1

Diversity, community composition and dynamics of 1

non-pigmented and late-pigmenting rapidly growing 2

mycobacteria in an urban tap water production and 3

distribution system 4

Authors 5

S. Dubrou1, J. Konjek

2,3, E. Macheras

2,3, B. Welté

5, L. Guidicelli

1, E. Chignon

1, M. Joyeux

5, JL. 6

Gaillard2,3

, B. Heym2,3

, T. Tully4*

and G. Sapriel2,#3*

7

Affiliations 8

1 Laboratoire d’Hygiène de la Ville de Paris, 11 rue George Eastman 75013 Paris, France. 9

2 Service de Microbiologie, Hôpital Ambroise Paré (Assistance Publique – Hôpitaux de Paris), 9 10

avenue Charles de Gaulle, 92104 Boulogne-Billancourt, France. 11

3 EA 3647, UFR des Sciences de la Santé Paris Ile-de-France Ouest, Université de Versailles Saint-12

Quentin-en-Yvelines, 78280 Guyancourt, France. 13

4 CNRS/UPMC/ENS – UMR 7625, Laboratoire Écologie & Évolution, Université Pierre et Marie 14

Curie, Case 237, 7 Quai St Bernard, 75005 Paris, France. 15

5 Eau de Paris, Direction de la recherche, du développement et de la qualité de l’eau, 9 rue Schoelcher 16

75675 Paris cedex 14, France. 17

# Corresponding author E-mail : guillaumme.sapriel@uvsq.fr 18

* Both authors contributed equally to this work 19

Copyright © 2013, American Society for Microbiology. All Rights Reserved.Appl. Environ. Microbiol. doi:10.1128/AEM.00900-13 AEM Accepts, published online ahead of print on 8 July 2013

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

2

Keywords 20

Mycobacteria, rpoB-based identification, mapping, hierarchical clustering, principal component 21

analysis, urban water distribution system, M. chelonae. 22

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

3

ABSTRACT 23

Non-pigmented and late-pigmenting rapidly growing mycobacteria (RGM) have been reported to 24

commonly colonize water production and distribution systems. However, there is little information 25

about the nature and distribution of RGM species within the different parts of such complex networks 26

or about their clustering into specific RGM species communities. We conducted a large-scale survey 27

between 2007 and 2009 in the Parisian urban tap water production and distribution system. We 28

analyzed 1418 water samples from 36 sites, covering all production units, water storage tanks and 29

distribution units; RGM isolates were identified using rpoB sequencing. We detected eighteen RGM 30

species and putative new species, most isolates being Mycobacterium chelonae and Mycobacterium 31

llatzerense. Using hierarchical clustering and principal component analysis, we found that RGM were 32

organized into various communities correlating with water origin (groundwater or surface water) and 33

location within distribution network. Water treatment plants were more specifically associated with 34

species of the M. septicum group. On average, M. chelonae dominated network sites fed by surface 35

water and M. llatzerense those fed by groundwater. Overall M. chelonae prevalence index increased 36

along the distribution network and was associated with a correlative decrease in the prevalence index of 37

M. llatzerense, suggesting competitive or niche exclusion between these two dominant species. Our 38

data describe the great diversity and complexity of RGM species living in the inter-connected 39

environments that constitute the water production and distribution system of a large city, and highlight 40

the prevalence index of the potentially pathogenic species, M. chelonae, in the distribution network. 41

42

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

4

INTRODUCTION 43

Non-pigmented and late-pigmenting rapidly growing mycobacteria (RGM) are ubiquitous in soil and 44

water environments (1-6). Most are harmless saprophytes, but some, such as Mycobacterium chelonae, 45

and Mycobacterium abscessus, are opportunistic pathogens that are causing increasing concern (7-9). 46

Potentially pathogenic RGM are associated with a wide spectrum of diseases in humans, including 47

pulmonary tract, skin, soft-tissue and disseminated infections (10-14) mostly in patients with 48

predisposing conditions (e.g. lung diseases, chronic obstructive pulmonary disease, cystic fibrosis, 49

genetic predisposition, or immunosuppressive therapy). Contamination and hypersensitive reactions are 50

often due to environmental exposure ( associated with, for example, hot tubs, metalworking fluids, or 51

contaminated dust) (15-19) and many outbreaks have been reported over the last decade following 52

invasive medical procedures (20-25). Tap water may be an important source of contamination in urban 53

environments (13, 18, 26-30). 54

RGM are commonly recovered from water treatment and distribution systems (31-38), probably 55

because they can form biofilms (2, 39-41) and resist chlorination and oligotrophic conditions (42, 43). 56

Mycobacterium chelonae and Mycobacterium fortuitum were the species most frequently detected in 57

previous studies (31, 34, 35, 40, 44). However, these previous studies involved only a small number of 58

samples collected from a limited part of the water system considered, and provide only qualitative 59

information. Most of these studies addressed specifically RGM, and few used molecular methods 60

allowing an accurate identification of isolates to the species level (e.g. rpoB or hsp65 sequencing) (31, 61

34, 40, 44). Thus, although there are many reports of RGM detection in water treatment and 62

distribution systems, there are no rigorous and quantitative descriptions of the diversity and spatial 63

distribution of RGM species within these complex systems, and no robust information about their 64

clustering into communities. 65

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

5

Le Dantec et al. (44) reported a survey conducted in 2000-2001 and analyzed the occurrence of non 66

tuberculous mycobacteria (NTM) in the southern part of the Paris urban water system. NTM detection 67

rates were found to differ between two water treatment plants and to increase along the distribution 68

network. However, nearly 55% of the NTM isolates were not identified to the species level and only 69

three RGM species were detected: M. chelonae, M. fortuitum and Mycobacterium peregrinum. This 70

study included only one third of the distribution units and did not include a dataset representative of 71

proximal parts of the Parisian distribution network. Also, the small number of samples prevented 72

statistical analysis. 73

We report a large-scale systematic survey of RGM in the Parisian water treatment and distribution 74

systems, using an rpoB-based identification method (45, 46). 75

We specifically addressed the following questions: 76

i) What is the prevalence index of RGM in the different parts of the network? 77

ii) What RGM species are present within the network and what is the prevalence index of each? 78

iii) Do some RGM species accumulate in particular parts of the network and is there a significant link 79

between these groups of species and the origin of the water (groundwater, surface water)? 80

We also tested the validity of hypotheses and results from previous small sample studies, by using an 81

extensive sampling procedure to calculate a statistical index of RGM prevalence allowing reliable 82

comparisons. 83

MATERIALS AND METHODS 84

The Parisian drinking water supply system. Both surface and ground water feed the 85

