Regulation of glycolysis and the shift between different fermentation modes of Lactococcus lactis have been extensively studied 12345678910111213. Key glycolytic enzymes have been characterized and concentrations of glycolytic intermediates in cell extracts have been obtained in many cases since the eighties 11. However, despite the wealth of available metabolic information for L. lactis, the key question of what controls the glycolytic flux in this organism cannot yet be answered unambiguously 11. When growing on rapidly metabolized sugars, this species shows homolactic metabolism in which more than 90% of metabolized sugar is converted to lactic acid. A deviation from homolactic fermentation is observed under aerobic conditions 1415 or during the metabolism of galactose 16 or maltose 17. The mechanisms underlying the shift from homolactic to mixed acid fermentation have been the object of considerable controversy so far and a full explanation has yet to be put forward.

Although sugar metabolism is a central issue in L. lactis physiology studies, growth on glucose as the sole carbon source is the case for a relatively small number of studies 35181920 the majority carried out mainly with lactose. However, Luesink et al. 921 showed that growth on glucose resulted in higher activities of the glycolytic key enzymes phosphofructokinase (PFK), pyruvate kinase (PYK), and L-lactate dehydrogenase (LDH), the genes of which form the tricistronic las operon. Further on, Even et al. 3, using a novel DNA macroarray technology, showed that several genes of glycolysis were expressed to higher levels on glucose and that genes of the mixed acid pathway were expressed to higher levels on galactose. Also, data given by Even et al. 3 on specific rates of growth, substrate consumption, and product formation (lactate, formate, acetate, and ethanol) during growth of L. lactis IL1403 on two different synthetic media (MCD and MS10R) with glucose or galactose as carbon source, show that glucose supports higher growth rates, sugar consumption rates and lactate production rates in both media than galactose.

The above-mentioned studies were carried out under anaerobic conditions. Aeration however, has been shown to strongly influence the cellular content of key enzymes. The negative effect of oxygen on expression of the pfl gene (encoding pyruvate formate lyase) is well known 2223, and the pfl gene product is known to be very sensitive to oxygen 232425. Expression of the adhE gene (encoding alcohol dehydrogenase) is known to be reduced by aeration 26. The levels of the three key glycolytic enzymes were found to be lower in L. lactis cells grown under aerobic conditions 15. In contrast, the in vitro specific activities of α-acetolactate synthase (ALS) and the pyruvate dehydrogenase (PDH) complex have been reported to increase with aeration 727. Most studies with L. lactis have been carried out under totally anaerobic conditions or, in a few cases, under totally aerobic conditions 714272829. Intermediate oxygen concentrations have been applied, to the best of our knowledge, only in the works by Jensen et al. 7 and Nordkvist et al. 19 with L. lactis ssp. cremoris. Both studies were carried out under microaerobic conditions (5 % dissolved oxygen tension-DOT, relative to saturation with air) and with glucose as the sole source of carbon. Comparisons were made at different aeration levels in the work of Nordkvist et al. 19. The maximum specific growth rate decreased with increasing aeration, while an optimum yield of lactate on glucose was obtained under microaerobic conditions.

Despite the fact that many reports describe the regulatory role of PTS (phosphotransferase system)- and non-PTS-sugars, there is not much information concerning the regulation of uptake of glucose itself and very limited information on growth rates of lactococcal strains on this carbohydrate. Information on the characteristics and properties of transport systems for glucose seem also to be rare. The physiology of L. lactis exposed to different concentrations of glucose and the levels of expression of key enzymes in the central pathway of energy production (PFK, PYK, and PDH) is also non-investigated. Recent results reported by Andersen et al. 30 concerning the role of PFK on glycolytic flux in L. lactis, show that this enzyme plays an important role since glycolytic and lactate fluxes were decreased proportionally by a twofold reduction of PFK activity. A key role was also attributed to PFK with regard to the glycolytic flux control by Neves et al. 15.

