Joint effects of global warming and pollutants on damselflies

However, we poorly know the roles of thermal adaptation, life history and the rapid evolution at an expanding range front in shaping the vulnerability to contaminants under global w^armi

Khuong DINH VAN

Dissertation presented in partial fulfilment of the requirements for the degree of Doctor in Science (Biology)

October 2014

Prof Dr Robby Stoks

University of Leuven, Belgium

Prof Dr Lieven Bervoets

University of Antwerp, Belgium

Dissertation presented in partial fulfilment of the requirements for the degree of Doctor in Science (Biology)

October 2014

Examination committee

Prof Dr Em Frans Ollevier

University of Leuven (KU Leuven), Belgium

Prof Dr Luc De Meester

University of Leuven (KU Leuven), Belgium

Dr Lizanne Janssens

University of Leuven (KU Leuven), Belgium

Prof Dr Tom Wenseleers

University of Leuven (KU Leuven), Belgium

Dr Hélène Arambourou

IRSTEA - Laboratoire d'écotoxicologie, France

Prof Dr Matthias Liess

UFZ - Helmholtz Centre for Environmental Research, Germany

© 2014 KU Leuven, Groep Wetenschap & Technologie

Uitgegeven in eigen beheer, Khuong Dinh Van, Leuven

Alle rechten voorbehouden Niets uit deze uitgave mag worden vermenigvuldigd en/of openbaar gemaakt worden door middel van druk, fotokopie, microfilm, elektronisch o f op welke andere wijze ook zonder voorafgaandelijke

schriftelijke toestemming van de uitgever.

All rights reserved No part o f the publication may be reproduced in any form by print, photoprint, microfilm, electronic or any other means without vvritten

permission from the publisher.

Cover photo: Ischnura elegans larva (Photo credit: Christophe Brochard).

ISBN: 978-90-8649-762-1

D/2014/10705/70

The research presented in this dissertation was funded by the Vietnam

International Educational Development (VIED) and the IRO-expat supplement.

Time had flown when I realized that I have been in the lab for four years Completion of a doctoral programme in a prestigious university like University o f Leuven is just a dream come true that 1 did not dare to think about during my childhood To achieve this, 1 have been received great supports and assistance from many people “around the world” and I highly appreciate all these valuable supports

I would like to express my deep gratitude to my promotor Professor Robby Stoks who provided me a best opportunity to start a scientific career as an eco(toxico)logist 1 am always inspired by his ideas to improve the quality of research both in conducting experiments and writing manuscripts With the respect I would like to thank you very much for your support

A special thank is also sent to Lizanne who always provides me a great technical and scientific assistance, even when she was extremely busy with her experiments and manuscripts I also thank Lizanne for a translation of the summary o f my thesis into Flemish Lizanne also helped me time after time to solve the housing problems in Leuven I highly appreciate all your assistance

1 would also like to express my gratitude to my co-promotor Professor Lieven Bervocts for the important contributions to my research 1 am very happy with all the great results from your assistance

i would also like to thank ail the assessors and jury members Professor Luc De Meester, Professor Matthias Liess, Professor Frans Ollevier, Dr Hélène Arambourou, and Professor Tom Wenseleers for their constructive comments for my thesis

[ would like to thank R.ia for a great assistance in laboratory works Ria is always available to answer my questions Ria also helped me and my family a lot in personal life 1 also thank Conny, Eddy, Melissa, Geert and Ronny for all the administrative and technical

assistance I would like to thank F a n s for solving a lot o f computer issues, particularly when I had a ‘‘crashed hardware" or installed new windows and softwares.

Special thanks are sent to E3 group for always willing to help and helped me a lot during high workload o f the experiments I would like to thank Eleven and Hajnl for great

temperatures in Redingenhof I would like to thank Sara and Hốlồne for excellent discussions

providing assisầnce during my study and preparations for my public defence All o f you are not only my best colleagues ever but also my great friends in Leuven

eggs o f Jschnura elegans damselfly and their contribution to the edition and revision o f the

chapter 1 A thank is also sent to Christophe Brochard for the pem ission to use his photos in

my thesis I would like to thank Maarten De Jonge and Steven Joosen for technical assistance when 1 worked with radioactive materials and analysing metal concenfrations at the University o f Antwerp

1 would like to thank the Internationalization team for “happy hours" together

I would also thank Dr Pham Quoc Hung, Dr Nguyen Tan Sy, Professor Nguyen Dinh Mao and leaders o f the Nha Trang University who have always provided me strong supports and always encourage me during my study abroad 1 would appreciate all their supports

I would like to thank Professor Hoang Tung at International University, Vietnam who always provides me supports and advices since I worked at Nha Trang University I also thank him for recommending me to KU Leuven for my PhD study This was one o f the most important recommendations for me so far

I would like to thank colleagues at the section o f Ecology, Evolution and C o n so latio n and friends in Leuven who have provided various kinds o f assistance or make m y life here

more memorable, particularly Anna, Frans and Aister for lunches together at Alma; Quan,

1 would like administrative officers at the Vietnam International Educational Development, particularly Ms Bui Hong Thuy, Ms Nguyen Thi Ha Anh, and Ms Ngo Thi Dieu Linh for handling my reports and scholarship 1 would like to thank Mr Edmundo Guzman at the International Office, KU Leuven for assisting with IRO expat

A special thank is also sent to mv parents Đinh Công ưấn and Hoàng Thị Bon and parents-in-law Vũ Văn Tiến and Đoàn Thị Hiền, brothers and sisters who always encourage

me during my time abroad My total gratitude goes to my wife VŨ Thị Thùy Minh and daughter Đinh Vũ Minh Ngọc for their endless love and support while they have to go throu,gh all the challenges in life without my assistance for a long time

D edication

I dedicate this thesis to my wife, daughter and parents

Global warming and contamination are two major threats to freshwater biodiversity, which, moreover, have the potential to interact synergistically To persist in a warming world,

organisms have to either adapt in situ to the temperature increase or shift their range to track

their moving climate niche However, we poorly know the roles of thermal adaptation, life history and the rapid evolution at an expanding range front in shaping the vulnerability to contaminants under global w^arming In this study, i investigated the separate and joint effects

of global warming and contaminant stress on damselflies focusing on two aspects: (i) a 4.C temperature increase and how this is further modulated by thermal and latitudinal adaptation and (ii) a poleward range expansion and associated life history evolution

