Molecular Typing using PCR-RFLP Reveals Diversity of Environmental Mycobacteria Agent of Buruli Ulcer in Ivory Coast, Cote d’Ivoire (West Africa)

Buruli ulcer is a neglected skin disease caused by Mycobacterium ulcerans (MU), and affects 1000 people each year worldwide and particularly in African countries. Eradication of Buruli ulcer is difficult because of the lack of early diagnostic in rural endemic regions, and the unknown of the disease in national health care system. In the rural wetlands and marshes in Central and West Africa, children are most affected. MU belongs to environmental Mycolactone Producing Mycobacteria (MPMs) and presents high genetic diversity of the circulating strains. Diagnoses were applied by culture and genome detection by Polymerase Chain Reaction (PCR). The PCR became a gold method to confirm clinical and environmental samples for MU. The MIRU-VNTR method and the Restriction Fragment Length Polymorphism (RFLP) combined with MIRU-VNTR markers were used to discriminate Mycobacteria from different sources. The aim of this study was to investigate the molecular diversity of different mycobacteria sources, from environment, culture strains, and clinical samples by using combined MIRU-VNTRTyping and RFLP. A total of 26 samples (water, sediment, mycobacteria strains, and swab) from endemic sites were first confirmed by IS2404 or Ziehl-Neelsen staining. The samples were analyzed by PCR typing for 4 specific markers (MIRU-1, VNTR6, VNTR19, ST-1) and the amplicons were digested with MspI restriction endonuclease and separated by 3% agarose gel electrophoresis for PCR-RFLP analysis. Our results showed different amplification by VNTR-MIRU-typing. For environmental samples a low amplification was detected by 25% for PCR-MIRU-1. Culture strains and clinical samples had amplification rate of 35.7% and 62.5% respectively. ST-1 had the best amplified marker for culture strains by 71.4%, while clinical samples have good amplification rate by 62.5 % for all markers. PCR-RFLP-profiles of clinical samples were identical, while environmental and mycobacteria strains showed different PCR-RFLP-profiles. We have developed a PCR-RFLP sensitive, easier and inexpensive approach to confirm genotyping of nontuberculosis mycobacteria in endemic countries for environmental screening. We suggest mutation in repeat loci of VNTR-MIRU sequence and the adaptation of mycobacteria from environment to human. This study confirms the circulation of several genotypes of mycobacteria in Ivory Coast.


Introduction:
Environmental mycobacteria are responsible for many infections in humans. One of them, Mycobacterium ulcerans (MU), is the agent of a skin disease called Buruli ulcer (BU). BU infection is the third most important mycobacteria infection worldwide after Leprosis and Tuberculosis [1]. BU occurs in rural regions in West and Central Africa, Latin America, Australia and Asia [2,3]. Since 2009, over 1000 cases of BU have been reported in Côte d'Ivoire each year [4]. Several studies have demonstrated the evidence of environmental transmission of MU from environmental samples [5][6]. MU genomes have been detected in endemic sites in aquatic environment, like rivers, wetlands and stagnant water spots [7][8][9][10][11][12][13][14] and in small mammals [15,16]. MU is a slow growing mycobaterium and the first positive results appears between 6 to 12 weeks, after several culture transition in various media [17]. For a faster diagnosis of MU, molecular detection is gold standard [18][19][20]. The virulence gene of MU is coded for Mylolactone, a toxin responsible for the destruction of skin tissues [21]. Further genetic analyses have confirmed high genomic diversity of MU and other MPMs, by using Variable Number Tandem Repeat (VNTR) and Mycobacterial Interspersed Repetitive Units (MIRU) [22][23][24][25] has demonstrated the high specificity and sensitivity of several VNTR-MIRU markers in clinical and environmental matrices.
Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis is a widely applied method to detect gene mutations, which allows distinguishing mutant-type and wild-type sequences via destructing or generating enzyme restriction sites through PCR and subsequent electrophoresis separation of differential fragments [26]. Compared to other methods, PCR-RFLP offers a simple operation, higher sensitivity and reproducibility, and no complex equipment requirements [27].
The objective of this study is to elucidate the genomic diversity of environmental Mycobacteria in several matrices, by using combined PCR-RFLP-method applied on four MIRU-VNTR markers.

