Expression of heat-shock protein genes in Apis mellifera meda (Hymenoptera: Apidae) after exposure to monoterpenoids and infestation by Varroa destructor mites (Acari: Varroidae)

Heat shock proteins ( hsps ) protect proteins in eukaryotic cells from damage. Expression of hsps in insects subject to different environmental stimuli is poorly characterized. Here, levels of expression of the hsps genes ( hsp 40, hsp 70, and hsp 90) were recorded in Apis mellifera Linnaeus (Hymenoptera: Apidae) workers after exposure to sublethal concentrations of thymol, eucalyptol, α-pinene, trans -anethole, diallyl disulfide and infestation with Varroa mites. Our results show a dose-dependent up-regulation in the levels of all the hsps tested after the bees were treated with thymol, eucalyptol and α-pinene. Although these up-regulated expressions were statistically significant for hsp 70 and hsp 90 when the bees were treated with thymol and eucalyptol, they were not significant when treated with α-pinene. In addition, significant down-regulated expressions of the hsp genes were recorded in the diallyl disulfide treatment. The transcriptions of all the hsps tested were significantly down-regulated when pupae were infested with different numbers (0-5) of Varroa mites. Thus, it is likely that hsps can be used as biomarkers of survival when honey bees are under toxic and pathogenic stress, but this needs to be confirmed.


INTRODUCTION
The synthesis of most proteins in insects and other animals declines when they are exposed to external stressors like toxins, heat, cold and anoxia, although the levels of heat shock protein (hsps) increase, which bind aberrant proteins and aid in their refolding.Hsp function is associated with co-chaperones such as the J-domain protein and hsp40 (King & MacRae, 2015).When cells are subject to biotic and abiotic stress they synthesize hsps that function as molecular chaperones.Under normal conditions, these molecules assist proteins during secretion, degradation, folding, intracellular localization, assembly, etc. (Zhao & Jones, 2012).Hsps are classifi ed based on their molecular weight in kilo daltons (kDa), such as hsp40, hsp60, hsp70, hsp90, hsp100 and small hsps (shsps).Hsp40 is characterized by a 70 amino acid domain (J domain) called DnaJ (Carmel et al., 2011).Hsp40 acts as hsp70ʼs co-chaperone, stimulates hsp70-ATPase activity, and delivers unfolded proteins to hsp70.The other hsps like hsp60, hsp70 and hsp90 fold the proteins in ATP-dependent way (Szabo et al., 1996;King & MacRae, 2015).
The responses of hsp genes in useful insects like honey bees under stress have not been investigated in great detail.There are at least 36 hsps genes in the honey bee genome Varroa mites.To determine the toxicity of the above monoterpenoids, 1 μL of each concentration was poured on a piece of fi lter paper, which was attached to the inner surface of a Petri dish lid.Acetone was used as the solvent.After the solvent evaporated, the lids of Petri dishes were sealed tightly with Parafi lm to avoid the volatilization of the monoterpenoids.The negative control consisted of a test group that was not exposed to the monoterpenoids.The positive control was treated only with acetone.All treatments and controls were incubated at 35 ± 2°C and 75 ± 2% RH under a 12L : 12D photoperiod.Five hours after the start of the treatments the numbers of dead Varroa mites were counted.The mites that did not move when touched with a fi ne soft brush were counted as dead.The toxicity experiments were done using a completely randomized design with three replicates.

Exposure of honey bees to monoterpenoides
Before commencing this experiment, honeybee adult workers were fed on a piece of cotton dental wick that was soaked with 20% sucrose solution.Bees were kept at 28 ± 2°C and 58 ± 5% relative humidity for three days.
Subsequently, the bees were exposed to monoterpenoids using a feeding method adapted from Koo et al. (2015).The lethal concentrations of the four monoterpenoids that caused 10-90% mortality of the Varroa mites were used to determine their side effects on honeybee adult workers.Prior to exposure, a series of concentrations of the monoterpenoids used in each treatment were individually dissolved in 20% sucrose solution and then poured into cotton wicks using a micropipette (2 ml).This sucrose source was put in a plastic Petri dish (50 mm), placed at the bottom of a container and then bees (n = 5) were released into the container.Before exposure, the bees were placed in a refrigerator (4°C) for approximately 10 min to immobilize them.For each monoterpenoid treatment, 5 groups of 5 bees (total 25 bees) were kept in the plastic containers.In the control only a solution of sucrose was used.The treatments and control were kept in an incubator (24 h darkness; 28 ± 2°C, and 58 ± 5% RH) for fi ve days.On the fi fth day, the surviving bees were collected and the levels of expression of the hsp40, hsp70 and hsp90 genes determined.

