Tarantulas of Australia: phylogenetics and venomics Renan Castro Santana Master of Biology and Animal Behaviour Bachelor of Biological Sciences

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2018 School of Biological Sciences

Undescribed species from Bradshaw, Northern Territory Abstract Theraphosid spiders (tarantulas) are venomous arthropods found in most tropical and subtropical regions of the world. Most Australian tarantula species were described more than 100 years ago and there have been no taxonomic revisions. Seven species of theraphosids are described for Australia, pertaining to four genera. They have large geographic distributions and they exhibit little morphological variation. The current taxonomy is problematic, due to the lack of comprehensive revision. Like all organisms, tarantulas are impacted by numerous environmental factors. Their venoms contain numerous peptides and organic compounds, and reflect theraphosid niche diversity. Their venoms vary between species, populations, sex, age and even though to maturity. Tarantula venoms are complex cocktails of toxins with potential uses as pharmacological tools, drugs, and bioinsecticides. Although numerous toxins have been isolated from venoms of tarantulas from other parts of the globe, Australian tarantula venoms have been little studied. Using molecular methods, this thesis aims to document venom variation among populations and species of Australian tarantulas and to better describe their biogeography and phylogenetic relationships. The phylogenetic species delimitation approach used here predicts a species diversity two to six times higher than currently recognized. Species examined fall into four main clades and the geographic disposition of those clades in Australia seems to be related to precipitation and its seasonality. Australian tarantulas are shown to be non- monophyletic. Species may have immigrated multiple times between South-east Asia (SEA) and Australia. Australian species belonging to the genera Phlogius and Coremiocnemis are shown to be paraphyletic. However, the inclusion of “true” Selenocosmia species, e.g., from southeast Asia, is necessary to make more sound decisions regarding synonyms and diagnoses. Haplotypes of P. crassipes reveals highly genetically structured genes and the absence of isolation by distance. Phlogius crassipes is composed of at least two cryptic species, and here its range and diagnosis are updated. Phlogius strenuus is transferred from Selenocosmia and its range and diagnosis are updated. With the new range of Phlogius strenuus, it is clear that populations, genetic diversity, and even entire species are at risk, based upon the assumption that the species might be harvested sustainably is rejected. I found 26 phylogenetic clades using 2% divergence of the nucleotides of the 16S barcode region. Using whole venom profiling, the venom of individuals within a clade varies as much as across clades. However, venom belonging to phylogenetic groups contained at least one unique peptide that could be utilised to distinguish it from other phylogenetic groups. These results suggest that venom variation between phylogenetic groups may be useful in ascertaining species identity. I show that the venom of P. strenuus changes continuously during development and throughout adulthood, i.e., there are ontogenetic changes in venom composition. Intraspecific and interspecific venom variation can demonstrate how fast evolving and adaptive venom can be. The presence of unique peptides within each phylogenetic group is evidence that the specificity of venom could be used as an identification method in future. In addition, unique peptides within potential species and within each maturity stage present extra evidence that a single species can highly contribute to amplify our chemical library, which can increase possibilities of making new pharmacological discoveries. Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, financial support and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my higher degree by research candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis and have sought permission from co-authors for any jointly authored works included in the thesis.

Publications during candidature

Santana, R. C., Perez, D., Dobson, J., Panagides, N., Raven, R. J., Nouwens, A., Jones, A., King, G. F., Fry, B. G. (2017). Venom profiling of a population of the theraphosid spider Phlogius crassipes reveals continuous ontogenetic changes from juveniles through adulthood. Toxins, 9(4), 116. https://doi.org/10.3390/toxins9040116

Santana, R. C. (2015). Community structure and composition of litter spiders (Arachnida: Araneae) and influence of macro-climatic factors on Parque Ecológico Jatobá Centenário, Morrinhos, Goiás, Brazil. Journal of Threatened Taxa, 7(10), 7612–7624. https://doi.org/10.11609/JoTT.o4030.7612-24

Publications included in this thesis

Santana, R. C., Perez, D., Dobson, J., Panagides, N., Raven, R. J., Nouwens, A., Jones, A., King, G. F., Fry, B. G. (2017). Venom profiling of a population of the theraphosid spider Phlogius crassipes reveals continuous ontogenetic changes from juveniles through adulthood. Toxins, 9(4), 116. https://doi.org/10.3390/toxins9040116 – incorporated as Chapter 5.

Contributor Statement of contribution Renan Castro Santana (Candidate) Conception and design (50%) Analysis and interpretation (50%) Drafting and production (57%) David Perez Conception and design (0%) Analysis and interpretation (4 %) Drafting and production (1%) James Dobson Conception and design (0%) Analysis and interpretation (4%) Drafting and production (3%) Nadia Panagides Conception and design (0%) Analysis and interpretation (4%) Drafting and production (3%) Robert J. Raven Conception and design (5%) Analysis and interpretation (10%) Drafting and production (10%) Amanda Nouwens Conception and design (5%) Analysis and interpretation (4%) Drafting and production (3%) Alun Jones Conception and design (10%) Analysis and interpretation (4%) Drafting and production (3%) Glenn F. King Conception and design (15%) Analysis and interpretation (10%) Drafting and production (10%) Bryan G. Fry Conception and design (15%) Analysis and interpretation (10%) Drafting and production (10%)

Manuscripts included in this thesis

None

Contributions by others to the thesis

Chapter one: Dr Robert J. Raven – editing draft. Dr Bryan G. Fry – editing draft. Dr Lyn Cook – editing draft.

Chapter two: Dr Robert J. Raven – financial support for DNA sequencing, assisted with field collection and editing draft. Dr Bryan G. Fry – financial support for DNA sequencing. Dr Lyn Cook – financial support for DNA sequencing, assisted creating ideas, analysis of data and draft dediting to end of section. Dr Jason Bond – financial support for DNA sequencing. Ethan Briggs – assisted with field collection and laboratory work.

Chapter three: Dr Robert J. Raven – financial support for DNA sequencing, assisted field collection, creating ideas and draft edition. Dr Bryan G. Fry – financial support for DNA sequencing. Dr Lyn Cook – financial support for DNA sequencing, assisted creating ideas, analysis of data and draft edition.

Chapter four: Dr Robert J. Raven – financial support for DNA sequencing and assisted with draft edition. Dr Bryan G. Fry – financial support for DNA sequencing and venom proteomics, assisted with draft edition. Dr Lyn Cook – financial support for DNA sequencing, assisted creating ideas, analysis of data and draft edition. Kimberley Biggs – assisted with data analysis and draft production. Dr Simon Blomberg – assisted with statistical analysis. Dr Alun Jones – assisted with mass spectrometry runs and analysis.

Chapter five: Contributions listed under section “Publications included in this thesis”

Statement of parts of the thesis submitted to qualify for the award of another degree

None

Research Involving Human or Animal Subjects

None Acknowledgements

Firstly, I would like to thank my supervisors, Dr Robert Raven, Dr Lyn Cook, Associate Professor Bryan Fry and Professor Glenn King, for all support and mentoring during the course of this study. Their indispensable advice was key to the completion of this thesis.

I would like to thank all my laboratory colleagues, especially James Dobson, Nadia Panagides, Ethan Briggs, Alicia Toon and Paul Lin, that helped me through this journey. Without your initial guidance, I could not have achieved this.

A great thank you to my field assistants, Daniel Dashevsky, Ethan Briggs, Perry Bennion and Alexandre Dal Bo. The laughs and adventures will be always remembered.

Thanks to the BushBlitz program of the Australian Department of Environment and Energy, without which this research would not have been possible.

A huge thank you to my wife, Aline Sampaio Villa, to keep me sane, focused and persistent though my PhD. To help share the load of life in good and bad moments and to help me to see a brighter future.

Financial support

This research was supported by The University of Queensland – Research training tuition fees offset and The University of Queensland – International Living Scholarship. UQ funding to Cook Lab, Venom Evolution Lab and Glenn King Lab which partly funded the research. The Department of Terrestrial Biodiversity of Queensland Museum which partially funded the research.

Keywords spider, venom, theraphosid, biogeography, ontogeny, taxonomy, systematic

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 060301, Animal Systematics and Taxonomy, 40% ANZSRC code: 060302, Biogeography and Phylogeography, 30% ANZSRC code: 060109, Proteomics and Intermolecular Interactions, 30%

Fields of Research (FoR) Classification

FoR code: 0603, Evolutionary Biology, 80% FoR code: 0601, Biochemistry and Cell Biology, 20%

Table of Contents

Chapter One – Introduction ...... 1 Tarantulas (Theraphosidae) ...... 2 Molecular phylogenetic studies on spiders ...... 5 Molecular phylogenetic studies of tarantulas ...... 7 Species concepts used for tarantulas ...... 10 Systematics of Australian tarantulas ...... 10 Tarantula venomic studies ...... 13 Geographic and taxa bias on tarantula venom studies ...... 15 Aims and hypothesis ...... 16 Thesis outline ...... 16 Chapter Two - Digging up the true biodiversity of Australian tarantulas: a first step to identify their whole diversity ...... 19 Abstract ...... 20 Introduction ...... 20 Material and methods ...... 22 Results ...... 26 Discussion ...... 32 Conclusion ...... 36 Acknowledgement ...... 36 Chapter Three - Australia's largest spider, the Eastern Tarantula (Theraphosidae: Phlogius crassipes), is a cryptic species complex ...... 37 Abstract ...... 38 Introduction ...... 38 Materials and methods ...... 40 Results ...... 45 Discussion ...... 50 Taxonomy ...... 54 Chapter Four - A study of whether venom profiles match phylogenetic groups in Australian tarantulas (Theraphosidae: Selenocosmiinae) ...... 61 Abstract ...... 62 Introduction ...... 62 Results ...... 63 Discussion ...... 66 Materials and Methods ...... 71 Chapter Five - Venom profiling of a population of the theraphosid spider Phlogius crassipes reveals continuous ontogenetic changes from juveniles through adulthood ...... 74 Abstract ...... 75 Introduction ...... 75 Results ...... 76 Discussion ...... 83 Conclusions ...... 85 Materials and Methods ...... 86 Chapter Six - General Discussion ...... 90 Species diversity of tarantulas in Australia ...... 91 Biogeography of tarantulas in Australia ...... 92 Variation of tarantula venoms in Australia ...... 93 References ...... 95 Appendices ...... 121 Chapter Two ...... 122 Chapter Three ...... 132 Chapter Four ...... 145 Chapter Five ...... 158

List of Figures & Tables

Chapter One Table 1.1. Taxonomic history of tarantulas of Australia. Current species, synonyms and taxonomic studies citing each species ...... 12 Chapter Two Table 2.1. List of species of Australian tarantulas ...... 21 Table 2.2. Genes, primer sequences and conditions used in this analysis ...... 23 Table 2.3. Running conditions settings for each partition including size and model...... 26 Figure 2.1. Maximum Credibility tree generated from concatenated gene matrix under Birth-Death model by BEAST2 ...... 29 Figure 2.2. PCA using three environmental variable collected from ALA ...... 30 Figure 2.3. Map using the three most responsive layers from PCA: precipitation on warmest quarter, precipitation on autumn and precipitation on summer ...... 30 Chapter Three Table 3.1. Genes, primer sequences and conditions used in this analysis ...... 41 Table 3.2. Running conditions settings for each partition including size and model ...... 44 Figure 3.1. Maximum credibility tree of mitochondrial genes of P. crassipes ...... 46 Figure 3.2. Haplotype network of 16S of P. crassipes ...... 47 Figure 3.3. Linear regression of mitochondrial genes (COI and 16S) of all P. crassipes sequenced ...... 47 Table 3.3. Statistical values of coalescence analysis performed by StarBeast ...... 48 Figure 3.4. ML tree of concatenated data of Australian Phlogius group ...... 49 Figure 3.5. PCA of morphometric values ...... 49 Figure 3.6. LDA of morphometric values...... 50 Figure 3.7. IBD of morphology ...... 50 Figure 3.8. Pictures of Phlogius crassipes female ...... 57 Figure 3.9. Pictures of Phlogius crassipes male ...... 58 Figure 3.10. Pictures of Phlogius strennus female ...... 59 Figure 3.11. Pictures of Phlogius strennus male ...... 60 Chapter Four Figure 4.1. Maximum credibility tree of Australian Theraphosidae ...... 64 Figure 4.2. Discriminant analysis of all samples of Australian Theraphosidae ...... 66 Figure 4.3: T-test analysis of the isotope with 1284.1 m/z amongst all phylogenetic groups...... 67 Table 4.1: Unique isotopes and their corresponding peptide determined to be evident within the each phylogenetic group using t-test analysis on MarkerView ...... 67 Chapter Five Figure 5.1. Combined liquid chromatography (LC) chromatograms ...... 78 Figure 5.2. Principal component analysis and discriminant analysis (PCA-DA) analysis of Phlogius crassipes population ...... 78 Table 5.1. Theraphotoxins from P. crassipes matched against UniProt arthropod database ...... 79 Table 5.2. Simpson’s similarity index for all peptides and all theraphotoxins ...... 80 Table 5.3. Permutation analyses of shotgun binary matrix data ...... 81 Figure 5.3. 1D SDS PAGE gel of representatives from a population of Phlogius crassipes and clustering analyses ...... 82 Figure 5.4. Simple Correspondence Analysis of representatives of group size from a population of P. crassipes ...... 83 Appendices Supplementary material - Chapter Two Supplementary figure 2.1. 16S gene MCC tree of AMC...... 123 Supplementary figure 2.2. 16S gene MCC tree of AEC ...... 123 Supplementary figure 2.3. 16S gene MCC tree of CYQC ...... 124 Supplementary figure 2.4. 16S gene MCC tree of TKC ...... 124 Supplementary figure 2.5. 18S gene NJ tree of AMC ...... 125 Supplementary figure 2.6. 28S gene NJ tree of AMC ...... 125 Supplementary figure 2.7. EF1γ gene NJ tree of AMC ...... 126 Supplementary figure 2.8. 18S gene NJ tree of AEC ...... 126 Supplementary figure 2.9. 28S gene NJ tree of AEC ...... 127 Supplementary figure 2.10. EF1γ gene NJ tree of AEC ...... 127 Supplementary figure 2.11. 18S gene NJ tree of CYQC ...... 128 Supplementary figure 2.12. 28S gene NJ tree of CYQC ...... 128 Supplementary figure 2.13. EF1γ gene NJ tree of CYQC ...... 129 Supplementary figure 2.14. 18S gene NJ tree of TKC ...... 129 Supplementary figure 2.15. 28S gene NJ tree of TKC ...... 130 Supplementary figure 2.16. EF1γ gene NJ tree of TKC ...... 130 Supplementary figure 2.17. PCA including 20 environmental variables collected from ALA ...... 131 Supplementary Table 2.1. Environmental variables ...... 131 Supplementary material - Chapter Three Supplementary figure 3.1. ABGD of 16S and CO1 from Phlogius samples...... 133 Supplementary figure 3.2. IBD of 18S and 28S of Phlogius samples ...... 133 Supplementary figure 3.3. IBD of EF1γ sequences of Phlogius samples ...... 134 Supplementary figure 3.4. ML tree of COI and 16S genes ...... 134 Supplementary figure 3.5. ML tree of 18S and 28S genes ...... 135 Supplementary figure 3.6. ML tree of EF1γ gene ...... 135 Supplementary figure 3.7. ML tree of 5 genes from Phlogius ...... 136 Supplementary figure 3.8. Haplotype network of 18S and 28S on QLD map ...... 136 Supplementary figure 3.9. PCA of morphometric values from Phlogius ...... 137 Supplementary figure 3.10. LDA of morphometric values from Phlogius ...... 137 Supplementary figure 3.11. PCA of morphometric values from males and females Phlogius...... 138 Supplementary table 1. Morphometric values of Phlogius ...... 138 Supplementary material - Chapter Four Supplementary table 4.1. Sample information of tarantulas ...... 146 Supplementary figure 4.1. PCA analysis of all spider venom ...... 147 Supplementary figure 4.2. PCA of all venom samples excluding Clade 1 ...... 148 Supplementary figure 4.3. Chromatogram group A ...... 148 Supplementary figure 4.4. Chromatogram group B ...... 149 Supplementary figure 4.5. Chromatogram group C ...... 149 Supplementary figure 4.6. Chromatogram group D ...... 149 Supplementary figure 4.7. Chromatogram group E ...... 150 Supplementary figure 4.8. Chromatogram group F ...... 150 Supplementary figure 4.9. Chromatogram group G ...... 150 Supplementary figure 4.10. Chromatogram group H ...... 151 Supplementary figure 4.11. Chromatogram group I ...... 151 Supplementary figure 44.12. Chromatogram group J ...... 151 Supplementary figure 4.13. Chromatogram group K ...... 152 Supplementary figure 4.14. Chromatogram group L ...... 152 Supplementary figure 4.15. Chromatogram group M ...... 152 Supplementary figure 4.16. Chromatogram group N ...... 153 Supplementary figure 4.17. Chromatogram group O ...... 153 Supplementary figure 4.18. Chromatogram group P ...... 153 Supplementary figure 4.19. Chromatogram group Q ...... 154 Supplementary figure 4.20. Chromatogram group R ...... 154 Supplementary figure 4.21. Chromatogram group S ...... 154 Supplementary figure 4.22. Chromatogram group T ...... 155 Supplementary figure 4.23. Chromatogram group U ...... 155 Supplementary figure 4.24. Chromatogram group V ...... 155 Supplementary figure 4.25. Chromatogram group W ...... 156 Supplementary figure 4.26. Chromatogram group X ...... 156 Supplementary figure 4.27. Chromatogram group Y ...... 156 Supplementary figure 4.28. Chromatogram group Z ...... 157 Supplementary material - Chapter Five Supplementary Tables ...... 159

List of Abbreviations used in the thesis (RP) HPLC Reverse phase - High Performance Liquid Chromatography °C Degrees Celsius °S Degrees south µg Micrograms µl Microliters µm Micrometre 1D One dimension ABGD Automatic Barcode Gap Discovery ACN Acetonitrile ADP Adenosine diphosphate AEC Australian Eremean clade AICM Akaike Information Criterion through Markov chain Monte Carlo ALA The Atlas of Living Australia AMC Australian monsoon region clade AMD Adenosine monophosphate aq Aqueous ATP Adenosine Tri-Phosphate BLAST Basic Local Alignment Search Tool British Museum of Natural History, now the Natural History Museum, BMNH London bp Base pairs BS Bootstrap support values The Convention on International Trade in Endangered Species of CITES Wild Fauna and Flora CNS Central Nervous System COI Cytochrome Oxidase 1 COII Cytochrome Oxidase 2 CRISP-2 Cysteine-rich secretory proteins 2 CTAB Cetyl trimethylammonium bromide CYQC Cape York and North Eastern Queensland clade Da Dalton DA Discriminant analysis DNA Deoxyribonucleic acid DP Declustering potential DTT Dithiothreitol EFy1 Elongation Factor 1-gamma EFy2 Elongation Factor 2-gamma FA Formic acid FDR False Discovery Rate Fig. Figure GMYC Generalised mixed Yule coalescent model GPS Global Positioning System GS1 Nebuliser gas 1 GS2 Heater gas 2 HKY Hasegawa-Kishino-Yano HPLC High Performance Liquid Chromatography I Invariant sites IBD Isolation by Distance ISSR Inter Simple Sequence Repeats ITS1 Internal transcribed spacer 1 ITS2 Internal transcribed spacer 2 IUCN International Union for Conservation of Nature JC Jukes Cantor kDa Kilodaltons Km Kilometres kV Kilovolt LC Liquid Chromatography LCMS or LC/MS Liquid Chromatography and Mass Spectrometry LC-MS/MS Liquid Chromatography - Tandem Mass Spectrometry LDA Linear discriminant analysis m/z Mass-to-charge ratio MAFFT Multiple Alignment using Fast Fourier Transform MALDI Matrix-assisted laser desorption/ionization MCC Maximum Clade Credibility MCG Museo Civico Genova ML Maximum Likelihood mL or ml Millilitre mM Micromole mm Millimetre mPTP Multi-rate Poisson tree processes MS Mass Spectrometry ms Milliseconds mtDNA Mitochondrial DNA NA Not applicable NGS Next Generation Sequencing NJ Neighbour-Join nm Nanometre NSW New South Wales NT Northern Territory nuDNA Nuclear DNA OL Olkola large OM Olkola medium OS Olkola small OXS Olkola extra small PCA Principal component analysis PCA-DA Principal component analysis and discriminant analysis PNG Papua New Guinea PNS Peripheral Nervous System QLD Queensland QM Queensland Museum RJR Robert J. Raven RNA Ribonucleic acid rRNA Ribosomal RNA SA South Australia SDS-PAGE Sodium Dodecyl Sulfate - Polyacrylamide Gel Electrophoresis SEA South-east Asia TKC Top End and Kimberley clade TOF-MS Time-of-flight mass spectrometry

U1-TRTX-Cv1a U1-theraphotoxin-Cv1a

U1-TRTX-Spl1a U1-theraphotoxin -Spl1a

U3-TRTX-Cg1a U3-theraphotoxin-Cg1a

U3-TRTX-Cg1b U3-theraphotoxin-Cg1b

U8-TRTX-Hs1b U8-theraphotoxin-Hs1b UHPLC Ultra High Performance Liquid Chromatography USA The United States of America V Volts WA Western Australia WAM Western Australia Museum μ-TRTX-Phlo1a µ-theraphotoxin-Phlo1a μ-TRTX-Phlo1b µ-theraphotoxin-Phlo1b μ-TRTX-Phlo2a µ-theraphotoxin-Phlo2a τ-TRTX-Gr1b τ-theraphotoxin-Gr1b

Chapter One – Introduction

1

Tarantulas (Theraphosidae)

Biology Tarantulas (Family Theraphosidae) are part of the Infraorder Mygalomorphae that also includes trapdoor, wishbone, funnel webs and curtain-web spiders. Mygalomorphae are characterised by having no more than six spinnerets, having two pairs of book lungs, paraxial (parallel) chelicerae and venom glands located in the chelicerae (Coddington & Levi, 1991; King, 2004; Raven, 1985). The parallel fangs of tarantulas and other mygalomorphs mean that spiders are only capable of striking in a downward direction: an action that requires their prey to be pushed against a hard substrate (such as the ground or tree trunks) prior to envenomation. The presence of two pairs of book lungs has led to the two current hypotheses on burrowing habits: (1) the surface area of the book lungs makes the spiders more susceptible to desiccation, and burrowing helps regulate humidity and temperature in the burrow (Davies & Edney, 1952; Schmitz, 2013); (2) unlike most araneomorph spiders that contain a single pair of book lungs and tracheae, mygalomorphs have only book lungs, which do not have efficient air exchange with the hemolymph and therefore lead a more sedentary lifestyle (suggested by: Anderson, 1970). Commonly known as bird-eating spiders, whistling or barking spiders, theraphosid spiders are the largest spiders in the world (Coddington & Levi, 1991). Tarantulas can be found all over the world in tropical and subtropical regions (Foelix, 2011). Their size allows these spiders to prey on large invertebrates and small vertebrates, such as frogs, snakes, lizards, birds and mammals (Punzo & Henderson, 1999). The burrowing habits of tarantulas requires them to be ambush predators. Hunting is more likely to occur at night, when the spider sits at the entrance of the burrow awaiting prey to pass by (Minch, 1978). They rarely leave the burrow, enlarging the tunnel as they grow. It is believed that most Australian tarantulas live in a single burrow throughout their entire lives, moving only if disturbed or during mating season (male only). Like all spiders, tarantulas are sexually dimorphic: female are normally bigger and bulkier than males. Not a lot is known about the maturation time of Australian tarantula species, but females and males of American tarantulas take between 8 to 13 years to reach maturity (Baerg, 1963). The oldest recorded mygalomorph, an Australian idiopid, lived 43 years. Males have a bulb located at the distal segment of each palp, which are used to transfer sperm to the female. When males reach adulthood and environment conditions are favourable, they leave their burrows looking for potential mates in mature 2 females. After mating, males usually do not live long, with laboratory registered records of up to 1 year after maturity is reached (Baerg, 1963). In nature, it is likely that male lifespans are even shorter. Females on the other hand can live a further 7 years after maturity (Baerg, 2018). Dispersal in Mygalomorph spiders is limited, as they rely on walking (Coddington, 2005). In close proximity, it is possible to find siblings and family relatives of one maternal female. The young then disperse looking for potential retreats under logs and rocks.