In chapters I-IV, 1 describe the results of common garden warming experiments where

larvae o f replicated populations from low, central and high latitudes of Ischnura elegans were

reared at 20.C and at 24.C corresponding to the mean summer water temperatures in ponds in

difference also corresponds to the predicted temperature increase by 2100 under iPCC scenario A IFI This space-for-timc substitution approach allows testing whether local thermal adaptation may shape the future susceptibility of contaminants under global warming at higher latitudes I found a strong latitude-specific vulnerability to contaminants for both the metal zinc and the pesticide chlorpyrifos Indeed, high-latitude populations showed a stronger reduction in escape swimming speed when exposed to zinc, indicating that the metal contamination may make them more vulnerable to predation Importantly, at high temperature high-latitude populations were also more vulnerable to zinc in term of survival and activity levels while these effects were not present or much less pronounced in low-latitude

vii

populations, indicating that gradual thermal adaptation may mitigate the future effects of metals on high-latitude populations under global warming In contrast, the low-latitude populations showed a stronger growth reduction at high chlorpyrifos concentrations and high temperature This result suggests that the faster life history, more than three generations per year, o f low-latitude populations makes them more vulnerable to organophosphate pesticides like chlorpyrifos At higher temperature, the negative effects o f chlorpyrifos were also stronger for predator behaviors while these effects were similar for antipredator behaviours at both temperatures These results highlight the complexity o f contaminant effects on predator- prey interactions being differentially temperature-dependent pending on the trophic level.

In chapters V and VI, I tested for the role o f the evolution o f a faster life history and a higher investment in flight morphology during range expansion in driving the vulnerability to two widespread pesticides, the organophosphate chlorpyrifos and the pyrethroid esfenvalerate,

in the edge populations of the damselfly Coenagrion scitulum, a currently poleward moving

damselfly 1 did this by comparing the vulnerability o f replicated core and edge populations reared under common garden conditions from the egg stage Besides testing for effects o f the pesticides during the larval stage I was particularly interested in carry-over effects in the adult stage in traits related to range expansion, namely flight ability and immune function I found evidence for the faster life history o f edge populations in term o f growth rate and development time while the higher investment in flight-related traits was only present at high density in the absence o f pesticides Nevertheless, the edge populations showed a consistently higher vulnerability to pesticides both during the exposure period and with regard to carry-over effects across metamorphosis; the mortality rate was higher in esfenvalerate-exposed edge larvae and the chlorpyrifos-induced reductions in flight muscle mass and phenoloxidase activity were only present in the edge animals

viii

The here presented resuits provide a comprehensive picture on the vulnerability of damselflies both at the larval and adult stage to different types of contaminants under global wanning I, for the first time, documented that thermal adaptation may mitigate the effects of metals such as zinc at higher latitude by applying a space-for-time substitution approach I also showed that the faster life history o f the multivoltine animals at low latitudes may make them more vulnerable to pesticides This research is also the first one that studied and documented that rapid range expansion may increase the vulnerability o f edge populations to pesticides These results emphasize the need for spatially explicit risk assessment and associated conservation tools The documented higher mortality and the carry-over effects from larval exposure on flight- and immune-related traits indicate that contamination may slow down the future range expansion in poleward moving species.

investering in vlieggerelateerde kenmerken wanneer ze opgekweekt werden bij hoge densiteit

en in afwezigheid van het pesticide Dieren aan het front waren gevoeliger voor het pesticide, zowel tijdens de blootstelling als na metamorfose in het adulte stadium: de sterflesnelheid was hoger in de aan esfenvaieraat blootgestelde dieren van het front en de door chioorpyrifos veroorzaakte daling in vliegspiermassa en fenoloxidase-activiteit was enkei aanwezig in dieren van het front

De resultaten voorgesteld in deze thesis bieden inzicht in de kwetsbaarheid van waterjuffers, in het larvale en adulte stadium voor polluenten met een verschillend werkingsmechanisme bij klimaatsopwarming Ik kon met behulp van een “space-for-time substitution approach” voor de eerste keer aantonen dat thermale adaptatie een invloed kan hebben op de gevoeligheid voor metalen zoals zink op hogere latitudes Mijn resultaten gaven ook aan dat de snellere levensgeschiedenisstrategie van multivoltiene dieren op lage latitudes hen kwetsbaarder maakte voor pesticiden In deze thesis werd ook voor de eerste keer

aangetoond dat een snelle areaaluitbreiding de gevoeligheid van populates aan het front voor

pesticiden verhoogt Deze resultaten illustreren het belang van ruimtelijk georienteerde risico- evaluatie en conservatiestrategiegn De hogere mortaliteit en de overdracht van effecten van larvale blootstelling op vlieg en immuungerelateerde kenmerken geven aan dat verontreiniging de toekomstige areaaluitbreiding van een noordwaarts opschuivende soo.rt kan vertragen

xii

Table of contents

S u m m a ry vii

S a m e n v a ttin g x

G eneral in tro d u c tio n 1

PART 1: V u ln e ra b ility to conta m in a nts under a 4.C increase in te m p e ra tu re under global w a rm in g 19

Chapter I Susceptibility to a metal under global warming is shaped by thermal adaptation along a latitudinal gradient 21

Chapter II Local adaptation and the potential effects of a contaminant on predator avoidance and antipredator responses under global warming: a space-for-time substitution approach 42

Chapter III Temperature- and latitude-specific individual grov،rth rates shape the vulnerability of damselfiy larvae to a widespread pesticide 64

Chapter IV Warming increases chlorpyrifos effects on predator but not anti-predator behaviours 100 PART 2: V u ln e ra b ility to co nta m in a nts under global w a rm ing at an expanding f r o n t 121

Chapter V Does rapid evolution of faster larval growth rates at an expanding range front increases vulnerability to a widespread pesticide? 123

Chapter VI Increased vulnerability to an insecticide at the expansion front in a poleward moving damselfiy Coenagrion scitulum 155

G eneral d iscu ssio n 193

R e fe re n ce s 213

P u b licatio n s 249

xiii

2012; M o et al., 2013; Campos et al., 2014), crustaceans (e.g Delorenzo et al.١ 2009;

