This protocol details the enrichment of native mycobacterial extracellular vesicles (mEVs) from axenic cultures of Mycobacterium smegmatis (Msm) and how mCherry (a red fluorescent reporter)-containing recombinant MsmEVs can be designed and enriched. Lastly, it verifies the novel approach with the enrichment of MsmEVs containing the EsxA protein of Mycobacterium tuberculosis.
Most bacteria, including mycobacteria, generate extracellular vesicles (EVs). Since bacterial EVs (bEVs) contain a subset of cellular components, including metabolites, lipids, proteins, and nucleic acids, several groups have evaluated either the native or recombinant versions of bEVs for their protective potency as subunit vaccine candidates. Unlike native EVs, recombinant EVs are molecularly engineered to contain one or more immunogens of interest. Over the last decade, different groups have explored diverse approaches for generating recombinant bEVs. However, here, we report the design, construction, and enrichment of recombinant mycobacterial EVs (mEVs) in mycobacteria. Towards that, we use Mycobacterium smegmatis (Msm), an avirulent soil mycobacterium as the model system. We first describe the generation and enrichment of native EVs of Msm. Then, we describe the design and construction of recombinant mEVs that contain either mCherry, a red fluorescent reporter protein, or EsxA (Esat-6), a prominent immunogen of Mycobacterium tuberculosis. We achieve this by separately fusing mCherry and EsxA N-termini with the C-terminus of a small Msm protein Cfp-29. Cfp-29 is one of the few abundantly present proteins of MsmEVs. The protocol to generate and enrich recombinant mEVs from Msm remains identical to the generation and enrichment of native EVs of Msm.
Despite the development and administration of a wide range of vaccines against infectious diseases, even to this day, ~30% of all human deaths still occur from communicable diseases1. Before the advent of the Tuberculosis (TB) vaccine – Bacillus Calmette Guerin (BCG) – TB was the number one killer (~10,000 to 15,000/100,000 population)2. With the administration of BCG and easy access to first and second-line anti-TB drugs, by 2022, TB-related deaths have dramatically dropped to ~1 million/year by 2022 (i.e., ~15-20/100,000 population1). However, in TB endemic populations of the world, TB-related deaths continue to stand at ~100-550/100,000 population1. While experts recognize several reasons leading to these skewed numbers, BCG-mediated protection not lasting for even the first decade of life appears to be the prominent reason3,4,5,6,7. Consequently, given the renewed 'Sustainable Development Goals' of the UN and the 'End TB Strategy', of WHO, there is a concerted global effort to develop a much superior vaccine alternative to BCG that perhaps provides lifelong protection from TB.
Towards that objective, several groups are currently evaluating modified/recombinant BCG strains, non-pathogenic and attenuated mycobacterial species other than BCG, and subunit candidates8,9,10,11,12,13,14,15,16,17,18. Typically, subunit vaccines are liposomes selectively loaded with few purified (~1-6) full-length or truncated immunogenic proteins of the pathogen. However, because of their spurious folding into non-native conformations and/or random non-functional interactions between the loaded proteins, subunits often lack native and germane epitopes and hence, fail to sufficiently prime the immune system14,19,20.
Consequently, extracellular vesicles (EVs) of bacteria have picked up pace as a promising alternative21,22,23,24,25,26. Typically, bacterial EVs (bEVs) contain a subset of their cellular components, including some portions of nucleic acids, lipids, and hundreds of metabolites and proteins27,28. Unlike liposomes where a few purified proteins are artificially loaded, bEVs contain hundreds of naturally-loaded, natively-folded proteins with a better propensity to prime the immune system, especially without the boost/aid of adjuvants and Toll-like receptor (TLR) agonists27,28,29. It is in this line of research that we and others have explored the utility of mycobacterial EVs as potential subunit boosters to BCG30. Despite concerns that bEVs lack uniform antigen loads, EVs from attenuated Neisseria meningitidis have successfully protected humans against serogroup B meningococcus31,32.
