The evolutionary significance of parthenogenesis and sexual reproduction in

the Australian spiny leaf insect,

Extatosoma tiaratum

Yasaman Alavi

Submitted in total fulfilment of the requirements of the degree of

Doctor of Philosophy

October 2016

School of BioSciences

Faculty of Science

The University of Melbourne

i

ii Abstract

The costs and benefits of sexual reproduction has long been a subject of debate in biology. The paradox arises from the fact that theoretically, sex is associated with many costs, yet it is the most prevalent mode of reproduction in the tree of life.

Facultative parthenogenetic systems, in which females can reproduce both sexually, and in the absence of sperm, parthenogenetically, provide suitable systems to compare costs and benefits of reproductive modes, while minimizing confounding effect that are not directly related to reproductive modes. In this thesis, I used the Australian

Phasmatid, Extatosoma tiaratum, to investigate the evolutionary significance of facultative parthenogenesis, and compare fitness consequences of sex and parthenogenesis.

The evolutionary significance of facultative parthenogenesis is unknown but male or sperm limitations are potential factors. I investigate male mating frequency and variation in ejaculate size and quality in E. tiaratum. I show that most, but not all, males are able to mate multiply, but ejaculate size decreases with increased number of matings. In addition, ejaculate size increased with increasing time interval between matings, suggesting that E. tiaratum males require time to replenish ejaculate reserves.

These findings suggest male sperm limitation may be an important factor influencing the evolution of parthenogenesis in this system.

iii The cytological mechanism of parthenogenesis determines the genetic diversity and heterozygosity levels of the offspring and is thus an important component of the comparison between reproductive modes. Using microsatellite markers that I developed for E. tiaratum, I investigated the cytological mechanism of parthenogenesis in this species. I demonstrated that the most likely mechanism of parthenogenesis in E. tiaratum is automixis with terminal fusion, resulting in substantial loss of heterozygosity in the first generation parthenogenetic offspring.

Based on these results I predicted parthenogenesis to be associated with fitness reduction in E. tiaratum.

I then investigated the fitness consequences of sex and parthenogenesis in E. tiaratum, in terms of offspring immunity and reproductive success. I show that parthenogenesis is associated with fitness reduction both in terms of immune function and reproductive success. Females derived from sexually conceived mothers have higher immune response compared to females derived from parthenogenetically conceived mothers. A female’s reproductive success is substantially higher if she is conceived sexually and produces sexual offspring. In fact, the cost of parthenogenesis on reproductive success is high enough to eliminate the two-fold cost of sex in E. tiaratum. Nevertheless, the cost of parthenogenesis can be compensated in the next generation, if parthenogenetically females produce daughters sexually.

The results of this thesis offer a new perspective in understanding the evolution and maintenance of facultative parthenogenesis in E. tiaratum, and suggest that females

iv enjoy the benefit of parthenogenesis allowing them to reproduce in the absence of sufficient or preferred sperm, while mediating the costs of homozygosity by reproducing sexually.

v Declaration

This is to certify that:

The thesis comprises only my original work towards the PhD except where indicated in the Preface.

Due acknowledgement has been made in the text to all other material used.

The thesis is fewer than 100 000 words in length, exclusive of tables, maps bibliographies and appendices.

…………………………………………

Yasaman Alavi

vi Preface

Yasaman Alavi (YA) was the primary author for all manuscripts (presented as chapters) within this thesis. These manuscripts were written with assistance and editorial changes by supervisors Mark Elgar (MAE), Therésa Jones (TMJ), and

Andrew Weeks (ARW). The ideas concerning the design of the experimental aspects of this thesis were originally conceived by YA but refined and developed in collaboration with MAE and TMJ. YA carried out the laboratory experiments and field samplings and collected all the data used within this thesis. YA performed analyses of all data with advice and suggestions provided by MAE, TMJ and ARW.

Chapter 2-5 of this thesis are written and presented as separate papers. Chapter 2 has been published as Alavi, Y., Elgar, A. M., and Jones, M. T. (2016). Male Mating

Success and the Effect of Mating History on Ejaculate Traits in a Facultatively

Parthenogenic Insect (Extatosoma tiaratum). Ethology, 122: 1-8. Chapter 3 has been accepted for publication as Alavi, Y., van Rooyen, A., Elgar, M. A., Jones, M. T., and

Weeks, A. R. (2016). Novel microsatellite markers suggest the mechanism of parthenogenesis in Extatosoma tiaratum is automixis with terminal fusion. Insect

Science, in press. The manuscripts as presented within this thesis are largely unchanged from the published material.

vii Acknowledgments

I would like to thank my amazing supervisors for their help and support during the production of this thesis. First of all, I am grateful to Mark Elgar, for giving me this wonderful opportunity to come to Melbourne and work on this project, and for his ongoing support and patient guidance. I have learned a lot from him, both in science and also on the philosophy of life. I am truly thankful to Therésa Jones for her constant and brilliant ideas, and for always giving me great advice and motivation. I have learned a lot from her, especially on experimental design and statistical analysis. I am also grateful to Andrew weeks for his helpful advice on chapters 1 and 3 of this thesis.

I would like to thank all members of the Elgar and Jones lab for their help and support over the past years. Especially, I am thankful to Gareth Hopkins for his comments on

Chapter 2 and 4 of my thesis, and to Dany Zemeitat, Qike Wang, Bernie Wittwer and

Eunice Tan for their help in maintaining the insect population in my absence. Also thanks to Joanna Durrant for showing me how to run immune assays.

I am thankful to Patrick Honan, Chloe Miller and Maik Fiedel from the Museum

Victoria, for their help in the field collections. Many thanks to Anthony Van Rooyen for helping with all the molecular lab work during the development and use of the microsatellite markers. Thanks to Mandy Parfitt, Uoda Khammy, Stephen Frankenberg and Rohan Long for borrowing me equipment for lab work.

viii Many thanks to the Holsworth Wildlife Research Endowment for funding this project, and the Melbourne School of Graduate Research and the School of BioSciences for providing funding for conferences.

Last but not least, I would like to thank my parents Salimeh and Kambiz, and my brother Behrang, for their everlasting love and support, without which this would not have been possible.

ix Table of contents

Abstract iii Declaration vi Preface vii Acknowledgments viii Table of contents x List of Figures xiv List of Tables xviii

Chapter 1 Introduction: the evolutionary significance of sex and parthenogenesis 1

Modes of reproduction 2 Parthenogenesis: modes, mechanisms and emergence 3 The paradox of sex: theories and empirical evidence 7 What is missing? 12 Facultatively parthenogenetic species as models 13 The study system, an Australian Phasmatid 14 The evolutionary significance of facultative parthenogenesis in E. tiaratum 17 References 27

Chapter 2 Variation in male mating success in a facultatively parthenogenic insect (Extatosoma tiaratum) 35

Summary 36 Introduction 37 Methods 40 Experimental animals 40 Experimental design 40 Preparation of the sperm solution 41

x Sperm density assay 42 Sperm viability assay 42 Statistical analyses 43 Results 45 Male mating history and survival 45 Effects of multiple mating on ejaculate size and quality 46 Discussion 50 References 54

Chapter 3 Novel microsatellite markers suggest the mechanism of parthenogenesis in Extatosoma tiaratum, is automixis with terminal fusion 58

Summary 59 Introduction 60 Methods 63 Laboratory stock population 63 Field collection 63 Next-generation sequencing and de novo genome assembly 64 Microsatellite isolation and characterization 64 Patterns of inheritance 66 The mechanism of parthenogenesis 67 Results 68 Next-generation sequencing and de novo genome assembly 68 Microsatellite isolation and characterization 68 Patterns of allele inheritance 69 Mechanism of parthenogenesis 70 Discussion 77 References 81

Chapter 4 Sex versus parthenogenesis; immune function in a facultatively parthenogenetic Phasmatid (Extatosoma tiaratum) 85

xi

Summary 86 Introduction 87 Methods 91 Animal maintenance and culturing 91 Determining reproductive history 92 Haemolymph collection and processing 92 Haemocyte concentration 93 Lytic activity assay 94 Phenoloxidase (PO) assay 95 Statistical analysis 95 Results 97 Determining reproductive history 97 Immune function assays 97 Discussion 106 References 110

Chapter 5 Sex versus parthenogenesis; female reproductive success in a facultatively parthenogenetic Phasmatid (Extatosoma tiaratum) 114

Summary 115 Introduction 116 Methods 121 Animal maintenance 121

Deriving F1 from F0 females 121

F1 fecundity 125

F2 Juvenile survival 125 Offspring survival of field- collected females 126 Microsatellite heterozygosity and modes of conception 127 Statistical analysis 128 Results 129 Microsatellite heterozygosity proportions and modes of conception 129

xii F1 body size, adult lifespan and fecundity 133

F2 Juvenile survival 136 Offspring survival of field-collected females 138 Discussion 140 References 145

xiii List of Figures

Chapter 1

Figure 1-1

Cytological mechanisms involved in different modes of automictic

parthenogenesis and their consequences in terms of inbreeding, for loci located

close to the centromere with low recombination rates (left) and far away from the

centromere with high recombination rates (right) (Modified from Pearcy et al

2006). Larger circles represent nuclei; lines represent chromatids with crosses

marking the location of the centromere. Black and white circles show alleles at a

given locus. Maternal genotype is heterozygous and parthenogenesis results in

homozygous offspring. 5

Figure 1-2

Adult Extatosoma tiaratum (a) female; (b) male; and (c) copulating pair. In the

copulating pair, the male is above the female and the visible spermatophore is

being transferred to the female 16

Figure 1-3

Extatosoma tiaratum sperm 20× under light (a) and fluorescent microscopy (blue

excitation filter at λ = 490 nm, b and c); b) shows dead sperm cells stained by

propidium iodide and c) shows viable sperm cells stained by SYBR14

(LIVE/DEAD assay, Molecular Probes; Sigma, Australia, L-7011) 20

xiv Figure 1-4

A schematic illustration of consequences of sex (continuous lines) and

parthenogenesis (dashed lines) in terms of reproductive success across

generations: a) an obligately sexual lineage with constant reproductive success, b)

an obligate sexual lineage that fails to mate successfully becomes extinct, c) an

obligately parthenogenetic lineage has reduced reproductive success at each

generation, and d) a facultative parthenogenetic lineage alternating between sex

and parthenogenesis will recover its reproductive success with sex 26

Chapter 2

Figure 2-1

The relationship between spermatophore number and ejaculate traits; a)

spermatophore mass, b) sperm density (Ln transformed), and c) viable sperm

proportion; Levels not connected by same letters are significantly different based

on post hoc Tukey’s tests. 47

Figure 2-2

The relationship between a) spermatophore mass and body mass (r2 =

0.19, P= 0.0002), b) spermatophore mass and mating interval (r2 = 0.45, P

<0.0001) c) sperm density (Ln transformed) and mating interval (r2 = 0.24, P

<0.0001) 48

xv Chapter 4

Figure 4-1

The interaction between F0 and F1 MOCs on logarithm of a) haemocyte

concentration and b) lytic activity in F1 female E. tiaratum; levels not connected

by same letters are significantly different based on post hoc Tukey’s test. 103

Figure 4-2

The effect of week of sampling on average logarithm of (a) haemocyte

concentration and (b) lytic activity in the F1 E. tiaratum offspring (across

treatments) 104

Figure 4-3

The interaction between week of sampling, and (a) F0 and (b) F1 modes of

conception, on average PO activity of F1 females; continuous lines represent

sexual and dashed lines represent parthenogenetic modes of conception. Levels

not connected by same letters are significantly different based on post hoc

Tukey’s test. 105

Chapter 5

Figure 5-1

The experimental set up showing mating treatments of F0 and F1 females and data

collected for F1 and F2 generations 123

xvi Figure 5-2

Mating status of F0 (top) and F1 (bottom) females during each time period; a proportion of F1 females produced by F0 females in the second time period were used as F1 females (N = 23) 124

Figure 5-3

The effect of mode of conception on life history trait measures: a) body size

(tarsus length), b) adult lifespan, c) weekly egg number, of F1 females, and d) development time of F2 second instar females to adulthood 135

Figure 5-4

The weekly oviposition rate of F1 females grouped by F1 mode of conception

(sex/parthenogenesis) and F1 mating status (mated/virgin) 135

Figure 5-5

The effect of trans-generational reproductive mode (F0 MOC – F1 MOC) on the total number of eggs produced by F1 females; Levels not connected by same letters are significantly different based on post hoc Tukey’s test. 136

Figure 5-6

Mean proportion of hatched and moulted progeny derived from eggs produced by mated (grey) and virgin (white) F1 (a) and FieldF0 b) females; Asterisks indicate significant differences. 139

xvii List of Tables

Chapter 1

Table 1-1

the empirical evidence for some of the models explaining the evolution sex 10

Chapter 2

Table 2-1

The effect of male body mass, male age, mating number and mating interval on a)

spermatophore mass, b) sperm density (Ln transformed) and c) viable sperm

proportion across males 46

Chapter 3

Table 3-1

Primer sequences, repeat motifs, size ranges, genotyping error rates and

Genebank accession number for 18 microsatellite markers isolated from

Extatosoma tiaratum 71

Table 3-2

Characterization of 15 microsatellite markers estimated from 11 individuals from

the Crystal Cascades, North Queensland, Australia 72

Table 3-3

xviii Microsatellite profiles of captive-bred E. tiaratum parents and offspring obtained

from three families 73

Table 3-4

Number of homozygous and heterozygous offspring produced

parthenogenetically by heterozygous mothers and rates of transition to

homozygosity (R) 75

Table 3-5

G-tests for observed and expected rates of transition to homozygosity (R)

depending on the mechanism of parthenogenesis (P = 0.05). R is not significantly

different from expected rates under automixis with terminal fusion for any loci. 76

Chapter 4

Table 4-1

Number of heterozygous microsatellite loci and the inferred mode of conception

(MOC) for F0 and F1 100

Table 4-2

Average haemocyte concentration (per ul), lytic activity (Δ absorbance) and PO

activity (Δ absorbance) in F1 females with either sexual or parthenogenetic

MOCs, derived from F0 with either sexual or parthenogenetic MOCs 101

Table 4-3

xix The effect of F0 and F1 mode of conception (MOC) and week of sampling (3-, 5-

and 7- weeks) on a) Log haemocyte concentration, b) Lytic activity and c) PO

activity 102

Chapter 5

Table 5-1

Number of heterozygous microsatellite loci and the inferred mode of conception

(MOC) for F0 and F1 females. Three out of five F0 females that had been

conceived parthenogenetically, produced progeny with parthenogenetic MOCs

(shown in bold) following mating. 130

Table 5-2

Number of heterozygous microsatellite loci and the inferred mode of conception

(MOC) of F2 female progeny, derived from mated F1 females. All females with

parthenogenetic MOCs, produced some offspring with parthenogenetic MOCs

(shown in bold). 131

Table 5-3

Number of heterozygous microsatellite loci and the inferred mode of conception

(MOC) for field-collected females (FieldF0), and their female offspring (FieldF1)

produced following mating. All FieldF0 females produced offspring with sexual

MOCs following mating. 132

Table 5-4

xx The effect of F1 mode of conception (MOC) and F1 mating status (virgin or mated) on F1 : a) body size, b) lifespan, c) weekly oviposition rate, and d) total eggs produced 134

Table 5-5

The effect of F0 and F1 mode of conception (MOC), and F1 mating status (virgin or mated) on F2: a) hatching frequency, b) moulting frequency, and c) development time 137

xxi

Chapter 1

Introduction: the evolutionary significance of sex

and parthenogenesis

1 Modes of reproduction

The capacity to reproduce is a core feature of living organisms, and a fundamental requirement for evolution. There are two general modes of reproduction: asexual, in which offspring are produced by one parent and inherit the genetic material of this mother, and sexual, where two parents are involved in offspring production and half of the offspring’s genome is inherited from each parent. Sexual reproduction involves chromosomal recombination and segregation – processes that shuffle genomes and can

(but need not) create variable offspring (Otto 2009). Asexual reproduction occurs through a diverse range of mechanisms from fission, fragmentation and budding

(vegetative reproduction) to the production of mitotic and meiotic eggs. As vegetative reproduction in plants does not result in the production of seeds, a necessity for dispersal, vegetative reproduction is considered more similar to growth than reproduction (Vrijenhoek 1998). In this thesis, I focus on parthenogenesis, where the development of unfertilised eggs typically results in the production of exclusively female offspring via a form of asexual reproduction. Parthenogenesis is also referred to as unisexual reproduction, especially where the exact mode of parthenogenesis is not known (Suomalainen 1976).

Although primitive forms of genetic mixing exist in prokaryotes, sexual reproduction in eukaryotes is a more complex process involving the persistence of diploidy, sex chromosomes and gonads. The oldest sexually reproducing organism reported from fossil records is the red algae Bangiomorpha pubescens (ca. 1200 Ma) (Butterfield

2 2000). Sexual reproduction has subsequently spread and the majority of eukaryotic organisms reproduce sexually. Nevertheless, a small proportion (about one in every

1000 taxa) has lost sex entirely and reproduces parthenogenetically (Vrijenhoek 1998).

These parthenogenetic taxa are distributed across the tree of life, occurring in diverse phyla including plants, rotifers, nematodes and arthropods (Bell 1982).

Parthenogenetic taxa generally occupy terminal nodes in phylogenetic trees, and were previously referred to as evolutionary dead ends (Simon et al 2003; but see Schwander and Crespi 2009, Schwander et al 2011).

Parthenogenesis: modes, mechanisms and emergence

Suomalainen (1976) classified parthenogenesis into two general types: 1) tychoparthenogenesis, often referred to as occasional or accidental parthenogenesis, in which a small proportion of unfertilised eggs may develop into a normally sexually reproducing species, and is common in arthropods; and 2) normal parthenogenesis, which can be further categorized as a) obligate parthenogenesis in which the eggs always develop parthenogenetically throughout all generations; b) cyclic parthenogenesis, that is where a species is parthenogenetic in alternating generations; c) paedogenesis in which the eggs of the individuals at the larval stage develop parthenogenetically; or, d) facultative parthenogenesis, in which a single female can produce both fertilised and unfertilised eggs that develop into viable offspring.

There are two main cytological mechanisms of parthenogenesis, which determine the chromosome numbers and levels of genetic diversity in the resulting offspring: 1)

3 generative or haploid parthenogenesis, in which a chromosome reduction takes place and the resulting offspring are haploid, and 2) somatic parthenogenesis in which the offspring are diploid or polyploid. Somatic parthenogenesis is more common and occurs through either apomixis or automixis. In apomixis, the egg cells are produced mitotically and the resulting offspring are genetically identical to their mother. In automixis, chromosomal reduction takes place through meiosis, and diploidy is restored through: a) gamete duplication, b) fusion of the resulting sister nuclei

(terminal fusion), c) fusion of non-sister nuclei (central fusion), or d) random fusion.

These processes are illustrated in Figure 1-1. In automixis with gamete duplication, heterozygosity is entirely lost, and all loci are homozygous in the offspring. With terminal fusion, most loci lose heterozygosity, except loci located far away from the centromere with high recombination rates. Automixis with central fusion has similar outcomes to apomixis, as heterozygosity is restored in most loci except those located far away from the centromere, which have high recombination rates.

Parthenogenetic lineages are typically derived from sexual ancestors and the mode of this transition often determines the type and the mechanism of parthenogenesis

(Schwander and Crespi 2009). This has important consequences for the resulting lineages in terms of genetic diversity and hence ecological fitness (Bell 1982). First, spontaneous mutations in genes involved in sexual reproduction (gamete production, egg fertilisation or mating behavior) can cause parthenogenesis (see Simon et al 2013).

4 Locus close to centromere Locus far from centromere (Without recombination) (With recombination)

Meiosis I

Meiosis II

Gamete duplication Gamete duplication Terminal fusion Terminal fusion

Central fusion Central fusion

Random fusion Random fusion

Figure 1-1 Cytological mechanisms involved in different modes of automictic parthenogenesis and their consequences in terms of inbreeding, for loci located close to the centromere with low recombination rates (left) and far away from the centromere with high recombination rates (right) (Modified from Pearcy et al 2006). Larger circles represent nuclei; lines represent chromatids with crosses marking the location of the centromere. Black and white circles show alleles at a given locus. Maternal genotype is heterozygous and parthenogenesis results in homozygous offspring.

These mutations can initially cause automixis by the preservation of meiotic oocyte production and diploidy restoration via one of the mechanisms described above

(Suomaleinen 1976, Bell 1982, Schwander and Crespi 2009). Spontaneous mutations can create apomictic lineages either directly from sexuality or through a stepwise process via automixis (Suomaleinen 1976, Schwander and Crespi 2009).

The second mechanism of transition to parthenogenesis is through infection with intracellular bacteria. The most well-known intracellular bacteria, which can induce parthenogenesis, are Wolbachia from the proteobacteria family (Stouthamer et al

5 1993) and Cardinium from the bacteroidetes group (Zchori-Fein et al 2001, Weeks et al 2003). Wolbachia is common (although not restricted) to species with haplo-diploid sex determination and can cause production of homozygous female offspring through gamete duplication (Stouthamer 1999). Hybridization is another mode of transition to asexuality common in many vertebrates and invertebrates (Bell 1982). Hybridization often results in polyploidy and the resulting polyploids can suffer reduced viability resulting from complications in chromosomal pairing during meiosis (White 1964,

Barton and Gale 1993).

The transition from sexual reproduction to automixis does not involve drastic cytological changes because the fusion of male and female gametes is replaced by fusion of female gametes (Schwander and Crespi 2009), and can occur through spontaneous mutations. Such mutations are relatively common in invertebrate taxa and may result in tychoparthenogenesis (development of unfertilised eggs in sexual species) (Bell 1982, Schwander et al 2010). Higher rates of automictic egg development have been induced by artificial selection in laboratory populations of the sexual reproducing Drosophila mercatorum (Templeton et al 1976). A similar process can occur through natural selection if there is an adaptive advantage for automictic reproduction, for example when there is a high risk of mating failure. However, because automixis often results in loss of hetereozygosity (and reduced fitness), strictly automictic lineages are unlikely to win the competition with their sexual ancestors (Schwander and Crespi 2009).

6 The paradox of sex: theories and empirical evidence

The prevalence of sexual reproduction in nature has been referred as “the queen of problems in evolutionary biology” (Bell 1982), having challenged biologists during the past century and remains largely unresolved. The paradox of sex is that sexual reproduction is far more costly compared with asexual reproduction in terms of time, energy and risk of predation and infection (Arnqvist and Nilsson 2000, Lehtonen

2012), and yet it is the dominant mode of reproduction across the tree of life. The most well recognised cost of sex is the ‘twofold cost’ or ‘cost of males’, which highlights the fact that sexual females cannot reproduce on their own, and must therefore produce males (Maynard Smith 1978, Bell 1982). Hence, ceteris paribus (all else being equal), a non-sexual mutant in a sexual species with an equal male-female ratio would double its relative representation in the next generation, and the mutation would therefore spread in the population (Maynard Smith 1978). Another cost of sex is that favorable gene combinations are broken up by recombination (Nei 1967, Lynch and Deng 1994).

Selection favours combinations of adaptive genes under constant environmental conditions, and breaking these up through recombination may be disadvantageous.

Further costs of sex include those associated with searching for suitable mating partners (Gascoigne et al 2009), and a greater exposure to predation and infection

(Otto et al 2009).

Darwin first noted the importance of sexual reproduction in creating the essential variation for evolution (Darwin 1960 pp 41 and 42) even though he later discarded it

7 as a reason for why sex exists (see Shön et al 2009). Weismann (1889) was the first to propose a theory for the evolution of sex based on the advantages of creating variable offspring. Since then, numerous theoretical models attempt to explain the predominance of sexual reproduction by describing conditions under which the benefits of sex outweigh the costs. Most of these theories rely on the advantages of genetic mixing, i.e. recombination and segregation. For example, the Fisher-Muller hypothesis (Fisher 1930, Muller 1932) suggests that sex allows the combining of beneficial mutations into the same genome. Similarly, sex may be advantageous under adaptations to novel environments, because it breaks apart selection interference between loci and increases the efficacy of selection (Hill and Robertson 1966).

Another potential advantage of sex is that it releases Muller’s ratchet (Muller 1964), caused by the accumulation of irreversible deleterious mutations in finite asexual populations.

One of the most well studied explanations is parasite resistance hypothesis for sex

(also known as the Red Queen model) that emphasises the advantage of genetic mixing in the co-evolutionary arms race between hosts and parasites (Van valen 1973, Jaenike

1978, Hamilton 1980). In this model, hosts benefit from producing sexual offspring with rare genetic combinations, which are more likely to be resistant to the common parasite genotypes. While the importance of sexual reproduction for maintaining an effective immune system against coevolving parasites has been established (e.g. Otto and Nuismer 2004, Decaestecker et al 2007, Morran et al 2011, Vergara et al 2014,

Auld et al 2016), it remains unclear whether sex has any immediate advantages, either in terms of increasing the efficacy of individual immune response, or through an

8 enhanced ability to resist diseases and infections. Exposure to parasites and pathogens may decrease host reproductive fitness due to reduced longevity and/or reproductive out put (Moore 2002), while immune function may prevent such costs (Roitt et al

1996).

Alternate models suggest immediate advantages and offer evolutionary explanations for the maintenance of sex (Hartfield and Keightley 2012). For example, sexual reproduction may have direct advantages because it repairs damaged DNA through recombination (Bernstein et al 1988) or because it facilitates transmission of selfish genes known as transposable elements (Goddard et al 2001).

While empirical evidence for the maintenance of sex models is not as abundant as theory, there is some support for these ideas in some species (Table1-1).

