Proteomics and Food Safety

Read Also

The role of bacteria in food processing and food safety

Foodborne illnesses result in numbers of hospitalizations and even deaths. Each year in the USA, about 325 000 hospitalizations and 5000 deaths caused by food poisoning are registered. Unfortunately, microorganisms and microbial toxins, especially foodborne ones as weapons of mass destruction still remain a threat. In food technology and biotechnology, careful monitoring of microbial contamination in the final product as well as monitoring of the production process and cleaning and sanitation are one of the most essential factors of the manufacturing process (44). The identification, confirmation, and quantification of bacteria and bacterial toxins in food are important analytical problems. The most common bacteria that cause food poisoning are Staphylococcus aureus, Campylobacter jejuni, some Salmonella and Staphylococcus species, some Bacillus strains and Escherichia coli O157:H7 strain. There are well-established and sensitive methods for detection of bacteria and their toxins available, mostly based on immunochemical methods. Proteomics and genomics technologies offer further, more sensitive and specific methods for identification of microbial food contaminants and their toxins, and for monitoring of cleaning and sanitation (45-48).

There are only few investigations that follow changes of proteomics of contaminating bacteria during food processing and equipment sanitation. The use of high hydrostatic pressure (HHP) technology is a new method for food preservation. Proteins are known to be the most important target of high pressure in living organisms (49) and HHP inhibits the growth of microorganisms by inactivating key enzymes that are involved in DNA replication and transcription enzymes and modifying both microbial cell walls and membranes (50). However, some bacteria such as Bacillus cereus can survive HHP treatment. Martinez-Gomariz et al. (51) analyzed changes in the proteome of this model organism during the HHP treatment. They found quantitative differences and identified some of differently expressed proteins. As expected, the expression of some proteins involved in nucleotide metabolic process was changed, but some other proteins such as those involved in carbohydrate catabolic process and transport, refolding, amino acid biosynthesis and bacterial ciliary and flagellar motility were also differentially expressed.

In a remarkable study, Boehmer at al. (52) follow proteomic changes in whey samples from a group of cows before and 18 h after infection with E. coli. Due to decreased milk production and quality, discarded milk and cattle mortality, such infections can cause mastitis, which is the most costly disease that affects the dairy industry. The aim of this study is the identification of biomarkers for evaluation of the efficacy of adjunctive therapies in decreasing inflammation associated with mastitis. Higher expression of some acute phase proteins such as transthyretin and complement C3 were found in whey samples 18 h after bacterial infection, but also some antimicrobial peptides and further acute phase α-1-acid glycoprotein were also detected. These biomarkers are candidate for future research into the effect of bacterial inflammation during mastitis.

As mentioned above, biofilm formation is an important fact that has to be taken into consideration during design of cleaning of stainless containers and other surfaces in food processing facilities. This problem has already been discussed in a review paper about microbial proteomics (5). In biofilms, some microorganisms such as sporogen bacterium Bacillus cereus (53-55), the Gram-positive bacterium Listeria monocytogens (56) and some pathogenic E. coli strains (57) can survive on the surface of stainless containers and other surfaces in the manufacturing facility, even under cleaning and sanitizing conditions. Better knowledge of biofilm formation and conditions that cause its degradation is necessary to prevent contamination by the above listed bacteria (58). Other biofilm-forming bacteria, such as Staphylococcus species (59) can survive food processing and cause human and animal infection. Incorporation of microorganisms is a kind of the natural way for their immobilization, and the high density of biofilms gives them better ability to survive aggressive treatment, but also a substantial, biocatalytic potential. The use of immobilized bacterial cells and bacterial biofilms for biosensors for food quality analysis and fermentation process control has been discussed elsewhere (5,18,60), and use of immobilized yeasts in brewing and winemaking processes will be presented later. In summary, in addition to physiological and genomic analyses, proteomic analysis of biofilm-forming microbial cells gives valuable information about their behavior during food processing and storage, symbiosis, possible infection and potential food poisoning, their defense against antimicrobial agents, and the potential to survive the cleaning and sanitation process (5,18,58).

