Introduction

Cobalamin (vitamin B12) is a structurally complex molecule consisting of four linked pyrrole rings and the cobalt ion in the center. There are four chemical forms of cobalamin that differ in the upper ligand: hydroxocobalamin (OHB12), methylcobalamin (CH3B12), deoxyadenosylcobalamin (AdoB12), and most chemically stable, cyanocobalamin (CNB12).

The chemical synthesis of cobalamin involves approximately 70 reactions. Microbial synthesis, which can be aerobic or anaerobic, involves fewer steps (Fig. 1). De novo synthesis involves about 30 reactions starting from glutamate. The salvage pathway is shorter than de novo synthesis, and it involves 12 genes1, 2. Pseudomonas denitrificans, Propionibacterium shermanii, Sinorhizobium meliloti, Eschericha coli and Bacillus megaterium are the main producers of CNB12 at the industrial scale1. Organisms that use vitamin B12 in their metabolism, and at the same time do not have the gene repertoire enabling its biosynthesis, use exogenous cobalamin actively transported through dedicated ABC transporters3, 4.

Figure 1
figure 1

Synthesis of vitamin B12 in bacteria (figure adapted with permission from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5282855/ under the terms of the Creative Commons Attribution 4.0 International License).

Full size image

Cobalamin influences cell metabolism via two mechanisms. It acts as a cofactor for enzymes, and regulates gene expression through so-called riboswitches. There are three major subfamilies of vitamin B12-dependent enzymes: AdoB12-dependent isomerases, CH3B12-dependent methyltransferases, and dehalogenases. The isomerases are the largest subfamily of B12-dependent enzymes. They play important roles in fermentation pathways. An example of B12-dependent isomerases is methylmalonyl-CoA mutase (MCM), found in bacteria and humans. Together with methylmalonyl-CoA epimerase, the enzyme is involved in converting propionate to succinate through the methylmalonyl-CoA pathway. Here, the enzyme catalyzes the reversible isomerization of l-methylmalonyl-CoA to succinyl-CoA using AdoCbl as a cofactor. Another common B12-dependent isomerase is ribonucleotide reductase (NrdZ). The enzyme catalyzes the conversion of ribonucleotides to deoxyribonucleotides for DNA replication and repair. AdoCbl adenosyl ribose is required to allow hydrogen transfer to the catalytic thiol group5. The B12-dependent methyltransferases play an important role in amino acid metabolism and CO2 fixation in anaerobic microorganisms. The most extensively studied B12-dependent methyltransferase is methionine synthase (MetH). This enzyme is responsible for the regeneration of methionine from homocysteine via the vitamin B12-dependent pathway and is involved in the folate pathway (Fig. 2). The methyl group of methylcobalamin is transferred to homocysteine forming methionine6. Vitamin B12-dependent dehalogenases are present in anaerobic bacteria. Reductive dehalogenases have a vital role in detoxifying aromatic and aliphatic chlorinated organic compounds7.

Figure 2
figure 2

Synthesis of methionine in M. tuberculosis. In the presence of vitamin B12 the riboswitch represses the translation of mRNA of metE. Cobalamin binds as a cofactor to MetH protein, and the latter provides methionine necessary for the cell. In turn, in the absence of vitamin B12, MetH is not functional. metE transcripts are efficiently transcribed to MetE protein, which provides methionine.

Full size image

Riboswitches were first detected in E. coli (Nou and Kadner, 1998). They are metabolite binding domains in specific mRNAs. Although riboswitches were mostly identified in prokaryotes, they are also present in higher organisms. They react to changes in the environment, such as changes in temperature, pH, or cofactor presence. Ligand binding allows for allosteric rearrangement of the mRNA and results in post-transcriptional control of gene expression. Cobalamin riboswitches repress gene expression by binding the ligand and preventing the mRNA's binding to ribosomes8.

The genus Mycobacterium accommodates bacterial species that carry genes presumably involved in the synthesis of vitamin B12. Mycobacteria are split into five phylogenetic clades, namely “tuberculosis-simiae”, “terrae”, “triviale”, “fortuitum-vaccae”, and “abscessus-chelonae”. There are several important pathogens in these groups. The most infamous one is M. tuberculosis, a causative agent of tuberculosis. This is an obligatory intracellular human pathogen with a complex life cycle. As shown previously, nearly all genes required for aerobic cobalamin synthesis are identifiable in M. tuberculosis, except for the cobF coding for precorin-6a synthase9.

Further research on M. tuberculosis confirmed the presence of two vitamin B12-dependent riboswitches in its genome, encoded by the Rv1133c and Rv0256c genes10. Rv1133c encodes a riboswitch regulating the metE gene expression of the cobalamin-independent methionine synthase (Fig. 2)11. The second cobalamin sensitive riboswitch at Rv0256c affects the PPE2-cobQ1-cobU operon. Rv0256c (PPE2) encodes a PPE2 family protein, while CobQ1 and CobU are presumably involved in vitamin B12 synthesis.

