Fermentation of food products: optimised lactic acid bacteria strains with reduced potential to accumulate biogenic amines

Biogenic amines (BA) are a major cause of food poisoning. Their appearance in food is the result of unwanted microbiological activity. Understanding the metabolic pathways and enzymes involved in BA production by LAB is essential for the understanding of the mechanism behind the spoilage of food and beverages. BA are the end product of metabolic pathways that provide the microorganisms with metabolic energy, in the form of either ATP or proton motive force. Pathways producing proton motive force are simple pathways involving the decarboxylation of amino acids yielding the BA. The ATP producing pathways are more complicated and involve deimination of guanidine groups. Both types of pathways involve transport proteins embedded in the cytoplasmic membrane, which are responsible for both the uptake of the precursor and the excretion of the BA in the cell. We have expressed three of these so-called precursor/product exchangers in a heterologous expression system and have functionally characterised them. The three transporters are: - The histidine/histamine exchanger HdcP of Lactobacillus hilgardii 0006; - The tyorisine/tyramine exchanger TyrP of Lactobacillus brevis; - The agmatine and putrescine exchanger AgmP of Lactobacillus brevis. The transporters were expressed in the lactic acid bacterium Lactococcus lactis under control of the inducible nisin promoter. The structural genes coding for the transporters were cloned in vector pNZ8048 yielding plasmids pNZhdcP, pNZtyrP and pNZagmP, which encodes the exchangers with an additional N-terminal His-tag that was added for expression studies. The plasmids were transformed to L. lactis strain NZ9000 and the expressed proteins characterised in resting cells or membrane vesicles prepared from the cells. The histidine transporter HdcP was identified as a membrane protein using antibodies raised against the N-terminal His-tag. It showed up in the membrane as a single protein with a apparent molecular mass of 40 kDa on SDS-PAGE (calculated molecular weight 55 kDa). L. lacis cells expressing HdcP had a twelve fold higher uptake activity of histidine relative to the control cells. Chase experiments demonstrated the ability of the transporter to catalyse homologous histidine/histidine exchange and heterologous histidine histamine exchange. The hdcP gene of L. hilgardii is the first gene that was experimentally shown to encode for a histidine/histamine exchanger. The tyrosine transporter TyrP was identified as a protein with an apparent molecular weight of 44 kDa on SDS-PAGE (calculated molecular weight 54 kDa). Tyrosine uptake in cells of L. lactis expressing TyrP was stimulated 15-fold relative to cells lacking TyrP. Internalised tyrosine could be chased with excess tyrosine and with tyramine demonstrating the tyrosine/tyramine exchange capacity of TyrP. Studies with membrane vesicles with a right-side-out orientation demonstrated that TyrP had a low uniporter activity transporting tyrosine or tyramine in the absence of cotransport. The exchange activity between tyrosine and tyramine, which is very efficient, was shown to be electrogenic. The agmatine transporter AgmP was expressed in L. lactis as a protein with an apparent molecular weight of 35 kDa (theoretical molecular weight of 53 kDa). L. lactis cells expressing AgmP catalysed uptake of agmatine and putrescine to high levels, while the control cells were devoid of this activity. The recombinant cells showed no higher uptake activity for arginine or ornithine demonstrating that the agmatine deiminase pathway and the arginine deiminase pathway make use of distinct transporters. The uptake was strongly dependent on the membrane potential component of the proton motive force suggesting that the transporter catalyses electrogenic uniport of the divalent, positively charged substrates. Agmatine/putrescine exchange was demonstrated by chasing accumulated putrescine with agmatine and vice versa. The properties of the exchangers HdcP, TyrP and AgmP are those typical for precursor/product exchangers. They catalyse efficient exchange between two structurally related compounds. HdcP and TyrP operate in decarboxylation pathways that generate proton motive force and, therefore, catalyse electrogenic exchange. AgmP operates in a deiminase pathway that generates ATP and, therefore, catalyses electroneutral exchange. The transporters have in common that they catalyse unidirectional transport of the substrates at a much lower activity. The studies provide a framework to scientist in the LAB scientific community to identify and characterise other transporters in LAB that are potential BA producers.

Biogenic amines (BA) are the product of decarboxylation pathways in food bacteria and a major cause of food poisoning. Substrates of decarboxylation pathways are amino acids that are converted into the corresponding amines or amino acids (e.g. histidine/histamine, aspartate/alanine or tyrosine/tyramine). The pathways consist of a transporter that catalyses the translocation of the substrate into the cell coupled to the secretion of the metabolic end product of the same substrate out of the cell, and a decarboxylase residing in the cytoplasm. In order to prevent the formation of BA by the organisms, it is essential to understand the physiological relevance of the pathways for the organisms. The benefit of the pathway for the microorganisms may be in the generation of metabolic energy and/or acid stress resistance. The pathways function as indirect proton pumps and, therefore, generate proton motive force. The two components of the proton motive force are generated in the two separate steps of the pathway. The transporter exchanges two substrate that differ in charge, which results in membrane potential, and the decarboxylation steps consumes a cytoplasmic proton, which results in a pH gradient across the membrane. The alkalinising effect of the pathway is also responsible for the acid stress function. While the different properties of the enzymes in the pathway have been established in vitro, the physiological benefit can only be studied in vivo by studying the performance of the pathway as a whole. For this purpose we have constructed a recombinant Lactococcus lactis strain containing a synthetic operon encoding for a tyrosine decarboxylation pathway. A synthetic operon was constructed containing the tyrosine decarboxylase tyrDC and the tyrosine/tyramine exchanger tyrP in the L. lactis expression vector pNZ8048 under control of the nisin promoter. The two genes originated from Lactobacillus brevis. The tyrDC gene was placed upstream of the tyrP gene and was tagged with a His-tag. Expression of the TyrDC protein was demonstrated by immunoblotting using antibodies against the His-tag. Expression of the TyrP protein was demonstrated by preparing right-side-out membranes and measuring tyrosine/tyramine exchange activity. Two assays were developed to assay for tyrosine decarboxylase activity of the cells. In one assay, radioactive tyrosine labelled at the C1 position was added to the cells. The C1 atom of tyrosine is released from the molecule as carbon dioxide upon decarboxylation. As carbon dioxide is easily released from the suspension, decarboxylase activity may be monitored from the decrease of the radioactivity from the suspension. Samples are taken in time and the radioactivity is measured in a liquid scintillation counter. The assay is simple and fast but requires significant conversion. In the second assay, the cell suspension is incubated with uniformly labeled tyrosine. Samples are taken in time and spotted on a TLC plate that is developed in a mixture of propanol, butanol and water. Radioactive spots on the plates are visualized using a phosphoimager. The assay is time consuming, but allows for a good monitoring of the purity of the substrate and the products developed. The tyrosine decarboxylation activity of L. lactis cells expressing the pathway was about half the activity of cells expressing the tyrosine decarboxylase alone. L. lactis cells producing the TyrDC protein showed essentially no tyrosine decarboxylation activity suggesting that the produced protein was inactive. In agreement, fractionation of the cells followed by detection of the protein in the different fractions demonstrated that all protein was present as inclusion bodies. Several approaches were followed to reduce the formation of inclusion bodies in order to improve the decarboxylation activity of the cells. The approaches included optimising the inducer concentration, changes to the composition of growth medium, lower temperature during growth and a lower initial pH. Several conditions independently resulted in significant improvement of the activity of the cells: lowering of the growth temperature to 25C, lowering of the initial pH of the growth medium to 6, addition of KCl to the growth medium, and addition of compatible solutes. The common factor in these conditions is stress. The activity of the cells could be further improved by combining the different conditions. Ultimately, a cell suspension at an optical density of 50 could convert 1mM of tyrosine in 24 h.

