Common and novel transcriptional routes to behavioral maturation in worker and male honey bees

Behavioral and genetic correlates of heterogeneity in learning performance in individual honeybees, Apis mellifera

Learning an olfactory discrimination task leads to heterogeneous results in honeybees with some bees performing very well and others at low rates. Here we investigated this behavioral heterogeneity and asked whether it was associated with particular gene expression patterns in the bee's brain. Bees were individually conditioned using a sequential conditioning protocol involving several phases of olfactory learning and retention tests. A cumulative score was used to differentiate the tested bees into high and low performers. The rate of CS+ odor learning was found to correlate most strongly with a cumulative performance score extracted from all learning and retention tests. Microarray analysis of gene expression in the mushroom body area of the brains of these bees identified a number of differentially expressed genes between high and low performers. These genes are associated with diverse biological functions, such as neurotransmission, memory formation, cargo trafficking and development.

The buzz within: the role of the gut microbiome in honeybee social behavior

Gut symbionts influence the physiology and behavior of their host, but the extent to which these effects scale to social behaviors is an emerging area of research. The use of the western honeybee (Apis mellifera) as a model enables researchers to investigate the gut microbiome and behavior at several levels of social organization. Insight into gut microbial effects at the societal level is critical for our understanding of how involved microbial symbionts are in host biology. In this Commentary, we discuss recent findings in honeybee gut microbiome research and synthesize these with knowledge of the physiology and behavior of other model organisms to hypothesize how host-microbe interactions at the individual level could shape societal dynamics and evolution.

OUP accepted manuscript

Many species have separate haploid and diploid phases. Theory predicts that each phase should experience the effects of evolutionary forces (like selection) differently. In the haploid phase, all fitness-affecting alleles are exposed to selection, whereas in the diploid phase, those same alleles can be masked by homologous alleles. This predicts that selection acting on genes expressed in haploids should be more effective than diploid-biased genes. Unfortunately, in arrhenotokous species, this prediction can be confounded with the effects of sex-specific expression, as haploids are usually reproductive males. Theory posits that, when accounting for ploidal- and sex-specific expression, selection should be equally efficient on haploid- and diploid-biased genes relative to constitutive genes. Here, we used a multiomic approach in honey bees to quantify the evolutionary rates of haploid-biased genes and test the relative effects of sexual- and haploid-expression on molecular evolution. We found that 16% of the honey bee’s protein-coding genome is highly expressed in haploid tissue. When accounting for ploidy and sex, haploid- and diploid-biased genes evolve at a lower rate than expected, indicating that they experience strong negative selection. However, the rate of molecular evolution of haploid-biased genes was higher than diploid-based genes. Genes associated with sperm storage are a clear exception to this trend with evidence of strong positive selection. Our results provide an important empirical test of theory outlining how selection acts on genes expressed in arrhenotokous species. We propose the haploid life history stage affects genome-wide patterns of diversity and divergence because of both sexual and haploid selection.

Searching beyond the streetlight: Neonicotinoid exposure alters the neurogenomic state of worker honey bees

Neonicotinoid insecticides have been implicated in honey bee declines, with many studies showing that sublethal exposure impacts bee behaviors such as foraging, learning, and memory. Despite the large number of ecotoxicological studies carried out to date, most focus on a handful of worker phenotypes leading to a "streetlight effect" where the a priori choice of phenotypes to measure may influence the results and conclusions arising from the studies. This bias can be overcome with the use of toxicological transcriptomics, where changes in gene expression can provide a more objective view of how pesticides alter animal traits. Here, we used RNA sequencing to examine the changes in neurogenomic states of nurse and forager honey bees that were naturally exposed to neonicotinoids in the field and artificially exposed to neonicotinoids in a controlled experiment. We found that neonicotinoid exposure influenced the neurogenomic state of foragers and nurses in different ways; foragers experienced shifts in expression of genes involved in cognition and development, while nurses experienced shifts in expression of genes involved in metabolism. Our study suggests that neonicotinoids influence nurse and forager bees in a different manner. We also found no to minimal overlap in the differentially expressed genes in our study and in previously published studies, which might help reconcile the seemingly contradictory results often reported in the neonicotinoid literature.

