Characterization of Three Lactic Acid Bacteria and Their Isogenic ldh Deletion Mutants Shows Optimization for Y_ATP (Cell Mass Produced per Mole of ATP) at Their Physiological pHs

Abstract

Several lactic acid bacteria use homolactic acid fermentation for generation of ATP. Here we studied the role of the lactate dehydrogenase enzyme on the general physiology of the three homolactic acid bacteria Lactococcus lactis, Enterococcus faecalis, and Streptococcus pyogenes. Of note, deletion of the ldh genes hardly affected the growth rate in chemically defined medium under microaerophilic conditions. However, the growth rate was affected in rich medium. Furthermore, deletion of ldh affected the ability for utilization of various substrates as a carbon source. A switch to mixed acid fermentation was observed during glucose-limited continuous growth and was dependent on the growth rate for S. pyogenes and on the pH for E. faecalis. In S. pyogenes and L. lactis, a change in pH resulted in a clear change in Y_ATP (cell mass produced per mole of ATP). The pH that showed the highest Y_ATP corresponded to the pH of the natural habitat of the organisms.

Comparative analyses, as demonstrated by comparative genomics and bioinformatics, are extremely powerful for (i) transfer of information from (experimentally) well-studied organisms to other organisms and (ii) when coupled to functional and phenotypic information, insight into the relative importance of components to the observed differences and similarities. The central principle is that important aspects of the functional differences between organisms derive not only from the differences in genetic components (which underlies comparative genomics) but also from the interactions between their components. Although this type of analysis is much discussed, only a very few studies focus on cross-species comparisons.

Here we study three relatively simple and highly related lactic acid bacteria (LAB) which nevertheless exhibit stark and important differences in their functional relationship with humans: these organisms are homofermentative lactic acid bacteria, namely, Lactococcus lactis, the major microorganism used in the dairy industry (21); Enterococcus faecalis, a major LAB in the human intestinal microbiota and a (fecal) contaminant in food and water as well as a contributor to food fermentation (16); and Streptococcus pyogenes, an important human pathogen (9, 15). These organisms have a similar primary metabolism but persist in completely different environments (milk, feces, skin/mucous membranes/blood).

L. lactis is by far the best-studied lactic acid bacterium (3, 13, 14, 23, 24, 30), and a kinetic model for its complete glycolysis, including some branching pathways, has been developed (2, 10). For the other two lactic acid bacteria, less information is readily available.

Generally, observations made with L. lactis are quickly translated to other LAB. With this approach we are able to separate organism-specific observations from observations that are general for LAB.

Here three LAB and their isogenic ldh deletion strains were characterized with respect to growth rates, catabolic flux distribution, ATP demand, and their ability to utilize different carbon sources. Our data identified important differences in the physiologies of these three LAB.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

