Concilier Pratiques Et Régulations Naturelles »
UNIVERSITE MONTPELLIER 2- Habilitation à Diriger des Recherches
Gestion agroécologique des bioagresseurs :
« Concilier pratiques et régulations naturelles »
Thierry BREVAULT
CIRAD, Département Persyst UPR 115 AIDA « Agroécologie et Intensification Durable des cultures Annuelles » Montpellier, France
BIOPASS, Centre commun ISRA-IRD de Bel-Air Dakar, Sénégal
Remerciements
C’est a e un grand plaisir que je remercie ici tous eu ui ’o t aid à pose les jalo s de ette t ajectoire scientifique. Me i au appo teu s ui o t a ept d’ alue e oi e. Pensée particulière pour mes mentors : Guy, Serge, Pascal, Maurice, Samuel, Yves, et pou tous les oll gues et tudia ts a e les uels j’ai pa tag u e g a de di e sit d’e p ie es et de te ai s. Sans oublier ma famille, maillon indispensable.
Valor da biodiversidade : Apis mellifera adansonii, Cabuca, rio Corubal, Guiné-Bissau, 2017.
Table des matières
1. Curriculum vitae ...... 1 1.1. Parcours universitaire et professionnel ...... 1 1.2. Implication dans des projets de recherche ...... 3 1.3. Activités de formation ...... 6 1.3.1. Co-encadrements de doctorants ...... 6 . . . E ad e e ts d’ tudia ts e aste ...... 7 1.3.3. Enseignement et transfert de compétences pour les partenaires ...... 9 1.4. Expertise et évaluation de la recherche ...... 10 1.5. Vie collective ...... 10 1.6. Publications ...... 12 . . . A ti les da s des e ues à fa teu d’i pa t ...... 12 1.6.2. Articles en préparation ...... 15 . . . A ti les da s d’aut es e ues ...... 15 . . . Chapit es d’ou ages ...... 16 1.6.5. Communications à congrès ou réunions ...... 16 . . . Rappo ts d’e pe tise ...... 21 1.7. Synthèse bibliométrique ...... 22
2. Bilan des activités de recherche ...... 23 2.1. Contexte ...... 23 2.2. Problématique et approche ...... 23 2.3. Systèmes biologiques ...... 26 2.4. Diversité génétique et résistance aux insecticides chez le puceron du coton ...... 27 2.4.1. Diversité génétique et spécialisation écologique ...... 28 2.4.2. Carte de résistance aux insecticides ...... 30 2.4.3. Sélection par les traitements insecticides et compétition ...... 30 2.4.4. Implications pour la gestion des pucerons en culture cotonnière ...... 31 2.5. Traits de la résistance aux pyréthrinoïdes chez la noctuelle du coton ...... 32 2.5.1. Epidémiologie de la résistance ...... 32 2.5.2. Mécanismes de résistance ...... 34 2.5.3. Héritabilité et coût de la résistance ...... 34 2.5.4. Implications pour la gestion de la résistance ...... 35 2.5.5. Traitements insecticides en culture cotonnière et résistance des vecteurs ...... 36 2.6. Stratégies de gestion de la résistance au coton transgénique Bt...... 37
2.6.1. Co t i utio des efuges atu els à l’ olutio de la sista e ...... 37 2.6.2. Limites du coton Bt à deux toxines ...... 39 2.6.3. Des refuges dans les semences pour retarder la résistance ...... 40 2.6.4. Résistance des insectes aux cultures Bt : leçons du terrain...... 41 2.7. Conclusion...... 43
3. Projet de recherche. La biodiversité au service de la régulation des bioagresseurs ...... 45 3.1. Concept de régulation écologique ...... 45 3.2. Observer autrement ...... 47 3.2.1. Relations entre biodiversité et régulation ...... 47 . . . Le pa sage o e ad e d’ tude ...... 48 3.2.3. La prise en compte du socio-agroécosystème ...... 49 3.3. Agir autrement ...... 50 . . Cas d’ tude : La i euse de l’ pi de il ...... 50 3.4.1. Contexte ...... 50 3.4.2. Hypothèses ...... 52 3.4.3. Objectifs ...... 53 3.4.4. Approche méthodologique ...... 53 3.4.5. Collaborations scientifiques ...... 56
4. Conclusion ...... 57
5. Références bibliographiques ...... 59
6. Liste des tirés à part joints ...... 63
1. Curriculum vitae
Thierry Brévault, né le 10 mai 1969 à Vitré (Ille-et-Vilaine) Nationalité française
Etablissement et Unité de recherche Centre de coopération Internationale en Recherche Agronomique pour le Développement (Cirad), Département Performances des systèmes de production et de transformation tropicaux (Persyst) UPR Agroécologie et Intensification Durable des cultures Annuelles (AIDA) http://ur-aida.cirad.fr/ Equipe Caractérisation et gestion intégrée des risques d'origine biotique (Carabe) TA-B 115/02, Avenue Agropolis, 34398 Montpellier cedex 5 Téléphone : 04 67 61 55 00
Adresse professionnelle Laboratoire Biopass (Biologie des populations et écologie des communautés animales des écosystèmes sahélo-soudaniens) Centre commun Isra-IRD de Bel-Air, Route des Hydrocarbures, BP 1386, Dakar, Sénégal Téléphone : +221 33 849 33 31 Adresse électronique : [email protected]
Expertise et champ disciplinaire Entomologiste spécialiste en écologie appliquée à la gestion des insectes d'intérêt agricole Biologie des populations-écologie : écologie comportementale, écologie du paysage, écologie des communautés, écologie évolutive, écologie fonctionnelle (biodiversité) Page Web : http://agents.cirad.fr/index.php/Thierry+BREVAULT/
1.1. Parcours universitaire et professionnel
1988-1993 Diplô e d’I gé ieur e Agriculture E ole Sup ieu e d’Ag i ultu e d’A ge s ESA M oi e de fi d’ tudes e ad pa Gu Rodet Inra, Station de recherches de Zoologie et d’Apidologie, Do ai e St Paul, Mo tfa et). Pollinisation du trèfle blanc porte-graines par l’a eille do esti ue, Apis mellifera (Hymenoptera, Apidae).
1994-1995 Coopérant du Service National Cirad, Département Cultures Annuelles Ingénieur de recherche au Programme Arachide du Centre National de Recherches Agronomiques de Bambey (CNRA) de l’I stitut S galais de Re he hes Ag i oles Is a . Diagnostic agronomique, tests de traitements de semences en protection à la levée, lutte biotechnique contre la bruche de l’a a hide, Caryedon serratus (Coleoptera, Bruchidae).
1996-1997 Diplô e d’Etudes Approfo dies (DEA) de Biologie de l’E olutio et E ologie Montpellier SupAgro
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 1
Mémoire encadré par Serge Quilici (Cirad Réunion, Département Productions Fruitières et Horticoles, Programme Arboriculture fruitière). Ecologie visuelle chez la mouche de la tomate, Neoceratitis cyanescens (Diptera, Tephritidae).
1997-1999 Thèse de Doctorat en Biologie de l’E olutio et E ologie Montpellier SupAgro, Ecole doctorale de Biologie Intégrative Thèse encadrée par Eric Thibout CNRS, I stitut de Re he he su la Biologie de l’I se te, Tours) et Serge Quilici (Cirad Réunion, Département Productions Fruitières et Horticoles, Programme Arboriculture fruitière). Etude des m a is es de lo alisatio de l’hôte hez la mouche de la tomate, Neoceratitis cyanescens (Diptera, Tephritidae). Programme Régional de Recherche appliquée sur les mouches des fruits, soute u pa l’U io Eu op e e et coordonné par la Co issio de l’O a I die .
Depuis 2000 Chercheur au Cirad Entomologiste spécialiste en écologie appliquée à la gestion des insectes d'intérêt agricole
2000-2001 Comité national interprofessionnel de l'arachide (CNIA), Dakar, Sénégal Département Cultures annuelles, Programme cultures alimentaires Coordinateur d’u projet de relance de la filière arachide de bouche en collaboration avec l’I stitut S galais de Re he hes Ag i oles Is a et l’Institut de Technologie Alimentaire (ITA) au Sénégal.
2001-2007 Institut de Recherche Agricole pour le Développement (Irad), Garoua, Cameroun Département Cultures annuelles, Programme coton Responsable scientifique de l’ uipe Gestio i t g e des a ageu s en culture cotonnière de la Section Coton de l’I ad.
2008-2010 University of Arizona, Department of Entomology, Tucson, USA Département Persyst, UPR Systèmes de Culture Annuels (SCA) Responsable de projet su les a is es d’ olutio de la résistance des insectes aux cultures transgéniques insecticides.
2011-2012 Cirad, Montpellier, France Département Persyst, UPR SCA Cha g d’a i atio s ie tifi ue au sei de l'équipe Caractérisation et gestion intégrée des risques d'origine biotique (Carabe).
2013-2017 Laboratoire Biopass, Centre commun Isra-IRD, Dakar, Sénégal Département Persyst, UPR AIDA Chargé de recherche en écologie appliquée à la gestion des insectes ravageurs des cultures.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 2
1.2. Implication dans des projets de recherche
J’ai o e à ’i t esse de p s à l’ ologie des i se tes e 1993 lors de mon stage de Mémoi e de fi d’ tudes d’i g ieu à la Station de recherches de zoologie et d’apidologie de Montfavet, portant sur la pollinisation entomophile des cultures de trèfle blanc porte- graines. De à , j’ai o up u poste de Coopérant du Service National au Cirad, en détachement auprès du « Programme arachide » de l’Institut Sénégalais de Recherches Agricoles (Isra). J’y ai conduit des expérimentations sur les arthropodes du sol responsables de d gâts à la le e, ai si u’u diagnostic agronomique en parcelles paysannes pour l’ide tifi atio des o t ai tes à la p odu tio d’a a hide. E , j’ai ejoi t l’ uipe « Mouches des fruits » du Cirad à La Réunion, pour y effectuer mon stage de DEA, puis un Doctorat en Biologie de l’ olutio et ologie. J’ai o duit des tudes en écologie chimique et comportementale pour le d eloppe e t d’u s st e de pi geage, dans le cadre d’u Programme régional de recherche appliquée sur les mouches des fruits (Encadré 1). Ce p e ie pa d’e p ie e ’a o fo t da s o i t t pou la e he he e ologie appliquée à la gestion des insectes d'intérêt agricole en milieu tropical.
J’ai t e ut au « Programme Coton » du Cirad en février 2000, sur un poste de chercheur en entomologie, spécialiste de la protection intégrée des ravageurs du cotonnier. Détaché au « Programme Cultures alimentaires », j’ai assu l’i t i de la coordination scientifique d’u e uipe du Ci ad da s le ad e d’un p ojet de ela e de la p odu tio d’a a hide de bouche au Sénégal (Encadré 1). Mes recherches ont porté principalement sur les facteurs de contamination des graines d’a a hide pa l’aflato i e, depuis le ha p jus u’à la commercialisation, avec des expérimentations dans le bassin arachidier en saison des pluies, en station et en parcelles paysannes, dans la vallée du fleuve en irrigué en saison sèche, et avec les opérateurs de la filière dans les unités de sto kage et de d o ti age de l’a a hide.
De juin 2001 à décembre 2007, j’ai o up les fo tio s de espo sa le de l’ uipe « Protection intégrée du cotonnier » de la Section Coton à l’Irad au Cameroun, et d’a i ateu s ie tifi ue du olet « Protection intégrée des cultures et résistances » du p og a e d’appui à la e he he gio ale pou le d eloppe e t du a le des savanes d’Af i ue e t ale Encadré 1). Cette affe tatio s’i s i ait da s le ad e d’u e de a de de la Société de Développement de la culture cotonnière (Sodécoton) et de l’o ga isation de producteurs de coton (OPCC) au Cameroun, d’un appui de la recherche pour évaluer le risque de résistance chez les ravageurs clés des systèmes cotonniers et proposer un plan d’a tio le as h a t. J’ai o t i u au o tage d’u la o atoi e gional spécialisé dans la surveillance et le diag osti de la sista e le age d’i se tes, ioessais et diagnostic moléculaire), puis au développement de stratégies de gestion de la résistance à l’usage de la fili e. J’ai gale e t o duit des a ti it s de criblage de ressources génétiques pour la résistance variétale aux piqueurs suceurs, et d’e p i e tatio su l’effet des s st es de semis sous couverture végétale sur la macrofaune du sol. Une partie significative de mes activités a été consacrée à la formation des partenaires du Sud, des collègues de la recherche du Tchad et de Centrafrique, aux cadres techniques de la Sodécoton, et à la production de fiches techniques à destination des agents techniques de vulgarisation et des organisations de producteurs.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 3
De à , j’ai t a ueilli o e he heu asso i au Department of Entomology de l’U i e sit d’A izo a à Tu so Etats-Unis), internationalement reconnu pour son expertise sur les cultures génétiquement modifiées résistantes aux insectes. Mes travaux ont permis de ett e e ide e e tai s a is es d’adaptatio des populatio s d’i se tes ravageurs, avec des retombées significatives pour la gestion durable de la résistance en cultures de coton GM, notamment pour les pays du Sud (Encadré 1). Trois articles majeurs publiés dans PNAS, Evolutionary Applications et Nature Biotechnology abordent les avancées et les limites des stratégies actuellement préconisées pour gérer la résistance des insectes dans les cultures transgéniques Bt. Ce s jou a pe is d’e i hi le po tefeuille de olla o atio s ie tifi ue de l’ uipe et de fai e onnaître les compétences du Cirad outre- Atlantique. J’ai pa ti ip à de o euses u io s s ie tifi ues, do t la Conférence internationale Resistance organisée par le Rothamsted Research Centre (Harpenden, UK, 2008), la réunion du groupe de recherche IOBC-WPRS su l’i pa t e i o e e tal des o ga is es g ti ue e t odifi s , et les o g s de l’ESA (Entomological Society of America) à Reno (2008) et à San Diego (2010).
En poste à Montpellier en 2011-2012, je me suis investi dans la vie collective de l’ ta lisse e t et de l’u it de e he he SCA. J’ai ainsi contribué à la labellisation du Dispositif de e he he et d’e seig e e t e Pa te a iat Divecosys (Diversité des systèmes de production et gestion agro-écologique des bio-ag esseu s e Af i ue de l’Ouest, http://www.divecosys.org/), avec la coordination de la rédaction du projet scientifique. J’ai participé au Groupe de travail « Entomologie au Cirad ». Ce travail avait comme objectif de la ifie les e jeu et les fo de e ts de l’e to ologie au Ci ad, de d esse u tat des lieu d og aphie, fo es et fai lesses , d’ide tifie des f o ts de e he hes p io itai es et d’e d dui e des p opositio s d’ olutio et de e ou elle e t des o p te es. Enfin, j’ai pa ti ip au a age e t et à l’a i atio s ie tifi ue de l’ uipe Ca a e (18 agents), à t a e s i la ed a isatio de l’espa e de olla o atio e t e he heu s de l’ uipe, ii la mutualisation les forces sur quelques thématiques de recherche prioritaires et (iii) l’e ploitatio de la i hesse du dispositif g og aphi ue o e sou e de a ia ilit pou la conception de systèmes de culture, suite à une aluatio de l’u it de e he he.
Depuis 2013, je suis en poste au sein du laboratoire « Biologie des populations et écologie des communautés animales des écosystèmes sahélo-soudaniens » (Biopass), qui associe l’IRD-UMR CBGP (Centre de Biologie pour la Gestio des Populatio s , l’Is a I stitut S galais de Re he hes Ag i oles , l’U ad U i e sit Cheikh A ta Diop de Daka , Département de Biologie animale) et le Cirad (UPRs Aida et Hortsys), sur le Centre commun Isra-IRD de Bel-Air à Dakar (Sénégal). Mes activités de recherche portent sur les processus de régulation écologique des insectes ravageurs dans deux systèmes de production agricole contrastés: (i) cultures vivrières (mil) dans les parcs agroforestiers du bassin arachidier et (ii) cultures maraîchères dans la zone des Niayes, avec un focus sur l’i asio e te de la mineuse de la tomate, Tuta absoluta. Je ’intéresse notamment à l’effet des p ati ues ag i oles et du o te te pa sage su la st u tu e des o u aut s d’e e is atu els et leur fonction de régulation des populations de ravageurs (Encadré 1). Ces activités pa ti ipe t au positio e e t s ie tifi ue et à l’a i atio du DP Di e os s, avec l’e ad e e t s ie tifi ue de deu th ses de do to at et u e pa ti ipatio à l’e seig e e t en Master 2 de l’U i e sit Cheikh A ta Diop de Daka et de l’E ole Natio ale Sup ieu e Agronomique de Thiès.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 4
Encadré 1. Principaux projets conduits et partenariats
Programme Régional de recherche appliquée sur les mouches des fruits (PRMF). Fonds Européen de D eloppe e t FED de l’Union Européenne et de la Co issio de l’O a I die (7.ACP.RPR.400 N° REG RIN 7502), 1996-2000, k€. Partenaires : Service de la Protection des Végétaux de la Direction de l'Agriculture et de la Forêt de La Réunion, du Mi ist e de l’Ag i ultu e et des Ressou es Ma i es MARM des Se helles, du Mi ist e de l’Ag i ultu e de Mau i e, de la Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture (NAFA).
Projet de relance de la filière arachide de bouche au Sénégal. Fonds Stabex COM de l’Union Européenne et Co it atio al i te p ofessio el de l’a a hide (CNIA), 1999-2003, k€. Partenaires : Institut Sénégalais de Recherches Agricoles (Isra) et Institut de Technologie Alimentaire (ITA).
Projet d'Appui à la Recherche régionale pour un développement durable des savanes d'Afrique centrale (Prasac-Ardesac). Ministère français des Affaires Etrangères et Européennes (FSP, Fonds de Solidarité Prioritaire) et Communauté économique et monétaire d'Afrique centrale (Cemac), 2004- 2009, k€. Partenaires : Institut de Recherches Agricoles pour le Développement (Irad), Universités de Ngaoundéré et de Dschang au Cameroun, Institut Centrafricain de la Recherche Agronomique (Icra) et Université de Bangui, Institut Tchadien de Recherche Agronomique pour le Développement (Itrad).
Risk Assessment for Insect Resistance to Pyramided Bt Cotton. U.S. Department of Agriculture (USDA) National Research Initiative Competitive Grants Program -Project 2007-02227- et USDA Biotechnology Risk Assessment Grant Award 2011-33522- , k€. Partenaires: Department of Entomology, University of Arizona, Tucson, USA.
Biodiversité et gestion des bioagresseurs dans les paysages agricoles (BioBio). PI, Programme d’e elle e pou l’E seig e e t et la Re he he au Sud IRD , 2013- , k€. Partenariat avec le Prof. K. Diarra, Université Cheikh Anta Diop de Dakar, Equipe Production et protection intégrées en agroécosystèmes horticoles (2PIA).
Renforcement de la régulation écologique des insectes ravageurs des cultures de céréales sèches (Recor). Co-PI, P og a e de p odu ti it ag i ole e Af i ue de l’Ouest PPAAO/WAAPP), 2013- 201 , k€. Pa te a iat a e l’Isra et le Centre de Suivi Ecologique (CSE) au Sénégal.
Apport de la télédétection pour le renforcement de la régulation écologique des ravageurs des cultures de céréales sèches au Sénégal (Trecs). Programme Dynafrique Tosca-Cnes, 2014-2017, 160 k€. Partenariat avec l'Isra et le CSE au Sénégal, l'UMR CBGP et l'UMR Tétis (Territoires, Environnement, Télédétection et Information Spatiale) à Montpellier.
Simulating pest spread due to ecological and anthropogenic dynamics (Spread). Co-PI, USAID – Virginia Tech IPM Innovation Lab, 2015-2019, k€. Pa te a iat a e Vi gi ia Te h, l’INRA-UMR ISA Institut Sophia Agrobiotech à Sophia Antipolis.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 5
1.3. Activités de formation
Depuis le d ut de a a i e, j’ai e ad de o eu tudia ts plus de de i eau master depuis 2003, soit plus de 2 par an). Je suis a tuelle e t i pli u da s l’e ad e e t s ie tifi ue de deu do to a ts de l’E ole do to ale « Sciences de la vie, de la Santé et de l'Environnement » de l’U ad au S gal. Si j’ai pu t a s ett e à es tudia ts uel ues l e ts de thode, de fle io et d’o ga isatio pou po d e à u e uestio de e he he, j’ai e etou le plus sou e t fi i de leu appli atio et de leu a gatio à conduire les expérimentations, sur des terrains parfois très difficiles. Une bonne partie de ces stages ont abouti à une publication ou à une communication à conférence.
1.3.1. Co-encadrements de doctorants
2007-2010 : Joseph ACHALEKE, Faculté d'Agronomie et des Sciences Agricoles, Université de Dschang, Cameroun. Résistance aux pyréthrinoïdes chez la noctuelle Helicoverpa armigera en Afrique centrale: des mécanismes à la gestion. Directeur Pr. Richard Ghogomu, Université de Ds ha g. T ois pu li atio s à IF a e l’ tudia t e p e ier auteur, trois a e l’ tudia t e co-auteur.
2012-2016 : Mamadou DIATTE, Faculté des Sciences et Techniques, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Stratégies de gestion durable des principaux ravageurs de la tomate dans les périmètres maraîchers des Niayes au Sénégal. Directeur Pr. Karamoko Diarra (Ucad). Allo atio de e he he de l’Is a. U e pu li atio à IF a e l’ tudia t e premier auteur, trois a e l’ tudia t e o-auteur.
2012-2016 : Babacar LABOU, Faculté des Sciences et Techniques, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Distribution des populations de la teigne Plutella xylostella, du borer Hellula undalis, et de leurs ennemis naturels dans les cultures de chou des Niayes au Sénégal. Directeur Pr. Karamoko Diarra (Ucad). Deux publications à IF a e l’ tudia t e premier auteur.
2014-2017 : Elhadji Serigne SYLLA, Faculté des Sciences et Techniques, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Invasion de la mineuse de la tomate, Tuta absoluta, en Afrique sub-saharienne : dynamique, niche écologique et potentiel de régulation biologique. Directeur Pr. Karamoko Diarra (Ucad). S jou e alte a e à l’UMR ISA, Sophia a tipolis, France. Bourse de thèse du Sud du Cirad. Trois publications à IF a e l’ tudia t e p e ie auteur, trois a e l’ tudia t e o-auteur.
2015-2018 : Ahmadou SOW, Faculté des Sciences et Techniques, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Régulation naturelle des populations de la mineuse de l'épi de mil, Heliocheilus albipunctella, dans le bassin arachidier au Sénégal. Directeur Pr. Mbacké Sembène (Ucad). S jou e alte a e à l’UMR CBGP, Mo tpellie , France. Bourse du Service de Coop atio et d’A tio Cultu elle de l’a assade de F a e au S gal. Une publication à IF en préparation a e l’ tudia t e premier auteur.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 6
1.3.2. Encadrements d’étudia ts e master
2016 - Mariama KASSE, Master 2 Biologie animale, Spécialité Ecologie et gestion des écosystèmes, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Réponse comportementale de la noctuelle, Helicoverpa armigera, à des cotonniers écimés. - Amy MBODJ, Master 2 Biologie animale, Spécialité Entomologie, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Effet de l’ i age des oto ie s su l’i ide e des he illes de la apsule et l’ la o atio du e de e t. - Madiémé NDIAYE, Master 2 Biologie animale, Spécialité Ecologie et gestion des écosystèmes, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Incidence et régulation naturelle des populatio s de la i euse de l’ pi de il, Heliocheilus albipunctella, dans le bassin arachidier au Sénégal. - Oumar SEYDI, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Facteurs affectant la dynamique des populations de la mineuse de la tomate, Tuta absoluta, en saison des pluies.
2015 - Demba Dalla SOW, Master 2 Biologie animale, Spécialité Ecologie et gestion des écosystèmes, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Réponse de la noctuelle, Helicoverpa armigera (Lepidoptera, Noctuidae), à l'écimage des cotonniers
2014 - Amadou Oury DIALLO, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Résistance aux insecticides de la noctuelle, Helicoverpa armigera (Lepidoptera, Noctuidae). - Mousl DIAW, M oi e d’I g ieu des T a au Statisti ues. E ole Natio ale de la Statisti ue et de l’A al se E o o i ue. Proposition de méthodes statistiques pour la modélisation des relations entre pratiques agricoles, paysage et abondance d'insectes. - Coumba Souna FAYE, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Résistance aux insecticides de la teigne du chou, Plutella xylostella (Lepidoptera, Plutellidae) - Bayo LY, Master 2 Biologie animale, Spécialité Entomologie, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Caractérisation spatiale de l'incidence de la mineuse de l'épi de mil, Heliocheilus albipunctella, et de sa régulation naturelle dans le bassin arachidier au Sénégal.
2013 - Abdourahmane. DABO, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Mise au point d'un protocole de mesure de la résistance aux insecticides pour la teigne du chou, Plutella xylostella (Lepidoptera, Plutellidae). - Abdrahmane DIA, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Incidence des mouches des fruits (Diptera, Tephritidae) en cultures maraîchères dans la zone des Niayes. - Solange GUERIN, P ojet de fi d’ tudes pou le Diplô e d'I g ieu Ag o o e, Sp ialit Agrogéomatique. Université de Toulouse, Institut National Polytechnique de Toulouse,
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 7
ENSAT. Pratiques agricoles, paysage et régulation naturelle des insectes ravageurs des cultures. Etude de cas dans les Niayes au Sénégal. - Déthié NGOM, Master 2 Biologie animale, Spécialité Entomologie, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. D a i ue et a ia ilit spatiale de l’i festatio au ha p des gousses d’a a hide pa la u he, Caryedon serratus Olivier (Coleoptera, Bruchidae). - Ahmadou SOW, Master 2 Biologie animale, Spécialité Entomologie, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Systèmes agroforestiers en zone sèche et régulation naturelle des insectes ravageurs des cultures: étude de cas dans le bassin arachidier au Sénégal. - Etienne TENDENG, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Mise au point d'un test de mesure de la résistance aux insecticides chez la noctuelle de la tomate, Helicoverpa armigera (Lepidoptera, Noctuidae).
2012 - Gaëlle BERNADAS, Master 1 césure, Bordeaux Sciences Agro. Gestion et analyses des données spatialisées sous ArcGIS. - Serigne SYLLA, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Diagnostic des principaux insectes ravageurs des cultures de tomate plein champ dans les Niayes au Sénégal. - Alima NIANG FALL, Master 2 Gestion durable des agroécosystèmes horticoles, Université Cheikh Anta Diop de Dakar (Ucad), Sénégal. Diagnostic des principaux insectes ravageurs du chou et du parasitisme associé dans la zone des Niayes au Sénégal. - Mélanie MARCHAND, Master 1 Biodiversité Végétale Tropicale, Université de Montpellier-SupAg o, F a e. Effet de l’ i age su les i te a tio s e t e le otonnier et ses insectes ravageurs.
2007 - Adeli e BERTRAND, I g ieu Ag o o e de l’ENSA de Re es. Volontariat International en Entreprise (VIE).
2006 - Laurent COUSTON, Master AgroParisTech. Volontariat International en Entreprise (VIE). - Yakouba OUMAROU, Master 2 Biotechnologies et productions végétales, Option entomologie, Faculté des Sciences Agronomiques (FASA), Université de Dschang, Cameroun. Activité et rémanence d’i se ti ides pou le o t ôle des he illes de la capsule (Lepidoptera, Noctuidae) en culture cotonnière. - Julien TRIBOT, Master 2 Biologie et Technologies du Végétal, Spécialité professionnelle technologies du végétal et productions spécialisées, I stitut Natio al d’Ho ti ultu e d’Angers (INH), France. La résistance aux insecticides peut-elle expliquer la faible diversité génétique du puceron Aphis gossypii en culture cotonnière ?
2005 - Daphné LINDERME, M oi e de fi d’ tudes d’I g ieu e Ag i ultu e, I stitut Sup ieu d’Ag i ultu e de Beau ais (ISAB), France. Processus de colonisation des
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 8
parcelles de cotonniers par le puceron Aphis gossypii Glover (Hemiptera : Aphididae) au Nord Cameroun. - NADAMA, M oi e de fi d’ tudes d’I g ieu ag o o e, FASA, U i e sit de Ds ha g, Cameroun. Influence de trois modes de gestion des sols sur le profil de la macrofaune du sol en parcelles cotonnières paysannes au nord du Cameroun. - Sébastien PICAULT, I g ieu e Ag i ultu e de l’E ole sup ieu e d’Ag i ultu e d’Angers (ESA). Volontariat International en Entreprise (VIE).
2004 - Simon BIKAY, M oi e de fi d’ tudes d’Ingénieur agronome, FASA, Université de Dschang, Cameroun. Inventaire de la macrofaune en culture cotonnière sous quatre modes de gestion des sols : cas de Windé Pintchoumba (Nord) et Zouana (Extrême-Nord). - Guilain BIGOUNDOU, M oi e de fi d’ tudes d’I g ieu ag o o e, FASA, U i e sit de Dschang, Cameroun. Surveillance et gestion de la résistance aux pyréthrinoïdes chez Helicoverpa armigera (Hübner) en système cotonnier au Nord-Cameroun.
2003 - Nicolas GERARD, Ingénieur Agronome de Montpellier SupAgro. Volontariat International en Entreprise (VIE). - Yakouba OUMAROU, Maîtrise en Biologie et physiologie animales, Université de Ngaoundéré, Cameroun. Surveillance de la résistance aux pyréthrinoïdes chez Helicoverpa armigera en système cotonnier au Nord-Cameroun. - Louis DAYANG, Maîtrise en Biologie et physiologie animales, Université de Ngaoundéré, Ca e ou . Effi a it et a e e d’i se ti ides s st i ues appli u s e t aite e t des semences de coton pour la lutte contre les insectes piqueurs suceurs.
1.3.3. Enseignement et transfert de compétences pour les partenaires
2017 : Cours- o f e e au Maste de Biologie a i ale de l’U i e sit Cheikh A ta Diop de Dakar. Pollinisation, abeilles et biodiversité.
2017 : Enseignement à l’E ole Natio ale Sup ieu e Ag o o i ue de Thi s ENSA su la régulation écologique des ravageurs des cultures. Cou s de h da s le ad e d’u odule su l’intensification écologique des systèmes de culture sahéliens.
2015 : Formation professionnelle des agents de la Direction de la Protection des Végétaux et de ep se ta ts d’o ga isatio s de p odu teu s a aî he s au S gal. Cou s de h su la iologie, l’ ologie et le o t ôle du a ageu i asif Tuta absoluta.
2011 et 2012 : Enseignement au Master complémentaire en protection des cultures tropicales et subtropicales, Université de Liège, Gembloux Agro-Bio Tech, Université catholique de Louvain et Montpellier SupAgro. Cours de 3 h sur la gestion de la résistance des insectes ravageurs aux insecticides, en particulier dans le cas des cultures Bt transgéniques.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 9
2002 à 2007 : Fo atio p ofessio elle des age ts d’appui te h i ue de la Sodécoton au Cameroun. Cours et travaux pratiques de 16 h sur les techniques de surveillance et de gestion de la résistance des ravageurs aux insecticides.
1.4. Expertise et évaluation de la recherche
Membre du Comité scientifique du Haut Conseil des Biotechnologies depuis 2015 (http://www.hautconseildesbiotechnologies.fr/fr), instance française indépendante chargée d’ lai e la d isio pu li ue sur des questions relatives aux biotechnologies, notamment les organismes génétiquement modifiés.
Evaluation de projets de recherche pour l’INRA M tap og a e E ose et l’I te atio al Foundation for Science (IFS grantees).
Relecteur pour des journaux scientifiques en entomologie, écologie, et sciences multidisciplinaires : Scientific Reports, PLoS One, Agriculture Ecosystems and Environment, Crop Protection, Bulletin of Entomological Research, Entomologia Experimentalis et Applicata, International Journal of Pest Management, Bioinvasion records, Cahiers Agricultures.
1.5. Vie collective
Organisation d’ateliers et conférences - Co-organisation d’u atelie de o e tatio da s le ad e du projet Biophora (Action incitative Cirad). Campus international de Baillarguet, Montpellier, 23 mai 2016. Cet atelier a permis de réunir localement ou par visioconférence (La Réunion, Guadeloupe, Sénégal, Kenya, Costa Rica et Burkina Faso) 25 chercheurs et deux chargés de valorisation du Cirad. Un vrai front de recherche existe sur le biocontrôle assisté par des entomovecteurs, avec beaucoup de questions sur chaque étape du processus de contamination et de diffusion des biocides, et sur les effets non intentionnels de la technologie. - A i atio d’u atelie s ie tifi ue da s le ad e du p ojet TopCot A tio i itati e Cirad). Laboratoire Farce, Université de Neuchâtel, Suisse, 18-19 mai 2016. Cet atelier a réuni une douzaine de chercheurs et techniciens (UPR Aida, UPR Hortsys, UMR PVBMT, IER et Laboratoire Farce) autour des mécanismes de défense naturelle des plantes. Il a pe is i d’ ta li u a e as d’ tudes à o dui e e su les a is es de d fe se des pla tes i pli u es da s le as de l’ i age du cotonnier, (ii) de partager des app o hes i o a tes et des thodes, iii d’a o e des olla o atio s et de pose les p e ie s jalo s d’u p ojet de e he he su l’i du tio de a tio s de d fe se du oto ie pa l’ i age et so utilisatio e protection des cultures. - Participation aux Ateliers Systèmes de Culture du Département Persyst. Co-auteur de la réflexion sur les « Perspectives de recherches sur les bio-régulations ». Juin 2015. - Me e du Co it lo al d’o ga isatio du s i ai e du DP Divecosys « Stratégies de protection des cultures : du modèle biologique au territoire », tenu du 2 au 4 juin 2015 su le a pus de l’U i e sit Cheikh A ta Diop de Daka UCAD . Plus d’u e soi a tai e
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 10
de pa ti ipa ts, e us du B i , de Côte d’I oi e, de France, du Mali, du Sénégal ou du Togo, o t p se t des etou s d’e p ie e et des sultats de e he he e ati e de gestion agroécologique des bioagresseurs. Un atelier sur le thème « stratégies de protection des cultures : du modèle biologique au territoire », a permis, à partir d'une analyse fine des systèmes biologiques et techniques en jeu, de réfléchir aux déclinaisons de l’a tio i di iduelle de l'ag i ulteu du ha p à l’e ploitatio , à elle olle ti e au sei de filières, organisations de produ teu s, et aut es seau à l’ helle d’u te itoi e. Cette fle io a d ou h su l’ide tifi atio de gaps de o aissa e e p i s e te es de uestio s ou d’h poth ses de e he he pou l’ ta lisse e t d'u e graine de projet collaboratif. - Co-organisatio d’u atelie de fo atio i titul « M thodes et outils d’e ploitatio de données spatialisées. Application aux études d’ ologie et à la gestio des populatio s de bioagresseurs ». Cet atelier, financé conjointement par le Cirad et le projet Peers-BioBio, s’est te u au Ce t e de Sui i E ologi ue CSE , du au o e e à Daka . Il a réuni des chercheurs, des professionnels et des étudiants en thèse ayant déjà inscrit dans leurs activités une approche spatiale pour répondre à une problématique de gestion des bioagresseurs. Responsable pédagogique : V. Soti (UPR Aida). - Co t i utio à l’a i atio s ie tifi ue d’u atelie d’ itu e de p ojet, te u du au octobre 2012 à Dakar (Sénégal), autour du DP Divecosys. Ce séminaire a rassemblé 13 chercheurs de France, du Sénégal, du Bénin et du Mali autour des questions suivantes : Qui sommes-nous ? Que voulons-nous faire ensemble ? Quel cadre conceptuel pour nos travaux ? Quelles sont nos questions de recherche ? Quels axes structurants ? - Contributio à l’atelie du DP Di e os s i titul « Atelie d’ ha ges et d’ itu e de projet. Thème "habitat et plantes compagnes », 18-20 octobre 2011 à Cotonou (Bénin). Cet atelie a u i u e ua a tai e de he heu s et d’e seig a ts-chercheurs de différents pays : Sénégal, Mali, Burkina Faso, Niger, Togo, Bénin et France. - O ga isatio d’u atelie s ie tifi ue i titul « To a ds a ultis ale app oa h fo improving pest management », 4-5 octobre 2011, Cirad Montpellier. Une soixantaine de chercheurs de différentes institutions de recherche (20) ont échangé sur les outils d’ aluatio du is ue, de su eilla e et de d te tio , ai si ue des thodes de gestio du a le des i se tes d’i t t ag i ole, di al ou t i ai e. Les 25 communications sont en lig e su l’a hi e ou e te HAL http://hal.archives-ouvertes.fr/cirad-00645926/.
Appui scientifique et groupes de travail - Mission Cameroun, 9-12 mai 2011. Bilan-diagnostic des activités de l’ uipe Ca a e au Nord Cameroun (P. Prudent et J. Sorèze). Evaluation des perspectives sur les thématiques scientifiques prioritaires et la dimension du partenariat, en adéquation avec le projet de l’ uipe Ca a e. - Mission Mali, 5-7 décembre 2011. Point sur les activités de recherche développées par A. Re ou et ses oll gues au sei du P og a e oto de l’IER. - A i atio d’u p ojet de e he he su l’ i age des oto ie s o e a i e hi i ue à la po te des No tuidae, a e la ise e pla e d’a ti ités de recherche complémentaires au Mali et au Cameroun (2011-2012). - Participation au Groupe de travail « Entomologie au Cirad », 21-22 novembre 2011. P se tatio d’u e o u i atio su « U e isio ag o ologi ue pou la gestio des insectes ravageurs » en collaboration avec A. Ratnadass (UPR Hortsys).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 11
Participation à comités de thèse 2015-2018 : Membre du comité de thèse d’Ibrahima Thiaw, Université Gaston Berger, Centre de Suivi Ecologique, Dakar, Sénégal. Co- o eptio d’u pa sage fo tio el pou u e gestio ag o ologi ue d’u a ageu des ultu es : le as de la i euse de l’ pi de il au Sénégal. 2012-2014 : Membre du comité de thèse de Toulassi Nurbel, Université de La Réunion, UMR PVBMT. Relations insecte-plante chez deux espèces de mouches des fruits nuisibles aux Cucurbitaceae. 2012-2014 : Membre du comité de thèse de Noëlline Tsafack, ENSA Toulouse, INRA-UMR Dynafor, Toulouse, France. Co t i utio de l’a al se pa sag e à l’ tude de l’i festatio des parcelles de coton par la noctuelle, Helicoverpa armigera, au Nord Bénin.
Participation à jury de recrutement au Cirad - Poste 2338 « Agroécologue spécialiste de la dynamique de populations et de la régulation des bioagresseurs » pou l’UPR Aida décembre 2015). - Poste 2381 « Systématicien entomologiste maîtrisant les approches morphologiques et moléculaires » pou l’UMR CBGP d e e . - Poste « E to ologiste sp ialiste e Biologie des populatio s » pou l’UMR Bioagresseurs (23 août 2011). - Poste 1706 « Chercheu e ag o o ie du pa sage » pou l’UR SCA -22 juin 2011). - Poste « Ag o o e ologue » pou l’UR SCA -22 juin 2011).
1.6. Publications Les noms des étudiants encadrés en Doctorat ou en Master sont soulignés. Les tirés à part des publications grisées sont annexés au document.
1.6.1. Arti les da s des revues à fa teur d’i pa t (indexées dans Web of Science)
1. Labou B, Brévault T, Sylla S, Diatte M, Bordat D & Diarra K (2017) Spatial and temporal incidence of insect pests in cabbage farmer fields in Senegal. International Journal of Tropical Insect Science (sous presse). 2. Sylla S, Brévault T, Bal AB, Chailleux A, Diatte M, Desneux N & Diarra K (2017) Rapid spread of the tomato leafminer, Tuta absoluta (Lepidoptera, Gelechiidae), an invasive pest in sub-Saharan Africa. Entomologia Generalis (sous presse). 3. Diatte M, Brévault T, Sylla S, Tendeng E, Sall-Sy D & Diarra K (2017) Arthropod pest complex and associated damage in field-grown tomato in Senegal. International Journal of Tropical Insect Science (sous presse). 4. Sylla S, Brévault T, Diarra K, Bearez P & Desneux N (2016) Life-history traits of Macrolophus pygmaeus with different prey foods. Plos One 11: e0166610. 5. Labou B, Brévault T, Bordat D & Diarra K (2016) Determinants of parasitoid assemblages of the diamondback moth, Plutella xylostella, in cabbage farmer fields in Senegal. Crop Protection 89: 6– 11. 6. Sylla S, Brévault T, Streito J-C & Diarra K (2016) First Record of Nesidiocoris tenuis (Reuter) (Heteroptera: Miridae), as a predator of the tomato leaf miner, Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae), in Senegal. Egyptian Journal of Biological Pest Control 26: 851–853. 7. Brévault T, Tabashnik BE & Carrière Y (2015) A seed mixture increases dominance of resistance to Bt cotton in Helicoverpa zea. Scientific Reports 5: 9807.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 12
8. Brévault T, Renou A, Vayssières J-F, Amadji G, Assogba-Komlan F, Diallo MD, De Bon H, Diarra K, Hamadoun A, Huat J, Marnotte P, Menozzi P, Prudent P, Rey J-Y, Sall D, Silvie P, Simon S, Sinzogan A, Soti V, Tamo M & Clouvel P (2014) DIVECOSYS: Bringing together researchers to design ecologically-based pest management for small-scale farming systems in West Africa. Crop Protection 66: 53–60. 9. Brévault T, Sylla S, Diatte M, Bernadas G & Diarra K (2014) Tuta absoluta Meyrick (Lepidoptera: Gelechiidae): a new threat to tomato production in sub-Saharan Africa. African Entomology 22: 441–444. 10. Brévault T, Heuberger S, Zhang M, Ellers-Kirk C, Ni X, Masson L, Li X, Tabashnik BE & Carrière Y (2013) Potential shortfall of pyramided transgenic cotton for insect resistance management. Proceedings of the National Academy of Sciences of the United States of America 110: 5806– 5811. 11. Silvie PJ, Renou A, Vodounnon S, Bonni G, Adegnika MO, Hema O, Prudent P, Soreze J, Ochou GO, Togola M, Badiane D, Ndour A, Akantetou PK, Ayeva B & Brévault T (2013) Threshold-based i te e tio s fo otto pest o t ol i West Af i a: What’s up ea s late ? C op P ote tio : 157–165. 12. Tabashnik BE, Brévault T & Carrière Y (2013) Insect resistance to Bt crops: lessons from the first billion acres. Nature Biotechnology 31: 510–521. 13. Tsafack N, Menozzi P, Brévault T, Soti V, Deconchat M & Ouin A (2013) Effects of landscape context and agricultural practices on the abundance of cotton bollworm Helicoverpa armigera in cotton fields: A case study in northern Benin. International Journal of Pest Management 59: 294– 302. 14. Brévault T, Nibouche S, Achaleke J & Carrière Y (2012) Assessing the role of non-cotton refuges in delaying Helicoverpa armigera resistance to Bt cotton in West Africa. Evolutionary Applications 5: 53–65. 15. Brévault T, Carletto J, Tribot J & Vanlerberghe-Masutti F (2011) Insecticide use and competition shape the genetic diversity of the aphid Aphis gossypii in a cotton-growing landscape. Bulletin of Entomological Research 101: 407–413. 16. Heuberger S, Crowder DW, Brévault T, Tabashnik BE & Carrière Y (2011) Modeling the effects of plant-to-plant gene flow, larval behavior, and refuge size on pest resistance to Bt cotton. Environmental Entomology 40: 484–495. 17. Achaleke J & Brévault T (2010) Inheritance and stability of pyrethroid resistance in the cotton bollworm Helicoverpa armigera (Lepidoptera: Noctuidae) in Central Africa. Pest Management Science 66: 137–141. 18. Brévault T & Quilici S (2010) Interaction between visual and olfactory cues during host finding in the tomato fruit fly Neoceratitis cyanescens. Journal of Chemical Ecology 36: 249–259. 19. Brévault T & Quilici S (2010) Flower and fruit volatiles assist host-plant location in the tomato fruit fly Neoceratitis cyanescens. Physiological Entomology 35: 9–18. 20. Carletto J, Martin T, Vanlerberghe-Masutti F & Brévault T (2010) Insecticide resistance traits differ among and within host races in Aphis gossypii. Pest Management Science 66: 301–307. 21. Houndete TA, Ketoh GK, Hema OSA, Brévault T, Glitho IA & Martin T (2010) Insecticide resistance in field populations of Bemisia tabaci (Hemiptera: Aleyrodidae) in West Africa. Pest Management Science 66: 1181–1185. 22. Ryckewaert P, Deguine J-P, Brévault T & Vayssières JF (2010) Fruit flies (Diptera: Tephritidae) on vegetable crops in Reunion Island (Indian Ocean): state of knowledge, control methods and prospects for management. Fruits 65: 113–130. 23. Achaleke J, Martin T, Ghogomu RT, Vaissayre M & Brévault T (2009) Esterase-mediated resistance to pyrethroids in field populations of Helicoverpa armigera (Lepidoptera: Noctuidae) from Central Africa. Pest Management Science 65: 1147–1154. 24. Achaleke J, Vaissayre M & Brévault T (2009) Evaluating pyrethroid alternatives for the management of cotton bollworms and resistance in Cameroon. Experimental Agriculture 45: 35– 46.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 13
25. Brévault T, Couston L, Bertrand A, Theze M, Nibouche S & Vaissayre M (2009) Sequential pegboard to support small farmers in cotton pest control decision-making in Cameroon. Crop Protection 28: 968–973. 26. Brévault T, Oumarou Y, Achaleke J, Vaissayre M & Nibouche S (2009) Initial activity and persistence of insecticides for the control of bollworms (Lepidoptera: Noctuidae) in cotton crops. Crop Protection 28: 401–406. 27. Brévault T, Prudent P, Vaissayre M & Carrière Y (2009) Susceptibility of Helicoverpa armigera (Lepidoptera: Noctuidae) to Cry1Ac and Cry2Ab2 insecticidal proteins in four countries of the West African cotton belt. Journal of Economic Entomology 102: 2301–2309. 28. Brévault T, Guibert H & Naudin K (2009) Preliminary studies of pest constraints to cotton seedlings in a direct seeding mulch-based system in Cameroon. Experimental Agriculture 45: 25– 33. 29. Brévault T & Quilici S (2009) Oviposition preference in the oligophagous tomato fruit fly, Neoceratitis cyanescens. Entomologia Experimentalis et Applicata 133: 165–173. 30. Carletto J, Lombaert E, Chavigny P, Brévault T, Lapchin L & Vanlerberghe-Masutti F (2009) Ecological specialization of the aphid Aphis gossypii Glover on cultivated host plants. Molecular Ecology 18: 2198–2212. 31. Brévault T, Duyck P-F & Quilici S (2008) Life-history strategy in an oligophagous tephritid: the tomato fruit fly, Neoceratitis cyanescens. Ecological Entomology 33: 529–536. 32. Brévault T, Achaleke J, Sougnabe SP & Vaissayre M (2008) Tracking pyrethroid resistance in the polyphagous bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae), in the shifting landscape of a cotton-growing area. Bulletin of Entomological Research 98: 565–573. 33. Brévault T, Carletto J, Linderme D & Vanlerberghe-Masutti F (2008) Genetic diversity of the cotton aphid Aphis gossypii in the unstable environment of a cotton growing area. Agricultural and Forest Entomology 10: 215–223. 34. Brévault T & Quilici S (2008) Sexual attraction, male courtship and female remating in the tomato fruit fly, Neoceratitis cyanescens. Journal of Insect Behavior 21: 366–374. 35. Chouaibou M, Etang J, Brévault T, Nwane P, Hinzoumbe CK, Mimpfoundi R & Simard F (2008) Dynamics of insecticide resistance in the malaria vector Anopheles gambiae s.l. from an area of extensive cotton cultivation in Northern Cameroon. Tropical Medicine & International Health 13: 476–486. 36. Nibouche S, Brévault T, Klassou C, Dessauw D & Hau B (2008) Assessment of the resistance of cotton germplasm (Gossypium spp.) to aphids (Homoptera, Aphididae) and leafhoppers Ho opte a : Ci adellidae, T phlo i ae : ethodolog a d ge eti a ia ilit . Pla t B eedi g 127: 376–382. 37. Brévault T, Bikay S, Maldes JM & Naudin K (2007) Impact of a no-till with mulch soil management strategy on soil macrofauna communities in a cotton cropping system. Soil & Tillage Research 97: 140–149. 38. Brévault T & Quilici S (2007) Visual response of the tomato fruit fly, Neoceratitis cyanescens, to colored fruit models. Entomologia Experimentalis et Applicata 125: 45–54. 39. Brévault T & Quilici S (2007) Influence of habitat pattern on orientation during host fruit location in the tomato fruit fly, Neoceratitis cyanescens. Bulletin of Entomological Research 97: 637–642. 40. Nibouche S, Gozé E, Babin R, Beyo J & Brévault T (2007) Modeling Helicoverpa armigera (Hubner) Lepidopte a : No tuidae da ages o otto . E i o e tal E to olog : 151–156. 41. Brévault T & Quilici S (2000) Relationships between temperature, development and survival of different life stages of the tomato fruit fly, Neoceratitis cyanescens. Entomologia Experimentalis Et Applicata 94: 25–30. 42. Brévault T & Quilici S (2000) Diel patterns of reproductive activities in the tomato fruit fly, Neoceratitis cyanescens. Physiological Entomology 25: 233–241. 43. Brévault T & Quilici S (1999) Factors affecting behavioural responses to visual stimuli in the tomato fruit fly, Neoceratitis cyanescens. Physiological Entomology 24: 333–338.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 14
44. Rodet G, Vaissière BE, Brévault T & Torre Grossa JP (1998) Status of self-pollen in bee pollination efficiency of white clover (Trifolium repens L.). Oecologia 114: 93–99.
1.6.2. Articles en préparation
45. Diatte M, Brévault T, Sall-Sy D & Diarra K (2017) Parasitoid control of the tomato fruitworm, Helicoverpa armigera, in smallholder farmer fields in Senegal. International Journal of Pest Management (soumis). 46. Sylla S, Brévault T, Diarra K & Desneux N (2017) Geographic variation of host preference in the invasive tomato leafminer Tuta absoluta: implications for host range expansion. Journal of Pest Science (en préparation). 47. Soti V, Billand C, Lelong C, Goebel R & Brévault T. (2017) Processing high resolution satellite imagery to design a field sampling plan for pest monitoring. Environmental Entomology (en préparation). 48. Brévault T, Sow A, Thiaw C, Thiaw I & Soti V (2017) Effect of crop management and landscape on the incidence and crop damage of the millet head miner. Agriculture Ecosystems and Environment (en préparation). 49. Sow A, Brévault T, Delvare G, Hara J, Be oit L, Cœu D’a ie A, Gala M, Thiaw C & Sembène M (2017) A review of natural enemies of the millet headminer, Heliocheilus albipunctella : morphological and molecular identification using high-throughput sequencing. Biological Control (en préparation).
1.6.3. Articles dans d’autres revues
1. Thiaw I, Soti V, Goebel F-R, Sow A, Brévault T & Diakhaté M (2017) Effect of landscape diversity on biocontrol of the millet head miner, Heliocheilus albipunctella (Lepidoptera: Noctuidae) in Senegal. Landscape management for functional biodiversity, IOBC-WPRS Bulletin 122: 38-42. 2. Brévault T & Bouyer J (2014) From integrated to system-wide pest management: Challenges for sustainable agriculture. Outlooks on Pest Management 25: 212-213. 3. Tabashnik BE, Brévault T & Carrière Y (2014) Insect resistance to genetically engineered crops: Successes and failures. Information Systems for Biotechnology 1-4. 4. Renou A, Togola M, Téréta I & Brévault T (2012) First steps towards "Green" cotton in Mali. Outlooks on Pest Management 23: 173-176. 5. Tsafack N, Menozzi P, Brévault T, Deconchat M & Ouin A (2012) Is there a landscape effect on moth pest (H. armigera) abundance and infestation rate in cotton fields in North Benin? Landscape management for functional biodiversity, IOBC-WPRS Bulletin 75 : 191-195 6. Brévault T, Prudent P, Vaissayre M (2008) Baseline susceptibility of Helicoverpa armigera (Hübner) to Bt toxins Cry1Ac and Cry2Ab2 in West Africa. IOBC/WPRS Bulletin 33: 37-42. 7. Naudin K, Bikay S, Maldes JM & Brévault T (2008) Impacts des SCV sur la macrofaune, cas du coton au Nord Cameroun. Terre Malgache 26: 149-151. 8. Vassal JM, Brévault T, Achaleke J & Menozzi P (2008) Genetic structure of the polyphagous pest Helicoverpa armigera (Lepidoptera: Noctuidae) across the sub-saharan cotton belt. Communications in Agricultural and Applied Biological Sciences 73: 433-437. 9. Brévault T & Achaleke J (2005) Status of pyrethroid resistance in the cotton bollworm, Helicoverpa armigera, in Cameroun. Resistant Pest Management 15: 4-7. 10. Nibouche S, Beyo J, Brévault T & Staetz C (2002) Negative cross insensitivity in a dimethoate resistant strain of cotton aphid Aphis gossypii Glover in Northern Cameroon. Resistant Pest Management 12: 25-26.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 15
11. Brévault T, Asfom P, Beyo J, Nibouche S & Vaissayre M (2002) Assessment of Helicoverpa armigera resistance to pyrethroid insecticides in northern Cameroon. Communications in agricultural and applied biological sciences 67: 641-646. 12. Brévault T, Quilici S & Glénac S (1999) Mouche de la tomate à l'île de la Réunion : utiliser les signaux émis par la plante hôte pour piéger les femelles. Phytoma-La Défense des Végétaux 515: 35-36.
1.6.4. Chapitres d’ouvrages
1. Silvie P, Brévault T & Deguine JP (2017) IPM case studies : cotton. In : Aphids as crop pests. Van Emden Helmut F, Harrington R. Wallingford : CABI Publishing (sous presse). 2. Renou A & Brévault T (2016) Ravageurs et maladies du cotonnier, et gestion intégrée des ravageurs. In : Crétenet Michel (ed.), Gourlot Jean-Paul (ed.). Le cotonnier. Versailles, Ed. Quae, p. 109-154. 3. Brévault T, Soti V, Thiaw C & Clouvel P (2015) Maîtriser les paysages et les processus écologiques propres à cette échelle. In : Escadafal Richard (ed.), Masse Dominique (ed.), Chotte Jean-Luc (ed.), Scopel Eric (ed.). L'ingénierie écologique pour une agriculture durable dans les zones arides et semi-arides d'Afrique de l'Ouest. Montpellier : CSFD, Agropolis International, p. 46-49. 4. Quilici S, Atiama-Nurbel T & Brévault T (2014) Plant odors as fruit fly attractants. In : Shelly, T.E. ; Epsky, N. ; Jang, E.B. ; Reyes-Flores, J. ; Vargas, R.I. (eds.). Trapping and the Detection, Control, and Regulation of Tephritid Fruit Flies. Dordrecht : Springer, p. 119-144. 5. Thiéry D, Brévault T, Quilici S, Dormont L & Schatz B (2013) Recherche de la plante hôte à distance. In: Sauvion Nicolas (ed.), Calatayud Paul-André (ed.), Thiéry Denis (ed.), Marion-Poll Frédéric (ed.). Interactions insectes-plantes. Marseille : IRD [Marseille], p. 319-346. 6. Brévault T (2003) Datasheet on alien invasive species, Neoceratitis cyanescens Bezzi. CABI Crop Protection Compendium. 7. Brévault T & Nibouche S (2002) Résistance des insectes aux insecticides en Afrique de l'ouest et du centre. Atelier sur la résistance des insectes aux insecticides en Afrique de l'ouest et du centre, 2002-03-06/2002-03-07, Maroua (Cameroun).
