Biofilm Formation in Milk Production and Processing Environments; Influence on Milk Quality and Safety


  • Sophie Marchand,

    1. Authors Marchand, De Block, De Jonghe, Heyndrickx, and Herman are with the Inst. for Agricultural and Fisheries Research—Technology and Food Sciences Unit (ILVO-T&V), Brusselsesteenweg 370, 9090 Melle, Belgium. Author Coorevits is with the Faculty of Applied Engineering Sciences, Dept. of Biochemistry and Brewing, Univ. College Ghent, Schoonmeersstraat 52, 9000 Gent, Belgium; and Faculty of Science, Dept. of Biochemistry and Microbiology, Laboratory of Microbiology, Ghent Univ., K. L. Ledeganckstraat 35, 9000 Ghent, Belgium. Direct inquiries to author Marchand (E-mail: ).
    Search for more papers by this author
  • Jan De Block,

    1. Authors Marchand, De Block, De Jonghe, Heyndrickx, and Herman are with the Inst. for Agricultural and Fisheries Research—Technology and Food Sciences Unit (ILVO-T&V), Brusselsesteenweg 370, 9090 Melle, Belgium. Author Coorevits is with the Faculty of Applied Engineering Sciences, Dept. of Biochemistry and Brewing, Univ. College Ghent, Schoonmeersstraat 52, 9000 Gent, Belgium; and Faculty of Science, Dept. of Biochemistry and Microbiology, Laboratory of Microbiology, Ghent Univ., K. L. Ledeganckstraat 35, 9000 Ghent, Belgium. Direct inquiries to author Marchand (E-mail: ).
    Search for more papers by this author
  • Valerie De Jonghe,

    1. Authors Marchand, De Block, De Jonghe, Heyndrickx, and Herman are with the Inst. for Agricultural and Fisheries Research—Technology and Food Sciences Unit (ILVO-T&V), Brusselsesteenweg 370, 9090 Melle, Belgium. Author Coorevits is with the Faculty of Applied Engineering Sciences, Dept. of Biochemistry and Brewing, Univ. College Ghent, Schoonmeersstraat 52, 9000 Gent, Belgium; and Faculty of Science, Dept. of Biochemistry and Microbiology, Laboratory of Microbiology, Ghent Univ., K. L. Ledeganckstraat 35, 9000 Ghent, Belgium. Direct inquiries to author Marchand (E-mail: ).
    Search for more papers by this author
  • An Coorevits,

    1. Authors Marchand, De Block, De Jonghe, Heyndrickx, and Herman are with the Inst. for Agricultural and Fisheries Research—Technology and Food Sciences Unit (ILVO-T&V), Brusselsesteenweg 370, 9090 Melle, Belgium. Author Coorevits is with the Faculty of Applied Engineering Sciences, Dept. of Biochemistry and Brewing, Univ. College Ghent, Schoonmeersstraat 52, 9000 Gent, Belgium; and Faculty of Science, Dept. of Biochemistry and Microbiology, Laboratory of Microbiology, Ghent Univ., K. L. Ledeganckstraat 35, 9000 Ghent, Belgium. Direct inquiries to author Marchand (E-mail: ).
    Search for more papers by this author
  • Marc Heyndrickx,

    1. Authors Marchand, De Block, De Jonghe, Heyndrickx, and Herman are with the Inst. for Agricultural and Fisheries Research—Technology and Food Sciences Unit (ILVO-T&V), Brusselsesteenweg 370, 9090 Melle, Belgium. Author Coorevits is with the Faculty of Applied Engineering Sciences, Dept. of Biochemistry and Brewing, Univ. College Ghent, Schoonmeersstraat 52, 9000 Gent, Belgium; and Faculty of Science, Dept. of Biochemistry and Microbiology, Laboratory of Microbiology, Ghent Univ., K. L. Ledeganckstraat 35, 9000 Ghent, Belgium. Direct inquiries to author Marchand (E-mail: ).
    Search for more papers by this author
  • Lieve Herman

    1. Authors Marchand, De Block, De Jonghe, Heyndrickx, and Herman are with the Inst. for Agricultural and Fisheries Research—Technology and Food Sciences Unit (ILVO-T&V), Brusselsesteenweg 370, 9090 Melle, Belgium. Author Coorevits is with the Faculty of Applied Engineering Sciences, Dept. of Biochemistry and Brewing, Univ. College Ghent, Schoonmeersstraat 52, 9000 Gent, Belgium; and Faculty of Science, Dept. of Biochemistry and Microbiology, Laboratory of Microbiology, Ghent Univ., K. L. Ledeganckstraat 35, 9000 Ghent, Belgium. Direct inquiries to author Marchand (E-mail: ).
    Search for more papers by this author


Abstract:  Bacteria in milk have the ability to adhere and aggregate on stainless steel surfaces, resulting in biofilm formation in milk storage tanks and milk process lines. Growth of biofilms in milk processing environments leads to increased opportunity for microbial contamination of the processed dairy products. These biofilms may contain spoilage and pathogenic microorganisms. Bacteria within biofilms are protected from sanitizers due to multispecies cooperation and the presence of extracellular polymeric substances, by which their survival and subsequent contamination of processed milk products is promoted. This paper reviews the most critical factors in biofilm formation, with special attention to pseudomonads, the predominant spoilage bacteria originating from raw milk. Biofilm interactions between pseudomonads and milk pathogens are also addressed, as emerging risks and future research perspectives, specifically related to the milk processing environment.


Raw milk is an ideal culture medium for microorganisms. Because the microbial load of milk may hold spoilage and/or health risks, the manufacture of milk and milk products is subject to very stringent rules. These rules cover the way in which livestock is kept and milked, milk storage facilities, preparation methods, additives, processing equipment, and the transport tanks that move milk from the farm to the processing plants (Anonymous 2005; Anonymous 2006; Anonymous 2007; Anonymous 2011). On its journey from the farm to the consumer, milk comes into contact with the walls of the equipment in which it is being processed and transported. Since the European (and American) legislation has strict regulations concerning materials coming into contact with foods (Anonymous 2004; FDA 2007; EFSA 2008) and milk processing necessitates hygienic equipment material resistant to corrosion in alkaline and/or acidic conditions (Boulangé-Petermann and others 1997), the dairy industry has employed stainless steel for more than 60 years in almost all segments of the dairy chain. The development of stainless steel in the dairy industry is explained by the fact that it corresponds exactly to the requirements expected of materials in contact with food: 1) the material has to be chemical, bacteriological, and organoleptical neutral with regard to the food product, 2) the material should be easy to clean so that the hygiene and appearance of the food product are guaranteed, and 3) it has to be durable, including corrosion and aging (Anonymous 2004; Bremer and others 2009). Other factors also contribute to the preference of the dairy industry for stainless steel. These include its mechanical characteristics, expansion coefficient, thermal conductivity, and ease of use (Bremer and others 2009). It is difficult to find alternative products to compete with stainless steel in the milk industry, because of the processing conditions. However, in some manufacturing operations, alternative materials can be employed, but their use is still limited and restricted to certain applications. Examples of nonmetal materials used are elastomers (also known as rubbers) and plastics. They are often used in conveyer belts, containers, seals, gaskets, or cutting boards. Rubbers, such as ethylene propylene diene monomer rubber (EPDM), nitril butyl rubber (NBR, known as Buna-N®), silicon rubber, or fluoroelastomer (Viton) are used in both closed equipment (seals gaskets, membranes, fittings, and containers) and in open equipment such as conveyer belts (Faille and Carpentier 2009). Among these materials, the most frequently used gasket materials in milk processing equipment are EPDM and NBR (Faille and Carpentier 2009). A wide range of plastics is also available, but only a few of them are food-approved, such as polypropylene (PP), polycarbonate (PC), high-density polyethylene (HDPE), unplasticized polyvinyl chloride (PVC), and fluoropolymers such as polytetrafluoroethylene (PTFE, Teflon®). The latter, used for gaskets in the food industry, is porous and lacks resilience and must thus be used with care (Faille and Carpentier 2009).

