WO2011098820A1 - Phage-based limulus amoebocyte lysate assay for the rapid detection of bacteria - Google Patents

Abstract

A phage-based Limulus amoebocyte lysate (LAL) assay which is able to detect very low levels of viable bacteria in water and vegetable samples is described.

Description

PHAGE-BASED LIMULUS AMOEBOCYTE LYSATE ASSAY FOR THE RAPID DETECTION OF BACTERIA

This invention relates to the rapid detection of bacteria. More particularly, the present invention describes an assay for the rapid detection of selected, target bacteria in a mixed population or other sample of unknown biological load.

In environmental microbiology, the need for rapid methods to detect specific or target bacteria and to confirm their viability or metabolic activity has been widely acknowledged. It is well known that traditional culture methods for detecting indicator and pathogenic bacteria in food and water detect and confirm the presence of viable bacteria, but the culture methods used to grow sufficient numbers of bacteria for detection are slow.

Methods for rapidly detecting bacteria have been used which utilise the bioluminescent phenomenon of the luciferin-luciferase enzyme reaction in the presence of ATP (Ulitzur and Kuhn, 1989; Walker et al. 1992; Reiprich et al. 2002), or a detectable marker, often the enzyme luciferase, which can be introduced into bacteriophages and then used for bacterial detection (Jassim et al. 1990, 1993, and 1996; Stewart et al. 1998; Favrin et al. 2001 and 2003). These assays generally include a lysing reagent to break open the bacterial cells and release the intracellular ATP; thus the results only give a measure of microbial load rather than are indicative of the presence of specific or target pathogens within the microflora. Three methods, classic bacterial cultures, PCR analysis, and immunoassay, are available for the detection of E. coli in water and the environment (Frampton and Restaino 1993). Culture methods are usually laborious and expensive and require a minimum of 2-3 days to perform (Dey and Lattuada 1998). Although PCR assays may be useful for the examination of human or animal fecal samples, for example, Meng et al. (1996) described a PCR technique that could detect as few as 25 CFU of E. coli within 3h, their usefulness for diagnosis is limited due to their inability to differentiate between viable and non-viable bacteria (Sachse 2004). For immunoassays, although are sensitive, these assays are laborious, expensive, and can not definitely differentiate between viable and dead cells (Chapman et al. 1997). On the other hand, the phage-based LAL assay of the present invention is relatively simple and rapid; it targets only the viable cells at unrivalled specificity/qualitatively due to the use of E. co//-designed specific phages.

The Limulus amoebocyte lysate (LAL) test is widely used to measure lipopolysaccharides (LPS) or endotoxin. LAL is an aqueous extract of blood cells (amebocytes) from the horseshoe crab, Limulus polyphemus (Levin and Bang 1964) and the LAL test is based on an enzymatic reaction triggered by a trace amount of endotoxin or lipopolysaccharide, which is a membrane component of Gram-negative bacteria (Rossignol et al. 2006). This assay is based on the initial research of Levin and Bang (1964) that revealed the role of endotoxin in the extracellular coagulation of Limulus blood. Chromagenic LAL tests use a pyroenzyme from the LAL, a colourless substrate, and an E. coli endotoxin standard (Rokosz et al. 2003)or pyrochrome which is a versatile quantitative chromogenic reagent that may be used to perform either kinetic or endpoint assays in microplate readers.

Newer modifications of the LAL test are chromogenic and quantitative, therefore offering not only greater precision, but also considerably shorter assay times of as little as 10 min (Rossignol et al. 2006; Sakata et al. 2009). Although LAL assay has been used for testing endotoxin contamination in medical devices and parenteral solutions, this assay has not been used as a basis for a rapid detection test for specific bacteria and has only been used for the detection of non-specific mesophilic bacteria. Additionally, this assay has not been used to test for Gram-positive bacteria because of the requirement for the presence of endotoxin.

Using LAL assay coupled with phage assay specific detection of bacteria is not reported in the literature.

Rhee and Kang (2002) used chromogenic LAL endpoint as a rapid assay for the enumeration of total mesophilic microbial loads and coliforms as a means to assess the microbiological quality to detect >10³ CFU ml^-1 in raw milk samples. Siragusa et al. (2000) showed also that LAL assay was found to be an accurate and rapid means of gauging levels of beef carcass mesophilic non-specific microbial contamination.

However, while each of the techniques described detects and enumerates generic Gram-negative bacteria, they lack both the precision and the specificity needed to target specific bacteria in a sample, thus they are not reliable for use in the food, water, and medical industries. In conventional tests, the use of chemical extractants to liberate endotoxin is not specific, it was therefore necessary to develop specific lytic bacteriophages to confer the required specificity.

Therefore it is an object of the present invention to provide such an assay, using specific lytic bacteriophages to facilitate the detection of specific organisms in a sample.

The present invention will be described with reference to the use of coliphages to detect the presence of E.coli in a sample. However, it is not intended that the invention be limited to the use of coliphages as the invention finds equal utility with all bacteria where a specific bacteriophage may be used. For example, a need exists for reliable, rapid and specific detection assays for many bacteria such as environmental Enterobacteriaceae, Pseudomonas spp., Moraxella catarrhalis, Helicobacter pylori, Stenotrophomonas spp., Legionella spp., Acetic acid bacteria, Hemophilus influenzae, Neisseria meningitidis, Neisseria gonorrhoeae, Acinetobacter baumanii, Vibrio cholerae, and Campylobacter spp., which are all Gram- negative bacteria important in various industries for example, food, water, medicine and the like where rapid detection is desirable. For example, in the food and beverage industries, a reliable, rapid specific test would allow for positive release of food and drink product where the manufacturer could ascertain that a product was not contaminated when it left the factory. The method of the present invention may also be used for rapid detection of Gram- positive bacteria by using phages specific for certain cell membrane moieties, for example, at teichoic or teichuronic acids, and peptidoglycan layer

The present inventors therefore used specific coliphages in a LAL assay to produce a very specific, sensitive assay for the detection of contaminant E. coli bacteria in a tested sample.

Detection of E. coli in drinking or swimming water is necessary and is widely used to predict any human sewage or animal faecal contamination (Wang and Fiessel, 2008). Moreover, E. coli detection is considered as more specific than the detection of faecal conforms (Murphy et al. 2008). Until now, most water and food industry safety laboratories use the time-consuming classical methods of E. coli diagnosis which eventually take time between 12 to 24 h (Blanch et al. 2007; Brown et al. 2008).

In the preferred embodiment, the present invention uses the LAL assay reaction to detect strain-specific phage-based lysis of target E. coli bacteria.

Accordingly, the present invention provides an assay for the detection of bacteria, the assay comprising the steps of (i) introducing bacteriophage specific for a target species of bacteria to be detected into a sample, (ii) incubating the bacteriophage-sample mixture for a period of time sufficient to achieve lysis of the target bacteria, (iii) adding Limulus amoebocyte lysate labelled enzyme, (iv) incubating the mixture produced in step (iii), and (v) detecting the release of the label.

An advantage of the use of bacteriophage is that only viable cells are detected since only viable cells are susceptible to bacteriophage infection. Only viable cells are of a concern in the consideration of infection or contamination. Hence, the disadvantage of ATP-based or adenylate kinase- based assays (AKBA) in which viable, dead and lysed cell products are detected is that the potential for producing false positives (non-viable cells) is significantly reduced using the method of the present invention.