9 Table 1-1 the empirical evidence for some of the models explaining the evolution of sex Supporting Type of Theory tested Models proposed by Species benefits of References study sex

Adaptation to new Fisher 1930, Muller 1932, Hill and Saccharomyces Wolf et al 1987, Goddard et al environment Robertson 1966 cerevisiae yes Laboratory 2005 Drosophila melanogaster yes Laboratory Rice and Chippindale 2001 Chlamydomonas reinhardtii yes Laboratory Kaltz and Bell 2002 Brachionus calyciflorus yes Laboratory Becks and Agrawal 2012

Accumulation of deleterious Salmonella mutation Muller 1930 typhimurium yes Laboratory Andresson and Hughes 1996 Saccharomyces cerevisiae yes Laboratory Zeyl and Bell 1997 Aphid species yes Laboratory Normark and Moran 2000 Arabidopsis no Laboratory Wright et al 2002 Aspergillus nidulans yes Laboratory Bruggeman et al 2003 Daphnia pulex yes Laboratory Paland and Lynch 2006 Caenorhabditids species no Laboratory Cutter and Payseur 2003 Caenorhabditis elegans yes Laboratory Morran et al 2009 Potamopyrgus antipodarum yes Field Neiman et al 2010 Timema species yes Field Henry et al 2012

Van valen 1973, Jaenike 1978, Potamopyrgus Lively 1987, Jokela and Lively The Red Queen Hamilton 1980 antipodarum yes Field 1995 Psychid moths yes Field Kumpulainen et al 2004 Tribolium castaneum yes Laboratory Fischer and Schmid-Hempel 2005

10 Daphnia no Laboratory Killick et al 2006 Daphnia yes Field Decaestecker et al 2007 Potamopyrgus Koskella and Lively 2009, Jokela antipodarum yes Laboratory et al 2009 Potamopyrgus antipodarum yes Field King et al 2009 Caenorhabditis elegans yes Laboratory Morran et al 2011 bag warm moths no Field Elzinga et al 2012

DNA repair Bernstein et al 1988 Bacillus subtilis n o Laboratory Redfield 1993 Haemophilus influenzae no Laboratory Redfield 1993 Streptococcus pneumoniae no Laboratory Engelmoer and Rozen 2011

11 What is missing?

A significant challenge for studies that investigate the relative costs and benefits of sexual versus parthenogenetic reproduction is to resolve potentially confounding effects that are associated with, but not directly related to the two reproductive modes.

For example, parthenogenetic populations may be hybrids of multiple sexual populations, and thus have different ploidy levels (see Kearney 2005). Alternatively, these populations may occupy different spatial or temporal spaces compared with sexual populations (e.g. Lynch 1984, Hadany and Otto 2009). Cyclic parthenogenetic species may resolve some of these issues and thus provide more promising models

(e.g. rotifers: Becks and Agrawal 2010; Daphnia: Allen and Lynch 2008, Decaestecker et al 2009; aphids: Bulmer 1982, Simon et al 2002), but many of these species alter their reproductive modes according to local environmental conditions, and/or have different life histories that may include resting eggs.

An ideal species to explore the costs and benefits of sex is one in which extrinsic cofounding effects (i.e. not directly related to reproductive modes) are minimised, and the differences in progeny produced through each mode can be attributed to the effect of the reproductive mode itself.

12 Facultatively parthenogenetic species as models

Facultatively parthenogenetic species, in which females can produce, throughout their lives, both unfertilised yet viable eggs and sexually fertilised eggs provide an opportunity to evaluate the outcomes of sex and parthenogenesis. Such species are free from the confounding effects outlined above, including differences in ploidy levels, differences related to local adaptations, or differences in diapause duration of sexual and parthenogenetic eggs. In facultatively parthenogenetic species, the difference between the two reproductive modes is directly (and likely exclusively) associated with the effects of genetic mixing (recombination and segregation) and seminal fluid components. Accordingly, facultative parthenogens have been identified as suitable species for studying the evolution and maintenance of sex and parthenogenesis (e.g.

Corely and Moore 1999, Corely et al 2001, Matsuura and Nishida 2001, Matsuura et al

2004, Schneider and Elgar 2010, Burke et al 2015, Kobayashi and Miaguni 2016).

Facultative parthenogenesis is taxonomically widespread; occurring in insects, reptiles, fish, birds and sharks (Booth et al 2012), but its adaptive significance is poorly understood. Historically, it was thought to be advantageous in environments where the likelihood of sexual reproduction was reduced through a lack of available males

(Stalker 1956). Under these circumstances, the ability to reproduce parthenogenetically would allow a female to ‘rescue’ some degree of fitness, even in the absence of fertilisation, albeit with potentially lower hatching rates and offspring viability compared with sexual reproduction (e.g. Stalker 1956, Gerristen 1980, Kramer and

13 Templeton 2001). Thus, facultative parthenogenesis is predicted when population densities and male encounter rates are low (Courchamp et al 2008, Gascoigne et al

2009). Reduced fertilisation success due to insufficient transfer of sperm during mating has also been suggested as a potential selective force (Elzinga et al 2011), although direct empirical evidence for this hypothesis is yet to be provided.

The study system, an Australian Phasmatid

The Phasmatodea (stick insects) comprises approximately 500 genera, including obligate sexual, facultative parthenogenetic and obligate parthenogenetic species

(Whiting et al 2003). Stick insects therefore represent a promising group for studying the evolution of sexual and parthenogenetic reproduction. While the ancestral mode of reproduction is sexual, many species have lost sex entirely or partially (facultative parthenogenetic). Recent evidence suggests that Wolbachia infection is unlikely to be the cause of parthenogenesis in this group (Perez-Ruiz et al 2015). The transition between sex and parthenogenesis has been studied in a number of stick insects, including species from the Timema genus (Schwander and Crespi 2009, Schwander et al 2010, Henry et al 2012), the Bacillus genus (Mantovani et al 1997, Scali et al 2003,

Schön et al 2009) and the common tea-tree stick insect from New Zealand, Clitarchus hookeri (Morgan-Richards et al 2009).

The Australian spiny leaf insect (or Macleay’s Spectre), Extatosoma tiaratum is a facultative parthenogenetic phasmatid found in the eastern seaboard rainforests of

14 southern Queensland and northern New South Wales, Australia. It is easy to rear in captivity, and feeds on a range of plants species including Callicoma, Eucalyptus and

Rosaceae (Brock and Hasenpusch 2009). Males and females are sexually dimorphic.

The apterous (wingless) females (Figure 1-2a) are the larger sex (average female length is ~12 cm; male length is ~8 cm, Bian et al 2015), while males (Figure 1-2b) are winged and occasionally fly (, Brock and Hasenpusch 2009). Females reach sexual maturity after six immature instars, compared with five instars for males (Carlberg

1986). Two chromosomal numbers have been reported for populations of this species, both with XO males and XX female sex chromosomes (35 and 36 chromosomes in central New South Wales and 37 and 38 chromosomes in the north coast of New South

Wales, Craddock 1972).

Adult females commence producing eggs approximately four weeks after moulting, irrespective of whether they have mated (Carlberg 1983). Females oviposit continuously throughout their adult life (Carlberg 1983). Mated females produce offspring of both sexes, while virgin females produce unfertilised eggs that are all female. Virgin females remain receptive to mating even after they have commenced oviposition, and it is possible to manipulate the reproductive mode of any given female. Copulation lasts up to 19 hours, during which time the male attaches an external spermatophore to the female’s genitalia (Figure 1-2c, Carlberg 1983).

15 a b c

Figure 1-2 Adult Extatosoma tiaratum (a) female; (b) male; and (c) copulating pair. In the copulating pair, the male is above the female and the visible spermatophore (white sphere close to the final female tergite) is being transferred to the female.

The timing of reproduction and the number of eggs produced by females of E. tiaratum is related to their juvenile social environment, suggesting that parthenogenesis is strategic in this species (Schneider and Elgar 2010). Females reared in the presence of males (but provided with no opportunity to mate) delayed commencing oviposition and produced fewer eggs compared with females reared in the absence of males (Schneider and Elgar 2010). Virgin females reared in the presence of males delay oviposition and produce fewer eggs compared with females reared in the absence of males (Schneider and Elgar 2010). Virgin females that have not commenced oviposition produce chemical signals that attract males, but ovipositing females fail to do so (Schneider and Elgar 2010). It is therefore hypothesised that parthenogenesis is an adaptive response to male limitation in this species (Schneider and Elgar 2010). The rates of parthenogenesis in E. tiaratum in their natural environment remains unknown.

16 Extatosoma tiaratum provides a powerful model to investigate the costs and benefits of sex and parthenogenesis. As focal E. tiaratum females are capable of reproducing through both reproductive modes, comparisons are not confounded by ploidy levels or ecological factors (see above). Rather, the mode of egg conception (sexual or parthenogenetic) appears to be a plastic response that varies with the presence or absence of sperm, and is unlikely to involve physiological differences between the two types of eggs (sexual and parthenogenetic). Furthermore, it is possible to create females with similar genetic backgrounds (sister pairs) and age, and assign them to sexual or parthenogenetic reproductive modes, an experimental design that allows comparisons between the consequences of sex and parthenogenesis, while minimizing confounding effects.

The evolutionary significance of facultative parthenogenesis in E. tiaratum

My PhD research had two main aims. First, I investigated the evolutionary significance of facultative parthenogenesis in Extatosoma tiaratum. Here, I specifically asked why do female E. tiaratum continue producing males, if sexual reproduction is costly as predicted by the theory (Otto 2009, Lehtonen et al 2012, Meirmans et al

2012)? Conversely, if parthenogenesis is associated with fitness costs as shown in some species, how is the ability to reproduce parthenogenetically maintained? Second,

I used this unique reproductive system to compare the relative fitness outcomes of sex and parthenogenesis, while minimizing confounding effects that have challenged previous studies. I demonstrate that the ceteris paribus assumption does not hold, thus

17 the two-fold cost of sex is absent in E. tiaratum. Parthenogenetically conceived females are half as fecund as sexually conceived females, either as a result of a loss of heterozygosity or due to parthenogenetic itself. However, the cost of parthenogenesis can be compensated in the next generation if parthenogenetically conceived females produce daughters sexually. Therefore, female E. tiaratum can enjoy the benefit of parthenogenesis allowing them to reproduce in the absence of sufficient or preferred sperm, while mediating the costs of homozygosity by reproducing sexually.

While facultative parthenogenesis is taxonomically widespread, its evolutionary significance is poorly understood. One view is that parthenogenesis is an adaptation to male- (Stalker 1956) or sperm- (Elzinga et al 2011) limitation, as selection is expected to favour parthenogenesis when a large proportion of eggs remain unfertilised (Kramer and Templeton 2001, Schwander et al 2010). Few if any, studies have investigated either the degree of male multiple mating or the capacity of males to produce sperm at each mating attempt in facultatively parthenogenetic species: data that are important for investigating the link between the evolution of parthenogenesis and males sperm limitation. In chapter 2, I investigated male lifetime mating frequency and variation in ejaculate size and quality with multiple mating in E. tiaratum. I provided male E. tiaratum with weekly mating opportunities with different females throughout their adult life, and measured mating frequency and ejaculate size and quality. I show that most, but not all males are able to mate multiply, but that ejaculate size (both spermatophore size and sperm density), and quality (estimated as the proportion of viable sperm, see Figure 1-3) decrease with increased number of matings (Alavi et al

18 2016). In addition, ejaculate size increased with increasing time interval between matings, suggesting that E. tiaratum males require time to replenish ejaculate reserves.

The results of chapter 2 add to the growing body of evidence for male physiological constraints on ejaculate production (see Wedell et al 2002), and suggest sperm limitation may be one of the factors influencing the evolution of facultative parthenogenesis from obligate sexuality in E. tiaratum. Sperm limited males are likely to practice mate choice or differentially allocate ejaculate to preferred females (Wedell et al 2002). Male prudence may limit female reproductive success, and therefore can exacerbate the selection for parthenogenetc egg development. A more direct approach to finding the link between sperm limitation and the evolution of parthenogenesis would be to allow male E. tiaratum with different mating histories (virgin, once mated, twice mated, etc) to copulate with virgin females, and investigate the proportion of fertilised eggs.

While male and sperm limitation may explain the evolution of facultative parthenogenesis from obligate sexuality, they are perhaps insufficient to explain its maintenance. To understand the maintenance of both sex and parthenogenesis in E. tiaratum, it is necessary to know the cytological mechanism of parthenogenesis as this determines the genetic diversity and heterozygosity levels of the offspring and may thus have important implications for individual-level fitness.

19 a b c

Figure 1-3 Extatosoma tiaratum sperm 20× under light (a) and fluorescent microscopy (blue excitation filter at λ = 490 nm, b and c); b) shows dead sperm cells stained by propidium iodide and c) shows viable sperm cells stained by SYBR14 (LIVE/DEAD assay, Molecular Probes; Sigma, Australia, L-7011)

In chapter 3, I describe the cytological mechanism of parthenogenesis in E. tiaratum, by comparing the level of heterozygosity of five mothers with that of their parthenogenetic offspring (following Pearcy et al 2006). I first isolated 15 nuclear microsatellite markers using specimens of E. tiaratum collected from a natural population, and used them to calculate the proportion of homozygous offspring produced by heterozygous mothers for each locus. These data revealed that the most likely mechanism of parthenogenesis in E. tiaratum is automixis with terminal fusion, resulting in substantial losses of heterozygosity in the first generation parthenogenetic offspring (Alavi et al 2016).

Loss of microsatellite heterozygosity may be associated with loss of heterozygosity at fitness related loci, and therefore may be an important component determining

20 reproductive fitness (Väli et al, 2008). Based on these data, I predicted that parthenogenesis would lead to demonstrable fitness costs in E. tiaratum. This prediction was tested with two experiments; in chapter 4, I explored the consequences of reproductive modes for female immune function; in Chapter 5, I assessed female reproductive success over two generations.

Immune function determines the ability of individuals to defend themselves against injury and disease, and thus can have a significant impact on fitness. In Chapter 4, I investigated the effect of trans-generational reproductive modes on female immune function in E. tiaratum. I derived sexual or parthenogenetic first generation female offspring from mothers conceived either sexually or parthenogenetically (inferred from heterozygosity levels). I then measured immune function of the first generation female offspring over multiple time points using three commonly used innate immune function assays: haemocyte concentration, lysozyme-like (lytic) activity and phenoloxidase activity. I showed that a female’s haemocyte concentration and lytic activity is related to her maternal mode of conception. Females derived from sexually conceived mothers had significantly higher haemocyte concentration and lytic activity than females derived from parthenogenetically conceived mothers. In addition, both haemocyte concentration and lytic activity declined with age, suggesting physiological constraints on maintaining immune function over time (Siva-Jothy et al 2005).

The results of chapter 4 suggest there may be a negative, trans-generational effect of parthenogenesis on immune function in E. tiaratum, however the approach taken does

21 not permit a direct assessment of causality. A more powerful approach would be to directly infect sexual and parthenogenetic offspring with a pathogen, and measure fitness outcomes such as variation in offspring production and survival. Such a study across generations will be a powerful test of the Red Queen model. The importance of trans-generational genetic effects on immune function highlighted in chapter 4 may explain some of the inconsistencies we see across studies exploring the effect of inbreeding on immune function (e.g. Rantala and Roff 2007 negative correlations,

Stevens et al 1997, Gerloff et al 2003 no effects and Gershman et al 2010, Drayton and

Jennions 2011, Franke and Fischer 2013 positive correlations).

Finally in chapter 5, I compared the effect of sex and parthenogenesis on female reproductive success across two generations of E. tiaratum. I derived sexual or parthenogenetic first generation female offspring from mothers conceived either sexually or parthenogenetically (siblings of those used in chapter 4). I then assigned first generation sisters to either a virgin or a mated treatment and derived second generation offspring. I then investigated the effect of mode of conception and mating status on female development time, body size, lifespan, fecundity and offspring survival (proportion hatched and proportion moulted to second instar). I demonstrated that several life history traits (development time, body size, lifespan and egg production), depended on the female’s mode of conception, whereas the survival of her offspring was related to her mating status. Sexually conceived females were typically larger, developed to adults 20 days earlier, lived twice as long and produced twice as many eggs as females conceived parthenogenetically. Offspring hatching and moulting frequency was higher in mated females compared to virgins. There was also evidence

22 for an indirect cost of parthenogenesis in terms of the ability of females to utilise sperm. Parthenogenetically conceived, mated females were more likely to produce parthenogenetic offspring compared to sexually conceived, mated females.

The results of chapter 5 contradict the ceteris paribus assumption of the two-fold cost of sex, adding to the growing body of empirical support suggesting that this assumption does not always hold (e. g. Hong and Ando 1998, Kramer and Templeton

2001). Two consecutive generations of parthenogenesis severely reduces the reproductive prospects of females, including reduced oviposition rate and lifespan, as well as offspring hatching frequency, moulting frequency to second instar and development time to adulthood. The substantial loss of microsatellite heterozygosity as a result of parthenogenesis highlighted in chapter 3 may be responsible, particularly if the loss occurs at fitness related loci (Väli et al, 2008) and increases the expression of recessive deleterious alleles (Charlesworth and Charlesworth, 1999; Roff, 2002).

The findings of chapter 4 and 5 provide valuable insights into the maintenance of facultative parthenogenesis in this species. Even though parthenogenesis is associated with substantial fitness costs, such costs may be recovered with a single generation of sexual reproduction. For example, if a parthenogenetically conceived female that is likely small, and has low fecundity, mates and utilises sperm to fertilise her eggs, her offspring will have better chances of survival and increased fecundity. If she was derived from a sexually conceived mother, her offspring may have increased immune function, which likely increases their chances of survival and lifespan. Those

23 parthenogenetically conceived females that fail to reproduce sexually, due to 1) a lack of males, 2) mating with a sperm-depleted male, or 3) failure to utilize sperm, will have substantially lower reproductive success. Furthermore, the granddaughters of such females may have compromised immune function, which may reduce their survival and reproductive success.

The findings of chapter 2 and 5 generate some predictions regarding reproductive strategies of males and females in E. tiaratum. It is shown that 1) males are physiologically constrained with respect to their ability to mate and the amount of ejaculate they transfer to females, and 2) females vary in their quality (fecundity) based on their mode of conception. These are two of the three basic factors that facilitate the evolution of male mate choice (see Bondouriansky 2001). The third factor is the likelihood of future mating opportunities. Populations of E. tiaratum are expected to be female-biased, because only female offspring are produced through parthenogenesis. Therefore, a male that is able to distinguish between sexually and parthenogenetically produced females, for example based on their size (sexually conceived females are typically larger), and copulates with the former, will increase his reproductive success because such females have higher fecundity. The fact that parthenogenetically conceived, mated females are more likely to reproduce parthenogenetically compared to sexually conceived mated females (Chapter 5), suggests that either parthenogenetically conceived females are constrained in their ability to produce sexual offspring, or that males are investing differentially depending on a female’s mode of conception. Even though spermatophores were transferred to all females, males may have manipulated the size and/or quality of the sperm transferred

24 to parthenogenetically conceived females. For females, the cost of remaining unfertilised depends on their reproductive history. A female that is parthenogenetically conceived will have substantially lower reproductive success if she remains a virgin compared to a sexually conceived virgin female. Therefore, such parthenogenetically produced females have more incentive to invest in pheromone production to attract males. An experiment comparing male response to pheromone produced by females with sexual and parthenogenetc histories (extending the approach of Schneider and

Elgar 2010) will be informative.

Overall, the findings of this thesis offer a new perspective in understanding the evolution and maintenance of facultative parthenogenesis in E. tiaratum. Ancestral sexually reproducing female lineages would have been stable, given they were provided with sufficient sperm to fertilise their eggs (Figure 1-4a). If they failed to mate, due to factors such as low male encounter rates/female-biased sex ratios, or male sperm limitation/ prudence, they would become extinct (Figure 1-4b). A female lineage that acquired the ability to reproduce parthenogenetically, would reduce its reproductive success at each generation and would also eventually become extinct

(Figure 1-4c). However, a facultative parthenogenetic female lineage that could persist reproducing parthenogenetically until a mating opportunity arose, would recover from homozygosity, and the associated fitness costs of parthenogenesis (Figure 1-4d, but see

Gerber and Kokko 2016). Female lineages alternating between sex and parthenogenesis may have comparable reproductive success to obligate sexual lineages, while being less susceptible to extinction. Cycles of male abundance or seasonal fluctuations in population densities (Gascoigne et al 2009) may induce such

25 alternation between reproductive modes in natural populations. In addition, females may have some level of control over their reproductive mode, for example by adjusting pheromone signalling (Schneider and Elgar 2010) or post-copulatory regulation of sperm usage (Eberhard 1996). Alternatively, E. tiaratum females may adopt a mixed reproductive strategy by fertilising only a proportion of their eggs following mating to maximize their reproductive success. Indeed, the benefits of facultative parthenogenesis along with the widespread evidence for female mating failure

(Rhainds 2010) raise the question why this reproductive system is not more widespread in insects.

Figure 1-4 A schematic illustration of consequences of sex (continuous lines) and parthenogenesis (dashed lines) in terms of reproductive success across generations: a) an obligately sexual lineage with constant reproductive success, b) an obligate sexual lineage that fails to mate successfully becomes extinct, c) an obligately parthenogenetic lineage has reduced reproductive success at each generation, and d) a facultative parthenogenetic lineage alternating between sex and parthenogenesis will recover its reproductive success with sex

26 References

Alavi, Y., Elgar, M. A., and Jones, T. M. (2016). Male Mating Success and the Effect of Mating History on Ejaculate Traits in a Facultatively Parthenogenic Insect (Extatosoma tiaratum). Ethology, 122(7), 523-530. Alavi, Y., van Rooyen, A., Elgar, M. A., Jones, T. M., and Weeks, A. R. (2016). Novel microsatellite markers suggest the mechanism of parthenogenesis in Extatosoma tiaratum is automixis with terminal fusion. Insect Science, in press. Andersson, D. I., and Hughes, D. (1996). Muller’s ratchet decreases fitness of a DNA- based microbe. Proceedings of the National Academy of Sciences, 93(2), 906– 907. Arnqvist, G., and Nilsson, T. (2000). The evolution of polyandry: multiple mating and female fitness in insects. Animal Behaviour, 60(2), 145–164. Auld, S. K. J. R., Tinkler, S. K. and Tinsley, M. C. (2016). Sex as a strategy against rapidly evolving parasites. Proceeding of the Royal Society. B. 283: 20162226 Barton, N. H., and Gale, K. S. (1993). Genetic analysis of hybrid zones. Oxford: Oxford Univ. Press. Becks, L., and Agrawal, A. F. (2011). The effect of sex on the mean and variance of fitness in facultatively sexual rotifers. Journal of Evolutionary Biology, 24(3), 656–664. Bell, G. (1982). The masterpiece of nature: the evolution and genetics of sexuality. Book, Berkely: University of California. Bernstein, H., Hopf, F. A., and Michod, R. E. (1988). Is meiotic recombination an adaptation for repairing DNA, producing genetic variation, or both. The Evolution of Sex, 139–160. Bonduriansky, R. (2001). The evolution of male mate choice in insects: a synthesis of ideas and evidence. Biological Reviews, 76(3), 305–339. Booth, W., Smith, C. F., Eskridge, P. H., Hoss, S. K., Mendelson, J. R., and Schuett, G. W. (2012). Facultative parthenogenesis discovered in wild vertebrates. Biology Letters, 8(6), 983–985. Brian Charlesworth, and Deborah Charlesworth. (1999). The genetic basis of inbreeding depression. Genetical Research, 74(3), 329–340. Brock, P. D., and Hasenpusch, J. W. (2009). Complete Field Guide to Stick and Leaf Insects of Australia. Bruggeman, J., Debets, A. J. M., Wijngaarden, P. J., deVisser, J. A., and Hoekstra, R. F. (2003). Sex slows down the accumulation of deleterious mutations in the

27 homothallic fungus Aspergillus nidulans. Genetics, 164(2), 479–485. Burke, N. W., Crean, A. J., and Bonduriansky, R. (2015). The role of sexual conflict in the evolution of facultative parthenogenesis: a study on the spiny leaf stick insect. Animal Behaviour, 101, 117–127. Journal Article. Butterfield, N. J. (2000). Bangiomorpha pubescens n. gen., n. sp.: implications for the evolution of sex, multicellularity, and the Mesoproterozoic/Neoproterozoic radiation of eukaryotes. Paleobiology, 26(3), 386–404. Carlberg, U. (1983). Copulation in Extatosoma tiaratum (MacLeay)(Insecta: Phasmida). Zoologischer Anzeiger. 5(6), 340-356

Carlberg, U. (1986). Culturing stick-and leaf-insects (Phasmida)--a review. Zeitschrift Fur Versuchstierkunde, 29(1–2), 39–63. Corley, L. S., Blankenship, J. R., and Moore, A. J. (2001). Genetic variation and asexual reproduction in the facultatively parthenogenetic cockroach Nauphoeta cinerea: implications for the evolution of sex. Journal of Evolutionary Biology, 14(1), 68–74. Corley, L. S., and Moore, A. J. (1999). Fitness of alternative modes of reproduction: development constraints and the evolutionary maintenance of sex. Proceedings of the Royal Society B: Biological Sciences, 266(14-18). Courchamp, F., Berec, L., and Gascoigne, J. (2008). Allee effects in ecology and conservation. Environ. Conserv, 36, 80–85. Craddock, E. (1972). Chromosomal diversity in the Australian Phasmatodea. Australian Journal of Zoology, 20(4), 445–462. Cutter, A. D., and Payseur, B. A. (2003). Rates of deleterious mutation and the evolution of sex in Caenorhabditis. Journal of Evolutionary Biology, 16(5), 812– 822. Darwin, C. R. (1960) First notebook [B] (July 1837–February 1838). In de Beer G (ed) Darwin’s notebooks on transmutation of species. Part I. Bull British Mus (Nat Hist). Histor Ser 2(2):23–73, pp. 42–73 Decaestecker, E., Gaba, S., Raeymaekers, J. A. M., Stoks, R., Van Kerckhoven, L., Ebert, D., and De Meester, L. (2007). Host–parasite Red Queen dynamics archived in pond sediment. Nature, 450, 870–873. Drayton, J. M., and Jennions, M. D. (2011). Inbreeding and measures of immune function in the cricket Teleogryllus commodus. Behavioral Ecology, 22(3), 486– 492. Elzinga, J. A., Chevasco, V., Grapputo, A., and Mappes, J. (2011). Influence of male mating history on female reproductive success among monandrous Naryciinae (Lepidoptera: Psychidae). Ecological Entomology, 36(2), 170–180.