Mate Choice, Sex Roles and Sexual Cognition: Neuronal Prerequisites Supporting Cognitive Mate Choice

Across taxa, mate choice is a highly selective process involving both intra- and intersexual selection processes aiming to pass on one’s genes, making mate choice a pivotal tool of sexual selection. Individuals adapt mate choice behavior dynamically in response to environmental and social changes. These changes are perceived sensorily and integrated on a neuronal level, which ultimately leads to an adequate behavioral response. Along with perception and prior to an appropriate behavioral response, the choosing sex has (1) to recognize and discriminate between the prospective mates and (2) to be able to assess and compare their performance in order to make an informed decision. To do so, cognitive processes allow for the simultaneous processing of multiple information from the (in-) animate environment as well as from a variety of both sexual and social (but non-sexual) conspecific cues. Although many behavioral aspects of cognition on one side and of mate choice displays on the other are well understood, the interplay of neuronal mechanisms governing both determinants, i.e., governing cognitive mate choice have been described only vaguely. This review aimed to throw a spotlight on neuronal prerequisites, networks and processes supporting the interaction between mate choice, sex roles and sexual cognition, hence, supporting cognitive mate choice. How does neuronal activity differ between males and females regarding social cognition? Does sex or the respective sex role within the prevailing mating system mirror at a neuronal level? How does cognitive competence affect mate choice? Conversely, how does mate choice affect the cognitive abilities of both sexes? Benefitting from studies using different neuroanatomical techniques such as neuronal activity markers, differential coexpression or candidate gene analyses, modulatory effects of neurotransmitters and hormones, or imaging techniques such as fMRI, there is ample evidence pointing to a reflection of sex and the respective sex role at the neuronal level, at least in individual brain regions. Moreover, this review aims to summarize evidence for cognitive abilities influencing mate choice and vice versa. At the same time, new questions arise centering the complex relationship between neurobiology, cognition and mate choice, which we will perhaps be able to answer with new experimental techniques.

Sex-specific molecular specialization and activity rhythm-dependent gene expression in honey bee antennae

We performed an RNA-seq-based comparison of gene expression levels in the antennae of honey bee drones and time-trained foragers (workers) collected at different times of the day and different activity states. Interestingly, olfaction-related genes [i.e. odorant receptor (Or) genes, odorant binding protein (Obp) genes, carboxyl esterase (CEst) genes, etc.] showed stable gene expression differences between drone and worker antennae. Drone antennae showed higher expression of 24 Or genes, of which 21 belong to the clade X which comprises the receptor for the major queen pheromone compound 9-ODA. This high number of drone-biased Or genes suggests that more than previously thought play a role in sex-pheromone communication. In addition, we found higher expression levels for many non-olfaction-related genes including nitric oxide synthase (NOS), and the potassium channel Shaw In contrast, workers showed higher expression of 67 Or genes, which belong to different Or clades that are involved in pheromone communication as well as the perception of cuticular hydrocarbons and floral scents. Further, drone antennae showed higher expression of genes involved in energy metabolism, whereas worker antennae showed higher expression of genes involved in neuronal communication, consistent with earlier reports on peripheral olfactory plasticity. Finally, drones that perform mating flight in the afternoon (innate) and foragers that are trained to forage in the afternoon (adapted) showed similar daily changes in the expression of two major clock genes, period and cryptochrome2 Most of the other genes showing changes with time or onset of daily flight activity were specific to drones and foragers.

Extreme Differences in Recombination Rate between the Genomes of a Solitary and a Social Bee

Social insect genomes exhibit the highest rates of crossing over observed in plants and animals. The evolutionary causes of these extreme rates are unknown. Insight can be gained by comparing recombination rate variation across the genomes of related social and solitary insects. Here, we compare the genomic recombination landscape of the highly social honey bee, Apis mellifera, with the solitary alfalfa leafcutter bee, Megachile rotundata, by analyzing patterns of linkage disequilibrium in population-scale genome sequencing data. We infer that average recombination rates are extremely elevated in A. mellifera compared with M. rotundata. However, our results indicate that similar factors control the distribution of crossovers in the genomes of both species. Recombination rate is significantly reduced in coding regions in both species, with genes inferred to be germline methylated having particularly low rates. Genes with worker-biased patterns of expression in A. mellifera and their orthologs in M. rotundata have higher than average recombination rates in both species, suggesting that selection for higher diversity in genes involved in worker caste functions in social taxa is not the explanation for these elevated rates. Furthermore, we find no evidence that recombination has modulated the efficacy of selection among genes during bee evolution, which does not support the hypothesis that high recombination rates facilitated positive selection for new functions in social insects. Our results indicate that the evolution of sociality in insects likely entailed selection on modifiers that increased recombination rates genome wide, but that the genomic recombination landscape is determined by the same factors.

