Predatory bacteria as potential biofilm control and eradication agents in the food industry

Abstract

Biofilms are a major concern within the food industry since they have the potential to reduce productivity in situ (within the field), impact food stability and storage, and cause downstream food poisoning. Within this review, predatory bacteria as potential biofilm control and eradication agents are discussed, with a particular emphasis on the intraperiplasmic Bdellovibrio-and-like organism (BALO) grouping. After providing a brief overview of predatory bacteria and their activities, focus is given to how BALOs fulfill four attributes that are essential for biocontrol agents to be successful in the food industry: (1) Broad spectrum activity against pathogens, both plant and human; (2) Activity against biofilms; (3) Safety towards humans and animals; and (4) Compatibility with food. As predatory bacteria possess all of these characteristics, they represent a novel form of biofilm biocontrol that is ripe for use within the food industry.

Introduction

The food industry is a broad business network covering the various aspects of food production, food processing, and distribution. In this industry, bacterial contamination can cause serious problems at each stage, such as impacting the food production process, thereby lowering productivity or quality, leading to food spoilage during processing and distribution and potentially causing food poisoning within the consumers. There is also an economic cost. For instance, in New Zealand, a country of barely 4 million people, the cost associated with food-borne diseases was estimated to be $88.8 million NZD (Scott et al., 2000), while the global costs associated with biofilms incurred in all sectors in 2019 were reported to be more than $5 trillion USD annually, with $324 billion USD in food and agriculture alone (Cámara et al., 2022). To prevent bacterial contamination, therefore, the food industry needs to make every effort to introduce good agricultural and manufacturing practices, including managing hygiene and adopting various methods for microbial inactivation, such as thermal or radiation treatments, or the use of chemicals, such as hypochlorous acid. Despite these efforts, however, the problem of bacterial contamination persists, causing significant financial loss and deaths each year. Among the various factors that make it difficult to prevent contamination, one major concern is bacterial biofilms (Galie et al., 2018; Srey et al., 2013; Van Houdt and Michiels, 2010).

Biofilms, which represent the form in which most bacteria in nature are present and are defined as three-dimensional structures created by microbes within them as they gather and grow (Flemming and Wingender, 2010; Tolker-Nielsen et al., 2015). Biofilms are often found on surfaces and tend to be highly resistant to common bacterial eradication methods due to their structural and physiological properties, which are defined by extracellular polymeric substances (EPS). The bacterial EPS is a complex structure of macromolecules, such as polysaccharides (i.e., poly-N-acetylglucosamine), proteins and DNA, that not only surrounds the bacterial cells and hold them fast, but also limits the penetration of antibiotics and other antibacterials into the biofilm and, as a result, allows the bacteria within to survive treatment (Flemming and Wingender, 2010). Furthermore, the bacterial cells inside experience stress conditions such as limited oxygen and nutrient availabilities, quorum sensing mechanisms, etc., that forces them to enter a less active state, thereby increasing their resistance by not responding to drugs and treatments (Kim et al., 2009; Borriello et al., 2004; Nguyen et al., 2011; Zhao et al., 2020). As a result of both factors, i.e., EPS and low activity, bacteria present in biofilms are reported to be 10- to 1000-times more resistant to antibiotics (Mah and O'Toole, 2001), making the removal of the bacteria located within more difficult.

Biofilms are found in all stages of the food industry, and their incomplete sterilization in food processing plants may contaminate food or lead to serious cases of food poisoning. In response, methods for effective biofilm removal are a topic of constant discussion (Carrascosa et al., 2021; Galie et al., 2018). In this review, therefore, recent work related to predatory bacteria and their applications as novel biofilm biocontrol agents is discussed, highlighting some of the clear benefits found with these unique microorganisms.

Predatory bacteria and their life cycles

As a group of microbes, predatory bacteria have a particular lifestyle in that they grow by preying on other microbes. Although many people, including microbiologists, are not familiar with them, these predators exist ubiquitously throughout the environment and have been found in diverse locales, including rivers (Fry and Staples, 1976; Jang et al., 2022b; Pineiro et al., 2013), soil (Jurkevitch et al., 2000; Klein and Casida Jr, 1967; Oyedara et al., 2016), the ocean (Baer et al., 2004; Williams and Li, 2018), wastewater treatment plants (Cohen et al., 2021; Feng et al., 2016; Jurkevitch, 2020; Mun et al., 2022) and even within animals (Guo et al., 2017; Kelley and Williams, 1992; Schwudke et al., 2001). Moreover, various predatory “families” and their different lifestyles have been identified and studied. Predatory bacteria are commonly sub-categorized into four distinct groups by their predatory lifestyles (Table 1)—wolf-pack, epibiotic, intraperiplasmic and cytoplasmic—which are defined by the location of the predator in relation to its prey. Each of these is briefly described below.