Parisian water treatment and distribution systems (Fig. 1). Two of the three surface water treatment 86

plants are fed by the Seine river (plants W and Y) and one by the Marne river (plant X). The treatment 87

process differs between plants: i) plant W: pre-ozonation, coagulation flocculation settling, rapid sand 88

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

6

filtration, ozonation, filtration through granular activated carbon, chlorination; ii) plants X and Y: pre-89

ozonation, contact coagulation on biolite, rapid sand filtration, slow sand filtration, ozonation, filtration 90

through granular activated carbon, chlorination. The groundwater, carried by three aqueducts comes 91

from several springs in catchment areas 150 km south of Paris (Aqueducts A and Β) and 175 km west 92

of Paris (Aqueduct C). It is filtered through granular activated carbon and then chlorinated (A, B) or 93

treated with powdered activated carbon, then chlorinated, and subjected to ultrafiltration (C). 94

The treated water is stored in tanks with a storage capacity of 90,000 to 200,000 m3. The water 95

generally remains in these tanks for about two to three days. There are nine storage tanks (T1-T9) 96

which are supplied either with groundwater (T1, T4), surface water (T5-T9) or both (T2, T3, see Fig. 97

A1 for the detailed network topology). 98

On release from storage, the water flows into the water distribution system (~1,600 km of cast-iron 99

pipelines) where the average residence time is six hours. This network is divided into 16 areas of 100

homogeneous water origin known as “distribution units” (Fig. 1). We studied 15 distribution units 101

(units D1 to D15) for which the water origin is clearly known. 102

Sampling protocol and locations. Water samples were collected monthly from 36 points in 103

the production (plants W, X, Y and aqueducts A, B and C), storage (tanks T1 to T9) and distribution 104

(units D1 to D15) systems (Table A1). At plants, samples were collected immediately following each of 105

the following treatment steps: ozonation, granular activated carbon filtration and chlorination steps. 106

Samples from aqueducts were collected after chlorination. Water from storage tanks was collected at 107

the tank exits. Water from the distribution system (2 to 4 sampling sites per distribution unit) was 108

sampled from the drinking network pipes. In total, 1418 water samples were collected during the three-109

year study period: 325 samples from the production system (plants and aqueducts), 246 samples from 110

the storage system (tanks) and 847 samples from the distribution system, corresponding to an average 111

of 39 samples for each sampling site (i.e. one sample per month on average). 112

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

7

RGM culture and isolation from water samples. Water samples were collected in sterile 113

bottles and processed within 12 hours of collection. Samples were treated as previously described (44). 114

A 1-liter volume of collected water was filtered through a 0.45 µm pore-size cellulose membrane 115

(Millipore). The membrane was immersed in 10 mL of the same water sample and sonicated for 10 116

min, then decontaminated by addition of sodium lauryl sulfate to 3% and NaOH to 1%. After 117

neutralization, bacteria were concentrated by centrifugation and equal volumes were used to inoculate 118

each of six Löwenstein-Jensen and Coletsos agar slants. Cultures were incubated at 30°C and 37°C and 119

examined three times during the first week and then weekly for two months. Suspected RGM colonies 120

were streaked for isolation on agar slants and then stained by the Kinyoun-modified Ziehl-Neelsen 121

method. Acid-fast bacteria yielding non-pigmented colonies in less than seven days were subjected to 122

rpoB sequencing (see below). 123

rpoB-based identification. Partial rpoB sequencing was performed as described (47). 98% of 124

tested isolates could be amplified by PCR, and every amplified DNA could be analyzed by sequencing. 125

BLAST was used to compare sequences with a local bank of NTM rpoB sequences extracted from 126

GenBank. RGM species identification was based on an identity threshold of 97% as described by 127

Adékambi et al. (48-50). Sequences displaying less than 97% identity with any known RGM sequence 128

were considered to be new rpoB RGM sequence types (labeled ParisNewRGM with a specific code 129

number). 130

Alignments and phylogenetic analysis. The website www.phylogeny.fr was used for 131

phylogenetic analyzes (51). MUSCLE (http://www.drive5.com/muscle/) was used to align rpoB 132

sequences and a conserved stretch of 567 bp was selected with Gblocks (52). A representative set of 133

RGM rpoB sequences was chosen for alignment and tree construction (Table 1). A distance tree was 134

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

8

constructed by the neighbor-joining method with 1000 bootstrap replicates and the Kimura 2 135

substitution model. Similar results were obtained with the maximum likelihood method. 136

Prevalence index. For each 1-liter samples collected from the network, the RGM culture and 137

isolation method (see above) provided a binomial response for all RGM species grouped together (or 138

for each species individually): 0 when no RGM was detected, and 1 when RGM were detected. These 139

binomial variables were analyzed with generalized linear models (GLM) to estimate the RGM 140

detection probability - the estimated probability [0-1] of detecting RGM (or a particular RGM species) 141

in a 1-L sample of water. This estimation method was used rather than directly computing the 142

proportion of positive samples (the number of positive samples divided by the total number of samples 143

studied) for two reasons. (1) It provides unbiaised estimators of the detection probabilities, which is not 144

always the case if the proportion of positive samples is used, especially when the samples are not all 145

independent. In such situation, generalized linear mixed models (GLMM) should be applied to provide 146

unbiased estimators by taking into account the non independence of the samples, for example those 147

collected at the same sampling location. (2) GLM and GLMM provide both reliable estimators of the 148

RGM detection probability and their associated 95% confidence intervals; such confidence intervals are 149

required for statistically valid comparison of the prevalence index between sampling locations. 150

(53)(8)(8)[53]. The prevalence index was calculated from the results of the isolation and identification 151

of RGM at each site. Each individual RGM isolation is not in itself quantitative, but repeated isolation 152

analyses at each sampling site can be considered as repeated random sampling, allowing statistical 153

estimation of the RGM detection probability (here called the 'prevalence index'). This prevalence index 154

is not the same as a proportion, usually measured as the number of positive samples divided by the total 155

number of samples studied and expressed as a percentage. 156

Prevalence indexes were estimated using generalized linear models (Fig. 2A, GLM function of R (54), 157

link=logit) or generalized mixed linear models (Fig. 2B, GLMM PQL function from the MASS library 158