Taxonomy The family Theraphosidae currently includes 144 genera and 967 species (World Spider Catalog, 2018). The principal synapomorphies are a digitiform apical segment of posterior lateral spinnerets, well-developed tarsal scopulae, claw tufts and distinct maxillary lobes (Raven, 1985). The family is presently split into 11 subfamilies: Ornithoctoninae, Thrigmopoeinae, Harpactirinae, Stromatopelminae, Ischnocolinae, Schismatothelinae, Selenogyrinae, Eumenophorinae, Selenoscomiinae, Theraphosinae and Aviculariinae (Guadanucci, 2014; Raven, 1985). Raven (1985) was the first in recent times to form a phylogenetic (cladistic) hypothesis of relationships of the subfamilies. Four subfamilies are characterised by having stridulatory organs, while the remainder are diagnosed by a combination of other characters. The stridulatory apparatus differs in each family by location of its parts. Guadanucci (2014) reviewed the subfamilies with special attention to the Ischnocolinae and corroborated the results of Raven (1985) for most of the subfamilies, but divided the paraphyletic Ischnocolinae, describing a new subfamily (Schismatothelinae). Australia has one subfamily of theraphosid spiders, Selenocosmiinae, which is shared with Asia and the Americas (Guadanucci, 2014; Raven, 1985). Members of Selenocosmiinae are characterized by the presence of a lyra on the prolateral face of the maxillae formed by clavate paddles and opposed by strikers on the chelicerae retromargin (Raven, 1985) (Chapter 3: Figs 3.8g, 3.10g). Both structures can be rubbed against each other producing a hissing sound. The function of the sound is still unknown for tarantulas but it is hypothesized to be a defensive display (Henning, Weberz, Hinkel, & Schottler, 2002; Kirchner & Röschard, 1999). The phylogeny of many mygalomorph families is not resolved. There is disagreement about the placement of some families and their sister groups and about the possible paraphyletic nature of some families as currently defined (Bond, Hendrixson, Hamilton, & Hedin, 2012; Coddington & Levi, 1991; Hedin & Bond, 2006; Raven, 1985). 3

The Theraphosidae is no exception: although the family is well established as a sister group of Barychelidae; included genera and species are far from being resolved (e.g. genera Haplopelma, Cyriopagopus and Melopoeus (World Spider Catalog, 2018). Raven (1985) described the systematics of theraphosids as a “taxonomic nightmare”. Since then little has been done to improve this status. Many studies have been done on small generic revisions and new species descriptions, but most fail to address phylogenetic relationships, e.g. (Nunn et al., 2016; Pérez-Miles, 2000; Reichling & West, 1996; West & Nunn, 2010a, 2010b). Selenocosmiine tarantulas from SEA have been recently studied, but each has been based only on a small number of samples of narrow scope, based primarily on the collections of Rick West (e.g., Nunn et al., 2016; West & Nunn, 2010a). Variability of diagnostic characters has not been consistently addressed and rigorous techniques to ensure repeatability and to manage exceptions have been ignored. As a result, the generic boundaries so formed have been inconsistent with some authors rejecting their own prior conclusions, e.g., the holotype of the type species of Pseudocnemis davidgohi West, Nunn, & Hogg, 2012 is a paratype of Coremiocnemis hoggi West & Nunn, 2010a. In their review of Phlogiellus, Nunn et al. (2016) stated that one of the characters that had been used in the initial cladogram (West et al., 2012) to define Phlogiellus was 200--350 cuspules on the labium. Nunn et al. (2016) then emended that to a maximum of 320 cuspules but within that paper, described Phlogiellus pelidnus Nunn et al., 2016 as having “>320”. The actual count of the number of cuspules on the labium requires a rigorous method which was not given. Certainly, the difference between 300 and 700 cuspules is evident. However, no studies on intraspecific variability were given in either paper. Indeed, the possible sample sizes were always less than 5 and no statistics were given. Also, in Nunn et al. (2016), a second character used in the cladogram was considered non-diagnostic for Phlogiellus but the data for the cladogram (from West et. al, 2012) was not changed and re-run. So, the defining characters of the genus were reduced from 3 to 2 and the limits of one of the two remaining characters (the number of labial cuspules) was changed, thus compromising its validity. This persistent lack of rigour greatly undermines the value of morphology in phylogenetic studies on the group. Cryptic morphological species have been found on the family Theraphosidae with close species showing lack of taxonomically informative characters, so morphologically- based taxonomy is not optimal at resolving phylogenetic issues (Hamilton, Hendrixson, & Bond, 2016; Hebert, Ratnasingham, & DeWaard, 2003; Mendoza & Francke, 2017). The

4 development of molecular tools to estimate phylogenies has brought some light to this matter and is more rigorous.

Molecular phylogenetic studies on spiders More than half a century ago, the first DNA sequence was obtained (Sinsheimer, 1959). However, just 40 years ago, modern DNA sequencing methods were introduced (Maxam & Gilbert, 1977; Sanger, Nicklen, & Coulson, 1977). Since then, the speed and capacity of DNA sequencing has developed rapidly (Hutchison, 2007). In the 1970s, DNA sequences started to be used to understand evolution and the phylogeny of vertebrates and invertebrates (Davidson, Galau, Angerer, & Britten, 1975). However, the first studies on spider molecular phylogenetics were only done in the late twentieth century (Gillespie, Croom, & Palumbi, 1994; Huber et al., 1993). More than 280,000 spider DNA/RNA genomic sequences have been deposited in databases such as GenBank. Half of these sequences are from the family Theridiidae, which includes the medically important redback and black widow spiders. 96% of all sequences are from Araneomorphae spiders. Almost 8,000 mygalomorph spider species sequences can be found at GenBank. These numbers show how fast genomic data has increased and become important to studies involving spiders. However, the bias toward a few species and families is notable. Half of all mygalomorph spiders sequenced are Theraphosidae, but even they are still a long way from having sequences for all known species. Most studies of molecular phylogeny have been on araneomorph spiders, but various families of mygalomorph spiders have been studied from limited samples (Bond et al., 2012; Hamilton, Hendrixson, Brewer, & Bond, 2014; Hendrixson & Bond, 2004; Rix et al., 2017; Rix, Main, Raven, & Harvey, 2018; J. D. Wilson, Hughes, Raven, Rix, & Schmidt, 2018; J. S. Wilson, Gunnell, Wahl, & Pitts, 2013; Wishart & Rowell, 2008). In a number of studies by Bond, results have been highly unstable with major relationship differences between successive papers (Bond & Beamer, 2006; Bond et al., 2012; Hedin & Bond, 2006; Hedin, Derkarabetian, Ramírez, Vink, & Bond, 2018; Hendrixson, Guice, & Bond, 2015). The majority of research is confined to North American spiders. The highly diverse tropical areas of the planet are still understudied and under sampled. Before my study, the sequences of only two species of Australian tarantulas were deposited in GenBank. Sequencing of genes such as 16S, 18S, 28S, COI, COII, EFy1, EFy2, ITS1 and ITS2 using traditional Sanger sequencing methods has been largely used in spider studies 5

(Brewer, Cotoras, Croucher, & Gillespie, 2014). Each of these genes present advantages and disadvantages. Mitochondrial genes, such as COI, COII and 16S, reflect the matrilineal history as the mitochondrial genome in spiders is inherited from the mother (D.- X. Zhang & Hewitt, 2003). As a consequence, molecular species delimitation tools can overestimate the diversity of species and evolutionary relationships can be oversimplified (Bond, 2004; Bond, Hedin, Ramirez, & Opell, 2001). Ribosomal nuclear genes, such as 18S, 28S, ITS1 and ITS2, have a few thousand copies in the genome (Hwang & Kim, 1999). 18S is characterised as a slow evolving gene that better resolves phylogenetic issues in deeper clades (Hwang & Kim, 1999). 28S is slightly faster evolving than 18S making it ideal for medium depth phylogenetic analysis. Both 18S and 28S present variable regions that are fast evolving that can be virtually impossible to produce an unambiguous alignment (Bond & Hedin, 2006; Hwang & Kim, 1999). Another problem with ribosomal nuclear genes, including ITS1 and ITS2, is the presence of divergent paralogues of which some can be pseudogenes (Buckler, Ippolito, & Holtsford, 1997). The amplifications of different paralogues from different individuals can result in incorrect phylogenetic outcomes. Nuclear protein coding genes, such EFy1 and EFy2, are slow evolving genes like the ribosomal 18S (C.-P. Lin & Danforth, 2004). EFy1 and EFy2 coding property results in fewer alignment issues (Danforth, Lin, & Fang, 2005), but they also present orthologous and paralogous copies (Sanderson & Shaffer, 2002). The different characteristics of each gene makes the production of a phylogenetic reliable, homologous and unambiguous alignment challenging. Earlier studies tended to use a single locus gene on phylogenetic analysis leading to inconclusive or incorrect results due to deep coalescence, genetic structuring and/or lack of resolution of the used genes (e.g. Bond, 2004; Bond, Beamer, Lamb, & Hedin, 2006; Bond et al., 2001b). As a result, recent studies have been devoted to the use of multi-loci data, incorporating more resolution and less biases in the result. The increased accessibility to primers of different gene regions from non-model species has also played an important role in multi-loci data acquisition (Thomson, Wang, & Jhonson, 2010). Genetic phylogeny has helped to resolve issues in theraphosid taxonomy when morphological analyses appear to founder or be highly variable or unresolved. Molecular- based phylogenies are still not strongly supported by most taxonomists, as most taxonomists have concerns about COI unreliability when used alone (Hamilton et al., 2014; Hendrixson et al., 2013a) and that each analysis produces divergent results, especially at the family level. Next Generation Sequencing techniques appear to have a greater 6 resolution than traditional Sanger technique, working on multiple taxonomic levels, but it is still a new tool and its weaknesses have yet to be established fully (Brewer et al., 2014). As most taxonomic studies still rely on traditional tools, it is important to recognise that different genes perform differently depending on the taxonomic level. Family level hypotheses normally use nuclear genes as they do not show high variability at the species and genus level. Mitochondrial genes, however, are highly modified between species making them ideal to identify species and generic diversity, but they can be strongly geographically structured (Bond, Hedin, Ramirez, & Opell, 2001a). Sequences obtained using traditional tools are short; thus, multiple gene sequences are needed to achieve more accurate results.

Molecular phylogenetic studies of tarantulas More than 1561 theraphosid gene sequences are now available in GenBank. Almost 1400 of these sequences are from American tarantulas. There are 162 sequences from Asian species of which 154 sequences represent four species of the Asian genus Haplopelma. Only two sequences are of Australian species, putatively of the genus Selenotholus. Over 1300 (1319) sequences of tarantulas on GenBank are only from mitochondrial genes. Of these, 1161 and 154 are sequences of cytochrome oxidase subunit 1 (COI) and 16S ribosomal RNA (rRNA), respectively. The reason that mitochondrial genes are being used more frequently in genetic studies simply reflects that they are more abundant, and thus easier to amplify (Brewer et al., 2014). However, several problems are linked with mitochondrial genes: they behave as a single locus as they lack recombination; they are responsive to small population sizes, due to their haploid characteristic; and interchange between sexes cannot be retrieved, as they are maternally inherited (Brewer et al., 2014). Genetic variation of Brachypelma vagans (Mexican orange knee tarantula), an IUCN protected species from Belize, was studied between two populations using two mitochondrial and one nuclear gene (Longhorn, Nicholas, Chuter, & Vogler, 2007). The nuclear gene was invariable between populations and the results of population structure were based only on the mitochondrial genes. A Multi-loci coalescent approach was not done to verify the links between populations. Most of the conclusions of this study were based on the biased mDNA genetic structure. The increase of sample and population number and the use of more nuclear genes should provide a better understanding of the population’s genetic diversity and connections.

7

Brachypelma vagans was also the target species used to adapt the Inter Simple Sequence Repeats (ISSR) technique to spiders analysing genetic variation between populations (Machkour‐M’Rabet et al., 2009). This technique is widely used in genetic of populations. However, it has low reproducibility and there is a possibility that the fragments amplified are not homologous. Dr Jason E. Bond and his collaborators have published several papers focusing on family, genera and species delimitation using molecular data (Bond et al., 2012; Hamilton, Formanowicz, & Bond, 2011; Hamilton et al., 2016, 2014a; Hedin & Bond, 2006; Hendrixson, DeRussy, Hamilton, & Bond, 2013a; Hendrixson et al., 2015). In addition to attempting to solve deep relations of mygalomorph families (Bond et al., 2012; Hedin & Bond, 2006), several of these publications used different molecular tools to resolve the complicated genus Aphonopelma in the USA. A study using just CO1 revealed bias with respect to the number of specimens and populations sampled (Hamilton et al., 2014a). Using tree-based species delimitation methods, the authors disregarded the problems related to the use of COI alone. In another study, the same author delimited Aphonopelma hentzi, via the mitochondrial genes CO1 and 16S, identifying species and even population connections (Hamilton et al., 2011). However, they claimed to have identified cryptic diversity based on less than five individuals using mDNA alone. Hendrixson et al. (2015) combined COI sequences with ecological, geographical and morphological data to show species delimitation and geographical boundaries, discovering three putative new species. Three of the five populations analysed had less than five individuals. The species delimitations were based on the mDNA, morphometrics and breeding periods. However, morphometric measurements were not compensated based on spider size and breeding periods were based on only two seasons sampling. As a result, the most reliable data was the COI phylogeny, that as previously noted, is not ideal for species delimitation in isolation. Although use of the mtDNA only is discouraged, it can be used in multi-loci approaches to reveal more precise species-level relationships. For example, in the study which Mojave Desert tarantulas were analysed using the barcoding gene (COI) and three nuclear genes (ITS1, ITS2 and 5.8S) showing molecular evidence for two species of Aphonopelma mojave, which were separated by a river (Hendrixson et al., 2013a). However, this study failed to analyse deep coalescence and basal nodes did not present significant support. As result, the putative new species were not described.

8

The last Aphonopelma paper is a monograph of more than 300 pages (Hamilton et al., 2016) treating 55 species of Aphonopelma in the USA. NGS was used in combination with morphometric analyses, biogeography and behaviour. The anchored enrichment technique was used to provide a molecular cladogram that was able to highlight both deep and shallow relationships (Hamilton et al., 2016). The delimitation of 29 species was done using behaviour, biogeographical range and molecular data. The study used an immense genetic dataset, bringing some closure to the taxonomic status of these species. However, the author presented a dichotomous key beginning with the US state in which the tarantulas occur. The use of political boundaries in the genus in the absence of morphological characters defining the species causes a biological flaw. There has been a large geographic bias in studies of tarantula molecular phylogeny. Most research and even redundant data has been published on USA theraphosids, mostly on the genus Aphonopelma. Aphonopelma is so far the most studied tarantula genus. This genus has its distribution through Central and North America. However, specimens outside the USA have never been phylogenetically studied. All conclusions about the molecular phylogeny of Aphonopelma have been done based on incomplete studies of the genus (Hamilton et al., 2016). As the taxonomy of tarantulas is still chaotic and genetic phylogeny results are highly dependent on sample and taxa numbers, the conclusions made by previous research are fluid. Molecular phylogenetic researchers have a variety of methods to analyse data and interpret results. Most of these methods rely on phenetic or cladistic analyses. While phenetics are based on relative similarity of characters, cladistics use assumptions of ancestral and descendent characters to create an evolutionary tree (Hennig, 1966; Pichardo, 1978). Both methods have been widely used in tarantula phylogenetic studies. However, a bias towards a method depending on the research aim is notable. Studies of population genetic and structuring use mainly phenetic tools, while studies aiming to reveal relationships and species delimitation used cladistics analysis (e.g. Hamilton et al., 2016; Machkour‐M’Rabet et al., 2009). Studies examining both evolutionary relationships and population structure use both phenetics and cladistics methods (e.g. Hamilton et al., 2014). The use of morphology alone in tarantula systematics has decreased since the development and affordability of molecular tools. It is likely that the lack of taxonomically informative characters is the main reason for this change. Molecular data has revealed 9 many cryptic species in the group, but also shown that some morphological features were taxonomically non-informative (e.g. Hamilton et al., 2016; Hendrixson et al., 2013, 2015). The studies including molecular and morphological data have used morphology just as supporting evidence for the molecular result and whenever possible to create identification keys (see Hamilton et al., 2016; Hendrixson et al., 2013, 2015). The inclusion of morphological and molecular data in a single matrix has only been used once in tarantulas (Bond et al., 2012).

Species concepts used for tarantulas Species concepts play an important role in taxonomic research. It generates controversial in one of the oldest scientific dilemmas: what is it a species? The disagreement between many scientists about the nature of species has created many definitions (De Queiroz, 2005). There are at least 26 different species concepts and each one can lead to different results (see Hausdorf, 2011; Wilkins, 2009). The immense variety of organisms on the planet makes this task even harder. Some concepts can be applied to some organisms but not to others. For example, parthenogenetic species cannot be delimited by the biological species concept (Y. P. Lin, Edwards, Kondo, Semple, & Cook, 2017). The failure to use a proper species concept depending on the species studied can cause confusion and controversy, affecting conservation and other research areas. Many studies of tarantulas taxonomy that use only morphological traits fail to specify the species concept used (e.g. Nunn et al., 2016; West & Nunn, 2010a). The decision based on “the gut” of a competent taxonomist is called taxonomic species concept (Mayden, 1997) and it is often the major cause of improper classification. The most recent phylogenetic studies of tarantulas that includes species descriptions specify the species concept used (e.g. Hamilton et al., 2016; Hendrixson et al., 2013; Ortiz & Francke, 2017). Hamilton et al. (2016) used the unified concept of a species (De Queiroz, 2005) that define each species as an independent evolving metapopulation. De Queiroz (2007) recognizes that diverse methods can define species at distinct phases of the speciation continuum. However, Hamilton et al. (2016) jump between different species concepts when the first chosen is not suitable. Such practice goes against the scientific method, fitting the data to get to a result, where in reality, the data should be analysed and results interpreted according to the chosen methodology.

Systematics of Australian tarantulas 10

The subfamily Selenocosmiinae was named and diagnosed by Simon (1889) with refinements by (Raven, 1985). Four genera and seven species are described for Australia (World Spider Catalog, 2018) (Table 1.1). Selenocosmia Ausserer, 1871 is diagnosed by the presence of long spines on the outer side of the chelicerae. A large, oval shaped and hairless lyra (Chapter 3: Figs 3.8g, 3.10g) on the maxillae is formed by several rows of numerous short paddles. Three rows of short spines (intercheliceral pegs) are located on the upper cheliceral inner face and leg I is longer and thicker than leg IV (Raven, 1985; Smith, 1986). Four Australian species are included in Selenocosmia, including S. crassipes (L. Koch, 1874), S. subvulpina Strand 1907, S. stirlingi Hogg, 1901 and S. strenua (Thorell, 1881) (World Spider Catalog, 2018); however, the Australian Faunal Directory, the National Database (http://www.environment. gov.au/biodiversity/abrs/online-resources/fauna/afd/home) lists S. crassipes in the genus Phlogius Simon, 1887. Selenotypus plumipes Pocock, 1895 was described from a single specimen collected near Townsville, Queensland. The species is the only one in the genus. Selenotypus was diagnosed by the eye pattern, labium shape and the thicker and longer fourth leg (Pocock, 1895). The species S. plumipes Pocock 1895 was named for the very long hairs located on posterior legs (Pocock, 1895). Selenotholus Hogg, 1902 was established because of the unusual thoracic fovea, which is recurved (procurved in all other Selenocosminae) and similar thickness of all legs (Hogg, 1902). The holotype of S. foelshei Hogg 1902 has its type (original) locality at Palmerston, Northern Territory (Hogg, 1902). However, it is believed that the location is incorrect because no similar specimen has been found at the location for more than a century. At the time Hogg described this species, Darwin city was called Palmerston (Powell, 2012) and many type localities of species denoted only the port of departure rather than the sampling area (Raven, 1982; Raven & Hebron, 2018). Coremiocnemis was described by Simon in 1892 from south-east Asia. The genus consists of tarantulas with long paddle-shaped lyrae, cracked tarsi, intercheliceral pegs and a third claw at least on the fourth leg (Raven, 2005; West and Nunn, 2010). Several species are known from Southeast Asia, but only C. tropix Raven 2005 is described from Australia. This genus were recently revised and new morphological characters included in the genus diagnosis (West & Nunn, 2010a). With this new diagnosis, the Australian species C. tropix is distinguished from other species of the genus by lacking bilobular spermathecae, lanceolate embolus without distal flaring and long recurved setal bushes on metatarsus and tarsus IV (West & Nunn, 2010a). 11

Species description in the 19th and early 20th centuries were often lacking in detail. Most of these descriptions were short and include morphological characters that are now known to be uninformative. Often rudimentary drawings were included only in a few descriptions and even then were not informative. No obvious species concept was applied; species boundaries were assumed only by morphological differences.

Table 1.1. Taxonomic history of tarantulas of Australia. Current species, synonyms and citations for each species.

Taxonomic Species Synonyms Locality studies

Selenocosmia stirlingi Selenocosmia stirlingi Crown Point (Main, 1985) (Hogg, 1901) Hogg, 1901 Station, SA (Smith, 1986) Selenocosmia stalkeri (Smith, 1987) Hirst, 1907 Alexandria, NT (Schmidt, 1995) (Schmidt, 2002) (Schmidt, 2003) (Schmidt, 2015)

Phlogius crassipes Phrictus crassipes Bowen, QLD (Simon, 1887) (L. Koch, 1874) L. Koch, 1874 (Main, 1982) Ischnocolus lucubrans Port Mackay, (Main, 1985) L. Koch, 1874 QLD (Smith, 1986) Selenocosmia vulpina Cape Upstart, (Smith, 1987) Hogg, 1901 QLD (Schmidt, 2002) Selenocosmia crassipes (Schmidt, 2003) Hogg, 1901 (Schmidt, 2015)

Selenocosmia strenua Phrictus strenuus Somerset, Cape (Thorell, 1881) Thorell, 1881 York

Selenocosmia Selenocosmia subvulpina North subvulpina Strand, 1907 Queensland Strand, 1907

Selenotholus foelschei Selenotholus foelschei Palmerston, NT Hogg, 1902 Hogg, 1902

Selenotypus plumipes Selenotypus plumipes Major’s Creek, (Hogg, 1901) Pocock, 1985 Pocock, 1985 Townsville, QLD

Coremiocnemis tropix Coremiocnemis tropix Atherton, QLD Raven, 2005 Raven, 2005

A few studies have dealt with Australian tarantula species after their original description (Table 1.1). These studies were not revisions, using only a single or few species and each lacking an explicitly stated species concept. The poor uninformative 12 original descriptions resulted in species being synonymized and transferred to different genera multiple times (Table 1.1).

Tarantula venomic studies Tarantulas represent over 50% of all mygalomorph species (with venom data) on ArachnoServer (Herzig et al., 2011). Toxin sequences are available for 21 of the 132 genera of extant theraphosid spiders (Herzig et al., 2011; World Spider Catalog, 2018). The first studies on tarantula venom occurred in the 1970s. All studies of tarantula venom from this decade focused mainly on North American tarantulas, apart the inclusion of Pterinochilus sp. from Africa and a short communication on Brachypelma emilia from Mexico (Chan, Geren, Howell, & Odell, 1975; Odell, Ownby, Cabblness, Hudlburg, & Herrero, 1979; Perret, 1977a, 1977b; Schanbacher, Lee, Hall, et al., 1973; Schanbacher, Lee, Wilson, Howell, & Odell, 1973). At this point, hyaluronidase and nucleotides were noted in the venom and the synergistic activity of venom compounds was reported. Also in 70’s, proteolytic activity was found to be related to saliva contamination and venom regeneration intervals were assessed by Perret (1977a, 1977b). Perret’s (1977a, 1977b) findings helped to improve venom extraction and reach more accurate results by establishing venom regeneration time interval to venom milking and venom milking techniques free of contamination. Four types of polyamines were detected from four species of tarantulas, being three species from the USA and one from Honduras. The main polyamine found was spermine, but traces of putrescine, cadaverine and spermidine were also detected (Cabbiness et al., 1980). Two of these species caused skeletal muscle necrosis in rats which was related to the effects of the polyamines (Ownby & Odell, 1983). Aphonopelma hentzi was the first model species used to study tarantula venoms. By the end of the 1980s, many compounds had been identified in tarantula venoms (free amino acids, AMD, ADP, ATP, polyamines, hyaluronidase and toxic peptides) (Savel- Niemann & Roth, 1989). A follow-up study using species of Aphonopelma and Harpactirella, from the USA and South Africa, respectively, showed the presence of acylpolyamine toxins and linked these compounds with temporary paralysis of insects (Skinner, Dennis, Lui, Carney, & Quistad, 1990). The effect of venom from the South American tarantula, Grammostola rosea, was assessed in ganglion neurons and heart muscle (Lampe et al., 1993; Pascarel, Cazorla, Le Guennec, Orchard, & White, 1997; Piser, Lampe, Keith, & Thayer, 1994). For the first time, tarantula venom was shown to be able to modulate the activity of voltage-gated 13 calcium channels, being a voltage-dependent inhibitor. Selenocosmia stirlingi was the first Australian tarantula to be studied; it examined its venom toxicity in rats (Atkinson, 1993). Six species of the genus Brachypelma from Central America were fingerprinted and their differences analysed by species (Escoubas, Célérier, & Nakajima, 1997). Population and species similarity were found and a group of species in the genus were identified by venom composition. The use of venom on a blind identification was successfully done when correlated with post-identification using morphology. The first decade of the 21st century was rich in studies on tarantula venom. The increase of popularity was possibly due to the development of modern, cheaper and easier techniques to process proteomics results and/or venom compounds affecting vertebrate ion channels with potential for new drug discovery. Apart from a few studies releasing new fingerprint proteomics results of tarantulas species, most of the research was done on fragments or/and specific toxins and relating them to their specific function in vertebrate systems and cells. The volume of new data was high and, since the year 2000, at least five review papers have been published on spider venoms (Escoubas, Diochot, & Corzo, 2000; Escoubas & Rash, 2004; King, 2004; Sannaningaiah, Subbaiah, & Kempaiah, 2014; Vassilevski, Kozlov, & Grishin, 2009). One of the first studies of this millennium on tarantula venom investigated the use of modern techniques, such as reverse-phase (RP) HPLC and matrix-assisted laser desorption/ionisation (MALDI) mass spectrometry (MS), for studying venom composition (Escoubas, Corzo, Whiteley, Célérier, & Nakajima, 2002). Sex, age and species variability were reported. In South America, the venom of Lasiodora parahybana is one of the most extensively studied. Venom fingerprinting was done using MALDI and HPLC tools. Individual and age variation was assessed, showing compounds that are specific to juveniles and different intensity of components between different aged spiders. Also, specific compounds were also found to delimit species (biomarkers) (Guette, Legros, Tournois, Goyffon, & Célérier, 2006). A partial characterization of the venom of the Brazilian tarantula Vitalius dubius was done using electrophoresis, liquid chromatography (LC), enzymatic essays, immunoblotting and ELISA (Rocha-e-Silva, Sutti, & Hyslop, 2009). Proteomic characterization of Acanthoscurria paulensis using HPLC and MALDI techniques was assessed and toxicity investigated on frogs and mice (Mourão et al., 2013). Four species of Australian tarantulas and two Mexican tarantulas had their venom insecticidal activity assayed and the venom proteome analysed using electrophoresis, LC 14 and MS techniques (Gentz, Jones, Clement, & King, 2009; Herzig & Hodgson, 2008). Variation in venom of the Australian tarantula Coremiocnemis tropix was explored, regarding ontogeny, sex, moulting, including pharmacological and insecticidal activities (Herzig, 2010; Herzig & Hodgson, 2009). Intensive studies have been done on Asian tarantulas. Chilobrachys guangxiensis, Haplopelma hainanum and Haplopelma schmidti had their venom-gland transcriptome sequenced and venom proteomics assessed using old and modern tools (Chen, Deng, et al., 2008; Chen, Zhao, et al., 2008; Jiang et al., 2010; Liao et al., 2007; Y.-Y. Zhang et al., 2015). As a result, 315 of all 448 toxins of tarantulas listed on ArachnoServer are from these three species. Apart of the three species listed above, the only two species with recorded venom- gland transcriptomes are the South American Grammostola rosea and the African Pelinobius muticus, composing a majority of the remaining toxins on the database (Diego- García, Peigneur, Waelkens, Debaveye, & Tytgat, 2010; Ostrow et al., 2003). A very comprehensive review of tarantula venoms was published in 2004 (Escoubas & Rash, 2004). The authors discussed tools used to obtain venom fingerprints, toxicity of tarantula venom, structure and sequence of some toxins and their functions. (For those interested in more complete information of tarantula venom biochemistry, this review is an obligatory read).