The identification of genes involved in the uptake of vitamin B12 from the environment in M. tuberculosis was performed by random mutagenesis12. Deletion of the Rv1819c gene, encoding ABC transporter BacA, abolished the ability to transport vitamin B12. Moreover, deletion of the bacA did not affect the infectivity of tubercle bacilli, albeit virulence was reduced during prolonged infection4.

MetH, MutB, and NrdZ are the three cobalamin-dependent proteins of M. tuberculosis. Studies involving these proteins cast doubt whether the reference strain of M. tuberculosis H37Rv synthesizes cobalamin. Savvi et al. showed that M. tuberculosis H37Rv could not use propionate as a carbon source by using the methylmalonyl pathway without enriching the medium in vitamin B1213. Warner et al. showed that B12 supplementation is necessary for the growth of the ΔmetE mutant of M. tuberculosis H37Rv, which also requires the MetH cobalamin-dependent enzyme11. Both of these reports suggest that M. tuberculosis cannot synthesize cobalamin and relies on cobalamin scavenged from the host. In turn, the M. tuberculosis clinical strain CDC1551 was surmised to be able to synthesize cobalamin by demonstrating a truncated variant of MetH. It depends on MetE to synthesize cobalamin. Therefore, it is hyper susceptible to sulfonamides, which block the folate cycle where MetH is involved in the salvage pathway. When M. tuberculosis CDC1551 was carrying MetH of H37Rv in trans, the authors observed that the strain presented standard susceptibility to sulfonamides. They suspected that intracellular levels of cobalamin in M. tuberculosis CDC1551 allow for activation of MetH14. In 2018, we showed that genes presumably involved in vitamin B12 synthesis and metabolism are under purifying selective pressure, suggesting functionality of pathway15. Ignatov et al. showed that genes involved in vitamin B12 synthesis upregulate their expression during mycobacterial persistence, reached by growing bacteria in a medium deprived of K+16. In summary, information regarding the possibility of vitamin B12 synthesis in M. tuberculosis is chaotic. As for other mycobacteria, information is scarce. Vitamin B12 synthesis was confirmed in Mycolicibacterium smegmatis17, 18. One manuscript, published in 1977, currently not available for a full read online, reported the presence of vitamin B12 in the cells of M. smegmatis, Mycolicibacterium fortuitum, Mycobacterium asiaticum, Mycobacterium phlei, and Mycobacterium bovis BCG using Lactobacillus leichmannii ATCC7830 tube method19.

Here we present the results of our investigation on the presence of vitamin B12 in mycobacteria. The purpose of this study was to describe vitamin B12 production in phylogenetic order of Mycobacterium. We show that vitamin B12 is present in the cells of various non-tuberculous species. Interestingly, we could not identify vitamin B12 in several strains of M. tuberculosis cultured under different growth conditions, even though the genes responsible for vitamin B12 synthesis are actively expressed.

Results and discussion

Gene expression of vitamin B12 synthesis genes

We aimed to identify the genes involved in vitamin B12 synthesis in NTM included in this study (Table 1). We used whole-genome sequencing data and its annotation found in the major bioinformatics databases. The available data provided an incomplete indication of loci involved in the vitamin B12 biosynthesis pathway, as it is for M. tuberculosis. The precision of annotation, covering the entire extent of variability of proteins serving particular functions, is still to be developed.

Table 1 Identification of genes involved in vitamin B12 metabolism in various mycobacteria species based on publicly available annotations in major databases.
Full size table

We used RNA-Seq data available at ENA Database to estimate gene expression through transcripts per million base pair (TPM) values for genes involved in vitamin B12 synthesis in M. tuberculosis, M. abscessus subsp. abscessus, and M. smegmatis (Table 2). TPM values inform about the level of basal transcription of genes, and are not to be confused with relative gene expression in different conditions. The average gene expression for M. abscessus and M. smegmatis was 201.88 ± 547.4 TPM and 147.33 ± 607.04 TPM, respectively. In comparison, the average expression of genes predicted to be involved in vitamin B12 synthesis was 94.943 ± 9.483 TPM and 76.669 ± 29.645 TPM, respectively. For M. tuberculosis we investigated gene expression level in cells grown in rich broth20, in medium supplemented with cholesterol21, and in human macrophages22. The above conditions' average gene expression was 256.02 ± 551.112 TPM, 256.01 ± 764.53 TPM, and 256.02 ± 1039.71 TPM, respectively. Simultaneously, the average expression of genes predicted to be involved in vitamin B12 aerobic synthesis was lower, 114.114 ± 77.666 TPM, 54.189 ± 35.772 TPM, 145.871 ± 159.664 TPM, respectively. Their overall expression level was comparable to DnaG primase, an essential protein involved in DNA replication (104.986 ± 2.321, 91.056 ± 42.023, and 118.236 ± 98.324, respectively)23.