Table 1.

Classification of predatory bacteria according to their lifestyle

Subclass	Predatory strain	References
‘Wolf-pack’ predators	Lysobacter spp.	Seccareccia et al. (2015)
	Myxobacteria; Corallococcus spp., Myxococcus spp., Pyxidicoccus spp.	Livingstone et al. (2017)
Epibiotic predators	Pseudobdellovibrio exovorus	Koval et al. (2013), Pasternak et al. (2013)
	Micavibrio aeruginosavorus	Pasternak et al. (2013), Wang et al. (2011)
	Vampirococcus (Candidate Phyla Radiation)	Esteve et al. (1983), Moreira et al. (2021)
	Vampirovibrio chlorellavorus	Coder and Starr (1978), Hovde et al. (2019)
Intraperiplasmic predators	Bacteriovorax stolpii	Seideler et al. (1972)
	Bdellovibrio bacteriovorus	Im et al. (2019), Sathyamoorthy et al. (2019), Shilo (1969)
	Halobacteriovorax spp.	Baer et al. (2004)
	Peredibacter starrii	Seideler et al. (1972)
	Pseudobacteriovorax antillogorgiicola	McCauley et al. (2015)
Cytoplasmic predators	Daptobacter	Guerrero et al. (1986)

The first group is ‘wolf-pack’ predators. These are generally opportunistic predators, capable of growth axenically but also possess the ability to predate on other microbes. Wolf-pack predators, of which myxobacteria and Lysobacter are examples (Livingstone et al., 2017; Seccareccia et al., 2015), use social gliding motility to encounter prey. While it is not clear if wolf-pack predators need to physically touch their prey as they secrete membrane vesicles and secondary metabolites that can kill the prey from a distance (Evans et al., 2012; Xiao et al., 2011), this form of predation is often viewed as being social since a minimum number of predators are needed to secrete sufficient enzymes and metabolites necessary to ensure prey lysis (McBride and Zusman, 1996). This perspective, however, has been called into question as microscopic evidence shows individual myxobacterial cells can also lyse their prey (Berleman and Kirby, 2009). A key benefit of wolf-pack predators that arises since these predators neither attach to nor invade their prey, but rather secrete antibacterial enzymes and substances, is they prey on a wide range of bacterial species (Livingstone et al., 2017). On the other hand, they are not obligate predators, and under starvation conditions also form multicellular biofilms called fruiting bodies (Muñoz-Dorado et al., 2016; Shimkets, 1999).

In contrast with wolf-pack predators, those located within the next two groupings all tend to be obligate predators, requiring prey bacteria for their growth and survival. The first is epibiotic predators, which include many different of bacterial species, including Micavibrio aeruginosavorus (Pasternak et al., 2013; Wang et al., 2011), Bdellovibrio exovorus (Koval et al., 2013; Pasternak et al., 2013), Vampirococcus (Candidate Phyla Radiation) (Esteve et al., 1983; Moreira et al., 2021) and Vampirovibrio chlorellavorus (Coder and Starr, 1978; Hovde et al., 2019). As a group, these microbes predate others by attaching to the surface of susceptible prey and consuming them while remaining outside of the prey cell. This is achieved in four stages, namely, (1) attachment, where they recognize and adhere to the surface of the prey, (2) pore formation, where the cell wall of the prey bacterium is dissolved, (3) degradation and utilization of the prey’s macromolecules, and (4) septation, during which the elongated predator septates and releases progeny via binary fission (Koval et al., 2013).

Within the second obligate predatory grouping, i.e., intraperiplasmic, the most studied is Bdellovibrio bacteriovorus, a Gram-negative bacterium belonging to the Deltaproteobacteria that a very motile due to a single polar flagellum (Im et al., 2019; Sathyamoorthy et al., 2019; Shilo, 1969). In contrast with wolf-pack and epibiotic predators, intraperiplasmic predation is quite complex and consists of several different stages [attachment, invasion, bdelloplast formation, elongation, septation via segmentation (not binary fission) and, finally, lysis of the prey and release of the predatory progeny (Fenton et al., 2010; Rotem et al., 2015)] (Fig. 1). Among the characterized predatory bacteria, various genus’ have similar intraperiplasmic lifestyles, including Peredibacter (Seideler et al., 1972), Bacteriovorax (Seideler et al., 1972), Halobacteriovorax (Baer et al., 2004), and Pseudobacteriovorax (Koval et al., 2013).

Fig. 1.

