Introduction

Table olive fermentation is the oldest practice by our ancestors to preserve vegetables and to also produce different flavours and textures. Additionally, fermented table olives remain an important economy for many production countries and a component of the Mediterranean diet (and recommended as part of the Healthy Eating Pyramid published in 2010, https://dietamediterranea.com/). The high nutritional value of fermented table olives (e.g., their content of carbohydrates, fiber, minerals, vitamins, fatty acids, and amino acids) and their role as potential source of probiotic lactobacilli of vegetable origin1,2,3,4,5 make them very attractive from an economic and social point of view. Lactobacillus genus is the most representative and heterogeneous member of lactic acid bacteria (LAB) group currently consisting of 237 species (as of December 2018 in www.bacterio.net) since they harbour in their genome a plethora of genes involved with a wide array of functional properties6,7. Lactobacillus spp. are principal bacteria in olive fermentation processes, possessing many biochemical and physiological traits to ferment several carbohydrates and tolerate stress8. These phenotypes are important as the brine environment represent harsh conditions for bacterial growth with low nutrient availability, saltiness, low pH and the presence of antimicrobials (e.g., phenolic compunds and oleuropein); thus, highly robust L. plantarum and L. pentosus are frequently isolated from the end of olive fermentation1,8,9. Furthermore, Perpetuini, et al.10 demonstrated by transposon mutagenesis that the high capacity of L. plantarum and L. pentosus to survive in the hostile, brine environments was due to critical genes encoding proteins involved in carbohydrate metabolism, membrane structure and function, and gene-expression regulation. They further suggested that the obaD gene, which encodes a putative membrane protein strictly specific to L. pentosus/L. plantarum species, may be one of the key elements involved in their efficient adaptation to several conditions in many fermented food processes and natural ecosystems10.

Aloreña green table olive fermentation is a spontaneous process relying on L. pentosus strains and yeasts1,9. Resistance, persistence and predominance of Lactobacillus spp. in green table olive fermentation is due to their genetic variation and plasticity related to their chromosome and plasmids. In fact, L. pentosus species isolated from olive fermentation harbours the largest genome recognized to date and several plasmids (range: n = 5 to 7)11,12,13. However, L. plantarum contains the largest plasmids among the genus Lactobacillus14,15 such as L. plantarum 16, which harbors 10 plasmids ranging 6.46–74.08 kb16. Most of the Lactobacillus plasmids are cryptic; however, they possess important properties such as antibiotic resistance, exopolysaccharide production, antimicrobial activity, bacteriocin synthesis, bacteriophage resistance, carbohydrate metabolism, host colonization and probiotic activity17,18,19,20,21,22. On the other hand, megaplasmids were also detected in Lactobacillus sp., up to 490 kb23. In this study, we analyzed in silico five plasmids harboured by L. pentosus MP-10 isolated from naturally fermented Aloreña green table olives2,9,12. Moreover, we aimed to better understand the underlying functional and probiotic properties of these plasmids using curing plasmid experiments; in particular, we examined their physiological traits in metal tolerance and biosorption, antimicrobial activity and adaptation to gastro-intestinal conditions to determine possible probiotic applications of this bacterium.

Results

General features of L. pentosus MP-10 plasmids

We have already reported the sequencing of L. pentosus MP-10 genome12, which consisted of a single circular chromosome of 3,698 kbp and five plasmids ranging 29–46 kbp (accession numbers FLYG01000001 to FLYG01000006). Sequence annotation was done using the Prokka annotation pipeline, version 1.1124 as previously reported by Abriouel, et al.12. The general features of the circular five plasmids2 are reported in Table 1. The average GC content of L. pentosus MP-10 plasmids ranged 39.52–42.50%, slightly lower than the host chromosome (with GC value of 46.32%). Furthermore, the GC contents of L. pentosus MP-10 plasmids were among the highest of known L. pentosus plasmids. All open reading frames in L. pentosus MP-10 plasmids are greater than 34 amino acids (Tables 2–6). Blast search for homology revealed lower identity with other plasmids in the database; however depending on coverage percentage, some regions harboring several genes in L. pentosus MP-10 plasmids were highly related with plasmids of Lactobacillus species isolated from foods like fermented olives, kimchi, koumiss, tofu or raw sausages, and also from human saliva (Table 1).

