This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Raynaud, S.
Right arrow Articles by Loubiere, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Raynaud, S.
Right arrow Articles by Loubiere, P.
Agricola
Right arrow Articles by Raynaud, S.
Right arrow Articles by Loubiere, P.

 Previous Article  |  Next Article 

Applied and Environmental Microbiology, December 2005, p. 8016-8023, Vol. 71, No. 12
0099-2240/05/$08.00+0     doi:10.1128/AEM.71.12.8016-8023.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Metabolic and Transcriptomic Adaptation of Lactococcus lactis subsp. lactis Biovar diacetylactis in Response to Autoacidification and Temperature Downshift in Skim Milk

Sandy Raynaud,1 Rémi Perrin,2 Muriel Cocaign-Bousquet,1 and Pascal Loubiere1*

Laboratoire Biotechnologie-Bioprocédés, UMR 5504 INSA/CNRS & UMR 792 INSA/INRA, Institut National des Sciences Appliquées, 135 Avenue de Rangueil, 31077 Toulouse cedex 4, France,1 Soredab, La Tremblaye, La Boissière Ecole, France2

Received 13 July 2005/ Accepted 27 August 2005


arrow
ABSTRACT
 
For the first time, a combined genome-wide transcriptome and metabolic analysis was performed with a dairy Lactococcus lactis subsp. lactis biovar diacetylactis strain under dynamic conditions similar to the conditions encountered during the cheese-making process. A culture was grown in skim milk in an anaerobic environment without pH regulation and with a controlled temperature downshift. Fermentation kinetics, as well as central metabolism enzyme activities, were determined throughout the culture. Based on the enzymatic analysis, a type of glycolytic control was postulated, which was shared by most of the enzymes during the growth phase; in particular, the phosphofructokinase and some enzymes of the phosphoglycerate pathway during the postacidification phase were implicated. These conclusions were reinforced by whole-genome transcriptomic data. First, limited enzyme activities relative to the carbon flux were measured for most of the glycolytic enzymes; second, transcripts and enzyme activities exhibited similar changes during the culture; and third, genes involved in alternative metabolic pathways derived from some glycolytic metabolites were induced just upstream of the postulated glycolytic bottlenecks, as a consequence of accumulation of these metabolites. Other transcriptional responses to autoacidification and a decrease in temperature were induced at the end of the growth phase and were partially maintained during the stationary phase. If specific responses to acid and cold stresses were identified, this exhaustive analysis also enabled induction of unexpected pathways to be shown.


arrow
INTRODUCTION
 
Lactic acid bacteria (LAB) are gram-positive microorganisms that have a real economic impact due to their use in many different food transformation processes. The bacterium Lactococcus lactis, which is commonly used as a starter bacterium in the manufacture of different dairy products, such as cheese, butter, and buttermilk, is generally considered a model for metabolism regulation, physiology, and genetic studies in LAB. In industrial conditions, as well as in natural ecological niches (plants, animals, and gastrointestinal tracts), different physicochemical stresses (including acid, thermic, oxidative, and osmotic stresses) are encountered, leading to suboptimal growth. Acid stress turns out to be the major stress that L. lactis is confronted with during fermentative processes. L. lactis, which is generally considered a homofermentative bacterium, is able to convert more than 90% of the sugar metabolized into lactic acid, whose accumulation is responsible for the progressive acidification of the medium, which leads to autoinhibition of lactococcal growth. This process, which is generally attributed to bacterial membrane permeability to the nondissociated form of lactic acid, increases when the medium is acidified and results in a decrease in the internal pH (12). This cytoplasmic acidification is known to be responsible for growth inhibition (11, 21) or a loss of cell viability (10, 14, 26, 30) due to limitation of the activity of biological reactions. Like lactic acid production, cold stress also plays an important role in a number of food processes through chilling (ripening) and product preservation, which is generally done at low temperatures. A decrease in the culture temperature can result in a decrease in the growth rate or growth arrest. Like acid stress, cold shock induces specific responses and nonspecific responses (biosynthesis of "stress proteins," for example) at the same time, which are implicated in resistance and adaptive processes required for maintenance of cell viability in such stress conditions.

The growth and metabolism of lactococci in milk, a complex culture medium, and the physicochemical modifications of the medium that occur during the process, mainly due to acidification, are responsible for the textural and organoleptic characteristics of cheeses. Understanding lactococcal metabolic behavior in milk during acid and cold conditions appears to be very important for better control and reproducibility of milk fermentations as a part of the cheese-making process.

The availability of the complete genome sequence of L. lactis subsp. lactis IL1403 (2), as well as number of lactococcal plasmids, enables metabolic and transcriptomic studies to be done. A global analysis that integrated transcript and enzymatic data into the cellular growth context of an L. lactis subsp. cremoris strain in a defined nutritional medium under autoacidification conditions was proposed by Even et al. (5), but this analysis was restricted to the central metabolism genes. More recently, the first dynamic genome-wide transcriptome analysis of a LAB was performed during carbon starvation of L. lactis IL1403 (32), which enabled all the gene expression changes due to starvation to be quantified.

An L. lactis subsp. lactis biovar diacetylactis strain of industrial origin exhibiting efficient growth in milk was used throughout this study. With this culture, which was grown in skim milk and during autoacidification and a controlled temperature downshift, growth, pH, substrates, and fermentation products were monitored, glycolytic enzyme activities were measured, and changes in the transcript profiles were determined.


arrow
MATERIALS AND METHODS
 
Organism and growth conditions.
The bacterium used throughout this work was an L. lactis subsp. lactis biovar diacetylactis dairy strain, strain LD61, which was provided by Soredab. This strain, which contains plasmids that allow optimal growth in milk (lactose, protease, and citrate utilization), was grown in non-heat-sterilized skim milk (Lait G; Standa Industrie) and in anaerobic conditions under a nitrogen atmosphere in a 20-liter fermentor (Setric Génie Industriel, Toulouse, France) with an agitation speed of 250 rpm. The temperature in the fermentor was maintained at 34°C until the culture pH reached 5.2 (after 8 h of growth) and then was slowly decreased to 12°C in approximately 10 h. The culture was inoculated with cells from precultures grown in a complex M17 broth (Difco) supplemented with lactose (50 g/liter) as the carbon source. Preculture cells were harvested during the exponential phase, washed in phosphate buffer (66 mM KH2PO4, 43 mM K2HPO4), and concentrated in order to obtain an initial optical density at 580 nm of 0.01 to 0.02 in the fermentor.