At least theoretically, the best EVs that could boost BCG well are the EVs enriched from pathogenic bacteria. However, enriching EVs generated by pathogenic mycobacterium is expensive, time-consuming, and risky. Additionally, pathogen-generated EVs may be more virulent than protective. Given the potential risks, here, we report a well-tested protocol for the enrichment of EVs generated by axenically grown Msm, an avirulent mycobacterium.
However, despite encoding several pathogen protein orthologs, avirulent mycobacteria lack several vaccine antigens/pathogenic protein epitopes necessary to sufficiently prime the immune system towards protection33. Therefore, we also explored constructing and enriching recombinant EVs of Msm through molecular engineering, such that a significant portion of any pathogenic protein of interest expressed and translated in Msm, must reach its EVs. We hypothesized that one or more of the top 10 abundant proteins of Msm EVs when fused to the protein of interest will aid in such translocation.
While we were beginning to standardize the enrichment of mycobacterial EVs (mEVs) in our laboratory, in 2011, Prados-Rosales et al. first reported the visualization and enrichment of mEVs in vitro30. Later, in 2014, the same group published a modified version of their 2011 method34. In 2015, Lee et al. also reported an independently standardized method for mEV enrichment again from axenic cultures of mycobacteria35. Combining both protocols34,35 and incorporating a few of our modifications after thorough standardization, we describe here a protocol that helps routinely enrich mEVs from axenic cultures of mycobacteria36.
Here, we particularly detail the enrichment of Msm-specific EVs, which is an extension of a published protocol36 for the enrichment of mycobacterial EVs in general. We also detail how to construct recombinant mEVs (R-mEVs) that contain the mCherry protein (as a red fluorescent reporter) and EsxA (Esat-6)37,38,39 a predominant immunogen and a potential subunit vaccinogen of Mycobacterium tuberculosis. The protocol for enriching the R-mEVs remains identical to the one we have described for enriching native EVs from Msm.
1. Growth conditions of Mycobacterium smegmatis, Escherichia coli, and their derivatives
2. Enrichment of Msm mEVs by employing density gradient centrifugation
3. Construction and enrichment of recombinant mEVs.
NOTE: One of the 10 most abundant proteins (identified by mass spectrometry) of Msm EVs is Cfp-2930. Given its small size (29 kDa), simple secondary structure40, localization to the membrane41, and propensity to be secreted into spent media in axenic cultures (e.g., as a culture filtrate protein; secreted by both Msm and Mtb42,43, here, it has been exploited to deliver a red fluorescent reporter and a protein of interest (EsxAMtb) into mEVs. To achieve this,
We use M. smegmatis (Msm) as the model mycobacterium to demonstrate the enrichment of both native and recombinant mEVs (R-mEVs). This schematically summarized mEVs enrichment protocol (Figure 1) also works for the enrichment of R-mEVs of Msm and native EVs of Mtb (with minor modifications as in protocol notes of 1.2). Visualization of the enriched mEVs requires negatively staining them under a transmission electron microscope36 (Figure 2A). Typically, Msm-specific EVs separate out in the 4th-6th 1 mL fractions of the 13 mL 6-60% density gradient (Figure 2B). Approximately 80-100 µg protein equivalent of EVs are routinely obtained from 2 L of mid-logarithmic axenic cultures of Msm. Their diameters typically range between 20 nm and 250 nm (Figure 2C).
One long-term goal of our laboratory is to evaluate if mEVs of different mycobacteria could potentially act as a subunit booster to the existing vaccine, BCG. Since enriching mEVs generated by pathogenic bacteria is time-consuming, risky, and expensive, exploiting native and recombinant EVs from avirulent mycobacteria may work as a suitable alternative. Hence, we aim to not only standardize the protocol to enrich the mEVs of Msm but also to construct and enrich its R-mEVs.
To construct R-mEVs, we first selected the top 10 abundant proteins of Msm EVs (Table 1; we identified them from detailed mass spectrometry analyses of the Msm EVs)30. We hypothesized that if a foreign protein of interest is translationally fused to any of them, it should be able to localize into mEVs. Then, we shortlisted Cfp-29 among the 10 because it is the smallest among them (~29 kDa), is membrane-localized, and is a culture filtrate protein with a relatively simple secondary structure40,41,42,43. We translationally fused mCherry's (fluorescent protein) N-terminal end to the C-terminal end of Cfp-29 and evaluated its loading/delivery into mEVs. A portion of the enriched mEVs of Msm turned pink (Figure 3), indicating Cfp-29's ability to carry a foreign protein of interest into Msm EVs.
