Selection of mutants tolerant of oxidative stress from respiratory cultures of Lactobacillus plantarum C17


  • T. Zotta,

    Corresponding author
    1. Istituto di Scienze dell'Alimentazione-CNR, Avellino, Italy
    • Correspondence

      Teresa Zotta, Istituto di Scienze dell'Alimentazione-CNR, via Roma 64, 83110 Avellino, Italy.


    Search for more papers by this author
  • R.G. Ianniello,

    1. Scuola di Scienze Agrarie, Forestali, Alimentari e Ambientali, Università degli Studi della Basilicata, Potenza, Italy
    Search for more papers by this author
  • A. Guidone,

    1. Scuola di Scienze Agrarie, Forestali, Alimentari e Ambientali, Università degli Studi della Basilicata, Potenza, Italy
    Search for more papers by this author
  • E. Parente,

    1. Istituto di Scienze dell'Alimentazione-CNR, Avellino, Italy
    2. Scuola di Scienze Agrarie, Forestali, Alimentari e Ambientali, Università degli Studi della Basilicata, Potenza, Italy
    Search for more papers by this author
  • A. Ricciardi

    1. Istituto di Scienze dell'Alimentazione-CNR, Avellino, Italy
    2. Scuola di Scienze Agrarie, Forestali, Alimentari e Ambientali, Università degli Studi della Basilicata, Potenza, Italy
    Search for more papers by this author



Lactobacillus plantarum is a lactic acid bacterium involved in the production of many fermented foods. Recently, several studies have demonstrated that aerobic or respiratory metabolism in this species leads to improved technological and stress response properties.

Methods and Results

We investigated respiratory growth, metabolite production and stress resistance of Lact. plantarum C17 during batch, fed-batch and chemostat cultivations under respiratory conditions. Sixty mutants were selected for their ability to tolerate oxidative stress using H2O2 and menadione as selective agents and further screened for their capability to growth under anaerobic, respiratory and oxidative stress conditions.

Dilution rate clearly affected the physiological state of cells and, generally, slow-growing cultures had improved survival to stresses, catalase production and oxygen uptake. Most mutants were more competitive in terms of biomass production and ROS degradation compared with wild-type strain (wt) C17 and two of these (C17-m19 and C17-m58) were selected for further experiments.


This work confirms that, in Lact. plantarum, respiration and low growth rates confer physiological and metabolic advantages compared with anaerobic cultivation.

Significance and Impact of the Study

Our strategy of natural selection successfully provides a rapid and inexpensive screening for a large number of strains and represents a food-grade approach of practical relevance in the production of starter and probiotic cultures.


Lactic acid bacteria (LAB) are oxygen tolerant anaerobes used as starter, adjunctive or probiotic cultures in the production of many fermented and functional foods to improve organoleptic properties, increase safety and shelf-life and promote human health. During their use in foods, LAB are subjected to physical and chemical stresses that may impair growth, cell viability and fermentation capabilities and the type of metabolism mainly affects the robustness to harmful conditions by changing technological properties of strains: aerobic cultivation and the presence of heme and menaquinone in the medium induces in some LAB species (Lactococcus lactis, Streptococcus agalactiae, Enterococcus faecalis and Lactobacillus plantarum) the activation of an electron transport chain (ETC, respiratory metabolism), the synthesis of ROS (reactive oxygen species) detoxifying enzymes as well as an improved stress tolerance and biomass production (Lechardeur et al. 2011; Pedersen et al. 2012). The shift towards aerobic and respiratory pathway has been studied and characterized in the dairy starter Lc. lactis, allowing the selection of strains with improved technological and stress response properties, and this has been successfully exploited for an industrial scale production by Chr. Hansen A/S (Pedersen et al. 2005). More recently, several authors (Brooijmans et al. 2009a; Mazzeo et al. 2012; Quatravaux et al. 2006; Stevens et al. 2008; Watanabe et al. 2012a,b; Zotta et al. 2012) have demonstrated that in the versatile species Lact. plantarum the presence of functional aerobic or respiratory metabolism could also be useful for new applications in food technology and starter or probiotic development. However, most of these studies referred to the model strain Lact. plantarum WCFS1 and only a few (Guidone et al. 2013a; Watanabe et al. 2012a,b; Zotta et al. 2012) have considered the effect of oxygen and cofactor supplementation on the stress response of this species.

We previously (Zotta et al. 2013) investigated the effect of temperature, aeration and hemin–menaquinone supply on the growth, oxygen-related enzymes and stress behaviour of Lact. plantarum C17 during batch fermentations in comparison with anaerobic cultivation. Overall, we confirmed the recent observations (Watanabe et al. 2012a; Zotta et al. 2012) for which aerobic and respiratory lifestyle offers advantages compared with anaerobic growth.

The aim of this work was to evaluate the effect of respiration on the stress tolerance of Lact. plantarum C17 under more controlled growth conditions, using fed-batch and chemostat cultivations. This type of propagation has been widely exploited in several LAB species to reach an increased biomass and metabolite (organic acids, bacteriocins) production (Callewaert and De Vuyst 2000; Castro et al. 2007; Hwang et al. 2011; Mende et al. 2012; Peng et al. 2006; Racine and Saha 2007), but few authors have investigated the usefulness of fed-batch and continuous growth to understand the mechanisms of aerobic or respiratory metabolism and stress response (Dressaire et al. 2008; Papagianni et al. 2007; Tseng and Montville 1992) in LAB. Therefore, we combined chemostat cultivations and respiratory growth to achieve high-cell-density cultures (which could be of practical relevance for starter development) and to clarify some metabolic and stress-related aspects in Lact. plantarum.

