*3.5. Antagonistic E*ff*ect of Probiotic Cells against the Biofilm-Forming, S. aureus*

Next, it was hypothesized that the probiotic cells (from the mutual biofilm) could antagonize pathogenic bacteria, for instance, *S. aureus*, which is known as a robust biofilm-forming bacterium, especially a submerged type of biofilm. It was consequently found that the conditioning supernatant (CSN) obtained during the growth of the probiotic cells strongly inhibited biofilm formation by *S. aureus* (Figure 7A). This result indicates that there might be an induction in producing an antimicrobial substance(s) during the generation of the mutual biofilm. Interestingly, it seems that a major impact of this inhibitory effect was related to *B. subtilis* cells, although there was a modest contribution by the cells of *L. plantarum*. A further quantitation of the surface-adhered cells confirmed a potent inhibitory effect of the CSN against biofilm formation by pathogenic *S. aureus* (Figure 7B). Moreover, the microscopic visualization of the augmented biofilm phenotypes confirmed once again the anti-staphylococcal properties of the CSN (Figure 7C). Importantly, it was further confirmed that the CSN did not cause

growth inhibition of *S. aureus* (Figure 7D), which indicates the biofilm-specific mode of action of CNS against this pathogenic bacterium.

**Figure 6.** Bio-coating maintains LAB survivability during gastrointestinal digestion in vitro following freeze-drying. Mono- and dual-species cultures of *B. subtilis* and LAB cells were generated in MMRS medium during bacterial growth for 8 h in 37 ◦C, 50 rpm. Survival rates of LAB cells were determined based on CFU counts following freeze-drying and during gastro-intestinal digestion in vitro. \* *p* < 0.05 for comparison of the control and tested samples. Error bars represent standard deviation (SD). BFD—before freeze-drying; AFD—after freeze-drying; GP—gastric phase; IP—intestinal phase. The survival rates are shown for (**A**) *L. plantarum*, (**B**) *P. acidilactici*, and (**C**) *L. rhamnosus*.

It was further hypothesized that the growth media would affect the ability of the CSN to inhibit *S. aureus* biofilm formation. Therefore, the antibiofilm activity of the CNS produced by *B. subtilis* cells in the MMRS medium was compared with that produced in LB medium. Apparently, growth of the *Bacillus* cells in the MMRS induced the antibiofilm effect of the CSN (Figure 8). Thus, a significantly higher inhibition on *S. aureus* biofilm formation was found by the CSN from the MMRS medium compared to that produced in the LB medium. Accordingly, the CSN from MMRS medium showed around a three-log reduction in the *S. aureus* adherence onto the surface compared to that from LB (Figure 8A). The inhibitory effect of the CSN was further confirmed microscopically by testing a submerged biofilm of *S. aureus* cells using live–dead staining (Figure 8B).

**Figure 8.** Type of growth medium governs the inhibitory effect on the *S. aureus* biofilm formation. (**A**) quantification of the *S. aureus* cells attached to the surface using CFU method, and (**B**) the CLSM imaging of the *S. aureus* biofilm formation in the presence or absence of the CSN, following growth in TSB medium with 10% supernatant at 37 ◦C for 24 h. \* *p* < 0.05 for comparison of the CSN from LB to control; \*\* *p* < 0.05 for comparison of CSN from MMRS vs. LB. Error bars represent standard deviation (SD). Live cells are stained green (SYTO-9) and dead cells are stained red (propidium iodide). Scale bar = 50 μm.

According to recent findings, *B. subtilis* could affect *S. aureus* biofilm formation via signaling interference [4]. It was, thus, hypothesized that the production of either fengycins [4] or surfactin [37] could explain the antibiofilm activity of the CSN produced by *B. subtilis.* Both factors are produced by *B. subtilis*, and they could affect the *S. aureus* cells through interfering with inter- or intra-cellular signaling. It was subsequently found that expression of the genes encoding for these factors by *B. subtilis* was notably upregulated in the MMRS compered to LB medium (three- and two-fold induction in the expressions of *fenA* and *srfA*, respectively) (Figure 9). This result suggests the involvement of the regulatory pathways associated with those genes in the observed antibiofilm phenotype.