Preferably, the Limulus amoebocyte lysate labelled enzyme is labelled with a chromogenic, colorimetric or other optically detectable label. More preferably, the Limulus amoebocyte lysate labelled enzyme comprises a Limulus amoebocyte lysate pyrochrome reagent. The standard Pyrochrome test is read at 405 nm (Rokosz et al. 2003). The used LAL Pyrochrome contains an aqueous extract of amebocytes of Limulus polyphemus, dextran (stabilizer), EDTA, CaC^, MgC^, buffer and chromogenic substrate (Boc-Leu- Gly-Arg-p-nitroanilide). In the presence of endotoxin, factors in LAL are activated in a proteolytic cascade that results in the cleavage of a colorless artificial peptide substrate present in Pyrochrome LAL. Proteolytic cleavage of the substrate liberates p-nitroaniline (pNA), which is yellow and absorbs at 405 nm. The test is performed by adding a volume of Pyrochrome to a volume of specimen and incubating the reaction mixture at 37°C. The greater the endotoxin concentration in the specimen, the faster pNA will be produced (Lindsay et al., 1989). It is preferred that the optical detection of the label may be carried out with conventional laboratory equipment, such as a colorimeter or a spectrophotometer, or, especially for field applications, by eye.

The incubation in step (ii) is continued for sufficient time to allow the bacteriophage to lyse the target bacteria. It is preferred that the incubation time is kept short and so an ideal incubation time would be calculated as the time needed to produce sufficient endotoxin or other target substrate for detection according to the detection sensitivity or thresholds of the Limulus amoebocyte lysate labelled enzyme used. In turn, this depends on allowing sufficient (the minimum) numbers of target bacteria present in the sample to be lysed to produce enough endotoxin or other target substrate for detection. However, in a practical application it is preferred that the incubation time is about 30 minutes.

It is preferred that the incubation is conducted at or close to 37°C to speed the lysing of the target cells. However, the incubation may be conducted at ambient temperature, and this may prolong the actual time needed to achieve the abovementioned lysis. Hence, in the most preferred embodiment the incubation is conducted at 37°C for about 30 minutes.

The incubation in step (iv) is continued for sufficient time to allow the Limulus amoebocyte lysate labelled enzyme to react with the endotoxin or other target substrate released during cell lysis. Under laboratory conditions, this is likely to be no longer than an hour. Preferably, the incubation is of between 10 and 40 minutes, and preferred incubation times are 10, 20, 30 or 40 minutes according to sample size, suspected bacterial load and the nature of the target bacteria.

Preferably, the target bacteria is a Gram-negative bacterial strain. More preferably, the target bacteria is selected from the group comprising Enterobacteriaceae, Pseudomonas spp., Moraxella spp., Helicobacter spp., Stenotrophomonas spp., Legionella spp., Acetic acid bacteria, Hemophilus spp., Neisseria spp., Acinetobacteria spp., Vibrio spp., and Campylobacter spp. or mixtures thereof. For example, mixtures of bacteriophages specific for various strains of Enterobacteriaceae may be incubated in a sample suspected of containing Enterobacteriaceae. Also, mixtures of bacteriophages specific for each strain may be used to ascertain a preliminary indication of the presence of bacteria before determining which species of bacteria is present.

Preferably, the bacteriophages are highly selective for the more notable strains of the abovementioned pathogens, for example E.coli, Moraxella catarrhalis, Helicobacter pylori, Stenotrophomonas maltophilia, Legionella pneumophila, Acetic acid bacteria, Hemophilus influenzae, Neisseria meningitidis, Neisseria gonorrhoeae, Acinetobacter baumanii, Vibrio cholerae, and Campylobacter jejunii or mixtures thereof.

More preferably, the target bacteria are E.coli, and ideally EHEC E.coli.

In the most preferred embodiment for E.coli a mixture of highly specific lytic bacteriophages designed using the method of co-pending application number PCT/GB 2009/051641 is used to test for the presence of multiple strains of E.coli.

To advantage, the present inventors have found that the assay of the present invention provides both a quantitative and qualitative measure of bacteria within a period of less than one working day, ideally less than 70 minutes. In this way it should be possible to identify rapidly bacterial contamination in food or in any other sample, the biocidal potential of the bacteria and possibly a flawed or faulty sterilization process (which has resulted in the contamination or infection); thereby conferring the ability to take correction action immediately. Real time or near real time methods for detecting microorganisms are essential for implementation of a Hazard Analysis Critical Control Point (HACCP) program in any food and beverage plant (Cutter et al. 1996; Northcutt and Russell 1996). The LAL coupled phage assay of the present invention is able to detect <100 cells with high accuracy of qualitativity and specificity to a strain of Gram-negative bacteria within 70 min.

In an additional, optional, method step, the amount of endotoxin released during cell lysis is measured by the Limulus amoebocyte lysate labelled enzyme reaction to provide a quantitative detection assay for Gram- negative bacteria in that the yield of label detected in step (v) is indicative of the bacterial load lysed by the bacteriophage and hence the bacterial load of the sample. The sample may be any material suspected of or liable to have been infected by Gram-negative bacteria, for example foodstuffs, beverages, water including that from waterways, lakes or ponds, soil, medical or veterinary samples including samples from the body, medical or veterinary devices including implants and equipment, agricultural samples such as supplies, soil, fields, rice fields or water sources, brownfield sites, or other land areas. Preferably, the sample is a liquid sample or is made into a liquid sample for example by forming a slurry, suspension, homogenate or emulsion. Notable samples include, but are not limited to urine, stool, blood, pleural fluid, potable water, drinking water reservoirs, rivers, foodstuffs and beverages.

Embodiments of the invention will now be described by way of example only, with reference to and as illustrated by the accompanying drawings of which,

Figure 1 is a diagram showing the plotting of log AW test (continuous plot) and ΔΙΙΙ NegCon (dashed plot) in LAL assay versus log™ CFU/well for 10 known EHEC samples at pyrochrome incubation times 10, 20, 30, and 40 min. Dark bordered rectangle area on right is magnified showing the minimal detection threshold of LAL at bacterial titre 10² CFU for incubation time 40 min. The dotted two-headed arrow shows the difference in log Δΐυ between the negative control and test of 40 min pyrochrome incubation. The greater difference between log™ AWJtest and log™ AWJNegCon the higher positive detection achieved;

Figure 2 shows a Linear regression plot between known X values (the measured AIU of endotoxin) and the known Y values (Log™ CFU) for pyrochrome incubation times 10, 20, 30, and 40 min (graphs a, b, c, and d respectively);

Figure 3 is a diagram showing the plotting of log™ ARLUtesf (continuous plot) and ARLVNegCon (dashed plot) in AKBA assay versus log™ CFU/well for 10 known EHEC samples at ADP incubation times 10, 20, 30, and 40 min. Dark bordered rectangle area is magnified showing, unlike LAL, no minimal detection threshold at bacterial titre 10² CFU. The greater difference between log™ ARLUtesf and log™ ARLD NegCon the higher positive detection achieved, and Figure 4 shows a comparison in sensitivity (a and b) and specificity (c and d) of positive detection of known EH EC between LAL and AKBA assays at bacterial burdens 10³ (a and c) and 10⁴ (b and d) CFU for incubation times 10, 20, 30, and 40 min. It was shown that sensitivity and specificity of LAL were higher than that of AKBA at all tested incubation times and bacterial burdens. The sensitivity for LAL and AKBA increased with incubation time while the specificity did not increase or (d) decreased.