Gene duplication and the evolution of phenotypic diversity in insect societies

Gene duplication is an important evolutionary process thought to facilitate the evolution of phenotypic diversity. We investigated if gene duplication was associated with the evolution of phenotypic differences in a highly social insect, the honeybee Apis mellifera. We hypothesized that the genetic redundancy provided by gene duplication could promote the evolution of social and sexual phenotypes associated with advanced societies. We found a positive correlation between sociality and rate of gene duplications across the Apoidea, indicating that gene duplication may be associated with sociality. We also discovered that genes showing biased expression between A. mellifera alternative phenotypes tended to be found more frequently than expected among duplicated genes than singletons. Moreover, duplicated genes had higher levels of caste-, sex-, behavior-, and tissue-biased expression compared to singletons, as expected if gene duplication facilitated phenotypic differentiation. We also found that duplicated genes were maintained in the A. mellifera genome through the processes of conservation, neofunctionalization, and specialization, but not subfunctionalization. Overall, we conclude that gene duplication may have facilitated the evolution of social and sexual phenotypes, as well as tissue differentiation. Thus this study further supports the idea that gene duplication allows species to evolve an increased range of phenotypic diversity.

Transcriptomic analysis of instinctive and learned reward-related behaviors in honey bees

We used transcriptomics to compare instinctive and learned, reward-based honey bee behaviors with similar spatio-temporal components: mating flights by males (drones) and time-trained foraging flights by females (workers), respectively. Genome-wide gene expression profiling via RNA sequencing was performed on the mushroom bodies, a region of the brain known for multi-modal sensory integration and responsive to various types of reward. Differentially expressed genes (DEGs) associated with the onset of mating (623 genes) were enriched for the gene ontology (GO) categories of Transcription, Unfolded Protein Binding, Post-embryonic Development, and Neuron Differentiation. DEGs associated with the onset of foraging (473) were enriched for Lipid Transport, Regulation of Programmed Cell Death, and Actin Cytoskeleton Organization. These results demonstrate that there are fundamental molecular differences between similar instinctive and learned behaviors. In addition, there were 166 genes with strong similarities in expression across the two behaviors - a statistically significant overlap in gene expression, also seen in Weighted Gene Co-Expression Network Analysis. This finding indicates that similar instinctive and learned behaviors also share common molecular architecture. This common set of DEGs was enriched for Regulation of RNA Metabolic Process, Transcription Factor Activity, and Response to Ecdysone. These findings provide a starting point for better understanding the relationship between instincts and learned behaviors. In addition, because bees collect food for their colony rather than for themselves, these results also support the idea that altruistic behavior relies, in part, on elements of brain reward systems associated with selfish behavior.

Integrating the Dimensions of NGSS within a Collaborative Board Game about Honey Bees

The current reform in U.S. science education calls for the integration of three dimensions of science learning in classroom teaching and learning: Science and Engineering Practices, Crosscutting Concepts, and Disciplinary Core Ideas. While the Next Generation Science Standards provide flexibility in how curriculum and instruction are structured to meet learning goals, there are few examples of existing curricula that portray the integration of these dimensions as "three-dimensional learning." Here, we describe a collaborative board game about honey bees that incorporates scientific evidence on how genetic and environmental factors influence variations of traits and social behavior and requires students to collaboratively examine and use a system model. Furthermore, we show how students used and evaluated the game as a model in authentic classroom settings.

Developmental Transcriptome for a Facultatively Eusocial Bee, Megalopta genalis

Transcriptomes provide excellent foundational resources for mechanistic and evolutionary analyses of complex traits. We present a developmental transcriptome for the facultatively eusocial bee Megalopta genalis, which represents a potential transition point in the evolution of eusociality. A de novo transcriptome assembly of Megalopta genalis was generated using paired-end Illumina sequencing and the Trinity assembler. Males and females of all life stages were aligned to this transcriptome for analysis of gene expression profiles throughout development. Gene Ontology analysis indicates that stage-specific genes are involved in ion transport, cell-cell signaling, and metabolism. A number of distinct biological processes are upregulated in each life stage, and transitions between life stages involve shifts in dominant functional processes, including shifts from transcriptional regulation in embryos to metabolism in larvae, and increased lipid metabolism in adults. We expect that this transcriptome will provide a useful resource for future analyses to better understand the molecular basis of the evolution of eusociality and, more generally, phenotypic plasticity.