Table 1 General features of circular plasmids from L. pentosus MP-10.
Full size table
Table 2 Genes determined in pLPE-1 plasmid of Lactobacillus pentosus MP-10 isolated from naturally fermented Aloreña table olives.
Full size table
Tablze 3 Genes determined in pLPE-2 plasmid of Lactobacillus pentosus MP-10 isolated from naturally fermented Aloreña table olives.
Full size table
Table 4 Genes determined in pLPE-3 plasmid of Lactobacillus pentosus MP-10 isolated from naturally fermented Aloreña table olives.
Full size table
Table 5 Genes determined in pLPE-4 plasmid of Lactobacillus pentosus MP-10 isolated from naturally fermented Aloreña table olives.
Full size table
Table 6 Genes determined in pLPE-5 plasmid of Lactobacillus pentosus MP-10 isolated from naturally fermented Aloreña table olives.
Full size table

Figure 1 shows the frequency of KEGG functional annotations by BlastKOALA (KEGG tool; last updated March 4, 2016), which assigned plasmid genes to KEGG annotations corresponding to environmental information processing (pLPE-3, pLPE-4 and pLPE-5), genetic information processing (pLPE-2, pLPE-3, pLPE-4 and pLPE-5), carbohydrate metabolism (pLPE-3 and pLPE-5), amino acid metabolism (pLPE-3 and pLPE-5), cellular processes (pLPE-1, pLPE-2, pLPE-4 and pLPE-5), nucleotide metabolism (pLPE-2), metabolism of cofactors and vitamins (pLPE-3), and enzyme families (pLPE-3).

Figure 1
figure 1

BlastKOALA results of functional gene-categories predicted in Lactobacillus pentosus MP-10 plasmids and their frequencies.

Full size image

In silico analysis of plasmid properties in L. pentosus MP-10

Analysis of the annotated CDs of each L. pentosus MP-10 plasmid revealed the presence of five genes involved in mobilization (mobA gene) distributed in all plasmids except the pLPE-2 plasmid (Tables 2–6). These genes are likely required for plasmid relaxation and mobilization by conjugative plasmids. Also, conjugation-related genes were found, e.g., traG in pLPE-4 (traG_1 and traG_2) and pLPE-5 (traG_3) plasmids (Tables 5 and 6). A gene encoding for a bacteriophage peptidoglycan hydrolase that may have been involved in growth was found in pLPE-4 (XX999_00013 and XX999_00049) and pLPE-5 (XX999_03566) plasmids (Tables 5 and 6).

The presence of mobile genetic elements in L. pentosus MP-10 plasmids (pLPE-2, pLPE-3, pLPE-4 and pLPE-5) was already reported by Abriouel et al.2 such as four putative transposon Tn552 DNA-invertase bin3 (four different genes of the same family), transposase DDE domain proteins (4 genes in pLPE 2 and pLPE5 plasmids), transposases of the mutator family (3 genes in pLPE2, pLPE3 and pLPE5 plasmids) and transposases (2 genes in pLPE-2 and pLPE-3 plasmids). Concerning integrases, one phage integrase family protein (pLPE-1 plasmid) and 9 integrase core domain proteins were detected in pLPE-2, pLPE-3 and pLPE-5 plasmids (Tables 3, 4 and 6). A gene pinR coding for DNA invertase from prophage was detected in pLPE-5 plasmid (Table 5).

Chloride- (clcA_2) and sodium- (nhaS3_4) transport genes harboured by pLPE-2 plasmid (Table 3) indicated that this plasmid was involved in salt-tolerance in brine solutions (plasmid curing experiments). Furthermore, a copy of the same genes clcA_1, nhaS3_1, nhaS3_2 and nhaS3_3 were also found in L. pentosus MP-10 chromosome with the aim to potentiate chloride and sodium tolerance in brines.

Genes related to carbohydrate metabolism were found on plasmids (besides on the chromosome) such as L-Lactate dehydrogenase in pLPE-5 plasmid (ldh_7 and ldh_8 genes) (Table 6), genes involved in glucose uptake and metabolism such as glcU_1 and gdhIV_1 genes in pLPE-3 plasmid (Table 4), and a gene involved in xylan catabolic process (axeA1_3) in pLPE-5 (Table 5). However, another gene involved in xylan catabolic process (XX999_00089) was only detected in pLPE-3 plasmid, but not on the chromosome (Table 4).

Toxins reported in L. pentosus MP-10 plasmids include mazF-toxin encoding gene (XX999_03521) detected in pLPE-1 plasmid, genes coding for Zeta toxins in pLPE-3 (XX999_00053) and pLPE-4 (XX999_00024) plasmids, and also for antitoxins such as RelB antitoxin (XX999_00026) in pLPE-4 plasmid and the bifunctional antitoxin/transcriptional repressor RelB in pLPE-5 plasmid (XX999_03554) (Tables 2, 4–6). MazF toxin is a desirable property in probiotic bacteria, and is only detected in plasmid DNA of L. pentosus MP-10, not in the chromosome. However, L. pentosus MP-10 has to protect itself from the MazF toxin without any MazE antitoxin. On the other hand, RelB antitoxins were found both on plasmids and on the chromosome; however, no RelB toxins were detected. Zeta toxins were detected both on the chromosome (one gene) and also on plasmid DNA (two genes); however, no antitoxin was detected.

