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 Figure 5. 656 GSC gene and associated network absences across species. GSC genes with 657 absences in at least one included species are highlighted in dark grey. 658 Associated interaction network (I.N.) genes are indented, italicized, and 659 highlighted in light grey. Gene presence is indicated by light blue and absence is 660 indicated by black. First, GSC genes with no absences in their interaction 661 networks are arranged by increasing interaction network size. Next, GSC genes 662 with absences in their interaction networks are arranged in the same manner. 663 GSC genes with no interaction networks (six genes) are excluded from this 664 figure. 665 666 667 668 669 670 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 21 671 672 673 674 675 676 677 678 679 680 681 682 683

684 Fertility and germline cytological assays of bam null mutants in D. americana reveal that 685 bam is essential for GSC differentiation in both males and females. This demonstrates 686 that bam’s essential function in GSC differentiation is not novel to the D. melanogaster 687 species group, and its evolutionary origin in this function likely occurred just prior to or 688 just after the origin of the genus Drosophila. Bam functional differences in D. teissieri 689 and D. ananassae likely represent lineage-specific GSC functional losses. Similar 690 losses may exist among the many other non-tested taxa in this species-rich genus. 691 These analyses of bam function show that high amino acid sequence variability 692 between species does not necessarily imply functional divergence. D. americana bam 693 and D. melanogaster bam share only 35% sequence identity, yet both execute the same 694 essential function in GSC daughter differentiation. In contrast, D. teissieri bam and D. 695 melanogaster bam share 75% sequence identity and they are functionally distinct. 696 Conservation of sequence does not necessarily imply conservation of function and 697 divergence of sequence does not necessarily imply a divergence in function. A classic 698 example of this dynamic includes cytochrome c (Garrido et al. 2006). Cytochrome c is 699 one of the most conserved proteins among eukaryotes with an amino acid sequence 700 conservation of 70-90% between species as divergent as yeast and mammals. In many 701 yeasts and lower eukaryotes, cytochrome c functions solely as an electron carrier in the 702 mitochondrial respiratory chain while in mammals the same (or nearly identical) 703 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 22 cytochrome c also acquired an additional role as an apoptotic signal by helping to 704 trigger caspase activation and programmed cell death. This extra “moonlighting” 705 function in apoptosis is a dramatic divergence in biological role despite a highly 706 conserved structure (Garrido et al. 2006). 707 The resolution and depth of our analysis was made possible due to the large collection 708 and phylogenetic density of high-quality genomes for a growing number of species in 709 the genus Drosophila in combination with the depth of functionally defined genes within 710 D. melanogaster. Large-scale gene ortholog identification is mostly not possible in 711 species with only low-quality genomes available, and, potentially as a result, there are 712 currently very few large-scale comparative analyses of gene orthologs, essential or 713 nonessential (Carranza et al. 2018, Bergmiller et al. 2012, Leeuwen et al. 2020, Deng et 714 al. 2024). However, as more high-quality genomes are produced and published for more 715 species, this type of comparative ortholog analysis becomes more feasible. Even when 716 using only species with available high-quality genomes for comparative analysis, efforts 717 must be made to ensure potential orthologs and absences are properly identified. For all 718 of the GSC regulating genes in D. melanogaster that we analyzed (366 from Yan et al. 719 (2014) plus bgcn and Yb), ortholog predictions vary dramatically based on identification 720 strategy. Using Ensembl alone, we predicted 85% (311 of 366) of GSC regulating genes 721 absent in at least one included species while we only predicted ~8% (30 of 366) absent 722 based on the more detailed evaluation incorporating