Implication Des Microorganismes Dans Les Biotransformations Et Processus De Transfert Des Métaux Et Métalloïdes Dans Les Environnements Contaminés Par Les Mines

UNIVERSITE MONTPELLIER

Habilitation à Diriger des Recherches

Spécialité : Sciences de la Terre et de l’Environnement

Implication des microorganismes dans les biotransformations et processus de transfert des métaux et métalloïdes dans les environnements contaminés par les mines

par

Odile BRUNEEL

Chargée de Recherches à l’IRD

HydroSciences Montpellier UMR 5569

Soutenue le 29 mars 2016 devant le jury composé de

Bernard OLLIVIER Directeur de Recherche IRD, UMR 7294, Marseille Rapporteur Philippe NORMAND Directeur de Recherche CNRS UMR 5557, Lyon Rapporteur Pascale BAUDA Professeur Université de Lorraine, UMR 7360, Metz Rapporteure Pascal SIMONET Directeur de Recherche CNRS, UMR 5005, Lyon Examinateur Michel LEBRUN Professeur Université de Montpellier, UMR LSTM Examinateur

Ecole Doctorale : Systèmes Intégrés en Biologie, Agronomie, Géosciences, HydroSciences, Environnement SOMMAIRE

I CURRICULUM VITAE...... 3 Diplômes et formation...... 3 Parcours professionnel...... 3 Responsabilités récentes, animations scientifiques, comités...... 4 Collaborations récentes...... 4 Evaluation de la recherche...... 5 II CONTRATS DE RECHERCHE ET FINANCEMENTS ...... 6 III ENCADREMENT D’ETUDIANTS ET ENSEIGNEMENT…...... 8 Encadrement d’étudiants...... 8 Activité d’enseignement...... 9 IV PUBLICATIONS ET COMMUNICATIONS...... 10 Synthèse de la production scientifique...... 10 Publications...... 11 Communications, conférences et poster...... 14 V ACTIVITE DE RECHERCHE…...... 15 Préambule ...... 15 Travaux antérieurs ...... 19 Travaux actuels...... 47 VI PROJET DE RECHERCHE ...... 55 VII REFERENCES BIBLIOGRAPHIQUES……….…...... 60 VIII ANNEXES : SELECTION DE 5 PUBLICATIONS……….……...... 67

I CURRICULUM VITAE

Odile BRUNEEL

IRD-CR1 Née le 1er avril 1973 [email protected] Mariée, un enfant

Affectation actuelle : En expatriation depuis février 2012 au Laboratoire de Microbiologie et Biologie Moléculaire Université Mohammed V, Faculté des Sciences, Av Ibn Batouta BP1014 Rabat, Maroc

Laboratoire HydroSciences Montpellier, UMR5569 (CNRS/IRD/UM) Université de Montpellier, CC0057 (MSE), 163 rue Auguste Broussonet 34090 Montpellier, France

Domaine de recherche : Implication des microorganismes dans les biotransformations et processus de transfert des métaux et métalloïdes dans les environnements contaminés par les mines

DIPLOMES ET FORMATIONS • 2004 : Doctorat en Sciences de l’eau dans l’Environnement Continental, Ecole Doctorale Sciences de la Terre et de l’Eau. Laboratoire HydroSciences Montpellier. Université Montpellier II • 2001 : DESS Diagnostic, Prévention et Traitements en Environnement, Faculté Libre des Sciences de Lille, Mention Bien • 1997 : DEA de Biologie, option Biologie des Protistes de Clermont-Ferrand I et II

PARCOURS PROFESSIONNEL Recherche • Février 2012 - aujourd’hui : en affectation au sein du Laboratoire de Microbiologie et Biologie Moléculaire, Faculté des Sciences, Université Mohammed V, Rabat, Maroc • Depuis Octobre 2008: Chargée de Recherches 1ère classe à l’IRD • Octobre 2004 - aujourd’hui : Chargée de Recherches à l’IRD au sein du Laboratoire HydroSciences Montpellier (UMR 5569, CNRS-Université Montpellier-IRD) • 2001-2004 : Recherche en géomicrobiologie à l’Université Montpellier II dans le cadre de ma thèse. Laboratoire HydroSciences Montpellier, UMR 5569

Activités salariées • 1999-2000 : Professeur des écoles en CE2 à Djibouti (Afrique de l’Est) dans le cadre d’une coopération civile d’aide au développement

RESPONSABILITES RECENTES, ANIMATIONS SCIENTIFIQUES, COMITES • Représentante par Intérim de l’IRD au Maroc (août 2014-aujourd’hui) • Membre du comité de pilotage du réseau SICMED Mistrals « Activités minières dans le bassin méditerranéen – Interactions contaminants métalliques / écosystèmes − Interfaces avec la santé, l’environnement » • Membre élue depuis 2012 de la commission scientifique sectorielle n°1 (CSS1, Sciences physiques et chimiques de l’environnement planétaire) de l’IRD

COLLABORATIONS RECENTES Instituto de Biologıa Molecular y Biotecnologıa (Volga Iñiguez), Facultad de Ciencias Puras, Universidad Mayor de San Andres, C. 27 Campus Universitario Cota Cota, La Paz, Bolivie (laboratoire soutenus par le DSF de l'IRD dans le cadre du programme "jeunes équipes") Laboratoire de Microbiologie et Biologie Moléculaire (LMBM, L. Sbabou, J. Aurag et A. Filali-Maltouf), Faculté des Sciences, Université Mohamed V, Rabat, Maroc Laboratoire de Physiologie et Biotechnologie Végétale (LPBV, A. Smouni, M. Fahr), Faculté des Sciences, Université Mohamed V, Rabat, Maroc Equipe de recherche E2G, (R. Hakkou) Département des Sciences de la terre, Faculté des Sciences et Techniques de Guéliz, Université de Cadi Ayyad, Avenue Abdelkarim Elkhattabi, Gueliz, P.O. Box 549, Marrakech, Maroc Laboratoire Géoexplorations et Géotechniques (A. Ddekayir), Département de Géologie, Faculté des Sciences, BP. 11201, Zitoune, Meknès, Maroc Institut de Minéralogie et de Physique des milieux Condensés (IMPMC, G. Morin), UMR CNRS 7590, UPMC, 4 Place Jussieu, 75252 Paris, France Laboratoire AMPERE (E. Navarro), UMR CNRS 5005, Ecole Centrale de Lyon, Université de Lyon, 36 avenue Guy de Collongue, 69134 Ecully, France et Laboratoire des Symbioses Tropicales et Méditerranéennes, LSTM, UMR 113, TA A-82/J Campus de Baillarguet, 34398 Montpellier, France Laboratoire Biochimie et Physiologie Moléculaire des Plantes (BPMP, Patrick Doumas), 2, place Pierre Viala, 34060 Montpellier, France Equipe Environnement et Microbiologie (EEM, R. Duran, B. Lauga), UMR 5254 IPREM- EEM, Pau, France Laboratoire de Génétique Moléculaire, Génomique et Microbiologie (GMGM, P. Bertin, F. Ploetze), UMR 7156, Univ Louis Pasteur–CNRS, Strasbourg, France

EVALUATION DE LA RECHERCHE Participaption à un jury de thèse L. Giloteaux en décembre 2009

Participation à différents jurys de Licence, M1 et M2 tous les ans depuis 2005

Evaluations pour les journaux: FEMS Microbiology Ecology, Microbial Ecology, Extremophiles, Environmental Science and Pollution Research, Geomicrobioloy Journal, Journal of Applied Microbiology, PLOS ONE

Evaluations de projets : ANR (Blanc, JC), Ec2co (Microbiologie environnementale), FRB (Fondation pour la recherche sur la biodiversité

II CONTRATS DE RECHERCHE ET FINANCEMENT Contrats de recherches nationaux et internationaux • 2003-2005. Projet labélisé RITEAU (Ministère de l’Industrie, 677 k€). As5 : Mise au point d’un procédé biologique de potabilisation des eaux arséniées. Partenaires : IRH Environnement, BEFS-PEC, LMCP UMR 7590 (G. Morin). • 2004-2006. Projet ECODYN (AC, FNS, ECCO, 30 k€). « Processus de transfert et écotoxicité de l’arsenic et des métaux associés dans un hydrosystème en aval d’un drainage minier. Contrôles physico-chimiques et microbiologiques ». Partenaires : UMR 7590-CNRS- Universités Paris 6 et 7-IPG (G. Morin), LCABIE, UMR 5034, CNRS Université de Pau (O. Donard), LEM, Université de Pau (R. Durand), CB UPR 9043, Marseille (V. Bonnefoy), INERIS (J-M. Porcher), BRGM (M. Motelica), ECOLAG, UMR 5119 CNRS (C. Aliaume) • 2004-2006. Projet PICS CNRS (21 k€), Université de Huelva, Espagne). « Signature de l’activité bactérienne dans les précipités riches en fer des drainages miniers acides ». Partenaires: Departamento de Geologia, Universidad de Huelva, Espagne (JM. Nieto) • 2006-2007. PAI Protea (Ministères des Affaires Etrangères et de l’Education Nationale, de l’Enseignement Supérieur et de la Recherche, 10 k€) avec l’Afrique du Sud. « Metal and metalloid biotransformations in South African acid mine drainage systems”. Partenaires : Department of Biotechnology (D. Cowan), Department of Chemistry (L. Petrik) University of Western Cape, Capetown, Afrique du Sud • 2006-2008. Projet EC2CO-3BIO (INSU, CNRS, 40 k€) « Biologie, biominéraux et biotransformations dans les eaux acides minières ». Partenaires : IMPMC, UMR CNRS7590, Paris (G. Morin), IPREM, UMR 5254, CNRS- Université de Pau (R. Duran) • 2007-2008. Coordinatrice pour HSM du P2R Safe-Water (Afrique du Sud, 15 k€). « Study of the metal and metalloid biotransformations in South African acid mine drainage”. Partenaires : Department of Biotechnology (D. Cowan), Department of Chemistry (L. Petrik) University of Western Cape, Capetown, Afrique du Sud • 2007-2009. Coordinatrice du projet EC2CO-MicroBien (INSU CNRS, 90 k€) « Impact des microorganismes sur les transformations des métaux et métalloïdes dans des drainages miniers riches en sélénium ». Partenaires : IMPMC, UMR CNRS 7590, Paris (G. Morin), IPREM, UMR 5254, Pau (R. Duran) • 2007-2010. Projet ANR RARE, programme blanc (Agence Nationale pour la Recherche, 166 k€) « Reactivity of an arsenic-rich ecosystem: an integrated genomics approach ». Partenaires : GMGM, UMR 7156 de Strasbourg (P. Bertin) ; IMPMC, UMR 7590, Paris (G. Morin) ; IPREM, UMR 5254, Pau (R. Duran) • Depuis 2009- aujourd’hui. OSU OREME (INSU, INEE, 12 k€/an). Tâche d’Observation 1 (TO1, environ 10 k€/ans) « Suivi des processus hydrobiogéochimiques de transfert des métaux et métalloïdes issus des activités minières sur le site de Carnoulès ». Labelisée par l’OSU OREME (http://www.oreme.univ-montp2.fr/spip.php?rubrique41). Système d’Observation Pollution et adaptabilité biologique en aval des anciens sites miniers • 2009-2011. Bourse de thèse présidente environnée de l’UM2 (Université Montpellier II, 30 k€, A. Volant). « Etude des processus microbiens et géochimiques de mobilisation et de piégeage des éléments métalliques issus des activités minières ». • 2010-2011. Projet FRB MIGR’AMD (Fondation pour la Recherche sur la Biodiversité, 40 k€). « Microbial biogeography of acid mine drainage: a study of genetic

6 diversity and species diversity from an evolutionary perspective ». Partenaires : IPREM, UMR 5254, Pau (B. Lauga). Responsable du projet pour HSM • 2011-2012. Projet Ec2co Microbien. (INSU CNRS, 38 k€) « Rôle des bactéries du genre Thiomonas dans les transformations de polluants métalliques au sein d’écosystèmes miniers ». Partenaires : GMGM, UMR 7156, Strasbourg (F. Ploetze) et LSMBO, UMR 7509, Strasbourg (C. Carapito). Responsable du projet pour HSM • 2011-2013. Projet MISTRALS (INSU CNRS). Mediterranean integrated studies at regional and local scales. Aide au montage d’un réseau sur les activités minières dans le bassin méditerranéen – Interactions contaminants métalliques / écosystèmes - interfaces avec la santé, l’environnement et la société (Coordinateur : P. Doumas, UMR BPMP Montpellier). Membre du comité de pilotage • 2012-2013. Coordinatrice du Projet Ec2co ECODYN/MICROBIEN (INSU CNRS, 60 k€). « Etude des interactions plantes-microorganismes dans un contexte de réhabilitation de sites miniers au Maroc: mécanismes adaptatifs et effets sur le devenir des polluants métalliques ». Partenaires : LMBM, Rabat (L. Sbabou, J. Aurag, A. Filali-Maltouf) ; LPBV, Rabat (A. Smouni, M. Fahr) ; AMPERE, Lyon-LSTM, Montpellier (E. Navarro) ; UMR BPMP, (P. Doumas) ; • 2012-2014. Projet de coopération CNRS-CNRST avec le Maroc (CNRS, 4 k€). « Dynamique des métaux et métalloïdes et processus biogéochimiques mis en jeu dans les lacs de carrière du district minier de Zeïda (Maroc) ». Partenaire : Laboratoire d'Ingénierie Géologique de Meknes (A. Dekayir) • 2014-2017. Projet ANR ECO-TS IngECOST-DMA (ANR, Programme Eco- technologies & Eco-Services, 850 k€). « Ingénierie écologique appliquée à la gestion intégrée de stériles et drainages miniers acides riches en arsenic ». Partenaires : BRGM, Orléans (F. Battaglia-Brunet, C. Joulian) ; IMPMC, UMR 7590, Paris (G. Morin) ; Sol Environnement, Nanterre (A. Esnault) ; IRH, Gennevilliers (G. Grapin)

III ENCADREMENT D’ETUDIANTS ET ENSEIGNEMENT ENCADREMENT D’ETUDIANTS Licence et IUT • Julie Mougeot (2008) 2ème année DUT Génie Biologique option Analyses Biologiques et Biochimiques. UM2, Montpellier • Sandrine Gavalda (2009) 2ème année DUT Génie Biologique option Analyses Biologiques et Biochimiques. UM2, Montpellier • Cédric Bocquet (2009) 1ère année BTS Analyses Biologiques et Biotechnologie. Castelnaudary • Blandine Luce (2009-2010, 19 oct au 19 fév). Licence professionnalisante Géosciences, Traitement et Prévention des Pollutions 5GTPT). UM2, Montpellier • Aurélia Aidi (2010) 2ème année DUT Génie Biologique option Analyses Biologiques et Biochimiques. UM2, Montpellier • Coencadrement L. Rubini (2011) 2ème année DUT Génie Biologique, option Analyses Biologiques et Biochimiques. UM2, Montpellier • Co-encadrement M. Dufour (2012) 2ème année DUT Génie Biologique, option Analyses Biologiques et Biochimiques. UM2, Montpellier • Co-encadrement L. Causse (2013) 2ème année DUT Génie Biologique, option Analyses Biologiques et Biochimiques. UM2, Montpellier • Encadrement Keltoum Ouassal (2013) Licence en Sciences de la Vie, module PFE. Faculté des sciences. Université Mohammed V, Rabat, Maroc

Masters • Noémie Pascault (2006) Master 2 BGAE (Biologie, Géologie, Agroressources et Environnement). Spécialité BIMP (Biodiversité et Interactions Microbiennes et Parasitaires), parcours SM (Systèmes Microbiens). UM2, Montpellier Publication : Bruneel et al., 2008. Archaeal diversity in a Fe-As rich acid mine drainage at Carnoulès (France). Extremophiles. 12, 563-571 • Amélie Bardil (2007) Master 1 BGAE TEE, R2E (Terre, Eau et Environnement/Recherches en Eau). UM2, Montpellier • Amélie Bardil (2008) Master 2 BGAE, TEE, R2E (Terre, Eau et Environnement/Recherches en Eau). UM2, Montpellier Publication : Bruneel et al., 2011. Characterization of the active bacterial community involved in natural attenuation processes in arsenic-rich creek sediments. Microbial Ecology. 61, 793-810 • Anouk Favri (2008) Master 2 Pro Gestion des Ressources en Eaux. Parcours dans le cadre de la formation permanente. UM2, Montpellier • Camila Cordier (2010) Master 2 BGAE. Spécialité BIMP (Biodiversité et Interactions Microbiennes et Parasitaires), parcours SM (Systèmes Microbiens), UM2, Montpellier • Encadrement d’une étudiante chilienne, V. Verdejo Parada (2011) Master 1. Mention BGAE, spécialité BIMP, UM2, Montpellier • Encadrement Ikram Dahmani (2013) Master 2 BIOGECO (Biodiversité, Gestion et Conservation). Faculté des sciences, Université Mohammed V, Rabat, Maroc • Coencadrement Najoua Mghazli. 2015. Master 2 Production Végétale. Faculté des sciences, Université Mohammed V, Rabat, Maroc

Thèses • Co-encadrement, pour la partie microbiologie, de la thèse de Doctorat d’Ingénieur du CNRS de Marion Egal (2005-2008). Directrice de thèse : F. Elbaz Poulichet (DR CNRS chimiste à HSM). Encadrement par C. Casiot. Encadrement personnel effectif : 5% • Co-encadrement de thèse d’Aurélie Volant (2009-2012) soutenue le 12/12/12. Ecole Doctorale SIBAGHE. Directeurs thèse : F. Elbaz Poulichet (DR CNRS chimiste, HSM) et P. Bertin (Professeur au laboratoire GMGM, Strasbourg). Encadrement personnel effectif: 90% Publications : Bruneel et al. (2011) Characterization of the active bacterial community involved in natural attenuation processes in arsenic-rich creek sediments. Microbial Ecology. 61, 793-810 Volant et al. (2012) Archaeal diversity: temporal variation in the arsenic-rich creek sediments of Carnoulès Mine, France. Extremophiles. 16, 645-657 Volant et al. (2014) Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and others drivers along an acid mine drainage. FEMS Microbiology Ecology. 90, 247-263 • Co-encadrement thèse Ikram Dahmani. Décembre 2013. Faculté des sciences. Université Mohammed V, Rabat, Maroc. Directeur de thèse : J. Aurag (LMBM, Rabat), co-encadrement avec L. Sbabou (LMBM, Rabat) et E. Navarro (AMPERE, Lyon - LSTM, Montpellier). Encadrement personnel effectif : 60%

ACTIVITES D’ENSEIGNEMENT Université de Pau et des Pays de l’Adour ; Master 2. Module écologie moléculaire bactérienne. (3 heures/an, de 2004-2006) Université Montpellier 2 ; DEA SEEC – Etude du rôle des microorganismes dans le transfert de la pollution minière (2 heures/an, 2005-2008) Université Montpellier 2 ; M2R Fenec, L3Pro GPTP. Sortie terrain sur l’ancien site minier de Carnoulès (Gard) (3 à 9 h/an 2006-2011). Université Paris VI ; DEA puis M2R Sciences de l’Univers, Ecologie, Environnement – Parcours Hydrologie, Hydrogéologie, Stage terrain et visite de l’ancien site minier de Carnoulès (4 à 8 heures/an, de 2003-2011) Université Mohamed V de Rabat, Faculté des Sciences, Masters PV & BioGéCo (Module de Microbiologie du sol) (4 à 8 heures/an, 2012-2015)

IV PUBLICATION ET COMMUNICATION SYNTHESE DE LA PRODUCTION SCIENTIFIQUE

Bibliométrie ISI (mai 2015) Results found 31 Sum of the Times Cited without self-citations 690 Average Citations per Item 26 h-index 17

Journal Nb d’articles IF 2013 The ISME Journal 1 9.267 PLoS Genetics 1 8.167 Environmental Microbiology 1 6.24 Environmental Science and Technology 3 5.481 Water Research 1 5.323 Geochimica et Cosmochimica Acta 2 4.250 Applied and Environmental Microbiology 2 3.952 FEMS Microbiol Ecol 2 3.875 Chemosphère 1 3.499 Vaccine 1 3.485 Chemical Geology 2 3.482 Microbial Ecology 1 3.118 Science of the Total Environment 1 3.163 Environmental Chemistry 1 3.035 Journal of Applied Microbiology 1 2.386 Extremophiles 2 2.174 Applied Geochemistry 2 2.021 Aquatic Geochemistry 1 1.809 Geomicrobiology Journal 1 1.804 Environmental Science: Processes and Impacts. 1

PUBLICATIONS Revues à Comité de lecture dans des revues indexées (ISI & Pubmed) 1. Bruneel O, Personné J-C, Casiot C, Leblanc M, Elbaz-Poulichet F, Mahler BJ, Le Flèche A, Grimont PAD (2003) Mediation of arsenic oxidation by Thiomonas sp. in acid mine drainage (Carnoulès, France). Journal of Applied Microbiology. 95, 492-499 2. Morin G, Juillot F, Casiot C, Bruneel O, Personné J-C, Elbaz-Poulichet F, Leblanc M, Ildefonse P, and Calas G (2003) Bacterial formation of tooeleite and mixed arsenic(III) or arsenic(V)-Iron(III) gels in the Carnoulès acid mine drainage, France. A XANES, XRD, and SEM study. Environmental Science and Technology. 37, 1705-1712 3. Casiot C, Morin G, Juillot F, Bruneel O, Personné JC, Leblanc M, Duquesne K, Bonnefoy V, Elbaz-Poulichet F (2003) Bacterial immobilization and oxidation of arsenic in acid mine drainage (Carnoulès creek, France). Water Research. 37, 2929-2936 4. Duquesne K, Lebrun S, Casiot C, Bruneel O, Personné J-C, Leblanc M, Elbaz-Poulichet F, Morin G, and Bonnefoy V (2003) Immobilization of arsenite and ferric iron by Acidithiobacillus ferrooxidans in acid mine drainage. Applied and Environmental Microbiology. 69, 6165-6173 5. Casiot C, Leblanc M, Bruneel O, Personné J-C, Koffi K, Elbaz-Poulichet F (2003) Geochemical processes controlling the formation of As-rich waters within a tailings impoundment. Aquatic Geochemistry. 9, 273-290 6. Casiot C, Bruneel O, Personné J-C, Leblanc M, Elbaz-Poulichet F (2004) Arsenic oxidation and bioaccumulation by the acidophilic protozoan, Euglena mutabilis, in acid mine drainage (Carnoulès, France). Science of the Total Environment. 320, 259-267 7. Mévelec MN, Bout D, Desolme B, Marchand H, Magne R, Bruneel O, Buzoni-Gatel D (2005) Evaluation of protective effect of DNA vaccination with genes encoding antigens GRA4 and SAG1 associated with GM-CSF plasmid, against acute, chronical and congenital toxoplasmosis in mice. Vaccine. 23, 4489-4499 8. Casiot C, Lebrun S, Morin G, Bruneel O, Personné JC, Elbaz-Poulichet F (2005) Sorption and redox processes controlling arsenic fate and transport in a stream impacted by acid mine drainage. Science of the Total Environment. 347, 122-30 9. Bruneel O, Duran R, Koffi K, Casiot C, Fourçans A, Elbaz-Poulichet F, Personné J-C (2005) Microbial diversity in a pyrite-rich tailings impoundment (Carnoulès, France). Geomicrobiology Journal. 22, 249 - 257 10. Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F, Personné JC (2006) Diversity of microorganisms in Fe-As-rich acid mine drainage waters of Carnoulès, France. Applied and Environmental Microbiology. 72, 551-556 11. Casiot C, Pedron V, Bruneel O, Duran R, Personné J-C, Grapin G, Drakides C, Elbaz- Poulichet F (2006) A new bacterial strain mediating As oxidation in the Fe-rich biofilm naturally growing in a groundwater Fe treatment pilot units. Chemosphère. 64, 492-496 12. Bruneel O, Pascault N, Egal M, Bancon-Montigny C, Goni M, Elbaz-Poulichet F, Personné J-C, Duran R (2008) Archaeal diversity in a Fe-As rich acid mine drainage at Carnoulès (France). Extremophiles. 12, 563-571 13. Egal M, Elbaz-Poulichet F, Casiot C, Motelica-Heino M, Negrel P, Bruneel O, Nieto JM, Sarmiento AM (2008) Iron isotopes in acid mine waters and iron-rich solids from the Tinto-Odiel Basin (Iberian Pyrite Belt, Southwest Spain). Chemical Geology. 253, 162–171

14. Benzerara K, Morin G, Yoon TH, Miot J, Tyliszczak T, Casiot C, Bruneel O, Farges F, and Brown Jr GE (2008) Nanoscale study of As transformations by bacteria in an acid mine drainage system. Geochimica and Cosmochimica Acta. 72, 3949-3963 15. Casiot C, Egal M, Bruneel O, Cordier M-A, Bancon-Montigny C, Gomez E, Aliaume C, Elbaz-Poulichet F (2009) Hydrological and geochemical controls on metals and arsenic in a Mediterranean river contaminated by acid mine drainage (the Amous River, France); preliminary assessment of impacts on fish (Leuciscus cephalus). Applied Geochemistry. 24, 787-799 16. Egal M, Casiot C, Morin G, Parmentier M, Bruneel O, Lebrun S, Elbaz-Poulichet F (2009) Kinetic control on the formation of tooeleite, schwertmannite and jarosite by Acidithiobacillus ferrooxidans strains in an As(III)-rich acid mine water. Chemical Geology. 265, 432-441 17. Arsène-Ploetze F, Koechler S, Marchal M, Coppee J-Y, Chandler M, Bonnefoy V, Barakat M, Barbe V, Battaglia -Brunet F, Brochier-Armanet C, Bruneel O, G. Bryan C, Cleiss J, Heinrich-Salmeron A, Hommais F, Joulian C, Krin E, Lieutaud A, Lièvremont D, Michel C, Muller D, Ortet P, Proux C, Siguier P, Roche D, Rouy Z, Salvignol G, Slyemi D, Talla E, Weiss S, Weissenbach J, Médigue C, Bertin PN (2010) Structure, function and evolution of the Thiomonas spp. genome inferred from sequencing and comparative analysis. PLoS Genetics. 6 (2) e1000859 18. Egal M, Casiot C, Morin G, Elbaz-Poulichet F, Cordier MA Bruneel O (2010) An updated insight into the natural attenuation of As concentrations in Reigous Creek (southern France). Applied Geochemistry. 25, 1949–1957 19. Bruneel O, Volant A, Gallien S, Chaumande B, Casiot C, Carapito C, Bardil A, Morin G, Brown Jr GE, Personné JC, Le Paslier D, Schaeffer C, Van Dorsselaer A, Bertin PN, Elbaz-Poulichet F, Arsène-Ploetze F (2011) Characterization of the active bacterial community involved in natural attenuation processes in arsenic-rich creek sediments. Microbial Ecology. 61, 793-810 20. Bertin PN, Heinrich-Salmeron A, Pelletier E, Goulhen- Chollet F, Arsène-Ploetze F, Gallien S, Lauga B, Casiot C, Calteau A, Vallenet D, Bonnefoy V, Bruneel O, Chane-Woon- Ming B, Cleiss-Arnold J, Duran R, Elbaz-Poulichet F, Fonknechten N, Giloteaux L, Halter D, Koechler S, Marchal M, Mornico D, Schaeffer C, Thil Smith AA, Van Dorsselaer A, Weissenbach J, Médigue C and Le Paslier D (2011) Metabolic diversity between main microorganisms inside an arsenic-rich ecosystem revealed by meta- and proteo-genomics. The ISME Journal. 5, 1735-1747. Cet article a fait l’objet d’une note dans la revue Sciences (2011, vol 332, p1128) 21. Casiot C, Egal M, Bruneel O, Verma N, Parmentier M, Elbaz-Poulichet F (2011) Predominance of aqueous Tl(I) species in the river system downstream from the abandoned Carnoulès mine (Southern France). Environmental Science & Technology. 45, 2056-2084 22. Volant A, Desoeuvre A, Casiot C, Lauga B, Delpoux S, Morin G, Personné JC, Héry M, F Elbaz-Poulichet, Bertin P and Bruneel O (2012) Archaeal diversity: temporal variation in the Arsenic-Rich Creek Sediments of Carnoulès Mine, France. Extremophiles. 16, 645-657 23. Giloteaux L, Duran R, Casiot C, Bruneel O, Elbaz-poulichet F and Goñi-urriza M (2013) A survey of sulfate reducing bacteria in a heavily arsenic contaminated acid mine drainage (Carnoulès, France). FEMS Microbiol Ecol. 83 724–737

24. Maillot F, Morin G, Juillot F, Bruneel O, Casiot C, Ona-Nguema G, Wang Y, Lebrun S, Aubry E, Vlaic G, Brown GE Jr (2013) Structure and reactivity of As(III)- and As(V)-rich schwertmannites and amorphous ferric arsenate sulfate from the Carnoulès acid mine drainage, France: comparison with biotic and abiotic model compounds and implications for As remediation. Geochimica et Cosmochimica Acta. 104, 310-329 25. Resongles E, Casiot C, Elbaz-Poulichet F, Freydier R, Bruneel O, Piot C, Delpoux S, Volant A, Desoeuvre A (2013) Fate of Sb(V) and Sb(III) species along a gradient of pH and oxygen concentration in the Carnoulès mine waters (Southern France)". Environmental Science: Processes and Impacts. 15, 1536-1544 26. Adra A, Morin G, Ona-Nguema G, Maillot F, Casiot C, Bruneel O, Lebrun S, Juillot F, Brest J (2013) Arsenic Scavenging by Al-Substituted Ferrihydrites in a Circumneutral pH River Impacted by the Acid Mine Drainage of Carnoulès, Gard, France. Environmental Science and Technology. 47, 12784-12792 27 Héry M, Casiot C, Resongles E, Gallice Z, Bruneel O, Desoeuvre O, Delpoux S (2014) Release of arsenite, arsenate and methyl-arsenic species from streambed sediment impacted by acid mine drainage : a microcosm study. Environmental Chemistry. 11, 514-524 28 Volant A, Bruneel O, Desoeuvre A, Héry M, Casiot C, Bru N, Delpoux S, Fahy A, Javerliat F, Bouchez O, Duran R, Bertin PN, Elbaz-Poulichet F and Lauga B (2014) Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and others drivers along an acid mine drainage. FEMS Microbiology Ecology. 90, 247-263

Publications soumises Doumas P, Munoz M, Banni M, Becerra S, Bruneel O, Casiot C, Cleyet-Marel J-C, Gardon J, Noak Y, Sappin-Didier V. Polymetallic pollution from abandoned mines in Mediterranean regions: a multidisciplinary approach of environmental risks. Soumis à Regional Environmental Change

Publications en cours de soummission Idir Y, Sbabou L, Bruneel O, Filali-Maltouf A and Aurag J. Characterization of root- nodule bacteria isolated from Hedysarum spinosissimum L, growing in mining sites of Northeastern region of Morocco. Sera soumis à Environmental Science and Pollution Research

Publication dans des revues non indexées Casiot C, Héry M, and Bruneel O (2012) Pollution by mine drainage: towards biological treatment? In: Water at the Heart of Science. IRD Edition, Marseille Benyassine EM, Dekayir A, Héry M, Delpoux S, Desoeuvre A, Bruneel O, Benhassou H, Rouai M, Casiot C (2013) Contrasted arsenic speciation in two alkaline pit lakes from the abandoned Pb mining area of Zeida (Moulouya, Morocco). International Journal Clean-Soil, Air, Water. Special Focus Issue on Emerging Pollutants in Euro-Mediterranean and MENA Countries

COMMUNICATIONS, CONFERENCES ET POSTER Communication orale (5 dernières années) Bruneel O, Casiot C, Personné J-C, Volant A, Vadapalli VRK, Petrik L, Cowan DA, Morin G, Duran R, and Elbaz-Poulichet F. Impact des microorganismes sur les transformations des métaux et métalloïdes dans des drainages miniers d’Afrique du Sud. Proceeding, réunion de restitution du programme Ec2co. Toulouse, France. 22-26 November 2010 Bertin PN, Heinrich-Salmeron A, Pelletier E, Goulhen- Chollet F, Arsène-Ploetze F, Gallien S, Lauga B, Casiot C, Calteau A, Vallenet D, Bonnefoy V, Bruneel O, Chane-Woon- Ming B, Cleiss-Arnold J, Duran R, Elbaz-Poulichet F, Fonknechten N, Giloteaux L, Halter D, Koechler S, Marchal M, Mornico D, Schaeffer C, Thil Smith AA, Van Dorsselaer A, Weissenbach J, Médigue C and Le Paslier D. Diversity of metabolic interactions inside an arsenic-rich microbial ecosystem revealed by meta- and proteo-genomics. BAGECO11, 11th Conference on Bacterial Genetics and Ecology. Corfu, Greece. 29 May - 2 June 2011 Morin G, Ona-Nguema G, Juillot F, Maillot F, Wang Y, Egal M, Bruneel O, Casiot C, Elbaz-Poulichet F, Calas G, Brown JR. How biogenic nano-iron oxides can control the fate of pollutants. Goldschmidt. Prague, République Tchèque. 14-19 août 2011 Casiot C, Delpoux S, Desoeuvre A, Volant A, Egal M, Resongles E, Hery M, Freydier R, Elbaz-Poulichet F, Cadot E, Gardon J, Bruneel O. Spéciation et processus de transfert de métaux et métalloïdes dans les eaux minières: exemple du site de Carnoulès dans le Gard. Premières rencontres du Réseau "Environnements Miniers Méditerranéens".Montpellier, France. 14-16 mai 2012 Bruneel O, Desoeuvre A, Volant A, Héry M, Casiot C, Delpoux S, Freydier R, Elbaz – Poulichet F. Impact des microorganismes sur le transfert des contaminants métalliques dans les environnements miniers. Premières rencontres du Réseau "Environnements Miniers Méditerranéens". Montpellier, France. 14-16 mai 2012 Casiot C, Bruneel O, Hery M, Delpoux S, Desoeuvre A, Volant A, Resongles E, Freydier R, Elbaz-Poulichet F. Speciation and transfer processes of metals /metalloids in mining water : exemple of studies at the Carnoulès mining site (Gard). 4th SPECIATION seminar; Biological, environmental and nuclear speciation. Montpellier, France. May 29-31, 2012 Bruneel O, Volant A, Dahmani I, Sbabou L, Navarro I, Héry M, Désœuvre A, Casiot C, Filali-Maltouf A. Study of diversity using next generation sequencing. 4ème Congrès de l'Association Marocaine de Microbiologie (AMM) et 16ème Congrès de l'Association Africaine pour la Fixation Biologique de l'azote (AABNF) sur le thème BIOFERSOL, Biofertilisation des sols et développement durable en Afrique. Maroc, Rabat. 03-07 novembre 2014 Dekayir A, Benyassine M, Casiot C, Hery M, Bruneel O, El Hachimi ML, Rouai M. Contamination des eaux de lacs de carrières de la mine abandonnée de Zeida (Maroc). Colloque SICMED. Tunisie, Tunis. 18-20 novembre 2014

Poster (5 dernières années) Bruneel O, Casiot C, Personné J-C, and Elbaz-Poulichet F. The Carnoulès mine (Gard, France). Generation of as-rich acid mine drainage and natural attenuation processes. 5ème

Colloque International « Contamination Métallique : Impact sur l’Environnement, la Santé et la Société ». Oruro, Bolivie. 13-15th October 2010 Elbaz-Poulichet F, Casiot C, Bruneel O, Egal M, Morin G, Miot J, Benzerara K, Duran R, Goni-Urriza, M.; Giloteaux, L. Biologie, biominéraux et biotransformations dans les eaux acides minières – 3BIO. Colloque de restitution EC2CO. Toulouse, France. 23-25 novembre 2010 Bertin P.N., Heinrich-Salmeron A., Pelletier E., Goulhen- Chollet F., Arsène-Ploetze F., Gallien S, Lauga B, Casiot C, Calteau A, Vallenet D, Bonnefoy V, Bruneel O, Chane-Woon- Ming B, Cleiss-Arnold J, Duran R, Elbaz-Poulichet F, Fonknechten N, Giloteaux L, Halter D, Koechler S, Marchal M, Mornico D, Schaeffer C, Thil Smith AA, Van Dorsselaer A, Weissenbach J, Médigue C and Le Paslier D. Diversity of metabolic interactions inside an arsenic-rich microbial ecosystem revealed by meta- and proteo-genomics. Colloque Restitution ANR “Des molécules aux écosystèmes”. Montpellier France, 13-14th September 2011 Lauga B, Volant A, Bruneel O, Fahy A, Laoudi S, Casiot C, Nieto JM and Duran R. MIGRAMD : Microbial biogeography of Acid Mine Drainage: a study of genetic diversity and species diversity from an evolutionary perspective. Colloque FRB "Les Ressources Génétiques face aux nouveaux enjeux environnementaux, économiques et sociétaux". Montpellier, France. 21-22 September 2011 Lauga B, Volant A, Bruneel O, Fahy A, Laoudi S, Casiot C, Nieto M and Duran R. Microbial biogeography of Acid Mine Drainage: a study of specific diversity and molecular diversity, in Colloque Jacques Monod "Génomique écologique intégrative". Roscoff, France. 15-19 octobre 2011 Javerliat F, Volant A, Laoudi S, Bruneel O, Fahy A, Casiot C, Iniguez V, Nieto JM, Duran R and Lauga B. Microbial biogeography of Acid Mine Drainage: a study of genetic diversity and species diversity from an evolutionary perspective, Colloque Génomique Environnementale. Lyon, France. 28-30 November 2011 Volant A, Bruneel O, Desoeuvre A, Casiot C, Bru N, Delpoux S, Héry M, Javerliat F, Fahy A, Elbaz-Poulichet F, Duran R, Bertin P and Lauga B. Spatio-temporal dynamics of bacterial community in the very As-rich creek waters of Carnoulès mine, France. In Ecole Thématique Expert Génomique Environnementale. Aussois, France. 23-27 Avril 2012 Volant A, Bruneel O, Desoeuvre A, Casiot C, Bru N, Delpoux S, Héry M, Javerliat F, Fahy A, Elbaz-Poulichet F, Duran R, Bertin P and Lauga B. Spatiotemporal dynamics of bacterial community in the very As-rich creek waters of Carnoulès mine, France. ISME. Copenhague, Danemark, 19-24 August 2012 Benyassine EM, Dekayir A, Héry M, Delpoux S, Desoeuvre A, Bruneel O, Benhassou H, Rouai M, Casiot C.Contrasted arsenic speciation in two alkaline pit lakes from the abandoned Pb mining area of Zeida (Moulouya, Morocco). International Symposium On Emerging Pollutants in Irrigation Waters: Origins, Fate, Risks, and Mitigation. Tunisia : Hammamet.25- 28 November 2013

V ACTIVITE DE RECHERCHE

PREAMBULE Lors de mon stage de DEA, effectué en 1997 au sein du CJF INSERM 93-09, Immunologie des Maladies Infectieuses (Tours), j’ai travaillé sur la vaccination génique contre la toxoplasmose en utilisant le gène SAG1. Suite à ce stage et désirant depuis longtemps faire de la coopération mais n’ayant pas trouvé d’opportunités en biologie, je suis partie 2 ans en tant que professeur des écoles en CE2, à Djibouti, dans le cadre d’une coopération civile d’aide au développement. Cette expatriation ayant été extrêmement enrichissante et afin de pouvoir trouver du travail plus facilement dans la coopération, j’ai entrepris, à mon retour en France en 2000, une formation plus appliquée, un DESS « Diagnostics, Prévention et Traitements en Environnement ». Déjà très intéressée par l’Institut de Recherche pour le Développement, j’ai effectué mon stage de fin d’étude de 6 mois en avril 2002 au laboratoire HydroSciences Montpellier, UMR 5569 (IRD, CNRS, Université de Montpellier 1 et 2). Ce stage a porté sur l’identification des microorganismes présents dans les eaux de drainage de la mine de Carnoulès (Gard) et sur l’isolement d’espèces actives sur l’arsenic et le fer. La finalité de ces travaux était, à terme, d’aider au développement de procédés passifs de bio-réhabilitation des effluents miniers et industriels en utilisant ces microorganismes. Très intéressée par ce sujet et par les potentialités d’application de ce travail, j’ai finalement poursuivi cette étude par une thèse au sein de cette même UMR. Après ma thèse, soutenue en avril 2004, j’ai été recrutée à l’IRD en octobre de la même année pour travailler sur l’implication des microorganismes dans les environnements miniers. Mon programme de recherche s’intitule : «Etude des processus microbiens et géochimiques de transfert des métaux et métalloïdes issus des activités minières».

L'écologie microbienne suscite un engouement très important car les microorganismes, bien qu’invisibles à l’œil nu, sont essentiels à la vie sur terre. Ces microorganismes catalysent en effet les transformations uniques et indispensables aux cycles biogéochimiques de la biosphère de part leurs activités métaboliques. Ils produisent les composants essentiels de la planète et représentent le plus grand réservoir de nutriments terrestre, comme le nitrogène et le phosphore et séquestrent également environ 50% du carbone total des organismes vivants (Whitman et al., 1998). Ils sont également les principaux recycleurs de matières en décomposition, rendant disponible différents types de composés sous forme organique, permettant ainsi la survie et le fonctionnement des écosystèmes (Whitman et al., 1998 ; Falkowski et al., 2008). Parce que les microorganismes sont présents dans les 3 domaines du vivant (Archaea, Bacteria et Eukarya) et qu’ils représentent les groupes les plus diversifiés d’organismes sur Terre, une connaissance de leur diversité est primordiale pour la compréhension du fonctionnement des écosystèmes et des processus planétaires (Pace, 1997 ; Behnke et al., 2011).

L'exploitation minière est vitale pour l'économie mondiale, mais l'extraction des composés métalliques génère de grandes quantités de déchets. Actuellement, le volume est estimé à plusieurs milliers de millions de tonnes par an, mais est en augmentation exponentielle en raison de la demande qui ne cesse de croître et de l'exploitation de gisements de faibles teneurs (Hudson-Edwards and Dold, 2015). En l’absence d’une gestion extrêmement rigoureuse des sites miniers, ces derniers sont une source de nuisance importante, en raison de la présence de composés très toxiques comme le plomb, l’arsenic ou le mercure Leur accumulation tout au long de la chaîne alimentaire génère des problèmes importants pour la végétation, la santé animale et humaine. Lorsque des minéraux sulfurés sont présents dans ces déchets, ils peuvent former, en présence d’eau et d’oxygène, des effluents acides, riches en métaux et métalloïdes, appelés Drainages Miniers Acides (DMA) (Langmuir, 1997). Ces drainages de mine sont considérés comme l’une des plus importantes et pernicieuses forme de pollution des eaux provenant de l’activité minière à travers le monde et représentent d’importants impacts environnementaux et sociaux économique (Hallberg, 2010) avec des coûts de traitements estimés à plusieurs milliards de dollars. Même s’il est très difficile d’estimer l’impact des DMA à travers le monde, il a été suggéré que plus de 12 000 km de cours d’eau étaient touchés par les DMA rien qu’au Royaume Uni (Hallberg, 2010). Le problème de ces DMA est leur potentiel de menace à long terme, avec une production généralement étalée sur des dizaines, voire des centaines d’années après la fermeture des mines (Younger, 1997 ; Hallberg, 2010). Bien que ces milieux soient très hostiles en raison des conditions extrêmes de vie en termes de pH et de concentration en métaux et métalloïdes toxiques, de nombreux microorganismes (Bactéries, Archaea et Eucaryotes), naturellement présents, sont capables de s’y développer (Baker and Bandfiel, 2003 ; Jonhson and Hallberg, 2003). Ces microorganismes adaptés jouent un rôle essentiel car ils sont impliqués dans les mécanismes biogéochimiques contrôlant le comportement des métaux et métalloïdes, qui sont présents dans l’environnement sous différentes formes chimiques qui n’ont ni la même toxicité, ni la même mobilité. Les réactions d’oxydoréduction ou de méthylation sont généralement très lentes et nécessitent la plupart du temps une catalyse qui est bien souvent assurée par les microorganismes. Par exemple, le rôle clé de l’activité des microorganismes (et notamment ceux qui oxydent le fer) est connu depuis longtemps dans les réactions d’oxydation de la pyrite à l’origine de l’apparition des DMA (Edwards et al., 2000a ; Sand et al., 2001 ; Vera et al., 2013). Selon certains auteurs, l’activité microbienne serait à l’origine d’environ 75% de la production des DMA (Edwards et al., 2000b ; Baker and Banfield, 2003). Ces mêmes organismes qui oxydent le fer sont également susceptibles de promouvoir, dans l’eau, la formation d’oxydes de fer qui favorisent l’immobilisation des métaux en les coprécipitant ou en les adsorbant (Johnson and Hallberg, 2003, 2005 ; Johnson, 2014). Les bactéries sulfato-réductrices sont aussi capables d’immobiliser des métaux en favorisant la précipitation directe de sulfures métalliques généralement insolubles (Johnson and Hallberg, 2005). De plus, certains processus métaboliques vont également modifier la mobilité de l’élément toxique et/ou sa toxicité. Par exemple, la forme oxydée As(V) produite par Thiomonas sp. est considérée comme 60 fois moins toxique pour les organismes supérieurs que la forme réduite As(III) à pH acide. Ces quelques exemples illustrent le rôle des microorganismes dans les processus de

précipitation, complexation, adsorption, remise en solution et la distribution des différentes formes chimiques en solution. De plus, les stériles miniers sont généralement constitués de particules très fines facilement transportées par la pluie et le vent, polluants les terres agricoles, les cours d’eau ou les puits environnants et générant un problème de santé publique majeur pour les populations alentours (Mendez and Maier, 2008). Depuis une quinzaine d’années, des travaux se sont intéressés à l’utilisation de plantes pour limiter l’impact de cette pollution ; les déchets pouvant être immobilisés par la mise en place d’un couvert végétal (phytostabilisation) ou être accumulés dans les tissus végétaux (phytoextraction, Ma et al., 2011). Alors que l'établissement d'un couvert végétal sur ces déchets minier reste un défi, les microorganismes peuvent fortement accélérer le processus de phytostabilisation en influençant la croissance des plantes grâce à différents mécanismes (fixation d’azote, solubilisation du phosphate, production de phytohormones, etc., Rajkumar et al., 2012). Ils peuvent aussi intervenir directement sur la mobilisation/immobilisation des métaux et métalloïdes dans le sol (production de sidérophores, d’enzymes, etc. ou transformation rédox de ces éléments (Ma et al., 2011 ; Rajkumar et al., 2012). Les microorganismes jouent ainsi un rôle primordial dans ces environnements. Leur connaissance présente donc un intérêt fondamental majeur pour la gestion et la remédiation des sites contaminés.

Mon programme de recherche s’inscrit dans le cadre de l’axe 1 (Biogéochimie, Contaminant, Santé) de l’UMR HydroSciences Montpellier qui aborde les questions de pollution et de toxicité pour les écosystèmes aquatiques. Cet axe s’intéresse également aux aspects de bioréhabilitation et de recyclage des eaux. L’étude des pollutions d’origine minière a commencé il y a maintenant une 20aine d’années au laboratoire. D’abord principalement centrée sur les aspects purement géochimiques puis microbiologiques, cette équipe s’intéresse maintenant également à l’impact de ces polluants métalliques sur la santé grâce au recrutement d’un médecin et d’une géographe épidémiologiste. Au sein de cette équipe, je m’intéresse à la partie microbiologie et principalement au rôle des microorganismes dans le transfert des polluants inorganiques. Ce travail inclut à la fois de la microbiologie classique par isolement mais aussi de la biologie moléculaire et maintenant de la génomique. L’équipe de microbiologie comprend actuellement une Assistante Ingénieure depuis 2010 ainsi qu’une maître de conférences recrutée en janvier 2011.

TRAVAUX ANTERIEURS J’ai commencé ma première activité de recherche en 1997 au sein du CJF INSERM 93-09, Immunologie des Maladies Infectieuses (Tours), à l’occasion de mon stage de DEA sur la vaccination génique contre la toxoplasmose en utilisant le gène SAG1. La toxoplasmose est une maladie infectieuse très répandue qui touche les animaux à sang chaud dont l’Homme en raison de la présence d’un parasite protozoaire, Toxoplasma gondii. Généralement bénigne chez l’homme et asymptomatique dans 90% des cas, ce parasite peut menacer la vie lors d’une immunodépression ou peut avoir de graves conséquences pour le fœtus lors de la contamination d’une femme pendant la grossesse (primo-infection) en raison de la transmission transplacentaire du parasite. Durant mon stage, des essais vaccinaux, utilisant de l’ADN codant une des protéines de Toxoplasma gondii, le gène SAG1 (vaccination à ADN), ont été réalisés chez la souris qui présente des formes de toxoplasmose très proches de la toxoplasmose humaine. Cette étude a montré une bonne réponse du système immunitaire de la souris avec la production d’anticorps mais un taux de survie très faible lors de l’immunisation intramusculaire1.

Depuis mon stage de DESS en avril 2001, je m’intéresse à l’implication des microorganismes dans les biotransformations et processus de transfert des métaux et métalloïdes dans les drainages miniers acides. Ce programme de recherche a pour objectif de mieux comprendre les processus biogéochimiques qui contrôlent les transferts de métaux et métalloïdes, en particulier l’arsenic, et d’étudier de manière pluridisciplinaire et intégrée le fonctionnement de ces environnements extrêmes. Il se situe en effet à l’interface de la microbiologie, de la géochimie, mais également de l’hydrogéologie et de la minéralogie. En microbiologie, pour l’étude du chantier de Carnoulès, ce programme de recherche fédère différentes compétences apportées par plusieurs équipes d’autres laboratoires comme la métagénomique ou la métaprotéomique (collaboration étroite avec l’EEM de Pau (R. Duran, B. Lauga) et le laboratoire GMGM de Strasbourg (P. Bertin, F. Ploetze) ou la minéralogie (partenariat avec G. Morin, IMPMC, IPGP de Paris). Ces recherches incluent également la caractérisation physicochimique approfondie de ces environnements extrêmes par les chimistes du laboratoire (C. Casiot, F. Elbaz-Poulichet, MA. Cordier puis S. Delpoux). Ces approches combinées permettent d’obtenir une vision globale et intégrée des processus complexes qui conditionnent les interactions entre les microorganismes et leur environnement.

1 Mévelec MN, Bout D, Desolme B, Marchand H, Magne R, Bruneel O, Buzoni-Gatel D (2005) Evaluation of protective effect of DNA vaccination with genes encoding antigens GRA4 and SAG1 associated with GM-CSF plasmid, against acute, chronical and congenital toxoplasmosis in mice. Vaccine. 23, 4489-4499 19

Introduction à l’étude des drainages miniers acides L’intérêt pour l’étude de ces écosystèmes extrêmes est multiple. Comme nous l’avons vu, ces microorganismes présentent tout d’abord un grand intérêt pour la gestion des déchets miniers et leur connaissance est primordiale pour mieux gérer leurs impacts sur l’environnement et est également critique pour pouvoir continuer l’exploitation des ressources minérales, dont la demande ne cesse de croitre à travers le monde (Hallberg, 2010). Les microorganismes qui peuplent ces écosystèmes extrêmes sont généralement constitués de communautés simplifiées par rapport aux environnements plus hospitaliers (Tyson et al., 2004 ; Denef et al., 2010). Ceci est dû notamment aux pressions de sélection qu’impose l’adaptation des microorganismes à ces environnements, ainsi que par le nombre limité de sources d’énergie disponible dans le milieu (Baker and Banfield, 2003). Ces environnements sont donc colonisés par des espèces dites spécialistes, généralement peu abondantes, ce qui en font d’excellents modèles pour étudier la dynamique des microorganismes dans le temps et/ou l’espace, d’identifier les paramètres qui les gouvernent, d’étudier leurs capacités d’adaptation, de mieux comprendre leurs interactions et d’explorer les fonctions qu’elles exercent (Denef et al., 2010). Les DMA représentent également des habitats fragmentés, qui possèdent chacun des conditions physicochimique (T°, pH, concentration en oxygènes) et/ou des concentrations en métaux et métalloïdes différents, permettant ainsi d’aborder des questions particulières de biogéographie (Hallberg, 2010 ; Kuang et al., 2012). L’étude de la diversité de ces environnements extrêmes suscite également un intérêt important du fait que ces écosystèmes peuvent représenter un réservoir de nouvelles biomolécules ayant un intérêt biotechnologique. Enfin, les similarités qui existent entre la minéralogie de ces environnements, comme celui du Rio Tinto en Espagne, et de la planète Mars (vaste dépôts de sulfates et d’oxydes de fer) ont conduit à l’idée que les propriétés de ces acidophiles pourraient être similaires à ceux susceptibles d’être retrouvés sur Mars (Amils et al., 2007).

