Accumulation of Polyphosphate in Lactobacillus spp. and Its Involvement in Stress Resistance

Abstract

Polyphosphate (poly-P) is a polymer of phosphate residues synthesized and in some cases accumulated by microorganisms, where it plays crucial physiological roles such as the participation in the response to nutritional stringencies and environmental stresses. Poly-P metabolism has received little attention in Lactobacillus, a genus of lactic acid bacteria of relevance for food production and health of humans and animals. We show that among 34 strains of Lactobacillus, 18 of them accumulated intracellular poly-P granules, as revealed by specific staining and electron microscopy. Poly-P accumulation was generally dependent on the presence of elevated phosphate concentrations in the culture medium, and it correlated with the presence of polyphosphate kinase (ppk) genes in the genomes. The ppk gene from Lactobacillus displayed a genetic arrangement in which it was flanked by two genes encoding exopolyphosphatases of the Ppx-GppA family. The ppk functionality was corroborated by its disruption (LCABL_27820 gene) in Lactobacillus casei BL23 strain. The constructed ppk mutant showed a lack of intracellular poly-P granules and a drastic reduction in poly-P synthesis. Resistance to several stresses was tested in the ppk-disrupted strain, showing that it presented a diminished growth under high-salt or low-pH conditions and an increased sensitivity to oxidative stress. These results show that poly-P accumulation is a characteristic of some strains of lactobacilli and may thus play important roles in the physiology of these microorganisms.

INTRODUCTION

Inorganic polyphosphates (poly-P) are ubiquitous molecules in living organisms that consist of linear chains from a few to hundreds of phosphate molecules linked by phosphoanhydride bonds (1–3). In some cases this polymer is intracellularly accumulated in characteristic metachromatic inclusions, which were first described a century ago in Spirillum volutans and classically named as volutin granules (4). Poly-P was initially considered a phosphate and energy storage molecule, but subsequent studies have demonstrated its involvement in diverse physiological and regulatory mechanisms in bacteria (1–3). These included the responses to nutrient starvation and other stresses such as oxidative, acid, osmotic, or UV stress (2, 5). In addition, poly-P has been demonstrated to play roles in motility, competence, biofilm formation, or virulence (6, 7). The mechanisms by which poly-P influences all of these processes are still unclear in most cases and could involve a variety of aspects such as gene expression regulation (8), mRNA stability (9), protein turnover (10), or ATP homeostasis (11).

In bacteria, the main enzyme responsible for poly-P synthesis is polyphosphate kinase (Ppk; EC 2.7.4.1), which catalyzes the ATP-dependent formation of a phosphoanhydride bond between a poly-P chain and orthophosphate (3, 12). A second class of Ppk (Ppk2) exists, first characterized in Pseudomonas aeruginosa, which can also synthesize poly-P, but it mainly functions as a polyphosphate-dependent ATP/GTP-regenerating enzyme (13). The pivotal role of Ppk in poly-P synthesis and accumulation is exemplified by the fact that microorganisms with a big potential for poly-P accumulation, such as those involved in biological phosphate removal in activated sludges, carry several copies (up to four) of ppk genes in their genomes (14). The reverse process of poly-P hydrolysis can be carried out by several enzymes, but the most important are the exopolyphosphatases (EC 3.6.1.11), which split phosphate from the ends of poly-P chains (15). Two types of exopolyphosphatases have been characterized: the enzyme encoded by the ppx gene and the guanosine pentaphosphate (pppGpp) phosphohydrolase (GppA; EC 3.6.1.40). Both enzymes are related in sequence, but while Ppx acts on long-chain poly-Ps with high processivity, GppA is also able to hydrolyze pppGpp to ppGpp (15, 16). These two regulatory compounds [referred to as (p)ppGpp] are alarmones that trigger the stringent response induced by phosphate or amino acid starvation (17).

Lactobacillus is a major genus of the lactic acid bacteria (LAB), an economically important bacterial group associated with many food fermentation processes (18). Also, their members are normal inhabitants of human and animal mucosal surfaces, where they are considered beneficial microorganisms, and some of them are used as probiotics. There are very few studies about poly-P in Lactobacillus. The first citation dates from the early 1960s and describes the presence of poly-P in Lactobacillus casei, among other microorganisms (19). Also, while studying the ultrastructure of the membrane of Lactobacillus plantarum, Kakefuda et al. reported the presence of poly-P granules in its cytoplasm (20). Poly-P was later described in L. plantarum as a molecule complexing Mn(II) as a counter ion. In this species, which lacks superoxide dismutase, Mn(II) participates in a particular mechanism of nonenzymatic defense against oxygen toxicity by scavenging O₂⁻ (21, 22). The second reported example of poly-P inclusions in lactobacilli came from the observation that 60% of the Lactobacillus strains isolated from mozzarella cheese whey, particularly L. casei and Lactobacillus fermentum strains, formed inclusion bodies that were stained by the Neisser method and displayed a characteristic yellow fluorescence after DAPI (4′,6′-diamidino-2-phenylindole) staining (23). Recently, poly-P has been identified as an extracellular factor synthesized by Lactobacillus brevis which helps to maintain intestinal homeostasis in in vitro and in vivo models of intestinal inflammation by inducing cytoprotective heat shock proteins (24), thus providing a new functional role for poly-P.

