• Aucun résultat trouvé

Occurrence and diversity of both bacterial and fungal communities in dental unit waterlines subjected to disinfectants.

N/A
N/A
Protected

Academic year: 2021

Partager "Occurrence and diversity of both bacterial and fungal communities in dental unit waterlines subjected to disinfectants."

Copied!
11
0
0

Texte intégral

(1)

doi: 10.1093/femspd/ftw094

Advance Access Publication Date: 13 September 2016 Research Article

R E S E A R C H A R T I C L E

Occurrence and diversity of both bacterial and fungal

communities in dental unit waterlines subjected to

disinfectants

Damien Costa

1

,

2

,

, Anne Mercier

1

,

3

, Kevin Gravouil

3

, J ´er ˆome Lesobre

4

,

Julien Verdon

1

and Christine Imbert

1

,

5

1

Team Water Microbiology, Ecology and Biology of Interactions, CNRS, University of Poitiers, 86073 Poitiers

Cedex, France,

2

Bacteriology and Hospital Hygiene Laboratory, University Hospital of Poitiers 86021 Poitiers

Cedex, France,

3

Cooperative Laboratory Thanaplast SP-EBI-Carbios Bioplastics, Ecology and Biology of

interactions, CNRS UMR 7267, University of Poitiers 86073 Poitiers Cedex, France,

4

Microorganisms laboratory:

Genome and environment, UMR CNRS 6023, University Blaise Pascal 63177 Aubi `ere Cedex, France and

5

University of Medicine Pharmacy of Poitiers 86073 Poitiers Cedex, France

Corresponding author: Team Water Microbiology, Ecology and biology of interactions, CNRS, University of Poitiers, 86073 Poitiers Cedex, France. Tel/Fax: +33(2) 38-64-40-22; E-mail:damien.costa@univ-poitiers.fr

One sentence summary: This work focuses on the microbial ecology encountered in dental unit waterlines with specific interest given to human pathogens and influence of treatments.

Editor: Ake Forsberg

ABSTRACT

Chemical disinfectants are widely advocated to reduce the microbial contamination in dental unit waterlines (DUWL). However, until now their efficacy has been poorly examined after long-term application. In this study, through quantitative PCR and high-throughput sequencing, both bacterial and fungal communities were profiled from 8- to 12-year-old DUWL treated with disinfectants commonly used by European dentists. Water was collected from the tap water supplying units to the output exposure point of the turbine handpiece following a stagnation period and dental care activity. Results showed that (i) the unit itself is the principal source of microbial contamination and (ii) water stagnation, DU maintenance practices and quality of water supplying DU appeared as parameters driving the water quality. Despite disinfecting treatment combined to flushing process, the microbial contamination remained relevant in the studied output water, in association with a high bacterial and fungal diversity. The occurrence of potentially pathogenic microorganisms in these treated DUWL demonstrated a potential infectious risk for both patients and dental staff. A disinfectant shock before a prolonged

stagnation period could limit the microbial proliferation inside DUWL. Necessity to proceed to regular water quality control of DUWL was highlighted.

Keywords: dental unit waterlines (DUWL); disinfection; Oxygenal 6C; CalbeniumC; pyrosequencing; pathogens

Received: 4 June 2016; Accepted: 9 September 2016

C

 FEMS 2016. All rights reserved. For permissions, please e-mail:journals.permissions@oup.com

1

(2)

INTRODUCTION

Dental unit waterlines (DUWL) provide an environment con-ducive to microbial attachment and biofilm formation (Walker et al. 2000; Szyma ´nska 2005b; Szyma ´nska, Sitkowska and Dutkiewicz2008; Puttaiah et al.2012). DUWL are predominantly colonized by environmental microorganisms including bacte-ria, fungi and protozoa. However, some microorganisms com-monly found in the oral cavity, such as oral streptococci, or on skin, such as Staphylococcus aureus, have been also identi-fied in DUWL (Petti and Tarsitani2006; Petti et al.2013; Porteous et al.2014; Costa et al.2015). This suggests various origins from microbial contamination: microorganisms can be provided by the water supplying the dental chair unit, or through the suck back of biological fluids from oral cavity of patients resulting of a malfunction of antiretraction devices; finally, the continuous biofilm detachment or fragmentation also participates in this contamination (Pankhurst 2003; Wirthlin, Marshall and Row-land2003; Coleman et al.2009; O’Donnell et al.2011). There-fore, many studies are focused on the microbial contamina-tion of DUWL (Szyma ´nska2005a,b; Szyma ´nska, Sitkowska and Dutkiewicz2008; Kumar et al.2010; Arvand and Hack2013; Petti et al.2013; Costa et al.2015). Infections associated with this con-tamination are rarely reported (Martin 1987; Ricci et al.2012) maybe because of the difficulty to prove that the exposure to contaminated DUWL is at the origin of the infection or maybe because of the usually limited clinical significance of such infec-tions that do not result in case reports. Thus, infecinfec-tions related to contaminated dental output water may be more frequent than the number of corresponding case reports. In fact, both pa-tients and dental staff are regularly exposed to this infectious risk due to inhalation of aerosols produced during dental cares. In addition, dental unit water (DUW) may be ingested or may contaminate surgical wounds. Moreover, the quality of DUW is of importance since vulnerable patients (such as elderly, diabet-ics or immunocompromised patients) are frequent. The assess-ment of risks related to DUW is taken into account in guide-lines from governmental agencies such as the American Den-tal Association (ADA) and/or the Center for Disease Control and Prevention, and the current recommendation is that the dental output water as water delivered to dental unit (DU) be kept at or below existing standards for potable drinking water, i.e. with aerobic heterotrophic microbial density inferior to 500 CFU per mL (colony-forming units per millilitre) (American Dental As-sociation2012; Coleman et al.2014). Unexpectedly, in France as in most European countries, no standard defines the microbial quality of the dental output water delivered to patients.