Geographic and taxa bias on tarantula venom studies Venom proteomics research has not been as biased as phylogenetic studies. Of all 21 genera listed on ArachnoServer, 8 are from America, 7 from Africa and 6 from Australasia (Herzig et al., 2011). More genera and species have had their venoms fingerprinted, but they lack details of toxin structure and function and they are not listed in the database. Venom studies of American tarantulas have been done on the genera Grammostola, Aphonopelma, Brachypelma, Psalmopoeus, Euathlus, Theraphosa, Thrixopelma, Lasidora, Vitalius and Sericopelma. Genera from Africa that have been studied include Harpactirella, Pterinochilus, Ceratogyrus, Eucratoscelus, Heteroscodra, Hysterocrates, Pelinobius, Stromatopelma. Australasian region has eight genera studied: Chilobrachys, Haplopelma, Plesiophrictus, Coremiocnemis, Phlogius, Selenotypus, Selenocosmia and Selenotholus. Even with ~20% of all theraphosid genera represented in the ArachnoServer database, research on tarantula venom still provides only a limited picture of the total 15 diversity, variation, evolution and function of the toxins found in this spider family. Most genera listed above are represented by only a single species studied, meaning that venom from only ~3% of the theraphosid species in the world have been studied. If you consider that tropical regions have much of their arachnofauna undiscovered, it is possible that venom studies have covered less than 1% of all tarantula species.

Aims and hypothesis

Based on the gap of knowledge described on the sections above, this thesis aims to: 1- Discover the phylogenetic diversity of Australian tarantulas. a. Identify the species richness of tarantulas in Australia; i. As part of this aim, I will test the hypothesis that there are seven species of tarantula in Australia. b. Assess the species status of P. crassipes; i. I will specifically test the current hypothesis that P. crassipes is a single species. c. Characterize biogeographic patterns of Australian tarantulas; 2- Conduct further investigations of venoms of Australian tarantulas. a. Describe the venom profiles of Australian tarantulas; b. Correlate venom profile with phylogenetic diversity of Australia tarantulas; i. I will test the current hypothesis: venom profile differs among phylogenetic groups, given that venom genes evolve faster than housekeeping genes. c. Identify changes in venom according to developmental stage of the spider; i. I will test the hypothesis that venoms will have a different composition according to the developmental stage of the spider, because this is what has been found in other studies of tarantulas.

Thesis outline

Chapter Two This chapter addresses aims 1a and 1c by a molecular approach. Nuclear and mitochondrial genes were amplified and analysed separate and together to determine 16 species relationship and gene flow. Maximum likelihood, Bayesian inference and maximum parsimony analysis were used to determine relationships. Tree methods to delimited species were used on mitochondrial genes. Neighbour-joining trees were used to assess gene flow of nuclear genes. Climatic variables and geographic position of main clades were examined for patterns of correlation using PCA.

Chapter Three This chapter focuses on aim 1b using molecular and morphological methods. Multi- locus approaches were used, including mitochondrial and autosomal genes. Haplotype networks were built using mitochondrial genes to review genetic structure. Isolation by distance was assessed on nuclear and mitochondrial genes by linear regression of sample location and genetic distance. Barcode gap and species delimitation methods were applied on mitochondrial DNA to identify possible cryptic species. Maximum likelihood and Bayesian inference were performed on autosomal genes to assess gene flow. Species coalescent tests were done on nuclear genes to test the cryptic species hypothesis. Morphometric analysis were also performed using PCA, DA and IBD.

Chapter Four This study refers to aims 2a and 2b by phylogenetic and venomic analysis. Phylogenetic analysis are done using autosomal and mitochondrial genes. Concatenated gene matrix were processed on BEAST and relationships of individuals were assessed. Barcode gap were used on mitochondrial genes to delimit potential species. Venom was processed using LC/MS. Chromatograms were overlaid and similarities analysed. PCA and multivariate analysis and were performed on a list of venom isotopes and peptides to identify pattern within and between phylogenetic clades. T-tests were done to find specific venom compounds within each potential species and group of potential species.

Chapter Five This chapter addresses aims 2a and 2c by venomic analysis of a population of Phlogius strenuus. Spiders from the same population were collected, carapace were measured and venom was extracted. Individuals were grouped in sizes to perform statistical analysis. Venom were analysed by electrophoresis, LC/MS and LC/MS/MS. Electrophoresis bands presence and intensity were assessed and analysed by Rho similarity. LC/MS chromatograms were overlayed and similarities analysed. PCA-DA were performed on venom isotope list to evaluate differences on venom composition. 17

LC/MS/MS results were run against arthropod database to identify peptides. Peptide lists were analysed using simple permutation analyses and similarity of Simpson.

18

Chapter Two

Digging up the true biodiversity of Australian tarantulas: a first step to identify their whole diversity

Contributions: Dr Robert J. Raven contributed with financial support for DNA sequencing, assisted with field collection and editing drafts; Dr Bryan G. Fry contributed with financial support for DNA sequencing; Dr Lyn Cook contributed with financial support for DNA sequencing, assisted creating ideas, analysis of data and draft editing; Dr Jason Bond contributed with financial support for DNA sequencing; Ethan Briggs assisted with field collection and laboratory work. Lachlan McIntyre, Eric Vanderduys, Volker Herzig, Kieran Palmer, Eamon Amsters and Kieran Aland donated specimens; and Daniel Dashevsky, Perry Bennion and Alexandre Dal Bo assisted with field collection.

19

Abstract Currently four genera and seven species of tarantulas (Theraphosidae) are described from Australia. Apart from one species (Coremiocnemis tropix), all were described more than a century ago and there has been no comprehensive revision since. It is therefore likely that there is more species diversity than currently recognised. Here, I use DNA sequencing and a phylogenetic species delimitation approach to assess the actual species richness of tarantulas in Australia. I find there are 19 undescribed species and possibly many more. Specimens fall into four main clades that appear to be largely geographically distinct, with distributions seemingly related to precipitation and its seasonality. Furthermore, Australian tarantulas do not form a monophyletic group because a specimen from Southeast Asia is inserted within one of the major Australian clades. A more in-depth taxonomic study incorporating morphological features needs to be undertaken to delineate and describe new species and genera.

Introduction The Theraphosidae (commonly known as the tarantula family) is the most diverse family of mygalomorphs, currently comprising 144 genera and 967 species (World Spider Catalog, 2018). The Theraphosidae is well established as monophyletic and the sister group of Barychelidae (Brush-foot trapdoor spiders) (Bond, Hendrixson, Hamilton, & Hedin, 2012; Hedin & Bond, 2006), but the species diversity of the group and relationships within the family are far from resolved. For example, the Asian genus Cyriopagopus (Simon, 1887) and several species within it has changed genera (Raven, 1985; Smith & Jacobi, 2015). Eleven subfamilies are currently recognised with the highest species diversity in the Americas (Guadanucci, 2014). Australia’s tarantulas are all classified in the subfamily Selenocosmiinae, which is shared with Asia and America. Four genera and seven species are described for Australia at present. Selenocosmia Ausserer 1871 and Coremiocnemis Simon 1892 are shared with South East Asia, while the monotypic genera Selenotholus Hogg 1902 and Selenotypus Pocock 1895 are Australian endemics (World Spider Catalog, 2018). Apart from Coremiocnemis tropix, all Australian tarantulas were described more than a century ago (Table 2.1). Meanwhile, species have otherwise only been mentioned notably Phlogius crassipes (Main, 1982; Schmidt, 1995). Schmidt (1995), working more on Asian Selenocosmiinae, considered he had found the long lost type of Selenocosmia javanensis (the type species of Selenocosmia) and, on the basis of poorly explained differences between that and Australian species placed in Selenocosmia, restored Phlogius. Raven (2000) noted that the material on which that

20 decision was based was not Walckenaer's type and rescinded the restoration. However, Raven (2005: 17), in his diagnosis of Coremiocnemis (and description of Coremiocnemis tropix Raven 2005), tacitly agreed with Schmidt's concept of Selenocosmia, viz. "[Coremiocnemis] Differs from Selenocosmia in having the maxillary lyra consisting of long shafted paddles with long distal blades (Figs 14, 23). Coremiocnemis is a selenocosmiine theraphosid with intercheliceral peg spines (Figs 6, 37), maxillary lyra consisting of long paddles with long distal blades (Figs 14, 23), cracked fourth tarsi, and a third claw on the fourth leg." All Australian theraphosids have long-shafted paddles on the lyra and only Coremiocnemis is known to have intercheliceral leg spines. Thus, all Australian theraphosid species included in Selenocosmia were implicitly misplaced. Generic relationships within Selenocosmiinae are largely unknown (West, Nunn, & Hogg, 2012). New genera and species have been described without including suitable or pertinent material (Raven, 2000; Schmidt, 1995). Selenocosmiinae tarantulas from SEA have been recently studied but each has been based only on small samples of narrow scope based primarily on the collections of Rick West (e.g., Nunn et al., 2016; West & Nunn, 2010a). Equally, variability of diagnostic characters have either been ignored or not been consistently addressed and rigorous techniques of ensuring repeatability and managing exceptions have been ignored. As a result, the generic boundaries so formed have been inconsistent and no sooner than the first paper had been published a second paper by the same authors rejects those decisions, e.g., the holotype of the type species of Pseudocnemis davidgohi West, Nunn, & Hogg, 2012 is a paratype of Coremiocnemis hoggi West & Nunn, 2010a.

Table 2.1. List of species and authors of Australian tarantulas.

Species Author

Selenocosmia stirlingi (Hogg, 1901)

Phlogius crassipes (L. Koch, 1874)

Selenocosmia strenua (Thorell, 1881)

Selenocosmia subvulpina Strand, 1907

Selenotholus foelschei Hogg, 1902

Selenotypus plumipes Pocock, 1985

Coremiocnemis tropix Raven, 2005

21

Recent revisions of tarantula genera in northern and central America incorporating phylogenetic analyses have revealed previously undescribed species, but they have also resulted in synonymies as high variation among individuals of the same species has been recognised (e.g. Hamilton, Hendrixson, & Bond, 2016; Mendoza & Francke, 2017; Ortiz & Francke, 2017). Furthermore, many described species of theraphosid appear to be cryptic species complexes (Hamilton et al., 2016; Ortiz & Francke, 2017). With large variability of some characters, and little or no variation of other characters, morphology-only taxonomy is not promising for resolving taxonomic issues (Hamilton et al., 2016; Ortiz & Francke, 2017). During the last decade, a number of revisions on Australian mygalomorph have been published. The Idiopidae is the family which have had most of the attention, with revisions of family and genera level across the country (Harrison et al., 2017; Rix, Bain, et al., 2017; Rix, Cooper, et al., 2017; Rix, Main, Raven, & Harvey, 2018; Wilson, Hughes, Raven, Rix, & Schmidt, 2018). In a study of biogeography of the Idiopidae, Rix et al. (2017) found that consequential reduction of population size and fragmentation of habitat specialist species occurred due to the historical biogeographic evolution of the Australian Eremean region. At the same time, arid specialist clades expanded. The genus Euoplos has had several cryptic species uncovered from eastern Australia, many being sympatric and with a history of incorrect matches of males and females (Wilson et al., 2018). However, the most studied region is the southwestern Australia hotspot with two idiopid genera and one genus of the Migidae (Cooper, Harvey, Saint, & Main, 2011; Harvey, Main, Rix, & Cooper, 2015; Rix, Bain, et al., 2017; Rix et al., 2018). In these revisions, the authors exposed several new species with small endemic distributions. Doubt surrounds the taxonomy of Australian tarantulas due to the absence of their comprehensive revision and the continental size and biome diversity of Australia. This project aims to use a phylogenetic approach to reveal tarantula biodiversity in Australia and provide a biogeographic overview of the group.

Material and methods

Sampling Field sampling was undertaken from 2012 to 2016 across Queensland and the Northern Territory using previous locations recorded by museums. When possible, three specimens were collected per site. Ninety-five tarantulas were collected from diverse locations though Queensland and Northern Territory and are deposited in the respective museums (Queensland Museum, QM, and the Museum and Art Gallery of the Northern

22

Territory, MAGNT). GPS data for all localities were recorded and have been added to the QM and MAGNT databases, but in this thesis, GPS coordinates are not given due to the conservation concern that this would release important information to illegal collectors and threaten species or populations. The Western Australian Museum (WAM) kindly donated 23 tissue samples from Western Australia and five from the Northern Territory. Private individuals donated eight specimens as acknowledge bellow.

DNA extraction, PCR and sequencing Because of their different modes of inheritance, mitochondrial (mtDNA) and nuclear (nuDNA) DNA can provide complementary information about past gene flow. Female tarantulas do not move far from the maternal burrow, thus there should be strong geographic structuring based on mtDNA within species. Nuclear genes, however, are transferred by both males and females. Phylogenetic agreement among genes from both genomes can be interpreted as a lack of mixing of genes and decisions can then made about whether this is evidence of long-term reproductive isolation: equivalent to applying a biological species concept in which evidence for a long-term lack of gene flow between populations is interpreted as evidence for reproductive isolation. Genomic DNA was extracted from muscle tissue of one leg of a specimen using either a CTAB/Chloroform protocol (as per Lin et al., 2013) or Qiagen DNeasy Tissue Kit (Qiagen, Inc., Valencia, CA, USA) following the manufacturer's instructions. We amplified one region of mtDNA (16S) and three nuDNA genes: nuclear large subunit ribosomal RNA (LSU rDNA, 28S), nuclear small subunit ribosomal RNA (SSU rDNA, 18S) and Elongation factor 1- gamma (EF1γ) using the primers and PCR protocols listed in Table 2.1. PCR products were prepared for sequencing by adding a mix of exonuclease I and Antarctic Phosphatase. Sequencing was performed by Macrogen Inc., Republic of Korea.

Table 2.2. Genes, primer sequences and conditions used in this analysis.

Gene Primer sequence Annealing Primer Direction Reference region 5’ to 3’ temperature

Elongation EF1γ ATTGCBGCNCA F Factor 1- F78 GTAYAGYGG Step down (Ayoub, Garb, Hedin,

gamma EF1γ CCTTGRTTGAAY 58°C to 40°C & Hayashi, 2007) R (EF1γ) R1258 TTCTTTCC

23

EF1γ CAGTGGATTAG F TaF CTTTGCTG David Ortiz, pers. 48°C EF1γ CGCATTTTATCT comm. R TaR AGCCTCTG nuclear CTGGTTGATCCT 2880 F small GCCAGTAG (von Dohlen & subunit 55°C TCCAAGAATTTC Moran, 1995) rRNA 2883 R ACCTCT (18S)

GAGAGTTMAAS (Dowton & Austin, nuclear S3660 F AGTACGTGAAA 1998) large C subunit 55°C (Whiting, Carpenter, rRNA TCGGARGGAAC A335 R Wheeler, & Wheeler, (28S) CAGCTACTA 1997)

C12054 CGCCTGTTTAW 16S 16Sa- F (Arabi, Cruaud, CAAAAACAT ribosomal mod Couloux, & 48°C RNA C12054 CTGGCTYACGC Hassanin, 2010; (16S) 16S- R CGGTYTGAACT Simon et al., 1994) rRN CA

Phylogenetic analyses

Sequence traces were edited manually using the software Gene Studio v.2.2 (Gene- Studio, Inc., Suwanee, Georgia). Polymorphic sites in autosomal genes were phased using DNA Sequence Polymorphism v.6.10.01 (Rozas et al., 2017). Sequences were aligned using the Guidance2 Server with the MAFFT algorithm (Sela, Ashkenazy, Katoh, & Pupko, 2015). Alignments were then manually edited using software MEGA7 v.7.0.18 (Kumar, Stecher, & Tamura, 2016). Areas of ambiguous alignment were found in the 16S and 28S genes. These alignments were further processed using Gblocks v.0.91b (Castresana, 2000; Talavera & Castresana, 2007) with settings to allow smaller final blocks, less strict flanking positions 24 and gaps positions within the final blocks on both 16S and 28S. Fasta files were modified to fit each different software package using ALTER (Glez-Pena, Gomez-Blanco, Reboiro-Jato, Fdez-Riverola, & Posada, 2010). JModelTest (Santorum, Darriba, Taboada, & Posada, 2014) was used to determine the best fit model for each gene (Table 2.2). Models were manually changed when over-parameterisation was detected, typically by failure to converge on a parameter estimate. Codon position of Elongation Factor 1-gamma (EF1γ) were partitioned on 1st and 2nd position, and 3rd position. After checking for congruence among gene regions using NJ trees (i.e., no supported conflict), analyses of concatenated dataset were run using all specimens with at least three genes sequenced. Birth-Death models were run in BEAST2 (Bouckaert et al., 2014) using 100 million generations, sampling every 10,000 generations. Two runs were performed for each evolutionary model and convergence were assessed using Tracer v1.6 (Rambaut, Suchard, Xie, & Drummond, 2014). Runs were combined using LogCombiner and the maximum clade credibility (MCC) tree was chosen using TreeAnnotator from the BEAST2 package. Maximum likelihood and Bayesian inference analysis of concatenated matrices were run using IQTree (Trifinopoulos, Nguyen, von Haeseler, & Minh, 2016) and MrBayes v3.2.6 (Ronquist & Huelsenbeck, 2003), respectively. Bayesian inference ran for 100 million generations, sampling every 10,000 generations. Convergence of runs was assessed by Tracer v1.6 (Rambaut et al., 2014) and RWTY R package (R Core Team, 2017; D. Warren, Geneva, & Lanfear, 2016; D. L. Warren, Geneva, & Lanfear, 2017). A Maximum Parsimony tree of the concatenated matrix was generated using software MEGA7 v.7.0.18 with a thousand replicates for bootstrap (Kumar et al., 2016). The use of multiple methods can increase the confidence of the results as each analysis has its own assumptions (Alfaro, 2003). In the case of BEAST and MrBayes, both uses Bayesian inference, but MrBayes needs an outgroup and does not generate ultrametric trees that are necessary to perform further analysis of species delimitation. Supported deeper clades from concatenated trees were analysed separately. Two runs were performed for 16S gene alone using BEAST2 on a run of 10 million generation with 10% burn-in under Birth and Death model (Bouckaert et al., 2014). Convergence was assessed using Tracer v1.6 (Rambaut et al., 2014). Runs were combined using LogCombiner and the maximum clade credibility (MCC) tree was chosen using TreeAnnotator from the BEAST2 package. GMYC (Generalized mixed Yule Coalescent) (Fujisawa & Barraclough, 2013) and mPTP (Multi-rate Poisson tree processes) (Kapli et al., 2017) species delimitation algorithms were then performed. The analysis of each deeper 25 clade on its own, allowed us to use the 16S gene sequence without any ambiguous alignment, i.e. Gblock did not need to be used (Table 2.3). NJ trees of nuclear genes were used to assess congruence between GMYC and mPTP results. As in the 16S analysis, 28S genes did not present any ambiguous alignment when the deeper clades were analysed separately.

Table 2.3. Running conditions settings for each partition including size and model.

Partitions Length (bp) GBlock Length (bp) Model

EF1γ (1st and 2nd codon position) 456 - HKY

EF1γ (3rd codon position) 228 - HKY

18S 778 - JC

28S 774 731 HKY

16S 425 388 HKY+I

Biogeography Geographic coordinates were uploaded at the spatial portal from The Atlas of Living Australia (ALA.org.au). Twenty environmental variables, including 11 precipitation, 8 temperature and 1 soil moisture, were extracted for each sampling location (Supplementary table 2.1). Temperature and precipitation variables are the climate factors that most influence distribution and behaviour of tarantulas (Campbell & Engelbrecht, 2018; Kotzman, 1990; Stradling, 1994). Climate variables were chosen based on extreme factors as well as season and overall patterns. Soil moisture was included taking in consideration tarantula burrowing behaviour (Framenau, Baehr, & Zborowski, 2014; Schmitz, 2013). A Principal component analysis (PCA) was run using all 20 variables on statistical software PAST (Hammer, Harper, & Ryan, 2001). Coefficients of each explanatory variable were assessed based on component 1 and 2 of PCA. Variables that explained most of the variation between points were assessed and used to run a second PCA. Map layers tools at ALA were used to generate a map including the most significant explanatory variables and clade location points.

Results

26

Phylogenetic relationship of Australia tarantulas The most successfully amplified gene was 16S (100%) with a lower success rate in nuclear genes (EF1γ missing six sequences; 28S missing five; and 18S missing one sequence). The concatenated matrix of all four genes included 103 specimens, of which only 12 lacked a gene (i.e., included only 3 of the 4 loci). The mitochondrial gene 16S presented 93 parsimony informative sites whereas the nuclear genes presented 5 (18S), 39 (28S) and 45 (EF1y) parsimony informative sites. GBlocks excluded 37 and 43 positions from the 16S and 28S alignments respectively, which were mostly regions of ambiguous alignment due to indels. The NJ trees for each gene only present conflict on the shallow nodes of branches, where it is assumed to be interspecies genetic variation. Those sequences were not excluded from the analysis, as they show the genetic variability in populations. Phylogenies estimated using concatenated data lacked support at deeper nodes of all analyses performed (Fig. 2.1). Four undescribed taxa were not matched with any of the clades. Three of these taxa present morphological features that resemble each other and differentiate them from the other taxa (results not shown). The remaining unsupported taxon group is shown in figure 2.1 as sister taxon of the North East Queensland and Cape York clade (CYQC). This single specimen resembles morphologically some of the other undescribed specimens in the CYQC well supported clade (morphological data not shown). Due to the lack of support for these four taxa, they were excluded from further analysis. Four major clades were repeatedly recovered with support (Fig. 2.1). However, relationships within each of these major clades are still uncertain as many nodes lacked support. Selenocosmia effera (Simon, 1891), the only Southeast Asian species used in this project, is placed in the CYQC, but its exact position in the clade is unclear. The Australian Monsoon clade (AMC) presents specimens from the Northern Territory and Queensland (Fig. 2.3). AMC specimens from the east of Top End are a well- supported clade. Specimens from the western side of Top End and from Tanami Desert are also a well-supported group. Two samples from the Townsville area are separated, but one of them has a supported relationship with a sample from Talaroo (south centre of Cape York). The remaining specimens are scattered on the clade without significant support. Australian Eremean region clade (AEC) is the only clade recovered by all analysis. This clade is divided in three well-supported minor clades. One of the minor clades presents only specimens from Western Australia, while the other two present NT specimens and a single specimen from NSW (Fig. 2.3). The two minor clades present in NT almost overlap 27 with the closest samples being only 35 km distant. The presence of Broken Hill sample in Burt Plains clade makes very likely both clades, Alice Springs and Burt Plains, overlap in some point. AMC presents a specimen that is morphologically similar to Selenotypus plumipes and it is from the type locality of this species. AEC presents specimens that are like Selenocosmia stirlingi. However, they are all morphologically similar and confident match were not possible. CYQC has two Australian described species: Coremiocnemis tropix and Phlogius crassipes. Phlogius crassipes is separated in two clades with Coremiocnemis tropix and the other undescribed species nested in between them. Top End and Kimberley region clade (TKC) specimens could not be matched with any of the current Australian described species. Selenotholus foelshei and Selenocosmia subvulpina did not correspond to any of the specimens analysed in this project. Selenocosmia strenua resembles Phlogius crassipes and the morphological differentiation of these species is impossible.

Biogeography of Australian tarantulas The major clade distributions were consistent with some of the major terrestrial ecoregions in Australia. Thus, clade names have been designed to reflect the ecoregions in which members of the clade occur. In the Australian monsoon tropics, there are tree clades: Kimberley and Top End clade, Australian Monsoon Tropics clade and Cape York and North Queensland clade. The Australian Monsoon Tropics clade is the only which occur along its all extension, while the other clades are confined to the regions from which they take their names. Taking in consideration that there is probably a correlation between some environmental characteristics and the distribution of each major clade, PCA was run using 20 variables that could influence the distribution and occurrence of tarantulas (Supplementary material). Coefficients of each explanatory variable were assessed based on components 1 and 2. Three variables explained most of the variation between points: precipitation in warmest quarter, precipitation in autumn and in summer (Supplementary Table 2.1 and Fig. 2.2). Environment of AEC is separated from the other clades by the lower precipitation during summer, autumn and warmest quarter. Loading plots of component 1 shows that these variables are highly correlated, which explain that AEC is distributed over a wide geographical range, but the environmental factors variation where they occur seems to be very narrow. It is possible that AEC is a climate specialized clade, not occurring on a broad range of climate. The Environment of TKC lies fully within space of AMC. However, while AMC clade 28

Figure 2.1. Maximum Credibility tree generated from concatenated gene matrix under Birth-Death model by BEAST2. Squares symbolize support values, based on legend on top left. Acronyms on square legend means: Birth-Death (BD); Maximum Parsimony (P); Maximum Likelihood (ML); and Bayesian Inference (B). Clades names are shown on the right.

