Recent global declines in bee health have elevated the need for a more complete understanding of the cellular stress mechanisms employed by diverse bee species. We recently uncovered the biomarker lethal (2) essential for life [l(2)efl] genes as part of a shared transcriptional program in response to a number of cell stressors in the western honey bee (Apis mellifera). Here, we describe another shared stress-responsive gene, glycine N-methyltransferase (Gnmt), which is known as a key metabolic switch controlling cellular methylation reactions. We observed Gnmt induction by both abiotic and biotic stressors. We also found increased levels of the GNMT reaction product sarcosine in the midgut after stress, linking metabolic changes with the observed changes in gene regulation. Prior to this study, Gnmt upregulation had not been associated with cellular stress responses in other organisms. To determine whether this novel stress-responsive gene would behave similarly in other bee species, we first characterized the cellular response to endoplasmic reticulum (ER) stress in lab-reared adults of the solitary alfalfa leafcutting bee (Megachile rotundata) and compared this with age-matched honey bees. The novel stress gene Gnmt was induced in addition to a number of canonical gene targets induced in both bee species upon unfolded protein response (UPR) activation, suggesting that stress-induced regulation of cellular methylation reactions is a common feature of bees. Therefore, this study suggests that the honey bee can serve as an important model for bee biology more broadly, although studies on diverse bee species will be required to fully understand global declines in bee populations.

Pollination services provided by diverse bee species are critical for the health of both agricultural and ecological systems. In the face of global bee declines (Zattara and Aizen, 2021), characterization of the mechanisms used by diverse bee species to respond to stress will contribute to a more complete understanding of the determinants of pollinator health (López-Uribe et al., 2019). In the honey bee, mortality at the individual and colony levels is likely the result of multiple interacting stressors, including nutritional stress due to loss of appropriate forage, chemical poisoning from pesticides, changes to natural living conditions brought about through large-scale beekeeping practices, myriad environmental changes due to climate change, and infection by insect parasites and pathogenic microbes (Hristov et al., 2020; Steinhauer et al., 2018). In multiple species, including the model invertebrates Drosophila melanogaster and Caenorhabditis elegans, these same environmental factors are known to contribute to organismal disease by negatively impacting the cellular process of protein synthesis, folding and degradation (also known as proteostasis) (Powers and Balch, 2013). As proteostasis is a likely interaction point for the stressors causing disease in honey bees, we have previously characterized a number of model cell stress responses in Apis mellifera, or the western honey bee. We identified triggers and responses for the honey bee unfolded protein response (UPR) of the endoplasmic reticulum (ER) (Adames et al., 2020; Johnston et al., 2016), which responds to proteostatic perturbation in this compartment, the honey bee integrated stress response (ISR) (Flores et al., 2021; Snow, 2020), which responds to amino acid deprivation or ribosome dysfunction, and the Heat Shock Response (HSR) (Bach et al., 2021; McKinstry et al., 2017), which responds to disruption of proteostasis in the cytoplasm. Using well-defined abiotic stressors to trigger these responses to perturbed proteostasis, we recently used an unbiased approach to identify novel genes altered in expression during induction of these proteostatic network responses in the honey bee using transcriptome profiling (RNASeq) (Adames et al., 2020; Bach et al., 2021; Flores et al., 2021). Among these genes, we described the induction of members of the l(2)efl family of small heat shock protein genes (Shih et al., 2021), which have been proposed as stress biomarkers (Cho et al., 2022; McAfee et al., 2020a,b; Shih et al., 2021) and antiviral genes (Brutscher et al., 2017; McMenamin et al., 2018) in honey bees. However, other potential common stress genes were not explored.

In addition to understanding how honey bees are affected by and respond to various stressors, there is great interest in illuminating the factors that influence the health of non-Apis bee species (López-Uribe et al., 2019). While new genome resources allow for comparison of ‘omics’-based studies between bee species (Trapp et al., 2017), functional studies to support hypotheses generated using ‘omics’ data are equally important (Grozinger and Zayed, 2020). Evidence suggests that studies using honey bees cannot take the place of research on other bee species because of the diverse life histories within this phylogenetic group (Wood et al., 2020). For example, sterile workers of the eusocial honey bee might be expected to have a cellular stress response that diverges significantly from that of solitary bees constrained by reproduction and somatic maintenance trade-offs. Although over 90% of all bee species are solitary (Danforth et al., 2019), the volume of functional studies using solitary bee species is far below that performed with honey bees and bumble bees, especially those focused on the cell and molecular level.

Here, we built on our work in honey bees by characterizing a novel target gene we hypothesized to be part of a common stress response in honey bees. We found that transcription of glycine N-methyltransferase (Gnmt for gene and mRNA versus GNMT for protein in accordance with D. melanogaster convention) is induced by both abiotic and biotic stressors. GNMT plays a key role in the methionine cycle (Fig. 1A), in which methionine is first converted to S-adenosylmethionine (SAM), and then S-adenosylhomocysteine (SAH), and finally homocysteine before being recycled back to methionine (Sanderson et al., 2019). GNMT is thought to be the main regulatory switch controlling SAM levels and the SAM/SAH ratio, and thus is considered to be a key metabolic switch controlling cellular methylation reactions. We then characterized the canonical UPR in a distantly related solitary bee species, the alfalfa leafcutting bee (Megachile rotundata) (Hedtke et al., 2013) and determined whether this putative ‘common’ stress response gene, Gnmt, is induced by the UPR in this species in order to compare stress responses in bee species with different life histories.

Fig. 1.

Gnmt is induced by diverse stressors in honey bees. (A) Schematic diagram of the methionine cycle and the role of glycine N-methyltransferase (GNMT) in the cycle. MAT, methionine adenosyltransferase; dcSAM, decarboxylated S-adenosylmethionine; SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine; AHCY, adenosylhomocysteinase; BHMT, betaine-homocysteine methyltransferase; MS, methionine synthase. (B–G) Transcript levels of Gnmt relative to β-actin in midgut tissue from adult bees captured at the landing board after thermal stress [4 h at either 45°C (n=10) or 35°C (control temperature, n=7); B], or after pharmacological induction of endoplasmic reticulum (ER) stress (n=7 DMSO vehicle control, C; n=8 tunicamycin, T; C), aminoacyl tRNA synthetase (AARS) inhibition (n=8 DMSO vehicle control; n=8 halofuginone, H; D), ribosome stress (n=8 control, n=8 cycloheximide, CHX; E), proteasome inhibition (n=8 control, n=8 ixazomib, Ixa; F) or oxidative stress (n=8 control, n=8 sodium arsenite, NaArs; G). Data (means±s.e.m.) represent expression values of the genes of interest calculated using the 2−ΔCT method for individual bees. Statistical significance is indicated by asterisks (*P<0.05, **P<0.01).

Fig. 1.

Gnmt is induced by diverse stressors in honey bees. (A) Schematic diagram of the methionine cycle and the role of glycine N-methyltransferase (GNMT) in the cycle. MAT, methionine adenosyltransferase; dcSAM, decarboxylated S-adenosylmethionine; SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine; AHCY, adenosylhomocysteinase; BHMT, betaine-homocysteine methyltransferase; MS, methionine synthase. (B–G) Transcript levels of Gnmt relative to β-actin in midgut tissue from adult bees captured at the landing board after thermal stress [4 h at either 45°C (n=10) or 35°C (control temperature, n=7); B], or after pharmacological induction of endoplasmic reticulum (ER) stress (n=7 DMSO vehicle control, C; n=8 tunicamycin, T; C), aminoacyl tRNA synthetase (AARS) inhibition (n=8 DMSO vehicle control; n=8 halofuginone, H; D), ribosome stress (n=8 control, n=8 cycloheximide, CHX; E), proteasome inhibition (n=8 control, n=8 ixazomib, Ixa; F) or oxidative stress (n=8 control, n=8 sodium arsenite, NaArs; G). Data (means±s.e.m.) represent expression values of the genes of interest calculated using the 2−ΔCT method for individual bees. Statistical significance is indicated by asterisks (*P<0.05, **P<0.01).

Honey bee colonies and caging experiments

Honey bees were collected from outbred colonies in New York, NY, USA, consisting of a typical mix of Apis mellifera Linnaeus 1758 subspecies found in North America, at different times during the months of April–October. Source colonies were visually inspected for symptoms of common bacterial, fungal and viral diseases of honey bees. For experiments analyzing diverse stressors, adult bees were collected from the landing board and kept in 2.2×8.6×21.3 cm acrylic cages with a sliding door machined at Carleton Labs, Columbia University, USA. Cages were maintained in an incubator at 35°C in the presence of PseudoQueen (Contech, Victoria, BC, Canada) as a source of queen mandibular pheromone (QMP). For these experiments, bees were fed 33% sucrose via a modified 1.5 ml screw-cap tube, with sucrose solution alone or containing model stressors that are summarized in Table 1. These compounds and their effects in insects have been characterized in the model insect D. melanogaster (Chow et al., 2013; Deliu et al., 2017; Lundgren et al., 2005; Sykiotis and Bohmann, 2008) and honey bees (Adames et al., 2020; Flores et al., 2021; Johnston et al., 2016; Shih et al., 2021), with the exception of halofuginone, which has solely been used in honey bees (Flores et al., 2021; Snow, 2020), and ixazomib, for which the honey bee effects have not yet been published (J.W.S., unpublished observations). Specifically, we used 24 µmol l−1 tunicamycin to trigger the UPR, 200 µmol l−1 halofuginone to activate the ISR, 0.5 mg ml−1 cycloheximide to inhibit translation and 200 µmol l−1 ixazomib to perturb proteasome function. DMSO is the solvent for tunicamycin, halofuginone and ixazomib, so equivalent amounts of DMSO were added to the food of the control group for these treatments. For the model stressor of heat shock, which triggers the HSR, landing board bees were maintained for 4 h in cages at 35 or 45°C. We also investigated the impact of oxidative stress (induced by 0.001% sodium arsenite for 24–48 h). After treatment, midgut tissue [and additionally for thermal stress head tissue (predominantly brain and sensory organ tissue), thorax tissue (predominantly flight muscle) and abdominal wall tissue (predominantly fat body)] were dissected for gene expression analysis. Dissected tissues were immersed in RNAlater (Invitrogen, San Diego, CA, USA) for storage at −20°C prior to RNA extraction or flash frozen in liquid nitrogen for storage at −80°C for metabolite analysis. For UPR induction experiments with 7 day old honey bees, approximately 30 newly emerged bees were collected after hatching from a capped brood frame overnight in an incubator at 35°C (unless otherwise stated), added to the acrylic cages described above, and maintained in the same conditions as landing board bees above. At 7–9 days post-eclosion, honey bees were treated with tunicamycin in the same manner as alfalfa leafcutting bees below. All honey bee experiments were performed with a minimum of three independent trials, with similar results.