The health-promoting properties of some bacterial species that colonize the human gastrointestinal tract have been documented in clinical trials and they are gaining popularity as food additives (61). Bifidobacteria and lactobacilli are the most popular microorganisms that are added as live bacteria to food preparations under the generic name of probiotics (61-63). The proteomic map of Bifidobacterium longum, a strict fermentative anaerobe, was first performed about five years ago (64,65). The topics of the following investigations included the survival mechanisms of this bacterium focused on altered protein expression following bile salt, heat or osmotic shock, which these bacteria are exposed to in the human gastrointestinal tract and during the food manufacturing process (66-68, for review see 69). These studies can also be used as a model for survival of other bacteria under similar conditions (69,70).

 

Prions

All prion diseases or transmissible spongiform encephalopathies (TSEs) are characterized by the deposition of an abnormal conformation (PrP^Sc) of a normal cellular protein (PrP^C) in neural tissues in humans and animals. The different protein conformations are associated with different physicochemical properties (71). PrP^C is relatively soluble and protease-sensitive, while PrP^Sc is relatively insoluble and protease-resistant. TSEs include scarpie in sheep and goats, and bovine spongiform encephalopathy (mad cow disease or BSE). Human form of this disease is infectious Creutzfeldt-Jakob disease (CJD) caused by the consumption of meat and meat products of prion infected animals (71,72). The outbreaks of BSE and infectious variant CJD have prompted the need for reliable screening methods for prion infections as part of the safety control for meat and meat products. Identification of prion proteins is usually a time-consuming process and includes immunoaffinity techniques, combined with one- and two-dimensional electrophoresis and mass spectrometry (73,74). Although intensive studies have been performed, it is still long way to identifying reliable biomarkers for prion infection. Detection of prion-binding proteins did not give further revealing information about the biology of prions and the pathogenesis of TSE (72, 74-76). One of potential biomarker candidates is ubiquitin. This protein could be identified in the cerebrospinal fluid of CJD patients (77). However, this recent study has been performed only with a small number of samples, and ubiquitin as a highly abundant protein cannot be taken in consideration as a reliable biomarker. Herbst et al. (78) used a multidisciplinary approach to identify antemortem markers for prion disease. This rather complex strategy combines matrix-assisted laser desorption/ionization Fourier transform mass spectrometry (MALDI-FTMS), mass fingerprinting and bioinformatics for identification of candidate biomarkers in infected animals. Again, results of this study are still rather limited, and true positive rate was relatively low. More promising is recently published study by Nomura et al. (79). This group reported detection of autoantibodies in the sera of cattle with bovine spongiform encephalopathy. These autoantibodies were directed against glial fibrilary acidic proteins, and could be detected only in the serum of TSE-infected animals.

Tsiroulnikov et al. (80) presented a method for decontamination of animal meat and bone meal by use of bacterial proteolytic enzymes. Nattokinase from Bacillus subtilis that has been used for fermentation of boiled soybeans is also able to degrade prion proteins and potentially prevent prion infection (81). However, it is still a safety risk, if such contaminated animal food is used, and prion detection and elimination of diseased animals and contaminated meat (74-79) is a much safer way to prevent these kinds of foodborne diseases.

Allergens and toxic components

Proteins are responsible for many allergic reactions. The most threatening allergic reaction, anaphylaxis, is most frequently caused by peanuts or tree nuts (82). That is also the reason that most proteomic investigations towards identification and quantification of allergens were performed on food of plant origin (83). Milk and milk products, as well as seafood and processed food are other kinds of food that cause allergies (35). However, there are only few investigations of animal proteins involved in these adverse reactions.

