INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The toxic cellular effects of oxygen are a major factor in aging and mortality (3, 28). Oxygen toxicity is attributed to the activity of reactive oxygen species that attack proteins, lipids, and nucleic acids (15). Effects of oxygen have been extensively studied by use of bacterial models, principally with the facultatively respiring bacterium Escherichia coli (see references 7 and 13 for reviews). In this model, respiration itself is implicated as a source of oxidative damage in E. coli (8, 9, 18, 20, 27, and 36). It has been suggested that the shutdown of respiration in nutrient-limited conditions may reduce reactive oxygen species levels and thereby improve E. coli survival. Recent evidence further suggests that survival is favored by shifting cells to anaerobic conditions during entry into stationary phase (9).

Current information on the effects of oxygen is mainly based on respiring organisms. As such, the question of what anaerobes do in the presence of oxidative stress has been explored little. It is presumed that these organisms cope with stress in much the same way as aerobes, except that their defense systems, which may include superoxide dismutases (SODs) and catalases, may be more limited. However, there has been no demonstration to date that responses of anaerobes to an oxidative environment are predictable from the behavior of respiring bacteria.

The effects of oxygen have been examined with Lactococcus lactis, a gram-positive facultative anaerobe with a fermentative metabolism that can use different sugars to produce mainly L-(+)-lactic acid (19). Oxygenation of cultures results in an altered redox state and greater NADH oxidase activity (24, 25, 35); as a consequence, sugar fermentation is shifted toward mixed fermentation, and acetic acid, formic acid, CO₂, ethanol, and acetoin, as well as lactic acid, are produced (25).

Despite its classification as an anaerobe and studies that have focused nearly entirely on its fermentative metabolism, results obtained about 30 years ago suggested that L. lactis is able to undergo respiratory growth, provided that heme is added to aerated cultures; this view was supported by a demonstration of altered metabolic end products, cytochrome formation, and hemin-dependent oxygen uptake (34). However, more recent studies of an L. lactis subsp. diacetylactis strain suggested that respiration does not occur under these conditions, as cytochromes could not be detected (21). To date this question has not been further explored, and the consequences of respiratory growth have not been analyzed.

The toxic effects of oxygen on L. lactis growth and survival have been revealed by several studies under fermentation conditions. Growth is reportedly inhibited by oxygen (5), and prolonged aeration of lactococcal cultures can lead to cell death and DNA degradation (10). Oxygen toxicity may be due to formation of hydrogen peroxide and hydroxyl radicals (1, 10). Unlike E. coli, L. lactis possesses a single SOD (31) and no catalase. It was found that the addition of exogenous catalase improved survival of L. lactis cells exposed to oxygen (10). These results suggest that L. lactis may not be fully equipped to withstand the toxic effects of an oxidative environment.

Our studies on oxygen toxicity led us to dissect the positive effects of the addition of exogenous catalase on growth and survival of L. lactis. As catalase contains a heme nucleus (in which iron is complexed with a porphyrin molecule), we first examined the effects of oxygen in the presence of heme. We confirmed that L. lactis is capable of respiratory growth, in agreement with earlier work (34). Respiration conditions result in improved growth and a spectacular increase in long-term survival compared to growth under conventional fermentation conditions. The observed phenotypes require an intact cydA gene, which encodes cytochrome d oxidase. Under respiration conditions, fermentation occurs during initial growth, while respiration is greatest during the late exponential phase.

(An initial oral communication of respiration and its effects on lactococci was first presented at the Lactic Acid Conference in Veldhoven, Holland, in September 1999.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains, plasmids, and growth conditions. The strains used were L. lactis MG1363 (16); MG1363 sodA (31) and hemZ and cydA (this work) derivatives; and L. lactis subsp. lactis and L. lactis subsp. cremoris strains IL-1403, IL-582, IL801, IL896, Z105, Z106, and Z191 (CNRZ strain collection; kindly provided by P. Tailliez [37]). The two subspecies are closely related (over 80% identity between the subspecies strains, as determined using BLAST for gene analogues). Plasmid pKatE encodes the Bacillus subtilis katE gene (12) under the control of a constitutive promoter (P. Duwat and S. Sourice, unpublished data).