28 Engelmoer, D. J. P., and Rozen, D. E. (2011). Competence increases survival during stress in Streptococcus pneumoniae. Evolution, 65(12), 3475–3485. Fischer, O., and Schmid-Hempel, P. (2005). Selection by parasites may increase host recombination frequency. Biology Letters, 1(2), 193–195. Fisher, R. A. (1930). The genetical theory of natural selection: a complete variorum edition. Oxford University Press. Franke, K., and Fischer, K. (2013). Effects of inbreeding and temperature stress on life history and immune function in a butterfly. Journal of Evolutionary Biology, 26(3), 517–528. Gascoigne, J., Berec, L., Gregory, S., and Courchamp, F. (2009). Dangerously few liaisons: a review of mate-finding Allee effects. Population Ecology, 51(3), 355– 372. Gerber, N., and Kokko, N. (2016) Sexual conflict and the evolution of asexuality at low population densities. Proceedings of the Royal Society B: Biological Sciences, 283, 20161280. Gerloff, C. U., Ottmer, B. K., and Schmid Hempel, P. (2003). Effects of inbreeding on immune response and body size in a social insect, Bombus terrestris. Functional Ecology, 17(5), 582–589. Gerritsen, J. (1980). Sex and parthenogenesis in sparse populations. American Naturalist, 115(5), 718–742. Gershman, S. N., Barnett, C. A., Pettinger, A. M., Weddle, C. B., Hunt, J., and Sakaluk, S. K. (2010). Inbred decorated crickets exhibit higher measures of macroparasitic immunity than outbred individuals. Heredity, 105(3), 282–289. Goddard, M. R., Godfray, H. C. J., and Burt, A. (2005). Sex increases the efficacy of natural selection in experimental yeast populations. Nature, 434(7033), 636–640. Hadany, L., and Otto, S. P. (2009). Condition dependent sex and the rate of adaptation. The American Naturalist, 174(S1), S71–S78. Hamilton, W. D. (1980). Sex versus non-sex versus parasite. Oikos, 282–290. Hartfield, M., and Keightley, P. D. (2012). Current hypotheses for the evolution of sex and recombination. Integrative Zoology, 7(2), 192–209. Henry, L., Schwander, T., and Crespi, B. J. (2012). Deleterious mutation accumulation in asexual Timema stick insects. Molecular Biology and Evolution, 29(1), 401– 408. Hill, W. G., and Robertson, A. (1966). The effect of linkage on limits to artificial selection. Genetical Research, 8(3), 269–294. Hong, Z. D., and Ando, Y. (1998). Parthenogenesis in three species of genus Oxya. Japanese Journal of Applied Entomology and Zoology (Japan), (42), 65–69.

29 Jaenike, J. (1978). An hypothesis to account for the maintenance of sex within populations. Evol. Theory, 3, 191–194. Jokela, J., and Lively, C. M. (1995). Spatial variation in infection by digenetic trematodes in a population of freshwater snails (Potamopyrgus antipodarum). Oecologia, 103(4), 509–517. Kaltz, O., Bell, G., and Morgan, M. (2002). The ecology and genetics of fitness in Chlamydomonas. XII. Repeated sexual episodes increase rates of adaptation to novel environments. Evolution, 56(9), 1743–1753. Kearney, M. (2005). Hybridization, glaciation and geographical parthenogenesis. Trends in Ecology and Evolution, 20(9), 495–502. Killick, S. C., Carlsson, A. M., West, S. A., and Little, T. J. (2006). Testing the pluralist approach to sex: the influence of environment on synergistic interactions between mutation load and parasitism in Daphnia magna. Journal of Evolutionary Biology, 19(5), 1603–1611. King, K. C., Delph, L. F., Jokela, J., and Lively, C. M. (2009). The geographic mosaic of sex and the Red Queen. Current Biology, 19(17), 1438–1441. Kobayashi, K., and Miyaguni, Y. (2016). Facultative parthenogenesis in the Ryukyu drywood termite Neotermes koshunensis. Scientific Reports, 6. Koskella, B., and Lively, C. M. (2009). Evidence for negative frequency-dependent selection during experimental coevolution of a freshwater snail and a sterilizing trematode. Evolution, 63(9), 2213–2221. Kramer, M. G., and Templeton, A. R. (2001). Life-history changes that accompany the transition from sexual to parthenogenetic reproduction in Drosophila mercatorum. Evolution, 55(4), 748–761. Kumpulainen, T., Grapputo, A., and Mappes, J. (2004). Parasites and sexual reproduction in psychid moths. Evolution, 58(7), 1511–1520. Lehtonen, J., Jennions, M. D., and Kokko, H. (2012). The many costs of sex. Trends in Ecology and Evolution, 27(3), 172–178. Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by parasitism. Nature, 328(6130), 519–521. Lynch, M. (1984). The limits to life history evolution in Daphnia. Evolution, 38(3), 465–482. Lynch, M., and Deng, H.-W. (1994). Genetic slippage in response to sex. American Naturalist, 144(2), 242–261. Mantovani, B., Tinti, F., Bachmann, L., and Scali, V. (1997). The Bag320 satellite DNA family in Bacillus stick insects (Phasmatodea): different rates of molecular evolution of highly repetitive DNA in bisexual and parthenogenic taxa. Molecular Biology and Evolution, 14(12), 1197–1205.

30 Matsuura, K., Fujimoto, M., and Goka, K. (2004). Sexual and asexual colony foundation and the mechanism of facultative parthenogenesis in the termite Reticulitermes speratus (Isoptera, Rhinotermitidae). Insectes Sociaux, 51(4), 325– 332. Matsuura, K., and Nishida, T. (2001). Comparison of colony foundation success between sexual pairs and female asexual units in the termite Reticulitermes speratus (Isoptera: Rhinotermitidae). Population Ecology, 43(2), 119–124. Meirmans, S., Meirmans, P. G., and Kirkendall, L. R. (2012) The costs of sex: facing real-world complexities. The Quartely Review of Biology, 87, 19–40. Moore, J. (2002). Parasites and the behaviour of animals. Oxford University Press, Oxford, U.K. Morgan-Richards, M., Smissen, R. D., Shepherd, L. D., Wallis, G. P., Hayward, J. J., Chan, C., Chambers, G. K., and Chapman, H. M. (2009). A review of genetic analyses of hybridisation in New Zealand. Journal of the Royal Society of New Zealand, 39(1), 15–34. Morran, L. T., Parmenter, M. D., and Phillips, P. C. (2009). Mutation load and rapid adaptation favour outcrossing over self-fertilisation. Nature, 462, 350–352. Morran, L. T., Schmidt, O. G., Gelarden, I. A., Parrish, R. C., and Lively, C. M. (2011). Running with the Red Queen: Host-Parasite Coevolution Selects for Biparental Sex. Science, 333, 216-218. Muller, H. J. (1932). Some genetic aspects of sex. The American Naturalist, 66(703), 118–138. Muller, H. J. (1964). The relation of recombination to mutational advance. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 1(1), 2–9. Nei, M. (1967). Modification of linkage intensity by natural selection. Genetics, 57(3), 625. Neiman, M., Hehman, G., Miller, J. T., Logsdon, J. M., and Taylor, D. R. (2010). Accelerated mutation accumulation in asexual lineages of a freshwater snail. Molecular Biology and Evolution, 27(4), 954–963. Normark, B. B., and Moran, N. A. (2000). Testing for the accumulation of deleterious mutations in asexual eukaryote genomes using molecular sequences. Journal of Natural History, 34(9), 1719–1729. Otto, S. P., Nuismer, S. L. (2004). Species interactions andthe evolution of sex. Science 304, 1018–1020 Otto, S. P. (2009). The evolutionary enigma of sex. The American Naturalist, 174(S1), S1–S14. Paland, S., and Lynch, M. (2006). Transitions to asexuality result in excess amino acid substitutions. Science, 311(5763), 990–992.

31 Pearcy, M., Hardy, O., and Aron, S. (2006). Thelytokous parthenogenesis and its consequences on inbreeding in an ant. Heredity, 96(5), 377–382. Journal Article. Perez-Ruiz, M., Martinez-Rodriguez, P., Herranz, J., and Bella, J. L. (2015). A survey of Wolbachia, Spiroplasma and other bacteria in parthenogenetic and non- parthenogenetic phasmid (Phasmatodea) species. European Journal of Entomology, 112(3), 1–10. Rantala, M. J., and Roff, D. A. (2007). Inbreeding and extreme outbreeding cause sex differences in immune defence and life history traits in Epirrita autumnata. Heredity, 98(5), 329–336. Redfield, R. J. (1993). Evolution of natural transformation: testing the DNA repair hypothesis in Bacillus subtilis and Haemophilus influenzae. Genetics, 133(4), 755–761. Rhainds, M. (2010). Female mating failures in insects. Entomologia Experimentalis et Applicata, 136(3), 211–226. Rice, W. R., and Chippindale, A. K. (2001). Intersexual ontogenetic conflict. Journal of Evolutionary Biology, 14(5), 685–693. Roff, D. A. (2002). Inbreeding Depression: Tests of the Overdominance and Partial Dominance Hypotheses. Evolution, 56(4), 768–775. Roitt, I., J. Brostoff, and D. Male. 1996. Immunology. Mosby, Lon-don Scali, V., Passamonti, M., Marescalchi, O., And Mantovani, B. (2003). Linkage between sexual and asexual lineages: genome evolution in Bacillus stick insects. Biological Journal of the Linnean Society, 79(1), 137–150. Schneider, A., and Elgar, M. A. (2010). Facultative sex and reproductive strategies in response to male availability in the spiny stick insect, Extatosoma tiaratum. Australian Journal of Zoology, 58(4), 228–233. Schön, I., Martens, K., and Van Dijk, P. (2009). Lost sex: the evolutionary biology of parthenogenesis. Springer Science and Business Media. Schwander, T., and Crespi, B. J. (2009). Multiple direct transitions from sexual reproduction to apomictic parthenogenesis in Timema stick insects. Evolution, 63(1), 84–103. Schwander, T., Henry, L., and Crespi, B. J. (2011). Molecular Evidence for Ancient Asexuality in Timema Stick Insects. Current Biology, 21(13), 1129–1134. Schwander, T., Vuilleumier, S., Dubman, J., and Crespi, B. J. (2010). Positive feedback in the transition from sexual reproduction to parthenogenesis. Proceedings of the Royal Society B: Biological Sciences, 277(1686), 1435–1442. Simon, J.-C., Rispe, C., and Sunnucks, P. (2002). Ecology and evolution of sex in aphids. Trends in Ecology and Evolution, 17(1), 34–39.

32 Simon, J., Delmotte, F., Rispe, C., and Crease, T. (2003). Phylogenetic relationships between parthenogens and their sexual relatives: the possible routes to parthenogenesis in animals. Biological Journal of the Linnean Society, 79(1), 151–163. Smith, J. M., and Maynard-Smith, J. (1978). The evolution of sex. Book, Cambridge Univ Press. Stalker, H. D. (1956). On the evolution of parthenogenesis in Lonchoptera (Diptera). Evolution, 345–359. Stevens, L., Yan, G., and Pray, L. A. (1997). Consequences of inbreeding on invertebrate host susceptibility to parasitic infection. Evolution, 51(6), 2032– 2039. Stouthamer, R., Breeuwert, J. A., Luck, R. F., and Werren, J. H. (1993). Molecular identification of microorganisms associated with parthenogenesis. Nature, 361(6407), 66–8. Suomalainen, E., Saura, A., and Lokki, J. (1976). Evolution of parthenogenetic insects. In Evolutionary biology, 209–257. Springer. Templeton, A. ., Hamton, C. L., and Sing, C. F. (1976). The population genetics of parthenogenetic strains of Drosophila mercatorum. II. The capacity for parthenogenesis in a natural, bisexual population. Genetics, 82, 527–542. Väli, Ü., Einarsson, A., Waits, L., and Ellegren, H. (2008). To what extent do microsatellite markers reflect genome wide genetic diversity in natural populations? Molecular Ecology, 17(17), 3808–3817. Van Valen, L. (1977). The red queen. The American Naturalist, 111(980), 809–810. Vergara, D., Jokela, J., Lively, C. M. (2014). Infection dynamics in coexisting sexual and asexual host populations: support for the Red Queen hypothesis. American Naturalist. 184, 22–30. Vrijenhoek, R. C. (1998). Animal clones and diversity. Bioscience, 48, 617–628. Wedell, N., Gage, M. J. G., and Parker, G. A. (2002). Sperm competition, male prudence and sperm-limited females. Trends in Ecology and Evolution, 17(7), 313–320. Weeks, A. R., Velten, R., and Stouthamer, R. (2003). Incidence of a new sex-ratio- distorting endosymbiotic bacterium among arthropods. Proceedings of Biological Sciences, 270(1526), 1857–1865. Weismann, A. (1889). The significance of sexual reproduction in the theory of natural selection. Essays upon Heredity and Kindred Biological Problems, 1, 255–332. White, M. J. D. (1964). Cytogenetic mechanisms in insect reproduction. London: Royal Entomological Society of London.

33 Whiting, M. F., Bradler, S., and Maxwell, T. (2003). Loss and recovery of wings in stick insects. Nature, 421(6920), 264–267. Wolf, H. G., Wöhrmann, K., and Tomiuk, J. (1987). Experimental evidence for the adaptive value of sexual reproduction. Genetica, 72(2), 151–159. Wright, S. I., Lauga, B., and Charlesworth, D. (2002). Rates and patterns of molecular evolution in inbred and outbred Arabidopsis. Molecular Biology and Evolution, 19(9), 1407–1420. Zchori-Fein, E., Gottlieb, Y., Kelly, S. E., Brown, J. K., Wilson, J. M., Karr, T. L., and Hunter, M. S. (2001). A newly discovered bacterium associated with parthenogenesis and a change in host selection behavior in parasitoid wasps. Proceedings of the National Academy of Sciences of the United States of America, 98(22), 12555–60. Zeyl, C., and Bell, G. (1997). The advantage of sex in evolving yeast populations. Nature, 388(6641), 465–468.

34

Chapter 2

Variation in male mating success in a facultatively

parthenogenic insect (Extatosoma tiaratum)

Published as:

Alavi, Y., Elgar, A. M., and Jones, M. T. (2016). Male Mating Success and the Effect of Mating History on Ejaculate Traits in a Facultatively

Parthenogenic Insect (Extatosoma tiaratum). Ethology, 122: 1-8.

35 Summary

Males can typically increase their lifetime reproductive success by mating with multiple females. However, recent studies across a broad range of species have demonstrated physiological constraints on male multiple mating. In this study, we investigate male mating capacity in Extatosoma tiaratum, a facultative parthenogenetic phasmatid. Sperm limitation is thought to be one factor favouring the evolution and maintenance of parthenogenetic reproduction, but studies on male mating ability in facultative parthenogenetic species are extremely rare. To explore whether male reproductive success varies across matings, we provided males with weekly mating opportunities with different females throughout their lives. We then observed mating success, and the variation in ejaculate size and quality within each mating. We showed that most, but not all, males can mate multiply, however the amount of ejaculate produced is variable and depends upon male body mass and mating history.

36 Introduction

Traditionally, mating was assumed to be inexpensive for males, due to the relative small size of sperm compared with the larger eggs produced by females (Trivers

1972). Accordingly, males were assumed to have an unlimited supply of sperm allowing them to mate with all available fertile females (Bateman 1948). Moreover, studies documenting male multiple mating and even extreme male mating frequencies for a range of taxa were relatively common (Andersson 1994). However, mounting evidence demonstrates that both sperm production and copulation are energetically costly, and thus male reproductive potential may be physiologically constrained

(Dewsbury 1982, Wedell et al 2002, Scharf et al, 2013). Thus, while in theory selection should favor high male mating frequency, the reality is that this may not always be physiologically possible, even in the absence of female discrimination

(Dewsbury 1982).

Empirical evidence for physiological constraints on male mating, in terms of declines in ejaculate size and quality or reduced male lifespan has been documented for both vertebrates and invertebrates (vertebrates: e.g. Huber et al 1980, Nakatsuru and

Kramer 1982, Warner et al 1995, Preston et al 2001, invertebrates: e.g. Christenson

1989, Van Voorhies 1992, Rigaud and Moreau 2004). In the insects, sperm depletion following multiple mating is taxonomically widespread and both the quantity of sperm

(Watanabe et al 1998, Torres-Vila and Jennions 2005, Marcotte et al 2005, Wedell

2010, Elzinga et al 2011, Paukku and Kotiaho 2005, Rönn et al 2008, Damiens et al

37 2002, Damiens and Boivin 2005, Partridge and Farquhar 1981, Pitnick 1993, Jones

2001, Oliver and Cordero 2009) and quality of sperm (Dowling and Simmons 2012) may decline across successive matings. In cases with moderate or no decrease in ejaculate production, male mating history may affect other aspects of male reproductive success, such as longevity or mating frequency (e.g. Oliver and Cordero

2009, Lewis et al 2011, Salehialavi et al 2011).

Facultative parthenogenesis, in which females have the ability to reproduce both sexually and, if they remain unmated, parthenogenetically, is taxonomically widespread (Kramer and Templeton 2001, Matsuura et al 2004, Booth et al 2012), and unusually common among some insect orders, such as the Phasmatodea (Maynard

Smith 1978). The adaptive significance of facultative parthenogenesis is poorly understood, but it has been proposed as a potential adaptation to male- or sperm- limitation (Matsuura and Nishida 2010, Schneider and Elgar 2010, Schwander et al

2010, Elzinga et al 2011). However, few, if any, studies of facultatively parthenogenetic species have investigated either the degree of male multiple mating or the capacity of males to produce sperm at each mating attempt, yet such data are pivotal for understanding the evolution and maintenance of parthenogenetic reproduction.

In this study, we investigated male multiple mating in Macleay’s Spectre, Extatosoma tiaratum, a facultatively parthenogenetic Australian insect (Phasmatodea). Extatosoma tiaratum occurs in the rainforests of tropical and subtropical Queensland and northern

38 New South Wales (Gurney 1947). Females oviposit continuously throughout their adult lifespan, and un-fertilised eggs develop into female offspring (Carlberg 1983).

Parthenogenetic reproduction appears to be a strategic response to the absence of males, since females delay oviposition when reared with males at the juvenile stages

(Schneider and Elgar 2010). Copulations typically last up to 19 hours, during which time a male attaches a spermatophore to the female’s terminal genital segment

(Carlberg 1983). To explore male multiple mating, we provided males with a mating opportunity each week, and asked whether the variation in ejaculate size and quality depended upon male mating history. We expected males to mate with multiple females, as male biased sex ratios, a condition required for the evolution of monogyny

(male monogamy), is not expected in facultative parthenogenetic populations

(Fromhage et al 2005). However, in line with previous data on other invertebrates, including insects, we also expected that males would be limited in the amount of sperm they could transfer and thus we predicted declines in ejaculate characteristics over successive matings.

39 Methods

Experimental animals

A stock population was established from individuals obtained from various insect breeders across the Melbourne region (Victoria, Australia). Males and females were maintained in mesh cages (46 × 46 × 91 cm3) in one of two climate-controlled laboratories under identical conditions (24-26° C; 50% humidity; 12:12hr light:dark cycle). Males and females were reared in different laboratories to ensure there were no pheromone-derived influences on reproductive behaviour that could affect the current mating environment (see Schneider and Elgar 2010). All individuals were provided with the leaves of various species of Eucalyptus, ad libitum, which were lightly sprayed with water daily and replaced regularly as required. Twenty final instar male offspring were selected from the stock population. Experimental males were maintained until approximately four weeks after moulting to adulthood. Following their final moult, adult males were uniquely marked with a drop of non-toxic acrylic paint on their left tarsus and kept in a single mesh cage.

Experimental design

Two weeks after reaching adulthood, each of the 20 adult males was allocated a weekly mating opportunity (every 5-7 days) for the duration of his life. At each mating opportunity, a male was transferred to a mating cage containing at least ten sexually

40 mature females (virgin females were added to mating cages weekly to maximize female receptivity). This experimental protocol ensured that males had the opportunity to find preferred mates and intra-sexual competition between males did not interfere with the opportunity to mate. The cages were monitored five hours after dark and males were removed if they had not mounted any females. Copulating pairs were checked every 30 minutes for a successful mating (defined as spermatophore transfer), and the externally transferred spermatophore was removed from the female’s genitalia using fine forceps once it was fully produced, but prior to any sperm transfer. Each spermatophore was weighed (to the nearest 0.0001g) and digital images of each spermatophore were obtained from three different perspectives. The diameter was measured using ImageJ (1.46r) software, and the volume was estimated by assuming a spherical shape. Each copulating pair was kept in a separate container until the male detached. Males were weighed before and after mating, and tarsus length was also measured as an estimate of body size.

Preparation of the sperm solution

The sperm solution was prepared by first cutting the neck of the spermatophore using micro-scissors. The spermatophore was transferred into a 1.5 mL microcentrifuge tube containing 80 μl of 0.04 Beadle saline (128.3 mM NaCl, 4.7 mM KCl, and 23 mM

CaCl2) and squeezed gently before being left for one hour to ensure complete transfer of sperm into Beadle solution. The solution was then gently mixed and sperm density and viability was measured via two separate methods.

41 Sperm density assay

A total of 1 μl of the sperm solution was pipetted into a 200 μl microcentrifuge tube and diluted 1:100 in distilled water. 10 μl of the diluted sperm solution was pipetted on the well of a haemocytometer. Sperm were visualized using light microscopy ® Leica

DM 2500 (Leica Microsystems GmbH, Wetzlar, Germany), at 200× magnification, with all sperm within five predetermined grid squares counted. Sperm density was calculated by multiplying mean haemocytometer count by its dilution factor to calculate sperm density. Total sperm count was calculated as the product of spermatophore volume and sperm density.

Sperm viability assay

We used the ® LIVE/DEAD assay (Molecular Probes, Sigma, Australia, L-7011) to estimate sperm viability (see Damiens et al 2002). We pipetted 5 μl of the diluted sperm onto a glass slide and added 10 μl of 1:50 diluted 1mM SYBR14. The slide was incubated at room temperature in the dark for ten minutes before adding 2 μl of 2.4 mM Propidium Iodide followed by an additional 10 minute incubation. Samples were observed under fluorescent microscopy 30 minutes after staining (blue excitation filter at λ= 490 nm; 20× magnification). Ten images were taken from different field views

(200× magnification) on each slide and the proportion of live to dead sperm was quantified. At least 500 live spermatozoa were counted per sample.

42 Statistical analyses

We used general linear mixed models (GLMM including male ID as a random effect) in JMP version 12 to examine the effect of mating number on ejaculate size

(spermatophore mass and sperm density). To investigate the effect of mating number on sperm viability, we used the non-parametric Wilcoxon test weighted by the total number of sperm counted per sample. To remove the potential problem of the first mating interval being recorded as zero, and thus biasing models where we specifically needed to include the first interval, we added seven days (the minimum number of days we permitted a male to rest between mating opportunities) to the time until successful production of a spermatophore following the first mating opportunity. Thus, if a male mated and transferred a spermatophore on his first mating opportunity, his first mating interval would be recorded as 7; however if he failed to transfer a spermatophore on this attempt and mated on his subsequent attempt (approximately seven days later) his first mating interval would be recorded as 7 + 7 = 14 days. For all subsequent mating intervals, the actual number of days between the current and the previous spermatophore produced was taken as the mating interval. Spermatophores that were not removed immediately after production were excluded from analyses of sperm quantity and quality (11 spermatophores from the first mating and one from the third mating). As few males mated more than five times, spermatophores from the fifth and any subsequently matings were pooled for the spermatophore mass and sperm density analyses. For sperm viability analysis we excluded 6+ spermatophores, as the models with and without these spermatophores were similar and we were unable to include male ID as a random effect in the nonparametric model. Where possible, we

43 included mating interval (the total number of days between two consecutive matings), male body mass before each mating, male age and tarsus length as covariates in all models; terms were dropped where P > 0.10. Data were transformed where necessary to improve normality. Unless otherwise stated, all presented averages are means ± standard errors.

44 Results

Male mating history and survival

During their adult lifespan of 16.6 ± 3.7 weeks, males mated on average 4 ± 0.4 times

(minimum = 1 mating, maximum = 8 matings, median = 4 matings, interquartile range

= 3.2 matings). Nine of twenty virgin males mated at their first mating opportunity and the average age at first mating was 23.9 ± 1.8 days post final moult (minimum = 14 days, maximum = 48 days, median = 22 days, interquartile range = 8 days). On average, males lost 6.4 ± 0.7 % of their body mass during copulation and transferred a spermatophore that was roughly 1.4 ± 0.04 % of their body mass. Male body size

(tarsus length) did not influence either the total number of spermatophores produced

(GLM with Poisson error distribution and log-link: F1, 17 = 0.20, P = 0.66), or the age at first mating (F1, 17 = 0.95, P = 0.34). Average male mating interval was 16.2 ± 1.2 days (minimum = 7 days, maximum = 48 days, median = 13 days, interquartile range =

14 days). Proportional hazards survival analysis revealed that male survival was

2 comparable for all males regardless of the number of spermatophores produced (χ 1, 16

2 = 0.44, P = 0.50), their average body mass (χ 1, 16 = 0.16, P = 0.69), and the average

2 mass of spermatophores produced (χ 1, 16 = 0.6, P = 0.44). Male identity (random effect) did not account for more than 14% of the explained variation in any of our models.

45 Effects of multiple mating on ejaculate size and quality

Spermatophore mass –Spermatophore mass varied significantly with spermatophore number (Table 2-1a). Post-hoc Tukey’s tests revealed that, on average, the mass of spermatophores 1-3 were comparable but that spermatophore mass declined between the third and subsequent spermatophores (Figure 2-1a). Spermatophore mass was positively related to male body mass (Figure 2-2a), male mating interval (Figure 2-2b) and male age (Table 2-1a).

Table 2-1 The effect of male body mass, male age, mating number and mating interval on a) spermatophore mass, b) sperm density (Ln transformed) and c) viable sperm proportion across males

Model Parameters β ± SE DF Statistic Probability

a) Spermatophore mass

Male body mass 0.006 ± 0.002 1 F= 11.10 0.002 Male age 9.13 ± 4.12 × 10-5 1 F= 4.90 0.03 Spermatophore number 4 F= 10.40 <0.0001

Mating interval 0.0001± 4.43 × 10-5 1 F= 18.79 <0.0001

b) Sperm density

Spermatophore number 4 F= 4.96 0.002 Mating interval 0.02 ± 0.007 1 F= 6.51 0.01

c) Sperm viability (individual non-parametric models)

Male body mass n = 59 r = 0.09 0.50 s Male age n = 59 r = 0.27 0.04 s Spermatophore number n = 59 χ 2 = 7.29 0.12 4 Mating interval n = 59 r = -0.14 0.28 s

46 a) n= 9 n= 19 n= 13 n= 11 n= 16

BC A AB C C

b)

AB A ABC BC C

c)

Figure 2-1 The relationship between spermatophore number and ejaculate traits; a) spermatophore mass, b) sperm density (Ln transformed), and c) viable sperm proportion; Levels not connected by same letters are significantly different based on post hoc Tukey’s tests.

47 a)

b)

c)

Figure 2-2 The relationship between a) spermatophore mass and body mass (r2 = 0.19, P= 0.0002), b) spermatophore mass and mating interval (r2 = 0.45, P <0.0001) c) sperm density (Ln transformed) and mating interval (r2 = 0.24, P <0.0001)

48 Sperm density – Sperm density varied significantly with spermatophore number and was positively correlated with mating interval (Table 2-1b, Figure 2-2c). Post hoc

Tukey’s tests revealed sperm density was highest in the first and second spermatophores produced and lowest in the fifth or more spermatophores (Figure 2-

1b). We repeated the above analysis using total sperm count (sperm density multiplied by spermatophore volume) and achieved qualitatively similar results (results not presented).

