A protocol for culturing environmental strains of the Buruli ulcer agent, Mycobacterium ulcerans

Contaminations and fastidiousness of M. ulcerans may have both hamper isolation of strains from environmental sources. We aimed to optimize decontamination and culture of environmental samples to circumvent both limitations. Three strains of M. ulcerans cultured onto Middlebrook 7H10 at 30 °C for 20 days yielded a significantly higher number of colonies in micro-aerophilic atmosphere compared to ambient atmosphere, 5% CO2 and anaerobic atmosphere. In a second step, we observed that M. ulcerans genome uniquely encoded chitinase, fucosidase and A-D-GlcNAc-diphosphoryl polyprenol A-3-L-rhamnosyl transferase giving M. ulcerans the potential to metabolize chitine, fucose and N-acetyl galactosamine (NAG), respectively. A significant growth-promoting effect of 0.2 mg/mL chitin (p < 0.05), 0.01 mg/mL N-acetyl galactosamine (p < 0.05), 0.01 mg/mL fucose (p < 0.05) was observed with M. ulcerans indicating that NAG alone or combined with fucose and chitin could complement Middlebrook 7H10. Finally, the protocol combining 1% chlorhexidine decontamination with micro-aerophilic incubation on Middlebrook 7H10 medium containing chitin (0.2%), NAG (0.01%) and fucose (0.01%) medium and auto-fluorescence detection of colonies allowed for the isolation of one mycolactone-encoding strain from Thryonomys swinderianus (aulacode) feces specimens collected near the Kossou Dam, Côte d’Ivoire. We propose that incubation of chlorhexidine-decontaminated environmental specimens on Middlebrook 7H10-enriched medium under micro-aerophilic atmosphere at 30 °C may be used for the tentative isolation of M. ulcerans strains from potential environmental sources.

cultures are usually positive within a 6-12-week incubation, but a much longer incubation period of up to nine months may be necessary to obtain isolates, illustrating the fastidiousness of this microbe 11,12 .
The fact that thousands M. ulcerans isolates have been made from clinical sources sharply contrasts with the fact that several attempts to culture M. ulcerans from many specimens of flora and fauna remained unsuccessful 13,14 . However, numerous PCR-based investigations indicated that potential reservoirs or host carriers were localized in aquatic environments where M. ulcerans may be able to colonize different ecological niches eventually scattered along a food chain [15][16][17][18][19] . Culture of M. ulcerans from environmental samples is tedious and understanding the ecology of M. ulcerans has been severely hampered by the extreme difficulty of culturing the organism directly from the environment 14,20,21 . Cultures of collected diverse samples (water, soil, fish, rodents, biting flies, reptiles) from Buruli ulcer endemic areas failed to yield M. ulcerans a long time ago 13 , though testing of samples by molecular biology found M. ulcerans DNA 7,[22][23][24][25][26][27][28][29][30] . Finally, only one environmental M. ulcerans (M. ulcerans 00-1441 from a Buruli ulcer endemic area in Benin, West Africa) isolate has been firmly confirmed on Löwenstein-Jensen medium after 15-day of incubation in BACTEC 12b broth and three successive passages in mouse footpad P1, P2 and P3 for nine months, six months and 12 months, followed by culture on Löwenstein-Jensen for two months 21 . Three additional reported strains included two IS2404-PCR positive strains from two samples of aquatic plants and two wild aquatic insects collected in a Buruli ulcer endemic area of Côte d'Ivoire 31,32 and two M. ulcerans strains from moss and soil in Ghana 33 . None of these strains have been deposited in public collection.
The fact that only a few environmental isolates have been made after such a long experiment, suggests that contamination by fast-growing bacteria or mycobacteria and fungi of the environmental samples along with poorly appropriate culture media limited the isolation of M. ulcerans from environmental sources. Also when contaminants are present in the sample, they limit and inhibit the growth of M. ulcerans. Contaminants not eliminated by the decontamination method eventually rot or transferring the culture medium rendering impossible the incubation of cultures on a long period necessary for the isolation of M. ulcerans. There is thus a need to develop new protocols and innovative media to improve the recovery of M. ulcerans in primary culture from environmental sources in order to assess the viability of the pathogen in these sources 34 . The availability of the complete genome sequence of M. ulcerans is a boon to better known metabolic activities which could support the development of innovative culture media, as previously reported for some other fastidious pathogens 35 .