Fermentation analysis.
After transparization of the milk culture with a solution of EDTA (15 mM, pH 12.3), the cells were harvested, washed three times with a 0.9% NaCl solution, and resuspended before bacterial growth was estimated by spectrophotometric measurement at 580 nm (1 U of optical density at 580 nm was equivalent to 0.30 g [dry weight] per liter) as described by Mercade et al. (22). Precipitation of proteins in samples with a solution of barium hydroxide and 0.3 M zinc sulfate was performed before the concentrations of substrates (lactose and citrate) and fermentation products (lactate, formate, acetate, and ethanol) were measured by high-pressure liquid chromatography (3).

Preparation of crude extracts and enzyme assays.
A volume of culture corresponding to 115 mg (dry weight) cells was treated with 26.8 mM EDTA (pH 12), centrifuged (4°C, 10 min, 8,000 rpm), and washed three times with 0.2% (vol/vol) KCl (28). Cells were then resuspended in 5 ml of 45 mM Tris-15 mM carballylate buffer containing glycerol (20%), MgCl2 (4.5 mM), and dithiotreitol (1 mM). Cell disruption by sonication was followed by cell debris removal; the supernatant was used for all enzyme assays, and the protein concentrations of enzymatic extracts were determined by the method of Lowry et al. (18) with bovine serum albumin as the standard. All enzymes were assayed at 34°C and pH 7.2 immediately after cell disruption in the supernatant of the extract. The assays were based on coupling of the enzyme activity to consumption or production of NADH, which was monitored at 340 nm, as previously described by Even et al. (5). NADH oxidase activity was measured by a method derived from the method of Cocaign-Bousquet and Lindley (3) by using a reaction mixture containing Tris-HCl buffer (100 mM, pH 7.2), MnSO4 (5 mM), and NADH (0.6 mM). Specific enzyme activities (mmol · h–1 · g protein–1) were converted to whole-cell activities (mmol · h–1 · g [dry weight] of cells–1) using the measured protein content of L. lactis (42%).

ATPase extraction and assay.
The method used for extraction of ATPase was based on the method described by Drici-Cachon et al. (4), adapted to milk culture conditions (22). A volume of culture corresponding to 20 mg (dry weight) of cells was treated with EDTA (2 volumes, 26.8 mM, pH 12), centrifuged (4°C, 10 min, 6,000 x g), washed, and resuspended in MgCl2 (1 mM). The level of ATPase activity, based on ATP hydrolysis to ADP and Pi at 37°C, followed by an assay of the Pi formed, was estimated as previously described (6).

Extraction of total RNA, hybridization of labeled cDNA, and detection.
A volume of culture corresponding to 6 mg (dry weight) of cells was harvested, treated with EDTA (2 volumes, 26.8 mM, pH 12), centrifuged (4°C, 5 min, 8,000 x g), washed with 1 ml TE buffer (10 mM Tris-HCl [pH 8], 1 mM EDTA), resuspended in 500 µl TE buffer, and frozen in liquid nitrogen. Cells were stored at –80°C before extraction, and after cell disruption at 4°C (three 1-min cycles separated by 2-min cooling periods) with a mini bead beater (Biospec Products) with glass beads (0.6 g), 50 µl of 10% sodium dodecyl sulfate, and 500 µl of phenol (pH 4.7) and elimination of the cell debris and phenol, RNA was extracted with an Rneasy midi kit (QIAGEN), including the DNase I treatment described in the manufacturer's instructions. RNA was spectrophotometrically quantified (at 260 and 280 nm), and the quality was controlled on an agarose electrophoresis gel under denaturing conditions. cDNA was synthesized from RNA in a mixture containing 20 µg of total RNA, random hexamer primers (500 ng; Sigma Genosys), and Eurogentec L. lactis open reading frame (ORF)-specific primers (500 ng). Reverse transcription was performed for 1 h at 42°C with SuperScriptII reverse transcriptase (300 U; Life Technologies), dATP, dGTP, and dTTP (0.3 mM each), and [{alpha}-33P]dCTP (50 µCi; Amersham Biosciences). Then the reaction was performed for an additional 1 h at 42°C with unlabeled dCTP (0.12 mM) to finish cDNA synthesis, and the remaining RNA was hydrolyzed by RNase H (2.4 U; 20 min; 37°C; Life Technologies). Before hybridization, labeled cDNA was purified on Microspin G25 columns (Amersham Biosciences) used according to the manufacturer's instructions.

A set of PCRs (Eurogentec) specific for L. lactis IL1403 and some plasmid genes of industrial relevance were spotted in duplicate on positively charged nylon membranes by using the "Transcriptomic Platform" of Toulouse Genopole. A total of 2,053 of 2,310 ORFs identified in the genome (89%) and 63 plasmid ORFs were effectively available on these membranes. Nylon membranes were hybridized for 14 to 16 h at 68°C with labeled cDNA that had been denatured previously (10 min, 95°C). After hybridization, dried membranes were exposed to a phosphorimager screen for 3 days and scanned with a phosphofluoroimager (Storm 860; Molecular Dynamics).