Parmi les éléments toxiques des DMA, l’arsenic pose un problème particulier parce qu’il est fortement assimilable par les organismes vivants du fait que ses propriétés chimiques sont très voisines de celles du phosphore et du soufre, qui sont des éléments essentiels à la vie (Yammura and Amachi, 2014). Chez l’être humain, il est toxique et induit de nombreuses pathologies dont des cancers et ne devrait pas dépasser 10 µg.l-1 selon l’OMS (Yamanaka et Okada, 1994 ; McClintock et al., 2012 ; Jiang et al., 2013). L’arsenic présent dans l’eau de boisson représente un problème mondial majeur qui touche plusieurs millions de personnes, en particulier au Bangladesh où 35-77 millions de personnes sont concernés (Nordstrom, 2000; Argos et al., 2010 ; Yunus et al., 2011) mais cela touche également de nombreux pays comme les Etats-Unis, la Chine, le Mexique, l’Espagne ou le Canada, etc. (Jiang et al., 2013). Malgré sa faible abondance dans la croûte terrestre (0.0001%), il est largement distribué dans l’environnement où il est souvent associé avec les minerais métalliques sulfurés comme le cuivre, le plomb ou l’or, etc. (Oremland and Stolz, 2003). Dans les sols, il est généralement retrouvé à des concentrations inférieures à 15 mg.kg-1 (Yammura and Amachi, 2014). Bien que l’arsenic existe sous 4 états d’oxydation différents (V, III, 0, -III) avec une multitude de formes organiques et inorganiques, l’arséniate (As(V)) et l’arsénite (As(III) sont les formes inorganiques prédominantes dans l’environnement (Ormeland and Stolz, 2005), avec l’As(III) considéré comme plus toxique que l’As(V) (Lièvremont et al., 2009 ; Yammura

and Amachi, 2014). L’As(V) est généralement présent sous forme d’oxyanions chargés - 2- négativement (H2AsO4 / HAsO4 ) à pH modéré et a ainsi tendance à être fortement adsorbé sur la surface de nombreux minéraux chargés positivement, comme les oxydes et hydroxydes de fer et d’aluminium. L’As(III) est quand à lui généralement présent sous une forme non 0 chargée (H3AsO3 ) dans l’environnement et est donc habituellement moins adsorbé et donc plus mobile que l’As(V) (Yammura and Amachi, 2014). Dans les environnements aérobies, l’As(V) est souvent la forme prédominante alors qu’en conditions anoxiques, c’est la forme As(III) qui prédomine. Certains microorganismes ne sont pas seulement résistant à l’As mais le métabolisent activement via des réactions de méthylation, déméthylation, oxydation ou réduction, modifiant ainsi les formes redox de l’As et utilisant certaines de ces étapes pour générer de l’énergie (Oremland and Stolz, 2005 ; Stolz et al., 2010). A ce jour, de nombreux microorganismes, principalement des bactéries capables d’oxyder ou de réduire l’As, ont été isolés d’environnements contaminés par l’arsenic (Oremland and Stolz, 2003 ; Lièvremont et al., 2009). La réduction de l’arséniate comprend une voie de détoxification (gène arsC) ainsi que la respiration (gènes arrA/B). L’organisation de l’opéron ars varie fortement entre les taxons et les gènes de base inclus arsR, arsB et arsC tandis que arsD et arsA peuvent également parfois être trouvés (Oremland and Stolz, 2003). Les gènes arrA/B codent une enzyme réductase active durant la respiration anaérobie, utilisant l’As(V) comme accepteur final d’électron (Costa et al., 2014). L’oxydation microbienne de l’As(III), décrite pour la première fois en 1918, peut être médiée par 2 enzymes distinctes, Aio (comprenant gènes aox, aso et aro) très étudié et Arx récemment décrite par Zargar et al. (2012). L’oxydation aérobie de l’As(III) est catalysée par une arsénite oxidase qui utilise l’O2 comme accepteur terminal d’électrons et qui est codée par les gènes aioB/A (Lett et al., 2012 ; Costa et al., 2014). ArxAB est détectée chez des bactéries oxydant As(III) en conditions anoxiques, où la réduction du nitrate ou du chlorate est couplée à l’oxydation de l’As(III) (Oremland et al., 2009 ; Sun et al., 2010 ; Costa et al., 2014). Certains membres du genre Ectothiorhodospira sont également capables d’utiliser l’As(III) comme donneur d’électrons pour la croissance phototrophe anoxygénique (Kulp et al., 2008). Parce que ces processus de réduction d’As(V) ou d’oxydation d’As(III) affectent directement la spéciation et la mobilité de l’As, l’activité microbienne joue un rôle clé dans les cycles biogéochimiques de ce métalloïde et peuvent être utilisés pour dépolluer les sols et les eaux pollués par l’arsenic (Yammura and Amachi, 2014 ; Costa et al., 2014 ; Sarkar et al., 2014).

L’ancien site minier de Carnoulès a constitué pour le laboratoire HydroSciences un cadre privilégié pour l’étude des interactions entre les microorganismes et les polluants métalliques et notamment l’arsenic (Leblanc et al., 1996). L’intérêt dans ce site réside également dans le fait qu’un système de remédiation naturel est présent où, près de 99% de l’arsenic va précipiter et être piégé dans des minéraux de fer le long des 1,5 km du Reigous, petit ruisseau alimenté par les drainages miniers acides du stérile de Carnoulès (Leblanc et al., 1996). Enfin, la proximité géographique de ce site avec le laboratoire HydroSciences est un facteur non négligeable étant donné les nombreux allers-retours nécessaires pour étudier cet environnement sur le long terme. Ce site atelier est, depuis 2009, un site d’observation de

l’Observatoire des Sciences de l’Univers OREME (tâche d’observation intitulé « Suivi des processus hydrobiogéochimiques de transfert des métaux et métalloïdes issus des activités minières sur le site de Carnoulès »). Les connaissances acquises sur ce site avec mes collègues du laboratoire HSM, M. Leblanc, (géologue), J.-C. Personné puis A. Desoeuvre (AI depuis 2010) et M. Héry en 2011 (microbiologistes) ; F. Elbaz-Poulichet et C. Casiot (géochimistes) en association avec G. Morin (minéralogiste à l’IMPMC, Paris) et en collaboration avec 2 laboratoires de microbiologie, l’EEM de Pau (R. Duran, B. Lauga) puis le laboratoire GMGM de Strasbourg (P. Bertin, F. Ploetze) ont permis de mieux comprendre cet écosystème et ont ainsi contribué à fédérer différents groupes de recherches sur ce site.

Description du site minier de Carnoulès La mine de Carnoulès est située dans les Cévennes dans le Sud de la France et a été définitivement fermée en 1962.

Figure 1. Localisation et carte du site minier de Carnoulès. D’après Bruneel et al., 2005

Au Sud-Est du Massif Central, le long des Cévennes, un horizon conglomératique de 3 à 5 m d’épaisseur contenant de la marcasite (et/ou de la pyrite), de la galène, de la barytine et accessoirement de la sphalérite, des sulfo-arséniures (proustite, arsénopyrite) et des sulfure d’antimoine (freibergite) est présent au niveau de la mine de Carnoulès (Leblanc et al., 1996). Le gisement de 2.5 Mt contenait 3.5% Pb et 0.8% Zn et a été principalement exploité à ciel ouvert puis définitivement abandonné en 1962. Le stérile actuel, d’environ 1.2 Mt qui est confiné derrière une digue, comporte les déchets d’après traitement qui contiennent encore environ 0.7% de Pb et 10% de sulfure de Fe, (Leblanc et al., 1996).

Figure 2. Schéma simplifié du dépôt de stériles miniers de Carnoulès avec en (a) la localisation du forage instrumenté, des carottages réalisés sur le site (T1, T4) et du système de drainage et en (b) une coupe dans le dépôt montrant les différents horizons, la couverture d’argile en surface, les sables gris fins et riches en pyrite, les sables grossiers et le socle composé de quartzites du Trias (b). D’après Casiot et al., 2003a

Ce stérile a une superficie de 5500 m2 et une épaisseur de 10 à 24 m. Il est recouvert d’une couche d’argile de 0.3 m d’épaisseur. En dessous, il est constitué majoritairement de sables à pyrite contenant 75% de quartz et entre 5 et 15 % de pyrite qui contient de 1 à 4% d’As. Les minéraux secondaires incluent le K-feldspath, la biotite, la barytine et la galène (Alkaaby et al., 1985). Ces matériaux sont généralement très fins (taille moyenne des grains de 30 µm) et peu perméables, excepté près du fond où une couche de 2 à 3 mètres d’épaisseur contient du matériel ferrugineux relativement grossier (200 µm) (Koffi et al., 2003). L’oxydation des sulfures est limitée dans la partie supérieure du dépôt, contrairement à ce qui est généralement constaté dans d’autres stériles miniers et est probablement dû à la présence d’une couverture argileuse peu perméable et à la faible conductivité hydraulique du matériau qui limite l’infiltration des eaux de pluie (Koffi et al., 2003). Dans la partie inférieure du dépôt au contraire, les sulfures sont partiellement oxydés en liaison avec la présence d’un drain et la circulation d’eaux à la base du stock, dans une zone à matériaux plus grossiers très probablement en raison de la présence de sources enterrées présentes sous le stérile (Koffi et

al., 2003). Le niveau de l’eau se situe entre 1 et 10 m sous la surface en fonction de la localisation dans le stérile et de la saison. L’eau qui circule dans le stock de déchet donne naissance au ruisseau du Reigous dont la source apparaît à la base de la digue qui retient les déchets. La masse d’arsenic contenu dans ce stock de stériles est estimée à 3000 t. Compte tenu de la masse annuelle d’As rejetée par la source acide (6 t), la durée de vie du système est estimée à au moins 500 ans (Leblanc et al., 2002). Les études physicochimiques ont montré que le débit à la source est relativement faible (0.2 à 1 l.s-1) mais ces eaux coulent toute l’année. Elles sont pratiquement anoxiques à la source mais en quelques dizaines de mètres, on observe une augmentation de la concentration en oxygène. Les flux annuels d’arsenic, calculés au cours de 2 années hydrologiques aux caractéristiques différentes, varient de 2 à 6 t. Les concentrations en As diminuent rapidement en aval, juste avant le confluent avec l’Amous, elles sont en moyenne de 6 mg.l-1 avec de très fortes variations saisonnières (Leblanc et al., 1996). Ces diminutions sont à attribuer en partie à des dilutions avec de petits rus latéraux mais surtout à la précipitation de l’arsenic et à la formation de sédiments riches en fer et en arsenic. Les variations saisonnières du système du Reigous sont fortement marquées : en période d’étiage, les sédiments arséniés s’accumulent mais, en période de fortes pluies (printemps, automne), les sédiments sont érodés et transportés, entraînant une forte augmentation du flux d’arsenic avec un transport essentiellement sous forme particulaire (Leblanc et al., 2002).

A mon arrivée au laboratoire HydroSciences dans le cadre de mon stage de DESS en 2002, les travaux de Leblanc et al. (1996) avaient permis de mettre en évidence à Carnoulès, dans le ruisseau du Reigous qui draine le site, la formation de précipités contenant près de 20% d’arsenic autour de structures bactériennes, mais les processus géochimiques et microbiologiques à l’origine de la formation de ces solides n’étaient pas connus. Durant ce stage, sous l’encadrement de Jean Christian Personné, j’ai isolé une vingtaine de colonies bactériennes dans les eaux du stock de déchets miniers ainsi que dans les eaux le long du Reigous et j’ai commencé leurs études en laboratoire et en particulier, leurs activités sur l’oxydation du fer et de l’arsenic. La grande majorité de ces souches ont été identifiées comme étant des bactéries des genres Thiomonas et Acidithiobacillus ferrooxidans. Ce travail a permis de décrire plusieurs souches de Thiomonas et de montrer pour la première fois que des souches pures de Thiomonas étaient capables d’oxyder l’arsenic2.

Suite à ce travail, j’ai débuté en décembre 2002 une thèse intitulée « Contribution à l'étude des mécanismes couplés géochimiques et bactériologiques de transfert de la pollution minière sur le site de Carnoulès (Gard) » sous l’encadrement de Jean Christian Personné et de François Elbaz Poulichet, ma directrice de thèse. Bien qu’apportant des informations très

2 Bruneel O, Personné J-C, Casiot C, Leblanc M, Elbaz-Poulichet F, Mahler BJ, Le Flèche A, Grimont PAD. (2003) Mediation of arsenic oxidation by Thiomonas sp. in acid mine drainage (Carnoulès, France). Journal of Applied Microbiology. 95, 492-499 24

intéressantes sur le métabolisme des souches isolées, les techniques classiques d’isolement de souches pures et la caractérisation de leurs activités en laboratoire ne permettent pas de comprendre un écosystème étant donné que près de 99% des organismes ne peuvent être pour l’instant isolés par des approches culturales (Rappé and Giovannoni, 2003). J’ai donc rapidement été amenée à travailler avec le Laboratoire d’Ecologie Moléculaire (R. Duran, EA 3525, Ecologie Moléculaire Microbiologie de l’Université de Pau) pour mettre en œuvre une approche moléculaire qui n’était pas disponible, à l’époque, au Laboratoire HydroSciences.

Processus de génération du drainage minier acide riche en As de Carnoulès Les eaux de drainage de mines sont générées par l’exposition des minerais sulfurés, telles que la pyrite (FeS2) à l’oxygène et à l’eau (Johnson and Hallberg, 2003 ; Vera et al., 2013). De nombreux métaux sont présents sous forme de minerais sulfurés, comme la galène (PbS) ou la sphalérite et sont également souvent associés à la pyrite qui est le minerai sulfuré le plus commun. Le fer ferrique (Fe(III)) est le principal oxydant des minerais sulfurés (Baker and Banfield, 2003) : 3+ 2+ 2− + FeS2 + 14 Fe + 8 H2O → 15 Fe + 2 SO4 + 16 H La régénération du Fe(III), selon l’équation ci-dessous, est l’étape limitante de l’oxydation des minerais et nécessite de l’oxygène (Singer and Stumm, 1970) : 2+ + 3+ 14Fe + 3.5 O2 + 14H → 14Fe + 7H2O A pH supérieur à 4, l’oxydation du fer ferreux se produit chimiquement en présence d’oxygène ou biologiquement mais à des pH inférieures à 4, le taux d’oxydation chimique est très lent, voir négligeable et c’est l’activité des microorganismes oxydant le fer qui va avoir un rôle pivot dans la génération des DMA (Baker and Bandfield, 2003 ; Vera et al., 2013). De plus, en raison des faibles pH rencontrés dans ces environnements (jusque -3 comme dans la mine de Richmond aux Etats Unis (Californie, Nordstrom et al 2000), la solubilité des métaux est plus importante et les DMA contiennent donc généralement de très fortes concentrations en métaux et métalloïdes qui vont varier en fonction de la minéralogie de la roche d’origine (Hallberg, 2010).

Des études réalisées au sein du piézomètre S5, situé approximativement au centre du stérile minier en 2001 et 2002, ont montré de très fortes variations de la chimie sur une année qui semblaient être liées au niveau de la nappe et aux concentrations en oxygène dissous3. En période de remontée de la nappe, le niveau d’oxygène est très élevés (7-9 mg.l-1), le pH est acide (1.8), et de très fortes concentrations de fer (proche 20000 mg.l-1) et d’As (jusque 12000 mg.l-1, concentrations parmi les plus importantes au monde) ont été relevées avec les espèces oxydantes qui dominent (As(V) et Fe(III)). Ces teneurs très élevées ont été attribuées à la dissolution de phases secondaires, en particulier des hydroxysulfates de fer contenant jusque 10% d’As, présents dans le stock de déchets. A l’inverse, lorsque le niveau de la nappe diminue et que le milieu devient pratiquement anoxique (DO = 0.5 mg.l-1), le pH remonte autour de 4 et les concentrations en As et en Fe diminuent fortement et se stabilisent (autour

3 Casiot C, Leblanc M, Bruneel O, Personné J-C, Koffi K, Elbaz-Poulichet F (2003) Geochemical processes controlling the formation of As-rich waters within a tailings impoundment. Aquatic Geochemistry. 9, 273-290 25

de 3000 mg.l-1 pour Fe et 750 mg.l-1 pour As) avec As et Fe principalement sous forme réduite As(III), Fe(II). Dans le cadre de ma thèse et pour tenter de mieux comprendre la génération de ces eaux acides et riches en métaux ainsi que les variations associées, j’ai initié l’étude des bactéries présentes par des approches moléculaires ciblant l’ARNr 16S par les techniques de clonage- séquençage. Ces analyses nous ont permis de mettre en évidence que la diversité était faible comparée à des eaux non polluées avec un total de 5 taxons identifiés ici4.

Table 1. Inventaire des fragments d’ADNr 16S des clones présents en octobre 2001 (S5Oct) et janvier 2002 (S5Jan) dans les eaux du stock de déchets miniers, groupés selon l’analyse RFLP et l’analyse phylogénétique. D’après Bruneel et al., 2005

Ce travail a également montré que ce sont curieusement des groupes proches des bactéries sulfato-réductrices (BSR, Desulfosarcina variabilis) qui dominent et ceci principalement lorsque les concentrations en oxygène sont élevées et le pH très bas alors que ces bactéries sont pourtant connues pour préférer les conditions anoxiques. En octobre, quand le taux d’oxygène est faible, on trouve des organismes dont les séquences sont affiliées à Desulfosarcina variabilis (représentant environ 27% du nombre total de clones) associés à des séquences apparentées à Acidithiobacillus ferrooxidans, Thiobacillus et Acidimicrobium alors qu’en janvier, lorsque les conditions sont très oxygénées et le pH très acide, la communauté bactérienne est composée essentiellement de Desulfosarcina variabilis (représentant environ 95%) associée à Acidithiobacillus ferrooxdians et Thiobacillus spp.

4 Bruneel O, Duran R, Koffi K, Casiot C, Fourçans A, Elbaz-Poulichet F, Personné J-C (2005) Microbial diversity in a pyrite-rich tailings impoundment (Carnoulès, France). Geomicrobiology Journal. 22, 249 - 257 26

Etude du système de remédiation présent sur le site de Carnoulès Contrairement aux composés organiques qui peuvent être dégradés en composés simples et sans risque pour la santé comme le CO2 ou l’H2O du fait de leur minéralisation, la remédiation des métaux et métalloïdes implique seulement leur retrait de la solution dans le milieu aquatique (Bahar et al 2013). Cette remédiation de l’eau est due à des réactions biotiques et abiotiques qui font que ces composés toxiques deviennent insolubles et précipitent, s’accumulant dans des sédiments composés généralement d’une variété d’(oxyhydr)oxydes et d’hydroxysulfates de fer tel que la jarosite, la schwertmannite ou la ferrihydrite (Johnson and Hallberg, 2005; Johnson, 2014). Ces processus de précipitation résultent en grande partie de l’oxydation et de la précipitation du fer, qui est souvent le principal métal soluble présent dans le DMA, et de l’adsorption d’autres métaux et métalloïdes comme le Pb, l’U ou As sur les minéraux sulfurés formés (Hallberg, 2010). Comme l’oxydation abiotique du Fe(II) est un processus très lent dans les eaux acides, les microorganismes oxydants le fer qui catalysent ces réactions jouent un rôle pivot dans les processus de remédiation (Rowe and Johnson, 2008; Hallberg, 2010; Johnson, 2014).

A Carnoulès, les études de Leblanc et al. (1996, 2002) avaient révélé la présence d’un système de remédiation naturel efficace dans les eaux du Reigous qui permettait de limiter les concentration de l’As en aval du système. Pour identifier les processus chimiques et microbiologiques qui influencent le transfert de l’As dans le Reigous, des échantillons d’eau ont été prélevés lors de 8 campagnes de prélèvements en 2001. Les stations de prélèvement étaient situées dans les 30 premiers mètres du ruisseau où aucun apport latéral d’eau n’avait été observé.

Figure 3. Coupe montrant la localisation des stations de prélèvements dans les 30 premiers mètres du ruisseau du Reigous (1, A, C, E, F, 2). Le temps d’écoulement des eaux entre les stations 1 et 2 est d’environ 1 heure. D’après Casiot et al., 2003b

Les analyses physicochimiques réalisées dans le ruisseau du Reigous qui draine le site ont montré que l’arsenic en solution est essentiellement sous forme réduite (As(III)) à la source du Reigous5. Sur les 30 premiers mètres, 20 à 60% de l’arsenic coprécipite en liaison avec

5 Casiot C, Morin G, Juillot F, Bruneel O, Personné JC, Leblanc M et al. (2003b) Bacterial immobilization and oxidation of arsenic in acid mine drainage (Carnoulès creek, France). Water Res. 37, 2929-2936 27

l’oxydation du Fe(II) en Fe(III). Les formes méthylées sont absentes. Le taux de précipitation est variable selon les saisons. Il semble plus important pendant la saison humide lorsque la teneur en oxygène dans les eaux à la source est plus élevée.

Pour tenter de mieux comprendre ce système de remédiation, des études moléculaires combinées (comprenant des analyses de clonage et séquençage par la méthode de Sanger ainsi que des analyses t-RFLP) ont été réalisées afin d’identifier les communautés bactériennes présentes6.

Table 2. Inventaire des fragments d’ADNr 16S des clones présents dans les eaux du ruisseau du Reigous groupés selon l’analyse RFLP et l’analyse phylogénétique. D’après Bruneel et al, 2006

a abondance relative des clones dans chaque librairies

Ces analyses ont montré, comme au sein du stérile, une diversité faible avec l’identification de 2 à 4 taxons par échantillons. De plus, comme attendu en raison de la chimie de l’eau, les résultats ont mis en évidence que les eaux du ruisseau acide étaient largement dominées par des bactéries impliquées dans le cycle du fer et du soufre. Les séquences affiliées à la bactérie neutrophile qui oxyde le Fe, Gallionella ferruginea, sont largement dominante. Cette bactérie pourrait jouer un rôle important dans la remédiation naturelle observée dans le ruisseau acide en immobilisant l’As par coprécipitation avec le Fe(III).

La structure et la spéciation de l’As dans les sédiments du Reigous ont été également caractérisées par des analyses spectroscopiques et minéralogiques (XRD, XANES et SEM)7. Cette étude a mis en évidence des variations spatiales et temporelles des précipités formés dans le ruisseau. Pendant la saison humide, les précipités présents dans les 10 premiers mètres du ruisseau consistent essentiellement en tooéléite (un minéral rare de Fe6(AsO3)4- (SO4)(OH)4•4H2O) associée à des précipités amorphes d’As(III)-Fe(III). Pendant la saison sèche, la formation d’un oxyhydroxyde de Fe(III)-As(V) amorphe prédomine.

6 Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F, Personné JC (2006) Diversity of microorganisms in Fe-As-rich acid mine drainage waters of Carnoulès, France. Applied and Environmental Microbiology. 72, 551-556 7 Morin G, Juillot F, Casiot C, Personné JC, Elbaz-Poulichet F, Leblanc M, Ildefonse P, Calas G (2003) Bacterial formation of tooeleite and mixed arsenic(III) or arsenic(V)-iron(III) gels in the Carnoulès Acid Mine Drainage, France. A XANES, XRD, and SEM study. Environmental Science and Technology. 37, 1705-1712

Dans la poursuite des études réalisées dans le cadre du DESS, le rôle des bactéries isolées (Thiomonas et Acidithiobacillus ferrooxidans) dans les réactions de précipitation de l’As et du Fe a été testé grâces à des études en laboratoire. Six souches (B1 à B6) isolées à partir de l’eau du ruisseau ont été inoculées individuellement dans l’eau de la source. B1, B2, B3 et B6 sont des souches du genre Thiomonas et B4 et B5 sont des Acidithiobacillus ferrooxidans. En parallèle, les précipités formés par les souches B1 à B6 ont été analysés.

Figure 3. Essais en laboratoire présentant le pourcentage d’As total (AsT), de Fe(II) et d’As(III) éliminés après une semaine d’incubation de différentes souches de microorganismes (B1 à B6) isolées à partir de l’eau du Reigous avec les concentration en As(V) en solution en fin d’expérience. S1: eau de la source avant incubation et SA: témoin stérile. D’après Casiot et al., 2003b

Ces études ont montré que 3 souches de Thiomonas ont la capacité d’oxyder l’As(III) dans l’eau du Reigous: B2, B3 et B6. C’est la souche B6, qui du fait de l’oxydation simultanée de l’arsenic et du fer, entraîne le plus grand battement d’As (87%)5. Les précipités du témoin abiotique, tout comme dans ceux des bactéries B1, B2, B3, B4 et B6, sont constitués essentiellement d’hydroxydes sulfates ferriques d’As(V)7. Une souche d’Acidithiobacillus ferrooxidans (bactérie B5) permet la formation de tooéléite nanocristalline associée à un mélange de composés d’oxyhydroxydes d’As(III)/As(V)-Fe(III) amorphes. Des études avec d’autres souches d’Acidithiobacillus ferrooxidans ont également montrées que ce genre était capable de faire précipiter rapidement l’arsenic avec le Fe(III) en milieux synthétiques sous forme de schwertmannite8.

La bioremédiation de l’arsenic peut s’appuyer sur l’activité des microorganismes qui ont la capacité de détoxifier, mobiliser ou immobiliser l’As à travers différents processus comme l’oxydation, la réduction, la biométhylation, la sorption ou la complexation (Oremland et al., 2005 ; Bahar et al., 2013). Etant donné que l’As(III) est fortement toxique et mobile dans l’environnement, une stratégie de remédiation intéressante consiste généralement à le convertir en As(V), forme moins toxique et mobile qui a tendance à se fixer sur différents types de matrices (Bahar et al., 2013).

8 Duquesne K, Lebrun S, Casiot C, Bruneel O, Personné J-C, Leblanc M, Elbaz-Poulichet F, Morin G, and Bonnefoy V (2003) Immobilization of arsenite and ferric iron by Acidithiobacillus ferrooxidans in acid mine drainage. Applied and Environmental Microbiology. 69, 6165-6173 29

Aux cotés des bactéries, les eucaryotes, présents dans ces drainages miniers acides, ont également développé des stratégies de résistance vis-à-vis de cet élément toxique dans l’environnement. Euglena mutabilis est un protozoaire photosynthétique, communément rencontré dans les eaux minières acides qui semble bien adapté aux conditions extrêmes qui y règnent (Brake et al., 2001). Cet organisme peut jouer un rôle important dans les DMA en contribuant à l’apport d’oxygène par leur activité photosynthétique, en séquestrant le fer et probablement d’autres métaux par précipitation intracellulaire et en apportant de la matière organique (Brake et al., 2001). A Carnoulès, les euglènes sont présentes en grand nombre dans le Reigous et sont visibles grâce à la présence de tapis verts caractéristiques, pouvant atteindre une épaisseur d’environ 1 cm. L’étude de la dynamique saisonnière de ces biofilms a montré qu’elle n’est pas liée aux variations des concentrations en polluants métalliques mais à l’effet de l’érosion mécanique du sédiment du fait des fortes précipitations9. Cultivés en milieu synthétique en présence de 0.2 à 300 mg/l d’As(III), les euglènes issues du site de Carnoulès accumulent l’As à l’intérieur de leurs cellules sous forme d’arsénite et d’arséniate dont les concentrations varient en fonction des concentrations en As(III) du milieu de culture. L’arsenic est également adsorbé à la surface de la cellule sous forme d’As(V).

9 Casiot C, Bruneel O, Personné J-C, Leblanc M, Elbaz-Poulichet F (2004) Arsenic oxidation and bioaccumulation by the acidophilic protozoan, Euglena mutabilis, in acid mine drainage (Carnoulès, France). Science of the Total Environment. 320, 259-267 30

L’étude de cet écosystème s’est poursuivie après ma thèse dans le cadre de plusieurs projets de recherche. Une étude sur les Archaea a été réalisée suite aux résultats de l’étude moléculaire réalisée sur les bactéries présentes dans les eaux souterraines au sein du stock de déchets miniers. Historiquement, on pensait en effet que les bactéries représentaient les principaux microorganismes impliqués dans les processus de lixiviation à l’origine de la formation des DMA mais il a été montré depuis plusieurs années maintenant que les Archaea sont elles aussi susceptibles de jouer un rôle majeur, de part de leurs capacités à oxyder le fer dans les processus de génération et/ou de remédiation des DMA (Edwards et al., 2000c ; Baker and Banfield, 2003 ; Bini, 2010).

Table 3. Inventaire des fragments d’ADNr 16S des clones présents dans les eaux au sein du stock de déchets miniers et dans le DMA du Reigous. D’après Bruneel et al., 2008

a abondance relative des clones dans chaque librairie

Cette étude a révélé que l’ensemble des séquences retrouvées était affilié au phylum des Euryarchaeota, tandis que les Crenarchaeota n’étaient pas du tout présentes10. Ce travail a

10 Bruneel O, Pascault N, Egal M, Bancon-Montigny C, Goni M, Elbaz-Poulichet F, Personné J-C, Duran R (2008) Archaeal diversity in a Fe-As rich acid mine drainage at Carnoulès (France). Extremophiles. 12, 563-571 31

également montré que la structure des communautés d’Archaea dans l’aquifère au sein du stock de déchet minier était très différente de celle présente dans les eaux du ruisseau du Reigous drainant le site. Au sein de la nappe qui draine le stock de déchets, les séquences majoritaires sont proches de Ferroplasma acidiphilum, un microorganisme acidophile, sans paroi cellulaire, oxydant le fer et connu pour son rôle majeur dans le lessivage (Golyshina and Timmis, 2005). Dans les eaux du Reigous, par contre, ce sont des séquences affiliées à un groupe de Thermoplasmatales non cultivé, le clone YAC1, qui est largement dominant.

Une étude en collaboration avec le laboratoire EEM de Pau s’est également intéressée aux bactéries sulfato-réductrice présentent sur ce site pour essayer de déterminer l’influence des paramètres environnementaux sur la structure de ces communautés par analyse t-RFLP et étude des gènes dsrAB11. Ce travail réalisé sur une période de 3 ans a permis de mettre en évidence la présence prédominante de la famille Desulfobulbaceae dans le système et a montré que la dynamique des bactéries sulfato-réductrices semble être liée aux fluctuations spatio-temporelles du pH, du fer et des formes spécifiques de l’arsenic.

Pour tenter de mieux comprendre le fonctionnement du système de remédiation présent dans le drainage minier acide de Carnoulès, les concentrations en arsenic et métaux ont été suivis sur une période de plus de 4 ans sur l’ensemble du ruisseau. Cette étude a mis en évidence que les variations saisonnières semblaient être liées aux précipitations avec une augmentation des concentrations durant les mois secs12. Environ 30% de l’As initialement présent en solution se trouve sous forme d’As(III) qui coprécipite avec le fer dans les 40 premiers mètres du ruisseau. La minéralogie de ces précipités varie spatialement et saisonnièrement. Dans les 40 premiers mètres, on trouve des composés amorphes d’As(V)- As(III)-Fe(III) associés à de la tooéléite alors que plus en aval, ces phases d’oxyde de Fe sont remplacées par de la schwertmannite et de la ferrihydrite12. Des travaux en collaboration avec l’IMPMC de Paris, ont permis de montrer que ces minéraux d’arsenic se forment généralement en étroite association avec des cellules bactériennes dans le milieu extracellulaire ou dans le périplasme des cellules ainsi qu’autour d’abondantes vésicules organiques d’origine inconnue13. Les conditions de formation de ces minéraux de fer très riches en As(III) et As(V) et l’implication des bactéries dans ces processus ont été par ailleurs étudiées en laboratoire. Des bactéries du genre Acidithiobacillus ferrooxidans inoculées dans l’eau du Reigous conduisent à la formation d’un assemblage de minéraux (schwertmannite, tooéléite) qui est différent de celui obtenu en conditions abiotiques où l’on trouve généralement de la jarosite. De plus, la proportion des minéraux formés semble différer selon la souche d’Acidithiobacillus

11 Giloteaux L, Duran R, Casiot C, Bruneel O, Elbaz-poulichet F and Goñi-urriza M (2013) A survey of sulfate reducing bacteria in a heavily arsenic contaminated acid mine drainage (Carnoulès, France). FEMS Microbiol Ecol. 83 724–737 12 Egal M, Casiot C, Morin G, Elbaz-Poulichet F, Cordier MA Bruneel O (2010) An updated insight into the natural attenuation of As concentrations in Reigous Creek (southern France). Applied Geochemistry. 25, 1949–1957 13 Benzerara K, Morin G, Yoon TH, Miot J, Tyliszczak T, Casiot C, Bruneel O, Farges F, and Brown Jr GE (2008) Nanoscale study of As transformations by bacteria in an acid mine drainage system. Geochimica et Cosmochimica Acta. 72, 3949-3963 32

ferrooxidans et la taille de l’inoculum de départ14. Ces résultats semblent montrer que les bactéries peuvent influencer la composition minéralogique du précipité formé en intervenant sur la cinétique d’oxydation du fer, principalement durant les premiers stades d’incubation. La tooéléite, par exemple, ne semble se former que lorsque la cinétique d’oxydation du fer est lente et que le rapport As(III)/Fe(III) est élevé (≥ 0.8) avec des concentrations en Fe(III) du même ordre de grandeur que celles d’As(III). Pour mieux comprendre l’implication des microorganismes du genre Thiomonas, une collaboration avec le laboratoire GMGM de Strasbourg a permis de séquencer le génome de l’une de ces souches, Thiomonas sp. 3As, ce qui a permis de révéler les adaptations spécifiques de cet organisme lui permettant de survivre et de résister aux concentrations élevées de métaux et métalloïdes dans ces environnements extrêmes15. De plus, 8 souches différentes incluant 5 souches de la même espèce, ont également été comparées par hybridation génomique comparative. Le génome du genre Thiomonas semble avoir évolué à travers le gain ou la perte d’ilots génomiques, comme ceux conférant la résistance à l’As (opéron ars) par exemple. Cette capacité a permis à cette espèce de s’adapter à son environnement et suggère aussi que l’environnement influence l’évolution génomique de ces bactéries. Ces résultats soulignent de plus la variabilité très importante qui peut exister à l’intérieur d’un même groupe taxonomique, élargissant le concept d’espèces.

Etude des sédiments présents dans le DMA du Reigous En raison de la précipitation des éléments toxiques en solution, présent en grande quantité dans ces DMA, les sédiments de ces cours d’eau agissent comme des puits et accumulent de fortes quantités de composés métalliques toxiques. Cependant, ces éléments peuvent également être relargués dans l’eau en fonction de changements dans la chimie des sédiments, de l’évolution du régime hydrologique ou de l’activité microbienne et peuvent ainsi représenter une source potentielle de métaux et de métalloïdes toxiques (Park et al., 2006; Butler, 2011; Héry et al., 2014). Etudier les microorganismes présents et leurs fonctions dans de tels écosystèmes est donc également très important pour comprendre le devenir des polluants.

Toujours en collaboration avec le GMGM de Strasbourg (F. Ploetze), une étude s’est intéressée aux populations actives de ces écosystèmes présentes à la fois dans l’eau et les sédiments du ruisseau du Reigous. L’utilisation d’une méthode de clonage-séquençage nous a permis d’identifier les différentes populations présentes et une étude de métaprotéomique

14 Egal M, Casiot C, Morin G, Parmentier M, Bruneel O, Lebrun S, Elbaz-Poulichet F( 2009). Kinetic control on the formation of tooeleite, schwertmannite and jarosite by Acidithiobacillus ferrooxidans strains in an As(III)-rich acid mine water. Chemical Geology. 265, 432-441 15 Arsène-Ploetze F, Koechler S, Marchal M, Coppee J-Y, Chandler M, Bonnefoy V, Barakat M, Barbe V, Battaglia -Brunet F, Brochier-Armanet C, Bruneel O, G. Bryan C, Cleiss J, Heinrich-Salmeron A, Hommais F, Joulian C, Krin E, Lieutaud A, Lièvremont D, Michel C, Muller D, Ortet P, Proux C, Siguier P, Roche D, Rouy Z, Salvignol G, Slyemi D, Talla E, Weiss S, Weissenbach J, Médigue C, Bertin PN (2010) Structure, function and evolution of the Thiomonas spp. genome inferred from sequencing and comparative analysis. PLoS Genetics. 6 (2) e1000859 33

nous a permis d’identifier les membres actifs de ces environnements16. Ce travail a été réalisé en partie dans le cadre de la thèse d’Aurélie Volant (2009-2012, bourse environnée) que j’ai principalement encadré, intitulée « Etude des communautés microbiennes (Bactéries, Archaea et Eucaryotes) et de leurs variations spatiotemporelles dans la mine de Carnoulès fortement contaminée en arsenic ».

Table 4. Inventaire des fragments d’ADNr 16S des clones présents dans les eaux de drainage et les sédiments au point COWG, 30 mètres en aval de la source dans le ruisseau du Reigous. D’après Bruneel et al, 2011

Les analyses taxonomiques des banques de gènes codant pour l’ARNr 16S ont permis de montrer que la diversité bactérienne est plus faible dans l’eau que dans les sédiments avec 11 souches identifiées dans l’eau contre 13 souches dans les sédiments. Un total de 17 groupes taxonomiques différents ont été identifiés avec seulement 7 genres présents à la fois dans l’eau et les sédiments. La plupart des ces bactéries étaient affiliées à des β-protéobactéries telles que Gallionella ou Thiomonas mais également à des γ- protéobactéries (tel que Acidithiobacillus ferrooxidans), des α-protéobactéries (Acidiphilium), des δ-protéobactéries (Desulfomonile limimaris), des Nitrospira (Leptospirillum ferrooxidans), des Actinobacteria et des Firmicutes. Il s’agit majoritairement d’espèces trouvées communément dans les DMA avec une majorité impliquée dans les cycles du fer, de l’arsenic et du soufre. Les bactéries impliquées dans l’oxydation de l’As(III) sont affiliées à Thiomonas, celles impliquées dans l’oxydation de Fe(II) sont affiliées à Gallionella, At ferrooxidans, Ferrimicrobium, Leptospirillum, Sideroxydans lithotrophicus, et Ferrovum myxofaciens alors que la réduction

16 Bruneel O, Volant A, Gallien S, Chaumande B, Casiot C, Carapito C, Bardil A, Morin G, Brown Jr GE, Personné JC, Le Paslier D, Schaeffer C, Van Dorsselaer A, Bertin PN, Elbaz-Poulichet F, Arsène-Ploetze F (2011) Characterization of the active bacterial community involved in natural attenuation processes in arsenic-rich creek sediments. Microbial Ecology. 61, 793-810 34

du fer a été mise en évidence pour les bactéries des genres Acidiphilium, Desulfuromonas svalbardensis, Acidocella, Rhodoferax ferrireducens, At ferrooxidans et Ferrimicrobium acidiphilum. Concernant le cycle du soufre, des populations capables d’oxyder des composés soufrés inorganiques ont été mis en évidence telles que Thiobacillus, Thiomonas ou At ferrooxidans. Des bactéries sulfato-réductrices comme Desulfomonile limimaris ou Desulfuromonas svalbardensi pourraient être impliquées dans la consommation du sulfate. L'oxydation de l'arsenic associée à l'oxydation du fer et l'oxydation du soufre pourrait contribuer à la co-précipitation de ces éléments et expliquerait l'atténuation de la contamination arséniée constatée dans le Reigous. L’étude par métaprotéomique faite au niveau des sédiments a révélé que les genres oxydants le fer comme Gallionella et Acidithiobacillus et oxydants l’As comme Thiomonas comptent parmi les membres métaboliquement actifs de la communauté procaryote du Reigous.

Nous avons également caractérisé la communauté d’Archaea présente dans ces sédiments et avons étudié sa dynamique temporelle en utilisant la technique de clonage-séquençage du gène codant pour l’ARNr16S. Les Archaea restent pour l’instant assez mal connues et peu étudiées en raison notamment de difficultés d’isolement qui font que les connaissances sur leurs métabolismes ne concernent pour l’instant qu’un nombre restreint de souches. Ainsi, la diversité de ces organismes et leurs rôles physiologique au sein des DMA sont assez obscur et les études moléculaires restent indispensables.

Figure 4. Arbre phylogénétique basé sur le gène codant pour l'ARNr 16S représentant l'affiliation taxonomique de la communauté des Archaea présente dans les sédiments du ruisseau du Reigous au point COWG. Le nombre entre parenthèses indique le nombre de séquences de clones pour la période d'échantillonnage représenté par un symbole (Avril 2006, Octobre 2008, Janvier 2009 et Novembre 2009 ; Volant et al., 2012)

L'affiliation taxonomique des Archaea a montré un faible degré de diversité avec uniquement 2 phylums détectés: les Thaumarchaeota (contenant la grande majorité des séquences) et les Euryarchaeota17. Contrairement aux Archaea retrouvées dans les eaux du

17 Volant A, Desoeuvre A, Casiot C, Lauga B, Delpoux S, Morin G, Personné JC, Héry M, F Elbaz-Poulichet, Bertin P and Bruneel O (2012) Archaeal diversity: temporal variation in the Arsenic-Rich Creek Sediments of Carnoulès Mine, France. Extremophiles. 16, 645-657 36

stock de déchet minier et dans les eaux du DMA, nous n’avons pas retrouvé ici de microorganismes directement impliqués dans le cycle du fer ou du soufre. Un grand nombre de séquences sont affiliées à Thermogymnomonas acidicola, une Archaea hétérotrophe retrouvée dans d’autres DMA et qui pourrait jouer un rôle important dans l’écosystème en utilisant les composés organiques qui peuvent être toxiques pour certains autotrophes (Hallberg, 2010 ; Yang et al., 2014). Des microorganismes affiliés à des Archaea méthanogènes, impliquées dans le cycle du carbone, telle que Methanomassiliicoccus luminyensis, ont également été identifiées. Enfin, des séquences apparentées à Candidatus Nitrososphaera viennensis et Candidatus nitrosopumilus sp., des Archaea impliquées dans l’oxydation de l’ammonium, une étape clé du cycle de l’azote, ont été décris. L’ensemble de ces microorganismes pourrait donc contribuer conjointement avec les bactéries au processus de remédiation observé in situ. Cette étude a également permis de mettre en évidence des modifications importantes de la structure et de la composition de la communauté d’Archaea au cours du temps qui sont probablement liées à des modifications de l’environnement.

Une étude, en collaboration avec l’IMPMC de Paris (G. Morin), s’est également intéressée aux conditions de formation des minéraux de fer riches en arsénite As(III) et arséniate As(V) identifiés sur le site, tel que la schwertmannite qui joue un rôle très important dans la rétention de l’As et la remédiation de ce métalloïde. Ce travail a montré que l’oxydation bactérienne de l’arsenic, en favorisant la formation d’As(V)-schwertmannite ou d’arséniate ferrique, améliore grandement l’immobilisation de l’As dans la phase solide18. Une autre étude avec l’IMPMC de Paris s’est également intéressée à la structure de la ferrihydrite, un oxyhydroxyde de fer qui est également impliqué dans la rétention de l’As. C’est la phase minérale prédominante présente dans les sédiments de la rivière Amous (pH 6−7) qui se forme, après la confluence, après neutralisation avec les DMA du Reigous. Ces travaux montrent que cet oxyhydroxyde de fer pourrait également jouer un rôle important dans la séquestration de l’As dans les environnements miniers19. Des études géochimiques ont également porté sur le thallium et ont montré que la forme réduite du thallium Tl(I) est largement prédominante dans le DMA de Carnoulès et est peu adsorbé sur les particules de ferrihydrite, qui se forment dans la rivière Amous en aval du Reigous, impliquant une forte mobilité du thallium dans l’hydrosystème aval20.

18 Maillot F, Morin G, Juillot F, Bruneel O, Casiot C, Ona-Nguema G, Wang Y, Lebrun S, Aubry E, Vlaic G, Brown GE Jr (2013) Structure and reactivity of As(III)- and As(V)-rich schwertmannites and amorphous ferric arsenate sulfate from the Carnoulès acid mine drainage, France: comparison with biotic and abiotic model compounds and implications for As remediation. Geochimica et Cosmochimica Acta. 104, 310-329 19 Adra A, Morin G, Ona-Nguema G, Maillot F, Casiot C, Bruneel O, Lebrun S, Juillot F, Brest J (2013) Arsenic Scavenging by Al-Substituted Ferrihydrites in a Circumneutral pH River Impacted by the Acid Mine Drainage of Carnoulès, Gard, France. Environmental Science and Technology. 47, 12784-12792 20 Casiot C, Egal M, Bruneel O, Verma N, Parmentier M, Elbaz-Poulichet F (2011) Predominance of aqueous Tl(I) species in the river system downstream from the abandoned Carnoulès mine (Southern France). Environmental Science & Technology. 45, 2056-2084 37

Etude de la diversité fonctionnelle des communautés du Reigous par métagénomique La diversité fonctionnelle des microorganismes présents au sein du drainage minier acide de Carnoulès a été étudiée dans le cadre du projet ANR RARE initié par P. Bertin (GMGM de Strasbourg). Au départ prévu sur l’eau, le manque de matériel biologique a conduit la réalisation de cette étude dans les sédiments de surface au point COWG. Le séquençage massif et le réassemblage de l’ADN, réalisé par le Génoscope d’Evry, ont conduit à la reconstruction de 7 pseudo-génomes microbiens (CARN1 à CARN7) présents dans cet environnement21.

Tableau 5. Analyse phylogénétique des pseudogénomes présents dans le sédiment de la mine de Carnoulès réalisée à l’aide de 27 marqueurs universels ou du gène de l’ARNr 16S avec RDP. D’après Bertin et al., 2011

(1) Pour le 16S, l’organisme le plus proche a été obtenu par recherche à l’aide du logiciel BLAST sur la base de donnée NCBI nr. Seuls les microorganismes ayant un pourcentage de similarité >90 sont indiqués. La recherche des 27 marqueurs universels a été réalisé selon Ciccarelli et al., (2006). (2) Absence du gène de l’ARNr 16S

Cette analyse a confirmé la présence de souches identifiées par les études antérieures réalisées par PCR/Clonage/séquençage comme Thiomonas, Acidithiobacillus ferrooxidans, Thiobacillus sp. ou Gallionella. L’utilisation conjointe de la métagénomique et de la métaprotéomique a également permis de mettre en évidence les relations entre les microorganismes ainsi que les fonctions importantes dans cet environnement.

21 Bertin PN, Heinrich-Salmeron A, Pelletier E, Goulhen- Chollet F, Arsène-Ploetze F, Gallien S, Lauga B, Casiot C, Calteau A, Vallenet D, Bonnefoy V, Bruneel O, Chane-Woon-Ming B, Cleiss-Arnold J, Duran R, Elbaz-Poulichet F, Fonknechten N, Giloteaux L, Halter D, Koechler S, Marchal M, Mornico D, Schaeffer C, Thil Smith AA, Van Dorsselaer A, Weissenbach J, Médigue C and Le Paslier D (2011) Metabolic diversity between main microorganisms inside an arsenic-rich ecosystem revealed by meta- and proteo-genomics. The ISME Journal. 5, 1735-1747. Cet article a fait l’objet d’une note dans la revue Sciences (2011, vol 332, p1128) 38

Figure 5. Modèle de la communauté bactérienne de Carnoulès mettant en évidence les fonctions majeures identifiées par le séquençage du métagénome et la métaprotéomique présentes au sein du sédiment de COWG. Les interactions entre les organismes et les composés biologiques ou chimiques sont indiquées par des flèches. Les microorganismes CARN sont numérotés de 1 à 7.

Ce travail a permis de mettre en évidence différentes activités et interactions au sein de cette communauté comme la capacité de certains microorganismes à fixer l’azote (particulièrement peu abondant dans ce type d’environnement) et le carbone inorganique par les microorganismes autotrophes permettant le développement des microorganismes hétérotrophes. La capacité de formation de biofilms, connue pour apporter une meilleure résistance face aux différents stress environnementaux (Harrison et al., 2007, Marchal et al., 2012), a été révélée ainsi que la présence de flagelles et de capsules. Les mécanismes énergétiques impliquants l’arsenic, le fer et le soufre ont été identifiés ainsi que le recyclage et le transport de la matière organiques : acides aminés, vitamines et nucléosides. En particulier, ces études ont permis l’identification d'un nouveau phylum, ‘Candidatus Fodinabacter comunificans’ qui pourrait exercer à Carnoulès un rôle indirect mais important en participant au recyclage de la matière organique comme les acides aminés ou les nucléosides provenant notamment de microorganismes eucaryotes retrouvés sur le site (Halter et al., 2012a). Ces approches génomiques ont ainsi permis de mieux comprendre le rôle des microorganismes dans l’atténuation naturelle de l’arsenic sur ce site et d’attribuer à des organismes spécifiques, dont des organismes non encore cultivés, des fonctions importantes.

Utilisation des nouvelles technologies de séquençage pour l’étude des DMA Depuis quelques années, les avancées dans le domaine des techniques de séquençage haut débit, encore appelé séquençage massif ou de nouvelle génération, ont révolutionné la biologie moléculaire et ont ouvert une nouvelle aire dans les recherches concernant les études sur la biodiversité (Sogin et al., 2006; Behnke et al., 2011). En raison de la rapidité d’obtention et du coût relativement faible (et qui ne cesse de diminuer) pour produire des millions de séquences, il est maintenant possible d’explorer en profondeur la diversité et la complexité des communautés microbiennes. Dans les études de diversité, ces techniques apparaissent comme essentielles car elles permettent d’avoir une profondeur de séquençage et une vue quasi exhaustive des microorganismes présents. Ces techniques permettent donc de s’intéresser également aux populations minoritaires qui peuvent malgré tout jouer un rôle crucial dans les processus biogéochimiques alors que l’on ne pouvait que très difficilement les étudier avant selon les méthodes traditionnelles de biologie moléculaire comme le clonage- séquençage (Behnke et al., 2011).

C’est ce type d’analyses qui a été utilisé dans le cadre de 2 études qui se sont intéressées aux microorganismes (bactéries et eucaryotes) présents dans les sédiments miniers de la mine de Carnoulès. Comme nous l’avons vu précédemment, des études antérieures principalement basées sur des souches isolées, ont suggéré un rôle déterminant de l’activité bactérienne dans la co- précipitation de l’As avec le Fe(III) et le sulfate et la formation de phases amorphes d'oxyhydroxydes associées à des minéraux comme la tooéléite, la schwertmannite ou la ferrihydrite. Cette étude a permis l’analyse de la diversité bactérienne présente dans différents types de sédiments le long des 1500 m du Reigous, en utilisant, pour la première fois une approche de pyroséquençage 454 ciblant le gène codant pour l’ARNr 16S. Le but était d’identifier les communautés bactériennes présentes et d’étudier leurs dynamiques spatiales en fonction de la structure minéralogique des sédiments (comprenant notamment des sites très riches en tooéléite et en schwertmannite) permettant de comprendre si la dynamique de ces communautés est liée aux changements dans les sédiments. Cette approche a permis de générer un total de 53075 séquences de bonne qualité après normalisation, conduisant à l'identification de 966 OTU, mettant en évidence une diversité beaucoup plus importante que précédemment observé. Il est également à noter qu’une grande majorité de cette diversité est du à la présence d’OTUs rares (371, encore appelés singletons) et observés une seule fois pour l’ensemble des séquences. Ceci suggère qu’une part importante de la diversité observée se réfère à des taxons présents à une très faible abondance donnant naissance au concept de biosphère rare (Pedrós-Alió, 2007). En dépit de l’importance de cette biosphère rare dans de nombreuses études, son rôle écologique et fonctionnel reste mal compris actuellement (Galand et al., 2009). Pour certains auteurs, ces organismes pourraient devenir dominants et actifs suite à des modifications des conditions environnementales et pourraient permettre aux processus biogéochimiques d’être maintenus limitant ainsi les effets des modifications de l’environnement (Sogin et al, 2006 ; Behnke et al., 2011 ; Bachy and Worden, 2014). D'autres études s’interrogent également sur l'exactitude des estimations de la richesse en OTUs générée par le séquençage à haut débit, qui pourrait correspondre à des erreurs de séquençage (Huse et al., 2010).

450 400 350 300 S1 250 COWG 200 GALm 150

NumberOTUs of GAL 100 CONF 50 0 0 2000 4000 6000 8000 10000 12000 Number of sequences

Figure 6. Courbes de raréfaction des séquences bactériennes des gènes codants pour l'ARNr 16S présents dans les sédiments de la mine de Carnoulès et basé sur le nombre d’OTU calculés à 97% d'identité. Le nombre total de séquences analysées est tracé en fonction du nombre d’OTU observé.

Les courbes de raréfaction ont tendance à atteindre une asymptote pour la plupart des échantillons, ce qui suggère que la majorité des phylotypes bactériens présents ont été identifiés, ce qui est confirmé par la couverture très élevée (de 98 à 100%) pour tous les échantillons.

Table 6. Estimation de la richesse en OTU, des indices de diversité et de la couverture estimée pour les 5 échantillons de sédiments. Les résultats sont présentés pour les données normalisées, rééchantillonnées au hasard pour avoir une taille d’échantillon égale entre les sites.

a Les OTUs ont été définis à 97% d’identités b Somme des probabilités des classes observées calculées comme suit (1 - (n / N)), où n est le nombre de séquences uniques (singletons) et N est le nombre total de séquences c Prend en compte le nombre et la régularité des espèces Les valeurs entre parenthèses sont des intervalles de confiance à 95%

Au total, 15 phylums ont pu être identifiés ici pour l’ensemble des échantillons, ce qui est bien plus important que ceux retrouvés dans les analyses antérieures obtenues par clonage- séquençage qui n’excédaient généralement pas 5 phylums. L'analyse phylogénétique a révélé que la grande majorité des séquences (65%) appartenaient au phylum des Proteobacteria avec

une prédominance des bactéries oxydant de fer, représentées principalement par des séquences proches de Gallionella ou Acidithiobacillus ferrooxidans. Cette analyse quasi exhaustive des taxons présents a également révélé la présence de genres abondants encore jamais détectés auparavant par les analyses de clonage/séquençage, comme les membres des Comamonas, Stenotrophomonas ou Pseudoxanthomonas avec certains d'entre eux impliqués dans l'oxydation de l’As, un métabolisme importante impliqué dans la précipitation de As. Cependant, aucun paramètre évident ne semble lier les communautés à la structure des sédiments. Ce travail devrait être prochainement soumis dans le journal FEMS Microbiology Ecology.

Une seconde étude s’est intéressée aux communautés eucaryotes. Les communautés de bactéries des DMA ont été extensivement étudiées depuis plusieurs 10aines d’années avec les premières études de diversité qui remontent au milieu des années 1990 (Goebel et Stackebrandt, 1994 ; Kuang et al., 2012). Paradoxalement, les communautés eucaryotes de ces environnements ont été très peu étudiées bien qu’un intérêt croissant leur soit porté de part leur rôle écologique potentiellement important dans ces écosystèmes. Certains sont par exemple susceptibles de modifier l’abondance, la composition et l’activité des communautés microbiennes procaryotes par de nombreux mécanismes comme la prédation (Baker et al., 2004, 2009; Gadanho et al., 2006). D’autres sont connus pour jouer un rôle important dans le cycle du carbone et le recyclage des nutriments ce qui est primordiale dans ces environnements oligotrophes (Baker et al., 2004). Certains peuvent également apporter de l’oxygène au milieu par leurs activités photosynthétiques ou encore séquestrer des polluants métalliques dans les matrices extracellulaires ou à l’intérieur de la cellule (Brake et al., 2001). Concernant la mine de Carnoulès, aucune analyse de diversité n’avait encore été réalisée sur la communauté eucaryotes au niveau moléculaire bien que des études précédentes se soient intéressées au protozoaire photosynthétique Euglena mutabilis9 (Halter et al., 2012a, 2012b). L’objectif de ce travail était d’identifier les communautés eucaryotes présentes dans les sédiments du Reigous et d’étudier leur distribution spatiale le long du ruisseau par pyroséquençage 454 des gènes codant pour l’ARNr 18S.

Table 7. Répartition taxonomiques des séquences eucaryotes présentes dans les sédiments de la mine de Carnoulès. Le chiffre entre parenthèses représente l'abondance relative (%) des taxons.

*Calculé par rapport au nombre total de séquences présentes dans cette étude. Autres champignons corresponds aux Blastocladiomycota et à des champignons non classés

Les analyses phylogénétiques ont révélé la présence de 14 taxons essentiellement dominés par 6 groupes (représentant 91 % des séquences totales) affiliés aux phyla des Ascomycètes, Basidiomycètes, Alveolates, Stramenopiles, Streptophytes et Chlorophytes. Parmi ces groupes, les champignons constituaient à eux seuls près de 60 % des séquences et sont apparus être le groupe majoritaire sur l’ensemble des sédiments prélevés (ce qui est en accord avec les résultats obtenus par Baker et al. (2009)), suivis dans une moindre mesure par les Alveolates et les Stramenopiles. La majorité des séquences obtenues dans cette étude se sont révélées être apparentées à des taxons trouvés précédemment dans d’autres DMA tels que les Chlorophytes, les Streptophytes ou les Champignons, etc. (Amaral-Zettler et al., 2002, 2011). Le pyroséquençage a également permis de mettre en évidence de nouveaux taxons non détectés auparavant dans ce type de milieu tels que les Apusozoaires, les Centroheliozoaires et les Jakobides. Ces travaux ont également permis de mettre en évidence une structuration spatiale des communautés eucaryotes qui semble être liée en partie à la physicochimie de l’eau (arsénite, fer et potentiel redox). Ce travail devrait être prochainement soumis à la revue Environmental Microbiology.

