- Email: [email protected]
Recognition of mite-infested brood by honeybee (Apis mellifera) workers may involve thermal sensing
Author’s Accepted Manuscript Recognition of mite-infested brood by honeybee (Apis mellifera) workers may involve thermal sensing Daniel Bauer, Jakob Wegener, Kaspar Bienefeld www.elsevier.com/locate/jtherbio
PII: DOI: Reference:
S0306-4565(17)30495-3 https://doi.org/10.1016/j.jtherbio.2018.04.012 TB2102
To appear in: Journal of Thermal Biology Received date: 19 November 2017 Revised date: 22 February 2018 Accepted date: 24 April 2018 Cite this article as: Daniel Bauer, Jakob Wegener and Kaspar Bienefeld, Recognition of mite-infested brood by honeybee (Apis mellifera) workers may involve thermal sensing, Journal of Thermal Biology, https://doi.org/10.1016/j.jtherbio.2018.04.012 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1 Recognition of mite-infested brood by honeybee (Apis mellifera) workers may involve thermal sensing Daniel Bauer, Jakob Wegener, Kaspar Bienefeld* Bee Research Institute, F.-Engels-Straße 32, 16540 Hohen Neuendorf, Germany
*
Corresponding author: Tel. 0049-3303-293830; fax 0049-3303-293840
[email protected] [email protected] [email protected]
Abstract Hygienic behavior, i.e. the removal of diseased or damaged brood by worker honey bees (Apis mellifera), is seen as one of the principal behavioral elements of this species’ social immunity. Identification of the stimuli that trigger it would be helpful in searching for biochemical and molecular markers of this important breeding trait. While many studies at the genomic, transcriptomic, and behavioral level have pointed to the implication of chemical cues, we here hypothesized that thermal cues are alternatively/additionally involved. To test this hypothesis, we first measured whether infestation by the mite Varroa destructor (a condition known to induce hygienic behavior) leads to a thermal gradient between affected and unaffected brood. We found that infested brood cells were between 0.03 and 0.19 °C warmer than uninfested controls. Next, we tested whether artificially heating an area of a brood comb would increase the removal of infested or uninfested brood as compared to an unheated control area, and found that this was not the case. Finally, we investigated whether the heating of individual brood cells, as opposed to comb areas, would influence brood removal from cells adjacent to the heated one. This was the case for uninfested, though not for infested cells. We conclude that infestation by V. destructor leads to a heating of brood cells that should be perceivable by bees, and that small-scale temperature gradients can influence brood removal. This makes it appear possible that thermal cues play a role in triggering hygienic behavior of honey bees directed at varroa-infested larvae/pupae, although our results are insufficient to prove such an involvement.
Keywords: Varroa destructor; uncapping; thermography; gradient sensing 1. Introduction The ectoparasitic mite Varroa destructor is seen as the most important pest of the Western honey bee (Apis mellifera) in both economic and ecological terms (Le Conte et al., 2007; Le Conte et al.,
2 2010; Neumann and Carreck, 2010; van der Zee et al., 2015). Not surprisingly therefore, intense efforts at selecting varroa-resistant lines of honeybees are under way in many parts of the world. At the center of many of these breeding programs is “hygienic behavior” (Rothenbuhler, 1964), the capacity of worker bees to recognize and remove diseased brood. As Varroa destructor only reproduces on honeybee brood, hygienic behavior directed against varroa-infested cells greatly reduces reproductive success of the parasite, despite the fact that most mites survive brood removal and can later infest other cells (Spivak and Reuter, 2001). Much of the damage caused by varroa infestation is linked to the transmission of viruses such as deformed wing virus (DWV) (Di Prisco et al., 2016; Le Conte et al., 2010; Nazzi et al., 2012), and replication of DWV in infested brood increases the likelihood of removal of the brood concerned (Schöning et al., 2012). Varroa infests honeybee brood shortly before brood cells are sealed with wax by worker bees and bee larvae enter pupation. This means that the stimuli triggering hygienic behavior must be sensed by worker bees across the wax cap as well as the pupal cocoon. Results of several transcriptomic, genomic, and proteomic studies make it appear likely that increased hygienic behavior is linked to alterations in the capacity of bees to perceive chemical cues (Hu et al., 2016; Masterman et al., 2000; Mondet et al., 2015; Spötter et al., 2016; Swanson et al., 2009) – “hygienic” lines and individuals show an increased expression of certain odorant binding proteins, as well as of proteins involved in the processing of olfactory information. Bees selected for strong hygienic behavior were also experimentally shown to better discriminate between the odors of healthy and diseased brood (Masterman et al., 2000). Martin et al. (2002) have even identified candidate substances (mostly long-chained alkanes, alkenes or fatty acids) originating from the mites that are more strongly perceived by “hygienic” than by “non-hygienic” breeds of bees. However this does not preclude the existence of additional, nonolfactory components of the stimulus. Although honeybee biology offers examples of behaviors that seem to be triggered by chemical cues alone (e.g. the suppression of worker reproduction; Traynor et al., 2014), combinations of olfaction/gustation with other cues seems to be the more common case. Examples include the mating behavior of drones (Gary and Marston, 1971), navigation of foragers to floral resources (e.g. Ravi et al., 2016), and the recruitment of foragers by other foragers within the hive (Wenner et al., 1991). Temperature and temperature gradients are one source of information guiding honeybee behavior. The most evident illustration for this is the control of heating, ventilation and water evaporation for thermal homeostasis of the brood nest - in contrast to adult bees, honeybee brood is extremely stenothermic, and brood nest temperature is regulated by worker bees within the range of 32-36°C through both active cooling and heating (Kronenberg and Heller, 1982; Lindauer, 1954; Petz et al., 2004; Stabentheiner et al., 2010). Interactions between individual workers may equally involve thermal sensing, as hypothesized for the recruitment of new foragers to a particular source of nectar by other foragers returning from this source (Stabentheiner, 2001). Honeybee workers can be trained to associate certain temperatures with a food reward (Hammer et al., 2009), and thermal gradients as small as 0.2°C have been shown to evoke a behavioral response (Heran, 1952). Insects can use two different types of thermal cues to guide their behavior, absolute temperature and temperature gradients. Homeostasis of the honeybee brood nest temperature is an example of the former mechanism – bees heat or cool the nest without reference to a fixed external standard (Stabentheiner et al., 2010), apparently using a genetically-fixed internal standard. In Drosophila melanogaster, the biochemical basis of such an internal standard has recently been elucidated – its central elements appear to be thermosensitive transient receptor protein cation channels in the membranes of neuronal cells (Abram et al., 2016). In contrast, behaviors like thermal orientation
3 clearly require the perception of temperature gradients. Both honeybees and leafcutter ants (Atta vollenweideri) can be trained to associate temperature gradients with food rewards (Hammer et al., 2009; Kleineidam et al., 2007), and Atta-ants were shown to possess neurons that are sensitive to temperature changes below 0.1°C over a wide range of external temperature (Ruchty et al., 2010). Here we investigated whether thermal sensing is important for the triggering of Apis mellifera hygienic behavior. We started by measuring the temperature of infested an uninfested brood cells during development, in order to test the existence of a temperature difference that may serve as a discriminatory signal to hygienic workers. Next, we compared cell opening/brood removal in areas of a comb that either were or were not artificially heated. Finally, to test whether hygienic bees are exploiting small-spaced temperature gradients in the identification of diseased brood, we compared the removal of infested and non-infested brood as a function of the distance to an individual heated cell.
2. Materials and methods 2.1 Thermo-controlled hive box and colony Brood is usually situated at the center of the honeybee colony, surrounded by combs of pollen and honey, and brood outside this position is either neglected by workers or covered by thick layers of bees that ensure heating and insulation (Winston, 1987). To prevent this and enable video and thermographic observation of brood combs, an entire queenright colony of the subspecies A. m. carnica, composed of approximately 7.000 worker bees, three brood frames and three frames of food stores was therefore placed within an incubator set at 33.0°C. This temperature was chosen because it is within the range observed in the brood nest of unmanipulated colonies (Büdel, 1960), and leaves some space for biologically- or experimentally-induced increases, without running the risk of overheating brood. A flight hole was produced that allowed bees to leave and enter the incubator. External control over the hive temperature allowed us to place experimental brood combs at the border of the colony and still avoid crowding of bees on the brood. It also made observations independent of variations of outside temperature. An infrared video camera (PD/DX4-285GE; Kappa optronics, Germany; with objective Fujinon HF16HA-1B) and in some experiments also a thermographic camera (T335; FLIR systems, USA) were placed inside the incubator, directly facing the experimental comb (see photograph S1 for a depiction of the observational system). This thermocontrolled hive box was used for all three experiments.