Other coding genes for several functions, such as a serine-rich adhesin for platelets precursor (sraP gene), were detected in pLPE-2 plasmid but not on the chromosome (Table 3); genes coding for vitamin biosynthesis such as panE_1 and panE_2 genes coding for 2-dehydropantoate 2-reductase (biosynthesis of vitamin B5), a gene XX999_00068 coding for prephenate dehydratase (biosynthesis of phenylalanine, tyrosine and tryptophan), were detected on the pLPE-3 plasmid (Table 4) and also on the chromosome.

Regarding their responses to stress, in-silico analysis of plasmid sequences revealed the presence of yhdN_1 gene coding for a general stress protein 69 (in pLPE-3, Table 4) and several genes coding for metal tolerances, such as cadmium [cadmium resistance transporter (XX999_03594) and a putative positive regulator of cadmium resistance (cadC)] and two operons of arsenic resistance (in pLPE-5, Table 6). One ars operon consists of arsR_3 (arsenical resistance operon repressor ArsR) and arsB [arsenical pump membrane protein (ArsB)], but lacks arsC gene (arsenate reductase ArsC); the other ars operon contains arsA [arsenical pump-driving ATPase (ArsA)] and arsD gene [arsenical resistance operon trans-acting represor (ArsD)] in pLPE-5 (Table 6). The synteny of arsenic-resistance genes was examined by comparing the annotated sequences of pLPE-5 and pWCFS103 plasmids (aligned by MAUVE algorithm) from L. pentosus MP-10 and L. plantarum WCFS1, respectively. Comparison revealed that the synteny of genes was similar (Fig. 2), being arsenic operons in pLPE-5 of L. pentosus MP-10 composing of two copies each gene: arsB [coding for trivalent As(III) efflux permease ArsB], arsA [coding for trivalent As(III)-stimulated ATPase ArsA], arsD [coding for trivalent As(III) metallochaperone ArsD] and arsR_3 gene [a trivalent As(III)-responsive repressor (ArsR)]. On the other hand, arsC gene (arsC2 coding for reductase ArsC), as a part of ars operon with arsB and arsR genes, was found in L. pentosus MP-10 chromosome, as well as two arsR gene copies (arsR_1 and arsR_2).

Figure 2
figure 2

MAUVE visualization of the alignment of the pLPE-5 plasmid from L. pentosus MP-10 with the pWCFS103 plasmid from L. plantarum WCFS1. Arsenic- and cadmium-resistance genes are indicated.

Full size image

In vitro detection of functional properties in L. pentosus MP-10 plasmids

Effect of plasmid curing on growth of L. pentosus MP-10

The MIC of acridine orange (AO) was of 0.15 mg/ml; as such, we used 0.1 mg/ml as the sub-MIC for plasmid curing in this strain. After confirming L. pentosus MP-10 being cured of plasmids (data not shown), we compared the growth kinetics of uncured and cured L. pentosus MP-10C. The presence of plasmids did not affect the growth in MRS broth at 37 °C in any experimental conditions: presence/absence of 6.5% NaCl, different pH ranges (1.5 to 7.0), nor the presence of bile salts (1.8 or 3.6%) -no differences in 600 nm absorbances were detected over 24 h of incubation- (Figs S1–A,B, S2). In a similar manner, pH monitoring during their incubation also did not exhibit any significant differences between cured and uncured strains in regards to their acidification capacity (Fig. S1–C). Furthermore, no differences in the growth were detected between the cured and uncured L. pentosus MP-10 strains in the presence of xylan as the only carbohydrate source (Fig. S1–D). However, at high salt concentration of 8% usually found in brine, significant differences were detected between the cured and uncured L. pentosus MP-10 strains, with the uncured strain being the most tolerant (Fig. S1–E).

Table 7 shows that curing had no significant effect on the growth of uncured and cured L. pentosus MP-10 in the presence of phenolic compounds naturally present in the brines; both the cured and uncured strains tolerated more than 200 mg/ml of olive-leaf extract.

Table 7 Antibiotic and biodice susceptibility, and probiotic properties of cured and uncured L. pentosus MP-10 isolated from Aloreña Green table olives.
Full size table

Effect of plasmid curing on metal tolerance

Plasmid annotations predicted gene clusters involved in arsenate- and/or arsenite-, and cadmium resistance. First, we precisely determined metal concentrations that inhibit the visible growth of the wildtype L. pentosus MP-10; results showed that this strain tolerated high concentrations of metals depending on the metal with 1 < MIC < 4096 μg/ml, and tolerances were observed to be in order Fe > [Al/Cu/Co] > Zn > Cd > Hg (Table 8). When we compared the uncured and the cured L. pentosus MP-10, we found that mercury and cadmium exibited different MICs among strains by 2–8 fold increase (Table 8) in those uncured; as such, plasmids have a key role in mercury and cadmium tolerances.