Some non-chemical methods are recommended to reduce microbial contamination in DUWL output water such as the use of microbial filters and/or antiretraction valves, the achievement of regular flushing of DUWL or the sterilization of handpieces (Direction G ´en ´erale de la Sant ´e2006; O’Donnell et al.2011; Dal-lolio et al.2014; Smith and Smith2014). In addition, an efficient way to reduce microbial contamination and control biofilms in DUWL is the use of chemical treatments, especially based on hy-drogen peroxide and silver ions, chlorine dioxide, chlorhexidine, peracetic acid or citric acid (Schel et al.2006; Szyma ´nska2006). These agents can be used intermittently (e.g. daily or once to twice weekly) or continuously in DUWL supply water (O’Donnell et al.2011). Even if a number of studies have evaluated the effi-cacy of disinfectant agents to control the microbiological water quality inside DUWL, only a few were performed in real practice conditions after long-term application.

The aim of this study was to investigate both bacterial and fungal communities in DUWL treated with a disinfectant in real conditions, in actual daily working DUs and according to the flow from the water supply (incoming water) to the outer exposure point (output water). This study focused on DUWL subjected to CalbeniumC

or Oxygenal 6C

, two disinfectants commonly used by European dentists, for which no or a limited number of stud-ies were reported (Walker et al.2003; Schel et al.2006; Szyma ´nska

2006).

METHODS

DUs and disinfectants

Three 8- to 12-year-old DUs (named Unit 1 to 3) in operation in Poitiers (France) used daily for dental care to patients were inves-tigated. The tubing was never changed in the DUs displayed in this study. All the DUWL were supplied from the municipal wa-ter system of Poitiers, France (Supporting Information, Table SA) and subjected to flushing at the start of each day to reduce mi-crobial accumulation subsequently to overnight stagnation. In addition, as presented in the Table1, DUWL of Units 1 and 2 were continuously treated with 2% of Calbenium (Airel-Quetin, Champigny-sur-Marne, France) and DUWL of Unit 3 were con-tinuously treated with 0.3% of Oxygenal 6 (KaVo, Biberach, Germany). Importantly, for Unit 3, an additional treatment was implemented each week (on Friday) before the long water stag-nation period consequently to the inactivity of the weekend. The procedure consisted of a cycle of 45 min with Oxygenal 6 at 3% (v/v) circulated inside DUWL. Calbenium is a complex mixture based on ethylene diamine tetreacetic acid, benzalkonium chlo-ride, sodium tosylchloramide, allantoin, aspartam, sorbitol and spearmint or lemon aroma, whereas Oxygenal 6 is composed of hydrogen peroxide and silver ions.

Water sampling

DUW samples (1 L) were collected at two different time periods in routine dental practices from (i) tap water upstream from the unit (here named incoming water or IW); (ii) turbine handpiece output water after a 48-h stagnation period (after the weekend), on Monday morning and before the beginning of the working day (named output water after stagnation or OWS); and (iii) tur-bine handpiece output water immediately after dental care of the last patient of the sampling day (named output water after activity or OWA). In order to neutralize the residual disinfectant, water samples were stored in sterile bottles with 3 mL of sodium thiosulfate (18 g L−1) at 4◦C for no longer than 24 h before anal-ysis. During sampling, care measures were taken to ensure that the microorganisms did not originate from the sampler.

Total aerobic and aero-tolerant cultivable microbial biomass in DUW

Each DUW sample was analyzed by direct plate spread in duplicate using a WASP 2 spiral plater (Biomerieux, Marcy l’Etoile, France) on R2A (Difco, BD, Le Pont de Claix, France), medium specifically recommended in standard methods for heterotrophic plate counts of treated potable water (Reasoner and Geldreich1985). The number of CFU was counted 7 days af-ter plate incubation under aerobic conditions at 28◦C, as recom-mended by the manufacturer.

(3)

Table 1. Description of the DUs.

Origin Age (years) IW treatment OW treatment Concentration

Unit 1 Public clinic 12 Softener Calbenium Continuous 2.0% v/v

Unit 2 Public clinic 8 None Calbenium Continuous 2.0% v/v

Unit 3 Hospital 12 None Oxygenal 6 Continuous 0.3% v/v+ each week

(Friday) for 45 min 3.0% v/v

Total water DNA extraction

DNA was directly extracted from the DUWL samples filtered over a sterile 0.22-μm polycarbonate membrane (Sartorius, Dourdan, France) under laminar flow hood, as described by Costa et al. (2015). DNA extraction from water guaranteed DNA and RNA free filtered on polycarbonate membrane in the same conditions as for DUW samples was performed as negative control.

DNA concentrations of extracts were quantified in dupli-cate using SYBR Green I dye (Invitrogen, Villebon-sur-Yvette, France) and a standard curve of HindIII-digestedλ DNA frag-ments (Promega, Charbonni `eres-les-Bains, France) on a Light-Cycler 480 Instrument (Roche Applied Science, Meylan, France). Aliquots of 0.05 ngμL−1diluted DNA were stored at –20C ready

for molecular applications.

Quantitative polymerase chain reaction

Quantitative polymerase chain reaction (qPCR) was performed in a LightCycler FastStart DNA MasterPLUS Sybr Green I mix

(Roche Applied Science) in a LightCycler 480 Instrument (Roche Applied Science). All qPCR reactions were performed in dupli-cate for each DUW sample using the epMotionR 5070 pipetting robot (Eppendorf, Montesson, France). The 10μL reaction mix-ture contained 0.5μM of each primer, 1X of LightCycler FastStart DNA MasterPLUSSybr Green I mix (Roche Applied Science), 5μL

of H2O and 2μL of a 0.05 ng μL−1DNA-diluted template.

The copy number of the bacterial 16S rRNA gene was quan-tified using the 341F–515R primer set (Baker, Smith and Cowan

2003) as described in Costa et al. (2015). The FF390—FR1 primer set was used to access the copy number of the fungal 18S rRNA gene according to the amplification protocol previously described by Chemidlin Pr ´evost-Bour ´e et al. (2011). A standard curve was generated by performing serial dilutions of a known amount of a standard sample containing a fragment of the 18S rRNA gene of Candida albicans ATCC 3153. The standard curve encompassed 2× 102 to 2× 106 copies of the target sequence

per well. No-template control was run for each qPCR assay. The number of bacterial and fungal rRNA copies derived from the qPCR measurements was normalized to account for the fil-tered volume of water for each respective sample.