Table 2 Expression of genes involved in cobalamin metabolism in M. tuberculosis H37Rv based on RNA-Seq data.
Full size table

Studies with Propionibacterium sp. showed the crucial role of cobA gene in regulating the level of synthesis of vitamin B12. Vitamin B12 was shown to regulate the cobA operon through a riboswitch in its 5′ untranslated region (5′ UTR)24. Similarly, M. tuberculosis contains a PPE2-cobQ1-cobU operon, containing vitamin B12 synthesis genes and controlled by a riboswitch. Taken the ubiquity of vitamin B12 riboswitches across Prokarytotes, the mechanisms where the level of vitamin B12 synthesis genes seem to be controlled by the synthesis product might be common25. Presented results show that cobQ1 and cobU of M. tuberculosis are actively expressed in a rich broth and in the presence of cholesterol. Expression of cobQ1 was not observed in macrophages. The level of reading coverage of the mycobacterial genome is relatively low. We assume that the low coverage results from natural technical difficulties of isolating mycobacterial RNA from the Eukaryotic cells that have RNA of their own26. Since reads of cobU are present, we suspect that the absence of cobQ1 reads in macrophages is due to too low coverage.

Vitamin B12 concentration in non-tuberculous mycobacteria

We measured the concentration of vitamin B12 per mg of protein in cell lysates obtained from 7H9 medium cultures of various species of NTM spread across the phylogenetic tree (Fig. 3A)27. On average, mycobacterial cells contained 33.044 ng of cobalamin per mg of protein. The median level of vitamin B12 across analyzed cells was 29.217 ng per mg of protein. The lowest concentration of vitamin B12 was detected for M. innocens (3.704 ± 0.643 ng/mg of protein). The highest concentration of vitamin B12 was detected in M. attenuatum (90.211 ± 13,769 ng/mg of protein). Results regarding relatively high production of vitamin B12 in M. phlei (77.712 ± 10.597 ng/mg of protein), when compared with other species of mycobacteria, are in line with previous findings from 197719. When vitamin B12 concentration was normalized to protein content, we detected a higher concentration of vitamin B12 in mycobacteria than it was previously detected in P. aeruginosa. There, analyses by HPLC–MS detected from 0.32 to 3.72 ng of vitamin B12 per mg of protein, depending on culture conditions and strain28.

Figure 3
figure 3

Cobalamin concentration in cell lysates of non-tuberculous mycobacteria. Cobalamin was detected in cell lysates of non-tuberculous mycobacteria by immunoassay. Cells were grown in 7H9 medium supplemented with OADC, Tween 80, and CoCl2. We cultured cells until the suspension reached OD600 = 1. Next, cells were harvested and washed with fresh medium without supplements to remove residual medium proteins from the surface. The pellet was re-suspended in Tris buffer and disrupted by beat-beating. The suspension was spinned. We used supernatant to estimate the concentration of vitamin B12 and protein content. Results were obtained from three independent cultures; each lysate was analyzed in two technical replicates. (A) cobalamin concentration in cell lysates of non-tuberculous mycobacteria, when normalized in ng per mg of protein, (B) cobalamin concentration in non-tuberculous mycobacteria cell lysates, when normalized in ng per ml of culture, (C) cobalamin concentration in cell lysates of clinical strains of M. abscessus complex. Whiskers represent SD.

Full size image

There are different approaches to the normalization of vitamin B12 concentration in bacteria. To further compare our results with other bacterial species, we also normalized our data regarding vitamin B12 concentration to ml of culture (Fig. 3B). When calculated in such a way, we obtained from 0.049 to 1.2 ng of vitamin B12 per ml of culture. In comparison, Pseudomonas freudenreichii produced from 20 to 125 ng of vitamin B12/ ml of culture, depending on culture conditions29. B. megaterium produced from 0.26 ng/ml of culture to 204 ng/ml of culture, also depending on the culture conditions. Due to relatively low concentration of vitamin B12, expensive growth media, long culture time, and difficulties to disrupt the cells, we conclude that mycobacteria are not attractive alternative producers of vitamin B12 at the industrial scale.

Importantly, we show that NTM can produce vitamin B12, and synthesis is shared across the phylogenetic tree. The sensitivity of the immunoassay detection was suitable for the detection of vitamin B12 in mycobacterial cells. This observation is an important reference point for results obtained for M. tuberculosis.

