Mycobacterium llatzerense, a waterborne Mycobacterium, that resists phagocytosis by Acanthamoeba castellanii

Nontuberculous mycobacteria (NTM) are environmental bacteria increasingly associated to public health problems. In water systems, free-living amoebae (FLA) feed on bacteria by phagocytosis, but several bacteria, including many NTM, are resistant to this predation. Thus, FLA can be seen as a training ground for pathogenic bacteria. Mycobacterium llatzerense was previously described as frequently associated with FLA in a drinking water network. The present study aimed to characterize the interactions between M. llatzerense and FLA. M. llatzerense was internalised by phagocytosis and featured lipid inclusions, suggesting a subversion of host resources. Moreover, M. llatzerense survived and even multiplied in presence of A. castellanii. Using a genomic-based comparative approach, twelve genes involved in phagocytosis interference, described in M. tuberculosis, were identified in the M. llatzerense genome sequenced in this study. Transcriptomic analyses showed that ten genes were significantly upregulated during the first hours of the infection, which could partly explain M. llatzerense resistance. Additionally, M. llatzerense was shown to actively inhibit phagosome acidification. In conclusion, M. llatzerense presents a high degree of resistance to phagocytosis, likely explaining its frequent occurrence within FLA in drinking water networks. It underscores that NTM should be carefully monitored in water networks to prevent human health concerns.


M. llatzerense is resistant to A. castellanii predation.
In a previous study, we showed the frequent co-occurrence of Mycobacterium llatzerense (80% to 90% of identified NTM sequences) in association with FLA isolated from a drinking water network 33 . This finding encouraged a better characterisation of M. llatzerense interactions with Acanthamoeba castellanii. In this study, environmental isolates of M. llatzerense were collected from endpoint sites of the above-mentioned drinking water network, provided by treated groundwater. Their resistance to A. castellanii was determined through a droplet test, for comparing the ability of bacterial isolates to grow in presence or absence of A. castellanii seeded on the agar medium. This ability was assessed for five M. llatzerense environmental isolates, as well as a strain of Mycobacterium septicum, moderately resistant to A. castellanii predation, and isolated from the same drinking water network. E. coli was used as a control, as it is a commonly used food source for FLA (Fig. 1). The five environmental M. llatzerense isolates developed similarly with or without the presence of A. castellanii lawn on the nutritive medium (Fig. 1). In comparison, the ability of M. septicum to grow was strongly impaired in the presence of A. castellanii, while E. coli growth was completely inhibited (Fig. 1). Based on these observations, M. llatzerense displayed a resistance to A. castellanii predation. As no difference in A. castellanii resistance were observed between M. llatzerense environmental isolates, one representative isolate, M. llatzerense EDP_4, was selected for the subsequent experiments.

M. llatzerense persists and grows in co-culture with A. castellanii. To confirm M. llatzerense
resistance ability, its fate within A. castellanii was investigated during co-culture lasting up to 96 hours. This time course was chosen as it represents the longer incubation time inducing less than 10% A. castellanii encystment in the non-nutritive PAS buffer, the medium used for infection experiments. After 4 h of infection, about 10% of A. castellanii were infected, and reached 40% after 16 h ( Fig. 2A). This infection rate was maintained without significant differences up to 96 h post infection. During the time course of the experiment, M. llatzerense viability was assessed using CFU counting. Viable mycobacteria were recovered at all time points of the experiments, demonstrating their resistance to A. castellanii phagocytosis (Fig. 2B). A significant growth was even observed, starting at 72 h (P < 0.05; unpaired t-test) and reaching a 1.5 log 10 growth at 96 h post-infection (P < 0.01; unpaired t-test) (Fig. 2B).