Additionally, because the most common methods to enhance the competitiveness of strains are based on site-directed mutagenesis, using a range of recombinant DNA techniques, we tried to improve the phenotypic properties of wild-type strain Lact. plantarum C17 through spontaneous mutagenesis using high concentration of ROS (reactive oxygen species) generators to select respiratory cells with competitive phenotypes.

Material and methods

Strains and culture conditions

Lactobacillus plantarum C17, isolated from Caciocavallo cheese, was selected because of its stress tolerance (Parente et al. 2010), functional properties (Guidone et al. 2013b; Ciocia et al. 2013) and ability to shift towards aerobic/respiratory metabolism (Guidone et al. 2013a; Zotta et al. 2013).

The strain was maintained as freeze-dried stock in reconstituted 11% (w/v) skim milk containing 0·1% (w/v) ascorbic acid (RSM) in the Culture Collection of Scuola di Scienze Agrarie, Forestali, Alimentari e Ambientali, Università degli Studi della Basilicata.

For routine propagation (16 h, 35°C), a complex basal medium (WMB, Zotta et al. 2012) was used, while for batch, fed-batch and continuous cultivations a modified (without sodium acetate 5 g l−1) WMB (mWMB) was used.

Continuous fermentations

Chemostat cultivations were carried out in a 3-l glass fermentor (Applikon, Schiedam, the Netherlands), filled with 1·5 l of mWMB, under respiratory promoting conditions (air 0·2 v/v per min, with 2·5 μg ml−1 hemin and 1 μg ml−1 menaquinone supplementation) at 35°C (optimal temperature of growth for Lact. plantarum C17; ezControl controller, Applikon, Schiedam, the Netherlands) and constant pH of 6·5. Agitation was performed with a minimum impeller speed of 200 rev min−1 (2 Rushton turbines, diameter 45 mm), pH was controlled by automatic addition of sterile 1 : 1 Na2CO3/NaOH solution (4 eq l−1), while foaming was controlled by adding 1% (v/v) Antifoam A solution. Concentration of dissolved oxygen (DO%) was measured by a polarographic electrode (Applisens, Applikon, Schiedam, the Netherlands) and controlled at ≥30% by automatically varying stirrer speed.

The bioreactor was inoculated (5% v/v) with an overnight aerobic WMB preculture and operated batchwise (10 g l−1 initial glucose) until early stationary growth phase was reached, continuous feeding (40 g l−1glucose in the feed, SF) of fresh medium was started at a flow of 36 ml h−1. Fed-batch operation was used until the volume (V) reached 2 l. Volume was brought at 1·5 l and kept constant using inflow and overflow tubes connected to a peristaltic pump (model 502S, Watson-Marlow Ltd., England). Fresh medium (SF = 10 g l−1) was fed into the fermentor, and continuous cultures were operated at dilution rates (D) of 0·31, 0·17, 0·07 and 0·04 h−1 by varying the flow rate (F). Sampling was started after at least three culture volumes were passed through the vessel and, to ensure the steady state of culture, the optical density at 650 nm (OD650) of three samples taken from the outlet medium at 30-min intervals was measured. If the OD650 values were stable, three additional samples were taken and used for the subsequent analyses.

Chemical and biochemical analyses

Cell dry weight (measured after drying the washed biomass at 105°C for 24 h) was used to estimate the biomass yield per unit substrate consumed (YX/S, g g−1).

Residual glucose (as reducing sugar) in culture supernatants was measured using the 3,5-dinitrosalicylic acid (DNSA) method (Miller 1959), while enzymatic kits (R-Biopharm AG, Darmstadt, Germany) were used to quantify lactic and acetic acid. Hydrogen peroxide (H2O2) concentration in supernatants and catalase activity in whole cells were measured as described by Risse et al. (1992).

The activities of enzymes related to aerobic metabolism (pyruvate oxidase, POX; NADH oxidase, NOX; NADH peroxidase, NPR) were measured in cell-free extracts (FastPrep-24 Instrument, MP Biomedicals, Santa Ana, CA) according to Quatravaux et al. (2006) at both 25 and 37°C (in vitro assay temperature).

Oxygen uptake

The oxygen uptake of batch, fed-batch and steady-state cells was measured in situ by temporarily closing the air supply in the bioreactor and by monitoring the decrease in dissolved oxygen concentration (DO%) for 2 min at 35°C (15-sec intervals).

DO% values were transformed into μmol l−1 using the Henry's law, and the specific oxygen uptake rate (μmol O2/min/g biomass) was calculated.

Stress treatments

Tolerance of heat, freezing, freeze-drying and oxidative stresses was evaluated as described in Zotta et al. (2013). For heat treatments, batch, fed-batch and steady-state cells (whole cells recovered by centrifugation at 12 000 g, 5 min and washed and re-suspended in 20 mmol l−1 phosphate buffer pH 7·0, PB7; final OD650 = 1·0) were exposed at 55°C for 0, 5, 10, 15 and 30 min in PB7, and the number of survivors was estimated by pour plate counts in WMA (35°C, 48 h, anaerobiosis). The kinetics of inactivation were fitted using a Weibull model (van Boekel 2002).