93 Lactococcus lactis, 22 Enterococcus spp. and 74 Lactobacillus spp. strains isolated from Spanish traditional cheeses were screened by colorimetric methods to identify tyramine and histamina producers. Twenty two tyramine-producing strains were confirmed. Most of them (19) belong to Enterococcus genus, two were Lactobacillus curvatus and one Lb. brevis. Histamine is not as abundant in dairy products as tyramine, and in fact, the screening has not yielded any histamine producers. The results were confirmed by the PCR methods developed in this project. More strains of lactic acid bacteria (LAB) isolated from different kind of European cheeses were tested to identify tyramine-producing and histamine-producing strains. A surprisingly large percentage (40%) of analysed strains were tyramine producers. A significant lower proportion (9%) were histamine producers. An Argentinean collection of strains isolated from artisanal cheeses and dairy products was analysed for the production of biogenic amines. A total of 32 citrate fermenting LAB strains have been studied. Among the other 30 LAB, 26 showed to be tyramine producers and they were identified by biochemical and molecular methods as Enterococcus. Detection of the genes encoding the histidine and tyrosine decarboxylase has been performed using the PCR method developed in this project. The DNA of all the strains yielded the expected tdc amplicon and, none gave a positive result for detection of hdc gene. These results indicate a high incidence of tyramine producer Enterococcus strains in cheeses from Europe and from Argentina. Commercial starter cultures (7 Lactococcus lactis , 8 Streptococcus thermophilus , 3 Lactobacillus helveticus, 1 Brevibacterium casei and 1 undefined mesophilic) were also tested by colorimetric and ELISA methods for their ability to decarboxylate histidine. All the strains were found to be negative in terms of their ability to decarboxylate histidine in the conditions assayed. 63 Lb. hilgardii strains from different kind of wine were also analysed by colorimetric methods for their ability to produce tyramine and the results confirmed by PCR. 22 strains (34 %) were positive. Interestingly, these strains do not originate from the same type of wine. In contrast, they are disseminated in all types of wines (white, red, dry, sweet, with high ethanol level, from different regions). Moreover, any wine can contain tyramine producing and non-producing strains. The qPCR assay developed in this project was used to analyse 264 wine samples collected in wineries of Bordeaux s area during winemaking at the end of the malolactic fermentation. The results showed that 98% of the wines contained histamine-producing bacteria; up to a concentration of 5.106/ml. Histamine producers were present at concentrations above 1000 cells/ml in more than 72% of the samples. The population of histamine-producing bacteria can be more or less important at different locations. However, there are places where the wines contain very variable quantities of histamine-producing bacteria. Consequently, it is not certain that their distribution is related to specific geographical sites. Differences of wine composition and winemaking practices have most likely an important impact on the content in histamine-producing cells.

