The Role of Resuscitating Promoting Factors in Facilitating Infection of Mycobacterium smegmatis with Virulent Soil Bacteriophages

Bacterial strains and culture conditions

All bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli DH5α was grown in LB broth or on LB agar. Mycobacterium smegmatis strains were grown in 7H9 (Difco) Middlebrook supplemented with 0.5% glycerol, 0.2% glucose, 0.085% NaCl and 0.05% Tween-80 or tyloxapol or on 7H10 (Difco) agar medium supplemented with 0.5% glycerol, 0.2% glucose and 0.085% NaCl. Hygromycin was added at 200 μg/mL for E. coli and 50 μg/mL for M. smegmatis as needed.

Drug susceptibility and permeability testing

M. smegmatis mc²155 strain (wildtype, WT), the rpf deletion mutant (Δrpf) and the various rpf complemented derivatives (Table 1) were grown to exponential phase. Cultures were adjusted to optical density (OD)_{600 nm} = 0.4, serially diluted (10⁻¹ to 10⁻⁵), and appropriate dilutions plated on 7H10 agar to assess colony forming units (CFU/mL). In addition, 10 μL of each dilution was aliquoted onto 7H10 media supplemented with 0.004% sodium dodecyl sulfate (SDS) or 3 μg/mL vancomycin and un-supplemented plates served as controls. The plates were incubated at 37°C for 4 days before scoring for growth. The experiment was performed in triplicate.

Growth kinetics of the M. smegmatis wild type, rpf deletion mutant and complemented strains

Triplicate cultures of the WT, Δrpf and the various rpf complemented derivatives (Table 1) were inoculated at an OD_600 nm of 0.05 in 7H9 media with tyloxapol using overnight precultures grown in the same media. Cultures were incubated with shaking at 37°C and OD_600 nm was first measured 16 h, then at 3- and 16-h intervals over a 48-h period. The OD_600 nm measurements for the three biological replicates were averaged and plotted against time.

Preparation of M. smegmatis for bacteriophage infection

A single colony of each M. smegmatis strain was grown overnight in 5 mL of 7H9 medium at 37°C with shaking at 100 rpm. Pre-cultures were diluted 1:100 into 50 mL 7H9 and grown to mid log phase (OD_600 nm = 0.8) or late log phase (OD_600 nm = 2). Cultures were harvested by centrifugation at 4 500 × g, washed twice with 30 mL of mycobacteriophage (MP) buffer (10 mM Tris pH 7.5, 10 mM MgSO₄, 1 mM CaCl₂, and 68.5 mM NaCl), and resuspended in the appropriate volume of MP buffer to an OD_600 nm of 2.

Purification and amplification of mycobacteriophages

To isolate mycobacteriophages, 2 mL of soil samples were resuspended in 10 mL of MP buffer, shaken (200 rpm) for 2 h at 37°C and the supernatant filtered using a Membrane filtration (MF)-Millipore® Membrane 0.22 μm filter. Ten 50 μL aliquots were mixed with 450 μL of the WT strain (OD_600 nm = 2) at room temperature for 30–60 min. Following addition of 4.5 mL of Mycobacteriophage top agar (MPTA), mixtures were overlaid on 7H10 plates and incubated at 37°C for 2–3 days. MPTA contains 7H9 (Difco) Middlebrook liquid medium, supplemented with 1.14% LB-agar, 0.2% glycerol, 0.2% glucose, 0.085% NaCl, and 0.1% CaCl_2. Single plaques were picked and resuspended in 1 mL of MP buffer. The suspension was filtered, serially diluted (10⁻¹–10⁻¹⁰) and 450 μL of WT cells were added to 50 μL of each dilution of the mycobacteriophage (multiplicity of infection [MOI] 1:10, mycobacteriophage to bacteria) to ascertain the dilution yielding confluent (‘lacey’) plaques. High titer lysate was prepared by overlaying several ‘lacey’ plates with 3 mL of MP buffer each and incubating the plates for 4–6 h at room temperature with intermittent manual agitation. The pooled mycobacteriophage suspensions were collected, filter sterilized and the plaque-forming units (PFU)/mL assessed as described above. The mycobacteriophage lysates were stored at 4°C.