Tolerance of freezing (cells resuspended in RSM, OD650 = 1·0, storage at −20°C in 50% w/w glycerol solution, thawing) and freeze-drying (cells re-suspended in RSM, OD650 = 1·0, freeze-drying and storage at −20°C) was also evaluated by pour plating the samples on WMA (35°C, 48 h, anaerobiosis) after 30 days of storage.

To estimate the survival after oxidative stress, batch, fed-batch and steady-state cells were exposed (30 min, 35°C) to different H2O2 concentrations (from 0·8 to 0·0015 mol l−1) and the survivors (if any) were cultivated in WMB (pH 6·8, 16 h, 35°C, microplate experiment) and spread plated on WMA (35°C, 48 h, anaerobiosis). Additionally, the tolerance of menadione (a superoxide generator) was also evaluated by spread plate counting on WMA containing menadione (0·3–0·018 mmol l−1).

For continuous cultures, two replicate samples taken at 60-min distance were used and the stress treatments were performed in duplicate.

Selection of random mutants

The mutants were selected from batch, fed-batch and chemostat (D = 0·31, 0·17, 0·07 and 0·04 h−1) cultures. At the end of each cultivation step, the cells were (i) exposed (30 min, 35°C) to different H2O2 concentrations (from 0·8 to 0·0015 mol l−1) and cultivated (100 μl; by spread plating) on WMA (35°C, 48 h, anaerobiosis) and (ii) directly cultivated (100 μl; by spread plating) on WMA containing menadione (0·3–0·018 mmol l−1). The survivors to the highest H2O2 (0·8 or 0·4 mol l−1) and menadione (0·075 or 0·036 mmol l−1) concentrations were randomly picked (10 colonies for each growth conditions, including 5 H2O2-survivors and 5 menadione-survivors), and a total of 60 tolerant mutants were collected and stored (glycerol 25% v/v) at −24°C until the use.

Qualitative screening

All mutants were screened (cultivation in 96-well microplate) and compared with the wild-type strain C17 for their ability (i) to grow in anaerobiosis (AN; static cultivation in WMB, initial pH 6·5, at 35°C for 16 h, using Generbox jars, bioMérieux SA, Marcy-l'Etoile, France, and AnaeroGen bags, Oxoid; measurement of OD650), (ii) to grow in respiratory promoting conditions (RS; shaking cultivation in WMB with 2·5 μg mL−1 hemin and 1 μg ml−1 menaquinone, WMB+HM, initial pH 6·5, at 35°C for 16 h; agitation of microplates on a rotary shaker at 150 rev min−1; measurement of OD650) and (iii) to tolerate oxidative stress conditions (static cultivation in WMB with 0·16 g l−1 bromocresol purple, WMB+BCP, containing 0·8 or 0·4 mol l−1 H2O2 and in WMB+BCP with 0·075 or 0·036 mmol l−1 menadione, at 35°C for 16 h, using Generbox jars, bioMérieux SA, Marcy-l'Etoile, France, and AnaeroGen bags, Oxoid; change of colour from purple to yellow was considered as positive result). The effect of preculture (anaerobic or respiratory inoculum) on the growth and stress resistance was also evaluated.

Quantitative selection

Twenty strains were selected and grown (35°C, up to 24 h) under static (screw-cap tubes filled with WMB containing 10 g l−1 glucose and 0·1 mol l−1 MOPS, buffered WMB, initial pH 6·5; anaerobiosis) and shaking (250 ml baffled flasks with 50 ml buffered WMB+HM, agitation on a rotary shaker, 150 rev min−1; respiration) conditions, using anaerobic or respiratory inoculum as appropriate. Increase in OD650 and pH values was measured after 16 and 24 h of growth at 35°C.

Growth experiment on the best performing mutants

The two mutants C17-m19 and C17-m58 were further selected and used for a more complex anaerobic and respiratory growth experiments in the presence of either glucose or maltose (10 g l−1) as carbon source. Maximum specific growth rate (μmax), pH value and biomass production were measured after 8 and 24 h of incubation at 35°C.

Reagents and culture media

Unless otherwise noted, all reagents were obtained from Sigma-Aldrich s.r.l. (Milan, Italy) and culture media and ingredients from Oxoid (Basingstoke, Hampshire, UK).

Statistical analysis

All statistical and graphical analyses were performed using Systat 13.0 for Windows (Systat Software Inc., Richmond, CA).


Growth, chemical and biochemical analyses

Growth parameters, sugar consumption and organic acid productions of Lact. plantarum C17 are shown in Table 1.

Table 1. Growth parameters and metabolic production of Lactobacillus plantarum C17 during batch, fed-batch and continuous cultivations, under respiratory conditions (air 0·2 v/v per min, with 2·5 μg ml−1 hemin and 1 μg ml−1 menaquinone supplementation), at 35°C and constant pH 6·5
Growth conditionD or μmax (h−1)SS0 r S Y X/S Y P/S (lactic acid) Lactic acid (g l−1)Acetic acid (g l−1)
  1. Mean values±standard error are shown.

  2. D = dilution rate was used for fed-batch and continuous cultivations; μmax = maximum specific growth rate was used for batch cultivation.

  3. SS0 = consumed glucose, g.

  4. rS = rate of glucose consumption, g l−1 h−1.

  5. ν = specific rate of glucose consumption, g g−1 h−1).