Histamine is the biogenic amine the most frequently involved in food intoxications. It is produced in fermented beverages and foods by strains of lactic acid bacteria (LAB) that belong to diverse species. All these strains share a similar enzyme, the histidine decarboxylase (HDC) catalysing the conversion of histidine to histamine. Genes coding for the HDC of some LAB were identified during the past two decades. However, it was not known if they were associated with other genes. We have sequenced a 5732 bp DNA region surrounding the HDC gene of the wine LAB Lactobacillus hilgardii IOEB 0006. Four complete open reading frames were detected in this sequence. They code for the HDC, a protein of unknown function, a putative histidyl-tRNA synthetase and a membrane transporter. A biochemical characterisation of the transporter was performed in the Laboratory of Molecular Microbiology of the University of Groningen. It was demonstrated that this enzyme is a histidine/histamine exchanger. The combination of the HDC and the histidine/histamine exchanger forms a typical amino acid decarboxylation system involved in metabolic energy production and/or acid stress resistance. The role of the two other proteins is unknown but it is supposed that they play a role in the histamine-producing pathway. Further works have shown that the four genes of this pathway are located on a large plasmid of L. hilgardii. This plasmid was unstable since it could be lost by LAB grown in specific culture media. To determine if this histamine-producing pathway was conserved in other LAB of wine and fermented foods, we have analysed five other LAB strains producing histamine and belonging to different species: Oenococcus oeni IOEB 9204 from wine, Lactobacillus 30a from horse digestive tract, Lactobacillus buchneri DSM 5987 from cheese, Lactobacillus sakei LTH 2076 from sauerkraut and Tetragenococcus muriaticus LMG 18498 from squid liver sauce. A similar four-gene cluster was detected in all the strains except in L. 30a in which the gene coding for the histidyl-tRNA synthetase was apparently missing. As found in L. hilgardii, the genes were located on a plasmid in L. sakei, T. muriaticus and O. oeni. In contrast, they were detected on the chromosome in L. buchneri and L. 30a. These differences of genetic locations may be important for the stability and the dissemination of the histamine-producing pathway. Sequencing of the histamine decarboxylase cluster of Lactobacillus buchneri B301. To characterize the genes responsible for histamine synthesis in L. buchneri B301, a PCR reaction was performed using the oligonucleotides JV16HC and JV17HC (le Jeune et al., 1995). The sequence of the obtained amplicon was similar to those of hdcA genes found in databases; it may therefore encode HdcA. This 0.3 kb internal hdcA fragment was located in the L. buchneri chromosome. A reverse PCR strategy was designed to reveal the complete sequence of hdcA gene and the flanking regions. A 5,775 bp DNA fragment was sequenced. The sequence analysis showed the presence of four complete ORFs: hdcA. Analysis of the amino acid sequence showed the protein can be included in the group of bacterial histidine decarboxylases that use a covalently bound pyruvoyl residue as a prosthetic group. hdcB. Downstream of hdcA, a second ORF was found in the same orientation. Comparisons of the deduced amino acid sequence with those present in databases revealed similarities of 89 %, 81 % and 69 % with HdcB of Tetragenococcus muriaticus, O. oeni, and Lactobacillus 30A, respectively. Some authors have associated this protein with the regulation of the operon or transport. However, the protein appears to have no transmembrane motifs. hisS. Downstream of hdcB, another ORF was found in the same DNA strand. The deduced protein showed similarity to the class II histidyl-tRNA synthetases. As in other aminoacyl-tRNA synthetases, the region upstream of the start codon of hisS contains a putative promoter region and a leader region with the sequence features of the tRNA-mediated anti-termination system. hdcC. Upstream of hdcA an additional ORF was identified in the same orientation; this was designated hdcC. Analyses of HdcC hydropathy predicted 11 hydrophobic segments long enough to form a transmembrane helix. To determine whether the product of the hdcC gene was a membrane protein, as suggested by sequence analysis, it was expressed in Lc. lactis NZ9000 under the control of the nisin promoter. The results allow the conclusion that the protein encoded by hdcC is located in the membrane. This result, together with the sequence data, suggests that hdcC could be responsible for histamine/histidine exchange in L. buchneri. To our knowledge, this is the first time such a gene has been described.

Lactic Acid Bacteria (LAB) are essential in the Dairy Industry as starters; nevertheless, the metabolic activity of some strains can produce toxic substances known as Biogenic Amines (BA). Ingestion of foods with high levels of BA cause serious upheavals, that could even jeopardize the life of the consumer, specially in those people with deficiencies in the intestinal amino-oxidases, the detoxification enzymes. The most frequent and concentrated BA in cheeses is tyramine. Tyramine is formed via decarboxylation of its corresponding amino acid, tyrosine, through the action of enzymes produced by microorganisms present in the food. The result summarised here is a fast, sensitive and easy method to detect, at any point of the food chain, tyramine-producing LAB. The production of this BA is more related to strain than to species and the producer bacteria can be either contaminants of fermented foods or constitute part of starter cultures. Therefore, it is very important to determine which strains produce this undesirable compound, to avoid that they could be used in the starter cultures. Although several qualitative and quantitative methods have been developed to determine BA production, demands of consumers for better and healthier foods have resulted in an increased interest in the development of rapid and sensitive methods for detecting BA-producing microorganisms in food. The present method allows fast and sensitive identification of tyramine-producing strains by PCR, by use of specific tdcA primers. The tdcA gene encodes the tyrosine descarboxylase, which catalyses the synthesis of tyramine. The comparison of the tdcA sequence from Enterococcus durans IPLA655 with those included in the databases, led to the design of specific primers for amplification of an internal fragment of the tdcA gene from different genera, allowing the identification of any tyramine-producing LAB. Since, it is unnecessary to purify the DNA of the strains to be tested (an isolated colony can be used as a template) these new method is fast and easy. Moreover, the proposed PCR technique offers the possibility of detecting potential tyramine-producing strains in milk, at any point of milk fermentation, and in final products such as cheese.

Tyramine is a biogenic amine (BA) frequently detected in wines and responsible of health diseases. The early detection of tyramine-producing lactic acid bacteria (LAB) could make it possible to prevent tyramine production in wine. This is of great interest not only for public health but also from an economic point of view, because products that exceed recommended limits of BA can be rejected from the market. Several methods to detect tyramine-producing LAB have been described. The most reliable one consists in growing LAB in an appropriated medium containing the precursor amino acids and, after the required time of growth, tyramine production is determined by HPLC. However, this method is not easy to use, requires expensive equipment and necessitates several days to be performed. Identification of the tyramine-producing pathways in Lactobacillus brevis IOEB 9809 and Carnobacterium divergens AN 508, and comparison with the known sequences of Enterococcus faecalis and Enterococcus faecium allowed us to design PCR primers, named TD2 and TD5, located in the tyrDC gene. The TD2-TD5 primer set allowed for the amplification of a 1100-bp DNA fragment in all LAB species above-mentioned. We have screened 209 LAB strains of the IOEB collection with these primers in order to determine if they can detect tyramine-producing strains of wine. The ability of each strain to produce histamine was confirmed by HPLC. The results obtained by PCR and HPLC were identical. A total of 37 positive strains were detected. They belong to species as diverse as L. brevis, L. hilgardii, L. casei, L. fructivorans and Pediococcus parvulus. This PCR assay allows early detection of tyramine-producing bacteria during the manufacture of wine or other foods and also it permits the selection of commercial starter strains unable to synthesise tyramine. However, this assay is limited by the impossibility to quantify the population of tyramine-producing bacteria present in a sample. To circumvent this limitation it would be interesting to develop a similar assay based on real time quantitative PCR.

