It should be remembered that methods need to be accredited by bodies such as AOAC. Simply using one’s own invented technique is not good practice.

Subsections 
Conventional methods 
ISO 6579 Salmonella 2002 method 
Rapid detection techniques 
Separation and concentration techniques 
Immunomagnetic separation 
Impedance (conductance) microbiology 
Enzyme immunoassays and latex agglutination tests 
Nucleic acid probes and PCR 
ATP bioluminescence techniques and hygiene monitoring 

Conventional methods
Conventional methods for microbial detection primarily require the target organism to form a colony on a growth medium (Oxoid Ltd), see also fluorogenic and chromogenic media . These methods require little capital investment and expenditure on consumables is not excessive. However they are labour intensive with regard to media preparation and recording the results. Additionally as a period of incubation is required the procedure requires a time period measured in days. Since the target organism may be in the minority of the total microflora, or sublethally injured, recovery stages may be incorporated prior to selective procedures. The limits of detection are determined by the procedure and appropriate legislation. Hence an approved protocol for Salmonella detection in processed foods must be able to detect one Salmonella cell in 25 grams of material. Limitations of traditional detection techniques have hampered epidemiological studies of food poisoning outbreaks, for example until relatively recently small round structured viruses (SRSV) could only be detected by electron microscopy within 48 hours of illness. Since these conventional procedures are relatively laborious and time-consuming various rapid methods have been developed. Principal rapid methods include immunoassays (ELISA), latex agglutination, impedance microbiology, immunomagnetic separation, luminescence and gene probes linked to the polymerase chain reaction. These procedures either separate and concentrate the target cell or cell fragment to a detectable level by a non-growth step (i.e. immunomagnetic separation). Alternatively the end-detection method does not rely upon an incubation period for colony formation (i.e. luminescence). These rapid methods will be considered later. For detailed conventional protocols there are many standard reference sources; 
Bacteriological Analytical Manual , 
Microbiological Laboratory Guidebook. 
Compendium of methods for the microbiological examination of foods 
Practical Food Microbiology (PHLS UK). 
Modifications of the conventional plating method have included the Petrifilm technology (3M). Accredited Petrifilm products can be checked out here. 

2002 ISO method for Salmonella 
The new ISO 6579:2002 standard for the detetion of Salmonella in food was published in October (2002) and has three changes compared with the 1993 version 

Selenite cystine (SC) broth is replaced by Muller Kauffmann tetrathionate novobiocin broth (MKTTn) which shows superior selection to SC for Salmonella Typhi and Paratyphi (in addition selenite is hazardous to human health). 
Rappaport Vassiliadis (RV) broth has been replaced by Rappaport Vassiliadis Soya (RVS) broth. 
XLD is the first isolation medium rather than BGA. Nevertheless BGA is a standard alternative medium if its use is appropriate for the type of sample being tested. 
The new method has improved recovery of potentially fatal and invasive enteropathogenic strains of Salmonella such as S. Typhi and Paratyphi. 

Rapid detection techniques
Due to the prolonged and laborious nature of traditional methods, numerous rapid and automated methods have been developed and marketed. The definition of `rapid’ has not been formally defined, but it generally means any method which yields results quicker than the standard method. AOAC validated methods can be accessed here. 

The most frequently used rapid and automated methods in industry are Enzyme Linked Immunosorbent Assays (ELISA), impedance (or conductance), immunomagnetic separation (IMS) and bioluminescence. These techniques may be used on there own or in combination. Therefore a rapid separation method such as immunomagnetic separation can be used in conjunction with a rapid end detection method such as ELISA. These methods have been developed either to (a) replace the enrichment step (which requires a prolonged growth period) with a concentration step, i.e. immunomagnetic separation or (b) to replace the end-detection method, which is usually colony development and hence requires a prolonged incubation period, i.e. impedance microbiology and bioluminescence. 

Current methods require approximately 105 organisms/ml for reliable detection. Since the regulatory requirement is the ability to detect 1 cell in 25 gram of food a concentration factor of 107 is necessary. This is equivalent to 2 hours of polymerase chain reaction (PCR) amplification (5 minute cycle) and 2 hours of infection period for Salmonella phages carrying the lux gene. Hence current rapid methods have a minimum period measured in hours. The ideal is an instantaneous or `real time’ in vitro method. Potentially the bioluminescence technique coupled with ultra sensitive luminometers offers such a truly rapid technique. 

