THÈSE / UNIVERSITÉ DE RENNES 1 sous le sceau de l’Université Bretagne Loire
pour le grade de DOCTEUR DE L’UNIVERSITÉ DE RENNES 1 Mention : Biologie Ecole doctorale Vie Agro Santé
présentée par Bastien Labarrere
Préparée à l’Unité de Recherche CNRS 6553 Ecobio Ecosystèmes, Biodiversité, Evolution Université de Rennes 1
Thèse soutenue à Rennes Comment les plantes le 16 janvier 2017 répondent et s’adaptent devant le jury composé de : aux changements Annabel PORTÉ Chargée de Recherche INRA – Université de climatiques: Bordeaux1 / rapporteur Jacqui SHYKOFF étude aux marges froides Directrice de Recherche CNRS – Université Paris Sud / rapporteur (subantarctique) Sophie NADOT Professeur – Université Paris Sud / examinatrice Cécile SULMON Maître de Conférences – Université de Rennes 1 / examinatrice Françoise HENNION Chargée de Recherche CNRS - Rennes / directrice de thèse Andreas PRINZING Professeur- Université de Rennes 1 / co- directeur de thèse
Remerciements
Tout d’abord, je tiens à remercier Françoise et Andreas, pour m’avoir donné l’opportunité de faire cette thèse. Merci à vous deux pour votre présence, votre aide et vos conseils. Merci de m’avoir fait progresser pendant 3 ans. Merci de votre patience pour les relectures. Merci de m’avoir encadré tout en me laissant le champ de tester ce que j’avais envie. Merci tout simplement de m’avoir permis de vivre cette aventure, enrichissante à tous niveaux. Merci de m’avoir permis de découvrir des personnes et des lieux exceptionnels, même si certains ruisseaux ont été plus difficiles à traverser que d’autres.
Merci aux membres du jury d’avoir accepté d’évaluer mon projet.
Merci à Pete, Claudia et Mathias, de m’avoir chaleureusement accueilli pendant près de 3 mois, et de nous avoir reçu, comme des rois, à Jena, malgré des résultats sportifs ne jouant pas en notre faveur. Merci également à Pete, Claudia, Mathias, Nicole et Trish pour leur aide concernant les analyses transcriptomiques.
Merci à Valérie Gouesbet et Oscar de leur aide préciseuse pour les manips. Merci à Valérie de m’avoir aidé lors des manips en labo, et particulèrement de t’être aussi bien occupée de mes choux quand j’étais loin. Merci à Oscar de m’avoir conseillé et aidé pour l’extraction ARN. Merci également à Jean Luc, Thierry et Fouad pour leur aide.
Merci à Françoise Lamy pour ton aide sur le terrain. Une belle aventure humaine. Merci à Nina, Pierre, Yann et toutes ceux qui nous ont aidé sur le terrain. Une pensée aux compagnons d’aventure notamment les Talisker et les Pop’eleph.
Merci à Francisco et à toute l’équipe pédagogique pour m’avoir encadré lors de la mission d’enseignement. Merci à Valérie, Sandra, Isabelle, Fabienne, Tifenn et Badia sans qui rien ne serait possible. Merci de vous occuper aussi bien de nous. Merci à tous les membres d’Ecobio, que j’ai croisé et cottoyé pendant ces 3 années.
Merci aux anciens, qui sont partis, Kevin H et Maxime D, Grégoire, qui m’ont accueilli à Rennes.
Merci à Maud, Morgane, Tiphaine, Alix, Kevin P, Nathan. Merci particulièrement à Kevin et Maxime, pour les moments partagés dans le bureau. Merci à Kevin de ne pas m’avoir trop fait bavé en envoyant des photos de pulled pork. Maxime, occupe-toi bien de la relève.Merci aux espagnols, Juan et Jose qu’il est bon d’entendre un accent du sud.
Merci particulièrement à mes petites stagiaires. Emeline et Marine, Je vous souhaite de réussir, vous le méritez. J’espère avoir pu vous aider, Un grand merci également à Diana, AA, Micka, Alexis, Adrien, pour ces moments passés ensemble.
Enfin, Merci surtout à Flore, pour beaucoup de choses, mais tu le sais. Merci à tous ceux que j’ai dû certainement oublié, et qui n’ont pas vu leur nom apparaitre. Merci enfin à mes proches, qui m’entourent au quotidien
SOMMAIRE
Introduction ...... 1 1. Changement climatique aux îles Kerguelen...... 1 1.1. Présentation des îles Kerguelen ...... 1 1.2. Conditions environnementales des îles Kerguelen et changement climatique ...... 3 1.3. Présentation des espèces étudiées ...... 5 2. Potentiel adaptatif des espèces face aux modifications de l’environnement ...... 10 2.1. Définition de potentiel adaptatif ...... 10 2.2. Variation des traits ...... 11 2.2.1Notions générales : ...... 11 2.2.2 Variation établie : ...... 12 a) Variation inter et intra population ...... 12 b) Le cas des métabolites de réponse aux stress ...... 14 2.2.3 Variation potentielle : plasticité ...... 16 a) Définition ...... 16 b) Plasticité et bénéfices… ...... 16 c) … accompagnés de coûts ...... 17 3. Forte intégration phénotypique : quelle influence sur le potentiel adaptatif des espèces ? ...... 18 3.1. Définition de l’intégration phénotypique ...... 18 3.2. Influence de l’environnement sur l’intégration phénotypique ...... 20 3.3. Influence de l’intégration sur la variabilité des traits (établie et potentielle) ...... 21 Chapitre 1. Relation entre intégration phénotypique, variabilité des traits intra-population et performance des plantes dans des environnements naturels ...... 25 Chapitre 1A. High phenotypic integration is correlated with high trait variation in both natural and controlled conditions, in sub-Antarctic plant populations ...... 26 RESUME : ...... 27 INTRODUCTION ...... 30 MATERIALS AND METHODS ...... 35 RESULTS ...... 45 DISCUSSION ...... 48 CONCLUSION ...... 52 REFERENCES ...... 54
Chapitre 1B. Exploring performance of species from Iles Kerguelen: relationships to environment and phenotypic integration ...... 74 RESUME ...... 75 INTRODUCTION ...... 76 MATERIALS AND METHODS ...... 79 RESULTS ...... 84 DISCUSSION ...... 87 CONCLUSION ...... 91 REFERENCES ...... 92 Chapitre 2. Variations of secondary metabolites among natural populations suggest functional redundancy and versatility within sub-Antarctic Ranunculus species ...... 100 RESUME ...... 101 INTRODUCTION ...... 103 MATERIALS AND METHODS ...... 108 RESULTS ...... 115 DISCUSSION ...... 120 CONCLUSION ...... 125 REFERENCES ...... 128 Chapitre 3. Plant plasticity is constrained by performance costs, complex environments and weakly integrated phenotypes ...... 148 RESUME ...... 149 INTRODUCTION ...... 151 MATERIALS AND METHODS ...... 155 RESULT ...... 160 DISCUSSION ...... 163 CONCLUSION ...... 169 REFERENCES ...... 171 Chapitre 4 : Etude préliminaire de la variation du transcriptome, en relation avec l’environnement et le phénotype chez P. antiscorbutica ...... 188 INTRODUCTION ...... 188 MATERIELS ET METHODES ...... 190 RESULTATS ET INTERPRETATIONS ...... 192 DISCUSSION GENERALE ...... 198 1. Avancées conceptuelles ...... 199 1.1. Le degré d’intégration phénotypique est indépendant de l’environnement ...... 199
1.2. L’Intégration augmente la variabilité des traits établie et potentielle mais pas la performance...... 201 1.3. Flexibilité fonctionnelle des métabolites ...... 203 1.4. Plasticité plus importante ne signifie pas meilleure performance...... 204 1.5. Perspectives ...... 205 2. Potentiel adaptatif des espèces de Kerguelen ...... 209 2.1. Variation intra-population : probablement pas menacée par les changements climatiques ...... 209 2.2. Variation inter-population : faible variation chez R. moseleyi ...... 210 2.3. Variation inter-population : performance plus faible en milieux secs et chauds...... 211 2.4. Communauté avoisinante : pas d’effets sur les traits ou la performance ...... 213 2.5. Réponses plastiques aux changements environnementaux : pas un témoin de performance ...... 214 2.6. Réchauffement et ombrage combinés limitent la plasticité et la performance ...... 215 Conclusion : le potentiel adaptatif dépend de la persistance des habitats ...... 217 REFERENCES ...... 218
Introduction générale
Introduction 1. Changement climatique aux îles Kerguelen
1.1. Présentation des îles Kerguelen
Les îles Kerguelen (49°20’00” S, 69°20’00” E) sont situées dans l’Océan Indien austral, dans la zone subantarctique (Lebouvier et Frenot, 2007 ; Figure 1). Les îles Kerguelen font partie de la
Province biogéographique de l’Océan Indien Austral (« South Indian Ocean Province » ou SIOP, Chown et al. 1998), qui regroupe également les îles Crozet, Marion, Prince Edward et Heard
(Smith, 1984). Partie du territoire Français des TAAF (Terres Australes et Antarctiques
Françaises), cet archipel est extrêmement isolé et distant d’environ 3500 km des côtes de l’Afrique du Sud et de l’Australie (Lebouvier et Frenot, 2007). D’origine volcanique, les îles Kerguelen ont
émergé au début du Tertiaire il y a 115 millions d’années, et reposent sur un plateau océanique basaltique (Nicolaysen et al. 2000). Ces îles ont une superficie de 7215 km², (avoisinant la superficie de la Corse) et sont constituées d’une île principale (appelée Grande Terre) autour de laquelle on trouve 300 îles et îlots pour la plupart très proches de l’île principale. L’archipel est dominé par le Mont Ross (1850m) et est en partie recouvert par la calotte glaciaire Cook à l’Ouest.
Les îles Kerguelen abritent d’importants dénivelés qui s’accompagnent d’une diminution très importante des températures, la diminution de la température avec l’altitude est ainsi estimée à environ 0.8°C par 100m (Hennion et al., 2006). A titre de comparaison la diminution de température est de 0.6°C par 100m dans les Alpes (Douguédroit et Saintignon, 1970). Les conditions environnementales trouvées au sein des îles Kerguelen sont très variées (Wagstaff et
Hennion, 2007).
La forêt tempérée originelle qui a connu son apogée il y a 20 millions d’années (Philippe et al.,
1998) laisse aujourd’hui place à une végétation pauvre, dépourvue de ligneux, de type toundra. On compte 22 espèces de phanérogames autochtones (Frenot et al., 2001). A titre de comparaison, la
France métropolitaine en compte près de 3000. Cette flore comporte une composante d’origine 1
Introduction générale magellanique et une composante d’origine australo-néozélandaise, mais également 6 espèces endémiques de la Province de l’Océan Indien Austral (Hennion et Walton, 1997a). Elle est composée uniquement de plantes pérennes de types mégaherbes, graminées et formes en coussins ainsi que de bryophytes. Les communautés végétales présentes sur l’archipel sont relativement simples (Pansu et al., 2015). L’Homme est seulement présent depuis la fin du XVIIIe siècle et en faibles contingents (Frenot et al, 2006). Malgré tout, il a eu un impact important sur la flore locale, impact qui s’est intensifié après l’installation de la base permanente de Port-aux-Français (Grande
Terre) dans les années 1950, avec l’introduction de nombreuses espèces végétales originaires des régions tempérées (86 espèces, Frenot et al. 2006). Si l’impact des espèces végétales introduites est jusqu’à présent limité, un élément majeur pour la flore autochtone a été l’introduction du lapin en 1874, ce qui a très profondément affecté les communautés végétales natives des îles Kerguelen
(Chapuis et al., 2004). Il est à noter que notre étude se situe dans des régions exemptes de lapins.
Différents habitats sont reconnus dans l’archipel (Chastain, 1958 ; Smith, 1984 ; Figure 3), principalement discriminés par l’exposition au vent et le degré de drainage des sols (Hennion,
1992). Deux habitats sont particulièrement représentés : (i) les prairies herbacées de basse altitude : zones humides et protégées du vent, abritant une végétation dense (Hennion et al., 2006 ; Figure
3) (ii) les fell-fields : plateaux exposés aux vents et à l’érosion, caractérisés par des sols pauvres à faible capacité de rétention d’eau (Figure 3). Ces habitats sont principalement trouvés en altitude, et abritent une végétation ouverte dominée par des plantes en coussin.
2
Introduction générale
1.2. Conditions environnementales des îles Kerguelen et changement climatique
Les îles Kerguelen sont caractérisées par des températures froides et stables avec une moyenne annuelle de 4,6°C (2.1°C en août à 7.7°C en février), accompagnées de pluies abondantes (800-
3200mm annuels) (Frenot et al., 2006 ; Lebouvier et Frenot, 2007). Une autre caractéristique du climat subantarctique est représentée par les vents violents (Lebouvier et Frenot, 2007), qui modèlent les communautés végétales (Hennion, 1992 ; Le Roux et al., 2008). Actuellement, ces
îles font face à un changement climatique rapide et intense (Lebouvier et al., 2011). Des enregistrements météorologiques sont réalisés depuis plus de 50 ans par Météofrance à la base de
Port-aux-Français. On note une augmentation moyenne annuelle de 1,3°C depuis 1960, accompagnée d’une baisse du nombre de jours de gel annuels (baisse de 20 à 30 jours en 20 ans ; Frenot et al., 2006 ; Lebouvier et al., 2011). Cette augmentation de température peut paraître faible, mais elle est très importante relativement aux moyennes et amplitudes thermiques locales.
Egalement, on note une diminution importante des précipitations en quantité (diminution de 100 à
250mm annuels en 20 ans) et en fréquence (Chapuis et al., 2004 ; Frenot et al., 2006). Le Roux et al. (2008) travaillant sur les îles Marion, situées également dans la province SIOP, ont montré que le changement climatique se caractérisait également par une augmentation de la vitesse du vent
(vitesse moyenne et vitesse maximale). L’augmentation de la vitesse du vent devrait avoir un effet plus important sur la végétation dans les fell-fields d’altitude, très exposés, et aurait également pour effet une augmentation de l’évaporation des milieux (Le Roux et al., 2008). De plus, une diminution des précipitations et une augmentation de l’évaporation devraient entraîner l’augmentation de la salinité des milieux (Le Roux et al., 2008).
Les modifications des conditions environnementales entraînent d’ores et déjà des épisodes de sécheresse estivale à Kerguelen (Hennion, 1992 ; Frenot et al., 2006). Les effets du changement climatique sur les espèces végétales sont supposés être plus importants aux fortes latitudes
(Bergstrom et Chown, 1999 ; Le Roux et McGeoch 2007).
3
Introduction générale
a
Figure 1. (a) Localisation des îles Kerguelen dans l’Océan Indien austral (b) régions
échantillonnées dans les îles Kerguelen, d’après une carte IGN.
4
Introduction générale
De fait, les espèces autochtones de l’archipel – qui poussent habituellement dans des conditions particulièrement froides et humides - montrent d’ores et déjà des signes évidents de stress hydrique lors des épisodes de sécheresse estivale (Hennion, 1992 ; Chapuis et al., 2004 ; Lebouvier et al.,
2011). Particulièrement, les espèces inféodées aux milieux humides mais bien drainés, telles que la Brassicaceae Pringlea antiscorbutica R.Br. ex Hook.f. ou possédant des racines peu profondes, telles que les espèces de Ranunculus, devraient être les plus affectées par le changement climatique
(Le Roux et al., 2008).
A contrario, les espèces végétales introduites, provenant de régions tempérées, pourraient profiter de ces nouvelles conditions (Frenot et al., 2001). Les conditions climatiques froides des
îles Kerguelen sont un frein à l’installation et au développement de nouvelles espèces (Sinclair et al., 2003). Cependant, si le réchauffement persiste, les espèces introduites pourraient augmenter considérablement leur distribution sur l’île ainsi que leur impact sur la flore locale (Frenot et al.,
2006). On note d’ores et déjà une régression de l’autochtone Acaena magellanica au profit du pissenlit introduit Taraxacum officinale, régression qui serait due aux sécheresses estivales (Frenot et al., 2006).
1.3. Présentation des espèces étudiées
Dans cette thèse nous étudierons 4 espèces autochtones des îles Kerguelen, à savoir la Brassicaceae
Pringlea antiscorbutica R.Br. ex Hook.f. et les trois espèces de Ranunculus : Ranunculus biternatus Smith ; Ranunculus pseudotrullifolius Skottsb. ; et Ranunculus moseleyi Hook.f.
(Figure 2). Pringlea antiscorbutica aussi appelé le Chou de Kerguelen, est une espèce endémique de la Province Sud de l’Océan Indien (Van der Putten et al., 2010). Le Chou de Kerguelen est une espèce pérenne à reproduction sexuée uniquement (Hennion et Walton 1997a). Les individus matures produisent une quantité importante de graines (Hennion et Walton 1997a). Ces graines peuvent être dispersées par le vent (Schermann-Legionnet et al., 2007), mais également 5
Introduction générale
Ranunculus biternatus Smith Ranunculus pseudotrullifolius Skottsb.
Ranunculus moseleyi Hook.f. Brassicaceae Pringlea antiscorbutica R.Br. ex Hook.f.
Figure 2. Présentation des espèces étudiées
6
Introduction générale par l’eau de mer entre les différentes îles de l’archipel grâce à l’existence d’une gaine mucilagineuse gonflant au contact de l’eau (Hennion et Walton 1997a). P. antiscorbutica est autocompatible, et sa reproduction est essentiellement autogame (Schermann-Legionnet et al.,
2007). La phénologie de la floraison, notamment la déhiscence des fruits, varie selon l’altitude des individus, étant plus précoce à basse altitude (janvier) qu’à haute altitude (mars-avril) (Hennion et
Walton 1997a). P. antiscorbutica est présente dans une large gamme d’habitats à Kerguelen : on trouve la plante des pelouses littorales exposées aux embruns aux champs d’altitude (fell-fields) exposés aux vents jusqu’à environ 1000m, en passant par des prairies luxuriantes intérieures abritées du vent (Chastain, 1958). Ses habitats sont humides mais bien drainés (Hennion et Walton,
1997a). En cohérence avec cette distribution, P. antiscorbutica montre une forte tolérance aux températures froides, notamment aux épisodes de gel-dégel, ainsi qu’une tolérance aux embruns et
à l’inondation par les marées (Hennion et Walton, 1997b ; Hennion et Bouchereau, 1998 ; Hummel et al., 2004a,b).
Les trois espèces de renoncules étudiées présentent des distributions différentes. Ranunculus biternatus est circumpolaire australe, R. pseudotrullifolius uniquement magellanique et à
Kerguelen, tandis que R. moseleyi est décrite comme endémique des îles Kerguelen et des îles du
Prince Edward (Van der Putten et al. 2010). Cependant, la présence de R. moseleyi dans les îles du
Prince Edward, indiquée dans les années 1970, est remise en cause. En effet, R. moseleyi n’a plus
été retrouvée dans ces îles depuis des décennies (Steven Chown, comm. pers. à Françoise Hennion), ce qui ferait de R. moseleyi une espèce endémique stricte de Kerguelen (Lehnebach et al., 2017).
Les trois espèces de Ranunculus sont pérennes, et occupent des habitats différents sur l’archipel.
Ranunculus biternatus occupe une grande diversité d’habitats aussi bien aquatiques que terrestres
7
Introduction générale jusqu’à 500 m d’altitude (Hennion and Walton, 1997a; Réserve Naturelle des Terres Australes et
Antarctiques françaises and IPEV Programme 136, unpublished data).
8
Introduction générale
Côte rocheuse avec communauté à Crassula Mare proche du littoral avec Ranunculus Eboulis et ruisseau de fell-field avec communauté moschata et Ranunculus pseudotrullifolius pseudotrullifolius et Acaena magellanica à Azorella selago, Acaena magellanica et Pringlea (Ile Australia) (Isthme Bas) antiscorbutica et bryophytes (Val Studer, Mont Crozier, altitude 400m)
Mare asséchée avec Ranunculus moseleyi et Acaena magellanica Prairie de mégaherbes à basse altitude avec Prairie intérieure à Acaena magellanica Pringlea antiscorbutica (Ile Australia) avec mares à Ranunculus biternatus (Port-Jeanne d’Arc, altitude 230m) (Val Studer, altitude 80m)
Figure 3. Présentation des milieux d’étude 9
Introduction générale
Ranunculus pseudotrullifolius, halophile, est trouvée dans des habitats proches de la côte soit terrestre sur le rivage, soit aquatique dans des mares peu profondes, et ponctuellement jusqu’à des altitudes modérées autour de 200m (Hennion et al., 1994 ; observation personnelle ; données de distribution Réserve Naturelle des TAAF / Programme IPEV 136, non publié). Ranunculus moseleyi, strictement aquatique, se trouve sur les berges de petits ruisseaux et les bords d’étendues d’eau douce (Hennion et al., 1994). Les populations de ces trois espèces forment des groupes denses de rosettes avec production de stolons. Les graines peuvent être dispersées par les cours d’eau ou le vent. Les trois espèces montrent une part importante d’autogamie par cléistogamie des fleurs immergées en conditions aquatiques, ainsi que de reproduction végétative (Hennion et al.,
1994). La floraison est observée de décembre à avril chez R. biternatus et R. pseudotrullifolius et de janvier à avril chez R. moseleyi (Hennion, 1992).
Les quatre espèces étudiées, indigènes des îles Kerguelen, sont inféodées à des milieux humides et froids (Hennion et al., 1994 ; Hummel et al., 2004b). Ces espèces sont décrites comme particulièrement vulnérables face au changement climatique (Le Roux et al., 2008). On note d’ores et déjà une mortalité importante des plantules de P. antiscorbutica (Hennion, 1992) qui pourrait
être due à la récurrence de sécheresses estivales (Frenot et al., 2006). L’objectif de cette thèse est de déterminer le potentiel adaptatif de ces espèces face au changement climatique.
2. Potentiel adaptatif des espèces face aux modifications de l’environnement 2.1. Définition de potentiel adaptatif
Il est difficile de déterminer avec certitude le devenir d’une espèce. On peut néanmoins définir la propension d’une espèce à faire face à d’éventuels changements environnementaux comme le potentiel adaptatif de l’espèce (Blow et Hoffmann, 2005 ; Valladares et al., 2007 ; Nicotra et al., 10
Introduction générale
2010 ; Bolnick et al., 2011). Le potentiel adapatif repose sur (i) la diversité existante entre individus et (ii) la réponse plastique individuelle. Une forte diversité entre individus augmente la probabilité qu’au moins un individu soit selectionné par le nouvel environnement (Blow et Hoffmann, 2005).
Une forte réponse plastique permet un ajustement des individus aux nouvelles conditions environnementales (Valladares et al., 2007). Le potentiel adaptatif sensu stricto s’intéresse aux caractères héritables des individus. Les réponses qui ne sont pas adaptatives sensu stricto parce que non héritables, peuvent cependant potentiellement faciliter des changements héritables (West-
Eberhardt, 2003). Le potentiel adaptatif au sens large considère tout ce qui contribue à la capacité d’une espèce à faire face aux possibles modifications de son environnement... Dans cette thèse, nous traiterons du potentiel adaptatif des espèces au sens large.
Le potentiel adaptatif peut être estimé à travers la variation des traits des individus, qu’il s’agisse de variation des traits (morphologique ou métabolomique) établie entre les individus (Nicotra et al., 2010), ou de capacité de réponse des individus, définie par la plasticité phénotypique (Sultan,
2000).
2.2. Variation des traits 2.2.1Notions générales :
La valeur des traits d’un individu est définie par l’histoire évolutive de l’individu (e.g. Powell et al. 2009; Prinzing et al. 2001) mais peut être également modelée par son environnement. Il est ainsi bien établi que la valeur des traits des individus peut varier, au sein d’une espèce, entre des populations soumises à différentes conditions environnementales (Van Valen, 1973; Schlichting,
1986). Une forte variation de valeurs des traits participe au potentiel adaptatif, qu’il s’agisse de
11
Introduction générale variation intra ou inter populations établie entre individus (Nicotra et al., 2010), ou de la capacité de réponse plastique des individus (West-Eberhardt, 2003).
2.2.2 Variation établie : a) Variation inter et intra population La variation des valeurs de traits existante entre populations témoigne de la capacité de l’espèce à faire face à de multiples environnements. De plus, les conditions environnementales prévalant dans les différentes populations témoignent du panel d’environnements d’ores et déjà supporté par l’espèce. Cette gamme environnementale donne une indication sur les conditions environnementales qui pourront être supportées dans le cadre du changement climatique, sans la représenter complètement (réserve de variabilité). Enfin, étudier les différences de traits entre populations soumises à différentes conditions environnementales permet de déterminer quels facteurs environnementaux influencent les traits et la performance des individus, donnant des indications sur la réponse des plantes aux changements climatiques. Dans le chapitre 1, nous déterminerons quels facteurs environnementaux influencent la performance des espèces des îles
Kerguelen.
La variation des traits existant au sein d’une population permet à la population de faire face aux changements environnementaux et participe également au potentiel adaptatif des espèces (Bolnick et al., 2011). Une forte variabilité des valeurs de traits existant à l’intérieur d’une population augmente la probabilité qu’une valeur de trait soit sélectionnée par le nouvel environnement. La variation des traits existant à l’intérieur d’une population pourrait être influencée par l’environnement (Debat et David, 2001 ; Bolnick et al., 2011). Notamment, des facteurs environnementaux stressants (Gicquiaud et al., 2002) peuvent limiter la variation intra-population des traits. En effet, des conditions stressantes peuvent augmenter la pression de sélection sur une
12
Introduction générale valeur de trait optimale et ainsi diminuer la variation existante des traits. Jusqu’à présent, peu d’auteurs ont étudié l’influence de l’environnement sur le niveau de la variation intra population des traits. Dans le Chapitre 1, nous essaierons de déterminer si des facteurs environnementaux limitent la variation intra-population des traits aux îles Kerguelen.
13
Introduction générale
b) Le cas des métabolites de réponse aux stress
Le potentiel adaptif peut être observé à différents niveaux d’un organisme : au niveau morphologique, mais également au niveau du métabolome, chaque niveau apportant des informations complémentaires. Le métabolome comprend deux classes de composés : métabolites primaires et secondaires. Les métabolites primaires (glucides, protéines et lipides) assurent le fonctionnement de base de l’organisme. Les métabolites secondaires, longtemps considérés comme de simples déchets du métabolisme (Hartmann, 2007), ont, en définitive, des rôles importants dans la réponse des plantes à leur environnement (Croteau et al., 2000 ; Groppa et Benavides, 2008).
Les métabolites secondaires sont en effet à l’interface entre l’environnement et le phénotype de la plante et sont très impliqués dans la protection des plantes contre les stress environnementaux
(Teuscher et Lindequist, 2010 ; Tiburcio et al., 2014). Parmi les métabolites secondaires, polyamines et flavonoïdes (notamment les flavonols) sont particulièrement reconnus pour être impliqués dans la réponse des plantes aux stress (Agati et al., 2012 ; Tiburcio et al., 2014 ; Hennion et al., 2016) et ce rôle semble aussi concerner certaines amines aromatiques (Hennion et al., 2006,
2012) . Les polyamines sont de plus des régulateurs du développement chez les plantes
(Bouchereau et al., 1999 ; Groppa et Benavides, 2008) avec de tels rôles montrés récemment aussi chez des amines aromatiques (Hennion et al., 2016).
Tout en étant spécifique de chaque taxon, la composition en amines (polyamines et amines aromatiques) varie en fonction des conditions environnementales chez P. antiscorbutica, R. biternatus, R. pseudotrullifolius et R. moseleyi (Hennion et al., 2012). Egalement, il a été montré que la composition en polyamines a un rôle de protection contre les stress environnementaux chez
P. antiscorbutica (Hennion et Martin-Tanguy 2000; Hummel et al 2004b). Par ailleurs, les quercétines sont les seuls flavonols présents chez ces espèces de renoncules (Hennion et al., 1994).
14
Introduction générale
Leur composition varie entre différents environnements chez R. biternatus, R. pseudotrullifolius et
R. moseleyi (Hennion et al., 1994).
Nous cherchons maintenant à déterminer dans quelle mesure amines et quercétines peuvent permettre aux différentes espèces de plantes des îles Kerguelen de faire face au changement climatique. Nous déterminerons dans cette étude comment les plantes font d’ores et déjà face à leurs conditions environnementales entre populations et quels environnements naturels peuvent
être perçus comme stressants par les plantes (par exemple faible saturation en eau, salinité, ou altitude). Dans le chapitre 2, nous déterminerons, à partir de deux familles de composés (amines et quercétines), la relation entre métabolites de réponse au stress et conditions environnementales dans les différentes populations. Nous déterminerons également les relations des compositions en polyamines et quercétines avec le phénotype.
Les métabolites secondaires montrent une forte variabilité fonctionnelle (Croteau et al., 2000;
Pichersky et Gang, 2000; Wink, 2013). Ainsi, différents métabolites peuvent partager une même fonction (i.e. redondance fonctionnelle ; Wink, 2003 ; Hanada et al., 2011). Egalement, un métabolite donné peut avoir différents rôles au sein d’une même plante, dans des organes ou des environnements différents (i.e. versatilité fonctionnelle ; Wink, 2003; Lehmann et al., 2010; Di
Ferdinando et al., 2014). Cette variabilité fonctionnelle peut avoir des conséquences importantes sur la façon dont les plantes font face à leur environnement. Jusqu’alors, redondance et versatilité fonctionnelles n’ont pas été étudiées au niveau intra-spécifique, entre des populations distantes mais soumises à des environnements similaires. Un autre objectif du chapitre 2 sera de déterminer,
à partir de deux familles de composés (amines et quercétines) dans quelle mesure une variabilité fonctionnelle (redondance ou versatilité) est trouvée au sein des 3 espèces de Ranunculus R. biternatus, R. pseudotrullifolius et R. moseleyi.
15
Introduction générale
2.2.3 Variation potentielle : plasticité
a) Définition
Si le concept de plasticité est connu depuis le début du XXème siècle (Johansen, 1909), il faut attendre Bradshaw (1965) avant que la plasticité ne soit clairement présentée. La plasticité est alors définie comme la différence des valeurs de traits d’un même génotype dans différents environnements. Des précisions ont, depuis, été apportées à cette définition (Sultan, 2000). Par exemple, on parle de plasticité réversible (i.e. flexibilité) ou irréversible (i.e. plasticité développementale ; Piersma et Drent, 2003). Le développement des analyses moléculaires a permis de mieux comprendre les mécanismes sous-jacents à la plasticité, notamment les processus
épigénétiques (Bird, 2002). Les études de génétique quantitative ont également permis de déterminer que la plasticité pouvait être héritable (Pigliucci, 2005). La plasticité est maintenant considérée comme un caractère quantitatif soumis à sélection. Au sens large, la plasticité représente la capacité d’un individu à ajuster la valeur de ses traits à différents environnements (West-
Eberhard, 2003).
b) Plasticité et bénéfices…
Le changement climatique entraine des modifications rapides des conditions environnementales.
Les plantes, en tant qu’organismes sessiles, n’ont d’autre choix que de faire face à ces modifications de leur environnement (Sultan, 2000). La plasticité phénotypique est considérée comme un moyen de faire face aux modifications rapides de l’environnement (Valladares et al., 2007) et participe fortement au potentiel adaptatif des espèces comme nous l’avons défini ci-dessus. Les espèces végétales montrent des capacités plastiques pour de nombreux traits morphologiques (Sultan,
2000 ; Richards et al., 2006). La plasticité a été fortement étudiée dans le cadre des invasions biologiques. 16
Introduction générale
Une plus forte plasticité est majoritairement trouvée chez l’espèce envahissante en comparaison avec l’espèce indigène (Davidson et al., 2011). Cette différence de plasticité permettrait à l’espèce envahissante d’exprimer soit (i) un phénotype avantageux dans plus d’environnements que l’espèce indigène (i.e. Jack-of-all-trades) (ii) un phenotype plus avantageux que l’espèce indigène, dans un environnement favorable (i.e. Master-of-some) (iii) les deux à la fois (i.e. Jack-and-Master ;
Richards et al., 2006). Un caractère permettant aux individus d’exprimer une performance élevée dans de nombreux environnement devrait être particulièrement sélectionné. Au contraire, beaucoup d’espèces montrent une faible plasticité (Valladares et al., 2007), et de plus en plus d’auteurs soulignent ses coûts et ses limites (Valladares et al., 2007 ; Auld et al., 2009 ; Valladares et
Niinemets, 2008).
c) … accompagnés de coûts
La plasticité, fréquemment montrée comme adaptative, (Dudley, 2004), peut également être mal- adaptative (Van Kleunen et Fischer, 2005; Ghalambor et al., 2007). De plus en plus d’auteurs remettent en cause l’idée, communément acceptée, que des individus très plastiques montrent une meilleure performance en réponse à un changement que des individus peu ou pas plastiques
(Valladare et al., 2005; Ghalambor et al., 2007; Valladares et al., 2007). La réponse plastique à un changement environnemental repose sur toute une machinerie physiologique incluant la perception d’un signal environnemental, la transduction de ce signal, et l’expression d’une réponse phénotypique (Vinocur et Altman, 2005). Cette cascade physiologique et morphologique induit des coûts pour l’organisme, qui sont définis en deux catégories. On distingue les coûts constitutifs (ou de maintenance) des coûts induits (ou de production ; Sultan et Spencer, 2002 ; Chevin et al., 2010).
Les coûts constitutifs font références à l’entretien de la machinerie physiologique ou l’acquisition d’informations environnementales. Les couts constitutifs sont indépendants de l’environnement et 17
Introduction générale présents chez tous les individus plastiques (Auld et al., 2009). Les coûts induits font référence aux coûts de la modification phénotypique lors d’une réponse plastique, et dépendent de l’intensité de la réponse (Sultan et Spencer, 2002 ; Chevin et al., 2010). Les coûts associés à la plasticité pourraient affecter la performance des individus plastiques. Au-delà des modèles théoriques, peu d’études empiriques ont montré des couts (constitutifs et induits) de plasticité et encore moins ont analysé leur influence sur la performance des individus (Pigliucci, 2005 ; Valladares et al., 2007;
Auld et al., 2009).
Dans le chapitre 3, nous déterminerons les capacités plastiques des différentes espèces, face à des conditions environnementales susceptibles d’être trouvées dans le cadre d’un changement climatique. Nous chercherons à mettre en évidence les couts éventuels (constitutifs ou induits) de la plasticité et déterminer si une réponse plastique est liée à une meilleure ou une moins bonne performance.
3. Forte intégration phénotypique : quelle influence sur le potentiel adaptatif des espèces ?
3.1. Définition de l’intégration phénotypique
Les différents traits d’une plante ne varient pas indépendamment les uns des autres (Murren, 2002).
La variation des traits est aussi influencée par des interactions intrinsèques entre traits.
L’intégration phénotypique représente l’interdépendance des traits morphologiques et résulte de la nécessité d’un organe ou d’un organisme à maintenir des valeurs de traits cohérentes. L’intégration phénotypique est formellement définie comme le nombre et la force des corrélations entre les traits d’un individu (Pigliucci et al., 2003; Murren, 2012; Armbruster et al., 2014 ; Figure 4).
L’intégration phénotypique pourrait résulter de contraintes intrinsèques génétiques, développementales ou fonctionnelles (Pigliucci et Kaplan, 2010 ; Murren, 2012). Des modèles
18
Introduction générale génétiques conceptuels identifient la pléiotropie ou les déséquilibres de liaison comme des causes possibles de l’intégration (Conner, 2002 ; Murren, 2012). La pléiotropie est définie comme la capacité d’un gène à influencer l’expression de multiples autres gènes, qui voient alors leurs expressions corrélées. Les déséquilibres de liaison impliquent quant à eux, une interconnexion physique entre plusieurs gènes. L’intégration phénotypique peut également résulter de contraintes développementales ou fonctionnelles (Armbruster et al., 2014), différents traits ayant une croissance coordonnée afin de respecter la cohérence architecturale développementale ou fonctionnelle de l’organe dans lequel ils sont impliqués et in fine de l’organisme.
L’intégration phénotypique a principalement été étudiée dans le cadre de l’interaction entre traits floraux et pollinisateurs (Murren, 2012). Il a cependant été mis en évidence que l’intégration concernait aussi bien les traits floraux que végétatifs (Armbruster et al., 2014). La notion d’intégration est étroitement reliée à la notion de modularité fonctionnelle (Diggle, 2014). Ainsi, différents traits au sein d’un même module fonctionnel (tel que le module floral), sont plus fortement corrélés que des traits appartenant à différents modules fonctionnels. La modularité reflète différents niveaux d’intégration fonctionnelle au sein d’un individu. On observe alors un degré décroissant d’intégration entre les traits d’un organe ; les traits de différents organes rattachés
à la même fonction ; et in fine les traits de l’organisme entier. La modularité, permet le maintien de la cohérence fonctionnelle des organes, tout en autorisant la variation indépendante de traits impliqués dans des fonctions différentes.
Si l’intégration phénotypique est définie au niveau de l’individu, elle s’estime cependant au niveau de plusieurs individus (Schlichting, 1989 ; Rapson et Maze, 1994 ; Pigliucci, 2003). En effet, il est techniquement nécessaire d’observer une variation de traits pour estimer un degré d’intégration. Il est important de préciser que - si l’on excepte un degré de variance nul (qui ne
19
Introduction générale permet pas d’estimer l’intégration) - le degré de variance des traits n’influence cependant en rien le degré d’intégration. Contrairement à l’intuition générale, le degré de variation des traits n’est pas, mathématiquement positivement (ou négativement) relié au degré d’intégration phénotypique
(Figure 1, Chapitre 1).
. . Influence de l’environnement sur l’intégration phénotypique
L’intégration phénotypique pourrait résulter de pressions environnementales extrinsèques à l’organisme. Les pressions de sélection s’appliquant sur les différents traits d’un organisme, pourraient favoriser la combinaison de certaines valeurs de traits au sein d’une population
(Schlichting 1989). On note cependant peu de preuves empiriques de l’influence de l’environnement sur le degré d’intégration. Il a été montré que les stress abiotiques, en augmentant les pressions de sélection, peuvent augmenter le degré d’intégration (voir Gianoli et Palacio-Lopez,
2009). Cependant, ces études sont menées en laboratoire et étudient une réponse des plantes à court terme, face à l’application d’un unique facteur de stress abiotique. Ces études sont très éloignées des conditions naturelles des plantes, qui sont généralement exposées à une combinaison de facteurs environnementaux, qui plus est, relativement similaire depuis des générations. Il a été suggéré que la variation corrélée de différents facteurs environnementaux chacun influençant différents traits pouvait expliquer l’intégration phénotypique.
20
Introduction générale
Jusqu’alors, peu d’auteurs ont étudié la variation du degré d’intégration phénotypique en conditions naturelles. Parmi ces auteurs, Hermant et al. (2013) ont montré une variation du degré d’intégration phénotypique entre 18 espèces des îles Kerguelen. Afin de comprendre l’influence de l’environnement sur l’intégration phénotypique, il est maintenant nécessaire d’étudier l’intégration au niveau des populations. En effet, les individus d’une même espèce sont soumis à des conditions environnementales différentes. Les individus d’une même population sont, quant à eux, soumis à des conditions environnementales relativement similaires. Comparer les niveaux d’intégration entre populations exposées à différents environnements permettra de déterminer l’influence de l’environnement sur l’intégration. Boucher et al. (2013) ont montré une variation de l’intégration phénotypique entre populations d’une même espèce, qui pourrait être expliqué par des proxys environnementaux. Un des travaux de cette thèse sera de déterminer, la relation entre facteurs environnementaux et degré d’intégration chez plusieurs espèces végétales.
3.3. Influence de l’intégration sur la variabilité des traits (établie et potentielle)
Les conséquences écologiques de l’intégration phénotypiques n’ont que rarement été étudiées
(Gianoli et Palacio-Lopez, 2009). Pourtant, l’intégration des traits d’un organisme pourrait influencer son interaction avec l’environnement à l’instar d’autres paramètres comme la valeur moyenne de ces traits (Jones et al., 2013). Hornoy et al. (2011) ont par ailleurs montré que l’intégration des traits, plutôt que leur moyenne, expliquait les capacités d’invasion d’Ulex europaeus. Il revient de déterminer l’influence de l’intégration phénotypique sur la variation des traits (Murren, 2012). Particulièrement, Hermant et al. (2013) ont montré que les espèces végétales des îles Kerguelen présentant un plus fort degré d’endémisme montrent un degré d’intégration plus important.
21
Introduction générale
a) b)
Trait 2 Trait Trait 2 Trait
Trait 1 Trait 1
Faible intégration Forte intégration phénotypique phénotypique
Figure 4. Schémas représentant différents degrés d’intégration phénotypique (i.e. degré de corrélation entre les traits) chez un groupe d’individus. Un point représente un individu, caractérisé par deux traits représentés en abscisse (trait 1) et ordonnée (trait 2). Le schéma (a) représente une faible intégration phénotypique, les deux traits concernés variant indépendamment. A l’inverse le schéma (b) représente une forte intégration phénotypique, on remarque que lorsque la valeur du trait 1 augmente, la valeur du trait 2 augmente aussi.