Collection of honey bees infested with Varroa mite
Honey bee pupae were collected from 10 colonies (Sistan region, Zabol, Iran) infested with Varroa mites.Brood cells were uncapped individually using fi ne forceps and after counting the number of adult and nymphal mites in the cells, bee pupae were collected from cells with 0, 1, 2, 3, 4, 5 mites.These pupae were kept at -80°C until RNA extraction.

Gene expression of hsps
After fi ve days of exposure, the abdomens of bees (5 samples per each concentration and experiment) were randomly selected and cryogenized in liquid nitrogen.TRIzol® reagent (Invitrogen, USA) was used to extract total RNA according to the manufacturer's instructions.The RNA was subsequently washed twice with 1 mL 75% ethanol, centrifuged at 12,000 g for 7 min and dried at room temperature.The RNA pellet was re-suspended in nuclease-free water (20 μL).In addition, the RNA pellet was treated with DNase I (Invitrogen) to remove any contaminating genomic DNA.Immediately after extraction, RNA quantity and purity was evaluated using a spectrophotometer (Nanodrop2000c, Thermo Fisher Scientifi c).All samples had an excellent quality at 260 nm (200 ng RNA/sample).The RNAs were visualized on 1.5% agarose gel in TBE (pH 8.3).
With 1 μg of extracted RNA for each treatment, complementary DNA (cDNA) was synthesized using a QuantiTect Reverse Transcription Kit (Qiagen).The synthesized cDNAs were adjusted to 50 μL with sterile water and stored at -80°C prior to further Because of the signifi cant threat to honey bee health posed by Varroa mites, it is important that beekeepers monitor and control the levels of mites in their colonies.Recently, Varroa mites became resistant to synthetic insecticides as a result of their exclusive and consistent use.As a consequence, plant-derived compounds are being seriously considered for use as bio pesticides in Varroa management (Colin, 1990;Carayon et al., 2013;Xavier et al., 2015).These bio pesticides occur in plants belonging to certain plant families, such as Amaryllidaceae, Apiaceae, Asteraceae, Lamiaceae, Lauraceae, Myrtaceae, Piperaceae, Poaceae, Rutaceae, Zingiberaceae, etc. (Sahayaraj, 2014).
In this study, a carefully controlled and fully crossed laboratory experimental design was used to determine the effects of fi ve monoterpeoids (thymol, eucalyptol, α-pinene, diallyl disulfi de and trans-anethole) and the mite Varroa on the level of expression of different hsp genes (hsp40; hsp70; hsp90) in honey bee workers.Here, we attempt to answer the following questions: (1) Are there differences in the patterns of expression of the hsp genes in honey bees exposed to monoterpenoids?(2) Are there differences in the expression of these hsp genes in honey bees infested with Varroa mites?

Collection of mites and insects
Adults of Varroa destructor (Anderson & Trueman) (Acari: Varroidae) and A. mellifera workers (10-15 days old) were collected from colonies of A. mellifera meda located at Zabol (Sistan va Baluchestan, Iran) during July to August 2015.These collections did not involve endangered or protected species of honey bees and no specifi c permissions were required since the responsible beekeeper granted access to his apiary.Adults were collected from frames without brood, anesthetized with carbon dioxide (CO 2 ), transferred to the laboratory and placed in plastic containers (12 × 12 × 5 cm 3 ).