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

9

(55), link=logit) fitted to the binomial data for the presence/absence of RGM in 1-liter samples. We 159

used mixed models to estimate the mean prevalence index in the whole network. Location in the 160

network (production, plant, storage or distribution) was used as the random effect (56). These models 161

provide unbiased estimates of prevalence values with their 95% confidence intervals (95% CI) for the 162

network as a whole or for each position in the network, depending on the analysis. This approach was 163

applied to all RGM species considered together or to each RGM species independently. Two prevalence 164

values were considered to differ significantly from each other if their 95% CI did not overlap. The 165

estimated detection probabilities are referred to 'prevalence indexes'. 166

Hierarchical clustering. We identified putative groups of RGM species and estimated the 167

distances between them at the different collection sites by applying hierarchical clustering to a distance 168

matrix built as follows. We estimated the prevalence index of each species at each of the 36 collection 169

sites with a binomial linear model as described above. We included the 12 most prevalent species for 170

the cluster analysis and excluded sampling points with fewer than six identified isolates (four sampling 171

points, all at water treatment plants were thus excluded). This prevalence index matrix was then used 172

for the cluster analysis (pvclust function, with the average method and correlation distance) (57). This 173

generated a distance tree which was used to identify groups of different RGM species assemblages. 174

RGM groups were determined (bootstrap value >99%). The hypothesis of over-representation of 175

sampling points of surface water or groundwater origin was tested. A binomial test allowed 176

determination of associated p values. 177

Principal component analysis. We performed a principal component analysis of the 178

prevalence index matrix (FactoMineR package, PCA function) (58). Only species occurring more than 179

three times were considered. Correlation coefficients were determined (dimdesc function) with p < 5%. 180

Confidence ellipses for each RGM group were calculated (ellipse function) with p < 5%. 181

182

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

10

183

Results 184

RGM prevalence. The overall RGM prevalence index at sampling sites was 3.5 to 100% (Fig. 2A), 185

with an overall mean [95% CI] value of 72% [40-92%] across the whole network. Of the water 186

treatment units, plant W had the lowest RGM prevalence index for all steps of the water treatment 187

process. In each plant, the RGM prevalence was highest after filtration through granular activated 188

carbon. On average, the RGM prevalence index increased from the production units exits (54% [46-189

62]) to the tanks (75% [69-80]) and distribution units (96% [94-97]], Fig. 2). The RGM prevalence 190

index was significantly lower after the terminal chlorination step for plant W (3.5% [0.2-15]) and plant 191

X (48% [30-66], Fig. 2A). However, the RGM prevalence index in chlorinated water from aqueducts A 192

and B was quite high (90% and 100% respectively). The RGM prevalence index at three water storage 193

tanks was significantly lower than average (T1, T3, T4), whereas at two it was higher (T5 and T6). The 194

mean RGM prevalence index was above 80% for all distribution units (Fig. 2A and Table A1). 195

RGM species diversity and prevalence. A total of 643 RGM isolates were recovered from 196

the network; rpoB amplification and sequencing were successful for 630 (98%). Eighteen RGM 197

“species” (i.e., displaying an rpoB sequence type differing by at least 3% from any other rpoB 198

sequence) were identified: nine previously described RGM species, one recently described rpoB 199

sequence type (rpoB sequence NLJvIW_016 (59)), and eight new RGM rpoB sequence types 200

(ParisRGMnew1-8) (Fig. 2B & Fig 3). ParisRGMnew1, ParisRGMnew3 and ParisRGMnew4 were 201

recovered several times from independent water samples. ParisRGMnew1 and ParisRGMnew3 202

constitute a new group with NLJvIW_016 and ParisRGMnew7 (Fig. 3). ParisRGMnew4 is close in the 203

rpoB tree to species of the M. fortuitum group (including M. alvei, M. septicum and M. peregrinum) 204

(Fig. 3). 205

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

11

M. chelonae and M. llatzerense were by far the most frequently recovered species (average prevalence 206

of 25% and 20%, respectively; Fig. 2B). Six other RGM had mean prevalence index values between 207

1.5 and 5.3%: M. setense, M. salmoniphilum, M. peregrinum, M. septicum, ParisRGMnew4 and 208

ParisRGMnew1. The other RGM were found at very low frequencies (<1%) (Fig. 2B). 209

RGM species composition groups. Hierarchical cluster analysis identified four groups with 210

specific RGM species composition within the network (RGM groups 1 to 4, Fig. 4A and 4B). Group 1 211

clustered away from the others and comprised all sites at water treatment plants (upstream of the exit). 212

The three other groups included all distribution network sampling sites (production exits, storage, and 213

distribution units). The sampling sites for water of groundwater origin were over-represented in group 2 214

(p = 0.01), sampling sites for water of mixed origin were over-represented in group 3 (p = 0.1), and 215

those of surface water origin were over-represented in group 4 (p = 0.03). Almost all sampling sites for 216

water of groundwater origin were in group 2, whereas almost all sampling sites for water of surface 217

water and mixed origin were in groups 3 and 4. Thus, RGM species composition allows the description 218

of a limited number of groups that correlate with the origin of the water (groundwater vs surface water) 219

and the location in the network (water treatment plants vs distribution units). 220

Structure of RGM species composition. The RGM species forming the four groups 221

identified by hierarchical clustering analysis and their relationships with RGM species content were 222

further studied by principal component analysis (PCA) with the same dataset. The four groups were 223

clearly distinct (Fig. 5A). Group 1 was located in the upper-right panel and was associated with 224

sampling sites at water treatment plants. This direction was defined by a specific RGM species group 225

associating M. peregrinum, ParisRGMnew_4 and, to a lesser extent, M. septicum (correlation 226

coefficients on the first dimension: 0.59, 0.62 and 0.43, respectively; Fig. 5B). 227

Group 2 was in the upper-left panel, and included all sampling sites in the water distribution network 228

with groundwater origin. This direction was characterized by another group of species: M. llatzerense, 229

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

12

M. salmoniphilum and ParisRGMnew_3 (correlation coefficients on the first dimension: -0.69, -0.54 230

and -0.5, respectively; Fig. 5B). 231

Group 3 was in the centre of the panel (mainly sampling sites for water of mixed origin), and thus was 232

not characterized by an RGM species composition differing from that found on average in the water 233

system. 234

Group 4 was in the lower-right panel (mainly sampling sites for water of surface origin). In the first 235

dimension, this direction was defined by the same species as group 1 (i.e, M. peregrinum, 236

ParisRGMnew_4, M. septicum); in the second dimension, this direction was defined by M. chelonae 237

(correlation coefficient: -0.76) and, to a lesser extent, with NLJvIW_016 (-0.4). Interestingly, all but one 238

sampling sites of the water distribution network for water of surface origin were located in the lower 239

panel (right and left) and were defined by a higher occurrence of M. chelonae and NLJvIW_016. 240