M. llatzerense is internalised by A. castellanii. To gain insights into the interaction between M. llatzerense
and A. castellanii, infected amoeba cultures were observed using transmission electron microscopy (TEM). M. llatzerense bacilli were mostly found intracellularly in vacuoles, which suggested an active internalisation process (Fig. 3 their cytoplasm, which were more frequently observed after 24 h and 72 h (Fig. 3B-D respectively). These inclusions were further confirmed to be of lipidic nature (supplementary Figure 1). In comparison, non-internalised bacilli did not harbour such inclusions (Fig. 3A). Lipid droplets from the host were also observed in the vicinity of M. llatzerense-containing vacuoles (Fig. 3C). Internalisation of bacteria by phagocytic cells such as FLA is mostly mediated by phagocytosis. Indeed, this was confirmed for M. llatzerense, as its internalisation was abrogated by the treatment of A. castellanii with cytochalasin D, a potent inhibitor of actin polymerisation (Fig. 4). Thus, it was shown that M. llatzerense internalisation by A. castellanii is indeed mediated by phagocytosis. This NTM species was able to persist intracellulary for at least 96 h.  Each ratio was calculated with respect to the same condition without A. castellanii, where no significant growth was observed for M. llatzerense alone (data not shown). Statistical tests were performed using multiple unpaired t-tests (P < 0.05*P < 0.01**).
Scientific RepoRts | 7:46270 | DOI: 10.1038/srep46270 M. llatzerense shares phagosomal resistance genes with M. tuberculosis. In order to screen for conserved protein encoding genes involved in phagocytosis resistance, M. llatzerense whole genome was sequenced. The collection of 72′ 881′ 926 high quality bases allowed to draw the 6,7 Mb draft genome of M. llatzerense, assembled in 235 contigs. This genome is in the size range of previously described Mycobacteria genomes, comprised between 3.21 Mb and 7.3 Mb (M. leprae and M. tusciae respectively; Table 1). The genome size of M. llatzerense EDP_4 appeared slightly larger than another previously published genome from the same species (strain CLUC14; 6.1 Mb), while harbouring a similarly high GC content of 66.6%, versus 66.3% for M. llatzerense CLUC14. Further analyses between the two assemblies indicated an average nucleotide identity of 98.12%, hence confirming they belong to the same species (Supplementary figure 2) 42 . M. llatzerense EDP_4 genome was screened for genes encoding proteins documented for their role in interfering with phagosome maturation in M. tuberculosis, using BLASTp searches. The M. tuberculosis H37rv genome was used for this comparative approach because it represents the most richly annotated genome to date within the Mycobacterium genus. In addition, gene functions were widely investigated using this strain, giving thus a highly valuable insight into the involvement of many genes in pathogenicity and intracellular survival 21 .
Among an initial set of thirteen coding genes, eleven highly similar, as well as one moderately similar genes were identified in M. llatzerense genome, and are described below (Table 2). Comparable results were obtained by using the M. llatzerense CLUC14 genome (Supplementary Table 2). Rv3707c, PPE10, Cut2, and GlyA1 were shown to be involved in in phagosome maturation blocking, as suggested by transposition mutagenesis in M. bovis Bacille Calmette-Guérin (BCG); mutant deleted for these genes were significantly enriched in acidified phagosome 43 . NdK of M. tuberculosis was shown to block phagosome maturation in murine macrophages through disruption of Rab5 and Rab7 GTPases, thus blocking phagosomes fusion with endosomal and lysosomal compartments 44 . Secreted phosphatases PtpA and PtpB were involved in mycobacterial pathogenesis; PtpA was notably characterised for inhibiting the recruitment of vacuolar-H + -ATPases machinery at phagosome membranes 25 . The SecA2 protein, part of the sec system, was also demonstrated essential for blocking phagosome maturation for M. tuberculosis infecting macrophages. PhoP and FbpA were both indirectly involved in blocking phagosomal maturation, as they regulate independently other virulence factors such as type 7 secretion systems (PhoP) or the cell wall lipid composition (FbpA) 23,45 . In addition, type 7 secretion system locus ESX-3 was identified in M. llatzerense genome, whose heterodimeric effector EsxG-EsxH was recently involved in disruption of M. tuberculosis-containing phagosomes addressing to the lysosome 24 . Phagosomal resistance genes are induced upon phagocytosis by A. castellanii. In order to investigate the potential role of these genes in M. llatzerense resistance to A. castellanii, their expressions were quantified by qRT-PCR at 0 h, 2 h, 4 h and 8 h post infection, corresponding to the early stage of M. llatzerense internalisation (Fig. 5). Overall, M. llatzerense internalisation significantly impacted the global transcriptional profile of selected genes (ANOVA, F = 85.608, df = 1-7, P = 0.001). After 2 h, only three of the selected genes were significantly upregulated, corresponding to secA2 (8.23 folds), fbpA (7.28 folds) and phoP (6.23 folds). After 4 h, significant upregulations of transcription levels were observed for 8 other genes, namely cut2, esxG, fbpA, phoP, ptpA, ptpB, rv3707c and secA2. After 8 h, all genes except glyA1 and ppe10 were significantly upregulated. Only genes that were upregulated after 2 h reached more than 10-times fold changes (fbpA, phoP and secA2; Fig. 5). In contrast, glyA1 and ppe10 did not show any significant change in transcription levels throughout the experiment. The early response of M. llatzerense to phagocytosis by A. castellanii is reflected by the stimulation of fbpA, phoP and secA2 transcription. After 4 hours of infection, M. llatzerense transcriptional response to internalisation was prominent, while a late response is observed after 8 h through the significant upregulation of esxH and ndK. Taken together, the transcriptional increase of conserved virulence factors, shared between M. tuberculosis and M. llatzerense, is in favour of their involvement in the resistance to A. castellanii phagocytosis.

