Assessment of Aerobic and Respiratory Growth in the Lactobacillus casei Group

One hundred eighty four strains belonging to the species Lactobacillus casei, L. paracasei and L. rhamnosus were screened for their ability to grow under aerobic conditions, in media containing heme and menaquinone and/or compounds generating reactive oxygen species (ROS), in order to identify respiratory and oxygen-tolerant phenotypes. Most strains were able to cope with aerobic conditions and for many strains aerobic growth and heme or heme/menaquinone supplementation increased biomass production compared to anaerobic cultivation. Only four L. casei strains showed a catalase-like activity under anaerobic, aerobic and respiratory conditions and were able to survive in presence of H2O2 (1 mM). Almost all L. casei and L. paracasei strains tolerated menadione (0.2 mM) and most tolerated pyrogallol (50 mM), while L. rhamnosus was usually resistant only to the latter compound. This is the first study in which an extensive screening of oxygen and oxidative stress tolerance of members of the L. casei group has been carried out. Results allowed the selection of strains showing the typical traits of aerobic and respiratory metabolism (increased pH and biomass under aerobic or respiratory conditions) and unique oxidative stress response properties. Aerobic growth and respiration may confer technological and physiological advantages in the L. casei group and oxygen-tolerant phenotypes could be exploited in several food industry applications.


Introduction
The Lactobacillus casei group includes three closely related species (Lactobacillus casei, L. paracasei and L. rhamnosus), involved in different food and health-related applications [1][2][3]. Their wide ecological distribution (human host, vegetable, meat and dairy products) and the potential role as probiotics makes these species interesting for the development of new functional foods and relevant for several genetic and physiological studies.
The taxonomy of the L. casei group is controversial and many studies have addressed the identification and genotypic characterization of strains belonging to the species L. casei, L. paracasei and L. rhamnosus [1], [4]. Recently, comparative genomic studies [5][6][7] highlighted the heterogeneity of L. casei group, suggesting that genome diversification contributes to ecological niche adaptation in these species. However, the presence of genes not always reflects their functionality and the physiological and technological properties of strains mainly depend on expressed features in a given set of conditions.
Like other lactic acid bacteria (LAB), L. casei, L. paracasei and L. rhamnosus are considered oxygen-tolerant anaerobes with fermentative metabolism, which normally lack both catalase and an active electron transport chain (ETC). The growth condition and the type of metabolism significantly affect the stress responses in LAB and, recently, several studies have demonstrated that conditions which promote aerobic and respiratory growth (presence of oxygen, heme and/or menaquinone in the substrate) induce in Lactococcus lactis and L. plantarum useful traits (improved biomass production and stress tolerance) for industrial and biotechnological applications [8], [9].
Supplementation with heme may promote the synthesis of a heme-dependent catalase and a bd-type cytochrome oxidase [8], [9]. Catalase protects cells against oxidative stress by degrading hydrogen peroxide (H 2 O 2 ), while cytochrome bd oxidase, the final component of the minimal respiratory chain in LAB [8], [9], contributes to energy supply (through extra ATP generation) and depletion of intracellular oxygen. The biogenesis of catalase and cytochrome bd oxidase seems to be uncorrelated [10] and, in some cases, species lacking respiratory capability (L. sakei) encode a heme-dependent catalase (kat), while well characterised respiring species (Lc. lactis) lack the kat gene.
Menaquinone (vitamin K2), which acts as an electron shuttle in respiratory chain of LAB, is found in many vegetable and meat products and dietary intake may contribute to human health [11]. The ability to synthesize quinones varies among LAB: Lc. lactis subsp. cremoris MG1363, using the complete (mena)quinones biosynthesis complex menFDXBEC, produces and exploits menaquinone for respiratory growth, while other species (including L. plantarum) need exogenous menaquinone supplementation to perform respiration [9].
To date, with the exception of studies on L. plantarum [12][13][14][15][16][17][18][19], reports on the aerobic and/or respiratory metabolism in other species of Lactobacillus are rare [20][21][22]. Moreover, data on stress response mechanisms of L. casei group are limited to a small number of strains and conditions and, generally, have been carried out on cells grown in anaerobiosis.
L. casei, L. paracasei and L. rhamnosus are widely distributed in plant, animal and human-associated habitats, in which oxygen, heme and menaquinone may be present. Tolerance of oxygen and oxidative stresses may be important in the survival in different environments, including the gut, and during preservation of starter and probiotic cultures [8], [9]. In this work, we investigated the capability of a diverse collection of L. casei, L. paracasei and L. rhamnosus strains to cope with the presence of oxygen, ROS (reactive oxygen species) generating compounds, heme and menaquinone. The shift towards aerobic and respiratory growth has been considered, for the first time, in these species in order to identify respiration-competent strains and exploit the oxygentolerant phenotypes for development of improved starter and probiotic cultures. One hundred eighty four strains belonging to the species Lactobacillus casei, L. paracasei and L. rhamnosus, isolated from different sources (Table 1), were used. All strains were identified to the species level using a polyphasic approach (SDS-PAGE, DGGE-PCR, specific-PCR, multiplex-PCR, High Resolution Melting Analysis) and genotyping was performed by RAPD-PCR, Rep-PCR, Sau-PCR (Iacumin et al. 2014, unpublished data).