Table 1.

Stressors applied to Apis mellifera and Megachile rotundata in this study

Stressors applied to Apis mellifera and Megachile rotundata in this study
Stressors applied to Apis mellifera and Megachile rotundata in this study

Vairimorpha (Nosema) ceranae infection of honey bees

A single biotic stressor, infection by the microsporidia parasite Vairimorpha ceranae was used. Vairimorpha ceranae spores were obtained from infected individuals from a colony and serially passaged through bees. To isolate spores for new infections, midguts from infected bees were individually crushed in 0.5 ml H2O with a pestle and spore number was assessed by light microscopy. Homogenized midguts were washed with water and resuspended in 1 mol l−1 sucrose solution. For experiments with newly eclosed bees, the soak method of inoculation was used as previously described (Parrella et al., 2024). Briefly, a 50 ml conical tube was filled to the 20 ml mark with newly eclosed bees (∼65 bees). Sucrose solution (1 ml) containing V. ceranae spores (2×106 ml−1) was added to the conical tube, which was mixed by gentle inversion at 5 min intervals for a total of 30 min of exposure. Bees were then transferred to the cages described above and fed sucrose solution (and supplied with a ∼5 g pollen substitute patty) for the remainder of the experiment. At 9 days post-infection, honey bee midguts were dissected and placed in RNAlater (Invitrogen). To verify infection levels, qPCR using cDNA generated as detailed above was used to determine the relative amount of V. ceranae β-actin relative to honey bee β-actin levels.

NMR metabolite quantification

Excised midguts were stored initially in individual micro-centrifuge tubes (−78°C, dry ice) for transport and subsequently stored at −80°C. A recently determined optimal aqueous metabolite extraction protocol for diverse bee samples (brain, head, body, whole bee) was used here (McDevitt et al., 2021). To each tube containing a midgut, 1 ml of a 2:1 v/v solution of acetonitrile/water was added along with two steel beads [1/8 inch and 1/16 inch (∼3.18 and 1.59 mm)]. A reciprocating bead-beating apparatus was used for thorough homogenization followed by high-speed centrifugation (10 min, 18,000 rcf) to pellet cell debris and precipitate biomacromolecules. The resulting supernatant was isolated and subjected to rotary vacuum centrifugation [3 h, 1 torr (∼133 Pa), no heating], resulting in small dry pellets that were frozen at −80°C. Each pellet was resuspended with vortex agitation in 350 μl of NMR buffer [99% D2O, 0.1 mmol l−1 TSP (IUPAC: sodium 3-trimethylsilylpropionate), 100 mmol l−1 phosphate, pH 7.4] and transferred to a ShigemiTM susceptibility matched NMR tube. Because of small broad background (residual protein/lipid) signals, one-dimensional presaturation-CPMG (100 cycles, 1 ms interpulse delay) 1H NMR spectra were acquired [352 transients, 8 steady-state scans, 3 s presaturation, 4 s acquisition, 7 s recycle, t(π/2)=7.6 µs, 66 min each]. All spectra were obtained on a 14.1 T spectrometer (600 MHz for 1H, Varian Inc., Palo Alto, CA, USA; DDR1/VNMRS generation console, vnmrj 4.2) using an inverse triple resonance probe. Spectra were profiled manually by J.R.-C. and D.S.R. using the Chenomx NMR Suite 8.1 (Chenomx, Edmonton, AB, Canada) and are reported as concentrations (sample volume 0.350 μl). Full results and analyses of these data are forthcoming in another publication (N.C.R., J.R.-C., H.V.K., R. M. Cho, N. L. Lovinger, F. Wague, A. J. Loptakin, N. G. Texeira, M.E.F., D.R. and J.W.S., unpublished). Each sample was shimmed to an initial TSP line width of 0.8 Hz or lower. Analysis employed line broadening (0.3 Hz), phasing, baseline correction (spline), and reference deconvolution of the 0 ppm TSP peak.

Alfalfa leafcutting bee pupa source storage and incubation conditions

Alfalfa leafcutting bee, Megachile rotundata (Fabricius 1787), prepupae in alfalfa leaf cocoons were acquired from a distributor (Crown Bees, Woodinville, WA, USA) and were maintained at 4°C until ready for use in caging experiments. For caging experiments, cocoons were placed in a single layer (approximately 100 cells per box) into 7×9×13 cm plastic boxes with loose fitting lids to allow for free air exchange and maintained in a 30°C incubator in complete darkness with a glass tray of water to ensure that appropriate humidity (typically 50–60%) was maintained. Because male alfalfa leafcutting bees typically eclose at 18–21 days when maintained at 30°C, 1.5 ml modified Eppendorf tubes with 30% sterile sucrose solution were placed into the containers at 18 days to provide a food source for any organisms that eclosed overnight. Eppendorf tubes were modified by piercing two opposing holes with a heated tack near the tip of the tube. After 18 days, the containers were checked 1–2 times daily. Males were removed immediately after eclosion to minimize chances of mating. Female alfalfa leafcutting bees typically eclose at 21–24 days when maintained at 30°C.

Alfalfa leafcutting bee adult caging conditions and chemical treatments

Eclosed females were immediately transferred (4–6 per cage) to the acrylic cages described above, via InsectaVac Aspirator (Bioquip, Rancho Dominguez, CA, USA), and the incubators were maintained on a 12 h light, 12 h dark cycle, at 30°C. While many bees emerged from cells completely before transfer, some bees were still inside their uncapped cell during transfer, so cages usually had some uncapped cells. Survival was noted once daily, and any dead bees were removed immediately. Each cage was provided with one modified 1.5 ml microcentrifuge tube with 30% sterile sucrose solution, and the alfalfa leafcutting bees were fed ad libitum, with the sucrose solution tube replaced once every 2–3 days. For UPR induction, at 7–9 days of age, bees were fed 30% sterile sucrose solution with or without tunicamycin at 30 µmol l−1 (Sigma, St Louis, MO, USA) as previously performed for honey bees (Adames et al., 2020; Johnston et al., 2016; Shih et al., 2021). The solute for tunicamycin is DMSO, so equivalent amounts of DMSO were added to the food of the control group in tunicamycin experiments. After 48 h, bees were cold anesthetized, and midguts were removed and placed into RNAlater (Invitrogen) for storage prior to gene expression analysis of individual bees. All alfalfa leafcutting bee tunicamycin experiments were performed for a minimum of four independent trials with similar results.

Alfalfa leafcutting bee dye feeding assay

The dye feeding assay was developed based on a protocol for analyzing the feeding patterns of D. melanogaster (Edgecomb et al., 1994). At 7–9 days of age, bees were fed 30% sterile sucrose solution alone or containing 0.16% Acid Blue 9 (TCI America, Portland, OR, USA). After 48 h, bees were cold anesthetized, and midguts were removed and placed in 200 μl water. Following dissection, midguts were crushed using sterile plastic pestles (Fisher, Waltham, MA, USA). Tubes containing the crushed midguts were centrifuged at 6000 rpm for 5 min and the absorbance of the resulting supernatants was determined using a Spark 10M spectrophotometer (Tecan, Männedorf, Switzerland) at 628 nm.

Ortholog screening of the M. rotundata genome

UPR pathway gene candidates from A. mellifera were used to find orthologs in the alfalfa leafcutting bee genome using the BLAST family of search functions (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

RNA isolation, reverse transcription and quantitative PCR for gene expression analysis

RNA was prepared from bees from the described populations by manually crushing the tissue of interest with a disposable pestle in Trizol Reagent (Invitrogen) and extracting the RNA as per the manufacturer's instructions. RNA was subsequently DNAseI treated by RQ1 RNase-Free DNase (Promega, Madison, WI, USA) and quantified. cDNA was synthesized using approximately 1 µg of RNA with the iScript cDNA Synthesis Kit (BioRad, Hercules, CA, USA). Typically, 1 µl of cDNA was then used as a template for quantitative PCR to determine the level of expression of genes of interest using the iQ SYBR Green Supermix (BioRad) in an iCycler thermal-cycler (BioRad). Primer sequences for transcripts of genes of interest are in Table S1A or from Adames et al. (2020), Johnston et al. (2016) and Shih et al. (2021). The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the level of expression of that gene relative to β-actin using the 2−ΔCT method (Schmittgen and Livak, 2008).

Alfalfa leafcutting bee Xbp1 mRNA splicing

cDNA from above was used as a template for PCR using primers (forward, 5′-GCCAGAGTCGTGTGCTGTAG-3′ and reverse 5′-GCTCCTGGGGTTAGTGTCAG-3′) that spanned the predicted Xbp1 splice sites. PCR products were run on a 2.5% agarose gel to separate spliced from unspliced Xbp1. DNA was purified from bands representing amplicons from unspliced and spliced forms of the Xbp1 transcript using a Gel Extraction Kit (Qiagen, Hilden, Germany), cloned into pDrive (Qiagen) and sequenced (Genewiz, South Plainfield, NJ, USA). The honey bee splicing assay was performed as in Johnston et al. (2016).

Statistical analysis

For statistical analysis of gene expression, GraphPad Prism was used. Data were log10 transformed and compared using unpaired t-tests with Welch's correction when values fitted normal distributions or Mann–Whitney U non-parametric tests when values did not fit normal distributions. Normality was assessed using Shapiro–Wilk tests. When more than two groups were being compared, data were compared using one-way ANOVA with Tukey's multiple comparison test when values fitted normal distributions (Fig. S1B) or a Kruskal–Wallis test with Dunn's multiple comparisons test when there was a non-normal distribution (Fig. S4E,F). In the statistical treatment of targeted NMR metabolomic data, IBM SPSS 28.0.1.0 (142) was used, all outliers were retained, and there were no missing values. The sarcosine concentrations of the control and halofuginone groups were compared with an independent samples t-test following Levene's test and the Shapiro–Wilk test. Detailed descriptions of the statistical analysis used for figures can be found in Table S1B.