Table 8 Tolerance of cured and uncured L. pentosus MP-10 isolated from Aloreña Green table olives to heavy metals.
Full size table

The removal of different metals was shown in Table 8, which demonstrated that L. pentosus MP-10 was able to remove different metals, thus exhibiting high removal capacity of mercury (81.74% ± 2.04), cadmium (67.10% ± 0.88) and aluminium (57.14% ± 0.99). However, the cured L. pentosus MP-10C demonstrated statistically significant reduced performance. Metal removal differences between the uncured and the cured L. pentosus MP-10 highlight the role of plasmids to remove iron, cadmium, aluminium, cobalt, copper, zinc and mercury (Table 8).

To understand how L. pentosus MP-10 interact with selected metals, SEM analysis was performed and showed the biosorption potential of the uncured L. pentosus MP-10 (Fig. 3). The micrographs and EDX spectra obtained before and after the biosorption process showed clearly that the cell morphology of the uncured L. pentosus MP-10 changed and exhibited the presence of bright particles on the surface of the bacteria exposed to some metals. Regarding cadmium, mercury and zinc, we couldn´t detect these metals by EDX analysis. Furthermore, in the presence of either aluminium, cobalt, copper, mercury or zinc, higher potential for biofilm formation was observed. These results, confirmed by EDX analyses, support that these metals remained adsorbed entirely on the cell surface.

Figure 3
figure 3

SEM (A,C,E,G,I,K,M,O) and EDX (B,D,F,H,J,L,N,P) analysis of uncured L. pentosus MP-10 without metal (A,B) and with Al (C,D), Cd (E,F), Co (G,H), Cu (I,J), Fe (K,L), Hg (M,N) and Zn (O,P).

Full size image

Effect of plasmid curing on antimicrobial resistance and probiotic features

We determined the MIC of different antibiotics and biocides between uncured and cured strains, and the results did not show any significant differences in response between both strains except for clindamycin, which exibited 20 fold increase in the MIC in the uncured L. pentosus MP-10. Thus, plasmids have no role in the suceptibility to the antibiotics and biocides tested, except clindamycin (Table 7).

Regarding the probiotic features, the uncured and the cured L. pentosus MP-10 had performed similarly in auto-aggregation and co-aggregation with all pathogens tested (Table 7), which suggest that plasmids had neither any role in auto-aggregation nor co-aggregation processes. Regarding acid and bile tolerance, no differences were detected between the uncured and the cured L. pentosus MP-10 (Table 7).

Adhesion to mucin was measured in both the uncured and the cured L. pentosus MP-10, and the results showed a statistically significant increase in adhesion capacity to mucin in the uncured L. pentosus MP-10 (Table 7).

Discussion

Olive brine represents a stressful environment for the growth and survival of many bacteria due to the harsh conditions (i.e., high salt concentration, presence of phenolic compounds and low-nutrient availability), which provide selective pressures for the maintenance of LAB. As such, L. plantarum and L. pentosus have the genetic tools to survive and grow in the hostile olive-brine conditions10, and these genetic traits are widely distributed on both the chromosome and the plasmids, with several genes having multiple copies to enhance their adaptability and fitness in different ecological niches.

In this study, L. pentosus MP-10, isolated from Aloreña green table olives, harboured five plasmids with an average GC content (39.52–42.50%) slightly lower than the host chromosome (46.32%), this difference was less than 10% as reported by Nishida25 for the majority of plasmids. pLPE-5 had remarkably the lowest average GC content (39.52%) than the other four plasmids (pLPE-1, pLPE-2, pLPE-3 and pLPE-4), suggesting it is possibly a recent acquisition from another bacterium. In-silico analysis of plasmid sequences revealed the presence of genes involved in mobilization (mobA) and conjugation (traG) distributed in several plasmids, which suggest their role in gene mobilization and secretion using a type-IV secretion mechanism26. Furthermore, mobile genetic elements (e.g., transposon, transposase, integrase and invertase) were also found in several plasmids2 suggesting a frequent genetic diversification among the L. pentosus MP-10. Furthermore, bacteriophage peptidoglycan hydrolases were found in pLPE-4 and pLPE-5 plasmids; these lysozyme-like proteins may play a key role in L. pentosus MP-10 growth, its cell-wall structure, and immunomodulatory properties as reported by Rolain, et al.27.

