Effect of Immune Inducers on Nosema ceranae Multiplication and Their Impact on Honey Bee (Apis mellifera L.) Survivorship and Behaviors

Simple Summary Nosema disease of honey bees is caused by the fungus Nosema ceranae, which multiplies and damages cells lining the digestive tract, impairing food digestion and debilitating the bees. Current control involves using antibiotics, which is undesirable because of possible antibiotic resistance of the fungus and contamination of honey. In this study, the natural compounds flagellin, zymosan, chitosan and peptidoglycan were investigated as alternatives for controlling Nosema ceranae infections and for their effect on bee survivorship and behaviors. Chitosan and peptidoglycan reduced infection and increased survivorship of infected bees. However, neither compound altered the bees’ hygienic behavior, which was also not affected by the infection. Chitosan treated bees collected more pollen and nectar than healthy and infected bees. Memory in the bees was temporarily impaired by chitosan but was not affected by peptidoglycan, nor was it affected by Nosema ceranae. This study shows that chitosan and peptidoglycan provide benefits by partially reducing Nosema ceranae infection while increasing survivorship of honey bees. Also, chitosan and peptidoglycan increased the collection of pollen and nectar, which may improve bee health and colony productivity. These benefits could result in more honey produced, more crops pollinated and more healthy bee colonies. Abstract Nosema disease is a major disease of honey bees caused by two species of microsporidia, Nosema apis and N. ceranae. Current control involves using antibiotics, which is undesirable because of possible antibiotic resistance and contamination. In this study, flagellin, zymosan, chitosan, and peptidoglycan were investigated as alternatives for controlling N. ceranae infections and for their effect on bee survivorship and behaviors. Chitosan and peptidoglycan significantly reduced the infection, and significantly increased survivorship of infected bees, with chitosan being more effective. However, neither compound altered the bees’ hygienic behavior, which was also not affected by the infection. Chitosan significantly increased pollen foraging and both compounds significantly increased non-pollen foraging compared to healthy and infected bees. Memory retention, evaluated with the proboscis extension reflex assay, was temporarily impaired by chitosan but was not affected by peptidoglycan, nor was it affected by N. ceranae infection compared to the non-infected bees. This study indicates that chitosan and peptidoglycan provide benefits by partially reducing N. ceranae spore numbers while increasing survivorship compared to N. ceranae infected bees. Also, chitosan and peptidoglycan improved aspects of foraging behavior even more than in healthy bees, showing that they may act as stimulators of important honey bee behaviors.


Treatment of Bees with PAMPs and N. ceranae
Newly-emerged bees were randomly assigned to treatments and individually marked on the thorax with different colors of enamel paint, according to treatment. Subsequently, the marked bees were individually force fed 10 µL of solutions using a micropipetter (2-20 µL, Eppendorf, Westbury, NY, USA) containing either 50% sucrose syrup (negative control), 50% sucrose syrup containing 10 4 spores/µL of N. ceranae (positive control), or 50% sucrose syrup containing 10 4 spores/µL of N. ceranae plus 0.125 µg/µL zymosan, 1.0 µg/µL peptidoglycan, 1.0 µg/µL chitosan, or 0.1 µg/µL flagellin. The PAMP doses were chosen based on the following studies. The peptidoglycan dose was from Evans and Lopez [22] who fed it in sugar syrup to induce innate immune related genes in A. mellifera, the zymosan dose was based on Tanji et al. [28], who fed it to flesh flies, Sarcophaga peregrine, in sugar syrup to induce innate immunity, the chitosan dose was from Hashimoto et al. [29], who injected chitosan in sterile dH 2 O to induce innate immune related genes in the eri-silkworm, Samia cynthia ricini, and the flagellin dose was from Samakovlis et al. [24], who added it to a cell line culture of Drosophila melanogaster to induce an immune related gene.
After feeding, each bee was transferred to a screened hoarding cage (10 × 15 × 15 cm) until 60 bees of the same treatment had been introduced. Each cage was provided with two gravity feeders, one containing sterile dH 2 O and the other 50% sucrose syrup for the bees to feed ad libitum. The cages of all treatments were kept in an incubator at 32 • C and 60% RH for 20 days post treatment (dpt). Each treatment was tested with three biological replications.