Gnmt is upregulated by diverse proteotoxic stresses in honey bees

Members of the l(2)efl family of small heat shock proteins were previously identified as gene targets common to responses to diverse stressors in honey bees using transcriptome profiling (RNASeq) (Adames et al., 2020; Bach et al., 2021; Flores et al., 2021). When we analyzed these gene lists to define additional targets of a possible gene expression program shared by all stress responses, we observed Gnmt as a differentially expressed gene in all data RNASeq sets (Table S1C). Gnmt encodes a central enzyme within methionine metabolism, catalyzing the transfer of a methyl group from SAM to glycine to form sarcosine and SAH. GNMT is thought to be the principal regulator of the SAM/SAH ratio in cells in diverse organisms and therefore serves as a critical metabolic switch regulating cellular methylation reactions (Luka et al., 2009) (Fig. 1A).

To confirm these findings, we first examined heat shock-dependent induction of Gnmt in midgut tissue from honey bees kept at either at 35 or 45°C for 4 h and observed increased levels of Gnmt transcripts after heat shock in this tissue (Fig. 1B). We also confirmed that aminoacyl tRNA synthetase (AARS) inhibition (via halofuginone feeding), or ER stress (induced by tunicamycin feeding) increase Gnmt expression after 48 h of treatment compared with controls (Fig. 1C,D). We were interested in determining whether other forms of cellular stress also induced Gnmt expression. We showed that ribosome inhibition (by cycloheximide feeding; Fig. 1E) and proteasome inhibition (by ixazomib feeding; Fig. 1F) caused upregulation of Gnmt in midgut tissue relative to control bees receiving only sucrose solution after 48 h of treatment. Finally, we showed that induction of oxidative stress (via sodium arsenite feeding for 24 h) caused upregulation of Gnmt in midgut tissue relative to control bees receiving sucrose solution alone (Fig. 1G).

While orally administered treatments appear to only reliably affect the digestive tract, we can apply thermal stress to all tissues equally. Thus, we examined Gnmt expression in multiple tissues after heat shock, including head tissue (predominantly brain and sensory organ tissue), midgut, thorax tissue (predominantly flight muscle) and abdominal wall (predominantly fat body) tissues, after exposure to 35 or 45°C for 4 h (Fig. S1A). We found significant induction only in midgut tissue, suggesting that Gnmt induction may be part of a stress response that is tissue specific. We also took advantage of a previous dataset in which we had matched samples from attendant workers and reproductive queens (Shih et al., 2020) maintained at 35 or 45°C for 4 h to determine whether Gnmt is induced in queens. We found that Gnmt is increased in queen midguts to a degree comparable to that of workers following thermal stress (Fig. S1B).

Thus far, all of the stressors examined have been abiotic in nature. Therefore, we wanted to examine whether a biotic stressor, such as an infection, might lead to increased Gnmt expression. We looked at Gnmt expression in honey bees that had been infected with V. ceranae, a microsporidia species that causes individual disease in honey bees and can contribute to colony collapse (Snow, 2022). Bees were infected with V. ceranae 1 day post-eclosion via a recently described soak method or left uninfected (Parrella et al., 2024). On day 9 post-infection, we harvested midgut tissue and first measured levels of a V. ceranae transcript to confirm robust infection by this parasite (Fig. 2A). We then measured Gnmt transcript levels in these samples and found that expression of this gene was increased in infected bees when compared with uninfected bees (Fig. 2B).

Fig. 2.

Gnmt is induced by Vairimorpha ceranae infection in honey bees. (A) Vairimorpha (Nosema) ceranae infection levels (assessed by V. ceranae β-actin levels relative to honey bee β-actin) 9 days after newly eclosed bees were exposed to V. ceranae (V.c., n=6) or left uninfected (control, C; n=6). (B) Transcript levels of Gnmt relative to honey bee β-actin in midgut tissue from control or V. ceranae-infected bees. Data (means± s.e.m.) represent expression of the genes of interest calculated using the 2−ΔCT method for individual bees. Statistical significance is indicated by asterisks (**P<0.01).

Fig. 2.

Gnmt is induced by Vairimorpha ceranae infection in honey bees. (A) Vairimorpha (Nosema) ceranae infection levels (assessed by V. ceranae β-actin levels relative to honey bee β-actin) 9 days after newly eclosed bees were exposed to V. ceranae (V.c., n=6) or left uninfected (control, C; n=6). (B) Transcript levels of Gnmt relative to honey bee β-actin in midgut tissue from control or V. ceranae-infected bees. Data (means± s.e.m.) represent expression of the genes of interest calculated using the 2−ΔCT method for individual bees. Statistical significance is indicated by asterisks (**P<0.01).

Changes in metabolome correlate with transcriptional changes in methylation cycle

Increased Gnmt mRNA levels would be expected to impact levels of reactants and products in the reaction it catalyzes. Sarcosine (along with SAH) is a product of the reaction catalyzed by GNMT in which glycine is methylated using SAM as a methyl donor (Fig. 3A). We first confirmed that the effects of halofuginone on Gnmt gene expression in newly eclosed honey bees were similar to those observed above in older bees (Fig. 3B). We then used a NMR strategy for sampling the aqueous metabolome in either whole bees or bee organs, which has been pursued previously (McDevitt et al., 2021; Pizzorno et al., 2021), to examine the midgut levels of metabolites in response to halofuginone feeding. We found that the amino acid sarcosine was increased in the midgut (P<0.001, Fig. 3C). Full results and analysis of these data are forthcoming in another paper (N.C.R., J.R.-C., H.V.K., R. M. Cho, N. L. Lovinger, F. Wague, A. J. Loptakin, N. G. Texeira, M.E.F., D.R. and J.W.S., unpublished).

Fig. 3.

Halofuginone induces integrated stress response (ISR) target genes and impacts survival in newly eclosed honey bees. (A) Schematic diagram of the methionine cycle reaction catalyzed by GNMT. (B) Transcript levels of Gnmt relative to β-actin in midgut tissue from newly eclosed bees feeding on sucrose solution containing halofuginone (H; n=8) or DMSO vehicle (control, C; n=8) for 4 days. Data (means±s.e.m.) represent expression of the genes of interest calculated using the 2−ΔCT method for individual bees. (C) A box and whisker plot (median, first and third quartiles, no outliers, whiskers denote minimum/maximum) of midgut sarcosine concentration (relative to a 350 μl volume) for control and halofuginone-treated bees. Statistical significance is indicated by asterisks (**P<0.01, ***P<0.001).

Fig. 3.

Halofuginone induces integrated stress response (ISR) target genes and impacts survival in newly eclosed honey bees. (A) Schematic diagram of the methionine cycle reaction catalyzed by GNMT. (B) Transcript levels of Gnmt relative to β-actin in midgut tissue from newly eclosed bees feeding on sucrose solution containing halofuginone (H; n=8) or DMSO vehicle (control, C; n=8) for 4 days. Data (means±s.e.m.) represent expression of the genes of interest calculated using the 2−ΔCT method for individual bees. (C) A box and whisker plot (median, first and third quartiles, no outliers, whiskers denote minimum/maximum) of midgut sarcosine concentration (relative to a 350 μl volume) for control and halofuginone-treated bees. Statistical significance is indicated by asterisks (**P<0.01, ***P<0.001).

UPR pathway components are conserved in the alfalfa leafcutting bee

To determine whether the novel stress-responsive gene Gnmt behaves similarly in other bee species, we first characterized one of the model cellular responses used in honey bees in lab-reared adults of the solitary alfalfa leafcutting bee (Megachile rotundata) in comparison with age-matched honey bees. We chose the cellular response to ER stress triggering the UPR because of the high conservation of this pathway. Analysis of the M. rotundata genome revealed that the core signaling components of the IRE1, PERK and ATF6 pathways are conserved (Table S1D). We tested a number of putative UPR target genes based on previous results from the honey bee (Adames et al., 2020; Johnston et al., 2016). We first confirmed that alfalfa leafcutting bees had robust survival in our caging system and found ∼80% survival of eclosed adults for 9 days, with most death occurring early post-eclosion (Fig. S2A). We then confirmed that bees were consuming sucrose solution at 7 days post-eclosion using Acid Blue 9 dye (Fig. S2B). Once conditions were established, we exposed cages of bees at 7 days post-eclosion to either sucrose solution containing DMSO vehicle control or sucrose solution containing 30 µmol l−1 tunicamycin. Many of the putative target genes we studied before (Adames et al., 2020; Johnston et al., 2016) [including genes from each of the following groups: chaperones (Hsc70-3, Gp93, p58ipk), UPR pathway components (Ire1) and disulfide bond generating enzymes (CaBP)] were upregulated significantly in the alfalfa leafcutting bee digestive tract after 48 h of 30 µmol l−1 tunicamycin treatment relative to bees consuming sucrose solution alone (Fig. 4A–E). We chose β-actin as a reference gene (Xu et al., 2021) and found that its levels were not altered by treatment (Fig. 4F). Two additional reference genes (Rps18 and Gapdh) were tested and their expression did not differ significantly by treatment (tunicamycin treated versus untreated, Fig. S3C,D) and similar patterns of induced expression were found when these other reference genes were used for normalization (data not shown).

Fig. 4.

Alfalfa leafcutting bee unfolded protein response (UPR) target genes are induced by tunicamycin. (A–E) Transcript levels of UPR target genes representing chaperones Hsc70-3 (A), Gp93 (B), p58ipk (C), UPR pathway component Ire1 (D) and the disulfide bond-generating enzyme Cabp1 (E) relative to β-actin in midgut tissue from individual control (DMSO vehicle, C; n=5) or 30 μmol l−1 tunicamycin-treated bees (T; n=5) after 48 h. The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the expression level of that gene relative to β-actin using the 2−ΔCT method. (F) Threshold cycle number for β-actin for these bees. Data are means±s.e.m. Statistical significance is indicated by asterisks (*P<0.05, **P<0.01).

Fig. 4.