Given this ability of Cfp-29 (Figure 3), we then evaluated EsxA (Esat-6), a major immunogenic protein37,38,39 delivery into Msm EVs. Again, we generated two independent translational fusions to the C-terminus of Cfp-29-only EsxA and EsxA + 3X FLAG tag (3X FLAG fused to the C-terminus of EsxA. As expected, we observed Mtb's EsxA in Msm EVs (lanes 5, 6, and 10, Figure 4A,B), albeit in low quantities. Interestingly, Cfp-29::EsxA::3XFLAG was much more stable (lanes 3 and 4 plus 8 and 9; Figure 4A) and accumulated at higher levels in mEVs (lanes 3 and 4 plus 8 and 9; Figure 4B). In summary, we demonstrate the design, construction, and enrichment of R-mEVs that contain a foreign protein of interest (Figure 3 and Figure 4).
Figure 1: Schematic representation of mycobacterial extracellular vesicle enrichment. 'Days' (in red font, top of the figure) refers to the days necessary to enrich the mEVs from Msm (specifically for protocol steps 1 and 2). "Steps" (in black font, bottom of the figure) refer to the protocol steps as described in the protocol section. For enriching mEVs from Mtb, although all steps are similar, it takes at least 10 days for the different steps of culturing (step 1) to the first step of centrifugation (step 2). Subsequent steps require identical duration (as indicated in red font). Before the density gradient, the mEVs pellet + extracellular complexes look colorless when enriching both native mEVs and R-mEVs. However, it would appear pinkish to blue (see Figure 3) when enriching mEVs containing mCherry. Post the density gradient, the mEVs would appear white (native and R-mEVs) or pink (if containing mCherry) in the density gradient tube. The colors of mEVs in the figure are only for indicating clarity and do not represent the exact colors. See Figure 3 for more clarity. Abbreviations: Msm = M. smegmatis; Mtb = M. tuberculosis; 0.45 and 0.22 µm = pore size of disposal filter units. Please click here to view a larger version of this figure.
Figure 2: Mycobacterial EVs are circular, concentrate with density gradients, and vary in dimensions. (A) A representative image of Msm EVs upon their visualization with negative staining and viewing under a transmission electron microscope. Scale bar = 200 nm. (B) A representative image of how mEVs appear in the ultracentrifuge tube upon performing a 6-60% density gradient. If the 6-60% gradient is not accurately layered, the positioning of the major (top open arrow) and minor (bottom open arrow) bands of mEVs can significantly alter, thus altering the 1 mL fraction number. (C) A representative image upon nanotracking analysis of enriched mEVs of Msm. Nanotracking analyses reveal the proportions and concentrations of different-sized mEVs and the total number of mEVs within the preparation. On average, with the protocol detailed here, ~1-3 × 1010 EVs are enriched from 2 L of Msm and Mtb. Please click here to view a larger version of this figure.
Figure 3: Different steps indicative of enrichment of mEVs expressing mCherry. (A) Representative image showing Msm axenic culture expressing Cfp-29::mCherry. (B) Representative image of the bacterial pellet post centrifugation of Msm axenic culture expressing Cfp-29::mCherry. (C) Representative image of culture filtrate concentrate obtained post concentrating the spent media of Msm axenic culture expressing Cfp-29::mCherry through centricon concentrators. (D) Representative image of mEVs + extracellular complexes pellet post ultracentrifugation of culture filtrate concentrate obtained from Msm axenic culture expressing Cfp-29::mCherry. The pellet would, however, remain colorless if the mEVs were to be either native or recombinant (upon fusion to Cfp-29) for any foreign protein(s) except mCherry. (E) Representative images of mCherry containing mEVs. Note that not all mEVs are pink, indicating that not all mEVs contain Cfp-29. Good: A representative image indicating different bands of mEVs typically obtained after enriching for Cfp-29::mCherry EVs. Since mCherry-containing EVs are denser, they separate into later fractions. Poor: A representative image indicating poor separation (the white mEVs separate out in the first/second fraction and the mCherry-containing mEVs separate out in the 10th fraction (possibly because of poor resuspension of mEVs pellet, i.e., protocol step 2.3.2). Please click here to view a larger version of this figure.