L. lactis NZ9000 and the lactate dehydrogenase (LDH)-deficient strain NZ9010 (11, 17), E. faecalis V583 and V583 Δldh-1 (12), and S. pyogenes M49 591 and M49 591 Δldh were grown in batch cultures at 37°C in 96-well plates in either Todd-Hewitt broth supplemented with 0.5% (wt/vol) yeast extract (Oxoid) (THY medium) or a chemically defined medium (CDM) specifically designed to support the growth of all three LAB (pH 7.4). The CDM-LAB medium (12) contained the following per liter: 1 g K₂HPO₄, 5 g KH₂PO₄, 0.6 g ammonium citrate, 1 g acetate, 0.25 g tyrosine, 0.24 g alanine, 0.125 g arginine, 0.42 g aspartic acid, 0.13 g cysteine, 0.5 g glutamic acid, 0.15 g histidine, 0.21 g isoleucine, 0.475 g leucine, 0.44 g lysine, 0.275 phenylalanine, 0.675 g proline, 0.34 g serine, 0.225 g threonine, 0.05 g tryptophan, 0.325 g valine, 0.175 g glycine, 0.125 g methionine, 0.1 g asparagine, 0.2 g glutamine, 10 g glucose, 0.5 g l-ascorbic acid, 35 mg adenine sulfate, 27 mg guanine, 22 mg uracil, 50 mg cystine, 50 mg xanthine, 2.5 mg d-biotin, 1 mg vitamin B₁₂, 1 mg riboflavin, 5 mg pyridoxamine-HCl, 10 μg p-aminobenzoic acid, 1 mg pantothenate, 5 mg inosine, 1 mg nicotinic acid, 5 mg orotic acid, 2 mg pyridoxine, 1 mg thiamine, 2.5 mg lipoic acid, 5 mg thymidine, 200 mg MgCl₂, 50 mg CaCl₂, 16 mg MnCl₂, 3 mg FeCl₃, 5 mg FeCl₂, 5 mg ZnSO₄, 2.5 mg CoSO₄, 2.5 mg CuSO₄, and 2.5 mg (NH₄)₆Mo₇O₂₄. Both media were buffered with either 100 mM morpholineethanesulfonic acid buffer or 100 mM morpholinepropanesulfonic acid buffer for growth at pH 6.5 and 7.5, respectively. Cultures were grown statically to provide anaerobic (or rather low microaerophilic) conditions (7). Group A streptococcal (GAS) mutants harboring recombinant pUC18Erm1 plasmids (4) were maintained in medium containing 5 mg/liter erythromycin or 60 mg/liter spectinomycin. Escherichia coli DH5α isolates transformed with pUC18Erm1 derivatives were grown on Luria-Bertani broth supplemented with 300 mg/liter erythromycin and/or 100 mg/liter spectinomycin. All E. coli cultures were grown at 37°C under ambient air conditions.

Chemostat cultures.

L. lactis NZ9000, E. faecalis V538, and S. pyogenes M49 591 wild-type strains and their ldh-negative mutants were grown in anaerobic glucose-limited chemostat cultures in CDM-LAB medium (12). Cultures were grown in a Biostat Bplus fermentor unit with a total volume of 750 ml for S. pyogenes at a stirring rate of 100 rpm. L. lactis and E. faecalis were grown in Applikon-type fermentors at a stirring rate of 400 rpm in a total culture volume of 1,000 ml. The temperature was kept at 37°C for all three organisms.

The pH was maintained at the indicated value by titrating with sterile 2 M NaOH. Growth rates were controlled by the medium dilution rates (D; 0.05 h⁻¹ or 0.15 h⁻¹). Culture volume was kept constant by removing culture liquid at the same rate that fresh medium was added. The cultures were considered to be in steady state when no detectable glucose remained in the culture supernatant and the optical densities (ODs), dry weights, and product concentrations of the cultures were constant on two consecutive days. All chemostat results showed a carbon balance of 80% ± 10% on the basis of glucose consumption and organic acid formation. This concurs with previously published data obtained from continuous cultures (13).

Construction of recombinant vectors and GAS strains.