2.2 Measurement of the temperature of varroa-infested and uninfested brood cells To test whether varroa infestation induces a change in honeybee brood temperature, a comb covered with freshly capped brood on only one side was procured from another colony. Sixteen type K thermoelements (d = 0.5 mm; l = 40 mm: precision ± 0.1 °C) were calibrated in a water bath and carefully inserted from the back into non-adjacent capped brood cells. Strong white light was used to candle the brood cells and avoid damaging the larvae within. To prevent movements of the thermoelements within the cells after introduction, the wires exiting the elements were glued to a
4 piece of plastic tubing which was mounted behind the brood comb. Next, eight of the cells containing thermoelements were artificially infested with varroa mites. These were procured from brood of the same stage, stemming from a heavily-infested colony. To introduce mites, the cell caps were carefully opened with a razor blade and the parasites placed inside using a moistened brush (Bienefeld et al., 2015). Two mites per cell were introduced in order to reduce heterogeneity of results linked to reproduction or non-reproduction of the parasites within the cells (Harbo and Harris, 2005), or transmission or non-transmission of viruses by the mites (Bowen-Walker et al., 1999). The other eight thermoelements were mounted inside control cells, i.e. capped brood cells that were treated identically to the infested cells except that no mite was introduced. The brood comb was then placed at the outside of the broodnest of the experimental colony, with the brood facing the video camera. Temperature was recorded at intervals of 10 seconds, using a data logger (SQ2040 4F16; Grant, GB). Recordings started after mite introduction and lasted for 144 h (until day 15-16 of development, counted from the moment of egglaying). Video observations of the infested cells were recorded throughout this period, in order to keep track of hygienic behavior performed on them. After the end of the observations, the state of the brood (dead/alive) and the mites (presence/absence of the introduced mite and of any offspring) were determined. The experiment was repeated 7 times. Brood cells that were emptied by the bees during the recordings were excluded from the analysis. Results from a total of 33 infested and 54 non-infested cells entered the data analysis.
2.3 Effect of heating of brood areas on the frequency of hygienic behavior As we found that infestation by varroa repeatably led to an increase of temperature of the cells concerned, we tested whether an elevation of the temperature of an area of the experimental comb would lead to an increase in the frequency of hygienic behavior. A brood comb containing freshlycapped brood (day 10 after egg-laying) on one side and mostly empty cells on the other was procured from a donor colony. The wax cells on the empty side were removed. A copper plate covering 10 x 10 cells was mounted against the back side of the brood. The plate was mounted on a copper tube that was connected to an Arctic A10 tempered bath circulator (ThermoFisher, USA), pumping tempered water resin through the tube. The temperature of the plate and brood was monitored through thermoelements that were attached to the plate, and through thermographic observations of the brood surface. In this way, temperature at the surface of capped cells was elevated by 0.5 ± 0.3 °C (calculated from randomly chosen thermograhic images from each repetition of the experiment). A second, unheated copper plate was mounted on the back side of the comb at a distance of approx. 5 cm from the heated plate, and the cells in contact with this second plate were used as the non-heated control. Nine non-adjacent cells were artificially infested with 2 varroa mites per cell within each of the two areas of brood (heated and non-heated). Hygienic behavior was monitored through infrared video observations as in experiment 1, and the fate of mites in cells that were still capped was determined after the end of observations (day 15 of development). The experiment was repeated seven times, three times with bees from a line selected for hygienic behavior and four with bees from a closely-related, but largely non-hygienic line (see Hu et al., 2016, for a description of the lines).
2.4 Effect of small-scale temperature gradients on the frequency of hygienic behavior
5 To test the hypothesis that small-scale temperature differences are involved in triggering hygienic behavior, the following experiment was performed. The setup for this experiment is shown in figure 1. On a frame of freshly-capped brood, four capped cells were chosen that defined a rectangle of approximately 8 x 6 cm. Only these four individual cells were heated from the back, by introducing copper bolts (d = 3 mm) from behind. The pupae were previously removed from the cells. To heat the bolts in a precisely controlled way, they were soldered to a copper plate (d = 0.2 mm) that was attached to the back side of the comb. This plate was heated via a heating foil (Conrad Electronic, Germany) which was glued onto it and connected to an adjustable power supply. To prevent heating of other than the manipulated brood cells by the foil/copper plate, several layers of insulating foil (TIGA-Med, Germany) were attached between the copper plate and the comb. The position of the heated cells was marked on transparent sheets placed on the comb for later identification. Around each heated cell, sixteen additional cells were marked. Of these, four were directly adjacent to the heated cell, while the twelve others were separated from it by 1, 2, or 3 unheated cells (see figure 1 for a schematic representation of the marking scheme). Evaluation of thermographic images from 3 of the 4 repetitions (see S3 for an example) show that the surface of the heated cell was approximately 1.1°C warmer than the background temperature of the comb. The gradient between the heated and the directly adjacent cell was approximately 0.7°C ± 0.4. Cells separated from the heated cell by 1, 2, or 3 other cells were 0.8, 0.9, and 1.0 degrees cooler than the heated cell. Of the sixteen cells surrounding each heated cell, eight were artificially infested with two varroa mites/cell, while the other eight were not. Another cell, positioned at a distance from the heated cells, was equipped with an unheated wire, and cells adjacent to it or separated from it by 1, 2, or 3 other cells served as controls.
2.5 Data analysis Measurements of brood temperature inside infested and uninfested cells were treated by splitting the duration of the video observations (144 h) into blocks of 12 h, for which the mean temperature within each measured cell was calculated. This condensed dataset was analysed by repeated measurement ANOVA (SAS Procedure “mixed”; SAS, 2012). Because the temperature gradient depends on whether surrounding cells were empty or containing brood, the number of empty cells adjacent to the measured cell was included in the statistical model. The ANOVA therefore included the factors “time”, “infestation” (yes/no), “experiment” (repetitions of the experiment; 1 through 7), number of empty adjacent cells (1 to 6), as well as the interaction between time and “infestation”. Data on the likelihood of brood removal in areas of heated/unheated brood (experiment 2) were analyzed using the chi²-test. For this, data from the seven repetitions of the experiment were pooled after verification that this was admissible. Data on the speed and likelihood of brood removal as a function of the distance to a single, heated cell (experiment 3) were submitted to a Cox-regression (IBM, 2011), including the factors of “infestation” (yes/no), “experiment” (repetitions of the experiment, 1 through 13), and “distance from heated cell” (1 through 4). Interactions between “infestation” and “distance” were also included.
6
3. Results 3.1 Temperature of infested and uninfested brood The average course of temperature during development of infested and uninfested brood is shown in figure 2. During all 12 time intervals, temperature was slightly higher in the infested cells. The overall effect of infestation of brood temperature was highly significant (F = 45.2; df = 1; P < 0.001), as were the effects of repetition (F = 108.4; df = 6; P < 0.001), of “number of empty adjacent cells” (F = 23.15; df = 5; P < 0.001) and of time (F = 25.44; df = 11; P < 0.001). The temperature difference due to infestation appeared to be greatest 24-72 h post-infestation, when it averaged 0.09°C, and again 120 – 144 h post-infestation, when it reached 0.17 – 0.19°C. The interaction between factors time and infestation was however not significant (P = 0.14). 3.2 Effect of artificial heating of brood areas on the likelihood of removal The frequency of removal of infested brood in areas of the comb heated or not by 0.5 – 1.0 °C was not significantly different (χ²-test; χ² = 0.30; total n = 111; P = 0.64); this is true regardless of the type of worker bees used (selected for hygienic behavior: total n = 54; χ² = 0.02; P = 0.89; unselected: total n = 57; χ² = 0.63; P = 0.43). Not surprisingly, selected bees showed a higher proportion of infested brood removal (75.9% versus 50.0%; χ² = 9.30; P = 0.002). As in the case of infested cells, uninfested cells were not more likely to be removed in warmed compared to non-warmed areas of the comb (total n = 241; χ² = 0.30; P = 0.58). Infestation had a highly significant influence on the frequency of removal (χ² = 146.73; total n = 352; P < 0.001). 3.3 Effect of narrow-spaced temperature gradients on brood removal The frequency and timing of brood removal around artificially-heated individual cells is given in figure 3. Again, there was a highly significant effect of infestation status on the course of removal (Coxregression; Wald = 30.5; df = 2; P < 0.001), with infested cells being preferentially removed. The proximity of brood to the heated cell also influenced removal (Wald = 24.1; df = 3; P < 0.001), with only cells directly neighboring the heated cells showing an increased likelihood. The effect of the proximity of the heated cells was however only measurable in the case of uninfested cells (comparison between figure 3 a. and b.), as expressed by a significant interaction-term for the factors “infestation” and “distance” (Wald = 13.5; df = 6; P = 0.04). Only one out of 17 uninfested larvae/pupae neighboring a cell containing an unheated copper wire was removed, which was significantly different from the removal rate of uninfested brood around the heated wires (13 out of 38; χ² = 5.3; P = 0.03), showing that removal was triggered by heat rather than by manipulation or the presence of the wire.
