Food and Fertilizer Technology Center - publications

Mar. 04, 2016

Promising New Technologies for Assuring Food Safety

Promising New Technologies for Assuring Food Safety


Hitoshi Nagashima

National Food Research Institute,

National Agriculture and Food Research Organization

Tsukuba, Ibaraki, Japan




Four promising new technologies for assuring food safety are introduced. 1) Fluorescence fingerprint technique was applied to detect food hazards. It is nondestructive and quick measurement. As examples of food hazards, detection of mycotoxins in wheat flour and prediction of aerobic bacteria population on beef surface are shown. 2) High electric field alternating current (HEF-AC) technique provided effective inactivation of B. subtilis spores in orange juice. HEF-AC technology can be scaled up and can be applied to consecutive processing, making it a suitable inactivation technology for practical use. Besides, HEF-AC retained more fragrance and nourishment components than ultra-high temperature. 3) The development of the multiplex PCR detection kit for four pathogenic bacteria in food samples is shown. When the kit was used to detect the pathogenic bacteria, one cell per 25 g of sample was detected within 24 h. There was excellent agreement between the multiplex PCR assay and the conventional method. The detection kit will be valuable as a screening method for foods contaminated with these pathogens. 4) Liquid chromatography Orbitrap–mass spectrometry is a powerful technique that has very high sensitivity and selectivity. Its application enabled detection of several mycotoxin derivatives without chemical standards.


Keywords: fluorescence fingerprint, high electric field alternating current, multiplex PCR detection, liquid chromatography orbitrap–mass spectrometry



Food safety is a scientific discipline which covers handling, transportation, storage and retailing of food in ways that preventS foodborne illnesses. This includes numerous routines that should be followed to avoid potential health hazards. Potential risks of food safety in fresh fruit and vegetable products are contamination of harmful biological or chemical agents during handling, transportation, storage and retailing. In order to share experiences on research and development to ensure safety of fresh fruit and vegetables, I introduce here four promising new technologies for assuring food safety.

Fluorescence fingerprint

Fluorescence is well known technique in analytical chemistry. Its measurement is made by a pair of stimulus and response, which are excitation light and fluorescence spectra.  However, if we can have more information from the sample, we could have more precise or more identification at the same time. That is the reason we introduced fluorescence fingerprint (FF), which, in other words, is also termed as excitation emission matrix.  Fig.1 shows the principle of data acquisition for fluorescence fingerprint.  Scanting of excitation wavelength produces a lot of fluorescence spectra. They can be a three dimensional volume data consisting of an excitation wavelength axis, an emission wavelength axis and a fluorescence intensity axis.  This is FF.  The top view of the graph (contour map) shows an original pattern which reflects samples and is called optical property.  Conventional fluorescence analysis usually uses only one peak.  However, not only peak but also other points could have some related information to the objective.  In addition, as all data is recorded in digital value, we can extend the analyzing area to the whole area using a powerful computing PC (Shibata et al. 2011; Kokawa et al. 2011; 2012; Oto et al. 2012).


Fig. 1.  Fluorescence fingerprint


The quantification models were developed using partial least squares (PLS) regression with leave-one-out cross validation to the FF data of the calibration samples. The performance of PLS models depends on the number of latent variables (LVs) used. The optimum number of LVs was determined by minimizing the root-mean-square error of the prediction of cross-validation. The calibration model was applied to the validation dataset to evaluate the accuracy of the model. The fitting of the calibration model to the calibration and validation datasets was finally evaluated by the coefficient of determination (R2), standard error of calibration (SEC), and standard error of prediction (SEP) .