^n illm a nn et al., 2013), fishes (e.g Dietrich et al 2014; Papoulias et al.١ 2014; Sovova et al.١

2014), amphibians (e.g Rohr and Palmer, 2013: Johnson et al.١ 2013) Moreover, there is increasing evidence that global w a rn in g atid contaminants may interact in a synergistic way (whereby the combined impact is larger than the sum o f the individual impacts, sensu Folt et al., 1999; Darling and COte, 2008) (Brook et al., 2008; Sokolova and Lannig, 2008; Noyes et al., 2009; Moe et al., 2013)

In this chapter 1 w ill introduce the key topics o f my thesis: global warming, contaminants and especially their potential interplay Additionally, 1 w ill provide info m ation

on the used approaches, study species and measاlred traits when quantifying the effects o f global warming and contaminants, l.dnally, 1 cvill provide an (overview o f aims and outlines o f

my PhD project

؛ Global warming

Global warming is mainly characterized by the increase in surface temperature (IPCC, 2007)

W hile surface temperatures increased on average 0.6"C during the 20؛' century (Walther et ah.١

2002; Meehl and Tebaldi, 2004), the rate o f warming is expected to be considerably higher

AIFI; IPCC, 2007) There is a general consensus that anthropogenic forces such as the emission of greenhouse gasses are responsible for the acceleration of warming on earth during the 20؛؛ and 21^؛ century (IPCC, 2007; Ghommem et al., 2012) A great concern is that global ١

warming may increase the risk of extinction in 15-37% of species by 2050 (Thomas et al., 2004) To survive under global warming, organisms must either adapt in situ to the temperature rise by physiological adjustments (Chown and Gaston, 2008), or shift their range poleward and thereby track their temperature envelope (Parmesan et al., 1999; W alther et al., 2002; Chen et al., 2011)

General introduction

K CD U.

T - - r ٠ C ٠، ٠-C ١J r

-c، ٥ ٥ < < < <

Figure I Scenarios for global warming during the century projected by IPCC (2007)

Solid lines are multi-model global averages o f surface warming (relative to 1980-1999) for

Shading denotes the ± 1 standard deviation range of individual model annual averages The grey bars at right indicate the best estimate o f temperature rises by 2100 for each scenario (Source: adapted from IPCC, 2007) In current thesis I focused on scenario A IFI which predicts a 4.C increase in the mean temperature under global warming by 2100

2

General introduction

2 Global warming and contaminants

2.1 //7 ،؟zYw population persistence

In a warming and contaminated world, the ability o f populations to survive in situ will depend

not only on the ability to deal with the temperature increase itself, for example through physiological adjustments (Chown and Gaston, 2008), but also on the ability to deal with changed interactions with contaminants The emerging picture is one of increased vulnerability (toxicity) to contaminants at higher temperatures (Sokolova and Lannig, 2008; Noyes et ah, 2009; but see for pyrethroids Harwood et al., 2009), with vulnerability being the effect of a chemical at a controlled exposure duration (Rohr et al., 2011) However, the potential role of gradual local thermal adaptation in shaping the susceptibility to contaminants under global warming at high-latitude sites remains to be assessed Moreover, at higher temperatures and in the presence o f contaminants also interactions with predators change, potentially adding another level for complexity for local population persistence

While we have some knowledge about how species are differentially susceptible to contaminants across large-scale environmental gradients (Chapman et al., 2006; Clements et al., 2012), this is largely lacking for populations within species (but see Cherkasov et al.١

2010) Consequently, current risk assessment necessarily has to extrapolate from studies at the local scale to assess susceptibility across a species' range This is, however, potentially misleading as organisms often evolve different life histories along environmental gradients (Corbet et al., 2006) and therefore can be expected to also differ in susceptibility This is particularly relevant in the context o f joint effects of global warming and contaminants in which thermal adaptation may differentially shape susceptibility to contaminants along natural temperature gradients (Cherkasov et al., 2010; Мое et al., 2013)

General introduction

Many species increase the number o f generations per year at lower latitudes (e.g., damselflies Corbet et al., 2006) This typically results in less time to grow and develop per generation and the evolution of higher individual growth rates, i.e higher increases in individual body mass during the larval stage at lower latitudes (e.g., Ragland and Kingsolver, 2007; Nygren et al., 2008; Stoks et al., 2012) While the importance o f population growth rates and the associated recovery rates across generations in shaping the vulnerability o f populations to contaminants is well established (e.g., Barnthouse, 2004; Liess and von der Ohe, 2005; Kamo et al., 2011), the potential effect of individual growth rates in shaping the vulnerability o f individuals to contaminants during their lifespan has received less attention This is, however, directly relevant for the standard toxicity tests that typically evaluate the vulnerability o f a species to a contaminant based on individual effects (individual survival and individual growth rates) within a single generation (Walker et al., 2006) Life history theory and, more specifically, energy allocation trade-offs predict that animals with higher individual growth rates will allocate more energy to growth and less to other functions, including defence (Congdon et al., 2001; Sibly and Calow, 1989) This generates the untested hypothesis that animals from low-latitude populations with a faster life history should, therefore, show a lower tolerance to contaminants compared to animals from high-latitude populations

To study how organisms may gradually evolve under global warming I use an indirect space-for-time substitution (sensu Fukami and Wardle, 2005) where populations o f the same species are compared along a natural temperature gradient (e.g Sarmento et al., 2010) Assuming that other environmental factors do not differ systematically across latitudes, a careful choice o f regions can provide an insight into how closely ecological responses to current temperatures in low-latitude regions will match predicted temperatures in 2100 in the high-latitude regions (IPCC, 2007; De Block et al., 2013) The performance o f conspecific

indi\’iduals from the cooler region along the gradient can be tested at the current and predicted temperature under global warming and at the same time can be compared with the performance o f organisms at the warmer region of the gradient that have thermally adapted to the higher local temperature.

