Pathogen detection and gut bacteria identification in Apis cerana indica in Thailand

Pathogen infection of honeybees can lead to economic losses in apiculture. The earlier the pathogen contamination can be found the better it will likely be for the treatment of the infected colony and prevention of the spread of the pathogen within and between colonies. A total of 50 colonies of Apis cerana were sampled in Samut-Songkhram (five colonies) and Chumphon (45 colonies) provinces in the central and the south of Thailand, respectively. Diagnostic multiplex polymerase chain reaction (PCR) revealed that 20, 6, 4, 20 and 0% of the samples were infected by Paenibacillus larvae, Nosema ceranae, Nosema apis, Ascosphaera apis and Sacbrood virus (Morator aetatulus), respectively. Positive amplified PCR products of target genes were sequenced. A phylogenetic tree was constructed by the neighbor-joining (NJ) distance based, using the Phylogenetic Analysis Using Parsimony (PAUP 4.0b10) program. The phylogenetic relationship of each pathogen, based on the partial sequence of the 16S rRNA of P. larvae, N. ceranae and N. apis, and of the 5.8S rRNA of A. apis, revealed a significant difference from the non-Thai isolates of these pathogens, but no significant geographical isolation between the different Thai apiaries, although it separated some into closely related clusters. In order to reduce the use of antibiotics in an apiary, bacteria in the gut of healthy bees were focused. Interestingly, Bifidobacterium species, Lactobacillus species, Bacillus species, Lactobacillus spp. and other lactic acid bacteria, were isolated from larvae and adult workers, but gave conflicting preliminary identities based on their biochemistry-morphology versus sequence analysis of a partial fragment (1.4 kb) of their 16S rRNA.

Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License is a diverse array of pathogen-caused bee diseases, some of which can cause serious morbidity and mortality to bee colonies and lead to the economic loss to bee farmers.Thus, it would be much better if pathogen infections could be detected at an early stage.
After Apis cerana indica foragers return to a hive from foraging, they will dance to communicate about the food source (Oldroyd and Wongsiri 2006).Also, other bees inside the nest perform a cleaning behavior in order to remove contaminating objects including mites and some microorganisms in the flower pollen.However, contaminated microorganisms within the bees and in the nectar stored in forager's honey crop still remain (Anderson et al., 2013).
For decades, a diverse array of antibiotics and chemicals has been used to try to control bee disease.For example, tetracycline has been used to inhibit Paenibacillus larvae growth to limit American foulbrood (AFB) and Melissococcus plutonius growth to limit European foulbrood (EFB) (Martel et al., 2006).For microsporidia (Nosema) infections, treatment with fumagillin B has been found to be very effective, including in the control of infections by treating the hive box with fumagillin B (Higes et al., 2011).For chalkbrood disease, which is caused by Ascosphaera apis, many reagents have been used to eliminate the spores, such as sodium propionate, sorbic acid (Taber et al., 1975), citral and geraniol (Gochnauer et al., 1979).
However, the use of antibiotics and antimycotics to treat bee diseases has decreased in general due to public concern over the residual antimicrobial levels in the bee products and bacterial resistance to these chemicals (Kochansky et al., 2001).Many alternative sources were reported.For example, botanical compounds from plant extracts, such as cinnamon oil, bay oil, clove oil and origanum oil, have been used to control A. apis infections and prevent chalkbrood disease.The most effective compounds in these plant extracts (oils) were reported to be azadirachtin, thymol and α-terpinene (Calderone and Shimanuki, 1994).
In addition, three strains of Flourensia (Flourensia tortuosa, Flourensia riparia and Flourensia fiebrigii) isolated from Andiroba and Copaiba plant extracts were recently found to inhibit the in vitro growth of P. larvae, as also could the essential oils from these plants (Reyes et al., 2013).However, as an alternative approach, it is viewed that effective probiotics from the bees themselves or their products that can inhibit their own pathogens would be a better option.Prebiotics, probiotics and balance of intestinal flora in bee gut are capable of preventing and combating disease.Their imbalance results in abnormal function of the intestine and easily causes illness (Patruica and Mot, 2012).Along these lines, five strains of Bacillus and 11 phylotypes of lactic acid bacteria (LAB) from the genera Lactobacillus and Bifidobacterium, isolated from the honey and gut of Apis mellifera, were found to be able to inhibit the in vitro growth of 17 strains of P. larvae (Alippi and Reynaldi, 2006;Forsgren et al., 2010).In addition, the feeding of this LAB mixture to bee larvae caused a significant decrease in the number of infected colonies (Alippi and Reynaldi 2006;Forsgren et al., 2010).
In this research, A. cerana, which is an economic and native honey bee species to Thailand, was used.The diagnosis and distribution of some pathogens known to cause serious bee diseases were screened for using diagnostic multiplex polymerase chain reaction (PCR) and reverse transcriptase PCR (RT-PCR), and their phylogenetic relationships were evaluated.In addition, the bacteria residing in the gut of healthy bee larvae and adult workers were isolated and preliminarily identified, since they may include probiotic isolates.

