ATP is the universal energy molecule found in animals, plants, and microorganisms. ATP rapid hygiene monitoring tests have been employed in the food industry to ensure that adequate cleanliness is being maintained. However, because ATP is hydrolyzed to ADP and AMP by metabolic processes, by heat treatment, or under acidic or alkaline conditions, total adenylate (ATP+ADP+AMP [A3]) could be a more reliable sanitation indicator of food residues that may cause biofilm formation and allergen contamination. Therefore, a novel hygiene monitoring system to measure A3 was developed based on the luciferin-luciferase assay with the combination of two enzymes, pyruvate kinase and pyruvate phosphate dikinase, that can convert ADP into ATP and recycle AMP into ATP, respectively. The newly developed A3 assay system afforded stable bioluminescence signals and equivalent linear calibration curves between relative light units (RLU) and the amounts of ATP, ADP, and AMP, respectively. To verify the significance of the A3 method, the ratios of ATP, ADP, and AMP in various food samples were determined; large amounts of ADP and AMP were found in a variety of foods, such as meat, seafood, dairy, nuts, fruits, vegetables, and fermented foods. Sanitation monitoring of stainless steel exposed to raw meat was also examined, and the A3 method achieved a 200-RLU level, the typical benchmark value, after complete washing with detergent and rinsing. In contrast, a conventional ATP method showed less than 200 RLU after only a light cold and hot water rinse. In conclusion, the A3 assay appeared to be suitable for detection of adenylates from food residues that are not detected by the conventional ATP assay.
Proper and thorough environmental hygiene and employee sanitation education are critical to provide high quality food products that are safe for consumption. Food must not contain pathogens at levels likely to cause illness or levels of microorganisms likely to reduce shelf life. Moreover, because food allergies are also a major health issue, the unintended presence of allergens in food should be prevented. In this regard, the hazard analysis and critical control point (HACCP) system, a preventive method of ensuring food safety, has attracted a lot of attention. Recently, implementation of hazard analysis and risk-based preventive controls (HARPC) under the Food Safety Modernization Act has been made mandatory in the United States. The basic premise is to identify possible hazards in advance and set up a system of procedures, inspections, and records to minimize hazards, including contaminants such as microorganisms and allergens.
Traditionally, visual inspection and microbial counts using agar plates have been carried out to assess the cleanliness of food contact surfaces. Though immediate visual assessment is an easy method to determine whether equipment is clean, it depends on accessibility, lighting, surface, and the visual acuity of inspectors (8), and evidence suggests that it is an inadequate indicator of cleanliness (10). Microbial methods can provide either qualitative or quantitative information, but incubation takes 24 to 48 h. HACCP and HARPC require more rapid results because there is a possibility that a large amount of unsatisfactory or unsafe food may have been produced and sold before a defect is discovered (4, 10). Therefore, there has been increased interest in the use of rapid test methods in the production area rather than in the laboratory. Protein swabbing tests and lateral flow immunoassays are rapid testing options. The swab-based protein test is instrument free and is convenient to use to detect food debris and a broad range of allergens without specificities; however, previous studies have suggested that it may not be effective if residual surface contamination is low in protein (10). Lateral flow immunoassays have high sensitivities and specificities and are suitable for easy detection and determination of allergens. However, a survey by the U.S. Food and Drug Administration shows the extensive use of multiple major food allergens, including the big eight (milk, eggs, fish, crustacean shellfish, tree nuts, peanuts, wheat, and soybean) in the food industry (6), and it implies that multiple immunological tests are necessary to check for each allergen. Moreover, it is also known that fermentation processes, which reduce detectable allergen structures through associated proteolysis reactions (2, 11), can cause false negatives or inaccurately low measurement values.