General introduction

2.2 Range expansion

Besides adapting locally to the temperature increase, populations may, however, also react to global w arning by moving polewards and thereby track their temperature envelope (Pamiesan et al., 1999؛ Chen et a!., 2011) This alternative scenario also asks for an evolutionary perspective as multivariate trait evolution occurs in the edge populations at expanding range fronts (Burton et al., 2010) At expanding range edges, key life history traits

in the non-dispersive laiwal stage such as individual larval growth rates are predicted to increase (Phillips, 2009), because genotypes that are able to produce a larger number of generations in a given amount of time will be favoured by selection at the initial low conspeciBc densities at the expansion front (Phillips, 2009؛ Phillips et al٠١ 2010) The few

studies that looked at this, indeed documented the evolution of faster individual growth rates

at the edge populations in tadpoles of the cane load Bujo marirtus (Phillips, 2009) and larvae

o f the damselfly Coenagrion sciiulum (Therry et al., 20I4c) Additionally, theory predicts the

evolut.ion of increased dispersal ability at the range front (Shine et al., 2011) In line with this several studies documented an increased investment jn dispersal-related traits at the range front ( Hill et al., 2011) The higher investment in a rapid life history and dispersal-related traits is expected to come at a lower investment in other traits This generates the untested prediction that the rapid evolutionary changes in edge populations are traded off against defenc:e and repair against contaminants

Another particu!ar!y relevant topic in the context o f range expansion is whether changed vulnerability to contaminants in the larval stage, that is driven by evolutionary change in larval growth rate, can bridge metamorphosis and thereby negatively affect adult traits such as flight-related traits and the ability to deal with parasites Many species that are expanding their range have a complex life cycle with a larval stage where growth and development occur separated by metamorphosis from an adult stage where reproductioil and dispersal occur (Stoks and Cordoba-Aguilar, 2012) There is increasing evidence that negative effects o f contaminants may be carried over from the larval stage to the adult stage (Rohr and Palmer, 2005؛ Distel and Boone, 2009؛ Janssens and Stoks, 2013b؛ Janssens et al., 2014b) Camy-over effects o f contaminants on flight ability and on immune fiinction are directly relevant to study in the context o f range expansion as these fraits increase rapidly at the expansion front, likely in association with selection to accelerate the rate o f range expansion (Hill et al., 2011 ؛ Theiry et al., 2014c) Any negative effects on these traits may thereby slow down the rate o f fiirther range expansion.

Genera! Introduction

3 Study species

Damselflies have a complex life cycle with an aquatic larval stage where growth occurs and a terrestrial adult stage where dispersal and reproduction occur (Stoks and Cordoba-Aguilar,2012) Damselflies occur abundantly in small water bodies adjacent to agricultural areas where they can be especially impacted by pesticide run-off During the larval stage they cannot escape exposure to high temperatures and contaminants making damselflies particularly vulnerable to global warming (Hassail et al., 2007؛ Hassall and Thompson, 2008) and several contaminants (Beketov, 2004؛ Rubach et al., 2011) They are, therefore, useful organisms to study both direct effects and the interactions between contaminants and climatic

General introduction

stressors during the larval exposure period, and carry-over effects across metamorphosis to the adult stage Damselfly larvae are also important intermediate predators (Johnson, 1991; Jonsson et al., 2007), therefore any effects of contaminants and climatic stressors on damselflies likely will cascade through food webs (Stoks et al., in review), making them relevant organisms to study predator and antipredator behaviours in the context of the effects

Damselflies are showing strong range shifts toward higher latitudes under global warming (Hickling et al., 2006; Hassall and Thompson, 2010) The poleward range expansion induces rapid evolution o f traits such as increased growth rates at the range front (Therry et a l, 2014b), which are expected to increase the vulnerability to contaminants and thereby potentially slow down any further range expansion

In this research I chose to study two damselfly species: Ischnura elegans and

Coenagrion scituliim Each species was chosen to study the vulnerability to contaminants

under a particular aspect o f global warming: (i) the roles o f thermal adaptation and life history along a latitudinal gradient in shaping the vulnerability to contaminants in the widely

distributed Ischnura elegans, (ii) the cost o f a rapid evolution and a higher investment in

flight-related traits at an expanding front in shaping vulnerability to pesticides in the currently

poleward moving Coenagrion scitulum.

l٦he damselfly I elegans was chosen to study the role of thermal adaptation and life

history on the vulnerability to contaminants and predators because this species has a wide distribution in both low-latitude (southern France), central-latitude (Belgium) and high- latitude (southern Sweden) areas in Europe (Gosden et al.١ 201 1 ١ see also Fig 2) and previous

studies have shown patterns o f thermal adaptation for growth rate (Shama et al., 2011; Stoks

et al., 2012) and for predator-prey interactions (De Block et al., 2013) Along the latitudinal gradient, this species has a gradual change in voltinism, the number of generations per year,

7

with three to four generations per year in southern France, two generations per year in

changes are associated with a faster life history (growth and development) at low latitudes to compensate for the shorter larval growth period per generation (Shama et al.١ 2011; Stoks et al., 2012) At each latitude two randomly chosen populations from lentic shallow water bodies were sampled (see chapters I-IV for detailed information about the study populations)

General introduction

Figure 2 A larva o f Ischnura elegans damselfly and the map of sampling populations at low

(LI: +45٥43١N, +5٥22١E and L2: -l·45٠42١N, +5٥25١E), middle (M l: + 5 0 5 TN, 4٥41١E and M2: +50٥57١N, +5.27.E), and high (H I: +56٥39١N, +16٠20١E and H2: 5 6 4 0 ’N, 16٥17١E) latitudes (Photo credit for the damselfly larva: Khuong Dinh Van; the map of sampling populations is adapted from http://www.enchantedlearninu.cQm/freoiiraphv)

General introduction

Figure 3 An adult male o f damselfly Coena^rion scitidum (a) and the map of sampling

populations (b) Cl (+ 46.43'N , t |٥20T،) and C2 { ^ 50٥26١N, 4 1٠34١E) represent two core populations El ( ٣ 49٥14١N, 7٥I6١E) and E2 (-^ 46٥43١N, + l٥20١E) represent two edgepopulations (see chapter V and VI for detailed information) (Photo credit: Christophe Brochard for the adult male: Lieven Therry for the map)

The damselfly Coenagrion scitulum (Rambur, 1842) is a Mediterranean daniseltly

preferring small ponds (Dijkstra, 2006) u p to the I990s the northern range limit was sitاfated

in northern France, after which a north-eastward range expansion has occurred (W asscher and Goudsmits, 2010؛ Swaegers et al., 2013) In 2010, the northern-most limit o f the expanding range margin was situated in the southern parts ofT he Netherlands, and the northeastern limit

in Western G em any (Wasscher and Goudsmits, 2010) Edge populations have a higher laical growth rate, shorter development time (duration from hatching to metamorphosis) and higher flight-related traits (flight muscle mass and fat content) (Therry et al., 2014b; Therry et al.,