Proteomic strategies used in order to achieve more detailed and comprehensive characterization of food allergens are referred as ‘allergenomics’ (84). The common procedure for detection of proteins involved in allergic reactions is protein extraction (e.g. with 8 M urea with 4 % CHAPS, buffered with 40 mM Tris. HCl, pH=7-8), electrophoretic separation (SDS-PAGE or 2D electrophoresis), and detection of IgE binding proteins by immunoblotting. After tryptic digestion, the IgE binding proteins as potential allergens can be identified by mass spectrometry (84,85). This very effective, but also time-consuming method is similar to the method presented in Fig. 2. Stevenson et al. (82) use gel-free, label-free quantitative approach for identification of peanut allergens. Quantitative evaluation was achieved by peak integration and spectral counting in comparison with a protein standard. The workflow of this analytical procedure is shown in Figure 2. In the future, this method could be useful for high-throughput profiling of proteins, including seed allergens. However, more standardization and validation are still necessary.

Figure 2. (ref. 82)

Most allergies in the USA are caused by peanuts and peanut containing food products (82), and peanut proteins that may cause allergies are well characterized. Chassaigne et al. (86) use 2D electrophoresis, immunoblotting and high-resolution mass spectrometry for allergen detection in peanut seeds. They detected several isoforms of main allergens Ara h 1, Ara h 2, Ara h 3/4. Proteomic analyses show different contents of these allergens in different peanut varieties, and also the presence of several fragments of these proteins (87). As shown by Stevenson et al. (82), these proteins are absent from genetically engineered peanut seeds.

Bässler et al. (88) use a multidimensional protein fractionation strategy and LC-MS/MS for the molecular characterization of tomato seed allergens. In subsequent in silico modeling, high homology between epitopes of known allergens from walnut (89), cashew nut (90) and buckwheat (91) was found. Further proteomic analyses of plant proteomes were performed to detect allergens in wheat flour (83), maize (92) and sesame seeds (93).

By use of sophisticated quantitative proteomics technology, Chassaigne et al. (86) showed that genetically engineered peanut seeds contained significantly reduced amount of main allergens. Genetically modified (GM) tomato and soybean plants are approved for food use by the US Food and Drug Administration. During the assessment procedure, the allergic properties of the gene donor and the recipient organisms are considered in order to determinate the appropriate testing strategy. The amino acid sequence of the encoded protein was compared to all known allergens to assess whether the protein is a known allergen (88) to indicate a probability of allergic cross-reactivity and formation of neo-allergens. Further risk of food allergenicity is the stability of the protein in acidic environment in the presence of stomach protease pepsin had also to be tested. These tests were followed by in vitro and in vivo binding assays to human IgE, and no adverse reactions were found (94). However, some residual risk after long-term consumation of such food still remains, and further studies regarding allergenic potential of GM plants were performed. In subsequent proteomic study, GM versus non-modified soybean samples were compared, and 2 new potential allergens were indeed identified. In a short-term study, none of the individuals tested reacted differently to the GM versus non-modified samples (95). After this study, a residual risk of allergies after long-term consumption of GM crops still remains.

Food of animal origin, especially seafood and milk products can also cause allergies. However, proteomics tools have only been sparingly applied in the investigation of allergens in these products. It is well known that changes in the main milk protein casein such as carbonylation (36) or forming of covalent complexes between casein micelles and β-lactoglobulin (96) and modification of other proteins (97) during the production process, mainly heating, can cause induction of allergies to milk products, but a thorough proteomic and ‘allergenomic’ investigation has still to be performed. In their review about the use of proteomics as a tool for the investigation of seafood and other marine products, Piñeiro et al. (9) recommend the use of proteomics for detection of allergens in food of this origin. However, there are still only few studies in this field. Taka et al. (98) characterized an allergenic parvalbumin from frog by the use of LC-ESI-MS. The main crustacean allergens are proteins tropomyosin and arginine kinase (99,100). Tropomyosin is a myofibrillar protein of 35-38 kDa, and proteins from six species of crustaceans have also been cloned (101). Arginine kinase from some commercially relevant shrimp species was characterized by use of proteomic methods (102). Some additional shrimp allergens such as sarcoplasmatic calcium binding protein (SCP) have also been detected (103,104). Interestingly, this protein was previously detected as allergen in crayfish Procambarus clarkii (105). This finding further confirms the thesis of Bässler et al. (88) about shared epitopes in allergens of different origin.

