The Phenotypic Effects of Royal Jelly on Wild-Type D. melanogaster Are Strain-Specific

The role for royal jelly (RJ) in promoting caste differentiation of honeybee larvae into queens rather than workers is well characterized. A recent study demonstrated that this poorly understood complex nutrition drives strikingly similar phenotypic effects in Drosophila melanogaster, such as increased body size and reduced developmental time, making possible the use of D. melanogaster as a model system for the genetic analysis of the cellular mechanisms underlying RJ and caste differentiation. We demonstrate here that RJ increases the body size of some wild-type strains of D. melanogaster but not others, and report significant delays in developmental time in all flies reared on RJ. These findings suggest that cryptic genetic variation may be a factor in the D. melanogaster response to RJ, and should be considered when attempting to elucidate response mechanisms to environmental changes in non-honeybee species.


Introduction
In the social honeybee Apis mellifera, the royal jelly (RJ) diet is known to be responsible for the development of female larvae into queen rather than worker bees [1][2][3]. Fertile queen bees are markedly larger, develop more rapidly, and have significantly longer lifespans than sterile worker bees [1][2][3]. Interestingly, differentiation of honeybee larvae into queens rather than workers is not the result of underlying differences in genotype, but instead has been shown to be a direct consequence of epigenetic changes brought about through larval ingestion of RJ [4,5]. These effects of RJ were previously shown to be conserved in Drosophila melanogaster, as rearing D. melanogaster in the presence of RJ decreased developmental time, increased body size, and increased fertility and lifespan of the female progeny [6]. It was concluded from these findings that RJ acts across insect species, implying a functionally conserved mechanism of action.
It is interesting to note, however, that this previous study demonstrated phenotypic and physiological effects of RJ using only one wild-type D. melanogaster strain, Canton-S (Can-S). Multiple wild-type strains of D. melanogaster exist, and significant differences in responses to allelic mutations and environmental stimuli have been observed both between and even within certain wild-type strains [7][8][9]. As a result, wild-type organisms utilized in a study must be strictly controlled to avoid confounding experimental variables, however, doing so often provides a limited view of what occurs in nature, as the alleles in a genome usually vary among individuals [10][11][12][13]. Furthermore, cryptic genetic variation (CGV), which does not contribute to the normal range of phenotypes observed in a population, can significantly modify an existing phenotype in response to environmental changes [13]. Such a complex phenomenon could be particularly important to consider when introducing an abnormal food source, such as RJ, into a D. melanogaster population.
The importance of CGV and wild-type strain differences became apparent when we attempted to replicate previously reported RJ-fed D. melanogaster findings [6]. Despite multiple attempts, we were only able to replicate increases in the size of RJ-fed flies when using the Can-S D. melanogaster strain. Furthermore, in contrast to a previous report [6], we observed increased developmental times for every D. melanogaster strain reared on RJ containing food. Our findings suggest that the differential responses to RJ may be the result of CGV, and reaffirm the importance of considering genetic background dependence as a critical determinant capable of modifying phenotypes and/or behaviors that arise after environmental challenges.

