Characterization of genomic, physiological, and probiotic features of Lactiplantibacillus plantarum DY46 strain isolated from Turkish fermented turnip juice (Shalgam)


 A new Lb. plantarum strain DY46 was isolated from a traditionally fermented non-alcoholic beverage called shalgam from the Southern region of Anatolia following incubation on MRS agar at 30°C for 5 days. DY46 is gram-positive, short rod and catalase-negative. This bacterium fermented 22 of the 49 substrates tested on API CH50 fermentation panels. Whole-genome sequencing was performed using the Illumina Miseq platform to learn more about the metabolic capabilities of DY46. The sequences were assembled into a 3.32 Mb draft genome using PATRIC 3.6.8. consisting of 153 contigs, and preliminary genome annotation was performed using the RAST algorithm. The DY46 genome consists of a single circular chromosome of 3,332,827 bp that is predicted to carry 3219 genes, including 61 tRNA genes, 2 rRNA operons. The genome has a GC content of 44.3% includes 98 predicted pseudogenes, 25 complete or partial transposases and 3 intact prophages. DY46 genome also predicted to carry genes of Plantaricin-E, Plantaricin-F and Plantaricin-K showing the antimicrobial potential of this bacterium which can be linked-to in vitro antagonism tests that DY46 can inhibit Salmonella Typimirium ATCC14028, Klebsiella pneumonie ATCC13883, and Proteus vulgaris ATCC8427. The acid and bile tolerance of DY46 revealed this strain could potentially pass through the stomach and reach into the gut to provide probiotic therapeutic affects on health.

from pellet was carried out using the PureLink Genomic DNA Mini Kit (Invitrogen, Thermo-Fisher Scienti c, USA) per manufacturer's instructions for gram-positive bacteria. The quality and concentration of genomic DNA were checked by a Qubit 3.0 uorometer (Thermo-Fisher Scienti c, US) and agarose gel (1.5%). Identi cation of the test strain was carried out by full-length nucleotide sequencing of the 16S rDNA gene. Universal bacterial primers 27F (5'-AGAGTTTGATCCTGGCTCAG-3') and 1492R (5'-GGTTACCTTGTTACGACTT-3') were employed (Ni et al. 2015). Ampli cation reactions were prepared with a total volume of 30µL containing, 3 units of EasyTaq DNA polymerase (TransGen Biotech, Beijing, China), 3µL of 10x EasyTaq Buffer, 20µM of forward primer, 20 µM reverse primer, 0.9µL DMSO (3%), 2.4 µL 2.5mM dNTP, 22.1 µL nuclease-free water and 50ng template DNA. Reaction mixes were ampli ed with a thermal cycler (ABI Veriti 96, Thermo-Fisher Scienti c, California, USA) under the following conditions: the initial denaturation of DNA for 5 minutes at 94℃ was followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 52°C for 20s, extension at 72°C for 90s, and followed by a 5 min nal extension at 72°C. The PCR products were run on agarose gel (1.5%) and sequenced by the Sanger method (Ficus Biotechnology, Ankara, Turkey). The obtained sequences were searched against the Basic Local Alignment Search Tool (NCBI-BLAST). After con rming the DY46 strain via 16S rDNA, the whole genome sequencing libraries were constructed using Nextera XT DNA Library Preparation Kit and sequencing was ful lled by Illumina Miseq platform as paired-end (PE) 2x250 bases read. The low-quality reads were ltered and assembled in the genome assembly service of PATRIC 3.6.8. (https://patricbrc.org/app/Assembly2) with an auto strategy (Davis et al. 2020).