Histamine is the most frequently biogenic amine (BA) involved in food intoxications. It is usually present at a low concentration in wine, but its toxic effect is strengthened by the presence of alcohol and other BAs of wine. Although there is no rule, some countries recommend upper limits of histamine ranging from 2 to 10mg per litre. Histamine is produced by lactic acid bacteria (LAB) containing a specific enzyme, a histidine decarboxylase. Genes encoding the histidine decarboxylase of several LAB were described recently. Based on these gene sequences we have designed PCR primers which are useful for detecting histamine-producing LAB by quantitative PCR. In addition, we have optimised an existing procedure for the extraction of microbial DNA from wine samples. The combination of both, the DNA extraction procedure and the quantitative PCR, allows for the detection and quantification of histamine-producing LAB in wine. Results obtained by this method were confirmed by other methods: plate counting, Southern blotting, quantitative PCR specific for another LAB gene, and HPLC determination of produced histamine. This method appeared to be very accurate and sensitive, since it can detected as few as 2 histamine-producing cells per millilitre of wine and the quantification is precise up to 2x10e7 histamine-producing cells per millilitre. This method requires less than 8 hours to be performed and it can be carried out for many samples simultaneously. It is useful to detect histamine-producing LAB in must and wine at all stages of the winemaking in order to avoid histamine spoilage of wine.

Putrescine (1,4- diaminobutane) is generally the most abundant biogenic amine (BA) present in wine. It is undesirable because it confers the odour of bad breath, rotten meat. Bacteria produce putrescine from ornithine via an ornithine decarboxylase (ODC), or from agmatine via the reactions of an agmatine deiminase (AgDI) and a putrescine transcarbamylase (PTC). The ODC enzyme and its odc coding gene are well characterised. They were described in several bacteria including lactic acid bacteria (LAB) of wine. In contrast, bacterial AgDI and PTC and their corresponding genes were recently described for the first time, and they were only detected in two strains of Pseudomonas aeruginosa and Streptococcus mutans. The genes of both bacteria differ significantly in terms of sequence and arrangement. During the sequencing of the genes of the tyramine-producing pathway of Lb. brevis IOEB 9809 we have identified the 5 extremity of an extra gene sharing similarities with genes present in the genome sequences of Pediococcus pentosaceus and Listeria monocytogenes and with the PTC gene of S. mutans, suggesting that it was a PTC gene. In P. pentosaceus and L. monocytogenes the putative PTC gene belongs to a six-gene cluster coding for enzymes potentially involved in the production of putrescine from agmatine. To determine whether L. brevis 9809 contains the same gene cluster we have sequenced the DNA region located downstream of its putative PTC gene. We have obtained the sequence of six genes similar as those detected in P. pentosaceus and L. monocytogenes. This sequence differs from the S. mutans operon, because it contains an extra AgDI gene and the gene of a transcriptional regulator located at the 3 extremity. We have cloned, expressed, purified and characterised the two putative AgDI and a membrane transporter. Results confirmed that one of the putative AgDI was actually an AgDI, which converts agmatine to carbamoyl-putrescine, whereas the second enzyme was inactive under our production conditions. The membrane transporter was studied by the Laboratory of Molecular Microbiology of the University of Groningen. It proved to catalyse the exchange of agmatine and putrescine. We concluded that the six-gene cluster detected in L. brevis and also present in P. pentosaceus and L. monocytogenes is an agmatine deiminase pathway involved in putrescine production. It is related to the AgDI pathway detected in S. mutans, but it differs in terms of enzymes encoded and gene arrangements. This pathway could be widely distributed amongst LAB and contribute to the presence of putrescine in wine.

One important problem for genetic manipulation of natural strains of lactic acid bacteria (LAB) is that some of them are refringent to electrophoration. An alternative DNA transfer system is conjugation. Utilisation of this type of horizontal transfer from gram-positive and gram-negative bacteria to LAB has been investigated during the development of this project. Methods for plasmid transfer by mobilization to Lactobacillus, Lactococcus lactis and Oenococcus oeni have been standardised. Moreover, transfer by one of the methods of pMV158GFP plasmid has allowed standardisation of fluorescent detection of L. lactis in cheese matrix. In addition, a method for transfer by electrophoration and detection of the pCIT264 plasmid encoding the citrate transport system has been developed. The method is based on the acid resistance conferred by the citrate transport and metabolism. The new recombinant strain CRL30[pCIT264] has been characterised and the results revealed that indeed is a new L. lactis food grade-strain expressing the citrate transport system. Furthermore, in this study there have been developed and/or validated expression vectors. A food-grade expression system has been developed for Lactobacillus casei. A food-grade strain with nisRK stably integrated into the genome was constructed, in order to implement the nisin-controlled expression system (NICE) in this bacteria. Expression of b-glucuronidase (gus) reporter gene was employed to optimize the system, which has been successfully used to produce the main antigenic protein from Norwalk virus. Moreover, the studies performed in this study have shown that the pLS1RGFP vector is suitable for gene inducible overexpression in L. lactis. The usage of pLS1RGFP in L. lactis was standardised, and its utililisation for controlled expression of housekeeping enzymes validated by cloning of the lactococcal rnc gene in it, and characterization of the RNase III encoded by the constructed recombinant plasmid. Finally, the vector has been validated for expression of genes of industrial interest. The gtf of the Pediococcus parvulus 2.6, which synthesises an exopolysaccharide with prebiotic properties, was cloned in the vector under the control of the PM promoter. Immunological and enzymatic studies revealed that overexpression of the gene conferred to Gram-positive bacteria, including L. lactis, the ability to synthesise and secrete the exopolysaccharide. In addition, they demonstrated that indeed the gtf gene product is a glycosyltransferase bound to the cytoplasmic membrane.