4. Discussion Our data show a slight but systematic increase of brood temperature by Varroa destructor – infestation of approximately 0.03 to 0.19°C. The size of this increase is near the resolution of the thermoelements used (0.1 °C). As however the effect was consistently observed over seven repetitions of the experiment, it seems certain that it really exists. The cause of this increase is
7 unclear. Although many insects, including honeybees, are known to adjust their body temperature in response to infection or parasitization, they are generally achieving this through behavioral changes leading to increased uptake of heat from the environment (“behavioral fever”; e.g. Abram et al., 2016; Anderson et al., 2013; Starks et al., 2000; Campbell et al., 2010; Stabentheiner et al., 2007) – a mechanism that seems unlikely in the case of Apis mellifera brood. It is known however that viral replication – which is often triggered by varroa parasitization (Bowen-Walker et al., 1999; Gisder et al., 2009) - is frequently accompanied by intensified metabolic acitivity in the host, an effect that is attributed to the manipulation of host metabolism by the virus (Sanchez and Lagunoff, 2015). An increase in metabolic heat may also result from the immune response of the host to mite infestation and increased viral replication (Nazzi et al., 2012). Viral replication is often intense already during the first 24 h after infestation by the mite (Nazzi et al., 2012), which might explain why the temperature gradient between infested an non-infested cells already showed 12-24 h after mite introduction in our experiment. Varroa-induced viral replication is known to greatly influence the probability of hygienic behavior directed against varroa-infested cells (Schöning et al., 2012). The size of the temperature increase observed is only slightly below that of the smallest temperature differences shown to induce a behavioral change in honeybees. Heran (1952) showed that bees changed their position in a continuous temperature gradient if the temperature was lowered by 0.25 °C, and Hammer et al. (2009) were able to train honeybees to distinguish sources of heat that differed by 0.4°C. It therefore seems possible that the increase in temperature linked to varroa infestation could be big enough to be perceived by worker bees. The fact that honeybees can regulate broodnest temperature despite changes of outside temperature with a precision of approximately ± 1.5°C (Büdel, 1960) shows that they are capable to sense absolute temperature without calibration by comparison to a stable external standard, and use this information for behavioral adaptations. In experiment 2, we therefore tested whether such changes of absolute temperature would influence the frequency of hygienic behavior. The results show that the heating of brood areas by 0.5 – 1.0°C, i.e. slightly more than the observed natural difference between infested and non-infested cells, does not change the frequency of removal of either uninfested or infested cells. Strictly speaking, this is no proof that absolute temperature is not an element of the stimulus leading to brood removal, because 1) due to technical limitations regarding the precision of the bath circulator and especially of the heat transfer from the copper plate to the brood, we were not able to experimentally induce precisely the same temperature increase that had been observed in experiment 1 (max. 0.19°C) in the areas of brood artificially heated in experiment 2. 2) it cannot be excluded that in infested cells, the natural increase of temperature due to the infestation already leads to the maximum behavioral response, so that a further increase is not possible. 3) in the case of uninfested cells, the lack of the olfactory components of the stimulus may have suppressed the hypothetical effect of absolute temperature. Nevertheless, the combination of results from the heated area-experiment (experiment 2) with the experiment on small-scale temperature gradients (experiment 3) make it unlikely that absolute temperature is relevant for removal, because in experiment 3, an increase of temperature of the same order of magnitude in individual, uninfested cells did increase their likelihood of brood removal. This shows that the absolute size of the temperature difference was likely not the reason
8 for non-removal in experiment 2, and that removal can also be triggered without the presence of the supposed chemical stimuli produced by infestation/viral replication. Moreover, it can also be questioned whether absolute temperature would be a sufficiently reliable signal for triggering hygienic behavior. The measurements performed during experiment 1 (supplementary materials 2 and 3) show that temperature variations within the brood cells due to brood development may be larger than the difference between infested and non-infested cells, and it is known that temperature can also vary between different areas of the A. mellifera brood nest (Büdel, 1960). Compared to absolute temperature, thermal gradients between neighboring cells can be expected to be more independent of the overall position within the broodnest. They are also less likely to be masked by differences due to the brood stage, because brood of the same stage is usually (although not always) clustered on the combs. Known thermal gradients within the brood nest include those between the bottom and rim of (empty) cells (δ = -0.1 to 1.3; Stabentheiner et al., 2010) as well as those between endothermic and ectothermic adult bees (especially their thoraces; approximately 2 – 8°C; Basile et al., 2008; Stabentheiner et al., 2010), but these do not directly concern the surface of capped brood cells, where signal pickup for hygienic behavior is likely to occur. Figure 3 shows that the removal of brood can indeed be influenced by small-spaced gradients. The fact that the stimulating effect of small-scale heating in the experiment was limited to uninfested cells may indicate that in the case of infested cells, the gradient produced through infestation is already sufficient to trigger the maximum behavioral response. Alternatively however, it may mean that nonthermal cues emanating from infested brood are sufficient to trigger the maximum behavioral response, which cannot be increased any further by an additional thermal cue. To distinguish between these two hypotheses, an experiment involving the artificial cooling of individual infested cells would be of great interest. Although small-scale heating increased the likelihood of brood removal, this increase was insufficient to close the gap between infested and non-infested brood. Theoretically, this may be due to the fact that the thermal or spatial size of the artificially-induced gradient were not quite equal to that of the natural gradient. A more likely explanation is that thermal cues, if they are at all involved in triggering removal, may only be one component of a complex stimulus involving other elements. As pointed out in the introduction, at least some of these additional elements of the removal stimulus are likely to be perceived by olfaction (Hu et al., 2016; Masterman et al., 2000; Mondet et al., 2015; Schöning et al., 2012; Spötter et al., 2016). The hypothesis of thermal cues as exclusive triggers of the removal of varroa-infested brood is also made unlikely by the steady increase of removal frequency with the developmental stage of brood (regardless of heating or infestation status; figure 3), whose kinetics differ from those of both absolute brood temperature and the size of the temperature gradient between infested and uninfested cells (figure 2). Our study focuses on the stimuli triggering the removal of brood infested by Varroa destructor. "Varroa-sensitive hygiene” (VSH; Harbo and Harris, 2005) stands out from other forms of honeybee hygienic behavior, because at the difference to many other stressors affecting the brood (such as the fungus Ascosphaera apis or the bacterium Paenibacillus larvae), varroa rarely kills its host (Rosenkranz et al., 2010). Chemical and thermal cues emitted by dead larvae/pupae can evidently be expected to be different from those resulting from sublethal damage. Our results are therefore directly relevant only for the latter case. While we showed that varroa-infestation leads to a subtle increase in temperature, death of the larva/pupa most likely will lead to a (more pronounced) cooling.
9 Identification of the complete stimulus governing honeybee hygienic behavior towards varroainfested brood could stimulate the search for genetic variants of A. mellifera that either show increased sensitivity for this stimulus, or produce a stronger stimulus if infested. We have shown here that infestation by Varroa destructor leads to a slightly increased temperature of the infested brood cells, and that bees are able to use small-spaced temperature gradients for decisions regarding the removal or non-removal of brood cells. Our results are insufficient however to conclude whether or not such thermal cues are relevant in the case of hygienic behavior towards varroa-infested larvae/pupae.
5. Acknowledgements We thank Mario Neumann, Uwe Gerber and Fred Zautke for help with IR-recordings and the construction of the experimental hive box. Funding: The work of DB was made possible by a grant from the F.-W. Schaumann-foundation. Parts of the study were additionally funded by the European Commission (FP7 KBBE program 2013.1.3-02, SmartBees, Grant Agreement Number 613960).