Figure 4: Recombinant mEVs containing EsxA (ESAT-6), an Mtb-encoded immunogen in Msm total cell lysate and Msm generated EVs. (A) Coomassie gel and (B) western analysis (detected with ESAT6-specific polyclonal antibody) images showing the accumulation of fused ExsA in both Msm total cell lysate and mEVs expressing either cfp-29::esxA::3XFLAG (+ 3XFLAG, lanes 3 and 4 (total lysate) and lanes 8 and 9 (mEVs)) or cfp-29::esxA (- 3XFLAG, lanes 5 and 6 (total lysate) and lane 10 (mEVs)). Open and filled arrows indicate accumulated Cfp-29::ExsA::3XFLAG and Cfp-29::ExsA, respectively. Please click here to view a larger version of this figure.
Primers: | ||||||||||
Sl # | Gene | Primer | Sequence (5' to 3') | Source | To clone into | |||||
1 | cfp-29 | Forward | CAGTTCGAA(BstBI)ATGAACAACCTCTATCGC | Msm genomic DNA | pMV261 | |||||
Reverse | GAAAAGCTT(HindIII)GGGGGTCAGCGCGACAG | Msm genomic DNA | pMV261 | |||||||
2 | mCherry | Forward | GAAAAGCTT(HindIII) ggcggcggtggctcg (G4S linker)ATGGTGAGCAAGGGCGAGGAGG | Lab collection | pMV261 | |||||
Reverse | TGTGTTAAC(HpaI)CTACTTGTACAGCTCGTCC | Lab collection | pMV261 | |||||||
3 | esxA | Forward | GAAAAGCTT(HindIII)ggcggcggtggctcg (G4S linker) ATGACAGAGCAGCAGTGGAATTTCGCGGGTATCGAG | Mtb genomic DNA | pMV261 | |||||
Reverse | TGTGTTAAC(HpaI)TCATGCGAACA TCCCAGTGACGTTGCCTTCGGTCG |
Mtb genomic DNA | pMV261 | |||||||
4 | exsA-3X FLAG | Forward | GAAAAGCTT(HindIII)ggcggcggtggctcg (G4S linker) ATGACAGAGCAGCAGTGGAATTTCGCGGGTATCGAG | Mtb genomic DNA | pMV261 | |||||
Reverse | TGTGTTAAC(HpaI)TCA cttgtcgtcgtcgtccttgtagtcgatgtcgtg gtccttgtagtcaccgtcgtggtccttgtagtc (3XFLAG) TGCGAACATCCCAGTGACGTTG CCTTCGGTCGAAGCCATTGCCTGACC |
Mtb genomic DNA | pMV261 |
Table 1: Primer sequences. Forward and Reverse primers for amplifying cfp-29, mCherry, esxA, and esxA::3X FLAG and cloning into shuttle vector pMV261 are listed.
Supplementary File 1: A: pMV261, its circular map, main features, and complete nucleotide sequence. Yellow highlight-hsp60 promoter nucleotide sequence; Green and cyan highlight-restriction sites into which mCherry, esxA, and esxA::3X FLAG amplicons were cloned. B, C, and D: Nucleotide sequences of cfp-29, mCherry, and esxA, respectively. Green highlight-stop codon of cfp-29, mCherry, and esxA. Please click here to download this File.
Since developing a novel TB vaccine that is superior to and can replace BCG remains a formidable challenge, as an alternative, several groups are pursuing the discovery of different subunit TB vaccines that can boost BCG's potency and extend its protective duration48,49. Given the increasing attention to bacterial EVs (bEVs) as potential subunits and as natural adjuvants50,51, consistent enrichment of sufficient quantities of mEVs for their downstream testing/analysis has become an important step. It is in consideration of these research questions that this protocol aims to enrich mEVs from axenic cultures and their recombinant versions.