Exposure of Varroa mites to monoterpenoides
Colonies of A. mellifera that were heavily infested with Varroa mites and had not been treated with varroacides for six months were used as the source of Varroa mites.Varroa mites were removed from infested honey bee workers as described by Ariana et al. (2002).The infested adult honey bees (~1000-1500) were shaken directly from bee frames into an apparatus composed of an outer and inner cylinder.CO 2 was released into the inner cylinder for 5 min (fl ow rate of 5 L/min).After the bees were anesthetized, the apparatus was shaken several times to separate the mites from the adult honey bees.After the honeybees in inner cylinder recovered from the anesthesia they were returned to their mother colony.The mites were placed on white paper and approximately 200 Varroa mites were placed in each Petri dish (diameter = 60 mm), with each dish constituting a replicate.Varroa mites in each dish were provided with fi ve young bee pupae.Based on preliminary experiments, the lethal concentrations of four of the monoterpenoids, thymol, eucalyptol, α-pinene and trans-anethole, were 1, 2, 5, 12, and 30 ppm, while concentrations of 0.5, 1, 1.5, 2, and 2.5 ppm were used to assay the lethal effects of diallyl disulfi de on analysis.To perform Real-time PCR (RT-PCR), each reaction was prepared with a cDNA sample (50 ng) for each treatment as a template, 10 pM of gene-specifi c primers (Koo et al., 2015), SYBR green master mix (Invitrogen) and nuclease free water to a fi nal volume of 25 μL.For the fi rst PCR cycle, the reaction mixture was initially denatured at 95°C for 10 min (1 cycle), followed 15 s at 95°C for denaturing, 1 min for annealing at 60°C, and was extend at 72°C (35 s) for a total of 40 cycles.Fluorescence was measured during the annealing step.A dissociation step (95°C for 15s, 60°C for 30s, 95°C for 15s) was conducted to validate the amplifi cation of a single product in each PCR reaction.To obtain cDNAs of the pupae infested with mites, samples were washed three times with sterile deionized water and dried on fi lter paper.Individual pupae (5 samples per treatment, 3 replications) were randomly selected and homogenized in a 1.5 ml tube.Then the homogenate was treated with extract RNA.Synthesis cDNA and RT-PCR were conducted based on a procedure previously described by Koo et al. (2015).Actin was used as a reference gene to normalize the target genes (Lourenco et al., 2008).To obtain the relative quantities of hsp genes, threshold cycle (ct) values were used.Data were analyzed using the 2 -∆∆ct method (2 -∆∆ct = 2 -(∆ct treatment-∆ct reference) , Livak & Schmittgen, 2001;Qiao et al., 2015) using IQ5 Optical system software (Bio-Rad).In the negative control, no cDNA template was used.We replicated the reactions three times.Signifi cant differences among the gene expression levels of the treatments were analyzed using SPSS 16.0 software (Chicago, IL).In addition, one-way analysis of variance was used to compare the mean values (p > 0.05).

