The Response of the Honey Bee Gut Microbiota to Nosema ceranae Is Modulated by the Probiotic Pediococcus acidilactici and the Neonicotinoid Thiamethoxam

The honey bee Apis mellifera is exposed to a variety of biotic and abiotic stressors, such as the highly prevalent microsporidian parasite Nosema (Vairimorpha) ceranae and neonicotinoid insecticides. Both can affect honey bee physiology and microbial gut communities, eventually reducing its lifespan. They can also have a combined effect on the insect’s survival. The use of bacterial probiotics has been proposed to improve honey bee health, but their beneficial effect remains an open question. In the present study, western honey bees were experimentally infected with N. ceranae spores, chronically exposed to the neonicotinoid thiamethoxam, and/or supplied daily with the homofermentative bacterium Pediococcus acidilactici MA18/5M thought to improve the honey bees’ tolerance to the parasite. Deep shotgun metagenomic sequencing allowed the response of the gut microbiota to be investigated with a taxonomic resolution at the species level. All treatments induced significant changes in honey bee gut bacterial communities. Nosema ceranae infection increased the abundance of Proteus mirabilis, Frischella perrara, and Gilliamella apicola and reduced the abundance of Bifidobacterium asteroides, Fructobacillus fructosus, and Lactobacillus spp. Supplementation with P. acidilactici overturned some of these alterations, bringing back the abundance of some altered species close to the relative abundance found in the controls. Surprisingly, the exposure to thiamethoxam also restored the relative abundance of some species modulated by N. ceranae. This study shows that stressors and probiotics may have an antagonistic impact on honey bee gut bacterial communities and that P. acidilactici may have a protective effect against the dysbiosis induced by an infection with N. ceranae.


Introduction
The holobiont is defined as the functional association of a pluricellular host and its associated microbial communities [1,2].It is becoming increasingly clear that the host has a significant impact on the microbiota composition and vice versa.For instance, the gut microbiota can affect animal neurodevelopment and behavior and may even have effects on the collective behaviors of social organisms, as is suggested for honey bees [3,4].Honey bee microbiota assists its host in essential functions, including the digestion of nutrients [5].It supports general honey bee health and is involved in bee immunity, such as resistance to pathogens [6][7][8][9][10][11].
The composition of the honey bee gut microbiota is considered to be relatively simple, as it is represented by only five to nine bacterial phylotypes [8,12].The majority of the core microbiota includes two proteobacteria: Snodgrassella alvi and Gilliamella apicola, with the latter separated into two species: Gilliamella apicola and Gilliamella apis [13].It also includes the actinobacterium Bifidobacterium and two Lactobacillus-related clusters: Lactobacillus Firm-4, now referred to as Bombilactobacillus spp.(with B. mellis and B. mellifer), and Lactobacillus Firm-5, corresponding to the amended Lactobacillus genus (with L. apis, L. helsingborgensis, L. kimbladii, L. kullabergensis, and L. melliventris) [14,15].The proteobacteria Frischella perrara, Bartonella apis, and Parasaccharibacter apium, as well as a Gluconobacterrelated species group known as Alpha2.1, are also members of the gut microbiota but are less abundant in proportion and usually have a lower prevalence [8].They are considered rare members but participate in the composition of the dominant set of the honey bee gut microbiota.Although other non-core species are able to colonize the honey bee, these less frequently observed taxa are thought to appear in bees that have been stressed or highly exposed to invasive colonizers in the environment [16,17].
As a holobiont, the honey bee is exposed to a variety of biotic and abiotic stressors that affect both the insect and its gut microbiota [10,18,19].Among the biotic threats, the microsporidia Nosema ceranae is a highly prevalent intestinal parasite known to affect the gut microbiota [12,[20][21][22].This ubiquitous parasite is transmitted by its spores via the feco-oral route between bees, particularly during cleaning activities [23].Nosema ceranae infection is known to significantly increase the relative abundance of G. apicola and S. alvi and to decrease that of Alphaproteobacteria, Bifidobacterium spp., and Lactobacillus spp.Such dysbiosis could sensitize the honey bee to opportunistic pathogens [20,21].
The gut microbiota may also be impacted by abiotic threats like pesticides and especially neonicotinoid insecticides.Neonicotinoids are widely used for pest management but unfortunately also affect non-target insects like honey bees [24,25].A chronic exposure to thiacloprid transiently reduced the absolute abundance of Bombella apis and Lactobacillus Firm-5 [26].Both imidacloprid and thiamethoxam reduced the relative abundance of Lactobacillus spp.regardless of the season [22].The thiamethoxam metabolite clothianidin was found to cause dysbiosis in the gut microbiota [27].
The use of endogenous and exogenous bacterial strains as potential probiotics has been suggested as a strategy to enhance honey bee health by mitigating the detrimental effects on the gut microbiota of various pathogens, including N. ceranae [28,29], although their mode of action remains to be determined.A higher survival, a decrease in N. ceranae proliferation, and an increase in honey bee population in colonies were reported when supplementing with Protexin ® (ADM Protexin Ltd., Somerset, UK), a formulation based on Enterococcus faecium [30,31].The lactic acid bacterium Pediococcus acidilactici appeared to improve the survival of honey bees that were infected by N. ceranae and/or exposed to other compounds [32,33], suggesting that it can be used as a potential probiotic.
This study aimed to decipher the impact of N. ceranae and thiamethoxam, alone or in combination, on the honey bee gut microbiota and to test whether P. acidilactici was able to counteract some alterations.Honey bee workers were infected with N. ceranae spores, orally exposed to a low dose of thiamethoxam, and treated with P. acidilactici alone or in combination.The modulations of the gut microbiota were studied at the species level using whole-genome shotgun metagenomics.