Adsorption assays

A single colony of the WT, Δrpf, and Δrpf::rpfAB, was inoculated in 10 mL of 7H9 medium and incubated overnight at 37°C with shaking at 100 rpm. Pre-cultures were diluted 1:100 in 100 mL and grown overnight to mid log phase (OD _{600 nm} = 0.8) or late log phase (OD _{600 nm} = 2). Cells were harvested by centrifugation at 4 500 × g, washed twice with 40 mL of MP buffer and the pellets resuspended in the appropriate volume of MP to OD_600 nm of 2. For infection, 2.7 mL aliquot of each strain was mixed with 300 μL of high titer mycobacteriophage lysate (∼ 10⁹ PFU/mL) at an MOI of 1:10 and incubated at 37°C with gentle shaking at 30 rpm. At time points 0-, 15-, 30-, 60-, 90-, and 120-min post infection, 300 μL of the mycobacteriophage cell suspension was collected, vortexed for 5 s where appropriate and centrifuged at 20,000 × g for 30 s. Supernatant containing un-adsorbed bacteriophages, was carefully aspirated, diluted serially (10⁰–10⁻¹⁰), and 50 μL of each dilution was used to infect 450 μL WT cells (OD₆₀₀ = 2). After 15 min of incubation at room temperature, MPTA overlays were poured on 7H10 plates. Plates were incubated at 37°C for 2–3 days and the percentage PFU/mL relative to WT was calculated with data representing the average of three biological repeats.

Negative staining transmission electron microscopy

Transmission electron microscopy was performed at the Electron Microscope Unit, University of Cape Town, South Africa. Briefly, carbon-coated copper grids (Agar Scientific, UK) were made hydrophilic using a EMS100 Glow Discharge Unit (Electron Microscopy Sciences, USA). Samples were applied to the grids and negatively stained with 2% uranyl acetate (SPI Supplies, USA). Imaging was performed on a FEI T20 transmission electron microscope (Thermo Fisher [formerly FEI], Eindhoven, Netherlands) at 200 kV and images were collected using a Gatan US 1000 CCD camera (Ametek Inc., USA).

Extraction, whole genome sequencing, and annotation of mycobacteriophage DNA

Details are provided in the Supplementary Data.

Construction of M. smegmatis strains expressing single M. tuberculosis rpfA or rpfB homologues

In a previous study, we generated the pMV-RPFAB integrating plasmid with rpfA and rpfB genes from M. tuberculosis. ¹⁹ This vector was further manipulated to generate pMV-RPFAΔint and pMV-RPFBΔint, as described in the Supplementary Data.

Data analysis

Graphs and figures were generated using Microsoft PowerPoint, Excel, GraphPad Prism (version 10.2.3) or Biorender.com software. GraphPad Prism (version 10.2.3) was used for all statistical analysis and data comparisons.

Isolation, purification, and characterization of lytic mycobacteriophages

Two mycobacteriophages were isolated from damp soil collected in Johannesburg, South Africa: MontyDev from under a lavender bush (Benoni, GPS co-ordinates −26.1777°, 28.2838°) and Bora from a flowerbed (Braamfontein, GPS co-ordinates −26.1947°, 28.0323°). Both mycobacteriophages formed circular, turbid zones of clearance, on a lawn of WT M. smegmatis cells, with MontyDev ∼1–2 mm (Fig. 1A) and Bora in ∼1 mm in size (Fig. 1B). Transmission electron microscopy revealed that both mycobacteriophages are siphoviral with long, non-contractile tails (Fig. 1C and D). MontyDev has a ∼71 nm diameter icosahedral capsid and a ∼284 nm long tail, whereas Bora has a prolate capsid, measuring ∼163 nm long and ∼40 nm wide, with a ∼247 nm long tail.

OPEN IN VIEWER

FIG. 1.