Sperm viability – Overall, the proportion of viable sperm was relatively high (Median

= 0.96, interquartile range = 0.21, N = 20 males and 4 ± 4.4 spermatophores produced per male). The proportion of viable sperm was not correlated with spermatophore number (Figure 2-1c), male body mass, or mating interval (Table 2-1c). However, the proportion of viable sperm increased with male age (Table 2-1c).

49 Discussion

Our results showed that while males of E. tiaratum are capable of multiple mating, they did not mate with every opportunity: males typically mated in less than 30% of the weekly mating trials in which they were provided with access to females. The data also suggest that males are limited in the amount of ejaculate they are able to invest over successive matings: both spermatophore mass and sperm density decreased with increasing number of matings.

Our results add to the growing body of evidence for male physiological constraints on ejaculate production. Firstly, there was variation in the amount of ejaculate transferred across matings. Spermatophore size was the largest in the second mating and declined in the final mating. Sperm density was the highest at the first and second mating and again significantly declined in later matings. Secondly, the positive relationship between both spermatophore mass and sperm density, with mating interval suggests that males need time to replenish their sperm reserves following mating. Contrary to the results for ejaculate quantity, we found no evidence for reduction in ejaculate quality: sperm viability was comparable across all spermatophores produced. The latter result is perhaps unsurprising as selection on sperm viability is often strong in polyandrous insects (Hunter and Birkhead 2002, Garcı́a-González and Simmons

2005). This lack of a difference may have arisen due to small sample sizes, but it is also possible that other components of ejaculate quality such as sperm size and

50 velocity or any trade-offs between sperm traits (not measured here), were affected by male mating history.

Our data do not fully support the traditional view of male reproductive success (sensu

Bateman 1948): E. tiaratum males were limited in both the number of matings achieved and the amount of ejaculate provided to each mate. The former result is particularly interesting given that males were provided with ten females simultaneously. We suggest it is unlikely that a lack of available females could explain the observed low mating frequency. We instead suggest two mutually non-exclusive alternatives: either, all ten females found the male unattractive at a given mating opportunity and rejected his attempt, or males were physiologically constrained and thus unable to produce an ejaculate and/or copulate successfully. While we are unable to discount either mechanism entirely, the second explanation seems the more parsimonious, given the positive correlations between mating interval with spermatophore mass and sperm density, and the fact that more than half of the virgin males did not mate when first presented with a mating opportunity. However, we note that the females may also have discriminated against sperm-depleted males and thus avoided them. While E. tiaratum females cease sexual signalling (releasing a sex pheromone) once they commence egg production (Schenider and Elgar 2010), female receptivity does not appear to be directly correlated with sexual signalling because ovipositing females will mate if they encounter a male. Accordingly, we do not expect ovipositing to influence male mating frequency.

51 Lifetime male mating success was not related to male body size, suggesting that females do not discriminate between males according to their size, a common sexually selected trait (Jennions et al 2001). However, males appear to be sexually immature following their final moult to adulthood. Few males mated in their first attempt, and the average age at first mating was 23.9 ± 1.8 days post final moult. Whether this latency period prior to the first mating is a product of the need to acquire somatic resources, or to ensure sperm maturation following sexual maturity at the adult stage of the lifecycle is unknown. However, males typically lost 6.4% of their body mass following each mating, highlighting that mating is physically costly for males both directly in terms of investment in the mating act and spermatophores themselves, and indirectly through lost foraging opportunities (males do not feed during copulation, which lasts up to 19 hours). This may have significant consequences for E. tiaratum males in particular, as their natural diet predominantly comprises Eucalyptus leaves with low nutritional value (Moore et al 2004).

Sperm limitation may be a more significant factor favouring the maintenance of parthenogenesis than is generally appreciated, since mating with sperm depleted males can influence female reproductive success (Wedell and Ritchie 2004, Jones et al 2006,

Lauwers and Van Dyck 2006, Elzinga et al 2011). Males of E. tiaratum are capable of multiple mating, consistent with the likely female-biased populations in facultative parthenogenetic species. However, male physiological constraints may affect female fertilisation success either because males are unwilling to mate, or they transfer insufficient sperm. Such constraints are likely more important in populations with low mate encounter rates (due to stochastic changes in population densities or

52 environmental factors, Gascoigne et al 2009). Theoretically, parthenogenetically produced offspring may have reduced fitness compared with sexually produced offspring (Maynard Smith 1986, Kondrashov 1988), however parthenogenesis may rescue maternal fitness if the alternative is mating failure. By producing female offspring through parthenogenesis, females will increase their fitness, especially if some of their daughters can find a mate and reproduce sexually. Future research might be profitably aimed at investigating potential links between sperm limitation and parthenogenetic reproduction by studying egg fertilisation patterns and mate encounter rates in natural, ecologically varied, populations. In addition, there may be energetic benefits to females from receiving a spermatophore, even though this is likely negligible given the relatively small size of the spermatophore.

53 References

Andersson, M. B. (1994). Sexual selection. Princeton Univ. Press. Princeton, NJ. Bateman, A. J. (1948). Intra-sexual selection in Drosophila. Heredity, 2(3), 349–368. Booth, W., Smith, C. F., Eskridge, P. H., Hoss, S. K., Mendelson, J. R., and Schuett, G.W. (2012). Facultative parthenogenesis discovered in wild vertebrates. Biol. Lett. 8(6), 983–985. Carlberg, U. (1983). Copulation in Extatosoma tiaratum (MacLeay)(Insecta: Phasmida). Zool. Anz. 5(6), 340-356 Christenson, T. (1989). Sperm depletion in the orb-weaving spider Nephila clavipes (Araneae, Araneidae). J. Arachnol. 115–118. Damiens, D., Bressac, C., Brillard, J. P., and Chevrier, C. (2002). Qualitative aspects of sperm stock in males and females from Eupelmus orientalis and Dinarmus basalis (Hymenoptera : Chalcidoidea) as revealed by dual fluorescence. Physiol. Entomol. 27(2), 97–102. Damiens, D., and Boivin, G. (2005). Male reproductive strategy in Trichogramma evanescens: sperm production and allocation to females. Physiol. Entomol. 30(3), 241–247. Dewsbury, D. A. (1982). Ejaculate cost and male choice. Amer. Nat. 119(5), 601–610. Dowling, D. K., and Simmons, L.W. (2012). Ejaculate economics: testing the effects of male sexual history on the trade-off between sperm and immune function in Australian crickets. PLoS One. 7: e30172. Elzinga, J. A. Chevasco, V., Grapputo, A., and Mappes, J. (2011). Influence of male mating history on female reproductive success among monandrous Naryciinae (Lepidoptera: Psychidae). Ecol. Entomol. 36(2), 170–180. Fromhage, L., Elgar, M. A., and Schneider, J. M. (2005). Faitful without care: the evolution of monogamy. Evolution. 59(7), 1400–1405.

Garcı́a-González, F., and Simmons, L. W. (2005). Sperm viability matters in insect sperm competition. Curr. Biol. 15(3), 271–275. Gascoigne, J., Berec, L., Gregory, S., and Courchamp, F. (2009). Dangerously few liaisons: a review of mate-finding Allee effects. Popular Ecol. 51, 355–372. Gurney, A. B. (1947). Notes on some remarkable Australasian walking sticks, including a synopsis of the genus Extatosoma (Orthoptera: Phasmatidae). Ann. Entomol. Soc. Am. 40(3), 373–396. Gwynne, D. T. (2008). Sexual conflict over nuptial gifts in insects. Annu. Rev.

54 Entomol. 53, 83–101. Huber, M. H. R., Bronson, F. H., and Desjardins, C. (1980). Sexual activity of aged male mice: correlation with level of arousal, physical endurance, pathological status, and ejaculatory capacity. Biol. Rep. 23(2), 305–316. Hunter, F. M., and Birkhead, T. R. (2002). Sperm viability and sperm competition in insects. Curr. Biol. 12(2), 121–123. Jennions, M. D., Moller, A. P., and Marion, P. (2001). Sexually Selected Traits and Adult Survival: A Meta-Analysis. Q. Rev. Biol. 76(1), 3–36. Jones, T. M. (2001). A potential cost of monandry in the lekking sandfly Lutzomyia longipalpis. J. Insect. Behav. 14(3), 385–399. Jones, T. M., McNamara, K. B., Colvin, P. G. R, Featherson, R., and Elgar, M. A. (2006). Mating frequency, fecundity and fertilisation success in the hide beetle, Dermestes maculatus. J. Insect. Behav. 19(3), 357–371. Kondrashov, A. S. (1988). Deleterious mutations and the evolution of sexual reproduction. Nature. 336, 435-440. Kramer, M. G., and Templeton, A. R. (2001). Life history changes that accompany the transition from sexual to parthenogenetic reproduction in Drosophila Mercatorum. Evolution. 55(4), 748–761. Lauwers, K., and Van Dyck, H. (2006). The cost of mating with a non-virgin male in a monandrous butterfly: experimental evidence from the speckled wood, Pararge aegeria. Behav. Ecol. Sociobiol. 60(1), 69–76. Lewis, Z., Sasaki, H., and Miyatake, T. 2011: Sex starved: do resource-limited males ensure fertilisation success at the expense of precopulatory mating success? Anim. Behav. 81(3), 579–583. Marcotte, M., Delisle, J., and McNeil, J.N. (2005). Impact of male mating history on the temporal sperm dynamics of Choristoneura rosaceana and C. fumiferana females. J. Insect. Physiol. 51(5), 537–544. Matsuura, K., Fujimoto, M., and Goka, K. (2004). Sexual and asexual colony foundation and the mechanism of facultative parthenogenesis in the termite Reticulitermes speratus (Isoptera, Rhinotermitidae). Insectes Sociaux. 51(4), 325– 332. Matsuura, K., and Nishida, T. (2001). Comparison of colony foundation success between sexual pairs and female asexual units in the termite Reticulitermes speratus (Isoptera: Rhinotermitidae). Pop. Ecol. 43(2), 119–124. Maynard Smith, J. (1986). Contemplating life without sex. Nature. 324, 300-301. Maynard Smith, J. (1978). The evolution of sex, New York: Cambridge Univ. Press. Moore, B. D., Wallis, I. R., Wood, J. T., and Foley, W. J. (2004). Foliar nutrition, site

55 quality, and temperature influence foliar chemistry of tallowwood (Eucalyptus microcorys). Ecol. Monographs. 74(4), 553–568. Nakatsuru, K., and Kramer, D. L. (1982). Is sperm cheap? Limited male fertility and female choice in the lemon tetra (Pisces, Characidae). Science. 216(4547), 753– 755. Oliver, C., and Cordero, C. (2009). Multiple mating reduces male survivorship but not ejaculate size in the polygamous insect Stenomacra marginella (Heteroptera: Largidae). Evol. Ecol. 23(3), 417–424. Partridge, L., and Farquhar, M. (1981). Sexual activity reduces lifespan of male fruitflies. Nature. 294, 580-582. Paukku, S., and Kotiaho, J. S. (2005). Cost of reproduction in Callosobruchus maculatus: effects of mating on male longevity and the effect of male mating status on female longevity. J. Insect. Physiol. 51(11), 1220–1226. Pitnick, S. (1993). Operational sex ratios and sperm limitation in populations of Drosophila pachea. Behav. Ecol. Sociobiol. 33(6), 383–391. Preston, B. T, Stevenson, I. R., Pemberton, J. M., and Wilson, K. (2001). Dominant rams lose out by sperm depletion. Nature. 409, 681–682. Rigaud, T., and Moreau, J. (2004). A cost of Wolbachia–induced sex reversal and female–biased sex ratios: decrease in female fertility after sperm depletion in a terrestrial isopod. Proc. R. Soc. Lond. Ser. B. Biol. Sci. 271, 1941–1946. Rönn, J. L., Katvala, M., and Arnqvist, G. (2008). Interspecific variation in ejaculate allocation and associated effects on female fitness in seed beetles. J. Evol. Biol. 21(2), 461–470. Salehialavi, Y., Fritzsche, K., and Arnqvist, G. (2011). The cost of mating and mutual mate choice in 2 role–reversed honey locust beetles. Behav. Ecol. 22(5), 1104– 1113. Scharf, I., Peter, F., and Martin, O.Y. (2013). Reproductive trade-offs and direct costs for males in arthropods. Evol. Biol. 40(2), 169–184. Schneider, A., and Elgar, M.A. (2010). Facultative sex and reproductive strategies in response to male availability in the spiny stick insect, Extatosoma tiaratum. Aust. J. Zool. 58(4), 228–233. Schwander, T. Vuilleumier, S., Dubman, J., and Crespi, B. J. (2010). Positive feedback in the transition from sexual reproduction to parthenogenesis. Proc. R. Soc. Lond. Ser. B. Biol. Sci. 277, 1435–1442. Torres-Vila, L. M., and Jennions, M. D. (2005). Male mating history and female fecundity in the Lepidoptera: do male virgins make better partners? Behav. Ecol. Sociobiol. 57(4), 318–326. Trivers, R. (1972). Parental investment and sexual selection. Sexual Selection and the

56 Descent of Man, New York: Aldine de Gruyter. Van Voorhies, W. A. (1992). Production of sperm reduces nematode lifespan. Nature 360, 456–458 Warner, R. R., Shapiro, D. Y., Marcanato, A., and Petersen, C. W. (1995). Sexual conflict: males with highest mating success convey the lowest fertilisation benefits to females. Proc. R. Soc. Lond. Ser. B. Biol. Sci. 262, 135–139. Watanabe, M., Wiklund, C., and Bon’no, M. (1998). The effect of repeated matings on sperm numbers in successive ejaculates of the cabbage white butterfly Pieris rapae (Lepidoptera: Pieridae). J. Insect. Behav. 11(4), 559–570. Wedell, N. (2010). Variation in male courtship costs in butterflies. Behav. Ecol. Sociobiol. 64(9), 1385–1391. Wedell, N., Gage, M. J. G., and Parker, G. A. (2002). Sperm competition, male prudence and sperm-limited females. Trends. Ecol. Evol. 17(7), 313–320. Wedell, N., and Ritchie, M. G. (2004). Male age, mating status and nuptial gift quality in a bushcricket. Anim. Behav. 67(6), 1059–1065

57

Chapter 3

Novel microsatellite markers suggest the

mechanism of parthenogenesis in Extatosoma

tiaratum, is automixis with terminal fusion

Published as:

Alavi, Y., van Rooyen, A., Elgar, M. A., Jones, M. T., and Weeks, A. R.

(2016). Novel microsatellite markers suggest the mechanism of parthenogenesis in Extatosoma tiaratum is automixis with terminal fusion.

Insect Science. In press

58 Summary

Parthenogenetic reproduction is taxonomically widespread and occurs through various cytological mechanisms, which have different impacts on the genetic variation of the offspring. Extatosoma tiaratum is a facultatively parthenogenetic Australian insect

(Phasmatodea), in which females oviposit continuously throughout their adult lifespan irrespective of mating. Fertilised eggs produce sons and daughters through sexual reproduction and unfertilised eggs produce female offspring via parthenogenesis.

Here, we developed novel microsatellite markers for E. tiaratum and characterised them by genotyping individuals from a natural population. We then used the microsatellite markers to infer the cytological mechanism of parthenogenesis in this species. We found evidence suggesting parthenogenesis in E. tiaratum occurs through automixis with terminal fusion, resulting in substantial loss of microsatellite heterozygosity in the offspring. Loss of microsatellite heterozygosity may be associated with loss of heterozygosity in fitness related loci. The mechanism of parthenogenetic reproduction can therefore affect fitness outcomes and needs to be considered when comparing costs and benefits of sex versus parthenogenesis.

59 Introduction

Thelytokous parthenogenesis, the production of females from unfertilised eggs, is widespread among insects, including Lepidoptera, Orthoptera, Hemiptera, Coleoptera and Hymenoptera (Simon et al 2003). Parthenogenetic species are often derived from close sexual ancestors and can be obligate, cyclic or facultative parthenogens (Simon et al 2003). Thelytokous parthenogenetic females are often referred to as an evolutionary paradox because, all else being equal, they can reproduce twice as fast as sexual females by producing female only offspring; yet they are often out-competed by their sexual counterparts (Williams 1975, Smith and Maynard-Smith 1978, Bell 1982).

Facultatively parthenogenetic species, which can reproduce parthenogenetically when unmated, provide an opportunity to compare directly the consequences of sex and parthenogenesis. The cytological mechanism of parthenogenesis is an important component of this comparison, because each mechanism has a different outcome on the individual-level heterozygosity, which has important implications for individual fitness and genetic diversity at the population-level.

The two most common cytological mechanisms of parthenogenesis are apomixis and automixis, each resulting in different levels of heterozygosity (Suomalainen et al

1987). Apomictic parthenogenesis occurs through mitosis and the resulting offspring are identical clones of their mother (Suomalainen et al 1987). This mechanism involves no recombination and thus the ratio of heterozygote to homozygote loci in the offspring is identical to that of their mother (Pearcy et al 2006). In contrast, meiosis

60 takes place in automictic parthenogenesis and diploidy is restored through one of the following mechanisms: gamete duplication, central fusion (fusion of non-sister nuclei), terminal fusion (fusion of the resulting sister nuclei), or random fusion (Suomalainen et al 1987, Pearcy et al 2006). Automixis with central fusion has similar outcomes to apomixis as heterozygosity is restored in most loci, except for those loci far from the centromere, which have high recombination rates. In automixis with gamete duplication, heterozygosity is entirely lost, resulting in homozygosity at all loci

(Suomalainen et al 1987, Pearcy et al 2006). In terminal fusion, loci close to the centromere may be heterozygous as a result of crossing over between the chiasma and telomere (Rabeling and Kronauer 2013). Changes in heterozygosity levels can directly affect fitness (see Chapman et al 2009), so it is important to determine the mechanism of parthenogenesis when comparing it with sexual reproduction.

Polymorphic nuclear genetic markers such as microsatellite loci, can be used to determine the cytological mechanism of parthenogenesis by estimating the rate of transition (R) of heterozygous loci in mothers to homozygous loci in their offspring for each locus (Pearcy et al 2006). If parthenogenesis is derived through apomixis, no heterozygous loci are expected to change to the homozygous state and R = 0. In contrast, with gamete duplication R is always expected to be one (complete homozygosity), while with automixis the expected value of R varies depending on the position of each locus relative to the centromere. For loci closer to the centromere

(with lower recombination rates), a higher transition rate to homozygosity is expected

(see Pearcy et al 2006). Conversely, loci located far away from the centromere have lower transition rates due to recombination. Accordingly, the range of the expected

61 rates of transition to homozygosity (r) is between 0 and 1/3 for central fusion, 1/3 for random fusion and between 1/3 and 1 for terminal fusion (Pearcy et al 2006).

Here we investigate the mechanism of parthenogenetic reproduction in a facultatively parthenogenetic insect, Extatosoma tiaratum (Phasmatodea). Extatosoma tiaratum is native to the rainforests of tropical and subtropical Queensland and northern New

South Wales, Australia (Gurney 1947). Females commence ovipositing irrespective of mating (Schneider and Elgar 2010), and continue to oviposit throughout their adult phase, producing viable sons and daughters via fertilised eggs, and only daughters from unfertilised eggs by parthenogenesis (Carlberg 1983). We isolated 18 nuclear microsatellite markers for E. tiaratum and investigated their patterns of bi-allelic inheritance in progeny from sexual matings, as well as inheritance patterns in parthenogenetically produced progeny, to determine the cytological mechanism of parthenogenesis. In arthropods, bacterial endosymbiotic infection (most commonly by

Wolbachia) may be responsible for parthenogenetic reproduction (resulting in gamete duplication, Stouthamer et al 1999). However, we did not screen for Wolbachia infection in this study, as this bacteria is unlikely to be responsible for parthenogenesis in the Phasmatodea order and has never been associated with parthenogenesis in facultatively parthenogenetic species (Perez-ruiz et al 2015).

62 Methods

Laboratory stock population

A stock population of E. tiaratum was derived from eggs and juveniles obtained from existing cultures maintained by the Melbourne Zoological gardens and the Melbourne

Museum (both in Victoria, Australia), however the exact origin of these populations was unknown. All insects were housed in cylindrical containers (height: 23 cm; diameter: 25 cm), enclosed by a fine mesh. The cages were kept in climate-controlled conditions (24-26°C; 50% humidity; 12:12hr Light:Dark cycle). Males and females were housed separately. All eggs were incubated under climate-controlled conditions in plastic containers (37 × 36 × 30cm) containing moist sand (sprayed with water weekly). First instar nymphs were collected after hatching and stored in 100% Ethanol in -20 °C for later DNA extraction.

Field collection

We collected E. tiaratum from a natural population at Crystal Cascades, north

Queensland, Australia (16°57′42″S 145°40′46E) in December 2014 over three nights.

The insects were collected manually from the lower accessible branches of trees along a 2 km trail. The insects (N = 6 females, 5 males) were all at the juvenile stages (1st-5th instar) at the time of collection. Extatosoma tiaratum is nocturnal and inconspicuous

(see Bian et al 2016), making field collections challenging. Nevertheless, as the

63 laboratory stock was of unknown origin, we used wild, outbred insects to characterize our microsatellite markers. Following collection, the insects were moved to the laboratory and housed individually under similar conditions to the laboratory population before DNA was extracted at the adult stage.

Next-generation sequencing and de novo genome assembly

DNA was extracted from limb tissue (legs) of a single first instar individual (1-7 days old) from the stock population, using a QIAGEN DNA Easy kit (Qiagen), following the manufacturers recommendations. The 454 next generation sequencing platform

(Schuster 2008) was used to identify microsatellite loci for E. tiaratum. Approximately

10 µg genomic DNA was extracted from limb tissue as above. The DNA was then nebulized, ligated with 454 sequencing primers and tagged with a unique oligo sequence allowing sequences to be separated from pooled species DNA sequences using post-run bioinformatic tools. The sample was then analysed using high throughput DNA sequencing on 1/16 of a 70 X 75 mm PicoTiterPlate using the Roche

GS FLX (454) system (Margulies et al 2005) at the Australian Genome Research

Facility (AGRF, Brisbane, Australia).

Microsatellite isolation and characterization

64 Unique sequence contigs possessing microsatellite motifs were identified using the software GDD2 (Meglécz et al 2010). Primer 3 (Rozen and Skaletsky 2000) was used to design optimal primer sets for each unique contig where possible. A selection of 48 contigs including di-, tri-, and tetra-nucleotide repeats, were used for subsequent analysis. Loci were initially screened for polymorphism using template DNA from eight individuals from the laboratory population. Loci were pooled into ten groups of four, labelled with unique fluorophores (FAM, NED, VIC, PET) and co-amplified by multiplex PCR using a Qiagen multiplex kit (Qiagen) and an Eppendorf Mastercycler

S gradient PCR machine (see Blacket et al 2012). Genotyping was subsequently performed using an Applied Biosystems 3730 capillary analyzer (AGRF) and product lengths were scored manually and assessed for polymorphisms using GeneMapper version 4.0 (Applied Biosystems).

Polymorphic loci were selected and pooled into three groups for multiplexing based on observed locus specific allele size ranges, and checked using DNA obtained from wing and leg tissue of 11 field collected individuals from Crystal Cascades, north

Queensland. Reaction matrices for all PCR amplification consisted of 5 µL Qiagen multiplex mix, 4 µL of primer mix (0.2 µM of each primer) and 2 µL of template

DNA. PCR conditions consisted of an initial 15 minutes denaturing step at 94 °C, followed by 40 cycles of 94 °C for 30 seconds, 59 °C for 1:30 minutes, and 72 °C for

1:00 minutes, with a final extension step of 60 °C for 30 minutes.

65 Microsatellite profiles of the field specimens were examined using GeneMapper version 4.0 and alleles scored manually. The Excel Microsatellite Toolkit was used to estimate expected (HE) and observed (HO) heterozygosities and number of alleles (NA), while examination of conformation to Hardy–Weinberg equilibrium (HWE), the inbreeding coefficient (FIS) and linkage disequilibrium estimates between all pairs of loci were conducted using GENEPOP (online version: http://genepop.curtin.edu.au/).

Significance was adjusted for multiple comparisons using the sequential Bonferroni procedure where necessary (Rice 1989). Finally, all loci were assessed using MICRO-

CHECKER to check for null alleles (Van Oosterhout et al 2004). Null alleles are created by failure in PCR amplification and can be caused by mutations in primer binding regions or because PCR conditions are not ideal (Selkoe et al 2006). The frequency of null alleles was calculated using the Brookfield method (Brookfield

1996). In order to calculate genotyping error rate, we genotyped 33 individuals drawn randomly from the laboratory population in duplicate and determined the number of mismatches (Hoffman and Amos 2005).

Patterns of inheritance

To investigate inheritance patterns of the microsatellite loci, we collected sexually produced eggs from three families derived from the laboratory stock. Each male- female pair was housed in a separate cage in climate-controlled conditions (as above), until copulation took place (determined by the presence of a spermatophore – a protein capsule containing spermatozoa that is transferred from the male to the female). Eggs

66 were collected a week after copulation and were incubated as above. A subset of early instar nymphs (N = 45) from three distinct families was used to determine patterns of microsatellite allele inheritance. DNA was extracted and microsatellite markers were amplified as above. PCR products were sent to AGRF for genotyping. Microsatellite profiles were then scored manually using GeneMapper version 4.0. We estimated the probability of conformance to Mendelian expectations using G-tests (α = 0.05).

Significance values were adjusted for multiple comparisons using the sequential

Bonferroni correction method where necessary (Rice 1989).

The mechanism of parthenogenesis

Five virgin females from the stock population were isolated and housed in individual cages as above. These virgin females had no contact with males from the early juvenile phase of their lifecycle thus ensuring that eggs produced were created parthenogenetically. Females commenced oviposition approximately 30 days after the final juvenile moult. Eggs were collected weekly and incubated as above. First instar nymphs (N = 169 from five virgin mothers) were collected upon hatching. DNA was extracted and microsatellite markers were amplified as above. PCR products were genotyped at AGRF and microsatellite profiles were scored as above.