Honey Bee Rearing and Experimental Procedures
Apis mellifera (genotype Buckfast) workers were sampled from brood frames from three different colonies (Col1, Col2, and Col3) of the same apiary (UMR 6023, Clermont Auvergne University, Clermont-Ferrand, France) in September 2021.They were anesthetized with CO 2 and cohorts of 80 worker bees were distributed in Pain-type cages containing a 5 mm piece of TempQueen to mimic five queens' mandibular pheromones (Intko Supply Ltd., Chiliwack, BC, Canada) and maintained at 35 • C. Honey bees were fed ad libitum with 50% sucrose syrup (w/v) complemented with 1% (w/v) protein supplement (Provita'Bee, ATZ Dietetic, Mas-Cabardès, France).Seven experimental conditions were tested for each colony replicate: (1) untreated control (C), (2) treatment with P. acidilactici (P), (3) chronic exposure to 1.5 ng/mL thiamethoxam (T), (4) infection by N. ceranae (N), (5) infection by N. ceranae then treatment with P. acidilactici (NP), ( 6) infection by N. ceranae then chronic exposure to 1.5 ng/mL thiamethoxam (NT), and (7) chronic exposure to 1.5 ng/mL thiamethoxam and treatment with P. acidilactici (TP).Nosema ceranae spores were freshly purified from experimentally infected honey bees using a Percoll gradient to obtain 99.9% purity [34].Nosema ceranae-infected groups were collectively infected by adding ca. 10 5 spores per bee in sugar syrup [21].When the feeders containing N. ceranae spores were empty (within 12 h), they were removed and replaced with feeders containing only sucrose syrup.The homofermentative Pediococcus acidilactici MA18/5M (Bactocell ® , Lallemand SAS, Blagnac, France) was grown overnight in MRS medium at 37 • C. The cell concentration was determined by measuring absorbance at 600 nm (absorbance value of 1 at 600 nm ≈ 1.6 × 10 8 CFU/mL) and the sucrose syrup was supplemented daily with bacteria for a final concentration of 10 4 CFU/mL.Honey bees were fed ad libitum with sucrose syrup supplemented or not with thiamethoxam and/or P. acidilactici based on their experimental condition.For the TP condition, thiamethoxam and P. acidilactici were administered in different feeders at 1.5 ng/mL and 10 4 CFU/mL, respectively.The feeders were changed and sucrose consumption and mortality were monitored daily.All treatments were repeated for the three colonies (i.e., replicates).

