Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Lactic acid bacteria (LAB) are the main agents ensuring the safety and organoleptic properties of fermented foods. LAB encompass more than 10 genera which include a large number of species (above 100). However, the amount of LAB species used as starters for food fermentation is limited. In the dairy industry,Lactobacillus delbrueckii ssp. bulgaricusandStreptococcus thermophilusare the only two species involved in yogurt production, whereasLactococci(L. lactis,L. cremoris,L. diacetylactis) and otherLactobacilli(L. helveticus,L. acidophilus,L. casei) are also involved in cheese making. In fermented meat products,Lactobacillus sakei,Lactobacillus curvatus,Lactobacillus plantarum,Pediococcus pentosaceus, andPediococcus acidilacticiare the main starters.Oenococcus oeniis the emblematic wine starter. In addition, the natural microflora of the components used for food fermentation or present in the final fermented products may also encompass a great diversity of LAB species. However, most of the scientific studies have focused on starters. Concerning proteomics, its use to study the adaptation or functions of starters in fermented products is mainly restricted to laboratory conditions or to a few model matrices. This limitation results from a major technical problem: food matrices, whatever their animal origin (milk, meat) or vegetal origin are rich in proteins and peptides that may interfere with the study of LAB involvement. Indeed, many proteomic analyses dealing with food refer to the food matrix rather than to the microbes present in it (D’Alessandro and Zolla2012; Sentandreu and Sentandreu2011; Molina et al.2002). In many fermented products, the biomass represented by bacteria is negligible when compared to the proteins and peptides from the food components, and proteomic analyses of these products therefore do not require a previous removal of bacteria. As examples of such studies to investigate meat quality, see Vallejo-Cordoba et al. (2010) and Nakamura et al. (2010). In the present chapter we present studies dealing with a proteomic approach that allows exploring the ability of lactic acid bacteria to grow in fermented food models, their important functions for the quality of the products, and their ability to resist the harsh conditions encountered during fermentation processes. A few proteomics examples related to food spoilage and food safety, or to bacteria fingerprinting are also presented.

2 Generic Analyses in Laboratory Model Conditions: Adaptability to Conditions Mimicking the Food Processes

2.1 Introduction

In nature, the ability to respond quickly to stress is essential for survival. Indeed, under adverse conditions, bacteria set up cellular responses and defense mechanisms that significantly improve their chances of successful adaptation to harsh and/or sudden environmental changes. LAB, as do other bacteria, have to face various and often simultaneous stresses during starter handling and storage and also during food processes, in particular physicochemical stresses such as freezing, cold, heat, extreme pH, osmotic pressure, oxidative agents, or high hydrostatic pressure (HHP). A better comprehension of stress resistance is necessary to understand the adaptative response to these unfavorable conditions and thus to predict the potential functions that would be essential for the survival/maintenance of starters under such industrial conditions. This would allow and rationalize the preparation of adapted strains and their improvement for specific industrial applications. Table15.1proposes a synthetic summary of proteomic studies reported in the literature, and dealing with LAB starters of dairy products (L. lactis,L. delbrueckii ssp. bulgaricus,L. helveticus,L. casei,L. acidophilus,S. thermophilus), sourdough (L. sanfranciscensis), sausage (L. sakei), or a starter that can be found in all of these products (L. plantarum). In these articles, authors investigated adaptation of the bacteria to conditions encountered during fermentation processes. Table15.1summarizes only proteins whose expression was upregulated, representing therefore functions and regulatory mechanisms necessary for bacteria to overcome these stressful conditions.

Table 15.1 Synthetic overview of proteins up-regulated by different starters grown under various conditions. The LAB species studied are used for various food types: sourdough, dairy products (yogurt, cheese), meat (sausage) or for all of these. The type of stress applied is listed, as well as the experimental conditions. Proteins are listed according to their functional category
Full size table

We analyzed these data to search for: (1) proteins commonly upregulated under several stressful conditions and/or upregulated by several species thus representing a global stress signature, (2) species-specific responses, and (3) stress-specific responses.