Pyrosequencing of 16S and 18S rRNA gene sequences

Bacterial and fungal diversities were determined for each DUW sample by 454 pyrosequencing of ribosomal genes. Extracted DNA from each duplicated DUW sample was pooled prior to the pyrosequencing. Each sample was prepared from the same DNA input quantity in order to normalize the libraries and achieve even representation of each library in the pyrosequencing re-sults. The V3–V5 region of the bacterial 16S rRNA genes was amplified using the 341F and 926R primers (Baker, Smith and Cowan2003) as described in Costa et al. (2015). The fungal 18S rRNA genes were amplified using the FF390 and FR1 primers

as described by Terrat et al. (2015). The PCRs were performed in triplicate for each sample and pooled before purification of the PCR products using the PCR clean-up kit according to the manufacturer’s instructions (Macherey-Nagel, Hoerdt, France) and quantification using SYBR Green I dye (Invitrogen). Purified PCR products were then specifically tagged in a second PCR of seven cycles, conducted under similar amplification conditions, using primers containing pyrosequencing adaptors and 10 base pair multiplex identifiers barcode added to one primer at the 5position to specifically identify each sample. Finally, bacte-rial and fungal PCR products were purified and quantified using the Quant-iT PicoGreen dsDNA Assay Kit (Life Technologies,R

Saint-Aubin, France). In addition, size and purity of amplicons were checked using the MultiNA bioanalyzer (Shimadzu, Marne-la-Vall ´ee, France). Preparation of PCR amplicon library from neg-ative controls of DNA extraction and amplification failed, ex-cluding contamination bias. Equimolar PCR products (1× 109

moleculesμL−1) were pooled for a single sequencing reaction

run for each community (i.e. bacteria and fungi). The Lib-L kit (Roche Applied Science) was used for emPCR and the unidirec-tional sequencing of the amplicon library. The two pyrosequenc-ing runs were conducted in a GS Junior 454 Sequencer (Roche Ap-plied Science) following the manufacturer’s recommendations. The raw datasets are available on the EBI database system in the Sequence Read Archive, under study accession number PR-JEB12425 (http://www.ebi.ac.uk/ena).

Bioinformatics analysis of 16S and 18S rRNA gene sequences

Pyrosequencing data were analyzed using the GnS-PIPE (ver-sion 1.1.11) pipeline described by Terrat et al. (2015), based on the parameters detailed in SM (Supporting Information, Ta-ble SB). Briefly, raw reads were sorted according to identifier se-quences. All reads with mismatches in the primer sequence, ambiguities in the sequence or sequences inferior to a mini-mal length were discarded. Rigorous dereplication (i.e. cluster-ing of strictly identical sequences) was performed uscluster-ing a PERL program. The retained dereplicated reads were then aligned us-ing INFERNAL alignment (Cole et al. 2009) and clustered into operational taxonomic units (OTU) as described by Terrat et al. (2015). All the retained high-quality reads were taxonomically assigned according to the Silva r114 reference database (Quast et al.2013). OTUs were clustered with 5% sequence dissimilar-ity cut-off at the genus level to obtain reliable representation of bacterial and fungal communities through taxonomic clas-sification as previously explained in Costa et al. (2015). Dur-ing the analysis, all sDur-ingletons correspondDur-ing to reads detected only once and not clustered (that might be artefacts such as PCR chimeras and large sequencing errors produced by the PCR and the pyrosequencing) were checked taking into ac-count the quality of their taxonomic assignments (Terrat et al.

2015). In order to avoid biased community comparisons, the sample reads were reduced by random selection closed to the

(4)

lowest datasets. The retained high-quality reads were used for taxonomy-independent analyses including estimation of di-versity indices and taxonomy-based analyses using similarity approaches against Silva r114 reference databases and post-processed using R package (R Development Core Team 2004). Heatmaps were built up from the relative abundance values of the most dominant bacterial and fungal genera across the samples (relative abundance>1%) using the gplots R package (Warnes et al.2015).

Statistical analysis

CFU and qPCR data were analyzed using non-parametric tests with significance assessed at the level of P< 0.05. All statistical analyses were performed under R project (R Development Core Team 2004).

RESULTS AND DISCUSSION

Dental microbiological water quality investigations were per-formed thanks to the voluntary participation of 10 practitioners from hospitals and public dental clinics of the region of Poitiers (France). Among the DU put at our disposal, three 8- to 12-year-old DUs were continuously treated with disinfectants (Calbe-nium or Oxygenal 6). Except for the disinfectant application, the three offices presented similar features in terms of type of den-tal cares carried out and sterilization procedures that limited the influence to other parameters than disinfectant on the mi-crobial community encountered inside DUWL. These DUs were sampled to investigate occurrence and diversity of both bacte-rial and fungal communities during routine dental practice and accordingly to the flow from the water supply (incoming wa-ter: IW) to the outer exposure point (output water after stagna-tion period: OWS and output water after activity period: OWA). The IW analysis reflected the microbiological quality of the wa-ter supplying DUs. As wawa-ter stagnation promotes growth and the proliferation of biofilm within DUWL (Kumar et al.2010), the OWS can be considered as an indicator of the microor-ganisms present in the biofilm in the absence of feasibility to collect DUWL tubing. Microorganisms present in both OW sam-ples (OWS and OWA) may originate from IW or from the re-traction of oral fluids in DUWL. Contrary to the OWS, the OWA was exposed all day long to a chemically treated water flow and consequently was considered as an indicator of the treatment efficiency.

