Introduction

Honeybees collect pollen as their unique source of proteinaceous nutrients, which provide essential nitrogen for the proper physiological functioning of the colony. The quantity and quality of the pollen available in the honeybee habitat is crucial for the welfare and survival of bee populations [1].

Recent studies showed that honeybee workers preferably consume fresh pollen [2]. Since the pollen supply is irregular and greatly dependent on seasonal conditions, stored pollen is important for colony survival [3] because a shortage of pollen damages brood-rearing efficacy [4]. However, the high humidity (20–30%) of pollen collected by honeybees creates good conditions for bacterial and fungal growth [5]. Therefore, it is essential to preserve and protect pollen from microbial spoilage in the hive. A few descriptive studies have been carried out in recent decades in order to characterize the bacterial community associated with beebread [6,7,8,9] as well as the fungal composition of pollen [10,11,12] and honey [13]. However, much less is known about the functions of microbial lactic acid bacteria (LAB) associated with pollen and beebread.

The useful biochemical properties of LAB have been exploited practically since Neolithic times and are applied in various fields. For instance, LAB proteolytic and antimicrobial preservation systems are widely used in the food industry [14] and in medicine as probiotics [15]. In recent decades, endogenous and exogenous LAB also have been proposed as useful tool for control of honeybee bacterial diseases [16, 17]. It should be noted that, as confirmed by comparative genomics, particular LAB species residing in specific niches have different biochemical/metabolic activities, including their proteolytic systems [18].

As demonstrated by Gilliam et al., indigenous bacteria (reported as Bacillus species) can contribute to the efficient consumption of pollen thanks to enzymatic predigestion during beebread storage [10, 19].

Apart from proteins, pollen contains starches, whose content in pollen grains varies according to the plant species [20], ranging from 0 to 22% [21]. According to Hrassnigget al., nectar foragers use their own enzymes to degrade amylose to glucose [22]. However, similarly to their proteolytic ability, the amylolytic potential of honeybee colony LAB and its possible involvement in the decomposition of starch in pollen remain largely unknown and unexplored. Another LAB property is the antifungal action. Despite the suggestions that beebread might result from bacterial fermentation and that bees use competing flora in the pollen stores to reduce fungal and bacterial spoilage [23, 24], the hypothesis about competing flora, especially an inhibitory effect towards fungal microflora, has not been experimentally confirmed until now. Though it is well documented that at least four fungal species: Ascosphaera apis, Aspergillus spp., Nosema apis, and Nosema ceranae can infect hives and cause diseases in honeybee colony members [25].

The study objective was to reveal antifungal, proteolytic, and amylolytic properties honeybee LAB using different lab models.

We investigated the antimycotic effect of honeybee endogenous LAB on beebread fungal strains in the simulated beebread substrate (SBS). A putative functional role of honeybee LAB as pollen protein and starch processors/decomposers was studied using proteolytic and amylolytic activity detection methods. We reasoned that it would be informative to detect whether bacterial strains isolated from non-beebread but from a colony environment could grow in SBS. Therefore, apart from strains isolated from beebread, we used LAB isolated from worker honeybee crop and queen intestine, which increased the diversity of the tested bacterial species. Moreover, a recent study indicated the occurrence of the same bacterial species in different honeybee environments [26].

Material and Methods

Strains and Growth Conditions

The LAB strains (Table 1) isolated and identified during previous studies [27] were propagated in de Man, Rogosa and Sharpe (MRS) broth medium (Oxoid) supplemented with 0.1% l-cysteine and 2.0% fructose [8] prior to their use in the experiments. Fungal strains were isolated from beebread samples obtained from different combs of one honeybee colony. Beebread samples were suspended in physiological solution and spread on yeast malt (YM) agar (Difco). Plates were incubated at 34 °C for 48 h and colonies exhibiting different morphologies were purified by plating.

