Lactate dehydrogenase has no control on lactate production but has a strong negative control on formate production in Lactococcus lactis


P. R. Jensen, Section of Molecular Microbiology, BioCentrum, Technical University of Denmark, Building 301, DK-2800 Lyngby, Denmark. Tel.: +45 45 252510. Fax: +45 45 932809. E-mail: Peter.R.Jensen@BioCentrum.DTU.DK


A series of mutant strains of Lactococcus lactis were constructed with lactate dehydrogenase (LDH) activities ranging from below 1% to 133% of the wild-type activity level. The mutants with 59% to 133% of lactate dehydrogenase activity had growth rates similar to the wild-type and showed a homolactic pattern of fermentation. Only after lactate dehydrogenase activity was reduced ninefold compared to the wild-type was the growth rate significantly affected, and the ldh mutants started to produce mixed-acid products (formate, acetate, and ethanol in addition to lactate). Flux control coefficients were determined and it was found that lactate dehydrogenase exerted virtually no control on the glycolytic flux at the wild-type enzyme level and also not on the flux catalyzed by the enzyme itself, i.e. on the lactate production. As expected, the flux towards the mixed-acid products was strongly enhanced in the strain deleted for lactate dehydrogenase. What is more surprising is that the enzyme had a strong negative control ( inline image=−1.3) on the flux to formate at the wild-type level of lactate dehydrogenase. Furthermore, we showed that L. lactis has limited excess of capacity of lactate dehydrogenase, only 70% more than needed to catalyze the lactate flux in the wild-type cells.


lactate dehydrogenase

Lactococcus lactis plays an important role in dairy fermentations, mainly in the production of cheeses. In these fermentation processes, lactose is present at high concentrations (50 g·L−1) and is converted through glycolysis to lactic acid, with minor amounts of other compounds being produced in addition (homolactic fermentation). The resulting low pH contributes to the texture and flavor of cheeses and inhibits the growth of other bacterial species. Under conditions where sugar becomes limiting for growth of Lactococcus lactis, the metabolism shifts to mixed-acid products, i.e. formate, acetate and ethanol along with smaller amounts of lactate [1,2].

Work has been performed in the past to study the mechanisms involved in the shift between the two different fermentation modes in L. lactis. In the presence of excess sugar, the concentration of fructose 1,6-bisphosphate, the triose-phosphates, pyruvate, and the NADH/NAD+ ratio are high, whereas the concentration of phosphoenolpyruvate and inorganic phosphate are relative low [3–6]. In contrast, when sugar is limiting the concentration of these metabolites and cofactors are reversed to the opposite, high or low level. Particularly, the level of fructose-1,6-bisphosphate, which is known to activate both pyruvate kinase and lactate dehydrogenase, has been suggested to play a key role in the regulation of the fermentation mode [1].

Work has also been performed to determine the factors that control the flux through glycolysis by applying metabolic control analysis [7,8]. Based on inhibitor titration, Poolman et al. [9] suggested that glyceraldehyde 3-phosphate dehydrogenase had a large amount of control over the glycolytic flux during nongrowing conditions. However, it is still unclear whether the enzyme is important when the flux through glycolysis is high.

Recently, we showed that a mere twofold reduction of the activity of phosphofructokinase in L. lactis had a strong negative influence on the growth rate and the glycolytic flux [10]. Lactate dehydrogenase is the last enzyme in the pathway converting sugar to lactate in L. lactis. As expected, disruption of the ldh gene in L. lactis MG1363 [11] or in the derivative NZ3900 [12] diverts the majority of pyruvate towards mixed-acid products. However, it is still unclear to what extent this enzyme controls the metabolic fluxes in the wild-type L. lactis cells.

In this study, we use a new method for subtle and stable modulations of lactate dehydrogenase activity around wild-type levels in growing cells and we quantify the control exerted by lactate dehydrogenase on the growth rate and the metabolic fluxes in L. lactis. We show that lactate dehydrogenase has virtually no control on the growth rate, the glycolytic flux, and the flux to lactate, but the enzyme has a high negative control on the flux towards formate.

Materials and methods

Bacterial strains and plasmids

The bacterial strains and plasmids used in this study are listed in Table 1.