1.6.5. Communications à congrès ou réunions
1. Brévault T, Sow A, Thiaw I, Thiaw C, Delvare G & Soti V (2016) Tree-crop agroforestry systems promote natural control of the millet head miner, Heliocheilus albipunctella. Poster. 25 th International Congress of Entomology: Entomology without Borders, 2016-09-25/2016-09-30, Orlando (Etats-Unis). 2. Clouvel P, Goebel FR, Soti V & Brévault T (2016) Paradigm shift in agriculture: agro-ecological transition & ecological intensification, two case studies in Reunion Island and Senegal. International Symposium on Energy Challenges and Mechanics: Towards a Big Picture (ECM6). 6, 2016-08-14/2016-08-18, Inverness (Royaume-Uni). 3. Renou A, Téréta I, Togola M, Brévault T & Goebel FR (2016) Le retour d'une ancienne pratique : l'écimage des cotonniers. In : Gourlot Jean-Paul (ed.), Fruteau De Laclos Anne-Laure (ed.), Sigrist Jean-Charles (ed.), Ndoye Ousmane (ed.), Fortuno Sophie (ed.), Gérardeaux Edward (ed.). Atelier innovations techniques et indicateurs de durabilité sur la culture du coton. 2015-09-14/2015-09- 18, Dakar (Sénégal). 4. Sylla S, Diarra K, Desneux N & Brévault T (2016) Invasion of the tomato leafminer, Tuta absoluta in West Africa: Spatial dynamics, ecological niche, and potential for biological control. 25 th International Congress of Entomology. 25, 2016-09-25/2016-09-30, Orlando (Etats-Unis).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 16
5. Tereta I, Brévault T, Sissoko F, Goebel FR, Renou A (2016) Cotton topping as a way to reduce farmer's reliance on insecticides in Mali. Poster. World Cotton Research Conference. 6, 2016-05- 02/2016-05-06, Goiania (Brésil). 6. Thiaw I, Diakhate M, Goebel FR, Menozzi P, Brévault T & Soti V (2016) Apport de la télédétection à l'écologie du paysage au service de la régulation naturelle des ravageurs des cultures. In : Atelier innovations techniques et indicateurs de durabilité sur la culture du coton. 2015-09-14/2015-09- 18, Dakar (Sénégal). 7. Brévault T (2015) Arrêt sur un programme de recherche à Biopass : La biodiversité au service de régulation des insectes ravageurs de cultures. UMR CBGP, 9 juillet 2015, Montpellier (France). 8. Brévault T (2015) Repenser la gestion des ravageurs des cultures. Réunion interrégionale de o e tatio des pa te ai es de ise e œu e du p ojet oto : Contribuer à la compétitivité et à l’i te sifi atio du a le des fili es cotonnières africaines par le développement des capacités en Gestion Intégrée de la Production et des Déprédateurs GIPD. GCP/RAF/482/EC. Hôtel Faidherbe, 18-19 février 2015, Dakar (Sénégal). 9. Sylla S, Brévault T & Diarra K (2015) Invasion de la mineuse de la tomate, Tuta absoluta, au Sénégal: dynamique des populations, plantes-hôtes et ennemis naturels. Poster. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06- 02/2015-06-04, Dakar (Sénégal). 10. Clouvel P, Martin P & Brévault T (2015) Introduction à l'atelier de réflexion : "stratégies de protection des cultures du modèle biologique au territoire". Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 11. Dia A, Brévault T, Fall M & Diarra K (2015) Incidence des mouches des fruits (Diptera, Tephritidae) en cultures maraîchères dans la zone des Niayes au Sénégal. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 12. Diatte M, Brévault T, Sylla S, Sall-Sy D, Coly EV & Diarra K (2015) Incidence de deux ravageurs clés des cultures de tomate plein champ dans la zone maraîchère des Niayes au Sénégal. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06- 02/2015-06-04, Dakar (Sénégal). 13. Diatte M, Brévault T, Labou B, Sylla S, Sall D & Diarra K (2015) Pratiques phytosanitaires dans la zone maraîchère des Niayes. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 14. Labou B, Brévault T, Niang Fall A & Diarra K (2015) La teigne du chou, Plutella xylostella (Lepidoptera, Plutellidae), ravageur clé dans la zone maraîchère des Niayes au Sénégal. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06- 02/2015-06-04, Dakar (Sénégal). 15. Ngom D, Touré M, Diagne L, Thiaw C & Brévault T (2015) Dynamique et variabilité spatiale de l'infestation au champ des gousses d'arachide par la bruche, Caryedon serratus (Coleoptera, Bruchidae). Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 16. Ly B, Thiaw C, Sow A & Brévault T (2015) Dynamique de vol et incidence des populations de la mineuse de l'épi de mil, Heliocheilus albipunctella (Lepidoptera, Noctuidae). Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 17. Soti V, Diaw M, Sow A, Thiaw I, Thiaw C & Brévault T (2015) Effet des pratiques culturales et du contexte paysager sur l'abondance des populations de la mineuse de l'épi de mil, Heliocheilus albipunctella (Lepidoptera, Noctuidae), dans la zone de Bambey au Sénégal. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 18. Sow A, Ly B, Thiaw C & Brévault T (2015) Régulation naturelle de la mineuse de l'épi de mil, Heliocheilus albipunctella, dans la zone de Bambey au Sénégal. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 17
19. Tendeng E, Brévault T, Diatte M, Faye C, Dabo A, Diallo AO, Diarra K (2015) Résistance aux insecticides chez deux ravageurs clés des cultures maraîchères au Sénégal : la noctuelle de la tomate (Helicoverpa armigera) et la teigne du chou (Plutella xylostella). Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 20. Téréta I, Togola M, Brévault T, Renou A & Goebel FR (2015) Un espoir pour la diffusion d'une pratique écologique de protection des cotonniers au Mali. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). 21. Carrière Y, Welch K, Brévault T & Tabashnik BE (2014) Understanding evolution of resistance to pyramided Bt crops in Helicoverpa zea. In : Entomological Society of America. 2014-11-16/2014- 11-19, Portland (Etats-Unis). 22. De Bon H, Brévault T, Arodokoun DY, Diallo MD, Diarra K, Hamadoun A, Menozzi P, Sall D, Sinzogan A, Tamo M & Vayssières JF (2014) DIVECOSYS: A scientific partnership Platform for ecologically-based pest management in West Africa. Poster. Final report for the international symposium on agroecology for food security and nutrition, 2014-09-18/2014-09-19, Rome (Italie). 23. Diarra K & Brévault T (2014) Seasonal abundance of major cabbage insect pests and their natural enemies in Senegal. In : INRA ; UM2. 17ème Colloque de Biologie de l'Insecte. 17, 2013-10- 07/2013-10-09, Montpellier (France). 24. Brévault T, Ngom D, Soti V, Sow A & Thiaw C (2013) Régulation écologique des insectes ravageurs des cultures vivrières: de la parcelle au paysage. Atelier Divecosys sur l'écologie du paysage au service de la gestion des bio-agresseurs des cultures en Afrique de l'ouest, 2013-12-03/2013-12- 06, Cotonou (Bénin). 25. Brévault T, Ndoye O, Diatte M, Sylla S, Bernadas G & Gueye S, Diarra K (2013) Tuta absoluta, une nouvelle menace pour la production de tomate au Sénégal. In : CORAF/WECARD. Atelier de partage de l'information sur la propagation et de définition des moyens de lutte contre Tuta absoluta, 2013-05-07/2013-05-09, Dakar (Sénégal). 26. Clouvel P, Martin P, Brévault T, Renou A, Soti V, Birman D & Goebel FR (2013) Méthodes et concepts issus de l'écologie du paysage. Séminaire Réseau PIC INRA-CIRAD, 2013-02-04/2013-02- 06, Paris (France). 27. Diarra K & Brévault T (2013) Formation et recherche en agroécologie horticole au Sénégal : exemples de deux études de cas portant sur les ravageurs du chou (Plutella xylostella) et de la tomate (Tuta absoluta) dans la zone des Niayes. Réunion du Réseau africain de travail sur les rongeurs soudano-sahéliens, 2013-11-12/2013-11-16, Saint-Louis (Sénégal). 28. Thiaw C & Brévault T (2013) Renforcement de la régulation écologique des insectes ravageurs des céréales sèches et cultures associées (RECOR). Réunion du Réseau africain de travail sur les rongeurs soudano-sahéliens, 2013-11-12/2013-11-16, Saint-Louis (Sénégal). 29. Marchand M, Dormont L, Schatz B, Téréta I, Renou A & Brévault T (2012) Topping-mediated VOCs emitted by cotton plants: Electroantennogram response of two major noctuid pests. Poster. 6th International courses in Chemical Ecology "Frontiers of Chemical Ecology 2012" (ICE-12), 2012-11- 26/2012-12-07, Iena (Allemagne). 30. Brévault T (2011) Interventions sur seuils en culture cotonnière : exemples de détermination et d'application de seuils. Atelier FSP-Coton, 2011-02-17, Cotonou (Bénin). 31. Brévault T, Nibouche S, Achaleke J & Carrière Y (2011) Can natural refuges delay insect resistance to Bt cotton. In : Venugopalan M.V. (ed.), Balasubramanya R.H. (ed.), Kranthi Sandhya (ed.), Blaise (ed.). World Cotton Research Conference-5, Mumbai, 7-11 November 2011. World Cotton Research Conference. 5, 2011-11-07/2011-11-12, Mumbai (Inde). 32. Brévault T, Clouvel P & Martin P (2011) Spatial ecology remote sensing and GIS spatially-explicit models area-wide IPM : Towards a multi-scale approach for improving insect pest management. In : CIRAD. Atelier CIRAD Quels outils pour un changement d'échelle dans la gestion des insectes d'intérêt économique ? Workshop "Towards a Multi-Scale approach for Improving Pest Management", 2011-10-04/2011-10-05, Montpellier (France).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 18
33. Silvie P, Adegnika MO, Akantetou AK, Ayeva B, Bonni G, Brévault T, Gautier C, Hema OSA, Houndété TA & Togola M (2011) Cotton pest management programmes using threshold-based interventions developed by CIRAD and its partners in sub-Saharan african countries. In : Venugopalan M.V. (ed.), Balasubramanya R.H. (ed.), Kranthi Sandhya (ed.), Blaise (ed.). World Cotton Research Conference. 5, 2011-11-07/2011-11-12, Mumbai (Inde). 34. Bertrand A, Brévault T, Thézé M & Vaissayre M (2010) De la LEC à la LOIC : comment aider les paysans à prendre en charge la protection phytosanitaire de leurs parcelles de coton ? In : Seiny Boukar L. (ed.), Boumard Philippe (ed.). Savanes africaines en développement : Innover pour durer, 2009-04-20/2009-04-23, Garoua (Cameroun). 35. Quilici S, Brévault T, Rousse P, Hurtrel B, Duyck PF, Jacquard C, Delatte H, Deguine JP, Lereculeur A, Wattier C, Atiama-Nurbel T, Franck A, Simiand C, Chiroleu F (2010) Tritrophic interaction in the complexes of fruit flies damaging fruit and vegetable crops in Reunion island. Poster. International Symposium on Fruit Flies of Economic Importance. 8, 2010-09-26/2010-10-01, Valence (Espagne). 36. Sougnabé SP, Yandia A, Acheleke J, Brévault T, Vaissayre M & Ngartoubam LT (2010) Pratiques phytosanitaires paysannes dans les savanes d'Afrique centrale. In : Seiny Boukar L. (ed.), Boumard Philippe (ed.). Savanes africaines en développement : Innover pour durer, 2009-04-20/2009-04- 23, Garoua (Cameroun). 37. Brévault T, Carletto J & Vanlerberghe-Masutti F (2008) Host specificity and insecticide resistance in the cotton aphid, Aphis gossypii. World Cotton Research Conference. 4, 2007-09-10/2007-09- 14, Lubbock (Etats-Unis). 38. Brévault T, Achaleke J & Sougnabé N (2008) Tracking seasonal dynamics of pyrethroid resistance in African populations of Helicoverpa armigera. World Cotton Research Conference. 4, 2007-09- 10/2007-09-14, Lubbock (Etats-Unis). 39. Aboubakary A, Mathieu B, Beyo J, Woin N, Brévault T & Ratnadass A (2007) Protection insecticide du sorgho repiqué (Muskuwaari) contre les dégâts de foreurs des tiges (Sesamia cretica) au Nord- Cameroun. Séminaire régional sur l'agroécologie et les techniques innovantes dans les systèmes de production cotonniers : 24 au 28 septembre 2007, Maroua (Cameroun). 40. Brévault T (2007) Mieux comprendre le système de vie des ravageurs du coton, pour mieux les maîtriser... Journées d'Animation Scientifique des UPRs Coton, 2007-07-10/2007-07-12, Montpellier (France). 41. Naudin K, Bikay S, Maldes JM & Brévault T (2007) Impact of direct seeding and mulch on soil marofauna in cotton field in North Cameroon. FFEM Séminaire International Les sols tropicaux en semis direct sous couvertures végétales, 2007-12-03/2007-12-08, Antananarivo (Madagascar). 42. Brévault T, Carletto J, Picault S & Vanlerberghe-Masutti F (2006) Cotton races in Aphis gossypii evidenced by microsatellite markers and life history traits. Proceedings of the Beltwide Cotton Conference 2006, 2006-01-03/2006-01-06, San Antonio (Etats-Unis). 43. Aboubakary A, Ratnadass A, Mathieu B, Brévault T & Woin N (2005) Insecticidal protection of transplanted sorghum (Muskuwaari) from stem borer (Sesamia spp) damage in Northern Cameroon. Conférence Internationale sur les Ravageurs en Agriculture. 7, 2005-10-26/2005-10- 27, Montpellier (France). 44. Achaleke J, Brévault T, Blondin L & Vassal JM (2005) Interbreeding in Helicoverpa armigera populations from different host plants estimated by resistance status and microsatellite markers = Conférence Internationale sur les Ravageurs en Agriculture. 7, 2005-10-26/2005-10-27, Montpellier (France). 45. Achaleke J, Brévault T, Blondin L, Vassal JM (2005) Helicoverpa armigera. Truly polyphagous : Interbreeding in Helicoverpa armigera populations from different host plants estimated by resistance status and microsatellite markers. In : AFPP. 7ème Conférence Internationale sur les Ravageurs en Agriculture. 7, 2005-10-26/2005-10-27, Montpellier (France). 46. Brévault T (2005) Epidémiologie de la résistance chez H. armigera et stratégies de gestion. Mécanismes et stratégies de gestion de la résistance des insectes d'intérêt agricole aux insecticides en Afrique et en Asie : Atelier GeRICo - CFC/ICAC. Ouagadougou (Burkina Faso).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 19
47. Brévault T (2005) Surveillance de la sensibilité aux insecticides chez les piqueurs suceurs Aphis gossypii et Bemisia tabaci. Mécanismes et stratégies de gestion de la résistance des insectes d'intérêt agricole aux insecticides en Afrique et en Asie. Atelier GeRICo - CFC/ICAC. 014, 2004-12- 06/2004-12-10, Ouagadougou (Burkina Faso). 48. Brévault T, Bikay S & Naudin K (2005) Macrofauna pattern in conventional and direct seeding mulch-based cropping systems in North Cameroon. In : FAO. 3rd World Congress on Conservation Agriculture : Linking Production, Livelihoods and Conservation, 2005-10-03/2005-10-07, Nairobi (Kenya). 49. Brévault T, Carletto J, Linderme D, Vanlerberghe-Masutti F (2005) Pucerons du coton . Enquête d'identité : diversité génétique des populations d'A. gossypii en culture cotonnière au Nord Cameroun. In : AFPP. Conférence Internationale sur les Ravageurs en Agriculture. 7, 2005-10- 26/2005-10-27, Montpellier (France). 50. Brévault T & Naudin K (2005) Factors affecting cotton seedling in mulch-based cropping systems in North Cameroon. In : FAO. 3rd World Congress on Conservation Agriculture : Linking Production, Livelihoods and Conservation, 2005-10-03/2005-10-07, Nairobi (Kenya). 51. Brévault T, Carletto J, Linderme D & Vanlerberghe-Masutti F (2005) Pucerons du coton : enquête d'identité = Cotton aphids: Identity survey. Conférence Internationale sur les Ravageurs en Agriculture. 7, 2005-10-26/2005-10-27, Montpellier (France). 52. Lançon J, Wery J, Rapidel B, Gaborel C, Gérardeaux E, Ballo D, Brévault T & Fadegnon B (2004) Prototyping integrated cotton crop management systems for specific ecological and socioeconomic constraints in Western Africa. In : Jacobsen Sven-Erik (ed.), Jensen Christian Richardt (ed.), Porter John R. (ed.). Congress of the European Society for Agronomy. 8, 2004-07- 11/2004-07-15, Copenhagen (Danemark). 53. Brévault T, Beyo J, Nibouche S & Vaissayre M (2003) La résistance des insectes aux insecticides : problématique et enjeux en Afrique centrale = Insect resistance to insecticides in central Africa. In : Jamin Jean-Yves (ed.), Seiny-Boukar Lamine (ed.), Floret Christian (ed.). Savanes africaines : des espaces en mutation, des acteurs face à de nouveaux défis, 2002-05-27/2002-05-31, Garoua (Cameroun). 54. Lançon J, Wery J, Rapidel B, Angokaye M, Ballo D, Brévault T, Cao TV, Deguine JP, Dugué P, Fadegnon B, Fok M, Gaborel C, Gérardeaux E, Klassou C & Yattara A (2003) Prototyping crop management systems for specific cotton growing conditions. In : Swanepoel A. (ed.). World Cotton Research Conference. 3, 2003-03-09/2003-03-13, Cape Town (Afrique du Sud). 55. Nibouche S, Beyo J, Brévault T, Crétenet M, Gozé E, Jallas E, Martin P & Moussa AA (2003) Utilisation du modèle Cotons R - Simbad pour la définition de seuils d'intervention contre les chenilles de la capsule du cotonnier. In : Jamin Jean-Yves (ed.), Seiny-Boukar Lamine (ed.), Floret Christian (ed.). Savanes africaines : des espaces en mutation, des acteurs face à de nouveaux défis, 2002-05-27/2002-05-31, Garoua (Cameroun). 56. Beyo J, Brévault T, Nibouche S & Asfom P (2003) Suivi de la sensibilité de la chenille des capsules du cotonnier Helicoverpa armigera aux insecticides pyréthrinoïdes au Nord Cameroun. In : Jamin Jean-Yves (ed.), Seiny-Boukar Lamine (ed.), Floret Christian (ed.). Colloque Savanes africaines : des espaces en mutation, des acteurs face à de nouveaux défis, 2002-05-27/2002-05-31, Garoua (Cameroun). 57. Beyo J, Brévault T, Nibouche S, Asfom P (2002) Sensibilité aux insecticides pyréthrinoïdes chez Helicoverpa armigera au Nord Cameroun. In : Brévault Thierry (ed.), Nibouche Samuel (ed.). Résistance des insectes aux insecticides en Afrique de l'ouest et du centre. Atelier sur la résistance des insectes aux insecticides en Afrique de l'ouest et du centre, 2002-03-06/2002-03-07, Maroua (Cameroun). 58. Nibouche S, Beyo J, Brévault T, Crétenet M, Gozé E., Jallas E, Martin P & Moussa AA (2002) Cotons - Simbad : a tool for establishing cotton bollworm economic damage thresholds. In : Congress of the European Society for Agronomy. 7, 2002-07-15/2002-07-18, Cordoba (Espagne).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 20
59. Quilici S, Brévault T & Hurtrel B (2000) Major research achievements in Reunion within the Indian Ocean regional fruit fly programme. Indian Ocean Commission Regional Fruit Fly Symposium, 5-9 juin 2000, Flic-en-Flac (Mauritius). 60. Brévault T & Quilici S (1999) Orientation responses of the tomato fruit fly, Neoceratitis cyanescens, to visual and olfactory stimuli during host finding. In : Rendon P. (ed.), De Galan A. (ed.). 3rd meeting of the working group on fruit flies of the western hemisphere. Guatemala City (Guatémala). 61. Brévault T & Quilici S (1999) Basic patterns of mating activities in the tomato fruit fly, Neoceratitis cyanescens. In : Final FAO/IAEA Research Co-Ordination Meeting on "Medfly Mating Behaviour Studies under Field Cage Conditions", 1999-06-29/1999-07-03, Antigua (Guatémala). 62. Brévault T & Quilici S (1998) Factors affecting behavioral responses of tomato fruit fly females to visual stimuli. Fifth international symposium on fruit flies of economic importance. International Symposium on Fruit Flies of Economic Importance. 5, 1998-06-01/1998-06-05, Penang (Malaisie).
1.6.6. Rapports d’expertise
2016 - HCB. Avis relatif à une demande d'autorisation de mise sur le marché du maïs génétiquement modifié 1507 x 59122 pour la culture, l'importation, la transformation, et l'alimentation humaine et animale. http://www.hautconseildesbiotechnologies.fr/fr/avis/avis-relatif-a- demande-dautorisation-mise-sur-marche-mais-genetiquement-modifie-1507-x-59122 - HCB. Avis relatif à une demande d'autorisation de mise sur le marché du maïs génétiquement modifié Bt11 x 59122 x MIR604 x 1507 x GA21 pour l'importation, la transformation, et l'alimentation humaine et animale.
2015 - HCB. Avis en réponse à la saisine du 24 avril 2014 de Messieurs Bernard Accoyer et Jean Bizet suite à la p opositio de loi elati e à l’i te di tio de la ise e ultu e des a i t s de maïs génétiquement modifié sur le territoire français. http://www.hautconseildesbiotechnologies.fr/fr/avis/avis-reponse-a-saisine-24-avril-2014-messieurs-bernard-accoyer- jean-bizet-suite-a-proposition - HCB. A is elatif à u e de a de d’auto isatio de ise su le a h du soja g ti ue e t odifi MON à des fi s d’i po tatio , de t a sfo atio , et d’ali e tatio hu ai e et a i ale. http://www.hautconseildesbiotechnologies.fr/fr/avis/avis-relatif-a- demande-dautorisation-mise-sur-marche-soja-genetiquement-modifie-mon-87751 - PASE II. Missio d’appui au olet e he he-développement du CNRA. Intensification du a le des s st es d’e ploitation. Stratégies de protection intégrée de la culture oto i e. A al se des a uis de la e he he et p opositio d’a tio s. - 2013 - PIP-COLEACP. R isio et a tualisatio de l’iti ai e te h i ue de culture de la tomate dans les pays de la CEDEAO.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 21
1.7. Synthèse bibliométrique Publications indexées dans ISI Web of Knowledge depuis 10 ans (2007-2016).
Figure 1. Nombre de publications par année (à gauche) et nombre de citations par année (à droite). Données actualisées en avril 2017 à partir de la bse ISI Web of Knowledge (h-index : 13).
Tableau 1. Pu li atio s da s des jou au à fa teu d’i pa t IF . IF Premier ou Nom du journal Catégorie Publications (5 ans) dernier auteur Nature Biotechnology Biotechnology 41.39 1 PNAS Multidisciplinary sciences 10.29 1 1 Molecular Ecology Ecology 6.23 1 Scientific Reports Multidisciplinary sciences 5.53 1 1 Evolutionary Applications Evolutionary biology 4.63 1 1 PloS One Multidisciplinary sciences 3.54 1 Soil & Tillage Research Soil science 3.37 1 1 Oecologia Ecology 3.36 1 Pest Management Science Entomology 3.12 4 3 Journal of Chemical Ecology Ecology 3.05 1 1 Tropical Medicine & International Health Public and environmental health 2.80 1 Ecological Entomology Entomology 2.10 1 1 Agricultural and Forest Entomology Entomology 1.90 1 1 Journal of Economic Entomology Entomology 1.85 1 1 Bulletin of Entomological Research Entomology 1.81 3 3 Crop Protection Agronomy 1.79 5 4 Entomologia Experimentalis Et Applicata Entomology 1.67 3 3 Plant Breeding Agronomy 1.63 1 Environmental Entomology Entomology 1.60 2 1 Physiological Entomology Entomology 1.40 2 2 Experimental Agriculture Agronomy 1.29 2 2 Fruits Horticulture 1.05 1 Journal of Insect Behavior Entomology 1.05 1 1 International Journal of Pest Management Entomology 1.01 1 African Entomology Entomology 0.65 1 1 International Journal of Tropical Insect Science Entomology 0.52 1 Egyptian Journal of Biological Pest Control Entomology 0.23 1 TOTAL 41 28
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 22
2. Bilan des activités de recherche
2.1. Contexte
Les i se tes a ageu s des ultu es ep se te t u e o t ai te ajeu e à l’i te sifi atio de la production agricole, particulièrement dans les pays du Sud, où ils constituent une menace permanente pour la sécurité alimentaire. Si le recours aux pesticides a permis de lever partiellement cette contrainte dans les filières les plus encadrées (e.g. le coton) ou à haute valeur ajoutée (e.g. l’horticulture), elle menace la viabilité de ces systèmes de p odu tio pa la pe te d’effi a it de o o e de ol ules suite à l’appa itio de populations résistantes, et par une rupture des équilibres biologiques, avec pour conséquence des épisodes plus fréquents de pullulation. De même, la simplification des paysages agricoles et la fragmentation ou la suppression des habitats naturels, dues en pa ti ulie à l’a oissement de la pression anthropique sur les ressources, le changement climatique, ou encore les invasions biologiques, induisent de nouvelles perturbations qui participent à la modification des interactions entre les espèces. Toutes ces pressions environnementales ont des conséquences sur les processus écologiques et évolutifs au sein des agroécosystèmes, u’il o ie t de ieu o p e d e pou e p di e ou g e l’ olutio , et d’i t g e da s la o eptio de s st es de p odu tio ag i ole innovants, capables de répondre aux enjeux majeurs que sont l’aug e tatio de la production et la préservation de la biodiversité.
2.2. Problématique et approche
La plupart des organismes et des populations doivent répondre à des perturbations de leur environnement, menaçant parfois leur existence. Leur capacité à répondre de manière phénotypique et génotypique à ces défis et à développer des mécanismes d'adaptation est donc cruciale (Fig. 2). Les activités de recherche que j’ai pu o dui e, depuis le début de ma carrière, o t po t p i ipale e t su les p o essus d’adaptatio des populatio s d’i se tes ravageurs, en réponse à une pression de sélection imposée par les insecticides appliqués sur les cultures, ou par les cultures transgéniques produisant leurs propres molécules i se ti ides. L’o je tif de es e he hes ise à ieu anticiper et gérer les conséquences de ces pratiques agricoles sur les agroécosystèmes. Plus généralement, les leçons tirées de l'étude de ces processus d’adaptatio doi e t guider la conception de systèmes de protection des cultures plus durables. Les utilisateurs des produits de cette recherche sont les agriculteurs des pays du Sud, ais aussi l’e se le des a teu s o e s pa la durabilité des systèmes de production agricole et la pérennité des filières agricoles concernées.
Figure 2. M a is es d’adaptatio des o ga is es au st ess imposés par les perturbations environnementales. Adapté de Bijlsma & Loeschcke (2005).
La sista e d’u e populatio i se tes peut être définie
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 23
comme une diminution héréditaire de la sensibilité à un insecticide (Tabashnik et al. 2009). Une très faible p opo tio d’i di idus da s u e populatio d’i se tes i les poss de atu elle e t u all le de sista e de l’o d e de pou ou oi s . C’est lorsque cet allèle, par le biais de la sélection des individus porteurs augmente en fréquence dans la populatio au fil des g atio s, u’o pa le d’ olutio de la sista e. Il s'agit donc d'une adaptation au nouvel environnement créé par la présence des insecticides. Les individus résistants sont porteurs d'une ou plusieurs mutations génétiques codant pour des protéines qui interagissent avec l'insecticide. Ainsi, les protéines mutées empêchent l'i se ti ide d’attei d e sa i le, pa e e ple e le d g adant, ou en modifiant cette cible, pe etta t au i se tes po teu s de es utatio s de su i e à des doses d’i se ti ide normalement létales. Le développement de la résistance dans les populations d'insectes est un phénomène évolutif basé sur les mécanismes de la sélection naturelle. Ce n'est pas l'utilisation d'insecticides qui crée la résistance, mais elle sélectionne les individus porteurs d'allèles de résistance (apparus par mutation spontanée ou par migration) qui survivent au traitement et dont la descendance sera avantagée. Si la pression insecticide est maintenue sur plusieurs générations, alors la fréquence des individus porteurs d'allèles de résistance augmente progressivement à chaque génération. La apa it d’adaptatio des populations d’i se tes ravageurs aux insecticides représente donc un enjeu majeur pour la santé des plantes.
Mes a ti it s de e he he o t po t p i ipale e t su les fa teu s d’ olutio de la résistance aux insecticides (ou aux plantes transgéniques insecticides) chez les populations de deux ravageurs clés, selon deux grands axes d’a al se :
les bases génétiques de la résistance : fréquence initiale des allèles de résistance (variabilité génétique), nombre de gènes impliqués, dominance (avantage sélectif procuré en présence de traitements) et le coût associé aux gènes de résistance (Carrière et al. 2010). Elles d te i e t la itesse d’appa itio de la sista e. Lo s u’u all le conférant une résistance apparaît dans une population, il est initialement au stade hétérozygote. Sa probabilité de se fixer dans la population va dépendre en partie de son deg de do i a e. Si le oût de la sista e est le , l’all le de sista e a i dui e un désavantage sélectif en absence d’i se ti ides, entraînant son élimination des populations polymorphes. Ceci est important pour prédire la trajectoire évolutive des allèles de résistance aux pesticides parce que le coût de la résistance peut parfois être assez le et o t e ala e l’a a tage su la fit ess e p se e de l’age t s le tif, empêchant ainsi sa fixation dans la population.
le système de vie du a ageu , d fi i o e l’e se le des l e ts de l’ os st e ui d te i e t l’e iste e, l’a o da e et l’ olutio de ses populatio s (Kennedy & Storer 2000). Ces éléments peuvent être regroupés en quatre grandes catégories : les t aits de ie de l’i di idu à la populatio ode de ep odu tio , st at gie d og aphi ue, te ps de g atio , o ilit , ga e d’hôtes, etc.), et pour la population étudiée, les caractéristiques biotiques (disponibilité et qualité des ressources da s le te ps et l’espa e, e e is atu els, etc.) et abiotiques (climat, etc.) du milieu. Ils d te i e t le deg d’e positio de la populatio d’i se tes aux insecticides, dans l'espace et dans le temps. Les insectes qui ont un temps de développement court, de nombreuses générations par an et une forte prolificité (stratégie r) deviennent plus
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 24
rapidement résistants que les insectes au développement long, avec un nombre limité de générations et de descendants (stratégie k). Le degré d'isolement des populations est aussi un facteur important de l'évolution de la résistance, les flux géniques entre populations favorisant la diffusion des gènes de résistance. Le mode de reproduction est égale e t u fa teu i po ta t i flua t da s l’ olutio de la sista e, a il est à la base de la structure génétique des populations.
Pou o p e d e l’ olutio de la sista e, il faut aussi prendre en compte l'historique des traitements insecticides qui peuvent avoir présélectionné des gènes de résistance, de l'insecticide utilisé, de la dose, de la fréquence et de la distribution spatiale des applications, de sa rémanence (plus un insecticide est rémanent plus le nombre de générations d'insectes soumis à la pression sélective sera élevé), et de l'existence d'autres sources de pressions sélectives (insecticides appli u s su d’aut es ultu es hôtes). Ai si, lo s u’u e sista e apparaît dans une population, son évolution est déterminée par de nombreux facteurs d’o d e g ti ue, iologi ue, ologi ue et op atio els. Mieu o aît e es fa teu s et leu effet peut pe ett e de p oi l’ olutio de la sista e da s les populatio s des ravageurs concernés et de concevoir des stratégies permettant de retarder son évolution.
Encadré 2. Les grandes fo es oluti es i pli u es da s l’adaptatio
Quatre forces évolutives agissent sur la structure génétique des populations au fil des générations : la mutation, la sélection naturelle, la dérive génétique et la migration.
La mutation a un rôle primordial car elle permet des modifications dans les génomes par utatio s po tuelles, i se tio s, d l tio s, ou dupli atio s. C’est la seule fo e oluti e qui produit de nouveaux allèles, générant de la a ia ilit g ti ue. Lo s u’u e utatio o f e u ph ot pe a a tageu à l’i di idu ui la po te, la s le tio atu elle est la fo e qui le favorise dans la population. La sélection naturelle peut ainsi avantager (sélection positive) ou désavantage gati e u all le, odifie la aleu d’u a a t e phénotypique (directionnelle), réduire sa variance (stabilisante) ou favoriser la coexistence de plusieurs phénotypes distincts pour un même caractère (disruptive). En modifiant la distribution des caractères phénotypiques et des allèles associés, la sélection naturelle est la fo e oluti e à l’o igi e de l’adaptatio .
La dérive génétique est un élément stochastique qui peut influencer de façon non dirigée et al atoi e les f ue es all li ues d’une génération à une autre. Elle est particulièrement élevée dans les populations à faibles effectifs. Il est important de ne pas négliger les effets aléatoires de la dérive génétique, car ils peuvent dans certain cas porter la même signature génétique que la sélection naturelle.
La migration est un des facteurs susceptibles d’a oi le plus d’i pa t da s l’ olutio de la résistance dans les populations. Elle favorise la dispersion de la résistance à grande échelle, mais peut aussi a oi l’effet i e se à une échelle locale. En effet, elle peut contrecarrer les changements de fréquence allélique causés par la sélection en réintroduisant des allèles se si les, li ita t ai si les possi ilit s d’adaptatio lo ale, s’il e iste des populatio s « source » d’i se tes sensibles qui se maintiennent dans des zones non traitées. La migration dépend de la capacité de dispersion ainsi que des préférences écologiques des ravageurs concernés.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 25
Mes a ti it s de e he he su les p o essus d’adaptatio des populatio s d’i sectes ont porté sur la diversité génétique et la résistance aux insecticides chez le puceron du coton, su l’ pidémiologie, les bases génétiques et les mécanismes de la résistance des populations de chenilles de la capsule aux pyréthrinoïdes en Afrique centrale, et sur les stratégies de gestion de la résistance, notamment dans le cas des cultures de coton transgénique Bt.
2.3. Systèmes biologiques
U e des lefs de l’ la o atio du e de e t et de la ualit de la p odu tio oto i e en zone de savanes africaines est la protection contre les insectes ravageurs de la culture. A la fin des années 1990, les premiers cas de résistance de chenilles de la capsule et du puceron du coton ont été signalés en Afrique de l’Ouest Martin et al. 2000, Nibouche et al. 2002). Mes recherches ont porté principalement sur ces deux ravageurs clés de la culture cotonnière dans les agroécosystèmes de savanes africaines :
la noctuelle du coton, Helicoverpa armigera (Hübner) (Lepidoptera, Noctuidae), dont les larves, appelées chenilles de la capsule, s’atta uent aux organes fructifères du cotonnier, p o o ua t l’a s issio des outo s et jeu es apsules. H. armigera est une espèce cosmopolite et polyphage largement répandue dans les régions tropicales, subtropicales et tempérées du globe. La présence de ce ravageur a été récemment observée en Amérique du Sud, dans les cultures de soja au Brésil (Czepak et al. 2013). L’esp e a u e importance grandissante e Eu ope Liste A de l’OEPP , ota e t e F a e où elle migre en été depuis le bassin méditerranéen, infligeant des dégâts en cultures de maïs doux et haricots dans le sud-ouest. Da s les zo es oto i es d’Af i ue de l’Ouest, la colonisation des parcelles de coton s’effe tue g ale e t au d ut de la flo aiso des cotonniers (50 à 60 jours après la levée), à partir de populations adultes en provenance de plantes-hôtes locales (e.g. maïs ou adventices comme Cleome spp.) ou issues de migrations à plus longue distance (Fig. 3). Après deux à trois générations sur le coton, les populations du ravageur se reportent en début de saison sèche sur d’aut es plantes cultivées dans les périmètres irrigués (tomate, maïs) ou spontanées comme Hyptis suaveolens, ta dis u’u e p opo tio e o e al connue entre en diapause ou migre vers des zones plus favorables (Brévault et al. 2008). Des larves ont été retrouvées sur 217 espèces de plantes appartenant à 50 familles (Nibouche 1999).
A B E
C D
Figure 3. A) Papillon femelle, Helicoverpa armigera ; B) chenille dans une capsule de coton ; C) chenille sur une tomate ; D) paysage agricole ; E) système de vie (Photos T. Brévault).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 26
le puceron du coton, Aphis gossypii Glover (Hemiptera, Aphididae), qui, en prélevant la sève pour se nourrir, affaiblit les jeunes cotonniers et, en fin de cycle, altère la qualité de la fi e pa des d pôts de iellat au o e t de l’ou e tu e des apsules (Fig. 4). A. gossypii est une espèce cosmopolite et polyphage largement répandue dans les régions tropicales, subtropicales et tempérées du globe. Sa reproduction est caractérisée par une parthénogénèse apomictique asso i e à u e f o dit le e de l’o d e de descendants par femelle) et à une durée de développement larvaire très courte (de l’o d e d’u e se ai e . La olo isatio d’u oto ie , ui peut t e t s p o e, dès le stade plantule, se fait par des ailés. Deux périodes de croissance de la population sur les cultures de coton sont généralement observées: la première au début du cycle de culture après une période sans pluie et la seconde à la fin de la saison des pluies avant la récolte. En saison sèche, les populations du ravageur se concentrent sur les plantes maraîchères des périmètres irrigués. Plus de 658 plantes-hôtes, appartenant à 420 genres et 103 familles, ont été recensées en Afrique sub-saharienne (Deguine et al. 1997).
A B E
C D
Figure 4. A) Pucerons (formes ailés, adultes et larves aptères), Aphis gossypii ; B) jeune feuille de cotonnier gaufrée suite aux piqûres de pucerons ; C) feuille de poivron infestée ; D) fibre de coton souillée par le miellat et le développement de fumagine ; E) système de vie (Photos T. Brévault).
Ces deux insectes ravageurs sont responsables de pertes importantes de production, non seulement en culture cotonnière pendant la saison des pluies, mais aussi en cultures maraîchères en saison sèche (Renou & Brévault, 2016). La noctuelle du coton s’atta ue à une grande variété de plantes cultivées en Afrique, en particulier le coton, le maïs et la tomate. Le puceron du coton est considéré comme un ravageur clé des cucurbitacées (melon, concombre, courgettes, pastèque), malvacées (coton, gombo, hibiscus), et solanacées cultivées (pomme de terre, piment, poivron, aubergine). Ces deux insectes ravageurs sont également réputés pour une forte aptitude à développer des résistances aux insecticides, avec 763 cas répertoriés dans le monde pour 49 matières actives, et 268 cas pour 50 matières actives, pour H. armigera et A. gossypii, respectivement (APRD 2016).
2.4. Diversité génétique et résistance aux insecticides chez le puceron du coton
Les agroécosystèmes à base de cultures annuelles sont caractérisés par une importante hétérogénéité dans l'espace et dans le temps. La disponibilité des ressources variant de façon importante tout au lo g de l'a e et d’u e a e su l’aut e, la dynamique des populations des ravageurs dans un paysage agricole est dictée par la capacité des insectes à se disperser et à exploiter différents habitats. Il est généralement admis que les insectes ont
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 27
tendance à se spécialiser sur des ressources prévisibles, c'est-à-dire abondantes et faciles à trouver dans le temps, alors que l'imprévisibilité de l'apparition de l'hôte devrait favoriser la polyphagie (Jaenike 1990). En réalité, les espèces polyphages tendent à se spécialiser au niveau de la population (Futuyma & Moreno 1988, Thompson 1994). En présence de ressources substantielles mais limitées dans le temps, ce qui est le cas des cultures annuelles, ces populations sont capables de se déplacer sur de grandes distances, tandis que d'autres se déplacent localement selon une séquence de plantes-hôtes (Loxdale & Lushai 1999). Ces populatio s doi e t gale e t s’adapte à des essou es g ale e t soumises à des traitements insecticides.
En Afrique soudano-sahélienne, les champs de coton occupent un espace important du paysage agricole pendant la saison de pluies (juin à novembre) et abritent une grande diversité d'insectes ravageurs dont les populations sont en partie contrôlées par l'utilisation d'insecticides (Renou & Deguine 1992). La saison sèche implique un changement rapide de ce paysage agricole, avec une pénurie soudaine de ressources pour les insectes ph tophages. Seuls les petites p i t es i igu s o ti ue t d’off i des essources et jouent probablement un rôle clé dans les systèmes de vie de ces ravageurs. Le puceron du coton, A. gossypii, en raison d'un taux d’a oisse e t élevé et de sa capacité à se disperser par des morphes ailés, a le potentiel de coloniser rapidement divers agroécosystèmes. De plus, sa démographie exponentielle associée à un mode de reproduction parthénogénétique favorise la sélection rapide de mécanismes de résistance aux insecticides (Ai et al. 2003). Des cas sévères de résistance au diméthoate (famille des organophosphorés) ont été signalés dès 1996 au Cameroun (Deguine 1996).
E olla o atio a e l’Inra (UMR Isa, Sophia Antipolis), nous avons réalisé un échantillonnage pour évaluer la diversité génétique des populations de pucerons du coton dans le paysage agricole du Nord Cameroun, en relation avec la disponibilité des plantes- hôtes et les traitements insecticides (Brévault et al. 2008). Celui- i s’i s i ait da s u e étude plus large de la structuration génétique des populations du puceron A. gossypii (Carletto et al. 2009). Les pucerons ont été collectés en saison des pluies sur des parcelles de coton traitées et non traitées et en saison sèche sur des plantes de la famille des Malvaceae, Cucurbitaceae ou Solanaceae, cultivées dans des petits périmètres irrigués, ainsi que sur des plantes spontanées. Des marqueurs microsatellites ont été utilisés pour caractériser les individus échantillonnés. Leur carte de résistance aux insecticides a été estimée par des bioessais et par le diagnostic moléculaire de deux mutations dans le gène ace -1 associées à une insensibilité de l'acétylcholinestérase aux carbamates et aux insecticides organophosphorés, complétés par des tests enzymatiques.
2.4.1. Diversité génétique et spécialisation écologique
Considéré comme une espèce hautement polyphage, le puceron du coton (aussi appelé puceron du melon) est en fait composé de populations sympatriques génétiquement différenciées. Un échantillonnage mondial de pucerons a démontré une structuration génétique des populations par les plantes-hôtes (Carletto et al. 2009). De plus, une expérience de transfert de clones sur différentes plantes-hôtes a révélé l'existence de fortes différences de performance selon la plante-hôte testée (Carletto et al. 2009). Cinq races
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 28
d’hôtes dominées par des clones asexués (cucurbitacées, coton, pomme de terre, aubergine, et piment ou poivron) ont été identifiées sans ambiguïté (Fig. 5). Ces différentes races d’hôte ont pu évoluer à partir de populations ancestrales généralistes selon un processus de radiation adaptative. La faible variabilité génétique au sein de chaque race d’hôte à l'échelle mondiale pourrait plaider en faveur d'un événement de spécialisation récent. Cette faible diversité pourrait également être la conséquence d'une compétition pour la ressource entre différents clones spécialisés sur les mêmes plantes. Néanmoins, cette analyse concerne presque exclusivement des pucerons échantillonnés sur des cultures. Il est possible qu'un niveau élevé de variabilité génétique existe in natura, ce qui pourrait représenter un énorme potentiel d'adaptation pour cette espèce.
Figure 5. A gauche : échantillonnage du puceron Aphis gossypii dans différentes régions du monde. Le nom des génotypes détectés chez les pucerons collectés sur les différentes plantes est indiqué lorsque leur fréquence (entre parenthèses) était supérieure à 0,20. A droite : arbre de distribution des 44 génotypes multilocus montrant un regroupement selon les plantes-hôtes.
La capacité du puceron du coton à se déplacer et à coloniser différents habitats en réponse à la variation spatio-temporelle des ressources, lui permet d'exploiter des paysages agricoles instables. L'analyse de la diversité génétique de pucerons collectés dans le nord du Cameroun en parcelles de coton et en cultures maraîchères pendant deux années consécutives, a révélé une très faible diversité génétique, avec seulement 11 génotypes multilocus identifiés (Brévault et al. 2008). L’a al se de la p o i it génétique de ces génotypes a o t l’o u e e de i lignées (Fig. 5). Les pucerons collectés sur des cucurbitacées pendant la saison sèche appartenaient à la lignée génétique C, a e d’hôte spécialisée sur les cucurbitacées (Vanlerberghe-Masutti et al. 1999, Charaabi et al. 2008). La majorité des pucerons collectés sur des solanacées cultivées, aubergine et morelle noire, appartenaient au génotype Auber, tandis que ceux collectés sur poivron présentaient plutôt le génotype PsP4 (Brévault et al. 2008). Les lignées Burk et Ivo ont été collectées pendant la saison des pluies dans les champs de coton, et sur deux autres malvacées (gombo et oseille de Gui e pe da t la saiso s he. Elles appa tie e t p o a le e t à u e a e d’hôte spécialisée sur les plantes de la famille des malvacées (Brévault et al. 2008). Le génotype Burk1 caractérise la plupart des pucerons des cultures de coton collectés dans différents pa s d'Af i ue de l’Ouest, de Madagas a et du B sil, alo s u’Ivo a été détecté uniquement en Afrique sur les cultures de coton (Carletto et al. 2009). Le génotype Burk1 a également
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 29
été retrouvé en saison sèche en faible abondance, sur des plantes spontanées appartenant aux Capparidaceae, Convolvulaceae ou Euphorbiaceae. Les ultu es de go o et d’oseille de Guinée (genre Hibiscus) constituent probablement des réservoirs pour ces génotypes pendant la saison sèche (Brévault et al. 2008). Les aut es a es d’hôte se d eloppa t aussi sur Hibiscus, il est possible que ce soit un hôte ancestral à partir duquel la spécialisation s’est faite su d’aut es pla tes.
2.4.2. Carte de résistance aux insecticides
Pour évaluer la résistance aux insecticides parmi les races d’hôtes identifiées au Cameroun, des bioessais biologiques ont été réalisés. Des tests de diagnostic moléculaire (PCR-RFLP) et enzymatique ont été également conduits pour identifier les mécanismes impliqués dans la résistance aux insecticides (Carletto et al. 2010). Les six clones testés (Auber, Burk1, Ivo, PsP4, C4 et C9) sont tous sensibles à l'acétamipride (néonicotinoïde) et au carbosulfan (carbamate). Tous les clones portent la mutation S431F dans le gène de l'acétylcholinestérase1 et sont résistants au diméthoate (organophosphoré). Auber, PsP4 et Burk1 portent également la mutation A302S, ce qui leur confère une résistance modérée aux organophosphorés (Carletto et al. 2010). Auber et Burk1 sont très résistants à la cyperméthrine, insecticide de la famille des pyréthrinoïdes. Cette résistance est probablement associée à la mutation ponctuelle super-kdr (M918L) dans le gène du canal de sodium (gène para) ou à la détoxification enzymatique par des estérases et des oxydases. Une résistance multiple à une large gamme d'insecticides et des mécanismes de résistance multiples, en particulier pour le génotype Burk1, peuvent expliquer dans une certaine mesure la faible diversité génétique observée chez A. gossypii.
2.4.3. Sélection par les traitements insecticides et compétition
Les traitements insecticides ont un effet sur la diversité génétique des populations d’A. gossypii. Un changement de fréquence des génotypes en faveur de Burk1 a été observé lors du processus de colonisation des parcelles de coton, entre l'infestation initiale et le pic d'infestation au début de la saison, en particulier lorsque du diméthoate (insecticide organophosphoré) était appliqué (Brévault et al. 2008). Dans les parcelles traitées avec le diméthoate, toutes les plantes initialement infestées par Ivo ont été fortement colonisées par Burk1.
Des expériences sur des plantes entières au laboratoire et dans des cages en plein champ ont permis de tester l’effet des traitements insecticides, de la compétition inter-clonale, de l'adaptation à la plante-hôte ou des conditions climatiques (températures élevées et l'humidité relative faible en saison sèche), sur la diversité génétique du puceron dans les parcelles de coton (Brévault et al. 2011). Les deux génotypes Ivo et Burk1 ont une performance équivalente lorsque transférés séparément sur des plants de coton (Gossypium
1 L'insensibilité de l'acétylcholinestérase (AChE) des souches hautement résistantes de A. gossypii aux carbamates et aux organophosphorés résulte de deux mutations ponctuelles sur le gène ace 1 : la première changeant le codon de sérine en position 431 en un codon de phénylalanine (S431F) et la seconde changeant le codon d'alanine en position 302 en un codon de serine (A302S) (Andrews et al. 2004, Toda et al. 2004).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 30
hirsutum ou G. arboreum , de go o ou d’oseille de Gui e. Lo s ue t a sf s simultanément, Ivo se comporte mieux que Burk1, suggérant l’e iste e p o a le de oûts liés à la résistance qui pourraient affecter la performance des pucerons Burk1 en situation de compétition. En revanche, Burk1 se développe mieux en présence de traitements insecticides. Dans les expériences où les pucerons ont été autorisés à se déplacer vers les plantes voisines, Burk1 était mieux représenté que Ivo sur les plantes peu peuplées, ce qui indique que la dispersion peut être un moyen d'éviter la compétition sur les plantes très infestées. La plupart des pucerons de coton recueillis sur des plantes-hôtes de coton ou de relais au début de la saison de culture étaient Burk1 (> 90%), ce qui suggère une grande capacité de dispersion et reflète probablement une fréquence élevée sur les plantes-hôtes relais (Brévault et al. 2011).
2.4.4. Implications pour la gestion des pucerons en culture cotonnière
Dans le paysage agricole de sa a es d’Af i ue de l’Ouest, l'utilisation d'insecticides à large spectre sur le coton et les plantes-hôtes relais a conduit à la prévalence d'un génotype résistant à différentes classes d'insecticides, et possédant des traits caractéristiques d’esp es envahissantes. La résistance aux insecticides combinée à un mode de reproduction clonal pourrait expliquer la faible diversité génétique observée et une prévalence du génotype résistant Burk1 dans les parcelles de coton. La persistance du génotype sensible Ivo pourrait être maintenue g â e à l’hétérogénéité spatiale et temporelle du paysage agricole, incluant des cultures-hôtes non traitées et certaines plantes spontanées dans des habitats non cultivés.
L’ide tifi atio de a es d’hôtes hez A. gossypii, de plantes-réservoirs (gombo et oseille de Guinée), et d’u e a te de se si ilit au i se ti ides, do e des éléments essentiels pour repenser les stratégies de gestion des pucerons en Afrique sub-saharienne. Nous avons o fi ue la p essio e e e pa l’utilisatio d’i se ti ides à la ge spe t e e ultu e cotonnière et maraîchère et la compétition inter-clonale pouvaient expliquer la faible diversité génétique des populations du puceron et la prévale e d’u génotype multi- résistant. Le développement de stratégies de gestion intégrée basées sur la restauration du rôle des ennemis naturels a été entériné par la filière coton au Cameroun, notamment par l’utilisatio d’i se ti ides spécifiques. En outre, les modalités de contrôle du puceron du coton en saison sèche, dans les périmètres maraîchers, doit se faire en concertation avec les acteurs de la filière coton. La disponibilité et la distribution des cultures de gombo et d’oseille de Gui e doit être prise en compte dans les stratégies de gestion intégrée du puceron du coton, o e pla tes elais e i te saiso à l’ helle du pa sage ag i ole, ou comme plantes piège ou banque (Shelton & Badenes-Perez 2006) pour le contrôle biologique des pucerons dans les champs de coton.
Publications majeures 1. Brévault T, Carletto J, Linderme D & Vanlerberghe-Masutti F (2008) Genetic diversity of the cotton aphid Aphis gossypii in the unstable environment of a cotton growing area. Agricultural and Forest Entomology 10: 215–223.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 31
2. Carletto J, Lombaert E, Chavigny P, Brévault T, Lapchin L & Vanlerberghe-Masutti F (2009) Ecological specialization of the aphid Aphis gossypii Glover on cultivated host plants. Molecular Ecology 18: 2198–2212. 3. Carletto J, Martin T, Vanlerberghe-Masutti F & Brévault T (2010) Insecticide resistance traits differ among and within host races in Aphis gossypii. Pest Management Science 66: 301–307. 4. Brévault T, Carletto J, Tribot J & Vanlerberghe-Masutti F (2011) Insecticide use and competition shape the genetic diversity of the aphid Aphis gossypii in a cotton-growing landscape. Bulletin of Entomological Research 101: 407–413.
2.5. Traits de la résistance aux pyréthrinoïdes chez la noctuelle du coton
De pa leu i pa t su la p odu tio ag i ole et leu apa it d’adaptatio à des pe tu atio s i sta ilit ou a fa tio des essou es à l’ helle du pa sage ag i ole, fragmentation des habitats naturels, traitements insecticides, etc.), les lépidoptères phytophages représentent un enjeu majeur pour la gestion de la santé des plantes. La capacité à développer une résistance aux insecticides a placé la noctuelle du coton, H. armigera, comme l'un des ravageurs les plus importants en Afrique, en Asie, en Europe du Sud et en Australie. L'utilisation à grande échelle d'insecticides de la famille des pyréthrinoïdes pour contrôler ce ravageur a plus ou moins rapidement conduit à une résistance généralisée dans plusieurs régions de production de coton dans le monde. En Afrique, la résistance aux pyréthrinoïdes a été diagnostiquée en 1996 en Afrique du Sud (van Jaarsveld et al. et à l’Ouest du o ti e t Vassal et al. , Martin et al. 2000). Dans les zones de savanes africaines, le coton occupe parfois une partie importante du paysage ag i ole. L'appli atio g alis e d’i se ti ides de la fa ille des p th i oïdes, efficaces et peu coûteux, exerce une forte pression de sélection pendant le cycle de culture en saison des pluies. En outre, l'utilisation des mêmes insecticides sur les cultures maraîchères des petits périmètres irrigués en saison sèche, en particulier de tomate, contribue probablement à la s le tio d’i di idus po teu s de g es de sista e.
2.5.1. Epidémiologie de la résistance
Au Cameroun, les insecticides de la famille des pyréthrinoïdes ont été largement utilisés sur le coton pendant environ 20 ans, en raison de leur efficacité pour le contrôle d'une large gamme de ravageurs. Le programme de lutte recommandé a été conçu pour fournir aux petits producteurs de coton un programme de traitements systématiques basé sur un calendrier. Environ six traitements insecticides (dont quatre avec des pyréthrinoïdes) sont appliqués au cours du cycle de culture, à intervalle de deux semaines à partir de 45 jours après la levée des cotonniers (premiers boutons floraux).
Un réseau de surveillance a été mis en place en 1999, a e l’aide de la Sod oto et du projet Prasac-Ardesac, pour la détection et le suivi épidémiologique (étude de la distribution et des déterminants de la résistance) de la résistance aux pyréthrinoïdes chez la noctuelle du coton (Brévault et al. 2002). Une première étude a été conduite de 2002 à 2006 pour évaluer le niveau de la résistance aux pyréthrinoïdes des populations de la noctuelle, à une échelle régionale. De même, un suivi de sa dynamique à une échelle locale a été réalisé, à partir de la collecte de larves sur une séquence de différentes plantes-hôtes comprenant des plantes
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 32
cultivées et sauvages. Sur la période étudiée, le niveau global de résistance aux pyréthrinoïdes des populatio s d’H. armigera, bien que variable entre sites, a régulièrement augmenté, pour atteindre une situation inquiétante en 2005 (Brévault et al. 2008) (Fig. 6). Plusieurs échecs de contrôle observés dans les champs des agriculteurs au cours de la campagne 2004, étaient imputables à la résistance (Brévault & Achaleke 2005). Des tests o duits à pa ti de su des populatio s d’H. armigera du Tchad et du Nigéria ont montré que la résistance concernait l'ensemble de la zone cotonnière de l'Afrique centrale.
Figure 6. A gauche : Localisation des zones d'échantillonnage de populations d'Helicoverpa armigera. A droite : Survie (%) des larves collectées sur coton au début (août) et à la fin (octobre) de la période d'infestation et exposées à une dose diagnostic (30 µg) de cyperméthrine. Les barres indiquent une valeur maximale et les nombres entre parenthèses se rapportent au nombre total de tests effectués en août et octobre, respectivement.
Pourquoi la résistance a-t-elle augmenté si rapidement sans changements significatifs dans le paysage agricole ou dans la nature des insecticides utilisés ? L'h poth se d’u e diffusio des allèles de résistance depuis des populations d'Afrique de l'Ouest, où la résistance aux pyréthrinoïdes avait été confirmée plus tôt (Martin et al. 2000), ne semble pas être valide, un mécanisme différent de résistance ayant été identifié en Afrique centrale (Achaleke et al. 2009). Cette augmentation de la résistance pourrait s'expliquer par différents facteurs non exclusifs : (i) le respect des recommandations phytosanitaires réduisant la proportion de refuges non traités, (ii) la reconstitution des populations en début de saison des pluies à partir de populations résiduelles se maintenant sur les cultures maraîchères ou en diapause localement (peu ou pas de migration), et (iii) le renforcement de la pression de sélection par l'utilisation généralisée de pyréthrinoïdes sur des plantes-hôtes cultivées dans les zones d’ igration.
La surveillance de la résistance à l'échelle locale sur une séquence de plantes-hôtes a montré que la résistance peut être utilisée comme marqueur pour évaluer la spécialisation des populations d’H. armigera sur certaines plantes-hôtes (Fig. 7). La similitude entre la fréquence de résistance des larves recueillies simultanément à partir d'hôtes synchrones traités et non traités, tels que coton-maïs (août), coton-Hyptis (octobre) et tomate-maïs (décembre et février), confirme l'hypothèse de l'absence d'isolement reproductif entre les populations échantillonnées sur différentes plantes-hôtes (Achaleke et al. 2005). Le niveau de résistance au début d'une saison de coton (août, année n) était très similaire à celui obtenu à la fin de la saison précédente du coton (octobre, année n-1). Ces résultats indiquent que les traitements insecticides en culture cotonnière jouent probablement un rôle clé dans l'augmentation saisonnière et interannuelle de la résistance, même si
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 33
l'utilisation intensive d'insecticides sur les cultures locales de tomate participe aussi à la s le tio d’all les de sista e da s les populatio s siduelles au ou s de la saiso s he (Brévault et al. 2008). En outre, on note u e a se e d’aug e tatio saiso i e de la sista e à u e helle lo ale e et , suite à l’e lusio de la cyperméthrine pe da t le pi d’i festatio des populatio s d’H. armigera (Brévault et al. 2008).
Figure 7. Survie (%) des larves d’Helicoverpa armigera collectées sur une séquence de plantes-hôtes et exposées à une dose diagnostic (30 µg) de cyperméthrine (dans un rayon de 25 km autour de Garoua, Cameroun). Les barres indiquent que les valeurs maximales/minimales et les nombres entre parenthèses indiquent le nombre de tests (, Cleome ; , maïs ;, coton; , Hyptis; , tomate).
2.5.2. Mécanismes de résistance
Des bioessais (dose-réponse) ont été effectués sur des populations de larves collectées à l’ helle gio ale su les principales plantes-hôtes (coton et tomate). Des techniques biochimiques ont été utilisées pour identifier les mécanismes impliqués dans la résistance aux pyréthrinoïdes. La résistance au p th i oïdes o se e hez les populatio s d’H. armigera en Afrique centrale dès 2004 était principalement associée à la détoxification par une augmentation de la p odu tio d’est ases (Achaleke et al. 2009). La connaissance des mécanismes impliqués dans la résistance et l’a se e de résistance croisée au spinosad et à l'indoxacarbe est un élément clé pour concevoir de nouvelles stratégies de gestion de la résistance. En Afrique de l'Ouest, les populations de H. armigera ont développé une résistance ta oli ue pa la su p odu tio d’o dases Ma ti et al. . Un réseau régional impliquant des structures de recherche et de développement, ainsi que des opérateurs de la filière, a été créé avec succès pour gérer la résistance (Martin et al. 2005). Ses actions se sont concentrées sur l'utilisation régionale, rationnelle et concertée des pesticides dans l'ensemble de la zone cotonnière, basée sur une plus grande concertation entre acteurs de la filière et sur un renforcement des capacités des producteurs.
2.5.3. Héritabilité et coût de la résistance
Une étude des bases génétiques de la résistance d’H. armigera aux pyréthrinoïdes a été entreprise afin de mieux comprendre l'évolution de la résistance. Une souche sensible a été croisée avec une souche résistante (facteur de résistance à la cyperméthrine = 474) sélectionnée à pa ti d’u e population collectée dans des parcelles de tomate au nord du Cameroun.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 34
Des croisements entre les individus de la souche sensible, de la souche résistante et des hybrides (F1) indiquent que la résistance est transmise comme un caractère dominant (DLD = 0,86) et conférée par un seul gène autosomique. La dominance de la résistance diminue à esu e ue la dose de pe th i e aug e te, sugg a t u’elle est i o pl te e t −1 dominante (DML = 0,73) à la dose la plus élevée testée μg AI.g larve, dose tuant 100% des individus de la souche sensible). Bien que la dominance n'ait pas été évaluée au champ, les résultats suggèrent que les allèles de résistance peuvent se propager rapidement dans les populations de terrain dans la mesure où les phénotypes rr et rs survivent aux traitements insecticides. Enfin, le niveau de résistance (DL50) de la descendance hybride F1 a diminué de manière significative sur cinq générations en l'absence d'exposition aux pyréthrinoïdes, indiquant un coût biologique important de la résistance.
2.5.4. Implications pour la gestion de la résistance
L'a al se de l’ pid iologie, des a is es et des ases g ti ues de la résistance a apporté des indications essentielles pour concevoir une stratégie de gestion de la résistance des populatio s d’H. armigera aux pyréthrinoïdes. La sélection rapide des allèles de résistance liée à leur caractère dominant, et la diminution rapide de la résistance en l’a se e de p essio i se ti ide, suggèrent que l’exclusion temporelle des pyréthrinoïdes fe t es d’e lusio est un premier levier d’a tio pou o t e a e l’ olutio de la résistance. D’aut es esu es o e l'utilisation d’alte ati es au pyréthrinoïdes en saison sèche dans les cultures de tomate à l’ helle lo ale ou même régionale, et la mise en place de refuges naturels exempts d'insecticides (maïs en particulier) sont évoquées (Brévault et al. 2008).
Au Cameroun, u pla d’a tio basé sur l’e lusio des pyréthrinoïdes lorsque le coton est la seule plante-hôte disponible dans le paysage agricole (fenêtre du début septembre au début octobre), et leur remplacement par des insecticides ne présentant pas de résistance croisée (e.g. indoxacarbe), a été lancé dès 2006. Par ailleurs, un programme de traitements sur seuil développé sous fo e d’ oles pa sa es Lutte su o se atio i di iduelle des he illes, LOIC) a été mis en place (Fig. 8). Ce p og a e asse le aujou d’hui plus de producteurs qui utilisent de façon autonome une planchette de comptage pour décider de t aite ou o leu pa elle su la ase d’o se atio des he illes (Bertrand et al. 2009, Brévault et al. 2009).
Figure 8. Planchette de comptage pour la lutte sur observation individuelle des chenilles de la apsule LOIC au Ca e ou . L’ pi gle est a a e d’u t ou e s la d oite pou ha ue pla t o se et d’u t ou e s le haut pou chaque chenille comptabilisée. En cas de dépassement du seuil (zone rouge), un traitement à pleine dose est appliqué, ou pas de traitement dans le cas contraire (zone verte). La zone intermédiaire est u e zo e d’i d isio laissa t au planteur la décision de traiter ou non sa parcelle.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 35
2.5.5. Traitements insecticides en culture cotonnière et résistance des vecteurs
Durant mon séjour dans le Nord du Cameroun, j’ai eu l’oppo tu it d’ t e asso i à une tude o duite pa des oll gues de l’IRD et de l’O ea Laboratoire de Recherche sur le Paludisme, Organisation de Coordination pour la lutte contre les Endémies en Afrique Centrale), visant à explorer les possibles interactions entre traitements insecticides en culture cotonnière et résistance aux insecticides des populations de moustiques.
Des larves d’Anopheles gambiae ont été collectées en 2005 avant (mi-juin), pendant (mi- août) et à la fin (début octobre) du programme de traitements phytosanitaires du coton. Les larves ont été échantillonnées dans des sites de reproduction situés dans les champs de coton à Gaschiga et Pitoa, et à Garoua, une zone urbaine sans coton qui a servi de témoin (Chouaibou et al. 2009). Des tests de sensibilité aux insecticides ont été effectués avec différents insecticides. Une diminution significative de la sensibilité au DDT et aux pyréthrinoïdes, impliquant des mécanismes de détoxification, a été observée au cours de la campagne. E l’a se e d’effet di e t o e u e aug e tatio de la f ue e des individus portant une mutation Kdr, il semble que les perturbations environnementales causées par l'utilisation d'insecticides en agriculture, seuls ou en combinaison avec des polluants et des xénobiotiques naturels, permettent aux populations locales de moustiques de se constituer un arsenal enzymatique augmentant leur tolérance aux insecticides. L'interaction entre espaces agricoles et la santé humaine par la voie des vecteurs du paludisme, illustre l'étendue des effets non intentionnels de l'application de pesticides en milieu tropical. Cette étude renvoie à l'atelier organisé en 2011 : « Towards a multiscale approach for improving pest management » impliquant les deux communautés scientifiques travaillant l'une, sur la protection des cultures, et l'autre sur les maladies à transmission vectorielle (§ 1.5).
Publications majeures 1. Brévault T, Achaleke J, Sougnabé SP & Vaissayre M (2008) Tracking pyrethroid resistance in the polyphagous bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae), in the shifting landscape of a cotton-growing area. Bulletin of Entomological Research 98: 565–573. 2. Achaleke J, Martin T, Ghogomu RT, Vaissayre M & Brévault T (2009) Esterase-mediated resistance to pyrethroids in field populations of Helicoverpa armigera (Lepidoptera: Noctuidae) from Central Africa. Pest Management Science 65: 1147–1154. 3. Achaleke J & Brévault T (2010) Inheritance and stability of pyrethroid resistance in the cotton bollworm Helicoverpa armigera (Lepidoptera: Noctuidae) in Central Africa. Pest Management Science 66: 137–141. 4. Brévault T, Couston L, Bertrand A, Thézé M, Nibouche S & Vaissayre M (2009) Sequential pegboard to support small farmers in cotton pest control decision-making in Cameroon. Crop Protection 28: 968-973. 5. Chouaibou M, Etang J, Brévault T, Nwane P, Hinzoumbe CK, Mimpfoundi R & Simard F (2008) Dynamics of insecticide resistance in the malaria vector Anopheles gambiae s.l. from an area of extensive cotton cultivation in Northern Cameroon. Tropical Medicine & International Health 13: 476–486.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 36
2.6. Stratégies de gestion de la résistance au coton transgénique Bt
De à , j’ai t a ueilli o e he heu asso i au D pa te e t d’E to ologie de l’U i e sit d’A izo a UA , i te atio ale e t e o u pou so e pe tise su les impacts des cultures génétiquement modifiées résistantes aux insectes. Pour le Cirad, ce positionnement revêtait un intérêt en termes de montage de collaborations scientifiques et de t a sfe t de o p te es pou l’a o pag e e t scientifique et technique des filières cotonnières en Afrique, au o e t de l’i t odu tio o e iale du oto Bt au Burkina Faso. Le coton transgénique Bt produit une ou plusieurs toxines de Bacillus thuringiensis ciblant les lépidoptères ravageurs. Son adoption a permis dans certaines situations de dui e l’utilisatio d'insecticides, de participer à la suppression régionale de ravageurs, de favoriser l’a tio des ennemis naturels et d’a oît e les rendements ou de réduire leur variabilité (Tabashnik et al. 2013). E Af i ue de l’Ouest, son adoption devait permettre de réduire le recours aux traitements insecticides et de gérer la sista e d’H. armigera aux pyréthrinoïdes. Cependant, le développement de résistance aux toxines Bt au sein des populations des ravageurs ciblés, ou e o e l’i pa t su les o ga is es o i les, so t des préoccupations majeures pour la durabilité à long terme de ces cultures.
Pou l’UA, o e p ie e su la sista e au i se ti ides hez la noctuelle du coton H. armigera constituait une réelle opportunité pour le développement d’ tudes sur un modèle biologique voisin (H. zea) sur le uel ils ’a aie t pas d’e pe tise, mais qui faisait partie des ravageurs à risque dans la zone cotonnière du Sud-Est des Etats-Unis. J’ai o duit u e première étude sur la ise au poi t d’u e thodologie d’ aluatio de la contribution des refuges naturels à la gestion de la résistance de la noctuelle du coton, a a t l’i t odu tio du coton Bt à grande échelle sur un territoire donné. J’ai ensuite conduit deux études au laboratoire et e se e pou d te i e l’effet d’u e di i utio de l’e p essio de la to i e dans les plantes au cours de la saison ou d’u mélange de semences Bt et non Bt sur l’ olutio de la sista e.
2.6.1. Contribution des refuges naturels à l’évolutio de la résistance
La stratégie dite « refuges » a été largement adoptée pour retarder l'évolution de la résistance au coton Bt. L'idée qui sous-tend cette stratégie est que des individus sensibles (ss) produits par des plantes-hôtes ’e p i a t pas les to i es Bt, oto o -Bt (refuges dits structurés) ou plantes-hôtes alternatives cultivées ou spontanées (refuges dits naturels) se dispe se t et s’a ouple t a e de a es individus résistants (rr) issus des cultures Bt. Cette stratégie est particuliè e e t effi a e pou eta de l’ olutio de la résistance lorsque son héritabilité est récessive, car la descendance hétérozygote (rs) ne survit pas sur les cultures Bt. À l'inverse, si la résistance est dominante, la descendance survit sur les cultures Bt, et les refuges sont moins efficaces pour retarder la résistance (Tabashnik et al. 2008). Parce que la noctuelle du coton, H. armigera, est polyphage et très mobile, il est souvent admis que la p se e d’u e g a de di e sit de pla tes-hôtes alternatives (autres que le coton) dans les paysages ag i oles o stitue u e sou e d’i di idus se si les pe etta t de eta de l’ olutio de la sista e (Huang et al. 2010, Qiao et al. 2010). Cepe da t, s’il est elati e e t si ple d’ alue la p odu tio d’i di idus pa ces plantes-hôtes alternatives (e.g. Baker et al. 2008), il est plus difficile de mesurer les mouvements des papillons issus de ces refuges naturels vers les champs de coton Bt, tout au long du cycle de culture, et donc la
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 37
contribution effective de ces refuges naturels dans l'évolution de la résistance. Dans cette étude, nous avons capturé des papillons d’H. armigera aux abords de parcelles de coton, à l’aide de pi ges à ph o o e, dans trois paysages agricoles de la zone cotonnière ouest- africaine (Cameroun) où le coton Bt était absent. Nous avons utilisé des marqueurs biogéochimiques (isotopes de carbone comme signature de plantes-hôtes en C3 ou en C4, et gossypol comme signature du cotonnier comme plante-hôte) pour déterminer la contribution des refuges naturels comme source de papillons tout au long du cycle de culture du coton (Brévault et al. 2012)
L'utilisation de marqueurs biogéochimiques pour quantifier le mouvement des insectes constitue un outil précieux pour évaluer le rôle des refuges naturels pour retarder l'évolution de la sista e d’H. armigera au coton Bt. Pour les paysages étudiés, la contribution des plantes-hôtes alternatives comme source de papillons a été variable dans le temps et dans l’espa e, mais au moins équivalente à un refuge traité de 7,5% de coton non-Bt (Fig. 9).
Figure 9. (A-C) Pourcentage de papillons d’Helicoverpa armigera capturés dans des parcelles de coton issus de plantes-hôtes alternatives dans trois paysages agricoles (Guider, Djalingo et Tcholliré) au Cameroun. (D) Séquence de plantes-hôtes du ravageur dans la zone étudiée. Les courbes représentent l'occurrence temporelle et relative en surface des plantes hôtes.