In Japan, the most common contaminated mycotoxin in wheat is deoxynivalenol (DON). Concomitant contamination of other mycotoxins such as nivalenol (NIV) and zearalenone (ZEA) are also found. The wheat samples were artificially contaminated with Fusarium graminearum in the field. Four different levels (Low, medium, medium-high, and high) of contaminated wheat were harvested. They were ground into flour by the milling machine. To predict quantitative contamination level, PLS regression was applied. Actual contamination level was measured by HPLC-UV. Fig. 2 is the prediction of DON concentration in contaminated wheat flour. Both calibration and validation datasets show significant correlations between actual values and predicted values. However, degree of contamination is different from DON.  FF also reflects on these contaminations.  There could be created the model to predict for NIV and ZEA from the identical FF. The results of NIV and ZEA prediction have good correlations with chemical analyses, indicating that the FF can predict DON, NIV, and ZEA concentrations at the same time.


Fig. 2.  Prediction of DON


Lean beef pieces were purchased from a local meat store and they were cut into 45 x 45 x 8 mm pieces. Samples were stored aerobically by putting them into sterilized plastic Petri dishes with lids. Each lot of beef samples were stored in an incubator at 15 °C and analyzed after 0, 12, 24, 36 and 48 h of storage. Samples were placed between quartz plate and an acrylic plate (Fig. 3a) and mounted in the sample holder in the spectrophotometer. A Fluorescence spectrophotometer mounted with a front-surface sample holder was used to measure FF. FFs were measured in the range of 200 ~ 900 nm for both excitation and emission wavelengths. Four locations (Fig. 3b), cross marks, No.1 - 4) were measured for one sample at room temperature. A total of 240 FFs (4 lots x 5 different time of storage x 3 samples x 4 positions) were collected. After FF measurements, 40 mm squared areas on both quartz plate and beef sample were wiped with a sterile swab (Fig. 3b, shaded area). To ensure adequate sampling, the sample was swabbed in a horizontal pattern and again in a vertical pattern, while being rotated between the index finger and thumb in a back and forth motion. Serial dilutions of the swab sample were prepared with the phosphate buffer solution in which the swab was immersed, then aerobic plate count (APC [CFU/cm2], CFU: Colony forming unit) were determined by incubating 1 ml of appropriate dilution on PetrifilmTM Aerobic count plates for 48 hr at 35 °C . A total of 60 APCs (4 lots x 5 different time of storage x 3 samples) were determined through the entire experiment.  

PLS regression was applied to FF to develop a model for the prediction of aerobic plate count (APC). In this case for the beef meat, prediction model for the aerobic plate count was made with seven latent variables (LV), which gave best result with highest correlation and lowest SEC.  From the result for validation set, good correlation (R2 = 0.819) and small SEP (SEP = 0.752 log [CFU/cm2]) were obtained and the accuracy of the model was verified.


Fig. 3.    Sample preparation for FF and APC measurement (Cross mark No. 1-4: FF measurement positions, Shaded area: swab both surfaces of meat and quartz plate for APC measurement)


High electric field alternating current

Heating has been generally used for inactivating micro-organisms in foods, but heat treatment of foods also destroys delicate fragrance components and useful functionality components. Therefore, a high electric field pulse, a high-intensity light pulse and radioactive rays have been researched and developed both domestically and abroad as non-thermal inactivation methods, but these methods are expensive and their use is limited to food industry applications that demand large-scale processing.

Internal heating caused by an electric current has been used for 100 years and can be divided into two types, microwave heating and ohmic heating. Microwave heating technology has spread from industrial use to home use with products employing electromagnetic energy at a frequency of 2.45 GHz for heating food. The ohmic heating method is older than microwave heating and was reportedly used to inactivate micro-organisms in milk in 1920. However, ohmic heating using a high frequency of around 20 kHz has become a useful technology in the food industry and has been used to process fish cake since 1990 because of the increased stability and increased energy efficiency. It had been believed for a long time that micro-organisms in food were inactivated by the electrical effects of ohmic heating. However, ohmic heating did not induce electrical effects for inactivation, and Imai et al. reported on the characteristics of the breakdown of the cell membrane when an electric field was used in the ohmic heating of vegetables, where the voltage in a cell was close to 1 V.