Surfaces of equipment used in food and beverage (such as milk) processing and handling are commonly contaminated by microorganisms, even following cleaning and disinfection procedures (Gibson and others 1999; Marouani-Gadri and others 2010). These contaminating microorganisms appear as adherent microorganisms or as more complex structures called biofilms. Adherent spores and bacteria, as well as biofilms, can be observed on every surface of food industry plants such as stainless steel surfaces (Figure 2), floors, belts, or rubber seals (Costerton and others 1995; Kumar and Anand 1998).

Figure 2–.

Adhesion of various microorganisms on stainless steel surfaces. Adapted from Faille and Carpentier (2009). (A–B) Adhesion of Staphylococcus caprea and Pseudomonas fluorescens under static condition on stainless steel with a 2B finish (horizontally immersed). (C–D) Adhesion of Bacillus cereus spores on substrata with irregular topography, vertically immersed in a spore suspension (presence of scratches, flaws, … ).

Understanding Biofilms

An important reservoir of microbial contamination that has received relatively little attention in the dairy industry is the microbial biofilm. In milk storage and dairy processing operations, as well as in numerous other industrial systems, besides being present in the raw material, most bacteria are associated with surfaces (Mittelman and others 1990; Mosteller and Bishop 1993; Mittelman 1998). The attachment of “pioneering” bacteria with subsequent development of biofilms in milk processing environments is a potential source of contamination of finished products that may shorten the shelf life or facilitate transmission of diseases (Hood and Zottola 1995; Lindsay and others 2002; Brooks and Flint 2008). Despite the fact that bacteria are predominantly present in biofilms, for many years studies on bacterial physiology have focused primarily on the planktonic state. Now, however, it is well established that bacteria are able to switch between different habitation modes: single cells (the planktonic or free floating state) and biofilms. In addition, it has been established that for each planktonic bacterium detected, there might be close to 1000 organisms present in biofilms (Momba and others 2000). A biofilm is defined as a sessile microbial community characterized by adhesion to a solid surface and by production of a matrix that surrounds the bacterial cells and includes extracellular polysaccharides (EPSs), proteins and DNA (Wingender and others 2001; Whitchurch and others 2002; Costerton and others 2003; Bjarnsholt and others 2009). Biofilm development is a result of successful attachment and subsequent growth of microorganisms on a surface (Figure 1). Under suitable conditions, a biofilm in a milk processing environment develops initially through accumulation of organic matter on a metal surface, which is then colonized by bacteria. Transition from planktonic mode to biofilm mode is regulated by a variety of environmental and physiological triggers, such as quorum sensing, nutrient availability, and cellular stress. A biofilm community may comprise single and/or multiple species of bacteria and form a single layer or 3-dimensional structures. Biofilms are large, complex, and organized bacterial ecosystems in which water channels are dispersed providing passages for nutrient, metabolite, and waste product exchange (Sauer and others 2007). Biofilm communities can even provide (in analogy with apoptosis in higher eukaryotes) the selective pressure that is required for programmed cell death, by eliminating damaged individuals from the population (Bayles 2007). Because of competition reduction and the release of nutrients from the dead and lysed cells, nutrient availability is more easily maintained for the healthy individuals that remain (Bayles 2007). Programmed death and lysis of the bacterial cells probably occurs as a function of their spatial orientation within the biofilm. In addition, the released genomic DNA is a structural component of the biofilm matrix, which supports the notion that cell lysis contributes to the stability of the overall biofilm structure (Bayles 2007).

Figure 1–.

Stages of bacterial biofilm development. Adapted from Stoodley and others (2002). Stage 1: Initial attachment of cells to the surface. Stage 2: Production of EPS resulting in more firmly adhered “irreversible” attachment. Stage 3: Early development of biofilm architecture. Stage 4: Maturation of biofilm architecture. Stage 5: Dispersion of single cells from the biofilm. The bottom panels show each of the 5 stages of development represented by a photomicrograph of P. aeruginosa when grown under continuous-flow conditions on a glass substratum.

According to Mittelman (1998), the making of a mature biofilm may take several hours to several weeks, depending on the system under development. For example, in an experiment with Pseudomonas aeruginosa, a common biofilm former on medical devices, it has been established that attachment to stainless steel took place within 30 s of exposure; in an industrial water simulation experiment in a biofilm annular reactor, the colony forming units (CFUs) within the biofilm increased approximately 5-fold, from 420 to 2123 CFU/15 cm2, as the incubation time was prolonged from 24 to 96 h (Florjanic and Kristl 2011). More importantly, in dairy equipment biofilms, the development is also very rapid (8–12 h) (Scott and others 2007; Bremer and others 2009), with numbers of up to 106 bacteria per cm2 being recorded in the generation section of a pasteurizer after 12 h of operation (Bouman and others 1982; Bremer and others 2009). While a biofilm can spread at its own rate by ordinary cell division, it will also periodically release “pioneer” cells to colonize downstream sections of piping. The biological, chemical, and physiological factors that drive detachment are complex and incompletely understood (Chambless and Stewart 2007). Multiple factors are probably associated with attachment and detachment processes, depending on the availability of nutrients or oxygen (Chandy and Angles 2001; Rice and others 2005), shear–stress (Mittelman 1998; Guillemot and others 2006; Lee and others 2008; Florjanic and Kristl 2011), quorum sensing (Rice and others 2005), microbial metabolic activity, and microbial gene expression (Kaplan and others 2003; Kaplan and others 2004). Biofilm detachment has been divided into 3 processes: erosion, abrasion, and sloughing (Garny and others 2008). Erosion (result of fluid shear forces) and abrasion (collision of particles) refer to the continuous detachment of single cells or small cell clusters and affect the total biofilm surface. Sloughing refers to the instant loss of large parts of the biofilm, therefore affecting the entire biofilm and not only the biofilm surface (Morgenroth 2003). Depending on the strength of the biofilm, sloughing can even lead to a complete loss of the biofilm. Several detachment processes may occur simultaneously (Telgmann and others 2004). However, the original biofilm structure and magnitude and the detachment force might have a strong influence on the frequency and extent of a specific detachment process.

Biofilms are characterized by environmental conditions and the surfaces colonized, the bacterial genes activated and required to form and maintain the biofilm, and the types of extracellular products that are concentrated in the biofilm matrix. There are many different types of biofilms and even one bacterium may make several different types of biofilms under different environmental conditions. Here, we review the diverse array of biofilms formed in milk processing environment, with special attention to pseudomonads, the predominant spoilage bacteria originating from raw milk. Biofilm interactions between pseudomonads and milk pathogens will also be addressed, as well as emerging risks and biofilm control strategies specifically related to the milk processing environment.