If not inactivated or degraded during processing, some food components such as plant lectins constitute a possible risk, since consumption of raw or incorrectly processed beans can cause outbreaks of gastroenteritis, nausea, diarrhoea, and even more severe side reactions. Most plant lectins are secretory proteins. After secretion, they accumulate either in vacuoles or in the cell wall and intercellular spaces, mostly in seeds. Lectins such as concanavalin A, phytohemagglutinin, pea lectin and flavin are present in quite high levels and accumulate in vacuoles in cotyledons (106, 107). Most lectins show high specificity to distinct sugars, but they also have an extensive homology in primary structure, also from unrelated species. On the other hand, a plant species such as castor bean may contain structurally related lectins with different toxicity. Castor bean lectin ricin shows relatively weak agglutination, but very high toxicity for humans and animals; Ricinus communis agglutinin is weakly toxic, but a strong agglutinin (106).

Ricin and Phaseolus vulgaris lectin are two most common lectins that cause food poisoning (106,108). In humans, consumption of other raw beans can also cause gastroenteritis, nausea and diarrhoea (109). On the other hand, bean extracts enriched with in lectins or lectin-related amylase inhibitors are used as active ingradients of so-called ‘weight-blockers’ in dietetic preparations (110,111). Proteomic strategies to quantitative analysis of potentially harmful lectins in raw and processed food in dietary preparations include the use of chromatographic or electrophoretic strategies combined with mass spectrometry (LC-MS/MS, MALDI-TOF MS or MALDI-TOF/TOF MS). Affinity chromatography with immobilized glycoproteins or oligosaccharides can be used for enrichment of lectins. Lectin-enriched fraction can be further separated, e.g. by cation-exchange chromatography, followed by tryptic digestion and protein identification by mass spectrometry (106). However, these methods are still complex, expensive and time consuming. After detection of these potentially harmful components by proteomic methods, specific, ‘food based’ protocols, e.g. ELISA or other simple and fast protocols for their detection and quantitative determination can be developed.

In order to increase muscle accretion and reduce fat deposition, cattle are treated by anabolic steroids (112). All biochemical events that are caused by steroid use are oriented towards anabolic metabolism, resulting in a lower tyrosine aminotransferase as a marker of catabolism and a higher muscle building (113). Use of steroids can be detected by genomic or proteomic methods (114-116). In a study performed on calves, differential expression of adenosin kinase and reticulocalbin in the liver of calves treated with anabolic compound was found (116). It was also shown that metabonomics can be effectively used to study the different disruptive metabolic response in cattle after the use of anabolic steroids (10,112). Several biomarkers such as trymethilamine-N-oxide, dimethylamine, hippurate, creatine and creatinine were detected in urine of cattle treated with anabolic steroids. These urinary biomarkers characterize the biological fingerprint of anabolic treatment.

Pharmacological practices that are used to increase protein production in livestock can be detected by metabonomic and proteomic techniques that can be used as alternative techniques for screening analysis of veterinary drugs in animal products (116,117). Long-time and low-dose treatment administration of antibiotics, mainly tetracyclines, has also been used as a growth promoter in livestock production. This use is banned today in the European Union (118). One of the main reasons is that systematic antibiotic use promotes the development of resistant bacterial population (119). As already discussed above, the use of proteomics to elucidate molecular mechanisms of meat quality is well established (28). Gratacós-Cubarsí et al. (120) demonstrated that after administration, tetracyclines are rapidly degraded, but in the muscle of pigs treated with tetracyclines, several differentially expressed proteins were detected. Five spots in 2D electrophoresis that belong to differentially expressed proteins and candidate biomarkers for tetracycline treatment were identified as enzymes involved in muscle metabolism and two novel porcine proteins (120). Similar differences were also observed in composition of egg proteins from treated and non-treated chicken (121).