Characterization of mycobacteriophage plaque morphology and virion morphotype. (A) Plaques produced by phage MontyDev on a lawn of WT. (B) Plaques produced by phage Bora on a lawn of WT. (C) Mycobacteriophage MontyDev virions contain a ∼71 nm wide icosahedral head and a non-contractile tail with a length of ∼284 nm. (D) Mycobacteriophage Bora virions contain a ∼163 nm long and ∼40 nm wide prolate head, and a non-contractile tail with a length of ∼247 nm. WT, wildtype.

Genomic analysis of mycobacteriophages MontyDev and Bora

Whole genome sequencing confirmed that both mycobacteriophages are siphoviral with MontyDev in cluster R and Bora in cluster O. No integrase genes were detected in both the genomes, suggesting they are virulent mycobacteriophages. Neither mycobacteriophage genome contain tRNA or tmRNA reading frames. Predicted gene functions include structural, packaging, DNA replication, lytic, packaging and assembly, and hypothetical proteins (Supplementary Table S1). MontyDev has 71,250 bp, a G + C content of 56.0%, 101 predicted open reading frames, of which 70 (69.3%) are hypothetical proteins. All open reading frames except one (MONTYDEV_42) are transcribed from the forward, positive strand (Fig. 2A, Supplementary Table S1). The genome of Bora has 71,494 bp, a G + C content of 65.3%, and 124 predicted open reading frames, of which 82 (66.1%) are hypothetical proteins. Of these 124 genes, 54 are transcribed from the forward and 70 from the reverse strand (Fig. 2B, Supplementary Table S1). Both MontyDev and Bora contain the endolysins Lysin A and Lysin B, while the Holin protein is present only in Bora (Fig. 2A and B, Supplementary Table S1). Sequencing and annotation information for both mycobacteriophages is available at the Sequence Read Archive (BioProject numbers PRJNA1108268 and PRJNA1108264, respectively), the Actinobacteriophage Database (https://phagesdb.org) under https://phagesdb.org/phages/MontyDev/ and https://phagesdb.org/phages/Bora/, and Genbank (Accessions numbers PP763599 and PP763600, respectively).

OPEN IN VIEWER

FIG. 2.

Graphic representation of the genomes of mycobacteriophages MontyDev and Bora. The images were produced based on output from the Phamerator website (https://phamerator.org/phages). Nomenclature follows the guidelines as stipulated at the SEA-PHAGES Program (https://phagesdb.org/media/workflow/protocols/pdfs/Function_Assignments_best.pdf). (A) The genome of MontyDev (Cluster R) comprises 71,250 bp and 101 predicted genes. (B) The genome of Bora (Cluster O) comprises 71,494 bp and 124 predicted genes. Peptidoglycan lytic genes in these genomes are Lysin A (MONTYDEV_35 and BORA_64) Holin (BORA_66), tape measure protein (MONYTDEV_27) and minor tail protein, D-ala-D-ala carboxypeptidase (BORA_57). Lysin B (MONTYDEV_39 and BORA_65) breaks the ester bond between arabinogalactan and the mycolic acid layer. Blocks represent predicted open reading frames. Those above the ruler are transcribed in the forward orientation, below the ruler in reverse orientation. Colors: green, structural function; blue, nucleic acid metabolic function; red, lytic function; purple, other function; gray, hypothetical protein/unknown function.

OPEN IN VIEWER

FIG. 2.

(Continued).

Susceptibility of M. smegmatis strains with and without rpfs to MontyDev and Bora

We previously generated M. smegmatis strains in which a mutant (Δrpf) devoid of all its chromosomal rpf genes (rpfA, rpfB, rpfE and rpfE2) was complemented with combinations of the five M. tuberculosis rpf genes with their cognate promoters. ²² In our prior work, we generated a set of plasmids containing different combinations of the rpf genes from M. tuberculosis. The homology between M. tuberculosis and M. smegmatis rpfA and rpfB alleles at the protein level is high (see Supplementary Fig. S1). The rpfA gene from the two organisms share a 44.7% identity (entire length of 515 amino acids [aa]) with a 90.7% similarity in transglycosylase domain PF06737 (68/75 aa), whilst the rpfB alleles share a 74.4% identity (entire length 375 aa), with 90.7% similarity in transglycosylase domain PF06737 (68/75 aa). AlphaFold predictions for RpfA generate space filling models which all contain the conserved PF06737 domain (blue, high confidence in ribbon model; green block, low expected error) (Supplementary Fig. S2). RpfA and RpfB show highly conserved predictions for the entire length of the protein, while RpfC, RpfD, and RpfE contain the conserved PF06737 domain flanked by regions of lower confidence (Supplementary Fig. S2).