  6. YX/S = biomass yield coefficient (biomass yield, g, relative to total sugar consumed, g).

  7. YP/S = lactic acid yield (lactic acid produced, g, relative to total sugar consumed, g).

  8. Growth condition: All cultures were grown at 35°C in aerobiosis (in the presence of 2·5 μg ml−1 haemin and 1 μg ml−1 menaquinone).

Batch0·5773·94 ± 0·202·28 ± 0·120·09 ± 0·010·34 ± 0·011·33 ± 0·110·03 ± 0·00
Fed-batch0·07023·74 ± 0·251·70 ± 0·020·17 ± 0·000·47 ± 0·0211·21 ± 0·360·00 ± 0·00
Continuous0·31127·36 ± 0·268·51 ± 0·080·13 ± 0·000·47 ± 0·0012·97 ± 0·000·00 ± 0·00
Continuous0·17335·33 ± 0·096·12 ± 0·010·13 ± 0·000·57 ± 0·0219·97 ± 0·700·04 ± 0·00
Continuous0·07037·28 ± 0·052·62 ± 0·000·13 ± 0·000·70 ± 0·0125·92 ± 0·230·00 ± 0·00
Continuous0·04136·72 ± 0·051·52 ± 0·000·11 ± 0·000·79 ± 0·0028·90 ± 0·140·00 ± 0·00

Fed-batch and chemostat cultivations improved biomass yield (Yx/s) compared with the batch growth. Highest D rate cultivations resulted in higher residual glucose and lower production of total lactic acid (Yp/s), similarly in fed-batch cultures for which metabolite concentrations were significantly different from those measured during slow growth.

Although cultivations were carried out under respiratory conditions, acetate (the main product of pyruvate oxidase-acetate kinase, POX–ACK, pathway) was not detected in supernatants.

With the exception of cells cultivated in batch, high catalase activities (from 19 mkatal g−1 of biomass in 0·170 h−1 D rate to 28 mkatal g−1 of biomass in the lowest 0·040 h−1 D rate) were measured (Table 2) and therefore low levels of H2O2 were only found (0·02 mmol l−1) in batch cultures. Catalase activity was significantly and inversely correlated (r = −0·958) with specific growth rate.

Table 2. Catalase activity, tolerance to oxidative stress and oxygen uptake of Lactobacillus plantarum C17 during batch, fed-batch and continuous cultivations
Growth conditionD or μmax (h−1)CatalaseaOxidative stress tolerancebOxygen uptakec
H2O2 (mol l−1)Menadione (mmol l−1)Maximum specific rate of O2 uptake (in situ)Stirrer speed parameters (rev min−1)
Mean ± SDMinMax
  1. Mean values±standard error (se) are shown.

  2. D = dilution rate; μmax = maximum specific growth rate.

  3. a

    Catalase activity was expressed as mkatal g−1 of biomass.

  4. b

    The maximum concentration of H2O2 and menadione for which colony growth was evident is reported; col, average number of colonies counted on WMA plates and from which mutants were picked.

  5. c

    Oxygen uptake was measured in situ closing temporarily (maximum 2 min) the air sparger of stirred bioreactor and was expressed as μmol O2 g−1 of biomass, relative to residual glucose, mmol l−1. Stirrer speed parameters (mean values±standard deviation, minimum and maximum reached values) are reported to show the changes in oxygen consumption during growth conditions.

Batch0·5770·10 ± 0·000·4 (33 col)0·018 (spread)0·020 ± 0·00200·2 ± 1·0200206
Fed-batch0·07025·5 ± 0·440·8 (13 col)0·075 (72 col)0·108 ± 0·00246·2 ± 20·5200288
Continuous0·31119·7 ± 0·370·4 (41 col)0·036 (54 col)0·020 ± 0·00341·5 ± 71·7200479
Continuous0·17319·4 ± 0·780·2 (20 col)0·036 (85 col)0·054 ± 0·00351·5 ± 89·9200457
Continuous0·07123·2 ± 0·470·8 (37 col)0·075 (110 col)0·107 ± 0·01309·8 ± 108·4200546
Continuous0·04128·1 ± 1·070·8 (186 col)0·075 (149 col)0·087 ± 0·00296·2 ± 99·7200536

Activities of the enzymes involved in aerobic metabolism

As shown in a previous study (Zotta et al. 2013), the activities of POX and NPR were significantly (Tukey's HSD, < 0·005) affected by the assay temperature (Fig. 1). POX was completely inhibited in vitro at 37°C, while NPR slightly decreased in this condition. NOX activity was not particularly correlated to assay temperature, even if a small increase was observed at 37°C in fed-batch and 0·07 h−1 D rate cells. NPR activity, but not NOX activity, showed a significant negative correlation with specific growth rate (r = −0·829). Contrary to the NADH-dependent enzymes, mainly detected during fed-batch and low (0·07 and 0·04 h−1) D rate growth (at both 25 and 37°C), the highest POX activities were measured in fast-growing cells (D = 0·31 and 0·17 h−1) when significant amounts of residual glucose (Table 1) were still available in the substrate.

Figure 1.