The complex lifecycle of intraperiplasmic predatory bacteria, showing the different stages involved

The final grouping, cytoplasmic predators, is represented by only a single predator, i.e., Daptobacter, and is characterized by this microbe penetrating through both membranes and entering into the cytoplasm of its prey, members of the Chromatiaceae (Guerrero et al., 1986). Much like the wolf-pack predators, Daptobacter is also a facultative predator, capable of growing axenically in the absence of prey, and also divides by binary fission (Guerrero et al., 1986).

Despite the diversity of predatory bacteria, both taxonomically and based on their predatory activities, the one common feature shared by all is their ability to kill other bacteria and hydrolyze their macromolecules. This they achieve using the veritable arsenal of proteases, nuclease and other hydrolytic enzymes that are encoded within their genomes (Inoue et al., 2022a; Oyedara et al., 2018; Pasternak et al., 2012; Rendulic et al., 2004; Williams et al., 2019). It should come as no surprise, therefore, that predatory bacteria and their activities have garnered recent attention, with applications in diverse scientific fields being explored, including biofilm removal (Bratanis et al., 2020; Dwidar et al., 2012b; Kadouri and O'Toole, 2005), prevention of membrane biofouling (Kim et al., 2013, 2014), bioplastic recovery (Martínez et al., 2016, 2013) and even sludge treatment (Feng et al., 2017; Yan et al., 2022). However, a majority of the application-based studies that have been published, particularly for the intraperiplasmic Bdellovibrio-and–like organisms (BALOs), focus heavily on their use as potential biocontrol agents (Choi et al., 2017; Cloeckaert et al., 2013; Pérez et al., 2020). Stemming from these previous works, this review will discuss the potential application of predatory bacteria within the food industry (production, process, distribution) and, in particular, as a biofilm eradication agent through the lens of previous studies.

Potential of predatory bacteria as biofilm control and eradication agents for the food industry

Within food processing industries, microbial contamination can occur at multiple stages throughout production, processing, and distribution. When this occurs, it not only leads to declines in productivity but also poses a threat to public health due to increased potential for the development and spread of pathogens associated with food poisoning. Consequently, effective methods for preventing bacterial contamination and spread are paramount in the food industry, with key aspects of these methods including (1) broad activity against pathogens, (2) effectiveness at biofilm removal, (3) safety and (4) compatibility with foods.

While conventional treatments to control bacterial contamination, such as heat, chlorine, ozone or UV, are widely used, these come with some limitations as they may not be very effective at eradicating biofilms (de Carvalho, 2017), may impact the quality of the food being processed (Sert et al., 2020), can negatively affect the equipment being used (Greene et al., 1999) or offer limited scope and applicability due to other reasons, such as toxicity (Baggio et al., 2020). In the following sections, therefore, the importance of each criterion and how predatory bacteria address each is discussed.

Predatory bacteria consume diverse food pathogens

Treatments used to reduce bacterial contamination in the food industry typically have broad-spectrum bactericidal capabilities. This is necessary because food-associated pathogens are diverse (Table 2). For example, Salmonella or E. coli O157:H7, two bacterial strains that cause food poisoning, are present in the breeding farms and enter the food processing pipeline via the animals and meat in a process coined “farm-to-fork” (Collineau et al., 2020; Rothrock et al., 2021; Wilson et al., 2018; Zhang et al., 2022). Once present within the meat processing facilities, these bacteria can grow on the surface of the meat and exposed equipment, form a biofilm in a matter of hours and eventually spread to other foods (Dourou et al., 2011; Silagyi et al., 2009; Wang et al., 2013), potentially causing downstream problems such as food poisoning. Similarly, Pectobacteria spp. and Dickeya spp., are both crop pathogens (Czajkowski et al., 2009; Davidsson et al., 2013). While both of these bacterial species may infect crops in the field and reduce the overall harvest yields, they can also cause serious problems in already harvested produce during its storage (Blancard, 2012; Czajkowski et al., 2011), leading to additional losses in food production and distribution.

Table 2.