Table 1 LAB strains
Full size table

Identification of Fungal Isolates

Fungal genomic DNA was extracted from the mycelium of the fungal isolates grown for 3 to 7 days on malt agar using the Qiagen DNeasy Plant Mini Kit (Qiagen Ltd., Crawley, UK) according to the manufacturer’s instructions. Identification of each fungal strain was based on the sequencing of amplified products corresponding to specific genetic targets. These target genes were chosen after presumptive morphological identification based on macroscopic and microscopic observations of the isolates. For yeast isolates (BB01, BB02, BB03, BB04), the discriminant polymorphic D1-D2 region of the LSU 26S rDNA gene was targeted and amplified using the NL1 and NL4 primers [28]. For mold isolates, they were initially identified by sequencing of the internal transcribed spacer (ITS) region, including ITS1-5.8S-ITS2, using the ITS4 and ITS5 primers [29]. Then, molds belonging to the Penicillium or Aspergillus genera were further identified by partial sequencing of the β-tubulin gene using the Bt2a and Bt2b primers [30], while isolates belonging to the Cladosporium (Po3A, Po3C, Po11A, Po11B) genus were further identified by amplification and sequencing of the actin gene using the primers ACT-512F/ACT-783R [31]. Sequencing was carried out by Eurofins Genomics (Germany) using the BigDye Terminator Sequencing kit (Applied Biosystems, CA) and an ABI sequencer (Applied Biosystems). The sequences obtained were compared to the GenBank database using the Blastn program [32] with default settings.

Antifungal Activity Assay

  1. 1.

    Bacterial and Fungal Growth on Simulated Beebread Substrate

In order to perform antifungal activity assays, a specific SBS was prepared to mimic closely the natural condition of pollen fermentation while maintaining a semi-fluid state to allow aliquot sampling. Commercial polyfloral dry pollen was moistened (1/1 w/w) with a sugar solution resembling nectar (SRN), which was composed of a 20% carbohydrate mixture (glucose, fructose, and sucrose in equal amounts) dissolved in distilled water. SBS was sterilized by heat treatment at 70 °C for 15 min repeated four times, each period followed by cooling and 1 h at room temperature, and instantly used for experiments.

Prior to the experiment, 2 ml of sterile SBS was placed in 24-well cell culture flat-bottom plates (Corning Costar, Corning, NY) and 25 strains (representing 11 LAB species: Lactobacillus kunkeei, Lactobacillus sp., Lactobacillus casei, Lactobacillus kullabergensis, Lactobacillus helsingborgensis, Fructobacillus fructosus, Fructobacillus pseudoficulneus, Fructobacillus tropaeoli, Bifidobacterium asteroides, Enterococcus faecalis, Enterococcus durans) were inoculated in the substrate and incubated for 48 h at 34 °C. Their ability to grow on SBS was checked by plating serial tenfold dilutions on MRS agar plates and following incubation for 48 h at 34 °C under anaerobic conditions. The pH of SBS was also measured with a flat surface pH electrode. Fungal strains used as indicators were selected by seeding on the surface of SBS and incubation for 10 days at 34 °C. Their growth was checked visually and by plating the samples on YM agar.

  1. 2.

    Antifungal Activity Assay

As growth of the fungi differed, two different approaches were used to study their sensitivity towards honeybee LAB:

  1. (a)

    For Aspergillus niger Po1 and Zygosaccharomyces rouxii BB02strains (showing superficial visible growth), SBS was placed in 24-well plates as described above. Bacterial cell suspensions were prepared by centrifugation (10,000×g, 8 min, 4 °C) of overnight MRS broth cultures and resuspension of the pellets in 1 mM sodium phosphate buffer pH 7.0. LAB cells were inoculated in wells containing 2 ml of sterile SBS at a final concentration of 103 cells/ml (for Aspergillus niger Po1 spores/ml).

Fungal suspensions were prepared by scraping cells grown on the surface of the YM agar plate after incubation for 1 week at 34 °C and suspending them in 0.1% Tween 20 solution in distilled sterile water. Five microliters of fungal suspension (104 cell/ml) was spotted on each well over the SBS LAB culture. The growth of fungi was monitored by observing mycelium production for 2 weeks during incubation at 34 °C in aerobic conditions.

  1. (b)

    In the case of strain Candida sp. BB01, fungi and LAB cells were inoculated in SBS at an initial concentration of 103 cells/ml. The mixed culture was incubated at 34 °C for 9 days in aerobic conditions. Each day, a sample was taken for plating and CFU counting, diluted in 1 mM sodium phosphate buffer pH 7.0, and spread on potato dextrose agar (PDA) and MRS agar plates.

Plates were incubated for 48 h at 34 °C in aerobic conditions for the PDA agar plates and at 37 °C in anaerobic conditions for the MRS agar plates.

During the antifungal activity assay, three wells per combination, fungi with bacteria, bacteria and fungi alone, and sterile SBS substrate, were used as controls.

At the same time, the antifungal activity was checked using the agar overlay method as described by Rouse et al. [33]. All experiments related to the antifungal assay were performed in triplicate.