RBBH and synteny. Due to the 723 significant discrepancy in predictions, the predicted presence or absence of orthologs 724 via Ensembl must be further interrogated by additional strategies including RBBH and 725 syntenic evaluation before confidently characterizing predictions as orthologs that are 726 truly absent. As an example, benign gonial cell neoplasm (bgcn) is crucial gene that is 727 broadly conserved, including a human ortholog, that was predicted to be absent in the 728 initial RBBH evaluation. However, using follow-up BLASTn and syntenic evaluation we 729 found the bgcn ortholog split over short contigs that made initial identification difficult. In 730 contrast, Yb, an additional essential GSC regulating gene in D. melanogaster was 731 computationally predicted absent in D. obscura and D. pseudoobscura. Follow-up 732 syntenic evaluation and sequencing confirmed these absences of Yb from their syntenic 733 regions in these two species. 734 The 30 GSC (out of 366) GSC regulating genes that are confidently predicted to be 735 absent represent significant flexibility in essential GSC regulating genes beyond bam 736 across species. These results add to a growing body of evidence that suggests the 737 general ortholog assumption of shared function is not universal and provide examples of 738 developmental systems drift (DSD) in a process critical for early reproduction. These 30 739 genes are also not equally distributed across functional categories. While there are no 740 predicted absences in proteasome genes, two of six (33%) genes involved in the 741 kinetochore and spindle have predicted absences. This variable flexibility across 742 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 23 functional categories suggests there may be some networks that are more flexible in 743 their regulation while others are particularly intolerant. Additionally, further analysis of 744 these 30 genes with absent orthologs shows that these genes generally have 745 interaction networks including genetic and physical interactors. Absences in orthologs of 746 interaction network genes also generally occur in the same species that their related 747 GSC gene is absent. This could potentially indicate that having a network of interacting 748 partners more easily enables gene loss by compensating for the function of the absent 749 gene through redundancy or alternative regulation while essential genes without large 750 interaction networks could generally be more difficult to replace (Albalat 2016, Wagner 751 2008, Carroll 2008). One gene, Prosalpha1, is absent in the most species of the 752 included GSC genes. In addition to having a large interaction network (47 genes), it also 753 has a very similar gene, Prosalpha1R, within the surrounding syntenic region. 754 Prosalpha1R could possibly facilitate alternative execution of Prosalpha1’s essential 755 function. 756 Our findings here provide additional evidence that orthologous genes do not necessarily 757 function identically even when critical for regulating essential systems. Despite the 758 essential nature of proper gamete development, what has been described as 759 developmental systems drift (DSD) is occurring in genes critical for regulating the 760 earliest stages of this process across Drosophila and closely related outgroups. 761 Spermatogenesis regulating genes further downstream of GSC regulation have also 762 been evaluated across some Drosophila species. Results showcase similar functional 763 flexibility with duplications or losses of sperm nuclear basic proteins as well as the 764 emergence of many functional de novo genes (Lee U 2025, Chang C 2023). Some of 765 the particular genes involved in the process change, but the final phenotype is 766 maintained across species. 767 We conclude that ortholog function must be evaluated in a species-specific manner. 768 These 30 GSC genes and Yb that show absences in some species represent the 769 baseline of functional flexibility for the D. melanogaster essential GSC regulating genes, 770 but many other GSC regulating genes still present, like bam, could have variable 771 function across species. The extent of developmental systems drift occurring beyond 772 GSC regulating genes in these 17 species is largely unknown, but further evaluation of 773 ortholog functional flexibility focusing on different organisms and phenotypes could 774 reveal differences in the key genes regulating any number of traits across species. 775