Dynamic analysis of expression data.
Four growth conditions (beginning and end of both the growth and postacidification phases) were studied simultaneously, and the data were quantified independently three times. Thus, samples were obtained in the growth phase; the first sample, the reference sample, was taken at the onset of growth (6 h), before the temperature decreased, while the second sample was taken at 8.5 h, when the temperature had started to decrease (32°C, pH 5.13) and when the growth was significantly reduced. The next samples were taken in the stationary phase either at the beginning of the phase (15 h, 17°C, pH 4.85) or at the end of the phase (176 h, 12°C, pH 4.64). Hybridization signals were quantified, tested statistically, and assigned to gene names with the Bioplot software (perfected by S. Sokol in the Transcriptomic Platform of Toulouse). After the local background was removed, the mean intensity of the duplicates was normalized by using the mean intensity of all spots. For each gene, the average intensity was calculated from the three independent repetitions in each of the four conditions studied. Genes with average intensities below the detection limit for more than two conditions were not studied further. Expression ratios were calculated with average normalized intensities using the beginning of the growth phase as a reference. Student's test was used, and a P value of <0.05 was considered significant (i.e., there was 95% probability that expression of the gene in the two conditions was significantly different). A total of 702 genes that exhibited at least one significant variation of expression (P < 0.05) during the culture were selected, and the corresponding expression ratios were imported into the GeneSpring software (Silicon Genetics) to find clusters of genes that had similar expression profiles.


arrow
RESULTS AND DISCUSSION
 
Autoacidification, cold stress, and growth behavior.
The industrial L. lactis subsp. lactis biovar diacetylactis strain LD61 was grown in conditions as similar as possible as the conditions used in some cheese-making processes (milk, uncontrolled pH, temperature downshift). Thus, two physicochemical stresses that were likely to modulate bacterial growth and metabolism overlapped during the fermentation. Although acidic stress was directly dependent upon bacterial growth and organic acid production, the cold stress was imposed by a controlled decrease in temperature. The culture was maintained for 580 h to characterize the complete postacidification phase, but no change in the pH value was observed after 180 h of cultivation.

The growth phase ended after about 11 h of culture, when the temperature was only 27°C and the pH had decreased from 6.41 (initial value) to 4.94 (Fig. 1). The maximal biomass concentration produced was 1.02 g/liter. During this phase, the consumption of 18.5 mM lactose and all of the citrate (7.8 mM) led to simultaneous production of 88.5 mM lactic acid and 12.7 mM acetic acid. The citrate present in milk was rapidly consumed from the start of the culture by L. lactis subsp. lactis biovar diacetylactis strain LD61; no citrate was present in the milk after 6 h of culture.



View larger version (21K):
[in this window]
[in a new window]
  		FIG. 1. Changes in substrate and fermentation product concentrations, biomass, external pH, and temperature (°C) during the first 180 h of culture of L. lactis LD61 in skim milk under uncontrolled pH conditions and with a temperature downshift.

The temperature decrease continued after the growth arrest, and the temperature was 12°C after 17.5 h of culture. During the stationary phase, milk acidification continued (postacidification phase), although at a lower rate, which resulted in a final pH of 4.64; 19.9 mM lactic acid and 2.5 mM acetic acid were produced during this period, while 7.1 mM lactose was consumed.

As previously demonstrated (23), the quantity of carbon biomass produced by L. lactis can be considered equivalent to the carbon consumed from amino acids, even if there is not a direct metabolic relationship (24). Consequently, one can expect to observe an equilibrated catabolic carbon balance when values are calculated only from the substrates consumed and the products formed (i.e., without taking into account amino acids and biomass). Indeed, the global catabolic carbon balance calculated for the total culture was 100.4%, illustrating the very good carbon recovery from substrates to products. Moreover, the vast majority of the products formed was lactic acid, which was in total agreement with homolactic metabolism.

In skim milk, the growth of L. lactis LD61 was never exponential, as illustrated by the constant decline in the specific growth rate (Fig. 2). Like the growth rate, the specific rates of lactose consumption and lactate production decreased gradually during the culture. At the time of growth arrest, these catabolic rates were very low, and later they were still low and decreased until approximately 180 h.



View larger version (22K):
[in this window]
[in a new window]
  		FIG. 2. Specific growth rate (µ), lactose consumption rate (qLactose), and lactate production rate ({nu}Lactate) during the first 180 h of culture of L. lactis LD61 in skim milk under uncontrolled pH conditions with a temperature downshift. 1, growth phase; 2, postacidification phase.

Transcriptome profiling.
In order to obtain a chronological view of gene expression and the stress responses of L. lactis LD61 faced with acid and cold stresses in milk, the transcriptome was analyzed and compared for four samples taken during different stages of culture. A decrease in the total intracellular RNA content of 47% was observed at 8.5 h of culture, and the level was maintained in the stationary phase (Table 1), although no RNA degradation was observed on an agarose gel, even late in the culture (176 h) (data not shown). Moreover, the proportion of mRNA in the total RNA, which was estimated by the intensity of the whole membrane by taking into account the finding that the amount of RNA used for cDNA synthesis and hybridization on each membrane was constant, decreased regularly from 100% in the exponential phase to 38% at 176 h. This phenomenon was confirmed by the reduction in the number of spots detected on the membranes, since only 26.1% of the genes were still detected at the end of the culture.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Total RNA concentration, relative mRNA proportion, and number of genes detected on whole-genome nylon membranes of L. lactis LD61 cultivated in milk under uncontrolled pH conditions with a temperature downshift

The 702 genes selected by statistical analysis which showed at least one significant variation compared to the first point analyzed in the culture (6 h) were classified into six expression profile clusters. The first three clusters contained genes that were induced at 8.5 h or 15 h and were overexpressed until the end of the culture. The intensities of the 75 overexpressed genes of cluster 1 were stable until the end of the culture. The intensities increased until the end of the postacidification phase for the 17 genes of cluster 2, while for cluster 3 (46 genes) the intensities decreased from the beginning of the postacidification phase after overexpression during the growth phase; however, in both cases the genes remained overexpressed compared with the reference point. For these three clusters, the maximal intensity means were similar (between 2.1 and 2.3), and the maximal value, the value for the gapA gene, was 7.7. The next two clusters contained genes that were transiently overexpressed, until 15 h of culture for cluster 4 (19 genes) or only during the growth phase for cluster 5 (163 genes), and then underexpressed at the end of the fermentation. Finally, cluster 6 contained more than one-half of the genes analyzed (382 genes), and these genes were underexpressed throughout the culture.

A functional analysis of the different clusters was performed by taking into account the categories established by Bolotin et al. (2) for the strain IL1403 genome. Of the 702 selected genes, 246 encode hypothetical proteins having unknown functions. These genes are uniformly distributed among the different clusters since they represent 33.1 to 37.3% of each cluster. These values are similar to the proportion of hypothetical genes (35.8%) in the entire L. lactis genome (2), indicating that the two categories of genes, known and unknown, are equally involved in the stress response.