2014c) Therefore, c scitlumum was chosen to study the role o f a rapid evolution in life

history and a higher investment in flight-related in shaping the vulnerability to pesticides I

VI for more detail) Two core populations are occu^ing within the distribution of the species (Dijkstra, 2006) The edge populations were found recently when sampled (ca 4-5 years, Therry et al., 2014c) All four populations were separated >150 km from each other and there are no geographical obstacles (e.g mountains) to prevent further range movements

Most study populations are in natural areas without agriculture and therefore were unlikely to be exposed to pesticides (Coors et al., 2009) or metals (zinc concentrations in the sampling ponds were below detection limits, see chapter 1 and II) One of the study

in Saarland, Germany Any local adaptation to pesticides would, however, be unlikely in

Coenagrion damselflies given their high levels o f gene flow (Johansson et al., 2013).

Genera! intr٥duct!on

10

4 Traits under investigation

To get a multifaceted picture of the effects of contaminants under global warming, I studied effects on life history, behaviour and physiology

4 1 Life-history and behavioural traits

I quantified following key life-history traits in damseltlies: larval survival, larval growth rate, larval development rate and mass at emergence (Stoks and Cordoba-Aguilar, 2012) As aquatic ecosystems are strongly shaped by predator-prey interactions (Wellborn et al., 1996) 1 also studied effects on predator and antipredator traits 1 focused on two general activities (mobility and foraging), predator avoidance (activity reductions under predation risk) and a key antipredator trait (escape swimming) of damselfly larvae (Stoks et al., 2003; Gyssels and Stoks 2005; Stoks and McPeek, 2006)

General introduction

4.2 Physiological defence and detoxification variables

To start unraveling physiological mechanisms underlying the susceptibility to warming and contaminants, 1 quantified (i) bioaccumulation of interna! metal concentrations, (ii) key detoxification mechanisms for metals (metallothionem concentrations, Amiard et al., 2006) and for the pesticides (glutathione S-transferase activity, Domingues et al., 2010) and (iii)

Metallothioneins are non-enzymatic proteins with a low molecular weight, high cysteine content, no aromatic amino acids and heat stability (Amiard et al., 2006) Given a high cysteine content, containing thiol groups (-SH), metallothioneins can bind both essential (Cu and Zn) and non-essential metals (e.g., Cd, Hg) (Amiard et al., 2006) Metallothioneins are upregulated in metal-exposed animals and are thought to be involved in the detoxification

11

of non-essentia! meta!s and excessive amounts of essential metals (Amiard et al., 2006؛ De Jonge et al.١ 2011).

Glutathione S-transferase (GST) is a diverse family of enzymes involved in the detoxification of a wide range o f xenobiotics including insecticides (Enayati et al., 2005؛ Domingues et al., 2010) GST can metabolize insecticides by facilitating their reductive dehydrochlorination or by conjugation reactions with reduced glutathione, to produce water- soluble metabolites that are more easily excreted (Enayati et al., 2005؛ Domingues et al.,2010) The level of glutathione S-transferase activity is typically higher in pesticide-exposed animals (Domingues et al., 2010), including damselfiy larvae (Janssens and Stohs, 2013a)

Acetylcholinesterase is the target enzyme that is inhibited by organophosphate insecticides such as chlorp^ifos (see above and also Christensen et al., 2009) Pesticide- inhibited AChE activity has been documented for many vertebrates (Fulton and Key, 2001) and invertebrates (Fulton and Key, 2001؛ Domingues et al., 2010), including for damselfiy

la ^ a e (Janssens and Stoks, 2013a) AChE activity levels c o rd a te with survival (Fulton and Key, 2001) and individual g r o ^ rate (Janssens and Stoks, 2013a)

Genera! introduction

4.3 Flight- and immune-related traits

I quantified pesticide effects on three flight-related traits known to affect flight ability in

Coenagrion damselfiies: body mass, relative flight muscle mass and fat content (Gyulavdri et

al., 2014؛ Therry et al., 2014c)

1 also studied two important immune-related traits in insects; the activity of phenoloxidase and the encapsulation response (Rantala et al., 2011؛ Gonzalez-Santoyo and Cordoba-Aguilar, 2012) Phenoloxidase (PO) plays a crucial role in the melanization process

in which PO converts phenols to quinones, which subsequently polymerize to form melanin

12

(Gonzalez-Santoyo and Cordoba-Aguilar, 2012) Melanin is deposited onto foreign targets and responsible for defence against pathogens, and tissue repair (Gonzalez-Santoyo and Cordoba-Aguilar, 2012) Another, non-enzymatic way o f quantifying immune response in insects is measuring the encapsulation response by inserting a monofilament, which mimics

an ectoparasitic water mite attack in damselflies (Rantala et al.١ 2000; Rolff and Siva-Jothy,

2004; Rantala et al., 2010) Damselflies defend themselves against ectoparasitic water mites

by encapsulating the feeding tubes o f the externally located mites (Rantala et ai., 2010) in a similar way as they encapsulate the implant

General introduction

5 Aims and outlines o f the research

The focus of this thesis is to study separate and joint effects o f global warming and contaminant stress on damselflies focusing on life history responses and traits related to the interactions with natural predators and prey (behaviour) and with parasites (PO and encapsulation), and traits related to dispersal (flight-related traits) I quantified both stressor effects during the larval exposure period as well as the delayed effects bridging metamorphosis

Part I: Vulnerability to contaminants under a 4°C increase in temperature under global warming

In chapters I to IV I describe the results o f common garden warming experiments where

larvae o f replicated populations of low, central and high latitudes of / elegans were reared at

20.C and at 24.C Temperatures o f 20.C and 24.C reflect the mean summer water temperatures in ponds in southern Sweden and Southern France, respectively (De Block et al

13

2013) Importantly, the 4٥C temperature difference also corresponds to the predicted

temperature increase by 2100 under IPCC scenario A lF i (IPCC, 2007) This setup allows testing the idea that local thermal adaptation may predictably shape the susceptibility of animals along a latitudinal gradient and to explore the future effects of contaminants under global warming at higher latitudes