During detailed mass spectrometric proteome analyses of Msm EVs, Prados-Rosales et al. identified the 10 most abundant proteins30. We further hypothesized that upon sufficient molecular engineering, one among them should be sufficient to carry a foreign protein of interest into mEVs. Interestingly, Cfp-29 stood out as the best possible option because of its various features39,40,41,42. Our data indeed show that it is sufficient enough to carry mCherry and EsxA (albeit EsxA required a small tag at its C-terminus to be more stable) and accumulate them in the mEVs. Recently, in 2021, Tang et al. demonstrated that Cfp-29 is an encapsulin with the capability to carry dye-decolorizing peroxidase (DyP)-type peroxidases40. Perhaps, Cfp-29 can carry other foreign proteins too.
We explored EsxA because it is a prominent Mtb immunogen37,38,39, can be protective as a subunit vaccine in animal model37,38,39, is small in size; and its ortholog (MSMEG_0066) is interestingly absent in Msm EVs. Although we do not discuss this here, we have successfully generated R-mEVs for a few other Mtb proteins (uniquely present in Mtb and not encoded by Msm), including Antigen 85B (Rv 1886c). At the same time, interestingly, we also failed to generate R-mEVs for a few others, including full-length Rv2660c and Rv0288, possibly because the proteins are toxic to Msm. We conclude so because, despite cloning the correct nucleotide sequences and performing repeated transformations, no tranformants of Msm emerged. Since EsxA required a 3XFLAG tag at its C-terminus end for better accumulation in mEVs, we added a 3XFLAG tag to all other Mtb-encoded proteins that we evaluated. Despite the tag, Msm did not survive, indicating some Mtb proteins remain toxic despite the tag fusion. We speculate that in these cases, cloning just the predicted epitope regions or stitching several epitopes together may pay off (currently under evaluation in our laboratory). We used a small five amino acid (four Glycine and one Serine) linker between Cfp-29 and mCherry/EsxA to minimize the negative influence on the folding of the protein of interest44,45. We predict that without this linker, the protein of interest folding would be strongly influenced by Cfp-29 folding.
This enrichment protocol is easily extendable to the enrichment of Mtb-generated EVs. We also speculate that enriching mEVs from environmental mycobacteria should be also feasible with this protocol. Despite the protocol being simple and straightforward, it still requires 7-8 days to enrich Msm-generated EVs and R-mEVs. In contrast, it requires 15-20 days for the enrichment of Mtb-generated EVs. Concentrating EVs to a smaller volume is time-consuming and expensive and we are currently exploring tangential flow filtration and other approaches to solve the 'time' issue.
Lastly, we use Sauton's and not 7H9 to enrich mEVs because the 'ADC' supplement contains high amounts of bovine serum albumin that may interfere with any downstream uses of mEVs. This protocol may be easily extendable to any other specific media (e.g., low-iron media) that has to be used to evaluate the mEVs composition. Alternatively, for certain applications, at the last step (i.e. protocol step 2.7.5), instead of HEPES buffer, one could use sterile saline when injecting the same into mice or guinea pigs for various in vivo-based analyses.
The authors have nothing to disclose.
The authors sincerely thank Prof. Sarah M. Fortune for kindly sharing M. smegmatis mc2155 stock. They also acknowledge Servier Medical Art (smart.servier.com) for providing some basic elements for Figure 1. They sincerely acknowledge the support of the rest of the lab members for their patient adjustments during the long use of the incubator shakers, centrifuges, and ultracentrifuges for mEV enrichment. They also acknowledge Mr. Surjeet Yadav, the laboratory assistant, for always making sure the necessary glassware and consumables were always available and handy. Lastly, they acknowledge the administrative, the purchase, and the finance teams of THSTI for their constant support and help in the seamless execution of the project.