Food-associated pathogens and their susceptibility to predation

Pathogen	Source	Disease target	Predation in liquid culturesa	Biofilm removal by predationb	References
Gram-negative
Acidovorax avenae	Fruits, vegetables	Crop; seedling blight, bacterial fruit blotch (BFB)	+++	NDc	McNeely et al. (2017), Schaad et al. (2003)
Aeromonas spp.	Seafood	Aqua farming; crustaceans	++	Bdellovibrio  +	Chu and Zhu (2010), Dashiff et al. (2011), Hoel et al. (2019)
Agrobacterium tumefaciens	Fruits, vegetables	Crop; crown gall disease	±	ND	Escobar and Dandekar (2003), McNeely et al. (2017)
Burkholderia cepacia complex	Raw milk	Human	++	Micavibrio  +	Dashiff et al. (2011), McNeely et al. (2017), Moore et al. (2001)
Campylobacter	Vegetables, poultry, eggs, dairy	Human and Livestock	±	ND	Humphrey et al. (2007), Markelova (2010b)
Erwinia carotovora	Vegetables	Crop; soft rot	++	ND	Toth and Birch (2005), Yair et al. (2003)
Escherichia coli O157:H7	Vegetables, meat, dairy, and seafood	Human and Livestock	++	Bdellovibrio  +	Fratamico and Cooke (1996)
Escherichia coli STEC (non O157:H7)	Vegetables, meat, dairy, seafood	Human and Livestock	++	ND	Baker et al. (2016), Ottaviani et al. (2019)
Pseudomonas spp.	Fruits, vegetables, meat, dairy	Crop, Human and Livestock	+++	Bdellovibrio  +  Micavibrio  +	Dashiff et al. (2011), González‐Rivas et al. (2018)
Salmonella enterica serovar Typhimurium	Fruits, vegetables, poultry, eggs	Human and Livestock; gastroenteritis, septicemia	+++	Bdellovibrio  +	Atterbury et al. (2011), Fratamico and Cooke (1996), Nguyen et al. (2014), Im et al. (2018)
Serratia spp.	Vegetables, meat, dairy	Human and Livestock	±	ND	Dashiff et al. (2011), Kurz et al. (2003)
Shigella spp.	Fecal contaminated food and water	Human; diarrhea	+++	Micavibrio  +	Dashiff et al. (2011), Lima et al. (2015), Willis et al. (2016)
Yersinia enterocolitica	Meat, dairy	Human and Livestock	+++	ND	Monnappa et al. (2014)
Vibrio cholerae	Seafood	Human; cholera	++	Bdellovibrio  +	Cao et al. (2015), Reidl and Klose (2002), Wucher et al. (2021)
Vibrio parahemolyticus	Seafood	Human; gastroenteritis	+++	Bdellovibrio  +	Dashiff et al. (2011), Kongrueng et al. (2017), Letchumanan et al. (2014), Richards et al. (2012)
Vibrio vulnificus	Seafood	Human; cellulitis, septicemia	+++	ND	Diaz (2014), Richards et al. (2012)
Xanthomonas spp.	Vegetables	Crop; spots, blights	++	ND	Boch and Bonas (2010), Odooli et al. (2021)
Gram-positive
Clostridium botulinum	Honey, canned food	Human; botulism	ND	ND	Schneider et al. (2011)
Listeria monocytogenes	Vegetables, meat, dairy	Human; listeriosis	ND	ND	Genigeorgis et al. (1991)
Staphylococcus aureus	Meat, dairy	Human	–	Bdellovibrio  +	Im et al. (2018), Monnappa et al. (2014)

^aActivity: (–)—no activity, (±)—weak to no activity, ( +)—weakly positive activity, (++)—good activity, (+++)—strong activity

^bActivity ( +) Removal of the biofilm was observed by the listed predator based on a qualitative evaluation

^cND—Not done or evaluated

Studies on the prey spectrum and predation ability of predatory bacteria have been most actively conducted with B. bacteriovorus. According to Dashiff et al. (2011), one predatory strain, B. bacteriovorus 109J, attacked and reduced the viabilities (1- to 7-log) of numerous pathogens, including major food-associated pathogens, such as Salmonella, Aeromonas, Escherichia, Enterobacter, Citrobacter, Shigella, Vibrio, Yersinia and Serratia. While Campylobacter, another pathogen commonly associated with food poisoning, was not preyed upon by B. bacteriovorus 109J, Markelova (2010a) reported both Campylobacter jejuni and Helicobacter pylori were susceptible to another predatory strain, B. bacteriovorus 100NCJB. Similar to B. bacteriovorus, the epibiotic predator M. aeruginosavorus also has broad spectrum activities against many bacterial species and is capable of reducing Citrobacter, Enterobacter, Escherichia, Shigella and Yersinia populations all by as much as 2-log (Dashiff and Kadouri, 2011).

One limitation of the above predators, however, is their inability to kill Gram-positive bacterial strains, such as Enterococcus faecalis and Staphylococcus aureus (Dashiff et al., 2011; Im et al., 2018; Monnappa et al., 2014). This is not true for all predators as wolf-pack predators are capable of attacking and consuming Gram-positive strains. In a recent study, Inoue et al. (2022b) isolated two novel predatory strains, Bacteriovorax stolpii HI3 and Myxococcus sp. MH1, from a freshwater pond and compared their activities against a diverse assortment of bacterial species, including Gram-positive strains (e.g., S. aureus and Bacillus spp.). While B. stolpii HI3 was active against many of the Gram-negative strains tested (26 out of 45), it did not predate any of the eight Gram-positive bacterial strains. In contrast, Myxococcus sp. MH1 preyed on all 53 strains tested, both Gram-positive and Gram-negative (Inoue et al., 2022b). While Myxococcus sp. MH1 clearly offers a broader spectrum activity than B. stolpii HI3, one caveat associated with using Myxococcus sp. MH1 is the time required—a week was needed for predation to occur with several of the prey tested, limiting its potential as a biocontrol agent within food industries.