Dairy Practice—Mechanisms of Biofilm Formation

Bacterial attachment and the formation of biofilms appear to take place in different stages, such as formation of a conditioning layer, bacterial adhesion, bacterial growth, and biofilm expansion (Kumar and Anand 1998; O'Toole and others 2000; McLandsborough and others 2006; Kokare and others 2009). In dairy operations, the conditioning film mainly consists of organic milk components. This first stage occurs within the first 5–10 s after placement of an otherwise clean surface into a fluid environment (Mittelman 1998). The conditioning also alters the physicochemical properties of the surface, such as surface free energy, changes in hydrophobicity, and electrostatic charges, which may affect the subsequent order of microbial events (Dickson and Koohmaraie 1989). The formation of conditioning films can be influenced by the material type contacting the milk. As an example, certain materials are known for their “theta surface,” which is a characteristic expression of outermost atomic features least retentive of depositing proteins, and identified by the bioengineering criterion of having a measured critical surface tension (CST) between 20 and 30 mN/m (Baier 2006). The most effective atomic group exposures for theta surface results are intrinsically hydrophobic, closely packed methyl, CH3, terminals, or repeating CH2CF2 runs in polyvinylidene fluoride (PVDF) (Baier 2006). Unfortunately, most of these materials are not (yet) approved, at least in Europe, for contact with food (Anonymous 2004). Therefore, further research is needed before applications in the food industry become possible. During the second stage of biofilm formation, single bacterial cells are transported to surfaces and reversible bonds are formed between the cell wall and the substratum. Bacterial attachment is mediated by fimbriae, pili, flagella, and bacterial extracellular polymeric substances (EPSs) that act to form a bridge between bacteria and the conditioning film (Kokare and others 2009). The chemical structure of the EPS varies among different types of organisms and is also dependent on environmental conditions (Momba and others 2000). While there is some debate about the influence on surface roughness on bacterial attachment (Sreekumari and others 2005; Oliveira and others 2006; Silva and others 2008), there appears to be a general agreement about the importance of using surfaces with minimal cracks and crevices in order to reduce bacterial adherence and biofilm growth and to enhance cleaning effectiveness (Bremer and others 2009). Once established, biofilms accelerate corrosion and material detoriation (Storgards and others 1999a). Dead ends, corners, cracks, crevices, gaskets, valves, and joints are all possible points for biofilm formation (Storgards and others 1999a; Storgards and others 1999b). Biofilms do not possess a uniform structure (Wimpenny and others 2000; McLands-borough and others 2006). The structures that are formed depend on a large variety of intrinsic and extrinsic factors such as species, temperature, flow conditions, pH, presence of salts, nutrients, and so on (McLandsborough and others 2006). Next to contact material, temperature plays an important role in the adhesion of bacteria to surfaces. In general, higher temperatures (37 °C in comparison with 4, 12, and 22 °C) seem to increase cell surface hydrophobicity and subsequently bacterial attachment (Cappello and Guglielmino 2006; Di Bonaventura and others 2008). The effect of flow conditions has not been well studied in the dairy industry, but from studies in other systems, it is known that biofilms grow denser under high than under low shear conditions (Stoodley and others 2002; Bremer and others 2009). Contrary to expectations, both laminar and turbulent flow conditions have been observed to enhance bacterial attachment (by bringing bacteria closer to a surface) when compared to static conditions (Rijnaarts and others 1993). It has been speculated that turbulent flow may push bacterial cells onto the surface, thus enhancing probability of adhesion and biofilm formation (Donlan and Costerton 2002). With regard to pH, it has been shown that the pH of the surrounding solution influences the interaction between bacterial cells and the metal surface; the bacteria–metal adhesion force appears to reach the highest value when the pH of the solution is near the isolelectric point of the bacteria, that is, at the zero point charge (Sheng and others 2008). Stronger ionic strength in the solution, on the other hand, results in a higher bacteria–metal adhesion force, which is due to the stronger electrostatic attraction force between the positively charged metal surface and negatively charged bacterial surface (Sheng and others 2008). From this, it can be deduced that the higher the adhesion force, the more bacteria will attach to a particular surface. Concerning the effect of the presence of nutrients or certain milk components on the adhesion of bacteria, some conflicting statements can be retrieved from the literature. On the one hand, it is stated that milk proteins coated on stainless steel, rubber, and dairy equipment reduce bacterial adhesion (Speers and Gilmour 1985; Helke and others 1993; Bernbom and others 2009), while, on the other hand, certain bacteria (for example, Bacillus cereus) appear to need certain milk components before adhesion can occur (Shaheen and others 2010). These contradictory findings might be explained by the fact that different milk types (skim compared with whole milk and heated (100 min for 30 min) compared with unheated) were used in the experimental setups. Evidently, denatured milk proteins may have other characteristics than undenatured proteins naturally present in refrigerated whole raw milk. Second, whole milk contains natural surfactants and phospholipids, both surface-active compounds that can be retrieved in the fat globules of milk. The ability of B. cereus spores to adhere and act as an initiation stage for biofilm formation on a wide variety of materials commonly encountered in food processing plants is also well known (Peng and others 2001; Faille and others 2001; Heyndrickx and others 2010). The strong adhesion properties of B. cereus spores have been attributed to the hydrophobic character of the exosporium (Peng and others 2001; Faille and others 2001), which varies from species to species (Tauveron and others 2006) and to the presence of appendages on the surface of the spores (Vanloosdrecht and others 1989). Thick biofilms of B. cereus were shown to develop on stainless steel coupons at the air–liquid interface, while biofilm formation was much lower in submerged systems (Wijman and others 2007). This suggests that B. cereus biofilms develop particularly in partly filled industrial storage and piping systems and these biofilms act as a shelter for spore formation that can be subsequently released by dispersal into the food production system. Spores embedded in biofilms are protected against disinfectants such as chlorine, chlorine dioxide, and peroxyacetic-acid-based sanitizer (Ryu and Beuchat 2005). In some dairies, persistent silo tank contamination, heat exchange equipment contamination, or postpasteurization contamination are important sources of B. cereus (te Giffel and others 1996a; te Giffel and others 1996b). For pasteurized and extended shelf life (ESL) milk, the filling machine has been shown as the main source of recontamination, with the filler nozzles, aerosols, and the water at the bottom of the filling machine being of particular concern (Rysstad and Kolstad 2006). While for most of Bacillus strains, negative effects of whole milk on biofilm formation have been observed (Flint and others 1997a; Wong 1998), the study of Shaheen and others (2010) illustrated that B. cereus was capable of forming biofilms in whole milk, but not in water-diluted milk. The results of that latter study suggest that any surface-active compound found in whole milk might work as a surfactant needed for biofilm formation by certain strains of B. cereus. However, the effects of surfactants on biofilm formation might be strain-specific and generalization of results concerning (anti-)adhesive properties of milk compounds should thus be avoided and evaluated carefully.

Several groups have reported on the ability of bacteria to attach to surfaces commonly found in the milk processing environment, such as rubber and stainless steel (Czechowski 1990; Krysinski and others 1992; Suarez and others 1992). Scanning electron micrographs revealed that food-borne pathogens and spoilage microorganisms can accumulate as biofilms on aluminum, Buna-N and Teflon seals, and nylon materials typically found in food processing environments (Herald and Zottola 1988a; Herald and Zottola 1988b; Mafu and others 1990; Blackman and Frank 1996). More importantly, during heat processes above 65 °C, whey proteins in milk begin to denature and aggregate, which can lead to a faster adherence than proteins in their native state. This protein adherence can change the surface properties of stainless steel (De Jong 1997), increasing the likelihood of bacteria attaching to a surface and creating an environment that encourages bacterial attachment to an extent where in one study, it was discovered that fouled surfaces attracted 10–100 times more vegetative cells and spores of G. stearothermophilus to the surface than the clean stainless steel (Flint and others 2001). The milking equipment can be contaminated by milk spoilers and pathogens through the dairy farm and processing environment, but also through the rinsing water used in the milking machines (Oliver and others 2005). Microorganisms originating from rinsing water (especially Pseudomonas, Aeromonas, and Legionella spp. [Momba and others 2000]) can form biofilms that are difficult to eradicate and can act as a harbor and/or substrate for other microorganisms less prone to biofilm formation, thus increasing the probability of pathogen survival and further dissemination during milk processing (Lomander and others 2004). Table 1 demonstrates typical problem areas within dairies. Teixeira and others (2005) also illustrated that the short rubber milking tube (of the cluster in automatic milking machines) is one of the points more prone to biofilm formation. The cluster, which attaches to the udder of the cow, consists of 4 teatcup assemblies (each having a shell, a rubber liner, and a short milk and short pulse tube), a claw, a long milk tube, and a long pulse tube. All these constituents are made of rubber, stainless steel, or plastic. Other possible hazards include biofilm accumulation and microbial colonization in milk pipelines, storage tanks, and milk silos (Shaheen and others 2010), as well as fouling of heat exchangers (Giffel and others 1997; Flint and others 1997b; Flint and others 1999; Flint and others 2000) and adhesion of spores on packaging material surfaces (Kirtley and Mcguire 1989).