M. llatzerense is not degraded by A. castellanii and resides in poorly acidified phagosomal compartments.
In order to strengthen the results depicting significant changes of M. llatzerense gene expression during the early infection stage, we aimed to characterise its intracellular survival. Following internalisation,  www.nature.com/scientificreports/ the resistance of M. llatzerense to A. castellanii degradation was assessed, by stimulating the autophagic machinery using rapamycin. Autophagy was described as a major pathway for blocking tuberculous mycobacteria proliferation in host cells, as its activation greatly affects mycobacteria survival 46,47 . M. llatzerense survival was monitored for up to four hours after rapamycin treatment. Bacterial viability significantly decreased at 0, 1 and 4 hours after rapamycin treatment, as half of the bacilli survived compared to control condition (Fig. 6A). While significant, even in condition where bacterial degradation through phagocytosis and autophagy were strongly stimulated, we did not reach M. llatzerense total elimination, confirming its strong ability to resist A. castellanii degradation.
In addition, the characterisation of M. llatzerense-containing phagosomes was undertaken, in order to correlate the pattern of gene expression with an observable phenotype. Therefore, the acidification of M. llatzerense-containing phagosomes was evaluated using a pHrodo dye, which act as a fluorescent indicator of acidic environment. When using alive M. llatzerense for infecting A. castellanii, 40.52 ± 2.65% of the bacterial cells were positive for pHrodo labelling (Fig. 6B, Supplementary figure 3). In contrast, infection of A. castellanii with heat-killed mycobacteria resulted in a significant increase of pHrodo labelled bacterial cells (94.95 ± 2.25%; P < 0.001, unpaired t-test). These results, coupled with previous observations at the transcription level, strongly suggest that M. llatzerense is actively blocking phagosomal maturation, ultimately resulting in M. llatzerense persistence in poorly acidified compartments.   Discussion NTM are frequently found in aqueous environments along with FLA, facing therefore a potentially strong predatory pressure, which suggests that NTM resistance stand as an evolutionary advantage to persist in such environments 7,10,48,49 . Accordingly, recent genomic findings indicated that mycobacteria may have maintained a sympatric life with protozoa such as FLA 50 . NTM, as environmental representatives of the Mycobacterium genus have been repeatedly reported to develop resistance to FLA predation 16,19,20 . As a fitting illustration of this statement, the recently characterised species M. llatzerense was recurrently found in drinking water, and in association with FLA. Therefore, we undertook the characterisation of resistance mechanisms employed by M. llatzerense to A. castellanii phagocytosis.
In the present study, the ability of M. llatzerense to persist and even grow in presence of FLA was described. As for other mycobacteria, A. castellanii internalised M. llatzerense through phagocytosis. Once internalised, M. llatzerense were found in vacuoles corresponding to phagosomes. The observed infectivity of M. llatzerense was comparable to those of M. marinum and M. avium infecting Acanthamoeba spp., as reported previously 26,51 . While the resistance of a wide range of NTM has been observed, their capacity to grow in presence of amoebae has been much less described. Nonetheless, it was possible to evaluate that M. llatzerense growth in presence of A. castellanii was comparable to this of M. avium infecting the same FLA host, while higher than the growth observed for M. fortuitum and M. marinum 52,53 .