We first infected the various strains with MontyDev at a MOI of 1:10 to assess the role of rpf genes for mycobacteriophage infection. The Δrpf mutant and the Δrpf complemented with RpfCDE ^Mtb (Δrpf::RpfCDE ^Mtb ) yielded lower PFUs/mL compared to WT (Fig. 3A). However, combinatorial addition of rpfAB^Mtb to Δrpf (Δrpf::RpfAB ^Mtb ) showed a significant recovery in PFUs/mL, albeit still less than WT. The addition of rpfAB^Mtb to Δrpf::RpfCDE ^Mtb to generate a strain containing all the M. tuberculosis rpfs (Δrpf::RpfABCDE ^Mtb ) resulted in similar plaques levels as Δrpf::RpfAB ^Mtb suggesting that RpfA and RpfB together contribute to mycobacteriophage infection (Fig. 3A).

OPEN IN VIEWER

FIG. 3.

Infection of Mycobacterium smegmatis strains with lytic phages MontyDev and Bora. The mycobacteriophages were incubated with the respective bacterial cells at a MOI of 1:10 (1 mycobacteriophage:10 bacteria) and allowed to adsorb for at least 15 min. PFU were assessed by mixing the bacteriophage and cell suspension with top agar and overlaid on 7H10 plates. (A) Bar graph depicting normalized PFU in the WT, Δrpf mutant, and the Δrpf mutant complemented with combinations of rpfA–E infected with MontyDev. (B) Bar graph depicting normalized PFU in the WT, Δrpf mutant, and the Δrpf mutant complemented with combinations of rpfA and rpfB infected with MontyDev. (C) Comparison of the plaque forming units for WT, Δrpf mutant, and the Δrpf mutant complemented with combinations of rpfA and rpfB infected with MontyDev (open bars) and Bora (hatched bars). Error bars indicate standard error of the mean of three experiments. *p < 0.05, **p < 0.01, ***p < 0.005, ****p ≤ 0.0001 and ns = not significant. Graphs were generated using GraphPad Prism software, version 10. MOI, multiplicity of infection; PFU, plaque forming units; Rpf, resuscitation-promoting factor.

To evaluate the individual role of RpfA and RpfB in bacteriophage infection, each gene from M. tuberculosis was cloned into integrative plasmids and electroporated into Δrpf to generate strains Δrpf::RpfA ^Mtb and Δrpf::RpfB ^Mtb , respectively (Table 1). Quantitative polymerase chain reaction (PCR) confirmed the expression of the introduced genes in Δrpf (Supplementary Fig. S3). Only expression of rpfA and rpfB was assessed in WT since M. tuberculosis does not have the equivalent of the M. smegmatis rpfE and rpfE2 homologues. In M. smegmatis, expression of the rpfA^Msm transcript was ∼10-fold higher than the rpfB^Msm transcript (Supplementary Fig. S3). Complemented Δrpf strains expressed all M. tuberculosis rpf genes, albeit at varying degrees. We expected some differences in rpf gene expression in the recombinant M. smegmatis strains, as native M. tuberculosis promoters were used. Indeed, the rpfC^Mtb transcript was most abundant, while expression of the remaining M. tuberculosis genes was more comparable in each of the complemented strains, and to the M. smegmatis rpfA and rpfB genes (Supplementary Fig. S3). Despite these variations, our prior work using these heterologous complemented strains restored biofilm forming defects in M. smegmatis rpf deficient mutants. ¹⁹ Based on these data we tested the effect of rpf gene deletion on drug susceptibility and permeability. The Δrpf mutant showed a 2–3 log reduction in growth in the presence of 3 μg/mL vancomycin and 0.004% SDS which was rescued by rpfA and rpfB individually and in combination to WT levels (Supplementary Fig. S4). Comparison of growth kinetics showed no significant viability defect between the Δrpf and the three complemented strains Δrpf::RpfAB^Mtb, Δrpf::RpfA ^Mtb and Δrpf::RpfB ^Mtb . The WT strain displayed a small but significant difference in growth rate between 10 and 30 h; however, this difference was subsequently abrogated. The WT strain showed reduced growth in stationary phase; however, this was minor and likely due to clumping that routinely occurs in mycobacterial cultures at this stage of growth (Supplementary Fig. S5).