Projet MIGRAMD et analyse de biogéographie (France, Bolivie et Espagne) Le projet FRB, MIGRAMD, intitulé : « Microbial biogeography of Acid Mine Drainage: a study of genetic diversity and species diversity from an evolutionary perspective », porté par l’EEM de Pau (B. Lauga) nous a permis d’aborder une nouvelle notion, celle de biogéographie. L’objectif de ce projet était d’évaluer la diversité spécifique des microorganismes présents dans les DMA de 4 pays (Espagne, Portugal, France et Bolivie) plus ou moins riches en arsenic et séparés géographiquement pour mieux comprendre leurs organisations spatiales et leurs répartitions afin d’appréhender les processus qui les mettent en

place. Parallèlement à cette approche spatiale, la dynamique spatiotemporelle des communautés bactériennes a été étudiée dans les eaux le long du continuum du DMA de Carnoulès afin d’identifier les paramètres physicochimiques structurant l’assemblage des bactéries. C’est dans le cadre de cette dernière partie que s’est focalisé le travail de thèse d’Aurélie Volant. La configuration du DMA de Carnoulès est en effet telle que la contamination du Reigous (qui prend sa source au sein du stock de déchet minier) s'atténue le long du continuum, mettant en évidence un important gradient spatial des conditions physicochimiques. De plus, les fortes contraintes physico-chimiques qui s'exercent sur ces écosystèmes donnent l'opportunité d'étudier l'effet des pressions de sélection sur la biodiversité. Les méthodes de séquençages haut-débit, de part la profondeur de séquençage qu’elles permettent semblaient ici un outil de choix pour aborder ces problématiques. Concernant l’approche spatiale, douze DMA (eau et sédiments) ont ainsi été échantillonnés dans 3 régions du monde, l’Amérique du Sud (Bolivie), la péninsule Ibérique (Espagne et Portugal) et la France (DMA du Reigous à Carnoulès). Leurs principaux paramètres physicochimiques (pH, température, conductivité, concentrations en éléments métalliques et sulfates, etc) ont été caractérisés et la spéciation a été réalisée pour As et Fe). Les gènes codant pour l'ARNr 16S des bactéries et des Archaea ont été amplifiés par PCR puis séquencés sur pyroséquençage Roche 454 à la plateforme de génomique de Toulouse. Ce travail, s’est fait également en association avec l’équipe « Instituto de Biologıa Molecular y Biotecnologıa, Universidad Mayor de San Andres » de La Paz en Bolivie. Cette partie de l’étude est en cours d’analyse par l’EEM de Pau.

Concernant l’étude spatio-temporelle de Carnoulès, 6 campagnes de prélèvement ont été réalisées de novembre 2007 à janvier 2010 au niveau de 5 points de prélèvements, soit 30 échantillons au total22. Les paramètres physico-chimiques ont été caractérisés et l’étude à été réalisé en combinant une technique à empreinte moléculaire, la T-RFLP et le pyroséquençage 454 à la plateforme de génomique de Toulouse. Les analyses physico-chimiques ont montré qu’en moyenne 60% des concentrations en sulfate, 96% de celles en fer et 99% de celles en arsenic étaient précipitées le long des 1500 mètres du ruisseau du Reigous. Le pyroséquençage a permis de générer un total de 66016 séquences qui ont permis l’identification de 6801 OTUs incluant 4629 singletons représentant 68% des séquences. Vingt trois phylums bactériens ont été identifiés sur l’ensemble des échantillons analysés et le phylum largement majoritaire (68%) est celui des Protéobactéries. Les 3 OTUs majoritairement présents sont apparentés aux espèces trouvées précédemment comme Gallionella ferruginea, Acidithiobacillus ferrooxidans et Thiobacillus sp. et confortent ainsi les études antérieures. Cette étude a également permis de mettre en évidence des genres jamais identifiés jusqu’à présent à Carnoulès comme Ignavibacterium, Ralstonia ou Paludibacter, etc.

22 Volant A, Bruneel O, Desoeuvre A, Héry M, Casiot C, Bru N, Delpoux S, Fahy A, Javerliat F, Bouchez O, Duran R, Bertin PN, Elbaz-Poulichet F and Lauga B (2014) Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and others drivers along an acid mine drainage. FEMS Microbiology Ecology. 90, 247-263 44

Figure 7. Analyse Canonique par Correspondance (ACC) corrélant la structure des communautés bactériennes avec les paramètres physicochimiques présents dans les différents sites incluant l’arsenic (As), le fer (Fe), la conductivité (Cond), la température (T), l’oxygène dissous (DO), le potentiel redox (Eh), le pH et le sulfate. La structure des communautés correspond à l’abondance des OTUs obtenus à partir des données de T-RFLP (a) ou du pyroséquençage (b). Les principaux clusters ont été entourés. D’après Volant et al., 2014

Une séparation spatiale a également pu être clairement mise en évidence par les données de pyroséquençage et une analyse canonique des correspondances a identifié 3 paramètres physicochimiques (la concentration en arsenic, la température et le potentiel redox) comme des facteurs potentiellement responsables de cette structuration.

Conclusion sur l’étude des drainages miniers acides En conclusion sur cette partie, de nombreuses questions de recherche ont été abordées et les résultats ont permis une amélioration notable des connaissances acquises sur ce site, comme le souligne la vingtaine de publications issues des travaux de notre équipe et de ses partenaires citées dans les paragraphes précédents. Cela démontre également l’intérêt de ces travaux pluridisciplinaires qui ont permis d’obtenir une vision globale et intégrée des processus complexes qui conditionnent les interactions entre les microorganismes et leur environnement. La combinaison d’études de terrain et d’expériences en laboratoire a permis des avancées importantes dans la compréhension des processus d’oxydation et de précipitation de Fe et As et le rôle des microorganismes dans ces processus est désormais mieux compris. L’aptitude de souches, isolées du site, comme Acidithiobacillus ferrooxidans et Thiomonas à oxyder le fer ou l’arsenic a été montré en laboratoire, ainsi que leurs rôles dans la formation de minéraux particuliers. Ces résultats démontrent également qu’en dépit des conditions extrêmes qui règnent à Carnoulès, des communautés microbiennes complexes (Bactéries, Archaea et Eucaryotes) coexistent, interagissent et influencent directement ou indirectement le cycle de

certains métaux et métalloïdes et en particulier l’arsenic. Les résultats obtenus soulignent également l’intérêt de l’utilisation des nouvelles techniques de séquençage permettant de mettre en évidence une diversité majoritairement sous estimée par les techniques classiques de clonage-séquençage mais ouvre aussi la voie à de nombreuses interrogations concernant par exemple le rôle écologique des nombreux taxons rares mis en évidence. Un Projet ANR ECO-TS IngECOST-DMA, « Ingénierie écologique appliquée à la gestion intégrée de stériles et DMA riches en arsenic » a été accepté sur Carnoulès en 2014 pour une durée de 4 ans. Il a pour but la mise en place et l’étude d’un système de bioremédiation (intégrant un système utilisant la capacité des bactéries qui oxydent le fer à précipiter les polluants associé à un système utilisant la capacité des bactéries sulfato-réductrices à former des sulfures de métaux insolubles) qui pourra s’appuyer sur les connaissances déjà acquises sur ce site et qui permet d’aborder ici des études de remédiation appliquées.

TRAVAUX ACTUELS Depuis mon expatriation en février 2012 au sein du Laboratoire de Microbiologie et de Biologie Moléculaire de l’Université Mohammed V de Rabat au Maroc dirigé par le professeur Karim Filali-Maltouf, je m’intéresse à une thématique un peu différente qui est l’étude des interactions plantes-microorganismes dans la mise en place d’un couvert végétal. Ce travail a été initié à l’origine dans le cadre du Laboratoire Mixte International « Biotechnologie Végétale et Microbienne (Directeurs K. Filali-Maltouf et G. Béna) » dans l’axe thématique «Identification de plantes et microorganismes tolérants aux polluants métalliques dans les sites miniers marocains». Ce travail a été réalisé en grande partie dans le cadre d’un projet Ec2co (2012-2013) sur l’«Etude des interactions plantes-microorganismes dans un contexte de réhabilitation de sites minier: mécanismes adaptatifs et effets sur le devenir des polluants métalliques» dont je suis porteur. Comme nous l’avons vu, les activités minières sont très polluantes et ont un impact important sur l'environnement et la santé que ce soit lors de l’extraction du minerai, de sa transformation ou du fait de la production de milliers de tonnes de déchets. Ces déchets sont généralement composés de particules très fines et souvent riches en divers composés toxiques. L'activité minière a été l'un des piliers de l'économie marocaine et a entrainé l’accumulation de milliers de tonnes de résidus pour la plupart abandonnés à l'air libre. Les conditions climatiques du bassin méditerranéen (vents violents et périodes de pluies intenses qui succèdent à des périodes très sèches) favorisent le lessivage et la dissémination des polluants et rendent difficile l’installation d’un couvert végétale. Plusieurs technologies ont été développées pour dépolluer les sols contaminés par les polluants métalliques comme leur extraction par des moyens chimiques ou physiques ou encore l’élimination physique du sol qui est confiné dans des sites d'enfouissement. Mais ces techniques sont souvent très coûteuses à la fois d’un point de vue économique et environnemental et peuvent fortement altérer les qualités physiques, chimiques et biologiques des sols (Glick, 2010). Depuis une quinzaine d’années, des travaux se sont intéressés à l’utilisation de plantes pour dépolluer ces sols. Les polluants peuvent en effet être soit stabilisés dans le sol pour les rendre moins bioassimilables (phytostabilisation), soit être accumulés dans les tissus végétaux (phytoextraction) ou encore être transformés en formes volatiles (phytovolatilisation) (Khan, 2005 ; Kavamura and Esposito, 2010). Le principal obstacle à ces techniques est dû au fait que la plupart des résidus miniers sont de mauvais substrats pour la croissance des plantes en raison à la fois de la présence de métaux toxiques en concentrations élevées, de la présence parfois de pH acides, de la salinité souvent élevée, du manque de matière organique et de nutriments essentiels, de la mauvaise structure du sol et d’une mauvaise rétention de l’eau (Kid et al., 2009 ; de-Bashan et al., 2010). Ces résidus restent généralement dépourvus de couverture végétale pendant des décennies ou plus (Mendez and Maier, 2008 ; de-Bashan et al., 2010). Pour remédier ces environnements extrêmes et en raison des concentrations élevées en polluants métalliques, c’est la phytostabilisation qui est généralement préférée, c'est-à-dire la création d’une couverture végétale qui va limiter l’érosion éolienne et hydrique en stabilisant et précipitant les éléments métalliques au niveau des racines tout en limitant leur accumulation dans les feuilles (de-Bashan et al., 2010, Bolan et al., 2014). Alors que l'établissement d'un couvert végétal sur des sols contaminés par des polluants chimiques reste un défi, les microorganismes peuvent fortement accélérer le processus de

phytostabilisation (de-Bashan et al., 2010, Ma et al., 2011). La croissance des plantes est en effet fortement influencée par les microorganismes qui peuvent intervenir à plusieurs niveaux (fixation d’azote, solubilisation du phosphate, production de phytohormones ou d’antibiotiques, etc.). Certains microorganismes ont de plus la capacité à agir sur la mobilisation/immobilisation des métaux et métalloïdes dans le sol par la production de sidérophores, d’enzymes, ou d’acides organiques, etc. composés qui peuvent modifier ces éléments par acidification, chélation, précipitation, oxydoréduction ou méthylation, etc. (Rajkumar et al., 2012). De nombreux microorganismes sont en effet connus pour favoriser la croissance des plantes (effet PGPB, Plant Growth Promoting Bacteria, de-Bashan et al., 2010, Das et al., 2014). Plusieurs études ont ainsi montré que des rhizobactéries appartenant aux genres Achromobacter, Arthrobacter, Azotobacter, Bacillus, Pseudomonas, ou Serratia favorisaient la croissance des plantes dans des environnements contaminés par des métaux et métalloïdes (Ma et al., 2011). Parmi ces microorganismes, on peut distinguer les microorganismes telluriques, présents dans le sol; les rhizobactéries localisées à proximité immédiate des racines et les bactéries endophytes qui colonisent les tissus internes des plantes sans causer d’infections.

Contrairement à la pollution par les mines dans les régions tempérées, il n’existe que très peu d'études sur l'impact environnemental des activités minières dans les régions arides et semi-arides (González et al., 2011). Les sites que nous étudions dans le cadre de ce projet se situent dans le district minier de la ville d’Oujda, au Nord-Est du Maroc près de la frontière algérienne.

Figure 8. Situation géographique des régions étudiées et localisation des stations de prélèvements. D’après Smouni et al., 2010

Les sites d’études comprennent les digues de lavage des mines abandonnées de Pb et Zn de Touissit et Boubker ainsi que les scories de l’ancienne fonderie de Oued El Heimer. Ces déchets qui représentent plusieurs millions de tonnes constituent des digues de très grande superficie.

Figure 9. Description des sites de Oued El Heimer, Touissit et Boubker. A) Scories plombifères déposées sur de grandes surfaces aux abords de la fonderie. B) Digue de sable aux abords du village de Touissit. C) Digue de sable à proximité d’un champ de blé dans la région de Boubker. D’après Smouni et al., 2010

Malgré un climat austère et une forte teneur en polluants, une flore tolérante parvient à s’y développer. Ces plantes, ainsi que les microorganismes associés sont à priori adaptés aux conditions édapho-climatiques de ces régions et présentent donc une ressource pour le développement de stratégies de réhabilitation, notamment par la phytostabilisation et la mise en place d’un couvert végétal qui limiterait l’érosion éolienne et hydrique. L’intérêt pour cette région est multiple. D’une part, le périmètre étudié est fortement impacté par une pollution polymétallique aussi bien au niveau des terres agricoles que des cours d’eau et des puits (Smouni et al., 2010). L’index de pollution des échantillons prélevés dans ces environnements est généralement très élevé du fait de la présence simultanée de plusieurs polluants (As, Cd, Cu, Ni, Pb et Zn) avec de très importantes teneurs en Pb, Zn et As (respectivement jusque 7 g/kg, 2 g/kg et 187 mg/kg) (Smouni et al., 2010). Enfin, la présence de sites divers et originaux (stériles couverts et digues nues, revégétalisés ou non, de façon naturelle ou par action humaine) permet de comparer leurs impacts sur l’association plantes-

microorganismes qui les a colonisés sur une échelle de temps variée. Cette situation constitue ainsi une source d’une incroyable diversité floristique et microbienne.

L’objectif de ce projet de recherche consiste à étudier à la fois les plantes et les microorganismes associés et doit permettre : (i) d’étudier les mécanismes de résistance et d’accumulation notamment pour le Pb et le Zn de 2 plantes endémiques (une plante hyperaccumulatrice de plomb, Hirschfeldia incana et une légumineuse Hedysarum spinosissimum) (ii) d’identifier les communautés microbiennes capables de se développer sur ces différents environnements et de mieux comprendre leurs mécanismes de résistance et d’adaptation et enfin (iii) d’isoler de nouveaux microorganismes à la fois rhizosphériques et symbiotiques (présents dans les nodules de H spinosissimum) et de mieux appréhender leurs mécanismes d’action sur la croissance des plantes et la mobilisation/immobilisation des métaux et métalloïdes. Cette étude combinée des plantes et des microorganismes devrait permettre à terme de proposer une collection de plantes et de microorganismes résistants à ces polluants et susceptibles d’être des outils efficaces pour établir un programme de phytoremédiation. Elle pourrait ainsi avoir un impact sociétal important en accélérant significativement les processus de réhabilitation de ces zones contaminées.

C’est un travail pluridisciplinaire qui associe à la fois des végétalistes, des microbiologistes et une géochimiste (P. Moulin, Ingénieur IRD, US IMAGO). Il a été réalisé en collaborations avec des laboratoires marocains : le Laboratoire de Microbiologie et Biologie Moléculaire (LMBM, L. Sbabou et J. Aurag) dans lequel s’effectue actuellement mon expatriation ainsi que le Laboratoire de Physiologie et Biotechnologie Végétale (LPBV, A. Smouni, M. Fahr) de l’Université de Rabat. Ce projet comprend également l’implication de partenaires français: le Laboratoire des Symbioses Tropicales et Méditerranéennes (AMPERE-LSTM, E. Navarro) ainsi que le laboratoire de Biochimie et Physiologie Moléculaire des Plantes de Montpellier (BPMP, P. Doumas, F. Auguy). Ce travail s’inscrit également dans le cadre du réseau SICMED «Environnements Miniers Méditerranéens » coordonné par P. Doumas (BPMP, Montpellier).

Dans le cadre de ce projet de recherche, mon travail a essentiellement porté sur l’étude de la diversité des microorganismes et sur leur rôle dans le transfert des métaux et métalloïdes dans ces environnements extrêmes. Ce travail est réalisé dans le cadre de la thèse de I. Dahmani qui a débutée en décembre 2013 et que je coordonne avec 2 encadrants marocains, J. Aurag, L. Sbabou ainsi que E. Navarro. Ces travaux comprennent l’utilisation de nouvelles techniques de séquençage associant études taxonomiques et études fonctionnelles. L’analyse des données de pyroséquençage réalisée sur 24 sols miniers en triplicat (associant sols nus et sols rhizosphériques) est en cours par bioinformatique à l’aide du logiciel Mothur (http://www.mothur.org/wiki). Cette technique de pyroséquençage haut débit, permet d'appréhender de manière la plus exhaustive possible la diversité microbienne globale et nous permet également de pouvoir accéder à la biosphère “rare” qui semble jouer un rôle important dans l’adaptation à ces environnements pollués.

L’étude fonctionnelle par métagénomique (travail en collaboration avec E. Navarro, AMPERE-LSTM) nous permettra d’étudier la résistance des microorganismes et les mécanismes importants impliqués dans de tels écosystèmes, comme la fixation du CO2 atmosphérique ou le métabolisme de l’azote, etc. La comparaison des métagénomes issus de divers environnements permettra une analyse de type fonctionnel (présence/absence/diversité de gènes impliqués dans les fonctions du sol). Les échantillons de l’analyse métagénomique ont été choisis parmi les 24 échantillons utilisés pour le pyroséquençage et comprennent l’étude des sols qui font l’objet d’isolements pour l’étude des bactéries PGPB (travail effectué par des membres de l’équipe LMBM) afin de pouvoir comparer l’activité des bactéries isolées à celles de l’analyse fonctionnelle.

L’analyse des données du pyroséquençage a commencé a donné ses premiers résultats. Un des soucis rencontré lors de cette étude a été la génération d’un fichier conséquent (près de 5 Go pour le fichier ssh) qui n’a pas permis de réaliser l’ensemble des opérations sur ordinateur et il a fallu l’utilisation d’un serveur à distance au niveau du laboratoire de Montpellier pour finaliser les analyses. Les séquences brutes générées par la technique 454 GS-FLX Titanium auprès de la plateforme de séquence MR DNA (Molecular Research LP, Texas, EU, http://mrdnalab.com) ont été analysées en utilisant la version 1.33.2 du logiciel mothur, (http://www.mothur.org) (Schloss et al., 2009). Ces séquences ont été traitées par la commande "shhh.flows" en utilisant l'algorithme PyroNoise (Quince et al. 2009, 2011). Le prétraitement des séquences non alignées a inclus la suppression des codes barres, des deux amorces, de toutes les séquences ambigües (contenant au moins un nucléotide «N», ainsi que toutes celles qui contenaient plus de 8 homopolymères). Les séquences identiques (100%) ont ensuite été regroupées pour accélérer le traitement des données et les séquences représentatives ont été alignées sur la base de données de référence SILVA (bactéries et archées) en utilisant l'algorithme de Needleman-Wunsch (Needleman & Wunsch, 1970). Les séquences mal alignées ont ensuite été éliminées. Une autre étape de criblage (pré-cluster) a été appliquée pour réduire les erreurs dues au pyroséquençage, par regroupement des séquences qui ne présentent qu’une base de différence sur 100 pb par rapport à une séquence de référence présente en plus grand nombre dans le groupe (Huse et al., 2010). Les séquences chimériques ont été détectées et supprimées en utilisant le programme Uchime Chimera (Edgar et al., 2011) et les séquences d’Archaea ou les organites des organismes eucaryotes comme les chloroplastes ont également été retirés de l'ensemble de données. Ces analyses sont en cours et je ne m’étendrai donc pas trop dessus. Les premiers résultats montrent que l’étude des 72 échantillons de sols (24 sites en triplicat) a permis d’obtenir un total de 1545801 séquences brutes ayant une longueur moyenne d’environ 400 pb. Après le nettoyage et l’ensemble des traitements, 743501 séquences de bonne qualité (d’environ 172 pb) ont été récupérées. Après normalisation, 101 016 séquences correspondant à 6966 OTUS dont 3640 OTUs rares (représentant 52% des séquences) ont pu être identifiés.

Table 8. Estimation de la richesse en OTUs, des indices de diversité et de la couverture estimée pour les 5 échantillons de sédiments. Les résultats sont présentés pour les données normalisées, rééchantillonnées au hasard pour avoir une taille d’échantillon égale entre les sites.

Sampling stations N° reads N° of N° of Singletons Good's Chao1 Richness Shannon after quality normalized OTUa coverageb diversityc filtering reads MINE DE BOUBKER OMF12BoGrRh1a 16966 1403 428 62 82% 934 (785; 1146) 5.14 (5.06; 5.23) OMF12ToGrRh1b 15417 1403 397 40 85% 715 (617; 857) 5.06 (4.98; 5.15) OMF12ToGrRh1c 16215 1403 433 62 82% 855 (732; 1028) 5.18 (5.09; 5.26) OMF12BoGrRh3a 16543 1403 396 43 84% 769 (655; 933) 4.92 (4.83; 5.02) OMF12BoGrRh3b 9265 1403 362 33 87% 617 (535; 737) 4.86 (4.77; 4.95) OMF12BoGrRh3c 6182 1403 368 25 87% 623 (540; 745) 4.99 (4.90; 5.07) OMF12BoGrNu4a 5796 1403 132 13 98% 178 (151; 240) 4.23 (4.17; 4.28) OMF12BoGrNu4b 10148 1403 147 18 97% 198 (169; 266) 4.36 (4.31; 4.42) OMF12BoGrNu4c 5781 1403 123 13 98% 170 (141; 248) 4.27 (4.22; 4.32) OMF12BoGHRh5a 15933 1403 363 43 85% 663 (569;800) 4.77 (4.68; 4.86) OMF12BoGHRh5b 17654 1403 390 50 84% 761 (648;923) 4.92 (4.84; 5.01) OMF12BoGHRh5c 6858 1403 362 31 86% 624 (541; 746) 4.83 (4.74; 4.91) OMF12BoGHNu6a 1479 1403 10 1 100% 12 (10; 25) 1.39 (1.36; 1.43) OMF12BoGHNu6b 2548 1403 27 2 100% 33 (28;66) 2.79 (2.75; 2.83) OMF12BoGHNu6c 1403 1403 26 10 99% 92 (48; 223) 2.20 (2.15; 2.24) OMF12BoHeRh13a 12336 1403 291 27 88% 697 (550; 927) 4.56 (4.48; 4.64) OMF12BoHeRh13b 9130 1403 280 32 89% 514 (431; 642) 4.42 (4.33; 4.51) OMF12BoHeRh13c 16533 1403 273 26 89% 526 (434; 669) 4.40 (4.31; 4.49) OMF12BoHeRh15a 16622 1403 336 45 86% 760 (618;973) 4.70 (4.61; 4.78 OMF12BoHeRh15b 4431 1403 326 36 88% 555 (477;673) 4.75 (4.66; 4.83) OMF12BoHeRh15c 8442 1403 322 32 88% 534 (463; 642) 4.66 (4.57; 4.74) MINE DE TOUISSIT OMF12ToGrRh17a 13495 1403 426 39 82% 862 (734; 1043) 5.04 (4.95; 5.13) OMF12ToGrRh17b 9853 1403 443 48 82% 901 (768; 1088) 5.10 (5.01; 5.19) OMF12ToGrRh17c 5823 1403 407 35 84% 760 (653; 913) 4.96 (4.87; 5.06) OMF12ToGrNu18a 14943 1403 213 39 92% 396 (323; 517) 3.93 (3.84; 4.01) OMF12ToGrNu18b 15623 1403 233 42 91% 441 (361; 573) 4.01 (3.92; 4.10) OMF12ToGrNu18c 5824 1403 211 28 93% 322 (277; 399) 4.00 (3.92; 4.09) OMF12ToGrRh19a 11327 1403 444 50 82% 815 (708; 965) 5.12 (5.03; 5.21) OMF12ToGrRh19b 10389 1403 408 36 85% 728 (629; 870) 5.12 (5.03; 5.21) OMF12ToGrRh19c 9623 1403 429 46 84% 740 (647; 873) 5.20 (5.12; 5.29) OMF12ToGrNu20a 11811 1403 307 45 88% 641 (526; 818) 4.48 (4.39; 4.57) OMF12ToGrNu20b 11922 1403 310 43 88% 656 (536; 839) 4.48 (4.39; 4.57) OMF12ToGrNu20c 8779 1403 309 51 88% 595 (497; 746) 4.48 (4.38; 4.57) OMF12ToHeRh21a 13553 1403 465 75 77% 1312 (1071; 1650) 4.92 (4.82; 5.02) OMF12ToHeRh21b 11746 1403 449 80 80% 1019 (857; 1246) 4.92 (4.82; 5.02) OMF12ToHeRh21c 11950 1403 444 73 81% 864 (744; 1032) 4.99 (4.90; 5.09) OMF12ToHeNu22a 5971 1403 449 49 81% 863 (745; 1028) 5.20 (5.12; 5.29) OMF12ToHeNu22b 15394 1403 440 54 81% 1026 (853; 1271) 5.11 (5.02; 5.19) OMF12ToHeNu22c 12936 1403 443 56 81% 954 (807; 1162) 5.11 (5.02; 5.20) OMF12ToHeRh23a 4035 1403 532 58 78% 1067 (923; 1264) 5.62 (5.55; 5.69) OMF12ToHeRh23b 9998 1403 552 70 78% 967 (858; 1115) 5.72 (5.65; 5.79) OMF12ToHeRh23c 7476 1403 552 61 78% 975 (864; 1126) 5.69 (5.62; 5.76) OMF12ToGrRh25a 1775 1403 385 28 86% 681 (585; 821) 4.98 (4.89; 5.07) OMF12ToGrRh25b 3785 1403 375 28 87% 576 (512; 669) 4.87 (4.78; 4.96)

OMF12ToGrRh25c 3707 1403 414 44 84% 773 (665; 929) 5.09 (5.01; 5.18) OMF12ToGrNu26a 11614 1403 448 53 82% 931 (787; 1136) 5.28 (5.20; 5.36) OMF12ToGrNu26b 10035 1403 434 47 83% 750 (658; 881) 5.18 (5.10; 5.27) OMF12ToGrNu26c 8394 1403 430 50 83% 924 (774; 1140) 5.14 (5.06; 5.23) OMF12ToGrRh27a 13447 1403 445 74 82% 792 (693; 929) 5.08 (4.99; 5.17) OMF12ToGrRh27b 6646 1403 415 51 84% 697 (613; 816) 5.02 (4.93; 5.11) OMF12ToGrRh27c 19157 1403 458 62 81% 905 (779; 1081) 5.13 (5.04; 5.22) SITE DE OUED EL HEIMER OMF12OhGrRh29a 18557 1403 243 26 91% 457 (374; 594) 4.40 (4.33; 4.48) OMF12OhGrRh29b 6486 1403 270 25 89% 523 (431; 666) 4.49 (4.41; 5.57) OMF12OhGrRh29c 14358 1403 292 27 88% 576 (477; 728) 4.56 (4.48; 4.64) OMF12OhGrNu30a 6869 1403 106 4 96% 212 (157; 327) 2.43 (2.33; 2.53) OMF12OhGrNu30b 7341 1403 174 27 93% 368 (285; 513) 2.86 (2.75; 2.98) OMF12OhGrNu30c 5523. 1403 112 13 96% 245 (175; 392) 2.25 (2.13; 2.36) OMF12OhGrRh31a 7651 1403 541 105 74% 1350 (1136; 1641) 5.31 (5.22; 5.41) OMF12OhGrRh31b 12625 1403 565 119 73% 1370 (1164; 1646) 5.42 (5.33; 5.51) OMF12OhGrRh31c 9757 1403 556 108 74% 1345 (1141; 1619) 5.37 (5.27; 5.46) OMF12OhHeRh33a 14507 1403 586 110 73% 1377 (1177; 1644) 5.68 (5.60; 5.76) OMF12OhHeRh33b 19202 1403 589 113 72% 1363 (1169; 1621) 5.66 (5.58; 5.74) OMF12OhHeRh33c 19009 1403 572 107 74% 1371 (1164; 1651) 5.64 (5.56; 5.72) OMF12OhHeNu34a 8970 1403 475 67 80% 1007 (856; 1217) 5.38 (5.30; 5.45) OMF12OhHeNu34b 9217 1403 485 86 79% 1003 (860; 1201) 5.37 (5.29; 5.45) OMF12OhHeNu34c 16679 1403 510 91 78% 1053 (905; 1255) 5.49 (5.41; 5.57) OMF12OhHeRh35a 5205 1403 492 68 80% 869 (766; 1009) 5.44 (5.36; 5.51) OMF12OhHeRh35b 16874 1403 536 100 75% 1238 (1056; 1485) 5.46 (5.38; 5.54) OMF12OhHeRh35c 16636 1403 527 126 75% 1260 (1065; 1525) 5.49 (5.41; 5.57) OMF12OhHeNu36a 1904 1403 404 57 86% 637 (565; 742) 5.34 (5.27; 5.41) OMF12OhHeNu36b 4089 1403 464 63 82% 838 (732; 986) 5.45 (5.38; 5.52) OMF12OhHeNu36c 5329 1403 425 39 85% 661 (590; 762) 5.35 (5.28; 5.42 a OTUs définis avec un seuil de 97% de similarités entre les séquences b Somme des probabilités des classes observées calculées selon (1 - (n/N)), où n représente le nombre de singleton et N est le nombre total de séquences c Prend en compte le nombre et la régularité des espèces Les valeurs entre parenthèses représentent les intervalles de confiance à 95%

L’indice de diversité de Shannon varie fortement entre les échantillons et est compris entre 1.39 et 5.72 et le taux de couverture est compris entre 72 et 100%.

Figure 10. Composition des différents phylums basés sur la classification des séquences des gènes codants pour les ARNr16S des bactéries présents dans les 3 sites étudiés. L'affiliation des séquences est basée sur la classification RDP.

Soixante dix-sept pourcent des séquences ont pu être classées au niveau du phylum et 17 phylums ont pu être identifiés. Les Actinobacteria et les Protéobactéria sont les phylums les plus importants représentant respectivement environ 35 et 25% de l’ensemble des séquences suivi par les Bactéroidetes (7%), les Acidobactéria (5%), les Gemmatimonadetes (3%) et le groupe TM7 (environ 2%). Le reste des phylums représente moins de 1%. Les 5 genres les plus abondants sont Sphingomonas, Arthrobacter, Rubrobacter, Nocardioides et Pseudonocardia. Les genres Sphingomonas et Arthrobacter ont déjà été détectés dans des stériles miniers (Schippers et al., 2010, Chen et al., 2013). Le genre Sphingomonas est souvent retrouvé dans la phyllosphere de différentes plantes et pourrait être impliqué dans la protection de plantes comme Arabidopsis thaliana contre certains pathogènes (Innerebner et al., 2011). Plusieurs études ont également montrées l’effet bénéfique du genre Arthrobacter sur la croissance des plantes dans des environnements contaminés par des métaux et métalloïdes (Ma et al., 2011). Un important travail d'analyses statistiques et bioinformatiques est maintenant nécessaire afin de décrire la diversité spécifique de chaque site, de comparer les assemblages microbiens entre les sites, de révéler les relations inter-spécifiques et d’identifier les facteurs environnementaux déterminant la composition de la communauté.

VI. PROJET DE RECHERCHE Ces thématiques sur lesquelles j'ai acquis aujourd'hui certaines connaissances vont définir le cadre de mes recherches pour les années à venir.

Je souhaite développer 3 grands volets dans le cadre de l’étude sur les environnements miniers: - Caractériser la diversité taxonomique et fonctionnelle en utilisant les nouvelles technologies de séquençages associées à des méthodes tels que la métagénomique ou la métatranscriptomique afin d’explorer la structure et la dynamique des communautés microbiennes (Bactéries, Archaea et Eucaryotes) et obtenir ainsi une vue d'ensemble de leurs potentiels métaboliques dans le but de mieux comprendre le fonctionnement de ces écosystèmes particuliers, - Continuer l’étude des interactions plantes-microorganismes initiée lors de mon expatriation qui permettent d’aider à la mise en place de solutions pratiques dans le cas de la mise en place d’un couvert végétal, - Poursuivre l’isolement de microorganismes afin de pouvoir les étudier en laboratoire ce qui permet d’avoir une meilleure connaissance de leurs réelles capacités métaboliques.

Caractérisation de la diversité taxonomique et fonctionnelle Ces dernières années, les nouvelles technologies de séquençage ont fortement accélérées les recherches en biologie et microbiologie et ont permis la production de très grands volumes de séquences en diminuant fortement les prix, comparées aux méthodes de séquençage traditionnelles (Knief et al., 2014). Ces développements récents permettent maintenant de répondre à des questions qui n’étaient pas concevables il y a seulement quelques années en raison essentiellement des limitations techniques et financières comme par exemple : qu’elles sont les communautés microbiennes présentes, que font elles, comment arrivent t’elles à se développer dans ces environnements, comment évoluent t’elles en fonction des perturbations et des changements de leurs environnements, comment interagissent elles entre elles et avec leur environnement, comment peuvent elles affecter le développement des plantes (Knief et al., 2014)? Le recours aux disciplines « méta-omiques », permet en effet d’avoir une meilleure compréhension de l'écologie microbienne. La métagénomique est par exemple devenue une des disciplines scientifiques les plus actives. Cette approche permet désormais l’analyse de communautés microbiennes qui semblaient largement hors de portée il y a encore quelques années comme les organismes non cultivés et permet d'obtenir une vue d'ensemble du potentiel métabolique des communautés présentes. Des études comme la métatranscriptomique basées sur l'expression des gènes sont également très intéressantes pour apporter de nouvelles connaissances sur la dynamique fonctionnelle des communautés microbiennes, mieux identifier les facteurs environnements qui régulent leurs activités et mettre en exergue les fonctions phares essentielles pour le fonctionnement de la communauté (Gifford et al., 2011 ; Carvalhais et al., 2012). Cependant, il faut malgré tout bien garder à l’esprit que, comme la plupart des techniques de biologie, ces technologies ne sont pas non plus exempts de biais. Concernant les études de diversité par les techniques de séquençage à haut débit (pyroséquençage ou technologie

Illumina, etc.), il est à noter par exemple la faible taille des amplicons générés pour l’instant ainsi que le taux important d’erreurs de séquençage, que l’utilisation de nouveaux algorithmes tente de corriger (Huse et al., 2010 ; Knief, 2014). Il reste également les biais inhérents à la biologie moléculaires comme ceux au niveau de l’extraction de l’ADN ou ceux générés lors de l’étape de PCR. L’utilisation des séquenceurs de 3ème générations permettront peut être de limiter ce dernier biais dans les prochaines années. Concernant la métagénomique par exemple, même un séquençage « en profondeur » d’un environnement ne permet d’accéder qu’à une petite fraction de la variabilité génétique réellement présente en identifiant principalement les membres les plus abondants (Gilbert and Dupont, 2011).

Je souhaite, dans la poursuite de mes travaux, continuer à m’investir dans l’étude des drainages miniers acides. Comme déjà présenté précédemment, les conditions de vie extrêmes de ces environnements (pH acides, concentrations élevés en métaux et métalloïdes toxiques qui diffèrent d’un site à un autre) et les communautés simplifiées qui les caractérisent permettent d’aborder un certain nombre de questions fondamentales et contribuent à mieux comprendre la structure des communautés microbiennes et leurs profils de diversité. Mais c’est surtout l’aspect appliqué qui m’intéresse en raison de l’implication de ces organismes dans les processus de génération et/ou de remédiation de ces DMA qui peuvent avoir des applications concrètes dans de nombreux pays du Sud, en accord avec les objectifs de l’IRD. Une meilleure caractérisation de la diversité génétique mais aussi fonctionnelle de ces communautés microbiennes ainsi que l’étude de leurs interactions entre elles et avec leur environnement est en effet une étape essentielle à la compréhension du fonctionnement de ces écosystèmes pour pouvoir développer à terme des stratégies pour remédier à ces pollutions. Pour ce faire, il est également indispensable de prendre en compte l’ensemble de la communauté microbienne comprenant les organismes procaryotes et eucaryotes car, jusqu’à présent, la majorité des études réalisées sur ces écosystèmes se sont focalisées sur les bactéries.

L’ancien site minier de Carnoulès, par ses caractéristiques comme les concentrations exceptionnelles en As ainsi que par la présence d’un gradient spatial de pollution résultant de processus naturels de remédiation demeurera un site d’étude important pour mes travaux, notamment dans le cadre de projets comme l’ANR ECO-TS IngECOST-DMA ou au travers de l’observatoire OSU OREME.

Les projets de recherche que je souhaite développer au cours des prochaines années concerneront surtout les pays du Sud avec par exemple le Maroc, en continuation avec les travaux lancés depuis 3 ans maintenant dans le cadre de mon expatriation et pays avec lequel je souhaite poursuivre les collaborations dans le futur. L’étude des DMA va se faire par exemple dans le cadre d’un projet débuté récemment concernant les drainages de mine non pérennes présents dans un nouveau chantier, la mine de Kettara au Maroc qui se fait en association avec l’équipe de recherche E2G (R. Hakkou) de la Faculté des Sciences et Techniques de Guéliz à Marrakech. La mine de Kettara, située à environ 30 km au nord-est de Marrakech a été exploitée pour sa pyrrhotite de 1964 à 1981 et a produit 5.2 Mt de pyrrhotite concentré contenant une

moyenne de 29% en poids de sulfures ce qui a généré d’importants stériles miniers répartis, sans aucune protection, sur de très grande surfaces autour de la mine (Lghoul et al., 2014). C’est l’une des mines qui pose le plus de problèmes autour de Marrakech en raison notamment de la présence de ces DMA qui se forment à chaque pluie importante et qui expose directement la population du village minier de Kettara avoisinant, qui comprends environ 2000 habitants. Le climat est semi aride avec une moyenne de 250 mm de pluies annuelles qui surviennent généralement sur de courtes périodes et avec une forte intensité.

Figure 10. Présentation de la mine de Kettara: (a) localisation, (b) effluent de DMA, and (c) minéraux secondaires, (d) vue panoramique. D’après Lghoul et al. 2014

Cette mine a été extensivement étudiée comme en atteste les nombreuses publications depuis quelques années (Hakkou et al., 2008a, 2008b, 2009 ; Lghoul et al., 2014 et références citées). Ces travaux ont été réalisés en grande partie dans le cadre d’une chaire de recherche IDRC maroco-canadienne, entre l’équipe de recherche E2G (R. Hakkou) de la Faculté des Sciences et Techniques de Marrakech et l’Institut de recherche sur les mines et l’environnement (UQAT) situé à Québec au Canada.

Cette étude microbiologique, se fait en partenariat avec l’équipe de recherche E2G à Marrakech dans le cadre du Master 2 de N. Mghazli qui vient de débuter (co-encadrement avec L. Sbabou du LMBM de Rabat) et qui a pour but d’identifier les communautés procaryotes (Bactéries et Archaea) présentes dans ces déchets miniers pour tenter d’identifier les communautés responsables de la génération de ces drainages miniers acides. Les prélèvements ont été réalisés sur 9 points répartis sur l’ensemble du site et une analyse par séquençage Illumina est en cours.

Un autre aspect de la mobilisation des métaux et métalloïdes concerne le cas particulier de l’As qui fait actuellement peser une lourde menace sur la santé de nombreuses personnes à travers le monde et ce, principalement dans les pays du Sud, où plusieurs millions de personnes consomment des eaux de boisson contaminées par ce toxique (Nordstrom, 2000 ; Jiang et al., 2013). Comme nous l’avons vu, les microorganismes sont fortement impliqués dans les processus de transfert de ce polluant dans l’environnement. Des travaux sur l’étude des transferts de l’As (d’origine minière ou géogénique, présent naturellement dans la roche) vers le milieu aquatique et les impacts sanitaires associés sont traités dans le cadre du LMI Picass-Eau (« Prédire l’Impact du Climat et des usAges sur les reSSources en Eau en Afrique SUbsaharienne"). Ce projet prévoit notamment d’aborder la question de la mobilisation de l’arsenic vers la ressource en eau sur le bassin du Nakambé au Burkina Faso. Il s’agit de déterminer les facteurs qui influencent la variabilité spatiale de l’arsenic (facteurs géologiques, hydrogéologiques, physico-chimiques et microbiologiques) dans les aquifères de la région de Ouahigouya dans le Nord du Burkina Faso et d’étudier l’implication des microorganismes dans ces systèmes et l’impact sur la santé des populations exposés. Ces travaux de recherche seront développés dans le cadre de l’ANR BALWASA (Basement aquifers for a local water service in Africa), si elle est acceptée. Cet ANR a été déposé cette année par P. Genthon, un hydrogéologue de HydroSciences. Ces travaux se feront plus spécifiquement dans le cadre du Work Package n° 3 intitulé « Arsenic contamination and health near Ouahigouya » en collaboration notamment avec F. Lalanne de l’Institut International d’Ingénierie de l’Eau et de l’Environnement et de P. Genthon.

Etude des interactions plantes–microorganismes dans un contexte de phytoremediation et de réhabilitation des environnements miniers au Maroc Le travail initié dans le cadre du projet Ec2co va se poursuivre dans les prochains mois avec un volet important concernant l’analyse métagénomique des populations bactériennes présentes dans ces environnements et se fera dans le cadre de la thèse de Ikram Dahmani en collaboration avec Isabelle Navarro (AMPERE, Lyon-LSTM, Montpellier). Cette nouvelle approche permettra de mieux comprendre comment les communautés bactériennes s’adaptent à leur environnement et interagissent avec lui (identification des gènes de résistance ou d’oxydation à l’arsenic, etc.) et comment elles peuvent affecter le développement des plantes (fixation d’azote par exemple, etc.).

D’autres études sur la revégétalisation pourront également se mettre en place, par exemple sur la mine de Kettara, dans le cadre d’un projet de recouvrement de ces déchets utilisant des déchets de mines de phosphates, basiques, qui est à l’étude actuellement (Lghoul et al. 2014) afin de limiter les infiltrations d’eau et la formation de DMA. Si ce revêtement est mis en place sur l’ensemble du site, ce qui est prévu dans le cadre de la chaire, une couverture végétale devra être apportée sur le long terme. Des études de diversité et de métagénomiques permettraient d’identifier les communautés de microorganismes présentes et de mieux comprendre leurs interactions avec les plantes. Ce travail est à combiner également avec l’isolement de souches bactériennes pour étudier leurs activités bénéfiques sur les plantes naturellement présentes (solubilisation du phosphate, production de sidérophores ou d’auxine,

fixation d’N, etc.) afin de pouvoir proposer à terme une collection de plantes et de microorganismes résistants à ces polluants et susceptibles d’être des outils efficaces pour établir un programme de phytoremédiation.

Isolement de microorganismes et étude en laboratoire de leurs capacités métaboliques Bien que les méthodes moléculaires apportent des informations indispensables du fait qu’un faible pourcentage de microorganismes de l’environnement peuvent actuellement être cultivés en laboratoire, les méthodes culturales restent indispensables pour une meilleure connaissance des organismes et de leurs interactions réelles avec l’environnement en utilisant des études physiologiques. Des études d’isolement réalisées à Carnoulès sur les sédiments du Reigous (Delavat et al., 2012), ont par exemple souligné l’importance du maintient des méthodes culturales pour l’identification précise et la compréhension du rôle fonctionnel des microorganismes dans leurs environnements. Les études réalisées sur les souches de Thiomonas et d’Aciditiobacillus ferrooxidans ont également bien montré l’intérêt de ces travaux pour la compréhension de leur rôle réel dans l’environnement et le système de remédiation présent à Carnoulès. De plus, les méthodes culturales permettent de contourner certains biais inhérents aux approches moléculaires comme la résistance de certaines bactéries à la lyse cellulaire ou bien la difficulté à détecter les microorganismes appartenant à la biosphère rare.

Dans le cadre d’une collaboration avec le laboratoire GMGM de Starsboug (P. Bertin), un projet Ec2co devrait être soumis cette année qui va spécifiquement s’intéresser, entre autre, à l’étude des microorganismes difficiles à cultiver. Ces travaux vont se focaliser plus spécifiquement sur des organismes comme Gallionella ferruginea que nous n’avons pas pu isoler pour l’instant malgré de nombreux essais ou encore pour tenter de cultiver le pseudogénome CARN1 qui semble avoir un rôle important au sein de l’écosystème et qui est retrouvé en assez grand nombre dans l’eau et les sédiments et détecté depuis l’utilisation des méthodes de séquençage à haut débit. Ce projet à pour but de trier par cytométrie les microorganismes selon des critères taxonomiques et/ou fonctionnels avant d’en séquencer le génome afin d’étudier le métabolisme de ces microorganismes ; d’isoler par des approches de culture in situ des populations non cultivées et enfin de déterminer la dynamique et l'activité des populations microbiennes en fonction des variations contrôlées des paramètres physicochimiques.

VII REFERENCES BIBLIOGRAPHIQUES

Alkaaby A, Leblanc M, Perissol M (1985) Minéralisation diagénétique précoce (Pb-Zn- Ba) dans un environnement détritique continental : cas tu Trias de Carnoulès (Gard. France), C.R. Acad. Sci. Paris 300 pages, pp 919-922 Amaral-Zettler LA, Gomez F, Zettler, Keenan BG, Amils R and Sogin ML (2002) Microbiology: Eukaryotic diversity in Spain's river of fire. Nature. 417, 137-137 Amaral-Zettler LA, Zettler ER, Theroux SM, Palacios C, Aguilera A and Amils R (2011) Microbial community structure across the tree of life in the extreme Río Tinto. 496 ISME J. 5, 42-50 Amils R, Gonzalez-Toril E, Fernandez-Remolar D, Gomez F, Aguilera A, Rodriguez N, Malki M, Garcıa-Moyano A, Fairen AG, de la Fuente V, Sanz, JL (2007). Extreme environments as Mars terrestrial analogs: the Rio Tinto case. Planetary and Space Science. 55, 370-381 Argos M, Kalra T, Rathouz PJ, Chen Y, Pierce B, Parvez F, et al. (2010) Arsenic exposure from drinking water, and all-cause and chronic-disease mortalities in Bangladesh (HEALS): a prospective cohort study. The Lancet. 376, 252-258 Arsène-Ploetze F, Koechler S, Marchal M, Coppee J-Y, Chandler M, Bonnefoy V, Barakat M, Barbe V, Battaglia-Brunet F, Brochier-Armanet C, Bruneel O, G. Bryan C, Cleiss J, Heinrich-Salmeron A., Hommais F., Joulian C., Krin E., Lieutaud A., Lièvremont D., Michel C, Muller D, Ortet P, Proux C, Siguier P, Roche D, Rouy Z, Salvignol G, Slyemi D, Talla E, Weiss S, Weissenbach J, Médigue C, Bertin PN (2010) Structure, function and evolution of the Thiomonas spp. genome inferred from sequencing and comparative analysis. PLoS Genetics. 6: (2) e1000859 Bachy C and Worden AZ (2014) Microbial Ecology: Finding Structure in the Rare Biosphere. Current Biology. 24, R315-R317 Bahar MM, Megharaj M, and Naidu R (2013) Bioremediation of arsenic-contaminated water: recent advances and future prospects. Water Air Soil Pollut. 224, 1722 Baker BJ and Banfield JF (2003) Microbial communities in acid mine drainage. FEMS Microbiol. Ecol. 44, 139-152 Baker BJ, Lutz MA, Dawson SC, Bond PL and Banfield JF (2004). Metabolically active eukaryotic communities in extremely acidic mine drainage. Appl Environ Microbiol. 70, 6264-6271 Baker BJ, Tyson G.W, Goosherst L and Banfield JF (2009) Insights into the diversity of eukaryotes in acid mine drainage biofilm communities. Appl Environ Microbiol. 75: 2192- 2199 Behnke A, Engel M, Christen R, Nebel M, Klein RR, and Stoeck T (2011). Depicting more accurate pictures of protistan community complexity using pyrosequencing of hypervariable SSU rRNA gene regions. Environmental Microbiology. 13, 340-349 Bertin PN, Heinrich-Salmeron A, Pelletier E, Goulhen- Chollet F, Arsène-Ploetze F, Gallien S, Lauga B, Casiot C, Calteau A, Vallenet D, Bonnefoy V, Bruneel O, Chane-Woon- Ming B, Cleiss-Arnold J, Duran R, Elbaz-Poulichet F, Fonknechten N, Giloteaux L, Halter D, Koechler S, Marchal M, Mornico D, Schaeffer C, Thil Smith AA, Van Dorsselaer A, Weissenbach J, Médigue C and Le Paslier D (2011) Metabolic diversity between main

microorganisms inside an arsenic-rich ecosystem revealed by meta- and proteo-genomics. The ISME Journal. 5, 1735-1747 Bini E (2010) Archaeal transformation of metals in the environment. FEMS Microbiol Ecol. 73, 1-16 Bolan N, Kunhikrishnan A, Thangarajan R, Kumpiene J. Park J, Makino T, Kirkham MB, Scheckel, K (2014) Remediation of heavy metal (loid) s contaminated soils–to mobilize or to immobilize? Journal of hazardous materials. 266, 141-166 Brake SS, Dannelly HK, Connors KA, Hasiotis ST (2001) Influence of water chemistry on the distribution of an acidophilic protozoan in an acid mine drainage system at the abandoned Green Valley coal mine, Indiana. Appl. Geochem. 16, 1641-1652 Bruneel O, Duran R, Koffi K, Casiot C, Fourçans A, Elbaz-Poulichet F, Personné J-C (2005) Microbial diversity in a pyrite-rich tailings impoundment (Carnoulès, France). Geomicrobiology Journal. 22, 249 - 257 Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F, Personné JC (2006) Diversity of microorganisms in Fe-As-rich acid mine drainage waters of Carnoulès, France. Applied and Environmental Microbiology. 72, 551-556 Bruneel O. Pascault N., Egal M., Bancon-Montigny C., Goni M., Elbaz-Poulichet F., Personné J.-C., Duran R. (2008) Archaeal diversity in a Fe-As rich acid mine drainage at Carnoulès (France). Extremophiles. 12, 563-571 Bruneel O, Volant A, Gallien S, Chaumande B, Casiot C, Carapito C, et al. (2011) Characterization of the active bacterial community involved in natural attenuation processes in arsenic-rich creek sediments. Microb Ecol. 61, 793-810 Butler BA (2011) Effect of imposed anaerobic conditions on metals release from acid- mine drainage contaminated streambed sediments. Water Res. 45, 328-336 Carvalhais LC, Dennis PG, Tyson GW, Schenk, PM (2012). Application of metatranscriptomics to soil environments. Journal of microbiological methods. 91, 246-251 Casiot C, Leblanc M, Bruneel O, Personné J-C, Koffi K, Elbaz-Poulichet F (2003a) Geochemical processes controlling the formation of As-rich waters within a tailings impoundment. Aquatic Geochemistry. 9, 273-290 Casiot C, Morin G, Juillot F, Bruneel O, Personné JC, Leblanc M et al. (2003b) Bacterial immobilization and oxidation of arsenic in acid mine drainage (Carnoulès creek, France). Water Res. 37, 2929-2936 Chen LX, Li JT, Chen YT, Huang LN, Hua ZS, Hu M, Shu WS (2013) Shifts in microbial community composition and function in the acidification of a lead/zinc mine tailings. Environmental microbiology. 15, 2431-2444 Ciccarelli FD, Doerks T, von Mering C, Creevey CJ, Snel B and Bork P (2006) Toward automatic reconstruction of a highly resolved tree of life. Science. 311, 1283- 1287 Costa PS, Scholte LLS, Reis MP, Chaves AV, Oliveira PL, Itabayana LB, Suhadolnik MLS, Barbosa FAR., Chartone-Souza E, Nascimento AMA. (2014) Bacteria and genes involved in arsenic speciation in sediment impacted by long-term gold mining. Plos one. 9, e95655 Das S, Jean J-S, Kar S, Chou M-L, Chen C-Y (2014) Screening of plant growth- promoting traits in arsenic-resistant bacteria isolated from agricultural soil and their potential implication for arsenic bioremediation. J Hazard Mater. 272, 112–120

de-Bashan LE, Hernandez J-P, Nelson KN, Bashan Y, Maier R (2010) Growth of Quailbush in acidic, metalliferous desert mine tailings: effect of Azospirillum brasilense Sp6 on biomass production and rhizosphere community structure. Microb. Ecol. 60, 915-27 Delavat F, Lett MC and Lievremont D (2012) Novel and unexpected bacterial diversity in an arsenic-rich ecosystem revealed by culture-dependent approaches. Biol Direct. 7, 28 Denef VJ, Mueller RS, and Banfield JF (2010) AMD biofilms: using model communities to study microbial evolution and ecological complexity in nature. The ISME journal. 4, 599-610 Edgar RC, Haas BJ, Clemente JC, Quince C and Knight R (2011) UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27: 2194-2200 Edwards KJ, Bond PL, Druschel GK, McGuire MM, Hamers RJ & Banfield JF (2000a) Geochemical and biological aspects of sulfide mineral dissolution: lessons from Iron Mountain, California. Chem Geol. 169, 383–397 Edwards KJ, Bond PL and Banfield JF (2000b) Characteristics of attachment and growth of Thiobacillus caldus on sulphide minerals: a chemotactic response to sulphur minerals? Environ. Microbiol. 2, 324-332 Edwards KJ, Bond PL, Gihring TM, Banfield JF (2000c). An archaeal iron-oxidizing extreme acidophile important in acid mine drainage. Science. 287, 1796-1799. Falkowski PG, Fenchel T, Delong EF (2008) The microbial engines that drive earth's biogeochemical cycles. Science 320:1034-1039 Gadanho M, Libkind D and Sampaio JP (2006) Yeast diversity in the extreme acidic environments of the Iberian Pyrite Belt. Microbial ecology. 52, 552-563 Galand PE, Casamayor EO, Kirchman DL and Lovejoy C (2009) Ecology of the rare microbial biosphere of the Arctic Ocean. Proc Natl Acad Sci USA. 106, 22427–22432 Gifford SM, Sharma S, Rinta-Kanto JM and Moran MA (2011) Quantitative analysis of a deeply sequenced marine microbial metatranscriptome. ISME Journal. 5, 461-472 Gilbert JA and Dupont CL (2011) Microbial metagenomics: beyond the genome. Annual Review of Marine Science, 3, 347-371. Glick BR (2010) Using soil bacteria to facilitate phytoremediation Biotechnology Advances. 28, 367-374 Goebel BM and Stackebrandt E (1994) Cultural and phylogenetic analysis of mixed microbial populations found in natural and commercial bioleaching environments. Appl Environ Microbiol. 60, 1614–1621 Golyshina OV and Timmis KN (2005). Ferroplasma and relatives, recently discovered cell wall-lacking archaea making a living in extremely acid, heavy metal-rich environments. Environ Microbiol. 7, 1277-1288 González V, García I, del Moral F, de Haro S, Sánchez JA, Simón M (2011) Impact of unconfined sulphur-mine waste on a semi-arid environment (Almería, SE Spain) Journal of Environmental Management. 92, 1509-1519 Hakkou R, Benzaazoua M., Bussière B (2008a) Acid mine drainage at the abandoned Kettara mine (Morocco): 1. Environmental characterization. Mine Water Environ. 27, 145– 159