Example 1 : Phage-based Lumulus Amebocvte Lvsate (LAL) assay

Materials and Methods

Preparation of the anti- E. coli phage cocktail

Media

Luria broth (LB): tryptone 10 g I^"1 (HiMedia, Mumbai, India), yeast extract 5 g 1 (HiMedia, Mumbai, India), and sodium chloride 10 g I^-1 (HiMedia, Mumbai, India) at pH 7.2 were used in all the protocols. L-agar (LA), consisted of the above with the addition of 14 g I^-1 agar (HiMedia, Mumbai, India) was used for culture maintenance. Bacterial dilutions from 18 h LB cultures grown at 37°C were carried-out in phosphate buffered saline (PBS, Oxoid, UK). For plaque assay, the 'soft layer agar' used was LB prepared in Lambda-buffer [6 mmol 1 Tris pH 7.2, 10 mmol M Mg(SO₄) ₂-7H₂O, 50 pg mM gelatin (Oxoid, UK)] and supplemented with 4 g I^-1 agar bacteriology No. 1 (HiMedia, Mumbai, India).

Bacterial strains

Four hundred and thirty clinical isolates of EHEC and non-EHEC E. coli were obtained from human sources, hospital inpatients, (Microbiology laboratories, Hospital Serdang and Hospital Kajang in Selangor, Malaysia) including documented sporadic cases of haemorrhagic colitis, non-haemorrhagic colitis, urinary tract infections, infected wounds, vaginitis, and bacteraemic cases. Furthermore, other clinical isolates were obtained from animal sources. They were reconfirmed by using Microbact GNB 12A system (Oxoid, UK), a microtitre well-scaled chemical test. Microbact system has 100% sensitivity for identifying E. coli from other Enterobacteracea bacteria. In addition, several E. coli reference strains were used: one EHEC NTCC 129001 and five non-EHEC (two are generic strains; ATCC 12799 and NTCC 9001 , three human enteropathogenic strains (EPEC); ATCC 12810, ATCC 25922, and ATCC 35218 (zoonotic). Both E. coli clinical isolates and representative NTCC and ATCC E. coli strains were used throughout phage isolation, propagation, optimization and designing. The strains were maintained on L-agar plates and transferred bimonthly. All cultures were stored at -20°C in 15% glycerol. Prior to investigation a stock culture of the bacteria was maintained on LA plate. One loopful of the bacterial strain was inoculated into a 100 ml Erlenmeyer flask containing 10 ml of LB and incubated for 18 h at 37°C and 90 rev min^-1 in an incubator shaker (Innova 4000, New Brunswick Scientific). For experimental tests appropriate serial dilutions were made in LB.

Bacteriophages

All wild bacteriophages (phage) used in this study were isolated from and specifically designed for 430 clinical isolates and 6 reference strains of EHEC and non-EHEC E. coli according to the phage breeding method of PCT/GB2009/051641 derived from UK Patent Application No. 0822068.3. The phage master mix was composed of 140 phages that were previously isolated, designed, and produced by two types of novel design techniques (Jassim et al. 2010). The first technique is the chemical vertical designing which is characterized by enhancing the lytic infective criteria of the designed phages in order to obtain optimized biokinetic potential. The second technique is the chemical horizontal design which is characterized by altering the specificity of the designed phages to be reoriented to new strains of E. coli leading to wider coverage of target bacteria (Jassim et al. 2010). The resultant phages were mixed together forming what is called the 'phage master mix'.

Samples processing

Artificially inoculated samples

Ten clinical isolates of known Enterohaemorrhagic E. coli (EHEC) O157:H7 obtained from human inpatients Hospital Serdang and Hospital Kajang in Selangor state, Malaysia were artificially inoculated into two different samples, plain water and lettuce. Lettuce samples were soaked in 500 ml of PBS (Sigma, USA) supplemented with 10⁷ CFU ml^-1 of mixed strains of EHEC for 2 h at 20°C. Then, 100 g cut portions of lettuce was placed in a Stomacher bag and 100 ml of sterile PBS was added. The bag was placed in Stomacher and washed for 2 min. The soiled PBS was collected and centrifuged for 2500 x g for 5 min at room temperature. The bacterial pellets were resuspended in 10 ml of sterile PBS and washed 3 times via subsequent centrifugation 2500 x g for 5 min at room temperature. The concentration of E. coli was measured by the standard plating method on LA for 18 h. For both plain water and lettuce PBS washings the bacterial titres were adjusted for LAL and AKBA assays to be 10^1"7 CFU/microplate well.

Naturally contaminated samples

Thirty out of 74 water samples, obtained from low-medium hygiene domestic lakes (ponds) were shown to harbour different titres of unknown E. coli bacteria which were shown to be non-EHEC by using Mckonckey Sorbitol Agar. These 30 samples were used to detect the presence of E. coli bacteria via the rapid phage-based LAL detection test of the present invention. Two hundred millilitre of lake water was filtered through a syringe sterile filter membrane [25 mm diameter / 0.45 μιτι pore size; Millipore (Canada) Ltd, Mississauga, ON, Canada].

To recover the bacteria, the membrane filter was transferred into a 5-ml sterile glass test tube containing 1 ml of PBS and the tube was vortexed vigorously for 1 min and centrifuged at 2500 x g for 5 min. This was repeated 3 times. The pellet from the last wash was re-suspended in 1 ml of PBS. The bacterial titre was checked by using the standard plating method on L-agar plates in order to be compared later with the quantitative results of rapid detection testing.

Principle of LAL coupled with phage-based detection

LAL assay is a technique used to measure the endotoxin content in a sample expressed in endotoxin unit (EU) or international unit (IU). In the present innovation LAL was used as a reporter for E. coli in a sample. The principle was to exploit the oozing of lipopolysaccharides (LPS) from damaged cell walls of target bacteria caused by the lytic effect of specific phages and accordingly detect the presence of target bacteria. The phages used were a mixture of highly lytic designed phages, namely, phage master mix against pathogenic E. coli strains (Jassim et al. 2010). LAL assay was used to measure the difference between the extracellular endotoxin level in a sample containing combination of the phage master mix and target bacteria before and after the burst time of phages. Test samples were compared with positive control, using chemical extractant instead of phages, negative control, and incompatible phage: bacteria mixture, and the control standard endotoxin (CSE) solution which serves for the quantization of the measured endotoxin using the linear regression equation of the standard curve.

Strategy of LAL testing

The phage master mix was ultracentrifuged using Beckman L2-65 ultracentrifuge (Beckman Instruments, Inc., Fullerton, USA) 80,000 x g for 8 h twice to get phage samples devoid of endotoxin contamination. Bacterial samples, plain water or washing PBS (Sigma, USA) of lettuce that contain known E. coli contaminant, were centrifuged 2500g for 5 min and resuspended to the original volume (1 ml). This was repeated 3 times in order to obtain Bacteria-free endotoxin samples. Bacterial samples were subjected to standard plating methods to measure the bacterial concentration. The titres of target bacteria samples were used in LAL testing; 10^{1 "7} CFU/microplate well. On the other hand, the multiplicity of infection (MOI) of the used phage master mix was 100. Variable incubation times of the LAL pyrochrome reagent (Associates of Cape Cod incorporated, USA) were tested, namely 10, 20, 30, and 40 min. Four main objectives were principally pursued; can LAL testing be used to detect a specific strain or species of bacteria in a very short time, the second objective was to determine the minimal bacterial concentration that can be detected, the third objective was to determine the optimal time for the detection of each bacterial concentration, and the fourth objective was to render the innovated LAL assay into a precise quantitative assay.