Sample collection
In total, 10 apiaries were selected and were sampled in May and November, 2012.One apiary was in Samut-Songkhram province (Central Thailand) and the other nine were in Chumphon province (Southern Thailand).In each apiary, five colonies were chosen at random and sampled where, for each colony, 50 each of larvae, pupae and adults were randomly collected.However, for the analysis of gut bacteria, three colonies within the apiary in Samut-Songkhram province only were selected at random and 50 each of healthy larvae and adults were collected for analysis.

Disease diagnosis by morphology
The presence of abnormal morphological characters of larvae and adults were observed externally, including the color and body shape.Additionally, the presence or absence of Varroa mite was observed at mites adhering to the bee's body (Genersch, 2010).

DNA and RNA extraction
DNA from the larva, pupa and adults was extracted using a DNA extraction mini kit (cat.# 51304, QIAgen), while RNA was extracted from those samples using a RNeasy mini kit (cat.# 74104, QIAgen), both as per the manufacturer's instructions.Eluted samples were kept at -20°C until used.The concentration of DNA or RNA in each sample was calculated from the absorbance at 260 nm and the purity of the sample was estimated from the ratio of absorbance at 260 and 280 nm.The quality of DNA and RNA was observed by 0.8% (w/v) agarose-Tris-borate-EDTA (TBE) gel electrophoresis or 1.2% (w/w) formaldehyde/agarose gel electrophoresis, respectively, followed by ethidium bromide (EtBr) staining and ultraviolet (UV)-transillumination.

Primer design
Total DNA was amplified by multiplex PCR using the designed primers shown in Table 1.In order to detect P. larvae infection, the  (Piccini et al., 2002).DNA from P. larvae spore which was used as a positive control was a gift from Dr. P. Chantawannakul, Chiang Mai University, Thailand.In order to detect the microsporidian Nosema species infections, two pairs of primers were designed from the SSU rRNA, the first (MITOC-F3/MITOC-R3) for Nosema ceranae (accession # FJ 425736.1),and the second (APIS-F4/APIS-R4) for N. apis (accession # JX 860435.1).These were used in conjunction with the primer pair RpS5-F5/RpS5-R5 to amplify a fragment of the host ribosomal protein S5 (RpS5) (designed from A. mellifera; accession # XM 393226.3)as an internal control (Hamiduzzaman et al., 2010).For the detection of A. apis, the specific primer pair A. apis-F8/A.apis-R8 was designed from the 5.8S rRNA of A. apis (accession # HQ 905552.1)and used in conjunction with the RpS5-F5/RpS5-R5 primer pair as an internal control (James and Skinner, 2005).
For the detection of Sacbrood virus (SBV), the specific primer pair SBV-F6/SBV-R6 designed from the polyprotein (pol) of SBV (accession # HG 779873.1)was used in conjunction with the 28S-F7/28S-R7 primer pair to amplify a fragment of the 28S rRNA of A. cerana (designed from A. mellifera; accession # X 93388.1)as an internal control (Yang et al., 2012).
For the detection of P. larvae, the PCR was performed with an initial 94°C for 1 min followed by 35 cycles of 94°C for 1 min, 45°C for 1 min and 64°C for 2 min before a final 72°C for 7 min.For Nosema, the PCR conditions were an initial 94°C for 2.5 min, 10 cycles of 94°C for 15 s, 55°C for 30 s and 72°C for 45 s, followed by 20 cycles of 94°C for 15 s, 50°C for 30 s and 72°C for 50 s, and then a final 72°C for 7 min.For the detection of A. apis, the PCR was performed at 94°C for 1 min, followed by 30 cycles of 94°C for 45 s, 45°C for 1 min and 72°C for 1 min, and then a final 72°C for 7 min.After PCR amplification, all PCR products were observed following resolution by 0.8% (w/v) agarose gel electrophoresis, EtBr staining and UV-transillumination.

