**4.1.2 Phage typing**

354 Salmonella – A Dangerous Foodborne Pathogen

Serotyping is easy to perform and standardized antisera are commercially available. However, it only allows the assignment of *Salmonella* strains to a specific serotype, and no

> *Salmonella* polyvalent antisera (A-E)

+ -

Serotype definition

Non-*Salmonella*

Detection of *Salmonella* flagellar antisera

During the 1980's, a tremendous increase in *S. enteritidis* was identified, particularly in the Northeastern U.S. (Rodrigue et al.,1990). Studies linked *S. enteritidis* to contaminated shell eggs or foods that contained eggs (Mishu et al., 1994). During 1987-1997, five serotypes accounted for 66% of all clinical infections in which a *Salmonella* isolate was identified to the serotype level. *S. typhimurium* accounted for 24% of these isolates, *S. enteritidis* (22%), *S. heidelberg* (9%), *S. newport* (5%) and *S. hadar* (4%) followed (Olsen et al., 2001). When clinical outbreaks were distinguished from sporadic infections, *S. enteritidis* was implicated in 55%

In Tunisia, from 1994 to 2004, 16.214 *Salmonella* isolates were reported to the national Centre of Enteropathogenic bacteria at Pasteur Institute, Tunis, Tunisia. (Ridha et al., 2007). The largest proportion of *Salmonella* isolates was from human origin (n=6815) followed by isolates from food (n=5539). During the surveillance period, the top five reported *Salmonella*  serotypes were: *Enteritidis*, *Anatum*, *Corvallis*, *Braenderup*, and *Livingstone.* These five serotypes accounted for 3479 strains of all *Salmonella* isolates from food. (Ridha et al., 2007). Finally, *Salmonella* isolates reported from environmental origin cam in last position (n=1611)

Serological analysis usually remains the first step in an epidemiological investigation of *Salmonella* and may be sufficient for epidemiological investigations associated with uncommon serotypes (Threlfall & Frost, 1990). However, smaller labs often do not have access to the pools of serum required for this analysis and may need to rely on other techniques to analyze isolates. The multiplex PCR, an easier molecular method, has been developed to differenciate between the most common serotypes of *Salmonella enterica* subsp.

of *Salmonella* cases associated with a clinical outbreak (Olsen et al., 2001).

after isolates from animal origin (n=2249) (Ridha et al., 2007).

*enterica* (Imen et al. 2010).

further differentiation between strains of the same serotype is achieved.

*Salmonella* group A,B, C, D, and E

Detection of *Salmonella* somatic with monovalent antisera

Fig. 2. Serotyping analysis scheme for *Salmonella*

Individual isolates of many *Salmonella* serotypes vary in their susceptibility to lysis by different bacteriophages and this has led to a typing scheme based on reactivity to a panel of bacteriophage. Therefore, a *Salmonella* strain is subjected to a specified set of typing phages and the lytic pattern obtained commonly allows the assignment to a specific phage type. The strains exhibiting a lytic pattern that does not correspond to a known phage type are classified as RDNC (= Reacting with the typing phage, but lytic pattern Did Not Correspond to any recognized phage types).

Phage typing is mostly performed for serotypes such as *S. Typhimurium*, *S. Enteritidis*, *S. Typhi* or *S. Paratyphi*, although phage typing systems are also available for a number of additional serotypes, including *S. Virchow*. Phage typing has led to the discrimination of over 200 *S. typhimurium* phage types (Threlfall & Frost, 1990) and, together with antimicrobial susceptibility analyses, led to detection of several large-scale, international epidemics including the dissemination of a multi-drug resistant clone of *S. typhimurium*  DT104, (definitive phage type, DT, 104) (Threlfall, 2000). In Denmark, phage typing as described by the World Health Organization (WHO) Collaborative Centre for phage typing of Salmonella (Health Protection Agency (HPA), Colindale, United Kingdom) has been applied for surveillance of S. Enteritidis and S. Typhimurium in humans, food and food production animals. Phage typing has proven to be an important tool for strain characterisation and the results obtained have been used since the mid-90s in surveillance, source attribution and outbreak investigations (Baggesen & Wegener, 1994; Hald et al., 2007)

In general, phage typing is only performed by the National Reference Centers, since only these institutions have access to the defined sets of typing phages. The interpretation of the results requires considerable experience (Riley, 2004). Although, phage typing in *Salmonella* epidemiology has been used since the 1950s, the stability of phage types can be limited by phage type conversion (Rabsch et al., 2002), even during an outbreak (Mmolawa et al., 2002). This is due to the acquisition of a temperate phage or a plasmid. Besides, host-controlled phage defence mechanisms such as restriction/modification systems and phage adsorption inhibition are also responsible for the phage typing difficulties of a *Salmonella* strain.

By means of a sterile inoculation loop, the test culture was inoculated into a test tube containing 4 mL double strength nutrient broth with a special care for heavy inoculum to give visible turbidity for *S. Enteritidis* and a very light inoculum for *S. Typhimurium* to give a barely visible turbidity.