Table 1–.  Overview of biofilm problem areas at dairy farms and dairy processing plants (Wirtanen 2004; Teixeira and others 2005; Agarwal and others 2006; Gunduz and Tuncel 2006).
  Type of bacteria
Sampling pointsMaterialsPseudomonasAeromonasStaphylococcusBacillusLABEnterobacteriaceaeListeria
Balance tankSteel++
Aging tankSteel+
Feeding unit*+++
Conveyer belt of packaging machineRubber+++
Floor drain*++
Ultrafiltration membranesSteel+
Silo, welded jointsSteel+++*
Air separators, insideSteel++++*
Tank truck, valve, gasketRubber+++*
Tank truck, air separatorSteel++*
Tank truck, air separator, gasketRubber+++*
Bulk tank outlet on farmSteel*****+
Rubber linersRubber*****+
Short milking tubeRubber++++

Environments, which select for monospecies biofilms (such as those of thermophilic bacilli) in dairy processing plants, are typically the sections with elevated temperatures (40 to 65 °C) (Stadhouders and others 1982; Flint and others 1997a; Murphy and others 1999). Examples are preheating and evaporation sections of milk powder plants, plate heat exchangers used during the pasteurization process, centrifugal separators operated at warm temperatures (45–55 °C), recycle loops in butter manufacturing plants, and cream heaters in anhydrous milk fat plants (Burgess and others 2010). A typical problem in the manufacture of milk powder is the high levels of Anoxybacillus flavithermus and Geobacillus spp. The spores of these organisms are very heat-resistant, with the vegetative cells able to grow in temperature of up to 65 °C (Palmer and others 2010). The bacteria are normally present in low levels in raw milk, but may reach 105 CFU/g in the final product after 15–20 h of plant operation (Hinton and others 2002; Ruckert and others 2004). The limited residence time of the milk during milk powder manufacture cannot explain the number of thermophiles found in the final product. Suggestions are made that biofilm formation on the milk evaporator and consequent sloughing off into the product line is responsible for the high contamination levels of the final product (Hinton and others 2002; Palmer and others 2010). Strategies (such as shorter production lengths and the use of sanitizers) to prevent thermophilic biofilm formation had limited success, partly due to limited knowledge on the structure and composition of those biofilms in milk processing operations (Burgess and others 2010).

In a model pasteurizer, thermophilic streptococci were detected on the walls of the cooling section at levels of 107 cells/cm2, and subsequent research in processing facilities indicated that thermophilic streptococci could be frequently isolated from the cooling section of pasteurizers (Bouman and others 1982). The attachment of resistant Streptoccocus thermophilus occurs mainly to heat exchanger plates in the downstream side sections of pasteurizers giving rise to the contamination of pasteurized milk (Driessen and others 1984). Recontamination of consumer packages of pasteurized milk with Gram-negative psychrotrophic bacteria, on the other hand, was associated with rinsing water in and around the filling machine during the filling operation (Eneroth and others 1998; Dogan and Boor 2003). This suggests that bacteria could have formed biofilms in the rinse water system. Biofilms that can develop on the sides of gaskets may also be a possible source of postpasteurization contamination (Austin and Bergeron 1995). Langeveld and others (1995) heated milk in a laboratory-scale stainless steel tube heat exchanger and found that, as a result of release from the tube walls, the concentration of bacteria in the milk could increase by a factor of 106. These authors demonstrated a relationship between the density of bacteria on the tube walls and the concentration of cells in the milk after heating. The observations of Scott and others (2007) on commercial milk evaporators showed that in a plant cleaned according to standard industrial practice and processing high-quality milk (<100 thermophiles per mL) that resides, on average, 20–30 min in the plant, might result in outflowing milk that can contain up to 106 cells per mL within 18 h of run commencement. The authors concluded that it was not possible for these numbers to have been produced during the transit of the milk through the plant and must thus have originated from cells immobilized on plant internal surfaces. Other locations where biofilms often arise are ultrafiltration and reverse osmosis membranes (Tang and others 2009a, 2009b, 2010). Membrane separation technology is often used for the removal of bacteria from skim milk in the production of ESL milk, concentration of casein micelles, and recovery of serum proteins from whey. Membrane biofouling caused by microbial attachment leads to decreased membrane flux and increased filtration pressure, and subsequently, increased operation cost due to frequent cleaning and replacement of clogged membranes (Liao and others 2004; Le-Clech and others 2006). Tang and others (2009a) illustrated that the bacterial isolates recovered from such membranes in dairy plants predominantly belonged to Pseudomonas, Bacillus, and Klebsiella genera. Furthermore, the attachment of the different isolates appeared highly variable and there was an enhanced adherence in the presence of whey. In an ice cream plant, most of the biofilm formations were seen on the conveyer belt of the packaging machine 8 h after the beginning of the production (Gunduz and Tuncel 2006). Most of the Gram-negative biofilm-forming bacteria were identified as Proteus, Enterobacter, Citrobacter, Shigella, Escherichia, Edwardsiella, Aeromonas, Plesiomonas, Moraxella, Alcaligenes, and Pseudomonas species. Gram-positive biofilm-forming isolates consisted of Staphylococcus, Bacillus, Listeria, and lactic acid bacteria such as Streptococcus, Leuconostoc, and Pediococcus (Gunduz and Tuncel 2006) species. A brief overview of bacteria isolated at different sampling points in dairy processing plants is given in Table 1.

Which organisms attach to the surface is a function of the planktonic population present in the raw material and the processing conditions in the particular equipment. Heat-sensitive Pseudomonas and Listeria species are most likely to be found in pipes and silos holding milk prior to pasteurization, whereas thermophilic biofilms may form in heated equipment. In terms of their effect on product acceptability, biofilms can contain a dual risk: product detoriation and disease transmission, respectively, through spoilage bacteria and pathogens.

Risks posed by spoilage bacteria and pathogens—the role of pseudomonads

Bacterial spoilage still causes significant losses for the dairy industry. Milk contamination with psychrotrophic microorganisms is of particular concern to the dairy industry as dairy products are stored and distributed at temperatures permissive for the growth of these organisms. Psychrotrophic bacteria are ubiquitous in nature and can be isolated from soil, water, and vegetation (Cousin 1982). The psychrotrophic population in refrigerated raw milk includes both Gram-positive and Gram-negative genera; they comprise representatives of Pseudomonas, Aeromonas, Acinetobacter, Serratia, Alcaligenes, Achromobacter, Enterobacter, Flavobacterium, Klebsiella, Bacillus, Arthrobacter, Clostridium, Lactobacillus, Listeria, Staphylococcus, Corynebacterium, Microbacterium, and Micrococcus (Cousin 1982; Champagne and others 1994; Lafarge and others 2004; Munsch-Alatossava and Alatossava 2006). The conditions during storage and transport in refrigerated tanks cause the raw milk microbiota to change from predominantly Gram positives to predominantly Gram negatives during bacterial growth. Gram-negative bacteria usually account for more than 90% of the microbial population in cold raw milk that has been stored (Cousin 1982).