Procedure

Test samples

Fifty microlitre phage master mix composed of 172 highly lytic designed coliphages including 22 EHEC-specific designed phages, were added to 50 μΙ of the test bacterial sample (10^1"7 CFU) in a sterile 96 wells microplate (Sterilin, UK). The phage:bacteria mixture was incubated for 30 min (the burst time) at 37°C. Afterwards, 50 μΙ of the LAL pyrochrome reagent were added to the test phage:bacteria samples then followed incubation for 10, 20, 30 and 40 min at 37°C. At the end of the pyrochrome incubation period, OD was measured by spectrophotometry (Bio-Rad Laboratories, Hercules, Canada) at wavelength 405 nm which was named as ODtest2. ODtestI was calculated at 405 nm by the summation of the OD value of 50 μΙ phage alone plus 50 μΙ of bacterial sample alone for the corresponding pyrochrome incubation period at 37°C. Therefore, ODtestI represents the summation of background OD value of the phage and target bacteria. Triplicate wells for each test sample were conducted.

Controls and standards

The negative control samples, in triplicate, were prepared similar to the test samples save for using incompatible phage:bacteria combinations. This ensures that no lytic reaction might take place between the phages and bacteria used in the negative control wells. ODnegCon l value was measured at 405 nm by the summation of the OD value of 50 μΙ phage alone plus 50 μΙ of bacterial sample alone for the corresponding pyrochrome incubation period at 37°C. ODnegCon2 was measured at 405 nm after adding the pyrochrome reagent for the corresponding period which was added after 30 min from mixing the incompatible bacteria and phages together.

Triplicate of positive control samples were prepared by mixing target bacteria with chemical extractant, benzalkonium chloride (BKC) (0.5 mg ml^-1 ; Merck, Germany) for 15 min as a lysing agent to be compared with phage lytic potential. ODposCon l was measured at 405 nm after adding the pyrochrome reagent for the corresponding incubation period before adding BKC to the target bacteria. ODposCon2 was measured after adding to the target bacteria the pyrochrome reagent for the corresponding incubation period. The control standard endotoxin (CSE) (Associates of Cape Cod incorporated, USA) was used to quantify the measured OD values in terms of IU of endotoxin. Serial concentrations of CSE were prepared; 0.005, 0.05, 0.5, 5, 50 IU ml^-1. Regression line equation was applied to measure the predicted IU endotoxin concentration for each OD reading. Interpretation of the assay

By linear regression equation, OD was converted to IU of endotoxin. The difference in values between; \ test1 and \ tets2 (for each pyrochrome incubation period and for each bacterial concentration) was evaluated in terms of significance. AOD represents the net increase in OD value after exposing target bacteria to phages (specific biological extractants) or chemical extractants, BKC. Accordingly, ΔΙΙΙ represents the net amount of endotoxin liberated extracellularly in response to the used extractants. Moreover, the difference magnitude (ΔΙΙΙ) of test (AlUtesf), negative control (AW negCon) and positive control (AW posCon) samples were compared with each other. The positive detection of target bacteria was considered when \ test2 was significantly higher than \ test1 and the AW test was close to AW PosCon and significantly higher than AW NegCon. Since the number of target bacteria is known, the ΔΙΙΙ per bacterium (ΔΙΙΙ-bac) was calculated. ΔΙΙΙ-bac represents the amount of endotoxin which can be liberated from each bacterial cell of E. coli using the designed phage master mix-based lysis for certain pyrochrome incubation period. Accordingly, the predicted number (pr. no.) of bacteria is calculated by dividing the ΔΙΙΙ over ΔΙΙΙ-bac or by using the linear regression equation between ΔΙΙΙ of endotoxin and log CFU. Hence, LAL detection test is a specific quantitative assay for bacterial detection. Furthermore, the sensitivity and specificity of LAL assay were calculated as well as the correlation coefficient (r) between the number of the detected bacteria and the real number shown by plating method.

The performance of the formulated LAL assay was compared to that of the AKBA of luciferin-luciferase. Because phage-based AKBA is usually of limited detection threshold, up to 10³ - 10⁴ CFU and because of it is expensive settings, designing a new assay able to override these drawbacks became necessary. Moreover, the newly designed phage-based LAL assay and AKBA were tested in conjunction with of use of the newly designed 172 phages (Jassim et al. 2010) in a phage master mix. Therefore, qualitative and quantitative E. co//-specific rapid detection testing now becomes attainable because of the wide coverage of almost all pathogenic strains of E. coli by the designed phage master mix. Comparative Example 1 : Phage-based adenylate kinase bioluminescence assay

Principle of the assay

Adenylate kinase bioluminescence assay (AKBA) was conducted on the same artificially inoculated lab samples of EHEC that were used for LAL assay in order to apply a reliable comparison between these two assays. Rapid cleanliness testing using ATP and Adenylate kinase bioluminescence have become widely accepted methods to monitor the hygienic status of food production lines and verify effective cleaning procedures (Kyriakides and Patel 1994; Russell 1995). Adenylate kinase is a key intracellular enzyme with a role to equilibrate concentrations of the adenine nucleotides within the cell, by the reaction shown below:

Adenylate kinase:

ADP + ADP « » ATP + AMP

Mg2⁺

Squirrell and Murphy (1994) proposed the use of the intracellular enzyme adenylate kinase as a bacterial cell marker in place of ATP. Adenylate kinase is present in both eukaryotes and prokaryotes and has a low molecular weight (20-30 kDa). It is most abundant in the mitochondria of tissues such as liver and muscle in which there is considerable energy turnover (Siekevitz and Potter 1953).

Procedure

Test samples

Triplicate of each test sample and controls were subjected to AKBA assay using bioluminescence white 96 microplates (Sigma, USA). The preparation of samples conducted was the same as that carried-out for LAL assay regarding the purification of phages and bacterial suspensions. Fifty microliter of 1 :1 v/v phage master mix with target bacteria at MOI 100 were incubated for 30 min at 37°C in order to let phages lyse completely target bacteria. Afterwards, 50 μΙ of 10 mmol M ADP (Sigma, USA) and 50 μΙ of buffer (50 mmol ¹ Tris+ 15 mmol I ¹ MgCI_2;Merck, Germany) at pH 7.4 were added to the mixture of phage:bacteria and incubated for 10, 20, 30, and 40 min at 37°C. At the end of the AKBA incubation period, 50 μΙ of a luciferin- luciferase mixture (Calbiochem, USA) in 25 mmol I^"1 Hepes buffer (Merck, Germany) were added at semi-dark environment and within seconds light emission reaction was read using endpoint assay of GloMax 96 microplate luminometer (Promega, USA).

Controls and standards

Triplicates of negative controls (incompatible phage:bacteria mixtures) and triplicate of positive control samples were prepared by mixing target bacteria with chemical extractant, benzalkonium chloride (BKC) (0.5 mg ml^-1 ; Merck, Germany) for 15 min as a lysing agent to be compared with phage lytic potential were prepared. Just before adding luciferin-luciferase mixture, the positive control was diluted 1 :50 to avoid BKC inhibitory effect on luciferase enzyme. The dilution factor (1 :50) were taken into consideration for bioluminescence readings. Linear regression equation was used for the standard curve of the standard ATP solution to get the RLU of the bioluminescence reaction.

Interpretation of the assay

The differences in relative light units (RLU) values between; RLVtestl and RLUtete2, RL\JnegCon 1 and RL\JnegCon2, RWposCon I and RL\JposCon2 (for each ADP incubation period and for each bacterial concentration) were evaluated in terms of significance. The difference magnitude (ARLU) of test (ARLUfesf), negative control (ARLU negCon) and positive control (ARLVposCon) samples were compared with each other. The positive detection of target bacteria was considered when RLUtes£2 was significantly higher than RLVtestl and the ARLUtesf was close to ARLU PosCon and significantly higher than ARLVNegCon. Moreover, like LAL assay, the sensitivity and specificity of AKBA assay was checked.