2.2 Global Stress Signatures

From all data summarized in Table15.1, general stress proteins are among the most important families of proteins characterizing the bacterial response to various stress conditions: these proteins are induced whatever the stress or the food matrix, and in all species. As an example, several of the Clp proteins both respond to different stresses in one bacterial species, but also respond to the same stress in different bacterial species: ClpL is upregulated after both oxidative-, heat-, cold- and lactose-starvation stresses inS. thermophilus. ClpP responds to lactose starvation (which provokes growth arrest) inS. thermophilus, and also to the entry in the stationary phase inL. plantarum. It also responds to a cold shock in bothS. thermophilusandL. acidophilus. It is known that a pre-adaptation to one stress condition helps bacterial cells to become resistant to other stresses. This reflects the multiple stresses that can be encountered during food processes. In addition, this may explain the multiple protections ensured by stress-response proteins. For example, a pre-adaptation ofL. delbrueckiissp.bulgaricusto acid confers a better cryotolerance.

Another relevant feature in Table15.1is that both general stress proteins and proteins described as specific for a stress, such as oxidative stress-response proteins, actually respond to a large panel of stressful conditions. For instance, Dpr is indeed upregulated after an oxidative stress inS. thermophilus, but also after acid stress in this bacterium and after conditions leading to growth arrest inS. thermophilus,L. plantarum, andL. helveticus.

Proteins involved in carbohydrate transport and metabolism generally respond to conditions reflecting growth rate such as entry to late exponential or stationary phases, lactose starvation, or decrease of growth temperature. Some of those, however, are also upregulated after various stresses. In addition, upregulation of proteins involved in nitrogen metabolism is mostly observed under conditions of growth arrest (entry to late exponential or stationary phases, lactose starvation) in different LAB species. A similar, although less pronounced, tendency is observed for proteins involved in nucleotide transport and metabolism. This certainly shows that a central metabolism is required for bacteria switching from optimal conditions for fast growth, to harsher conditions. This implies that bacteria should not be considered as dormant when they enter a famine state after feast conditions.

Interestingly, upregulation of proteins involved in translation or transcription arises after various stress conditions, rather than under growth-arrest conditions. In addition, these upregulated proteins are also typical of the response ofL. sakeiand also, to a lesser extent ofL. sanfranciscensis, after HHP treatment.

2.3 Species-Specific Responses

Among the data available in the literature, only a few species comparisons can be performed. Growth under acidic conditions or acid stress is among the conditions that have been best described in lactic acid bacteria. Indeed, LAB producing lactic acid and low pH is one characteristic of fermented products. For that reason, many proteomic studies report the effect of low pH in several starter species. In Fig.15.1, the response of three dairy LAB (L. lactis, L. delbrueckiissp. bulgaricus, andS. thermophilus) to one stress, acidic conditions are compared in order to detect if a species-specific response exists.L. lactisis the most responsive species, with 61 proteins upregulated followed byL. delbrueckissp.bulgaricuswith 21 over-expressed proteins andS. thermophiluswith 12 over-expressed proteins. Strikingly, a substantial species-specific answer is observed. Indeed, only four upregulated proteins are shared by the three species: three general stress proteins (GroES, GroEL, HSP) and lactate dehydrogenase.S. thermophilusandL. delbrueckissp.bulgaricushave no other common upregulated protein,L. delbrueckissp.bulgaricusandL. lactishave one, andL. lactisandL. delbrueckissp.bulgaricusshare five. InS. thermophilus, the functions induced after an acid shock may contribute to the intracellular pH homeostasis with the upregulation of H+-ATPase responsible for proton extrusion, and urease able to produce NH3. These results clearly indicate that each species has a specific protein pattern. Each species over-expresses similar functional categories, but the targeted proteins are specific for each bacterium.

Fig. 15.1
figure 00151

Venn diagram representing the response ofS. thermophilus, L. lactis, andL. bulgaricusto acid conditions. The up-regulated proteins common to several species or species-specific are presented

Full size image

In Table15.1we can also observe that, in the oxidative stress-response protein category,L. lactisover-expresses AhpC, SodA, and Tpx, whereasL. delbrueckii ssp. bulgaricusandS. thermophilusover-express only an oxidoreductase and Dpr, respectively.