Currently, the predominant Gram-negative microorganisms limiting the shelf life of ultra heat-treated (UHT) processed fluid milk at 4 °C are Pseudomonas spp., especially P. fragi, P. lundensis, and P. fluorescens-like organisms (Craven and Macauley 1992; Ternström and others 1993; Marchand and others 2009a; De Jonghe and others 2011). Pseudomonas spp. can grow to high numbers and can form biofilms during refrigerated storage. Many of them produce heat-stable extracellular lipases, proteases, and lecithinases that contribute to milk spoilage (Shah 1994; Sorhaug and Stepaniak 1997; Marchand and others 2009b). Furthermore, many of these enzymes remain active even following thermal processing steps that destroy their producing organisms (Garcia and others 1989; Sorhaug and Stepaniak 1997; Marchand and others 2009b). Degradation of milk components through various enzymatic activities can reduce the shelf life of processed milk. For example, digestion of casein by proteases can lead to bitter of flavors and the clotting and gelation of milk (Chen and others 2003; Datta and Deeth 2003). Lipases hydrolyze tributyrin and other milk fat glycerides to yield free fatty acids, which cause milk to taste rancid, bitter, unclean, and soapy. Lecithinases degrade milk fat globule membrane phospholipids and increase the susceptibility of milk fat to the action of lipases (Cousin 1982; Shah 2000). The hydrolytic products of milk fats and proteins always decrease the organoleptic quality of fluid milk products.

In raw milk holding equipment, 2 distinct but connected phases are available for microbial growth: the liquid phase, in which planktonic cells proliferate, and the solid/liquid interface (such as milk-covered cooling tank walls) where cells can attach and form biofilms (Wong and Cerf 1995; Somers and others 2001). Each phase constitutes a unique habitat and cells can move from one to the other, depending on growth stage, nutrient availability, and flow shear forces (Stoodley and others 2002). Pseudomonas spp. and Streptococcus spp. are among the bacteria most frequently isolated from surfaces in the food industry (Sundheim and others 1992; Mettler and Carpentier 1998; Flint and others 1999; Flint and others 2000; Simões and others 2008). While streptococci form predominantly monospecies biofilms on heat exchanger plates in the downstream side of the sections of pasteurizers (Bouman and others 1982; Driessen and others 1984; Flint and others 1999), Pseudomonas spp. are more likely to produce multispecies biofilms on the walls of milk cooling tanks or pipelines prior to heat processing. The development of a single-species biofilm may occur due to the fact that heat-sensitive species are killed during pasteurization leaving only heat-resistant species such as Streptococcus bovis and Streptococcus thermophilus (Bouman and others 1982; Flint and others 2000). In addition to the risk of being a severe contamination source to subsequent milk batches passing the biofilm region, Pseudomonas biofilms may attract and/or shelter other (spoilage or pathogenic) bacteria. In this regard, Simoes and others (2009) illustrated that dual biofilms of P. fluorescens and B. cereus were about 5 times more metabolically active than P. fluorescens monospecies biofilms. In terms of viability, P. fluorescens was more tolerant to antimicrobials than B. cereus in single-species biofilms. Moreover, bacteria were more susceptible to antimicrobials in single-species biofilms than in dual-species biofilms (Simoes and others 2009). Kives and others (2005) reported on the cocultivation of Lactococcus lactis ssp. cremoris and Pseudomonas fluorescens in refrigerated milk. Compared to each monospecies biofilm, the dual-species biofilms showed a more developed structure in which both species were maintained. The benefit was most significant for L. lactis, a poor biofilm former, which probably benefited from the enhanced attaching potential provided by the quickly developing matrix originating from P. fluorescens. In addition, the latter strain consumed much of the available oxygen in the biofilm, which was an additional advantage for the anaerobic L. lactis. In return, P. fluorescens utilized the lactic acid produced by L. lactis as a nutrient source. This interdependence led to compact masses of P. fluorescens entrapping L. lactis cells. Besides, P. fragi has been shown to enhance the attachment of L. monocytogenes to glass surfaces (Sasahara and Zottola 1993). The enhancement was attributed to polysaccharide production by P. fragi. Also, Flavobacterium spp. have been shown to promote biofilm formation of L. monocytogenes (Bremer and others 2001). Probably, P. fragi and Flavobacterium spp. act as primary colonizers of the surface, making adhesion easier for L. monocytogenes. Lindsay and others (2002) demonstrated an enhancement of B. cereus cell attachment by 0.5–1 log cfu/cm2 in a binary biofilm with P. fluorescens. In return, B. cereus appeared to protect P. fluorescens from the sanitizers used in this study.

Another important feature of milk-spoiling Pseudomonas biofilms might be the altered phenotype of the inhabiting strains. One process involved in phenotypic diversification is phase variation, which is usually a reversible, high-frequency phenotype switching corresponding to differential expression of one or several genes. The genes implicated in phase variations encode the GacA/GacS 2-component regulatory system (van den Broek and others 2005b), which regulates secondary metabolism, exo-enzyme production, quorum sensing, motility, and, not surprisingly, biofilm formation (Lapouge and others 2008). Phase variation, which is inducible by environmental factors such as temperature (Schwan and others 1992; Gally and others 1993), medium composition (White-Ziegler and others 2000), and stress conditions (White-Ziegler and others 2002), can influence the growth characteristics and extracellular enzyme production in Pseudomonas spp. (Chabeaud and others 2001; van den Broek and others 2005a). Since phase variation seems to be induced in the biofilm growth mode, this can hold important implications toward the production of milk-spoiling enzymes by pseudomonads present in dairy biofilms. Workentine and others (2010) characterized 2 distinct colony morphology variants from biofilms of P. fluorescens mutants missing the GacS sensor kinase. These variants produced more biofilm cell mass and displayed a change in amino acids and metabolites produced through glutathione biochemistry. In laboratories, these types of colony morphology variants are recovered at increasing frequencies when biofilms are exposed to stressors such as oxidative agents, antibiotics, and metal ions (Davies and others 2007; Harrison and others 2007; Boles and Singh 2008). This suggests that phenotypic switching might play an important role in the survival of a biofilm population during environmental stresses. Since biofilms are frequently exposed to sanitizers during cleaning of dairy processing equipment, phenotypic switching may occur on a regular basis and even influence the enzyme production by pseudomonads adding an additional spoiling factor to the subsequently processed milk batch. This might certainly be the case if such Pseudomonas biofilms are present in the cooling equipment in farms or holding silos in the dairy factory. Since the enzymes might be released from the biofilms into the milk, without bacterial detachment, the contamination might go unnoticed until problems arise with the shelf life of the heat-treated dairy products. This might be of special importance to UHT-processed milk since the Pseudomonas enzymes are heat-resistant and withstand the heating conditions applied (Dogan and Boor 2003; Marchand and others 2009b). Therefore, it is clear that dairy processing equipment should be checked regularly for biofilm formation and cleaned efficiently in order to prevent milk-spoiling events or consumer exposure to pathogens.