Thus, M. chelonae and NLJvIW_016 are associated with sampling sites fed by surface water, and M. 241

llatzerense, M. salmoniphilum and ParisRGMnew_3 with sampling sites fed by groundwater. This was 242

confirmed by a PCA only including sampling sites in the distribution network (data not shown). 243

244

Variations of M. chelonae and M. llatzerense prevalence along the water 245

distribution network. We compared the prevalence values for M. chelonae and M. llatzerense at 246

different levels of the distribution network (water production plants, water production exits, storage 247

tanks and distribution units) according to the water origin (Fig. 6). At every level of the distribution 248

network, the prevalence index of M. chelonae was higher for sampling sites for water of surface than 249

groundwater origin; consistently, sampling sites for mixed water origin had intermediate prevalence. 250

Irrespective of the water origin, M. chelonae prevalence index increased along the water distribution 251

network, with mean values of between 40 and 45% in distribution units fed by surface water and mixed 252

water, respectively, and around 25% for those fed by groundwater. By contrast, the prevalence index of 253

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

13

M. llatzerense was higher for parts fed by groundwater at both the production and storage levels, but 254

decreased substantially between the storage and distribution levels to about 25%, a value very similar 255

to that found for parts fed by surface water or mixed water. 256

To determine if there was a relationship between the presence of M. chelonae and that of M. 257

llatzerense, we plotted variations of their prevalence index estimates for each point which could be 258

univocally linked to an upstream point in the network (Fig. 7). As expected, the prevalence index of M. 259

chelonae significantly increased with distance along the network (M. chelonae variations significantly 260

higher than 0 on the x axis, p < 0.05), and this increase was linearly correlated with the decrease of M. 261

llatzerense prevalence index (Pearson's product-moment correlation, R = -0.76, 95% confidence 262

interval [-0.91, -0.44], t15 = -4.54, p < 0.001, Fig. 7). Thus, the increase of M. chelonae prevalence 263

index with progression along the water distribution network was associated with a decline in M. 264

llatzerense prevalence index, indicating RGM communtity species rearrangements along water 265

distribution network. 266

267

DISCUSSION 268

This is the first large-scale survey describing RGM diversity, prevalence, and species composition in an 269

entire water treatment and distribution system. We report the first extensive sampling procedure in the 270

water network of a large city, in this case Paris. This large-scale sampling strategy enabled us to 271

estimate mean RGM detection probabilites, and their associated confidence intervals both locally (at 272

each tested location of the water system) and for the entire system. This was done for each species 273

independently and for all RGM grouped together. The resulting estimated RGM prevalence indexes 274

were suitable for reliable comparisons between species and between sampling sites. RGM were 275

ubiquitous (overall prevalence index >70%), with a continuous increase along the network to 276

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

14

prevalence index values exceeding 95% in the distribution units. The presence of RGM within water 277

production and distribution networks and their increase in the terminal ends of the networks have been 278

reported by studies on NTM as a whole (40, 60). However, our study specifically focusing on RGM 279

shows that the prevalence index and species diversity are much higher than reported in previous 280

studies. For example, the 2000-2001 study on NTM in the Parisian water network reported only 16% of 281

samples to be RGM-positive, and the isolation of only three different RGM species (44). Other studies 282

on NTM in drinking water networks similarly reported no more than three or four RGM species. We 283

detected nine RGM species, and possibly eight putative new RGM species on the basis of their rpoB 284

sequences (>3% divergence from any other rpoB sequence (1-6)). Only one of the eight sequences 285

possibly corresponding to putative new RGM species has been reported previously (NLJvIW_016 rpoB 286

sequence, see below) (37). We are currently characterizing four of them —NLJvIW_016, 287

parisRGMnew_1, parisRGMnew_3, and parisRGMnew_4— in more detail. 288

There are various possibly reasons why we detected so many RGM species. First, the water system of a 289

large city like Paris may be seen as being composed of different ecosystems that may be associated 290

with specific RGM species communities. Studies focusing on a restricted part of the water system (like 291

all previous studies) are very likely to be subject to strong sampling bias, and to fail to detect species 292

that are associated with other parts of the water system. Second, some species were detected at a very 293

low frequency, and were only detected because of the high number of isolates analyzed (630 isolates 294

subjected to rpoB sequencing). Previous small sampling strategies (31, 34, 35, 44) thus gave a biased 295

picture of the RGM community in the water system, underestimating species diversity. Third, recent 296

improvements in molecular identification of RGM allowed us to identify an unprecedented proportion 297

of the isolates by rpoB sequencing, and to discover several putative new species (61). 298

M. chelonae and M. llatzerense are by far the most prevalent RGM species in the Paris water 299

distribution network. M. chelonae has been found in urban water systems worldwide (38, 39, 59), but 300

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

15

its prevalence index and spatial distribution have never been estimated. M. llatzerense has been isolated 301

from tap water samples (59, 62), but has not previously been detected in water distribution networks 302

(34, 38, 44, 59), most probably because this species had not been described at the time these studies 303

were performed (62). Consistent with our results, M. peregrinum (2, 31) M. septicum (31, 35, 37) and 304

M. salmoniphilum (60, 63) (37, 63) have been isolated from similar water systems. These observations 305

suggest that the RGM species community groups identified in the Parisian water distribution system 306

may be common to other urban water distribution networks. However, differences in water origin, 307

climate or treatment processes may affect RGM community groups. Indeed, some studies in Mexico, 308

Greece and South African cities reported very different RGM species communities (31, 35, 64). 309

Differences in RGM isolation protocols may also contribute to these observed differences. 310

Our data show clearly that the RGM species present in the water distribution network form two types 311

of communities that are determined by the origin of the water (surface versus groundwater). This 312

possibility was suggested by the 2000-2001 survey (44), which described differences between the 313

NTM species detected in the distribution units studied; however, it was not possible to determine 314

whether NTM species compositions where shaped by the production units per se, or by the water origin 315

or both. Our study clarifies this point and demonstrates, with robust statistics, that water origin is a 316

major factor shaping the structure of RGM communities (M. llatzerense, M. salmoniphilum and 317

ParisRGMnew3 for units fed with groundwater versus M. chelonae and NLJvW_016 for those fed with 318

surface water). Our data are consistent with the study of van Ingen et al. in the Netherlands, which 319

reported the frequent recovery of M. llatzerense and M. salmoniphilum from household tap water of 320

groundwater origin (37). The association of M. llatzerense and M. salmoniphilum with groundwater 321

may reflect adaptation to lower temperatures for these species, which have a more psychrophilic profile 322

than other mycobacteria, such as M. chelonae and M fortuitum (62, 63). Groundwater temperatures are 323

not subject to seasonal variation and are substantially lower than surface water temperatures (13°C on 324

average with maxima of 15°C, versus 15°C on average with maxima of 21°C, respectively in water 325