In addition to possessing broad spectrum activities against many different pathogens, one other clear benefit of predatory bacteria is their ability to also mitigate antibiotic resistant populations. This was demonstrated by several different groups where CDC-priority resistance markers (colistin- and carbapenem-resistance) and multidrug-resistant pathogens were employed (Dharani et al., 2018; Jang et al., 2022a; Sun et al., 2017). One of these studies, Jang et al. (2022a), delved deeper into the predation process and demonstrated that not only does predation kill the pathogens, E. coli and K. pneumoniae in their study, it also significantly removed the antibiotic resistance gene pools, i.e., mcr-1, bla_KPC-2 and bla_OXA-51, by as much as 99.3%.

Predatory bacteria effectively remove bacterial biofilms

Current estimates place approximately 80% of all bacteria throughout nature are present within biofilms (Flemming and Wuertz, 2019). Within food industries, biofilms can be found on the surfaces of the food, stainless steel worktops or conveyor belts used in processing (Galie et al., 2018). While these biofilms may be resistant to many forms of treatment, including antibiotic/chemical treatments, based on studies conducted, they likely are quite susceptible to predatory bacteria and their activities. The first example of biofilm predation reported was by Kadouri and O'Toole (2005), where B. bacteriovorus 109J was capable of significantly removing E. coli and Pseudomonas fluorescens biofilms in either static or flow cell environments. The same group expanded on these results with a different predatory strain, M. aeruginosavorus, and its activities against biofilms of several different pathogenic bacterial species, including K. pneumoniae and Pseudomonas aeruginosa (Kadouri et al., 2007). In subsequent studies, Im et al. (2018) demonstrated BALOs also dismantle Salmonella enterica biofilms, a concern particularly in poultry industries (Joseph et al., 2001; Merino et al., 2019), while Chanyi and Koval (2014) explored the role of the predator’s pili in biofilm predation efficiencies, finding the loss of one pili gene (pilT1) had no observable effect but the other (pilT2) abolished the ability of their predator, B. bacteriovorus 109JA, to effectively remove E. coli biofilms (Fig. 2a). Table 2 lists many of the studies where predation was used to reduce biofilms of food-borne Gram-negative pathogens. However, the published results often employed qualitative, rather than quantitative, assessments to determine if predation was effective against these pathogens.

Fig. 2.

Removal of biofilms by intraperiplasmic predatory bacteria. (A) Susceptible Gram-negative bacterial biofilms are consumed by the predator, reducing the viability of the prey present within. (B) In contrast, Gram-positive bacteria are not predated on, meaning their viabilities do not decrease when the predator encounters their biofilms but are rather dispersed as the predator hydrolyzes the extracellular polymeric substances composing the extracellular matrix

While Gram-positive bacterial species, such as Staphylococcus spp., Streptococcus spp., Listeria spp., and Enterococcus spp., are not predated on by intraperiplasmic predatory strains (Dashiff et al., 2011; Im et al., 2017, 2018), their biofilms are dispersed by both the predators and their secreted hydrolytic enzymes (Im et al., 2018; Monnappa et al., 2014) (Fig. 2b). In fact, the study by Im et al. (2018) found this occurs even when the predator, B. bacteriovorus HD100, was washed to remove any free proteases that may be present in the media. Through transcriptomic analyses, they demonstrated the predator utilizing the biofilm’s EPS as a source of amino acids and secrete proteases in response (Im et al., 2018). Aside from these, other studies considered additional predator-pathogenic bacterial biofilm combinations, including the removal of Stenotrophomonas maltophilia biofilms by B. exovorus (Chanyi et al., 2016) and Staphylococcus epidermidis biofilms by Lysobacter gummosus, a wolf-pack predator (Gökçen et al., 2014).