22
Introduction générale
Nous cherchons donc à déterminer l’influence d’un fort degré d’intégration sur l’écologie des espèces, notamment sur leur capacité de variation des traits, qu’il s’agisse de variation établie au sein des populations, ou de potentiel plastique.
L’intégration des traits peut avoir des répercussions majeures sur la capacité de variation de ces derniers. En théorie l’intégration phénotypique peut diminuer ou augmenter la variation des traits, aussi bien établie que plastique. Un trade-off est généralement supposé entre intégration et variation des traits. En effet, un fort degré d’intégration peut être relié à une faible variation des traits, car les traits ne peuvent pas varier de façon indépendante. Gianoli et Palacio-Lopez (2009) ont montré que l’intégration était négativement corrélée à la plasticité de certains traits chez deux espèces végétales. D’un autre côté, chez les organismes fortement intégrés, la modification extrême de la valeur d’un trait s’accompagne d’une variation extrême de la valeur des autres traits. Chez des organismes faiblement intégrés la variation extrême d’une valeur de trait, indépendante des autres traits peut nuire à la cohérence et l’intégrité de l’organisme. Un degré d’intégration plus important autorise alors une variation des traits plus importante. Egalement, une variation environnementale peut entrainer la variation de la valeur d’un unique trait (Pigliucci, 2001, Richards et al., 2006). Or, chez un organisme intégré, la variation de la valeur d’un trait va imposer la variation de la valeur des autres traits –contrairement aux organismes moins intégrés - augmentant ainsi la variation des traits globale de l’organisme. Sur le terrain nous testerons la relation entre intégration et variation des traits à l’intérieur des populations chez des espèces indigènes des îles Kerguelen et nous confirmerons nos observations en laboratoire. Dans le chapitre 3, nous testerons l’influence du degré d’intégration phénotypique sur les capacités plastiques de ces espèces en conditions contrôlées. De plus nous déterminerons la relation entre intégration et performance, sur le terrain
(Chapitre 1) et en conditions contrôlées (Chapitre 3).
23
24
Chapitre 1. Relation entre intégration phénotypique, variabilité des traits intra-population et performance des plantes dans des environnements naturels
A. High phenotypic integration is correlated with high trait variation in both natural and controlled conditions, in sub-Antarctic plant populations (American Journal of Botany, invited for resubmission)
B. Exploring performance of species from the Iles Kerguelen: relationships to environment and phenotypic integration
25
Chapitre 1A. Integration increases trait variation
Chapitre 1A. High phenotypic integration is correlated with high trait variation in both natural and controlled conditions, in sub-Antarctic plant populations
26
Chapitre 1A. Integration increases trait variation
RESUME :
La variation des traits existante à l’intérieur des populations contribue au potentiel adaptatif des espèces. La variation d’un trait à l’intérieur d’une population, peut être contrainte par des facteurs environnementaux stressants, et modulée par le degré d’intégration phénotypique, i.e. degré de corrélation entre les traits. Une forte intégration phénotypique peut contraindre la variation des traits (car les traits ne peuvent pas varier indépendamment) ou la favoriser (car la valeur extrême d’un trait est cohérente avec les valeurs extrêmes des autres traits). On fait l’hypothèse que la variation des traits intra-population (i) diminue avec les stress abiotiques, (ii) peut augmenter ou diminuer en fonction du degré d’intégration phénotypique. Les espèces indigènes des îles Kerguelen arborent un fort degré d’intégration phénotypique. Nous étudions des populations naturelles de quatre espèces végétales indigènes des îles Kerguelen, le long de gradients environnementaux. De plus, pour distinguer l’influence de l’environnement de l’influence de l’intégration sur la variation des traits, une expérience en conditions contrôlées a également été menée. Nous avons montré que les facteurs environnementaux contrôlent la moyenne mais non la variation des traits (ou leur intégration) à l’intérieur des populations. A l’inverse, l’intégration phénotypique ne diminue pas, mais peut favoriser la variation des traits. Ce résultat est montré en conditions naturelles et en conditions contrôlées au niveau des traits végétatifs et floraux. Ainsi la variation des traits entre les organismes dépend de l’intégration intrinsèque des individus plutôt que de l’environnement. De plus, contrairement à ce qui est communément suggéré, l’intégration phénotypique favorise la variation des traits. De ce fait, l’intégration phénotypique contribue positivement au potentiel adaptatif des espèces.
ABSTRACT
Traits express the response of plants to their environment and trait variation within a population contributes to its capacity to cope with environmental changes. Variation of a given trait within a population, in turn, might be constrained by particularly stressful environments or might be controlled by high phenotypic integration, i.e. strong correlation among traits. High phenotypic integration might both decrease trait variation (because traits cannot vary independently) and increase trait variation (because extreme values of one trait are consistent with extreme values of other traits). Here we hypothesize that trait variation within populations (i) declines with abiotic stress, (ii) either declines or increases with the integration among traits. We studied four plant species, known to show a high degree of phenotypic integration, in natural populations sampled across various environmental gradients in sub-Antarctic Iles Kerguelen. Moreover, to disentangle environmentally induced from intrinsic phenotypic integration, we studied plants also under controlled conditions. We found that environmental gradients control means but not variation or integration of traits within populations. In contrast, trait variation often increased (and never declined) with phenotypic integration. This result is found both in situ and under controlled conditions and for both vegetative and floral traits. These results suggest that variation of traits among organisms may depend on intrinsic phenotypic integration within organisms, more than on particular environments. Thereby phenotypic integration may contribute to the evolutionary capacity of populations to respond to environmental changes. 27
Chapitre 1A. Integration increases trait variation
High phenotypic integration is correlated with high trait variation
in both natural and controlled conditions, in sub-Antarctic plant
populations1
Bastien Labarrere2*, Andreas Prinzing2, Richard C. Winkworth3 & Françoise Hennion2
American Journal of Botany, invited for resubmission
2 UMR 6553 Ecobio, Université de Rennes 1, CNRS, Av du Général Leclerc, F-35042 Rennes,
France, [email protected], *corresponding author, fax : (+33) 2 23 23 50 26
3 Institute of Fundamental Sciences, Massey University, PO Box 11222, Palmerston North,
New Zealand.
Keywords : phenotypic integration ; trait variation ; plant species ; environmental gradients ; soil pH; soil water saturation; soil conductivity; sub-Antarctic; Iles Kerguelen
Phenotypic integration enhances trait variation in plants
Acknowledgements
28 Chapitre 1A. Integration increases trait variation
The authors thank participants in the 2011-2012 summer campaign in Kerguelen : Philippe
Choler (LECA, Grenoble, France) and field volunteers Marine Pouvreau, Françoise Cardou,
Julie Vingère and Marion Lombard; we thank Réserve Naturelle TAAF and personnel of Center
ECOLEX, Rennes (Valérie Gouesbet, Jean-Luc Foulon, Thierry Fontaine and Fouad Nassur).
Field work was supported by IPEV (programs 1116 PlantEvol, F. Hennion and 136 Ecobio, M.
Lebouvier). French-New Zealand collaboration was supported by CNRS InEE (PICS
“AntarctBiodiv”, F. H.) and University Rennes 1 (Direction de la Recherche et de l’Innovation,
Service des Affaires Internationales). This research is linked to CNRS Zone-Atelier Antarctique and SCAR programmes AntEco and AnT-ERA. B.L. was supported by a PhD grant from
Ministry of Research and Education (France). Also, we would like to thank Editors and
Reviewers whose corrections helped us to improve our article.
29 Chapitre 1A. Integration increases trait variation
INTRODUCTION
Traits are at the interface between an organism and its environment and it is well established that, within species, some trait values may vary between populations to fit environmental variation between locations (Bell, 1987; Van Valen, 1973; Schlichting, 1986). Trait variation that occurs within populations provides the raw materials for natural selection (Bolnick et al.,
2011). High trait variation within populations increases the probability that a particular trait value is available in the population. Albeit some traits may have a single optimal value across environments, it is generally accepted that a high availability of trait values facilitates establishing a trait optimum matching the local environment, and trait variation increases the capacity of populations to respond to local selection, e.g. due to new climatic conditions
(Nicotra et al., 2010). Trait variation within populations may result from genetic (Kanaga et al.,
2008) or environmental factors (Debat and David, 2001; Bolnick et al., 2011). Hence, low genetic diversity (Mitchell-Olds and Schmitt, 2006), low plasticity and environmental stresses
(Gicquiaud et al., 2002) may limit trait variation in populations. Stressful environmental conditions may favor one optimal trait state, resulting in a decrease in trait variation within populations.
Values of the different traits of a plant are not independent from each other (Murren et al.,
2002) and internal interactions among traits may also affect trait variation. Internal interactions among traits are known as phenotypic integration, basically resulting from the need for different traits to be mutually consistent across an organism. Phenotypic integration is formally defined as the number and strength of correlations among traits within an individual (Pigliucci et al.,
2003; Murren, 2012; Armbruster et al., 2014) and may result from genetic constraints or plastic responses to the environment (Murren et al., 2002; Armbruster et al., 2014). Conceptual models describe links between genetic correlations and trait correlations (Pigliucci and Kaplan, 2010) and genomic studies identified pleiotropy or linkage disequilibrium as possible causes of
30 Chapitre 1A. Integration increases trait variation integration (Conner, 2002; Murren, 2012). Specifically, pleiotropic genes influence the variation of multiple traits, and genetic correlations caused by linkage disequilibrium may induce trait correlations.
Integration has mainly been studied in the interaction between floral traits and pollinators
(Murren, 2012). Integration may involve both vegetative and floral traits (Armbruster et al.,
2014; Murren, 2012) and may be dissociated in different functional modules in plants (Diggle,
2014; Murren, 2012). Modularity increases the possibility of trait variation, allowing traits that are not functionally related to vary independently but maintaining consistency among functionally related traits (Diggle, 2014). Traits involved in the same functions are then more integrated than traits involved in different functions. Different trait modules within plants may respond to different environments without mutually interfering. Integration among traits may control interactions between an organism and its environment, possibly no less than do mean values of traits (Jones et al., 2013). For instance, studying both vegetative and reproductive traits in Ulex europaeus, Hornoy et al. (2011) showed that plants differ in integration rather than trait means between native and invaded regions, integration being lower in the invaded region. The overall influence of environment on integration and the influence of integration on trait variation remain to be analyzed (Murren, 2012).
Studying the relation between phenotypic integration and environment, one study showed that integration may increase with increasing abiotic stress in the laboratory (see Gianoli and
Palacio-Lopez, 2009). Modifications of phenotypic integration induced by abiotic stress suggest that integration is driven by physiological or environmental constraints rather than being genetically based. Such experiments usually involve a single abiotic stress imposed on individuals that otherwise experience non-stressing conditions. Yet, plants in natural populations experience a very different situation as they are exposed to combinations of abiotic and biotic stresses and have often experienced broadly similar conditions over many
31 Chapitre 1A. Integration increases trait variation generations. These differences make it difficult to generalize the results obtained under laboratory experiments to plants in natural environments. Few authors have yet studied trait variation and integration in the field. Among these authors, Hermant et al. (2013) worked on trait variation and integration in a sample of plant species growing in sub-Antarctic Iles
Kerguelen. They showed in situ that phenotypic integration may vary among species, being highest in species endemic to the Iles Kerguelen. The study attributed one value of phenotypic integration by species, calculating the mean of integration among different populations of a species. Yet, individuals among populations of a given species probably experience more different environmental conditions than do individuals within populations. To determine how environmental factors influence the degree of integration within a species, it is necessary to compare the degree of integration within a population with the environment of that population.
Boucher et al. (2013) showed differences in the degree of integration among populations of one species, somewhat related to the environment. Multi-species, multi-gradient analyses of the relationship between integration and the environment still appear to be missing.
Conceivably the maintenance of organism consistency and hence phenotypic integration may have important outcomes in terms of observed trait variation, albeit this relationship has not been studied so far. In theory, trait variation may be either high or low, independent of the degree of phenotypic integration (Figure 1). We hypothesize that, on one hand, a high degree of phenotypic integration across the individuals of a population may be related to a low variation of each of the traits because integration prevents traits from varying independently. In such case there would be a trade-off between trait variation within populations and integration. On the other hand, we may hypothesize that a high degree of integration may be associated with a high trait variation considering that integration ensures that an extreme value of one trait is consistent with the extreme values of other traits and facilitates variation of each trait. Organisms with extreme values of one trait are less likely to be eliminated by selection if these values are
32 Chapitre 1A. Integration increases trait variation consistent with the values of other traits. Testing these hypotheses requires combining studies in nature and in controlled conditions because an observed relationship between trait integration and trait variation may simply result from local microenvironmental variation within and among natural populations (Schlichting, 1989). To our knowledge, no tests of the relationship between trait variation and integration exist so far. Determining whether phenotypic integration facilitates or constrains trait variation will help to understand the influence of phenotypic integration on the capacity of populations to cope with environmental changes.
Beyond pairwise interactions among environment, trait variation and integration, we aim to understand how environment, trait variation and integration influence each other simultaneously and thus determine both direct and indirect interactions. We may expect four extreme scenarios summarized in Figure 2. Environment may primarily drive integration which itself affects variation, so that the effect of environment on trait variation is mainly indirect.
Alternatively, we may expect that the environment primarily drives trait variation which itself affects integration, so that the effect of environment on integration is mainly indirect. We also may expect that environment drives both trait variation and integration without any causal relationship between the latter two. Finally, we may expect that the environment affects neither trait variation nor integration and that the latter two influence each other, that is to say that trait variation is mainly determined by intrinsic interactions among traits (i.e. phenotypic integration).
This study will use the Iles Kerguelen system as they are of great interest to examine relationships among trait variation, integration and the abiotic environment. Located in the southern Indian Ocean the Iles Kerguelen are characterized by a cold and oceanic climate
(Lebouvier and Frenot, 2007). They harbor a wide range of biotopes and environmental conditions (Wagstaff and Hennion, 2007) and a small native flora (22 species, Frenot et al.,
2001) and thus species-poor plant communities (Pansu et al., 2015). Furthermore, a previous
33 Chapitre 1A. Integration increases trait variation study showed that plant species endemic to the Iles Kerguelen exhibit high phenotypic integration (Hermant et al., 2013) and are therefore potentially good models for our study.
Finally, determining the significance of such high phenotypic integration for trait variation may help to better understand to which degree Kerguelen plants are able to cope with current and future climate change.
In the field, we measured phenotypes and characterized the environments occupied by populations of four plant species native to Iles Kerguelen. In parallel we cultivated plants under controlled conditions to test if correlations found in nature are also observed in homogenous and constant conditions. We investigated three species of Ranunculus (i.e., R. biternatus, R. pseudotrullifolius and R. moseleyi) and the Brassicaceae Pringlea antiscorbutica growing across gradients in soil conditions. For P. antiscorbutica we also considered altitude. We asked
(i) do phenotypic integration or trait variation within populations change with local abiotic conditions, and (ii) does phenotypic integration relate to trait variation and is the relationship positive or negative?
34 Chapitre 1A. Integration increases trait variation
MATERIALS AND METHODS
1. Species under study-
The Iles Kerguelen (49°20’00” S, 69°20’00” E) are situated in the Southern Indian Ocean within the sub-Antarctic region (Lebouvier and Frenot, 2007; Figure 3). We studied the three
Ranunculus species (i.e., R. biternatus, R. pseudotrullifolius and R. moseleyi), of which R. biternatus is austral circumpolar, R. pseudotrullifolius magellanic and on Kerguelen, and R. moseleyi is a strict Kerguelen endemic (Lehnebach et al., 2017) and Pringlea antiscorbutica a native Brassicaceae endemic to the South Indian Ocean Province (eg. Van der Putten et al.,
2010). All are perennial plants. On the Iles Kerguelen these species occupy different habitats
(Hennion and Walton, 1997a). Ranunculus biternatus is widespread on the island occurring in habitats below 500m above sea level (Hennion and Walton, 1997a; Réserve Naturelle des
Terres Australes et Antarctiques françaises and IPEV Programme 136, unpublished data). In contrast, R. moseleyi and R. pseudotrullifolius have more restricted distributions. The latter, being halophilous, occurs within a short distance of the coast, occupying peaty or sandy shorelines and pond margins (Hennion and Walton, 1997a), and sporadically in ponds up to about 200m a.s.l. (Réserve Naturelle des Terres Australes et Antarctiques françaises and IPEV
Programme 136, unpublished data). Ranunculus moseleyi is strictly aquatic, growing only in freshwater lakes and ponds (Hennion and Walton, 1997a). On the Iles Kerguelen P. antiscorbutica, an endemic species from the Southern Indian Ocean, occurs in habitats ranging from coastal meadows to montane fell-fields (Hennion and Bouchereau, 1998).
2. Collection of data at field sites-
2.1. Plant measurements
35 Chapitre 1A. Integration increases trait variation
Plants were sampled from December 2011 to February 2012. Ranunculus species were sampled in two regions in Iles Kerguelen: Isthme Bas, a large flat isthmus and Ile Australia, a large island (Figure 3). Populations of Ranunculus species were sampled across soil abiotic gradients such as soil water saturation, pH and conductivity. Sampling was designed to maximize the range of abiotic conditions for each species so that environmental signals in traits can be expected. We defined ‘population’ as a continuous group of plants living in a same site and subject to similar environmental conditions. Twelve populations were sampled for each
Ranunculus species (R. biternatus, R. pseudotrullifolius and R. moseleyi). For P. antiscorbutica,
9 populations were sampled across altitudinal gradients in three regions across Iles Kerguelen:
Ile Australia, Ile Guillou and Mont Crozier (Figure 3). These regions span different elevations across Iles Kerguelen, and hence our sampling design allowed us to study populations from different altitudes within and among regions. We covered a wide part of the range of altitudinal gradient in Iles Kerguelen studying populations from 0 to 585 meters above sea level.
In each population we selected 10 individuals and measured plant height, the length and width of the largest leaf, the number of leaves and flowers, the flowering stage and the size of the largest flower or inflorescence (largest diameter for Ranunculus, length of the club for P. antiscorbutica). Flowering stage of the individuals was estimated following Hennion et al.
(2012). The traits listed above are likely all related to plant size. We hence used two traits that we expect to be independent of plant size: (i) leaf shape (leaf width / leaf length). (ii) leaf dry matter content (LDMC), calculated for 20 leaves were sampled from a minimum of 15 individuals. Leaves were collected then directly put in distilled water in airtight bags for rehydration. They were then weighed before and after 48 hour-drying at 80°C following
Cornelissen et al. (2003). We found that some traits (flowering stage, number of flowers and the size of largest flower or inflorescence) were highly redundant. We thus only kept the continuously varying trait ‘size of largest flower or inflorescence’ as floral trait. Likewise, we
36 Chapitre 1A. Integration increases trait variation found low variability of LDMC within and among populations; hence we did not use this trait in our analysis.
2.2. Environmental measurements
In each population we measured soil water saturation, pH and conductivity. Three samples of soil, each of 20mL, were collected at the rhizosphere level of the measured plants. To determine soil water saturation, half of each soil sample was dried at 105°C for 48 hours and weighed before and after drying (Hermant et al., 2013). The other half of each soil sample was mixed with a known volume of distilled water and then was left for 18 to 24 hours to permit sedimentation of soil particles. Soon after, pH was determined using a pH meter (BASIC 20
PLUS CRISON, resolution 0.01pH). After another 18 to 24 hours of sedimentation, conductivity was determined using a conductivity meter (CONSORT K810, resolution 0.1 µS cm-1) (Hermant et al., 2013). Temperature was determined using loggers installed at soil surface close to plants in six P. antiscorbutica populations, that recorded every 30 minutes from January to February, i.e. a period of high plant development (Aubert et al., 1999).
We used multiple species that were sampled in a roughly similar range of environments to explore consistency of observed relationships across species (Appendix 1). Precisely, R. biternatus was found in a range of soil water saturation from 20% to 99% whereas R. pseudotrullifolius and R. moseleyi were found in soil water saturation from respectively 37% to
97% and from 37% to 99%, correspondingly to R. biternatus being known as the most terrestrial species (Hennion et al., 1994). P. antiscorbutica was found in conditions of soil water saturation from 10% to 66%. R. pseudotrullifolius was found in conditions with highest mean conductivity
(2155.1 mS.cm-1, Appendix 1), which was consistent with its halophilous status (Hennion and
Walton, 1997a), and the populations sampled for P. antiscorbutica had conditions with lower mean conductivity (448.31 mS.cm-1, Appendix 1). The four species occurred in a similar range
37 Chapitre 1A. Integration increases trait variation of pH and variation of pH found within species was low (Appendix 1). The environmental distributions of species we observed are consistent with the known ecology of species (Hennion and Walton, 1997a) and our results now provide precise values and ranges of soil conditions.
The widespread R. biternatus was found in the largest range of environmental conditions, moreover, R. biternatus and P. antiscorbutica were found in drier conditions than R. pseudotrullifolius and R. moseleyi.
The Ranunculus species are amphibious. Ranunculus populations were hence sampled in terrestrial and aquatic environments. We note that aquatic stands showed higher means for conductivity and obviously soil water saturation, but did not differ from terrestrial stands in either pH or variability of environmental factors (soil water saturation, pH, conductivity;
Appendix 2).
3. Controlled conditions-
We used controlled conditions to verify that the relationship between integration and trait variation was not due to covariation of environmental conditions in the field. To grow plants in controlled conditions, we used seeds from different natural populations in P. antiscorbutica, R. biternatus and R. pseudotrullifolius. No seeds were available for R. moseleyi. Seeds from P. antiscorbutica were collected in the field in 2013 in three regions: Ile Australia and Mont
Crozier in the same populations where the phenotypic measurements had been performed in situ, and Ile Mayes (Figure 3). Ten individuals were randomly selected within each population.
In each of the ten individuals, fifteen siliques were then randomly collected from the median part of the longest infructescence where seed size is maximum and least variable (Hennion,
Schermann and Atlan, unpublished). We also collected seeds from six P. antiscorbutica plants growing in a phytotron in Rennes in 2013. A total of 38 populations were used in P. antiscorbutica. In R. biternatus and R. pseudotrullifolius, seeds were collected in 2014 in
38 Chapitre 1A. Integration increases trait variation respectively 13 and 8 populations in Isthme Bas, in the same populations where the phenotypic measurements had been performed in situ. All seeds were stored dry with silica gel at 4°C until their use according to established protocols (e.g. Hennion and Walton, 1997a).
Seeds were sterilized and germinated in Petri dishes at 25°C for P. antiscorbutica and 21°C for Ranunculus species, under low light (1kLux) following established protocols (Hennion and
Walton, 1997a, 1997b). To get homogeneous design we kept a constant number of seedlings from each population. In each species, three hundred seedlings were planted in vermiculite substrate and fertilized with ½ Hoagland solution (Hoagland and Arnon, 1938). For the R. pseudotrullifolius, NaCl was added to Hoagland solution to reach a concentration of 1g.L-1, being consistent with salinity found in the field. Plants were grown hydroponically in a phytotron in Rennes under conditions (light exposure, photoperiod, temperature and air humidity) mimicking austral summer in Iles Kerguelen at best, according to previous protocols
(see Hennion et al., 2006). Plants were measured after 5 months of growth, traits considered were height and diameter of the rosette, and in the 2 longest leaves: length and width of lamina and length of petiole. Note that we found trait differences among plants from different field populations, despite the fact that plants were grown in the same conditions.
4. Data analyses-
4.1. Mean and variance of environment and traits
We calculated variance of traits using the coefficient of variation (CV), i.e. Standard deviation
(SD)*mean-1 (Lovie, 2005). We chose to use CV rather than SD as we found a highly significant positive relationship between SDs and mean of the traits that is to say that analyzing SD was similar to analyzing trait mean. As most traits were related to plant size, calculating CVs also corrected for possible effects of plant size. In R. pseudotrullifolius, we yet found that SD of leaf length and leaf width was not correlated to the corresponding trait mean (p=0.16; adj. r²=0.11
39 Chapitre 1A. Integration increases trait variation and p=0.19; adj. r²=0.078) and dividing SD by the mean might theoretically introduce a bias.
We hence conducted an additional analysis using instead of CVs, SDs with trait means in covariable. We found results to be similar to an analysis of CVs and chose to present only results using CV to be consistent with results on other traits and species.
The estimate of variance within a population may not be certain and robust when based on a small sample, inclusion or exclusion of a single or few individuals might change the results.
This poses a problem if uncertainty in the estimation of variance within populations is as large as or larger than differences of variances among populations. We hence explored the importance of this problem in two ways.
First, we explored robustness of our estimates of within-population variances by bootstrap resampling: resampling a sample 1000 times with replacement, and recalculating variances for each resample. For a given trait and a given species, we calculated for each population the SD across the bootstrap resamples – providing a measure for the uncertainty of the estimation of within-population variance. We found that uncertainties of estimation of within population variances (i.e. SDs across bootstrap resamples) were smaller than the variation of SDs among populations (Appendix 4). This indicates that uncertainty in the estimation of variances within populations will not hide major across-populations patterns, despite the relatively small sample size per population. We also used these bootstrap resamples to redo the analyses we done based on the initial samples, i.e. relating variances within populations to integrations of traits within populations. We even aimed at testing the significance of these relationships, which makes little sense for the most extreme bootstrap resamples. We hence randomly selected 30 resamples yielding variances within +/-1 SD of the mean. For each of these resamples we tested relationships between trait variances and trait integrations. We found that the results of these analyses of bootstrapped samples confirmed those of the initial samples: the significant relationship remained significant in 100%, and the two non-significant relationships remained
40 Chapitre 1A. Integration increases trait variation non-significant in 77% and 80% of the bootstrap resamples. This result indicates that initial results are robust to inclusion or exclusion of individuals, despite the limited size per sample.
In a second approach to explore the robustness of our results despite small sample size per population we increased the number of individuals per population by merging pairs of populations. Our sampling design was composed of pairs of nearby populations and we hence merged each pair of two nearby populations of 10 individuals into one of 20. We recalculated variances and CVs across these 20 individuals and again tested relationships between CVs and integrations. The observed relationships were very similar to those found in the initial analyses across more, smaller populations (Appendix 5).
At a more general level, small samples will, if anything, result in random errors of population estimates, possibly hiding existing patterns, but are very unlikely to generate patterns. In short, small samples reduce power to detect patterns. However, we found multiple significant correlations between trait variation and integration, each with high r². Overall, in our study system, sampling 10 individuals per population permitted robust estimations of intra-population variance, and captured differences among populations.
4.2. Phenotypic integration
Phenotypic integration was calculated as the percentage of variance explained by the first axis of a Principal Component Analysis (PCA) performed on traits described above (Hermant et al., in 2013). PCA was performed using a correlation matrix. We also calculated integration using the average correlation among pairs of traits (García-Verdugo et al., 2009). We found similar results with both methods and we chose to proceed the analysis with results based on PCA as this method tends to be easier to interpret going from 0 to 100 percent. Trait values were scaled and PCA conducted using the FactoMineR package of R 3.0.2 (Lê et al., 2008; R Core Team,
2012) In the total absence of trait variation among individuals, phenotypic integration is not
41 Chapitre 1A. Integration increases trait variation defined (Figure 1). In our sampling there is no population that shows no trait variation among individuals, and we hence could identify phenotypic integration for each population.
To explore whether phenotypic integration varied depending on the traits incorporated in its calculation, we calculated different integrations that include different traits, including traits representing the same or different plant modules, or traits related or unrelated to plant size.
Specifically, we calculated integrations based on: only traits representing the vegetative module; traits representing both the vegetative and the floral modules; only traits related to plant size; and traits both related and unrelated to plant size (leaf shape). Traits included in the different estimations of integration are presented in Appendix 3. We, first, found strong positive relationships between the different integrations across populations (r² = 0.4 to r² = 0.77, on average = 0.58, N=36, using linear regressions). Moreover, we found similar relationships between (i) integrations calculated from different sets of traits and (ii) environmental factors or trait variation (Appendix 3). Hence, module identity of traits and their dependence or independence from body size do not seem to strongly affect our measures of integration or how integration relates to environment or variance. We chose to use integration based on plant height, leaf length, leaf width and flower size, which is representative of the overall degree of integration of the plant within and across modules.
In the laboratory, we used a somewhat different set of traits to calculate integration than in the field. Specifically, we did not account for floral traits as due to the short duration of the experiment, plants did not flower. Instead, we used height of the plant and length and width of lamina and length of petiole of the 2 longest leaves. Note that even if traits used in the calculation of integration are different between the field and the laboratory, we found similar relationships between trait variation and integration (Figure 4, 5 & 6).
42 Chapitre 1A. Integration increases trait variation
We admit that existing measures of integration assume that integration results from continuous relationships among quantitative traits, and do not capture discontinuous relationships such as canalization of characters into a few distinct combinations.
4.3. Hypothesis Testing
The relationships between trait means, CVs or integration and environmental conditions in the field were tested using simple linear regression. We tested the relationship between integration and CVs independently in the field and in controlled conditions using the same method.
Analyses were conducted in R 3.0.2 (R Core Team, 2012). We verified our results with multiple regression analyses simultaneously accounting for the three environmental factors and found similar results. We present only simple regression analyses as they are more robust given higher degrees of freedom. For multiple testing on the same data set, p-values were corrected using sequential Bonferroni’s correction (Cabin and Mitchell, 2000). In all regression analyses we verified the assumptions of the analyses using residual plots. In P. antiscorbutica, populations were sampled across different regions characterized by contrasted environmental conditions.
Phenotypic differences among regions appear to not be entirely explained by differences in measured environmental conditions and might result from differentiation among regions. Thus, the region of the sample was then taken into account when relating traits to the environment
(Table 1, 2 & 3). We avoided comparing ratios between traits and integration as ratios reflect to some degree phenotypic integration and comparisons of such ratios to integration would be circular. To analyze indirect relationships between environment, integration and trait variation we conducted path analysis using piecewiseSEM package (Shipley, 2013) in R 3.0.2 (R Core
Team, 2012). We chose to perform an integrative model using a single, integrative measure of trait variation instead of performing various models containing the variation of each different trait. We carried out a Principal Component Analysis based on the variation of traits (plant size,
43 Chapitre 1A. Integration increases trait variation leaf length, leaf width and flower size) of populations and used the coordinates of a population on the first axis of the PCA as an integrative measure of trait variation in that population.
4.4. Random model of phenotypic integration and trait variation
We used a null model to investigate whether purely numerical and not biological effects might also result in a correlation between phenotypic integration and trait variation. We used 1000 randomly simulated populations each containing 10 individuals characterized by four traits with values assigned at random within a range from 0 to 10. We calculated phenotypic integration and trait variation for each of these simulated populations as explained above. Different populations showed different degrees of integration or trait variation, but we found no significant relation between integration and trait variation (Table 4). This same model also provided information on the mean degree of integration in random populations where traits are not biologically correlated. Note that in the above null model we preferred to create random variables over randomly assigning empirical data as this is more conservative. We also created a null model with populations that showed either low or high trait variances and then calculated integrations for populations with either low or high variances. We found that integration did not differ significantly between populations with high or low trait variation (Table 5).
44 Chapitre 1A. Integration increases trait variation
RESULTS
1. Degree of phenotypic integration-
All our measured traits varied among individuals within all populations. Therefore, we could estimate phenotypic integration for every population (Figure 1). Estimates of phenotypic integration for field populations of R. biternatus, R. pseudotrullifolius and R. moseleyi were, respectively, 68.51% (CV 0.21; SD 14.07), 65.39% (CV 0.21; SD 13.56) and 66.69% (CV 0.17;
SD 11.13). In P. antiscorbutica in the field, mean degree of integration was 59.07% (CV 0.14;
SD 8.26). In P. antiscorbutica, R. biternatus and R. pseudotrullifolius grown in controlled chambers the average integration was respectively 52.00% (CV 0.22; SD 11.52), 58.03% (CV
0.23; SD 13.09), 66.72% (CV 0.21; SD 13.70). In our randomized model, the degree of integration was 44.40% (CV 0.14; SD 6.35). We found that all estimated phenotypic integration values were significantly higher than in the random model (p<0.001) for all studied species.
2. Theoretical relationship between phenotypic integration and trait variation-
Our random model revealed no significant relationship between integration and trait variation
(Table 4). Nor did integration differ between populations with high versus low trait variation
(Table 5).
3. Relationships with environment- 3.1 Trait mean
Table 1 summarises the relationship between plant traits and environmental variation. Briefly, plants in R. biternatus populations growing at sites with higher soil water saturation and pH, had longer and broader leaves. Those at higher soil conductivity also had broader leaves and were taller. For R. pseudotrullifolius, plants in populations growing at sites with higher soil water saturation were taller with longer and broader leaves while those at sites with higher soil
45 Chapitre 1A. Integration increases trait variation conductivity had broader leaves. In R. moseleyi, we found no significant trait variation with varying soil conditions. In P. antiscorbutica, plants growing at sites with higher soil conductivity were shorter, with shorter leavesand smaller flowers. We found no relationship between any trait and either altitude or temperature. We found no relationship between leaf shape ratio and anyenvironmental variable in any of the four species.
3.2. Trait variation
We found no significant relationship between trait variation within populations and soil conditions in any of the four species (Table 2). Overall, we found that all non-significant relationships are far from being significant, with low adjusted r² values (Table 2). Low adjusted r² values indicate that the absence of relation between trait variation and environment is not due to a lack of statistical power.
3.3. Phenotypic integration
We found no significant relationship between phenotypic integration and any of the environmental variables tested for any of the species, including temperature variation (Table
3). Overall, we found that all non-significant relationships are far from being significant with low adjusted r² values (Table 3). Even with a larger sample size these relationships would not have been ranked significant.
4. Relationship between trait variation and phenotypic integration-
In the field, a positive relationship was found between variation in leaf length and width with phenotypic integration in R. pseudotrullifolius and variation in leaf length and flower size with phenotypic integration in P. antiscorbutica (Figure 4 and 5). No significant relationship was found in R. biternatus and R. moseleyi (Figure 4).
46 Chapitre 1A. Integration increases trait variation
Figure 6 summarises the covariation between trait variation and integration for plants grown in controlled chambers. For all three tested species, we found cases of positive relationships between trait variation and integration. In P. antiscorbutica, length of the longest leaf, length and width of 2nd longest leaf increase significantly with phenotypic integration (Figure 6). In R. pseudotrullifolius, lengths of the two longest leaves and width of the longest leaf increase significantly with phenotypic integration (Figure 6). In R. biternatus, variation of leaf width increases with phenotypic integration (Figure 6).
Note that the null model indicates no relationship whatsoever between integration and trait variation (adj r² < 0.001; Table 4).
5. Path analysis between environment, integration and trait variation in the field-
We used path analysis to test whether environment has an indirect effect on trait variation through a direct effect on integration which in turn affects variation. Conversely we also tested whether environment has an indirect effect on integration through a direct effect on trait variation itself affecting integration. We found no significant indirect effects of environment but we did find a significant and positive direct relationship between integration and trait variation in R. pseudotrullifolius (p= 0.006). We did not find any significant relationship between integration and our integrated measure of trait variation in R. biternatus, R. moseleyi and P. antiscorbutica.
47 Chapitre 1A. Integration increases trait variation
DISCUSSION
Natural selection works on trait variation within species. However, different traits often cannot vary independently but are mutually phenotypically integrated. Phenotypic integration reflects the degree to which different traits and different parts of organisms vary consistently. In this study we analyzed natural populations living in different environments. We aimed to determine the influence of the environment on trait variation and integration as well as the relation between trait variation and phenotypic integration. We showed that some trait means are related to the environment, while this was not the case for trait variation or integration among traits.
Furthermore, we found positive or null relationships between integration and trait variation, rather than a negative relationship that would be expected if integration constrained trait variation as commonly supposed (Murren et al., 2002). We found this relationship both in the field and in a much more homogeneous environment in the lab.
1. Environment and how it relates to trait means and variation-
Here we showed that trait means related to plant size vary with environmental conditions in R. biternatus, R. pseudotrullifolius and P. antiscorbutica. Specifically, plants were smaller (plant height and leaf size) in drier conditions (R. biternatus and R. pseudotrullifolius), or in acid or saline conditions (R. biternatus and P. antiscorbutica). Small mean trait values in populations may result from selection - plants with small trait values being more performant - or result from a limitation of plant size due to environmental stresses. We suggest that, here, small trait values may indicate conditions that induce stresses in plants (Grime, 2006), either salt (Zhu, 2001) or drought (Farooq et al., 2009) stresses. In the controlled environment, we also found trait differences among plants from different field populations. These differences cannot result from different resource condition as plants were grown under the same conditions. Differences of trait means among populations may either result from different adaptations in different
48 Chapitre 1A. Integration increases trait variation populations or from different maternal effects, resulting in different resource allocations.
Environment may influence trait variation within populations (Debat and David, 2001).
Stressful environmental conditions may favor one optimal trait state, resulting in a decrease in trait variation within populations. Indeed, several authors found that stressful conditions could hinder trait variation (Gicquiaud et al. 2002). We, in contrast, did not show any relationships between trait variation within populations and mean environmental values such as soil water saturation, pH and conductivity in two of four species (R. biternatus and R. pseudotrullifolius), even if these environmental factors have an impact on trait means. In the two remaining species
P. antiscorbutica and R. moseleyi we found, respectively, only one and two trait variation- environment relationships, respectively, being only marginally significant. These results are consistent with a study on Ranunculus species, Cook and Johnson (1968) which found no significant relation between trait variation and environment despite the fact that Ranunculus trait means are known to be highly dependent on the environment (Garbey et al, 2004; Van
Kleunen et al, 2005).
2. Phenotypic integration and environment-
We showed distinct differences in the degree of phenotypic integration among the different natural populations. Integration is hence not a fixed value in species. Phenotypic integration may result from genetic correlations (Conner, 2002), but also from developmental constraints
(Murren et al. 2002). Phenotypic analyses of integration take into account both genetic correlations and developmental constraints, and hence tend to be more powerful than genetic analyses of integration (Conner, 2002). Environment may also have an influence on phenotypic integration (Armbruster et al., 2014). Several authors have shown in the laboratory that integration may vary in response to the environment (Callahan and Waller, 2000; Pigliucci and
Hayden, 2001). Yet, in a field study, Murren et al (2002) found that integration does not vary
49 Chapitre 1A. Integration increases trait variation across environments. Here also, looking at degree of integration in four species across a range of three environmental factors in the field, we found no variation in integration associated with environmental variation. This difference between field and laboratory studies may reflect differences in either biological or temporal scales. In the laboratory, studies use sudden and short-term exposure of individuals to an extreme environment and individuals may then become less integrated. Conversely, in the field, we analyzed established populations exposed for long to a given environmental constraint and past selection may have maintained only organisms with higher level of integration reflecting a higher organism consistency (Callahan and Waller,
2000). This would explain the lack of a relationship between environmental factors and phenotypic integration in the field. Another possible explanation for differences in the integration - environment relationship between laboratory and field studies would be that a loss of integration in response to environmental stress in the lab reflects a lack of synchrony among the responses of different traits. This lack of synchrony would hence only be a temporary phenomenon. In contrast in the field, what we measure in these perennial plants is the final outcome after all traits have responded and integration has been reestablished.
Overall, environment did not explain differences of the degree of phenotypic integration found among populations in the field. Even the least severe environments within Iles Kerguelen do not lead to a reduction of the degree of integration. We might speculate that even under least stressful conditions realized on Iles Kerguelen, phenotypic integration remains useful.
Consequently, current changes in climate conditions on Iles Kerguelen might require breaking current phenotypic integration and establishing new patterns of integration. Whether this transitory loss of integration currently occurs should be investigated in the future.
3. Phenotypic integration and trait variation-
50 Chapitre 1A. Integration increases trait variation
In highly integrated organisms, traits must covary in a prescribed way, suggesting that phenotypic integration constrains trait variation. In fact, Gianoli and Palacio-Lopez (2009) found such a negative relationship between trait variation and phenotypic integration at the level of individuals in two perennial plant species in the lab. Such a negative relationship would imply that under current climate change populations might not be able to change mean values of traits while maintaining consistency among traits. Studying natural populations, we found no negative relationship between phenotypic integration and trait variation. Moreover, we showed that phenotypic integration may even increase with trait variation. Differences with results found by Gianoli and Palacio-Lopez (2009) may result from the fact that they observed short term response while we observe established trait variation which is potentially a longer- term response. Overall, in the field, high integration within populations does not mean limited variation of traits within populations and may not be considered as a constraint to trait change
(Murren, 2002). In contrast, in populations with high phenotypic integration, an extreme value of a given trait matches an extreme value of another trait, allowing high trait variance. Part of trait variation may result from plant plasticity. Ranunculus species are known to show high plasticity (Garbey et al, 2004), for example heterophylly, a form of adaptive plasticity (Wells and Pigliucci, 2000). Van Kleunen et al. (2000) showed in Ranunculus species that plasticity has costs in plants. For instance, the plastic change of a trait value may reduce the consistency among traits of the entire organism, rendering plasticity maladaptive. High phenotypic integration might prevent such maladaptive plasticity as traits cannot vary independently from each other.In the field, we found a positive relationship between integration and trait variation for R. pseudotrullifolius and P. antiscorbutica but not for R. biternatus and R. moseleyi. Note that the three Ranunculus species show vegetative reproduction while P. antiscorbutica only shows sexual reproduction. Hence relationship between integration and trait variation may not be due to a particular way of reproduction of species studied.In parallel with field investigations,
51 Chapitre 1A. Integration increases trait variation we conducted laboratory experiments, growing seeds from the field under controlled conditions. We found that although growing in the same conditions, populations with higher integration showed higher trait variation, being consistent with the positive relationship found in the field. The fact that positive integration-trait variation relationships were observed both, across populations in the field – being subject to different environments - and populations in the laboratory – being subject to the same environment - suggests that the integration-trait variation relationship is independent of the environment. Hence, we showed that the positive relationship found in the field was unlikely to result from local microenvironmental variation within and among populations (as proposed by some authors Schlichting, 1989; Armbruster,
2014). Results found in both the field and under controlled conditions demonstrate that the relation between environment, trait variation and integration corresponds to the scenario shown in Figure 2d and detailed in Figure 7, with phenotypic integration being the main factor influencing trait variation in populations. Phenotypic integration is still rarely considered in studies of trait variation that usually focus on the impact of the environment. Moreover, phenotypic integration is commonly suggested as a likely constraint on trait variation. However, we showed that phenotypic integration controls trait variation more than does the environment.