Sampling, DNA Purification, and Shotgun Metagenomic Sequencing
Sixteen days after treatment initiation, six whole guts, excluding crops, were dissected and pooled in 1 mL PBS containing ca. 0.3 g of 1 mm glass beads.The absence or presence of N. ceranae spores in the guts was confirmed by bright field microscopy.To reduce the quantity of eukaryotic DNA from both the host and the parasite, bacterial enrichment was performed as previously described [35].DNA was extracted using a phenol-chloroform method.DNA was quantified using Qubit TM 3.0 Fluorometer (dsDNA HS assay kit, Thermo Fisher Scientific Inc, Waltham, MA, USA) and 200 ng total DNA were randomly fragmented by sonication using a Covaris E220 (Covaris ® , Woburn, MA, USA) with a target fragment length of 500 bp.Paired-end libraries were prepared using the Illumina TruSeq ® DNA Nano library kit.Libraries were sequenced to obtain 2 × 150 bp paired-end reads using a NovaSeq 6000 apparatus (Illumina ® , San Diego, CA, USA).

Data Processing and Taxonomical Analysis
Sequencing read quality control was performed using fastp with default parameters [36].Reads from A. mellifera and N. ceranae were removed from the dataset using bowtie2 [37].We also verified the absence of N. apis by mapping all the reads on a 382 bp fragment of the N. apis gene encoding the small subunit ribosomal RNA.No read mapped this region.All the samples were randomly subsampled to 3.88 million reads based on the sample with the fewest reads.The relative abundance of microbial communities was assessed using MetaPhlAn4 (Version Jan 21) with default parameters [38].Data were filtered by keeping species that met at least one of the following criteria: (i) The species was present in at least one sample with a relative abundance higher than 0.1%, (ii) the species had a relative abundance lower than 0.1% but was present in at least half of the samples, or (iii) the species had a relative abundance lower than 0.1% but was present in all the replicates of a same condition.Significant biomarker and discriminant feature discovery was performed using LEfSe with default parameters using a linear discriminant analysis (LDA) score ≥ 2 between conditions [39].Principal component analyses (PCA), biplots, and other graphs were generated using R packages ggplot2 (Version 3.3.6),corrplot (V.0.92), FactoMineR (V.2.7), factoextra (V.1.0.7), and rgl (V.0.110.2).The raw data were deposited in NCBI Sequence Read Archive with BioProject accession number PRJNA977416.

Statistical Analysis
Statistical analysis was performed using Prism (Version 10.0.3) statistical software.Honey bee survival was analyzed using the Kaplan-Meier method followed by the log-rank Mantel-Cox test.The pairwise Mann-Whitney U test was used to compare the infection levels between experimentally infected and not experimentally infected groups.