Bacterial and fungal abundances in DUWL: influence of disinfectant

Occurrence of bacterial and fungal communities was estimated by quantitative PCR targeting the 16S rRNA and 18S rRNA genes, respectively. Results sustained colonization by both bacteria and fungi in the DUWL, with levels varying according to each DU and to the water flow inside DU (Table2a and b). Based on the num-ber of rRNA gene copy in microbial genomes, fungi would be less prevalent and significantly less abundant than bacteria in sampled DUWL. Interestingly, 18S rRNA gene copies from Amoe-bae spp. were also recovered from our samples by qPCR (Sup-porting Information, Table SC) based on the protocol from Le Calvez et al. (2012). Due to their encystment ability, free-living amoebae (FLA) may protect some microorganisms from envi-ronmental stress, as already described for Legionella pneumophila (Greub and Raoult2004; Dey et al.2009), Candida spp (Barbot et al.2014) or Mycobacterium spp (Delafont et al.2014). Thus, oc-currence of FLA may contribute to the propagation of microor-ganisms able to resist or even escape to phagocytosis, includ-ing pathogens, and to their efficient recolonization elsewhere in DUWL. Altogether, qPCR investigations showed the coloniza-tion of DUWL by a wide range of microorganisms including bac-teria, fungi and protists despite flushing process and disinfect-ing treatment performed in the workdisinfect-ing DU. Higher 16S and 18S rRNA gene copy number per 100 mL of water was quantified in OWS compared to IW (P< 0.05) in the studied DU, with the ex-ception of the bacterial colonization in OWS of Unit 3 (Table2a and b). This highlights the critical influence of the stagnation period that significantly increases microbial abundance inside DUWL, as previously reported (Santiago et al.1994; Arvand and Hack2013; Costa et al.2015). Remarkably, the number of rRNA gene copy per 100 mL of water recovered in OWA remained unex-pectedly higher than in IW (P< 0.05) in all studied DUs, whereas OWA was exposed all day long to a chemically treated water flow. Surprisingly, the IW of Unit 3 showed a discrepancy in the number of 16S rRNA gene copy per 100 mL of water, signifi-cantly higher (P< 0.05) than the ones from the two other DUs. After biological analyses and new in situ plumbing investiga-tions, this bacterial contamination could be attributed to a lo-cal degradation of the water distribution pipe, confirming the interest of regular water quality control of DUWL. All molec-ular results were consistent with the plate counts in current use that reflected only the aerobic and aero-tolerant culturable fraction of the DUW microbial communities. Although two DUs

Table 2. Bacterial (a) and fungal (b) abundance in DUW estimated using 16S rRNA and 18S rRNA gene quantification by qPCR, respectively. Each DUW sampled at two different time periods was analyzed in triplicate. Each value represents the mean± standard deviation.

Bacterial 16S rRNA gene copies 100 mL−1of water

Treatment IW OWS OWA

(a) Unit 1 Calbenium 6.58 (±0.39) × 103 3.42 (±0.73) × 106a 9.16 (±2.70) × 106b Unit 2 Calbenium 4.51 (±1.35) × 102 4.32 (±0.16) × 105a 9.42 (±10.40) × 102b Unit 3 Oxygenal 6 1.93 (±0.30) × 106 7.00 (±0.77) × 105a 2.92 (±0.10) × 106b (b) Unit 1 Calbenium ND 1.34 (±0.00) × 103a 1.69 (±0.63) × 104b Unit 2 Calbenium 2.59 (±0.20) × 102 5.05 (±0.53) × 103a 9.50 (±6.83) × 102b Unit 3 Oxygenal 6 4.59 (±2.71) × 102 1.37 (±0.32) × 104a 2.47 (±0.18) × 104b ND: not detected.

aValue significantly different from the corresponding IW values (Mann-Whitney test, P< 0.05). bValue significantly different from the corresponding OWS values (Mann-Whitney test, P< 0.05).

(5)

Table 3. Total aerobic and aero-tolerant cultivable microbial biomass in DUW using plate count on R2A medium. Each DUW sampled at two different time periods was analyzed in duplicate, as recommended by the manufacturer of the WASP 2 spiral plater (Biomerieux). Each value represents the mean± standard deviation.

CFU mL−1of water

Treatment IW OWS OWA

Unit 1 Calbenium ≤500 7.10 (±1.60) × 103 4.90 (±0.54) × 103

Unit 2 Calbenium ≤500 7.13 (±1.16) × 103 4.33 (±0,15) × 103

Unit 3 Oxygenal 6 7.60 (±0.30) × 103 2.60 (±0.40) × 104 >5.00 × 104a

aUpper limit of quantification.

displayed less than 500 CFU per mL in IW (Table3), none of the DUWL reached the microbial water quality level recommended by the ADA for OW (American Dental Association2012), even if all DUWL were submitted to disinfection.

Altogether, qPCR and CFU analyses combined approaches showed that the microbial colonization of DUWL changed ac-cordingly to the water circulation within the DU and that stagna-tion was a significant contributory factor in the contaminastagna-tion of output water (O’Donnell et al.2011; Coleman et al.2014).

Remarkably, for Unit 1 and Unit 3, subjected to Calbenium and Oxygenal 6 disinfectants respectively, the number of bacte-rial and fungal rRNA gene copy per 100 mL of water were higher in OWA than in OWS (P< 0.05) whereas in Unit 2, also treated with Calbenium, a significant decrease was observed between OWS and OWA (P< 0.05). These inconsistent results suggested that a range of factors inherent to each DU may limit or interfere with the activity of the disinfectant. For example, suck back of oral fluids or deterioration of the microbial quality of the sup-plying water may be an additional source of microorganisms in DUWL (Walker and Marsh2007; O’Donnell et al.2011; Coleman et al.2014).

Until now, to our knowledge, no comprehensive evaluation of the efficacy of Calbenium in DUWL under routine dental prac-tices was done. However, Walker et al. (2003) showed that an overnight Oxygenal 6 treatment on a laboratory DUW 14-day-old biofilm model resulted in a loss of 100% of plate count mi-croorganisms and in a reduction of 99.2% of biofilm coverage of dental tubing. In general dental practices, Schel et al. (2006) have demonstrated the efficacy of Oxygenal 6 application in a continuous mode over six other disinfectants to maintain a wa-ter quality level consistent with ADA guidelines in 91% of the treated DUWL studied. However, no information about the char-acteristics of the studied DUWL was provided, such as the DU origin and age that may confound comparisons of results. Con-sequently, the present study reinforces current interests in

eval-uating the efficacy of treatments under routine dental practices after long-term application of disinfectant, such as Calbenium or Oxygenal 6 since they are commonly used by European den-tists.

Bacterial and fungal genus richness, diversity and evenness indices in DUWL: influence of disinfectant

Pyrosequencing was used to characterize both bacterial and fungal communities according to the water route inside DUWL treated with a disinfectant. A total sum of 37 010 (with a mean length of 385 bases) and 28 619 reads (with a mean length of 316 bp) for bacteria and fungi, respectively, were obtained from the nine samples (IW, OWS and OWA for each studied DU) af-ter 454 pyrosequencing quality filaf-tering. Afaf-ter bioinformatic fil-ters and a homogenization step, a total of 252 and 81 different representative bacterial and fungi genera, respectively, were as-signed, indicating a high microbial diversity in DUWL. The com-plexity of the bacterial and fungal communities in the three water groups from the three DUWL was investigated based on richness (number of genus-level OTUs), Shannon and Evenness indices (Table4a and b).