29 are very sparse in the graph, TKC is more restricted to the negative side of the Component 2, i.e. TKC clade are more negatively influenced by the precipitation on warmest quarter. AMC occupies a large area in the PCA graph, meaning that specimens and probably species of this clade are not geographically confined by climate variables as the other clades. CYQC mostly separate from rest but some overlap by AMC specimens from Townville and Cooktown area (Fig. 2.3). By the position of CYQC on the graph, it is possible to conclude that species within this clade prefer regions with wet summer and autumn.

Figure 2.2. PCA using three environmental variables collected from ALA. Component 1 corresponds to 84% of the variance and Component 2 corresponds to 12%. Colours correspond to deeper clade. AEC on blue, TKC on green, AMC on purple and CYQC on red.

Figure 2.3. Distribution map of the specimens used in this research. Colours correspond to deeper clade. AEC on blue, TKC on green, AMC on purple and CYQC on red. 30

Species delimitation The four major clades were analysed separately regarding species delimitation. The independent assessment permitted us to use the 16S and 28S without any exclusion from Gblock, as the alignment did not have any ambiguous regions. GMYC recovered 43 species, while mPTP retrieved 15 species (Supplementary figures 2.1-2.4). Taking the more conservative results of mPTP, it still recovered twice the number of currently described species in Australia. The Australian monsoon region clade (AMC) has 4 to 20 species (Supplementary figure 2.1). It is the clade with most the discrepancy between GMYC and mPTP results. GMYC recovered almost all terminals as species, apart from the clades from Kings Plains (QLD), Camooweal (QLD), Douglas-Daly (NT), Adelaide River (NT), Wongalara (NT) and Tanami desert (WA). The only similarity between GMYC and mPTP were the clades from Kings Plains and Camooweal. mPTP recognised as the specimens from Wongalara, Tanami and Maningrida (NT) a single species. The remaining specimens from QLD, ranging from Townsville to Longreach, south to Goondiwindi, and Top End samples were also combined as a single species. The Australian desert clade (AEC) has 2 to 4 species (Supplementary figure 2.2). GMYC found 4 species. All specimens in the southern inland area of Western Australia, from the border of South Australia to the Shark Bay area, are considered one species. The only individual analysed from the Pilbara is a species on its own. Burt Plains (NT), Tennant Creek (NT) and Broken Hill (NSW) samples make a third species. The fourth species is composed of individuals around Alice Springs (NT). mPTP recovered all NT and NSW samples as a single species and all WA samples as a second species. In the Cape York and North Eastern Queensland clade (CYQC), GMYC found 13 species (Supplementary figure 2.3). Among them are 2 species from Glennie Tablelands, one from The Tip of Cape York, one from Jardine River region, one from Iron Range, one from Cooktown, Coremiocnemis tropix from Cairns and five species that are supposed to be Phlogius crassipes. mPTP differences are one species from Glennie Tablelands, Cairns and Cooktown samples are grouped in one species and P. crassipes is divided in only 2 species. Using GMYC results, Top End and Kimberley clade (TKC) has 6 species (Supplementary figure 2.4). One species from Bonaparte Archipelago (WA), one from Derby region (WA), one from Darwin (NT), one from Daly River (NT) and two from Bradshaw (NT). mPTP surprisingly recovered the entire clade as a single species.

31

To resolve non-matching results between GMYC and mPTP, NJ trees of each nuclear gene were constructed. Each tree was analysed to identify allele sharing between putative species. The autosomal ribosomal gene 18S has very low variation of the samples analysed (Supplementary figures 2.5, 2.8, 2.11 and 2.14). In the AMC group, this gene grouped Kings Plains clade with Selenotypus plumipes clade and separated a sample from Townsville, while the remaining samples did not have resolution to be separated. AEC had one specimen from Alice Springs with one change, all the rest were identical sequences. The northern populations of P. crassipes were classed as a distinct species by mPTP from CYQC was detached from the other samples. The more diverse nuclear ribosomal gene 28S was mostly congruent with 16S results, apart from some inconsistencies (Supplementary figures 2.6, 2.9, 2.12 and 2.15). On AMC, S. plumipes and its sister taxa from Talaroo were separated and a sample from Judbarra (NT) and Middle Point (NT) were mixed. AEC present diverse of misplacements, mixing samples from WA and NT. On the other hand, CYQC presented resolution that separated all species recovered by mPTP, with the exemption of Cooktown and Cairns specimens that are disunited. TKC also presented a good resolution differing form most of the clades, except for those from Bradshaw and Darwin. The only protein coding gene used in this study (EF1y) also presented some inconsistencies (Supplementary figures 2.7, 2.10, 2.13 and 2.16). AMC had a specimen from Bachelor (NT) nest within Middle Point specimens; a sample from Muttaburra (QLD) has one allele within Barcaldine (QLD) clade and the other allele within Townsville sample; an allele from Middle Point is nested within Mt Bundey (NT) samples; presented incorrect topology of samples from Tanami and Wongallara and from Judbarra and Adelaide River. AEC showed mixed clades from Burt Plain and Alice Spring and, on the base on the tree, samples from WA and NT were mixed due to low resolution. CYQC EF1y tree is consistent with the finds of 16S and 28S, except by the two individuals from Cooktown being separated and one of them grouped with Coremiocnemis tropix specimen. TKC EF1y results corroborate the results of GYMC, apart of the mixed haplotypes from Bradshaw clades.

Discussion

Species diversity of tarantulas in Australia Australia’s continental size and harsh conditions makes field work rather difficult, expensive and dangerous (Ridpath, Williams, & Haynes, 1991; Woinarski, Mackey, Nix,

32

Traill, & McMillan, 2007). Australian tarantulas burrowing habits also present a challenge on sampling, with records of burrows reaching further than 100 cm deep and there are no techniques to lure the spiders (Kotzman, 1990). Soil diversification, such as rocky and hard clay soils, are also difficult to dig. Our attempt to sample tarantulas for most places around Australia failed, due to lack of DNA samples at museums, scarcity of populations during field work and inability to explore the southern half of the country. Analyses of the results allows a few conclusions about the phylogenetic diversity, relationship and biogeography of these spiders. In the most conservative interpretation of these data, Australia has 19 tarantula species. Fifteen species scattered on four supported clades and four species with inconclusive relationships. The eastern range of AMC overlaps its range with that of CYQC, more precisely on the southeast of Cape York and Townsville, which is the type locality of the only described species of AMC, Selenotypus plumipes Pocock, 1901. The genus Selenotypus is presently monotypic and endemic to Australia but species of the AMC could be placed in Selenotypus. A more in-depth taxonomic study can review morphological features to estimate the inclusion of those species into the genus. The absence of AMC specimens on the extreme north of Cape York is indicative that AMC is an exclusively Australian clade. However, only a study containing specimens from Asia can test this hypothesis. AMC presented GMYC results that recognized each sampled locality as a separated species. It can be the result of over estimation promoted by this method, as found by other researches (e.g. Opatova, Bond, & Arnedo, 2013; Satler, Carstens, & Hedin, 2013). GMYC method is also susceptible to sample size (Hamilton, Hendrixson, Brewer, & Bond, 2014) and to genetic structure, which is characteristic of mitochondrial DNA on mygalomorph spiders (e.g. Hedin, Carlson, & Coyle, 2015). However, the tarantula’s habitat specificity and sexual behavioural discrepancy, such as errant males and sedentary females, may explain GMYC results, as these tarantulas characteristics can promote geographic fragmentation of populations over time and space (Coyle & Icenogle, 1994; Main, 1987), generating a high biodiversity (Bond, Beamer, Lamb, & Hedin, 2006). On the other hand, mPTP results seem to under estimate AMC biodiversity. This unexpected result goes against most of the literature that found mPTP to overestimate biodiversity (Hedin et al., 2015; Ortiz & Francke, 2016, 2017). Due to geographic amplitude of the samples and we think that the real diversity of these clade is somewhere between GMYC and mPTP results. Two species had coherent results between both methods, Camooweal and Kings Plains. With certain confidence, we can say that these are new undescribed species. There are many locations to be sampled, which could uncover other 33 populations or putative species in Australia. The increase of sample size is needed to generate results that can solve the species status of the other specimens. Species recovered by both species delimitation methods have a large distribution area on AEC. The two species retrieved by mPTP are split in two by GMYC. Due to this inconsistency between both methods, no conclusive results could be achieved. The underestimation of mPTP can be seen again in the results of TKC. This method retrieves only one species for the entire clade. GMYC results seem to be more realistic based on 28S and EF1γ NJ trees. Derby and Fitzroy Crossing species has different allele composition from the remaining specimens. The same is found for the Bonaparte species. Both species are putative new undescribed species. TKC Bradshaw samples are divided in two minor clades. Bradshaw samples were collected in about 50 km apart, but they have enough genetic dissimilarity to be considered two species by GMYC. One group of samples was collected on top of a tableland with large rock being part of substrate and tarantulas burrowing in between gaps. These spiders seem to be much smaller (visually only, data not yet collected) than their sister group collected in lowland with substrate composed of soil and fragmented rocks. Kimberley is one of the most biodiversity-rich regions of Australia, being one of the Australian biodiversity hotspots (Australian Government, 2018). However, this region lacks sampling effort and the addition of more specimens for this region can solve species inconsistences and uncover new species. CYQC includes three known species: Coremiocnemis tropix, Phlogius crassipes and Selenocosmia effera. The presence of S. effera in the group indicates that the Australian tarantulas are not monophyletic, but tarantulas have migrated between Australian mainland and Southeast Asia, which corroborate the prediction of Main (1981). The direction of the tarantula migration is, however, difficult to predict only with the data of this project, but it has been speculated that Australian tarantulas migrated to Australia from the north (Main, 1981). Three species, apart from S. effera, are recovered by both species delimitation methods. Iron Range species appears to have a confined distribution in the tropical rainforest of Lockhart region. Tip of Cape York species were found in an unusual habitat, under logs on the most northern beach of Australia. Jardine River species were found just along the riverbank, but being sympatric with Phlogius crassipes. All those species appears to have a diminutive size (results not show). This dwarfism can be found in other species of tarantulas, such as the Aphonopelma anax (Chamberlin, 1940) group in USA (Hamilton et al., 2016). Phlogius crassipes was divided in two geographically distinct clades. However, Coremiocnemis tropix and another two undescribed species are nested in within P. 34 crassipes clades. It is probable that P. crassipes is composed of multiple cryptic species. Phylogenetics has been notable in relation to its ability to expose morphological cryptic species (e.g. Hamilton et al., 2016; Ortiz & Francke, 2017). Coremiocnemis tropix only shares with Asian Coremiocnemis, the presence of intercheliceral pegs (Raven, 2005; West & Nunn, 2010). Australian Coremiocnemis lack many of the genus synapomorphies, such as bilobular spermathecae, lanceolate embolus without distal flaring and long recurved setal bushes on metatarsus and tarsus IV (West & Nunn, 2010).

Geography of clades All variables that may explain the distribution of the Australian tarantula clades are related to precipitation and its seasonality. On a continental bioregionalization study using vegetation, precipitation was also the most important variable (González-Orozco et al., 2014). The same study found six Australian bioregions which tree can be correlated with the tarantulas’ distribution (González-Orozco et al., 2014). The Northern bioregion of Australia is composed of three major sandstone regions (Bowman et al., 2010; González-Orozco et al., 2014). Cape York Peninsula is the most eastern sandstone region and it has been in contact with Papua New Guinea (PNG) several times through the Sahul shelf (Hall, 2012). The presence of wet tropical rainforests and shared flora between Cape York and SEA (Crayn, Costion, & Harrington, 2015; Hall, 2012) makes very likely those places to share sister taxa in the CYQC, as observed with Selenocosmia effera. However, Selenocosmia javanensis (type specimen of Selenocosmia genus) is from Java (Walckenaer, 1837), located western of Wallace line (Evans, 2012), characteristic break of taxa sharing. A future study including Sunda and Sahul taxa can elucidate the relationship within Selenocosmiinae clades and their evolutionary history. During lowered sea level that exposed Sahul shelf, Cape York was also in total connection with the other two northern sandstone regions: Kimberley and Arnhem Land (Bowman et al., 2010). TKC appears to be the only clade restricted to those sandstone regions (Fig. 2.3). We speculate that soil composition may affect this clade distribution, but the only variable used here was annual mean of soil moisture and it did not show to have a significant impact. The presence of TKC on both Top End and Kimberley regions could be a reflection of the total connection between both lands during Sahul shelf exposition. An increase in sample size and localities and a phylogeographic study including molecular dating could test this hypothesis, showing the divergence dates of the species and the clade and checking if they match with any of the Sunda-Sahul connections. 35

The distribution of the AMC is mainly in the Northern bioregion, but also includes individuals/taxa from the northern part of the Northern Desert bioregion (González-Orozco et al., 2014), those two bioregions can be together considered as monsoonal tropics (Bowman et al., 2010; Crisp, Cook, & Steane, 2004). The monsoonal tropics are characterized by high summer precipitation and dry winters composed mainly of savannah vegetation (Crisp et al., 2004). Those seasonalities can be the determinant on AMC distribution. Eremaean bioregion, which is characterised by arid and semi-arid zone, overlaps with most of AEC distribution (Byrne et al., 2008). AEC northern distribution enters Northern Desert bioregion, which also includes some specimens of AMC (González-Orozco et al., 2014). The division of Eremaean and monsoonal tropics can be made not only by the amount of rainfall, but also by summer-winter precipitation patterns (Burbidge, 1960). It is possible that AEC is not only adapted to arid climate, but also to winter rainfall pattern. Rainfall period is well known to influence mating season of tarantulas (Campbell & Engelbrecht, 2018; Ferretti, Pérez-Miles, & González, 2010). Mating seasonality of AMC and AEC can be addresses in future studies which could explain geographic limitations of both clades based on behavioural aspects of the spiders and climate factors.

Conclusion Even though sampling was not comprehensive or exhaustive, this is unequivocally the most complete study on the group at the moment. The Australian theraphosid spiders’ diversity is underestimated. This study shows that the species count is at least double that currently described and highlighted seven current undescribed species. The four supported clades that probably constitute genera are geographically distributed in a way that coincide with Australian ecoregions, mirroring the biogeographical development of Australia.

Acknowledgement We would like to thank Mark Harvey from WAM for the donation of samples of specimens. Lachlan McIntyre, Eric Vanderduys, Volker Herzig, Kieran Palmer, Eamon Amsters and Kieran Aland for the sample donations.

36

Chapter Three

Australia's largest spider, the Eastern Tarantula (Theraphosidae: Phlogius crassipes), is a cryptic species complex

Contributions: Dr Robert J. Raven contributed with financial support for DNA sequencing, assisted field collection, creating ideas and draft edition; Dr Bryan G. Fry contributed with financial support for DNA sequencing; and Dr Lyn Cook contributed with financial support for DNA sequencing, assisted creating ideas, analysis of data and editing of drafts.

37

Abstract Phlogius crassipes, Australia's largest spider, is one of the most harvested spider species by the pet trade in Australia and yet its taxonomy is unrevised, and its habitat preferences, population structure, reproductive strategies, distribution and behaviour are unknown. Establishing its correct identity in a repeatable process is critical to its continued survival. Until now, the species has been considered as having a very wide distribution in north-eastern Australia and little morphological variation. However, spiders previously considered P. crassipes are shown to occur in highly genetically structured populations and, given the absence of isolation by distance and concordance across mitochondrial and nuclear genomes, they likely represent multiple species. Using phylogenetic methods and multi-locus data from the mitochondrial and nuclear genomes, and morphometric analyses of morphological characters, this study tests the hypothesis that it is a single biological species and rejects that. At the northern end of its range, a second species, Phlogius strenuus, was known but not ever later recorded. At least two species, one northern (Phlogius strenuus) and one southern (P. crassipes) are diagnosable using molecular data, but no consistent morphological or colour difference could be found between the species. With the transfer of Phlogius strenuus from Selenocosmia, the first ever species diagnoses along with known distributions of P. crassipes and P. strenuus are given. These changes have implications for conservation. P. crassipes is less widespread than previously thought; it has been confused with a second cryptic species (P. strenuus) and levels of harvesting of both need to be examined for sustainability.

Introduction Despite being large and well known to the general public, Australian tarantulas (bird-eating, barking or whistling spiders) are poorly known scientifically and taxonomically. There have been no comprehensive taxonomic revisions of the group and, to date, DNA sequence data have not been applied to test current classifications or assess species boundaries. The lack of understanding of species boundaries places Australian tarantulas at risk of over-collection by the pet trade; the Queensland Department of Environment and Heritage estimates that about 10,000 tarantulas are collected from the wild each year, mostly from Queensland (EHP Qld). Without good knowledge of the taxonomy of tarantula species in Queensland, there is a risk that species might be harvested unsustainably. Tarantulas (Theraphosidae) are generally long-lived spiders. Females can live up to 20 years and each will spend her entire adult life in or near her burrow (Ibler, Michalik, &

38

Fischer, 2013). Males, on the other hand, leaves their burrow after reaching adulthood to look for females and can live up to eight years (Criscuolo, Font-Sala, Bouillaud, Poulin, & Trabalon, 2010; Main, 1984). Juveniles disperse relatively short distances and normally move only a few metres from the maternal burrow (Shillington & McEwen, 2006). Such sex-biased and small-scale dispersal can generate strong genetic structure within and among populations, and breaks in distributions can lead to speciation events (Avise, 2000). Indeed, geographically disjunct, cryptic species of mygalomorph spiders, including theraphosids, have recently been recognised on other continents using molecular genetics (e.g., Bond, 2004; Hamilton, Formanowicz, & Bond, 2011; Hendrixson, Guice, & Bond, 2015) The largest tarantula in Australia, the Eastern Tarantula (Phlogius crassipes (L. Koch, 1874)) has, until now, been considered to have an extensive distribution from the tip of Cape York Peninsula to Rockhampton in Queensland (QM records; Main, 1985; Raven, 2000) (Fig. 3.1). This is a latitudinal range of about 1,600 km and covers habitats ranging from coastal wet forest, rainforests, open eucalypt forest and scrublands. Across this same region, many other non-flying animal taxa exhibit disjunct distributions between sister species (e.g., Bryant & Krosch, 2016; Edwards et al., 2016), likely due to barriers formed by several major landscape features in the region, e.g., Black Mountain Corridor (Bell, Yeates, Moritz, & Monteith, 2004) and Burdekin Gap (Hugall & Stanisic, 2011). Given the range of habitats in which the Australian Eastern tarantula occurs and the noted biogeographic breaks across which it is distributed, it is hypothesised that Phlogius crassipes, as currently recognised, is actually a cryptic species complex. Further compounding the complex taxonomic status of Phlogius crassipes are several changes in its generic placement over the past few decades. Originally described as Phrictus crassipes, it was soon transferred to Phlogius. Phlogius crassipes was transferred to Selenocosmia by Hogg (1902), which was later followed (notably by Main 1982, 1985), but Schmidt (1995) restored the species to Phlogius. Raven (2000) showed the material interpreted by Schmidt as the presumed holotype of the type species of Selenocosmia was actually collected well after the species had been described and thus incorrect, and returned the species to Selenocosmia. Later, Raven (2005) acknowledged the difference of bristles on the maxillary lyra between Selenocosmia javanensis (the type species of Selenocosmia) and that of Phlogius and implicitly restored the genus Phlogius and its type species Phlogius crassipes. In Australia, the species is formally named Phlogius crassipes (AFD, 2009) but remains in Selenocosmia in the World Spider Catalog.

39

An investigation of the distribution of P. crassipes is urgently needed as conservation management and allied decisions are based on documented species status. This study aims to assess the genetic status of Phlogius crassipes and test whether the current concept of this species is valid under a biological species concept. In addition to morphometric analyses, mitochondrial and nuclear genes were obtained from specimens across the known range of the species. Five genes were analysed in a phylogenetic framework and morphometric analyses were conducted to test the current hypothesis that P. crassipes is a single biological species.

Materials and methods

Species concept I apply a biological species concept to assess the species status of P. crassipes as it is currently recognised. Because of their different modes of inheritance, mitochondrial (mtDNA) and nuclear DNA can provide complimentary information about past gene flow. Given female tarantulas do not move far from the maternal burrow, it could be expected that there would be strong geographic structuring based on mtDNA. Nuclear genes, however, are spread by both males and females. Phylogenetic agreement among genes from both genomes can be interpreted as a lack of mixing of genes and decisions can then made about whether this is evidence of long-term reproductive isolation: equivalent to applying a biological species concept in which evidence for a long-term lack of gene flow between populations is interpreted as evidence for reproductive isolation (e.g., Lin, Edwards, Kondo, Semple, & Cook, 2017; Lin, Kondo, Gullan, & Cook, 2013). Significant shifts in branching rates determined in a coalescence framework, and morphological differentiation, are considered additional evidence of long-standing reproductive isolation.

Sample information Prior to this study, 237 samples (separate registrations) including over 68 males and 169 females plus 69 samples only of juvenile specimens of P. crassipes were held by the Queensland Museum, and these were used to provide an initial estimate of the distribution of the species. Field sampling was conducted during 2012-2016 across this distribution (Figure 3.1) and, when possible, three specimens were collected per site. All newly collected specimens are deposited in the Queensland Museum (Supplementary material). All localities were GPS recorded and included in the Queensland Museum database, but here exact GPS coordinates are not given due to the conservation concern

40 that this would release important information to illegal collectors and threaten species or populations.

DNA extraction, PCR and sequencing The DNA of 59 tarantulas was extracted and sequenced: 46 specimens of Phlogius crassipes, six specimens of Coremiocnemis, three individuals of Selenotholus and three individuals of Selenotypus. Genomic DNA was extracted from muscle tissue of one leg using either a CTAB/Chloroform protocol (Lin et al., 2013) or a Qiagen DNeasy Tissue Kit (Qiagen, Inc., Valencia, CA, USA). We amplified two regions of mtDNA (16S and COI) and three nuDNA regions (28S, 18S and EF1g) using the primers and PCR protocols listed in Table 3.1. PCR products were prepared for sequencing by adding a mix of exonuclease I and Antarctic Phosphatase. Sequencing was performed by Macrogen Inc., Republic of Korea.

Table 3.1. Genes, primer sequences and PCR conditions used in this analysis.

Gene Primer sequence Annealing Primer Direction Reference region 5’ to 3’ temperature

EF1g ATTGCBGCNCA F (Ayoub, Garb, F78 GTAYAGYGG Step down Hedin, & EF1g CCTTGRTTGAAY 58°C to 40°C Elongation R Hayashi, 2007) Factor 1- R1258 TTCTTTCC

gamma* EF1g CAGTGGATTAG (EF1γ) F TaF CTTTGCTG David Ortiz, 48°C EF1g CGCATTTTATCT pers. comm. R TaR AGCCTCTG

CTGGTTGATCCT 18S 2880 F GCCAGTAG (von Dohlen & ribosomal 55°C TCCAAGAATTTC Moran, 1995) RNA (18S) 2883 R ACCTCT

28S S3660 F GAGAGTTMAAS 55°C (Dowton &

41

ribosomal AGTACGTGAAA Austin, 1998) RNA (28S) C (Whiting, Carpenter,

TCGGARGGAAC Wheeler, & A335 R CAGCTACTA Wheeler, 1997)

C12054 CGCCTGTTTAW (Arabi, Cruaud, 16Sa- F Couloux, & 16S CAAAAACAT mod Hassanin, ribosomal 48°C 2010; C. RNA (16S) C12054 CTGGCTYACGC 16S- R CGGTYTGAACT Simon et al., rRN CA 1994)

GGTCAACAAATC LCO F ATAAAGATATTG (Folmer, Black, Cytochrome 1490 G Hoeh, Lutz, & c oxidase 1 50°C TAAACTTCAGGG Vrijenhoek, (COI) HCO R TGACCAAAAAAT 1994) 2198 CA

Nineteen museum samples collected in the last 15 years failed to yield quality DNA to amplify full-length genes. New internal primers for 16S were designed to amplify small fragments (90 and 132 base pairs) allowing us to match those sequences with fresh material for full-length sequences (Briggs et al., in prep). A neighbour-joining tree was used to match museum specimens to clades derived using sequences from fresh material collected for this project.

Phylogenetic analyses Sequence traces were edited using the software Gene Studio v.2.2 (Gene-Studio, Inc., Suwanee, Georgia). Polymorphic sites in autosomal genes were phased using DNA Sequence Polymorphism v.6.10.01 (Rozas et al., 2017). Sequences were aligned using the Guidance2 Server with the MAFFT algorithm (Sela, Ashkenazy, Katoh, & Pupko, 2015) and then edited manually using MEGA7 v.7.0.18 (Kumar, Stecher, & Tamura, 2016).

42

Areas of ambiguous alignment were found in the 16S and 28S genes when outgroups were included. These alignments were further processed using Gblocks v.0.91b (Castresana, 2000; Talavera & Castresana, 2007) with settings to allow smaller final blocks and gaps positions within the final blocks. FASTA files were modified to fit each different software package using ALTER (Glez-Pena et al., 2010). PartitionFinder (Lanfear, Frandsen, Wright, Senfeld, & Calcott, 2016) and BModelTest v1.0.4 (Bouckaert, 2015) were used to determine the best fit model for each gene or gene partition (Table 3.2). Partition finder models were used on IQTree and MrBayes softwares and BModelTest on BEAST2. Models were also manually changed when over-parameterisation was detected by failure to converge on a parameter estimate (Table 3.2). 18S and 28S were concatenated and analysed together because they are closely linked in the ribosomal tandem arrays. COI and 16S were also concatenated when both were included in an analysis given they are linked as part of the non-recombining mitochondrial chromosome. Maximum likelihood analyses were performed for each locus (mtDNA, rDNA, EF1g) and two concatenated datasets (one containing only specimens for which all genes were successfully sequenced, 5-gene dataset, and a second comprising all specimens for which at least four genes were successfully sequenced, 4plus dataset) using IQTree (Trifinopoulos, Nguyen, von Haeseler, & Minh, 2016). Bayesian inference was performed on concatenated datasets using MrBayes v3.2.6 (Ronquist & Huelsenbeck, 2003). Tracer v1.6 (Rambaut, Suchard, Xie, & Drummond, 2014) and the RWTY R package (R Core Team, 2017; D. Warren, Geneva, & Lanfear, 2016; D. L. Warren, Geneva, & Lanfear, 2017) were used to assess convergence of all Bayesian analyses and over- parameterization. The six specimens of Selenotypus and Selenotholus were used as outgroups on IQTree and MrBayes analysis. These specimens were chosen based on the results of chapter two.