 A high electric field pulse, a non-thermal inactivation technology, inactivates micro-organisms in foods using high-voltage pulse sterilization with a very narrow pulse width (less than 10 μs) and high electric field strength (more than 10 kV/cm). 

A potential difference is induced between the two ends of the cell membrane when a high-strength electric field is used on a cell for sterilization. A hole subsequently opens in a local fragile site of the membrane by electricity perforation through a mechanism called electroporation. Hulsheger et al. reduced Escherichia coli two orders of magnitude when they applied 30 pulses of 30 μs width in a 12 kV/cm electric field. Qin et al. succeeded in reducing Escherichia coli six orders of magnitude by applying 60 pulses with 3 μs width in a 40 kV/cm field. Pothakamury et al. applied 50 pulses at 16 kV/cm to Staphylococcus aureus in a food model of milk and reduced the bacteria by four orders of magnitude. Electroporation is known to be generated on a cell membrane when an electric potential exceeding 1 V per cell is applied. Uemura et al. developed a high electric field alternating current (HEF-AC) technology that combined ohmic heating and a high electric field. HEF-AC was originally designed to inactivate Escherichia coli in liquid foods by Uemura and Isobe (Uemura and Isobe, 2002). The inactivation was caused by a combination of electric field effect and Ohmic heating effect.  Geveke et al. applied a 20 kHz, 18 kV/cm electric field called a radio frequency electric field to E. coli in apple juice at a moderately low temperature of 50°C, reducing the E. coli to 3 log by the high electric field effect. High-voltage pulses were not able to inactivate spores (Geveke et al., 2006). Uemura et al. used a HEF-AC with an electric field of 10 kV/cm on Bacillus subtilis spores that were added to orange juice and reduced the number of bacteria by four orders of magnitude by heating the electrode exit to 120°C (Uemura and Isobe, 2003).  Inoue et al. used a HEF-AC on various microorganisms, including the highly heat-resistant spores that were added to a model liquid, and reduced the bacteria by over three orders of magnitude (Inoue et al., 2007).  With HEF-AC, rapid heating at temperatures higher than 100°C was required to inactivate B. subtilis spores in a short time.  In this study, we inactivate B. subtilis spores in a orange juice by a practical scale HEF-AC and compare the quality change of HEF-AC treated orange juice with a ultra high temperature (UHT) treated one.

Fig. 4. outlines the HEF-AC setup.  Raw orange juice in a tank is fed at a constant flow rate of 100 L/hour.  The internal pressure in the pipe between the feeder pump and the relief valve is controlled at 0.5 MPa by the pressure of the valve.  The AC power supply had a 2,000 V maximum output voltage, 50 kW maximum output power and a 20 kHz square-wave AC.  An electric treatment unit was constructed with a parallel plate electrode made of titanium (6.0 mm in width and 32 mm in length, with 4.0 mm between electrodes) and a surrounding insulator made of Teflon (Fig. 5).

Fig. 4. HEF-AC setup


Fig. 5.  Longitudinal section of electrode unit


Inactivation of B.subtilis spores

 An electric field of 2.8 kV/cm to 3.0 kV/cm was applied to 106 cfu/mL B. subtilis spores in the orange juice (Fig. 5); the outlet temperature and the sterilization effect are presented in Fig. 6. The sterilization effect increased when the electric field strength applied by the HEF-AC increased, and the electrode exit temperature rose from 110°C to 120°C.  The sterilization effect increased with increasing outlet temperature, and the spores were reduced by four orders of magnitude at 120°C.