In both Calbenium-treated Units 1 and 2, significant de-creases of the number of bacterial genera and Shannon index were observed in the OWA samples compared to the OWS ones suggest efficacy of Calbenium to minimize the bacterial diversity in DUWL.

While a constant decrease of Shannon and Evenness indices was observed in Unit 1 from IW to OWA, Evenness index re-mained stable in Unit 2, suggesting a distinct influence of Calbe-nium on the bacterial community of these two DUs, in favor of a few but highly dominant bacterial species in Unit 1. Considering bacterial biomass and diversity metrics, a similarity was shown in IW of Units 1 and 2, even if Unit 1 was equipped with a water

Table 4. Bacterial (a) and fungal (b) genus richness, diversity and evenness indices in DUW. Duplicate extracted DNA from each DUWL sample was pooled according to the water group prior to the pyrosequencing.

Number of genus Shannon Evenness

Treatment IW OWS OWA IW OWS OWA IW OWS OWA

(a) Unit 1 Calbenium 82 126 89 2.79 2.43 1.46 0.63 0.50 0.32 Unit 2 Calbenium 69 139 89 2.65 3.48 2.96 0.63 0.70 0.66 Unit 3 Oxygenal 6 225 94 200 3.87 2.42 3.56 0.71 0.53 0.67 (b) Unit 1 Calbenium 21 30 58 2.14 2.43 2.20 0.70 0.71 0.54 Unit 2 Calbenium 51 9 14 2.92 1.20 0.91 0.74 0.55 0.34 Unit 3 Oxygenal 6 35 14 27 2.19 0.68 1.17 0.61 0.26 0.35

(6)

softener unlike Unit 2 (Table1). This observation is in agreement with the results of Hambsch, Sacr ´e and Wagner (2004) suggest-ing that a water softener would not be a constant source of mi-crobial contamination of DUWL.

Remarkably, for the Oxygenal-treated Unit 3, the observed bacterial richness in the IW was the highest observed among the three studied DUs. A 58.2% decrease of the bacterial gen-era was observed in OWS compared to IW. Nevertheless, genus richness, Shannon and Evenness indices observed in the OWA reverted to values close to the ones observed in the IW. This re-sult suggested that Oxygenal 6 used as a shock cycle for 45 min before the weekend could be effective to control the negative in-fluence of water stagnation on the bacterial diversity in DUWL. It also highlighted the impact of contaminated supplying water on the OWA water microbial quality and, in this context, the subse-quent limited efficacy of Oxygenal 6 to reduce the bacterial load in OW.

The fungal community analyses showed different results in both Calbenium-treated Units 1 and 2. A reduction of the fungal genus number was observed in OWS compared to IW only in Unit 2, suggesting a lower fungal diversity in the output water only for this Unit. In Unit 1, a 2-fold increase of the number of fungal genera was observed in OWA compared to OWS. Evenness and Shannon evolution patterns suggested a dominance of a few fungal species in both OW (mainly in OWA) in Unit 2 instead of in Unit 1.

Thus, this discrepancy observed in diversity metrics sug-gested that the efficacy of Calbenium may be affected by dif-ferent factors such as the quality of the supply water and the presence of biofilms in DUWL tubing.

Remarkably, for the Oxygenal 6-treated Unit 3, a similar evo-lution pattern was observed for both bacterial and fungal diver-sity indices, highlighting the interest of shock treatment before a long water stagnation period and the influence of the microbial quality of the water supplying this DUWL.

Through these three distinct case studies, our results con-firmed that microbial richness and diversity may be strongly linked to technical, maintenance and practice characteristics of each DU but also dependent on the microbial quality of the wa-ter supplying the DU. Inconsistent results observed from the in-dices of both bacterial and fungal community diversity clearly suggested that the long-term disinfection efficacy should be op-timized to reduce both density and diversity of the microbial communities in DUWL.

Bacterial and fungal communities’ composition in DUWL: influence of disinfectants

Proteobacteria was a prevalent bacterial phylum in each stud-ied DUWL representing over 60% of the total sequences, al-though with levels varying according to the DU and to wa-ter flow inside DUs (Supporting Information, Table SD). This phylum was commonly detected in drinking water and dis-tribution system (Shi et al. 2013; Holinger et al. 2014; Huang et al.2014); their relative abundance may be due to a higher tolerance to chlorine compared to other phyla (Huang et al.

2014). Members of Actinobacteria, Firmicutes, Bacteroidetes and Chlamydia were also found as previously observed in DUWL (Costa et al.2015). At the class level, Gammaproteobacteria, Al-phaproteobacteria, Betaproteobacteria and Actinobacteria were the most represented in the studied DUWL. Ascomycota and Ba-sidiomycota were dominant at fungal phyla level and Saccha-romycetes, Dothideomycetes, Sordariomycetes, Agaricomycetes

and Tremellomycetes at the class level, also in varying propor-tions according to the DU and location (Supporting Information, Table SD).

Bacterial and fungal genera representing at least 1% of the total abundance of one of the three studied water groups were considered in each condition (Figs 1 and 2). In Unit 1, sub-jected to continuous Calbenium treatment, Halomonas (27% of the total sequences), Shewanella (13%), Propionibacterium (12%) and Sphingomonas (12%) were dominant bacterial genera de-tected in IW (Fig.1). Stenotrophomonas was the sole dominant genus in both OW, representing 42% and 76% of the total bacte-rial sequences in OWS and OWA, respectively (Fig.1). Concern-ing the fungal contamination, Mortierella (13%), Saccharomyces (13%) and Galactomyces (8%) were dominant in IW (Fig.2). The detection of the genus Candida, which was also dominant in IW (31% of the total sequences), varied according to the wa-ter flow inside DUWL; Candida represented 12% of the total se-quences in OWS and decreased to 5% in OWA. Phoma (20%) and Flammulina (28%) were also dominant in OWS and OWA, re-spectively. Nevertheless, unknown fungi represented 16% and 37% of the total sequences and unclassified fungi, 34% and 20% of the total sequences in OWS and OWA, respectively (Fig.2).

