The Lactobacillus strains (Table 1) used in this work were confirmed for riboflavin production by an array of analytical methods viz. Polymerase chain reaction (PCR) based method (presence of riboflavin biosynthesis genes), Spectrophotometric method, Microbiological assay method, and High Performance Liquid Chromatography in our previous studies (Thakur and Tomar, 2015a; Thakur et al., 2016b). All the strains stored previously at -80°C in MRS supplemented with glycerol (20% v/v) were routinely cultured on de Man-Rogosa -Sharp (MRS) medium (Sigma- Aldrich, St. Louis, MO, USA) for this study.

Purified PCR products (HiPuraA^TM purification kit, Himedia, India) were used for cloning of all the four genes. The cloning vector used in this study was _PTZ57R/T clone vector amp (InstClone PCR cloning kit, Stratagene, USA). The clones were transformed into competent cells of Escherichia coli (E. coli) (XL1 blue). The successful clones were picked from Luria broth+ ampicillin plates and amplified for target genes by colony PCR method followed by plasmid isolation. The positive clones were identified by PCR analysis of plasmid DNA by using primers used in our previous study (Thakur et al., 2016b). The nucleotide sequencing was performed by sequencing services provided by Xcelris Labs, Ltd, Ahmedabad, India. The chromatograms of sequences obtained were analyzed and converted to Fasta using Bio-Edit Software. Nucleotide sequence similarity searches were performed for the obtained sequences by matching with previously published complete genome of Lactobacillus species of interest.

Non-denaturing PAGE was used to detect the difference is size of all the four Rib genes amplified in different lactobacilli. Silver staining was used to view the band pattern in the PAGE after the final gel run.

The test isolates were washed thrice with saline solution (0.85% m/v NaCl), resuspended in this solution and used to inoculate at 2% (v/v) riboflavin-free culture medium (Riboflavin Assay Medium, Difco, Becton, Dickinson, and Co., Sparks, MD, USA), reconstituted skim milk and whey based medium and then incubated without agitation at 37°C for 18 h. The optical density of selected isolates was observed before harvesting them for RNA isolation. The log count/ml was checked at lag, log, and stationary phases of growth.

The primers (Table 2) were designed by aligning sequences of riboflavin operon using CLUSTALW program. The Lactobacillus fermentum IFO3956 strain was considered for primer selection: GenBank accession number NC_010610. The house keeping genes for normalizing real time reaction were synthesized for REC gene essential for the repair and maintenance of DNA and TUF gene encoding elongation factor from the database of genome in NCBI.

Selected Lactobacilli were grown in 10 ml of MRS, Riboflavin Assay Medium (devoid of riboflavin) (RAM), Skim milk and Whey based medium. The RNA was extracted after 6, 12, 18, and 24 h of incubation using TRIzol reagent followed by cell lysis by lysozyme (10 mg/100 ml) (Sigma, USA) (Ramiah et al., 2007). The quality of isolated RNA was checked (Sambrook and Russell, 2001). RNA was quantified and its purity of RNA was judged and used for reverse transcription. The cDNA was prepared with cDNA kits (RevertAid^TM First strand c-DNA synthesis kit, Fermentas, India), according to the manufacturer’s instructions. SYBR Green I Master mix (Roche) on 2 μL of diluted cDNA (1:1) using exon-spanning primers, 5 μL of SYBR green buffer 2X (Roche) and 2.5 pmol of each primer (Table 2) for a total volume reaction of 10 μL were used for qualitative PCR. The amplification was run in Lightcycler^® 480 II, and the results were analyzed using Lightcycler^® 480 II software release 1.50 SP3. The PCR conditions were as follows: initial denaturation at 95°C for 10 min, followed by 40 cycles of amplification at 63°C for 30 s and 72°C for 30 s. At the end of the each run a melting curve was achieved from 70 to 95°C and continuous fluorescence measurement was taken. A melting curve analysis was performed in order to verify the specificity of real-time PCR (RT-PCR) and finally, a cooling step to 4°C was achieved. Fluorescence was measured once every cycle after the extension step using filters for SYBR Green I (excitation at 465 nm and emission at 510 nm). To calculate the threshold cycle value, the normalized fluorescence data was converted to a log scale and threshold value was determined. The quantitative data of RT-PCR is generated on the basis of number of cycles required for optimal amplification generated fluorescence to reach a specific threshold of detection (the quantification cycle:Cq values) (Bustin et al., 2009). The comparative critical threshold (ΔΔCT) method was used to calculate the relative expression ratios in which the amount of target RNA is adjusted to a reference (internal target RNA) (Livak and Schmittgen, 2001). The Prism 7.00 was used to analyze the RT-PCR data sets.

The roseoflavin-resistant strains of KTLF1 was performed (according to Burgess et al., 2004, 2006) by exposing wild strain to increasing concentrations of roseoflavin (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in RAM. Further experiments were carried out by using subsequent inoculum in 1 mL of RAM supplemented with roseoflavin. From the culture grown at the maximum range of roseoflavin, 15 separated colonies were randomly isolated after spreading onto MRS agar plates, and those stocks were stored at -80°C in CDM roseoflavin-free supplemented with 20% of glycerol (Russo et al., 2014).

