Engineering Mycobacterium smegmatis to produce oxygenic mycolic acids of Mycobacterium tuberculosis

Introduction

Sepsis is a life-threatening condition that arises from the uncontrolled spread of infection, leading to organ damage and ultimately to death in the absence of prompt medical attention. In 2020, the WHO reported nearly 50 million cases and 11 million sepsis-related deaths, making it responsible for 20% of global mortality (WHO and Sepsis, 2024). In addition, the survivors may face long-term complications because of innate immune exhaustion, a dysfunctional state of innate immune cells (Kisha Pradhan et al., 2021). Although vaccines have been used in the prevention of severe sepsis, their effectiveness is limited in addressing acute sepsis and its associated immune dysfunction. Thus, there is an urgent need of therapeutic approaches that would modulate immune function under septic conditions and in this context Mycobacterium tuberculosis var. bovis Bacille Calmette-Guerin (BCG) has attracted attention, as M. tuberculosis in known to dampen human immune response (Pallavi Chandra et al., 2022; Haoran Wang, 2023) and the BCG vaccine seems to have beneficial impacts beyond preventing tuberculosis (Gong et al., 2022; Wang et al., 2022; Pi et al., 2024; Pittet et al., 2024). There are studies that show correlation between reduced disease severity of COVID-19 and BCG vaccination history (Gong et al., 2022; Wang et al., 2022) and further analysis has shown that while this correlation is not valid for COVID-19, there are other viral diseases where it is applicable (Wang et al., 2022; Pi et al., 2024; Pittet et al., 2024). All these apparent protections have been linked to the concept of trained immunity, where, with BCG vaccination, innate immune cells develop a memory-like state leading to “rewiring” of the innate immune memory that helps tackle future infections (Chen et al., 2023; Melvin Joy et al., 2021; Gong and Wu, 2021).

The recent studies, that build on the above-mentioned observations, have focused on the possible use of trehalose 6,6′-dimycolate (TDM), cell wall component in all M. tuberculosis (M. tuberculosis, Mtb) strains including the M. tuberculosis var. bovis Bacille Calmette-Guerin (BCG), the vaccine strain, in treating sepsis (Haoran Wang, 2023; Wu et al., 2024; Decout et al., 2017; Harland et al., 2008). TDM, also known as cord factor, is composed of a core trehalose sugar molecule with two long carbon chains called mycolic acids (MAs). It functions as an adjuvant and shows promise in enhancing immune response, potentially helping to restore homeostasis of the innate immune system (Wu et al., 2024). MAs, β-hydroxy fatty acids with long alpha-alkyl side chain, form the backbone of TDM and are an essential structural component of the mycobacterial cell wall. These compounds are found esterified into either arabinogalactan or trehalose of mycobacterial cell wall as well as free lipids within the cell wall (Haoran Wang, 2023; PaweŁczyk and Kremer, 2014). Given the TDM’s above-mentioned activities, MAs represent themselves as candidates to be explored for their abilities to restore immune homeostasis (Wu et al., 2024). MAs exist in six different forms depending on the carbon chain length and oxygenation states and these are called α-, α’-, hydroxy-, methoxy-, epoxy- and keto- MAs (Figure 1). Of these α-, methoxy-, and keto-MAs are found in BCG, although some of the strains derived from the original vaccine stock lack methoxy-MAs due to a point mutation in a specific methyl transferase gene, mma3 (Behr et al., 2000; Fatima et al., 2024; Fujita et al., 2005; Szewczyk Rafał et al., 2025).