In chapter I, I studied the thermal dependence o f the susceptibility o f Ischnura elegans

damselfly larvae to zinc with regard to survival and gro\vth rate I also quantified food intake, internal zinc concentrations, and metallothioneins (Amiard et al., 2006) to start unraveling mechanisms underlying the latitude-specific susceptibility to warming and contaminants

In chapter II, I assessed the vulnerability o f Ischnura elegans damselfly larvae to zinc

with regard to mobility, foraging, predator avoidance and antipredator responses under global warming I focused on two general activities (mobility and foraging), predator avoidance (activity reductions under predation risk) and a key antipredator trait (escape swimming) o f damselfly larvae (Stoks et al., 2003; Gyssels and Stoks 2005; Stoks and McPeek, 2006) Zinc has been known to alter the predator avoidance and antipredator responses of aquatic species (Atchison et al., 1987; Mogren and Trumble, 2010)

In chapter III, I tested for the role o f thermal adaptation and life history evolution of

Ischnura elegans damselfly larvae along a latitudinal thermal gradient in shaping the

vulnerability to the pesticide chlorpyrifos with regard to survival and growth rate I also quantified food intake and AChE and GST activity (Domingues et al., 2010) to start unraveling mechanisms underlying the temperature- and latitude-specific individual growth rates in shaping vulnerability to chlorpyrifos

In chapter IV, I tested whether any temperature modulation o f the effect o f the

pesticide chlorpyrifos will occur to the same extent for interactions o f Ischnura elegans

General introduction

14

damselfly !arvae with higher (antipredator traits) and lower (predator traits) trophic levels I quantified general predator behaviours (mobility, foraging) and two key antipredator responses ofdam selfiy larvae (activity reductions under predation risk and escape swimming performance) (Stoks et al., 2003: Stoks and McPeek, 2006).

Gene٢a! Int٢٥duc^؛on

Part 2: Vulnerability to contaminants under global warming at, an expandingfront

In chapters ٧ and VI, I tested for the role of the evolution of a faster life history and higher

investment in flight morphology during range expansion in driving the vulnerability to two

widespread pesticides in the edge populations of Coenagrion scitulum damselfly, a currently

poleward moving damselfly (Swaegers et al.١ 2013) I did this by comparing the vulnerability

o f replicated core and edge populations reared under common garden conditions from the egg stage Besides testing for effects of the pesticide during tlie larval stage I was particularly interested in carry-over effects in the adult stage in traits related to range expansion, namely flight ability (chapters V and VI) and immune function (chapter V)

In chapter V, 1 tested for effects of the pesticide chlorpyrifos in an indoor experiment where la ^ a e were reared individually Given that trade-ofifs may only be visible at suboptimal conditions I included two food treatments Besides testing tOr efl'ects of the pesticide during the larval stage, I was particularly interested in carry-over effects to the adult stage in traits related to range expansion, namely flight ability and immune function I, therefore, focused on three sets of ecologically relevant traits: life histoiy (larval growth and survival, mass at emergence, metamorphosis success), flight-related traits (mass at emergence, flight muscle mass and fat content, Therry et al., 2014c) and immune-related traits (phenoloxidase activity

2012) Specifically, I tested whether the evolution of faster growth rates in edge populations is associated with an increased vulnerability to contaminants

15

In chapter VI I tested for a higher vulnerability to esfenvalerate, a pyrethroid, at low and high densities in an outdoor container experiment I included low and high densities because densities are initially lower at the expansion front and pesticide effects may be mediated by density Larvae were reared in group at low and high densities from late summer

to fall and winter before exposing to esfenvalerate in the spring Besides testing for effects on larval survival, growth and development, I also tested for potential carry-over effects bridging metamorphosis on a set o f flight-related traits (body mass, relative flight muscle mass and fat content), that may be especially relevant for dispersal ability

General introduction

6 Contaminants studied

Trace metals and pesticides occur extensively in aquatic systems worldwide (Schwarzenbach

et al., 2006; Eurostat, 2007) Main sources o f metal contamination in aquatic ecosystems are wastewater from mining and industries, and urban runoff (Mance, 1987) Very high metal concentrations are still being observed in water bodies (Sola et al., 2004; Nieto et al., 2007; Ruiz et al., 2014) and negative effects o f metals on aquatic organisms are being documented (Sokolova and Lannig, 2008; Cherkasov et al., 2010; Vellinger et al., 2012; Kimberly and Salice, 2014) Another important type o f contaminants in aquatic ecosystems are pesticides Pesticide use in agriculture in Europe is high (Eurostat, 2007) For example, the annual application of pesticides to food crops in the European Union exceeds 140,000 tons (Eurostat,2007) Pesticide use is predicted to increase in the future due to climate change (Kattwinkel et al., 2011) Pesticide contamination in aquatic habitats often occurs through runoff from adjacent agricultural areas (Jergentz et al., 2004)

I here studied the effects o f one trace metal, zinc, and two pesticides with a different mode o f action: the organophosphate insecticide chlorpyrifos and the pyrethroid insecticide

16

esfenvalerate Zinc is one o f the most common contaminants in aquatic systems (Mance, 1987; B rix et al., 2011) Zinc is an essential metal, but an excessive amount o f zinc can provoke toxicity to aquatic animals (B rix et al., 2011; Brinkman and Johnston, 2012) by inhibiting calcium ion regulation, leading to hypocalcemia (Hogstrand and Wood, 1996) A t very high concentrations, zinc can damage gills leading to hypoxia (Hogstrand and Wood, 1996) Uptake o f zinc occurs mainly through the body surface and digestive tracts in aquatic insects (Hare, 1992).