Interestingly, intracellular M. llatzerense harboured many intracytoplasmic lipid inclusions. Lipid bodies can represent a form of energy storage, but are also thought to be involved in host inflammatory and immune responses 54 . Similarly to the observations described in this study, it was shown that M. marinum was able to actively subvert host lipid bodies when infecting D. discoideum 55 . Although this type of observation needs to be repeatedly observed among various mycobacteria, it is likely that host lipid subversion represents a shared mechanism among the genus Mycobacterium.
In order to explore the genomic basis of M. llatzerense resistance to A. castellanii, a comparative genomic approach was undertaken. M. llatzerense whole genome sequencing revealed a large-sized 6.7 Mb genome, which is in accordance with its environmental and generalist lifestyle 56 . Screening for protein encoding genes involved in phagosome maturation blocking underlined the presence of twelve candidates shared with M. tuberculosis H37rv (reviewed in ref. 21). During M. llatzerense internalisation by A. castellanii, an early increase in fbpA, phoP and secA2 transcription levels was observed. FbpA is supposedly involved in trehalose 6,6′ dimycolate (TDM) synthesis, an important surface glycolipid also designated as "cord factor" 45 . The significant upregulation of fbpA might lead to an enrichment of TDM, which was correlated with M. tuberculosis intracellular survival 57 . The early overexpression of phoP gene in M. llatzerense could suggest a role in blocking phagosome acidification. To support this hypothesis, M. tuberculosis mutants deficient for PhoP expression failed at blocking phagosome acidification, while still able to prevent phagolysosomal fusion 23 . The secA2 gene showed one of the earliest and strongest overexpression following internalisation. This gene, as part of the sec system, was recently suggested to be involved in secretion of acid phosphatase PtpA and PtpB 22,25 . In M. llatzerense the upregulation of SecA2 transcription was parallel to those of PtpA and PtpB, which is in accordance with the proposed hypothesis. Indeed, a thorough investigation will be required to confirm the role of the Sec system in PtpA and PtpB secretion. Transcription analysis of ndk highlighted a late uperegulation. In M. tuberculosis, NdK was shown to inhibit phagosomal fusion with early endosomes and lysosomes 44 . The activity of T7SS ESX3 variant was indirectly monitored through EsxG and EsxH transcription, which were significantly upregulated at 4 h and 8 h post infection, respectively. These genes, encoding an ESX3 heterodimeric effector, were linked to the inhibition of phagolysosome fusion 24 . Structural and functional analyses of EsxG-EsxH heterodimer also suggested a role in iron and zinc scavenging [58][59][60] . Cut2 and Rv3707c transcription was also significantly upregulated at 4 h post infection, however their roles are not yet clearly elucidated 43 .
In contrast to the significant changes observed in transcription levels of the other selected genes, both glyA1 and ppe10 were not differentially transcribed throughout the experiment. The roles for these genes have been identified by the characterisation of mutants unable to prevent phagosomal acidification 43 . PE-PPE proteins are thought to be involved in Mycobacterium pathogenicity 61 , therefore the expression profile of PP10 best BLAST hit identified in M. llatzerense seemed surprising. It is however worth noting that only a moderate sequence similarity (40% identity) was identified between PPE10 from M. tuberculosis and the corresponding protein in M. llatzerense. Furthermore, M. llatzerense EDP_4 was shown to harbour a total of 27 PE-PPE genes in its genome. It would be thus relevant to analyse the expression of these other genes during infection, in order to evaluate their importance in FLA phagocytosis resistance. So far the secreted hydroxymethyltransferase GlyA1 has only been described in M. bovis BCG for blocking phagosome maturation. As no significant difference in expression of GlyA1 was observed in our study, we cannot confidently conclude on its role in the survival of M. llatzerense infecting A. castellanii. Transcription analyses indeed provide strong arguments in favour of the involvement of significantly modulated genes. However, it has to be kept in mind that functional analyses are required to confirm the importance of such genes in M. llatzerense resistance to A. castellanii phagocytosis, such as phenotypic characterisation of M. llatzerense survival within FLA when impaired for selected gene's expression, through mutagenesis or interference experiments.