The level of vitamin B12 concentration in the NTM cells is variable, and it depends on the cell line

M. abscessus complex is a group of non-tuberculous mycobacteria. It is an emerging human pathogen often associated with the infection of cystic fibrosis patients. It consists of three subspecies M. abscessus subsp. abscessus, M. abscessus subsp. massiliense and M. abscessus subsp. bolletii. We measured the concentration of vitamin B12 per mg of protein in cell lysates obtained from 7H9 medium cultures of various clinical strains of M. abscessus subsp. abscessus and M. abscessus subsp. bolletii (Fig. 3C). We detected vitamin B12 in cells of all of the analyzed strains. On average, cells contained 19.842 ng of cobalamin per mg of protein. The median level of vitamin B12 across analyzed cells was 18.121 ng per mg of protein. The lowest concentration of vitamin B12 was detected for M. abscessus subsp. abscessus strain A5 (14.861 ± 1.848). The highest concentration of vitamin B12 was detected in M. abscessus subsp. abscessus strain A7 (31.582 ± 1.071). The difference in concentration between the highest and the lowest producing strain was statistically significant (p < 0.01, t = 11.07, df = 3).

We observed up to a twofold difference in the level of vitamin B12 synthesis across distinct strains of the same species. Strain variability in cobalamin concentration was observed previously in Pseudomonas aeruginosa, where the concentration of vitamin B12 ranged from 0.84 to 3.72, hence changed four-fold, depending on a strain28. Possible sources of the variability in the production of vitamin B12 in different strains are mutations either in the promoter regions of genes involved in the synthesis or directly in coding sequences, resulting in enzymes with altered reaction rates30.

Strain variability in the level of vitamin B12 production is important in the context of M. tuberculosis. Data presented in previous manuscripts suggested indirectly that certain strains of M. tuberculosis may be capable of cobalamin synthesis, while others are probably not14, 31. As in other species, mycobacteria do show a certain spread in the level of vitamin B12 synthesis that probably can be attributed to the genetic background rather than the environmental factors or stage of the growth.

Increased concentration of vitamin B12 in mycobacterial cells under starvation results from accumulation rather than increased production

In our previous study, we showed that cells of M. smegmatis grown in a medium deprived of nutrients contain an approximately eightfold amount of vitamin B12 when compared to cultures grown in a rich broth. An increase in vitamin B12 concentration was also observed in stationary phase cultures17. A similar observation was made in P. aeruginosa. There, vitamin B12 concentration increased from non-detectable during exponential growth to 0.32–0.67 ng/mg of protein in stationary phase cultures, depending on a strain. The concentration further increased up to 3.72 in conditions of continuous-flow growth28.

Here, we show that the reason behind the increased concentration of vitamin B12 in starved cells of M. smegmatis mc2 probably results from accumulation rather than increased synthesis. We estimated the relative gene expression of genes involved in cobalamin synthesis in starved cells compared to cells in the logarithmic phase (Fig. 4). We observed that the expression of genes involved in vitamin B12 synthesis was either constitutive (0 to 1-fold change in relative expression to sigA) for cobG, cobL, cobO and cobD or repressed (> 3-fold change) for cobU and cobN.

Figure 4
figure 4

Relative gene expression of genes involved in cobalamin biosynthesis in M. smegmatis. Data across samples was normalized to sigA. Bars represent log fold change. Whiskers represent SD.

Full size image

Accumulation of vitamin B12 in starved cells and old cultures of M. smegmatis is important in the context of cobalamin detection in M. tuberculosis. M. smegmatis is a model organism for studying the biology of mycobacteria, including M. tuberculosis32. Bacteria of the same phylogenetic order are likely to maintain the same biological pathways and mechanisms. Indeed, increased expression of cobalamin synthesis genes was reported in dormant cultures of M. tuberculosis16. Therefore, if cobalamin was to be present in the cells of M. tuberculosis, it was more likely to be identified in prolonged, starved, or dormant cultures.

Lack of observable vitamin B12 production in M. tuberculosis

We tested the contents of cells of M. tuberculosis for vitamin B12 by immunoassay (Table 3). We included laboratory strain of M. tuberculosis H37Rv and five clinical strains of M. tuberculosis, here grouped into a group of “clinical strains”. As a negative control strain, we used M. tuberculosis deficient in cobIJ gene. Predicted function of cobIJ is precorrin-2 C20-methyltransferase/precorrin-3B C17-methyltransferase. The gene product is required at the early stage of vitamin B12 synthesis (Fig. 1).