Fig. 6.   Viability loss of B. subtilis spores in orange juice using HEF-AC at different outlet temperature

            (n: Viable counts at indicated temperature, N0: Initial viable counts)


Quality change of orange juice

The comparison of the content of linalool and limonene was documented. They are fragrance components of orange juice after the HEF-AC and UHT treatments.  These results demonstrate that 24% more linalool and 15% more limonene remained in the orange juice after HEF-AC treatment than after UHT treatment. The results the effects on β-carotene, hesperidin and L-ascorbic acid content in orange juice after HEF-AC treatment and UHT treatment indicate that 25% more β-carotene, 18% more hesperidin and 8% more L-ascorbic acid remained in the orange juice after HEF-AC treatment than after UHT treatment.

The results revealed that HEF-AC treatment clearly retained more functional components of orange juice compared to conventional UHT treatment. Most likely, this was due to the fact that the holding time, in case of HEF-AC was almost ten times shorter than in the case of UHT treatment, so that HEF-AC preserved these functional components.

Multiplex PCR detection of pathogenic bacteria

Foodborne illnesses caused by Salmonella spp., Listeria monocytogenes, or Escherichia coli O157:H7 are a major public health concerns worldwide.  There are approximately 1.4 million cases of various illnesses annually, resulting in 1,000 deaths.  To prevent these outbreaks, the ability to rapidly detect these pathogens in food is critical.

Reliable detection techniques are a prerequisite for the detection and identification of these pathogenic bacteria in foods, food sources, and food processing plants. Because the conventional culture method for detecting pathogens (Fig. 7., Left side) is time consuming, results are frequently not available until the food has been either released to the market or consumed, thus increasing the risk of transmission of pathogens. Pathogens are often present in very low numbers against a background of indigenous microflora, rendering the recovery of target organisms difficult. Rapid and sensitive assays with high specificity are required for the detection of pathogenic bacteria in foods and other types of samples. The polymerase chain reaction (PCR) is a biochemical technology in molecular biology used to amplify a single copy or a few copies of a piece of DNA across several orders of magnitude, generating thousands to millions of copies of a particular DNA sequence. PCR-based methods have the potential for the rapid and sensitive detection of foodborne pathogens. Because PCR can target unique genetic sequences, such as the virulence genes of microorganisms, it has the advantage of being an extremely specific assay.


Fig. 7. Schematic representation of detection procedure of conventional culture method and multiplex PCR method


The multiplex PCR method is capable of determining the presence of Salmonella spp., L. monocytogenes, and E. coli O157:H7 directly from enrichment cultures by targeting the specific DNA sequences of each pathogen (Kawasaki et al., 2005).  This multiplex PCR method was used to detect pathogens in spiked pork samples, and the detection sensitivity for each pathogen was 1 CFU per 25 g food sample after enrichment for 24 h. Moreover, there was excellent agreement between the results of the multiplex PCR and the conventional culture method in naturally contaminated meat samples.

The schematic representation of the conventional culture method and the multiplex PCR detection method is shown in Fig. 7. This multiplex PCR method is simple to use, and the results are typically available within 24 h compared to at least four days for the conventional culture method.  The multiplex PCR detection protocol consists of three steps: 1) pre-enrichment culturing, 2) simple DNA extraction, and 3) multiplex PCR detection. Firstly, an enrichment medium allows the simultaneous growth of Salmonella spp., L. monocytogenes, and E. coli O157:H7 for subsequent detection of each pathogen using a multiplex PCR assay with similar sensitivity. Secondly, a simple DNA extraction method ensures the high sensitivity of the multiplex PCR assay.  Finally, the multiplex PCR component concentrations have been optimized for the specific detection of Salmonella spp., L. monocytogenes, and E. coli O157:H7. The results of limit of individual target strains for multiplex PCR shows that the correct PCR product was clearly detected in each of the target genomic DNA samples estimated to contain 1 cell/PCR (103 CFU/ml of culture) (Kawasaki et al. 2005).