Bioinformatic analyses
Genome annotation and comprehensive genome analysis were performed by using NCBI Prokaryotic Genome Annotation Pipeline (PGAP) and PATRIC 3.6.8. platform (Tatusova et al. 2016; Davis et al. 2020). A circular genome map of the strain genome was generated with CG view server (Stothard and Wishart 2005). The calculation of orthologous average nucleotide identity values (OrthoANI) of the DY46 and other compared L.plantarum, L.pentosus and L. paraplantarum strains were implemented by OrthoANI tool v0.93.1 (Lee et al. 2016). Prediction of metabolic pathways of L. plantarum DY46 was carried out using BlastKOALA for scanning against the KEGG database (Kanehisa et al. 2016). The bacteriocin production responsible gene cluster prediction was ful lled using the BAGEL 4 webserver (http://bagel4.molgenrug.nl/) (Heel et al. 2013). Following, each member of predicted gene clusters was con rmed via the NCBI protein BLAST suite (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Theoretical isoelectric point (pI) and molecular mass (MW) of the predicted peptides were calculated by Expasy Compute pI/MW online tool (https://www.expasy.org/resources/compute-pI-mw). The prophage regions on the genome of DY46 were identi ed and annotated with the PHASTER-Phage Search Tool Enhanced Release (Arndt et al. 2016). In order to identify the genes transferred horizontally, all protein-coding genes obtained from PHASTER were screened against the non-redundant protein (NR) database by performing protein-BLAST. If a gene's homologous protein was found to match a microorganism other than L. plantarum ≥ 80%, that gene was considered horizontally transferred

Carbohydrate fermentation
The carbohydrate fermentation patterns of the DY46 strain were determined using an API 50 CHL kit (BioMérieux, Marcy l'Etoile, France) consisting of 49 different carbohydrate tests in accordance with the manufacturer's protocols.

Determination of antibiotic susceptibility
Antibiogram assays were conducted to nd out the resistance or sensitivity of the DY46 strain against commonly used antibiotics. Ready to use commercial antibiotic disks (methicillin, vancomycin, amikacin, kanamycin, azithromycin, tetracycline, penicillin G (Bioanalyse); ampicillin, oxacillin, carbenicillin, amoxycillin, streptomycin, erytromycin, rifampacin (oxoid)) were employed for antibiotic susceptibility testing of L. plantarum DY46. The application of disk diffusion assay was performed according to a modi ed Kirby-Bauer method (Sharma et al. 2017). Interpretation of inhibition zone (mm) results was carried out with regard to Clinical and Laboratory Standards Institute's performance standards for antimicrobial testing (CLSI M100-S22, 2012). Results with an inhibition zone less than or equal to 14mm were noted resistant (R). Additionally, inhibition zones greater than 20 mm were considered sensitive (S) and between 15-19 mm were accepted as semi-sensitive or intermediate (I).

Probiotic Properties
To determine probiotic properties of the DY46: β-haemolysis, cell surface hydrophobicity, cellular auto-aggregation, GABA-production capacity tests and antibacterial activity assay against several pathogens were performed, respectively. On the other hand, the growth kinetics of the DY46 were analyzed at different pH and bile concentrations. Evaluation of the β-haemolytic activity of the DY46 was performed using 5% sheep blood containing Columbia agar plate. The isolate was streaked on the Columbia agar followed by incubation at 37℃ for 48 h under anaerobic conditions (Angmo et al. 2016). Cell surface hydrophobicity and auto-aggregation assays were performed in compliance with Goel et al. (2020). Antibacterial activity assay was executed by the agar well diffusion method (Mishra and Prasad 2005). The supernatant of 18-20 h grown DY46 was analyzed against E. coli O157:h7 (ATCC 43895), S. aureus (ATCC 25923), B. Cereus (ATCC 33019), S. Typhimurium (ATCC 14028), P. mirabilis (ATCC 29906), P. vulgaris (ATCC 8427) and K.pneumoniae (ATCC 13883). Growth characteristics of the DY46 were investigated at different pH and bile concentrations. First, the overnight grown fresh culture of DY46 was prepared in MRS broth at 30°C. Second, the MRS medium was prepared to create media with different conditions. Third, Oxyrase (Sigma-Aldrich, USA) was added to the prepared media to reduce the oxygen level. The pH of the MRS medium was adjusted to a value between 6.8-8.4 which is optimum for oxyrase. After adjusting the pH value, oxyrase was added to the medium in proportions according to McMahon et al. (2020). MRS medium with ve different pH values (pH 2, 3, 4, 5, 7) was prepared using 3N HCl and 3N NaOH. For the preparation of bile concentrations of a different pre-separated MRS medium, the pH was adjusted to 6.5. Four different bile concentrations (0.3%, 0.5%, 1% and control (no bile)) were prepared using ox bile extract (Sigma, Germany). After preparing all media, they were incubated at 36.5°C for 30 minutes to activate oxyrase. Growth measurements were performed in HIDEX Sense Microplate Reader (Hidex, Finland) using 96 well-plates with lid. Each well was inoculated with 200 µl from media containing the previously indicated ratios (McMahon et al. 2020).
Samples were inoculated in quadruplicate. Spectrophotometric measurements were carried out at 30°C and 37°C with a 300 rpm orbital shake. An optical density (OD) measurement at a wavelength of 600 nm was performed every 20 minutes at 72 hours post-inoculation.