**Figure 9.** Relative expression of *B. subtilis* genes related to antagonistic activity during growth in the different media. The real-time (RT)-PCR analysis was performed for quantitation of *fenA* and *srfA* gene expression in *B. subtilis* cells grown in either LB or MMRS medium as described in the Methods. \* *p* < 0.05 compared to control. Error bars represent standard deviation (SD).

#### **4. Discussion**

The importance of healthy commensal microbiota for the mammalian host is evident; thus, there is widespread use of probiotics for preventing and treating various health problems in humans, as well as in animals. Nonetheless, maximizing the survivability of probiotic cells during their formulation, as well delivery, remains a significant challenge. Probiotic bacteria are supposed to go through a long route starting with processing, through shelf life and the passage through the GIT, which includes dealing with extreme conditions [38]. Since these processes may affect cell survivability, an effective way(s) of delivering probiotic bacteria to the mammalian host would be highly useful.

It is now well established that biofilm formation represents one of the most favorable microbial lifestyles within often challenging natural environments [39]. The biofilm provides bacterial cells protection against challenging environmental conditions such as changes in shear forces, extreme temperatures, desiccation, extreme pH, and antimicrobial agents [40–42]. It was, therefore, proposed to generate a protective bio-coating system for probiotic cells for a possibility of a more efficient delivery to the mammalian host. The most appropriate candidate for this mission appeared to be the robust biofilm-forming *B. subtilis*, since it naturally colonizes the mammalian gut and is considered to be harmless to mammals including humans [19,20]. It was confirmed that there are no antagonistic interactions during the generation of this complex multispecies system; thus, no antagonism was observed between *B. subtilis* cells and LAB, or during the formation of symbiotic biofilm bundles through inducing the expression of *tapA* operon (involved in the matrix production) by *B. subtilis*. It should be emphasized that the tested LAB strains belong to different genera with a different origin. Nonetheless, it was possible to generate cooperating and protected multispecies biofilms, which indicates the feasibility of using this bio-coating system for a wide range of probiotic species.

As suggested throughout the study, the generated bio-coating system increased the LAB survivability during drying processes, which points to the feasibility of using this system for processing probiotic cells for further food or biotechnological applications. The drying process is commonly used as a means for storage and distribution, which lowers the expense and inconvenience of using a cool chain. Although water is essential for bacteria living and the drying processes damage cell structure and viability, the long-time preservation and retention of viability during storage is often enhanced by lowering the water activity [34]. The robust biofilm matrix, produced by *B. subtilis*, contains polysaccharides (PS) that presumably have an important role in protecting bacteria during drying processes. The PS could provide bacteria a hydrated microenvironment. Thus, through the drying process, the PS layers may serve as a barrier on the cell surface and prevent water removal [18,43]. In the case of *L. rhamnosus*, which could not survive the 40 h of desiccation, the bio-coating process enabled a very significant increase in survivability during the desiccation process. This finding highly suggests the possibility of protecting desiccation-sensitive probiotic cells using this bio-coating system.

Freeze-drying is the most common drying method for long-term preservation of microorganisms, in the microbiological industry, thanks to optimal protection of cell viability [36]. Usually, before the freeze-drying process, protective agents like skim milk, sucrose, or other sugar types are added to the drying medium to prevent cell damage during the drying process and storage of freeze-dried cells [44]. Some studies showed that biofilm PS can also be used as a protective agent [43,45], which is in agreement with the findings of current study. We observed higher survivability during freeze-drying for *L. plantarum* and *L. rhamnosus* cells following their growth through the bio-coating system. However, we did not observe a significant increase in survival rates for *P. acidilactici* following freeze-drying. One of the possible explanations for this result could be related to the possible resistance of this environmental isolate to desiccation stress due to its adaptation to the udder environment (from where it was isolated).