Separation and concentration techniques
In order to discriminate the target organism from other cells (procaryotic and eucaryotic) a separation step is normally required. This subsequently generates a large quantity of material of which only a portion is used for further analysis unless a concentration step is also used. For example homogenising a food sample in a blender (Stomacher, etc) dilutes the material ten-fold, generating large volume of material (250 ml), yet the detection procedure may only require a few millilitres. By concentrating the target organism the detection period should be shortened and more efficient. 

Membrane filtration - Direct Epifluorescent Technique (DEFT) and Hydrophobic Grid Membrane
Membranes can be made from nitrocellulose, cellulose acetate esters, nylon, polyvinyl chloride and polyester (Sharpe, 1994). They are very thin and hence can be directly mounted on a microscope. Membrane filters are used in modified conventional techniques for a variety of purposes: 
Concentration of target organism from a large volume to improve detection limits. 
Remove growth inhibitors 
Transfer organism between growth media without physical injury through resuspension 
. The sensitivity of the DEFT results from the concentration of cells by membrane filtration before staining. Its ability to distinguish live and dead bacteria comes from the use of the nucleophilic fluorochrome acridine orange, which fluoresces at different colours in cells during different phases of growth. The dye fluoresces red with RNA and green with DNA. Generally, viable cells fluoresce orange-red while dead cells fluoresce green. In 1991 the ISO-GRID?method for Salmonella was given AOAC approval for use with all foods (Method 991.12). 

Immunomagnetic separation
The isolation stage can be shortened by replacing the selective enrichment stage with a non-growth related procedure. Immunomagnetic separation (IMS) uses superparamagnetic particles are coated with antibodies against the target organism to selectively isolate the organism from a mixed population. The target organism can be detected using standard microbiological procedures; see Table 4.3. for a review of applications. Superparamagnetic are used since they only exhibit magnetic properties in the presence of an external magnetic field. The application of immunomagnetic separation has been reviewed for its use in medical microbiology by 豯svik et al. (1994) and for applied microbiology by Safar韐 et al. (1995). 

Dynabeads?(Dynal A/S, Oslo, Norway) produce superparamagnetic, polystyrene-based particles with diameters of 2.8 祄 (Dynabeads?M-280) and 4.5 祄 (Dynabeads?M-450; Fig. 4.2). Both M-280 and M-450 carriers can be obtained in non-activated and tosylactivated form and therefore can be coated with the antibody of the user’s choice. Additionally coated Dynabeads?with covalently immobilized streptavidin or secondary antibodies against selective primary antibodies are commercially available. An alternative magnetic sorbent is BioMag (Metachem Diagnostics Ltd, Northampton). The particles (0.5 - 1.5 祄 in diameter) are formed by silanized magnetic iron oxide and carry amino-, carboxy- or thiol- groups on the surface. Other sources are Polysciences Ltd (Northampton , UK), Scipac (Sittingbourne, UK) and Sepadyn (Indianapolis, USA). Alternatively, microbial cells can be made magnetisable by the direct adsorption of submicron particles of magnetic iron oxides (ferro fluids) on to their surface (Safar韐 et al 1995). 

Impedance (conductance) microbiology
Impedance microbiology detects microbes either directly due to the production of ions from metabolic end products or indirectly from carbon dioxide liberation (Don Whitley Scientific Ltd ). The detailed physics of the process will not be covered here since the topic has been reviewed by Kell & Davey (1990) and Silley & Forsythe (1996). The direct method monitors impedance (reciprocal of conductance) changes of the growth medium. Microbial metabolism produces ionic end products (organic acids and ammonium ions) from the growth medium and therefore increases the conductivity of the medium. The method is not applicable to media with high ionic strength, such as listeria selection broths, since the initial conductance values is outside the range of the instruments. The indirect technique is a more versatile method in which a potassium hydroxide bridge (solidified in agar) is formed across the electrodes. The test sample is separated from the potassium hydroxide bridge by a headspace. During microbial growth carbon dioxide accumulates in the headspace and subsequently dissolves in the potassium hydroxide. The resultant potassium carbonate is less conductive and it is this decrease in conductance change which is monitored. The time taken for a conductance change to be detectable (`time to detection’) is dependent upon the inoculum size. Essentially the equipment has algorithms which determine when the rate of conductance change is greater than the preset threshold. Initially reference calibration curve is constructed using known numbers of the target organism are constructed. Subsequently the microbial load of subsequent samples will be automatically determined. The limit of detection is a single viable cell since, by definition, the viable cell will multiply and eventually cause a detectable conductance change. Much of the early work on impedance microbiology was in the food and dairy industries. I have written a supplemental Powerpoint presentation on impedance microbiology showing the various types of impedance measurement available (Click for list of presentations) 

