**5. Molecular components of** *Salmonella* **biofilms formed on abiotic surfaces**

Curli fimbriae (formerly designated as thin aggregative fimbriae or Tafi) and cellulose are the two main matrix components (exopolymeric substances, EPS) in *Salmonella* biofilms (Gerstel & Römling, 2003). When co-expressed on Congo Red (CR) agar plates, curli fimbriae and the exopolysaccharide cellulose form the characteristic rdar (red, dry and rough) morphotype (also called rugose or wrinkled) (Römling, 2005). Their syntheses are co-regulated by a complex regulatory system. The LuxR type regulator CsgD protein stimulates the production of curli through transcriptional activation of the *csgBAC* (formerly *agfBAC*) operon, while the activation of cellulose production is indirect through the regulator AdrA which is a member of the GGDEF protein family regulated by *csgD* (Römling et al., 2000). García et al. (2004) demonstrated that most GGDEF proteins of *S*. Typhimurium are functionally related, probably by controlling the levels of the same final product, cyclic di-GMP, a secondary messenger that seems to regulate a variety of cellular functions including cellulose production and biofilm formation. The co-expression of curli fimbriae and cellulose leads to the formation of a highly hydrophobic network with tightly packed cells aligned in parallel in a rigid matrix and enhances biofilm formation on abiotic surfaces (Jain & Chen, 2007). Solomon et al. (2005) showed that 72% of 71 *S. enterica* strains, originating from produce, meat or clinical sources and belonging to 28 different serovars, expressed the rdar morphotype, with curli- and cellulose-deficient isolates being least effective in biofilm formation on polystyrene microtiter plates. White et al. (2006) showed that rdar morphotype significantly enhanced the resistance of *Salmonella* to dessication and sodium hypochlorite, suggesting that this phenotype could play a role in the transmission of *Salmonella* between hosts. However, aggregation via the rdar morphotype does not seem to be a virulence adaptation in *S*. Typhimurium, since competitive infection experiments in mice showed that nonaggregative cells outcompeted rdar-positive wild-type cells in all tissues analyzed (White et al., 2008).

A variety of environmental cues such as nutrients, oxygen tension, temperature, pH, ethanol and osmolarity can influence the expression of the transcriptional regulator CsgD, which regulates the production of both cellulose and curli (Gerstel & Römling, 2003). Transcription of *csgD* is dependent upon the stationary phase-inducible sigma factor RpoS, and is maximal in the late exponential or early stationary phase of growth (Gerstel & Römling, 2001). For an extensive overview on the current understanding of the complex genetic network regulating *Salmonella* biofilm formation, reader is advised to refer to the recently published review of Steenackers et al. (2011). When *csgD* is not expressed the morphotype is a conventional smooth and white (saw) colony, which does not produce any extracellular matrix (Römling et al., 1998b). In wild type *Salmonella* strains, rdar morphotype is restricted to low temperature (below 30°C) and low osmolarity conditions, but biogenesis of curli

The persistence of *Salmonella* within the food chain has become a major health concern, as biofilms of this pathogen formed in food processing environments can serve as a reservoir for the contamination of food products. The development of materials to be used for foodcontact surfaces with improved food safety profiles continues to be a challenge. One approach which has been developed to control microbial attachment is the manufacture of food-contact materials incorporating antimicrobial compounds. Triclosan-impregnated kitchen bench stones (silestones), although prone to bacterial colonization, were found to reduce *S*. Enteritidis biofilm development on them and also the viability of cells within the