Metabolic product determinations. Biochemical measurements were performed with 24-h culture supernatants of cells grown as described above. Measurements of glucose, acetate, lactate, and ethanol were obtained according to kit supplier's instructions (Boehringer GmbH, Mannheim, Germany). Levels of diacetyl and acetoin were determined by gas chromatography (HRGC5160 Carlo Erba Instrument; ThermoQuest, Les Ulis, France) with an FFAP column (12 m; inner diameter, 0.52 mm; phase thickness, 1 µm; J and W Scientific, Folsom, Calif.). Conditions for static headspace analysis were as follows. Five milliliters of a culture supernatant saturated with NH₂SO₄ (5 g) and containing 0.25 ml of 22% H₂SO₄ plus 0.2 ml of a 0.15% aqueous solution of 2-pentanol as an internal standard was placed in a 22-ml vial fitted with a Minimert valve (Interchim). The mixture was warmed to 60°C for 15 min before samples (0.1 ml) were removed with a gas syringe. Chromatography was performed with an initial oven temperature of 55°C (2 min). The temperature was increased by 30°C min¹ up to 145°C and then by 5°C min¹ up to 150°C (kept there for 4 min). The flow rate of the carrier gas (H₂) was 5 ml min¹. The temperature of the injector, in the splitless mode, was 145°C; that of the flame ionization detector was 200°C. Data were recorded with a Kontron PC integration pack. Concentrations were calculated from standard curves.

Oxygen detection. Dissolved oxygen was determined using an oxygen meter (Ponselle, Versailles, France). Twenty-four-hour aerated (oxygenated) cultures grown with or without hemin (Ox/H or Ox/No H, respectively) were centrifuged, washed twice in saline buffer (50 mM Tris-HCl, 0.15 M NaCl [pH 7]), and resuspended at an OD₆₀₀ of 0.5 in saline buffer containing 1% glucose. Aerated buffer devoid of bacteria was used as the reference for oxygen-saturated medium, and aerated water was used for calibrations. Thirty-milliliter samples were used for measurements. After cells were resuspended in buffer, all samples were vortexed and allowed to reequilibrate for 3 min before measurement at the first time point. Measurements were obtained at room temperature. Data represent the means of two independent experiments. Measurements varied by a maximum of 10%.

Mutant constructions. Alignments of microbial cydA and hemZ genes (http://www.ncbi.nlm.nih.gov/BLAST/unfinishedgenome.html and http://www.ncbi.nlm.nih.gov/BLAST/) were used to design several pairs of degenerate primers predicted to isolate respective internal gene fragments. The primers used in the final experiments were as follows: 5'-CARTTYGGNATGAAYTGG-3' and 5'-CATRATNCTRAANGTCCA-3' for cydA (designed to amplify cydA open reading frame positions 75 through 334, as defined for ~470-amino-acid full-length cytochrome d ubiquinol oxidase subunit I) and 5'-ATHTGGACNGAYGARGG-3' and 5'-ARNGGDATNCCRTGRTA-3' for hemZ (designed to amplify hemZ open reading frame positions 70 through 200, as defined for ~310-amino-acid full-length ferrochelatase lyase). PCRs using Thermus aquaticus Taq polymerase were carried out as follows: 5 min at 96°C, followed by 30 cycles of 15 s at 96°C, 15 s at 50°C, and 15 s at 72°C. Fragment sizes were ~780 bp for cydA and ~390 bp for hemZ. After end filling in using T4 polymerase plus the Klenow fragment, purified fragments were cloned into the SmaI site of pRV300 (23). Sequences were confirmed to correspond to putative cydA and hemZ genes.

Northern blotting experiments. Cultures were inoculated with a 1/200 dilution of an overnight culture of MG1363 grown without aeration in M17-glu and then were grown under Ox/H and Ox/No H conditions. Samples were removed at various cell densities, corresponding to about 3, 4, 6, and 8 h after the start of growth. Cell concentrations were all adjusted to a cell density equal to an OD₆₀₀ of 1 before cell samples were prepared. RNA preparation and Northern blot analyses were carried out as previously described (29). The cydA-specific probe corresponds to a 777-bp PCR-amplified internal fragment, as described above. A 16S ribosomal DNA (rDNA)-specific probe was used to ensure that the same amounts of total RNA were present in the wells (data not shown). Probes were labeled by nick translation using Ready-to Go (Pharmacia) and [³²P]dCTP.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