Table 1. Bacterial strains and plasmids. The feature of a plasmid is indicated first by the vector ligated ‘::’ to the insert. The restriction endonuclease used for digestion is shown. Kilobases (kb) and base pairs (bp) denote the sizes of inserts. The coordinates in brackets are the sequence number of the las operon accession L07920. Abbreviations: AmpR, ampicillin-resistance gene; ErmR, erythromycin-resistance gene; TetR, tetracyclin-resistance gene; MCS, multiple cloning sites.
Strain or plasmidRelevant characteristicsReference or source
L. lactis
 MG1363 L . lactis subsp. cremoris is a prophage-cured and plasmid-free derivative of 
 MBP12MG1363 derivative with ldh out of frameThis study
 MBP50MBP12 attB:: pHWA202 (ldh), ErmRThis study
 MBP51MBP12 attB:: pMBP34 (CP2-ldh), ErmRThis study
 MBP52MG1363 attB:: pHWA202 (ldh), ErmRThis study
 HWA236MG1363 attB:: pHWA227 (CP7-ldh), ErmRThis study
 HWA239MG1363 attB:: pHWA230 (CP29-ldh), ErmRThis study
 HWA242MG1363 attB:: pHWA231 (CP25-ldh), ErmRThis study
 HWA248MBP12 attB:: pHWA227 (CP7-ldh), ErmRThis study
 HWA251MBP12 attB:: pHWA230 (CP29-ldh), ErmRThis study
E. coli
 XLI-Blue MRF′Cloning host, restriction-minus cells, (mcrA)183 (mcrCB-hsdSMR-mrr)173 endA1
supE44 thi-1 recA1 gyrA96 relA1 lac[F′proAB lacIqZΔM15 Tn5 (Kanr)]
 ABLE KCloning host which lowers the plasmid copy number, C lac (LacZω) [KanrmcrA
mcrCB mcrF mrr hsdR(rκ)][F′proAB lacIqZΔM15 Tn10 (Tetr)]
 pMU2901pJDC9 harbouring pyk′- ldh on a 6.0-kb EcoRI fragment from L. lactis subsp. lactis 
LM0230, (1785 → ?), carrying the 3′end of pyk, the coding region of ldh and
further downstream of the las operon. Accession M95919, ErmR

 pUC18 E . coli cloning vector, pBR322 ori, MCS in lacZ′, AmpRNew England Biolabs
 pRCI E . coli subcloning vector, ErmR[20]
 pG+host4Derivative of pGK12 harbouring oriTS of the broad-host-range replicon pWV01
used for homologous recombination, ErmR
 pBF12αVector for site-specific integration in L. lactis
pMOSblue-T::332bp PCR of attP from the phage TP901-1, shuttle vector
between E. coli and L. lactis, AmpR, ErmR
 pLB95Derivative of pG+host8 harbouring oriTS of the broad-host-range replicon
pWV01, orf1 encoding the integrase of TP901-1, TetR
 pCP2pAK80 derivative carrying constitutive promoter CP2-lacLM, ErmR[22]
 pCP7pAK80 derivative carrying constitutive promoter CP7-lacLM, ErmR[22]
 pCP25pAK80 derivative carrying constitutive promoter CP25-lacLM, ErmR[22]
 pCP29pAK80 derivative carrying constitutive promoter CP29-lacLM, ErmR[22]
 pMBP2pUC18 HincII-HindIII::pMU2901 DraI-HindIII, 778 bp (+3026→+3804bp), AmpRThis study
 pMBP4pMBP2 EcoRI-SmaI::pMU2901 EcoRI-DraI, 1.24 kb (+1785→+3025bp), AmpRThis study
 pMBP6Integration vector to inactivate ldh
pG+host4 EcoRI-HindIII::pMBP4 EcoRI-HindIII, 2044 bp with Δldh, ErmR
This study
 pHWA78 ldh subcloned in pUC18
pUC18 SmaI::pMU2901HaeIII, 1.32 kb of ldh (+2838→?), AmpR
This study
 pHWA97 ldh subcloned pRC1
pRC1 EcoRI-SalI::pHWA78 EcoRI-SalI, 1.35 kb, ErmR
This study
 pHWA202 ldh cloned in the vector for site-specific integration in L. lactis
pBF12αXbaI::pHWA97 SpeI-XbaI, 1.37 kb, AmpR, ErmR
This study
 pMBP34CP2 in front of ldh cloned in the vector for site-specific integration in L. lactis
pHWA202 SalI-XbaI::pCP2 XhoI-XbaI, 88 bp, AmpR, ErmR
This study
 pHWA227CP7 in front of ldh cloned in the vector for site-specific integration in L. lactis
pHWA202 SalI-XbaI::pCP7 XhoI-XbaI, 74 bp, AmpR, ErmR
This study
 pHWA230CP29 in front of ldh cloned in the vector for site-specific integration in L. lactis
pHWA202 SalI-XbaI::pCP29 XhoI-XbaI, 89 bp, AmpR, ErmR
This study
 pHWA231CP25 in front of ldh cloned in the vector for site-specific integration in L. lactis
pHWA202 SalI-XbaI::pCP25 XhoI-XbaI, 103 bp, AmpR, ErmR
This study

Growth media and growth conditions

Escherichia coli strains were grown aerobically at 37 °C in Luria–Bertani broth [13]. L. lactis strains were cultivated as batch cultures in the Biostat Q fermentor system (B. Braun Biotech International, Melsunge, Germany) in 1-L vessels with a working volume of 600 mL of SA medium [14] supplemented with 0.25% glucose. The temperature was set to 30 °C, and the cultures were stirred at 200 r.p.m. The pH was maintained at 7.0 by the addition of 2 m NaOH. The medium in the fermentor was inoculated with an exponentially growing preculture to an initial D600 ranging from 0.02 to 0.04. Samples were withdrawn for measurements of biomass, fermentation products, and enzyme activities by pipetting. The cell density was correlated to the cell mass of L. lactis to be 0.36 g (dry weight) per L at a D600 of 1.