In fact, integration promotes trait variation within populations. Trait variation, in turn is known to increase the capacity of populations to face climate change (Bolnick et al., 2011).
CONCLUSION
Overall, we showed in both natural populations and in controlled conditions that there is no trade-off between capacity of traits to vary and their capacity to be mutually integrated in a highly consistent organism. Contrary to a common and intuitive assumption, phenotypic integration may even increase with trait variation, an extreme value of a given trait is then associated with an extreme value of another trait. In highly integrated organisms, traits may
52 Chapitre 1A. Integration increases trait variation share the same developmental pathway (Schlichting and Pigliucci, 1998) and this developmental organization may facilitate the simultaneous variation of different traits, helping to respond to combined stresses. Furthermore, we found that trait variation within populations only rarely related to environment, but is often and positively related to phenotypic integration.
Phenotypic integration, by increasing trait variation, may be a major actor that contributes to the capacity of species to respond to environmental changes. The observed interactions among environment, integration and trait variation may thus be summarized by scenario in Figure 7.
To date, only few studies are interested in the influence of phenotypic integration on trait variation and to our knowledge we are the first to show that integration may influence trait variation more than does the environment, and consistently so between the field and a controlled laboratory. A new step of investigation could be to determine if the degree of integration is fixed in populations or plastic and to determine the influence of integration on plant performance under conditions likely to be found in the next decades.
53 Chapitre 1A. Integration increases trait variation
REFERENCES Armbruster, W. S., Pélabon, C., Bolstad, G. H., & Hansen, T. F. (2014). Integrated phenotypes: understanding trait covariation in plants and animals.Philosophical Transactions of the Royal Society of London B: Biological Sciences, 369(1649), 20130245. Aubert, S., Assard, N., BOUTIN, J. P., Frenot, Y., & DORNE, A. J. (1999). Carbon metabolism in the subantarctic Kerguelen cabbage Pringlea antiscorbutica R. Br.: environmental controls over carbohydrates and proline contents and relation to phenology. Plant, Cell & Environment, 22(3), 243- 254. Bell, G., & Smith, J. M. (1987). Short-term selection for recombination among mutually antagonistic species. Nature, 328(6125), 66-68. Bolnick, D. I., Amarasekare, P., Araújo, M. S., Bürger, R., Levine, J. M., Novak, M., ... & Vasseur, D. A. (2011). Why intraspecific trait variation matters in community ecology. Trends in ecology & evolution, 26(4), 183-192. Boucher, F. C., Thuiller, W., Arnoldi, C., Albert, C. H., & Lavergne, S. (2013). Unravelling the architecture of functional variability in wild populations of Polygonum viviparum L. Functional ecology, 27(2), 382- 391. Cabin, R. J., & Mitchell, R. J. (2000). To Bonferroni or not to Bonferroni: when and how are the questions. Bulletin of the Ecological Society of America, 246-248. Callahan, H. S., & Waller, D. M. (2000). Phenotypic integration and the plasticity of integration in an amphicarpic annual. International journal of plant sciences, 161(1), 89-98. Conner, J. K. (2002). Genetic mechanisms of floral trait correlations in a natural population. Nature, 420(6914), 407-410. Cook, S. A., & Johnson, M. P. (1968). Adaptation to heterogeneous environments. I. Variation in heterophylly in Ranunculus flammula L. Evolution, 496-516. Cornelissen, J. H. C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D. E., ... & Poorter, H. (2003). A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian journal of Botany, 51(4), 335-380. Debat, V., & David, P. (2001). Mapping phenotypes: canalization, plasticity and developmental stability. Trends in Ecology & Evolution, 16(10), 555-561. Diggle, P. K. (2014). Modularity and intra-floral integration in metameric organisms: plants are more than the sum of their parts. Philosophical Transactions of the Royal Society of London B: Biological Sciences,369(1649), 20130253. Farooq, M., Wahid, A., Kobayashi, N., Fujita, D., & Basra, S. M. A. (2009). Plant drought stress: effects, mechanisms and management. Sustainable Agriculture, 153-188. Frenot, Y., Gloaguen, J. C., Masse, L., & Lebouvier, M. (2001). Human activities, ecosystem disturbance and plant invasions in subantarctic Crozet, Kerguelen and Amsterdam Islands. Biological conservation, 101(1), 33-50. Garbey, C., Thiébaut, G., & Muller, S. (2004). Morphological plasticity of a spreading aquatic macrophyte, Ranunculus peltatus, in response to environmental variables. Plant Ecology, 173(1), 125-137. García-Verdugo, C., Granado-Yela, C., Manrique, E., de Casas, R. R., & Balaguer, L. (2009). Phenotypic plasticity and integration across the canopy of Olea europaea subsp. guanchica (Oleaceae) in populations with different wind exposures. American Journal of Botany, 96(8), 1454-1461. Gianoli, E., & Palacio López, K. (2009). Phenotypic integration may constrain phenotypic plasticity in plants. Oikos, 118(12), 1924-1928. Gicquiaud, L., Hennion,‐ F., & Esnault, M. A. (2002). Physiological comparisons among four related Bromus species with varying ecological amplitude: polyamine and aromatic amine composition in response to salt spray and drought. Plant Biology, 4(6), 746-753. Grime, J. P. (2006). Plant strategies, vegetation processes, and ecosystem properties. John Wiley & Sons. Hennion, F., Fiasson, J. L., & Gluchoff-Fiasson, K. (1994). Morphological and phytochemical relationships between Ranunculus species from Iles Kerguelen. Biochemical systematics and ecology, 22(5), 533- 542. 54
Chapitre 1A. Integration increases trait variation
Hennion, F., & Walton, D. W. (1997a). Ecology and seed morphology of endemic species from Kerguelen phytogeographic zone. Polar Biology, 18(4), 229-235. Hennion, F., & Walton, D. W. (1997b). Seed germination of endemic species from Kerguelen phytogeographic zone. Polar Biology, 17(2), 180-187. Hennion, F., & Bouchereau, A. (1998). Accumulation of organic and inorganic solutes in the subantarctic cruciferous species Pringlea antiscorbutica in response to saline and cold stresses. Polar Biology, 20(4), 281-291. Hennion, F., Frenot, Y., & Martin Tanguy, J. (2006). High flexibility in growth and polyamine composition of the crucifer Pringlea antiscorbutica in relation to environmental conditions. Physiologia Plantarum, 127(2), 212-224.‐ Hermant, M., Prinzing, A., Vernon, P., Convey, P., & Hennion, F. (2013). Endemic species have highly integrated phenotypes, environmental distributions and phenotype–environment relationships. Journal of Biogeography, 40(8), 1583-1594. Hoagland, D. R., & Arnon, D. I. (1938). Growing plants without soil by the water-culture method. Circ. Calif. Agric. Exp. Stn. Hornoy, B., Tarayre, M., Hervé, M., Gigord, L., & Atlan, A. (2011). Invasive plants and enemy release: evolution of trait means and trait correlations in Ulex europaeus. PLoS One, 6(10), e26275. Jones, C. S., Martínez-Cabrera, H. I., Nicotra, A. B., Mocko, K., Marais, E. M., & Schlichting, C. D. (2013). Phylogenetic influences on leaf trait integration in Pelargonium (Geraniaceae): Convergence, divergence, and historical adaptation to a rapidly changing climate. American journal of botany, 100(7), 1306-1321. Kanaga, M. K., Ryel, R. J., Mock, K. E., & Pfrender, M. E. (2008). Quantitative-genetic variation in morphological and physiological traits within a quaking aspen (Populus tremuloides) population. Canadian Journal of Forest Research, 38(6), 1690-1694. Lê, S., Josse, J., & Husson, F. (2008). FactoMineR: an R package for multivariate analysis. Journal of statistical software, 25(1), 1-18. Lebouvier, M. & Frenot, Y. (2007). Conservation and management in the French sub-Antarctic islands and surrounding seas. In Papers and proceedings of the royal society of Tasmania (Vol. 141, No. 1, pp. 23- 28). Lebouvier, M., Laparie, M., Hulle, M., Marais, A., Cozic, Y., Lalouette, L., ... & Renault, D. (2011). The significance of the sub-Antarctic Kerguelen Islands for the assessment of the vulnerability of native communities to climate change, alien insect invasions and plant viruses. Biological Invasions, 13(5), 1195-1208. Lehnebach, C, Winkworth, R.C., Becker, M., Lockhart, P.J. & Hennion, F. Around the pole: evolution of sub- Antarctic island Ranunculus. Journal of Biogeography early view 15th February 2017, DOI: 10.1111/jbi.12952. Lovie, P. (2005). Coefficient of variation. Encyclopedia of statistics in behavioral science. Mitchell-Olds, T., & Schmitt, J. (2006). Genetic mechanisms and evolutionary significance of natural variation in Arabidopsis. Nature, 441(7096), 947-952. Murren, C. J., Pendleton, N., & Pigliucci, M. (2002). Evolution of phenotypic integration in Brassica (Brassicaceae). American Journal of Botany, 89(4), 655-663. Murren, C. J. (2012). The integrated phenotype. Integrative and comparative biology, 52(1), 64-76. Nicotra, A. B., Atkin, O. K., Bonser, S. P., Davidson, A. M., Finnegan, E. J., Mathesius, U., ... & van Kleunen, M. (2010). Plant phenotypic plasticity in a changing climate. Trends in plant science, 15(12), 684-692. Pansu, J., Winkworth, R. C., Hennion, F., Gielly, L., Taberlet, P., & Choler, P. (2015). Long-lasting modification of soil fungal diversity associated with the introduction of rabbits to a remote sub-Antarctic archipelago. Biology letters,11(9), 20150408. Pigliucci, M., & Hayden, K. (2001). Phenotypic plasticity is the major determinant of changes in phenotypic integration in Arabidopsis. New Phytologist, 152(3), 419-430. Pigliucci, M. (2003). Phenotypic integration: studying the ecology and evolution of complex phenotypes. Ecology Letters, 6(3), 265-272.
55
Chapitre 1A. Integration increases trait variation
Pigliucci, M., & Kaplan, J. (2010). Making sense of evolution: The conceptual foundations of evolutionary biology. University of Chicago Press. Schlichting, C. D. (1986). The evolution of phenotypic plasticity in plants. Annual review of ecology and systematics, 667-693. Schlichting, C. D. (1989). Phenotypic integration and environmental change. BioScience, 39(7), 460. Schlichting, C. D., & Pigliucci, M. (1998). Phenotypic evolution: a reaction norm perspective. Sinauer Associates Incorporated. Shipley, B. (2013). The AIC model selection method applied to path analytic models compared using a d- separation test. Ecology, 94(3), 560-564. Van der Putten, N., Verbruggen, C., Ochyra, R., Verleyen, E., & Frenot, Y. (2010). Subantarctic flowering plants: pre glacial survivors or post glacial immigrants?. Journal of biogeography, 37(3), 582-592. Van Kleunen. (2000). Clonal integration in Ranunculus reptans: by product or adaptation?. Journal of Evolutionary‐ Biology, 13(2), 237-248.‐ Van Kleunen, M., & Fischer, M. (2005). Constraints on the evolution ‐of adaptive phenotypic plasticity in plants. New Phytologist, 166(1), 49-60. Van Valen, L. (1973). A new evolutionary law, Evol, Theory 1: 1-30.. 1974. Two modes of evolution. Nature, 252, 298-299. Wagstaff, S. J., & Hennion, F. (2007). Evolution and biogeography of Lyallia and Hectorella (Portulacaceae), geographically isolated sisters from the Southern Hemisphere. Antarctic Science, 19(04), 417-426. Wells, C. L., & Pigliucci, M. (2000). Adaptive phenotypic plasticity: the case of heterophylly in aquatic plants. Perspectives in Plant Ecology, Evolution and Systematics, 3(1), 1-18. Zhu, J. K. (2001). Plant salt tolerance. Trends in plant science, 6(2), 66-7
56
Chapitre 1A. Integration increases trait variation
Table 1. Relation between trait means and abiotic conditions. Adjusted r squared and p-values after sequential Bonferroni's correction are presented. Significant p-values are indicated in bold, signs of significant relationships are indicated. In Ranunculus species, sample size (N) =12 populations, in P. antiscorbutica N =9 populations.
R. pseudotrullifolius Plant height Leaf length Leaf width Number of leaves Leaf shape P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign Soil water saturation 0.006 0.52 + 0.0081 0.50 + 0.02 0.42 + 0.75 0.089 0.14 0.13 pH 0.60 -0.035 0.69 -0.082 0.68 -0.080 0.19 0.12 0.42 0.026 Conductivity 0.29 0.033 0.12 0.19 0.031 0.42 0.84 -0.12 0.20 0.53
R. biternatus Plant height Leaf length Leaf width Number of leaves Leaf shape P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign Soil water saturation 0.072 0.29 0.0003 0.77 + 0.0036 0.61 + 0.52 -0.066 0.46 0.042 pH 0.071 0.27 0.0049 0.54 + 0.022 0.39 + 0.72 -0.085 0.72 0.08 Conductivity 0.0035 0.61 + 0.046 0.30 0.029 0.40 + 0.46 -0.045 0.62 0.015
R. moseleyi Plant height Leaf length Leaf width Number of leaves Leaf shape P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² Sign Soil water saturation 0.13 0.34 0.40 0.016 0.48 0.32 0.72 -0.085 0.6 -0.07 pH 0.95 -0.10 0.24 0.28 0.086 0.36 0.80 -0.093 0.81 0.09 Conductivity 0.098 0.21 0.20 0.12 0.45 0.015 0.80 -0.020 0.74 -0.087
P. antiscorbutica Plant height Leaf length Leaf width Flower size Leaf shape (Site in covariable) P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² Sign Soil water saturation 0.60 0.51 0.52 0.78 0.86 0.57 0.45 0.35 0.61 -0.14 pH 0.80 0.49 0.63 0.49 0.15 0.74 0.71 0.24 0.49 -0.26 Conductivity 0.020 0.85 - 0.0025 0.93 - 0.28 0.73 0.0021 0.91 - 0.66 0.24 Altitude 0.68 0.46 0.78 0.44 0.97 0.57 0.86 0.23 0.57 -0.1 Temperature 0.98 0.44 0.97 0.43 0.97 0.57 0.98 0.22 0.49 -0.05 (only Site) (0.04) (0.54) (0.046) (0.52) (0.019) (0.64) (0.12) (0.35) (0.56) (-0.05) 57
Chapitre 1A. Integration increases trait variation
Table 2. Relation between trait variation and abiotic conditions. Adjusted r squared and p-values after sequential Bonferroni's correction are presented. Significant p-values are indicated in bold, signs of significant relationships are indicated. In Ranunculus species N=12, in P. antiscorbutica N =9.
R. pseudotrullifolius Plant height Leaf length Leaf width Number of leaves Leaf shape P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign Soil water saturation 0.11 0.19 0.630. -0.04 0.63 -0.074 0.68 -0.080 0.45 0.12 pH 0.49 -0.02 0.66 -0.05 0.81 -0.094 0.75 -0.088 0.41 -0.025 Conductivity 0.48 -0.0064 0.87 -0.12 0.27 0.089 0.75 -0.11 0.05 0.33
R. biternatus Plant height Leaf length Leaf width Number of leaves Leaf shape P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign Soil water saturation 0.18 0.14 0.84 -0.11 0.37 0.038 0.94 0.12 0.42 0.14 pH 0.090 0.22 0.76 -0.089 0.13 0.094 0.93 -0.11 0.33 0.15 Conductivity 0.69 -0.091 0.32 0.043 0.57 -0.032 0.57 -0.078 0.54 0.097
R. moseleyi Plant height Leaf length Leaf width Number of leaves Leaf shape P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign Soil water saturation 0.049 0.29 0.15 0.15 0.73 -0.086 0.72 -0.044 0.44 -0.032 pH 0.45 -0.0002 0.65 -0.076 0.054 0.28 0.66 -0.078 0.39 0.14 Conductivity 0.48 -0.014 0.61 -0.037 0.98 -0.10 0.48 -0.043 0.84 0.03
P. antiscorbutica Plant height Leaf length Leaf width Flower size Leaf shape (Site in covariable) P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign P adj. r² sign
Soil water saturation 0.48 0.089 0.77 0.51 0.98 0.023 0.74 0.67 0.82 -0.3 pH 0.71 -0.048 0.64 0.53 0.49 0.13 0.67 0.67 0.78 0.32 Conductivity 0.049 0.57 0.47 0.53 0.51 0.16 0.39 0.73 0.92 -0.18
Altitude 0.83 -0.024 0.94 0.48 0.81 0.0082 0.43 0.71 0.84 -0.31 Temperature 0.94 -0.080 0.85 0.50 0.73 0.0054 0.66 0.69 0.92 -0.31 (only Site) (0.34) (0.01) (0.034) (0.56) (0.26) (0.15) (0.01) (0.71) 0.48 -0.10
58
Chapitre 1A. Integration increases trait variation
Table 3. Relation between phenotypic integration and abiotic conditions. Adjusted r squared and p-values after sequential Bonferroni's correction are presented. Significant p-values are indicated in bold, signs of significant relationships are indicated. In Ranunculus species N=12, in P. antiscorbutica N =9.
R. pseudotrullifolius Integration P. antiscorbutica Integration P adj. r² sign (Site in covariable) P adj. r² sign Soil water saturation 0.71 0.050 Soil water pH 0.97 -0.10 saturation 0.98 -0.14 Conductivity 0.74 -0.010 pH 1 -0.13 Conductivity 0.90 0.23 R. biternatus Integration Altitude 1 -0.024 P adj. r² sign Temperature 1 -0.097 Soil water saturation 1.39 -0.042 (only Site) (0.19) (0.24) pH 0.97 -0.10 Conductivity 1 -0.096
R. moseleyi Integration P adj. r² sign Soil water saturation 0.84 -0.095 pH 0.32 0.11 Conductivity 0.060 0.38
59
Chapitre 1A. Integration increases trait variation
Table 4. Exploring the relationship between trait variation (CV) and phenotypic integration expected at random, based on 10000 individuals belonging to 1000 populations. Adjusted r squared and p-values after sequential Bonferroni's correction are presented. Note that integration does not relate to trait variation, indicating that integration and trait variation are not related for purely numerical reasons.
Trait 1 Trait 2 Trait 3 Trait 4 P adj. r² P adj. r² P adj. r² P adj. r² Phenotypic integration 1 -0.0020 1 -0.0015 1 -0.00072 1 -0.0016
Table 5. Anova test of the influence of trait variation (high or low) on phenotypic integration. N= 80 populations of 10 individuals. Lack of relationship indicates that all scenarios from Figure 1 are theoretically equally possible.
ANOVA test Df Sum Sq Mean Sq F value Pr(>F) Effect of degree of trait variation on integration 3 17.5 5.83 Residuals 76 1875 24.67 0.24 0.87
60
Chapitre 1A. Integration increases trait variation
Appendix 1 Measured range of environmental factors in species studied.
Soil water saturation (%) pH Conductivity (mS.cm-1) min. max. mean sd. min. max. mean sd. min. max. Mean sd. R. biternatus 20.00 99.88 69.91 23.71 4.84 7.14 5.75 0.63 0.00 4.6 1.13 1.50 R. pseudotrullifolius 36.75 96.75 82.27 16.82 5.12 6.39 5.67 0.34 0.00 4.67 2.16 1.63 R. moseleyi 37.37 99.84 77.51 20.28 5.27 6.82 6.00 0.52 0.00 5.97 1.51 1.96 P. antiscorbutica 9.76 65.51 34.37 20.96 5.01 7.13 6.02 0.75 0.00 1.42 0.45 0.59
61
Chapitre 1A. Integration increases trait variation
Appendix 2. Mean and standard deviation of soil water saturation, pH and conductivity in terrestrial or aquatic stands in Ranunculus species.
Species conditions Soil water saturation (%) pH Conductivity (mS.cm-1) Population mean sd. mean sd. Mean sd. Number R. biternatus terrestrial 65.33 26.35 5.58 0.65 0.56 0.60 12 aquatic 73.85 24.51 5.90 0.63 1.51 1.83 12 R. pseudotrullifolius terrestrial 79.11 21.65 5.75 0.22 1.98 1.67 12 aquatic 85.44 11.37 5.59 0.43 2.38 1.79 12 R. moseleyi terrestrial 73.71 14.26 5.68 0.42 0.86 0.54 12 aquatic 80.68 25.16 6.32 0.42 2.05 2.59 12
62
Chapitre 1A. Integration increases trait variation
Appendix 3. Relation between phenotypic integration and trait variation (coefficient of variation) or environmental factors (mean value) in populations. Species identity is included as covariable. N= 36 populations (10 individuals by population). The combination of traits used in the study is given in bold.
Trait variation / Environmental factors
flower size
flower size
flower size
63
Chapitre 1A. Integration increases trait variation
Appendix 4. Trait variation and associated standard deviation in populations of Ranunculus species. Standard deviation of trait variations are calculated by bootstrap (1000 resample). Note that most trait variation differences among populations persist even after accounting for standard deviation, indicating robustness of comparisons among trait variations.
64
Chapitre 1A. Integration increases trait variation
Appendix 5. Relation between phenotypic integration and trait variation (coefficient of variation). Species identity is included as covariable. N= 18 populations (20 individuals by population). Relationships are very similar to those found for 36 populations of 10 individuals, indicating robustness of these analyzes despite small sample size.
Trait variations Plant height Leaf length Leaf witdth P adj. r² P adj. r² sign P adj. r² Phenotypic integration 0.91 -0.07 0.016 0.24 + 0.12 0.009
65
Chapitre 1A. Integration increases trait variation
Fig.1 Theorical relationship between variance and integration (covariance). In the case of null variance integration is not calculated.
Fig.2 Theoretical extreme scenarios of the relationship among environment, phenotypic integration and trait variation. Arrows in bold and dotted lines represent respectively strong direct and weak indirect relationships. See Introduction for explanations.
Fig.3 (a) Location of Iles Kerguelen in the Southern Indian Ocean; (b) sampled sites in Iles Kerguelen (map modified from Google maps).
Fig.4 Relationship between phenotypic integration and variation of individual traits in Ranunculus species in the field. P values after sequential Bonferroni’s correction and adjusted r squared are indicated. In each species, N =12.
Fig.5 Relationship between phenotypic integration and variation of individual traits in Pringlea antiscorbutica in the field. P values after sequential Bonferroni’s correction and adjusted r squared are indicated, N =9.
Fig.6 Relationship between phenotypic integration and variation of individual traits in R. biternatus, R. pseudotrullifolius and P. antiscorbutica in controlled conditions. P values after sequential Bonferroni’s correction and adjusted r squared are indicated. In R. biternatus N =13, in R. pseudotrullifolius N =8 and in P. antiscorbutica N =38.
Fig.7 Summarized relationship between environment, trait variation and phenotypic integration.
66
Chapitre 1A. Integration increases trait variation
Figure 1
67
Chapitre 1A. Integration increases trait variation
Figure 2
E viro e t E viro e t
I tegratio Trait variatio
Trait variatio I tegratio
a
d E viro e t E viro e t
Trait variatio I tegratio Trait variatio I tegratio
68
Chapitre 1A. Integration increases trait variation
Figure 3
and Ile Guillou
69
Chapitre 1A. Integration increases trait variation
Figure 4
70
Chapitre 1A. Integration increases trait variation
Figure 5
71
Chapitre 1A. Integration increases trait variation
Figure 6
72
Chapitre 1A. Integration increases trait variation
Figure 7
E viro e t
no relationship
Trait variatio I tegratio positive relationship
73
Chapitre 1B. Exploring performance of species from Iles Kerguelen: relationships to environment and phenotypic integration
74
Chapitre 1B. Exploring plant performance
RESUME
Les îles Kerguelen, situées dans la région subantarctique, sont exposées à un changement climatique intense caractérisé par une augmentation des températures et une diminution des précipitations. Les espèces végétales indigènes semblent particulièrement sensibles à ces modifications environnementales. L’objectif de cette étude est d’explorer les conditions environnementales (notamment température et teneur en eau) qui diminuent d’ores et déjà la performance des plantes sur le terrain, afin de pouvoir estimer la capacité des espèces à faire face au changement climatique. Egalement, il est suggéré que le degré d’intégration phénotypique (i.e. degré de corrélation entre les traits) a un effet positif sur la performance des plantes. Nous déterminerons la relation entre intégration et performance des plantes. Chez trois espèces de Ranunculus indigènes des îles Kerguelen, nous avons étudié des populations naturelles le long d’un gradient de saturation en eau du sol, et d’un gradient d’altitude servant de proxy au gradient de température. Dans chaque population nous avons caractérisé l’environnement abiotique et biotique, ainsi que l’intégration et la performance des plantes. Nous avons montré que l’intégration phénotypique n’est pas reliée à la performance des plantes. Une diminution de la teneur en eau diminue systématiquement la performance des plantes chez les trois espèces étudiées, contrairement à la température qui peut augmenter ou diminuer la performance des plantes. Ainsi, les changements climatiques en cours devraient avoir un effet négatif sur la performance des espèces étudiées. Nous suggérons que la persistance d’habitats humides favorables aux renoncules sera un facteur prépondérant de la survie des espèces .
75
Chapitre 1B. Exploring plant performance
Exploring performance of species from the Iles Kerguelen: relationships to environment and phenotypic integration
Bastien Labarrere, Andreas Prinzing & Françoise Hennion
INTRODUCTION
Iles Kerguelen located in the sub-Antarctic region are facing fast and intense climate change
(Lebouvier et al., 2011). Increased temperature and decreased rainfall have been noted in the last decades (Frenot et al., 2006; Lebouvier et al., 2011). Plants in Iles Kerguelen experience stable environment including low temperatures and abundant rainfall, and hence appear to be particularly sensitive to climate change (Frenot et al., 2006). Signs of stress have been observed in plants during summer periods for several decades (Hennion, 1992; Chapuis et al., 2004). In addition, Iles Kerguelen are isolated in the South Indian Ocean, standing around 3500km from
Africa, the same distance from Australia (Lebouvier and Frenot, 2007). Therefore, isolation does not allow gene flow from continental populations that could introduce adaptation and dispersal is unlikely to offer an escape route for their floras. Instead the plants on these islands will need to cope with the changes of their environment to survive.
Within species, trait values may vary among populations such that plants are adapted to their local environments (Van Valen, 1973; Schlichting, 1986). Such trait variation may result from genetic differences (Kanaga et al., 2008) or phenotypic plasticity (Sultan, 2000). Variation in plant performance among different environmental conditions show how well plants cope with environmental variation and are an indicator of how plants might respond to environmental change. Particularly, determining how temperature and dryness already impact plant performance might be an indicator of plant capacity to cope with climate change.
76
Chapitre 1B. Exploring plant performance
Traits of an organism do not vary independently but are correlated one to each other. The number and strength of correlations among traits across individuals is defined as the degree of phenotypic integration (Schlichting 1989; Cheverud 1996; Wagner and Altenberg 1996;
Pigliucci et al. 2003). Phenotypic integration may result from genetic, developmental or environmental constraints (Conner, 2002; Murren, 2012; Armbruster, 2014). Phenotypic integration maintains consistency across variation in trait values within plants (Schlichting,
1989; Murren, 2012). Therefore, phenotypic integration has been suggested to enhance plant performance, yet evidence is lacking (Gianoli and Palacio-Lopez, 2009; Mallitt et al., 2010).
Iles Kerguelen exhibit strong environmental gradients (Hennion et al., 2006), and plants from different populations experience different local environments (Hennion et al., 2006;
Labarrere et al., unpubl.). We studied three Ranunculus species (R. biternatus, R. pseudotrullifolius and R. moseleyi) indigenous to Iles Kerguelen. Ranunculus biternatus is austral circumpolar, R. pseudotrullifolius magellanic and on Kerguelen, and R. moseleyi is a strict Kerguelen endemic (Lehnebach et al., 2017). Ranunculus species, with shallow roots and living in wet habitats appear to be particularly sensitive to climate change (Le Roux et al.,
2008). Earlier studies showed some variation in performance of these species in relation with environmental conditions in the field (Hennion et al., 1994) or under cultivation (Hennion and
Walton, 1997a). Here we explore which environmental factors may impact plant performance in the field and concentrate particularly on temperature and dryness, as well as investigating the effect of phenotypic integration on plant performance. We hypothesize that both R. biternatus and R. pseudotrullifolius, being of magellanic origin, may perform better at higher temperatures, whereas R. moseleyi, strict Kerguelen endemic, may have a narrower sub-
Antarctic climate temperature optimum.
We led a baseline study, investigating the influence of biotic and abiotic environmental factors on plant populations on Iles Kerguelen. We studied plant populations across altitudinal
77
Chapitre 1B. Exploring plant performance
and soil water saturation gradients, using altitudinal gas a proxy for temperature (Schortemeyer et al., 2015). In each population we measured morphological traits and plant performance, and we characterized the abiotic environment (temperature, soil water saturation, pH, conductivity) and the biotic environment (species richness and diversity of the surrounding community).
Furthermore, populations in the field may have adapted to their local environments over many generations. Long term adjustment to local conditions may differ from the need to cope rapidly with climate change. Transplantation experiments show how plants cope with rapid environmental change. Therefore, we coupled in situ observations with transplant experiments to a laboratory room to measure plant response to a rapid increase in temperature. We therefore investigated (i) the relationship between phenotypic integration and plant performance in situ
(ii) how plant performance varies on a temperature and dryness gradient in situ (iii) how plants respond to warmer conditions in laboratory room.
78
Chapitre 1B. Exploring plant performance
MATERIALS AND METHODS
1. Species under study
The Iles Kerguelen (49°20’00” S, 69°20’00” E) are located in the Southern Indian Ocean within the sub-Antarctic region (Lebouvier and Frenot, 2007; Figure 3). We studied three perennial
Ranunculus species: R. biternatus, R. pseudotrullifolius and R. moseleyi. On the Iles Kerguelen these species occupy different habitats: Ranunculus biternatus is widespread, occurring in many habitat types up to 500m above sea level (Hennion and Walton, 1997a; Réserve Naturelle des Terres Australes et Antarctiques françaises and IPEV Programme 136, unpublished data).
In contrast, the aquatic R. moseleyi and halophilous R. pseudotrullifolius have more restricted distributions, R. pseudotrullifolius occurs only near the coast in peaty or sandy shorelines and ponds, and sporadically in ponds up to about 200m a.s.l. (Réserve Naturelle des Terres
Australes et Antarctiques françaises and IPEV Programme 136, unpublished data). Ranunculus moseleyi is strictly aquatic, growing only in freshwater lakes and ponds (Hennion and Walton,
1997a).
2. Collection of data at field sites-
2.1. Plant measurements
From December 2015 to January 2016, we sampled Ranunculus populations in three regions in
Iles Kerguelen: Port-Jeanne-d’Arc, Port-Douzième and Val Studer (Figure 3, general introduction). Ranunculus biternatus was sampled in the three regions while R. pseudotrullifolius was sampled in Port-Jeanne-d’Arc and R. moseleyi in Port-Jeanne-d’Arc and
Val Studer. Within regions, we sampled populations across water saturation and altitudinal gradients. In this purpose, we sampled two locations per region at different altitudes. In each location, we sampled two nearby populations: one in low water saturation condition (i.e. “dry population”) and the other in high water saturation condition (i.e. “wet population”).
79
Chapitre 1B. Exploring plant performance
Ranunculus moseleyi is strictly aquatic, hence dry population in this species means a population temporarily dried out. In each population, sampling was performed across a small area so as to ensure that plants belonging to a same population were subject to roughly similar environmental conditions. Twelve populations were sampled for R. biternatus, two in R. pseudotrullifolius and four in R. moseleyi. In each population we selected 10 individuals and measured plant height and diameter and petiole length, lamina length and lamina width of the largest leaf.
2.2. Estimation of plant performance
We used the percentage of non-senescent leaves as a proxy of plant performance (Zamora et al., 1998). Also, we used specific leaf area (SLA) as an indicator of stress (Cornelissen et al.,
2003; Vaieretti et al., 2007; Sonnier et al., 2010). To determine SLA, 20 leaves were sampled from a minimum of 10 individuals. Leaves were collected then put in distilled water in airtight bags for rehydration after several days at cool temperature in the dark. Surface and weight of leaves were estimated. Then, leaves were dried for 48 hours at 80°C, and weighed another time, following Cornelissen et al. (2003). SLA was calculated as the ratio between leaf surface and leaf dry mass (m².kg-1) (Garnier et al., 2001).
2.3. Environmental measurements
We measured soil water saturation, pH and conductivity in each population. In wet populations we measured in situ the pH using a pH meter (BASIC 20 PLUS CRISON, resolution 0.01pH), and collected 50mL of water. Conductivity in the water was determined in the laboratory using a conductivity meter (CONSORT K810, resolution 0.1 µS cm-1) (Hermant et al., 2013). In dry populations, soil water saturation and conductivity were measured either in situ using a TDR
Trime probe, or in the laboratory. For the laboratory measurement, 50mL of soil were collected.
In the laboratory, the soil was mixed with a known volume of distilled water and then left 18 to
24 hours to permit sedimentation of soil particles. Soon after, pH was determined using a pH 80
Chapitre 1B. Exploring plant performance meter (BASIC 20 PLUS CRISON, resolution 0.01pH) (Hermant et al., 2013). After another 18 to 24 hours of sedimentation, conductivity was determined using a conductivity meter
(CONSORT K810, resolution 0.1 µS cm-1) (Hermant et al., 2013). In order to explore soil nutrient content, we collected 50mL of soil (or water) in each population and determined the
C/N ratio in LAS laboratory. Temperature was determined using loggers installed at soil surface close to the populations, which recorded every 30 minutes from January to February. Altitude was determined using GPS eTrex 10 Garmin.
We characterized the plant community for each population, recording each species present and its abundance, and calculating species richness (number of species) and diversity using
Shannon index (Keylock, 2005). Data collected in the field allowed us to determine the relation between surrounding community and plant traits only in R. biternatus. To determine the relation between surrounding community and traits in R. pseudotrullifolius and R. moseleyi, we added data collected in 2011-2012, using same protocols.
3. Collection of data in laboratory room-
3.1. Plant collection and measurement
We collected three of the ten measured individuals from each population and transplanted them to a laboratory room at Port-aux-Français where they were cultivated for 3 months. Plants were grown in a sheltered room at 10°C, and were watered once a week. The experiment involved
36 plants from the 12 populations in R. biternatus, 12 plants from four populations in R. moseleyi and six plants from two populations in R. pseudotrullifolius. Conditions in the laboratory room were warmer than in the field (Appendix 1). Plant survival was recorded after the 3 month experiment, and surviving plants were measured following the same protocol as in the field. Only 12 plants from four populations of R. biternatus, three plants from one population of R. moseleyi, and six plants from two populations of R. pseudotrullifolius survived.
81
Chapitre 1B. Exploring plant performance
4. Data analyses-
4.1. Phenotypic integration
In each population, we calculated phenotypic integration. Phenotypic integration was calculated as the percentage of variance explained by the first axis of a Principal Component Analysis
(PCA) performed on traits (Hermant et al., in 2013). Traits considered were plant height and diameter, petiole length, lamina length and lamina width of the largest leaf. PCA was performed using correlation matrix. Trait values were scaled and PCA conducted using the FactoMineR package of R 3.0.2 (Lê et al., 2008; R Core Team, 2012).
5. Hypothesis Testing-
To determine local differences of performance between dry and wet populations within locations we used ANOVA using R 3.0.2 (R Core Team, 2012). Likewise, we used ANOVA to determine local differences of performance between low and high altitude within regions
(local gradient). The relationships between integration or environments and performance across regions were tested using simple linear regressions, using the same software. For multiple testing on the same data set, p-values were corrected using sequential Bonferroni’s correction
(Cabin and Mitchell, 2000). In all regression analyses we verified the assumptions of the analyses using residual plots. We investigated whether traits or performance were higher or lower in the laboratory room using a paired t-test (R. 3.0.2). We used linear regressions to determine the relationships between surrounding community (richness or diversity) and target- plant traits or performance. However, the biotic environment might not be independent from the abiotic environment so we used redundancy analyses (rda function in R. 3.0.2) to determine relationships between surrounding community (richness or diversity) and target-plant traits or performance with abiotic environment as a co-variable. We found similar results with or
82
Chapitre 1B. Exploring plant performance without the abiotic environment as a co-variable (data not shown) so we chose to show results from linear regressions.
83
Chapitre 1B. Exploring plant performance
RESULTS
1. Influence of environment on plant performance across regions
In R. biternatus, we found no relationships between environmental factors and SLA (Table 1).
In contrast, the percentage of non-senescent leaves increased significantly with pH and decreased with conductivity (Table 1). The percentage of non-senescent leaves also increased with soil water saturation and C/N ratio but relationships were only marginally significant
(Table 1). Also, we found no effect of the surrounding community on plant size or performance in any of the species (Table 4 and 5).
2. Influence of phenotypic integration on plant performance across regions
In R. biternatus, we found no relationships between phenotypic integration and the percentage of non-senescent leaves or SLA (p>0.05; mean adj.r²= 0.012; respectively df= 12; 6). We found similar results in dry or wet populations, as well as in low or high altitude populations (data not shown).
3. Influence of dry local conditions on plant performance at different altitudes
For all three species and all situations (altitudes and regions), the percentage of non-senescent leaves was higher in wet populations than in dry populations with one exception, that of R. biternatus at Val Studer, where the percentage of non-senescent leaves was higher in dry than in wet populations at high altitude (Figure 1; Table 2).
4. Influence of altitude on plant performance in wet or dry populations
In R. biternatus, percentage of non-senescent leaves was lower at high altitude in wet conditions but no relation was found in dry conditions, across regions (Figure 1; Table 3). However the opposite pattern was found in Port-Jeanne-d’Arc, with more non-senescent leaves at high
84
Chapitre 1B. Exploring plant performance altitude in dry but not in wet populations (Figure 1; Table 3). In contrast in Val Studer, percentage of non-senescent leaves decreases with altitude in wet populations but not in dry populations (Figure 1; Table 3). In Port-Douzième, there is no relationship between percentage of non-senescent leaves and altitude in wet or dry populations (Figure 1). In R. moseleyi, percentage of non-senescent leaves decreases with altitude in dry or wet conditions (Figure 1;
Table 3).
5. Percentage of plant survival in laboratory room
R. biternatus and R. moseleyi suffered high mortality in the growth room. Survival did not depend on field measures of SLA, percentage of non-senescent leaves, mean integration or environmental conditions (altitude or soil water saturation) of the source population (p>0.05; mean adj.r²=0.1; respectively df= 15 and 5). In R. pseudotrullifolius, survival was of 100% but this involved only two populations.
6. Trait and performance differences between laboratory room and the field
In R. biternatus, we found no difference in percentage of non-senescent leaves between the laboratory room and the field (p>0.05; df=15; mean difference=0.083%; SD=22.18). Plant traits
(Plant height, lamina length and width) were significantly higher after 3 months in the laboratory room (respectively p=0.0039; 0.00021; 0.0029; df=15). In R. pseudotrullifolius the percentage of non-senescent leaves was significantly higher in the laboratory room than in the field (mean difference =16.67%; SD=15.31; p=0.022; df=6). Plant height and lamina length were not significantly different after 3 months in the laboratory room (p>0.05; df=6). In contrast, lamina width was higher in the laboratory room (p=0.042; df=6). In R. moseleyi, the percentage of non-senescent leaves decreased non-significantly to 7.46% lower (SD=26.43
(p>0.05; df=5) in the laboratory room than in the field before transplantation. Plant traits (Plant
85
Chapitre 1B. Exploring plant performance height, lamina length and width) were significantly higher after 3 months in the laboratory room
(respectively p=0.016; 0.015; 0.019; df=5).