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

16

storage tanks (data not shown)). No “thermophilic” RGM (e.g., species from the M. smegmatis group) 326

were isolated in our survey, consistent with temperature having an influence on the composition of 327

RGM communities. Other biotic and abiotic factors, such as non mycobacterial microbial flora, 328

treatment parameters, nitrate concentrations or total organic carbon, which often differ between surface 329

and groundwater, may also be involved. We also observed a tendency of communities to be composed 330

of phylogenetically related species. For example, the species of the community associated with surface 331

water treatment plants – M. septicum, M. peregrinum and ParisRGMnew_4 – all belonged to the M. 332

fortuitum phylogenetic group. 333

The collection of a representative dataset from all along the water distribution network (production site 334

exits, water storage tanks and distribution units) allowed, for the first time, to track “vertical” changes 335

along the distribution network. Our results demonstrate, with robust statistical support, changes in 336

species composition along water distribution network, as suggested by the 2000-2001 survey (44). The 337

most notable feature was the increase of M. chelonae prevalence index and the parallel decrease of M. 338

llatzerense prevalence index as water progressed along the network. This could result from the 339

competitive exclusion of M. llatzerense by M. chelonae and/or from changes in environmental 340

conditions (temperature, chlorine concentration, etc.) favoring the development of M. chelonae. The 341

propagation of M. chelonae along the network may involve its particular ability to form biofilms (39, 342

41) and/or to resist chlorination (42). Various environmental factors influencing overall RGM presence 343

have been described (65). We are currently studying physico-chemical variables that may influence 344

RGM species composition dynamics, and more specifically the adaptive success of M. chelonae, in this 345

water distribution network. 346

ACKNOWLEDGMENTS 347

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

17

We thank Dr Joël Pothier, Dr Guillaume Achaz, and all the members of the 'Atelier de bioinformatique 348

de Jussieu' for helpful advice for statistical analysis. We also thank Dr Jean-François Humbert for 349

careful reading of the manuscript. 350

351

REFERENCES 352 1. Bland, C. S., J. M. Ireland, E. Lozano, M. E. Alvarez, and T. P. Primm. 2005. Mycobacterial 353 ecology of the Rio Grande. Applied and environmental microbiology 71:5719-5727. 354 2. Covert, T. C., M. R. Rodgers, A. L. Reyes, and G. N. Stelma, Jr. 1999. Occurrence of 355 nontuberculous mycobacteria in environmental samples. Applied and environmental 356 microbiology 65:2492-2496. 357 3. Falkinham, J. O., 3rd. 2009. Surrounded by mycobacteria: nontuberculous mycobacteria in 358 the human environment. Journal of applied microbiology 107:356-367. 359 4. Lavania, M., K. Katoch, D. Parashar, P. Sharma, R. Das, D. S. Chauhan, V. D. Sharma, and 360 V. M. Katoch. 2008. Predominance of Mycobacterium fortuitum-chelonae complex in 361 Ghatampur field area, endemic for leprosy. Indian journal of leprosy 80:323-330. 362 5. Lee, E. S., Lee, M. Y., Han, S.H., & Ka, J.O. 2008. Occurence and molecular differentiation of 363 environmental mycobacteria in surface waters. J Microbiol Biotechnol 18(7):1207-1215. 364 6. Parashar, D., R. Das, D. S. Chauhan, V. D. Sharma, M. Lavania, V. S. Yadav, S. V. Chauhan, 365 and V. M. Katoch. 2009. Identification of environmental mycobacteria isolated from Agra, 366 north India by conventional & molecular approaches. The Indian journal of medical 367 research 129:424-431. 368 7. Parrish, S. C., J. Myers, and A. Lazarus. 2008. Nontuberculous mycobacterial pulmonary 369 infections in Non-HIV patients. Postgraduate medicine 120:78-86. 370 8. Primm, T. P., C. A. Lucero, and J. O. Falkinham, 3rd. 2004. Health impacts of 371 environmental mycobacteria. Clinical microbiology reviews 17:98-106. 372 9. van Ingen, J., M. J. Boeree, P. N. Dekhuijzen, and D. van Soolingen. 2009. Environmental 373 sources of rapid growing nontuberculous mycobacteria causing disease in humans. Clinical 374 microbiology and infection : the official publication of the European Society of Clinical 375 Microbiology and Infectious Diseases 15:888-893. 376 10. Griffith, D. E., T. Aksamit, B. A. Brown-Elliott, A. Catanzaro, C. Daley, F. Gordin, S. M. 377 Holland, R. Horsburgh, G. Huitt, M. F. Iademarco, M. Iseman, K. Olivier, S. Ruoss, C. F. 378 von Reyn, R. J. Wallace, Jr., and K. Winthrop. 2007. An official ATS/IDSA statement: 379 diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. American 380 journal of respiratory and critical care medicine 175:367-416. 381 11. Ingram, C. W., D. C. Tanner, D. T. Durack, G. W. Kernodle, Jr., and G. R. Corey. 1993. 382 Disseminated infection with rapidly growing mycobacteria. Clinical infectious diseases : an 383 official publication of the Infectious Diseases Society of America 16:463-471. 384 12. Lai, C. C., C. K. Tan, C. H. Chou, H. L. Hsu, C. H. Liao, Y. T. Huang, P. C. Yang, K. T. Luh, and 385 P. R. Hsueh. 2010. Increasing incidence of nontuberculous mycobacteria, Taiwan, 2000-386 2008. Emerging infectious diseases 16:294-296. 387 13. Merlani, G. M., and P. Francioli. 2003. Established and emerging waterborne nosocomial 388 infections. Current opinion in infectious diseases 16:343-347. 389