from this dataset suggest some potential patterns related to ortholog loss of 776 essential genes, but more extensive analysis needs to be executed to make more 777 concrete assessments of the prevalence and characteristics of ortholog functional 778 flexibility. 779 780 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 24 781 Data Availability 782 All raw and filtered data are available in a public repository at 783 www.github.com/lukearnce/bam_null_ortholog. These data include the following: 784 McDonald-Kreitman test for D. americana bam 785 PRANK inputs and control parameters for the D. americana bam MK test with 786 contingency table values, the P-value of the χ2, alpha, and proportion of fixations 787 predicted to be due to positive selection. 788 D. americana bam CRISPR Null fertility assays 789 D. americana male and female estimation statistics and significance as an effect size 790 outside of the 95% confidence intervals 791 GSC ortholog identification 792 Initial RBBH ortholog analysis results, Full genomes, custom scripts to execute forward 793 and reverse BLASTp, filtered ortholog hits, interaction network genes, RBBH with 794 syntenic evaluation gene hits, and sequence verification of predicted ortholog absences 795

796 We thank Jolie Carlisle for her thoughtful comments on an earlier version of the 797 manuscript. This work was supported by National Institute of Health (United States) 798 R01-GM095793 to Charles F. Aquadro. 799 800 801 802 803 804 805 806 807 808 809 810 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 25 811 812 813 814 815 816 817 818 819 820 821

822 Albalat R, Canestro C 2016 Evolution by gene loss. Nature Reviews Genetics 17, 379-823 391. https://doi.org/10.1038/nrg.2016.39 824 Arnce LR, Bubnell JE, Aquadro CF 2025 Comparative Analysis of Drosophila Bam and 825 Bgcn Sequences and Predicted Protein Structural evolution. Journal of Molecular 826 Evolution https://doi.org/10.1101/2024.12.17.628990 827 Bauer DuMont et al. 2007 Recurrent Positive Selection at Bgcn, a Key Determinant of 828 Germ Line Differentiation, Does Not Appear to be Driven by Simple Coevolution with Its 829 Partner Protein Bam. Molecular Biology and Evolution, 24(1):1882-191. 830 https://doi.org/10.1093/molbev/msl141 831 Bergmiller T et al. 2012 Patterns of Evolutionary Conservation of Essential Genes 832 Correlate with Their Compensability. PLOS Genetics, 833 https://doi.org/10.1371/journal.pgen.1002803 834 Carlisle J et al. 2024 Recurrent Independent Pseudogenization Events of the Sperm 835 Fertilization Gene ZP3r in Apes and Monkeys. Journal of Molecular Evolution, 92:695-836 702. https://doi.org/10.1007/s00239-024-10192-x 837 Carranza S et al. 2018 Diversity, distribution and conservation of the terrestrial reptiles 838 of Oman (Sauropsida, Squamata). PLOS One, 839 https://doi.org/10.1371/journal.pone.0190389 840 Carroll, S. 2008 Evo-devo and an expanding evolutionary synthesis: a genetic theory of 841 morphological evolution. Cell 134(1)25-36. https://doi.org/10.1016/j.cell.2008.06.030 842 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 26 Chang C et al. 2023 Expansion and loss of sperm nuclear basic protein genes in 843 Drosophila correspond with genetic conflicts between sex chromosomes. eLife, 844 https://doi.org/10.75554/eLife.85249 845 Charlesworth J and Eyre-walker A 2008 The McDonald-Kreitman test and slightly 846 deleterious mutations. Molecular Biology and Evolution, 25(6):1007-15. 847 https://doi.org/10.1093/molbev/msn005 848 Choi JY and Aquadro CF 2015 Molecular Evolution of Drosophila Germline Stem Cell 849 and Neural Stem Cell Regulating Genes. Genome Biology and Evolution, 7(11):3097-850 114. https://doi.org/10.1093/gbe/evv207 851 Civetta A et al. 2006 Rapid evolution and gene-specific patterns of selection for three 852 genes of spermatogenesis in Drosophila. Molecular Biology and Evolution, 23(3):655-853 62. https://doi.org/10.1093/molbev/msj074 854 Deng D et al. 2024 Functional Divergence in Orthologous Transcription Factors: Insights 855 from AtCBF2/3/1/ and OsDREB1C Molecular Biology and Evolution, 41(5):msae089. 