LAB used in food production or as probiotics are subjected to harsh conditions derived from industrial processes or from the passage through the gastrointestinal tract. Stress adaptation and survival thus play a crucial role in the robustness and functionality of strains (25, 26). Owing to the participation of poly-P in the mechanisms of stress response in bacteria but also as a likely probiotic factor in a Lactobacillus strain, 20 species of the genus Lactobacillus were screened in the present study for the formation of poly-P. More than half of the strains tested harbor poly-P inclusions in their cytoplasms. The functionality of the ppk gene in poly-P synthesis was investigated in L. casei BL23, and the effect of its mutation on the resistance to several stresses was assessed.

MATERIALS AND METHODS

Strains and culture media.

Lactobacillus strains (Table 1) were routinely grown in MRS liquid medium (Difco) under static conditions at 37°C. For poly-P granules, observation MEI medium (27) was used containing 0.5% yeast extract, 0.5% tryptone, 0.4% K₂HPO₄, 0.5% KH₂PO₄, 0.02% MgSO₄·7H₂O, 0.005% MnSO₄, 1 ml of Tween 80 liter⁻¹, 0.05% cysteine, and 0.5% glucose. A similar medium was formulated without the addition of K₂HPO₄ and KH₂PO₄ (LP-MEI). Escherichia coli DH10B and Lactococcus lactis MG1363 were used as hosts for cloning, and they were grown in Luria-Bertani medium at 37°C under vigorous shaking and in M17 medium (Oxoid) supplemented with 0.5% glucose at 30°C under static conditions, respectively. Antibiotics used for plasmid selection were ampicillin at 100 μg ml⁻¹ for E. coli and erythromycin and chloramphenicol at 5 μg ml⁻¹ for L. casei and Lactococcus lactis. When these two last antibiotics were used in combination, they were applied at 2.5 μg ml⁻¹ each.

TABLE 1.

Presence of poly-P granules in different Lactobacillus strains and its correlation with the occurrence of ppk genes

Strain	Origin	Poly-Pa	Sequenced genomesb	ppkc
L. acidophilus ATCC 9224	Sour milk	++	Yes	Yes
L. acidophilus ATCC 4356	Human intestine	—	Yes	Yes
L. acidophilus CNRZ216	Rat intestine	+	Yes	Yes
L. animalis CECT4061	Rumen	—	Yes	No
L. brevis NCIB8561	Beer	++	Yes	Yes
L. brevis ATCC 14869	Human feces	+/–	Yes	Yes
L. buchneri BL152	Mouse feces	+/–	Yes	Yes
L. casei BL23	Laboratory strain	++	Yes*	Yes
L. casei ATCC 334	Emmental cheese	++	Yes*	Yes
L. coryniformis ATCC 25602	Silage	++	Yes*	Yes
L. coryniformis ATCC 25600	Air of dairy barn	++	Yes*	Yes
L. crispatus ATCC 33820	Human eye	—	Yes	No
L. curvatus ATCC 25601	Milk	—	Yes	No
L. farciminis ATCC 29644	Sausage	+	Yes	Yes
L. fermentum ATCC 14931	Fermented beets	+/–	Yes*	Yes
L. gasseri 4B2	Human feces	—	Yes	No
L. gasseri ATCC 33323	Human feces	—	Yes*	No
L. intestinalis BL260	Rat cecum	+	No	—
L. intestinalis ATCC 49335	Rat intestine	+	No	—
L. johnsonii BL261	Rat cecum	—	Yes	No
L. johnsonii ATCC 11506	Human intestine	—	Yes	No
L. johnsonii ATCC 33200	Human blood	—	Yes*	No
L. murinus BL262	Rat cecum	—	No	—
L. murinus ATCC 35020	Rat intestine	—	No	—
L. plantarum ATCC 14917	Pickled cabbage	++	Yes*	Yes
L. plantarum WCFS1	Human intestine	++	Yes*	Yes
L. reuteri BL259	Rat cecum	+	Yes	Yes
L. rhamnosus ATCC 7469	Human mouth	++	Yes	Yes
L. rhamnosus GG	Human intestine	++	Yes*	Yes
L. salivarius ATCC 11741	Saliva	—	Yes	No
L. sakei ATCC 15521	Sake	—	Yes	No
L. sakei 23K	Sausage	—	Yes*	No
L. sakei 160*1K	Sausage	—	Yes	No
L. taiwanensis DSM21401	Silage	—	No	—

^aPresence of poly-P granules in bacterial cells grown in MEI medium detected by Neisser straining: ++, granules in almost all cells (>95%); +, granules in less than half of the cells; +/–, granules in a few cells (<10%); –, no granules.