Metabolic and transcriptomic analysis of catabolic control.
The two major phenotypic events observed during the culture were concomitant reductions in the specific growth rate (µ) and specific metabolic rates (qLactose and {nu}Lactate), first until the growth arrest and then until the end of the postacidification phase for the catabolic flux (Fig. 2). To further analyze this biochemical behavior, the specific activities of central metabolic enzymes were measured throughout the culture in order to determine metabolic bottlenecks in the glycolytic flux control. Based on the changes in the activities during the culture, these enzymes were classified into three groups (Table 2). The glucokinase (GLK) and lactate dehydrogenase specific activities remained constant during the growth phase, and they decreased 20- and 3-fold in the growth arrest phase and the postacidification phase, respectively. The majority of the enzymes, including glucose-6-phosphate isomerase, phosphofructokinase (PFK), triose phosphate isomerase, 3-phosphoglyceratekinase (PGK), and pyruvate kinase, had similar profiles, in which the activities were stable until there was a great decrease at 176.5 h. Other enzymes showed increased activity either during the growth phase (glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) or when the culture entered the postacidification phase (fructose-1,6-bisphosphate aldolase [FBA], phosphoglycerate mutase [PMG], enolase [ENO], and membrane ATPase).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Enzyme specific activities and in vivo corrected activities compared to the carbon flux in cell extracts of L. lactis LD61 cultivated in skim milk under uncontrolled pH conditions with a temperature downshift

The global decreases in concentration observed for different enzymes (glucose-6-phosphate isomerase, PGK, lactate dehydrogenase, and pyruvate kinase) more or less late during the fermentation was in total agreement with the underexpression of the corresponding genes (Table 2). Indeed, underexpression in the postacidification phase was observed for transcripts of the pyk and ldh genes. These genes, which along with pfk (missing on membranes) belong to the "las" operon (16, 17), and the transcriptional activator ccpA (underexpressed), which is implicated in the regulation of "las" operon transcription (19), exhibited the same expression profile during culture. For two of the four enzymes that exhibited increased specific activities during culture (PMG, FBA, ENO, and GAPDH), a time lag was observed between gene expression and enzyme concentration. The PMG and FBA concentrations increased at 15 h, whereas their transcripts were transiently induced at 8.5 h (ratios, 1.8 and 1.3). The ENO concentration increased substantially after growth ended, which is consistent with the changes in the transcripts of the duplicated genes enoA (ratios, 2.2, 1.4, and 1.7) and enoB (ratios, 1.6, 3.3, and 2.8). enoA is generally considered the functional glycolytic gene (12), and the enoB product may also be implicated in glycolysis. In the same way, the GAPDH specific activity transiently increased at the end of the growth phase, in agreement with the very strong induction of the gapA gene (ratios, 2.3, 4.6, and 7.7). Although gapB, which was not detected here, is considered the essential glycolytic gene (12), it seems that the gapA product could be involved in vivo in these multiple-stress conditions. Finally, GLK exhibited a different behavior since the glk gene was transiently induced at 8.5 h with a ratio of 2, whereas the enzyme concentration varied little during the growth phase and decreased substantially in the stationary phase, suggesting possible involvement of posttranscriptional regulation.

Besides modifications in their concentrations, the activities of all these central metabolic enzymes were substantially modified in vivo, as a function of the intracellular pH and temperature during the fermentation. The pH and temperature effects on enzyme activities were measured for each enzyme (data not shown), which allowed the estimated in vivo enzyme specific activities to be calculated. The estimated values were compared with the glycolytic flux (Table 2) to identify the possible control points of the central metabolism; i.e., an activity-on-flux ratio of around 1 indicated that the enzyme is probably involved in flux control. The glycolytic flux control seemed to be shared by almost all of the enzymes at the beginning of the culture (there was a very large excess of only triose phosphate isomerase compared to the flux), while five enzymes (GLK, PFK, GAPDH, PGK, and ENO) had estimated in vivo activities close to the flux after the growth arrest (Table 2), indicating that these enzymes were probably involved in the slowing down of the flux and control. These observations were reinforced by the dynamic transcriptomic analysis. Indeed, it has been observed previously that a flux bottleneck is generally associated with accumulation of the metabolites just upstream of an enzyme (8). In our experiment, the transcriptomic analysis showed that there was overexpression of some genes, which allowed a change in the flux from the central metabolism, probably to avoid an increase in the concentration of metabolites that was too high just upstream of the bottleneck. The dhaK and dhaM genes, which allowed diversion of excess dihydroxyacetone phosphate to fatty acid production, were transiently overexpressed, thus confirming that there was control at the triose level (GAPDH, PGK, and ENO). Five genes of the Leloir pathway, galE (ratios, 1.2, 0.6, and 0.4), galK (ratios, 2.6, 2, and not determined), galM (ratios, 1.7, 1.9, and 1.5), galT (ratios, 2.4, 2.5, and 0.7), and lacZ (ratios, 2.9, 2.4, and not determined), which were induced until the beginning of the postacidification phase by plasmid genes involved in lactose consumption by the phosphotransferase system (PTS), were underexpressed. These transcriptional events seemed to indicate that there was a shift from the PTS and the tagatose pathway, the preferred and most efficient cell system for lactose catabolism, to the permease and the Leloir pathway. Such a reorientation of the lactose consumption pathway may have been due to inhibition at the PTS level (which was not measured) or at the level of PFK, an enzyme which has been identified as an enzyme involved in a possible glycolysis-limiting step (1). This shift was also confirmed by the transient induction of three (cpsM, rmlB, rmlC) of the four genes involved in sugar-nucleotide synthesis from glucose-1-phosphate resulting from the Leloir pathway, thus preventing more carbon entry at the glucose-6-phosphate level in conditions in which the PFK activity is not high enough to ensure a high glycolytic carbon flux.

The combination of metabolic and transcriptomic data presented in this work has great potential for identification of genomic and metabolic events that occur in a bacterial culture. In most cases, the observed regulations and limitations observed at the enzymatic level could be confirmed by the transcriptomic analysis. In other cases, a transcriptomic response was observed, which was not necessarily directly related to a significant carbon flux modification. The transcriptome can thus be considered a strong and sensitive tool for analyzing the metabolic control of a pathway, since genes involved in peripheral pathways were induced just upstream of the postulated carbon bottleneck. Finally, another interesting aspect of the transcriptome illuminated in our analysis was the identification of the functional gene when duplicated genes for the same function were present in the genome sequence (e.g., gap or eno genes).