Activities of the main enzymes related to aerobic metabolism, measured in cell-free extracts of batch, fed-batch and steady-state cultures of Lactobacillus plantarum C17 grown under respiratory promoting conditions (air 0·2 v/v per min, with 2·5 μg ml−1 hemin and 1 μg ml−1 menaquinone supplementation). Assay was carried out at both 25 and 37°C. POX, pyruvate oxidase; NOX, NADH oxidase; NPR, NADH peroxidase. Data were expressed as μkatal mg−1 of protein. As NOX and NPR, one katal corresponds to the oxidation of 1 mol of NADH per min at T°C; as POX, one katal corresponded to the production of 1 mol of H2O2 per min at T°C. Letters (a, b, c, d, e, f) on multiplot bars indicate significant differences (Tukey's HSD, P < 0·005) in enzymatic activities among growth conditions, asterisks (*) indicates significant differences between assay temperature. Standard error bars are shown.

Oxygen uptake

The concentration of dissolved O2 (DO,%) was kept constant at 30% by changing the stirrer speed in bioreactor. The consumption of O2 was assessed by analysing the stirrer speed profiles during different growth or by measuring in situ (closing temporarily the air sparger) the decrease in DO% (Table 2).

With exception of batch cultivation, for which no change in stirrer speed was observed, fed-batch and continuous cultures needed significant shifts in stirrer speed parameters (maximum values and standard deviation from average values) to maintain the required DO%, indicating O2 uptake during growth. The largest stirrer speed changes were observed in low D rate cultivations. Consistent with these results, the measurements carried out in situ confirmed greater O2 consumption (μmol O2 g−1 of biomass relative to mmol l−1 of residual glucose) by slow-growing cells and to a lesser extent in those from batch and high D rate cultures. However, despite to the weak changes in stirrer speed, fed-batch cells had the highest capability of O2 utilization.

Stress tolerance

Thermal inactivation

The kinetics of thermal inactivation were estimated with the Weibull model, which provided a good fit for most cultivations (R2 ranging from to 0·90 to 0·99). Weibull modelling showed that exponential growth in batch culture and high D rate cultivations significantly (< 0·005) impaired heat survival decreasing the time to reach 3-log-cycle reduction (t3D; Fig. 2). Fed-batch cells and slow-growing cells (D rates of 0·041 and 0·07 h−1) exhibited the greatest heat stress tolerance, with a time of log-cycle reduction higher than that estimated for the fast-growing (0·31 and 0·17 h−1; t3-D increase from 1·5 to 4 times) and batch (t3-D increase from 9 to 17 times) cells. As shown in Fig. 2, t3-D values were inversely correlated with growth rates (i.e. the higher the growth rate, the lower the survival), and positively correlated with catalase and NPR activities. A square root transformation of t3-D provided the best fit with specific growth rate (R2 = 0·79).

Figure 2.

Relationship between t3D values (time to reach 3-log-cycle reduction; estimated with the Weibull model, van Boekel 2002) and dilution or maximum specific growth rates (μmax, h−1) of batch, fed-batch and steady-state cells of Lactobacillus plantarum C17 grown under respiratory promoting conditions (air 0·2 v/v per min, with 2·5 μg ml−1 hemin and 1 μg ml1 menaquinone supplementation). The continuous line was plotted using distance weighted least squares (DWLS) smoothing.

Tolerance of oxidative stress

The resistance to H2O2 was significantly improved in fed-batch and slow-growing (Table 2), in agreement with the higher values of catalase and NPR activities detected in these cultures. Additionally, the number of survivors revealed a satisfactory tolerance to menadione (up to 0·075 mmol l−1 in fed-batch and low D rates of 0·041 and 0·07 h−1), with a significant correlation (r = 0·90) between specific growth rate and survival, even if Lact. plantarum lacks the sod gene (encoding for the antioxidant enzyme superoxide dismutase). NPR activity showed a significant correlation with tolerance of H2O2 (r = 0·87) and menadione (r = 0·95).

Tolerance of freezing and freeze-drying

Culture conditions, type of preservation treatment (freezing vs freeze-drying) and interaction all had a significant effect (P < 0·001) on survival. The slowest-growing cells (D = 0·041 h−1) exhibited the highest survival to freezing (<0·4 log-cycle reduction) and freeze-drying (<0·2 log-cycle reduction) (Fig. 3). However, batch and 0·070 h−1 D rate cultures also had high levels of tolerance, with less than 1-log-cycle reduction. Freezing, compared with freeze-drying, reduced the number of survivors for slow-growing cells, otherwise, it offered a net advantage compared with freeze-drying.

Figure 3.

Tolerance of freezing (black bars) and freeze-drying (grey bars) stresses in batch, fed-batch and steady-state cells of Lactobacillus plantarum C17 grown under respiratory conditions (air 0·2 v/v per min, with 2·5 μg ml−1 hemin and 1 μg ml−1 menaquinone supplementation). Survival was expressed as reduction in log (N/N0), where N0 and N are the number of viable cells before and after exposure to stresses. Letters on plot bars indicate significant differences (Tukey's HSD, P < 0·001) in survival to freezing (a, b, c, d, e) and freeze-drying (A, B, C, D) among growth conditions. Standard error bars are shown.

Selection of random mutants

Due to the impossibility to obtain recombinant strains from Lact. plantarum C17 by site-directed mutagenesis (because of its intrinsic resistance to erythromycin and tetracycline, and of a high membrane hydrophobicity; data not shown; Guidone et al. 2013b), we tried to exploit the selective pressure due to ROS (H2O2 and menadione, a superoxide generating compound) to increase the frequency of mutants with enhanced oxidative stress tolerance and respiratory capability. We randomly picked colonies from plates with the highest tolerated concentration of H2O2 or menadione (Table 2) and tested all mutants for their robustness to respiratory metabolism and oxidative stress conditions.