Alfalfa leafcutting bee unfolded protein response (UPR) target genes are induced by tunicamycin. (A–E) Transcript levels of UPR target genes representing chaperones Hsc70-3 (A), Gp93 (B), p58ipk (C), UPR pathway component Ire1 (D) and the disulfide bond-generating enzyme Cabp1 (E) relative to β-actin in midgut tissue from individual control (DMSO vehicle, C; n=5) or 30 μmol l−1 tunicamycin-treated bees (T; n=5) after 48 h. The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the expression level of that gene relative to β-actin using the 2−ΔCT method. (F) Threshold cycle number for β-actin for these bees. Data are means±s.e.m. Statistical significance is indicated by asterisks (*P<0.05, **P<0.01).

Additionally, we tested similarly aged honey bees (starting 7 days post-eclosion) for induction of these target genes and found that chaperones (Hsc70-3, p58ipk, Gp93), UPR pathway components (Ire1) and disulfide bond generating enzymes (CaBP) were upregulated significantly in the honey bee digestive tract after 48 h of 30 µmol l−1 tunicamycin treatment relative to bees consuming sucrose solution alone (Fig. S3A–F) as reported previously for older bees (Adames et al., 2020; Johnston et al., 2016).

Xbp1 mRNA splicing occurs in response to tunicamycin treatment

Xbp1 mRNA splicing is a critical step that is often monitored for evidence of IRE1 pathway activation. Using the non-canonical intron structures predicted for the alfalfa leafcutting bee Xbp1 mRNA (Johnston et al., 2016) to design primers spanning the predicted intron, we were able to differentiate unspliced and spliced Xbp1 mRNA via PCR and gel analysis (Fig. 5A). Alfalfa leafcutting bees fed 30 µmol l−1 tunicamycin for 48 h had two abundant PCR products of ∼116 and ∼93 base pairs (bp), representing unspliced (Xbp1u) and spliced (Xbp1s) mRNA species, respectively, while bees fed sucrose solution containing vehicle had a single dominant PCR product of ∼116 bp at 48 h (band identities confirmed by purification and sequencing). In addition to splicing after tunicamycin treatment, total Xbp1 mRNA was upregulated significantly in the alfalfa leafcutting bee digestive tract after 48 h of tunicamycin treatment relative to bees consuming sugar syrup alone (Fig. 5B). To examine splicing more quantitatively, we designed a primer set that amplifies only the spliced version of the Xbp1 transcript (described in more detail for honey bees below). Using this primer set and the primer set that amplifies all Xbp1, we observed that while total Xbp1 transcript increased modestly after tunicamycin treatment, the ratio of spliced to total Xbp1 increased significantly after tunicamycin treatment (Fig. 5B–D).

Fig. 5.

Xbp1 mRNA splicing in alfalfa leafcutting bees is induced by tunicamycin. (A) Xbp1 mRNA splicing by IRE1 in midgut tissue from individual bees fed sucrose solution with (T; n=5) or without (DMSO vehicle control, C; n=5) 30 μmol l−1 tunicamycin for 48 h. Xbp1u, unspliced Xbp1; Xbp1s, spliced Xbp1. (B,C) Transcript levels of total Xbp1 (B) or spliced Xbp1 alone (C) relative to β-actin in midgut tissue from individual control (n=5) and 30 μmol l−1 tunicamycin-treated bees (n=5) after 48 h. (D) Ratio of spliced Xbp1 to total Xbp1 in midgut tissue from individual control (n=5) and 30 μmol l−1 tunicamycin-treated bees (n=5) after 48 h. The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the expression level of that gene relative to β-actin using the using the 2−ΔCT method. Data are means± s.e.m. Statistical significance is indicated by asterisks (**P<0.01).

Fig. 5.

Xbp1 mRNA splicing in alfalfa leafcutting bees is induced by tunicamycin. (A) Xbp1 mRNA splicing by IRE1 in midgut tissue from individual bees fed sucrose solution with (T; n=5) or without (DMSO vehicle control, C; n=5) 30 μmol l−1 tunicamycin for 48 h. Xbp1u, unspliced Xbp1; Xbp1s, spliced Xbp1. (B,C) Transcript levels of total Xbp1 (B) or spliced Xbp1 alone (C) relative to β-actin in midgut tissue from individual control (n=5) and 30 μmol l−1 tunicamycin-treated bees (n=5) after 48 h. (D) Ratio of spliced Xbp1 to total Xbp1 in midgut tissue from individual control (n=5) and 30 μmol l−1 tunicamycin-treated bees (n=5) after 48 h. The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the expression level of that gene relative to β-actin using the using the 2−ΔCT method. Data are means± s.e.m. Statistical significance is indicated by asterisks (**P<0.01).

We again tested similarly aged honey bees for Xbp1 splicing using conventional PCR and observed Xbp1 mRNA splicing after tunicamycin treatment (Fig. S4A) in accordance with our previous findings in older bees (Adames et al., 2020; Johnston et al., 2016). We then tested the honey bees that were 7 days post-eclosion using the Xbp1s primer set and the primer set that amplifies total Xbp1 and observed that while total Xbp1 transcript increased modestly after tunicamycin treatment, the ratio of spliced to total increased substantially after tunicamycin treatment (Fig. S4B–D). Using the primer set that amplifies only the spliced version of the Xbp1 transcript, we observed that the amount of Xbp1s increased dramatically in a dose-dependent manner after tunicamycin treatment in a way similar to the amount of the canonical target gene Hsc70-3 (Fig. S4E,F).

UPR pathway induces select l(2)efl genes in the alfalfa leafcutting bee

We previously found that members of the l(2)efl family of genes are induced by diverse stressors in honey bees (Shih et al., 2021). We showed that two of these (410087a and 724367) could be useful as stress biomarker genes in honey bees responding to diverse types of stress (Cho et al., 2022; Shih et al., 2021). Additionally, we found that M. rotundata and Osmia lignaria, which are both in the Megachilidae family, share shsp gene number and structure with A. mellifera (Shih et al., 2021). We tested whether UPR activation induced the transcription of the homologs of two of the l(2)efl genes shown to be UPR targets in honey bees. We found that only one of the l(2)efl genes (100881126 corresponding to 410087a in the honey bee) was upregulated significantly in the alfalfa leafcutting bee digestive tract after 48 h of 30 µmol l−1 tunicamycin treatment relative to bees consuming sucrose solution alone (Fig. 6A,B). Expression of the l(2)efl gene 100876534 (corresponding to 724367 in the honey bee) was not increased by tunicamycin treatment. Interestingly, when we tested similarly aged honey bees, we found both of the tested l(2)efl genes (410087a and 724367) were upregulated significantly in the honey bee digestive tract after 48 h of 30 µmol l−1 tunicamycin treatment relative to bees consuming sucrose solution alone (Fig. 6C,D).

Fig. 6.

Select alfalfa leafcutting bee and honey bee l(2)efl genes are induced by tunicamycin. (A,B) Transcript levels of the l(2)efl genes 100881126 (A) and 100876534 (B) relative to β-actin in midgut tissue from individual control (DMSO vehicle, C; n=5) or 30 μmol l−1 tunicamycin-treated (T; n=5) alfalfa leafcutting bees after 48 h. (C,D) Transcript levels of the l(2)efl genes 410087a (C) and 724367 (D) relative to β-actin in midgut tissue from adult honey bees captured at the landing board and fed sucrose solution with (T; n=10) or without (DMSO vehicle control, C; n=10) 30 μmol l−1 tunicamycin for 48 h. Data (means±s.e.m.) represent expression values of the genes of interest calculated using the 2−ΔCT method for individual bees. Statistical significance is indicated by asterisks (**P<0.01).

Fig. 6.

Select alfalfa leafcutting bee and honey bee l(2)efl genes are induced by tunicamycin. (A,B) Transcript levels of the l(2)efl genes 100881126 (A) and 100876534 (B) relative to β-actin in midgut tissue from individual control (DMSO vehicle, C; n=5) or 30 μmol l−1 tunicamycin-treated (T; n=5) alfalfa leafcutting bees after 48 h. (C,D) Transcript levels of the l(2)efl genes 410087a (C) and 724367 (D) relative to β-actin in midgut tissue from adult honey bees captured at the landing board and fed sucrose solution with (T; n=10) or without (DMSO vehicle control, C; n=10) 30 μmol l−1 tunicamycin for 48 h. Data (means±s.e.m.) represent expression values of the genes of interest calculated using the 2−ΔCT method for individual bees. Statistical significance is indicated by asterisks (**P<0.01).

ER stress activates the Gnmt gene in alfalfa leafcutting bees

To determine whether Gnmt was upregulated in response to stress in the midgut of the distantly related and solitary alfalfa leafcutting bee species, we measured the expression of this gene after tunicamycin treatment of females 7 days post-eclosion. We found that expression of Gnmt was increased in response to tunicamycin treatment in alfalfa leafcutting bees (Fig. 7A). To confirm that these genes can be induced in honey bees following this protocol, we also examined midguts of age-matched honey bees (7 days post-eclosion) after 30 μmol l−1 tunicamycin treatment for 48 h in age-matched honey bees and observed induction of Gnmt in response to tunicamycin in these bees as well (Fig. 7B).

Fig. 7.

Alfalfa leafcutting bee and honey bee Gnmt is induced by tunicamycin. Transcript levels of Gnmt relative to β-actin in midgut tissue from individual control (DMSO vehicle, C) or 30 μmol l−1 tunicamycin (T)-treated alfalfa leafcutting bees (n=5 control, n=5 tunicamycin; A) and honey bees (n=10 control, n=10 tunicamycin; B) after 48 h. The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the expression level of that gene relative to β-actin using the 2−ΔCT method. Data are means±s.e.m. Statistical significance is indicated by asterisks (**P<0.01).

Fig. 7.

Alfalfa leafcutting bee and honey bee Gnmt is induced by tunicamycin. Transcript levels of Gnmt relative to β-actin in midgut tissue from individual control (DMSO vehicle, C) or 30 μmol l−1 tunicamycin (T)-treated alfalfa leafcutting bees (n=5 control, n=5 tunicamycin; A) and honey bees (n=10 control, n=10 tunicamycin; B) after 48 h. The difference between the threshold cycle number for β-actin and that of the gene of interest was used to calculate the expression level of that gene relative to β-actin using the 2−ΔCT method. Data are means±s.e.m. Statistical significance is indicated by asterisks (**P<0.01).