**5. Molecular components of** *Salmonella* **biofilms formed on abiotic surfaces**  Curli fimbriae (formerly designated as thin aggregative fimbriae or Tafi) and cellulose are the two main matrix components (exopolymeric substances, EPS) in *Salmonella* biofilms (Gerstel & Römling, 2003). When co-expressed on Congo Red (CR) agar plates, curli fimbriae and the exopolysaccharide cellulose form the characteristic rdar (red, dry and rough) morphotype (also called rugose or wrinkled) (Römling, 2005). Their syntheses are co-regulated by a complex regulatory system. The LuxR type regulator CsgD protein stimulates the production of curli through transcriptional activation of the *csgBAC* (formerly *agfBAC*) operon, while the activation of cellulose production is indirect through the regulator AdrA which is a member of the GGDEF protein family regulated by *csgD* (Römling et al., 2000). García et al. (2004) demonstrated that most GGDEF proteins of *S*. Typhimurium are functionally related, probably by controlling the levels of the same final product, cyclic di-GMP, a secondary messenger that seems to regulate a variety of cellular functions including cellulose production and biofilm formation. The co-expression of curli fimbriae and cellulose leads to the formation of a highly hydrophobic network with tightly packed cells aligned in parallel in a rigid matrix and enhances biofilm formation on abiotic surfaces (Jain & Chen, 2007). Solomon et al. (2005) showed that 72% of 71 *S. enterica* strains, originating from produce, meat or clinical sources and belonging to 28 different serovars, expressed the rdar morphotype, with curli- and cellulose-deficient isolates being least effective in biofilm formation on polystyrene microtiter plates. White et al. (2006) showed that rdar morphotype significantly enhanced the resistance of *Salmonella* to dessication and sodium hypochlorite, suggesting that this phenotype could play a role in the transmission of *Salmonella* between hosts. However, aggregation via the rdar morphotype does not seem to be a virulence adaptation in *S*. Typhimurium, since competitive infection experiments in mice showed that nonaggregative cells outcompeted rdar-positive

A variety of environmental cues such as nutrients, oxygen tension, temperature, pH, ethanol and osmolarity can influence the expression of the transcriptional regulator CsgD, which regulates the production of both cellulose and curli (Gerstel & Römling, 2003). Transcription of *csgD* is dependent upon the stationary phase-inducible sigma factor RpoS, and is maximal in the late exponential or early stationary phase of growth (Gerstel & Römling, 2001). For an extensive overview on the current understanding of the complex genetic network regulating *Salmonella* biofilm formation, reader is advised to refer to the recently published review of Steenackers et al. (2011). When *csgD* is not expressed the morphotype is a conventional smooth and white (saw) colony, which does not produce any extracellular matrix (Römling et al., 1998b). In wild type *Salmonella* strains, rdar morphotype is restricted to low temperature (below 30°C) and low osmolarity conditions, but biogenesis of curli

biofilm (Rodrigues et al., 2011).

wild-type cells in all tissues analyzed (White et al., 2008).

fimbriae occurs upon iron starvation at 37°C. Römling et al. (2003) showed that the majority (more than 90% of 800 strains) of human disease-associated *S*. Typhimurium and *S*. Enteritidis (isolated from patients, foods and animals) displayed the rdar morphotype at 28°C, but just rarely at 37°C. Interestingly, mutants in the *csgD* promoter have also been found expressing rdar morphotype independently of temperature (Römling et al., 1998b).

Curli fimbriae are amyloid cell-surface proteins, and are involved in adhesion to surfaces, cell aggregation, environmental persistence and biofilm development (Austin et al., 1998; Collinson et al., 1991; White et al., 2006). The *csg* (curli subunit genes) genes (previously called *agf* genes) involved in curli biosynthesis are organized into two adjacent divergentlytranscribed operons, *csgBAC* and *csgDEFG* (Collinson et al., 1996; Römling et al., 1998a). Knocking out the gene encoding for the subunit of thin aggregative fimbriae, AgfA, results in pink colony formation, the pdar (pink, dry and rough) morphotype, which is characterised by production of cellulose without curli (Jain & Chen, 2007). Solano et al. (2002) stressed the importance of the applied biofilm system since they noticed that curli were not essential for biofilm mediated glass adherence under adherence test medium (ATM) conditions, while they were indispensable to form a tight pellicle under LB conditions.

In addition to curli, the second component of the extracellular matrix of the *Salmonella* biofilms is cellulose, a *β*-1→4-D-glucose polymer, which is biosynthesized by the *bcsABZCbcsEFG* genes (bacterial cellulose synthesis) (Zogaj et al., 2001). Both operons are responsible for cellulose biosynthesis in both *S*. Enteritidis and *S*. Typhimurium (Jain & Chen, 2007; Solano et al., 2002). Cellulose production impaiment generates a bdar (brown, dry and rough) morphotype on congo red (CR) agar plates, characteristic of the expression of curli. Solano et al. (2002) showed that cellulose is a crucial biofilm determinant for *Salmonella*, under both LB and ATM conditions, without however affecting the virulence of the bacterium. Additionally, cellulose-deficient mutants were more sensitive to chlorine treatments, suggesting that cellulose production and biofilm formation may be an important factor for the survival of *Salmonella* in hostile environments. Prouty & Gunn (2003) identified its crucial importance for biofilm formation on glass coverslips. However, cellulose was not a major constituent of the biofilm matrix of *S*. Agona and *S*. Typhimurium strains isolated from the feed industry, but it contributed to the highly organized matrix structurization (Vestby et al., 2009a). Malcova et al. (2008) found that cellulose was not crucial for *S*. Enteritidis adherence and biofilm formation on polystyrene.