Metabolic profile within L. pentosus MP-10 plasmids include carbohydrate enzymes such as L-lactate dehydrogenase, glucose uptake and metabolism and xylan catabolic enzymes. L-lactate dehydrogenase was codified by two genes (ldh_7 and ldh_8) located on pLPE-5 plasmid; however, six L-lactate dehydrogenase (ldh_1, ldh_2, ldh_3, ldh_4, ldh_5 and ldh_6) and four D-lactate dehydrogenase (XX999_00315, XX999_00955, XX999_02047 and XX999_02719) coding genes were also present on the chromosome. Both enantiomers (L-lactate and D-lactate) are produced by L. pentosus MP-10 being D-and L-lactate dehydrogenases involved in the reversible metabolism of D- and L-lactate, respectively. This finding is of great interest suggesting that the use of L. pentosus MP-10 as a probiotic may help human to metabolise D-lactate obtained from exogenous sources (e.g., diet and the carbohydrate-fermenting bacteria normally present in the gastrointestinal tract) since mammalian cells lack sufficient D-lactate dehydrogenase required to utilise D-lactic acid—leading to chronic fatigue syndrome and D-lactic acidosis or D-lactate encephalopathy associated with short bowel syndrome28,29,30. Further, L-lactate dehydrogenase genes present on the plasmids may enhance their metabolic activity during the fermentation process to produce more L-lactate and energy. However, the presence of L-lactate dehydrogenase (ldh_7 and ldh_8) coding genes on pLPE-5 plasmid did not enhance the acidification capacity, as results were similar after 8 and 24 h incubation in both cured and uncured L. pentosus MP-10, suggesting that these genes either have a minor role in lactate production or they are regulated. Further experiments, based on differential relative expression of ldh gene in both the cured and uncured L. pentosus MP-10 strains, revealed low expression level in the cured strain (Fig. S3), thus the low activity of lactate dehydrogenase gene in the cured strain is enough to give rise to a substantial lactate accumulation in the fermentation broth in a manner similar as the uncured strain. Regarding glucose uptake and metabolism, glcU_and gdhIV genes were over-expressed in the uncured L. pentosus MP-10 indicating the role of plasmid in this process (Fig. S3).

Among defense mechanisms found on plasmids, gene encoding the mazF toxin (pLPE-1), Zeta toxins (pLPE-3 and pLPE-4), and also antitoxins such as RelB antitoxin (pLPE-4) and the bifunctional antitoxin/transcriptional repressor RelB (pLPE-5) were detected in L. pentosus MP-10 plasmids. RelBE and MazEF are known as sequence-specific endo-ribonucleases that inhibit the global translations of cellular mRNAs31. MazF toxin is a desirable trait for probiotic bacteria, as its antimicrobial property inhibits several pathogens in foods and the gastrointestinal tract32. However, L. pentosus MP-10 must protect itself from the mazF toxin, as no MazE antitoxin was detected. Either their protection relies on other mechanisms because mazF is functional being only expressed in the uncured strain (Fig. S3). On the other hand, genes for RelB antitoxins were found both on plasmids and on the chromosome; however, no RelB-toxin genes were detected. So this antitoxin may contribute a greater defense against other bacteria possessing RelB toxins, possibly increasing its competitiveness and survival in several ecological niches including gastrointestinal tract. This feature was mainly linked to plasmid being relB antitoxin gene over-exopressed in the uncured strain (Fig. S3). Zeta toxins, which are kinases that kill bacteria through global inhibition of peptidoglycan synthesis33, are detected both on the chromosome and also on plasmid DNA of L. pentosus MP-10, however no antitoxin was detected. Overall, L. pentosus MP-10 harbored in their plasmids incomplete toxin-antitoxin systems unlike what occur naturally in bacterial genomes, since several toxins or antitoxins were detected without self protection.

Data obtained by in-silico analysis suggests that plasmid-borne genes mediate the persistence of lactobacilli under olive fermentation conditions and enhance their probiotic properties; however, this hypothesis requires further studies for confirmation. As such, plasmid curing experiments carried out with L. pentosus MP-10 showed several differences between the uncured and the cured strains regarding metal tolerances, removal and mucin adhesion. However, plasmid curing did not affect their tolerance to gastro-intestinal conditions (e.g., acids and bile salts); neither their ability to grow under determined conditions (i.e., different pH intervals, bile salts or sodium chloride of 6.5%) nor their colony morphology were changed after plasmid curing (data not shown). However, at high concentration of chloride of 8% (commonly added to brines), L. pentosus pLPE-2 plasmid plays a key role in salt tolerance. In this sense, the results suggest that the plasmids did not govern the fermentation of carbohydrates under these conditions, however different results were obtained by Adeyemo and Onilude34 which showed that plasmid curing had a significant negative effect on growth, physiological characteristics and colony morphology of L. plantarum isolated from fermented cereals. In this study, plasmids in L. pentosus MP-10 may confer a selective advantage, providing other physiological properties in certain environments such as gut and brines and thus allowing metal tolerance and removal, salt tolerance and adherence to mucin and thus their persistence in competitive ecological niches. Mucin adhesion declined in the cured L. pentosus MP-10 since a serine-rich adhesin for platelets precursor gene (sraP, detected in pLPE-2 plasmid) may be involved in mucin adhesion mechanisms similarly as reported by Hevia, et al.34 for an extracellular serine/threonine-rich protein as a novel aggregation-promoting factor with affinity to mucin in Lactobacillus plantarum NCIMB 8826. The role of L. pentosus MP-10 plasmids in mucin adhesion was confirmed by relative expression gene analysis as reported by Pérez Montoro et al.35, since recA and pgm genes considered as potential biomarkers of mucin adhesion were over-expressed in the uncured strain (Fig. S3). However, auto-aggregation and co-aggregation with some pathogens were not changed after plasmid curing of L. pentosus MP-10.