L. lactis growth period is extended in aerated medium containing heme. The growth of L. lactis strain MG1363 in M17-glu was compared under Ox/H or Ox/No H conditions and conditions of nonoxygenation with or without hemin (No Ox/H or No Ox/No H, respectively) (Fig. 1). Exponential-phase growth kinetics were similar for the first 8 h under all four conditions. After this time, the No Ox/H, Ox/No H, and No Ox/No H cultures entered stationary phase and attained a final pH of 4.4. In contrast, growth of the Ox/H culture continued, with a final biomass about twofold greater than those of the other three cultures and a final pH of ~6. The pH decline during the first 8 h of growth was similar for all cultures; after this time, the Ox/H culture pH started to rise. These results suggest that L. lactis metabolism is altered by growth in Ox/H conditions; differences are first detected just before control cultures enter stationary phase.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Altered growth and pH of L. lactis grown under Ox/H conditions. An overnight 30°C MG1363 culture in M17-glu was used to inoculate media (at a 1/1,000 dilution) for growth and pH measurements under four conditions at 30°C: Ox/H, No Ox/H, Ox/No H, and No Ox/No H. As growth curves and final pHs for the No Ox/H, Ox/No H, and No Ox/No H conditions were similar, only data for the No Ox/No H (open squares; solid line for growth, broken line for pH) and Ox/H (closed circles; solid line for growth, broken line for pH) conditions are shown. Growth of all cultures was saturated at 24 h. Experiments were performed six times and yielded comparable results. Results of a representative experiment are shown.

Ox/H conditions result in markedly improved long-term survival. Strain MG1363 was cultured under Ox/H, No Ox/H, Ox/No H, and No Ox/No H conditions. The cell populations of the latter three cultures at 24 h reached ~2 × 10⁹ per ml, while that of the Ox/H culture was 7 × 10⁹ per ml. After 80 h, survival of the three control cultures dropped 10⁵-, 10⁶-, and 10⁸-fold for the No Ox/H, Ox/No H, and No Ox/No H cultures, respectively. In marked contrast, the viability of the Ox/H culture remained within twofold the initial cell number.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Improved long-term survival of L. lactis after growth under Ox/H conditions. MG1363 cells were grown under four conditions at 30°C as described in the legend to Fig. 1: Ox/H (closed circles), No Ox/H (closed squares), Ox/No H (open circles), and No Ox/No H (open squares). After 24 h, cultures were transferred to 4°C. Cell viability was determined by plating dilutions on solid M17-glu at the indicated times. Experiments were performed four times and yielded comparable results. Results of a representative experiment are shown.

Metabolism is altered under Ox/H compared to fermentative growth conditions, particularly late in growth. During L. lactis fermentation, glucose is metabolized to form mainly lactic acid, and the final pH is ~4.5. However, while the Ox/H culture initially produced a pH decrease, the final pH was significantly higher (Fig. 1). As this result suggested altered metabolism, particularly late in growth, we compared glucose consumption and the amounts of expected fermentation products for Ox/H, Ox/No H, and No Ox/No H cultures (Fig. 3). The three cultures displayed similar glucose consumption and lactic acid accumulation during the first 6 to 7 h of growth, indicating that fermentation occurs in all cases. After this time, the Ox/H culture could be distinguished from the other cultures by continued rapid glucose consumption, arrest, and subsequent reduction in lactic acid accumulation. (Final amounts of lactic acid under Ox/H conditions were sometimes as low as 1 g/liter; data not shown.) Acetoin was observed to accumulate early during growth in the Ox/H culture, and the amounts were clearly different from those in the other cultures; the greatest increment in acetoin compared to that in the other cultures occurred after ~8 h of growth. Similarly, higher acetate levels in the Ox/H culture showed the greatest difference late in growth. These results show that (i) the higher final pH of the Ox/H culture is correlated with reduced amounts of acid end products; (ii) the Ox/H culture undergoes fermentation in the beginning of growth, as reflected by lactic acid accumulation; and (iii) a metabolic shift in the Ox/H culture is most pronounced late in growth. The above results are consistent with the results of earlier studies (34) that correlated the altered metabolism with respiration. They further suggest that the respiration-like metabolism occurs late in growth.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Metabolic alterations in Ox/H cultures are greatest in late exponential phase. L. lactis MG1363 was grown aerobically in M17-glu in the presence or absence of hemin or anaerobically in the absence of hemin. Samples were taken at various intervals to monitor growth, glucose consumption, and lactate, acetate, acetoin, and diacetyl accumulation (see Materials and Methods). Determinations were performed at least twice, with a maximum deviation between measurements of 10%.