67 Results

Next-generation sequencing and de novo genome assembly

A total of 133,048 sequence reads (average length = 435.3 ± 0.33 bp) were obtained from the 454 Next Generation Sequencing (NGS) run. Previous studies indicate that these figures are not excessive as they are commonly achieved by NGS analyses using only 1/16th of a 70 × 75 mm picoTitre Plate (Miller et al 2012, 2013). Nonetheless these data represent approximately 4.4 % of the ~1.3 Gb stick insect genome (Soria-

Carrasco et al 2014).

Microsatellite isolation and characterization

A total of 3,257 unique sequence contigs possessing microsatellite motifs were identified by GDD2 analysis of which 2,361 contigs had optimal priming sites adjacent to microsatellite motifs. Forty-eight contigs were selected for subsequent analysis, 33 of which contained di-nucleotide repeats, 9 containing tri-nucleotide repeats, and 6 containing a tetra-nucleotide repeat. The initial screening analysis found 18 loci to be polymorphic, 11 were monomorphic and 19 failed to amplify. The 18 polymorphic loci were then pooled into three groups for multiplexing based on observed locus specific allele size ranges (Table 3-1). The genotyping error rate was 0.02 ± 0.005

(mean ± SE) for the 18 isolated microsatellite markers, with values > 0.03 for loci

ET17, ET27 and ET39 (0.07, 0.05 and 0.06 respectively, Table 3-1). These three loci

68 were thus excluded from further analyses. The majority of the loci were characterized by moderate to high levels of genetic variation with an average of 4.7 ± 0.58 alleles per locus while microsatellite heterozygosity estimates ranged between 0 and 1 (mean ±

SE = 0.5 ± 0.07) among loci in the field collected specimens (Table 3-2). Locus ET29 was monomorphic in all field-collected samples. Linkage disequilibrium analysis confirmed marker independence, indicating no evidence of significant linkage between loci, while MICRO-CHECKER analysis suggested that there may be null alleles present at loci ET8, ET20, ET24 and ET40 (Table 3-2). All loci conformed to Hardy–

Weinberg expectations according to Fisher’s exact tests and FIS estimates did not show any evidence for heterozygote excess in the Crystal Cascades population (Table 3-2).

Patterns of allele inheritance

The genotypes of 45 progeny from three families derived from the stock population showed conformity with Mendelian bi-allelic segregation patterns at all 15 microsatellite loci (Table 3-3). Family A had 31 distinct alleles in total with two polymorphic (heterozygous) loci in the mother’s and eight in the father’s genome.

Family B had a total of 32 alleles with eight polymorphic loci in the mother’s and six in the father’s genome. Family C had 29 distinct loci in total with two polymorphic loci in the mother’s and seven in the father’s genome. ET26 was the only monomorphic locus in all three families.

69 Mechanism of parthenogenesis

The proportion of homozygous offspring produced parthenogenetically by heterozygous mothers ranged between 0.44 – 1, depending on the locus (Table 3-4).

Polymorphism was rare in offspring at most loci, despite at least one heterozygous mother for each locus (except loci ET2, ET20, ET29 and ET40 which were homozygous in all five mothers). Therefore most loci had homozygosity transition rates (R) close to 1, with progeny from heterozygous mothers becoming mostly homozygous. The only exception is locus ET6 with 62 heterozygous offspring out of a total of 111 (R = 0.44). The observed R value for this locus is significantly different from the expected rates under apomixis or automixis with gamete duplication, central fusion and random fusion (Table 3-5). None of the loci had R values significantly different to those expected under automixis with terminal fusion. Therefore, the most likely mechanism of parthenogenesis in E. tiaratum is automixis with terminal fusion.

70 Table 3-1 Primer sequences, repeat motifs, size ranges, genotyping error rates and Genebank accession number for 18 microsatellite markers isolated from Extatosoma tiaratum

Genebank Repeat Locus Primer sequences (5'–3') Size (bp) Error rate Accession motif No. Multiplex 1 ET2 AGAGCAACTCAGCTGACGAA ACT 90 0 KP938274 TCTCATCCTCTTGATCATTGTTT ET14 AGATACCACGTGGCAAGCAT AG 188 0.01 KP938277 GTGACGAGTCTTCGCCAGTT ET16 ATGGTCGACGAGCGTACAAT AC 103 0 KP938275 CTGCTTCTGCAAAGAGTGGG ET26 CATGGCTACAACCTTGGCTC AGGT 404 0 KP938279 TACTGATCCCACCATTTGGC ET30 TTGAAAGTGCCCAGACCTGT AC 150 0.01 KP938276 TAACACGGTGCATTGTCCAG ET40 TCATACGATTGCAGAACCGA AG 229 0.01 KP938278 CAGGCAATGAGCGATAGAGAA

Multiplex 2 ET6 GTACTGAGACTTCACGAACTAAATGT AG 292 0 KP938284 TGACATTGCTGGTCAACAGG ET11 TCTTCTTACGCCCACGACAT AG 149 0.03 KP938282 AGAGGGTTGATTCCGTGTGA

ET15 AGATACCACGTGGCAAGCAT AC 319 0 KP938285 CAGTACACGAGTGGTGGTCG ET24 TTCGGAAGATAGCATTACATTTAACA AG 90 0 KP938280 GGCGGGAAACTTTCGAGTAG ET27 TTCCTACAGCTCTGACTATCCG AAT 131 0.05 KP938281 TGATCACAGTCAAAGTTGGACAATA ET29 GCAAGCGACTTAAGCTTCCC AC 202 0 KP938283 TCACCTCGAGACCAGCAAAT Multiplex 3 ET8 CTGCAAGTAGCCAACTTCTCTG AAT 121 0.03 KP938288 ATGCTGTTGCTTGGAATGCT ET17 TATCTCGAGGAAGACGAGCG AC 130 0.07 KP938289 AGAGACGTGGTCAGTCAGCC ET20 TCCCTATGACATCACCACCC AG 214 0.02 KP938291 CAGACAACGAAACGAGATGG ET35 CACCTCCTTCCTCCTTCACA AC 108 0 KP938287 AACATGGCTTGTGATGCATTT ET39 GCCTCAGTGTGGGAAGTTGT AG 98 0.06 KP938286 CTGACAGGTACTGGCTGGCT ET42 ACTTTCACCTGTCCGTGCTG AAT 181 0 KP938290 GAATTTCGGAGTGGCTTCAA

71 Table 3-2 Characterization of 15 microsatellite markers estimated from 11 individuals from the Crystal Cascades, North Queensland, Australia

Frequency of Locus N H /H HWE F A O E IS null alleles

ET2 2 0.36/0.31 1 -0.2 0 ET6 2 0.27/0.51 1 0.37 0.29 ET8 6 0.40/0.68 0.35 0.5 0.27 ET11 6 0.64/0.80 1 0.26 0.11 ET14 9 0.64/0.74 1 0.14 0.009 ET15 2 0.27/0.25 1 -0.12 0 ET16 5 0.90/0.69 1 -0.32 0 ET20 8 0.50/0.86 1 0.59 0.4 ET24 8 0.60/0.86 0.48 0.34 0.12 ET26 7 0.90/0.78 1 -0.13 0.12 ET29 1 NA NA NA 0 ET30 2 0.18/0.17 0.74 0 0 ET35 3 0.45/0.58 1 0.19 0.07 ET40 6 0.54/0.84 0.38 0.32 0.14 ET42 4 0.70/0.73 1 0.32 0.22

Number of alleles (NA), observed (HO) and expected (HE) heterozygosities, Hardy-Weinberg equilibrium p-values (HWE) and inbreeding coefficient (FIS). p-values indicate the probability that HO differs from HE. Locus ET29 is monomorphic in this population, therefore HO, HE, HWE and FIS values are not available (NA).

72 Table 3-3 Microsatellite profiles of captive-bred E. tiaratum parents and offspring obtained from three families

Famil Locus Mother Father Offspring Exp Ratio N p-value y

ET14 A 212, 212 212, 212 212, 212 (12) 1 12 a

B 200, 212 212, 222 212, 212 (10) 200, 222 (8) 200, 212 (5) 212, 222 (4) 1:1:1:1 27 1.00

C 212, 212 212, 212 212, 212 (6) 1 6 a

ET16 A 128, 128 128, 128 128, 128 (12) 1 12 a

B 128, 128 126, 126 126, 128 (27) 1 27 a

C 128, 128 126, 128 128, 128 (3), 126, 128 (3) 1:1 6 1.00

ET2 A 97, 97 97, 109 97, 97 (5) 97, 109 (7) 1:1 12 1.00

B 97, 97 97, 109 97, 97 (14) 97, 109 (13) 1:1 27 1.00

C 97, 97 97, 109 97, 97 (2) 97, 109 (4) 1:1 6 1.00

ET26 A 410, 410 410, 410 410, 410 (12) 1 12 a

B 410, 410 410, 410 410, 410 (27) 1 27 a

C 410, 410 410, 410 410, 410 (6) 1 6 a

ET30 A 167, 167 173, 177 167, 173 (5) 167, 177 (7) 1:1 12 1.00

B 167, 167 173, 177 167, 173 (9) 167, 177 (15) 1:1 24 1.00

C 167, 167 167, 177 167, 167 (2) 167, 177 (3) 1:1 5 1.00

ET40 A 256, 256 254, 258 254, 256 (6) 256, 258 (6) 1:1 12 1.00

B 256, 256 254, 254 254, 256 (27) 1 27 a

C 256, 256 256, 258 256, 256 (3) 256, 258 (3) 1:1 6 1.00

ET11 A 170, 170 170, 174 170, 170 (5), 170, 174 (7) 1:1 12 1.00

B 170, 170 174, 174 170, 174 (27) 1 27 a

C 170, 170 170, 174 170, 170 (2), 170, 174 (4) 1:1 6 1.00

ET15 A 340, 340 338, 340 340, 340 (7) 338, 340 (5) 1:1 12 1.00

B 340, 344 338, 340 340, 340 (8) 338, 344 (8) 340, 344 (6) 338, 340 (5) 1:1:1:1 27 1.00

C 340, 340 340, 340 340, 340 (6) 1 6 a

ET 24 A 120, 120 112, 112 112, 120 (12) 1 12 a

B 104, 104 112, 112 104, 120 (27) 1 27 a

C 104, 120 124, 124 104, 124 (4) 120, 124 (2) 1:1 6 1

ET29 A 229, 229 229, 235 229, 235 (6) 229, 229 (6) 1:1 12 1.00

B 229, 229 229, 229 229, 229 (27) 1 27 a

C 229, 229 229, 229 229, 229 (6) 1 6 a

ET6 A 338, 310 308, 308 338, 310 (10) 308, 308 (2) 1:1 12 1.00

B 310, 310 308, 310 338, 310 (10) 310, 310 (17) 1:1 27 1.00

73 C 310, 310 308, 308 308, 310 (6) 1 6 a

ET20 A 289, 289 237, 239 239, 289 (5) 1:1 5 0.56

B 285, 289 239, 289 289, 289 (12) 285, 289 (3) 239, 289 (1) 1:1:1:1 16 0.84

C 289, 289 239, 243 243, 289 (1) 243, 289 (1) 1:1 2 1.00

ET35 A 126, 126 126,128 126, 126 (5) 1:1 5 0.56

B 128, 130 130, 130 130, 130 (12) 128, 130 (1) 1:1 13 0.56

C 126, 126 128, 128 126, 128 (4) 1 4 a

ET42 A 200, 203 206, 206 200, 206 (5) 203, 206 (4) 1:1 9 1.00

B 200, 203 206, 206 200, 206 (10) 203, 206 (7) 1:1 17 1.00

C 200, 203 200, 206 200, 206 (1) 1:1:1:1 1 1.00

ET8 A 139, 139 139, 142 139, 142 (6) 139, 139 (5) 1:1 11 1.00

B 139, 139 139, 139 139, 139 (27) 1 27 a

C 139, 139 142, 142 139, 142 (3) 1 3 a

Expected ratios under Mendelian segregation, number of offspring (N) and p-values obtained from G- tests are shown. Frequency of each genotype is shown in parenthesis under offspring. a - observed and expected genotypes were exactly the same, and therefore a G-tests could not be performed.

74 Table 3-4 Number of homozygous and heterozygous offspring produced parthenogenetically by heterozygous mothers and rates of transition to homozygosity (R)

Informative loci Hetero mother Homo offspring Hetero offspring R

ET6 4 49 62 0.44

ET8 1 22 0 1.00

ET11 2 64 0 1.00

ET14 3 91 1 0.99

ET15 5 158 1 0.99

ET16 1 36 0 1.00

ET24 2 71 0 1.00

ET26 3 89 1 0.99

ET29 3 101 0 1.00

ET30 1 34 0 1.00

ET42 1 150 1 0.99

75 Table 3-5 G-tests for observed and expected rates of transition to homozygosity (R) depending on the mechanism of parthenogenesis (P = 0.05). R is not significantly different from expected rates under automixis with terminal fusion for any loci. * Represents significant differences between the observed and expected R.

Automixy Locus Gamete duplication Central fusion Random fusion Terminal fusion Apomixy r=1 r=0-1/3 r=1/3 r=1/3-1 r=0 ET6 **** 1 1 1 **** ET8 1 **** **** 1 **** ET11 1 **** **** 1 **** ET14 1 **** **** 1 **** ET15 1 **** **** 1 **** ET16 1 **** **** 1 **** ET24 1 **** **** 1 **** ET26 1 **** **** 1 **** ET29 1 **** **** 1 **** ET30 1 **** **** 1 **** ET42 1 **** **** 1 ****

76 Discussion

We isolated 18 microsatellite markers for E. tiaratum, three of which had high genotyping error rates possibly caused by stutter patterns (loci ET 17, ET27 and ET39,

Table 3-1). The remaining 15 microsatellite markers were independent and conform to

Mendelian expectations. Our preliminary analysis suggests that the population in

Crystal Cascades is moderately genetically diverse (average 4.7 ± 0.58 alleles per locus) with only one monomorphic locus (ET29). This level of microsatellite polymorphism is greater than that found in two sexually reproducing populations of another phasmid, Bacillus rossius, (with one and three monomorphic loci out of five,

Andersen et al 2005). The presence of multiple polymorphic loci in the genome of individuals from the Crystal Cascades population suggests that all the individuals were produced sexually, given the mechanism of parthenogenesis (automixis with terminal fusion) will lead to high individual homozygosity. The population was in HWE, with

FIS values indicating no evidence of heterozygosity excess or deficit (Table 3-2).

However, larger samples sizes are required to confirm these results, as there was a suggestion that four loci (ET8, ET20, ET24 and ET40) contained null alleles.

Microsatellite analysis using 15 microsatellite markers suggests that the most likely mechanism of parthenogenesis in E. tiaratum is automixis with terminal fusion. Most heterozygous loci in mothers (except ET6) were homozygous in their parthenogenetically produced offspring, thereby excluding either apomixis, automixis with central fusion, or random fusion mechanisms of parthenogenesis. Locus ET6 had

77 high transition rates to heterozygosity, which distinguishes the mechanism from gamete duplication. Further cytogenetic analyses are required to confirm this mechanism. The lower rate of transition to homozygosity in locus ET6 suggests that this locus might be located further away from the centromere and therefore have higher recombination rates (given that the mechanism is automixis with terminal fusion). Although endosymbiotic infection is unlikely to be responsible for the occurrence of parthenogenetic reproduction in Phasmatodea (Perez-ruiz et al 2015), we are unable to completely rule out this possibility. Nonetheless, our data indicate that parthenogenesis leads to severe loss of microsatellite heterozygosity in E. tiaratum.

Loss of microsatellite heterozygosity may be associated with loss of heterozygosity at fitness related loci and therefore could be an important component of determining reproductive fitness (Väli et al 2008). Loss of heterozygosity can result in inbreeding depression when homozygosity is associated with increased genetic load due to the accumulation of recessive deleterious alleles (partial dominance hypothesis:

Charlesworth and Charlesworth 1999, Roff 2002). The genetic load is expected to purge rapidly in isolated parthenogenetic populations, resulting in lineages with improved parthenogenetic capacity (Kramer and Templeton 2001). However, this pattern may not be likely to occur in E. tiaratum with high rates of sexual reproduction.

Studies in various organisms report correlations between loss of heterozygosity and reduced fitness (Chapman et al 2009). Such fitness costs may explain why most obligate or cyclic parthenogenetic species reproduce by apomixis (e.g. Johnson and leefe 1999, Mark Welch and Meselson 2000, Delmotte et al 2001, Tsutsui et al 2014,

78 Vavre et al 2004) while automixis is mainly observed among facultative or geographic

(only in certain populations) parthenogens (e.g. Matsuura et al 2004, Pearcy et al 2006,

Sekiné and Tojo 2009, Kellner and Heinze 2011). Apomictic populations may persist over a greater number of generations while avoiding inbreeding depression, whereas automictic populations would quickly become inbred (Schwander and Crespi 2009).

The specimens collected from Crystal Cascades showed high heterozygosity for most loci in all individuals, suggesting they were produced sexually. While this is a small sample size from a single population, this observation is consistent with the prediction that automictic parthenogenesis occurs sporadically, for example in cases of mating failure (Stalker 1956, Schwander et al 2010). It is noticeable that other factors apart from inbreeding depression may be responsible for reduced fitness in parthenogenetic eggs, for example development constraints associated with an absence of paternal factors (see Engelstaedter 2008).

In the stick insect order Phasmatodea, obligate parthenogenetic species such as

Bacillus whitei, Sipyloidea sipylus, Carausius morous, and obligate parthenogenetic species of Timema genus reproduce by apomixis (Pijnacker 1966, 1967, Marescalchi et al 1991, Schwander and Crespi 2009). In contrast, automixis has been reported in the facultatively parthenogenetic Bacillus rossius and B. atticus (Pijnacker 1969,

Marescalchi et al 1993). This observation is consistent with the above argument that apomictic lineages are generally more viable as they are more successful at maintaining genetic diversity. The transition from sexual reproduction to automixis does not involve drastic cytological changes and occurs at low frequencies in many normally sexually reproducing species (Schwander et al 2010). This frequency can

79 increase once there is a selective advantage to automixis, for example through mating failure (Schwander et al 2010). The adaptive significance of automictic parthenogenesis is not fully understood, but potential fitness costs associated with a loss of heterozygosity at fitness related loci, may be evident at a diverse range of life history traits.

80 References

Andersen, D. H., Pertoldi, C., Loeschcke, V., and Scali, V. (2005). Characterization of microsatellite loci in the stick insects Bacillus rossius rossius, Bacillus rossius redtenbacheri and Bacillus whitei (Insecta: Phasmatodea). Molecular Ecology Notes, 5, 576–578. Bell, G. (1982). The masterpiece of nature: the evolution and genetics of sexuality. University of California, Berkely. U.S.A. Bian, X., Elgar. M. A., and Peters, R. (2016). The swaying behavior of Extatosoma tiaratum: motion camouflage in a stick insect? Behavioral Ecology, 27, 83–92. Blacket, M. J., Robin, C., Good, R. T., Lee, S. F., and Miller, A. D. (2012). Universal primers for fluorescent labelling of PCR fragments - an efficient and cost- effective approach to genotyping by fluorescence. Molecular Ecology Resources, 12(3), 456–63. Brookfield J. F. Y. (1996). A simple new method for estimating null allele frequency from heterozygote deficiency. Molecular Ecology, 5, 453–455. Carlberg, U. (1983). Copulation in Extatosoma tiaratum (MacLeay) (Insecta: Phasmida). Zoologischer Anzeiger. 5(6), 340-356. Chapman, J. R., Nakagawa, S., Coltman, D. W., Slate, J., and Sheldon, B.C. (2009). A quantitative review of heterozygosity-fitness correlations in animal populations. Molecular Ecology, 18(13), 2746–2765. Charlesworth, B. and Charlesworth, D. (1999). The genetic basis of inbreeding depression. Genetical Research, 74(03), 329–340. Delmotte, F., Leterme, N., Bonhomme, J., Rispe, C., and Simon, J.C. (2001). Multiple routes to asexuality in an aphid species. Proceedings of the Royal Society of London, Biological Sciences, 268(1483), 2291–9. Engelstädter, J. (2008). Constraints on the evolution of asexual reproduction. BioEssays 30, 1138–1150. Gurney, A. B. (1947). Notes on some remarkable Australasian walkingsticks, including a synopsis of the genus Extatosoma (Orthoptera: Phasmatidae). Annals of the Entomological Society of America, 40(3), 373–396. Hoffman, J. I., and Amos, W. (2005). Microsatellite genotyping errors: detection, approaches, common sources and consequences for paternal exclusion. Molecular Ecology, 14, 599–612. Johnson, S. G., and Leefe, W. R. (1999). Clonal diversity and polyphyletic origins of hybrid and spontaneous parthenogenetic Campeloma (Gastropoda: Viviparidae) from the south-eastern United States. Journal of Evolutionary Biology, 12(6),

81 1056–1068. Kellner, K., and Heinze, J. (2011). Mechanism of facultative parthenogenesis in the ant Platythyrea punctata. Evolutionary Ecology, 25(1), 77–89. Marescalchi, O., Pijnacker, L. P., and Scali, V. (1991). Cytology of parthenogenesis in Bacillus whitei and Bacillus lynceorum (Insecta Phasmatodea). Invertebrate Reproduction and Development, 20(1), 75–81. Marescalchi, O., Pijnacker, L. P., and Scali, V. (1993). Automictic parthenogenesis and its genetic consequence in Bacillus atticus atticus (Insecta Phasmatodea). Invertebrate Reproduction and Development, 24(1), 7–12. Mark Welch, D., and Meselson, M. (2000). Evidence for the evolution of bdelloid rotifers without sexual reproduction or genetic exchange. Science, 288(5469), 1211–1215. Matsuura, K., Fujimoto, M., and Goka, K. (2004). Sexual and asexual colony foundation and the mechanism of facultative parthenogenesis in the termite Reticulitermes speratus (Isoptera, Rhinotermitidae). Insectes Sociaux, 51(4), 325– 332. Meglécz, E., Costedoat, C., Dubut, V., Gilles, A., Malausa, T., Pech, N., and Martin, J.F. (2010) QDD: a user-friendly program to select microsatellite markers and design primers from large sequencing projects. Bioinformatics (Oxford, England), 26(3), 403–4. Miller, A. D., van Rooyen, A., Ayres, R. M., and Raadik, T. A., Fairbrother, P. and Weeks, A. R. (2012). The development of 24 polymorphic microsatellite loci for the endangered barred galaxias, Galaxias fuscus, through next generation DNA sequencing. Conservation Genetics Resources, 4(3), 613–616. Miller, A. D., Van Rooyen, A., Sweeney, O. F., Whiterod, N. S., and Weeks, A. R. (2013). The development of 10 novel polymorphic microsatellite markers through next generation sequencing and a preliminary population genetic analysis for the endangered Glenelg spiny crayfish, Euastacus bispinosus. Molecular Biology Reports, 40(7), 4415–4419. Pearcy, M., Hardy, O., and Aron, S. (2006). Thelytokous parthenogenesis and its consequences on inbreeding in an ant. Heredity, 96(5), 377–382. Perez-Ruiz, M., Martinez-Rodriguez, P., Herranz, J., and Bella, J. L. (2015). A survey of Wolbachia, Spiroplasma and other bacteria in parthenogenetic and non- parthenogenetic phasmid (Phasmatodea) species. European Journal of Entomology, 112(3), 1–10. Pijnacker, L. P. (1966). The maturation divisions of the parthenogenetic stick insect Carausius morosus Br. (Orthoptera, Phasmidae). Chromosom, 19(1), 99–112. Pijnacker, L. P. (1967). Oogenesis in the parthenogenetic stick insect Sipyloidea sipylus Westwood (Orthoptera, Phasmidae). Genetica, 38(1), 504–515.

82 Pijnacker, L. P. (1969). Automictic parthenogenesis in the stick insect Bacillus rossius Rossi (Cheleutoptera, phasmidae). Genetica, 40(1), 393–399. Rabeling, C., and Kronauer, D.J.C. (2013). Thelytokous parthenogenesis in eusocial Hymenoptera. Annual Review of Entomology, 58, 273–292. Rice, W. R. (1989). Analyzing tables of statistical tests. Evolution, 43(1), 223–225. Roff, D. A. (2002). Inbreeding depression: tests of the overdominance and partial dominance hypotheses. Evolution, 56(4), 768–775. Rozen, S., and Skaletsky, H. (2000). Primer3 on the WWW for general users and for biologist programmers. Bioinformatics Methods and Protocols, 132, 365–386. Schneider, A., and Elgar, M. A. (2010). Facultative sex and reproductive strategies in response to male availability in the spiny stick insect, Extatosoma tiaratum. Australian Journal of Zoology, 58(4), 228–233. Schwander, T., and Crespi, B.J. (2009). Multiple direct transitions from sexual reproduction to apomictic parthenogenesis in Timema stick insects. Evolution, 63(1), 84–103. Schwander, T., Vuilleumier, S., Dubman, J., and Crespi, B.J. (2010). Positive feedback in the transition from sexual reproduction to parthenogenesis. Proceedings of the Royal Society B: Biological Sciences, 277(1686), 1435–1442. Schuster, S. C. (2008). Next-generation sequencing transforms today’s biology. Nature Methods 5, 16–18. Sekiné, K., and Tojo, K. (2009). Automictic parthenogenesis of a geographically parthenogenetic mayfly, Ephoron shigae (Insecta: Ephemeroptera, Polymitarcyidae). Biological Journal of the Linnean Society, 99(2), 335–343. Simon, J., Delmotte, F., Rispe, C., and Crease, T. (2003). Phylogenetic relationships between parthenogens and their sexual relatives: the possible routes to parthenogenesis in animals. Biological Journal of the Linnean Society, 79(1), 151–163. Smith, J. M., and Maynard-Smith, J. (1978). The evolution of sex. Cambridge University Press. Cambrige, UK. Soria-Carrasco, V., Gompert, Z., Comeault, A. A., Farkas, T. E., Parchman, T. L., Jonston, J. S., Buerkle, B. C., Feder, J. L., Bast, J., Schwander, T., Egan, S. P., Crespi, B. J., and Nosil, P. (2014). Stick insect genomes reveal natural selection’s role in parallel speciation. Science, 344, 738–42. Stalker, H. D. (1956). On the evolution of parthenogenesis in Lonchoptera (Diptera). Evolution, 345–359. Stouthamer, R., Breeuwer, J. A. J., and Hurst, G.D.D. (1999). Wolbachia Pipientis: microbial manipulator of arthropod reproduction. Annual Review of Microbiology, 53(1), 71–102.