86
Chapitre 1B. Exploring plant performance
DISCUSSION
Iles Kerguelen face intense and rapid climate change characterized by an increase of temperature and a decrease of rainfall (Lebouvier et al., 2011). Plant species in Iles Kerguelen have been growing for a long time in stable environment characterized by cold and rainy conditions, and might be impacted by such changes. Also, correlations among traits within organisms, defined as phenotypic integration might influence plant performance. Particularly, phenotypic integration is suggested to increase plant performance but evidence is lacking
(Malitt et al., 2010). We found that phenotypic integration did not influence plant performance in the field. Also, we showed in the field that low water saturation had a more severe impact on plant performance than altitude. Moreover, we showed that performance did not decrease in response to a slight increase of temperature in a laboratory room.
1. Phenotypic integration does not influence plant performance
The consequences of integration on plants remain to be analysed (Murren, 2012). Particularly, phenotypic integration is suggested to increase performance of organisms but evidence is lacking (Malitt et al., 2010). In this study, we showed no relationship between plant integration and performance. The effect of phenotypic integration on plant performance might be higher in stressful conditions (Schlichting, 1989). Organisms that show higher integration are supposed to be internally more consistent, and hence more able to cope with stressful conditions. We tested the relationship between integration and performance in different environments. We found no relationships between plant integration and performance, whatever the altitude or soil water saturation of the populations. The consequences of phenotypic integration on plant performance still remain to be deciphered.
87
Chapitre 1B. Exploring plant performance
2. Influence of environment on plant performance
In R. biternatus, we showed that performance was generally higher in wet populations than in dry populations, whatever the altitude. Only in one of the three regions studied (i.e. Val Studer), at high altitude, plant performance was lower in a wet than in a dry population.Altitudinal sites in Kerguelen, notably in Val Studer, are subject to numerous freeze-thawing cycles within a year (Hennion et al., 2006). Under these conditions, submerged plants in the wet populations may be subject to important freezing and thawing, negatively affecting plant performance. In contrast, in Port-Jeanne-d’Arc and Port-Douzième, plant performance did not differ with altitude in wet populations. In dry populations, performance was lower at low altitude in Port-
Jeanne-d’Arc, but no difference was found in other regions. At high altitude, temperature is lower than at low altitude and the thermal gradient is strong (Hennion et al., 2006). Hence, in
Port-Jeanne-d’Arc, temperature impacts plant performance in dry populations but not in wet populations. Also in R. biternatus, we found across regions that soil acidity and conductivity particularly decrease plant performance. Ongoing increase of temperature and dryness lead to stronger soil evaporation and may have for consequence an increase of soil conductivity in sub-
Antarctic islands, which are subject to salt sprays (Le Roux et al., 2008). Increase of soil conductivity is less commonly studied compared to warming or dryness, but might be a major issue for plants in the next decades (Le Roux et al., 2008).
Overall, an increase of temperature and dryness might impact the performance of R. biternatus. Ranunculus biternatus was described to occur up to 300m a.s.l. in Iles Kerguelen
(Hennion and Walton, 1997a), however in recent and larger scale surveys the species was sporadically found up to around 500m (Réserve Naturelle des Terres Australes et Antarctiques françaises and IPEV Programme 136, unpublished data). Ranunculus biternatus might reach cooler altitudinal environments in Iles Kerguelen to cope with climate warming. However, an
88
Chapitre 1B. Exploring plant performance increase of wind speed has been noted in the last decades in sub-Antarctic Marion Island, having for effect an increase of evaporation, particularly at high altitude (Le Roux et al., 2008);
In R. moseleyi, performance was higher at higher temperature (i.e. low altitude) in wet or dry (i.e. temporarily emerged) populations. Also, performance was lower in dry populations whatever the temperature (i.e. low or high altitude). Hence, simultaneous increase of temperature and dryness induced by climate change might both increase and decrease plant performance. It should be noted that R. moseleyi is strictly aquatic (Hennion and Walton,
1997a). Climate change in Kerguelen is characterized by a significant reduction of the total level of precipitations (Lebouvier et al., 2011). Indeed, mean annual precipitations decreased by 100mm to 250mm in the two past decades (Frenot et al., 2006). Climate drying may reduce the number of pools, as shallow ones dry up during summer (Hennion et al., 1994; Le Roux et al., 2008). Such reduction in pools might impact populations of R. moseleyi. Overall, plants might be able to cope with an increase of temperature, but survival of R. moseleyi might be more dependent on the persistence of wet habitats.
In R. pseudotrullifolius, we showed that plant performance was higher under wet conditions.
Ranunculus pseudotrullifolius, being halophilous, occurs near the coast, in areas receiving salt sprays on sandy shorelines and in ponds (Hennion and Walton, 1997a). Consequently, R. pseudotrullifolius might be impacted by both an increase of temperature and the decrease of precipitations. Ranunculus pseudotrullifolius was only described at low altitude (Aubert de la
Rüe, 1964), but here we found this species at 230m a.s.l. Nevertheless, we suggest that the need to grow in saline environments might hinder this species ability to colonise high altitude sites.
3. Response of plants to a rapid increase of temperature, in a laboratory room
We transplanted plants from their populations in the field to the laboratory, to an unheated room subject to slightly warmer and drier conditions, sheltered from the wind. We could not identify the factors that influenced survival and plant performance in the laboratory room, finding no
89
Chapitre 1B. Exploring plant performance effect of environmental variables from the source populations or degree of phenotypic integration of the source population.
In R. biternatus and R. moseleyi, plant performance did not significantly increase or decrease in the laboratory room. In contrast, we found that plants in R. pseudotrullifolius were more performant in the laboratory room than in the field. Note that the experiment in the laboratory room lasted 3 months, and that changes in performance might appear in a longer term.
Cultivation experiments in a common garden for 4 years at Brest, under a cool oceanic climate
(much warmer than our current experimental conditions), showed higher records of survival in
R biternatus than in R pseudotrullifolius, in turn higher than in R moseleyi (Hennion and
Couderc, unpubl.). Our results show that under slightly higher and drier conditions for short (3 months) duration, R. biternatus, R. moseleyi and R. pseudotrullifolius did not decrease their performance.
90
Chapitre 1B. Exploring plant performance
CONCLUSION
Overvall, R. moseleyi, endemic to Iles Kerguelen, shows a limited distribution whereas R. biternatus being austral and circumpolar is distributed across a wider range of temperature across the southern hemisphere. Yet, within Iles Kerguelen R. moseleyi surprisingly performs better under warmer conditions in the field contrary to R. biternatus. We showed in the three species that plants from drier populations are less performant. Differences of performance observed among populations in relation with different local environments may be an indicator of how plants will face environmental changes. Hence, ongoing dryness and to a lesser extent warming are suggested to decrease plant performances. In addition, dryness and warming may increase soil conductivity, possibly even increasing the impact on plants. In Iles Kerguelen species have the possibility to reach cooler altitudinal environments. Yet, climate change in the sub-Antarctic tends to increase wind speed, resulting in increased dryness due to evaporation, especially at high altitudes (Le Roux et al., 2008). Hence, we suggest that the three Ranunculus species in Iles Kerguelen might not avoid, but will have to cope with environmental changes.
Persistence of freshwater habitats suitable to the three Ranunculus species might be the major component of species survival facing climate change in Iles Kerguelen.
91
Chapitre 1B. Exploring plant performance
REFERENCES
Aubert De La Rüe E. (1964) Observations sur les caracteres et la repartition de la vegetation des Iles Kerguelen. Comité national français des recherches antarctiques Armbruster W.S., Pélabon C., Bolstad G.H. & Hansen T.F. (2014) Integrated phenotypes: understanding trait covariation in plants and animals. Phil. Trans. R. Soc. B, 369, 20130245. Cabin, R. J., & Mitchell, R. J. (2000). To Bonferroni or not to Bonferroni: when and how are the questions. Bulletin of the Ecological Society of America, 81, 246-248. Chapuis J.-L., Frenot Y. & Lebouvier M. (2004) Recovery of native plant communities after eradication of rabbits from the subantarctic Kerguelen Islands, and influence of climate change. Biological Conservation, 117, 167-179. Cheverud, J. M. (1996). Developmental integration and the evolution of pleiotropy. American Zoologist, 36, 44-50. Conner J.K. (2002) Genetic mechanisms of floral trait correlations in a natural population. Nature, 420, 407-410. Cornelissen, J. H. C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D. E., ... & Pausas, J. G. (2003). A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian journal of Botany, 51, 335-380. Frenot Y., Lebouvier M., Gloaguen J.-C., Hennion F., Vernon P. & Chapuis J.-L. (2006) Impact des changements climatiques et de la fréquentation humaine sur la biodiversité des îles subantarctiques françaises. Belgeo. Revue belge de géographie, 363-372. Garnier, E., Shipley, B., Roumet, C., & Laurent, G. (2001). A standardized protocol for the determination of specific leaf area and leaf dry matter content. Functional ecology, 15, 688-695. Gianoli E. & Palacio López K. (2009) Phenotypic integration may constrain phenotypic plasticity in plants. Oikos, 118, 1924-1928. Hennion F. (1992) Etude‐ des caractéristiques biologiques et génétiques de la flore endémique des îles Kerguelen. Hennion F., Fiasson J.L. & Gluchoff-Fiasson K. (1994) Morphological and phytochemical relationships between Ranunculus species from Iles Kerguelen. Biochemical systematics and ecology, 22, 533-542. Hennion F. & Walton D.W.H. (1997a) Seed germination of endemic species from Kerguelen phytogeographic zone. Polar Biology, 17, 180-187. Hennion F. & Walton D.W.H. (1997b) Ecology and seed morphology of endemic species from Kerguelen phytogeographic zone. Polar Biology, 18, 229-235. Hennion F., Frenot Y. & Martin Tanguy J. (2006) High flexibility in growth and polyamine composition of the crucifer Pringlea antiscorbutica in relation to environmental conditions. Physiologia Plantarum‐ , 127, 212-224. Hermant M., Prinzing A., Vernon P., Convey P. & Hennion F. (2013) Endemic species have highly integrated phenotypes, environmental distributions and phenotype–environment relationships. Journal of Biogeography, 40, 1583-1594. Kanaga M.K., Ryel R.J., Mock K.E. & Pfrender M.E. (2008) Quantitative-genetic variation in morphological and physiological traits within a quaking aspen (Populus tremuloides) population. Canadian Journal of Forest Research, 38, 1690-1694. Keylock, C. J. (2005). Simpson diversity and the Shannon–Wiener index as special cases of a generalized entropy. Oikos, 109, 203-207. Le Roux P.C. & McGeoch M.A. (2008) Changes in climate extremes, variability and signature on sub-Antarctic Marion Island. Climatic Change, 86, 309-329.
92
Chapitre 1B. Exploring plant performance
Lê S., Josse J. & Husson F. (2008) FactoMineR: an R package for multivariate analysis. Journal of statistical software, 25, 1-18.
Lebouvier M., Laparie M., Hulle M., Marais A., Cozic Y., Lalouette L., Vernon P., Candresse T., Frenot Y. & Renault D. (2011) The significance of the sub-Antarctic Kerguelen Islands for the assessment of the vulnerability of native communities to climate change, alien insect invasions and plant viruses. Biological Invasions, 13, 1195-1208. Lebouvier M. & Frenot Y. (2007) Conservation and management in the French sub-Antarctic islands and surrounding seas. Mallitt K.L., Bonser S.P. & Hunt J. (2010) The plasticity of phenotypic integration in response to light and water availability in the pepper grass, Lepidium bonariense. Evolutionary ecology, 24, 1321-1337. Murren C.J. (2012) The integrated phenotype. Integrative and comparative biology, 52, 64- 76. Pigliucci M. (2003) Phenotypic integration: studying the ecology and evolution of complex phenotypes. Ecology Letters, 6, 265-272. Schlichting C.D. (1986) The evolution of phenotypic plasticity in plants. Annual review of ecology and systematics, 667-693. Schlichting C.D. (1989) Phenotypic integration and environmental changes. BioScience, 39, 460-464. Schortemeyer, M., Evans, J. R., Bruhn, D., Bergstrom, D. M., & Ball, M. C. (2015). Temperature responses of photosynthesis and respiration in a sub-Antarctic megaherb from Heard Island. Functional Plant Biology, 42, 552-564. Sultan S.E. (2000) Phenotypic plasticity for plant development, function and life history. Trends in plant science, 5, 537-542. Sonnier, G., Shipley, B., & Navas, M. L. (2010). Quantifying relationships between traits and explicitly measured gradients of stress and disturbance in early successional plant communities. Journal of Vegetation science, 21, 1014-1024. Vaieretti, M. V., Díaz, S., Vile, D., & Garnier, E. (2007). Two measurement methods of leaf dry matter content produce similar results in a broad range of species. Annals of Botany, 99, 955-958. Van Valen L. (1973) A new evolutionary law. Evolutionary theory, 1, 1-30. Wagner, G. P., & Altenberg, L. (1996). Perspective: complex adaptations and the evolution of evolvability. Evolution, 967-976. Zamora, R., Gómez, J. M., & Hódar, J. A. (1998). Fitness responses of a carnivorous plant in contrasting ecological scenarios. Ecology, 79, 1630-1644.
93
Chapitre 1B. Exploring plant performance
Table 1. Linear regression indicating relationships between SLA or percentage of non-senescent leaves and environmental factors in R. biternatus. P-values (P), adj.r² (R²) and the sign of the relation (sign) are indicated.
Altitude Temperature Soil water saturation pH Conductivity C/N Df P R² sign P R² sign P R² sign P R² sign P R² sign P R² sign SLA 0.49 0.049 0.35 0.002 0.59 0.034 0.27 0.13 0.35 0.002 0.59 0.16 10 Non-senescent leaves (%) 0.86 0.074 0.78 0.07 0.091 0.14 + 0.0024 0.48 + 0.0007 0.71 - 0.077 0.19 + 15
Table 2. ANOVA indicating difference of plant performance (percentage of non-senescent leaves) between dry and wet populations at low or high altitudes, in the different locations. P-values (P), F value (F) and degrees of freedom (Df) are indicated. “” : performance is lower/higher in dry populations.
low altitude high altitude Df Performance in dry vs wet populations P F P F R. biternatus Across regions 0.0001 42 < 0.014 6.8 < 58 Port-Jeanne-d'Arc 0.0001 125 < 0.0015 14 < 18 Port-Douzième 0.04 4.85 < 0.0001 240 < 18 Val Studer 0.0001 26.5 < 0.043 4.75 > 18 R. moseleyi Across regions 0.001 15.6 < 0.032 7.5 < 18 R. pseudotrullifolius Port-Jeanne-d'Arc NA NA 0.036 5.14 < 18
94
Chapitre 1B. Exploring plant performance
Table 3. ANOVA indicating difference of plant performance (percentage of non-senescent leaves) between low and high altitude in dry or wet populations, in the different locations. P-values (P), F value (F) and degrees of freedom (Df) are indicated. “” : performance is lower/higher at low altitude.
dry population Wet population Df Performance at low vs high altitude P F P F R. biternatus Across regions 0.19 1.69 0.018 5.9 > 58 Port-Jeanne-d'Arc 0.002 12.1 < 0.17 2 18 Port-Douzième 0.15 2.28 0.15 2.25 18 Val Studer 0.4 0.5 0.0001 30 > 18 R. moseleyi Across regions 0.009 8.4 > 0.02 3.8 > 18
95
Chapitre 1B. Exploring plant performance
Table 4. Linear regressions indicating the relation between surrounding community (species richness and diversity) and target plant traits or performance in R. biternatus in 2015-2016 campaign. P-values (P), F values (F) and degree of freedom (Df) are indicated.
Non- Number of Leaf Phenotypic Plant height Leaf width senescent Leaf spread leaves length integration leaves (%) Surrounding community P F P F P F P F P F P F P F Df Shannon diversity 0.22 1.74 0.38 0.80 0.20 2.0 0.56 0.54 0.27 0.026 0.26 1.47 0.22 1.70 10 Species richness 0.79 0.075 0.38 0.69 0.36 0.87 0.56 0.47 0.14 0.10 0.77 0.10 0.91 0.012 10
96
Chapitre 1B. Exploring plant performance
Table 5. Linear regressions indicating the relation between surrounding community (species richness and Shannon index diversity) and target plant traits or performance in R. biternatus, R. pseudotrullifolius and R. moseleyi, in 2011-2012 campaign. P-values (P), F values (F) and degree of freedom (Df) are indicated.
Number of Phenotypic Plant height Leaf length Leaf width Df leaves integration P F P F P F P F P F R. biternatus Shannon diversity 0,82 0.12 0,38 0.01 0,33 0,36 0,36 0.37 0.30 0.14 11 Species richness 0,92 0.09 0,072 4.20 0,35 0,38 0,39 0.32 0.62 0.25 11 R. pseudotrullifolius Shannon diversity 0.05 4.75 0,33 1.06 0,14 0,12 0.07 4.0 0.08 3.70 11 Species richness 0.09 3.37 0,14 2.0 0,089 0,06 0.09 3.4 0.13 2.20 11 R. moseleyi Shannon diversity 0,51 0.21 0,69 0.03 0,36 0,19 0,19 2.7 0.24 0.02 11 Species richness 0.70 0.42 0.70 0.15 0.16 2.75 0.07 4.31 0.86 0.03 11
97
Chapitre 1B. Exploring plant performance
R. biter atus R. oseleyi R. pseudotrullifolius
c c c b b c 100 b 100 b 100 a a b b b a a 80 80 80 b
a a 60 60 60
40 40 40 20 20 20
non-senescent (percentageleaves) of Performance 05m 153m 87m 153m 80m 180m VS PJDA PJDA PJDA P12 VS 80m 153m 231m
Figure 1. Mean performance (as the percentage of non-senescent leaves) in the different populations in R. biternatus, R. moseleyi and R. pseudotrullifolius. The region and the altitude of the populations are indicated. Red colour= dry population; blue colour= wet population. In each populations N=10. Within regions, a/b/c/d indicate locations that differ significantly (p<0.05) using ANOVA. Red = dry populations, blue = wet population. Abbreviations: PJDA: Port-Jeanne-D’Arc;
P12: Port-Douzième; VS: Val Studer.
98
Chapitre 1B. Exploring plant performance
Appendix 1. Altitudes and mean summer temperatures in the different sampled locations, and in the laboratory room.
Altitude Mean summer Region (m) Temperature (°C) Port-Jeanne-d'Arc 30 9.19 153 6.83 Port-Douzième 87 7.34 150 6.83 Val Studer 80 8.73 180 6.94 Laboratory room 0 10
99
Chapitre 2. Variations of secondary metabolites among natural populations suggest functional redundancy and versatility, the case of sub- Antarctic Ranunculus species
100
Chapitre 2. Variation of secondary metabolites
RESUME
Le métabolome est connu pour être influencé à la fois par la phylogénie et l’environnement. Le métabolome agit également sur le phénotype, qui a une action en retour sur le métabolome. Bien que la macroévolution du métabolome soit de mieux en mieux déterminée, sa microévolution est encore peu connue. Notamment, la variation intra-spécifique du métabolome dans la nature, intra ou inter-populations a été peu étudiée. La distribution géographique des populations pourrait être un facteur important agissant sur la variation inter-population du métabolome. La microévolution entre région peut affecter la relation entre les métabolites et l’environnement ou le phénotype. Chez trois espèces de Ranunculus, des populations naturelles ont été étudiées le long de deux gradients environnementaux similaires, situés dans deux régions géographiquement distantes des îles Kerguelen. Deux classes de métabolites, quercétines et amines, connues pour répondre aux stresses environnementaux et également, pour les amines, influencer le phénotype, sont étudiées. Nous avons montré que la composition des métabolites (amines et quercétines) varie en fonction de l’environnement, mais également indépendamment de l’environnement. Selon la région, des métabolites différents répondent à un même environnement ou influencent un même trait, suggérant une redondance fonctionnelle intra-spécifique. De plus, selon la région, on observe une relation différente, voire opposée, entre un métabolite donné et un environnement ou trait donné, suggérant une versatilité fonctionnelle intra-spécifique. Ces résultats suggèrent que l’environnement, mais également une microévolution neutre influencent la composition des métabolites à l’intérieur des espèces, et ce chez deux classes de métabolites différentes. De plus, nous suggérons une redondance et une versatilité fonctionnelle intra-spécifique, au niveau de leur réponse à l’environnement ou de leur relation avec le phénotype. Ces résultats apportent de nouvelles perspectives dans la compréhension des réponses évolutives des plantes à des changements environnementaux.
ABSTRACT
The metabolome is known to be shaped by both phylogeny and environment, and to interact with the phenotype. While macroevolution of the metabolome is being more and more deciphered, the microevolutionary differentiation among regions remains little studied. Specifically, nothing is known on whether microevolutionary differentiation among regions affects the relation between metabolites and environments or phenotypes. We studied populations of three Ranunculus species distributed across two similar environmental gradients in two distinct geographical regions in Iles Kerguelen. We used two metabolite classes, quercetins and amines, both known to respond to stresses and the latter also known to influence morphology. We showed both environment-dependent and environment- independent variation of metabolites (amines or quercetins) among populations. We showed that depending on regions, different metabolites may be related to the same environment or the same trait, suggesting metabolite redundancy within species. Moreover, depending on regions, a given metabolite may show different or even opposite relations with the same environment or the same trait, suggesting metabolite versatility within species. Our results suggest that non-selected microevolutionary differentiation contributes to shaping metabolite composition within species, for two metabolites classes. Moreover we suggest that metabolites may be functionally redundant and versatile within species, both in their response to environments and in their relation with the phenotype. These findings open new perspectives for understanding evolutionary responses of plants to environmental changes.
101
Chapitre 2. Variation of secondary metabolites
Variations of secondary metabolites among natural populations suggest
functional redundancy and versatility
– the case of sub-Antarctic Ranunculus species
Bastien Labarrere1*, Andreas Prinzing1, Emeline Chesneau1 & Françoise Hennion1
1 UMR 6553 Ecobio, Université de Rennes 1, CNRS, Av du Général Leclerc, F-35042 Rennes,
France, [email protected], *corresponding author, fax : (+33) 2 23 23 50 26
Keywords : Metabolites; amines; quercetins; environment; phenotype; redundancy; versatility; plants
102
Chapitre 2. Variation of secondary metabolites
INTRODUCTION
Plants are sessile organisms that have to face changes in their environment. The metabolome, being at the interface between plant phenotype (e.g. morphological traits) and its environment, plays a major role in sustaining environmental constraints. Metabolites are classified in two wide categories i.e. primary and secondary metabolites (Croteau et al., 2000). Primary metabolites are ubiquitous across lineages and essential to plant physiology, including for instance amino acids, carbohydrates or poly-alcohols (Croteau et al., 2000; Pichersky and Gang, 2000). Secondary metabolites were initially considered as waste products (Hartmann, 2007) before identifying their major roles in the interactions between the plant and its environment (Croteau et al., 2000, Groppa and Benavides, 2007). Just like metabolites in general, secondary metabolites play important roles in plant lifespan, responding to environment or influencing the phenotype, hence highly affecting individual fitness and survival (Croteau et al., 2000; Teuscher and Lindequist, 2010, Tiburcio et al., 2014). Secondary metabolites are involved in protection against environmental stresses, competition or attraction to pollinators (Croteau et al., 2000; Wahid and Ghazanfar, 2006, Wink
2013). Some secondary metabolites are also involved in vegetative and floral development in plants
(Ober, 2005; Tiburcio et al., 2014). Yet, while the understanding of the physiological functions of secondary metabolites in plants has increased in the past decades (Wink, 2013), detailed effects of secondary metabolites on plant morphological traits are rarely known (Ober, 2005).
Beyond responding to the environment, the composition of secondary metabolites in plants also evolved concomitantly to lineages and species as shown by recent macroevolutionary studies
(Fiehn, 2002; Hartmann, 2007; Wink et al., 2013; Tiburcio et al., 2014). Different secondary metabolites are hence restricted to different taxonomic groups (Croteau et al., 2000; Pichersky and
Gang, 2000) and the base levels of compounds depend on each species (Fiehn, 2002; Hennion et
103
Chapitre 2. Variation of secondary metabolites al., 2012). Thanks to development of molecular analyses the macroevolution of more and more metabolites is being deciphered (Hartmann, 2007; Tohge et al. 2013; Pathania et al., 2016; Peng et al., 2016; Xie et al., 2016). Yet, it still remains to determine how secondary metabolites evolve within species, across plants under both: abiotic environmental selection regimes and neutral differentiation among populations in different regions (Tohge et al. 2013; Sulmon et al. 2015).
Recently, Hennion et al. (2012) showed differences in amine composition among environments within species, and consistent across species, suggesting a possible link between the evolution of amine composition and the differentiation of lineages among environments.
In nature, plants form populations and plants within populations are in general more closely related while plants among populations may show stronger evolutionary differences. Several authors showed differences in metabolite composition among populations within species (Davey et al., 2009; Hennion et al., 2012; El-Backry et al., 2014). Yet, present environments may be more different among populations than within, and differences in metabolite composition among populations might hence be dependent or independent of present environments, but might reflect past environments. Only few authors relate metabolite differences among populations with abiotic environmental differences among populations (Davey et al., 2009; Hennion et al., 2012), dissociating metabolite variations that are dependent on present-day environments- and such that are independent. This latter part of variation in metabolite composition that would be independent of environmental factors would hence likely reflect neutral microevolutionary differentiation among populations or responses to past environmental selection pressures.
High diversity of secondary metabolites goes along with functional diversity (Croteau et al.,
2000; Pichersky and Gang, 2000; Wink, 2013). First, functional convergence is observed through metabolite macroevolution (Wink 2003; Tiburcio et al., 2014). Depending on lineage or species, different metabolites may respond to the same environment or be related to the same trait. So-called 104
Chapitre 2. Variation of secondary metabolites
“functional redundancy” is observed across all metabolites and plant lineages (Wink, 2003). Yet, functional redundancy of secondary metabolites might also well be observed among plants within species, but this remains to be investigated (Hanada et al., 2011) . Second, the inverse has also been observed: metabolite versatility, a given metabolite may have different roles within a given plant species in different organs or in different environments (e.g. Wink, 2003; Lehmann et al., 2010; Di
Ferdinando et al., 2014). However, whether redundancy or versatility might also emerge among populations that occupy similar environments but distant regions remains unknown. Such features would further increase the functional diversity of secondary metabolites. We hypothesize that, within species, neutral microevolutionary differentiation among distant but environmentally similar regions exists, and it includes the function of metabolites. Specifically, we hypothesize that functional redundancy or versatility exist within species among plant populations distributed in such distant regions. We predict that depending on the region, different metabolites might respond to the same environmental factor or affect the same trait in plants, suggesting functional redundancy. Moreover, depending on the region, the same metabolite might respond to different environmental factors or affect different traits in plants, suggesting functional versatility.
Amines and quercetins are two widespread groups of secondary metabolites. Amines include aliphatic polyamines, acetyl conjugates, and aromatic amines. Polyamines are low molecular weight polycationic molecules with amino groups (Tiburcio et al., 2014; Hennion et al., 2016).
Polyamines are ubiquitous in plants, involved in various internal processes such as growth control,
DNA replication and cell differentiation, organ development and leaf senescence, and described as
“growth regulators” (Bouchereau et al., 1999; Groppa and Benavides 2007). Polyamines are also known to respond to external environment (Alcazar et al., 2006) and are involved in the protection of plants against stresses such as drought, UV, salinity or heat (Tiburcio et al., 2014; Hennion et 105
Chapitre 2. Variation of secondary metabolites al., 2016). More characteristic of animals and bacteria, acetyl polyamines were described in several plant species (Hennion, Frenot & Martin Tanguy 2006; Lou et al., 2016). Acetyl polyamines were found to accumulate in response to several‐ abiotic or biotic stresses (Hennion, Frenot & Martin
Tanguy 2006; Hennion et al., 2016) and roles of protection against stress are now being deciphered‐
(Jammes et al., 2014; Lou et al., 2016). Aromatic amines are monoamines with either an arylalkyl or an indol radical, both widespread in plants (Smith, 1980). The physiological roles of aromatic amines remain less well understood in plants than that of polyamines, however many aromatic amines were also found to respond to abiotic stresses, and roles in development and growth were shown for some (see Hennion et al., 2016 and references therein). Quercetins are compounds with variable phenolic structures, they belong to flavonols among flavonoidsand are widely distributed across plant taxa (Agati et al., 2012) and were detected in our Ranunculus species of interest
(Hennion et al., 1994). Quercetins are mainly antioxidants and are involved in the protection of plants against abiotic stresses (Agati et al., 2012). The relationship between flavonoid (including quercetin) concentration and growth remains unclear (Treutter, 2005), albeit a trade-off is commonly suggested (Treutter, 2005). Overall, amines and quercetins include numerous various metabolites that respond to the environment and are related to the phenotype. Amines and quercetins are hence suitable compounds to analyse metabolite microevolution and their functional redundancy and versatility between regions within species.
We aim to determine differences in secondary metabolite composition or function in natural populations of plants located in distant but environmentally similar regions. This aim requires several populations of given species distributed across wide environmental gradients within each of multiple regions. These conditions are realized in sub-Antarctic Iles Kerguelen (Hennion et al.,
2012). Located in the southern Indian Ocean the Iles Kerguelen harbor a wide range of abiotic 106
Chapitre 2. Variation of secondary metabolites environmental conditions (Wagstaff and Hennion, 2007) and distinct regions across which plants are distributed. We know from Hennion et al. (2012) that amine composition of populations varies in relation to both species and the environment (9 species) in Iles Kerguelen. Previous work on flavonoids of the three Ranunculus species growing in Iles Kerguelen showed that quercetins were the only flavonols in these species and that composition of populations varied in relation to both species and the environment (Hennion et al., 1994).
We sampled populations of the three Ranunculus species native to Iles Kerguelen across environmental gradients in different geographical regions. We focused on environmental factors known to have the major effects on these plants (Hennion et al., 1994; Rott et al., 2006). We measured plant metabolite composition (amines and quercetins) and phenotype, and characterized the abiotic environments of populations. We asked whether (i) environment alone explains differences in metabolite composition among populations within species or whether environment- independent microevolutionary differentiation among regions also shapes metabolite composition,
(ii) metabolites show redundancy within species, different metabolites being related to the same environments or traits in the two different regions, (iii) metabolites show versatility within species, the same metabolites being related to different environments or traits in the two different regions.
107
Chapitre 2. Variation of secondary metabolites
MATERIALS AND METHODS
1. Plant collection-
The Iles Kerguelen (49°20’00” S, 69°20’00” E) are situated in the Southern Indian Ocean within the sub-Antarctic region (Lebouvier and Frenot, 2007) (Fig. 1). These islands are characterized by permanently low temperature (4.6°C annual mean), strong and permanent winds (10 m.s-1 annual mean), high precipitation (annual mean of 760 mm in the studied regions; Lebouvier and Frenot,
2007; Lebouvier et al., 2011).
We studied three Ranunculus species (i.e., R. biternatus, R. pseudotrullifolius and R. moseleyi), of which R. biternatus is austral circumpolar, R. pseudotrullifolius magellanic and on Kerguelen, and R. moseleyi is a strict Kerguelen endemic (Lehnebach et al., 2017). All are perennial plants.
On the Iles Kerguelen these species occupy different habitats (Hennion and Walton, 1997).
Ranunculus biternatus is widespread on the island occurring in habitats below 500m above sea level. In contrast, R. pseudotrullifolius and R. moseleyi have more restricted distributions. The first one, being halophilous, occurs within a short distance of the coast, occupying peaty or sandy shorelines and ponds (Hennion and Walton, 1997). Ranunculus moseleyi is strictly aquatic, growing only in freshwater lakes and ponds (Hennion and Walton, 1997).
Plants were sampled in 36 populations (12 populations per species) equally distributed across two regions in Iles Kerguelen: Isthme Bas, a large flat isthmus (about 30 km2) and Ile Australia, a large island (about 20 km2) (Fig. 1). Plants across the two regions were sampled at similar altitudes and subject to similar temperatures. We defined ‘population’ as a continuous group of plants living in a same site. Sampling was performed across a small area so as to ensure that plants assigned to a same population were subject to roughly similar environmental conditions.
108
Chapitre 2. Variation of secondary metabolites
The entire sampling was performed in summer during a short period, 6 weeks from the end of
December 2011 until early February 2012, to avoid variation in metabolite contents due to seasonal differences (F. Hennion et al., unpublished data). Plants were sampled between 11 hours and 17 hours to avoid bias from daily variation of metabolism (Tiburcio et al., 1990 ; Fujihara and
Yoneyama, 2001). Per population, an average of five individual plants, of the most frequent size in the local population were sampled. Per individual, we measured phenotype and collected two to four leaves to quantify amines and quercetins. Because metabolite composition may vary within a given plant at a given moment between leaves of different developmental stages (Foster and
Walters, 1991; Fujihara and Yoneyama, 2001), we sampled an appropriate and constant leaf developmental stage, i.e., young developing, fully photosynthetic leaves as in previous work (e.g.,
Hennion et al., 2006; Hennion et al., 2012). The samples collected were frozen in liquid nitrogen and stored at − 80 ° C then lyophilized and ground to powder.
2. Measurements-
2.1. Plant measurements
In each individual we measured plant height, the length and width of the largest leaf, the numbers of leaves and flowers, the flowering stage and the size of the largest flower (largest diameter). Flowering stage of the individuals was estimated following Hennion et al. (2012). To determine leaf dry matter content (LDMC), in each population a total of 20 leaves were sampled from a minimum of 15 individuals and processed following Cornelissen et al. (2003). Leaves were collected then directly put in distilled water in airtight bags for rehydration. They were then weighed before and after 48 hour-drying at 80°C. Some traits (flowering stage, number of flowers and the size of largest flower) brought highly redundant information. We thus only kept the
109
Chapitre 2. Variation of secondary metabolites
continuously varying trait ‘size of largest flower’ as floral trait. Likewise, we found low variability of LDMC among populations; hence we did not use this trait in our analysis.
2.2. Determination and quantitation of free amines and acetylated polyamines
This determination followed Hennion et al. (2012), with modified quantities as follows. Five- ten mg of powdered samples were thoroughly mixed with 100 to 200μL of 1 mmol.L-1 HCl supplemented with 10 μmol.L-1 heptanediamine, as an internal standard, on a magnetic stirring plate (2000 rpm) for 1 h at 4 ° C. The homogenates were then centrifuged for 15 min at 10000g at
4 °C, and the three supernatants collected. The pellets were further extracted twice with 100 to 200
μL of 1 mmol.L-1 HCL and 10 μmol.L-1 heptanediamine. After short stirring, the homogenates were centrifuged for 15 min at 10 000g at 4°C. The combined supernatants were used as the crude extracts for characterization and determination of free and acetylated amines and polyamines and stored frozen at − 20 °C before chromatographic analyses. HPLC and fluorescence spectrophotometry were used to separate and quantify amines prepared as their dansyl derivatives according to Smith and Davies (1985) with some modifications as follows. Aliquots (200 μL) of the supernatant were added to 200 μL of saturated sodium carbonate and 600 μL of dansyl chloride in acetone (7.5 mg.mL-1) in a 5mL tapered reaction vial. After a brief vortexing, the mixture was incubated in darkness at room temperature for 16 h. Excess dansyl chloride was converted to dansylproline by 30 min incubation after adding 300 μL (150 mg.mL-1) of proline. Dansylated amines were extracted in 1 mL ethylacetate. The organic phase was collected then evaporated to dryness, and the residue was dissolved in methanol and stored in glass vials at − 20 °C. Standards were processed in the same way, and two to 50 nmol (per assay) were dansylated for each standard alone or in combination. Free and acetylated amines were quantified after yield correction with the internal standard and calibration with external standards. The HPLC column was packed with reverse phase SpherisorbODS2 C18 (particle size 5 μm; 4.6 × 250 mm, Waters, Milford, USA). 110
Chapitre 2. Variation of secondary metabolites
The mobile phase consisted of a solution of 17.5 mmol.L-1 potassium acetate (pH 7.17) as eluent
A and acetonitrile as eluent B. The solvent gradient, modified according to Hayman et al. (1985) was as described by Jubault et al. (2008). The flow rate of the mobile phase was 1.5 mL.min-1. For fluorescence detection of dansyl amines, an excitation wavelength of 366 nm was used with an emission wave length of 490 nm. Identities of peaks of amines were confirmed by spiking the sample with known amounts of authentic standards (Sigma, St. Louis, Missouri, USA). The HPLC design consisted of a thermoelectron pump (SpectraSystem P1000 XR, Thermo Fisher, San Jose,
California, USA) and (Spectra-Series AS100) autosampler with a 20 μL injection loop, and detection through an FP-2020 Plus fluorometer (Jasco, Inc., Easton, Maryland, USA). Signals were computed and analyzed using Azur software (Datalys, St Martin d’Hères, France).
2.3. Determination and quantitation of quercetins
Quercetins were the sole flavonols detected in Ranunculus species from Iles Kerguelen
(Hennion et al., 1994). We weighed about 10 mg of plant powder in an Eppendorf tube, and added
1 mL of methanol acidified with 1% formic acid. The tube was vortex-agitated first and put in ultra-sonic bath for 5 min. The tube was then centrifuged briefly and 900 µL of the supernatant was removed using a 1mL plastic syringe, and filtered using a PTFE 13 mm 0.45 µm syringe filter.
The methanol extract was then poured in an injection vial for UPLC analysis; 2 µL of the extract were injected in the Waters UPLC_PDA_ESI_TQD system for flavonol quantitation. The reversed phase column, an Acquity Waters C18 BEH (2.1 x 150 mm) 1.7 µm, was maintained at 30°C.The solvents used for the binary gradient were A: ultra-pure water with 0.1 % formic acid, B: acetonitrile with 0.1 % formic, the flow was 0.4 mL/min. The gradient applied was 98% A from 0 to 0.2 min, 10% A from 0.2 to 14 min, 14 to 15 min isocratic 10% A, 15 min to 17 min 98% A, 17 to 20 min isocratic 98% A. The photo diode array detector scanned from 190 to 600 nm and 111
Chapitre 2. Variation of secondary metabolites flavonols were detected at 350 nm, external quantitation with some flavonol standards was applied.
The identity or structure of flavonols were confirmed with the triple quadrupole mass detector in full scan negative mode or targeted fragmentation, the capillary voltage was 2.9 kV, the cone voltage was 37 V, the source temperature was maintained at 150°C and the desolvatation temperature at 400°C, the desolvatation gaz flow was 800 L/h.
2.4. Amines characterized
We characterized 15 different amines which belonged to four biochemical categories: aliphatic amines and their acetylated conjugates, phenylalkylamines and indolalkylamines. The detected aliphatic amines were: agmatine (Agm), diaminopropane (DAP), putrescine (Put), cadaverine
(Cad), spermidine (Spd), spermine (Spm), N8-acetylspermidine (N8Ac-Spd) and N1-acetylspermine
(N1Ac-Spm). Phenylalkylamines were phenylethylamine (Phe), octopamine (Oct), 3-methoxy-4- hydroxy phenylethylamine (3M4OHPhe), tyramine (Tyr), and dopamine (Dop). Indolalkylamines were tryptamine (Try) and serotonin (Ser). All 15 compounds described were present in the three species, in the two regions. Thus, differences in amine composition between species or regions reflected shifts in concentrations and not qualitative differences.
2.5. Quercetins characterized
Following Hennion et al. (1994), we performed an analysis of flavonols. Quercetins characterized were: quercetin 3-diglucoside-7-glucoside (Q-3GL), quercetin 3-(caffeyl-glucosyl)glucoside-7- glucoside (Q-3GL+Caf), quercetin 3-(ferulyl-glucosyl)glucoside-7-glucoside (Q-3GL+Fer), quercetin 3-(caffeyl-xylosyl)glucoside-7-glucoside (Q-2GL+Xyl+Caf), quercetin 3-(ferulyl- xylosyl)glucoside-7-glucoside (Q-2GL+Xyl+Fer), quercetin 3-xylosylglucoside-7-glucoside (Q-
2GL+Xyl), quercetin 3-xylosylglucoside (Q-GL+Xyl), quercetin 3-diglucoside (Q-2GL) and 112
Chapitre 2. Variation of secondary metabolites isoquercitrin (IQC). All 9 compounds described were present in the three species, in the two regions. Thus, differences in quercetin composition between species or regions reflected shifts in concentrations and not qualitative differences.
2.6. Environmental measurements
In each population we measured soil water saturation, pH and conductivity. Three samples of soil, each of 20mL, were collected at the rhizosphere level of the measured plants. To determine soil water saturation, half of each soil sample was dried at 105°C during 48 hours and weighed before and after drying (Hermant et al., 2013). The remainder soil was mixed with known volume of distilled water and then was left 18 to 24 hours to permit sedimentation of soil particles. Readily after, pH was determined using a pH meter (BASIC 20 PLUS CRISON, resolution 0.01pH). After another 18 to 24 hours of sedimentation, conductivity was determined using a conductivity meter
(CONSORT K810, resolution 0.1 µS cm-1) (Hermant et al., 2013).