Results And Discussion
The Genome of the DY46 strain The whole genome of the L. plantarum DY46 strain composed of a circular chromosome of 3,332,827 bp with a GC ratio of 44.3395%, a total of 3.219 genes, comprising of 3.054 protein-coding sequences, 61 tRNAs, 2 rRNAs, 4 non-coding RNAs and 98 pseudogenes ( Fig. 1.  As expected, LL441 (cheese) and RI-113 (salami) showed a greater genetic distance than other strains due to their isolation sources (Gonzalez et al. 1994; Flórez and Mayo 2018; Inglin and Meile 2020). L. plantarum can be found in many different environments and shares its ecological niche with L. pentosus and L. paraplantarum and other facultative heterofermentative members of the genus Lactobacillus (Stiles and Holzapfel 1997). Besides, L. plantarum, L. pentosus and L. paraplantarum display very close phenotypes and are genotypically similar due to their rRNA have as same as sequence identity (> 99%). Therefore, these species cannot be discriminated from each other using 16s rDNA sequence analysis (Parente et al. 2010). According to Ortho ANI results which were shown in Fig. 2. L. paraplantarum and L. pentosus exhibit 85.89% and 79.93% similarity with the DY46 strain, respectively. It is usually reported that the ANI value should be above 95-96% to consider that the genomes of the two species are the same (Lee et al. 2016). This con rms that the DY46 strain belongs to the Lb. plantarum species.  (Table S21). It is important to note that DY46 cannot metabolize D-Sorbitol, although its genome has sorbitol-6-phosphate 2-dehydrogenase (srlD) and the glucitol/sorbitol phosphotransferase system (srlB, srlE and srl A ) genes. This might be due to the lack of expression of the abovementioned genes encoding speci c enzymes required to metabolize sorbitol (Buron-Moles et al. 2019). Lactiplantibacillus plantarum is a protean and resilient species that can grow on a wide range of carbohydrate sources. This phenotypic character is associated with genes involved in carbohydrate metabolism and transport. Most of the transporters involved in carbohydrate metabolism are located in the phosphoenolpyruvate-dependent sugar phosphotransferase system (PTS) (Gänzle and Follador 2012; Gao et al. 2020). The entire PTS of DY46 strain that was encoded by its genome, comprises PTS System Enzyme I (general enzyme gene, ptsI), phosphocarrier protein HPr gene (ptsH), 26 complete/incomplete substrate-speci c enzyme II (EII) complexes genes (Table S24). In the genome of the DY46 strain, the genes of glucose-glucoside, fructose-mannose-sorbose, glucitol, Lactose-N, N'-diacetylchitobioseβ-glucosides and N-Acetylglucoseamine EII complex families were observed as multiple copies, while L-ascorbate, sorbose, mannitol and galactitol EII complex genes were found as single copies. Additionally, the DY46 possesses several other carbohydrate transporter encoding genes on its genome. However, their substrate speci city is unknown and could not be predicted. According to Kleerebezem et al. (2003) various sugar transporter systems are known to import more than one substrate. Lb. plantarum is classi ed into the facultative heterofermentative Lactobacillus species, which utilize the sugars by way of glycolysis (Embden-Meyerhof-Parnas (EMP) pathway) or the phosphoketolase (PK) pathway, leading to homolactic or heterolactic fermentation routes, respectively . DY46 genome carries 6-phosphofructokinase 1, fructose-bisphosphate aldolase, glucose-6-phosphate isomerase, transketolase and phosphoketolase genes, which are encoding the key enzymes of EMP and PK pathways (Eiteman and Ramalingam 2015). The genes encoding enzymes related in the intact EMP and PK pathways were predicted in the genome of the DY46 strain and listed in Table S23. Furthermore, 1-phosphofructokinase enzyme encoding fruK gene was detected which is used as the key gene for differentiation of hetero-and homofermentative lactobacilli species