Statistical analysis

SPSS version 12.0 and Microsoft Excel 2000 were used for the analysis of the current study statistics. The IU of endotoxin and RLU of bioluminescence were measured by using linear regression equation of the standard curve, which was repeated at every run, of CSE and reference ATP in LAL and AKBA assays, respectively. The used equation for linear regression was Y = a + bX, where X is the explanatory variable and Y is the dependent variable, b is the slope of the line, and a is the intercept. Student /-test was used to compare pair-wisely among the mean ΔΙΙΙ- or ARLU-tesf, -PosCon and - NegCon values at different incubation times and for different bacterial titres in LAL and AKBA assays, respectively. For lake water samples, Pearson correlation coefficient (r) was calculated to measure the correlation between the predicted number of bacteria measured by LAL assay (pr. no.) and the real bacterial number measured by standard plating method. For the artificially inoculated samples, the sensitivity and the specificity of the LAL and AKBA assays were calculated from the test and the negative control samples respectively as LAL and AKBA results were compared with that of the standard plating method (golden standard). For the unknown E. coli samples of lake water, 74 samples (30 positive for E. coli and 44 negative for E. coli) were used to measure the sensitivity and specificity of the assay for this category of samples. P value less than 0.01 was considered as significant for Pearson (r) and t- tests.

Results

Phage-based LAL detection testing

The Positive detection of target bacteria was achieved when the difference between \ test1 and IUtes/2 was significant and AW test was insignificantly lower than AW posCon and significantly higher than AW negCon. The results in Table 1 shows the positive detections as indicated in bold font. Two sets of triplicates of ten strains of EHEC O157:H7 were artificially inoculated into water and lettuce and were used to optimize the protocols of LAL assay for E. coli detection. Bacterial concentrations 10^1"4 CFU/well were only shown in Table 1 as the higher concentrations, whilst 10^5"7 CFU/well are not critical for the evaluation of rapid detection assays are presented in Fig. 1 .

The quantitative measurement of endotoxin liberation can be used to count the number of detected bacteria. Since the standard curve of CSE was used for each run, AOD was converted to AIU of endotoxin. AlU-bac was calculated for each bacterial concentration and for each pyrochrome incubation time. The relationship between log of bacterial burdens of EHEC (10² to 10⁴ CFU) and logio of AIU was linear (Fig. 1 ). Therefore, the generic mean ΔΙΙΙ-bac for bacterial concentrations (10² to 10⁴ CFU) at each pyrochrome incubation time was calculated out of mean ΔΙΙΙ-bac of three bacterial burdens, namely 10², 10³, and 10⁴ CFU. Accordingly, this generic mean is statistically reliable only for bacterial burdens (10² - 10⁴ CFU). The ΔΙΙΙ-bac of higher bacterial concentrations should be of different value.

The mean AlU-bac for bacterial burdens (10² - 10⁴ CFU):

- 40 min is 0.00388 IU

- 30 min is 0.0027 IU

- 20 min is 0.0024 IU

- 10 min is 0.002 IU

The mean of AlU-bac represents the amount of endotoxin detected per target bacterial cell using the innovated protocols for each pyrochrome incubation time. Accordingly, the mean AIU of every bacterial concentration and every pyrochrome incubation time was divided by the generic mean AlU-bac of the corresponding pyrochrome incubation time to obtain the predicted CFU/well, or pr. no., of the tested samples.

In addition, another quantifying method, a linear regression plot was conducted between the measured AIU of endotoxin and the known Logio CFU for pyrochrome incubation times 10, 20, 30, and 40 min (Fig. 2) in order to predict the logio CFU/well based on regression equation Y = a + bX, where X is the explanatory variable and Y is the dependent variable, b is the slope of the line, and a is the intercept. The correlation coefficient and regression index for LAL assay at 10, 20, 30, and 40 min were all above 0.9 indicating a very high linear relationship between the measured AIU of endotoxin and the actual logio CFU/well. This allows the currently formulated phage-based LAL assay to be reproduced by other sets of experiments and as follows:

- 40 min: y= 1 .89+ 0.073^* X, regression index R² 0.92 (P<0.01 )

- 30 min: y= 1 .89+ 0.077^* X, regression index R² 0.92 (P<0.01 )

- 20 min: y= 1 .91 + 0.08^* X, regression index R² 0.93 (P<0.01 )

- 10 mi: y= 1 .99+ 0.084^* X, regression index R² 0.93 (P<0.01 ) The pr. No. figures of target bacteria using either the generic mean ΔΙΙΙ- bac or the linear regression methods were very close (P>0.05). Therefore, both methods can be used to quantify the CFU of the detected EHEC.

Table 1 shows that phage-based LAL assay proved to be a very sensitive assay for the detection of target bacteria as low as 10² bacterial cell of EHEC. It was shown that the higher the bacterial concentration the lower pyrochrome time is needed (Table 1 and Fig. 1 ). For bacterial concentration 10² CFU/well, pyrochrome incubation time of 40 min was needed as minimum time for the detection of target bacteria. The minimal pyrochrome incubation time for bacterial concentrations 10³ and 10⁴ was 10 min but the lowest best time, in terms of sensitivity/specificity, was 20 min. On the other hand, the ΔΙΙΙ-bac was shown to be different at each pyrochrome incubation time, increasing with the increase of pyrochrome incubation time. Therefore, for quantization purposes, certain ΔΙΙΙ-bac must be used for each pyrochrome incubation time in order to get as precise as possible the predicted number of target bacteria. It is noteworthy to mention that there was no significant difference in terms of AIU and sensitivity/specificity between EHEC-inoculated plain water and EHEC-inoculated lettuce washing PBS.

After optimizing phage-based LAL protocols, they were challenged against 30 samples, in triplicates, of low hygiene lake water. These samples were contaminated with unknown wild strains of E. coli. The E. coli burden was enumerated by standard plating method along conducting LAL assay. Accordingly, 30 samples were grouped into 3 categories according to the bacterial titres:

E. coli at titre of 50-10² CFU/microplate well

E. coli at titre of 10²- 10³ CFU/microplate well

E. coli at titre >10³ CFU/microplate well

These tests were a proof challenge for two main aspects. First, evaluate the optimized protocols so far achieved. Second, evaluate the coverage of the designed coliphages against unknown environmental E. coli strains. The pyrochrome incubation times used were only 20, 30, and 40 min. The results shown in Table 2 provided evidence that the used designed phage master mix covered well the randomly selected samples of lake water E. coli wild strains. The sensitivity and specificity of the LAL assay for the detection for unknown wild strains of E. coli (Table 2) was very close to these for the known laboratory EHEC strains (Table 1 ). This indicated strongly that the positive detection of both wild EHEC and non-EHEC bacteria using the current phage-based LAL assay was almost the same. Moreover, this provided evidence that the so-called "phage master mix" used (known from PCT/GB2009/051641 ) and the LAL assay protocols of the present invention were both reliable and highly efficient for detecting a wide range of different unknown environmental E. coli strains within the rapid time frame of 50 to 70 min. For both EHEC and the environmental E. coli strains the predicted number of bacterial cells was so close to the actual mean log CFU/well, the difference was less than log™ 0.7 as shown by the standard plating method highlighting the quantitative reliability of LAL assay. The Pearson correlation coefficient (r) was calculated between the pr. no. and the real bacterial titre, shown by the standard plating method, for each bacterial titre detected and at each incubation period. It was found that the lowest r was +0.86 and the highest was +0.92. This provided extra evidence for the significant and strong positive correlation between pr. no. and the real bacterial number which gives more reliability to use LAL assay of the present invention as a quantitative and qualitative assay for E. coli rapid detection.