Tolerance of Oxidative Stresses
A preliminary test was performed to select the concentration of H 2 O 2 or ROS generators (menadione or pyrogallol) which provided the best discrimination between sensitive and tolerant strains. Ten strains (final OD 450 = 2.0) were randomly chosen and cultivated (inoculum 10% v/v, 16 h at 37uC in microplates) in MRS with 0.16 g/L bromocresol purple (MRS-BCP), pH 6.8, containing H 2 O 2 (ten two-fold dilutions from 880 to 1.7 mM) or pyrogallol (ten two-fold dilutions from 200 to 0.4 mM) or spotted (5 mL) on MRS agar plates containing menadione (9 two-fold dilutions from 0.4 to 0.0015 mM). Change of colour from purple to yellow (H 2 O 2 and pyrogallol) or spot development (menadione) was considered as positive results. Appropriate concentrations of H 2 O 2 (1 and 2 mM), pyrogallol (25 and 50 mM), menadione (0.15 and 0.2 mM) were selected and all strains were exposed to oxidative stresses as described above.

Assessment of Respiratory Growth in Selected Strains
The effect of menaquinone supplementation was further evaluated in 60 strains (shown in boldface in Table 1) selected on the basis of stress response properties (heat, acid, osmotic, bile; Reale et al. 2014, unpublished data) and capability to tolerate oxygen and ROS (this study). WMB medium [17] was used to ensure the absence of heme (i.e. meat extract in MRS) during anaerobic and unsupplemented aerobic growth and a lower glucose concentration (10 g/L instead of 20 g/L; WMB10) was used to reduce the effect of carbon catabolite repression, if any, on the shift towards aerobic and respiratory metabolism [12]. Cultivation was carried out in 24-well microplates in: a. anaerobiosis (AN), b. aerobiosis (AE) and c. respiratory promoting condition (RS; AE cultivation in presence of 2.5 mg/mL hemin and 1 mg/mL menaquinone). Microplates (1 mL substrate/well) were inoculated with 20 mL of standardized (final OD 650 = 1.0)    AN, AE and RS cell suspensions (final OD 650 = 1.0) were exposed (30 min, 37uC) to serial dilutions of H 2 O 2 (ten two-fold dilutions from 880 to 1.7 mM). The survivors (if any) were cultivated in microplates as described before. Change of colour from purple to yellow and turbidity were considered as indication of the presence of survivors. Catalase-like activity was measured on the AN, AE and RS cell free extracts (obtained by mechanical lysis in FastPrep-24 Instrument, MP Biomedicals, Santa Ana, California, USA; 5 cycles of 60 sec at speed 6.0) according to the modified protocol of Risse et al. [27]. Briefly, AN, AE and RS samples were first incubated (15 min, 37uC) with 16 mM H 2 O 2 (final concentration) and successively (10 min, 37uC) with a mixture containing 4-amino-antipyrine (3 mmol/L), sodium 3,5dichloro-2 hydroxybenzenesulfonate (10 mmol/L) and peroxidase (0.28 U/mL). The residual amounts of H 2 O 2 were spectrophotometrically measured at 510 nm. One mkatal (mkat) was defined as the amount of enzyme required to degrade 1 mmol H 2 O 2 /s. All measurements were run in duplicate.  