ATP swabbing tests have attracted attention because they enable rapid verification of sanitation processes on site for implementation of HACCP and HARPC programs. ATP is the universal energy molecule found in all living things. The presence of ATP on surfaces indicates improper cleaning and the presence of contamination, including organic debris and bacteria. This implies a potential for the surface to harbor bacteria, provide a source of nutrients for bacterial growth, and support the formation of a biofilm. Using ATP tests can also enhance employee awareness of hygiene through immediate feedback about the effectiveness of their processes. Moreover, the test is versatile because it can screen for food residues that may contain residual allergens. Conventional ATP test systems, however, have a weakness in that ATP is degraded to ADP and AMP by heat, acids or alkalis, and enzymes, as shown in Figure 1 (1, 5, 9). Because they can only detect ATP, these conventional systems can miss residual contamination.
The aim of this study was to develop a novel hygiene monitoring system based on the detection of total adenylate (ATP+ADP+AMP [A3]). The amounts of ATP, ADP, and AMP in each food were assayed. Food residue detection sensitivities of an A3 assay system were compared with those of commercially available ATP test kits, and the assays were compared using sanitation monitoring of a stainless steel model.
MATERIALS AND METHODS
Analytical grade ATP·2Na, ADP·K, and AMP were purchased from Oriental Yeast (Tokyo, Japan).
Monitoring degradation of ATP under heat and acid or alkaline conditions
Bioluminescence reagents for the measurements of ATP, ATP+AMP, and A3 were prepared based on a previous report (13). Five micromolar ATP solutions in 50 mM phthalic acid (pH 4.0), 50 mM sodium phosphate buffer (pH 6.9), and 50 mM glycocholic acid–NaCl–NaOH (pH 11.3) were prepared. These solutions were subdivided into glass tubes with lids and heated at 80°C. After the time-dependent samples (100 μL) were diluted 100-fold, 10-μL aliquots were mixed with 100 μL of each bioluminescence reagent in ATP-free plastic tubes (Lumitube, Kikkoman Biochemifa, Tokyo, Japan), and the resulting luminescence was measured using the Lumitester C-110 (measurement time: 10 s, measurement range: 0 to 999,999 RLU; Kikkoman Biochemifa), which was equipped with a photomultiplier as a photon sensor. The measurement output was relative light units (RLU). Measurements were performed at 23 ± 1°C. The experiment was repeated three times, and the means were reported. The same procedures were performed using each solution without ATP to check backgrounds. Because the assay results showed RLU for ATP, ATP+AMP, and A3, RLU derived from ADP and AMP were calculated by the values of A3 − (ATP+AMP) and (ATP+AMP) − ATP, respectively. The abundance ratios were expressed as relative values, with RLU values of ATP before heating being normalized to 100%.
Preparation of A3 detection devices
A3 detection devices were manufactured according to the design of the LuciPac Pen (Kikkoman Biochemifa), a device for the detection of ATP+AMP. Pyruvate kinase was added to the ATP+AMP assay liquid reagent, which includes luciferase and pyruvate orthophosphate dikinase, in the same manner as previously reported (12, 13). Obtained liquid reagent was freeze-dried and pulverized, and then devices for the A3 assay were assembled. The bioluminescence of the device was then measured using the luminometer, Lumitester PD-30 (measurement time: 10 s, measurement range: 0 to 999,999 RLU; Kikkoman Biochemifa), which was equipped with a photodiode as a photon sensor.
Time course of luminescence of A3 method
One hundred microliters of 5 × 10−8 M ADP solution was pipetted onto a dry swab. After the swab stick was returned to the main tube and inserted completely, time measurements were started. The main body was shaken to mix the samples, extraction solution, and reagent thoroughly. The device was then inserted immediately into the Lumitester PD-30, and the resulting luminescence was measured by pushing the main button at 20 s. (The result was shown after 10 s.) Rereadings at 1-min intervals after starting the reaction were carried out for 5 min. Measurements were performed at 23 ± 1°C. The remaining RLU were expressed as relative values, with RLU values at 20 s after starting the reaction being normalized to 100%. The experiment was repeated five times, and the means were reported.