In the perspective of achieving a culture-based field investigation of M. ulcerans, we aimed at improving the decontamination of samples along with the composition of culture media in order to optimize the chance of recovering additional environmental M. ulcerans isolates.
Cultures at 37 °C. remained negative during the same period of incubation.  (Fig. 3).
After 1% chlorhexidine decontamination of freshwater sample, the culture onto sheep-blood Columbia agar remained sterile without contamination. The same freshwater sample mocked-inoculated with M. ulcerans strain (CU001 and ATCC33728 separately) and chlorhexidine-decontaminated yielded autofluorescent colonies of M. ulcerans, starting at day 15 after inoculation onto the chitin, fucose and NAG growth promoters media. Colony counting was done on day 30 (Fig. 4). Several colonies obtained during this step were confirmed by RT-PCR  Detecting IS 2404, IS 2606 and KR-B targets in environment samples, Côte d'Ivoire. While the negative controls remained negative and all the samples were free of PCR inhibition, 5/12 water without plant debris (41.66%) were positive for KR-B gene, insertion sequence IS2404 and IS2606; five (41.66%) were positive for KR-B and IS2606; one was positive for IS2404 and IS2606 and one was negative. Among the eight water with plants debris, two (25%) were positive for KR-B and IS2404; one was positive for IS2404 and IS2606; one was positive for IS2404 and one was positive for IS2606; and three were negative. No Thryonomys swinderianus's aulacode feces were definitely found positive according to our criteria while one was positive for IS2404 and IS2606, one was positive for IS2404, one was positive for IS2606 and eight were negative (Table 1). Further, the calculated values for ΔCt (IS2606 -IS2404) from PCR-positive environmental samples were ≤ 3.32 (95% CI = 0.43-1.70). These low values suggest that environmental mycobacteria here detected to harbor the two IS2606 and IS 2404 insertion sequences, are not M. ulcerans which is acknowledged to harbor two orders of magnitude more copies of IS 2404 than IS2606 [ref. 7 ].
Culturing M. ulcerans micro-colony in environment samples, Côte d'Ivoire. The culture of the 31 samples decontaminated with 1% chlorhexidine yielded one auto-fluorescent micro-colony isolated from one IS2404-positive aulacode feces sample after 45-day incubation on the DZ medium at 30 °C under a micro-aerophilic atmosphere condition. This unique micro-colony yielded Ziehl-Neelsen -positive bacilli identified as a MPM after the RT-PCR positivity for KR-B, IS2404 and IS2606 in the presence of negative controls (Fig. 5). However, calculated value for ΔCt (IS2606-IS2404) of 2. 97 was not in favor that the colony was of M. ulcerans, but rather in favor of other mycobacteria carrying the three detected molecular targets 7 . Sub-culturing onto the DZ medium of the micro-colony material remaining after identification failed.

Discussion
We here report on a culture protocol for the tentative recovery of viable M. ulcerans mycobacteria from environmental sources. Temperature of incubation is a crucial point for the culture of mycobacteria as previously reported 36 . Indeed, M. ulcerans strains have optimal growth between 28-32 °C; they are very sensitive to higher temperatures, a temperature of 41 °C over a period of 24 hours kills more than 90% of the bacilli 13 . This observation gives indications for the storage of environmental samples for M. ulcerans culture in endemic countries where the ambient temperature may reach values in-between 37 °C-45 °C, suggesting that samples should be stored at lower temperatures into any appropriate transport medium before inoculation. We then observed that growth of the M. ulcerans strains here investigated was significantly more rapid in micro-aerophilic atmosphere than in ambient atmosphere and in a 5% CO 2 -enriched atmosphere. It was previously suggested that M. ulcerans is capable of growth under aerobic but not anaerobic conditions 3 and to survive anaerobic conditions 37 . During our experiment we observed a significant growth of the three strains of M. ulcerans cultured under anaerobic conditions after day 20 of incubation. Indeed, breaking anaerobic atmosphere may occur during which a suitable atmospheric condition for the growth of mycobacteria was created during colonies count and/or pocket replacement every five days. Further, a possible activity of the cydA locus in M. ulcerans may sustain the ability of this strains to survive under low-oxygen conditions 37 .