A2 type Biosafety Cabinet | Thermo Fisher Scientific, USA | 1300 series | |
Bench top Centrifuge | Eppendorf, USA | 5810 R | |
BstB1, HindIII, HpaI | NEB, USA | NEB | |
Cell densitometer | GE Healthcare, USA | Ultraspec 10 | |
Citric Acid | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Dibasic Potassium Phosphate | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Double Distilled Water | Merck, USA | ~18.2 MW/cm @ 25 oC | |
Electroporation cuvettes | Bio-Rad, USA | 2 mm | |
Electroporator | Bio-Rad, USA | Electroporator | |
EsxA-specific Ab | Abcam, UK | Rabbit polyclonal | |
Ferric Ammonium Citrate | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Floor model centrifuge | Thermo Fisher Scientific, USA | Sorvall RC6 plus | |
Glassware | Borosil, INDIA | 1 L Erlenmeyer flasks | |
Glycerol | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
HEPES and Sodium Chloride | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Incubator shakers | Thermo Fisher Scientific, USA | MaxQ 6000 & 8000 | |
L-Asparagine | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Luria Bertani Broth and Agar, Miller | Hi Media, INDIA | Hi Media | |
Magnesium Sulfate Heptahydrate | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Magnetic stirrer | Tarsons, INDIA | Tarsons | |
mCherry-specific Ab | Abcam, UK | Rabbit monoclonal | |
Microwave | LG, INDIA | MC3286BLT | |
Middlebrook 7H9 Broth | BD, USA | Difco Middlebrook 7H9 Broth | |
Middlebrook ADC enrichment | BD, USA | BBL Middlebrook ADC enrichment | |
Nanodrop | Thermo Fisher Scientific, USA | Spectronic 200 UV-Vis | |
NEB5a | NEB, USA | a derivative of DH5a | |
Optiprep (Iodixanol) | Merck, USA | Available as 60% stock solution (in water) | |
PCR purification kit | Hi Media, INDIA | Hi Media | |
pH Meter | Mettler Toledo, USA | Mettler Toledo | |
Plasmid DNA mini kit | Hi Media, INDIA | Hi Media | |
Plate incubator | Thermo Fisher Scientific, USA | New Series | |
Plasmid pMV261 | Addgene, USA * *The plasmid is no more available in this plasmid bank |
Shuttle vector | |
Proof-reading DNA Polymerase | Thermo Fisher Scientific, USA | Phusion DNA Plus Polymerase | |
Q5 Proof-reading DNA Polymerase | NEB, USA | NEB | |
Refrigerated circulating water bath | Thermo Fisher Scientific, USA | R20 | |
Middlebrock 7H11 Agar base | BD, USA | BBL Seven H11 Agar base | |
SOC broth | Hi Media, INDIA | Hi Media | |
Sodium Hydroxide | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
T4 DNA Ligase | NEB, USA | NEB | |
Tween-80 | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
Ultracentrifuge | Beckman Coulter, USA | Optima L100K | |
Ultracentrifuge tubes – 14 mL | Beckman Coulter, USA | Polyallomer type – ultra clear type in SW40Ti rotor | |
Ultracentrifuge tubes – 38 mL | Beckman Coulter, USA | Polypropylene type– cloudy type for SW28 rotor | |
Ultrasonics cleaning waterbath sonicator | Thermo Fisher Scientific, USA | Sonicator – bench top model | |
0.22 µm Disposable filters | Thermo Fisher Scientific, USA | Nunc-Nalgene | |
30-kDa Centricon concentrators | Merck, USA | Amicon Ultra centrifugal filters – Millipore | |
3X FLAG antibody | Sigma-Aldrich, Merck, USA | Sigma Aldrich | |
400 mL Centrifuge bottles | Thermo Fisher Scientific, USA | Nunc-Nalgene | |
50 mL Centrifuge tubes | Corning, USA | Sterile, pre-packed | |
Bacteria | |||
Strain | |||
Escherichia coli | NEB, USA | NEB 5-alpha (a derivative of DH5α). | |
Msm expressing cfp29::mCherry | This study | MC2 155 | |
Msm expressing cfp29::esxA | This study | MC2 155 | |
Msm expressing cfp29::esxA::3X FLAG | This study | MC2 155 | |
Mycobacterium smegmatis (Msm) | Prof. Sarah M. Fortune, Harvard Univ, USA | MC2 155 |