Species delimitation BEAST2 (Bouckaert et al., 2014) was used to generate two ultrametric trees from concatenated 16S and COI data for Phlogius to use in coalescence-based species delimitation. Only one representative of each haplotype was used. Runs comprised 20 million generations, 25% burn-in, and either a Yule or Birth and Death tree prior (Bouckaert et al., 2014). The maximum clade credibility (MCC) tree was found for each using Tree Annotator v1.8.4 (Rambaut & Drummond, 2010). The MCC trees were then 43 used with a single threshold Generalized Mixed Yule-Coalescent (GMYC) (Fujisawa & Barraclough, 2013) and Multi-rate Poisson Tree Processes (mPTP) (Kapli et al., 2017) species delimitation approaches. The potential species identified by the coalescence approaches, and a species hypothesis based on a COI barcode gap of 6% for tarantulas (Hamilton et al., 2011), were used as alternative hypotheses for testing using nuclear loci in StarBEAST (Heled & Drummond, 2010) in BEAST2 (Bouckaert et al., 2014). The Coalescent constant populations were selected and clock-rate and population mean options left as default. Harmonic mean, AICM and likelihood mean were assessed after two runs using Tracer v1.6 (Rambaut et al., 2014). Log files were checked using Tracer v1.6 and merged using LogCombiner v2.4.7 (Bouckaert et al., 2014) with a 10% burn-in. Hypotheses were evaluated by determining whether their likelihoods were significantly different. I calculated significance in two ways: firstly using Harmonic means and Bayes Factors, and secondly using the AICM, as per Lin et al. (2017).

Table 3.2. Substitution model applied to each partition in Maximum Likelihood and Bayesian analyses.

Length Partitions IQTree MrBayes BEAST2 (bp)

EF1γ 705 HKI+I HKI+I 123321

18S 845 JC JC 121123

28S 807 HKI+I HKI+I 123453

16S 393 F81+I F81+I 123145

COI (1st and 2nd codon 440 F81 F81 123121 positions)

COI 220 F81+I F81+I 121323 (3th codon positions)

Isolation by distance

44

Linear distributions, such as that of P. crassipes down the east coast of Australia, can lead to a pattern of isolation by distance (IBD) (Wright, 1943). IBD can occur when populations that are geographically closer are more likely to interbreed than those that are further apart. Because sampling disjunct populations across a species with IBD can appear as if there is reduced gene flow, and might be misinterpreted as species-level differences, we tested for a pattern of IBD using linear regression between pairwise genetic distance (combined mtDNA: 16S and COI) and geographic distance in PAST v3.11 (Hammer, Harper, & Ryan, 2001). To visualise any structure geographically, haplotype networks were constructed for mtDNA using TCS (Clement, Snell, Walker, Posada, & Crandall, 2002) in popART (popart.otago.ac.nz).

Morphometric analyses Twenty-six characters (Supplementary Material Table 3.1) were scored for 81 sexually mature preserved spiders with little or no damage from the collection in the Queensland Museum (accession numbers in Supplementary Material Table 3.1), including 29 that were also used for DNA. Measurements for all carapace and leg features were taken dorsally. All measurements were converted into ratios by dividing each by carapace length, thus standardising for body size. Measurements from males and females were analysed separately using Principal Component Analysis (PCA) in PAST v3.11 (Hammer et al., 2001) to assess whether there was congruence with genetic groups. Discriminant Analysis (DA) was also applied to determine whether genetic groups could be differentiated morphologically. IBD was assessed using linear regression on pairwise Principal Component scores of morphometrics (R Core Team, 2017; Vavrek, 2011) and geographic distance between samples on statistical software R (Bivand, Pebesma, & Gomez-Rubio, 2013; Pebesma & Bivand, 2005; R Core Team, 2017).

Results

Phylogenetic analyses Not all spiders yielded enough and/or quality of DNA to sequence all genes: only 31 specimens were successfully sequenced for all five genes (5PLUS-gene dataset: 25 P. crassipes, all three Selenotholus and all three Selenotypus). An additional 21 P. crassipes and six Coremiocnemis were obtained for four genes, creating an alignment of 58 specimens in the 4PLUS-gene dataset. Thirty-one P. crassipes were amplified for both

45

16S and COI (mtDNA), but only 26 were included in species delimitation analyses after identicals were removed. Forty-six specimens were included in the 16S-only analyses. There was up to 7.8% divergence within P. crassipes in 16S. A 6% COI barcode gap identified two geographically separated lineages of P. crassipes as potential species: a northern lineage (north of 17.5°S) and a southern lineage (between latitudes 18.2°S and 23.5°S) (Supplementary material Fig. 3.1), which corresponds geographically to the Townsville-Mareeba gap seen in the IBD plot. GMYC and mPTP also proposed the northern lineage as a distinct species (red, Fig. 3.1) but split the southern lineage in two: a southernmost lineage (blue, Fig. 3.1) between 18.2-19.5°S and a central lineage (green, Fig. 3.1) between 20-23.5°S. The level of support varied between delimitation methods and tree: mPTP recognised three species on both starting trees and, although GMYC also identified three putative species none had posterior probability (PP) values >0.95 (Fig. 3.1).

Figure 3.1. Maximum credibility tree under a birth-death tree prior generated by BEAST2 of concatenated mitochondrial genes from P. crassipes. Values in black represent GMYC/mPTP support using the Yule tree prior. Values in red represents GMYC/mPTP support on the Birth-Death tree prior. GMYC posterior probability are given in decimals. Clades identified as species using mPTP are indicated by an asterisk. Map shows locality of specimens in each clade, with muted colour representing museum samples sequenced only for small fragment of 16S.

46

The haplotype network of 16S shows the strong north–south phylogeographic structuring (Fig. 3.2). Plots of genetic versus geographic distance (IBD) show two distinct classes (Fig. 3.3), with a gap in genetic distance between 4-7% that does not correspond to geographic distance: there is no cline across the dataset as a whole. The break in genetic divergence is consistent with within-species diversity of <4.5% and between- species divergence of >7%. The step-wise break in genetic divergence at the 250 km distance class corresponds to pairwise distances between the Townsville region (red- orange haplotypes in Fig. 3.2) and the Mareeba region (green haplotypes in Fig. 3.2).

Figure 3.2. Haplotype network of 16S of Phlogius crassipes. Similar haplotypes are coloured in similar colours. The size of the circles in the network is proportional to the number of individuals with that haplotype. Bars on branches indicate the number of steps between haplotypes. The map shows the proportion of each haplotype as each sampling location.

Figure 3.3. Plot of genetic versus geographic distance for mitochondrial DNA (COI and 16S) including all P. crassipes sequenced. 47

The two species under the barcode concept and the three species under coalescent methods formed the two species hypotheses for comparison. AICM indicated no significant difference between the 2-species and 3-species hypotheses, but Bayes Factors (based on Harmonic means) indicated positive support for the 3-species hypothesis over the 2- species hypothesis (Table 3.3).

Table 3.3. Test of the 2- and 3-species hypotheses in *BEAST. Hypothesis AICM Harmonic mean Likelihood mean

2 species 15328.4 -7639.92 -7611.65

3 species 15328.03 -7637.03 -7611.61

Significance p > 0.83 -2.89

Maximum likelihood and Bayesian Inference of each autosomal gene recovered both the northern and the southern clades, but did not separate the southern clade into two (Supplementary Material Fig. 3.4–3.6). However, P. crassipes was rendered paraphyletic by having Coremiocnemis nested inside it (Fig. 3.4; Supplemental material Fig. 3.4 – 3.6). The six specimens of Coremiocnemis seem to represent at least three species (Fig. 3.4), only one of which is currently described (C. tropix). All three lineages of Coremiocnemis are nested inside Phlogius, with good support (Fig. 3.4). One is sister to the northern clade of P. crassipes, and the other two form a sister group to the southern clades of P. crassipes (Fig. 3.4).

Morphological analyses

All 19 museum samples were assigned to a clade represented by DNA from newly collected specimens using the mini-barcode (16S fragment). Males and females of P. crassipes are morphologically distinct, although there is some overlap (Supplementary material Fig. 3.11). PCA did not separately cluster the northern and southern lineages in either females or males (Fig. 3.5), but discriminant analysis (LDA) could distinguish males of the northern and southern clades (Fig. 3.6) and the two southern clades from each other (Supplementary material Fig. 3.10).

48

Figure 3.4. Maximum likelihood tree of the concatenated 4plus-gene dataset. Red clade represents the northern lineage, and the green and blue clades represent the two southern lineages. Black tips are species identified as other genera. Habitus of Phlogius and Coremiocnemis tropix (from Raven 2005) are shown on the right, with the distinguishing character of Coremiocnemis (intercheliceral pegs) also shown.

Figure 3.5. PCA of morphometric values from females (left) and males (right) Blue polygon represents southern species and red polygon represents northern species.

49

Figure 3.6. LDA of morphometric values from females (left) and males (right). Blue bars represents southern species and red bars represents northern species.

Figure 3.7. IBD of morphology female (R-squared: 0.0412; p-value: 1.626e-11) (left) and male (R-squared: 0.0001892; p-value: 0.745) (right).

Discussion

Is Phlogius crassipes a cryptic species complex? We have found strong evidence that P. crassipes, as currently recognised, represents more than a single biological species. Species delimitation methods based on mtDNA recovered at least two putative species (Fig. 3.1): a northern lineage and a southern lineage. In taxa like tarantulas, where females do not disperse far from the natal burrow, it could be expected that there would be a strong pattern of IBD (Isolation by Distane) in mtDNA markers: haplotypes and haplotype clusters should be strongly geographically patterned. Indeed, this is what we find when looking at the geographic distribution of haplotypes (Fig. 3.2). However, for P. crassipes, the recognition of putative

50 mtDNA species is not the result of the uneven sampling across the north-south distribution of P. crassipes because no cline was detected in tests of IBD (Fig. 3.3). Instead, the IBD showed a distinct break in genetic divergence between 4.5-7% that was not correlated with increased geographic distance. If it were, there would be a general positive slope among the coordinates whereas there are two classes of coordinates: one (lower points) with low genetic divergence and a positive slope at distances less than 750 km, which is likely within-species IBD, and one (upper points) with no correlation at distance classes greater than 250 km, which is likely to represent the between-species distances. Furthermore, the split into northern and southern lineages based on analyses of mtDNA are reciprocated in each of the two nuclear genes. Given that nuclear genes are more slowly evolving than mtDNA, sort independently via sexual reproduction, and have a deeper expected time to coalescence (Avise, 2000), the congruence between the mtDNA genes and each of the nuclear genes indicates that there has been no gene flow for a very long time (2Ne generations, or 2 x the effective population size x generation time). Taking a purely statistical approach that ignores mating systems, the chance of obtaining the same two groups with a random assignment of nuclear alleles to the two mtDNA groups is very small (P<10-6). Perhaps the strongest evidence for more than one species under the current concept of P. crassipes is that species belonging to the genus Coremiocnemis are nested among specimens currently recognised as P. crassipes (Fig. 3.4). That is, the northern and southern lineages of P. crassipes are not each other's closest relative, and each is more closely related to a clade of Coremiocnemis than to each other. Although there is no necessity for a species to be monophyletic, indeed it is expected that all species will pass through a phase of non-monophyly early in divergence (Avise, 2000), the sum of evidence presented here (mtDNA, congruence with nuDNA, and non-monophyly) indicate that P. crassipes is not a single biological species.

How many species are in the Phlogius crassipes species complex? All lines of evidence indicate that there are at least two species under the current concept of P. crassipes but mPTP, and GMYC to a lesser extent, indicate that the southern lineage represents two species. However, neither of the nuclear genes recovered two separate southern clades (Supplementary material Fig. 3.5-3.6) and, although the two clades are present in analysis of concatenated data (Fig. 3.4), this is likely driven by the mtDNA genes. None of the clades were identified a priori by morphometric analyses of either females or males (Figs 3.5-3.7). This seems to be a common pattern in tarantulas, 51 as many of the species that have recently been discovered and named are cryptic, having first been recognised using molecular tools (e.g. Bond & Beamer, 2006; Hamilton, 2009). Here, we found that males of the northern and southern lineages are morphologically distinct when analysed using a discriminant function (Fig. 3.6), as are males of the two southern lineages (Supplementary material Fig. 3.10). This means a set of ratios obtained on males could be useful for morphological identification of either two or three species. Nevertheless, most collected specimens of P. crassipes are females or juveniles because males are rare as they leave their burrow to reproduce and most likely die in a few months after reaching adulthood (Kotzman, 1990). The up-shot of this is that identification to species, whether two or three are recognised, will require molecular tools in most cases. At this stage, we prefer to take a conservative approach and recognise only two species under the current concept of P. crassipes: erecting a new species at this stage might result in a future synonym. More specimens from across the entire distribution of P. crassipes, especially from between Bowen and Townsville, and more nuclear loci are needed. The latter are especially important to be able to assess the probably of ongoing gene flow, and new methods such as targeted gene capture (e.g. Hamilton, Hendrixson, & Bond, 2016) and DArTseq might provide relatively cost efficient ways of obtaining hundreds or thousands of nuclear loci.

What are the species? The holotype of P. crassipes (holotype at Zoological Museum Hamburg) has been examined by Dr Robert Raven (Queensland Museum), but no measurements or tissue for DNA were taken that could be included in statistical or molecular analysis, but descriptive morphological notes were recorded and pictures were taken. Given that the type locality of P. crassipes is Bowen, QLD (Koch, 1874), the southern clade is here considered to be representative of P. crassipes. This means that, although P. crassipes has been considered to occur as far north as the tip of Cape York Peninsular, it is now recognised as occurring only between the latitudes of 18.2°S and 23.5°S. We provide a revision diagnosis and description of P. crassipes below. Selenocosmia strenua (syn. Phrictus strenuus) is morphologically very similar to P. crassipes and, indeed, its description overlaps that of P. crassipes. At times it has been transferred to Phlogius (Schmidt, 1995; E. Simon, 1887; Thorell, 1881) and then to Selenocosmia (Rainbow, 1899; Raven, 2000). Most specimens collected from its type location have been identified as P. crassipes (e.g. Hawkeswood, 2003; Schmidt, 1995; 52

Queensland Museum labelled specimens). Here, we recognise the similarities between P. crassipes and S. strenua and consider S. strenua the northern cryptic sister species of P. crassipes. We transfer S. strenua to genus Phlogius, increasing its known range from 17.5°S to the tip of Cape York Peninsular.

Do geographic breaks in lineages correspond to known biogeographic barriers? Phlogius crassipes and P. strenuus distributions are separated by the Burdekin- Lynd barrier (Ford, 1986). This phylogeographic barrier has been reported for birds and is characterized by cooler climate, high precipitation and elevation within the Wet tropics (R. D. Edwards et al., 2017; Ford, 1986). The well described dry Black Mountain Corridor (Bryant & Krosch, 2016) seems to have no effect on P. strenuus genetic structure. The P. strenuus distribution suggests that they invaded the tropical rainforests of the Wet Tropics region. While the P. crassipes occurs on lowland in the southern half of the Wet Tropics, through to the dry area of Burdekin basin to rainforests around Airlie Beach, it reaches its southern distribution in the forests in the Rockhampton region. The secondary split of P. crassipes into a southern and a central distributed clade coincides with the Burdekin Basin, which has been reported a geographic barrier for other taxa (e.g., R. D. Edwards et al., 2017). Characterised by reduced rainfall, phylogeographic breaks have been reported for frogs, lizards and birds (Chapple, Hoskin, Chapple, & Thompson, 2011; James & Moritz, 2000; Joseph & Moritz, 1994). The central clade appears to be confined to more humid areas, such as Magnetic Island, Alligator Creek and Wallaman, whereas the southern P. crassipes specimens are distributed in an area with diverse vegetation types, such as dry woodland and closed humid creek edges forests. Some terrestrial vertebrates have speciated due to breaks in populations by emergence of a geographic barrier, which intersects P. crassipes range (e.g. Hoskin & Couper, 2014; Potter, Cooper, Metcalfe, Taggart, & Eldridge, 2012; Wilmer & Couper, 2015). Taking in consideration that invertebrates tend to be more habitat-specific than vertebrates, it is expected that populations of P. crassipes present breaks along their distribution. Invertebrates are normally subject to microclimate differences and have smaller habitat range (Hurd & Fagan, 1992; Santana, 2015; Uetz, 1991). Snails and beetles of the fragmented and small tropical rainforest of north-east Queensland have undergone speciation events due to expansion and shrinkage of their habitat (Bell, Moritz, Moussalli, & Yeates, 2007; Bell et al., 2004; Hugall, Moritz, Moussalli, & Stanisic, 2002). However, as a burrowing animals, tarantulas can regulate their environment via control of temperature and humidity to some degree (Mason, Tomlinson, Withers, & Main, 53

2013). Adult males leave their burrows when the ambient conditions are suitable. Although in doing so, they may not be regulated by microclimate factor as other invertebrates, which could explain their diverse distribution, they are still susceptible to extreme climate and/or combination of factors from some phylogeographic barriers.

Implications for generic classification Our results show strong evidence of paraphyly of both Phlogius and Coremiocnemis (Fig. 3.4). Coremiocnemis was originally described from South-east Asia and it has several morphological features that differ from the only species of Coremiocnemis currently described from Australia (C. tropix, West & Nunn, 2010). South- east Asian species of Coremiocnemis present bilobular spermathecae, lanceolate embolus without distal flaring and long recurved setal bushes on metatarsus and tarsus IV. The only character shared between Australian Coremiocnemis and the congeneric South-east Asian species is the presence of intercheliceral pegs. Australian species of Coremiocnemis resemble P. crassipes, with the only remarkable difference being the presence of intercheliceral pegs in C. tropix. The morphological difference between South- east Asian and Australian Coremiocnemis indicates doubt in the placement of the Australian species. It is likely that C. tropix belongs in Phlogius but the low sampling number of this species included in this project do not allow enough data to address the placement of C. tropix or the two other lineages sampled here.

Conservation Collection of tarantula can be challenging. It can take several hours or even days of fieldwork to locate a population. In this study, many locations, such as some locations around Townsville, Mackay and Rockhampton, sampled previously were re-visited but no burrows were found. Phlogius crassipes and P. strenuus have large linear distributions that do not overlap. Linearly distributed populations along the species range may serve as genetic corridor where the genetic diversity of the species is preserved. The unsustainable harvesting by the pet trade and increase of human activities on their habitats can create a gap between populations, breaking the gene flow. It is important that habitats are preserved and collection is regulated by policy.

Taxonomy

Genus Phlogius Simon, 1887

54

Phrictus L. Koch, 1874: 488. Type species by monotypy, Phrictus crassipes L. Koch, 1874; holotype female in ZMH. Phlogius Simon, 1887d: 195, replacement name for Phrictus L. Koch, 1874, preoccupied in the Hemiptera. Removed from the synonymy of Selenocosmia by Schmidt, 1995k; explicitly replaced in synonymy by Raven (2000) but implicitly accepted by Raven (2005: 17), https://biodiversity.org.au/afd/taxa/Phlogius. Diagnosis: Phlogius differs from Selenocosmia in the numerous plesiomorphic paddle-shaped (rather than short domed) lyra setae and in lacking the intercheliceral pegs. Differs from Australian Coremiocnemis by absence of interchecliceral pegs. Differ from South East Asian Coremiocnemis by lack of intercheliceral pegs, monolobular spermathecae, distal flaring embolus and lack of plumose and thicker forth leg. Males and females differ from other Australian theraphosid genera by the much thicker tibia I in a proportion of about 1.3 times thicker than tibia IV. Diagnostic description: Males and females from the southern and northern ends of the range have a dark to black sternum and leg coxae and this contrast with the lighter orange cuticle of the maxillae (Fig. 3.8-3.11). The colour of the sternum and legs coxae is partially from the dark prostrate mat hairs, but is certainly distinct in the darker cuticle colour, also. The colour correlates well with the dark brown femora and light brown patella—tarsi. This combination occurs in specimens from Rockhampton to Bamaga. (N.B: these colour differences are retained in preserved specimens.) Throughout that range, and often sympatric, are spiders in which the contrast is very weak to absent, both in the femora and sternum-maxillae. The contrast is inconsistent and less distinct in spiders from Townsville and Cooktown.

Phlogius crassipes (L. Koch, 1874) New combination (Figs 8--9) Phrictus crassipes L. Koch, 1874a: 489, plate. 37, fig. 5; Simon, 1887d: 195. Holotype female, Bowen, northeastern Queensland, examined by RJR. Selenocosmia vulpina Hogg, 1901: 246, fig. 28, holotype male, Cape Upstart (near Bowen, northeast Queensland, in BMNH 46/73, examined by RJR. First synonymised by Main (1985b: 47). Selenocosmia crassipes: Main, 1982c: 588; Main, 1985b: 47. Selenocosmia vulpina: Smith, 1986b: 127, fig. 48h; Smith, 1987d: 127, fig. 48h.

55

Phlogius crassipes: Schmidt, 1995k: 10, figs 1–3. 5 (transfer from Selenocosmia, rejected by Raven, 2000 implicitly rescinded by Raven, 2005); Schmidt, 2002h: 1, fig 5; https://biodiversity.org.au/afd/taxa/Phlogius_crassipes Material examined. Queensland Museum collection number: S2674; S2677; S2687; S2692; S2695; S10510; S2703; S2703; S2698; S10465; S10468; S10474; S10482; S10489; S10491; S10496; S10502; S105043; S105045; S105049; S10507; S10510; S105877; S105880; S10616; S10631; S10673; S10674; S10690; S10695; S10716; S10749; S34610; S4553; S63893; S67299; S6801; S97175; Diagnosis: Males and females are readily recognised by the dark brown femora contrasting with the patellae-tarsi on legs I-IV and the dark or light brown sternum and coxae with distinctly contrasting paler orange brown maxillae. Variation, sizes. Carapace length: adult females, 13.70-24.23; adult males, 14.75-20.82. Carapace width: adult females, 11.06-20.18; adult males, 12.25-18.80. Other measures found in the Supplementary material. Diagnostic COI nucleotides: 5T, 38G, 89A, 161A, 209G, 218A, 285A, 380A, 473A, 485T, 503A, 515A, 578G Remarks. Ischnocolus lucubrans L. Koch, 1874 is explicitly excluded from the synonymy and may be better placed in Phlogiellus.

Phlogius strenuus (Thorell, 1881). New combination Phrictus strenuus Thorell, 1881: 253. Syntypes (1 female, 1 juv., in MCG examined by RJR. Selenocosmia strenua: Main, 1985: 48. (Figs 10–11) Material examined. Queensland Museum collection number: S5383; S5086; S5462; S5394; S5366; S5451; S5431; S5375; S5384; S5450; S5390; S5454; S5441; S5378; S5405; S10466; S10467; S10470; S10509; S10661; S10677; S10682; S10683; S10687; S10698; S10704; S10715; S17729; S20389; S21324; S35388; S38795; S48240; S50862; S56432; S62159; S63729; S87203; S87217; S91202. Diagnosis: Males and females are readily recognised by the dark brown femora contrasting with the patellae--tarsi on legs I--IV and the dark brown sternum and coxae with distinctly contrasting paler orange brown maxillae. Variation, sizes. Carapace length: adult females, 13.00-24.17; adult males, 15.18-19.92. Carapace width: adult females, 10.74-21.45; adult males, 13.44-17.47. Other measures found in the Supplementary material. 56

Diagnostic COI nucleotides: 5G, 38A, 89T, 161G, 209A, 218G, 285T, 380G, 473T, 485A, 503G, 515T, 578A.

Figure 3.8. Phlogius crassipes (L. Koch, 1874), female, QM S63893: a. Prosoma, dorsal view; b. sternum, maxillae, labium and coxae, ventral view; c. Abdomen, ventral view; d. Carapace, dorsal view showing pilosity; e. Legs I, II, prolateral view; f. Legs III and IV, prolateral view; g. Maxilla, anterior face showing lyra; h. Spermathecae, ventral view; i. Eyes, dorsal view; j. Maxilla, posterior face.

57

Figure 3.9. Phlogius crassipes (L. Koch, 1874), male, QM S2698. a. Prosoma, dorsal view; b. sternum, maxillae, labium and coxae, ventral view; c. Abdomen, ventral view; d. Leg I, prolateral view; e, g. Palpal tibia, bulb and tarsus, lateral (e) and ventral (g) views; f. Embolus tip in widest view.

58

Figure 3.10. Phlogius strenuus (Thorell, 1881) female, QM S10698: a. Prosoma, dorsal view; b. sternum, maxillae, labium and coxae, ventral view. c. Abdomen, ventral view. d. Legs I, II, prolateral view; e. Legs III and IV, prolateral view; f. Eyes, dorsal view; g. Maxilla, anterior face showing lyra; h. Spermathecaem ventral view; i. Maxilla, posterior face; J. Carapace, dorsal view showing pilosity.

59

Figure 3.11. Phlogius strenuus (Thorell, 1881) male, QM S10698: a. Prosoma, dorsal view; b. sternum, maxillae, labium and coxae, ventral view; c. Abdomen, ventral view; d, e. Palpal tibia, bulb and tarsus, lateral (d) and ventral (e) views; f. Leg I, prolateral view; g. Embolus tip in widest view.