N. ceranae Infection Intensity
N. ceranae infections were determined by randomly collecting 10 bees from each cage at 10 or 20 dpt as described above. Two samples from each macerate were analyzed and the numbers of N. ceranae spores/bee were averaged. After this experiment, the two PAMPs that significantly restrained the proliferation of N. ceranae spores, were selected for PER, survivorship, and behavioral testing in observation hives.

Observation Hives
Three observation hives were installed to study survivorship, hygienic, and foraging behavior of the experimental bees. Each observation hive (47.0 × 4.1 × 96.5 cm) was placed in a separate windowless room and was connected to a runway ramp (65 × 10 × 2 cm), which in turn was connected to an opening to allow the bees to forage normally outside the building. The ramp was covered with glass to allow the observation of foraging behavior. The runway was designed so that outgoing bees were deterred by a barrier from leaving the hive via the return lane [30]. Four newly-drawn combs with about 4000 worker bees of variable ages and a queen unrelated to the experimental bees were installed into each hive. Bees were emerged and treated as above, except for the N. ceranae + zymosan and N. ceranae + flagellin treatments because these PAMPs were less able to reduce spore loads in the treated bees of the cage experiments. A total of 500 bees per each of the four treatments were introduced into each observation hive (2000 experimental bees). The bees of the four treatments were co-fostered in common hives, which is a way of controlling environmental influences and to better separate the effects of the different treatments [31]. For hygienic behavior observations, a small door (10 × 10 cm) was installed at the bottom of the hive over the plexiglass to facilitate introducing a comb section containing frozen brood at the bottom of the lower frame.

Survivorship
The survivorship of the experimental bees in the observation hives was determined as per Unger and Guzman-Novoa [32] by performing visual censuses of the bees of each treatment during early mornings when bees were not foraging. To determine acceptance, the first census was done 24 h after introducing the bees to each of the observation hives and then continued daily from day 8 to 31. The survivorship of the bees for each treatment was estimated by calculating the proportion of bees present in the hive per census day over the initial number of accepted bees.

Hygienic Behavior
Hygienic behavior was evaluated by introducing to the hive a 4 × 4 cm section of a comb containing freeze-killed capped brood. The comb of the lower frame of each observation hive had an open square area of 4 × 4 cm that had been cut out using a knife. This open area faced the 10 × 10 cm lower plexiglass door of the observation hive, through which the comb section containing the frozen brood was introduced and fixed to the comb using toothpicks.
Two hygienic tasks of the bees, cell uncapping and brood removal, were recorded during 4 h of observation per day and three repetitions of the assay were conducted during three days, when the bees were 12, 14, and 15 days old. The proportion of bees of each treatment participating in hygienic tasks were calculated by dividing the number of bees uncapping cells or removing brood over the number of bees remaining alive for each treatment during the days of observation.

Foraging Behavior
Foraging behavior of the bees was measured for 2 h in the early morning and 2 h in late afternoon, each of three days per week on weeks 2-4 post-introduction to the observation hives. Foraging was assessed by recording the number of bees of each treatment that entered the observation hive through the ramp. Bees were classified as carrying or not carrying pollen on their pollen baskets. The proportion of bees participating in each foraging task (foraging trips with or without pollen) for each of the treatments, was calculated. This calculation was performed by dividing the number of forager bees with or without pollen over the number of bees remaining alive for each treatment per observation day.
After four weeks, when both hygienic and foraging assessments were completed, the remaining live marked bees were collected from the observation hives and frozen. Then, 30 randomly chosen bees from each treatment were sampled and N. ceranae infection intensity was determined by counting spores as described previously, to ensure that infections had developed in the experimental bees.