Pou si ule l’ olutio de la sista e, ous avons pris comme exemple le coton Bt cultivé dans plusieurs régions du monde, dont le Burkina Faso, qui produit deux toxines, Cry1Ac et Cry2Ab, actives contre les lépidoptères phytophages (Showalter et al. 2009). Ces cotons à deux toxines ont le potentiel de retarder plus efficacement la résistance que des cotons à une seule toxine plantés séquentiellement ou en mosaïque (Roush 1998, Zhao et al. 2003). Notre modèle incorporant les paramètres biologiques de H. armigera indique que la présence de plantes-hôtes alte ati es da s les pa sages tudi s ’est pas suffisa te pou retarder sig ifi ati e e t l’ olutio de la résistance. En particulier, lorsque la concentration d'une toxine (ici Cry1Ac) diminue au cours du cycle dans la plante, la résistance au coton Bt survient rapidement dans les situations où les refuges de coton non Bt sont rares et où la résistance à Cry2Ab n'est pas récessive, car la résistance est essentiellement dépendante de la résistance à une toxine (ici Cry2Ab). Le déclin au cours du
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 38
cycle de la culture de la concentration d'une toxine (ici Cry1Ac) invalide l'une des hypothèses fondamentales de la stratégie dite « pyramide », à savoir que les insectes résistants à une to i e so t tu s pa l’aut e toxine (Carrière et al. 2010). Ainsi, la présence de refuges naturels en abondance dans le paysage agricole ’e t aî e pas nécessairement une production suffisante d’i di idus sensibles tout au long de la saison de culture du coton, pour retarder efficacement la résistance au coton Bt. Les simulations soulignent aussi l’i po ta e de mieux connaître les patrons de migration du ravageur, en début et en fin de saiso des pluies, o e l du assage g ti ue et de l’ olutio de la sista e.
2.6.2. Limites du coton Bt à deux toxines
Pour retarder l'évolution de la résistance des ravageurs aux cultures transgéniques produisant des protéines insecticides Bt, la stratégie pyramidale pe et d’o te i des plantes qui produisent deux toxines ou plus, actives pour le même ravageur. Une condition déterminant la performance des plantes à deux toxines est que les insectes résistants à une toxine soie t tu s pa l’aut e toxine. Sous cette hypothèse et elle d’u e sista e récessive, seuls les insectes homozygotes pour la résistance aux deux toxines peuvent survivre sur une plante à deux toxines. On s'attend donc à ce que de tels individus doublement résistants soient rares dans des populations qui n'ont pas été exposées préalablement à l'une ou l'autre des toxines.
Dans cette étude, nous avons constaté que cette condition ’ tait pas nécessairement validée car la sélection en laboratoire pour la résistance à Cry1Ac d'une souche de H. zea collectée au champ augmentait de manière significative la survie des larves sur le coton à deux toxines Cry1Ac et Cry2Ab par rapport à la souche non sélectionnée (Brévault et al. 2013). E out e, ous a o s o se u e di i utio sig ifi ati e de l’e p essio des to i es C A et C A au ou s du le du oto ie , a e pou effet, la pe te d’effi a it insecticide des cotonniers en fin de cycle (Fig. 10). Nous avons également trouvé des signes de résistance croisée entre les toxines Cry1A et Cry2A, à partir de os sultats et d’u e analyse des données provenant de 21 expériences de sélection.
Figure 10. Survie (IC 95%) des la es d’Helicoverpa zea d'une souche dérivée du champ (GA), d'une souche résistante à Cry1Ac (GA-R) et de leur descendance F1, exposées à des cotonniers non-Bt, Cry1Ac ou Cry1Ac et Cry2Ab.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 39
L'i o po atio de do es e pi i ues da s u od le d’ olutio de la sista e o t e que les écarts observés par rapport aux conditions idéales réduisent considérablement les avantages de la stratégie pyramidale pour des ravageurs comme H. zea qui ont une faible sensibilité aux toxines Bt (en particulier en fin de cycle avec la diminution des concentrations) et qui, dans certains cas, ont été largement exposés à l'une des toxines de la pyramide avant que les plantes à deux toxines ne soient adoptées.
Le défaut d’u redundant killing et l’e iste e d’u e sista e o essi e à C A , remettent en cause les recommandations de gestion de la résistance basées sur des conditions idéales : efficacité complète ou presque des deux toxines et résistance récessive. En outre, la gestion de la résistance aux cultures Bt chez les ravageurs peu sensibles aux toxines Bt pourrait être améliorée en prenant mieux en compte ce défaut, dû en particulier à la résistance croisée et à la diminution saisonnière de la concentration en toxines.
2.6.3. Des refuges dans les semences pour retarder la résistance
De o eu t a au de e he he o t o t ue l’i pla tatio de refuges peut retarder l'adaptation des insectes aux cultures Bt. Cependant, l'échelle spatiale optimale pour la plantation de ces refuges reste encore mal définie. Aux Etats-Unis, le règlement impose de planter les refuges de plantes non-Bt en blocs, sous forme de champs, de bandes ou de lignes séparés de la culture Bt. En 2010, les règlements ont été modifiés pour inclure des refuges plantés avec des mélanges de semences Bt et non Bt. Les mélanges de semences ont plusieurs avantages par rapport aux refuges en blocs, dont celui de contraindre les agriculteurs à respecter la mise en place de refuges. Toutefois, les résultats de modélisation suggèrent que lorsque les larves sont capables de se déplacer entre les plantes, les mélanges de semences pourraient accélérer l'évolution de la résistance en augmentant la dominance de la résistance, notamment par une survie des larves hétérozygotes supérieure à celle des larves homozygotes sensibles (Mallet & Porter 1993), avec des implications pour la durabilité des cultures Bt.
Nous avons conduit une s ie d’e p i e tatio s en serre pour comparer la dominance de la résistance au coton Bt produisant Cry1Ac (ci-après dénommé coton Bt) planté en culture pure (100%) par rapport à un mélange de plantes Bt et non Bt. Nous avons testé trois souches de H. zea: une souche dérivée du champ (GA) qui a été exposée aux toxines Bt uniquement sur le terrain, une souche résistante (GA-R) dérivée de GA et sélectionnée dans le laboratoire pour la résistance à Cry1Ac, et la descendance F1 obtenue à partir de croisements réciproques entre ces souches. Nous avons mesuré la survie de ces trois souches pour calculer la dominance (h), qui varie de 0 pour une résistance complètement récessive à 1 pour une résistance complètement dominante. La dominance de la résistance d’H. zea au coton Bt Cry1Ac a été significativement plus élevée dans un mélange de semences (h = 0,76) que dans la culture pure Bt (h = 0,48). La survie des larves a été significativement plus élevée pour GA-R que pour F1 dans le bloc de coton Bt, mais elle n'a pas varié de manière significative entre GA-R et F1 dans le mélange (Fig. 11). L'incorporation de ces données dans un modèle de génétique des populations suggère que l'augmentation observée de la dominance pourrait accélérer l'évolution de la résistance de 2 à 4,5 fois.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 40
Figure 11. Survie (IC 95%) des larves d’Helicoverpa zea dans des mélanges de cotonniers Bt (78%) avec des cotonniers non-Bt (22%) ou en culture pure (100%) avec trois souches : (GA) souche sensible dérivée du champ, (GA-R) souche résistante sélectionnée au laboratoire, et (F1) leur descendance. L'astérisque indique une survie de 0% de GA en présence de coton Bt uniquement.
Les résultats obtenus ont fourni la première preuve expérimentale que la dominance de la résistance aux ravageurs est plus élevée dans un mélange de semences Bt et non Bt que dans une culture pure de Bt (Brévault et al. 2015). Nous avons également montré que l’aug e tatio de la dominance provient de l'augmentation de la survie des individus hétérozygotes par rapport à celle des individus homozygotes sensibles, en présence de plantes non Bt. La preuve expérimentale et les simulations appuient les résultats de modélisation antérieurs indiquant que les mélanges de semences peuvent être moins efficaces que les refuges en blocs pour la gestion de la résistance aux cultures Bt chez des ravageurs mobiles tels que H. zea, pour lesquels les larves se déplacent généralement entre les plantes. En plus du mouvement des larves entre les plantes, une faible sensibilité intrinsèque aux toxines augmente les chances que la dominance de la résistance soit plus élevée dans les mélanges de plantes Bt et non Bt que dans les cultures Bt pures. Dans ce cas de figure, le projet de certains fournisseurs de semences de proposer aux agriculteurs un mélange de semences Bt et non Bt (refuge in a bag) pour éviter aux agriculteurs d’a oi à pla te u e pa elle efuge, ’est pas fo e t u e o e id e pou eta de l’ olutio de la résistance.
2.6.4. Résistance des insectes aux cultures Bt : leçons du terrain
Depuis , les ag i ulteu s du o de e tie o t se plus de illio s d’he ta es de maïs et de cotonniers génétiquement modifiés pour produire les protéines insecticides de la bactérie Bacillus thuringiensis (Bt). Ces plantes o t pe is de dui e l’utilisation des insecticides chimiques pour lutter contre les insectes cibles, mais elles ont aussi entraîné, o e le p oit la th o ie de l’ olutio , la s le tio d’i se tes sista ts. E jui , notre papier publié dans la revue Nature Biotechnology, rapportait que cinq espèces d’i se tes a aie t d elopp u e sista e à es p ot i es i se ti ides p oduites pa les plantes transgéniques (Tabashnik et al. 2013).
La t s g a de ajo it des populatio s d’i se tes i les ’o t pas d elopp de sistance aux plantes GM insecticides. On parle de cas de sista e, a l’ olutio d’u e sista e se f e à u e ou plusieu s populatio s d’i se tes, da s u o te te do . Elle e sig ifie i ue tous les i di idus de la populatio ou de l’esp e so t devenus résistants, ni que la situation est irréversible. La base de données Arthropod Pesticide Resistance
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 41
(http://www.pesticideresistance.org/search.php) répertorie plus de 3600 cas de résistance aux insecticides conventionnels pour les insectes de l’o d e des l pidopt es do t as de résistance à la seule deltaméthrine). Si l’aug e tatio des as de sista e s’e pli ue pa l’e te sio des supe fi ies ulti es e pla tes Bt, ui so t pass es de .1 à 66 millions d’he ta es e t e à , et donc par une exposition accrue des ravageurs aux toxines, on constate aussi des disparités dans la rapidité avec laquelle ces résistances sont apparues. Dans certains cas (maïs Cry1F/Spodoptera frugiperda, Puerto Rico), la résistance est survenue en trois a s, alo s ue da s d’aut es aïs C A /Ostrinia nubilalis, Etats-Unis), elle ’a toujours pas été détectée au bout de quinze ans (Fig. 12).
Figure 12. Evolution des surfaces plantées en cultures Bt à l'échelle mondiale et des cas de résistance au champ. La résistance a été rapportée pour cinq principaux ravageurs : Helicoverpa zea (2002), Spodoptera frugiperda (2006), Busseola fusca (2007), Pectinophora gossypiella (2008) et Diabrotica virgifera (2009). Quelques nouveaux cas ont été détectés depuis, chez D. virgifera (Gassmann et al. 2016), ou encore chez P. gossypiella pour le coton Cry1Ac/Cry2ab en Inde (Fabrick et al. 2015).
Nous avons analysé les conditions dans lesquelles cette résistance apparaît, mais surtout les facteurs qui retardent son apparition. Il est confirmé, tout d’a o d, u’en accord avec la th o ie de l’ olutio , l’effi a it des ultu es Bt a des ha es de du e si les all les de résistance sont initialement rares dans la population d’i se tes et si l’h ita ilit de ette résistance est récessive, ’est à-di e si seuls les i se tes ui poss de t deu opies de l’all le de résistance survivent sur les plantes Bt. L’ tude d o t e l’i t t des efuges se s à proximité des cultures Bt, avec des plantes qui ne produisent pas de toxines Bt. Dans ces efuges, les i se tes o sista ts peu e t su i e puis s’a ouple a e les a es i di idus qui ont survécu sur les cultures Bt. En cas de résistance récessive, leur descendance ne surviv a pas su la ultu e Bt : ’est u e so te de dilutio g ti ue de la sista e. E Australie, où la réglementation a été appliquée strictement, moins de 1% d’i di idus résistants ont été recensés dans les populations de H. armigera et de H. punctigera sur le coton Bt, alors que, dans le sud des États-Unis, où la réglementation était beaucoup moins o t aig a te, plus de % d’i di idus sista ts o t t d te t s pou e tai es populations de H. zea.
L’ tude soulig e toute l’i po ta e des esu es p ises e a o t pou p e i l’appa itio des résistances. La mise en culture sur de grandes surfaces de plantes transgéniques i se ti ides doit t e a o pag e d’u e politi ue de ise e pla e de refuges, dont l’i po ta e et la nature doivent être estimées en fonction des conditions locales (par e e ple la f ue e d’all les de sista e à la to i e Bt d jà p se ts da s les populatio s d’i se tes, la dose de to i e p oduite pa la ultu e Bt, et l’e iste e de efuges aturels -des plantes ou des cultures conventionnelles exemptes de toxines, de la même espèce ou d’aut es esp es- sur lesquels peuvent se développer des insectes sensibles) avant même
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 42
l’i pla tatio d’u e ultu e Bt. Pou eta de l’appa itio des sista ces, il est mentionné le recours à des plantes transgéniques exprimant deux toxines Bt. Mais là encore, certaines o es p ati ues s’i pose t pou o se e l’effi a it des toxines, par exemple éviter le cas où les plantes à deux toxines sont cultivées en e te ps ue des pla tes ’e p oduisa t u’u e seule des deu , ou lo s ue les i se tes o t d jà a uis u e sista e à l’u e. Si l’adaptatio des i se tes au ultu es Bt est i lu ta le, il est possi le de la eta de e etta t e œu e u e gestio intégrée des ravageurs, qui associe dispositifs de surveillance et mesures de gestio de la sista e des populatio s d’i se tes.
Publications majeures 1. Brévault T, Nibouche S, Achaleke J & Carrière Y (2012) Assessing the role of non-cotton refuges in delaying Helicoverpa armigera resistance to Bt cotton in West Africa. Evolutionary Applications 5: 53–65. 2. Brévault T, Heuberger S, Zhang M, Ellers-Kirk C, Ni X, Masson L, Li X, Tabashnik BE & Carrière Y (2013) Potential shortfall of pyramided transgenic cotton for insect resistance management. Proceedings of the National Academy of Sciences of the United States of America 110: 5806–5811. 3. Brévault T, Tabashnik BE & Carrière Y (2015) A seed mixture increases dominance of resistance to Bt cotton in Helicoverpa zea. Scientific Reports 5: 9807. 4. Tabashnik BE, Brévault T & Carrière Y (2013) Insect resistance to Bt crops: lessons from the first billion acres. Nature Biotechnology 31: 510–521.
2.7. Conclusion
L’adaptatio des populatio s d’i se tes a ageu s à l’appli atio d’i se ti ides ou à l’i t odu tio de a i t s ulti es sista tes da s l’ag o os st e, o stitue u e elle t od le pou l’ tude des p o essus micro-évolutifs et de sélection naturelle en action. L’app o he d’écologie évolutive et la composition de diff e ts outils d’a al se, de simulation ou de diagnostic, apportent des éléments pour une meilleure compréhension des p o essus à l’o igi e des évolutions observés dans les systèmes biologiques. Cette o p he sio des p o essus d’adaptatio off e des l s pou p di e l’ olutio des populations de ravageurs à une perturbation du milieu et concevoir des modes de gestion durable.
D’u poi t de ue appli u , les résultats de mes travaux de recherche ont participé à la conception de stratégies de gestion de la résistance. Ainsi, l’ide tifi atio d’u e a e d’hôte « coton » chez le puceron A. gossypii, de plantes-réservoirs (gombo et oseille de Guinée) et d’u e a te de se si ilit au i se ti ides, ont-ils été déterminants pour repenser les stratégies de gestion de ce puceron dans plusieurs pays de la zone cotonnière ouest- africaine. Au Cameroun, les traitements contre les pucerons ne sont déclenchés aujou d’hui u’à pa ti d’u seuil d’i te e tio et utilise t des insecticides sans résistance croisée et spécifiques, pe etta t de p se e l’a tio des e e is atu els. Les apports de connaissances sur l’ pid iologie, les mécanismes et les bases génétiques de la résistance aux pyréthrinoïdes chez la noctuelle, H. armigera, ont été essentiels pour concevoir une stratégie d’IRM Insect resistance management) appropriée en Afrique centrale. Au Cameroun, u pla d’a tio basé sur la substitution des pyréthrinoïdes par des insecticides plus sélectifs et un programme de traitements sur seuil développé sous fo e d’ oles paysannes ont été mis en place à partir de 2006.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 43
En termes de modélisation à caractère générique, les travaux sur l’adaptatio des populatio s d’i se tes a ageu s au oto génétiquement modifié ont permis de mettre au point une méthodologie pour évaluer l’i po ta e des refuges naturels, comme outil de gestion de l' olutio de la sista e d’i se tes a ageurs aux cultures Bt, avant même leur implantation. Trois articles majeurs abordent les limites des stratégies actuellement préconisées pour gérer la résistance, comme les plantes à plusieurs toxines (pyramides) ou le mélange de semences Bt à des semences non Bt. Le sujet est sensible, notamment dans les pa s de Sud, où l’o peut ai d e u e efficacité réduite des pyramides en présence d’u e aisse su sta tielle de l’e p essio des toxines par les plantes au cours du cycle cultural ou d’une diminution de la pureté variétale liée à la pollinisation croisée ou à des e eu s de a ipulatio au ou s du p o essus de p odu tio et d’utilisatio des se e es.
Dans les différentes a ti it s de e he he ue j’ai pu o dui e su l’adaptatio des populations de ravageurs aux insecticides ou aux variétés résistantes Bt, la migration est apparue comme un facteur déterminant dans les processus d’ olutio de la sista e. En effet, cette force évolutive peut participer à la dilution des allèles de résistance en réintroduisant des allèles sensibles dans une population, limitant ainsi les possibilités d’adaptatio lo ale. Ces flux de gènes dépendent de la présence dans le paysage agricole de ressources alternatives capables de supporter la production de suffisamment d’individus sensibles i.e. ’i posa t pas la e p essio de s le tio , mais aussi de la capacité de dispersion des individus. Pour le puceron du coton, la prise en compte de la distribution des ressources à l’ helle du pa sage et de la saison, est primordiale pour gérer son incidence en culture cotonnière. De même, la distribution spatiale et temporelle des parcelles de coton et des plantes-hôtes alternatives, participe à l’adaptatio des populations d’H. armigera en déterminant leur deg d’e position aux insecticides, selo u’elles jouent le rôle de refuge (e.g. les cultures non traitées de maïs) ou au contraire imposent la même pression sélective (e.g. les cultures de tomate, dans les périmètres maraîchers en saison sèche, traitées avec les es i se ti ides u’e ultu e oto i e).
Les questions liées à la contribution des refuges naturels ou à la mise en place de refuges structurés pour la gestion de la résistance aux cultures transgéniques insecticides, illustrent bien la nécessité de prendre en compte le paysage, comme i eau d’o ga isatio pour traiter des processus écologiques. Les flux liés à l’utilisatio de essou es distribuées dans l’espa e et da s le te ps, ui o stitue t u e osaï ue d’ha itats, d te i e t le fonctionnement écologique et la dynamique spatio-te po elle des populatio s d’i se tes. C’est da s ette app o he d a i ue des flu et de leu o s ue e su les populatio s de ioag esseu s, e lie a e la dispo i ilit des essou es, ue s’i s it le p ojet de recherche présenté ci-après.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 44
3. Projet de recherche. La biodiversité au service de la régulation des bioagresseurs
Le recours aux insecticides pour protéger les cultures contre les insectes ravageurs a montré ses limites, du fait de leur effet délétère sur l’environnement, la santé humaine et la biodiversité, et le développement de résistances dans les populations de bioagresseurs. Par ailleurs, la si plifi atio des pa sages sulta t de l’i te sifi atio ag i ole a o t i u à la fragmentation des habitats non cultivés, souvent supports de la biodiversité fonctionnelle. L’ osio de ette iodi e sit et la pe te des se i es de gulatio ui l’a o pag e, accroît la sensibilité des systèmes agricoles aux bioagresseurs. De plus, le changement climatique et les invasions biologiques induisent de nouvelles perturbations des écosystèmes qui affectent les interactions entre les espèces. Toutes ces perturbations ont des conséquences sur la biodiversité et sur les processus écologiques et évolutifs qui en dépendent. Il convient de mieux comprendre ces effets afin de gérer durablement les populations de bioagresseurs.
Dans les zones semi-arides tropicales de l'Afrique de l'Ouest, la réduction des précipitations et les changements dans l’utilisatio des terres (expansion des terres agricoles au détriment des espaces naturels, réduction de la diversité cultivée, déboisement, feux de brousse, déclin des systèmes mixtes d'élevage, etc.) ont entraîné une simplification des paysages agricoles (Agra 2014). Ces changements sont considérés comme des causes importantes du déclin de la biodiversité dans les paysages agricoles, la végétation semi-naturelle résiduelle étant le plus souvent concentrée dans les arbres d’i t t et les parcours pour les animaux. La façon dont ces évolutions ont pu affecter certains services écosystémiques tels que la régulation naturelle des bioagresseurs des cultures reste peu documentée. En Europe, la simplification des paysages agricoles favorise souvent les pullulations de bioagresseurs par la concentration des ressources autour de quelques espèces cultivées (Tscharntke et al. 2007), tandis que la fragmentation et la disparition d'habitats semi-naturels ont un impact négatif su les o u aut s d’e e is atu els et leur fonction de régulation (Chaplin-Kramer et al. 2011, Veres et al. 2013). La capacité à prévoir les changements du fonctionnement écologique des agroécosystèmes en réponse aux perturbations du milieu devient de plus en plus prégnante, avec la nécessité d'améliorer la productivité et la durabilité des systèmes de production agricole.
3.1. Concept de régulation écologique
La régulation écologique des bioagresseurs est un des services écosystémiques fournis par la biodiversité (Crowder & Jabbour 2014). Cette gulatio s’e e e via les ressources utilisées par le bioagresseur dans son habitat (régulation bottom-up), et via les ennemis naturels (régulation top-down) tels que prédateurs, parasitoïdes et pathogènes (Fig. 1). Au niveau du champ cultivé, les caractéristiques variétales et les pratiques culturales (e.g. variétés résistantes ou tolérantes, asso iatio d’esp es, fertilisation, etc.) constituent des leviers d’a tio o ilisa les pa l’ag i ulteu pour opposer des barrières physiques ou chimiques aux bioagresseurs ou stimuler les mécanismes de défense ou de compensation de la culture (Ratnadass et al. 2012). Au niveau du paysage agricole, l’a a ge e t spatial des ultu es peut retarder leur colonisation par les bioagresseurs, tandis que les espaces non cultivés
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 45
peu e t o stitue des ha itats pou diff e tes esp es d’e e is atu els et do servir de support à la régulation écologique des insectes ravageurs des cultures (Rusch et al. 2010). On parle alors de lutte biologique par conservation, qui vise à protéger et favoriser les populatio s d’e e is atu els présents dans les agroécosystèmes, grâce à l’a age e t des pratiques ou des habitats fournissant des ressources alimentaires, des proies ou des hôtes alternatifs, et des a is ou des sites d’hi e atio La dis et al. . D’aut es facteurs comme la compétition entre individus et les composantes abiotiques du milieu participent aussi aux processus de régulation des bioagresseurs (Fig. 13).
L’a ti atio des se i es de régulation écologique des bioagresseurs constitue une voie prometteuse pour réduire la dépendance des agricultures aux pesticides et inventer des modèles de gestion agroécologique des systèmes de protection des cultures. Son pilotage passe par une meilleure connaissance du fonctionnement des populations de bioagresseurs dans les agroécosystèmes (Gurr et al. 2003). Ceci suppose d’i t g e les traits de vie de l’i di idu à la populatio d og aphie, capacités de dispersion, etc. , l’a s au essou ces da s le te ps et l’espa e ultu es, végétation semi-naturelle, etc.), les interactions intra- et interspécifiques dans le système biologique étudié (compétition, relations trophiques, etc.), les perturbations induites par les pratiques culturales (e.g. les traitements insecticides) et les st u tu es spatiales ui d te i e t la atu e et l’i te sit des p o essus ologi ues considérés (e.g. les mouvements entre habitats).
Figure 13. La régulation écologique des insectes ravageurs des cultures est l'un des services écosystémiques (SE) fournis pa la iodi e sit . Elle peut s’e e e pa l’a tio des e e is atu els (top- down) ou par la qualité des ressources et leur disponibilité selon la composition et la structure du paysage (bottom-up). Certaines ressources sont aussi utilisées par les ennemis naturels qui peuvent y trouver des proies ou hôtes alternatifs et de la nourriture. Ce système multi- trophique est soumis à des perturbations liées aux interventions te h i ues de l’ag i ulteur et à des facteurs du milieu. Les flèches en pointillés représentent les interactions trophiques.
La iodi e sit fo tio elle ta t g tale u’a i ale i i les l e ts et les propriétés du paysage et les ennemis naturels) représente un facteur de résistance et de résilience de l’ os st e fa e au pe tu atio s e i o e e tales (Snyder & Tylianakis 2012). Conserver, activer ou restaurer sa fonction de régulation écologique des insectes ravageurs des ultu es est u d fi à ele e pou l’adaptatio des s st es de p odu tio ag i ole au changements environnementaux. L’ide tifi atio et la o ilisatio des le ie s d’a tio reposent sur une meilleure connaissance des conditions tant écologiques que sociales du milieu considéré. P odui e aut e e t suppose d’o se e et d’agi autrement.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 46
3.2. Observer autrement
L'emploi de « autrement » s'adresse ici au changement de posture de l'entomologie d'une approche circonscrite au modèle biologique, c'est-à-dire l'interaction stricte entre le ravageur et la culture sous l'effet des pratiques culturales, ou entre le ravageur et un ou plusieurs ennemis naturels, vers une approche des systèmes écologiques impliqués dans la régulation des populations de bioagresseurs. Pour aborder la complexité de ces systèmes, nous proposons d’o se e autrement les modèles biologiques qui nous intéressent, en adopta t ota e t les thodes et o epts issus de l’ ologie des o u aut s et de l’ ologie du pa sage. Ce changement nécessite une approche systémique permettant de (i) caractériser les interactions multi-trophiques entre et au sein des communautés du système biologique étudié et (ii) d’i t g e l’e se le des ha itats ulti s et o ulti s o e cadre d’a al se des p o essus ologi ues. Au-delà des aspects biophysiques, observer autrement implique aussi de (iii) prendre en compte la perception des acteurs du territoire dans une approche participative de la gestion collective des ressources et des processus d'innovation.
3.2.1. Relations entre biodiversité et régulation
La régulation écologique des bioagresseurs est un service écosystémique complexe généralement associé positivement à la richesse ou à la diversité des communautés d’ennemis naturels, mais pas toujours. Des effets positifs de la richesse spécifique sont observés lorsque les espèces agissent de manière complémentaire pour la régulation des bioagresseurs ou lorsqu'une ou plusieurs espèces facilitent la capture de proies par une autre, de sorte que l'effet combiné de plusieurs espèces dépasse la mortalité engendrée par une espèce seule (Crowder & Jabbour 2014). Des effets positifs peuvent également se produire lo s u’u e communauté plus riche en espèces assure une régulation malgré des perturbations, une ou plusieurs espèces ’ ta t pas affe t es pa ette pe tu atio . Pa ailleurs, les ennemis naturels peuvent exercer une protection de la plante en modifiant simplement le comportement des herbivores (e.g. comportement de fuite). A l’oppos , l'augmentation de la richesse spécifique peut avoir un effet négatif sur la régulation en présence de prédation intra-guilde ou d’i te f e es o po te e tales e t e e e is naturels. Dans une méta-analyse, Letourneau et al. (2009) ont montré que 71% des études impliquant des arthropodes ont mis en évidence des effets positifs de la richesse en ennemis naturels sur la régulation, avec des effets particulièrement forts dans les systèmes agricoles. Cependant, cette étude ne montre pas dans quelle mesure la richesse des communautés favorise le contrôle des bioagresseurs par complémentarité ou facilitation, ou par la plus grande probabilité de contenir une ou quelques espèces clés.
Les réseaux trophiques sont une description des communautés biologiques centrée sur les interactions entre les consommateurs et les ressources. La caractérisation des réseaux trophiques dans lesquels les ressources primaires, les bioagresseurs et les ennemis naturels interagissent est essentielle pour mieux comprendre les relations entre biodiversité et régulation écologique, et promouvoir le contrôle biologique des bioagresseurs. Il est donc essai e d’o se e d’aut es t aits de la iodi e sit , o e l’ uita ilit eveness), la diversité génétique ou la diversité fonctionnelle des communautés. Crowder et al. (2010) ont
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 47
montré que l'équitabilité des communautés de prédateurs et d'agents pathogènes améliorait le contrôle du doryphore Leptinotarsa decemlineata (Coleoptera, Chrysomelidae), en réduisant la compétition intra-spécifique. En outre, la diversité à différents niveaux trophiques doit aussi être prise en compte, les agroécosystèmes comprenant généralement plusieurs espèces de ravageurs ou de plantes hôtes potentielles, comme les mauvaises herbes associées aux cultures. Ainsi, Mollot et al. (2012) ont montré que la présence de ressources alternatives dans les bananeraies du fait de l’enherbement induit un changement de régime alimentaire des prédateurs généralistes, une augmentation de leur abondance et une plus forte prédation des œufs du ha a ço Cosmopolites sordidus (Coleoptera, Curculionidae). Encore peu d'études ont examiné les effets de la diversité à de multiples niveaux trophiques sur la régulation de bioagresseurs.
L'avènement des technologies de séquençage à haut débit (NGS) ouvre de nouvelles perspectives pour une identification rapide et peu coûteuse des espèces (barcoding) et la reconstitution des relations trophiques plante-herbivore, proie-prédateur ou hôte- parasitoïde (Symondson & Harwood 2014). Le séquençage de l'ADN dans les contenus i testi au fou it u i sta ta du gi e ali e tai e de p dateu s ou d’he i o es. Ces technologies ouvrent également la possibilité de détecter les endoparasitoïdes ou les hyperparasitoïdes, quel que soit leur stade dans les larves hôtes (Rougerie et al. 2011).
3.2.2. Le paysage comme adre d’étude
Le paysage est une représentation de systèmes écologiques en interaction (supérieur à l’ os st e où se d oule t et so t contrôlés un certain nombre de processus (Burel & Baudry 1999). Il se caractérise donc par son organisation, son hétérogénéité, sa diversité et sa d a i ue, sous l’i flue e d’a ti it s hu ai es ou de perturbations naturelles. Les flux d’o ga is es et de ati e au sei du paysage et les échanges entre systèmes écologiques so t o sid s o e des l s de o p he sio des p o essus ologi ues ui s’ d oule t, ta t au i eau de l’i di idu ue de la populatio . La o ple it et l’h t og it de ces structures spatiales et de leur dynamique, bien que reconnues, ont longtemps été gommées pa les ologues. La d a he adopt e e ologie du pa sage i t g e l’o jet d’ tude, le pa sage, ses d te i a ts, le ilieu et la so i t , et ses effets su les p o essus écologiques étudiés. On peut ainsi étudier la façon dont des plantes ou des animaux réagissent selon les propriétés locales du paysage, mais aussi comment les activités humaines ou les perturbations naturelles modifient ces propriétés. L’ ologie du pa sage s’i t esse au elatio s e t e les pat o s d’o ga isatio s spatiales et les p o essus écologiques, c'est-à-di e au auses et au o s ue es de l’h t og it spatiale. L’ide tifi atio de a is es de o t ôle des p o essus ologi ues tudi s pe et de prédire leur évolution et de d fi i des l e ts de gestio et d’a age e t de l’espa e nécessaires pour atteindre tel ou tel objectif, e.g. aug e te l’effi ie e de la gulatio écologique des bioagresseurs.
La composition et la configuration du paysage agricole, considéré comme une mosaïque d'habitats, déterminent le potentiel des ressources, mais aussi leur accessibilité (Mazzi & Dorn 2012). La fragmentation des habitats peut influencer les fonctions écologiques telles que les cascades trophiques ou la dispersion des insectes. La structure du paysage détermine
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 48
également la diversité, l'abondance et l'efficacité des ennemis naturels des bioagresseurs par la disponibilité de ressources (refuges, sources alimentaires, hôtes alternatifs, etc.) que les habitats cultivés et non cultivés peuvent fournir (Tscharntke et al. 2005). Généralement, la régulation des ravageurs augmente avec la complexité du paysage (Bianchi et al. 2006, Chaplin-Kramer et al. 2011).
Le paysage permet donc d’app he de o joi te e t les ilieu a th opis s et atu els et d’ lai e les p o essus de d isio da s la gestio des essou es iodi e sit , disponibilité ou la su eilla e de pullulatio s d’i se tes. Il appa aît i o tou a le si l’o eut s’i t esse à la du a ilit des s st es ologi ues et au p o essus ui s’ atta he t, comme la o se atio de la iodi e sit fo tio elle et des se i es de l’ os st e. L’utilisatio d’outils ode es d’aide au diag osti et à la prévision, comme la télédétection t aite e t d’i ages satellites), l’i age ie aérienne à l'aide de drones équipés de caméras à haute résolution, les s st es d’i fo atio g og aphi ue SIG et les plateformes de modélisation spatiale, permettent aujourd’hui de ieu app he de ces systèmes complexes.
3.2.3. La prise en compte du socio-agroécosystème
Notre approche s’i s it da s le ad e g al de l'ag o ologie, d fi ie o e l'appli atio de o epts et de p i ipes d’ ologie à la o eptio de systèmes agricoles durables (Altieri . Not e p i ipal o je tif est de o t i ue à l’ e ge e de s st es ag i oles résilients et productifs basés sur une mobilisation accrue des services de régulation écologique des bioagresseurs aux différentes é helles d’a tio , du ha p ulti au pa sage agricole. L’a itio des travaux de recherche sur les systèmes agricoles et la complexité des processus d'adoption de méthodes agroécologiques, conduit à interagir en amont du processus d’i o atio avec les acteurs des territoires concernés (Opdam et al. 2016). Les pratiques agricoles et les scénarii d’a age e t du paysage proposés e ue d’aug e te l’effi ie e du contrôle biologique des bioagresseurs, o ilise t l’a tio olle ti e, et à e titre doivent être co-conçus avec les acteurs. En outre, une approche intégrée de la gestion des bioagresseurs nécessite davantage de compétences, nécessite des systèmes de surveillance et de décision et entraîne des coûts opérationnels plus élevés que le seul usage d'i se ti ides. Passe d’u e gestio i di iduelle à u e gestio olle ti e et o e t e de la gestion des bioagresseurs requiert une vision systémique et l’i pli atio des a teu s, pour la conception de solutions durables empruntes des processus de changement inhérents à l’a ti it des so i t s u ales da s la du e.
La prise en compte du socio-agroécosystème requiert aussi l’i t g atio de différentes disciplines, en particulier des sciences humaines et sociales, pour explorer les logiques d’a tio des a teu s, la i ulatio de l’i fo atio , et les modes d'organisation collective de gestion du paysage (Bodin & Crona 2009). U e eilleu e o aissa e des seau d’a teu s doit pe ett e d’ide tifie les aiso s pou les uelles les i di idus s’e gage t à modifier leu s p ati ues, et les le ie s de l’a tio olle ti e.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 49
3.3. Agir autrement
Le contrôle des bioagresseurs peut être réalisé par une action sur le milieu, en perspective d’u e supp essio au niveau du champ cultivé ou du verger, par voie chimique (e.g. traitements insecticides), biotechnique (e.g. pièges à phéromones), biologique (e.g. lâchers de parasitoïdes) ou mécanique (e.g. travail du sol) (Brévault et al. 2014). Il peut également être envisagé sur une échelle plus large, en perspective d’u e gulatio des populations de bioagresseurs au niveau de la plantation ou du bassin de production, par la modification de la quantité et de la qualité des ressources (bottom-up) ou pa l’a age e t des ha itats au sein du paysage pour favoriser l’a tio des e e is atu els (top-down). La démarche que nous proposons pour une gestion écologique des bioagresseurs consiste à composer (i) l’a tio su le ilieu au t a e s des p ati ues et des i o atio s te h i ues respectueuses de la biodiversité fonctionnelle et ii l’a tio su la structure du paysage pour activer les processus écologiques impliqués dans les services de régulation des bioagresseurs. Cette démarche implique u ha ge e t fo da e tal des o tou s de l’a tio , d'une approche conventionnelle, souvent individuelle et indépendante de traitement des symptômes à l’ helle de la pa elle, à une organisation collective de la gestion de services écosystémiques incluant des dimensions écologiques, économiques et sociales (Cong et al. 2014). La o positio , la taille et la st u tu e des seau d’a teu s i pli u s da s la gestio des essou es atu elles à l’ helle du territoire (agriculteurs, organisations paysannes, décideurs politiques, etc.), et la mise en place de règles de gestion des services e i o e e tau , o stitue t le pilie de l’a tio . Notre vision du succès est que les petits producteurs passent de l’adoptio de pa uets te h ologi ues à l’adoptio de pa sages fonctionnels localement co-conçus.
L'agriculture peut contribuer à la préservation de la biodiversité fonctionnelle, qui en retour délivre des services écosystémiques tels que la régulation des bioagresseurs. L'amélioration de nos connaissances sur ces processus écologiques est essentielle pour adapter les pratiques agricoles. L'agriculture devra intégrer de nouvelles approches inspirées de la nature, afin de réduire la dépendance à l'égard des pesticides, et garantir la résistance et la résilience des systèmes de production face aux perturbations induites par les changements environnementaux.
3.4. Cas d’étude : La i euse de l’épi de il
3.4.1. Contexte
En Afrique subsaharienne, le mil est l'une des céréales les plus cultivées avec une production annuelle de plus de 10 millions de tonnes (FAO 2015), fournissant de la nourriture à plus de 60 millions de personnes. Parce que le mil peut se développer dans des environnements contraints (sécheresse et faible fertilité des sols) où d'autres cultures échouent généralement, il a un grand potentiel pour contribuer au défi de sécurité alimentaire dans les zones arides, comme o posa te d’u e ag i ultu e li ato-intelligente. Au Sénégal, la production de mil est concentrée dans le « bassin arachidier », où l'agriculture est dominée par des systèmes à base de céréales (mil) et de légumineuses (arachide et niébé), qui constituent la principale source d’ali e tatio et de revenus pour les communautés rurales.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 50
Cependant, un ravageur clé du mil, la i euse de l’ pi, Heliocheilus albipunctella de Joannis (Lepidoptera, Noctuidae), représente un obstacle ajeu à l’i te sifi atio de la p odu tio (Ba et al. 2013). Au Sénégal, ce ravageur a commencé à causer des dégâts dans les cultures de il jus u’à % de pe te de e de e t e g ai s suite à u e lo gue p iode de s he esse au d ut des a es . L’esp e est o ophage et u i olti e. Les adultes émergent du sol un à deux mois après le début de la saison des pluies. Après accouplement, les femelles pondent sur les épis au stade de la floraison. Après éclosion, les jeunes larves perforent les glumes et dévorent l'intérieur des fleurs, tandis que celles plus âgées coupent les p do ules flo au , fo a t ai si des gale ies su l’ pi selo u t a e spi ale caractéristique (Fig. 14). Au terme de leur développement, les larves se nymphosent dans le sol (chrysalides) où elles restent en diapause pendant toute la saison sèche. Malgré son incidence élevée (30 à % d’ pis i fest s da s la zo e de Ba e e -2014), aucun t aite e t i se ti ide ’est effe tu . Son impact sur la production en grain est contenue (2- 20% de perte) grâce à une très forte régulation naturelle de ses populations. Des o se atio s au ha p o t o t u’en l'absence de régulation naturelle des ravageurs, les pertes de grains auraient dépassé 90% dans un quart des champs (Sow et al. 2015). Ceci en fait un système biologique particulièrement perti e t pou alue l’effet de perturbations environnementales sur la structure des réseaux trophiques et les processus associés, comme la régulation écologique des bioagresseurs.
A B E A B
C D C D
Figure 14. A fe elle adulte d’Heliocheilus albipunctella en train de pondre sur un épi ; B) larve sur un épi ; C) parasitoïde larvaire, Cardiochiles sp. ; D) champ de mil ; E) système de vie (Photos T. Brévault).
L'impact de la régulation naturelle du ravageur, mais aussi sa variabilité d’a o da e d’u ha p à l’aut e, fo t de la i euse de l’ pi de il un système biologique très pertinent pour identifier les déterminants de la régulation écologique da s u o te te d’ag ofo este ie traditionnelle. En outre, beaucoup d’e seig e ents sont attendus sur le fonctionnement des populatio s d’i se tes da s es agroécosystèmes exempts de pesticides.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 51
3.4.2. Hypothèses
En l'absence de toute application d'insecticide par les agriculteurs dans les systèmes de culture à base de mil, notre première hypothèse est que la mortalité induite par les ennemis naturels est une composante majeure de la régulation des populatio s de la i euse de l’ pi de mil. Les prédateurs incluent les arthropodes [Hemiptera (Anthocoridae, Reduviidae), Coleoptera (Carabidae), Hymenoptera (Vespidae, Formicidae) et Dermaptera], et probablement aussi des vertébrés tels que les chauves-souris et les oiseaux. Les parasitoïdes recensés appartiennent aux ordres des Hyménoptères (Braconidae, Ichneumonidae, Trichogrammatidae) et des Diptères (Tachinidae) (Sow et al. 2017). Il s’agi a i i d’ alue l’i pa t de ces e e is atu els su l’a o da e des populatio s du a ageu i le, et de caractériser la diversité et la structure des communautés impliquées.
Notre deuxième h poth se est ue la atu e et l’i te sit des p o essus ologi ues impliqués dans la régulation écologique des ravageurs dépendent des caractéristiques du milieu et de la structure du paysage (utilisation des terres, abondance et configuration des habitats semi-naturels, etc.). Dans les agroécosystèmes à base de mil, les champs de case (à proximité des habitations) sont utilisés pour les céréales et bénéficient de la fertilisation organique par le fumier et les ordures ménagères (Tschakert & Tappan 2004). Les champs de brousse, plus lointains et généralement moins fertiles, sont consacrés aux rotations avec des légumineuses à graines et ne reçoivent pas nécessairement d'engrais (Manlay 2002). On peut pe se ue l’a lio atio de la fe tilit dans les champs de case devrait favoriser la régulation biologique des insectes ravageurs (Altieri et al. 2012). D’aut e pa t, la composition et la configuration du paysage agricole déterminent le potentiel des ressources, mais aussi leur accessibilité. La fragmentation des habitats peut affecter la dispersion et la distribution des ennemis naturels et les fonctions écologiques telles que la régulation. Les paysages agricoles dans le bassin arachidier sont structurés par des systèmes de parcs agroforestiers traditionnels (SAF), dans lesquels des arbres, principalement l'acacia Faidherbia albida, sont dispersés dans les champs cultivés. Si l'impact des SAF sur la fertilité des sols a été largement démontré, son effet sur les ravageurs des cultures et le contrôle biologique ’ont pas été étudiés (Andres & Bhullar 2016). À l'échelle du paysage, les SAF offrent une plus grande diversité de niches écologiques aux arthropodes dans le temps et dans l'espace qu'une simple mosaïque de cultures annuelles.
Notre troisième hypothèse est que l'utilisation des ressources par les acteurs, sur un territoire partagé, interagit avec les propriétés des paysages recherchées pour la régulation biologique de la mineuse de l’ pi. L'engagement des acteurs derrière notre approche (et l'i pa t ult ieu e se a effe tif ue si les solutio s so t possi les d’u poi t de ue organisationnel, discutées en fonction de leur intérêt pour la communauté, et conçues conjointement. La biomasse cultivée aussi bien que la biomasse semi-naturelle (arbres, pâturages, etc.) représentent des ressources ou des habitats potentiels pour les ravageurs et leurs ennemis naturels. Cependant, comme le montrent les études récentes menées en Afrique de l'Ouest, cette ressource est l'objet de tensions entre les agriculteurs et les éleveurs dans les territoires ruraux (Andrieu et al. 2015, Diarisso et al. 2015). Ces études concluent sur la nécessité d'une approche concertée de la gestion des ressources sur le territoire. Ici, la lutte contre les insectes ravageurs des cultures est un processus « invisible », o e elui de l’i festatio à ses d uts e.g. les œufs . Co e le sugg e t Opda et al.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 52
(2016), notre contribution consiste à rendre les services écosystémiques « visibles » afin de permettre leur inclusion dans les processus de décisions individuelles et collectives concernant la régulation naturelle des bioagresseurs.
3.4.3. Objectifs
Les objectifs généraux sont (i) de développer des solutions innovantes de gestion des ravageurs des cultures en s'appuyant sur une gestion collective et coordonnée de paysages agricoles multifonctionnels, et (ii) d’apprendre des environnements agricoles exempts de pesticides (bio-inspiration) comment la biodiversité et les processus écologiques associés doivent être gérés pour restaurer ou stimuler la régulation naturelle des populations de ravageurs, en particulier dans les paysages dits « immuno-déprimés ».
Les objectifs spécifiques de notre projet de recherche sont les suivants: • Identifier la dynamique du paysage et les forces en jeu, en interaction avec la gestion des ressources par les acteurs. • Evaluer les effets des pratiques culturales et de la structure du paysage (en particulier les habitats semi-naturels) sur l’i ide e de la i euse de l’ pi, la structure du réseau trophique et l’effi a it du o t ôle iologi ue asso i . • Identifier les règles génériques établies entre biodiversité et structure du réseau trophique pour restaurer ou stimuler le contrôle biologique des ravageurs. • D te i e l’ helle spatiale fo tio elle pou la gestion des communautés d’i se tes tudi es. • Identifier les règles générales pour co-concevoir des pratiques et des modes d’a age e t des ha itats dans les territoires explorés, en tenant compte des multi- fonctionnalités paysagères et des s e a ii d’ olutio de l’utilisation des terres.
Le projet proposé doit contribuer à l'invention de modèles « verts » d'agriculture qui respectent la préservation de l'environnement, l'adaptation au changement climatique et les objectifs de sécurité alimentaire. Les recommandations comprendront des plans agro- environnementaux et des règles de gestion collective conçues conjointement avec les acteurs des territoires concernés.
3.4.4. Approche méthodologique
Le projet est structuré selon quatre volets d’a ti it s (Fig. 15).
Figure 15. Volets d’a ti it s du programme de recherche. Volet 1 : Dynamique et structure du paysage. Volet 2 : D te i a ts pa sage s de l’i ide e des ravageurs et du contrôle biologique. Volet 3 : Réseaux trophiques et contrôle biologique. Volet 4 : Co-conception de paysages fonctionnels.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 53
Le volet 1 vise à identifier les facteurs socio-économiques du changement d'utilisation des terres et des trajectoires de paysage au cours des dernières décennies dans les territoires explorés et les conséquences possibles sur l'incidence des ravageurs. La dynamique spatiale et temporelle du paysage sera analysée à l'aide de séries chronologiques satellitaires et d'archives de projets de recherche ou de développement. Des enquêtes sur l’o upatio des sols seront menées pour comprendre l'évolution récente du paysage et mettre en évidence l'utilisation des ressources par les acteurs des territoires considérés. Cette première étape caractérisera l'évolution des espaces cultivés et de la végétation semi-naturelle, en mettant l'accent sur les arbres et les parcours qui contribuent à la conservation de la biodiversité et fournissent des services d'approvisionnement clés (bois de chauffage, médicaments, fourrage, etc.). Une thèse portant sur la gestion des ressources naturelles par les communautés u ales da s la zo e d’ tude, en regard de la régulation naturelle de la i euse de l’ pi de il, a démarré en 2015 (§ 1.5), sur un financement du Programme Dynafrique Tosca-Cnes (Encadré 1, page 5). Elle est encadrée par une collègue géographe de l’UPR Aida (V. Soti) basée au Centre de Suivi Ecologique (CSE) à Dakar.
Une deuxième étape de l'analyse sera effectuée avec des images satellites à très haute résolution afin d'évaluer la disponibilité spatio-temporelle des ressources pour les ravageurs et leurs ennemis naturels selon leur cycle biologique. Une typologie des paysages sera réalisée à partir de cette analyse pour dresser un plan d'échantillonnage des parcelles selon nos hypothèses de recherche concernant l'effet des pratiques agricoles et du paysage sur l'incidence et les dégâts de la i euse de l’ pi de il, mais aussi sur la régulation écologique. Un article méthodologique su la o st u tio d’u pla d’ ha tillo age à pa ti du t aite e t d’i ages satellites THRS est e p pa atio § 1.6.2 n°47).
Les champs de mil sélectionnés seront suivis dans le volet 2 pour identifier les déterminants de l'incidence des ravageurs et de l’efficience du contrôle biologique. L'abondance du ravageur cible, H. albipunctella, et le taux de parasitisme associé, seront estimés. Les dégâts seront évalués en mesurant le nombre et la longueur des spirales sur les épis de mil, tandis que les pertes de rendement seront estimées par des relations allométriques. L'exclusion des ennemis naturels (pose de manchons autour des épis infestés) sera effectuée sur des plantes sentinelles pour évaluer l'impact global du contrôle biologique (Biocontrol Service Index [BSI], voir Woltz et al. 2012). Les pratiques agricoles (préparation du sol, date de semis, variété, désherbage, cultures intercalaires, etc.) dans les pa elles d’o se atio seront évaluées à travers les entretiens avec les agriculteurs. Nous utiliserons des méthodes statistiques spatialement explicites permettant l'exploration des données à plusieurs échelles. Ce volet a démarré en 2014 avec le suivi de parcelles de mil dans deux zones du assi a a hidie , d’u e th se et de t ois stages de aste § . . et . . , dans le cadre d’un projet de renforcement de la régulation écologique des insectes ravageurs des cultures de céréales sèches (Recor) soutenu par le Programme de productivité agricole en Afrique de l’Ouest, en pa te a iat a e l’Is a et le CSE au S gal (Encadré 1, page 5). Nos premières observations montrent que l'abondance de la végétation semi-naturelle (principalement des arbres) favorise la régulation naturelle de la i euse de l’ pi de il, mais les processus écologiques sous-jacents sont encore inconnus (Brévault et al. 2016, Thiaw et al. 2017). Un a ti le su l’effet des p ati ues ultu ales et du pa sage su l’i ide e du a ageu est en préparation (§ 1.6.2 n°48).
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 54
Le volet 3 utilisera des outils moléculaires à haut débit pour documenter la structure des réseaux trophiques (y inclus ressources primaires, autres ravageurs et leurs ennemis naturels) et les processus écologiques sous-jacents du contrôle biologique. Nous proposons de décrire la structure des réseaux trophiques en utilisant une approche en deux étapes: (i) élaboration d'une base de données de référence de codes-barres ADN des plantes hôtes, proies, prédateurs, parasitoïdes et hyperparasitoïdes, et (ii) évaluation de la prédation et du parasitisme par la détection des proies dans l'intestin des prédateurs, et la détection de parasitoïdes dans les oeufs et les larves des herbivores, respectivement. Les données sur les spécimens échantillonnés ainsi que les séquences seront stockées dans une base. Les réseaux trophiques seront caractérisés en utilisant des méthodes d'analyse de réseau (Brandes & Erlebach 2005). La variation entre les structures des réseaux trophiques aidera à identifier comment les caractéristiques du paysage peuvent affecter la fonction de régulation des ravageurs (Tixier et al. 2013). Une thèse sur ce volet a démarré en 2015 (§ . . , a e u s jou e alte a e à l’UMR CBGP. U p e ie at i le su l’ide tifi atio o phologi ue et ol ulai e des e e is atu els de la i euse de l’ pi est e préparation (§ 1.6.2 n°49).
Les résultats des volets 1 à 3 seront intégrés dans un modèle multi-agent spatialement explicite du volet 4 utilisant le langage de modélisation Ocelet (Degenne et al. 2009) pour simuler les dynamiques et flux de populations de ravageurs, le contrôle biologique et les pertes de rendement en interaction avec la structure du paysage et les pratiques agricoles. Le modèle comprendra trois sous- od les: ha ge e t d’assole e t ou de st u tu e paysagère, (2) croissance et rendement de la culture, et (3) dynamique des populations de ravageurs modulée par les ennemis naturels et autres facteurs biotiques et abiotiques (Fig. 16). Il sera utilisé pour tester la résistance du paysage à la i euse de l’ pi de il, en simulant différents scénarii de dynamique ou de patrons de paysage. Cet outil devrait permettre de tester des hypothèses génériques pour la régulation écologique et la gestion des ravageurs dans les agroécosystèmes.
Figure 16. Diagramme de représentation du modèle spatialement explicite (Plateforme de modélisation Ocelet p se ta t des e e ples d’i te a tio s e t e e tit s. SNV : végétation semi- naturelle.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 55
Les connaissances recueillies dans le volet 1 sur la dynamique du paysage et ses principaux moteurs, seront utilisées pour l'opérationnalisation de la conception de paysages multifonctionnels et de règles de gestion collective, en collaboration avec les acteurs, notamment sous forme de jeux de rôles (D'Aquino & Bah 2014).
3.4.5. Collaborations scientifiques Le projet p oit d’i t g e une large gamme d'échelles et de compétences complémentaires des unités de recherche en France (UPR Aida, UMR CBGP, Green et Tétis) et au Sénégal (CSE et Biopass). Le projet proposé réunira des spécialistes du paysage et de la biodiversité, des agroécosystèmes et de la transition agroécologique, de l'écologie des insectes et du contrôle des ravageurs, de la gestion de l'environnement et de l'innovation. Cela impliquera un large éventail de disciplines allant de la systématique moléculaire et de l'écologie, de l'entomologie, de l'écologie des communautés, de l'écologie fonctionnelle et de l'écologie du paysage, à la modélisation écologique, à la géographie, à la gestion du paysage, à la socio-écologie et à l'ingénierie agroécologique.
Publications majeures 1. Brévault T & Bouyer J (2014) From integrated to system-wide pest management: Challenges for sustainable agriculture. Outlooks on Pest Management 25: 212-213. 2. Brévault T, Renou A, et al. (2014) DIVECOSYS: Bringing together researchers to design ecologically- based pest management for small-scale farming systems in West Africa. Crop Protection 66: 53– 60. 3. Brévault T, Soti V, Thiaw C & Clouvel P (2015) Maîtriser les paysages et les processus écologiques propres à cette échelle. In : Escadafal R. (ed.), Masse D. (ed.), Chotte J.-L. (ed.), Scopel E. (ed.). L'ingénierie écologique pour une agriculture durable dans les zones arides et semi-arides d'Afrique de l'Ouest. Montpellier : CSFD, Agropolis International, p. 46-49. 4. Thiaw I, Soti V, Goebel F-R, Sow A, Brévault T & Diakhaté M (2017) Effect of landscape diversity on biocontrol of the millet head miner, Heliocheilus albipunctella (Lepidoptera: Noctuidae), in Senegal. Landscape management for functional biodiversity, IOBC-WPRS Bulletin 122: 38-42.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 56
4. Conclusion
Mieux comprendre les processus qui sous-tendent la fonction de régulation écologique est un enjeu majeur pour élaborer des stratégies de gestion durable des ravageurs, avec une perspective d'enseignements forts pour la transition agroécologique des systèmes de production agricole au Sud comme au Nord. Ma contribution active au dispositif de e he he et d’e seig e e t e pa te a iat Di e os s Brévault et al. 2014) offre l'opportunité d'adosser l'app o he d’ ologie appli u e à la gestio des ioag esseu s que je soutiens, au monde de l'enseignement et du développement agricole, au Sénégal et au-delà en Afrique de l'Ouest.
D’aut es ha tie s s’ou e t e pa all le ou au croisement de ces activités de recherche. Ils s’i s i e t da s ot e d a he g ale de o positio de p ati ues et d’i o atio s techniques pour limiter l'incidence des bioagresseurs (§ 3.3). Complémentaires à la mobilisation des processus de régulations naturelles telle que nous l'avons présentée, ces chantiers ont pour objet d’ la gi le champ des pratiques (cf. titre du mémoire) par l'élicitation des défenses naturelles des plantes et des techniques de biocontrôle, dont je livre ici deux exemples emblématiques tirés des travaux menés en collaboration avec d'autres collègues :
L’i du tion de défenses naturelles des plantes (qui modifient la qualité de la ressource dans notre schéma, Fig. 13 page 46). Des expérimentations conduites au Mali depuis plusieurs années par l'IER (I. Téréta) et le Cirad (A. Renou) o t o t ue l’ i age des cotonniers permet de réduire sig ifi ati e e t l’i ide e de plusieu s i se tes ravageurs, en particulier les he illes de la apsule. L’ i age tel ue e o a d pa l'IER est une opération manuelle qui consiste en la suppression de la cime des cotonniers, une dizaine de jours après le début de la floraison. Les premiers résultats obtenus par notre équipe au Sénégal montrent un effet de dissuasion sur la ponte des femelles de noctuelles (tests de réponse comportementale en grande cage) et une augmentation de la prédation (proies sentinelles au champ), suite à l’ i age des oto ie s. Ils ouvrent aussi de ou eau f o ts de e he he su l’effet du nectar extra-floral, en quantité et en qualité, comme facteur potentiel de recrute e t d’e e is atu els et d’aug e tatio du contrôle biologique. Cette activité est en développement à Montpellier par une jeune chercheure récemment recrutée dans notre équipe (A. Llandres-Lopez, UPR Aida), qui sera affectée au Sénégal fin 2017.
Le développement de nouveaux systèmes de biocontrôle. Le projet exploratoire Biophora (Biocontrol Phoretic Agents), duquel je suis co-porteur, met l'accent sur l’utilisatio du concept de phorésie (du grec phoros : porter) : t pe d’i te a tio e t e deu organismes où l’u pho o te est t a spo t pa l’aut e age t pho ti ue . U e des appli atio s consisterait à utiliser les individus sauvages (auto-dissémination à partir de pièges) ou à produire et à relâcher des insectes (mâles stériles ou encore des auxiliaires des cultures), comme « convoyeurs » de biocides pour le contrôle de ravageurs ou de vecteurs con- ou même hétéro-spécifiques. Les principaux systèmes biologiques visés dans ce projet sont les mouches des fruits (Tephritidae), les mouches tsé-tsé (Glossinidae) et les moustiques Culi idae , ui so t diffi iles à i le de pa la atu e, l’a essi ilit ou la di e sit de leurs habitats. De nombreuses questions de recherche se posent sur le choix des agents
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 57
phorétiques, des biocides, de l’ helle de l’a tio , et sur l’ aluatio des is ues. Ce p ojet s’adosse au t a au d elopp s pa J. Bou e UMR Ast e da s le ad e d’u e bourse du Conseil européen de la recherche (ERC). Une jeune chercheure récemment affectée à Biopass (A. Chailleux, UPR Hortsys) démarre des études en ce sens sur les mouches des fruits.
"L’ha ilitatio à diriger des re her hes sa tio e la re o aissa e du haut iveau scientifique du candidat, du caractère original de sa démarche dans un domaine de la science, de son aptitude à maîtriser une stratégie de recherche dans un domaine scientifique ou technologique suffisamment large et de sa capacité à encadrer de jeunes chercheurs".