Table 3 Types of cultures tested for cobalamin in cells of M. tuberculosis H37Rv, five clinical strains of M. tuberculosis, and ∆cobIJ.
Full size table

We screened M. tuberculosis cell lysates derived from cultures grown in various conditions. The growth conditions aimed to mimic the environments that can be found during M. tuberculosis infection cycle. All cultures were supplemented with cobalt to evade the blockade of synthesis due to insufficient cobalt concentration. First, we investigated the possibility of de novo synthesis of vitamin B12. We tested logarithmic phase cultures, stationary phase cultures, and acidified cultures that would mimic the infection's active stage. For granuloma conditions, we tested starved cultures, persister cultures, and hypoxic cultures. ELISA immunoassay detected less than one ng of vitamin B12 per one ml of lysate in all of the samples. The samples were considered negative for vitamin B12 based on the cut-off value of the sensitivity of the test. Taken that vitamin B12 tends to accumulate in the cells during prolonged growth, our results suggest that it is unlikely that there is an ongoing de novo synthesis of vitamin B12 inside M. tuberculosis cells.

Next, we wanted to see if M. tuberculosis might rely on substances widely present in the host to produce vitamin B12. We supplemented the growth medium with uroporphyrinogen III, which is a precursor of heme in the human body and a precursor of vitamin B12 in bacteria (Fig. 1). We tested cell lysates from logarithmic phase cultures, stationary phase cultures, persister cell cultures, and hypoxic cultures. ELISA immunoassay detected less than one ng of vitamin B12 per one ml of lysate in all of the samples. Hence, the samples were considered negative for vitamin B12 based on the cut-off value of the test's sensitivity.

M. tuberculosis metE promoter responds to vitamin B12 concentration present in the host

We used the GFP reporter system to see if the cells of M. tuberculosis could respond to vitamin B12 concentration found within the host (Fig. 5). Vitamin B12 concentration in the human body is between 0.2 and 0.9 μg/ml. We constructed a series of mutants, H37Rv::attB + rsB12, ∆bacA::attB + rsB12, ∆cobIJ::attB + rsB12, carrying the gene of GFP under the control of the metE promoter, controlled by a vitamin B12-dependent riboswitch. In our model, the presence of vitamin B12 in the cells prevents translation of gfp transcript, which results in diminished fluorescence. Of note, distinct clones of the same cell lines showed a different level of basal fluorescence without supplementation of medium without vitamin B12. Therefore, the fluorescence level could not be reliably compared between different cell lines due to the distinct basal expression of GFP in the clones. However, green fluorescence levels could be relatively compared within one clone of the cell line when considering different concentrations of vitamin B12 in the growth medium. We tested various concentrations of vitamin B12. We observed that supplementation of the growth medium with vitamin B12 gradually diminished gene expression of the green fluorescence protein of M. tuberculosis H37Rv and ∆cobIJ from 100 to 23.93% and 23.70%, respectively. In turn, the green fluorescence expression of ∆bacA was not affected, and it remained constant at approximately 100%. Similarly, the autofluorescence level was constant for the control strain M. tuberculosis H37Rv, which lacked the reporter system. M. tuberculosis H37Rv GFP expression diminished to 70.29% in the presence of 0.5 μg/ml of vitamin B12 (p = 0.01, t = 6.64, df = 3). Hence, M. tuberculosis metE promoter is responsive to vitamin B12 concentration found in the human body. Our results confirm the role of BacA as the transporter of vitamin B1233. Further, our results indirectly confirm the lack of vitamin B12 production in M. tuberculosis H37Rv, because the wild type strain and the knock-out strain similarly showed a decrease in fluorescence corresponding to increasing concentration of supplemented vitamin B12.

Figure 5
figure 5

The evaluation of the protein expression of gfp gene under the control of vitamin B12- dependent riboswitch, based on the green fluorescence. We used a Guava flow cytometer to evaluate the green fluorescence in cells of H37Rv::attB + rsB12, ∆bacA::attB + rsB12, ∆cobIJ::attB + rsB12, carrying the gene of GFP under the control of the metE promoter, controlled by a vitamin B12-dependent riboswitch. We observed that the green fluorescence of strains H37Rv::attB + rsB12 and ∆cobIJ::attB + rsB12 significantly differed from the control strains grown in medium without exogenous vitamin B12 supplementation. Each strain was analyzed in samples collected from three cultures. Statistical analysis was performed with paired t-test. Whiskers represent SD.