To complete transcription analyses, an investigation at the phenotypic level was performed. Most notably, a particular focus was given on the potential role of the autophagic machinery in modulating the infection. Autophagy, firstly described as a process maintaining cellular homeostasis through nutrients and organelles recycling, is also characterised as an antimicrobial mechanism, allowing the degradation of intracellular pathogens 62 . The autophagic machinery has been shown to play a role in the degradation of intracellular tuberculous mycobacteria, as it promotes phagosome maturation 46 . Unsurprisingly, autophagy induction in M. llatzerense -infected A. castellanii resulted in a significant decrease of bacterial survival. However, efficient M. llatzerense killing appeared nuanced, as 50% of intracellular mycobacteria were still viable after autophagy induction. The partial survival of M. llatzerense following autophagy activation might reflect its inherent resistance to advert conditions such as acidic and oxidative stresses. This finding is in accordance with a previous study describing M. avium ability to multiply within phagolysosomes, while M. tuberculosis was only able to persist in such condition 63 . In addition, by using a pH-sensitive fluorescent dye, it was possible to put in evidence that live mycobacteria were significantly less found in acidified compartments than heat-killed mycobacteria. This result strengthened the body of evidence regarding the ability of M. llatzerense to actively block A. castellanii phagocytosis process, and corroborate the transcription profiles of the analysed genes.
In conclusion, the work presented here provides insights into mechanisms employed by M. llatzerense, an environmental NTM, to resist A. castellanii predation. M. llatzerense was shown to infect efficiently A. castellanii through phagocytosis, remained viable and was able to grow for up to 96 h. M. llatzerense intracellular localisation within A. castellanii indicated persistence in vacuoles, as well as a potential subversion of host lipid metabolism. A comparative genomic approach identified twelve conserved virulence factors, shared between M. llatzerense and M. tuberculosis. Analyses of transcriptional profiles highlighted that all selected genes except glyA1 and ppe10 were significantly upregulated during the early stage of A. castellanii infection. This led to suggest that M. llatzerense is very likely able to actively hamper phagosomal acidification, while resisting to autophagy-induced advert conditions. Taken together, this study clearly defined M. llatzerense as an amoeba resistant bacteria. M. llatzerense was shown to establish an elaborate and efficient strategy not only to survive but to take advantage as well of the presence of FLA for its growth. This capability might very well explain the frequent recovery of this NTM in several drinking water networks.

Methods
Isolation and culture of microorganisms. NTM were isolated from Paris drinking water network.
Briefly, water samples (1 L) were filtered through a 0.45 μ m cellulose membrane, sonicated for 10 minutes, and decontaminated using 3% lauryl sulfate and 1% NaOH, as described previously 64 . Bacterial pellets were spread onto Lowenstein Jensen medium and incubated at 30 °C. Colonies were subcultured on Middlebrook 7H10 agar plates. To ensure the purity of isolated strains, isolates were decontaminated once again according to the Löwenstein method as described previously, and spread onto Middlebrook 7H10 plates supplemented with 10% OADC (Oleic acid, Albumin, Dextrose, Catalase; Becton-Dickinson) 65 . Ziehl-Neelsen cold staining was performed in order to confirm the presence of acid-fast bacilli. Briefly, heat-fixed samples were stained for 3 h in cold carbol Fuchsin. Samples were rinsed and treated using 25% sulphuric acid treatment for 45 s followed by absolute ethanol for 5 min. A counterstaining was performed by incubating samples with 0.2% methylene blue for 2 min. All mycobacteria isolates were maintained on Middlebrook 7H10 plates at 30 °C. Escherichia coli strain K-12 strain was grown and maintained on LB-agar (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, 15 g/L agar) plates at 37 °C. Acanthamoeba castellanii strain ATCC 30234 was cultured axenically in PYG medium (20 g/L proteose peptone, 1 g/L yeast Extract, 1 g/L sodium citrate, 0.1 M glucose, 0.4 mM CaCl 2 , 4 mM MgSO 4 , 2.5 mM Na 2 HPO 4 , 2.5 mM KH 2 PO 4 , 50 μ M Fe(NH 4 ) 2 (SO 4 ) 2 , pH 6.5) at 30 °C. Droplet tests were performed by seeding homogeneously 4.10 6 A. castellanii trophozoïtes on Middlebrook 7H10 agar supplemented with 10% (v/v) OADC (120X120 mm Petri dishes). Mycobacteria suspensions were prepared in PAS buffer, supplemented with 0.005% triton X-100 to avoid clumping, and homogenised by three passages using a 27 gauges needle. The suspension was spotted (from 1 × 10 8 to 1 × 10 3 bacteria/mL) on agar medium in presence or absence of A. castellanii as prepared previously, and incubated at 30 °C or 37 °C for optimal bacterial growth.