Je ne suis pas certain de répondre seul aux critères de haut niveau scientifique ou de caractère original de ma démarche, mais ce que je revendique dans ce travail, c'est une contribution à l'approche pluridisciplinaire de la gestion agroécologique des bioagresseurs, avec des collègues écologues, spécialistes de la télédétection, géomaticiens, modélisateurs, agronomes, sociologues ou géographes, et 'est à e tit e ue j’esp e t e e esu e de transmettre la flamme à de jeunes chercheurs.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 58
5. Références bibliographiques
Agra (2014) Africa Agriculture Status Report 2014. Climate change and smallholder agriculture in sub- Saharan Africa. Nairobi, Kenya, 216 pp. Ai Y , Qiu XH. & He FQ (2003) A review of insecticide resistance mechanism of Aphis gossypii. Entomological Knowledge 40: 385 – 391. Altieri MA (1987) Agroecology: The scientific basis of alternative agriculture. Boulder, CO: Westview Press. 143 pp. Altieri M, Ponti L & Nicholls CI (2012) Soil fertility, biodiversity and pest management. In: Biodiversity and insect pests: Key issues for sustainable management. Edited by GM Gurr, SD Wratten, WE Snyder, and DMY Read. John Wiley & Sons, Ltd. 72-84. Andres C & Bhullar GS (2016) Sustainable Intensification of Tropical Agro-Ecosystems: Need and Potentials. Frontiers in Environmental Science 4, 5. Andrews MC, Callaghan A, Field LM, Williamson MS & Moores GD (2004) Identification of mutations conferring insecticide-insensitive AChE in the cotton-melon aphid. Aphis gossypii Glover. Insect Molecular Biology 13:555–561. Andrieu N, Vayssières J, Corbeels M, Blanchard M, Vall E & Tittonell P (2015) From farm scale synergies to village scale trade-offs: Cereal crop residues use in an agro-pastoral system of the Sudanian zone of Burkina Faso. Agricultural Systems 34: 84-96. APRD (2016) Arthropod Pesticide Resistance Database Michigan State University. http://www.pesticideresistance.org accessed 03.23.2017. Consulté le 20/03/2017. Ba M, Baoua I, Ndiaye M, Dabire-Binso C, Sanon A & Tamò M (2013) Biological control of the millet head miner Heliocheilus albipunctella in the Sahelian region by augmentative releases of the parasitoid wasp Habrobracon hebetor: effectiveness and farmers' perceptions. Phytoparasitica 41: 569-576. Baker GH, Tann CR & Fitt GP (2008) Production of Helicoverpa spp. (Lepidoptera, Noctuidae) from different refuge crops to accompany transgenic cotton plantings in eastern Australia. Australian Journal of Agricultural Research 59:723–732. Bertrand A, Brévault T, Thézé M & Vaissayre M (2010) De la LEC à la LOIC : comment aider les paysans à prendre en charge la protection phytosanitaire de leurs parcelles de coton ? In : Seiny Boukar L. (ed.), Boumard Philippe (ed.). Savanes africaines en développement : Innover pour durer, 2009-04-20/2009-04-23, Garoua (Cameroun). Bianchi F., Booij CJH & Tscharntke T (2006). Sustainable pest regulation in agricultural landscapes: a review on landscape composition, biodiversity and natural pest control. Proceedings of the Royal Society of London B: Biological Sciences 273: 1715–1727. Bijlsma R & Loeschcke V (2005) Environmental stress, adaptation and evolution: an overview. Journal of Evolutionary Biology 18: 744-749. Bodin O & Crona BI (2009). The role of social networks in natural resource governance: What relational patterns make a difference? Global Environmental Change 19: 366–374. Brandes U & Erlebach T (2005) Network Analysis - Methodological Foundations. LNCS Tutorial 3418, Springer Verlag, 471 pp. Brévault T, Asfom P, Beyo J, Nibouche S & Vaissayre M (2002) Assessement of Helicoverpa armigera resistance to pyrethroids in northern Cameroon. Mededelingen Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen 67: 641–646. Brévault T & Achaleke J (2005) Status of pyrethroid resistance in the cotton bollworm, Helicoverpa armigera, in Cameroon. Resistant Pest Management Newsletter 15: 4–7. Brévault T, Sylla S, Diatte M, Bernadas G & Diarra K (2014) Tuta absoluta Meyrick (Lepidoptera: Gelechiidae): a new threat to tomato production in sub-Saharan Africa. African Entomology 22:441–444. Brévault T, Sow A, Thiaw I, Thiaw C, Delvare G & Soti V (2016) Tree-crop agroforestry systems promote natural control of the millet head miner, Heliocheilus albipunctella. Poster. 25 th
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 59
International Congress of Entomology: Entomology without Borders, 2016-09-25/2016-09-30, Orlando (Etats-Unis). Burel F & Baudry J (1999) Ecologie du paysage: concepts, méthodes et applications, Lavoisier, Paris. 360 pp. Carrière Y, Crowder D & Tabashnik BE (2010) Evolutionary ecology of insect adaptation to Bt crops. Evolutionary applications 3: 561-573. Chaplin-Kramer R, O'Rourke ME & Blitzer EJ (2011) A meta-analysis of crop pest and natural enemy response to landscape complexity. Ecology Letters 14: 922-932. Charaabi K, Carletto J, Chavigny P, Marrakchi M, Makni M & Vanlerberghe-Masutti F (2008) Clonal diversity of the melon aphid Aphis gossypii (Glover) in Tunisia is structured by host plants. Bulletin of Entomological Research 98: 333–341. Cong R-G, Smith HG, Olsson O & Brady M (2014) Managing ecosystem services for agriculture: Will landscape-scale management pay? Ecological Economics 99: 53–62. Crowder DW, Northfield TD, Strand MR & Snyder WE (2010) Organic agriculture promotes evenness and natural pest control. Nature 466: 109–112. Crowder D & Jabbour L (2014) Relationships between biodiversity and biological control in agroecosystems: Current status and future challenges. Biological Control 75: 8-17. Czepak C., Albernaz KC, Vivan LM, Guimaraes HO & Carvalhais T (2013) First reported occurrence of Helicoverpa armigera (Hübner) (Lepidoptera: Noctuidae) in Brazil. Agricultural Research in the Tropics. 43: 110-113. D’A ui o P & Bah A (2014) Multi-level participatory design of land use policies in African drylands: a method to embed adaptability skills of drylands societies in a policy framework. Journal of Environmental Management 132: 207-219. Degenne P, Lo Seen D, Parigot D, Forax R, Tran A, Ait Lahcen A, Curé O & Jeansoulin R (2009) Design of a domain specific language for modelling processes in landscapes. Ecological Modelling 220: 3527-3535. Deguine J-P, Martin J, Merlier H & Leclant F. (1997) Inventaire des plantes-hôtes d’Aphis gossypii Glover (Hemiptera, Aphididae) en Afrique. CIRAD-CA. 18 pp. Deguine J-P (1996) The evolution of insecticide resistance in Aphis gossypii Glover (Hemiptera: Aphididae) in Cameroon. Resistant Pest Management Newsletter 8: 13-14 . Diarisso T, Corbeels M, Andrieu N, Djamen Nana P & Tittonell P (2015) Biomass transfers and nutrient budgets of the agro-pastoral systems in a village territory in south-western Burkina Faso. Nutrient Cycling in Agroecosystems 101: 295-315. Fabrick JA; Unnithan Gopalan C, Yelich AJ et al. (2015) Multi-Toxin Resistance Enables Pink Bollworm Survival on Pyramided Bt Cotton. Ssientific Reports 12,5: 16554. FAO (2015) FAOSTAT database. http://faostat3.fao.org/home/E (consulté le 23/06/2016). Futuyma DJ & Moreno G (1988) The evolution of ecological specialization. Annual Review of Ecology and Systematics 19: 207–233. Gassmann AJ, Shrestha RB, Jakka SRK et al. (2016) Evidence of Resistance to Cry34/35Ab1 Corn by Western Corn Rootworm (Coleoptera: Chrysomelidae): Root Injury in the Field and Larval Survival in Plant-Based Bioassays Journal of Economic Entomology 109: 1872-1880. Goergen G, Lava Kumar P, Sankung SB, Togola A & Tamò M (2016) First Report of outbreaks of the fall armyworm Spodoptera frugiperda (JE Smith) (Lepidoptera, Noctuidae), a vew alien invasive pest in West and Central Africa. PLoS ONE 11(10): e0165632. Gurr GM, Wratten SD & Luna JM (2003) Multi-function agricultural biodiversity: pest management and other benefits. Basic and Applied Ecology 4: 107-116. Huang J K, Mi JW, Lin H et al. (2010) A decade of Bt cotton in Chinese fields: assessing the direct effects and indirect externalities of Bt cotton adoption in China. Science China-Life Sciences 52:981-991. Jaenike J (1990) Host specialization in phytophagous insects. Annual Review of Ecology and Systematics 21: 243-273.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 60
Kennedy GG & Storer NP (2000) Life systems of polyphagous arthropod pests in temporally unstable cropping systems. Annual Review of Entomology 45: 467-493. Landis DA, Wratten SD & Gurr GM (2000) Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology 45: 175-201. Letourneau DK, Jedlicka JA, Bothwell SG, Moreno CR (2009) Effects of natural enemy biodiversity on the suppression of arthropod herbivores in terrestrial ecosystems. Annual Review of Ecology, Evolution, and Systematics 40: 573-592. Loxdale HD & Lushai G (1999) Slaves of the environment: the movement of herbivorous insects in relation to their ecology and genotype . Philosophical Transactions of the Royal Society Series B, Biological Sciences 354: 1479-1495 . Mallet J & Porter P (1992) Preventing insect adaptation to insect-resistant crops—Are seed mixtures or refugia the best strategy. Philosophical Transactions of the Royal Society Series B, Biological Sciences 250: 165-169. Manlay RJ (2002) Carbon, nitrogen and phosphorus allocation in agro-ecosystems of a West African savanna III. Plant and soil components under continuous cultivation. Agriculture, Ecosystems and Environment 88: 249-269. Ma ti T, O hou O hou G, Hala N’Klo F, Vassal JM & Vaissa e M P eth oid esista e i the cotton bollworm, Helicoverpa armigera (Hübner), in West Africa. Pest Management Science 56: 549–554. Martin T, Chandre F, Ochou Ochou G, Vaissayre M. & Fournier D (2002) Pyrethroid resistance mechanisms in the cotton bollworm Helicoverpa armigera (Lepidoptera: Noctuidae) from West Africa. Pesticide Biochemistry and Physiology 74: 17–26. Martin T, Ochou Ochou G, Djihinto A, Traore D, Togola M, Vassal JM, Vaissayre M & Fournier D (2005) Controlling an insecticide resistance bollworm in West Africa. Agriculture, Ecosystems and Environment 107: 409–411. Mazzi D & Dorn S (2012) Movement of insect pests in agricultural landscapes. Annals of Applied Biology 160: 97–113. Mollot G, Tixier P, Lescourret F, Quilici S &Duyck P-F (2012) New primary resource increases predation on a pest in a banana agroecosystem. Agricultural and Forest Entomology 14: 317-323. Nibouche S (1999). Helicoverpa armigera (Lepidoptera, Noctuidae, Heliothinae). Les déprédateurs du cotonnier en Afrique tropicale et dans le reste du monde.CIRAD-CA, 51 pp. Nibouche S , Beyo J. & Brévault T (2002) Negative cross insensitivity in a dimethoate resistant strain of cotton aphid Aphis gossypii Glover in northern Cameroon. Resistant Pest Management Newsletter 12: 25-26. Opdam P, Coninx I, Dewulf A, Steingröver E, Vos C & van der Wal M (2016). Does information on landscape benefits influence collective action in landscape governance? Current Opinion in Environmental Sustainability 18: 107–114. Qiao FB, Huang JK, Rozelle S & Wilen J (2010) Natural refuge crops, buildup of resistance, and zero- refuge strategy for Bt cotton in China. Science China-Life Sciences 53:1227–1238. Ratnadass A, Fernandes P, Avelino J & Habib R (2012) Plant species diversity for sustainable management of crop pests and diseases in agroecosystems: a review. Agronomy for Sustainable Development 32: 273-303. Renou A & Deguine J-P (1992) Ravageurs et protection de la culture cotonnière au Cameroun . Coton et Fibres Tropicales 13 : 7-52. Renou A & Brévault T (2016) Ravageurs et maladies du cotonnier, et gestion intégrée des ravageurs. In : Crétenet Michel (ed.), Gourlot Jean-Paul (ed.). Le cotonnier. Versailles, Ed. Quae, 109-154. Rougerie R, Smith MA, Fernandez-Triana J, Lopez-Vaamonde C, Ratnasingham S & Hebert PDN (2011) Molecular analysis of parasitoid linkages (MAPL): gut contents of adult parasitoid wasps reveal larval host. Molecular Ecology 20 : 179-186. Rusch A, Valantin-Morison M, Sarthou JP & Roger-Estrade J (2010) Biological control of insect pests in agroecosystems: effects of crop management, farming systems and semi-natural habitats at the landscape scale. A review. Advances in Agronomy 109: 219-259.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 61
Showalter AM, Heuberger S, Tabashnik BE & Carrière Y (2009) A primer for using transgenic insecticidal cotton in developing countries. Journal of Insect Science 9: 1–39. Snyder WE & Tylianakis JM (2012) The ecology of biodiversity-biocontrol relationships. In: Biodiversity and insect pests: Key issues for sustainable management. Edited by GM Gurr, SD Wratten, WE Snyder, and DMY Read. John Wiley & Sons, Ltd. 23-40. Sow A, Ly B, Thiaw C & Brévault T (2015) Régulation naturelle de la mineuse de l'épi de mil, Heliocheilus albipunctella, dans la zone de Bambey au Sénégal. Séminaire DIVECOSYS sur la gestion agroécologique des bioagresseurs en Afrique de l'Ouest, 2015-06-02/2015-06-04, Dakar (Sénégal). Sow A, Brévault T, Delvare G et al. (2017) A review of natural enemies of the millet headminer, Heliocheilus albipunctella : morphological and molecular identification using high-throughput sequencing. Biological Control (en préparation). Symondson WOC & Harwood J.D., 2014. Special issue on molecular detection of trophic interactions: Unpicking the tangled bank. Molecular Ecology 23: 3601-3604. Tabashnik BE, Gassmann AJ, Crowder DW & Carrière Y (2008) Insect resistance to Bt crops: Evidence versus theory. Nature Biotechnology 26: 199–202. Tabashnik BE, van Rensburg JBJ & Carrière Y (2009) Field-Evolved Insect Resistance to Bt Crops: Definition, Theory, and Data. Journal of Economic Entomology 102: 2011-2025. Thompson JN (1994) The Coevolutionary Process. University of Chicago Press, Chicago, Illinois. 376 pp. Tixier P., Peyrard N, Aubertot J-N et al. (2013) Modelling interaction networks for enhanced ecosystem services in agroecosystems. Advances in Ecological Research 49: 437-480. Toda S, Komazaki S, Tomita T & Kono Y (2004) Two amino acid substitutions in acetylcholinesterase associated with pirimicarb and organophosphorus insecticide resistance in the cotton aphid Aphis gossypii Glover (Homoptera: Aphididae). Insect Molecular Biology 13: 549–553. Tschakert P & Tappan G (2004) The social context of carbon sequestration: considerations from a multi-scale environmental history of the Old Peanut Basin of Senegal. Journal of Arid Environments 59: 535-564. Tscharntke T, Klein AM, Kruess A, Steffan-Dewenter I & Thies C (2005) Landscape perspectives on agricultural intensification and biodiversity - ecosystem service management. Ecology Letters 8: 857-874. Tscharntke T, Bommarco R, Clough Y et al. (2007) Conservation biological control and enemy diversity on a landscape scale. Biologial Control 43: 294-309. Van Jaarsveld MJ, Basson NCJ & Marais P (1997) Synthetic pyrethroid resistance in field strains of Helicoverpa armigera (Hübner) (Lepidoptera: Noctuidae) in South Africa. African Plant Protection 4: 15–18. Vanlerberghe-Masutti F & Chavigny P (1998) Host-based genetic differentiation in the aphid Aphis gossypii Glover, evidenced from RAPD fingerprints. Molecular Ecology 7: 905–914. Vassal JM, Vaissayre M & Martin T (1997) Decrease in the susceptibility of Helicoverpa armigera Hu e Lepidopte a: No tuidae to p eth oid i se ti ides i Côte d’I oi e. Resista t Pest Management Newsletter 9: 14–15. Veres A, Petit S, Conord C & Lavigne C (2013) Does landscape composition affect pest abundance and their control by natural enemies? A review. Agriculture, Ecosystems and Environment 166: 110- 117. Woltz JM, Isaacs R & Landis DA (2012) Landscape structure and habitat management differentially influence insect natural enemies in an agricultural landscape. Agriculture, Ecosystems and Environment 152: 40-49.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 62
6. Liste des tirés à part joints
1. Brévault T, Carletto J, Linderme D & Vanlerberghe-Masutti F (2008) Genetic diversity of the cotton aphid Aphis gossypii in the unstable environment of a cotton growing area. Agricultural and Forest Entomology 10: 215–223. 2. Brévault T, Achaleke J, Sougnabé SP & Vaissayre M (2008) Tracking pyrethroid resistance in the polyphagous bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae), in the shifting landscape of a cotton-growing area. Bulletin of Entomological Research 98: 565–573. 3. Brévault T, Nibouche S, Achaleke J & Carrière Y (2012) Assessing the role of non-cotton refuges in delaying Helicoverpa armigera resistance to Bt cotton in West Africa. Evolutionary Applications 5: 53–65. 4. Brévault T, Heuberger S, Zhang M, Ellers-Kirk C, Ni X, Masson L, Li X, Tabashnik BE & Carrière Y (2013) Potential shortfall of pyramided transgenic cotton for insect resistance management. Proceedings of the National Academy of Sciences of the United States of America 110: 5806–5811. 5. Tabashnik BE, Brévault T & Carrière Y (2013) Insect resistance to Bt crops: lessons from the first billion acres. Nature Biotechnology 31: 510–521. 6. Brévault T, Renou A, et al. (2014) DIVECOSYS: Bringing together researchers to design ecologically- based pest management for small-scale farming systems in West Africa. Crop Protection 66: 53– 60.
Scène de travail familial au champ en début de saison des pluies, Bambey, Sénégal, 2016.
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 63
Dossie d’Ha ilitatio à Di ige des Re he hes – Thierry Brévault 64 Agricultural and Forest Entomology (2008), 10, 215–223 DOI: 10.1111/j.1461-9563.2008.00377.x
Genetic diversity of the cotton aphid Aphis gossypii in the unstable environment of a cotton growing area
T. Brévault , J. Carletto * , D. Linderme and F. Vanlerberghe-Masutti * CIRAD, UPR Cotton Farming Systems, Garoua, Cameroon; CIRAD, UPR Cotton Farming Systems, Montpellier, F-34398 France; IRAD PRASAC-ARDESAC, Garoua, Cameroon and *UMR 1112 INRA-UNSA R.O.S.E., Sophia Antipolis F-06903, France
Abstract 1 Spatial and temporal habitat heterogeneity represented by annual crops is a major factor influencing population dynamics of phytophagous insect pests such as the cotton aphid Aphis gossypii Glover. We studied the effects of instability of the cotton agroecosystem resulting from the temporary availability of the plant re- source and the repeated use of insecticides on the genetic variability of the cotton aphids. 2 Samples of A. gossypii were collected in cotton plots, treated or not with insecti- cides and from vegetable crops (Malvaceae, Cucurbitaceae and Solanaceae) within the cotton growing area of northern Cameroon. The genetic structure of the samples was assessed using eight microsatellite markers. Insecticide resist- ance was estimated through the detection of two mutations in the ace -1 gene that are associated with insensitivity of acetylcholinesterase to carbamate and organo- phosphate insecticides. 3 The results obtained show that both host plants and insecticides act in genetic structuring of A. gossypii. Ninety-three percent of aphids collected on cotton were characterized by the same microsatellite multilocus genotype, Burk1 , which also displays the insecticide resistant alleles. 4 During the dry season, the cotton crop season after, the genotype Burk1 was principally found on two other malvaceous cultivated plants, rosella and okra, acting as suitable reservoir plants. The ability of the cotton aphid to move among asynchronous suitable habitats in response to changes in resource availability enables the pest to exploit unstable cropping systems. An understanding of the cotton aphid life system may aid to improve strategies for integrated resistance management. Keywords Agricultural landscape, Aphis gossypii , cotton , genetic diversity, host plant specialization, insecticide resistance, microsatellite markers.
Introduction great distances ( Stinner et al. , 1983; Loxdale & Lushai, 1999), whereas others may move more locally by tracking a Agroecosystems are characterized by environmental hetero- sequence of temporarily suitable host plants ( Brandenburg & geneity in space and time. Because the availability and suita- Kennedy, 1982; Shelton & North, 1986 ). Consequently, many bility of the crops vary dramatically throughout the year and polyphagous pests exploit a sequence of crops that offer a over years, insect pests in annual cropping systems are more substantial but time limited resource, and uncultivated or likely to be polyphagous and mobile than phytophagous in- scarce hosts that permit them to bridge the critical period. sects living in natural habitats ( Kennedy & Storer, 2000 ). Insect pests of annual crops are generally r-strategists, hav- Insect pest population dynamics in an agricultural landscape ing high capacities for population increase and are poor com- is dictated by the ability of the insects to disperse to and to petitors ( Wallner, 1987). Moreover, many of them are very exploit different habitat patches. Populations may move over responsive to seasonal temperatures or rainfall patterns, so that their cyclic outbreaks are determined by the variations in Correspondence: T. Brévault. Tel: + 33 4 67 61 55 00; fax: + 33 4 both food availability and weather. Furthermore, the suitabil- 67 61 56 66; e-mail: [email protected] ity of crop patches for the development and survival of the
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society 216 T. Brévault et al. offspring of immigrant insects is strongly affected by cultural by crop season and insecticide treatments. Aphids were practices, particularly insecticide use ( Kennedy & Storer, sampled on insecticide-treated and untreated cotton plots, 2000 ). and on different vegetable crops belonging to the Malvaceae, In sub-Saharan west and central Africa, cotton fields oc- Cucurbitaceae and Solanaceae families or on weeds. cupy a significant part of the agricultural landscape during Microsatellite markers were used to study the amount and the rainy growing season (mid-May to November) and har- variation of clonal diversity in A. gossypii and to reveal any bour a great diversity of insect pest species whose popula- transition of clones from, for example, cotton fields at the tions are partially controlled by the use of insecticides end of the rainy season to vegetable crops and weeds during ( Renou & Deguine, 1992 ). The dry season involves a rapid the dry season. Moreover, molecular markers associated with change in this agricultural landscape, with a sudden short- the adaptive mutations S431F and A302S were used to fol- age of resources for phytophagous insects. At this time, low variation in the genetic diversity of these populations in small irrigated plots generally located in the vicinity of response to insecticide pressure. towns and villages probably play a key role in the life sys- tems of cotton insect pests. The cotton aphid Aphis gossypii Glover is a cosmopolitan and polyphagous species widely Materials and methods distributed in tropical, subtropical and temperate regions. In west and central Africa, this aphid colonizes cotton crops Aphid sampling from mid-July to November, whereas small surfaces of irri- gated vegetable crops and opportunist weeds are mainly in- Aphids were collected in 2003 and 2004 from two cotton 2 fested by A. gossypii from November to late April ( Deguine fields in Djalingo (9°23Ј N, 13°45Ј E; 4800 m ) and Kodek 2 et al., 1999). Two periods of population growth on cotton (10°66 ЈN, 14°41Ј E; 4000 m ), with the two sites separated by crops are generally observed: the first in the beginning of 250 km. Half of each field received insecticide treatment at a the crop season after a period without rain and the second at weekly interval throughout the growing season with an orga- the end of the crop season when the rainy season is over. As nophosphate (dimethoate, 400 g/ha) whereas the other part a sap feeder and virus vector, this aphid may directly affect was left without any insecticide treatment. No seed treatment plant health, particularly at the early cotton crop stages. At was applied. From the twenty-first day after sowing, cotton the latter stages, it represents a strong threat to cotton fibre plants were inspected daily. Apterous individuals were col- quality because of honeydew contamination on open bolls lected from 40 – 50 plants at the beginning of the cotton grow- (sticky cotton). This aphid species is characterized by a ing season as soon as the first colonies were detected (25 July strictly clonal reproduction all the year round throughout 2003 and 27 July 2004 in Djalingo, 5 August 2003 and 2 most of its range in Africa ( Blackman & Eastop, 1984; Ebert August 2004, in Kodek). The same plants were sampled at & Cartwright, 1997 ). Because of a high rate of increase and the peak of the first infestation (24 August 2003 and 13 its ability to disperse by winged morphs, this pest has the po- August 2004 in Djalingo, 27 August 2003 and 17 August tential to rapidly colonize diverse agroecosystems. Analysis 2004 in Kodek) and later on at the peak of the second infesta- of the genetic diversity of A. gossypii by the use of molecu- tion (23 October 2003 in Djalingo and 28 October 2003 in lar markers has shown that the clonal diversity is structured Kodek). Additional sampling was performed throughout the by its host plants ( Vanlerberghe-Masutti & Chavigny, 1998; dry season in irrigated areas near Garoua (9°31Ј N, 13°40 ЈE) Fuller et al., 1999; Vanlerberghe-Masutti et al., 1999). close to Djalingo, and Maroua (10°60 ЈN, 14°32 Ј E) close to Furthermore, exponential demography associated with par- Kodek. Aphids were collected on vegetable crops (Solanaceae, thenogenesis in this aphid species favoured the rapid selec- Cucurbitaceae and Malvaceae), which represent the main host tion for insecticide resistance mechanisms as a consequence plants of A. gossypii in northern Cameroon, and on weeds of intense insecticide use ( Delorme et al. , 1997; Ahmad that are uncommon host plants. In each plot, apterous adults et al., 2003; Ai et al., 2003 ). Severe cases of resistance to were collected from plants at the peak of infestation. Each dimethoate have been reported for over a decade in northern aphid was placed in a 1.5 ml vial containing ethanol (95%). Cameroon ( Deguine, 1996; Nibouche et al. , 2002 ). As for many other insects, resistance to carbamate and organophos- Genetic analyses with microsatellite markers phorous insecticides in A. gossypii is due to structural altera- tions of acetylcholinesterase (AChE) ( Moores et al. , 1996; A total of 1176 aphids were genotyped for the eight microsat- Delorme et al., 1997). It has been demonstrated that AChE ellite loci specific of A. gossypii as described by Vanlerberghe- insensitivity of highly-resistant strains of A. gossypii results Masutti et al. (1999). For DNA extraction, aphids were washed from two point mutations in the ace 1 gene: the first one twice in a NaCl 0.65% (w/v) solution, before crushing in a changing the serine codon at position 431 into a pheny l- 0.5 mL tube containing 55 l of 5% Chelex (w/v). Each tube alanine codon (S431F) and the second changing the alanine was heated at 56 °C for 30 min and at 100 °C for 7 min, then codon at position 302 into a serine codon (A302S) (Li & vortexed, centrifuged and stored at – 20 °C. Microsatellite loci Han, 2002; Andrews et al., 2004; Toda et al., 2004). amplification was performed in two polymerase chain reac- The present study aimed to assess the genetic structure and tion (PCR) reactions, the first comprising a multiplex PCR. functioning of cotton aphid populations in the heterogeneous The forward primer of each microsatellite locus was labelled agricultural landscape of northern Cameroon, marked by un- with a fluorescent dye (FAM, NED, PET, VIC). Different dyes stable plant resource availability and suitability as affected were chosen for loci having the same allele size to analyse
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 Clonal diversity in unstable environment 217 simultaneously the eight microsatellite loci (Ago24-FAM, step of 7 min at 72 °C. Ten microlitres of the PCR product Ago53-VIC, Ago59-NED, Ago66-VIC, Ago69-NED, Ago84- were digested in two distinct reactions respectively with 2 U PET, Ago89-PET and Ago126-FAM). The multiplex PCR re- of Ssp I and 2 U of Nae I restriction enzymes at 37 °C for 3 h. action using the specific primers of the seven loci Ago53, DNA fragments were separated by electrophoresis in a 1.5% Ago59, Ago66, Ago69, Ago84, Ago89 and Ago126 was per- agarose gel at 100 mA for 45 min in TBE 1X and revealed af- formed in a final volume of 5 l containing 1X reaction ter ethidium bromide (0.5 g/ml) staining under ultraviolet buffer Qiagen Multiplex PCR Master Mix (containing dNTP- lights. The susceptible codon in position 431 (serine) is in- m mix, HotStartTaq DNApolymerase and 3 m MgCl2 as the fi- cluded in the SspI restriction site, whereas the susceptible nal concentration; Qiagen Inc., Valencia, California), 0.2 m codon in position 302 (alanine) is part of the Nae I restriction of each primer and 1 l of Chelex DNA extraction diluted at site; therefore, only the susceptible alleles are cut. The size 1 : 10. The cycling conditions on an Eppendorf thermocycler of the various bands was established by comparison with a were: initial denaturation at 95 °C for 15 min followed by 25 100– 900 bp size marker (Smart Ladder; Eurogentec). cycles of 30 s at 95 °C, 90 s at 56 °C and 30 s at 72 °C and a fi- nal extension for 30 min at 60 °C. The second PCR using the primer specific of the locus Ago24 was performed in the same Results conditions except for the primer concentration (0.1 m ) and the thermocycler programming: 5 min at 95 °C, 35 cycles of Genetic diversity on cotton 30 s at 95 °C, 45 s at 62 °C and 30 s at 72 °C, and a final elon- gation of 7 min at 72 °C. One microlitre of each of the two The genetic variability detected at the eight microsatellite loci PCR reactions was mixed with 0.15 l of GeneScan ™ - over the sample of 612 aphids collected from cotton crops 500LIZ Size Standard and 10 l of Hi-Di Formamide permitted the identification of only three multilocus geno- (Applied Biosystems, Foster City, California) for denaturing. types: Burk1 , Burk2 and Ivo ( Table 1 ). Burk2 was detected in PCR products were separated and detected by capillary elec- only one isolated individual, whereas Burk1 was largely pre- trophoresis with an ABI 3100 automated sequencer (Genetic dominant whatever the time of sampling, the site or the insec- Analyzer). Results were interpreted using the software ticide treatment ( Table 2). Its frequency varied over the range STRand 2.2.241 (Acid Nucleic Analysis Software, http:// 73 – 100%. There was a significant effect of the insecticide www.vgl.ucdavis.edu/informatics/STRand/) that identifies treatment on the frequency of Burk1 and Ivo at the start of the allele sizes at each microsatellite locus by comparison with infestation ( 2 = 6.99, P < 0.01) and the early season peak of the standard size. Each individual was assigned a multilocus infestation ( 2 = 17.24, P < 0.01), whereas the year had no sig- genotype representing the combination of the alleles at the nificant effect. The site had a significant effect only at the eight loci. Individuals sharing the same multilocus genotype early season peak of infestation ( 2 = 10.28, P < 0.01). At the were considered to belong to the same clone. Changes in gen- start of the infestation, Burk1 was generally more frequent in otype frequencies between two consecutive sample collections treated plots than in untreated plots ( 2 = 5.97, P < 0.05) and on marked plants in treated and nontreated cotton fields, were its frequency in treated plots increased systematically statistically analyzed using a chi-square test followed by ( 2 = 7.13, P < 0.01). At the end of the cropping season in Marascuilo procedure using XLSTAT (Fahmy, 2006). 2003, Burk1 was the only genotype recovered in treated or To investigate the relationships among the multilocus gen- untreated cotton plots and its frequency was significantly otypes, we used the program populations , version 1.2.28 higher than at the beginning of the season ( 2 = 12.96, ( http://www.cnrs-gif.fr/pge/bioinfo) to generate a matrix of P < 0.01). A second period of population growth at the end of pairwise genetic distances based on the Allele Shared the rainy season in 2004 was not observed ( Table 2). Distance ( Jin & Chakraborty, 1993 ) and to construct a Table 3 shows the changes in the clonal composition of the Neighbour-joining tree. Bootstrap values were computed by infesting population on the marked plants from the initial in- resampling loci and are given as percentages over 2000 festation to the early peak in treated and untreated fields. In replications. plots treated with dimethoate, all the plants initially infested by Ivo were then heavily colonized by Burk1. The reverse situation was rarely observed. In untreated plots, 68.4% of Detection of S431F and A302S mutation the plants initially infested by Ivo were colonized by Burk1 on ace1 gene at the time of the early peak of infestation, whereas the re- A PCR-RFLP procedure to detect the two mutations S431F verse was observed for only 10.4% of the plants. and A302S on ace 1 gene was adapted from that described by Andrews et al. (2004) . The PCR was performed in a final Genetic diversity on irrigated crops and weeds volume of 25 l containing 1X of reaction buffer, 1.5 mm of m m MgCl2 , 0.25 m of each dNTP, 1.2 m of RESF2 and RESR2 Eight additional multilocus genotypes were identified from primer ( Andrews et al., 2004), 0.25 U of Taq polymerase the analysis of 564 aphids collected in vegetable crops and (GoldStar; Eurogentec, Belgium), 3 l of DNA Chelex ex- weeds during the dry season ( Table 1 ). Two genotypes were traction diluted at 1 : 5. Amplifications were carried out in a found to diverge from Burk1 by one allele at locus Ago59 thermocycler (Gene Amp PCR System 2700; Applied and were named Burk3 and Burk4 . These two genotypes Biosystems) for 5 min at 94 °C followed by 40 cycles of characterized, respectively, one individual collected from 1 min at 95 °C, 1 min at 58 °C and 1 min at 72 °C and a final okra and two individuals collected from rosella ( Table 4).
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 218 T. Brévault et al.
Table 1 Identification of multilocus genotypes of Aphis gossypii in cotton, vegetable crops and weeds, assessed by allele combination at eight microsatellite loci: identification of mutations S431F and A302S in ace 1 gene for each multilocus genotype
Microsatellite loci ace -1 gene Multilocus genotype Ago24 Ago53 Ago59 Ago66 Ago69 Ago84 Ago89 Ago126 N Freq S431F A302S n
Burk1 153-155 116-116 163-211 145-152 107-115 118-118 154-160 166-176 870 0.740 RR RR 288 Burk2 153-155 113-116 163-211 145-152 107-115 118-118 154-160 166-176 1 8.5 × 10 –4 RR RR 1 Burk3 153-155 116-116 163-209 145-152 107-115 118-118 154-160 166-176 1 8.5 × 10 –4 RR RR 1 Burk4 153-155 116-116 163-213 145-152 107-115 118-118 154-160 166-176 2 1.7 × 10 –3 RR RR 2 Ivo 153-153 113-116 161-175 152-154 113-113 104-124 154-154 166-176 54 0.050 RS SS 40 C9 153-157 116-116 182-182 152-152 109-114 112-118 150-150 176-176 97 0.082 RR SS 10 C10 153-153 116-116 182-188 152-152 109-111 112-112 150-150 176-176 2 1.7 × 10 –3 RR SS 2 C11 153-153 116-116 180-182 152-152 109-114 112-116 150-150 176-176 31 0.026 RR SS 5 C12 153-153 116-116 182-188 152-152 109-111 112-116 150-150 176-176 2 1.7 × 10 –3 RR SS 2 Auber 155-158 113-116 161-161 147-152 107-107 110-110 150-150 180-180 60 0.051 RR SS 6 PsP4 107-107 116-116 163-163 152-152 115-115 107-115 150-150 166-176 56 0.048 RR RR 6
Freq, frequency of the multilocus genotype; N , total number of aphids exhibiting the multilocus genotype; n , total number of aphids analyzed for the two mutations S431F and A302S; R, resistant allele; S, susceptible allele.
Most of the aphids (96%) infesting these two cultivated genotype Burk1 and 3% had the genotype Auber . All aphids plants exhibited the multilocus genotype Burk1 . By contrast, sampled on spontaneous plants belonging to Capparidaceae, out of the 131 aphids collected on cucurbits (water melon, Convolvulaceae or Euphorbiaceae displayed the genotype melon or cucumber), only one individual displayed the geno- Burk1 ( Table 4 ). type Burk1 and all the other aphids were distributed between the four closely-related multilocus genotypes, C9 , C10 , C11 Genetic relationships among the or C12, among which C9 was the most frequently observed multilocus genotypes (68%) ( Table 4 ). Among the aphids collected on eggplants, 90% of those sampled in 2004 displayed the genotype Burk1 , The four Burk genotypes that derived from one another by whereas 100% displayed the genotype Auber in 2005. This one or two alleles over the eight microsatellite loci ( Tables 1 genotype Auber characterized 70% of the aphids collected and 5 ) clustered in the phylogenetic network shown in Fig. 1 . from another Solanum species, the black nightshade, whereas A second cluster, very divergent from the Burk cluster and Burk1 and Ivo represented 8% and 22%, respectively. Ninety supported by a high bootstrap value, comprised the C geno- percent of the aphids collected from sweet peppers exhib - types that differed by two to four alleles. The other multilocus ited the divergent multilocus genotype PsP4 , 7% had the genotypes, Ivo , Auber and PsP4 , corresponded to three
Table 2 Relative frequencies of Aphis gossypii genotypes collected in treated and nontreated cotton fields
Early season Late season Initial infestation peak infestation peak infestation
Year Site Insecticide Genotype n Freq n Freq n Freq
2003 Djalingo Treated Burk1 25 1 26 1 26 1 Nontreated Burk1 26 0.765 24 0.727 32 1 Ivo 8 0.235 9 0.273 0 0 Kodek Treated Burk1 27 0.931 28 1 27 1 Ivo 2 0.069 0 0 0 0 Nontreated Burk1 26 0.920 26 0.960 26 1 Ivo 1 0.035 1 0.040 0 0 Burk2 1 0.035 0 0 0 0 2004 Djalingo Treated Burk1 25 0.962 43 1 Ivo 1 0.038 0 0 Nontreated Burk1 39 0.830 35 0.833 Ivo 8 0.170 7 0.167 Kodek Treated Burk1 26 0.929 26 1 Ivo 2 0.071 0 0 Nontreated Burk1 28 0.903 28 1 Ivo 3 0.097 0 0
Freq, frequency of individuals.
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 Clonal diversity in unstable environment 219
Table 3 Genotype switch between two consecutive sample 292 aphids displaying either the multilocus genotype Burk1 , collections on marked plants, in treated and nontreated cotton Burk2, Burk3 or Burk4 were homozygous for the resistant fields mutation at both positions, as were the aphids exhibiting the genotype PsP4 . Aphids having the genotype Ivo were hetero- Initial Early peak Insecticide infestation infestation N Freq N zygous for the S431F mutation and homozygous for the sus- s t ceptible codon at position 302. Aphids belonging either to Treated Burk1 Ivo 4 0.041 98 the C genotypes or to Auber were homozygous for the resist- Ivo Burk1 5 1 5 ant mutation at position 431 and homozygous for the suscep- Nontreated Burk1 Ivo 11 0.104 106 tible codon at position 302. Ivo Burk1 13 0.684 19
Freqs , frequency of switches on the ( Nt ) total number of clones at the initial infestation. Discussion phylogenetic lineages that diverged from each other and from Low clonal diversity and host specificity the Burk and C lineages by 9 –12 alleles. Analysis of the polymorphism observed at eight microsatel- lite loci in 1176 aphids of the species A. gossypii collected from cotton crops, vegetable crops and weeds in northern Detection of S431F and A302S mutation Cameroon, revealed a very low genetic diversity: only 11 on ace1 gene multilocus genotypes were identified. Moreover, the genetic A subsample of 363 aphids collected from cotton or irrigated relatedness of these genotypes suggests that the genotypic vegetable crops was genotyped for the two adaptive muta- diversity of the sample was distributed into five genetic tions S431F and A302S on the ace 1 gene ( Table 1 ). All the lineages ( Fig. 1 ). Therefore, 74% of the individuals fell into
Table 4 Relative frequencies of Aphis gossypii genotypes collected on vegetables and weeds
Family Host plant Site Date of sampling Genotype n Freq
Irrigated Cucurbitaceae water melon Citrullus Garoua 25 February 2005 C11 23 0.719 vegetable lanatus (Thunb.) C9 9 0.281 crops Melon Cucumis melo L. Garoua 25 February 2005 C9 33 1 Garoua 3 March 2005 C9 29 0.879 C10 2 0.061 C11 1 0.030 Burk1 1 0.030 Cucumber Cucumis sativus L. Garoua 3 March 2005 C9 24 0.727 C11 7 0.212 C12 2 0.061 Malvaceae Okra Abelmoschus esculentus Garoua 11 February 2004 Burk1 31 1 (L.) Moench 5 February 2005 Burk1 30 0.938 C9 2 0.063 Maroua 16 May 2003 Burk1 30 0.970 Burk3 1 0.030 Maroua 11 March 2005 Burk1 32 1 Rosella Hibiscus sabdariffa L. Garoua 3 February 2004 Burk1 31 0.969 Ivo 1 0.031 5 February 2005 Burk1 57 0.867 Ivo 3 0.067 Burk4 2 0.067 Solanaceae Eggplant Solanum melongena L. Garoua 25 February 2004 Burk1 28 0.903 Auber 3 0.097 5 February 2005 Auber 31 1 Sweet pepper Capsicum Garoua 6 January 2004 PsP4 30 1 annuum L. 5 February 2005 PsP4 26 0.900 Burk1 2 0.069 Auber 1 0.034 Black nightshade Solanum Garoua 1 April 2005 Auber 25 0.690 nigrum L. Ivo 8 0.220 Burk1 3 0.080 Weeds Capparidaceae Aristolochia albida Duch. Garoua 7 April 2005 Burk1 40 1 Convolvulaceae Ipomoea eriocarpa R. Br. Garoua 7 April 2005 Burk1 8 1 Euphorbiaceae Euphorbia hirta L. Garoua 7 April 2005 Burk1 8 1
Freq, frequency of individuals.
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 2 2 0 T. Brévault et al.
Table 5 Genetic distances between pairs of all 11 Aphis gossypii multilocus genotypes
Burk1 Burk2 Burk3 Burk4 Ivo C9 C10 C11 C12 Auber
Burk1 Burk2 0.0625 Burk3 0.0625 0.1250 Burk4 0.0625 0.1250 0.0625 Ivo 0.6250 0.5625 0.6250 0.6250 C9 0.6250 0.6875 0.6250 0.6250 0.7500 C10 0.6875 0.7500 0.6875 0.6875 0.6875 0.2500 C11 0.6875 0.7500 0.6875 0.6875 0.6875 0.1875 0.1875 C12 0.6875 0.7500 0.6875 0.6875 0.6875 0.2500 0.0625 0.1250 Auber 0.7500 0.6875 0.7500 0.7500 0.7500 0.7500 0.7500 0.7500 0.7500 PsP4 0.5625 0.625 0.5625 0.5625 0.7500 0.5625 0.5625 0.5625 0.5625 0.7500 the lineage Burk , 11% into the C lineage, whereas the re- whereas Ivo is detected on cotton crops only in Western maining 15% were equally distributed into Ivo, Auber and Africa (J. Carletto, unpublished data). This strongly suggests PsP4 genotypes. In addition, all the aphids found to feed on that Burk and Ivo clones are specialized on cultivated malva- cucurbits during the dry season belonged to the C genetic ceous host plants (cotton, okra, rosella) and they could rep- lineage, which is specific to the cucurbit host race ( Fuller resent a second host race in the species A. gossypii beside the et al. , 1999; Vanlerberghe-Masutti et al., 1999; Charaabi et al. , cucurbit host race. In this aphid species, which reproduces 2008 ). The Burk and Ivo genetic lineages represented 100% mainly by apomictic parthenogenesis, specialization may oc- of the aphids sampled during the rainy season in cotton fields cur if fitness variation associated with host plant already ex- ( Table 2 ) and 99% of the aphids collected on two other mal- ists because selection will act rapidly among asexual vaceous crops (okra and rosella) during the dry season populations, retaining the genotypes having the best per- (Table 4 ). The genotype Burk1 also characterized the few formance on a particular host plant. Host-plant specializa- aphids found in very low densities on weeds and it was occa- tion associated with genetic divergence of aphid populations sionally found at high frequency among aphids collected is particularly well documented in the pea aphid from eggplant in Garoua in 2004 ( Table 4 ). This suggests that Acyrthosiphon pisum (Harris) (Via, 1999; Via et al. , 2000; Burk1 is rather generalist. However, this genotype has never Hawthorne & Via, 2001; Simon et al., 2003; Frantz et al. , been detected among the numerous A. gossypii collected on 2006). However, the processes that are responsible for host eggplant crops in Burkina Faso (J. Carletto, unpublished specialization in aphids are unknown. In particular, differ- data), nor on Solanaceae (potato, tomato and green pepper) ences in physiological and behavioural traits that affect re- or Malvaceae (Hibiscus syriacus L.) in Tunisia (Charaabi source and habitat use by the aphid host races have to be et al., 2008 ). Furthermore, the genotype Burk1 is found to elucidated. It has long been suggested that landing of alate characterize most of the aphids from cotton crops in different aphids on plants is visually mediated and that chemical cues countries from Western Africa, Madagascar and Brazil, are detected afterwards. However, several laboratory and field studies have indicated that plant volatiles influence the orientation behaviour of walking and flying aphids prior to host contact ( Pickett et al., 1992). In particular, winged fe- male gynoparae are able to respond to volatile cues released by their specific primary host plant ( Powell & Hardie, 2001 ). Our observations showed that cotton-adapted genotypes were present at the very early stages of cotton crop infestation but we have no evidence that A. gossypii alate virginoparae (asexual females) are selected in response to cotton olfactory cues. Mechanisms for plant acceptance or rejection by aphids are essentially based on nutritional cues relying on plant sec- ondary metabolites that are assessed during stylet penetra- tion and probing of peripheral plant tissues before contact with the phloem. However, the active compounds that might stimulate or arrest parturition by clones are not identified ( Powell et al. , 2006 ).
Influence of insecticide treatments and Figure 1 Neighbour-joining tree based on shared allele distances competition among clones calculated for 11 multilocus genotypes of Aphis gossypii from various crops in Cameroon. Bootstrap values are given as a The screening of 363 A. gossypii individuals for the presence percentage over 2000 replications. of the mutation S431F in their ace 1 gene sequence revealed
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 Clonal diversity in unstable environment 2 2 1 that all individuals were homozygous for the mutation except Dry season relay hosts and implications those belonging to the multilocus genotype Ivo , which were for aphid management heterozygous. The second mutation A302S was found in The cotton-adapted genotypes Burk1 and Ivo were detected aphids having either a Burk genotype or the PsP4 genotype in high proportions in samples collected from other cultivated (Table 1). The insecticide resistant phenotypes of these indi- Malvaceae, okra and rosella. These crops are heavily infested viduals are not known but Andrews et al. (2004) monitored and are present all year round over much of northern biochemical assays of acetylcholinesterase insecticide sensi- Cameroon. They may constitute excellent reservoirs for cot- tivity of three strains of A. gossypii that were genotyped for ton-adapted genotypes during the dry season when cotton these two mutations. They showed that a phenylalanine at po- hosts are no longer available, particularly for resistant geno- sition 431 was highly correlated with insensitivity to pirimi- types because of intensive insecticide use on the vegetable carb (350-fold) and to a wide range of carbamates and crops. Some weeds were also found to host Burk1 during the organophosphates (three- to 25-fold). The second mutation, a dry season, but their importance appears to be low compared serine at position 302, was always found in conjunction with with okra and rosella because very few individuals were the first one and greatly increased pirimicarb insensitivity found. The ability of the cotton aphid to move among and (4500-fold) and slightly increased insensitivity to the other utilize different and asynchronous habitat patches in response insecticides (nine- to 150-fold). According to their results to changes in availability and suitability enables the pest to ( Andrews et al. , 2004), we suggest that, in the present study, exploit unstable agricultural landscapes. These characteris- aphids belonging to the genotypes Burk or PsP 4 that were tics determine the location and timing of damaging pest pop- homozygous for both mutations exhibited the highest insecti- ulations ( Kennedy & Storer, 2000 ). From this point of view, cide resistance level and aphids with the genotype Ivo proba- the availability and distribution of okra and rosella crops in bly the least. This could explain why a shift of genotype the vicinity of the cotton agroecosystem could significantly frequency in favour of Burk1 was observed during the coloni- affect the cotton aphid life system and should be taken into zation process in the cotton field, from the initial infestation account in cotton pest management practices. Cropping of to the peak of infestation in the early season, particularly these two relay host plants could either be avoided or they when dimethoate insecticide was applied on cotton plots. could be used for trap cropping purposes ( Shelton & However, a similar trend was observed in untreated plots. It Badenes-Perez, 2006) and/or as parasitoid inoculated ‘banker’ could be hypothesized that Burk1 winged aphids migrated plants ( Bennison, 1992 ) for biological control of aphids in into these untreated plots from the neighbouring treated plots cotton fields. but also from surrounding farmers’ fields that received insec- The present study of the temporal dynamics of genotypic ticides according to the local cotton pest management pro- diversity of A. gossypii populations in the cotton growing gramme. In northern Cameroon from 2003 onward, area in Cameroon over two consecutive years revealed ex- A. gossypii management has relied essentially on the specific tremely low genetic diversity resulting from the adaptation use of neonicotinoids (acetamiprid, 10 g/ha), but organophos- of the aphids to two heavy selection pressures, the distribu- phates (profenofos, 150 g/ha) and pyrethroids (cypermethrin, tion of host plants and the use of insecticides. Such a shap- 36 g/ha) are still largely deployed throughout the season to ing of genetic variability has been shown in populations of control the bollworms Helicoverpa armigera (Hübner). As the polyphagous aphid M. persicae ( Zamoum et al. , 2005; shown by Brazzle et al. (1997) for Bemisia argentifolii Vorburger, 2006). Bellows & Perring in cotton, insecticides applied to control bollworms may enhance aphid populations. Furthermore, the reduction in clonal diversity could also be the result of a competition between the two genotypes, Burk1 Acknowledgements and Ivo , if we assume that aphids carrying Burk1 have a better We express our sincere gratitude to the cotton company of performance than Ivo on cotton whatever the insecticide pres- Cameroon SODECOTON for its financial support. This sure. Fuller et al. (1999) observed a significant decrease in study would not have been carried out without the scientific clonal diversity over a crop season in A. gossypii populations and material assistance of M. Yaya (LANAVET, Garoua), in commercial cucumber glasshouses and competition between F. Simard (OCEAC/IRD, Yaounde) and N. Leroudier (IPMC/ two Curcurbitaceae-specific genotypes of A. gossypii was ex- CNRS, Sophia-Antipolis). We thank M. Vaissayre (CIRAD) perimentally demonstrated ( Rochat et al., 1999). Nevertheless, for reading the manuscript and we are grateful to four anon- analysis of two A. gossypii samples collected in untreated iso- ymous reviewers for their comments and helpful sugges- lated irrigated cotton fields during the dry season (Maroua, tions. The first two authors of this publication made equal Cameroon, 14 March 2005) and in wild cotton plants contributions. (Ziguinchor, Senegal, 250 km from the cotton-growing area, 22 April 2005) revealed a predominance of the genotype Ivo , with a frequency of 95% and 100%, respectively (T. Brévault, unpublished data). Therefore, it may be hypothesized that, in References the absence of insecticide treatment, the genotype Burk1 has a Ahmad , M. , Arif , M.I. & Denholm , I . ( 2003 ) High resistance of field much lower fitness than Ivo because the insecticide adaptive populations of the cotton aphid Aphis gossypii Glover (Homop- mutations have a cost, as shown for the peach potato aphid tera: Aphididae) to pyrethroid insecticides in Pakistan . Journal of Myzus persicae (Foster et al. , 1999; Guillemaud et al. , 2003 ). Economic Entomology , 96 , 875 – 878 .
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 222 T. Brévault et al.
Andrews , M.C. , Callaghan , A. , Field , L.M. , Williamson , M.S. & Kennedy , G.G. & Storer , N.P . ( 2000 ) Life systems of polyphagous Moores , G.D . ( 2004 ) Identification of mutations conferring insec- arthropod pests in temporally unstable cropping systems . Annual ticide-insensitive AChE in the cotton-melon aphid, Aphis gossypii Review of Entomology , 45 , 467 – 493 . Glover . Insect Molecular Biology , 13 , 555 – 561 . Li , F. & Han , Z.J . ( 2002 ) Two different genes encoding acetylcholin- Ai , Y. , Qiu , X.H. & He , F.Q . (2003 ) A review of insecticide resis- esterase existing in cotton aphid ( Aphis gossypii) . Genome , 45 , tance mechanism of Aphis gossypii . Entomological Knowledge , 1134 – 1141 . 40 , 385 – 391 . Loxdale , H.D. & Lushai , G . ( 1999 ) Slaves of the environment: the Bennison , J.A . ( 1992 ) Biological control of aphids on cucumbers, movement of herbivorous insects in relation to their ecology and use of open rearing systems or “banker plants” to aid establish- genotype . Philosophical Transactions of the Royal Society Series ment of Aphidus matricarieae and Aphidoletes aphidimyza . Me- B, Biological Sciences , 354 , 1479 – 1495 . delingen van de Faculteit Landbouwwetenschappen, Universiteit Moores , G.D. , Gao , X.W. , Denholm , I. & Devonshire , A.L . ( 1996 ) Gent , 57 , 457 – 466 . Characterization of insensitive acetylcholinesterase in insecticide- Blackman , R.L. & Eastop , V.F . ( 1984 ) Aphids on the World’s Crops, resistant cotton aphids, Aphis gossypii Glover (Homoptera: Aphi- An Identification Guide . John Wiley & Sons , U.K . didae) . Pesticide Biochemistry and Physiology , 56 , 102 – 110 . Brandenburg , R.L. & Kennedy , G.G . ( 1982 ) Intercrop relationships Nibouche , S. , Beyo , J. & Brévault , T . ( 2002 ) Negative cross insensi- and spider mite dispersal in a corn/peanut agro-ecosystem . Ento- tivity in a dimethoate resistant strain of cotton aphid Aphis gossy- mologia Experimentalis et applicata , 32 , 269 – 276 . pii Glover in northern Cameroon . Resistant Pest Management Brazzle , J.R., Heinz , K.M. & Parella , M.P . ( 1997 ) Multivariate ap- Newsletter , 12 , 25 – 26 . proach to identifying patterns of Bemisia argentifolii (Homoptera: Pickett , J.A. , Wadhams , L.J. , Woodcock , C.M. & Hardie , J . ( 1992 ) Aleyrodidae) infesting cotton . Environmental Entomology , 26 , The chemical ecology of aphids . Annual Review of Entomology , 473 – 482 . 37 , 67 – 90 . Charaabi , K. , Carletto , J. , Chavigny , P. , Marrakchi , M. , Makni , M. & Powell , G. & Hardie , J . ( 2001 ) The chemical ecology of aphid host Vanlerberghe-Masutti , F . (2008 ) Genotypic diversity of the cotton- alternation: how do return migrants find the primary host plant? melon aphid Aphis gossypii (Glover) in Tunisia is structured by Applied Entomology and Zoology , 36 , 259 – 267 . host plants . Bulletin of Entomological research , DOI: 10.1017/ Powell , G. , Tosh , C.R. & Hardie , J . ( 2006 ) Host plant selection by S0007485307005585/ . aphids: behavioral, evolutionary and applied perspectives . Annual Deguine , J.-P. ( 1996 ) The evolution of insecticide resistance in Aphis Review of Entomology , 51 , 309 – 330 . gossypii Glover (Hemiptera: Aphididae) in Cameroon . Resistant Renou , A. & Deguine , J.P . ( 1992 ) Ravageurs et protection de la cul- Pest Management Newsletter , 8 , 13 – 14 . ture cotonnière au Cameroun . Coton et Fibres Tropicales, Série Deguine , J.-P. , Martin , J. & Leclant , F . ( 1999 ) Extreme polyphagy of ‘Etudes, documents et synthèses’ , 13 , 7 – 52 . Aphis gossypii Glover (Hemiptera: Aphididae) during the dry sea- Rochat , J. , Vanlerberghe-Masutti , F. , Chavigny , P. , Boll , R. & son in northern Cameroon . Insect Science and its Application , 19 , Lapchin , L . ( 1999 ) Inter-strain competition and dispersal in 23 – 36 . aphids: evidence from a greenhouse study . Ecological Entomol- Delorme , R. , Augé , D. , Bethenod , M.-T. & Villatte , F . ( 1997 ) Insecti- ogy , 24 , 450 – 464 . cide resistance in a strain of Aphis gossypii from Southern France . Shelton , A.M. & Badenes-Perez , F.R . ( 2006 ) Concepts and applica- Pesticide Science , 49 , 90 – 96 . tions of trap cropping in pest management . Annual Review of En- Ebert , T.A. & Cartwright , B . ( 1997 ) Biology and ecology of Aphis tomology , 51 , 285 – 308 . gossypii Glover (Homotera: Aphididae) . Southwestern Entomolo- Shelton , A.M. & North , R.C . (1986 ) Species composition and phe- gist , 22 , 116 – 153 . nology of Thysanoptera within field crops adjacent to cabbage Fahmy , T. ( 2006 ) XLSTAT, version 2006. Addinsoft, Paris . fields . Environmental Entomology , 15 , 512 – 519 . Foster , S.P. , Woodcock , C.M. , Williamson , M.S. , Devonshire , A.L. , Simon , J.C. , Carre , S. , Boutin , M. , Prunier-Leterme , N. , Sabater- Denholm , I. & Thompson , R . (1999 ) Reduced alarm response by Munoz , B. , Latorre , A. & Bournoville , R . ( 2003 ) Host-based diver- peach-potato aphids, Myzus persicae (Hemiptera: Aphididae), gence in populations of the pea aphid: insights from nuclear with knock-down resistance to insecticides (kdr) may impose a markers and the prevalence of facultative symbionts . Proceedings fitness cost through increased vulnerability to natural enemies . of the Royal Society of London Series B, Biological Sciences , 270 , Bulletin of Entomological Research , 89, 133 – 138 . 1703 – 1712 . Frantz , A. , Plantegenest , M. , Mieuzet , L. & Simon , J.C . ( 2006 ) Eco- Stinner , R.E. , Barfield , C.S. , Stimac , J.L. & Dohse , L . ( 1983 ) Dis- logical specialization correlates with genotypic differentiation in persal and movement of insect pests . Annual Review of Entomol- sympatric host-populations of the pea aphid . Journal of Evolu- ogy , 28 , 319 – 335 . tionary Biology , 19 , 392 – 401 . Toda , S. , Komazaki , S. , Tomita , T. & Kono , Y . (2004 ) Two amino Fuller , S.J. , Chavigny , P. , Lapchin , L. & Vanlerberghe-Masutti , F . acid substitutions in acetylcholinesterase associated with pirimi- ( 1999 ) Variation in clonal diversity in glasshouse infestations of carb and organophosphorous insecticide resistance in the cotton the aphid, Aphis gossypii, in southern France . Molecular Ecology , aphid Aphis gossypii Glover (Homoptera: Aphididae). Insect 8 , 1867 – 1877 . Molecular Biology , 13 , 549 – 553 . Guillemaud , T. , Brun , A. , Anthony , N . et al . ( 2003 ) Incidence of in- Vanlerberghe-Masutti , F. & Chavigny , P . ( 1998 ) Host-based genetic secticide resistance alleles in sexually-reproducing populations of differentiation in the aphid species Aphis gossypii Glover, evidenced the peach-potato aphid Myzus persicae (Hemiptera: Aphididae) from RAPD fingerprints. Molecular Ecology , 7 , 905 – 914 . from southern France . Bulletin of Entomological Research , 93 , Vanlerberghe-Masutti , F. , Chavigny , P. & Fuller , S.J . ( 1999 ) Charac- 289 – 297 . terization of microsatellite loci in the aphid species Aphis gossypii Hawthorne , D.J. & Via , S . ( 2001 ) Genetic linkage of ecological spe- Glover . Molecular Ecology , 8 , 865 – 702 . cialization and reproductive isolation in pea aphids . Nature , 412 , Via , S . ( 1999 ) Reproductive isolation between sympatric races of 904 – 907 . pea aphids. I. Gene flow restriction and habitat choice . Evolution , Jin , L. & Chakraborty , R . ( 1993 ) Estimation of genetic distance and 53 , 1446 – 1457 . coefficient of gene diversity from single-probe multilocus DNA fin- Via , S. , Bouck , A.C. & Skillman , S . (2000 ) Reproductive isolation gerprinting data . Molecular Biology and Evolution , 11 , 120 – 127 . between divergent races of pea aphids on two hosts. II. Selection
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 Clonal diversity in unstable environment 2 2 3
against migrants and hybrids in the parental environments . Evolu- Zamoum , T. , Simon , J.C. , Crochard , D. , Ballanger , Y. , Lapchin , L. , tion , 54 , 1626 – 1637 . Vanlerberghe-Masutti , F. & Guillemaud , T . ( 2005 ) Does insecti- Vorburger , C . ( 2006 ) Temporal dynamics of genotypic diversity re- cide resistance alone account for the low genetic variability of asexually reproducing populations of the peach-potato aphid veal strong clonal selection in the aphid Myzus persicae . Journal Myzus persicae ? Heredity , 94 , 630 – 639 . of Evolutionary Biology , 19 , 97 – 107 . Wallner , W.E . (1987 ) Factors affecting insect population dynamics: differences between outbreak and non-outbreak species . Annual Accepted 18 November 2007 Review of Entomology , 32 , 317 – 340 . First published online 18 June 2008
© 2008 The Authors Journal compilation © 2008 The Royal Entomological Society, Agricultural and Forest Entomology, 10, 215–223 Bulletin of Entomological Research (2008) 98, 565–573 doi:10.1017/S0007485308005877 Ó 2008 Cambridge University Press Printed in the United Kingdom First published online 1 July 2008
Tracking pyrethroid resistance in the polyphagous bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae), in the shifting landscape of a cotton-growing area
T. Bre´vault1,2 *, J. Achaleke2, S.P. Sougnabe´3 and M. Vaissayre1 1CIRAD, UPR Annual Cropping System, Montpellier, F-34398 France: 2IRAD, PRASAC/ARDESAC, Cotton Research Section, PO Box 415, Garoua, Cameroon: 3ITRAD, PRASAC/ARDESAC, Rainfed Crops Research Program, N’Djamena, Chad
Abstract
In cotton-growing areas of Central Africa, timing of host crops and pest management practices in annual rainfed cropping systems result in a shifting mosaic of habitats that influence the dynamics and resistance of Helicoverpa armigera (Hu¨ bner) populations on spatial scales, both within and across seasons. From 2002 to 2006, regional and local resistance was monitored among cotton fields and among the major host plants of the bollworm. From 2002, pyrethroid resistance increased within and across cotton-growing seasons to reach a worrying situation at the end of the 2005 growing season. Cotton crops played a fundamental role in the increase in seasonal resistance, even if the intensive use of insecticides on local tomato crops strongly concentrated resistance alleles in residual populations throughout the off-season. Due to the relative stability of resistance in H. armigera populations despite a long off-season, we believe that after the dispersal of the moths southwards at the end of the growing season, reverse migration mainly accounts for the reconstitution of populations at the onset of the following growing season. In addition, local resistance monitoring in 2005 and 2006 showed that it was possible to control the increase in resistance by temporarily stopping the use of pyrethroids during the period of peak infestation of cotton by H. armigera. On the other hand, the similar resistance frequency of populations sampled from sprayed and unsprayed synchronous hosts confirmed the absence of reproductive isolation between adults. As a result, diversity in cropping systems should be encouraged by planting alternative host plants to provide a mosaic of habitats, which in return would provide insecticide-free refuges. The implications for insecticide resistance management in annual cropping systems are discussed. Keywords: Helicoverpa armigera, bollworm, insecticide resistance, pyrethroid, life system, host plant, cotton, Central Africa
(Accepted 28 December 2007)
*Author for correspondence: CIRAD, TA B-10/02, Avenue Agropolis, 34 398 Montpellier Cedex 5, France Fax: (+33)467615666 E-mail: [email protected] 566 T. Bre´vault et al.
Introduction Dec tomato Jan maize The use of pyrethroid insecticides to control the boll- worm, Helicoverpa armigera (Hu¨ bner), in cotton-producing Nov Feb areas has, more or less, rapidly led to widespread resistance: g successively Australia (Gunning et al., 1984; Forrester et al., Hyptis 1993), China (Shen et al., 1992), India (McCaffery et al., 1989; Armes et al., 1992; Kranthi et al., 2001), Thailand (Ahmad Oct Mar & McCaffery, 1988), Turkey (Ernst & Dittrich, 1992) and Pakistan (Ahmad et al., 1995, 1997) have been concerned. In Africa, pyrethroid resistance was diagnosed in the 1996–97 season in the southern (Van Jaarsveld et al., 1997) and cotton western regions of the continent (Vassal et al., 1997; Martin Sept Apr et al., 2000) where populations developed metabolic resis- maize tance to pyrethroids via overproduction of cytochrome Rainfed P450-dependent monoxygenases (or MFOs) (Martin et al., crops 2002). In Cameroon, where pyrethroid insecticides had been widely used on cotton for approximately 20 years because of Aug May their efficacy in controlling a wide range of cotton pests at Cleome low cost, a monitoring network was set up in 1999 for the Jul Jun early detection of bollworm resistance to pyrethroids (Bre´vault et al., 2002). Several control failures in farmers’ Fig. 1. Typical host plant succession of Helicoverpa armigera in fields in the 2004 cotton-growing season were undoubtedly the cotton-growing area of Central Africa. Band thickness is due to pyrethroid resistance (Bre´vault & Achaleke, 2005). approximately proportionate to host abundance. g, generation Because of its high polyphagy and its capability of time. dispersing over great distances, H armigera has the potential to rapidly colonize diverse cropping systems and to develop insecticide resistance (Fitt, 1989). In the cotton growing area Similarly, the indiscriminate use of insecticides on commer- of Central Africa, including on northern Cameroon and cial vegetable crops, such as tomatoes, which are mostly Chad, H. armigera is a major pest of cotton and vege- cultivated in small irrigated plots during the dry season, table crops, particularly tomato, and affects production by significantly contributes to the selection of resistant genes. destroying flowering and fruit-bearing organs. Due to the On the other hand, large areas of untreated maize crops and annual sequence of dry and rainy seasons, the agricultural wild hosts may provide relatively insecticide-free refuges landscape, thus, consists of a shifting mosaic of habitats. (Roush, 1989). At the beginning of the rainy season (June to mid-July), a While the relationship between heavy insecticide use and common wild plant that grows in degraded soils and along insecticide resistance in cotton pests have been well docu- roadsides, Cleome viscosa L. (Capparidaceae), provides a mented in several cotton-growing areas worldwide, overall nursery for the first generation of H. armigera (fig. 1). From pyrethroid resistance has not been established at the regional mid-July to the end of October, the bollworm simultaneously scale of Central Africa. In addition, few detailed studies colonizes rainfed crops, such as silking maize (1 generation) have dealt with the impact of the life system of polyphagous and cotton (up to three generations), as a function of the insects (particularly the sequence of host plants) on the evol- pattern of planting dates. Although millet, groundnut, ution of insecticide resistance in annual cropping systems. cowpea and sorghum are known as host plants in several This paper reports the 2002 to 2006 dynamics of pyrethroid different cropping systems in the world, local varieties resistance in field-collected H. armigera larvae from different cultivated in Central Africa are not significantly infested. At sampling sites in the Central African cotton-growing area the end of the rainy season (late October), populations can and through a sequence of crop and wild hosts at a local persist on wild plants, such as Hyptis suaveolens (L.) Poit. scale in northern Cameroon. Understanding the population (Lamiaceae) (Gadpayle et al., 2004). During the dry season dynamics of the polyphagous bollworm in a shifting agri- (late October through March), local populations can persist cultural landscape, both in time and space, is a key step in as diapausing individuals, by dispersing to local small developing sustainable strategies for local and regional irrigated sites under tomato and maize crops or by migrat- insecticide resistance management (IRM). ing over long range distances to find suitable habitats (Nibouche, 1994). H. armigera, thus, exploits a sequence of crops that offer a substantial but time-limited resource Materials and methods (maize and cotton) and uncultivated or scarce hosts (Cleome, Hyptis and tomato) that enable the bollworm to bridge the Vial tests critical period (Nibouche et al., 2007). This succession of host To assess the spatial and temporal patterns of resistance plants is likely to affect temporal and spatial population in the cotton-growing area, H. armigera larvae were collected dynamics (Kennedy & Storer, 2000) and, consequently, the in cotton fields from four sampling areas in Cameroon (fig. 2) gene flow that determines the pattern of insecticide resis- at the beginning (mid-August) and at the end (early October) tance. As a rainfed crop, cotton occupies a considerable part of the infestation period from 2002 to 2006. Four additional of the Central African agricultural landscape. Each growing sampling areas in Chad and Nigeria were monitored to season, the widespread application of cypermethrin in the validate our results at a broader regional scale. Local cotton-growing area exerts continuous selection pressure. resistance was monitored from July 2003 to February 2007 Dynamics of pyrethroid resistance in Helicoverpa armigera 567
Cotton growing area of Central Africa
sampling area
Maroua Chad Kaélé
Gombe Garoua Pala Sarh Nigeria
Moundou Mayo Galké
Cameroon 0 200 400 km
Fig. 2. Location of sampling areas on the survey of cypermethrin resistance of Helicoverpa armigera in the cotton-growing region of Central Africa, including on Nigeria, Cameroon and Chad. by collecting larvae on successive major host plants (fig. 1) expressed in micrograms per gram of larva, was applied in an area located 25 km from Garoua (923 N, 1345 E). on the thorax with an Arnold micro-applicator (Burkard, Resistance to pyrethroid insecticides was assessed by vial UK). A minimum of five doses were applied to build a tests (McCutchen et al., 1989) adapted to bollworm larvae regression line representative of the relation between dosage (Vaissayre et al., 2002). This method is based on the use of and mortality. For each test, a control population was treated glass vials whose internal wall has been coated with an with pure acetone. Treated larvae were kept at a constant insecticide. For this experiment, cypermethrin was chosen as temperature of 25+2 C and a L14 : D10 photoperiod cycle. the active ingredient as the most commonly used pyrethroid Mortality was observed 48 hours later. Larvae were con- in Central Africa. We used technical grade cypermethrin sidered dead if they gave no coordinated response to (94%, Arysta Life Science). Two treatments were tested in the stimulation by touch with a blunt needle. laboratory in 40 ml vials: 10 untreated vials, used as control, and 30 vials, each treated with 30 mg of cypermethrin, a diagnostic dose reported to kill all the individuals of a Data analysis susceptible strain (BK77) and 60–80% of a resistant popu- Vial test results were expressed as the percentage sur- lation collected in Benin in 1997 (Vaissayre et al., 2002). vival, corrected by the mortality observed in the control Fourth instar larvae of between 10 and 15 mm long were using Abbott’s (1925) formula. Square root transformed data collected from host plants not less than five days after the were statistically compared by ANOVA using SAS GLM last insecticide application (if any). They were then placed (SAS Institute, 1989) followed by Duncan’s multiple range individually in vials without food and maintained at test. For topical application bioassays, the analysis of data ambient temperature. The number of dead larvae was re- was performed with the WINDL software (Cirad, France), corded 24 hours later. For each sampling area, a minimum of according to Finney’s log-probit method (Finney, 1971). The two replications were carried out in different villages, except lethal concentration (LC50) was expressed as micrograms of when there were not enough larvae. At least two replications active ingredient per gram of larva. For each strain, the were performed for each survey, depending on the number resistance factor (RF) was calculated by dividing the LC50 of of larvae collected. the tested strain by that of the susceptible reference strain BK77 (Martin et al., 2000).