Most of mutants (from 80%, when anaerobic precultures were used, to 100% when aerobic precultures were used) were able to grow better than wild-type (wt) strain C17 under both anaerobic and respiratory conditions (Fig. 4, section S3 of the graph) and, generally, had higher tolerance of oxidative stress, especially in the presence of menadione. The type of preculture significantly changed the fitness of strains, giving improved growth performance (higher OD ratio mutant/wt, Fig. 4, section S3 of the graph) and stress resistance with anaerobic inocula. To confirm the significance of data analysis, one-sample t-test, with Bonferroni protection, was carried out on the mutant/wt OD ratios for all growth conditions (P value <0·05) and the alternative hypothesis H1: ratio>1·0 (corresponding to the wt/wt OD ratio) was used to calculate the 95% confidence interval. Confidence intervals (CI at 95%) calculated on the raw OD data of wt C17 were, respectively, ±0·013 and ±0·015 for anaerobic and respiratory growth starting with anaerobic inocula and, respectively, ±0·014 and ±0·013 for anaerobic and respiratory growth starting with aerobic precultures.

Figure 4.

Comparison between the 60 mutants and the wild-type (wt) strain Lactobacillus plantarum C17 on the basis of growth capability (OD ratio, mutant/wt) under anaerobic (AN) and respiratory (RS) promoting conditions (cultivation in the presence of 2·5 μg ml−1 hemin and 1 μg ml−1 menaquinone). Black triangles and grey circles indicate, respectively, the use of anaerobic or respiratory precultures. Plot was divided into 4 sections by dashed lines because OD ratio > 1 indicates the capability to grow better than the wt strain.

The distribution of potential respiratory mutants is shown in Fig. 5a. Only few mutants (section S3 of the graph) exhibited the common traits (concurrent increase of OD650 nm and pH values; ratios RS/AN > 1; Brooijmans et al. 2009b) of respiratory pathway when cultivated in WMB with hemin and menaquinone supplementation. The inability of most mutants to shift towards respiratory metabolism may explain the phenotypic advantage (first screening; Fig. 4) observed in the mutants obtained from anaerobic precultures. With exception of C17-m56 mutant, respiratory phenotypes were evident after 24 h of incubation. Several mutants showed an increased biomass production, compared with wt strain (Fig. 5b), in both growth conditions and times of incubation (section S3 of the graph), while others only in anaerobiosis (section S2) or in the presence of O2 and cofactors (section S4).

Figure 5.

Panel (b) Distribution of 20 selected mutants and their wild-type (wt) strain Lactobacillus plantarum C17 cultivated for 16 h (grey circles) and 24 h (black triangles) under anaerobic (AN; static screw-cap tubes) and respiratory (RS) promoting conditions (shaking buffled flasks, in presence of 2·5 μg ml−1 hemin and 1 μg ml−1 menaquinone); the ratios of OD650 nm (OD ratio RS/AN) and pH (pH ratio RS/AN) values measured in both growth conditions are plotted; graph was divided into 4 sections by dashed lines to better represent the possible respiratory strains (both OD and pH ratios higher than 1). Panel (b) Comparison between the 20 selected mutants and the wt strain on the basis of biomass production (Biomass ratio, mutant/wt) after 16 h (grey circles) and 24 h (black triangles) of incubation under anaerobic (AN) and respiratory (RS) promoting conditions. Plot was divided into 4 sections by dashed lines since “Biomass ratio” >1 indicates the capability to grow better than the wt strain.

The mutants C17-m19 (higher biomass under respiratory conditions) and C17-m58 (higher biomass in both conditions) were selected to carry out a growth experiment in the presence of glucose and maltose (a non-PTS sugar for which the repression of pox genes by CcpA is not observed; Lorquet et al. 2004). Results are shown in Table 3. Both mutants were more competitive than the wild-type in terms of growth rate and biomass production, especially when cultivated on maltose and under respiratory conditions. Differences in biomass were more evident after 24 h of incubation and the mutant C17-m58 exhibited the highest level of biomass. In the presence of maltose, even if the biomass was higher, a significant increase in pH was observed, suggesting a greater shift towards the aerobic/respiratory metabolism.

Table 3. Growth parameters of Lactobacillus plantarum C17 (wild-type) and its mutants C17-m19 and C17-m58 under static (anaerobiosis) and shaking (aerobiosis promoting respiration) conditions
StrainsGrowth conditionSugarμmax (h−1)8 h24 h
pHX (g l−1)pHX (g l−1)
  1. Mean values±standard error are shown.

  2. μmax = maximum specific growth rate, estimate ± standard error.

  3. Growth condition: AN = anaerobiosis, cultivation in screw-cap tubes filled with buffered WMB containing 0·1 mol l−1 MOPS, initial pH 6·5; RS = aerobiosis promoting respiration (shaking in buffled flasks with buffered WMB pH 6·5, containing 2·5 μg ml−1 haemin and 1 μg ml−1 menaquinone).