With respect to metals, which are considered non-biodegradable and non-thermodegradable and are of high concern in both developing and developed countries because of their impact on the environment and health (water and food), the wild strain L. pentosus MP-10 showed greater tolerance to their increased concentrations (MICs higher than 1 mg/ml, except for cadmium and mercury) of iron, cobalt, copper, aluminium and zinc. This suggests that high contamination of metals in the environment from natural and anthropogenic sources36 may be tolerable by the bacteria. The self-protective mechanisms displayed by L. pentosus MP-10 as a response to metals is promoted by their architecture (cell wall and membrane) and also by their resistance determinants located on the chromosome and the plasmids. Moreover, several chromosomally encoded cation transporters (e.g., encoded by czcD gene) have a predicted substrate range, including cadmium, cobalt and zinc; although the increased resistance towards different metals are displayed by plasmids (especially the pLPE-5 plasmid). Similar results were obtained by van Kranenburg et al.22, which reported that the plasmid-borne (pWCFS103) cadC gene coding for a transcription regulator of the cadmium operon was responsible of the increased resistance to cadmium in L. plantarum WCFS1. Furthermore, the synteny of ars genes in both L. pentosus MP-10 and L. plantarum WCFS122 was similar suggesting their evolutionary relatedness. Arsenic and cadmium are among the most toxic elements widely ocurring in the environment, often a threat to food and water supply. Arsenic is known as a group A “known” carcinogen according to the United States Environmental Protection Agency (USEPA) and contributes to a range of other illnesses such as cardiovascular and peripheral vascular diseases, neurological disorders, diabetes mellitus and chronic kidney disease37,38,39. Detoxification of this metal was earlier established by bacteria. Thus, tolerance of L. pentosus MP-10 is necessary to prevent damage to their cells.

The ability of L. pentosus MP-10 to bind different metals was demonstrated by SEM and EDX analysis. This is of great importance with regards to their application as an adjunct to improve food safety and quality by bioquenching metals and probiotically reduce metal toxicity among human intestinal microbiota and thus protecting the host40. Also, we demonstrated that L. pentosus MP-10 contributed to metal removal, especially mercury and cadmium (81 and 67%, respectively).

Metal- and antibiotic-resistance genes often co-exist on the same plasmid, however in this case, we did not find any genes coding for clindamycin resistance on plasmids, which was the only antibiotic with different susceptibility after plasmid curing. Thus, clindamycin resistance in L. pentosus MP-10 may rely on other plasmid-associated genes that we could not deciphered yet.

Conclusions

In-silico analysis of L. pentosus MP-10 plasmids suggests that plasmid-borne genes mediate the persistence of lactobacilli under olive-fermentation conditions and enhance their probiotic properties with genes encoding for carbohydrate metabolism, defense mechanisms, metal tolerance and mobilization increasing subsequently its competitiveness and survival in several ecological niches. Plasmid curing demonstrated the role of plasmids in the increased metal tolerance, and bioremoval of several metals (e.g., iron, aluminium, cobalt, copper, zinc, cadmium and mercury). This probiotic property by L. pentosus MP-10 should be exploited to detoxify metals in intestines; basically they could bioquench the metals in the gut thus reducing their toxic exposure to humans and animals, in the food matix and in environmental bioremediation.

Materials and Methods

Bacteria and growth conditions

Lactobacillus pentosus MP-10 isolated from naturally-fermented Aloreña green table olives1 were cultured in de Man Rogosa and Sharpe (MRS) broth (Fluka, Madrid, Spain) at 37 °C for 24 h. Pathogenic bacteria used in this study included Listeria innocua CECT 910, Staphylococcus aureus CECT 4468, Escherichia coli CCUG 47553, and Salmonella Enteritidis UJ3449, which were cultured in Tryptone Soya Broth (TSB; Fluka, Madrid, Spain) at 37 °C for 24 h. Cultures were maintained in 20% glycerol at −20 °C and −80 °C for short- and long-term storage, respectively.

In silico analysis of L. pentosus MP-10 plasmid sequences

The genome sequence of L. pentosus MP-10 consisted of a single circular chromosome of 3,698,214 bp, with an estimated mol% G + C content of 46.32% and 5 plasmids ranging 29–46 kb (accession numbers FLYG01000001 to FLYG01000006) were annotated using the Prokka annotation pipeline, version 1.11 (Seemann, 2014) as previously reported by Abriouel, et al.12. The predicted CDSs of plasmids2,12 were annotated by using BLAST (Basic Local Alignment Search Tool) and the associated GO (Gene Ontology) terms were obtained by using Swiss-Prot database.