Bacteriocin biosynthesis
According to the results of the whole genome search of the DY46 strain against the BAGEL database, the gene cluster responsible for bacteriocin biosynthesis consists of 26 genes and its total length is approximately 24.2 kb (Fig. 3). In this gene cluster, transport-related genes, immunity protein and plantaricin precursor genes and several core genes (Pln E, Pln F and pln K) are encoded and all protein sequences con rmed by protein BLAST (Table S1). L. plantarum DY46 strain was found to have the same core genes encoding Class II bacteriocins as ATCC8014 (Yu et al. 2020). The plnEF locus is widely distributed among L. plantarum strains isolated from various ecological niches. The well-studied plnEF loci have also been reported in L. plantarum WCFS1, NC8, JDM1, C11, V90, J51 and J23 strains, respectively (Tai et al. 2015). On the other hand, the isoelectric points (pI) and amino acid lengths of the pln E and pln F mature peptides that do not have the GG leader sequence that we detected (Table S2) are identical to pln E and pln F bacteriocins that were previously reported in WCFS1, NC8, J23, J51, C11 and V90 strains (Diep et al. 2009). Normally, plnJ and pln K peptides, which are subunits of plantaricin JK, are encoded in the same gene cluster/operon and more effective together, whereas in the present study only the pln K peptide was detected with 95.35% percent identity (Todorov 2009). Moreover, the pI (8.59) and length of the mature pln K peptide (28aa) detected have differed signi cantly from the mature plnK peptides (pI:10.52; 32aa) as previously reported in C11, NC8, V90 and WCFS1 (Diep et al. 2009). Apart from the core genes, the presence of secretion genes pln H (HlyD, accessory protein for ABC-transporter; (Accession no: WP_027821501.1), pln G (LanT, Bacteriocin ABC-transporter; WP_027821502.1) have been veri ed in the plantaricin gene cluster. Bacteriocin ABC transporters are involved in the transport of the mature peptide through the cell membrane, which is formed by deleting the leader peptide sequence from prebacteriocin (Havarstein et al. 1995). The accessory protein (also called the accessory factor) is another necessary component for the ABC transporter system-dependent translocation process (Nes et al. 1996). Another gene cluster member identi ed is the putative Na + /H + antiporter protein (orf00033: AFM80194.1), which sustainably maintains intracellular proton balance and leading to the enabling of ATP required for ABC transporters (Jia et al. 2017). In addition, the bacteriocin gene cluster contains genes encoding orf00020 (CAAX amino terminal protease family protein; EFK30757.1) and orf 00028 (bacteriocin immunity protein; WP_127526380.1) immunity proteins that play a role in protecting bacteria from their mature bacteriocins (Todorov 2009). The other members of the bacteriocin biosynthetic gene cluster were listed in supplemental table S1.