Phage-based AKBA detection testing

AKBA was conducted on the same above ten strains of EHEC-artificially inoculated samples. The same criteria of determining the positive detection of target bacteria in LAL assay were pursued in AKBA. Positive detection of target bacteria was achieved when the difference between RLVtestl and RLUtes/2 was significant and ARLUtesf was insignificantly lower than ARLVposCon and significantly higher than ARLVnegCon.

AKBA was conducted on the same 10 known EHEC strains using the same designed phage master mix for 10, 20, 30, and 40 min incubation times at 37°C. The used EHEC concentrations were of wide range 10² to 10⁷ CFU/well (Fig. 3). However, only 10² to 10⁴ CFU/well were shown in Table 2 as higher concentrations are not critical for AKBA evaluation. Unlike LAL assay, AKBA was not capable to detect E. coli at bacterial titre of 10² CFU/well. The minimal threshold of E. coli titre detected by AKBA was 10³ CFU/well at incubation time 20 min at sensitivity/specificity 74/78. For bacterial concentration 10⁴ CFU/well, AKBA was capable to detect target bacteria within just 10 min at relatively low sensitivity/specificity, 72/78. Nevertheless, 20 min incubation period of ADP, for bacterial titre 10⁴ CFU/well, gave higher sensitivity/specificity, 85/83.

In general, it was shown that LAL assay appeared more sensitive and specific than AKBA in all bacterial titre by detecting low bacterial concentrations, up to 10² CFU/well (Fig. 4). The sensitivity of positive detection in both LAL and AKBA, at 10^{3 " 4} CFU/well, was increasing with assay incubation period. However, the rate of increase slowed down after the incubation period of 30 min (Fig. 4 a and b). On the other hand, the specificity of positive detection in LAL and AKBA, at 10³ CFU/well, was slightly increasing with assay incubation time (Fig. 4 c) while, at 10⁴ CFU, it was decreasing with assay incubation time (Fig. 4 d). This indicates that specificity of LAL and AKBA assays does not decrease with increase of assay incubation time at lower bacterial concentrations, less than 10⁴ CFU. On the contrary, it decreases clearly with increase of assay incubation time at higher bacterial concentrations which provides evidence that, in general, diluting samples to 10^{3 " 4} CFU/well for assay incubation periods 20 - 30 min are considered optimal for both LAL and AKBA assays in terms of sensitivity and specificity. Like LAL assay, there was no difference in terms of ARLU and sensitivity/specificity between water and lettuce washing PBS samples.

The present inventors designed the ultimate specificity conferred by using highly specific and lytic designed phages against E. coli bacteria in formulating high sensitivity/specificity LAL assay in comparison with the AKBA control test. The LAL assay of the invention is able to detect target bacteria specifically within only 70 min. Therefore, this method provides a specific rapid detection assay of E. coli bacteria or of any other Gram-negative bacteria. The performance of phage-based LAL assay was compared with the well known phage-based AKBA. Although, the methodology of AKBA is not new, the use of a mixture of highly specific and lytic phages, 172 designed phages including 22 EHEC-specific phages, against E. coli bacteria via novel non-genetic phage design technique is considered innovative.

The minimal threshold of EHEC concentration detected by AKBA was 10³ CFU/well at incubation time 20 min at sensitivity/specificity 74/78. On the other hand, the minimal threshold for positive detection of EHEC at LAL assay was 10² CFU/well, which is one log lower than that of AKBA, and was of higher sensitivity/specificity, 88/81 . Accordingly, for LAL assay, the overall time needed for detecting the minimal level of EHEC bacteria (<100 CFU) at good sensitivity/specificity was 70 min. This threshold of bacterial detection was found not possible to be achieved using AKBA. In addition, when the minimal threshold of AKBA was compared to the corresponding incubation time and bacterial concentration in LAL, namely 10³ CFU/well for 20 min, LAL sensitivity/specificity, 91/86, proved again to be superior on the corresponding AKBA sensitivity/specificity, 74/78. Therefore, LAL assay proved to be superior to AKBA in terms of the sensitivity, specificity, and minimal detection limit. Another advantage of LAL over AKBA, the LAL materials, reagents, and instruments are much cheaper than of AKBA. The spectrophotometer is cheaper and more readily available than a luminometer and endotoxin pyrochrome is cheaper and more stable during storage than luciferin: luciferase enzyme complex. Moreover, false positive results in LAL assay are lower than in AKBA as ATP contamination takes place more easily than that from endotoxin because ATP contamination might take place from any mammalian or prokaryotic cells (all of which contain ATP).

It is noteworthy that it has been reported previously that ATP method could only detect 10⁵ CFU ml^" with a 50 μΙ sample size and when the bacterial sample size increased to 2 ml, there was a 1 log increase in sensitivity (Trudil et al. 2000). In general the ATP detection limit ranges from 10⁴-10⁵ CFU (Dostalek and Branyik 2005; Wilson et al. 2007; Noda et al. 2008) in which it does not provide sufficient sensitivity for some industrial and clinical applications. By the same token, AK assay, employing lytic phages to release intracellular AK reported a detection limit of 10⁴ CFU ml^" for both E. coll and Salmonella newport (Blasco et al. 1998). Nonetheless, the results from the phage-based AKBA for E. coll and EHEC from water and lettuce samples demonstrated the AK system could readily detect 10³ CFU from 50 μΙ sample size. Therefore, we can conclude that the designed phages technology has demonstrably improved phage test sensitivity for the AKBA and has subsequently increased the RLU by 1 to 2 log without the need of increasing the sample size, adjusting the voltage setting of the instrument or using nucleic acid testing formats as previously reported (Trudil 2000; Wilson et al. 2007; Noda et al. 2008).

LPS or endotoxin is present in all living Gram-negative cells; therefore, this technology can be adapted to a portable spectrophotometer that provides quantitative and qualitative results to provide an equally rapid, accurate means of detecting and enumerating any Gram-negative bacteria. The most significant advantage of LAL assay as compared to AKBA is that the LAL assay can only be applied to Gram-negative bacteria while AKBA can be used for almost all bacteria.

The only instrument needed is a portable spectrophotometer that can be used in field and phage stocks and can be kept for years without fastidious requirements. Using microplate spectrophotometer for LAL assay guarantees the ability to conduct at least 30 tests per h including the negative and positive controls. This might be highly recommended for the largest public water systems (serving millions of people) where at least 480 samples of water per month must be taken to examine water cleanliness (EPA 2006).

To design a quantitative rapid detection assay for E. coli bacteria, the amount of endotoxin liberated per bacterial cell, AlU-bac, was measured. As seen in Fig. 1 , the graph of log AIU measured in range between log 10² and log™ 10⁴ CFU was linear. Therefore, finding the generic mean AlU-bac for each incubation time was feasible which was calculated of mean AlU-bac at 10², 10³, and 10⁴ CFU. Since the target bacteria were of a known titre, the validity of using such generic mean AlU-bac was tested. The predicted number (pr no.) was used to yield the pr no. of the target bacteria in a sample. It gave close figures and the differences were almost less than Iogio0.7, from the real CFU of target bacteria. For the lake water samples, rwas shown to be consistently positive and higher than +0.84 which fairly indicated a high correlation coefficient between the pr. no. and the real bacterial number shown by the standard plating method. In addition, linear regression equation was used to predict the log CFU/well. The correlation coefficient and regression index for LAL assay at 10, 20, 30, and 40 min times were all above 0.9 indicating a very high linear relationship between the measured AIU of endotoxin and the actual log™ CFU/well. Therefore, the currently formulated phage-based LAL assay can be reproduced, as a quantitative as well as qualitative assay, easily by other researchers and other sets of experiments.