Table 1), were selected for further kinetic growth studies. All strains were re-cultivated at 37uC in: a. anaerobiosis (AN, screwcap tubes filled with WMB10 containing 0.1 M MOPS, buffered WMB10, initial pH 6.8), b. aerobiosis (AE; 250 mL baffled shaking flasks with 50 mL buffered WMB10, agitation on a rotary shaker, 150 rpm) and c. respiration (RS, AE growth in presence of 2.5 mg/mL hemin and 1 mg/mL menaquinone). Anaerobic precultures were used as inocula (5 log cfu/mL). Samples were aseptically withdrawn (30 min-interval) until to the late exponential growth phase (10 h) and after 24 h of incubation. Viable counts were performed on WMA (WMB pH 6.8 with 1.2% w/v agar), using a Whitley Automated Spiral Plater 2 (WASP2; Don Whitley Scientific Limited, UK), and colonies were enumerated with the EasyCount 2 colony counter (bioMérieux) after 48 h of incubation at 37uC in anaerobisosis. Lag time and maximum specific growth rates (m max ) were estimated with the primary biphasic model of Baranyi and Roberts [25] using the DMFit v 2.0 program [26].
H 2 O 2 in supernatants and catalase activity in cell free extracts (obtained by mechanical lysis as described above) were measured as described by Risse et al. [27]. The activities of enzymes related to the oxygen metabolism (pyruvate oxidase, POX; NADH oxidase, NOX; NADH peroxidase, NPR) were measured as described by Quatravaux et al. [13]. Protein concentration was measured using the Bradford method [28]. Oxygen uptake by AN, AE and RS cells was evaluated using the resazurin assay [24]. All growth experiments and analytical measurements were run in duplicate.
Unidirectional (genes vs genomes) sequence similarity was detected using the IMG tools modifying, for each selected gene and genome, the BLAST cut-offs parameters (E-value, minimum % of identity).

Statistical Analyses
Statistical (analyses of variance, correlations, two-way contingency tables) and graphic analyses were performed using SYSTAT 13.0 for Windows (Systat Software Inc., Richmond, CA, USA), while the Matrix Hierarchical Cluster Analysis (normalized data, Pearson distance, Average linkage UPGMA method) was obtained with PermutMatrix program v. 1.9.3 (LIRMM, France).

Heterogeneity of L. casei Group in the Aerobic Growth and ROS Tolerance
All strains of L. casei group were able to cope with aerobic conditions, even if a large variability in growth behaviour was found (Figures 1a and 1b). For many (about 70%) strains (mostly belonging to the species L. casei and L. rhamnosus) the presence of oxygen and heme-supplementation enhanced growth compared to anaerobic cultivation, while for some L. paracasei strains aerobiosis and heme apparently impaired growth. Heterogeneity in OD and pH values, between anaerobic and aerobic conditions, was less noticeable after 42 h of incubation (data not shown).
All strains grown in heme-supplemented aerobiosis were also tested for their tolerance of ROS generating compounds to evaluate if aerobic growth (adaptation to oxygen and activation of antioxidant enzymes) and heme (synthesis of heme-dependent catalase and cytochrome bd oxidase) improved tolerance of oxidative stresses.
With the exception of 4 L. casei strains (CI4368, N87, N811, N2014), which surprisingly showed a catalase-like activity and survived to 2 mM of H 2 O 2 , none of the strains tolerated even the lowest (1 mM) H 2 O 2 concentration used in this study. All strains (except LMG6904 and LC3) of L. casei and the 16% of L. paracasei survived to 0.2 mM menadione, while only 5 strains L. rhamnosus were tolerant. Most strains (100% of L. casei, 97% of L. rhamnosus and 76% of L. paracasei), instead, tolerated well the exposure to   (Table 2). Oxidative stress tolerance was not significantly associated to the isolation source.