To discriminate the riboflavin producing isolates and riboflavin biosynthesis genes on the basis of mRNA expression levels in different media and at different time and principal component analysis (PCA) treatment, IBM SPSS Statistics 21.0 software program (IBM, Armonk, NY, USA) was used (ANOVA followed by a Tukey’s post hoc test). P < 0.05 was considered as statistically significant.

Purified amplicons were ligated into PTZ57R/T cloning vector. The ligates were transformed into competent cells of E. coli XL1 blue strain. The recombinant clones showed white (recombinant) colonies on LB agar plates supplemented with ampicillin (100 μg/ml). Ten randomly selected recombinant clones were analyzed by colony PCR for Rib genes. Consequently, positive clones were used for plasmid DNA analysis and isolated plasmids were further confirmed for their size by PCR and subsequently sequenced (Supplementary File). The sequences obtained from the isolates were compared by BLAST analysis for similarity check with three reference strains of Lactobacillus submitted to NCBI GenBank (Supplementary File). No size variation was found in the studied genes in different lactobacilli after staining the PAGE with silver staining (Supplementary File).

The CFU/ml was observed for KTLF1 at different growth phases (Figure 1) and the mRNA and the cDNA were prepared as described above. During the early growth, expression levels remain low and get elevated at later phase. The expression levels of Rib 1, Rib 4 were significantly higher in all the media used at four time intervals (Figure 2), whereas Rib2 and Rib 3 have shown almost constant regulation with different variables (Isolates, Media, and Time) (Figure 2). After incubation in different media at different time, the level of mRNAs expression was found to be changed in all the tested isolates. Particularly in KTLF1 strain, the mRNA expression level increased significantly in MRS at 12–18 h, in RAM at 6–12 h, in Milk at 6 h and in whey the upregulation was observed at 12 h (Figure 3). Among all the media used, RAM has shown increase in relative expression followed by MRS, Milk and Whey (Figure 4). Overall, there was no statistical difference in expression levels of Rib2 and Rib3 at different variables but a gradual decrease in the mean expression with time was recorded. Figure 5 revealed the marked difference of increase in the intensity of RT-PCR products at different variables in KTLF1 in MRS and RAM which correspond significantly to change in mRNA expression profile. In order to study the effect of different parameters mentioned on riboflavin production, relative expression levels of the Rib 1, Rib 2, Rib 3, and Rib 4 genes were calculated in comparison with lowest riboflavin producing strain obtained from our previous study. The gene expression profiles for each strain grown under these conditions were different (Figure 2). Among the five tested strains, KT2 has shown significant variation in gene expression profile across the different incubation time intervals. The Rib 3 and Rib4 genes had a steady transcript level (fold change equal to one) in all the tested strains except in KT2 strain. The most significant difference was found at 6 and 24 h. This showed that expression of these genes was not much affected with the strain in two different media. Whereas, in KTLF1 (Figure 3), effect of media on different genes was significant in Milk and whey however, the relative expression levels of these genes was steady in MRS and RAM across different incubation. In particular, the expression levels of these genes were found to be lowered with increased incubation time in all the four media. In KTLF1, all the genes have shown significant variation in their expression in milk, Whey and MRS (Figure 4). In milk and whey, the fold increase was found in descending order with the increased incubation times, whereas, in whey media, except Rib 1, remaining three genes have shown steady expression over different incubation periods. In MRS and RAM, except Rib 1 and Rib 4, remaining two genes have shown significant change in expression profile.

The multivariate analysis was used for comparison of experimental data obtained for five lactobacilli for mRNA expression of four Rib genes (PCA). Two dimensional plots are drawn in Figure 6. In plot 1 (Figure 6A), contribution of the media with respect to variance is shown. The first two components present the total variance of 53.7%. The discrimination of isolates along PC1 is mainly due to RAM and Whey. Whereas along PC2, the isolate KT2 is discriminated vis-à-vis other isolates. In plot 2 (Figure 6B), the first two components present the total variance of 73.7%. The discrimination along samples along PC1 is mainly due to Rib 1 and Rib 4, whereas Rib 2 and Rib 3 are responsible for discrimination along PC2.

By following the procedure described by Burgess et al. (2006), KTLF1 was exposed to roseoflavin, a structural analog of riboflavin, which induces mutations in riboflavin-producing strains leading to a novel producer phenotype of the vitamin. Roseoflavin resistant variants were isolated and inoculated in the riboflavin-free medium and incubated for 16 h at 30°C. Both wild type and variant strains were re-inoculated in MRS, RAM, Milk and Whey based media at 37°C, 24 h and harvested for RNA isolation and riboflavin production. From five isolated variants, only KTLF1 variant [KTLF1 (4)] was able to up regulate mRNA expression of Rib genes and the increase in riboflavin production was more than 3.5-fold (4.5 mg/L) as compared with wild type strain in a culture medium without riboflavin (Figure 7A). The spike in mRNA expression level was observed after the exposure of roseoflavin in all the four genes across the media at different times of incubation (Figure 7B). The maximum up regulation was observed in Rib1 followed by Rib4. The fold increase in expression was lower as to that of other wild type strains (Figure 7B).