For the construction of an S. pyogenes M49 591 ldh-knockout strain, a 2,977-bp fragment comprising the l-lactate dehydrogenase gene (ldh) and 1,000 bp of the upstream and 993 bp of the downstream flanking sequences was PCR amplified from chromosomal DNA of GAS M49 591 using the forward/reverse primer pair 5′-CAC TTG AGC TCT ATT GAC GCC ATA GGG AAA-3′/5′-CCA ACG CAT GCG CAA AGA AGT GGT TCT GAT-3′. The resulting PCR fragment was digested with SacI and SphI and ligated into the equally treated pUC18Erm1 vector (4). The resulting plasmid was used as a template for an outward PCR with primer pair 5′-TAA TCG GAT CCG AGA CTT CGG TCT CTT TTT-3′/5′-AGT GCA GTC GAC TCT AAA CAT CTG CTT AAT-3′ binding to the flanking regions. Thus, the resulting PCR product comprised the whole plasmid, including the upstream and downstream flanking regions of the ldh gene but excluding the ldh gene itself. After restriction of this fragment with BamHI and SalI, it was ligated with an equally treated PCR fragment (primer pair 5′-GGC GGC GTC GAC TTG ATT TTC GTT CGT GAA TAC ATG-3′/5′-GGC GGC GGA TCC CCA ATT AGA ATG AAT ATT TCC CAA A-3′) comprising the spectinomycin resistance gene aad9 from plasmid pSF152 (26). The resulting recombinant plasmid, pUCerm-ldh-ko, was transformed into S. pyogenes M49 591, and assays for double-crossover events were performed by selection for erythromycin-sensitive but spectinomycin-resistant transformants. The correct replacement of the ldh gene by the aad9 gene in the respective transformants was confirmed by appropriate PCR assays and LDH activity assays. For all PCR amplifications, a Phusion high-fidelity PCR kit (Finnzymes) was used. For L. lactis and E. faecalis, the ldh deletion strains NZ9010 Δldh-1 and V538 Δldh-1 respectively, have been published previously (11, 12).

Analysis of carbon fluxes.

Steady-state bacterial dry weight was measured as described previously (1). Glucose, pyruvate, lactate, formate, acetate, succinate, and ethanol were determined by high-pressure liquid chromatography (HPLC; LKB) with a Rezex organic acid analysis column (Phenomenex) at a temperature of 45°C with 7.2 mM H₂SO₄ as the eluent, using a RI 1530 refractive index detector (Jasco) and AZUR chromatography software for data integration. Discrimination between d- and l-lactate was performed using a d-/l-lactate assay kit (Megazyme).

Aspartic acid, serine, glutamic acid, glycine, histidine, arginine, threonine, alanine, proline, cysteine, tyrosine, valine, methionine, lysine, isoleucine, leucine, and phenyalanine were determined by HPLC (Agilent) by use of the Waters AccQ Tag method. Fluorescence was analyzed using a Hitachi F-1080 fluorescence detector set to 250 nm excitation, and emission was recorded at 395 nm.

Substrate utilization assays.

For substrate utilization assays, bacteria were grown overnight in a chemically defined medium (12), pelleted by centrifugation, washed twice in phosphate-buffered saline (pH 7.4), and suspended in glucose-free CDM-LAB medium. Optical densities were adjusted to 0.05, and 100 μl bacterial suspension was applied to each well of Biolog phenotype microarray plates PM1 and PM2. The microarray plates were incubated for 24 h at 37°C in a 5% CO₂ atmosphere, and the optical densities of each well were measured. The optical densities in well A1 of the arrays containing no carbon source were subtracted from all values. Optical densities in the wells containing α-d-glucose were set equal to 100%, and all other values were related accordingly.

Calculation of specific ATP synthesis rates.

The rate of substrate-level ATP synthesis [q_{ATP (SLP)}] is stoichiometrically coupled to the rate of lactate, acetate, and ethanol synthesis as follows: 1 glucose + 2ADP + 2P_i → 2 lactate + 2ATP and 1 glucose + 3ADP + 3P_i → 1 acetate + 1 ethanol + 3ATP. The energy required for maintenance (qATP_maintenance) was estimated by extrapolation of the linear line of D plotted against the total energy (qATP_total) to D equal to 0 (see Fig. S1 in the supplemental material for an example). The qATP at the maximal specific growth rate (qATPμ_max) was estimated by extrapolating the same line to the D at which the specific organism has its maximal specific growth rate. This method is adapted from previously published methods (5, 8, 19, 27). Here we assumed a constant qATP_maintenance, since (i) qATP_maintenance is very small and would show only a small contribution to the qATP at μ_max and (ii) no consensus on the calculation of qATP_maintenance at μ_max exists (27).