Chlorpyrifos (CPF) (Fig 4a) is an organophosphate insecticide that is among the most frequently used pesticides worldwide (Eaton et al., 2008) CPF is highly toxic to aquatic organisms (reviewed in Barron and Woodburn, 1995) CPF is known to prevent the functioning o f the e n z ^ e acetylcholinesterase (AChE), in breaking down the neurotransmitter acetylcholine (Christensen et al., 2009) The resulting accumulation o f acetylcholine in the synaptic cleft causes overstimulation o f the neuronal cells, which leads to neurotoxicity and eventually death (Christensen et al., 2009)

Genera! introduction

(b)

ا ٦

ا ء

)1

Figure 4 Chemical structure o f (a) chlorpyrifos, an organophosphate pesticide, and (b)

esfenvalerate, a pyrethroid (source: http://www.pesticideinfo.org)

Esfenvalerate (Fig 4b) is one o f the most w idely applied pyrethroid pesticides during the past Pwo decades (Beketov and Liess, 2005; Spurlock and Lee, 2008) Pyrethroids typically damage the nervous system (Cold and Forbes, 2004) through interference with ion

17

channels in the nerve axons (Solomon et al., 2001) Esfenvalerate has shown to be highly toxic to aquatic invertebrates (reviewed in Rasmussen et al., 2013b), including damselfly larvae (Beketov, 2004).

CPF and esfenvalerate is also highly hydrophobic (e.g., log Kow: CPF = 4.7 Smith, 198land esfenvalerate > 6, Kelley, 2004) and rapidly partition to suspended particles and sediments (e.g., Кос: CPF > 4960 g/ml United States Environmental Protection Agency,

2011, and esfenvalerate = 215,000 Kelley, 2004) The concentrations o f these pesticides rapidly decrease in water For example, the half-life o f CPF was ca 3.8 ~ 9.1h in a man-made pond (Mazanti et al., 2003) and esfenvalerate concentrations were reduced more than 90% after 24h (Stampfli et al., 2011)

Genera! introduction

18

Chapter I

Susceptibility to a metal under global warming is shaped

by thermal adaptation along a latitudinal gradient

Chapter I

Khuong Dinh Van* Lizanne Janssens* Sara Debecker, Maarten De Jonge, Philippe Lambret Viktor Nilsson-Ortman Lieven Bervoetsand Robby Stoks

Published in Global Change Biology (2013) 19: 2625-2633

*Joint first authors

21

Abstract

Global warming and contamination represent two major threats to biodiversity that have the potential to interact synergistically There is the potential for gradual local thermal adaptation and dispersal to higher latitudes to mitigate the susceptibility o f organisms to contaminants and global warming at high latitudes Here, 1 applied a space-for-time substitution approach to

study the thermal dependence o f the susceptibility o f Ischnura elegans damselfly larvae to

zinc in a common garden warming experiment (20.C and 24.C ) with replicated populations from three latitudes spanning >1500 km in Europe We observed a striking latitude-specific effect of temperature on the zinc-induced mortality pattern; local thermal adaptation along the latitudinal gradient made Swedish, but not French, damselfly larvae more susceptible to zinc

at 24.C Latitude- and temperature-specific differences in zinc susceptibility may be related to the amount o f energy available to defend against and repair damage since Swedish larvae showed a much stronger zinc-induced reduction o f food intake at 24.C The pattern o f local

strongly magnify the impact o f a contaminant such as zinc at higher latitudes unless there is thermal evolution and/or migration o f lower-latitude genotypes Our results underscore the critical importance o f studying the susceptibility to contaminants under realistic warming scenarios taking into account local thermal adaptation across natural temperature gradients

22

Chapter I

Introduction

Global warming and contamination have been identified as major threats to biodiversity

2008) which, moreover, have the potential to interact synergistically (Brook et al., 2008; Sokolova and Lannig, 2008; Noyes et al., 2009; Мое et al., 2013) The ability o f populations

to survive locally under global wanning will therefore depend not only on the ability to deal with the temperature increase itself, for example through physiological adjustments (Chown and Gaston, 2008), but also on the ability to deal with changed interactions with contaminants The emerging picture is one o f increased susceptibility (toxicity) at higher temperatures (Sokolova and Lannig, 2008; Noyes et al., 2009), with susceptibility being the effect o f a chemical at a controlled exposure duration (Rohr et al., 2011) However, the potential for gradual local thermal adaptation and dispersal to higher latitudes that may mitigate the susceptibility to contaminants under global warming at high-latitude sites remains to be assessed (Dunne et al., 2004; Fukami and Wardle, 2005)

One approach to determine how organisms may gradually evolve under global warming is to use an indirect space-for-time substitution (sensu Fukami and Wardle, 2005) where populations of the same species are compared along a natural temperature gradient (e.g Sarmento et al., 2010) The perfomiance of conspecific individuals from the cooler region along the gradient can be tested at the current and predicted temperature under global warming and at the same time can be compared with the performance o f organisms at the wanner region o f the gradient that have thermally adapted to the higher local temperature This approach has been used only rarely (but see De Block et al., 2013) and to our knowledge never in the context o f the combined impact of global warming and another key anthropogenic stressor such as a contaminant While we have some knowledge about how species are differentially susceptible to contaminants across large-scale environmental

23

gradients (Chapman et al 2006; Clements et al., 2012), this is largely lacking for populations within species (but see Cherkasov et al., 2010) Consequently, current risk assessment necessarily has to extrapolate from studies at the local scale to assess susceptibility across a species’ range This is, however, potentially misleading as organisms often evolve different life histories along environmental gradients (Blanckenhom and Demont, 2004) and therefore can be expected to also differ in susceptibility (Kamo et al., 2011) This is particularly relevant in the context of joint effects o f global warming and contaminants in which besides life history evolution also thermal adaptation may further differentially shape susceptibility to contaminants along natural temperature gradients (Cherkasov et al., 2010; Мое et al., 2013).

Here we study the thermal dependence o f the susceptibility o f an aquatic insect to a metal in a common garden warming experiment with replicated populations from three latitudes spanning >1500 km in Europe to test the idea that local thermal adaptation may predictably shape the susceptibility o f animals along a latitudinal gradient and to explore the future effects o f metals under global warming at higher latitudes Freshwater ecosystems are especially vulnerable to global warming (Woodward et al., 2010) and to contaminants (Bronmark and Hansson, 2002) because most aquatic animals cannot escape these stressors

We documented how warming and contaminants affect two fitness-related traits in damselflies (Stoks and Cordoba-Aguilar, 2012): survival ^ d growth rate To start unraveling mechanisms underlying the latitude-specific susceptibility to warming and metals (a recently highlighted priority in ecotoxicology, Clements et al., 2012) we quantified food intake, bioaccumulation (internal metal concentrations) and a key detoxification mechanism (metallothionein concentrations, Amiard et al., 2006) As model species we chose damseifly larvae, important intermediate predators in aquatic food webs (Johnson, 1991; Jonsson et al.,2007), that are particularly vulnerable to global warming (Hassall et al., 2007; Hassall and Thompson, 2008), and are showing strong range shifts toward higher latitudes under global

