Drone Biology Part 2: Why Do Males Come From Unfertilized Eggs?
Male drones are produced when a queen lays unfertilized eggs, whereas the female workers come from fertilized eggs. It’s not the presence of a sperm alone that produces workers, for Cape honey bees (A.mellifera capensis) workers can lay unfertilized eggs that develop into female workers without mating, so what is determining the bees’ sex?
Early studies found the gene csd (“Complementary Sex Determiner”) to be responsible for -you guessed it- sex determination. Bees with one copy of csd became drones, whereas those with two different copies (“alleles”) became workers (or queens, if they were fed a rich & excessive diet).
Drones emerge from a hemizygous genotype (possessing one copy of the csd allele) which if you recall highschool biology, makes them “haploid” or “1n”. Atleast 100 csd alleles have been identified, which translates to atleast 4950 possible combination pairs1, all producing the same result: male or female.
This part 2 of drone biology ended up having little to do with drones, instead examining how gene combinations determine sex, influence egg development & where the Cape honey bees fit into this discourse.
csd Regulates Fem Protein Modifications
The Complementary Sex Determiner gene is based on a heterozygous vs homozygous/hemizygous genotype. Heterozygous means the egg contains two different copies of the csd allele (producing a female), homozygous means an egg contains two identical copies of the allele (which produce a diploid drone), whereas hemizygous genotypes contain only one copy of the csd allele (producing a normal drone). The review by Seiler & Beye state that around 11-19 csd allels exist in a single colony (“population”), with over 150 csd alleles identified in A. mellifera as a whole (reviewed in Seiler & Beye 2024).
More highschool biology: the csd gene codes for a Csd protein. In this case, when an egg produces two different Csd proteins (i.e. “heterozygous”), the genes for female sex determination are activated. When there is a combination of two identical Csd proteins (“homozygous”), whether by inbreeding or luck, a diploid male is produced. Now, if this egg has one Csd protein? You guessed it, it’ll produces a regular drone.
Csd Activation Triggers fem Gene Activity
A heterozygous Csd pairing triggers the splicing of the “feminizer” gene transcripts (fem; great name!), producing Fem proteins that ultimately produce a female sex (reviewed in Seiler & Beye 2024). No Fem proteins are expressed in males, and it is the presence of Fem that leads to further downstream signalling to continue the female development process (ovaries, stinger and so on).
Two examples of the downstream sex-specific signalling involve the activation of doublesex (dsx) and glubschauge (glu) genes by Fem. dsx is involved in reproductive organ development (testes and endophallus versus ovaries and spermatheca), whereas glu plays a role in producing the compound eyes. In a drone’s case, the absence of Fem proteins means that dsx & glu remain unspliced (“unalerted”), leading to the development of the drone’s large compound eye along with his male reproductive organs.
In a nutshell (eggshell?), it is the heterozygosity of Csd that produces worker bees, and the hemizygosity that produces drones.
Having 4950 csd Combinations is Evolutionary Favourable
When a single csd allele was mutated (“removed”) in a heterozygous, female-destined egg, there was a shift to male development. When a haploid csd genotype egg (i.e. destined to be a diploid male) had one copy of the allele mutated, the egg shifted to female-sex regulation (albeit an incomplete female expression there was no fem splicing [because the eggs had already gone past that stage of csd-fem activation since they started out as homozygous for csd] Otte et al., 2023).
So how is that 4950 possible csd combinations from 100 csd alleles only produce two sexes? The short answer is that sex is not determined by variation in the allele itself (its “polymorphisms”), rather due to recognition of the alleles at its CC domain2 (Otte et al., 2023, also see diagram above). This conserved region in the csd gene establishes a single binding site and active state, so that no matter what gene csd combination is encountered, the results will remain the same- two different csd produce a female, two identical csd produce a diploid drone, and a single csd produces a regular bloke. Otte & colleagues proposed that it is favourable for the number of csd alleles to continue increasing with evolutionary time, as it decreases the chances of forming an identical pairing and thus diploid drones (which are evolutionary dead ends).
An Egg is Activated After it Passes Through the Oviduct
The work of Sasaki & Obara showed that honey bee eggs begin cell division and embryogenesis after passing through the queen (or worker’s) oviduct (Sasaki & Obara, 2002). This movement from the lateral oviduct through to the common oviduct “distorts” or “damages” the egg, which stimulates it to start dividing. It is thought this mechanical distortion works in addition to the environmental changes, going from from being inside the queen’s (or worker’s) body and into that of the hive environment (“osmotic shock”).