RESULTS

Deletion of ldh does affect growth rate of lactic acid bacteria in rich medium but not in CDM-LAB medium.

L. lactis, E. faecalis, and S. pyogenes are referred to as LAB because of the fact that in the presence of glucose, lactate is produced as the main fermentation product. This metabolic pathway is relatively inefficient, since only two ATP molecules are generated from one glucose molecule (Fig. 1). All three LAB possess the genetic make up for mixed acid fermentation (6, 18, 20, 28), a more effective way of fermentation generating three ATP molecules per molecule of glucose (Fig. 1). All three genomes reveal (at least) two genes encoding an LDH. S. pyogenes possesses one (l-LDH) (18), E. faecalis two (l-LDH, l-LDH2) (20), and L. lactis three (l-LDHA, l-LDHB, l-LDHX) (21) l-lactate dehydrogenases. E. faecalis and S. pyogenes each encode one additional d-lactate dehydrogenase (d-LDH). In both L. lactis and E. faecalis it has been shown that the l-LDH is responsible for over 95% of total lactate synthesis (7, 12). This could also be confirmed for S. pyogenes in the present study (data not shown).

FIG. 1.

Schematic representation of the anaerobic pathways of glucose catabolism in LAB.

In all three LAB, the main ldh gene (encoding the l-LDH responsible for over 90% of the total lactate flux in the wild-type strains) was removed (11, 12). The resulting ldh deletion strains were analyzed in a batch growth setup in CDM-LAB or rich THY medium at pH 6.5 and pH 7.5 under low microaerobic conditions. As expected, all wild-type strains performed complete homolactic acid fermentation under all conditions. None of the three ldh deletion strains showed a significant difference with respect to growth rate compared to that of the wild-type counterpart (Table 1) in CDM-LAB medium, except for E. faecalis grown at pH 6.5.

TABLE 1.

Maximal specific growth rates of the three lactic acid bacteria and their ldh deletion mutants^a

Medium	pH	Maximum specific growth rate (h−1) of strain:
		L. lactis		E. faecalis		S. pyogenes
		NZ9000	NZ9010	V583	V583 Δldh-1	M49 591	M49 591 Δldh
CDM-LAB	6.5	0.43 ± 0.06	0.45 ± 0.06	0.74 ± 0.01	0.63 ± 0.01b	0.43 ± 0.01	0.39 ± 0.04
	7.5	0.49 ± 0.01	0.49 ± 0.06	0.79 ± 0.07	0.89 ± 0.07	0.39 ± 0.02	0.35 ± 0.01
THY	6.5	0.76 ± 0.05	0.59 ± 0.05b	1.14 ± 0.05	0.99 ± 0.05b	0.86 ± 0.13	0.69 ± 0.01b
	7.5	0.81 ± 0.09	0.72 ± 0.01	1.16 ± 0.02	1.01 ± 0.02b	0.57 ± 0.06	0.53 ± 0.03

^aStrains were grown in 96-well plates at 37°C under low microaerobic conditions. Values indicate the average μ_max ± standard deviation.

^bSignificantly different, P < 0.05 (two-tailed Mann-Whitney U test).

In THY medium at both pHs, all three wild-type LAB showed higher growth rates than their isogenic counterparts with ldh deletions, although not to the extent observed previously for L. lactis in MRS medium (7). Only small pH-dependent differences in the maximal growth rates were observed for wild-type L. lactis and E. faecalis. S. pyogenes, however, showed a significantly lower specific growth rate at pH 7.5 in rich medium but only a slightly lower specific growth rate at pH 7.5 in CDM-LAB medium. All strains with ldh deletions grew 10 to 20% slower than the wild-type strains, except for S. pyogenes at pH 7.5, where deletion of ldh did not result in a significant decrease in growth. In late stationary phase, all three ldh deletion strains grew to a higher optical density and the activity of LDH remained below 10% of total fermentation activity for all the three ldh deletion strains (data not shown).