Full size image

Previous reports suggested that the ability to synthesize vitamin B12 by M. tuberculosis was restricted in M. cannetti like ancestor9, 10. M. tuberculosis is an obligate pathogen with possible access to vitamin B12 from the host. It is, therefore, possible that the genes involved in vitamin B12 synthesis in the genome of M. tuberculosis are remnants from a more independent ancestor. The genes of the vitamin B12 biosynthesis pathway in Mycobacterium leprae, another obligate pathogen of Mycobacterium genus, evolved into pseudogenes10. The most probable explanation is that M. tuberculosis does not synthesize vitamin B12 anymore. However, there was still not enough time since the abrogation of the pathway to accumulate mutations that would entirely degrade the pathway, impede the expression of genes and convert them into pseudogenes. It seems that the disruption of the pathway might have taken place relatively recently, as cobF encoding region was found in two M. tuberculosis strains found in the African Great Lakes region, representing Lineage 8 of M. tuberculosis complex31. To be precise, it cannot be excluded that M. tuberculosis does synthesize vitamin B12, but the level of vitamin concentration is undetectable by the immunoassay we used. Finally, it remains to be established whether M. tuberculosis is able to synthesize cobamides other than vitamin B1234.

Summary

We conclude that mycobacteria are generally capable of vitamin B12 synthesis, with the likely exception of M. tuberculosis. Our results are direct evidence of vitamin B12 production in these clinically important group of bacteria.

Materials and methods

Bacterial strains

We analyzed the level of vitamin B12 in several type strains and clinical strains of non-tuberculous mycobacteria (Table 4). Mycolicibacterium porcinum, M. fortuitum, and Mycobacteroides abscessus complex were isolated and differentiated in Canada35. Further, we included a laboratory strain of M. tuberculosis H37Rv, its genetically modified derivatives ΔcobIJ and ΔbacA, and five clinical strains of M. tuberculosis isolated in Lodz, Poland, between 2006 and 2008. Each strain belonged to a different clade. We chose strains: 321 (spoligotype 35, clade H4), 404 (spoligotype 46, clade U (likely H)), 663 (spoligotype 50, clade H3), 216/8 (spoligotype 42, clade LAM9) and 218/8 (spoligotype 1253, clade S)36. We used Escherichia coli Top10 for cloning.

Table 4 List of strains used in this study.
Full size table

Bacterial cultures

E. coli Top10

Bacteria were cultured at 37 °C for 18–20 h in liquid or solid Luria–Bertani broth. Where necessary, the media were supplemented with antibiotics or other supplements at the following concentrations: kanamycin (Bioshop) 50 μg/ml; ampicillin (Bioshop) 100 μg/ml, X-gal 40 μg/ml (BioShop), sucrose 2% (Sigma Aldrich) at 37 °C.

Non-tuberculous and tuberculous mycobacteria

Where necessary, media were supplemented with 10% oleic acid albumin dextrose catalase growth supplement (OADC) (Becton–Dickinson), 0.05% Tween 80 (Sigma), tyloxapol 0.015% (Sigma-Aldrich), cobalt chloride 12 μg/ml (Sigma Aldrich), kanamycin 25 μg/ml (BioShop); X-gal 40 μg/ml (BioShop), sucrose 2% (Sigma Aldrich), vitamin B12 (adenosylcobalamin) 10 μg/ml (Sigma Aldrich), uroporfirynogen III octamethyl ester 1 μg/ml (Sigma Aldrich) (dissolved in 25% DMSO), OAD (0.05% Oleic Acid, 5% bovine serum albumin, fraction V, 2% glucose, and 0.85% NaCl). All cultures were started at OD600 = 0.1.

Unless stated otherwise, cultures and seed cultures of non-tuberculous mycobacteria and M. tuberculosis were cultivated in 7H9 broth supplemented with OADC, Tween 80, and cobalt chloride. Cultures of non-tuberculous mycobacteria were started at OD600 = 0.05 and carried out until they reached OD600 = 1. M. tuberculosis cultures were started at OD600 = 0.1. For seeding, the appropriate amount of logarithmic phase culture (OD600 = 0.8) was spinned down, washed in fresh medium, spinned again, and re-suspended in fresh medium.

For testing vitamin B12 concentration, non-tuberculous mycobacteria were cultured in 7H9 broth supplemented with OADC, Tween 80, and cobalt chloride until they reached OD600 = 1. Starved cultures of M. smegmatis were cultured as described previously17.