The efficiency of using a phage mixture composed of 172 highly lytic coliphages, including 22 EHEC-specific phages, was challenged in the current LAL assay. Thirty lake water samples, that proved to be E. coli contaminated, were subjected for the innovated phage-based LAL assay. Interestingly, the lowest detection limit of E. coli titre, the sensitivity, and the specificity of LAL assay for these samples, containing unknown strains of non-EHEC, were closely similar to these of the known 10 laboratory strains EHEC. This provided strong evidence for the wide and reliable coverage of the used phage master mix for the environmental E. coli bacteria. It is noteworthy to mention that the used phage master mix of designed phages was prepared on hundreds of clinical isolates of human pathogenic E .coli (EHEC and non- EHEC) as well as environmental isolates of E. coli (Jassim et al. 2010). This mixture was found to be satisfactory in yielding acceptable sensitivity and specificity results, 84/75 and 92/81 at 50-100 and 10² to 10³ CFU/well respectively, for the detection of unknown non-EHEC strains. In addition, there was no significant difference in sensitivity, specificity, and the minimal detection limit between the known EHEC isolates and the unknown non- EHEC environmental E. coli bacteria which pinpoints to the feasibility of using phage-based LAL assay in detecting both clinical and hygienic E. coli bacteria. So, the designed phage-based LAL assay is capable of detecting specifically EHEC and non-EHEC bacteria at very low titres <10² CFU, within <70 min along with adequate quantitative potential for the detected bacteria.

The present inventors have provided a novel and rapid phage-based detection test for Gram-negative bacteria, and E. coli in particular, comprising a LAL assay having a detection limit of <10² CFU at <70 min. In comparison with a known rapid detection test, AKBA, LAL was shown to have a detection limit (<10² CFU) one log lower, higher sensitivity, and higher specificity than AKBA which showed a detection limit of 10³ CFU.

Utilizing phage design method (Jassim et al. 2010) has substantially improved test sensitivity for the AKBA or ATP assay by < 2 logs more than previously reported by other researchers. This improvement in the test AKBA or ATP detection limits would subsequently improve instituting proactive measures for quality assurance, i. e., implementing HACCP programs system allow for increased detection limits as well as specific identification.

Table 1. Phage-based LAL assay values, AWJtest, AWJPosCon, and AWJNegCon for 2 sets of 10 known EHEC at titres adjusted to log₁₀ 10¹ to 10⁴ CFU/well. The positive detections typed in bold with sensitivity, specificity, and predicted number (pr. No.) of the detected E. coli.

Pyrochrome incubation period

*sn/sp: sensitivity/specificity for the positive detections † pr no.: Logio predictive CF U/well.

Table 2. Phage-based LAL assay AWJtest, AWJPosCon, and AWJNegCon values for 30 unknown E. coli bacteria in lake water samples at different titres/well. The positive detections typed in bold with sensitivity, specificity, the predicted number of detected E. coli, pr. No., and the correlation coefficient r.

^*sn/sp: sensitivity/specificity for the positive detections

† pr-no.: Logio predictive CFU/well

t {ή: The Pearson correlation coefficient between pr. no. and the real standard plating number of bacteria. (+) and (-) signs for positive and negative r

Table 3. Phage-based AKBA assay ARLUtesf, ARLUPosCon, and ARLUNegCon values for 2 sets of 10 known EHEC bacteria at titres adjusted to 2, 3, and 4 log CFU/well. The positive detections typed in bold with sensitivity, specificity of the positively detected E. coli.

^*sn/sp: sensitivity/specificity for the positive detections

References

Blanch, A.R., Galofre, B., Lucena, F., Terradillos, A., Vilanova, X. And Ribas, F. (2007) Characterization of bacterial coliform occurrences in different zones of a drinking water distribution system. J AppI Microbiol 102, 71 1 - 721 .

Blasco, R., Murphy, M.J., Sanders, M.F. and Squirrell, D.J. (1998) Specific assays for bacteria using phage mediated release of adenylate kinase. J.

AppI. Micro. 84; 661 -666.

Brown, J.M., Proum, S. and Sobsey, M.D. (2008) Escherichia coli in household drinking water and diarrheal disease risk: evidence from

Cambodia. Water Sci Technol 58, 757-763.

Chapman, P.A., Cerdan-Malo, AT., Siddons, C.A. and Harkin, M.A. (1997)

Use of a commercial enzyme immunoassay and confirmation system for detecting Escherichia coli O157 in bovine fecal samples. AppI. Environ.

Microbiol 63, 2549-2553.

Cutter, C.N., Dorsa, W.J. and Siragusa, G.R. (1996) A rapid microbial ATP bioluminescence assay for meat carcasses. Dairy Food and Environmental

16(11), 726-736.

Dey, B.P. and Lattuada, CP. (1998) Microbiology laboratory guidebook, 3rd ed., vol. 1 . U.S. Department of Agriculture, Washington, D.C

Dostalek, P. and Branyik, T. (2005). Prospects for Rapid Bioluminescent Detection Methods in the Food Industry - a Review. Czech J. Food Sci. 23; 85-92.

EPA. (2006) Basic Information about E. coli 0157:H7 in Drinking Water

(http://www.epa.gov/safewater/ contaminants/ecoli.html)

Favrin, S.J., Jassim, S.A.A. and Griffiths, M.W. (2001 ) Development and optimization of a novel immunomagnetic separation-bacteriophage assay for the detection of Salmonella enterica Serovar enteritidis in broth. AppI

Environ Microbiol 67, 217-224.

Favrin, S.J., Jassim, S.A.A. and Griffiths, M.W. (2003) Application of a novel immunomagnetic separation-bacteriophage assay for the detection of

Salmonella enteritidis and Escherichia coli O157:H7 in food. Int J Food

Microbiol 85, 63-71 .

Frampton, E.W. and Restaino, L. (1993) Methods for Escherichia coli identification in food, water and clinical samples based on beta- glucuronidase detection. J AppI Bacteriol lA, 223-233.

Jassim, S.A.A., Abdulamir, A.S., and Abu Bakar, F. 2010. Methods for bacteriophage design. WO/2010/064044

Jassim, S.A.A., Ellison, A., Denyer, S.P. and Stewart, G.S. (1990) In vivo bioluminescence: a cellular reporter for research and industry. J Biolumin

Chemilumin 5, 1 15-22.

Jassim, S.A.A., Camprubi, S., Tomas, J.M., Williams, P., Stewart, G.S.A.B. and Denyer, S.P. (1993) In vivo bioluminescence for studying bacterial adhesion and in vitro phagocytosis. In Bioluminescence and

Chemiluminescence ed. Szalay, A. A., Kricka, L.J. and Stanley, P. pp.

491 -495. New York: John Wiley and Sons,

Jassim, S.A.A., Stewart, G.S.A.B. Denyer, S.P., Park, S.F., Rostas-Mulligan,

K. and Ress, C. (1996) Methods for rapid microbial detection. US patent

5498525. Kyriakides, A.L. and Patel, P.D. (1994) Luminescence techniques for microbiological analysis of foods. In Rapid Analysis Techniques in Food Microbiology ed. Patel, P. pp. 196-231 . London: Blackie Academic & Professional.

Levin, J. and Bang, F.B. (1964) The role of endotoxin in hte extracellular coagulation of Limulus blood. Bull Johns Hopkins Hosp.1 15: 265.

Lindsay, G.K., Roslansky P.F. and Novitsky, T. (1989) Single-Step,

Chromogenic Limulus Amebocyte Lysate Assay for Endotoxin J. Clinic.

Microbiol 27, 947-951 .