Gnmt expression may be regulated by FOXO and ATF4 in response to stress

Based on the diversity of stresses able to promote gene induction in honey bees, we were interested in exploring the potential trans factors and cis elements responsible for regulating increased gene expression. In both flies and mice, Gnmt is increased after a reduction in IIS signaling (Tain et al., 2020), likely mediated by the forkhead box (FOXO) transcription factor. We observed a number of consensus sites for FOXO binding (TKTTYACY; Bai et al., 2013) near the Gnmt transcriptional start site (TSS) in both honey bees and alfalfa leafcutting bees (Fig. S5).

FOXO is the key transcription factor involved in the IIS pathway responding to both nutrient levels and stress (Semaniuk et al., 2021a,b). We also looked for binding sites for other transcription factors associated with the stress pathways we found to induce Gnmt expression. We found one putative ATF4-binding CREB response element binding site (defined as either TGACGT or TTKCATCAK) downstream of the Gnmt TSS in both species. The transcriptional regulator controlling the oxidative stress response, NRF2 (and its homologs), is responsible for promoting resistance to oxidative stress in C. elegans, D. melanogaster and mammals (reviewed in Sykiotis and Bohmann, 2008). NRF2-responsive genes possess the consensus binding site TGAYNNNGC, which makes up the core of the antioxidant responsive element (ARE). We found evidence of a single ARE in the honey bee and alfalfa leafcutting bee Gnmt promoters and an additional ARE site downstream of the TSS in alfalfa leafcutting bees (Fig. S5). When examining the mRNA for regulatory features, we found that the predicted transcripts of both the honey bee and alfalfa leafcutting bee Gnmt genes possess significant 5′ and 3′ UTR sequences with the 5′ UTR possessing an upstream open reading frame (uORF; Fig. S5).

GNMT as a stress-inducible metabolic switch regulating cellular methylation

Here, we report the identification of Gnmt as a novel stress-inducible gene upregulated by diverse stressors using unbiased approaches in honey bees. We found that Gnmt expression is induced by both abiotic and biotic stressors. In bees that were infected with V. ceranae, a microsporidia species that causes individual disease in honey bees and can contribute to colony collapse (Snow, 2022), Gnmt expression was increased after 9 days of infection when compared with similarly aged bees that were not infected. Vairimorpha ceranae is documented to cause many changes to infected epithelial cells at the biochemical and cell biological level (reviewed in Snow, 2022) and defining whether infection-induced induction of Gnmt is beneficial or detrimental in response to infection will be important. sHSPs represent another class of potential biomarkers identified by our lab using this same approach (Shih et al., 2021) that have now been shown by multiple groups to be induced by abiotic stress (Cho et al., 2022; Goblirsch and Adamczyk, 2022; McAfee et al., 2020a,b; Parrella et al., 2024; Shih et al., 2021) and biotic stress as they comprise part of the antiviral response in multiple bee species (Brutscher et al., 2017; McMenamin et al., 2018). We extend this finding to show that Gnmt expression is also triggered by stress in alfalfa leafcutting bees. GNMT plays a key role in the methionine cycle (Sanderson et al., 2019), in which it is thought to be the main regulatory switch controlling the SAM/SAH ratio. Evidence for the key role of GNMT in this cycle is found in diverse organisms. In humans, deficiency in the Gnmt gene causes hypermethioninemia with increased levels of SAM (Mudd et al., 2001). Mice lacking Gnmt have elevated methionine levels, a dramatically increased SAM to SAH ratio, and multiple other metabolic abnormalities (Hughey et al., 2018; Luka et al., 2006). Fruit flies deficient in Gnmt also have an increased SAM/SAH ratio (Obata and Miura, 2015; Obata et al., 2014; Tain et al., 2020) and increased sarcosine is one of the most pronounced observed metabolic changes (Obata and Miura, 2015; Obata et al., 2014). In accordance with the results from Drosophila, we also found an increase in sarcosine levels. These results show that by at least one metric, our observation of stress-induced Gnmt expression results in the expected biochemical changes.

The current thought is that effects of methionine and its cycle may primarily impact transmethylation, which refers to the myriad of methylation reactions performed by diverse enzymes using SAM as a methyl donor. In fact, most methylation in cells takes place by SAM-dependent methyltransferases. These methyltransferases catalyze the often reversible methylation reactions of DNA, RNA, proteins and small molecules (Abdelraheem et al., 2022). Some key examples include the methylation of DNA which directs epigenetic regulation of gene expression, the methylation of RNA which regulates many of the processes involved in RNA handling (e.g. splicing and protein synthesis), the methylation of proteins to alter their function (e.g. histone methylation to regulate gene expression) and the methylation of diverse small molecules in the service of metabolism or to alter a molecule's chemical properties. Methylation of RNA molecules is one type of methylation reaction that is highly conserved, SAM dependent and involved in regulating many aspects of RNA biology (Zhang et al., 2021). Interestingly, the N6-methyladenosine (m6A) methylation of RNA has been implicated cellular stress in many contexts including cellular stress (Wilkinson et al., 2021) and has recently been shown to be involved in honey bee biology (Wang et al., 2021). Specifically, differential m6A methylation was shown to post-transcriptionally regulate gene expression in honey bee larval development and caste differentiation (Wang et al., 2021). Further work will be necessary to disentangle potential changes in the methylation of the molecules above resulting from stress-induced increases in Gnmt expression and activity in bees.

GNMT’s role in the regulation of methionine pathway metabolism is thought to be significant in terms of governing these SAM-dependent transmethylation reactions. By using SAM as the methyl donor for conversion of glycine to sarcosine, Gnmt activity reduces SAM levels in favor of SAH. As SAM is the primary methyl donor for a whole host of cellular methylation reactions (Parkhitko et al., 2020), a concomitant reduction in transmethylation would be expected after Gnmt induction. Although changes in cellular methylation reactions are likely a primary consequence of Gnmt upregulation and enzymatic conversion of SAM and glycine to sarcosine and SAH, it is important to note that the methionine cycle (Sanderson et al., 2019) is directly linked to a number of other metabolic processes which might also be affected. These pathways include the transsulfuration pathway/gamma-glutamyl cycle (Bonifácio et al., 2021), the one carbon metabolism/folate cycle (Ducker and Rabinowitz, 2017) and polyamine biosynthesis (Casero et al., 2018). Gnmt overexpression may play key roles in the regulation of these aspects of cell biology as well, and future work will be required to define the consequences of increased Gnmt expression on cell biology in bees.

In the fly, GNMT-mediated regulation of SAM levels is driven by genetic lesions that induce tissue damage (Kashio et al., 2016; Obata et al., 2014). Dietary methionine acting via SAM levels has also been implicated in regulating protein synthesis and tissue regeneration in the fly (Obata et al., 2018). Prior to our study, Gnmt upregulation had not previously been associated with cellular stress responses. It is likely that damage occurring via genetic means or caused by the triggers of cellular stress responses used here (such as thermal stress, ER stress, etc.) could result in the activation of a non-specific stress pathway with shared target genes such as Gnmt. One likely candidate is the IIS pathway. Multiple lines of evidence suggest that Gnmt levels are controlled by insulin in flies (Obata and Miura, 2015; Obata et al., 2014; Tain et al., 2020) and the IIS pathway may be activated during these model stress responses in bees. This link with the IIS pathway fits with other data showing that Gnmt expression increases with age and that its overexpression increases lifespan (in flies: Obata and Miura, 2015; Tain et al., 2020). Other genetic and chemical perturbations have also implicated the methionine pathway as being a critical mediator of lifespan extension (Parkhitko et al., 2019, 2020). For example, S-adenosyl-l-homocysteine hydrolase (SAHH), which converts SAH to adenosine and homocysteine, as mediated by inactivation of its negative regulator, also results in increased lifespan (Parkhitko et al., 2016). In the fly, the fat body is thought to be the primary organ for GNMT-mediated control of systemic SAM levels (Obata and Miura, 2015; Obata et al., 2014; Tain et al., 2020). Similarly, in mammals, Gnmt regulates physiological SAM levels through action in the liver (Luka et al., 2009), which is often considered functionally analogous to the insect fat body. We observed increased Gnmt expression and metabolite differences in the digestive tract. While there are many potential reasons for this difference, the most likely appears to be that the perturbations used in our system cause a tissue-specific response in the midgut rather than a systemic alteration of SAM levels. Supporting this idea, most of our perturbations were localized to the midgut via drug delivery. SAM levels have already been shown to have a special role in regulating stem cell proliferation in the fly midgut (Obata et al., 2018). Although this effect appeared not to be dependent on GNMT, it might be expected that proteotoxic stressors shown here might influence SAM levels via a different mechanism from the DNA damage perturbation used (Obata et al., 2018). To interpret these differences, the use of experiments designed to allow for more direct comparison of the two models is warranted.

Shared stress responses in diverse bee species

Our study supports the idea that there is considerable overlap in the stress responses of disparate bee species. Furthermore, our results confirm that honey bees can provide a useful model for learning about the mechanism of stress resistance in other bee species. In addition to finding Gnmt as a stress-induced target gene in both bee species, we also observed significant conservation in the function of the canonical UPR in the two different bee species with disparate life histories. Based on the high conservation of UPR architecture and apparent function in honey bees as compared with other insects (Adames et al., 2020; Johnston et al., 2016), it is to be expected that such a conserved pathway would be quite similar in another bee species. The alfalfa leafcutting bee possesses the same unusual sequence at the 5′ end of the Xbp1 mRNA non-canonical intron stem-loop recognized by IRE1 as other bee species (Johnston et al., 2016). Despite the validation of honey bees as a model for other bees shown here, previously published work suggests that studies using honey bees cannot take the place of research on other bee species (Wood et al., 2020), and our study provides some support for this idea as well. Specifically, we observed differences in UPR induction of select l(2)efl genes between the two species. The expression levels of these sHSP (at both the transcript and protein levels) represent candidate biomarkers of stress in honey bees (Cho et al., 2022; McAfee et al., 2020a,b; Shih et al., 2021). With some exceptions, our previous phylogenetic analysis of the sHSP proteins in other bee species shows a high degree of conservation of this family among bees with different life histories. Critical for the current work, we observed that M. rotundata and O. lignaria, which are both in the Megachilidae family, share shsp gene number and structure with A. mellifera (Shih et al., 2021). Thus, the expectation would be that these genes would be induced by stressors similarly in these two bee species. However, contrary to our expectations, we found that only one of the two examined l(2)efl genes was induced in alfalfa leafcutting bees despite both being strongly induced in honey bees of the same age. It will be of interest to explore potential causes and consequences of such differential responses. While only relatively subtle differences were found using the candidate gene approach described here, it seems reasonable to speculate that a more unbiased approach, such as RNASeq, might yield more differences of obvious biological significance in the stress responses of these two species. However, more importantly, these data suggest that there are real differences in cellular responses between bee species. While honey bees can provide important insight into overall bee biology, studies on diverse bee species (Wood et al., 2020) will be required to fully understand global declines in bee populations (Zattara and Aizen, 2021).

