Presence of galactose in precultures induces lacS and leads to short lag phase in lactose-grown Lactococcus lactis cultures

Lactose conversion by lactic acid bacteria is of high industrial relevance and consistent starter culture quality is of outmost importance. We observed that Lactococcus lactis using the high-affinity lactose-phosphotransferase system excreted galactose towards the end of the lactose consumption phase. The excreted galactose was re-consumed after lactose depletion. The lacS gene, known to encode a lactose permease with affinity for galactose, a putative galactose–lactose antiporter, was upregulated under the conditions studied. When transferring cells from anaerobic to respiration-permissive conditions, lactose-assimilating strains exhibited a long and non-reproducible lag phase. Through systematic preculture experiments, the presence of galactose in the precultures was correlated to short and reproducible lag phases in respiration-permissive main cultivations. For starter culture production, the presence of galactose during propagation of dairy strains can provide a physiological marker for short culture lag phase in lactose-grown cultures. Electronic supplementary material The online version of this article (10.1007/s10295-018-2099-0) contains supplementary material, which is available to authorized users.

3 Supplementary Table S1 β-galactosidase assay results. Bioreactor batch samples were assessed along with the necessary positive and negative controls to rule out the contribution of extracellular β-galactosidases in galactose excretion (mean values of two technical replicates from two biological replicates, respectively). The relatively higher absorbance values in bioreactor samples from the later phase of growth can be interpreted as the increasing levels of compounds interfering with the assay towards the end of the lactose phase, such as lactate, ethanol, and acetate.

Sample and description
Absorbance at 410 nm  Supplementary Text S1 Off-line analytics of bioreactor cultures.

Biomass concentration
The optical density (OD600) was measured in polystyrene semi-micro cuvettes using a Genesis 20 spectrophotometer (Thermo Scientific, Waltham, MA, USA) in its linear range. The conversion factor between dry weight (DW) and optical density were determined using cell samples from cells grown in shake flasks. 5 mL of the broth volume was harvested at the early stationary phase of growth, vacuum filtered through a filter disc (0.45 µm PESU Membrane, Sartorius Stedim Biotech GmbH, Göttingen, Germany) and subsequently washed with 2x5 mL deionized water. Dry weight measurements were performed in both biological and technical triplicates. The conversion factor was found to be 0.38±0.03 g DW L -1 (OD600) -1 .

Extracellular metabolites
Lactose, galactose and extracellular metabolites were analyzed using HPLC (Dionex Ultimate 3000 unit, Thermo cientific, Waltham, MA, USA) with an Aminex HPX-87P 300 x 7.8 mm column (Bio-Rad, Hercules, CA, USA). The oven temperature was set to 80°C and the eluent was 5 mM H2SO4 at a flow rate of 0.6 mL min -1 . A Shodex RI-101 refractive index detector (Showa Denko K.K., Tokyo, Japan) was used to quantify the sugars and the extracellular metabolites. Galactose was also analyzed using NMR (Spedia-NMR TM , Spinnovation Analytical integrated NMR services, Oss, the Netherlands) to confirm the HPLC results (Table   S2).

Calculation of yields, rates and carbon balances
Based on the HPLC data of the consumed lactose and the produced extracellular metabolites, the extracellular metabolite yields were calculated as Cmol/Cmol yields. The measured amount of extracellular 7 metabolite produced at the time of lactose depletion was divided by the measured initial lactose concentration. For the biomass yield, the measured g L -1 DW biomass concentration at the time of lactose depletion was converted to Cmol using previously published biomass composition and molar mass of L.
lactis [2]. The same biomass composition was used when lactose and galactose were applied, and under anaerobic, aerobic and respiration-permissive conditions. Specific sugar consumption rates of lactose and galactose were calculated using the volumetric sugar consumption rates and the biomass dry weight concentrations ( Fig. 5B, D, F). Carbon balances on lactose consumed were calculated as the sum of the Cmol/Cmol yields of the extracellular metabolites produced (lactate, acetate, acetoin).

Calculation of the off-gas rates
The volumetric carbon dioxide evolution rate (CER, mmol (L h) -1 ) and oxygen uptake rate (OUR, mmol (L h) -1 ) were calculated as where F (L h -1 ) is the volumetric aeration rate at 1 atm and 20 °C, 22.41 (L mol -1 ) is the molar gas volume at 1 atm and 20 °C, 0.2095 is the oxygen molar fraction in the inlet air, is a compensation factor to relate the outlet gas flow rate to the aeration rate based on an inert gas balance, 0.0004 is the carbon dioxide molar fraction in air, and V (L) is the reactor volume. In order to compensate for the moisture content of the off-gas, the moisture fraction of the off-gas ( 2 ) was approximated using the oxygen wet 8 (%) value. The oxygen wet (%) was the measured oxygen content of the off-gas before inoculation in batch cultures.
Supplementary Text S2 Lag phase analysis of the main bioreactor cultures.

Lag phase calculations
The end of the lag phase was determined as the time when the first derivative of the CO2 (%) signal average had a non-zero value. The CO2 (%) and O2 (%) in the off-gas were monitored continuously using BlueSens gas analyzers (BlueSens Gas Technology GmbH, Herten, Germany) to calculate the volumetric CO2 evolution rate (CER, mmol·L -1 ·h -1 ) and the O2 uptake rate (OUR, mmol·L -1 ·h -1 ).  The samples were treated with RNAprotect reagent and frozen at -80°C.   RT-is ok if Cq difference is > 4-5 (bacterial samples, sufficient difference, else "NOT OK")  primer efficiency is ok in the range 80-110% (evaluation of the primer design -all primer pairs passed in the dilution range 50x-1000x) The qPCR data was evaluated with the MxPro software (Agilent Technologies Inc., Santa Clara, CA, USA) and Excel.
 Cq determination: threshold was manually set at the take-off point of the amplification curves in order to align the thresholds of multiple plate qPCR study  Possible outliers in terms of the shape of amplification curves were removed.

Statistical methods for results significance
The expression fold-difference between aerobic and respiration-permissive (aerobic hemin-supplemented) galactose batch samples was tested with a one-tailed t-test at a 95% confidence interval (alpha 0.05): Are the