Genetic evidence for respiration. Aerobic respiration requires a functional electron transfer chain formed by cytochromes with oxygen as a terminal electron acceptor, resulting in the production of ATP (17). In E. coli, two terminal oxidases may be produced when cells are grown under aerobic conditions. Cytochrome bo, encoded by the cyoABCDE operon, is induced under oxygen-rich conditions, and cytochrome bd, encoded by the cydAB operon, is induced in stationary phase or under oxygen-limited conditions (6). The above results suggested that respiratory-like growth in L. lactis occurs preferentially in late logarithmic phase. Furthermore, L. lactis may grow under oxygen-poor conditions in nature (14). We thus hypothesized that cytochrome bd, which is active under oxygen-limited conditions, may be present.

View larger version (15K): [in this window] [in a new window]   FIG. 4.   Effects of cydA and sodA mutations on L. lactis long-term survival under respiration conditions. MG1363 and cydA and sodA mutants were grown at 30°C as described in the legend to Fig. 2 under Ox/H (closed circles), Ox/No H (open circles; for MG1363 and sodA strains), and No Ox/No H (open squares) conditions. After 24 h, cultures were transferred to 4°C. Cell viability was determined by plating dilutions on solid M17-glu at the indicated times. Experiments were performed at least twice; the results of a representative experiment are shown.

cydA expression is induced in the late growth phase under respiration conditions. Our results indicate that respiratory growth is a late-exponential-phase event. We examined whether cydA expression is growth regulated in L. lactis. For this purpose, Northern blotting was performed with mRNA samples extracted at different times during growth from cells cultured under aeration or respiration conditions (Fig. 5). RNA concentrations were monitored using a 16S rDNA probe (data not shown). The cydA-specific probe revealed the presence of two cydA transcripts. As a potential cydA-cydB operon was revealed by diagnostic sequencing of L. lactis strain IL-1403 (4), we consider it likely that the two transcripts correspond to cydA mRNA and the bicistronic mRNA. The cydA transcripts were detected in all samples and under both growth conditions. However, the expression of cydA mRNA was induced under respiration conditions after ~6 h of growth (corresponding to an OD₆₀₀ of 4.1) but not under aeration conditions. This late induction of cydA is in keeping with our results suggesting that respiration is favored during the late growth phase. Nevertheless, the detection of cydA transcripts under other growth conditions may indicate that cytochrome activity is limited by the absence of a necessary cofactor; heme uptake also may be a growth-phase-regulated event.

View larger version (54K): [in this window] [in a new window]   FIG. 5.   Expression of cydA is induced under Ox/H conditions in late-exponential-phase cells. L. lactis strain MG1363 was grown under aeration and respiration conditions (see Materials and Methods), and samples were removed for total RNA extraction at different OD600s during growth (times, from left to right, correspond to approximately 3, 4, 6, and 8 h after the start of growth). Northern blot analysis was performed using an internal cydA fragment as a probe. Membranes were subsequently dehybridized and rehybridized with a 16S rDNA probe to verify that the amounts deposited in the wells were identical. Northern blotting was performed three times and yielded results similar to those shown here.

cydA is directly involved in L. lactis oxygen uptake. Respiration involves oxygen consumption from a medium. It was previously shown that oxygen is consumed by L. lactis and that oxygen uptake is abolished in the presence of KCN, a respiration inhibitor (34). If cytochrome d oxidase is required for respiration, then a cydA mutant would be expected to be deficient for oxygen uptake. To address this question, we compared oxygen depletion from a medium containing either wild-type or cydA resting cells that had been cultured under aeration conditions in the presence or absence of hemin (Fig. 6). Note that this experiment is designed to determine relative oxygen consumption between strains in a resting state and is not a measure of respiration rate. Rapid oxygen depletion from the medium was observed for wild-type cells that had been grown under Ox/H conditions but not under Ox/No H conditions. In contrast, no oxygen depletion was observed for cydA cells regardless of the growth conditions. These results show that cydA is required for oxygen utilization in L. lactis.

View larger version (19K): [in this window] [in a new window]   FIG. 6.   Heme-dependent oxygen consumption by L. lactis is abolished in the mutant cydA strain. Aerated cultures of MG1363 and the cydA mutant, grown with or without hemin (H), were harvested after overnight growth. Oxygen consumption was evaluated with cells resuspended in saline buffer containing 1% glucose (see Materials and Methods). Oxygen remaining in the medium was measured using an Oxymeter. Ox/H and Ox/No H conditions are represented by closed and open symbols, respectively. Duplicate samples showed less than a 10% difference in measurements. Experiments were performed twice and yielded comparable results.