In view of the above and to extend our knowledge of the metabolic behavior of L. lactis, we performed batch cultivation experiments with L. lactis growing on glucose under different aeration conditions. Having evaluated the effects of aeration, microaerobic conditions were chosen to perform a series of batch fermentations with different initial glucose concentrations and fed-batch fermentations in which glucose was added at a rate suitable to maintain a stable concentration throughout the runs (glucostat). Specific growth rates and the yield on glucose of biomass, lactate, formate, ethanol, acetate and pyruvate were estimated. Specific uptake rates of glucose were determined, as well as the levels of expression of the las operon enzymes and the pools of intracellular metabolites for the range of 13.75 to 555 mM glucose concentrations. To further investigate the regulation of glucose uptake by L. lactis, the dynamics of glucose transport through the cell membrane were also studied.

Results showed that the amount of provided oxygen determines the outcome of fermentation. The equimolar quantities of the by-products formate, acetate, and ethanol produced under anaerobic conditions indicate that the PFL (pyruvate formate-lyase) pathway was the only active pathway of pyruvate metabolism besides LDH, this meaning that there was no flux through the PDH (pyruvate dehydrogenase) enzyme complex. As expected, a different profile was obtained in the presence of oxygen and this can be explained by the activity of NADH oxidases 19. Since under aerobic conditions neither formate nor ethanol was formed, the fluxes through PFL and alcohol dehydrogenase should be zero, this meaning that the PDH enzyme complex was responsible for the first step in the conversion of pyruvate to acetate. This is in accordance with the work of Jensen et al. 7 who showed that the specific in vitro activities of PDH and NOX (NADH oxidases) increase with aeration. At 5 % DOT microaerobic conditions, product formation was extremely homolactic and the highest yield of lactate on glucose was obtained. Formate was not formed, indicating an inactive PFL enzyme, and the very small amounts of ethanol in the broth should have resulted from activity of the PDH complex. Results of the same trend have been reported by Nordkvist et al. 19 for L. lactis ssp. cremoris MG1363.

Glucose conversion under microaerobic conditions was complete (Fig. 1) and the maximum specific growth rate was higher compared to that of semiaerobic and aerobic fermentations. The yield of biomass on glucose was the lower obtained within the set of conditions. This must be caused by an increased carbon distribution towards lactate formation since specific glucose uptake rates were high enough in the microaerobic fermentation with an initial glucose concentration of 55 mM (Fig. 4). A quite different situation was obviously obtained from the described by Jensen et al. 7 with L. lactis ssp. cremoris under microaerobic conditions with a 55 mM initial glucose concentration, where a 5 % DOT in the bioreactor changed the situation of glucose limitation prevalent under anaerobic conditions to one of glucose excess. The reason for the limited glucose consumption was not investigated.

Since microaerobic conditions supported higher specific rates of sugar consumption in the homofermentative mode, we carried out a series of batch fermentations with 5 % DOT, in which the concentration of glucose in the medium differed within the range of 13.75 to 555 mM. These fermentations were used to extract reference points for fed-batch cultivations. Studying the effects of glucose concentration in batch culture, observations unavoidably include the effects of a continuously changing sugar level during the process. This can be avoided with fed-batch culture operated as glucostat, a tool of choice for investigating the effects of glucose concentration levels. A small number of studies 6133132 have shown that the type and concentration of external sugar affects the concentrations of intermediate metabolites and cofactors and consequently the fermentation productivities. Although these reports are valuable, limitations may exist by the application of anaerobic batch culture only, or by comparisons within a narrow range of concentrations, or even by the use of non-proliferating cells rather than actively fermenting cells. In this study, a series of glucostat experiments were carried out to determine the effects of a stable glucose concentration level, within a broad concentration range, on fermentation rates, key glycolytic enzyme activities, and concentrations of important intracellular metabolites under microaerobic conditions.