To verify whether a similar μ_max also signified that in a mixed culture deletion of ldh does not represent a disadvantage, the S. pyogenes M49 591 wild type and its ldh deletion strain were cocultivated in THY medium (pH 7.5). A 52%/48% distribution of wild-type and mutant bacteria was shown after 18 h of cocultivation of both strains in THY medium and subsequent plating of serial dilutions on THY agar plates with and without spectinomycin. This indicates that the lack of an l-LDH represented no significant disadvantage to the organism under the conditions tested.

Effect of deletion of ldh in LAB on substrate utilization.

The decrease in maximal specific growth rate of the ldh deletion in rich medium might be due to differences in the ability to utilize carbon sources other than glucose. To assess the impact of the ldh knockout on the ability of LAB to utilize different substrates as a carbon source, Biolog phenotype microarrays were applied. Using these arrays, growth of the strains on 190 different carbon substrates was evaluated. Comparison of the substrate utilization of the three strains and their isogenic ldh deletion strains showed that there were 11 carbon sources on which all wild-type strains were able to grow to at least 10% of the optical densities reached by growth on glucose (for complete lists, see Tables S1 to S3 in the supplemental material). These substrates were α-d-glucose, d-mannose, maltose, maltotriose, N-acetyl-d-glucosamine, d-fructose, d-trehalose, d-glucosamine, sucrose, salicin, and dextrin. However, S. pyogenes showed optimal growth on glucose and sucrose (101.1%) (Table 2). With all the other C sources tested, S. pyogenes ended up at lower ODs after 24 h of growth. For E. faecalis there was no carbon source that led to an equal or even better growth yield compared to that achieved with glucose. In contrast, the growth yield of L. lactis was the same or improved compared to that with glucose on gentiobiose (128.3%) and d-cellobiose (114.5%).

TABLE 2.

Substrate utilization of S. pyogenes M49 591, E. faecalis V583, and L. lactis NZ9000 and their ldh deletion mutants^a

Substrate	Final optical densities of strains compared to that with glucose substrate (%)
	L. lactis		E. faecalis		S. pyogenes
	NZ9000	NZ9010	V583	V583 Δldh-1	M49 591	M49 591 Δldh
Gentiobiose	128.3 ± 12.3	98.3 ± 12.5b	63.3 ± 8.9	70.5 ± 4.1	1.4 ± 5.0	5.7 ± 7.9
d-Cellobiose	114.5 ± 6.3	67.8 ± 13.4b	68.4 ± 1.9	72.4 ± 0.9b	6.2 ± 6.9	6.9 ± 11.2
d-Trehalose	124.4 ± 26.5	42.9 ± 41.2b	68.8 ± 4.5	69.9 ± 0.9	90.4 ± 16.5	38.4 ± 9.8b
Sucrose	23.2 ± 35.2	6.5 ± 8.5	45.8 ± 2.6	46.0 ± 2.9	101.1 ± 20.4	38.6 ± 11.3b
Maltose	72.5 ± 9.6	35.8 ± 18.9b	93.5 ± 2.2	96.7 ± 5.3	52.3 ± 14.1	40.5 ± 15.2
Maltotriose	107.9 ± 21.6	70.3 ± 30.8	85.2 ± 13.0	102.8 ± 3.5b	87.4 ± 16.5	78.9 ± 9.0
d-Glucosaminic acid	3.3 ± 8.1	2.7 ± 2.4	12.8 ± 3.1	9.1 ± 1.0b	0.6 ± 1.6	1.4 ± 2.5
d-Mannose	117.8 ± 26.3	105.5 ± 5.7	93.8 ± 6.8	101.3 ± 9.1	82.0 ± 17.7	41.3 ± 14.7b

^aOut of the 190 tested carbon sources, only those with significant differences between at least one mutant and wild-type pair are shown. Optical densities of the cultures grown on glucose were set equal to 100% for all strains, and optical densities for growth on all other substrates were related to this value.