  4. Sugar: GLU, glucose (10 g l−1); MAL, maltose (10 g l−1).

  5. X = net production of biomass (XtX0: Xt is the final biomass, X0 is the initial biomass).

C17 (wild-type)ANGLU0·54 ± 0·024·19 ± 0·011·60 ± 0·024·14 ± 0·021·59 ± 0·02
MAL0·53 ± 0·014·30 ± 0·011·82 ± 0·004·12 ± 0·001·70 ± 0·00
RSGLU0·49 ± 0·024·23 ± 0·011·34 ± 0·084·49 ± 0·001·43 ± 0·08
MAL0·52 ± 0·024·33 ± 0·041·58 ± 0·014·70 ± 0·012·52 ± 0·01
C17-m19 (mutant)ANGLU0·53 ± 0·024·14 ± 0·001·74 ± 0·044·12 ± 0·001·64 ± 0·04
MAL0·52 ± 0·014·25 ± 0·011·93 ± 0·014·12 ± 0·011·99 ± 0·01
RSGLU0·61 ± 0·024·18 ± 0·001·49 ± 0·034·52 ± 0·001·85 ± 0·03
MAL0·53 ± 0·024·27 ± 0·011·73 ± 0·004·72 ± 0·022·41 ± 0·00
C17-m58 (mutant)ANGLU0·58 ± 0·014·17 ± 0·001·77 ± 0·034·11 ± 0·011·73 ± 0·03
MAL0·50 ± 0·014·29 ± 0·001·92 ± 0·014·15 ± 0·011·96 ± 0·01
RSGLU0·60 ± 0·014·18 ± 0·011·57 ± 0·024·49 ± 0·021·79 ± 0·02
MAL0·55 ± 0·014·29 ± 0·011·80 ± 0·024·71 ± 0·012·74 ± 0·02


We studied, for the first time, growth, enzymatic activities, metabolite production and stress tolerance of a Lact. plantarum strain grown under controlled respiratory conditions in batch, fed-batch and continuous culture. Although aeration and supplementation with heme and menaquinone did result in significant oxygen consumption and in a relatively low lactic acid yield especially under some conditions, little or no acetic acid was produced. Shift towards aerobic metabolism in the absence of supplementation with heme or menaquinone generally results in Lact. plantarum in the conversion of pyruvate into acetate via pyruvate oxidase (POX) and acetate kinase (ACK) activities (Goffin et al. 2006; Quatravaux et al. 2006). However, as demonstrated in a previous study (Zotta et al. 2013), POX activity was significantly affected by temperature (being active only at 25°C, in vitro assay), justifying the limited or null production of acetic acid (in vivo). The loss of POX stability at temperatures beyond 32°C was demonstrated by Risse et al. (1992), and Tittmann et al. (1998) showed that the activation of the ternary complex FAD-thiamine-pyruvate oxidase is favoured by low temperature (more at 10°C compared to 25°C). In this study, we observed that POX and NPR (which promotes the degradation of H2O2 with water formation using NADH+H+ as a donor; Quatravaux et al. 2006) were similarly regulated by oxygen and temperature, while NOX activity (which leads the oxidation of NADH+H+ generating H2O2; Quatravaux et al. 2006) was not necessarily related to aerobic growth and temperature. In this study, the activity of enzymes, which use NADH as a cofactor, NOX and NPR, was positively correlated with low growth rate, while POX activity was mainly detected in fast-growing cells, when significant amounts of residual glucose were still available, in contrast with the previous observations about carbon catabolite repression of pox gene (Lorquet et al. 2004). Stevens et al. (2008), also, found that poxF gene could be expressed in exponential phase (when sugar are not limiting for the growth), suggesting that other factors should be taken in account in the regulation of oxygen pathway and/or that the control of aerobic metabolism is strain-specific.

Metabolism via POX–ACK activities generates H2O2 and CO2, while NOX activity generates H2O2. In this study, H2O2 was detected at low concentration only in batch supernatants (exponential growth phase), because of catalase activity in the other growth conditions and/or of low POX activity. Previously, Guidone et al. (2013a) and Zotta et al. (2013) demonstrated that Lact. plantarum C17 was unable to produce the enzyme under anaerobic (without oxygen and heme supplementation) or nonsupplemented aerobic (with oxygen but without heme supplementation) cultivation, confirming the presence of a heme-dependent catalase (Abriouel et al. 2004) in this strain.

Although limited or no POX activity was found (in vitro test), a significant O2 consumption was measured in chemostat cultures, suggesting that the O2 uptake was mainly related to respiratory pathway (functionality of cytochrome oxidase and activation of ET chain). Fed-batch and cultivation at low D rates increased O2 uptake by respiratory cells. The nutrient depletion that occurs during slow-growing conditions, presumably, induces the cells to a greater demand for energy supply that may be satisfied through the extra ATP generation (Pedersen et al. 2012) resulting from the activation of ET chain, which in turn stimulates the consumption of O2 as electron acceptor.

Among the stress conditions to which LAB are subjected, oxidative stress is one of the most significant for the ecological, technological and health-related implications. Oxidative damage is mainly related to the presence of oxygen; using a range of flavin-dependent oxidases, in fact, LAB can produce reactive oxygen species (ROS; hydrogen peroxide H2O2, superoxide anion O2, hydroxyl radical ˙OH), which are toxic to cell structures, resulting in the breaking of peptide bonds, depolymerization of nucleic acids, oxidation of membrane lipids, polysaccharides and fatty acids. To overcome the noxious effects of oxygen and ROS, some LAB have developed defence systems mainly based on the synthesis of antioxidant enzymes, such superoxide dismutase, catalase, peroxidases and oxidases (Casselin et al. 2011). H2O2 is also known to be mutagenic in Lact. plantarum (Machielsen et al. 2010). Our results show that, in Lact. plantarum, respiratory pathway and low growth rate cultivation significantly alleviate oxidative stress, compared with batch and high growth rate continuous cultivations, and that this is correlated with NPR activity and, to a lesser extent, with catalase. Respiration and limited carbon supply, favouring O2 elimination by activation of ET chain, reduce ROS generation and protect cells from harmful conditions. This evidence was supported by the findings of Rezaiki et al. (2004) who demonstrated an increased oxidative stress tolerance in a sod gene mutant of Lc. lactis under respiratory state and, more recently, by Watanabe et al. (2012a) who reported the positive effect of respiration on the survival to H2O2 of Lact. plantarum WCFS1.