Thus, there is a need for a facile system for the generation of milligram amounts of MAs produced by BCG. However, a rather slow growth rate (doubling time, 16–24 h; (Ohara, 2012)) makes it harder to generate the needed materials in large amounts from this organism. On the other hand, Mycobacterium smegmatis mc²155 (Msm), which is not pathogenic, grows relatively fast (doubling time, 3–4 h; (Wang et al., 2021)), and is genetically tractable, has the potential to remove above-mentioned roadblocks. It carries most of the genes for MA synthesis (producing α-, and α’-, epoxy-MAs) and an introduction of a few specific Mtb genes into it causes the recombinants to produce hydroxy-, methoxy-, and keto-MAs (Figure 2; (Purwantini and Mukhopadhyay, 2013; Fatima et al., 2024; Melvin Joy et al., 2021)). We report how this system has helped us to generate milligram amounts of hydroxy-, methoxy-, and keto-MAs from laboratory scale cultures.

Methods

DNA, bacteria and growth conditions

The oligonucleotides, plasmids, and bacteria used in this study are described in Tables 1–3 respectively. E. coli NEB 10-beta was obtained from New England Biolabs (NEB; Ipswich, MA). Since, mycolic acid oxygenation genes from BCG and M. tuberculosis H37Rv (Mtb) are either identical at the DNA sequence level or encode proteins of identical primary structures, Mtb chromosomal DNA was used as the template for the amplification of genes of interest towards the construction of targeted recombinant M. smegmatis (M. smegmatis, Msm) strains. The Mtb chromosomal DNA was obtained from National Institute of Health’s TB Vaccine Testing and Research Materials Contract (TBVTRMC) facility at Colorado State University, and M. smegmatis mc²155 was a gift from Prof. W. Jacobs, Jr, Albert Einstein College of Medicine (New York, NY). All E. coli strains were grown in lysogeny broth (LB) (Bertani, 2004) and on LB solid medium (agar, 1.5–2%) for 16 h at 37 °C. All M. smegmatis strains were grown in liquid and on solid (with agar, 1.5–2%) Middlebrook 7H9 medium (DUBOS and Middlebrook, 1947) supplemented with glycerol at the final concentration of 0.5% for 3 days at 37 °C; liquid medium contained Tween 80 (final concentration, 0.05%). E. coli and Msm strains carrying pSMT3-based plasmids were selected on hygromycin (100 μg/mL).

TABLE 1. Oligonucleotides.

TABLE 2. Plasmids.

TABLE 3. Bacterial strains.

Construction of M. tuberculosis gene expression vectors and bacterial strains

The plasmids were based on pSMT3, a mycobacteria-E. coli shuttle vector that confers hygromycin resistance to these organisms (Purwantini and Mukhopadhyay, 2013; Garbe et al., 1994), and pEP-rv0132c which was previously developed by cloning Mtb fhmad gene along with its native promoter into pSMT3 (Purwantini and Mukhopadhyay, 2013). The relevant DNA fragments were PCR-amplified from the Mtb chromosomal DNA using NEB Q5 High-Fidelity DNA polymerase with GC enhancer (NEB) and then combined with the pSMT3 or pEP-rv0132c backbone (generated via PCR) via the Gibson assembly method (Gibson et al., 2009) using the NEBuilder HiFi DNA Assembly kit (NEB). The following genomic regions of Mtb were amplified along with their promoters (encoded proteins): rv0645c-rv0642c (rv0645c, mycolic acid methyltransferase, MmaA1; rv0644c, cyclopropane mycolic acid synthase, MmaA2; rv0643c, methoxy mycolic acid synthase, MmaA3; rv0642c, hydroxy mycolate synthase, MmaA4) (Behr et al., 2000); rv0643c-v0642c (MmaA3 and MmaA4) (Behr et al., 2000); rv0132c (F₄₂₀-dependent hydroxy mycolic acid dehydrogenase, fHMAD, previously known as Fgd2) (Purwantini and Mukhopadhyay, 2013). Following are the details of the Mtb DNA elements that were assembled to obtain the expression vectors (starting plasmid, DNA element cloned, and resultant vector name): pSMT3, mmaA1-4 operon along with the respective promoter (P_mmaA1), and pUB-A1.2.3.4; pSMT3, mmaA3-4 operon along with the respective promoter (P_mmaA3), and pUB-A3.4; pEP-rv0132c (which contains P_fhmad and fhmad), mmaA1-4 operon along with the respective promoter (P_mmaA1), and pUB-f_A1.2.3.4; pEP-rv0132c, mmaA3-4 operon along with the respective promoter (P_mmaA3), and pUB-f_A3.4. The assembled plasmids were transformed into E. coli NEB 10-beta via electroporation (1.8 kV and 0.1 cm cuvette) using Electroporator 2,510 (Eppendorf North America; Hauppauge, NY) and were then purified from the resulting recombinant E. coli strains. After confirming its authenticity via sequencing to entirety at Plasmidsaurus Inc. (Louisville, KY 40213), each plasmid was introduced into M. smegmatis mc²155 via electroporation (2.5 kV and 0.2 cm cuvette) (Snapper et al., 1990). The maps of the plasmids that were constructed in the study are shown in Figure 2.