Plate screening of fermentation modes

Screening of the ldh mutant strains (additional ldh gene integrated on the chromosome) was performed on plates with bromcresol green (as a pH indicator) to distinguish between the pattern of product formation as homolactic (small green colonies) or mixed-acid fermentation (large white colonies) as reported previously [15]. The screening procedure was performed during growth of transformed strains of MG1363 [16] or MBP12 with pHWA202 derivatives and pLB95 [17] to obtain the ldh mutants. The cultures were plated on M17 broth [18] with 1% glucose, 10 mg·L−1 bromcresol green, and incubated at the permissive temperature 28 °C with erythromycin and tetracycline.

Curing of plasmid encoding the integrase

The ldh mutant strains were subsequently cured of pLB95 by growing at the nonpermissive temperature 35 °C on plates containing M17 broth with 2% glucose, 50 mm NaCl, and 10 mg·L−1 bromcresol green. The resulting colonies were tested for the loss of pLB95 on plates with or without tetracycline at 30 °C.


Antibiotics were used at the following concentrations: ampicillin, 100 µg·mL−1 (for selection of pUC18 and pBF12α derivatives in E. coli); erythromycin, 2 µg·mL−1 (for selection for maintenance of pHWA202 derivatives integrated in L. lactis) and 200 µg·mL−1 (for selection of pRC1, pG+host4, and pAK80 derivatives and subsequently pMU2916 in E. coli); tetracycline, 2 µg·mL−1 (for selection of pLB95 encoding the integrase orf1 gene necessary for integration in L. lactis) and 10 µg·mL−1 (for selection of E. coli strain ABLE K); kanamycin 50 µg·mL−1 (for selection of E. coli strain XLI-Blue MRF′).


Restriction enzymes, T4 DNA ligase and DNA polymerase Klenow fragment were purchased from Pharmacia Biotech and used as specified by the manufacturer. The restriction enzyme SpeI was purchased from Gibco/BRL Life Technologies.

Construction of plasmids used to alter the expression
of ldh

The plasmids constructed in this research are listed in Table 1. To obtain a strain, in which the chromosomal copy of ldh was disrupted, a temperature sensitive integration vector was employed using the following procedure. The plasmid pMU2901 was digested with DraI and HindIII giving an internal fragment of ldh (778 bp). This fragment was cloned in pUC18 digested with HincII and HindIII, resulting in pMBP2. The DNA region containing the 3′ end of pyk and 5′ end of ldh was obtained from pMU2901 digested with EcoRI and DraI. Subsequently, this 1.24-kb fragment was cloned in pMBP2 digested with EcoRI and SmaI, resulting in pMBP4. These ldh fragments with 16 bp of the polylinker of pUC18 in between were transferred by digesting pMBP4 with EcoRI and HindIII and inserted into pG+host4 [19] digested with EcoRI and HindIII, resulting in pMBP6.

Cloning of ldh in the integration vector pHWA202 was carried out using a three-step procedure as follows. The plasmid pMU2901 was digested with HaeIII resulting in a 1.32-kb fragment, which was cloned in pUC18 digested with SmaI, resulting in pHWA78. The ldh gene from pHWA78 was subcloned into pRC1 [20] by digesting the plasmids with EcoRI and SalI, yielding pHWA97. Subsequently, pHWA97 was digested with SpeI and XbaI and the gene was inserted in the integration vector pBF12α[21] digested with XbaI. The last cloning step resulted in pHWA202. The ldh gene was equipped with synthetic promoters by digesting the plasmids pCP2, pCP7, pCP29, and pCP25 [22] with XhoI and XbaI, and cloning the promoter fragments in pHWA202 digested with the SalI and XbaI. This cloning step resulted in the plasmids pMBP34, pHWA227, pHWA230, and pHWA231.

Quantification of glucose and fermentation products

The quantification of glucose, pyruvate, lactate, formate, acetoin, acetate, and ethanol was performed by using an HPLC system from Shimadza Corporation, Kyoto, Japan as described previously [10].

Measurement of lactate dehydrogenase activity

Enzyme activity was measured in permeable cells, which resulted in stable enzyme extracts. The enzyme preparation was analogous to the standard procedure used to determine β-galactosidase activity [23] with modifications in extract buffer as described previously [10]. Lactate dehydrogenase activity was measured by using a modified procedure of [24]. The final concentration in the reaction mixture was 50 mm triethanolamine with a pH of 7.5, 1 mm fructose 1,6-bisphosphate, 0.2 mm NADH, and 10 mm sodium pyruvate was used to initiate the reaction. The activity was determined from the rate of NADH oxidation at A340 nm at 28 °C using a Specord M500 spectrophotometer (Zeizz, Jena, Germany). The protein concentration in the cell extract was determined at A280 nm and correlated with A280 nm, 1 cm=1 to 1 mg protein per ml.