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

18

14. Wallace, R. J., Jr., J. M. Swenson, V. A. Silcox, R. C. Good, J. A. Tschen, and M. S. Stone. 390 1983. Spectrum of disease due to rapidly growing mycobacteria. Reviews of infectious 391 diseases 5:657-679. 392 15. Cayer, M. P., M. Veillette, P. Pageau, R. Hamelin, M. J. Bergeron, A. Meriaux, Y. Cormier, 393 and C. Duchaine. 2007. Identification of mycobacteria in peat moss processing plants: 394 application of molecular biology approaches. Canadian journal of microbiology 53:92-99. 395 16. Kankya, C., A. Muwonge, B. Djonne, M. Munyeme, J. Opuda-Asibo, E. Skjerve, J. Oloya, 396 V. Edvardsen, and T. B. Johansen. 2011. Isolation of non-tuberculous mycobacteria from 397 pastoral ecosystems of Uganda: public health significance. BMC public health 11:320. 398 17. Mangione, E. J., G. Huitt, D. Lenaway, J. Beebe, A. Bailey, M. Figoski, M. P. Rau, K. D. 399 Albrecht, and M. A. Yakrus. 2001. Nontuberculous mycobacterial disease following hot 400 tub exposure. Emerging infectious diseases 7:1039-1042. 401 18. Nakanaga, K., Y. Hoshino, Y. Era, K. Matsumoto, Y. Kanazawa, A. Tomita, M. Furuta, M. 402 Washizu, M. Makino, and N. Ishii. 2011. Multiple cases of cutaneous Mycobacterium 403 massiliense infection in a "hot spa" in Japan. Journal of clinical microbiology 49:613-617. 404 19. Wilson, R. W., V. A. Steingrube, E. C. Bottger, B. Springer, B. A. Brown-Elliott, V. Vincent, 405 K. C. Jost, Jr., Y. Zhang, M. J. Garcia, S. H. Chiu, G. O. Onyi, H. Rossmoore, D. R. Nash, and 406 R. J. Wallace, Jr. 2001. Mycobacterium immunogenum sp. nov., a novel species related to 407 Mycobacterium abscessus and associated with clinical disease, pseudo-outbreaks and 408 contaminated metalworking fluids: an international cooperative study on mycobacterial 409 taxonomy. International journal of systematic and evolutionary microbiology 51:1751-410 1764. 411 20. Duarte, R. S., M. C. Lourenco, S. Fonseca Lde, S. C. Leao, L. Amorim Ede, I. L. Rocha, F. S. 412 Coelho, C. Viana-Niero, K. M. Gomes, M. G. da Silva, N. S. Lorena, M. B. Pitombo, R. M. 413 Ferreira, M. H. Garcia, G. P. de Oliveira, O. Lupi, B. R. Vilaca, L. R. Serradas, A. Chebabo, 414 E. A. Marques, L. M. Teixeira, M. Dalcolmo, S. G. Senna, and J. L. Sampaio. 2009. 415 Epidemic of postsurgical infections caused by Mycobacterium massiliense. Journal of 416 clinical microbiology 47:2149-2155. 417 21. Kim, H. Y., Y. J. Yun, C. G. Park, D. H. Lee, Y. K. Cho, B. J. Park, S. I. Joo, E. C. Kim, Y. J. Hur, 418 B. J. Kim, and Y. H. Kook. 2007. Outbreak of Mycobacterium massiliense infection 419 associated with intramuscular injections. Journal of clinical microbiology 45:3127-3130. 420 22. Kluger, N., C. Muller, and N. Gral. 2008. Atypical mycobacteria infection following 421 tattooing: review of an outbreak in 8 patients in a French tattoo parlor. Archives of 422 dermatology 144:941-942. 423 23. Munayco, C. V., C. G. Grijalva, D. R. Culqui, J. L. Bolarte, L. A. Suarez-Ognio, N. Quispe, R. 424 Calderon, L. Ascencios, M. Del Solar, M. Salomon, F. Bravo, and E. Gotuzzo. 2008. 425 Outbreak of persistent cutaneous abscesses due to Mycobacterium chelonae after 426 mesotherapy sessions, Lima, Peru. Revista de saude publica 42:146-149. 427 24. Murillo, J., J. Torres, L. Bofill, A. Rios-Fabra, E. Irausquin, R. Isturiz, M. Guzman, J. 428 Castro, L. Rubino, and M. Cordido. 2000. Skin and wound infection by rapidly growing 429 mycobacteria: an unexpected complication of liposuction and liposculpture. The 430 Venezuelan Collaborative Infectious and Tropical Diseases Study Group. Archives of 431 dermatology 136:1347-1352. 432 25. Winthrop, K. L., M. Abrams, M. Yakrus, I. Schwartz, J. Ely, D. Gillies, and D. J. Vugia. 433 2002. An outbreak of mycobacterial furunculosis associated with footbaths at a nail salon. 434 The New England journal of medicine 346:1366-1371. 435 26. Carbonne, A., F. Brossier, I. Arnaud, I. Bougmiza, E. Caumes, J. P. Meningaud, S. 436 Dubrou, V. Jarlier, E. Cambau, and P. Astagneau. 2009. Outbreak of nontuberculous 437