Hakkou R, Benzaazoua M, Bussière B, (2008b) Acid mine drainage at the abandoned Kettara mine (Morocco): 2. Mine waste geochemical behavior. Mine Water Environ. 27, 160– 170 Hakkou R, Benzaazoua M, Bussière B (2009) Laboratory evaluation of the use of alkaline phosphate wastes for the control of acidic mine drainage. Mine Water Environ. 28, 206–218 Hallberg KB (2010) New perspectives in acid mine drainage microbiology. Hydrometallurgy. 104, 448–453 Halter D, Goulhen-Chollet F, Gallien S, Casiot C, Hamelin J et al. (2012a) In situ proteo-metabolomics revealed metabolite secretion by the acid mine drainage bioindicator, Euglena mutabilis. The ISME J. 6, 1391-1402 Halter D, Casiot C, Heipieper HJ, Plewniak F, Marchal M, Simon S, et al. (2012b) Surface properties and intracellular speciation revealed an original adaptive mechanism to arsenic in the acid mine drainage bio-indicator Euglena mutabilis. Applied microbiology and biotechnology. 93, 1735-1744 Harrison JJ, Ceri H, Turner RJ (2007). Multimetal resistance and tolerance in microbial biofilms. Nature Reviews Microbiology. 5, 928-938 Héry M, Casiot C, Resongles E, Gallice Z, Bruneel O, Desoeuvre O, Delpoux S (2014) Release of arsenite, arsenate and methyl-arsenic species from streambed sediment impacted by acid mine drainage : a microcosm study. Environmental Chemistry. 11, 514-524 Hudson-Edwards KA and Dold B (2015). Mine Waste Characterization, Management and Remediation. Minerals. 5, 82-85 Huse SM, Welch DM, Morrison HG and Sogin ML (2010) Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environ Microbiol. 12, 1889-1898 Innerebner G, Knief C, Vorholt JA (2011) Protection of Arabidopsis thaliana against leaf-pathogenic Pseudomonas syringae by Sphingomonas strains in a controlled model system. Appl. Environ. Microbiol. 77, 3202–10 Jiang JQ, Ashekuzzaman SM, Jiang A, Sharifuzzaman SM, Chowdhury SR (2013) Arsenic contaminated groundwater and its treatment options in Bangladesh. International journal of environmental research and public health. 10, 18-46 Johnson DB and Hallberg KB (2003) The microbiology of acidic mine waters. Research in Microbiology. 154, 466-473 Johnson DB, Hallberg KB (2005) Acid mine drainage remediation options: a review. Sci Total Environ. 338, 3–14 Johnson DB, Hallberg KB (2008) Carbon, iron and sulfur metabolism in acidophilic micro-organisms. Adv. Microb. Physiol. 54, 201–255 Johnson DB (2014) Recent Developments in microbiological approaches for securing mine wastes and for recovering metals from mine waters. Minerals. 4, 279-292 Kavamura VN, Esposito E (2010) Biotechnological strategies applied to the decontamination of soils polluted with heavy metals Biotechnology Advances. 28, 61-69 Khan AG (2005) Role of soil microbes in the rhizospheres of plants growing on trace metal contaminated soils in phytoremediation Journal of Trace Elements in Medicine and Biology. 18, 355-364

Kidd P, Barceló J, Pilar Bernal M, Navari-Izzo F, Poschenrieder C, Shilev S, Rafael C, Monterroso C (2009) Trace element behaviour at the root–soil interface: Implications in phytoremediation. Environmental and Experimental Botany. 67, 243- 259 Knief C (2014) Analysis of plant microbe interactions in the era of next generation sequencing technologies. Frontiers in plant science. 5 Koffi K, Leblanc M, Jourde H, Casiot C, Pistre S, Gouze P and Elbaz-Poulichet F (2003) Reverse oxidation zoning in mine tailings generating arsenic-rich acidic waters (Carnoulès, France). Mine water and the environment. 22, 7-14 Kuang JL, Huang LN, Chen LX, Hua ZS, Li SJ, Hu M., et al., Shu WS (2012) Contemporary environmental variation determines microbial diversity patterns in acid mine drainage. The ISME journal. 7, 1038-1050 Langmuir D (1997) Acid mine waters. In. McConnin, R., Aqueous Environmental Geochemistry, Prentice-Hall, Inc., New Jersey, pp. 457-478 Leblanc M, Achard B, Othman DB, Luck JM, Bertrand-Sarfati J, Personné JC (1996) Accumulation of arsenic from acidic mine waters by ferruginous bacterial accretions (stromatolites). Applied Geochemistry. 11, 541-554 Leblanc M, Casiot C, Elbaz-Poulichet F and Personné C (2002) Arsenic removal by oxidizing bacteria in an heavily arsenic contaminated acid mine drainage system (Carnoulès, France). Geological Society of London Special Publication « Mine Water Hydrogeology and Geochemistry » (Younger P.L. and Robin N.S. Eds). In Geological Society, London, Spec. Publication. 198 pages, 267-274 Lett MC, Muller D, Lièvremont D, Silver S, Santini J (2012) Unified nomenclature for genes involved in prokaryotic aerobic arsenite oxidation. J Bacteriol. 194, 207–8 Lièvremont D, Bertin PN, and Lett M-C (2009) Arsenic in contaminated waters: biogeochemical cycle, microbial metabolism and biotreatment processes. Biochimie. 91, 1229-1237 Lghoul M, Maqsoud A, Hakkou R, and Kchikach A (2014) Hydrogeochemical behavior around the abandoned Kettara mine site, Morocco. Journal of Geochemical Exploration. 144, 456-467 Ma Y, Prasad MNV, Rajkumar M, Freitas H. (2011) Plant growth promoting rhizobacteria and endophytes accelerate phytoremediation of metalliferous soils Biotecnology advances. 29, 248-258 McClintock TR, Chen Y, Bundschuh J, Oliver JT, Navoni J, Olmos V, Villaamil Lepori E, Ahsan H, Parvez, F. (2012) Arsenic exposure in Latin America: Biomarkers, risk assessments and related health effects. Science of the Total Environment. 429, 76-91 Marchal M, Briandet R, Halter D, Koechler S, DuBow MS, Lett MC, Bertin PN (2011). Subinhibitory arsenite concentrations lead to population dispersal in Thiomonas sp. PloS one. 6(8), e23181 Mendez MO, Maier RM (2008) Phytoremediation of mine tailings in temperate and arid environments. Rev Environ Sci Biotechnol. 7, 47–59 Needleman SB and Wunsch CD (1970) A general method applicable to the search for similarities in the amino acid sequence of two proteins. J Mol Biol. 48, 443-453 Nordstrom (2002) Public health-worldwide occurrences of arsenic in ground water. Science. 296, 2143–2145

Oremland RS and Stolz JF (2003) The ecology of arsenic. Science. 300, 939-944 Oremland RS and Stolz JF (2005). Arsenic, microbes and contaminated aquifers. Trends in microbiology. 13, 45-49 Oremland RS, Kulp TR, Blum JS, Hoeft SE, Baesman S, Miller LG, Stolz JF (2005). A microbial arsenic cycle in a salt-saturated, extreme environment. Science. 308, 1305-1308 Oremland RS, Saltikov CW, Wolfe-Simon F, Stolz JF (2009) Arsenic in the evolution of earth and extraterrestrial ecosystems. Geomicrobiol J. 26, 522–536 Pace NR (1997) A molecular view of microbial diversity and the biosphere. Science. 276, 734–740 Park JM, Lee JS, Lee JU, Chon HT, Jung MC (2006) Microbial effects on geochemical behavior of arsenic in As-contaminated sediments. J Geochem Explor. 88, 134-138 Pedrós-Alió C (2007) Ecology, dipping into the rare biosphere. Science. 315, 192–193 Quince C, Lanzen A, Curtis TP, Davenport RJ, Hall N, Head IM et al. (2009) Noise and the accurate determination of microbial diversity from 454 pyrosequencing data. Nat Methods. 6, 639–641 Quince C, Lanzen A, Davenport RJ & Turnbaugh PJ (2011) Removing noise from pyrosequenced amplicons. BMC Bioinformatics. 12, 38 Ren WX, Li PJ, Zheng L, Fan SX, Verhozina VA (2009). Effects of dissolved low molecular weight organic acids on oxidation of ferrous iron by Acidithiobacillus ferrooxidans. Journal of hazardous materials. 162, 17-22 Rajkumar M, Ae N, Prasad MNV, Freitas H. (2010) Potential of siderophore-producing bacteria for improving heavy metal phytoextraction. Trends Biotechnol. 28, 142-149 Rajkumar M, Sandhya S, Prasad MNV, and Freitas H (2012) Perspectives of plant- associated microbes in heavy metal phytoremediation. Biotechnology advances. 30, 1562- 1574 Rappé MS and Giovannoni SJ (2003) The uncultured microbial majority. Annual Rev. Microbiol. 57, 369-94 Rowe OF, Johnson DB (2008) Comparison of ferric iron generation by different species of acidophilic bacteria immobilized in packed-bed reactors. Syst Appl Microbiol. 31, 68–77 Sand W, Gehrke T, Jozsa P-G, Schippers A (2001) (Bio)chemistry of bacterial leaching- direct vs. indirect bioleaching. Hydrometallurgy. 59, 159-175 Sarkar A; Kazy KS; Sar P (2014) Studies on arsenic transforming groundwater bacteria and their role in arsenic release from subsurface sediment. Environ Sci Pollut Res. 21, 8645– 8662 Schippers A, Breuker A, Blazejak A, Bosecker K, Kock D, Wright TL (2010) The biogeochemistry and microbiology of sulfidic mine waste and bioleaching dumps and heaps, and novel Fe(II)-oxidizing bacteria. Hydrometallurgy. 104, 342–350 Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. (2009) Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol. 75, 7537-7541 Singer PC, Stumm W (1970) Acid mine drainage: the rate-determining step. Science. 167, 1121–1123

Sogin ML, Morrison HG, Huber JA, Welch DM, Huse SM, Neal PR, et al. (2006) Microbial diversity in the deep sea and the underexplored ‘rare biosphere’. Proc Natl Acad Sci USA. 103, 12115–12120 Smouni A, Ater M, Auguy F, Laplaze L, El Mzibri M, Berhada F, Filali-Maltouf A, Doumas P. (2010) Evaluation de la contamination par les éléments-traces métalliques dans une zone minière du Maroc oriental Cah. Agric. 19, 1-7 Stolz JF, Basu P, and Oremland RS (2010) Microbial arsenic metabolism: new twists on an old poison. Microbe. 5, 53–59 Sun W, Sierra-Alvarez R, Milner L, Field JA (2010) Anaerobic oxidation of arsenite linked to chlorate reduction. Appl Environ Microbiol. 76, 6804–6811 Tyson GW, Chapman J, Hugenholtz P, Allen EE, Ram RJ, Richardson PM, and Banfield JF (2004) Community structure and metabolism through reconstruction of microbial genomes from the environment. Nature. 428, 37-43 Vera M, Schippers A, Sand W (2013) Progress in bioleaching: fundamentals and mechanisms of bacterial metal sulfide oxidation—Part A. Appl Microbiol Biotechnol. 97, 7529–7541 Volant A, Desoeuvre A, Casiot C, Lauga B, Delpoux S, Morin G, Personné JC, Héry M, Elbaz-Poulichet F, Bertin P and Bruneel O (2012) Archaeal diversity: temporal variation in the arsenic-rich creek sediments of Carnoulès Mine, France. Extremophiles. 16, 645-657. Volant A, Bruneel O, Desoeuvre A, Héry M, Casiot C, Bru N, Delpoux S, Fahy A, Javerliat F, Bouchez O, Duran R, Bertin PN, Elbaz-Poulichet F and Lauga B (2014) Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and others drivers along an acid mine drainage. FEMS Microbiology Ecology. 90, 247-263 Yang Y, Li Y, Sun QY (2014) Archaeal and bacterial communities in acid mine drainage from metal-rich abandoned tailing ponds, Tongling, China. Transactions of Nonferrous Metals Society of China. 24, 3332-3342 Whitman WB, Coleman DC and Wiebe WJ (1998) Prokaryotes: the unseen majority. Proceedings of the National Academy of Sciences. 95, 6578-6583 Yamamura S and Amachi, S (2014) Microbiology of inorganic arsenic: from metabolism to bioremediation. Journal of bioscience and bioengineering. 118, 1-9 Yamanaka K and Okada S (1994). Induction of lung-specific DNA damage by metabolically methylated arsenics via the production of free radicals. Environmental Health Perspectives. 102, 37-40 Younger PL (1997) The longevity of minewater pollution: a basis for decision-making. Science of the Total Environment. 194, 457-466 Yunus M, Sohel N, Hore SK, Rahman M (2011) Arsenic exposure and adverse health effects: a review of recent findings from arsenic and health studies in Matlab, Bangladesh. The Kaohsiung journal of medical sciences. 27, 371-376 Zargar K, Conrad A, Bernick DL, Lowe TM, Stolc V, et al. (2012) ArxA, a new clade of arsenite oxidase within the DMSO reductase family of molybdenum oxidoreductases. Environ Microbiol. 14, 1635–45

VIII ANNEXES : SELECTION DE 5 PUBLICATIONS

Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F, Personné JC (200) Diversity of microorganisms in Fe-As-rich acid mine drainage waters of Carnoulès, France. Applied and Environmental Microbiology. 72, 551-556

Bruneel O, Pascault N, Egal M, Bancon-Montigny C, Goni M, Elbaz-Poulichet F, Personné J-C, Duran R (2008) Archaeal diversity in a Fe-As rich acid mine drainage at Carnoulès (France). Extremophiles. 12, 563-571

Bruneel O, Volant A, Gallien S, Chaumande B, Casiot C, Carapito C, Bardil A, Morin G, Brown Jr GE, Personné JC, Le Paslier D, Schaeffer C, Van Dorsselaer A, Bertin PN, Elbaz-Poulichet F, Arsène-Ploetze F (2011) Characterization of the active bacterial community involved in natural attenuation processes in arsenic-rich creek sediments. Microbial Ecology. 61, 793-810

Volant A, Desoeuvre A, Casiot C, Lauga B, Delpoux S, Morin G, Personné JC, Héry M, Elbaz-Poulichet F, Bertin P and Bruneel O (2012) Archaeal diversity: temporal variation in the Arsenic-Rich Creek Sediments of Carnoulès Mine, France. Extremophiles. 16, 645-657

Volant A, Bruneel O, Desoeuvre A, Héry M, Casiot C, Bru N, Delpoux S, Fahy A, Javerliat F, Bouchez O, Duran R, Bertin PN, Elbaz-Poulichet F and Lauga B (2014) Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and others drivers along an acid mine drainage. FEMS Microbiology Ecology. 90, 247-263

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jan. 2006, p. 551–556 Vol. 72, No. 1 0099-2240/06/$08.00ϩ0 doi:10.1128/AEM.72.1.551–556.2006 Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Diversity of Microorganisms in Fe-As-Rich Acid Mine Drainage Waters of Carnoule`s, France O. Bruneel,1* R. Duran,2 C. Casiot,1 F. Elbaz-Poulichet,1 and J.-C. Personne´1 Laboratoire Hydrosciences Montpellier, UMR5569, Universite´Montpellier 2, Place E. Bataillon, Case MSE, 34095 Montpellier cedex 05,1 and Laboratoire d’Ecologie Mole´culaire-Microbiologie, EA 3525, Universite´de Pau et des Pays de l’Adour, Avenue de l’Universite´, IBEAS, BP 1155, F-64013 Pau cedex,2 France

Received 14 March 2005/Accepted 28 September 2005

The acid waters (pH 2.7 to 3.4) originating from the Carnoule`s mine tailings contain high concentrations of · dissolved arsenic (80 to 350 mg · liter؊1), iron (750 to 2,700 mg · liter؊1), and sulfate (2,000 to 7,500 mg liter؊1). During the first 30 m of downflow in Reigous creek issuing from the mine tailings, 20 to 60% of the dissolved arsenic is removed by coprecipitation with Fe(III). The microbial communities along the creek have been characterized using terminal-restriction fragment length polymorphism (T-RFLP) and 16S rRNA gene library analyses. The results indicate a low bacterial diversity in comparison with unpolluted water. Eighty percent of the sequences obtained are related to sequences from uncultured, newly described organisms or recently associated with acid mine drainage. As expected owing to the water chemistry, the sequences recovered are mainly related to bacteria involved in the geochemical Fe and S cycles. Among them, sequences related to uncultured TrefC4 affiliated with Gallionella ferruginea, a neutrophilic Fe-oxidizing bacterium, are dominant. The description of the bacterial community structure and its dynamics lead to a better understanding of the natural remediation processes occurring at this site.

The processing of sulfide-rich ores in the recovery of base toxic, bacteria in acid mine waters may be useful in AMD metals, such as copper, lead, zinc, and gold, has produced large bioremediation or that of some other industrial effluents. In quantities of pyrite wastes (20). When exposed to rain, this order to develop remediation processes or optimize them, fur- material generates acid mine drainage (AMD) which contains ther knowledge of the bacteria living in the extreme environ- large amounts of sulfate, iron, arsenic, and heavy metals. De- ment of AMD is required. spite their toxicity, such waters host organisms, both pro- This study aims to investigate the microbial community of a karyotes and eukaryotes, which are able to cope with the pol- small creek (the Reigous, present at Carnoule`s, France). The lution (2, 33). Some of them have the capacity to modify the Carnoule`s mine (Fig. 1) has been inactive since 1962. Its ex- physicochemical conditions of the water either by detoxifica- ploitation has left about 1.5 megatons of tailings containing tion or by metabolic exploitation. For example, efficient oxida- 0.7% Pb, 10% FeS2, and 0.2% As. The tailings are contained tion of As by bacteria has been reported in AMD or in chem- behind a dam. Water percolating through the tailings emerges ically somewhat similar waters like those from hot springs (3, 7, at the base of the dam, forming the head of the Reigous creek. 21, 25, 30). Because of their elevated Fe concentration, the The head waters of the creek are characterized by low pH (2.7 development of iron-oxidizing bacteria is favored in AMD (16) to 3.4) and high concentrations of As (100 to 350 mg · literϪ1), where Acidithiobacillus ferrooxidans and Leptospirillum fer- Ϫ1 2Ϫ Fe (750 to 2,700 mg · liter ), and SO4 (2,000 to 7,500 mg · rooxidans are often observed (2). literϪ1). Owing to its ability to oxidize Fe, the bacterial consortium in The As and Fe behavior in creek water has been intensively AMD plays a major role in the immobilization of the elements studied (8, 22, 23, 24). As(III) is the dominant As type, whereas that exhibit a strong affinity for solid Fe oxide phases such as Fe occurs as Fe(II). Along the first 30 m of the creek (about 1 h Sr, Cs, Pb, U (14), and As (8, 24). In addition, the ability of of residence time), the bacterially mediated oxidation of Fe(II) several bacterial strains in AMD to oxidize As further contrib- leads to the coprecipitation of 20 to 60% of the dissolved As. utes to reduction of its toxicity in water, because As(III) is The precipitate which contains up to 22% of As is mainly considered to be more toxic than As(V) (28) and because composed of As(III)-Fe(III) oxy-hydroxide in the wet season arsenate adsorbs more strongly than arsenite to Fe(III) oxides while As(V)-Fe(III) oxy-hydroxide compounds predominate in and hydroxides at acidic pH (5, 26). the dry season. Several phenotypes of Acidithiobacillus ferrooxi- Owing to their tolerance of heavy metals and the ability of dans some to promote transformations that make some metals less have been isolated, and their role in the oxidation of Fe(II) and the coprecipitation of As has been demonstrated in laboratory experiments (8, 11). Additionally, Bruneel et al. (7) * Corresponding author. Mailing address: Laboratoire Hydro- isolated at this site three different strains of Thiomonas spp. sciences Montpellier, UMR5569, Universite´Montpellier 2, Place E. closely related to Thiomonas sp. strain Ynys1 able to promote Bataillon, Case MSE, 34095 Montpellier cedex 05, France. Phone: 33-4-67-14-36-59. Fax: 33-4-67-14-47-74. E-mail: [email protected] As oxidation in laboratory conditions. -montp2.fr. The present study combines terminal-restriction fragment

551 552 BRUNEEL ET AL. APPL.ENVIRON.MICROBIOL.

Purified PCR products (600 to 700 ng) were digested with 12 U of enzyme HaeIII or HinfI (New England Biolabs). The lengths of terminal-restriction fragments (T-RFs) from the digested PCR products were determined by capil- lary electrophoresis on an ABI prism 310 (Applied Biosystems). About 50 ng of the digested DNA from each sample was mixed with 10 ␮l of deionized form- amide and 0.25 ␮l of 6-carboxytetramethylrhodamine size standard, denatured at 94°C for 2 min, and immediately chilled on ice prior to electrophoresis. After an injection step of 10 s, electrophoresis was carried out for up to 30 min, applying a voltage of 15 kV. T-RFLP profiles were performed using GeneScan software (ABI). Dominant operational taxonomic units represent T-RFs whose fluorescence was higher than 100 fluorescence units for at least one sample. Predictive diges- tions were made on the RDP web site (http://rdp.cme.msu.edu/html/index.html) using the T-RFLP Analysis Program. Cloning and restriction analysis. To further characterize the bacterial populations inhabiting the creek in each sampling period and sampling point, the bacterial diversity was analyzed by cloning PCR amplified 16S rRNA genes. For S1 and COWG, libraries were constructed for each sampling period. For COWA, a library was constructed only for October, since the comparison between Oc- tober and January T-RFLP profiles showed mainly a disappearance of T-RFs in January. Bacterial 16S rRNA genes were amplified with unlabeled 8F and 1489R primers. These PCR products were cloned in Escherichia coli TOP10 using the pCR2.1 Topo TA cloning kit (Invitrogen, Inc.). Cloned 16S rRNA gene fragments were amplified using the primers TOP1 (5Ј-GTGTGCTGGAATTCGCC CTT-3Ј) and TOP2 (5Ј-TATCTGCAGAATTCGCCCTT-3Ј), located on the vector and surrounding the inserted PCR fragment, and then were digested with the enzyme HaeIII or HinfI. Restriction profiles were analyzed using 2.5% agarose gel electrophoresis (small-fragment resolution agarose; QA agarose; QBioge`ne, Inc.). Sixty clones from each library were analyzed and grouped according to their RFLP patterns (HaeIII and HinfI digestion). Only sequences from dominant groups were determined 16S rRNA gene sequencing. Partial sequences of the 16S rRNA gene (from 8 to 336 according to E. coli numbering) were determined by the dideoxy nucleotide chain termination method using a BigDye cycle sequencing kit (Applied Biosystems) on an ABI PRISM 310 Genetic analyzer (Applied Biosystems). DNA sequence analyses were performed via the infobiogen server (http://www FIG. 1. Map of the Carnoule`s mining site and location of sampling .infobiogen.fr) by using the FASTA, BLAST, ALIGNN, and CLUSTALW pro- stations. Sampling stations were Reigous spring (S1), 3 m downstream grams (1, 13, 29). Phylogenetic trees were constructed by using the PHYLIP of the spring (COWA), and 30 m downstream of the spring (COWG). computer package (13). The confidence level of the phylogenetic tree topology was evaluated by performing 100 bootstrap replications with the SEQBOOK program. length polymorphism (T-RFLP) analysis in order to investigate the dynamics of the bacterial communities and 16S rRNA gene RESULTS library analysis to identify the dominant bacterial group. Bacterial community structures. The results of the T-RFLP MATERIALS AND METHODS analysis of bacterial community structure are presented in Fig. 2. The average T-RF number was relatively small (about 10) Sampling procedure and physicochemical determinations in situ. Water samples for molecular analysis of microbial populations were collected in October both in October and January, reflecting low bacterial diversity. 2002 and January 2003 in the spring and at two other locations in the creek over The bacterial population characterized by a 216-bp (Ϯ 2 bp) a distance of 30 m (Fig. 1). A volume of 200 ml of water was filtered through a T-RF was generally the most abundant, except at station S1 in ␮ sterile 0.22- m nucleopore filter. These filters were then transferred to a tube, October. Abundance of this 216-bp T-RF generally increased frozen in liquid nitrogen, and stored at Ϫ20°C until further analysis. DNA isolation. Genomic DNA was extracted from filtered water using the between October and January, except at station COWG, where UltraClean Soil DNA Isolation kit according to the recommendation of the the variations were minor. manufacturer (MoBio Laboratories, Inc.). All extracted genomic DNA samples Composition of bacterial communities. The most represen- were stored at Ϫ20°C until further processing. tative sequences of the dominant clones are summarized in T-RFLP analysis. Ј Ј Primers 8F (5 -AGAGTTTGATCCTGGCTCAG-3 ) and Table 1, and the phylogenetic analysis of all the obtained 1489R (5Ј-TACCTTGTTACGACTTCA-3Ј) (19, 31) were used for T-RFLP analysis to assess the bacterial community structures. Forward (8F) and reverse sequences are presented in Fig. 3 to 5. (1489R) primers were fluorescently labeled with tetrachlorofluorescein phos- The most abundant sequence types are positioned within the phoramidite and hexachlorofluorescein phosphoramidite (E.S.G.S. Cybergene beta subdivision of the Proteobacteria (Table 1). They were ␮ Group), respectively. The PCR amplification mixture contained 12.5 l Hot Start recovered at all stations during both sampling periods, ac- Taq polymerase master mix (QIAGEN), 0.5 ␮l of each primer (20 ␮M), and 10 ng of DNA template. A final volume of 50 ␮l was adjusted with distilled water. counting for 5 to 28% of the clones in October and more than 16S rRNA gene amplification reactions were cycled in a PTC200 thermocycler 65% in January. Numerous clone sequences of this group dis- (MJ Research) with a hot start step at 94°C for 15 min followed by 35 cycles of played around 95% homology with a sequence isolated from 94°C for 1 min, 52°C for 1.5 min, and 72°C for 1 min, with a final extension step an acid- and iron-rich stream in the United Kingdom (Gen- at 72°C for 10 min. The amount of PCR product was determined by comparison Bank accession no. AY766002) (unpublished data). The phy- to known concentrations by the “dots method” (Smartlader; Eurogentec) after migration on agarose gel. PCR products were purified with the GFX PCR DNA logenetic analyses (Fig. 3) did not allow affiliation of the clone purification kit (Amersham-Pharmacia). sequences with any representative of the subdivision. The clos- FIG. 2. Seasonal comparison of bacterial community T-RFLP fingerprints from the AMD of Carnoule`s,France, in October and January samples.

FIG. 3. Phylogenetic analysis of 16S rRNA gene sequences afﬁliated with the Gallionella division from the AMD of Carnoule`s, France. Clone names in boldface correspond to sequences found in October (Oct) and January (Jan) within the three stations along the Reigous Creek, S1, COWA, and COWG.

553 554 BRUNEEL ET AL. APPL.ENVIRON.MICROBIOL. b 202 160 217 217 202 217 202 217 202 217 219 202 251–197 250–197 T-RFs (bp) a 8 8 5 8 5 5 4 5 1 10 28 22 66 70 of clones (%) Relative abundance % 96 94 94 95 94 94 91 97 93 95 91 91 Similarity (sulfate reduction) (sulfate reduction) (sulfate reduction) (sulfate reduction) (sulfide oxidation) (iron oxidation) (iron oxidation) (iron oxidation) (iron oxidation) (iron oxidation) (sulfate reduction) c spp. (sulfide/iron oxidation) (sulfide/iron oxidation and iron reduction) 99 (sulfide/iron oxidation and iron reduction) 95 Closest relative (postulated metabolism) and COWG) ␦ - Proteobacteria Desulfobacterium indolicum Actinobacteria Actinomycetales ␥ - Proteobacteria A. ferrooxidans ␤ - Proteobacteria Gallionella ferruginea ␤ - Proteobacteria Gallionella ferruginea ␦ - Proteobacteria Desulfobacterium indolicum ␤ - Proteobacteria Gallionella ferruginea ␦ - Proteobacteria Desulfobacterium indolicum ␤ - Proteobacteria Gallionella ferruginea ␦ - Proteobacteria Desulfobacterium indolicum ␤ - Proteobacteria Gallionella ferruginea ␤ - Proteobacteria Thiobacillus plumbophilus ␦ - Proteobacteria Desulfomonile tiedjei ␥ - Proteobacteria A. ferrooxidans FIG. 4. Phylogenetic analysis of 16S rRNA gene sequences affiliated with the Desulfobacterium division from the AMD of Carnoule`s, France. Clone names in boldface correspond to sequences found in October (Oct) and January (Jan) within the three stations along the Reigous Creek, S1, COWA, and COWG. AJ877944 AJ877952 AJ877934

est relative (91%) is Gallionella ferruginea, a neutrophilic iron- oxidizing bacterium. The sequences representing the second-most abundant type are positioned within the delta subdivision of the Proteobacte- S1Oct7 S1Oct43 AJ877956 S1Oct11 AJ877924 COWAOct77 AJ877946 COWGOct52 AJ877947 S1Jan50 AJ877949 COWGJan29 AJ877957 COWGJan58 AJ877951 COWGJan20 AJ877953 ria (Table 1, Fig. 4). These sequences were more abundant in October, representing 10, 8, and 5% of the clones at S1, COWA, and COWG, respectively, than in January, with 5 and 1% at S1 and COWG. In October, all the clones were similar Station Clone Accession no. Phylogenetic group COWA COWAOct18 AJ877926 COWG COWGOct61 AJ877929 COWG COWGJan9 AJ877935 (more than 90% similarity) to clones found in an AMD at Iron Mountain (4). In contrast, the clones of January were similar (94% similarity) to those found in a forested wetland impacted Corresponds to the relative abundance of clones for each library. T-RFs correspond to expected terminal fragments with HaeIII predictive digestion. The phylogenetic group, closest relatives, and postulated metabolism and relative abundance of clones are indicated. by sulfate-rich waters from coal piles (6). As for the main a b c TABLE 1. Most representative sequences of bacterial clones found in October in the three stations of the Reigous Creek (S1, COWA, and COWG) and in two stations in January (S1 period Sampling October S1 S1Oct9 January S1 S1Jan1 sequence, the phylogenetic analyses did not allow the afﬁlia- VOL. 72, 2006 DIVERSITY OF MICROORGANISMS IN MINE DRAINAGE 555

The next most abundant sequence types, representing 8% of the clones at S1 in October and 4% at COWG in January (Table 1), were firmly positioned in the Acidithiobacillus ferrooxidans group (Fig. 5). Three sequences are related (99% similarity) to uncultured A. ferrooxidans KF/GS-JG36-22 (27) isolated in waste piles of a uranium mine, and one sequence was related to A. ferrooxidans DSM 2392 (Fig. 5). The next group, representing 8% of the clones at S1 in October, was associated with the Actinobacteria group, with 94% similarity with sequences recovered in forested wetland exposed to coal effluent (6). The phylogenetic analysis could not affiliate the sequence with any isolated bacterium (data not shown). The last sequence found, detected only in January at COWG, representing 5% of the clones, was firmly positioned in the Thiobacillus group with 91% similarity with T. plumbophilus DSM 6690. These strains were isolated from a ura-

nium mine, and they grew by oxidation of H2S, galena (PbS),

and H2 (10). Finally, sequences closely related to 16S rRNA genes from a chloroplast of Euglena spp. were also detected (data not shown). This was not surprising, since the 16S rRNA gene of chloroplasts is closely related to the bacterial 16S rRNA gene and therefore can be ampliﬁed by primers 8F and 1489R. Moreover, this is consistent with previous work indicating that these organisms are able to accumulate and oxidize As in the cell (9).

DISCUSSION In the Reigous creek, the low bacterial diversity as revealed by molecular-based methods is consistent with the results of Baker and Banfield (2) in a similar environment. This may reflect the limited number of different electron donors and acceptors available in AMD and the toxicity of heavy metals and low pH. Numerous sequences in the libraries are related to sequences previously found in AMD, indicating that the clone libraries were not contaminated. Nevertheless, 80% of the sequences could not be closely related to cultured organisms, suggesting that they may constitute new taxa. As long as the bacterial strains were not isolated, their physiological role in the creek ecology will remain uncertain. Both molecular methods revealed that the dominant population (216-bp [Ϯ2 bp] T-RF) can be related to Gallionella ferruginea sequences, as indicated by predictive digestion (217 bp) and 16S rRNA gene library analyses. Gallionella ferruginea FIG. 5. Phylogenetic analysis of 16S rRNA gene sequences affili- is a neutrophilic bacterium that oxidizes Fe. It has been shown ated with the Acidithiobacillus division from the AMD of Carnoule`s, to efficiently remove Fe, As(III), and As(V) in water (17). It is France. Clone names in boldface correspond to sequences found in possible that an acid-tolerant relative of this bacterium has the October (Oct) and January (Jan) within the three stations along the ability to oxidize iron under acid pH conditions. In the creek, Reigous Creek, S1, COWA, and COWG. Strains in boldface (CC1, B5, the abundance of this population was much more significant in B4, and B9) represent the bacteria isolated in the Carnoule`s Creek. January (more than 65%) than in October (less than 30%). Such variations are consistent with the occurrence of higher Fe tion of the clone sequences with any representative of the and As precipitation rates in the rainy seasons than in other subdivision. The closest relative was Desulfobacterium indoli- seasons, as reported by Casiot et al. (8). In addition to the cum, a sulfate-reducing bacterium (18). For the clone from Gallionella ferruginea sequences, the library analyses show the COWG in January that was phylogenetically distant from the presence of other uncultured bacterial groups related to the Fe others (Fig. 4), the closest relative is Desulfomonile tiedjei,a cycle, such as the Actinobacteria group. Members of this group, sulfate-reducing bacterium (12). previously reported in AMD, are iron-oxidizing, heterotrophic, 556 BRUNEEL ET AL. APPL.ENVIRON.MICROBIOL. acidophilic bacteria capable of autotrophic growth. Some of Euglena mutabilis, in acid mine drainage (Carnoule`s, France). Sci. Total them may play a synergistic role, removing organic carbon (4). Environ. 320:259–267. 10. Drobner, E., H. Huber, R. Rachel, and K. O. Stetter. 1992. Thiobacillus Finally, A. ferrooxidans constituted a minor group in the Fe- plumbophilus spec. nov., a novel galena and hydrogen oxidizer. Arch. Mi- oxidizing bacterial population contrary to expectations from crobiol. 157:213–217. 11. Duquesne, K., S. Lebrun, C. Casiot, O. Bruneel, J.-C. Personne´, M. Leblanc, previous findings based on isolation and culturing techniques F. Elbaz-Poulichet, G. Morin, and V. Bonnefoy. 2003. Immobilization of (8). arsenite and ferric iron by Acidithiobacillus ferrooxidans in acid mine drain- With respect to bacteria involved in S cycling, the sequences age. Appl. Environ. Microbiol. 69:6165–6173. 12. El Fantroussi, S., J. Mahillon, H. Naveau, and S. N. Agathos. 1997. Intro- recovered, in addition to A. ferrooxidans, are related to mem- duction of anaerobic dechlorinating bacteria into soil slurry microcosms and bers of the Desulfobacterium genera, which contains sulfate- nested-PCR monitoring. Appl. Environ. Microbiol. 63:806–811. reducing bacteria (18). As the water of the Carnoule`s creek is 13. Felsenstein, J. 1993. PHYLIP-phylogeny inference package (version 3.5c), Department of Genetics, University of Washington, Seattle, WA. fully oxygenated, the presence of bacteria from this group, 14. Ferris, F. G., R. O. Hallberg, B. Lyveń, and K. Pedersen. 2000. Retention of which is characterized by anaerobic respiration, may be sur- strontium, cesium, lead and uranium by bacterial iron oxides from a subterranean environment. Appl. Geochem. 15:1035–1042. prising. Nevertheless, this is in agreement with several studies 15. Friese, K., K. Wendt-Potthoff, D. W. Zachmann, A. Fauville, B. Mayer, and that have recently reported sulfate- and iron-reducing bacteria J. Veizer. 1998. Biogeochemistry of iron and sulfur in sediments of an acidic under acidic conditions (15, 32). mining lake in Lusatia, Germany. Water Air Soil Poll. 108:231–247. 16. Hallberg, K. B., and D. B. Johnson. 2003. Novel acidophiles isolated from Considering the small population, sequence analyses indi- moderately acidic mine drainage waters. Hydrometallurgy 71:139–148. cate the presence of bacteria from the ␤ subdivision of the 17. Katsoyiannis, I. A., and A. I. Zouboulis. 2004. Application of biological Proteobacteria affiliated with the Thiobacillus group. A member processes for the removal of arsenic from groundwater. Water Res. 38:17– 26. of this group was recently described as a galena and hydrogen 18. Kuever, J., M. Konneke, A. Galushko, and O. Drzyzga. 2001. Reclassification oxidizer (11). A Thiomonas sp., which has been isolated and of Desulfobacterium phenolicum as Desulfobacula phenolica comb. nov. and description of strain SaxT as Desulfotignum balticum gen. nov., sp. nov. Int. shown to be very active in the oxidation of As (7), was not J. Syst. Evol. Microbiol. 51:171–177. detected by molecular techniques, probably reflecting its low 19. Lane, D. J. 1991. rRNA sequencing, p. 115–175. In G. M. E. Stachenbradt abundance. (ed.), Nucleic acid techniques in bacterial systematics. Wiley, Chichester, United Kingdom. Analyses of the most abundant clones strongly indicate that 20. Langmuir, D. 1997. Acid mine waters, p. 457–478. In R. McConnin (ed.), further efforts have to be exerted to fully understand AMD Aqueous environmental geochemistry. Prentice-Hall, Englewood Cliffs, N.J. systems. They include isolation and identification of the organ- 21. Langner, H. W., C. R. Jackson, T. R. Mcdermott, and W. P. Inskeep. 2001. Rapid oxidation of arsenite in a hot spring ecosystem, Yellowstone National isms represented by these clones in order to define their eco- Park. Environ. Sci. Technol. 35:3302–3309. physiological roles. 22. Leblanc, M., B. Achard, D. Benothman, J. Bertrand-Sarfati, J. M. Luck, and J. C. Personne´. 1996. Accumulation of arsenic from acidic mine waters by ferruginous bacterial accretions (stromatolites). Appl. Geochem. 11:541– ACKNOWLEDGMENTS 554. 23. Michard, G., and J. Faucherre. 1970. Etude geóchimique de l’alte´ration des This study was financed by the project GEOMEX from the CNRS, minerais sulfure´s de St. Se´bastien d’Aigrefeuille. Chem. Geol. 6:63–84. the project Environnement Vie et Socie´te´s (CNRS-INSU), the ACI- 24. Morin, G., F. Juillot, C. Casiot, O. Bruneel, J.-C. Personne´, F. Elbaz-Pou- Ecologie Quantitative, and the ACI-ECCODYN (French Ministry of lichet, M. Leblanc, P. Ildefonse, and G. Calas. 2003. Bacterial formation of Research). tooeleite and mixed arsenic(III) or arsenic(V)-Iron(III) gels in the Car- noule`s acid mine drainage, France. A XANES, XRD, and SEM study. Environ. Sci. Technol. 37:1705–1712. REFERENCES 25. Oremland, R. S., and J. F. Stolz. 2003. The ecology of arsenic. Science 1. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. 300:939–944. Basic local alignment search tool. J. Mol. Biol. 215:403–410. 26. Sadiq, M. 1997. Arsenic chemistry in soils: an overview of thermodynamic 2. Baker, B. J., and J. F. Banfield. 2003. Microbial communities in acid mine predictions and field observations. Water Air Soil Poll. 93:117–136. drainage. FEMS Microbiol. Ecol. 44:139–152. 27. Selenska-Pobell, S., G. Kampf, K. Flemming, G. Radeva, and G. Satchanska. 3. Battaglia-Brunet, F., M.-C. Dictor, F. Garrido, C. Crouzet, D. Morin, K. 2001. Bacterial diversity in soil samples from two uranium waste piles as Dekeyser, M. Clarens, and P. Baranger. 2002. An arsenic(III)-oxidizing determined by rep-APD, RISA and 16S rDNA retrieval. Antonie Leeuwen- bacterial population: selection, characterization, and performance in reac- hoek 79:149–161. tors. J. Appl. Microbiol. 93:656–667. 28. Squibb, K. S., and B. A. Fowler. 1983. The toxicity of arsenic and its com- 4. Bond, P. L., S. P. Smriga, and J. F. Banfield. 2000. Phylogeny of microoor- pounds, p. 233–269. In B. A. Fowler (ed.), Biological and environmental gansims populating a thick, subaerial, predominantly lithotrophic biofilm at effects of arsenic. Elsevier, New York, N.Y. an extreme acid mine drainage site. Appl. Environ. Microbiol. 66:3842–3849. 29. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: 5. Bowell, R. J. 1994. Sorption of arsenic by iron oxides and oxyhydroxides in improving the sensitivity of progressive multiple sequence alignment through soils. Appl. Geochem. 9:279–286. sequence weighting, position-specific gap penalties and weight matrix choice. 6. Brofft, J. E., J. V. McArthur, and L. J. Shimkets. 2002. Recovery of novel Nucleic Acids Res. 22:4673–4680. bacterial diversity from a forested wetland impacted by reject coal. Environ. 30. Wakao, N., H. Koyatsu, Y. Komai, H. Shimokawara, Y. Sakurai, and H. Microbiol. 4:764–769. Shiota. 1988. Microbial oxidation of arsenite and occurrence of arsenite- 7. Bruneel, O., J.-C. Personne´, C. Casiot, M. Leblanc, F. Elbaz-Poulichet, B. J. oxidizing bacteria in acid mine water from a sulfur-pyrite mine. Geomicro- Mahler, A. Le Fle`che, and P. A. D. Grimont. 2003. Mediation of arsenic biol. J. 6:11–24. oxidation by Thiomonas sp. in acid mine drainage (Carnoule`s, France). 31. Weisburg, W. G., S. M. Barns, D. A. Pelletier, and D. J. Lane. 1991. 16S J. Appl. Microbiol. 95:492–499. ribosomal DNA amplification for phylogenetic study. J. Bacteriol. 173:697– 8. Casiot, C., G. Morin, F. Juillot, O. Bruneel, J. C. Personne´, M. Leblanc, K. 703. Duquesne, V. Bonnefoy, and F. Elbaz-Poulichet. 2003. Bacterial immobili- 32. Wielinga, B., J. K. Lucy, J. N. Moore, O. F. Seastone, and J. E. Gannon. zation of arsenic in acid mine drainage (Carnoule`s creek, France). Water 1999. Microbiological and geochemical characterization of fluvially depos- Res. 37:2929–2936. ited sulfidic mine tailings. Appl. Environ. Microbiol. 65:1548–1555. 9. Casiot, C., O. Bruneel, J.-C. Personne´, M. Leblanc, and F. Elbaz-Poulichet. 33. Zettler, L. A. A., F. Go´mez,E. Zettler, B. G. Keenan, R. Amils, and M. L. 2003. Arsenic oxidation and bioaccumulation by the acidophilic protozoan, Sogin. 2002. Eukaryotic diversity in Spain’s river of fire. Nature 417:137. Extremophiles DOI 10.1007/s00792-008-0160-z

ORIGINAL PAPER

Archaeal diversity in a Fe–As rich acid mine drainage at Carnoule`s (France)

O. Bruneel Æ N. Pascault Æ M. Egal Æ C. Bancon-Montigny Æ M. S. Gon˜i-Urriza Æ F. Elbaz-Poulichet Æ J.-C. Personne´ Æ R. Duran

Received: 26 November 2007 / Accepted: 9 March 2008 Ó Springer 2008

Abstract The acid waters (pH = 2.73–3.4) that originate Keywords Microbial diversity Arsenic from the Carnoule`s mine tailings (France) are known for Acid mine drainage Mine tailings their very high concentrations of As (up to 10,000 mg l-1) and Fe (up to 20,000 mg l-1). To analyze the composition of the archaeal community, (their temporal variation inside Introduction the tailing and spatial variations all along the Reigous Creek, which drains the site), seven 16S rRNA gene The processing of sulfide-rich ores in the recovery of base libraries were constructed. Clone analysis revealed that all metals such as copper, lead, zinc, and gold, has produced the sequences were affiliated to the phylum Euryarchaeota, large quantities of pyrite wastes (Langmuir 1997). When while Crenarchaeota were not represented. The study exposed to rain, this material generates acid mine drainage showed that the structure of the archaeal community of the (AMD) which contains large quantities of sulfate, iron, aquifer of the tailing stock is different to that of the Rei- arsenic and heavy metals. Despite their toxicity, these gous Creek. Irrespective of the time of sampling, the most waters are colonized by iron- and sulfur-oxidizing pro- abundant sequences found inside the tailing stock were karyotes and form stable microbial communities with related to Ferroplasma acidiphilum, an acidophilic and obligate acidophilic eukaryotes (fungi, yeasts, algae and ferrous-iron oxidizing Archaea well known for its role in protozoa; Johnson 1998; Zettler et al. 2002). The metabolic bioleaching. Inversely, in Reigous Creek, a sequence activities of such communities lead to solubilization affiliated to the uncultured Thermoplasmatales archaeon, (leaching) of the heavy metals from the sulfidic ores and clone YAC1, was largely dominant. This study provides a pollution of surface and subsurface waters fed by the run- better understanding of the microbial community associ- off. ated with an acid mine drainage rich in arsenic. For several decades, bacteria-like Acidithiobacillus or Leptospirillum have been considered to be the principal acidophilic sulfur- and iron-oxidizing microorganisms in Communicated by J.N. Reeve. AMD. They were believed to be responsible for pyrite oxidation and for the release of associated metals. How- O. Bruneel (&) N. Pascault M. Egal C. Bancon-Montigny F. Elbaz-Poulichet J.-C.Personne´ ever, during the last 10 years, several studies have Laboratoire Hydrosciences Montpellier, UMR 5569, evidenced the presence of archaeal communities in acidic IRD, CNRS, Universite´s Montpellier 1 et 2, waters (Edwards et al. 2000; Dopson et al. 2004). Previ- Universite´ Montpellier 2, Place E. Bataillon, ously, Archaea were renowned for their ability to inhabit Case MSE, 34095 Montpellier Cedex 05, France e-mail: [email protected] extreme environments and specialized niches but their widespread presence in non-extreme environments, such as M. S. Gon˜i-Urriza R. Duran marine and terrestrial soils, was also recently revealed Equipe Environnement et Microbiologie (Chaban et al. 2006). UMR CNRS 5254, IPREM, EEM, Universite´ de Pau et des Pays de l’Adour, Avenue de l’Universite´, IBEAS, Archaeal communities are often better adapted to low BP 1155, 64013 Pau Cedex, France pH, high concentrations of total and ferrous iron and other 123 Extremophiles metals, and moderately elevated temperatures than classi- The source of the Reigous Creek, now located at the foot cal bioleaching mesophilic bacteria (Acidithiobacillus spp. of the dike retaining the mining spoil, is acid (pH 2.7–3.4) and Leptospirillum spp.). Archaea were seen as numeri- and very rich in dissolved arsenic and iron (80–350 and cally significant members in these environments (Bond 750–2,700 mg l-1 respectively, Leblanc et al. 2002) pre- et al. 2000; Edwards et al. 2000; Johnson and Hallberg dominantly in their reduced forms: As(III) and Fe(II). The 2003). Furthermore, it has been suggested that Archaea water discharge is comprised between 0.8 and 1.7 l s-1. could play a major role in the generation of AMD (Baker In the Reigous Creek, As(III) is the dominant As species and Banfield 2003) with oxidation of iron. Some members whereas Fe occurs as Fe(II). Along the first 30 m of the of the Archaea that respire As(V) like Pyrobaculum creek (about 1 h residence time), the microbial mediated aerophilum and Pyrobaculum arsenaticum have been dis- oxidation of Fe(II) leads to the coprecipitation of 20–60% covered (Huber et al. 2000; Oremland and Stolz 2003). of the dissolved As. As-rich (up to 20%) yellow sediments Furthermore, Pyrobaculum arsenaticum, forms realgar cover the bottom of the creek. The precipitate is mainly

(As2S2) as a precipitate under organotrophic conditions in composed of amorphous Fe(III)–As(III) associated with the presence of thiosulfate and arsenate. These findings tooeleite, a rare nanocrystal mineral of Fe(III)–As(III) suggest that Archaea may play a significant role in the during the winter period and with amorphous Fe(III)– biogeochemical cycling of arsenic (Huber et al. 2000; As(V) the rest of the year (Casiot et al. 2003b; Morin et al. Chaban et al. 2006). 2003). Bacteria play an essential role in the oxidation of Fe Highly acidic environments are relatively scarce world- and As (Casiot et al 2003b). Bacterial diversity is lower wide and are generally associated with mining activities. than in unpolluted water. Sequences related to G. ferrugi- The oxidation by meteoric water of the pyrite-rich wastes nea, a neutrophilic Fe-oxidizing bacterium, are dominant from the abandoned Pb–Zn Carnoule`s mine generates low (Bruneel et al. 2006). pH (2.7–3.4) water containing high concentrations of As The biogeochemical processes that occur in the Car- and Fe, up to 10,000 and up to 20,000 mg l-1, respectively noule`s spoil heaps are more complex than those in the (Casiot et al. 2003a). We previously characterized the creek. The general hydrochemistry and aquifer hydrody- bacterial communities and showed that populations related namics have already been broadly characterized (Koffi to sulfate-reducing bacteria and Gallionella ferruginea et al. 2003; Casiot et al. 2003a). The spoil heaps are cov- seem to play a key role in AMD functioning (Bruneel et al. ered by an impermeable layer of clay which prevents 2005, 2006). To know how a system is structured and how it rainwater percolation from the surface towards the unsat- functions, we first have to address the diversity of the whole urated zone. The aquifer originates from former natural community. We used a molecular phylogenetic approach to springs that were buried under the tailings (Koffi et al. characterize the microbial structure and infer a corre- 2003). Therefore, the primary region of oxidation is located sponding ecosystem function where appropriate. The aim of at the base of the tailing, where the oxygen rich rainwater the present study was to investigate the archaeal community can penetrate directly. The dominant organisms (27–65%) in water samples from an AMD very rich in As, to improve are related to Desulfosarcina variabilis a sulfate-reducing our understanding of the implication of these microorgan- bacterium. Acidithiobacillus ferrooxidans represent the isms in AMD functioning. This is the first molecular second most important group (8–14%). analysis of the archaeal community present in the Carnoule`s Cultivable bacterial strains of A. ferrooxidans and mine system. Thiomonas (shown to be very active in the oxidation of As) were identified both in the tailing stock and in the Reigous Creek (Bruneel et al. 2003). Materials and methods Sampling and analysis Description of the study site Three surveys were carried out in November 2004, April The lead and zinc mine of Carnoule`s, which has been 2005, and September 2005 in the tailing stock. Ground- abandoned since 1963, produced 1.2 MT of spoil material waters were collected in a borehole (S5, between 10 and containing sand, sulfide minerals, heavy metals (Pb, Zn, Tl) 12 m deep) located in the center of the tailings. Samples and metalloids (As, Sb). The material is deposited in the were also taken along the Reigous Creek, (collecting middle of and across the upstream part of a creek (the downstream seepage waters from the surroundings) in Reigous) at the site of its natural spring. The Reigous November 2005, at the spring (S1), 30 m downstream from collects downstream seepage waters from the surroundings the spring (station COWG), 150 m downstream (COWS), before joining, at 1.5 km, the relatively pristine Amous and 1,500 m (CONF) upstream from the confluence river. between the Reigous and the Amous river. Water samples 123 Extremophiles

(300 ml) were filtered through sterile 0.22 lm Nuclepore 16S rRNA gene sequencing filters that were then transferred to cryotubes, frozen in liquid nitrogen, and stored at -80°C until further analysis. Partial sequences of the 16S rRNA gene were determined The main physicochemical parameters [pH, T°C, dis- by the dideoxy nucleotide chain-termination method using solved oxygen (DO), etc.] were measured at the sampling the BigDye 3.1 kit (Applied Biosystems) on an ABI points. Measurements of pH and water temperature were PRISM 3730XL Genetic analyzer (Applied Biosystems). made in the field with an Ultrameter Model 6P (Myron L Sequences were checked for chimeras using the CHI- 125 Company, Camlab, Cambridge). Water samples were MERA CHECK function of the Ribosomal Database immediately filtered through 0.22 lm Millipore mem- Project II (Maidak et al. 2001). DNA sequence analyses branes fitted on Sartorius polycarbonate filter holders. were performed using the BLAST, ALIGNN, and CLU- Samples for total Fe and As determination were acidified to STALW programs (Altschul et al. 1990; Felsenstein 1993; pH = 1 with HNO3 (14.5 M), and stored at 4°C in poly- Thompson et al. 1994). A phylogenetic tree was con- ethylene bottles until analysis. The samples for Fe and As structed using the PHYLIP computer package (Felsenstein speciation and sulfate determination were stored in the dark 1993). The confidence level of the phylogenetic tree and analyzed within 24 h. topology was evaluated by performing 100 bootstrap replications with the SEQBOOK program. All the sequences DNA isolation obtained were submitted to the EMBL databases under accession numbers AM765808 to AM765809 and Genomic DNA was extracted from filtered water using the AM778965 to AM778977. UltraClean Soil DNA Isolation Kit according to the recommendations of the manufacturer (MoBio Laboratories Chemical analysis Inc., USA). All the extracted genomic DNA samples were stored at -20°C until further processing. The determination of total dissolved As was performed by hydride generation atomic fluorescence spectrometry (HG- PCR amplification AFS). Analyses of As species were carried out using coupled anion-exchange chromatography–HG-AFS. This Amplification of archaeal 16S rRNA genes was obtained method, described by Bohari et al. (2001), has a detection using primers Arch21F (50-TTCCGGTTGATCCYGCCG limit of 2.3 nM for As(III) and 6.1 nM for As(V). The GA-30) and Arch958R (50-YCCGGCGTTGAMTCCAA precision is better than 5%. Total dissolved Fe was deter- TT-30) (Delong 1992). The PCR amplifications were per- mined by flame atomic absorption spectrometry. Fe(II) was formed as previously described (Bruneel et al. 2006). The determined using colorimetry at 510 nm after complexa- amount of PCR product was determined by comparison to tion with 1,10-phenanthrolinium chloride solution in known concentrations after migration on agarose gel. buffered samples (pH 4.5) (Rodier et al. 1996). The detection limit is 0.2 lM and the precision better than 5%. Archaeal 16S rRNA gene library analysis The sulfate concentration was determined after precipitation of BaSO4 with BaCl2 and spectrophotometric Archaeal 16S rRNA gene libraries were constructed to measurement at 650 nm. characterize the archaeal populations. Archaeal 16S rRNA genes were amplified with Arch21F and Arch958R prim- Rarefaction analysis, diversity index, and coverage ers. These PCR products were cloned in E. coli TOP 10 values using the pCR2.1 Topo TA cloning kit (Invitrogen, Inc.), according to the manufacturer’s instructions. Cloned 16S PAST (PAleontological STatistics v 1.19) software from rRNA gene fragments were reamplified using the primers the website http://folk.uio.no/ohammer/past/ was used for TOP1 (50-GTGTGCTGGAATTCGCCCTT-30) and TOP2 different diversity indices (Rarefaction analysis, Taxa, (50-TATCTGCAGAATTCGCCCTT-30) located on the Total clones, Singletons, Dominance, Coverage, Shannon, vector and surrounding the inserted PCR fragment, and Equitability, and Simpson) for each clone library. To per- then digested with the enzymes HaeIII or HinfI. Restriction form rarefaction analysis, the total number of clones profiles were analyzed using 2.5% agarose gel electro- obtained compared with the number of clones representing phoresis (small-fragment resolution agarose; QA agarose, each unique phylotype was used to produce the rarefaction QBioge`ne, Inc.). Around 60–70 clones from each library curves. Coverage values were calculated to determine how were analyzed and grouped according to their RFLP pat- efficiently the libraries described the complexity of a terns (HaeIII and HinfI digestion). The sequences of clones theoretical community like an original archaeal commu- from dominant groups were determined. nity. The coverage (Good 1953) value is given as 123 Extremophiles

C = 1 - (n1/N) where n1 is the number of clones that and 20,000 mg l-1 for As and Fe, respectively) in the occurred only once in the library. tailing stock (Casiot et al. 2003a) and from 80 to 350 mg l-1 for As, 750 to 2,700 mg l-1 for Fe, and 2,000 to 7,500 mg l-1 for sulfate in the head waters of the Results Reigous creek (Leblanc et al. 2002).