If the egg was destined to become a worker, then sperm is released from the spermatheca and into the common oviduct, and the now-fertilized egg is released into the cell. Sasaki & Obara built on previous work that showed queens fertilized eggs within 10-28 seconds after they were relocated into the common oviduct and found that if an egg remained unfertilised for a longer period (~30 minutes), it will no longer accept sperm and will continue developing into a drone.
Drone-Destined Eggs Do Not Undergo Traditional Meiosis
Yet more reviews of high school biology: Meiosis is the process by which gametes (“sex cells” are made). Broadly, it is a two-part process: Meiosis I whereby the cell is cloned, and meiosis II whereby that cloned cell’s chromosomes are “halved”, giving us the haploid (i.e. a single set of unpaired chromosomes, referred to as “n”); male sperm or female ova.
In Hymenoptera, meiosis I does not occur (it is “aborted”), as the drone’s cells are already haploid. When it’s time for gamete production, meiosis I functions more like mitosis, whereby the sperm cell is cloned, producing two haploid copies. In meiosis II, cloned cell division takes place, and these two cells become four, producing four haploid sperms with a genetic content identical to the rest of the drone’s body.
“So what”, you ask? It would mean the concept of “dominant” and “recessive” traits is unapplicable- diploidy (“2n”) is what makes this possible, as one of the “n”s can express a dominant gene, while the other expresses a “recessive” gene. In haploidy, what you’re born with is what you’re passing on- if the drone sperm is expressing a gene for more aggression, then when he mates with a queen and the queen uses his sperm for an egg, that worker has a higher chance of turning out more aggressive.
Another drawback is that the drone is more susceptible to diseases as it does not have a second set of genes that could carry some sort of innate resistance. If a queen was to mate with a single drone, this would lead to strong genetic bottlenecking- you won’t get much diversity from a drone. This bottleneck is counteracted by Her Majesty’s liberal polyandry in the drone congregation areas (which in turn are made up from drones of different colonies).
There is an upside to this genetic blockage – it would mean that any drone that is born from a mother queen with genetic diseases does not survive and will not pass on these faulty genes to a new colony’s queen.
Evolutionary Benefits of Haploid Drones
From the above example, the loss of recessive and dominance has its benefits: consider a recessive, deleterious trait- let’s use “one pair of wings” as an example (recall that bees have two pairs). In a diploid population, the gene may be carried but not expressed (“heterozygosity”). If the diploid individual mates with another carrier, there is a chance for their offspring to end up with one pair of wings. In haploidy, this trait is always expressed in the individual, and as it is deleterious, the drone dies, as only those with the “two pairs of wings” trait can reproduce, thus leaving us with colonies that only contain bees with the correct number of wings.
Sidenote: Viable Diploid Males are Found in other Hymenopterans
Euodynerus foraminatus is a potter wasp species that births viable diploid males who also produce fertile offspring. “Diploid males in E. foraminatus are sexually competent, father diploid daughters, and can transmit either allele at a locus through their sperm. Furthermore, daughters of these diploid males also mate normally and are fully fertile.” (Cowan & Stahlhut, 2004).
The authors propose that because male and female broods develop in the same nest, they have adapted to mate with one another after maturing (Cowan & Stahlhut, 2004). Perhaps this insures the wasps against travelling long distances or searching for a mate in nature (Where diseases and other wasp-sized natural disasters await!). If diploid males were a dead end for the species, then it’s expected this population would have died out, however they are still around, and unlike in honey bees, these diploid matings do not produce triploid offspring.
Reproductive Genes & Gender Identity
Non-technical readers are welcome to skip this section
From a molecular perspective, there are five major genes involved in male vs female sex determination: csd, fem & dsx from above, as well as tra2 (Transformer Optimus Prime, a well-studied gene in Drosophila) & rbp1 (retinol binding protein 1; Pan et al., 2022). The latter two, Tra2 & rpb1 are RNA-splicing factors involved in worker ovary development. Pan & colleagues found additional genes involved in ovarian development, with functions like vitellogenin absorption, polar body formation and as transport proteins. However, since this article is about drones, this is about all the coverage they’ll get.