To develop an innovative culture medium for improve tentative isolation and growth of M. ulcerans, we thought that comparative genome analysis reveal unique metabolic features and clues to enrich a Middlebrook 7H10 medium. Indeed, this approach has been successfully used to design a new culture medium for the fasditious pathogen Tropheryma whipplei, another Actinobacterium 35 . Accordingly, we incorporated chitin, fucose and N-acetylgalactosamine into a Middlebrook 7H10 medium in order to experimentally test their effect on the growth of M. ulcerans. The experimental data were authenticated by the negativity of the negative controls used in every experiment and the reproducibility of data over three different M. ulcerans strains culture and triplicate experiments. We observed that the N-acetyl-galactosamine-enriched Middlebrook 7H10 medium and three growth-promoters-Middlebrook 7H10 medium yielded similar results indicative that both media could be used for tentative isolation of MPM including M. ulcerans from environmental sources. Chitin is the second most abundant organic and renewable source in nature, after cellulose 38 . This linear homopolymer can be hydrolyzed at β-1,4-linkages by the enzymatic action of glycoside hydrolase enzymes, the chitinases (E.C. 3.2.1.14) and the N-acetylglucosaminidases (E.C. 3.2.1.52) 39,40 . Chitin is found in the structure of fungi, crustaceans (crabs, lobsters), insects, mollusks, cephalopods, fishes such as zebrafish (Danio rerio) and amphibians [41][42][43] . The derivatives of chitin play a crucial role in the interaction between higher plants and symbiotic bacteria; suggesting that chitin synthesis may serve roles other than the production of skeletal material 42 . N-acetyl galactosamine has been recognized as a minor covalently-bound amino sugar component of the cell wall of some slow-growing mycobacteria and orthologs of polyprenyl-phospho-N-acetyl-galactosaminyl synthase (ppgS), which are found in the genomes of slowly-growing mycobacteria including M. bovis, M. bovis BCG, M. leprae, M. marinum and M. avium subsp. paratuberculosis, as well as in M. abscessus; but not in the genomes of other rapidly growing Mycobacterium species such as Mycobacterium smegmatis 44,45 .
Chitinases are chitin-degrading enzymes belonging to the glycoside hydrolase family 18 (GH18) and 19 (GH19) 39 (www.cazy.org). They act in a synergistic to perform the complete enzymatic hydrolysis of chitin to  46,47 . The GH18 family is widely distributed in all kingdoms, including viruses, bacteria, plants, fungi and animals 38 . Bacterial chitinases and chitin-binding proteins (CBPs) play a fundamental role in the degradation of the ubiquitous biopolymer chitin, and the degradation products serve as an important nutrient source for marine-and soil-dwelling bacteria [48][49][50][51] . M. ulcerans genomes encode for a GH18 compatible with a putative chitinase activity 3 . N-acetylglucosaminidases belong to glycoside hydrolase family 20 (GH20) 39 . α-l-Fucosidases are enzymes involved in metabolism of α-l-fucosylated molecules, compounds with a fundamental role in different life essential processes including development 52 . These enzymes play a fundamental role in the degradation of the ubiquitous biopolymer 39,46,50 . The degradation products serve as an important nutrient source for bacteria in the nature 51 . M. ulcerans may obtain energy and carbon from the degradation of plant saccharides which were demonstrated to stimulate M. ulcerans growth in vitro 14,53 . It was recently shown that chitin promoted growth of M. ulcerans 43 . Fucose was here tested after we observed that M. ulcerans contained an alpha-L-fucosidase cytoplasmic protein involved in carbohydrate transport and metabolism 39 .
It was proved that some green algae extracts stimulate the growth of M. ulcerans 14 . In a later study however, the growth of M. ulcerans was indirectly observed by using quantitative PCR kinetics, instead of the simple observation of growing colonies as reported here 43 .