The inhibitory effects of disinfectants to the A3 assay
Sodium hypochlorite solution (food additives grade, Wako Pure Chemical, Osaka, Japan) was diluted with water to 500 ppm of effective chlorine concentration. Ethanol solution (76.9 to 81.4%, disinfection grade) was from Wako Pure Chemical. Osvan (10%, w/v; benzalkonium chloride solution, Takeda Pharmaceutical Company, Osaka, Japan) was diluted 100-fold with water to prepare 0.1% benzalkonium chloride according to manufacturer's guidance. Ten microliters of these disinfectant solutions or water and 10-μL aliquots of 5 × 10−6 M adenylate solutions were pipetted onto the moistened swabs with 80 μL of water. After the swab stick was returned to the main tube and inserted completely, the main body was shaken to mix the samples, extraction solution, and reagent thoroughly. The device was then inserted immediately into the Lumitester PD-30, and the resulting luminescence was measured. All measurements were made at 23 ± 1°C, and data were recorded electronically. The RLU were expressed as relative values, with RLU values without disinfectants being normalized to 100%. The experiment was repeated five times for each aliquot, and the means were reported.
Preparations for calibration curves for ATP, ADP, and AMP solutions
To achieve an objective evaluation, the preparations for standard curves of the A3 system were carried out by an independent third-party testing laboratory, Food Safety Net Services (San Antonio, TX). ATP, ADP, and AMP solutions were prepared in nuclease-free water at 10−4 M concentrations according to contents information in the supplier's certificates of analyses (ATP·2Na purity: 99.8%; ATP contents: 86.1%; contaminants of ADP and AMP: 0.333 and 0.034%. ADP·K purity: 99.9%; ADP contents: 84.0%; contaminants of ATP: 0.008%. AMP purity: 97.3%; AMP contents: 92.5%; contaminants of ADP: 0.02%). These solutions were serially diluted 10-fold with water to obtain solutions with lower concentrations of ATP, ADP, and AMP. Dry swabs were premoistened with 100 μL of nuclease-free water, and then 10-μL aliquots were pipetted onto swabs. The measurements were performed as described above. The experiment was repeated 10 times for each aliquot, and the means were reported.
As the references, standard curves for the conventional ATP method and ATP+AMP measurement were originally carried out using the LuciPac II–Lumitester C-110 and LuciPac Pen–Lumitester PD-30, respectively. All measurements were performed as the study of the A3 systems.
Measurement of the ratios of ATP, ADP, and AMP in foods and beverages
All foods and beverages used in this study were commercially available. Solid foods (10 g) were homogenized with Milli-Q water (90 mL; Millipore, Bedford, MA) with a blender, resulting in 10-fold diluted samples. Serial 10-fold dilutions were carried out for all samples, including liquid foods and beverages. The amounts of ATP, ATP+AMP, and A3 were assayed using the LuciPac II–Lumitester C-110, LuciPac Pen–Lumitester PD-30, and established A3 monitoring system, respectively. One hundred microliters of each suspended sample was pipetted onto a dry swab, and the assays were performed as described above. The measurements were repeated three times, and the means were obtained. The ratios of ADP, AMP, and A3 to ATP were expressed as relative values, with the values of ATP being normalized to 1.
Evaluation of detection performances of A3 and ATP monitoring test kits
To achieve an objective evaluation, this study was also carried out by the independent third-party testing laboratory, Food Safety Net Services.
Three popular commercially available ATP detection systems (ATP-1, ATP-1H, and ATP-2) were operated in accordance with manufacturers' instructions. ATP-1 and ATP-1H were produced by the same manufacturer, and ATP-1H was the combination of a high-sensitivity device and a luminometer. Dry swabs for A3 detection were premoistened with 100 μL of nuclease-free water before use, and measurement procedures were performed in the same manner as for the calibration curves study.
The three devices were compared for their ability to detect ATP, ADP, and AMP solutions (10−7 M) and different types of food matrices. The measurements of ATP, ADP, and AMP solutions were repeated 10 times. Blank sample measurements consisting of nuclease-free water were repeated 20 times, and the means were reported.