The culture was incubated by shaking at 200 rpm at 37°C for 1-1.5 h for *S. Enteritidis* and for *S. Typhimurium* 1.5 h without agitation to obtain a very light growth in early log phase. After incubation, it was flooded over the surface of double strength nutrient agar using a flooding pipette and the excess of culture was removed. As soon as the surface of agar dried, the appropriate typing phages at routine test dilutions were applied to the dried surface by a multipoint inoculation loop. When the phage spots dried, the agar plate was incubated at 37°C for 18 h. At the end of the incubation, the agar plate was read using a magnifying glass through the bottom of the plate (Ward et al., 1987).

Phage susceptibilities were evaluated by means of the plaque number, size and transparency. The pattern was compared with known phage type patterns in the database and defined. If the culture did not react with any of the typing phages, it was defined as non-typable (NT); and if the culture reacted with the typing phages, but gave a different

Laboratory Typing Methods for Diagnostic of

strain specific patterns.

(Lyytikainen et al., 2000).

technique (Liebisch & Schwarz, 1996).

to improve Salmonella spp. surveillance systems.

pattern, even within an outbreak (Echeita & Usera, 1998).

**4.2.2 PFGE (pulsed field gel electrophoresis)** 

Salmonella Strains, the "Old" Organism That Continued Challenges 357

Plasmid profiling is most useful in an outbreak setting that is limited temporally and geographically (Mendoza & Landeras, 1999). Furthermore, this technique will only be

PFGE has been considered as the "gold standard" among other molecular typing methods. By cutting the bacterial DNA with rare-cutting restriction endonucleases and running with special electrophoresis separation technique which use pulsed currents that change polarity at defined intervals, it separates the large fragments of DNA up to 12000 kb and yields

The choice of restriction endonuclease is somewhat empiric, but the most commonly used enzymes in *Salmonella* have been XbaI, SpeI and NotI. Comparisons of patterns from multiple enzymes can elucidate new subtypes and increase the discriminatory power of this

PFGE of 60 *S. enteritidis* isolates revealed 28 different XbaI restriction profiles and 26 with SpeI, yet when the patterns generated from both enzymes were combined, 32 different pulsed-field types could be identified (Ridley et al., 1998). PFGE was used to determine whether molecular subtyping was able to detect unsuspected clusters or outbreaks of *S. typhimurium* (Bender et al., 2001). In fact, during a four-year period, 16% of isolates were linked to common source outbreaks. Of these, the authors felt that 62% of outbreak strains would have been missed without the use of PFGE molecular subtyping (Bender et al., 2001). PFGE has also been used to track outbreak strains occurring across national boundaries

PFGE is characterized by a high degree of reproducibility both within and between laboratories (Swaminathan et al., 2001). The recent introduction of computerized gel-based data collection and analysis systems allows better standardization between laboratories thus creating the ability to rapidly compare restriction fragment patterns from isolates analyzed from remote locations (Swaminathan et al., 2001). Large databanks that house PFGE patterns from isolates around the world will greatly enhance *Salmonella* outbreak detection. PulseNet, a molecular subtyping network for foodborne bacterial disease surveillance, has been active in developing standardized PFGE protocols and establishing a national database. An outbreak of *S. agona* linked to contaminated cereal was identified in 1998. PFGE, in association with PulseNet, was used to identify cases in adjoining states that were not initially thought to be at risk (Swaminathan et al., 2001). In fact, combining typing methods such as PFGE and information from food chains, it was possible to identify related strains and common source of contamination. This type of approach may be useful in order

PFGE, however, is not always successful. Some serotypes, especially those with certain distinct phage types, can be so genetically homogeneous that multiple genotypic techniques fail to discriminate outbreak from non-outbreak strains. Ahmed et al. (Ahmed et al., 2000) evaluated PFGE to differentiate *S. enteritidis* DT8 strains that developed during a Canadawide outbreak of gastroenteritis that was eventually traced to contaminated cheese. Successful discrimination was only achieved with a combination of intensive epidemiological, genotypic and phenotypic methods (Ahmed et al., 2000). Additionally, certain serotypes may be more susceptible to genetic rearrangements that can alter the PFGE

successful if the serotype of interest carries multiple plasmids of differing sizes.

pattern other than those in the database, it was considered as reacting with the typing phages, but lytic pattern did not correspond to any recognized phage types, so called RDNC (= Reacting with the typing phage, but lytic pattern Did Not Correspond to any recognized phage types). But, we must note that phage typing analyses needs typing phage sets to be performed.

In bref, phage typing can play an important role in surveillance and control of the common *Salmonella* serotypes. However, this requires strengthened efforts to make the system available to more laboratories internationally, possibly a simplification of the system to enhance its robustness even though this may slightly compromise its discriminatory power, and finally improved external and internal quality assurance systems.