Determination of Proteolytic Activity on Ultra-Heat Treated Skimmed Milk

To examine the proteolytic activity on milk of all 25 LAB strains, the method described by El-Ghaish et al. [34] was used, with some modifications. Taking into account that these LAB strains live in fructose-rich niches (an environment containing honey), a 4% w/v mixture of equal amounts of glucose, fructose, and sucrose was added to standard ultra-heat treated (UHT) skimmed milk (Matin Léger) containing a reduced amount (0.5%) of lactose.

LAB strains were grown at 37 °C in MRS broth medium (Oxoid) supplemented with 0.1% l-cysteine and 2.0% fructose until the late exponential growth phase. Five % (v/v) of this culture was inoculated into modified UHT skimmed milk and incubated for 24 and 48 h at 37 °C. Controls were prepared by inoculation of an equivalent volume of sterile broth in modified UHT skimmed milk. To assess proteolytic activity after 24 and 48 h of incubation, samples of fermented milk were diluted tenfold with sample loading buffer and analyzed by SDS-PAGE [35].

Proteolytic activity was measured in the same manner using a casein substrate instead of UHT skimmed milk: Na-caseinate (Sigma-Aldrich GmbH, Germany) 5 mg/ml was dissolved in SRN composed of a 5% carbohydrate mixture (glucose, fructose, and sucrose in equal amounts) dissolved in distilled water. Proteolytic activity was analyzed by SDS-PAGE. Tests were carried out in triplicate.

Determination of Proteolytic Activity on the Pollen Substrate

The same LAB were checked for their proteolytic activity on the pollen substrate. Pollen was gathered manually from flowering maize tassels in paper bags. The pollen powder was stored at − 20 °C after dehydration for 24 h at room temperature [36]. Fifty milligrams per milliter of pollen was suspended in SRN composed of a 5% carbohydrate mixture (glucose, fructose, and sucrose in equal amounts) dissolved in distilled water. The substrate was sterilized by pasteurization as described above. One % w/w of overnight bacterial culture was then added, mixed, and incubated for 1 and 2 weeks. Aliquots of 200 μl of pollen suspension were diluted in an equal volume of sample loading buffer [34]. The samples were then heated in a boiling bath for 3 min and centrifuged (10,000×g, 8 min, 4 °C) before the content of supernatants was analyzed by 12% SDS PAGE.

The proteolytic strain Enterococcus faecalis HH22 [34] was used as a positive control for activity assays both in milk and on pollen. Tests were carried out in triplicate.

Amylolytic Activity Assay

The LAB strains studied were plated onto modified MRS agar (supplemented with 0.1% l-cysteine) in which glucose was replaced by 2% (w/v) soluble starch (MRS-starch medium) and incubated at 37 °C for 48 h. Starch hydrolysis was detected after treatment of the plates with iodine solution (KI, 0.6% w/v, I2 0.3% w/v) dissolved in sterile distilled water) and observation of color changes around the bacterial lawn. The ability to hydrolyze starch was also determined by inoculating the strains into MRS-starch broth. After incubation at 37 °C for 48 h, the broth pH was measured and compared to the initial pH value. Tests were carried out in triplicate.

Ethical Issues

The article complies with all points of ethical standards, since it does not describe any experiments on animals or humans concerning only the development or inhibition of microbial communities.

Results

Fungal Isolate Identification

In total, 17 different fungal isolates were collected from beebread samples.

Five fungal strains were identified as yeasts. The LSU 26S rDNA gene sequence of the BB01 strain shared 98% similarity with its closest relative Candida floricola, indicating that it could be a potential new species. In fact, according to Peterson and Kurtzman [37], unlike conspecific strains, separate species have more than 1% substitutions in the D1/D2 region. The existence of a new species should be further investigated through multiple genetic target analyses and phenotypic approaches. The same can be concluded for strain BB04, whose LSU 26S rDNA gene sequence shared 97% similarity with Lecythophora mutabilis (GenBank accession number AB100628.1). The sequences of other fungal strains had 100% similarity with corresponding species. The BB02 strain was identified as Zygosaccharomyces rouxii and BB03 as Rhodotorula slooffiae. The 13 other fungi corresponded to four Cladosporium cladosporioides strains (Po3A, Po3C, Po11A, Po11B), four Alternaria sp. strains (Po2A, Po2C, Po5, Po6), one Alternaria infectoria (Po8), one Aspergillus niger (Po1), one Penicillium oxalicum (Po4), one Aureobasidium pullulans (Po9), and one Nigrospora sphaerica (Po7). One strain was identified as bacteria Streptomyces sp. (Po10).