Honey Bee Survival in Response to Treatments
Honey bee workers were infected with N. ceranae, chronically exposed to a solution of 1.5 ng/mL of thiamethoxam, and/or fed with the lactic acid bacterium P. acidilactici (10 4 CFU/mL).The previous treatments were applied alone or in combination.None of the treatments led to significant differences in syrup consumption.Honey bees consumed, on average, 368 CFU/bee/day of P. acidilactici and 0.1 ng/bee/day of thiamethoxam.
Infection with N. ceranae significantly reduced survival of the honey bees, whether they were exposed to thiamethoxam or not, compared to uninfected control worker bees (Figure 1).Thiamethoxam and P. acidilactici administered alone did not significantly affect honey bee survival.Thiamethoxam associated with P. acidilactici led to a significant decrease in survival (p-value = 0.0021).In contrast, introducing P. acidilactici to infected honey bees appeared to boost overall survival (82.2% of survival in NP versus 78.2% in N at day 16), although the difference was not statistically significant (Table S1).
Microorganisms 2024, 12, x FOR PEER REVIEW 4 of 13 relative abundance lower than 0.1% but was present in at least half of the samples, or (iii) the species had a relative abundance lower than 0.1% but was present in all the replicates of a same condition.Significant biomarker and discriminant feature discovery was performed using LEfSe with default parameters using a linear discriminant analysis (LDA) score ≥ 2 between conditions [39].Principal component analyses (PCA), biplots, and other graphs were generated using R packages ggplot2 (Version 3.3.6),corrplot (V.0.92), Fac-toMineR (V.2.7), factoextra (V.1.0.7), and rgl (V.0.110.2).The raw data were deposited in NCBI Sequence Read Archive with BioProject accession number PRJNA977416.

Statistical Analysis
Statistical analysis was performed using Prism (Version 10.0.3) statistical software.Honey bee survival was analyzed using the Kaplan-Meier method followed by the logrank Mantel-Cox test.The pairwise Mann-Whitney U test was used to compare the infection levels between experimentally infected and not experimentally infected groups.

Honey Bee Survival in Response to Treatments
Honey bee workers were infected with N. ceranae, chronically exposed to a solution of 1.5 ng/mL of thiamethoxam, and/or fed with the lactic acid bacterium P. acidilactici (10 4 CFU/mL).The previous treatments were applied alone or in combination.None of the treatments led to significant differences in syrup consumption.Honey bees consumed, on average, 368 CFU/bee/day of P. acidilactici and 0.1 ng/bee/day of thiamethoxam.
Infection with N. ceranae significantly reduced survival of the honey bees, whether they were exposed to thiamethoxam or not, compared to uninfected control worker bees (Figure 1).Thiamethoxam and P. acidilactici administered alone did not significantly affect honey bee survival.Thiamethoxam associated with P. acidilactici led to a significant decrease in survival (p-value = 0.0021).In contrast, introducing P. acidilactici to infected honey bees appeared to boost overall survival (82.2% of survival in NP versus 78.2% in N at day 16), although the difference was not statistically significant (Table S1).
PCA based on the relative abundance of bacterial species revealed a separation of the infected group (N) from the control group (C), N. ceranae and thiamethoxam (NT), and N. ceranae and P. acidilactici (NP) (Figure 3).Furthermore, the samples of group N appeared to be closer to each other compared to the other treatments, suggesting a lower within-group variability.Of note, pairwise comparisons showed no significant differences in alpha and beta diversity indices between the tested conditions.some samples, such as Arsenophonus spp., which represented 7.8% in one sample (TP2, bees co-exposed to thiamethoxam and P. acidilactici), whereas it was undetected in all other samples (mean of 0.02 ± 0.04%).
PCA based on the relative abundance of bacterial species revealed a separation of the infected group (N) from the control group (C), N. ceranae and thiamethoxam (NT), and N. ceranae and P. acidilactici (NP) (Figure 3).Furthermore, the samples of group N appeared to be closer to each other compared to the other treatments, suggesting a lower withingroup variability.Of note, pairwise comparisons showed no significant differences in alpha and beta diversity indices between the tested conditions.