Extreme recombination frequencies shape genome variation and evolution in the honeybee, Apis mellifera

Meiotic recombination is a fundamental cellular process, with important consequences for evolution and genome integrity. However, we know little about how recombination rates vary across the genomes of most species and the molecular and evolutionary determinants of this variation. The honeybee, Apis mellifera, has extremely high rates of meiotic recombination, although the evolutionary causes and consequences of this are unclear. Here we use patterns of linkage disequilibrium in whole genome resequencing data from 30 diploid honeybees to construct a fine-scale map of rates of crossing over in the genome. We find that, in contrast to vertebrate genomes, the recombination landscape is not strongly punctate. Crossover rates strongly correlate with levels of genetic variation, but not divergence, which indicates a pervasive impact of selection on the genome. Germ-line methylated genes have reduced crossover rate, which could indicate a role of methylation in suppressing recombination. Controlling for the effects of methylation, we do not infer a strong association between gene expression patterns and recombination. The site frequency spectrum is strongly skewed from neutral expectations in honeybees: rare variants are dominated by AT-biased mutations, whereas GC-biased mutations are found at higher frequencies, indicative of a major influence of GC-biased gene conversion (gBGC), which we infer to generate an allele fixation bias 5 – 50 times the genomic average estimated in humans. We uncover further evidence that this repair bias specifically affects transitions and favours fixation of CpG sites. Recombination, via gBGC, therefore appears to have profound consequences on genome evolution in honeybees and interferes with the process of natural selection. These findings have important implications for our understanding of the forces driving molecular evolution.

Evolution results from changes in allele frequencies in populations. The main forces that cause such changes are natural selection and random genetic drift. However, an additional process, GC-biased gene conversion (gBGC), associated with meiotic recombination, affects the probability that alleles are passed from one generation to the next. The honeybee, Apis mellifera, has extremely high recombination rates—more than 20 times to those observed in humans. However, the reason for this is unknown and the effects of such high recombination rates on evolution are not well understood. Here we use patterns of genetic variation in the genomes of 30 honeybees to infer variation in the rate of recombination across the genome. We find that recombination rates and levels of genetic variation are strongly correlated, which is indicative of a pervasive impact of natural selection on genetic variation. We also infer a major role of DNA methylation in determining recombination rates in genes. Patterns of genetic variation appear to be strongly skewed due to the effects of gBGC, suggesting that recombination generates a bias in transmission of alleles during meiosis. This process seems to be interfering with the efficacy of selection at removing deleterious alleles and favouring beneficial ones. Recombination therefore has a huge impact on genetic variation and evolution in honeybees and appears to play a dominant role in genome evolution.

Causes and consequences of crossing-over evidenced via a high-resolution recombinational landscape of the honey bee

Background

Social hymenoptera, the honey bee (Apis mellifera) in particular, have ultra-high crossover rates and a large degree of intra-genomic variation in crossover rates. Aligned with haploid genomics of males, this makes them a potential model for examining the causes and consequences of crossing over. To address why social insects have such high crossing-over rates and the consequences of this, we constructed a high-resolution recombination atlas by sequencing 55 individuals from three colonies with an average marker density of 314 bp/marker.

Results

We find crossing over to be especially high in proximity to genes upregulated in worker brains, but see no evidence for a coupling with immune-related functioning. We detect only a low rate of non-crossover gene conversion, contrary to current evidence. This is in striking contrast to the ultrahigh crossing-over rate, almost double that previously estimated from lower resolution data. We robustly recover the predicted intragenomic correlations between crossing over and both population level diversity and GC content, which could be best explained as indirect and direct consequences of crossing over, respectively.

Conclusions

Our data are consistent with the view that diversification of worker behavior, but not immune function, is a driver of the high crossing-over rate in bees. While we see both high diversity and high GC content associated with high crossing-over rates, our estimate of the low non-crossover rate demonstrates that high non-crossover rates are not a necessary consequence of high recombination rates.

Electronic supplementary material

The online version of this article (doi:10.1186/s13059-014-0566-0) contains supplementary material, which is available to authorized users.