Growth and nitrogen metabolism.
The functions closely linked to growth exhibited similar transcriptional changes. Indeed, more than 73% of the genes detected involved in cell wall biosynthesis were underexpressed starting at 8.5 h. Similarly, of the eight "dna" genes essential for replication, the levels of 7 were too low to be detected, and dnaC was repressed throughout the culture. The other functions, cell division, transcription, and translation, were repressed as well, but they were repressed later. The genes implicated in cell division were repressed globally either starting at 8.5 h (ftsA, ftsH, ftsQ, ftsW1, and the cell division regulator ezrA) or starting at 15 h (ftsX, ftsZ, gidA, and mesJ). Just two of the five genes coding for RNA polymerase subunits (rpoA and rpoE) were clustered and were shown to be strongly underexpressed in the postacidification phase (ratios, 0.3 and not determined) despite induction at 8.5 h. The decreasing expression of RNA polymerase subunits genes was probably related to the decline in transcription and the decrease in mRNA concentration in the postacidification phase. Translation was also repressed, since 33 of the 61 genes detected which have been implicated in this function were underexpressed in these conditions, and 24 exhibited, after brief weak induction (ratios between 1.05 and 1.5), a great reduction in expression ratios (maximal ratios, 0.5). Only four genes involved in translation were overexpressed throughout the culture. One of these genes, rpmH, is the only gene coding for a ribosomal 50S subunit protein, L34, whose level was almost doubled at 8.5 h and remained high during the postacidification phase (ratio at 176 h, 1.4). During carbon starvation, another ribosomal protein, L33 (rpmGC genes), showed a similar induction pattern during the stationary phase (32). It seems that these ribosomal proteins play a special role in maintenance of translation after growth arrest. Although translation should have been strongly decreased during the stationary phase, it seemed to be still active at the beginning of this phase since the in vivo concentrations of some glycolytic enzymes (FBA, PMG, and ENO) increased at 26.5 h.

Milk is a very particular medium, in which the use of caseins as amino acid sources and for growth requires functional extracellular proteases, oligopeptide transporters, and intracellular peptidases. In our conditions, 70% of the genes coding for oligopeptide transporters (oppA, oppB, oppC, oppD, oppF, optD, and optS) and eight genes coding for extracellular proteases, including the plasmid gene prtP, were underexpressed after the decrease in growth. Only four genes involved in this function were transiently weakly induced at 8.5 h (maximal ratio, 1.4) but strongly underexpressed after the growth arrest since none of these genes were detected at the end of the culture. The majority of the genes detected that were involved in amino acid biosynthesis were also repressed, and the few genes which were induced code for enzymes involved either in the biosynthesis of histidine, one of the less prominent amino acids in milk (15), or in aspartate (metE) and glutamate (argG) family biosynthesis pathways, which have specific functions directly linked to the physicochemical stresses encountered during fermentation (see below).

Acid- and cold-stress-specific responses.
During fermentation a number of physiological responses directly linked to acid and cold stresses were brought to the fore. First, autoacidification induced pathways known to be involved in protection against cytoplasm acidification. The arcA, arcB, arcC1, arcC2, and arcD1 genes, linked to arginine metabolism, were strongly induced at 8.5 h, with abundance ratios ranging from 1.9 to 3.4; these results are in agreement with the previous demonstration of induction of the arginine pathway in acid conditions (6). These genes encode proteins involved in arginine deamination (i.e., the arginine deiminase [ADI] pathway) (27) and the associated arginine/ornithine aniporter, whose role is ATP and NH3 production that enables cytoplasm alkalinization. The kinA/llrA two-component system is known to be a transcriptional activator of the ADI pathway (25), but in our conditions no induction of these genes was observed. We propose that either the induction of this system took place during the growth phase and could not be detected or the kinC/llrC system involved in global stress responses (25) and transiently induced in our experiment was responsible for ADI activation. In addition to the ADI pathway, two genes coding for reversible enzymes involved in arginine peripheral functions were also induced; argG is involved in biosynthesis or catabolism of arginine, and argS plays a role in loading and unloading of tRNAArg. Also, the overexpression of four genes, potA, potB, and potD implicated in the transport of the biogenic amines spermidine and putrescine produced from ornithine (the ADI product) and metk implicated in the biosynthesis of S-adenosylmethionine necessary for spermidine synthesis, is interesting. The induction of this pathway is of interest since the synthesis of biogenic amines has consequences for the organoleptic qualities and safety of cheese products. However, the genes responsible for the two last steps of the synthesis of these biogenic amines (ornithine decarboxylase and spermidine synthase) have not been identified yet in the L. lactis IL1403 chromosome, and although biogenic amine production has been found in some LAB, the functionality of this pathway in the LD61 strain has not been demonstrated, and to our knowledge, this strain does not produce biogenic amines. Two other systems that allow proton consumption were also induced. (i) The ATPase activity, which enables proton extrusion coupled to ATP hydrolysis, increased during the postacidification phase, with a maximal value at 97 h, while no transcript induction was observed for the atp genes coding for ATPase subunits, suggesting that ATPase activity could be regulated at the posttranscriptional level. (ii) The citrate decarboxylation pathway of L. lactis is known to be activated in acid stress conditions (7, 20). This activation can be seen at the transcript level since four genes, citC, citE, citF, and citM involved in the decarboxylation pathway, were transiently induced late in the growth phase with expression ratios between 1.3 and 1.6. However, this response was independent of the citrate availability since citrate was exhausted from the medium very early in the culture as a consequence of the plasmid citrate assimilation activity characteristic of L. lactis subsp. lactis biovar diacetylactis strains (33, 34). It should be noted that citrate permease induction was probably linked to citrate availability since the corresponding transcript was underexpressed in the late growth phase.