The doubling time obtained with the culture of M. ulcerans strains onto each growth promoter was less than two days and was in agreement with previously reported values. The doubling time was estimated to be of approximately 36 h 54 , 1-2 days 55 and 3.3 ± 0.56 days 14 . In the Dubos medium (a liquid medium), the doubling time was less than 48 hours at 33 °C 56 and 44 hours in the Dubos medium without serum 57 . A doubling time estimated to be between three and four days has been reported 58 . In mice, the doubling time was approximately 3.5 days 59-61 . The discrepancy may be first due to the strain of M. ulcerans, secondly to the calculation method 57 or the type of culture medium. In contrast, M. marinum has a doubling time of 6-11 hours 3 .
Culture onto 5% sheep-blood Columbia agar of freshwater sample after chlorhexidine decontamination remained sterile so, we deduced that the chlorhexidine decontamination method has been effective. Culture onto growth promoters' media of chlorhexidine decontaminated freshwater sample mocked-inoculated with M. ulcerans strains allowed to isolate M. ulcerans colonies confirmed by RT-PCR.
Here, IS2404, IS2606 and KR-B molecular targets were detected by RT-PCR in water, plants debris and aulacodes feces, all collected in Côte d'Ivoire. The fact that the reproducible Ct values we obtained for IS2404 and IS2606 are not in line with those expected from M. ulcerans pure DNA, suggests these natural specimens do not contain purely M. ulcerans but rather other mycobacteria harboring these insertion sequences in variable numbers. M. ulcerans's DNA has been detected previously in two feces from the aulacode collected in Côte d'Ivoire 22 and in small mammal (Mastomys) in Ivory Coast 27 suggesting that these animals may shelter and vehicle M. ulcerans.
Growth promoters allow the isolation of micro-colony which was positive for IS2404, IS2606 and KR-B. However, direct sub-culture onto the growth-promoters for more biological material failed. This failure could be explained by the very low inoculum remaining for the culture after we realized Zielh-Neelsen staining and the RT-PCR for the identification of the micro-colony. Our positive samples were collected in a Buruli ulcer endemic region in the centre of Côte d'Ivoire, closely related but different from a site where we previously PCR-amplified M. ulcerans DNA in the feces of aulacodes 22 . Therefore, this interesting yet limited observation warrants further field studies for confirmation.

Conclusions
A 1% chlorhexidine decontamination and addition of appropriate concentration of N-acetyl galactosamine or combination of chitin, N-acetyl galactosamine and fucose to the standard Middlebrook 7H10 culture medium promoted the growth of M. ulcerans under microaerophilic atmosphere at 30 °C. These protocols are proposed as a first-line protocol for the tentative isolation of M. ulcerans strains during field campaigns in Buruli ulcer endemic areas.

Methods
Ethics statement. The study has been conducted with collection references strains of M. ulcerans and no experiment or test has been performed on patients or/and animals.
M. ulcerans strains. Three strains of M. ulcerans isolated from different geographic origins were used throughout the study. M. ulcerans strain Cu001 (a gift from Prof Vincent Jarlier, Centre National de Référence des Mycobactéries, Paris France) was from Côte d'Ivoire 62 , M. ulcerans ATCC 33728 isolated from Japan and M. ulcerans ATCC 25900 belongs to the Borstel collection (Schröder 5392) and was probably isolated in Africa. The identification of these three strains were ensured by rpoB gene sequence analysis prior to experiments 63 . M. ulcerans strains were sub-cultured at 30 °C onto Middlebrook 7H10 agar medium supplemented with 10% (v/v) oleic acid/albumin/dextrose/catalase (OADC) (Becton Dickinson, Sparks, MD, USA) and 0.5% (v/v) glycerol (Sigma-Aldrich, Lyon, France) in a microaerophilic atmosphere until we have enough colonies to prepare inoculum. Then, a mycobacterial suspension was prepared by placing a loopful of colonies in a tube containing sterile phosphate buffered saline (PBS, pH 6.5) and sterile glass beads. The tube was vigorously vortexed in order to separate any bacterial aggregates and adjusted with PBS in setting the mycobacterial concentration of inoculum to 0.5 McFarland standards in order to obtain a final suspension containing 10 7 acid-fast bacteria (AFB)/mL using a turbidimeter (Biolog Inc., Hayward, U.S.A). This suspension was shown to be free of clumps by microscopic examination after Ziehl-Neelsen staining. Then, a 10 6 AFB/mL working suspension was prepared.