Proboscis Extension Reflex (PER)
Memory retention was assessed with the PER assay by individually feeding newly emerged bees with the treatments as described previously, then transferring 80 into a screened hoarding cage (10 × 15 × 15 cm). Bees were provided with sterile dH 2 O and 50% sucrose syrup ad libitum and were kept in an incubator at 32°C and 60% RH for up to 20 days.
The PER test was performed with 13-to 19-day-old bees following standard protocols [33]. On the first day of evaluation, each bee was manually removed from its cage, and transferred to a glass vial (1 × 4 cm), which was then capped. The vial was placed on ice for 3 min to reduce the bee's motion to Insects 2020, 11, 572 5 of 14 facilitate its manipulation. Then, to immobilize the bee, its abdomen was introduced into a plastic tube (6 × 3 cm) of a stationery marker and wrapped with Parafilm (American National Can, Neenah, WI, USA) around its thorax; the head of the bee was free to move. Groups of 10 to 15 bees held in tubes placed in a plastic rack in an up-standing position were permitted to rest during 24 h before performing the PER test, and were fed three times per day with 10 µL of 50% sucrose syrup.
Experimental bees were trained to memorize the odor of clove oil (Sigma). A 30-mL syringe containing a piece of filter paper (Whatman No. 1) soaked with 20 µL of clove oil was used to project scented air to the bee held in the plastic tube at 5 cm from the syringe tip. The syringe plunger was slowly pushed to project the clove oil scent towards the bee. To make the scent flow, an aluminum foil ventilation pipe (10 × 100 cm) was placed about 1 cm behind the immobilized bee and a fan (Hawaiian Breeze, Canadian Tire, Guelph, ON, Canada) was installed with its rear end facing the opposite edge of the pipe so that the air was exhausted out of the testing arena. PER involved training the bees three times, followed by three PER measurements and ending with a sugar response (SR) evaluation. Training was done as per Denker et al. [34]. PER measurements were performed at 2, 24, and 48 h after the last training to evaluate short, medium, and long-term memory, respectively [33]. Tests comprised only the exposure to the scent for 5 s without touching the bee's antennae and without feeding. If the bee extended its proboscis, the event was recorded as positive. Bees were fed daily 2 h before or after the tests. Immediately after the 48 h PER measurement, a SR evaluation was done by touching the antenna of the bee with a toothpick containing a small drop of 50% sucrose syrup. Results for bees with no ability to extend the proboscis during the SR evaluation were discarded from the analyses. A minimum of 50 bees were required to pass the SR evaluation for each of the treatments. Three replications were performed.

Statistical Analyses
Percentage or proportion data from survivorship, hygienic behavior and foraging behavior were arcsine square-root transformed to normalize their distributions before analyzing them. The transformed data were then subjected to two-way analyses of variance (ANOVA). N. ceranae spore counts data were square-root transformed before being analyzed by repeated-measures ANOVA. Means were compared with Fisher Protected LSD tests when significance was detected. The proportion of surviving bees at 30 dpt in the observation hives as well as the PER data were subjected to contingency table analysis with Chi square tests. These analyses were performed using R 3.3.1 (R Development Core Team, Auckland, New Zealand).

Bee Survivorship
Significant effects for treatment and age were detected when the proportions of surviving bees at the sampled days were analyzed (F3, 200 = 18.80, p < 0.0001 and F24, 200 = 44.31, p < 0.0001, respectively; Figure 1). No noticeable differences in the proportion of surviving bees between the treatments was evident up to 21 dpt, but a higher proportion of the negative control bees remained alive between 22 and 31 dpt followed by N. ceranae + chitosan, N. ceranae + peptidoglycan and N. ceranae + no treatment. By 31 dpt, the proportion of live bees for non-inoculated control, N. ceranae + chitosan, N. ceranae + peptidoglycan and N. ceranae + no treatment, was 0.41, 0.22, 0.16, and 0.12, respectively, which were each significantly different from each other (p < 0.05 based on χ 2 tests). Thus, while chitosan and peptidoglycan treatments both reduced spore counts to levels not significantly different from each other (Table 1), chitosan gave a greater improvement to survivorship (Figure 1).

Hygienic Behavior
No differences between non-inoculated control, N. ceranae + no treatment, chitosan or peptidoglycan treatments were found for the proportion of bees conducting cell uncapping (F 3, 32 = 0.23, p > 0.05; Figure 2A) or brood removal (F 3, 32 = 0.45, p > 0.05; Figure 2B). However, there were effects of colony with bees in colony one uncapping and removing significantly more brood than bees in the other two colonies (F 2, 33 = 70.38, p < 0.0001, F 2, 33 = 26.74, p < 0.0001, respectively). No treatment × colony interactions were detected for any of the two hygienic behavior variables (p > 0.05).

Hygienic Behavior
No differences between non-inoculated control, N. ceranae + no treatment, chitosan or peptidoglycan treatments were found for the proportion of bees conducting cell uncapping (F3, 32 = 0.23, p > 0.05; Figure 2A) or brood removal (F3, 32 = 0.45, p > 0.05; Figure 2B). However, there were effects of colony with bees in colony one uncapping and removing significantly more brood than bees in the other two colonies (F2, 33

Memory Retention
The only differences between treatments for the percentage of bees that responded positively during the PER assay was for a lower percentage for the N. ceranae + chitosan treatment at 2 and 24 h (Table 2). However, at 48 h, no differences were detected for any of the treatments.