Curve fitting and calculation of control coefficients

Experimental data of the lactate dehydrogenase activity and the corresponding effect on the glycolytic, lactate, and formate fluxes and growth rates were calculated as the percentage change relative to the wild-type. The experimental data points were fitted to the following functions by using the programs datafit 6.1.10 (Oakdale Engineering, Oakdale, PA, USA) and prism 3.0 (Graphpad software, San Diego, CA, USA). The dependence of the glycolytic flux (JG) on the lactate dehydrogenase activity (x) was fitted to a one phase exponential association function: JG(x)=a+b·(1−exp(−c·x)), with a=58.95, b=41.87, and c=0.03153. The lactate flux was fitted to two functions: JL1(x)=a·b(1/x)·xc, with a=590.4, b=2.538×10−13, and c=−0.3206. As the above function excludes the data point at 0.9% activity, a third ordered polynomial function was fitted in addition to the flux to lactate: JL2(x)=a·x3+b·x2+c·x+d, with a=8.12×10−5, b=0.02638, c=2.78, and d=5.0. The flux towards the mixed-acid branches was here estimated from the formate flux and two different functions could describe the data. The first is a one phase exponential decay function: JF1(x)=a+b·exp(−c·x), with a=49.35, b=686.5, and c=0.02869. A second function could also describe the exponential data points, but differs significantly for the points above the wild-type: JF2(x)=exp(a+b·x+c·x2), with a=6.604, b=−0.02780, and c=6.629×10−5. The dependence of the growth rate (µ) on the lactate dehydrogenase activity was equal to the function, which was used to describe the glycolytic flux: µ(x)=a+b·(1−exp(−c·x)), with a=66.52, b=32.07, and c=0.03908, respectively.

To calculate control coefficients of lactate dehydrogenase the fitted function was differentiated and the derivatives were scaled by multiplying with the lactate dehydrogenase activity and divided by the respective fluxes using the equation:

inline image

For simplicity (J) refers here to either a flux or a growth rate. Using this equation control coefficients at the wild-type level were calculated as a function of the lactate dehydrogenase activity by intrapolating the response from the ldh mutants close to the wild-type. To unravel the control outside the wild-type level fitted curves were drawn using data from the ldh mutants above and below the wild-type level.


Construction of strains with modulated expression of ldh

The experimental approach chosen to access the control exerted by lactate dehydrogenase was to modulate the expression of ldh using site-specific chromosomal integration of an additional ldh gene transcribed from synthetic promoters [22]. The integration was made in the wild-type and in a strain deficient in lactate dehydrogenase and hence, expression levels both above and below the wild-type were obtained.

In order to obtain a strain devoid of lactate dehydrogenase activity, the ldh reading frame was disrupted by inserting 16 bp from the pUC18 polylinker in the beginning of the ldh coding region. The disruption was obtained by a double crossover using the temperature sensitive plasmid pMB6 (Fig. 1A) in which the pUC18 polylinker was flanked by a fragment containing the 3′ end of pyk and the 5′ end of ldh and an internal fragment of ldh. The resulting lactate dehydrogenase deficient strain MBP12 is shown in Fig. 1B. Subsequently, a series of plasmids were constructed, which carried the intact ldh gene transcribed from synthetic promoters (pMBP34, pHWA227, pHWA230, and pHWA231). The plasmids also harbor the attachment site attP of the TP901-1 phage, an E. coli replicon, and a selectable antibiotic marker for both E. coli and L. lactis. Figure 1C shows the structure of pHWA227.

Figure 1.

Strategy used to modulate the expression of ldh. (A) Schematic presentation of the chromosomal las operon in MG1363 and DNA fragments cloned in pMBP6. The bent arrow indicates the native promoter Plas of the operon. The dashed black line is the DNA fragment with the 3′ end of pyk and 5′ end of ldh. A hatched gray line indicates the internal part of ldh. MCS as a small black box denotes the 16-bp insert from the multiple cloning sites of pUC18. (B) After selection for erythromycin resistance at nonpermissive temperature and subsequently excision of pMBP6 resulted in the lactate dehydrogenase deficient strain MBP12. (C) The site-specific integration vector pHWA227 is shown where ldh is transcribed from the constitutive promoter CP7 (dotted black box). The black box signifies the attP region of TP901-1. By selecting for erythromycin resistance pHWA227 was site-specific integrated at attB on the chromosome of MG1363 (A) and MBP12 (B) resulting in the ldh mutants strains HWA236 (D) and HWA248 (E), respectively. The site-specific integration of pHWA277 between attP and attB generates the sequences attL and attR.