Co-cultures and phagocytosis assays.
For the assessment of mycobacteria viability, A. castellanii monolayers in 12-wells plates were infected at a MOI of 1 for 16 h. Supernatant was discarded, cells washed twice and resuspended in 1 mL of fresh PAS buffer at 0, 24, 72 and 96 hours. Trophozoites were lysed and mycobacteria enumeration was performed as described above.
For evaluating the number of infected cells, A. castellanii monolayers in 12-wells plates were infected at a MOI of 1. Trophozoites were washed three times using PAS buffer before processing. Cells were centrifuged at 800 g for 5 minutes, supernatant was discarded, and the pellet resuspended in 50 μ L of PAS, spread on a glass slide, heat-fixed and processed for Ziehl-Neelsen staining. Percentages of infected amoebae were calculated on the basis of at least 200 amoebae cells par slides, in a minimum of five different fields. In order to determine internalisation mechanisms of M. llatzerense, A. castellanii monolayers in 12-wells plates were infected at a MOI 10 for 4 h, in co-incubation with cytochalasin D at 100 μ M in PAS buffer. For induction of autophagy A. castellanii trophozoïtes were infected at a MOI of 10 for 4 h, and treated for 1 h using 1 μ M or 10 μ M rapamycin in PAS.
Evaluation of phagosomal acidification. M. llatzerense suspension prepared as described previously were pre-labelled using pHrodo TM Red succinimidyl ester (ThermoFischer Scientific) following manufacturer recommendations, but excluding the methanol washing step. Briefly, bacterial suspensions were incubated for 1 h in 0.1 M sodium bicarbonate buffer containing 20 μ M pHrodo TM Red succinimidyl ester, in the dark. Bacteria were then pelleted by centrifugation at 12000 g for 5 min, and washed twice in PAS buffer supplemented with 0.005% triton X-100. In one condition, mycobacteria were killed before the initial incubation by heating the suspension at 85 °C for 15 min, in order to assess the importance of bacterial viability in subsequent analyses. A. castellanii monolayers were infected by either alive of heat-killed labelled mycobacteria at a MOI of 10, in 6 wells plates. After 2 h of incubation, trophozoites were detached, fixed with 2% paraformaldehyde for 15 min in the dark. Cells were pelleted by centrifugation at 800 g for 10 min, and resuspended in a 30 μ L of SlowFade Diamonds antifade mountant (ThermoFischer Scientific) with Hoechst at a concentration of 400 ng/mL. Samples were examined using a confocal laser scanning microscope (SP8, Leica). To estimate the fraction of M. llatzerense residing in acidified phagosomes, a ratio was calculated based on the total number of bacteria (as detected by Hoechst staining) divided by the number of bacteria being labelled as well by pHrodo, which fluoresces only in acidic environment.
Nucleic acid extraction. Isolated M. llatzerense colony was suspended in phosphate buffer saline, subsequently bead-beaten in tubes containing 500 mg of small diameter glass beads (100 μ m) and 4 glass beads of 2 mm diameter (Sigma) using Fastprep apparatus for 30 s, put on ice for 1 min, and bead-beaten again using the same parameters. The suspension was then processed for DNA extraction using NucleoSpin Tissue kit (Macherey-Nagel), following manufacturer recommendations for bacterial DNA extraction.
For RNA extraction cells were collected, suspended in 400 μ L of resuspension solution (sterile ultrapure water with 10% Glucose, 12.5 mM Tris, 5 mM EDTA, pH 7.6) and transferred to a tube containing 400 mg of small diameter glass beads (100 μ m) and 500 μ L of acid phenol. Samples were bead-beaten as described above and centrifuged for 5 min at 14000 g, 4 °C. Supernatants were kept on ice and transferred into clean tubes. 1 mL of TRIzol ® reagent (Life Technology) was added to each sample, and incubated at room temperature for 5 min.