83 Suomalainen, E., Saura, A., and Lokki, J. (1987) Cytology and evolution in parthenogenesis. CRC Press. Boca Raton, U.S.A. Tsutsui, Y., Maeto, K., Hamaguchi, K., Isaki, Y., Takami, Y., Naito, T., and Miura, K. (2014). Apomictic parthenogenesis in a parasitoid wasp Meteorus pulchricornis, uncommon in the haplodiploid order Hymenoptera. Bulletin of Entomological Research, 104(03), 307–313. Van Oosterhout, C., Hutchinson, W. F., Wills, D. P. M., and Shipley, P. (2004). MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes, 4(3), 535–538. Vavre, F., de Jong, J. H., and Stouthamer, R. (2004). Cytogenetic mechanism and genetic consequences of thelytoky in the wasp Trichogramma cacoeciae. Heredity, 93(6), 592–6.

Williams, G. C. (1975). Sex and evolution. Princeton University Press. Princeton, U.S.

84

Chapter 4

Sex versus parthenogenesis; immune function in a

facultatively parthenogenetic Phasmatid

(Extatosoma tiaratum)

85 Summary

Facultative parthenogenetic species in which females can alternate between sex and parthenogenesis are suitable models to investigate the costs and benefits of sex and parthenogenesis. Theory predicts sexual reproduction to be advantageous in the co- evolutionary arms race between hosts and pathogens (the Red-queen hypothesis).

While empirical evidence for the advantages of sex in resisting co-evolving pathogens has been demonstrated in various invertebrate taxa, studies investigating the direct effect of sex on immune function are extremely rare. Here, we used the facultatively parthenogenetic Australian phasmatid, Extatosoma tiaratum, to investigate the effect of both maternal and offspring mode of conception (sexual or parthenogenetic) on offspring immune function (haemocyte concentration, lytic activity and phenoloxidase activity). We show that when parthenogenesis persists beyond one generation in E. tiaratum, it may have negative effects on immune function in terms of haemocyte concentration and lytic activity. Moreover, immune response decreases across consecutive sampling weeks, suggesting there are physiological constraints with respect to mounting an immune responses in close time intervals.

86 Introduction

Facultative parthenogenesis in which females can reproduce both sexually and parthenogenetically occurs in isolated clades across a broad taxonomic range, including insects, reptiles, fish, birds and sharks (Booth et al 2012). The adaptive significance of facultative parthenogenesis is poorly understood, but it may provide benefits in environments where the likelihood of sexual reproduction is lower (Stalker

1956; Gerritsen 1980). In contrast, sexual reproduction may persist in such species because of the significant fitness costs associated with parthenogenetic reproduction arising as a consequence of lower genetic variability in parthenogens (Stalker 1956,

Carson 1967, Lamb and Wiley 1979, Hong and Ando 1998). The necessary empirical studies comparing the outcomes of alternative reproductive modes on life history traits are rare and focus mainly on traits directly associated with reproductive fitness (e.g.

Corley et al 2001, Matsuura et al 2004, Matsuura and Kobayashi 2007, Sekine and

Tojo 2010). Facultative parthenogenetic species have been identified as suitable models to investigate the evolution of sexual reproduction, because they provide the opportunity to compare directly sexual and parthenogenetic progeny from the same female. Cyclic parthenogenetic systems (such as Daphnia, rotifers, aphids, etc) are also useful, with the drawback that sexual and parthenogenetic eggs are typically different; resistant, resting eggs are produced via sex and non-resistant eggs are produced via parthenogenesis.

The importance of sexual reproduction for maintaining an effective immune system

87 has been broadly acknowledged (Van Valen 1973, Jaenike 1978, Hamilton 1980). The argument is based on the idea that hosts benefit from producing sexual progeny with reshuffled genomes, which are more likely to provide resistance to co-evolving pathogens. Empirical evidence for the advantage of sex in resisting infection typically cites correlations between the prevalence of sex and pathogen/parasites either in natural populations (e.g. Lively 1987, Kumpulainen et al 2004, Decaestecker et al

2007, King et al 2009), or in studies that employ experimental evolution (e.g. Koskella and Lively 2009, Morran et al 2011, Kerstes et al 2012). However, it remains unclear whether sex has an immediate advantage in terms of increasing the efficacy of an immune response.

Immune function determines the ability of individuals to defend themselves against injury and disease and as such it is likely to have a highly significant impact on fitness.

Invertebrates lack the highly specific immunological memory typical of the adaptive vertebrate immune system. However, recent experimental evidence suggests that the innate immune system has memory-like features analogous to adaptive immunity and thus the capacity and degree to which it responds may shift when an animal is exposed to repeated immunological challenges (Kurtz 2005, Little et al 2005). Insect immunity in particular is composed of both humoral and cellular components (see Beckage

2008), and is relatively easy to measure. There are three commonly used measures for innate immune responses in invertebrates. 1) Haemocyte concentration - a measure of the density of haemocytes within insect haemolymph. These are responsible for the core cellular defence pathways, such as phagocytosis and encapsulation (Ribeiro and

Brehélin 2006); 2) lysozyme-like (lytic) activity – a measures of the capacity of the

88 immune system to lyse bacterial cell walls and thus clear a bacterial infection

(Beckage 2008); and 3) phenoloxidase (PO) activity, which is an important component of the humoral defence pathway, and is involved in the encapsulation response following a foreign challenge such as infection by an endoparasite (Kanost and

Gorman 2008). It is broadly assumed that these measures provide a realistic indication of an invertebrate’s ability to resist disease and infection (Adamo 2004). Direct evidence supporting this assumption includes positive correlations between these immune measures and fitness traits such as survival or offspring production (e.g.

Rantala et al 2002, Siva-Jothy et al 2005, Lawniczak et al 2007) but whether these also vary with mode of reproduction has been rarely investigated.

Here, we used the facultatively parthenogenetic Australian phasmatid, Extatosoma tiaratum, to investigate the trans-generational effect of alternative reproductive modes

(sex and parthenogenesis) on immune function. The relative efficacy of immune function in females with sexual and parthenogenetic modes of conception can be compared directly in E. tiaratum. Female E. tiaratum oviposit continuously throughout their adult lives; unfertilised eggs develop into female offspring via parthenogenesis; fertilised eggs develop into males and female offspring (Carlberg 1983). The cytological mechanism of parthenogenesis (automixis with terminal fusion) results in substantial loss of heterozygosity in the F1 progeny (Chapter 3; Alavi et al 2016), which makes it possible to infer how females, of unknown history, were produced. We investigated the effect of both maternal and offspring modes of conception on offspring haemocyte concentration, lysozyme-like (lytic) activity and phenoloxidase

(PO) activity. In addition, we asked whether immune function varied longitudinally

89 across the sampling periods. The main reason to use this protocol is that in the absence of a direct challenge (e.g. a bacterial challenge), we can use wounding as a low level but potentially realistic challenge that should promote upregulation of the immune system. The experiment further permits us to explore whether variation, if it does exist, is comparable for sexually and parthenogenetically produced offspring.

90 Methods

Animal maintenance and culturing

A stock population of E. tiaratum was established from eggs and juveniles obtained from the Melbourne Museum Victoria, Australia. Females were housed in individual cylindrical containers (height: 23 cm; diameter: 25 cm), enclosed by a fine mesh. All insects were provided with ad libitum fresh leaves of various species of Eucalyptus that had been lightly sprayed with water. Eggs were incubated in plastic boxes (16 ×10

×5 cm) with 5 mm of sand and misted with water weekly. All insect containers and egg boxes were kept under climate-controlled conditions (24-26°C; 50% humidity;

12:12hr Light: Dark cycle).

Adult females (yet to commence oviposition) from the stock population (hereafter defined as F0 experimental females) were assigned to either virgin (n = 5) or mated treatments (n = 7). Females in the virgin treatment were kept isolated, while females in the mated treatment were provided with adult males every week until mating occurred and a spermatophore was visibly attached to the female abdomen. We derived first generation female juveniles (hereafter defined as F1 females) from eggs produced by these mated and virgin F0 females (ensuring that eggs were taken from females at a similar age), and raised a proportion though to adulthood (n = 10 female offspring from virgin F0 females; 11 female offspring from mated F0 females). All F1 females remained virgins during the experiment and thus varied only in their mode of

91 conception and not their mating status.

Determining reproductive history

We followed the methods developed in chapter 3 (Alavi et al 2016), to infer the mode of conception (MOC), for F0 and F1 females used in the experiment; the latter were assayed to confirm that F1 females with a mated mother were produced sexually rather than parthenogentically. The cytological mechanism of parthenogenesis (automixis with terminal fusion) results in substantial loss of microsatellite heterozygosity, such that 14 out of 15 developed markers are usually homozygous in parthenogenetically produced females (Alavi et al 2016). We genotyped 15 microsatellite markers in all F0 females and F1 offspring used in the experiment (see Chapter 3) and inferred the mode of conception, based on the ratio of the proportion of heterozygous to homozygous microsatellite loci (less than 0.1 = parthenogenetic MOC i.e. fusion of two egg cells; >

0.1 = sexual MOC i.e. fusion of egg cell and sperm).

Haemolymph collection and processing

To assess the relationship between maternal (F0) and offspring (F1) mode of conception and offspring (F1) immune function, we collected haemolymph from all virgin adult F1 females at three time periods: (i) 3-weeks post final moult to adult – this was before the start of oviposition and thus provided baseline immune function

92 prior to egg production; (ii) 5-weeks post final moult – here, all females had commenced ovipositing; and (iii) 7-weeks post final moult. For each sample, a small puncture was made in the joint between the leg and abdomen using a 23G sterile needle (Becton Dickinson and Co.; Melbourne, VIC, Australia). A total of 6 μl of haemolymph was pipetted from the created bubble; transferred to a 0.5 ml Eppendorf tube (Sarstedt, Mawson Lakes, SA, Australia) containing 80 μl of Phosphate buffer saline (PBS; 11.9 mM phosphate, 137 mM NaCl, 2.7 mM KCl, pH 7.4) and the solution was stored at −80 0C for later processing (following Siva-Jothy et al 2005,

Jones et al 2015). The haemolymph collection procedure was minimally invasive and did not visibly affect experimental animals, however the expectation was that the wounding provided them with a challenge and they would respond by upregulation of immune function.

Haemocyte concentration

We determined haemocyte concentration for each sample within 15 mins following haemolymph collection and before freezing the haemolymph. A 10 μl aliquot of haemocyte and PBS solution was pipetted onto a Neubauer haemocytometer

(Blaubrand, Wertheim, Germany) and examined at × 40 magnification on an Olympus

BX50 stereomicroscope (Olympus, Tokyo, Japan). The total number of haemocytes was counted and haemocyte concentration was calculated (Following Jones et al

2015).

93 Lytic activity assay

Lytic activity was assessed using methods similar to previously published insect studies (see Jones et al2015). We assessed lytic activity by adding 70 μl Micrococcus lutus (lysodeikticus) bacteria solution (3 mg/ml PBS) to 20 μl of the haemolymph-PBS solution in a round bottom 96-well plate (Sarstedt, Mawson Lakes, SA, Australia). We measured the change in absorbance at 490 nm, over 180 minutes at 30 0C using a micro-plate spectrometer (PerkinElmer Enspire 3.0 Multimode Plate Reader;

Melbourne, VIC, Australia). Measurements were taken at the start and end of the 120 minutes period. As suspended particles settle over time, we manually mixed the samples immediately prior to the final read. As lysozyme in the haemolymph degrades the bacteria, the turbidity of the solution decreases over time, and so the total change in absorbance was calculated by subtracting the final read from the initial, higher values showing greater lytic activity. Each sample was run in duplicate with a series of controls on each plate (n = 2 plates): two sample controls (pure PBS instead of hemolymph - PBS solution) and 12 PBQCs (pooled biological quality controls). The

PBQCs were created from two separate pools of hemolymph - PBS solution aliquoted into smaller amounts and included in duplicate at the start, middle and end of each plate. The PBQCs were used to calculate the mean coefficient of variation within the plates (18.2 ± 0.03 %).

94 Phenoloxidase (PO) assay

To explore variation in PO activity, we added 7 μl of 1.3 mg / ml bovine pancreas α- chymotrypsin to 5 μl of haemolymph-PBS solution in a round bottom 96-well

(following Jones et al 2015). The samples were then incubated for 20 minutes at room temperature to convert all pro-PO to active PO (Moreno-Garcia et al 2013). We then added 90 μl of L-DOPA to each well and measured the change in absorbance at 490 nm measured over 120 min at 26oC. PO converts L-DOPA to dopachrome, causing the samples to darken over time. At the end of the assay, we calculated the total change in absorbance by subtracting the initial read from the final read, higher values showing greater total PO content. As above, we ran each sample in duplicate and had two sets of controls on each plate (n = 2 plates). The coefficient of variation was 4.9 ± 0.04%.

Statistical analysis

Analyses were performed using JMP version 12.0.0 (SAS institute, NC, USA). We ran two sets of general linear mixed models for each immune function measure; first, we investigated baseline haemocyte concentration, lytic activity and Po activity only in the offspring derived from mated mothers. The purpose of this analysis was to distinguish between the effect of mating per se from the effect of the mode of conception (mated females can reproduce offspring parthenogenetically, see below).

Second, we ran overall models including all females and the three sampling periods.

We investigated the effect of sampling period (week), F0 mode of conception (sexual

95 and parthenogenetic) and F1 mode of conception (sexual and parthenogenetic) on each immune function measure. Individual identity was added as a random effect to all models. To account for the initial baseline differences at week three, we weighted overall models by the corresponding immune function measure at 3-week sampling points. Haemocyte concentration and lytic activity were Ln transformed to improve normality. Full or maximal models initially included all main effects and interactions and were then reduced using hierarchical removal of any term with a significance of P

> 0.1.

96 Results

Determining reproductive history

The microsatellite heterozygosity data revealed that among the F0 females (N = 12), seven (n = 5 allocated to the virgin treatment and n = 2 allocated to the mated treatment) had been conceived sexually and five (all allocated to the mated treatment) had been conceived parthenogenetically (Table 4-1). All F1 females derived from virgin F0 were conceived parthenogenetically (with no heterozygous loci); while three of the mated F0 females also produced parthenogenetic daughters. The remaining four mated F0 females all produced sexual daughters (heterozygosity proportion = 0.35 ±

0.04).

Immune function assays

The average immune function measures are shown in Table 4-2. Our initial set of analyses showed that among the F1 females derived from mated F0 females, baseline lytic activity was significantly higher in those with sexual compared with parthenogenetic modes of conception (F1,10 = 7.94, P = 0.02). There was a trend toward higher baseline PO activity in females with sexual modes of conception (mean ± standard error PO activity = 0.86 ± 0.07) compared to parthenogenetic females (mean

± standard error PO activity = 0.92 ± 0.05; F1,10 = 4.44, P = 0.06). Baseline haemocyte

97 concentration was comparable between F1 females with sexual and parthenogenetic modes of conception (F1,10 = 1.10, P = 0.32).

The overall models showed that both haemocyte concentration (Table 4-3a, Figure 4-

1a) and lytic activity (Table 4-3b, Figure 4-1b) are significantly higher in the F1 offspring derived from F0 females, with sexual modes of conception compared to parthenogenetic modes of conception. Both haemocyte concentration (Table 4-3a,

Figure 4-2a) and lytic activity (Table 4-3b, Figure 4-2b) decreased over the three sampling periods. While there was no significant interaction between F0 and F1 modes of conception for haemocyte concentration (F1, 61 = 0.09, P = 0.80), this interaction was significant for lytic activity (F1, 62 = 6.61, P = 0.02). Post hoc Tukey’s tests revealed that F1 offspring derived from sexually conceived F0, that were parthenogenetically conceived (F0 - F1 MOC = sexual – parthenogenetic), have the highest lytic activity, while parthenogenetic – parthenogenetic females have the lowest lytic activity (Figure

4-1b).

The pattern of PO activity was less clear. There was a positive correlation between PO activity and haemocyte concentration (± SE = 0.31 ± 0.1; P = 0.002, Table 4-3c) and significant interactions of week, with F0 and F1 modes of conception (Table 4-3c). This interaction did not show any clear directional patterns (Figure 4-3a and b). In a separate model investigating the effect of heterozygosity proportion and week of sampling (only including sexually produced females), we found a significant positive

98 correlation between PO activity, and both heterozygosity proportion (F1, 23 = 8.90, P =

0.02) and week of sampling (F2, 22 = 4.61, P = 0.03).

99 Table 4-1 Number of heterozygous microsatellite loci and the inferred mode of conception (MOC) for F0 and F1

F No. het F No het F 0 F MOC F mating status 1 F MOC 1 loci 0 0 loci 1

1 5 Sex Virgin 0 Parthenogenesis 2 5 Sex Virgin 0 Parthenogenesis 3 4 Sex Virgin 0 Parthenogenesis 4 4 Sex Virgin 0 Parthenogenesis 5 7 Sex Virgin 0 Parthenogenesis 6 7 Sex Virgin 0 Parthenogenesis 7 8 Sex Virgin 0 Parthenogenesis 8 8 Sex Virgin 0 Parthenogenesis 9 3 Sex Mated 3 Sex 10 3 Sex Mated 3 Sex 11 6 Sex Virgin 0 Parthenogenesis 12 6 Sex Virgin 0 Parthenogenesis 13 0 Parthenogenesis Mated 5 Sex 14 0 Parthenogenesis Mated 5 Sex 15 0 Parthenogenesis Mated 3 Sex 16 0 Parthenogenesis Mated 7 Sex 17 0 Parthenogenesis Mated 0 Parthenogenesis 18 0 Parthenogenesis Mated 0 Parthenogenesis 19 0 Parthenogenesis Mated 0 Parthenogenesis 20 3 Sex Mated 8 Sex 21 3 Sex Mated 7 Sex 22 2 Sex Mated 6 Sex

100 Table 4-2 Average haemocyte concentration (per ul), lytic activity (Δ absorbance) and PO activity (Δ absorbance) in F1 females with either sexual or parthenogenetic MOCs, derived from F0 with either sexual or parthenogenetic MOCs

Haemocyte F MOC F mate status F MOC N Lytic activity PO activity 0 0 1 concentration

Sex Virgin Parthenogenesis 10 4574 ± 512 0.20 ± 0.03 0.94 ± 0.07

Sex Mated Sex 4 3750 ± 593 0.10 ± 0.01 1.04 ± 0.09

Parthenogenesis Mated Parthenogenesis 3 2547 ± 359 0.05 ± 0.01 0.82 ± 0.02

Parthenogenesis Mated Sex 4 2173 ± 349 0.11 ± 0.03 0.92 ± 0.05

101 Table 4-3 The effect of F0 and F1 mode of conception (MOC) and week of sampling (3-, 5- and 7- weeks) on a) Log haemocyte concentration, b) Lytic activity and c) PO activity

Model Parameters β ± SE DF Statistic Probability

a) Log haemocyte concentration

F MOC 1 F = 0.43 0.04 0 2, 61 Week 2 F = 10.44 0.0002 3, 60

b) Log lytic activity (Δ absorbance)

F 0 MOC 1 F 2, 61 = 5.25 0.03 F1 MOC 1 F2, 61 = 0.06 0.81 Week 2 F3, 60 = 8.14 0.001 F0 MOC × F1 MOC 1 F4, 59 = 6.61 0.02

c) PO activity (Δ absorbance)

F MOC 1 F = 0.46 0.50 0 2, 61 F MOC 1 F = 1.56 0.22 1 2, 61 Week 2 F = 1.32 0.28 3, 60 Log haemocyte concentration 0.31 ± 0.1 1 F1, 62 = 10.22 0.002 F MOC × Week 1 F = 5.42 0.009 0 4, 59 F1 MOC × Week 1 F4 , 59 = 3.77 0.03

102

Figure 4-1 The interaction between F0 and F1 MOCs on logarithm of a) haemocyte concentration and b) lytic activity in F1 female E. tiaratum; levels not connected by same letters are significantly different based on post hoc Tukey’s test.

103

Figure 4-2 The effect of week of sampling on average logarithm of (a) haemocyte concentration and (b) lytic activity in the F1 E. tiaratum offspring (across treatments)

104

Figure 4-3 The interaction between week of sampling, and (a) F0 and (b) F1 modes of conception, on average PO activity of F1 females; continuous lines represent sexual and dashed lines represent parthenogenetic modes of conception. Levels not connected by same letters are significantly different based on post hoc Tukey’s test.

105 Discussion

This study has two main findings; first, parthenogenesis may have a negative trans- generational effect on at least two components of immune function in E. tiaratum

(haemocyte concentration and lytic activity). Mode of conception (sexual or parthenogenetic) of a female is related to her daughter’s ability to up regulate immune function, and this effect is likely a result of the genetic consequences of reproductive modes, rather than the effect of copulation per se (i.e. the effect of seminal fluid components). Second, haemocyte concentration and lytic activity of E. tiaratum decline with multiple sampling, perhaps a consequence of the physiological costs of mounting an immune response in close time intervals.

Assuming that up regulation of the immune function translates into increased fitness, sexual reproduction has a potential adaptive advantage compared to parthenogenesis in terms of immune response in E. tiaratum. As haemocyte concentration is the core cellular response in innate immunity, and lytic activity determines the ability to resist bacterial infection, individuals with higher levels of these measures are generally expected to have better immune function (Adamo 2004). Immune function can directly influence reproductive success if it results in increased adult lifespan (Rolf 2002). Because female

E. tiaratum produce eggs continuously (on average 17.2 ± 0.94 eggs per week, N = 29 females, YA unpublished data) over their 226 ± 12 day lifespan (N = 46 females, YA

106 unpublished data), lifetime reproductive success is a function of lifespan. Being sexually conceived is thus expected to provide females with a competitive advantage over being conceived via parthenogenesis. It is noticeable however, that increased immune function may trade-off with other physiologically costly process such as egg production, and subsequently influence reproductive fitness (Schwenke et al 2016). Therefore, a higher immune response may not necessarily correlate with higher reproductive success in E. tiaratum.

Copulation affects female immune response in various insect species (see Morrow and

Innocenti 2012). Mated females may upregulate genes involved in immune response as a result of exposure to sperm (Morrow and Innocenti 2012) and seminal fluid components

(Rolf 2002), or pathogens transferred during copulation (Lawniczak et al 2007). Such effects may influence the offspring immune function as non-genetic parental effects

(Grindstaff et al 2003). However, the association between reproductive modes and immune function in E. tiaratum is likely independent of the effect of either seminal fluid components or pathogens, as the parthenogenetically conceived offspring produced by mated females had lower baseline lytic activity compared to the offspring conceived sexually. Alternatively, this pattern may be a result of the genetic component of reproduction, potentially indicating a selective advantage to genetic mixing.

The mechanism underlying the effect of parthenogenesis on immune function in E.

107 tiaratum is unclear. One possibility is an indirect link through a general reduction in body condition, which can affect immune function pathways. Parthenogenesis results in substantial loss of microsatellite heterozygosity, in E. tiaratum, which may reflect genome wide heterozygosity (Väli et al 2008). Accordingly, parthenogenetically conceived offspring may exhibit reduced body condition, either through the direct effect of increased homozygosity at fitness related loci, or indirectly because of disruption to other physiological processes resulting in lower body condition (Charlesworth and

Charlesworth 1999, Tomkins et al 2004, Drayton and Jennions 2011). Lower body condition can affect costly physiological processes including immune function (e.g.

Ahtiainen et al 2004, Rantala and Kortet 2004) and more specifically the production of haemocyte cells, lytic enzymes and phenoloxidase. The positive correlation between a female’s heterozygosity proportion and her PO activity is to our knowledge shown for the first time in this study. The mechanism by which the negative effect of parthenogenesis arises following two consecutive generations of parthenogenesis remains to be investigated.

Age related declines in immune function are not uncommon in invertebrates (e.g.

Zerofsky et al 2005, Beckage 2008, Siva-Jothy et al 2005), but may not necessarily explain the decline in haemocyte concentration and lytic activity in this study. The four- week sampling period (3-7 weeks post final moult) represents a relatively short (< 15%) time interval for female E. tiaratum with average adult lifespan of 226 ± 12 days (N = 46,

YA unpublished data). Therefore, the gradual decrease in haemocyte concentration and

108 lytic activity is unlikely to be caused by the effect of aging in such a short proportion of female lifespan. Instead, this pattern suggests that the haemolymph sampling procedure may elicit an immune challenge. Considering that mounting an immune response is physiologically costly in insects (see Schwenke et al 2016) and the diet of E. tiaratum is mostly comprised of Eucalypt leaves with low nutritional value (Moore et al 2004), this reduction may indicate that the insects are physiologically constrained in keeping a high immune response in short time intervals.

The current study explores a rarely studied aspect of the relationship between sexual reproduction and the ability to resist diseases, and suggests a short-term advantage to sexual reproduction in terms of general immunity, in addition to the previously described long-term advantages in co-evolving with specific pathogens/ parasites (Van Valen 1973

Jaenike 1978, Hamilton 1980). Further, the importance of trans-generational investigations of genetic effects on immune function was highlighted, suggesting some of the inconsistencies in the effect of inbreeding on immune function (e.g. Rantala and Roff

2007 negative correlations, Stevens et al 1997, Gerloff et al 2003 no effects and Gershman et al 2010, Drayton and Jennions 2011, Franke and Fischer 2013 positive correlations) may be resolved by trans-generational studies.

109 References

Adamo, S. A. (2004). Estimating disease resistance in insects: phenoloxidase and lysozyme-like activity and disease resistance in the cricket Gryllus texensis. Journal of Insect Physiology, 50(2), 209–216. Ahtiainen, J. J., Alatalo, R. V, Kortet, R., and Rantala, M. J. (2006). Immune function, dominance and mating success in drumming male wolf spiders Hygrolycosa rubrofasciata. Behavioral Ecology and Sociobiology, 60(6), 826–832. Alavi, Y., van Rooyen, A., Elgar, M. A., Jones, T. M., and Weeks, A. R. (2016). Novel microsatellite markers suggest the mechanism of parthenogenesis in Extatosoma tiaratum is automixis with terminal fusion. Insect Science, in press. Beckage, N. E. (2008). Parasitoid polydnaviruses and insect immunity. Insect Immunology, 243–270. Booth, W., Smith, C. F., Eskridge, P. H., Hoss, S. K., Mendelson, J. R., and Schuett, G. W. (2012). Facultative parthenogenesis discovered in wild vertebrates. Biology Letters, 8(6), 983–5. Carlberg, U. (1983). Copulation in Extatosoma tiaratum (MacLeay)(Insecta: Phasmida). Zoologischer Anzeiger. Carson, H. L. (1967). Selection for parthenogenesis in Drosophila mercatorum. Genetics, 55(1), 157–71. Charlesworth, B., and Charlesworth, D. (1999). The genetic basis of inbreeding depression. Genetical Research, 74(3), 329–340. Corley, L. S., Blankenship, J. R., and Moore, A. J. (2001). Genetic variation and asexual reproduction in the facultatively parthenogenetic cockroach Nauphoeta cinerea: implications for the evolution of sex. Journal of Evolutionary Biology, 14(1), 68–74. Decaestecker, E., Gaba, S., Raeymaekers, J. A. M., Stoks, R., Van Kerckhoven, L., Ebert, D., and De Meester, L. (2007). Host–parasite “Red Queen”dynamics archived in pond sediment. Nature, 450(7171), 870–873. JOUR. Drayton, J. M., and Jennions, M. D. (2011). Inbreeding and measures of immune function in the cricket Teleogryllus commodus. Behavioral Ecology, 22(3), 486–492. Franke, K., and Fischer, K. (2013). Effects of inbreeding and temperature stress on life history and immune function in a butterfly. Journal of Evolutionary Biology, 26(3), 517–528.