856 https://doi.org/10.1093/molbev/msae089 857 DuMont, V.L., White, S.L., Zinshteyn, D. and Aquadro, C.F. 2021. Molecular population 858 genetics of Sex-lethal (Sxl) in the D. melanogaster species group - a locus that 859 genetically interacts with Wolbachia pipientis in Drosophila melanogaster. G3 860 Genes|Genomes|Genetics 11(8), jkab197. https://doi.org/10.1093/g3journal/jkab197 861 Egea R et al. 2008 Standard and generalized McDonald-Kreitman test: a website to 862 detect selection by comparing different classes of DNA sites. Nucleic Acids Research, 863 36: 157-162. https://doi.org/10.1093/nar/gkn337 864 Eyre-Walker A et al. 2006 The distribution of fitness effects of new deleterious amino 865 acid mutations in humans. Genetics, 173(2):891-900. 866 https://doi.org/10.1534/genetics.106.057570 867 Flores HA, Bubnell JE, Aquadro CF, Barbash DA. 2015 The Drosophila bag of marbles 868 Gene Interacts Genetically with Wolbachia and Shows Female-Specific Effects of 869 Divergence. PLOS Genetics 185(4):613-27. 870 https://doi.org/10.1371/journal.pgen.1005453 871 Garrido C et al. 2006 Mechanisms of cytochrome c release from mitochondria. Cell 872 Death & Differentiation 13, 1423-1433. https://doi.org/10.1038/sj.cdd.4401950 873 Gleason R et al. 2018 Protecting and Diversifying the Germline. Genetics, 208(2):435-874 471. https://doi.org/10.1534/genetics.117.300208 875 Ho J et al. 2019 Moving beyond P values: data analysis with estimation graphics. 876 Nature Methods, 16(7):565-566. https://doi.org/10.1038/s41592-019-0470-3 877 Hopkins B, et al. Decoupled evolution of the Sex Peptide gene family and Sex Peptide 878 Receptor in Drosophilidae. PNAS 121(3):e2312380120. 879 https://doi.org/10.1073/pnas.2312380120 880 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 27 Insco ML, Leon A, Tam CH, McKearin DM, Fuller MT. 2009 Accumulation of a 881 differentiation regulator specifies transit amplifying division number in an adult 882 stem cell lineage. Proc Natl Acad Sci U S A. 106(52) 22311-22316. 883 https://doi.org/10.1073/pnas.0912454106 884 Insco ML, Bailey AS, Kim J, Olivares GH, Wapinski OL, Tam CH, et al. 2012 A self-885 limiting switch based on translational control regulates the transition from 886 proliferation to differentiation in an adult stem cell lineage. Cell Stem Cell. 887 11: 689–700. https://doi.org/10.1016/j.stem.2012.08.012 888 Ji S, Li C, Hu L, Liu K, Mei J, Luo Y, et al. 2017 Bam-dependent deubiquitinase complex 889 can disrupt germ-line stem cell maintenance by targeting cyclin A. Proc 890 Natl Acad Sci U S A. 114(24):6316-6321. https://doi.org/10.1073/pnas.1619188114 891 Kahney E et al. 2019 Regulation of Drosophila germline stem cells. Current Opinion in 892 Cell Biology, Vol. 60, 27-35. https://doi.org/10.1016/j.ceb.2019.03.008 893 Kersey PJ et al. 2010 Ensembl Genomes: Extending Ensembl across the taxonomic 894 space Nucleic Acids Research Vol. 38 D563-D569. https://doi.org/10.1093/nar/gkp871 895 Lamb AM, Wang Z, Simmer P, Chung H, Wittkopp P 2020 ebony Affects Pigmentation 896 Divergence and Cuticular Hydrocarbons in Drosophila americana and D. novamexicana 897 Front. Ecol. Evol. https://doi.org/10.3389/fevo.2020.00184 898 Lee U et al. 2025 Comparative single cell analysis of transcriptional bursting reveals the 899 role of genome organization on de novo transcript origination. bioRxiv preprint. 900 Leeuwen J et al. 2020 Systematic analysis of bypass suppression of essential genes. 901 Mol Syst Biol, 16:1-24. https://doi.org/10.15252/msb.20209828 902 Li F et al. 2021 Phylogenomic analyses of the genus Drosophila reveals genomic 903 signals of climate adaptation. Molecular Ecology Resources. 904 https://doi.org/10.1111/1755-0998.13561 905 Li Y, Minor NT, Park JK, McKearin DM, Maines JZ. 2009 Bam and Bgcn antagonize 906 Nanos-dependent germ-line stem cell maintenance. Proc Natl Acad Sci U S A. 907 106: 9304–9309. https://doi.org/10.1073/pnas.0901452106 908 Li Y, Zhang Q, Carreira-Rosario A, Maines JZ, Mckearin DM. 2013 Mei-P26 909 Cooperates with Bam, Bgcn and Sxl to Promote Early Germline Development in 910 the Drosophila Ovary. PLoS One. 8: 58301. 