Antibiotic resistance
Antibiotic susceptibility of the DY46 strain was evaluated as reported by the Clinical and Laboratory Standards Institute's performance standards. The zone of inhibition (ZOI) values of fourteen antibiotics tested against the DY46 strain was shown in Table S3 with resistome search match results. DY46 was found to be resistant (ZOI ≤ 14mm) to methicillin (5 µg Table S3 have been detected with PATRIC 3.6.8 and KEGG databases. The identi ed genes were found to be related to β-Lactams (9), Streptomycin (2), Vancomycin (7), Macrolides (1), Tetracyclines (1) and Rifampacin (1). It is commonly accepted that Lactobacillus species have very high resistance to aminoglycosides. Lb. plantarum is known to be resistant to vancomycin due to its intrinsic peptidoglycan precursors consisting of D-lactate instead of D-alanine at the C-terminus (Gueimonde et al. . In addition, streptomycin resistance responsible gibD and S12p genes were detected. However, no speci c genes were found for amikacin and kanamycin, even if resistance was present. This is because genotype and phenotype do not overlap completely (Zhang et al. 2012). Apart from these, RlmA (II) (23S rRNA (guanine(748)-N(1))-methyltransferase), S10p (SSU ribosomal protein S10p (S20e)), rpoB (DNA-directed RNA polymerase beta subunit) genes detected which were related with macrolides, tetrcyclines and rifamycins, respectively. Moreover, mecA (penicillin-binding protein 1A), pbp2a (penicillin-binding protein 2A) and PenP (beta-lactamase class A) major genes responsible for beta-lactam resistance were detected on the DY46 genome. Methicillin and oxacillin resistance is known to be associated with penicillin-binding proteins. Interestingly, resistance to Penicillin G was observed in DY46, although Lactobacillus are known to be susceptible. Some authors have reported Penicillin G-resistance in recent years in some strains of Lactobacillus rhamnosus, Lactobacillus reuteri, and Lactobacillus plantarum, which also con rms our study (Abriouel et al. 2015;Zheng et al. 2017). Because of the growing concern that foods and/or common bacteria may serve as potential reservoirs for antimicrobial resistance genes, probiotics must not carry transferable antimicrobial resistance genes to be used for humans or animals (Zhang et al. 2012). Because many Lactobacillus species have intrinsic resistance to many antimicrobial compounds, and such resistance is known not to be associated with any particular safety concerns. However, the intrinsic antibiotic resistance genes on the chromosome should not be anked by integrases and/or transposases. As a result of Protein BLAST screening for antibiotic resistance genes detected in this study, no evidence of horizontal gene transfer was found (Table S20).

Prophages and horizontal gene transfer
Prophage search results display 9 prophage regions (three intact, two questionable and four incomplete) found in the genome of the DY46 strain and summarized in Table 2. One of the three intact prophage regions showed similarity with Lactob_Sha1_NC_019489 (48.8Kb), (region 1), and the other two like Lactob_phig1e_NC_004305 (39.9Kb) and Staphy_SPbeta_like_NC_029119 (29.6Kb), region 2 and region9, respectively (Fig S1.). It was determined that Lactob_Sha_1 and Lactob_phig1e showed the highest protein matching among the identi ed prophages. These are the most common temperate prophages ever described in L. plantarum strains (Pei et al. 2020). All the prophage regions have attL/attR sequences and integrase except for region 6 (Paenib_PBL1c) and 7 (Bacill_vB_BtS_BMBtp14). In bacterial genomes, integrases are functional identi ers for phages, pathogenicity islands and integrative plasmids (Juhas et al. 2009;Liu et al. 2015). Three integrases (PP_00611(region 1), PP_01193(region 2) and PP_03267 (region 9)) were determined in the identi ed intact regions. The location of attL/attR sequences varies within intact phages. The attL sequences in regions 2 and 9 are located upstream of the integrase, while the attR sequence of region 1 was found downstream of the integrase. Additionally, attL and attR sequences of phage 1 and phage 9 were identical, but the attL and attR sequences of phage 2 are different from them. The Intact phage region 1 extends from 558,178 bp to 607,042 bp of the genome and includes 58 protein-coding sequences containing all prophage components from PP_00554 (transposase) to PP_00611 (phage integrase). The intact region 2 extends from 1,217,186 bp to 1,257,125 bp of the genome and consists of 49 protein-coding sequences containing prophage components from PP_01193 (phage integrase) to PP_01241. Moreover, the intact region 3 (9) is located between 3,302,034 bp to 3,331,708 bp of the genome and consists of 33 protein-coding sequences containing prophage components from PP_03263 (protease) to PP_03295. Unlike previously reported for L. plantarum WCFS1 (Ventura et al. 2003) and 5 − 2 strains , only the two intact phages (Sha1 and phig1e) were found to contain all packaging/head/tail gene clusters, DNA packaging genes and the lysis cassette. All components of the identi ed prophage elements have been listed in Table S4-S12.