In addition to the importance of understanding how honey bees are affected by and respond to various stressors, there is great interest in understanding the factors that influence the health of non-Apis bee species (López-Uribe et al., 2019). This is especially true for solitary bee species, which account for over 90% of all bee species (Danforth et al., 2019). Our previous work supported the findings of other groups that expression of l(2)efl genes could represent ideal biomarkers of stress in honey bees (McAfee et al., 2020a,b; Shih et al., 2021). This study suggests that one of these genes (100881126, homologous to honey bee 410087 V. ceranae) is induced at a very high level in response to at least one stress. Thus, while not currently validated as a valuable indicator of exposure to stressors naturally encountered by alfalfa leafcutting bees (or honey bees), our results suggest that sHSP may provide potential biomarkers of stress in some non-Apis bee species as well as honey bees.

Native to Eurasia, alfalfa leafcutting bees were originally imported into the USA for alfalfa crop pollination. Alfalfa pollination by alfalfa leafcutting bees is critical for alfalfa seed production in the 4.6 billion USD per year alfalfa hay industry (Pitts-Singer and Cane, 2011). Alfalfa leafcutting bees are also important pollinators in non-agricultural ecosystems and have been facing high levels of mortality due to factors of diverse origin (Potts et al., 2016). In the USA in particular, alfalfa leafcutting bee production is so depressed that a substantial proportion of bees required for pollination must be imported each year (James and Pitts-Singer, 2013). These stressors, all of which are implicated in the loss of other bee species, include nutritional stress due to loss of appropriate forage, chemical poisoning from pesticides, changes to normal living conditions brought about through intensive management practices, and infection by arthropod parasites and pathogenic microbes (Goulson et al., 2015). Yet, despite being the second most intensively managed pollinator in the USA, our understanding of cellular stress in this bee species is incomplete. Thus, establishment of experimental systems to allow for empirical studies of diverse bee species is highly desirable. Prior experiments in alfalfa leafcutting bees have largely focused on the impacts of stressors on larval or pupal development (e.g. Cambron et al., 2021; Cambron-Kopco et al., 2022; Melicher et al., 2019; Torson et al., 2017, 2019; Yocum et al., 2018) and very few studies have been undertaken to perform experiments on adult bees using either lab-reared (Bennett et al., 2013, 2015; Cervantes and López-Martínez, 2022; Hayes and López-Martínez, 2021; James et al., 2012) or wild-captured bees (Hamblin et al., 2017). Similarly, while important studies examining stress responses at the molecular level in alfalfa leafcutting bees have been performed (Xu and James, 2012; Yocum et al., 2005, 2018), work to understand stress at the cell and molecular level in adult alfalfa leafcutting bees is still in its nascent stages, in part because of the challenges of working with these and other solitary bee species. By characterizing the UPR in alfalfa leafcutting bees in comparison with the UPR in honey bees, we advance the alfalfa leafcutting bee species as a model for cellular stress responses in solitary bees and establish a system to explore differences in cellular stress responses between bee species to provide insight into the mechanisms of bee health declines.

Conclusions

In conclusion, we show that a key signal transduction event and a number of gene targets of the UPR are robustly induced upon ER stress in alfalfa leafcutting bees. We also provide evidence that honey bees can represent a model for better understanding cellular stress in alfalfa leafcutting bees. Discovery of Gnmt as a stress-activated target gene in honey bees accurately predicts it as a stress-inducible gene in the solitary bee species. The consequences of Gnmt induction by stress in these two bee species is likely related to its role as a key metabolic switch controlling cellular methylation reactions. Further work will be necessary to fully explore the potential changes in cellular methylation induced by stress in bees and to define whether these are beneficial or detrimental in the face of abiotic and biotic stressors. This study also suggests that the honey bee can serve as an important model for bee biology more broadly, although studies on diverse bee species will be required to understand global declines in bee populations. In particular, additional studies using more unbiased approaches with a greater number and diversity of bee species will be needed to fully understand the similarities and differences in the stress responses of different bee species.

The authors acknowledge the technical assistance of the Barnard College Cell Biology Laboratory classes (2019–2021) in the completion of preliminary experiments. In particular, we thank the Barnard College Summer Research Institute for support of undergraduate research students. We thank Brian Breczinski for supervision of the Bucknell regional NMR center.

Author contributions

Conceptualization: S.G.M., J.W.S.; Methodology: H.V.K., S.G.M., N.C.R., J.R.-C., D.R., M.E.F., J.W.S.; Validation: H.V.K., S.G.M., N.C.R., M.E.F., J.W.S.; Formal analysis: H.V.K., S.G.M., N.C.R., J.R.-C., V.A.C., D.R., C.A.M., M.M., M.E.F., J.W.S.; Investigation: H.V.K., S.G.M., N.C.R., J.R.-C., V.A.C., D.R., C.A.M., M.M., M.E.F., J.W.S.; Resources: N.C.R., J.W.S.; Data curation: H.V.K., S.G.M., N.C.R., J.R.-C., V.A.C., D.R., C.A.M., M.M., M.E.F., J.W.S.; Writing - original draft: H.V.K., S.G.M., M.E.F., J.W.S.; Writing - review & editing: H.V.K., S.G.M., N.C.R., J.R.-C., V.A.C., D.R., C.A.M., M.M., M.E.F., J.W.S.; Visualization: H.V.K., S.G.M., N.C.R., J.R.-C., V.A.C., D.R., C.A.M., M.M., M.E.F., J.W.S.; Supervision: D.R., J.W.S.; Project administration: J.W.S.; Funding acquisition: J.W.S.

Funding

This research was paid for by funds from Barnard College and Columbia University. Open access funding provided by Columbia University. Deposited in PMC for immediate release.

Data availability

All relevant data can be found within the article and its supplementary information.