N. ceranae Altered the Abundance of Some Gut Bacterial Species
PCA based on the relative abundance of species seemed to segregate control (C) and infected (N) samples, with F. perrara, G. apicola, and P. mirabilis strongly contributing to

N. ceranae Altered the Abundance of Some Gut Bacterial Species
PCA based on the relative abundance of species seemed to segregate control (C) and infected (N) samples, with F. perrara, G. apicola, and P. mirabilis strongly contributing to the clustering towards the infected group, whereas L. kimbladii, B. asteroides, and Bombilactobacillus spp.leaned toward the non-infected group (Figure 4A).LDA analysis highlighted the significantly enriched or depleted bacterial species in infected honey bee guts.The relative abundance of P. mirabilis was 20 times higher in infected honey bees (1.95 ± 1.03%) compared to uninfected ones (0.10 ± 0.03%).The abundance of G. apicola (2.02 ± 0.25% and 1.38 ± 0.18% in N and C, respectively) and F. perrara (2.75 ± 0.44% and 1.42 ± 0.89%) was also significantly increased in infected guts.In contrast, infected honey bees were significantly depleted of lactic acid bacteria (LAB), namely, B. asteroides, L. kimbladii, L. kullabergensis, L. melliventris, B. mellis, B. mellifer, and F. fructosus, with a decrease in relative abundance ranging from 2.3 times to 18.4 times compared to the control (Figure 4B).

Pediococcus acidilactici Attenuated the Modulation of Specific Lactic Acid Bacteria by Nosema ceranae
The addition of P. acidilactici to the syrup had no significant effect by itself on the composition of the gut microbiota in uninfected bees except for P. mirabilis (Table S2).In contrast, several species contributed to the separation of the infected samples (N) and the infected P. acidilactici-treated (NP) groups (Figure 5A).LDA indicated that B. asteroides, Klebsiella oxytoca, B. mellifer, and B. mellis were significantly increased when infected bees

Pediococcus acidilactici Attenuated the Modulation of Specific Lactic Acid Bacteria by Nosema ceranae
The addition of P. acidilactici to the syrup had no significant effect by itself on the composition of the gut microbiota in uninfected bees except for P. mirabilis (Table S2).In contrast, several species contributed to the separation of the infected samples (N) and the infected P. acidilactici-treated (NP) groups (Figure 5A).LDA indicated that B. asteroides, Klebsiella oxytoca, B. mellifer, and B. mellis were significantly increased when infected bees were fed with P. acidilactici (NP vs. N, Figure 5B).Two species, B. apis and Gottschalkia acidurici, were depleted (not shown for G. acidurici, as it was undetected in the NP group).The comparison of infected honey bees fed with P. acidilactici (NP) and the control group (C) showed that three species were more abundant in the NP group: P. mirabilis, G. apicola, and Morganella morganii, and one species was 14-fold less abundant: Lactobacillus kimbladii (Table S2).
were fed with P. acidilactici (NP vs. N, Figure 5B).Two species, B. apis and Gottschalkia acidurici, were depleted (not shown for G. acidurici, as it was undetected in the NP group).The comparison of infected honey bees fed with P. acidilactici (NP) and the control group (C) showed that three species were more abundant in the NP group: P. mirabilis, G. apicola, and Morganella morganii, and one species was 14-fold less abundant: Lactobacillus kimbladii (Table S2).).The values above the bars represent the average species abundance in the control honey bees relative to the relative abundance in percentage.G. acidurici is not on the graph bar because it is totally absent from the NP conditions, but it is significantly enriched in the N condition (LDA > 3).Species in bold are significantly different compared to the respective control (N vs. NP).

Thiamethoxam Also Counteracted Specific Alterations Induced by N. ceranae
The relative abundance of only one species, B. asteroides, was affected by the chronic and sublethal exposure to thiamethoxam, with a significant decrease compared to the control group (Table S2).PCA analysis showed that LAB species contributed to the separation of the N and NT groups (Figure 6A).A significant increase in the abundance of L. kullabergensis, A. kunkeii, B. mellis, B. mellifer, B. coryneforme, and B. asteroides and a significant decrease in F. perrara were observed in infected honey bees exposed to thiamethoxam (NT) compared to the unexposed group (N, Figure 6B).Interestingly, the abundance of these ).The values above the bars represent the average species abundance in the control honey bees relative to the relative abundance in percentage.G. acidurici is not on the graph bar because it is totally absent from the NP conditions, but it is significantly enriched in the N condition (LDA > 3).Species in bold are significantly different compared to the respective control (N vs. NP).