^bExistence of genome sequence data for strains of the same species. An asterisk (*) indicates that a particular strain has been sequenced.

^cPresence of a ppk gene in the sequenced genomes of strains of the same species.

Polyphosphate granules staining.

Lactobacillus strains were grown in MEI or LP-MEI plates (1.8% agar) at 30°C in anaerobic jars (AnaeroGen; Oxoid) for 72 h. Bacterial smears on microscope slides were air dried, and poly-P granules were specifically stained by the procedure of Neisser (28). The smears were covered with a freshly prepared solution consisting in 1 volume of 0.1% methylene blue dissolved in 5% ethanol plus 5% acetic acid, mixed with 2 volumes of 0.33% crystal violet in 10% ethanol. After 1 min, this solution was thoroughly washed with water and cells were stained for 1 min with a solution of 0.3% chrysoidin G in water. This solution was washed with water and, after drying, the slides were observed under an Eclipse 90i microscope (Nikon) at ×100 magnification.

Transmission electron microscopy.

Lactobacillus strains were grown for 48 h on MEI plates, and cells were recovered from the plates with phosphate-buffered saline (PBS; pH 7.4). After being washed with PBS, the bacteria were fixed with 3% paraformaldehyde plus 0.5% glutaraldehyde and embedded in agarose. Samples in agarose were washed with PBS, treated with 2% osmium tetroxide and dehydrated with ethanol. The samples were included in LR White resin (Electron Microscopy Sciences) that was polymerized at 60°C. Ultrathin sections (60 nm) were obtained with a Leica UC6 ultramicrotome and stained with uranyl acetate and lead citrate. Samples were observed with a JEOL JEM-1010 transmission electron microscope, and images were digitally captured with a MegaView III camera (Olympus).

Sequence analysis.

Similarity searches were performed with BLASTP with the genomic BLAST service provided by the National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi) using as query sequences Ppk, Ppx1, and Ppx2 encoded by Lactobacillus casei BL23 (LCABL_27820, LCABL_27830, and LCABL_27810, respectively). Genomic context analysis was performed on genomes deposited at the Microbial Genome Database for Comparative Analysis (MBGD; http://mbgd.genome.ad.jp/) (29). Briefly, an orthology table of all genes present in the genomes of representative strains of Lactobacillales was obtained using the clustering algorithm implemented in MBGD. The orthology table was queried for polyphosphate kinases and polyphosphatases in order to retrieve the corresponding genes. The genes were confirmed as Ppk or Ppx by checking the presence of typical conserved domains: the N-terminal, middle, and duplicated C-terminal domains in Ppk (30) and the N-terminal nucleotide-binding domain of the Ppx-GppA family and C-terminal HDc domains in Ppx (31, 32).

Mutant construction.

A nonreplicative plasmid for ppk disruption in L. casei BL23 was constructed by cloning a 516-bp fragment from the L. casei BL23 ppk gene (LCABL_27820 [33]) into pRV300 plasmid (34). The ppk gene was amplified with oligonucleotides PPK-F1 (5′-TTTTGAATTCTATTCTCTTCATCAAACCGC) and PPK-R1 (5′-TTTTCTCGAGAAGAGGATGCATCTGATTTG) using chromosomal DNA from L. casei BL23 strain as a template. The obtained PCR product was digested with EcoRI and XhoI (restriction sites are underlined in the oligonucleotide sequences), ligated to pRV300 digested with the same restriction enzymes, and transformed in E. coli. The constructed plasmid was used to transform L. casei BL23 by electroporation in a GenePulser electroporator (Bio-Rad) as previously described (35). After electroporation, cells were plated on MRS with erythromycin, and colonies were obtained after 3 days of incubation at 37°C. The disruption of ppk was confirmed by PCR on chromosomal DNA isolated from several erythromycin-resistant clones by using the oligonucleotide PPKcomp (5′-CAGGCGATGACAATGTC), which hybridizes at the 3′-end of ppk, and a forward oligonucleotide from pRV300 (5′-CAGGAAACAGCTATGAC). One ppk-disrupted clone was chosen and designated BL379.