Efficacy of Different Cleaners and Sanitizers on Dairy Biofilms

Biofilm control in dairy manufacturing plants generally involves a process called cleaning-in-place (CIP). This cleaning process is characterized by the cleaning of complete plant items or pipeline circuits without the need to dismantle or open the equipment and with little or no manual involvement from the operator (Bremer and others 2006). CIP can be defined as circulation of cleaning liquids through machines and other equipment in a cleaning circuit. The passage of the high-velocity flow of liquids over the equipment surfaces generates a mechanical scouring effect that dislodges milk deposits. This, however, only applies to the flow in pipes, heat exchangers, pumps, valves, separators, and so on. The normal technique for cleaning large milk storage tanks is to spray the detergent on the upper surfaces, and then allow it to run down the walls. The mechanical scouring is then often insufficient, but this effect can be improved to some extent by use of specially designed spray devices (Bylund 1995).

In the dairy industry, CIP systems generally involve the sequential use of caustic (sodium hydroxide) and acid (nitric acid) wash steps, and chemicals originally selected for their ability to remove organic (proteins and fat) and inorganic (calcium phosphate and other minerals) fouling layers (Kessler 1981). In some cases, sanitizers are also incorporated in the CIP system (Kessler 1981; Bylund 1995). The choice of the cleaning process is determined by the type and composition of the soiling matter as well as by the design of the equipment to be cleaned (Kessler 1981). In addition, dairy CIP programs differ according to whether the circuit to be cleaned contains heated surfaces or not. Examples of both cleaning programs are given in Table 3. The main difference between the 2 types is that acid circulation must always be included in the first type to remove encrusted protein and salts from the surfaces of heat-treatment equipment. To enhance cleaning effectiveness, caustic detergents and caustic additives have been developed, which contain surfactants, emulsifying agents, chelating compounds, and complexing agents (Bremer and others 2006). Traditionally, chlorine (sodium hypochlorite)-based sanitizers have been used, however, a wide variety of sanitizers including quaternary ammonium compounds, anionic acids, iodophores, and chlorine-based compounds are currently in use or being evaluated for use in CIP systems (Joseph and others 2001; Parkar and others 2004; Bremer and others 2006). The selection of detergents and disinfectants in the dairy industry depends on the efficacy, safety, and rinsability of the agent and whether it is corrosive or affects the sensory values of the processed products.

Table 3–.  Examples of dairy CIP programs, adapted from Bylund (1995)
CIP wash steps for circuits with pasteurizers and other equipment with heated surfaces (UHT, and others)
1Rinsing with warm water for about 10 min.
2Circulation of an alkaline detergent solution (0.5%–1.5%) for about 30 min at 75 °C.
3Rinsing out alkaline detergent with warm water for about 5 min.
4Circulation of (nitric) acid solution (0.5%–1.0%) for about 20 min at 70 °C.
5Postrinsing with cold water.
6Gradual cooling with cold water for about 8 min.
CIP wash steps for circuits with pipe systems, tanks, and other process equipment with no heated surfaces
1Rinsing with warm water for 3 min.
2Circulation of a 0.5%–1.5% alkaline detergent at 75 °C for about 10 min.
3Rinsing with warm water for about 3 min.
4Disinfection with hot water 90–95 °C for 5 min.
5Gradual cooling with cold tap water for about 10 min (normally no cooling for tanks).

A feature of CIP operations, evident in both industrial and laboratory-scale systems, is their variability in effectiveness in eliminating surface-adherent bacteria or biofilms (Austin and Bergeron 1995; Faille and others 2001; Dufour and others 2004). This variability is not surprising as a large number of factors can influence CIP effectiveness including the nature, age and composition of the biofilm, the cleaning agent composition and concentration, cleaning time, cleaning agent temperature, degree of turbulence of the cleaning solution, and the characteristics of the surface being cleaned (Stewart and Seiberling 1996; Changani and others 1997; Lelievre and others 2001; Lelievre and others 2002). Cleaning effectiveness is dependent on both product and processing plant-specific variables. The optimal CIP regime varies among dairy processing plants and also over time within a given plant. In this regard, Bremer and others (2006) have demonstrated that dairy biofilms consisting of Gram-positive spore-forming bacilli and thermoresistant streptococci were not adequately removed by a standard CIP procedure (water rinse, 1.0% sodium hydroxide at 65 °C for 10 min, water rinse, 1.0% nitric acid at 65 °C for 10 min, and water rinse). However, the authors found that when a caustic additive (containing chelating and sequestering agents and surface-active wetting agents) and a nitric acid blend (containing surfactants) were added, a 3.8 log reduction in the number of cells recovered from a stainless steel surface was achieved (Bremer and others 2006). This study thus illustrated that the effectiveness of a “standard” CIP can possibly be enhanced through testing and use of caustic and nitric blends. Hydrogen peroxide has been found to be effective in removing biofilms from equipment used in hospitals (MattilaSandholm and Wirtanen 1992). Wirtanen and others (1995) showed that the peroxide-based disinfectant was the most effective disinfectant against Pseudomonas biofilms when the microbiological activity was measured using conventional cultivation. The effect of hydrogen peroxide is based on the production of free radicals, which affect the biofilm matrix. The microbicidal effect of peracetic acid on microbes in biofilms was shown to be variable (Christensen 1989; Kramer 1997). Aldehydes did not break the biofilm, but rather seemed to improve its stability. The biofilm must be disrupted in some way before chemical agents such as peracetic acid and aldehydes can be used effectively (Wirtanen 2004). The effect of ozone treatments has been found to vary depending on the processing circumstances and the bacteria tested; ozonation proved very effective in the treatment of cooling water systems (Lin and Yeh 1993). Disinfectants are most effective in the absence of organic (such as fat-, sugar-, and protein-based) materials (Wirtanen 2004). Organic substances, pH, temperature, concentration, and contact time generally control the efficacy of disinfectants (Mosteller and Bishop 1993). The disinfectants must be effective, safe, and easy to use and also easily rinsed off from surfaces, leaving no toxic residues or traces that affect the sensory attributes of the food product. In a study by Lequette and others (2010), the cleaning efficiency of polysaccharidases and proteolytic enzymes against biofilms of bacterial species found in food industry processing lines was analyzed. Two serine proteases and an α-amylase appeared to be the most efficient enzymes. Proteolytic enzymes promoted biofilm removal of a larger range of bacterial species than polysaccharidases, while more specifically, the serine proteases were more efficient in removing Bacillus biofilms and the polysaccharidases were better at removing P. fluorescens biofilms (Lequette and others 2010). Solubilization of enzymes with a buffer containing surfactants and dispersing and chelating agents enhanced the efficiency of polysaccharidases and proteases in removing biofilms of Bacillus and P. fluorescens, respectively (Lequette and others 2010). Considering these results, a combination of enzymes targeting several components of EPS, surfactants, and dispersing and chelating agents could be a good alternative to chemical cleaning agents.