Biogenic amines (BA) are the product of decarboxylation pathways in food bacteria and a major cause of food poisoning. Substrates of decarboxylation pathways are amino acids that are converted into the corresponding amines or amino acids (e.g. histidine/histamine, aspartate/alanine, tyrosine/tyramine), but also di/tricarboxylates that are converted into monocarboxylates (e.g. malate/lactate, citrate/lactate). The latter are beneficial in fermentation processes. The pathways consist of a transporter that catalyses the translocation of the substrate into the cell coupled to the secretion of the metabolic end product of the same substrate out of the cell, and a decarboxylase. In most cases the end product is the direct decarboxylation product of the substrate and, therefore, both are structurally related. The transporters are at the gate of the pathways and any unwanted activity is best attacked there. We aim to develop inhibitors of BA production for food preservatives, and to engineer the transporters to prevent inhibition by naturally occurring compounds of the beneficial citrate and malate pathways. Both goals would benefit from a detailed structural model of the binding site of the transporter. The principle of binding both substrate and end-product was studied by determining the global structure of the transporters and by identifying and characterising the binding site of the citrate and malate transporters MleP and CitP. These studies involve the membrane topology, domain structure and the residues on the protein that co-ordinate the different groups of the substrates. The studies made use of bioinformatics approaches in analyzing the gene families to which the transporters belong and of experimental approaches to identify residues in the binding site and membrane topologies. The citrate and malate transporters are members of the 2-hydroxycarboxylate transporter (2HCT) family. The structural model of the proteins of the 2HCT family shows two homologous domains consisting of 5 transmembrane segments (TMS) each, and separated by a large hydrophilic loop that resides in the cytoplasm. An additional TMS is present at the N-terminus of the protein. This domain is not part of the two-domain structure. A major consequence of the odd number of TMSs in the two domains is that they have opposite orientations in the membrane. The N-terminus of the N-terminal domain is in the periplasm, while the N-terminus of the C-terminal domain is in the cytoplasm. Homologous domains with inverted topologies turn out to be a common structural motif in membrane proteins. The corresponding loops between the fourth and fifth TMS in each domain fold into pore-loop structures. Both regions contain an extraordinarily high fraction of residues with small side chains, which may reflect a compact packing of the loops between the transmembrane segments. A sequence motif GGxG is located at the centre of the pore-loop regions and may represent the vertex of the loops. The loop between TMS VIII and IX in the C-terminal domain folds into an amphipathic surface helix. This feature is not observed in the N-terminal domain. The pore-loop structures in the two domains enter the protein from opposite sites of the membrane and are likely to contact each other in the three dimensional structure where they form the substrate binding site and the translocation site. Close to this structural arrangement, an arginine residues in TMS XI in the C-terminal domain is highly conserved in the family of proteins. The residue is interacting with one of the carboxylate groups on the substrates and places the substrate binding site at the cytoplasm/membrane interface. Interaction between a second carboxylate group and the hydroxyl group on the substrates are essential for binding, but the interacting residues on the proteins are still elusive.

Decarboxylation of metabolites is a desirable step in dairy, meat and wine industries, but the conversion of amino acids in biogenic amines results in the production of unwanted toxic products. The pathways consist of a membrane transporter and a cytosolic decarboxylase. The benefit of the pathway for the bacteria could be the generation of metabolic energy and/or homeostasis of intracellular pH. Regulation of the intracellular pH is a crucial parameter for the cellular functioning in low pH conditions. We aim to understand the mechanisms behind the energy generation and resistances provided by decarboxylation pathway. Consequently, due to their participation in the homoeostasis of intracellular pH, the decarboxylation pathways of malic and citric acids are of interest. In these study, we analysed the effects of L-malic acid and citric acid on tolerance to acidic conditions and on regulation of intracellular pH in Oenococcus oeni. Bacterial cultures growth with addition of various organic acids and pH conditions revealed that L-malate enhanced on the growth at pH equal or below 4.5. On the other hand, the presence of citrate in the medium led to complete inhibition of the growth at pH 3.2. Metabolism of hexoses, but not of organic acids was impaired by low pH. Regulation of intracellular pH revealed that both L-malic and citric acids participated in the enhancement of the transmembrane pH gradient. For growth performed at pH 5.3 or 4.5, maintenance of a pH potential above 1 unit due to L-malate metabolism was dependent on the presence of either L-malate or citrate during growth. Interestingly, at pH 3.2, only cells grown with L-malate produced a pH potential over 1.5 unit whatever the buffer used. Finally, study of the expression of genes involved in the metabolism of organic acids showed that the presence of L-malate in the medium increased the amount of mleP mRNA at pH 4.5 and 3.2, whereas the expression of citrate genes was neither affected by pH nor by the presence of organic acids. A 20 bp chromosomal deletion within the mae gene was achieved in L. lactis through a two step homologous recombination process. Growth of parental and mutant strains was performed with or without citrate. The mutant remained able to metabolise the citrate. Acetate and diaceyl/acetoin production observed during growth showed that the mutant was still able to convert citrate into oxaloacetate and acetate. Results suggest that oxaloacetate was decarboxylated to pyruvate at least for a part, although the deletion mutant lacks oxaloacetate decarboxylase encoded by mae. In contrast, a delay in the lag phase was observed in M17 medium during growth of the mutant with citrate and its growth rate was at least 20% lower compared to the growth rate of the parental strain. This suggested that generation of ATP was affected in the mutant strain. One hypothesis would be that the deletion of mae would hinder the ability to generate the delta-pH in the presence of citrate. This hypothesis was analysed by comparing the capacity of parental and mutant cells to generated transmembrane proton gradient in presence of citrate. The behaviour of the parental and mutant strains was different. Our result revealed that the intracellular pH of the mutant was identical in the presence or absence of citrate. These results suggest that the decarboxylation reaction play a role by consuming a proton in the cytoplasm, which results in a proton gradient over the cytoplasmic membrane.

Histamine is the most frequently biogenic amine (BA) involved in food intoxications. It is usually present at a low concentration in wine, but its toxic effect is strengthened by the presence of alcohol and others BA of wine. Although there is no rule, some countries recommended upper limits of histamine ranging from 2 to 10 mg per litre. Histamine is produced by lactic acid bacteria (LAB) containing a histidine decarboxylase and the other enzymes of the histamine-producing pathway. This pathway is present in LAB of diverse species, but only in some strains in each species. It was detected in a few strains of the species Oenococcus oeni, which is required for the manufacture of wines. A method based on real time quantitative PCR has been developed to detect and quantify the population of histamine-producing LAB present in wine. We have used this method to determine the populations of histamine-producing LAB in diverse wines of the region of Bordeaux. A total of 264 wines were collected in wineries at the end of the malolactic fermentation during the production of the vintage 2005. Analyses indicated that 98% of the wines contained populations of histamine-producing bacteria ranging from 1 to 5.10e6 cells per millilitre. Only 2% of the wines were apparently devoid of histamine-producers. The concentrations of histamine were determined in some of the wines. It was found that significant concentrations of histamine (above 2mg/l) were present when the population of histamine-producing LAB was above 1000 cells per ml. This level was reached or exceeded in 72% of the analysed wines. The results were plotted on a map of Bordeauxs area to determine if there was a correlation between the measured populations of histamine producing LAB and the geographical origin of the wines. This map suggests that the population of histamine-producing bacteria can be more or less important from place to place. However some places contained wines with highly variable populations of histamine-producing bacteria. We concluded that wines of some places could actually be easily spoiled by histamine-producing LAB but that parameters unrelated to the geography were also important for determining the final levels of histamine-producing LAB. Differences of wine composition and winemaking practices could be involved.