Latasa et al. (2005) also reported another matrix component, BapA, a large cell-surface protein required for biofilm formation of *S*. Enteritidis. This protein was found to be loosely associated with the cell surface, while it is secreted through the BapBCD type I protein secretion system, encoded by the *bapABCD* operon. The expression of *bapA* was demonstrated to be coordinated with the expression of curli and cellulose through the action of *csgD* (Latasa et al., 2005). Also, these authors demonstrated that a *bapA* mutant strain showed a significant lower colonization rate at the intestinal cell barrier and consequently a decreased efficiency for organ invasion compared with the wild-type strain.

Motility was found to be important for *Salmonella* biofilm development on glass (Prouty & Gunn, 2003) and polyvinyl chloride (PVC) (Mireles et al., 2001). On the contrary, Teplitski et al. (2006) noticed that the presence of the flagellum on the surface of the cell, functional or not, is inhibitory to biofilm formation on polystyrene, as mutants lacking intact flagella, showed increased biofilm formation compared to the wild-type. Flagella were not found to be important for *S*. Typhimurium rdar expression on Congo Red (CR) agar plates (Römling & Rohde, 1999). Solano et al. (2002) noticed that flagella affect *S*. Enteritidis biofilm development

Attachment and Biofilm Formation by Salmonella in Food Processing Environments 167

use a variety of autoinducing polypeptides (AIPs). AHLs are synthesized and recognized by QS circuits composed of LuxI and LuxR homologues, respectively (Whitehead et al., 2001). Both AHLs and AIPs are highly specific to the species that produce them. A third QS system is proposed to be universal, allowing interspecies communication, and is based on the enzyme LuxS which is in part responsible for the production of a furanone-like compound,

Bacteria use QS communication circuits to regulate a diverse array of physiological activities, such as genetic competence, pathogenicity (virulence), motility, sporulation, bioluminescence and production of antimicrobial substances (Miller & Bassler, 2001). Yet, a growing body of evidence demonstrates that QS also contributes to biofilm formation by many different species (Annous et al., 2009; Davies et al., 1998; Irie & Parsek, 2008; Lazar, 2011). As biofilms typically contain high concentration of cells, autoinducer (AI) activity and QS regulation of gene expression have been proposed as essential components of biofilm

To date, three QS systems have been identified in *S. enterica* and are thought to be mainly implicated in the regulation of virulence (SdiA, luxS/AI-2 and AI-3/epinephrine/ norepinephrine signaling system) (Boyen et al., 2009; Walters & Sperandio, 2006). Firstly, the LuxR homologue SdiA has been characterized in *Salmonella*, but there does not appear to be a corresponding signal-generating enzyme similar to LuxI in this species (Ahmer et al., 1998). Since *Salmonella* does not possess a luxI homologue, it cannot produce its own AHLs (Ahmer, 2004). However, *Salmonella* SdiA can detect AHLs produced by a variety of bacterial species, leading to the suggestion that SdiA can be used in interspecies communication within a mixed-species community (Michael et al., 2001; Smith & Ahmer 2003). Till now, SdiA is known to activate the expression of the *rck* operon and the *srgE* gene (Ahmer et al., 1998; Smith & Ahmer, 2003). In contrast to the function of SdiA in *E. coli* adherence to HEp-2 epithelial cells and also biofilm formation on polystyrene (Lee et al., 2009; Sharma et al., 2010), no direct link between SdiA and *Salmonella* biofilms has been reported. Interestingly, Chorianopoulos et al. (2010) demonstrated that cell-free culture supernatant (CFS) of the psychrotrophic spoilage bacterium *Hafnei alvei*, containing AHLs among other unknown metabolites, negatively influenced the early stage of biofilm formation by *S.* Enteritidis on stainless steel. Similarly, Dheilly et al. (2010) reported the inhibitory activity of CFS from the marine bacterium *Pseudoalteromonas* sp. strain 3J6 against biofilm formation on glass flow cells by *S. enterica* and other Gram-negative bacteria. Taking into account that *Salmonella* possess SdiA, a receptor of AHLs which may be produced by resident flora on food-contact surfaces (Michael et al., 2001; Smith & Ahmer, 2003; Soares & Ahmer, 2011), the effect of AHLs on biofilm formation by this pathogen in multispecies real