Enzyme immunoassays and latex agglutination tests
The combination of enzyme-linked immunosorbent assay (ELISA) and monoclonal antibodies (McAb) have found widespread use in food microbiology. ELISA is most commonly preformed using McAb coated microtitre trays to capture the target antigen. The captured antigen is then detected using a second antibody which may be conjugated to an enzyme. Addition of a substrate enables the presence of the target antigen to be visualised. ELISA methods offer specificity and potential automation. A wide range of enzyme-linked immunosorbent assays (ELISA) are commercially available, especially for Salmonella, pathogenic E. coli and Listeria. The technique generally requires the target organism to be 106 cfu/ml, although a few tests report a sensitivity limit of 104. Hence the conventional pre-enrichment and even selective enrichment might be required prior to testing. Immunoassays may be performed either as the ’standard’ ELISA tray (96 well) format, or the antibodies can be immobilized onto a solid support in the form of a dip stick. See the TECRA products for examples of both approaches. Also the AOAC approved tests can be checked out HERE. 

Nucleic acid probes and PCR 
Food microbiology is slowly being transformed by the opportunities of genetic engineering and the polymerase chain reaction (Stewart, 1997). Subsequently food pathogens can be detected without such an emphasis on selective media. The polymerase chain reaction (PCR) technique amplifies DNA by using a heat stable DNA polymerase in a repetitive cycle of heating and cooling (Saiki et al., 1988). The target DNA is a known DNA sequence for example a food pathogen or the gene for a particular toxin. The sample DNA is mixed with the PCR buffer, Taq, deoxyribonucleoside triphosphates and two primer DNA sequences (c. 20 to 30 nucleotides long). In the first step the reaction mixture is heated to 94oC for 5 minutes to separate the double-stranded target DNA. After cooling, to approximately 55oC for 30 seconds, the primers anneal to the complementary sequence on the target DNA. Subsequently the temperature is raised to 72oC for 2 minutes and the Taq polymerase extends the primers, using the complementary strand as a template. The reaction mixture is reheated to 94oC to separate the double-stranded DNA. Subsequently the replicated target sites act as new templates for the next cycle of DNA copying. After the cycle of heating and cooling is repeated 30 to 40 times the target DNA will have been amplified to a theoretical maximum of 109 copies. The true amount is usually less due to enzyme denaturation but approximately 100 礸 of purified DNA is obtained. The resulting DNA is stained with ethidium bromide and visualised by agarose gel electrophoresis with UV transillumination (312 nm). Negative control samples omitting DNA must be used in order to check for contamination of the PCR reaction by extraneous DNA. Many detection kits have been developed which are specific for food pathogens. A European study (involving 35 laboratories) on PCR methods for foodborne bacteria (Salmonella, Campylobacter, EHEC, Listeria monocytogenes and Yersinia enterocolitica) is underway, they are producing guidelines and kit validation methods, as well as thermocycler recommendations (although a pop-up appears requesting a password, I just clicked it off and obtained the protocol details). 

In time we may have microarrays (also known as DNA chips or DNA arrays, but the technique is not necessarily limited to DNA sequences) which will enable a single sample to be ananlysed for 1000’s of micro-organisms at teh same time. A potential example is the use of 16S signature sequences.