RT-PCR amplification
Since SBV is a RNA virus, it was detected via RT-PCR.Reactions were performed in a final volume of 25 µl composed of 12.5 µl of Emerald Amp GT PCR master mix (2x Premix), 5 µl of RNA template (200 ng), 400 nM of each primer, 0.5 µl of 200 µM dNTP mix and 0.5 µl of AMV Reverse Transcriptase (cat.# A1250, Promega).The thermal cycling was performed at an initial 48°C for 45 min for RT-mediated cDNA synthesis, and then PCR amplified at 94°C for 2 min followed by 30 cycles of 93°C for 1 min, 50°C for 30 s and 72°C for 1 min before a final 72°C for 7 min.After the amplification was done, the RT-PCR products were observed by 0.8% (w/v) agarose-TBE gel electrophoresis, EtBr staining and UVtransillumination.

Phylogenetic analysis
The targeted PCR product was purified using a PCR product purification kit (cat.# 28104, Qiagen) and then commercially direct sequenced by AIT biotech (Singapore).The obtained nucleotide sequence was then searched for similar sequences in the NCBI GenBank database using the BLASTn algorithm.The obtained nucleotide sequences and the similar ones from GenBank along with the selected outgroup sequences were aligned using the Clustal X program.
The phylogenetic relationship among the different isolates of each pathogen was analyzed by the neighbor-joining (NJ) distance based, using the Phylogenetic Analysis Using Parsimony (PAUP 4.0b10) program (Swofford 2000).The support for the nodes was evaluated by bootstrap analysis with 1,000 times.

Bacterial culture
The isolation of gut bacteria from healthy appearing larvae or adult A. cerana bees followed the method of Yoshiyama and Kimura (2009).In brief, the larval body and the dissected gut of adult bees were surface sterilized by immersion into 10 ml of 70% (v/v) ethanol for 1 min, followed by 10 ml of 6% (w/v) sodium hypochlorite (NaOCl) for 1 min, and then, rinsed with 10 ml of sterile distilled water for 1 min.After that, either three larvae or the guts from three healthy bees were homogenized in 700 µl of phosphate buffer saline (PBS) pH 7.2, and then clarified by centrifugation at 2,817× g for 1 min at room temperature.The supernatant was harvested, serially diluted and each serial dilution was spread at 200 µl per plate onto a duplicate set of triplicate Brain Heart Infusion (BHI) agar (cat.# 241830, Difco) plates.The first set of triplicate plates was then cultured for aerobic bacteria at 37°C for 18 to 24 h, whilst the second set of triplicate plates was cultured for anaerobic bacteria by incubation in an anaerobic jar with a gas pak (cat.# AN 25US, Oxoid) at 37°C for four to five days.

Bacterial colony morphology
The colony morphology, in terms of the color, luster and smooth, convex or jagged edges, was observed and recorded.The number of colonies (Nc) was counted from three randomly selected 1 cm 2 boxes on each 63.6 cm 2 plate and the colony forming units (CFUs)/ml was then evaluated as: CFU/ml = [(Nc) × (the area of plate (cm 2 )) × (the serial dilution factor) / 0.6].
In addition, Gram staining of bacteria from representative colonies of which each with different morphology was performed.

Biochemical characterization (substrate fermentation) of bacterial colonies
The ability of three representative colonies from each different morphological bacterial isolate (section 2.5.2) to ferment different substrates was observed using the API 20A test (cat.# 20300, Biomerieux) for anaerobic bacteria and the API 50 CH test (cat.# 50300, Biomerieux) for aerobic bacteria.