60

Chapter Four

A study of whether venom profiles match phylogenetic groups in Australian tarantulas (Theraphosidae: Selenocosmiinae)

Contributions: Dr Robert J. Raven contributed with financial support for DNA sequencing and assisted with draft editing; Dr Bryan G. Fry contributed with financial support for DNA sequencing and venom proteomics, assisted with draft editing; Dr Lyn Cook contributed with financial support for DNA sequencing, assisted creating ideas, analysis of data and draft editing; Kimberley Biggs assisted with data analysis and draft production; Dr Simon Blomberg assisted with statistical analysis; Dr Alun Jones assisted with mass spectrometry runs and analysis.

61

Abstract Species identifications performed using morphological characters alone may be unreliable, and DNA phylogenetic methods require invasive techniques. Alternative techniques are needed. Spider venoms are composed of a range of different compounds that reflect the wide range of habitat and niches that spiders inhabit, as well as their prey and predator populations. Their venom is known to vary between species, populations, sex, age and even maturity. It has been suggested that spider venom might provide the basis for a new method of identification as each species contains a unique cocktail of molecules. However, before this method can be developed and implemented, a more comprehensive understanding of the relationship between venom of different phylogenetic groups is required. Here we show that each phylogenetic group generated by 90 Australian tarantulas contain at least one unique peptide that could be utilised to distinguish it from other phylogenetic groups. Whole-venom profiling revealed that individuals within a phylogenetic group vary more than between-group variation. However, our analysis revealed some unique peptides in each group. Our data suggest that venom variation between phylogenetic groups may be a useful method for species identification of morphological challenging species. This new technique could be used in a myriad of different situations that require spider identification, such as quarantine identification and ecological research.

Introduction Spiders have developed a number of adaptations in order to exist in a diverse range of habitats and niches. One such adaptation is the ability to hunt a wide range of prey (Sandoval, 1994; Uetz, 1992). This ability has ultimately led to spiders developing complex venom (Casewell, Wüster, Vonk, Harrison, & Fry, 2013; Escoubas, Diochot, & Corzo, 2000; Escoubas, Sollod, & King, 2006). Spider venoms are composed of three broad functional groups: neurotoxins affect the central nervous system (CNS) and peripheral nervous system (PNS); lytic peptides; and enzymes (Rash & Hodgson, 2002). These functional groups, however, are composed of a myriad of different compounds, including inorganic salts, amino acids, neurotransmitters, peptides and proteins (Rash & Hodgson, 2002; Saez et al., 2010). Diverse factors can influence the presence of these different components within spider venom (Escoubas et al., 2006; Herzig, 2010; Santana et al., 2017). These include sex, maturity, location and phylogeny (Escoubas, Célérier, & Nakajima, 1997; Herzig, 2010; Herzig & Hodgson, 2009; Santana et al., 2017). Two studies led by Escoubas

62

(Escoubas et al., 1997; Escoubas, Corzo, Whiteley, Célérier, & Nakajima, 2002) have reported possible identification of species using venom. These variations highlight the uniquely complex nature of spider venom and demonstrate why spider venom profiling may be able to be utilised to determine the species of spider. Currently, only two methods of determining the species of a spider are available. The first involves morphology which is inexpensive but occasionally unreliable (Hebert, Cywinska, Ball, & DeWaard, 2003; Montes de Oca, D’Elía, & Pérez-Miles, 2016). The second method involves DNA sequencing which is more reliable but highly invasive (Barrett & Hebert, 2005). Therefore, a new method of classification that combines the reliability of DNA sequencing with the non-invasive nature of morphological classification would be valuable. Spider venom may potentially provide the basis for such a method as closely related spiders are hypothesized to contain similar elements within their venom (Chippaux, Williams, & White, 1991; Nentwig, Friedel, & Manhart, 1992; Schroeder et al., 2008). To develop the method, a more comprehensive understanding of the relationship between phylogenetic groups and venom is required. Phylogenetics provides a reliable tool to assess relationship and potential species boundaries. Here we compare the venoms of spiders from different phylogenetic groups within the family Theraphosidae. We sought to identify unique peptides from each phylogenetic group that could be used to differentiate these groups from one another.

Results

Phylogeny A maximum credibility tree was produced to determine the phylogenetic relationships between selected Australian theraphosid spiders. We used a concatenated matrix containing 16S, 18S, 28S and EF1γ genes on BEAST2 (SCIEX). Twenty-six different phylogenetic groups were determined based upon a 16S barcoding gap of 0.02 (Fig. 4.1). Almost all phylogenetic groups detected using this barcoding gap were also grouped together within the maximum credibility tree. Only the venom samples for phylogenetic group “O” remained separate. All twenty-six phylogenetic groups, however, could be further grouped into three deeper clades using the high posterior probabilities evident within the maximum credibility tree (Fig. 4.1).

63

Figure 4.1: Maximum credibility tree of Australian Theraphosidae generated by BEAST2 using a concatenated matrix. Twenty-six different phylogenetic groups were determined to be evident, using a 16S barcoding gap of 0.02. Three deeper clades could also be determined based on the high posterior values evident within the maximum credibility tree. Phlogius crassipes is represented by group A and B. Phlogius strennus is group D. Coremiocnemis tropix is group E. Selenocosmia stirlingi is possibly in the orange group. Selenotypus plumipes is group R.

LC/MS Chromatograms produced from LCMS analysis of the venom samples were overlaid and compared based on the phylogenetic groups identified. These chromatograms allowed for a visual comparison of the composition of the venoms from the same and different phylogenetic groups. The peptide peaks for all samples within each phylogenetic group’s chromatogram aligned with minimal threshold adjustments. Only two samples were required to be offset, likely due to small differences between the buffers used. Adjustment was done by aligning the first peak that normally includes only salts and small hydrophilic compounds. The similar peptide peak patterns evident within the samples of

64 each phylogenetic group’s chromatogram therefore suggests that similar peptides may be present in venom of the same phylogenetic group (Supplementary material). Furthermore, different peptide peak patterns were observed between venom samples from different phylogenetic groups. However, there was considerable variation in peak intensity within all chromatograms, showing the great individual variation in quantity of these compounds.

DA and PCA Analysis Discriminant Analysis (DA) and Principal Components Analysis (PCA) analysis were conducted on the LC/MS results, retention time and masses, of venom samples in order to determine whether the venoms of different phylogenetic groups were significantly different from one another as indicated by the comparison of chromatogram patterns. PCA separated groups A, B, C, D and E from the other phylogenetic groups (Supplementary material). However, the remaining phylogenetic groups were not separated in the PCA analysis. Furthermore, groups A, B, C, D and E remained mixed whilst separating from the other major groups. Groups A, B, C, D and E were then removed from the analysis to observe the differences between the remaining phylogenetic groups (Supplementary material). The phylogenetic groups F, X, Y and Z separated in one direction while V and W were driven in another direction. This suggests that these phylogenetic groups are different from each other as well as the remaining phylogenetic groups. Once again, however, F, X, Y and Z did not segregate from one another and neither did V and W. Discriminant analysis reproduce similar results. Eight groups were separated by the analysis, but the remaining 18 phylogenetic groups remain mixed (Fig. 4.2).

Multivariate Analysis Peptide masses as well as their start and stop times were recovered using the LC/MS reconstruction tool in Analyst TF v1.6 (SCIEX). A multivariate analysis was performed to determine whether the phylogenetic groups could be distinguished from one another when considering the compound mass and LC/MS time. The multivariate analysis indicated that the variance between different phylogenetic groups (0.0003) was smaller than the variation between specimens within the same group (0.06). t-Test Analysis

65

t-Tests in MarkerView v1.3 (SCIEX) compare the isomers present within one phylogenetic group with the remaining phylogenetic groups in order to highlight the isomers that differ most between groups. This technique was used to determine whether unique isomers (i.e., peptides), could be identified within each phylogenetic group. The t- tests were conducted on the three deeper clades. Only the deeper clade one contained a single unique isomer, while the other two deeper clades lacked unique isomers.

Figure 4.2. Discriminant analysis of all samples of Australian theraphosid venoms. Letters represent phylogenetic groups.

However, when the t-tests were performed on the individual phylogenetic groups, all groups aside from A and B were found to have at least one unique isomer in their phylogenetic group (Table 4.1). Whilst phylogenetic groups A and B did not contain isomers that were only unique to their group, there was one unique isomer that was only evident in groups A and B together (Figure 4.3). Using the charge and retention times of the unique isomers, where possible, to match with the peptides recovered by LC/MS reconstruction using Analyst TF v1.6 (SCIEX), a list of unique peptides and isomers for each phylogenetic group was created (Table 4.1).

Discussion Seven species of theraphosid spiders are currently known from Australia (EHP, 2018; World Spider Catalog, 2018). However, if each phylogenetic group determined

66 within this investigation represents a species, than this diversity is highly underestimated. This underestimation in the diversity of Australian tarantulas is likely due the difficulty of resolving the species by morphology alone as well as the remote locations of some of these spiders which makes them difficult to accurately sample (Hamilton, Hendrixson, & Bond, 2016; Ortiz & Francke, 2017; Ridpath, Williams, & Haynes, 1991; Woinarski, Mackey, Nix, Traill, & McMillan, 2007).

Figure 4.3: t-test analysis of the isomer with 1284.1 m/z amongst all phylogenetic groups. The mass of this isomer’s corresponding peptide was determined to be 5027.

Table 4.1: Unique isomers and their corresponding peptide determined to be evident within the each phylogenetic group using t-test analysis on MarkerView.

Phylogenetic LC Peptide Phylogenetic LC Peptide m/z m/z Group time Mass Group time Mass

A N 955.8 23.0 3690

1284.1 24.8 3850 982.0 25.0 7105 B O 1081.4 17.2 4321

1297.1 23.7 3883

C 708.4 10.1 3541 P 918.7 23.8 4544

67

1925.8 19.8 3630

1272.9 9.3 3816 D 1853.0 24.5 3706 Q 1760.2 23.1 4351

1002.0 26.5 7029 E 634.1 21.6 1262 R 1139.6 33.3 6832

1554.3 13.4 7303

F 1587.7 25.0 1268 S 1315.7 23.7 3945

818.2 17.6 2034 G 1201.8 8.1 3604 T 1163.3 21.1 3487

1241.3 24.6 3635 H 958.6 4.5 3834 U 1834.2 9.2 3667

1988.8 18.7 3954

547.3 11.1 4328 1021.1 28.1 1036 I 610.9 4.0 3660 V 1236.6 16.5 3707 787.4 12.3 3933

787.9 16.2 4337 984.1 10.7 3889 874.9 28.8 1272 J 1489.5 5.6 2973 W 927.8 35.1 1235 2182.0 23.2 3635 1181.3 16.2 3583

K 682.2 22.5 4085 X 855.9 19.6 4342

1239.0 21.6 4190 L 806.8 19.3 3996 Y 1817.0 16.4 6890

971.0 16.1 3633 M 1672.0 27.9 6676 Z 1438.4 7.6 2828

The 16S mitochondrial gene can be geographically biased and overestimate the number of clades/species (Cooper, Harvey, Saint, & Main, 2011; Simon et al., 1994). The two venom samples that were determined to be in phylogenetic group O by 16S gene analysis were collected at the same location (a few metres apart). The separation of Group 68

“O” samples on the MCC tree could be due to several assumptions that the Bayesian inference method needs as input (Lemmon & Moriarty, 2004). The 16S genes, however, varied more between samples than did other genes. In previous studies, greater variation has been shown to provide better resolution on the tips of phylogenetic trees (Arnedo & Ferrández, 2007; Wang, Yang, Wang, & Yu, 2017). Furthermore, the 16S gene has been previously identified to be reliable for determining phylogeny (Arnedo & Ferrández, 2007; Barr et al., 2009; Nicolas et al., 2012; Wang et al., 2017). Due to the lack of support on the topology separating group O samples, phylogenetic groups using the 16S barcoding gap were followed. The spiders used herein were collected at a number of remote locations throughout Australia and their phylogeny was based on multi-loci concatenated analysis; this further indicates that the diversity of Australian tarantulas has been significantly underestimated. Although a comprehensive phylogenetic analysis could answer this matter, it is beyond the scope of this project. Visual comparison of chromatograms has been used in a range of different projects (e.g. Escoubas et al., 2002; Santana et al., 2017). This technique has the advantage of being cheap, fast and even untrained people can visually compare chromatograms. The presence and absence of peaks on chromatograms can be observed between phylogenetic groups but individual variation of peak intensity has been evident in most groups. The individual variation can be related to sex, maturity, location and phylogeny (Escoubas et al., 2006; Herzig, 2010; Santana et al., 2017). Despite the chromatographic differences, the venoms of each phylogenetic group could not be distinguished from one another using either DA, PCA or multivariate analysis. The PCA conducted in MarkerView compared the isomers evident within each venom sample. Some isomers of the same phylogenetic groups have the same m/z but a slightly different retention time from one another. The software does not “recognise” these isomers as the same, producing a false negative result (Tsuyama, Mizuno, & Masujima, 2012). Therefore, a multivariate analysis of the 50 most intense peptides was conducted. Peptides that had a mass of 2 Da from one another and were within 0.5 min of their start and stop time were altered to have the same value in order to provide a normalized result. However, more variation was detected between venom samples within the same phylogenetic groups than venom samples from different phylogenetic groups. Various research has been conducted on the composition of spider venom and the factors that influence its components. For example, de-Oliveria et al. (1999) hypothesized that male spiders generally have a more diverse range of peptides within their venom than that in 69 female spiders, due to the wandering behaviour of males (during breeding season). Furthermore, Santana et al. (2017) determined that the components evident within the venom of Australian tarantulas changes ontogenetically. Most spiders collected in our study were females, but some males or sub-adults were also used. This may have contributed to a greater variation between venom samples within the same phylogenetic groups. The initial deeper clade t-test analysis was uninformative as non-unique isomers were identified within each clade. This may be due to the rapid rate in which venom genes evolve compared to those of “house-keeping” genes used for determining phylogeny (Casewell et al., 2013). Spiders hunt a wide variety of prey but are also a common prey themselves to other spiders, birds and mammals (Théry & Casas, 2002). For this reason, spiders presumably have highly complex venom containing diverse peptides (Casewell et al., 2013). For this wide variety of peptides to develop, the genes that encode for these elements must evolve at a faster rate in comparison to “house-keeping” genes (Escoubas et al., 2002). This therefore suggests why unique isomers were not identified amongst the deeper clades. Whilst the spiders within these clades may be close phylogenetically, the venom of these spiders has most likely diverged more strongly. This putative faster rate of evolution in spider venom genes also explains why almost all the phylogenetic groups contained at least one unique isomer/peptide. The only groups in which unique peptides were not evident were groups A and B. However, when these groups were combined, the new group contained one peptide that was not evident within any other phylogenetic group. This suggests that groups A and B may actually be one phylogenetic group. This leads to the conclusion that spider venom profiling may provide characteristics suitable for use in identifying the species of unknown spiders. This process could be useful in quarantine inspections. Monitoring imports is highly important as invasive species of spiders may cause disturbances to native arthropod species (Vink, Derraik, Phillips, & Sirvid, 2011). Also, regulatory guidelines require the establishment of the diversity of the known fauna, and thus the ability to recognise imported new fauna versus existing fauna. Furthermore, Australian tarantulas are highly sought on the international black market and are therefore frequently illegally exported (Alacs & Georges, 2008). Therefore, if the venom of a spider found is profiled, the phylogenetic group it belongs to can be identified, ensuring that foreign species of spiders in imports can be immediately dealt with appropriately. Furthermore, spiders found within illegal exports can be potentially relocated back to their natural habitat.

70

However, before this practical use can become a reality, further experimentation is needed with a larger sample size of spiders and a more diverse range of sexes and ages. Whilst sub-adults and male spiders were included within our results, most of the venoms studied were collected from mature female spiders. As previously discussed, factors such as sex and maturity can influence the composition of a spider’s venom. Further experimentation that includes venom samples from more male and juvenile spiders must be conducted in order to ensure that unique peptides can still be identified within groups. Finally, this experimental process should also be conducted on different families of spiders to confirm that these results do not just apply to the family Theraphosidae in Australia. Consequently, based on the present results, it is unlikely that the species of an unknown spider can be determined when comparing its entire venom profile to that of other spider species. However, despite the various factors that can alter the elements evident within spider venom, a strong result of at least one unique peptide within almost every phylogenetic group was produced, suggesting that further investigations of spider venom may provide the basis for a new method of identifying species. Venom gland transcriptome sequencing should be done in a future to reveal the peptide composition for each unique compound revealing variation and relationship between venom samples.

Materials and Methods

Collection of theraphosid spiders Theraphosid spiders were collected by RCS with others from various areas throughout Queensland, New South Wales, Western Australia and the Northern Territory. The captured spiders were stored in a ventilated container at a constant temperature in a dark room where they were fed weekly. Multiple individuals from each species were sought from each area. However, due to the inaccessibility of some locations, there are gaps in our knowledge of Australian theraphosid species.

Venom and DNA Extraction Venom was extracted from each spider using electro-forceps with the voltage adjusted between 6 V and 12 V depending on the size of the spider. Venoms were then diluted in 1 ml of MilliQ water before their absorbance was measured at 280 nm using a Thermo Scientific™ NanoDrop 2000 spectrophotometer. Aliquots of 20 μg were separated and dried using a SpeedVac vacuum concentrator. The venom samples were then stored at –80oC until further use.

71

After providing venom, the third leg of every spider was then removed through autotomy. Each leg was stored at –20˚C in 100% ethanol. Genomic DNA was extracted from the muscle tissue of the removed legs using the CTAB/Chloroform protocol and a Qiagen DNeasy Tissue Kit (Lin, Kondo, Gullan, & Cook, 2013).

Phylogenetic analysis PCR amplification was performed using four genes: 16S, 18S, 28S and EF1γ. For the 16S sequences, the primers C12054 16Sa-mod and C12054 16S-rRN were utilised with an annealing temperature of 48°C. For the 18S sequences, the primers 2880 and 2883 were utilised within an annealing temperature of 55°C. The 28S sequences were amplified using an annealing temperature of 55°C with the primers 3660 and A335. Finally, the EF1γ gene amplification utilised the external primers EF1γF78 and EF1γR12 using step-down (58°C to 40°C), and internal primers EF1γTaF and EF1γTaR at an annealing temperature of 48°C. Sequencing was then performed by Macrogen Inc., Korea. All samples for the four genes were aligned using Guidance2 (Sela, Ashkenazy, Katoh, & Pupko, 2015). The alignments were then manually edited and models for each gene were identified using Mega7 (Kumar, Stecher, & Tamura, 2016). To determine the phylogenetic relationships evident a concatenated analysis of all four genes was performed by BEAST2 using one hundred million generations with data collection every 10,000 repeats. To group specimens, the barcode gap of 16S were analysed using p distance on ABGD (Puillandre, Lambert, Brouillet, & Achaz, 2012). A weak barcode gap of 0.16 to 0.25 was identified. The aligned 16S sequences were then utilised to produce a neighbour joining tree as 16S sequences have been determined to be a reliable marker (Barr et al., 2009; Nicolas et al., 2012; Wang et al., 2017). Groups were determined using a barcode gap of 0.02 as previous studies conducted found the mean interspecific and intraspecific divergence to be 0.0155 and 0.3107, respectively (Barr et al., 2009).

LCMS Analysis For each venom sample, 25 μl of 20 μg/μL venom underwent ultra-high- performance liquid chromatography (UHPLC)–mass spectrometry (MS) using a Zorbax 300SB C18 with a gradient of 2–40% Buffer B over 35 min, 40–98% Buffer B over 2 min, then 2 min isocratic elution at 98% Buffer B. Buffer B was composed of 90% acetonitrile and 0.1% FA in water. Buffer A was 0.1% FA in water. The UHPLC was directly coupled to a triple quadrupole time-of-flight mass spectrometer (Triple TOF 5600 mass spectrometer

72

(SCIEX, Concord, Ontario, Canada)) operated in positive ion acquisition mode. Data were acquired for 46 min over the m/z range of 350–2000 with a cycle time of 0.5 s.

Venom analysis The mass of the peptides within each venom sample was determined by utilising the LC/MS reconstruction function within Analyst TF v1.6 (SCIEX). Based on the phylogenetic groups, these chromatograms were grouped with PeakView v2.2 (SCIEX). Some samples within the chromatograms were required to be offset in order to align with the remaining samples. These offsets were required due to potential differences in the buffers utilised throughout the LC/MS process as different samples were run using different buffer batches. LC/MS results were then input into the software MarkerView v1.3 (SCIEX) set to a minimum retention time of 2 min, a maximum retention time of 36 min, a noise threshold of 100, a mass tolerance of 2 Da and a maximum number of peaks of 5000. PCA was then conducted. The principle components that explained more than 5% of the variation within the results were analysed. Phylogenetic groups that consistently separated from the remaining groups were then removed and further PCA analysis was conducted. A multivariate analysis was also conducted in R v3.4.1 utilising the packages nlme and reshape2 (Francis, 2017; Lenth, 2016). Before running the multivariate analysis, peptide masses within the LC/MS analysis data that were within 2 Da of one another and had start and stop times within 0.5 min were manually changed to the same values in order to ensure accurate statistical analysis. MarkerView v1.3 (SCIEX) was then utilised again to conduct t-tests on all samples. Each phylogenetic group was run against the remaining groups and the first 50 isomers were investigated for being only evident with the group analysed. The unique isomers identified were matched to their corresponding peptide manually through the comparison of the retention time of the isomers with the start and stop times of the peptide peaks within the mass chromatograms. The charge of the isomers within Analyst TF v1.6 (SCIEX) was also utilised to identify their corresponding peptides by calculating and comparing potential peptide mass.

73

Chapter Five

Venom profiling of a population of the theraphosid spider Phlogius crassipes reveals continuous ontogenetic changes from juveniles through adulthood

Note: This chapter is the identical reproduction of article published in the journal Toxins. However, early findings of this thesis changed the species name of this chapter to Phlogius strenuus.

Citation: Santana, R. C., Perez, D., Dobson, J., Panagides, N., Raven, R. J., Nouwens, A., Jones, A., King, G. F., Fry, B. G. (2017). Venom profiling of a population of the theraphosid spider Phlogius crassipes reveals continuous ontogenetic changes from juveniles through adulthood. Toxins, 9(4), 116. https://doi.org/10.3390/toxins9040116

Contribution: David Perez, James Dobson, Nadia Panagides assisted with laboratory work and result analysis; Robert J. Raven assisted with draft editing; Amanda Nouwens and Dr Alun Jones assisted with mass spectrometry runs and analysis; Dr Simon Blomberg assisted with statistical analysis; Dr Bryan G. Fry and Dr Glenn F. King contributed with financial support for venom proteomics and assisted with draft editing;

74

Abstract Theraphosid spiders (tarantulas) are venomous arthropods found in most tropical and subtropical regions of the world. Tarantula venoms are a complex cocktail of toxins with potential use as pharmacological tools, drugs and bioinsecticides. Although numerous toxins have been isolated from tarantula venoms, little research has been carried out on the venom of Australian tarantulas. We therefore investigated the venom profile of the Australian theraphosid spider Phlogius crassipes and examined whether there are ontogenetic changes in venom composition. Spiders were divided into four ontogenic groups according to cephalothorax length, then the venom composition of each group was examined using gel electrophoresis and mass spectrometry. We found that the venom of P. crassipes changes continuously during development and throughout adulthood. Our data highlight the need to investigate the venom of organisms over the course of their lives to uncover and understand the changing functions of venom and the full range of toxins expressed. This in turn should lead to a deeper understanding of the organism’s ecology and enhance the potential for biodiscovery.

Introduction Spiders are abundant generalist predators that provide effective population control of other arthropods, most notably phytophagous insects (Roth, 1993). They are found worldwide in all terrestrial ecosystems, except Antarctica. Araneae (Arthropoda: Arachnida) is one of the most speciose animal orders in the world (Raizer, Japyassú, Indicatti, & Brescovit, 2005), with >46,000 species, encompassing 4029 genera and 113 families (World Spider Catalog, 2018). Araneae is divided into two suborders, Mesothelae and Opisthothelae, with the latter suborder further subdivided into the infraorders Mygalomorphae and Araneomorphae. Mygalomorph spiders, known as primitive spiders, have two pairs of book lungs and a pair of parallel downward-facing chelicerae, each containing venom glands (Framenau, Baehr, & Zborowski, 2014). Theraphosidae is the most speciose mygalomorph family, with 140 genera encompassing 983 species (World Spider Catalog, 2018). Commonly known as tarantulas, bird-eating spiders, whistling or barking spiders, theraphosid spiders are the largest spiders in the world (Coddington & Levi, 1991). Like almost all spiders, they possess venom glands (King, 2004). The main function of spider venom is prey acquisition, with predator deterrence as a secondary function (Escoubas, Diochot, & Corzo, 2000; Sannaningaiah, Subbaiah, & Kempaiah, 2014; Vassilevski, Kozlov, & Grishin, 2009). Tarantula envenomations typically cause only mild symptoms in humans. However,

75 species of the genera Latrodectus, Loxosceles, Phoneutria, Atrax, Illawarra and Hadronyche can cause sickness and even death (Escoubas, Sollod, & King, 2006; Gray, 2010). Spider venom is a complex cocktail of different molecules including inorganic salts, small molecules such as amino acids, neurotransmitters and larger polyamines, peptides, and proteins (Escoubas et al., 2000; Sannaningaiah et al., 2014; Vassilevski et al., 2009). The different compounds in spider venom are thought to work synergistically, thus increasing venom efficiency (Chan, Geren, Howell, & Odell, 1975; Nagaraju, Mahadeswaraswamy, Girish, & Kemparaju, 2006; Vassilevski et al., 2009). It has been estimated that each spider venom contains more than 100 peptides, with some containing more than 1000, which provides an immense natural pharmacopeia (Escoubas et al., 2006). Spider venom composition may be influenced by different predatory niche factors such as sex, diet, habitat and climate, making the venom possibly species-specific and enlarging the opportunity for drug discovery (Vassilevski et al., 2009). Early studies of spider venoms focused on medically important spiders with a view towards development of antivenom (Escoubas et al., 2006). Since the turn of the century, our increased understanding of the molecular composition of spider venoms and the mode of action of spider-venom compounds has led to the development of spider-venom peptides as pharmacological tools, drugs and bioinsecticides (King, 2011; King & Hardy, 2013; Saez et al., 2010; Smith et al., 2015). The venom ontogeny of spiders has been studied in species of the infraorder Araneomorphae, with a focus on differences related to size, sex and defense behaviour (Binford, 2000; R. M. G. De Andrade, De Oliveira, Giusti, Da Silva, & Tambourgi, 1999; Herzig, Ward, & Dos Santos, 2004; Nelsen, 2013). In contrast, although differences in venom composition between sexes have been investigated in mygalomorphs using modern molecular tools, ontogenetic variation in venom has only been studied using differences in LD50 and venom yield (Herzig, 2010; Herzig & Hodgson, 2009; Palagi et al., 2013). In other studies, focusing on different aspects of venom, ontogenetic variation in venom composition was only reported on briefly (Escoubas, Corzo, Whiteley, Célérier, & Nakajima, 2002; Guette, Legros, Tournois, Goyffon, & Célérier, 2006). Thus, this study aims to provide, for the first time, a full venom proteomics characterization of ontogenetic variation within a population of mygalomorph spiders, namely the Australian tarantula Phlogius crassipes.