Preparation and analysis of mycolic acid methyl esters (MAMEs)

The MAMEs were generated as described previously (Figure 3) (Purwantini and Mukhopadhyay, 2013; Besra, 1998). In brief, M. smegmatis cells were pelleted from a liquid culture via centrifugation at 10,000 x g and lipids present therein were saponified by incubation of the cells at 110 °C with 15% tetrabutyl ammonium hydroxide (TBAH) overnight; the treatment broke the ester linkages between MAs and mycobacteria cell wall components. This step was followed by the addition of 5% iodomethane solution in dichloromethane into the mixture and shaking it at room temperature for 5 h to methylate the free MAs. The resulting dichloromethane layer that contained MAMEs were recovered and washed three times with equal volumes of water, 0.1 M HCl, and water, in sequence. Then the solution was dried under a nitrogen stream. The dried powder was dissolved in a toluene-acetonitrile mixture (2:1, v/v) and from this solution MAMEs were precipitated with an addition of acetonitrile that changed the toluene-acetonitrile ratio to 2:3 (v/v). The MAME pellet was recovered via centrifugation at 16,900 x g at 4 °C and dissolved in dichloromethane. This solution was then stored in −20 °C until further use.

The MAME solutions were analyzed for their composition via thin layer chromatography (TLC) on an aluminum-backed silica gel plate (10 × 10 cm, Merck 5735-silica gel 60F254) by eight developments with petroleum ether (80:20)-diethyl ether mixture (9:1) as described previously (Purwantini and Mukhopadhyay, 2013; Besra, 1998). The silica plate was sprayed with 5% molybdophosphoric acid in 100% ethanol and charred at 110 °C to visualize the bands of individual MAME groups.

Growth studies of M. smegmatis strains expressing M. tuberculosis mycolic acid oxygenation genes

The organisms were grown in 500 mL Erlenmeyer’s flasks with baffles in 100 mL 7H9 medium with glycerol, Tween 80 and hygromycin at concentrations described above. A 5 mL culture was developed in a 15 mL culture tube overnight in the same medium to an optical density at 600 nm (OD₆₀₀) of 1, as measured by using a Genesys™ 50 UV-visible spectrophotometer (ThermoFisher Scientific, Waltham, MA) and a cuvette with a 1 cm light path. This overnight culture was used to inoculate the 100 mL culture to provide a starting OD₆₀₀ value of 0.01, and the inoculated culture was incubated at 37 °C in a shaking incubator (Controlled environment incubator shaker, New Brunswick scientific co. Inc.; Edison, NJ) operating at 250 rpm. The growth was followed by withdrawing 800 mL of culture every 3 h and measuring the respective OD₆₀₀ value (Figure 4).