Thiamethoxam Also Counteracted Specific Alterations Induced by N. ceranae
The relative abundance of only one species, B. asteroides, was affected by the chronic and sublethal exposure to thiamethoxam, with a significant decrease compared to the control group (Table S2).PCA analysis showed that LAB species contributed to the separation of the N and NT groups (Figure 6A).A significant increase in the abundance of L. kullabergensis, A. kunkeii, B. mellis, B. mellifer, B. coryneforme, and B. asteroides and a significant decrease in F. perrara were observed in infected honey bees exposed to thiamethoxam (NT) compared to the unexposed group (N, Figure 6B).Interestingly, the abundance of these species did not significantly differ compared to the untreated control group (NT vs. C), suggesting that thiamethoxam counteracted the effect of N. ceranae for these species.Only two species showed different abundance between the infected and thiamethoxam-exposed group (NT) and the control group (C): P. mirabilis was more abundant and L. kimbladii was five times less abundant in the NT group (Table S2).
species did not significantly differ compared to the untreated control group (NT vs. C), suggesting that thiamethoxam counteracted the effect of N. ceranae for these species.Only two species showed different abundance between the infected and thiamethoxam-exposed group (NT) and the control group (C): P. mirabilis was more abundant and L. kimbladii was five times less abundant in the NT group (Table S2).

Discussion
The taxonomic assignation based on numerous genomic markers allowed for unambiguous assignments at the species level and for a more accurate estimation of relative abundance in comparison to the DNA metabarcoding approach based on 16S rRNA gene [38].For example, it allowed the response of the different Gilliamella and Lactobacillus species to be distinguished.
The taxonomic microbial composition of the untreated bees (C group) differed from most of those previously reported, which showed a large domination by Lactobacillus and Bombilactobacillus genera followed by Gilliamella spp.and Snodgrassella alvi [8].In this work, the bee guts were largely dominated by Bartonella apis.A high level of Bartonella had been reported in overwintering honey bees and was associated with a lower abundance