A plasmid for the expression of L. casei BL23 ppk gene was constructed by cloning the ppk gene into pT1NXCAT, a pT1NX derivative (36) in which the erythromycin resistance cassette has been replaced by a chloramphenicol resistance cassette (M. J. Yebra, unpublished results). The ppk gene, together with its translational initiation signals, was amplified by PCR from chromosomal DNA of L. casei BL23 with oligonucleotides PPK-F2 (5′-TGAAGATCTAGGAAGAAACATTTTATGAAG) and PPK-R2 (5′-AATACTAGTTTCATGCGTCATCGTCCTCCTTG). The PCR product was digested with BglII and SpeI (restriction sites introduced for cloning in the oligonucleotides are underlined) and ligated to pT1NXCAT digested with the same enzymes. The ligation mixture was transformed into Lactococcus lactis MG1363 by electroporation as described previously (37) with selection for chloramphenicol resistance and the constructed plasmid (pT1ppk), in which ppk is expressed from the lactococcal P1 promoter (36), was transferred to L. casei BL379. L. casei BL23 and BL379 strains were also transformed with the pT1NXCAT vector.

Polyphosphate isolation and electrophoresis.

Poly-P was isolated by its resistance to hydrolysis with hypochlorite (38). The different L. casei strains were grown in MEI plates for 48 h at 37°C. The bacterial cells were recovered from the plates with 2 ml of 0.9% NaCl and washed with an additional volume of 0.9% NaCl. Bacterial cells equivalent to 1 ml of a cell suspension at an optical density at 595 nm (OD₅₉₅) of 3 were lysed in 1 ml of 5% sodium hypochlorite with gentle agitation for 45 min at room temperature. Insoluble material was pelleted by centrifugation at 16,000 × g for 5 min at 4°C and washed twice with 1 ml of 1.5 M NaCl plus 1 mM EDTA at 16,000 × g for 5 min, 4°C. Poly-P was extracted from the pellets with two consecutive washes with 1 ml of water and a centrifugation step at 16,000 × g for 5 min at 4°C between them. Poly-P in the pooled water extracts was precipitated by adding NaCl to 0.1 M and 1 volume of ethanol, followed by incubation on ice for 1 h. After centrifugation at 16,000 × g for 10 min, the poly-P pellet was resuspended in 50 μl of water and analyzed by polyacrylamide electrophoresis. To this end, 15-μl samples were mixed with 1.5 μl of 10× loading dye (50% sucrose, 0.125% bromophenol blue, 450 mM Tris-borate [pH 8.3], 13.5 mM EDTA) and loaded onto 8% polyacrylamide gels containing 8 M urea in Tris-borate-EDTA (TBE) buffer. The gels were run in TBE buffer at 100 V and stained with a solution of 0.05% toluidine blue, 25% methanol, and 5% glycerol, followed by distaining with a solution of 25% methanol and 5% glycerol (39).

Resistance to stress conditions.

To study the effect of the inactivation of the ppk gene of L. casei BL23 on the response to stress, bacteria from overnight cultures in MRS of different L. casei strains were washed with 0.1% peptone two times and resuspended in 0.1% peptone. These suspensions were used to inoculate fresh MRS, MEI, and LP-MEI media without antibiotics and containing either 0.6 M NaCl, 0.5% bile (bovine; Sigma), or 5% ethanol or pH adjusted to 4 at an OD₅₉₅ of 0.05. Bacterial growth was monitored in 96-well microtiter plates (250 μl of medium per well) in a POLARstar plate reader (BMG Labtech) at 37°C. The growth response to high temperature was also tested in MRS, MEI, and LP-MEI at 42°C. To test resistance to oxidative stress, 10⁶ CFU were spread on MRS, MEI, and LP-MEI agar medium in 90-mm-diameter petri dishes. Subsequently, Whatman 3MM filter paper discs (5 mm in diameter) containing 5 μl of a solution of 50 mM plumbagin (5-hydroxy-3-methyl-1,4-naphthoquinone; Sigma) were layered on the plates. After 2 days of incubation at 37°C, the halos of growth inhibition were measured.

Statistical analysis.

Results are indicated as means ± the standard deviation (n = 3 to 17). The significances of the difference of the means in experiments carried out with wild-type L. casei BL23 and the ppk mutant were calculated using a Student t test with Prism 5.03 software (GraphPad Software, San Diego, CA).

RESULTS

Several species of Lactobacillus accumulate polyphosphate as volutin granules.