In addition to the protective capability, *B. subtilis* demonstrated potent antimicrobial activity against pathogenic *S. aureus*. This result was not surprising since *B. subtilis* was recently explored for its probiotic functionality on many levels. It was shown that *B. subtilis* could stimulate an immune response in humans, as well as maintain a favorable balanced microbiota, and decrease infection and diarrhea via

the synthesis of antimicrobial agents [46]. Production of antimicrobial agents is one of the antagonistic properties of probiotic bacteria, and indeed *B. subtilis* produces a wide diversity of substances, which influence a broad spectrum of pathogens via different mechanisms [27]. Several studies suggested either the growth inhibition or depression of *Staphylococcal* virulence by the antagonistic activity of *B. subtilis* [4,47]. In agreement with the literature, the current study showed that most of the antimicrobial activity of the multispecies biofilm system was due to substances produced by *B. subtilis*. Importantly, this activity was specific to the mitigation of biofilm formation by *S. aureus* rather than inhibition of its growth. Notably, the relatively modest effect of the CSN derived from *L. plantarum* on biofilm formation by *S. aureus* cells might still have an important role in mitigating this problematic pathogen. The synergistic activity of substances produced by different probiotic species might provide a further antimicrobial effect against such persistent pathogens.

Findings of this study further indicated that the inhibitory effect of the CSN is associated with the production of secondary metabolites, for instance, lipopeptides by *B. subtilis*. In this regard, the inhibitory effect could be related to their chemical structure [30] or to their ability to inhibit quorum sensing [4]. Lipopeptides produced by *B. subtilis* function firstly as quorum-sensing interrupters (fengycins), by inhibiting the quorum-sensing regulatory system [4], and surfactin, by regulating the autoinducer-2 (AI-2) activity [38]. Thus, these lipopeptides could inhibit quorum sensing via a different mechanism. The other antimicrobial mode of action of lipopeptides is related to the similarity of their chemical structure to surface-active agents, which might impair the ability of cells to attach to the surface and form a biofilm structure [30]. According to the results presented in this study, it appears that the growth of *B. subtilis* in the MMRS medium triggers the production of antimicrobial lipopeptides. This finding is indeed conceivable since the MMRS medium notably induces biofilm formation in *B. subtilis* [33], which is highly related to the production of antimicrobial lipopeptides [32].

Taken together, the data shown in this study suggest of the robust probiotic functionality of *B. subtilis* (i) in protecting the probiotic LAB during their exposure to unfavorable environmental conditions, such as desiccation and acid stresses, and (ii) strong anti-biofilm activity against pathogenic bacteria such as *S. aureus*.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-2607/7/10/407/s1: Figure S1: The CSLM images of LAB cells following their monoculture growth; Figure S2: Growth in dual-species biofilm does not change medium acidification rate by LAB cells; Table S1: Bacterial strains used in this study; Table S2: Primers used for RT-PCR analyses.

**Author Contributions:** Conceptualization was originated by M.S. and developed together with H.K.; developing a methodology during the investigation and the data analysis accomplished by H.K. and M.S.; the experiments were performed by H.K; the writing, review & editing of the manuscript accomplished by H.K and M.S.; the supervision and funding acquisition were done by M.S. All authors approved the final version of the manuscript.

**Funding:** This work was partially supported by the Nitzan Grant No. 4210376 of the Chief Scientist of the Ministry of Agriculture and Rural Development (Israel). Hadar Kimelman is a recipient of an Excellence Scholarship for M.Sc. students granted by the Agricultural Research Organization, as well as by The Hebrew University in Jerusalem.

**Acknowledgments:** This study forms part of Hadar Kimelman's M.Sc. project. We would like to acknowledge Doron Steinberg (from the Hebrew University of Jerusalem) for helpful discussions. We also would like to thank Yulia Kroupitski, Yigal Achmon, Carmel Hutchings, and Bat-Chen Cohen (from the Shemesh lab, ARO), Shamay Yaacoby (from the Department of Ruminant Science, ARO), and Shlomo Blum, Marcelo Fleker, and Oleg Krifucks (from the National Mastitis Reference Center, Kimron Veterinary Institute) for their help in bacterial isolation and characterization. We are also grateful to Ram Reifen and Zipi Berkovich (from The Hebrew University of Jerusalem) for their assistance performing experiments in the in vitro GIT system. We also acknowledge members of the Shemesh lab for helpful discussions and technical assistance. We also wish to thank Eduard Belausov (from the ARO) and Dr. Einat Zelinger (from the Hebrew University) for excellent technical assistance with the microscopy analyses.

**Conflicts of Interest:** The authors declare no conflicts of interest.