3. Statistical analyses
We related to determine differences of total contents of metabolite (amines or quercetins) among populations that are no due to environmental differences among populations we conducted
ANOVA analyses with the environment as a co-variable using in R 3.0.2 software (R Core Team,
2012). To determine differences of total metabolite concentrations (amines or quercetins) among populations that due to environmental differences among populations we regressed concentration against environmental predictors. To determine the relationships between compositions of metabolite (amines or flavonoids) and environment or phenotype we used redundancy analyses, using cca function (Ter Braak, 1986; Legendre and Legendre, 2012) in R 3.0.2 software (R Core
Team, 2012). We determined relationships between metabolite composition and single 113
Chapitre 2. Variation of secondary metabolites environmental factors or single traits using redundancy analyses with rda function. Also we determined relationships between metabolite composition and the overall environment (i.e. taking into account simultaneously the different environmental factors) or the overall phenotype (i.e. taking into account simultaneously the different traits) using redundancy analyses with rda function.
To determine whether metabolite-environment or metabolite-phenotype relationships differ between regions, we tested an interaction term of region in the relationships between total metabolite contents and either (i) the environment or (ii) the phenotype among populations across regions.
For multiple testing on the same data set, p-values were corrected using sequential Bonferroni’s correction (Cabin and Mitchell, 2000). In all regression analyses we verified the assumptions of the analyses using residual plots.
114
Chapitre 2. Variation of secondary metabolites
RESULTS
1. Total contents of metabolites (Amines or Quercetins)
1.1 Populations differ in total metabolite contents independently from environment
In order to identify an effect of population on total metabolite contents that was not due to environmental differences among populations we conducted an ANOVA with environmental conditions as co-variables. We found that total contents of amines or quercetins differed significantly among populations within and across regions, i.e. independent of the environment
(Table 1).
1.2. Environments partly explain variation of total metabolite contents across populations
Total amine content increased with conductivity in R. biternatus, R. pseudotrullifolius and R. moseleyi (Table 2). In R. biternatus and R. pseudotrullifolius, total amine contents also increased with respectively soil water saturation and pH (Table 2).
Total quercetin contents decreased with soil water saturation in R. biternatus. In R. pseudotrullifolius, total quercetin contents increased with soil water saturation and decreased with pH and conductivity (Table 2). In R. moseleyi, total quercetin contents increased with soil water saturation and pH (Table 2).
1.3. Total metabolite contents partly explain variation of phenotypes across populations
Total amine content was positively related with plant height in R. biternatus, and negatively in R. moseleyi (Table 3). Total amine content was negatively related to the number of leaves in R. biternatus.
115
Chapitre 2. Variation of secondary metabolites
Total quercetin content was negatively related to plant height in R. pseudotrullifolius (Table 3).
Total quercetin content was positively related to the number of leaves in R. biternatus and negatively in R. pseudotrullifolius (Table 3).
1.4. Differences between regions in metabolite-environment relationships
We showed significant relationships between total amine or quercetin content and environment within regions (Table 2). We found a significant effect of regions on the amine-environment or quercetin-environment relationships in R. pseudotrullifolius and R. moseleyi (Table 6).
1.5. Differences between regions in metabolite-phenotype relationships
We found significant relationships between total amine or quercetin content and the phenotype within regions (Table 3). We found a significant effect of regions on the amine-phenotype or quercetin-phenotype relationships in R. pseudotrullifolius and R. moseleyi (Table 6).
2. Metabolite composition (amines or quercetins)
2.1. Populations differ in metabolite composition partly independently from environment
In order to identify an effect of population on metabolite composition that was not due to environmental differences among populations we conducted redundancy analyses with environmental conditions as co-variables. We found that metabolite (amines or quercetins) composition differed significantly among populations across regions within species, i.e independently of the environment (Table 7).
2.2. Environment partly explain variation of metabolite compositions across populations
116
Chapitre 2. Variation of secondary metabolites
Amine composition was significantly related to overall environment in R. biternatus and R. pseudotrullifolius (respectively N = 58 and 57, p=0.003 ; r²=0.26 and p<0.001 ; r²=0.33), contrary to R. moseleyi (N = 47, p>0.05 ; r²=0.096). Testing the relationships between amine composition and single environmental factors, we found that soil water saturation and conductivity significantly influence amine composition in R. biternatus (p=0.013; r²=0.071 and p=0.015; r²=0.08) and in R. pseudotrullifolius (p=0.004; r²=0.084 and p=0.002; r²=0.22). Amine composition is also influenced by soil pH in R. pseudotrullifolius (p=0.014; r²=0.078). ). The relationship between a given compound and a given environmental variable differed among species (Fig. 2).
Quercetin composition was significantly related to overall environment in R. moseleyi
(p=0.002; r²=0.35) and not in R. biternatus or R. pseudotrullifolius (p>0.05). No single environmental factor influenced significantly the composition in quercetins (Table 4).
2.3. Metabolite compositions partly explain variation of phenotypes across populations
Amine composition was significantly related to the phenotype in R. pseudotrullifolius and R. moseleyi in contrast to R. biternatus (Table 5). Traits influenced by amine composition were mainly plant height and the number of leaves. Different compounds are related to different traits (Fig. 4) and the relationships between a given compound and a given trait differed among species (Fig. 4).
Quercetin composition was significantly related to the phenotype in R. biternatus, R. pseudotrullifolius and R. moseleyi (Table 5). Traits influenced by quercetin composition were mainly plant height and the number of leaves. Different compounds were related to different traits
(Fig. 5) and the relationships between a given compound and a given trait differed among species
(Fig. 5).
117
Chapitre 2. Variation of secondary metabolites
2.4. Differences between regions in metabolite-environment relationships
Amine composition was significantly related to the environment within each of the two regions studied in R. pseudotrullifolius and R. moseleyi (Table 4). In R. biternatus, the relation was significant in Isthme Bas and marginally significant in Ile Australia (Table 4). We found relationships between amine composition and particular environmental factors in R. biternatus, R. pseudotrullifolius and R. moseleyi (Table 4; Fig. 6). The relationships between a given compound and a given environmental variable differed among regions within species, some even being opposite (Figs. 2 and 6). Although not always being significant, environment mostly explained distinctly higher variances in amine composition within than across regions (Table 4; Fig. 2).
The relationship between quercetin composition and the environment was significant in Ile
Australia but not in Isthme Bas in R. biternatus and R. pseudotrullifolius (Table 4) while we observed the opposite in R. moseleyi. We found relationships between quercetin composition and particular environmental factors in R. biternatus and R. pseudotrullifolius (Table 4). The relationships between a given compound and a given environmental variable differed among regions within species, some even being opposite (Fig. 3). Although not always being significant, environment mostly explained distinctly higher variances in quercetin composition within than across regions (Table 4; Fig. 3).
2.5. Differences between regions in metabolite-phenotype relationships
Amine composition was significantly related to the phenotype in Ile Australia in R. pseudotrullifolius and R. moseleyi (Table 5). In R. biternatus, and R. pseudotrullifolius we found a significant relationship in Isthme Bas (Table 5; Fig. 7). We found relationships between amine composition and particular traits in R. biternatus, R. pseudotrullifolius and R. moseleyi (Table 5;
118
Chapitre 2. Variation of secondary metabolites
Fig. 7). The relationships between a given compound and a given trait differed among regions within species, but no opposite relationship was observed (Fig. 4). Although not always significant, phenotype mostly explained distinctly higher variances in amine composition within than across regions (Table 5; Fig. 4).
Quercetin composition was significantly related to the phenotype in R. pseudotrullifolius and R. moseleyi in Ile Australia (Table 5). We also found a significant relationship in R. pseudotrullifolius in Isthme Bas. We found relationships between quercetins and traits in R. biternatus, R. pseudotrullifolius and R. moseleyi (Fig. 5). The relationships between a given compound and a given trait differed among regions within species, but no opposite relationship was observed (Fig.
5). Although not always significant, phenotype mostly explained distinctly higher variances in quercetin composition within than across regions (Table 5; Fig. 5).
119
Chapitre 2. Variation of secondary metabolites
DISCUSSION
While the macroevolution of secondary metabolites is being more and more deciphered, their microevolution at the intraspecific level remains far more obscure. We investigated how secondary metabolites vary at the intraspecific level, where plants are subject to both environmental constraints and neutral microevolution. We hypothesized that microevolution within species could affect metabolite composition but also metabolite functions – i.e. the way metabolites respond to the environment or influence the phenotype. In a multi-species study, we investigated two secondary metabolite families (amines and quercetins) among natural populations in sub-Antarctic Iles Kerguelen. Populations were distributed across similar environmental gradients in two distinct regions. This pattern allows to determine differences between regions in environment-metabolite or trait-metabolite relationships. We found that variation of amine or quercetin composition among populations or regions is not only shaped by environment, suggesting that neutral microevolution, or evolution under past environmental selection pressures, also shapes metabolite composition within species. Moreover, we showed that depending on regions, different metabolites may be involved in the same function, which supports the hypothesis of metabolite redundancy within species. Also, depending on regions, a given metabolite may show different or even opposite functions, which supports the hypothesis of metabolite versatility within species. Our results hence suggest that microevolution within species shapes secondary metabolite composition but also influences metabolite functions through metabolite redundancy and versatility.
Overall relation between total metabolite contents and environment or traits
Amine and quercetin compositions found in the three species were consistent with previous work (Hennion et al. 1994, 2012). We know from respectively Hennion et al. (2012) and
120
Chapitre 2. Variation of secondary metabolites
Hennion et al. (1994) that amine or quercetin compositions in natural populations of
Ranunculus species from Iles Kerguelen differ among environments and species. Here, we deciphered functions of amines and quercetins in natural plant populations, determining that amine or quercetin compositions were related both to environmental factors and traits within species. In particular, we showed in a multi-species analysis in a natural environment that amines and quercetins have an opposite effect on the phenotype, the former relates to increased growth, the latter to decreased growth.
Amines are known to be involved in plant growth and development but also to respond to external environmental stresses (Groppa and Benavides 2007; Alcazar et al., 2010; Tiburcio et al., 2014). Yet, few studies assessed both amine response to environment and their relation with the phenotype in nature (Hummel et al. 2004). Here we showed that amines are related to both environment and plant phenotype in natural populations. Moreover, we showed that amines increase with size of the overall plant and its organs in natural populations. Yet, relations between traits and amines differ among species.
Quercetins are involved in plant response to stress (Agati et al., 2012). Yet, the relationships, if any, between quercetin contents and plant trait values still remain unclear, although it has been suggested that quercetins may limit growth (Treutter, 2005). Here we showed that quercetins are related to both environment and plant phenotype in natural populations. For instance, quercetin 3-diglucoside-7-glucoside levels decrease with soil water content and are negatively related to plant height. Moreover, we found that different quercetins were related to different modules in plants such as vegetative traits or floral traits and a plant ratio such as leaf shape. Also, in R. moseleyi quercetin contents increased with leaf dry matter content (LDMC) which is known to be a proxy of stress tolerance in plants (Vaieretti et al., 2007). Hence, we suggest a possible link between quercetin contents and plant stress tolerance. In addition, we found that size of the overall plant and its organs are negatively related to quercetin contents,
121
Chapitre 2. Variation of secondary metabolites hence supporting the hypothesis of a trade-off between quercetin contents and plant growth as suggested in Treutter (2005).
Differences in composition among populations
We found variation of amine or quercetin composition among populations or regions within species. Only few authors differentiate metabolite variation that is environment-dependent and metabolite variation that is environment-independent (Davey et al., 2009; Hennion et al., 2012).
Here, we showed that the different populations within species have the same pool of compounds but differ in the total contents of the amines or quercetins and in the levels of the different amines or quercetins. Moreover, we showed that such differences in amine or quercetin compositions among populations are only partially explained by the environment, suggesting that also environment-independent i.e. either neutral microevolution or evolution under other selection pressures than present-day environments, shapes amine or quercetin composition within plant species. Note that the amount of the variation explained by environment is different between species but also between regions within species, which we discuss below.
Differences in relationship of metabolites with environment or traits between regions
(1) Metabolite redundancy-
We found that the relationships between secondary metabolites and environments or traits differed among regions within species. Hence, depending on regions, the same environmental factor or the same trait was related to different metabolites. That different metabolites have the same function in plants growing in different regions suggests a functional redundancy among metabolites within species. We found this pattern in two metabolite families (amines and quercetins). Metabolite redundancy among plants within species has been little studied so far
(Sulmon et al., 2015). This concept has been documented more at the interspecific level. Groppa
122
Chapitre 2. Variation of secondary metabolites and Benavides (2007) reported that different secondary metabolites responded to given environments among species across studies - even if many authors emphasized that different studies are often hardly comparable.
Both in amine and quercetin families, the compounds are closely related one to each other, share metabolic pathways and may derive one from each other (Treutter, 2005; Tiburcio et al.,
2014). Moreover, metabolites belonging to the same family often show similar or at least overlapping functions in plants (Treutter, 2005; Tiburcio et al., 2014; Hennion et al., 2016).
Developmental and functional proximity among metabolites belonging to the same family may provide a possible mechanism for the rise of metabolite redundancy.
Overall, we found evidence for functional redundancy of metabolites within species i.e. the same function is related to different metabolites in different regions. Still, it remains to understand the causes of such redundancy. All the metabolites are found in all the populations studied across regions: hence, differences in metabolite-environment or metabolite-trait relationships between regions are not due to the presence or absence of different metabolites between regions. In contrast, each population potentially synthesizes several compounds that are not used to respond to environment or to affect traits. Producing unused compounds may have a cost in plants and it remains to understand costs and benefits of metabolite redundancy.
Finally, it is possible that metabolites might also be redundant among populations within regions, which could explain some non-significant relationships found between given metabolites and given environments across populations within regions.
(2) Metabolite versatility-
Secondary metabolites may be involved in different functions in different organs or in different environments within the same plant species, i.e. metabolite versatility (Wink, 2013;
Lehmann et al., 2010; Di Ferdinando et al., 2014). Here we extended the concept of metabolite
123
Chapitre 2. Variation of secondary metabolites versatility, showing that the same metabolites may have different functions in different regions subject to a similar environment. Surprisingly, a given metabolite may even have opposite functions in different regions, i.e. depending on regions, the same metabolite may either increase or decrease with a given environmental factor or trait. Due to different or opposite effects of the same metabolite in different regions, an analysis across regions comes to the false conclusion that this metabolite likely has no effect at all. We do not exclude that versatility may also exist among populations within regions. As suggested across regions, we propose that some metabolite-environment relationships across populations within regions falsely appear to be unrelated due to metabolite versatility among populations. Hence, we strongly emphasize the necessity to take into account versatility while analyzing metabolite functions. Overall, the capacity of a metabolite to have different roles increases the capacity of a species to respond to the environment. This is all the more important in a given species, with only a limited pool of metabolites at its disposal.
Metabolite redundancy and versatility as a result of microevolution driven by distance rather than environment
Both regions had roughly the same environmental gradients. Hence we may probably exclude the environmental differences as cause of metabolite redundancy and versatility. We recognize that metabolite-environment relationships found across populations within regions may result from either metabolite plasticity or adaptation among populations within regions.
Still, we have indication from previous work that some part of the amine composition of plants in natural populations is heritable or persistent for a few years (Hennion et al., 2012; Hennion et al., 2016). Future research should try to dissociate such plasticity and adaptation processes.
Differences in the patterns of variation between distant but environmentally similar regions suggest neutral microevolution between regions. It should be noted that differences may also
124
Chapitre 2. Variation of secondary metabolites be, in theory, due to microevolution under selection pressures other than present-day environments, i.e. past environmental conditions. However, there is some indication for persistence of present-day sub-Antarctic physical and climatic conditions in Kerguelen since last glacial maximum (Van Der Putten et al., 2010). Further studies may determine whether differences of relationships between regions are heritable or not. Note that populations in the three Ranunculus species exhibit important rates of self-pollination and vegetative reproduction
(Hennion et al., 1994). Self-pollination and vegetative reproduction within populations may reduce gene flows across populations and therefore may increase microevolutionary differentiation among populations. We suggest that species’ reproduction modes induce high microevolutionary distances among populations and may favor redundancy and versatility.
Further studies may aim to determine factors that may increase or decrease metabolite redundancy and versatility in plants.
CONCLUSION
Overall, our study helps to understand how secondary metabolites evolve within species. We showed that variation of secondary metabolite composition among populations was only partially related to environment, suggesting that neutral microevolution also shapes metabolite composition within species. We showed differences in metabolite-environment and metabolite- trait relations among regions. Hence, we suggest metabolite redundancy and versatility among regions within species. Metabolite redundancy and versatility are found between regions with similar environments suggesting that both are shaped by neutral microevolution. We emphasize that metabolite redundancy and versatility within species contribute to the high functional diversity of secondary metabolites, which constitute an impressive palette of tools for plants to sustain environment and its changes. We found patterns suggesting metabolite redundancy and versatility in all three species studied, in two distinct families of secondary metabolites (i.e.
125
Chapitre 2. Variation of secondary metabolites amines and quercetins), related to all environmental parameters and all morphological variables. Hence, the observed patterns of metabolite redundancy and versatility likely do not result from a special case but may be extendable to other species, locations or secondary metabolite families. Further studies may aim to understand why particular compounds only have effects in particular regions, and why a given compound may have different functions in the same environment. Also, further studies may investigate costs and benefits of metabolite redundancy and versatility.
126
Chapitre 2. Variation of secondary metabolites
Acknowledgements
Field work at Kerguelen was supported by IPEV (programmes 136 Ecobio, M. Lebouvier and
1116 PlantEvol, F. Hennion). We thank volunteers Marine Pouvreau, Françoise Cardou, Julie
Vingère and Marion Lombard, Richard Winkworth (IFS, Massey University, New Zealand)
Philippe Choler (LECA, Grenoble, France), the IPEV logistics team, and Réserve Naturelle
TAAF for help during the 2011-2012 summer campaign. We thank Nathalie Marnet who performed the flavonoid analyses at P2M2 facility (INRA, Le Rheu). This research is linked to
CNRS Zone-Atelier Antarctique, to CNRS LIA “AntarctPlantAdapt” (F.H.) with New Zealand, and the Scientific Committee on Antarctic Research programmes AntEco and AnT-ERA. B.L. was supported by a PhD grant from Ministry of Research and Education (France).
127
Chapitre 2. Variation of secondary metabolites
REFERENCES Agati, G., Azzarello, E., Pollastri, S., & Tattini, M. (2012). Flavonoids as antioxidants in plants: location and functional significance. Plant Science, 196, 67-76. Alcázar, R., Marco, F., Cuevas, J. C., Patron, M., Ferrando, A., Carrasco, P., ... & Altabella, T. (2006). Involvement of polyamines in plant response to abiotic stress. Biotechnology letters, 28(23), 1867-1876. Alcázar, R., Altabella, T., Marco, F., Bortolotti, C., Reymond, M., Koncz, C., ... & Tiburcio, A. F. (2010). Polyamines: molecules with regulatory functions in plant abiotic stress tolerance. Planta, 231(6), 1237-1249. Bouchereau, A., Aziz, A., Larher, F., & Martin-Tanguy, J. (1999). Polyamines and environmental challenges: recent development. Plant Science, 140, 103-125. Cabin R.J. & Mitchell R.J. (2000) To Bonferroni or not to Bonferroni: when and how are the questions. Bulletin of the Ecological Society of America, 81, 246-248. Cornelissen, J. H. C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D. E., ... & Poorter, H. (2003). A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian journal of Botany, 51(4), 335-380. Croteau, R., Kutchan, T. M., & Lewis, N. G. (2000). Natural products (secondary metabolites). Biochemistry and molecular biology of plants, 24, 1250-1319. Davey, M. P., Woodward, F. I., & Quick, W. P. (2009). Intraspecfic variation in cold-temperature metabolic phenotypes of Arabidopsis lyrata ssp. petraea. Metabolomics, 5(1), 138-149. Di Ferdinando, M., Brunetti, C., Agati, G., & Tattini, M. (2014). Multiple functions of polyphenols in plants inhabiting unfavorable Mediterranean areas. Environmental and Experimental Botany, 103, 107-116. El-Bakry, A. A., Hammad, I. A., & Rafat, F. A. (2014). Polymorphism in Calotropis procera: preliminary genetic variation in plants from different phytogeographical regions of Egypt. Rendiconti Lincei, 25(4), 471-477. Fiehn, O. (2002). Metabolomics–the link between genotypes and phenotypes. Plant molecular biology, 48(1-2), 155-171. Foster, S. A., & Walters, D. R. (1991). Polyamine concentrations and arginine decarboxylase activity in wheat exposed to osmotic stress. Physiologia Plantarum, 82(2), 185-190. Fujihara, S., & Yoneyama, T. (2001). Endogenous levels of polyamines in the organs of cucumber plant (Cucumis sativus) and factors affecting leaf polyamine contents. Physiologia plantarum, 113(3), 416-423. Groppa, M. D., & Benavides, M. P. (2008). Polyamines and abiotic stress: recent advances. Amino acids, 34(1), 35-45. Hanada, K., Sawada, Y., Kuromori, T., Klausnitzer, R., Saito, K., Toyoda, T., ... & Hirai, M. Y. (2011). Functional compensation of primary and secondary metabolites by duplicate genes in Arabidopsis thaliana. Molecular biology and evolution, 28(1), 377-382. Hartmann, T. (2007). From waste products to ecochemicals: fifty years research of plant secondary metabolism. Phytochemistry, 68(22), 2831-2846. Hayman, A. R., Gray, D. O., & Evans, S. V. (1985). New high-performance liquid chromatography system for the separation of biogenic amines as their Dns derivatives. Journal of Chromatography A, 325, 462-466.
128
Chapitre 2. Variation of secondary metabolites
Hennion, F., Fiasson, J. L., & Gluchoff-Fiasson, K. (1994). Morphological and phytochemical relationships between Ranunculus species from Iles Kerguelen. Biochemical systematics and ecology, 22(5), 533-542. Hennion F. & Walton D.W.H. (1997) Ecology and seed morphology of endemic species from Kerguelen phytogeographic zone. Polar Biology, 18, 229-235. Hennion, F., Frenot, Y., & Martin‐Tanguy, J. (2006). High flexibility in growth and polyamine composition of the crucifer Pringlea antiscorbutica in relation to environmental conditions. Physiologia Plantarum, 127(2), 212-224. Hennion, F., Bouchereau, A., Gauthier, C., Hermant, M., Vernon, P., & Prinzing, A. (2012). Variation in amine composition in plant species: How it integrates macroevolutionary and environmental signals. American journal of botany, 99(1), 36-45. Hennion, F., Litrico, I., Bartish, I. V., Weigelt, A., Bouchereau, A., & Prinzing, A. (2016). Ecologically diverse and distinct neighbourhoods trigger persistent phenotypic consequences, and amine metabolic profiling detects them. Journal of Ecology, 104(1), 125- 137. Hermant, M., Prinzing, A., Vernon, P., Convey, P., & Hennion, F. (2013). Endemic species have highly integrated phenotypes, environmental distributions and phenotype–environment relationships. Journal of Biogeography, 40(8), 1583-1594. Hummel, I., El Amrani, A., Gouesbet, G., Hennion, F., & Couée, I. (2004). Involvement of polyamines in the interacting effects of low temperature and mineral supply on Pringlea antiscorbutica (Kerguelen cabbage) seedlings. Journal of experimental botany, 55(399), 1125-1134. Jammes, F., Leonhardt, N., Tran, D., Bousserouel, H., Véry, A. A., Renou, J. P., ... & Leung, J. (2014). Acetylated 1, 3‐diaminopropane antagonizes abscisic acid‐mediated stomatal closing in Arabidopsis. The Plant Journal, 79(2), 322-333. Jubault, M., Hamon, C., Gravot, A., Lariagon, C., Delourme, R., Bouchereau, A., & Manzanares- Dauleux, M. J. (2008). Differential regulation of root arginine catabolism and polyamine metabolism in clubroot-susceptible and partially resistant Arabidopsis genotypes. Plant Physiology, 146(4). Lebouvier M. & Frenot Y. (2007). Conservation and management in the French sub-Antarctic islands and surrounding seas. Lebouvier, M., Laparie, M., Hulle, M., Marais, A., Cozic, Y., Lalouette, L., ... & Renault, D. (2011). The significance of the sub-Antarctic Kerguelen Islands for the assessment of the vulnerability of native communities to climate change, alien insect invasions and plant viruses. Biological Invasions, 13(5), 1195-1208. Legendre, P., & Legendre, L. F. (2012). Numerical ecology (Vol. 24). Elsevier. Lehmann, S., Funck, D., Szabados, L., & Rentsch, D. (2010). Proline metabolism and transport in plant development. Amino acids, 39(4), 949-962. Lehnebach, C., Winkworth, R.C., Becker, M., Lockhart, P.J. & Hennion, F. (2017) Around the pole: evolution of sub-Antarctic island Ranunculus. Journal of Biogeography early view 15th February 2017, DOI: 10.1111/jbi.12952. Lou, Y. R., Bor, M., Yan, J., Preuss, A. S., & Jander, G. (2016). Arabidopsis NATA1 acetylates putrescine and decreases defense-related hydrogen peroxide accumulation. Plant physiology, pp-00446. Ober, D. (2005). Seeing double: gene duplication and diversification in plant secondary metabolism. Trends in plant science, 10(9), 444-449.
129
Chapitre 2. Variation of secondary metabolites
Pathania, S., Bagler, G., & Ahuja, P. S. (2016). Differential network analysis reveals evolutionary complexity in secondary metabolism of Rauvolfia serpentina over Catharanthus roseus. Frontiers in Plant Science, 7. Peng, M., Gao, Y., Chen, W., Wang, W., Shen, S., Shi, J., ... & Jian, W. (2016). Evolutionarily Distinct BAHD N-acyltransferases are Responsible for Natural Variation of Aromatic Amine Conjugates in Rice. The Plant Cell, tpc-00265. Pichersky, E., & Gang, D. R. (2000). Genetics and biochemistry of secondary metabolites in plants: an evolutionary perspective. Trends in plant science, 5(10), 439-445. Rott, E., Gross, E., & Schwienbacher, E. (2004). Small-scale heterogeneity of Ranunculus trichophyllus in Lake Tovel (microhabitat, morphology, phenolic compounds and molecular taxonomy). Acta Biologica, 81(Supplemento 2), 359-367.Samal A., Wagner A. & Martin O.C. (2011) Environmental versatility promotes modularity in genome-scale metabolic networks. BMC systems biology, 5, 1. Smith, T. (1980) Plant amines. Secondary plant products (eds B. EA. & C. BV.), pp. 433-456. Springer-Verlag, Berlin, D. Smith, M. A., & Davies, P. J. (1985). Separation and quantitation of polyamines in plant tissue by high performance liquid chromatography of their dansyl derivatives. Plant Physiology, 78(1), 89-91. Sulmon, C., Van Baaren, J., Cabello-Hurtado, F., Gouesbet, G., Hennion, F., Mony, C. & Gérard, C. (2015). Abiotic stressors and stress responses: What commonalities appear between species across biological organization levels? Environmental Pollution, 202, 66-77. Teuscher, E., & Lindequist, U. (2010). Biogene Gifte: Biologie-Chemie-Pharmakologie- Toxikologie. Wissenschaftl. Verlag-Ges. Ter Braak, C. J. (1986). Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology, 67(5), 1167-1179. Tiburcio, A. F., Kaur-Sawhney, R., & Galston, A. W. (1990). Polyamine metabolism. Miflia BJ, Lea P J. The Biochemistry of Plants, Intermediary Nitrogen Metabolism, 16, 283-285. Tiburcio, A. F., Altabella, T., Bitrián, M., & Alcázar, R. (2014). The roles of polyamines during the lifespan of plants: from development to stress. Planta, 240(1), 1-18. Tohge, T., Watanabe, M., Hoefgen, R., & Fernie, A. R. (2013). The evolution of phenylpropanoid metabolism in the green lineage. Critical reviews in biochemistry and molecular biology, 48(2), 123-152. Treutter, D. (2005). Significance of flavonoids in plant resistance and enhancement of their biosynthesis. Plant Biology, 7(06), 581-591. Vaieretti, M. V., Díaz, S., Vile, D., & Garnier, E. (2007). Two measurement methods of leaf dry matter content produce similar results in a broad range of species. Annals of Botany, 99(5), 955-958. Van der Putten, N., Verbruggen, C., Ochyra, R., Verleyen, E., & Frenot, Y. (2010). Subantarctic flowering plants: pre‐glacial survivors or post‐glacial immigrants? Journal of biogeography, 37(3), 582-592. Wagstaff, S. J., & Hennion, F. (2007). Evolution and biogeography of Lyallia and Hectorella (Portulacaceae), geographically isolated sisters from the Southern Hemisphere. Antarctic Science, 19(4), 417. Wahid, A., & Ghazanfar, A. (2006). Possible involvement of some secondary metabolites in salt tolerance of sugarcane. Journal of plant physiology, 163(7), 723-730. Wink, M. (2003). Evolution of secondary metabolites from an ecological and molecular phylogenetic perspective. Phytochemistry, 64(1), 3-19. 130
Chapitre 2. Variation of secondary metabolites
Wink, M. (2013). Evolution of secondary metabolites in legumes (Fabaceae).South African Journal of Botany, 89, 164-175. Xie, L., Liu, P., Zhu, Z., Zhang, S., Zhang, S., Li, F., ... & Sun, R. (2016). Phylogeny and Expression Analyses Reveal Important Roles for Plant PKS III Family during the Conquest of Land by Plants and Angiosperm Diversification. Frontiers in Plant Science, 7.
131
Chapitre 2. Variation of secondary metabolites
1 Table 1. Differences among populations in the total contents of amines or quercetins. For each species, ANCOVAs 2 were conducted either across or within regions with “population” as factor and environmental conditions as co- 3 variables. Each line represents a separate ANCOVA and only the result for the factor “population” is shown. Error df 4 are given for each analysis. (See Table 7 for analysis of compositions). 5
6
7
8
9
10
11 12 13 14 15 16 17 18 19
132
Chapitre 2. Variation of secondary metabolites
20 Table 2. Relationship between abiotic conditions and total levels of amines or quercetins. Simple regression analyses, indicated are adjusted r squares and p-values after sequential 21 Bonferroni's correction. Significant p-values are indicated in bold, signs of significant relationships are indicated. Amine analyses: R. biternatus, sample size (N) = 58; R. 22 pseudotrullifolius N=57; R. moseleyi N=46. Quercetin analyses: R. biternatus, sample size (N) = 52; R. pseudotrullifolius N=48; R. moseleyi N=26. 23
24 25 26 27 28 29 30 31 32 33 34
133
Chapitre 2. Variation of secondary metabolites
35 Table 3. Relationship between total levels of amines or quercetins and traits. Simple regression analyses, indicated are adjusted r squares and p-values after sequential Bonferroni's 36 correction. Significant p-values are indicated in bold, signs of significant relationships are indicated. Amine analyses: R. biternatus, sample size (N) = 58; R. pseudotrullifolius N=57; 37 R. moseleyi N=46. Quercetin analyses: R. biternatus, sample size (N) = 52; R. pseudotrullifolius N=48; R. moseleyi N=26.
38 39
134
Chapitre 2. Variation of secondary metabolites
40 Table 4. Multivariate relationship between a given environmental variable and compositions of either a) all amines or b) all quercetins. Simple multivariate redundancy analyses 41 relating a given environmental variable to concentrations of all compounds. “Overall environment” represents scores along the first axis of a PCA calculated across the “sol water 42 saturation”, “pH” and “conductivity” Indicated are r squared and p-values after sequential Bonferroni's correction. Significant p-values are indicated in bold. Sample sizes (N) are 43 presented in the table. (See Fig. 2 & 3 for pairwise relationships between individual environmental variables and individual compounds, and Fig. 6 for an example of the full 44 multivariate relationships between all environmental variables and all compounds). 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Table 5. Multivariate relationship between a given trait and compositions of either a) all amines or b) all quercetins. Simple multivariate redundancy analyses relating a given trait to 61 concentrations of all compounds. “Phenotype” represents scores along the first axis of a PCA calculated across the remaining traits. Indicated are r squared and p-values after sequential 62 Bonferroni's correction. Significant p-values are indicated in bold. Sample size (N) are presented in the table. (See Fig. 4 & 5 for pairwise relationships between individual traits and 63 individual compounds, and Fig. 7 for an example of the full multivariate relationships between all traits and all compounds). 64 65 66 67 68 69 70
135
Chapitre 2. Variation of secondary metabolites
71 72 73 74 75 76 77 78
136
Chapitre 2. Variation of secondary metabolites
79 Table 6. Effects of region on the relationships between total metabolite contents (amines or quercetins) and a) the environment, or b) the phenotype among populations. The table 80 shows the interaction term between region and either a given environmental variable or a given trait. Interaction terms are taken from a multiple regression analyses that contain also 81 the respective raw variables. “Overall environment” as explained in Tab. 4, “phenotype” as explained in Tab. 5. Indicated are p-value and t-value of the interaction term, and adjusted 82 r squared of the multiple regression. Significant p-values after sequential Bonferroni's correction are indicated in bold. Amine analyses: R. biternatus, sample size (N) = 58; R. 83 pseudotrullifolius N=57; R. moseleyi N=46. Quercetin analyses: R. biternatus, sample size (N) = 52; R. pseudotrullifolius N=48; R. moseleyi N=26.
Effect of region on relationship between total metabolite contents and… a) Amines …overall environment …soil water saturation …pH …conductivity P-value t-value adj. R² P-value t-value adj. R² P-value t-value adj. R² P-value t-value adj. R² R. biternatus 0.69 -0.77 0.12 0.16 -1.44 0.12 0.27 -1.1 0.02 0.078 2.29 0.23 R. pseudotrullifolius 0.23 -1.21 0.32 0.069 2.34 0.34 0.95 -0.061 0.31 0.14 -1.51 0.45 R. moseleyi 0.16 1.42 0.35 0.84 0.21 0.26 0.93 -0.093 0.24 0.001 4.68 0.48
Quercetins …overall environment …soil water saturation …pH …conductivity Effect of region on relationship with metabolite P-value t-value adj. R² P-value t-value adj. R² P-value t-value adj. R² P-value t-value adj. R² R. biternatus 0.27 1.12 0.001 0.075 1.82 0.016 0.98 0.022 0.02 0.23 1.22 0.013 R. pseudotrullifolius 0.022 -2.65 0.12 0.0001 -4.53 0.29 0.022 -2.38 0.088 0.12 -1.59 0.079 R. moseleyi 0.12 1.95 0.14 0.057 2.53 0.49 0.17 1.42 0.14 0.39 -0.88 0.16 84 Effect of region on relationship between total metabolite contents and… b) Amines …phenotype …plant height …number of leaves Effect of region on relationship with metabolite P-value t-value adj. R² P-value t-value adj. R² P-value t-value adj. R² R. biternatus 0.91 0.015 0.042 0.89 0.88 0.025 0.094 -2.17 0.066 R. pseudotrullifolius 0.99 -0.17 0.26 0.31 0.45 0.27 0.0010 -3.29 0.39 R. moseleyi 0.046 3.064 0.35 0.042 2.4 0.32 0.41 0.22 0.26
Quercetins …phenotype …plant height …number of leaves Effect of region on relationship with metabolite P-value t-value adj. R² P-value t-value adj. R² P-value t-value adj. R² R. biternatus 0.63 -0.07 0.008 0.76 -0.87 0.022 0.61 -0.39 0.17 R. pseudotrullifolius 0.77 0.29 0.025 0.87 0.16 0.10 0.022 2.82 0.13 R. moseleyi 0.17 -1.83 0.18 0.08 0.28 0.11 0.57 1.62 0.10 85
137
Chapitre 2. Variation of secondary metabolites
86 Table 7. Differences among populations in the composition of amines or quercetins. For each species, redundancy analyses were conducted either across or within regions, with 87 “population” as factor and as dependent variables the concentrations of all amines (top) or all quercetins (bottom). Environmental variables (soil water saturation, pH and conductivity) 88 were included as co-variables. Each line represents a separate redundancy analysis and only the result for the factor “population” is shown (See Table 1 for analysis of total contents). 89 90 91
138
Chapitre 2. Variation of secondary metabolites
92 Figure 1. Sampled regions in Iles Kerguelen (modified from IGN map). 93 Figure 2. Relationships between individual amines and individual environmental variables from simple regression analyses, within and across regions. Sign of 94 significant p-values (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 58; in R. pseudotrullifolius N=57; in R. moseleyi N=46. 95 Abbreviation: Agm: Agmatine; Put: Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; 96 Oct: Octopamine; N1Ac.Spm: N1-acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine. (See Tab. 4 and Fig. 6 97 for increasingly complete multivariate relationships between environmental variables and compounds). 98 Figure 3. Relationships between individual quercetins and individual environmental variables from simple regression analyses, within and across regions. Sign of 99 significant p-values (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 52; in R. pseudotrullifolius N=48; in R. moseleyi N=26. 100 Abbreviation: Q-3GL: Quercetin 3-diglucoside-7-glucoside; Q-3GL+caf: Quercetin 3-(caffeyl-glucosyl)glucoside-7-glucoside; Q-3GL+Fer: quercetin 3-(ferulyl- 101 glucosyl)glucoside-7-glucoside; Q-2GL+Xyl+Caf: quercetin 3-(caffeyl-xylosyl)glucoside-7-glucoside; Q-2GL-Xyl+Fer: Quercetin 3-(ferulyl-xylosyl)glucoside-7-glucoside; Q- 102 2GL-Xyl : Quercetin 3-xylosylglucoside-7-glucoside; Q-GL-Xyl : Quercetin 3-xylosylglucoside; Q-2GL : Quercetin 3-diglucoside; IQC: Isoquercitrin. (See Tab. 4 for multivariate 103 relationships between environmental variables and compounds). 104 Figure 4. Relationships between individual traits and individual amines from simple regression analyses, within and across regions. Sign of significant p-values 105 (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 58; in R. pseudotrullifolius N=57; in R. moseleyi N=46. Abbreviation: Agm: 106 Agmatine; Put: Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; Oct: Octopamine; 107 N1Ac.Spm: N1-acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine. (See Tab. 5 and Fig. 7 for increasingly 108 complete multivariate relationships between traits and compounds). 109 Figure 5. Relationships between individual traits and individual quercetins from simple regression analyses, within and across regions. Sign of significant p-values 110 (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 52; in R. pseudotrullifolius N=48; in R. moseleyi N=26. Abbreviation: Q-3GL: 111 Quercetin 3-diglucoside-7-glucoside; Q-3GL+caf: Quercetin 3-(caffeyl-glucosyl)glucoside-7-glucoside; Q-3GL+Fer: quercetin 3-(ferulyl-glucosyl)glucoside-7-glucoside; Q- 112 2GL+Xyl+Caf: quercetin 3-(caffeyl-xylosyl)glucoside-7-glucoside; Q-2GL-Xyl+Fer: Quercetin 3-(ferulyl-xylosyl)glucoside-7-glucoside; Q-2GL-Xyl : Quercetin 3- 113 xylosylglucoside-7-glucoside; Q-GL-Xyl : Quercetin 3-xylosylglucoside; Q-2GL : Quercetin 3-diglucoside; IQC: Isoquercitrin. (See Tab. 5 for multivariate relationships between 114 traits and compounds). 115 Figure 6. Full relationship between all environmental variables and all amine compounds, within and across regions, exemplified for R. moseleyi. Environmental 116 factors are soil water saturation, pH and conductivity. In redundancy analyses, axes are constrained to show, at best, variance explained by amines. Individual, amines (in bold) and 117 environmental factor (with arrows) are indicated. N= are 28 in Isthme Bas; 30 in Australia; 58 across regions. P values (p) are indicated. Abbreviation: Agm: Agmatine; Put: 118 Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; Oct: Octopamine; N1Ac.Spm: N1- 119 acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine (see Tab. 4 and Fig. 2 & 3 for more detailed analysis, for 120 the full set of species, and for both amines and quercetins) 121 Figure 7. Full relationship between all environmental variables and all amine compounds, within and across regions, exemplified for R. biternatus. Traits are Plant 122 height, Number of leaves, Flower size, Leaf length, Leaf width, LDMC. In redundancy analyses, axes are constrained to show, at best, variance explained by amines. Individual, 123 traits (in bold) and amines (with arrows) are indicated. N= are 13 in Isthme Bas; 28 in Australia; 41 across regions. P values (p) are indicated. Abbreviation: Agm: Agmatine; Put: 124 Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; Oct: Octopamine; N1Ac.Spm: N1- 125 acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine (see Tab. 5 and Fig. 4 & 5 for more detailed analysis, for 126 the full set of species, and for both amines and quercetins) 127 128 129
139
Chapitre 2. Variation of secondary metabolites
130 Figure 1. Sampled regions in Iles Kerguelen (modified from IGN map). 131
Isthme Bas
Ile Australia
140
Chapitre 2. Variation of secondary metabolites
132 Figure 2. Relationships between individual amines and individual environmental variables from simple regression analyses, within and across 133 regions. Sign of significant p-values (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 58; in R. pseudotrullifolius N=57; in 134 R. moseleyi N=46. Abbreviation: Agm: Agmatine; Put: Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; 135 Tyr: Tyramine; Oct: Octopamine; N1Ac.Spm: N1-acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine. (See 136 Tab. 4 and Fig. 6 for increasingly complete multivariate relationships between environmental variables and compounds). 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
141
Chapitre 2. Variation of secondary metabolites
152 Figure 3. Relationships between individual quercetins and individual environmental variables from simple regression analyses, within and across 153 regions. Sign of significant p-values (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 52; in R. pseudotrullifolius N=48; in 154 R. moseleyi N=26. Abbreviation: Q-3GL: Quercetin 3-diglucoside-7-glucoside; Q-3GL+caf: Quercetin 3-(caffeyl-glucosyl)glucoside-7-glucoside; Q-3GL+Fer: quercetin 3- 155 (ferulyl-glucosyl)glucoside-7-glucoside; Q-2GL+Xyl+Caf: quercetin 3-(caffeyl-xylosyl)glucoside-7-glucoside; Q-2GL-Xyl+Fer: Quercetin 3-(ferulyl-xylosyl)glucoside-7- 156 glucoside; Q-2GL-Xyl : Quercetin 3-xylosylglucoside-7-glucoside; Q-GL-Xyl : Quercetin 3-xylosylglucoside; Q-2GL : Quercetin 3-diglucoside; IQC: Isoquercitrin. (See Tab. 4 157 for multivariate relationships between environmental variables and compounds). 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
142
Chapitre 2. Variation of secondary metabolites
173 Figure 4. Relationships between individual traits and individual amines from simple regression analyses, within and across regions. Sign of 174 significant p-values (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 58; in R. pseudotrullifolius N=57; in R. moseleyi N=46. 175 Abbreviation: Agm: Agmatine; Put: Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; 176 Oct: Octopamine; N1Ac.Spm: N1-acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine. (See Tab. 5 and Fig. 7 177 for increasingly complete multivariate relationships between traits and compounds). 178 179 180 181 182 183 184 185 186 187 188 189 190 191
143
Chapitre 2. Variation of secondary metabolites
192 Figure 5. Relationships between individual traits and individual quercetins from simple regression analyses, within and across regions. Sign of 193 significant p-values (<0.05) after sequential Bonferroni's correction are presented. In R. biternatus, sample size (N) = 52; in R. pseudotrullifolius N=48; in R. moseleyi N=26. 194 Abbreviation: Q-3GL: Quercetin 3-diglucoside-7-glucoside; Q-3GL+caf: Quercetin 3-(caffeyl-glucosyl)glucoside-7-glucoside; Q-3GL+Fer: quercetin 3-(ferulyl- 195 glucosyl)glucoside-7-glucoside; Q-2GL+Xyl+Caf: quercetin 3-(caffeyl-xylosyl)glucoside-7-glucoside; Q-2GL-Xyl+Fer: Quercetin 3-(ferulyl-xylosyl)glucoside-7-glucoside; Q- 196 2GL-Xyl : Quercetin 3-xylosylglucoside-7-glucoside; Q-GL-Xyl : Quercetin 3-xylosylglucoside; Q-2GL : Quercetin 3-diglucoside; IQC: Isoquercitrin. (See Tab. 5 for multivariate 197 relationships between traits and compounds). 198
144
Chapitre 2. Variation of secondary metabolites
Figure 6. Full relationship between all environmental variables and all amine compounds, within and across regions, exemplified for R. moseleyi. Environmental variables are soil water saturation, pH and conductivity. In redundancy analyses, axes are constrained to show, at best, variance explained by environmental variables. Individual, amines (in bold) and environmental variables (with arrows) are indicated. N= are 28 in Isthme Bas; 30 in Australia; 58 across regions. P values (p) are indicated. Abbreviation: Agm: Agmatine; Put: Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; Oct: Octopamine; N1Ac.Spm: N1-acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3-methoxy-4-hydroxy phenylethylamine (see Tab. 4 and Fig. 2 & 3 for more detailed analysis, for the full set of species, and for both amines and quercetins)
145
Chapitre 2. Variation of secondary metabolites
Figure 7. Full relationship between all amine compounds and all traits, within and across regions, exemplified for R. biternatus. Traits are Plant height, Number of leaves, Flower size, Leaf length, Leaf width, LDMC. In redundancy analyses, axes are constrained to show, at best, variance explained by amines. Individual, traits (in bold) and amines (with arrows) are indicated. N= are 13 in Isthme Bas; 28 in Australia; 41 across regions. P values (p) are indicated. Abbreviation: Agm: Agmatine; Put: Putrescine; Spm: Spermine; Spd: Spermidine; Cad: Cadaverine; DAP: 1,3-diaminopropane; Dop: Dopamine; Ser: Serotonin; Tyr: Tyramine; Oct: Octopamine; N1Ac.Spm: N1-acetylspermine; N8Ac.Spd: N8-acetylspermidine; Try: Tryptamine; 3M4OHPhe: 3- methoxy-4-hydroxy phenylethylamine (see Tab. 5 and Fig. 4 & 5 for more detailed analysis, for the full set of species, and for both amines and quercetins)
146
147
Chapitre 3. Plant plasticity is constrained by performance costs, complex environments and weakly integrated phenotypes
148
Chapitre 3. Costs and limits of plasticity
RESUME
Les plantes peuvent faire face au réchauffement climatique par une réponse plasticité de leurs traits. Cependant la plasticité peut devenir un coût pour la plante, et être limitée par l’intégration des traits ou par la modification d’autres facteurs environnementaux tels que l’ombrage. Les preuves empiriques de tels coûts et limites de la plasticité sont encore manquantes. Nous étudions trois espèces végétales provenant des îles Kerguelen en subantarctique, région faisant face à un réchauffement très intense. Pour une population d’origine nous avons identifié (i) la plasticité des plantes en réponse à différentes températures ou conditions d’exposition lumineuse, (ii) la performance photosynthétique ou la sénescence morphologique et (iii) l’intégration phénotypique des plantes. Nous avons trouvé que les plantes ayant un fort degré de plasticité avaient une performance plus faible. Egalement, les plantes provenant de populations d’origine plus intégrées étaient plus plastiques. De plus, un ombrage réduit la capacité de réponse plastique des plantes à un réchauffement. Qui plus est, un ombrage ou un réchauffement ne diminuent pas la performance des plantes, contrairement à une exposition simultanée à ces deux facteurs. Ainsi, l’intégration phénotypique peut favoriser, et non limiter, la plasticité des traits. Cependant, nos résultats suggèrent que la réponse plastique au réchauffement peut être limitée par une augmentation parallèle de l’ombrage de la part de compétiteurs. De plus, la plasticité, à cause de ses couts inhérents, ne garantit pas le succès des individus dans leur nouvel environnement.