Aqueous chemistry Composition of archaeal communities

The physicochemical composition of the water is presented 16S rRNA gene library analyses were performed to identify in Table 1. The pH inside the piezometer was between 3.73 the dominant groups of archaeal populations. The most and 5.78. The temperature varied from 15.5 to 20.6°C and representative sequences of the dominant clones are sum- was relatively stable throughout the year (Koffi et al. marized in Table 2 and the phylogenetic filiations of the 2003). The DO was quite low particularly in April 2005 sequences obtained are presented in Fig. 1. DNA could be (between 0.1 and 0.2 mg l-1). The concentration of As extracted from all sampling sites except the S5 borehole in inside the tailing stock varied greatly. As(III) was pre- November 2005. In the Carnoule`s mine drainage, numer- dominant, comprised between 78 and 277 mg l-1, and ous sequences in the libraries are related to sequences As(V) varied between 42 and 66 mg l-1. The concentra- previously found in AMD, showing that the clone libraries tion of Fe(II) (Fe(III) not detected, data not shown) varied were not contaminated. greatly, i.e. between 778 and 1,299, and sulfate between Clones analysis revealed that all the sequences were 3,264 and 4,195 mg l-1. The concentrations of As(III), Fe affiliated to the phylum Euryarchaeota, while Cre- 2- and SO4 were highest in November 2004. narchaeota were not represented. The most abundant In the Reigous creek, the 2.5 pH at the spring increased sequence types present in the water of the tailing (S5) along the creek to reach 3.43 at COWS and 3.25 just before displayed from 99 to 100% homology with Ferroplasma the confluence with the Amous (CONF), 1.5 km away. The acidiphilum strain DR1, that was detected in microbial DO content was 1 mg l-1 in the spring but it increased consortia from AMD and in industrial bioleaching envi- along the creek to reach 5–6 at COWS and 3–4 mg l-1 at ronments (Dopson et al. 2004, AY22042). They were CONF. Dissolved As and Fe concentrations decreased at recovered in the groundwater in November 2004 and April varying degrees along the course of the creek, (30 mg l-1 2005, representing a large majority of the clones (65–72%). for As(III), 39 mg l-1 for As(V), 879 mg l-1 for Fe(II) and The second most abundant group (9% in November 2004 4,388 mg l-1 for sulfate at the spring station (S1) but only but 65% in September 2005) was similar (99–100%) to the 0.53 for As(III and V), 25 mg l-1 for Fe(II) and 749 for uncultured archaeon clone ant h4 (Table 2, Fig. 1) found in sulfate at the CONF station. These elements are removed two anaerobic sludges (DQ462728, unpublished). The by coprecipitation with Fe(III). This process results from sequences representing the second most abundant type in bacterially mediated As- and Fe-oxidation (Casiot et al. April 2005 (15%) were similar (91%) to clones of the 2003b). Furthermore, the increase in pH as a result of uncultured archaeon clone YAC1 (Table 2, Fig. 1) found dilution by unpolluted tributaries after COWG also con- in communities of different hot springs (DQ237924, tributes to an increase in As and Fe precipitation. During unpublished). In September 2005, the second most this sampling period, the concentrations of As, Fe and important group (20%), (Table 2), was related to the 2- SO4 were not particularly high in comparison to the uncultured archaeon clone ASL1 found in AMD (Baker concentrations usually found in these waters (up to 10,000 and Banfield 2003; AF544224).

Table 1 Physico-chemical characteristics of the water (mg l-1) during the sampling in S5, S1, COWG, COWS and CONF

2- Sampling station Sampling period pH (±SD) T (°C) DO (±SD) As(III) (±SD) As(V) (±SD) Fe (II) (±SD) SO4 (±SD)

Tailing S5 November 2004 5.78 (±0.05) 15.5 2 277 (±14) 42 (±2) 1299 (±104) 4195 (±420) stock April 2005 4.05 (±0.05) 17.3 0.1–0.2 128 (±6) 66 (±3) 784 (±62) 3264 (±326) September 2005 3.73 (±0.05) 20.6 4–5 78 (±4) 53 (±3) 778 (±62) 3629 (±363) Reigous S1 November 2005 2.5 (±0.05) 14.6 1 30.0 (±0.8) 39 (±2) 879 (±70) 4388 (±441) Creek COWG 2.74 (±0.05) 10.6 5–6 22.0 (±0.8) 22.0 (±0.8) 501 (±40) 1785 (±182) COWS 3.43 (±0.05) 7.2 5–6 4.5 (±0.2) 1.50 (±0.08) 95 (±8) 902 (±90) CONF 3.25 (±0.05) 6.7 3–4 0.53 (±0.02) 0.53 (±0.02) 25 (±2) 749 (±75) SD Standard deviation

123 Extremophiles

Table 2 Archaeal clones found in Carnoule`s mine drainage with closest match organism or clone name, percent similarity, phylogenetic group, closest relative and percent number of each group compared to the total number of clones Sampling station Sampling period Clones Phylum Closest relative Number of bp Relative (accession number) identical and abundance of % similarity clones (%)a

Tailing stock S5 November 2004 S5Nov04 73 Euryarchaeota F. acidiphilum strain 100 72 DR1 (AY222042) S5Nov04 82 Euryarchaeota Uncultured archaeon 100 9 clone ant h4 (DQ303256) April 2005 S5Apr05 12 Euryarchaeota F. acidiphilum strain 99 65 DR1 (AY222042) S5Apr05 47 99 S5Apr05 45 Euryarchaeota Uncultured archaeon 91 15 clone YAC1 (DQ237924) September 2005 S5Sep05 53 Euryarchaeota Uncultured archaeon 99 65 clone ant h4 (DQ303256) S5Sep05 56 Euryarchaeota Uncultured archaeon 97 20 clone ASL1 (AF544224) Reigous Creek S1 November 2005 S1Nov05 90 Euryarchaeota F. acidiphilum strain 99 54 DR1 (AY222042) S1Nov05 58 Euryarchaeota Uncultured archaeon 93 21 clone YAC1 (DQ237924) COWG CGNov05 19 Euryarchaeota Uncultured archaeon 93 59 clone YAC1 (DQ237924) CGNov05 94 93 CGNov05 32 Euryarchaeota F. acidiphilum strain 100 4 DR1 (AY222042) COWS CSNov05 10 Euryarchaeota Uncultured archaeon 92 93 clone YAC1 (DQ237924) CSNov05 20 Euryarchaeota Uncultured archaeon 99 6 clone ant g10 (DQ303253) CONF CFNov05 6 Euryarchaeota Uncultured archaeon 94 74 clone YAC1 (DQ237924) a The abundance of clones was calculated for each library

In the Reigous creek during the sampling campaign in related (99% similarity) to the uncultured archaeon clone November 2005, the most abundant group (21% at the ant g10 isolated in macroscopic filaments from an extre- spring S1, 59% at COWG, 93% at COWS and 74% at mely acidic environment, Tinto River (DQ303253, CONF) was related (92–94%) to the uncultured archaeon unpublished). Phylogenetic analyses (Fig. 1) did not enable clone YAC1. These clones were found in low abundance affiliation of the clone sequences with any representative of (15%) in the groundwater and only in April 2005. The the subdivision. The closest relative (91%) was Thermo- second most abundant group in the creek was similar (99– plasma sp. SO2 (AB262009, unpublished). 100%) to F. acidiphilum, also numerically significant members in Carnoule`s tailing stock. The abundance of this Rarefaction analysis, diversity index and coverage group decreased along the creek, representing 54% of the values of the clone libraries analyzed clones at the spring S1, but only 4% at COWG and was undetected at COWS and CONF. The least abundant Table 3 shows Dominance, Shannon, Equitability, Simp- sequences (6%) found only at the COWS station was son index and Coverage values calculated for each library. 123 Extremophiles

Fig.1 Phylogenetic analysis of S5Apr05 12 (AM778965) 16S rRNA gene sequences S1Nov05 90 (AM778970) afﬁliated with Archaea CGNov05 32 (AM778974) members from the AMD of Carnoule`s (France). Clone Ferroplasma acidiphilum strain DR1 (AY222042) names in bold correspond to Ferroplasma acidiphilum strain YT DSM 12658T (AJ224936) sequences found in the Uncultured archaeon ASL32 (AF544222) 57 Carnoule`s mine drainage S5Nov04 73 (AM765808) Uncultured archaeon ant c8 (DQ303251) S5Apr05 47 (AM778966) Ferroplasma acidarmanus (AF145441) 76 Uncultured archaeon ant c7 (DQ303250) Ferroplasma sp. MT17 (AF513710) Uncultured archaeon ant h10 (DQ303255)

97 S5Sep05 53 (AM778968) S5Nov04 82 (AM765809) 63 Uncultured archaeon ant h4 (DQ303256) Ferroplasma sp. JTC3 (AY830840) 76 Uncultured archaeon MS14 (AF232925) 71 Ferroplasma cyprexacervatum (AY907888) Thermoplasma volcanium (AF339746) Thermoplasma sp. S02 (AB262009) 29 Uncultured archaeon ASL1 (AF544224) 75 Uncultured archaeon ARCP1-28 (AF523940) 38 29 Uncultured archaeon ant g4 (DQ303254) S5Sep05 56 (AM778969) 17 Uncultured archaeon ant g10 (DQ303253)

65 CSNov05 20 (AM778976) 99 Uncultured archaeon AS1 (AF544219) Uncultured archaeon ant b7 (DQ303249) 53 Unidentified archaeon pISA42 (AB019742) Uncultured euryarchaeote pLM14A-1 (AB247822) Uncultured Thermoplasmatales archaeon OPPD020 (AY861955) Uncultured archaeon YAC1 (DQ237924) 50 CFNov05 6 (AM778977) 66 49 CGNov05 94 (AM778973)

95 S1Nov05 58 (AM778971) CGNov05 19 (AM778972)

37 8 S5Apr05 45 (AM778967) 16 CSNov05 10 (AM778975) Archaeoglobus fulgidus strain VC-16 (X05567) Sulfolobus solfataricus (D26490) 50 Sulfurisphaera ohwakuensis DSM 1242T (D85507) 96

76 Metallosphaera hakonensis (D86414) T 53 Acidianus infernus DSM 3191 (D85505) 61 Acidianus ambivalens DSM 3772T (D85506) Uncultured archaeon PMA5 (DQ399817) Uncultured archaeon ZAR100 (AY341269) Acidithiobacillus caldus (X72851)

0.05

To estimate diversity coverage and to determine whether a 0.93). In November 2005, the COWS library showed lower sufﬁcient number of clones from each library had been diversity indices (Shannon: 0.5704; Simpson: 0.2397) than sequenced, rarefaction analysis was performed. The gen- the other libraries (Shannon ranging from 1.151 to 1.604; erated curves were near saturation (data not shown), Simpson from 0.4488 to 0.6545). Inversely, in November consistent with the high coverage values (between 0.82 and 2005, the COWS library presented a higher Dominance

123 Extremophiles

Table 3 Diversity indices calculated for the seven clone libraries from different stations in Carnoule`s mine drainage Clone library Taxa Total clones Singletons Dominance Coverage(C) Shannon (H) Equitability Simpson (1-D)

S5 November 2004 9 66 5 0.4913 92 1.151 0.5241 0.5087 S5 April 2005 11 66 6 0.4564 90 1.277 0.5323 0.5436 S5 September 2005 10 61 6 0.4512 90 1.234 0.5361 0.5488 S1 November 2005 14 57 10 0.3456 82 1.604 0.6077 0.6545 COWG November 2005 13 69 9 0.3989 86 1.424 0.5552 0.6011 COWS November 2005 6 61 4 0.7603 93 0.5704 0.3183 0.2397 CONF November 05 12 61 6 0.5512 90 1.189 0.4785 0.4488 index (0.7603) than the other libraries (from 0.3456 to in the world. Furthermore, some strains of this genus like 0.5512). Ferroplasma acidarmanus Fer1 was shown to be an arsenic-hypertolerant acidophilic archaeon (Gihring et al. 2003; Baker-Austin et al. 2007). This strain, isolated from Discussion the Iron Mountain mine, California, was able to grow with up to 10 g arsenate per litre but his growth was reduced In the AMD site of Carnoule`s, more than 65% of the ar- with 5 and 10 g of arsenite per litre. This population, which chaeal sequences could not be closely related to cultured is more acid-resistant than iron- and sulfur-oxidizing bac- organisms, suggesting that they may constitute new taxa. teria, is in fact known to mobilize metals from sulfide ores, Only sequences close to F. acidiphilum were related to e.g. pyrite, arsenopyrite and copper-containing sulfides. cultured organisms. Rarefaction data and percent coverage According to Golyshina and Timmis (2005) Ferroplasma calculations suggested that the archaeal 16S rRNA gene spp. are probably the major players in the biogeochemical libraries reach saturation. cycling of sulfur and sulfide metals in highly acidic envi- Whatever the sampling period, the water of S5 inside the ronments, and may have considerable potential for tailing stock, where intensive pyrite oxidation takes place, biotechnological applications such as biomining and bio- was numerically dominated by sequences clearly related to catalysis under extreme conditions. These results are F. acidiphilum, or to the uncultured clone ant h4 which consistent with those of Edwards et al. (2000) at the Iron showed more than 98% similarity with F. acidiphilum. Mountain acid-generating site (United State), where the This isolate was an acidophilic, mesophilic, ferrous-iron microbial community is dominated (85%) by an archaeon oxidizing, cell-wall lacking microbe that became the basis of the genus Ferroplasma. For these authors, the presence of a new archaeal lineage: the new genus Ferroplasma of this population and other closely related Thermoplas- within the new family Ferroplasmaceae, in the order matales suggests that these acidophiles are important Thermoplasmatales, which includes the families Thermo- contributors to acid mine drainage and may substantially plasmaceae and Picrophilaceae (Golyshina and Timmis impact iron and sulfur cycles. The growth of F. acidiphi- 2005). These two populations represented 81% of clones in lum occurs between 20 and 45°C with an optimum at 35°C November 2004, 65% in April 2005, and 65% in and at pH 1.3–2.2 with an optimum at pH 1.7 (Golyshina September 2005. Previous analysis of the bacterial com- et al. 2000). Surprisingly, we detected this population in a munity in the Carnoule`s tailing showed that the dominant less acidic environment (3.73–5.7). Isolation and charac- population was related to the sulfate-reducing bacteria terization of members of this population are needed to Desulfosarcina variabilis (Bruneel et al. 2005). This pop- determine their physiological capabilities especially at the ulation could not clearly explain the leaching of the pH range found in Carnoule`s waters. Carnoule`s tailing as it is well known that it was mostly The clone sequences from the Reigous Creek were acidophilic ferrous iron-oxidizing microorganisms that related to the same groups detected in the tailing S5 but the were found to be involved in the production of acid mine abundance of each varied. The dominant population in the drainage (Baker and Bandfield 2003). Iron oxidizing bac- Reigous Creek (21% of total clones at the spring S1, 59% teria like A. ferrooxidans and Sulfobacillus spp. were also at COWG, 93% at COWS and around 74% CONF) was present in the Carnoule`s mine tailing but represented a related to the uncultured archaeon clone YAC1 found in minor population (Bruneel et al. 2005). Thus, F. acidiph- communities in different hot springs. Phylogenetic analy- ilum could explain the intensive leaching observed in the ses (Fig. 1) did not enable affiliation of the clone sequences Carnoule`s tailing and the high concentration of As, up to with any cultured representative of the subdivision and this 10,000 mg l-1, one of the highest concentrations reported clone could thus represent a new species. The closest

123 Extremophiles relative (91%) was the uncultured Thermoplasmatales Acknowledgments The study was financed by the EC2CO pro- archaeon found in the Yellowstone geothermal ecosystem gramme (Institut National des Sciences de l’Univers, CNRS). We thank Marjorie Cloez for identification of the archaeal population in (Spear et al. 2005). The order Thermoplasmatales includes the site, and Marie Ange Cordier for assistance in analysis of the families Ferroplasmaceae, Thermoplasmaceae and physical–chemical parameters. Picrophilaceae (Golyshina and Timmis 2005). The known members of the Thermoplasmales are all acidophilic. Some groups, like the family Ferroplasmacea within this order, References are capable of iron oxidation (Edwards et al. 2000; Go- lyshina and Timmis 2005). A previous study of bacterial Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215:403–410 populations in the Carnoule`s creek showed that the domi- Baker BJ, Banfield JF (2003) Microbial communities in acid mine nant bacterial population was related to G. ferruginea,a drainage. FEMS Microbiol Ecol 44:139–152 neutrophilic bacterium that oxidizes Fe (Bruneel et al. Baker-Austin C, Dopson M, Wexler M, Sawers RG, Stemmler A, 2006). Consistent with previous observations demonstrat- Rosen BP, Bond PL (2007) Extreme arsenic resistance by the acidophilic archaeon ‘‘Ferroplasma acidarmanus’’ Fer1. Ex- ing that G. ferruginea efficiently remove As (III and V) in tremophiles 11:425–434 water by coprecipitation with Fe (Katsoyiannis and Zou- Bohari Y, Astruc A, Astruc M, Cloud J (2001) Improvements of boulis 2004), this population may play a key role in the hydride generation for the speciation of arsenic in natural remediation process observed in the Reigous creek (Casiot freshwater samples by HPLC-HG-AFS. J Anal At Spectrom 16:774–778 et al. 2003b). If the uncultured archaeon clone YAC1 Bond PL, Smriga SP, Banfield JF (2000) Phylogeny of microoor- oxidizes Fe, this population could play a role in the natural gansims populating a thick, subaerial, predominantly lithotrophic remediation processes occurring in the Reigous Creek in biofilm at an extreme acid mine drainage site. Appl Environ association with G. ferruginea, but until the archaeal Microbiol 66:3842–3849 Bruneel O, Personne´ J-C, Casiot C, Leblanc M, Elbaz-Poulichet F, strains are isolated, their physiological role in the creek Mahler BJ, Le Fle`che A, Grimont PAD (2003) Mediation of ecology will remain uncertain. Environmental genome data arsenic oxidation by Thiomonas sp. in acid mine drainage like those obtain with analysis of assembled random (Carnoule`s, France). J Appl Microbiol 95:492–499 shotgun sequence data can also provide detailed insight Bruneel O, Duran R, Koffi K, Casiot C, Fourçans A, Elbaz-Poulichet F, Personne´ J-C (2005) Microbial diversity in a pyrite-rich into the metabolic potential of uncultivated organisms tailings impoundment (Carnoule`s, France). Geomicrobiol J (Tyson et al. 2005). 22:249–257 Our study demonstrated the existence of a complex Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F, Personne´ JC (2006) prokaryotic community in the Carnoule`s AMD where Diversity of acidophilic microorganisms involved in Fe–As immobilisation and oxidation in Carnoule`s acid mine drainage bacterial and archaeal populations are present. Both phyl- (France). Appl Environ Microbiol 72:551–556 otype communities were significantly altered in terms of Casiot C, Leblanc M, Bruneel O, Personne´ J-C, Koffi K, Elbaz- size and structure with microhabitats varying inside the Poulichet F (2003a) Geochemical processes controlling the AMD particularly in underground water from the tailing formation of As-rich waters within a tailings impoundment. Aquatic Geochem 9:273–290 and in the Reigous and the small creek draining the site. Casiot C, Morin G, Juillot F, Bruneel O, Personne´ JC, Leblanc M, The occurrence of different dominant communities is likely Duquesne K, Bonnefoy V, Elbaz-Poulichet F (2003b) Bacterial associated with the formation of environmental gradients immobilization of arsenic in acid mine drainage (Carnoule`s of temperature, pH, oxidation–reduction potential, etc. creek, France). Water Res 37:2929–2936 Chaban B, Ng SYM, Jarrell KF (2006) Archaeal habitats, from the Other methods such as fluorescence in situ hybridization extreme to the ordinary. Can J Microbiol 52:73–116 (FISH) will help to clearly assess the relative proportion of Delong EF (1992) Archaea in coastal marine environments. Proc Natl population. However, this method has not been widely Acad Sci USA 89:5685–5689 applied to samples of thermophilic archaea and may be Dopson M, Baker-Austin C, Hind A, Bowman JP, Bond PL (2004) Characterization of Ferroplasma isolates and Ferroplasma limited by cross-hybridization. Furthermore, methods such acidarmanus sp. nov., extreme acidophiles from acid mine as metagenomic research (study of the entire genetic drainage and industrial bioleaching environments. Appl Environ composition of communities of an environment) could help Microbiol 70:2079–2088 to study the total diversity, physiology, ecology and phy- Edwards KJ, Bond PL, Gihring TM, Banfield JF (2000) An archaeal iron-oxidizing extreme acidophile important in acid mine logeny of microbial population but all of the approaches drainage. Science 287:1796–1799 that are available today have advantages and limitations Felsenstein J (1993) PHYLIP-Phylogeny Inference Package (version (Pontes et al. 2007). Only, the isolation of archaeal strains 3.5c). Department of Genetics, University of Washington, at the Carnoule`s mine will extend our understanding of the Seattle Gihring TM, Bond PL, Peters SC, Banfield JF (2003) Arsenic ubiquity of archaea in such environments, and help eluci- resistance in the archaeon ‘‘Ferroplasma acidarmanus’’: new date the microbial component driving the biogeochemical insights into the structure and evolution of the ars genes. processes present in this and other extreme AMD sites. Extremophiles 7:123–130

123 Extremophiles

Golyshina OV, Timmis KN (2005) Ferroplasma and relatives, Geological Society of London Special Publication, Kluwer recently discovered cell wall-lacking Archaea making a living Academic Publishers, Dordrecht, pp 464 in extremely acid, heavy metal-rich environments. Environ Maidak BL, Cole JR, Lilburn TG, Parker CT, Jr Saxman PR, Farri RJ Microbiol 7:1277–1288 (2001) The RDP-II (Ribosomal Database Project). Nucleic Acids Golyshina OV, Pivovarova TA, Karavaiko GI, Kondrateva TF, Moore Res 29:173–174 ERB, Abraham WR, Lunsdorf H, Timmis KN, Yakimov MM, Morin G, Juillot F Casiot C Bruneel O Personne´ J-C, Elbaz-Poulichet Golyshin PN (2000) Ferroplasma acidiphilum gen. nov., sp nov., F, Leblanc M, Ildefonse P, Calas G (2003) Bacterial formation of an acidophilic, autotrophic, ferrous-iron-oxidizing, cell-wall- tooeleite and mixed arsenic(III) or arsenic(V)-iron(III) gels in lacking, mesophilic member of the Ferroplasmaceae fam. nov., the Carnoule`s acid mine drainage, France. A XANES, XRD, and comprising a distinct lineage of the Archaea. Int J Syst Evol SEM study. Environ Sci Technol 37:1705–1712 Microbiol 50:997–1006 Oremland RS, Stolz JF (2003) The ecology of arsenic. Science Good IJ (1953) The population frequencies of species and the 300:939–944 estimation of the population parameters. Biometrika 40:237–264 Pontes DS, Lima-Bittencourt CI, Chartone-Souza E, Nascimento Huber R, Sacher M, Vollmann A, Huber H, Rose D (2000) AMA (2007) Molecular approaches: advantages and artifacts in Respiration of arsenate and selenate by hyperthermophilic assessing bacterial diversity. J Ind Microbiol Biotechnol 34:463– Archaea. Syst Appl Microbiol 23:305–14 473 Johnson DB (1998) Biodiversity and ecology of acidophilic micro- Rodier J, Broutin JP, Chambon P, Champsaur H, Rodier L (1996) organisms. FEMS Microbiol Ecol 27:307–317 L’analyse de l’eau. Dunod, Paris, p 1383 Johnson DB, Hallberg KB (2003) The microbiology of acidic mine Spear JR, Walker JJ, McCollom TM, Pace NR (2005) Hydrogen and waters. Res Microbiol 154:466–473 bioenergetics in the Yellowstone geothermal ecosystem. Proc Katsoyiannis IA, Zouboulis AI (2004) Application of biological Natl Acad Sci USA 102:2555–2560 processes for the removal of arsenic from groundwater. Water Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: Res 38:17–26 improving the sensitivity of progressive multiple sequence Koffi K, Leblanc M, Jourde H, Casiot C, Pistre S, Gouze P, Elbaz- alignment through sequence weighting, position-specific gap Poulichet F (2003) Reverse oxidation zoning in mine tailings penalties and weight matrix choice. Nucleic Acids Res 22:4673– generating arsenic-rich acidic waters (Carnoule`s, France). Mine 4680 Water Environ 22:7–14 Tyson GW, Lo I, Baker BJ, Allen EE, Hugenholtz P, Banfield JF Langmuir D (1997) Acid mine waters. In: McConnin R (eds) (2005) Genome-directed isolation of the key nitrogen fixer Aqueous environmental geochemistry. Prentice-Hall, New Jer- Leptospirillum ferrodiazotrophum sp. nov. from an acidophilic sey, pp 457–478 microbial community. Appl Environ Microbiol 71:6319–6324 Leblanc M, Casiot C, Elbaz-Poulichet F, Personne´ C (2002) Arsenic Zettler LAA, Go´mez F, Zettler E, Keenan BG, Amils R, Sogin ML removal by oxidizing bacteria in a heavily arsenic contaminated (2002) Eukaryotic diversity in spains river of fire. Nature acid mine drainage system (Carnoule`s, France). In: Younger PL, 417:137 Robins NS (eds) Mine water hydrogeology and geochemistry.

123 Microb Ecol (2011) 61:793–810 DOI 10.1007/s00248-011-9808-9

ENVIRONMENTAL MICROBIOLOGY

Characterization of the Active Bacterial Community Involved in Natural Attenuation Processes in Arsenic-Rich Creek Sediments

Odile Bruneel & Aurélie Volant & Sébastien Gallien & Bertrand Chaumande & Corinne Casiot & Christine Carapito & Amélie Bardil & Guillaume Morin & Gordon E. Brown Jr & Christian J. Personné & Denis Le Paslier & Christine Schaeffer & Alain Van Dorsselaer & Philippe N. Bertin & Françoise Elbaz-Poulichet & Florence Arsène-Ploetze

Received: 7 August 2010 /Accepted: 20 January 2011 /Published online: 12 February 2011 # Springer Science+Business Media, LLC 2011

Abstract Acid mine drainage of the Carnoulès mine with Fe(III), in agreement with previous studies, which (France) is characterized by acid waters containing high suggest a role of microbial activities in the co-precipitation concentrations of arsenic and iron. In the first 30 m along of As(III) and As(V) with Fe(III) and sulfate. To investigate the Reigous, a small creek draining the site, more than 38% how this particular ecosystem functions, the bacterial of the dissolved arsenic was removed by co-precipitation community was characterized in water and sediments by

Electronic supplementary material The online version of this article (doi:10.1007/s00248-011-9808-9) contains supplementary material, which is available to authorized users. O. Bruneel (*) : A. Volant : C. Casiot : A. Bardil : G. E. Brown Jr C. J. Personné : F. Elbaz-Poulichet Surface and Aqueous Geochemistry Group, Laboratoire HydroSciences Montpellier, Department of Geological and Environmental Sciences, UMR5569 (CNRS-IRD-Universités Montpellier I et II), Stanford University, Université Montpellier II, CC MSE, Stanford, CA 94305-2115, USA Place Eugène Bataillon, 34095 Montpellier Cedex 05, France e-mail: [email protected] G. E. Brown Jr B. Chaumande : P. N. Bertin : F. Arsène-Ploetze Stanford Synchrotron Radiation Laboratory, SLAC, Génétique Moléculaire, 2575 23 Sand Hill Road, MS 69, Génomique Microbiologie, Menlo Park, CA 94025, USA UMR7156, Université de Strasbourg/CNRS, 28 rue Goethe, 67083 Strasbourg Cedex, France D. Le Paslier S. Gallien : C. Carapito : C. Schaeffer : A. Van Dorsselaer Génomique Métabolique, UMR8030, CNRS, Laboratoire de Spectrométrie de Masse Bio-organique, 2 rue Gaston Crémieux, Institut Pluridisciplinaire Hubert Curien, 91057 Evry Cedex, France UMR7178 (CNRS-Université de Strasbourg), 25 rue Becquerel, 67087 Strasbourg, France D. Le Paslier G. Morin Commissariat à l’Energie Atomique (CEA), Institut de Minéralogie et de Physique des Milieux Condensés Direction des Sciences du Vivant, Institut de Génomique, (IMPMC), UMR7590 (CNRS - Universités Paris 6&7 - IPGP), Genoscope, Laboratoire de Génomique Comparative, 140, rue de Lourmel, 2 rue Gaston Crémieux, 75015 Paris, France 91057 Evry Cedex, France 794 O. Bruneel et al.

16S rRNA encoding gene library analysis. Based on the However, to be able to develop remediation processes and/or results obtained using a metaproteomic approach on sedi- optimize existing processes, further knowledge is required ments combined with high-sensitivity HPLC-chip spec- about how these bacteria function in situ. In particular, it is trometry, several GroEL orthologs expressed by the of crucial importance to determine which bacteria are viable community were characterized, and the active members of andactiveinsuchecosystems. the prokaryotic community inhabiting the creek sediments The AMD of Carnoulès mine in Southern France is a were identified. Many of these bacteria are β-proteobacteria highly suitable site for analyzing how microorganisms such as Gallionella and Thiomonas,butγ-proteobacteria contribute to the transformation of metals and metalloids in such as Acidithiobacillus ferrooxidans and α-proteobacteria situ, since efficient natural remediation processes are known such as Acidiphilium, Actinobacteria,andFirmicutes were to occur at this site [12, 18]. This former mine generated also detected. around 1.2 Mt of tailings containing 0.7% Pb, 10% FeS2, and 0.2% As. Water percolating through the tailings forms the head of the Reigous creek. This creek is acidic (pH Introduction around 3) and highly contaminated with As (100 to 350mgL−1). The behavior of As and Fe in the Reigous Acid Mine Drainage (AMD) is one of the most serious forms creek has been intensively studied [18, 23, 48]. In the creek of water pollution in industrial and post-industrial areas spring, As(III) is the main As species present and Fe occurs worldwide [38]. AMD is generated when the wastes from in the form of Fe(II) [18]. Along the first 30 m of the creek the mining and processing of sulfide ores (such as pyrite or (about 1 h residence time), the oxidation of Fe(II) leads to arsenopyrite) come into contact with oxygenated water [5]. the co-precipitation of more than 38% of the dissolved As AMD is often characterized by pH values of 2–4. Such [18, 23]. Arsenic accounts for up to 22% of the total dry waters generally contain high levels of iron, toxic metals weight of the sediments formed along the first 10 m along (such as aluminum, manganese, lead, cadmium, and zinc), the creek. In the wet season, approximately 30 m down- and metalloids (arsenic) [5, 32, 48]. AMD can still occur stream of the spring, these sediments are mainly composed hundreds of years after mine closure and tens of thousands of of As(III)–Fe(III) oxyhydroxysulfates, whereas As(V)–Fe kilometers of groundwater, streams, lakes, and estuaries (III) oxyhydroxysulfates compounds predominate during the throughout the world have been directly impacted [40]. In dry season [48]. Because of the very high molar As/Fe ratio several cases of AMD, natural remediation has been (up to 0.3) existing in the dissolved phase of the Carnoulès observed, as for example at the Carnoulès site in France creek, the mineralogical content of the sediments differs and the Rio Tinto site in Spain [18, 53]. In such AMD, toxic significantly from that classically observed at most AMD compounds are accumulated in sediments consisting of a [47], especially along the first 50 m of the creek. Several variety of iron (oxyhydr)oxides and hydroxysulfates such as strains of Thiomonas and Acidithiobacillus ferrooxidans jarosite, schwertmannite, and ferrihydrite [48]. Natural have been isolated from Reigous creek waters, and based remediation of metal pollutants is generally due to the on the results of laboratory experiments, it has been occurrence of abiotic reactions and/or microbial activities suggested that these bacteria may contribute to the oxidation that make these toxic compounds insoluble and lead them to of Fe(II) and the co-precipitation of As [14, 18, 21–23, 48]. accumulate in sediments [32, 40]. This toxic compound Preliminary analyses have shown that the DNAs of both precipitation processes mainly involve the oxidation and bacteria are present in the Reigous creek, as well as that of precipitation of iron, which is often the main soluble metal Gallionella sp., Thiobacillus sp., and some sulfate-reducing present in AMD, and the adsorption of other metals and bacteria [12]. Archaea have also been found to occur in the metalloids by the ferric minerals formed [32, 51]. Indeed, Reigous creek (Ferroplasma acidiphilum and sequences many elements such as Sr, Cs, Pb, U, and As show a strong affiliated to uncultured Thermoplasmatales archaeon) as well affinity for solid iron oxide [18, 27, 48]. Abiotic oxidation of as a eukaryotic microorganism, Euglena mutabilis [12, 13, Fe(II) proceeds very slowly in acidic (pH 3.5) waters [51]. In 15]. However, the bacterial population inhabiting the As-rich contrast, iron-oxidizing bacteria catalyze the reaction and Reigous sediments has never been characterized so far. It, thus accelerate the formation of solid iron oxide [32, 41, 51]. therefore, seemed to be necessary not only to identify the In addition, several bacteria contribute to the immobilization bacteria present in this creek but also to determine which of arsenic via their ability to oxidize this metalloid [6, 14, 18, members of this community are viable and, therefore, 22, 48], arsenate (As(V)) being adsorbed more strongly than perform metabolic activities in situ. arsenite (As(III)) by Fe(III) oxides and hydroxides at acidic The aim of this study was to describe the bacterial pH levels [10]. Thiomonas strains show a high arsenite populations occurring in both the sediments and waters at oxidation capacity, and these metabolic activities have been the disused Carnoulès site and to identify the bacteria at extensively analyzed under laboratory conditions [6, 14, 22]. work. For this purpose, three complementary approaches Active Bacteria in Arsenic-Rich Sediments 795 were used. First, chemical and mineralogical studies were samples) was used to check analytical accuracy and performed in order to determine the arsenic species present. precision. The results showed that the recovery rate A 16S rRNA encoding gene library was then analyzed in obtained was within ±5%. order to identify the bacterial population present in the Analyses of inorganic arsenic species (As(III), As(V)) creek sediments and waters. Lastly, based on the findings were carried out using anion-exchange chromatography obtained using a metaproteomic approach combined with (25 cm×4.1 mm i.d. Hamilton PRP-X100 column with high-sensitivity mass spectrometry methods, the active Varian ProStar gradient solvent delivery system) coupled to species inhabiting the sediments were identified. a hydride generation (VGS 200, FISONS, France) with an atomic fluorescence spectrometry detector (Excalibur, PSAnalytical, GB) [17]. The detection limit obtained was Methods 172 ng L−1 for As(III) and 458 ng L−1 for As(V), with a precision better than 5%. Total dissolved Fe was deter- Sampling and Analysis mined by flame atomic absorption spectrometry. Fe(II) concentration was determined using colorimetry at 510 nm Samples were collected from Reigous creek in April 2006 after complexation with 1,10-phenanthrolinium chloride at COWG station located 30 m downstream from the solution in buffered samples (pH 4.5) [50] (detection limit: spring. This sampling was part of a long-term monitoring of 11 μgL−1; precision better than 5%). Sulfate concentration the physicochemistry of the Reigous Creek water [23]. The was determined after precipitation of BaSO4 with BaCl2 main physicochemical parameters (pH, temperature, and and spectrophotometric measurement at 650 nm [50]. dissolved oxygen concentrations) were measured in situ at this sampling point. The 5-cm deep sediments on the Solid Sample Characterization bottom of the creek and a thin column (less than 10 cm) of running water covering the sediments were sampled. Solid XAFS data were gathered on the laboratory samples and samples were removed with a sterile spatula from the the sample taken at COWG on April 2006, at 10 K in surface of the sediments. Water samples (300 ml) were transmission mode on a bending magnet D44 at the LURE immediately filtered through 0.22 μm Millipore membranes synchrotron (Orsay, France), and in fluorescence mode on fitted on Sartorius polycarbonate filter holders (for water the 11-2 wiggler beamline at SSRL (Stanford, CA), chemical analysis) or through sterile 0.22-μm Nucleopore respectively. Experiments and data reduction were previ- filters that were then transferred to a collection tube (Nunc), ously reported [48, 49]. frozen in liquid nitrogen, and stored at −80°C until DNA extraction (for 16S rRNA encoding gene analysis). Sam- DNA Isolation, 16S rRNA Encoding Gene Cloning, pling was repeated three times. For total Fe and As Restriction Analysis, and Sequencing determination, filtered water was acidified to pH=1 with

HNO3 (14.5 M) and stored at 4°C in polyethylene bottles Genomic DNA was extracted in triplicate from filtered until analysis. For As and Fe speciation, a 10 μl aliquot of water and sediments using the UltraClean Soil DNA filtered sample water was added to either 0.5 ml of 5% (v/v) Isolation Kit according to the manufacturer’s recommenda- 0.25 M EDTA solution for As speciation [7] or a mixture of tions (MoBio Laboratories Inc., USA). These triplicates were 0.5 ml acetate buffer (pH 4.5) and 1 ml of 1,10- pooled before PCR amplification. All the genomic DNA phenanthrolinium chloride solution for Fe speciation [50]. samples extracted were stored at −20°C until further process- The vials were completed to 10 ml with deionized water. ing. Bacterial diversity was analyzed by cloning PCR The samples used for arsenic speciation and Fe(II) and amplified 16S rRNA encoding genes. Bacterial 16S rRNA sulfate determination were stored in the dark and analyzed encoding genes were amplified with 8F (5′-AGAGTTTGA within 24 h. TCCTGGCTCAG-3′) and 1489R primers (5′-TACCTTGT TACGACTTCA-3′)[43, 57], as previously described [12]. Chemical Analysis These PCR products were cloned into Escherichia coli TOP 10 strain using the pCR2.1 Topo TA cloning kit (Invitrogen, The determination of total dissolved As was performed by Inc.). Cloned 16S rRNA encoding gene fragments were re- ICP-MS using Thermo X7 series with a conventional amplified using the primers TOP1 (5′-GTGTGCT external calibration procedure. Indium was used as internal GGAATTCGCCCTT-3′) and TOP2 (5′-TATCTGCAGA standard to correct for instrumental drift and possible ATTCGCCCTT-3′) that anneal to the vector and surround matrix effects. It was not necessary to correct interference the inserted PCR fragment and then digested with HaeIII or with chloride because of the extremely high As levels HinfI enzymes. Restriction profiles were analyzed using present. Certified reference material SLRS-4 (freshwater 2.5% agarose gel electrophoresis (small fragment resolution 796 O. Bruneel et al. agarose; QA agarose, QBiogène, Inc.). Around 200 clones were washed in 10 mL of saline buffer and agitated from each library were analyzed and grouped according overnight at 4°C. After 10 min of decantation, 7.5 mL of to the RFLP patterns obtained. PAST (Paleontological supernatant were added without mixing to 17.5 mL of STatistics v 1.19) software from the website http://folk. Nycodenz solution (Axis-Shield, Dundee, Scotland), and uio.no/ohammer/past/ was used to calculate different then centrifuged for 30 min at 10,000×g. The cellular diversity indices (rarefaction analysis, taxa, total clones, fraction (nycodenz/sample interface) was removed and singletons, dominance, coverage, shannon, equitability, washed by adding two volumes of NaCl 0.9% and and simpson, Table 1) for each clone library. The total centrifuged for 15 min at 10,000×g at 4°C. Proteins were number of clones obtained compared with the number of extracted from this cellular fraction as previously described clones representing each unique phylotype was used to [58], further purified using the 2-D Clean-up kit (GE produce the rarefaction curves. Coverage values were Healthcare), and resuspended in rehydration buffer calculated to determine how efficiently the libraries described (364gL−1 thiourea, 1,000 g L−1 urea;25gL−1 CHAPS, the complexity of a theoretical community such as an original 0.6% v/v IPG buffer Pharmalyte, 10 g L−1 of DTT, and bacterial community. The coverage [29] value is given as C= 0.01% bromophenol blue). Protein concentrations were

1—(n1/N)wheren1 is the number of clones that occurred quantified using the 2-D Quant kit (GE Healthcare). These only once in the library. Rarefaction analysis showed that the proteins were separated by 2-D gel electrophoresis as curves generated were near saturation (data not shown) and previously described [58] and finally stained with silver consistent with the high coverage values of the two clone nitrate. Gels were analyzed using an Image Scanner, libraries (97.8 for the sediment and 98.6 for the water). This LabScan v 3.0 (GE Healthcare), and the ImageMaster indicated that the clone libraries were sufficiently sampled. 2D platinum software program (v. 6.01, GE Healthcare). Partial sequences of the clones from dominant groups The spots selected were cut out of the 2-D gels and stored at were determined by the dideoxy nucleotide chain- −20°C. Eighty-one spots were analyzed by performing termination method using the BigDye 3.1 kit (Applied nanoLC-Chip-MS/MS. Biosystems) on an ABI PRISM 3730XL Genetic analyzer (Applied Biosystems). The PINTAIL program [4] was used In-Gel Digestion, Mass Spectrometry Analysis, and Protein to check the presence of chimerae. Sequences were also Identification examined manually for chimerae, which were excluded from further analyses. These sequences were compared Unless otherwise specified, all chemicals were obtained from with known sequences (NCBI database) using the BLAST, Sigma (St. Louis, MO, USA). In-gel digestion of gel spots ALIGN, and CLUSTALW programs [1, 26, 55]. All was performed as previously described [58]. The resulting sequence data obtained were submitted to the EMBL peptides were analyzed by performing nanoLC-MS/MS on databases under accession numbers (FR676963-FR677013; an Agilent 1100 Series HPLC-Chip/MS system (Agilent AM988784-AM988794; AM988796; AM988798; Technologies, Palo Alto, USA) coupled to an HCT Ultra ion AM988801-AM988805; AM988807-AM988809). trap (Bruker Daltonics, Bremen, Germany). The MS/MS data were analyzed using the MASCOT 2.2.0 algorithm Preparation of Protein Extracts and Gel Electrophoresis giving a maximum of one missed cleavage, with a mass Analysis tolerance of 0.5 Da for MS and MS/MS data and carbamidomethylation of cysteines and oxidation of methio- Prokaryotes were separated from sediments and eukaryotes nines were specified as the variable modifications. MS/MS using a Nycodenz density gradient. It should be noted that data searches were performed against two in-house generated the main population determined from the DNA directly databases. The first database was composed of the protein extracted from sediments was similar to that identified after sequences of all the organisms related to the groups Nycodenz treatment (data not shown), which suggests that identified by performing 16S rRNA encoding gene analysis this treatment did not result in enrichment of the sample by on the Reigous creek sediments and water (α-, β-, δ-, any particular microorganisms. Ten grams of sediments and γ-proteobacteria, Bacilli, Clostridia, Actinobacteria,

Table 1 Diversity indices calculated from the two clone libraries in sediment and water at the Carnoulès mine drainage creek

Clone library Taxa Total no. of clones Singletons Dominance (D) Coverage (C) Shannon (H) Equitability Simpson (1-D)

Sediments 15 229 5 0.5503 97.8 1.146 0.4231 0.4497 Water 18 221 3 0.2551 98.6 1.949 0.6744 0.7449 Active Bacteria in Arsenic-Rich Sediments 797

Nitrospira), as well as unclassified bacteria from http://beta. COWG reached 38%, corroborating the typical removal uniprot.org/, Thiomonas sp. from http://www.genoscope.cns. rates measured during the long-term monitoring study [23]. fr/ (FP475956–FP475957), Euglenozoa,andViridiplantae. The pale-yellow loosely packed sediments previously The second database included all bacterial and archaeal observed during the dry season at the sampling point GroEL chaperonins (12501 and 291 sequences, respectively) chosen in this study (COWG ~30 m downstream of extracted from the Uniprot database (http://www.uniprot.org/ the spring), consisted of an amorphous Fe(III)–As(V) uniprot). hydroxysulfate mineral with an As/Fe molar ratio of 0.5 To assess the false positive rate in the protein identifi- to 0.6 [48]. However, various other biominerals may be cation, a target-decoy database search was performed [25]. formed from dissolved Fe(II) and As(III) [47, 48]. The With this approach, peptides are matched against a database exact nature and structure of the sediment sample studied consisting of the native protein sequences detected in the was, therefore, further investigated. X-ray powder diffrac- database (target) and the sequence-reversed entries (decoy). tion (data not shown) and X-ray absorption spectroscopy Protein identification was confirmed when at least two data obtained at the As–K edge (Fig. 1) showed that arsenic peptides with a minimum Mascot ion score of 30 were was present in these samples in an amorphous Fe(III)–As detected. In the case of one-peptide hits, the score of the (V) hydroxysulfate phase as previously observed [48]. unique peptide had to be greater than the 95% significance These analyses showed that despite the presence of a minor Mascot threshold level. All the proteins identified were As(III) impurity, the oxidized arsenic form As(V) predom- added to the “InPact” proteomic database developed at our inates in this sediment (Fig. 1a). In a previous study, it has laboratory (http://inpact.u-strasbg.fr/~db/)[8]. been shown that the catalytic oxidation of As(III) by Thiomonas sp. strains accelerates such As–Fe precipitation Phylogenetic Analyses process [48]. Therefore, using extended X-ray absorption fine structure analysis at the As–K edge (Fig. 1b), the A search for GroEL homologs and 16S rRNA encoding structure observed in our samples was compared with that sequences was carried out in the Uniprot and NCBI of the minerals obtained after As(III) oxidation by the databases, respectively. A total number of 530 reviewed Thiomonas sp. strain B2 in bioassays in which sterilized GroEL bacterial sequences were retrieved from the RefSeq Carnoulès Creek water was used [48]. These comparisons database. These sequences were 500–550 amino acids in (Fig. 1b, c) showed that the molecular structure of the length. Only one sequence representative of each genus amorphous Fe(III)–As(V) hydroxysulfate phase observed in (259 sequences in all) was kept. GroEL and 16S rRNA these sediments was similar to that of the Fe(III)–As(V) encoding sequences were aligned using ClustalW [55]. hydroxysulfate obtained in the presence of Thiomonas, Alignments were checked by hand and positions with more further supporting its role in situ. than 1% of gaps were automatically removed. Neighbor- joining trees were constructed with 185 amino acids in the Composition of Bacterial Communities in Reigous Creek case of the GroEL sequences and with 310 nt in that of the Water and Sediments 16S rRNA encoding sequences. Trees were drawn up using the iTOL website (http://itol.embl.de/)[44]. Two 16S rRNA encoding gene libraries were constructed (Table 2), containing 229 clones in the sediment library and 221 in that of the water library. The Shannon index (see Results and Discussion “Methods”) and equitability values were greater in the water library than in the sediment library (Table 1), which Physical and Chemical Characteristics of Samples suggests that the bacterial diversity was lower in the sediment than in the water samples. Eleven different The physicochemistry of the Reigous Creek water at the species were identified in the sediments and 13 in the time of sampling was typical of that revealed during the water (Table 2, Fig. 2). long-term monitoring study [23]. The water sample was Several of the bacteria identified in the present study in acid (pH=3.28) and moderately oxygenated (dissolved both the waters and the sediments have been previously oxygen concentration=3.5±0.5 mg L−1); it contained detected in the Reigous waters [12, 14, 15, 18, 21, 22]. 2− extremely high concentrations of SO4 (2700± Classified first by abundance order, several sequences were 300 mg L−1), Fe (620±30 mg L−1), and As (140± affiliated to Thiobacillus sp. ML2-16. This bacterium 4mgL−1), with a large predominance of Fe(II) (90±10% has been frequently reported to occur in AMD [5]. of total Fe concentration) and equal proportion of As(III) The presence of several strains affiliated to this and As(V). The removal of As during the course of the proteobacteria is in agreement with the results of previous Reigous Creek from its source to the sampling station studies and shows that these bacteria persist in this 798 O. Bruneel et al.