References Abram, P. K., Boivin, G., Moiroux, J. and Brodeur, J. (2016). Behavioural effects of temperature on ectothermic animals: unifying thermal physiology and behavioural plasticity. Biological Reviews 92, 1859-1876. Anderson, R. D., Blanford, S., Jenkins, N. E. and Thomas, M. B. (2013). Discriminating Fever Behavior in House Flies. PLoS ONE 8, e62269. Basile, R., Pirk, C. W. W. and Tautz, J. (2008). Trophallactic activities in the honeybee brood nest – Heaters get supplied with high performance fuel. Zoology 111, 433-441. Bienefeld, K., Zautke, F. and Gupta, P. (2015). A novel method for undisturbed long-term observation of honey bee (Apis mellifera) behavior – illustrated by hygienic behavior towards varroa infestation. Journal of Apicultural Research 54, 541-547. Bowen-Walker, P. L., Martin, S. J. and Gunn, A. (1999). The Transmission of Deformed Wing Virus between Honeybees (Apis melliferaL.) by the Ectoparasitic MiteVarroa jacobsoniOud. Journal of Invertebrate Pathology 73, 101-106. Büdel, A. (1960). Bienenphysik. In Biene und Bienenzucht, eds. A. Büdel and E. Herold), pp. 115-180. Munich: Ehrenwirth. Campbell, J., Kessler, B., Mayack, C. and Naug, D. (2010). Behavioural fever in infected honeybees: parasitic manipulation or coincidental benefit? Parasitology 137, 1487-1491. Gary, N. E. and Marston, J. (1971). Mating behaviour of drone honey bees with queen models (Apis mellifera L.). Animal Behaviour 19, 299-304. Gisder, S., Aumeier, P. and Genersch, E. (2009). Deformed wing virus: replication and viral load in mites (Varroa destructor). Journal of General Virology 90, 463-467. Hammer, T. J., Hata, C. and Nieh, J. C. (2009). Thermal learning in the honeybee, Apis mellifera. Journal of Experimental Biology 212, 3928-3934. Harbo, J. and Harris, J. (2005). Suppressed mite reproduction explained by the behaviour of adult bees. J Apicultural Res 44. Heran, H. (1952). Untersuchungen über den Temperatursinn der Honigbiene (Apis mellifica) unter besonderer Berücksichtigung der Wahrnehmung strahlender Wärme. Zeitschrift für vergleichende Physiologie 34, 179-206. Hu, H., Bienefeld, K., Wegener, J., Zautke, F., Hao, Y., Feng, M., Han, B., Fang, Y., Wubie, A. J. and Li, J. (2016). Proteome Analysis of the Hemolymph, Mushroom Body, and Antenna Provides
10 Novel Insight into Honeybee Resistance against Varroa Infestation. Journal of Proteome Research 15, 2841-2854. IBM. (2011). IBM SPSS statistics 20 core system user's guide. Armonk: IBM Corporation. Kleineidam, C. J., Ruchty, M., Casero-Montes, Z. A. and Roces, F. (2007). Thermal radiation as a learned orientation cue in leaf-cutting ants (Atta vollenweideri). Journal of Insect Physiology 53, 478-487. Kronenberg, F. and Heller, H. C. (1982). Colonial thermoregulation in honey bees (Apis mellifera). Journal of Comparative Physiology 148, 65-76. Le Conte, Y., de Vaublanc, G., Crauser, D., Jeanne, F., Rousselle, J.-C. and Bécard, J.-M. (2007). Honey bee colonies that have survived Varroa destructor. Apidologie 38, 566-572. Le Conte, Y., Ellis, M. and Ritter, W. (2010). Varroa mites and honey bee health: can Varroa explain part of the colony losses? Apidologie 41, 1-11. Lindauer, M. (1954). Temperaturregulierung und Wasserhaushalt im Bienenstaat. Zeitschrift für vergleichende Physiologie 36, 391-432. Martin, S. J., Elzen, P. J. and Rubink, W. R. (2002). Effect of acaricide resistance on reproductive ability of the honey bee mite Varroa destructor. Experimental and Applied Acarology 27, 195-207. Masterman, R., Smith, B. H. and Spivak, M. (2000). Brood odor discrimination abilities in hygienic honey bees (Apis mellifera L.) using proboscis extension reflex conditioning. Journal of Insect Behavior 13, 87-101. Mondet, F., Alaux, C., Severac, D., Rohmer, M., Mercer, A. R. and Le Conte, Y. (2015). Antennae hold a key to Varroa-sensitive hygiene behaviour in honey bees. Scientific Reports 5, Article nb. 10454. Nazzi, F., Brown, S. P., Annoscia, D., del Piccolo, F., Di Prisco, G., Varricchio, P., della Vedova, G., Cattonaro, F., Caprio, E. and Pennacchio, F. (2012). Synergistic parasite-pathogen ineractions mediated by host immunity can drive the collapse of honeybee colonies. PLOS Pat ens 8, e1002735. Neumann, P. and Carreck, N. L. (2010). Honey bee colony losses. Journal of Apicultural Research 49, 1-6. Petz, M., Stabentheiner, A. and Crailsheim, K. (2004). Respiration of individual honeybee larvae in relation to age and ambient temperature. Journal of Comparative Physiology B 174, 511518. di Prisco, G., Annoscia, D., Margiotta, M., Ferrara, R., Varricchio, P., Zanni, V., Caprio, E., Nazzi, F. and Pennacchio, F. (2016). A mutualistic symbiosis between a parasitic mite and a pathogenic virus undermines honey bee immunity and health. Proceedings of the National Academy of Sciences 113, 3203-3208. Ravi, S., Garcia, J. E., Wang, C. and Dyer, A. (2016). The answer is blowing in the wind: free flying honeybees can integrate visual and mechano-sensory inputs for making complex foraging decisions. The Journal of Experimental Biology 219 (Pt21), 3465-3472. Rothenbuhler, W. C. (1964). Behaviour genetics of nest cleaning in honey bees. I. Responses of four inbred lines to disease-killed brood. Animal Behaviour 12, 578-583. Ruchty, M., Roces, F. and Kleineidam, C. J. (2010). Detection of Minute Temperature Transients by Thermosensitive Neurons in Ants. Journal of Neurophysiology 104, 1249-1256. Sanchez, E. L. and Lagunoff, M. (2015). Viral Activation of Cellular Metabolism. Virology 479, 609-618. SAS. (2012). Version 9.4. Cary, NC: SAS Institute Inc. Schöning, C., Gisder, S., Geiselhardt, S., Kretschmann, I., Bienefeld, K., Hilker, M. and Genersch, E. (2012). Evidence for damage-dependent hygienic behaviour towards Varroa destructorparasitised brood in the western honey bee, Apis mellifera. Journal of Experimental Biology 215, 264271. Spivak, M. and Reuter, G. S. (2001). Varroa destructor Infestation in Untreated Honey Bee (Hymenoptera: Apidae) Colonies Selected for Hygienic Behavior. Journal of Economic Entomology 94, 326-331.
11 Spötter, A., Gupta, P., Mayer, M., Reinsch, N. and Bienefeld, K. (2016). Genome-Wide Association Study of a Varroa-Specific Defense Behavior in Honeybees (Apis mellifera). Journal of Heredity 107, 220-227. Stabentheiner, A. (2001). Thermoregulation of dancing bees: thoracic temperature of pollen and nectar foragers in relation to profitability of foraging and colony need. Journal of Insect Physiology 47, 385-392. Stabentheiner, A., Kovac, H. and Schmaranzer, S. (2007). Thermal Behaviour of Honeybees During Aggressive Interactions. Ethology : formerly Zeitschrift fur Tierpsychologie 113, 995-1006. Stabentheiner, A., Kovac, H. and Brodschneider, R. (2010). Honeybee Colony Thermoregulation – Regulatory Mechanisms and Contribution of Individuals in Dependence on Age, Location and Thermal Stress. PLoS ONE 5, e8967. Starks, P. T., Blackie, C. A. and Seeley, T. D. (2000). Fever in honeybee colonies. Naturwissenschaften 87, 229-231. Swanson, J. A., Torto, B., Kells, S. A., Mesce, K. A., Tumlinson, J. H. and Spivak, M. (2009). Odorants that induce hygienic behavior in honeybees: Identification of volatile compounds in chalkbrood-infected honeybee larvae. Journal of Chemical Ecology 35, 1108-1116. Traynor, K. S., Le Conte, Y. and Page, R. E. (2014). Queen and young larval pheromones impact nursing and reproductive physiology of honey bee (Apis mellifera) workers. Behavioral Ecology and Sociobiology 68, 2059-2073. van der Zee, R., Gray, A., Pisa, L. and de Rijk, T. (2015). An Observational Study of Honey Bee Colony Winter Losses and Their Association with Varroa destructor, Neonicotinoids and Other Risk Factors. PLoS ONE 10, e0131611. Wenner, A. M., Daniel, E. M. and Friesen, L. J. (1991). Recruitment, Search Behavior, and Flight Ranges of Honey Bees. American Zoologist 31, 768-782. Winston, M. L. (1987). The biology of the honey bee. Cambridge and London: Harvard University Press.
12 Figures S1: Photograph depicting the setting of the experiment to measure the temperature within varroainfested brood cells, as well as to observe hygienic behavior.
S2: Thermographic image of the experimental comb used to observe the influence of heating an area of brood on brood removal by worker bees.
Area 1 depicts the position of the heated copper plate, situated behind the comb. Area 2 indicates the position of the unheated copper plate, defining the area from which control cells were chosen. Black and blue triangles give the positions with maximum and minimum temperature within the two areas, three of which indicate the thorax of bees crawling on the comb.
S3: Thermographic image of the experimental comb used to determine the influence of narrow-scale temperature gradients on the removal of brood.
The four crosses indicate the position of cells that were heated from within through copper wires. The infested and uninfested cells for the observations were arranged around these heated cell (see figure 1).
13 Figure 1: Design of the experimental comb used for studying the effect of small-scale heat gradients on hygienic behavior 1a. position of monitored infested/uninfested cells on the comb
14 1b. setup for heating individual cells
A: camera for thermographic recording B: camera for infrared recording C: triple insulation D: copper plate E: heating foil F: wax foundation G: infested pupae H: uninfested pupae I: copper bolt, soldered to the copper plate (D) and introduced from behind into the heated cell J: controllable power supply
15 Figure 2: Temperature within infested and uninfested brood cells
The graph depicts means from seven replications of the experiment with a total of 87 individually measured cells. The overall effect of mite infestation was significant (repeated measurement-ANOVA with infestation status as covariate, P < 0.001), as was the effect of time post-infection (P < 0.001).
16 Figure 3: Effect of the distance to an individual heated cell on the kinetics of removal of infested and uninfested brood
The graph depicts the data from 4 repetitions of the experiment, with a total of 304 observed cells. Arrows indicate the survival at the end of the observations. Data were analysed by Cox-regression. Infestation status had a highly significant effect on brood removal (P < 0.001). The effect of the distance from the heated cell was also significant (P = 0.001). There was also a significant interaction between these two factors (P = 0.007), with only the combination of uninfested brood and cells directly adjacent to the heated cells increasing removal.