Unlike the effect of acid stress, the response to cold stress was only partial. The decrease in temperature during the culture probably had a physical effect on the bacteria since two genes, fabZ and thiL, which support the synthesis of unsaturated and long-chain fatty acids that protect against membrane rigidity, were induced. However, the decrease in temperature did not provoke a large cold shock response since neither of the two L. lactis csp genes, cspD and cspE, was induced, and only three genes encoding known cold-induced proteins involved in different cellular processes (36), llrC, ptsH, and osmC, were overexpressed.

Cross-responses to stresses.
Genes normally induced by different stresses were also induced in our conditions. The groES and grpE genes encoding chaperones involved in the heat shock response (13) were transiently induced with a weak range (ratios, 1.6 and 1.4), although clpB and clpE, whose products are ATP-linking subunits of the Clp protease, were highly induced throughout the culture, with maximal ratios of 3.1 and 2.2, respectively. However this heat-shock-type response was only partial since the other genes of the heat shock regulon, clpP, clpC, clpX, and dnaK (groEL was not on the membrane), were either underexpressed (clpP and clpC) or not detected. The ctsR gene, which is involved in the regulation of clp genes (35), was underexpressed during the culture, and hrcA, which encodes a transcriptional regulator of chaperone genes (13), was constitutive. The induction of the ytgH gene, which encodes a general stress protein homologous to the gls24 product of Enterococcus faecalis (9), was induced at 8.5 h, like cstA involved in carbon starvation. Although the culture was maintained in anaerobic conditions, a number of genes for aerobic metabolism were transiently overexpressed. Genes involved in the oxygen response, like genes encoding oxidoreductases (ycgD, yphC, and yugB), NADH oxidase (noxC), and NADH dehydrogenase (noxA), genes coding for cofactors important in aerobiosis (gpo, trxA, and trxB2), and genes of metabolic pathways specific for the presence of oxygen (butA, butB, qor, cydA, pdhB, and pdhD) were induced. However, since anaerobiosis was maintained, no NADH oxidase activity was detected in cellular extracts throughout the culture, and a gene like sodA, which is essential for survival in the presence of oxygen, was not induced. Some genes linked to UV stress and DNA damage repair or degradation, including uvrA, recJ, xseA, gyrA, recN, and ssbB, were induced at 8.5 h, probably to protect against mutagen effects of acid stress.

An original response was observed with regard to phosphate starvation. Of the 10 genes encoding high-affinity phosphate and phosphonate transporters, 8 (pstA, pstB, pstC, pstE, phnA, phnB, phnC, and phnE) were strongly overexpressed (with ratios between 1.6 and 3.1), generally until the beginning of the postacidification phase. The pst genes belong to the Pho regulon induced during phosphate starvation in Bacillus subtilis. Milk is a medium that is relatively rich in phosphate, but a decrease in the pH, which is known to have a negative effect on phosphate transport (28, 29), may diminish phosphate import, thus mimicking a starvation state for the cell and then inducing high-affinity transporters. Moreover, cellular phosphate limitation may have an effect on acid stress resistance because a link between resistance to low pH and mutation of phosphate transporters has been observed by Rallu et al. (31).

Other responses.
Some pathways which do not seem to have a link with the stresses encountered in the culture were also overexpressed. The guaC, purA, purC, purD, purE, purF, and purL genes, belonging to the purine biosynthesis pathway, were overexpressed at 8.5 h with ratios between 1.4 and 5.8, and genes for the salvation pathway (xpt, deoB, and prsA), as well as pbuX encoding a xanthine permease and trxA and trxB2 involved in metabolism of thioredoxin, a cofactor that plays a role in nucleotide conversion, were also induced. There should be concerted opposite regulation of the purine and pyrimidine metabolic pathways since the udp, pyrH, carB, and dut genes of the pyrimidine biosynthesis pathway were repressed or considered constitutive for thyA. The activation of the purine biosynthesis pathway may be linked either to the milk medium, which is known to contain particularly low levels of these compounds, or to the different physicochemical stresses. Only the guaA and dukA genes, which are involved in GMP biosynthesis, were underexpressed in these conditions. However, there may be a relationship between the repression of GMP synthesis and acid stress, since a link between deletion of guaA and acid stress resistance was observed in L. lactis by Rallu et al. (31).

Activation of many prophages was seen throughout the culture, sometimes with high ratios (1.2 to 3), as observed by Redon et al. (32) for L. lactis in carbon starvation conditions. This response seemed to be dependent upon the culture conditions or the stresses, since four prophages (pi1, pi2, pi3, and ps2) were induced in both cultures, while the ps3 prophage was specifically induced in carbon starvation conditions (32) and repressed during fermentation, whereas the ps1 prophage was never induced. The reason for induction of these genes and their potential role in the stress response remain to be determined.

Concluding remarks.
Transcriptomic analysis combined with phenotypic and metabolic characterization of a bacterium cultivated in controlled conditions is a powerful tool for obtaining an exhaustive view of cell behavior. In the present study, the absence of data related to the intracellular metabolite concentrations due to the use of milk medium was overcome by transcriptomic data, which enabled illumination of activation of peripheral pathways upstream of the supposed metabolic bottlenecks.

Of the many stress answers identified, arginine metabolism seems to be an essential weapon against stress since many genes involved in pathways around this amino acid were induced, perhaps including the synthesis of biogenic amines. Our exhaustive analysis also highlighted original answers not directly related to the stresses that the cells were faced with, for example, the phosphate starvation response, oxidative stress responses, the concerted induction of purine biosynthesis pathways, and prophage induction. This global approach with well-controlled culture conditions will undoubtedly provide new insights for understanding lactic acid bacteria in the food processing context.