Thirty-four strains belonging to 20 distinct Lactobacillus species (Table 1) were screened for the presence of volutin granules by the Neisser staining method, which reveals these intracellular accumulations by the characteristic metachromasia of poly-P. Preliminary analyses showed that the use of MEI medium, which contains a higher concentration of phosphate, instead of MRS resulted in a higher proportion of Neisser-positive cells (data not shown). Therefore, MEI medium was chosen for further studies. Intracellular granules stained in black were detected in 18 of the tested strains belonging to 11 different species (Table 1). Figure 1A shows microscope images of stained representative strains comprising both poly-P⁻ and poly-P⁺ examples. The abundance of granules in the bacterial cell populations of poly-P⁺ strains varied, and ranged from granules detected in few cells to granules detected in practically all cells. Short cells usually carried one granule, whereas long cells normally had a variable number of granules homogeneously distributed along their whole length. For some strains (especially those from L. plantarum), the presence of granules was evident irrespective of the phosphate concentration in the medium (bacteria cultured in MEI or LP-MEI), although their frequency and intensity were lower under low-phosphate conditions (Fig. 1B). On the contrary, for the rest of the poly-P-producing strains the number of granules was dramatically lowered in LP-MEI, and in some cases the occurrence of Neisser-positive cells was a rare event. In some species (e.g., L. casei), these Neisser-positive cells were characterized for being longer and for the presence of more-dense dark granules (Fig. 1B), suggesting a heterogeneity in the physiological state within the cell population.

FIG 1.

Poly-P granules in several Lactobacillus strains. Bacterial smears were stained with Neisser staining and observed at ×100 magnification. (A) Examples of poly-P⁻ and poly-P⁺ strains are presented. (B) Poly-P granules in L. plantarum WCFS1 and L. casei BL23 strains grown under high- or low-phosphate conditions (MEI or LP-MEI, respectively).

Figure 2 shows electron microscopy micrographs of selected poly-P⁺ lactobacilli. Under the electron microscope, bacterial cells grown in MEI medium contained circular and oval holes in their cytoplasms, with diameters ranging from a few nanometers to >500 nm. These holes occupied a substantial portion of the intracellular volume, which in some cases led to a deformation in the bacterial cell shape (Fig. 2). Although poly-P inclusions are reported as electron-dense granules that fade under the electron beam, numerous examples show that these granules are usually chipped or torn out during the preparation of thin sections, leaving characteristic holes behind (22, 40, 41). Therefore, the existence of holes was attributed to the presence of poly-P granules. This was confirmed by the fact that, contrarily to cells grown in MEI medium, these holes were not observed in preparations from L. casei BL23 grown in LP-MEI (Fig. 3).

FIG 2.

Electron micrographs of selected Lactobacillus strains. Different strains were grown in MEI medium, fixed, and stained, and ultrathin preparations were observed using transmission electron microscopy. Under these conditions, the positions that were occupied by poly-P granules appeared as holes in the bacterial cytoplasm. The horizontal scale bars represent 0.5 μm.

FIG 3.

Electron micrographs of L. casei BL23 cells grown in MEI and LP-MEI media. The characteristic holes caused by the removal of the poly-P inclusions are not observed in bacteria grown in LP-MEI. The horizontal scale bars represent 5 μm.

Presence of ppk genes in Lactobacillus.

A BLAST search on sequenced genomes from Lactobacillus revealed that formation of poly-P granules in the tested strains for which their genome sequences were available fully correlated with the presence in the corresponding genomes of ppk genes which encode putative Ppks (Table 1). Conversely, inspection of poly-P⁻ strains with available genome sequences revealed that all of them lacked ppk. For the rest of poly-P⁺ strains, where no genome sequence was available, ppk genes were always found in all of the available genomes of different strains belonging to the same species, except for Lactobacillus intestinalis, for which no genome data are available. The only example of a poly-P⁻ strain belonging to a species where ppk was found was Lactobacillus acidophilus ATCC 4356 (Table 1). In addition, no poly-P granules were detected in the tested strains of Lactobacillus taiwanensis and Lactobacillus murinus, which do not have any sequenced representative strain. The products of ppk from lactobacilli were ∼700 amino acids in length. They carried the typical Ppk N-terminal and middle domains and two duplicated C-terminal domains structurally similar to the catalytic domains of phospholipase D, which are normally present in Ppks (30).

Figure 4 shows a schematic representation of the gene clusters carrying ppk genes found in Lactobacillales. Apart from Lactobacillus, in the order Lactobacillales ppk genes were also found in Aerococcus urinae, Oenococcus oeni, and some species of Leuconostoc and Streptococcus. In most cases the ppk gene was flanked by two genes (ppx1 and ppx2) encoding two putative exopolyphosphatases. Both gene products belonged to the Ppx-GppA family of phosphatases but shared a low percentage of identical residues (usually <20%). Gene ppx2 encoded a protein that was ∼200 amino acids shorter than the product of ppx1 that lacked the C-terminal part that corresponds to the metal-dependent HDc phosphohydrolase domain (31), which is present in this exopolyphosphatase class. Regarding Lactobacillales encoding just one Ppx, Aerococcus urinae and Lactobacillus delbrueckii harbored a ppx1 homolog, whereas Oenococcus oeni, Streptococcus parauberis, and Streptococcus uberis harbored a ppx2 homolog.

FIG 4.