Topical bioassays Results Each insect colony was established with a minimum of Resistance among cotton fields 300 larvae collected from cotton or tomato crops around Gaschiga and Djalingo, two neighbouring localities in the From 2002 onwards, the annual survey of cotton crops vicinity of Garoua (fig. 2). First generation larvae of each showed a significant increase in the resistance frequency of colony were used for toxicological tests. The insecticide H. armigera cotton populations to cypermethrin (F = 100.5, solution used for the topical application was prepared from P < 0.001), culminating at the end of the 2005 growing season the above-mentioned technical grade insecticide diluted (fig. 3). Pyrethroid resistance increased systematically with- with acetone. Fourth instar larvae were allocated into weight in cotton-growing seasons (F = 45.1, P < 0.001), except in classes from 16 to 40 mg, and then individually held in six 2006. It was also observed that the mean survival rate at the cell-boxes containing artificial diet. A quantity of insecticide beginning of one cotton season (August, year n) was very solution (0.6–1.0 ml), depending on larva weight class similar to that obtained at the end of the previous cotton 568 T. Bre´vault et al.
100 until October 2004, the resistance frequency of bollworm populations from Mayo Galke´ was still significantly lower 90 a a 80 ab than that of areas sampled in the central (Garoua and Pala) 70 and northern area (Maroua) of the cotton-growing region. b b Surveys from October 2005 showed that Mayo Galke´ still 60 c displayed the lowest resistance, but significant differences 50 were only observed between Mayo Galke´ and Moundou. In 40 de cd October 2006, pyrethroid resistance grew significantly worse 30 e in the central part of the cotton growing area, particularly 20 f Pala and Kae´le´. At the beginning of cotton infestation by Mean survival rate (%) 10 H. armigera (August), significant spatial differences were 0 only detected in 2005 and 2006, with the lowest resistance 2002 2003 2004 2005 2006 frequencies observed in samples from Mayo Galke´. (9,9) (8,11) (7,11) (8,13) (12,13)
Fig. 3. Mean survival rate (%)ofHelicoverpa armigera field- collected larvae exposed to 30 mg cypermethrin at the beginning Resistance among host plants (August) and at the end (October) of infestation of cotton crops. To measure seasonal resistance dynamics, local scale Bars indicate maximum value and the numbers in parentheses monitoring was conducted of larvae collected from a se- refer to the total numbers of vial tests conducted in August and October, respectively. Bars followed by different letters are quence of major host plants within 25 km around Garoua. In significantly different (ANOVA, P < 0.05). Numbers in parenth- the 2003/04 season, larvae sampled from Cleome and maize eses indicate the number of vial tests. at the beginning of the rainy season (July and August), presented a low resistance frequency (fig. 4). However, a significant increase of resistance to cypermethrin was season (October, year n–1). On the other hand, the resistance recorded through the sequence of host plants (F = 14.1, frequency of cypermethrin was seen to vary between P < 0.001), including cotton, Hyptis and tomato. Interest- sampling sites (table 1). While no difference was observed ingly, populations sampled from unsprayed Hyptis (October between populations in August 2002, the first sign of 2003) or irrigated maize (December 2003 and February pyrethroid resistance was detected in October of the same 2004) showed comparable resistance frequency to larvae year in the central part of the cotton-growing region from sprayed plants of cotton (F = 0.01, P = 0.909) or tomato (Garoua, Cameroon). In the following growing season (F = 2.7, P = 0.181), respectively. Between March and May (2003), only bollworm populations collected from the south- 2004, the scarcity of host plants and larvae restricted vial ern part, including Mayo Galke´ (Cameroon) and Moundou tests. This dry transition period was associated with a (Chad), could be considered as not resistant (% survival marked decrease in resistance frequency in populations < 20). Although overall resistance continuously increased colonizing Cleome and maize at the beginning of the
Table 1. Mean survival rate (%)ofHelicoverpa armigera field-collected larvae exposed to 30 mg cypermethrin at the beginning (August) and at the end (October) of infestation of cotton crops in the cotton-growing area of Central Africa.
Year Month Nigeria Cameroon Chad F P Gombe Maroua Kae´le´ Garoua Mayo Galke´ Pala Moundou Sarh Aug- 6.9 (3) 6.1 (3) 13.0 (3) 3.6 (3) 1.0 0.455 2002 Oct- 16.3 (5) 19.1 (3) 22.9 (22) 11.2 (3) ab ab a b 3.3 0.035 Aug- 22.7 (3) 17.3 (6) 23.9 (12) 14.4 (6) 2.1 0.134 2003 Oct- 25.0 (2) 29.8 (6) 42.3 (8) 33.5 (33) 13.2 (6) 15.7 (2) b ab a ab c c 10.7 < 0.001 Aug- 19.9 (4) 26.9 (4) 29.7 (10) 20.8 (2) 1.1 0.398 2004 Oct- 40.7 (2) 52.6 (6) 47.4 (7) 53.8 (16) 35.6 (6) 56.7 (2) ab a ab a b a 5.4 0.001 Aug- 56.5 (6) 53.0 (12) 36.3 (5) 61.7 (2) a a b a 4.5 0.014 2005 Oct- 54.9 (2) 61.3 (5) 68.1 (6) 64.5 (12) 54.2 (6) 72.0 (8) 83.3 (2) 69.0 (6) b b ab ab b ab a ab 2.5 0.034 Aug- 43.6 (6) 61.6 (5) 59.3 (13) 45.5 (4) 76.2 (7) 75.9 (2) 65.4 (3) c b b c a a ab 11.5 < 0.001 2006 Oct- 58.3 (2) 54.6 (2) 75.2 (8) 59.7 (29) 57.2 (5) 82.0 (10) 66.5 (6) 66.7 (4) c c ab c c a bc bc 10.6 < 0.001
Values in the same row followed by different letters are significantly different (ANOVA, P < 0.05). Numbers in parentheses indicate the number of vial tests. Dynamics of pyrethroid resistance in Helicoverpa armigera 569
100
90 (5) 80 (8) (2) (2) (2) (9) (4) 70 (6) (4) (3) (4) (8) (8) 60 (3) (8) (8) (8) (10) (17) (2) (8) (7) (8) (8) (11) (8) 50 (2) (2) (10) (2)
40 (8) (2) (8) Mean survival rate (%) 30 (8) (8) (8) (8) (8) 20 (8) 10
0 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 10 11 12 10 11 12 10 11 12 10 11 12 2003/04 2004/05 2005/06 2006/07
Fig. 4. Mean survival rate (%)ofHelicoverpa armigera field-collected larvae exposed to 30 mg cypermethrin through a sequence of crop and wild host plants (within the range of 25 km of Garoua, Cameroon). Bars indicate maximum/minimum values and numbers in parentheses indicate the number of vial tests ( , Cleome; m, maize; L, cotton; K, Hyptis; *, tomato). following rainy season in June 2004 (F = 30.6, P < 0.001). In three last seasons. Finally, except in 2004/05, the resistance the 2004/05 season, a significant increase in resistance was frequency decreased during the dry transition period from also recorded through the cotton-tomato sequence (F = 19.1, March to June. P < 0.001), although a significant decrease in resistance fre- To complete the field survey, larval resistance was quency was observed on irrigated tomato crops from assessed in the laboratory by topical application on reared December 2004 to February 2005 (F = 11.0, P = 0.045). On the F1 colonies originating from cotton or tomato crops in two other hand, the similarity in the frequency of resistant villages situated within a range of 25 km from Garoua. LC50 larvae from synchronously sprayed and unsprayed host measurements confirmed the results of vial tests with plants was confirmed in the case of the cotton-maize in a significant increase of resistance in wild populations of August 2004 (F = 0.4, P = 0.536), the cotton-Hyptis in October H. armigera across years and with a marked increase in 2004 (F = 1.9, P = 0.194) and the tomato-irrigated maize in resistance in populations sampled on tomato (table 2). February 2005 (F = 2.3, P = 0.226) but not the tomato-maize in December 2004 (F = 8.5, P = 0.043). The 2005/06 and 2006/07 seasons recorded a new pattern. At the beginning of the Discussion rainy season, first populations sampled on Cleome presented a high resistance frequency. For the first time, no further Resistance among cotton fields increase in cypermethrin resistance was observed in larvae The 2002 field survey of pyrethroid resistance in the sampled on cotton (2004/05, F = 3.0, P = 0.103; and 2005/06, cotton bollworm, H. armigera, revealed pyrethroid resistance F = 0.02, P = 0.890), whereas a significant increase in resis- in some local populations of the bollworm, H. armigera,in tance was observed in larvae sampled on tomato in Cameroon. By 2004, resistance frequency of field-sampled December (2004/05, F = 19.6, P < 0.001; and 2005/06, F = 8.2, larvae had gradually increased to reach a worrying situation. P = 0.017). Resistance frequency in tomato populations Laboratory bioassays on field-sampled populations showed significantly decreased from December to February (F = 6.9, that most control failures reported by farmers in 2004 were P = 0.022; F = 5.4, P = 0.036) in the 2004/05 and 2005/06 definitely due to pyrethroid resistance. Subsequent monitor- seasons, but not in the 2006/07 season (F = 0.24, P = 0.636). ing results indicated that pyrethroid resistance in H. armigera Based on this four-year study, we observed that the over- concerned the entire cotton-growing area of Central Africa. all resistance frequency of local H. armigera populations The cotton pest management programme in northern significantly increased (F = 27.3, P < 0.001), depending on Cameroon was designed to provide small-scale cotton the season (F = 49.7, P < 0.001, 2003/04 < 2004/5 < 2005/06 = growers with a simple and cheap calendar-based spraying 2006/07) and on the host plant (F = 27.4, P < 0.001, Cleome = programme. Approximately six insecticide treatments are maize < cotton < Hyptis < tomato), and that there was a applied. The spraying programme begins from about day significant interaction between the two parameters (F = 4.2, 45 after seedlings emerge (usually at squaring), with two P < 0.001). Unexpectedly, cotton significantly selected consecutive endosulfan applications at biweekly interval, H. armigera larvae for cypermethrin resistance during the followed, irrespective of the pest status or level of infesta- first two seasons only. Tomato systematically selected tion, by applications of a pyrethroid (usually cypermethrin) resistance until December, but resistance frequency did not associated with an organophosphate (usually profenofos) increase and sometimes even decreased in February in the (Vaissayre et al., 1984; Ochou et al., 1998). In Chad, 570 T. Bre´vault et al.
Table 2. Response of Helicoverpa armigera populations to topically applied cypermethrin.
x1 Season Native host Date of Location LC50 mgg 95% confidence Group RF plant sampling larva interval Cotton Oct–02 Djalingo 10.1 6.0–21.5 e 14 2002/03 Djalingo 4.2 1.8–6.4 e 6 Tomato Feb–03 Gaschiga 5.3 1.5–7.7 e 8 Cotton Oct–03 Djalingo 10.5 4.9–15.3 de 15 2003/04 Tomato Feb–04 Gaschiga 24.8 19.5–37.8 bc 35 Djalingo 28.8 14.5–43.2 bcd 41 Cotton Oct–04 Gaschiga 24.1 17.0–61.6 bc 34 2004/05 Dec–04 Gaschiga 24.3 17.3–29.1 c 35 Tomato Feb–05 Gaschiga 33.6 11.3–65.4 bcd 48 Djalingo 38.8 20.1–56.1 bc 55 Cotton Oct–05 Gaschiga 44.0 27.0–71.0 bc 63 2005/06 Djalingo 1010 115–8873 a 1443 Tomato Dec–05 Gaschiga 82.0 34.7–194 b 117
x1 LC50: lethal concentration expressed in mgg larva. RF (resistance factor) = LC50 of the tested strain/LC50 of the reference strain BK77. Groups were built based on overlapping confidence intervals. pyrethroids-profenofos mix is used throughout the cotton- alleles in the absence of a significant biological fitness cost growing season, while control methods in Nigeria are and reconstitution of populations at the onset of the rainy unknown. In South India, although recommended in- season from diapausing populations vs. migrating popu- secticides were becoming inefficient, the use of alternative lations; (ii) maintenance of selection pressure by the use of chemistry that would have dramatically increased the cost of pyrethroids on intermediate cultivated host plants, such as pest control was rejected in small-scale farming systems, tomato crops; and (iii) disappearance of refuge patches with where farmers preferred to increase the number of sprays. the continual reduction of natural ecosystems in the land- To control the high pressure from resistant H. armigera scape (Madden et al., 1995; De Souza et al., 1995). In Australia, larvae, farmers had applied over 30 sprays as opposed to the Daly & Fitt (1990) considered that fitness cost was irrelevant recommended 6–10 sprays (Ramasubramanian, 2004). In to explain the rapid off-season drop in resistance frequen- Australia, conventional cotton production relies on chemical cies. In India, De Souza et al. (1995) noted that diapause- pesticides for pest control with 8–15 sprays applied to inducing conditions occurred before spraying against conventional crops (Fitt, 2000). H. armigera had been completed and, in many areas, coincided Why did resistance increase so rapidly without any with a peak density. These authors suggested that this significant changes in the agricultural landscape or in phenomenon may have led to a lower resistance frequency insecticide use in previous years? It can be hypothesized in the proportion of the diapausing population, i.e. a refuge that resistant genes were recently introduced by resistant in time as opposed to a refuge in space. Their model showed immigrant moths from West Africa, where resistance to that a significant drop in resistance could be achieved with pyrethroids in H. armigera was confirmed earlier (Vassal only 1–2% of the H. armigera population entering diapause. et al., 1997; Martin et al., 2000). In the 1998–2002 period, vial tests with 30 mg cypermethrin performed on field-collected Resistance among host plants larvae at the end of the cotton season gave survival rates of 1–35% in Ivory Coast (Martin et al., 2003), whereas values Field monitoring at a local scale across the sequence of above 20% were obtained in neighbouring countries such as host plants showed that resistance frequency can be used as Mali, Burkina Faso and Benin (Vaissayre et al., 2002). If we a marker to assess the gene flow of H. armigera between host consider that control is very likely to fail as soon as survival plants. The similarity between the resistance frequency rate exceeds 20%, then most sampling sites in Cameroon of larvae collected at the same time from treated and should have been considered as pyrethroid-resistant from untreated hosts, such as cotton-maize (August), cotton- the end of the 2003 growing season, with the exception of the Hyptis (October) and tomato-maize (December and Febru- southern cotton-growing area (Mayo Galke´). Knowing that ary), as well as the annual trend, confirmed the hypothesis of the use of pyrethroid was homogeneous throughout the the absence of reproductive isolation between populations cotton-growing area of Cameroon, this particular situation sampled on different host plants (Achaleke et al., 2005). is probably explained by landscape components where Furthermore, Martin et al. (2002) showed, in Ivory Coast, that predominant maize crops and natural ecosystems served the same resistance mechanism (MFO-mediated) was as insecticide-free refuges, which would have diluted re- involved in strains from different host plants, particularly sistance frequency in the latter area. Resistance worsened in cotton and tomato. 2005, followed by a relative stabilization in October 2006. Results from the 2003 and 2004 surveys of host plants Generally, the resistance of larvae samples did not decrease confirmed the key role of cotton in the seasonal increase in between the end of one season (October n) and the sub- resistance. It can be assumed that dispersal of H. armigera sequent early season (August n+1), despite a long dry occurs only at particular times during or after the cropping season of six to seven months without cotton crops. This season, when resources are no longer available, which stability could be explained by different non-exclusive eventually modifies resistance patterns across the cotton- factors that remain to be explored: (i) persistence of resistance growing area. Han et al. (1999), in China, and Madden et al. Dynamics of pyrethroid resistance in Helicoverpa armigera 571
(1995), in India, reported a seasonal pattern with a maximum colonizing Cleome at the beginning of the rainy season. Long at the end of the cropping season. Similarly, in West Africa, range migration from more humid areas, associated with air the resistance level to pyrethroids was generally highest movements (Inter Tropical Front), may be more consistent, in populations collected at the end of cotton treatments as shown by the relative spatial homogeneity of resistance and decreased during the dry season (Martin et al., 2003). frequency in populations sampled on cotton at the beginning However, in 2005 and 2006, our results showed that local of infestation (August). resistance frequency did not increase within the cotton Temporal positioning of host crops and pest management season, probably due to the temporary exclusion of cyper- practices in annual cropping systems result in a shifting methrin. In Cameroon, this strategy for the management mosaic of habitats that influence resistance dynamics at of pest resistance have been implemented since the 2005 local and regional scales, both within and across seasons cotton season in some cotton areas where resistance had (Kennedy & Storer, 2000; Sequeira & Playford, 2001). This caused severe pest outbreaks. In 2006, cypermethrin was information can help improve system-based resistance temporarily replaced by endosulfan or indoxacarb in the management by preserving the efficacy of both synthetic whole cotton area from September 9 to 22, the period which and natural insecticides, including Bt toxins expressed by coincides with the onset of a predictable peak in H. armigera transgenic crops (Gustafson et al., 2006). The results of the infestation of cotton. On the other hand, due to unfavourable study clearly showed that the window strategy that pre- conditions, the sowing of maize was abnormally delayed scribes endosulfan (mid-July to mid-August) at the begin- and extended in both years; so that, susceptible stages of ning of the cotton spraying programme did not overcome cotton and maize were very synchronous throughout resistance in Cameroon. According to our results, strategies September (J. Achaleke, unpublished data). If the usual for managing resistance should exclude the use of pyre- planting dates had been respected, cotton would have been throids on one generation if cotton is the only available host the only available host in September. We hypothesize that plant in the agricultural landscape (window from early maize might have served as a consistent refuge acting as a September to early October), or on the last generation (mid- reservoir of non-resistant moths (De Souza et al., 1995; Han September to mid-October) when populations experience et al., 1999; Wu et al., 2004). bottlenecks in food resources associated with southward In our survey, a significant increase in resistance emigration. Furthermore, alternative insecticide with no frequency was recorded in tomato fields in irrigated sites. cross resistance (spinosad, indoxacarb, etc.) should be more In 2005, strong resistance to cypermethrin was observed widely used to control H. armigera in high value cash crops, in a population collected on Gaschiga (RF > 1000). Vege- such as tomato. In Australia, the threat of pyrethroid table crops that are mostly located in the vicinity of towns resistance in H. armigera is countered by a maximum of offer conditions that are favourable for the rapid evolution of two consecutive pyrethroids sprays to coincide with peak resistance. Indeed, irrigated sites form very small ‘oases’ bollworm damage (Forrester et al., 1993). In the absence of surrounded by desert landscapes that probably prevent gene reproductive isolation in populations of different hosts, flow from outside while insecticides that are generally diversified cropping systems should be encouraged with the intended for cotton are indiscriminately applied on tomato planting of alternative host plants (e.g. maize, sorghum, crops. However, these crops offer a weak carrying capacity cowpea, sunflower, etc.) to provide a greater mosaic of for adult production due to high temperatures, high in- habitats, which, in return, would increase insecticide-free secticide use and limited acreage. The decreased resistance refuges or trap crops (Sequeira, 2001) with enhanced bio- observed in February 2005 and 2006 may have resulted from logical control (Fitt, 2000; Grundy et al., 2004). On the other the emergence of late season diapausing moths from cotton hand, uncultivated hosts, such as Cleome and Hyptis, which fields. serve as nurseries at the beginning and at the end of the The resistance frequency of the populations colonizing cropping season, should be systematically destroyed prior to Cleome at the beginning of the rainy season was quite the flowering stage. similar to that of cotton populations in the previous year. Priority actions should focus on the rational and This observation indicates that populations colonizing concerted use of pesticides in cotton areas at the regional Cleome may be recruited via diapausing individuals that scale, with cooperation between the agrochemical industry, survive the hot and dry season and emerge with early rains. research and extension services in communicating with and However, Nibouche (1994) showed that diapause occurring explaining strategies to farmers (Denholm & Rowland, 1992; in H. armigera in West Africa was of limited duration and did Martin et al., 2005). Resistance in H. armigera populations not allow the pest to survive the whole dry season. In fact, should be monitored at the onset of the rainy season on the pest possibly survives the dry season by migrating. Cleome to evaluate the initial resistance frequency and, Rainfed crop areas are re-colonized at the beginning of every subsequently, across the season to evaluate programmes. On rainy season (May) by moths produced by populations that the whole, reduced reliance on insecticides should be based colonize vegetable crops located at varying distances from on sampling systems and thresholds, to better target sprays the northern region during the off-season. The reverse (Silvie et al., 2001), as well as on the use of more selective migration (from rainfed crops grown in the north to vege- insecticides and the integration of a range of non-chemical table crops southwards) occurs at the beginning of the dry tactics (Fitt, 2000; Vaissayre et al., 2006). season (November). To conclude, the increment acreage of More fundamental research is needed to understand the cotton crops linked to the disappearance of insecticide-free genetic basis of resistance mechanisms, including fitness refuges in the agricultural landscape may have contributed cost, and the role of ecological factors that favour the in- to a certain extent to strengthening the selection of resistant crease in frequency and the further spread of resistance alleles. Even if some irrigated sites, where tomato crops alleles in H. armigera populations. In particular, detailed are grown, strongly concentrate resistance, densities are knowledge is required on the origin of populations that first probably too low to contribute significantly to populations colonize Cleome at the beginning of the rainy season and, 572 T. Bre´vault et al. subsequently, colonize cotton (diapause vs. immigration) Ernst, G. & Dittrich, V. (1992) Comparative measurements of and, more generally, on the role of selection, drift and resistance to insecticides in three closely related Old and migration in the genetic structuring of H. armigera popu- New World bollworm species. Pesticide Science 34, 147–152. lations at a regional scale. Such results could help identify Finney, D.J. (1971) Probit Analysis. 3rd edn, 333 pp. London, practical ways of reducing insecticide exposure and increase Cambridge University Press. the use of host plants as potential refuges in IRM prog- Fitt, G.P. (1989) The ecology of Heliothis species in relation to rammes for sustainable cotton production. agroecosystems. Annual Review of Entomology 34, 17–52. Fitt, G.P. (2000) An Australian approach to IPM in cotton: integrating new technologies to minimise insecticide dependence. Crop Protection 19, 793–800. Acknowledgements Forrester, N.W., Cahill, M., Bird, L.J. & Layland, J.K. (1993) Management of pyrethroid and endosulfan resistance in We express our sincere gratitude to the Cameroon cotton Helicoverpa armigera (Lepidoptera: Noctuidae) in Australia. development corporation SODECOTON for providing the Bulletin of Entomological Research (Suppl.) 1, 1–132. financial and technical assistance that permitted the gather- Gadpayle, J.G., Rajgure, Vennila, S., Bambawale, O.M., Deole, ing of field data on the resistance status of H. armigera.We S.A., Karanjkar, P.P., Panchbhai, P.R. & Biradar, V.K. also acknowledge PRASAC-ARDESAC Project who enabled (2004) New record of Hyptis suaveolens Poit. (Lamiaceae completion of the regional aspect of this work. We are (Labiatae): Lamiales) as a host plant of Helicoverpa armigera. grateful to the two anonymous reviewers for comments and Annals of Plant Protection Sciences 12, 441–442. helpful suggestions. Grundy, P.R., Sequeira, R.V. & Short, K.S. (2004) Evaluating legume species as alternative trap crops to chickpea for management of Helicoverpa spp. (Lepidoptera: Noctuidae) References in central Queensland cotton cropping systems. Bulletin of Abbott, W.S. (1925) A method of computing the effectiveness of Entomological Research 94, 481–486. an insecticide. Journal of Economic Entomology 18, 265–267. Gunning, R.V., Easton, C.S., Greenup, L.R. & Edge, V.E. (1984) Achaleke, J., Bre´vault, T., Blondin, L. & Vassal, J.M. (2005) Synthetic pyrethroid resistance in Heliothis armigera Interbreebing in Helicoverpa armigera populations from (Hubner) in Australia. Journal of Economic Entomology 77, different host plants estimated by resistance status and 1283–1287. microsatelite markers. CD-ROM, 5 pp. in 7e`me Confe´rence Gustafson, D.I., Head, G. & Caprio, M.A. (2006) Modeling the Internationale sur les Ravageurs en Agriculture. ANPP, 26–27 impact of alternative hosts on Helicoverpa zea: adapatation October 2005, Montpellier, France. to Bollgard cotton. Journal of Economic Entomology 99, 2116– Ahmad, M. & McCaffery, A.R. (1988) Resistance to insecticides 2124. in Thailand strain of Heliothis armigera (Hu¨ bner) (Lepidop- Han, Z., Wang, Y., Zhang, Q., Li, X. & Li, G. (1999) Dynamics of tera: Noctuidae). Journal of Economic Entomology 81, 45–48. pyrethroid resistance in a field population of Helicoverpa Ahmad, M., Arif, M.I. & Ahmad, Z. (1995) Monitoring in- armigera in China. Pesticide Science 55, 462–466. secticide resistance of Helicoverpa armigera (Lepidoptera: Kennedy, G.G. & Storer, N.P. (2000) Life systems of poly- Noctuidae) in Pakistan. Journal of Economic Entomology 88, phagous arthropod pests in temporally unstable cropping 771–776. systems. Annual Review of Entomology 45, 467–493. Ahmad, M., Arif, M.I. & Attique, M.R. (1997) Pyrethroid Kranthi, K.R., Jadhav, D., Wanjari, R., Kranthi, S. & Russel, D. resistance of Helicoverpa armigera (Lepidoptera: Noctuidae) (2001) Pyrethroid resistance and mechanisms of resistance in Pakistan. Bulletin of Entomological Research 87, 343–347. in field strains of Helicoverpa armigera (Lepidoptera: Armes, N.J., Jadhav, D.R., Bond, G.S. & King, A.B.S. (1992) Noctuidae). Journal of Economic Entomology 94, 253–263. Insecticide resistance in Helicoverpa armigera in South India. Madden, A.D., Holt, J. & Armes, N.J. (1995) The role of Pesticide Science 34, 355–364. uncultivated hosts in the spread of pyrethroid resistance in Bre´vault, T. & Achaleke, J. (2005) Status of pyrethroid resistance Helicoverpa armigera populations in Andhra Pradesh, India: in the cotton bollworm, Helicoverpa armigera, in Cameroon. a simulation approach. Ecological Modelling 82, 61–74. Resistant Pest Management Newsletter 15, 4–7. McCaffery, A.R., King, A.B.S., Walker, A.J. & El-Nayir, H. Bre´vault, T., Asfom, P., Beyo, J., Nibouche, S. & Vaissayre, M. (1989) Resistance to synthetic pyrethroids in the bollworm (2002) Assessement of Helicoverpa armigera resistance to Heliothis armigera from Andhra Pradesh, India. Pesticide pyrethroids in northern Cameroon. Mededelingen Faculteit Science 27, 65–76. Landbouwkundige en Toegepaste Biologische Wetenschappen, McCutchen, B.F., Plapp, F.W., Nemec, S.J. & Campanhola, C. Universiteit Gent 67, 641–646. (1989) Development of diagnostic monitoring techniques Daly, J.C. & Fitt, G.P. (1990) Resistance frequencies in over- for larval pyrethroid resistance in Heliothis spp. (Lepidop- wintering pupae and the first spring generation of Heli- tera: Noctuidae) in cotton. Journal of Economic Entomology coverpa armigera (Lepidoptera: Noctuidae): selective 82, 1502–1507. mortality and immigration. Journal of Economic Entomology Martin, T., Ochou Ochou, G., Hala N’Klo, F., Vassal, J.M. & 83, 1682–1688. Vaissayre, M. (2000) Pyrethroid resistance in the cotton Denholm, I. & Rowland, M.W. (1992) Tactics for managing bollworm, Helicoverpa armigera (Hu¨ bner), in West Africa. pesticide resistance in arthropods: theory and practice. Pest Management Science 56, 549–554. Annual Review of Entomology 37, 91–112. Martin, T., Chandre, F., Ochou, O.G., Vaissayre, M. & Four- De Souza, K., Holt, J. & Colvin, H. (1995) Diapause, migration nier, D. (2002) Pyrethroid resistance mechanisms in the and pyrethroid-resistance dynamics in the cotton boll- cotton bollworm Helicoverpa armigera (Lepidoptera: Noctui- worm, Helicoverpa armigera (Lepidoptera: Noctuidae). Ecolo- dae) from West Africa. Pesticide Biochemistry and Physiology gical Entomlogy 20, 333–342. 74, 17–26. Dynamics of pyrethroid resistance in Helicoverpa armigera 573
Martin, T., Ochou Ochou, G., Vaissayre, M. & Fournier, D. crops in central Queensland: implications for resistance (2003) Monitoring of insecticide resistance in Helicoverpa management. Australian Journal of Entomology 40, 264–269. armigera (Hubner) from 1998 to 2002 in Cote d’Ivoire, West Shen, J., Tan, J., Zhou, B., Jin, C. & Tan, F. (1992) Pyrethroid Africa. Resistant Pest Management Newsletter 12, 51–55. resistance in Heliothis armigera (Hubner) (Lepidoptera: Martin, T., Ochou, O.G., Djihinto, A., Traore, D., Togola, M., Noctuidae) in China. Resistant Pest Management Newsletter Vassal, J.M., Vaissayre, M. & Fournier, D. (2005) Control- 4, 22–24. ling an insecticide resistance bollworm in West Africa. Silvie, P., Deguine, J.P., Nibouche, S., Michel, B. & Vaissayre, Agriculture, Ecosystems and Environment 107, 409–411. M. (2001) Potential of threshold-based interventions for Nibouche, S. (1994) Cycle e´volutif de Helicoverpa armigera cotton pest control by small farmers in West Africa. Crop (Hu¨ bner, 1808) (Lepidoptera, Noctuidae) dans l’Ouest du Protection 20, 297–301. Burkina Faso: biologie, e´cologie et variabilite´ ge´ographique Vaissayre, M., Sement, G. & Trijau, J.P. (1984) Aspects des populations. PhD thesis, Ecole Nationale Supe´rieure phytosanitaires de la culture cotonnie`re en Coˆte d’Ivoire, Agronomique, Montpellier, France. d’apre`slere´seau d’essais a` trois niveaux de protection. Nibouche, S., Gue´rard, N., Martin, P. & Vaissayre, M. (2007) Coton et Fibres Tropicales 39, 1–7. Modelling the role of refuges for sustainable management Vaissayre, M., Vassal, J.M., Irving, S. & Staetz, C. (2002) A vial of dual-gene Bt cotton in West African smallholder farming test method for the survey of pyrethroid resistance in systems. Crop Protection 26, 828–836. Helicoverpa armigera in West Africa. Resistant Pest Manage- Ochou, O.G., Martin, T. & Hala, N.F. (1998) Cotton insect pest ment Newsletter 12, 20–22. problems and management strategies in Coˆte d’Ivoire, Vaissayre, M, Ochou, G.O., Hema, O.S. & Togola, M. (2006) West Africa. pp. 833–837 in Proceedings of the world Cotton Quelles strate´gies pour une gestion durable des ravageurs Research Conference, 6–12 September, Athens, Greece. du cotonnier en Afrique subsaharienne? Cahiers Agricul- Ramasubramanian, T. (2004) Magnitude, mechanism and tures 15, 80–84. management of pyrethroid resistance in Helicoverpa armi- Van Jaarsveld, M.J., Basson, N.C.J. & Marais, P. (1997) gera Hubner in India. Journal of Entomology 1, 6–11. Synthetic pyrethroid resistance in field strains of Helicoverpa Roush, R.T. (1989) Designing resistance management programs: armigera (Hu¨ bner) (Lepidoptera: Noctuidae) in South how can you choose? Pesticide Science 26, 423–441. Africa. African Plant Protection 4, 15–18. SAS Institute (1989) SAS/STAT User’s guide, version 6. 4th Vassal, J.M., Vaissayre, M. & Martin, T. (1997) Decrease in the edn., vol. 2. 846 pp. Cary, NC, USA, SAS Institute Inc. susceptibility of Helicoverpa armigera (Hu¨ bner) (Lepidop- Sequeira, R. (2001) Inter-seasonal population dynamics and tera: Noctuidae) to pyrethroid insecticides in Coˆte d’Ivoire. cultural management of Helicoverpa spp. in a central Resistant Pest Management Newsletter 9, 14–15. Queensland cropping system. Australian Journal of Experi- Wu, K., Feng, H. & Guo, Y. (2004) Evaluation of maize as a mental Agriculture 41, 249–259. refuge for management of resistance to Bt cotton by Sequeira, R. & Playford, C. (2001) Abundance of Helicoverpa Helicoverpa armigera (Hu¨bner) in the Yellow River cotton- (Lepidoptera: Noctuidae) pupae under cotton and other farming region of China. Crop Protection 23, 523–530. Evolutionary Applications
Evolutionary Applications ISSN 1752-4571
ORIGINAL ARTICLE Assessing the role of non-cotton refuges in delaying Helicoverpa armigera resistance to Bt cotton in West Africa
Thierry Bre´ vault,1,2 Samuel Nibouche,3 Joseph Achaleke4 and Yves Carrie` re2
1 CIRAD, UPR 102, F-34398 Montpellier, France 2 Department of Entomology, University of Arizona, Tucson, AZ, USA 3 CIRAD, UMR PVBMT, F-97410 Saint-Pierre, La Re´ union, France 4 IRAD, PRASAC-ARDESAC, Garoua, Cameroon
Keywords Abstract Bacillus thuringiensis, biogeochemical markers, insect resistance management, Non-cotton host plants without Bacillus thuringiensis (Bt) toxins can provide genetically engineered crops, polyphagous refuges that delay resistance to Bt cotton in polyphagous insect pests. It has pest, Bt cotton, refuge strategy. proven difficult, however, to determine the effective contribution of such refuges and their role in delaying resistance evolution. Here, we used Correspondence biogeochemical markers to quantify movement of Helicoverpa armigera moths Thierry Bre´ vault, CIRAD, TA B 102/02, Avenue from non-cotton hosts to cotton fields in three agricultural landscapes of the Agropolis, 34398 Montpellier cedex 5, France. Tel.: +33 4 67 61 55 85; fax: +33 4 67 56 66; West African cotton belt (Cameroon) where Bt cotton was absent. We show e-mail: [email protected] that the contribution of non-cotton hosts as a source of moths was spatially and temporally variable, but at least equivalent to a 7.5% sprayed refuge of Received: 14 July 2011 non-Bt cotton. Simulation models incorporating H. armigera biological param- Accepted: 7 August 2011 eters, however, indicate that planting non-Bt cotton refuges may be needed to First published online: 7 October 2011 significantly delay resistance to cotton producing the toxins Cry1Ac and Cry2Ab. Specifically, when the concentration of one toxin (here Cry1Ac) doi:10.1111/j.1752-4571.2011.00207.x declined seasonally, resistance to Bt cotton often occurred rapidly in simula- tions where refuges of non-Bt cotton were rare and resistance to Cry2Ab was non-recessive, because resistance was essentially driven by one toxin (here Cry2Ab). The use of biogeochemical markers to quantify insect movement can provide a valuable tool to evaluate the role of non-cotton refuges in delaying the evolution of H. armigera resistance to Bt cotton.
Busseola fusca (Fuller), the fall armyworm, Spodoptera fru- Introduction giperda (J.E. Smith), the pink bollworm, Pectinophora gos- Cotton is widely grown in West Africa, where it helps sypiella (Saunders), and the cotton bollworms, sustain millions of resource-poor farmers and rural com- Helicoverpa zea (Boddie) and H. punctigera (Wallengren), munities. Transgenic cotton producing the Bacillus thur- respectively, evolved resistance to Cry1Ab corn in South ingiensis (Bt) toxins Cry1Ac and Cry2Ab was recently Africa, Cry1F corn in Puerto Rico, Cry1Ac cotton in introduced to Burkina Faso (James 2009) to increase agri- India, Cry1Ac and Cry2Ab in the United States, and cultural profitability. Such Bt cotton is referred to as a Cry2Ab in Australia (Van Rensburg et al. 2007; Tabashnik ‘pyramid’ because it produces two distinct Bt toxins et al. 2008, 2009; Bagla 2010; Carrie`re et al. 2010; Downes active against some lepidopteran pest species (Roush et al. 2010). In turn, field-evolved resistance was reported 1998; Showalter et al. 2009). Insect resistance, however, to result in increased crop damage by B. fusca, H. zea, can reduce the effectiveness of Bt crops and is therefore a S. frugiperda, and P. gossypiella (Matten et al. 2008; major concern for the long-term sustainability of Bt Tabashnik et al. 2009; Monsanto 2010; Storer et al. 2010). crops. Indeed, some populations of the cereal stem borer, Furthermore, monitoring data from China and India
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 53 Resistance to Bt cotton Bre´ vault et al. indicate that the frequency of resistance to Cry1Ac cotton resistance (Feng et al. 2010). In West Africa, the absence is increasing in some populations of H. armigera (Hu¨b- of genetic structure among H. armigera populations ner) (Liu et al. 2009; Tabashnik et al. 2009), a major pest observed by Nibouche et al. (1998) and Vassal et al. of cotton throughout the West African cotton belt, where (2008) suggests significant moth movement (>500 km) it has already evolved resistance to pyrethroid insecticides from southern regions to the cotton belt at the beginning (Martin et al. 2005; Bre´vault et al. 2008). of the growing season in June–July and reverse migration Management of insect resistance to Bt crops requires south at the end of the growing season in October– production of abundant susceptible individuals in refuges November. A small proportion of moths also enter dia- of non-Bt host plants that disperse and mate with the pause locally during the dry season (Nibouche 1994). As rare resistant survivors in Bt fields (Gould 1998; Tabash- documented in Agrius convolvuli L. (Lepidoptera: Sphingi- nik et al. 2008, 2009; Carrie`re et al. 2010). Because the dae) (Bowden 1973), migrating moths probably follow cotton bollworm, H. armigera, is polyphagous and mobile the seasonal movements of the intertropical convergence (Forrester et al. 1993; Bre´vault et al. 2008; Vassal et al. zone. 2008), non-cotton host plants in West Africa could The purpose of this study was to evaluate the effective reduce the reliance on refuges of non-Bt cotton to delay contribution in space and time of non-cotton refuges to resistance. Here, non-cotton host plants refer to a ‘non- the pool of H. armigera moths in three agricultural land- structured refuge’ (i.e., host crops and wild host plants), scapes of the West African cotton belt (Cameroon), using as opposed to a ‘structured refuge’ (i.e., non-Bt cotton biogeochemical markers to quantify movement of moths planted as part of a licensing agreement). While some from non-cotton hosts to cotton fields throughout the studies have evaluated the production of H. armigera by cropping season, prior to the introduction of Bt cotton. non-cotton host plants elsewhere (Green et al. 2003; Wu We also used a two-locus population genetics model et al. 2004; Ravi et al. 2005; Baker et al. 2008), movement incorporating realistic estimates of key H. armigera bio- of moths from non-cotton hosts to cotton fields has logical parameters and seasonal decline of Cry1Ac pro- never been quantified in space and time. Nevertheless, it duction to evaluate how short- and long-range movement is often assumed that cotton refuges are not required to from non-cotton refuges may affect the evolution of resis- delay H. armigera resistance to Cry1Ac/Cry2Ab cotton in tance to Cry1Ac/Cry2Ab cotton in each agricultural land- agroecosystems where small fields of diversified crops and scape. Results indicated that supplementing non-cotton patches of non-cultivated hosts are close together (Ravi refuges with refuges of non-Bt cotton would provide a et al. 2005; Wu and Guo 2005; Huang et al. 2010; Liu robust strategy to delay the evolution of H. armigera et al. 2010; Qiao et al. 2010), such as in West Africa. resistance to Bt cotton in West Africa. Simulation models suggest that pyramided plants have the potential to delay resistance more effectively than sin- Materials and methods gle-toxin plants used sequentially or in mosaics, even with relatively small refuges (Roush 1998; Zhao et al. 2003). Sampling These models, however, assume that production of both Three sampling locations, where all cotton grown was toxins Cry1Ac and Cry2Ab remains constant throughout non-Bt cotton, were selected in Cameroon to represent the growing season at levels that kill most target insects. the typical range of conditions encountered in the West Nevertheless, the concentration of Cry1Ac in cotton gen- African cotton belt (Fig. 1, Table 1). The agricultural erally declines when plants start producing flowers and landscape (cultivated vs. uncultivated area) and the abun- bolls, while Cry2Ab levels could remain more constant dance of cotton in the cropping system (Guider > Djal- (Adamczyk et al. 2001; Bird and Akhurst 2005; Kranthi ingo > Tchollire´) differed significantly between the three et al. 2005; Olsen et al. 2005; Showalter et al. 2009; Carri- sampling locations. At each location, moths were cap- e`re et al. 2010). Accordingly, the seasonal decline in the tured with six pheromone (97% (Z)-11-hexadecenal concentration of one toxin (here Cry1Ac) could invalidate and 3% (Z)-9-hexadecenal) traps (Biosyste`mes, Cergy one of the fundamental assumptions of the pyramid strat- Pontoise, France) modified from the Hartstack nylon-mesh egy (i.e., the killing of insects resistant to one toxin by 60-cm-diameter cone trap (Hartstack et al. 1979). One trap another toxin) and thus accelerate resistance evolution was set per cotton field, and traps were separated by a (Carrie`re et al. 2010). Furthermore, as pointed out by distance of 0.5–2 km. Traps were inspected daily to pre- Bourguet et al. (2010), long-range migration has received serve the quality of moths, and pheromone lures were little attention in theoretical models of resistance evolu- changed every 2 weeks. Eighteen moth collections tion. Yet, the dilution effect of resistance alleles that (6 months, three locations) were performed from June to migrating moths such as H. armigera could exert on local November 2006 (N = 3380). Moths were preserved in 95% populations may significantly delay the evolution of ethanol and stored at )20°C for subsequent analyses.
54 ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 Bre´ vault et al. Resistance to Bt cotton
West African cotton belt sampling location
N B. FASO GU DJ TC
CAMEROON 0 150 300 km
Figure 1 Sampling locations of Helicoverpa armigera moths in Cameroon (GU, Guider; DJ, Djalingo; TC, Tchollire´ ). Transgenic cotton producing Bt toxins was recently introduced to West Africa, but only in Burkina Faso. Other Bt crops such as Bt corn have not yet been released in West Africa.
Table 1. Main agronomic characteristics of the three sampling locations: Guider, Djalingo, and Tchollire´ .
Percentage Cotton area of cotton Annual per farmer area planted Seed-cotton rainfall (mm) Cotton area (ha) (ha) Jun-30 yield (kg/ha) Sampling location Main crops and landscape 2006 Avg.* 2006 Avg. 2006 Avg. 2006 Avg. 2006 Avg.
Guider Sorghum, corn, cotton, peanut 1122 1042 14 280 14823 0.6 0.6 67 47 947 1139 Djalingo Peanut, corn, cotton, sorghum 942 983 10 054 11 084 0.7 0.9 89 83 924 1043 Tchollire´ Corn, peanut, cotton, wildlife reserve 967 1159 3197 2999 0.8 0.9 98 86 1065 1154
Information on cropping systems was obtained from SODECOTON data (Direction de la production agricole, Garoua, Cameroun (2006)) in a circu- lar area (25 km radius) around each sampling location. *Average of 2004–2007 growing seasons. Although crops such as sorghum and peanut are known hosts of Helicoverpa armigera in several regions of the world, varieties grown in West Africa are seldom infested. Corn is generally 3–4 times more abundant than cotton in the cotton belt of Cameroon.
then lyophilized for 30 min to remove remaining mois- Biogeochemical analyses ture. The remainder of the moth was placed in a separate Moths were analyzed for 13C/12C carbon isotope signa- ethanol-filled vial for subsequent gossypol analyses. Each tures of natal host photosynthetic type (Deniro and forewing (approximately 1 mg) was tightly folded into a Epstein 1978; Gould et al. 2002) and gossypol (Rojas 5 · 9 mm tin capsule (ThermoQuest, Milan, Italy), indi- et al. 1992), a phytochemical which is uniquely produced vidually placed in a 96-well plate and assigned a specific in the lysigenous glands of cotton (Gossypium spp.) and number. Automated isotopic ratio mass spectrometric closely related species (Jaroszewski et al. 1992). Gossypium analyses were conducted at the Scotland Research Insti- arboreum and G. herbaceum (A genome), G. barbosanum tute (SCRI, Dundee, UK). Forewings were combusted, and G. anomalum (B genome), and G. barbadense are and constituent gases were separated on a gas chromato- occasionally found in West Africa, but these plants are graph column linked to a mass spectrometer. The output rare compared to cultivated G. hirsutum (Valicek 1979). from the mass spectrometer analysis is a ratio, which can For isotope signatures, one forewing of each moth was be converted to a d13C value using Pee Dee Belemnite clipped off and placed on paper towels for 30 min at (PDB) as a reference (Hood-Nowotny and Knols 2007). ambient temperature to enable ethanol to evaporate and Wings from moths reared on common weeds Cleome
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 55 Resistance to Bt cotton Bre´ vault et al. viscosa L. (Capparidaceae) and Hyptis suaveolens (L.) Poit. migrants colonizing the cotton-growing area in June–July (Lamiaceae), as well as on field-grown tomato, cotton, or primarily comprised susceptible individuals, unless ele- corn in the laboratory had d13C values ranging from vated frequency of resistance alleles occurred in the cot- )28.6& to )26.3& (N=15), )27.0& to )25.6& ton belt owing to use of Bt cotton there and important (N=15), )27.7& to )23.4& (N=30), )25.1& to movement occurred between the cotton belt and southern )23.0& (N=15), and )9.3& to )8.2& (N = 15), regions in October–November. respectively. There was no overlap between C3- and C4 The model has two main compartments: the rain-fed (corn)-reared moths. Results from these analyses enabled host plants in the cotton belt and the off-season host us to classify any moth with a value of less than )20.0& plants in southern regions. Both compartments exchanged as having fed on a C3 plant and any moth with a value moths by migration in June–July and October–November. of more than )15.0& as having fed on a C4 plant. The percentage of moths migrating from the south and Moth abdomens were analyzed at Monsanto labs contributing to the first generation in the cotton belt is (Monsanto Company, Creve Coeur, MO) for bound gos- MR1. The percentage of moths emigrating from the cot- sypol using high-pressure chromatography coupled with a ton belt and contributing to the first generation in the triple quadrupole mass spectrometer (Orth et al. 2007). southern regions is MR2 (Figs 2D and S1). Rain-fed host Gossypol was always detected in moths reared on cotton plants in the cotton belt encompass three subcompart- in the laboratory (N = 15). Furthermore, moths reared in ments: non-cotton refuges (subcompartment 1; cultivated the laboratory on C. viscosa (N = 15) and H. suaveolens and wild non-cotton hosts), non-Bt cotton refuges (sub- (N = 15), as well as on field-grown tomato (N = 30) and compartment 2), and Bt cotton (subcompartment 3). As corn (N = 15) had no detectable levels of gossypol. In cultivated landscapes in West Africa usually form a mosaic analyses of moths trapped in cotton fields, <2% of blank of small fields, we assumed random mating between moths samples (i.e., without moths) yielded false-positive results originating from the three subcompartments. (i.e., 1 of 51). We categorized host plants as cotton According to Gustafson et al. (2006), the number of (which is a C3 plant), non-cotton C3 plants (e.g., weeds moths produced per surface area in each subcompartment such as Cleome spp. and Hyptis sp. and tomato), and C4 of a region during one generation, the total number of plants (e.g., corn). Results from gossypol analyses were produced moths, is as follows: confirmed by isotopic ratio analyses with an accuracy of M ¼ A :E LS:LB ð1Þ 99.5%. A total of 658 moths were analyzed for both stable h h h h carbon isotopes and gossypol residues (Table S1). where Ah is the relative area of the region occupied by the host type h (subcompartment), Eh is the relative (to unsprayed non-Bt cotton, i.e., E = 1) number of effective Simulation model 2 eggs (eggs that would produce reproductive adults in the The population genetics model (Fig. S1) incorporated absence of mortality owing to Bt toxin or insecticide estimates of the key biological parameters for H. armigera sprays), LS is the proportion of larvae surviving insecti- to simulate changes in the frequency of two resistance cide sprays (only in non-Bt cotton fields), according to a alleles owing to selection by Cry1Ac/Cry2Ab cotton. It calendar-based spraying program commonly used in West was adapted from the model of Nibouche et al. (2007) Africa (Vaissayre et al. 2006; Bre´vault et al. 2009), and and specifically incorporated data on moth movement LBh considers the proportion of larvae surviving ingestion between non-cotton refuges and cotton fields obtained in of the Bt toxins (only in Bt cotton fields) and fitness cost this study. The model was written in R version 2.8.1 (R (on all host types), averaged across the nine genotypes Development Core Team R 2008). We estimated the time (ss1ss2, ss1rs2, ss1rr2, rs1ss2, rs1rs2, rs1rr2, rr1ss2, rr1rs2, to resistance as the number of years required for H. armi- and rr1rr2—where s and r stand for susceptibility and gera survival to exceed 20% on Bt cotton (Sawicki 1987). resistance alleles, respectively) and weighted by their rela- The evolution of resistance was modeled over four gen- tive abundance: erations per growing season, from July to October, based LB ¼ LB : f ð2Þ on the life cycle of H. armigera on cotton in West Africa h X hg hg g (Nibouche et al. 2007; Bre´vault et al. 2008). The model also accounted for immigration of moths from southern where LBhg is the survival of genotype g on host type h and regions to the cotton belt in June–July, and initiation of fhg is the relative abundance of genotype g on host type h. diapause in the cotton belt or emigration south in Survival of genotype g on Cry1Ac/Cry2Ab cotton dur- October–November (Nibouche 1994; Nibouche et al. ing the course of the growing season was calculated from 1998). We assumed that Bt crops were not cultivated in empirical data (see Table 2 and parameter estimation southern regions (James 2009). Accordingly, the pool of below), according to Finney’s formula (1971):
56 ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 Bre´ vault et al. Resistance to Bt cotton
(A) Guider Refuges C4 plants Non-cotton C3 plants (B) Djalingo 100 100
75 75
50 50 % moths % moths
25 25
0 0 Jun. Jul. Aug. Sep. Oct. Nov. Jun. Jul. Aug. Sep. Oct. Nov.
(C) Tcholliré (D) Host plant sequence Corn (C4) 100
75
MR1 Cotton MR2 50 (C3/gossypol+)
% moths Weeds (C3) Weeds 25 (C3)
Jun. Jul. Aug. Sep. Oct. Nov. 0 Jun. Jul. Aug. Sep. Oct. Nov. Moth emergence from cotton
Figure 2 (A–C) Moths trapped in cotton fields (%) that originated from non-cotton host plants. Remaining moths (100 – % indicated by bar) originated from cotton. Moths were trapped at three locations (Guider, Djalingo, and Tchollire´ ) in Cameroon in 2006. (D) Typical sequence of He- licoverpa armigera host plants in the West Africa cotton belt throughout the cropping season. Curves represent temporal occurrence and relative area of host plants.
In the presence of Bt cotton, the proportion of moths LB3g ¼ LB3g1:LB3g2:LFCg ð3Þ produced by refuges (fref, see Fig. S1) is obtained from eqn (1) as follows: where LB3g1 and LB3g2 are survival of genotype g to the Cry1Ac and Cry2Ab toxins, respectively, and LFCg takes A1 :E1 :LB1 þPref :AC:LS: LB2 into account fitness costs associated with resistance to f ¼ ref : :LB þ :AC: LS: LB þðÞ1À :AC:E :LB Cry1Ac (see parameter estimation below). Also, survival of A1 E1 1 Pref 2 Pref 3 3 genotype g on non-Bt hosts considered fitness costs (see ð6Þ parameter estimation below): Combining eqns (5 and 6) results in the following:
LB1g ¼ LB2g ¼ LFCg ð4Þ fref ¼ : : : fnoncot LS LB1þð1Àf noncotÞ ðÞPref LSLB2 The relative area of cotton in the region is AC, and the : : : : : : : fnoncot LS LB1þð1Àf noncotÞ ½Pref LS LB2þðÞ1ÀPref E3 LB3 relative area of cotton devoted to non-Bt cotton is Pref. ð7Þ Accordingly, the relative area planted to non-Bt cotton (A2) and Bt cotton (A3) is, respectively, A2 = Pref. AC and Eqn (7) allows the calculation of the monthly propor- ) A3 =(1 Pref). AC. In the absence of Bt cotton, Pref =1 tion of moths produced in refuges in the presence of Bt and the observed proportion of moths produced by cotton based on the observed percentage of gossypol- non-cotton hosts (fnon cot) is obtained from eqn (1) as positive moths quantified in this study (Fig. 2 and S1). follows: Thus, data on area or carrying capacity of the different
A1 : E1 host plants are not needed to calculate the proportion of fnon cot ¼ ð5Þ moths produced by refuges. A1 : E1 þ AC : LS
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 57 Resistance to Bt cotton Bre´ vault et al.
Table 2. Standard values of empirical parameters used to model the evolution of Helicoverpa armigera resistance to Bt cotton at three locations in the cotton belt of Cameroon. Sensitivity analyses were performed to evaluate effects of variation in several of these parameters (see Materials and methods).
Parameter Definition Value References fnoncot Proportion of moths originating from non-cotton hosts Fig. 1 Present study LS Survival of larvae to insecticide sprays in non-Bt cotton 0.20 Bre´ vault et al. (2009)
LB3 ss1* Survival of ss1 larvae on Cry1Ac cotton in August, September, and October 0.02, 0.17, 0.37 Kranthi et al. (2005) c, h Fitness cost and dominance of cost (Cry1Ac) 0.34, 0.33 Bird and Akhurst (2004) Fitness cost and dominance of cost (Cry2Ab) 0.00, 0.00 Mahon and Young (2010)
E3 Number of effective eggs produced by adults surviving on Bt cotton 0.60 Mahon and Olsen (2009)
(relative to non-Bt cotton E2) p0 Initial allele frequency (Cry1Ac) 0.0003 Mahon et al. (2007b) Initial allele frequency (Cry2Ab) 0.0033 Mahon et al. (2007b) RF,b Resistance factor and slope (Cry1Ac) 63, 1.0 Akhurst et al. (2003) Resistance factor and slope (Cry2Ab) 6830, 0.76 Mahon et al. (2007a)
DLC Dominance of resistance (Cry1Ac) 0.26 Akhurst et al. (2003) Dominance of resistance (Cry2Ab) 0.00 Mahon et al. (2007a) MR1 Proportion of moths migrating from southern regions and 0.98 colonizing the cotton belt MR2 Proportion of moths from the cotton belt contributing to 0.20 the pool of migrants moving south
*Survival on Bt cotton. The West African cotton belt is colonized at the beginning of the growing season (June–July) by moths migrating from the south, and moths from the cotton belt return south at the end of the growing season (October–November). Modification of the frequency of r alleles in the cotton homozygote, 1 for heterozygote, and 0 for susceptible belt owing to the immigration of moths from southern homozygote). As a result of the absence of selection pres- regions at the beginning of June was computed as follows: sure on non-Bt hosts, in accordance with eqn (3), the equation is the same for refuge subcompartments 1 (non- 0 : : f cbrk ¼ fcbrk ðÞ1 À MR1 þfsrrk MR1 ð8Þ cotton hosts) and 2 (non-Bt cotton).
The frequency of the rk allele in moths parents of a where fcbr is the frequency of the rk allele in moths of k generation was computed as follows: the cotton belt prior to immigration, f 0cbrk the frequency modified by immigration, and fsrri the frequency of the frk ¼ fref : frk1 þðÞ1 À fref : frk2 ð11Þ rk allele in moths migrating from the southern regions. Modification of the frequency of r alleles in the southern where fref is the proportion of moths produced in refuges regions owing to the immigration of moths from the cotton (eqn 7), and frk1 and frk2 the frequency of rk allele in moths belt at the end of October was computed as follows: emerging, respectively, from refuges and Bt cotton (eqn 10). Because the observed proportion of moths produced by 0 : : f srrk ¼ fsrrk ðÞ1 À MR2 þfcbrk MR2 ð9Þ non-cotton hosts sometimes resulted from small samples (Table S1), we used Monte Carlo simulations (Peterson where fsrrk is the frequency of the rk allele in moths of and Hunt 2003) to assess the impact of uncertainty in esti- 0 the southern regions prior to immigration, f srrkthe fre- mating this parameter on the number of years to achieve quency modified by immigration, and fcbrkthe frequency >20% survival on Cry1Ac/Cry2Ab cotton. In the Monte of the rk allele in moths migrating from the cotton belt. Carlo simulations, the proportion of moths produced by The frequency of rk allele in moths emerging from ref- non-cotton hosts (fnon cot) in a given generation and region uges or from Bt cotton was computed as follows: was sampled repeatedly from a Student’s t distribution. : : This distribution was computed (rt random generation P LBhg fhg ng g function in R) from the observed proportion of moths frkh ¼ : : ð10Þ 2 P LBhg fhg produced by non-cotton hosts and the sample sizes used g in each generation and region in the study (Table S1, where LBhg is the survival of genotype g in subcompart- Fig. 2). The random values generated by the Monte Carlo ment h computed from eqns (3) or (4), fhg the frequency procedure were used in 1000 simulations to evaluate the of genotype g in eggs of the considered generation, and trajectory of resistance and the variability of this trajectory ng the number of rk alleles in genotype g (2 for resistant in each region.
58 ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 Bre´ vault et al. Resistance to Bt cotton
all subcompartments, where h is the dominance of fit- Parameter estimation ness cost (Table 2).