The use of DNA chips for food microorganisms will probably increase very soon because of the desire of industry to understand the genomic basis of variable phenotypic and technological properties.


arrow
ACKNOWLEDGMENTS
 
We thank Sophie Mondeil for her technical assistance with measurement of in vitro enzyme activities and determination of the effects of pH and temperature on the enzyme activities.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Laboratoire Biotechnologie-Bioprocédés, UMR 5504 INSA/CNRS & UMR 792 INSA/INRA, Institut National des Sciences Appliquées, 135 Avenue de Rangueil, 31077 Toulouse cedex 4, France. Phone: (33) 561 559 438. Fax: (33) 561 559 400. E-mail: loubiere{at}insa-toulouse.fr. Back


arrow
REFERENCES
 
    1
  1. Andersen, H., C. Solem, K. Hammer, and P. R. Jensen. 2001. Twofold reduction of phosphofructokinase activity in Lactococcus lactis results in strong decrease in growth rate and in the glycolytic flux. J. Bacteriol. 183:3458-3467.[Abstract/Free Full Text]
    2. 2
  2. Bolotin, A., P. Wincker, S. Mauger, O. Jaillon, K. Malarme, J. Weissenbach, S. D. Ehrlich, and A. Sorokin. 2001. The complete genome sequence of the lactic acid bacterium Lactococcus lactis ssp. lactis IL1403. Genome Res. 11:731-753.[Abstract/Free Full Text]
    3. 3
  3. Cocaign-Bousquet, M., and N. D. Lindley. 1995. Pyruvate overflow and carbon flux within the central metabolic pathway of Corynebacterium glutamicum during growth on lactate. Enzyme Microbiol. Technol. 79:108-116.
    4. 4
  4. Drici-Cachon, Z., J. Guzzo, J.-F. Cavin, and C. Diviès. 1996. Acid tolerance in Leuconostoc oenos. Isolation and characterisation of an acid-resistant mutant. Appl. Microbiol. Biotechnol. 44:785-789.[CrossRef]
    5. 5
  5. Even, S., N. D. Lindley, and M. Cocaign-Bousquet. 2001. Molecular physiology of sugar catabolism in Lactococcus lactis IL1403. J. Bacteriol. 183:3817-3824.[Abstract/Free Full Text]
    6. 6
  6. Even, S., N. D. Lindley, P. Loubiere, and M. Cocaign-Bousquet. 2002. Dynamic response of catabolic pathways to autoacidification in Lactococcus lactis: transcript profiling and stability in relation to metabolic and energetic constraints. Mol. Microbiol. 45:1143-1152.[CrossRef][Medline]
    7. 7
  7. Garcia-Quintans, N., C. Magni, D. de Mendoza, and P. Lopez. 1998. The citrate transport system of Lactococcus lactis subsp. lactis biovar diacetylactis is induced by acid stress. Appl. Environ. Microbiol. 64:850-857.[Abstract/Free Full Text]
    8. 8
  8. Garrigues, C., P. Loubière, N. D. Lindley, and M. Cocaign-Bousquet. 1997. Control of the shift from homolactic to mixed-acid fermentation in Lactococcus lactis: predominant role of the NADH/NAD+ ratio. J. Bacteriol. 179:5282-5287.[Abstract/Free Full Text]
    9. 9
  9. Giard, J. C., A. Rince, H. Capiaux, Y. Auffray, and A. Hartke. 2000. Inactivation of the stress- and starvation-inducible gls24 operon has a pleiotrophic effect on cell morphology, stress sensitivity, and gene expression in Enterococcus faecalis. J. Bacteriol. 182:4512-4520.[Abstract/Free Full Text]
    10. 10
  10. Hartke, A., S. Bouche, J. C. Giard, A. Benachour, P. Boutibonnes, and Y. Auffray. 1996. The lactic acid stress response of Lactococcus lactis subsp. lactis. Curr. Microbiol. 33:194-199.[CrossRef][Medline]
    11. 11
  11. Hutkins, R. W., and N. L. Nannen. 1993. pH homeostasis in lactic acid bacteria. J. Dairy Sci. 76:2354-2365.[Abstract]
    12. 12
  12. Jamet, E., S. D. Ehrlich, F. Duperray, and P. Renault. 2001. Study of the duplicated glycolytic genes in Lactococcus lactis IL1403. Lait 81:115-129. (In French.)[CrossRef]
    13. 13
  13. Kilstrup, M., S. Jacobsen, K. Hammer, and F. K. Vogensen. 1997. Induction of heat shock proteins DnaK, GroEL, and GroES by salt stress in Lactococcus lactis. Appl. Environ. Microbiol. 63:1826-1837.[Abstract]
    14. 14
  14. Kim, W. S., J. Ren, and N. W. Dunn. 1999. Differentiation of Lactococcus lactis subspecies lactis and subspecies cremoris strains by their adaptive response to stresses. FEMS Microbiol. Lett. 171:57-65.[CrossRef][Medline]
    15. 15
  15. Lindmark-Mansson, H., R. Fondèn, and H.-E. Pettersson. 2003. Composition of Swedish dairy milk. Int. Dairy J. 13:409-425.[CrossRef]
    16. 16
  16. Llanos, R. M., A. J. Hillier, and B. E. Davidson. 1992. Cloning, nucleotide sequence, expression, and chromosomal location of ldh, the gene encoding L-(+)-lactate dehydrogenase, from Lactococcus lactis. J. Bacteriol. 174:6956-6964.[Abstract/Free Full Text]
    17. 17
  17. Llanos, R. M., C. J. Harris, A. J. Hillier, and B. E. Davidson. 1993. Identification of a novel operon in Lactococcus lactis encoding three enzymes for lactic acid synthesis: phosphofructokinase, pyruvate kinase, and lactate dehydrogenase. J. Bacteriol. 175:2541-2551.[Abstract/Free Full Text]
    18. 18
  18. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275.[Free Full Text]
    19. 19
  19. Luesink, E. J., R. E. van Herpen, B. P. Grossiord, O. P. Kuipers, and W. M. de Vos. 1998. Transcriptional activation of the glycolytic las operon and catabolite repression of the gal operon in Lactococcus lactis are mediated by the catabolite control protein CcpA. Mol. Microbiol. 30:789-798.[CrossRef][Medline]
    20. 20
  20. Magni, C., D. de Mendoza, W. N. Konings, and J. S. Lolkema. 1999. Mechanism of citrate metabolism in Lactococcus lactis: resistance against lactate toxicity at low pH. J. Bacteriol. 181:1451-1457.[Abstract/Free Full Text]
    21. 21
  21. Mercade, M., N. D. Lindley, and P. Loubiere. 2000. Metabolism of Lactococcus lactis subsp. cremoris MG 1363 in acid stress conditions. Int. J. Food Microbiol. 55:161-165.[CrossRef][Medline]
    22. 22
  22. Mercade, M., S. Raynaud, M. Cocaign-Bousquet, and P. Loubière. Physiological analysis of lactic acid bacteria in skim milk: methods improvement and development. Submitted for publication.
    23. 23
  23. Novak, L., M. Cocaign-Bousquet, N. D. Lindley, and P. Loubière. 1997. Metabolism and energetics of Lactococcus lactis during growth in various complex or synthetic media. Appl. Environ. Microbiol. 63:2665-2670.[Abstract]
    24. 24
  24. Novak, L., and P. Loubière. 2000. The metabolic network of Lactococcus lactis: distribution of 14C-substrates between catabolic and anabolic pathways. J. Bacteriol. 182:1136-1143.[Abstract/Free Full Text]
    25. 25
  25. O'Connell-Motherway, M., D. van Sinderen, F. Morel-Deville, G. F. Fitzgerald, S. D. Ehrlich, and P. Morel. 2000. Six putative two-component regulatory systems isolated from Lactococcus lactis subsp. cremoris MG1363. Microbiology 146:935-947.[Abstract/Free Full Text]
    26. 26
  26. O'Sullivan, E., and S. Condon. 1997. Intracellular pH is a major factor in the induction of tolerance to acid and other stresses in Lactococcus lactis. Appl. Environ. Microbiol. 63:4210-4215.[Abstract]
    27. 27
  27. Poolman, B., A. J. Driessen, and W. N. Konings. 1987. Regulation of arginine-ornithine exchange and the arginine deiminase pathway in Streptococcus lactis. J. Bacteriol. 169:5597-5604.[Abstract/Free Full Text]
    28. 28
  28. Poolman, B., A. J. Driessen, and W. N. Konings. 1987. Regulation of solute transport in streptococci by external and internal pH values. Microbiol. Rev. 51:498-508.[Free Full Text]
    29. 29
  29. Poolman, B., R. M. Nijssen, and W. N. Konings. 1987. Dependence of Streptococcus lactis phosphate transport on internal phosphate concentration and internal pH. J. Bacteriol. 169:5373-5378.[Abstract/Free Full Text]
    30. 30
  30. Rallu, F., A. Gruss, and E. Maguin. 1996. Lactococcus lactis and stress. Antonie Leeuwenhoek 70:243-251.
    31. 31
  31. Rallu, F., A. Gruss, S. D. Ehrlich, and E. Maguin. 2000. Acid- and multistress-resistant mutants of Lactococcus lactis: identification of intracellular stress signals. Mol. Microbiol. 35:517-528.[CrossRef][Medline]
    32. 32
  32. Redon, E., P. Loubière, and M. Cocaign-Bousquet. 2005. Transcriptome analysis of the progressive adaptation of Lactococcus lactis to carbon starvation. J. Bacteriol. 187:3589-3592.[Abstract/Free Full Text]
    33. 33
  33. Sesma, F., D. Gardiol, A. P. de Ruiz Holgado, and D. de Mendoza. 1990. Cloning of the citrate permease gene of Lactococcus lactis subsp. lactis biovar diacetylactis and expression in Escherichia coli. Appl. Environ. Microbiol. 56:2099-2103.[Abstract/Free Full Text]
    34. 34
  34. Smith, M. R., J. Hugenholtz, P. Mikoczi, E. de Ree, A. W. Bunch, and J. A. M. de Bont. 1992. The stability of the lactose and citrate plasmids in Lactococcus lactis subsp. lactis biovar. diacetylactis. FEMS Microbiol. Letters. 96:7-12.
    35. 35
  35. Varmanen, P., H. Ingmer, and F. K. Vogensen. 2000. ctsR of Lactococcus lactis encodes a negative regulator of clp gene expression. Microbiology 146:1447-1455.[Abstract/Free Full Text]
    36. 36
  36. Wouters, J. A., M. Mailhes, F. M. Rombouts, W. M. de Vos, O. P. Kuipers, and T. Abee. 2000. Physiological and regulatory effects of controlled overproduction of five cold shock proteins of Lactococcus lactis MG1363. Appl. Environ. Microbiol. 66:3756-3763.[Abstract/Free Full Text]