Probiotic properties
When new probiotic strains are discovered, certain characterization tests are required to con rm probiotic properties. Therefore, probiotic characterization tests were performed to con rm the probiotic properties of the DY46. The β-haemolysis test results showed that DY46 does not have β-hemolytic activity. The cell surface hydrophobicity of DY46 characterized by using xylene. As shown in Fig. 4A, the cell surface hydrophobicity of DY46 appears to increase in direct proportion to the bile salt concentration. Cell surface hydrophobicity of the DY46 was determined as 33%, 38.5% and 46.1% at 3, 5 and 10g/L bile salt concentrations, respectively. The cell surface hydrophobicity of the control sample was found to be at 4.38%. However, similar to present work, it has been reported in previous studies that some lactobacilli, including L. acidophilus and L. johnsonii strains displayed surface hydrophobicity as low as 2-5% (Rijnaarts et al. 1993;Schillinger et al. 2005). Kaushik et al. (2009) reported that such large differences in cell surface hydrophobicity could occur due to differences in the expression level of cell surface proteins, depending on environmental conditions and/or bacterial strain.
Auto-aggregation is an important bacterial characteristic in different ecological niches, especially in human and animal mucosa where probiotics confer health bene ts. The auto-aggregation capacity is an important factor for maintaining su cient numbers of probiotic strains under adverse conditions of the oral cavity and gastrointestinal tract. The cellular auto-aggregation test results presented in Fig. 4B  Overall, better growth kinetics achieved at 37ºC for pH conditions tested though lower pH values reduced or decreased growth completely. Although a shorter lag phase seen at 37ºC with all bile concentrations evaluated, DY46 cells better-tolerated bile salt at 30ºC with higher µ max and nal cell concentrations obtained. The DY46 can be resisting and proliferating under adverse conditions with moderately lower pH values and bile concentration mimicking the human GIT.
In conclusion, whole-genome sequencing and physiological characterization of Lb. plantarum DY46 isolated from Shalgam has been performed to determine probiotic properties of this novel strain. Genome analysis revealed this strain follows a facultative heterofermentative sugar metabolism where hexoses are cleaved thru glycolysis versus pentoses are hydrolyzed via the pentose phosphate pathway. Genome evidence predicted DY46 could biosynthesize Plantaricin-E, Plantaricin-F, Plantaricin-K showing the antimicrobial potential of DY46 which was con rmed by in vitro antagonistic activity test that supernatants of DY46 culture provided inhibition zones against K. pneumonia ATCC 13883, P. vulgaris ATCC 8427, S. Thypimirium ATCC 14028. Also, DY46 is tolerant to acid and bile concentrations mimicking human gastrointestinal conditions. Overall, Lb. plantarum DY46 is a promising bacterium possessing certain probiotic traits con rmed by in vitro analysis and perhaps a potential dietary supplement candidate that might provide therapeutic bene ts to the host.   The predicted gene cluster responsible from the biosynthesis of Plantaricins by using BAGEL4 webserver.