Discussion
This study showed that chitosan and peptidoglycan provide benefits by partially reducing N. ceranae spore numbers while increasing survivorship compared to N. ceranae infected bees. Also, chitosan and peptidoglycan improved aspects of foraging behavior even more than in healthy bees, showing that they may act as stimulators of important honey bee behaviors.
N. ceranae-infected bees lived significantly less time than non-infected bees. Decreased survival of bees caused by N. ceranae infection in this study agrees with several other studies conducted with individual bees in cages [4,[35][36][37] or colonies of bees [38][39][40]. Bees infected with N. ceranae live less time than non-infected bees probably because the fungus impairs their digestive functions and consequently affects nutrient absorption. Hence, infected bees would gain less energy from nutrients compared to healthy bees [36] and may eventually die of starvation and lack of energy. Furthermore, Mayack et al. [41] suggested that N. ceranae causes significant regulatory changes in the highly conserved octopamine hormone stress pathway, affecting the bee's life span and possibly their age polyethism. This study showed that both chitosan and peptidoglycan reduced N. ceranae spore counts by more than 60%, which confirms a previous report on chitosan's partial suppression of N. ceranae infections in honey bees [42]. However, although both PAMPs similarly reduced N. ceranae spore counts, chitosan was more effective in increasing survivorship of N. ceranae infected bees. Perhaps chitosan has other beneficial effects for the bees reflected in survivorship. For example, chitosan treatment of bees can reduce oxidative stress [43].
One effect of infection by N. ceranae appears to be the suppression of the immune system of honey bees. For example, expression of the immune related gene, hymenoptaecin, has been reported to be down-regulated at 3 and 6 dpi [44], 2 dpi [45] and 5 dpi [25]. The beneficial effects of 1.0 µg/µL chitosan and peptidoglycan on honey bees may be related to changes in gene expression. For bees inoculated with N. ceranae, expression of hymenoptaecin was significantly higher with peptidoglycan and chitosan than non-treated control bees at 5 dpt. Expression of defensin 2 in bees inoculated with N. ceranae was significantly higher with peptidoglycan and chitosan compared to non-treated control bees by 1 and 2 dpt, respectively. For blue cheese, significantly higher expression was detected in N. ceranae infected bees by 1 dpt with chitosan and 10 dpt with peptidoglycan compared to infected non-treated bees. Thus, one effect of these treatments may be to reverse the suppression of the honey bee immune system by N. ceranae.
This is the first study to evaluate the hygienic behavior of N. ceranae infected bees, as well as the effect of chitosan and peptidoglycan treatments on this behavior. Hygienic behavior provides resistance to honey bees against several brood diseases and to the parasitic mite V. destructor [46][47][48][49]. Thus, any pathogen that could impair the ability of honey bees to defend themselves against brood diseases and parasites would have a negative effect on honey bee colonies. However, no significant effect of N. ceranae infections or chitosan and peptidoglycan treatments on the two variables of hygienic behavior analyzed in this study was detected. By the age when the bees were evaluated for hygienic behavior, the impact of N. ceranae infections may not have yet been significant on bee health as it takes at least two weeks for spore counts to reach high levels [50]. Thus, testing older bees might have shown an effect. It was noteworthy that neither chitosan nor peptidoglycan had any negative effect on the hygienic behavior of the bees. Thus, their use would not affect this important behavior.
For foraging behavior, N. ceranae infection did not affect either the proportion of bees that performed pollen or non-pollen foraging trips. Conversely, Kralj and Fuchs [10] reported a reduced number of foraging trips for N. ceranae infected bees compared to non-infected bees in one of their experiments, which was believed to be due to impaired orientation skills of N. ceranaeand N. apis-infected foragers compared to healthy bees. However, their results were inconsistent because in a second identical experiment, they did not find differences in foraging trips between infected and non-infected bees. In another study, Alaux et al. [39] found that Nosema infections did not reduce foraging behavior in young bees, but it was reduced in 20-day-old bees or older. One difference between the studies is that this study calculated foraging relative to the number of bees that were alive at observation times, whereas Alaux et al. [39] and Kralj and Fuchs [10] counted the number of live bees relative to the starting number of bees. Thus, studies using older bees may have reported differences in foraging activity because the number of infected bees at observation times was lower than that of non-inoculated bees. Another factor that may affect the impact of N. ceranae infection on foraging behavior is the inoculum dose. Odemer et al. [51] found that infection increased foraging behavior, but they fed a lower dose (1.4 × 10 4 spores of N. ceranae per bee) than the 1.0 × 10 5 spores per bee used in this study, as well as Alaux et al. [39] who used 2.0 × 10 4 spores per bee and Kralj and Fuchs [10] who used 1.2 × 10 5 spores with bulk feeding. Thus, a number of factors, such as bee age and pathogen dose, could be responsible for the differences in results. This emphasizes the need for standardized conditions to be able to compare studies of how Nosema infections affect the foraging behavior of honey bees.