While many of the above studies used polystyrene as the substrate for their biofilms, predatory strains are also effective against biofilms on other surfaces, including stainless steel (Fratamico and Cooke, 1996), silicon (Dwidar et al., 2012a), membrane filters (Kim et al., 2013) and even eukaryotic cells (Dwidar et al., 2013), all of which should be considered as they are used in the food industry. While the first three illustrate predators can be used to remove biofilms from a variety of abiotic surfaces, including the removal of E. coli O157:H7 and Salmonella spp. from stainless steel surfaces like those found in food processing plants (Fratamico and Cooke, 1996), the study by Dwidar et al. (2013) proved bacterial biofilms present on the surfaces of eukaryotic cells are also effectively removed while the underlying eukaryotic cells are not harmed by the predators or their activities. This is important as it suggests predators like B. bacteriovorus can be used within the food industry to reduce the presence of pathogenic species and their biofilms on the surfaces of meat, as was demonstrated recently (Ottaviani et al., 2019), without impacting the quality of the food product, or causing harm to the workers. This latter idea is discussed further later in this review.

Co-applications of predatory bacteria with other antimicrobials have synergetic effects

Whereas many studies report on the ability of different predators and their enzymes to remove bacterial biofilms, one limitation is that this removal is never complete as some of the biofilm, and its bacteria, survive. In response, researchers have combined the activities of predators with other antimicrobials with the goal of making them more effective than when used alone. For instance, the study by Dwidar et al. (2012a) used carbon dioxide (CO₂) particles alongside predation to remove biofilms of E. coli. Not only did they achieve much better removal when these two methods were used together, but the predator itself reduced the E. coli viabilities by 50,000-fold when used alone. In contrast, use of CO₂ aerosols alone to disperse the biofilm saw only a 50% reduction in the E. coli viabilities (Singh et al., 2015), meaning the number of bacteria being dispersed without predation was much higher, effectively increasing the chances for secondary exposures to occur. Another group found the use of either DNase I or DspB (a poly-N-acetylglucosaminidase) improved the ability of B. bacteriovorus 109J to remove biofilms (Dashiff and Kadouri, 2011). In contrast, co-applications with proteinase K reduced the effectiveness of this predator. The authors shared that while the exact reasons for proteinase K inhibition are not clear (prey-associated and/or predator-associated changes), the use of trypsin did not have the same negative impacts, suggesting the detrimental effects seen are specific for proteinase K and its activities.

In addition to using physical (CO₂ aerosols) and enzymatic treatments, other groups also explored the use of predatory bacteria alongside antibiotics. Although the focus was not on biofilms, one such study considered the co-application of B. bacteriovorus HD100 with violacein (Im et al., 2017). As noted above, B. bacteriovorus strains only predate on Gram-negative bacterial species, meaning Gram-positive bacteria are not targeted. To circumvent this, Im et al. (2017) selected violacein, a hydrophobic secondary metabolite produced by various bacterial species that has antibacterial activities against Gram-positive bacterial species as it attacks their membranes (Aruldass et al., 2015; Cauz et al., 2019; Choi et al., 2015, 2021). Im et al. (2017) found that when used against a mixed population consisting of both Gram-positive and Gram-negative bacterial pathogens (S. aureus, Acinetobacter baumannii, Bacillus cereus and K. pneumoniae), violacein or the predator alone reduced the total pathogen viabilities by 19% and 68%, respectively, but led to a 99.98% reduction when used together, illustrating the potential use of predators alongside chemical antimicrobials. A similar concept was employed by Chanyi et al. (2016) in their study where they evaluated the potential co-application of B. exovorus with either ciprofloxacin or kanamycin in the removal of S. maltophilia biofilms. They reported that, whereas kanamycin inhibited predation of the biofilms, ciprofloxacin tended to have no observable impact, positive or negative, on the activities of B. exovorus. The reported effects of kanamycin are not too surprising given other translation inhibitors, i.e., streptomycin, chloramphenicol and puromycin, were all shown previously to allow the predator to attach to their prey but prevented predation from occurring (Varon and Shilo, 1968).

Safety of predatory bacteria shown in in vitro and in vivo tests

A third area that requires consideration is the safety of predatory bacteria. While examples of safety were indirectly reported several decades ago, such as in the study by Lenz and Hespell (1978) where it was shown BALOs do not grow within eukaryotic cells (rabbit ova), the last two decades have seen a plethora of studies exploring this issue, with heavy emphasis once more given to the BALO predators. Examples of this are two recent studies where the impact of several different BALO strains, including B. bacteriovorus, M. aeruginosavorus and a natural isolate of Bacteriovorax stolpii, on numerous human and mouse cell lines were evaluated (Gupta et al., 2016; Monnappa et al., 2016). Both studies concluded none of the predatory strains evaluated cause any loss in viability and that they led to only mild cellular responses (i.e., induced cytokine production levels), even when added at very high densities (1,000 predators per human or mouse cell).