Abdelraheem
,
E.
,
Thair
,
B.
,
Varela
,
R. F.
,
Jockmann
,
E.
,
Popadić
,
D.
,
Hailes
,
H. C.
,
Ward
,
J. M.
,
Iribarren
,
A. M.
,
Lewkowicz
,
E. S.
,
Andexer
,
J. N.
et al. 
(
2022
).
Methyltransferases: functions and applications
.
Chembiochem
23
,
e202200212
.
Adames
,
T. R.
,
Rondeau
,
N. C.
,
Kabir
,
M. T.
,
Johnston
,
B. A.
,
Truong
,
H.
and
Snow
,
J. W.
(
2020
).
The IRE1 pathway regulates honey bee Unfolded Protein Response gene expression
.
Insect Biochem. Mol. Biol.
121
,
103368
.
Bach
,
D. M.
,
Holzman
,
M. A.
,
Wague
,
F.
,
Miranda
,
J. J. L.
,
Lopatkin
,
A. J.
,
Mansfield
,
J. H.
and
Snow
,
J. W.
(
2021
).
Thermal stress induces tissue damage and a broad shift in regenerative signaling pathways in the honey bee digestive tract
.
J. Exp. Biol.
224
,
jeb242262
.
Bai
,
H.
,
Kang
,
P.
,
Hernandez
,
A. M.
and
Tatar
,
M.
(
2013
).
Activin signaling targeted by insulin/dFOXO regulates aging and muscle proteostasis in Drosophila
.
PLoS Genet.
9
,
e1003941
-
e16
.
Bennett
,
M. M.
,
Petersen
,
K.
,
Yocum
,
G.
,
Rinehart
,
J.
,
Kemp
,
W.
and
Greenlee
,
K. J.
(
2013
).
Effects of extended prepupal storage duration on adult flight physiology of the alfalfa leafcutting bee (Hymenoptera: Megachilidae)
.
J. Econ. Entomol.
106
,
1089
-
1097
.
Bennett
,
M. M.
,
Cook
,
K. M.
,
Rinehart
,
J. P.
,
Yocum
,
G. D.
,
Kemp
,
W. P.
and
Greenlee
,
K. J.
(
2015
).
Exposure to suboptimal temperatures during metamorphosis reveals a critical developmental window in the solitary bee, Megachile rotundata
.
Physiol. Biochem. Zool.
88
,
508
-
520
.
Bonifácio
,
V. D. B.
,
Pereira
,
S. A.
,
Serpa
,
J.
and
Vicente
,
J. B.
(
2021
).
Cysteine metabolic circuitries: druggable targets in cancer
.
Br. J. Cancer
124
,
862
-
879
.
Brutscher
,
L. M.
,
Daughenbaugh
,
K. F.
and
Flenniken
,
M. L.
(
2017
).
Virus and dsRNA-triggered transcriptional responses reveal key components of honey bee antiviral defense
.
Sci. Rep.
7
,
6448
.
Cambron
,
L. D.
,
Yocum
,
G. D.
,
Yeater
,
K. M.
and
Greenlee
,
K. J.
(
2021
).
Overwintering conditions impact insulin pathway gene expression in diapausing Megachile rotundata
.
Comp. Biochem. Physiol. Part Mol. Integr. Physiol.
256
,
110937
.
Cambron-Kopco
,
L. D.
,
Yocum
,
G. D.
,
Yeater
,
K. M.
and
Greenlee
,
K. J.
(
2022
).
Timing of diapause initiation and overwintering conditions alter gene expression profiles in Megachile rotundata
.
Front. Physiol.
13
,
844820
.
Casero
,
R. A.
,
Stewart
,
T. M.
and
Pegg
,
A. E.
(
2018
).
Polyamine metabolism and cancer: treatments, challenges and opportunities
.
Nat. Rev. Cancer
18
,
681
-
695
.
Cervantes
,
L.
and
López-Martínez
,
G.
(
2022
).
Anoxia hormesis following overwintering diapause boosts bee survivorship and adult performance
.
Sci. Total Environ.
802
,
149934
.
Cho
,
R. M.
,
Kogan
,
H. V.
,
Elikan
,
A. B.
and
Snow
,
J. W.
(
2022
).
Paromomycin reduces Vairimorpha (Nosema) ceranae infection in honey bees but perturbs microbiome levels and midgut cell function
.
Microorganisms
10
,
1107
.
Chow
,
C. Y.
,
Wolfner
,
M. F.
and
Clark
,
A. G.
(
2013
).
Using natural variation in Drosophila to discover previously unknown endoplasmic reticulum stress genes
.
Proc. Natl. Acad. Sci. USA
110
,
9013
-
9018
.
Danforth
,
B. N.
,
Minckley
,
R. L.
,
Neff
,
J. L.
and
Fawcett
,
F.
(
2019
).
The Solitary Bees
.
Princeton University Press
.
Deliu
,
L. P.
,
Ghosh
,
A.
and
Grewal
,
S. S.
(
2017
).
Investigation of protein synthesis in Drosophila larvae using puromycin labelling
.
Biol. Open
6
,
1229
-
1234
.
Ducker
,
G. S.
and
Rabinowitz
,
J. D.
(
2017
).
One-carbon metabolism in health and disease
.
CMET
25
,
27
-
42
.
Edgecomb
,
R. S.
,
Harth
,
C. E.
and
Schneiderman
,
A. M.
(
1994
).
Regulation of feeding behavior in adult Drosophila melanogaster varies with feeding regime and nutritional state
.
J. Exp. Biol.
197
,
215
-
235
.
Flores
,
M. E.
,
McNamara-Bordewick
,
N. K.
,
Lovinger
,
N. L.
and
Snow
,
J. W.
(
2021
).
Halofuginone triggers a transcriptional program centered on ribosome biogenesis and function in honey bees
.
Insect Biochem. Mol.
139
,
103667
.
Goblirsch
,
M.
and
Adamczyk
,
J. J.
(
2022
).
Using the honey bee (Apis mellifera L.) cell line, AmE-711, to evaluate insecticide toxicity
.
Environ. Toxicol. Chem.
42
,
88
-
99
.
Goulson
,
D.
,
Nicholls
,
E.
,
Botías
,
C.
and
Rotheray
,
E. L.
(
2015
).
Bee declines driven by combined stress from parasites, pesticides, and lack of flowers
.
Science
347
,
1255957
.
Grozinger
,
C. M.
and
Zayed
,
A.
(
2020
).
Improving bee health through genomics
.
Nat. Rev. Genet.
21
,
277
-
291
.
Hamblin
,
A. L.
,
Youngsteadt
,
E.
,
López-Uribe
,
M. M.
and
Frank
,
S. D.
(
2017
).
Physiological thermal limits predict differential responses of bees to urban heat-island effects
.
Biol. Lett.
13
,
20170125
-
4
.
Hayes
,
T.
and
López-Martínez
,
G.
(
2021
).
Resistance and survival to extreme heat shows circadian and sex-specific patterns in A cavity nesting bee
.
Curr. Res. Insect. Sci.
1
,
100020
.
Hedtke
,
S. M.
,
Patiny
,
S.
and
Danforth
,
B. N.
(
2013
).
The bee tree of life: a supermatrix approach to apoid phylogeny and biogeography
.
BMC Evol. Biol.
13
,
138
-
113
.
Hristov
,
P.
,
Shumkova
,
R.
,
Palova
,
N.
and
Neov
,
B.
(
2020
).
Factors associated with honey bee colony losses: a mini-review
.
Vet. Sci.
7
,
166
.
Hughey
,
C. C.
,
Trefts
,
E.
,
Bracy
,
D. P.
,
James
,
F. D.
,
Donahue
,
E. P.
and
Wasserman
,
D. H.
(
2018
).
Glycine N-methyltransferase deletion in mice diverts carbon flux from gluconeogenesis to pathways that utilize excess methionine cycle intermediates
.
J. Biol. Chem.
293
,
11944
-
11954
.
James
,
R. R.
and
Pitts-Singer
,
T. L.
(
2013
).
Health status of Alfalfa leafcutting bee larvae (Hymenoptera: Megachilidae) in United States Alfalfa Seed Fields
.
Environ. Entomol.
42
,
1166
-
1173
.
James
,
R. R.
,
McGuire
,
M. R.
and
Leland
,
J. E.
(
2012
).
Susceptibility of adult alfalfa leafcutting bees and honey bees to a microbial control agent, Beauveria bassiana
.
Southwest. Entomol.
37
,
13
-
21
.
Johnston
,
B. A.
,
Hooks
,
K. B.
,
McKinstry
,
M.
and
Snow
,
J. W.
(
2016
).
Divergent forms of endoplasmic reticulum stress trigger a robust unfolded protein response in honey bees
.
J. Insect Physiol.
86
,
1
-
10
.
Kashio
,
S.
,
Obata
,
F.
,
Zhang
,
L.
,
Katsuyama
,
T.
,
Chihara
,
T.
and
Miura
,
M.
(
2016
).
Tissue nonautonomous effects of fat body methionine metabolism on imaginal disc repair in Drosophila
.
Proc. Natl. Acad. Sci. USA
113
,
1835
-
1840
.
López-Uribe
,
M. M.
,
Ricigliano
,
V. A.
and
Simone-Finstrom
,
M.
(
2019
).
Defining pollinator health: assessing bee ecological, genetic, and physiological factors at the individual, colony, and population levels
.
Annu. Rev. Anim. Biosci.
8
,
annurev-animal-020518-115045-26
.
Luka
,
Z.
,
Capdevila
,
A.
,
Mato
,
J. M.
and
Wagner
,
C.
(
2006
).
A glycine N-methyltransferase knockout mouse model for humans with deficiency of this enzyme
.
Transgenic Res.
15
,
393
-
397
.
Luka
,
Z.
,
Mudd
,
S. H.
and
Wagner
,
C.
(
2009
).
Glycine N-methyltransferase and regulation of S-adenosylmethionine levels
.
J. Biol. Chem.
284
,
22507
-
22511
.
Lundgren
,
J.
,
Masson
,
P.
,
Mirzaei
,
Z.
and
Young
,
P.
(
2005
).
Identification and characterization of a Drosophila proteasome regulatory network
.
Mol. Cell. Biol.
25
,
4662
-
4675
.
McAfee
,
A.
,
Chapman
,
A.
,
Higo
,
H.
,
Underwood
,
R.
,
Milone
,
J.
,
Foster
,
L. J.
,
Guarna
,
M. M.
,
Tarpy
,
D. R.
and
Pettis
,
J. S.
(
2020a
).
Vulnerability of honey bee queens to heat-induced loss of fertility
.
Nat. Sustain.
3
,
367
-
376
.
McAfee
,
A.
,
Milone
,
J.
,
Chapman
,
A.
,
Foster
,
L. J.
,
Pettis
,
J. S.
and
Tarpy
,
D. R.
(
2020b
).
Candidate stress biomarkers for queen failure diagnostics
.
BMC Genomics
21
,
571
.
McDevitt
,
J. C.
,
Gupta
,
R. A.
,
Dickinson
,
S. G.
,
Martin
,
P. L.
,
Rieuthavorn
,
J.
,
Freund
,
A.
,
Pizzorno
,
M. C.
,
Capaldi
,
E. A.
and
Rovnyak
,
D.
(
2021
).
Methodology for single bee and bee brain 1H-NMR metabolomics
.
Metabolites
11
,
864
.
McKinstry
,
M.
,
Chung
,
C.
,
Truong
,
H.
,
Johnston
,
B. A.
and
Snow
,
J. W.
(
2017
).
The heat shock response and humoral immune response are mutually antagonistic in honey bees
.
Sci. Rep.
7
,
8850
.
McMenamin
,
A.
,
Daughenbaugh
,
K.
,
Parekh
,
F.
,
Pizzorno
,
M.
and
Flenniken
,
M.
(
2018
).