17 Highlights:
infestation of honeybee brood cells with Varroa destructor increased temperature of cell caps heating of entire areas of brood combs did not influence the likelihood of removal by worker bees existence of small-scale temperature gradients increased the likelihood of brood removal
PII: DOI: Reference:
S0306-4565(17)30495-3 https://doi.org/10.1016/j.jtherbio.2018.04.012 TB2102
To appear in: Journal of Thermal Biology Received date: 19 November 2017 Revised date: 22 February 2018 Accepted date: 24 April 2018 Cite this article as: Daniel Bauer, Jakob Wegener and Kaspar Bienefeld, Recognition of mite-infested brood by honeybee (Apis mellifera) workers may involve thermal sensing, Journal of Thermal Biology, https://doi.org/10.1016/j.jtherbio.2018.04.012 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1 Recognition of mite-infested brood by honeybee (Apis mellifera) workers may involve thermal sensing Daniel Bauer, Jakob Wegener, Kaspar Bienefeld* Bee Research Institute, F.-Engels-Straße 32, 16540 Hohen Neuendorf, Germany
*
Corresponding author: Tel. 0049-3303-293830; fax 0049-3303-293840
[email protected] [email protected] [email protected]
Abstract Hygienic behavior, i.e. the removal of diseased or damaged brood by worker honey bees (Apis mellifera), is seen as one of the principal behavioral elements of this species’ social immunity. Identification of the stimuli that trigger it would be helpful in searching for biochemical and molecular markers of this important breeding trait. While many studies at the genomic, transcriptomic, and behavioral level have pointed to the implication of chemical cues, we here hypothesized that thermal cues are alternatively/additionally involved. To test this hypothesis, we first measured whether infestation by the mite Varroa destructor (a condition known to induce hygienic behavior) leads to a thermal gradient between affected and unaffected brood. We found that infested brood cells were between 0.03 and 0.19 °C warmer than uninfested controls. Next, we tested whether artificially heating an area of a brood comb would increase the removal of infested or uninfested brood as compared to an unheated control area, and found that this was not the case. Finally, we investigated whether the heating of individual brood cells, as opposed to comb areas, would influence brood removal from cells adjacent to the heated one. This was the case for uninfested, though not for infested cells. We conclude that infestation by V. destructor leads to a heating of brood cells that should be perceivable by bees, and that small-scale temperature gradients can influence brood removal. This makes it appear possible that thermal cues play a role in triggering hygienic behavior of honey bees directed at varroa-infested larvae/pupae, although our results are insufficient to prove such an involvement.
Keywords: Varroa destructor; uncapping; thermography; gradient sensing 1. Introduction The ectoparasitic mite Varroa destructor is seen as the most important pest of the Western honey bee (Apis mellifera) in both economic and ecological terms (Le Conte et al., 2007; Le Conte et al.,
2 2010; Neumann and Carreck, 2010; van der Zee et al., 2015). Not surprisingly therefore, intense efforts at selecting varroa-resistant lines of honeybees are under way in many parts of the world. At the center of many of these breeding programs is “hygienic behavior” (Rothenbuhler, 1964), the capacity of worker bees to recognize and remove diseased brood. As Varroa destructor only reproduces on honeybee brood, hygienic behavior directed against varroa-infested cells greatly reduces reproductive success of the parasite, despite the fact that most mites survive brood removal and can later infest other cells (Spivak and Reuter, 2001). Much of the damage caused by varroa infestation is linked to the transmission of viruses such as deformed wing virus (DWV) (Di Prisco et al., 2016; Le Conte et al., 2010; Nazzi et al., 2012), and replication of DWV in infested brood increases the likelihood of removal of the brood concerned (Schöning et al., 2012). Varroa infests honeybee brood shortly before brood cells are sealed with wax by worker bees and bee larvae enter pupation. This means that the stimuli triggering hygienic behavior must be sensed by worker bees across the wax cap as well as the pupal cocoon. Results of several transcriptomic, genomic, and proteomic studies make it appear likely that increased hygienic behavior is linked to alterations in the capacity of bees to perceive chemical cues (Hu et al., 2016; Masterman et al., 2000; Mondet et al., 2015; Spötter et al., 2016; Swanson et al., 2009) – “hygienic” lines and individuals show an increased expression of certain odorant binding proteins, as well as of proteins involved in the processing of olfactory information. Bees selected for strong hygienic behavior were also experimentally shown to better discriminate between the odors of healthy and diseased brood (Masterman et al., 2000). Martin et al. (2002) have even identified candidate substances (mostly long-chained alkanes, alkenes or fatty acids) originating from the mites that are more strongly perceived by “hygienic” than by “non-hygienic” breeds of bees. However this does not preclude the existence of additional, nonolfactory components of the stimulus. Although honeybee biology offers examples of behaviors that seem to be triggered by chemical cues alone (e.g. the suppression of worker reproduction; Traynor et al., 2014), combinations of olfaction/gustation with other cues seems to be the more common case. Examples include the mating behavior of drones (Gary and Marston, 1971), navigation of foragers to floral resources (e.g. Ravi et al., 2016), and the recruitment of foragers by other foragers within the hive (Wenner et al., 1991). Temperature and temperature gradients are one source of information guiding honeybee behavior. The most evident illustration for this is the control of heating, ventilation and water evaporation for thermal homeostasis of the brood nest - in contrast to adult bees, honeybee brood is extremely stenothermic, and brood nest temperature is regulated by worker bees within the range of 32-36°C through both active cooling and heating (Kronenberg and Heller, 1982; Lindauer, 1954; Petz et al., 2004; Stabentheiner et al., 2010). Interactions between individual workers may equally involve thermal sensing, as hypothesized for the recruitment of new foragers to a particular source of nectar by other foragers returning from this source (Stabentheiner, 2001). Honeybee workers can be trained to associate certain temperatures with a food reward (Hammer et al., 2009), and thermal gradients as small as 0.2°C have been shown to evoke a behavioral response (Heran, 1952). Insects can use two different types of thermal cues to guide their behavior, absolute temperature and temperature gradients. Homeostasis of the honeybee brood nest temperature is an example of the former mechanism – bees heat or cool the nest without reference to a fixed external standard (Stabentheiner et al., 2010), apparently using a genetically-fixed internal standard. In Drosophila melanogaster, the biochemical basis of such an internal standard has recently been elucidated – its central elements appear to be thermosensitive transient receptor protein cation channels in the membranes of neuronal cells (Abram et al., 2016). In contrast, behaviors like thermal orientation
3 clearly require the perception of temperature gradients. Both honeybees and leafcutter ants (Atta vollenweideri) can be trained to associate temperature gradients with food rewards (Hammer et al., 2009; Kleineidam et al., 2007), and Atta-ants were shown to possess neurons that are sensitive to temperature changes below 0.1°C over a wide range of external temperature (Ruchty et al., 2010). Here we investigated whether thermal sensing is important for the triggering of Apis mellifera hygienic behavior. We started by measuring the temperature of infested an uninfested brood cells during development, in order to test the existence of a temperature difference that may serve as a discriminatory signal to hygienic workers. Next, we compared cell opening/brood removal in areas of a comb that either were or were not artificially heated. Finally, to test whether hygienic bees are exploiting small-spaced temperature gradients in the identification of diseased brood, we compared the removal of infested and non-infested brood as a function of the distance to an individual heated cell.
2. Materials and methods 2.1 Thermo-controlled hive box and colony Brood is usually situated at the center of the honeybee colony, surrounded by combs of pollen and honey, and brood outside this position is either neglected by workers or covered by thick layers of bees that ensure heating and insulation (Winston, 1987). To prevent this and enable video and thermographic observation of brood combs, an entire queenright colony of the subspecies A. m. carnica, composed of approximately 7.000 worker bees, three brood frames and three frames of food stores was therefore placed within an incubator set at 33.0°C. This temperature was chosen because it is within the range observed in the brood nest of unmanipulated colonies (Büdel, 1960), and leaves some space for biologically- or experimentally-induced increases, without running the risk of overheating brood. A flight hole was produced that allowed bees to leave and enter the incubator. External control over the hive temperature allowed us to place experimental brood combs at the border of the colony and still avoid crowding of bees on the brood. It also made observations independent of variations of outside temperature. An infrared video camera (PD/DX4-285GE; Kappa optronics, Germany; with objective Fujinon HF16HA-1B) and in some experiments also a thermographic camera (T335; FLIR systems, USA) were placed inside the incubator, directly facing the experimental comb (see photograph S1 for a depiction of the observational system). This thermocontrolled hive box was used for all three experiments.