In addition to the above in vitro studies, several groups explored the safety of BALOs within various animal models (Table 3). As discussed in the next section, predatory bacteria are present in and help protect shellfish. They have also been found in and isolated from other animals (Guo et al., 2017; Kelley and Williams, 1992; Schwudke et al., 2001), but their potential impact when provided artificially and at higher doses was not known. One of the first studies to explore this was conducted within chicks and explored the ability of the predator (B. bacteriovorus HD100) to also mitigate Salmonella enterica (Atterbury et al., 2011). They found, when the predator was added alone, it had no adverse effects on the growth and well-being of the chicks. Moreover, its addition after pre-dosing the chicks with S. enterica had beneficial impacts on the birds, significantly reducing both pathogen numbers in the gut cecal contents as well as abnormal cecal morphologies. Other in vivo studies have also been performed, including looking at the impacts of BALOs in the gut (Johnke et al., 2020; Shatzkes et al., 2017), on wound healing skin (Liu et al., 2022; Tajabadi et al., 2022), on corneal wound healing (Romanowski et al., 2016), when introduced into the lungs (Shatzkes et al., 2015, 2016) or in the bloodstream (Shatzkes et al., 2015). In every case, the predator had no ill effects on the animal host and, when tested alongside a bacterial pathogen, was often capable of significantly reducing the population of the latter.

Table 3.

Safety tests performed evaluating predatory bacteria

Animal host	Tested cells (organ or organelle)	Loss in vitality?	References
Bovine	Madin–Darby bovine kidney (MDBK) cells	–a	Boileau et al. (2011)
Chick	Gut (Oral gavage)	–	Atterbury et al. (2011)
Human	HaCaT (Keratinocytes)	–	Gupta et al. (2016), Monnappa et al. (2016)
	HepG2 (Liver epithelial)
	HK-2 (Kidney epithelial)
	MD (Spleen monocytes)
	THP-1 (Blood monocytes)
	T84, Caco2 (Colon carcinoma)
	NuLi-1 (Alveolar epithelial)
Mice	Blood vessel (Intravenous)	–	Shatzkes et al. (2015)
	Lung (Intranasal)	–	Shatzkes et al. (2015)
	Skin (Full-thickness skin wound)	–	Liu et al. (2022), Tajabadi et al. (2022)
	Raw 264.7 (Murine monocytes)
Rabbit	Eye (Ocular surface)	–	Romanowski et al. (2016)
Rat	Gut (Intrarectal)	–	Shatzkes et al. (2017)
	Lung (Intranasal)	–	Shatzkes et al. (2016)
Zebrafish	Larval microinjection (Hindbrain, Tail muscle, Caudal vein)	–	Willis et al. (2016)

^aActivity (–)—no loss in viability observed

BALOs have low immunogenic potential

Several potential reasons for the gentle nature of BALOs towards animal hosts are already known. For instance, Gram-negative outer membranes contain lipopolysaccharides (LPS), endotoxins that are recognized by toll-like receptor (TLR) proteins, leading to induced inflammatory responses and cytokine production levels within our cells (Akira and Hemmi, 2003). While BALOs are themselves Gram-negative bacteria, and also possess LPS within their outer membrane, their LPS has an altered structure as the phosphate groups have been replaced with mannose, effectively making them neutral in charge and much less immunogenic (Schwudke et al., 2003). Similarly, another component present within bacteria that induces immunomodulatory responses is flagellin, the sub-unit protein found within the flagellum (Hajam et al., 2017). BALOs once more also possess a flagellum but this is sheathed by the outer membrane (Thomashow and Rittenberg, 1985), which acts to mask the flagellin from host TLR proteins and immune response.

Methods to remove BALOs after treatment

While the above in vitro and in vivo studies all clearly illustrate the safe nature of BALOs, if the need to control their presence does arise, a simple application of soap (surfactant) will suffice based on the studies by Cho et al. (2019) and Jang et al. (2022b), As discussed in both of studies, BALOs are highly sensitive to detergents, with concentrations of sodium dodecyl sulfate (SDS) as low as 0.02% causing complete or near complete and instantaneous killing of the predator. As such, one can envision using BALOs to remove pathogens from fruits or vegetables and, with a quick wash in a detergent solution and rinse, effectively eliminate most or all of the predators before consuming the produce. One other option to reduce the viability of predatory bacteria on treated produce is to use radiation. This was demonstrated in the study by Olanya et al. (2020) where low doses of gamma irradiation were selected since they do not harm the produce (Niemira and Fan, 2014) but were able to kill B. bacteriovorus 109 present on butter lettuce (Lactuca sativa), with a 0.5 kGy treatment reducing the predator viabilities by 3- to 5-log.