This multiplex PCR detection method has high sensitivity, since one cell per PCR reaction tube was detectable. Therefore, the multiplex PCR is a useful method for the rapid screening contaminated food for Salmonella spp., L. monocytogenes, and E. coli O157:H7. There have been many reports on pathogen detection using PCR methods for various foods, such as chicken, milk, ground beef, etc. (Thomas et al., 1991; Croci et al., 2004). However, many of these reports described detection from pure cultures or from a specific food matrix. There have been few reports describing sample treatments conducted prior to PCR to remove PCR inhibitors from a variety of food matrices followed by the detection of pathogenic bacteria.

To evaluate the practical use of the multiplex PCR method for detecting the three pathogens in foods, the conventional culture method was compared to the PCR assay using 75 commercial food samples (Table 1). For E. coli O157:H7, one sample (pork intestine) was found positive using the multiplex PCR method, but this E. coli isolate was confirmed to be E. coli O55 by the conventional culture method and serological testing. Of the 75 food samples, 13 were found positive for Salmonella spp. by the multiplex PCR method, but only nine samples were positive by the conventional culture method. For L. monocytogenes, 15 samples were positive by the multiplex PCR assay compared to 14 samples by the conventional culture method.


Table 1. Results obtained with the multiplex PCR and conventional culture methods from the retail food samples


The multiplex PCR assay was also performed on spiked frozen food samples. The detection frequency of the pathogens from 28 samples stored at -20 0C for periods of two weeks and two months was documented.  The detection rate for each pathogen using multiplex PCR was higher than that of the conventional culture method in all the post-storage frozen samples. The detection rate decreased using both methods on samples stored for two months at -20 0C compared to those stored for two weeks.  This was evident particularly for Salmonella Enteritidis since detection by the culture method declined considerably after frozen storage.  The detection rate for each pathogen by multiplex PCR was greater than 75% for all food samples after frozen storage for two months (Kawasaki et al., 2009).

Detection of mycotoxin derivatives by high-resolution LC-Orbitrap MS

Liquid chromatography–mass spectrometry (LC-MS) is an analytical chemistry technique that combines the physical separation capabilities of liquid chromatography with the mass analysis capabilities of mass spectrometry. LC-MS is a powerful technique that has very high sensitivity and selectivity and so is useful in many applications. Its application is oriented towards the separation, general detection and potential identification of chemicals of particular masses in the presence of other chemicals (i.e., in complex mixtures), e.g., artificial chemicals from natural-products extracts, and pure substances from mixtures of chemical intermediates. The Orbitrap (Fig. 8), a new type of mass analyzer, has drawn attention due to its analytical performance in terms of resolution and mass accuracy. Here I show the case of mycotoxin derivatives as an example (Nakagawa et al. 2011; 2012; 2013).


Fig. 8. High resolution LC-Orbitrap MS


Fusarium fungi are known as plant pathogens infecting cereals, and some of them produce mycotoxins such as trichothecenes and zearalenone. For these mycotoxins, several glucoside derivatives are reported and known as “masked mycotoxins”.  Since the hydrolysis of masked mycotoxins releasing their aglycons was reported, these glucosides are considered to present an additional potential risk for mycotoxins. In addition, trichothecenes are sesquiterpenoid mycotoxins composed of some groups, we expected that glucoside derivatives derived from various trichothecenes would be found in nature. Screening of new masked mycotoxins was performed by means of LC-Orbitrap MS. With the accurate mass and high-resolution (AM/HR) measurement, the detection of compounds whose chemical standards are not available is possible. The identification was carried out on the basis of characteristic ions and fragmentation patterns observed with LC-Orbitrap MS. Masked mycotoxins derived from type B trichothecenes (fusarenon-X and NIV) in wheat grain that was artificially infected with Fusarium fungi were detected.  Masked mycotoxins derived from type A trichothecenes (T-2 toxin and HT-2 toxin) in commercially available corn powder reference material were also detected. Corn powder was naturally contaminated with type A trichothecenes.  Although the absolute structures were not clarified except for the T-2 toxin-glucoside, 3-OH glucosylation seems to be the most probable based on the fragment profiles and concomitant detection of deoxynivalenol-3-glucoside (DON3Glc) in the identical samples. The amount of these masked mycotoxins was estimated according to an extrapolation based on the molar ratio DON3Glc/DON in each sample.  These findings indicate that not only type B, but also type A trichothecenes are naturally glucosylated in plants such as wheat and corn.  Although the existence of these masked mycotoxins is not currently included in the risk evaluation, more analytical and toxicological studies are needed to determine their prevalence in foods and the relevance for human health.