Citrate is present in milk at concentrations of 8 to 9 mM and is cometabolised with sugars by strains of lactic acid bacteria (LAB), including Weisella paramesenteroides. The breakdown of citrate results in production of carbon dioxide (responsible for the texture of some cheeses) and production of the flavour compound diacetyl, which is essential for the quality of dairy products such as butter, buttermilk, and cottage cheese. Therefore, to improve and control production of aroma compounds by different LAB, it is important to understand the molecular mechanism which regulate the citrate metabolic pathways. The dairy Weissella paramesenteroides citrate fermenting J1 strain isolated from an Argentinean cheese has been the subject of this study. We have previously characterised its citMCDEFGRP (cit) operon, encoding the citrate permease and the two enzymes (citrate lyase and oxaloacetate decarboxylase), which catalyse conversion of citrate into pyruvate. The expression of both the cit operon and the upstream divergent gene citI are induced by the presence of citrate in the medium. In this work DNA protein interaction studies as well as in vitro transcription reveal that CitI, encoded by citI, recognises two A+T rich operator sites located between citI and citM and that the DNA-binding affinity of CitI is increased by citrate. Subsequently, this citrate-signal-propagation leads to activation of the cit operon through an enhanced recruitment of RNA polymerase (RNAP) to its promoters. Our results indicate that control of CitI by the cellular pools of citrate provides a mechanism for sensing the availability of citrate and adjust the expression of the cit operon accordingly. The results presented in this work suggest a scenario for the transcriptional activation of the Pcit and PcitI promoters in W. paramesenteroides. In the absence of citrate, CitI would bind to the operator sites stimulating the RNA polymerase to form complexes on both promoters. This would result in low expression levels of the citrate fermentation enzymes as well as of the citrate permease P, which will catalyse the uptake of citrate when this compound becomes available in the environment. Once transported inside the cell, citrate would bind to CitI enhancing the regulator affinity for its DNA operators and resulting in increased RNAP recruitment at Pcit and PcitI. This would result in transcriptional activation from both promoters. Activation of Pcit results in coordinated synthesis of the citrate fermentation enzymes and breakdown of citrate. Thus, our model predicts that a small increase in CitI transcription could account for the large increase of the cit mRNA detected in W. paramesenteroides growing in a citrate-supplemented medium. In W. paramesenteroides, Leuconostoc mesenteroides, Oenococcus oeni and Lactobacillus plantarum, citI is placed in opposite orientation with respect to the cit operon. In addition, in these bacteria two consensus CitI binding operators could be found in the intergenic region between citI and citM. These facts suggest a similar mechanism of transcriptional activation by CitI in conjunction with citrate in all these bacteria. Thus, CitI seems to play a pivotal role in citrate sensing and in the transcriptional stimulation of the operons involved in the transport and metabolism of citrate in LAB.

Biogenic amines (BA) are a major cause of food poisoning. Their appearance in foods is the result of unwanted microbiological activity. Understanding the metabolic pathways and enzymes involved in BA production by LAB is essential for the understanding of the mechanism behind the spoilage of foods and beverages. BA are the end product of metabolic pathways that provide the microorganisms with metabolic energy, mostly in the form of proton motive force. The pathways consist of a transport protein embedded in the cytoplasmic membrane and a metabolic enzyme that converts an amino acid into the corresponding BA by a decarboxylation reaction. The transporters are responsible for both the uptake of the precursor (the amino acid) and the excretion of the BA in the cell. The transporters are at the gate of the pathways and any unwanted activity of the fermenting organisms is best attacked there. We have screened a set of compounds for their ability to inhibit tyramine formation by Lactobacillus brevis by studying the substrate specificity of the tyorisine/tyramine exchanger TyrP. The tyrosine/tyramine exchanger TyrP was expressed in the lactic acid bacteria Lactococcus lactis under control of the inducible nisin promoter. The structural genes coding for the transporter was cloned in vector pNZ8048 yielding plasmids pNZtyrP, which encode the exchangers with an additional N-terminal His-tag that was added for expression studies. The plasmids were transformed to L. lactis NZ9000 strain and the substrate specificity of TyrP was studied in membrane vesicles with a right-side-out orientation using heterologous exchange and inhibition studies. BA are the direct decarboxylation product of amino acid, and, consequently, the transporters catalyse structurally related compounds. A set of 19 structural analogues of tyrosine modified at different parts of the molecule were assayed for their ability to compete with tyrosine in the TyrP transporter activities. The set included hydroxyphenyl propionic acid, hydroxyphenyl pyruvic acid, and hydroxyphenyl lactic acid there were modified at the C2 amino group, tyramine, L-tyrosinol, L-tyrosine methyl ester, and L-tyrosine hydrazide modifed at the C2 carboxylate group, D,L-a-methyl tyrosine modified at the C2 hydrogen atom, L-phenyl alanine, fluoro-L-phenyl alanine, chloro-L-phenylalanine,bromo-L-phenyl alanine, and iodo-L-phenyl alanine modifed at the para-hydroxyl group, and meta-DL-tyrosine, ortho-DL-tyrosine, and DL-dihydroxy-phenylalanine (DOPA) that are modified at the ring. Modifications of the C2 carboxylate group that is removed in the decarboxylation reaction were well tolerated by TyrP. The kinetics with the two physiological substrates L-tyrosine and its decarboxylation product tyramine were comparable. Analysis of the substrate specificity of TyrP in the exchange reaction demonstrates that the amino group and the phenyl ring with the para hydroxyl group provide the most important interaction sites of the molecules with the protein. All modifications of these groups resulted in at least a 10-fold reduction of the exchange rate. Remarkably, for many of the modifications the effect on the exchange rate was stronger than on the inhibition of counterflow, suggesting that translocation involves more critical interactions between substrate and protein than the initial binding step. A potent inhibitor of tyramine production by the tyrosine decarboxylation pathway of Lactobacillus brevis should bind with high affinity to the transporter and, preferentially, should translocate the compound not or only slowly. Heterologous exchange assays predominantly for the ability to translocate, while the inhibition of counterflow assay measures affinity. Therefore, a good inhibitor is poorly active in exchange and highly potent in inhibiting counterflow. With L-phenylalanine as the substrate, TyrP had a translocation activity that was far less than 10% under conditions where counterflow activity was inhibited by >95%. L-phenylalanine is one of the naturally occurring amino acids in foods. Adding additional L-phenyl alanine during the fermentations could prevent the formation of the BA tyramine.