Also,L. lactisshowed upregulation of more glycolysis proteins when compared to other species: under acidic stress conditions,L. lactisover-expresses twelve proteins whereasL. delbrueckiissp.bulgaricusandS. thermophilusover-express, respectively, six and one proteins of the carbohydrate metabolism category. In conclusion, for a given stress, a species-specific response indeed exists.

2.4 Stress-Specific Responses

Regarding data presented in Table15.1, we also considered the different stresses in order to evaluate a possible stress-specific response in LAB. As an example, we considered three stresses: heat, cold, or acid shock and compared the response of one LAB species,S. thermophilusto each of these three stresses.As shown in Fig.15.2, only one general stress protein (GroEL) is upregulated following each of the three stresses. Some proteins are common to the response to two different stresses. Indeed, four proteins (HSP, EF-Ts, EF-G, GroES) are upregulated under either heat or acid conditions; two other proteins (Gls24, EF-Tu) are upregulated either by cold or acid shock. Only one, ClpL, is common to heat and cold shock response. It thus appears that, in addition to a common central stress response,S.thermophiluselicits different protein sets in order to adapt to each environmental modification.

Fig. 15.2
figure 00152

Venn diagram representing the response ofS. thermophilusto three different stresses. The up-regulated proteins common to several stresses or stress specific are presented

Full size image

Regarding oxidative stress response,S. thermophilusover-expresses only stress-response proteins, five being general stress proteins, and the seven remaining specifically targeting the oxidative metabolism. Compared to other stresses applied, in which several functional protein categories are generally upregulated, this is the sole example of such a response focused on only a given functional category.

Finally, we noticed that during growth at low temperature two proteins characteristic of oxidative stress response (MsrA and OhrA) were upregulated inL. sakei, whereas inS. thermophilusthe response was mainly characterized by induction of general stress proteins, in particular the cold shock proteins CspA and B.

Concerning non-dairy LAB, HHP treatment has also been studied. Indeed HHP can be used as a preservative treatment, alternative to filtration or heat sterilization that is not applicable in some solid and raw food matrices (dry sausage, sourdough). After an HHP treatment,L. sakeiandL. sanfranciscensisshow quite different responses. Both over-express general stress proteins but different ones. The only common protein upregulated after HHP treatment in both bacteria is DnaK.

3 Analyses in Food Matrices

As mentioned above, proteomics studies of LAB behavior in food products has been mainly limited by the high protein content of food materials. A very few in situ studies have been reported, mainly concerning cheese.Yvon et al. (2011) have compared the ability in cheese making of twoL. lactisstrains with different properties. One strain was more proteolytic, as evaluated by growth in cheese matrix, amino acid production, and proteolytic enzymatic activities. For the two strains, regarding protein expression pattern, functional categories affected by growth in cheese were mainly related to acid stress response and amino acid starvation, indicating that strains have to cope with these stressful environmental conditions when grown in cheese. However, differences were observed between the two strains. A stronger growth limitation for one strain was linked with a lower proteolytic activity. This study underlines the importance of proteomics for evaluating dynamics of LAB population in cheese and consequent technological properties enlightening thatL. lactischeese adaptation depends on proteolytic activities.

Regarding the cheese process, several LAB can be involved. The Emmental cheese ecosystem, for instance, is composed of complex microflora including different LAB species (L. helveticus, L. bulgaricus, S. thermophilus) andPropionibacterium freundenreichii.During cheese ripening, bacterial proteins can be released due to bacterial cell lysis. These proteins contribute to cheese organoleptic and textural properties. Gagnaire et al. (2004) have performed a proteomic study in order to evaluate the respective part of each flora in the ripening process. Proteins released in the soluble fraction, after exclusion of milk and cheese proteins by chromatography, have been analyzed by two-dimensional gel electrophoresis. They could identify 21 proteins fromS. thermophilus, 17 fromL. delbrueckissp. bulgaricus, and 8 fromP. freundenrichii.These proteins were allocated to different functional categories such as stress proteins, DNA repair, or oxido-reduction, indicating a stress cellular status of the LAB during cheese ripening. Several peptidases were also identified, some being attributed toS. thermophilusand others to ­L. delbrueckissp.bulgaricus, enlightening the respective roles of these two species in the ripening process.