ABSTRACT
One response of plants to climate warming is plasticity of traits, but plasticity might come at a cost and might be limited by the integration among traits or by simultaneous shift of another environmental factor such as shading. Empirical evidence of such costs and limitations of plasticity is lacking. We studied three plant species from the sub-Antarctic, the region currently facing one of the strongest warming world-wide. For each of multiple source populations we estimated (i) plasticity by exposing seeds from a given source population to different temperature and light treatments, (ii) performance (photosynthesis or morphological senescence) and (iii) morphological integration across young plants. We found that plants from more plastic source populations performed poorly. Plants from more integrated source populations were more plastic. Exposure to shade rendered plants less plastic to warming. Moreover, simultaneous shading and warming, rather than shading or warming alone, reduced plant performance. Overall, phenotypic integration surprisingly may favour rather than limit plasticity. However, our results suggest that plasticity in response to climate warming might be limited by parallel increase in shading from competitors, and itself does not ensure success due to induced performance costs.
149
Chapitre 3. Costs and limits of plasticity
Plant plasticity is constrained by
performance costs, complex environments and weakly integrated phenotypes
Bastien Labarrere1*, Andreas Prinzing1, Marine Renaudon1 & Françoise Hennion1
1 UMR 6553 Ecobio, Université de Rennes 1, CNRS, Av du Général Leclerc, F-35042 Rennes,
France, [email protected], *corresponding author, fax : (+33) 2 23 23 50 26
Keywords: Phenotypic integration; phenotypic plasticity; performance; plant species; abiotic treatment
150
Chapitre 3. Costs and limits of plasticity
INTRODUCTION
Climate change alters the environments and plants, being sessile organisms, have to cope with these alterations (Sultan, 2000). Phenotypic plasticity is defined as the capacity of a given genotype to produce different phenotypes under different environmental conditions (Bradshaw, 1965;
Schlichting, 1986; Sultan, 2000). Phenotypic plasticity is considered as one of the major means by which plants cope with environmental changes (Valladares et al., 2007). Plasticity can be considered as part of a species’ adaptive potential to sustain environmental changes (Ghalambor et al., 2007) and many studies showed that plants are plastic for morphological traits (Sultan, 2000;
Richards et al., 2006). However, many species express little plasticity (Valladares et al., 2007), and we do not know what are the costs or limitations that prevent species from being more plastic
(Pigliucci, 2005).
If plasticity is costly and reduces performance, one would expect little or no plasticity (Auld et al., 2009). Plastic response requires physiological and morphological machinery for acquiring environmental information, signal transduction, and the expression of a phenotypic response
(Vinocur and Altman, 2005). Such machinery implies costs, which may be of two kinds: constitutive and induced (Chevin et al., 2010). Constitutive costs include the costs of maintaining the physiological and signal acquisition machinery, and must be paid independent of the plastic response (Auld et al. 2009). Induced costs include the costs of the phenotypic changes and depend on the intensity of the response (Sultan and Spencer, 2002; Chevin et al., 2010). There is increasing evidence that costs of plasticity might reduce plant fitness, and even challenge their survival
(Pigliucci, 2005; Valladares et al., 2007; Ghalambor et al., 2007; Valladares and Niinemets, 2008).
However, demonstrations of the relationship between plasticity and performance remain scarce and ambiguous (Givnish, 2002; Valladares et al., 2007; Auld et al., 2009).
151
Chapitre 3. Costs and limits of plasticity
Plants may show little plasticity also because plasticity is constrained. One limitation might be that in nature, plants face complex environments in which multiple factors vary. Plastic response to such complex environmental changes might be more difficult to achieve than responses to simple laboratory environments under which plasticity is often studied (Valladares et al., 2007). In a complex environment, different factors, for instance drought and shade, may lead to opposite plant responses (Sack and Grubb, 2002). Hence variation in combinations of factors leading to opposite responses may annihilate an overall response. Also, costs involved in the response to a first factor may limit responses to other factors, due to constraints on energy allocation (Chevin et al., 2010).
Different traits responding to different factors independently may decrease intrinsic consistency of the organism and decrease plant performance. Also, a phenotype advantageous under extreme value of a given environmental factor can be maladaptive under extreme value of another factor
(Valladares et al., 2007). Hence, variation of a first factor may increase performance of individuals that show plastic response, yet, variation of an additional factor may render such plastic response maladaptive - and then decrease performance of plastic individuals. It still remains to determine how plants respond to multiple environmental factors and the consequences on plant performance.
Another limit to plasticity might be the integration among traits. Plasticity is mostly studied analyzing the variation of a given trait independent of other traits (Valladares et al., 2007; Gianoli and Palacio-Lopez, 2009; Mallitt et al., 2010; Murren, 2012). Yet, traits cannot vary independently in plants (Pigliucci, 2003). Consistency of plant organs is achieved through a correlation among trait values that is defined as phenotypic integration. Phenotypic integration may result from genetic, developmental, functional and environmental constraints (Murren, 2012; Armbruster et al., 2014). Very few authors have investigated the influence of phenotypic integration on plasticity
(Gianoli and Palacio-Lopez, 2009, Mallitt et al., 2010). On one hand, strong correlation among traits may limit their independent plastic variation. Indeed, Gianoli and Palacio-Lopez (2009) 152
Chapitre 3. Costs and limits of plasticity showed, in two plant species, a negative relationship between integration and plasticity. On the other hand, in poorly integrated organisms, extreme plastic modification of one trait may not be consistent with the modification of other traits and may result in maladaptive plasticity. In contrast, in highly integrated organisms, extreme modification of one trait value is consistent with extreme modification of the value of other traits. Also, in highly integrated organisms the modification of one trait value goes along with modifications of the values of other traits, increasing the overall degree of plasticity of the organism. Phenotypic integration might hence both, limit and facilitate plasticity.
The sub-Antarctic region is subject to a rapid and intense climate change (Le Roux et al., 2008;
Lebouvier et al., 2011). In Iles Kerguelen, located in the sub-Antarctic region, mean annual temperature increased by 1.3°C since 1960, and annual rainfall decreased by 100 to 250mm, in twenty years (Lebouvier et al., 2011). Plant species in Iles Kerguelen, have been growing for a long time in stable environment including low temperatures and abundant rainfall, seem particularly sensitive to ongoing changes (Frenot et al., 2006). These species already show signs of stress during droughts (Chapuis et al., 2004). Moreover, the geographical isolation of Iles Kerguelen hinders migration opportunities. Plasticity and tolerance to changes may help prevent the extinction of populations and species. We studied three indigenous species from Iles Kerguelen: two Ranunculus species: R. biternatus and R. pseudotrullifolius (Ranunculaceae) and the Brassicaceae Pringlea antiscorbutica. Some Ranunculus species are highly plastic (Garbey et al, 2004), for example exhibiting heterophylly that is considered a form of adaptive plasticity (Wells and Pigliucci, 2000).
All three species show trait plasticity when grown under different conditions (Hennion et al., 1994), and Pringlea antiscorbutica shows decreased root or shoot growth in response to salt or heat
(Hummel et al., 2004a; Hummel et al., 2004b). Investigating Kerguelen plant plastic capacities and
153
Chapitre 3. Costs and limits of plasticity performance in response to environmental variation is of major interest to determine survival of species to ongoing climate change.
We studied species in a phytotron, subjecting plants to variation of temperature for the
Ranunculus species and temperature and light levels for P. antiscorbutica. We investigated variation in morphological traits and photosynthetic efficiency. We aimed to determine (i) whether the degree of plasticity differed between response to temperature, shading or both; (ii) whether phenotypic integration covaried with degree of plasticity; and (iii) whether the degree of plasticity covaried with plant performance.
154
Chapitre 3. Costs and limits of plasticity
MATERIALS AND METHODS
1. Species under study-
The Iles Kerguelen (49°20’00” S, 69°20’00” E) are situated in the Southern Indian Ocean within the sub-Antarctic region (Lebouvier and Frenot, 2007) (figure 2). We studied R. biternatus Smith,
R. pseudotrullifolius Skottsb., and Pringlea antiscorbutica. R.Br. ex Hook.f. All are perennial plants. On the Iles Kerguelen these species occupy different habitats (Hennion and Walton, 1997a).
Ranunculus biternatus is widespread on the island occurring in habitats below 500m above sea level (Hennion and Walton, 1997a; Réserve Naturelle des Terres Australes et Antarctiques françaises and IPEV Programme 136, unpublished data). In contrast, R. pseudotrullifolius being halophilous has a more restricted distribution and occurs within a short distance of the coast, occupying peaty or sandy shores and ponds (Hennion and Walton, 1997a), and sporadically in ponds up to about 200m a.s.l. (Réserve Naturelle des Terres Australes et Antarctiques françaises and IPEV Programme 136, unpublished data). On the Iles Kerguelen P. antiscorbutica, an endemic species from the Southern Indian Ocean, occurs in habitats ranging from coastal meadows to montane fell-fields (Hennion and Bouchereau, 1998).
2. Seed collection-
To grow plants under controlled conditions, we used seeds from different natural populations in the three species. Populations were defined as continuous groups of plants living at a given site.
Seeds from P. antiscorbutica were collected in the field in 2013 in three regions: Ile Australia,
Mont Crozier and Ile Mayes. Ten individuals were randomly selected within each population.
Fifteen siliques were then randomly collected from the median part of the longest infructescence where seed size is maximum and least variable, from each of the ten individuals (Hennion,
155
Chapitre 3. Costs and limits of plasticity
Schermann-Legionnet and Atlan, unpubl.). Seeds were also collected from P. antiscorbutica plants growing from two populations in a phytotron in 2013. In R. biternatus and R. pseudotrullifolius, seeds were collected in 2014 from 20 individuals across three populations in Isthme Bas, the same in which phenotypic measurements had been performed in 2012. All seeds were stored dry with silica gel at 4°C until their use according to an established protocol (Hennion and Walton, 1997b).
3. Controlled conditions- We grew plants in a phytotron at Rennes, France. Seeds were sterilized and germinated in Petri dishes at 25°C for P. antiscorbutica and 21°C for Ranunculus species, under low light (1kLux) following an established protocol (Hennion and Walton, 1997b). To render the overall design more homogeneous we kept a constant number of seedlings from each population. In each species, seedlings were planted in vermiculite and fertilized with ½ Hoagland solution (Hoagland and
Arnon, 1938). For the R. pseudotrullifolius, NaCl was added to Hoagland solution to reach a concentration of 1g.L-1, being consistent with the salinity in the field (Labarrere et al., unpubl).
Plants were grown hydroponically in a phytotron for five months under 14h daylight, with an irradiance of 15klux from a combination of Philips TLD 79 fluorescent and Philips Aquarelle TLD
89 tubes, at 9°C during the day and 4°C during the night, with a relative humidity of 75/85% day/night. These conditions mimick the austral summer in Iles Kerguelen (Hennion and
Bouchereau, 1998; Hennion et al., 2006).
4. Experimental treatments- After five months, these plants were assigned to different treatment groups. We established two replicates. In each replicate, 64 plants from five source populations from P. antiscorbutica were assigned to each of the four treatment combinations of a factorial experiment of control (9°C day and 4°C night) versus high (13°C day and 4 night) temperature and control (15klux) versus low 156
Chapitre 3. Costs and limits of plasticity light (5klux) treatments. For R. biternatus and R. pseudotrullifolius, there were insufficient seeds for all treatment combinations so only temperature was manipulated, as above : In each of the two replicates, we used 76 plants from R. biternatus (four from each of 19 mothers) per treatment and
56 plants from R. pseudotrullifolius (four from each of 14 mothers) per treatment. Experimental design is summarized in Figure 1. Note that we found significant trait differences among source mothers (or source populations) in the control, in the three species (Appendix 1). Plants were grown under their experimental conditions for an additional 3 months.
5. Trait measurements - We measured all plants after five months of growth, just before assigning them to treatments and again at the end of the three month treatment period for the following traits: height, diameter and number of leaves of the rosette, length and width of lamina and length of petiole of the 2 longest leaves. In R. biternatus and R. pseudotrullifolius, diameter was not measured. We also estimated two aspects of photosynthetic activity: leaf spread, the ratio between length of leaf that is not curled and maximum flattened leaf length, as an estimation of photosynthetic activity (Zamora et al.,
1998). Leaves that are curled indicated that the plant is not using its complete photosynthetic capacity, because the curled part of the leaf is not exposed to light and therefore does not contribute to photosynthesis. We further measured the maximum efficiency of photosystem II (Maxwell and
Johnson, 2000; Zhu et al., 2008), a useful performance parameter known to decrease with stress
(Maxwell and Johnson, 2000; Kalaji et al., 2011). Schortemeyer et al. (2015) showed that photosynthesis in P. antiscorbutica may respond to warming. Maximum PSII efficiency (Fv/Fm) were measured using a PAM chlorophyll fluorometer system (Heinz Walz, Effeltrich, Germany) in saturating pulse mode. After 30 minutes in the dark, the maximum fluorescence of leaves was measured under dark conditions (Fm) and a subsequent saturating pulse of light was applied.
157
Chapitre 3. Costs and limits of plasticity
Minimum fluorescence (F0) was determined under weak light before a pulse, and variable fluorescence (Fv=Fm–F0) was calculated.
6. Data analyses-
6.1. Trait means and plasticity
We investigated plant plasticity comparing relative change in plant trait values during the three months of growth under the experimental conditions between the different experimental conditions.
We aimed to remove the intrinsic trait value differences among plants which would not be induced by experimental conditions. Thus, we calculated trait value as: Standardized trait value Ts = (T1-
T0)/T0, with T0 being the measurement from just before the beginning of the treatment and T1 being the measurement carried out after three months of growth under the experimental conditions.
In P. antiscorbutica, we calculated the means of relative change in trait value across plants from the same population subjected to same condition. In R. biternatus and R. pseudotrullifolius, we calculated means across plants from the same mother subjected to same condition. In all species, we calculated - for each population (or mother) and treatment - trait plasticity as the absolute value of the difference between trait mean under treatment and trait mean under control. In P. antiscorbutica we also calculated the mean degree of plasticity across the three treatments.
6.2. Phenotypic integration
We calculated phenotypic integration across individuals from same source populations before and after the treatments in P. antiscorbutica. Samples sizes did not permit the same estimation for
Ranunculus species. Phenotypic integration was calculated as the percentage of variance explained by the first axis of a Principal Component Analysis (PCA) performed on traits (Hermant et al., in
2013; Labarrere et al., unpubl.). PCA was performed using correlation matrix. Traits were height 158
Chapitre 3. Costs and limits of plasticity and diameter of the rosette, and of the 2 longest leaves: length and width of lamina and length of petiole. Trait values were scaled and PCA conducted using the FactoMineR package of R 3.0.2 (Lê et al., 2008; R Core Team, 2012). We found no difference in degree of phenotypic integration before versus after the exposition to treatments (Appendix 2a) or among the different treatments when compared with controls (Appendix 2b).
6.3. Hypothesis Testing
The relationships between trait means, plasticity or integration were tested using simple linear regression. We determined differences of trait values among treatments using ANOVA. We used
MANOVA to determine whether plant phenotype (accounting for all traits) differed among treatments. The same traits were tested in MANOVA and ANOVA analyses. Analyses were conducted in R 3.0.2 (R Core Team, 2012). For multiple testing on the same data set, p-values were corrected using sequential Bonferroni’s correction (Cabin and Mitchell, 2000). In all regression analyses we verified the assumptions of the analyses using residual plots.
159
Chapitre 3. Costs and limits of plasticity
RESULT
1. Effects of treatments on plant traits-
In P. antiscorbutica, plants were significantly larger after three months of warmer temperatures than in control (p=0.005; F=5.9; Df=16; MANOVA) but significantly smaller than controls after being grown at low light, without, or with warmer temperatures (respectively: p=0.0043; F=6.8;
Df=16 and p=0.012; F=5.1; Df=16; MANOVA). We found no difference in plant phenotype between shading and combined warming-shading treatments (p=0.55; F=0.93: Df=16;
MANOVA). Trait differences between treatments and the control are indicated Table 1. Trait differences among plants are shown in Appendix 3.
In R. biternatus, number of leaves as well as petiole length and lamina width of leaf 2 were significantly larger in warming treatment compared to control (Table 2). In R. pseudotrullifolius, lamina length of leaf 1 was significantly larger in warming treatment than in control (Table 2).
Overall, phenotype was larger (marginally significant) in warming in R. biternatus but not in R. pseudotrullifolius (respectively: p=0.058; F=2.2; Df=17 and p=0.57; F=0.8; Df=12; MANOVA)
2. Effects of treatments on plant performance –
In P. antiscorbutica, we found that leaves were more curled in combined warming-shading treatment than in the other treatments or in the control (Table 3).
The Ranunculus species did not differ significantly between control and warming treatment for survival, photosynthetic efficiency or leaf curling (Table 4).
3. Effect of treatments on the degree of plasticity-
In P. antiscorbutica, we found that the degree of plasticity was higher in response to warming treatment than in response to shading treatment or combined warming-shading treatment
160
Chapitre 3. Costs and limits of plasticity
(respectively: p=0.05; F=2.19; Df=16 and p=0.036; F=2.21; Df=16; MANOVA). In contrast, the degree of plasticity did not differ between shading treatment and combined warming-shading treatment (p=0.45; F=0.46; Df=16; MANOVA). Specifically, degrees of plasticity of plant height and lamina width of the tallest leaf were higher in response to warming treatment than to shading treatment (Table 5). Degrees of plasticity of lamina width of leaf 1 and leaf 2 were higher in response to warming treatment than in response to combined warming-shading treatments (Table
5).
4. Effect of the degree of plasticity on performance-
In P. antiscorbutica, we found significant negative relationship between the degree of trait plasticity and plant performance (percentage of leaf spread and photosynthesis efficiency) in response to warming treatment (Figure 2). Also, we found a marginally significant negative relationship between trait plasticity and performance (percentage of leaf spread) in response to combined warming-shading treatment (Figure 3). We found no significant relationships between the degree of trait plasticity and plant performances in response to shading treatment (p-value>0.05; all adj.r²<0.1; simple linear regressions). To explore constitutive costs of being plastic we compared across-treatment means of plasticity per source population to performances in control condition but found no relationships (p-value>0.05; all adj.r²<0.1; simple linear regression).
In R. biternatus and R. pseudotrullifolius, we found no relationship between the degree of trait plasticity and plant performance.
5. Effect of phenotypic integration the degree of plasticity-
In P. antiscorbutica, we found a significant positive relationship between phenotypic integration and the degree of plasticity of leaf 1 length in shading treatment (Figure 4). Also, we found marginally significant positive relationships between phenotypic integration and the degree of
161
Chapitre 3. Costs and limits of plasticity
plasticity of leaf 2 length in shading treatment (Figure 4). We found no significant relationship between integration and the degree of trait plasticity in other treatments.
162
Chapitre 3. Costs and limits of plasticity
DISCUSSION
As sessile organisms, plants have to face changes of their environmental conditions. Phenotypic plasticity is considered as a major means for plants to face environmental change (Ghalambor et al., 2007). Yet, plasticity has costs and limits that are recognized but remain to be demonstrated
(Pigliucci, 2005; Valladares et al., 2007; Chevin et al., 2010; Auld et al., 2009). Also, plasticity is mainly studied analysing the independent response of traits to the variation of a single environmental factor (Valladares et al., 2007). Yet, in nature plants are subject to the variation of multiple factors. Moreover, traits are intercorrelated (i.e. phenotypic integration) and cannot vary independently. We investigated the relationships between phenotypic integration and phenotypic plasticity to two different environmental factors, temperature and shading. Moreover, we investigated the relationships between plasticity and plant performance in experimental treatments.
We showed that phenotypic plasticity may differ in response to different environmental factors and that variation in one factor can impede the phenotypic response to another factor. More highly integrated populations may also show higher plasticity. Moreover, we showed that plasticity in morphological traits is associated with lower values of some measures of performance suggesting induced costs of plasticity.
1. Trait values and performance under different treatments-
Iles Kerguelen are facing rapid and intense climate change, characterized by an increase of temperatures and a decrease of rainfalls in the past decades (Lebouvier et al., 2011). Plants in Iles
Kerguelen already exhibit signs of stress during dry summer periods (Frenot et al., 2006). Pringlea antiscorbutica shows decreased root and shoot growth in response to salt or heat (Hummel et al.,
2004a; Hummel et al., 2004b). Plant height, lamina and petiole length decreased in response to a
163
Chapitre 3. Costs and limits of plasticity decrease in the degree and frequency of submergence in R. biternatus and R. pseudotrullifolius
(Hennion et al., 1994). Moreover, R. biternatus showed larger achenes than in the field in
Kerguelen, when grown in temperate conditions (Hennion and Walton, 1997b). In this study, we showed that all species exhibit plastic morphological responses to abiotic changes. In P. antiscorbutica we found that traits (4 of 9 traits) increase in response to warming. Traits increasing in response to warming contrast with results shown in Hummel et al. (2004b). Yet, we compared plants in temperatures from 9 to 13°C while Hummel et al. (2004b) compared plants at 5°C to plants at 22°C. The higher temperature treatment used in Hummel et al. (2004b) may have imposed a stress leading to a decrease of morphological trait values in plants. Moreover, Hummel et al.
(2004b) worked on 15-day old seedlings. In R. biternatus, various traits increase in response to warming (3 on 8 traits). In contrast, in R. pseudotrullifolius, only few plastic traits increased in response to warming (1 on 8 traits).
Species that show plasticity in response to the abiotic factors that may change are more likely to sustain climate change (Valladares et al., 2007). Temperature increased by 1.3°C since 1960
(Lebouvier et al., 2011). Plasticity shown in P. antiscorbutica, R. biternatus and to a lesser extent in R. pseudotrullifolius in response to warming might favour their capacity to face climate change.
Moreover, we showed in the three species that an increase of temperature does not affect young, non-mature plant performances during the treatment. Seedling stages are considered to be more sensitive to environmental stresses than older stages (Clark & Wilson, 2003; Hummel et al.,
2004b). However, our treatment lasted for three months only and we recognize that a persistent increase of temperature might in the long term affect plant performance. Under well-watered cultivation in a common garden under a temperate climate at Brest (France), young plants of P. antiscorbutica, R. biternatus and R. pseudotrullifolius collected from Iles Kerguelen first showed increased growth in relation to the field but finally collapsed after several months (Hennion (com. 164
Chapitre 3. Costs and limits of plasticity pers.). In addition, in the field, plants are also exposed to the wind and different levels of soil water saturation or soil nutrient. Thus, increased temperature combined with these factors might affect plant performance.
Increased temperatures in Iles Kerguelen is suggested to favour expansion of invasive plant species (Frenot et al., 2006), which may compete native species, for instance for light resource. We investigated the effect of shading as well as the combined effect of warming and shading in P. antiscorbutica. We showed that traits decreased in response to shading or combined warming and shading.. Moreover, plant phenotypes did not differ between the two treatments. Yet, we showed that while shading does not affect plant performance, combined warming and shading reduces plant performance. Plants of P. antiscorbutica might be sensitive to the combination of climate change and shading due to colonization by invasive species. Note that P. antiscorbutica is found in a wide range of environments in Iles Kerguelen from coastal meadows to altitudinal fell-fields. Increase of temperature and light competition is expected to mainly affect P. antiscorbutica at low altitude.
At high altitude, environment is less suitable to colonization by aliens, and P. antiscorbutica less likely to be impacted by inter-specific light competition. Overall, warming or shading does not impact plant performance separately but their combination does. The effects of biotic and abiotic factors on plant performance are only rarely analysed together, but here we showed that competition for light (with natives and introduced species) might impact the capacity of plants to cope with warming. Importantly, by triggering plant growth, warming may itself be the cause of increased light competition.
2. Degrees of plasticity differ among experimental treatment-
In P. antiscorbutica, we investigated differences of degree of plasticity among treatments.
Particularly, we investigated to what extent simultaneous variation of two environmental factors affect the degree of plant plasticity - compared to the variation of single factor. Firstly, we found 165
Chapitre 3. Costs and limits of plasticity that the degree of plasticity is higher in response to warming than in response to shading. We yet recognize that the two factors do not have the same units and difference of plasticity might result from a difference of the intensity of changes, hardly comparable, between two factors. Secondly, we found that the degree of plasticity is lower in response to warming-shading than in response to warming. We discussed above that warming and shading have opposite effects on plant growth. A first explanation for the decreased degree of plasticity in response to multiple factors is hence that opposite effects of warming and shading may annihilate each other - resulting in a low degree of plasticity. Further study may investigate whether, in contrast, factors having the same effect on plant growth have cumulative effects on the degree of plasticity. Another explanation is that constraints of resource allocation might prevent traits to respond to multiple environmental factors, as suggested by Chevin et al. (2010). We showed above that the degree of plasticity is higher in response to warming than in response to shading. We might therefore expect the plant phenotype in warming-shading to be closer to the plant phenotype in warming than plant phenotype in shading. Yet, we found that plant phenotypes in warming-shading are closer to plant phenotypes in shading than plant phenotypes in warming. We hence suggest that, beyond their opposite effects, shading treatment impose a stronger constraint than warming treatment. Resource constraints imposed by shading might prevent plants from responding to temperature and might potentially explain lower performance discussed above.
Overall, we showed that shading decreases plant plastic response to warming. More generally simultaneous variation of multiple factors may annihilate plant plastic capacity. Climate change alters simultaneously multiple environmental factors, and plasticity might hence not be a mean to face such changes. Also, we emphasize the need to investigate plasticity in response to simultaneous shift of multiple factors – rather than shifts of single factors - to better estimate organism capacity to face environmental change. 166
Chapitre 3. Costs and limits of plasticity
3. Phenotypic integration may increase plasticity-
In highly integrated phenotypes, traits cannot vary independently (Pigliucci, 2003; Murren, 2012).
Therefore, phenotypic integration has been suggested to constraint trait response to environmental change (Gianoli and Palacio-Lopez, 2009), especially when various environmental conditions change simultaneously (Schlichting, 1989). Yet, very few authors analysed the influence of phenotypic integration on the degree of plasticity (Gianoli and Palacio-Lopez, 2009). Before determining the influence of phenotypic integration on the degree of plastic in response to a given abiotic factor, we must verify whether phenotypic integration responds or not to that abiotic factors.
High phenotypic integration decreasing in response to an abiotic factor may not be a constraint for traits to vary. Evidence of the relationship between phenotypic integration and abiotic conditions is lacking and scarce (Mallitt et al., 2010; Labarrere et al., unpubl.). Here, we showed that degree of integration does not vary in response to abiotic conditions.
Given that integration does not respond to treatments, and therefore might potentially influence trait plasticity. Integration might be a constraint for trait plastic response as shown in Gianoli and
Palacio-Lopez (2009). Alternatively, however, we suggest that integration might also favour trait plastic response as in highly integrated organisms, extreme variation of one trait value matches extreme variation of other traits values, allowing extreme trait modification to not reduce plant consistency (Labarrere et al., unpubl.). Here, we found no relationships between integration and plasticity in warming or warming-shading treatment, but we found that phenotypic integration increased plant plasticity in shading treatment. To our knowledge this is the first time such an increase of plasticity of traits by their integration has been evidenced.
167
Chapitre 3. Costs and limits of plasticity
4. Costs of plasticity: High degree of plasticity decreases performance-
Plasticity helps plants facing environmental changes (Richards et al., 2006), yet, plasticity also has costs which may compromise plant fitness (Valladares et al., 2007; Valladares and Niinemets,
2008). We investigated whether highly plastic populations had better or worse performance in P. antiscorbutica. We showed that more plastic populations had lower plant performance in high temperature treatment. Yet, we found no significant relationship between plasticity and performance in the other treatments. Note that plastic response to warming did not mean weaker growth but rather the opposite, and plasticity in response to warming is hence not the result of lower performance.
Costs of plasticity may be dissociated in two kinds of costs (Chevin et al., 2010; Auld et al.,
2009), constitutive and induced. Constitutive cost result from maintaining basis physiological machinery of plasticity while induced costs depends on the amount of phenotype changes (Sultan and Spencer, 2002). Constitutive costs have to be payed even without phenotypic change in a constant environment (Chevin et al., 2010; Auld et al., 2009). However, we found that performance under stable conditions did not depend on the capacity to plastically respond changing conditions.
Plasticity might hence have only little constitutive costs in our study system. Induced costs, in contrast, reflect the costs of morphological modification in response to environmental change. High induced costs might impose resource allocation constraints (Chevin et al., 2010), which does not allow plants to maintain the same degree of performance. Here we found that the degree of plasticity is negatively related to plant performance, in warming treatment. Above, we discussed that warming induced higher degree of plasticity than other treatments. We suggest that a high degree of plasticity generates important induced costs that impact plant performance. We also observed negative plasticity-performance relationship in response to combined warming-shading treatment. We suggest that most stressful environment, such as warming and shading combined, 168
Chapitre 3. Costs and limits of plasticity reduces energy available in plants therefore plants allocating resources in trait modification show decreased performance. Overall, we found no positive relationship between the degree of plasticity and performance suggesting that more plastic organisms do not inevitably better sustain environmental change. On the opposite, plants that show high degree of plasticity may perform poorly. Performance is poor once the environmental change occurs, not before, i.e. reflecting induced rather than constitutive costs.
CONCLUSION
Although the concept of plastic response to environmental change has long been studied, our understanding has to be improved about costs and limits of such plastic response (Pigliucci, 2005).
Here we showed that simultaneous shift of various environmental factors may limit plasticity, reducing the capacity to respond to climate warming under increased light competition (itself potentially resulting from warming). Also, we showed that, contrary to a common belief
(Schlichting, 1989), phenotypic integration does not limit, but may even increase trait plasticity.
Moreover, we showed that highly plastic individuals perform poorly. Due to induced costs, plant plastic response to environmental change does not guarantee plant success in the new environment.
Hence, poorly plastic individuals may be as, or even more, performant in the new environment than highly plastic individuals. We particularly emphasize the need to analyse organism performance while studying plasticity.
169
Chapitre 3. Costs and limits of plasticity
Acknowledgements
The authors thank volonteers of Programme IPEV 136 and Réserve Naturelle TAAF for help in collecting seeds. Field work was supported by IPEV (programmes 1116 PlantEvol, F. Hennion and
136 Ecobio, M. Lebouvier). We thank Cécile Sulmon for advice on photosynthesis measurement, particularly regarding the use of PAM chlorophyll fluorometer system. We thank personnel of
Center ECOLEX, Rennes (Valérie Gouesbet, Jean-Luc Foulon, Thierry Fontaine and Fouad
Nassur) for help in plant cultivation.
170
Chapitre 3. Costs and limits of plasticity
REFERENCES
Armbruster W.S., Pélabon C., Bolstad G.H. & Hansen T.F. (2014) Integrated phenotypes: understanding trait covariation in plants and animals. Phil. Trans. R. Soc. B, 369, 20130245. Auld J.R., Agrawal A.A. & Relyea R.A. (2009) Re-evaluating the costs and limits of adaptive phenotypic plasticity. Proceedings of the Royal Society of London B: Biological Sciences, rspb20091355. Bradshaw A.D. (1965) Evolutionary significance of phenotypic plasticity in plants. Advances in genetics, 13, 115-155. Cabin R.J. & Mitchell R.J. (2000) To Bonferroni or not to Bonferroni: when and how are the questions. Bulletin of the Ecological Society of America, 81, 246-248. Chapuis J.-L., Frenot Y. & Lebouvier M. (2004) Recovery of native plant communities after eradication of rabbits from the subantarctic Kerguelen Islands, and influence of climate change. Biological Conservation, 117, 167-179. Chevin L.-M., Lande R. & Mace G.M. (2010) Adaptation, plasticity, and extinction in a changing environment: towards a predictive theory. PLoS Biol, 8, e1000357. Clark D.L. & Wilson M.V. (2003) Post-dispersal seed fates of four prairie species. American Journal of Botany, 90, 730-735. Couée I., Hummel I., Sulmon C., Gouesbet G. & El Amrani A. (2004) Involvement of polyamines in root development. Plant Cell, Tissue and Organ Culture, 76, 1-10. Frenot Y., Lebouvier M., Gloaguen J.-C., Hennion F., Vernon P. & Chapuis J.-L. (2006) Impact des changements climatiques et de la fréquentation humaine sur la biodiversité des îles subantarctiques françaises. Belgeo. Revue belge de géographie, 363-372. Garbey C., Thiébaut G. & Muller S. (2004) Morphological plasticity of a spreading aquatic macrophyte, Ranunculus peltatus, in response to environmental variables. Plant Ecology, 173, 125-137. Ghalambor C.K., McKay J.K., Carroll S.P. & Reznick D.N. (2007) Adaptive versus non adaptive phenotypic plasticity and the potential for contemporary adaptation in new environments. Functional ecology, 21, 394-407. ‐ Gianoli E. & Palacio López K. (2009) Phenotypic integration may constrain phenotypic plasticity in plants. Oikos, 118, 1924-1928. Givnish T.J. (2002) Ecological‐ constraints on the evolution of plasticity in plants. Evolutionary ecology, 16, 213-242. Hennion F. & Bouchereau A. (1998) Accumulation of organic and inorganic solutes in the subantarctic cruciferous species Pringlea antiscorbutica in response to saline and cold stresses. Polar Biology, 20, 281-291. Hennion F., Bouchereau A., Gauthier C., Hermant M., Vernon P. & Prinzing A. (2012) Variation in amine composition in plant species: How it integrates macroevolutionary and environmental signals. American journal of botany, 99, 36-45. Hennion F., Fiasson J.L. & Gluchoff-Fiasson K. (1994) Morphological and phytochemical relationships between Ranunculus species from Iles Kerguelen. Biochemical systematics and ecology, 22, 533-542. Hennion F., Huiskes A.H.L., Robinson S. & Convey P. (2006) Physiological traits of organisms in a changing environment. In: Trends in Antarctic Terrestrial and Limnetic Ecosystems,
171
Chapitre 3. Costs and limits of plasticity
pp. 129-159. Springer. Hennion F. & Walton D.W.H. (1997a) Ecology and seed morphology of endemic species from Kerguelen phytogeographic zone. Polar Biology, 18, 229-235. Hennion F. & Walton D.W.H. (1997b) Seed germination of endemic species from Kerguelen phytogeographic zone. Polar Biology, 17, 180-187. Hermant M., Prinzing A., Vernon P., Convey P. & Hennion F. (2013) Endemic species have highly integrated phenotypes, environmental distributions and phenotype–environment relationships. Journal of Biogeography, 40, 1583-1594. Hoagland D.R. & Arnon D.I. (1938) Growing plants without soil by the water-culture method. Circ. Calif. Agric. Exp. Stn. Hummel I., El Amrani A., Gouesbet G., Hennion F. & Couée I. (2004) Involvement of polyamines in the interacting effects of low temperature and mineral supply on Pringlea antiscorbutica (Kerguelen cabbage) seedlings. Journal of experimental botany, 55, 1125- 1134. Hummel I., Gouesbet G., El Amrani A., Aïnouche A. & Couée I. (2004) Characterization of the two arginine decarboxylase (polyamine biosynthesis) paralogues of the endemic subantarctic cruciferous species Pringlea antiscorbutica and analysis of their differential expression during development and response to environmental stress. Gene, 342, 199- 209. Kalaji H.M., Bosa K., Kościelniak J. & Żuk-Gołaszewska K. (2011) Effects of salt stress on photosystem II efficiency and CO 2 assimilation of two Syrian barley landraces. Environmental and Experimental Botany, 73, 64-72. Lebouvier M. & Frenot Y. (2007) Conservation and management in the French sub-Antarctic islands and surrounding seas. Lebouvier M., Laparie M., Hulle M., Marais A., Cozic Y., Lalouette L., Vernon P., Candresse T., Frenot Y. & Renault D. (2011) The significance of the sub-Antarctic Kerguelen Islands for the assessment of the vulnerability of native communities to climate change, alien insect invasions and plant viruses. Biological Invasions, 13, 1195-1208. Lê S., Josse J. & Husson F. (2008) FactoMineR: an R package for multivariate analysis. Journal of statistical software, 25, 1-18. Le Roux, P. C., & McGeoch, M. A. (2008). Changes in climate extremes, variability and signature on sub-Antarctic Marion Island. Climatic Change, 86, 309-329. Mallitt K.L., Bonser S.P. & Hunt J. (2010) The plasticity of phenotypic integration in response to light and water availability in the pepper grass, Lepidium bonariense. Evolutionary ecology, 24, 1321-1337. Maxwell K. & Johnson G.N. (2000) Chlorophyll fluorescence—a practical guide. Journal of experimental botany, 51, 659-668. Murren C.J. (2012) The integrated phenotype. Integrative and comparative biology, 52, 64-76. Pigliucci M. (2003) Phenotypic integration: studying the ecology and evolution of complex phenotypes. Ecology Letters, 6, 265-272. Pigliucci M. (2005) Evolution of phenotypic plasticity: where are we going now? Trends in Ecology & Evolution, 20, 481-486. Richards C.L., Bossdorf O., Muth N.Z., Gurevitch J. & Pigliucci M. (2006) Jack of all trades, master of some? On the role of phenotypic plasticity in plant invasions. Ecology letters, 9, 981-993. Sack L. & Grubb P.J. (2002) The combined impacts of deep shade and drought on the growth and biomass allocation of shade-tolerant woody seedlings. Oecologia, 131, 175-185. 172
Chapitre 3. Costs and limits of plasticity
Schlichting C.D. (1986) The evolution of phenotypic plasticity in plants. Annual review of ecology and systematics, 667-693. Schlichting C.D. (1989) Phenotypic integration and environmental changes. BioScience, 39, 460- 464. Schortemeyer, M., Evans, J. R., Bruhn, D., Bergstrom, D. M., & Ball, M. C. (2015). Temperature responses of photosynthesis and respiration in a sub-Antarctic megaherb from Heard Island. Functional Plant Biology, 42, 552-564. Sultan S.E. (2000) Phenotypic plasticity for plant development, function and life history. Trends in plant science, 5, 537-542. Sultan S.E. & Spencer H.G. (2002) Metapopulation structure favors plasticity over local adaptation. The American Naturalist, 160, 271-283. Valladares F., Gianoli E. & Gómez J.M. (2007) Ecological limits to plant phenotypic plasticity. New Phytologist, 176, 749-763. Valladares F. & Niinemets Ü. (2008) Shade tolerance, a key plant feature of complex nature and consequences. Annual Review of Ecology, Evolution, and Systematics, 237-257. Vinocur B. & Altman A. (2005) Recent advances in engineering plant tolerance to abiotic stress: achievements and limitations. Current opinion in biotechnology, 16, 123-132. Wells C.L. & Pigliucci M. (2000) Adaptive phenotypic plasticity: the case of heterophylly in aquatic plants. Perspectives in Plant Ecology, Evolution and Systematics, 3, 1-18. Zamora R., Gómez J.M. & Hódar J.A. (1998) Fitness responses of a carnivorous plant in contrasting ecological scenarios. Ecology, 79, 1630-1644. Zhu X.-G., Long S.P. & Ort D.R. (2008) What is the maximum efficiency with which photosynthesis can convert solar energy into biomass? Current opinion in biotechnology, 19, 153-159.