We used published data to estimate model parameters to The relative number of effective eggs (Eh) depends on simulate the evolution of resistance at each of the three attractiveness of host plants for oviposition, fecundity of sampling locations (Table 2). The seasonal decline in adults that oviposit on the crop, and survival of larvae Cry1Ac concentration in Cry1Ac/Cry2Ab cotton resulted and pupae in the absence of Bt toxins or insecticides. We in a significant increase in survival of a H. armigera strain assumed that E3 = 0.6 (E2 = 1) to account for the limited with high frequency of a field-derived allele conferring fecundity of moths originating from Bt cotton (Mahon resistance to Cry2Ab (Mahon and Olsen 2009). We and Olsen 2009). Initial frequency of the Cry1Ac and assumed that mortality of ss1 larvae to Cry1Ac in Bt cot- Cry2Ab resistance alleles was set to 0.0003 and 0.0033, ton (LB3ss1) decreased throughout the growing season as respectively, according to Mahon et al. (2007b), but reported by Kranthi et al. (2005). Mortality of ss2 larvae higher values (0.003 and 0.033) were also modeled. to Cry2Ab in Bt cotton was then calculated to reproduce Pheromone trapping data and gossypol analyses the seasonal change in survival of genotype ss1ss2 on support the hypothesis that some moths migrate south to Cry1Ac/Cry2Ab cotton (Table S2) observed by Mahon non-cotton hosts instead of diapausing locally during and Olsen (2009). the dry season. Large trap catches in the cotton belt in Survival of genotypes rs1 and rr1 to Cry1Ac and rs2 the early growing season and in the southern regions and rr2 to Cry2Ab was computed with standard dose– at the end of the cotton-growing season cannot be mortality regressions as in the study of Nibouche et al. explained by local emergence. MR1 could be high because (2007). A theoretical concentration of Cry1Ac or Cry2Ab only 2% of moths trapped in the cotton belt in the early toxin in Cry1Ac/Cry2Ab cotton was calculated, given the growing season contained gossypol. Data on gossypol assumed mortality of ss1 or calculated mortality of ss2 content of moths trapped in the southern region from (Table 2). This theoretical concentration was then used to October to December indicate that MR2 could be below calculate the mortality of rs1, rs2, rr1, and rr2, given the 20% (Table 2). We also used a sensitivity analysis to resistance factor (RF), the slope of the dose–mortality determine the effect of migration (MR1 and MR2 = 0.1 regression (b), and the dominance of resistance for the and 0.9) on the evolution of resistance. lethal concentration LC50 (DLC : from 0 to 1, where 0 = completely recessive and 1 = completely dominant Results resistance). For Cry1Ac, we used RF = 63 and b = 1 and assumed that partially recessive resistance (DLC = 0.26) Movement of H. armigera moths from non-cotton hosts was the most likely scenario (Akhurst et al. 2003). For to cotton fields Cry2Ab, we used RF = 6830 and b = 0.76 and assumed Most moths trapped early in the growing season (June– that completely recessive resistance (DLC = 0) was the July) had signatures of C3 (79.7–88.3% of moths) and C4 standard level of dominance (Mahon et al. 2007a). (6.7–18.6%) non-cotton plants, but very few gossypol- Data indicate that genetic background of H. armigera positive moths were detected (Fig. 2A–C). When the first strains and characteristics of cotton plants affect the dom- moth generation emerged from cotton (August), 87.0– inance of resistance to Cry1Ac and Cry2Ab, which can 93.8% of moths still had signatures of C3 and C4 non- vary from recessive to partially dominant (Akhurst et al. cotton plants (Fig. 2A–C). The contribution of non-cot- 2003; Bird and Akhurst 2004, 2005; Mahon et al. 2007a, ton refuges to the pool of moths trapped in cotton fields 2008; Wu et al. 2009; Nair et al. 2010). We thus used a decreased during the second (September) and third sensitivity analysis to determine the effect of dominance (October) generations, particularly at Djalingo (20.0– of resistance to each toxin (DLC = 0, 0.1, 0.3, and 0.5) on 7.5%), and to a lesser extent at Tchollire´ (62.5–22.2%) the evolution of resistance. Survival of the genotypes and Guider (65.2–45.0%). At cotton harvest (November), under the various combinations of dominance was calcu- most moths originated from non-cotton C3 plants at lated with the RF and b values used above. Djalingo (93.1%) and Tchollire´ (96.6%), whereas moths Based on results from published studies, we assumed from cotton still contributed significantly to the pool of non-recessive fitness costs of resistance to Cry1Ac moths (50.0%) at Guider (Fig. 2A–C). and no fitness costs of resistance to Cry2Ab (Bird and Akhurst 2004, 2007; Mahon and Olsen 2009; Mahon Evolution of H. armigera resistance to Cry1Ac/Cry2Ab and Young 2010). Survival of the rr1 genotypes was cotton reduced by a factor LFCrr1 =1) c in all subcompart- ments, where c is the fitness cost. Survival of the rs1 Simulations showed that the evolution of resistance was genotypes was corrected by a factor LFCrs1 =1) hc in primarily driven by Cry2Ab resistance alleles, as the initial
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 59 Resistance to Bt cotton Bre´ vault et al. resistance allele frequency and the dominance of Cry1Ac Guider Djalingo Tcholliré resistance had little effect on the number of years to (A) 40 achieve >20% survival on Cry1Ac/Cry2Ab cotton, except in some cases when inheritance of resistance to Cry2Ab 30 was completely recessive (Tables S3 and S4). While the resistance allele r1 did contribute to survival on Cry1Ac/ 20 Cry2Ab cotton (Table S2; compare, for example, survival of ss1rr2 and rs1rr2), the fact that r1 did not appreciably 10 affect the time to resistance is not surprising because the Years to resistance ratio of survival of ss1 larvae on Cry1Ac cotton was above 20% in October when the concentration of Cry1Ac was 0 0 10 20 30 40 50 low (Table 2; see parameter LB3 ss1). Thus, the presence of Non-Bt cotton refuges (%) Cry2Ab was primarily responsible for the low survival of susceptible insects on Cry1Ac/Cry2Ab cotton throughout (B) 40 the growing season (i.e., from 0.6% in July to 7% in Octo- ber; Table S2). Results outlined below are therefore largely 30 insensitive to the initial frequency of resistance to Cry1Ac and the dominance of resistance to this toxin. Among-site variability affected the role of non-cotton 20 refuges in delaying resistance evolution (Fig. 3A,B; Table S3). In the absence of refuges (including non-cotton 10 refuges), resistance evolved in 3 years or less, except when resistance to Cry2Ab was completely recessive (dominance 0 of resistance, DLC = 0) and initial frequency of Cry2Ab 0 10 20 30 40 50 Non-Bt cotton refuges (%) resistance (p0) was 0.0033. With completely recessive resis- tance to Cry2Ab (dominance of resistance, DLC = 0), non- Figure 3 Effect of the abundance of sprayed refuges of non-Bt cot- cotton refuges were sufficient to delay resistance ‡9 years at ton (%) on the evolution of Helicoverpa armigera resistance to the three locations, irrespective of the initial frequency of Cry1Ac/Cry2Ab cotton at three locations in Cameroon: Guider (•), Cry2Ab resistance (Table S3). With partially recessive resis- Djalingo (h), and Tchollire´ (D). Simulations considered data on move- tance to Cry2Ab (DLC = 0.1) and initial resistance allele ment between non-cotton refuges and cotton fields measured at frequency of 0.0033 to Cry2Ab, non-cotton refuges delayed each site (Fig. 2A–C). For Cry2Ab, the initial resistance allele fre- resistance ‡32 years at Guider, ‡16 years at Tchollire´, and quency was 0.0033 (A) or 0.033 (B), and resistance was partially ‡8 years at Djalingo (Fig. 3A). With partially recessive recessive (DLC = 0.1, dashed line) or semi-dominant (DLC = 0.5, solid line). For Cry1Ac, the initial resistance allele frequency was 0.0003, resistance to Cry2Ab (DLC = 0.1) and higher initial resis- and resistance was partially recessive (DLC = 0.26) (Table 2). The crite- tance allele frequency of 0.033 to Cry2Ab, however, resis- rion for resistance evolution was >20% survival on Bt cotton. tance evolution was faster and non-cotton refuges delayed resistance ‡17 years at Guider, ‡9 years at Tchollire´, and £6 years at Djalingo (Fig. 3B). With higher dominance of Resistance evolution was significantly affected by pat-
Cry2Ab resistance (DLC = 0.3 or 0.5), sprayed refuges of terns of migration (Fig. 4, Table S6). When many moths 20% non-Bt cotton in addition to non-cotton refuges migrated north into the cotton belt but few returned south delayed resistance ‡8 years at Guider, £11 years at Tchol- (MR1 = 0.98, MR2 = 0.2 or MR1 = 0.90, MR2 = 0.1), lire´, and £8 years at Djalingo (Fig. 3B, Table S3). In a southern migrants diluted the frequency of resistance alleles worst-case scenario with an initial resistance frequency of and delayed resistance. Long-range migration, however, 0.033 and semi-dominant resistance to Cry2Ab did not delay resistance when many moths from the cotton
(DLC = 0.5), sprayed refuges of 50% non-Bt cotton delayed belt returned south (MR2 = 0.9), or the pool of migrants resistance 15 years at Guider, 8 years at Tchollire´, and from the south was small compared to the population over- 6 years at Djalingo (Fig. 3B). Monte Carlo simulations wintering in the cotton area (MR1 = 0.1). incorporating variability in the proportion of moths pro- duced by non-cotton hosts (f ) during each H. armi- non cot Discussion gera generation revealed similar trends in resistance evolution and confirmed that resistance evolution differed The adoption of transgenic Bt cotton in West Africa between sampling locations and according to the domi- raises novel and important issues related to the sustain- nance of resistance (compare Tables S3 and S5 or Fig. 3). ability of such technology in small-scale cropping systems.
60 ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 Bre´ vault et al. Resistance to Bt cotton
MR1/MR2 0/0 0.9/0.1 0.9/0.9 0.1/0.9 0.1/0.1 (e.g., C. viscosa) and early-planted corn. At this time, few 40 gossypol-positive moths were detected, and the few posi- tive moths trapped in cotton fields likely originated from 30 overwintering pupae or possibly from cotton left in fields from the previous growing season. The contribution of 20 non-cotton host plants to the pool of moths trapped in cotton fields decreased during the second (September) 10 and third (October) generations, particularly at Djalingo, Years to resistance and to a lesser extent at Tchollire´ and Guider. Given the 0 abundance of cotton in Guider > Djalingo > Tchollire´,a 0 1020304050greater abundance of non-cotton hosts in Guider than in Non-Bt cotton refuges (%) Djalingo could explain why there were proportionally more cotton-produced moths in Djalingo than in Guider. Figure 4 Effect of the abundance of sprayed refuges of non-Bt cot- At cotton harvest (November), most moths likely origi- ton (%) on the evolution of Helicoverpa armigera resistance to Cry1Ac/Cry2Ab cotton at Djalingo (Cameroon). Simulations consid- nated from late season weeds (e.g., H. suaveolens) at Djal- ered data on movement between non-cotton refuges and cotton ingo and Tchollire´, possibly reflecting high larval fields and patterns of long-range migration (Fig. 2). For Cry2Ab, the mortality in the last H. armigera generation on cotton, initial resistance allele frequency was 0.0033, and resistance was diapause, or reverse migration southward. At Guider, semi-dominant (DLC = 0.5). For Cry1Ac, initial resistance allele fre- where cotton is usually planted a few weeks later, moths quency was 0.0003, and resistance was partially recessive from cotton still contributed significantly to the pool of (D = 0.26) (Table 2). MR1 is the proportion of moths from southern LC moths. regions colonizing the cotton belt in June–July; MR2 the proportion of moths from the cotton belt contributing to the pool of migrants mov- Our seasonal assessment of H. armigera movement ing south in October–November. The criterion for resistance evolution indicates that non-cotton refuges were equivalent to was >20% survival on Bt cotton. Results of simulations for MR1/MR2 ‡7.5% non-Bt cotton refuges treated with insecticides values of 0/0 were almost identical to results obtained for 0.9/0.9, throughout the cotton-growing season. In simulations, 0.1/0.9 and 0.1/0.1. corn-produced moths were not distinguished from moths produced in other refuge types. However, from a manage- ment perspective, evaluation of moths from corn was Because H. armigera is polyphagous and highly mobile, it important because it is often assumed that moths from is often assumed that refuges of non-cotton host crops corn represent a large proportion of the pool of moths and wild host plants provide sufficient refuges to delay originating from non-cotton hosts and corn could be the evolution of resistance, thus reducing or even sup- used as a non-structured refuge. Even if non-cotton hosts pressing the need of non-Bt cotton refuges (Ravi et al. such as corn were important sources of susceptible moths 2005; Wu and Guo 2005; Liu et al. 2010; Qiao et al. at the three studied locations, moth production was not 2010). However, movement of H. armigera from non-cot- temporally synchronous with emergence of moths from ton hosts to cotton fields had never been quantified Bt cotton fields, especially during the second (September) directly. Rather, studies assessing the refuge potential of and third (October) generations. Accordingly, the pres- alternative host plants primarily compared insect densities ence of abundant non-cotton hosts in the agricultural between non-Bt cotton and non-cotton host crops (Green landscape does not imply that non-cotton hosts can pro- et al. 2003; Wu et al. 2004; Ravi et al. 2005; Baker et al. vide sufficient numbers of Bt-susceptible moths to effec- 2008), although such comparisons do not take into tively delay resistance to Bt cotton. Provided that fitness account movement from non-cotton hosts to cotton fields costs high or non-recessive, non-cotton hosts such as or the overall contribution of the non-cotton hosts sur- corn could, however, play a significant role in delaying rounding cotton fields. Our study addressed these resistance evolution. problems by quantifying movement between all potential Using the same biogeochemical markers as we did here, non-cotton hosts and cotton fields, in three contrasted Head et al. (2010) reported a low relative contribution of agricultural landscapes before the commercial release of moths from cotton (i.e., < c.a. 40% for any trapping Bt cotton. date) to H. zea populations near cotton fields during the Results show variability in the moth production of dif- period of H. zea emergence from Bt cotton in Arkansas, ferent host plants among sampling locations and through- North Carolina, and Mississippi. They also found that C4 out the cropping season. As expected, most moths hosts contributed > c.a. 15% of the H. zea moths trapped trapped in the early season had signatures of C3 and C4 on any given date during the period of moth emergence non-cotton plants, indicating sources from seasonal weeds from cotton in Arkansas, Georgia, Louisiana, Mississippi,
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 61 Resistance to Bt cotton Bre´ vault et al. and North Carolina. Based on these data, Head et al. resistance to the toxin Cry2Ab in H. armigera (Mahon (2010) concluded that refuges of non-Bt cotton will play and Olsen 2009; Mahon and Young 2010), or empirical a minor role in the management of resistance to Bt cot- data on seasonal changes in the movement of H. armigera ton. Because moth populations can decline at sites where from non-cotton host plants to cotton fields and on use of Bt crops is high (Carrie`re et al. 2003, 2004, 2010; regional variation in the contribution of non-cotton ref- Wu et al. 2008; Hutchison et al. 2010) and biogeochemi- uges. Here, we found a low efficacy of the pyramid strat- cal markers provide relative measures of the source egy when the concentration of Cry1Ac declined during potential of various refuges without addressing whether the growing season, resistance to Cry2Ab was non-reces- local moth populations are large enough to delay resis- sive, and only non-cotton refuges were available, despite tance, caution should be exerted when using biogeochem- the important but temporally and regionally variable ical markers to assess the role of particular refuges in moth contribution from non-cotton hosts to putative Bt regions where Bt crops are used. If the area occupied by cotton fields. Under the first two conditions, our results cotton and non-cotton refuges is small compared to the indicate that refuges of non-Bt cotton would be needed area occupied by Bt cotton, the production of moths to significantly delay resistance unless high and sustained from refuges could be insufficient to delay resistance and movement from non-cotton refuges to cotton fields refuges that are a relatively low source of moths could occurred during the growing season (e.g., Guider) or still be needed. long-range migration was more important northward Theory underlying the pyramid strategy predicts that than southward. two-toxin cotton will be most effective for delaying the While some H. armigera individuals overwinter in the evolution of resistance when each Bt toxin kills most sus- West African cotton belt, others immigrate from southern ceptible pests and resistance to each toxin is recessive regions to colonize the cotton-growing area in June–July throughout the growing season, abundant refuges and fit- or emigrate south from the cotton belt in October– ness costs are present, and selection with either of the November (Nibouche 1994; Nibouche et al. 1998). toxins does not cause cross-resistance to the other (Gould Because the extent of H. armigera migration and its vari- 1998; Gould et al. 2006; Tabashnik et al. 2008, 2009; ability remain poorly known, research on this topic could Showalter et al. 2009). Our model considered the seasonal be invaluable for the development of resistance manage- decline in mortality of a strain resistant to Cry2Ab on ment strategies in West Africa. Furthermore, it will also Cry1Ac/Cry2Ab cotton (Mahon and Olsen 2009), which be critical to assess the effects of seasonal changes in the paralleled the decline in Cry1Ac concentration generally production of Cry1Ac and Cry2Ab in African cotton cul- observed in Bt cotton during the course of the growing tivars on the survival and dominance of resistance in season. Such reduction in mortality of Cry2Ab-resistant H. armigera. Despite current uncertainty about these insects on Cry1Ac/Cry2Ab cotton invalidates one of the parameters, our empirical and simulation results suggest fundamental assumptions of the pyramid strategy, i.e., the that the use of non-Bt cotton refuges will enhance the killing of insects resistant to one toxin by the other toxin, management of H. armigera resistance to Bt cotton in and thus could accelerate resistance evolution (Carrie`re West Africa. et al. 2010). Seasonal declines in Cry1Ac-induced mortal- ity and more stable Cry2Ab-induced mortality, as mod- Conclusions eled here, necessarily generate stronger selection for resistance to Cry2Ab than Cry1Ac. Thus, it is not surpris- The evolution of resistance in target pests such as ing that our simulations showed that resistance to pyram- H. armigera could cut short the profitability of Bt cotton ided two-toxin Bt cotton was primarily driven by the in West Africa. The adoption of Bt cotton is expected to evolution of resistance to Cry2Ab when the concentration reduce and simplify pest management problems, includ- of Cry1Ac declined during the growing season. ing pyrethroid resistance in the cotton bollworm H. armi- In previous modeling work, based on simulations of a gera. A 3-year field trial in Burkina Faso indicates that ‘worst-case scenario’ (dominant resistance to both Cry1Ac the use of Bt cotton varieties containing the genes Cry1Ac Ò and Cry2Ab, suboptimal mortality induced by Cry2Ab, and Cry2Ab from Monsanto (Bollgard II) can increase high efficiency of insecticide sprays in non-Bt cotton ref- yield by 30% and reduce insecticide use by 60% (James uges, and high MR2), Nibouche et al. (2007) concluded 2008). If commercial results confirm these findings, the that Bt cotton should not be grown on more than 30% of use of biotech cotton will likely expand in the rest of the the total cotton cropping area to delay resistance evolu- West African cotton belt. Several countries have passed a tion by more than 10 years. In contrast to the present national biosafety law or are in the process to do so, to study, this earlier model did not incorporate recent esti- authorize the commercial release of Bt cotton. Biogeo- mates of key parameters that influence the evolution of chemical markers provide a valuable tool to evaluate the
62 ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 Bre´ vault et al. Resistance to Bt cotton role of a variety of refuges in delaying the evolution of Bre´vault, T., Y. Oumarou, J. Achaleke, M. Vaissayre, and S. Nibouche. resistance to Bt crops in polyphagous insect pests. Such 2009. Initial activity and persistence of insecticides for the control of markers could be useful to assess the role of non-Bt cot- bollworms (Lepidoptera: Noctuidae) in cotton crops. Crop Protection 28:401–406. ton vs. non-cotton refuges in delaying H. armigera resis- Carrie`re, Y., C. Ellers-Kirk, M. S. Sisterson, L. Antilla, M. Whitlow, T. J. tance in Burkina Faso and other countries that may Dennehy, and B. E. Tabashnik. 2003. Long-term regional adopt Bt cotton. suppression of pink bollworm by Bacillus thuringiensis cotton. Proceedings of the National Academy of Sciences USA 100:1519–1523. Acknowledgements Carrie`re, Y., P. Dutilleul, C. Ellers-Kirk, B. Pedersen, S. Haller, L. Antilla, T. J. Dennehy et al. 2004. Sources, sinks, and zone of influ- We thank T. Malausa for advice at early stages of the pro- ence of refuges for managing insect resistance to Bt crops. Ecological ject, M. Vaissayre, S. Heuberger, and D. Crowder for crit- Applications 14:1615–1623. ical review of the manuscript, and F. Chiroleu, G. Head, Carrie`re, Y., D. W. Crowder, and B. E. Tabashnik. 2010. Evolutionary R. Orth, C. Scrimgeour, and J.M. Vassal for technical ecology of adaptation to Bt crops. Evolutionary Applications 3:561– assistance. This work was supported by PRASAC-ARDE- 573. SAC and SODECOTON Co. Deniro, M. J., and S. Epstein. 1978. Carbon isotopic evidence for different feeding patterns in two Hydrax species occupying same Data archiving statement habitat. Science 201:906–908. Downes, S., T. Parker, and R. Mahon. 2010. Incipient Resistance of Data for this study are available in Dryad: doi:10.5061/ Helicoverpa punctigera to the Cry2Ab Bt toxin in Bollgard II Cotton. dryad.ts7sm. PLoS ONE 5:1–5. Feng, H., F. Gould, Y. Huang, Y. Jiang, and K. Wu. 2010. Modeling Literature cited the population dynamics of cotton bollworm Helicoverpa armigera (Hu¨bner) (Lepidoptera: Noctuidae) over a wide area in northern Adamczyk, J. J., L. C. Adams, and D. D. Hardee. 2001. Field efficacy China. Ecological Modelling 221:1819–1830. and seasonal expression profiles for terminal leaves of single and Forrester, N. W., M. Cahill, L. J. Bird, and J. K. Layland. 1993. double Bacillus thuringiensis toxin cotton genotypes. Journal of Management of pyrethroid and endosulfan resistance in Helicoverpa Economic Entomology 94:1589–1593. armigera (Lepidoptera: Noctuidae) in Australia. Bulletin of Akhurst, R. J., W. James, L. J. Bird, and C. Beard. 2003. Resistance to Entomological Research (Suppl.) 1:1–132. the Cry1Ac delta-endotoxin of Bacillus thuringiensis in the cotton Gould, F. 1998. Sustainability of transgenic insecticidal cultivars: inte- bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae). Journal grating pest genetics and ecology. Annual Review of Entomology of Economic Entomology 96:1290–1299. 43:701–726. Bagla, P. 2010. Hardy cotton-munching pests are latest blow to GM Gould, F., N. Blair, M. Reid, T. L. Rennie, J. Lopez, and S. Micinski. 2002. crops. Science 327:1439. Bacillus thuringiensis-toxin resistance management: stable Baker, G. H., C. R. Tann, and G. P. Fitt. 2008. Production of isotope assessment of alternate host use by Helicoverpa zea. Helicoverpa spp. (Lepidoptera, Noctuidae) from different refuge Proceedings of the National Academy of Science USA 99:16581–16586. crops to accompany transgenic cotton plantings in eastern Australia. Gould, F., M. B. Cohen, J. S. Bentur, G. G. Kennedy, and J. Van Duyn. Australian Journal of Agricultural Research 59:723–732. 2006. Impact of small fitness costs on pest adaptation to crop varie- Bird, L. J., and R. J. Akhurst. 2004. Relative fitness of CrylA-resistant ties with multiple toxins: a heuristic model. Journal of Economic and -susceptible Helicoverpa armigera (Lepidoptera: Noctuidae) on Entomology 99:2091–2099. conventional and transgenic cotton. Journal of Economic Green, W. M., M. C. de Billot, T. Joffe, L. van Staden, A. Bennett-Nel, Entomology 97:1699–1709. C. L. N. du Toit, and L. van der Westhuizen. 2003. Indigenous Bird, L. J., and R. J. Akhurst. 2005. Fitness of Cry1A-resistant plants and weeds on the Makhathini Flats as refuge hosts to and -susceptible Helicoverpa armigera (Lepidoptera: Noctuidae) on maintain bollworm population susceptibility to transgenic cotton transgenic cotton with reduced levels of Cry1Ac. Journal of (Bollgard (TM)). African Entomology 11:21–29. Economic Entomology 59:1166–1168. Gustafson, D. I., G. P. Head, and M. A. Caprio. 2006. Modeling the Bird, L. J., and R. J. Akhurst. 2007. Effects of host plant species on impact of alternative hosts on Helicoverpa zea adaptation to Bollgard fitness costs of Bt resistance in Helicoverpa armigera (Lepidoptera: cotton. Journal of Economic Entomology 99:2116–2124. Noctuidae). Biological Control 40:196–203. Hartstack, A. W., J. A. Witz, and D. R. Buck. 1979. Moth traps for the Bourguet, D., F. Delmotte, P. Franck, T. Guillemaud, X. Reboud, C. tobacco budworm (Lepidoptera: Noctuidae). Journal of Economic Vacher, and A. S. Walker. 2010. The skill and style to model the Entomology 72:519–522. evolution of resistance to pesticides and drugs. Evolutionary Head, G., R. E. Jackson, J. Adamczyk, J. R. Bradley, J. Van Duyn, J. Applications 3:375–390. Gore, D. D. Hardee et al. 2010. Spatial and temporal variability in Bowden, J. 1973. Migration of pests in the tropics. Communications in host use by Helicoverpa zea as measured by analyses of stable carbon Agricultural and Applied Biological Sciences 38:785–796. isotope ratios and gossypol residues. Journal of Applied Ecology Bre´vault, T., J. Achaleke, S. P. Sougnabe´, and M. Vaissayre. 2008. 47:583–592. Tracking pyrethroid resistance in the polyphagous bollworm, Hood-Nowotny, R., and B. G. J. Knols. 2007. Stable isotope methods Helicoverpa armigera (Lepidoptera: Noctuidae), in the shifting in biological and ecological studies of arthropods. Entomologia landscape of a cotton-growing area. Bulletin of Entomological Experimentalis et Applicata 124:3–16. Research 98:565–573.
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 63 Resistance to Bt cotton Bre´ vault et al.
Huang, J. K., J. W. Mi, H. Lin, Z. J. Wang, R. J. Chen, R. F. Hu, S. Nair, R., V. Kalia, K. K. Aggarwal, and G. T. Gugar. 2010. Inheritance Rozelle et al. 2010. A decade of Bt cotton in Chinese fields: assessing of Cry1Ac resistance and associated biological traits in the cotton the direct effects and indirect externalities of Bt cotton adoption in bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae). Journal China. Science China-Life Sciences 52:981–991. of Invertebrate Pathology 104:31–38. Hutchison, W. D., E. C. Burkness, P. D. Mitchell, R. D. Moon, T. W. Nibouche, S. 1994. Cycle e´volutif de Helicoverpa armigera (Hu¨bner, Leslie, S. J. Fleischer, M. Abrahamson et al. 2010. Areawide 1808) (Lepidoptera, Noctuidae) dans l’Ouest du Burkina Faso: suppression of European corn borer with Bt maize reaps savings to biologie, e´cologie et variabilite´ ge´ographique des populations. non-Bt maize growers. Science 330:222–225. PhD Thesis, Ecole Nationale Supe´rieure Agronomique, James, C. 2008. Global status of commercialized biotech/GM crops: Montpellier. 2008, ISAAA Brief No. 39. International Service for the Acquisition Nibouche, S., R. Bue`s, J. F. Toubon, and S. Poitout. 1998. Allozyme of Agri-biotech Applications (ISAAA), Ithaca, NY. polymorphism in the American cotton bollworm Helicoverpa armi- James, C. 2009. Global status of commercialized biotech/GM crops: gera (Hu¨bner) (Lepidoptera: Noctuidae): comparison of African and 2008, ISAAA Brief No. 41. International Service for the Acquisition European populations. Heredity 80:438–445. of Agri-biotech Applications (ISAAA), Ithaca, NY. Nibouche, S., N. Gue´rard, P. Martin, and M. Vaissayre. 2007. Jaroszewski, J. W., T. Stromhansen, S. Honorehansen, O. Thastrup, Modelling the role of refuges for sustainable management of and H. Kofod. 1992. On the botanical distribution of chiral forms dual-gene Bt cotton in West African smallholder farming systems. of gossypol. Planta Medicina 58:454–458. Crop Protection 26:828–836. Kranthi, K. R., S. Naidu, C. S. Dhawad, A. Tatwawadi, K. Mate, E. Olsen, K. M., J. C. Daly, H. E. Holt, and E. J. Finnegan. 2005. Patil, A. A. Bharose et al. 2005. Temporal and intra-plant variability Season-long variation in expression of Cry1Ac gene and efficacy of of Cry1Ac expression in Bt-cotton and its influence on the survival Bacillus thuringiensis toxin in transgenic cotton against Helicoverpa of the cotton bollworm, Helicoverpa armigera (Hubner) (Lepidop- armigera (Lepidoptera: Noctuidae). Journal of Economic Entomol- tera: Noctuidae). Current Science 89:291–298. ogy 98:1007–1017. Liu, F., Z. Xu, Y. C. Zhu, F. Huang, Y. Wang, H. Li, C. Gao et al. Orth, R. G., G. Head, and M. Mierkowski. 2007. Determining larval 2009. Evidence of field-evolved resistance to Cry1Ac-expressing Bt host plant use by a polyphagous lepidopteran through analysis of cotton in Helicoverpa armigera (Lepidoptera: Noctuidae) in northern adult moths for plant secondary metabolites. Journal of Chemical China. Pest Management Science 66:155–161. Ecology 33:1131–1148. Liu, C. X., Y. H. Li, Y. L. Gao, C. M. Ning, and K. M. Wu. 2010. Peterson, R. K., and E. H. Hunt. 2003. The probabilistic economic Cotton bollworm resistance to Bt transgenic cotton: a case analysis. injury level: incorporating uncertainty into pest management Science China-Life Sciences 53:934–941. decision-making. Journal of Economic Entomology 96:536–542. Mahon, R. J., and K. M. Olsen. 2009. Limited survival of a Cry2Ab- Qiao, F. B., J. K. Huang, S. Rozelle, and J. Wilen. 2010. Natural refuge resistant strain of Helicoverpa armigera (Lepidoptera: Noctuidae) on crops, buildup of resistance, and zero-refuge strategy for Bt cotton Bollgard II. Journal of Economic Entomology 102:708–716. in China. Science China-Life Sciences 53:1227–1238. Mahon, R. J., and S. Young. 2010. Selection experiments to assess R Development Core Team R. 2008. A Language and Environment for fitness costs associated with Cry2Ab resistance in Helicoverpa Statistical Computing. R Foundation for Statistical Computing, armigera (Lepidoptera: Noctuidae). Journal of Economic Vienna. Entomology 103:835–842. Ravi, K. C., T. M. Mohan, G. Manjunath, B. V. Head, D. P. A. Patil, Mahon, R. J., K. M. Olsen, K. A. Garsia, and S. R. Young. 2007a. K. Greba, and J. Premalatha. 2005. Relative abundance of Heli- Resistance to Bacillus thuringiensis toxin Cry2Ab in a strain of coverpa armigera (Lepidoptera: Noctuidae) on different host crops Helicoverpa armigera (Lepidoptera: Noctuidae) in Australia. Journal in India and the role of these crops as natural refuge for Bacillus of Economic Entomology 100:894–902. thuringiensis cotton. Environmental Entomology 34:59–69. Mahon, R. J., K. M. Olsen, S. Downes, and S. Addison. 2007b. Frequency Rojas, M. G., R. D. Stipanovic, H. J. Williams, and S. B. Vinson. 1992. of alleles conferring resistance to the Bt toxins Cry1Ac and Cry2Ab in Metabolism of gossypol by Heliothis virescens (F) (Lepidoptera: Australian populations of Helicoverpa armigera (Lepidoptera: Noctui- Noctuidae). Environmental Entomology 21:518–526. dae). Journal of Economic Entomology 100:1844–1853. Roush, R. T. 1998. Two-toxin strategies for management of insecticidal Mahon, R. J., K. M. Olsen, and S. Downes. 2008. Isolations of Cry2Ab transgenic crops: can pyramiding succeed where pesticide mixtures resistance in Australian populations of Helicoverpa armigera have not? Philosophical Transactions of the Royal Society of London (Lepidoptera: Noctuidae) are allelic. Journal of Economic B: Biological Sciences 353:1777–1786. Entomology 101:909–914. Sawicki, R. M. 1987. Monitoring and interpreting changes in insecti- Martin, T., O. G. Ochou, A. Djihinto, D. Traore, M. Togola, J. M. cide resistance. In Ford M. G., D. W. Holloman, B. P. S. Khambay, Vassal, M. Vaissayre et al. 2005. Controlling an insecticide resistance and R. M. Sawicki, eds. Combating Resistance to Xenobiotics, bollworm in West Africa. Agriculture, Ecosystems and Environment pp. 105–117. Ellis Horwood, Chichester. 107:409–411. Showalter, A. M., S. Heuberger, B. E. Tabashnik, and Y. Carrie`re. Matten, S. R., G. P. Head, and H. D. Quemada. 2008. How 2009. A primer for using transgenic insecticidal cotton in developing governmental regulation can help or hinder the integration of Bt countries. Journal of Insect Science 9:1–39. crops into IPM programs. In J. Romeis, A. M. Shelton, and G. G. Storer, N. P., J. M. Babcock, M. Schlenz, T. Meade, G. D. Thompson, Kennedy, eds. Integration of Insect-Resistant Genetically Modified J. W. Bing, and R. M. Huckaba. 2010. Discovery and Characteriza- Crops within IPM Programs, pp. 27–39. Springer, New York. tion of Field Resistance to Bt Maize: Spodoptera frugiperda (Lepidop- Monsanto. 2010. Cotton in India. http://www.monsantoindia.com/ tera: Noctuidae) in Puerto Rico. Journal of Economic Entomology monsanto/layout/pressreleases/mmb_pressrelease.asp [accessed on 103:1031–1038. February 2010].
64 ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 Bre´ vault et al. Resistance to Bt cotton
Tabashnik, B. E., A. J. Gassmann, D. W. Crowder, and Y. Carrie`re. Table S1. Name and geographic coordinates of sampling locations. 2008. Insect resistance to Bt crops: evidence versus theory. Nature The number of Helicoverpa armigera moths analyzed in each month Biotechnology 26:199–202. (2006) for stable carbon isotope ratio and gossypol is reported. Tabashnik, B. E., B. J. Van Rensburg, and Y. Carrie`re. 2009. Field- Table S2. Standard values of relative survival of Helicoverpa armi-
evolved insect resistance to Bacillus thuringiensis crops: definition, gera genotypes on Cry1Ac/Cry2Ab cotton (LB3g). We assumed a fitness
theory, and data. Journal of Economic Entomology 102:2011–2025. cost associated with resistance to Cry1Ac but not Cry2Ab and DLC Vaissayre, M., G. O. Ochou, O. S. A. Hema, and M. Togola. 2006. values of 0.26 and 0.0 for resistance to Cry1Ac and Cry2Ab, respec- Changing strategies for sustainable management of cotton pests in tively (Table S3). Other survival values were calculated in sensitivity sub-Saharan Africa. Cahiers Agricultures 15:80–84. analyses assessing the effect of dominance (see Methods).
Valicek, P. 1979. Wild and cultivated cottons. Coton et Fibres Tropi- Table S3. Effects of dominance of resistance (DLC), presence of cales 34:239–264. non-cotton refuges, abundance of non-Bt cotton refuges (Pref) and
Van Rensburg, B. J. 2007. First report of field resistance by stem borer, initial frequency of resistance alleles (po) on the number of years to Busseola fusca (Fuller) to Bt-transgenic maize. South African Journal achieve >20% survival on Cry1Ac/Cry2Ab cotton. Simulations consid- of Plant and Soil 24:147–151. ered movement of H. armigera from non-cotton hosts to cotton at Vassal, J. M., T. Bre´vault, J. Achaleke, and P. Menozzi. 2008. Genetic three locations in Cameroon, in the absence of long-range migration. structure of the polyphagous pest Helicoverpa armigera (Lepidoptera: Table S4. Frequency of resistance alleles r1 (Cry1Ac, top) and r2 Noctuidae) across the sub-Saharan cotton belt. Communications in (Cry2Ab, bottom) at the time when > 20% survival on Cry1Ac/Cry2Ab Agricultural and Applied Biological Sciences 73:433–437. cotton occurred (see Table S3). Simulations considered movement of Wu, K. M., and Y. Y. Guo. 2005. The evolution of cotton pest man- H. armigera from non-cotton hosts to cotton at three locations in agement practices in China. Annual Review of Entomology 50:31– Cameroon, in the absence of long-range migration.
52. Table S5. Effects of dominance of resistance (DLC), abundance of Wu, K., H. Feng, and Y. Guo. 2004. Evaluation of maize as a refuge non-Bt cotton refuges (Pref) and initial frequency of resistance alleles
for management of resistance to Bt cotton by Helicoverpa armigera (po) on the mean number of years (with 0.05- and 0.95-quantile) to (Hubner) in the Yellow River cotton-farming region of China. Crop achieve >20% survival on Cry1Ac/Cry2Ab cotton. For each generation Protection 23:523–530. and region, the proportion of moths produced by non-cotton hosts
Wu, K. M., Y. H. Lu, H. Q. Feng, Y. Y. Jiang, and J. Z. Zhao. 2008. (fnoncot) was sampled repeatedly (1000 iterations) from the Student dis- Suppression of cotton bollworm in multiple crops in China in areas tribution, according to Monte Carlo simulation (see Methods).
with Bt toxin-containing cotton. Science 321:1676–1678. Table S6. Effects of dominance of resistance to Cry2Ab (DLC), Wu, Y., J. M. Vassal, M. Royer, and I. Pieretti. 2009. A single linkage abundance of non-Bt cotton refuges (Pref) and initial frequency of the
group confers dominant resistance to Bacillus thuringiensis d-endo- Cry2Ab resistance allele (po) on the number of years to achieve >20% toxin Cry1Ac in Helicoverpa armigera. Journal of Applied Entomol- survival on Cry1Ac/Cry2Ab cotton at Djalingo (Cameroon). Simula- ogy 133:375–380. tions considered movement of H. armigera from non-cotton hosts to Zhao, J. Z., J. Cao, Y. X. Li, H. L. Collins, R. T. Roush, E. D. Earle, cotton and patterns of long-range migration. and A. M. Shelton. 2003. Transgenic plants expressing two Bacillus Please note: Wiley-Blackwell are not responsible for the content or thuringiensis toxins delay insect resistance evolution. Nature Biotech- functionality of any supporting materials supplied by the authors. Any nology 21:1493–1497. queries (other than missing material) should be directed to the corre- sponding author for the article.
Supporting Information
Additional Supporting Information may be found in the online version of this article: Figure S1. General description of the simulation model. Circled numbers refer to the corresponding equation number in Materials and Methods.
ª 2011 Blackwell Publishing Ltd 5 (2012) 53–65 65 Potential shortfall of pyramided transgenic cotton for insect resistance management
Thierry Brévaulta,1, Shannon Heubergerb, Min Zhangb, Christa Ellers-Kirkb, Xinzhi Nic, Luke Massond, Xianchiun Lib, Bruce E. Tabashnikb, and Yves Carrièreb,1 aCentre de coopération Internationale en Recherche Agronomique pour le Développement, Unité Propre de Recherche 102, F-34398 Montpellier, France; bDepartment of Entomology, University of Arizona, Tucson, AZ 85721; cU.S. Department of Agriculture–Agricultural Research Service Crop Genetics and Breeding Research Unit, Tifton, GA 31793; and dNational Research Council of Canada, Biotechnology Research Institute, Montreal, QC, Canada H4P 2R2
Edited by May R. Berenbaum, University of Illinois at Urbana–Champaign, Urbana, IL, and approved February 15, 2013 (received for review September 25, 2012)
To delay evolution of pest resistance to transgenic crops producing complete redundant killing and recessive resistance, only the insects insecticidal proteins from Bacillus thuringiensis (Bt), the “pyramid” homozygous for resistance to both toxins have high survival on a strategy uses plants that produce two or more toxins that kill the two-toxin cultivar. Such doubly resistant individuals are expected same pest. In the United States, this strategy has been adopted to be rare in populations that have not been exposed previously widely, with two-toxin Bt cotton replacing one-toxin Bt cotton. Al- to either toxin. though two-toxin plants are likely to be more durable than one- Several factors could reduce redundant killing and decrease toxin plants, the extent of this advantage depends on several con- the effectiveness of the pyramid strategy for pests such as ditions. One key assumption favoring success of two-toxin plants Helicoverpa zea. Redundant killing is reduced when some sus- is that they kill insects selected for resistance to one toxin, which is ceptible pests survive exposure to the toxins produced by a pyra- called “redundant killing.” Here we tested this assumption for a ma- mid (14). Survival of susceptible H. zea larvae on two-toxin cotton jor pest, Helicoverpa zea, on transgenic cotton producing Bt toxins can reach 5% during the growing season (5, 16–18), which could Cry1Ac and Cry2Ab. Selection with Cry1Ac increased survival on reduce the efficacy of the pyramid strategy. Furthermore, as Bt two-toxin cotton, which contradicts the assumption. The concentra- cotton plants age, toxin concentrations decline, which could in- SCIENCES AGRICULTURAL tion of Cry1Ac and Cry2Ab declined during the growing season, crease survival of pests that have inherently low susceptibility to Bt which would tend to exacerbate this problem. Furthermore, analysis toxins (19, 20). Redundant killing could also be undermined if of results from 21 selection experiments with eight species of lepi- selection for resistance to one of the toxins causes cross-resistance dopteran pests indicates that some cross-resistance typically occurs to the other toxin (5, 14, 19, 21). Cry1Ac and Cry2Ab have been between Cry1A and Cry2A toxins. Incorporation of empirical data considered a good combination for pyramided Bt crops because into simulation models shows that the observed deviations from they have low amino acid homology and bind to different target ideal conditions could greatly reduce the benefits of the pyramid sites in the larval midgut (22, 23). However, in field-derived strains strategy for pests like H. zea, which have inherently low suscepti- of H. zea and Helicoverpa armigera, responses to Cry1Ac and bility to Bt toxins and have been exposed extensively to one of the Cry2Ab were genetically correlated, indicating potential cross- toxins in the pyramid before two-toxin plants are adopted. For such resistance (5, 24–27). Although redundant killing is critical for pests, the pyramid strategy could be improved by incorporating em- the success of the pyramid strategy, little is known about factors pirical data on deviations from ideal assumptions about redundant affecting redundant killing in H. zea and other pests with low killing and cross-resistance. susceptibility to Bt toxins. For example, based on their modeling results, Onstad and Meinke (28) called for empirical evaluation genetically modified | sustainability of pyramids to develop resistance management plans. Here we examined the previously untested assumption of re- dundant killing in H. zea on cotton producing Cry1Ac and orn and cotton engineered to produce insecticidal proteins Cfrom Bacillus thuringiensis (Bt) have provided several benefits, Cry2Ab. We found that a strain selected for resistance to Cry1Ac including reduced insecticide use, regional pest suppression, pro- had increased survival on two-toxin cotton relative to its un- tection of natural enemies, and increased or less variable yields (1– selected parent strain, which contradicts the assumption of re- 3). Evolution of resistance by pests, however, is the most serious dundant killing. We also found evidence of cross-resistance threat to the continued efficacy of Bt crops. Significant increases between Cry1A and Cry2A toxins from an analysis of data from in the frequency of alleles conferring resistance to Bt toxins pro- 21 selection experiments including the results reported here. duced by transgenic crops have been reported in some populations Incorporation of empirical data into simulation models shows of at least seven target species (4–12). Analyses of monitoring data that the observed deviations from ideal conditions could greatly and field experiments suggest that refuges of host plants that do reduce the benefits of the pyramid strategy for H. zea. not produce Bt toxins and grow near Bt crops can reduce the risk of resistance (4, 5, 13). Such refuges of non–Bt host plants delay resistance by enabling survival of susceptible pests that can mate Author contributions: T.B. and Y.C. designed research; T.B., M.Z., C.E.-K., and Y.C. per- with resistant pests surviving on Bt crops. formed research; X.N. provided field-collected insects; X.N., L.M., X.L., and B.E.T. contrib- The first generation of Bt cotton, which was grown on a large uted new reagents/analytic tools; T.B., S.H., B.E.T., and Y.C. analyzed data; and T.B., S.H., B.E.T., and Y.C. wrote the paper. scale starting in 1996 in the United States, produced only one Bt fl toxin, Cry1Ac. This Cry1Ac cotton was progressively and com- Con ict of interest statement: B.E.T. received support for research that is not related to this publication from the following sources: Cotton Foundation, Cotton Inc., National pletely replaced in the United States from 2003 to 2011 by Cotton Council, Monsanto, and Dow AgroSciences. He is also a coauthor of a patent on “pyramided” plants that produce two Bt toxins, either Cry1Ac engineering modified Bt toxins to counter pest resistance, which is related to research and Cry2Ab or Cry1Ac and Cry1F (Fig. 1 and Fig. S1). This re- described by Tabashnik et al. (2011, Nature Biotechnology 29: 1128-1131). placement was spurred by the idea that the evolution of resistance This article is a PNAS Direct Submission. would be delayed substantially by two-toxin crops relative to one- 1To whom correspondence may be addressed. E-mail: [email protected] or ycarrier@ toxin crops (1). A central assumption of the two-toxin pyramid ag.arizona.edu. strategy is that insects resistant to one toxin will be killed by This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. the other toxin, which is called “redundant killing” (14, 15). With 1073/pnas.1216719110/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1216719110 PNAS Early Edition | 1of6 100 both Cry1Ac and Cry2Ab (Fig. 2). The odds of survival for GA-R 2 90 Cry1Ac Cry1Ac + Cry2Ab Cry1Ac + Cry1F relative to GA were 11 times higher on Cry1Ac cotton (χ = 31.5, = < 2 = 80 df 2, P 0.001) and 13 times higher on two-toxin cotton (χ = < 70 11.3, df 1, P 0.001) (Fig. 2). Furthermore, inheritance of 60 resistance on Cry1Ac cotton was not completely recessive. Sur- vival on Cry1Ac cotton was three times higher for the F1 progeny 50 2 = = = 40 of GA and GA-R relative to GA (χ 4.86, df 2, P 0.028) Bt cottonBt (%) 30 (Fig. 2). The value of h, which varies from 0 for recessive re- 20 sistance to 1 for dominant resistance, was 0.25. Survival on non–Bt cotton did not differ between GA (71%) 10 and GA-R (62%) (χ2 = 0.97, df = 1, P = 0.33). Thus, we did not 0 detect a significant fitness cost affecting this trait (Fig. 2). Sur- 2011 1996 1997 1998 1999 2001 2002 2003 2004 2005 2008 2009 2010 2012 2007 2000 2006 vival of GA-R was higher on non–BtcottonthanoneitherCry1Ac Year cotton (χ2 = 19.8, df = 1, P < 0.001) or two-toxin cotton (χ2 = 71.3, df = 1, P < 0.001), indicating incomplete resistance (I = 0.43 on Fig. 1. Percentage of total hectares of upland cotton planted to Bt cotton Cry1Ac cotton and 0.11 on two-toxin cotton). from 1996 to 2012 in the United States (U.S. Department of Agriculture– Agricultural Marketing Service, 1996–2012 crops). The non–Bt cotton per- Cross-Resistance Between Cry1A and Cry2A Toxins. centage is 100% minus the total height of each bar. To test the widely held assumption that cross-resistance does not occur between Cry1A and Cry2A toxins, we analyzed results of 21 experiments in Results which strains of eight major lepidopteran pests had been selected for resistance to a Cry1A toxin and subsequently evaluated for Effects of Laboratory Selection with Cry1Ac on Susceptibility to Bt cross-resistance to Cry2A, or vice versa. In 19 of 21 experiments, Toxins. We evaluated susceptibility to Bt toxins of three strains of selection with one toxin decreased susceptibility to the other toxin H. zea: a susceptible laboratory strain (LAB-S), a field-derived fi (Fig. 3 and Table S2). The overall pattern in the 21 experiments strain from Georgia that was exposed to Bt toxins only in the eld considered together indicates significant cross-resistance between (GA), and a resistant strain (GA-R) that was derived from the GA Cry1A and Cry2A toxins (signed-rank test, P = 0.0002). Analysis strain and selected in the laboratory with Cry1Ac in diet for nine of each selection experiment separately detected significant cross- generations. Comparison between the susceptible strain and the resistance in seven of 21 cases. In these seven cases, selection with GA strain suggests that exposure to Bt crops in the field had se- one toxin caused an average 140-fold increase in the LC50 of the lected for some resistance to Cry1Ac and Cry2Ab in the founders = fi other toxin (range 3- to 420-fold; Table S2). In the remaining of the GA strain. For the eld-derived GA strain relative to LAB- 14 cases, significant cross-resistance was not detected when each S, the concentration of toxin killing 50% (LC50) was 55-fold higher experiment was analyzed individually, but selection with one toxin for Cry1Ac and 14-fold higher for Cry2Ab (Table 1). decreased susceptibility to the other toxin in 12 of 14 cases, which Selection with Cry1Ac increased resistance to Cry1Ac and refutes the null hypothesis of no cross-resistance for these 14 cases caused strong cross-resistance to the closely related toxin Cry1Ab, (signed-rank test, P = 0.044). In these 14 cases, selection with one but not to the more distantly related toxin Cry2Ab. After nine toxin caused an average increase of 1.3-fold in the LC50 or IC50 generations of laboratory selection of GA-R, the LC50 of Cry1Ac (concentration causing 50% inhibition of growth) of the other for GA-R was 10 times higher than for GA and 560 times higher toxin (range = 0.32- to 2.2-fold; Table S2). than for LAB-S (Table 1). However, based on the conservative fi criterion of nonoverlap of 95% ducial limits (FLs), selection with Concentration of Cry1Ac and Cry2Ab Toxins in Terminal Leaves. We fi Cry1Ac did not signi cantly increase the LC50 of Cry2Ab (in μg next determined if the concentrations of Bt toxins varied between toxin per ml diet) for GA-R (62) relative to GA (31) (Table 1). By cultivars or due to plant aging to see how this might impact re- fi contrast, selection with Cry1Ac caused a statistically signi cant, dundant killing in pyramid plants. The concentration of Cry1Ac in 15-fold increase in the LC50 of Cry1Ab for GA-R (940) relative terminal leaves decreased significantly during the growing season to GA (63) (Table S1). in both one- and two-toxin cotton (F = 58.8, df = 4, P < 0.001) (Fig. 4A). In addition, the concentration of Cry1Ac was generally – Survival from Neonate to Adult on Bt and non Bt Cotton. We next higher in one- than two-toxin cotton (F = 17.5, df = 1, P < 0.001). determined if selection for resistance to Cry1Ac affected survival on Bt plants. Laboratory selection with Cry1Ac increased survival of GA-R relative to GA on Bt cotton producing only Cry1Ac or 90 Non-Bt cotton Cry1Ac cotton Cry1Ac and Cry2Ab cotton 80 Table 1. Responses of H. zea to Cry1Ac and Cry2Ab toxins incorporated in diet 70 95% Fiducial 60
limits 50
−1 Toxin Strain N LC50 (μg·ml ) Lower Upper Slope RR 40
Cry1Ac LAB-S 336 0.42 0.08 0.80 1.1 1.0 (%) Survival 30 GA 560 23 14 44 1.0 55 20 GA-R 560 230 140 480 0.9 560 Cry2Ab LAB-S 336 2.2 1.6 2.8 2.2 1.0 10 GA 336 31 21 54 1.5 14 0 GA-R 336 62 30 340 0.9 28 GA F1 GA-R GA F1 GA-R GA GA-R
LAB-S, susceptible LAB-S; GA, field-derived strain from Georgia; GA-R, Fig. 2. Survival (+ 95% CI) from neonate to adult of H. zea from a field- resistant strain derived from the GA strain and selected with Cry1Ac in the derived strain (GA), a resistant strain (GA-R), and their F1 progeny reared on laboratory; N, number of larvae tested; RR, resistance ratio, the LC50 of plant material from non–Bt cotton and Bt cotton producing Cry1Ac or both a strain divided by the LC50 of the susceptible LAB-S strain. Cry1Ac and Cry2Ab.
2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1216719110 Brévault et al. Discussion Redundant killing, which occurs when each toxin produced by a two-toxin Bt cultivar kills all insects resistant to the other toxin, is essential for optimal success of the pyramid strategy (14, 15, 20, 28). Here we found that this assumption did not apply, because labo- ratory selection for resistance to Cry1Ac of a field-derived strain of H. zea (GA) significantly increased survival of larvae from the GA-R strain on both one- and two-toxin Bt cotton. Furthermore, our analysis of 21 selection experiments with eight species of lepi- dopteran pests shows pervasive cross-resistance between Cry1A and Cry2A toxins. When susceptible insects can survive on a Bt crop, even alleles with small effects on resistance can increase survival and contribute to the evolution of resistance (19). Accordingly, in pests with low susceptibility to Cry1A and Cry2A toxins, cross- Fig. 3. Cross-resistance between Cry1A and Cry2A in 21 selection ex- resistance between Cry1A and Cry2A will generally hasten evolu- periments. Insect strains were selected for resistance to a Cry1A toxin and tion of resistance. In accord with results from other simulation subsequently evaluated for cross-resistance to Cry2A, or vice versa. The models (14, 15, 20, 28), the modeling results presented here (Figs. 5 CRR is the LC50 (or IC50) of the toxin not used for selection (e.g., Cry2Ab) for the strain selected with the other toxin (e.g., Cry1Ac) divided by the and 6) indicate that resistance to a two-toxin pyramid evolves faster
LC50 (or IC50) of the toxin not used for selection (e.g., Cry2Ab) for an when each toxin of a pyramid does not kill all individuals resistant unselected, control strain. The expected value of log (CRR) is 0 if cross- to the other toxin, as seen with H. zea (Fig. 2). The joint effects on resistance is absent and >0 if cross-resistance occurs. Nineteen of the 21 evolution of resistance to pyramids of both cross-resistance and a ratios were >0, indicating significant cross-resistance between Cry1A and lack of complete redundant killing have received limited attention Cry2A toxins. Insect strains were from the following species: Diatracea previously. Our results imply that, to advance resistance manage- saccharalis (Ds), H. armigera (Ha), Helicoverpa punctigera (Hp), Heliothis ment for pyramids, this issue needs more attention. virescens (Hv), H. zea (Hz), Pectinophora gossypiella (Pg), Plutella xylostella Increased survival of GA-R on two-toxin cotton likely occurred (Pg), and Trichoplusia ni (Tn). because the concentration of Cry2Ab was not sufficient to kill individuals resistant to Cry1Ac. Survival of GA-R was 6.7% on two-toxin cotton and 27% on one-toxin cotton, yielding 25% sur- SCIENCES AGRICULTURAL Seasonal changes in Cry1Ac concentration differed between the vival on two-toxin cotton relative to one-toxin cotton (6.7%/27%), × = = = cultivars (cultivar date interaction, F 2.71, df 4, P 0.040). which translates to 75% mortality on two-toxin cotton relative to fi The concentration of Cry1Ac was signi cantly lower in two- than one-toxin cotton. In the SP15 strain of H. armigera selected for one-toxin cotton 52 and 66 d after planting (DAP) (linear con- homozygous resistance to Cry2Ab, but not selected with Cry1Ac, trasts, P < 0.004), but did not differ significantly between the survival on cotton producing both toxins increased as the growing cultivars at 28, 80, and 95 DAP (linear contrasts, P > 0.05). The season progressed (29). The Bt concentration of plants was not Cry2Ab concentration also declined seasonally in terminal leaves measured in this case, but a seasonal decline in the concentration in two-toxin cotton (F = 45.9, df = 4, P < 0.001) (Fig. 4B). of Cry1Ac likely reduced levels of redundant killing (19). In the results reported here, a seasonal decline in the concentration of Simulation Results. The two-toxin pyramid strategy is expected to Cry2Ab may have reduced redundant killing of H. zea. This con- be most effective for delaying resistance when alleles conferring clusion is based on the assumption that seasonal declines in the resistance to each toxin are rare, inheritance of resistance to each concentration of Bt toxins measured in leaves paralleled declines toxin is recessive, redundant killing is complete, and sufficient in bolls and squares on which larvae were also fed. refuges are present (5, 14, 19). We used simulation modeling to Another factor that may have reduced redundant killing on evaluate the potential effects of deviations observed here from the two-toxin cotton is weak cross-resistance between Cry1Ac and Cry2Ab. After selection with Cry1Ac, the LC of Cry2Ab for the ideal conditions of complete redundant killing and completely 50 GA-R strain was about double relative to its parent strain GA, recessive resistance. The modeling results show that with a 10% fi – and this difference was not signi cant based on the overlap be- refuge of non Bt cotton, evolution of resistance to two-toxin Bt tween 95% FLs (Table 1). However, the criterion of nonoverlap cotton was greatly accelerated by either a lack of complete re- of 95% FLs is statistically conservative (30) and we cannot exclude dundant killing or by nonrecessive resistance (Fig. 5). As expected, the possibility of weak cross-resistance. Moreover, weak cross- resistance generally evolved slower with either lower initial resistance is consistent with the results from our analysis of resistance allele frequency or larger refuges (Figs. 5 and 6). 21 selection experiments, and the significant positive genetic
Fig. 4. Concentration (±95% CI) of (A) Cry1Ac and (B) Cry2Ab in fresh terminal leaves of cotton producing only Cry1Ac (DP 448 B) and cotton producing Cry1Ac and Cry2Ab (DP 164 B2RF). Thirty-eight, 52, 66, 80, and 95 DAP correspond to presquaring, squaring, early fruiting (first flower), fruiting, and late fruiting stages of cotton, respectively. Least squares means toxin con- centration and associated 95% CI for each cultivar and date were obtained from ANOVA.
Brévault et al. PNAS Early Edition | 3of6 Fig. 5. Simulated evolution of resistance to two-toxin cotton: effects of redundant killing, dominance, and initial frequency.
Resistance to toxins one and two was conferred by alleles r1 and r2 at independent loci, respectively. The initial r2 allele frequency was 0.001. The initial r1 allele frequency was either 0.001 (triangles) or 0.1 (circles). The time to resistance was the number of years until ≥25% of the population could survive on Bt cotton in the third generation of each year. Analogous to the parameter h representing dominance of resistance to one toxin conferred by a single locus, we define hp as domi- nance of resistance to two-toxin plants conferred by two loci. We also define the RKF, which varies from 0 for no redundant killing to 1 for complete redundant killing (see SI Materials and Methods for additional details). (A) With recessive re- sistance (hp = 0), the time for resistance to evolve was >1,000 y with complete redundant killing (RKF = 1)andaninitialr1 allele frequency of either 0.001 or 0.1. By contrast, when redundant killing was not complete (RKF = 0.64), recessive resistance evolved in 139 y with an initial r1 allele frequency of 0.001 and in 3 y with an initial r1 allele frequency of 0.1. (B) With partially recessive resistance (hp = 0.25) and initial r1 allele frequency = 0.001, resistance evolved in 15 y either with or without complete redundant killing. With hp = 0.25 and initial r1 allele frequency = 0.1, resistance evolved in 5 y with complete redundant killing and in 3 y without complete redundant killing. correlations between resistance to Cry1Ac and Cry2Ab found in (35–37) than Cry1Ac (38–42). In any case, seasonal declines in the field-derived strains of H. zea and H. armigera (5, 26). concentration of Bt toxins could be an important factor affecting Resistance of H. zea was not completely recessive to Cry1Ac redundant killing because there is extensive variation among cotton here (h = 0.25) or to Cry1Ac in diet as reported previously cotton cultivars in patterns of production of Bt toxins (43). (h = 0.83) (4, 31). As shown here and in other studies using sim- ulation models, nonrecessive resistance can accelerate the evolu- Conclusions tion of resistance to two-toxin cotton (14, 15, 20). Fitness costs, Previous experimental evidence on the pyramid strategy comes which occur when fitness on non–BthostplantsislowerforBt– primarily from a model system with diamondback moth and resistant than susceptible insects, can slow evolution of resistance, noncommercial Bt broccoli plants producing Cry1Ac and Cry1C particularly to two-toxin plants (19, 32). Although we did not detect (44, 45). Although most of the optimal conditions for pyramids asignificant fitness cost associated with resistance to Cry1Ac, we apply to this model system, they may not apply for some other pest- might have underestimated cost because we compared the resistant Bt crop combinations, particularly when pests have inherently low GA-R strain with its parent strain GA, which apparently contained susceptibility to one or more of the toxins in the pyramid (19, 20, some resistance alleles. We did detect incomplete resistance, in- 28, 46). Here we found several deviations from optimal conditions dicated by the lower survival of the GA-R strain on either Cry1Ac for H. zea and Bt cotton producing Cry1Ac and Cry2Ab. Our cotton or two-toxin cotton relative to non–Bt cotton (Fig. 2). The results show that the commercially available two-toxin Bt cotton incomplete resistance detected, which can help to delay resistance plants we tested did not cause complete redundant killing of H. zea (33, 34), was incorporated in our simulations (Table S3). [redundant killing factor (RKF) = 0.64]. Also, inheritance of The results in our study are mainly similar to previous results in resistance to Cry1Ac was not completely recessive (h = 0.25). terms of the relative toxicity of Cry1Ac and Cry2Ab to H. zea and The deviations from ideal conditions we found with H. zea and the concentrations of these toxins in Bt cotton plants (SI Materials two-toxin Bt cotton, which entail both lack of complete redundant and Methods). However, unlike our results showing similar sea- killing and nonrecessive resistance, are likely to accelerate re- sonal declines of about threefold in Cry1Ac and Cry2Ab in DP 164 sistance relative to the conditions examined in previous modeling B2RF, previous studies with other cultivars found that the seasonal studies focusing primarily on complete or nearly complete re- decline in toxin concentration was generally less for Cry2Ab dundant killing (RKF = 0.99–1), recessive inheritance (hp ≤ 0.05), or both (28, 32, 45, 47, 48) (see Table S3 for details). Our simulation results under ideal conditions (hp = 0andRKF = 1) correspond closely with the projected outcomes in simulations 40 hphp = 0.25, RKFRKF == 11 under the favorable assumptions examined previously. By contrast, h = 0.25, RKF = 0.64 hpp = 0.25, RKF = 0.64 the substantial deviations from ideal conditions based on empirical
hp = 0, RKF = 0.64 30 hp = 0, RKF = 0.64 = = data from H. zea (hp 0.25 and RKF 0.64) yielded much faster evolution of resistance in simulations (Figs. 5 and 6). Moreover, H. zea 20 populations were exposed extensively to Cry1Ac cotton be- fore and after two-toxin plants were introduced (Fig. 1 and Fig. S1). Because of cross-resistance between Cry1Ac and Cry1Ab (Table 10
Years to resistance Years S1), similar exposure to Cry1Ab corn is also problematic because it would tend to increase the frequency of resistance to Cry1Ac.
0 Field monitoring data show decreased susceptibility of H. zea 0 10 20 30 40 50 populations to both Cry1Ac and Cry2Ab in some regions in the Refuge (%) United States (4, 5). In light of these data and the deviations from optimal conditions summarized above, effective management of Fig. 6. Simulated effects of refuge percentage, dominance, and redundant resistance in this case may require relatively large refuges of non– killing on evolution of resistance to two-toxin Bt cotton. The RKF was either 1, which represents the ideal condition of complete redundant killing, or Bt host plants in conjunction with multiple control tactics as part 0.64, which reflects the higher survival on two-toxin Bt cotton for individuals of integrated pest management (4, 20, 46). Our results also suggest selected for resistance to one toxin (Cry1Ac) relative to susceptible individ- that management of resistance to Bt crop pyramids in pests with uals, based on empirical data for survival of the GA-R strain relative to the inherently low susceptibility to Bt toxins could be enhanced by GA strain (Fig. 2). Table S3 gives fitness values for each genotype under each addressing effects of cross-resistance, less than complete redundant set of conditions simulated. killing, and seasonal declines in the concentration of Bt toxins.