Proteomic analysis (MALDI-MS) has also been successfully used to monitor the extent of bacterial digestion of milk proteins during yogurt production (Fedele et al.1999). This study showed that: (1) changes in milk protein profiles were due to the action of the two yogurt LABS. thermophilusandL. delbrueckissp.bulgaricus, (2) casein hydrolysis varied among the strains tested, and (3) proteolysis was relevant when milk was fermented with mixtures of the two yogurt bacterial species, probably because of synergistic phenomena. This tool could be employed to determine the effectiveness of different yogurt LAB ratios for yogurt production, in terms of production of peptides that can bear probiotic properties.

Recently, a new methodology has been applied to the proteomic analysis of cheese ripening (Jardin et al.2012). The ITRAQ (isobaric tagging for relative and absolute quantification) method allows identification and quantification of proteins in a single LC-MS/MS run. It was used to follow Emmental-type cheese during ripening, by analyzing proteins present in the aqueous phase. By this method, the authors could overcome differences in the dynamics range between milk and bacterial proteins, and could identify proteins from both origins. Proteins from milk did not show significant increase in the aqueous fraction. Bacterial proteins identified were issued fromL. helveticusorS. thermophilus. A major increase in their concentration was observed between day 20 and day 69 of ripening. These bacterial proteins were mainly assigned to the functional category of stress proteins. This study thus confirmed the previous finding of a stress status for LAB during cheese ripening. It also opens fields for a simultaneous analysis of cheese and bacterial proteins and thus a better global understanding of cheese making.

No such deep in situ comprehensive approach has been reported for meat, even though some studies have attempted to approach meat matrix conditions. Fadda et al. (2010) have studied the response ofL. sakeigrown in a chemically defined medium after addition of either myofibrillar or sarcoplasmic proteins. Most of the proteins showing modified expression were observed in the presence of myofibrillar extract (16 proteins) compared to sarcoplasmic protein addition (6 proteins). Most proteins were less expressed in the presence of the meat extracts. General stress proteins were downregulated whereas proteins of energy metabolism, pyrimidine metabolism, and translation were upregulated indicating thatL. sakeiis really adapted to the presence of the meat substrate. In other words, this means that meat is indeed the favorite habitat ofL. sakeiand that laboratory media constitute a stressing environment. This situation is not observed with cheese bacteria that show a stress status during the cheese process.

4 Food Safety: Biogenic Amine Production

Proteomics have also been used as a tool either to detect or study biogenic amine production, or even for the fingerprinting of biogenic amine-producing strains or species. Although many fermented foods may be confronted with the risk of ­biogenic amine production, wine and dry fermented sausages are the most critical. In ­addition, fish products, fermented or not, are also susceptible to host bacteria able to produce biogenic amines such as histamine, tyramine, or putrescine. A proteomic analysis, combining two-dimension gel electrophoresis to MALDI-TOF and MS-MSde novosequencing, was used to determine proteins synthesized by twoLactobacillusstrains (Lactobacillussp. 30a ATCC33222 and aLactobacillussp. strain isolated from an amine-contaminated wine) under various laboratory growth conditions (Pessione et al.2005). The results helped to identify proteins involved in amino acid transport and conversion to biogenic amines, but could not confirm the hypothesis that amino acid accumulation induces biogenic amine synthesis, nor that this synthesis is a bacterial response to medium acidification.