^bSignificantly different, P < 0.05 (two-tailed Mann-Whitney U test).

The deletion of the ldh gene in S. pyogenes resulted in a significantly reduced growth yield of this strain on d-mannose (−49.6%), d-trehalose (−57.5%), and sucrose (−61.8%) as the carbon source in comparison to that of wild-type S. pyogenes (Table 2). For L. lactis the deletion of the ldh gene also resulted in hampered growth yield on d-trehalose (−65.5%), d-cellobiose (−46.8%), maltose (−50.7%), and d-gentiobiose (−23.4%). Almost no significant changes in the substrate utilization of the E. faecalis ldh knockout strain were observed, but a small significant reduction of the growth yield on d-glucosaminic acid (−29.2%) was detected, although the wild-type E. faecalis strain also showed small growth on this substrate (12.1% compared to that on glucose). The ldh knockout led to a small improvement of the growth yield on maltotriose (+20.7%) and d-cellobiose (+5.8%) compared to that of the E. faecalis wild-type strain, which is probably the result of increased efficiency of ATP formation from pyruvate in the ldh deletion strain. For the other substrates tested, no significant differences were observed.

ATP demand under glucose-limited continuous growth conditions is strongly pH and organism dependent.

ATP demand can be estimated by performing growth in continuous cultures and subsequently allows calculation of the qATP_total from the formed fermentation products (see Materials and Methods and Fig. S1 in the supplemental material). This allows estimation of the qATPμ_max and qATP_maintenance. Furthermore, this could give indications on the role of qATP in the pyruvate flux distribution.

In order to determine the pH and growth rate dependency of the flux distribution (i.e., homolactic acid, acetate, ethanol, acetoin, and butanediol formation) and the cell mass (g) produced per mol of ATP generated by substrate catabolism (Y_ATP) of the LABs under defined continuous conditions, all three strains were grown as anaerobic glucose limited chemostat cultures in CDM-LAB medium under conditions that varied in growth rate and pH. Glucose limitation was verified by HPLC analysis and by a linear correlation between changes of the glucose concentration in the medium and cell density; i.e., a 2-fold decrease in the glucose concentration resulted in a 2-fold decrease in biomass (data not shown). None of the 17 amino acids except arginine (data not shown) was completely consumed for all three organisms. Arginine was consumed completely. Under these energy-limited growth conditions, this is likely due to use of arginine for ATP formation by formation of ornithine.

E. faecalis mainly showed mixed acid fermentation under all conditions, while L. lactis and S. pyogenes mainly exhibited homolactic acid fermentation (Table 3). Only S. pyogenes showed more mixed acid fermentation at lower dilution rates at both pH 6.5 and pH 7.5. E. faecalis and L. lactis did not show a significant growth rate-dependent change in fermentation pattern at these growth rates. Mixed acid fermentation did show a strong pH dependency for E. faecalis, with a more homolactic acid fermentation phenotype occurring at pH 6.5. For L. lactis and S. pyogenes, no significant pH-dependent differences were observed.

TABLE 3.

Relative flux distribution in the three lactic acid bacteria at two dilution rates and two pHs during continuous cultivation in glucose-limited CDM-LAB medium^a

Strain	Dilution rate (h−1)	pH	Lactate (mol/mol glucose)	Formate (mol/mol glucose)
L. lactis NZ9000	0.05	6.5	1.0 ± 0.3	0.6 ± 0.3
		7.5	1.1 ± 0.5	0.2 ± 0.2
	0.15	6.5	1.4 ± 0.4	0.4 ± 0.1
		7.5	1.5 ± 0.3	0.2 ± 0.1
E. faecalis V583	0.05	6.5	0.9 ± 0.3	0.4 ± 0.3
		7.5	0.2 ± 0.1	1.2 ± 0.1
	0.15	6.5	0.7 ± 0.1	0.7 ± 0.1
		7.5	0.2 ± 0.1	1.1 ± 0.1
S. pyogenes M49 591	0.05	6.5	0.8 ± 0.4	0.5 ± 0.1
		7.5	0.6 ± 0.3	0.5 ± 0.2
	0.15	6.5	1.4 ± 0.1	0.1 ± 0.1
		7.5	1.1 ± 0.3	0.2 ± 0.2

^aValues indicate mol product/mol glucose ± standard deviation.