Results 76

LCMS Analysis of Venom LCMS analysis of tarantula venoms was performed to provide a visual means of comparing the variation in venom composition between size groups. Peaks in the chromatogram correspond to the detection of a toxin or group of toxins with similar hydrophobicity. However, this method does not provide a comprehensive overview of venom complexity as most LC peaks will be comprised of multiple toxins. Both extreme size groups showed a smaller number of peaks (Olkola extra small (OXS) and Olkola large (OL) with 9 peaks) (Figure 5.1a,d) compared with the intermediate groups (Olkola small (OS) and Olkola medium (OM) with 11 peaks) (Figure 5.1b,c). Small variations in intensity were noted in all chromatograms, but could be due to individual variation. Clear differences in peaks were observed between the smallest (Figure 5.1a) and largest (Figure 5.1d) groups while the intermediate groups (Figure 5.1b,c) did not differ significantly. Group OXS lacks the peaks around 7, 14, and 18 min that are observed in chromatograms from all other groups. Also, the OXS group has peaks of lower intensity at 10 and 15 min compared to its closest group (OS). Group OL lacks the peaks at 13, 28 and 30 min that are present in chromatograms from all other groups and it differs in the intensity of peaks at 15 and 21 min. Principal component analysis and discriminant analysis (PCA-DA) (95.5%) showed a pattern of change in venom according to spider size (Figure 5.2), with gradual changes in venom pattern. However, it is clear that the venoms from tarantulas smaller than 10 mm (OXS) are different from tarantulas bigger than 10.1 mm (OS, OM and OL). Although specimens from the OS and OM groups are mixed in the central area of PCA-DA graph, OS spiders are more close to OXS spiders, while OM spiders are closer to OL. Adult tarantulas bigger than 15.5 mm (group OL) are dispersed over a larger area of the graph, but they overlap with OM in a small region.

Shotgun Sequencing of Venom Peptides and Proteins LC-MS/MS of whole venoms (shotgun) was conducted to identify toxins found in each size group and determine if any changes in venom were occurring over the lifespan of P. crassipes. For a small number of spiders, it was difficult to extract a significant amount of venom. This was especially true for smaller specimens from the OXS and OS groups, which limited sample numbers for some analyses. The individuals selected to perform the shotgun analyses followed by LC-MS/MS were representative of each group.

77

Figure 5.1. Combined liquid chromatography (LC) chromatograms of specimens from the (a) OXS – Olkola extra small group, (b) OS – Olkola small group, (c) OM – Olkola medium group, and (d) OL – Olkola large group. Each colour represent a different specimen. Vertical axis is Intensity and horizontal axis is Time (minutes).

Figure 5.2. Principal component analysis and discriminant analysis (PCA-DA) analysis of Phlogius crassipes population from Olkola Aboriginal land. Tarantulas are grouped according to cephalothorax size. Each dot represents a specimen.

78

MS analysis revealed a total of 12 peptides/proteins in the OXS group, 11 in OS, 8 in OM and 8 in OL. These results contrast with the LC/MS data presented above that showed more complexity in the venoms of OS and OM spiders (11 peaks), and less complexity in venom from OXS and OL spiders (9 peaks). However, as the LC/MS peaks can have one or more toxins with similar hydrophobicity overlapping in the chromatogram, the venom can be more complex in the early stages of life of tarantulas. On the other hand, the lack of a comprehensive database to match all the toxins in the venom may be responsible for the non-corroborative results between the shotgun and LC/MS analyses.

Table 5.1. Theraphotoxins from P. crassipes matched against UniProt arthropod database. Mass Toxins OXS OS OM OL 4,112 µ-theraphotoxin-Phlo1a 4,155 µ-theraphotoxin-Cg1a 4,146 µ-theraphotoxin-Phlo1b

8,773 U8-theraphotoxin-Hs1b

3,822 U1-theraphotoxin -Spl1a

8,869 U36-theraphotoxin-Cg1a 3,284 µ-theraphotoxin-Phlo2a

3,413 U1-theraphotoxin-Cv1a 3,712 Jingzhaotoxin F7-15.33 3,941 κ-theraphotoxin-Pg1b 3,955 κ-theraphotoxin-Pg1a CRISP-2-Grammostola 45,314 rosea 47,520 Hyaluronidase (Fragment) 6,940 κ-theraphotoxin-Cg3a 1 3,681 δ-theraphotoxin-Cg1a 1

4,334 U3-theraphotoxin-Cg1a

4,366 U3-theraphotoxin-Cg1b 4,150 τ-theraphotoxin-Gr1b

A BLAST of shotgun-derived sequences against the UniProt database revealed matches with almost 30 spider-venom components. Some of these were present in all of the P. crassipes samples, while others were found in just one representative tarantula. Almost 20 theraphosid toxins matched with the samples analysed (Table 5.1). Homologs

79 of CRISP-2, U3-TRTX-Cg1a and μ-TRTX-Phlo1b toxins were found on all specimens. μ-

TRTX-Phlo1b inhibits the human voltage-gated sodium channel NaV1.7, while the functions of U3-TRTX-Cg1a and CRISP-2 are still unknown. Clearly, the venom of small juveniles is more complex and includes a larger array of compounds, while large juveniles and female adults have less complex venom. However, some toxins are exclusive to each group, such as homologs of μ-TRTX-Phlo2a (OL), μ-TRTX-Phlo1a (OM), U8-TRTX-Hs1b

(OS), U1-TRTX-Spl1a (OS), U3-TRTX-Cg1b (OS), U1-TRTX-Cv1a (OXS) and τ-TRTX-Gr1b (OXS). All three μ-TRTX-Phlo toxins were described from another population of Phlogius crassipes (Chow, Cristofori-Armstrong, Undheim, King, & Rash, 2015), but we found only two of them in the larger individuals. U1-TRTX-Spl1a is found in venom from Selenotypus plumipes, a species that cohabits with a different population of P. crassipes. Jingzhaotoxin F7-15.33, two TRTX-Pg and the six TRTX-Cg toxins are found in the Asian tarantula

Chilobrachys guangxiensis, while U1-TRTX-Cv1a is found in venom from the Asian tarantula Coremiocnemis valida, both from the same subfamily of the Australian tarantulas (Selenocosmiinae). The other three toxins were originally described from spiders in different subfamilies from Asia and America. Shotgun similarities reveal that representatives of each spider group have different venoms, having a maximum of 40% similarity when compared to all peptides matched (Table 5.2). However, when only comparing against theraphotoxin matches, the similarity increased to 50%–80%. The venom of the spiders from small to large does not reflect a linear progression of juvenile venom to adult venom; rather, the venom changes without pattern between life stages. Venom of the group OL is more similar to the group OXS and OS than of the group OM on both tables. Permutation analysis also shows a statistical difference between all groups using the complete collection of peptides or just the matched theraphotoxins (Table 5.3).

Table 5.2. Left: Simpson’s similarity index for all peptides matched from the database. Right: Simpson’s similarity index for all theraphotoxins matched from the database (NA = not applicable). OXS OS OM OL OXS OS OM OL

OXS 1.000 NA NA NA OXS 1.000 NA NA NA

OS 0.400 1.000 NA NA OS 0.688 1.000 NA NA OM 0.300 0.300 1.000 NA OM 0.700 0.700 1.000 NA OL 0.300 0.300 0.233 1.000 OL 0.583 0.500 0.500 1.000

80

Table 5.3. Permutation analyses of shotgun binary matrix data. p > 0.05 indicates a significant difference. Left: permutation including all toxins matched by database search. Right: permutation including just spider toxins matched by database search (NA = not applicable). OXS OS OM OL OXS OS OM OL

OXS 0 NA NA NA OXS 0 NA NA NA

OS 0.9889 0 NA NA OS 0.3328 0 NA NA

OM 0.9999 0.3918 0 NA OM 0.3965 0.2194 0 NA

OL 0.9993 0.3901 0.5508 0 OL 0.7763 0.8015 0.3801 0

Another 130 matches were found in the UniProt database (Supplementary Table). Most of these were peptides from the insect orders Diptera and Hymenoptera. Proteins from assassin bugs and ticks were also matched despite their more distant phylogenetic relationship to tarantulas and their different feeding methods. These matches were mostly to similarities with intracellular proteins or to similar toxins found in bees, wasps, ants, assassin bugs and ticks to the spider toxins. This could be an example of convergent evolution or recruitment of the toxin by the ancestors of these orders.

1D SDS-PAGE Analysis of Venom While LC/MS separates toxins on the basis of hydrophobicity, 1D gel electrophoresis separates them according to molecular mass. The high molecular mass gel bands, which correspond to large proteins and enzymes, appear to be similar in all groups (Figure 5.3). However, the intensity of these bands differs between individuals, possibly indicating that these high molecular proteins are more important at different life stages than others. We observed considerably more variation in the quantity and intensity of bands with molecular masses smaller than 25 kDa. The highest number of gel bands were found in the group OS (9 bands), followed by groups OXS and OM (8), and the group OL (7). These results are consistent with the shotgun analysis, which pointed to higher complexity for venom from the smaller spiders (groups OXS and OS). The intensity of the nine visible bands was examined using GelQuantNET software. Clustering results showed a gradual change in band intensity according to spider size (Figure 5.3). Group OXS appears to be the most different with just over 40% similarity with

81 the other groups. The OS group is almost 40% different from OXS and 20% from OM and OL. OM and OL differs only 3%, but they are 50% different from OXS.

Figure 5.3. Left. 1D SDS PAGE gel of representatives from a population of Phlogius crassipes. Left lane is molecular marker (protein ladder) followed by OL, OM, OS and OXS specimens, respectively. Right. Clustering analyses with Rho similarity and 10,000 bootstrap values from one-dimensional gel electrophoresis of P. crassipes individuals of four different sizes.

LC-MS/MS Analysis of 1D Gel Bands In addition to the visual comparison that can be made from the 1D SDS PAGE gels, LC-MS/MS was also conducted on the excised bands. The data provided by this analysis can detect lower abundance toxins missed in the shotgun analysis due to ion suppression. However, the shotgun analysis of venom can detect proteins and peptide often of lower molecular mass that cannot be resolved by the gel. The SDS PAGE gel and LC-MS/MS analysis of excised gel bands complement each other and provide a clearer picture of the venom that cannot be matched by the shotgun analysis. The Simple Correspondence Analyses showed a high number of peptides exclusive to each spider, with a few being found in all specimens (Figure 5.4). Some peptides were found in venom from two individuals as can be noted between OXS and OS, and between OM and OL. Searches identified 443 protein matches in the UniProt database from all individuals and bands. Apart from a few matches to theraphotoxins, most of the proteins were from dipterans (flies, fruit flies and mosquitos), hymenopterans (ants, bees and wasps), 82 assassin bugs and ticks. Interestingly, high molecular mass bands from OM and OXS spiders returned sequences that matched with Sictox-LhiαIA2a (Loxosceles sp.) and α- latrocrustotoxin (Latrodectus sp.), respectively. A match with CRISP-2 of Grammostola rosea was found in all specimens analysed except the representative from the OL group (Supplementary Table).

Figure 5.4. Simple Correspondence Analysis of representatives of group size from a population of P. crassipes. Each number corresponds to a single venom compound identified by Protein Pilot. Axes correspond to the two dimensions created by the analysis.

Discussion A clear shift in venom profile is evident from the chromatograms obtained from the smallest (OXS) and largest (OL) spiders (Figure 5.1A,D). While visual inspection of the chromatograms appears to show that venoms of the OS and OM groups are most similar, cluster analysis revealed that OM venom is more closely related to that from OL spiders. Despite the lack of drastic variation in SDS PAGE gels of venoms from the four groups of spiders, the shotgun analysis revealed that the venom profile varies between each size group. In addition to venom changes from juvenile to adult, sampling of venom from older tarantulas showed that the venom continues to change even after adulthood, as evidenced in the PCA analysis (Figure 5.2). Ontogenetic shifts in tarantula venoms have been reported previously (Escoubas et al., 2002; Guette et al., 2006; Herzig, 2010). Venoms of the Indian theraphosid spider 83

Poecilotheria rufilata and the African theraphosid spider Pterinochilus murinus show variation between juveniles and adults, though this variation mostly resulted from changes in toxin abundance rather than differences in venom profile (Escoubas et al., 2002). In addition, some ontogenetic changes recorded by this study may have been masked by the grouping of spiders according to sexual maturity, ignoring the size variation in juveniles to adults. Differences in both toxin abundance and venom profile were observed between juvenile and adult specimens in the Brazilian theraphosid spider Lasiodora parahybana (Guette et al., 2006). However, the authors did not take into account spider size, but rather grouped spiders according to age, which may not reflect a change in prey size specialization. Spider size might, to some extent at least, determine the maximum-sized prey a tarantula can subjugate, which could lead to different micro-niches according to spider size. These different predator–prey interactions might impose different selection pressures that lead to variations in venom composition. Shifts in venom composition due to a change in diet have been observed in snakes where prey changes from one taxonomic class to another as the snakes mature (Andrade & Abe, 1999; Gibbs, Sanz, Chiucchi, Farrell, & Calvete, 2011; Jackson et al., 2016). The change in venom to target larger prey (i.e., small vertebrates in this case), might enable spiders to exploit more energy-rich resources as a trophic strategy. The lack of sexually mature males in the sampling has made this research incomplete. Thus, changes in venom composition due to selection pressure exerted from exposure to different predators in adulthood was beyond the scope of this study but should be the subject of future work. Sexually mature male mygalomorph spiders are known to leave their burrows in search of females, exposing them to reptilian, avian and mammalian predators. Therefore, it would benefit the spider to have defensive toxins that target these animals in addition to their invertebrate-specific predatory compounds. For example, it was recently shown that venom from the tarantula Heteroscodra maculata contains algogenic peptides that induce pain in vertebrate predators and which are presumably used for defense (Osteen et al., 2016). The venoms of male tarantulas tend to show greater variation in composition compared to female venoms (Escoubas et al., 2002; Herzig, 2010; Herzig & Hodgson, 2009; Rocha-e-Silva, Sutti, & Hyslop, 2009). Herzig demonstrated that male Coremiocnemis tropix yield considerably less venom than females, but their venom contains a higher variety of toxin components (Herzig, 2010; Herzig & Hodgson, 2009). However, in the South American tarantula Acanthoscurria atrox, female venom contains 84

~50% more compounds than male venom, with just 15% shared toxins between venom from both sexes (Mourão, Silva, Block Jr, & Schwartz, 2007). Brachypelma and Grammostola venom differs quantitatively and qualitatively between sexes (Escoubas et al., 2002). We predict that P. crassipes male venom likely has a different composition to the female venoms analyzed here, as the ecology of sexually mature males differs from juveniles and adult females. From a biodiscovery perspective, the lack of protein matches to the compounds identified by LC-MS/MS and shotgun analyses are due to the absence of comprehensive studies of tarantula venoms, especially Australian tarantulas. Transcriptome studies are needed to fully understand and characterize these spider venoms. With no transcriptome available for this species, identification of the venom proteome was limited. Moreover, without a transcriptome, false matches may occur. Researchers often use adult males or female spiders to find venom toxins that can be used to develop new drugs and bioinsecticides (Gentz, Jones, Clement, & King, 2009; Herzig & Hodgson, 2008; King, 2004). It is estimated that the venom of each species of spider contains at least 50 peptide toxins (Escoubas & Rash, 2004; Tedford, Sollod, Maggio, & King, 2004). If the venom of a single species such as P. crassipes changes multiple times over its lifespan, the number of toxins with the potential for biodiscovery could be far greater. Therefore, studies are needed on the venom of juveniles to determine the full pharmacological potential of spider venoms. Researchers studying spider venoms should consider pooling venom from a spectrum of sizes/ages to obtain a better representation of the full range of toxins that can be expressed in the venom throughout the life of the animal.

Conclusions This study revealed that P. crassipes venom changes continuously according to spider size, which we used as a proxy for age. Older Olkola spiders with a prosoma size over 15.5 mm continue to modify their venom composition even after becoming adults. Changes in venom composition during the lifetime of P. crassipes could be due to a change in the prey the spiders encounter at different life stages. Males are known to have an errant life stage where they leave their burrows in search of mates. During this time, the males are exposed to large mammalian and reptilian predators that they do not encounter when living in burrows or under logs and rocks. It is possible that their venom incorporates toxins that enable male spiders to defend themselves from predators during this life stage.

85

Materials and Methods

Spider Collection P. crassipes specimens were collected at Olkola land, central Cape York, north Queensland. To ensure that all specimens collected were from the same population, samples were taken within a 1000 m radius from Killarney station (–15.425849, 143.495126). Thirty-four spiders of different sizes were sampled using active manual search during day and night. GPS coordinates were recorded when spider burrows were located. Spiders were collected under rocks and logs or excavated from burrows. This expedition was a component of the BushBlitz nature discovery program (bushblitz.org.au). Spiders were transported alive to the Queensland Museum then stored in a dark room in containers containing moisturized sphagnum at a temperature of 24 ± 1 °C. They were fed house crickets (Acheta domestica) once per week.

Venom Collection Electrostimulation was used to extract the venom after one month of rehousing. Spiders were not fed for at least one week prior to venom extraction. The milking technique consisted of a forceps wire soldered to an adjustable electric converter that was used to stimulate the muscle surrounding the venom glands to contract and release venom. Voltage was 12 V for spiders with prosoma >13 mm, 9 V for spiders with prosoma of 10–13 mm, and 7.5 V for spiders with prosoma <10 mm. Venom was collected into a 1.5 mL microcentrifuge tube, lyophilized, then stored at –20 °C.

Venom Proteomics Spiders were separated into four groups according to cephalothorax length: group OXS ≤ 10 mm; 10 mm < OS > 13 mm; 13 mm ≤ OM ≥ 15.5 mm; and OL > 15.5 mm. These groups were defined by observation based on the size range of the food the spiders could safely handle. Lyophilized venoms were dissolved in 1 mL of deionised water and protein content quantified from absorbance at 280 nm measured on a NanoDrop 2000 spectrophotometer (Thermo Fisher, Sydney, Australia). Aliquots were then freeze dried and stored at –80 °C until further processing.

LC/MS (ESI uHPLC, Mass Spectrometry) The extracts were analyzed by ESI LC-MS on a Shimadzu Prominence uHPLC (Tokyo, Japan) coupled to a Triple TOF 5600 mass spectrometer (SCIEX, Concord, Ontario, Canada) equipped with a duo electrospray ion source. 5 µL of each extract was

86 injected onto a 2.1 mm × 100 mm Zorbax 300SB-C18 1.8 µm column (Agilent, Santa Clara, California, USA) at 300 µL/min. The samples were eluted from the HPLC column using a linear gradient of 2% solvent B for 1 min, 1%–40% solvent B over 35 min at 300 µL/min flow rate, followed by a steeper gradient from 40% to 98% solvent B in 5 min. Solvent B was held at 98% B for 2 min for washing the column and returned to 2% solvent B for equilibration prior to the next sample injection. Solvent A consisted of 0.1% formic acid (aq) and solvent B contained acetonitrile/0.1% formic acid (aq). The ionspray voltage was set to 5300 V, declustering potential (DP) 100 V, collision energy 8V, curtain gas flow 25, nebuliser gas 1 (GS1) 50, heater gas 2 (GS2) 50, gas temperature 500 °C and interface heater at 150 °C. The mass spectrometer acquired 500ms full scan Time-of-flight mass spectrometry (TOF-MS) data over the mass range 350–2000. The data was acquired and processed using Analyst TF 1.6 software (SCIEX, Concord, Ontario, Canada).

In-Solution Sample Preparation for LC-MS/MS For shotgun protein sequencing, whole crude venom from a representative individual of each group was reduced, alkylated, digested and submitted for LC-MS/MS analysis. This was performed by first dissolving 10 μg of lyophilized venom in 20 μL of 8 M urea. Samples were then reduced by incubation at 37 °C for 30 min following addition of 10 μL of 15 mM dithiothreitol (DTT), then cysteine residues were alkylated by adding 10 μL of 100 mM iodoacetamide (IAA) and incubating samples in the dark for 30 min at room temperature. 10 μL of 15 mM DTT was then added to quench excess IAA. 50 μL of 50 mM ammonium bicarbonate (ABC) followed by trypsin (1:50) were then added and the sample was incubated overnight at 37 °C to digest venom peptides and proteins. The samples were then desalted using a C18 ZipTip (EMD Millipore, Billerica, MA, USA) according to the manufacturer’s protocol. Dried samples were resuspended in mobile phase A (5% acetonitrile, 0.1% formic acid in MilliQ water).

1D SDS PAGE Non-Reduced and Reducing 1D gel electrophoresis was conducted using a representative individual from each group following a modified Laemmli protocol (Ali et al., 2013); 40 μg of venom was loaded per lane, and gels were run at room temperature at 100 V until the loading dye reached approximately 10 mm from the base of the gel. “Non-reducing gels” were run using crude venom. “Reducing gels” were performed by dissolving venom samples in 5 μL of 3×

87 sample loading buffer (15 μL total volume) with 50 mM DTT followed by 4 min incubation at 100 °C prior to loading onto the gel.

GeLC-MS/MS Samples were separated using RP-HPLC on a Dionex Ultimate 3000 RSLC nano- system. Samples were desalted on a Thermo PepMap 100 C18 trap (0.3 × 5 mm, 5 µm) for 5 min using a flow rate of 30 µL/min, followed by separation on an Acclaim PepMap RSLC C18 (150 mm × 75 µm) column at a flow rate of 300 nL/min. For bands cut out of 1D gels, a gradient of 10%–95% buffer B over 7 min was used, whereas for whole-venom samples processed using reduction/alkylation/ digestion (shotgun), a gradient of 10%–95% buffer B over 90 min was used. In both cases, buffer A was 1% ACN/0.1% FA and buffer B was 80% ACN/0.1% FA. Eluted peptides were directly analysed on an Orbitrap Elite mass spectrometer (Thermo) using a nanospray ionization electrospray interface. Source parameters included a capillary temperature of 275 °C; S-Lens RF level at 60%; source voltage of 2 kV and maximum injection times of 200 ms for MS and 150 ms for MS2. Instrument parameters included an FTMS scan across the m/z range 350–1800 at 60,000 resolution followed by information-dependent acquisition of the top 10 peptides across the m/z range 40–1800. Dynamic ion exclusion was employed using a 15 s interval. Charge- state screening was enabled with rejection of +1 charged ions and monoisotopic precursor selection enabled.

Bioinformatics and Statistical Analyses LC-MS data were processed using the LC-MS peptides reconstruction tool in Analyst TF v1.6. LC graphs were edited to include the peak mass of the most relevant peaks. Combined LC graphs by groups were plotted using PeakView v2.2. LC-MS files were input on Maker View v1.2.1 software, which generates an isotopes table for all samples. Isotope tables were classified according to elution time, intensity of peak and isotope mass. To avoid MS misreadings, isotopes that occurred in less than two samples were excluded and a maximum of 400 isotopes were selected to analyse. Marker View software performed a principal component analysis and discriminant analysis (PCA-DA). Data from shotgun and 1D spot LC-MS/MS were converted to mascot generic format (mgf) using the msConvert software (ProteoWizard v3.0.9576) and Protein Pilot™ v5.0 (Sciex) was used to search for sequence matches against all Arthropoda sequences (Uniprot and Trembl) obtained from www.uniprot.org downloaded on 21 January 2016. The Protein Pilot search conditions were set to include alkylation method (iodoacetamide),

88 tryptic digestion, and allowing for urea denaturation and FDR analyses for shotgun and 1D gel modifications for 1D gel spots, including artefacts induced by the preparation or analysis processes. This was done to maximize the identification of protein sequences. Spectra were inspected manually to eliminate false positives. The list of toxins generated for each group were examined using Simple Correspondence Analyses (R Core Team, 2017; Vavrek, 2011). Shotgun and 1D gel spots results were plotted on a separated binary matrix table, reflecting the presence and absence of peptides for each specimen. Peptides with more than 99% confidence for shotgun or 95% confidence for 1D spots and total coverage of minimum 2 for shotgun and 1 for 1D spots were included in the matrix. Similarity of Simpson were applied to verify percentage of difference between specimens using the freeware PAST v3.11 (Hammer, Harper, & Ryan, 2001). Shotgun matrix was analyzed by permutation of samples where performed and observed data were plotted against expected data using the Fossil package for R v3.2.3 (R Core Team, 2017; Vavrek, 2011). 1D spots LC-MS/MS data matrix was analyzed by simple correspondence analyses using Mass package for software R v3.2.3 (Venables & Ripley, 2002). The intensities of 1D reduced gel spots were analyzed using the software GelQuantNET 1.8.2. Results were plotted as a matrix table showing the intensity of each spot from the gel. Clustering analyses using Rho similarity and 10,000 bootstrap samples were applied to verify the difference between specimens using the freeware PAST v3.11 (Hammer et al., 2001).