In Unit 2, also subjected to continuous Calbenium treat-ment, Halomonas (30%), Propionibacterium (18%) and Shewanella (14%) were still the three dominant bacterial genera detected in IW, as observed in Unit 1 (Fig.1). These genera were also highly detected in OWS (Halomonas 23%, Propionibacterium 20% and Shewanella 14% of the total sequences), whereas Parvular-cula (14%), Barnesiella (14%), Sphingomonas (10%) and Sphingobium (9%) were dominant in OWS. Regarding the fungal contamina-tion, Mortierella (26%) and Malassezia (6%) genera were dominant in IW, Mrakia (50%) and Sporobolomyces (18%) in OWS, and Pseu-dallescheria (10%) and Cladosporium (6%) in OWA (Fig.2). Contrary to previous observation concerning Unit 1, Candida genus repre-sented 14% of the total sequences in IW of Unit 2 and its preva-lence increased according to the water flow inside DUWL, repre-senting 27% of the total sequence in OWS and reaching 77% in OWA.

Regarding Unit 3 subjected to both continuous and intermit-tent Oxygenal 6 treatments, the IW was dominated by Pseu-domonas (12% of the total sequences), Chlorobacterium (11%), Sph-ingomonas (7%) and Sphingobium (7%) genera (Fig.1). This ob-servation evidenced that, in this DU, the bacterial community composition of IW differed from that detected in the two other studied DUs. While 36% of the total sequences detected in OWS were affiliated to Legionella, this genus was not dominant in OWA. Both OW were dominated by Sphingomonas, represent-ing 22% and 17% of the total sequences in OWS and OWA, respectively, and Sphingobium genera (14% and 17%). Regard-ing the fungal contamination, the majority of the sequences were affiliated to unknown fungi (47%), Mrakia (22%) and Can-dida (8%) in IW; to environmental fungi (88%) in OWS; and to Mrakia (54%), environmental fungi (11%) and Candida (3%) in OWA (Fig.2).

For the first time, a next-generation DNA sequencing was used to concurrently investigate both bacterial and fungal com-munities in DUWL. Results showed that large and diverse bacterial and fungal communities were associated from the water supply to the outer exposure point of the three DUs. Despite flushing and disinfectant applications, more than 80 distinct fungal genera could colonize the DUWL, mainly be-longing to Ascomycota and Basidiomycota phyla. Results showed that the Saccharomycetes class was dominant in the core fungal

(7)

Figure 1. Heatmap representing relative abundance of bacterial communities in IW, OWS and OWA collected in the three studied DUs. Only genera representing at least 1% of relative abundance from the total sequences list in at least one condition (IW, OWS or OWA) for each DU were represented.

(8)

Figure 2. Heatmap representing relative abundance of fungal communities in IW, OWS and OWA collected in the three studied DUs. Only genera representing at least 1% of relative abundance from the total sequences list in at least one condition (IW, OWS or OWA) for each DU were represented.

(9)

microbiome of DUWL The Candida genera occurred in the three studied DUWL, whereas, until recently, this genus has been recovered only occasionally from DUWL (Walker et al. 2004; Szyma ´nska 2005). The infectious risk associated with Candida al-bicans is well documented, whereas other members of the Can-dida species are also involved in increasing number of infections (Arendrup2013). In contrast, Aspergillus and Penicillium, previ-ously described as prevalent in drinking water and DUWL, were not detected in our DUWL (Walker et al.2004; Szyma ´nska 2005; Sammon et al.2010; Mesquita-Rocha et al.2013; Oliveira et al.

2013). Occurrence of these two genera like Cladosporium may result from environmental contamination, as they are ubiqui-tously present on walls and ceilings, for example (Gniadek and Macura2007).

Our results highlighted the critical influence of the water stagnation on the microbial contamination and probably the as-sociated influence of the biofilm developed inside DUWL. In-deed, especially in the OWS condition, variable and partially specific bacterial and fungal communities were observed in each sampled DU whereas the bacterial and fungal commu-nities present in the IW condition were closer (in particular for Units 1 and 2); this suggest the development of a specific biofilm inside each DU and, consequently the singularity of each studied case. Moreover, results highlighted that OWA con-dition was influenced by the microbial community from both IW and OWS and consequently, by the own biofilm of each DU. In addition, the occurrence of potentially pathogenic bac-terial and fungal genera such as Stenotrophomonas, Propionibac-terium, Legionella, Halomonas, Pseudomonas, Shewanella, Candida, Cladosporium and Fusarium in DUWL even subjected to a dis-infectant reinforced the interest to control the microbiologi-cal water quality in dental output water. The studied continu-ous disinfecting protocols seemed partially effective to reduce the microbial density and diversity in DUWL after long-term application; however, their activity was insufficient to reach the microbial quality CFU threshold defined by the ADA in output waters (Puttaiah et al.2012). Results also suggested dis-tinct and variable disinfectant activity on the bacterial and fungal communities inhabiting DUWL. Interestingly, a shock treatment before a long water stagnation period may limit the microbial colonization of DUWL. Also, a lack of efficacy of dis-infecting treatments in DUWL may be explained by a short ex-posure time and frequency, an inadequate disinfectant and/or concentration, a malfunction of devices or a lack of compliance to disinfectant protocols (Montebugnoli et al.2005; O’Donnell et al.2007; Coleman et al.2014). The exopolysaccharide poly-mers excreted from microorganisms attached to the internal tubing and growing as a biofilm may also form a protective ma-trix to chemical water treatments as well as the stratification of the biofilm may reduce the exposure of all the microorgan-isms to disinfectants (Costerton, Stewart and Greenberg1999; O’Donnell et al.2011; Ramage et al.2012). Results from O’Donnell et al. (2007) showed that the lesser efficacy of some disinfectants based on hydrogen peroxide may result from the selection of disinfectant-tolerant species and especially, from the produc-tion by microorganisms of enzymes such as catalase that de-grade the active agent. Interestingly, in Unit 3 treated with hy-drogen peroxide (active agent of Oxygenal 6), the prevalence of two catalase positive bacterial genera (e.g. Sphingomonas and Sphingobium) was higher in OW than in IW suggesting a selec-tion of these species after exposure to disinfectant. Altogether, our results may suggest that different disinfection strategies are needed to improve the microbial quality of the DUWL output water.