This study is the first to find a positive effect of two natural compounds on a behavior of honey bees, indicating that they can have effects independent of their ability to restrain the proliferation of Nosema spores. Chitosan treatment increased both the number of pollen and non-pollen foraging trips, while peptidoglycan significantly increased only non-pollen foraging trips. Increased foraging activity due to chitosan and peptidoglycan might be beneficial because colonies would have more food stores for winter survival and honey production. The number of foraging trips has been shown to be directly related to honey production and overwintering survival of honey bee colonies [30,52]. However, further investigation is warranted to evaluate whether chitosan and peptidoglycan treatments can affect honey production and overwintering survival. Future studies should also test the compounds without N. ceranae infection to determine if they would also increase foraging.
The results from the PER assays in this work are consistent with those of two previous studies showing no effects of N. apis or N. ceranae infections on honey bee learning and memory using the PER assay [16,53]. Conversely, Gage et al. [54] showed that N. ceranae negatively impacts PER in honey bees. While Gage et al. [54] used the same level of inoculum, they tested 7-to 15-day-old bees versus 13-to 19-day-old bees in this study. Thus, the age of the bees could be having an impact on the results, warranting further investigation.
While there were no differences at 48 h, chitosan had a negative effect on PER at 2 and 24 h post training. In contrast, peptidoglycan never significantly affected bee memory. The age of the bees first evaluated by the PER assay was similar to that of bees assessed for foraging behavior during the first days of the observation trials. Thus, the negative effects of chitosan versus the lack of an effect of peptidoglycan on bee memory did not appear to be reflected in chitosan and peptidoglycan both improving aspects of foraging. Even though there is a correlation between learning ability and foraging behavior of honey bees [55], it may be that any impairment of learning by chitosan measured with the PER assay in this study was not sufficient to negatively affect aspects of learning related to foraging after chitosan treatment. One possibility is that chitosan negatively affected PER in the short term because chitosan can be a central nervous system depressant, slowing down brain activity in vertebrates [56]. Whether there are any long-term effects of that on behaviors related to learning remains to be investigated.
While PAMPs are widely studied for their effects on infections, this study demonstrates that they can also affect the learning ability and memory retention of bees as well as foraging behaviors of honey bees. Considering the importance of those functions for honey bees, it is key to study such compounds for the benefit of honey bees aside for their potential to control N. ceranae. While the level of control of the pathogen was limited, further work on these and other natural compounds, combinations, doses, formulations, and other factors may be able to increase the level of N. ceranae control. Examples of other compounds that could be tested include β-glucans, which are immune-modulator molecules that have been shown to reduce viral loads and increase survival rate of honey bees parasitized with Varroa destructor [57]. Natural compounds also have the benefit of reducing the risks of antibiotic honey contamination and development of antibiotic resistance.

Conclusions
This study showed that chitosan can reduce N. ceranae spore multiplication, improve survivorship of infected bees, and increase both pollen and non-pollen foraging without adversely affecting hygienic behavior. However, chitosan caused a significant decrease in memory retention for the first 24 h of the PER tests, although at 48 h no effect was detected. By comparison, peptidoglycan is also promising in that it achieved similar results, except for not improving survivorship of N. ceranae-infected bees and not increasing pollen foraging. However, peptidoglycan had the advantage of not showing a negative effect on memory retention. Although successful, the reduction in spore numbers with either compound is limited. A question is whether these compounds can be made to work better, such as by providing them continuously, rather than only one initial feeding, and testing at a colony level, where conditions are more difficult to control. Funding: This study was funded in part by a High Quality Personnel grant from the agreement between the University of Guelph and the Ontario Ministry of Agriculture, Food and Rural Affairs of Canada.