Genomic context of ppk genes identified in sequenced genomes of bacterial species belonging to Lactobacillales. The presence of ppk, ppx1, and ppx2 homologues and the occurrence of similar flanking genes in different strains and species is depicted. An asterisk above the strain name indicates the presence of truncated genes.

Inactivation of ppk in L. casei BL23 affects poly-P synthesis.

In order to prove the functionality of the ppk gene, a derivative strain of the genetically amenable L. casei BL23 with a disruption in this gene was obtained. To this end, an internal fragment of ppk from BL23 was cloned into an integrative vector that was used for gene disruption by single crossover integration in the genome. Neisser staining confirmed the absence of dark granules in the cytoplasm of the ppk-disrupted strain (BL379), further confirming the poly-P nature of these inclusions (Fig. 5A). Moreover, isolation of intracellular poly-P and its subsequent electrophoretical analysis revealed that, while a substantial amount of poly-P was extracted from L. casei wild type, the ppk mutant showed a defect in poly-P synthesis (Fig. 5B). Poly-P granule formation was restored by plasmid expression of ppk in the mutant strain, although their levels were different compared to the wild type (Fig. 5A). This result showed that the absence of poly-P granules in the ppk mutant was related to a lack of Ppk activity. A higher proportion of Neisser-positive cells was detected in MRS culture compared to MEI in the complemented strain. Whether this effect is due to regulation of ppk expression in the wild-type strain or an effect of the growth medium on the activity of the P1 promoter in the complemented strain has not been elucidated. In addition, the likely polar effect of the ppk mutation on the expression of ppx1 and ppx2, which encode putative poly-P hydrolyzing enzymes, may also have an impact on poly-P amounts.

FIG 5.

Effect of a mutation in ppk on poly-P synthesis in L. casei BL23. (A) Neisser staining of wild-type L. casei BL23 and strain BL379, disrupted in ppk, grown in different media. Staining of the BL379 strain complemented by the expression of ppk from a plasmid (pT1ppk) is also shown; (B) separation of poly-P isolated from BL23 and BL379 strains grown on MEI medium on a denaturing polyacrylamide gel. Poly-P was stained with toluidine blue.

Effect of a mutation in ppk in the resistance to stress in L. casei BL23.

L. casei wild type and the ppk-defective strain were grown in different media (MRS, MEI, and LP-MEI) and under different conditions (high salt, the presence of bile, the presence of ethanol, low pH, or high temperature) in order to evaluate their ability to grow under stress conditions (Fig. 6). In general, the ppk mutation did not have a strong effect on growth under the tested conditions. Growth under reference conditions was not affected except for LP-MEI medium, where a significant decrease in the growth rate of the mutant strain compared to the wild type was observed (0.257 ± 0.007 and 0.230 ± 0.011 h⁻¹ for the wild-type and ppk mutant strains, respectively; P = 0.0208; Fig. 6D). The extent of the effect of the ppk mutation under the different stress conditions varied depending on the growth medium. The growth rates under a high NaCl concentration of the ppk strain were lower than those measured for the wild-type strain in all media, although the effect was statistically significant only when grown in LP-MEI (0.176 ± 0.011 and 0.147 ± 0.014 h⁻¹ for the wild-type and ppk mutant strains, respectively; P = 0.0474; Fig. 6A, C, and D). Also, compared to the wild type, a significantly lower final OD was reached by the ppk strain in MEI medium with a high salt concentration (0.655 ± 0.011 and 0.537 ± 0.016 for the wild-type and ppk mutant strains, respectively; P = 0.0004; Fig. 6C). The most pronounced effect was observed when cells were grown in MRS medium at pH 4, where the ppk mutant displayed a reduced growth rate (0.022 ± 0.001 and 0.016 ± 0.001 h⁻¹ for the wild-type and ppk strains, respectively; P < 0.0001; Fig. 6B) and final OD (0.807 ± 0.06 and 0.489 ± 0.008 for the wild-type and ppk mutant strains, respectively; P = 0.0001; Fig. 6B). This defect could be partially restored by plasmid expression of ppk in the mutant strain (Fig. 6B). Growth in the presence of ethanol or bile or at a high temperature was not affected by the ppk mutation under our experimental conditions (data not shown).

FIG 6.

Effect of a mutation in ppk from L. casei on growth under different stress conditions. Growth curves of L. casei strains (BL23, wild type; BL379, ppk mutant, and BL379[pT1ppk]) are shown. (A) Growth in MRS medium; (B) growth in MRS medium adjusted to pH 4; (C) growth in MEI medium with or without 0.6 M NaCl or with pH adjusted to 4; (D) growth in LP-MEI medium with or without 0.6 M NaCl or with pH adjusted to 4;. The results are the means of at least three experiments. Standard deviations were always lower than 15%.