Honey bee and bumble bee antiviral defense
.
Viruses
10
,
395
.
Melicher
,
D.
,
Torson
,
A. S.
,
Anderson
,
T. J.
,
Yocum
,
G. D.
,
Rinehart
,
J. P.
and
Bowsher
,
J. H.
(
2019
).
Immediate transcriptional response to a temperature pulse under a fluctuating thermal regime
.
Integr. Comp. Biol.
59
,
320
-
337
.
Mudd
,
S. H.
,
Cerone
,
R.
,
Schiaffino
,
M. C.
,
Fantasia
,
A. R.
,
Minniti
,
G.
,
Caruso
,
U.
,
Lorini
,
R.
,
Watkins
,
D.
,
Matiaszuk
,
N.
,
Rosenblatt
,
D. S.
et al. 
(
2001
).
Glycine N-methyltransferase deficiency: a novel inborn error causing persistent isolated hypermethioninaemia
.
J. Inherit. Metab. Dis.
24
,
448
-
464
.
Obata
,
F.
and
Miura
,
M.
(
2015
).
Enhancing S-adenosyl-methionine catabolism extends Drosophila lifespan
.
Nat. Commun.
6
,
1
-
9
.
Obata
,
F.
,
Kuranaga
,
E.
,
Tomioka
,
K.
,
Ming
,
M.
,
Takeishi
,
A.
,
Chen
,
C.-H.
,
Soga
,
T.
and
Miura
,
M.
(
2014
).
Necrosis-driven systemic immune response alters SAM metabolism through the FOXO-GNMT axis
.
Cell Rep.
7
,
821
-
833
.
Obata
,
F.
,
Tsuda-Sakurai
,
K.
,
Yamazaki
,
T.
,
Nishio
,
R.
,
Nishimura
,
K.
,
Kimura
,
M.
,
Funakoshi
,
M.
and
Miura
,
M.
(
2018
).
Nutritional control of stem cell division through S-adenosylmethionine in Drosophila intestine
.
Dev. Cell
44
,
741
-
751.e3
.
Parkhitko
,
A. A.
,
Binari
,
R.
,
Zhang
,
N.
,
Asara
,
J. M.
,
Demontis
,
F.
and
Perrimon
,
N.
(
2016
).
Tissue-specific down-regulation of S-adenosyl-homocysteine via suppression of dAhcyL1/dAhcyL2 extends health span and life span in Drosophila
.
Genes Dev.
30
,
1409
-
1422
.
Parkhitko
,
A. A.
,
Jouandin
,
P.
,
Mohr
,
S. E.
and
Perrimon
,
N.
(
2019
).
Methionine metabolism and methyltransferases in the regulation of aging and lifespan extension across species
.
Aging Cell
18
,
179
-
118
.
Parkhitko
,
A. A.
,
Filine
,
E.
,
Mohr
,
S. E.
,
Moskalev
,
A.
and
Perrimon
,
N.
(
2020
).
Targeting metabolic pathways for extension of lifespan and healthspan across multiple species
.
Ageing Res. Rev.
64
,
101188
.
Parrella
,
P.
,
Elikan
,
A. B.
,
Kogan
,
H. V.
,
Wague
,
F.
,
Marshalleck
,
C. A.
and
Snow
,
J. W.
(
2024
).
Bleomycin reduces Vairimorpha (Nosema) ceranae infection in honey bees with some evident host toxicity
.
Microbiol. Spectr.
12
,
e0334923
.
Pitts-Singer
,
T. L.
and
Cane
,
J. H.
(
2011
).
The alfalfa leafcutting bee, Megachile rotundata: the world's most intensively managed solitary bee
.
Annu. Rev. Entomol.
56
,
221
-
237
.
Pizzorno
,
M. C.
,
Field
,
K.
,
Kobokovich
,
A. L.
,
Martin
,
P. L.
,
Gupta
,
R. A.
,
Mammone
,
R.
,
Rovnyak
,
D.
and
Capaldi
,
E. A.
(
2021
).
Transcriptomic responses of the honey bee brain to infection with deformed wing virus
.
Viruses
13
,
287
.
Potts
,
S. G.
,
Imperatriz-Fonseca
,
V.
,
Ngo
,
H. T.
,
Aizen
,
M. A.
,
Biesmeijer
,
J. C.
,
Breeze
,
T. D.
,
Dicks
,
L. V.
,
Garibaldi
,
L. A.
,
Hill
,
R.
,
Settele
,
J.
et al. 
(
2016
).
Safeguarding pollinators and their values to human well-being
.
Nature
540
,
220
-
229
.
Powers
,
E. T.
and
Balch
,
W. E.
(
2013
).
Diversity in the origins of proteostasis networks — a driver for protein function in evolution
.
Nat. Rev. Mol. Cell Biol.
14
,
237
-
248
.
Sanderson
,
S. M.
,
Gao
,
X.
,
Dai
,
Z.
and
Locasale
,
J. W.
(
2019
).
Methionine metabolism in health and cancer: a nexus of diet and precision medicine
.
Nat. Rev. Cancer
19
,
625
-
637
.
Schmittgen
,
T. D.
and
Livak
,
K. J.
(
2008
).
Analyzing real-time PCR data by the comparative C(T) method
.
Nat. Protoc.
3
,
1101
-
1108
.
Semaniuk
,
U.
,
Piskovatska
,
V.
,
Strilbytska
,
O.
,
Strutynska
,
T.
,
Burdyliuk
,
N.
,
Vaiserman
,
A.
,
Bubalo
,
V.
,
Storey
,
K. B.
and
Lushchak
,
O.
(
2021a
).
Drosophila insulin-like peptides: from expression to functions – a review
.
Entomol. Exp. Appl.
169
,
195
-
208
.
Semaniuk
,
U.
,
Strilbytska
,
O.
,
Malinovska
,
K.
,
Storey
,
K. B.
,
Vaiserman
,
A.
,
Lushchak
,
V.
and
Lushchak
,
O.
(
2021b
).
Factors that regulate expression patterns of insulin-like peptides and their association with physiological and metabolic traits in Drosophila
.
Insect Biochem. Mol.
135
,
103609
.
Shih
,
S. R.
,
Huntsman
,
E. M.
,
Flores
,
M. E.
and
Snow
,
J. W.
(
2020
).
Reproductive potential does not cause loss of heat shock response performance in honey bees
.
Sci. Rep.
10
,
19610
-
19618
.
Shih
,
S. R.
,
Bach
,
D. M.
,
Rondeau
,
N. C.
,
Sam
,
J.
,
Lovinger
,
N. L.
,
Lopatkin
,
A. J.
and
Snow
,
J. W.
(
2021
).
Honey bee sHSP are responsive to diverse proteostatic stresses and potentially promising biomarkers of honey bee stress
.
Sci. Rep.
11
,
22087
.
Snow
,
J. W.
(
2020
).
Prolyl-tRNA synthetase inhibition reduces microsporidia infection intensity in honey bees
.
Ann. Abeille
51
,
557
-
569
.
Snow
,
J. W.
(
2022
).
Nosema apis and N. ceranae infection in honey bees: a model for host-pathogen interactions in insects
. In
Microsporidia, Current Advances in Biology
(ed.
L. M.
Weiss
and
A. W.
Reinke
), pp.
153
-
177
.
Springer
.
Steinhauer
,
N.
,
Kulhanek
,
K.
,
Antúnez
,
K.
,
Human
,
H.
,
Chantawannakul
,
P.
,
Chauzat
,
M.-P.
and
Vanengelsdorp
,
D.
(
2018
).
Drivers of colony losses
.
Curr. Opin. Insect Sci.
26
,
142
-
148
.
Sykiotis
,
G. P.
and
Bohmann
,
D.
(
2008
).
Keap1/Nrf2 signaling regulates oxidative stress tolerance and lifespan in Drosophila
.
Dev. Cell
14
,
76
-
85
.
Tain
,
L. S.
,
Jain
,
C.
,
Nespital
,
T.
,
Froehlich
,
J.
,
Hinze
,
Y.
,
Grönke
,
S.
and
Partridge
,
L.
(
2020
).
Longevity in response to lowered insulin signaling requires glycine N-methyltransferase-dependent spermidine production
.
Aging Cell
19
,
e13043
.
Torson
,
A. S.
,
Yocum
,
G. D.
,
Rinehart
,
J. P.
,
Nash
,
S. A.
,
Kvidera
,
K. M.
and
Bowsher
,
J. H.
(
2017
).
Physiological responses to fluctuating temperatures are characterized by distinct transcriptional profiles in a solitary bee
.
J. Exp. Biol.
220
,
3372
-
3380
.
Torson
,
A. S.
,
Yocum
,
G. D.
,
Rinehart
,
J. P.
,
Nash
,
S. A.
and
Bowsher
,
J. H.
(
2019
).
Fluctuating thermal regimes prevent chill injury but do not change patterns of oxidative stress in the alfalfa leafcutting bee, Megachile rotundata
.
J. Insect Physiol.
118
,
103935
.
Trapp
,
J.
,
McAfee
,
A.
and
Foster
,
L. J.
(
2017
).
Genomics, transcriptomics and proteomics: enabling insights into social evolution and disease challenges for managed and wild bees
.
Mol. Ecol.
26
,
718
-
739
.
Wang
,
M.
,
Xiao
,
Y.
,
Li
,
Y.
,
Wang
,
X.
,
Qi
,
S.
,
Wang
,
Y.
,
Zhao
,
L.
,
Wang
,
K.
,
Peng
,
W.
,
Luo
,
G.-Z.
et al. 
(
2021
).
RNA m6A modification functions in larval development and caste differentiation in honeybee (Apis mellifera)
.
Cell Rep.
34
,
108580
.
Wilkinson
,
E.
,
Cui
,
Y.-H.
and
He
,
Y.-Y.
(
2021
).
Context-dependent roles of RNA modifications in stress responses and diseases
.
Int. J. Mol. Sci.
22
,
1949
.
Wood
,
T. J.
,
Michez
,
D.
,
Paxton
,
R. J.
,
Drossart
,
M.
,
Neumann
,
P.
,
Gérard
,
M.
,
Vanderplanck
,
M.
,
Barraud
,
A.
,
Martinet
,
B.
,
Leclercq
,
N.
et al. 
(
2020
).
Managed honey bees as a radar for wild bee decline?
Apidologie
51
,
1100
-
1116
.
Xu
,
J.
and
James
,
R. R.
(
2012
).
Temperature stress affects the expression of immune response genes in the alfalfa leafcutting bee, Megachile rotundata
.
Insect Mol. Biol.
21
,
269
-
280
.
Xu
,
J.
,
Welker
,
D. L.
and
James
,
R. R.
(
2021
).
Variation in Expression of Reference Genes across Life Stages of a Bee, Megachile rotundata
.
Insects
12
,
36
.
Yocum
,
G. D.
,
Kemp
,
W. P.
,
Bosch
,
J.
and
Knoblett
,
J. N.
(
2005
).
Temporal variation in overwintering gene expression and respiration in the solitary bee Megachile rotundata
.
J. Insect Physiol.
51
,
621
-
629
.
Yocum
,
G. D.
,
Childers
,
A. K.
,
Rinehart
,
J. P.
,
Rajamohan
,
A.
,
Pitts-Singer
,
T. L.
,
Greenlee
,
K. J.
and
Bowsher
,
J. H.
(
2018
).
Environmental history impacts gene expression during diapause development in the alfalfa leafcutting bee, Megachile rotundata
.
J. Exp. Biol.
221
,
jeb.173443
.
Zattara
,
E. E.
and
Aizen
,
M. A.
(
2021
).
Worldwide occurrence records suggest a global decline in bee species richness
.
One Earth
4
,
114
-
123
.
Zhang
,
W.
,
Qian
,
Y.
and
Jia
,
G.
(
2021
).
The detection and functions of RNA modification m6A based on m6A writers and erasers
.
J. Biol. Chem.
297
,
100973
.

Competing interests

The authors declare no competing or financial interests.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.