Assessment of Respiratory Growth in Selected Strains of L. casei Group
The effect of menaquinone and, thus, the activation of a possible respiratory pathway, were further evaluated on the growth 60 selected strains.
Growth and acid production in anaerobiosis, unsupplemented aerobiosis (AE) and respiration (RS) are compared in Figures 2a  (AE, 16 h), 2b (AE, 24 h), 2c (RS, 16 h) and 2d (RS, 24 h). Ratios between optical density (OD) and pH values measured in AE, RS and AN cultivations were calculated to identify oxygen-tolerant and respiration-competent phenotypes. Several strains (section S3 of the graphs) exhibited a concurrent increase of OD 650 and pH values (OD and pH ratios .1; these are the common traits of aerobic and respiratory growth; 20) when cultivated in presence of air (AE/AN ratios) or air and cofactors (RS/AN ratios), and some of them (14 strains) were able to consume oxygen in some or all conditions (black symbols in the section S3 of the graph). A smaller number of isolates (9 strains) grew better only under anaerobic conditions (section S2 of the graphs).
The effect of heme/menaquinone supplementation on growth (OD ratio between RS and AE conditions at 16 h and 24 h) of 51 oxygen-tolerant strains was also evaluated (Figure 3). The presence of respiratory cofactors generally improved growth compared to unsupplemented aerobiosis (Figure 3, section S3), but for a few strains seemed to impair growth (Figure 3, section S1).
The capability to consume oxygen in LAB may be related to activity of flavin-dependent oxidases (pyruvate oxidase, POX; NADH oxidase, NOX), which may be produced under both AE and RS condition, or to the activity cytochrome oxidase in respiratory ET chain (RS growth only). The time of resazurin discoloration (as indication of oxygen uptake) was shorter in presence of heme and menaquinone (ratio RS/AE of discoloration time ,1, at both 16 h and 24 h,), suggesting a boost of oxygen uptake by the cytochrome oxidase activity in respiratory cells. The reduction of resazurin measured at 16 h (late exponential phase) and 24 h (stationary phase) was uncorrelated in both aerobic (r = 0.282) and respiratory (r = 0.556) cells, suggesting that the activity of enzymes involved in the oxygen utilization may be related with the growth phase and physiological state of cells.

Classification of Strains on the Basis of Anaerobic, Aerobic and Respiratory Growth Pattern
To correlate the parameters of aerobic and respiratory growth (increases of biomass and pH, oxygen uptake), a Matrix Hierarchical Cluster Analysis (MHCA; Figure 4) was performed on the 60 selected strains using as variables the rate of resazurin reduction and the concurrent increase of biomass and pH values (OD and pH ratios AE/AN or RS/AN .1) measured in anaerobic, aerobic and respiratory cells, at both 16 and 24 h. A z-value transformation was used for all variables.
Classification generated 3 major clusters that allowed to distinguish the strains in: oxygen-sensitive anaerobes (cluster C1; exclusively L. paracasei), unable to consume oxygen and for which aerobic and respiratory conditions impaired growth compared to anaerobic cultivation; respiration-competent strains (cluster C2; exclusively L. casei and L. rhamnosus), with oxygen uptake and increased biomass production in both AE and RS and pH in RS conditions resulting, possibly, by the shift towards aerobic and respiratory pathways; a large heterogeneous group (cluster C3) including oxygen-tolerant anaerobes, with increased growth under AE and or RS conditions, but with limited oxygen consumption ability.