The toxic effect of plumbagin, a superoxide radical generator, was tested by measuring the inhibition halos produced in solid medium (Fig. 7). The ppk-disrupted strain was slightly more sensitive to oxidative stress and expression of the ppk gene from a plasmid only restored the wild-type sensitivity in MRS medium.

FIG 7.

Effect of a mutation in ppk from L. casei BL23 on growth under oxidative stress conditions. The diameters of the inhibition halos of plumbagin on agar plates are shown. The results are from at least 17 measurements. ***, P < 0.0003; **, P < 0.003.

DISCUSSION

While searching for the presence of poly-P in several species of Lactobacillus we came to the unexpected finding that >50% of the tested species were able to form intracellular poly-P inclusions. References to these inclusions in lactobacilli are scarce and restricted to L. plantarum (20, 22) and certain strains isolated from mozzarella cheese whey (23). It has to be noted that, as reported here and as described by others (42, 43), the presence of poly-P granules may largely depend on growth conditions, which might account for the general lack of reports about poly-P in lactobacilli. It is thus not surprising that LAB were generally considered unable to accumulate this compound (44). This situation is not unprecedented and, as an example, the industrially relevant Corynebacterium glutamicum was considered a poly-P⁻ microorganism in contrast to the pathogenic Corynebacterium diphtheriae. However, poly-P inclusions were finally reported for the former species (45).

The capacity of lactobacilli to form poly-P inclusions was correlated with the presence of ppk genes in the strains where genome sequences are available, which is in agreement with the fact that polyphosphate kinase has been described as the main enzyme responsible for poly-P synthesis (3). Our observations strongly suggest the occurrence of ppk⁺ and Δppk species in Lactobacillus, although a careful evaluation of each strain must be performed. Thus, L. acidophilus ATCC 4356 did not show poly-P inclusions, whereas they were detected in L. acidophilus ATCC 9224 and L. acidophilus CNRZ216. This could be the result of the absence of ppk or mutations in ppk in strain ATCC 4356, as also found in the ppk clusters of some sequenced strains (e.g., Lactobacillus rhamnosus Lc705 or Streptococcus parauberis KCTC11537, Fig. 4). The reason why ppk (and poly-P synthesis in substantial amounts) is detected in some bacterial species while it is absent from others is intriguing. A ppk gene is not harbored by important model organisms such as Bacillus subtilis (3), whereas it is present in Bacillus thuringiensis or Bacillus cereus. It seems that the presence of ppk genes within a given genus is restricted to particular species. In the case of lactobacilli, ppk genes were indistinctly absent or present among lactobacilli isolated from similar niches (e.g., human intestine and fermented foods from animal or vegetal origin; Table 1).

We experimentally established the link between ppk and poly-P by gene disruption in L. casei strain BL23. However, the fact that a microorganism harbors a ppk gene does not necessarily imply that it possesses the capacity to form poly-P intracellular inclusions. For example, E. coli synthesizes poly-P via Ppk although it generally does not form intracellular granules unless specific mutations (e.g., phoU) increase the phosphate uptake rate (46), showing that phosphate transport can be limiting for poly-P formation. In the present study, poly-P granules were always observed in ppk⁺ species, although their abundance was in general dependent on the growth medium and on the presence of large amounts of phosphate. One exception to this was L. plantarum, which formed poly-P inclusions under low-phosphate conditions, suggesting that this species possesses a superior capacity for phosphate uptake and accumulation. In this microorganism, a very efficient uptake system for Mn(II) can accumulate intracellular Mn(II) over 30 mM, and this transport is connected to phosphate transport and probably to poly-P accumulation, as shown by the reduced Mn(II) accumulation in phosphate-limited cells (21). Mn(II) is stored in a high-molecular-weight complex consisting of poly-P and proteins and serves as an alternative mechanism for oxygen tolerance by scavenging reactive oxygen species (22, 47).