L. lactis NZ9010 with the single ldh deletion showed increased activity of alternative LDH proteins, as was observed previously (7) and as was shown by an increase in homolactic acid fermentation during prolonged growth. Deletion of the main ldh of E. faecalis and S. pyogenes resulted in complete mixed acid fermentation under all conditions (data not shown) and the qATP_total values were similar to those for the cognate wild-type strains (data not shown). This indicates that deletion of ldh does not result in an overall increase in ATP-dissipating reactions.

qATP_maintenance and qATPμ_max (Table 4) were calculated as described above (see Fig. S1 in the supplemental material). qATP_maintenance did not show large pH- or species-dependent differences for any of the three LAB (Table 4), although in general the qATP_maintenance for L. lactis was found to be higher than that for the other two LAB. However, large differences were observed with respect to Y_ATP (27). The Y_ATP for S. pyogenes was almost 2-fold higher at pH 7.5 than at pH 6.5, while for L. lactis the Y_ATP at pH 7.5 was almost 2-fold lower that that at pH 6.5. For Enterococcus faecalis no significant pH dependence of Y_ATP was observed.

TABLE 4.

Physiological parameters of LAB grown in C-limited continuous cultures^a

Strain	pH	qATPmaintenance	YATP	qATPμmax
L. lactis NZ9000	6.5	5.0 ± 1.5	8.4 ± 1.0	33 ± 2.9
	7.5	8.2 ± 3.4	4.4 ± 0.5	94 ± 15
E. faecalis V583	6.5	2.0 ± 1.2	12.8 ± 2.6	49 ± 3.3
	7.5	2.0 ± 0.8	14.2 ± 3.3	47 ± 2.7
S. pyogenes M49 591	6.5	−2.6 ± 2.2	5.2 ± 0.7	88 ± 9.2
	7.5	2.9 ± 2.1	9.4 ± 2.3	38 ± 7.1

^aqATP_maintenance was calculated according to the methods applied by Tempest and Neijssel (27). Y_ATP was determined at a D of 0.15 since Y_ATP at low dilution rates is strongly influenced by qATP_maintenance. qATP at the maximal specific growth rate (qATPμ_max; for data on μ_max, see Table 1) was estimated by extrapolation of the slope for qATP_total to D, similar to μ_max.

DISCUSSION

Here we have studied the general physiological characteristics of three well-known LAB and their isogenic ldh deletion strains. Previously, Bongers et al. (7) reported that deletion of the main ldh in L. lactis did affect the μ_max under anaerobic but not under aerobic conditions in the rich M17 broth. In accordance with those data, we observed that in the rich THY medium the lack of the LDH enzyme also resulted in a reduced maximal growth rate at pH 6.5 for L. lactis and also for S. pyogenes and E. faecalis. In contrast to that, in a buffered chemically defined medium supplied with glucose as the main carbon source, deletion of ldh did not result in significantly lower growth rates for all three LAB under anaerobic/microaerobic conditions.