In all further experiments, colonies were observed and count by using a MZ-FLIII fluorescence microscope (Leica, Nanterre, France) equipped with a GFP filter and an ICA digital camera (Leica) to detect mycobacterial auto-fluorescence as previously described 64  Testing atmosphere conditions for culturing M. ulcerans. The first step consisted in the culture of M. ulcerans under different conditions. For each M. ulcerans strain, a 100 µL-volume of a 10 5 AFB/mL suspension corresponding to 10 4 AFB was cultured in triplicate onto Middlebrook 7H10 medium supplemented with 10% (v/v) OADC and 0.5% (v/v) glycerol then, incubated at 30 °C under four different atmospheric conditions. The atmospheric growing conditions were artificially created by the use of sealed plastic pouch that can hold ten Middlebrook 7H10 medium poured in 55-mm diameter Petri dishes. The conditions of culture to produce standard, microaerophilic (5% oxygen), anaerobic (total absence of free oxygen and 8-14% CO 2 ) and 5% CO 2 rich atmosphere conditions with a 15% final concentration of oxygen, all in closes pouch were made respectively by simple culture method, microaerophilic condition using CampyGen Compact (OXOID Ltd, Basingstoke, Hampshire, England), Anaerobie Poche System/AnaeroGen Compact (OXOID Ltd, Basingstoke, Hants RG24 8PW, UK) and CO 2 Rich atmosphere using CO 2 Gen Compact (OXOID Ltd). All inoculated cultures and negative control inoculated with sterile PBS were incubated at 30 °C. In parallel, we cultured onto Middlebrook 7H10 medium supplemented with 10% (v/v) OADC and 0.5% (v/v) glycerol, the three strains of M. ulcerans (10 4 AFB per Petrish dish), incubated at 37 °C. Cultures were all examined at day 5, day 10, day 15, day 20 and day 30 by a MZ-FLIII fluorescence microscope. Counting of fluorescent colonies was performed using the Leica Application Suite software.

Growth-promoters.
In a second step, we searched for genes encoding chitinase, N-acetyl galactosaminase (NAG) and fucosidase in completely sequenced Mycobacterium genomes (NCBI Gene Bank, last accessed in February 2016). To estimate the copy number of these genes, we considered protein functional description of the respective strains as documented in the gff files of NCBI Gene Bank bacterial genome repository. In each completely sequenced Mycobacterium genome, we counted the number of genes using the key terms "chitinase", "cellulase", "fucosidase" in their protein functional annotation column.
A total of 109 Mycobacterium genomes were screened in this analysis. The presence of one of these three genes was detected in 70 genomes and the presence of the three genes at once was detected only in the M. ulcerans genome.
Therefore, these three substances were tested as growth promoters for M. ulcerans. N-acetylgalactosamine (Sigma-Aldrich) and fucose (Sigma-Aldrich) were dissolved in sterile distilled water (Sigma-Aldrich) at a concentration of 1 mg/mL. They were then prepared on Middlebrook 7H10 agar medium supplemented with 10% (v/v) OADC and 0.5% (v/v) glycerol (Sigma) at a concentration of 0.1 mg/mL in the first step followed by a preparation with a final concentration of 0.01 mg/mL in a second step. These media were poured into 55-mm Petri dishes (Gosselin, Borre, France). Solubilization of chitin (Sigma-Aldrich) was achieved as previously described with few modifications by slowly dissolving chitin in 37% concentrated hydrochloric acid "HCl" (Sigma-Aldrich) 66 (Supplementary material). The obtained colloidal chitin was then dissolved into Middlebrook 7H10 medium supplemented with 10% (v/v) OADC and 0.5% (v/v) glycerol at a concentration of 0.2 mg/mL and this medium was poured into 55-mm Petri dishes. The purity of chitin, fucose and NAG delivered by Sigma-Aldrich was controlled by MALDI-TOF-MS before incorporation into the enriched culture media. For each of the three reagents, one microliter of the supernatant was spotted per spot onto the polished-steel MSP 96 target plate (Bruker Daltonics) and allowed to dry at room temperature. Each dry spot was then overlaid with 1 μL of matrix solution (saturated α-cyano-4-hydroxycinnamic acid in 50% acetonitrile and 2.5% trifluoroacetic acid) (Bruker Daltonics). The plate was air-dried for 5 minutes and loaded for manual processing into the MALDI-TOF mass spectrometer (delay: 170 ns; ion source 1 (IS1) voltage: 20 kV; ion source 2 (IS2) voltage: 16.65 kV; lens voltage: 7.20 kV; mass range: 0 kDa to 1 kDa) taking into account the known molar mass of reagents. Molar mass was identified by the m/z (mass/charge) parameter on spectra. The specific peaks were obtained for chitin (molar mass of 627.59 g/mol), N-acetyl galactosamine (molar mass of 221. 20 g/mol) and fucose (molar mass of 164.15 g/mol) confirming the presence and purity of the reagents.