MALDI-TOF mass spectrometric analysis of mycolic acid methyl esters

The two terminal lanes of a silica plate of a TLC with resolved MAs were cut out and developed as described above to view the MAs bands. Then, using these lanes as guides the silica layers carrying individual MA groups were scrapped off and extracted 4-times with dichloromethane to obtain the respective MA solutions (Figure 6). These solutions were analyzed via MALDI-TOF mass spectrometry at the Virginia Tech Mass Spectrometry Incubator (VT-MSI) employing a Bruker timsTOF flex MALDI-2 instrument operating in a positive ion mode, using α-Cyano-4-hydroxycinnamic acid (10 mg/mL in 50:50 water with 0.1% formic acid) as the matrix over an m/z range of 800–4,000. The MALDI-2 post-ionization laser and TIMS ion mobility separation were not used in this study. The instrument is capable of mass measurements with accuracies below 3 ppm. Samples were analyzed using a 355 nm Bruker smartbeam 3D laser, with spectra acquired from 10,000 accumulated laser short per spot. External calibration was performed using peptide standards using the instrument’s electrospray ion source and ESI-L Low Concentration Turning Mix (Agilent Technologies; Santa Clara, CA).

Results and discussion

Design basis for our MA production platform

It has been shown that the introduction of Mtb mmaA1-mmaA4 operon (rv0642-0645c) or a respective subset, mmaA3-mmaA4 (rv0642-0643c), into M. smegmatis causes the resulting recombinant organism to produce methoxy-MAs (Purwantini and Mukhopadhyay, 2013) (Figure 1). This process proceeds through an intermediate formation of hydroxy-MAs, since the expression of mmaA4 (rv0642c), which is also called hma, yields hydroxy-MAs in a recombinant M. smegmatis (Purwantini and Mukhopadhyay, 2013). Similarly, these hydroxy-MAs are converted to keto-MAs when the fHmad (rv0132c) gene of Mtb is provided in addition to mmaA4 (rv0642c) (Purwantini and Mukhopadhyay, 2013) (Figures 1, 5). Previous studies have identified two promoters for the mmaA1-4 operon, one of which, located upstream of mmaA1 (P_mmaA1) (Barkan et al., 2010), controls the entire operon, and the other is internal (P_mmaA3) and drives the expression of mmaA3-4 (Belley et al., 2004) (Figure 2).

The multi-copy plasmid pSMT3 provides mycobacterial hsp60 promoter (P_hsp60) which is strong and constitutive (Purwantini and Mukhopadhyay, 2013; Garbe et al., 1994; Gaora, 1998). In an earlier work, we successfully expressed the fhmad gene in Msm by using a pSMT3-based replicable plasmid, pEP-rv0132c, that carries the respective coding sequence along with its 500 bp upstream region (called here P_fhmad) (Purwantini and Mukhopadhyay, 2013); P_fhmad provided sufficient promoter activity in the heterologous host (Purwantini and Mukhopadhyay, 2013). Accordingly, we used pSMT3 and pEP-rv0132c for the construction of expression vectors (Figure 2). For methoxy-MA production we have developed two derivatives of pSMT3 (Figure 2). In pUB-A1.2.3.4 the expression of the cloned mmaA1-4 operon was driven by a 300 bp upstream region of this operon that provided P_mmaA1 activity (Figure 2). For the other, pUB-A3.4, a promoter (P_mmaA3) represented by a 400 bp DNA element occurring upstream of mmaA3 was used to express mmaA3-4 genes (Figure 2). Similarly, two vectors were designed for producing keto-MAs. In pUB-f_A3.4 and pUB-f_A1.2.3.4, fhmad was under the control of P_fhmad, and these plasmids carried, in addition, the expression constructs for mmaA3-4 and mmaA1-4 as described above, respectively (Figure 2).