RESULTS
Varroa mites were killed by all of the fi ve monoterpenoids tested under laboratory conditions (Table 1).The mortalities recorded were greater than 90% when the highest concentrations (30 and 2.5 ppm, respectively) of diallyl disulfi de and thymol were used.As expected, all of the honey bee workers exposed to concentrations of monoterpenoids that were lethal for Varroa mites survived.
The levels of hsp40, hsp70, and hsp90 mRNA in the honey bee workers and pupae responded to the different concentrations of monoterpenoids (Table 2).The results re- veal a dose dependent up-regulation in the levels of hsp70 and hsp90 when the bee workers were exposed to thymol, eucalyptol and α-pinene.Although these up-regulated expressions were statistically signifi cant for hsp70 and hsp90 when the bees were treated with thymol and eucalyptol, the level of expression of hsp40 was not signifi cantly increased when the bees were exposed to any of the monoterpenoids tested.In addition, no signifi cant up-regulations in the expression of hsp70 and hsp90 genes were induced by α-pinene.
We also studied the expression levels of hsp genes after exposure to trans-anethole and diallyl disulfi de.Significant down-regulated expressions of the three genes tested were recorded in the bee workers treated with diallyl disulfi de.Similarly, slightly down-regulated expressions of hsps were recorded in bees exposed to trans-anethole, but these changes were not signifi cant (Table 2).
The transcripts of all the hsps tested were signifi cantly down-regulated when bee pupae were exposed to different numbers of Varroa mites (Fig. 1).
Honey bees come in contact with various pollutants during foraging and as a consequence of controlling Varroa mites in apiaries.It is hypothesized that sub lethal concentrations of pesticides in the fi eld can reduce the fecundity of honey bee pests as well as protecting foraging honey bees against pests.However, it is emphasized that sub lethal doses of pesticides disturb the physiology and behaviour of honey bees (Dai et al., 2010;Smodis Skerl & Gregorc, 2010;Gregorc & Ellis, 2011).As a result, the contamination of honey bee colonies with either miticides or pesticides could affect their physiology and any changes in the physiological parameters may result in the colony collapse disorder.Synthetic miticides or pesticides also affect the expression of hsps in honey bees.Lethal concentrations of pesticides like imidacloprid (0.5-50 ppm) decrease the levels of hsps (hsp70, hsp90, and grp78) in honey bees (Koo et al., 2015).Despite their biological effi ciency and the widespread application of botanical pesticides such as essential oils, their effects as stressors on expression profi les of hsps in insects have not yet been documented.
Untreated honey bee colonies infested with Varroa mites only survive for one to two years.Recent reports that Varroa mites have developed resistance to synthetic miticides have resulted in more attention being paid to developing botanical toxins for controlling mites in apiaries.The development of novel miticidal compounds is essential in order to combat the increasing resistance of mites to conventional miticides.Hence, in the search for alternative miticides the miticidal activity of plant secondary metabolites has been evaluated (Colin, 1990;Greatti & Barbatinin, 1996;Ariana et al., 2002;Umpiérrez et al., 2011;Carayon et al., 2013;Singh, 2014;Li et al., 2017).It is well documented that plant essential oils can be used as fumigant miticides in the treatment of Varroa mites of which the monoterpenoids are probably the best-known (Carayon et al., 2013).The results mainly confi rm that all fi ve monoterpenoids in a vapor-phase toxicity bioassay can kill Varroa mites but not honey bees.Previous studies also show that plant essential oils, including camphor, eucalyptol, menthol and thymol, are effective against Varroa at concentrations of up to 50 mg.kg - and safe for honey bees (Imdorf et al., 1999).The possibility of using essential oils in honey bee colonies with no resultant contamination of colony products, has encouraged researchers to investigate the physiological, behavioural, and biochemical effects and properties of essential oils with proven miticidal activity.
In spite of the evidence that honey bees exposed to pesticides and infestation with Varroa mites express immune genes (Boncristiani et al., 2012), no study has investigated the effects of secondary metabolites and Varroa on the transcript levels of hsps in honey bees.It was hypothesized that these stimuli are associated with changes in hsp levels in treated and untreated samples.Our result demonstrate that the relative mRNA levels of the hsps tested in honey bees exposed to secondary metabolites and infestation with Varroa mites were changed either up or down in a way specifi c to each treatment.The levels of the hsps tested (hsp40, hsp70, and hsp90) gradually increased when the concentrations of thymol eucalyptol and α-pinene were increased (Table 2).This increase indicates that the chaperone actions of hsps are involved in ameliorating the toxicity of these monoterpenoids.Up-regulation of hsps promotes tolerance in stressed cells (Du Rand et al., 2015).Our results confi rm that sub lethal concentrations of thymol and eucalyptol induce an increase in the levels of hsps in honey bees.
Different thymol-based formulations can be used to effectively control Varroa mites (Umpiérrez et al., 2011).A thymol-based varroacide signifi cantly alter the level of gene expression, metabolic responses including immunity, detoxifi cation and development in honey bees (Boncristiani et al., 2012).Researchers report signifi cant changes in the level of expression of different hsp genes in different organisms in response to plant secondary metabolites.For example, Burt et al. (2007) show that Escherichia coli treated with carvacrol (1 Mm) produce a signifi cant amount of hsp60.Like thymol, carvacrol is a phenol, which inhibits the function of transient receptor potential-like channels in insects (Parnas et al., 2009).
In this study, the level of hsp gene expression induced by α-pinene is dose-dependent.This monoterpenoid produced no signifi cant increase in the level of heat shock proteins (p > 0.05).The levels of hsps in honey bees treated with trans-anethole were down-regulated, but insignifi cantly so.Hummelbrunner & Isman (2001) show that for Spodo ptera litura (F.) larvae trans-anethole synergizes the toxicity of thymol.Therefore, it is likely that such compounds are stressors for honey bees and if combined their toxicity is increased synergistically.It is assumed that the mechanism of toxicity of such monoterpenoids is the same or the target sites in insects are similar, but the level of understanding is poor.In an attempt to characterize the detoxifi cation mechanisms in stressed honey bees, Du Rand et al. (2015) report that active detoxifi cation of nicotine by A. mellifera scutellata is associated with metabolic and proteomic responses such as up-regulation of heat shock proteins.Sun et al. (2014) studied the level of expression of hsp90 in Apolygus lucorum (Heteroptera: Miridae) in response to different stressors such as pesticides, including cyhalothrin, imidacloprid, chlorpyrifos and emamectin benzoate.
Their results confi rm that after pesticide treatment using LD 50 s, the mRNA levels of AlHSP90 in surviving male and female adults are signifi cantly higher than in the controls.Transcriptional results confi rm that AlHSP90 is an important gene involved in the resistance to pesticides.
In this study, the potential for using diallyl disulfi de for controlling Varroa mites is emphasized.This compound and diallyl trisulfi de are the main components of garlic essential oil and are both insecticidal and bactericidal (Yang et al., 2009).The application of a mixed formulation of garlic with "sulfur and pepper" (Greatti & Barbatinin, 1996) and "neem, garlic, and tobacco extracts" (Qayyoum et al., 2013) have proved effective in managing Varroa mites.Despite the fact that garlic and its components (diallyl disulfi de, diallyl trisulfi de) are toxic for Varroa, the metabolic responses of honey bees to these substances have not been investigated.Our results indicate that the levels of the hsps tested decrease when the concentration of diallyl disulfi de is increased.This indicates that it may not be possible to use changes in the chaperone action of hsps in honey bees as biosensors of particular environmental stresses such as pesticide or botanicals (Feder & Hofmann, 1999).In addition, expression of the mRNA levels of hsps indicates that these genes are affected by the toxicity of this compound.Highly down regulated hsps in honey bees in response to diallyl disulfi de indicate that these proteins cannot prevent cell injury in honey bees.Therefore, its use for controlling Varroa mites is dependent on determining whether honey bees can tolerate this compound.
Our results demonstrate that an infestation of Varroa mites can result in signifi cantly lower transcript levels of all the hsps tested in honey bees.There was a signifi cant decrease in level of these genes in bees infested with 1 to 5 mites.There was a signifi cant down-regulation in the mRNA level of hsp90 in bees infested with one mite compared to the control.Our results show that the number of mites regularly infesting bee workers probably decrease the immune responses of bees, which need to be investigated in future studies.Zhu et al. (2013) investigate the roles of hsps (hsp20, hsp75 and hsp90) in the host-parasite interaction between Pieris rapae (Lepidoptera: Pieridae) and its parasitoid Pteromalus puparum (Hymenoptera: Pteromalidae).Their results indicate that in parasitized pupae, the hsp20 transcripts in response to P. puparum are signifi cantly higher than in the controls.In addition, the expression of hsp75 and hsp90 were down-regulated in parasitized pupae.They conclude that the hsps tested are involved in host-parasitoid interactions and the transcription of hsp could be a component of the syndrome of parasitized hosts (Zhu et al., 2013).Rinehart et al. (2002) and Shim et al. (2007) report similar interactions in Sarcophaga crassipalpis (Diptera: Sarcophagidae) parasitized by Nasonia vitripennis (Hymenoptera: Pteromalidae) and Plodia interpunctellaI (Lepidoptera: Pyralidae) parasitized by Bracon hebetor (Hymenoptera: Braconidae), respectively.In general, the expression of these hsp genes in response to parasitization indicates they have an important role in hostparasite interactions.Zhu et al. (2013) suggest that during immune defense, the production of reactive oxygen species (ROS) is often induced, which results in proteotoxicity of host cells, death of the parasitoid and protein denaturation.In addition, hsps as cytoprotective proteins are activated in response to an increase in ROS levels or toxic substances produced by pathogens.The regulation of the expression of hsp in an insect after parasitization could protect the host's tissues and the developing parasitoid from noxious agents produced by the host's immune response (Zhu et al., 2013).