The second QS system of *Salmonella* uses the LuxS enzyme for the synthesis of AI-2 (Schauder et al., 2001; Soni et al., 2008). The Lsr ABC transporter is known to be involved in the detection and transport of AI-2 into the cell (Taga et al., 2001), while the *rbs* transporter has recently been suggested as an alternative AI-2 uptake system (Jesudhasan et al., 2010). A *S.* Typhimurium *luxS* deletion mutant was impaired in biofilm formation on polystyrene (De Keersmaecker et al., 2005; Jesudhasan et al., 2010). However, this phenotype could not be complemented by extracellular addition of QS signal molecules, suggesting that AI-2 is not the actual signal involved in *Salmonella* biofilm formation (De Keersmaecker et al., 2005). To this direction, Kint et al. (2010) analyzed additional *luxS* mutants for their biofilm phenotype. Interestingly, a *luxS* kanamycin insertion mutant and a partial deletion mutant,

called autoinducer-2 (AI-2) (Schauder et al., 2001).

physiology (Kjelleberg & Molin, 2002; Parsek & Greenberg, 2005).

food processing environments needs to be further studied.

only under LB but not under ATM conditions. Stafford & Hughes (2007) showed that the conserved flagellar regulon gene *flhE*, while it is not required for flagella production or swimming, appeared to play a role in flagella-dependent swarming and biofilm formation on PVC. Kim & Wei (2009) noticed that flagellar assemply was important during biofilm formation on PVC in different (meat, poultry and produce) broths and on stainless steel and glass in LB broth.

Colanic acid, a capsular extracellular polysaccharide, essential for *S*. Typhimurium biofilm development on epithelial cells was found not to be required for *Salmonella* biofilm formation on abiotic surfaces (Ledeboer & Jones, 2005; Prouty & Gunn, 2003). Solano et al. (2002) showed that colonic acid was important to form a tight pellicle under LB conditions, while it was dispensable under ATM conditions. De Rezende et al. (2005) purified another capsular polysaccharide (CP) from extracellular matrix of multiresistant *S*. Typhimurium DT104 which was found to be important for biofilm formation on polystyrene centrifuge tubes and was detected at both 25°C and 37°C. This was comprised principally of glucose and mannose, with galactose as a minor constituent. Malcova et al. (2008) confirmed the importance of this capsular polysaccharide in the biofilm formation capacity of strains unable to produce either curli fimbriae or cellulose. Due to mucoid and brown appearance on Congo Red agar plates, their morphotype was designated as sbam (smooth, brown and mucoid).

However, other capsular polysaccharides can be present in the extracellular biofilm matrix of *Salmonella* strains (de Rezende et al., 2005; Gibson et al., 2006; White et al., 2003), and the exact composition depends upon the environmental conditions in which the biofilms are formed (Prouty & Gunn, 2003). Another component of the EPS matrix of *Salmonella* bile-induced biofilms, the O-antigen (O-ag) capsule, while it was found to be crucial for *S*. Typhimurium and *S*. Typhi biofilm development on gallstones, this was not necessary for adhesion and biofilm formation on glass and plastic (Crawford et al., 2008). The formation of this O-ag capsule was also found to be important for survival during desiccation stress (Gibson et al., 2006). Anriany et al. (2006) highlighted the importance of an integral lipopolysaccharide (LPS), at both the O-antigen and core polysaccharide levels, in the modulation of curli protein and cellulose production, as well as in biofilm formation, thereby adding another potential component to the complex regulatory system which governs multicellular behavior in *S*. Typhimurium. Mireles et al. (2001) observed that for *S*. Typhimurium LT2, all of the LPS mutants examined were able to form a biofilm on polyvinyl chloride (PVC) but none were able to attach to a hydrophilic surface such as glass. Kim & Wei (2009) noticed that a *rfbA* mutant of *S*. Typhimurium DT104, showing an aberrant LPS profile, was impaired in rdar expression, pellicle formation, biofilm forming capability on PVC in meat, poultry and produce broths and biofilm formation on stainless steel and glass.