An important bottleneck in the study of BA producing pathways at the beginning of the project was that (excepted hdcA) the structural genes coding for the decarboxilases and the transporters had not been identified. Cloning, molecular analysis, and genetic engineering of the genes involved in BA production in wine and fermented milk was performed to achieve the objectives. Sequencing of the tyrosine decarboxylase cluster of Enterococcus durans IPLA 655. Two degenerate primers were designed, based on the partial sequence of the tyrosine decarboxylase protein from Lb. brevis IOEB 9809. These were used to test E. durans IPLA 655 by PCR for the presence of the tyrosine decarboxilase gene. The obtained 820 bp amplicon was cloned and sequenced. The sequence analysis revealed an ORF which showed strong similarity to the tyrosine decarboxylase genes of E. faecalis JH2-2 (83% identity) and Lb. brevis IOEB 9809 (73% identity). This gene was designated tdcA. Using reverse PCR, the E. durans genome was walked on either side of this fragment. Upstream of tdcA, a complete ORF was found that shared 78% and 72% identity respectively with the tyrosyl t-RNA syntethases of E. faecalis JH2-2 and Lb. brevis IOEB 9809. This gene was designated tyrS. The analysis of its sequence revealed that the encoded tyrosyl-tRNA synthetase (TyrS) belongs to the class I aminoacyl-tRNA synthetases, characterized by HIGH and KMSKS motifs. The HIGH motif is perfectly conserved in E. durans IPLA 655 TyrS and the KMSKS motif is represented by the KFGKT sequence, as in E. coli, Bacillus subtilis, and E. faecalis. Downstream of tdcA, another ORF (tyrP) was found that shared 84% and 66% identity respectively with the antiporters of E. faecalis JH2-2 and Lb. brevis IOEB 9809. The deduced amino acid sequence of this protein (determined using the Sosui program) revealed a membrane protein structure with 11 transmembrane helices. This protein may be involved in tyrosine/tyramine exchange. Sequencing of the tyramine producing pathway of the wine LAB Lactobacillus brevis IOEB 9809:A strategy based on linker-mediated PCR was employed to sequence the L. brevis IOEB 9809 genomic sequence surrounding the 792 bp-fragment of tyrDC determined in a previous work. A 7979 bp DNA sequence was determined. It contains four complete and one partial open reading frames (ORFs) encoding polypeptides larger than 100 amino acids. The tyrDC gene codes for a protein of 626 amino-acids with a calculated molecular weight of 70.5 kDa in agreement with experimental results. The upstream ORF showed strong similarities with genes of tyrosyl-tRNA synthetase (tyrRS). Downstream of tyrDC were two ORFs related to genes of amino-acid permeases (tyrP) and Na+/H+ antiporters (nhaC), respectively, and the 5 -end of a fifth ORF similar to ornithine transcarbamylase genes (otc). Sequencing of the tyramine producing pathway of food LAB Carnobacterium divergens AN 508: To determine the presence of a tyrDC gene amonst known tyramine producing carnobacteria, PCRs were carried out using the degenerated primers P1-rev and P2-for (Lucas and Lonvaud-Funel, 2002). An amplification product of ~ 800 base pairs was obtained with the strains C. divergens AN 508 and C. piscicola AN 545. The PCR product for the two carnobacteria were sequenced (78.7% sequence similarity). They showed also high similarities with the published sequenced of Lb. brevis IOEB 9809 and Enterococcus faecalis. The complete tyrDC gene sequence of C. divergens was determined by use of inverse and conventional PCR. A 2.2kb PCR product was obtained and sequenced. It was located at the 3-end of the previously sequenced fragment of the tyrDC gene. To obtain the 5-end of this gene, a degenerated primer was designed on the basis of a comparison of the sequence of tyrRS genes identified in E. faecalis and other LAB because sequencing of the tyrDC cluster of E. faecalis had previously suggested the presence a tyrRS gene upstream of tyrDC . A 900-bp DNA fragment was successfully amplified using this strategy. After sequencing, it was confirmed that it contained the 5-end of the tyrDC gene and a fragment of a tyrRS gene. The complete sequence of the tyrDC cluster determined in C. divergens AN 508 comprises 3427 bp including the three genes tyrRS, tyrDC and tyrP.