The general metabolic pathways of L. pentosus MP-10 plasmids were reconstructed using BlastKOALA (last updated March 4, 2016) as part of the KEGG (Kyoto Encyclopedia of Genes and Genome) tool in the pathway database (http://www.genome.jp/kegg/pathway.html) for annotating genomes; here, we used the annotated genes predicted in each L. pentosus MP-10 plasmid as the input query.

To evaluate the alignment and the synteny of genes between the L. pentosus MP-10 and L. plantarum WCFS1 plasmid data sets, comparison was done by using Mauve algorithm in Lasergene’s MegAlign Pro software (Lasergene 14).

In vitro analysis of L. pentosus MP-10 plasmid properties

Plasmid curing

First, we determined the minimum inhibitory concentrations (MIC) of acridine orange (AO) to L. pentosus MP-10 using the broth micro-dilution method. Overnight cultures, grown in MRS broth at 37 °C for 24 h, were diluted 1/10 (v/v) in fresh MRS broth and 20 µl were added to each well of 96-well microtiter plates. 180 µl of MRS broth supplemented with AO at different concentrations (12.5–400 μg/ml) were then added to the wells and incubated at 37 °C under aerobic conditions for 24 h. Bacterial growth was evaluated by the presence of turbidity. MIC was defined as the lowest concentration of AO that inhibited visible growth. Each experiment was done in triplicate.

Plasmid curing (eliminating the plasmid from cells) of L. pentosus MP-10 was done as described by Adeyemo and Onilude41 with some modifications. Briefly, MRS broth (4 ml) supplemented with the sub-MIC of AO, as determined in this study, was inoculated with a selected colony of L. pentosus MP-10 grown onto MRS agar; then the cultures were incubated at 37 °C for 72 h. Serial dilutions of bacterial cultures in NaCl (0.85%) were plated onto MRS agar, and the resulting colonies, obtained after incubation for 48 h at 37 °C, were inoculated into MRS broth to obtain a pure culture. Cultures were maintained in 20% glycerol at −20 °C and −80 °C for short- and long-term storage, respectively.

To confirm that the resulting colonies were cured of plasmids, bacterial cultures (uncured and cured) were subjected to plasmid isolation as described by Abriouel, et al.42 and then visualized on 0.8% agarose gel electrophoresis (iNtRON Biotechnology) in 1xTBE (Tris-Boric acid-EDTA) buffer.

For additional confirmation, total genomic DNA (uncured and cured strains) was extracted using DNA Extraction Kit (Xtrem Biotech SL, Spain) according to the manufacturer´s instructions and tested for plasmid-borne genes. DNA quantification and quality assessment were carried out using a NanoDrop 2000 spectrophotometer (Thermo Scientific). DNAs were frozen at −20 °C until required and then subjected to PCR amplification of genes harboured by pLPE5, the biggest plasmid detected in L. pentosus MP-10. The PCR primers were designed in this study: Ars-pl5-F (5′-ATTATTTTGATCTCATTGATTTT-3′) and Ars-pl5-R (5′-TGAATAAACGAAACGGGAATGT-3′), yielding an amplicon of 570 bp. The 50 µl PCR mixture contained 20 ng of DNA, 0.5 μm of each primer (Ars-pl5-F and Ars-pl5-R), 200 μm of each deoxyribonucleoside triphosphate (Bioline), and 1 U of Taq DNA polymerase in 1X buffer according to the manufacturer’s instructions (Bioline). PCR was performed under the following conditions: one cycle at 95 °C for 3 min, 35 cycles at 95 °C for 30 s, 58 °C for 30 s, and 72 °C for 1 min and the final hold for 3 min at 72 °C. Analysis of PCR products was done by electrophoresis through a 1% agarose gel electrophoresis in 1xTBE (Tris-Boric acid-EDTA).

Effect of plasmid curing on growth, safety and functional properties of L. pentosus MP-10

Growth properties

To test whether there is any differences in growth between the uncured and the cured L. pentosus MP-10 strains, MRS broth was inoculated (1% v/v) with overnight cultures of each strain and then incubated at 37 °C for 24 h. Growth rates (OD600nm) were measured each hour using Microtiter plate reader (iMark Microplate Absorbance Reader, Bio-Rad instrument). Additionally, we measured pH at different time intervals (following 0, 8 and 24 h of incubation at 37 °C).

To determine the effect of pH on the growth of both strains, MRS broth was adjusted to different pH ranges (1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5 and 7.0) with phosphate buffer, and they were inoculated (1% v/v) overnight cultures of both strains and then incubated at 37 °C for 24 h, as described above.

To test whether brine conditions had an effect on the growth of the plasmid-cured versus uncured L. pentosus MP-10 strains in MRS broths under the following experimental conditions: unsupplemented vs. those supplemented with either 6.5% (or high concentration of 8%) NaCl or phenolic compounds, or modified MRS broth (without glucose) added with xylan (5 g/l) were inoculated with both strains as described above. Phenolic compounds were obtained from freshly pulverized olive leaves using RETSCH laboratory ball mills (Retsh MM 400). The leaf extracts were resuspended in LSM broth, centrifuged and the resulting supernatant was filtered (0.45 μm) and added at different concentrations (0.780 to 200 mg/ml) to MRS broth. The cultures were incubated at 37 °C for 24 h and the OD600nm was measured as described above.