Figure 1 X-ray absorption spectra at the As–K edge of the sample, space for the COWG sediment sample and the Thiomonas sp. strain (COWG April 11, 2006,) showing similarities with the X-ray B2 precipitate sample. c Corresponding Fourier transforms of the amorphous Fe(III)–As(V) hydroxysulfate phases obtained after incu- experimental and fit curves. Dotted lines experimental; solid lines bating sterilized Carnoulès water with the Thiomonas sp. strain B2 fitting curves. The local structure of both the COWG and the isolated at the site [48]. These phases reached a molar As/Fe ratio of laboratory Thiomonas sp. samples includes bidentate arsenate– 0.8, as described in [48]. a Linear least-squares fitting of XANES data oxygen–iron complexes characterized by ~1.5–2.0 Fe atoms at an showed that the largest arsenic fraction (90±2%) was in the As(V) As–Fe distance of 3.31±0.02 Å. A small arsenic fraction (10±2%) oxidation state. A small arsenic fraction (10±2%) was in the As(III) was in the As(III) oxidation state, which resulted in a slight decrease oxidation state, which resulted in a slight decrease in the amplitude in in the amplitude in b the EXAFS spectrum and in c the corresponding the EXAFS spectrum. b Shell by shell fit of the EXAFS spectra in k- Fourier transform ecosystem [12]. Secondly, bacteria affiliated to Gallionella donor [32]. Lastly, bacteria affiliated to Leptospirillum capsiferriformans were detected in both the water and ferrooxidans, an iron-oxidizing member of the Nitrospirae sediments. G. capsiferriformans is an oxygen-dependent [36], were also present. This bacterium has been found to ferrous iron-oxidizing bacterium that grows at circum- occur in several acidic environments and in biofilms neutral pH [59]. Relatives of Gallionella,suchasG. originating from AMD [5, 9, 32]. ferruginea in particular, have often been detected, some- Other newly characterized groups were identified only in times as the dominant group in microbial mine water sediments (Table 2). The presence in acidic mine waters of communities, including Carnoulès [12, 33–35]. Bacteria Acidocella sp., a non-iron-oxidizing heterotrophic acido- related to At. ferrooxidans as well as Thiomonas strains phile is quite common in AMD [33, 35 ]. Ferrimicrobium is have been previously isolated from this site [14, 18, 21, an iron-oxidizing heterotroph that can also use iron as an 22]. At. ferrooxidans, which was the first microorganism to electron acceptor [20]. Other prokaryotes detected in be isolated from an acidic leaching environment, occurs sediments were affiliated to Acidiphilium sp. CCP3, a ubiquitously in AMD, as does Thiomonas [32]. In addition non-iron-oxidizing heterotrophic acidophile that is also to these groups, other species that have not previously been quite common in AMD [33] and Dokdonella koreensis,a described at this site were detected in this study. Some of γ-proteobacteria. these species were found to occur in both sediments and Six newly characterized groups were identified only in water. For example, bacteria affiliated to the Firmicutes water. Some of these bacteria were related to Sideroxydans Alicyclobacillus sp. BRG 73 were identified. This genus, lithotrophicus LD-1, an oxygen-dependent ferrous iron- found in AMD [5], is characterized by moderately oxidizing bacterium that grows at circumneutral pH [59]; thermophilic, acidophilic, strictly aerobic, and endospore- Rhodoferax ferrireducens, a psychrotolerant, facultative forming bacilli [30]. Likewise, bacteria related to “Ferrovum anaerobic bacterium which is able to oxidize acetate with myxofaciens” PSTR were detected in both the waters and the reduction of Fe(III) [28]; and an Acidobacteriaceae sediments. “Ferrovum myxofaciens” is an autotrophic iron- bacterium, CH1. Members of Acidobacteria have previously oxidizer which predominates in some AMD and is able to been reported in AMD [5, 32]. Finally, three sulfate-reducing grow litho-autotrophically, using ferrous iron as an electron bacteria were related to strain JHA1, Desulfomonile limimaris, cieBcei nAsncRc Sediments Arsenic-Rich in Bacteria Active Table 2 Bacterial clones detected at the Carnoulès mine drainage with their phylogenetic group, the closest isolated relative and the relative abundance of each group versus the total number of clones (100%)

Sampling Clones Phylogenetic group Closest isolated relative (accession number) Percentage Relative of similarity abundance of clones (%)

Sediment CGA6Sd1a, 5a, 10b, 13c, 23c, 34c, 59b, 89c, 92c ß-Proteobacteria Thiobacillus sp. ML2-16 (DQ145970) 95–96 31 CGA6Sd4b, 36c, 48c ß-Proteobacteria Gallionella capsiferriformans ES-2 (DQ386262) 96 21 CGA6Sd13a γ-Proteobacteria Acidithiobacillus ferrooxidans DX-1 (EU084695) 99 10 CGA6Sd10a, 37a α-Proteobacteria Acidocella sp. M21 (AY765998) 99–100 8 CGA6Sd31c, 36a ß-Proteobacteria Thiomonas sp. PK44 (AY455806) 99 6 CGA6Sd6b, 18a, 27b γ-Proteobacteria Dokdonella koreensis NML 01–0233 (EF589679) 92 5 CGA6Sd32a Actinobacteria Ferrimicrobium sp. BGR 49 (GU167992) 99 3 CGA6Sd38c α-Proteobacteria Acidiphilium sp. CCP3 (AY766000) 99 3 CGA6Sd20a, 76c Firmicutes Alicyclobacillus sp. BGR 73 (GU167996) 92–99 3 CGA6Sd58b Nitrospirae Leptospirillum ferrooxidans Sy (AF356839) 99 2 CGA6Sd51c ß-Proteobacteria “Ferrovum myxofaciens” PSTR (EF133508) 100 2 Water CGA6Wt4c, 7a, 15c, 20a, 21b, 22b, 23a, 31b, 32a, 33c, ß-Proteobacteria Thiobacillus sp. ML2-16 (DQ145970) 94–96 26 45b, 54c, 63b, 73c CGA6Wt7c, 17b, 21a, 23c, 27b, 35b, 36b, 67c, 80c, 86c ß-Proteobacteria Gallionella capsiferriformans ES-2 (DQ386262) 89–97 18 CGA6Wt25c, 56b γ-Proteobacteria Acidithiobacillus ferrooxidans BGR:110 (GU168011) 100 10 CGA6Wt5a, 48c, 79b Acidithiobacillus ferrooxidans DSM 2392 (AJ459800) 91–92 CGA6Wt9a, 19a, 27a, 9b, 61c Firmicutes Alicyclobacillus sp. BGR 73 (GU167996) 91–99 8 CGA6Wt11a, 29c, 78c ß-Proteobacteria Sideroxydans lithotrophicus LD-1 (DQ386859) 94–97 8 CGA6Wt3a, 8c ß-Proteobacteria “Ferrovum myxofaciens” PSTR (EF133508) 98–99 4 CGA6Wt86b ß-Proteobacteria Thiomonas sp. PK44 (AY455806) 94 3 CGA6Wt15a Nitrospirae Leptospirillum ferrooxidans (AB510912) 94 3 CGA6Wt51b ß-Proteobacteria Rhodoferax ferrireducens T118 (CP000267) 99 3 CGA6Wt61b Acidobacteria Acidobacteriaceae bacterium CH1 (DQ355184) 96 3 CGA6Wt42c Sulfate-reducing bacterium JHA1 (EF442984) 82 3 CGA6Wt30a δ-Proteobacteria Desulfomonile limimaris (NR_025079) 87 3 CGA6Wt10a δ-Proteobacteria Desulfuromonas svalbardensis 60 (AY835390) 82 1

Sequences closely related to 16S rRNA genes from Euglena spp. chloroplast were also detected (data not shown). The 16S rRNA encoding gene of chloroplasts is closely related to the bacterial 16S rRNA encoding gene and can therefore be amplified by primers 8F and 1489R 799 800 O. Bruneel et al. an anaerobic dehalogenating bacterium from marine sediments Figure 2 Phylogenetic tree based on 16S rRNA-encoding sequences. b Sequences were aligned using ClustalW. Alignments were checked by [54]andDesulfuromonas svalbardensis 60, a psychrophilic, hand and positions with more than 1% of gaps were automatically Fe(III)-reducing bacterium isolated from Arctic sediments [56] removed. Neighbor-joining trees were drawn up with 310 nt using (Table 2). ITOL (http://itol.embl.de/)[44]. Accession numbers: see supplemen- All in all, 17 species of bacteria were identified in the tary data.Inred: bacteria identified based on the metaproteomic (GroEL identifications) approach; in blue: bacteria identified using the water and sediments sampled at the Reigous creek. Only 16S rRNA encoding gene library; in black: sequences detected in seven genera were found to be present in both phases, NCBI databases which are closely related to the bacteria present in the sixwerefoundonlyinwater,andfouronlyinthe Reigous sediment sediments (Table 2). Most of these species are common residents of AMD [5, 32]. This quite low bacterial diversity was probably due to the high concentration of Characterization of the Main Proteins Expressed toxic compounds in this AMD and was consistent with by the Sediment Community previous observations showing that the biodiversity of acidic, metal-rich mine waters is mainly restricted to Proteins expressed by this community were liable to corre- specialized prokaryotes and some eukaryotes such as spond to orthologs originating from diverse prokaryotes and Euglena [15, 52], which has been detected in this study to have similar amino acid sequences. For those reasons and to (data not shown). In the Reigous system, in both sedi- improve their characterization, proteins were separated by ments and water, the populations observed were mainly performing 2-D gel electrophoresis (supplementary Fig. 1). A involved in the Fe, As, and S cycles. The populations total number of 89 proteins were identified, 44% of which involved in Fe(II) oxidation were related to Gallionella, At. (39 proteins, Table 3) originated from bacteria and 39% (35 ferrooxidans, Ferrimicrobium, Leptospirillum, Sideroxydans proteins) from protists, while 15 proteins originated from lithotrophicus,or“Ferrovum myxofaciens” [32, 59], whereas higher plants, probably from decomposed plant debris ferric iron reduction has been described for populations like present at the Reigous creek. One third of the proteins Acidiphilium spp., Acidocella, Desulfuromonas svalbardensis, originated from protists, mainly consisting of E. mutabilis Rhodoferax ferrireducens,orevenAt. ferrooxidans and detected in our samples (data not shown) and could not be Ferrimicrobium acidiphilum [20, 28, 56]. Since the Carnoulès completely removed using the Nycodenz Gradient. E. creek spring contains mainly Fe(II) and As(III) in the form of mutabilis is a common inhabitant of AMD [2, 11, 15, 39] dissolved species [18], the Fe(III) and As(V) may be formed and was also present at the surface of sediments. These as the result of microbial oxidation processes via the activity Euglena proteins are involved in various metabolic processes, of acidophilic iron- and arsenite-oxidizing bacteria [24, 48]. In suggesting that this eukaryote may play a relevant role in this other words, the large amounts of soft pale-yellow As(V)–Fe ecosystem, in agreement with recent results (Gouhlen-Chollet (III) hydroxysulfate sediments analyzed here (Fig. 1)were and Bertin, unpublished data). Only proteins originating from probably formed by the joint activities of iron-oxidizing (e.g., bacteria were further analyzed in this study. The 2-D gel At. ferrooxidans or Gallionella) and arsenic-oxidizing (e.g., electrophoresis approach used here allowed identifying from a Thiomonas sp.) microorganisms. Concerning S cycling, we single sample only the most abundant cytoplasmic proteins. found populations able to oxidize the reduced inorganic sulfur Therefore, some relevant proteins that might be expressed in compounds, like Thiobacillus, Thiomonas,orAt. ferrooxidans this environment, such as rusticyanin from At. ferrooxidans, [32, 42, 46]. Sulfate-reducing bacteria such as Desulfomonile which is known to be involved in Fe(II) oxidation and was limimaris or Desulfuromonas svalbardensis were also found previously thought to possibly play a functional role at this to be present in sediments and may be involved in sulfate site [21, 24, 48], were not detected in this study. This consumption [54, 56]. membrane protein may not be resolved in the 2-D gel, or it Among these prokaryotes, some bacteria may be present may not be abundant at the sampling point. In addition, the but not functionally active, and it was, therefore, crucial to majority of the bacteria forming the Carnoulès community differentiate between dead or inactive cells and functional have never been grown and studied in vitro so far. Their cells. To determine which organisms play a significant role genome sequences, and hence their protein sequences, which in the natural remediation processes such as the Fe(II) are required for MS identification purposes, may not be oxidation processes observed at the study site, a metapro- available in public databases, except for those of At. teomic approach was used to list the bacterial population ferrooxidans, Thiomonas,andGallionella strains. All these expressing proteins, i.e., those which were active. The hypotheses may explain why so few bacterial proteins were metaproteomic approach was possible with sediments but identified in this study. failed with the water samples because larger numbers of The bacterial proteins identified originated from Aquificae bacterial cells were recovered from sediments than from were Actinobacteria, Deinococcus, Synergistetes, Firmicutes, water (data not shown). Bacteroidetes,andα-, β-, and γ-proteobacteria. Among the Active Bacteria in Arsenic-Rich Sediments 801

-proteobacteria Actinobacteria -proteobacteria Synergistetes -proteobacteria Firmicutes -proteobacteria Acidobacteria Nitrospirae Deinococcus/Thermus 802 Table 3 Bacterial proteins identified in the Reigous sediment’s microbial community

Phylum Class, Family, Genus Organism Level of Protein name Protein accession numbers Spot number b Peptide sequence discrimination

Aquificae Aquificae (class); Aquificales; Hydrogenobaculum sp. Species Putative uncharacterized protein A7WFJ8 25 IGAAVIGR Aquificaceae; Hydrogenobaculum Y04AAS1 Deinococcus- Deinococci; Deinococcales; Deinococcus deserti Species 60 kDa chaperoninsa C1CZP1 4, 6 APGFGDR Thermus Deinococcaceae; Deinococcus QLVFDEAAR Deinococci; Deinococcales; Deinococcus sp. A62 Species 60 kDa chaperoninsa A1YUK7 2 AVLVAIEEIK Deinococcaceae; Deinococcus Synergistetes Synergistia; Synergistales; Thermanaerovibrio Genus or 60 kDa chaperoninsa D1B621 4 IAQVASISANDK Synergistaceae; Thermanaerovibrio acidaminovorans below FGSPTITNDGVTIAK Firmicutes Bacilli; Lactobacillales; Lactobacillus Subspecies 60 kDa chaperoninsa Q70BV2, Q70BV5 4, 5 APGFGDR Lactobacillaceae delbrueckii subsp. YGAPTITNDGVTIAK Indicus or delbrueckii Bacilli; Bacillales; Bacillaceae Virgibacillus Family or 60 kDa chaperoninsa A9LHR1 4, 5, 6 AVEVAVK pantothenticus below Bacilli; Bacillales; Bacillaceae; Bacillus halodurans, Species 60 kDa chaperoninsa O50305, A8VUQ6, D3FSF9 2 NVTSGANPMVIR Bacillus pseudofirmus or selenitireducens Clostridia; Clostridiales; Alkaliphilus Genus or 60 kDa chaperoninsa A6TLJ1, A8HJ57 6 LSGGVAVIQVGAATETELK Clostridiaceae; Alkaliphilus metalliredigens or below ormlandii Clostridia; Clostridiales; Clostridium botulinum Species 60 kDa chaperoninsa A5I723, A7FYP3, A7GIN3, 4 APGFGDR Clostridiaceae; Clostridium B1IFD4, B1L1K0, C1FLV5, LGIDIIR C3KUC8, B1Q9U6, B1QI57 Clostridia; Clostridiales; Clostridium hiranonis Species 60 kDa chaperoninsa B6FW06 4, 5, 6 APGFGDR Clostridiaceae; Clostridium KALEEPLR VGAATEVEMK TNDIAGDGTTTATVLAQAIIR Clostridia; Clostridiales; Clostridium Species or 60 kDa chaperoninsa C7IKN8 4, 6 APGFGDR Clostridiaceae; Clostridium papyrosolvens subspecies FGSPTITNDGVTIAK Bacteroidetes Flavobacteria; Flavobacteriales; Gramella forsetii (strain Species ATP synthase; beta subunit A0M791 11 MPSAVGYQPTLATEMGAMQER Flavobacteriaceae; Gramella KT0803) Phosphoglycerate kinase A0M6J2 13, 20 LGDIYVNDAFGTAHR Actinobacteria Actinobacteria (class); Propionibacterium Species 60 kDa chaperoninsa A5JUG8 4, 5, 6 NVTAGANPIELK Actinobacteridae; Actinomycetales; freudenreichii Propionibacterineae; Propionibacteriaceae Actinobacteria (class); Stackebrandtia Suborder or 60 kDa chaperoninsa C4DUC7 6 GMNALADAVK Actinobacteridae; Actinomycetales; nassauensis below Glycomycineae; Glycomycetaceae; Stackebrandtia Actinobacteria (class); Streptosporangium Family 60 kDa chaperoninsa D2BBD1 4 APGFGDR Actinobacteridae; Actinomycetales; roseum GTFTSVAVK Streptosporangineae; Streptosporangiaceae; .Buele al. et Bruneel O. Streptosporangium Actinobacteria (class); Arthrobacter, Order ATP synthase; beta subunit A0JY64, A1R7V3, A3TGD9, 10, 31, 32, 33, 38 DVQNQDVLLFIDNIFR Actinobacteridae; Actinomycetales Janibacter, A5CQ60, A6W7G9 VALSALTMAEYFR Clavibacter or Kineococcus IGLFGGAGVGK Proteobacteria α-proteobacteria, Rhizobiales; Sinorhizobium medicae Genus or 60 kDa chaperoninsa A6UH06 2, 4 LVAAGMNPMDLK Table 3 (continued)

Phylum Class, Family, Genus Organism Level of Protein name Protein accession numbers Spot number b Peptide sequence discrimination cieBcei nAsncRc Sediments Arsenic-Rich in Bacteria Active Rhizobiaceae; Sinorhizobium/ below AAVEEGIVAGGGVALLR Ensifergroup; Sinorhizobium α-proteobacteria, Rhodospirillales; Acidiphilium cryptum Genus or 60 kDa chaperoninsa A5G1G2 5, 6 APGFGDR Acetobacteraceae; Acidiphilium below AAVEEGIVPGGGVALAR AVAAGMNPMDLK AGIIDPTK ENTTIVEGAGK α-proteobacteria, Rhodobacterales; Rhodobacterales Family or 60 kDa chaperoninsa B6AWC8 4, 5 APGFGDR unclassified Rhodobacterales bacterium below EIELADPFENMGAQLVK HTCC2083 SVAAGMNPMDLK β-proteobacteria, Burkholderiales; Acidovorax avenae Species 60 kDa chaperoninsa A1TKQ5, D1STJ1 4 APGFGDR Comamonadaceae; Acidovorax VGAATEVEMK AVTALVAELKK VTLADLGQAK AAVEEGIVAGGGVALLR β-proteobacteria, Burkholderiales; Bordetella petrii Species 60 kDa chaperoninsa A9I685 2, 4, 5 APGFGDR Alcaligenaceae; Bordetella AVEEPLR VGAATEVEMK EGVITVEDGK VQIEEATSDYDREK DLLPVLEQVAK VEDALHATR VQIEEATSDYDR β-proteobacteria Bordetella avium Species 50S ribosomal protein Q2L2M6 NC14 AEILDAIAGMTVLELSELIK (strain 197 N) L7/L12 DLVDGAPKPVK β-proteobacteria; Burkholderiales; Herminiimonas Species Glutathione-dependent A4G6P6 NC12 IIAIDTNPAK Oxalobacteraceae; Herminiimonas arsenicoxydans formaldehyde dehydrogenase TNLCVAVR (alcohol dehydrogenase class III), HEAR2039 Glutathione-independent A4G6Q5 NC3, NC11 LEDAPAAYK formaldehyde dehydrogenase, VIDYVGVDCR HEAR2048 GMTMGHEMTGEVIEVGSDVEVVK FPELITPQGK β-proteobacteria, Burkholderiales; Leptothrix cholodnii or Genus 60 kDa chaperoninsa B1XXY9, C7HZY6 2, 3, 4, 5, 36, 60 SFGAPTVTK unclassified Burkholderiales; Thiomonas YVAAGMNPMDLKR Burkholderiales Genera incertaesedis intermedia APGFGDR LQNMGAQMVK EGVITVEDGK VGAATEVEMK VIAEEVGLTLEK YVAAGMNPMDLK VTLADLGQAK IQIEEATSDYDREK VEDALHATR 803 IQIEEATSDYDR 804 Table 3 (continued)

Phylum Class, Family, Genus Organism Level of Protein name Protein accession numbers Spot number b Peptide sequence discrimination

AMLEDIAILTGGK AAVEEGIVAGGGVALLR Thiomonas 3As Genus 50S ribosomal protein L1; FP475956 38 VAVSSTMGIGVR THI3722 VDTATVNAAVAGQ β-proteobacteria, Burkholderiales; Limnobacter sp. Genus or 60 kDa chaperoninsa A6GTE5 4 APGFGDR Burkholderiaceae; Limnobacter MED105 below VGAATEVEMK GVNILANAVK β-proteobacteria, Burkholderiales; Methylibium Genus or 60 kDa chaperoninsa A2SCV1 2, 4 SFGAPTVTK unclassified Burkholderiales; petroleiphilum below YVAAGMNPMDLKR Burkholderiales Genera incertaesedis; Methylibium APGFGDR VQIEEATSDYDREK LQNMGAQMVK VGAATEVEMK EGVITVEDGK VIAEEVGLTLEK YVAAGMNPMDLK VTLADLGQAK VEDALHATR AAVEEGIVAGGGVALLR VQIEEATSDYDR VQIEEATSDYDR AMLEDIAILTGGK β-proteobacteria, Burkholderiales; Ralstonia pickettii Subspecies 60 kDa chaperoninsa B2U6M6 4 APGFGDR Burkholderiaceae; Ralstonia VGAATEVEMK DLLPILEQVAK AAVEEGIVAGGGVALLR β-proteobacteria, Burkholderiales; Verminephrobacter Genus 60 kDa chaperoninsa A1WL05 2, 4 APGFGDR Comamonadaceae; eiseniae EVVFGGEAR Verminephrobacter VGAATEVEMK EGVITVEDGK VIAEEVGLTLEK VQIEEATSDYDREK AVTALVAELKK VTLADLGQAK

VEDALHATR al. et Bruneel O. VQIEEATSDYDR AMLEDIAILTGGK β-proteobacteria, Methylophilales; Methylotenera mobilis Genus or 60 kDa chaperoninsa C6WTL6 5, 6 SVAAGMNPMDLK Methylophilaceae; Methylotenera below Table 3 (continued)

Phylum Class, Family, Genus Organism Level of Protein name Protein accession numbers Spot number b Peptide sequence discrimination cieBcei nAsncRc Sediments Arsenic-Rich in Bacteria Active TNDIAGDGTTTATVLAQAIIR β-proteobacteria, Gallionellales; Gallionella ferruginea Genus or 60 kDa chaperoninsa C5V7N1 2 VGAATEVEMK Gallionellaceae; Gallionella below GYLSPYFINNQDR DLLPVLEQVAK VEDALHATR γ-proteobacteria, Acidithiobacillales; Acidithiobacillus Species 60 kDa chaperoninsa B5EN19, B7J561 2, 3 APGFGDR Acidithiobacillaceae; ferrooxidans HALEGFK Acidithiobacillus AVIAGMNPMDLK GVNVLADAVK VVSEEIGMK VEDALHATR AMLEDMAILTGGR LESTTLADLGQAK γ-proteobacteria, Methylococcales; Methylococcus Order or 60 kDa chaperoninsa Q60AY0 2, 3, 4 APGFGDR Methylococcaceae; Methylococcus capsulatus below VGAATEVEMK VEDALHATR QIVANAGDEPSVVLNK γ-proteobacteria; Pseudomonadales; Pseudomonas Genus Outer membrane lipoprotein A2VC34, A4XVE5, O85409- 4, 15, 17, 19, 22, 26– LTATEDAAAR Pseudomonadaceae OprI 085430-O85432-O85437, 27, 40, 44, 49, 55, KADEALAAAQK O85439-O85444, Q3K906, 60, 62, 64–65, Q48K14, Q883S8 NC7-NC10 ADEALAAAQK ITATEDAAAR γ-proteobacteria, unclassified Gammaproteobacterium Group or 60 kDa chaperoninsa Q1YSA6 5 SVAAGMNPMDLK Gammaproteobacteria; OMG group; HTCC2207 clade AQIEDTSSDYDR SAR92 clade δ-proteobacteria, Myxococcales; Haliangium ochraceum Order or 60 kDa chaperoninsa D0LRR3 4, 6 APGFGDR Nannocystineae; Haliangiaceae; below VGAATEVEMK Haliangium GYLSPYFVTDSER Proteobacteria Several Proteobacteria Phylum Succinyl-CoA synthetase; beta Q3KFU6c NC1, NC3, NC11 LEGNNAELGAK subunit a MS/MS data were searched against an in-house database including all the bacterial and archaeal GroEL sequences obtained from Uniprot b Spots from 1 to 65 originated from a pH 4 to 7 gradient gel (Supplementary Fig. 1), spots from NC1 to NC14 originated from a pH 3 to 10 gradient gel (data not shown) c Several proteins may correspond to this identification: Q3KFU6, A1A8Y0, A1FGM7, A1JRB6, A1KTM6, A2UEB0, A3HHN5, A3M887, A4TNT8, A4W879, A4XV90, A5W114, A5WC33, A6BTC9, A6T6F6, A6V7K5, A7FKR4, A7MQX5, A7ZJA8, A7ZXY8, A8AJ84, A8GB83, P0A836, P0A837, P0A838, P0A839, P53593, P66869, P66870, Q02K73, Q0T6W6, Q0TJW6, Q1CAG1, Q1CFM0, Q1I7L3, Q1REJ8, Q21IW6, Q2NUM2, Q2SD35, Q324I4, Q32IK3, Q3Z476, Q48K68, Q4FVH9, Q4KFY6, Q4ZUW7, Q57RL3, Q5F878, Q5PCM7, Q66DA0, Q6D7G2, Q6F8L4, Q7N6V5, Q7NZ47, Q883Z4, Q88FB2, Q8ZH00, Q9JUT0, Q9JZP4 805 806 O. Bruneel et al.

39 bacterial proteins detected, there were two distinct ATP 16S rRNA-encoding gene analysis showed that these synthases, two distinct 50 S ribosomal proteins, one bacteria abound in this ecosystem. Bacteria affiliated to At. phosphoglycerate kinase, and one succinyl-CoA synthetase, ferrooxidans and Gallionella are able to oxidize iron. In which are involved in energy metabolism, translation, addition, many strains of the Thiomonas genus are able to glycolysis, and TCA cycle, respectively (Table 3). These oxidize As(III) into As(V) under laboratory conditions proteins may not play a specific role in this environment [6, 14, 22]. The fact that their proteins were detected shows since they are known to be housekeeping proteins in bacteria. that these bacteria were viable and metabolically active. This Among the other proteins, one outer protein OrpI, one finding supports the hypothesis that the oxidation of Fe(II) to uncharacterized protein with an unknown function, one Fe(III) catalyzed by iron-oxidizing microorganism such as glutathione-dependent, and one glutathione-independent At. ferrooxidans and Gallionella and oxidation of As(III) into formaldehyde dehydrogenases were identified. The latter As(V) by As(III) oxidizers such as Thiomonas,probably two proteins belong to the formaldehyde detoxification leads to the precipitation of the more or less ordered iron pathway. Although no correlation with the environmental oxy-hydroxides (Fe(III)–As(V) hydroxysulfate) detected in conditions might explain the functional specificities of these this study (Fig. 1)[48]. This finding is in agreement with proteins, it has been reported that the arsenite-oxidizing previous data showing that Gallionella ferruginea efficiently bacteria H. arsenicoxydans synthesizes alcohol dehydroge- remove Fe, As(III), and As(V) in water [41]. The present nase and glutathione-dependent formaldehyde dehydrogenase data show for the first time that this bacterium is active when grown in the presence of arsenic [58]. and probably plays a functional role in the sediments of It is worth noting that more than half of the identified the Reigous creek. Some heterotrophic bacteria such as bacterial proteins were similar to the 60-kDa GroEL chaper- Acidiphilium werealsofoundtobeactiveinthisAMD, onin. These data suggest that multiple chaperonins of various suggesting that they could cope with the low amount of genetic origins are expressed by the Reigous creek commu- organic carbon (dissolved organic carbon concentration nity. GroEL is known to be ubiquitously present in Bacteria 1.7±0.4 mg/L [16]) of Reigous creek water. It has previously and Archaea. These proteins are generally abundantly been suggested that these acidophilic heterotrophic bacteria expressed in bacterial cells, especially under stress conditions may be involved in organic carbon turnover processes [32]. such as those occurring in this particularly toxic environment Interestingly, these four bacteria (At. ferrooxidans, Thiomonas, [3]. The groEL gene is conserved in prokaryotes, and has Gallionella,andAcidiphilium) found to be active members of been found to be present in one copy in the majority of this AMD community have been previously identified in sequenced genomes, except in the case of some pathogens AMD, but some of them were thought to have different [45, 60]. Because of its conservation properties (supplemen- optimum pH levels. Indeed, G. ferruginea is a neutrophilic tary Fig. 2), this gene is often used as a phylogenetic marker bacterium which oxidizes Fe, but relatives of Gallionella, [31]. In addition, it has been previously established that some have often been detected in AMD [12, 33–35]. The strain of the amino acids stretch occurring in GroEL are specific to occurring at Carnoulès showed less than 97% homology with one genus or family of bacteria. These peptide sequences can G. capsiferriformans and its physiological characteristics are therefore be used as the signature of a specific phylogenetic probably different. It seems probable that an acid-tolerant group. These GroEL identifications (Table 3, Fig. 3) were, relative of this bacterium is able to oxidize iron under acid pH therefore, considered for use as a possible taxonomic tool in conditions. addition to the 16S rRNA-based taxonomic approach. To In addition to the bacteria belonging to these four genera, determine which bacteria in the whole community were other bacteria were also found to be active, but a active, i.e., able to express proteins, the organisms identified discrepancy was again observed between the bacteria using the 16S rRNA encoding gene library- and GroEL- identified based on the results of 16S rRNA encoding gene based approaches (Fig. 2)werecompared. analysis and the metaproteomic approach. The GroEL Most of the bacteria identified based on GroEL belonged to protein sequences of some bacteria identified using the five phyla divisions. Bacteria belonging to Deinococcus and 16S rRNA encoding gene library were not available in the Synergistetes did not feature among those identified based on Uniprot database, which might explain this discrepancy. the 16S rRNA encoding gene. One possible explanation for However, phylogenetic comparisons between the 16S this discrepancy may be that a PCR or cloning bias may rRNA and metaproteomic data obtained (Fig. 2) suggested have prevented those bacteria from being detected with this which of the bacteria present in this ecosystem may express method. These findings suggest that metaproteomic methods a GroEL identified in the metaproteomic study. Based on used as taxonomic tools can provide a useful complementary these comparisons, it seems likely that in addition to tool in addition to the 16S rRNA encoding gene approach. Thiomonas, At. ferrooxidans, Acidiphilium,andGallionella, Interestingly, Thiomonas, At. ferrooxidans, Acidiphilium,and clones related to β-, γ-, and δ-proteobacteria,suchas Gallionella, expressed proteins and were, therefore, active. Limnobacter or Methylococcus (At. ferrooxidans DSM Active Bacteria in Arsenic-Rich Sediments 807

A. avenae subsp. avenae A. avenae subsp. citrulli A. ferrooxidans ATCC53993 A. ferrooxidans ATCC23270T A. cryptum A. metalliredigens A. oremlandii B. halodurans B. pseudofirmus B. selenitireducens B. petrii C. botulinum C. hiranonis C. papyrosolvens D. desertii D. sp. A62 G. ferruginea Gamma-proteobacterium HTCC2207 H. ochraceum L. delbrueckii subsp. delbrueckii L. delbrueckii subsp. indicus L. cholodnii Limnobacter sp. MED105 M. petroleiphilum M. capsulatus M. mobilis P. freudenreichii R. picketii Rhodobacterales bacterium HTCC2083 S. medicae S. nassauensis S. roseum T. acidaminovorans Thiomonas 3As T. intermedia V. eiseniae V. pantothenticus

Figure 3 Alignment of the amino acid sequences of 60 kDa chaper- chaperonins expressed by the bacteria present at the study site are onins. Thirty-six sequences matched the MS results: 29 Proteins were unknown, the possibility cannot be ruled out that two identified peptides identified in Table 3: 22 of them unambiguously matched one protein were erroneously assigned to two distinct chaperonins, whereas these (the bacterial name of which is given in orange), whereas fifteen peptides may in fact have originated from one protein, the amino acid identifications matched at least two proteins (22 bacterial names given sequence of which has not yet been included in the databases. in black). Thiomonas sp. 3As data were added for the sake of Nevertheless, because of the high level of conservation observed in comparison (in blue). Thirty-seven sequences in all were aligned. These GroEL proteins (Supplementary Fig. 2), the proteins identified may be sequences were compared with ClustalW2 using the default parameters expressed by a close relative of the organism identified using the (http://www.ebi.ac.uk/Tools/clustalw2/index.html). Alignment was per- Uniprot database. GroEL originated from Acidiphilium cryptum; Acid- formed using the Jalview software program [19]. The blue highlighted ithiobacillus ferrooxidans ATCC23270T; Acidithiobacillus ferrooxidans letters correspond to identical residues among the 37 orthologs ATCC53993; Acidovorax avenae subsp. avenae; Acidovorax avenae (BLOSUM62 score option). Highlighted letters corresponding to the subsp. citrulli; Alkaliphilus metalliredigens; Alkaliphilus oremlandii; peptide sequences identified by MS allowed distinguishing several Bacillus halodurans; Bacillus pseudofirmus; Bacillus selenitireducens; orthologs. First, 23 proteins were identified because at least one Bordetella petrii; Clostridium botulinum; Clostridium hiranonis; Clos- identified peptide matched only this GroEL: these peptides (labeled in tridium papyrosolvens; Deinococcus desertii; Deinococcus sp. A62; red) corresponding to signature sequences specific to one species, Gallionella ferruginea; Haliangium ochraceum; Lactobacillus del- genus, or family of bacteria are located in the same homology regions brueckii subsp. delbrueckii; Lactobacillus delbrueckii subsp. indicus; but differed from other peptides by at least one amino acid substitution. Leptothrix cholodnii; Limnobacter sp. MED105; Methylibium petrolei- Secondly, 13 of the proteins identified were confirmed, since several philum; Methylococcus capsulatus; Methylotenera mobilis; Propioni- identified peptides (labeled in green) matched this GroEL. Each of these bacterium freudenreichii; Ralstonia picketii; Rhodobacterales bacterium individual peptides matched several proteins; however, only one protein HTCC2083; Sinorhizobium medicae; Stackebrandtia nassauensis; detected in Uniprot contained all these amino-acid sequences, which Streptosporangium roseum; Thermanaerovibrio acidaminovorans; Thi- suggests that a relative of this protein was probably present in this omonas 3As; Thiomonas intermedia; Verminephrobacter eiseniae; extract. Since the full amino-acid sequences of almost all the Virgibacillus pantothenticus; and Gammaproteobacterium HTCC2207 808 O. Bruneel et al.

2392 affiliated bacteria), Methylotenera (Thiobacillus- may contribute importantly to the remediation process related clones), D. koreensis (which is affiliated to the observedinsitu. γ-proteobacteria HTCC2207), Ferrimicrobium-like bacteria (which are related to the Actinobacteria S. nassauensis and In Conclusion S. roseum), and Haliangium (clones affiliated to sulfate- reducing bacteria JHA1, D. limimaris and D. svalbardensis) The active bacterial species inhabiting Carnoulès AMD may play a role in this ecosystem. All in all, the data ecosystem were identified in this study using high- obtained here show that Firmicutes, which could be affiliated sensitivity nanoLC-chip-MS/MS methods combined with to Alicyclobacillus ferrooxidans, are also active. These a 16S rRNA based phylogenetic approach. The meta- bacteria are able to oxidize ferrous iron [37]andmay, proteomicdataobtainedhereshowforthefirsttimethat therefore, participate in the transformation of the iron present Gallionella, Thiomonas, At. ferrooxidans,andAcidiphi- in high concentrations in these waters. The active popula- lium actively express proteins in situ. Previous hypotheses tion as a whole was not only composed of several iron- based on experiments performed under laboratory con- oxidizers in addition to At. ferrooxidans but also ditions [14, 18, 21, 22, 24, 48] suggest that microbial contained iron reducers, one known arsenite-oxidizer, activity may contribute to the arsenite oxidation and As sulfate-reducing, and sulfur compound oxidizers, and both immobilization occurring in the heavily contaminated autotrophic and heterotrophic bacteria. All these bacteria AMD at the Carnoulès mine via iron oxidation processes. Active Bacteria in Arsenic-Rich Sediments 809

Since these bacteria were found to be active and to express 7. Bednar AJ, Garbarino JR, Ranville JF, Wildeman TR (2002) proteins which are among the most abundant proteins Preserving the distribution of inorganic arsenic species in groundwater and acid mine drainage samples. Environ Sci Technol encountered at this site, it seems likely that the large 36:2213–2218 amounts of pale-yellow As(V)–Fe(III) hydroxysulfate 8. Bertin PN, Médigue C, Normand P (2008) Advances in sediments forming at Carnoulès, which were characterized environmental genomics: towards an integrated view of micro- – here, may result from the conjugate activities of iron- organisms and ecosystems. Microbiology 154:347 359 9. Bond PL, Smriga SP, Banfield JF (2000) Phylogeny of micro- oxidizing microorganisms (such as At. ferrooxidans, Alicyclo- organisms populating a thick, subaerial, predominantly lithotro- bacillus ferrooxydans, Ferrimicrobium,orGallionella)and phic biofilm at an extreme acid mine drainage site. Appl Environ arsenic-oxidizing microorganisms (such as Thiomonas sp.). Microbiol 66:3842–3849 Several bacteria may be responsible in situ for changing the 10. Bowell RJ (1994) Sorption of arsenic by iron oxides and oxyhydroxides in soils. Appl Geochem 9:279–286 ratio between the oxidized and reduced forms of iron, arsenic, 11. Brake SS, Hasiotis ST (2010) Eukaryote-dominated biofilms and and sulfur compounds, promoting the formation of the Fe their significance in acidic environments. Geomicrobiol J 27:534–558 (III)–As(V) hydroxysulfate precipitates detected in this study. 12. Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F, Personné JC These bacteria are therefore of prime importance in the partial (2006) Diversity of microorganisms in Fe-As-rich acid mine drainage waters of Carnoules, France. Appl Environ Microbiol but efficient natural process of remediation undergone by the 72:551–556 contaminated Carnoulès ecosystem. In addition, autotrophic 13. Bruneel O, Pascault N, Egal M, Bancon-Montigny C, Goni-Urriza iron, arsenic, and sulfur oxidizers may provide the organic MS, Elbaz-Poulichet F, Personné JC, Duran R (2008) Archaeal carbon sources required by the functional heterotrophs such diversity in a Fe-As rich acid mine drainage at Carnoules (France). Extremophiles 12:563–571 as Acidiphilium present in this ecosystem. All in all, the 14. Bruneel O, Personné JC, Casiot C, Leblanc M, Elbaz-Poulichet F, present findings provide evidence that the functional Mahler BJ, Le Flèche A, Grimont PA (2003) Mediation of arsenic genomics approach provides a useful means of describing oxidation by Thiomonas sp. in acid-mine drainage (Carnoulès, – bacterial communities such as those inhabiting the Reigous France). J Appl Microbiol 95:492 499 15. Casiot C, Bruneel O, Personné JC, Leblanc M, Elbaz-Poulichet F creek and determining their contribution to natural attenuation (2004) Arsenic oxidation and bioaccumulation by the acidophilic processes. It is now proposed to use approaches of this kind protozoan, Euglena mutabilis, in acid mine drainage (Carnoules, in future studies to complete this picture of the functional France). Sci Total Environ 320:259–267 processes at work in this ecosystem, as well as to investigate 16. Casiot C, Egal M, Bruneel O, Bancon-Montigny C, Cordier MA, Gomez E, Aliaume C, Elbaz-Poulichet F (2009) Hydrological and the role played by the less abundant active bacteria identified geochemical controls on metals and arsenic in a Mediterranean in this study. river contaminated by acid mine drainage (the Amous river, France); preliminary assessment of impacts on fish (Leuciscus – Acknowledgements The study was financed by the EC2CO program cephalus). Appl Geochem 24:787 799 (“Institut National des Sciences de l’Univers,” CNRS), the “Observatoire 17. Casiot C, Lebrun S, Morin G, Bruneel O, Personné JC, Elbaz- de Recherche Méditerranéen en Environnement” (OSU-OREME), and by Poulichet F (2005) Sorption and redox processes controlling the ANR 07-BLANC-0118 project (“Agence Nationale de la Recherche”). arsenic fate and transport in a stream impacted by acid mine – Sébastien Gallien and Aurélie Volant were supported by a grant from the drainage. Sci Total Environ 347:122 130 French Ministry of Education and Research. This work was performed in 18. Casiot C, Morin G, Juillot F, Bruneel O, Personné JC, Leblanc M, the framework of the “Groupement de recherche: Métabolisme de Duquesne K, Bonnefoy V, Elbaz-Poulichet F (2003) Bacterial l’Arsenic chez les Microorganismes” (GDR2909-CNRS). immobilization and oxidation of arsenic in acid mine drainage (Carnoulès creek, France). Water Res 37:2929–2936 19. Clamp M, Cuff J, Searle SM, Barton GJ (2004) The Jalview Java alignment editor. Bioinformatics 20:426–427 References 20. Coupland K, Johnson DB (2008) Evidence that the potential for dissimilatory ferric iron reduction is widespread among 1. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) acidophilic heterotrophic bacteria. FEMS Microbiol Lett Basic local alignment search tool. J Mol Biol 215:403–410 279:30–35 2. Amaral Zettler LA, Gomez F, Zettler E, Keenan BG, Amils R, 21. Duquesne K, Lebrun S, Casiot C, Bruneel O, Personné JC, Sogin ML (2002) Microbiology: eukaryotic diversity in Spain’s Leblanc M, Elbaz-Poulichet F, Morin G, Bonnefoy V (2003) River of Fire. Nature 417:137 Immobilization of arsenite and ferric iron by Acidithiobacillus 3. Arsène F, Tomoyasu T, Bukau B (2000) The heat shock response ferrooxidans and its relevance to acid mine drainage. Appl of Escherichia coli. Int J Food Microbiol 55:3–9 Environ Microbiol 69:6165–6173 4. Ashelford KE, Chuzhanova NA, Fry JC, Jones AJ, Weightman AJ 22. Duquesne K, Lieutaud A, Ratouchniak J, Muller D, Lett MC, (2005) At least 1 in 20 16S rRNA sequence records currently held Bonnefoy V (2008) Arsenite oxidation by a chemoautotrophic in public repositories is estimated to contain substantial anoma- moderately acidophilic Thiomonas sp.: from the strain isolation to lies. Appl Environ Microbiol 71:7724–7736 the gene study. Environ Microbiol 10:228–237 5. Baker BJ, Banfield JF (2003) Microbial communities in acid mine 23. Egal M, Casiot C, Morin G, Elbaz-Poulichet F, Cordier MA, drainage. FEMS Microbiol Ecol 44:139–152 Bruneel O (2010) An updated insight into the natural attenuation 6. Battaglia-Brunet F, Dictor MC, Garrido F, Crouzet C, Morin D, of As concentrations 3 in Reigous Creek (southern France). Appl Dekeyser K, Clarens M, Baranger P (2002) An arsenic(III)- Geochem 25:1949–1957 oxidizing bacterial population: selection, characterization, and 24. Egal M, Casiot C, Morin G, Parmentier M, Bruneel O, Lebrun S, performance in reactors. J Appl Microbiol 93:656–667 Elbaz-Poulichet F (2009) Kinetic control on the formation of 810 O. Bruneel et al.

tooeleite, schwertmannite and jarosite by Acidithiobacillus fer- 43. Lane DJ, Pace B, Olsen GJ, Stahl DA, Sogin ML, Pace NR (1985) rooxidans strains in an As(III)-rich acid mine water. Chem Geol Rapid determination of 16S ribosomal RNA sequences for 265:432–441 phylogenetic analyses. Proc Natl Acad Sci USA 82:6955–6959 25. Elias JE, Gygi SP (2007) Target-decoy search strategy for 44. Letunic I, Bork P (2007) Interactive tree of life (iTOL): an online increased confidence in large-scale protein identifications by mass tool for phylogenetic tree display and annotation. Bioinformatics spectrometry. Nat Methods 4:207–214 23:127–128 26. Felsenstein J (1992) Estimating effective population size from 45. Lund PA (2009) Multiple chaperonins in bacteria–why so many? samples of sequences: a bootstrap Monte Carlo integration FEMS Microbiol Rev 33:785–800 method. Genet Res 60:209–220 46. Moreira D, Amils R (1997) Phylogeny of Thiobacillus cuprinus 27. Ferris FG, Hallberg RO, Lyvén B, Pedersen K (2000) Retention of and other mixotrophic thiobacilli: proposal for Thiomonas gen. strontium, cesium, lead and uranium by bacterial iron oxides from nov. Int J Syst Bacteriol 47:522–528 a subterranean environment. Appl Geochem 15:1035–1042 47. Morin G, Calas G (2006) Arsenic in soils, mine tailings, and 28. Finneran KT, Johnsen CV, Lovley DR (2003) Rhodoferax former industrial sites. Elements 2:97–101 ferrireducens sp. nov., a psychrotolerant, facultatively anaerobic 48. Morin G, Juillot F, Casiot C, Bruneel O, Personné JC, Elbaz- bacterium that oxidizes acetate with the reduction of Fe(III). Int J Poulichet F, Leblanc M, Ildefonse P, Calas G (2003) Bacterial Syst Evol Microbiol 53:669–673 formation of tooeleite and mixed arsenic(III) or arsenic(V)-iron 29. Good IJ (1953) The population frequencies of species and the (III) gels in the Carnoules acid mine drainage, France. A XANES, estimation of population parameters. Biometrika 40:237–264 XRD, and SEM study. Environ Sci Technol 37:1705–1712 30. Goto K, Mochida K, Kato Y, Asahara M, Fujita R, An SY, Kasai 49. Ona-Nguema G, Morin G, Juillot F, Calas G, Brown GE (2005) H, Yokota A (2007) Proposal of six species of moderately EXAFS analysis of arsenite adsorption onto two-line ferrihydrite, thermophilic, acidophilic, endospore-forming bacteria: Alicyclo- hematite, goethite, and lepidocrocite. Environ Sci Technol bacillus contaminans sp. nov., Alicyclobacillus fastidiosus sp. 39:9147–9155 nov., Alicyclobacillus kakegawensis sp. nov., Alicyclobacillus 50. Rodier J, Broutin JP, Chambon P, Champsaur H, Rodi L (1996) macrosporangiidus sp. nov., Alicyclobacillus sacchari sp. nov. L’Analyse des Eaux. Dunod, Paris, p 1383 and Alicyclobacillus shizuokensis sp. nov. Int J Syst Evol 51. Rowe OF, Johnson DB (2008) Comparison of ferric iron Microbiol 57:1276–1285 generation by different species of acidophilic bacteria immobi- 31. Gupta RS (1998) Protein phylogenies and signature sequences: a lized in packed-bed reactors. Syst Appl Microbiol 31:68–77 reappraisal of evolutionary relationships among archaebacteria, 52. Rowe OF, Sanchez-Espana J, Hallberg KB, Johnson DB (2007) eubacteria, and eukaryotes. Microbiol Mol Biol Rev 62:1435– Microbial communities and geochemical dynamics in an extreme- 1491 ly acidic, metal-rich stream at an abandoned sulfide mine (Huelva, 32. Hallberg KB (2010) New perspectives in acid mine drainage Spain) underpinned by two functional primary production microbiology. Hydrometallurgy 104:448–453 systems. Environ Microbiol 9:1761–1771 33. Hallberg KB, Coupland K, Kimura S, Johnson DB (2006) 53. Sánchez España JS, LópezPamo E, Santofimia Pastor E, Reyes Macroscopic streamer growths in acidic, metal-rich mine waters Andrés J, Rubi J (2005) The natural attenuation of two acidic in north wales consist of novel and remarkably simple bacterial effluents in Tharsis and La Zarza-Perrunal mines (Iberian Pyrite communities. Appl Environ Microbiol 72:2022–2030 Belt, Huelva, Spain). Environ Geol 49:253–266 34. He Z, Xiao S, Xie X, Zhong H, Hu Y, Li Q, Gao F, Li G, Liu J, 54. Sun B, Cole JR, Tiedje JM (2001) Desulfomonile limimaris sp. Qiu G (2007) Molecular diversity of microbial community in nov., an anaerobic dehalogenating bacterium from marine sedi- acid mine drainages of Yunfu sulfide mine. Extremophiles ments. Int J Syst Evol Microbiol 51:365–371 11:305–314 55. Thompson JD, Plewniak F, Thierry J, Poch O (2000) DbClustal: 35. Heinzel E, Hedrich S, Janneck E, Glombitza F, Seifert J, rapid and reliable global multiple alignments of protein sequences Schlomann M (2009) Bacterial diversity in a mine water treatment detected by database searches. Nucleic Acids Res 28:2919–2926 plant. Appl Environ Microbiol 75:858–861 56. Vandieken V, Mussmann M, Niemann H, Jorgensen BB (2006) 36. Hippe H (2000) Leptospirillum gen. nov. (ex Markosyan 1972), Desulfuromonas svalbardensis sp. nov. and Desulfuromusa nom. rev., including Leptospirillum ferrooxidans sp. nov. (ex ferrireducens sp. nov., psychrophilic, Fe(III)-reducing bacteria Markosyan 1972), nom. rev. and Leptospirillum thermoferroox- isolated from Arctic sediments, Svalbard. Int J Syst Evol Micro- idans sp. nov. (Golovacheva et al. 1992). Int J Syst Evol biol 56:1133–1139 Microbiol 50(Pt 2):501–503 57. Weisburg WG, Barns SM, Pelletier DA, Lane DJ (1991) 16S 37. Jiang CY, Liu Y, Liu YY, You XY, Guo X, Liu SJ (2008) ribosomal DNA amplification for phylogenetic study. J Bacteriol Alicyclobacillus ferrooxydans sp. nov., a ferrous-oxidizing bacteri- 173:697–703 um from solfataric soil. Int J Syst Evol Microbiol 58:2898–2903 58. Weiss S, Carapito C, Cleiss J, Koechler S, Turlin E, Coppee J-Y, 38. Johnson DB (1995) Acidophilic microbial communities: candi- Heymann M, Kugler V, Stauffert M, Cruveiller S, Médigue C, Van dates for bioremediation of acidic mine effluents. Int Biodeterior Dorsselaer A, Bertin PN, Arsène-Ploetze F (2009) Enhanced Biodegrad 35:41–58 structural and functional genome elucidation of the arsenite- 39. Johnson DB, Hallberg KB (2003) The microbiology of acidic oxidizing strain Herminiimonas arsenicoxydans by proteomics mine waters. Res Microbiol 154:466–473 data. Biochimie 91:192–203 40. Johnson DB, Hallberg KB (2005) Acid mine drainage remediation 59. Weiss JV, Rentz JA, Plaia T, Neubauer SC, Merrill-Floyd M, options: a review. Sci Total Environ 338:3–14 Lilburn T, Bradburne C, Megonigal JP, Emerson D (2007) 41. Katsoyiannis IA, Zouboulis AI (2004) Application of biological Characterization of neutrophilic Fe(II)-oxidizing bacteria isolated processes for the removal of arsenic from groundwaters. Water from the rhizosphere of wetland plants and description of Res 38:17–26 Ferritrophicum radicicola gen. nov. sp. nov., and Sideroxydans 42. Kelly DP, Wood AP (2000) Confirmation of Thiobacillus paludicola sp. nov. Geomicrobiol J 24:559–570 denitrificans as a species of the genus Thiobacillus, in the beta- 60. Williams TA, Codoner FM, Toft C, Fares MA (2010) Two subclass of the Proteobacteria, with strain NCIMB 9548 as the chaperonin systems in bacterial genomes with distinct ecological type strain. Int J Syst Evol Microbiol 50(Pt 2):547–550 roles. Trends Genet 26:47–51 Author's personal copy

Extremophiles (2012) 16:645–657 DOI 10.1007/s00792-012-0466-8

ORIGINAL PAPER

Archaeal diversity: temporal variation in the arsenic-rich creek sediments of Carnoule`s Mine, France

A. Volant • A. Desoeuvre • C. Casiot • B. Lauga • S. Delpoux • G. Morin • J. C. Personne´ • M. He´ry • F. Elbaz-Poulichet • P. N. Bertin • O. Bruneel

Received: 16 February 2012 / Accepted: 3 May 2012 / Published online: 20 June 2012 Ó Springer 2012

Abstract The Carnoule`s mine is an extreme environment a low degree of biodiversity with two different phyla: Eur- located in the South of France. It is an unusual ecosystem due yarchaeota and Thaumarchaeota. The archaeal community to its acidic pH (2–3), high concentration of heavy metals, varied in composition and richness throughout the sampling iron, and sulfate, but mainly due to its very high concentration campaigns. Many sequences were phylogenetically related of arsenic (up to 10 g L-1 in the tailing stock pore water, and to the order Thermoplasmatales represented by aerobic or 100–350 mg L-1 in Reigous Creek, which collects the acid facultatively anaerobic, thermoacidophilic autotrophic or mine drainage). Here, we present a survey of the archaeal heterotrophic organisms like the organotrophic genus Ther- community in the sediment and its temporal variation using a mogymnomonas. Some members of Thermoplasmatales can culture-independent approach by cloning of 16S rRNA also derive energy from sulfur/iron oxidation or reduction. We encoding genes. The taxonomic affiliation of Archaea showed also found microorganisms affiliated with methanogenic Archaea (Methanomassiliicoccus luminyensis), which are involved in the carbon cycle. Some sequences affiliated with Communicated by F. Robb. ammonia oxidizers, involved in the first and rate-limiting step Electronic supplementary material The online version of this in nitrification, a key process in the nitrogen cycle were also article (doi:10.1007/s00792-012-0466-8) contains supplementary observed, including Candidatus Nitrososphaera viennensis material, which is available to authorized users. and Candidatus nitrosopumilus sp. These results suggest that Archaea may be important players in the Reigous sediments A. Volant A. Desoeuvre C. Casiot S. Delpoux J. C. Personne´ M. He´ry F. Elbaz-Poulichet O. Bruneel (&) through their participation in the biochemical cycles of ele- Laboratoire HydroSciences Montpellier, HSM, UMR 5569 ments, including those of carbon and nitrogen. (IRD, CNRS, Universite´s Montpellier 1 et 2), ´ Universite Montpellier 2, Place E. Bataillon, CC MSE, Keywords Archaea Diversity Arsenic 34095 Montpellier, France e-mail: [email protected] Acid mine drainage Lead and zinc mine

B. Lauga Equipe Environnement et Microbiologie, EEM, UMR 5254 Introduction (IPREM, CNRS), Universite´ de Pau et des Pays de l’Adour, BP 1155, 64013 Pau, France Acid mine drainage (AMD) water is a worldwide environ- G. Morin mental problem caused by active and abandoned mines Institut de Mine´ralogie et de Physique des Milieux Condense´s, (Johnson and Hallberg 2003). Mining and processing of sul- IMPMC, UMR 7590 (CNRS, Universite´ Pierre et marie curie/Paris 6), 4 place Jussieu, 75252 Paris, France fide-rich ores produce large amounts of pyrite-rich waste. In contact with meteoric water, oxidation of this material gen- P. N. Bertin erates AMD. These effluents are generally characterized by a Laboratoire de Geńe´tique Molećulaire, Geńomique, low pH and contain significant quantities of sulfates, metals Microbiologie, GMGM, UMR 7156 (Universite´ de Strasbourg, CNRS), De´partement Microorganismes, Geńomes, and metalloids including arsenic. AMD generation is mainly Environnement, 28 Rue Goethe, 67083 Strasbourg, France mediated by acidophilic iron-oxidizing microorganisms 123 Author's personal copy

646 Extremophiles (2012) 16:645–657

(Edwards et al. 1999). Natural remediation of AMD can be solid sulfidic wastes containing 0.7 wt% lead, 10 wt% iron observed at the Carnoule`s site (France) or at Rio Tinto (Spain) and 0.2 wt% arsenic, which are stored behind a 6 m high (Casiot et al. 2003;Sańchez Espanãetal.2005). This natural dam on the uppermost course of Reigous Creek. The remediation of metal pollutants is generally due to the seepage water, which percolates through the wastes, occurrence of abiotic reactions and/or microbial activities that emerges at the base of the tailings dam, and is the initial make these toxic compounds insoluble and lead to their source of Reigous Creek. The water is acidic (2 \ pH \ 3) accumulation in sediments (Hallberg 2010). These precipi- and rich in dissolved sulfate, iron and arsenic (2000–7700, tations mainly involve the oxidation and precipitation of iron 500–1000 and 50–350 mg L-1, respectively) the later being and the adsorption of other metals and metalloids by the predominantly in reduced forms: Fe(II) and As(III) (Casiot resulting ferric minerals. Sulfate-reducing bacteria also have et al. 2003). The arsenic concentration decreases within the the ability to reduce sulfate to sulfide, which then reacts with first 30 m of the creek mainly due to bacterial iron oxidation certain dissolved metals to form insoluble precipitates which leads to the coprecipitation of 20–60 % of dissolved (Hallberg 2010). In addition, several bacteria contribute to the arsenic (Casiot et al. 2003). Although the arsenic level immobilization of arsenic, thanks to their ability to oxidize remains high, its concentration subsequently decreases by this metalloid, arsenate As(V) being less soluble than arsenite around 95 % between the source of Reigous Creek and its As(III) (Bowell 1994). confluence with the Amous River, 1.5 km downstream. The The microbiology of AMD streams has been the subject of precipitates, which form around stromatolitic-like bacterial numerous studies. While a large amount of information is structures, are mainly composed of Fe(III)–As(III) in winter available on acidophilic bacteria indigenous to AMD, little is in the first 10 m and of amorphous Fe(III)–As(V) during the known about Archaea (Hallberg 2010). Several studies evi- rest of the year (Casiot et al. 2003; Morin et al. 2003). Many denced the presence of archaeal communities in acidic studies (including culture-dependent and independent) have waters (Edwards et al. 2000; Dopson et al. 2004). The been conducted on the bacterial communities inhabiting the Archaea reported in AMD systems include groups of sulfur Carnoule`s mine. Two of them focused specifically on sed- and/or iron-oxidizers, such as Sulfolobus, Acidianus, Met- iment. The active bacterial species were identified in the allosphaera, Sulfurisphaera, and Ferroplasma (Edwards sediments in the April 2006 library using high sensitivity et al. 2000; Golyshina et al. 2000; Baker and Banfield 2003). nanoLC-chip-MS/MS methods combined with a 16S rRNA It has consequently been suggested that Archaea could also based phylogenetic approach (Bruneel et al. 2011). This play a major role in the generation and remediation of AMD study showed that Gallionella, Thiomonas, Acidithiobacil- via oxidation of iron (Baker and Banfield 2003). Archaea lus ferrooxidans, and Acidiphilium actively expressed pro- may also play a role in the biogeochemical cycling of arsenic, teins in situ. Meta- and proteo-genomics approaches were for example, through the presence of Archaea that respire also used on sediments in the May 2007 library and allowed As(V) like Pyrobaculum aerophilum and Pyrobaculum reconstruction of seven bacterial strains (Bertin et al. 2011). arsenaticum (Huber et al. 2000; Oremland and Stolz 2003). These studies and previous results (Casiot et al. 2003; In a previous study, Bruneel et al. (2008) investigated the Morin et al. 2003) suggest that the large amounts of As(V)– archaeal community in water samples from Carnoule`s, an Fe(III) hydroxysulfate sediments forming at Carnoule`s may AMD very rich in metallic elements and especially arsenic result from the combined activities of iron-oxidizing compared to many others AMD. This study reported the microorganisms (such as At. ferrooxidans, Alicyclobacillus presence of Ferroplasma acidiphilum and sequences affili- ferrooxidans, Ferrimicrobium,orGallionella) and arsenic- ated to uncultured Thermoplasmatales archaeon. However, oxidizing microorganisms (such as Thiomonas sp.). the archaeal population that inhabits the arsenic-rich Reigous sediments has never been characterized. Thus, to improve Sampling procedure and measurement our understanding of AMD functioning, we characterized the of physicochemical properties archaeal communities present in sediment samples from the arsenic-rich AMD of the Carnoule`s mine (France) and their Four sampling campaigns were carried out in April 2006, temporal variations using a 16S rRNA encoding gene library. October 2008, January 2009 and November 2009. Samples were collected at the station called COWG (Carnoule`s Oxidizing Wetland, point G) located 30 m downstream of Materials and methods the spring (Bruneel et al. 2003). 5 cm deep pale yellow loosely packed sediments were collected at the bottom of Description of the study site the creek using a sterile spatula and pooled [global posi- tioning system (GPS) coordinates: 44107001.8000N/ The Pb-Zn Carnoule`s mine, located in southern France, was 4100006.9000E]. This sampling was done in three repli- closed in 1962. The mining extraction left about 1.2 Mt of cates. Solid phases were harvested by centrifugation and 123 Author's personal copy