Lactobacilli exhibited a genetic structure in which ppk was clustered with two exopolyphosphatase genes (ppx1 and ppx2). The only exception to this rule was L. delbrueckii, which appeared to carry deletions/insertions at the 3′-end of ppk (Fig. 4). In E. coli and other bacteria, ppk forms an operon with the ppx gene, which denotes coordination in the expression of both poly-P-synthesizing and -hydrolyzing activities (15). The question of how can the expression of these two enzymes with opposite functions lead to the accumulation of poly-P can be explained by the fact that the hydrolytic activity of Ppx is strongly inhibited by (p)ppGpp, resulting in blockage of poly-P turnover (2, 5). The existence of two exopolyphosphatase genes in an operon with ppk has never been described, but the occurrence of several polyphosphatases in the same microorganism is not exceptional. In E. coli poly-P can also be hydrolyzed by the product of gppA, which encodes a guanosine pentaphosphate (pppGpp) phosphohydrolase that is involved in the stringent response (16). Furthermore, two exopolyphosphatases, lacking the C-terminal HDc domain, are encoded in the genome of Actinomycetales (48, 49), although they are distantly located and not clustered with ppk. Compared to Ppx1, and similar to the Ppx proteins from Actinomycetales, Ppx2 from lactobacilli lacks the C-terminal HDc domain and only possesses the N-terminal nucleotide-binding domain of the sugar kinase/actin/Hsp70 superfamily present in members of the Ppx-GppA polyphosphatases (32). The specific function of the products of ppx1 and ppx2 in lactobacilli is unknown. The two exopolyphosphatases from the members of the Actinomycetales Corynebacterium glutamicum (Gg0488 and Cg1115 proteins [49]) and Mycobacterium tuberculosis (Rv0496 and Rv1026 proteins [48]) have been genetically and biochemically characterized. It has been shown that Cg1115 and Rv0496 can hydrolyze short-chain poly-P, but the function of Gg0488 and Rv1026 has not yet been established. Unlike GppA from E. coli, Rv0496 and Rv1026 from M. tuberculosis were not able to hydrolyze pppGpp to ppGpp, but they exhibited exopolyphosphatase and ATPase activities, respectively, which were inhibited by (p)ppGpp (48). Studies with Ppx from E. coli, have determined that the N-terminal domain is mainly responsible for the catalytic activity, whereas the C-terminal domain is involved in poly-P recognition and processivity of the enzyme (50). The existence in lactobacilli of two Ppx enzymes which belong to different classes (that from the E. coli Ppx and GppA prototypes and that from Actinomycetales) represents a new finding and suggests that both enzymes might have nonoverlapping substrates. The precise role of Ppx1 and Ppx2 and whether they are implicated in the metabolism of (p)ppGpp alarmones deserves further studies.

Many aspects of the response to stress and adaptation to particular niches have been deeply studied in lactobacilli, but, with the exception of the role of Mn(II) in oxygen tolerance (22, 47), the occurrence or participation of poly-P in the underlying mechanisms is totally unexplored. Similar to the case in other bacteria (3, 12), we showed that mutation of ppk in L. casei resulted in a decreased growth under a variety of stresses, with the most pronounced effect observed for acid pH. It has been hypothesized that poly-P might provide the cells with an extra buffering capacity to counteract cytoplasmic pH changes (2, 51). In this sense, Aprea et al. described the occurrence of poly-P granules in L. casei and L. fermentum strains isolated from mozzarella cheese whey and indicated that poly-P synthesis in these bacteria may account for a defense mechanism to counteract the low pH found in that environment (23). The survival under oxidative stress was also diminished in the L. casei ppk mutant. A mechanism based on Mn(II) for reactive oxygen species scavenging, such as the one present in L. plantarum (22, 47), has never been described in other lactobacilli. In E. coli, poly-P and (p)ppGpp are signaling factors regulating the expression of the alternative sigma factor S (RpoS), essential for the expression of genes involved in stress and stationary-phase response (52). Apart from RpoN (σ⁵⁴), involved in the expression of a sugar transport operon (53), no other additional sigma factor besides the housekeeping RpoD (σ⁷⁰) is present in L. casei. Nevertheless, (p)ppGpp has been involved in stress response in some LAB (54), and it plays a role in the modulation of poly-P levels by inhibiting the activity of Ppx (5). Based on this and on the present results, it is reasonable to think that a link between (p)ppGpp, poly-P, and stress exists in L. casei and probably in the rest of lactobacilli that carry ppk genes, but the particular mechanisms have yet to be evaluated.

It appears that poly-P in bacteria is not restricted to the “classical” poly-P-accumulating microorganisms. The unexpected discovery that poly-P accumulation is widespread in lactobacilli might have important implications for the technological processes where these bacteria are used. In addition to the likely role of poly-P synthesis in the stress adaptation of lactobacilli, this compound can also play a role in the functionality of some strains as probiotics. Poly-P synthesized by L. brevis SBC8803 interacts with integrin β₁ activating the p38 mitogen-activated protein kinase pathway, which results in the induction of HSP27 in intestinal epithelial cells and in the protection of the intestinal barrier function (24). Poly-P is found in culture supernatants of L. brevis, and it could be interesting to test whether other strains of probiotic lactobacilli possess the capacity to release poly-P with similar intestinal homeostatic effects. Finally, poly-P synthesis has been linked to the resistance to heavy metals in bacteria, acting as a chelating agent that enhances the metal retention capacity and reduces its toxicity (2, 55). A growing interest exists in the use of probiotic microorganisms as metal detoxification agents (56, 57). In this sense, poly-P could be involved in the ability of some strains to bind metals and may thus be a factor with relevance for detoxification in humans.