CONCLUSION

The present study indicated that each DUWL develop a hetero-geneous and complex ecological system with bacterial and fun-gal assemblages in populations, influenced by the conjunction of a range of factors. Through these three distinct case studies, results clearly suggested that patients and dental staff remained exposed to a potential infectious risk relevant even though DUWL were submitted to disinfectant associated to flushing pro-cesses. Also, our study demonstrated the necessity and impor-tance of considering concurrent approaches targeting density but also diversity of both bacterial and fungal communities for the evaluation of DUWL disinfectant efficacy. Despite the lim-ited number of studied DUs, our results clearly claim and sup-port local management efforts, the definition of European stan-dards for DUWL output water, regular microbial water quality investigations of DUWL during routine general dental practice and the compliance with existing preventive recommendations in offices.

SUPPLEMENTARY DATA

Supplementary data are available at FEMSPD online.

ACKNOWLEDGEMENTS

Authors gratefully acknowledge the practitioners from hos-pitals and public dental clinics of the region of Poitiers for their voluntary participation to our studies on the evalua-tion of the dental water quality. Authors are sincerely thank-ful for the scientific supports provided by S. Terrat and colleagues from the GenoSol platform (INRA, Dijon, France,

www2.dijon.inra.fr/plateforme genosol/) for the development of the 454 pyrosequencing. They also thank D. Debail for revision of the original English-language document.

FUNDING

This work was partly supported by the following 2015–2020 pro-grams: the State-Region Planning Contracts (CPER) and the Eu-ropean Regional Development Fund (FEDER).

Conflict of interest. None declared.

REFERENCES

American Dental Association (ADA). 2012. Statement on Dental Unit Waterlines. http://www.ada.org/en/member-center/ oral-health-topics/dental-unit-waterlines (20 December 2015, date last accessed).

Arendrup MC. Candida and candidaemia. Susceptibility and epi-demiology. Dan Med J 2013;60:B4698.

Arvand M, Hack A. Microbial contamination of dental unit water-lines in dental practices in Hesse, Germany: a cross-sectional study. Eur J Microbiol Immunol 2013;3:49–52.

Baker GC, Smith JJ, Cowan DA. Review and re-analysis of domain-specific 16S primers. J Microbiol Methods 2003;55: 541–55.

Barbot V, Costa D, Deborde M et al. Efficacy of dental unit disin-fectants against Candida spp. and Hartmannella vermiformis. Pathog Dis 2014;70:203–448.

Chemidlin Pr ´evost-Bour ´e N, Christen R, Dequiedt S et al. Valida-tion and applicaValida-tion of a PCR primer set to quantify fungal communities in the soil environment by real-time quantita-tive PCR. PLoS One 2011;6:e24166.

(10)

Cole JR, Wang Q, Cardenas E et al. The Ribosomal Database Project: improved alignments and new tools for rRNA analy-sis. Nucleic Acids Res 2009;37:D141–5.

Coleman DC, O’Donnell MJ, Miller AS et al. Minimising bial contamination in dental unit water systems and micro-bial control in dental hospitals. In: Walker JH (ed.). Decontam-ination in Hospitals and Healthcare. UK: Woodhead Publishing Limited University of Leeds, 2014, 166–207.

Coleman DC, O’Donnell MJ, Shore AC et al. Biofilm problems in dental unit water systems and its practical control. J Appl Mi-crobiol 2009;106:1424–37.

Costa D, Mercier A, Gravouil K et al. Pyrosequencing analysis of bacterial diversity in dental unit waterlines. Water Res 2015;81:223–31.

Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common cause of persistent infections. Science 1999;284: 1318–22.

Dallolio L, Scuderi A, Rini MS et al. Effect of different disinfection protocols on microbial and biofilm contamination of dental unit waterlines in community dental practices. Int J Environ Res Public Health 2014;11:2064–76.

Delafont V, Mougari F, Cambau E et al. First evidence of amoebae-mycobacteria association in drinking water network. Environ Sci Technol 2014;48:11872–82.

Dey R, Bodennec J, Mameri MO et al. Free-living freshwater amoebae differ in their susceptibility to the pathogenic bacterium Legionella pneumophila. FEMS Microbiol Lett 2009;290:10–7.

Direction G ´en ´erale de la Sant ´e. Guide de pr ´evention des infec-tions li ´ees aux soins r ´ealis ´es en chirurgie dentaire et stom-atologie. In: Minist`ere de la sant´e et des solidarit´es, Direction g´en´erale de la sant´e, Conseil sup ´erieur d’hygi `ene publique de France, Deuxi `eme ´edition 2006.

Gniadek A, Macura AB. Intensive care unit environment contam-ination with fungi. Adv Med Sci 2007;52:283–7.

Greub G, Raoult D. Microorganisms resistant to free-living amoe-bae. Clin Microbiol Rev 2004;17:413–33.

Hambsch B, Sacr ´e C, Wagner I. Heterotrophic plate count and consumer’s health under special consideration of water soft-eners. Int J Food Microbiol 2004;92:365–73.

Holinger EP, Ross KA, Robertson CE et al. Molecular analysis of point-of-use municipal drinking water microbiology. Water Res 2014;49:225–35.

Huang K, Zhang XX, Shi P et al. A comprehensive insight into bacterial virulence in drinking water using 454 pyrosequenc-ing and Illumina high-throughput sequencpyrosequenc-ing. Ecotox Environ Safe 2014;109:15–21.

Kumar S, Atray D, Paiwal D et al. Dental unit waterlines: source of contamination and cross-infection. J Hosp Infect 2010; 74:99–111.

Le Calvez T, MC Trouilh ´e, Humeau P et al. Detection of free-living amoebae by using multiplex quantitative PCR. Mol Cell Probes 2012;26:116–20.

Martin MV. The significance of the bacterial contamina-tion of dental unit water systems. Brit Dent J 1987;163: 152–4.

Mesquita-Rocha S, Godoy-Martinez PC, Gonc¸alves SS et al. The water supply system as a potential source of fungal infection in paediatric haematopoietic stem cell units. BMC Infect Dis 2013;13:289.

Montebugnoli L, Dolci G, Spratt DA et al. Failure of anti-retraction valves and the procedure for between patient flushing: a ra-tionale for chemical control of dental unit waterline contam-ination. Am J Dent 2005;18:270–4.

O’Donnell MJ, Boyle MA, Russell RJ et al. Management of dental unit waterline biofilms in the 21st century. Future Microbiol 2011;6:1209–26.