Methoxy-, hydroxy- and keto-MA production in recombinant M. smegmatis strains

The results from TLC analysis of mycolic acids methyl esters (MAMEs) showed that all four recombinant strains, carrying pUB-A3.4, pUB-f_A3.4, pUB-A1.2.3.4, and pUB-f_A1.2.3.4 (Figure 2), produced hydroxy-MAs (H-MAs) and methoxy-MAs (M-MAs) (Figure 5A, lanes 2–5). In contrast, the wild-type M. smegmatis mc²155, the control strain, exhibited only the native MA profile with α-, α’-, and epoxy-MAs (Figure 5A, lane 1). It is likely that the methoxy-MAs produced in M. smegmatis A1.2.3.4 and M. smegmatis f_A1.2.3.4 contained cyclopropanation because of the presence of mmaA1 and mmaA2 genes in the respective expression vectors (Figures 1, 2) (Barkan et al., 2010). The strains with pUB-f_A3.4 and pUB-f_A1.2.3.4 produced keto-MAs (K-MAs) (Figure 5A, lanes three and 5). In a TLC system used here, keto-MAs and epoxy-MAs exhibit similar but discernible migration patterns (Figure 5A) (Purwantini and Mukhopadhyay, 2013), and in the past via NMR analysis of materials recovered from the respective bands we have shown that epoxy-MAs are not present in a recombinant strain that produces keto-MAs (Purwantini and Mukhopadhyay, 2013).

Recovery of methoxy-, epoxy-, and keto-mycolic acids (MAs) from the recombinant strains and their authentication via mass spectrometry

The methoxy- and keto-MAs were recovered via scraping from the TLC plates where the MAMEs from M. smegmatis A3.4 and M. smegmatis f_A3.4 were resolved, respectively (Figure 6), and Table 4 shows the yields of these MAs. We did not recover methoxy-MAs from M. smegmatis A1.2.3.4 and M. smegmatis f_A1.2.3.4, as these strains likely produced cyclopropanated versions of these compounds; we will analyze these strains in the future. The process generated about 12–20 mg methoxy-MAs and 10–16 mg of keto-MAs per 100 mL culture of the recombinant M. smegmatis strains. The identities of these products were confirmed via MALDI mass spectrometry (Figure 7), and the m/z values were consistent with the published data (Purwantini and Mukhopadhyay, 2013; Fatima et al., 2024; Teramoto et al., 2015; Francü oise Laval et al., 2001; Cecilia et al., 2022). The recovered M-MAs and K-MAs carried 78–83 carbon atoms (Figure 7), and the mass of an M-MA species was one carbon mass unit higher than that of the corresponding K-MA. A -CH-OCH₃ unit of an M-MA is equivalent to a -C=O in the corresponding K-MA species and, consequently, the former MA form is expected to have ∼16 Da higher molecular mass compared to the latter (Figure 7; Table 5).

TABLE 4. Purification of Mycolic Acid Methyl Esters (MAMEs) from recombinant Mycobacterium smegmatis strains expressing Mycobacterium tuberculosis genes.

In each case, the quantification was performed by recovering material from a specific TLC, band (Figure 6A) and determining its weight. Each isolated preparation was examined for purity via TLC (Figure 6B).

TABLE 5. Observed m/z values in MALDI-MS analysis of recovered oxygenated mycolic acids and corresponding inferred number of carbons.

The presented m/z values are from data in Figure 7. The typical mass accuracy of the instrument used was less than 3 ppm.

The strain expressing mmaA1-4 genes produced more methoxy-MAs compared to that with mmaA3-4 (Table 4). For keto-MAs production the mmaA1-4 genes teamed up with fhmad better than mmaA3-4 offering a high yield of the compound (Table 4). Both observations suggest that the promoter of mmaA1-4 operon is stronger in driving the gene expression when compared to the internal promoter of mmaA3-4. Whether these effects were due to the activity of the cyclopropanation genes (mmaA1 and mmaA2) remains to be investigated. The MAs produced in Mtb exhibit the following relative levels: α-MA, 41.5%; M-MA, 33.5%; K-MA, 25% (Apoorva Bhatt et al., 2007). In our recombinant system with 4 M. smegmatis strains, the observed relative amounts were the following (cloned genes, % MA type): mmaA3-4, 25% α-MA, 14% E-MA, and 15% M-MA; mmaA3-4 and fhmad, 17% α-MA, 27% M-MA, and 14% K-MA; mmaA1-4, 25% α-MA, 18% E-MA, and 20% M-MA; mmaA1-4 and fhmad, 17% α-MA, 23% M-MA, and 18% K-MA; in each case the rest could be attributed to α′-MA and H-MA which were not quantified in our study. Interestingly, E-MAs and K-MAs were produced in similar proportions, suggesting a sort of replacement of the former by the latter by the action of the cloned fhmad gene. The mechanism underlying this effect has not been elucidated.