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

19

mycobacterial subcutaneous infections related to multiple mesotherapy injections. Journal 438 of clinical microbiology 47:1961-1964. 439 27. Dailloux, M., M. L. Abalain, C. Laurain, L. Lebrun, C. Loos-Ayav, A. Lozniewski, and J. 440 Maugein. 2006. Respiratory infections associated with nontuberculous mycobacteria in 441 non-HIV patients. The European respiratory journal : official journal of the European 442 Society for Clinical Respiratory Physiology 28:1211-1215. 443 28. Dytoc, M. T., L. Honish, C. Shandro, P. T. Ting, L. Chui, L. Fiorillo, J. Robinson, A. 444 Fanning, G. Predy, and R. P. Rennie. 2005. Clinical, microbiological, and epidemiological 445 findings of an outbreak of Mycobacterium abscessus hand-and-foot disease. Diagnostic 446 microbiology and infectious disease 53:39-45. 447 29. Lee, W. J., T. W. Kim, K. B. Shur, B. J. Kim, Y. H. Kook, J. H. Lee, and J. K. Park. 2000. 448 Sporotrichoid dermatosis caused by Mycobacterium abscessus from a public bath. The 449 Journal of dermatology 27:264-268. 450 30. Rahav, G., S. Pitlik, Z. Amitai, A. Lavy, M. Blech, N. Keller, G. Smollan, M. Lewis, and A. 451 Zlotkin. 2006. An outbreak of Mycobacterium jacuzzii infection following insertion of 452 breast implants. Clinical infectious diseases : an official publication of the Infectious 453 Diseases Society of America 43:823-830. 454 31. Castillo-Rodal, A. I., M. Mazari-Hiriart, L. T. Lloret-Sanchez, B. Sachman-Ruiz, P. 455 Vinuesa, and Y. Lopez-Vidal. 2012. Potentially pathogenic nontuberculous mycobacteria 456 found in aquatic systems. Analysis from a reclaimed water and water distribution system in 457 Mexico City. European journal of clinical microbiology & infectious diseases : official 458 publication of the European Society of Clinical Microbiology 31:683-694. 459 32. Gomila, M., A. Ramirez, and J. Lalucat. 2007. Diversity of environmental Mycobacterium 460 isolates from hemodialysis water as shown by a multigene sequencing approach. Applied 461 and environmental microbiology 73:3787-3797. 462 33. Hilborn, E. D., T. C. Covert, M. A. Yakrus, S. I. Harris, S. F. Donnelly, E. W. Rice, S. Toney, 463 S. A. Bailey, and G. N. Stelma, Jr. 2006. Persistence of nontuberculous mycobacteria in a 464 drinking water system after addition of filtration treatment. Applied and environmental 465 microbiology 72:5864-5869. 466 34. Santos, R., F. Oliveira, J. Fernandes, S. Goncalves, F. Macieira, and M. Cadete. 2005. 467 Detection and identification of mycobacteria in the Lisbon water distribution system. Water 468 science and technology : a journal of the International Association on Water Pollution 469 Research 52:177-180. 470 35. September, S. M., V. S. Brozel, and S. N. Venter. 2004. Diversity of nontuberculoid 471 Mycobacterium species in biofilms of urban and semiurban drinking water distribution 472 systems. Applied and environmental microbiology 70:7571-7573. 473 36. Shin, J. H., H. K. Lee, E. J. Cho, J. Y. Yu, and Y. H. Kang. 2008. Targeting the rpoB gene 474 using nested PCR-restriction fragment length polymorphism for identification of 475 nontuberculous mycobacteria in hospital tap water. J Microbiol 46:608-614. 476 37. van Ingen, J., H. Blaak, J. de Beer, A. M. de Roda Husman, and D. van Soolingen. 2010. 477 Rapidly growing nontuberculous mycobacteria cultured from home tap and shower water. 478 Applied and environmental microbiology 76:6017-6019. 479 38. Tsintzou, A., A. Vantarakis, O. Pagonopoulou, A. Athanassiadou, and M. 480 Papapetropoulou. 2000. Environmental Mycobacteria in Drinking Water Before and After 481 Replacement of the Water Distribution Network. Water, Air, & Soil Pollution 120:273-282. 482 39. Schulze-Robbecke, R., B. Janning, and R. Fischeder. 1992. Occurrence of mycobacteria in 483 biofilm samples. Tubercle and lung disease : the official journal of the International Union 484 against Tuberculosis and Lung Disease 73:141-144. 485

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

20

40. Torvinen, E., S. Suomalainen, M. J. Lehtola, I. T. Miettinen, O. Zacheus, L. Paulin, M. L. 486 Katila, and P. J. Martikainen. 2004. Mycobacteria in water and loose deposits of drinking 487 water distribution systems in Finland. Applied and environmental microbiology 70:1973-488 1981. 489 41. Williams, M. M., M. A. Yakrus, M. J. Arduino, R. C. Cooksey, C. B. Crane, S. N. Banerjee, 490 E. D. Hilborn, and R. M. Donlan. 2009. Structural analysis of biofilm formation by rapidly 491 and slowly growing nontuberculous mycobacteria. Applied and environmental 492 microbiology 75:2091-2098. 493 42. Le Dantec, C., J. P. Duguet, A. Montiel, N. Dumoutier, S. Dubrou, and V. Vincent. 2002. 494 Chlorine disinfection of atypical mycobacteria isolated from a water distribution system. 495 Applied and environmental microbiology 68:1025-1032. 496 43. Schwartz, T., S. Hoffmann, and U. Obst. 2003. Formation of natural biofilms during 497 chlorine dioxide and u.v. disinfection in a public drinking water distribution system. Journal 498 of applied microbiology 95:591-601. 499 44. Le Dantec, C., J. P. Duguet, A. Montiel, N. Dumoutier, S. Dubrou, and V. Vincent. 2002. 500 Occurrence of mycobacteria in water treatment lines and in water distribution systems. 501 Applied and environmental microbiology 68:5318-5325. 502 45. Adekambi, T., and M. Drancourt. 2004. Dissection of phylogenetic relationships among 19 503 rapidly growing Mycobacterium species by 16S rRNA, hsp65, sodA, recA and rpoB gene 504 sequencing. International journal of systematic and evolutionary microbiology 54:2095-505 2105. 506 46. Adekambi, T., T. M. Shinnick, D. Raoult, and M. Drancourt. 2008. Complete rpoB gene 507 sequencing as a suitable supplement to DNA-DNA hybridization for bacterial species and 508 genus delineation. International journal of systematic and evolutionary microbiology 509 58:1807-1814. 510 47. Macheras, E., A. L. Roux, S. Bastian, S. C. Leao, M. Palaci, V. Sivadon-Tardy, C. Gutierrez, 511 E. Richter, S. Rusch-Gerdes, G. Pfyffer, T. Bodmer, E. Cambau, J. L. Gaillard, and B. Heym. 512 2011. Multilocus sequence analysis and rpoB sequencing of Mycobacterium abscessus 513 (sensu lato) strains. Journal of clinical microbiology 49:491-499. 514 48. Adekambi, T., P. Berger, D. Raoult, and M. Drancourt. 2006. rpoB gene sequence-based 515 characterization of emerging non-tuberculous mycobacteria with descriptions of 516 Mycobacterium bolletii sp. nov., Mycobacterium phocaicum sp. nov. and Mycobacterium 517 aubagnense sp. nov. International journal of systematic and evolutionary microbiology 518 56:133-143. 519 49. Adekambi, T., P. Colson, and M. Drancourt. 2003. rpoB-based identification of 520 nonpigmented and late-pigmenting rapidly growing mycobacteria. Journal of clinical 521 microbiology 41:5699-5708. 522 50. Adekambi, T., M. Reynaud-Gaubert, G. Greub, M. J. Gevaudan, B. La Scola, D. Raoult, and 523 M. Drancourt. 2004. Amoebal coculture of "Mycobacterium massiliense" sp. nov. from the 524 sputum of a patient with hemoptoic pneumonia. Journal of clinical microbiology 42:5493-525 5501. 526 51. Dereeper, A., V. Guignon, G. Blanc, S. Audic, S. Buffet, F. Chevenet, J. F. Dufayard, S. 527 Guindon, V. Lefort, M. Lescot, J. M. Claverie, and O. Gascuel. 2008. Phylogeny.fr: robust 528 phylogenetic analysis for the non-specialist. Nucleic acids research 36:W465-469. 529 52. Edgar, R. C. 2004. MUSCLE: a multiple sequence alignment method with reduced time and 530 space complexity. BMC bioinformatics 5:113. 531