Chapter I

24

warming (Hickiing et al., 2006; Hassail and Thompson, 20ا0د We chose zinc as a

contaminant since it is one of the most common contaminants in aquatic systems (Brix et al.,

2011) Zinc is an essential metal, but an excessive amount ol'zinc can provoke toxicity to

aquatic animals (Brix et al., 2011) Uptake of zinc 0CCU٢S mainl)' through the body surface

and digestive tracts in aquatic insects (11 are, 1992)

Materials and methods

Study populations and rearing experiment

We studied populations o f the damselfly Jschnura elegans from three latitudes representing

the low-latitude (southern France), central (Belgium) and high-latitude (southern Sweden)

parts o f the range in Europe (Gosden et al., 2011) At each latitude two randomly chosen

populations were sampled, namely Salette (5٥2 2١23.92ب45ه43’30.58١٦١ل ,+١١E) and Arandon

(+45٥42١35.64N, +5٠25’47.28”E) for France; De Maten ( + 50٠57’3g.89”N, +5٥27١28.99"E)

These were all shallow waterbodies At each site water and sediment samples were collected

and analyzed All zinc concentrations o f water and sednnent sampies were below detection

limits (water: <2.5 pg/1, sediment: <146 pgZg dw)

individually in small jars with wet filter paper for oviposition Eggs were transferred to the

laboratory in Belgium Throughout the experiment, eggs and larvae o f all populations from

summer water temperatures in ponds in southern Sweden and Southern France, respectively

Chapte٢ ا

25

(De Block et al., 2013) Importantly, the 4.C temperature difference also cortesponds with the predicted temperature increase by 2100 under IPCC scenario AIFI (IPCC, 2007) Therefore, this setup is suitable for a space-for-time substitution to explore the effects o f global w arn in g

at higher latitudes Ten days after hatching, larvae were allocated individually to plastic vials (7.5 cm height, 3.5 cm diameter) filled to a h.eight o f 6 cm with aged tap water Vials were placed in temperature-controlled water baths, and regularly re-distributed among the three water baths per temperature All six water baths were placed in the same room to ensure equal

conditions (for example, light regimes) Larvae were fed Artemia nauplii ad libitum (459 ±

48, mean ± SE, n = 10) five days per week L a ^ a e were checked three times a week for molts into the final instar When molted into the final instar, larvae from each rearing temperature were introduced in the zinc exposure experiment at their respective temperature Note that by doing so, all la ^ a e had been frلاly acclimated to their experimental temperature (starting from the egg stage) before we tested effects o f temperature, zinc and their combination

Chapter ا

Experimental design

To test whether the effects o f zinc exposure depend on temperature and latitude of origin we set up a nested full factorial design with 2 populations per latitude x 3 latitudes x 2 temperatures (20.C and 24.C) x 3 zinc concentrations (0, 50, 150 mg zinc/1) The zinc

concentrations were based on a previous study on another coenagrionid damselfly (Argia sp.)

are extremely high but have been observed in natural water bodies in contaminated areas in Europe (e.g., Nieto et al., 2007) We prepared the zinc solution daily based on a stock solution

stock solution with synthetic water (for details see Janssens and Stoks, 2013c) that was also 26

used as control Growth rate and activity o f damselfly larvae in this synthetic water do not differ from those observed in natural pond water (Janssens and Stoks, 2013c).

individmlly to 100 ml glass jars Jar walls were covered by brown tape to make sure that the larvae could not see each other (damselfly larvae are cannibalistic and impose stress on each other; McPeek et al., 2001; De Block and Stoks, 2004a) Each jar was filled with 50 ml of one

o f the three zinc solutions (0, 50, 150 mg zinc/I) The medium in each jar was renewed daily

to minirr.ize the increase o f zinc concentrations due to water evaporation from the jars The measured zinc concentrations in the experimental vials with nominal concentrations of 50 and

150 mg zinc/1 (based on a pooled sample o f 10 vials) were 45.91 and 142.22 mg zinc/1 when the medium was freshly renewed; after 24h zinc concentrations were 47.82 and 160.9 mg

were fed daily the same amount o f Artemia as during the pre-exposure period Because some

mortality occurred before the exposure period, the numbers o f larvae exposed per combination o f latitude x temperature x zinc concentration varied between 30 and 44, except

sizes per response variable are shown in the figures

Response variables

We daily checked for mortality when refreshing the medium To obtain an estimate of growth rate each larva was weighted before (initial wet mass) and after (final wet mass) the 6-day exposure period to the nearest 0.01 mg using an electronic balance (Mettler Toledo® AB135- S) individual growth rates were calculated as (Infinaimass - Ininitiai mass)/6 days (see Stoks et al., 2012)

Chapter 1

27

On day six of the exposure period we quantified the food intake For logistic reasons it was not possible to measure this trait in all larvae Therefore, we restricted these measurements to all larvae from the low (France) and high (Sweden) latitudes To quantify food intake each larva was acclimated for 7 minutes in a plastic container (15 cm x 10 cm x12.5 cm) filled with 600 ml o f medium corresponding to the temperature and zinc treatment

o f a given larva Artemia nauplii (3560 ± 298, mean ± SE, n = 5) were then added to the

container and the number o f nauplii consumed during 7 minutes was monitored by direct

observation Food intake was expressed as the number o f Artemia nauplii eaten per mg larval

wet mass per min

At the end o f the exposure period all larvae were frozen at -80.C for later analyses

We quantified concentrations o f the detoxification protein metallothionein (MT) based on the protocol of Klein et al (1994) that was optimized for damselfly larvae (see Data SI) The MT concentrations were expressed as nmol/g larval wet mass In the same larvae in which the MT concentrations were measured, we also quantified internal zinc concentrations based on the protocol by Eraly et al (2010) (see Data SI) The internal zinc concentrations were expressed

as pg/g larval wet mass Ten larvae (five per population) from each o f the 18 treatment combinations (3 latitudes x 2 temperatures x 3 zinc concentrations) were used for M T and internal zinc measurements (total of 180 larvae)