110 Gerloff, C. U., Ottmer, B. K., and Schmid Hempel, P. (2003). Effects of inbreeding on immune response and body size in a social insect, Bombus terrestris. Functional Ecology, 17(5), 582–589. Gerritsen, J. (1980). Sex and parthenogenesis in sparse populations. American Naturalist, 718–742. Gershman, S. N., Barnett, C. A., Pettinger, A. M., Weddle, C. B., Hunt, J., and Sakaluk, S. K. (2010). Inbred decorated crickets exhibit higher measures of macroparasitic immunity than outbred individuals. Heredity, 105(3), 282–289. Grindstaff, J. L., Brodie, E. D., and Ketterson, E. D. (2003). Immune function across generations: integrating mechanism and evolutionary process in maternal antibody transmission. Proceedings of the Royal Society of London B: Biological Sciences, 270(1531), 2309–2319. Hamilton, W. D. (1980). Sex versus non-sex versus parasite. Oikos, 282–290. Hong, Z. D., and Ando, Y. (1998). Parthenogenesis in three species of genus Oxya. Japanese Journal of Applied Entomology and Zoology (Japan), (42), 65–69. Jaenike, J. (1978). An hypothesis to account for the maintenance of sex within populations. Evol. Theory, 3, 191–194. Kanost, M. R., and Gorman, M. J. (2008). Phenoloxidases in insect immunity. Insect Immunology, 1, 69–96. Kerstes, N. A. G., Bérénos, C., Schmid-Hempel, P., and Wegner, K. M. (2012). Antagonistic experimental coevolution with a parasite increases host recombination frequency. BMC Evolutionary Biology, 12(1), 18. King, K. C., Delph, L. F., Jokela, J., and Lively, C. M. (2009). The geographic mosaic of sex and the Red Queen. Current Biology, 19(17), 1438–1441. Koskella, B., and Lively, C. M. (2009). Evidence for negative frequency-dependent selection during experimental coevolution of a freshwater snail and a sterilizing trematode. Evolution, 63(9), 2213–2221. Kumpulainen, T., Grapputo, A., and Mappes, J. (2004). Parasites and sexual reproduction in psychid moths. Evolution, 58(7), 1511–1520. Kurtz, J. (2005). Specific memory within innate immune systems. Trends in Immunology, 26(4), 186–192. Lamb, R. Y., and Willey, R. B. (1979). Are parthenogenetic and related bisexual insects equal in fertility? Evolution, 33, 774–775. Lawniczak, M. K. N., Barnes, A. I., Linklater, J. R., Boone, J. M., Wigby, S., and

111 Chapman, T. (2007). Mating and immunity in invertebrates. Trends in Ecology and Evolution, 22(1), 48–55. Little, T. J., Hultmark, D., and Read, A. F. (2005). Invertebrate immunity and the limits of mechanistic immunology. Nature Immunology, 6(7), 651–654. Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by parasitism. Nature, 328(6130), 519–521. Matsuura, K., Fujimoto, M., and Goka, K. (2004). Sexual and asexual colony foundation and the mechanism of facultative parthenogenesis in the termite Reticulitermes speratus (Isoptera, Rhinotermitidae). Insectes Sociaux, 51(4), 325–332. Matsuura, K., and Kobayashi, N. (2007). Size, hatching rate, and hatching period of sexually and asexually produced eggs in the facultatively parthenogenetic termite Reticulitermes speratus (Isoptera: Rhinotermitidae). Applied Entomology and Zoology, 42(2), 241–246. Moore, B. D., Wallis, I. R., Wood, J. T., and Foley, W. J. (2004). Foliar nutrition, site quality, and temperature influence foliar chemistry of tallowwood (Eucalyptus microcorys). Ecological Monographs, 74(4), 553–568. Moreno-Garcia, M., Cordoba-Aguilar, A., Conde, R., and Lanz-Mendoza, H. (2013). Current immunity markers in insect ecological immunology: assumed trade-offs and methodological issues. Bull Entomol Res, 103(2), 127–139. Morran, L. T., Schmidt, O. G., Gelarden, I. A., Parrish, R. C., and Lively, C. M. (2011). Running with the Red Queen: Host-Parasite Coevolution Selects for Biparental Sex. Science, 333(6039), 216 LP-218. Morrow, E. H., and Innocenti, P. (2012). Female postmating immune responses, immune system evolution and immunogenic males. Biological Reviews, 87(3), 631–638. Rantala, M. J., Jokinen, I., Kortet, R., Vainikka, A., and Suhonen, J. (2002). Do pheromones reveal male immunocompetence? Proceedings of the Royal Society of London B: Biological Sciences, 269(1501), 1681–1685. Rantala, M. J., and Kortet, R. (2004). Male dominance and immunocompetence in a field cricket. Behavioral Ecology, 15(2), 187–191. Rantala, M. J., and Roff, D. A. (2007). Inbreeding and extreme outbreeding cause sex differences in immune defence and life history traits in Epirrita autumnata. Heredity, 98(5), 329–336. Ribeiro, C., and Brehélin, M. (2006). Insect haemocytes: What type of cell is that? Journal of Insect Physiology, 52(5), 417–429. Rolff, J. (2002). Bateman’s principle and immunity. Proceedings of the Royal Society of

112 London B: Biological Sciences, 269(1493), 867–872. Schwenke, R. A., Lazzaro, B. P., and Wolfner, M. F. (2016). Reproduction-immunity trade-offs in insects. Annual Review of Entomology, 61, 239–256. Sekiné, K., and Tojo, K. (2009). Automictic parthenogenesis of a geographically parthenogenetic mayfly, Ephoron shigae (Insecta: Ephemeroptera, Polymitarcyidae). Biological Journal of the Linnean Society, 99(2), 335–343. Siva-Jothy, M. T., Moret, Y., and Rolff, J. (2005). Advances in Insect Physiology Volume 32. Advances in Insect Physiology (Vol. 32). Elsevier. Stalker, H. D. (1956). On the evolution of parthenogenesis in Lonchoptera (Diptera). Evolution, 345–359. Stevens, L., Yan, G., and Pray, L. A. (1997). Consequences of inbreeding on invertebrate host susceptibility to parasitic infection. Evolution, 51(6), 2032–2039. Tomkins, J. L., Radwan, J., Kotiaho, J. S., and Tregenza, T. (2004). Genic capture and resolving the lek paradox. Trends in Ecology and Evolution, 19(6), 323–328. Van Valen, L. (1977). The red queen. The American Naturalist, 111(980), 809–810. Zerofsky, M., Harel, E., Silverman, N., and Tatar, M. (2005). Aging of the innate immune response in Drosophila melanogaster. Aging Cell, 4(2), 103–108.

113

Chapter 5

Sex versus parthenogenesis; female reproductive success in a facultatively parthenogenetic Phasmatid

(Extatosoma tiaratum)

114 Summary

The evolution of sexual reproduction remains a significant challenge for evolutionary biology because, in theory, sex is far more costly than asexual reproduction. While numerous theoretical and empirical studies have explored the relative costs and benefits of sex, the issue remains unresolved. One major empirical challenge is to find model taxa that allow minimizing the confounding effects that are associated with, but not directly related to, reproductive modes. Facultative parthenogenetic species in which unfertilised eggs develop into female offspring provide promising models. We used the facultative parthenogenetic Australian phasmatid, Extatosoma tiaratum, to compare the effect of sex and parthenogenesis on female fitness. We manipulated the mating status (virgin or mated) of females with known reproductive histories, and investigated the effect of trans- generational reproductive modes on life history traits and reproductive success. We show that life history traits and fecundity are dependent on a female’s mode of conception, while offspring survival depends on a female’s mating status: sexually conceived females are typically larger, produce more eggs and live longer than parthenogenetically conceived females. Offspring survival is higher for mated females compared to virgins. Furthermore, a female’s maternal mode of conception has little influence on her reproductive success.

Our results show that parthenogenesis is associated with fitness costs in E. tiaratum, but these costs can be recovered in the next generation, if parthenogenetically conceived females produce daughters sexually.

115 Introduction

The predominance of sexual reproduction in nature has long been a paradox in evolutionary biology, because in theory, sexual reproduction is far more costly compared with asexual reproduction. As only females can produce offspring, ceteris paribus (all else being equal), asexual females have a two-fold advantage by producing daughters only, compared with their sexual counterparts which produce both sons and daughters (Maynard

Smith 1978, Bell 1982). Additionally, there are other costs of sex, including breaking up beneficial gene combinations (Nei 1967, Lynch and Deng 1994), search costs associated with obtaining a partner (Gascoigne et al 2009), and an increased risk of predation and/or infection (Arnqvist and Nilsson 2000). Accordingly, asexual females should theoretically win when in direct competition against sexual females. Nevertheless, most animals and plants reproduce sexually, and parthenogenetic taxa generally occupy terminal nodes in phylogenetic trees (Suomalainen 1987, Simon et al 2003).

Theory describes various conditions under which sexual reproduction is beneficial. For example, sex can be advantageous in reducing the accumulation of deleterious mutations

(Muller’s ratchet hypothesis, Muller 1964); during adaptation to novel environments (Hill-

Robertson interference, Hill and Robertson 1966); or, when encountering co-evolving pathogens (Jaenike 1978, Hamilton 1980). Recently, the benefits of sex have been categorized according to the benefits of genetic mixing, such as recombination and

116 segregation (Otto 2009, Roze 2012). There is some empirical support for these ideas in particular species (mutation accumulation: Normark and Moran 2000, Bruggeman et al

2003, Paland and Lynch 2006; adaptation to novel environments: Wolf et al 1987, Kaltz and Bell 2002, Becks and Agrawal 2012; and, pathogen resistance: Lively 1987, 1990

Kumpulainen et al 2004, Decaestecker et al 2007, King et al 2009, Morran et al 2011).

However, evidence for the benefits of sex under more general conditions remains largely elusive.

Asexual reproduction is not free of costs. Parthenogenesis, the development of unfertilised eggs and the most common form of asexual reproduction in the animal kingdom, has fitness costs that often violate at least part of the ceteris paribus assumption of the two- fold cost of sex. In insects, parthenogenesis results in reduced hatch rate and survival compared with sexual reproduction (Stalker 1956, Carson 1967, Lamb and Willey 1979,

Hong and Ando 1998, Corley et al 2001, Kramer and Templeton 2001, Matsuura et al

2004). For these species, the maintenance of parthenogenesis represents another unsolved paradox (Templeton 1983).

One of the main challenges for empirical studies that investigate the relative costs and benefits of sexual versus parthenogenetic reproduction is to resolve potentially confounding effects. For example, parthenogenetic populations may be a hybrid of multiple sexual populations (see Kearney 2005), or they may occupy different spatial or

117 temporal spaces compared to sexual populations (e.g. Lynch 1984, Hadany and Otto

2009). Cyclic parthenogenetic species capable of both sexual and asexual reproduction

(e.g. rotifers; Becks and Agrawal 2010, Daphnia; Paland and Lynch 2006, Decaestecker et al 2009, aphids; Bulmer 1982, Simon et al 2002) provide promising models, but many of these species alter their reproductive modes according to the local environmental conditions (e.g. sexual reproduction and production of resting eggs in winter).

Furthermore, most of these species only retain a single reproductive mode (either sexual or asexual) within a single generation, rather than a female having the capacity to alternate between asexual and sexual modes (repeatedly) within her lifespan.

Facultative parthenogenetic species, in which females can produce both fertilised and unfertilised, but viable eggs, provide an opportunity to evaluate the outcomes of sex and parthenogenesis, without introducing typical confounding factors. In these species, the difference between the two reproductive modes is directly (and likely exclusively) associated with the effects of genetic mixing (recombination and segregation) and the presence of seminal fluid components. While widespread in the animal kingdom (see

Booth et al 2012), the evolutionary significance of facultative parthenogenesis is poorly understood. Traditionally, it is thought to occur in environments where sex is difficult or even impossible (Stalker 1956), often arising as a result of stochastic fluctuations in population density resulting in female-biased sex ratios and a general absence of males

(Stalker 1956, Schwander et al 2010). However, even in the presence of males, limitation

118 in delivering sufficient sperm for fertilisation may also be a potential selective force

(Elzinga et al, 2011, Chapter 2; Alavi et al 2016).

In this study, we investigated the effect of sexual versus parthenogenetic reproduction on female reproductive success in the facultatively parthenogenetic Australian phasmatid,

Extatosoma tiaratum. This species occurs in rain-forested areas of New South Wales and south eastern parts of Queensland, Australia (Brock and Hasenpusch 2009). Adult females commence oviposition at approximately four weeks following their final juvenile moult, and produce eggs continuously throughout their adult lifespan, irrespective of whether they mate or remain virgin (Carlberg 1983). Laboratory experiments demonstrated that virgin females produce pheromones to attract males and are capable of adjusting their oviposition rate according to the perceived abundance of males during the juvenile stages of their development, presumed to be a response to anticipated future mating opportunities

(Schneider and Elgar 2010).

This species thus provides a powerful model to investigate the trans-generational effects of alternating different reproductive modes, because it is possible to manipulate the mode of conception of both a focal female (determined by whether her mother mates or not) and her offspring’s mode of conception (determined by whether she mates or not).

Additionally, it is possible to infer the reproductive background of females of unknown origin in E. tiaratum, because the cytological mechanism of parthenogenesis (automixis with terminal fusion) results in substantial loss of heterozygosity (Chapter 3; Alavi et al

119 2016). Using females from a captive-bred stock population, we first took sister pairs derived from either sex or parthenogenesis, and randomly assigned each sister to either a virgin or mated treatment. This simple experimental design allowed us to document the survival and reproductive success associated with different modes of conception and mating status under similar genetic backgrounds. We confirmed that the rates of parthenogenesis in our captive population was not artificially elevated, by comparing the hatching frequency of eggs produced by virgin wild caught females with that of virgin captive bred females.

120 Methods

Animal maintenance

A stock population of E. tiaratum was established from eggs and juveniles obtained from the Melbourne Museum and local breeders in Victoria in 2012. All adult females were housed in cylindrical containers (height: 23 cm; diameter: 25 cm), enclosed by a fine mesh. All nymphs, juveniles and adult males were housed in plastic containers (30 ×20

×18 cm). All eggs were incubated in plastic boxes (16 ×10 ×5 cm) with 5 mm of sand and sprayed with a fine mist of water weekly. All insect containers and egg boxes were kept in climate-controlled conditions (24-26°C; 50% humidity; 12:12hr Light: Dark cycle), and individuals were given ad libitum misted leaves of various species of Eucalyptus.

Deriving F1 from F0 females

Juvenile environment influences female reproductive investment in E. tiaratum (Schneider and Elgar 2010). Because the females in our stock population (hereafter defined as F0 females) were from various sources, their juvenile rearing condition, such as nutrition and sex ratio, were likely different. In addition, we did not know the mode of conception

(MOC) of F0 females at the start of the experiment, before assigning them to mating treatments. Therefore, we derived first generation offspring (hereafter defined as F1) from

121 F0 and subsequently assigned them to either virgin or mated treatment based on their modes of conception in a balanced crossed design. This protocol allowed us to minimize potentially confounding effects associated with maternal rearing conditions and reproductive history.

Adult F0 females (yet to commence oviposition) from the stock population were assigned to either a virgin (n = 5 females) or the mated (n = 9 females) treatment before they initiated oviposition (Figure 5-1). Each female was transferred to an individual cylindrical container (as above), where she was held for the duration of her adult life. Approximately one month after commencing oviposition, each F0 experimental female in the mated treatment was provided with a single adult male (introduced into her container). The male was maintained with the female one day per week until mating occurred and a spermatophore was visible. This procedure allowed us to separate the effect of female age from mating treatment by creating two distinct time periods: i) the first time period corresponded to the first two months of reproduction (following the commencement of oviposition) when all females were virgin, and, ii) the second time period corresponded to the third and fourth months of reproduction where females in the mated treatments had mated (Figure 5-2). Males were removed from the female cages after copulation, and eggs produced in the second time period were collected and incubated as described above.

Nymphs were collected upon hatching and a subset of females was reared to adult stages under similar laboratory conditions as above (N = 23, 10 nymphs were derived from 5 virgin F0 and 13 nymphs were derived from 9 mated F0).

122

F1

Virgin Virgin Juveniles F0 F Mated 2

Juveniles Juveniles

Virgin Mated Juveniles Hatching frequency Mated Moulting frequency Body size Development time Adult lifespan Weekly oviposition rate Total eggs

Figure 5-1 The experimental set up showing mating treatments of F0 and F1 females and data collected for F1 and F2 generations

123 Time period

1 - 2 3 - 4 F0 months months

Virgin Virgin Virgin n = 5

Mated n = 9 Virgin Mated

Time period

1st 2nd 3rd F1 month month month

Virgin Virgin Virgin Virgin n = 13

Mated n = 10 Virgin Mated Mated

Figure 5-2 Mating status of F0 (top) and F1 (bottom) females during each time period; a proportion of F1 females produced by F0 females in the second time period were used as F1 females (N = 23)

124 F1 fecundity

The 23 isolated adult F1 females were used to assess the relative impact of trans- generational reproductive modes (F0 and F1) on F1 body size, lifespan, fecundity and offspring survival and development time. F1 females were assigned to either a virgin (n

= 13) or mated (n = 10) treatment. Of the 23 F1 females, 18 were in sister pairs, each sister assigned to either virgin or mating treatment in a crossed design. The remaining five females were unrelated and were randomly assigned to virgin or mated mating treatments. Eggs were collected over three time periods for F1 females: during the first month of oviposition when all females were virgin; during the second month of oviposition when the mated treatment females had mated, and during the third month of oviposition, similar to the second month in terms of the mating treatments (Figure

5-2). The eggs were collected and counted weekly over the entire reproductive lifespan of F1 females. The eggs produced by each F1 female for each time period were incubated in separate plastic containers as described above.

F2 Juvenile survival

Eggs produced in the first three months of oviposition were used to estimate F2 hatching frequency (number of F2 nymphs hatched divided by the total number of eggs). The egg containers were checked daily and the number of hatched nymphs was recorded. A proportion of the nymphs (n = 216) hatched in the second time period

125 were used to estimate moulting frequency (number of nymphs moulted to second instar divided by total number of nymphs monitored). In addition, we recorded the development time (number of days elapsed from the second instar to the final moult) for a proportion of F2 males (n = 13), and females (n = 17) derived from virgin or mated F1 females. The nymphs and juveniles derived from each F1 female were maintained in separate containers (two replicate containers per F1 female) with similar amounts of food and water, and at constant density of 5 - 10 individuals per container.

F1 tarsus length was measured after natural death and used as a measure of body size.

Offspring survival of field- collected females

Artificial selection can increase rates of parthenogenetic egg development (Templeton et al 1976). Captive breeding is likely to elevate rates of parthenogenetic egg development, as a result of female-biased sex ratios. To investigate this possibility, we compared juvenile survival of the progeny derived from field-collected females

(hereafter defined as FieldF0) to that of the stock females (F1). Males and females were collected as juveniles from a natural population in Crystal Cascades near Cairns (north

Queensland, Australia; 16°57′42″S 145°40′46″E) during December 2014. Juvenile

FieldF0 were transferred to the laboratory within two days where they were raised under standard conditions (see above) to adulthood. All females were virgins during the first month of oviposition (n = 7). In the second month of oviposition, each female was placed together with a virgin male, who had also been collected in the field (n =

7), until mating occurred and the spermatophore was transferred. Eggs were collected

126 before and after mating and hatching and moulting frequency of the nymphs (hereafter defined as FieldF1) were recorded as described above.

Microsatellite heterozygosity and modes of conception

Mating may not necessarily result in the production of sexual offspring in facultative parthenogenetic species, either because the sperm failed to transfer, or because females did not utilize the sperm. We used previously developed microsatellite markers

(Chapter 3; Alavi et al 2016) to investigate the microsatellite heterozygosity proportion and infer the mode of conception for for all F0 (n = 14 females), F1 (n = 23) and field- collected females (n = 7). In addition, we genotyped a proportion of F2 female offspring (3 - 5 offspring per female) derived from all mated F1 and FieldF0 females to investigate patterns of sperm utility in mated females.

The cytological mechanism of parthenogenesis (automixis with terminal fusion) in E. tiaratum results in substantial loss of microsatellite heterozygosity, with 14 out of 15 developed markers usually homozygous in parthenogenetically produced females

(Alavi et al 2016). We genotyped the 15 microsatellite markers and inferred the mode of conception based on the microsatellite heterozygosity proportion (the proportion of heterozygous to homozygous microsatellite loci) following methods described previously (Alavi et al 2016). If microsatellite heterozygosity proportion is close to zero, the mode of conception is most likely parthenogenetic and if the proportion is

127 more than 0.1, the mode of conception is sexual (less than 0.1 = parthenogenetic MOC i.e. fusion of two egg cells; > 0.1 = sexual MOC i.e. fusion of egg cell and sperm, see

Pearcy et al 2006).

Statistical analysis

All analyses were performed using JMP version 12.0.0 (SAS institute, NC, USA). We determined the effect of reproductive treatments on life history and reproductive success in separate general linear models (F1 weekly oviposition rate and total eggs, and F2 and FieldF1 hatching and moulting frequencies) and where data could not be transformed appropriately, non-parametric models (F1 body size and F2 development time). Adult lifespan of F1 females were analysed using Cox proportional hazard models (Kay 1977). F1 tarsus length was used as a measure of body size and female

(F0 and F1) identity was added as a random effect when appropriate. Models investigating hatching and moulting frequencies were weighted by the number of eggs per box, and the number of nymphs per container respectively. Models were first fitted using 2-way interactions and were subsequently reduced using hierarchical removal of any term with a significance of P > 0.1.

128 Results

Microsatellite heterozygosity proportions and modes of conception

The microsatellite heterozygosity data revealed that among the F0 females (N = 14), 8

(n = 5 allocated to the virgin and n = 3 to the mated treatment) had been conceived sexually (Table 5-1). The remaining 6, all allocated to the mated treatment, had been conceived parthenogenetically.

Among all mated females from the stock population (n = 9 F0 and 10 F1 females), seven females (n = 3 F0 and 4 F1) produced parthenogenetic offspring (Tables 5-1 and

5-2). Interestingly, all of the mated females that produced parthenogenetic daughters, were conceived parthenogenetically (n = 12 parthenogenetically conceived mated females, out of which 7 produced parthenogenetic offspring).

All field-collected females (FieldF0) had been conceived sexually and produced offspring with sexual modes of conception following mating (Table 5-3).

129 Table 5-1 Number of heterozygous microsatellite loci and the inferred mode of conception (MOC) for F0 and F1 females. Three out of five F0 females that had been conceived parthenogenetically, produced progeny with parthenogenetic MOCs (shown in bold) following mating.

F No. het F mating F No. het F F 0 F MOC 0 1 F MOC 1 0 loci 0 status loci 1

F1-1 F0-1 5 Sex Virgin 0 Parthenogenesis F1-2 F0-1 5 Sex Virgin 0 Parthenogenesis F1-3 F0-2 4 Sex Virgin 0 Parthenogenesis F1-4 F0-2 4 Sex Virgin 0 Parthenogenesis F1-5 F0-3 7 Sex Virgin 0 Parthenogenesis F1-6 F0-3 7 Sex Virgin 0 Parthenogenesis F1-7 F0-4 8 Sex Virgin 0 Parthenogenesis F1-8 F0-4 8 Sex Virgin 0 Parthenogenesis F1-9 F0-5 3 Sex Mated 2 Sex F1-10 F0-5 3 Sex Mated 4 Sex F1-11 F0-6 6 Sex Virgin 0 Parthenogenesis F1-12 F0-6 6 Sex Virgin 0 Parthenogenesis F1-13 F0-7 0 Parthenogenesis Mated 3 Sex F1-14 F0-7 0 Parthenogenesis Mated 3 Sex F1-15 F0-8 0 Parthenogenesis Mated 4 Sex F1-16 F0-8 0 Parthenogenesis Mated 5 Sex Parthenogenesi F1-17 F0-9 0 Parthenogenesis Mated 0 s F1-18 F0-10 0 Parthenogenesis Mated 6 Sex Parthenogenesi F1-19 F0-11 0 Parthenogenesis Mated 0 s Parthenogenesi F1-20 F0-12 0 Parthenogenesis Mated 0 s F1-21 F0-13 3 Sex Mated 3 Sex F1-22 F0-13 3 Sex Mated 4 Sex F1-23 F0-14 2 Sex Mated 5 Sex

130 Table 5-2 Number of heterozygous microsatellite loci and the inferred mode of conception (MOC) of F2 female progeny, derived from mated F1 females. All females with parthenogenetic MOCs, produced some offspring with parthenogenetic MOCs (shown in bold).