173
Chapitre 3. Costs and limits of plasticity
Table1. ANOVA analyses indicating trait differences between different abiotic treatments in P. antiscorbutica. P-values (P) after sequential Bonferroni's correction and F values (F) are indicated. "": trait is lower/larger in the first treatment indicated. N= 10 in each treatment, df=18.
Plant Leaf 1 Leaf 2 Plant height Diameter Number of leaves Petiol length Lamina length Lamina width Petiol length Lamina length Lamina width P F P F P F P F P F P F P F P F P F Control vs shading 0.30 2.1 0.001 27.8 > 0.001 24.8 > 0.50 0.47 0.10 6.09 > 0.007 15.7 > 0.88 0.02 0.10 6.84 > 0.29 1.2 Control vs warming-shading 0.17 1.14 0.06 8.4 > 0.028 21.8 > 0.25 4.54 0.83 0.046 0.008 11 > 0.11 2.85 0.69 0.16 0.21 1.7 Shading vs warming-shading 0.59 0.3 0.25 1.5 0.34 0.4 0.42 3.78 > 0.15 2.23 0.59 0.29 0.32 5.05 > 0.30 4.37 > 0.88 0.02 Control vs warming 0.75 0.10 0.18 2 0.001 16.8 < 0.14 2.33 0.02 10.72 < 0.001 28.8 < 0.30 1.14 0.05 8.1 < 0.001 28 <
174
Chapitre 3. Costs and limits of plasticity
Table 2. ANOVA analyses indicating trait differences between warming and control in R. biternatus and R. pseudotrullifolius. P-values (P) after sequential Bonferonni's correction and F values (F) are indicated. "": trait is lower/larger in the control. In R. biternatus, df= 36; in R. pseudotrullifolius, df=26.
Plant Leaf number 1 Leaf number 2 Plant height Number of leaves Petiol length Lamina length Lamina width Petiol length Lamina length Lamina width Control vs warming P F P F P F P F P F P F P F P F R. biternatus 0.14 3.62 0.0091 10.94 < 0.085 4.12 0.12 2.44 0.11 4.67 0.03 7.19 < 0.066 1.07 0.0003 5.8 < R. pseudotrullifolius 0.54 0.66 1 1.36 0.43 0.18 0.033 3.1 < 0.077 2.6 0.85 1 0.70 0.39 0.90 0.6
175
Chapitre 3. Costs and limits of plasticity
Table 3. ANOVA anylises indicating differences of performance (photosynthesis efficiency and leaf spread) between treatments in P. antiscorbutica. P-values (P) after sequential Bonferroni's correction and F values (F) are indicated. "": performance is lower/higher in the first treatment indicated. N= 10 in each treatment. Df=18.
Photosynthesis efficiency Leaf spread P F P F Control vs warming 0.11 2.8 0.90 0.014 Control vs shading 0.19 1.81 0.09 3.32 Control vs warming-shading 0.24 1.51 0.0002 31.74 > Shading vs warming-shading 0.58 0.32 0.02 8.02 > Warming vs shading 0.89 0.018 0.091 3.11 Warming vs warming-shading 0.42 0.69 0.0002 27.92 >
176
Chapitre 3. Costs and limits of plasticity
Table 4. ANOVA analyses indicating differences of performance (percentage of survival, photosynthesis efficiency and leaf spread) between warming and control in R. biternatus and R. pseudotrullifolius. P-values (P) after sequential Bonferroni's correction and F values (F) are indicated. "": performance is lower/higher in the control. In R. biternatus, df= 36; in R. pseudotrullifolius, df=26. Percentage of Photosynthesis survival efficiency Leaf spread P F P F P F R. biternatus Control vs warming 0.14 1.01 0.90 0.014 0.51 0.2 R. pseudotrullifolius Control vs warming 0.14 0.93 0,08 3.24 < 0.11 1.65
177
Chapitre 3. Costs and limits of plasticity
Table 5. ANOVA analyses indicating differences of the degree of plasticity between treatments in P. antiscorbutica. "" : trait is lower/larger in the first treatment indicated. P-values (P) and F values (F) are indicated. N=10 in each treatment. df=16.
Plant Leaf number 1 Leaf number 2 Plant height Number of leaves Petiole length Lamina length Lamina width Petiole length Lamina length Lamina width P F P F P F P F P F P F P F P F Warming vs shading 0.054 4.31 > 0.13 2.53 0.28 1.23 0.19 1.87 0.036 5.19 > 0.35 0.94 0.38 0.83 0.30 1.18 Warming vs warming-shading 0.13 2.55 0.54 0.38 0.87 0.03 0.35 0.91 0.012 7.97 > 0.72 0.13 0.06 3.97 0.017 6.99 > Shading vs warming- shading 0.64 0.22 0.43 0.64 0.23 1.54 0.71 0.15 0.41 0.73 0.25 1.45 0.34 0.95 0.31 1.09
178
Chapitre 3. Costs and limits of plasticity
Appendix 1. MANOVA analyses indicating trait differences among source mothers (R. biternatus and R. pseudotrullifolius) or source populations (P. antiscorbutica) in the control condition. Traits used in the analyses are: plant height, number of leaves, and in the 2 tallest leaves: petiole length, lamina length and lamina width. P-values (P), F values (F) and degree of freedom (Df) are indicated.
MANOVA P F Df R. biternatus 0.0001 1.55 125 R. pseudotrullifolius 0.021 1.28 88 P. antiscorbutica 0.0001 10.7 305
179
Chapitre 3. Costs and limits of plasticity
Appendix 2. ANOVA analyses indicating in P. antiscorbutica whether a) degree of phenotypic integration differs before and after application of treatments b) phenotypic integration differs among treatments. P-values (P), F values (F) and Degree of freedom (Df) are indicated.
a) ANOVA P F Df Warming 0.58 0.58 9 Shading 0.13 1.7 9 warming-shading 0.4 0.89 9
b) ANOVA P F Df Control vs warming 0.84 0.04 18 Control vs shading 0.32 1.06 18 Control vs warming- shading 0.29 1.19 18
180
Chapitre 3. Costs and limits of plasticity
Appendix 3. Principal component analysis (PCA) scattering populations based on the means of trait values of the population. Variable factor map and individual factor map from PCA are presented. The Variable Factor Map shows correlations among traits used in the analysis, and gives an understanding of how observations will be scattered along the individual factor map. The individual factor map shows trait differences among populations, and is interpreted based on the variable factor map. Different colors are used to dissociate populations in different treatments, as indicated in the legend. The first two axes of the analysis are presented, and percentages of variance explained by the axes are indicated.
181
Chapitre 3. Costs and limits of plasticity
Figure 1. Illustration of experimental design. The number of plants used per treatments, per species in one on two replicates are indicated.
Figure 2. Negative relationship between the degree of plasticity and performance in the warming treatment in P. antiscorbutica. P-values and adj.r² are indicated, N=10.
Figure 3. Negative relationship between the degree of plasticity and performance in the warming- shading treatment in P. antiscorbutica. P-values and adj.r² are indicated, N=9.
Figure 4. Positive relationship between phenotypic integration and the degree of plasticity of leaf
1 width or leaf 2 length in the shading treatment in P. antiscorbutica. N=9.
182
Chapitre 3. Costs and limits of plasticity
Figure 1
Shadi g treat e t War i g-Shadi g treat e t 16 plants from 5 source populations (80 16 plants from 5 source populations (80 plants) in P. antiscorbutica plants) in P. antiscorbutica
Co trol War i g treat e t 16 plants from 5 source populations (80 16 plants from 5 source populations (80 plants) in P. antiscorbutica plants) in P. antiscorbutica 4 plants from 19 mothers (76 plants) in R. 4 plants from 19 mothers (76 plants) in R. biternatus biternatus 4 plants from 14 mothers (56 plants) in R. 4 plants from 14 mothers (56 plants) in pseudotrullifolius R. pseudotrullifolius
183
Chapitre 3. Costs and limits of plasticity
Figure 2
spread
Degree of plasticity of leaf width (absolute difference between warming and control)
184
Chapitre 3. Costs and limits of plasticity
Figure 3
Degree of plasticity of leaf 2 width (absolute difference between warming-shading and control)
185
Chapitre 3. Costs and limits of plasticity
Figure 4
186
187
Chapitre 4: Etude préliminaire du transcriptome
Chapitre 4 : Etude préliminaire de la variation du transcriptome, en relation avec l’environnement et le phénotype chez P. antiscorbutica
INTRODUCTION
Les îles Kerguelen subissent un changement climatique rapide et intense (Frenot et al., 2006 ;
Lebouvier et al., 2011). Ces îles, localisées dans l’Océan Indien Austral à 3500km des continents les plus proches, sont extrêmement isolées et n’offrent ainsi pas de voie de migration possible. Les espèces végétales des îles Kerguelen n’auront alors pas d’autre possibilité pour se maintenir que celle de faire face aux changements environnementaux.
Au sein d’une espèce, les traits des individus peuvent varier en relation avec les conditions environnementales locales (Van Valen, 1973 ; Schichting, 1986). La variation des traits observée entre des individus soumis à différentes conditions locales peut résulter de différences génétiques entre individus (Kanaga et al., 2008) ou d’une réponse plastique (Sultan, 2000). Il est communément accepté que la variation des traits observée entre les individus d’une espèce contribue au potentiel adaptatif de l’espèce (Nicotra et al., 2010). Analyser comment les individus d’une espèce peuvent s’ajuster à différentes conditions environnementales donne une indication sur la capacité de cette espèce à faire face à des modifications environnementales futures. La relation existant entre les individus et un facteur environnemental peut servir de proxy pour estimer la réponse future des individus à une modification de ce facteur.
Le transcriptome représente l'ensemble des ARN transcrits, et traduit le niveau d’expression des différents gènes d’un organisme. L’étude du transcriptome témoigne à la fois de la diversité génétique des individus et de la réponse de ces individus à leur environnement. Le transcriptome est très réactif aux conditions environnementales, étant moins sujet à l’inertie de l’histoire de vie des individus que le phénotype. L’étude du transcriptome permet de déterminer les facteurs environnementaux, traits et gènes impliqués dans l’adaptation (Hoffmann et Daborn, 2007 ;
Voelckel et al., 2008). Lockhart, Voelckel, Becker et leurs collaborateurs, ont montré que l’étude
188
Chapitre 4: Etude préliminaire du transcriptome du transcriptome permettait de déterminer les gènes déterminant les traits fonctionnels sous sélection (Voelckel et al., 2008, 2012 ; Becker et al., 2013).
Pringlea antiscorbutica est une espèce végétale endémique de la province Sud de l’Océan Indien (Van der Putten et al., 2010). Cette espèce occupe des habitats très variés dans les Iles
Kerguelen. On trouve ainsi des populations de P. antiscorbutica des pelouses littorales exposées aux embruns aux prairies luxuriantes intérieures des versants abrités du vent, et jusqu’aux champs d’altitude (fell-fields) exposés aux vents, températures froides et aux nombreux cycles de gel/dégel (Hennion & Walton 1997b). Pringlea antiscorbutica colonise de préférence les environnements humides et bien drainés (Hennion & Walton 1997b).
Déterminer comment P. antiscorbutica répond et s’adapte à diverses conditions environnementales permettra de mieux appréhender son potentiel adaptatif. Nous étudierons le transcriptome et le phénotype de 8 individus de P. antiscorbutica vivant dans 3 populations soumises à des conditions environnementales que nous caractériserons. Aucune étude n’a à ce jour
été réalisée sur l’expression génétique de Pringlea antiscorbutica, ce qui nécessitera la création d’une banque EST de référence pour cette espèce. L’objectif est de déterminer les gènes dont l’expression varie entre les populations, et de mettre en relation cette variation avec l’environnement ou le phénotype.
189
Chapitre 4: Etude préliminaire du transcriptome
MATERIELS ET METHODES
1. Echantillonnage des individus-
Nous avons étudié 3 populations de P. antiscorbutica, situées dans deux régions différentes des
îles Kerguelen : l’île Australia (6 individus provenant de 2 populations), et Val Studer (2 individus provenant d’1 population). Les populations d’Australia sont situées à basse altitude (15m), tandis que la population de Val Studer est située à 285m d’altitude. L’échantillonnage a été réalisé sur une courte période (6 semaines) de mi- décembre 2011 à mi-janvier 2012. L’échantillonnage a été réalisé entre 11h et 17h pour se placer autour du maximum d’activité photosynthétique et éviter la variation journalière du transcriptome. Pour chaque individu, 2 à 4 feuilles, du même stade de développement (i.e jeune feuille photosynthétique) ont été prélevées, stockées dans du RNA later et conservées à -20°C. Pour chaque individu ont été mesurés la taille de la plante, son diamètre, la longueur et la largeur de ses feuilles. En laboratoire, l’extraction des ARNm été ensuite réalisée à l’aide du RNeasy Plant Mini Kit de Qiagen (Gehrig et al., 2000).
2. Mesures environnementales-
Dans chaque population, la teneur en eau et le pH du sol ont été mesurés. Trois échantillons de sol
(20mL) ont été prélevés par population, au niveau de la rhizosphère. Pour déterminer la teneur en eau (%), la moitié du sol a été pesée, puis séchée à 105°C pendant 48 et pesée une nouvelle fois
(Hermant et al., 2013). Le sol restant a été dilué avec un volume connu d’eau distillée puis laissé reposer 18 à 24h afin de permettre la sédimentation des particules du sol. Le pH a ensuite été mesuré à l’aide d’un pH mètre (BASIC 20 PLUS CRISON, résolution 0.01pH). La température a
été enregistrée, à l’aide de sondes situées au niveau des plantes, toutes les 30 minutes en janvier et février.
190
Chapitre 4: Etude préliminaire du transcriptome
3. Analyse des données-
Le séquençage des ARNm a été réalisé par la société Novogene. L’assemblage de novo a été réalisé
à l’aide du logiciel Trinity version 2.2.0. (Haas et al., 2013). Par population, les séquences assemblées ont été annotées à l’aide de blastx (version 2.2.27; e-value cutoff de 0,0001) en utilisant la référence : TAIR10 protein database. Le meilleur score a été utilisé pour l’annotation. Les séquences avec la même annotation entre les populations ont été comparées, et la plus longue séquence ajoutée à la librairie. Les séquences des différents individus ont été comparées à la librairie à l’aide du logiciel bowtie2 et le nombre d’occurrences pour chaque séquence a été déterminé à l’aide de HTSeq-count. Les gènes ayant un degré d’expression différent (DEGs) entre individus ont été déterminés à l’aide de DESeq2 (p-value ajustée de 0,05), et leur représentation a
été analysée à l’aide du logiciel Panther (http://pantherdb.org/). La représentation spatiale des différents individus a été réalisée à l’aide du package MADE4 (Culhane et al., 2005) sous R.3.3.1
(R Core Team). Les analyses en composante principale représentant les traits ou l’environnement ont été réalisées à l’aide du package FactomineR (Lê, Josse et Husson, 2008) sous R3.3.1 (R Core Team, 2012).
191
Chapitre 4: Etude préliminaire du transcriptome
RESULTATS ET INTERPRETATIONS
Inféodée aux habitats humides mais bien drainés, Pringlea antiscorbutica, est décrite comme particulièrement vulnérable face aux changements climatiques en cours (Hennion et Walton,
1997b ; Le Roux et al., 2008 ; Lebouvier et al., 2011). Cette espèce est trouvée dans des environnements variés dans les îles Kerguelen. Etudier comment les individus ont répondu ou se sont adaptés à différents environnements donne des indications sur la capacité de cette espèce à faire face aux changements en cours et à venir. Cette étude présente les résultats préliminaires sur la variation du transcriptome en relation avec l’environnement et le phénotype chez P. antiscorbutica.
Cette étude constitue la première analyse du transcriptome de Pringlea antiscorbutica.
L’assemblage de novo qui a été réalisé a ainsi permis de constituer la première librairie EST pour cette espèce. Cette librairie, qui servira de référence, constitue une étape préliminaire essentielle à la réalisation de futures analyses du degré d’expression du génome.
Nos premiers résultats montrent une forte corrélation entre l’expression du génome, le phénotype et les caractéristiques environnementales (Figure 1). Ce premier constat permet d’observer que le transcriptome est bien à l’interface entre environnement et phénotype.
On observe particulièrement une différence entre les populations d’Australia (Australia 1 et Australia 2), situées à basse altitude (respectivement 5 et 15m), et la population de Val Studer, située à haute altitude (285m) (Appendice 1). Cette différence est observée au niveau du phénotype et du transcriptome. Les plantes en altitude sont caractérisées par des valeurs de traits plus petites
(Figure 1). Au niveau du transcriptome, on observe que la population située en altitude exprime des gènes de tolérance au froid, en comparaison avec une population de faible altitude (Figure 3).
Egalement, les individus présents à Val Studer expriment des gènes de réponse à un manque d’eau.
Ceci est cohérent avec la teneur en eau du sol plus faible relevée à Val Studer – en comparaison avec Australia 2 (Figure 1 ; Figure 3). Egalement les individus de Val Studer expriment des glucosinolates. Les glucosinolates sont des métabolites secondaires, contenant un groupement
192
Chapitre 4: Etude préliminaire du transcriptome azote et sulfate (Wittstock et Halkier, 2002). Ces composés sont notamment caractéristiques des
Brassicaceae (Wittstock et Halkier, 2002) et ont été analysés chez P. antiscorbutica (Barillari et al., 2005). Voelckel et al. (2008) ont étudié la variation du transcriptome chez trois espèces de
Pachycladon le long de gradients environnementaux dans les Alpes de Nouvelle-Zélande. Ces auteurs ont montré une variation de l’expression de gènes responsables de la synthèse et de l’hydrolyse des glucosinolates (Voelckel et al., 2008). Voelckel et al. (2008) précisent que la variation des glucosinolates peut être une réponse à la pression d’herbivores ou à des facteurs abiotiques. Les îles Kerguelen sont dépourvues d’herbivores, et ici les patterns de glucosinolates devraient correspondent à des facteurs abiotiques.
La différence au niveau de l’environnement, du phénotype et du transcriptome est plus importante entre les populations des deux régions :Australia et Val Studer, qu’entre les deux populations d’une même région (Australia) (Figure 1). On note cependant que la teneur en eau est plus élevée à Australia 2 qu’à Australia 1. Cette différence se retrouve au niveau du transcriptome ou les individus d’Australia 1 expriment des gènes de réponse au manque d’eau (Figure 2). Ces résultats préliminaires constituent une première approche de la variation du transcriptome chez Pringlea antiscorbutica. L’étude en cours portera sur 17 individus issus de 6 populations sur le terrain. Cette première analyse a principalement permis la construction d’un transcriptome de référence pour cette espèce. De plus, les comparaisons préliminaires entre transcriptome, phénotype et environnement montrent des résultats encourageants. On note ainsi une correspondance étroite entre environnement, transcriptome et phénotype.
193
Chapitre 4: Etude préliminaire du transcriptome
Figure 1. Analyses multivariées représentant a) les différences entre populations caractérisées par des facteurs environnementaux, b) les différences entre individus (en noir) et population (en rose) caractérisées par les traits morphologiques et c) les différences entre les individus caractérisée par l’expression du génome.
194
Chapitre 4: Etude préliminaire du transcriptome
Figure 2. GO analyse, représentant les fonctions surexprimées chez les individus d’Australia 2 en comparaison avec les individus d’Australia 1. Les tons plus foncés représentent les fonctions plus exprimées dans la population d’Australia 2.
195
Chapitre 4: Etude préliminaire du transcriptome
Figure 3. GO analyse, représentant les fonctions surexprimées chez les individus de Val Studer en comparaison avec les individus d’Australia 2. Les tons plus foncés représentent des fonctions plus exprimées dans la population de Val Studer.
196
Chapitre 4: Etude préliminaire du transcriptome
Appendice 1. Teneur en eau, pH, altitude et température dans les différentes populations
Teneur en eau du Température moyenne pH Altitude (m) Population sol (%) (Janvier-Février; °C) Australia 1 9,9 5,52 15 9,8 Australia 2 44,1 5,02 15 9,8 Val Studer 9,8 6,81 285 6,9
197
DISCUSSION GENERALE
1. Avancées conceptuelles
2. Potentiel adaptatif des espèces de Kerguelen
198
Discussion générale
1. Avancées conceptuelles
1.1. Le degré d’intégration phénotypique est indépendant de l’environnement
La variation de la valeur d’un trait d’un organisme est majoritairement considérée indépendamment de celle des autres traits (Valladares et al., 2007). Or, les traits d’un organisme sont corrélés entre eux. On définit le degré d’intégration phénotypique comme le nombre et la force des corrélations entre les traits (Pigliucci, 2003). L’intégration phénotypique a principalement été étudiée au niveau des traits floraux dans le cadre de la relation entre fleur et insecte pollinisateur (Murren, 2012 ;
Armbruster et al., 2014). L’intégration phénotypique existe cependant également au niveau des traits végétatifs (Mallitt, 2010 ; Hornoy et al., 2011). L’intégration phénotypique pourrait avoir des bases génétiques, développementales, fonctionnelles ou environnementales (Murren, 2002 ;
Armbruster et al., 2014). Jusqu’à présent, les causes et les conséquences de l’intégration phénotypique sont encore peu connues (Gianoli, 2004 ; Gianoli et Palacio-Lopez, 2009 ; Mallitt et al., 2010). Dans cette thèse, nous avons étudié l’influence de l’environnement sur l’intégration phénotypique aussi bien in situ dans des populations naturelles, qu’en conditions contrôlées dans des populations exposées à des facteurs abiotiques simples ou combinés. De plus, nous avons étudié l’influence de l’intégration sur la variabilité des traits (incluant la variabilité potentielle, i.e. plasticité) et la performance des individus.
L’influence de l’environnement sur le degré d’intégration phénotypique reste encore peu connue
(Mallitt et al., 2010). Il a été montré que des conditions environnementales stressantes peuvent augmenter l’intégration phénotypique (Schlichting 1989 ; Gianoli, 2004 ; Garcia-Verdugo et al.,
2009). Cependant, ces études sont menées en conditions contrôlées, sur des plantes soudainement exposées à un stress abiotique simple. A notre connaissance, seuls Mallitt et al. (2010) ont étudié l’influence combinée de deux facteurs abiotiques (sécheresse et intensité lumineuse) sur le degré 199
Discussion générale d’intégration des individus chez l’espèce végétale Lepidium bonariense, toujours en conditions contrôlées. Ces auteurs n’ont pas trouvé de relation entre l’exposition à différent facteurs environnementaux et le degré d’intégration. Nos résultats vont dans la même direction : en conditions contrôlées nous n’avons pas trouvé de variation de l’intégration phénotypique en réponse à l’augmentation de la température, la diminution de l’intensité lumineuse ou la combinaison des deux facteurs chez Pringlea antiscorbutica. En conditions contrôlées réalistes, i.e. multifactorielles, l’intégration ne semble donc pas dépendre de l’environnement.
Dans des telles conditions contrôlées les plantes sont exposées à un changement environnemental brusque. Or, en conditions naturelles, les plantes peuvent être exposées aux mêmes contraintes environnementales depuis des générations. Si l’intégration phénotypique ne change pas instantanément en réponse à l’environnement, il se peut que l’environnement façonne l’intégration phénotypique à plus long terme. Peu d’études ont étudié l’intégration phénotypique en conditions naturelles. Cette thèse fait suite à l’étude de Hermant et al. (2013) qui ont montré que le degré d’intégration augmentait avec le degré d’endémisme chez 14 espèces végétales des îles
Kerguelen. A notre connaissance, aucune étude n’a jusqu’à lors été menée au niveau intra- spécifique. Or, au sein d’une même espèce, les plantes sont réparties en diverses populations, qui peuvent être exposées à différentes conditions environnementales. Un des objectifs de la thèse est donc d’appréhender les patrons d’intégration phénotypique en condition naturelle au niveau intra- spécifique. Nous avons étudié la relation entre le degré d’intégration trouvé au sein des populations et les conditions environnementales des populations (teneur en eau, pH, conductivité du sol, température moyenne ou altitude). Nous n’avons pas trouvé de relation entre le degré d’intégration phénotypique des individus d’une population et les conditions environnementales de la population chez 4 espèces végétales.
200
Discussion générale
1.2. L’Intégration augmente la variabilité des traits établie et potentielle mais pas la performance.
Les conséquences de l’intégration phénotypique sur les plantes sont encore mal connues (Gianoli et Palacio-Lopez, 2009 ; Mallitt et al., 2010). Nous avons analysé la relation entre le degré d’intégration phénotypique et la performance des plantes, ainsi que l’influence de l’intégration sur la variabilité des traits aussi bien existante que potentielle (i.e. plasticité). Premièrement, il a été suggéré que l’intégration phénotypique pouvait augmenter la fitness des individus (Schlichting,
1989). Nous avons observé l’influence du degré d’intégration phénotypique sur la performance des plantes, aussi bien in situ qu’en réponse à des changements de conditions environnementales en conditions contrôlées. Nous n’avons pas trouvé de lien entre intégration phénotypique et performance des plantes.
Deuxièmement, l’intégration phénotypique est communément perçue comme une contrainte pour la variation des traits des organismes. A notre connaissance, seuls Gianoli et Palacio-Lopez
(2009) ont étudié la relation entre intégration phénotypique et plasticité. Ces auteurs ont montré chez deux espèces végétales (Convolvulus chilensis et Lippia alba) une relation négative entre le degré d’intégration phénotypique et la plasticité moyenne des traits. En étudiant Pringlea antiscorbutica nous n’avons pas trouvé de relation négative entre le degré d’intégration phénotypique et la plasticité des différents traits des plantes. Au contraire, nous avons montré une relation positive entre intégration phénotypique et plasticité de traits, tel que la longueur du limbe.
Cette différence de résultat entre les deux études (relation positive ou négative) est trouvée en réponse au même traitement expérimental (diminution de luminosité). Cependant, nous avons
étudié l’influence de l’intégration phénotypique à un temps t sur le degré de plasticité des traits à un temps t+1, en vérifiant que le degré d’intégration n’était pas différent entre t et t+1. Nous testons si des populations ayant des degrés d’intégration différents ont des degrés de plasticité différents
201
Discussion générale en réponse à un changement environnemental. En revanche, Gianoli et Palacio-Lopez (2009), ont
étudié la relation entre intégration à t+1 et plasticité à t+1. L’étude ne précise pas le degré d’intégration avant le traitement expérimental ce qui ne permet pas de déterminer si une forte intégration a pour conséquence une faible plasticité. La corrélation négative entre intégration et plasticité à t+1 pourrait aussi s’expliquer par une baisse d’intégration provoquée par une forte plasticité notamment dans un organisme peu intégré dans lequel les traits varient indépendamment.
En bilan, nos résultats ne sont pas opposés à ceux de Gianoli et Palacio-Lopez (2009) qui eux n’ont pas teste une causalité
Troisièmement, la plasticité représente la réponse des individus à un changement environnemental rapide, cependant les plantes peuvent être soumises aux mêmes conditions environnementales depuis des générations. Sur le terrain, nous avons étudié l’influence de l’intégration phénotypique sur la variation des traits établie au sein des populations. Nous avons montré, dans le chapitre 1, que l’intégration phénotypique et non l’environnement influençait la variation des traits au sein des populations, qui plus est, de manière positive. Encore peu d’auteurs prennent en compte l’intégration phénotypique lorsqu’ils étudient la variation des traits des individus (Mallitt et al., 2010). Or, nous montrons que l’intégration phénotypique peut s’avérer être un facteur prépondérant, favorisant la variation des traits des individus.
Les travaux de cette thèse apportent, à notre connaissance, pour la première fois des résultats sur les conséquences du degré d’intégration phénotypiques en termes de variabilité des traits et de performance. Il est communément suggéré que l’intégration phénotypique peut augmenter la performance des individus mais diminuer leur capacité de variation des traits (Schlichting, 1989 ;
Gianoli et Palacio-Lopez, 2009; Mallitt et al., 2010). Dans cette thèse, nous n’avons pas trouvé de relation entre l’intégration phénotypique et la performance des plantes. Par contre, l’intégration phénotypique est positivement corrélée à la variation des traits trouvée in situ, et à la plasticité des 202
Discussion générale traits des individus. Ainsi, l’intégration phénotypique pourrait participer positivement au potentiel adaptatif des espèces. Particulièrement, il est nécessaire de ne plus considérer l’intégration phénotypique comme un frein à la variation des traits.
1.3. Flexibilité fonctionnelle des métabolites
Les métabolites secondaires permettent aux organismes de tolérer leur environnement et répondre
à ses modifications (Tohge et al., 2013 ; Tiburcio et al., 2014). Il existe une grande diversité de métabolites secondaires, qui s’accompagne d’une variabilité fonctionnelle (Wink, 2013).
Premièrement, une convergence fonctionnelle des métabolite secondaire a été observée entre lignées (Tiburcio et al., 2014). A travers les lignées, différents métabolites assurent la même fonction et sont dits fonctionnellement redondants (Wink, 2003). Cette redondance fonctionnelle a très peu été étudiée au sein des espèces (Sulmon et al., 2015). Dans le chapitre 2, nous avons montré qu’à l’intérieur des espèces, plusieurs métabolites pouvaient assurer une même fonction. De plus, différents métabolites secondaires peuvent être utilisés pour une même fonction entre des populations d’une même espèce. Secondement, des plantes vivant dans des environnements différents peuvent montrer une versatilité fonctionnelle (Wink, 2003). La versatilité d’un métabolite représente le fait que ce métabolite ait des fonctions différentes dans des organes différents d’une plante ou entre des plantes vivant dans des environnements différents (Wink, 2003
; Lehmann et al., 2010 ; Samal et al., 2011 ; Di Ferdinando et al., 2014). Dans le chapitre 2, nous avons montré que des métabolites donnés pouvaient également avoir des rôles différents, et parfois contraires, dans un même organe d’une même espèce et un même environnement entre des régions différentes. Nous sommes, à notre connaissance les premiers à mettre en évidence une redondance et une versatilité fonctionnelle de métabolites secondaires à un niveau intra-spécifique. Cette étude
203
Discussion générale souligne la variabilité fonctionnelle des métabolites. Redondance et versatilité fonctionnelles entre individus de différentes populations montrent qu’il est nécessaire de prendre en compte des individus provenant de différentes populations afin de déterminer les rôles des métabolites secondaires.
1.4. Plasticité plus importante ne signifie pas meilleure performance
La plasticité phénotypique représente la capacité des individus à ajuster la valeur de leurs traits en réponse à des changements environnementaux (Sultan, 2000). Les plantes, en tant qu’organismes sessiles, n’ont d’autres moyens que de faire face aux changements de conditions environnementales. La plasticité représente un moyen majeur par lequel les plantes font face aux changements environnementaux (Valladares et al., 2007). Si la plasticité est fréquemment montrée comme étant un atout pour les plantes, elle présente également des couts. La notion de coûts de plasticité est récente dans la littérature (Valladares et al., 2007 ; Auld et al., 2009 ; Valladares et
Niinemets, 2008). La réponse plastique repose sur une cascade physiologique et morphologique qui induit des coûts à l’organisme (Vinocur et Altman, 2005). De façon théorique, on distingue les coûts constitutifs, des coûts induits. Les coûts constitutifs, de perception et transduction du signal environnemental sont constants et permanant (Auld et al., 2009). Les couts induits font référence aux couts de la modification du phénotype en réponse à l’environnement (Chevin et al., 2010). Les coûts associés à la plasticité pourraient affecter la performance des individus plastiques. De plus en plus d’auteurs remettent alors en cause la relation positive communément acceptée entre plasticité et performance des individus (Valladares et al., 2005 ; Ghalambor et al., 2007 ; Valladares et al., 2007). A ce jour, peu d’études ont analysés les couts de la plasticité et il reste à déterminer si plasticité signifie réellement meilleure performance (Pigliucci 2005 ; Valladares et al., 2007 ;
204
Discussion générale
Auld et al., 2009). Dans le chapitre 3, nous avons montré chez Pringlea antiscorbutica que les plantes ayant une plus grande plasticité n’avaient pas une performance accrue. Au contraire, si le degré de plasticité est trop important, on observe une diminution de la performance liée au degré de plasticité. La baisse de performance observée lorsque le degré de plasticité est important pourrait
être imputée aux coûts induits (i.e. de modification phénotypique) qui deviendraient trop importants. En revanche, nous n’avons pas pu mettre en évidence de coûts constitutifs de la plasticité. Nous suggérons que les plantes étudiées sont toutes plastiques dans une certaine mesure et doivent s’acquitter des mêmes coûts constitutifs de perception et maintenance du signal environnemental. L’étude menée apporte des preuves expérimentales de la relation entre plasticité et performance des plantes. Elle met en lumière la nécessité de ne pas considérer la réponse plastique d’une plante à un changement environnemental comme preuve de sa capacité à rester performant face au changement.
1.5. Perspectives
Dans cette thèse, nous avons montré que le degré d’intégration phénotypique n’était pas déterminé par des facteurs environnementaux (Chapitre 1). Il reste alors à établir ce qui détermine le degré d’intégration phénotypique. L’intégration phénotypique peut-être influencée par des facteurs environnementaux que nous n’avons pas pris en compte in situ ou en conditions contrôlées.
Cependant, les facteurs pris en compte ont été sélectionnés pour avoir une influence prépondérante sur les plantes dans les îles Kerguelen. Si l’intégration phénotypique n’est ni plastique en laboratoire, ni en relation avec l’environnement in situ, on peut suggérer que les différents degrés d’intégration phénotypique entre populations sont le résultat d’une évolution neutre entre les populations. Des études futures pourraient déterminer les bases moléculaires de l’intégration, telles que la pléiotropie ou le déséquilibre de liaison. Egalement, Les conséquences de l’intégration
205
Discussion générale phénotypique chez les organismes restent encore à être élucidées. Un fort degré d’intégration définit une forte cohérence entre les valeurs des traits d’un organisme. Cette forte cohérence peut permettre à un organisme de mieux tolérer des stress environnementaux extrêmes, ce qui n’a pas
été testé dans le cadre de la thèse.
Ce projet de thèse a permis d’approfondir les connaissances sur la variabilité des métabolites secondaires, notamment leur variabilité fonctionnelle (Chapitre 2). Les résultats apportés soulèvent alors de nouvelles questions sur le fonctionnement des métabolites secondaires. Des études futures pourraient déterminer si les patrons que nous avons observés sont héritables ou dus à des réponses plastiques. Egalement il reste à déterminer pourquoi différents métabolites sont utilisés pour une même fonction (i.e. redondance fonctionnelle) entre deux régions à l’environnement similaire, alors que le même pool de métabolite est présent dans ces deux régions. Il reste à déterminer pourquoi un même composé réagit différement face à un même environnement dans deux régions distantes (i.e. versatilité fonctionnelle). De manière générale, des études ultérieures pourraient déterminer les facteurs, comme par exemple les flux de gêne, qui peuvent favoriser ou diminuer une redondance ou versatilité fonctionnelle. Ces résultats soulèvent la question des coûts et bénéfices d’une telle variabilité fonctionnelle.
206
Discussion générale
Figure 4. Ratio théorique de performance entre individus plastiques et non plastiques, en réponses
à différentes intensités de changements environnementaux. La ligne en pointillé indique le cas où la performance des individus plastique est égale à la performance des individus non plastiques.
207
Discussion générale
Dans le chapitre 3, nous avons essayé de déterminer des coûts et limites de la plasticité.
Notamment, nous avons montré que le degré de plasticité était plus important en réponse à un réchauffement qu’en réponse à un réchauffement accompagné d’un ombrage. Ces deux facteurs ont un effet opposé sur la croissance, qui explique en partie la baisse du degré de plasticité. Il serait intéressant de tester l’effet d’un réchauffement accompagné d’une augmentation de luminosité pour voir si les deux effets se cumulent, et augmentent le degré de plasticité. Egalement, des études ultérieures pourraient apporter des précisions sur les coûts de la plasticité, en étudiant la relation entre degré de plasticité et performance, face à des changements environnementaux d’intensité ou de complexité différentes.
Nous suggérons que les coûts de la plasticité pourraient dépendre de l’intensité du changement environnemental. Ainsi, la relation entre plasticité et performance dépendrait de l’intensité des changements, comme présenté dans la figure 4 p204. Dans un environnement stable, on peut s’attendre à ce que les individus plastiques soient moins performants que les individus non plastiques, dû aux couts constitutifs de la plasticité. Dans un environnement avec des changements d’intensité modérée, les individus qui présentent un fort degré de plasticité ajustant mieux leurs traits à l’environnement, devraient être plus performant que des individus peu ou non-plastiques.
Dans un environnement avec des changements de forte intensité, une forte plasticité entraîne une modification extrême des traits et de forts coûts induits associés. Les coûts induits de la plasticité peuvent alors contrebalancer et même dépasser ses bénéfices, suggérant une relation négative entre plasticité et performance. Dans le chapitre 3, nous avons essayé d’explorer ces hypothèses, nous avons cependant seulement montré un effet négatif de la plasticité sur la performance en réponse à un réchauffement.