Sequence analysis of the 16S rRNA gene
For representative isolates from each morphologically different colony type (section 2.5.3),colony PCR was performed.A single colony was picked into 20 µl of 10 mM TE buffer (pH 7.5) and subjected to three cycles of -20°C for 30 min followed by incubated 96°C for 5 min and vortexing.A 2 µl aliquot (prediluted to ~100 ng/µl) was then PCR amplified using the eu27F (5'-GAGAGTTTGATGCTGGCTCAG -3') and eu1495R (5'-CTACGGCTACCTTGTTACGA -3') primer pair (Yoshiyama and Kimura, 2009) in a final volume of 25 µl containing 12.5 µl of Emerald Amp GT PCR master mix (2x Premix), 8.5 µl of nuclease free distilled water and 1 µl of 10 µM of each primer.The PCR was performed by an initial 95°C for 1 min, followed by 35 cycles of 95°C for 1 min, 55°C for 1 min and 72°C for 1 min, and then a final 72°C for 7 min.After PCR, the amplified product was checked by 0.8% (w/v) agarose-TBE gel electrophoresis with EtBr staining and UV-transillumination.The expected size of PCR product was about 1.4 kb.

Bee morphology
The morphology of randomly sampled bee larvae, pupae and adults from 50 colonies was observed.In overall, an abnormal black and brown body color was found in a large number of larvae, whilst the pupa was a more prevalent infection target for mites.In contrast, neither mites (Varroa or any species, parasitic or also phoretic) nor an abnormal body color was observed in the sampled adults.

Multiplex PCR or RT-PCR
After each multiplex PCR, the control (host gene amplicon) bands of the expected size (Table 1) were observed in each case (Figure S1 in supplementary information).Interestingly, polymorphism of detected pathogens could be observed.The percentage of infected bees could then be estimated (Table 2).From the five colonies (one apiary) sampled at Samut-Songkhram province (Central Thailand), P. larvae, N. ceranae and N. apis infections were found in 2 (40%), 3 (60%) and 1 (20%) of the colonies, respectively, although even in the infected colonies, they were at a low frequency of individuals (1.5 to 2%).However, no A. apis or SBV infections were found.For the 45 colonies (five apiaries) sampled in Chumphon province (Southern Thailand), P. larvae, N. apis and A. apis infections were detected in 8 (17.8%), 1 (2.2%) and 10 (22.2%) of the colonies, respectively, although these also were at a typically low prevalence in each infected colony (0 to 5.5%).However, no N. ceranae infections or SBV infections were found, and so no SBV infections were detected in this study at all.

Similarity of partial DNA sequence
The partial sequences of the obtained 16S rRNA gene fragments obtained from the PCR amplifications for the 10 isolates of P. larvae, 3 isolates of N. ceranae and 2 isolates of N. apis, and that for the 5.8S rRNA gene of the 10 isolates of A. apis were searched against the GenBank database.The highest DNA identities to AY030079.1 for P. larvae, KC708007.1 for N. ceranae, U76706.1 for N. apis and HQ905552.1 for A. apis with 99 to 100% sequence identity.This supports the likely correct PCR amplification and ascribed infection status.

Phylogenetic tree construction
The sequences of the amplified pathogen gene fragment (16S or 5.8S rRNA), along with the most identical sequences from GenBank and that for the selected outgroups, were aligned using Clustal, manually checked and after removal of ambiguous aligned bases, were used to construct the respective pathogen phylogenetic trees by the NJ.
The aligned 497 bp fragments of the 16S rRNA gene from P. larvae from the infected bees in both Samut-Songkhram and Chumphon provinces revealed 74 differences of which 5 sites were parsimonious informative.Using the sequences from P. montaniterrae (accession code AB295646.1),P. larvae_Germany (accession code DQ079621.1),P. larvae_Japan (accession code NR_112166.1),P. larvae_Canada (accession code AF478106.1)and P. larvae_USA (accession code NC_022577.1)as outgroups, the obtained phylogeny revealed that the Thai P. larvae sequences were closely related to the recorded Paenibacillus genus from elsewhere (Figure 1), and all together formed six closely related clusters (Clusters I to VI).The samples from Samut-Songkhram fell into clusters III and IV whilst those from Chumphon fell into their own groups (clusters I, II, V and VI).The isolates of P. larvae in clusters I to VI originated from bees in different geographical areas, where the dominant plant species also differed, being palm, rambutan, longan, pineapple, coconut, and rubber trees, respectively.However, these six clusters are neither strongly divergent, nor are they well supported by bootstrap analysis (50 to 78% bootstrap support).Thus, P. larvae are not resolved by this gene fragment into separate geographical regions, which may reflect its recent pandemic spread or lack of geographic isolation.
With respect to the 16S rRNA gene fragment sequences of N. ceranae, the 202 bp contained 21 different sites of which 0 were parsimonious informative.Using N. bombycis (accession code AB093018.2) and N. ceranae_Indonesia (accession and A. apis_Canada (accession code U18362.1) as the outgroup, the obtained phylogeny revealed that the sequences of the 10 isolates from Chumphon province are as shown in Figure 4.The isolates from Chumphon province fell into four distinct groups (Clusters I-IV).
Cluster I contained the sequence of A. apis infecting bees found in palm trees in the La-mae district, whilst clusters II, III and IV contained A. apis isolates from bees found in rubber, cashew nut and mango trees, respectively.However, the relevance of the tree-geographic location upon the A. apis genotype and population genetics is not addressed due to insufficient samples to allow any meaningful inferences.