The plasmids pHWA227, pHWA230, and pHWA231 were integrated at the attB locus on the chromosome of MG1363 in the presence of the integrase (pLB95) by site-specific recombination between the attachment sites attP and attB. This resulted in the ldh mutant strains HWA236, HWA239, and HWA242, see Fig. 1D. In order to obtain expression levels below the wild-type level pMBP34, HWA227, and HWA230 were integrated in the lactate dehydrogenase deficient strain MBP12, resulting in the ldh mutant strains MBP51, HWA248, and HWA251, see Fig. 1E. As controls the parental plasmid pHWA202 without synthetic promoters in front of ldh was integrated in the lactate dehydrogenase deficient strain and MG1363, resulting in MBP50 and MBP52, respectively. The latter strain, which was used as the proper isogenic reference strain in the following physiological experiments, had properties identical to MG1363 with respect to lactate dehydrogenase activity, growth rate and metabolic fluxes. The ldh mutant strains were all cured of pLB95 by growing at nonpermissive temperature at 35 °C without selection of tetracycline. The phenotypes of MBP50 and MBP51 on the indicator plates (large white colonies) indicated that they were mixed-acid fermentative strains whereas the remaining ldh mutants appeared mainly homolactic (small green colonies).

The lactate dehydrogenase activity was subsequently measured in the mutant strains (Table 2). The ldh mutants had activities of lactate dehydrogenase ranging from less than 1% to 133% of the activity in MBP52. As seen from Table 2, several mutant strains were obtained having lactate dehydrogenase activities close to the wild-type level. The strains HWA242, HWA239, and HWA236 overexpressed ldh to 111, 122, and 133% of wild-type lactate dehydrogenase activity, respectively. Expression levels below the wild-type were achieved from MBP51, HWA251, and HWA248, which had 11, 59, and 86% of lactate dehydrogenase activity, respectively. The activity of the strain in which the chromosomal ldh was disrupted was determined to be below 1%.

Table 2. Specific activity of lactate dehydrogenase in ldh mutants. All mutants strains are MG1363 or MBP12 derivatives. The synthetic promoter transcribing ldh integrated at attB in the chromosome is shown. Pattern of fermentation was determined by analyzing the concentration of end products by HPLC. Specific activities are shown in nmol·min−1·mg protein−1. MBP52 was appointed to be the wild-type and the specific activity of LDH in the constructed strains is shown as the percentage of MBP52. Values are averages of at least two independent measurements and percentage standard deviations are shown in parenthesis.
StrainGenotypePattern of
Specific activities of LDH enzyme
nmol·min−1·mg−1% of wild-type
MBP52wt attB::ldhHomolactic102100 (8.7)
MBP50wt ldh attB::ldhMixed acid<1<1
MBP51wt ldh attB::CP2-ldhMixed acid1111 (0.9)
HWA251wt ldh attB::CP29-ldhHomolactic5859 (3.6)
HWA248wt ldh attB::CP7-ldhHomolactic8786 (6.2)
HWA242wt attB::CP25-ldhHomolactic113111 (0.2)
HWA239wt attB::CP29-ldhHomolactic124122 (7.5)
HWA236wt attB::CP7-ldhHomolactic136133 (13.2)

The influence of modulated lactate dehydrogenase activity on the fermentation pattern

The strains with lactate dehydrogenase activities around the wild-type level exhibited homolactic fermentation, and the fermentation remained homolactic with decreasing lactate dehydrogenase activity even with 59% lactate dehydrogenase activity. A major shift towards the production of mixed acids was observed when the activity was lowered to 11%.

The end-product formation of the ldh mutants is shown in Fig. 2. All strains except MBP50 and MBP51 produced similar amounts of the main product lactate (87%) and only minor amounts of formate and acetate. The strains HWA239 and HWA236 which overexpressed ldh to 122 and 133%, respectively yielded 7% more lactate than MBP52. The production of mixed-acid products was similar to the wild-type. The strains MBP50 and MBP51 showed a diversity of acidic products with acetate, formate and ethanol being predominant in addition to minor amounts of acetoin and lactate (Fig. 2). The shift from homolactic to mixed acids did not occur at one particular lactate dehydrogenase activity, but was a gradual transition from one fermentation mode to the other. The carbon recovery of the strains with low expression of ldh was poor (76% C-mol), which probably reflected that other end-products than the ones measured here were produced as reported previously [11,25].

Figure 2.

End-product formation of ldh mutants. The products produced by the ldh mutant strains during batch fermentation were analyzed by HPLC. Samples were withdrawn from cultures in stationary growth phase of a D600 of approximately 1.2, immediately after glucose has been exhausted. Thin bars signify the error in absolute values. The C-mol recovery of MBP50 and 51 ranged between 73 and 78%, whereas the other strains had a C-mol recovery of 82–85%.

Growth of the strains with altered lactate dehydrogenase activity

The growth of the ldh mutants was followed during batch fermentation. Figure 3 shows that the growth of the ldh mutant strains was exponential from a low cell density until the entry into the stationary growth phase. All the strains reach a yield on glucose (in terms of final D per mol of glucose per L) similar to the wild-type except for the strain deficient of lactate dehydrogenase and the strain with 11% of lactate dehydrogenase activity, which reached a slightly higher yield (108%). This probably reflects that the total ATP production during mixed-acid fermentation is higher compared with the homolactic strains. The specific growth rate was reduced by 28% (0.58 h−1) for the lactate dehydrogenase deficient strain compared to the wild-type (0.81 h−1). The remaining ldh mutant strains with lactate dehydrogenase acivities of 59% or higher had specific growth rates ranging from 0.79 to 0.81 h−1.