Extremophiles (2012) 16:645–657 647 dried under vacuum before mineralogical and spectro- the manufacturer’s recommendations (MoBio Laboratories scopic analyses. The main physicochemical parameters Inc., Carlsbad, CA, USA). These triplicates were pooled (pH, T °C, and dissolved oxygen concentration) of the before PCR amplification. All extracted genomic DNA running water at the sampling point were measured in the samples were stored at -20 °C until further analysis. field. pH and water temperature were measured with an Ultrameter Model 6P (Myron L 125 Company, Camlab, PCR amplification Cambridge). Water samples (500 ml) were immediately filtered through 0.22 lm Millipore membranes fitted on Amplification of archaeal 16S rRNA genes was obtained Sartorius polycarbonate filter holders. For total iron and by PCR using primers Arch21F (50-TTCCGGTTGATCC arsenic determination, the filtered water was acidified to pH YGCCGGA-30) and Arch958R (50-YCCGGCGTTGAMTC 0 1 with HNO3 (14.5 M) and stored at 4 °C in polyethylene CAATT-3 ) (Delong 1992). Two PCR protocols were used bottles until analysis. For arsenic and iron speciation, a due to major amplification difficulties. The first PCR 20 ll aliquot of filtered water sample was added to either a amplification mixture contained 2 ll of DNA template, mixture of acetic acid and EDTA (Samanta and Clifford 2 ll of both primers (20 lM), 25 ll of PCR Master Mix 2005) for arsenic speciation or to a mixture of 0.5 ml Ampli Taq Gold 360 (Applied Biosystems, Foster City, acetate buffer (pH 4.5) and 1 ml of 1,10-phenanthrolinium CA, USA). Sterile distilled water was added to reach a final chloride solution (Rodier et al. 1996) for Fe(II) determi- volume of 50 ll. The PCR conditions were as follows, an nation. The vials were completed to 10 ml with deionized initial denaturation step of 95 °C for 7 min followed by 35 water. The samples for iron and arsenic speciation and denaturation cycles at 95 °C for 1 min, an annealing cycle sulfate determination were stored in the dark and analyzed at 55 °C for 45 s and an extension cycle at 72 °C for within 24 h. Chemical analysis were carried out as previ- 1 min. Final extension was at 72 °C for 10 min. As ously described (Bruneel et al. 2011). amplification of the January 2009 sample failed with this protocol, another enzyme was used, the PCR Extender Solid sample characterization Polymerase Mix (5Prime, Hamburg, Deutschland) as well as for a part of the November 2009 sample, which was also The mineralogical composition of the solid samples col- very difficult to amplify. The second PCR amplification lected at COWG was qualitatively determined using powder mixture contained 2 ll of DNA template, 2 ll of both X-ray diffraction analysis (XRD). Data were collected with primers (20 lM), 2.5 ll of dNTPs 10 mM, 5 ll reaction Co K-alpha radiation on an X’Pert PRO P analytical dif- Tunning buffer 910 and 0.5 ll of PCR Extender Poly- fractometer equipped with an X’Celerator detector, in con- merase Mix (5Prime, Hamburg, Deutschland). Sterile dis- tinuous mode and a counting time of 4 h per sample. X-ray tilled water was added to reach a final volume of 50 ll. The absorption spectroscopy data were collected on the solid PCR conditions were as follows: initial denaturation step at phases sampled at COWG in October 2008, January 2009, 94 °C for 3 min followed by 35 denaturation cycles at and November 2009. X-ray absorption near edge structure 94 °C for 1 min, an annealing cycle at 55 °C for 1 min, (XANES) and extended X-ray absorption fine structure and an extension cycle at 72 °C for 1.5 min. Final exten- (EXAFS) spectra were recorded at a temperature 10–15 K sion was at 72 °C for 10 min. PCR products were purified in fluorescence mode on the FAME BM30B bending mag- with the GFX PCR DNA purification kit (Amersham- net beamline at ESRF (Grenoble, France). Data for the April Pharmacia). The PCR Extender polymerase mix creates 2006 COWG sample were previously collected at the 11-2 blunt ended products. For TA CloningÒ,30 A-overhangs wiggler beamline at SSRL (Stanford, CA) and analyzed in are needed on these PCR products, which are obtained with Bruneel et al. (2011). Experimental details and data reduc- a different Taq polymerase. To 25 ll of purified PCR tion procedures are reported in previous studies (Morin et al. product, 2.5 ll of buffer 109, 0.5 ll of dATPs 10 mM, and 2003; Ona-Nguema et al. 2005; Hohmann et al. 2011). 0.5 ll of Taq DNA polymerase (EurobiotaqÒ, Eurobio, XANES and EXAFS data were interpreted by linear com- France) were added. The PCR amplification mixture was bination fitting using a set of model compound spectra. This then incubated at 72 °C for 20 min. set includes As(V)- and As(III)–Fe(III) oxyhydroxides and oxyhydroxysulfates synthesized via biotic and abiotic Cloning and 16S rRNA gene sequencing pathways (Morin et al. 2003; Maillot 2011). The PCR products were cloned in E. coli TOP 10 strain DNA isolation using the pCR2.1 Topo TA cloning kit (Invitrogen, Inc., Carlsbad, CA, USA). Cloned 16S rRNA gene fragments Triplicate genomic DNA was extracted from sediments were re-amplified using the primers TOP1 (50-GTGTGCT using the UltraClean Soil DNA Isolation Kit according to GGAATTCGCCCTT-30) and TOP2 (50-TATCTGCAGAA 123 Author's personal copy

648 Extremophiles (2012) 16:645–657

TTCGCCCTT-30) that anneal to the vector and flank the values were calculated to determine how efficiently the inserted PCR fragment. A total of 340 clones from the four libraries described the complexity of a theoretical com- libraries were sequenced. Partial sequences of the clones munity like an original archaeal community. The coverage were determined by the dideoxy nucleotide chain-termi- (Good 1953) value is given as C = 1 - (n1/N) where n1 is nation method using the BigDye 3.1 kit (Applied Bio- the number of clones that occurred only once in the library. systems) on an ABI PRISM 3730XL Genetic analyzer To determine the significance of differences between (Applied Biosystems). The MALLARD program (Ashelford archaeal libraries, a LIBSHUFF statistical analysis was et al. 2006) was used to detect and then remove chimera. performed in Mothur following Singleton et al.’s (2001) Sequences were also examined manually for chimera, which method. A LIBSHUFF comparison of libraries yielded the were excluded from further analyses. Sequences were then following formula using the Bonferroni correction: aligned in Mothur (http://www.mothur.org)(Schlossetal. 0.05 = 1-(1 - a)k(k - 1), where a is the critical P value 2009) using the SILVA archaeal database as reference and k is the number of libraries. The critical P value is alignment. The same program was used to calculate a 0.0042 when four libraries are compared. If any comparison neighbor-joining (NJ) (Saitou and Nei 1987) distance of two libraries has a P value below or equal to 0.0042, then matrix using the Jukes-Cantor (JC) correction. The matrix there is 95 % confidence to believe that the two libraries was then used to assign sequences to operational taxo- concerned differ significantly in community composition. nomic units (OTUs) defined at 97 (species level) and 85 % Jaccard and Yue & Clayton theta tree clustering analysis (class level) cutoff using the furthest-neighbor algorithm. (Yue and Clayton 2005) were also performed in Mothur to Sequences were compared with the available databases identify community membership and structure relationships NCBI and Greengenes (http://greengenes.lbl.gov)by between the libraries. BLAST online searches (Altschul et al. 1990) and Mothur to identify their taxonomic identities. Representative sequences for each OTU defined at 97 % cutoff were Results identified using the tool implemented in Mothur and were submitted to the EMBL databases under accession numbers Physical and chemical characteristics of samples (HE653775–HE653816). The physicochemical characteristics of the waters are listed Phylogenetic analysis in Table 1. The physicochemistry of Reigous Creek water at these sampling periods was typical of that observed Archaeal 16S rRNA gene homologs were collected from during a previous long-term monitoring study (Egal et al. the database at NCBI using the BLAST program with 2010). The water samples were acid (pH = 2.91–3.28) default parameters; one representative of each OTU was and very rich in sulfate (1830–3400 mg L-1), iron selected, giving a dataset of 99 sequences for final analy- (510–1735 mg L-1), and arsenic (70–194 mg L-1), with sis. Multiple sequence alignment of partial prokaryotic predominance of the reduced forms Fe(II) and As(III). sequences was performed using Clustal W (Thompson Dissolved oxygen concentrations ranged from 3.5 to et al. 2000). A maximum likelihood phylogenetic recon- 7.86 mg L-1. The January 2009 sample showed the lowest struction was obtained using the PhyML program (Guindon iron, arsenic, and sulfate concentrations. and Gascuel 2003) with the GTR model, four evolutionary The nature and structure of the sediment samples were rates, a calculated proportion of invariant sites and calcu- investigated using mineralogical and spectroscopic meth- lated nucleotide frequencies (default parameters). Statisti- ods. XANES analyses at the arsenic K-edge showed that, cal likelihood at nodes was calculated via a likelihood-ratio despite the presence of an As(III) component equal to test (Anisimova and Gascuel 2006). 12–34 ± 5 % of total arsenic, the oxidized arsenic form As(V) predominated in all the sediments (Fig. S1). EXAFS Statistical analysis of diversity and comparison data (Fig. S2) showed that As(V) was mainly present in the of archaeal libraries samples in an amorphous Fe(III)–As(V) hydroxysulfate phase, as previously observed (Morin et al. 2003; Bruneel The Mothur software package was also used to generate et al. 2011), As(III) being likely sorbed to poorly ordered diversity indices and statistics (OTUs, total clones, single- schwertmannite. For January 2009, there was not enough tons, Chao1, Shannon, evenness, coverage) for each clone time exposure to X-ray beam in EXAFS analysis to record library as sequence similarity with a 97 % cutoff. The total this sample, however, based on the XANES data we can number of clones obtained compared with the number of assume that this sample should be similar to the others. clones representing each unique phylotype was used to XRD analyses (Fig. S3) showed that these arsenic-bearing produce the rarefaction curves at the 85 % level. Coverage phases were mixed with sandy components (quartz, 123 Author's personal copy

Extremophiles (2012) 16:645–657 649

Table 1 Physico-chemical characteristics of the water (mg L-1) during sampling at COWG

2- Sampling period pH (±SD) T (°C) DO (±SD) As(III) (±SD) As(V) (±SD) Fe (total) (±SD) Fe(II) (±SD) SO4 (±SD)

April 2006 3.28 (±0.05) 12.9 3.5 (±0.5) 69 (±3) 71 (±4) 620 (±30) 560 (±28) 2700 (±300) October 2008 3.13 (±0.05) 14.3 5.7 (±0.1) 133 (±7) 20 (±1) 1250 (±62) 1220 (±61) 3400 (±340) January 2009 2.91 (±0.05) 9.4 5.5 (±0.5) 43 (±2) 27 (±1) 510 (±25) 540 (±27) 1830 (±183) November 2009 3.26 (±0.05) 13.1 7.9 (±0.1) 161 (±8) 33 (±2) 1735 (±87) 1440 (±72) 3300 (±330) SD standard deviation

K-feldspar and micas) and that pyrite was only detected after October 2008.

Diversity analysis

A total of 340 clones obtained from the four independent 16S rRNA gene libraries were fully sequenced and phylogenetically analyzed. Thirteen sequences were identified as likely chimeras and excluded from further analyses. Sequencing and phylogenetic analysis of the 327 remaining cloned sequences led to the identification of 9 and 42 OTUs defined at two different levels of identity (85 and 97 %, respectively). Rarefaction curves calculated at the class level (85 % identity, the rank usually used for representing the microbial community) were near saturation (Fig. 1). Table 2 shows the Shannon, evenness, and Chao1 indices Fig. 1 Rarefaction curves of the archaeal 16S rRNA sequences from ` and the coverage values calculated for each library at 97 % Carnoules mine sediments at 85 % identity. The total number of sequenced clones is plotted against the number of OTUs observed in identity. The coverage values of the four clone libraries the same library (90, 88, 96 and 92, respectively, for April 2006, October 2008, January 2009, and November 2009) indicate that the clone libraries were sufficiently sampled. The estimations Fig. 2a) showed that April 2006 and November 2009 were of the diversity indices show that the structure and mem- more related to each other in this respect, whereas April bership composition of the archaeal community changed 2006 and October 2008 were more related to each other in over the sampling period. The Shannon diversity (H) and terms of community structure (Yue & Clayton index, Chao1 richness indices ranged from 1.37 to 2.57 and 6.5 to Fig. 2b). 30.5, respectively. The diversity (H = 1.37) and richness (Chao1 = 6.5) were significantly lower in January 2009 Phylogenetic analysis of archaeal community whereas November 2009 displayed the highest values (H = 2.57; Chao1 = 30.5), which is consistent with the Four 16S rRNA encoding gene libraries were constructed rarefaction curves. each containing a distinct archaeal community, which varied in composition and richness throughout the sam- Comparison of archaeal community pling campaigns. In April 2006, the 16S rRNA phylogenetic reconstruc- The overall community structure was analyzed for each tion (Fig. 3) showed that all the sequences corresponding sample using the Mothur software package. LIBSHUFF to 17 OTUs (OTUs 1–17) were affiliated to the phylum analysis was performed to compare the OTU compositions Euryarchaeota, as previously observed in the water samples of each clone library revealing a high degree of variation from Carnoule`s (Bruneel et al. 2008). The most abundant between individuals and showing that with Bonferroni OTU (OTU 1, 53 clones representing around 61 % of the correction, each library differed significantly from all sample) was affiliated to the order Thermoplasmatales others (Table 3). The resulting dendrograms of Jaccard and which contained 97 % of the sequences grouped in 15 Yue & Clayton theta similarity coefficient analysis (Fig. 2) OTUs. Within this order, the majority of the OTUs were identified one major cluster and one outlier (January 2009). closely related to uncultured clones from an acidic envi- The similarity in community membership (Jaccard index, ronment such as acidic mine water and sediments (Fig. 3).

123 Author's personal copy

650 Extremophiles (2012) 16:645–657

Table 2 Diversity indices and statistics calculated for the four clone libraries from COWG station at different sampling periods Clone library No. of sequences No. of OTUsa Singletons Good’s coverageb Shannon diversityc Chao1 richness

April 2006 87 17 9 90 1.63 24.2 October 2008 80 20 10 88 1.96 27.5 January 2009 47 6 2 96 1.37 6.5 November 2009 113 25 9 92 2.57 30.5 a OTUs were deﬁned at 97 % cutoff b Coverage: sum of probabilities of observed classes calculated as (1 - (n/N)), where n is the number of singleton sequences and N is the total number of sequences c Takes into account the number and evenness of species

Table 3 Community comparison using LIBSHUFF Y library Apr-06 Oct-08 Jan-09 Nov-09

X library Apr-06 – 0.0260 \0.0001* 0.1529 Oct-08 0.0001* – \0.0001* 0.0001* Jan-09 \0.0001* \0.0001* – 0.2222 Nov-09 \0.0001* \0.0001* \0.0001* – Fig. 2 Similarities in archaeal community membership (Jaccard a) * Signiﬁcant difference. Bonferroni correction P value = 0.0042 and in community structure (Yue & Clayton b) between samples. Values are based on 0.03 distances – Not compared

The BLAST affiliation (Table 4) showed that some of 22–27, and 31) which accounted for 86 % of the total these OTUs displayed 89–94 % similarity with Thermo- archaeal clones including the same most abundant OTU gymnomonas acidicola, a moderately thermophilic, acido- (OTU1; around 54 %, Fig. 3). The BLAST affiliation philic, strictly aerobic heterotroph that uses yeast extract, (Table 4) also revealed similarity of some OTUs with as well as glucose and mannose (in the presence of yeast Thermogymnomonas acidicola. Three OTUs (18, 21, and extract) as carbon and energy sources (Itoh et al. 2007). 28) affiliated with uncultured clones isolated from acidic Additionally, OTU 15 related to the uncultured clone environments (clone SALE1B1 and clone anta6) and from ORCMO 26 retrieved from a copper mine drainage (Rowe a forested wetland impacted by reject coal (clone ARCP2- et al. 2007, Fig. 3) was found. This OTU was assigned 12) (Brofft et al. 2002; Garcıá-Moyano et al. 2007, Fig. 3), to methanogenic lineage (Methanomicrobia, Fig. 3) with respectively, were assigned to Methanomicrobia. The the closest relative Methanomassiliicoccus luminyensis, a remaining OTUs (OTUs 29 and 30) were affiliated with methanogenic Archaea recently isolated from human environmental sequences originating from acidic soil and faeces (Dridi et al. 2011). However, the Greengenes clas- acidic hot springs, which likely represent uncultured lin- sification (Table 4) assigned this OTU to the order Ther- eages of Thaumarchaeota. moplasmatales. Lastly, an unknown group belonging A significant change in the archaeal community to the Euryarchaeota and represented by OTU 8 was appeared in the January 2009 library, when diversity detected. This group formed an independent branch that decreased and no cultured species were identified. Indeed, was distantly related to the identified groups and showed almost all the sequences (96 %) clustered in five OTUs low similarity with the uncultured archaeon clone hfm29 (OTUs 8, 32, 33, 34, and 35, Fig. 3) were related to the isolated from an iron-rich microbial mat (Kato et al. in uncultured archaeon clone hfmA029 previously found in press). April 2006. This group formed an independent branch that Twenty OTUs were retrieved from the October 2008 was far away from the remaining groups. This clone dis- library, 18 of which belonged to the Euryarchaeota and two played around 97 % similarity with Methanothermobacter to the Thaumarchaeota (Fig. 3). Like in April 2006, most thermautotrophicus, an autotrophic thermophilic methan- of the Euryarchaeota sequences were affiliated with the ogen recovered from an anaerobic sewage sludge digester Thermoplasmatales (OTUs 1, 6, 9, 11, 12, 17, 19, 20, (Zeikus and Wolee 1972). Remaining sequences grouped

123 Author's personal copy

Extremophiles (2012) 16:645–657 651

Fig. 3 Maximum likelihood tree of 16S rRNA gene homologs from scale bar gives the average number of substitutions per site. The the archaeal clones (in bold) along with a selection of representatives number in parenthesis indicates the number of clones for the sampling of archaeal diversity. Numbers at nodes indicate a LTR (approximate period which is represented by a symbol (star April 2006, square likelihood ratio test) branch support as computed by PhyML. The October 2008, circle January 2009 and diamond November 2009)

123 Author's personal copy

652 Extremophiles (2012) 16:645–657

Table 4 Identification number of the OTUs retrieved from the Reigous Creek sediment of Carnoule`s mine, taxonomic affiliation and representative sequence for each OTU OTU Number of Representative Taxonomic affiliation Closest relative (% of identity) ID sequences sequence Phylum Class Order

1 132 ArCMSdO8D35 Euryarchaeota Thermoplasmata Thermoplasmatales Aciduliprofundum sp. EPR07-39 (85 %) 2 15 ArCMSdA6A12 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (89 %) 3 2 ArCMSdA6A86 Euryarchaeota Thermoplasmata Thermoplasmatales Aciduliprofundum sp. EPR07-39 (85 %) 4 4 ArCMSdA6A46 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (90 %) 5 1 ArCMSdA6A17 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (91 %) 6 16 ArCMSdA6A67 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (92 %) 7 3 ArCoSdN9H63 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (93 %) 8 24 ArCMSdJ9B78 Euryarchaeota – – Clone hfmA029 (86 %) 9 4 ArCMSdA6A30 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (88 %) 10 3 ArCoSdN9D80 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (94 %) 11 7 ArCMSdO8B50 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (94 %) 12 4 ArCMSdO8B53 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (90 %) 13 1 ArCMSdA6A52 Euryarchaeota Thermoplasmata Thermoplasmatales Thermoplasma volcanium (84 %) 14 2 ArCMSdA6A84 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (90 %) 15 3 ArCoSdN9H35 Euryarchaeota Thermoplasmata – Methanomassiliicoccus luminyensis B10 (80 %) 16 1 ArCMSdA6A92 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (91 %) 17 4 ArCoSdN9H43 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (88 %) 18 5 ArCMSdO8A3 Euryarchaeota Thermoplasmata – Methanomassiliicoccus luminyensis B10 (82 %) 19 1 ArCMSdO8A13 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (89 %) 20 1 ArCMSdO8A16 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (87 %) 21 2 ArCoSdN9H67 Euryarchaeota Thermoplasmata – Methanomassiliicoccus luminyensis B10 (83 %) 22 1 ArCMSdO8A24 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (88 %) 23 2 ArCMSdO8A56 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (89 %) 24 4 ArCMSdO8A49 Euryarchaeota Thermoplasmata Thermoplasmatales Thermoplasma volcanium GSS1 (89 %) 25 1 ArCMSdO8A54 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (90 %) 26 3 ArCMSdO8A74 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (99 %) 27 1 ArCMSdO8A85 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (93 %)

123 Author's personal copy

Extremophiles (2012) 16:645–657 653

Table 4 continued OTU Number of Representative Taxonomic afﬁliation Closest relative (% of identity) ID sequences sequence Phylum Class Order

28 1 ArCMSdO8A89 Euryarchaeota Thermoplasmata – Methanomassiliicoccus luminyensis B10 (85 %) 29 5 ArCMSdJ9A29 – – – Candidatus Nitrosocaldus yellowstonii HL72 (84 %) 30 2 ArCMSdO8C25 – – – Candidatus Nitrososphaera gargensis (83 %) 31 1 ArCMSdO8E23 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (92 %) 32 15 ArCMSdJ9C75 Euryarchaeota – – Clone SVB_Fis_02_pl37c09 (86 %) 33 16 ArCMSdJ9C68 Euryarchaeota – – Clone hfmA029 (85 %) 34 1 ArCMSdJ9C55 Euryarchaeota Methanomicrobia Methanomicrobiales Clone hfmA029 (84 %) 35 2 ArCoSdN9A45 Euryarchaeota – – Clone hfmA029 (85 %) 36 15 ArCoSdN9B9 Thaumarchaeota No class Nitrososphaerales Candidatus Nitrososphaera sp. EN76 (96 %) 37 6 ArCoSdN9F14 Thaumarchaeota No class Cenarchaeales Candidatus Nitrosopumilus sp. NM25 (93 %) 38 11 ArCoSdN9D53 Euryarchaeota – – Clone TG_FD0.2_SA043 (100 %) 39 2 ArCoSdN9H79 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (90 %) 40 1 ArCoSdN9E14 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (88 %) 41 1 ArCoSdN9G7 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (93 %) 42 1 ArCoSdN9H80 Euryarchaeota Thermoplasmata Thermoplasmatales Thermogymnomonas acidicola JCM 13583 (92 %) OTU definition and taxonomic identification of representative sequences were done using mothur (Schloss et al. 2009; see ‘‘Materials and methods’’ for details). Only taxonomic affiliations with 100 % similarity are shown. The closest relative was obtained by BLAST search on NCBI nr database

in a single OTU (OTU 29) belonged to an unknown group unknown groups of the Euryarchaeota (Fig. 3). Among of Thaumarchaeota previously found in October 2008 and these, the most abundant sequences belonging to OTU 38, were affiliated with clone GBX-A-COQ1-158 isolated displayed a strong similarity (99 %) with an uncultured from an acidic hot spring (Fig. 3). archaeon clone LC15_L00B08 isolated from the monim- An increase in archaeal diversity was observed in the olimnion of a stratified lake (Gregersen et al. 2009). The November 2009 library, with 25 OTUs belonging to four other OTUs (8, 32, 33, and 35) were affiliated with the Euryarchaeota (69 clones corresponding to 61 % of the the uncultured archaeon clones hfmA029 mainly present sample) and to the Thaumarchaeota (22 clones corre- in the January 2009 library. The Thaumarchaeota detected sponding to 19 % of the sample, Fig. 3). The sequences in this study fell into different lineages clustered in three from the Euryarchaeota were distributed in 22 OTUs. OTUs. The first (OTU 36) belonged to Thaumarchaeota Fifteen were related to the order Thermoplasmatales, 11 of group I.1b and the 15 sequences within this OTU showed which (OTUs 1, 2, 4, 6, 7, 9, 10, 11, 17, 24, 26) were from 95 to 96 % similarity with Candidatus Nitrososph- previously found in the April 2006 and October 2008 aera viennensis a chemolithoautotrophic ammonia-oxi- libraries (Fig. 3). As in the results observed in these two dizing archaeon (Tourna et al. 2011). The second (OTU sampling periods, OTU 1 was also the most abundant 37) was assigned to Thaumarchaeota group I.1a and the group in the sample in November 2009 (32 %). Addi- sequences displayed 92–93 % similarity with Candidatus tionally, OTUs 15 and 21, also found in the two first Nitrosopumilus sp., another ammonia-oxidizing prokaryote libraries, were assigned to the order Methanomicrobia. (Matsutani et al. 2011). The last OTU (OTU 29), previ- The remaining five OTUs (8, 32, 33, 35, and 38), were not ously found in October 2008 could not be related to any shown to be related to any known species and formed known species.

123 Author's personal copy

654 Extremophiles (2012) 16:645–657

Discussion although none of the clones could be identified with high similarity ([97 %) as belonging to any cultured species. Archaeal 16S rRNA gene analysis of the sediment sampled This order is represented by thermoacidophilic organisms at the Reigous Creek showed that the Carnoule`s archaeal (Reysenbach 2001), which often derive energy from sulfur community includes the phylum Euryarchaeota and oxidation or reduction. So far, the order contains three Thaumarchaeota. The relatively low archaeal diversity families, each represented by one genus: the Thermoplas- revealed by molecular-based methods is consistent with the mataecae, the Picrophilaceae and the Ferroplasmaceae results of studies in similar environments (Bond et al. 2000; (Itoh et al. 2007). The Thermoplasmataecae comprises Baker and Banfield 2003; Bruneel et al. 2008;Sańchez- species like Tp. acidophilum that couple the oxidation of Andrea et al. 2011). This may reflect the limited number of organic carbon with reduction of elemental sulfur, whereas different electron donors and acceptors available in this members of the Ferroplasmaceae are strict iron-oxidizing AMD and the high concentration of toxic compounds along chemolithotrophs such as Ferroplasma acidiphilum (Itoh with the low pH. Most of the phylotypes identified in this et al. 2007). Microorganisms affiliated with methanogenic study were related to genera and species usually found in Archaea such as Methanomassiliicoccus luminyensis were extreme environments (hot springs, acidic springs, hydro- also identified. Methanogenic communities play an thermal vents, etc.) and showed similarities with sequences important role in the global carbon cycle, completing the obtained in previous studies of Tinto River and other AMD conversion of organic carbon into methane gas by utilizing

(Sańchez-Andrea et al. 2011; Garcıá-Moyano et al. 2007; the metabolic products of bacteria (CO2,H2, acetate, and Rowe et al. 2007, Fig. 3). formate) and other simple methyl compounds available in Regarding the dynamics of the archaeal community, our the environment (Sanz et al. 2011). Lastly, we found study showed that significant modifications in this com- microorganisms involved in ammonia oxidation, a key step munity occurred throughout the sampling period. All the in the nitrogen cycle (Brochier-Armanet et al. 2011), with sampling periods showed differences in community struc- presence of sequences affiliated to Candidatus Nitro- ture and membership although April 2006 and October sosphaera viennensis and Candidatus Nitrosopumilus sp. 2008 were more similar in terms of community structure. Until recently, ammonia oxidation, the first nitrification Similarity coefficient analysis showed that January 2009 step of the nitrogen cycle was thought to be carried out was very different from all the other sampling periods. only by autotrophic ammonia-oxidizing bacteria (AOB) In January 2009, the archaeal community changed and belonging to the Beta- and Gammaproteobacteria lineages diversity decreased. Almost all the sequences were related (Purkhold et al. 2000) occasionally supported by hetero- to an uncultured archaeon clone hfmA029 affiliated with trophic nitrifiers in soil environments (De Boer and methanogenic lineage (Methanothermobacter thermautot- Kowalchuk 2001). Ammonia-oxidizing Archaea (AOA) are rophicus). This clone, hfmA029, previously found in April members of the proposed novel Phylum Thaumarchaeota, 2006 (OTU 8) in only 2 % of the sample became the and are currently being indentified in almost all environ- dominant population in January 2009. The differences in ments (Brochier-Armanet et al. 2008). These Archaea may the archaeal community observed in January 2009 may thus play a major role in the nitrogen cycle in the Carnoule`s result from a modification in the composition of the sedi- sediments. ment, although the physicochemical analysis of the sedi- Previous studies focused on the bacterial communities ments appeared to be similar throughout the sampling inhabiting the Carnoule`s AMD sediment. These studies period, and consisted mainly of an amorphous Fe(III)– showed that the active population of bacteria also con- As(V) hydroxysulfate mineral. Indeed, XRD analyses tained iron reducers, sulfate-reducing, and sulfur com- revealed that pyrite first appeared in October 2008. Like- pound oxidizers, and both autotrophic and heterotrophic wise, since late 2007, a leakage of fine grey sulfide-reach bacteria (Bruneel et al. 2011). Statistical analysis of sands out of the tailings pile has been observed after the genomic and proteomic data demonstrated that both met- rainfall events that generally occur in September and abolic specificity and partnerships can co-exist in this October. This is probably due to the corrosion of the drains arsenic-rich sediment (Bertin et al. 2011). These processes at the bottom of the tailing stock that are responsible for the include the fixation of inorganic carbon and nitrogen by water discharge inside the mine tailing. In January 2009, several strains, in particular those belonging to the Thio- the sulfide sands, originated from the tailings stock, formed monas, Acidithiobacillus, and Gallionella related genera. a very thick layer (around 3 cm deep) in the bottom of the However, this study did not find evidence for the presence creek which could explain the change in the archaeal of archaeal species among the dominant organisms, sug- community. gesting that they may represent a small proportion of the In the Reigous sediment, most of the sequences were microbial community in the sediment. Despite the fact that phylogenetically related to the order Thermoplasmatales, we cannot really infer the implication of the Archaea 123 Author's personal copy

Extremophiles (2012) 16:645–657 655 detected in most of these metabolic pathways because #3033 and by SESAME IdF Grant #1775. Part of the field chemical most of them could not be affiliated to cultured species, data was acquired through the OSU OREME. we can point to their probable implication in a specific metabolism currently unknown in bacteria (Forterre et al. 2002), methanogenesis. Archaea are also involved in the nitrogen cycle (Candidatus Nitrososphaera viennensis and References Candidatus Nitrosopumilus sp.) and some of them may Ther- Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic also be involved in the sulfur and iron cycles ( local alignment search tool. J Mol Biol 215(3):403–410 moplasmatales). All these microorganisms may contribute Anisimova M, Gascuel O (2006) Approximate likelihood-ratio test to the remediation process observed in situ and could also for branches: a fast, accurate, and powerful alternative. Syst Biol be involved in the stability of this sediment by changing 55(4):539–552 Ashelford KE, Chuzhanova NA, Fry JC, Jones AJ, Weightman AJ the ratio between oxidized and reduced forms of iron, (2006) New screening software shows that most recent large 16S arsenic, and sulfur compounds, promoting the formation rRNA gene clone libraries contain chimeras. Appl Environ and/or dissolution of the Fe(III)–As(V) hydroxysulfate Microbiol 72(9):5734–5741 precipitates. Baker BJ, Banfield JF (2003) Microbial communities in acid mine drainage. FEMS Microbiol Ecol 44(2):139–152 Because isolation and phenotypic characterization of Bertin PN, Heinrich-Salmeron A, Pelletier E, Goulhen-Chollet F, many environmental Archaea are currently not possible, Arse`ne-Ploetze F, Gallien S, Lauga B, Casiot C, Calteau A, the physiological features and ecological significance of Vallenet D, Bonnefoy V, Bruneel O, Chane-Woon-Ming B, some Archaea detected in this AMD remain difficult to Cleiss-Arnold J, Duran R, Elbaz-Poulichet F, Fonknechten N, assess. Moreover, the fact that most of the archaeal Giloteaux L, Halter D, Koechler S, Marchal M, Mornico D, Schaeffer C, Smith AAT, Van Dorsselaer A, Weissenbach J, sequences were only distantly related (\94 % similarity) to Medigue C, Le Paslier D (2011) Metabolic diversity among main known archaeal species suggests that other taxa may exist. microorganisms inside an arsenic-rich ecosystem revealed by Additionally, the contradictions observed in the taxonomic meta- and proteo-genomics. ISME J 5(11):1735–1747 Bond PL, Smriga SP, Banfield JF (2000) Phylogeny of microorgan- affiliation resulting from the 16S rRNA phylogenetic isms populating a thick, subaerial, predominantly lithotrophic reconstruction (Fig. 3) and the Greengenes classification biofilm at an extreme acid mine drainage site. Appl Environ (Table 4) suggest that there is still a lack of information Microbiol 66(9):3842–3849 making the taxonomic identification difficult to assess. Bowell RJ (1994) Sorption of arsenic by iron oxides and oxyhydroxides in soils. Appl Geochem 9(3):279–286 Indeed, almost half of the 16S rRNA gene sequences Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P (2008) archived in GenBank database lacks clear taxonomic Mesophilic crenarchaeota: proposal for a third archaeal phylum, information (DeSantis et al. 2006). As a consequence, the Thaumarchaeota. Nat Rev Microbiol 6(3):245–252 different authors use different names for uncultured clus- Brochier-Armanet C, Forterre P, Gribaldo S (2011) Phylogeny and ters which lead to conflicting nomenclatures. Recently evolution of the Archaea: one hundred genomes later. Curr Opin Microbiol 14(3):274–281 developed high-throughput techniques (metagenomics, Brofft JE, McArthur JV, Shimkets LJ (2002) Recovery of novel metaproteomics, and microarrays) may help link the bacterial diversity from a forested wetland impacted by reject identity of AMD-promoting prokaryotes to their function in coal. Environ Microbiol 4(11):764–769 . Bruneel O, Personne JC, Casiot C, Leblanc M, Elbaz-Poulichet F, mining environments (Mohapatra et al. 2011; Bertin et al Mahler BJ, Le Fleche A, Grimont PA (2003) Mediation of 2011) in the absence of laboratory culture. In the future, arsenic oxidation by Thiomonas sp. in acid-mine drainage these new genomic tools should provide a more precise (Carnoules, France). J Appl Microbiol 95(3):492–499 assessment of the archaeal diversity that will probably lead Bruneel O, Pascault N, Egal M, Bancon-Montigny C, Gon˜i-Urriza M, Elbaz-Poulichet F, Personne´ JC, Duran R (2008) Archaeal to substantial changes in current archaeal phylogeny and diversity in a Fe–As rich acid mine drainage at Carnoule`s taxonomy (Brochier-Armanet et al. 2008; Schleper et al. (France). Extremophiles 12(4):563–571 2005) and to a better understanding of the evolution and Bruneel O, Volant A, Gallien S, Chaumande B, Casiot C, Carapito C, metabolic capacities of uncultured Archaea. In conclusion, Bardil A, Morin G, Brown GE, Personne´ JC, Le Paslier D, Schaeffer C, Van Dorsselaer A, Bertin PN, Elbaz-Poulichet F, to increase our insight into the functioning of these highly Arsene-Ploetze F (2011) Characterization of the active bacterial acidic environments and to elucidate the role of these community involved in natural attenuation processes in arsenic- microorganisms, both improving culture strategies for rich creek sediments. Microb Ecol 61(4):793–810 further physiological and metabolic characterization of Casiot C, Morin G, Juillot F, Bruneel O, Personne´ JC, Leblanc M, Duquesne K, Bonnefoy V, Elbaz-Poulichet F (2003) Bacterial newly detected species and a greater sequencing effort are immobilization and oxidation of arsenic in acid mine drainage still needed. (Carnoule`s creek, France). Water Res 37(12):2929–2936 De Boer W, Kowalchuk GA (2001) Nitrification in acid soils: micro- Acknowledgments The French CRG is gratefully acknowledged organisms and mechanisms. Soil Biol Biochem 33(7–8):853–866 for provision of beamtime on the FAME BM30B beamline. This work DeLong EF (1992) Archaea in coastal marine environments. Proc was supported by EC2CO CNRS/INSU program, by ACI/FNS Grant Natl Acad Sci USA 89(12):5685–5689

123 Author's personal copy

656 Extremophiles (2012) 16:645–657

DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Maillot F (2011) Structure locale des nano-oxyhydroxydes de fer(III) Huber T, Dalevi D, Hu P, Andersen GL (2006) Greengenes, a de type ferrihydrite et schwertmannite. PhD thesis, UPMC, Paris chimera-checked 16S rRNA gene database and workbench Matsutani N, Nakagawa T, Nakamura K, Takahashi R, Yoshihara K, compatible with ARB. Appl Environ Microbiol 72(7):5069– Tokuyama T (2011) Enrichment of a novel marine ammonia- 5072 oxidizing archaeon obtained from sand of an eelgrass zone. Dopson M, Baker-Austin C, Hind A, Bowman JP, Bond PL (2004) Microbes Environ 26(1):23–29 Characterization of Ferroplasma isolates and Ferroplasma Mohapatra BR, Douglas Gould W, Dinardo O, Koren DW (2011) acidarmanus sp. nov., extreme acidophiles from acid mine Tracking the prokaryotic diversity in acid mine drainage- drainage and industrial bioleaching environments. Appl Environ contaminated environments: a review of molecular methods. Microbiol 70(4):2079–2088 Miner Eng 24(8):709–718 Dridi B, Fardeau ML, Ollivier B, Raoult D, Drancourt M (2011) The Morin G, Juillot F, Casiot C, Bruneel O, Personne JC, Elbaz- antimicrobial resistance pattern of cultured human methanogens Poulichet F, Leblanc M, Ildefonse P, Calas G (2003) Bacterial reflects the unique phylogenetic position of archaea. J Antimicrob formation of tooeleite and mixed arsenic(III) or arsenic(V)- Chemother 66(9):2038–2044 iron(III) gels in the Carnoules acid mine drainage, France. Edwards KJ, Goebel BM, Rodgers TM, Schrenk MO, Gihring TM, A XANES, XRD, and SEM study. Environ Sci Technol Cardona MM, Hu B, McGuire MM, Hamers RJ, Pace NR, 37(9):1705–1712 Banfield JF (1999) Geomicrobiology of pyrite (FeS2) dissolu- Ona-Nguema G, Morin G, Juillot F, Calas G, Brown GE Jr (2005) tion: Case study at Iron Mountain, California. Geomicrobiol J EXAFS analysis of arsenite adsorption onto two-line ferrihy- 16(2):155–179 drite, hematite, goethite, and lepidocrocite. Environ Sci Technol Edwards KJ, Bond PL, Gihring TM, Banfield JF (2000) An archaeal 39(23):9147–9155 iron-oxidizing extreme acidophile important in acid mine Oremland RS, Stolz JF (2003) The ecology of arsenic. Science drainage. Science 287(5459):1796–1799 300(5621):939–944 Egal M, Casiot C, Morin G, Elbaz-Poulichet F, Cordier MA, Bruneel Purkhold U, Pommerening-Roser A, Juretschko S, Schmid MC, O (2010) An updated insight into the natural attenuation of As Koops HP, Wagner M (2000) Phylogeny of all recognized concentrations in Reigous Creek (southern France). Appl species of ammonia oxidizers based on comparative 16S rRNA Geochem 25(12):1949–1957 and amoA sequence analysis: implications for molecular diver- Forterre P, Brochier C, Philippe H (2002) Evolution of the Archaea. sity surveys. Appl Environ Microbiol 66(12):5368–5382 Theor Popul Biol 61(4):409–422 Reysenbach AL (2001) Order I. Thermoplasmatales ord. nov. In: Garcıá-Moyano A, Gonza´lez-Toril E, Aguilera A, Amils R (2007) Boone DR, Castenholz RW, Garrity GM (eds) Bergey’s manual Prokaryotic community composition and ecology of floating of systematic bacteriology, vol 1, 2nd edn. Springer, New York, macroscopic filaments from an extreme acidic environment, Rıó p 335 Tinto (SW, Spain). Syst Appl Microbiol 30(8):601–614 Rodier J, Broutin JP, Chambon P, Champsaur H, Rodi L (1996) Golyshina OV, Pivovarova TA, Karavaiko GI, Kondrateva TF, Moore L’Analyse des Eaux. Dunod, Paris, p 1383 ER, Abraham WR, Lunsdorf H, Timmis KN, Yakimov MM, Rowe OF, Sańchez Espanã J, Hallberg KB, Johnson DB (2007) Golyshin PN (2000) Ferroplasma acidiphilum gen. nov., sp. Microbial communities and geochemical dynamics in an nov., an acidophilic, autotrophic, ferrous-iron-oxidizing, cell- extremely acidic, metal-rich stream at an abandoned sulfide wall-lacking, mesophilic member of the Ferroplasmaceae fam. mine (Huelva, Spain) underpinned by two functional primary nov., comprising a distinct lineage of the Archaea. Int J Syst production systems. Environ Microbiol 9(7):1761–1771 Evol Microbiol 50 Pt 3:997–1006 Saitou N, Nei M (1987) The neighbor-joining method: a new method Good IJ (1953) The population frequencies of species and the for reconstructing phylogenetic trees. Mol Biol Evol 4(4): estimation of population parameters. Biometrika 40(3/4):237– 406–425 264 Samanta G, Clifford DA (2005) Preservation of inorganic arsenic Gregersen LH, Habicht KS, Peduzzi S, Tonolla M, Canfield DE, species in groundwater. Environ Sci Technol 39(22):8877–8882 Miller M, Cox RP, Frigaard NU (2009) Dominance of a clonal Sańchez Espanã J, Pamo EL, Pastor ES, Andre´s JR, Rubı´ JAM (2005) green sulfur bacterial population in a stratified lake. FEMS The natural attenuation of two acidic effluents in Tharsis and La Microbiol Ecol 70(1):30–41 Zarza-Perrunal mines (Iberian Pyrite Belt, Huelva, Spain). Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm Environ Geol 49(2):253–266 to estimate large phylogenies by maximum likelihood. Syst Biol Sańchez-Andrea I, Rodriguez N, Amils R, Sanz JL (2011) Microbial 52(5):696–704 diversity in anaerobic sediments at Rio Tinto, a naturally acidic Hallberg KB (2010) New perspectives in acid mine drainage environment with a high heavy metal content. Appl Environ microbiology. Hydrometallurgy 104(3–4):448–453 Microbiol 77(17):6085–6093 Hohmann C, Morin G, Ona-Nguema G, Guigner JM, Brown GE Jr, Sanz JL, Rodriguez N, Diaz EE, Amils R (2011) Methanogenesis in Kappler A (2011) Molecular-level modes of As binding to the sediments of Rio Tinto, an extreme acidic river. Environ Fe(III) (oxyhydr)oxides precipitated by the anaerobic nitrate- Microbiol 13(8):2336–2341 reducing Fe(II)-oxidizing Acidovorax sp. strain BoFeN1. Geo- Schleper C, Jurgens G, Jonuscheit M (2005) Genomic studies of chim Cosmochim Acta 75(17):4699–4712 uncultivated archaea. Nat Rev Microbiol 3(6):479–488 Huber R, Sacher M, Vollmann A, Huber H, Rose D (2000) Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister Respiration of arsenate and selenate by hyperthermophilic EB, Lesniewski RA, Oakley BB, Parks DH, Robinson CJ, Sahl archaea. Syst Appl Microbiol 23(3):305–314 JW, Stres B, Thallinger GG, Van Horn DJ, Weber CF (2009) Itoh T, Yoshikawa N, Takashina T (2007) Thermogymnomonas Introducing mothur: open-Source, platform-independent, com- acidicola gen. nov., sp. nov., a novel thermoacidophilic, cell munity-supported software for describing and comparing micro- wall-less archaeon in the order Thermoplasmatales, isolated bial communities. Appl Environ Microbiol 75(23):7537–7541 from a solfataric soil in Hakone, Japan. Int J Syst Evol Microbiol Singleton DR, Furlong MA, Rathbun SL, Whitman WB (2001) 57(11):2557–2561 Quantitative comparisons of 16S rRNA gene sequence libraries Johnson DB, Hallberg KB (2003) The microbiology of acidic mine from environmental samples. Appl Environ Microbiol 67(9): waters. Res Microbiol 154(7):466–473 4374–4376

123 Author's personal copy

Extremophiles (2012) 16:645–657 657

Thompson JD, Plewniak F, Thierry J, Poch O (2000) DbClustal: rapid and Yue JC, Clayton MK (2005) A similarity measure based on species reliable global multiple alignments of protein sequences detected by proportions. Commun Stat Theory Methods 34(11):2123–2131 database searches. Nucleic Acids Res 28(15):2919–2926 Zeikus JG, Wolee RS (1972) Methanobacterium thermoautotrophicus Tourna M, Stieglmeier M, Spang A, Ko¨nneke M, Schintlmeister A, sp. n., an anaerobic, autotrophic, extreme thermophile. J Bacte- Urich T, Engel M, Schloter M, Wagner M, Richter A, Schleper C riol 109(2):707–713 (2011) Nitrososphaera viennensis, an ammonia oxidizing archaeon from soil. Proc Natl Acad Sci USA 108(20):8420–8425

123 RESEARCH ARTICLE Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and other drivers along an acid mine drainage Aurelie Volant1, Odile Bruneel1, Angelique Desoeuvre1, Marina Hery 1, Corinne Casiot1,Noelle€ Bru2, Sophie Delpoux1, Anne Fahy3, Fabien Javerliat3, Olivier Bouchez4, Robert Duran3, Philippe N. Bertin5, Francßoise Elbaz-Poulichet1 &Beatrice Lauga3

1Laboratoire HydroSciences Montpellier, HSM, UMR 5569 (IRD, CNRS, Universites Montpellier 1 et 2), Universite Montpellier 2, Montpellier, France; 2Laboratoire de Mathematiques et de leurs Applications, UMR 5142 (CNRS), Universite de Pau et des Pays de l’Adour, Pau, France; 3Equipe Environnement et Microbiologie, EEM, UMR 5254 (IPREM, CNRS), Universite de Pau et des Pays de l’Adour, Pau, France; 4INRA Auzeville, Plateforme Genomique Chemin de Borde Rouge, Castanet-Tolosan, France; and 5Departement Microorganismes, Genomes, Environnement, Laboratoire de Gen etique Moleculaire, Genomique, Microbiologie, GMGM, UMR 7156 (Universite de Strasbourg, CNRS), Strasbourg, France

Correspondence: Odile Bruneel, Laboratoire Abstract HydroSciences Montpellier, UMR5569, Universite Montpellier 2, Place E. Bataillon, Deciphering the biotic and abiotic factors that control microbial community CC MSE, 34095 Montpellier, France. structure over time and along an environmental gradient is a pivotal question Tel.: (+33)4 67 14 36 59; in microbial ecology. Carnoules mine (France), which is characterized by acid fax: (+33)4 67 14 47 74; waters and very high concentrations of arsenic, iron, and sulfate, provides an e-mail: [email protected] excellent opportunity to study these factors along the pollution gradient of Reigous Creek. To this end, biodiversity and spatiotemporal distribution of Received 16 April 2014; revised 10 July 2014; accepted 16 July 2014. bacterial communities were characterized using T-RFLP fingerprinting and high-throughput sequencing. Patterns of spatial and temporal variations in bac- DOI: 10.1111/1574-6941.12394 terial community composition linked to changes in the physicochemical conditions suggested that species-sorting processes were at work in the acid mine Editor: Tillmann Lueders drainage. Arsenic, temperature, and sulfate appeared to be the most important factors that drove the composition of bacterial communities along this contin- Keywords uum. Time series investigation along the pollution gradient also highlighted spatial and temporal dynamics; bacterial habitat specialization for some major members of the community (Acidithioba- diversity; acid mine drainage; arsenic. cillus and Thiomonas), dispersal for Acidithiobacillus, and evidence of extinction/ re-thriving processes for Gallionella. Finally, pyrosequencing revealed a broader phylogenetic range of taxa than previous clone library-based diversity. Overall, our findings suggest that in addition to environmental filtering processes, additional forces (dispersal, birth/death events) could operate in AMD community.

which leads to bioleaching, is well known, as is their role Introduction in natural attenuation of such polluted waters (Edwards Acid mine drainage (AMD) is one of the most pernicious et al., 2000; Hallberg, 2010; Johnson, 2012). Despite the forms of pollution in the world and is widely recognized central role of microorganisms in such ecosystem func- as having costly environmental and socioeconomic tioning, our understanding of the mechanisms shaping impacts (Hallberg, 2010). AMD occurs when waste from microbial community structure and diversity in AMD the extraction and processing of sulﬁde ore comes into remains limited. As pointed out by Miller et al. (2009), contact with oxygenated water. Drainages are typically deciphering how microbial communities are patterned acidic and usually contain high concentrations of sulfate, along environmental gradients is a pivotal question in

MICROBIOLOGY ECOLOGY MICROBIOLOGY metals and metalloids including arsenic. Although per- microbial ecology. Although AMD is characterized by ceived as extreme environments hostile to life, a variety changing conditions over time and space, few studies of microorganisms are able to thrive in it. For some of were interested in comparing the microbial communities them, their role in the oxidation of sulﬁde minerals, along such environmental gradients. The AMD of

Carnoules in southern France provides an excellent result of microbiologically mediated As–Fe coprecipitation opportunity to investigate these fundamental questions of (Morin et al., 2003; Bruneel et al., 2006). 10–47% of Fe, microbial ecology. This site is characterized by an acid and 20–60% of As are removed from the aqueous phase pH (2–4) and high levels of metal and metalloids, in par- within the first 30 m of the creek. Beyond this point ticular As (up to 10 g L1 in the tailings stock pore (COWG sampling site, located 30 m downstream from water, and 100–350 mg L1 at the source of Reigous the spring, Fig. 1), the Reigous receives water from quar- Creek), and natural attenuation processes result in a ries and mine galleries, especially after rainfall events, strong spatial pollution gradient along the drainage. which strongly influence its acidity and metal content Indeed, nearly 95% of the arsenic in solution is removed (Egal et al., 2010). between the source of Reigous Creek, which emerges from the mine tailings, and its confluence with the Sampling procedure and measurement of Amous River, 1.5 km downstream. To our knowledge, physicochemical properties this AMD is one of the most As-rich AMDs reported to date (Morin & Calas, 2006). It is an outstanding example Six sampling campaigns were carried out in November of adaptation to life in an extreme environment. 2007, February 2008, October 2008, March 2009, Novem- In this study, we used a combination of molecular ber 2009, and January 2010 at five sampling sites, result- approaches to investigate the spatial and temporal dynam- ing in a set of 30 samples. Groundwater was collected ics of bacterial communities in relation to the physico- from a borehole (S5, between 10 and 12 m deep) located chemical parameters in Carnoules acid mine drainage. within the tailings. Water samples were also taken at four Using 16S rRNA gene pyrosequencing and terminal sites along Reigous Creek (collecting downstream seepage restriction fragment length polymorphism (T-RFLP), our water from the surroundings) at the spring (S1), 30 m aim was (1) to characterize the spatial dynamics of the downstream from the spring (COWG), 150 m down- structure and composition of the bacterial communities stream (GAL), and 1500 m downstream (CONF), just along an environmental gradient, (2) to evaluate the tem- before the confluence between Reigous Creek and the poral changes in the composition of the communities, Amous River (Fig. 1). Water samples (300 mL) were and (3) to determine whether their dynamics could be immediately filtered through sterile 0.22 lm Nuclepore linked to variations of environmental conditions. filters, which were transferred to a collection tube (Nunc), frozen in liquid nitrogen, and stored at 80 °C until DNA extraction. This sampling was carried out in Materials and methods triplicate. Measurements of water conductivity, temperature, redox potential, pH, and dissolved oxygen concen- Description of the study site tration were carried out as previously described (Bruneel The Pb–Zn Carnoules mine, located in southern France, et al., 2011). For chemical analyses, 500 mL water sam- produced 1.2 Mt of solid wastes that are stored behind a ples were immediately filtered through 0.22 lm Millipore dam and contain 0.7% Pb, 10% Fe, and 0.2% As. The membranes fitted on Sartorius polycarbonate filter hold- aquifer is not fed by vertical percolation of rainwater ers. For total Fe and As determination, the filtered water through the tailings, but rather originates from natural was acidified to pH 1 with HNO3 (14.5 M) and stored at springs that were buried under the tailings (Koffi et al., 4 °C in polyethylene bottles until analysis. A 20 lL ali- 2003). The water table is 1–10 m below the surface of the quot of the filtered water sample was added either to a tailings stock, depending on the season and location. mixture of acetic acid and EDTA (Samanta & Clifford, With the exception of temperature, which is almost con- 2005) for As speciation or to a mixture of 0.5 mL acetate stant with average values around 15 °C, the physicochem- buffer (pH 4.5) and 1 mL of 1,10-phenanthrolinium ical parameters of the groundwater vary as a function of chloride solution (Rodier et al., 1996) for Fe(II) determi- the hydrological conditions (Casiot et al., 2003b). In nation. The vials were filled to 10 mL with deionized 2001, the groundwater below the tailings contained extre- water. Samples destined for Fe and As speciation and sul- mely high levels of arsenic: up to 10 000 mg L1 (Casiot fate determination were stored in the dark and analyzed et al., 2003b). The water emerges at the bottom of the within 24 h. Chemical analyses were carried out as previ- dam, forming the source of the Reigous Creek. This ously described (Bruneel et al., 2011). AMD is acid (pH ≤ 3), with high concentrations of sulfate (2000–7700 mg L1), iron (500–1000 mg L1), DNA isolation and arsenic (50–350 mg L1). Iron and arsenic are mainly present in their reduced forms Fe(II) and As(III) Genomic DNA was extracted in triplicate from filtered (Casiot et al., 2003a). The natural attenuation of As is the water using the UltraClean Soil DNA Isolation Kit

Fig. 1. Map of Carnoules mine and location of sampling sites.