In all cases, experiments were done in triplicate.

Evaluation of metal tolerance

The sensitivity of both L. pentosus strains (MP-10 and MP-10C (cured)) towards metals: cadmium (CdSO4·8/3H2O), cobalt (CoCl2), copper (CuCl2·2H2O), iron (FeSO4·7H2O), mercury (HgCl2), aluminium (Al2O3), or zinc (ZnCl2) was tested in LSM broth supplemented with 0 to 10 mg/ml of each metal and then inoculated with 2% (v/v) of an overnight culture of each strain. After 24 h of incubation at 37 °C, the MIC from each metal exposure was determined as described above, which corresponded to the lowest concentration that completely inhibited visible growth.

To analyse the removal of metals by cured and uncured L. pentosus MP-10, MRS broth supplemented with ½MIC of each metal was inoculated with 2% (v/v) of an overnight culture of each strain and then incubated 24 h at 37 °C. After incubation, the bacterial cells were removed by centrifugation and kept for the subsequent examination of metal sorption. The resulting supernatants were filter sterilized using a 0.22 μm filter (Millipore, Spain) and then used to check metal removal. MRS broth added either with different metals (with ½MIC) or not were used as positive and negative controls, respectively. The positive controls (MRS broth with individual metal added: Fe at 2 mg/ml; Al, Co and Cu at 1 mg/ml; Zn at 0.5 mg/ml; Cd at 4 μg/ml and 0.5 μg/ml; and Hg at 1 μg/ml and 0.5 μg/ml) were considered “100%” baselines to calculate relative metal removal rates (as a percentage).

Metal concentrations were measured using 7900 ICP-Mass Spectrometer (Agilent, USA) with graphite tube atomizer and autosampler, a superior matrix tolerance and advanced collision/reaction cell (CRC) technology to remove the polyatomic interferences that can affect some of the trace elements. The spectrometer software was Agilent ICP-MS MassHunter Work Station, which provides simple autotuning functions, and a Method Wizard automates the method setup process.

Biosorption of metals by L. pentosus MP-10 was further examined using scanning electron microscope (SEM) coupled with energy dispersive X-ray spectroscopy before and after metal uptake. For this, a drop of the bacterial pellet, which had been previously exposed to a metals (as previously described), were disposed into microporous capsules (ANAME, Spain), dried and then dehydrated in a series of 20, 40, 60, 80, and 100% ethanol solutions (15 min each) before suspension in acetone for 1 h. After this, the capsules were subjected to critical-point drying before examination by SEM (FESEM, MERLIN de Carl Zeiss, Oxford).

Safety and probiotic properties

To determine differences in antimicrobial (antibiotic and biocide) susceptibility of L. pentosus MP-10C versus wild strain, we determined the MIC of several antimicrobials following the method previously described by Casado Muñoz, et al.42,43 using LSM broth (Oxoid).

To determine if plasmids further play a role in several probiotic peroperties, we analyzed acid- and bile- tolerances, auto-aggregation, co-aggregation with pathogens (L. innocua CECT 910, S. aureus CECT 4468, E. coli CCUG 47553, and S. Enteritidis UJ3449) and mucin adhesion in both L. pentosus strains (MP-10 and MP-10C) according to the methods reported by Pérez Montoro et al.35.

Gene expression analysis

To analyse the role of plasmid in several metabolic and probiotic properties, both the uncured and cured L. pentosus strains were subjected to RNA extraction using Direct-zol™ RNA Miniprep (Zymo Research, California, USA) according to the manufacturer’s instructions. RNA quantification and quality assessment were carried out by using a NanoDrop 2000 spectrophotometer (Thermo Scientific). RNAs were adjusted to a concentration of 500 ng/ml and frozen at −80 °C until required for analysis.

The expression of selected genes (Table S1) was determined by quantitative, real-time PCR (qRT-PCR) using SensiFASTTM SYBR & Fluorescein One-Step Kit (BIOLINE) as reported in Pérez Montoro et al.35.

Statistical analysis

All analyses were performed in triplicate. Statistical descriptors were calculated using Excel 2007 (Microsoft Corporation, Redmond, Washington, US), e.g., determining averages and standard deviations. Statistical comparison of growth and probiotic properties assays were conducted by analysis of variance (ANOVA) using Statgraphics Centurion XVI software (Statpoint Technologie, Warrenton, Virginia, US). The same software was used to perform Shapiro–Wilk and the Levene tests to check data normality and to perform 2-sided Tukey’s multiple contrast to determine the pair-wise differences between strains. Level of significance was set at P < 0.05.