We tested several types of cultures of M. tuberculosis for vitamin B12 concentration. Logarithmic phase cultures of M. tuberculosis were grown in 7H9 medium supplemented with OADC, Tween80, and cobalt chloride until they reached OD600 = 0.8. Stationary phase cultures were carried out in the same medium, and they were collected after 15 days of culture. Acidified cultures (pH 5.7) were carried out in 7H9 broth, supplemented with 5% bovine serum albumin, fraction V, 2% glucose, and 0.85% NaCl, Tween 80, and cobalt chloride. They were collected after 1 week of culture. Starved cultures were carried out in 7H9 broth supplemented with Tween 80, 0.5% glycerol, and cobalt chloride. They were collected after one week of culture. Hypoxic cultures were carried out as previously described37. In brief, starter cultures of M. tuberculosis grown on Dubos medium supplemented with OAD and cobalt chloride were tightly locked in flasks with 0.25 headspace ratio and cultured at 37 °C on a shaker for six weeks. Catalase is not recommended for use in hypoxia experiments because it influences redox balance, and redox stress is an important stress factor of hypoxia38. Methylene blue was added to control cultures as an indicator of oxygen depletion. After this time, the flasks were opened, cultures were spinned down and washed three times with 7H9 medium. A sample of the culture was plated as viability control, while the rest was lyzed. For persister cultures, we used a medium deprived of K+, as previously described16. Here, starter cultures grown on Sauton medium were spinned down and re-suspended in Sauton medium deficient in K+ supplemented with cobalt chloride. After two weeks, rifampicin was added to cultures at 5 μg ml−1 and the culture continued for the next four weeks at 37 °C.

Cloning strategy

All molecular cloning was performed in E. coli T10. Knock-out mutants of mycobacteria were obtained by the method of gene replacement through homologous recombination (Tables 5, 6; Fig. 6; Supplementary Figs. 1 and 2)39. Briefly, sequences flanking desired deletion were amplified by PCR. We used AccuPrime Pfx High Fidelity Polymerase (Invitrogen), and genomic DNA of M. tuberculosis H37Rv for the reaction. PCR products were introduced into pJET1.2 plasmid (Thermo Fisher Scientific) and sequenced. Following confirmation of cloning of proper sequence, we cut out flanking sequences using restriction enzymes, and sequentially introduced them into p2NIL plasmid, together with marker genes from pGOAL17. Plasmids were transformed into M. tuberculosis H37Rv thru electroporation. The cells underwent gene replacement by allelic exchange as described previously39. Similarly, episome plasmid containing green fluorescence protein (GFP) gene under the control of the riboswitch of M. tuberculosis metE gene was constructed with a similar procedure. We started with PCR amplification of products on genomic DNA of M. tuberculosis H37Rv and pJAM plasmid carrying gfp. Subsequently, we introduced sequences to pJET1.2, and we confirmed proper cloning by sequencing. Next, we used restriction digestion to cut out the sequences, and we introduced them into pMV306 episome plasmid. M. tuberculosis H37Rv and its derivative strains were transformed by electroporation.

Table 5 List of primers used in this study.
Full size table
Table 6 List of plasmids used in this study.
Full size table
Figure 6
figure 6

Southern blots confirming deletion of genes cobIJ and bacA in M. tuberculosis H37Rv. We used a gene replacement method through homologous recombination to obtain unmarked genetic mutants with large deletions inside the genes. Single cross-over (SCO) describes an intermediate step of mutagenesis. The images are cropped, hence altered lane numbering. Full-size images can be found in the supplementary data.

Full size image

Vitamin B12 ELISA

Bacterial cultures were spinned down, washed with fresh medium without supplements, spinned down again, and re-suspended in 0.01 M TRIS pH 7.5. The mixture was transferred to disruptor eppendorfs. Cells were disrupted twice using the MP disruptor system with the Quick prep adapter (MP Biomedicals) and 0.1 mm silica spheres (45 s, 6.0 m/s with 5 min intervals). Samples were spinned down, and lysates were transferred to new eppendorfs. As a principal, we normalized the results regarding vitamin B12 concentration to protein concentration in cell lysates. We wanted to avoid errors resulting from a different level of difficulty to disrupt mycobacterial cells of different species. Protein concentration in lysates was measured using Bradford reagent (BioShop) and estimated with a standard curve. In order to achieve a sufficient detection limit of vitamin B12, we only used the lysates that contained at least 0.5 mg of protein per ml, preferably between 1 and 2 mg of protein per ml.

Vitamin B12 ELISA (Demeditec) was performed according to manufacturer instructions. The test is based on the principle of the competitive enzyme-linked immunosorbent assay. The surface of a microtiter plate was covered with an antibody directed against vitamin B12 by the manufacturer. Samples and standards were mixed with a vitamin B12-peroxidase conjugate in the wells of the microtiter plate. Both enzyme-labeled and free vitamin B12 competed for the antibody binding sites. After one hour of incubation at room temperature, the wells were washed to remove the unbound material. A substrate solution was added, resulting in the development of a blue color. The color development was inhibited by the addition of a stop solution, and the color turned yellow. The yellow color was measured photometrically at 450 nm. The concentration of vitamin B12 was indirectly proportional to the color intensity of the test sample.

For each condition, we analyzed lysates from three independent cultures. Each sample was analyzed in duplicate wells, as recommended by the producer of the immunoassay. The minimum detection level of vitamin B12 was settled at 1 ng/ml based on our previous observations17, and all samples below this level were considered negative for vitamin B12. For purposes of enabling comparison with the results obtained from other species of bacteria, our results were also normalized to ml of culture.