High morphological diversity was observed among the Cladosporium cladosporioides and Alternaria sp. strains.

Identification of the fungal strains showing details of the target genes used and similarities found with the closest relatives in GenBank are given in Table 2.

Table 2 Identification of beebread fungal isolates using specific target gene sequencing
Full size table

Antifungal Activity Assay

Only 12 out of 25 LAB strains (Table 1) were able to grow on SBS, and changing substrate pH were chosen for the antifungal activity assay: Lactobacillus kunkeei 13p, L. kunkeei 1, Fructobacillus fructosus 49a, F. fructosus 32, Fructobacillus pseudoficulneus 54, F. pseudoficulneus 57, Fructobacillus tropaeoli 21p, F. tropaeoli 46, Lactobacillus casei 45, Enterococcus faecalis 43, E. faecalis 41, and Enterococcus durans 42s.

Only three fungal strains, Candida sp. (BB01), Z. rouxii (BB02), and A. niger (Po1), yielded detectable growth on SBS and were thus used as indicator strains for the LAB antifungal activity assay.

As seen in Fig. 1, the indigenous LAB strains tested showed different abilities to inhibit A. niger (Po1) growth on the pollen substrate surface. Visible growth appeared 1 week after depositing the suspension of A. niger spores on the substrate surface. While A. niger (Po1) growth was unaffected when the substrate was preinoculated and incubated with F. fructosus 32,F. pseudoficulneus 54, F. pseudoficulneus 57, E. faecalis 43, E. faecalis 41, E. durans 42s, and L. casei 45, its growth was inhibited by L. kunkeei 13p, L. kunkeei 1p, F. fructosus 49a, F. tropaeoli 21p, and F. tropaeoli 46p. In the case of Z. rouxii (BB02), fungal growth was only strongly inhibited by the L. kunkeei 1p and 13p strains. In the case of Candida sp. BB01 strain, growth was affected in different ways depending on the presence of the LAB strains. Two patterns of replication dynamics for bacteria-yeast co-incubation were mainly observed using the plating and CFU counting technique; these are illustrated in Fig. 2. When co-cultured with LAB strains: L. kunkeei 13p, L. kunkeei 1p, F. fructosus 49A, F. tropaeoli 21P, and F. tropaeoli 46, Candida sp. BB01 strain growth was inhibited; the concentration of yeast cells did not exceed 106 CFU/ml and dropped after 72 h of incubation (Fig. 2c). When inoculated alone or co-cultured with strains: L. casei 45, F. fructosus 32, F. pseudoficulneus 54, F. pseudoficulneus 57, E. durans 42S, E. faecalis 43, and E. faecalis 41, the concentration of yeast cells reached 107 CFU/ml within 72 h and remained constant during 120 h of incubation (Fig. 2d). Two patterns of SBS pH changes were induced by the tested LAB: strains partly inhibiting growth of Candida sp. BB01 strain acidified SBS more efficiently (pH 4.3–4.7) than strains that did not acidify SBS as much (pH 5–5.7). The same LAB strains tested against the same fungal targets using the agar overlay method showed no inhibitory activity.

Fig. 1
figure 1

Inhibition by different LAB strains of development of molds—A. niger Po1 spores and Z. rouxii (BB01) cells on SBS. a A. nige r (Po1). b Z. rouxii (BB01). Co-cultured with the following LAB: 1 L. kunkeei 13P, 2 L. casei 45, 3 E. durans 42S, 4 F. fructosus 49A, 5 F. pseudoficulneus 57, 6 F. tropaeoli 21P, 7 substrate inoculated with corresponding fungi only, 8. sterile substrate

Full size image
Fig. 2
figure 2

Coincubation of Candida sp. (BB01) and different LAB in SBS. The growth dynamics patterns of separated (a, b) and simultaneous (c, d) incubation of fungi Candida sp. BB01 and LAB (L. kunkeei 13p, F. pseudoficulneus 57). The values representing mean and standard deviation are calculated from three independent experiments

Full size image

Analysis of the changes in the simulated beebread proteins with SDS PAGE showed that Candida sp. (BB01) and Z. rouxii (BB02) strains altered protein profiles, but the phenomenon was inhibited when yeasts were co-cultured with endogenous LAB (Fig. 3, lines 1 and 3).