Gut bacteria in healthy bees
The potential use of natural probiotics as an alternative way to control bee pathogens is increasingly interesting challenging (Evans and Lopez, 2004).Here, the aerobic and anaerobic bacteria in the guts of healthy (asymptomatic) A. cerana bees were isolated and then characterized by their colony morphology, biochemistry (Gram stain and substrate fermentation) and analysis of the partial 16S rRNA gene sequence (Table 3).For the 16S rRNA sequence analysis, the nucleotide identity of the obtained sequences to the recorded ones in GenBank ranged between 90 and 99% over the 981 bp region examined.
However, the identification of just over half (10/17; ~59%) of the colonies was clearly different at the genus and higher taxonomic levels between the sequence analysis and the morphological-biochemical analysis, although it seemed that molecular analysis was more reliable.Nonetheless, the same class of taxonomy could be obtained by both methods.

DISCUSSION
The deleterious bee diseases that are frequently found in apiaries are AFB caused by P. larvae, microsporidosis caused by N. ceranae and N. apis, chalkbrood disease caused by A. apis, and sacbrood disease caused by the SBV.In order to detect these pathogen infections in bee colonies at an early stage prior to the onset of acute or more severe chronic symptoms, PCR (and RT-PCR for RNA viruses) have been used to detect low (asymptomatic) pathogen levels as well as to confirm the diagnosis and infection levels in both the bees and in their products, such as honey.Nowadays, several diverse applications of PCR have been introduced to bee farm management.For example, the detection of SBV contamination in A. mellifera adults using a one-step RT-PCR loop-mediated isothermal amplification primer set (RT-LAMP), where the SBV-specific primers were designed based on the SBV pol gene sequence.This method is simple, rapid and can be visualized by a simple color change from orange to green (Yang et al., 2012).Furthermore, a portable RT-PCR thermal cycler has been successfully developed to work in situ in apiaries that can be applied to detect many different types of bee pathogens in less than 8 min (Han et al., 2008).
Distribution and screening of pathogens in A. mellifera have been reported in many countries but very few in Thailand.However, no report in Thailand was focused on A. cerana although they are economically important as well.In this research, apiaries in Samut-Songkhram province and Chumphon province were the target.An apiary in Samut-Songkhram province is a natural apiary.A bee farmer has never used any antibiotics or chemicals at all so we used A. cerana in this location as control.In contrast, for the rest of apiaries, bee apiaries in Chumphon province are managed.Antibiotics or chemicals have been used.In addition, A. cerana are very popular for bee farmers in Chumphon province which led us to collect samples easier.In this study of Thai, A. cerana indica apiaries by diagnostic multiplex PCR, the number of colonies infected with P. larvae, N. ceranae, N. apis and A. apis were 4, 6, 2 and 0%, respectively, in Samut-Songkhram province and 16, 0, 2 and 20%, respectively, in Chumphon province.No SBV infections were found in all samples from either province.Thus, the numbers of infected colonies and individual bees in infected colonies found in this research was low, and lower than that which reported 91.1 and 57.1% P. larvae infected A. mellifera honey and colonies, respectively (Lauro et al., 2003).Whilst it is possible that the pathogen susceptibility depends on the bee species, it is also likely to depend upon local population genetics and demography of hostparasite interactions.Whilst A. mellifera is very good at foraging and can be managed well in an apiary, it has a relatively weak disease resistance.In contrast, the native (to Thailand) A. cerana is less efficient at foraging and less easy to manage in an apiary since it absconds from the allocated hive easily, but it is very tolerant to diseases, which might reflect that it has co-evolved to the local pathogens (Chen et al., 2000).
Varroa mites can cause a serious problem to bee colonies and can also transmit or enhance susceptibility to other pathogens.Of the seven known pathogenic bee viruses, five were detected in just 20 Varroa destructor mites by RT-PCR (Chantawannakul et al., 2006).These were SBV, Kashmir bee virus (KBV), acute paralysis virus (APV), deformed wing virus (DWV) and black queen cell virus (BQCV).Thus, the transfer of these viruses between colonies and bee species is possible.Subject to the limitation of the very small sample size of infected individuals and geographically sampled locations, N. ceranae and N. apis appeared allopatric being restricted to the central and southern region, respectively.Although it is possible that the developmental stage of the bee, in terms of larva, pupa and adult worker, was also an important determinant in the type of infection found, that is the different bee life stage have different susceptibilities, or survival rates or immunity to the different pathogens, the number of infected individuals of each developmental stage was too low in this study to allow any such reliable analysis.Moreover, the infection prevalence may depend on the geographic region.For example, a higher Nosema species infection level was previously reported for bees from the northern part of Thailand than in this research.Investigation of the N. ceranae and N. apis infection level in Apis dorsata, A. cerana, Apis florea and A. mellifera in apiaries and forests in northern Thailand revealed that 77.5% of A. mellifera were infected by N. ceranae compared to 22.2, 45.4 and 37.5% in the native honeybees A. cerana, A. florea and A. dorsata, respectively (Chaimanee et al., 2010).However, infection by N. apis was not found.In general, microsporidia infections (Nosema species) in bees are a severe global problem.High infection frequencies of N. ceranae in A. cerana have been found in China (61%), Taiwan (73%), Japan (75%), Indonesia (2.63%), Vietnam (2.63%),South Korea (2.63%) and the Solomon islands (21%), respectively (Chen et al., 2009;Botías et al., 2012).