In batch culture under microaerobic conditions, the initial glucose concentration of 138 mM supported the highest specific rates for growth, glucose uptake and lactate production. In fed-batch culture, maximal rates obtained by maintaining a continuous 55 mM glucose concentration. In batch culture, specific production rates for lactic acid have an optimum value reached rather early in the fermentation process (Fig. 3). However, the maximum values observed in batch runs occurred when the sugar had fallen to values that according to glucostat data were too low to give significant lactate production. Thus, there must be two aspects to the effect of the sugar. One is the level itself, the other arising from the dynamic situation with cells being exposed to a constantly changing glucose level in the bioreactor. The effect of continuous high glucose concentrations on cell viability may also explain the observed low specific production rates and biomass production levels when glucose exceeded the concentration of 277 mM (Fig. 5).

Glucostat data show that the concentration of glucose is related to its uptake and an optimum uptake rate is observed at 55 mM. Concentrations above 138 mM cannot support a high uptake rate (Fig. 6). The values shown in Fig. 6 obviously represent overall specific uptake rates and include all carbon metabolism of the cell, including excreted products and biosynthesis of cell cytoplasm. It appears from the results that as long as there is a "sink" for the carbon inside the cell, transport can increase in parallel with the sink. If that sink is saturated, then apparent transport will decrease. This is especially true at high glucose concentrations, where glucose appears to be entering the cell by simple diffusion.

Maintaining a continuous low glucose level such as 13.75 mM, specific glucose uptake rates are low and very interestingly, a shift towards mixed acid metabolism takes place. It is known so far for anaerobic culture that homolactic metabolism occurs on substrates supporting rapid growth in which significant amounts of glucose remain in the medium, and a mixed acid fermentation occurs when growth rates are lower and in true carbon-limited chemostats 633. Working the glucostat fed-batch mode under microaerobic conditions, a situation was arranged in which significant amounts of glucose were always present in the fermentation broth with glucose being the substrate supporting the highest fermentation rates 5. Therefore, the shift from homolactic to mixed acid fermentation could be directly correlated to the glucose uptake rate and consequently to the flux through glycolysis. Indeed, low glycolytic flux has commonly been ascribed to be the cause of mixed acid product formation 3683435. It has also been shown, that the glycolytic flux governs product formation only when the flux cannot meet the anabolic demand 4. In response to increased ATP demand the glycolytic flux increases, but if the latter is restricted, mixed acid formation occurs 4. This is in agreement with the data for ADP and ATP contents of cells presented in Fig. 8 but, the fact that the cells were growing under the glucostat fed-batch mode and glucose was always available but at a fixed, low concentration, points out that apart from the ATP consuming processes, sugar transport should also be involved in the mechanism underlying the phenomena, a fact rather ignored in previously published studies.

The shift from homolactic to mixed acid fermentation in L. lactis has been directly correlated to the glycolytic flux, estimated from the specific rates of sugar (glucose, galactose and lactose) consumption, in the work of Garrigues et al. 6. Under anaerobic conditions, the predominant role of the NADH/NAD⁺ ratio in controlling the shift was shown, as well as the relationship between glyceraldehyde phosphate dehydrogenase (GAPDH) activity and the NADH/NAD⁺ ratio. However, under conditions supporting less rapid growth with a diminished flux through glycolysis and a lower NADH/NAD⁺ ratio, such as growth on lactose and galactose, the major pathway bottleneck was identified at the level of sugar transport rather than GAPDH. The influence of GADPH on glycolysis has been discussed as either strictly controlling 36 or having such a role only when the glycolytic flux is high 6. Quite different regulatory aspects of glucose metabolism in the presence of oxygen have been reported by Neves et al. 15. It has been shown that the glycolytic flux was not primarily determined by the level of NADH in the cell, while the main point in that work was the observation that the decrease in the level of PFK activity (40%) was identical to the decrease in the glycolytic flux. A negative effect exerted by oxygen on the glycolytic flux was identified and attributed to depressed PFK activity. Working with another strain (L. lactis MG5627), the same group 11 observed a stimulation of glucose consumption under semiaerobic condition, a fact characterized as "apparent discrepancy", which showed that the level of oxygen during growth notably affected the cell metabolic machinery through different effects on gene expression.