16S signature sequences 
ASM News article 
16S sequences 
Ribosomal database project II 

ATP bioluminescence techniques and hygiene monitoring.
Since the molecule adenosine triphosphate (ATP) is found in all living cells (eucaryotic and procaryotic) its detection is indicative of living material being present (Companies : Biotrace, Celsis, Biocontrol and others). ATP can be detected rapidly by light emission through the combined use of the enzyme luciferase and a luminometer (Stanley, 1989). The technique has a detection limit of 1 pg ATP, which is equivalent to approximately 1000 bacterial cells; 10-15 g ATP per cell. This sensitivity limit applies to commercially available reagents. More sensitive assays are possible with research-grade reagents Sample analysis can be achieved in minutes and is therefore considerably faster than conventional colony counts. The use of ATP measurements as an indication of microbial presence was recognised in the 1960’s. However it has required improvements in luminometer design (reduced cost and portability) and reagent stability to make ATP bioluminescence a more accessible technique in a food factory. There are essentially three areas of application; hygiene monitoring, testing liquids such as final rinses from Clean in Place systems and assessing the bacteriological quality of foods. The first two applications measure the total ATP content of a sample. In contrast, for assessing the bacteriological quality of foods only the microbial ATP must be measured and this requires selective extraction procedures outlined below. 

The firefly (Photinus pyralis) luciferase reaction is the most efficient and best known bioluminescent reaction for ATP determination. It requires the co-factor D-luciferin adn Mg2+ for activity and this is an expensive component in the available commercial kits. A dioxetanone is formed by the activity of the luciferase enzyme with oxygen and a Mg-ATP complex. Subsequently, yellow-green light (max. 562nm) is emitted after a delay of approximately 300ms to a maximum within about one second (Grayeski, 1987). Nowadays a constant (slowly declining) light output is achieved using lower amounts of luciferase than early formulations which resulted in a flash of light. For hygiene testing the total ATP content of the sample is determined. This will include eucaryotic and microbial ATP. To determine the microbial ATP level selective extraction is used. First, non-microbial ATP is extracted with a non-ionic detergent (Triton X-100) and then destroyed by treating with a high level of (typically) potato ATPase for 5 minutes. Subsequently the microbial ATP is extracted using either trichloroacetic acid (5%), an organic solvent (ethanol, acetone or chloroform) which will require subsequent dilution to avoid luciferase inhibition or cationic detergents. Careful timing of mixing and reading is required to allow for luciferase inhibition (Simpson & Hammond, 1991). Since the level of ATP in eucaryotic cells is three orders of magnitude greater than bacterial cells this procedure is difficult to achieve reliably. 

Although it has been recognised for many years that ATP-bioluminescence could be exploited as a means of monitoring the hygienic status of materials, such as work surfaces, the technique required the development of luminometers to measure low light levels. Bioluminescence is now widely used to assess the hygienic quality of a work surface. Samples are taken by sweeping an ATP-free swab over a surface and then measuring the amount of ATP through a series of extract procedures. First generation instruments required the operator to swab a site (usually standardised at 10cm2) and then immerse the swab in a releasing agent, mix with the luciferase-luciferin solution and place in a cuvette for the light reading. Once the second reagents had been added the light emitted had to be measured immediately. Therefore there were a number of pipetting steps plus potential delays in reading the values which were all sources of error. Second generation instruments have the reagents in the swab handle, which are released over the swab tip when required and the whole unit is placed in the luminometer. 

Correlation between standard plate counts and light output from a ATP bioluminescence assay is frequently good (r>0.85). However precise correlation should not expected since the method also detects non-microbial ATP which in many real-life factory situations provides the majority of the ATP pool and non-culturable microbes. The absence (or low colony count) of microbial growth after sampling the test surface would indicate it was microbiologically clean. However food residues can serve enable sublethally injured cells to recover. Hence the combined detection of microbial ATP and food debris is an advantage. A positive ATP value infers the sample is contaminated but does not determine whether it is due to micro-organisms or food residues. Setting `in-house’ acceptance values for luminescence readings can be troublesome, for the reasons given above, if plate counts are used as the `golden rule’. A better approach is to set the level of bioluminescence criteria according to cleaning regimes, that is the values obtained after a good clean and after a deep clean. These readings should be monitored to ensure that standards are maintained or even improved. Subsequently ATP-bioluminescence has become established as a means of monitoring the cleaning regime especially at a Critical Control Point of a Hazard Analysis Critical Control Point (HACCP) procedure. The examples of applications is extremely broad with widespread acceptance.