Discussion
The taxonomic assignation based on numerous genomic markers allowed for unambiguous assignments at the species level and for a more accurate estimation of relative abundance in comparison to the DNA metabarcoding approach based on 16S rRNA gene [38].For example, it allowed the response of the different Gilliamella and Lactobacillus species to be distinguished.
The taxonomic microbial composition of the untreated bees (C group) differed from most of those previously reported, which showed a large domination by Lactobacillus and Bombilactobacillus genera followed by Gilliamella spp.and Snodgrassella alvi [8].In this work, the bee guts were largely dominated by Bartonella apis.A high level of Bartonella had been reported in overwintering honey bees and was associated with a lower abundance of the core bacterial taxa [43,44].It has been proposed that this shift could be due to pollen shortages since B. apis is able to use alternative energetic sources such as lactate and ethanol [44].Castelli et al. have shown that nutritional stress (i.e., monofloral vs. polyfloral pollen) could also increase the relative abundance of B. apis [45].Here, honey bees were collected at the transition between the summer and winter population, which can potentially explain the high abundance of B. apis.However, a geographic and/or ecotypic variation of the gut microbiota cannot be excluded, as suggested by [46].
Infection with N. ceranae significantly decreased the survival of the honey bees; however, this varied between colonies (Figure 1 and Table S1).Previous studies have shown that N. ceranae can induce a significant decrease in survival [32,33,47] but not consistently [48,49].The presence of N. ceranae increased the relative abundance of species belonging to the Proteobacteria phylum and decreased the relative abundance of those belonging to Firmicutes and Actinobacteria in comparison to the C group.This could be seen as a complete displacement of the gut microbiota [50].
N. ceranae might indirectly impact the microbiota by changing the morphology and the homeostasis of the bee gut.For instance, F. perrara, which was significantly increased in presence of N. ceranae, is known to trigger a morphological change in the epithelial surface in the gut by inducing scab formation that could be linked to immune responses [51].N. ceranae can modulate the expression of immune-related genes, including genes coding for antimicrobial peptides (AMPs) [33, [52][53][54][55], which could eventually impact the gut microbial communities.In return, the microbiota modulations following infection by N. ceranae could affect gut functions.The core species that increased (G.apicola and S. alvi) or decreased (Lactobacillus and Bombilactobacillus spp.) in response to infection are thought to participate in digestive functions by breaking down complex sugars and secreting fermentation products [8,56].Such modifications could thus alter the metabolites available for the host as well as its sensitivity to other pathogens.For instance, some lactic acid bacteria strains can produce antimicrobial molecules [8].P. mirabilis, whose relative abundance increased in infected honey bees, is thought to have beneficial effects against Paenibacillus larvae, the causative agent of American foulbrood [57].
The supplementation of infected honey bees with P. acidilactici seemed to partially restore their gut microbiota (Figure 5B), suggesting that it can act as a potential probiotic but with limited improvement in survival (Figure 1).Although P. acidilactici impacted some bacterial species in honey bees infected with N. ceranae, P. acidilactici itself was not detected.This suggests that the bacteria did not colonize or proliferate in the gut.P. acidilactici would restore the microbiota indirectly by producing modulating compounds rather than via direct competition.In broilers, a symbiotic supplement containing P. acidilactici reduces the amount of Clostridium perfringens in the intestine and raises C. perfringens antibodies in the mucosa, which lessens the severity of necrotic enteritis [58].In addition to restoring the gut microbiota, P. acidilactici can also reestablish the expression of genes encoding catalase and glutathione peroxidase, which were found to be significantly reduced by N. ceranae [33], suggesting a potential probiotic effect on the honey bee holobiont.
Thiamethoxam also appeared to mitigate the dysbiosis caused by N. ceranae except for P. mirabilis and L. kimbladii (Figure 6B and Table S2).It has been shown that thiamethoxam has an inhibitory effect on N. ceranae proliferation in the stingless bee Melipona colimana [54].However, in A. mellifera, the effects of neonicotinoids on the parasite's success seemed variable [59][60][61].In all of these cases, the combination of neonicotinoids and Nosema led to increased mortality.Moreover, other deleterious effects of thiamethoxam on the physiology and behaviors of honey bees have been described [24,[62][63][64].Thus, even if thiamethoxam may partially restore the microbiota of infected bees, it cannot be seen as beneficial for the holobiont.
The present work shows that P. acidilactici can offset the strongly altered gut microbiota of N. ceranae-infected bees.However, it is worth noting that such a restoration of the microbiota does not necessarily infer improved honey bee health.To prove the probiotic potential of P. acidilactici, it is necessary to assess its effects on A. mellifera colonies.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms12010192/s1, Figure S1: Prevalence of Nosema ceranae with or without experimental infection; Table S1: Prevalence of Nosema ceranae with or without experimental infection; Table S2: Species showing significant difference in relative abundance between the two conditions.

Figure 1 .
Figure 1.Cumulative proportion of survival of honey bees infected by Nosema ceranae (N), exposed to thiamethoxam (T) at 1.5 ng/mL, and exposed to Pediococcus acidilactici (P), alone or in combination.The Kaplan-Meier estimator was calculated on pooled replicates (log-rank Mantel-Cox test).

Figure 1 .
Figure 1.Cumulative proportion of survival of honey bees infected by Nosema ceranae (N), exposed to thiamethoxam (T) at 1.5 ng/mL, and exposed to Pediococcus acidilactici (P), alone or in combination.The Kaplan-Meier estimator was calculated on pooled replicates (log-rank Mantel-Cox test).