Applied and Environmental Microbiology, December 2005, p. 8016-8023, Vol. 71, No. 12
0099-2240/05/$08.00+0     doi:10.1128/AEM.71.12.8016-8023.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Mansour, S., Bailly, J., Landaud, S., Monnet, C., Sarthou, A. S., Cocaign-Bousquet, M., Leroy, S., Irlinger, F., Bonnarme, P. (2009). Investigation of Associations of Yarrowia lipolytica, Staphylococcus xylosus, and Lactococcus lactis in Culture as a First Step in Microbial Interaction Analysis. Appl. Environ. Microbiol. 75: 6422-6430 [Abstract] [Full Text]  
  • Nouaille, S., Even, S., Charlier, C., Le Loir, Y., Cocaign-Bousquet, M., Loubiere, P. (2009). Transcriptomic Response of Lactococcus lactis in Mixed Culture with Staphylococcus aureus. Appl. Environ. Microbiol. 75: 4473-4482 [Abstract] [Full Text]  
  • Sieuwerts, S., de Bok, F. A. M., Hugenholtz, J., van Hylckama Vlieg, J. E. T. (2008). Unraveling Microbial Interactions in Food Fermentations: from Classical to Genomics Approaches. Appl. Environ. Microbiol. 74: 4997-5007 [Full Text]  
  • Pedersen, M. B., Garrigues, C., Tuphile, K., Brun, C., Vido, K., Bennedsen, M., Mollgaard, H., Gaudu, P., Gruss, A. (2008). Impact of Aeration and Heme-Activated Respiration on Lactococcus lactis Gene Expression: Identification of a Heme-Responsive Operon. J. Bacteriol. 190: 4903-4911 [Abstract] [Full Text]  
  • Maligoy, M., Mercade, M., Cocaign-Bousquet, M., Loubiere, P. (2008). Transcriptome Analysis of Lactococcus lactis in Coculture with Saccharomyces cerevisiae. Appl. Environ. Microbiol. 74: 485-494 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Raynaud, S.
Right arrow Articles by Loubiere, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Raynaud, S.
Right arrow Articles by Loubiere, P.
Agricola
Right arrow Articles by Raynaud, S.
Right arrow Articles by Loubiere, P.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»