No genetic tradeoffs between hygienic behaviour and individual innate immunity in the honey bee, Apis mellifera

Many animals have individual and social mechanisms for combating pathogens. Animals may exhibit short-term physiological tradeoffs between social and individual immunity because the latter is often energetically costly. Genetic tradeoffs between these two traits can also occur if mutations that enhance social immunity diminish individual immunity, or vice versa. Physiological tradeoffs between individual and social immunity have been previously documented in insects, but there has been no study of genetic tradeoffs involving these traits. There is strong evidence that some genes influence both innate immunity and behaviour in social insects – a prerequisite for genetic tradeoffs. Quantifying genetic tradeoffs is critical for understanding the evolution of immunity in social insects and for devising effective strategies for breeding disease-resistant pollinator populations. We conducted two experiments to test the hypothesis of a genetic tradeoff between social and individual immunity in the honey bee, Apis mellifera. First, we estimated the relative contribution of genetics to individual variation in innate immunity of honey bee workers, as only heritable traits can experience genetic tradeoffs. Second, we examined if worker bees with hygienic sisters have reduced individual innate immune response. We genotyped several hundred workers from two colonies and found that patriline genotype does not significantly influence the antimicrobial activity of a worker’s hemolymph. Further, we did not find a negative correlation between hygienic behaviour and the average antimicrobial activity of a worker’s hemolymph across 30 honey bee colonies. Taken together, our work indicates no genetic tradeoffs between hygienic behaviour and innate immunity in honey bees. Our work suggests that using artificial selection to increase hygienic behaviour of honey bee colonies is not expected to concurrently compromise individual innate immunity of worker bees.

A worldwide survey of genome sequence variation provides insight into the evolutionary history of the honeybee Apis mellifera

The honeybee Apis mellifera has major ecological and economic importance. We analyze patterns of genetic variation at 8.3 million SNPs, identified by sequencing 140 honeybee genomes from a worldwide sample of 14 populations at a combined total depth of 634×. These data provide insight into the evolutionary history and genetic basis of local adaptation in this species. We find evidence that population sizes have fluctuated greatly, mirroring historical fluctuations in climate, although contemporary populations have high genetic diversity, indicating the absence of domestication bottlenecks. Levels of genetic variation are strongly shaped by natural selection and are highly correlated with patterns of gene expression and DNA methylation. We identify genomic signatures of local adaptation, which are enriched in genes expressed in workers and in immune system- and sperm motility-related genes that might underlie geographic variation in reproduction, dispersal and disease resistance. This study provides a framework for future investigations into responses to pathogens and climate change in honeybees.

De novo transcriptome hybrid assembly and validation in the European earwig (Dermaptera, Forficula auricularia)

BACKGROUND

The European earwig (Forficula auricularia) is an established system for studies of sexual selection, social interactions and the evolution of parental care. Despite its scientific interest, little knowledge exists about the species at the genomic level, limiting the scope of molecular studies and expression analyses of genes of interest. To overcome these limitations, we sequenced and validated the transcriptome of the European earwig.

METHODOLOGY AND PRINCIPAL FINDINGS

To obtain a comprehensive transcriptome, we sequenced mRNA from various tissues and developmental stages of female and male earwigs using Roche 454 pyrosequencing and Illumina HiSeq. The reads were de novo assembled independently and screened for possible microbial contamination and repeated elements. The remaining contigs were combined into a hybrid assembly and clustered to reduce redundancy. A comparison with the eukaryotic core gene dataset indicates that we sequenced a substantial part of the earwig transcriptome with a low level of fragmentation. In addition, a comparative analysis revealed that more than 8,800 contigs of the hybrid assembly show significant similarity to insect-specific proteins and those were assigned for Gene Ontology terms. Finally, we established a quantitative PCR test for expression stability using commonly used housekeeping genes and applied the method to five homologs of known sex-biased genes of the honeybee. The qPCR pilot study confirmed sex specific expression and also revealed significant expression differences between the brain and antenna tissue samples.

CONCLUSIONS

By employing two different sequencing approaches and including samples obtained from different tissues, developmental stages, and sexes, we were able to assemble a comprehensive transcriptome of F. auricularia. The transcriptome presented here offers new opportunities to study the molecular bases and evolution of parental care and sociality in arthropods.