2.2 Measurement of the temperature of varroa-infested and uninfested brood cells To test whether varroa infestation induces a change in honeybee brood temperature, a comb covered with freshly capped brood on only one side was procured from another colony. Sixteen type K thermoelements (d = 0.5 mm; l = 40 mm: precision ± 0.1 °C) were calibrated in a water bath and carefully inserted from the back into non-adjacent capped brood cells. Strong white light was used to candle the brood cells and avoid damaging the larvae within. To prevent movements of the thermoelements within the cells after introduction, the wires exiting the elements were glued to a
4 piece of plastic tubing which was mounted behind the brood comb. Next, eight of the cells containing thermoelements were artificially infested with varroa mites. These were procured from brood of the same stage, stemming from a heavily-infested colony. To introduce mites, the cell caps were carefully opened with a razor blade and the parasites placed inside using a moistened brush (Bienefeld et al., 2015). Two mites per cell were introduced in order to reduce heterogeneity of results linked to reproduction or non-reproduction of the parasites within the cells (Harbo and Harris, 2005), or transmission or non-transmission of viruses by the mites (Bowen-Walker et al., 1999). The other eight thermoelements were mounted inside control cells, i.e. capped brood cells that were treated identically to the infested cells except that no mite was introduced. The brood comb was then placed at the outside of the broodnest of the experimental colony, with the brood facing the video camera. Temperature was recorded at intervals of 10 seconds, using a data logger (SQ2040 4F16; Grant, GB). Recordings started after mite introduction and lasted for 144 h (until day 15-16 of development, counted from the moment of egglaying). Video observations of the infested cells were recorded throughout this period, in order to keep track of hygienic behavior performed on them. After the end of the observations, the state of the brood (dead/alive) and the mites (presence/absence of the introduced mite and of any offspring) were determined. The experiment was repeated 7 times. Brood cells that were emptied by the bees during the recordings were excluded from the analysis. Results from a total of 33 infested and 54 non-infested cells entered the data analysis.
2.3 Effect of heating of brood areas on the frequency of hygienic behavior As we found that infestation by varroa repeatably led to an increase of temperature of the cells concerned, we tested whether an elevation of the temperature of an area of the experimental comb would lead to an increase in the frequency of hygienic behavior. A brood comb containing freshlycapped brood (day 10 after egg-laying) on one side and mostly empty cells on the other was procured from a donor colony. The wax cells on the empty side were removed. A copper plate covering 10 x 10 cells was mounted against the back side of the brood. The plate was mounted on a copper tube that was connected to an Arctic A10 tempered bath circulator (ThermoFisher, USA), pumping tempered water resin through the tube. The temperature of the plate and brood was monitored through thermoelements that were attached to the plate, and through thermographic observations of the brood surface. In this way, temperature at the surface of capped cells was elevated by 0.5 ± 0.3 °C (calculated from randomly chosen thermograhic images from each repetition of the experiment). A second, unheated copper plate was mounted on the back side of the comb at a distance of approx. 5 cm from the heated plate, and the cells in contact with this second plate were used as the non-heated control. Nine non-adjacent cells were artificially infested with 2 varroa mites per cell within each of the two areas of brood (heated and non-heated). Hygienic behavior was monitored through infrared video observations as in experiment 1, and the fate of mites in cells that were still capped was determined after the end of observations (day 15 of development). The experiment was repeated seven times, three times with bees from a line selected for hygienic behavior and four with bees from a closely-related, but largely non-hygienic line (see Hu et al., 2016, for a description of the lines).
2.4 Effect of small-scale temperature gradients on the frequency of hygienic behavior
5 To test the hypothesis that small-scale temperature differences are involved in triggering hygienic behavior, the following experiment was performed. The setup for this experiment is shown in figure 1. On a frame of freshly-capped brood, four capped cells were chosen that defined a rectangle of approximately 8 x 6 cm. Only these four individual cells were heated from the back, by introducing copper bolts (d = 3 mm) from behind. The pupae were previously removed from the cells. To heat the bolts in a precisely controlled way, they were soldered to a copper plate (d = 0.2 mm) that was attached to the back side of the comb. This plate was heated via a heating foil (Conrad Electronic, Germany) which was glued onto it and connected to an adjustable power supply. To prevent heating of other than the manipulated brood cells by the foil/copper plate, several layers of insulating foil (TIGA-Med, Germany) were attached between the copper plate and the comb. The position of the heated cells was marked on transparent sheets placed on the comb for later identification. Around each heated cell, sixteen additional cells were marked. Of these, four were directly adjacent to the heated cell, while the twelve others were separated from it by 1, 2, or 3 unheated cells (see figure 1 for a schematic representation of the marking scheme). Evaluation of thermographic images from 3 of the 4 repetitions (see S3 for an example) show that the surface of the heated cell was approximately 1.1°C warmer than the background temperature of the comb. The gradient between the heated and the directly adjacent cell was approximately 0.7°C ± 0.4. Cells separated from the heated cell by 1, 2, or 3 other cells were 0.8, 0.9, and 1.0 degrees cooler than the heated cell. Of the sixteen cells surrounding each heated cell, eight were artificially infested with two varroa mites/cell, while the other eight were not. Another cell, positioned at a distance from the heated cells, was equipped with an unheated wire, and cells adjacent to it or separated from it by 1, 2, or 3 other cells served as controls.
2.5 Data analysis Measurements of brood temperature inside infested and uninfested cells were treated by splitting the duration of the video observations (144 h) into blocks of 12 h, for which the mean temperature within each measured cell was calculated. This condensed dataset was analysed by repeated measurement ANOVA (SAS Procedure “mixed”; SAS, 2012). Because the temperature gradient depends on whether surrounding cells were empty or containing brood, the number of empty cells adjacent to the measured cell was included in the statistical model. The ANOVA therefore included the factors “time”, “infestation” (yes/no), “experiment” (repetitions of the experiment; 1 through 7), number of empty adjacent cells (1 to 6), as well as the interaction between time and “infestation”. Data on the likelihood of brood removal in areas of heated/unheated brood (experiment 2) were analyzed using the chi²-test. For this, data from the seven repetitions of the experiment were pooled after verification that this was admissible. Data on the speed and likelihood of brood removal as a function of the distance to a single, heated cell (experiment 3) were submitted to a Cox-regression (IBM, 2011), including the factors of “infestation” (yes/no), “experiment” (repetitions of the experiment, 1 through 13), and “distance from heated cell” (1 through 4). Interactions between “infestation” and “distance” were also included.
6
3. Results 3.1 Temperature of infested and uninfested brood The average course of temperature during development of infested and uninfested brood is shown in figure 2. During all 12 time intervals, temperature was slightly higher in the infested cells. The overall effect of infestation of brood temperature was highly significant (F = 45.2; df = 1; P < 0.001), as were the effects of repetition (F = 108.4; df = 6; P < 0.001), of “number of empty adjacent cells” (F = 23.15; df = 5; P < 0.001) and of time (F = 25.44; df = 11; P < 0.001). The temperature difference due to infestation appeared to be greatest 24-72 h post-infestation, when it averaged 0.09°C, and again 120 – 144 h post-infestation, when it reached 0.17 – 0.19°C. The interaction between factors time and infestation was however not significant (P = 0.14). 3.2 Effect of artificial heating of brood areas on the likelihood of removal The frequency of removal of infested brood in areas of the comb heated or not by 0.5 – 1.0 °C was not significantly different (χ²-test; χ² = 0.30; total n = 111; P = 0.64); this is true regardless of the type of worker bees used (selected for hygienic behavior: total n = 54; χ² = 0.02; P = 0.89; unselected: total n = 57; χ² = 0.63; P = 0.43). Not surprisingly, selected bees showed a higher proportion of infested brood removal (75.9% versus 50.0%; χ² = 9.30; P = 0.002). As in the case of infested cells, uninfested cells were not more likely to be removed in warmed compared to non-warmed areas of the comb (total n = 241; χ² = 0.30; P = 0.58). Infestation had a highly significant influence on the frequency of removal (χ² = 146.73; total n = 352; P < 0.001). 3.3 Effect of narrow-spaced temperature gradients on brood removal The frequency and timing of brood removal around artificially-heated individual cells is given in figure 3. Again, there was a highly significant effect of infestation status on the course of removal (Coxregression; Wald = 30.5; df = 2; P < 0.001), with infested cells being preferentially removed. The proximity of brood to the heated cell also influenced removal (Wald = 24.1; df = 3; P < 0.001), with only cells directly neighboring the heated cells showing an increased likelihood. The effect of the proximity of the heated cells was however only measurable in the case of uninfested cells (comparison between figure 3 a. and b.), as expressed by a significant interaction-term for the factors “infestation” and “distance” (Wald = 13.5; df = 6; P = 0.04). Only one out of 17 uninfested larvae/pupae neighboring a cell containing an unheated copper wire was removed, which was significantly different from the removal rate of uninfested brood around the heated wires (13 out of 38; χ² = 5.3; P = 0.03), showing that removal was triggered by heat rather than by manipulation or the presence of the wire.