Figure 3.

Growth curves of ldh mutants. Cell density (D600) is shown as a function of time of the cultures. The time on the X-axis is relative in order to present the curves separately.

Validation of the experimental setup for control analysis

For the purpose of metabolic control analysis it is important that the gene expression level is stable, i.e. the specific enzyme activity should remain constant with increasing cell density. Samples were withdrawn during growth at different cell densities of HWA251 and the lactate dehydrogenase activity was measured (Fig. 4). Indeed, the strain had a constant specific lactate dehydrogenase activity during exponential growth and only a minor drop in activity took place in the stationary growth phase.

Figure 4.

Constitutive expression of ldh. Lactate dehydrogenase activity in strain HWA251 as percentage of the wild-type activity during growth (crosses). Growth of the culture is indicated as a broken line.

It is also important that the fluxes through the system are constant for the duration of the experiment. Samples were withdrawn at time intervals during growth experiment with the ldh mutants and the concentration of glucose and end-products was analyzed by HPLC. Figure 5 shows the glucose consumption, production of lactate and formate of the strains and these were indeed constant for all strains over a broad range of cell densities. The formate produced may be taken to represent the flux through the mixed-acid branches as lipoic acid, required for the activity of the pyruvate dehydrogenase complex, is absent from the growth medium used here.

Figure 5.

Consumption of glucose and production of end-products in ldh mutants. The consumption of glucose, production of lactate and formate (mm) are shown as a function of cell density as open triangles, closed diamonds and circles, respectively.

Determination of control exerted by lactate dehydrogenase

Metabolic fluxes were measured as the steady state consumption rate of glucose and the steady state production rate of lactate and formate during exponential growth (Fig. 5). The relative metabolic fluxes were calculated from the slopes of the curves for substrate and product concentrations plotted against the optical densities, multiplied by the specific growth rates. The fluxes were then plotted as a function of the relative lactate dehydrogenase activity see Fig. 6. Subsequently, to analyze the control by lactate dehydrogenase in L. lactis on the metabolic fluxes and the growth rate, control coefficients were calculated from the scaled slopes of fitted curves of the relative rates against relative enzyme activities.

Figure 6.

Metabolic fluxes, growth rate, and control coefficients of lactate dehydrogenase. Glycolytic flux, lactate flux, formate flux and growth rate are represented in A, B, C, and D, respectively. The relative metabolic fluxes and the relative growth rate are shown as a function of the relative lactate dehydrogenase activity (squares). Error bars indicate the standard deviation in percentage. The experimental data points are fitted to curves shown by penetrating lines and lines without symbols show the calculated control coefficients. In those cases where two different curves are used to describe the experimental data points, black and gray lines are used. Note that the formate production on Fig. 6C is plotted in log–log space.

Lactate dehydrogenase has no control on the glycolytic flux

Figure 6A shows the variation of the relative glycolytic flux plotted against the relative activity of lactate dehydrogenase. The lactate dehydrogenase deficient strain and the strain with 11% of lactate dehydrogenase activity had a glycolytic flux of 64 and 68% of the wild-type, respectively. Increasing the activity caused an increase of the glycolytic flux. At 59% or higher activities of lactate dehydrogenase the flux was almost constant. From the experimental data points a fitted curve was drawn using a one phase exponential association function. The calculated control coefficient from low activity of lactate dehydrogenase increased to a maximum value (0.18) at 26% of activity and subsequently the control decreased gradually towards zero with increasing activities. The control coefficient on the glycolytic flux by lactate dehydrogenase at the wild-type level was calculated from the equation JG(x) to be inline image=0.06. A third order polynomial function was also fitted (data not shown) to the data points, but resulted in a similar control coefficient (0.01) at the wild-type enzyme level. Independently of which of the fitted curves used, the data indicates that lactate dehydrogenase has virtually no control on the glycolytic flux of L. lactis at the wild-type enzyme level.

No control by lactate dehydrogenase on the production of lactate

When examining the lactate flux as a function of the lactate dehydrogenase activity (Fig. 6B), it was clear that even the mutant strain in which the ldh gene was disrupted had a small flux to lactate (4.7%). From this data point the lactate flux increased until 59% of lactate dehydrogenase activity and the flux was subsequently almost constant at higher activities of lactate dehydrogenase. The two functions fitted from the data points resulted in widely different control coefficients at low lactate dehydrogenase activities. Here, the third ordered polynomial function JL2(x) gave a better fit and the control was high (0.75), indicating that under these conditions lactate dehydrogenase does become rate limiting for the flux to lactate. Subsequently, the control decreased with increasing activity of lactate dehydrogenase. The calculated control coefficients were quite similar ranging from inline image=−0.03 to inline image=−0.06 at 100% of lactate dehydrogenase activity using the equations JL1(x) and JL2(x), respectively. This shows that no control was exerted by lactate dehydrogenase at the wild-type level on the lactate flux.