Multiple (6) genes were identified in testicular development and spermatogenesis. Identifying them individually doesn’t mean much, as an organism is a result of the whole being greater than the sum of the parts (something the neuro-socio-bio-psycho-oogaboogaologists don’t get). Until then, feel free to appreciate that there are multiple genes involved in a drone’s reproductive capacity, presented here as dot points based on the aforementioned work of Pan et al., 2022:
- dpp is involved in germline stem cell renewal and differentiation (making more “primordial sperms”)
- dicer1 is required in germ cell development & spermatogenesis (based on studies from mammals)
- stat92e is used in establishing and maintaining germline stem cells in the testes
- spata5 & spata17 are ATPases with diverse functions (breaking down ATP into ADP to release energy for various processes), some of which have been reported with male reproductive organ development
- Lastly, muc1 is a gene that codes for a membrane protein involved in spermatozoa maturation
As of this article’s publication in on 31st August 2025, there were no followup projects on the work of Pan et al., 2022. One followup question I’d like to know is whether increasing the expression of one of these genes (via drugs or [more likely] gene therapy) result in enhanced drone reproductive capacity with no adverse effects.
Note: Epigenetic differences are beyond the scope of this post & beekeeping in general. If you’re concerned that your drone larvae are malformed because of a lack DNA CpG methylation in the testicular germ cells… get another job.
What of A.m. capensis & Laying Workers?
In a modification of meiosis, diploidy is restored in A.m. capensis workers by fusion of components from the first polar body and the egg pronucleus, forming the zygote nucleus (“diploid egg”; Verma & Ruttner 1983). Cape workers can reactivate their ovaries and lay eggs when a queen is absent from the hive (or even in her presence, but far less frequently). As the eggs are laid diploid, they develop into female worker bees, clones of their mother, but not their queen (who remains their genetic “mother”).
The technical term for these laying workers is “central fusion automixis”, whereby instead of normal haploid gametes, two the central meiotic products fuse in anaphase II (reviewed in Mumoki et al., 2021). This restores diploidy without reducing recombination, so the csd heterozygosity is maintained (recall that workers are heterozygous for csd which gives a female genotype, and since cape workers are “cloning” their own cells, they are cloning their heterozygous csd configuration). When these eggs are laid, they behave as if they were fertilised- the signalling cascade triggers Fem & Tra2 activity and female embryonic development follows.
It has been proposed that a single recessive gene drives this thelytoky (“female birth”; Lattorff et al., 2005 citing Ruttner’s Taxonomy and Biogeography of Honeybees). Latroff and colleagues also reject the idea of it being a bacterial infection inherited from their mother queen that induces parthenogenesis, as this would have produced different phenotypes “including cytoplasmic incompatibility, male killing and induction of parthenogenesis”, rather than constantly producing the same results (i.e. diploid laying workers).
They identified the the gene in its recessive genotype as being the trigger for A.m. capensis workers to produce diploid eggs3; other bee subspecies are either homozygous or heterozygous for the th gene, and only produce haploid drone eggs (Lattorff et al., 2005).
A somewhat-related followup has now been published, describing the differences between male and female bees
Read the whole post and found puns, typos mistakes? Voice your displeasure with a strongly-worded message to admin@tgbee.com.au, or by filling this form.
Footnotes
- *For those who want the math, “4950 pairings” is based on how many unique unordered pairs that can be formed:
It’s not “100 x 100” (10,000) as that would lead to duplicates, e.g. pairing of csd1 & csd2 is the same as csd2 & csd1, so that eliminates half of them, giving us ~5000 “unique” pairing. Of those unique pairings, 100 combinations would be diploid (i.e. 50 pairs – 100 alleles divided by 2), which produces a diploid drone and is removed by workers in the larval stage, so in reality it becomes 9,900 /2, giving us 4950. ↩︎ - This is not to say that the binding takes place at these CC domains, for they are structural rather than functional. “CC” is short for “coiled-coil” domain, which is a protein structure that has two or more alpha helixes (“helices”) configurations wrapped around one another. Their presence allows a protein to form stable & robust complexes, and since it works so well, evolution imposes few or no modifications on them across species (i.e. are “highly conserved”). The CC domains have found their way into essential functions like cell divisions, transport, signalling and immune response; disrupt them (via mutations) and the consequences are lethal.
We can add one more “essential function” to the CC domain list: sex determination in A. mellifera. ↩︎ - It also enables the laying workers to produce queen-like pheromones and stimulate rapid ovarian activation, resulting in host hives being “fooled” into thinking that this Cape worker is really their “queen”. This bypasses worker policing & will lead to eventual colony collapse as the laying worker does not live long. ↩︎