O’Donnell MJ, Shore AC, Russell RJ et al. Optimisation of the long-term efficacy of dental chair waterline disinfection by the identification and rectification of factors associated with wa-terline disinfection failure. J Dent 2007;35:438–51.

Oliveira BR, Crespo MT, San Rom ˜ao MV et al. New insights con-cerning the occurrence of fungi in water sources and their potential pathogenicity. Water Res 2013;47:6338–47.

Pankhurst CL. Risk assessment of dental unit waterline contam-ination. Prim Dent Care 2003;10:5–10.

Petti S, Moroni C, Messano GA et al. Detection of oral strepto-cocci in dental unit water lines after therapy with air turbine handpiece: biological fluid retraction more frequent than ex-pected. Future Microbiol 2013;8:413–21.

Petti S, Tarsitani G. Detection and quantification of dental unit water line contamination by oral streptococci. Infect Cont Hosp Ep 2006;27:504–9.

Porteous NB, Bizra E, Cothron A et al. A survey of infection control teaching in U.S. dental schools. J Dent Educ 2014;78:187–94. Puttaiah R, Svoboda KK, Lin SM et al. Evaluation of an automated

dental unit water system’s contamination control protocol. J Contemp Dent Pract 2012;13:1–10.

Quast C, Pruesse E, Yilmaz P et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 2013;41:D590–6.

Ramage G, Rajendran R, Sherry L et al. Fungal biofilm resistance. Int J Microbiol 2012;2012:528521.

Reasoner DJ, Geldreich EE. A new medium for the enumeration and subculture of bacteria from potable water. Appl Environ Microb 1985;49:1–7.

Ricci ML, Fontana S, Pinci F et al. Pneumonia associated with a dental unit waterline. Lancet 2012;379:684.

Sammon NB, Harrower KM, Fabbro LD et al. Incidence and dis-tribution of microfungi in a treated municipal water supply system in sub-tropical Australia. Int J Environ Res Public Health 2010;7:1597–611.

Santiago JI, Huntington MK, Johnston AM et al. Microbial con-tamination of dental unit waterlines: short- and long-term effects of flushing. Gen Dent 1994;42:528–35.

Schel AJ, Marsh PD, Bradshaw DJ et al. Comparison of the effi-cacies of disinfectants to control microbial contamination in dental unit water systems in general dental practices across the European Union. Appl Environ Microb 2006;72:1380–7. Shi P, Jia S, Zhang XX et al. Metagenomic insights into

chlori-nation effects on microbial antibiotic resistance in drinking water. Water Res 2013;47:111–20.

Smith G, Smith A. Microbial contamination of used dental hand-pieces. Am J Infect Control 2014;42:1019–21.

Szyma ´nska J. Evaluation of mycological contamination of dental unit waterlines. Ann Agr Environ Med 2005a;12:153–5. Szyma ´nska J. Microbiological risk factors in dentistry.

Cur-rent status of knowledge. Ann Agr Environ Med 2005b;12: 157–63.

Szyma ´nska J. Bacterial decontamination of DUWL biofilm using Oxygenal 6. Ann Agr Environ Med 2006;13:163–7.

Szyma ´nska J, Sitkowska J, Dutkiewicz J. Microbial contamina-tion of dental unit waterlines. Ann Agr Environ Med 2008;15: 173–9.

Terrat S, Plassart P, Bourgeois E et al. Meta-barcoded evaluation of the ISO standard 11063 DNA extraction procedure to char-acterize soil bacterial and fungal community diversity and composition. Microb Biotechnol 2015;8:131–42.

(11)

Walker JT, Bradshaw DJ, Bennett AM et al. Microbial biofilm formation and contamination of dental-unit water sys-tems in general dental practice. Appl Environ Microb 2000;66: 3363–7.

Walker JT, Bradshaw DJ, Finney M et al. Microbiological evalua-tion of dental unit water systems in general dental practice in Europe. Eur J Oral Sci 2004;112:412–8.

Walker JT, Bradshaw DJ, Fulford MR et al. Microbiological evalu-ation of a range of disinfectant products to control mixed-species biofilm contamination in a laboratory model of

a dental unit water system. Appl Environ Microb 2003;69: 3327–32.

Walker JT, Marsh PD. Microbial biofilm formation in DUWS and their control using disinfectants. J Dent 2007;35:721–30. Warnes GR, Bolker B, Bonebakker L et al. Package ‘gplots’. 2015.

file:///M:/R/donesz-utilisation R heatmap.pdf(18 December 2015, date last accessed).

Wirthlin MR, Marshall GW, Rowland RW. Formation and decon-tamination of biofilms in dental unit waterlines. J Periodontol 2003;74:1595–609.

Figure

Table 1. Description of the DUs.
Table 2. Bacterial (a) and fungal (b) abundance in DUW estimated using 16S rRNA and 18S rRNA gene quantification by qPCR, respectively
Table 3. Total aerobic and aero-tolerant cultivable microbial biomass in DUW using plate count on R2A medium
Figure 1. Heatmap representing relative abundance of bacterial communities in IW, OWS and OWA collected in the three studied DUs
+2

Références

Documents relatifs

Bottle reservoirs allow the water to be cut off from the mains water, enable the choice of which type of water is used for the care (filtered mains water, distilled water, or

Abundance of the 10 main bacterial genus whose frequencies were significantly increased at T5 in the rhizosphere of inoculated plants compared to healty plants (p &lt; 0.05).

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

Because the fungi responded more strongly to soil nutri- ents, the In-Out fungal beta-diversity was higher on siliceous bedrock and at high elevations (Figure 5B; Table A2 in

A main finding of this study is that, although the landscape scale beta diversity patterns of the three microbial domains investigated are all related to plant community

40 جمابرلا فلتخبم رئازلجا ةيصخش نع عافدلا ىلع لمعتس تايعجم و بازحأ زورب لىإ يدؤيس يذلا يسايسلا لمجا و ةدعصلأا فلتمخ ىلع و تاحورطلأا و

Supplementary Data 7 | Differential abundance analysis (DESeq2) of bacterial ASVs applied to different factors, including plot, cultivar, fruit parts, harvest date, tree

According to our findings, the average number of counted bacterial strains in females with different physiological digestive stages has increased progressively from unfed to