208
Discussion générale
2. Potentiel adaptatif des espèces de Kerguelen
2.1. Variation intra-population : probablement pas menacée par les changements climatiques
Une forte variation des traits existant au sein des populations a été reconnue comme participant au potentiel adaptatif des populations, et ainsi des espèces (Nicotra et al., 2010 ; Bolnick et al., 2011).
L’environnement peut influencer la variation intra-population des traits (Debat et David, 2001), notamment des stress environnementaux peuvent diminuer cette variation (Gicquiaud et al., 2002).
Les îles Kerguelen sont soumises à un changement climatique rapide et intense caractérisé particulièrement par une augmentation des températures et une diminution des précipitations
(Lebouvier et al., 2011). La modification des conditions environnementales (notamment la température et la teneur en eau) peut entraîner un stress chez les espèces étudiées (Le Roux et al.,
2008), pouvant diminuer la variation intra-population des traits. De plus, les espèces étudiées sont caractérisées par un fort degré d’intégration phénotypique (Hermant et al., 2013). Or, il a été communément suggéré qu’un fort degré d’intégration esr relié à une faible variation des traits
(Schlichting, 1989 ; Gianoli et Palacio-Lopez, 2009). Dans le chapitre 1, nous avons cherché à savoir si des facteurs environnementaux ou le degré d’intégration phénotypique entraînaient une diminution de la variation intra-population des traits, et ainsi du potentiel adaptif des espèces. Nous avons montré que les facteurs environnementaux (tels que la température, la saturation en eau du sol, son pH et sa conductivité) n’influençaient pas la variation intra-population des traits, au contraire de l’intégration phénotypique, comme discuté précédemment. Plus précisément, nous avons montré chez P. antiscorbutica, R. biternatus et R. pseudotrullifolius que le degré d’intégration phénotypique était positivement corrélé à la variation intra-population des traits. En revanche, chez R. moseleyi, il n’y a pas de relation entre intégration et variation intra-population
209
Discussion générale des traits. Ainsi, globalement, le fort degré d’intégration des espèces étudiées ne serait pas un frein, bien au contraire, au potentiel adaptatif des espèces. Dans le cadre du changement climatique, une augmentation des températures ou une diminution de la saturation en eau par exemple n’entraîneraient pas de diminution de la variation intra-population des traits, et ainsi du potentiel adaptatif des espèces, dans les gammes environnementales étudiées.
2.2. Variation inter-population : faible variation chez R. moseleyi
Au sein d’une espèce, les traits peuvent varier entre populations présentant des conditions environnementales locales différentes. La capacité de faire face à différentes conditions environnementales participe au potentiel adaptatif de l’espèce. Dans le chapitre 1, nous avons montré que les traits des plantes varient en fonction des conditions environnementales chez P. antiscorbutica, R. biternatus et R. pseudotrullifolius, contrairement à R. moseleyi. La variation des traits observée entre populations exposées à des conditions environnementales différentes peut résulter d’une réponse plastique des individus ou d’une adaptation génétique aux conditions environnementales. Dans le chapitre 3, nous avons montré que les différences de traits entre populations persistent lorsque des graines provenant de différentes populations sont mises en culture dans les mêmes conditions environnementales chez P. antiscorbutica, R. biternatus et R. pseudotrullifolius. Ces différences entre populations peuvent être dues à des différences génétiques ou un effet maternel résiduel. On peut supposer que les différences entre populations naturelles sont le résultat à la fois d’adaptations génétiques et d’une réponse plastique des individus à l’environnement. Hennion et al. (1994) ont montré que la variabilité de la taille et la forme des feuilles chez R. biternatus, R. pseudotrullifolius et R. moseleyi comporte une part de réponse plastique principalement en rapport avec le degré et la fréquence de submergence, plasticité
210
Discussion générale moindre chez R. biternatus avec une part génétiquement fixée mais une amplitude globale de variabilité beaucoup plus importante.
On observe une relation entre variation inter-population des traits et gamme d’habitats occupée par l’espèce. Ainsi, chez les renoncules, R. biternatus est l’espèce qui présente la plus forte variation des traits en fonction de l’environnement – et ce, entre populations naturelles ou face à un changement environnemental appliqué de façon standardisée à de multiples espèces en laboratoire. Ranunculus biternatus est également l’espèce trouvée dans la plus large gamme environnementale. On trouve R. biternatus notamment dans des environnements plus secs que R. pseudotrullifolius et R. moseleyi. Egalement, Pringlea antiscorbutica présente une forte variation des traits en fonction de l’environnement et est trouvée dans la gamme environnementale la plus large. Ranunculus moseleyi, strictement aquatique et montrant une faible variation des traits entre populations apparaît comme étant particulièrement vulnérable face aux changements climatiques en cours.
2.3. Variation inter-population : performance plus faible en milieux secs et chauds.
Dans le chapitre 1, nous avons cherché les facteurs environnementaux qui influencent la performance des plantes. Particulièrement, en cohérence avec le changement climatique observé à
Kerguelen, nous nous sommes intéressés à l’influence d’une augmentation de la température et d’une diminution de la teneur en eau sur la performance des plantes. Nous avons montré que les plantes étaient moins performantes en terme de taille : plus petites (taille de la plante et de la plus grande feuille) en conditions plus sèches (R. biternatus et R. pseudotrullifolius), plus salées ou plus acides (R. biternatus et P. antiscorbutica). Ces résultats peuvent indiquer que ces conditions sont plus stressantes pour les plantes (Zhu, 2001 ; Grime, 2006 ; Farooq et al., 2009). Chez R. biternatus,
R. moseleyi et R. pseudotrullifolius, la performance (pourcentage de feuilles non-sénescentes) des
211
Discussion générale plantes est plus faible en milieux secs. Chez R. moseleyi, on observe une augmentation de la performance avec la température, chez R. biternatus au contraire, les plantes sont moins performantes lorsque la température est plus élevée. Les contraintes logistiques lors de l’échantillonnage ne nous ont pas permis d’étudier ces aspects chez P. antiscorbutica et R. pseudotrullifolius. Chez R. biternatus, nous avons également pu démontrer que la performance des plantes diminuait avec l’acidité et la conductivité des sols.
Déterminer comment la performance des individus varie le long de gradients environnementaux permet d’estimer l’impact des changements environnementaux à venir sur les plantes. De plus l’augmentation de l’évaporation des sols à prévoir sous le changement climatique à Kerguelen devrait engendrer une augmentation de la conductivité des sols ainsi qu’il a été observé sur l’île
Marion (Le Roux et al., 2008). D’après nos résultats, la diminution de la teneur en eau et l’augmentation de la conductivité des sols devraient alors diminuer la performance des espèces
étudiées à l’exception de R. pseudotrullifolius. En revanche, les îles Kerguelen offrent d’importants dénivelés et les espèces végétales pourraient palier à l’augmentation de la température en colonisant des environnements d’altitude (Bergstrom et Chown, 1999). Cependant, le changement climatique entraîne également l’augmentation de la vitesse du vent, favorisant l’évaporation du sol, particulièrement en altitude (Le Roux et al., 2008). Pringlea antiscorbutica, R. biternatus, R. pseudotrullifolius et d’autant plus R. moseleyi (uniquement aquatique) sont inféodés à des environnements humides (Hennion et al., 1994 ; 2006). Il est avéré sur l’île Marion que P. antiscorbutica (inféodée aux milieux bien drainés) et les Ranunculus (ayant des racines peu profondes) sont particulièrement sensibles à l’assèchement (Le Roux et al., 2008). De manière plus globale, les espèces végétales autochtones poussent habituellement dans des conditions environnementales humides, celles-ci caractérisant les habitats des îles Kerguelen. La teneur en eau des sols paraît être un élément prépondérant pour les espèces de Kerguelen. R. 212
Discussion générale pseudotrullifolius est une espèce halophile. La nécessité de pousser dans des environnements salins pourrait limiter les possibilités de colonisation des milieux d’altitude. Cette espèce est décrite uniquement côtière dans la littérature (Aubert de la Rüe, 1964), cependant elle est présente ponctuellement jusqu’à des altitudes modérées autour de 200m (observation personnelle, campagne d’été 2015-2016 IPEV PlantEvol ; données de distribution Réserve Naturelle des TAAF
/ Programme IPEV 136, non publié). Ainsi, Ranunculus pseudotrullifolius montre une amplitude
écologique encore à préciser et qui pourrait être déterminante face au changement climatique.
2.4. Communauté avoisinante : pas d’effets sur les traits ou la performance
Les communautés biotiques des îles Kerguelen sont considérées comme pauvres. De plus les populations de Ranunculus étudiées forment des tapis denses généralement monospécifiques.
Pringlea antiscorbutica est trouvé dans des habitats de basse altitude à végétation dense et pousse
également en altitude où la végétation est ouverte. Dans le chapitre 1, nous avons cherché à déterminer l’influence de la communauté avoisinante sur les traits et la performance des plantes.
Nous n’avons trouvé aucune relation entre la communauté avoisinante, que ce soit sa richesse ou sa diversité, et la moyenne ou l’intégration des traits chez R. biternatus, R. pseudotrullifolius ou R. moseleyi. De plus nous n’avons pas montré de relation entre la communauté avoisinante et la performance des plantes. Des résultats similaires ont été trouvés chez P. antiscorbutica (données non montrées). Ces résultats confortent l’hypothèse que les facteurs abiotiques - et notamment les caractéristiques du sol telles que teneur en eau, pH et conductivité – représentent les facteurs environnementaux limitants principaux aux îles Kerguelen. Ce constat pourrait cependant changer dans les prochaines années. Les conditions froides trouvées aux îles Kerguelen limitent les possibilités de colonisation d’espèces qui proviennent de zones plus tempérées (Frenot et al., 2006).
Le réchauffement climatique est alors supposé favoriser la colonisation des habitats des îles
213
Discussion générale
Kerguelen par des espèces introduites (Frenot et al., 2005, 2006). Ces espèces pourraient ainsi entrer en compétition et avoir une influence plus conséquente sur les espèces natives étudiées. On note d’ores et déjà le cas d’une plante introduite, le pissenlit qui colonise des habitats en lieu et place de l’espèce indigène Acaena magellanica (Frenot et al., 2006).
2.5. Réponses plastiques aux changements environnementaux : pas un témoin de performance
Nos études ont montré que les plantes étudiées répondent plastiquement aux changements environnementaux. Cependant, la capacité de réponse plastique des plantes ne témoigne pas forcément de leur capacité à rester performantes face aux changements. Il a été montré chez
Pringlea antiscorbutica une diminution de la croissance aérienne et racinaire en réponse à une exposition à la chaleur (Hummel et al., 2004a). Les traits morphologiques de R. biternatus et R. pseudotrullifolius sont connus pour être partiellement réversibles, on observe notamment une diminution de la taille foliaire lorsque le degré et la fréquence de submersion des plantes diminuent
(Hennion et al., 1994). Egalement, Hennion et Walton (1997a) ont trouvé que les individus de R. biternatus produisaient des akènes plus grands en conditions tempérées. Dans le chapitre 3, nous avons déterminé que P. antiscorbutica, R. biternatus et R. pseudotrullifolius montraient une réponse plastique face à une augmentation de température. Cette réponse plastique est caractérisée chez les 3 espèces par une augmentation de la valeur des traits sous une température plus élevée.
On note que davantage de traits répondent à l’augmentation de température chez R. biternatus par rapport à R. pseudotrullifolius. Chez P. antiscorbutica, nous avons également montré une réponse plastique à l’ombrage, caractérisée par une diminution des traits. Nous n’avons pas observé de différence de performance (pourcentage de feuilles non-sénescentes ou rendement photosynthétique) entre les plantes exposées à différentes conditions, et ce chez les 3 espèces
214
Discussion générale
étudiées. Nous avons cherché à savoir si les individus présentant un plus fort degré de plasticité
étaient plus performants chez P. antiscorbutica. Nous avons montré, bien au contraire, que le degré de réponse plastique pouvait être corrélé avec une baisse de la performance.
Les trois espèces étudiées ne montrent ici pas de baisse de performance face à une augmentation de la température. L’âge des plantes est ici à prendre en compte : 8 mois, il s’agit de plantes très jeunes, non matures. Pringlea antiscorbutica étant une espèce à grande longévité (Chapuis et al.,
2000), il est possible que la plasticité des individus âgés soit très différente. On ne connaît pas la durée de vie des trois espèces de renoncules étudiées mais il s’agit bien d’espèces pérennes. Il est
également possible que, chez ces jeunes plantes, la baisse de performance devienne significative à plus long terme. Hennion (com. pers.) a trouvé chez ces espèces que les plantes cultivées sous des conditions tempérées océaniques à Brest avec un arrosage optimal montrent dans les premiers mois une forte croissance mais également une sénescence accélérée suivie de mort après quelques mois.
2.6. Réchauffement et ombrage combinés limitent la plasticité et la performance
La modification simultanée de plusieurs facteurs environnementaux peut être plus stressante pour les plantes que la modification d’un seul facteur environnemental (Valladares et al., 2007). Dans le chapitre 3, nous avons déterminé la plasticité et la performance d’individus de P. antiscorbutica exposés à une augmentation simultanée de la température et de l’ombrage. Nous avons montré une réponse plastique des individus de P. antiscorbutica face à la modification simultanée des deux facteurs. Cependant, l’ombrage diminue la capacité de réponse plastique à la température. De plus, on observe une diminution de la performance des plantes exposées à deux facteurs, ce qui n’est pas le cas lorsqu’un seul facteur est modifié. L’ombrage entraîne une diminution de ressources pour la plante. Les contraintes d’allocation conséquentes à cette diminution de ressource peuvent diminuer
215
Discussion générale la capacité de réponse à d’autres facteurs environnementaux, mais également peuvent empêcher la plante de maintenir son niveau de performance.
Le changement climatique aux îles Kerguelen favorise la colonisation d’espèces végétales issues des milieux tempérés (Frenot et al., 2006). Le Chou de Kerguelen devrait alors faire face à une augmentation de la température accompagnée d’une augmentation de la compétition interspécifique notamment pour la lumière. Ce scenario entraînerait une diminution de la performance des individus de P. antiscorbutica. On note cependant que le Chou de Kerguelen est trouvé dans une large gamme environnementale aux îles Kerguelen, des pelouses littorales aux fell- fields d’altitude. Une augmentation de la compétition interspécifique est à prévoir localement à faible altitude où les communautés sont d’ores et déjà denses. En revanche, les habitats d’altitude, notamment caractérisés par des épisodes de gel-dégel fréquents ainsi que des vents violents, sont peu favorables à la colonisation par les espèces tempérées, et de fait ces espèces, sinon absentes, y restent d’importance très modérée (obs. pers.). Ainsi, les populations d’altitude de Pringlea antiscorbutica ne devraient être que peu menacées par des espèces introduites.
216
Discussion générale
Conclusion : le potentiel adaptatif dépend de la persistance des habitats
Les changements climatiques (e.g. teneur en eau, conductivité ou température) devraient avoir un impact négatif sur la performance des espèces étudiées. La modification simultanée de plusieurs facteurs environnementaux pourrait limiter la réponse plastique des plantes, et accentuer l’impact négatif sur les plantes. Face à une augmentation de la température, les habitats d’altitude pourraient apparaitre comme étant plus favorables. Cette hypothèse est cependant à nuancer par l’augmentation possible de la vitesse du vent, avec son action désséchante, qui affecterait particulièrement les plantes en altitude. Le réchauffement climatique favorise la colonisation d’espèces introduites. La compétition pour la lumière imposée par ces changements biotiques pourrait affecter la performance des espèces indigènes des îles Kerguelen. Les quatre espèces
étudiées (dans une moindre mesure chez R. pseudotrullifolius) ont la possibilité de coloniser des environnements d’altitude afin de pallier à l’augmentation de température et la compétition interspécifique. Ainsi, la persistance des habitats humides, particulièrement en altitude, pourrait
être un élément majeur de la survie des espèces indigènes des îles Kerguelen.
217
REFERENCES Agati G., Azzarello E., Pollastri S. & Tattini M. (2012) Flavonoids as antioxidants in plants: location and functional significance. Plant Science, 196, 67-76. Armbruster W.S., Pélabon C., Bolstad G.H. & Hansen T.F. (2014) Integrated phenotypes: understanding trait covariation in plants and animals. Phil. Trans. R. Soc. B, 369, 20130245. Aubert de la Rüe E. (1964) Observations sur les caractères et la répartition de la végétation des îles Kerguelen. C.N.F.R.A. Biologie, 1, 1-60. Auld J.R., Agrawal A.A. & Relyea R.A. (2009) Re-evaluating the costs and limits of adaptive phenotypic plasticity. Proceedings of the Royal Society of London B: Biological Sciences, rspb20091355. Becker M., Gruenheit N., Steel M., Voelckel C., Deusch O., Heenan P.B., McLenachan P.A., Kardailsky O., Leigh J.W. & Lockhart P.J. (2013) Hybridization may facilitate in situ survival of endemic species through periods of climate change. Nature Climate Change, 3, 1039-1043. Bergstrom D.M. & Chown S.L. (1999) Life at the front: history, ecology and change on southern ocean islands. Trends in Ecology & Evolution, 14, 472-477. Bird A. (2002) DNA methylation patterns and epigenetic memory. Genes & development, 16, 6- 21. Blows M.W. & Hoffmann A.A. (2005) A reassessment of genetic limits to evolutionary change. Ecology, 86, 1371-1384. Bolnick D.I., Amarasekare P., Araújo M.S., Bürger R., Levine J.M., Novak M., Rudolf V.H.W., Schreiber S.J., Urban M.C. & Vasseur D.A. (2011) Why intraspecific trait variation matters in community ecology. Trends in ecology & evolution, 26, 183-192. Boucher F.C., Thuiller W., Arnoldi C., Albert C.H. & Lavergne S. (2013) Unravelling the architecture of functional variability in wild populations of Polygonum viviparum L. Functional ecology, 27, 382-391. Bouchereau A., Aziz A., Larher F. & Martin-Tanguy J. (1999) Polyamines and environmental challenges: recent development. Plant Science, 140, 103-125. Bradshaw A.D. (1965) Evolutionary significance of phenotypic plasticity in plants. Advances in genetics, 13, 115-155. Chapuis J.-L., Frenot Y. & Lebouvier M. (2004) Recovery of native plant communities after eradication of rabbits from the subantarctic Kerguelen Islands, and influence of climate change. Biological Conservation, 117, 167-179. Chapuis J.L., Hennion F., Le Roux V. & Le Cuziat J. (2000) Growth and reproduction of the endemic cruciferous species Pringlea antiscorbutica in Kerguelen Islands. Polar Biology, 23, 196-204. Chastain A. (1958) La flore et la végétation des îles de Kerguelen: ṗolymorphisme des espèces australes (vol. 11). Éditions du Muséum. Chevin L.-M., Lande R. & Mace G.M. (2010) Adaptation, plasticity, and extinction in a changing environment: towards a predictive theory. PLoS Biol, 8, e1000357. Chown S.L., Gremmen N.J.M. & Gaston K.J. (1998) Ecological biogeography of Southern Ocean islands: species area relationships, human impacts, and conservation. The American Naturalist, 152, 562-575. Conner J.K. (2002) Genetic mechanisms‐ of floral trait correlations in a natural population.
218
Nature, 420, 407-410. Croteau R., Kutchan T.M. & Lewis N.G. (2000) Natural products (secondary metabolites). Biochemistry and molecular biology of plants, 24, 1250-1319. Culhane A.C., Thioulouse J., Perrière G. & Higgins D.G. (2005) MADE4: an R package for multivariate analysis of gene expression data. Bioinformatics, 21, 2789-2790. Davidson A.M., Jennions M. & Nicotra A.B. (2011) Do invasive species show higher phenotypic plasticity than native species and, if so, is it adaptive? A meta analysis. Ecology letters, 14, 419-431. Debat V. & David P. (2001) Mapping phenotypes: canalization, plasticity‐ and developmental stability. Trends in Ecology & Evolution, 16, 555-561. Di Ferdinando M., Brunetti C., Agati G. & Tattini M. (2014) Multiple functions of polyphenols in plants inhabiting unfavorable Mediterranean areas. Environmental and Experimental Botany, 103, 107-116. Diggle P.K. (2014) Modularity and intra-floral integration in metameric organisms: plants are more than the sum of their parts. Phil. Trans. R. Soc. B, 369, 20130253. Douguédroit A. & de Saintignon M.-F. (1970) Méthode d'étude de la décroissance des températures en montagne de latitude moyenne: exemple des Alpes françaises du sud. Revue de géographie alpine, 58, 453-472. Dudley S.A. (2004) The functional ecology of phenotypic plasticity in plants. Phenotypic plasticity: functional and conceptual approaches. Oxford University Press, Oxford, 151- 172. Farooq M., Wahid A., Kobayashi N., Fujita D. & Basra S.M.A. (2009) Plant drought stress: effects, mechanisms and management. In: Sustainable agriculture, pp. 153-188. Springer. Frenot Y., Chown S.L., Whinam J., Selkirk P.M., Convey P., Skotnicki M. & Bergstrom D.M. (2005) Biological invasions in the Antarctic: extent, impacts and implications. Biological reviews, 80, 45-72. Frenot Y., Gloaguen J.C., Masse L. & Lebouvier M. (2001) Human activities, ecosystem disturbance and plant invasions in subantarctic Crozet, Kerguelen and Amsterdam Islands. Biological conservation, 101, 33-50. Frenot Y., Lebouvier M., Gloaguen J.-C., Hennion F., Vernon P. & Chapuis J.-L. (2006) Impact des changements climatiques et de la fréquentation humaine sur la biodiversité des îles subantarctiques françaises. Belgeo. Revue belge de géographie, 363-372. García-Verdugo C., Granado-Yela C., Manrique E., de Casas R.R. & Balaguer L. (2009) Phenotypic plasticity and integration across the canopy of Olea europaea subsp. guanchica (Oleaceae) in populations with different wind exposures. American Journal of Botany, 96, 1454-1461. Gehrig H.H., Winter K., Cushman J., Borland A. & Taybi T. (2000) An improved RNA isolation method for succulent plant species rich in polyphenols and polysaccharides. Plant Molecular Biology Reporter, 18, 369-376. Ghalambor C.K., McKay J.K., Carroll S.P. & Reznick D.N. (2007) Adaptive versus non adaptive phenotypic plasticity and the potential for contemporary adaptation in new environments. Functional ecology, 21, 394-407. ‐ Gianoli E. (2004) Plasticity of traits and correlations in two populations of Convolvulus arvensis (Convolvulaceae) differing in environmental heterogeneity. International Journal of Plant Sciences, 165, 825-832. Gianoli E. & Palacio López K. (2009) Phenotypic integration may constrain phenotypic plasticity in plants. Oikos, 118, 1924-1928. ‐ 219
Gicquiaud L., Hennion F. & Esnault M.A. (2002) Physiological comparisons among four related Bromus species with varying ecological amplitude: polyamine and aromatic amine composition in response to salt spray and drought. Plant Biology, 4, 746-753. Grime J.P. (2006) Plant strategies, vegetation processes, and ecosystem properties. John Wiley & Sons. Groppa M.D. & Benavides M.P. (2008) Polyamines and abiotic stress: recent advances. Amino acids, 34, 35-45. Haas B.J., Papanicolaou A., Yassour M., Grabherr M., Blood P.D., Bowden J., Couger M.B., Eccles D., Li B. & Lieber M. (2013) De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nature protocols, 8, 1494-1512. Hanada K., Sawada Y., Kuromori T., Klausnitzer R., Saito K., Toyoda T., Shinozaki K., Li W.- H. & Hirai M.Y. (2011) Functional compensation of primary and secondary metabolites by duplicate genes in Arabidopsis thaliana. Molecular biology and evolution, 28, 377-382. Hartmann T. (2007) From waste products to ecochemicals: fifty years research of plant secondary metabolism. Phytochemistry, 68, 2831-2846. Hennion F. (1992) Etude des caractéristiques biologiques et génétiques de la flore endémique des îles Kerguelen. Thèse de Doctorat, Muséum National d’Histoire Naturelle, Paris, 264pp. Hennion F. & Bouchereau A. (1998) Accumulation of organic and inorganic solutes in the subantarctic cruciferous species Pringlea antiscorbutica in response to saline and cold stresses. Polar Biology, 20, 281-291. Hennion F., Bouchereau A., Gauthier C., Hermant M., Vernon P. & Prinzing A. (2012) Variation in amine composition in plant species: How it integrates macroevolutionary and environmental signals. American journal of botany, 99, 36-45. Hennion F., Fiasson J.L. & Gluchoff-Fiasson K. (1994) Morphological and phytochemical relationships between Ranunculus species from Iles Kerguelen. Biochemical systematics and ecology, 22, 533-542. Hennion F., Frenot Y. & Martin Tanguy J. (2006) High flexibility in growth and polyamine composition of the crucifer Pringlea antiscorbutica in relation to environmental conditions. Physiologia Plantarum‐ , 127, 212-224. Hennion F., Litrico I., Bartish I.V., Weigelt A., Bouchereau A. & Prinzing A. (2016) Ecologically diverse and distinct neighbourhoods trigger persistent phenotypic consequences, and amine metabolic profiling detects them. Journal of Ecology, 104, 125- 137. Hennion F. & Martin Tanguy J. (2000) Amines of the subantarctic crucifer Pringlea antiscorbutica are responsive to temperature conditions. Physiologia Plantarum, 109, 232-243. ‐ Hennion F. & Walton D.W.H. (1997a) Ecology and seed morphology of endemic species from Kerguelen phytogeographic zone. Polar Biology, 18, 229-235. Hennion F. & Walton D.W.H. (1997b) Seed germination of endemic species from Kerguelen phytogeographic zone. Polar Biology, 17, 180-187. Hermant M., Prinzing A., Vernon P., Convey P. & Hennion F. (2013) Endemic species have highly integrated phenotypes, environmental distributions and phenotype–environment relationships. Journal of Biogeography, 40, 1583-1594. Hoffmann A.A. & Daborn P.J. (2007) Towards genetic markers in animal populations as biomonitors for human induced environmental change. Ecology Letters, 10, 63-76.
‐ 220
Hornoy B., Tarayre M., Hervé M., Gigord L. & Atlan A. (2011) Invasive plants and enemy release: evolution of trait means and trait correlations in Ulex europaeus. PLoS One, 6, e26275. Hummel I., El Amrani A., Gouesbet G., Hennion F. & Couée I. (2004a) Involvement of polyamines in the interacting effects of low temperature and mineral supply on Pringlea antiscorbutica (Kerguelen cabbage) seedlings. Journal of experimental botany, 55, 1125- 1134. Hummel I., Quemmerais F., Gouesbet G., El Amrani A., Frenot Y., Hennion F. & Couée I. (2004b) Characterization of environmental stress responses during early development of Pringlea antiscorbutica in the field at Kerguelen. New phytologist, 162, 705-715. Johannsen W. (1909) Elemente der Erblichkeitslehre. Fischer. Jones C.S., Martínez-Cabrera H.I., Nicotra A.B., Mocko K., Marais E.M. & Schlichting C.D. (2013) Phylogenetic influences on leaf trait integration in Pelargonium (Geraniaceae): Convergence, divergence, and historical adaptation to a rapidly changing climate. American journal of botany, 100, 1306-1321. Kanaga M.K., Ryel R.J., Mock K.E. & Pfrender M.E. (2008) Quantitative-genetic variation in morphological and physiological traits within a quaking aspen (Populus tremuloides) population. Canadian Journal of Forest Research, 38, 1690-1694. Le Roux P.C. & McGeoch M.A. (2007) Fine-scale variation in the spatial association of plant species: A test of the stress-gradient hypothesis in the sub-Antarctic. South African Journal of Botany, 73, 297-298. Le Roux P.C. & McGeoch M.A. (2008) Changes in climate extremes, variability and signature on sub-Antarctic Marion Island. Climatic Change, 86, 309-329. Lebouvier M. & Frenot Y. (2007) Conservation and management in the French sub-Antarctic islands and surrounding seas. Lebouvier M., Laparie M., Hulle M., Marais A., Cozic Y., Lalouette L., Vernon P., Candresse T., Frenot Y. & Renault D. (2011) The significance of the sub-Antarctic Kerguelen Islands for the assessment of the vulnerability of native communities to climate change, alien insect invasions and plant viruses. Biological Invasions, 13, 1195-1208. Lehmann S., Funck D., Szabados L. & Rentsch D. (2010) Proline metabolism and transport in plant development. Amino acids, 39, 949-962. Lehnebach, C, Winkworth, R.C., Becker, M., Lockhart, P.J. & Hennion, F. Around the pole: evolution of sub-Antarctic island Ranunculus. Journal of Biogeography early view 15th February 2017, DOI: 10.1111/jbi.12952. Lê S., Josse J. & Husson F. (2008) FactoMineR: an R package for multivariate analysis. Journal of statistical software, 25, 1-18. Mallitt K.L., Bonser S.P. & Hunt J. (2010) The plasticity of phenotypic integration in response to light and water availability in the pepper grass, Lepidium bonariense. Evolutionary ecology, 24, 1321-1337. Murren C.J. (2002) Phenotypic integration in plants. Plant Species Biology, 17, 89-99. Murren C.J. (2012) The integrated phenotype. Integrative and comparative biology, 52, 64-76. Nicolaysen K., Frey F.A., Hodges K.V., Weis D. & Giret A. (2000) 40 Ar/39 Ar geochronology of flood basalts from the Kerguelen Archipelago, southern Indian Ocean: implications for Cenozoic eruption rates of the Kerguelen plume. Earth and Planetary Science Letters, 174, 313-328. Nicotra A.B., Atkin O.K., Bonser S.P., Davidson A.M., Finnegan E.J., Mathesius U., Poot P., Purugganan M.D., Richards C.L. & Valladares F. (2010) Plant phenotypic plasticity in a 221
changing climate. Trends in plant science, 15, 684-692. Pansu J., Winkworth R.C., Hennion F., Gielly L., Taberlet P. & Choler P. (2015) Long-lasting modification of soil fungal diversity associated with the introduction of rabbits to a remote sub-Antarctic archipelago. Biology letters, 11, 20150408. Philippe M. & Giret A. (1998) Bois fossiles tertiaires et quaternaires de Kerguelen (océan Indien austral). Comptes Rendus de l'Académie des Sciences-Series IIA-Earth and Planetary Science, 326, 901-906. Pichersky E. & Gang D.R. (2000) Genetics and biochemistry of secondary metabolites in plants: an evolutionary perspective. Trends in plant science, 5, 439-445. Piersma T. & Drent J. (2003) Phenotypic flexibility and the evolution of organismal design. Trends in Ecology & Evolution, 18, 228-233. Pigliucci M. (2001) Phenotypic plasticity: beyond nature and nurture. JHU Press. Pigliucci M. (2003) Phenotypic integration: studying the ecology and evolution of complex phenotypes. Ecology Letters, 6, 265-272. Pigliucci M. (2005) Evolution of phenotypic plasticity: where are we going now? Trends in Ecology & Evolution, 20, 481-486. Pigliucci M. & Kaplan J. (2010) Making sense of evolution: The conceptual foundations of evolutionary biology. University of Chicago Press. Powell J.R., Parrent J.L., Hart M.M., Klironomos J.N., Rillig M.C. & Maherali H. (2009) Phylogenetic trait conservatism and the evolution of functional trade-offs in arbuscular mycorrhizal fungi. Proceedings of the Royal Society of London B: Biological Sciences, 276, 4237-4245. Prinzing A. (2001) The niche of higher plants: evidence for phylogenetic conservatism. Proceedings of the Royal Society of London B: Biological Sciences, 268, 2383-2389. Rapson G.L. & Maze J. (1994) Variation and integration in the rare grass Achnatherum (Oryzopsis) hendersonii: phenotypic comparison with parapatric common congeners. Canadian journal of botany, 72, 693-700. Richards C.L., Bossdorf O., Muth N.Z., Gurevitch J. & Pigliucci M. (2006) Jack of all trades, master of some? On the role of phenotypic plasticity in plant invasions. Ecology letters, 9, 981-993. Samal A., Wagner A. & Martin O.C. (2011) Environmental versatility promotes modularity in genome-scale metabolic networks. BMC systems biology, 5, 1. Schermann-Legionnet A., Hennion F., Vernon P. & Atlan A. (2007) Breeding system of the subantarctic plant species Pringlea antiscorbutica R. Br. and search for potential insect pollinators in the Kerguelen Islands. Polar Biology, 30, 1183-1193. Schlichting C.D. (1986) The evolution of phenotypic plasticity in plants. Annual review of ecology and systematics, 667-693. Schlichting C.D. (1989) Phenotypic integration and environmental changes. BioScience, 39, 460- 464. Sinclair B.J., Addo Bediako A. & Chown S.L. (2003) Climatic variability and the evolution of insect freeze tolerance. Biological Reviews, 78, 181-195. Smith R.I.L. (1984)‐ Terrestrial plant biology of the sub-Antarctic and Antarctic. Sulmon C., Van Baaren J., Cabello-Hurtado F., Gouesbet G., Hennion F., Mony C., Renault D., Bormans M., El Amrani A. & Wiegand C. (2015) Abiotic stressors and stress responses: What commonalities appear between species across biological organization levels? Environmental Pollution, 202, 66-77. Sultan S.E. (2000) Phenotypic plasticity for plant development, function and life history. Trends 222
in plant science, 5, 537-542. Sultan S.E. & Spencer H.G. (2002) Metapopulation structure favors plasticity over local adaptation. The American Naturalist, 160, 271-283. Teuscher E. & Lindequist U. (2010) Biogene Gifte: Biologie-Chemie-Pharmakologie- Toxikologie. Wissenschaftl. Verlag-Ges. Tiburcio A.F., Altabella T., Bitrián M. & Alcázar R. (2014) The roles of polyamines during the lifespan of plants: from development to stress. Planta, 240, 1-18. Tohge T., Watanabe M., Hoefgen R. & Fernie A.R. (2013) The evolution of phenylpropanoid metabolism in the green lineage. Critical reviews in biochemistry and molecular biology, 48, 123-152. Valladares F., Arrieta S., Aranda I., Lorenzo D., Sánchez-Gómez D., Tena D., Suárez F. & Pardos J.A. (2005) Shade tolerance, photoinhibition sensitivity and phenotypic plasticity of Ilex aquifolium in continental Mediterranean sites. Tree Physiology, 25, 1041-1052. Valladares F., Gianoli E. & Gómez J.M. (2007) Ecological limits to plant phenotypic plasticity. New Phytologist, 176, 749-763. Valladares F. & Niinemets Ü. (2008) Shade tolerance, a key plant feature of complex nature and consequences. Annual Review of Ecology, Evolution, and Systematics, 237-257. Van der Putten N., Verbruggen C., Ochyra R., Verleyen E. & Frenot Y. (2010) Subantarctic flowering plants: pre glacial survivors or post glacial immigrants? Journal of biogeography, 37, 582-592. Van Kleunen M. & Fischer M.‐ (2005) Constraints on ‐the evolution of adaptive phenotypic plasticity in plants. New Phytologist, 166, 49-60. Van Valen L. (1973) A new evolutionary law. Evolutionary theory, 1, 1-30. Vinocur B. & Altman A. (2005) Recent advances in engineering plant tolerance to abiotic stress: achievements and limitations. Current opinion in biotechnology, 16, 123-132. Voelckel C., Heenan P.B., Janssen B., Reichelt M., Ford K., Hofmann R. & Lockhart P.J. (2008) Transcriptional and biochemical signatures of divergence in natural populations of two species of New Zealand alpine Pachycladon. Molecular ecology, 17, 4740-4753. Wagstaff S.J. & Hennion F. (2007) Evolution and biogeography of Lyallia and Hectorella (Portulacaceae), geographically isolated sisters from the Southern Hemisphere. Antarctic Science, 19, 417. West-Eberhard M.J. (2003) Developmental plasticity and evolution. Oxford University Press. Wink M. (2003) Evolution of secondary metabolites from an ecological and molecular phylogenetic perspective. Phytochemistry, 64, 3-19. Wink M. (2013) Evolution of secondary metabolites in legumes (Fabaceae). South African Journal of Botany, 89, 164-175. Wittstock U. & Halkier B.A. (2002) Glucosinolate research in the Arabidopsis era. Trends in plant science, 7, 263-270. Zhu J.-K. (2001) Plant salt tolerance. Trends in plant science, 6, 66-71.
223
Abstract
The adaptive potential of a species can be defined as its capacity to cope with environmental change. Adaptive potential increases with phenotypic variation, from the intra-individual to the inter-population level, but factors controlling and explaining this variation still remain poorly understood.We studied four plant species from Iles Kerguelen in the sub-Antarctic region which is currently facing one of the strongest climate changes worldwide. Plant species from Iles Kerguelen are known to show high phenotypic integration (i.e. strong correlation among traits), a phenomenon that has been suggested to constrain trait variation. For these species we studied what constrains phenotypic variation, considering the external environment, the internal phenotypic integration and the associated performance costs. We found that intra-individual variation, i.e. plasticity, may be constrained by complex environmental change and the performance costs it triggers. In contrast, plasticity may be favored by high degree of phenotypic integration (Chapter 3). We found that inter-individual variation within populations may not be constrained by environmental factors, but may be favored by high phenotypic integration (Chapter 1). We found inter-population variation within regions may be constrained by restricted environmental variation (Chapter 1). Finally, we studied secondary metabolites (amines and flavonols) that connect environmental variation to phenotypic variation. We found that compositions and functions of these metabolites vary among regions, probably reflecting evolutionary differentiation among regions (Chapter 2). Patterns of variation betweenregions suggest that within species metabolites may be functionally redundant or versatile, for which to our knowledge our results are the first hint. Overall, we suggest that climate change in Kerguelen will impact plant species performance, and that the persistence of suitable wet habitats will be determinant in species capacities to cope with such changes. Furthermore, this project identified so far underestimated factors which may favor the adaptive potential of species. Particularly, we emphasize that the adaptive potential of species may increase due to (i) phenotypic integration, (contrary to common suggestion) and (ii) metabolite redundancy or versatility (only poorly studied so far). Moreover, we evidenced, partly for the first time, multiple costs and limits of plasticity and suggest that plasticity does not guarantee plant success in the new environment.
Key words : Trait variation; Phenotypic integration; Plasticity; Plant species; Performance; Secondary metabolites; Redundancy; Versatility; Climate change; Iles Kerguelen
Résumé
Le potentiel adaptatif d’une espèce peut être défini par sa capacité à faire face aux changements environnementaux. Le potentiel adaptatif augmente avec la variation du phénotype, du niveau intra-individuel au niveau inter-populations. Cependant, les facteurs qui contrôlent et expliquent cette variation phénotypique sont encore relativement peu compris. Nous avons étudié quatre espèces végétales des îles Kerguelen, en subantarctique, une des régions les plus sévèrement affectée par le changement climatique. Les espèces végétales des îles Kerguelen montrent un fort degré d’intégration phénotypique (i.e. forte corrélation entre les traits), suggéré comme étant une contrainte pour la variation des traits. Chez ces espèces, nous avons étudié les facteurs qui peuvent contraindre la variation phénotypique : les facteurs environnementaux extrinsèques, l’intégration phénotypique intrinsèque et les coûts associés de performance. Nous avons montré que la variation intra-individuelle, i.e. plasticité, peut être contrainte par la modification simultanée de multiple facteurs environnementaux, et par les coûts de performance qui en résultent. En revanche, la plasticité peut être favorisée par un fort degré d’intégration phénotypique (Chapitre 3). Nous avons montré que la variation interindividuelle, à l’intérieur des populations, n’était pas contrainte par des facteurs environnementaux, mais pouvait cependant être favorisée par un fort degré d’intégration phénotypique (Chapitre 1). Egalement, nous avons trouvé que la variation inter- populations à l’intérieur d’une région peut être contrainte par une variation environnementale limitée (Chapitre 1). Finalement, nous avons étudié des métabolites secondaires (amines et flavonols) qui font le lien entre variation environnementale et variation phénotypique. Nous avons trouvé que la composition et la fonction de ces métabolites varient entre régions, suggérant une différentiation évolutive entre régions (Chapitre 2). Les patrons de variation entre régions, suggèrent au niveau intra-spécifique une redondance et une versatilité fonctionnelle des métabolites, que nous somme, à notre connaissance, les premiers à mettre en évidence. Nous suggérons que le changement climatique des îles Kerguelen va avoir un impact négatif sur la performance des espèces végétales. La persistance d’habitats humides favorables à ces espèces sera alors un facteur déterminant de la capacité des espèces à faire face au changement climatique. De plus, ce projet a permis d’identifier des facteurs jusqu’alors peu reconnus qui pourtant favorisent le potentiel adaptatif des espèces. En particulier, le potentiel adaptatif peut être favorisé par (i) le degré d’intégration phénotypique (contrairement à ce qui est communément suggéré) et (ii) la redondance et la versatilité des métabolites (qui a peu été étudiée jusqu’alors). Qui plus est, nous avons mis en évidence pour la première fois, plusieurs coûts et limites de la plasticité, suggérant qu’une réponse plasticité de la plante n’est pas une garantie de succès dans le nouvel environnement.
Mot clés : Variation phénotypique ; Intégration phénotypique ; Plasticité ; Espèces végétales ; Performance ; Métabolites secondaires ; Redondance ; Versatilité ; Changement climatique ; Iles Kerguelen