The phylogenetic analysis of bee pathogens has been investigated in several regions.For example, the genetic et al., 2003).However, these Thai A. apis isolates formed six different clusters (I to VI), with cluster III the most prevalent, and so the potential differential resistance to treatment of these A. apis isolates may be of some concern.
In this study, no SBV infections were detected across all the sampled bees, which contrasted to other studies in different bee species or geographical regions.For example, the low prevalence (0.90%) of detected SBV infections in A. mellifera in Germany (Hedtke et al., 2011), and the very high SBV infection frequencies in A. cerana in China (Ai et al., 2012), at 86% for adults, 33% for pupae and 79% for larvae, and in South Korea at 60.5% in larvae and 30.7% in adults (Choe et al., 2012).
It is also possible that the infection level depends on the bee developmental stage, but the number of infected bees in this study was too low to allow any meaningful discrimination of the infection frequency in different developmental life stage at each geographic location.With this caveat in mind, in this study, the frequency of infection by Nosema was slightly lower in larvae than in adults, A. apis infection was more prevalent in larvae and pupae than in adults and infection by P. larvae was the highest in the pupae.Elsewhere, as already mentioned earlier, SBV infection frequencies in A. cerana in China was higher in adults (86%) and then larvae (79%) than in pupae (33%) (Ai et al., 2012), whilst in South Korea the SBV infection frequency in A. cerana was also higher in larvae (60.5%) than in adults (30.7%) (Choe et al., 2012).Furthermore, pupae of A. cerana were highly susceptible to infection by P. larvae vegetative cells, whilst 1-d-old larvae were the most susceptible to infection by P. larvae spores (Chen et al., 2000).
The severity of pathogen infection has been reported to depend on the bee larval instar and form of the pathogen (Bamrick, 1967).Spores of P. larvae could infect 100 and 90% of intermediate (21 h old) and older (45 h old) A. mellifera larvae, respectively, but only 60% of younger (3 h old) larvae.In contrast, the vegetative cells of P. larvae could highly infect younger (77.5%) and intermediate (75%) larvae to a higher frequency than older larvae (55%).Moreover, 1-day-old larva could be infected by about 10 P. larvae spores compared to the millions of spores needed to infect 2-day-old larva (Alippi and Aguilar 1998).
In order to find a safe way to control pathogen infections in bees, the use of probiotics is an interesting but challenging possibility that is gaining interest.Thus, the cultivable aerobic and (facultative) anaerobic bacteria from the gut of healthy bees (larvae and adult workers) were isolated and characterized by the colony morphology, biochemistry (substrate fermentation ability and Gram stain) and analysis of a 16S rRNA fragment sequence.The discrepancy between the genus, and sometimes even higher taxonomic levels, classification of just over half (10/17) of the isolates between the biochemical and molecular classification is of concern and awaits resolution, but it is likely that Bacillus species were the most often obtained isolates.For example, from the molecular based classification of the 17 different morphologically discriminated colonies, six were Bacillus spp.(35.3%) (1/4 in larvae and 5/13 in adult workers) and they accounted for 44.5% of all CFUs (36.8 and 45.9% in larvae and adult workers, respectively).This finding agrees with Yoshiyama and Kimura (2009) who found mostly Bacillus spp.from the gut of A. cerana japonica, although Wu et al. (2013) found 11 different strains of Bifidobacterium species from the gut of A. cerana japonica.
In addition, members of the Firmicute, Actinobacteria, Alphaproteobacteria and Gammaproteobacteria have been isolated from the gut of A. cerana cerana in China (Jiang et al., 2013).In related bee species, 15 strains of Bacillus were isolated from A. mellifera gut and honey (Sabaté et al., 2009), whilst Bifidobacterium, Lactobacillus, Gluconacetobacter, Simonsiella, Serratia and Bartonella were isolated from the gut of worker A. mellifera capensis and A. mellifera scutellata bees (Jeyaprakash et al., 2003).Overall, broadly similar bacterial families have been isolated from different honey bee species.Although previously, the bacteria from the gut of bees has been found to exert antimicrobial activity against P. larvae (Evans and Armstrong, 2006) and antibiotic resistant bacteria (Aween et al., 2012), the potential probiotic activity, if any, of these 17 isolates from A. cerana indica in Thailand remains to be established.