Fig. 3
figure 3

SDS PAGE of the simulated beebread substrate inoculated with LAB and fungi. Arrows on lanes 2 (SBS inoculated with Z. rouxii BB02) and 3 (SBS inoculated with Candida sp. BB01) are showing alteration of protein profiles by fungi. Lanes 3 and 5 fungi were inoculated with L. kunkeei 13p where no protein alteration is visible. Lane 1—sterile SBS

Full size image

Proteolytic and Amylolytic Activity of LAB

No proteolytic activity was detected by the technique used. It was not observed in the pollen substrate even with highly proteolytic LAB strains used as positive controls, although they demonstrated proteolysis of other proteins used as substrates.

The strains studied did not grow or grew very poorly on the modified MRS agar plates in which glucose was replaced by starch. No change in pH was observed in the modified MRS broth either.

Discussion

The inability of the majority of fungal strains to grow on SBS indicates that they may be present in the beebread samples as sporadic microorganisms. Most of them are reported as ubiquitous organisms, some have already been isolated from nectar (Aureobasidium pullulans [38]) and pollen samples of different botanical origins (Penicillium spp., Alternaria spp., Aspergillus niger) [39]. The growth dynamics of Z. rouxii and Candida sp. in SBS suggest that they can alter the bee colony microbial environment. A recent study [40] also demonstrated the presence of obligate osmophilic fungi belonging to the genus Zygosaccharomyces in the beebread samples. The finding of A. niger in beebread may be even more important as it belongs to the genus Aspergillus, associated with the honeybee disease stonebrood [41]. The susceptibility of adult bees to aspergillosis has also been demonstrated in laboratory settings [42], and it has been reported that A. niger produces carcinogenic fumonisins [43] which has an implication also for humans, as pollen is food product for humans too.

According to results, the strain of A. niger is able to grow on SBS. The growth of this fungal pathogen in honeybee pollen stores could expose honeybee larvae to stonebrood. The LAB antifungal activities observed in the SBS experimental model used indicate highly competitive interactions between fungi and bacteria present in pollen collected and stored by honeybees. It should be mentioned that the same LAB tested against the same fungal targets using the agar overlay method showed no inhibitory activity, which demonstrates that they express this feature only in particular conditions (or in an ecological niche); it has already been hypothesized that on different food matrixes microorganisms can display different features [44]. The correlation between antifungal effect and amounts of lactic acid, acetic acid, and ethanol produced by heterofermentative fructophilic LAB was reported by Endo et al. [45]. No pronounced antifungal activity has been observed in homofermentative and “facultatively” heterofermentative LAB. Study of LAB metabolites produced in the “pollen environment” can explain nature of antifungal activities. LAB strains used in these experiments have been thoroughly checked in previous work for production of antimicrobial peptides—bacteriocins [17]. The antifungal ability of some endogenous honeybee LAB species expressed in SBS and the absence of proteolytic and amylolytic activities in these LAB suggests that they participate in the transformation of pollen into beebread and more probably in its conservation than in its predigestion [2]. Antifungal activity expressed by some particular LAB strains can be explained by the specific organic acid production profile of the tested bacteria. The lack of amylolytic activity in the LAB tested supports the opinion of Herbert et al. that the absence of starch in beebread is a result of an amylolytic enzyme added by honeybees to the pollen during its processing in the hive, as it is certainly present in fresh pollen [46].

The changes in protein composition occurring during the presence of yeasts in SBS and the inhibition of this phenomenon by endogenous LAB suggest that beebread biochemistry may not be a constant system, but dependent on its botanic origin and on a microbial competition factor. Hence, it can be concluded that endogenous LAB are adapted to this niche and that they probably have an impact on the protection against fungal flora and keeping in check pathogenic fungi such as A. niger. The results obtained highlight the importance of specific yeast and LAB in the control of the microbiological status of beebread, which is the key food that nurse bees feed to larvae. They also demonstrate the importance of studying pollen microbiota, since pollen is widely used not only as a food by humans but also in pharmaceutical preparations [47]. The potential danger to humans of mold/aflatoxin contaminating honeybee-collected pollen has recently been shown by Kostić et al. [48]. Interestingly, the Food and Agriculture Organization of the United Nations recommends the beebread conservation method [49] provided by Dany [50] who advises using dairy LAB starters to preserve pollen for human use, but the biopreservative effectiveness of the proposed exogenous LAB strains has not yet been studied.