Meng, J., Zhao, S., Doule, M. P., Mitchell, S. E. and Kresovich. (1996)

Polymerase chain reaction for detecting Escherichia coli O157:H7. Int J

Food Microbiol 32, 103-1 13.

Murphy, H.M., Payne, S.J. and Gagnon, G.A. (2008) Sequential UV- and chlorine-based disinfection to mitigate Escherichia coli in drinking water biofilms. Water Res 42, 2083-2092.

Noda, K., Matsuno, T., Fujii, H., Kogure, T., Urata, M., Asami, Y. and Kuroda

A. (2008) Single bacterial cell detection using a mutant luciferase.

Biotechnol Lett 30; 1051-1054.

Northcutt, J and Russell, S.M. (1996) Making HACCP happen in your plant.

Broiler Industry, July p. 24-35.

Reiprich, W.G., Lagrange, F., Plettenberg, H.K. Hoffmann, M. (2002) Rapid monitoring of superficial bacteria based on bioluminescence techniques on instant film. Biomed Tech (Berl) 47 Suppl 1 Pt 1 , 423-425.

Rhee, M.S. and Kang, D.H. (2002) Rapid and simple estimation of microbiological quality of raw milk using chromogenic Limulus amoebocyte lysate endpoint assay. J Food Prot 65, 1447-1451 .

Rokosz, A., Gorska, P., Michalkiewicz, J. and Luczak, M. (2003) [Biological activity of lipopolysaccharides from clinical Bacteroides fragilis strains isolated in Poland determined in reaction with limulus amoebocyte lysate].

Med Dosw Mikrobiol 55, 365-370.

Rossignol, D., Lynn, M., Wittek, A. and Rose, J. (2006) Elevated plasma levels of limulus amoebocyte lysate-reactive material. J Infect Dis 194,

1340-1341 .

Russell, S.M. (1995) Sanitation procedures and HACCP. Broiler Industry, October, p. 22-38.

Sachse, K. (2004) Specificity and performance of PCR detection assays for microbial pathogens. Mol Biotechnol 26, 61 -80.

Sakata, M., Fukuma, Y., Todokoro, M. and Kunitake, M. (2009) Selective assay for endotoxin using poly(epsilon-lysine)-immobilized Cellufine and Limulus amoebocyte lysate (LAL). Anal Biochem 385, 368-370.

Siekevitz, P. and Potter, V.R. (1953) The adenylate kinase of rat liver mitochondria. J Biol Chem 200, 187-196.

Siragusa, G.R., Kang, D.H. and Cutter, C.N. (2000) Monitoring the microbial contamination of beef carcass tissue with a rapid chromogenic Limulus amoebocyte lysate endpoint assay. Lett Appl Microbiol 31 , 178-183.

Squirrell, D.J. and Murphy, M.J. (1994) Adenylate kinase as a cell marker in bioluminescence assays. In Bioluminescence and Chemiluminescence; Fundamental and Applied Aspects ed. Campbell, A.K., Krika, I.J. and Stanley, P.E. pp. 486±489.Chichester: John Wiley & Sons. Stewart, G.S.A.B., Jassim, S.A.A., Denyer, S.P., Newby, P., Linley, K. and Dhir, V.K. (1998) The specific and sensitive detection of bacterial pathogens within 4 h using bacteriophage amplification. J AppI Bactenol 84, 777-783.

Trudil, D., L. Loomis, R. Pabon, J. A. K Hasan, and C. L. Trudil. (2000) Rapid ATP method for the screening and identification of bacteria in food and water samples. Biocatalysis-2000: Fundamentals & Applications, Moscow University Chemistry Bulletin 41 (6); 27-29.

Ulitzur, S. and Kuhn, J. (1989) Detection and/or identification of microorganisms in a test sample using bioluminescence or other exogenous genetically introduced marker. US patent 4,861 ,709.

Walker, A.J., Jassim, S.A., Holah, J.T., Denyer, S.P. and Stewart, G.S. (1992) Bioluminescent Listeria monocytogenes provide a rapid assay for measuring biocide efficacy. FEMS Microbiol Lett 70, 251 -5.

Wang, D. and Fiessel, W. (2008) Evaluation of media for simultaneous enumeration of total coliform and Escherichia coli in drinking water supplies by membrane filtration techniques. J Environ Sci (China) 20, 273- 277.

Wilson, S. Banin, S. and Stanley, C. (2007) Detecting ATP with nucleic acid amplification. IVDT, Nov 2007, p45. Category: Molecular Diagnostics.

Claims

1 . An assay for the detection of target bacteria, the assay comprising the steps of

(i) introducing bacteriophage specific for a target species of bacteria to be detected into a sample,

(ii) incubating the bacteriophage-sample mixture for a period of time sufficient to achieve lysis of the target bacteria,

(iii) adding Limulus amebocytes lysate labelled enzyme,

(iv) incubating the mixture produced in step (iii), and

(v) detecting the release of the label

where release of the label is indicative of the presence of target bacteria in the sample.

2. An assay according to claim 1 , in which the Limulus amebocyte lysate labelled enzyme is labelled with a chromogenic, colorimetric or other optically detectable label.

3. An assay according to claim 1 or claim 2, in which the Limulus amebocyte lysate labelled enzyme comprises a Limulus amebocytes lysate pyrochrome reagent.

4. An assay according to claim 2 or claim 3, in which the optical detection of the label is carried out by a colorimeter, a spectrophotometer or by eye.

5. An assay according to any one of the previous claims, in which the incubation in step (ii) is continued for about 30 minutes.

6. An assay according to any one of the previous claims, in which the incubation in step (ii) is conducted at or close to 37^QC.

7. An assay according to any one of the previous claims, in which the incubation in step (iv) is continued for sufficient time to allow the Limulus amebocytes lysate labelled enzyme to react with the endotoxin or other target substrate released during cell lysis.

8. An assay according to claim 7, in which the incubation is no longer than one hour.

9. An assay according to claim 7 or claim 8, in which the incubation is of between 10 and 40 minutes.

10. An assay according to any one of the previous claims, in which, the target bacteria is Gram-negative.

1 1 . An assay according to any one of the preceding claims, in which the target bacteria selected from the group comprising Enterobacteriaceae, Pseudomonas spp, Moraxella spp, Helicobacter spp, Stenotrophomonas spp, Legionella spp, Acetic acid bacteria, Hemophilus spp, Neisseria spp, Acinetobacteria spp, Vibrio spp, and Campylobacter spp or mixtures thereof.

12. An assay according to any one of the preceding claims, in which mixtures of bacteriophages are used to detect mixtures of target bacteria.

13. An assay according to any one of the preceding claims, in which a mixture of highly specific lytic coliphages is used to test for the presence of multiple strains of E.coli.

14. An assay according to any one of the previous claims, in which a quantitative and qualitative measure of bacteria is established in less than 70 minutes.

15. An assay according to any one of the previous claims further comprising the additional step of measuring the amount of endotoxin released during cell lysis by the Limulus amebocytes lysate labelled enzyme reaction to provide a quantitative detection assay for Gram- negative bacteria by measuring the yield of label detected in step (v).

16. An assay according to any one of the previous claims, in which the sample comprises foodstuffs, beverages, water including waterways, lakes or ponds, soil, medical or veterinary samples including samples from the body, medical or veterinary devices including implants and equipment, agricultural samples such as supplies or water sources, refuse, brownfield sites, or other land areas.

17. An assay according to any one of the preceding claims in which the sample is urine, stool, blood, pleural fluid, potable water, drinking water reservoir sample, riverwater, agricultural field sample, rice field sample, food or a beverage.