The target foods for this study were ground beef, raw sausage, ready-to-eat turkey, raw tuna, raw shrimp, raw whole egg, beer, peanuts, and cantaloupe. Portions (10 g) of each food matrix were mixed with 90 mL of sterile distilled water and were homogenized for 2 min in a Masticator Panoramic 400 mL (IUL Instruments, Barcelona, Spain) to generate the 10−1 dilution. Aliquots of homogenate were measured as suspensions. The beer sample was used without dilutions. Aliquots of 10 μL of each sample were pipetted onto the appropriate swabs for each type of detection device and food matrix (n = 5 for each condition).
Sanitation monitoring of stainless steel using ATP, ATP+AMP, and A3 methods
Six stainless steel coupons were exposed to three commercially available raw meat samples (chicken, beef, and pork) for two test repeats. These plates were washed three times. The first wash was a rinse with tap water (20°C) for 30 s; the second was with hot tap water (55°C) for 30 s. In the last rinse, the plate was cleaned using a sponge with dishwashing detergent and was rinsed with tap water (20°C). After each wash, square surfaces (10 by 10 cm [4 by 4 in.]) were swiped with three swabs premoistened with running tap water, and then ATP, ATP+AMP, and A3 assays were carried out as described above. The means for each raw meat were reported.
RESULTS AND DISCUSSION
Degradation of ATP under heat and acidic or alkaline conditions
First, chemical decomposition of ATP was tracked with a luciferase-based bioluminescence assay. Figure 2 demonstrates that ATP was degraded at 80°C under acidic, neutral, and alkaline conditions, as shown with the black lines. Under acidic conditions, ATP was likely converted through ADP (Fig. 2, blue) to AMP (Fig. 2, red). Under alkaline conditions, the degradation of ATP and accumulation of ADP were confirmed. In either case, the amounts of A3 (Fig. 2, green) were clearly maintained. Therefore, these data indicate that A3 detection is a more sensitive and reliable indicator of sanitation. Previous studies using gas chromatography–mass spectrometry also showed that ATP in water was readily degraded to ADP and AMP. After 30 to 120 s at 150°C, ADP was accumulated and a trace amount of AMP was generated (5), similar to results seen in this study. In food production facilities, heating for cooking and sanitation are indispensable. These data indicate that conventional ATP hygiene monitoring tests may show a passing test for sanitation effectiveness, even if heated food debris, which may contain low amounts of ATP, remains on the surface. Therefore, hygiene monitoring systems based on A3 detection look reasonable for the control of surface cleanliness in food production facilities.
The principle of the A3 detection method
The principles of detection of ATP, ATP+AMP, and A3 are shown in Figure 3. Firefly luciferase can produce light in the presence of ATP, luciferin, oxygen, and Mg2+ (Fig. 3, black) (12, 13). When the ATP is increased, the light intensity is increased, too. Consequently, ATP can be quantified by the detection of the light using a luminometer, known as the ATP method.
In the reaction of luciferase, ATP is degraded into AMP; this can inhibit the luciferase reaction, resulting in a decrease of light production. To detect AMP simultaneously and maintain the light production, ATP was regenerated from AMP using pyruvate orthophosphate dikinase reactions in the presence of phosphoenol pyruvate, inorganic pyrophosphate (PPi), and Mg2+ (Fig. 3, red) (12, 13). This is the basis of the ATP+AMP monitoring system.
Moreover, an additional enzymatic reaction was combined with ATP+AMP detection for the simultaneous detection of ADP. ADP was successively converted to ATP by a pyruvate kinase (PK) reaction in the presence of phosphoenol pyruvate, Mg2+, and K+ (Fig. 3, blue) (13). Consequently, a novel A3 detection device was successfully developed (Fig. 4, right). Its bioluminescence can also be measured using the luminometer (Fig. 4, left).
Time courses of bioluminescence signals were measured in the A3 assay. The A3 method demonstrated a stable reaction using 10−12 mol adenylates, and 90, 88, and 88% signals for ATP, ADP, and AMP, respectively, were maintained even after 5 min. This long-lived bioluminescence can be explained by the principle of the regeneration system of ATP from AMP (Fig. 3), which offers stable measurement regardless of mixing time.