100 μ L of chloroform/isoamyl alcohol (24:1 v/v) was added to each sample, shaken vigorously, incubated for 1 min and centrifuged for 5 min at 14000 g. Aqueous phase was transferred into a new tube, treated again with 200 μ L of chloroform/isoamyl alcohol as described above. The collected aqueous phase was washed with 500 μ L of isopropanol for 1 h on ice and centrifuged at 16000 g for 30 min. Supernatants were discarded and pellets washed with 1 mL of 70% ethanol, centrifuged 15 min at 16000 g. Pellets were air dried and suspended 50 μ L of ultrapure water, 7.5 mM sodium acetate, and 150 μ L of absolute ethanol. Samples were incubated at − 20 °C overnight and centrifuged for 30 min at 16000 g. Pellets were rinsed as described above, let air dried, and suspended in 30 μ L of ultrapure water. RNA samples were subsequently treated using TURBO TM DNase (Life Technologies) according to manufacturer recommendations, purified once again using phenol/chloroform and quantified by spectrophotometry. RNA were aliquoted, stored at − 80 °C, and one aliquot of each condition was used immediately for reverse transcription, using GoScript TM Reverse transcription System (Promega). The resulting cDNA were used for downstream quantification by qPCR.
DNA amplification and quantification. Amplifications of targeted loci from total DNA were done by PCR. Mixtures were prepared in a final volume of 25 μ L, containing 5 μ L of 5x GC buffer, 400 nM of each deoxynucleotide triphosphate, 1,5 mM MgCl2, 500 nM of each primer, 0.5 U of Phusion polymerase (Thermo Scientific), 2.5 μ L of template DNA and PCR grade water in sufficient quantity for 25 μ L. Primers U341F (5′ -CCTACGGGAGGCAGCAG-3′ ) U926R (5′ -CCGTCAATTCMTTTRAGT-3′ ) and Myco F (5′ -GGCAAGGTCACCCCGAAGGG-3′ ) Myco R (5′ -AGCGGCTGCTGGGTGATC-3′ ) were used for partial amplification of 16 S and rpoB genes, respectively 66 . PCR were carried out in a Master Cycler Thermocycler (Eppendorf). PCR mixtures were subjected to an initial denaturation at 98 °C for 30 sec, 35 cycles of denaturation at 98 °C for 10 sec, hybridization at 56 °C for 10 sec and elongation at 72 °C for 20 sec. A final extension step was applied at 72 °C for 5 min. Amplicons were checked on 1.5% agarose gel and stained with Midori Green (Nippon Genetics). cDNA was quantified by quantitative PCR (qPCR) using primers pairs targeting specific loci (Supplementary Table 1). Mixtures for quantification were performed in a final volume of 20 μ L, containing 10 μ L of 2X SYBR ® Green PCR Master Mix (Life Technologies), 250 nM of each primer, 2 ng of cDNA, and molecular grade water in sufficient quantity for 20 μ L. Experiments were carried out in a Viia7 real time thermocycler (Life Technologies), consisting in an initial denaturation step at 95 °C for 10 min, 45 cycles of denaturation at 95 °C for 15 sec, annealing for 15 sec at 63 °C to 65 °C, and elongation at 72 °C for 20 sec. Fusion curves were collected for each experiments. Quantification of transcripts were performed by calculating fold changes using 2 −ΔΔCt method, corrected by qPCR efficiency for each gene, and normalised to the control condition at 0 h, as well as the control gene expression of 16S rRNA 67 . Statistical significance was assessed using multiple t-tests coupled with the Holm-Sidak method, as implemented in Graphpad Prism 6.
Whole genome sequencing and annotation. Once identification and purity of mycobacterial isolates were confirmed by both 16 S and rpoB sequencing, total DNA were processed for whole genome sequencing, using GS junior 454 pyrosequencer (Roche). Library preparation, emulsion PCR and sequencing were performed following manufacturer's protocols. Raw sequences were then assembled using GS denovo Assembler software (Roche). The assembly was uploaded for annotation using RAST online annotation tool 68,69 . Sequences resulting from the assembly are available on the NCBI database under accession number ASM186516v1.
Electron microscopy. The localisation of M. llatzerense inside A. castellanii was investigated using transmission electron microscopy following the protocol described in 70 . Briefly, amoebae were carefully collected at 4 h, 24 h and 72 h after infection by M. llatzerense at a MOI of 1, and fixed in 3% glutaraldehyde and osmium tetroxide. Cells were dehydrated in successive baths of 50%, 70%, 90% and 100% acetone. Cells were embedded in araldite resin, and ultrathin section of 60 nm were cut and stained with uranylacetat and lead salts according to Reynolds. Ultrathin sections were observed using JEOL 1010 transmission electron microscope operating at 75 kV.