Citrate present in milk is cometabolised with sugars by strains of lactic acid bacteria (LAB), including Lactococcus lactis biovar diacetylactis (L. diacetylactis). The breakdown of citrate results in production of carbon dioxide (responsible for the texture of some cheeses) and production of the flavour compound diacetyl, which is essential for the quality of dairy products such as butter, buttermilk, and cottage cheese. In addition, citrate metabolism results in synthesis of lactic acid, which has an important role in the hygienic quality of fermented food, since impairs the growth of pathogens. Therefore, a bacterial strain, able to produce aroma compounds and at the same time been acid producer and acid resitant, will have a very good potential as starter for dairy fermentations. This work focused on the study of citrate metabolism by L. diacetylactis CRL264 isolated from artisanal argentinean cheese and its contribution to pH homeostasis and acid stress resistance. We used tightly controlled pH conditions and studied the effect of citrate metabolism on the growth parameters. This approach allowed assessing the cellular response to acid pH independently of the pH changes concomitant with the growth of batch cultures. The toxic effect of the organic lactic and acetic acids accumulated during the fermentation processes was also examined. The CRL264 strain harboring pCIT264 (which encodes the citrate permease P), and its derivatives CRL30 (CRL264 cured of pCIT264) and CRL30pCIT (CRL30 transformed with pCIT264) were studied. The growth characteristics of these strains on glucose with/without citrate were compared. Under controlled pH, growth on glucose plus citrate is characterised by higher specific growth rate but lower biomass production; at the onset of citrate exhaustion, the specific growth rate decreased 7-fold. Acetic acid was the primary cause for this decrease upon citrate depletion and, also for the decrease in biomass when higher concentrations of citrate were catabolised. In vivo 13C-NMR confirm that during co-metabolism all acetate formed is derived from citrate, and glucose is converted to lactate. Addition of arginine combined with citrate led to a remarkable enhancement of the final biomass, showing the beneficial effect of the extra ATP derived from arginine to counteract the deleterious effects of weak acids on the magnitude of the proton motive force. In line with this hypothesis, lower ATP levels were measured by 31P-NMR in cells metabolizing glucose plus citrate as compared to glucose alone. Given the crucial influence of citrate metabolism on the performance of CRL264 at low pH, its capacity for citrate uptake was analysed. Transport assays showed that in the parental strain CRL264 the citrate transport is 2.7 fold faster than in its derivative CRL30pCIT, correlating with a 2 fold higher copy number of pCIT264 than pCIT. This enhanced efficiency in the transport correlates with the better performance of CRL264 at acidic pH. Based on these results, we propose that increasing the transport capacity of CRL264 could be a useful strategy to improve furthers the performance of this strain under acidic conditions, a trait remarkably important in industrial fermentations. NMR allows distinguishing between intra and extracellular lactate. In this work, we obtained in vivo 13C-NMR data that showed a decrease of the intracellular lactate pool when a pulse of citrate was given to the cells. These results suggest that the transport system operating in CRL264 strain is a citrate/lactate exchanger system, which provides the cells with an advantage against lactate toxicity. The oxaloacetate decarboxylase activity results essential in the alcalinisation of the cytosol during citrate fermentation in L. diacetylactis. In this work, we demonstrated that citM gene from L. lactis CRL264 encodes an oxaloacetate decarboxylase. The enzyme exhibits high levels of similarity to malic enzymes from other organisms. CitM was purified and its oxaloacetate decarboxylase activity demonstrated by biochemical and genetic studies. The highest oxaloacetate decarboxylation activity was found at low pH in the presence of manganese, and the Km value for oxaloacetate was 0.52 +/- 0.03mM. In addition, in this work the citM-CDEFXG genes encoding the subunits of the citrate lyase of CRL264 have been characterised. The genes are transcribed as a single polycistronic mRNA of 8.6kb, and their expression is induced at low pH. This response to acid stress takes place at the transcriptional level and result in an increase of enzyme levels, demonstrated by proteomic analysis, and correlates with increased activity of citrate lyase. It is suggested that coordinated induction of the citrate transporter, CitP, and citrate lyase by acid stress provides a mechanism to make the cells more resistant to the inhibitory effects of the lactate fermentation product, which accumulates under these conditions.

Yoghurt was manufactured using the following culture combinations; Lactobacillus bulgaricus LB12 + Streptococcus thermophilus CHCC6483 (histamine positive); Lb. bulgaricus LB12 + S. thermophilus CHCC1524 (histamine positive); Lb. bulgaricus LB12 + S. thermophilus CHCC3050 (histamine negative). Immediately after fermentation (Day 1), and upon storage at 4°C (Day 20 and Day 50), samples were withdrawn and prepared for histamine determination by HPLC. Histamine was found to be present in the yoghurt, which contained Lb. bulgaricus LB12 together with either of the histamine positive S. thermophilus strains (CHCC6483 or CHCC1524); no histamine was detected for yoghurt manufactured with the histamine negative S. thermophilus strains (CHCC3050). Presumably, Lb. bulgaricus LB12 produces histidine, which is subsequently decarboxylated to histamine by the positive S. thermophilus strains. It was also found that the majority of the histamine produced was formed during the yoghurt fermentation itself, and only a minimal increase in histamine was observed upon storage of the yoghurts at 4ºC for up to 50 days. Based on this study it was concluded that the yoghurt manufactured with the histamine positive S. thermophilus strains (CHCC6483 or CHCC1524) is unlikely to pose a significant risk due to the low levels of histamine detected (approximately 20 mg/l). Nevertheless, if a histamine positive strain is used for yoghurt production then the availability of histidine should be limited as much as possible.

Biogenic Amines (BA) are toxic substances that appear in foods and beverages as a result of amino acid decarboxylation. The enzyme histidine decarboxylase (HdcA) catalyses the decarboxylation of histidine to histamine, the BA most frequently involved in food poisoning. A real-time quantitative polymerase chain reaction (real-time qPCR) assay for the direct detection and quantification of histamine-producing strains in milk and cheese has been developed. A set of primers was designed, based on the histidine decarboxylase gene sequence of different Gram-positive bacteria. The results showed that the proposed procedure is a specific and highly sensitive technique for detecting potential histamine-producing strains in culture media, milk and curd. The method has been also optimised to quantify the presence of histamine-producing microorganisms in cheeses and during the cheese-making process. Chromatographic methods (HPLC) verified the capacity of real-time qPCR to correctly quantify histamine accumulation. Moreover, the entire process only takes about two hours, and 96 samples can be processed simultaneously. This could be very beneficial for the dairy industry, since microbiological methods require more than 24 h to be completed. In addition, the procedure allows the identification of potential histamine-producing LAB, which would be useful, when screening for starter strains is been performed. Another important advantage of this method is that it can be used at any point in the manufacturing process, even when histamine is still undetectable by other methods. In conclusion, the proposed method offers a rapid and simple way of characterizing the LABs in different types of dairy substrate. It could be employed to prevent the selection of histamine-producing strains as starter cultures. It is much faster than any microbiological procedure for the detection and quantification of HDC+ strains in milk, curd and cheese, and even allows an estimation of the histamine content of these substrates. Since HDC+ LAB of different origins were detected in this study, the proposed method might be of use with other types of fermented foods and drinks.

Rezultaty

Udostępnij tę stronę

Pobierz