Testing growth-promoters for M. ulcerans culture. For each one of the M. ulcerans strains under study, a 100 µL-volume of a 10 5 AFB/mL suspension and PBS as negative control were cultured in parallel in triplicate onto each of the three enriched media and onto Middlebrooks 7H10 reference medium supplemented with 10% (v/v) OADC and 0.5% (v/v) glycerol. We seeded an inoculum of 10 4 AFB by Petri dish rather that a greater concentration of AFB in order to minimize bias of counting because in bacterium kinetic, population growth strongly depends on initial conditions. The 55-mm plates were incubated at 30 °C in microaerophilic atmosphere. Five days, ten days, fifteen days, twenty days and thirty days after inoculation, colonies were observed by a MZ-FLIII fluorescence microscope (Leica, Nanterre, France) equipped with a GFP filter and an ICA digital camera (Leica) to detect mycobacterial autofluorescence 64 . Counting of fluorescent colonies was performed using the Leica Application Suite software (Leica) 64 . Identification of colonies was confirmed by MALDI-TOF-MS as previously described 65,67 and by rpoB gene sequence analysis 63 .

Decontamination of river freshwater experimentally inoculated with M. ulcerans strains.
Freshwater obtained from the Huveaune River, a small river of southern France was used. In first time, we analyzed without decontamination formerly direct cultured samples of the river freshwater on 5% sheep-blood Columbia agar (COS, bioMérieux, La Balme-Les-Grottes, France) and after two days of culture we identified by MALDI-TOF-MS the growth bacterial colonies. The direct culture aimed to list bacteria and fungi contained in the native freshwater for the differential appreciation of germs between direct culture and culture after chlorhexidine decontamination. Secondarily, we decontaminated the freshwater of Huveaune River with 1% chlorhexidine Scientific RePoRTS | (2018) 8:6778 | DOI:10.1038/s41598-018-25278-y and culture the pellet onto 5% sheep-blood Columbia agar. The decontamination was processed as previously described 68 .
The next step consisted of on the chlorhexidine decontamination of freshwater experimentally inoculated with M. ulcerans suspension. Briefly, in 50 mL Corning Falcon conical centrifuge tubes (Becton Dickinson) containing 5 mL of river freshwater experimentally inoculated with one milliliter of 10 7 AFB/mL of each M. ulcerans Cu001 and M. ulcerans ATCC25900, the chlorhexidine decontamination was done as previously described 68 . Then, a 100-μL volume of pellet was inoculated in parallel on 5% sheep-blood Columbia agar, in triplicate on each of the growth-promoting media (chitin, N-acetyl galactosamine, fucose) and on the standard Middlebrook 7H10 medium. The negative control was consisted of seeding 100 µL of sterile PBS onto 5% sheep-blood Columbia agar, standard Middlebrook 7H10 medium and onto each of the growth-promoting media. Cultures and negative controls were examined at day 3, day 7 to appreciate the quality of the decontamination method and quality of the cultures. Colonies were counted at day 15 and day 30 using the Leica Application Suite software as described above. For the treatment of samples for seeding, water samples without and with plants debris were concentrated by centrifugation at 1,700 g for 15 min and the supernatants were removed. The resulting pellets from water without plant debris were suspended in PBS before decontamination. The resulting pellets from water with plant debris were suspended in PBS and mechanically disrupted with silica beads (0.5 mm diameter) and the obtained suspensions were transferred into 50 mL Corning Falcon conical centrifuge tubes before decontamination. 500 mg aulacode feces added into 50 mL Corning Falcon conical centrifuge tubes containing 5 mL of PBS were mechanically disrupted with silica beads (0.5 mm diameter) and the obtained suspensions were transferred into another 50 mL Corning Falcon conical centrifuge tubes before decontamination. Then, each sample was decontaminated using 1% chlorhexidine as previously described 68 . 100 µL of decontaminated pellet were seeded onto the chitin medium, NAG medium, fucose medium and DZ medium (Middlebrook 7H10 medium containing chitin (0.2%), NAG (0.01%) and fucose (0.01%)) and incubated for 24 weeks at 30 °C under a micro-aerophilic atmosphere condition using CampyGen Compact. 100 µL of sterile PBS was seeded in each media as negative controls and incubated in the same conditions. Colonies were screened with auto-fluorescence three times a month and the obtained colonies were identified by Real-Time PCR (RT-PCR) incorporating three independent gene targets, IS2404, IS2606 and KR-B, within the M. ulcerans genome 7,22,69,70 and internal positive control to determine the level of inhibition as previously described 7 .