Ethics Statement:
This work was carried out with the approval of the National Buruli Ulcer Control Program in Ivory Coast (PNLUB) and the CNR Buruli at Pasteur Institute no further permissions were required for the environmental sampling and clinical samples.

DNA Extraction from environmental and clinical samples
The Mobio Powersoil Kit (MOBIO, USA) was used with minor modifications to extract DNA from water and sediment samples. Briefly, 400 µL of samples were added to 60 µL of lysis C1 solution, following by horizontally shaking with the Mobio Vortex and then centrifuged for 30 seconds, 10000 rpm at room temperature. After several incubation by washing and centrifugation, DNA was resuspended in 100 µL of elution buffer. DNA extraction from clinical samples were performed by alcaline heat shock following by ethanol precipitation [28]. The samples were suspended in 2 ml sterile water and centrifuged for 10 min, 15000 rpm at 4°C. 200 µL of pellet were performed in DNA extraction. Each sample was performed twice in DNA extraction. All DNA templates were eluted with 50 μL DNA/ Rnase fee H2O and stored at −20°C until use. Negative and positive extraction samples were applied to verify the presence of inhibitors in the samples.

Culture of Mycobacteria from environmental samples
Cultures were performed at 22°C on 3 Lowenstein-Jensen (LJ) solid media: a simple LJ medium, a LJ medium supplemented with NaCl and LJ supplemented with Glycerol. All media were inoculated with 150 µL of inactivated samples. The samples were placed in the dark and in the light for 21 days. Each sample was performed twice in dark and in light incubation. Positive cultures were analyzes using Ziehl-Neelsen (ZN) staining test with microscopy for acid fast bacilli specific for mycobacteria [17]. 5 µL of extracted DNA were added to 45 µL PCR-mix containing µM each specific primer for MIRU-1, ST-1, VNTR6, VNTR19, 0.2 mM dNTPs (Sigma, USA), 1X Flexi Taq-polymerase buffer, 1.5 mM MgCl2 and 1Unit Go-Flexi Taq-Polymerase (Promega, Germany). The PCR reaction was described in previous studies [11,28]. PCR were tested by the 1/10 dilution of DNA, to test the presence of PCR inhibitors in the samples. Negative control (sterile DNase/RNase-free water) and positive control (181UB15 genomic DNA) were included in all PCR runs for the test quality and the detection of decontamination. The PCR was running in a thermocycler (GenAmp 9700, Applied Biosystems, USA).

PCR-RFLP for MIRU-VNTR
The analysis of MIRU-VNTR sequences of 4 markers was performed in the NCBI bank data, and analyzed with Geneious and RestrictionMapper 3.0 software (Geneious 8.1.7; http:// www.restrictionmapper.org/). Restriction maps performed with various restriction enzymes were compared to find the best fitting enzyme in order to have the clearest specific profile after digestion. MspI was found to make better DNA profiles. The RFLP was applied with MspI amplified DNA from PCR reactions of MIRU-1, VNTR6, VNTR19 and ST-1.
The amplified DNAs from PCR were purified by using Qiaquick kit (Qiagen, Germany) with manufactor's instructions. 5 µL purified PCR product sequence were added to 15 µL restriction digestion-Mix containing 1X enzyme buffer, 10µg/mL BSA and 0.1U/µL MspI (Promega, Germany). The reaction was incubated for 24 hours at 37°C.

Electrophoresis of DNA amplified products
5µL PCR-product were analyzed by electrophoresis in 1.5% Agarose gel, and migrated at 100 Volts. The size was estimated by comparison with a 100 bp ladder (Promega, Germany). For RFLP-product, 20 µL of digested with MspI PCR-product were analyzed by electrophoresis in 3% Agarose gel at 85 Volts.