Ultrasonic Cleaning

Most of the previously described cleaning and disinfection processes are well known and often used in food industry premises. A less familiar technique applicable in cleaning off place (COP) systems is ultrasonic cleaning. The use of ultrasound is one of the most recently studied promising cleaning methods (Kallioinen and Manttari 2011). Ultrasound is a form of energy generated by pressure/sound waves of frequencies that are too high to be detected by the human ear, namely, above 16 kHz (Jayasooriya and others 2004). During a sonication process, longitudinal waves are created when a sonic wave meets a liquid medium, thereby creating regions of alternating compression and expansion. These regions of pressure change cause cavitation to occur, and gas bubbles are formed in the medium. These bubbles have a larger surface area during the expansion cycle, which increases the diffusion of gas, causing the bubble to expand (Dolatowski and others 2007). A point is reached where the ultrasonic energy provided is not sufficient to retain the vapor phase in the bubble; therefore, rapid condensation occurs. The condensed molecules collide violently, creating shock waves. Depending on the frequency used and the sound wave amplitude applied, a number of physical, chemical, and biochemical effects can be observed, which enable a variety of applications. In ultrasonic cleaning, biofilm or foulant removal takes place as a result of mechanical actions, caused by ultrasound in the fluid medium, or as a result of chemical interactions of foulants with radicals, which are generated into the liquid through ultrasonic treatment (Ashokkumar and Grieser 1999; Lamminen and others 2004; Kallioinen and Manttari 2011). In the study by Oulahal and others (2004), 2 ultrasonic devices, a flat (T1) and a curved (T2) ultrasonic transducers, were developed to remove biofilms from opened and closed surfaces, respectively. The authors obtained total removal of Escherichia coli and Staphylococcus aureus milk model biofilms with the T1 transducer (10 s at 40 kHz), while the T2 transducer failed to completely remove these model biofilms: 30% and 60% removal for the E. coli and S. aureus biofilms, respectively (Oulahal and others 2004). When a chelating agent was combined with the ultrasound of transducer T2, complete removal was obtained in the E. coli biofilm, but no enhancement could be obtained in the S. aureus milk biofilm (Oulahal and others 2004). In the study by Baumann and others (2009), the efficacy of ultrasound and ozonation was determined using for the removal of L. monocytogenes biofilms from stainless steel chips. Ultrasound (20 kHz, 100% amplitude, 120 W) was applied for 30 or 60 s at a distance of 2.54 cm from a biofilm chip, while it was submerged in 250 mL of sterile potassium phosphate buffer (pH 7.0). Ozone was cycled through the 250 mL of potassium phosphate buffer containing the biofilm chip also for 30 or 60 s at concentrations of 0.25, 0.5, or 1.0 ppm. Each of the treatments alone resulted in a significant detachment of the cells, with ultrasound being the most effective. For the ozone in combination with ultrasound treatment, detachment was higher than by either treatment alone (Baumann and others 2009). It can be questioned that ultrasound is usable in large plants used for milk powder, cheese, or yogurt because of the necessary scaling-up. Nevertheless, for certain applications (such as cleaning of storage tanks) in milk processing, ultrasound may be a fruitful option.

Biofilm Resistance to Antimicrobial Agents

Bacteria in biofilms have intrinsic mechanisms that protect them from even the most aggressive environmental conditions, including the exposure to antimicrobials (Gilbert and others 2002; Cloete 2003; Davies 2003). Dynes and others (2009) investigated the effect of subinhibitory concentrations of 4 antimicrobial agents. Their results indicate that each antimicrobial agent elicited a unique response: P. fluorescens cells and biofilms changed their morphology and architecture, as well as the distribution and abundance of biomacromolecules, in particular the exopolymer matrix. Diversity in microbial communities leads to a variety of complex relationships involving inter- and intraspecies interactions (Berry and others 2006; Hansen and others 2007). The surface colonization by one type of bacterium can enhance the attachment of others to the same surface. This process allows the development of multispecies communities often possessing greater combined stability and resilience than that of each individual species (Moller and others 1998; Burmolle and others 2006). In this regard, Norwood and Gilmour (2000) investigated the effect of sodium hypochlorite on multispecies biofilms containing P. fragi, S. xylosus, and L. monocytogenes. In a constant-depth film fermenter, in the absence of sodium hypochlorite, the steady-state population of L. monocytogenes was only 1.5% of the total plate count, while the P. fragi proportion amounted to 59% and was significantly greater than that of S. xylosis (39.5%), showing a greater competitive advantage for the pseudomonad in the unchallenged biofilm. While all 3 planktonic cultures, subjected to 10 ppm free chlorine for 30 s, were completely eliminated, only a 2 log reduction in L. monocytogenes cells in the multispecies biofilm could be achieved after the biofilm was exposed to 1000 ppm free chlorine for 20 min. Their study confirmed that multispecies biofilms increased protective properties over monospecies biofilms. The authors attributed these observations to the shielding effect of increased numbers (or aggregation) of microorganisms but also to the production of greater amounts of EPS. Sommer and others (1999) also found an increase of Pseudomonas biofilm resistance to chlorine with increasing age of the biofilm. Here, it was speculated that metabolic change or the production of exocellular compounds might be responsible for the interaction with free chlorine or prevention of its diffusion in the biofilm. Also, Lindsay and others (2002) reported an increased resistance of P. fluorescens against a chlorine dioxide containing sanitizer through protection by the mere presence of the more tolerant B. cereus. Such shielding results from the physical protection or engulfment of the sensitive species by the tolerant one. Modifications in EPS composition and quantity also appear to influence bacterial resistance. Certainly, in pseudomonads, the capacity to alter EPS composition may be part of its intrinsic resistance to antimicrobials (Dynes and others 2009).

Monitoring, Detection, and Lab-Scale Biofilm Research

The bacterial enumeration of biofilms helps in identifying the type(s) of microorganisms involved in biofilm formation. The different methods employed for sampling and enumeration of biofilms in a dairy plant are swabbing, rinsing, agar flooding, and agar contact methods (Kumar and Anand 1998). The organisms found in biofilms, however, are not always easily cultured, resulting in an underestimate—or no detection at all—of the true biofilm population inside a liquid handling system (Wirtanen 1995). As an alternative, scraping (Frank and Koffi 1990), vortexing (Mustapha and Liewen 1989), and ultrasound are often used. Studies have shown that ultrasound generates sufficient cavitational bubble activity to remove biofilms from metal, glass, ceramic, and plastic surfaces (Stickler and Hewett 1991). The ultrasonic treatment for microbial recovery consists of immersing and agitating samples, usually in glass test tubes or flasks containing liquid, in a high-frequency ultrasonic bath (18–55 kHz) (Jeng and others 1990). The disadvantage of this lab-scale method, however, is that it is an invasive technique. Surface samples have to be cut out from the plant and studied in bath sonicators (Oulahal-Lagsir and others 2000). Besides, grooves, crevices, dead, ends, and corrosion patches are areas where biofilms readily occur but are hard to access, thus hampering sampling of such areas. In addition, some of the bacteria in biofilms in dairy environments are subjected to various stresses such as starvation, chemicals, heat, cold, and desiccation that may injure the cells and render them unculturable. Fortunately, a combination of disciplines can be used to develop data on biofouling in liquid operations. The biofouling monitoring systems can provide different levels of information according to their specific design (Flemming 2003; Tamachkiarow and Flemming 2003). For example, some are able to assess the biofilm dynamics, attachment/detachment events, but cannot differentiate between the constituents of such layers (such as biotic/abiotic) (Pereira and Melo 2009). More specific monitoring devices are able to characterize the chemical/biological composition of a given fouling layer, although they are too sophisticated and costly to be operated in an industrial setting (Pereira and Melo 2009). For example, to monitor the onset of buildup of fouling on internal surfaces of heat exchangers, a commercial sensor can be used to measure the local heat flux and temperature on the hot side of a plate-type heat exchanger. A real-time estimate of the fouling rate can be obtained by calculating the heat transfer coefficient normalized to its value at the beginning of the run (Bennet 2007). Another system to evaluate global fouling can be used by calculating the energy balance over a tubular-type heat exchanger. In that case, the overall heat transfer coefficient is calculated by measuring the inlet and outlet temperatures (Bennet 2007). Also, specular reflectance Fourier transform infrared (FTIR) spectroscopy can capture a fingerprint of the organic constituents of a fouling film and epifluorescence optical microscopy with the appropriate fluorescent nucleic acid stains can be used to determine whether the pipe wall fouling is biological or abiotic in nature. Different methods are available that report biofilm growth online, in real-time and nondestructively, but they all are based on physical methods. One example is a method that uses 2 turbidity measurement devices, one of which is constantly cleaned. The difference of signals is proportional to the biomass developing on the noncleaned window (Klahre and Flemming 2000). Another one is the fiber active device (FOS), which is based on a light fiber integrated in the test surface, measuring the scattered light of material deposited on the tip (Tamachkiarow and Flemming 2003). Fornalik (2008), on the other hand, has developed a fouling cell assembly in 316L-grade stainless steel that may be placed in dairy pipes and silos. Such assemblies enable monitoring of biofilm development without removal of the processing equipment out of the plant and can be used to generate objective data on the effectiveness of cleaning procedures (Fornalik 2008).