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

21

53. Bolker, B. M., M. E. Brooks, C. J. Clark, S. W. Geange, J. R. Poulsen, M. H. Stevens, and J. 532 S. White. 2009. Generalized linear mixed models: a practical guide for ecology and 533 evolution. Trends in ecology & evolution 24:127-135. 534 54. Ihaka, R., and R. Gentleman. 1996. R: A Language for Data Analysis and Graphics. Journal 535 of Computational and Graphical Statistics 5:299-314. 536 55. Ripley, B. D. 2002. Modern applied statistics with S. Springer. 537 56. Pinheiro, J. C., and D. M. Bates. 2000. Mixed effects models in S and S-PLUS. Springer 538 Verlag. 539 57. Suzuki, R., and H. Shimodaira. 2006. Pvclust: an R package for assessing the uncertainty 540 in hierarchical clustering. Bioinformatics 22:1540-1542. 541 58. Beh, E. J. 2012. Exploratory multivariate analysis by example using R. Journal of Applied 542 Statistics 39:1381-1382. 543 59. Vaerewijck, M. J., G. Huys, J. C. Palomino, J. Swings, and F. Portaels. 2005. Mycobacteria 544 in drinking water distribution systems: ecology and significance for human health. FEMS 545 microbiology reviews 29:911-934. 546 60. Falkinham, J. O., 3rd, C. D. Norton, and M. W. LeChevallier. 2001. Factors influencing 547 numbers of Mycobacterium avium, Mycobacterium intracellulare, and other Mycobacteria 548 in drinking water distribution systems. Applied and environmental microbiology 67:1225-549 1231. 550 61. Tortoli, E. 2006. The new mycobacteria: an update. FEMS immunology and medical 551 microbiology 48:159-178. 552 62. Gomila, M., A. Ramirez, J. Gasco, and J. Lalucat. 2008. Mycobacterium llatzerense sp. nov., 553 a facultatively autotrophic, hydrogen-oxidizing bacterium isolated from haemodialysis 554 water. International journal of systematic and evolutionary microbiology 58:2769-2773. 555 63. Whipps, C. M., W. R. Butler, F. Pourahmad, V. G. Watral, and M. L. Kent. 2007. Molecular 556 systematics support the revival of Mycobacterium salmoniphilum (ex Ross 1960) sp. nov., 557 nom. rev., a species closely related to Mycobacterium chelonae. International journal of 558 systematic and evolutionary microbiology 57:2525-2531. 559 64. Kormas, K. A., C. Neofitou, M. Pachiadaki, and E. Koufostathi. 2010. Changes of the 560 bacterial assemblages throughout an urban drinking water distribution system. 561 Environmental monitoring and assessment 165:27-38. 562 65. Jacobs, J., M. Rhodes, B. Sturgis, and B. Wood. 2009. Influence of environmental gradients 563 on the abundance and distribution of Mycobacterium spp. in a coastal lagoon estuary. 564 Applied and environmental microbiology 75:7378-7384. 565 566

567 568

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

TABLE 1. GI numbers of rpoB sequences used for alignment and distance tree construction.

6272066745477628345957507817276914887905062720659169264484213876752627206693154356056272066116101597421387673862720663312451981510384276021885531543560718357928031087710831543559521995809162720653307147893346230977120401028:1347927-1351880 3464271543459575645477708108796981:1071942-1075442 290745667148879066148879080148879078148879072148879074296011771118462219:c4634813-4631379 183980035:1213006-1216536

Mycobacterium gadium CIP 105388 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium goodii strain CIP 106349 RNA polymerase (rpoB) gene, partial cds Mycobacterium immunogenum strain CIP 106684 RpoB gene, complete cds Mycobacterium jacuzzii RpoB (rpoB) gene, partial cds Mycobacterium llatzerense partial rpoB gene, strain MG12 Mycobacterium moriokaense CIP 105393 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium mucogenicum strain FI-07073 rpoB gene, partial cds Mycobacterium murale strain DSM 44428 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium novocastrense CIP 105546 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium peregrinum strain FI-09182 RNA polymerase beta-subunit (rpoB) gene, partial cds Mycobacterium phlei CIP 105389 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium phocaicum strain CIP 108542 RNA polymerase beta subunit (rpoB) gene, partial cds Mycobacterium porcinum strain FI-08029 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium pulveris CIP 106804 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium salmoniphilum strain NVI_6609 RNA polymerase beta subunit (rpoB) gene, partial cds Mycobacterium senegalense strain ATCC BAA-849 RNA polymerase beta gene, partial cds Mycobacterium septicum strain D13 RpoB (rpoB) gene, partial cds Mycobacterium setense strain FI-09152 RNA polymerase beta-subunit (rpoB) gene, partial cds Mycobacterium smegmatis ATCC:14468 RpoB (rpoB) gene, partial cds Mycobacterium sp. M05 RNA polymerase subunit B (rpoB) gene, partial cds Mycobacterium tusciae strain CIP10667 RNA polymerase beta-subunit (rpoB) gene, partial cds Mycobacterium wolinskyi clone CRM-0267 RNA polymerase beta subunit (rpoB) gene, partial cds Mycobacterium barrassiae CIP 108545 RNA polymerase subunit B (rpoB) gene, complete cds Mycobacterium flavescens strain ATCC 14474 RNA polymerase subunit beta (rpoB), partial cds Mycobacterium mageritense strain ATCC 700351 RNA polymerase beta subunit (rpoB) gene, partial cdsMycobacterium vanbaalenii PYR-1 chromosome, complete genome Mycobacterium neoaurum strain ATCC 25795 RNA polymerase beta subunit (rpoB) gene, partial cds Mycobacterium farcinogenes strain 3753 RpoB gene, complete cds Mycobacterium rhodesiae strain CIP 106806 RNA polymerase (rpoB) gene, partial cdsMycobacterium sp. MCS chromosome, complete genome Mycobacterium sp. NL-JvIW-001 RNA polymerase beta subunit (rpoB) gene, partial cds Mycobacterium sp. MG20 partial rpoB gene, strain MG20 Mycobacterium sp. MHSD12 partial rpoB gene, strain MHSD12 Mycobacterium sp. MHSD11 partial rpoB gene, strain MHSD11 Mycobacterium sp. MHSD4 partial rpoB gene, strain MHSD4 Mycobacterium sp. MHSD5 partial rpoB gene, strain MHSD5 Mycobacterium sp. NL-JvIW-016 RNA polymerase beta subunit (rpoB) gene, partial cdsMycobacterium avium 104, complete genomeMycobacterium marinum M chromosome, complete genome

a : for complete genome, sequence coordinate on chromosome is indicated

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on Novem

ber 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from