Culture of mycobacteria
From 24 samples in culture, 62.5% (15/24) were positive and confirm by ZN staining specific for mycobacteria (Table 2). Water samples have the most positive cultures. All positive cultures were observed after 6 days, rapid growing mycobacteria, no slow-growing culture were positive from environmental samples. All positive mycobacteria were from endemic sites.

RFLP-Analyzing
From all including samples, only positive PCR-typing was analyzed in RFLP-Analyzing. We compared the digestion profile of the 500 bp or 380 bp of MIRU-1 from 3 matrices.

Discussion
M. ulcerans, the agent of BU is endemic in Ivory Coast, and its epidemiology unknown. The evidence of environmental transmission of MU and MPMs from molecular studies were essential to discriminating clinical and environmental or their interaction. In this study, we have demonstrated that specific markers from mycobacteria were not detected in different samples in the environment and in clinical samples. The high amplification rate by PCR was detected for ST-1 by 57% for all samples. Our results were similar with the findings of Kakou et al. [28] for the polymorphism of MU for clinical samples in Ivory Coast. The authors have found the circulation of 2 clones in the same endemic sites and the high amplification rate of ST-1 by 85%. Previous studies have detected different genotypes and profiles of MU in West Africa [13]. In this study, the using of endonuclease enzyme MspI to confirm the DNA profiles of amplified PCR-VNTR has demonstrated that clinical and environmental samples have difference in genotyping's markers.
We have detect only rapidly growing mycobacteria, in contrast of findings of Aboagye et al. [29], the have detected 60% of slow growing mycobacteria in Ghana. This could be explain by the use of another decontamination method in this study and the difficulties of the growth mycobacteria from environmental samples [29]. In Ivory Coast, molecular tests have confirm the BU incidence in the country [19] and reveals the diversity of circulating strains in humans in endemic sites, but the goal is to elucidate the correlation between all nontuberculosis mycobacteria in the environment involved in the incidence of BU in the country.
Recently, Ablordey et al. [30] have demonstrated using whole genome sequencing that two exotic genotypes were found in an endemic site in Ghana with highly restricted intra-strain genetic variation. Our finding shown the profiles difference between genotyping markers. MIRU-VNTR are variable repeat sequences with some new deletion or insertion in the repeat loci, we suggest the evidence of mutation in the MIRU-VNTR loci from the environment in human. This can been explain the difference in RFLP-method from clinical, culture or environmental matrices. VNTR-MIRU could be an adaptation genotyping marker of mycobacteria during infection from environment to human. The using of RFLP-method is consistent to detect low rate of gene mutations in cancer cells [31].
The high specificity and sensitivity of genotyping markers for MU and MPMs cannot be the evidence for detection in all matrices and in particular in endemic sites of MPMs. This study confirmed those VNTR profile variability between MU clinical strains in one hand, but also between environmental mycobacteria strains. The use of RFLP with MspI to confirm each amplified marker, gives better resolution of MIRU-VNTR-typing method. The RFLP-profiles confirm the difference in MIRU-VNTR markers. Chemlial et al.
[32] has demonstrated the high diversity of MU strains by using RFLP with IS2404. IS2404 is located both in chromosome and in plasmid in MU and was abundant copies in mycobacteria. Our study is only focused on chromosome genome for all genotyping markers to elucidate the polymorphism of mycobacteria.

Conclusion
Buruli ulcer is endemic in Ivory Coast, and the need to control and understand the transmission from the environment to human is very important. We used molecular tools to elucidate the diversity of circulating strains in different matrices. Our results confirm the existence of polyclonal infection of BU in the high endemic country and the spread of several mycobacteria strains in the environment. We have developed first a PCR-RFLP sensitive, easier and inexpensive approach to confirm genotyping of nontuberculosis mycobacteria in endemic countries for environmental screening. In perspective, we will explore the whole genome sequencing for circulating strains in Ivory Coast.