Due to the limitations of studying biofilm development in practical settings, most of biofilm research has been performed in laboratory-based model systems. Despite the many different types of biofilm model systems described in the scientific literature, none of them can be considered as the optimal, universally applicable model system. On the contrary, every researcher has to choose a particular model system that enables to address the specific research questions formulated in the beginning of the study. Since biofilm growth simulating devices are beyond the scope of this review, only a small oversight is given here (Table 2). For further information on this topic, the reader is directed toward other extensive reviews (McLandsborough and others 2006).

Table 2–.  Biofilm growth simulating devices. Primary use and limitations in biofilm research, adapted from McLandsborough and others (2006).
DevicePrimary UseLimitationsReferences
  1. *Already used in milk biofilm studies.

Microtiter plate*Useful in genetic studies because of high throughput screening of “static biofilms.”Only for early stages of biofilm formation.(O'Toole and Kolter 1998; Djordjevic and others 2002; Stepanovic and others 2004)
BioFluxUseful in genetic studies because of high troughput screening of “flow biofilms.”Limited to GFP-expressing bacteria.(Benoit and others 2010)
Polycarbonate membranes*Simple methodology and easy to use. Suitable for antimicrobial penetration tests.Since bacterial cultures are manually deposited on the membrane, biofilms do not naturally develop.(Anderl and others 2000; Werner and others 2004; Borriello and others 2004; Tang 2011)
Capillary reactorBiofilm structure is formed on glass capillary. Direct microscopic observation is feasible.Biofilm growth is limited to a single surface.(Mccoy and others 1981; Werner and others 2004)
Flow cell reactor*Biofilms can be studied under either laminar or turbulent flow in order to simulate the changes in fluid velocity that occur during the operation of industrial reactors.None for the purpose the method was designed for.(Pereira and others 2002; Simões and others 2008)
Robbin's device*Using a brass pipe, removable sections of the wall can be removed to test biofilm growth. Used in industrial biofouling.Just one type of material can be tested at a time.(Mccoy and others 1981; Mittelman 1998)
Modified Robbin's deviceAllows several materials to be tested. Used in industrial biofouling.Used for traditional biofilm cultures and not for genetic investigations.(Nickel and others 1985)
Calgary biofilm deviceHigh troughput. Rapid and reproducible assays in biofilm susceptibility to antibiotics.None for the purpose the method was designed for.(Ceri and others 1999)
Rotating disk reactorDifferent biomaterials can be used for colonization and shear forces can be controlled.High variability seen in biofilm formation between samples.(Okabe and others 1999)
CDC biofilm reactorUsed to follow biofilm formation (under moderate to high shear), characterize biofilm structure, and assess the effect of antimicrobial agents.The baffle rotation speed has to be carefully controlled.(Donlan and others 2004; Goeres and others 2005)
Rotating annular reactor*Application of a well defined shear fieldNone for the purpose the method was designed for(Camper and others 1996; Jang and others 2006)
Batch and batch-fed growth system*Suitable for a wide variety of biofilm experimentsHigh variability seen in biofilm formation between samples(Cerca and others 2004; Sirianuntapiboon and others 2005)
Constant-depth film fermentorBiofilm growth and resistance to antimicrobials in multispecies biofilmsIn order to reach a steady-state biofilm, the biofilm has to be grown in a chemostat.(Knowles and others 2005)
Animal modelsBiofilm formation and distribution in tissues can be monitored. Used in a biofilm model of chronic cystitis and prostatitis.Time-consuming and regulatory issues(Kadurugamuwa and others 2004)

After selection of an appropriate growth model to produce biofilms, the need arises to reliably quantify the number of cells in the developed biofilm and to determine the structure and composition of the biofilm. Biofilm structure development has been analyzed using light, fluorescence, differential interference contrast (DIC), transmission electron (TE), scanning electron (SE), atomic force (AF), and confocal laser scanning microscopy (CLSM) (Ceri and others 1999; Storgards and others 1999a; Djordjevic and others 2002; Donlan and Costerton 2002; Hunter and Beveridge 2005; Lagace and others 2006; Sigua and others 2010; Shaheen and others 2010). CLSM has been developed in the 1980s and allows examination of biofilms without the limitations imposed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Fully hydrated biofilms are analyzed by progressive laser scans at different focal planes within the sample. Computer analysis of the scanned images permits recreation of the 3-dimensional structure of the biofilm. The application of CLSM combined with a number of fluorescent stains provides an important and effective tool to analyze the composition and structure of hydrated biofilms in situ, nondestructively, and in real time (Lawrence and Neu 1999; Manz and others 1999). Viability and distribution of cells within the biofilm may be analyzed as well. When using epifluorescence or CLSM, the choice of suitable fluorescent stains is critical in order to increase the contrast between organisms and the exopolymers in the biofilms. Nucleic acid stains such as 4,6-diamino-2-phenylindole (DAPI) or acridin orange have been used to stain the DNA of cells regardless of their viability (Trachoo 2003). Other dyes sensitive to viable cells such as propidium iodine (PI) or 5-cyano-2,3-ditolyl tetrazolium chloride may be used to further resolve viable and dead cells (Donlan and Costerton 2002).

Conclusions and Research Perspectives

Biofilms are one of the main recontamination sources of milk. It has been established that for each planktonic bacterium detected, there might be close to 1000 organisms present in biofilms. In the dairy industry, mono- as well as multispecies biofilms can occur. Pathogenic bacteria can coexist within a biofilm with other environmental organisms; an example of this is L. monocytogenes surviving in Pseudomonas biofilms. Biofilms are difficult to remove from milk processing environments due to the production of EPS materials and the difficulties associated with cleaning complex processing equipment and processing environments. Since stringent cleaning protocols are available, cleaning procedures should be accurately applied, and ideally, the cleaning efficiency should be evaluated. There is far too little knowledge on persisting contamination sources and existing innovative cleaning and disinfection techniques. Therefore, it is important that research results in this area are thoroughly communicated with the industry. In addition, an objective “cleaning efficiency measuring system” should be developed, which in the end can lead to the issuance of directives for economical and technical optimalization of existing CIP systems. Biofilm control relies in the end on the design of storage and processing equipment, effective cleaning and sanitizing procedures, and the correct implementation and application. The management of these factors is important to ensure safe and good-quality milk and dairy products.

From a practical viewpoint, future research could also focus on coating strategies to reduce microbial attachment on dairy equipment and cleaners and sanitizers with fortified properties (for example, addition of EPS or protein-degrading enzymes). In addition, exploration of the exo-enzyme production by biofilm pseudomonads might be of interest when milk spoilage is under study. From a more fundamental research approach, it could be very useful to investigate the temperature effect on the development of Pseudomonas biofilms. Since both biofilm production and exo-enzyme production are under the same control in pseudomonads and exo-enzyme production is elevated at lower temperatures (certainly in milk), it could be interesting to check if milk Pseudomonas strains have a competitive advantage over the other milk bacteria due to an increased biofilm-forming capacity at lower temperatures.


The authors wish to thank the Agency for Innovation by Science and Technology (IWT) for financial support.