The inhibitory effects of disinfectants to the A3 assay
Disinfectants are used in cleaning to kill microorganisms, and these chemicals may be left on the surface. Because residues from certain sanitizing compounds are known to inhibit conventional ATP tests, the ability to detect A3 in the presence of different sanitizers was assessed. The target sanitizer compounds that were used for this study were sodium hypochlorite (500 ppm of effective chlorine concentration), ethanol (ca. 80%), and benzalkonium chloride (0.1%). The RLU measurement values in Table 1 indicate that, although these chemicals seem to affect the A3 assays to some extent, they were not fatal when 10% volumes of disinfectants were added to the moistened swabs. The retained measurement values of ATP, ADP, and AMP detections were almost the same, which shows that ATP measurement with luciferase has a dominant influence on the inhibitions by these disinfectants, and the conversion of ADP to ATP with PK and the recycling of AMP to ATP with pyruvate orthophosphate dikinase (Fig. 3) were not affected dramatically.
Calibration curves of A3 method for ATP, ADP, and AMP solutions
Calibration curves of the A3 method using reagent-grade ATP, ADP, and AMP standards are shown in Figure 5A. The A3 assay afforded equivalent linear calibration curves between RLU and the amounts of ATP, ADP, and AMP in logarithmic scales.
To study the ratios of the three adenylates in foods, the standard curves of ATP and ATP+AMP methods were compared as references (Fig. 5B and 5C). They also provided linear relationships between ATP concentration and RLU. The same amount of ATP gave almost the same amount of RLU in ATP, ATP+AMP, and A3 methods. These data demonstrated that they were applicable to estimate ratios of three adenylates in food materials and processed foods by comparison of their RLU values (Table 2).
ATP, ATP+AMP, and A3 measurement in food samples
Recently, ATP tests have been attracting considerable attention as tools to control food debris. There are two main reasons to detect food residues for sanitation control: food residues may harbor bacteria and promote the formation of biofilms, which can shield bacteria from the action of sanitizers (3), and they can also contain unwanted allergens (8). Accordingly, detection sensitivity of ATP, ATP+AMP, and A3 assays for food residues were compared.
Table 2 shows the measurement RLU values of ATP, ATP+AMP, and A3, and the ratios of ADP, AMP, and A3 to ATP in food, such as meat, seafood, dairy, nuts, grains, bakery, fruits, and vegetables. As can be seen, very little ATP was present in meat, egg, and seafood products. However, the amount of A3 was significant, and nearly all detections were more than two orders of magnitude greater in sensitivity than ATP alone. The data demonstrated that raw meat and fish contain mainly ADP, regardless of species. On the other hand, raw whole egg, shellfish (shrimp, oyster, scallop, and abalone), mollusk (squid), and salmon eggs contained a large amount of AMP. AMP was also abundant in processed meat and seafood, such as sausage, bacon, beef jerky, canned fish, and dried fish.
For dairy foods, such as milk and yogurt, RLU of the A3 method were approximately three to five times higher than those of the ATP method. More highly processed foods, such as butter, ice cream, cheese, and coffee creamer, showed a greater differential between the response of the A3 and the conventional ATP method. Notably, Camembert cheese, which has mold on its surface, contained a relatively high level of AMP. Thus, the A3 method appears to be more sensitive in dairy applications.
Many plant seeds, such as nuts and seeds, are widely known as allergens. The A3 method also showed remarkably high sensitivity for peanut, tree nuts (almond, cashew, hazelnut, walnut, and pistachio), soybean, wheat, barley, sesame, mustard, and processed foods made from nuts and grains (peanut butter, soymilk, chocolate, pasta, buckwheat noodles, and beer), although the ATP method showed very low responses. These adenylates mainly consisted of AMP. Interestingly, in wheat flour, the sensitivity of the A3 assay and the ratio of A3 increased after boiling, probably owing to improvement of extraction efficiency, chemical degradation of ATP, and inactivation of AMP deaminase. In general, because wheat is usually processed by heating in food manufacturing, this indicates that the A3 assay may be an advantageous method for sanitation monitoring in heat-processed foods that contain wheat. Baked goods, such as bread and biscuits that were made from wheat and also eggs and milk, were also good targets of the A3 method.
