Transgenic Testing Does Not Support a Role for Additional Candidate Genes in Wolbachia Male Killing or Cytoplasmic Incompatibility

Abstract

Endosymbiotic bacteria in the genus Wolbachia remarkably infect nearly half of all arthropod species. They spread in part because of manipulations of host sexual reproduction that enhance the maternal transmission of the bacteria, including male killing (death of infected males) and unidirectional cytoplasmic incompatibility (CI; death of offspring from infected fathers and uninfected mothers). Recent discoveries identified several genes in prophage WO of Wolbachia (wmk, cifA, and cifB) that fully or partially recapitulate male killing or CI when transgenically expressed in Drosophila melanogaster However, it is not yet fully resolved if other gene candidates contribute to these phenotypes. Here, we transgenically tested 10 additional gene candidates for their involvement in male killing and/or CI. The results show that despite sequence and protein architecture similarities or comparative associations with reproductive parasitism, transgenic expression of the candidates does not recapitulate male killing or CI. Sequence analysis across Wmk and its closest relatives reveals amino acids that may be important to its function. In addition, evidence is presented to propose new hypotheses regarding the relationship between wmk transcript length and its ability to kill a given host, as well as copy number of wmk homologs within a bacterial strain, which may be predictive of host resistance. Together, these analyses continue to build the evidence for identification of wmk, cifA, and cifB as the major genes that have thus far been shown to cause reproductive parasitism in Wolbachia, and the transgenic resources provide a basis for further functional study of phage WO genes.IMPORTANCE Wolbachia are widespread bacterial endosymbionts that manipulate the reproduction of diverse arthropods to spread through a population and can substantially shape host evolution. Recently, reports identified three prophage WO genes (wmk, cifA, and cifB) that transgenically recapitulate many aspects of reproductive manipulation in Drosophila melanogaster Here, we transgenically tested 10 additional gene candidates for CI and/or male killing in flies. The results yield no evidence for the involvement of these gene candidates in reproductive parasitism, bolstering the evidence for identification of the cif and wmk genes as the major factors involved in their phenotypes. In addition, evidence supports new hypotheses for prediction of male-killing phenotypes or lack thereof based on wmk transcript length and copy number. These experiments inform efforts to understand the full basis of reproductive parasitism for basic and applied purposes and lay the foundation for future work on the function of an interesting group of Wolbachia and phage WO genes.

Keywords: Drosophila; Wolbachia; cytoplasmic incompatibility; male killing; prophage WO; reproductive parasitism; transgenics.

FIG 1

Map of gene candidates assessed for reproductive parasitism across the wMel genome. Prophage WO regions are shown in their indicated colors. Gene positions, indicated by black lines, are approximate. The white line at the top indicates the first nucleotide position in the genome. The WOMelB prophage region is expanded below to show the relative positions of these genes. Genes are roughly to scale, with candidates shown in green and noncandidate genes in white. Different arrow directions indicate locations on opposite DNA strands.

FIG 2

Transgenic expression of WD0633 does not recapitulate male killing or CI. (A) Diagrams of protein architecture using domains indicated from SMART (59) and HHpred (60) databases. (B) Sex ratios of adult flies either expressing (Act5c-Gal4) or not expressing (CyO) the indicated genes. Each sample point represents the adult offspring produced by a replicate family of 10 mothers and 2 fathers, with expressing and nonexpressing flies of a given genotype being siblings. Bars represent the mean sex ratios. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction across either expressing or nonexpressing flies. ns, results were not statistically significant. (C) Hatch rate of embryos with infected (filled sex symbol) or uninfected (unfilled sex symbol) flies expressing or not expressing an indicated gene with the nanos-Gal4:VP16 gonad-specific driver. Bars represent the median hatch rate. Each dot represents the hatch rate of offspring of a single male and female. Black dots indicate a cross with WD0633, and gray dots indicate crosses without transgenes. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction.

FIG 3

Transgenic expression of wMel wmk homologs does not recapitulate male killing. (A) Nucleotide phylogeny of wMel wmk homologs. (B) Sex ratios of adult flies either expressing (Act5c-Gal4) or not expressing (CyO) the indicated genes. Each sample point represents the adult offspring produced by a replicate family of 10 mothers and 2 fathers, with expressing and nonexpressing flies of a given genotype being siblings. Bars represent the mean sex ratios. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction across either expressing or nonexpressing flies. (C) Amino acid alignment of Wmk (WD0626) and its homologs in wMel. Green highlights indicate amino acids unique to Wmk. Blue boxes indicate the NCBI-predicted position of the HTH DNA-binding domains (above the indicated amino acids). (D) Schematic of amino acids unique to WD0626 (Wmk) across the protein sequence, as indicated by black lines. Locations are approximate.

FIG 4

Additional male-killing gene candidates do not induce a biased sex ratio with transgenic expression. (A) Diagrams of protein architecture using the indicated domains from the SMART (59) and HHpred (60) databases. (B) Sex ratios of adult flies either expressing (Act5c-Gal4) or not expressing (CyO) the indicated genes. Each sample point represents the adult offspring produced by a replicate family of 10 mothers and 2 fathers, with expressing and nonexpressing flies of a given genotype being siblings. Bars represent the mean sex ratios. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction across either expressing or nonexpressing flies.

FIG 5

Coexpression of cifA and wmk neither enhances the wmk sex ratio bias nor recapitulates CI induction. (A) Sex ratios of adult flies either expressing (Act5c-Gal4) or not expressing (CyO) the indicated genes. Each sample point represents the adult offspring produced by a replicate family of 10 mothers and 2 fathers, with expressing and nonexpressing flies of a given genotype being siblings. Bars represent the mean sex ratios. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction across either expressing or nonexpressing flies. (B) Hatch rate of embryos with infected (filled sex symbol) or uninfected (unfilled sex symbol) flies expressing or not expressing an indicated gene with the nanos-Gal4:VP16 gonad-specific driver. Bars represent the median hatch rate. Each dot represents the hatch rate of offspring of a single male and female. Colors indicate the presence or absence of the transgenes as indicated in the key. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction.

FIG 6

Expression of wmk with an alternative upstream start codon results in loss of a sex ratio bias. (A) Diagram of locations of alternative start codons up to 100 bp upstream of wmk or its homologs in the indicated strains. Purple stripes indicate the codons (not to scale). *, wRec causes CI in its native host but can cause male killing when introgressed into a sister species. **, wBol1-b natively infects a nondrosophilid host, the Hypolimnas bolina blue moon butterfly, while all other strains in the diagram infect drosophilid hosts. (B) Diagram of hypothetical model where multiple wmk transcripts of various lengths are expressed. (C) Sex ratios of adult flies either expressing (Act5c-Gal4) or not expressing (CyO) the indicated genes. Each sample point represents the adult offspring produced by a replicate family of 10 mothers and 2 fathers, with expressing and nonexpressing flies of a given genotype being siblings. Bars represent the mean sex ratios. Statistics are based on a Kruskal-Wallis one-way ANOVA followed by Dunn’s correction across either expressing or nonexpressing flies.