Materials and Methods
Canton-S and Oregon-R Drosophila melanogaster were procured from Bloomington (stocks #1 and #5, respectively), and were maintained as isogenized laboratory strains. All flies were kept in an incubator with a 12:12 LD cycle at a temperature of 23-25.5°C, with an unregulated humidity ranging from 50-60%. Both strains were maintained on 2.5g of instant fly food (proprietary Carolina Formula 4-24 Medium, Blue; Burlington, NC) combined with 10 mL of water. Pure honeybee royal jelly (RJ) was a gift from Dr. Ryszard Maleszka, which he procured from his own laboratory's colonies. It was stored at 4°C and protected from light when not in active use.
Fly food for the rearing experiments was prepared by adding RJ to instant fly food through manual mixing of the appropriate amount of RJ (0%, 10%, 15%, 20%, or 25% as indicated, w/v) into 10 ml of water prior to addition of the water to 2.5g instant food.
To assess differences in body size in both strains of D. melanogaster in response to RJ, mating was conducted by initially placing five male and five female flies into a vial, and upon confirmation of larvae visibly present within the vial, removing parents. Vials were visually checked daily, and action was taken for each vial as needed. Eclosed flies were removed from the vial after each eclosion period, and whole body length of male and non-virgin female flies was subsequently measured using standard millimeter micro-calipers under a dissecting microscope. To ensure greater accuracy of measurements, females were not measured until they had completely shed their pupal waste products, but before they any signs of a distended abdomen. Measurements were made across multiple vials for each condition (of RJ%), until a sufficient number of flies had been measured for each condition.
Pilot experiments were conducted at Stanford University and replicated at Bridgewater State University, however, finding of no difference between initial preliminary results prompted collaborative data collection between the two universities thereafter. Data from both locations was gathered in an identical fashion and was combined for all analyses after no difference was observed between data collected at either location (as described below).
Length of time until the first eclosion was determined for at least two vials for both strains across all conditions (of RJ%) by visual inspection of the vials after removal of the parental generation (mating and parental generation removal as described above). Vials were visually inspected every dusk (1 hr before lights off) until flies began to eclose, and then continued to be inspected visually in this way for three consecutive days thereafter.
Statistical analyses were performed using one-way ANOVA, with Tukey HSD for subsequent pairwise comparisons (or Dunnett's T3 for comparisons with unequal variances found by a Levene's Test), for both time to first eclosion and body length in male and female RJ fed flies compared to control fed flies. Data obtained from both Stanford University and Bridgewater State University were combined for these analyses, as a Two-Way ANOVA revealed no significant difference between the results obtained in any experiments at the two different institutions (F 1,279 = 0.13, p = 0.72).
As a prior study reporting larger D. melanogaster size was conducted with Canton-S D. melanogaster (Can-S) [6], we hypothesized that the dramatic phenotypic effects of RJ could be a manifestation of CGV. To test this, we repeated our RJ experiments using Can-S flies. Interestingly, we found that female Can-S flies reared on 15% (n = 52) and 20% (n = 34) RJ exhibited a significantly longer body length compared to controls (n = 65), and that body length increased in a dose-dependent manner (F 3,210 = 26.63, p<0.001, Levene's p = 0.12, p<0.001 for 15% and 20% comparisons to 0% ; Fig 2A and 2B, S1C Fig, and S1 Data). No change in body length was observed for female flies reared on 10% RJ (n = 63) compared to controls. Subsequent pairwise comparisons showed that of the female Can-S flies, those reared on 10% RJ were significantly smaller than those reared on 15% (p = 0.001) or 20% RJ (p<0.001), but that flies reared on 15% RJ were not significantly smaller than those reared on 20% RJ (p = 0.33 ; Fig 2A and 2B).

Discussion
Our findings demonstrate that many of RJ's effects on honeybees appear to be preserved in Drosophila melanogaster, but only in some strains. This is consistent with a large body of evidence indicating that CGV can have significant experimental impact [7][8][9][14][15][16][17][18], and highlights how profoundly genetic background can influence phenotypic outcomes. As there is evidence that RJ differentially effects longevity, body size, and fecundity, of multiple other species, such as silkmoths [19], crickets [19], C. elegans [20], and mice [21], it will be important to consider how CGV may be influencing such observations. This will be particularly true for studies aiming to elucidate the mechanism by which RJ response has been retained across species. For example, as the EGFR signaling pathway has been specifically implicated in RJ response [6], observations that the phenotypic effects of EGFR knockout mice were genetic background dependent [18] will be important to take into consideration when investigating RJ response in mammals. Taken together, this emphasizes the importance of considering CGV in studies seeking to elucidate RJ's effects in non-honeybee species.
In contrast to our findings of body size differences in D. melanogaster, we failed to validate a previous report that RJ reduces D. melanogaster developmental time [6]. Rather, we consistently observed significant dose-dependent developmental delays in both D. melanogaster lines reared in the presence of RJ, a finding which was substantiated in work from Shorter et.al [22] published while this manuscript was under preparation. Further, we also failed to replicate a previous report of female-specific RJ effects in Can-S D. melanogaster [6], instead repeatedly finding that it increased body size of both sexes in this strain. Shorter et al. [22] confirmed this finding as well, not only in wt1118 Can-S B D. melanogaster, but also in three additional wild-type lines they collected [22]. In further keeping with our results, Shorter et al. similarly reported strainspecific differences in RJ response, finding that RJ decreased fly size in one of the four lines they observed. Though a number of experimental controls would be required to make the two studies fully comparable, our and Shorter et al.'s work demonstrate strain-specific responses to RJ, and highlight the importance of accounting for CGV in elucidating RJ's mechanism of function in species other than the honeybee.
Taken together, it is clear that the impacts of CGV on RJ studies should not be taken lightly. Although we investigated differential effects of RJ in only two wild-type strains of D. melanogaster in this work, it is likely that extending this study to a greater number of wild-type strains would uncover a number of additional phenotypic differences that could lend greater insight into RJ's mechanism of action. In summary, our findings indicate that obtaining a comprehensive understanding of RJ's effects will require a multidimensional approach that takes into account the diverse effects that standing genetic variations can produce. differences in size compared to controls while (B) male Ore-R flies reared on 20% RJ were slightly smaller than controls. In contrast, both (C) female and (D) male Can-S flies raised on 20% RJ were significantly larger than controls. However, a significant change in body length was not observed in (C) female flies raised on 10% RJ. (TIF)