Investigation of Factors Affecting the Lowest Adaptation to Aerobic Growth in L. paracasei
Since L. paracasei exhibited the lowest adaptation to the aerobic growth ( Figure 3, grey triangles; Figure 4, cluster C1 and C3), we investigated the duration of lag phase, the m max values, the reduction in aerobic and respiratory cells after 16 and 24 h of incubation (columns 1, 2, 3, 4); pH ratios .1, AE/AN and RS/AN, after 16 h and 24 h of incubation (columns 5, 6, 7, 8); OD ratios .1, AE/AN and RS/AN, after 16 h and 24 h of incubation (columns 9, 10, 11, 12). Row dendrogram: strain_species. Colour scale: from green (negative data; minimum value is 24.51) to red (positive data, maximum values is +4.51), indicates the change from the mean in standard deviation units. doi:10.1371/journal.pone.0099189.g004 Figure 5. Relationships between maximum specific growth rates (Y-axis) and lag phases (X-axis) estimated in 5 L. paracasei strains (B061, C4H8, R61, SP57, V3) cultivated in anaerobiosis (light grey circles), aerobiosis (grey triangles) and respiratory (black squares) promoting condition. doi:10.1371/journal.pone.0099189.g005 activities of catalase, POX, NOX and NPR, the production of H 2 O 2 and the capability to consume oxygen in 5 strains showing different phenotypes when cultivated in AN, AE and RS conditions. The biphasic model of Baranyi and Roberts [25] provided an excellent fit for all growth curves (R 2 from 0.975 to 0.998). Anaerobic inocula were used for all cultivations and, thus, the time of adaptation mainly depended on the presence of oxygen, heme and menaquinone. Tolerance of oxygen and utilization of respiratory cofactors differed among strains. Supplementation and, to a lesser extent, aerobiosis significantly increasing lag phase of strain C4H8, while no lag phase was observed in AN cultivation), but reduced the lag phase in strains R61 and B061. Heme and menaquinone delayed the entry in exponential phase for the strains SP57 and V3, while unsupplemented aerobiosis seemed to offer a net gain in their growth, compared to AN cultivation ( Figure 5).
Catalase, POX, NOX, NPR and oxygen consumption were undetectable at 10 h of incubation, suggesting that the maximum specific growth rate (m max ) and the cell number at the end of exponential growth phase were affected by oxygen and heme/ menaquinone inhibition, rather than by the production of enzymes related to the aerobic metabolism. Cell numbers measured at the end of exponential phase were significantly correlated to growth rates (r = 0.967) but, surprisingly, the lag phase and m max were positively correlated for some strains (C4H8 and V3; a longer adaptation period positively affected the growth rate) or completely uncorrelated for other (SP57, R61 and B061, Figure 5).
Contrarily to catalase (which was never detected), POX, NOX and NPR were detected after 24 h of cultivation. POX activity was found only in aerobic and respiratory growing cells (Figure 6a), suggesting that the enzyme is strictly related to aerobic growth. The presence of POX, that leads the conversion of pyruvate into acetate by POX-acetate kinase (ACK) pathway, was also confirmed by the increased pH in aerobic and respiratory cultures. Strain V3, which grew better only in anaerobiosis, was discarded from Figure 6 (a, b, c) because it was unable to synthesise any of the flavin-depend oxidases, which may explain its oxygen-sensitive phenotype. Cell numbers at the end of incubation (24 h) were significantly (p value ,0.05) correlated with the activities of POX (r = 0.694), NOX (r = 0.710) and NPR (r = 0.614) (Figure 6a, b, c) and, with exception of strain SP57 for which heme and menaquinone supplementation had a negative effect on growth, increased in aerobiosis and even more in heme and menqauinone supplenented aerobiosis, when significant levels of POX, NOX, NPR were measured. Although no catalase activity was found, H 2 O 2 was undetectable in aerobic/respiratory supernatants, probably because of degradation by NPR. Despite the presence of flavin-dependent oxidases, no singificant oxygen consumption was observed, suggesting that the strains were able to tolerate oxygen and inactivate ROS, but were unable to use it as a final electron acceptor in the respiratory chain (all strains belonged to cluster C3 in Figure 4).