Flow cytometer analysis

Samples of cultures were analyzed on the flow cytometer Guava EasyCyte Flow Cytometer with High Power Blue Laser (Merck) suitable for detection of bacteria. Unstained control samples were diluted to reach a concentration of 400–800 cells/μl. Cell suspensions were first run through the flow cytometer to set a population gate around the bacteria by using the forward-scatter versus side-scatter parameters. Next, the voltages in the green fluorescence channel were adjusted so that the fluorescence histogram of the unstained bacteria appeared within the first compartments of the logarithmic scale of fluorescence. Ten thousand events were collected at a set standard low event rate. We used Guava software to analyze the acquired data. For each strain, we analyzed data from three cultures.

qPCR

Bacterial cultures were spinned down, re-suspended in water, and three volumes of TriReagent was added (Bioshop). The mixture was transferred to disruptor eppendorfs. Cells were disrupted twice using the MP disruptor system with the Quick prep adapter (MP Biomedicals) and 0.1 mm silica spheres (45 s, 6.0 m/s with 5 min intervals). Samples were spinned down, and the supernatant was transferred to new eppendorfs. One volume of chloroform was added, samples were vigorously mixed and spinned down. The top phase was transferred to new eppendorfs and precipitated with 1 volume of isopropanol and 1/10 volume of sodium acetate. Following precipitation, samples were re-suspended in water and digested with Turbo DNase I (Invitrogen by Thermo Fisher Scientific) following the manufacturer's instructions. The RNA quantity was assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). cDNA was obtained using SuperScript III First-Strand Synthesis Super Mix kit with random hexamers (Invitrogen). qPCR was performed using SG qPCR Master Mix (2×), plus ROX Solution (Eurx), and synthetic primers (Table 5) on a 7900HT real-time PCR system (Applied Biosystems). Real-time PCR conditions were as follows: initial activation at 95 °C for 10 min, followed by 40 cycles at 94 °C for 15 s (denaturation), 62 °C for 30 s (annealing), 72 °C for 30 s (extension). The melting curve analysis was performed at the end of each qPCR reaction to verify a single, specific product was generated. The threshold cycle (CT) value for each studied gene was normalized to the expression of msmeg_2758 (sigA) (ΔCT) and converted to linear form (2 − ΔCT). The RNA samples for each strain were isolated from three independently grown cultures. Each sample for qPCR was run in triplicate.

RNA Seq

Raw RNA Seq reads were downloaded from European Nucleotide Archive Database (ENA). We analyzed gene expression of M. abscessus subsp. abscessus grown in 7H9 medium supplemented with OADC40, M. smegmatis grown in 7H9 medium with glucose41, and three experiments performed with M. tuberculosis H37Rv grown in 7H9 broth supplemented with OADC20, in a medium supplemented with cholesterol as a sole carbon source21 and in human THP-1 derived macrophages three days post-infection. Each experiment contained data for three replicates. Raw sequences were uploaded and processed with Geneious Prime 2021 (Biomatters, New Zealand). Reads were mapped to M. tuberculosis H37Rv accession number NC_000962 using Bowtie2 Geneious plug-in42. Gene expression analysis, through estimation of transcripts per kilobase million (TPM), was performed with Geneious.

Identification of loci in the whole genome sequencing data based on the genome annotation

We identified genes involved in vitamin B12 metabolism in the following strains: M. tuberculosis H37Rv (NC_000962), M. abscessus subsp. abscessus ATCC19977 (NC_010397), M. abscessus subsp. bolletii FLAC 003 (CP014950), M. conspicuum JCM 14738 (GCA_010730195), M. fortuitum CT6 (CP011269), M. gastri DSM 43505 (LQOX1000000), M. gordonae 24T (CP059165), M. innocens MK13 (LS999933), M. kansasii ATCC 12478 (GCA_000157895.1), M. persicum H48 (GCA_002705835), M. phlei CCUG 21000 (GCA_001582015), M. porcinum ACS 3670 (NZ_MBDY01000007.1), M. terrae NCTC 10856 (GCA_900187145), M. xenopi RIVM700366 (NZ_AJFI01000095.1), M. szulgai DSM 44166 (NZ_LQPW01000016.1) and M. smegmatis mc2 155 (CP009494). We screened the following databases: National Center of Biotechnology Information, Nucleotide and Protein (NCBI), Mycobrowser, STRING, UniProt, and we manually screened the sequences thru Geneious Prime (Biomatters, New Zealand).

Statistical analysis

Statistical analysis was performed with Develve Statistical Software, with paired t-test. The level of statistical significance was p < 0.05. All results are reported as the means ± SD unless otherwise stated.