4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1216719110 Brévault et al. Materials and Methods After 7 d or earlier if larvae had molted to third instar in less than 7 d, larvae Insect Strains, Rearing, and Selection. We used three strains of H. zea: a sus- were transferred individually to 470 mL clear plastic cups (Fabri-Kal) venti- lated with a 5-cm-diameter mesh-covered hole in the lid. These older larvae ceptible LAB-S obtained from Benzon Research Inc (Carlisle, PA), a field-derived were offered plant terminals bearing small bolls, leaves, and squares. Plant strain from Georgia that was exposed to Bt toxins only in the field (GA), and stems were inserted in the lid of a 30 mL cup of water placed at the bottom of a resistant strain derived from the GA strain that we selected in the laboratory the container, and the containers disposed in a growth chamber maintained with Cry1Ac in diet for nine generations (GA-R) as described below. at 27 ± 1 °C, 60 ± 10% RH, with a photoperiod of 14:10 (L:D). Plant material We provided moths with cotton balls wetted with a 10% dilution of honey and water were renewed as needed (generally twice a week) until the in water for feeding and cheesecloth for egg-laying. Eggs were harvested daily insects pupated within the container. Survival was recorded every 4 d. and larvae were reared on diet (Southland Products Inc.). Strains were main- tained at 27 ± 1°C,60± 10% relative humidity (RH), and 14 light (L):10 dark (D). Analysis of Cross-Resistance. We reviewed experiments in which insect strains The GA strain originated from 180 larvae collected in July 2008 from had been selected for resistance to a Cry1A toxin and subsequently evaluated “ ” Cry1Ab corn (Zea mays L.) hybrid DKC 6971 (MON810) near Tifton, Georgia, for cross-resistance to Cry2A, or vice versa. For each selection experiment and and was reared on diet without exposure to toxins. After two generations toxin not used for selection, the LC50 or IC50 (i.e., toxin concentration of laboratory rearing, we used a subset of insects from GA to start the GA-R inhibiting 50% of growth) of the strain selected for resistance was divided strain. We selected GA-R for resistance to Cry1Ac during each of nine non- by the LC50 or IC50 of the unselected control strain. The expected value consecutive generations by exposing at least 1,000 GA-R neonates to Cry1Ac in of this cross-resistance ratio (CRR) is 1 if cross-resistance is absent and >1if diet. In each selected generation, only larvae that reached third instar after 7 d cross-resistance is present. However, the logarithm of the CCR was used in of feeding on diet treated with Cry1Ac were transferred to non–Bt diet and statistical analyses to improve linearity and normality of this variable (54). reared to pupation to continue the strain. The concentration (μgofCry1Ac The expected value of log CRR is 0 if cross-resistance is absent and >0 if cross- −1 mL diet) increased progressively with 10–20 in selected generations 1–3and resistance is present. 100–1,000 in selected generations 4–9. Toxin Concentration in Plants. The concentrations of Cry1Ac and Cry2Ab in Bt Toxins. We used protoxin crystals with spores of Cry1Ab and Cry1Ac prepared cotton leaves were analyzed with ELISA (see SI Materials and Methods for by J. Sánchez as described previously (49) and Cry2Ab produced by a recombi- additional details). nant acrystalliferous strain of Bt ssp. kurstaki (HD73 cry–) that was transformed with the Cry2Ab gene from strain HD1 of Bt ssp. kurstaki (50). We used the Population Genetics Model. To simulate the evolution of H. zea resistance to protoxin preparations described above in all experiments, with one exception. two-toxin cotton, we used a deterministic model with two loci. Locus one For generations 4–9 of selection of GA-R, we needed higher toxin concentrations affected responses to Cry1Ac and locus 2 affected responses to Cry2Ab. Each and we used MVP II (Dow AgroSciences) containing 20% Cry1Ac protoxin (51). locus had two alleles: r1 and r2 conferring resistance and s1 and s2 susceptibility SCIENCES to Cry1Ac and Cry2Ab, respectively. We assumed initial gametic equilibrium AGRICULTURAL Diet Bioassays. We used diet incorporation bioassays (52) to assess the responses (15). The time to resistance was the number of years until ≥25% of the of LAB-S, GA, and GA-R to Cry1Ac, Cry2Ab, and Cry1Ab. Diet bioassays were population could survive on Bt cotton in the third generation of each year conducted simultaneously using generation 12 of GA-R and generation 14 of (see SI Materials and Methods for additional details). GA. Toxins were suspended in distilled water and incorporated into a bean- baseddiet(53)at45–55 °C (1:5 volume of toxin solution to diet). We dispensed Data Analysis. We used probit analysis to estimate the toxin concentration
0.5 mL of diet in each well of 128-cell bioassay trays (Bioserv) using a 25 mL causing 50% mortality (LC50), its 95% FLs, and slope of the concentration– Repeater Plus Pipettor (Eppendorf). For controls, distilled water was mixed mortality line (55). Log linear models were fit with WINDL 2.0 (56). LC50 values with the diet. All assays included six to seven toxin concentrations ranging were considered significantly different if their 95% FL did not overlap. We −1 −1 from 0 to 300 μg Cry1Ac or Cry1Ab mL diet, and 0–50 μgCry2AbmL diet. estimated the resistance ratio as the LC50 of a strain divided by the LC50 of the After the diet was dry, one <24 h-old neonate was transferred onto the susceptible LAB-S strain. Maternal effects and sex linkage affecting resistance diet surface of each cell. Trays were then covered with plastic ventilated covers to Cry1Ac cotton were evaluated by comparing survival of insects from ’ 2 and incubated at 27 ± 1°C,60± 10% RH, and a photoperiod of 14L:10D. We the reciprocal crosses with a Pearson s χ test. Survival on Cry1Ac cotton did fi replicated each combination of insect population and toxin concentration four not differ signi cantly between the F1 progeny from reciprocal crosses (GA 2 to five times, with 16 neonates per replicate. Mortality was recorded after 7 d. females × GA-R males and vice versa; χ = 0.83, df = 1, P = 0.36), indicating We considered larvae dead if stimulation with a blunt needle did not elicit autosomal inheritance. Accordingly, we pooled data from the two reciprocal a coordinated response. crosses for subsequent analyses. We compared survival of strains GA, GA-R, and their hybrid progeny (F1) reared on fresh plant material of non–Bt, Cry1Ac, and two-toxin cotton using logistic regression for binary data. These analyses Survival on Bt Cotton and non–Bt Cotton Plant Material. To evaluate survival on compared survival to adulthood of the strains separately for each cultivar. non–Bt, Cry1Ac and Cry1Ac + Cry2Ab cotton, neonates of GA, GA-R, and re- Dominance (h) was evaluated from the corrected survival of the F progeny ciprocal crosses between these strains were fed on cotton in the laboratory. 1 relative to that of parental strains GA and GA-R (57). Values of h range from Reciprocal crosses produced F1 progeny designated as F1 and F1 .F was a b 1a 0 (completely recessive resistance) to 1 (completely dominant resistance). obtained by crossing GA-R males with GA females, and F was obtained by 1b Fitness costs associated with resistance to Bt toxins occur if Bt–resistant crossing GA-R females with GA males. Reciprocal mass crosses were conducted individuals have lower fitness than Bt–susceptible individuals on non–Bt with a minimum of 30 mating pairs, and were only tested on non–Bt and plants (58). Potential fitness costs associated with Cry1Ac resistance were Cry1Ac cotton. Bioassays with plants were conducted simultaneously using evaluated by comparing survival of GA, F1, and GA-R on non–Bt cotton with insects from generation 13 and 15 of GA-R and GA, respectively. logistic regression. Incomplete resistance occurs when resistant individuals fi Survival on plant material was assessed in 2010, using eld-grown cotton have lower survival on a Bt crop than on a non–Bt crop (19, 58). Incomplete planted on June 5 on nutrient-rich heavy loam soil at the West Campus resistance was assessed by comparing survival of GA on Cry1Ac or two-toxin Agricultural Center of the University of Arizona. We used Bt cotton cultivars cotton with survival on non–Bt cotton with a Pearson’s χ2 test. Levels of DP 448 B, which produces only Cry1Ac, and DP 164 B2RF, which produces incomplete resistance (I) were calculated by dividing survival of GA-R on – Cry1Ac and Cry2Ab, and non Bt cotton cultivar DP 5415 as a control. Plants Cry1Ac cotton (or two-toxin cotton) by survival of GA-R on non–Bt cotton. fl were ood irrigated and did not receive any fertilizer, as we did not see The change in Cry1Ac concentration in leaves across the growing season fi evidence of nutrient de ciency in plants as the season progressed. Pest abun- was assessed using a two-way ANOVA including the effects of cultivar dance was low throughout the growing season and no insecticides were (Cry1Ac or two-toxin cotton), time (treated as a categorical variable), and the required to protect plants. interaction between these factors. Linear contrasts were used to compare the We started the insect feeding experiment 79 days after planting (DAP), concentration of Cry1Ac between cultivars on each date. The seasonal change when cotton was bearing bolls. For the first 7 d, larvae were placed individually in Cry2Ab concentration in leaves was assessed using a one-way ANOVA. in 30 mL clear plastic cups (ProPak@) that contained a 2-cm-diameter leaf disk The log-response ratio of the CRR was not normally distributed when from the terminal on a 5 mL mixture of 2% agar (2 g of agar per 100 ml of all data (Shapiro–Wilk test, P = 0.0009) or experiments not detecting sig- water) and 0.1% sorbic acid for moisture and microbial control (29). Cups were nificant cross-resistance (P = 0.026) were considered. In both cases, we used covered with a clear plastic lid and put in a growth chamber at 27 ± 1°C, a one-tailed Wilcoxon signed-rank test to test the hypothesis that the av- 60 ± 10% RH, and photoperiod 14L:10D. We used 60–240 neonates for each erage log-response ratio was greater than 0 (54). All statistical analyses combination of cotton type and insect type (GA, GA-R, and F1 progeny). were performed in JMP 9.0 (SAS Institute).
Brévault et al. PNAS Early Edition | 5of6 ACKNOWLEDGMENTS. We thank M. Hill, R. A. Garcia, C. M. Jones, and David Crowder for providing comments on this manuscript. This study was A. Mazza for technical assistance. We thank J. Sánchez from the group of supported by U.S. Department of Agriculture (USDA) National Research Initia- A. Bravo and M. Soberón (Universidad Nacional Autónoma de México, tive Competitive Grants Program Project 2007-02227 and USDA Biotechnology Cuernavaca) for providing Cry1Ab and Cry1Ac. We thank Mark Sisterson and Risk Assessment Grant Award 2011-33522-30729.
1. National Research Council (2010) The Impact of Genetically Engineered Crops on Farm 30. Payton ME, Greenstone MH, Schenker N (2003) Overlapping confidence intervals or Sustainability in the United States (National Academies Press, Washington, DC). standard error intervals: What do they mean in terms of statistical significance? J 2. Tabashnik BE, et al. (2010) Suppressing resistance to Bt cotton with sterile insect re- Insect Sci 3:34. leases. Nat Biotechnol 28(12):1304–1307. 31. Burd AD, Bradley JR, Van Duyn JW, Gould F (2000) Resistance of bollworm, Heli- 3. Lu Y, Wu K, Jiang Y, Guo Y, Desneux N (2012) Widespread adoption of Bt cotton and coverpa zea, to CryIA(c) toxin. Proceedings, 2000 Beltwide Cotton Conferences, 4-8 insecticide decrease promotes biocontrol services. Nature 487(7407):362–365. January 2000, San Antonio, TX, eds Dugger CP, Richter DA (National Cotton Council 4. Tabashnik BE, Gassmann AJ, Crowder DW, Carrière Y (2008) Insect resistance to Bt of America, Memphis, TN), pp 923–926. crops: Evidence versus theory. Nat Biotechnol 26(2):199–202. 32. Gould F, Cohen MB, Bentur JS, Kennedy GG, Van Duyn J (2006) Impact of small fitness 5. Tabashnik BE, Van Rensburg JBJ, Carrière Y (2009) Field-evolved insect resistance to Bt costs on pest adaptation to crop varieties with multiple toxins: A heuristic model. J – crops: Definition, theory, and data. J Econ Entomol 102(6):2011–2025. Econ Entomol 99(6):2091 2099. 6. Van Rensburg JBJ (2007) First report of field resistance by stem borer, Busseola fusca 33. Carrière Y, Tabashnik BE (2001) Reversing insect adaptation to transgenic insecticidal – (Fuller) to Bt-transgenic maize. S Afr J Plant Soil 24(3):147–151. plants. P Roy Soc B-Biol Sci 268(1475):1475 1480. 7. Bagla P (2010) India. Hardy cotton-munching pests are latest blow to GM crops. Sci- 34. Crowder DW, Carrière Y (2009) Comparing the refuge strategy for managing the ence 327(5972):1439–1439. evolution of insect resistance under different reproductive strategies. J Theor Biol – 8. Storer NP, et al. (2010) Discovery and characterization of field resistance to Bt maize: 261(3):423 430. Spodoptera frugiperda (Lepidoptera: Noctuidae) in Puerto Rico. J Econ Entomol 35. Addison SJ, Rogers DJ (2010) Potential impact of differential production of the Cry2Ab and Cry1Ac proteins in transgenic cotton in response to cold stress. J Econ 103(4):1031–1038. Entomol 103(4):1206–1215. 9. Gassmann AJ, Petzold-Maxwell JL, Keweshan RS, Dunbar MW (2011) Field-evolved 36. Adamczyk JJ, Jr., Adams LC, Hardee DD (2001) Field efficacy and seasonal expression resistance to Bt maize by western corn rootworm. PLoS ONE 6(7):e22629. profiles for terminal leaves of single and double Bacillus thuringiensis toxin cotton 10. Zhang HN, et al. (2011) Early warning of cotton bollworm resistance associated with genotypes. J Econ Entomol 94(6):1589–1593. intensive planting of Bt cotton in China. PLoS ONE 6(8):e22874. 37. Greenplate JT, et al. (2003) Partial characterization of cotton plants expressing two 11. Wan P, et al. (2012) Increased frequency of pink bollworm resistance to Bt toxin toxin proteins from Bacillus thuringiensis: Relative toxin contribution, toxin in- Cry1Ac in China. PLoS ONE 7(1):e29975. teraction, and resistance management. J Appl Entomol 127(6):340–347. 12. Downes S, Parker T, Mahon R (2010) Incipient resistance of Helicoverpa punctigera to 38. Adamczyk JJ, Sumerford DV (2001) Potential factors impacting seasonlong expression the Cry2Ab Bt toxin in Bollgard II cotton. PLoS ONE 5(9):e12567. of Cry1Ac in 13 commercial varieties of Bollgard cotton. J Insect Sci 1:13. 13. Carrière Y, et al. (2012) Large-scale, spatially-explicit test of the refuge strategy for 39. Adamczyk JJ, Hardee DD, Adams LC, Sumerford DV (2001) Correlating differences in – delaying insecticide resistance. Proc Natl Acad Sci USA 109(3):775 780. larval survival and development of bollworm (Lepidoptera: Noctuidae) and fall ar- 14. Roush RT (1998) Two-toxin strategies for management of insecticidal transgenic myworm (Lepidoptera: Noctuidae) to differential expression of Cry1A(c) delta- crops: Can pyramiding succeed where pesticide mixtures have not? Philos Trans R Soc endotoxin in various plant parts among commercial cultivars of transgenic Bacillus – Lond Ser B 353(1376):1777 1786. thuringiensis cotton. J Econ Entomol 94(1):284–290. 15. Gould F (1986) Simulation models for predicting durability of insect-resitant germ 40. Greenplate JT (1999) Quantification of Bacillus thuringiensis insect control protein – plasm: A deterministic diploid, two-locus model. Environ Entomol 15(1):1 10. Cry1Ac over time in Bollgard cotton fruit and terminals. J Econ Entomol 92(6): 16. Jackson RE, Bradley JR, Jr., Van Duyn JW, Gould F (2004) Comparative production of 1377–1383. Helicoverpa zea (Lepidoptera: Noctuidae) from transgenic cotton expressing either 41. Olsen KM, Daly JC, Holt HE, Finnegan EJ (2005) Season-long variation in expression of one or two Bacillus thuringiensis proteins with and without insecticide oversprays. J Cry1Ac gene and efficacy of Bacillus thuringiensis toxin in transgenic cotton against Econ Entomol 97(5):1719–1725. Helicoverpa armigera (Lepidoptera: Noctuidae). J Econ Entomol 98(3):1007–1017. 17. Storer NP (2005) Resistance management rationale and strategy for Widestrike insect 42. Siebert MW, et al. (2009) Quantification of Cry1Ac and Cry1F Bacillus thuringiensis protection (Cry1F/Cry1Ac) in cotton. Proceedings, 2005 Beltwide Cotton Conferences, insecticidal proteins in selected transgenic cotton plant tissue types. J Econ Entomol 4-7 January 2005, New Orleans, LA, ed Richter DA (National Cotton Council of 102(3):1301–1308. America, Memphis, TN). 43. Showalter AM, Heuberger S, Tabashnik BE, Carrière Y (2009) A primer for using 18. Greenberg SM, Li YX, Liu TX (2010) Effect of age of transgenic cotton on mortality of transgenic insecticidal cotton in developing countries. J Insect Sci 9:22. lepidopteran larvae. Southwest Entomologist 35(3):261–268. 44. Zhao JZ, et al. (2005) Concurrent use of transgenic plants expressing a single and two 19. Carrière Y, Crowder DW, Tabashnik BE (2010) Evolutionary ecology of insect adap- Bacillus thuringiensis genes speeds insect adaptation to pyramided plants. Proc Natl tation to Bt crops. Evol Appl 3(5-6):561–573. Acad Sci USA 102(24):8426–8430. 20. Brévault T, Achaleke J, Nibouche S, Carrière Y (2012) Assessing the role of non-cotton 45. Zhao JZ, et al. (2003) Transgenic plants expressing two Bacillus thuringiensis toxins refuges in delaying Helicoverpa armigera resistance to Bt cotton in West Africa. Evol delay insect resistance evolution. Nat Biotechnol 21(12):1493–1497. Appl 5(1):53–65. 46. Tabashnik BE, Gould F (2012) Delaying corn rootworm resistance to Bt corn. J Econ – 21. Gould F (1998) Sustainability of transgenic insecticidal cultivars: Integrating pest ge- Entomol 105(3):767 776. netics and ecology. Annu Rev Entomol 43:701–726. 47. Ives AR, Glaum PR, Ziebarth NL, Andow DA (2011) The evolution of resistance to two- – 22. Caccia S, et al. (2010) Binding site alteration is responsible for field-isolated resistance toxin pyramid transgenic crops. Ecol Appl 21(2):503 515. to Bacillus thuringiensis Cry2A insecticidal proteins in two Helicoverpa species. PLoS 48. Carrol MW, Head G, Caprio M (2012) When and where a seed mix refuge makes sense – ONE 5(4):e9975. for managing insect resitance to Bt plants. Crop Prot 38:74 79. fi 23. Hernández-Rodríguez CS, Van Vliet A, Bautsoens N, Van Rie J, Ferré J (2008) Specific 49. Gómez I, et al. (2006) Speci c epitopes of domains II and III of Bacillus thuringiensis Cry1Ab toxin involved in the sequential interaction with cadherin and aminopepti- binding of Bacillus thuringiensis Cry2A insecticidal proteins to a common site in the dase-N receptors in Manduca sexta. J Biol Chem 281(45):34032–34039. midgut of Helicoverpa species. Appl Environ Microbiol 74(24):7654–7659. 50. Bah A, van Frankenhuyzen K, Brousseau R, Masson L (2004) The Bacillus thuringiensis 24. Burd AD, Gould F, Bradley JR, Van Duyn JW, Moar WJ (2003) Estimated frequency of Cry1Aa toxin: Effects of trypsin and chymotrypsin site mutations on toxicity and nonrecessive Bt resistance genes in bollworm, Helicoverpa zea (Boddie) (Lepidoptera: stability. J Invertebr Pathol 85(2):120–127. Noctuidae) in eastern North Carolina. J Econ Entomol 96(1):137–142. 51. Tabashnik BE, et al. (2002) Inheritance of resistance to Bt toxin crylac in a field-derived 25. Jackson RE, Bradley JR, Van Duyn JW (2004) Performance of feral and Cry1Ac-selected strain of pink bollworm (Lepidoptera: Gelechiidae). J Econ Entomol 95(5):1018–1026. Helicoverpa zea (Lepidoptera: Noctuidae) strains on transgenic cottons expressing 52. Ali MI, Luttrell RG, Young SY, 3rd (2006) Susceptibilities of Helicoverpa zea and He- one or two Bacillus thuringiensis ssp kurstaki proteins under greenhouse conditions. J liothis virescens (Lepidoptera: Noctuidae) populations to Cry1Ac insecticidal protein. J – Entomol Sci 39(1):46 55. Econ Entomol 99(1):164–175. 26. Gao YL, Wu KM, Gould F, Shen ZC (2009) Cry2Ab tolerance response of Helicoverpa 53. Burton RL (1969) Mass Rearing the Corn Earworm in the Laboratory (U.S. Department armigera (Lepidoptera: Noctuidae) populations from CrylAc cotton planting region. J of Agriculture, New Orleans, LA) Vols 33–134. – Econ Entomol 102(3):1217 1223. 54. Hedges LV, Gurevitch J, Curtis PS (1999) The meta-analysis of response ratios in ex- 27. Ali MI, Luttrell RG (2007) Susceptibility of bollworm and tobacco budworm (Lepi- perimental ecology. Ecology 80(4):1150–1156. doptera: Noctuidae) to Cry2Ab2 insecticidal protein. J Econ Entomol 100(3):921–931. 55. Finney DJ (1971) Probit Analysis (Cambridge Unive Press, Cambridge, UK), 3rd Ed. 28. Onstad DW, Meinke LJ (2010) Modeling evolution of Diabrotica virgifera virgifera 56. CIRAD (1999) WINDL, Version 2.0 (Centre de coopération Internationale en Recherche (Coleoptera: Chrysomelidae) to transgenic corn with two insecticidal traits. J Econ Agronomique pour le Développement, Montpellier, France). Entomol 103(3):849–860. 57. Liu YB, Tabashnik BE (1997) Inheritance of resistance to the Bacillus thuringiensis 29. Mahon RJ, Olsen KM (2009) Limited survival of a Cry2Ab-resistant strain of Heli- toxin Cry1C in the diamondback moth. Appl Environ Microbiol 63(6):2218–2223. coverpa armigera (Lepidoptera: Noctuidae) on Bollgard II. J Econ Entomol 102(2): 58. Gassmann AJ, Carrière Y, Tabashnik BE (2009) Fitness costs of insect resistance to 708–716. Bacillus thuringiensis. Annu Rev Entomol 54:147–163.
6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1216719110 Brévault et al. REVIEW
Insect resistance to Bt crops: lessons from the first billion acres
Bruce E Tabashnik1, Thierry Brévault2 & Yves Carrière1
Evolution of resistance in pests can reduce the effectiveness of insecticidal proteins from Bacillus thuringiensis (Bt) produced by transgenic crops. We analyzed results of 77 studies from five continents reporting field monitoring data for resistance to Bt crops, empirical evaluation of factors affecting resistance or both. Although most pest populations remained susceptible, reduced efficacy of Bt crops caused by field-evolved resistance has been reported now for some populations of 5 of 13 major pest species examined, compared with resistant populations of only one pest species in 2005. Field outcomes support theoretical predictions that factors delaying resistance include recessive inheritance of resistance, low initial frequency of resistance alleles, abundant refuges of non-Bt host plants and two-toxin Bt crops deployed separately from one-toxin Bt crops. The results imply that proactive evaluation of the inheritance and initial frequency of resistance are useful for predicting the risk of resistance and improving strategies to sustain the effectiveness of Bt crops.
Transgenic crops are one of the most widespread and controversial Here we summarize the theory for managing pest resistance to Bt applications of biotechnology1–4. To reduce reliance on insecticide crops, outline new criteria for categorizing evidence of field-evolved sprays, scientists have genetically engineered corn and cotton plants to resistance, review the global status of resistance to Bt crops based on make insecticidal proteins encoded by genes from the common bacte- current field monitoring data, and test the correspondence between rium Bacillus thuringiensis (Bt)5. These Bt proteins kill some devastat- theoretical predictions and observed patterns of field-evolved resis- ing insect pests, but cause little or no harm to most other organisms, tance. The criteria for categorizing field-evolved resistance described including people4,5. Benefits of Bt crops include reduced insecticide and applied here explicitly acknowledge that resistance is not ‘all or use, pest suppression, conservation of beneficial natural enemies, none’, which facilitates objective classification of monitoring data and increased yield and higher farmer profits6–12. The area planted with may help to gauge management actions appropriately, depending on the Bt crops worldwide increased from 1.1 million hectares in 1996 to severity of resistance. Compared with previous reviews on this topic, 66 million hectares in 2011, with a cumulative total of more than 420 mil- the field monitoring data analyzed here are more recent and represent lion hectares (>1 billion acres) (Fig. 1). Bt corn accounted for 67% of corn more cases (24 in all), as well as larger and more diverse sets of Bt toxins planted in the United States during 2012 (http://www.ers.usda.gov/Data/ (six toxins from four Cry families) and pest species (13 species from BiotechCrops/) and Bt cotton accounted for 79–95% of cotton planted in two insect orders). Using data from 77 studies published as of 2012, we Australia, China, India and the United States during 2010 to 2012 (Fig. 2). report the first statistical analyses of the association between observed The remarkable ability of insects to adapt to insecticides and other global patterns of field-evolved resistance and predicted effects of two control tactics supports the conclusion that evolution of resistance by key biological parameters: dominance of resistance and initial resistance pests is the main threat to the continued success of Bt crops13–23. Many allele frequency. The results provide insights that can be used proactively previous reviews have addressed pest resistance to Bt crops13–23, includ- to improve resistance management. ing a 2011 mini-review emphasizing four successful cases of the high- dose⁄refuge resistance management strategy in North America22 and our Theory for managing pest resistance to Bt crops 2009 review of 17 cases involving 11 species of lepidopteran pests and The refuge strategy has been the primary approach used worldwide to four Bt toxins (B.E.T., Van Renburg, J.B.J. & Y.C.)21. Several papers have delay pest resistance to Bt crops and has been mandated in the United compared field outcomes for resistance to Bt crops with predictions from States, Australia and elsewhere8,16,23. Despite implementation of some theory, but the rigor of these previous comparisons has been limited by resistance management practices for conventional insecticides, the man- small sample sizes for both the field outcomes and the factors predicted dates for the refuge strategy are part of an unprecedented proactive effort to affect resistance19–22. to slow resistance to Bt crops that recognizes both their value and the strong threat of resistance. The concept underlying the refuge strategy is 1Department of Entomology, University of Arizona, Tucson, Arizona, USA. that most of the rare resistant pests surviving on Bt crops will mate with 2Centre de coopération Internationale en Recherche Agronomique pour le the relatively abundant susceptible pests from nearby refuges of host Développement, UPR 102, Montpellier, France. Correspondence should be plants without Bt toxins24–27. If inheritance of resistance is recessive, the addressed to B.E.T. ([email protected]). progeny from such matings will die on Bt crops, substantially delaying Received 24 October 2012; accepted 26 March; published online 7 June 2013; the evolution of resistance. This approach is sometimes called the ‘high- doi:10.1038/nbt.2597 dose refuge strategy’ because it works best if the dose of toxin for insects
NATURE BIOTECHNOLOGY VOLUME 31 NUMBER 6 JUNE 2013 1 REVIEW
70 7 specify that high-dose Bt plants should kill at least 99.99% of susceptible Bt crops 31 Resistant species insects in the field . 60 6 Mathematical modeling consistently predicts that resistance
✽ will evolve more slowly if the initial resistance allele frequency is 50 5 low15,24,25,32,33. Although several papers22,34–38 propose (without the- ported)
re oretical or empirical evidence) that the success of the refuge strategy 40 4 requires an initial resistance allele frequency ≤0.001, modeling results
30 3 imply that the refuge strategy can be useful with much higher resis- tance allele frequencies, particularly if fitness costs are associated with
Resistant species 24,28,33,39,40
crops (million ha) crops 20 2 resistance . For example, with recessive resistance and fitness Bt
(reduced efficacy efficacy (reduced costs, refuges delayed resistance substantially in a model with an initial 10 1 resistance allele frequency of 0.3 (Y.C. & B.E.T.)28. Refuge abundance can be measured for each pest in terms of the 0 0 percentage of its host plants that are non-Bt plants. If more than one
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 species of non-Bt host plant is available, the ‘effective’ refuge percentage can be estimated by adjusting for the relative abundance of susceptible Figure 1 Planting of Bt crops globally and field-evolved resistance. Planting pests produced on different host plant species41–45. The effective refuge of Bt crops globally each year and cumulative number of insect species with percentage can also be adjusted downward for the effects of treating field-evolved resistance and reduced efficacy reported. Planting of Bt crops refuges with insecticides, because such treatments reduce the ability increased from 1.1 million hectares (ha) in 1996 to 66 million ha in 2011 41 (ref. 2). Field-evolved resistance associated with reduced efficacy of Bt crops of refuges to delay resistance . Because pest movement and mating has been reported for five major target pests (year first detected): H. zea patterns interact with the distribution and abundance of refuges and Bt (2002), S. frugiperda (2006), B. fusca (2007), P. gossypiella (2008) and crop fields to affect evolution of resistance, spatially explicit approaches D. v. virgifera (2009) (Tables 1 and 2). *, For 2011, the number of species are useful for assessing refuge effectiveness46. with resistant populations may be underestimated because reports of field- Modeling results suggest that when inheritance of resistance is not evolved resistance typically are published 2 or more years after resistance is recessive, increasing refuge abundance can still substantially delay resis- first detected. tance. For example, results from a single-locus, two-allele model of a generic pest with an initial resistance allele frequency of 0.001 suggest eating Bt plants is high enough to kill all (or almost all) of the offspring that resistance can be delayed for >20 years with ≥5% refuges if resistance from matings between resistant and susceptible insects16 (B.E.T., Y.C. et is completely recessive (h = 0) and with >50% refuges if resistance is al.)15,27. Therefore, in theory, three key factors favor success of the refuge partially dominant (h ≥ 0.4) (ref. 20). strategy: first, recessive inheritance of resistance; second, low resistance First-generation Bt crops each produce a single Bt toxin, but many allele frequency; and third, abundant refuges of non-Bt host plants near second-generation Bt crops, named pyramids, produce two or more Bt crops16 (B.E.T., Y.C. et al.)15,21. Two additional factors predicted to delay 100 100 resistance are fitness costs and incomplete Australia 90 90 China 16 19,28,29 Cry1Ac Cry1Ac resistance (B.E.T., Y.C. et al.) . Fitness 80 80 Cry1Ac + Cry2Ab costs occur when fitness on non-Bt host plants 70 70 is lower for resistant insects than susceptible 60 60 insects, so that refuges select against resis- n (%) 50 50 to tance28,29. Incomplete resistance occurs when 40 40 30 30 resistant insects can complete development on Bt cot 20 20 Bt plants, but they are at a disadvantage com- 10 10 pared with resistant insects that develop on 0 0 non-Bt plants19,28. 100 100 The dominance of resistance on a Bt crop 90 India 90 United States Cry1Ac Cry1Ac plant can be measured in terms of the param- 80 80 Cry1Ac + Cry2Ab Cry1Ac + Cry2Ab or Cry1F eter h, which varies from 0 for completely 70 70 60 60 recessive to 1 for completely dominant16 (Liu, n (%) 50 50
30 to Y.B. & B.E.T) . When such direct data are not 40 40 available, dominance can be assessed indirectly 30 30 Bt cot by measuring survival of susceptible insects 20 20 on Bt plants31. This indirect assessment relies 10 10 on the idea that if Bt plants do not kill all or 0 0 96 98 nearly all homozygous susceptible insects, 1996199719981999200020012002200320042005200620072008200920102011 19 199719 19992000200120022003200420052006200720082009201020112012 they probably will not kill nearly all individu- Year planted Year planted als heterozygous for resistance. If so, survival is likely to be higher for the heterozygotes than Figure 2 Percentage of cotton hectares planted with Bt cotton producing one toxin (gray) or two toxins (black) in four countries. All Bt cotton produced Cry1Ac. In Australia and India, all two-toxin cotton for the homozygous susceptible insects, which produced Cry1Ac and Cry2Ab. In the United States from 2004 to 2012, 86% of two-toxin cotton yields nonrecessive inheritance of resistance produced Cry1Ac and Cry2Ab and 14% produced Cry1Ac and Cry1F. The ranking of each country in terms that accelerates adaptation16,27. Thus, the US of 2012 cotton production (percentage of world production) was 1 for China (27%), 2 for India (22%), 3 Environmental Protection Agency guidelines for the United States (15%) and 7 for Australia (3.4%) (see Supplementary Methods for details).
2 VOLUME 31 NUMBER 6 JUNE 2013 NATURE BIOTECHNOLOGY REVIEW
Box 1 An alternative definition of resistance
The Insecticide Resistance Action Committee (IRAC), composed resistance objectively and for proactive detection and responses of members from more than a dozen major agrochemical and to resistance. First, by the time a product has failed repeatedly, biotech companies, aims to promote resistance management it is usually too late to respond most effectively to resistance. strategies for insecticides and Bt crops to support sustainable Second, the “expected level of control” is not specified, which agriculture and improve public health (http://www.irac-online. allows variation in interpretation, including changes over time in org/about/irac/). IRAC defines resistance as “a heritable change expectations. Third, because the definition depends on the label in the sensitivity of a pest population that is reflected in the recommendation, resistance cannot occur in any species that is repeated failure of a product to achieve the expected level of not on the label, which excludes evolution of resistance in non- control when used according to the label recommendation for target pests and non-pest species96,104. By contrast, the term that pest species” (http://www.irac-online.org/about/resistance/). “field-evolved resistance” as defined here explicitly recognizes The first part of the IRAC definition, “a heritable change in the that resistance results from evolution, enables objective sensitivity of a pest population” and the definition of field-evolved identification of resistance, facilitates proactive detection and resistance (see main text) both emphasize a genetically based management of resistance, and applies to resistance in pest and decrease in susceptibility. The remainder of the IRAC definition beneficial organisms. Various other definitions of resistance have sets additional conditions that are problematic for identifying been proposed and discussed in depth elsewhere105,106. distinct Bt toxins that are active against the same pest21,47. The assump- Pest control problems associated with field-evolved resistance vary tion underlying this approach (which is not always true) is that selection from none to severe, depending on the frequency of resistant individu- for resistance to one toxin does not cause cross-resistance to the other als, the extent to which resistance increases survival in the field, the toxins in the pyramid, so that insects resistant to all toxins in the pyramid geographical distribution of resistant populations, the insect’s popula- are extremely rare47,48. Other factors favoring success of pyramids match tion density and the availability of alternative controls21. We define four those listed above for the refuge strategy, including abundant refuges categories of field-evolved resistance: 1) >50% resistant individuals and and the following conditions for each toxin in the pyramid: recessive reduced efficacy of the Bt crop in the field has been reported; 2) >50% inheritance of resistance, low initial resistance allele frequency, fitness resistant individuals and reduced efficacy is expected, but has not been costs and incomplete resistance21,39,47,48. Modeling results and small- reported; 3) 1–6% resistant individuals; and 4) <1% resistant individu- scale experiments with noncommercial Bt broccoli plants indicate that als. For categories 3 and 4, the percentage of resistant individuals is low resistance to pyramids evolves faster if one-toxin plants are grown con- enough that reduced efficacy of the Bt crop in the field is not expected. currently with two-toxin plants47. This occurs because the one-toxin We adopt terms used previously and refer to cases in category 3 as an plants select for resistance to each toxin separately, which reduces the ‘early warning’ of resistance52 and cases in category 4 as ‘incipient resis- advantage of the two-toxin plants47. tance’53. In principle, an additional category could be 6–50% resistant individuals, but none of the cases reviewed here was in that range. The Field-evolved resistance: criteria and categories fifth category is cases in which monitoring data show no statistically Field-evolved (or field-selected) resistance is a genetically based decrease significant decrease in susceptibility. in susceptibility of a population to a toxin caused by exposure of the Each case reviewed here involves evaluation of field-evolved resis- population to the toxin in the field14,21,49. A Web of Science search with tance to one Bt toxin in populations of one pest species from one ‘topic = field-evolved resistance’ identified 54 publications, starting with country. Although initial detection of resistance can be based on data two 1996 papers about resistance to Bt toxins produced by two indepen- from a single field population, all of the cases of field-evolved resis- dent research teams50,51 and including 31 papers published from 2010 to tance reviewed here entail evidence of genetically based, decreased 2012. These 54 publications were authored by >150 academic, govern- susceptibility from several field populations. To classify each case ment and industry scientists from five continents and have been cited of field-evolved resistance into one of the four categories we define >900 times, including 300 citations in 2012. Despite this widespread and above, we estimated the percentage of individuals resistant to a toxin increasing use of the term ‘field-evolved resistance’, some scientists favor based on survival on a diet treated with a ‘diagnostic concentration’ of alternative terms and definitions for resistance (Box 1). the toxin that kills all or nearly all susceptible individuals, or survival Natural genetic variation affecting responses to Bt toxins usually on intact Bt plants or parts of Bt plants containing that toxin. For occurs in insect populations, with some alleles conferring susceptibility many cases, we also report the resistance ratio, which is the concen- and others conferring resistance. Before an insect population is exposed tration of toxin killing 50% of insects tested (LC50) for a field-derived 16,19 to a Bt toxin, alleles conferring resistance are typically rare . Field- strain divided by the LC50 for a conspecific susceptible strain. Large evolved resistance occurs when exposure of a field population to a toxin increases in LC50 in field-selected populations yield resistance ratios increases the frequency of alleles conferring resistance in subsequent >10 and indicate that >50% of the population is resistant21. generations21. Thus, detecting resistance alleles without demonstrating that their frequency has increased in field populations does not consti- Resistance monitoring data tute evidence of field-evolved resistance21. Here we review 24 cases for which resistance monitoring data are pub- The primary goal of monitoring resistance to Bt crops is to detect lished in peer-reviewed journals for 13 species of major lepidopteran field-evolved resistance early enough to enable proactive management and coleopteran pests that are targeted by six Bt toxins in transgenic before field failures occur21. Resistance monitoring includes sampling corn and cotton in eight countries (Tables 1 and 2 and Supplementary and testing of insects that survive on Bt crops as well as insects from Tables 1–4). other sources, including non-Bt host plants. Failure to sample insects from Bt crops favors underestimation of the frequency of resistance, Reduced efficacy reported or expected. The cumulative number of which can postpone detection of resistance21. major pest species with field-evolved resistance to Bt toxins in crops
NATURE BIOTECHNOLOGY VOLUME 31 NUMBER 6 JUNE 2013 3 REVIEW
Table 1 Evaluation of field-evolved resistance in 24 cases involving 13 species of major pests targeted by Bt cropsa Pesta Bt crop Toxin Country Yearsb High dosec Low initial freq.d >50% resistant individuals and reduced efficacy reported B. fusca Corn Cry1Ab South Africa 8 No ?e D. v. virgifera Corn Cry3Bb USA 7 No No H. zea Cotton Cry1Ac USA 6 No No P. gossypiella Cotton Cry1Ac India 6f No ? S. frugiperda Corn Cry1F USA 3 No ? >50% resistant individuals and reduced efficacy expected H. zea Cotton Cry2Ab USA 2g No No 1–6% resistant individuals D. saccharalis Corn Cry1Ab USA 10 No No H. armigera Cotton Cry1Ac China 13 No No O. furnacalis Corn Cry1Ab The Philippines 5 No ? P. gossypiella Cotton Cry1Ac China 13 No ? <1% resistant individuals H. armigera Cotton Cry1Ac Australia 15 No Yes H. armigera Cotton Cry2Ab Australia 8 Yes No H. punctigera Cotton Cry2Ab Australia 8 Yes No No decrease in susceptibility D. grandiosella Corn Cry1Ab USA 6 ? Yes D. v. virgifera Corn Cry34/35Ab USA 4 No No H. punctigera Cotton Cry1Ac Australia 10 ? Yes H. virescens Cotton Cry1Ac USA 11 Yes No H. virescens Cotton Cry1Ac Mexico 11 ?? H. virescens Cotton Cry2Ab USAfa 2 Yes ? O. nubilalis Corn Cry1Ab USA 15 No Yes O. nubilalis Corn Cry1Ab Spain 4 ? ? P. gossypiella Cotton Cry1Ac USA 13 Yes No P. gossypiella Cotton Cry2Ab USA 5 Yes Yes S. nonagroides Corn Cry1Ab Spain 7 ? Yes aSee Table 2 and Supplementary Tables 1–6 for details. D. v. virgifera is a coleopteran; the other 12 pests are lepidopterans. bYears elapsed between the first year of commercialization in the region studied and: (i) for the six cases with >50% resistant individuals and reduced efficacy reported or expected (red and orange), the first year of field sampling that yielded evidence of resis- tance, or (ii) for all other cases, the most recent year of monitoring data reviewed here (see Table 2 and Supplementary Tables 1–4 for details). cBased on direct evaluation of recessive inheritance of resistance for 12 cases with relevant data and on survival of susceptible individuals on Bt plants for 7 cases without such direct data (Supplementary Table 5). dBased on an initial resistance allele frequency below the detection threshold; yes indicates initial screening did not detect any major resistance alleles (Supplementary Table 6). e”?” indicates answer could not be determined with available data. fExcludes years when Bt cotton was grown illegally in India before it was commercialized in 2002 (refs. 81,91,92). Resistance was first detected in samples collected in 2008, 6 years after commercialization. If illegal planting started in 2000, the total years elapsed would be 8. gMay reflect some cross-resistance caused by selection with Cry1Ac21,93,94. and reduced transgenic crop efficacy increased from one in 2005 to was detected fewer than 10 years after the Bt crop was commercial- five in 2010 (Figs. 1, 3 and 4). These five cases include resistance to ized. In a sixth case of field-evolved resistance of H. zea to Cry2Ab in Bt corn in three pests (Busseola fusca, Diabrotica virgifera virgifera Bt cotton, the percentage of resistant individuals exceeded 50% for and Spodoptera frugiperda) and resistance to Bt cotton in two pests several populations and reduced efficacy is expected, but has not yet (Helicoverpa zea and Pectinophora gossypiella; Tables 1 and 2 and been reported (Box 4). Boxes 2 and 3). In each of these five cases, field-evolved resistance Early warning: 1–6% resistant individuals. In four cases, the percentage of resistant individuals increased significantly to reach 1–6%, which is not expected to reduce efficacy in the field (Table 1 and Supplementary Table 2). For field populations exposed to Cry1Ac cotton in China for Reduced efficacy 13 years, maximum survival at a diagnostic concentration of Cry1Ac 52 54 1–6% resistant was 2.6% for Helicoverpa armigera and 5.6% for P. gossypiella , with <1% resistant 0% survival for susceptible control populations. For Ostrinia furnicalis Susceptible exposed to Cry1Ab corn in the Philippines, the maximum survival at a diagnostic concentration of Cry1Ab increased 14-fold from 0.4% in 2007 to 5.5% in 2009 (ref. 55). For Diatraea saccharalis in Louisiana, 20052010 data from F2 screens show the frequency of alleles conferring resistance to Cry1Ab corn increased eightfold from 0.0023 in 2004 to 0.018 in Figure 3 Resistance of major pest species to Bt crops in 2005 and 2010. 2009 (ref. 56). We estimate that the percentage of resistant individu- For each pest species, the color indicates the status of the most resistant population. In 2005, the only pest with resistant field populations was H. als in the populations sampled in Louisiana in 2009 was 1.0–2.4%, zea; the other eight pests evaluated were susceptible. Data for 2005 (n = based on the partial dominance of resistance (Supplementary Table 9 species) are from reference 21. Data for 2010 (n = 13 species) are from 5) and an estimated frequency of heterozygous individuals of 0.031 Table 1. (Supplementary Table 2).
4 VOLUME 31 NUMBER 6 JUNE 2013 NATURE BIOTECHNOLOGY REVIEW
Table 2 Bioassay data indicating field-evolved resistance to the toxins in Bt crops for five pests with >50% resistant individuals and reduced efficacy reporteda Bioassay data Strains Control Pest Cry toxin Country Year comm.b testedc Initial yeard Final yeare Parameter valuee Test valuef References B. fusca 1Ab S. Africa 1998 2 2006 2006 Max. surv.g 0.0% 64% 21,95 B. fusca 1Ab S. Africa 1998 8 2007 2007 Max. surv.h 0.0% 88% 96 H. zea 1Ac USA 1996 2 2002 2002 Max. surv.i 0.0% 52% 97,98 H. zea 1Ac USA 1996 64 1992 2004 Max. RRj 1.2 580 59,60 H. zea 1Ac USA 1996 197 2002 2006 Max. RRj 40k >1,000 59,97,99 P. gossypiella 1Ac India 2002l 6 2007 2009 Survivalm 2.0% 72% 100 P. gossypiella 1Ac India 2002l 2 2007 2009 Max. RRj 1.0 47 100 S. frugiperda 1F USA 2003 8 1990 2008 Max. RRn 1.0 >356 75 S. frugiperda 1F USA 2003 13 2010 2011 Max. RRo 1.0 >970 76 S. frugiperda 1F USA 2003 13 2010 2011 Survivalp 0.8% 90% 76 D. v. virgifera 3Bb USA 2003 9 2009 2009 Survivalq 17%r 52%s 101 D. v. virgifera 3Bb USA 2003 13 2010 2010 Survivalq 5%t 74%u 102 aCry1Ab, Cry1F and Cry3Bb produced by Bt corn; Cry1Ac produced by Bt cotton. bFirst year Bt crop was grown commercially in the location monitored. cTotal number of field-derived strains tested in bioassays. dInitial and final years during which field populations were sampled. eValue for parameter from one or more susceptible strains, based on initial year unless noted otherwise. fValue for parameter from one or more field-selected resistant strains, based on final year unless noted otherwise. gMaximum survival in the field at 18 days on Bt corn plants relative to non-Bt corn plants for a susceptible field population (control value) and a resistant field popula- tion (test value). hSurvival in the greenhouse at 35 days on Bt corn plants relative to non-Bt corn plants for the most susceptible field population (control value) and the most resistant field population (test value). iSurvival at 4 days on Bt cotton leaves relative to non-Bt cotton leaves for a susceptible lab strain (control value) and the most resistant field population (test value); in addition, four strains derived from the field in 2004 had >50% survival at a diagnostic concentration of Cry1Ac in 59 j k diet . Maximum resistance ratio, the highest LC50 for a field-derived strain divided by the LC50 for one or more susceptible strains. The maximum resistance ratio of 40 in 2002 reflects field-evolved resistance that was detected in that year. lBt cotton was grown illegally in Gujarat, India, for at least 2 years before it was com- mercialized81,91,92. mMean survival at a concentration of 1 microgram Cry1Ac per ml diet in lab bioassays; resistance detected in a population from Amreli, Gujarat, sampled in 2008. nField-selected resistant strains were derived from four populations sampled in Puerto Rico during 2007 to 2008 (test value); the control value is based on a strain derived from a field population sampled from Georgia in 2005. The most resistant field populations had <50% mortality and growth inhibition at the highest concentration tested, yielding a maximum resistance ratio >35 based on LC50 values and >356 based on the concentration causing 50% growth inhibition o (IC50). Field-selected resistant strains were derived from four populations sampled in Puerto Rico during 2010 and 2011 (test value); the control value is the mean for nine strains derived from US mainland populations sampled during 2010 and 2011. pMean survival at 7 days on Bt corn leaves relative to non-Bt corn leaves for the strains derived in 2010 and 2011 from Puerto Rico (test value) and the US mainland (control value). qMean survival on Cry3Bb corn plants relative to non-Bt corn plants in lab bioassays. rMean for five strains derived in 2010 from “control” fields in Iowa where severe corn rootworm damage was not seen. sMean for four strains derived in 2009 from four “problem” fields in Iowa where growers reported severe corn rootworm injury to Bt corn fields planted with Cry3Bb corn (3 fields) or a com- bination of Cry3Bb corn and Cry34/35Ab corn (1 field). tMean for six control strains derived before Cry3Bb corn was commercialized (1995–2001) from fields in four states. uMean for seven strains derived in 2010 from problem fields in Iowa.
Incipient resistance: <1% resistant individuals. For three cases that entail 2010–2011 (ref. 57). These results show that the statistically significant H. armigera and Helicoverpa punctigera exposed to Bt cotton in Australia, yet small rises in resistance allele frequency characteristic of incipient F1 and F2 screens detected statistically significant increases over time resistance do not necessarily indicate that further increases in resistance in the frequency of alleles conferring resistance to Cry1Ac or Cry2Ab, are imminent. yet the highest estimated percentage of resistant individuals was <1% (Table 1 and Supplementary Table 3). Because of the low percentage of No decrease in susceptibility. Monitoring data provide strong evidence resistant individuals after many years of extensive exposure to Bt cotton, of no decrease in susceptibility to toxins produced by Bt crops in 11 these three cases exemplify successful resistance management. cases (Table 1 and Supplementary Table 4). These cases include evi- Among these three cases, the maximum resistance allele frequency dence of no decreased susceptibility to each Bt toxin tested against all detected is 0.048, based on results of the F1 screen for H. punctigera populations examined for four species: Diatraea grandiosella, Heliothis resistance to Cry2Ab in Australia in 2008–2009 (refs. 53,57,58). Results virescens, Ostrinia nubilalis and Sesamia nonagrioides. In addition, the from F1 screens conducted in 2008–2009 also showed that the frequency of alleles conferring resistance to Cry2Ab was eight times higher in areas where Bt cotton was grown relative to non-cropping areas (P < 0.0001, ref. 53). Based on this difference between areas in the same season and a significantly increased frequency of resistance to Cry2Ab over time in Bt cot- ton growing areas, Downes et al.53 termed this Reduced efficacy >50% resistant “incipient resistance.” Based on recessive inher- 1–6% resistant itance and Hardy-Weinberg equilibrium, they <1% resistant Susceptible estimated 0.2% (0.0482) of H. punctigera larvae were resistant to Cry2Ab in 2008–2009, which is too low to reduce the efficacy of Bt cotton in Figure 4 Global status of field-evolved resistance to Bt crops. Each circle represents 1 of 24 cases the field. Moreover, the frequency of resistance involving evaluation of field-evolved resistance to one toxin in Bt corn or Bt cotton in populations of one to Cry2Ab did not increase from 2008–2009 to pest species from one country (Tables 1 and 2 and Supplementary Tables 1–4).
NATURE BIOTECHNOLOGY VOLUME 31 NUMBER 6 JUNE 2013 5 REVIEW
Box 2 Field-evolved resistance to Bt corn with reduced efficacy reported
Field-evolved resistance of S. frugiperda (fall armyworm) to Bt corn efficacy in the field preceded documentation of resistance with producing Cry1F occurred in 3 years in the United States territory of bioassays73–76,101,102,107. Puerto Rico75,76 (Tables 1 and 2). This is the fastest documented Bt corn producing Cry3Bb to kill beetles, particularly D. v. virgifera case of field-evolved resistance to a Bt crop with reduced efficacy (western corn rootworm), was first registered in the United States reported and is consistent with worst-case scenarios envisioned in in 2003 (ref. 108). By 2009, farmers planted Cry3Bb corn on 13 1997 by some experts32,68. It is also the first case of resistance million ha, which was 36% of all corn in the United States100,109. leading to withdrawal of a Bt crop from the marketplace. High levels Field and laboratory data show that control problems in the field of resistance persisted in Puerto Rico in 2011, 4 years after Cry1F during 2009 and 2010 were associated with resistance to Cry3Bb in corn had been voluntarily withdrawn from sales76. some Iowa populations of D. v. virgifera110–112. In ‘problem’ fields, Field-evolved resistance to Bt corn producing Cry1Ab occurred which had severe damage to Cry3Bb corn caused by rootworms, in B. fusca (maize stem borer) in South Africa in 8 years21,101 Cry3Bb corn had been planted for 3–7 years110,111. A 2011 field (Tables 1 and 2) and has some striking parallels with S. frugiperda study of two of the problem fields identified in 2009 found that resistance to Cry1F corn. In both cases, proactive resistance D. v. virgifera emergence did not differ significantly between monitoring was not conducted and anecdotal evidence of reduced Cry3Bb corn and non-Bt corn112. data show no decreased susceptibility of D. v. virgifera to Cry34/35Ab Testing theory with data and of P. gossypiella to Cry1Ac and Cry2Ab in Arizona. Three cases in The data from resistance monitoring studies reviewed here generally the United States show no decrease in susceptibility after ≥10 years of confirm the main predictions from the evolutionary theory underly- exposure to a toxin produced by a Bt crop: O. nubilalis to Cry1Ab in Bt ing the refuge and pyramid strategies for managing pest resistance to corn, and H. virescens and P. gossypiella to Cry1Ac in Bt cotton. Overall, Bt crops. As detailed below, resistance was less likely to evolve rapidly 5 of the 24 cases show no decrease in susceptibility after ≥10 years of if the high-dose standard was met (indicating recessive inheritance of exposure to the Bt crop; and 14 of the 24 cases (58%) show either no resistance), the initial resistance allele frequency was low, refuges were decrease in susceptibility (11 cases) or <1% resistant individuals (3 cases) abundant and Bt plants with two-toxin pyramids were grown separately after 2 to 15 years (mean = 9 years) (Table 1). from one-toxin Bt plants.
Box 3 Field-evolved resistance to Bt cotton with reduced efficacy reported
Both cases of field-evolved resistance to Bt cotton with reduced commercialization in that region59,117,118. The extensive evidence efficacy reported (Tables 1 and 2) have been controversial. confirming this case of resistance includes >50% survival at a In India, Bt cotton hybrids generated by crossing a Bt cotton diagnostic concentration for four strains derived from the field in cultivar with local non-Bt cotton cultivars were commercialized 2003 (refs. 59,115). in 2002, but illegal planting of Bt cotton hybrids began sooner One of the primary arguments disputing the conclusion of field- in the western state of Gujarat91,92. Resistance of P. gossypiella evolved resistance in this case was that “larval samples should (pink bollworm) to Bt cotton producing Cry1Ac (Bollgard) was not be collected from Bt crops” for resistance monitoring99. As first detected with laboratory bioassays of the offspring of insects noted above, testing insects sampled from Bt crops is critical collected from the field in 2008 in Gujarat97 (Table 2). Monsanto for monitoring resistance. Moreover, the evidence in this case (St. Louis) reported that their 2009 field monitoring confirmed documents resistance in samples from non-Bt crops as well as from P. gossypiella resistance to Cry1Ac in four districts of Gujarat113. Bt crops, including a strain derived from non-Bt cotton in 2004 This resistance seen in laboratory bioassays was associated with that had a resistance ratio >500 (refs. 59,115,119). Another unusually high abundance of larvae on Cry1Ac cotton and moths challenge was that the evidence of field-evolved resistance came caught in pheromone traps98,113,114. entirely from the laboratory99. However, “unacceptable levels of A prominent Indian entomologist challenged the conclusion boll damage” observed in problem fields were associated with of field-evolved resistance, claiming that resistance monitoring decreased susceptibility to Cry1Ac in laboratory bioassays115,117, should be based only on insects collected from non-Bt cotton, similar to the evidence from India97,98,113. yet Monsanto had collected larvae from Bt cotton plants98,114. In 2012, Luttrell and Jackson118 asserted that selection of Monsanto aptly countered this criticism by stating that their H. zea resistance to Cry1Ac in the laboratory before Bt cotton resistance monitoring based on insects collected from Bt was commercialized “argues against conclusions of field-evolved cotton in India is “standard practice”98. Indeed, testing insects resistance.” Yet, the selection experiment they cite60 demonstrates collected from Bt plants is an essential component of resistance that resistance alleles were present, but not common, before Bt 21,75,101,107,110–112,115 monitoring . cotton was commercialized. This is reflected in the low LC50 of Ironically, some of the dubious arguments disputing Monsanto’s Cry1Ac before laboratory selection for the strain derived from
report of P. gossypiella resistance to Cry1Ac cotton in India the field in 1992, and the >100-fold increase in LC50 of this mirror those offered by Monsanto and others99 to challenge strain caused by seven generations of selection with Cry1Ac60. documentation of H. zea (bollworm) resistance to Cry1Ac cotton In the United States, the registration for Cry1Ac cotton expired in the United States116. Researchers discovered the initial in September 2009 (ref. 21) and this product was replaced evidence of field-evolved resistance of H. zea to Cry1Ac cotton progressively from 2003 to 2011 by cotton that produces two Bt in the southeastern United States in 2002, 6 years after its toxins, either Cry1Ac and Cry2Ab or Cry1Ac and Cry1F (Fig. 2).
6 VOLUME 31 NUMBER 6 JUNE 2013 NATURE BIOTECHNOLOGY REVIEW
Box 4 Field-evolved resistance to Bt cotton with reduced efficacy likely
Like both cases of field-evolved resistance to Bt cotton producing 2011, the percentage of Bt cotton producing two toxins increased Cry1Ac (Box 3), the case of H. zea resistance to Cry2Ab in the from 0% to 90% (Fig. 2), whereas the number of sprayings against southeastern United States has been controversial. The initial H. zea on Bt cotton tripled121. data documenting resistance in this case show a significant In the five states of the midsouth region, sprays for H. zea per increase in the proportion of populations screened that had an hectare of Bt cotton were relatively low from 2004 to 2007 (mean
LC50 value greater than the diagnostic concentration of toxin = 0.75, s.e.m. = 0.04), compared with 2000–2003 and 2008– (150 mg Cry2Ab per ml diet), which indicates >50% survival at 2010 (data from ref. 118; mean = 1.2, s.e.m. = 0.09; t = 3.9, the diagnostic concentration21,93 (Supplementary Table 1). Based df = 9, P < 0.01). One explanation for this pattern is that fewer on this criterion, the percentage of H. zea populations tested that sprays were needed during 2004 to 2007 because two-toxin plants were resistant to Cry2Ab rose from 0% in 2002 to 50% in 2005, producing Cry1Ac and Cry2Ab initially had relatively high efficacy only 2 years after commercialization of Bt cotton producing against H. zea, but their efficacy declined because of resistance Cry2Ab and Cry1Ac21,93. The percentage of populations with a to Cry2Ab. An alternative hypothesis is that sprays increased resistance ratio >10 also increased from 0% in 2002 to 50% because of increased planting of corn, which is a preferred host in 2005 (refs. 21,93). Three populations sampled from non-Bt for H. zea118. However, we found no association between the area plants in Arkansas in 2005 had such low mortality in bioassays planted to corn and sprays for H. zea on Bt cotton in the midsouth 2 that LC50 values could not be calculated, but were estimated to from 1999 to 2010 (data from ref. 118; r = 0.01, df = 10, P = be >400 mg Cry2Ab per ml diet93. The decreased susceptibility to 0.76). We also found no association between the area planted to Cry2Ab detected in 2005, when cotton producing this toxin was corn and sprays for H. zea on all cotton in Arkansas, Georgia and not common (Fig. 2), suggests that resistance to Cry1Ac caused Mississippi from 2000 to 2011 (Supplementary Tables 7–9 and some cross-resistance to Cry2Ab93, which is consistent with data Supplementary Fig. 1). showing a genetic correlation between resistance to these two Because some susceptible individuals can complete toxins94. development in the field on cotton producing Cry1Ac and In addition, data from Arkansas show that mortality caused by a Cry2Ab122 and resistance to Cry1Ac in diet tests is associated diagnostic concentration of Cry2Ab decreased substantially in 2010 with increased survival on cotton leaves containing both compared with the previous 4 years for field populations relative toxins, the increased survival of field-selected strains on diet to a susceptible laboratory strain120. This evidence of resistance treated with diagnostic concentrations of Cry1Ac and Cry2Ab is to Cry2Ab coincided with higher abundance of H. zea in the field probably associated with increased survival in the field on cotton and increased insecticide sprays targeting H. zea on Bt cotton in plants producing both of these toxins21,119. Although Luttrell 2010 (ref. 120). For the entire United States, the mean number of and Jackson118 state that they did not find strong evidence insecticide sprays per hectare of Bt cotton directed primarily at of “sustained loss of field control or increased resistance H. zea nearly doubled in 2009–2011 (0.88, s.e.m. = 0.1) compared levels over time,” they conclude, “From a practical farm-level with 1999–2008 (0.48, s.e.m. = 0.03) (data from ref. 121; t-test, perspective, effective control of bollworm [H. zea] requires t = 4.9, df = 11, P < 0.001). In the United States from 1999 to supplemental insecticides, even on dual-gene Bt cottons.”
High dose. Field outcomes show that resistance was less likely to evolve The other two exceptions involve cases with <1% resistant individu- quickly if plants met the high-dose standard indicating that resistance als after 15 years of exposure to Bt crops: O. nubilalis and Cry1Ab corn was inherited as a recessive trait (Table 1 and Fig. 5). Available data in the United States and H. armigera and Bt cotton in Australia (Table enabled evaluation of this factor for 19 cases, based on direct assessment 1). In both cases, inheritance of resistance was completely recessive on of dominance (h) (12 cases) or indirect assessment derived from survival young plants (h = 0), but not on older plants (h = 0.31 for O. nubilalis of susceptible pests on Bt plants in the field (7 cases) (Supplementary and 0.63 for H. armigera)64–66. These direct estimates of dominance Table 5). Bt plants met the high-dose standard in six of nine (67%) cases with either no decrease in susceptibility or <1% resistant individuals, but not in any of the ten cases with ≥1% resistant individuals (Fisher’s exact test, P = 0.003; Table 1). A compelling contrast confirming the importance of the high-dose criterion is seen between the rapid evolution of resistance to Cry1Ac in Bt cotton by H. zea, but not by the closely related pest H. virescens (Table 1). Cry1Ac cotton met the high-dose standard against H. virescens, but not H. zea (Table 1). These two polyphagous pests that attack cotton were sampled from the same region and tested side-by-side in some studies59,60. Moreover, evolution of resistance to insecticides other than Bt toxins has been faster in H. virescens than H. zea61,62, which refutes High dose Not high dose the alternative hypothesis that resistance generally evolves faster in H. zea than H. virescens. One of the three exceptional cases in which the high-dose standard Figure 5 Resistance to Bt crops and dose criterion. Resistance evolved more slowly when the high-dose criterion was met (left, n = 6 cases) than when was not met and the percentage of resistant individuals was <1% involves it was not met (right, n = 13 cases). Red, >50% resistance and reduced D. v. virgifiera and Bt corn producing Cry34/35Ab (Table 1). In this case, efficacy reported; orange, >50% resistance and reduced efficacy expected; the monitoring data cover only 4 years since commercialization, during yellow, 1–6% resistant individuals; blue, <1% resistant individuals; which adoption of this product has been limited63. green, no decrease in susceptibility (see Table 1 and text for details).