(MoBio Laboratories Inc., Carlsbad, CA) according to ment (T-RF) profiles were obtained from the digested the manufacturer’s recommendations. These triplicate amplicons by suspending 1 lL aliquots in 8.75 lL form- extractions were pooled before PCR amplification. All amide with 0.25 lL of Genescan ROX 500 size standard genomic DNA extracts were stored at 20 °C until further (Applied Biosystems). T-RFs were separated on an ABI analysis. PRISM 3130xl Genetic Analyser (Applied Biosystems). Data were collected and analyzed using GENEMAPPER software (version 1.4, Applied Biosystem). To increase strin- Terminal restriction fragment length gency for the T-RF profiles of 16S rRNA genes, T-RFs polymorphism outside the range of the size standard (35–500 bp) were The 16S rRNA genes were amplified by PCR, and the discarded, and the background noise level was set at 30 bacterial community structure was identified by T-RFLP. fluorescence units. T-ALIGN software (Smith et al., 2005) The fluorescent labeled primers HEX 357F (50-hexa- was used to compare replicate profiles and to generate chloro-fluorescein-phosphoramidite-CCTACGGGAGGCA consensus profiles containing only T-RFs that occurred in GCAG-30) (Lane, 1991) and 926R (50-CCGTCAATTCMT replicate reactions. Consensus profiles were then aligned TTRAGT-30) (Muyzer & Ramsing, 1995), described as on the basis of the length of the T-RFs and individual peak universal within the bacterial domain, were used. Tripli- areas as previously described by Smith et al. (2005) with cate PCR amplifications were performed on each sample. the confidence interval set at 0.5, resulting in the genera- The reaction mixture contained 1 lL of DNA template, tion of data sets of aligned T-RFs that gave individual rela- 1 lL of both primers (10 lM), and 12.5 lL of PCR Mas- tive peak areas as a percentage of the overall profile. T-RFs ter Mix Ampli Taq Gold 360 (Applied Biosystems, Foster were included in the subsequent analysis if they represented City, CA). Sterile distilled water was added to obtain a > 1% of the cumulative peak height for the sample. final volume of 25 lL. PCR conditions were as follows: one cycle at 95 °C for 10 min, 35 cycles at 95 °C for Construction of the libraries, 454- 45 s, 55 °C for 45 s, and 72 °C for 45 s, followed by pyrosequencing, and sequence quality control 10 min at 72 °C. The 90 PCR products were purified with Illustra GFXTM PCR DNA and the Gel Band Purifica- The 16S rRNA genes were also amplified by PCR for tion Kit (GE Healthcare, Munich, Germany). The concen- multiplex pyrosequencing using barcoded primers. The tration of PCR product was determined by comparison primer pairs used, targeting the V3 to V5 variable regions with molecular markers (Smartlader, Eurogentec) after of the 16S rRNA gene, were 357F (50-AxxxCCTA migration on agarose gel. Approximately 100 ng of puri- CGGGAGGCAGCAG-30) and 926R (50-BxxxCCGTCAAT fied amplicon was digested in 10 lL reaction with 0.3 U TCMTTTRAGT-30). A and B represent the two FLX Tita- of enzyme HpaII or AluI (New England Biolabs Inc., Ips- nium adapters (A adapter sequence: 50-CGTATCGCCTC wich, MA) at 37 °C for 3 h. Terminal restriction frag- CCTCGCGCCATCAG-30; B adapter sequence: 50-CTAT

GCGCCTTGCCAGCCCGCTCAG-30), and xxx represent threshold) trained on the RDP taxonomic outline imple- the sample-specific barcode sequence. PCR was performed mented in Mothur and a modified bacterial database. In using 30–35 cycles under conditions identical to those silico T-RF prediction of the 16S rRNA gene sequences described above for T-RFLP. The number of cycles was obtained in this study was performed using the program varied with the samples to obtain a strong band with a TRiFLE (Junier et al., 2008), and predicted T-RFs were minimum number of cycles to respect the initial abun- linked to measured T-RFs from the microbial commu- dances of bacterial communities. The 90 PCR products nity profiles. with a proximal length of 569 bp were excised from 1% agarose gel and purified with the QIAquick Gel Extrac- Estimation of diversity and statistical analysis tion Kit (QIAGEN Inc., Valencia, CA). To minimize random PCR bias, triplicates were pooled in equimolar Diversity indices ratios prior to pyrosequencing. Pyrosequencing of the 30 amplicon libraries was performed on a GS-FLX-Titanium Nonparametric Chao1 and Shannon alpha diversity esti- sequencer (Roche 454 Life Sciences) at the GenoToul mates, as well as coverage and rarefaction curves, were genomic platform in Toulouse (France) using four sepa- calculated with MOTHUR v.1.30 for each sample. Analysis rate 1/8 region of a plate. of variance (ANOVA) was performed with Tukey’s tests to identify differences between sampling sites. Processing of pyrosequencing data and taxonomic classification Cluster analysis Preliminary quality checks, sorting, and trimming of the To compare community composition based on T-RFLP 454 reads were carried out using the NG6 pipeline and 454-pyrosequencing data, normalized OTUs abun- (http://vm-bioinfo.toulouse.inra.fr/ng6/). Tags were dances were square-root-transformed and pairwise dis- extracted from the 454 reads using the sff file (Roche similarities among samples were calculated using the software), and three kinds of analysis were performed as relative abundance-based Bray–Curtis index (BC). Non- described by Ueno et al. (2010): (1) BLAST search for metric multidimensional scaling (nMDS) analysis was E. coli, phage, and yeast contaminants, (2) read quality performed on the dissimilarity matrices to visualize pat- analysis, and (3) removal of sequences that were too long terns of community composition. Using the 454-pyrose- or too short (sequences with more or less than two stan- quencing data, we carried out a random sampling dard deviations from the mean), sequences containing procedure to make equal the number of sequences per more than 4% of N, low-complexity sequences and sample (486 sequences) and we removed singleton OTUs duplicated reads, using Pyrocleaner. The sequences were (sequences that only occurred once) to reduce the influ- then analyzed with the software Mothur version 1.30 ence of rare OTUs. One-way analysis of similarity (Schloss et al., 2009). Preprocessing of unaligned (ANOSIM) and multiple pairwise comparisons were used to sequences included removing sequences < 450 bp, all test whether there were significant differences in commu- sequences containing ambiguous characters, and nity composition in space. R-values > 0.75 are commonly sequences with more than eight homopolymers. We also interpreted as well-separated bacterial compositions; removed sequences that did not align over the same span R > 0.5 as overlapping, but clearly different; and R < 0.25 of nucleotide positions. Identical sequences were as practically not separable. grouped, and representative sequences were aligned against the SILVA bacterial and archaeal reference data- CCA base using the Needleman–Wunsch algorithm (Needle- man & Wunsch, 1970). Chimeric sequences were Canonical correspondence analyses (CCAs) were used to detected and removed using the implementation of Chi- explore variations in the bacterial communities under the mera Uchime. A further screening step (precluster) was constraint of our set of environmental variables. Explana- carried out to reduce sequencing noise by clustering tory variables were log(x + 1)-transformed where neces- reads differing by only one base every 100 bases (Huse sary to approximate normal distribution. This model was et al., 2010). The remaining high-quality reads were used tested with Monte Carlo permutation tests (499 random- to generate a distance matrix and were clustered into ized runs) to determine significance, and each environ- operational taxonomic units (OTUs) defined at 97% cut- mental parameter was tested by stepwise analysis to off using the average neighbor algorithm. Next, the detect significant predictors. All statistical analyses were OTUs were phylogenetically classified to genus level performed with R 3.0.1 (R Development Core Team, using the naive Bayesian classifier (80% confidence 2012) including the VEGAN package.

Results 96% of iron, and 99% of arsenic were removed from the aqueous phase between S1 and CONF sampling sites. Spatial and temporal analyses of the environmental data set Diversity and species richness estimators of bacterial communities The physicochemical characteristics of the water samples were determined for six sampling dates and at five differ- Hex-labeled PCR products were digested separately with ent sites in a borehole and along Reigous Creek (Fig. 1; two restriction enzymes. HpaII that produced the largest Supporting Information, Table S1). With the exception of numbers of T-RFs (data not shown) was used to assess pH, the environmental variables measured differed signifi- the differences in the microbial communities. T-RFLP cantly between sites (ANOVA, P < 0.05). A significant profiles generated showed a total of 43 different T-RFs decrease (Tukey’s test, P < 0.01) in temperature was for the five sites, and the number of T-RFs detected in observed in Reigous Creek, where the influence of air each sample varied from 2 to 17 (Fig. 3). Average T-RF temperature causes a larger range of values (Fig. 2a) in richness (number of T-RFs) and average Shannon diver- contrast to the temperature of the water in the borehole sity indices calculated from relative peak intensity data (S5) and at the source of the Reigous (S1), which did not were highest at S1 and COWG (H = 2.03 0.3 and differ significantly (13.3 1.4 °C and 13.6 1 °C, H = 1.96 0.3, respectively), while the lowest values respectively, and Tukey’s test, P = 0.99). All samples were were observed at GAL (H = 1.26 0.4; Fig. 4a). Values characterized by low pH (≤ 3.7). No significant variations at CONF were intermediate (H = 1.67 0.7). Although in pH (ANOVA, P = 0.49) were observed between the sam- bacterial community diversity varied among the sampling pling sites located downstream from the source (COWG, sites, the differences were not significant (ANOVA, GAL and CONF; Fig. 2b). Dissolved oxygen (DO) con- F = 2.46, P = 0.071). For each site, the bacterial commu- centrations in the water at the upstream sites presented a nity showed variations over time, but no particular trend mean of 1.0 1.0 mg L1 for S5 and 0.8 0.6 mg L1 could be identified. Some T-RFs were found in the for S1, denoting generally suboxic conditions at these majority of the profiles (e.g. T-RF 150), where they usu- sites. DO increased sharply between S1 and COWG ally accounted for a high proportion of the total T-RFs (mean of 6.1 1.8 mg L1) (Fig. 2c) and continued to (Fig. 3). Between one and three site-specific T-RFs were increase slightly all along the creek, to reach a mean of identified in all the sites (in red in Fig. 3). 10.5 1.3 mg L1 at CONF. The redox potential (Eh) A total of 99 441 sequence reads were generated in a showed average values of 558 85 mV at S5. Eh single run of 454-pyrosequencing from 30 independent increased along Reigous Creek from 512 39 mV at S1 16S rRNA gene libraries. Note that pyrosequencing of to 635 99 mV at CONF (Fig. 2d). In contrast, average two samples taken in February 2008 (S5 and S1) failed conductivity decreased from 7588 5203 lScm1 at S5 and were thus excluded from analysis. After trimming to 5121 631 lScm1 at S1 and reached minimum and processing with Mothur, 63 442 reads remained with 1 2 (1612 97 lScm ) at CONF (Fig. 2e). Sulfate (SO4 ) an average length of 530 bp. Clustering of the remaining concentrations were maximum in the groundwater at S5 sequences led to the identification of 6613 OTUs (includ- with a mean of 14 080 10 630 mg L1. After a sharp ing 4510 singletons) defined at 97% identity. Although decrease at S1 (average values of 2682 1180 mg L1), singletons represented 68% of the total number of OTUs, concentrations gradually decreased along Reigous Creek they only accounted for 7% of the total DNA sequences. (Fig. 2f). Dissolved Fe concentrations in the groundwater The results of rarefaction analysis along with the Chao1 at S5 exhibited average values of 4474 2855 mg L1. and the Shannon indexes and coverage values are listed Fe concentrations decreased from the source (S1, average in Table 1. In the resampled data set, Good’s coverage values of 1317 383 mg L1) to CONF, where Fe ranged from 69% to 97% with an average value of 85%, remained below 82 mg L1 (Fig. 2g). The proportion of indicating that the majority of bacterial phylotypes were Fe(III) (difference between Fe(total) and Fe(II)) was gen- recovered. Species richness (Chao1 index) of the bacterial erally negligible except at some sampling dates at S5 and communities presented significant variations along Rei- CONF (Table S1). At S5, concentrations of dissolved gous Creek (ANOVA, P = 0.001, F = 6.66) (Fig. 4b). The arsenic (As) exhibited an average value of 440 nonparametric estimators Chao1 ranged between 52 and 184 mg L1 (Fig. 2h). Along Reigous Creek, As concen- 495 estimated OTUs for all the sites considered (Table 1). trations decreased with increasing distance from the The highest average OTU richness was found at CONF source (average value of 175 71 mg L1), to values and S1 (Chao1 = 364 145 and 296 32, respectively), below 6 mg L1 at CONF (Fig. 2h), with predominance suggesting that an important number of OTUs were not of the reduced form As(III). An average of 65% of sulfate, revealed by the analysis of these two sites. Situated

(a)

(b)

(c)

(d)

Fig. 2. Variations in the main physicochemical parameters over the course of the study and boxplot of each variable per sampling site. Arrows indicate sampling dates for T-RFLP and pyrosequencing analysis. Note that some data are missing, as shown by the gaps in the curves. DO, dissolved oxygen; Eh, redox potential.

(e)

(f)

(g)

(h)

Fig. 2. Continued

Fig. 3. Relative abundance of terminal restriction fragments (T-RFs) derived from bacterial communities. Single T-RFs per sampling site are in red. Dominant T-RFs are in bold. Taxonomic afﬁliation of T-RFs was carried out by in silico T-RFLP analysis. T-RF 91 and 100 could not be assigned to any phylogenetic group; T-RF 125 represented Armatimonadetes gp4 and Chlorobi; T-RF 150 was mainly related to Deinococcus-Thermus, Spirochaetes and Actinobacteria; T-RF 162 was mainly related to A. ferrooxidans but could be assigned to other proteobacterial phylotypes detected in the AMD. N7: November 2007; O8: October 2008; M9: March 2009; N9: November 2009; J10: January 2010.

(a) (b)

Fig. 4. Average diversity and species richness index per group standard deviation calculated based on (a) T-RFLP proﬁles and (b) 454 pyrosequencing reads of the reduced data set based on the 16S rRNA genes.

ª 2014 Federation of European Microbiological Societies. FEMS Microbiol Ecol && (2014) 1–17 Published by John Wiley & Sons Ltd. All rights reserved ESMcoilEcol Microbiol FEMS Table 1. Estimated OTU richness, diversity indices, and estimated sample coverage for each 16S rRNA gene library. Results are presented for full data set reads (full data set) and for reduced communities bacterial of dynamics Spatiotemporal data sets without singletons and randomly resampled to make the sample size equal (reduced data set)

Reduced data set Full data set

Sampling No. of Obs. No. of Obs. * † ‡ * † ‡ && sites reads OTUs Chao1 Shannon Coverage reads OTUs Chao1 Shannon Coverage

21)1–17 (2014) S5 S5N7 486 79 116 (95; 163) 3.18 (3.04; 3.32) 92 2089 216 576 (435; 810) 3.16 (3.07; 3.24) 93 S5O8 486 32 52 (38; 98) 1.66 (1.52; 1.80) 97 2523 87 258 (165; 465) 1.49 (1.42; 1.56) 98 S5M9 486 52 89 (67; 146) 2.14 (1.99; 2.29) 94 2838 177 381 (298; 520) 2.20 (2.13; 2.28) 96 S5N9 486 92 142 (115; 199) 3.31 (3.17; 3.46) 91 1086 159 481 (334; 751) 3.22 (3.11; 3.34) 91 S5J10 486 70 102 (82; 150) 3.40 (3.29; 3.51) 94 2354 195 598 (423; 908) 3.41 (3.35; 3.48) 95 S1 S1N7 486 138 266 (209; 368) 3.86 (3.71; 4.01) 83 2719 436 1329 (1051; 1773) 3.88 (3.79; 3.96) 90 S1O8 486 131 289 (217; 422) 3.93 (3.80; 4.05) 83 3422 727 2396 (1994; 2925) 4.59 (4.52; 4.66) 85 S1M9 486 127 270 (204; 392) 3.82 (3.68; 3.95) 84 2057 365 1068 (844; 1397) 3.97 (3.89; 4.06) 88 S1N9 486 137 315 (237; 455) 3.48 (3.30; 3.65) 81 2021 392 1040 (848; 1313) 3.53 (3.41; 3.64) 86 S1J10 486 181 342 (277; 449) 4.48 (4.35; 4.61) 78 4573 845 2158 (1859; 2544) 4.81 (4.75; 4.88) 88 COWG CGN7 486 121 193 (159; 256) 3.75 (3.61; 3.89) 87 688 194 367 (301; 474) 4.06 (3.93; 4.19) 82 CGF8 486 105 201 (154; 295) 3.16 (2.99; 3.33) 87 2160 305 835 (659; 1099) 3.21 (3.11; 3.30) 90 CGO8 486 98 206 (152; 312) 2.90 (2.72; 3.07) 87 2119 255 1025 (715; 1544) 2.82 (2.72; 2.92) 92 CGM9 486 146 382 (279; 565) 3.87 (3.72; 4.01) 79 2756 511 1710 (1364; 2196) 4.11 (4.02; 4.19) 87 CGN9 486 112 349 (232; 578) 3.00 (2.81; 3.18) 84 1317 201 624 (455; 907) 2.80 (2.67; 2.93) 89 CGJ10 486 50 88 (65; 148) 1.44 (1.26, 1.62) 94 1827 136 472 (311; 781) 1.51 (1.41; 1.62) 95 GAL GLN7 486 91 144 (117; 200) 2.28 (2.08; 2.49) 90 1638 223 543 (423; 734) 2.23 (2.10; 2.35) 91 GLF8 486 65 224 (135; 430) 1.33 (1.14; 1.52) 90 1679 177 665 (456; 1030) 1.56 (1.44; 1.68) 92 GLO8 486 93 253 (172; 418) 2.57 (2.39; 2.75) 87 2093 236 614 (473; 839) 2.64 (2.54; 2.74) 93

ª GLM9 486 107 239 (178; 354) 2.36 (2.14; 2.57) 85 2489 356 1070 (843; 1403) 2.49 (2.37; 2.60) 90 04Fdrto fErpa irbooia Societies. Microbiological European of Federation 2014 ulse yJh ie osLd l ihsreserved rights All Ltd. Sons & Wiley John by Published GLN9 486 110 227 (171; 334) 2.60 (2.39; 2.82) 86 1682 285 715 (570; 934) 2.65 (2.51; 2.78) 88 GLJ10 486 83 201 (140; 331) 2.05 (1.85; 2.25) 88 2246 271 862 (654; 1183) 2.12 (2.01; 2.23) 91 CONF CFN7 486 239 492 (400; 635) 5.07 (4.97; 5.18) 70 3256 933 1995 (1764; 2290) 5.92 (5.87; 5.98) 84 CFF8 486 62 186 (115; 352) 1.46 (1.27; 1.65) 91 1768 240 972 (692; 1428) 1.94 (1.82; 2.06) 89 CFO8 486 191 335 (279; 426) 4.40 (4.25; 4.55) 77 1731 548 1163 (1004; 1377) 5.02 (4.93; 5.12) 80 CFM9 486 84 202 (142; 323) 1.87 (1.66; 2.07) 88 527 125 462 (306; 751) 2.28 (2.07; 2.50) 81 CFN9 486 233 477 (390; 612) 4.82 (4.69; 4.95) 69 1994 699 1708 (1467; 2026) 5.45 (5.37; 5.53) 77 CFJ10 486 180 495 (367; 711) 4.06 (3.89; 4.23) 74 5718 1343 3763 (3330; 4289) 4.79 (4.71; 4.86) 84 *OTUs were defined at 97% cutoff. † Takes into account the number and evenness of species. ‡ Coverage: sum of probabilities of observed classes calculated as (1 (n/N)), where n is the number of singleton sequences and N is the total number of sequences. Values in brackets are 95% confidence intervals. 9 10 A. Volant et al. between these two sites, COWG and GAL exhibited inter- sample analyzed (Fig. S1b–d). Most of the sequences mediate richness estimates (Chao1 = 236 110 and associated with Acidobacteria could not be classified to 215 39, respectively). The lowest richness was observed the order level except for Acidobacteriales and Holopha- in the tailing groundwater at S5 (Chao1 = 100 33). gales. As can be seen in Fig. S2, a relatively small number Bacterial OTU diversity, estimated by the Shannon index, of OTUs dominated at all sites (> 1% in total abundance also differed significantly between sites (ANOVA, F = 3.01, per sample). The most abundant OTUs were phylogeneti- P = 0.039), with values ranging from 1.33 to 5.07 cally related to Gallionella ferruginea (Gallionellales), (Table 1). In agreement with T-RFLP data analysis Acidithiobacillus ferrooxidans (Acidithiobacillales), and (Fig. 4a), the highest average diversity value was found at Thiobacillus sp. (Hydrogenophilales), collectively account- S1 (H = 3.91 0.36) and the lowest value at GAL ing for 41% of all the sequences. (H = 2.20 0.47) (Fig. 4b). As predictable, average T-RF When possible, T-RFs were assigned to a taxon or a diversity is lower than OTU diversity (c. 50%); indeed, group of taxon by in silico restriction of 16S rRNA gene taxon-specific resolution of pyrosequencing is much sequences. Among the five dominant T-RFs (91, 100, 125, higher than fingerprinting (Pilloni et al., 2012). Again, no 150, and 162 bp in size), T-RFs 91 and 100 could not be seasonal trend was observed. The same richness and assigned to any specific phylogenetic group. Armatimona- diversity patterns were observed in both the full and detes gp4 and Chlorobi were represented by T-RF 125, resampled data sets, although the richness estimator and T-RF 150 was mainly related to Deinococcus-Thermus, and Shannon index were higher in the full data set, due Spirochaetes, and Actinobacteria. While T-RF 162 was to the larger number of sequences (data not shown). mainly related to A. ferrooxidans, it could be assigned to other proteobacterial phylotype detected in this AMD. Taxonomic assignment of bacterial pyrosequencing reads and T-RFs Spatial and temporal variations in bacterial community structure At a confidence threshold of 80%, we were able to assign 56 426 of 63 442 qualified reads (that is, 89%) to a Spatiotemporal dynamics of bacterial populations were known phylum (Table S2) and 76% to a known order identified by T-RFLP analysis and 454-pyrosequencing of (Supporting Information, Fig. S1). Most of the unclassi- 16S rRNA genes (Fig. S3). fied reads (55% representing 9.6% to 37% of the qualified Although samples formed overlapping clusters on the reads of each sample) were associated with samples col- nMDS plot of the T-RFLP profiles, weak but significantly lected at S1. Altogether, 23 bacterial phyla were recovered different bacterial communities at the five sites were from our samples, with 4–8 different phyla found in sam- revealed (ANOSIM Global R = 0.2819, P < 0.001). S5 dif- ples collected at S5, 12–13 at S1, 9–14 at COWG, 10–13 fered significantly from the other sites, with some over- at GAL, and 9–20 at CONF (Table S2). Most of the bac- lapping communities (pairwise tests: r-values ranging terial sequences (86%) belonged to phyla that are most from 0.45 to 0.61, P < 0.05). These results highlighted often encountered in acid mine drainages worldwide changes in the structure of the bacterial communities (Proteobacteria, Actinobacteria, Firmicutes, Acidobacteria, between the tailing groundwater (S5) and the water in Bacteroidetes, and Nitrospirae). In addition, microorgan- Reigous Creek. The high dissimilarity observed within isms representing 0.5% of the total sequences were related each site revealed variations in community structure over to CARN1, ‘Candidatus Fodinabacter communificans’. time, especially at S5, GAL, and CONF (Fig. S3a). These Proteobacteria was the most abundant phylum in all the variations may have masked a spatial pattern. samples, accounting for 69.6% of all sequences retrieved. nMDS analyses of 454-pyrosequencing data also This phylum was represented by bacteria belonging to the showed that the composition of the bacterial communi- Alphaproteobacteria, Betaproteobacteria, Gammaproteobac- ties differed significantly along the spatial gradient from teria, Deltaproteobacteria, and Epsilonproteobacteria. The the sterile (S5) to the confluence (CONF) (Fig. S3b). Fur- most abundant classes in nearly all the samples were thermore, an ANOSIM test corroborated the nMDS plot Betaproteobacteria and Gammaproteobacteria (average val- data, revealing significantly different bacterial composi- ues of 63.4% and 30.4% of the pyrosequencing reads, tions in water as a function of the spatial location (Global respectively). There were dominated by Gallionellales and R = 0.6192, P < 0.001), except at GAL and COWG which Acidithiobacillales, respectively (Fig. S1a). Three other did not differ significantly (ANOSIM pairwise comparison phyla, Actinobacteria represented mainly by Acidimicrobi- r = 0.206, P = 0.37). Higher temporal variation at CONF ales and Actinomycetales, Firmicutes (principally Clostridi- was highlighted by the large cluster on the nMDS plot. ales and Bacilliales), and Acidobacteria, were also The marked temporal variations in the bacterial commu- abundant but their proportion varied depending on the nity at S5 and CONF highlighted by the two data sets

ª 2014 Federation of European Microbiological Societies. FEMS Microbiol Ecol && (2014) 1–17 Published by John Wiley & Sons Ltd. All rights reserved Spatiotemporal dynamics of bacterial communities 11 may be linked to the stronger seasonal fluctuation of S5 and S1 than in downstream samples. T-RF 150 (Dei- some physicochemical parameters at these sites, particu- nococcus-Thermus, Spirochaetes, and Actinobacteria – larly temperature at CONF (Fig. 2a) and pH, Eh, sulfate, related) was the most abundant phylotype in the tailings Fe, and As at S5 (Fig. 2b, d, f, g and h). groundwater (S5), accounting for up to 70% of the T- We investigated the four most abundant phyla to get a RFLP profiles. It was relatively abundant along the creek global view of the variations of bacterial communities especially at GAL. T-RF 162 (related to A. ferrooxidans along the creek (Fig. S1). Proteobacteria distribution var- but also to other proteobacterial phylotype) was not ied between samples, with a relative predominance of dominant at S5 and represented a minor fraction of the Acidithiobacillales in samples from S5 and S1, followed by bacterial community along the creek. the dominance of Gallionellales in the majority of other samples (Fig. S1a). A general increase in Betaproteobacte- Linking bacterial community structure to ria was observed in the downstream direction of Reigous environmental variables Creek. Among Actinobacteria, the Acidimicrobiales were present in all samples except those collected at CONF Canonical correspondence analysis (CCA) was performed where almost equal proportions of Acidimicrobiales and to elucidate the main relationships between physicochemi- Actinomycetales were retrieved (Fig. S1b). The Firmicutes cal variables and bacterial community structure and com- phylum revealed the dominance of Clostridiales in sam- position (Fig. 6). Samples were plotted in different areas ples from S5, whereas Bacillales were dominant at the of the diagram depending on their environmental charac- other sites, again except CONF. Different orders were teristics. The resolution of 454-pyrosequencing allowed to dominant at CONF over time, including Lactobacillales account for more variation than T-RFLP (36.4% and and Selenomonadales (Fig. S1c). No Acidobacteria were 20.5%, respectively) in the species–environment relation- retrieved at S5 in October 2008 (Fig. S1d). ship across the first two canonical axes. CCA axis 1 based We also assessed the dynamics of the dominant genera on T-RFLP data only separated the samples into two clus- (> 5% in total abundance per sample) (Fig. 5). The rela- ters, one containing the tailings site (S5) and the other tive abundance of genera at each site varied over the sam- grouping the sites along the creek (S1, COWG, GAL, and pling period. While the relative abundance of Gallionella CONF), with sulfate, DO, and pH being the strongest was almost constant in COWG and GAL, there was an determinants of bacterial community structure (Fig. 6a). important temporal variation in the other sites. This was In contrast, a higher resolution was observed with CCA evident in S5 where this genus was extinct and re-thrived axis 1 based on 454-pyrosequencing data, which was most over time (Fig. 5a). Although Gallionella was present in closely correlated with iron, arsenic, and conductivity and almost all sites for a sampling date, there was no clear rela- separated the sampling sites into three clusters (CONF, tionship between sites. Indeed, GAL and COWG exhibited GAL+COWG, and S1+S5) as a function of the pollution a relatively high proportion of Gallionella at almost all the gradient (Fig. 6b). The upstream site S5 was highly pol- sampling dates (as much as 85% of all pyrosequencing luted and little oxygenated, whereas the downstream site reads at GAL) without any link with the upstream sites S5 CONF was less polluted and characterized by a higher and S1 (Fig. 5b). In contrast, Acidithiobacillus represented redox potential (Eh). CCA axis 2 separated samples a minor fraction of the bacterial community, except at S5 according to water temperature. After the Monte Carlo where this OTU was dominant (24–72% of the pyrose- permutation test, the environmental variables significantly quencing reads). The relative abundance of Acidithiobacil- correlated with the canonical axes based on 454-pyrose- lus showed a decreasing trend along the continuum for quencing data were arsenic (F-ratio = 1.9, P = 0.01), tem- each sampling date (Fig. 5b). This genus also exhibited an perature (F-ratio = 1.4, P = 0.01), and sulfate (F- increase in October 2008 and March 2009 for S5 and S1 ratio = 1.4, P = 0.01). The differences between the two (Fig. 5a). Members of the Thiobacillus genus showed data sets were probably due to the power of 454-pyrose- higher proportion in COWG at each sampling date inde- quencing over T-RFLP for taxon resolution. Focusing on pendently of the other sites (Fig. 5b). The temporal varia- 454-pyrosequencing data, the influence of environmental tion of this genus was minor except an important increase variables on dominant OTUs (> 5% of total abundance at GAL in October 2008 (Fig. 5a). per sample) was also investigated (Fig. S4). Nineteen of T-RFLP profiles from the 30 samples were investigated the 23 dominant OTUs showed a strong correlation with to assess the dynamics of T-RFs (Fig. 3). Among the five the physicochemical parameters. Five OTUs (15, 16, 28, dominant T-RFs, T-RFs 91 and 100 represented a large 32, and 52) were strongly correlated with elevated DO and proportion of the T-RFLP profiles in almost all samples Eh and 14 with high temperatures and high concentrations except those from S5. T-RF 125 related to Armatimonade- of As, Fe, and sulfate. As indicated by the position of tes gp4 and Chlorobi was more abundant in samples from OTUs 1, 3, 4, and 11 on the graph, near the origin of the

(a)

(b)

Fig. 5. Relative abundance of the dominant genera (> 5% in total abundance per sample) presented by (a) sampling sites and (b) sampling date. N7: November 2007; O8: October 2008; M9: March 2009; N9: November 2009; J10: January 2010.

axes, none of the environmental variables measured in the study could explain their distribution and thus their niche. Discussion At the least polluted site (CONF), Gallionella, Ferrovum, This study combined a classical ﬁngerprinting method and Acidiferrobacter were the main genera detected, (T-RFLP) and a high-throughput barcoded pyrosequenc- whereas, at the most polluted sites (S5 and S1), a higher ing of 16S rRNA genes to investigate the diversity, spatial number of genera were codominant (Acidithiobacillus, distribution, and seasonal variation of bacterial communi- Ignavibacterium, Ralstonia, Leptospirillum, Gallionella, ties in Carnoules AMD (France), which is heavily con- Ferrovum, etc.). taminated with As.

(a) (b)

Fig. 6. Canonical correspondence analysis (CCA) correlating the bacterial community structure at each sampling site with arsenic (As), iron (Fe), conductivity (Cond), temperature (T), dissolved oxygen (DO), redox (Eh), pH and sulfate. The bacterial community structures correspond to OTU abundances from (a) T-RFLP data and (b) pyrosequencing data. The main clusters are highlighted.

phylotypes present. Nonetheless, additional sequencing Spatial and temporal variations in the effort would be required to exhaustively characterize the environmental data set and in the bacterial bacterial community, particularly for samples from the community least polluted site CONF, as shown by the lower coverage Monitoring the physicochemical parameters of Reigous values and the lack of asymptote in the rarefaction curves Creek confirmed previous results (Casiot et al., 2003a; (data not shown). Egal et al., 2010), showing a significant decrease in nMDS analyses revealed significant differences in the concentrations of dissolved As, Fe, and sulfate with composition of the bacterial communities in the five sites increasing distance from the source: 72% of sulfate, 96% along the AMD (Fig. S3). However, different clustering of iron, and 99% of arsenic had been removed by the patterns were obtained based on T-RFLP or pyrosequenc- time Reigous Creek flowed into the River Amous (Table ing data. With pyrosequencing, individual sequences can S1). In addition, the concentrations of As and Fe in the be classified at the genus level. In contrast, one T-RF can water from the tailings stock were much lower in 2007– correspond to several different bacterial phylotypes 2010 (average values of 440 184 mg L1 and 4474 (belonging to different genera or even different higher 2855 mg L1, respectively), than those measured in 2001 taxonomic levels). Such differences in the resolution of (up to 10 g L1 for As and around 20 g L1 for Fe, Casi- the two methods may explain the differences obtained in ot et al., 2003b), although these concentrations are still the cluster analyses (Hwang et al., 2012). Nevertheless, very high compared to other AMDs. changes in bacterial community structure between the Both molecular methods highlighted a higher bacterial tailings groundwater (S5) and the Reigous Creek were diversity than expected in this extreme habitat. T-RFLP revealed by the two sets of data, reflecting important dif- profiles showed for the five sites a total of 43 T-RFs rang- ferences in ecological conditions between the two habi- ing from 2 to 17 T-RFs per sampling site (Fig. 3). For py- tats. According to both methods, S1 was the most diverse rosequencing data, a total of 63 442 reads led to the bacterial community, while GAL was the least diverse. identification of 6613 OTUs, including 4510 singletons Therefore, bacterial diversity varied independently of the representing 68% of the total number of OTUs. As sampling site, suggesting that globally upstream commu- expected, a larger number of phylotypes were identified nities do not influence downstream communities. The using the pyrosequencing method leading to a significant apparent minimum effect of immigration suggests that increase in resolution. Average Good’s coverage was over species-sorting processes best describe bacterial commu- 89%, suggesting that the 16S rRNA gene sequences into nity structure in these connected environments, with local each sample represented the majority of the bacterial environmental factors driving the composition of the

FEMS Microbiol Ecol && (2014) 1–17 ª 2014 Federation of European Microbiological Societies. Published by John Wiley & Sons Ltd. All rights reserved 14 A. Volant et al. bacterial community at each site. Temporal variations of Gallionella-like organisms may be more tolerant to acid the bacterial communities could also be observed at each and metal than currently thought (Fabisch et al., 2013). site although no particular trend could be identified. In accordance with our results, temperature has been pre- However, the important temporal variation of the bacte- viously suggested as a primary factor controlling the rial community observed at S5 and CONF may be due to structure and dynamics of microbial communities in a higher seasonal fluctuation in physicochemical parame- AMD (Edwards et al., 1999) and in various natural envi- ters at these two sites, especially temperature at CONF, ronments like hot springs (Ward et al., 1998; Miller et al., and pH, Eh, sulfate, Fe, and As at S5. 2009) or marine environments (Fuhrman et al., 2008). However, our fine-scale investigation at the genus level Nevertheless, sulfate and arsenic concentrations have not of the bacterial communities along the Reigous Creek previously been shown to be significantly correlated with over time provided some important data and allowed to bacterial diversity in AMDs. Earlier studies identified dif- establish some hypothesis about community composition. ferent environmental predictors of microbial populations Indeed, Gallionella in contrast to Acidithiobacillus do not in AMD including conductivity and rainfall (Edwards seem to benefit from the seed bank provided by the most et al., 1999), pH (Lear et al., 2009), oxygen gradient upstream sites (S5 and S1). This suggests that Gallionella, (Gonzalez-Toril et al., 2011), and season (Streten-Joyce under a process that still need to be elucidated, extinct/ et al., 2013), which may result in site-specific physico- re-thrived at each site over time. In contrast, Acidithioba- chemical and geochemical characteristics (Kuang et al., cillus that is preferentially encountered upstream of the 2012). Furthermore, while many studies highlighted pH Reigous Creek or Thiobacillus that thrived at COWG as the most important factor structuring AMD communi- could be found at these sites, under conditions that ties (Kuang et al., 2012; Chen et al., 2013), our study reflect their preferential habitats. The presence of these produced no evidence of the influence of this parameter, organisms downstream of the sites would instead reflect probably due to the limited variation in pH among our dispersal from upstream sites. samples (average values of 2.5 0.8–3.2 0.3).

Physicochemical parameters shape the Composition of the bacterial communities composition of the bacterial community In this study, we were able to identify a wider phyloge- This work highlighted a spatial gradient of physicochemi- netic range of taxa than in any previous clone library- cal conditions linked to a significant shift in bacterial based diversity survey of the Carnoules AMD, including community composition along the continuum. Indeed, sequences of several previously undetected taxa. These canonical correspondence analysis of the whole pyrose- new taxa include members of the Bacteroidetes, Chlorobi, quencing data set indicated that arsenic, temperature, and Chloroflexi, Elusimicrobia, Chlamydiae, Cyanobacteria, Dei- sulfate were the factors that most influence the composi- nococcus-Thermus, Spirochaetes, Fibrobacteres, Fusobacteria, tion of the bacterial communities (Fig. 6b). The level of Gemmatimonadetes, Plantctomycetes, Verrumicrobia, and pollution affects also some dominant bacterial popula- of the uncultured OD1-PO11-TM7 clade. The majority of tions (> 5% of relative abundance). Gallionella, Ferrovum, phyla that were not previously detected on clone libraries and Acidiferrobacter were the dominant genera in water accounted only for < 1% of the pyrosequencing data, sampled at the least polluted site (CONF) and were cor- explaining why they were missed with the clone library related with high DO and Eh, whereas in water from the approach. The high rate of low-abundance populations most polluted sites (S5 and S1), a larger number of dom- (68% of singletons) increased the phylogenetic bacterial inant genera were detected (Acidithiobacillus, Ignavibacte- diversity. However, despite the preponderance of this rare rium, Ralstonia, Leptospirillum, Gallionella, and Ferrovum) biosphere in most studies, its ecological and functional whose relative abundance was correlated with higher tem- roles remain largely unexplored (Galand et al., 2009). perature and high concentrations of As, Fe, and sulfate Recent studies indicated that such organisms may be at a (Fig. S4). Thus, different members of a given genus such dormant or a spore stage, but in favorable conditions as Gallionella (OTUs 19, 15 and 1) or Ferrovum (OTUs they may become active and even dominant (Delavat 25 and 28) were correlated with different environmental et al., 2012). Thus, these taxa may play an important role parameters, suggesting that these OTUs correspond to in extreme habitats like AMD, buffering the effects of bacterial phylotypes with some specificity explaining these important environmental shifts (Sogin et al., 2006; Mon- different behaviors. Furthermore, the high abundance of chy et al., 2011). However, further investigations will be Gallionella-related sequences in these acidic ecosystems needed to determine whether they play a role in this eco- characterized by contrasted levels of pollution is consis- system and/or whether they reflect allochthonous input tent with results of a previous study suggesting that from surrounding environments. Moreover, the high-

ª 2014 Federation of European Microbiological Societies. FEMS Microbiol Ecol && (2014) 1–17 Published by John Wiley & Sons Ltd. All rights reserved Spatiotemporal dynamics of bacterial communities 15 throughput sequencing also questions the accuracy of oxidation of As (Bruneel et al. 2003). A metaproteomic OTU richness estimates, as sequencing errors and inade- approach also showed that Gallionella, Thiomonas, and A. quate clustering algorithms can lead to overestimates of ferrooxidans actively express proteins in situ, thus proba- community richness (Huse et al., 2010). The majority of bly playing a functional role in this AMD (Bruneel et al., the most abundant taxa detected in this study were 2011). These populations could play an important role in related to orders commonly encountered in AMD, most the efficient remediation process observed along this creek of which are known to be involved in Fe, As, and S by favoring the oxidation of Fe(II) and the co-precipita- cycles: namely Gallionellales (Betaproteobacteria), Acidi- tion of As (Casiot et al., 2003a; Bruneel et al., 2011). thiobacilliales (Gammaproteobacteria), Acidimicrobiales This work has increased our knowledge of bacterial (Actinobacteria), Hydrogenophilales (Betaproteobacteria), diversity and dynamics in acid mine drainage. Bacterial Burkholderiales (Betaproteobacteria), Nitrospiralles (Nitros- diversity in Carnoules AMD was revealed to be much pirae), Desulforomonadales (Deltaproteobacteria), and Des- higher than previously evidenced using clone library ulfobacterales (Deltaproteobacteria). The ecological role of techniques (Bruneel et al., 2011), as suggested by cul- previously detected taxa has been widely characterized ture-dependent methods (Delavat et al., 2012). Our study (Bruneel et al., 2005, 2006, 2011; Bertin et al., 2011) in revealed complex patterns of spatial and temporal varia- this ecosystem. A relatively small number of OTUs domi- tions in bacterial community composition, suggesting nated at each sampling site (Fig. S2) and the majority of that community composition reflects changes in physico- them were phylogenetically related to taxa previously chemical conditions. This investigation provided a first found in AMD (Gallionella ferruginea, Acidithiobacillus step to the study of spatial and temporal structure of ferrooxidans, and Thiobacillus sp.), as well in Carnoules bacterial communities and the factors that control it. To revealing their persistence in such ecosystems (Baker & improve our understanding of the functioning of this Banfield, 2003; Bruneel et al., 2006, 2011; Hallberg et al., ecosystem, future efforts should be oriented toward 2006; Heinzel et al., 2009; Hallberg, 2010). These three active communities and how they fluctuate in response genera varied in their relative abundance over the sam- to environmental changes. Such knowledge will help to pling period. Gallionella was present in high proportions determine their roles in the functioning of the AMD in almost all samples, mainly at GAL and COWG. In ecosystems and explain important assembly processes in contrast, except at S5 where this genus was dominant, microbial ecology. Acidithiobacillus accounted for a minor fraction of the bacterial community (Fig. 5). Furthermore, our study Acknowledgements confirmed the presence of relatives of a novel bacterial phylum, ‘Candidatus Fodinabacter communificans’ This study was financed by the FRB (Fondation pour la detected by a recent metagenomic investigation of Car- recherche sur la Biodiversite) program blanc AAP-IN-2009- « noules AMD and prominent in the active COWG com- 039, the Observatoire de Recherche Mediterraneen de l’ munity (Bertin et al., 2011; Fahy et al., unpublished Environnement » (OSU-OREME). A.V. was supported by data). The relatively high number of unclassified bacteria a grant from the French Ministry of Education and per sample (0.3–37%) supports the fact that many bacte- Research and F.J. by a grant from the Direction Generale ria remain to be cultured. These results thus corroborate de l’Armement (DGA). This work was performed within the main observations made in previous studies, except the framework of the Groupement de recherche: Metabol- for the predominance of organisms related to sulfate isme de l’Arsenic chez les Microorganismes (GDR2909- reducing bacteria (SRB) identified in water from the tail- CNRS). ings by Bruneel et al. (2005) using a cloning–sequencing approach. Instead, our results revealed the dominance of References Acidithiobacillales over SRB in these samples. The very low proportion of SRB populations in our study (on Baker BJ & Banfield JF (2003) Microbial communities in acid mine drainage. FEMS Microbiol Ecol 44: 139–152. average 0.1% of total abundance per sample) could be Bertin PN, Heinrich-Salmeron A, Pelletier E et al. (2011) partly due to differences in the physicochemical variables Metabolic diversity among main microorganisms inside an of the water, to the choice of a stringent similarity cutoff arsenic-rich ecosystem revealed by meta- and but also to the different primers used for PCR amplifica- proteo-genomics. ISME J 5: 1735–1747. tion. Furthermore, relatives of Thiomonas belonging to Bruneel O, Personne JC, Casiot C, Leblanc M, Elbaz-Poulichet the Burkholderiales order were retrieved and accounted F, Mahler BJ, Le Fleche A & Grimont P AD (2003) for < 1% of the total sequences (Fig. S1a). Despite their Mediation of arsenic oxidation by Thiomonas sp. in low abundance, several strains of Thiomonas sp. have acid-mine drainage (Carnoules, France). J Appl Microbiol 95: been previously isolated and shown to be active in the 492–499.

Bruneel O, Duran R, Koffi K, Casiot C, Fourcßans A, Hallberg KB (2010) New perspectives in acid mine drainage Elbaz-Poulichet F & Personne JC (2005) Microbial diversity microbiology. Hydrometallurgy 104: 448–453. in a pyrite-rich tailings impoundment (Carnoules, France). Hallberg KB, Coupland K, Kimura S & Johnson DB (2006) Geomicrobiol J 22: 249–257. Macroscopic streamer growths in acidic, metal-rich mine Bruneel O, Duran R, Casiot C, Elbaz-Poulichet F & Personne waters in North Wales consist of novel and remarkably JC (2006) Diversity of microorganisms in Fe-As-rich acid simple bacterial communities. Appl Environ Microbiol 72: mine drainage waters of Carnoules, France. Appl Environ 2022–2030. Microbiol 72: 551–556. Heinzel E, Hedrich S, Janneck E, Glombitza F, Seifert J & Bruneel O, Volant A, Gallien S et al. (2011) Characterization Schlomann€ M (2009) Bacterial diversity in a mine water of the active bacterial community involved in natural treatment plant. Appl Environ Microbiol 75: 858–861. attenuation processes in arsenic-rich Creek sediments. Huse SM, Welch DM, Morrison HG & Sogin ML (2010) Microb Ecol 61: 793–810. Ironing out the wrinkles in the rare biosphere through Casiot C, Morin G, Juillot F, Bruneel O, Personne JC, Leblanc improved OTU clustering. Environ Microbiol 12: 1889–1898. M, Duquesne K, Bonnefoy V & Elbaz-Poulichet F (2003a) Hwang C, Ling F, Andersen GL, LeChevallier MW & Liu WT Bacterial immobilization and oxidation of arsenic in acid (2012) Microbial community dynamics of an urban mine drainage (Carnoules creek, France). Water Res 37: drinking water distribution system subjected to phases of 2929–2936. chloramination and chlorination treatments. Appl Environ Casiot C, Leblanc M, Bruneel O, Personne JC, Koffi K & Microbiol 78: 7856–7865. Elbaz-Poulichet F (2003b) Geochemical processes controlling Johnson DB (2012) Geomicrobiology of extremely acidic the formation of As-rich waters within a tailings impoundment subsurface environments. FEMS Microbiol Ecol 81:2–12. (Carnoules, France). Aquat Geochem 9:273–290. Junier P, Junier T & Witzel KP (2008) TRiFLe, a program for Chen LX, Li JT, Chen YT, Huang LN, Hua ZS, Hu M & Shu in silico terminal restriction fragment length polymorphism WS (2013) Shifts in microbial community composition and analysis with user-defined sequence sets. Appl Environ function in the acidification of a lead/zinc mine tailings. Microbiol 74: 6452–6456. Environ Microbiol 15: 2431–2444. Koffi K, Leblanc M, Jourde H, Casiot C, Pistre S, Gouze P & Delavat F, Lett MC & Lievremont D (2012) Novel and Elbaz-Poulichet F (2003) Reverse oxidation zoning in mine unexpected bacterial diversity in an arsenic-rich tailings generating arsenic-rich acidic waters (Carnoules, ecosystem revealed by culture-dependent approaches. Biol France). Mine Water Environ 22:7–14. Direct 7: 28. Kuang JL, Huang LN, Chen LX, Hua ZS, Li SJ, Hu M, Li JT & Edwards KJ, Gihring TM & Banfield JF (1999) Seasonal Shu WS (2012) Contemporary environmental variation variations in microbial populations and environmental determines microbial diversity patterns in acid mine conditions in an extreme acid mine drainage environment. drainage. ISME J 7: 1038–1050. Appl Environ Microbiol 65: 3627–3632. Lane DJ (1991) 16S/23S rRNA sequencing. Nucleic Acid Edwards KJ, Bond PL, Druschel GK, McGuire MM, Hamers Techniques in Bacterial Systematic (Stackebrandt E & RJ & Banfield JF (2000) Geochemical and biological aspects Goodfellow M, eds), pp. 115–175. Wiley, New York. of sulfide mineral dissolution: lessons from Iron Mountain, Lear G, Niyogi D, Harding J, Dong Y & Lewis G (2009) California. Chem Geol 169: 383–397. Biofilm bacterial community structure in streams affected Egal M, Casiot C, Morin G, Elbaz-Poulichet F, Cordier MA & by acid mine drainage. Appl Environ Microbiol 75: 3455– Bruneel O (2010) An updated insight into the natural 3460. attenuation of as concentrations in Reigous Creek (southern Miller SR, Strong AL, Jones KL & Ungerer MC (2009) France). Appl Geochem 25: 1949–1957. Barcoded pyrosequencing reveals shared bacterial Fabisch M, Beulig F, Akob DM & Kusel€ K (2013) Surprising community properties along the temperature gradients of abundance of Gallionella-related iron oxidizers in creek two alkaline hot springs in Yellowstone National Park. Appl sediments at pH 4.4 or at high heavy metal concentrations. Environ Microbiol 75: 4565–4572. Front Microbiol 4: 390. Monchy S, Sanciu G, Jobard M et al. (2011) Exploring and Fuhrman JA, Steele JA, Hewson I, Schwalbach MS, Brown quantifying fungal diversity in freshwater lake ecosystems MV, Green JL & Brown JH (2008) A latitudinal diversity using rDNA cloning/sequencing and SSU tag gradient in planktonic marine bacteria. P Natl Acad Sci USA pyrosequencing. Environ Microbiol 13: 1433–1453. 105: 7774–7778. Morin G & Calas G (2006) Arsenic in soils, mine tailings, and Galand PE, Casamayor EO, Kirchman DL & Lovejoy C (2009) former industrial sites. Elements 2:97–101. Ecology of the rare microbial biosphere of the Arctic Ocean. Morin G, Juillot F, Casiot C, Bruneel O, Personne JC, P Natl Acad Sci USA 106: 22427–22432. Elbaz-Poulichet F, Leblanc M, Ildefonse P & Calas G (2003) Gonzalez-Toril E, Aguilera A, Souza-Egipsy V, Lopez Pamo E, Bacterial formation of tooeleite and mixed arsenic(III) or Sanchez Espana~ J & Amils R (2011) Geomicrobiology of La arsenic(V)-iron(III) gels in the Carnoules acid mine Zarza-Perrunal acid mine effluent (Iberian Pyritic Belt, drainage, France. A XANES, XRD, and SEM study. Environ Spain). Appl Environ Microbiol 77: 2685–2694. Sci Technol 37: 1705–1712.

Muyzer G & Ramsing NB (1995) Molecular methods to study Ward DM, Ferris MJ, Nold SC & Bateson MM (1998) A the organization of microbial communities. Water Sci natural view of microbial biodiversity within hot spring Technol 32:1–9. cyanobacterial mat communities. Microbiol Mol Biol Rev 62: Needleman SB & Wunsch CD (1970) A general method 1353–1370. applicable to the search for similarities in the amino acid sequence of two proteins. J Mol Biol 48: 443–453. Pilloni G, Granitsiotis MS, Engel M & Lueders T (2012) Supporting Information Testing the limits of 454 pyrotag sequencing: reproducibility, quantitative assessment and comparison to Additional Supporting Information may be found in the T-Rflp fingerprinting of aquifer microbes. PLoS ONE 7: online version of this article: e40467. Fig. S1. Composition of different phyla based on classifi- R Development Core Team (2012) R: A Language and cation of 16S rRNA gene sequences of bacteria from each Environment for Statistical Computing. R Foundation for sample using RDP Classifier: (a) Proteobacteria orders, Statistical Computing, Vienna, Austria. http://www. R-project.org. (b) Actinobacteria orders, (c) Firmicutes orders, and (d) Rodier J, Broutin JP, Chambon P, Champsaur H & Rodi L Acidobacteria orders. (1996) L’Analyse des Eaux. Dunod, Paris, p. 1383. Fig. S2. Histogram of the relative abundance of dominant Samanta G & Clifford DA (2005) Preservation of inorganic OTUs at the Carnoules sampling sites (G. ferruginea arsenic species in groundwater. Environ Sci Technol 39: subsp. capsiferriformans ES-2: NC_014394; A. ferrooxidans 8877–8882. strain HL1: JF815535; Thiobacillus sp. ML2-16: Schloss PD, Westcott SL, Ryabin T et al. (2009) Introducing DQ145970; Pseudomonas migulae: AY605698; A. ferrivo- mothur: open-source, platform-independent, rans SS3: NR_074660; Actinobacterium BGR 105: community-supported software for describing and GU168008; Acidobacteriaceae bacterium CH1: DQ355184; comparing microbial communities. Appl Environ Microbiol Ferrimicrobium sp. Py-F2: KC208496; Metallibacterium sp. – 75: 7537 7541. 911: HE858262; Alicyclobacillaceae bacterium Feo-D4-16- Smith CJ, Danilowicz BS, Clear AK, Costello FJ, Wilson B & CH: FN870323; Acidisphaera sp. nju-AMDS1: FJ915153; Meijer WG (2005) T-Align, a web-based tool for Betaproteobacterium OYT1: AB720115. comparison of multiple terminal restriction fragment Fig. S3. Nonmetric multidimensional scaling analysis of length polymorphism profiles. FEMS Microbiol Ecol 54: the composition of the bacterial community estimated by 375–380. Sogin ML, Morrison HG, Huber JA, Mark Welch D, Huse SM, (a) T-RFLP and (b) 454 pyrosequencing based on 16S Neal PR, Arrieta JM & Herndl GJ (2006) Microbial diversity rRNA genes. in the deep sea and the underexplored “rare biosphere”. P Fig. S4. (a) Ordination plot of CCA based on pyrose- Natl Acad Sci USA 103: 12115–12120. quencing data showing OTUs with relative abundance Streten-Joyce C, Manning J, Gibb KS, Neilan BA & Parry >5%. (b) Abundant OTUs and their correlation with DL (2013) The chemical composition and bacteria environmental variables and phylogenetic affiliation deter- communities in acid and metalliferous drainage from the mined by BLAST search. wet-dry tropics are dependent on season. Sci Total Environ Table S1. Physicochemical characteristics of the water at 443:65–79. each sampling site and sampling date. Ueno S, Le Provost G, Leger V et al. (2010) Bioinformatic Table S2. Relative abundance (in %) of total sequences of analysis of ESTs collected by Sanger and pyrosequencing bacterial 16S rRNA genes from each sample assigned to methods for a keystone forest tree species: oak. BMC different phyla. Genomics 11: 650.