High negative control by lactate dehydrogenase on the production of formate

The dependency of the formate flux on lactate dehydrogenase is presented in Fig. 6C in a log–log plot in order to visualize the changes in the formate flux around the wild-type enzyme level. As expected when ldh was not expressed, the flux to formate was highest and decreased as the expression of ldh increased to the wild-type level. Above the wild-type level of lactate dehydrogenase the flux decreased even further. The control coefficients calculated from the equation JF1(x) became increasingly negative with increasing lactate dehydrogenase activity and gained a minimum value at 82% of lactate dehydrogenase activity. Hereafter the control became rapidly less negative. When using the equation JF2(x) to describe the flux data, the derived curve was lower at 133% of activity and therefore also the control was more negative above the wild-type level. The calculated control coefficients ranged from inline image=−1.27 to inline image=−1.45 at 100% activity of lactate dehydrogenase using the equations JF1(x) and JF2(x), respectively. Independently of which function to apply to the formate flux, the conclusion is that lactate dehydrogenase has a high negative control on formate production at the wild-type enzyme level.

No control by lactate dehydrogenase on the growth rate

The influence of lactate dehydrogenase activity on the growth rate was also investigated. The ldh mutants having from 59 to 133% of lactate dehydrogenase activity had growth rates very close to the wild-type, whereas activities below 1–11% resulted in reduced growth rates (Fig. 6D). The control on the growth rate showed a quite similar pattern to the control determined on the glycolytic flux. A maximum of control was observed around 22% of lactate dehydrogenase activity (0.14) and subsequently the control decreased to zero as the activity increased. The low control by lactate dehydrogenase at low enzyme activities probably reflects that the flux from pyruvate can be rerouted through the mixed-acid branches. The control coefficient calculated from the curve fit of equation µ(x) was determined to inline image=0.03 at the wild-type level, indicating virtually no control by lactate dehydrogenase on the growth rate.


The aim of the present study was to investigate the control exerted by lactate dehydrogenase on the metabolic fluxes and the growth rate of L. lactis. The approach used was to modulate the lactate dehydrogenase activity around the wild-type level by changing the expression of the ldh gene, measure the metabolic responses and calculate the relevant control coefficients. Importantly, strains with lactate dehydrogenase activities both below and above the wild-type level were obtained in this study. As small changes in gene expression can shift the distribution of control, the system under investigation should remain at a steady state, and we verified that the lactate dehydrogenase activity of the mutant strains was proportional to the cell density during our experiments. The rate of glucose consumption and the rate of product formation were also at steady state with a linear relationship between concentration and cell density within the range investigated. Therefore, these strains should form a solid basis for determining the control exerted by lactate dehydrogenase on L. lactis physiology.

Physiological effects of large changes in the expression
of ldh

Enhancing the ldh expression resulted in a minor increase in the amount of lactate produced, which is in agreement with an earlier study; in L. lactis the expression of ldh[26] was increased on low copy-number plasmids from 1.05- to 1.8-fold, which is close to the range applied in this study and they found only a small increase in the production of lactate. Gradually lowering the expression of ldh from the wild-type level towards zero did not affect the amount of lactate produced until almost a twofold reduction had taken place, and after this point a proportional reduction in lactate production occurred. A low amount of lactate was still produced in the strain deficient in lactate dehydrogenase, which is in accordance with earlier observations with the strain MG1363 [11] or the derivative NZ3900 [12,25] in which lactate dehydrogenase was disrupted. Recently, several other genes with homology to ldh were identified in L. lactis[27], and the corresponding enzymes may also catalyze the formation of lactate. Alternative enzyme activities may also convert pyruvate to lactate; an enzyme related to lactate dehydrogenase is, e.g. l-hydroxyisocaproate dehydrogenase, which is encoded on the chromosome of L. lactis ssp. lactis IL1403 [27].

Enhancing the lactate dehydrogenase activity had virtually no effect on the growth rate and glycolytic flux in L. lactis. Reducing the activity also did not affect these properties, until almost half the activity had been reached. These data show that L. lactis has homeostatic properties, serving to maintain ATP production and growth rate constant. Reducing lactate dehydrogenase activity may, to some extent, simulate the presence of high concentrations of lactic acid in the growth medium, which is also expected to inhibit lactate dehydrogenase. It is therefore tempting to speculate that these homeostatic mechanisms may be part of the response that these cells normally use to cope with high concentrations of lactic acid in the growth medium.

Absence of control by lactate dehydrogenase on the growth rate and glycolytic flux

Lactate dehydrogenase exerted no control on the growth rate and the glycolytic flux at the wild-type activity of the enzyme. This phenomenon is likely to reflect the presence of homeostatic mechanisms aiming to keep the energy production constant in the cells. Such mechanisms may include high elasticity towards pyruvate for the enzymes at the pyruvate branch, and metabolic regulation of other enzymes in the glycolytic pathway. Homeostatic regulation may also take place at the level of expression of the corresponding genes.