NATURE BIOTECHNOLOGY VOLUME 31 NUMBER 6 JUNE 2013 7 REVIEW
are based on survival of F1 progeny on Bt plants to the adult stage for 5). Thus, available evidence suggests that low refuge abundance and non- H. armigera65,66, but only for 15 days for O. nubilalis, which might recessive inheritance of resistance accelerated evolution of resistance by overestimate h for this pest64. In the experiments with H. armigera and B. fusca to Bt corn. Bt cotton, the concentration of Cry1Ac was 75% lower in the old plants As with B. fusca, low refuge abundance and failure to meet the high- compared with the young plants66. dose criterion apparently accelerated evolution of S. frugiperda resistance In both of these cases, non-Bt crop refuges were abundant. For H. to Cry1F corn in Puerto Rico75,76. Based on mortality at the highest armigera in Australia, the mean percentage of non-Bt cotton was 73% concentration of Cry1F tested, resistance was partially recessive76 from 1996 to 2003 (range, 40–90%) when Cry1Ac cotton was planted, (h = 0.14), which does not meet the high-dose standard of at least 95% and 15% (range, 6–30%) from 2004 to 2011, when two-toxin Bt cotton mortality of heterozygotes31 (h ≤ 0.05). Although the levels of Cry1F in producing Cry1Ac and Cry2Ab replaced Cry1Ac cotton57 (Fig. 2). For Bt corn are “close to high dose” against this pest76, modeling results sug- O. nubilalis and Bt corn in the United States, the minimum percentage gest that such moderate doses can cause faster resistance evolution than of corn planted with non-Bt corn for any state for a given year was 24%, either higher doses that kill all or nearly all heterozygotes, or lower doses which occurred in Iowa in 2010 and 2012 (ref. 67). From 1996 to 2012, that allow substantial survival of susceptibles16 (B.E.T. & Croft, B.A.)25. the mean was 53% non-Bt corn in Iowa67. Overall, the results show that Cross-resistance to Cry1F caused by exposure to Cry1A toxins in sprays rapid evolution of resistance is less likely when the high-dose standard and Bt corn might have also promoted resistance to Cry1F-producing is met, and in some cases when this criterion is not met throughout the corn76. Evolution of resistance in this case was probably also accelerated growing season, resistance can be delayed for more than a decade with by continuous exposure to Bt corn during as many as 10 generations per abundant refuges. year, which translates to 30 generations of selection in 3 years76. A scarcity of refuges in India and China may have promoted faster Low initial resistance allele frequency. The monitoring data show that evolution of P. gossypiella resistance to Cry1Ac cotton in these two coun- rapid resistance evolution was less likely when the initial resistance allele tries compared with the United States (Table 1), where refuges have been frequency was low (Table 1). The initial resistance allele frequency was relatively abundant and high compliance with the refuge strategy was below the detection threshold (no major resistance alleles detected, esti- documented by our team in Arizona (Y.C., B.E.T et al.)77,78. Regulations mated frequency = 0) in 6 of 11 cases (55%) with either no decrease in in India mandate refuges of non-Bt cotton, but apparently compliance susceptibility or <1% resistant individuals, compared with 0 of 5 cases has been low79,80. China has not required non-Bt cotton refuges, and the with >1% resistant individuals (Fisher’s exact test, one-tailed P = 0.058; non-Bt cotton percentage decreased to 8% in 2008 and 6% in 2009 and Table 1 and Supplementary Table 6). 2010 in six provinces of the Yangtze River Valley54. With a criterion of an initial resistance allele frequency <0.001, Another factor accelerating P. gossypiella resistance in India and China however, the association with resistance was not significant. The ini- might be a lower concentration of Cry1Ac in the types of Bt cotton tial resistance allele frequency was <0.001 in 7 of 11 cases (64%) with grown there compared with the varieties grown in the United States. no decrease in susceptibility or <1% resistant individuals versus 2 of In side-by-side field trials conducted in China, the abundance of P. g o s- 5 cases (40%) with ≥1% resistant individuals (Fisher’s exact test, one- sypiella larvae was about five times higher on the predominant Bt cotton tailed P = 0.37). Moreover, in two of the three cases from the United variety grown in China compared with a Bt cotton variety grown on a States with no decrease in susceptibility for more than a decade, the limited basis in the United States54. Although the efficacy and toxin estimated initial resistance allele frequency was not <0.001; instead it concentrations have not been compared directly among the most popu- was 0.0015 for H. virescens in four southern states and 0.16 for P. g os - lar types of Bt cotton grown in these countries, survival of susceptible sypiella in Arizona68–71 (Table 1 and Supplementary Tables 4 and 6). In P. gossypiella in the field was higher in both India and China than the laboratory-selected strains of these pests, resistance to Cry1Ac is reces- United States, and the high-dose standard was met in the United States, sive68,71 (Supplementary Table 5). In addition, refuges were abundant but not in the two Asian countries (Table 1 and Supplementary Table 5). for the first decade in both of these cases. The mean statewide percent- It is unclear why P. gossypiella resistance to Cry1Ac is a much more seri- age of cotton planted to non-Bt cotton from 1996 to 2005 was 42% in ous problem in India than in China (Tables 1 and 2 and Supplementary Arizona9 and 50% in Arkansas, which had one of the highest adoption Table 2). However, unlike the true-breeding varieties of Bt cotton planted rates of Bt cotton of any state where H. virescens was monitored59,69,72. in China, the United States, and elsewhere, hybrids account for nearly These results support the prediction from modeling studies that even all Bt cotton planted in India80. In 2009, >500 Bt cotton hybrids were when the initial resistance allele frequency exceeds 0.001, resistance can approved for planting in India80. Some of these diverse Bt cotton hybrids be delayed substantially, particularly if inheritance of resistance is reces- and the unapproved Bt cotton grown in India81 may have lower toxin sive and refuges are abundant28. For P. gossypiella in Arizona, substantial concentrations than the Bt cotton varieties grown in China. fitness costs and incomplete resistance probably also helped to delay Comparing field outcomes and refuge abundance in Australia, China resistance28,71. and the United States for three congeneric pests (H. armigera, H. punc- tigera and H. zea) provides useful lessons for managing resistance when Refuges. Consistent with previous reviews based on relatively limited the high-dose standard is not met (Table 1 and Fig. 2). Cotton plants data20–22, the more extensive monitoring data reviewed here support producing Cry1Ac or both Cry1Ac and Cry2Ab do not meet the high- the prediction that abundant refuges can delay resistance. Results from dose standard for any of these three pests (Table 1 and Supplementary grower surveys in South Africa imply that the low abundance of refuges of Table 5). After more than a decade of exposure to Bt cotton, the fre- non-Bt corn hastened evolution of B. fusca resistance to Bt corn producing quency of resistant individuals remained <1% for H. armigera and Cry1Ab73,74. On average, from 1998 to 2004, fewer than 30% of the farm- H. punctigera in Australia for Cry1Ac and Cry2Ab, increased to between ers planting Bt corn in the Vaalharts area of South Africa complied with 1% and 5% for H. armigera in China for Cry1Ac, and exceeded 50% for contracts requiring them to plant non-Bt corn refuges73. In addition, pre- some populations of H. zea in the southeastern United States for both commercialization field data showing 2– 3% survival of susceptible larvae Cry1Ac and Cry2Ab (Table 1). on Cry1Ab corn relative to non-Bt corn indicate that this Bt corn does Of the three countries, Australia has applied the most stringent refuge not meet the high-dose standard against B. fusca (Supplementary Table requirements, which may have substantially delayed resistance. For cotton
8 VOLUME 31 NUMBER 6 JUNE 2013 NATURE BIOTECHNOLOGY REVIEW producing only Cry1Ac, the minimum percentage of non-Bt cotton cumulative duration of pest exposure to Bt crops, the number of pest required on each farm in Australia was 70% from 1996 to 2003 (ref. populations exposed and improved monitoring efforts. 57) versus 4% in the United States26,82. For two-toxin cotton, Australia Our review of field-evolved resistance to Bt crops based on monitor- requires 10% non-Bt cotton or the equivalent in terms of other non-Bt ing data for up to two decades from 24 cases in eight countries generally crop host plants on each farm83, whereas the United States has elimi- confirms the principles of resistance management based on evolutionary nated refuge requirements in most regions84. theory. As predicted, factors associated with sustained susceptibility to In China, however, virtually all Bt cotton planted produces only the Bt toxins in transgenic crops are a toxin concentration that meets the Cry1Ac (Fig. 2) and refuges of non-Bt cotton have not been required52. high-dose standard and thus renders inheritance of resistance recessive Nonetheless, non-Bt host plants other than cotton accounted for >92% (see Theory section above), a low initial frequency of resistance alleles, of the cropping area planted to H. armigera host plants from 1997–2006 and abundant refuges of non-Bt host plants near Bt crops that promote (ref. 7), which probably slowed resistance. Although H. zea in the United survival of susceptible insects. States also uses non-Bt host plants other than cotton, one of its major Before commercialization, scientists can evaluate insect responses alternative hosts is Bt corn producing Cry1Ab, which is expected to to Bt crops to determine if the high-dose standard is met and if the select for cross-resistance to Cry1Ac41,43. Taking this and other factors initial frequency of resistance is low, using the techniques described including insecticide sprays into account, the ‘effective refuge’ for H. zea in the studies reviewed here (Supplementary Tables 5 and 6). Because during the three generations in which it feeds on cotton was meticu- resistant strains are often not available before commercialization, the lously estimated as 39% in Arkansas for 2001–2005 (ref. 41). high-dose standard can be assessed proactively by measuring survival of susceptible insects on Bt crops31. In parallel, estimates of the frequency Pyramids. Field outcomes are consistent with the prediction that resis- of individuals with a genetically based decrease in susceptibility relative tance to pyramids will evolve faster if two-toxin plants are grown at the to conspecific individuals can be made proactively with bioassays of same time as plants producing only one of the toxins in the pyramid47. In field-derived strains using Bt plants, Bt plant parts, or diagnostic con- the United States, farmers planted one-toxin cotton producing Cry1Ac centrations of toxin in diet. Although F2 screens have been especially concurrently with two-toxin cotton producing Cry1Ac and Cry2Ab useful for detecting rare recessive resistance alleles, the modeling and from 2004 to 2010, whereas Australian growers completely replaced empirical results reviewed here do not support the idea that it is critical Cry1Ac cotton with two-toxin cotton during 2004 (Fig. 2). As noted to determine if the initial resistance allele frequency is <0.001. above, the frequency of resistance to both toxins has exceeded 50% for The relevant theory and data suggest that if the criteria for high dose some populations of H. zea in the United States, whereas it has remained and low initial frequency are met, resistance can be delayed with limited <1% for H. armigera and H. punctigera in Australia (Table 1). refuges. Conversely, if these criteria are not met, resistance is likely to In principle, faster evolution of resistance in H. zea than in H. armig- evolve rapidly unless refuges are abundant. Therefore, systematic assess- era could also reflect higher initial resistance allele frequencies or more ment of these criteria can be used proactively to enhance resistance dominant inheritance of resistance in H. zea21. The available data suggest management. Moreover, if reporting the assessment of these criteria that initial resistance allele frequencies for Cry1Ac and Cry2Ab were becomes standard practice, the data available for testing predictions will not significantly higher for H. zea than for H. armigera (Supplementary increase steadily, thereby facilitating refinements in resistance manage- Table 6), but resistance to both toxins appears to be more dominant in ment strategies. H. zea (Supplementary Table 5). In the past decade, farmers in the United States, India and Australia have shifted largely from planting first-generation transgenic plants pro- Conclusions ducing one Bt toxin to using ‘pyramids’ that produce two or more dis- From 2005 to 2010, the data available to assess the effectiveness of tinct Bt toxins active against a particular pest (Fig. 2). The limited field resistance management tactics for Bt crops increased dramatically and data available for pyramids confirm predictions from theory and small- the number of major target pests with some populations resistant to Bt scale experiments with a model system indicating that pyramids work crops and reduced efficacy reported surged from one to five (Tables 1 best when implemented proactively47, as has been done in Australia57. and 2 and Figs. 1, 3 and 4). The increase in documented cases of resis- Conversely, when a pyramid of two toxins is adopted after resistance is tance likely reflects increases in the area planted to Bt crops (Fig. 1), the no longer rare to one of the toxins, the benefits of this approach seem to
Table 3 Bt toxin pyramids used proactively and separately from one-toxin plants or remedially and concurrent with one-toxin plants Pest Crop Country Toxins in pyramid21,57,63 Resistance detecteda Proactive and separate from one-toxin plants H. armigera Cotton Australia Cry1Ac, Cry2Ab None H. punctigera Cotton Australia Cry1Ac, Cry2Ab None Remedial and concurrent with one-toxin plants D. v. virgifera Corn USA Cry3Bb, Cry34/35Ab Cry3Bb H. zea Cotton USA Cry1Ac, Cry2Ab Cry1Ac H. zea Cotton USA Cry1Ac, Cry1F Cry1Ac P. gossypiella Cotton India Cry1Ac, Cry2Ab Cry1Ac S. frugiperda Corn USA Cry1F, Cry1A.105b, Cry2Ab Cry1F aResistance detected to one of the toxins in a pyramid before the pyramid completely replaced single-toxin Bt crops producing one of the toxins in the pyramid. Monitoring data and references are provided in Supplementary Tables 3 and 4 for H. armigera and H. punctigera, and in Table 2 for the four other pests. bCry1A.105 is a chimeric Bt toxin with its amino acid sequence 99% identical to Cry1F for domain III, identical to Cry1Ab for domain I, and identical to Cry1Ac for domain II and C terminus95. Although data evaluating S. frugiperda responses to Cry1A.105 have not been reported, cross-resistance to Cry1A.105 is expected in Puerto Rico because populations there have been selected for resistance to each of its three parent toxins: Cry1F in Bt corn, and Cry1Ab and Cry1Ac in sprays75,76. For S. frugi- perda populations and families from Puerto Rico resistant to Cry1F, resistance ratios for Cry1Ab and Cry1Ac ranged from 12 to 89 (refs. 75,103).
NATURE BIOTECHNOLOGY VOLUME 31 NUMBER 6 JUNE 2013 9 REVIEW
be greatly reduced, as exemplified by resistance to Cry2Ab in Bt cotton (2012). for H. zea in the United States (Table 1). In several other cases, pyramids 12. Lu, Y., Wu, K., Jiang, Y., Guo, Y. & Desneux, N. Widespread adoption of Bt cotton and insecticide decrease promotes biocontrol services. Nature 487, 362–365 (2012). are also being used as a remedial tactic following documented resistance 13. Onstad, D. Insect Resistance Management: Biology, Economics, and Prediction to Bt crops producing only one of the toxins in the pyramid (Table 3). (Academic Press, London, 2008). 14. Heckel, D.G. Insecticide resistance after Silent Spring. Science 337, 1612–1614 Although all of the data reviewed here involve crystalline (Cry) Bt (2012). toxins, transgenic crops producing vegetative insecticidal proteins (Vips) 15. Tabashnik, B.E. Evolution of resistance to Bacillus thuringiensis. Annu. Rev. Entomol. from Bt are registered in the United States21 and their use is expected 39, 47–79 (1994). 85 16. Gould, F. Sustainability of transgenic insecticidal cultivars: integrating pest genetics in Australia . In addition, genetically engineered Cry toxins that were and ecology. Annu. Rev. Entomol. 43, 701–726 (1998). more effective than native Bt toxins against resistant strains of several 17. Ferré, J. & Van Rie, J. Biochemistry and genetics of insect resistance to Bacillus pests in laboratory tests might eventually broaden options for manag- thuringiensis. Annu. Rev. Entomol. 47, 501–533 (2002). 18 86 18. Pardo-López, L., Bravo, A. & Soberón, M. Bacillus thuringiensis insecticidal three- ing insects with transgenic plants (B.E.T. et al.) . Scientists are also domain toxins: mode of action, insect resistance and consequences for crop protec- developing transgenic plants that control insects by means of RNA inter- tion. FEMS Microbiol. Rev. 37, 3–22 (2013). 87–90 19. Carrière, Y., Crowder, D.W. & Tabashnik, B.E. Evolutionary ecology of insect adapta- ference and fusion proteins . tion to Bt crops. Evol. Appl. 3, 561–573 (2010). Even with a wider range of approaches used to engineer plants for 20. Tabashnik, B.E., Gassmann, A.J., Crowder, D.W. & Carrière, Y. Insect resistance to protection against insects, resistance management will continue to be Bt crops: evidence versus theory. Nat. Biotechnol. 26, 199–202 (2008). 21. Tabashnik, B.E., Van Rensburg, J.B.J. & Carrière, Y. Field-evolved insect resistance to essential. Based on the 24 cases reviewed here, pests can evolve resistance Bt crops: definition, theory, and data. J. Econ. Entomol. 102, 2011–2025 (2009). to toxins in Bt crops in as few as 2 years under the worst circumstances; 22. Huang, F., Andow, D.A. & Buschman, L. Success of the high-dose/refuge resistance under the best circumstances, however, efficacy can be sustained for management strategy after 15 years of Bt crop use in North America. Entomol. Exp. Appl. 140, 1–16 (2011). 15 years or more. Although regulations in the United States and else- 23. Tabashnik, B.E. Delaying insect resistance to transgenic crops. Proc. Natl. Acad. Sci. where mandate refuges of non-Bt host plants for some Bt crops, farmer USA 105, 19029–19030 (2008). 24. Georghiou, G.P. & Taylor, C.E. Genetic and biological influences in the evolution of compliance is not uniformly high and the required refuge percentages insecticide resistance. J. Econ. Entomol. 70, 319–323 (1977). may not always be large enough to achieve the desired delays in evolu- 25. Tabashnik, B.E. & Croft, B.A. Managing pesticide resistance in crop-arthropod com- tion of resistance3,63,73–76,79–81. Both in theory and practice, using Bt plexes: interactions between biological and operational factors. Environ. Entomol. 11, 1137–1144 (1982). crops in combination with other tactics as part of integrated pest man- 26. US Environmental Protection Agency. The Environmental Protection Agency’s White agement may be especially effective for delaying pest resistance9. We Paper on Bt Plant-pesticide Resistance Management
10 VOLUME 31 NUMBER 6 JUNE 2013 NATURE BIOTECHNOLOGY REVIEW
(2010). 77. Carrière, Y. et al. Long-term evaluation of compliance with refuge requirements for 45. Brévault, T., Nibouche, S., Achaleke, J. & Carrière, Y. Assessing the role of non-cotton Bt cotton. Pest Manag. Sci. 61, 327–330 (2005). refuges in delaying Helicoverpa armigera resistance to Bt cotton in West Africa. Evol. 78. Tabashnik, B.E. et al. Sustained susceptibility of pink bollworm to Bt cotton in the Appl. 5, 53–65 (2012). United States. GM Crops Food 3, 194–200 (2012). 46. Carrière, Y. et al. Large-scale, spatially explicit test of the refuge strategy for delaying 79. Stone, G.D. Biotechnology and the political ecology of information in India. Hum. insecticide resistance. Proc. Natl. Acad. Sci. USA 109, 775–780 (2012). Organ. 63, 127–140 (2004). 47. Zhao, J.-Z. et al. Concurrent use of transgenic plants expressing a single and two 80. Choudhary, B. & Gaur, K. Bt Cotton in India: A Country Profile. ISAAA Series of Bacillus thuringiensis genes speeds insect adaptation to pyramided plants. Proc. Biotech Crop Profiles (ISAAA, Ithaca, NY, 2010). Natl. Acad. Sci. USA 102, 8426–8430 (2005). 81. Herring, R.J. Stealth seeds: Bioproperty, biosafety, biopolitics. J. Dev. Stud. 43, 48. Brévault, T. et al. Potential shortfall of pyramided transgenic cotton for insect resis- 130–157 (2007). tance management. Proc. Natl. Acad. Sci. USA 110, 5806--5811 (2013). 82. US Environmental Protection Agency. Biopesticides registration action document— 49. National Research Council. Pesticide Resistance: Strategies and Tactics for Bacillus thuringiensis plant-incorporated protectants)
NATURE BIOTECHNOLOGY VOLUME 31 NUMBER 6 JUNE 2013 11 REVIEW
(2011). in Proceedings of the 2004 Beltwide Cotton Conferences, San Antonio, TX, 109. Monsanto Biotechnology Trait Acreage. Fiscal Years 1996 to 2009.
12 VOLUME 31 NUMBER 6 JUNE 2013 NATURE BIOTECHNOLOGY CFI declaration: B.E.T. is coauthor of a patent on engineering modified Bt toxins to counter pest resistance, which is related to published research (Nat. Biotechnol. 29, 1128–1131, 2011). Dow AgroSciences, Monsanto and Bayer CropScience did not provide funding to support this work, but may be affected financially by publication of this paper and have funded other work by B.E.T.
NATURE BIOTECHNOLOGY Crop Protection 66 (2014) 53e60
Contents lists available at ScienceDirect
Crop Protection
journal homepage: www.elsevier.com/locate/cropro
DIVECOSYS: Bringing together researchers to design ecologically- based pest management for small-scale farming systems in West Africa
Thierry Brevault a, b, Alain Renou a, Jean-François Vayssieres c, d, Guillaume Amadji e, Françoise Assogba-Komlan f, Mariama Dalanda Diallo g, Hubert De Bon c, Karamoko Diarra h, Abdoulaye Hamadoun i,Joel€ Huat c, k, Pascal Marnotte a, Philippe Menozzi a, k, Patrick Prudent a, Jean-Yves Rey c, j, Dieynaba Sall j, Pierre Silvie a, * Serge Simon c, f, Antonio Sinzogan e, Valerie Soti a, l, Manuele Tamo d, Pascal Clouvel a, a CIRAD, PERSYST, UPR AIDA, Montpellier, France b BIOPASS, ISRA-IRD-UCAD, Campus de Bel-Air, Dakar, Senegal c CIRAD, PERSYST, UPR HORTSYS, Montpellier, France d IITA, Benin Station, Cotonou, Benin e UAC, FSA, Cotonou, Benin f INRAB, PCM, Cotonou, Benin g UGB, UFR S2ATA, Saint-Louis, Senegal h UCAD, FST, Dakar, Senegal i IER, CRRA de Sotuba, Bamako, Mali j ISRA, CDH, Dakar, Senegal k AFRICARICE, Cotonou, Benin l CSE, Dakar, Senegal article info abstract
Article history: Crop pests are a major constraint to the intensification of agricultural production in the tropics, with Received 23 June 2014 novel issues related to global change (climate, land use, biological invasions, etc.), food security and Received in revised form preservation of natural resources and biodiversity. A research, extension and education network called 22 August 2014 DIVECOSYS (Diversity of cropping systems and ecologically-based pest management in West Africa) was Accepted 25 August 2014 launched in 2010 to synergize applied research actions in response to growing concerns on the Available online vulnerability of agricultural systems to pest management in West Africa. This scientific network brings together research and academic institutions, with expertise spanning a multidisciplinary perspective Keywords: Research network from biology and ecology to remote sensing, agronomy and integrated pest management. Its main sci- fi Biodiversity enti c objective is to explore the potential of biodiversity and ecological processes such as pest regu- Biological control lation, enabling novel ecologically-based models for productive systems, reduction of pesticide use, and Ecosystem services adaptation or resilience of farming systems in the face of environmental disruptions. From Northern Landscape Senegal to Southern Benin, the research group explores a wide range of ecoregions and socio-ecological Africa contexts, including stakeholders and their objectives, land use and agricultural practices, and manage- ment of biodiversity for enhancing biological control. Main challenges to be turned into opportunities include (i) encouraging collaborations amongst researchers from different scientific fields, (ii) fostering interactive research and synergies among research institutions and among countries, and (iii) developing an ecological engineering approach for the design of sustainable agricultural systems for smallholder farmers. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction
Between 1982 and 2007, the overall production of cereals and grain legumes was multiplied by 3 and 2.4 times respectively in * Corresponding author. Tel.: þ33 4 67 61 55 00. E-mail address: [email protected] (P. Clouvel). sub-Saharan Africa, where the population more than doubled over http://dx.doi.org/10.1016/j.cropro.2014.08.017 0261-2194/© 2014 Elsevier Ltd. All rights reserved. 54 T. Brevault et al. / Crop Protection 66 (2014) 53e60 the same period (Uhder et al., 2011). Such a rise in production was challenges for enhanced pest control. During this process, knowl- mostly achieved by extending cultivated areas, without any or only edge gaps are identified and addressed for better targeting limited increases in yields. In order to meet the demand, cultivated ecologically-based pest management for small-scale agricultural areas would need to increase 2e3-fold by 2050. Such an expansion systems in West Africa. leading to large-scale farming systems is barely feasible in most African countries due to the limited availability of arable land and 2. Dealing with emerging concerns for pest management water resources. In addition, such an extension would have a detrimental impact on natural areas where biodiversity must be In West Africa, farmer field schools (Settle and Garba, 2011), preserved. On the contrary, smallholder agriculture must be sup- threshold-based interventions (Renou et al., 2012; Silvie et al., ported as the backbone of global food security in most African 2013), insecticidal transgenic crops such as Bt cotton (Hema countries (Tscharntke et al., 2012). Hence, there is an impending et al., 2009), or biological methods, e.g. spraying bio-insecticides challenge of achieving efficient and productive agricultural land or releasing beneficial insects (Neuenschwander, 2001), have the use, while preserving natural resources. potential to help reducing the use of broad spectrum insecticides. Crop pests are a major constraint to the intensification of agri- However, concurrent emergence of new pest problems has been cultural production, especially in the tropical areas (Oerke, 2006; observed. Evolution of multiple resistance to pesticides (including FAO, 2012; Savary et al., 2012). They include all organisms harm- Bt-formulations) has been demonstrated in some populations of ful to crops: arthropods (insects and mites), fungi, bacteria, viruses, the whitefly, Bemisia tabaci (Gennadius) (Gnankine et al., 2013), along with nematodes, rodents, birds and weeds. Losses caused by and the cabbage diamondback moth, Plutella xylostella (L.) crop pests along the value chain of agricultural production remain (Grzywacz et al., 2010). Emerging pests have been observed as part abnormally high and threats are set to increase with climate change of the release of ecological niches, e.g. sucking pests replacing (Maxmen, 2013). The massive use of synthetic pesticides has shown bollworms on Bt cotton (Deguine et al., 2008). Expansion of the major limitations, including serious hazards for the environment host range has been reported in the cotton bollworm, Helicoverpa and human health, and sometimes evolution of resistance in target armigera (Hübner) (T. Brevault, unpublished data). Increased rate of organisms. In addition, detrimental effects on biodiversity can biological invasions has also been recorded (Youm et al., 2011), result in pest outbreaks due to the alteration of ecosystem services with recent cases such as the fruit fly, Bactrocera invadens Drew- such as natural pest regulation (Power, 2010; Meehan et al., 2011). Tsuruta-White (Vayssieres et al., 2009a; Goergen et al., 2011), or In West Africa, where family agriculture dominates, pesticide use the tomato leafminer, Tuta absoluta Meyrick (Brevault et al., 2014). remains limited to and reserved for certain high value cash crops Weeds form a particular group of pests, whose control takes up a (particularly cotton eGossypium hirsutum L., cocoa eTheobroma major share of producers' farming calendars. Herbicide use is cacao L., fruits and horticultural crops). Rainfed staple crops are increasing in West Africa, notably non-selective herbicides such as mainly grown without pesticides, with the exception of rice and glyphosate, whose generalized use could lead to the evolution of cowpea. However, the overall simplification of landscapes resulting weed resistance and could have negative consequences for the from the reduction of crop diversity and the expansion of mono- environment (Busi et al., 2013). Regarding the dynamics of insect culture, and in some cases from urban expansion, has contributed populations, weeds may also serve as a reservoir or a refuge for to the loss of biodiversity. As a result, cultivated ecosystems could pest insects (e.g. Cleome spp. for H. armigera or local Euphorbiaceae become more susceptible to pest outbreaks and biological in- for B. tabaci), or a source of food for natural enemies (Ratnadass vasions (Tscharntke et al., 2012), particularly in a context of global et al., 2012). warming and climate change (Pettorelli, 2012). The adaptation of crop protection to risks arising from global DIVECOSYS, a French acronym for ‘Diversite des systemes de environmental change is a major challenge for agricultural production et gestion agro-ecologique des bio-agresseurs en Afri- research. In this perspective, DIVECOSYS has based its approach in que de l'Ouest’ (Diversity of cropping systems and ecologically- the general framework of agro-ecology, defined as the application based pest management in West Africa), was launched in 2010 as of ecological concepts and principles to the design of sustainable a research, extension and education network, in response to con- agricultural systems (Altieri, 1987; Gliessman, 1990). Among the cerns related to the vulnerability of agricultural systems to pest various concepts and approaches underlying agro-ecology outbreaks in West Africa. This partner network brought together (Mendez et al., 2013), our research posture is close to that of more than fifty researchers, teachers, extensionists and research Tscharntke et al. (2005), who postulate that agriculture can fellows from eleven research and academic institutions based in contribute to conservation of high-diversity systems, which may in Benin, Mali, and Senegal (see list of authors). Its main scientific return provide ecosystem services such as pollination and bio- objective is to explore the potential of biodiversity and ecological logical control. Improving our knowledge on the relative impor- processes involved in pest regulation, enabling novel ecologically- tance of local (field scale) and landscape management of based models for productive systems (produce more sustainably), biodiversity is thus critical to better manage or enhance ecosystem reduction of pesticide use (produce more safely), and adaptation or services. resilience of farming systems in the face of global change and DIVECOSYS also considers side-effects of control strategies environmental disruptions (produce more durably). Main chal- across agricultural landscapes (Brevault and Bouyer, 2014). Even lenges to be turned into opportunities include (i) encouraging the ecologically sound pest management using biological control exchange of ideas and information amongst researchers from agents have the potential to disrupt ecosystem communities. Non- different scientific fields spanning from biology and ecology to intentional effects of exotic natural enemy releases on native spe- remote sensing and modelling, (ii) fostering interactive research cies need to be assessed before implementing biological control and synergies among research institutions and among countries, programs. Environmental risks include global or local extinction of and (iii) developing ecological engineering for the design of sus- a native species (target or non-target), large reductions in their tainable agricultural systems by combining agronomy, theoretical distribution or abundance, competition with other natural enemies, and applied ecology. and transmission of pathogens to native organisms (Van Lenteren We present here the paradigm and conceptual framework un- et al., 2006). However, because of our commitment to smallholder derlying research issues explored by the research group DIVE- farming systems, our ambition is to deliver not only technical rec- COSYS, according to the diversity of socio-ecological contexts and ommendations, but also to integrate in a near future local T. Brevault et al. / Crop Protection 66 (2014) 53e60 55 knowledge and social sciences for participatory design of pest main staple food crops in the south with scattered orchards (mango management strategies. and mainly citrus) and horticultural crops. Variability in rural population density (from 10 up to 300 inhab. kmÀ2) needs to be seen in line with local natural re- 3. Exploring the diversity of socio-ecological contexts sources. With low population densities, arid territories face con- straints linked to the availability of water resources for irrigation, to From Northern Senegal to Southern Benin, the research group low soil fertility and to land availability, sometimes associated with DIVECOSYS has been exploring a wide range of ecoregions, mainly salinization of cultivated soils. With a population density equal to À determined by rainfall patterns, from 250 mm in the North, up to 80e250 inhab. km 2, the Senegalese groundnut basin is more at 1,200e1400 mm in the southern areas (Fig. 1, Table 1). A wide range risk from a food security viewpoint than in the more populated of socio-ecological contexts is also addressed, including stake- zones in Benin. holders and their objectives, land use and agricultural practices DIVECOSYS is also interested in the cultivated diversity implied (pesticide use), and management of biodiversity (particularly pests in food security for populations in rural areas and urban centres and their natural enemies) for enhanced biological control. with which they interfere (peri-urban agriculture). Current Aridity is the main abiotic constraint in the Northern areas, with research focus on cereal food crops including millet -Pennisetum a short rainy season from JuneeJuly to October (4e5 months). glaucum (L.), sorghum eSorghum vulgare (L.), maize and rice -Oryza Intensification of agriculture and farming systems (cereals, grain sativa (L.), grain legumes such as groundnut eArachis hypogaea (L.), legumes, or cotton) has contributed to the simplification of land- or cowpea eVigna unguiculata (L.), fruit trees (mango, citrus, scapes, both in terms of land use and biodiversity, in the groundnut pineapple -Ananas comosus (L.) Merr., and cashew orchards), and basin in Senegal, on the Mandingue plateau in Mali, and to a lesser market gardening (tomato eSolanum lycopersicum L., cabbage extent in the cotton-producing area in Northern Benin. In some -Brassica oleracea L., onion -Allium cepa L., okra -Abelmoschus areas where access to water is possible during the dry season, e.g. in esculentus (L.), traditional vegetables, etc.). Cash crops (cotton, the Niayes or the Senegal River valley in Senegal or in the Niger cashew, mango, citrus, horticulture, etc.) are also investigated, as a Valley (North of Benin), patches of irrigated crops (vegetables, ce- major contribution to the income of rural populations, but also as reals, legumes, etc.), orchards (mango eMangifera indica L., citrus) parts of landscape mosaics and ecological processes. Insects are the and perennial ecosystems (forests, galleries) in lowlands and along most studied pests in DIVECOSYS, probably due to their economic irrigation channels, form potential habitats where some biodiver- importance among crop pests, but also partly due the predomi- sity can persist in space and time. Benin is dominated by shrub nance of entomologists within the research group. Life history traits cover (closed open, deciduous) followed by broad leaves tree cover of the wide range of pests encountered in these tropical areas, offer and mosaics of crops and savannahs. Cotton is the major cash crop a complexity gradient in terms of management, depending on their grown in the country followed by cashew nuts (Anacardium occi- mobility (local movement versus migration or diapause), their dentale L.). In central-Northern Benin, orchards (mainly mango) are trophic specialization (monophagy versus polyphagy), the extent of integrated in different types of tree savannahs alongside cashew their spatio-temporal distribution according to crop season. orchards most of the time. In the south, most of the original Farming systems in these ecoregions generally consist of a mosaic vegetation has been replaced by secondary grasslands and thickets. of cultivated and semi-natural habitats that vary over time. Maize (Zea mays L.) and cassava (Manihot esculenta Crantz) are the Remnant forests, hedges, field margins, dry residue from previous annual crops, grazing land, trees, shrubs and fallow are potential habitats from which insect pests (or natural enemies of pests) can colonize crops.
4. Designing pest management strategies from a three-pillar approach
4.1. From the integration of conventional and alternative tactics
The DIVECOSYS approach considers farming and ecological di- mensions for development of pest control strategies (Table 1). This includes chemical control (calendar or threshold-based spraying of conventional pesticides), host resistance (genotype), cultural con- trol (crop management such as sowing date, crop rotation, inter- cropping, sanitation, trap crops, etc.), biological control (antagonists, parasitoids, predators, pathogens, etc.), physical and mechanical control (screens, nets, traps, temperature, tillage, etc.) and behavioural control (sexual disruption, pheromone trapping, etc.). These tactics alone are no longer relevant for an effective and sustainable control system. The challenge of DIVECOSYS is to consider alternatives to chemical control as potentially efficient by themselves or complementary, and to combine them appropriately (synergies) to significantly reduce the use of pesticides. The approach we propose here rely on the action (i) at different stages of the pest life cycle or at different stages of targeted crops (from Fig. 1. Main ecoregions explored by the research group DIVECOSYS. An ecoregion is initial inoculum to population outbreaks), and (ii) at different defined here as a “recurring pattern of ecosystems associated with characteristic organizational levels of the landscape (field, mosaic composition, combinations of soil and landform that characterise that region” (Omernik, 2004). (1) fi Niayes, (2) Senegal river delta, (3) groundnut basin in Senegal, (4) North, (5) Centre, (6) spatial con guration of mosaic units). This landscape approach South and coast of Benin, (7) Mandingue plateau, and (8) Koutiala plateau in Mali. offers new opportunities to activating ecological processes for the 56 T. Brevault et al. / Crop Protection 66 (2014) 53e60
Table 1 Farming systems and ecological contexts explored.
Country Ecoregion Biome (residual cover %) Landscape Farming system Biological model Life traits
Senegal 1-Niayes Bush savannah (20e90) JuneeNov.: mosaic of crops Agropastoralism, market Cabbage moth Polyphagy Dec.eJune: patches of irrigated gardening orchards Fruit flies Oligophagy crops Tomato leafminer Resistance Invasion Migration 2-Senegal river delta Herbaceous savannah Patches of irrigated areas Agropastoralism, rice, Stem borers, Diapause (20e80) flood-recession sorghum, Sorghum bugs Diversity market gardening Tomato leafminer Invasion 3-Groundnut basin Tree savannah (0e20) Mosaic of crops under tree Agropastoralism, dry Groundnut weevil Host races park cover cereals, legumes Millet head miner Egg-laying Monophagy Diapause Behaviour Benin 4-North Tree savannah (30e40) Mosaic of crops under tree park, Cotton, cereals and Bollworms Polyphagy orchards. Savannahs and legumes Fruit flies Mobility dry forests Mangoes and cashew Scale insects Competition nuts Bugs, ants Transmission of Fruit borers plant pathogens Diversity 5-Centre Tree savannah (20e40) Patches of humid zones, orchards. Tubers, cereals, legumes, Aphids, thrips Oligophagy Savannah and forests lowland rice and market Cowpea bugs Polyphagy gardening Stem borers Mobility Mangoes, pineapple Weeds Competition and cashew nuts Fruit flies Transmission of plant pathogens Diversity 6-South and coast Tree savannah (10) and Savannah and gallery forest. Citrus and mangoes Fruit flies Polyphagy coastal forest (50) Patches of irrigated areas, Market gardening, Leaf rollers, ants Mobility orchards Maize, cassava Seed weevil Diversity Whiteflies Fecundity Nematodes Transmission of plant pathogens Mali 7-Mandingue plateau Herbaceous savannah Mosaic of crops Cereals, legumes Orchards Stem borers Oligophagy Monophagy (0e20) Market gardening Fruit flies Polyphagy Okra virus disease 8-Koutiala plateau Tree savannah (0e20) Mosaic of crops under tree park Cotton, cereals, legumes Bollworms Polyphagy Rice, market gardening Stem borers Diversity Orchards Fruit flies Weeds
This table summarizes the contributions of twenty researchers. control of mobile pests in space and time. In addition, landscape variability of pest populations and may limit or delay crop coloni- ecology provides a hierarchical and integrative ecological basis for zation (Mazzi and Dorn, 2012). They also determine abundance, dealing with issues of biodiversity and ecosystem functioning at diversity and effectiveness of natural enemy communities through multiple scales (Wu, 2006; Birch et al., 2011). the nature of resources (refuges, food source, alternative hosts, etc.)
4.2. To the activation of ecological processes
Activating the ecological processes involved in pest regulation is a major challenge for designing agro-ecological pest management for farming systems. Ecological regulation of insect pest pop- ulations may be achieved i) at the lower trophic level (bottom-up) by host plants through their genotype (tolerance) and distribution in the agricultural landscape or ii) at the higher trophic level (top- down) by predators and parasitoids (Gurr et al., 2003). Agricultural practices can affect the abundance of a pest population by acting on the environment (Fig. 2), with a potential of suppression within the targeted area (e.g. field, orchard). Agricultural practices can also affect the abundance of a pest population at the landscape scale considered as a mosaic of habitats, by modifying the quantity and quality of resources for pests or by offering suitable habitats for different species of natural enemies for biological control (Carriere et al., 2012; Woltz et al., 2012; Monteiro et al., 2013). Composition Fig. 2. Interactions between agricultural practices, landscape context and environ- fi and configuration of landscapes determine the potential of re- ment. To contribute to ecological intensi cation of agricultural systems in the explored systems, we propose to act (i) upon environment to reduce negative externalities sources, but also their accessibility, acting either as a biological associated to crop management, such as pesticide applications, and (ii) upon landscape corridor or, on the contrary, acting as physical and chemical bar- to activate ecological processes that underlie pest regulation. The biotic compartment riers. These landscape properties determine the spatio-temporal includes crop pests. T. Brevault et al. / Crop Protection 66 (2014) 53e60 57 that cultivated habitats, natural and semi-natural areas can provide Considering the diversity of socio-ecological systems explored (Tscharntke et al., 2005; Rusch et al., 2013; Veres et al., 2013). by our research network, we propose to distinguish three major Generally, ecological pest regulation increases with landscape territory scales depending upon provision of incentives through complexity (Bianchi et al., 2006; Chaplin-Kramer et al., 2011). markets where markets function, and development of market in- The abundance of a pest population results from dynamic in- stitutions where they do not exist: (i) worldwide market oppor- teractions with the biotic and abiotic environment, which occur tunities (fruit production, cotton), (ii) local market opportunities, over time and space steps that usually exceed the crop growth cycle especially for urban areas (vegetable and fruit production), and (iii) (multiannual variability) or the cultivated plot alone (landscape). regional market opportunities and on-farm consumption (cereals, According to Burel (1996) and Burel and Baudry (1999), the land- grain legumes). For stakeholders involved in (i), the emergence of scape, in constant interaction with the ecological processes that considerations relating to the environment, food safety (e.g. arise from it, is the result of the environmental and social dynamics contamination by pesticides), and risks of crop losses due to dis- that have developed there (Fig. 3). Given that, we postulate that the eases and insect pests, are incentives by the market or by the landscape is the appropriate level of analysis for a holistic approach payment of environmental services associated with the reduction to action, i.e. individual and collective pest control programs, their of pesticide use (Power, 2010). For (ii), the emergence of commer- efficiency in relation to the target biological system and their ex- cial niches and general requirements for product safety is one way ternalities, with a particular view to sustainable production to encourage the market to evolve towards food systems. For (iii), in management. the absence of incitation by the market, strategies should provide visible and immediate benefits to be adopted.
4.3. Taking into account local socio-ecological systems 5. Basic research issues Our ambition to contribute to the implementation of innova- tive strategies for sustainable pest management needs increased We distinguish here between applied research specific to the interactions with farmers to collectively identify technical action- particular socio-ecological contexts explored by the research group levers. Recent experiences highlight the complexity to adoption of DIVECOSYS in West Africa, and more theoretical or basic research agro-ecological methods by smallholders in developing countries needs which are knowledge gaps. The activation of ecological (Parsa et al., 2014). As reported by Giller et al. (2009), farmers in processes involved in pest regulation requires a thorough knowl- Sub-Saharan Africa often attribute a substantially higher value to edge of the life system of targeted pests (reproduction rate, local fi immediate costs and bene ts than those incurred or realized in movements and migration, competition, gene flow and meta- the future due to the constraints of production and food security population structure, etc.), of food webs including plants (pri- that they face. Yet, while farmers seek substantial, visible and mary producers), pests (primary consumers) and communities of fi immediate bene ts when considering adoption of conservation predators, parasitoids (secondary and tertiary consumers), and fi agriculture (CA), many of the bene ts of employing CA are only hyperparasitoids. It also assumes an understanding of the effect of realized in the longer term (FAO, 2008). Institutional elements agricultural practices on environment and on landscape multi- fi required for all successful strategies for agricultural intensi cation functional properties (connectivity, permeability, etc.) that govern include a stable macroeconomic environment, provision of in- short distance movements of targeted pests and access to re- centives through markets in areas where markets do function, sources, together with interacting communities of the considered development of market institutions where they do not, and public ecosystem (Box 1). and private investment in an appropriate mix of physical, human, Crossing representations of biological systems linked to natural, and social capital (Ehui and Pender, 2005). An integrated agronomy and ecology gives rise to more theoretical and basic approach to pest control is more knowledge intensive, requiring research issues, in order to produce knowledge, methods and tools monitoring and decision systems, and currently incurs higher (especially models) for individual and collective action. A few key operational costs than does the sole use of insecticides (Cook et al., research issues are listed below: 2007). A similar diagnostic could be made regarding imple- mentation of innovative strategies for integrated pest manage- What is the spatial structure and area of the targeted pest ment (IPM): namely the market as a major driver. In East Africa, populations? On what scale do sub-populations interact e some smallholder farmers use push pull strategies to protect genetically and demographically? their maize and sorghum crops (Khan and Pickett, 2004). Their What are the effects of crop management and landscape context fi success relies on ecological but also speci c social conditions, on the spatio-temporal abundance of targeted pests and the fi including reduced access to costly pesticides, increased bene ts effectiveness of biological control? with the introduction of multifunctional plants (grain legumes, What experimental designs are to be implemented to study and fi livestock feed, etc.) in cropping systems, signi cant increase model ecological processes, in relation to landscape attributes of crop yield, and well-thought information campaigns. and life system of targeted pests? What are the relevant spatial organizations for individual and collective action, with a view to sustainable production man- agement in a given area? In the context of global change, what are the scenarios of landscape evolution for the socio-ecological systems consid- ered? What are their foreseeable impacts on ecological pro- cesses, and how can habitats (cultivated or non-cultivated) be managed to preserve functional biodiversity and pest regulation services? As a result, what models might be developed to explore the abovementioned scenarios, notably with a view to Fig. 3. Action and assessment framework based on concepts derived from landscape ecology (Burel and Baudry, 1999). Landscape configuration results from the dynamics an ex ante assessment of emerging production systems as of environment and society which has developed there (black arrows). innovations? 58 T. Brevault et al. / Crop Protection 66 (2014) 53e60
Box 1 system (GIS) are used as modern tools for understanding the spatial Landscape and habitat connectivity. distribution of insects and unravelling ecological processes involved in pest regulation.
Connectivity is not a general property of a landscape; it has 6.2. Cotton bollworms to be defined in reference to the movements of the species to be managed (Burel and Baudry, 2005). It is therefore In Mali, seed cotton yield can be reduced by 30e40% due to necessary to take into account the movement methods and insect pests, especially bollworms (Lepidoptera, Noctuidae) such as ecological requirements of the species in question. A con- H. armigera, Diparopsis watersi Rothschild and Earias spp. Area- nected landscape for one species may have numerous wide implementation of threshold-based spraying programs (as barriers for another species. The landscape is dynamic, opposed to conventional calendar-based spraying) resulted to the varies from season to season; the time dimension has to be reduction of insecticide use by 27% from 1997 to 2008 in Mali taken into account. For example, fallow and rangelands (Renou et al., 2012; Silvie et al., 2013). To go further, research is offer resources and movement possibilities for anthophi- needed to define more simple and efficient rules for decision, from lous species; their mowing or grazing date affect their use, local to regional scales. Among alternatives to chemical sprays, hence their role in connectivity. topping cotton plants after the onset of flowering could be a way to Connectivity does not have an absolute ecological quality reduce bollworm infestations (Renou et al., 2011). Field and labo- value; it may be conducive to the movement of undesirable, ratory studies (plant volatile compounds) are currently underway parasitic, invasive or predatory species. To assess or restore to better understand mechanisms involved in the reduction of connectivity, it is therefore necessary to understand: (i) how bollworm infestations (PAFICOT, “Projet d'appui ala filiere Coton- a species moves within a landscape, (ii) how the manager Textile”). can facilitate or limit those movements. Of course, it is not a In Benin, a two-year study highlights the importance of matter of reproducing this process for all species, firstly considering both agricultural practices and landscape context to because species can be grouped by type of behaviour in identify ways to improve the management of bollworm pop- landscapes, secondly because management is often geared ulations (FSP 2006e43). Delayed sowing and frequent weeding towards a few target species (and the species on which they reduced the abundance of H. armigera in cotton fields. Avoiding depend). high concentration of cotton crops in the surrounding landscape (500 m) or tomato crops, as previous landcover, would also reduce H. armigera abundance (Tsafack et al., 2013). Modifying cultural practices or managing landscape at a collective scale to reduce How can landscapes be managed collectively? What governance bollworm infestation is a real challenge. is required to act upon complex socio-technical systems (e.g. horticultural production based on a plurality of stakeholders 6.3. Cabbage diamondback moth and markets, local versus export)? Strengthening ecosystem services such as regulation of insect pests by natural enemies is a potential way to reduce the depen- 6. Applied research issues dence of small farmers on pesticides, to limit field-evolved resis- tance to insecticides, and to design new models of ecologically- Research activities are undertaken with local stakeholders intensive agriculture. The BIOBIO research project (Biodiversity (farmer groups in Senegal and Benin, regional or local centres for and pest management in agricultural landscapes, AIRD) aims at agricultural development in Benin, cotton companies in Mali, etc.) evaluating the effect of crop management and landscape compo- who are actively involved in formulating research issues and sition and configuration on the abundance of the diamond back bringing out innovations. In the following, some examples of moth, P. xylostella, and the efficiency of natural regulation of its research and extension projects currently developed in West Africa populations by predators and parasitoids (Sow et al., 2013a, 2013b). are discussed in the light of ecological engineering for pest Sampling is carried out in a network of farmers' cabbage fields in management. the Niayes, the main vegetable-producing area in Senegal, across two years. Results should clarify the role of insecticide use and 6.1. Millet head miner landscape context on the abundance of larval populations on cab- bage and on the impact of biological control. From an applied In Senegal, the head miner, Heliocheilus albipunctella De Joannis perspective, results could lead to the implementation of novel (Lepidoptera, Noctuidae), is the major insect pest of millet. recommendations to stakeholders for integrated management of Strengthening ecological regulation of this pest opens an avenue to the cabbage diamondback moth both at the field and landscape explore to improve yield and quality of grain cereals, in collabora- scales. Research activities are implemented with local partners tion with small-scale farmers. At the level of the agricultural (farmer' association in Senegal) to facilitate exchange of ideas with landscape, agro-forestry systems offer greater arthropod diversity producers and dissemination of information. and more ecological niches in time and space than a mosaic of annual crops. They can act as a source or relay for insect pest 6.4. Mango fruit flies populations. They can also increase the effectiveness of biological control by providing populations of natural enemies with addi- Fruit flies (Diptera, Tephritidae) have become an increasingly tional resources, especially during the dry season. A research prevalent pest of fruit productions in Benin, but also in sub- project (WAAPP, West Africa Agricultural Productivity Program, Saharan countries, affecting not only mango production but also waapp.coraf.org/) assesses the importance of biodiversity as a fac- other important fruit value chains like citrus (Vayssieres et al., tor in ecological control of insect pests and as a factor of ecosystem 2010). A new invasive species, Bactrocera invadens, which prob- resilience in the face of environmental disturbances. Remote ably originated from Sri Lanka, has colonized East, Central and sensing, satellite image processing, and geographic information West Africa. This polyphagous species can infest more than 40 T. Brevault et al. / Crop Protection 66 (2014) 53e60 59 fruit species in BenineCameroon (Goergen et al., 2011) and more ecological regulation of crop pests in different socio-agro- than 30 fruit species in Senegal (Ndiaye et al., 2012). An IPM- ecosystems. One particular challenge is to measure the impor- package has been proposed to growers in the framework of the tance of biodiversity as a provider of ecosystem services such as West African Fruit Fly Initiative (WAFFI). Biological control with pest control, and as a factor of ecosystem resilience in the face of weaver ants, Oecophylla longinoda (Latreille), and parasitoids environmental change. The research group also contributes to the (Hymenoptera, Braconidae) is a way to control fruit flies on mango dissemination of advanced techniques and methods that enable orchards (Van Mele et al., 2007) and also citrus (J.-F. Vayssieres, research capacity building among research teams, including personal communication). Research efforts were conducted to- geographic information systems, spatial analysis, landscape ecol- wards a better knowledge of weaver ants, especially their preda- ogy, population genetics, ecological modelling, and scientific tory behaviour but also their repulsive effect on pests (Adandonon writing skills. By aggregating a high diversity of researchers, dis- et al., 2009; Van Mele et al., 2009). Other IPM tools for monitoring ciplines, and contexts, DIVECOSYS is a suitable platform to syner- of mango fruit flies and citrus fruit flies in pilot-orchards, sanita- gizing proof of concept ideas into advances for agricultural research tion activities in mango-citrus orchards, threshold-based bio- in Africa. insecticide bait sprays and conservation of parasitoids (Vayssieres From an applied perspective, research conducted within DIVE- et al., 2009b, 2011, 2012; Diatta et al., 2013), are required (Fig. 4). COSYS aims to develop ecologically-based pest management stra- For IPM to be more effective, an area-wide strategy should be tegies for small-scale farming systems in West Africa. Future deployed. DIVECOSYS should bring knowledge, know-how and research projects will include insecteplant interactions, insect and tools (GIS) to help researchers implementing a spatial perspective weed community ecology and biological control, pesticide impacts to their research. on biodiversity, landscape ecology and area-wide integrated pest management. As managers of their territories, stakeholders with 6.5. Stored product pests their own knowledge and innovation capacity are more and more strongly involved for designing and assessing novel and sustainable With the global increase in cereal, grain legume and tuber agricultural systems. production expected to cover the food requirements of an ever growing population, post-harvest losses are an important field of Acknowledgements study. Pests in stored foodstuffs, such as the beetles Prostephanus truncates (Horn), Sitophilus zeamais Motschulsky and S. oryzae (L.), We thank all the scientific partners for providing information weevils, and the moth Sitotroga cereallela (Olivier), cause certain and comments for advice at early stages of this manuscript. Prep- quantitative losses. In addition, they can be a source of fungal aration of this paper was supported by CIRAD AI N6 “Appui aux contamination in grain stocks, right from the field, involving fungi dispositifs de recherche et d'enseignement en partenariat”. such as Aspergillus flavus or Fusarium verticillioides, responsible for the production of mycotoxins of the aflatoxin and fumosinin groups References respectively, which are dangerous for human health. In addition to direct infestations of insects in the field, there are tuber exchanges Adandonon, A., Vayssieres, J.-F., Sinzogan, A., Van Mele, P., 2009. Density of pher- omone sources of the weaver ant Oecophylla longinoda affects oviposition between producers, which may contaminate storage barns. No behaviour and damage by mango fruit flies (Diptera: Tephritidae). Int. J. Pest research has been yet conducted on this issue within the DIVE- Manag. 55, 285e292. COSYS group, but the importance of post-harvest losses underlines Altieri, M.A., 1987. Agroecology: the Scientific Basis of Alternative Agriculture. fi Westview Press, Boulder, CO. the need to integrate this topic in the scienti c priorities. Bianchi, F., Booij, C.J.H., Tscharntke, T., 2006. Sustainable pest regulation in agri- cultural landscapes: a review on landscape composition, biodiversity and nat- 7. Conclusions ural pest control. Proc. R. Soc. B. Biol. Sci. 213, 1715e1727. Birch, A.N., Begg, G.S., Squire, G.R., 2011. How agro-ecological research helps to address food security issues under new IPM and pesticide reduction policies for The main objective of the research group DIVECOSYS is the global crop production systems. J. Exp. Bot. 62, 3251e3261. assessment and the understanding of processes that shape Brevault, T., Sylla, S., Diatte, M., Bernadas, G., Diarra, K., 2014. 2014. Tuta absoluta Meyrick (Lepidoptera: Gelechiidae): a new threat to tomato production in sub- Saharan Africa. Afr. Entomol. 22, 441e444. Brevault, T., Bouyer, J., 2014. From integrated to system-wide pest management: challenges for sustainable agriculture. Outlooks Pest Manag. 25, 212e213. Burel, F., 1996. Hedgerows and their role in agricultural landscapes. Crit. Rev. Plant Sci. 15, 169e190. Burel, F., Baudry, J., 1999. Ecologie du paysage. Concepts, methodes et applications. TEC & DOC, Paris, 362 p. Burel, F., Baudry, J., 2005. Habitat quality and connectivity in agricultural land- scapes: the role of land use systems at various scales in time. Ecol. Indic. 5, 305e313. Busi, R., Vila-Aiub, M.M., Beckie, H.J., Gaines, T.A., Goggin, D.E., Kaundun, S.S., Lacoste, M., Neve, P., Nissen, S.J., Norsworthy, J.K., Renton, M., Shaner, D.L., Trane, P.J., Wright, T., Yu, Q., Powles, S.B., 2013. Herbicide-resistant weeds: from research and knowledge to future needs. Evol. Appl. 6, 1218e1221. Carriere, Y., Goodell, P.B., Ellers-Kirk, C., Larocque, G., Dutilleul, P., Naranjo, S.E., Ellsworth, P.C., 2012. Effects of local and landscape factors on population dy- namics of a cotton pest. PLoS One 7, e39862. Chaplin-Kramer, R., O'Rourke, M.E., Blitzer, E.J., 2011. A meta-analysis of crop pest and natural enemy response to landscape complexity. Ecol. Lett. 14, 922e932. Cook, S.M., Khan, Z.R., Pickett, J.A., 2007. The use of push-pull strategies in inte- grated pest management. Annu. Rev. Entomol. 52, 375e400. Deguine, J.-P., Ferron, P., Russell, D., 2008. Sustainable pest management for cotton production. A review. Agron. Sustain. Dev. 28, 113e137. Fig. 4. From local suppression to activation of regulation processes for ecological pest Diatta, P., Rey, J.-Y., Vayssieres, J.-F., Diarra, K., Coly, E.V., Lechaudel, M., Grechi, I., management. Diversity of approaches and scales of action within DIVECOSYS. Along Ndiaye, S., Ndiaye, O., 2013. Effect of the physiological state and development the Y-axis, action is positioned according to a qualitative gradient of ecosystem service stage of fruit (citruses, mangos and papayas) on the laying preference of Bac- use. To illustrate this, the grey line connects different studies carried out on mango trocera invadens Drew Tsuruta & White (Tephritidae: Diptera). Fruits 68, fruit flies in Senegal and Benin. 507e516. 60 T. Brevault et al. / Crop Protection 66 (2014) 53e60
Ehui, S., Pender, J., 2005. Resource degradation, low agricultural productivity, and Savary, S., Ficke, A., Aubertot, J.N., Hollier, C., 2012. Crop losses due to diseases and poverty in sub-Saharan Africa: pathways out of the spiral. Agric. Econ. 32, their implications for global food production losses and food security. Food 225e242. Secur. 4, 519e537. FAO, 2012. Global Pact Against Plant Pests Marks 60 Years in Action. Online at: Settle, W., Garba, M.H., 2011. Sustainable crop production intensification in the http://www.fao.org/news/story/en/item/131114/icode/ (accessed 25.03.14.). Senegal and Niger River basins of francophone West Africa. Int. J. Agric. Sustain. FAO, 2008. Conservation Agriculture, 2008-07-08. http://www.fao.org/ag/ca/index. 9, 171e185. html (accessed 22.05.14.). Silvie, P.J., Renou, A., Vodounnon, S., Bonni, G., Obayomi Adegnika, M., Hema, O., Giller, K.E., Witter, E., Corbeels, M., Tittonell, P., 2009. Conservation agriculture and Prudent, P., Soreze, J., Ochou Ochou, G., Togola, M., Badiane, D., Ndour, A., smallholder farming in Africa: the heretics' view. Field Crop. Res. 114, 23e34. Komlan Akantetou, P., Ayeva, B., Brevault, T., 2013. Threshold-based in- Gliessman, S.R. (Ed.), 1990. Agroecology: Researching the Ecological Basis for Sus- terventions for cotton pest control in West Africa : what's up 10 years later. tainable Agriculture. Springer-Verlag, New York. Crop Prot. 43, 157e165. Gnankine, O., Bassole, I.H., Chandre, F., Glitho, I., Akogbeto, M., Dabire, R.K., Sow, G., Arvanitakis, L., Niassy, S., Diarra, K., Bordat, D., 2013a. Life history traits of Martin, T., 2013. Insecticide resistance in Bemisia tabaci Gennadius (Homoptera: Oomyzus sokolowskii Kurdjumov (Hymenoptera: Eulophidae), a parasitoid of Aleyrodidae) and Anopheles gambiae Giles (Diptera: Culicidae) could compro- the diamondback moth. Afr. Entomol. 21, 231e238. mise the sustainability of malaria vector control strategies in West Africa. Acta Sow, G., Arvanitakis, L., Niassy, S., Diarra, K., Bordat, D., 2013b. Performance of the Trop. 128, 7e17. parasitoid Oomyzus sokolowskii (Hymenoptera: Eulophidae) on its host Plutella Goergen, G., Vayssieres, J.-F., Gnanvossou, D., Tindo, M., 2011. Bactrocera invadens xylostella (Lepidoptera: Plutellidae) under laboratory conditions. Int. J. Trop. (Diptera: Tephritidae), a new invasive fruit fly pest for the Afrotropical region: Insect Sci. 33, 38e45. host plant range and distribution in West and Central Africa. Environ. Entomol. Tsafack, N., Menozzi, P., Brevault, T., Soti, V., Deconchat, M., Ouin, A., 2013. Effects of 40, 844e854. landscape context and agricultural practices on the abundance of cotton boll- Grzywacz, D., Rossbach, A., Rauf, A., Russell, D.A., Srinivasan, R., Shelton, A.M., 2010. worm Helicoverpa armigera in cotton fields: a case study in northern Benin. Int. Current control methods for diamondback moth and other brassica insect pests J. Pest Manag. 59, 294e302. and the prospects for improved management with lepidopteran-resistant Bt Tscharntke, T., Clough, Y., Wanger, T.C., Jackson, L., Motzke, I., Perfecto, I., vegetable brassicas in Asia and Africa. Crop Prot. 29, 68e79. Vandermeer, J., Whitbread, A., 2012. Global food security, biodiversity con- Gurr, G.M., Wratten, S.D., Luna, J.M., 2003. Multi-function agricultural biodiversity: servation and the future of agricultural intensification. Biol. Conserv. 151, pest management and other benefits. Basic Appl. Ecol. 4, 107e116. 53e59. Hema, O., Some, H.N., Traore, O., Greenplate, J., Abdennadher, M., 2009. Efficacy of Tscharntke, T., Klein, A.M., Kruess, A., Steffan-Dewenter, I., Thies, C., 2005. Land- transgenic cotton plant containing the Cry1Ac and Cry2Ab genes of Bacillus scape perspectives on agricultural intensification and biodiversity e ecosystem thuringiensis against Helicoverpa armigera and Syllepte derogata in cotton service management. Ecol. Lett. 8, 857e874. cultivation in Burkina Faso. Crop Prot. 28, 205e214. Uhder, C., Bricas, N., Zoungrana, B., Thirion, M.C., Roy-Macauley, H., Maraux, F., Khan, Z.R., Pickett, J.A., 2004. The ‘push-pull’ strategy for stemborer management: a Konate, G., Baris, P., Demay, S., de Raissac, M., Boirard, H., Remy, P., Simon, D., case study in exploiting biodiversity and chemical ecology. In: Gurr, G.M., 2011. Les cultures vivrieres pluviales en Afrique de l'Ouest et du Centre. In: Wratten, S.D., Altieri, M.A. (Eds.), Ecological Engineering for Pest Management: Elements d'analyse et propositions pour l'action. Coll. A Savoir, vol. 6. Editions Advances in Habitat Manipulation for Arthropods. CABI, Wallington, Oxon, UK, de l'AFD, CIRAD, IFAD Publication, Paris, 192 p. pp. 155e164. Van Lenteren, J.C., Bale, J., Bigler, F., Hokkanen, H.M.T., Loomans, A.J.M., 2006. Maxmen, A., 2013. Crop pests: under attack. Nature 501, 15e17. Assessing risks of releasing exotic biological control agents of arthropod pests. Mazzi, D., Dorn, S., 2012. Movement of insect pests in agricultural landscapes. Ann. Ann. Rev. Entomol. 51, 609e634. Appl. Biol. 160, 97e113. Van Mele, P., Vayssieres, J.-F., Van Tellingen, E., Vrolijks, J., 2007. Effects of an African Meehan, T.D., Werling, B.P., Landis, D.A., Grattona, C., 2011. Agricultural landscape weaver ant, Oecophylla longinoda, in controlling mango fruit flies (Diptera: simplification and insecticide use in the Midwestern United States. Proc. Natl. Tephritidae) in Benin. J. Econ. Entomol. 100, 695e701. Acad. Sci. U. S. A. 108, 11500e11505. Van Mele, P., Vayssieres, J.-F., Adandonon, A., Sinzogan, A., 2009. Ant cues affect the Mendez, V.E., Bacon, C.M., Cohen, R., 2013. Agroecology as a transdisciplinary, partic- oviposition behaviour of fruit flies (Diptera: Tephritidae) in Africa. Physiol. ipatory, and action-oriented approach. Agroecol. Sustain. Food Syst. 37, 3e18. Entomol 34, 256e261. Monteiro, L.B., Ricci, B., Franck, P., Lavigne, C., Toubon, J.F., Sauphanor, B., 2013. Vayssieres, J.-F., Korie, S., Ayegnon, D., 2009a. Correlation of fruit fly (Diptera Predation of codling moth eggs is affected by pest management practices at Tephritidae) infestation of major mango cultivars in Borgou (Benin) with abiotic orchard and landscape levels. Agric. Ecosyst. Environ. 166, 86e93. and biotic factors and assessment of damage. Crop Prot. 28, 477e488. Ndiaye, O., Vayssieres, J.F., Ndiaye, S., Diedhiou, P.M., Ba, C.T., Diatta, P., 2012. Sea- Vayssieres, J.-F., Sinzogan, A., Ouagoussounon, I., Korie, S., Thomas-Odjo, A., 2009b. sonality and range of fruit fly (Diptera: Tephritidae) host plants in orchards in Effectiveness of spinosad bait sprays (GF-120) in controlling mango-infesting Niayes and the Thies Plateau (Senegal). Fruits 311e331. fruit flies (Diptera Tephritidae) in Benin. J. Econ. Entomol. 102, 515e521. Neuenschwander, P., 2001. Biological control of the cassava mealybug in Africa: a Vayssieres, J.-F., Sinzogan, A., Adandonon, A., Korie, S., 2010. Diversity of fruit fly review. Biol. Control 21, 214e229. species (Diptera Tephritidae) associated with citrus crops (Rutaceae) in south- Oerke, E.C., 2006. Crop losses to pests. J. Agric. Sci. 144, 31e43. ern Benin in 2008e2009. Int. J. Biol. Chem. Sci. 4, 1881e1897. Omernik, J.M., 2004. Perspectives on the nature and definition of ecological regions. Vayssieres, J.-F., Wharton, R., Adandonon, A., Sinzogan, A., 2011. Preliminary in- Environ. Manag. 34, 27e38. ventory of parasitoids associated with fruit flies in mangoes, guavas, cashew, Parsa, S., Morse, S., Bonifacio, A., Chancellor, T.C.D., Condori, B., Crespo-Perez, V., pepper and wild fruit crops in Benin. Biocontrol 56, 35e43. Hobbs, S.L.A., Kroschel, J., Ba, M., Rebaudo, F., Sherwood, S.G., Vanek, S.J., Faye, E., Vayssieres, J.-F., Adandonon, A., N'Diaye, O., Sinzogan, A., Kooymann, C., Badji, K., Herrera, M.A., 2014. Obstacles to integrated pest management adoption in Rey, J.-Y., Wharton, R.A., 2012. Native parasitoids associated with fruit flies developing countries. Proc. Natl. Acad. Sci. U. S. A. 111, 3889e3894. (Diptera: Tephritidae) in cultivated and wild fruit crops in Casamance, Senegal. Pettorelli, N., 2012. Climate change as a main driver of ecological research. J. Appl. Afr. Entomol. 20, 308e315. Ecol. 49, 542e545. Veres, A., Petit, S., Conord, C., Lavigne, C., 2013. Does landscape composition affect Power, A.G., 2010. Ecosystem services and agriculture: tradeoffs and synergies. Proc. pest abundance and their control by natural enemies? A review. Agric. Ecosyst. R. Soc. B Biol. Sci. 365, 2959e2971. Environ. 166, 110e117. Ratnadass, A., Fernandes, P., Avelino, J., Habib, R., 2012. Plant species diversity for Woltz, J.M., Isaacs, R., Landis, D.A., 2012. Landscape structure and habitat man- sustainable management of crop pests and diseases in agroecosystems: a re- agement differentially influence insect natural enemies in an agricultural view. Agron. Sustain. Dev. 32, 273e303. landscape. Agric. Ecosyst. Environ. 152, 40e49. Renou, A., Ter eta, I., Togola, M., 2011. Manual topping decreases bollworm in- Wu, Y., 2006. Landscape ecology, cross-disciplinarity, and sustainability science. festations in cotton cultivation in Mali. Crop Prot. 30, 1370e1375. Landsc. Ecol. 21, 1e4. Renou, A., Ter eta, I., Togola, M., Brevault, T., 2012. First steps towards « green » Youm, O., Vayssieres, J.F., Togola, A., Robertson, S.P., Nwilene, F.E., 2011. International cotton in Mali. Outlooks Pest Manag. 23, 173e176. trade and exotic pests: the risks for biodiversity and African economies. Outlook Rusch, A., ValantinMorison, M., Sarthou, J.P., RogerEstrade, J., 2013. Effect of crop Agric. 40, 59e70. management and landscape context on insect pest populations and crop damage. Agric. Ecosyst. Environ. 166, 118e125.