911 https://doi.org/10.1371/journal.pone.0058301 912 Markow T and O’Grady P 2006 Drosophila: A Guide to Species Identification and Use. 913 Book 914 McDonald J and Kreitman M 1991 Adaptive protein evolution at the Adh locus in 915 Drosophila Nature 351(6328):652-4. https://doi.org/10.1038/351652a0 916 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 28 McKearin DM, Spradling AC. 1990 Bag-of-marbles: A Drosophila gene required to 917 initiate both male and female gametogenesis. Genes Dev. 4: 2242–2251. 918 https://doi.org/10.1101/gad.4.12b.2242 919 Ohlstein B, Lavoie CA, Vef O, Gateff E, Mckearin DM. 2000 The Drosophila Cystoblast 920 Differentiation Factor, benign gonial cell neoplasm, Is Related to DExH-box Proteins 921 and Interacts Genetically With bag-of-marbles. 155(4):1809-19. 922 https://doi.org/10.1093/genetics/155.4.1809 923 Ohlstein B, McKearin D. 1997 Ectopic expression of the Drosophila Bam protein 924 eliminates oogenic germline stem cells. Development. 124(18):3651-62. 925 https://doi.org/10.1242/dev.124.18.3651 926 Pan L, Wang S, Lu T, Weng C, Song X, Park JK, et al. 2014 Protein competition 927 switches the function of COP9 from self-renewal to differentiation. Nature. 928 514: 233–236. https://doi.org/10.1038/nature13562 929 Sgromo A, Raisch T, Backhaus C, Keskeny C, Alva V, Weichenrieder O, et al. 2018 930 Drosophila Bag-of-marbles directly interacts with the CAF40 subunit of the 931 CCR4–NOT complex to elicit repression of mRNA targets. Rna. 24: 381– 932 395. https://doi.org/10.1261/rna.064584.117 933 Shen R, Weng C, Yu J, Xie T. 2009 eIF4A controls germline stem cell self-renewal by 934 directly inhibiting BAM function in the Drosophila ovary. Proc Natl Acad Sci U S A. 106: 935 11623–11628. https://doi.org/10.1073/pnas.0903325106 936 Shivdasani AA, Ingham PW. 2003 Regulation of Stem Cell Maintenance and Transit 937 Amplifying Cell Proliferation by TGF-β Signaling in Drosophila Spermatogenesis. 938 Curr Biol. 13(23):2065-72. https://doi.org/10.1016/j.cub.2003.10.063 939 Swan A, Hijal S, Hilfiker A, Suter B 2001 Identification of new X-chromosomal genes 940 required for Drosophila oogenesis and novel roles for fs(1)Yb, brainiac and dunce 941 Genome Res. 11(1):67-77. https://doi.org/10.1101/gr.156001 942 Szakmary A, Reedy M, Qi H, Lin H 2009 The Yb protein defines a novel organelle and 943 regulates male germline stem cell self-renewal in Drosophila melanogaster J Cell Biol. 944 185(4): 613-627. https://doi.org/10.1083/jcb.200903034 945 Tekaia F 2016 Inferring Orthologs: Open Questions and Perspectives. Genomics 946 Insights 9:17-28. https://doi.org/10.4137/GEI.S37925 947 Ting X 2013 Control of germline stem cell self-renewal and differentiation in the 948 Drosophila ovary: concerted actions of niche signals and intrinsic factors. Wiley 949 Interdiscip Rev Dev Biol. 2: 261–273. https://doi.org/10.1002/wdev.60 950 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint 29 Tokusumi T. et al. 2011 Germ line differentiation factor Bag of Marbles is a regulator of 951 hematopoietic progenitor maintenance during Drosophila hematopoiesis. Development, 952 138(18):3879-84. https://doi.org/10.1242/dev.069336 953 True JR, Haag ES. 2001. Developmental system drift and flexibility in evolutionary 954 trajectories. Evol Dev. 3(2):109-119. https://doi.org/10.1046/j.1525-955 142x.2001.003002109.x 956 Vankuren NW, Long M. 2018 Gene duplicates resolving sexual conflict rapidly evolved 957 essential gametogenesis functions. Nature Ecology and Evolution 2, 705-712. 958 https://doi.org/10.1038/s41559-018-0471-0 959 Wagner A. 2008 Robustness and evolvability: a paradox resolved. Proc Biol Sci 960 275(1630):91-100. https://doi.org/10.1098/rspb.2007.1137 961 Weiss KM, Fullerton SM. 2000. Phenogenetic drift and the evolution of genotype- 962 phenotype relationships. Theor Popul Biol. 57(3):187-195. 963 https://doi.org/10.1006/tpbi.2000.1460 964 Yan D et al. 2014 A Regulatory Network of Drosophila Germline Stem Cell Self-965 Renewal Developmental Cell Vol 28 Issue 4 p. 459-473. 966 https://doi.org/10.1016/j.devcel.2014.01.020 967 Zakerzade R, et al. 2025 Diversification and recurrent adaptation of the synaptonemal 968 complex in Drosophila. PLoS Genetics 21(1):e1011549. 969 https://doi.org/10.1371/journal.pgen.1011549 970 971 972 973 974 975 976 977 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 6, 2025. ; https://doi.org/10.1101/2025.04.30.651540doi: bioRxiv preprint