F mating F No. het F F F MOC 1 2 F MOC 2 1 1 status loci 2

F2-1 F1-2 Parthenogenetic Mated 6 Sex F2-2 F1-2 Parthenogenetic Mated 0 Parthenogenesis F2-3 F1-2 Parthenogenetic Mated 0 Parthenogenesis F2-4 F1-2 Parthenogenetic Mated 0 Parthenogenesis F2-5 F1-2 Parthenogenetic Mated 0 Parthenogenesis F2-6 F1-3 Parthenogenetic Mated 0 Parthenogenesis F2-7 F1-3 Parthenogenetic Mated 0 Parthenogenesis F2-8 F1-3 Parthenogenetic Mated 0 Parthenogenesis F2-9 F1-3 Parthenogenetic Mated 0 Parthenogenesis F2-10 F1-3 Parthenogenetic Mated 0 Parthenogenesis F2-11 F1-6 Parthenogenetic Mated 0 Parthenogenesis F2-12 F1-6 Parthenogenetic Mated 0 Parthenogenesis F2-13 F1-6 Parthenogenetic Mated 0 Parthenogenesis F2-14 F1-6 Parthenogenetic Mated 0 Parthenogenesis F2-15 F1-6 Parthenogenetic Mated 0 Parthenogenesis F2-16 F1-8 Parthenogenetic Mated 4 Sex F2-17 F1-8 Parthenogenetic Mated 0 Parthenogenesis F2-18 F1-8 Parthenogenetic Mated 0 Parthenogenesis F2-19 F1-8 Parthenogenetic Mated 5 Sex F2-20 F1-9 Sex Mated 4 Sex F2-21 F1-9 Sex Mated 6 Sex F2-22 F1-9 Sex Mated 5 Sex F2-23 F1-12 Parthenogenetic Mated 5 Sex F2-24 F1-12 Parthenogenetic Mated 4 Sex F2-25 F1-12 Parthenogenetic Mated 5 Sex F2-26 F1-15 Sex Mated 5 Sex F2-27 F1-15 Sex Mated 3 Sex F2-28 F1-15 Sex Mated 5 Sex F2-29 F1-15 Sex Mated 5 Sex F2-30 F1-15 Sex Mated 6 Sex F2-31 F1-18 Sex Mated 7 Sex F2-32 F1-18 Sex Mated 6 Sex F2-33 F1-18 Sex Mated 4 Sex F2-34 F1-18 Sex Mated 5 Sex F2-35 F1-22 Sex Mated 6 Sex F2-36 F1-22 Sex Mated 6 Sex F2-37 F1-22 Sex Mated 6 Sex F2-38 F1-22 Sex Mated 5 Sex

131 Table 5-3 Number of heterozygous microsatellite loci and the inferred mode of conception (MOC) for field-collected females (FieldF0), and their female offspring (FieldF1) produced following mating. All FieldF0 females produced offspring with sexual MOCs following mating.

FF No. FF mating FF No. het FF FF FF 0 FF MOC 0 1 1 1 0 het loci 0 status loci MOC

FF1-1 FF0-1 2 Sex Mated 7 Sex FF1-2 FF0-2 2 Sex Mated 7 Sex FF1-3 FF0-3 2 Sex Mated 6 Sex FF1-4 FF0-4 2 Sex Mated 6 Sex FF1-5 FF0-5 2 Sex Mated 5 Sex FF1-6 FF0-2 5 Sex Mated 3 Sex FF1-7 FF0-3 5 Sex Mated 5 Sex FF1-8 FF0-4 5 Sex Mated 7 Sex FF1-9 FF0-5 5 Sex Mated 7 Sex FF1-10 FF0-3 6 Sex Mated 4 Sex FF1-11 FF0-4 6 Sex Mated 5 Sex FF1-12 FF0-5 6 Sex Mated 5 Sex FF1-13 FF0-6 6 Sex Mated 4 Sex FF1-14 FF0-4 8 Sex Mated 3 Sex FF1-15 FF0-5 8 Sex Mated 4 Sex FF1-16 FF0-6 8 Sex Mated 4 Sex FF1-17 FF0-7 8 Sex Mated 5 Sex FF1-18 FF0-8 8 Sex Mated 6 Sex FF1-19 FF0-5 7 Sex Mated 4 Sex FF1-20 FF0-6 7 Sex Mated 6 Sex FF1-21 FF0-7 7 Sex Mated 4 Sex FF1-22 FF0-8 7 Sex Mated 4 Sex FF1-23 FF0-9 7 Sex Mated 5 Sex

132 F1 body size, adult lifespan and fecundity

Body size – Body size was significantly influenced by mode of conception in F1 females (Table 5-4a). Females that were conceived sexually were typically larger

(tarsus length 26.72 ± 0.37mm) than females that were conceived parthenogenetically

(tarsus length 25.27 ± 0.28 mm, Figure 5-3a).

Adult lifespan – A proportional hazards survival analysis showed that F1 females conceived sexually survived significantly longer than those conceived parthenogenetically (Table 5-4b, Figure 5-3b).

Weekly oviposition rate – Weekly oviposition rate was significantly influenced by F1 mode of conception, but not week of oviposition or F1 mating status (Table 5-4c).

Over the first three months of oviposition, sexually conceived females produced more eggs weekly (29.67 ± 0.72 eggs) compared to parthenogenetic females (18.10 ± 0.68;

P < 0.0001; Figure 5-3c). There was a significant F1 mating status by week interaction, but there were no linear patterns to this relationship (Figure 5-4). F1 identity explained

23.86% of the variation in weekly oviposition rate.

Total number of eggs – Total number of eggs produced over entire F1 adult lifespan was significantly influenced by F1 body size and F1 mode of conception (Table 5-4d).

133 There was a positive relationship between female body size and total number of eggs produced (β = 1.14 ± 0.46). F1 females that were sexually conceived produced more eggs in total (865 ± 65.80) compared to parthenogenetically conceived F1 females (315

± 47.28 eggs; P < 0.0001; Table 5-4d). A post hoc Tukey’s tests investigating the effect of trans-generational reproductive modes (single factor with two levels: F0

MOC- F1 MOC) on F1 egg production showed a significant effect (F2, 21 = 18.53, P <

0.0001, Figure 5-5).

Table 5-4 The effect of F1 mode of conception (MOC) and F1 mating status (virgin or mated) on F1 : a) body size, b) lifespan, c) weekly oviposition rate, and d) total eggs produced

Model Parameters (F1) β ± SE DF Statistic Probability

a) Body size

2 F1 MOC 2 χ 2, 21 = 8.04 0.0047

b) Adult lifespan

2 F1 MOC 2 χ 2, 21 = 13.59 0.0002

c) Weekly oviposition rate

F1 MOC 2 F2, 320 = 28.15 < 0.0001 F1 mating status 2 F2, 320 = 2.91 0.10 Week -0.001 ± 0.005 1 F1, 321 = 2.28 0.13 F1 mating status × Week 0.01 ± 0.005 3 F3, 319 = 6.58 0.01

d) Total eggs

F1 MOC 2 F1, 22 = 32.4 0.0001 Body size 1.14 ± 0.46 1 F1, 22 = 6.08 0.02

134 Figure 5-3 The effect of mode of conception on life history trait measures: a) body size (tarsus length), b) adult lifespan, c) weekly egg number, of F1 females, and d) development time of F2 second instar females to adulthood

50

40

s

g

g 30

e

f Sex- Mated

o

r

e Sex- Virgin b 20 m Parth- Mated

u

N

10 Parth- Virgin

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Week

Figure 5-4 The weekly oviposition rate of F1 females grouped by F1 mode of conception

(sex/parthenogenesis) and F1 mating status (mated/virgin) – the shaded area indicates the time range in which females in the mated treatment mated (average mating latency from the start of oviposition = 36.10 ± 2.96 days, minimum 26 days, maximum 47 days)

135 1400 B A B A

1200

1000

800

600

400 Total eggs Totaleggs produced

200 n = 3 n = 5 n = 10 n = 5 0 PARTH-PARTH PARTH-SEX SEX-PARTH SEX-SEX

F0 MOC- F1 MOC

Figure 5-5 The effect of trans-generational reproductive mode (F0 MOC – F1 MOC) on the total number of eggs produced by F1 females; Levels not connected by same letters are significantly different based on post hoc Tukey’s test.

F2 Juvenile survival

F2 hatching frequency – Hatching frequency was significantly higher in the F2 progeny derived from mated (0.49 ± 0.17) than virgin F1 females (0.36 ± 0.17), but was not influenced by F0 or F1 modes of conception (Table 5-5a, Figure 5-6a).

F2 moulting frequency – Similar to hatching frequency, moulting frequency of F2 progeny was significantly higher in the progeny derived from mated (0.55 ± 0.08) than virgin F1 females (0.33 ± 0.07, Table 5-5b, Figure 5-6a).

136 F2 development time – Female nymphs that were conceived sexually, developed to adults significantly more quickly (155.16 ± 7 days) than those conceived parthenogenetically (174.45 ± 2.68 days, Figure 5-3d). The development time also depended upon progeny’s gender, with males developing more quickly than females

(males: 99.23 ± 3.43 days; females: 167.65 ± 3.69 days, Table 5-5c).

Table 5-5 The effect of F0 and F1 mode of conception (MOC), and F1 mating status (virgin or mated) on F2: a) hatching frequency, b) moulting frequency, and c) development time

Model Parameters (F2) DF Statistic Probability a) Hatching frequency

F0 MOC 2 F2, 68 = 3.24 0.09 F1 MOC 2 F2, 68 = 0.58 0.45 F1 mating status 2 F2, 68 = 14.83 0.001 F1 MOC × F1 mating status 4 F4, 64 = 3.24 0.09 b) Moulting frequency

F1 mating status 2 F2, 21 = 5.00 0.04 c) Development time

2 F1 MOC 2 χ 2, 17 = 5.16 0.02 2 F1 gender 2 χ 2, 28 = 21.42 0.0001

137 Offspring survival of field-collected females

Hatching frequency was comparable between eggs derived from FieldF0 before, and after mating (F2, 9 = 0.25, P = 0.62, Figure 5-6b). However, the proportion of nymphs that moulted into second instar larvae was significantly higher for progeny derived from mated (0.44 ± 0.08) than virgin females (0.17 ± 0.07, F2, 10 = 48.50, P = 0.002,

Figure 5-6b). Variance component estimates suggested that 39% of the variation in

FieldF1 moulting frequency is explained by FieldF1 identity. In a separate model, we investigated the relative hatching frequency (proportion of eggs hatched/ number of eggs) of eggs produced by virgin F1 (stock population) and FieldF0 (field-collected) females. There was a trend toward higher hatching frequency of FieldF0 eggs, but this difference was non- significant (F2, 16 = 3.20, P = 0.09).

138 a) F2

0.8 Mating status *** *

Virgin

0.6 Mated Proportion 0.4

0.2 Hatched Moulted

b) FieldF1

0.8 ***

0.6

Proportion 0.4

0.2 Hatched Moulted Survival

Figure 5-6 Mean proportion of hatched and moulted progeny derived from eggs produced by mated (grey) and virgin (white) F1 (a) and FieldF0 b) females; Asterisks indicate significant differences.

139 Discussion

This study has three main findings. First, parthenogenesis has substantial costs for female E. tiaratum, in terms of both life history and reproductive success; female body size, longevity, fecundity and offspring survival are all negatively influenced by parthenogenesis. Second, the reduction in fitness can be “rescued” by sex in the next generation; a female’s maternal mode of conception does not influence her body size, longevity, or reproductive success. Similarly, offspring survival is independent of a female’s mode of conception, and is determined by her mating status instead. Finally, we found evidence for a potential, indirect cost of parthenogenesis: the fertilisation rate of parthenogenetically conceived females was lower than that of sexually conceived females.

The crucial ceteris paribus assumption of the widely understood two-fold cost of sex

(Maynard Smith 1978; Bell 1982) was not supported by our experiments. Among females of similar genetic background and age, those conceived sexually, typically develop to adults 20 days earlier, survived twice as long and produced twice as many eggs as females conceived parthenogenetically. The reproductive success of sexually reproducing female E. tiaratum was substantially greater than that of parthenogenetically reproducing females, to the extent that it reduces and may even eliminate the two-fold cost of sex. These results are consistent with previous studies of parthenogenetic insect species, demonstrating that parthenogenesis negatively influences several reproductive life history traits, including body size, hatching

140 frequency, development time and longevity in stick insects (e.g. Salmon 1955,

Morgan-Richards et al 2010, Burke et al 2015) and other taxa including Blattodea

(Roth 1974, Corley et al 2001), Diptera (Carson et al 1967, Kramer and Templeton

2001), Ephemeroptera (Sekine and Tojo 2010) Hymenoptera (Matsuura et al 2004,

Matsuura and Kobayashi 2007) and Orthoptera (Hong and Ando 1998).

It is notable that the laboratory conditions were potentially benign in terms of predation and infection risk. Thus, the cost of parthenogenesis in terms of egg production and offspring survival may potentially be greater under natural conditions.

Hatching may be particularly prone to failure under more variable natural conditions, since moisture affects egg development rate (Carlberg 1991). Nevertheless, the survival of offspring of virgin females from the laboratory populations and field- collected populations was remarkably comparable, indicating that the rate of parthenogenetic egg development is not un-naturally elevated by captive breeding.

It is not clear why parthenogenesis results in reduced reproductive success. We cannot separate the effects of the mode of conception and seminal fluids on reproductive success in E. tiaratum because mated females can reproduce parthenogenetically (and indeed did in this experiment). In other words, the average higher survival of the progeny derived from mated females compared to virgin may arise because the F2 individuals had sexual modes of conception, and/or because all the progeny benefited from the seminal fluid delivered by the male. Parthenogenesis results in substantial

141 loss of microsatellite heterozygosity, which may reflect loss of heterozygosity at fitness related loci (Väli et al 2008), thereby increasing genetic load due to the expression of deleterious recessive alleles (heterozygosity advantage, Charlesworth and Charlesworth 1999, Roff 2002). Comparable correlations of heterozygosity levels with various life history, morphological and physiological traits have been documented in various taxa (see Chapman et al, 2009).

In addition to direct costs of parthenogenesis on female reproductive success, there may be indirect costs for E. tiaratum females in terms of reduced ability to fertilise eggs. More than half of the mated females that were themselves conceived parthenogenetically, produced parthenogenetic offspring, despite the presumed presence of sperm (as determined by the successful transfer of a spermatophore). Our experimental approach was unable to detect whether the females failed to fertilise eggs because there was insufficient sperm, or whether they instead exercised post- copulatory mate choice (Eberhard 1996). The former is possible because male E. tiaratum appear to be limited in their ability to transfer ejaculate (Alavi et al 2016).

Another possibility is that parthenogenetically conceived females are limited in their ability to obtain, store or utilize sperm, as a result of loss of heterozygosity and/or directly influenced by parthenogenesis itself (see Kramer et al 2002). The latter in particular is an intriguing possibility that remains to be tested.

142 Our results highlight the importance of considering simultaneously a female’s mode of conception and her mating status, when assessing variation in female reproductive success. We showed that life history traits including body size, fecundity and longevity reflect females’ mode of conception rather than their mating status. This likely explains the differences between our results with a previous study exploring the maintenance of parthenogenesis in E. tiaratum. In their study, Burke et al (2015) report reduced longevity and fecundity among mated compared to virgin females.

Unfortunately, the mode of conception of these females was not determined, and thus it is not possible to exclude the possibility that they were parthenogenetically conceived, thus had low heterozygosity levels which would lead, as was the case for the females in this present study, to lower levels of fecundity and longevity.

Here, we offer a different perspective in understanding the evolutionary significance of facultative parthenogenesis, which is traditionally viewed as the “best of a bad job” when sexual reproduction was reduced or not possible. On first appearance, the observed costs of parthenogenesis for E. tiaratum are consistent with this view, and are even sufficiently severe to ask why parthenogenesis persists. However, our trans- generational data suggest that facultative parthenogenesis, rather than being a fall back strategy in the absence of males, may maximize female reproductive success by combining the “best of both worlds”. Because sexual reproduction appears to entirely recover the cost of parthenogenesis in the next generation, alternating between sex and parthenogenesis may allow female E. tiaratum to enjoy the benefits of both reproductive modes. Parthenogenesis enables females to reproduce in the absence of

143 sufficient or preferred sperm, while sexual reproduction potentially provides the benefits of genetic reshuffling (Agrawal 2006). Cycles of male abundance or seasonal fluctuations in population density (Gascoigne et al 2009) may induce the alternation between reproductive modes in natural populations. Females may have some control over their reproductive mode, for example by adjusting pheromone signalling

(Schneider and Elgar 2010) or post-copulatory regulation of sperm usage (Eberhard

1996). Alternatively, E. tiaratum females may adopt a mixed reproductive strategy to maximize reproductive success, by fertilising only a proportion of their eggs following mating. Indeed, the benefits of facultative parthenogenesis along with the widespread evidence for female mating failure (Rhainds 2010) raise the question why this reproductive system is not more widespread in insects.

144 References

Agrawal, A. F. (2006). Evolution of sex: why do organisms shuffle their genotypes? Current Biology, 16(17), R696–R704. Alavi, Y., Elgar, M. A., and Jones, T. M. (2016). Male Mating Success and the Effect of Mating History on Ejaculate Traits in a Facultatively Parthenogenic Insect (Extatosoma tiaratum). Ethology, 122(7), 523–530. Alavi, Y., van Rooyen, A., Elgar, M. A., Jones, T. M., and Weeks, A. R. (2016). Novel microsatellite markers suggest the mechanism of parthenogenesis in Extatosoma tiaratum is automixis with terminal fusion. Insect Science, in press. Arnqvist, G., and Nilsson, T. (2000). The evolution of polyandry: multiple mating and female fitness in insects. Animal Behaviour, 60(2), 145–164. Becks, L., and Agrawal, A. F. (2011). The effect of sex on the mean and variance of fitness in facultatively sexual rotifers. Journal of Evolutionary Biology, 24(3), 656–664. Bell, G. (1982). The masterpiece of nature: the evolution and genetics of sexuality. Book, Berkely: University of California. Booth, W., Smith, C. F., Eskridge, P. H., Hoss, S. K., Mendelson, J. R., and Schuett, G. W. (2012). Facultative parthenogenesis discovered in wild vertebrates. Biology Letters, 8(6), 983–5. Brock, P. D., and Hasenpusch, J. W. (2009). Complete Field Guide to Stick and Leaf Insects of Australia. Bruggeman, J., Debets, A. J. M., Wijngaarden, P. J., deVisser, J. A., and Hoekstra, R. F. (2003). Sex slows down the accumulation of deleterious mutations in the homothallic fungus Aspergillus nidulans. Genetics, 164(2), 479–485. Bulmer, M. G. (1982). Cyclical parthenogenesis and the cost of sex. Journal of Theoretical Biology, 94(1), 197–207. Burke, N. W., Crean, A. J., and Bonduriansky, R. (2015). The role of sexual conflict in the evolution of facultative parthenogenesis: a study on the spiny leaf stick insect. Animal Behaviour, 101, 117–127. Carlberg, U. (1983). Copulation in Extatosoma tiaratum (MacLeay)(Insecta: Phasmida). Zoologischer Anzeiger. Carlberg, U. (1991). Effect of moisture on the length of the pre-embryonic diapause and the rate of the embryonic development in Extatosoma tiaratum MacLeay (Insecta: phasmida). Zoologischer Anzeiger, 227(1–2), 25–37.

145 Carson, H. L. (1967). Selection for parthenogenesis in Drosophila mercatorum. Genetics, 55(1), 157–71. Chapman, J. R., Nakagawa, S., Coltman, D. W., Slate, J., and Sheldon, B. C. (2009). A quantitative review of heterozygosity–fitness correlations in animal populations. Molecular Ecology, 18(13), 2746–2765. Charlesworth, B., and Charlesworth, D. (1999). The genetic basis of inbreeding depression. Genetical Research, 74(3), 329–340. Corley, L. S., Blankenship, J. R., and Moore, A. J. (2001). Genetic variation and asexual reproduction in the facultatively parthenogenetic cockroach Nauphoeta cinerea: implications for the evolution of sex. Journal of Evolutionary Biology, 14(1), 68–74. Decaestecker, E., Gaba, S., Raeymaekers, J. A. M., Stoks, R., Van Kerckhoven, L., Ebert, D., and De Meester, L. (2007). Host–parasite “Red Queen”dynamics archived in pond sediment. Nature, 450(7171), 870–873. Eberhard, W. G. (1996). Female control: sexual selection by cryptic female choice. Princeton University Press. Elzinga, J. A., Chevasco, V., Grapputo, A., and Mappes, J. (2011). Influence of male mating history on female reproductive success among monandrous Naryciinae (Lepidoptera: Psychidae). Ecological Entomology, 36(2), 170–180. Gascoigne, J., Berec, L., Gregory, S., and Courchamp, F. (2009). Dangerously few liaisons: a review of mate-finding Allee effects. Population Ecology, 51(3), 355– 372. Gerritsen, J. (1980). Sex and parthenogenesis in sparse populations. American Naturalist, 718–742. Hadany, L., and Otto, S. P. (2009). Condition dependent sex and the rate of adaptation. The American Naturalist, 174(S1), S71–S78. Hamilton, W. D. (1980). Sex versus non-sex versus parasite. Oikos, 282–290. Hill, W. G., and Robertson, A. (1966). The effect of linkage on limits to artificial selection. Genetical Research, 8(3), 269–294. Hong, Z. D., and Ando, Y. (1998). Parthenogenesis in three species of genus Oxya. Japanese Journal of Applied Entomology and Zoology (Japan), (42), 65–69. Jaenike, J. (1978). An hypothesis to account for the maintenance of sex within populations. Evol. Theory, 3, 191–194. Kaltz, O., Bell, G., and Morgan, M. (2002). The ecology and genetics of fitness in Chlamydomonas. XII. Repeated sexual episodes increase rates of adaptation to novel environments. Evolution, 56(9), 1743–1753.

146 Kay, R. (1977). Proportional hazard regression models and the analysis of censored survival data. Applied Statistics, 227–237. Kearney, M. (2005). Hybridization, glaciation and geographical parthenogenesis. Trends in Ecology and Evolution, 20(9), 495–502. King, K. C., Delph, L. F., Jokela, J., and Lively, C. M. (2009). The geographic mosaic of sex and the Red Queen. Current Biology, 19(17), 1438–1441. Kramer, M. G., and Templeton, A. R. (2001). Life-history changes that accompany the transition from sexual to parthenogenetic reproduction in Drosophila mercatorum. Evolution, 55(4), 748–761. Kramer, M. G., Templeton, A. R., Miller, K. G. (2002). Evolutionary implications of developmental instability in parthenogenetic Drosophila mercatorum. I. Comparison of several strains with different genotypes. Evolution and Development. 4, 223–233. Kumpulainen, T., Grapputo, A., and Mappes, J. (2004). Parasites and sexual reproduction in psychid moths. Evolution, 58(7), 1511–1520. Lamb, R. Y., and Willey, R. B. (1979). Are parthenogenetic and related bisexual insects equal in fertility? Evolution, 33, 774–775. Lively, C. M. (1987). Evidence from a New Zealand snail for the maintenance of sex by parasitism. Nature, 328(6130), 519–521. Lynch, M. (1984). The limits to life history evolution in Daphnia. Evolution, 465–482. Lynch, M., and Deng, H.-W. (1994). Genetic slippage in response to sex. American Naturalist, 242–261. Matsuura, K., Fujimoto, M., and Goka, K. (2004). Sexual and asexual colony foundation and the mechanism of facultative parthenogenesis in the termite Reticulitermes speratus (Isoptera, Rhinotermitidae). Insectes Sociaux, 51(4), 325– 332. Matsuura, K., and Kobayashi, N. (2007). Size, hatching rate, and hatching period of sexually and asexually produced eggs in the facultatively parthenogenetic termite Reticulitermes speratus (Isoptera: Rhinotermitidae). Applied Entomology and Zoology, 42(2), 241–246. Maynard Smith, J. (1978). The evolution of sex. Book, New York: Cambridge University Press. Morgan-Richards, M., Smissen, R. D., Shepherd, L. D., Wallis, G. P., Hayward, J. J., Chan, C., Chambers, G. K., and Chapman, H. M. (2009). A review of genetic analyses of hybridisation in New Zealand. Journal of the Royal Society of New Zealand, 39(1), 15–34. Morran, L. T., Schmidt, O. G., Gelarden, I. A., Parrish, R. C., and Lively, C. M.

147 (2011). Running with the Red Queen: Host-Parasite Coevolution Selects for Biparental Sex. Science, 333(6039), 216 LP-218. Muller, H. J. (1964). The relation of recombination to mutational advance. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 1(1), 2–9. Nei, M. (1967). Modification of linkage intensity by natural selection. Genetics, 57(3), 625. Normark, B. B., and Moran, N. A. (2000). Testing for the accumulation of deleterious mutations in asexual eukaryote genomes using molecular sequences. Journal of Natural History, 34(9), 1719–1729. Otto, S. P. (2009). The evolutionary enigma of sex. The American Naturalist, 174(S1), S1–S14. Paland, S., and Lynch, M. (2006). Transitions to asexuality result in excess amino acid substitutions. Science, 311(5763), 990–992. Rhainds, M. (2010). Female mating failures in insects. Entomologia Experimentalis et Applicata, 136(3), 211–226. Roff, D. A. (2002). Inbreeding Depression: Tests of the Overdominance and Partial Dominance Hypotheses. Evolution, 56(4), 768–775. Roth, L. M. (1974). Control of oötheca formation and oviposition in Blattaria. Journal of Insect Physiology, 20(5), 821–844. Roze, D. (2012). Disentangling the benefits of sex. PLoS Biology, 10(5), e1001321. Salmon, J. T. (1955). Parthenogenesis in New Zealand stick insects. In Transactions of the Royal Society of New Zealand, Vol. 82, 1189–1192. Schneider, A., and Elgar, M. A. (2010). Facultative sex and reproductive strategies in response to male availability in the spiny stick insect, Extatosoma tiaratum. Australian Journal of Zoology, 58(4), 228–233. Schwander, T., Vuilleumier, S., Dubman, J., and Crespi, B. J. (2010). Positive feedback in the transition from sexual reproduction to parthenogenesis. Proceedings of the Royal Society B: Biological Sciences, 277(1686), 1435–1442. Simon, J.-C., Rispe, C., and Sunnucks, P. (2002). Ecology and evolution of sex in aphids. Trends in Ecology and Evolution, 17(1), 34–39. Simon, J., Delmotte, F., Rispe, C., and Crease, T. (2003). Phylogenetic relationships between parthenogens and their sexual relatives: the possible routes to parthenogenesis in animals. Biological Journal of the Linnean Society, 79(1), 151–163. Stalker, H. D. (1956). On the evolution of parthenogenesis in Lonchoptera (Diptera). Evolution, 345–359.

148 Suomalainen, E., Saura, A., and Lokki, J. (1976). Evolution of parthenogenetic insects. In Evolutionary biology. Springer. Templeton, A. R. (1983). Natural and experimental parthenogenesis. Pp. 343–398 in M. Ashburner, H. L. Carson, and N. J. J. Thompson, eds. The genetics and biology of Drosophila. Vol. 3. Academic Press, New York. Väli, Ü., Einarsson, A., Waits, L., and Ellegren, H. (2008). To what extent do microsatellite markers reflect genome wide genetic diversity in natural populations? Molecular Ecology, 17(17), 3808–3817. Wolf, H. G., Wöhrmann, K., and Tomiuk, J. (1987). Experimental evidence for the adaptive value of sexual reproduction. Genetica, 72(2), 151–159.

149