Figure 1 .
Figure 1.The NJ tree of the aligned 497 bp nucleotide data set of the 16S rRNA gene fragment sequences from P. larvae.The 16S rRNA gene sequence of P. montaniterrae, P. larvae_Germany, P. larvae_Japan, P. larvae_Canada and P. larvae_USA were used as the outgroups for the analysis.CP and SS were abbreviated from Chumphon and Samut-Songkhram.The number under branches indicates the percentage of bootstrap support from 1,000 replications.Scale bar represents substitution or site.

Figure 2 .
Figure 2. The NJ tree of the aligned 202 bp nucleotide data set of the 16S rRNA gene sequences from N. ceranae.The 16S rRNA gene sequence of N. bombycis, N. ceranae_Indonesia, N. ceranae_Argentina, N. ceranae_USA, N. ceranae_Poland, N. ceranae_Australia, N. ceranae_France, N. ceranae_Germany, N. ceranae_Hiroshima, N. ceranae_Mie, N. ceranae_Saga and N. ceranae_Spain were used as the outgroups for the analysis.SS was abbreviated from Samut-Songkhram.The number under branches indicates the percentage of bootstrap support from 1,000 replications.Scale bar represents substitution or site.

Figure 3 .
Figure 3.The NJ tree of the aligned 251 bp nucleotide data set of the 16S rRNA gene sequences from N. apis.The 16S rRNA gene sequence of Encephalitozoon cuniculi, N. apis_Canada, N. apis_USA, N. apis_Spain, N. apis_Australia, N. apis_Tasmania, N. apis_Nova Scotia, N. apis_Poland, N. apis_Island, N. apis_Queensland, N. apis_Ireland, N. apis_UK and N. apis_China were used as outgroups for the analysis.CP and SS were abbreviated from Chumphon and Samut-Songkhram.The number under branches indicates the percentage of bootstrap supported with 1,000 replications.Scale bar represents substitution or site.

Figure 4 .
Figure 4.The NJ tree of the aligned 338 bp nucleotide data set of 5.8S rRNA gene sequences from A. apis and related species.The 5.8S rRNA gene sequence of Beauveria bassiana, A. apis_California, A. apis_Belgium, As. apis Denmark, A. apis_USA and A. apis_Canada were used as outgroups for the analysis.CP was abbreviated from Chumphon.The number under branches indicates the percentage of bootstrap support with 1,000 replications.Scale bar represents substitution or site.

Table 1 .
Primers used for the multiplex PCR or RT-PCR(*) detection of pathogens.

Table 2 .
Prevalence of pathogen infections in A. cerana in Thailand.