Many varieties of fruits and vegetables, including mushrooms, were also investigated. Although distributions of ADP and AMP likely depend on the species, parts, and processing, the results revealed that the A3 method is a reliable means of detection.
Comparison of detection performances of A3 and ATP monitoring test kits
The detection performance of the A3 and three commercially available ATP monitoring systems (ATP-1, ATP-1H, and ATP-2) were compared. This study was carried out by an independent third-party laboratory.
First, the detection of ATP, ADP, and AMP solutions (10−7 M, 10 μL, 10−12 mol per assay) were evaluated (Table 3). The blank means of A3, ATP-1, ATP-1H, and ATP-2 were 4.1, 0, 0.05, and 14.0, respectively. Not surprisingly, ATP could be detected by all systems. Regarding ADP and AMP, the A3 monitoring system could also detect them as shown in Figure 5A and Table 3, but other conventional ATP monitoring systems could not detect ADP and AMP at all because they do not use ATP generation from ADP and AMP.
Next, the three different types of assays were assessed for their abilities to detect the adenylate in different types of food matrices (ground beef, raw sausage, ready-to-eat turkey, raw tuna, raw shrimp, raw whole egg, beer, peanut, and cantaloupe; Fig. 6A through 6J). As demonstrated in the data of Table 2 using adenylate monitoring devices made by the same manufacturer, the A3 method showed much higher signal intensity than the other three ATP test kits in raw and processed meat and seafood, fermented foods, nuts, and fruits. In addition to the ability to detect ADP and AMP, efficient extraction of intracellular adenylates from food samples is another important factor to achieve good detection. The A3 system includes a larger amount of extraction solution in a compartment and uses mutant luciferase with increased resistance for the detergent (7). Because the A3 method also meets this requirement, it could achieve highly sensitive detection of food samples.
Sanitation monitoring of stainless steel using ATP, ATP+AMP, and A3 methods
Although the A3 method is very sensitive, there was still some doubt about whether the signal intensity of A3 decreased after proper washing. To clarify this question, sanitation monitoring with stainless steel was examined (Fig. 7A). Stainless steel coupons were exposed to raw chicken, beef, and pork, which are generally recognized as foods with significant risks of pathogen contamination and which contain high ratios of ADP, as shown in Table 2. They were washed step-by-step, and square areas were swabbed with three swabs premoistened with running tap water (same as used in the washing processes). RLU were assayed with ATP, ATP+AMP, and A3 methods using these swabs (Fig. 7B). Using the ATP method (shown by red lines), the first and second rinse steps with cold and hot water, respectively, produced less than 200 RLU. This is the typical benchmark value indicating a clean surface. This generated a false negative, which indicated that the surface was clean enough, although the washing was not completed. This result was consistent with the data that showed that the conventional ATP method hardly responded to raw meat (Table 2). The ATP+AMP method showed approximately 200-RLU even after a hot water rinse because AMP contents are not present in high amounts in raw meats. However, using the A3 method (shown by green lines), the residue remaining on the surface was recovered by swabbing, and the 200‐RLU level was not achieved until after complete washing with detergent and rinsing. These results showed that the A3 system is superior for monitoring surface cleanliness.
In summary, the new hygiene monitoring assay was successfully developed based on the luciferase assay with the combination of two additional enzymatic reactions. In this method, all three adenylate molecules, A3, can be detected simultaneously. A wide variety of food contains significantly large amounts of ADP and/or AMP compared with ATP. Accordingly, the A3 method is far more sensitive for detection of adenylate from food residue than conventional ATP assays. This new technology delivers more reliable results for determination of hygiene and sanitation levels in the food industry.
Appreciation is expressed for the technical support provided by Food Safety Net Services in the United States for their work in preparation of calibration curves of ATP, ADP, and AMP for the A3 methods and for comparison of the performance of ATP sanitation monitoring test kits and the A3 method for detection of food and adenylate solutions. Appreciation is also expressed to Oriental Yeast Co., Ltd. for their great cooperation in choosing adenylate reagents with high purities and provide purity information for this research.