In the present study, under conditions of increased glycolytic flux, as in the case of the 55 mM glucostat, the FBP pool (the PFK reaction product) was the largest observed (Fig. 7). It has been proposed 237 that FBP has an important role in the regulation of L. lactis metabolism acting as an allosteric regulator (activator) of PYK and LDH: high levels of FBP activate PYK and LDH and direct the flux towards lactate production, while low levels of FBP lead to LDH inactivation and inhibition relief of pyruvate-formate lyase by triose phosphates, resulting thus to a shift to mixed acid fermentation. Other researchers have questioned such a direct effect since the intracellular concentrations of FBP in general are sufficient to ensure full activation of LDH 6. The present work shows that the concentrations of FBP pools and the NADH/NAD⁺ ratio in the glucostat runs of 13.75 and 138 mM were almost identical, while neither the specific glucose uptake rates (Fig. 6) nor the fermentation patterns were similar. Moreover, much reduced FBP pools at very high glucose concentrations (Fig. 7) suggest that FBP cannot be regarded as a direct regulator of product formation, while they provide an indication of inhibition of PFK activity at such high glucose levels. While the FBP pool level cannot be directly connected to the glycolytic flux and the fermentation pattern, the explanation of the phenomena rather lies in the ATP demand of the cells and the glucose transport capacity of the microorganism.

Specific activities of the key glycolytic enzymes PFK, PYK and the LDH (Table 2), show that expression of the las operon genes in microaerobic glucostat cultures was influenced by the glucose level. The 55 mM glucose level supported the highest enzyme activities and this was reflected on the intracellular metabolites pools (Fig. 7). Values obtained in anaerobic batch cultures (Table 2, in parentheses) carried out with 55 mM initial glucose concentration are given for comparison and they are lower compared to those obtained in microaerobic glucostats. Low specific enzyme activities were obtained in the presence of high glucose concentration levels, such as 277 mM. The maximum glycolytic flux of 25.5 mmol g CDW^-1 h^-1 was observed in the 55 mM glucostat. That was reduced to 14 mmol g CDW^-1 h^-1 in the 277 mM glucostat, in which fewer ATP consuming processes took place in cells (Fig. 8). The 55% reduction in the glycolytic flux corresponds to a 56% reduction of PFK activity (Table 2). The explanation for the negative influence of elevated glucose levels on the glycolytic flux is likely to lie in part in depressed PFK activity. In addition, the course of FBP concentration over increasing glucose levels and the intracellular accumulation of unphosphorylated glucose (Fig. 7), refer to depressed PFK activity. As mentioned above, a strong influence of PFK on the glycolytic flux was identified in the works by Andersen et al. 30 and Neves et al. 15, in studies with the level of oxygen. A different approach in our work, through the glucose level, demonstrates, and validates, the regulatory role of PFK on glycolytic flux in L. lactis.

As with PFKs from Bacillus stearothermophilus and Escherichia coli, PFK from L. lactis is allosterically inactivated by excess ATP 38. Known inhibitors of PFKs in both prokaryotic and eucaryotic organisms include ATP, citrate and fatty acids, while known activators are ADP, AMP, cAMP, FBP, F6P, NH₄ and Pi 383940. ATP is the major effector and the activators for PFK are better described as deinhibitors of ATP because they reverse the effect of inhibitory concentrations of ATP. PFK is a tetrameric enzyme with two conformational states, R and T, which are in equilibrium. ATP is both a substrate and an allosteric inhibitor of PFK. Each subunit has two binding sites for ATP: a substrate site and an inhibitor site. The substrate site binds ATP equally well in either conformation but the inhibitor site binds ATP almost exclusively in the T state. The other substrate of PFK, F6P, preferentially binds to the R site. Consequently, at high concentrations, ATP acts as a heterotrophic allosteric inhibitor of PFK by binding to the T state, thereby shifting the T-R equilibrium in favour of the T state and thus decreasing PFK's affinity for F6P 40.