Isolation of
The presence of M. ulcerans DNA in environmental samples was analyzed by RT-PCR. We used the three primers (KR-B, IS2404 and IS2606) to improve the specificity of the M. ulcerans DNA detection in the environmental samples 71,72 . Primers and probes from Applied Biosystems that were selected from regions of the sequences for IS2404, IS2606 and KR present on the plasmid pMUM001 were used 7 . Probes IS2404TP and KR-BTP were labelled with the fluorescent dye 6-carboxyfluorescein (FAM) at the 5′ end and a non-fluorescent quencher at the 3′ end. Probe IS2606TP was labelled with the fluorescent dye VIC at the 5′ end and a non-fluorescent quencher at the 3′ end 7 ( Table 2). Total DNA was extracted from feces using the QIAmp ® DNA Stool kit according to the manufacturer's instructions (Qiagen, Stochach, Germany); total DNA was extracted from plant debris and sediment water using the NucleoSpin Tissue Kit (Macherey-Nagel, Hoerdt, France). M. ulcerans Cu001 DNA was extracted using a commercial Nucleospin Tissue kit (Macherey-Nagel, Hoerdt, France) which was used as a positive control in PCR-based identification of micro-colonies while distilled water was used as a negative control. The PCR inhibition was assessed by adding 10 μL of internal control into 190 μL of sample, as previously described 73 . Each IS2404, IS2606 and KR real-time PCR mixtures contained 5 μl of DNA or negative control, 20 μM of each primer, 5 μM of probe, 3.5 μL of sterile water and 10 μL of mastermix (Eurogentec) in a total volume of 20 μl. The RT-PCR program comprised one cycle at 50 °C for two minutes and 40 cycles at 95 °C for 15 seconds and 60 °C for  IS2404  IS2404 TF  19  59  AAAGCACCACGCAGCATCT  27746-27762  4/201   IS2404  IS2404 TR  18  AGCGACCCCAGTGGATTG  27787-27804   IS2404  IS2404 TP  6FAM-CGTCCAACGCGATC-MGBNFQ  27768-27781   IS2606  IS2606 TF  21  58  CCGTCACAGACCAGGAAGAAG  28912-28932  8/82   IS2606  IS2606 TR  21  TGCTGACGGAGTTGAAAAACC  28947-28969   IS2606  IS2606 TP  VIC-TGTCGGCCACGCCG- 22 , amplification was done in a CFX 96 ™ real time PCR thermocycler and detection system (BIO-Rad, Marnes-la-Coquette, France). Two negative controls were incorporated into each PCR run. All samples were tested in triplicate. A specimen was considered as positive for the detection of M. ulcerans when both the insertion sequences IS2606 and/or IS2404 and the KR-B detection were positive (Ct ≤ 40 cycles) in ≥2/3 replicates. The Ct cut-off value was chosen in order to increase the sensitivity of the detection as previously described 22 .
Statistics. The results of growth kinetics were expressed as mean value ± standard error of the mean (SEM) of counted colonies. Analysis of variance (ANOVA) was used for statistical analysis. P values lower than 0.05 (p < 0.05) were considered statistically significant. The doubling time was determined by calculating the average slope of the mycobacterial replication curve during the exponential phase of culture using the following formula: growing ratio: µx = lnxB-lnxA/tB-tA and doubling time: G = ln2/µx.