Both Δrpf::RpfA ^Mtb and Δrpf:RpfB ^Mtb , when infected with cluster R mycobacteriophage MontyDev, showed increased plaque formation, compared to Δrpf (∼25% vs. ∼10%). However, Δrpf::RpfAB ^Mtb showed more than double the plaque-forming ability compared with the single gene complements, indicating synergy between RpfA and RpfB (Fig. 3B). To assess if this synergy was broadly applicable in mycobacteriophages, we tested the susceptibility of the cluster O mycobacteriophage, Bora (Fig. 3C). Infection of Δrpf::RpfAB ^Mtb again yielded similar increased plaque-forming ability as observed with MontyDev, i.e., partial complementation of Δrpf (∼75%). However, the plaque-forming ability of Bora in Δrpf::RpfB ^Mtb was comparable to that of Δrpf::RpfAB ^Mtb (∼65% vs ∼75%), while the Δrpf::RpfA ^Mtb resulted in lower but similar levels of PFU/mL for both MontyDev and Bora (∼25%) (Fig. 3C).

To determine if the reduction in PFU/mL observed for the Δrpf mutant was due to an adsorption defect, we adapted an assay that monitors adsorption of phage over a period of 120 min, ²⁴ in mid and late log phase (OD_600 nm = 0.8 and OD_600 nm = 2.0, respectively), shown in Figure 4A. While MontyDev showed no adsorption differences between strains grown to late log phase, Bora exhibited some variation in adsorption under the same conditions. However, none of these differences reached statistical significance (Fig. 4B and C). To account for possible difference in cell wall structure for late log phase cells, we also tested adsorption at a lower optical density (OD_600 nm = 0.8, corresponding to mid log phase) and included a vortexing step to remove unbound bacteriophage. Under these conditions MontyDev showed a marginal reduction in adsorption with the Δrpf mutant after 60 min compared with WT, an effect that was reversed with genetic complementation however, these differences were not statistically significant (Fig. 4D). Bora displaced comparable levels of adsorption with the wild type (Fig. 4E).

OPEN IN VIEWER

FIG. 4.

Comparison of adsorption of mycobacteriophages to Mycobacterium smegmatis cells in mid and late log growth phase. (A) Schematic representation of the mycobacteriophage adsorption assay. Bacteriophages were incubated with bacterial cells at a MOI of 1:10 (1 mycobacteriophage:10 bacteria) and allowed to adsorb over a 120-min period. At various time points, samples were removed and the unbound bacteriophages in the supernatant were harvested and reinfected into the WT strain to quantify residual, unabsorbed bacteriophages. PFU were assessed by mixing the suspension with top agar and overlaid on 7H10 plates. Adsorption assessment of MontyDev (B) and Bora (C) in the WT, Δrpf mutant, and the Δrpf mutant complemented with rpfAB grown to late-log phase (OD_{600 nm} = 2.0) over a 120-min incubation period. Adsorption assessment of MontyDev (D) and Bora (E) with the WT, Δrpf mutant, and the Δrpf mutant complemented with rpfAB grown to mid-log phase (OD_{600 nm} = 0.8) over a 120-min incubation period. In these experiments a 5 second vortex step was included at each sampling point prior to centrifugation of the samples to remove loosely bound mycobacteriophages. The experiments were conducted in triplicate and the data are presented as percent (%) mycobacteriophage adsorbed over time. Error bars indicate standard error of the mean of three experiments. Graphs were generated using GraphPad Prism software, version 10.