The sensitivity of Lact. plantarum C17 to heat, freeze-thaw and freeze-drying stresses was also affected by specific growth rate, suggesting that slow-growing conditions or nutrient depletion lead to stationary phase-like state, which improve general stress response (GSR), activating the induction of cross-protection mechanisms (van de Guchte et al. 2002; Zangh et al. 2011). Freeze-drying, compared with freezing, has proved to be the most effective storage method for slow-growing cells, probably because of increased production of detoxifying enzymes (catalase, NPR) in these cultures, which may alleviate oxidative stresses during freeze-drying and during recovery of freeze-dried cells and may contribute to the survival of heat-stressed cells by relieving oxidative damage (van de Guchte et al. 2002; Zangh et al. 2011). Therefore, combining slow growth cultivations and respiratory pathway, it is possible to achieve high-cell-density cultures suitable for starter development and long-term storage.

A further purpose of this study was to enhance the robustness of Lact. plantarum C17 by natural selection based on oxidative stress tolerance and therefore 60 spontaneous mutants obtained during respiratory growth in continuous cultures were compared with the wild-type strain for their growth response and resistance to oxidative conditions. The best growing and tolerant strains derived from fed-batch and low D rate cultures; additionally and surprisingly, most of them were obtained in the presence of menadione as a selective agent. Although H2O2 is known to be a strong mutagenic agent in Lact. plantarum, it is unlikely that it contributed to random mutations during continuous cultivation, because it was undetectable because of high catalase activity. We therefore conclude that prolonged growth under respiratory conditions may select for mutants with improved fitness under aerobic conditions. The result was very interesting because the growth via respiration and limited cell-carbon supply could greatly improve the tolerance of superoxide anions in this important species, which lacks the sod gene in the genome. Archibald and Fridovich (1981) showed that aerobic cells of Lact. plantarum use an alternative mechanism to scavenge superoxide free radicals by accumulating high levels of intracellular manganese; successively, Watanabe et al. (2012b) confirmed the important role of manganese in preventing oxidative stress in respiratory cultures of Lact. plantarum WCFS1. Probably, Lact. plantarum C17 mutants adopted a similar defence system because the colonies exposed to high concentrations of ROS generators had an unusual silvery appearance.

The mutants C17-m19 (isolated from fed-batch culture, in the presence of menadione) and C17-m58 (isolated from lowest D rate culture, in the presence of menadione) with the best performances were tested to verify the role of different carbon sources in the activation of aerobic metabolism. As suggested by Lorquet et al. (2004), shift towards aerobic pathway is favoured by the presence of maltose compared with glucose. The non-PTS sugar supplementation also reduced acidification under anaerobic cultivation (early stationary growth phase) promoting biomass production in both wild-type and mutant strains, suggesting its use as a carbon source for high-cell-density production.

This work confirms that, in Lact. plantarum, respiration and low growth rates confer physiological advantages under unfavourable conditions. However, some aspects on the regulation and metabolic pathway of respiratory growth in this species still need to be clarified. The main questions are related to the regulation and activation of POX enzyme, because our results suggest that carbon catabolite repression, aeration and the presence of H2O2 are not the only factors affecting the growth via respiration, and the implication of temperature could be also possible. Additionally, the production of acetate as distinctive trait of the aerobic metabolism is controversial because oxygen utilization in Lact. plantarum is strain-specific (Guidone et al. 2013a) and the system POX–ACK is not similarly active in strains with different physiological properties; moreover, significant amount of acetate could be also detected in anaerobic cultures, as product of lactate-independent pathways, making difficult the discrimination between the two types of metabolism.

Our strategy of natural selection may be a tool of practical relevance: the mutants generated in this study, in fact, have improved growth and oxidative stress tolerance compared with the wild-type strain C17, although the nature of the mutations resulting in the improved phenotype is not clear. Because no selectable markers were used as in site-directed mutagenesis with integrative plasmids or transposons (Bron et al. 2004; Hüfner et al. 2007; Perpetuini et al. 2013; Sasikumar et al. 2013), the identification of the mutations would require whole-genome sequencing of the mutants and of the wild-type, which is beyond the scope of this work. The mutant C17-m58 has been selected for further comparative studies (chemostat cultivation in defined composition medium) to clarify some metabolic aspects about respiration in Lact. plantarum. Finally, because within the European Union the use of genetically modified LAB (GM-LAB) in food production is not yet fully applied and accepted either by legislation and consumers (Sybesma et al. 2006), the induction of spontaneous mutations by natural events (although generates uncontrolled and unknown gene modifications) represents a food-grade selection strategy.


This work was partly funded by Ministero dell'Istruzione, dell'Università e della Ricerca, Rome, Italy, PRIN n. 359 20088SZB9B.

Conflict of interest

No conflict of interest declared.