Predatory bacteria are compatible with foods

To prevent food spoilage or food poisoning, excessive contamination of raw materials and equipment with bacteria needs to be prevented. More than just controlling the presence of pathogens, the antibacterial agents used should also not harm or alter the raw materials or the equipment but, rather, be compatible with both. As illustrated in the above sections, predatory bacteria do not harm the underlying substrate and, as such, can be used on many different surfaces, including stainless steel.

As it currently stands, only a few studies used predatory bacteria with the aim of protecting, or removing pathogens associated with, foodstuffs (Table 4). One of the earliest of these explored the presence of Vibrio parahaemolyticus and Vibrio vulnificus, two pathogenic bacteria that infect shellfish, within the Eastern oyster, Crassostrea virginica (Richards et al., 2012). The authors found predatory bacteria (Bdellovibrio, Bacteriovorax and Micavibrio) naturally present in the water and in the oysters acted to suppress and reduce both pathogens. Their study also validated the earlier work by Kelley et al. (1997) who found predatory bacterial strains preferred to associate with surfaces, particularly the shells of oysters. Similar results were reported in another shellfish (black tiger shrimp; Penaeus monodon), where the addition of a predatory isolate (Bdellovibrionales bacterium BDHSH06) led to better survival and pathogen reductions of between 1.3- and 4.5-log in both the water and the intestines of the shrimp (Li et al., 2014).

Table 4.

Case studies of predatory bacteria being applied to protect food

Food	Predatory strain	Target pathogen	References
Crops
Cucumber	Corallococcus sp.	Fusarium oxysporum f. sp. cucumerinum (FOC)	Ye et al. (2020)
Mushroom	B. bacteriovorus	Pseudomonas tolaasii	Saxon et al. (2014)
Potato	B. bacteriovorus	Pectobacterium	Sason et al. (2022), Youdkes et al. (2020), Epton et al. (1989)
		Dickeya solani
		Erwinia carotovora
Rice	B. bacteriovorus	Xanthomonas oryzae	Uematsu (1980)
Soybean	B. bacteriovorus	Pseudomonas glycinea	Scherff (1973)
Seafood
Oyster	B. bacteriovorus	Vibrio spp.	Richards et al. (2012)
	B. stolpii
Shrimp	Bdellovibrio sp.	Vibrio spp.	Cao et al. (2019), Kongrueng et al. (2017), Lu et al. (2022)
	Bacteriovorax sp.

In addition to seafood, researchers have also tested the ability of BALOs and other predatory bacteria to protect produce. Examples include protecting mushrooms (Agaricus bisporus) from Pseudomonas tolaasii (Saxon et al., 2014) and potatoes (Solanum tuberosum) from the phytopathogens (Pectobacterium and Dickeya) (Sason et al., 2022; Youdkes et al., 2020). In both cases, the actions of the predators used reduced the formation of lesion (in the case of the mushrooms) and macerations on the potato slices, illustrating the potential benefits BALOs may offer in protecting and preserving produce. At least one study also explored their effects in situ. Ye et al. (2020) found, over a two-year study, that myxobacterium Corallococcus sp. strain EGB was capable of protecting cucumber (Cucumis sativus L.) plants from Fusarium wilt caused by Fusarium oxysporum f. sp. cucumerinum in both greenhouse and field studies. After application of Corallococcus sp. strain EGB to the soil, this predator not only reduced the Fusarium oxysporum f. sp. cucumerinum abundance but also migrated to the roots of the plants, where it modulated the soil microbiome, leading the authors to conclude this predator has the potential to be used as a biocontrol agent to prevent Fusarium wilt in situ. These results highlight several simple facts—predatory bacteria already exist ubiquitously in nature alongside food products (i.e., shellfish or crops) and their presence does not affect food quality but rather can protect these products from bacterial contamination and poisoning. Therefore, predatory bacteria could be useful to use for the food industry.

Predatory bacteria: potential biocontrol agents in the food industry

As noted above, bacterial contamination and biofilms are major concerns within the food industry as they reduce productivity and storage stability, and can cause downstream food poisoning when the contaminated food is consumed. Current estimates indicate global food and agricultural costs incurred due to unwanted bacterial biofilms exceed $320 billion USD each year, illustrating the extent of this problem. Work from diverse groups around the globe show predatory bacteria may be used as biocontrol agents to help mitigate this, with their broad spectrum activities against many pathogens and their biofilms, as well as their demonstrated safety towards plants and animals, including humans. One caveat, though, is this field, i.e., the use of predatory bacteria within the food industry to control bacterial contaminants and pathogens, is still relatively young and populated with only a small handful of studies. Considering the demonstrated ability of predatory bacteria to dismantle biofilms, however, as well as the description of new and novel predatory strains each year, this area clearly has immense potential for growth and development.

Declarations

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.