Comparison of the methoxy-MA preparation generated in the study with a commercially available product

In our TLC system, the migration pattern of the methoxy-MA standard obtained from Avanti Polar Lipids (Catalog no. 791281P) (Figure 5B, lane 1) was distinct from that of our methoxy-MA preparation (Figure 5B, lanes one and 3). This difference was attributed to the fact that our preparation contained methyl-esters of methoxy-MA (methoxy-MAMEs), and the purchased product was free lipids (acids). To confirm this assessment, we re-saponified our methoxy-MAs preparation with tetrabutyl ammonium hydroxide (TBAH) to generate free acids. The resultant de-esterified product exhibited a migration pattern that was similar to that of the material from Avanti Polar Lipids (Alabaster, AL) (Figure 5B lanes one and 2). This result in conjunction with our mass spectrometry data confirmed that we produced authentic methoxy-MAs of M. tuberculosis var. bovis BCG in our recombinant M. smegmatis strain.

Conclusion

Preliminary in vitro data from our laboratory tests indicate that the training of murine bone marrow-derived immune cells require MAs at a concentration of 10 μg/mL (Wu et al., 2024). This need can now be met by the genetically engineered and safe M. smegmatis strain-based system that we have developed as it can offer 10–20 mg of purified methoxy- and keto-MAs from a 100 mL culture (Table 4). The system can be scaled up easily to 100 s of liters (Biswarup Mukhopadhyay and Purwantini, 2000). Of the four engineered strains that we developed, two have the potential of producing cyclopropanated versions of the above-mentioned MAs (Figures 1, 2), and thereby, expanding the availability of candidates to four types of oxygenated MAs for downstream studies. Our system would also provide epoxy-MAs, if these are to be tested for immune modulation activities.

This applied study has revealed new facets of mycolic acid metabolism. The presence of fhmad (or fgd2 or rv0132c) boosted the level of M-MAs in a strain carrying either mmaA1-4 (Figure 5A, lanes pUB-A.1.2.3.4 versus pUB-f_A.1.2.3.4; lanes four versus 5) or mmaA3-4 (Figure 5A, lanes pUB-A.3.4 versus pUB-f_A3.4; lanes two versus 3). On the other hand, mmaA1-2, in the presence or absence of fhmad, lowers the level of M-MAs (Figure 5, pUB-f_A.3.4 versus pUB-f_A.1.2.3.4 and A.3.4 versus pUB-A.1.2.3.4; lanes three versus five and two versus 4); mmaA1-2 also lowers the levels of K-MAs (Figure 5A, pUB-f_A3.4 versus pUB-f_A1.2.3.4; lanes three versus 5). The molecular basis for these observations is not currently known.

Overall, this engineered platform provides an avenue for the generation of materials to the test the hypothesis that oxygenic mycolic acids could “rewire” the innate immune memory towards tackling sepsis with a manageable cost and for studying some aspects of mycolic acid modifications in M. tuberculosis in an amenable heterologous system.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Urszula Katarzyna Komarnicka, University of Wrocław, Poland

Reviewed by: Claire V Mulholland, Albert Einstein College of Medicine, United States

Neil Price, Agricultural Research Service (USDA), United States