Figure 4 .
Figure 4. Principal component analysis between C and N (PCA).(A) Based on the relative abundance (Metaphlan4) of the species in the gut microbiota infected by N. ceranae (N, pink) or the notinfected control group (C, black).(B) Mean log2 ratio of abundance in infected relative to control honey bees of species significantly enriched (positive values) or depleted (negative values) in response to the parasite (n = 3).The values above the bars represent the average species abundance in the control honey bees relative to the total relative abundance in percentage.Species in bold are significantly different compared to the respective control (N vs. C).

Figure 4 .
Figure 4. Principal component analysis between C and N (PCA).(A) Based on the relative abundance (Metaphlan4) of the species in the gut microbiota infected by N. ceranae (N, pink) or the not-infected control group (C, black).(B) Mean log2 ratio of abundance in infected relative to control honey bees of species significantly enriched (positive values) or depleted (negative values) in response to the parasite (n = 3).The values above the bars represent the average species abundance in the control honey bees relative to the total relative abundance in percentage.Species in bold are significantly different compared to the respective control (N vs. C).

Figure 5 .
Figure 5. Principal component analysis (PCA) between N and NP.(A) Shows the clustering based on the relative abundance (Metaphlan4) of the species in the gut microbiota between the group infection by Nosema ceranae (N, pink) versus the group infection by N. ceranae and fed with Pediococcus acidilactici (NP, green).(B) Mean log2 ratio of abundance in N. ceranae (N) and N. ceranae and P. acidilactici (NP) relative to control honey bees of species significantly enriched (positive values) or depleted (negative values) under either condition (n = 3).The values above the bars represent the average species abundance in the control honey bees relative to the relative abundance in percentage.G. acidurici is not on the graph bar because it is totally absent from the NP conditions, but it is significantly enriched in the N condition (LDA > 3).Species in bold are significantly different compared to the respective control (N vs. NP).

Figure 5 .
Figure 5. Principal component analysis (PCA) between N and NP.(A) Shows the clustering based on the relative abundance (Metaphlan4) of the species in the gut microbiota between the group infection by Nosema ceranae (N, pink) versus the group infection by N. ceranae and fed with Pediococcus acidilactici (NP, green).(B) Mean log2 ratio of abundance in N. ceranae (N) and N. ceranae and P. acidilactici (NP) relative to control honey bees of species significantly enriched (positive values) or depleted (negative values) under either condition (n = 3).The values above the bars represent the average species abundance in the control honey bees relative to the relative abundance in percentage.G. acidurici is not on the graph bar because it is totally absent from the NP conditions, but it is significantly enriched in the N condition (LDA > 3).Species in bold are significantly different compared to the respective control (N vs. NP).

Figure 6 .
Figure 6.Principal component analysis (PCA) between N and NT.(A) Shows the clustering based on the relative abundance (Metaphlan4) of the 15 species in the gut microbiota that contribute to separating the samples between the group infected by Nosema ceranae (N, pink) and the group infected and exposed to thiamethoxam (NT, cyan).(B) Mean log2 ratio of abundance in N. ceranae (N) and N. ceranae and thiamethoxam (NT) relative to control honey bees of species significantly enriched (positive values) or depleted (negative values) under either condition (n = 3).The values above the bars represent the average species abundance in the control honey bees relative to the total microbial content in percentage.Species in bold are significantly different compared to the respective control (N vs. NP).

Figure 6 .
Figure 6.Principal component analysis (PCA) between N and NT.(A) Shows the clustering based on the relative abundance (Metaphlan4) of the 15 species in the gut microbiota that contribute to separating the samples between the group infected by Nosema ceranae (N, pink) and the group infected and exposed to thiamethoxam (NT, cyan).(B) Mean log2 ratio of abundance in N. ceranae (N) and N. ceranae and thiamethoxam (NT) relative to control honey bees of species significantly enriched (positive values) or depleted (negative values) under either condition (n = 3).The values above the bars represent the average species abundance in the control honey bees relative to the total microbial content in percentage.Species in bold are significantly different compared to the respective control (N vs. NP).