4. Discussion Our data show a slight but systematic increase of brood temperature by Varroa destructor – infestation of approximately 0.03 to 0.19°C. The size of this increase is near the resolution of the thermoelements used (0.1 °C). As however the effect was consistently observed over seven repetitions of the experiment, it seems certain that it really exists. The cause of this increase is
7 unclear. Although many insects, including honeybees, are known to adjust their body temperature in response to infection or parasitization, they are generally achieving this through behavioral changes leading to increased uptake of heat from the environment (“behavioral fever”; e.g. Abram et al., 2016; Anderson et al., 2013; Starks et al., 2000; Campbell et al., 2010; Stabentheiner et al., 2007) – a mechanism that seems unlikely in the case of Apis mellifera brood. It is known however that viral replication – which is often triggered by varroa parasitization (Bowen-Walker et al., 1999; Gisder et al., 2009) - is frequently accompanied by intensified metabolic acitivity in the host, an effect that is attributed to the manipulation of host metabolism by the virus (Sanchez and Lagunoff, 2015). An increase in metabolic heat may also result from the immune response of the host to mite infestation and increased viral replication (Nazzi et al., 2012). Viral replication is often intense already during the first 24 h after infestation by the mite (Nazzi et al., 2012), which might explain why the temperature gradient between infested an non-infested cells already showed 12-24 h after mite introduction in our experiment. Varroa-induced viral replication is known to greatly influence the probability of hygienic behavior directed against varroa-infested cells (Schöning et al., 2012). The size of the temperature increase observed is only slightly below that of the smallest temperature differences shown to induce a behavioral change in honeybees. Heran (1952) showed that bees changed their position in a continuous temperature gradient if the temperature was lowered by 0.25 °C, and Hammer et al. (2009) were able to train honeybees to distinguish sources of heat that differed by 0.4°C. It therefore seems possible that the increase in temperature linked to varroa infestation could be big enough to be perceived by worker bees. The fact that honeybees can regulate broodnest temperature despite changes of outside temperature with a precision of approximately ± 1.5°C (Büdel, 1960) shows that they are capable to sense absolute temperature without calibration by comparison to a stable external standard, and use this information for behavioral adaptations. In experiment 2, we therefore tested whether such changes of absolute temperature would influence the frequency of hygienic behavior. The results show that the heating of brood areas by 0.5 – 1.0°C, i.e. slightly more than the observed natural difference between infested and non-infested cells, does not change the frequency of removal of either uninfested or infested cells. Strictly speaking, this is no proof that absolute temperature is not an element of the stimulus leading to brood removal, because 1) due to technical limitations regarding the precision of the bath circulator and especially of the heat transfer from the copper plate to the brood, we were not able to experimentally induce precisely the same temperature increase that had been observed in experiment 1 (max. 0.19°C) in the areas of brood artificially heated in experiment 2. 2) it cannot be excluded that in infested cells, the natural increase of temperature due to the infestation already leads to the maximum behavioral response, so that a further increase is not possible. 3) in the case of uninfested cells, the lack of the olfactory components of the stimulus may have suppressed the hypothetical effect of absolute temperature. Nevertheless, the combination of results from the heated area-experiment (experiment 2) with the experiment on small-scale temperature gradients (experiment 3) make it unlikely that absolute temperature is relevant for removal, because in experiment 3, an increase of temperature of the same order of magnitude in individual, uninfested cells did increase their likelihood of brood removal. This shows that the absolute size of the temperature difference was likely not the reason
8 for non-removal in experiment 2, and that removal can also be triggered without the presence of the supposed chemical stimuli produced by infestation/viral replication. Moreover, it can also be questioned whether absolute temperature would be a sufficiently reliable signal for triggering hygienic behavior. The measurements performed during experiment 1 (supplementary materials 2 and 3) show that temperature variations within the brood cells due to brood development may be larger than the difference between infested and non-infested cells, and it is known that temperature can also vary between different areas of the A. mellifera brood nest (Büdel, 1960). Compared to absolute temperature, thermal gradients between neighboring cells can be expected to be more independent of the overall position within the broodnest. They are also less likely to be masked by differences due to the brood stage, because brood of the same stage is usually (although not always) clustered on the combs. Known thermal gradients within the brood nest include those between the bottom and rim of (empty) cells (δ = -0.1 to 1.3; Stabentheiner et al., 2010) as well as those between endothermic and ectothermic adult bees (especially their thoraces; approximately 2 – 8°C; Basile et al., 2008; Stabentheiner et al., 2010), but these do not directly concern the surface of capped brood cells, where signal pickup for hygienic behavior is likely to occur. Figure 3 shows that the removal of brood can indeed be influenced by small-spaced gradients. The fact that the stimulating effect of small-scale heating in the experiment was limited to uninfested cells may indicate that in the case of infested cells, the gradient produced through infestation is already sufficient to trigger the maximum behavioral response. Alternatively however, it may mean that nonthermal cues emanating from infested brood are sufficient to trigger the maximum behavioral response, which cannot be increased any further by an additional thermal cue. To distinguish between these two hypotheses, an experiment involving the artificial cooling of individual infested cells would be of great interest. Although small-scale heating increased the likelihood of brood removal, this increase was insufficient to close the gap between infested and non-infested brood. Theoretically, this may be due to the fact that the thermal or spatial size of the artificially-induced gradient were not quite equal to that of the natural gradient. A more likely explanation is that thermal cues, if they are at all involved in triggering removal, may only be one component of a complex stimulus involving other elements. As pointed out in the introduction, at least some of these additional elements of the removal stimulus are likely to be perceived by olfaction (Hu et al., 2016; Masterman et al., 2000; Mondet et al., 2015; Schöning et al., 2012; Spötter et al., 2016). The hypothesis of thermal cues as exclusive triggers of the removal of varroa-infested brood is also made unlikely by the steady increase of removal frequency with the developmental stage of brood (regardless of heating or infestation status; figure 3), whose kinetics differ from those of both absolute brood temperature and the size of the temperature gradient between infested and uninfested cells (figure 2). Our study focuses on the stimuli triggering the removal of brood infested by Varroa destructor. "Varroa-sensitive hygiene” (VSH; Harbo and Harris, 2005) stands out from other forms of honeybee hygienic behavior, because at the difference to many other stressors affecting the brood (such as the fungus Ascosphaera apis or the bacterium Paenibacillus larvae), varroa rarely kills its host (Rosenkranz et al., 2010). Chemical and thermal cues emitted by dead larvae/pupae can evidently be expected to be different from those resulting from sublethal damage. Our results are therefore directly relevant only for the latter case. While we showed that varroa-infestation leads to a subtle increase in temperature, death of the larva/pupa most likely will lead to a (more pronounced) cooling.
9 Identification of the complete stimulus governing honeybee hygienic behavior towards varroainfested brood could stimulate the search for genetic variants of A. mellifera that either show increased sensitivity for this stimulus, or produce a stronger stimulus if infested. We have shown here that infestation by Varroa destructor leads to a slightly increased temperature of the infested brood cells, and that bees are able to use small-spaced temperature gradients for decisions regarding the removal or non-removal of brood cells. Our results are insufficient however to conclude whether or not such thermal cues are relevant in the case of hygienic behavior towards varroa-infested larvae/pupae.
5. Acknowledgements We thank Mario Neumann, Uwe Gerber and Fred Zautke for help with IR-recordings and the construction of the experimental hive box. Funding: The work of DB was made possible by a grant from the F.-W. Schaumann-foundation. Parts of the study were additionally funded by the European Commission (FP7 KBBE program 2013.1.3-02, SmartBees, Grant Agreement Number 613960).