Croci, L. et al. 2004. Comparison of PCR, electrochemical enzyme-linked immunosorbent assays, and the standard culture method for detecting Salmonella in meat products. Applied and Environmental Microbiolology 70: 1393.

Fujita, K. et al. 2010. Detection of deoxynivalenol using fluorescence excitation–emission matrix. Food and Bioprocess Technology 3:922.

Geveke, D.J. et al. 2006. Nonthermal inactivation of E. coli in fruit juices using radio frequency electric fields. Pp121-139 in: Advances in microbial food safety, eds. Juneja, V.K. et al., American Chemical Society, Washington DC, USA

Inoue, T. et al. 2007. Inactivation of various microorganism spores by high electric field AC treatment in saline.  Nippon syokuhin kougakkaishi (Journal of Food Engineering), 8, 3:123. [In Japanese with English Summary].

Kawasaki, S. et al. 2005. Multiplex PCR for simultaneous detection of Salmonella spp., Listeria monocytogenes, and Escherichia coli O157:H7 in meat samples. Journal of Food Protection 68:551.

Kawasaki, S. et al. 2009. Evaluation of a multiplex PCR system for simultaneous detection of Salmonella spp., Listeria monocytogenes, and Escherichia coli O157:H7 in food subjected to freezing. Foodborne Pathogens and Disease 6:81.

Kokawa, M. et al. 2011. Visualization of gluten and starch distributions in dough by fluorescence fingerprint imaging Bioscience, Biotechnology and Biochemistry 75:2112.

Kokawa, M. et al. 2011. Quantification of the distributions of gluten, starch and air bubbles in dough at different mixing stages by fluorescence fingerprint imaging. Journal of Cereal Science 55:15.

Kokawa, M. et al. 2012. Development of a quantitative visualization technique for gluten in dough using fluorescence fingerprint imaging Food and Bioprocess Technology 5, doi: 10.1007/s11947-012-0982-7

Lakowicz, J.R. 1990. Principles of Fluorescence Spectroscopy, 3rd edition, Springer-Verlag, New York,

Nakagawa, H. et al. 2011. Detection of a new Fusarium masked mycotoxin in wheat grain by high-resolution LC–OrbitrapTM MS. Food Additives and Contaminants 28: 1447.

Nakagawa, H. et al. 2012. The use of LC-Orbitrap MS for the detection of Fusarium masked mycotoxins: the case of type A trichothecenes. World Mycotoxin Journal 5: 271.

Nakagawa, H. et al. 2013. Detection of type A trichothecene di-glucosides produced in corn by high-resolution liquid chromatography-Orbitrap mass spectrometry. Toxins 5: 590.

Oto, N. et al. 2013. Non-destructive evaluation of ATP content and plate count on pork meat surface by fluorescence spectroscopy Meat Science 93:579.

Thomas, E.J. et al. 1991. Sensitive and specific detection of Listeria monocytogenes in milk and ground beef with the polymerase chain reaction. Applied and Environmental Microbiology 57:2576.

Uemura, K. and Isobe, S. 2002. Developing a new apparatus for inactivating Escherichia coli in saline water with high electric field AC. Journal of Food Engineering 53:203.

Uemura, K. and Isobe, S. 2003. Developing a new apparatus for inactivating Bacillus subtilis spore in orange juice with a high electric field AC under pressurized conditions. Journal of Food Engineering 56:325.




Read more

AgriculturalPolicy DragonFruitNetwork