Thus, the effects of ldh deletions on growth rates of lactic acid bacteria apparently depend not only on oxygen availability, as shown by Bongers et al. (7) for L. lactis, but also on the medium and pH. A major difference between M17, THY, and CDM-LAB media is the availability of carbon sources. While the complex rich media M17 and THY contain a variety of different potential carbon sources, in CDM-LAB medium, glucose is the only sugar component. It has been shown previously that utilization of sugars like maltose or galactose as the carbon source by LAB led to changes in product formation compared to that from growth on glucose (24, 25, 29). By screening a large variety of carbon sources, we could show that the ldh deletion mutants are not able to utilize all carbon sources as efficiently as their cognate wild types. This might contribute to the phenotypic differences observed during growth on rich medium. It seems that for growth on disaccharides, i.e., gentiobiose, d-cellobiose, d-trehalose, maltose, and sucrose, deletion of ldh in L. lactis results in impaired growth yields, while this is not observed in E. faecalis and only to a minimal extent in S. pyogenes, i.e., for d-trehalose and sucrose. Of note, all di- or trisaccharides linked by a β1,4-glucoside bond do not result in lower growth rates for S. pyogenes with the ldh deletion. Sugars that are linked by an α1,1-glucoside or α1,2-glucoside bond do result in strong retardation of growth yield upon deletion of ldh in S. pyogenes. However, the exact reasons for these differences remain unclear.

E. faecalis clearly showed the highest specific growth rate in both chemically defined medium and rich medium. L. lactis and S. pyogenes showed roughly similar growth rates under all conditions, with the exception that S. pyogenes grew slower in both media at pH 7.5.

The three lactic acid bacteria showed stark differences in growth under glucose-limited continuous conditions. E. faecalis showed mixed acid fermentation under almost all conditions but showed clearly more homolactic acid fermentation at low pH. Both S. pyogenes and L. lactis mainly showed homolactic acid fermentation, as was shown previously (24), under all conditions and showed no pH dependency with respect to mixed acid versus homolactic acid fermentation. This is in contrast to previous observations, caused by the strong differences in the amount of arginine in the medium (28) (data not shown). These data also indicate that the switch to mixed acid fermentation by these organisms is not caused by a decrease of their (relative) growth rate. This is especially exemplified by the clear pH dependence of the fraction of mixed acid fermentation at a D of 0.15 for E. faecalis, since the maximal specific growth rate does not show any pH dependency.

E. faecalis did not show a pH dependence of its Y_ATP, and the data shown here correlate very well with previously published data that showed a qATP_maintenance of 2 and a Y_ATP of about 13 (22) for glucose-limited continuous cultures grown at pH 7.0. Both S. pyogenes and L. lactis showed clear pH dependence with respect to their Y_ATP values, with each having the highest Y_ATP if growth is performed at or near the pH of their physiological environment, i.e., pH 6.5 for L. lactis (milk) and pH 7.5 for S. pyogenes (blood). This strongly indicates that the growth yields for these organisms are somehow optimized at their natural pH and quickly encounter (ATP-dissipating) difficulties at alternative pHs. The exact reason for or mechanism behind the observed differences in Y_ATP are beyond the scope of this study. Interestingly, no major impact for pH on the μ_max was observed for these lactic acid bacteria (except for S. pyogenes in rich medium). It seems, therefore, that μ_max is less dependent on the pH than Y_ATP.

Combining the maximal specific growth rate, qATP_maintenance, and Y_ATP allows calculation of the qATP at μ_max (see Table 4 and Materials and Methods for the formula used). This shows that qATP_total at μ_max is much higher for L. lactis and S. pyogenes under unfavorable conditions at pH 7.5 and 6.5, respectively. This indicates that growth at their natural pHs, 6.5 and 7.5, respectively, is not limited by ATP formation rates under the conditions tested here. The differences observed between the three wild-type lactic acid bacteria and their isogenic ldh deletion strains were strongly strain dependent. This leads to the (obvious) indication that observations made for a single species belonging to the order Lactobacillales cannot be translated to the other species in this specific order. The fact that these organisms are so closely related ensures, however, that these types of studies can more easily zoom in on the specific phenotypic differences and the physicochemical causes thereof. It may even help resolve the long-standing question as to what is the basis of the homolactic acid fermentation to mixed acid fermentation switch, since it is expected that a similar mechanism regulates this in all three organisms.