Evolution of recombination and genome structure in eusocial insects

Eusocial Hymenoptera, such as the European honey bee, Apis mellifera, have the highest recombination rates of multicellular animals.(1) Recently, we showed(2) that a side-effect of recombination in the honey bee, GC biased gene conversion (bGC), helps maintain the unusual bimodal GC-content distribution of the bee genome by increasing GC-content in high recombination areas while low recombination areas are losing GC-content because of biased AT mutations and low rates of bGC. Although the very high recombination rate of A. mellifera makes GC-content evolution easier to study, the pattern is consistent with results found in many other species including mammals and yeast.(3) Also consistent across phyla is the association of higher genetic diversity and divergence with high GC and high recombination areas.(4) (,) (5) Finally, we showed that genes overexpressed in the brains of workers cluster in GC-rich genomic areas with the highest rates of recombination and molecular evolution.(2) In this Addendum we present a conceptual model of how eusociality and high recombination rates may co-evolve.

Recombination is associated with the evolution of genome structure and worker behavior in honey bees

The rise of insect societies, marked by the formation of reproductive and sterile castes, represents a major unsolved mystery in evolution. Across several independent origins of sociality, the genomes of social hymenopterans share two peculiar attributes: high recombination and low but heterogeneous GC content. For example, the genome of the honey bee, Apis mellifera, represents a mosaic of GC-poor and GC-rich regions with rates of recombination an order of magnitude higher than in humans. However, it is unclear how heterogeneity in GC content arises, and how it relates to the expression and evolution of worker traits. Using population genetic analyses, we demonstrate a bias in the allele frequency and fixation rate of derived C or G mutations in high-recombination regions, consistent with recombination's causal influence on GC-content evolution via biased gene conversion. We also show that recombination and biased gene conversion actively maintain the heterogeneous GC content of the honey bee genome despite an overall A/T mutation bias. Further, we found that GC-rich genes and intergenic regions have higher levels of genetic diversity and divergence relative to GC-poor regions, also consistent with recombination's causal influence on the rate of molecular evolution. Finally, we found that genes associated with behavior and those with worker-biased expression are found in GC-rich regions of the bee genome and also experience high rates of molecular evolution. Taken together, these findings suggest that recombination acts to maintain a genetically diverse and dynamic part of the genome where genes underlying worker behavior evolve more quickly.

Understanding the relationship between brain gene expression and social behavior: lessons from the honey bee

Behavior is a complex phenotype that is plastic and evolutionarily labile. The advent of genomics has revolutionized the field of behavioral genetics by providing tools to quantify the dynamic nature of brain gene expression in relation to behavioral output. The honey bee Apis mellifera provides an excellent platform for investigating the relationship between brain gene expression and behavior given both the remarkable behavioral repertoire expressed by members of its intricate society and the degree to which behavior is influenced by heredity and the social environment. Here, we review a linked series of studies that assayed changes in honey bee brain transcriptomes associated with natural and experimentally induced changes in behavioral state. These experiments demonstrate that brain gene expression is closely linked with behavior, that changes in brain gene expression mediate changes in behavior, and that the association between specific genes and behavior exists over multiple timescales, from physiological to evolutionary.

Age-related learning deficits can be reversible in honeybees Apis mellifera

Many animals are characterized by declining brain function at advanced ages, including honeybees (Apis mellifera). Variation in honeybee social development, moreover, results in individual differences in the progression of aging that may be accelerated, delayed, and sometimes reversed by changes in behavior. Here, we combine manipulations of social development with a measurement of sensory sensitivity, Pavlovian (associative) learning, and a proteomic technique to study the brain of aged honeybees. First, we confirm that sensory sensitivity can remain intact during aging, and that age-associated learning deficits are specific to bees that forage, a behavior typically expressed after a period of nursing activity. These initial data go beyond previous findings by showing how foragers age in social groups of different age compositions and sizes. Thereafter, we establish that learning ability can recover in aged foragers that revert to nursing tasks. Finally, we use liquid chromatography coupled to tandem mass spectrometry (LC-MS(2)) to describe proteomic differences between central brains, from reverted former foragers that varied in recovery of learning performance, and from nurse bees that varied in learning ability but never foraged. We find that recovery is positively associated with levels of stress response/cellular maintenance proteins in the central brain, while variation in learning before aging is negatively associated with the amounts of metabolic enzymes in the brain tissue. Our work provides the strongest evidence, thus far, for reversibility of learning deficits in aged honeybees, and indicates that recovery-related brain plasticity is connected to cellular stress resilience, maintenance and repair processes.