References Abram, P. K., Boivin, G., Moiroux, J. and Brodeur, J. (2016). Behavioural effects of temperature on ectothermic animals: unifying thermal physiology and behavioural plasticity. Biological Reviews 92, 1859-1876. Anderson, R. D., Blanford, S., Jenkins, N. E. and Thomas, M. B. (2013). Discriminating Fever Behavior in House Flies. PLoS ONE 8, e62269. Basile, R., Pirk, C. W. W. and Tautz, J. (2008). Trophallactic activities in the honeybee brood nest – Heaters get supplied with high performance fuel. Zoology 111, 433-441. Bienefeld, K., Zautke, F. and Gupta, P. (2015). A novel method for undisturbed long-term observation of honey bee (Apis mellifera) behavior – illustrated by hygienic behavior towards varroa infestation. Journal of Apicultural Research 54, 541-547. Bowen-Walker, P. L., Martin, S. J. and Gunn, A. (1999). The Transmission of Deformed Wing Virus between Honeybees (Apis melliferaL.) by the Ectoparasitic MiteVarroa jacobsoniOud. Journal of Invertebrate Pathology 73, 101-106. Büdel, A. (1960). Bienenphysik. In Biene und Bienenzucht, eds. A. Büdel and E. Herold), pp. 115-180. Munich: Ehrenwirth. Campbell, J., Kessler, B., Mayack, C. and Naug, D. (2010). Behavioural fever in infected honeybees: parasitic manipulation or coincidental benefit? Parasitology 137, 1487-1491. Gary, N. E. and Marston, J. (1971). Mating behaviour of drone honey bees with queen models (Apis mellifera L.). Animal Behaviour 19, 299-304. Gisder, S., Aumeier, P. and Genersch, E. (2009). Deformed wing virus: replication and viral load in mites (Varroa destructor). Journal of General Virology 90, 463-467. Hammer, T. J., Hata, C. and Nieh, J. C. (2009). Thermal learning in the honeybee, Apis mellifera. Journal of Experimental Biology 212, 3928-3934. Harbo, J. and Harris, J. (2005). Suppressed mite reproduction explained by the behaviour of adult bees. J Apicultural Res 44. Heran, H. (1952). Untersuchungen über den Temperatursinn der Honigbiene (Apis mellifica) unter besonderer Berücksichtigung der Wahrnehmung strahlender Wärme. Zeitschrift für vergleichende Physiologie 34, 179-206. Hu, H., Bienefeld, K., Wegener, J., Zautke, F., Hao, Y., Feng, M., Han, B., Fang, Y., Wubie, A. J. and Li, J. (2016). Proteome Analysis of the Hemolymph, Mushroom Body, and Antenna Provides
10 Novel Insight into Honeybee Resistance against Varroa Infestation. Journal of Proteome Research 15, 2841-2854. IBM. (2011). IBM SPSS statistics 20 core system user's guide. Armonk: IBM Corporation. Kleineidam, C. J., Ruchty, M., Casero-Montes, Z. A. and Roces, F. (2007). Thermal radiation as a learned orientation cue in leaf-cutting ants (Atta vollenweideri). Journal of Insect Physiology 53, 478-487. Kronenberg, F. and Heller, H. C. (1982). Colonial thermoregulation in honey bees (Apis mellifera). Journal of Comparative Physiology 148, 65-76. Le Conte, Y., de Vaublanc, G., Crauser, D., Jeanne, F., Rousselle, J.-C. and Bécard, J.-M. (2007). Honey bee colonies that have survived Varroa destructor. Apidologie 38, 566-572. Le Conte, Y., Ellis, M. and Ritter, W. (2010). Varroa mites and honey bee health: can Varroa explain part of the colony losses? Apidologie 41, 1-11. Lindauer, M. (1954). Temperaturregulierung und Wasserhaushalt im Bienenstaat. Zeitschrift für vergleichende Physiologie 36, 391-432. Martin, S. J., Elzen, P. J. and Rubink, W. R. (2002). Effect of acaricide resistance on reproductive ability of the honey bee mite Varroa destructor. Experimental and Applied Acarology 27, 195-207. Masterman, R., Smith, B. H. and Spivak, M. (2000). Brood odor discrimination abilities in hygienic honey bees (Apis mellifera L.) using proboscis extension reflex conditioning. Journal of Insect Behavior 13, 87-101. Mondet, F., Alaux, C., Severac, D., Rohmer, M., Mercer, A. R. and Le Conte, Y. (2015). Antennae hold a key to Varroa-sensitive hygiene behaviour in honey bees. Scientific Reports 5, Article nb. 10454. Nazzi, F., Brown, S. P., Annoscia, D., del Piccolo, F., Di Prisco, G., Varricchio, P., della Vedova, G., Cattonaro, F., Caprio, E. and Pennacchio, F. (2012). Synergistic parasite-pathogen ineractions mediated by host immunity can drive the collapse of honeybee colonies. PLOS Pat ens 8, e1002735. Neumann, P. and Carreck, N. L. (2010). Honey bee colony losses. Journal of Apicultural Research 49, 1-6. Petz, M., Stabentheiner, A. and Crailsheim, K. (2004). Respiration of individual honeybee larvae in relation to age and ambient temperature. Journal of Comparative Physiology B 174, 511518. di Prisco, G., Annoscia, D., Margiotta, M., Ferrara, R., Varricchio, P., Zanni, V., Caprio, E., Nazzi, F. and Pennacchio, F. (2016). A mutualistic symbiosis between a parasitic mite and a pathogenic virus undermines honey bee immunity and health. Proceedings of the National Academy of Sciences 113, 3203-3208. Ravi, S., Garcia, J. E., Wang, C. and Dyer, A. (2016). The answer is blowing in the wind: free flying honeybees can integrate visual and mechano-sensory inputs for making complex foraging decisions. The Journal of Experimental Biology 219 (Pt21), 3465-3472. Rothenbuhler, W. C. (1964). Behaviour genetics of nest cleaning in honey bees. I. Responses of four inbred lines to disease-killed brood. Animal Behaviour 12, 578-583. Ruchty, M., Roces, F. and Kleineidam, C. J. (2010). Detection of Minute Temperature Transients by Thermosensitive Neurons in Ants. Journal of Neurophysiology 104, 1249-1256. Sanchez, E. L. and Lagunoff, M. (2015). Viral Activation of Cellular Metabolism. Virology 479, 609-618. SAS. (2012). Version 9.4. Cary, NC: SAS Institute Inc. Schöning, C., Gisder, S., Geiselhardt, S., Kretschmann, I., Bienefeld, K., Hilker, M. and Genersch, E. (2012). Evidence for damage-dependent hygienic behaviour towards Varroa destructorparasitised brood in the western honey bee, Apis mellifera. Journal of Experimental Biology 215, 264271. Spivak, M. and Reuter, G. S. (2001). Varroa destructor Infestation in Untreated Honey Bee (Hymenoptera: Apidae) Colonies Selected for Hygienic Behavior. Journal of Economic Entomology 94, 326-331.
11 Spötter, A., Gupta, P., Mayer, M., Reinsch, N. and Bienefeld, K. (2016). Genome-Wide Association Study of a Varroa-Specific Defense Behavior in Honeybees (Apis mellifera). Journal of Heredity 107, 220-227. Stabentheiner, A. (2001). Thermoregulation of dancing bees: thoracic temperature of pollen and nectar foragers in relation to profitability of foraging and colony need. Journal of Insect Physiology 47, 385-392. Stabentheiner, A., Kovac, H. and Schmaranzer, S. (2007). Thermal Behaviour of Honeybees During Aggressive Interactions. Ethology : formerly Zeitschrift fur Tierpsychologie 113, 995-1006. Stabentheiner, A., Kovac, H. and Brodschneider, R. (2010). Honeybee Colony Thermoregulation – Regulatory Mechanisms and Contribution of Individuals in Dependence on Age, Location and Thermal Stress. PLoS ONE 5, e8967. Starks, P. T., Blackie, C. A. and Seeley, T. D. (2000). Fever in honeybee colonies. Naturwissenschaften 87, 229-231. Swanson, J. A., Torto, B., Kells, S. A., Mesce, K. A., Tumlinson, J. H. and Spivak, M. (2009). Odorants that induce hygienic behavior in honeybees: Identification of volatile compounds in chalkbrood-infected honeybee larvae. Journal of Chemical Ecology 35, 1108-1116. Traynor, K. S., Le Conte, Y. and Page, R. E. (2014). Queen and young larval pheromones impact nursing and reproductive physiology of honey bee (Apis mellifera) workers. Behavioral Ecology and Sociobiology 68, 2059-2073. van der Zee, R., Gray, A., Pisa, L. and de Rijk, T. (2015). An Observational Study of Honey Bee Colony Winter Losses and Their Association with Varroa destructor, Neonicotinoids and Other Risk Factors. PLoS ONE 10, e0131611. Wenner, A. M., Daniel, E. M. and Friesen, L. J. (1991). Recruitment, Search Behavior, and Flight Ranges of Honey Bees. American Zoologist 31, 768-782. Winston, M. L. (1987). The biology of the honey bee. Cambridge and London: Harvard University Press.
12 Figures S1: Photograph depicting the setting of the experiment to measure the temperature within varroainfested brood cells, as well as to observe hygienic behavior.
S2: Thermographic image of the experimental comb used to observe the influence of heating an area of brood on brood removal by worker bees.
Area 1 depicts the position of the heated copper plate, situated behind the comb. Area 2 indicates the position of the unheated copper plate, defining the area from which control cells were chosen. Black and blue triangles give the positions with maximum and minimum temperature within the two areas, three of which indicate the thorax of bees crawling on the comb.
S3: Thermographic image of the experimental comb used to determine the influence of narrow-scale temperature gradients on the removal of brood.
The four crosses indicate the position of cells that were heated from within through copper wires. The infested and uninfested cells for the observations were arranged around these heated cell (see figure 1).
13 Figure 1: Design of the experimental comb used for studying the effect of small-scale heat gradients on hygienic behavior 1a. position of monitored infested/uninfested cells on the comb
14 1b. setup for heating individual cells
A: camera for thermographic recording B: camera for infrared recording C: triple insulation D: copper plate E: heating foil F: wax foundation G: infested pupae H: uninfested pupae I: copper bolt, soldered to the copper plate (D) and introduced from behind into the heated cell J: controllable power supply
15 Figure 2: Temperature within infested and uninfested brood cells
The graph depicts means from seven replications of the experiment with a total of 87 individually measured cells. The overall effect of mite infestation was significant (repeated measurement-ANOVA with infestation status as covariate, P < 0.001), as was the effect of time post-infection (P < 0.001).
16 Figure 3: Effect of the distance to an individual heated cell on the kinetics of removal of infested and uninfested brood
The graph depicts the data from 4 repetitions of the experiment, with a total of 304 observed cells. Arrows indicate the survival at the end of the observations. Data were analysed by Cox-regression. Infestation status had a highly significant effect on brood removal (P < 0.001). The effect of the distance from the heated cell was also significant (P = 0.001). There was also a significant interaction between these two factors (P = 0.007), with only the combination of uninfested brood and cells directly adjacent to the heated cells increasing removal.
17 Highlights:
infestation of honeybee brood cells with Varroa destructor increased temperature of cell caps heating of entire areas of brood combs did not influence the likelihood of removal by worker bees existence of small-scale temperature gradients increased the likelihood of brood removal