CONCLUSION
The potential for using selective and fully biodegradable compounds in the management of parasitic mites of honey bees is encouraging.Our results indicate that the expression of hsps in honey bees (except for hsp40) could serve as biomarkers for assessing the health status of honey beesexposed either to Varroa infestation or to monoterpenoids used to treat this infestation.Further research is required to verify our results and to determine the molecular mechanism of honey bee tolerance to external stressors like botanical toxins and infestation with mites.
ACKNOWLEDGEMENTS.The authors would like to thank local beekeepers and science offi cers for collecting the honey bees and assisting during this project.We greatly appreciate E.H. Hong for analyzing the data.

DECLARATION OF INTEREST.
The authors report no confl icts of interest.

Fig. 1 .
Fig. 1.Relative mRNA expression levels of hsps when honeybee pupae were exposure to Varroa mites.Relative levels of each hsp were measured by real-time polymerase chain reaction (RT-PCR) with the actin gene serving as a reference (control) to normalise data.The different letters above each bar indicate statistical difference analyzed using SPSS 16.0 software (Chicago, IL) followed by Duncan's multiple comparison test (p < 0.05).Values are presented as mean ± SE (n = 5, replication = 3).

Table 1 .
Mean percent mortality of Varroa mites after exposure to different concentrations of four monoterpenoids (thymol, eucalyptol, α-pinene and trans-anethole) under laboratory conditions.

Table 2 .
Relative level of mRNA expression of each hsp in honey bee workers previously exposed to each of the 5 monoterpenoids tested.
Relative levels of each hsp were measured by real-time polymerase chain reaction (RT-PCR) with the actin gene serving as a reference (control) to normalize data.The different letters indicate statistically signifi cant differences between treatments.Values are presented as mean ± SE (n = 5, replication = 3).