The low control by lactate dehydrogenase on the growth rate and glycolytic flux is in sharp contrast to our recent results, which showed that a mere twofold reduction of the activity of phosphofructokinase had a dramatic effect on L. lactis physiology [10]. The explanation here is that due to the presence of the mixed-acid branches, neither the growth rate, nor the glycolytic flux is entirely dependent on lactate dehydrogenase activity.

Absence of control by lactate dehydrogenase on lactate production

The fact that the control by lactate dehydrogenase on lactate production is very close to zero is perhaps a more surprising finding. Such a phenomenon was previously observed in E. coli where the essential enzyme, H+-ATPase (ATP synthase) also had no control on the growth rate [29] at its wild-type level. The lack of control by lactate dehydrogenase may reflect that the cell has an excess capacity of this enzyme that can be mobilized upon demand in order to maintain a constant lactate flux. Figure 7 shows a plot of the relative turnover number of the individual lactate dehydrogenase enzymes calculated by dividing the relative lactate flux with the relative lactate dehydrogenase activity. From the plot it can be seen that the individual enzymes gradually work faster when the total enzyme activity is reduced below the wild-type level, probably reflecting an increased intracellular pool of pyruvate. Around 60% relative lactate dehydrogenase activity the increase in relative turnover number stops. From this point, a further reduction in LDH activity does not result in an increase in relative turnover number, and at low activity of lactate dehydrogenase the turnover number is increased by less than twofold compared to the turnover number at the wild-type enzyme level. Thus, there is only a small excess capacity of lactate dehydrogenase activity in the wild-type cells and the Vmax appears to be reached already in the strain with 59% of lactate dehydrogenase activity. Together these results show that an enzyme can work relatively close to Vmax without being in control of its own flux, which emphasizes the fact that it is not possible to draw conclusions about flux control simply from the rate or properties of an enzyme [28].

Figure 7.

Relative turnover number of lactate dehydrogenase. The relative turnover number is calculated as the ratio of the relative lactate flux and the relative lactate dehydrogenase enzyme activity, and plotted as a function of the relative lactate dehydrogenase activity.

The sum of control coefficients by enzymes in a pathway with respect to a flux must by definition sum up to 1 [7], and control must therefore reside elsewhere in the system. In L. lactis glyceraldehyde 3-phosphate dehydrogenase [9] had a high control on the glycolytic flux. However, the latter determination of flux control was performed in nongrowing cells, which cannot be directly compared to growing cells. Another explanation could be that many glycolytic enzymes share the flux control and therefore most determined control coefficients will tend to be relatively small. Control may also be located outside the glycolytic pathway: In E. coli, it was recently shown that the majority of the control of the glycolytic flux resides in the processes that consume the ATP produced in glycolysis [29]. Recently, computer modeling and NMR studies of metabolite concentrations suggested that glycolysis in L. lactis[30] may also be limited by the reaction that consumes ATP.

Lactate dehydrogenase has a high negative control on the flux towards formate

Our results show that lactate dehydrogenase has an unusually high negative control on the flux to formate (C=−1.3) and hence on the production of mixed acids. Intuitively, part of the explanation for the high control coefficient could be the large difference in magnitude between the lactate and formate flux in the reference state; a small change in lactate production leads to a much higher relative change in formate production which translates to a high negative flux control. However, this is not necessarily so. If the pyruvate formate-lyase reaction was completely insensitive towards changes in the pyruvate pool then changes in lactate dehydrogenase might not have resulted in any change in formate production.

It has been demonstrated that the activity of lactate dehydrogenase in vitro is highly sensitive towards changes in the ratio of NADH/NAD+[6], and it was suggested that the ratio of NADH/NAD+ regulated the shift to mixed acid. However, neither the growth rate, nor the glycolytic flux changed when lactate dehydrogenase was modulated close to the wild-type level and it is therefore also unlikely that this modulation would have affected the NADH/NAD+ ratio to a significant extent under these conditions.

Formate production is representing an alternative route for ATP synthesis. Similarly, in the aerobic E. coli cell growing on glucose, there are two routes for ATP production, substrate level phosphorylation and oxidative phosphorylation. In this system, when the ATP synthase was reduced two homeostatic mechanisms became operative: the rate of substrate phosphorylation increased, and the respiration rate increased which led to a higher membrane potential and higher turnover number of the ATP synthase [31]. The latter mechanism appears to involve changes in gene expression of the respiratory components [32]. An interesting question is then whether the high negative control by lactate dehydrogenase on formate production also involves changes in gene expression; it has been found previously on several occasions that pyruvate formate-lyase activity is stimulated some threefold under conditions leading to mixed-acids production [6,33,34].


We sincerely appreciate the expert technical assistance of Katrine Madsen. We thank Bjarne Faurholm and Allan Hillier for donating the plasmid pBF12α and pMU2901, respectively. Martin Willemoës is acknowledged for suggestions and discussions. This work is part of the FØTEK program supported by the Danish Dairy Research Foundation (MFF) and the Center of Advanced Food Studies (LMC).