Ferrochelatase activity in L. lactis. Respiratory growth requires the addition of a heme compound to the aerated medium. We observed that the addition of the heme precursor PPIX confers effects equivalent to those conferred by hemin on the growth and survival of L. lactis. As PPIX allows respiratory growth, we postulated that L. lactis must express ferrochelatase, which charges PPIX with reduced iron to form heme. Degenerate primers were used to amplify from L. lactis strain MG1363 a 390-bp fragment that, as determined by sequencing, corresponds to a hemZ (ferrochelatase-encoding) gene. The fragment was used to construct a hemZ-disrupted mutant of MG1363. The hemZ-disrupted mutant was capable of respiratory-like growth upon heme addition to the medium. However, respiration did not occur when PPIX was added. These results show that at least the last gene of the heme biosynthesis pathway is functional in L. lactis. They also suggest that L. lactis has transport systems that allow protoporphyrin and iron uptake.

SodA is not required for long-term survival under respiration conditions. Superoxides are a major cause of stationary-phase mortality; for this reason, SODs are needed for bacterial survival in aerobic conditions (e.g., see reference 2). We examined the role of the unique SOD, SodA, in L. lactis. As expected, the sodA mutant survived very poorly under Ox/No H conditions (Fig. 4); only ~10⁴ to 10⁶ cell survivors were present in a 24-h aerated culture. In striking contrast, the sodA mutant behaved like the wild-type strain under respiration conditions; 50% of the cells were viable after 20 days of storage. Thus, in L. lactis, SodA is required for survival under aeration conditions but not under respiration conditions. These results support our observations that heme has considerable effects on L. lactis physiology, such that cells are less susceptible to oxidative stress and hence do not require SodA for survival when grown in the presence of a heme compound.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Capacity of a fermenting anaerobe to undergo respiration. The fermentative metabolism of L. lactis has been intensively studied, mainly due to its industrial relevance. Despite a past report suggesting that L. lactis can respire when oxygen and hemin are available (34), this observation has not been further documented since that time. We provide genetic and additional biochemical evidence for respiration and demonstrate for the first time that lactococci can be better adapted for growth and survival under these conditions than under fermentation conditions. The strong impact of respiration on lactococcal growth and long-term survival suggests that this mode of growth may be significant to lactococcal biology and should be further examined.

Respiration is a late-phase event. Several lines of evidence indicate that lactococcal metabolism is biphasic under respiration conditions, as tested here, growing first mainly via fermentation and then via respiration. (i) Lactic acid accumulation and a pH decline are observed during the first ~7 h of growth under respiration conditions; similar observations were previously described for an L. lactis subsp. diacetylactis strain, although respiration was ruled out (21). (ii) The cydA gene, which is required for respiration, is induced transcriptionally in late exponential phase in cells grown under respiration conditions. In E. coli, the corresponding cydA gene is also expressed in stationary phase (6). L. lactis genome analysis (4) has revealed the presence of cydA, cydB, cydC, and cydD open reading frames (the latter two are required for cytochrome d oxidase assembly [32]), while the cyoABCDE genes (required for cytochrome o oxidase synthesis) are absent. Note that although cydA transcription is induced late in Ox/H growth, transcripts are detected throughout growth. This finding suggests that other factors, e.g., hemin uptake, may limit cytochrome activity in early exponential phase.

Respiration confers long-term survival on L. lactis. The respiratory lifestyle of lactococci results in a greater growth yield and a remarkable improvement in survival (up to 10⁸-fold) compared to that under fermentation conditions. These results indicate a strong link between respiratory metabolism and long-term survival. Studies with E. coli have revealed the importance of numerous functions associated with long-term survival, several of which are involved in defense against oxygen (26, 32, 33). For example, the cydCD genes, which are needed for cytochrome d oxidase formation, also affect long-term survival or exit from stationary phase (32, 33). In lactococci, improved long-term survival after growth under respiration conditions requires an active cydA gene. We suggest that cytochrome d oxidase activity may serve as an oxygen trap to reduce oxygen toxicity and may also be involved in the activation of other survival genes. This hypothesis may explain why lactococci are generally observed to have poor long-term survival under conventional fermentation conditions (10), where heme and oxygen are absent.