The amount of work carried out in the area of glycolysis in L. lactis during the last years has been significant. In their very recent review, Neves et al. 11 described the progress made and the further research, which should be pursued for a better understanding of this organism. The need for conducting sugar transport studies was characteristically pointed out. In the present work, glucose uptake rate studies suggested that two uptake systems are involved in the transport of glucose through the cell membrane of L. lactis LM0230, a high affinity and low affinity system. Literature information on the transporters of glucose through the cell membrane in L. lactis is rather limited 41. Thompson 1213 described a high affinity system for glucose transport in L. lactis LM3, which corresponds to EII^Man. The existence of a low affinity system for glucose has not been reported earlier, however evidence for glucose transport by a permease has been reported 1320. The trend in glucose transport (Fig. 9) is shown to be saturating when higher glucose concentrations than approximately 40 mM were used. Under conditions in which both carriers are saturated, accumulation of unphosphorylated glucose inside cells could be a result of uncontrolled glucose entry by unfascilitated diffusion, a fact already described with other microorganisms 42. Accumulation of intracellular glucose and the low levels of phosphorylated sugars under such conditions refer to inhibition of the PFK enzyme.

Thus, we propose a regulation mechanism governed by the energy state of the cell, as this is expressed by the cellular quantities of ADP and ATP, by which L. lactis can handle the glycolytic flux under microaerobic conditions. ADP and ATP play central roles in the metabolic network and influence several steps of glycolysis because they are substrates and products of kinases and inhibitors of dehydrogenases. ATP acts as a free-energy donor to drive transport and biosynthesis and it is continuously regenerated from ADP by substrate level phosphorylation. ATP is invested in the upper part of the glycolytic pathway to generate a surplus in the lower part. Moreover, both ATP and ADT serve as precursors in the DNA and RNA synthesis, which has been shown to constitute an approximately 3 and 8% of L. lactis cells dry weight, respectively 34. It has also been shown that intracellular concentrations of ADP and ATP in growing L. lactis cells are tightly controlled (homeostatic control) at levels optimal for the cellular reactions 34. In the present study, under low glucose concentration conditions, provided in glucostat cultures, the glycolytic flux cannot meet the anabolic demand of the cells. There is glucose limitation and consequently energy limitation and the glucose transport capacity of the microorganism is not met, resulting in mixed acids formation. The FBP pool, through LDH and PYK control, does not directly influence product formation since low FBP concentrations were characteristic of both low (13.75 mM) and high (138 mM) glucose concentration levels in glucostat cultures. Therefore, under such conditions, the ATP demand and the glucose transport capacity of the cells are main regulators of the flux. Reversing the situation, by providing constant elevated glucose levels in the glucostat (e.g. above 55 mM), conditions in which glucose transport carriers are saturated, led to excess energy (formation of large intracellular pools of ADP and ATP) which the organism can handle through the allosteric properties of its enzymes. Excess ATP in this case, inhibits PFK activity slowing the glycolytic flux down. It can be suggested here that the extend to which ATP demand controls the glycolytic flux depends on how much excess capacity of glycolysis is present at cells.

The results in this study suggest that the control of the flux through the pathway of glycolysis resides to a large extent in processes outside the pathway itself, like glucose transport and the ATP consuming reactions. Depending on culture conditions, e.g. dissolved oxygen concentration and glucose concentration levels, the overall flux in L. lactis seems to be regulated by the ATP demand through the allosteric properties of key enzymes, with PFK having a significant influence on the control.

All authors contributed equally in this work