5 Identification, Fingerprinting, or Characterization of Bacterial Diversity by Proteomics

The MALDI-TOF technique can be used to fingerprint bacteria (Giebel et al.2010) including biogenic amine producers. As an example, the MALDI-TOF spectra obtained on bacterial extracts fromPseudomonasstrains isolated from fish products gave different fingerprints when obtained from biogenic amine producing or nonproducing strains (Fernández-No et al.2011). A similar approach was used aimed at identifying bacterial species present in seafood products, including the LAB speciesCarnobacterium divergens,Carnobacterium gallinarum, andCarnobacterium maltaromaticumand also twoStaphylococcusspecies (S. epidermitisandS. xylosus) which are naturally present or used as starter for the fermentation of sausage (Böhme et al.2011). These examples show that the pattern of MALDI-TOF spectra obtained from food bacteria can be used as a tool to differentiate strains or genus, or even to identify new isolates.

Proteomics have also been used as a tool for investigating LAB natural diversity. The meat-borne bacteriumL. sakeihas been described as a species showing high genomic diversity and 10 genotypic groups could be described (Chaillou et al.2009). Ten strains representing these clusters have been studied by a proteomic approach based on two-dimensional gel electrophoresis. This revealed a difference of up to 20% in the number of expressed proteins detected in gels, with specific strain protein patterns. Moreover, it was observed that the ten strains could split into two clusters, based on the pattern displayed by the four isoforms of the glyceraldehyde-3P-dehydrogenase (GapA) on gels. These two isoform families were associated with the twoL. sakeisubspecies previously reported in the literature asL. sakeisubsp.sakeiandL. sakeisubsp.carnosus.

Regarding the expression of this diversity, the metabolism ofL. sakeistrains has also been explored by proteomics in order to evaluate the potential of strains for technological applications. McLeod et al. (2010) have compared tenL. sakeifood isolates, when grown on either glucose or ribose as the carbon source, these two sugars being present in meat. A common regulation by ribose was observed for all strains but the study also pointed out differences between strains. For all strains, a total of ten proteins was upregulated after growth on ribose and six proteins were downregulated on glucose. A commercial starter strain and a putative b­iopreservative strain exhibited a different regulation in the utilization of these two carbon sources. One strain, isolated from fermented fish showed a higher level of expression of stress proteins when compared to the other strains.

Guillot et al. (2003) compared the 2-D gel electrophoresis profiles of twoL. lactisstrains. Apart from the whole set of glycolytic enzymes that were present in both strains, at conserved positions, this study revealed an important protein polymorphism. For strain NCDO763, a dairy strain belonging tocremorissubspecies, 26 proteins were not present in theL. lactissubsp.lactisIL1403 strain map. Among these, there were enzymes involved in lactose assimilation and amino acid metabolism (peptidases, amino acid synthases) that are of particular relevance for growth on milk and thus for cheese making.

These reports indicate that the proteomic approach can also be a useful tool for exploring bacterial diversity, in association with other methods such as genomics, to go deeper into LAB strain performance analysis for a better understanding of strain-specific technological properties.

6 Conclusion

Proteomic analyses of LAB from fermented foods have shown that bacteria use different strategies to face the various conditions encountered during the food production processes. When confronted with the same stress, different species will modify the protein they synthesize in different but similar ways: the functional categories that are affected by a specific stress are often similar, but the proteins that are upregulated after this stress may vary from species to species. The degree of similarity or difference between strains varies with the kind of stress or growth condition tested.

However, most of the proteomic results presented to date in this field issue from two-dimensional gel electrophoresis followed by mass spectrometry protein identification. One should not forget that the method mostly refers to cytoplasmic proteins, in a limited pI and MW window, and that only the emerging part of the iceberg is accessible to analysis. This narrow point of view is, however, informative about bacterial functions that have to be taken into consideration to understand and improve the production of fermented foods.

In addition, recent LAB whole genome sequencing studies have revealed large intraspecies diversity in both genome size and content. This means that stress response observed with one strain may be different in another strain. Whole mechanisms of bacterial adaptation to fermented food conditions are yet to be totally elucidated. Proteomics have allowed, nevertheless, the indication of some marker proteins that may be useful for the selection of well-adapted efficient starters for industrial applications.