89

Chapter Six

General Discussion

90

This thesis applies a molecular approach to explore the phylogenetic diversity, biogeography and venom variation of Australian tarantulas. As such, it is the most comprehensive research done on theraphosids in Australia and the only one that covers the entire theraphosid fauna of a continent. Sampling is far from ideal and several populations and potential species have not been collected due to inaccessibility of some areas. Australia’s size and hash conditions makes field work difficult, expensive and dangerous (Ridpath, Williams, & Haynes, 1991; Woinarski, Mackey, Nix, Traill, & McMillan, 2007). In general, the phylogenetic diversity of this family in Australia has been unknown and the fauna undescribed. Two cryptic species have been described from Queensland. The use of venom profiling to identify species was not successful across the board. However, unique peptides within the venoms were informative within each phylogenetic group. Ontogenetic venom variation within populations is also shown.

Species diversity of tarantulas in Australia Australian ecosystem diversity has been shown to drive plant and animal taxonomic diversity (Bell, Yeates, Moritz, & Monteith, 2004; Bowman et al., 2010). Clearly, the Australian tarantula fauna is far more diverse than represented by the seven described species (Raven, 1985, 2005; World Spider Catalog, 2018). Despite many expeditions and extensive sampling conducted as part of this thesis, some populations of tarantulas remain undocumented. The species delimitation approach used here predicted a diversity two to six times greater than that described. Australian species belonging to the genera Phlogius and Coremiocnemis were found to be paraphyletic. However, the inclusion of “true” Selenocosmia species, e.g., from South-east Asia, are needed to take more sound decision regarding synonyms and diagnosis. Genetically structured genes show a large genetic gap through the middle of the distribution that was previously known as that of Phlogius crassipes distribution. Thus indicating that P. crassipes is at least two cryptic species, the other being Phlogius (formerly Selenocosmia strenua) strenuus, a species that has never been identified since its discovery. Both species have an overlapping morphological description. Phlogius strenuus was then considered a cryptic species with northern distribution, while P. crassipes is the southern species. The updated distributions of P. crassipes and P. strenuus have tremendous implications to their conservation. The Queensland Department of Environment and 91

Heritage estimates that about 10,000 tarantulas are collected from the wild each year, mostly from Queensland (EHP Qld). It is believed that most of them are P. crassipes. With the new range of the species, it is clear that populations, genetic diversity and even entire species could be at risk, based upon the now invalid assumption that the species might be harvested sustainably. In future, genetic diversity in P. crassipes should be addressed, as more potential cryptic species may be uncovered. All potential species highlighted in this thesis should be revised taxonomically, including morphological analysis and increasing the amount of specimens and genetic data. Failure to address those questions could result in possible over-collecting by the pet trade, endangering some regionally endemic cryptic species (see Mendoza & Francke, 2017). The loss of spider species correspond to a loss of one of the most important natural predators and a venom that is considered a great pharmacopeia (Escoubas & Rash, 2004; Roth, 1993).

Biogeography of tarantulas in Australia All tarantula samples from Australia included herein were lie in one of four main phylogenetic clades that are known only from few biogeographic regions. The disposition of those clades in the mainland seems to be related to precipitation and seasonality. Similar results has been found on vegetation (González-Orozco et al., 2014), but evidence in invertebrates are scarce. The two most widespread clades (AEC and AMC) do not overlap in distribution and the gap correlates with the break between Eremaean and Australian Monsoon Tropics ecoregions (González-Orozco et al., 2014). Those two ecoregions are not only differentiated by the amount of rainfall but also by summer-winter precipitation patterns (Burbidge, 1960). Rainfall seasonality is well known to influence mating season of American tarantulas (Campbell & Engelbrecht, 2018; Ferretti, Pérez-Miles, & González, 2010) and it could be the driver that restricts each clade in each ecoregion in Australia. Australian tarantulas seem to have immigrated multiple times from South-east Asia (SEA) and Australia, which corroborates the hypotheses of Main (1981). A land connection from Cape York to Papua New Guinea (PNG) has emerged several times through the Sahul shelf (Hall, 2012), and that may have been the bridge for taxa shared between regions, e.g., shared flora (Crayn, Costion, & Harrington, 2015; Hall, 2012). Recent parochial revisions of SEA tarantulas, omitting the Australian fauna (Nunn, West, & Von Wirth, 2016; West & Nunn, 2010; West, Nunn, & Hogg, 2012), only add confusion to the 92 taxonomy. Generic revisions using molecular data within the Selenocosmiinae are urgently needed to provide a more robust understanding of evolution in this morphologically almost invariant and complex group. Even though the evolution of Selenocosmiinae implies that species and genera are shared between Australia and SEA, Australian species are endemic and not found elsewhere (World Spider Catalog, 2018). The limited understanding of the real biodiversity of tarantulas will reflect in diverse areas of research and environmental management. Conservation of some areas may be undervalued by the lack of formally named species. Ecological research using unnamed species could result in improper data interpretation. In addition, pharmacological studies using tarantula venoms might be difficult to replicate due to mismatched use of species in venom pooling.

Variation of tarantula venoms in Australia Understanding how the venom of different taxa differ is essential to improve research using spider venom. Diverse factors are known to influence the components within spider venom, such as species, population, sex and maturity (Escoubas, Célérier, & Nakajima, 1997; Escoubas, Corzo, Whiteley, Célérier, & Nakajima, 2002; Herzig & Hodgson, 2009). However, this thesis used advanced technology analysis that gives a more comprehensive and robust results increasing our understanding of venom variation in tarantulas. Differences within phylogenetic groups could be visually observed in chromatograms. It is known that venom composition evolves relatively faster than housekeeping genes (Casewell, Wüster, Vonk, Harrison, & Fry, 2013). Difference in venom of phylogenetic groups could be related to the fast adaptation of the venom to environmental conditions, including the type of prey available, as observed in snakes (Andrade & Abe, 1999; Gibbs, Sanz, Chiucchi, Farrell, & Calvete, 2011; Jackson et al., 2016). However, individual variation of venom within the phylogenetic groups was also documented. This variation can be attributed to the inadvertent use of juveniles and sexually mature male (Herzig & Hodgson, 2009). Adult male mygalomorph spiders are known to leave their burrows in search of females, exposing them to reptilian, avian and mammalian predators (Kotzman, 1990). Therefore, it would benefit the spider to have defensive toxins that target these animals in addition to their invertebrate-specific predatory compounds (Osteen et al., 2016). 93

The entire venom profiling results from LC/MS did not separate the species in the phylogenetic groups found exclusively by molecular data. That is probably due to individual variation within groups (Boevé, Kuhn-Nentwig, Keller, & Nentwig, 1995), but it also could be that some compounds of venom do not change deeply, being masked by the mass reconstruction software. Transcriptome sequencing of venom glands allied with proteomics of the venom can uncover small changes and provide a more complete result. Ontogenetic variation was found in tarantulas of the same population. Spider size might, to some extent at least, determine the maximum-sized prey a tarantula can subjugate, which could lead to different micro-niches according to spider size. These different predator–prey interactions might impose different selection pressures that lead to variations in venom composition. Shifts in venom composition due to a change in diet have been observed in snakes where prey changes from one taxonomic class to another as the snakes mature (Andrade & Abe, 1999; Gibbs et al., 2011; Jackson et al., 2016). The change in venom to target larger prey (i.e., small vertebrates in this case), might enable spiders to exploit more energy-rich resources as a trophic strategy. Intraspecific and interspecific venom variation can demonstrate how fast evolving and adaptive venom can be. The presence of unique peptides within each phylogenetic group is evidence that the specificity of venom could be used as identification method in the future. Such method is less invasive to the spider, and cheaper and faster than current alternative methods. However, it is still far from settled, as the fully understanding of venom variation on spiders is still far from being accomplished, more research on variation, peptide sequencing and inclusion of more taxonomic groups are needed. In addition, the presence of unique peptides within potential species and within each maturity stage present extra evidences that a single species can highly contribute to amplify our chemical library, which can increase our chances to make new pharmacological discoveries.

94

References

95

AFD. (2009). Australian Faunal Directory. Retrieved March 3, 2018, from https://biodiversity.org.au/afd/home Alacs, E., & Georges, A. (2008). Wildlife across our borders: a review of the illegal trade in Australia. Australian Journal of Forensic Sciences, 40(2), 147–160. https://doi.org/10.1080/00450610802491382 Alfaro, M. E. (2003). Bayes or Bootstrap? A Simulation Study Comparing the Performance of Bayesian Markov Chain Monte Carlo Sampling and Bootstrapping in Assessing Phylogenetic Confidence. Molecular Biology and Evolution, 20(2), 255–266. https://doi.org/10.1093/molbev/msg028 Ali, S. a., Yang, D. C., Jackson, T. N. W., Undheim, E. a B., Koludarov, I., Wood, K., … Fry, B. G. (2013). Venom proteomic characterization and relative antivenom neutralization of two medically important Pakistani elapid snakes (Bungarus sindanus and Naja naja). Journal of Proteomics, 89, 15–23. https://doi.org/10.1016/j.jprot.2013.05.015 Anderson, J. F. (1970). Metabolic rates of spiders. Comparative Biochemistry and Physiology, 33(1), 51–72. https://doi.org/10.1016/0010-406X(70)90483-4 Andrade, D. V, & Abe, A. S. (1999). Relationship of venom ontogeny and diet in Bothrops. Herpetologica, 55(2), 200–204. Retrieved from http://www.jstor.org/stable/3893080 Arabi, J., Cruaud, C., Couloux, A., & Hassanin, A. (2010). Studying sources of incongruence in arthropod molecular phylogenies: Sea spiders (Pycnogonida) as a case study. Comptes Rendus - Biologies, 333(5), 438–453. https://doi.org/10.1016/j.crvi.2010.01.018 Arnedo, M. A., & Ferrández, M.-A. (2007). Mitochondrial markers reveal deep population subdivision in the European protected spider Macrothele calpeiana (Walckenaer, 1805) (Araneae, Hexathelidae). Conservation Genetics, 8(5), 1147–1162. https://doi.org/10.1007/s10592-006-9270-2 Atkinson, R. K. (1993). A comparison of the toxicity and the venoms of twelve common Australian spider species on rodent vital organ systems. Comparative Biochemistry and Physiology - C Pharmacology Toxicology and Endocrinology, 106(3), 639–642. https://doi.org/10.1016/0742-8413(93)90220-F Australian Government. (2018). Australia’s 15 National Biodiversity Hotspots. Retrieved August 20, 2005, from http://www.environment.gov.au/biodiversity/conservation/ hotspots/national-biodiversity-hotspots Avise, J. C. (2000). Phylogeography: the history and formation of species. Harvard 96

university press. Ayoub, N. A., Garb, J. E., Hedin, M., & Hayashi, C. Y. (2007). Utility of the nuclear protein- coding gene, elongation factor-1 gamma (EF-1γ), for spider systematics, emphasizing family level relationships of tarantulas and their kin (Araneae: Mygalomorphae). Molecular Phylogenetics and Evolution, 42(2), 394–409. https://doi.org/10.1016/j.ympev.2006.07.018 Baerg, W. J. (1963). Tarantula Life History Records. Journal of the New York Entomological Society, 71(4), 233–238. Baerg, W. J. (2018). Tarantula Life History Records Author (s): W . J . Baerg Source: Journal of the New York Entomological Society, Vol. 71, No. 4 (Dec., 1963), pp. Published by: New York Entomological Society Stable URL: http://www.jstor.org/stable/25005880 Dec, 71(4), 233–238. Barr, N. B., Cook, A., Elder, P., Molongoski, J., Prasher, D., & Robinson, D. G. (2009). Application of a DNA barcode using the 16S rRNA gene to diagnose pest Arion species in the USA. Journal of Molluscan Studies, 75(2), 187–191. https://doi.org/10.1093/mollus/eyn047 Barrett, R. D. ., & Hebert, P. D. . (2005). Identifying spiders through DNA barcodes. Canadian Journal of Zoology, 83(3), 481–491. https://doi.org/10.1139/z05-024 Bell, K. L., Moritz, C., Moussalli, A., & Yeates, D. K. (2007). Comparative phylogeography and speciation of dung beetles from the Australian Wet Tropics rainforest. Molecular Ecology, 16(23), 4984–4998. https://doi.org/10.1111/j.1365-294X.2007.03533.x Bell, K. L., Yeates, D. K., Moritz, C., & Monteith, G. B. (2004). Molecular phylogeny and biogeography of the dung beetle genus Temnoplectron Westwood (Scarabaeidae: Scarabaeinae) from Australia’s wet tropics. Molecular Phylogenetics and Evolution, 31(2), 741–753. https://doi.org/10.1016/j.ympev.2003.09.010 Binford, G. J. (2000). Diversification of spider venoms: patterns and correlates of variation within individuals and across populations and phylogenies. PhD Thesis, The University of Arizona. Retrieved from http://search.proquest.com/docview/304587724?accountid=14723. Bivand, R. S., Pebesma, E., & Gomez-Rubio, V. (2013). Applied spatial data analysis with {R}, Second edition. Springer, NY. Retrieved from http://www.asdar-book.org/ Boevé, J. L., Kuhn-Nentwig, L., Keller, S., & Nentwig, W. (1995). Quantity and quality of venom released by a spider (Cupiennius salei, Ctenidae). Toxicon, 33(10), 1347– 1357. https://doi.org/10.1016/0041-0101(95)00066-U Bond, J. E. (2004). Systematics of the Californian euctenizine spider genus Apomastus 97

(Araneae : Mygalomorphae : Cyrtaucheniidae): the relationship between molecular and morphological taxonomy. Invertebrate Systematics, 18, 361. https://doi.org/10.1071/IS04008 Bond, J. E., & Beamer, D. a. (2006). A morphometric analysis of mygalomorph spider carapace shape and its efficacy as a phylogenetic character (Araneae). Invertebrate Systematics, 20(1), 1. https://doi.org/10.1071/IS05041 Bond, J. E., & Hedin, M. (2006). A total evidence assessment of the phylogeny of North American euctenizine trapdoor spiders (Araneae, Mygalomorphae, Cyrtaucheniidae) using Bayesian inference. Molecular Phylogenetics and Evolution, 41(1), 70–85. https://doi.org/10.1016/j.ympev.2006.04.026 Bond, J. E., Beamer, D. A., Lamb, T., & Hedin, M. (2006). Combining genetic and geospatial analyses to infer population extinction in mygalomorph spiders endemic to the Los Angeles region. Animal Conservation, 9(2), 145–157. https://doi.org/10.1111/j.1469-1795.2006.00024.x Bond, J. E., Hedin, M. C., Ramirez, M. G., & Opell, B. D. (2001). Deep molecular divergence in the absence of morphological and ecological change in the Californian coastal dune endemic trapdoor spider Aptostichus simus. Molecular Ecology, 10(4), 899–910. https://doi.org/10.1046/j.1365-294X.2001.01233.x Bond, J. E., Hendrixson, B. E., Hamilton, C. A., & Hedin, M. (2012). A reconsideration of the classification of the Spider infraorder mygalomorphae (Arachnida: Araneae) based on three nuclear genes and morphology. PLoS ONE, 7(6). https://doi.org/10.1371/journal.pone.0038753 Bouckaert, R. (2015). bModelTest : Bayesian site model selection for nucleotide data. BioRxiv, 1–17. https://doi.org/http://dx.doi.org/10.1101/020792 Bouckaert, R., Heled, J., Kühnert, D., Vaughan, T., Wu, C.-H., Xie, D., … Drummond, A. J. (2014). BEAST 2: A Software Platform for Bayesian Evolutionary Analysis. PLoS Computational Biology, 10(4), e1003537. https://doi.org/10.1371/journal.pcbi.1003537 Bowman, D. M. J. S., Brown, G. K., Braby, M. F., Brown, J. R., Cook, L. G., Crisp, M. D., … Ladiges, P. Y. (2010). Biogeography of the Australian monsoon tropics. Journal of Biogeography, 37(2), 201–216. https://doi.org/10.1111/j.1365-2699.2009.02210.x Brewer, M. S., Cotoras, D. D., Croucher, P. J. P., & Gillespie, R. G. (2014). New sequencing technologies, the development of genomics tools, and their applications in evolutionary arachnology. Journal of Arachnology, 42(1), 1–15. https://doi.org/10.1636/M13-78.1 Bryant, L. M., & Krosch, M. N. (2016). Lines in the land: a review of evidence for eastern 98

Australia’s major biogeographical barriers to closed forest taxa. Biological Journal of the Linnean Society, 119(2), 238–264. https://doi.org/10.1111/bij.12821 Buckler, E. S., Ippolito, A., & Holtsford, T. P. (1997). The evolution of ribosomal DNA: divergent paralogues and phylogenetic implications. Genetics, 145(3), 821–832. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9055091 Burbidge, N. T. (1960). The phytogeography of the Australian region. Australian Journal of Botany, 8(2), 75. https://doi.org/10.1071/BT9600075 Byrne, M., Yeates, D. K., Joseph, L., Kearney, M., Bowler, J., Williams, M. A. J., … Wyrwoll, K.-H. (2008). Birth of a biome: insights into the assembly and maintenance of the Australian arid zone biota. Molecular Ecology, 17(20), 4398–4417. https://doi.org/10.1111/j.1365-294X.2008.03899.x Cabbiness, S. G., Gehrke, C. W., Kuo, K. C., Chan, T. K., Hall, J. E., Hudiburg, S. a, & Odell, G. V. (1980). Polyamines in some tarantula venoms. Toxicon : Official Journal of the International Society on Toxinology, 18(1977), 681–683. https://doi.org/10.1016/0041-0101(80)90099-9 Campbell, H., & Engelbrecht, I. (2018). The Baboon Spider Atlas - using citizen science and the ‘fear factor’ to map baboon spider (Araneae: Theraphosidae) diversity and distributions in Southern Africa. Insect Conservation and Diversity, 11(2), 143–151. https://doi.org/10.1111/icad.12278 Casewell, N. R., Wüster, W., Vonk, F. J., Harrison, R. A., & Fry, B. G. (2013). Complex cocktails: The evolutionary novelty of venoms. Trends in Ecology and Evolution, 28(4), 219–229. https://doi.org/10.1016/j.tree.2012.10.020 Castresana, J. (2000). Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Molecular Biology and Evolution, 17, 540–552. Chan, T. K., Geren, C. R., Howell, D. E., & Odell, G. V. (1975). Adenosine triphosphate in tarantula spider venoms and its synergistic effect with the venom toxin. Toxicon, 13, 61–66. https://doi.org/10.1016/0041-0101(75)90159-2 Chapple, D. G., Hoskin, C. J., Chapple, S. N. J., & Thompson, M. B. (2011). Phylogeographic divergence in the widespread delicate skink (Lampropholis delicata) corresponds to dry habitat barriers in eastern Australia. BMC Evolutionary Biology, 11(1), 191. https://doi.org/10.1186/1471-2148-11-191 Chen, J., Deng, M., He, Q., Meng, E., Jiang, L., Liao, Z., … Liang, S. (2008). Molecular diversity and evolution of cystine knot toxins of the tarantula Chilobrachys jingzhao. Cellular and Molecular Life Sciences, 65(15), 2431–2444. https://doi.org/10.1007/s00018-008-8135-x 99

Chen, J., Zhao, L., Jiang, L., Meng, E., Zhang, Y., Xiong, X., & Liang, S. (2008). Transcriptome analysis revealed novel possible venom components and cellular processes of the tarantula Chilobrachys jingzhao venom gland. Toxicon, 52(7), 794– 806. https://doi.org/10.1016/j.toxicon.2008.08.003 Chippaux, J.-P., Williams, V., & White, J. (1991). Snake venom variability: methods of study, results and interpretation. Toxicon, 29(11), 1279–1303. https://doi.org/10.1016/0041-0101(91)90116-9 Chow, C. Y., Cristofori-Armstrong, B., Undheim, E. A. B., King, G. F., & Rash, L. D. (2015). Three peptide modulators of the human voltage-gated sodium channel 1.7, an important analgesic target, from the venom of an Australian tarantula. Toxins, 7, 2514–2533. Clement, M., Snell, Q., Walker, P., Posada, D., & Crandall, K. (2002). TCS: Estimating gene genealogies. Parallel and Distributed Processing Symposium, International Proceedings, 2, 184. Coddington, J. A. (2005). Phylogeny and classification of spiders. In Spiders of North America: an identification manual (pp. 18–24). https://doi.org/10.1006/bojl.2000.0418 Coddington, J. A., & Levi, H. W. (1991). Systematics and Evolution of Spiders (Araneae). Annual Review of Ecology and Systematics, 22(1), 565–592. https://doi.org/10.1146/annurev.es.22.110191.003025 Cooper, S. J. B., Harvey, M. S., Saint, K. M., & Main, B. Y. (2011). Deep phylogeographic structuring of populations of the trapdoor spider Moggridgea tingle (Migidae) from southwestern Australia: evidence for long-term refugia within refugia. Molecular Ecology, 20(15), 3219–3236. https://doi.org/10.1111/j.1365-294X.2011.05160.x Coyle, F. A., & Icenogle, W. R. (1994). Natural history of the Californian Trapdoor Spider genus Aliatypus (Araneae, Antrodiaetidae). The Journal of Arachnology, 22(3), 225– 255. Retrieved from http://www.jstor.org/stable/3705426 Crayn, D. M., Costion, C., & Harrington, M. G. (2015). The Sahul-Sunda floristic exchange: dated molecular phylogenies document Cenozoic intercontinental dispersal dynamics. Journal of Biogeography, 42(1), 11–24. https://doi.org/10.1111/jbi.12405 Criscuolo, F., Font-Sala, C., Bouillaud, F., Poulin, N., & Trabalon, M. (2010). Increased ROS production: A component of the longevity equation in the male mygalomorph, Brachypelma albopilosa. PLoS ONE, 5(10), 1–6. https://doi.org/10.1371/journal.pone.0013104 Crisp, M., Cook, L., & Steane, D. (2004). Radiation of the Australian flora: what can comparisons of molecular phylogenies across multiple taxa tell us about the evolution 100

of diversity in present-day communities? Philosophical Transactions of the Royal Society B: Biological Sciences, 359(1450), 1551–1571. https://doi.org/10.1098/rstb.2004.1528 Danforth, B. N., Lin, C.-P., & Fang, J. (2005). How do insect nuclear ribosomal genes compare to protein-coding genes in phylogenetic utility and nucleotide substitution patterns? Systematic Entomology, 30(4), 549–562. https://doi.org/10.1111/j.1365- 3113.2005.00305.x Davidson, E. H., Galau, G. A., Angerer, R. C., & Britten, R. J. (1975). Comparative aspects of DNA organization in Metazoa. Chromosoma, 51(3), 253–259. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/238803 Davies, M. E., & Edney, E. B. (1952). The Evaporation of Water from Spiders. Journal of Experimental Biology, 29(4), 571 LP-582. Retrieved from http://jeb.biologists.org/content/29/4/571.abstract De Andrade, R. M. G., De Oliveira, K. C., Giusti, A. L., Da Silva, W. D., & Tambourgi, D. V. (1999). Ontogenetic development of Loxosceles intermedia spider venom. Toxicon, 37(4), 627–632. https://doi.org/10.1016/S0041-0101(98)00200-1 de Oliveira, K. C., Gonçalves de Andrade, R. M., Giusti, A. L., da Silva, W. D., & Tambourgi, D. V. (1999). Sex-linked variation of Loxosceles intermedia spider venoms. Toxicon, 37(1), 217–221. https://doi.org/10.1016/S0041-0101(98)00130-5 De Queiroz, K. (2005). A unified species concept and its consequences for the future of taxonomy. Proceedings of the California Academy of Sciences, 56(June), 196–215. https://doi.org/10.1073/pnas.0502030102 De Queiroz, K. (2007). Species Concepts and Species Delimitation. Systematic Biology, 56(6), 879–886. https://doi.org/10.1080/10635150701701083 Dowton, M., & Austin, A. D. (1998). Phylogenetic relationships among the Microgastroid wasps (Hymenoptera: Braconidae): Combined Analysis of 16S and 28S rDNA genes and morphological data. Molecular Phylogenetics and Evolution, 10(3), 354–366. https://doi.org/10.1006/mpev.1998.0533 Diego-García, E., Peigneur, S., Waelkens, E., Debaveye, S., & Tytgat, J. (2010). Venom components from Citharischius crawshayi spider (Family Theraphosidae): Exploring transcriptome, venomics, and function. Cellular and Molecular Life Sciences, 67, 2799–2813. https://doi.org/10.1007/s00018-010-0359-x Edwards, R. D., Crisp, M. D., Cook, D. H., & Cook, L. G. (2017). Congruent biogeographical disjunctions at a continent-wide scale: Quantifying and clarifying the role of biogeographic barriers in the Australian tropics. PLoS ONE, 12(4), 1–17. 101

https://doi.org/10.1371/journal.pone.0174812 Edwards, S. V., Xi, Z., Janke, A., Faircloth, B. C., McCormack, J. E., Glenn, T. C., … Davis, C. C. (2016). Implementing and testing the multispecies coalescent model: A valuable paradigm for phylogenomics. Molecular Phylogenetics and Evolution, 94, 447–462. https://doi.org/10.1016/j.ympev.2015.10.027 EHP. (2018). Environment and Heritage Protection. Retrieved March 3, 2018, from https://www.ehp.qld.gov.au/ Escoubas, P., Célérier, M. L., & Nakajima, T. (1997). High-performance liquid chromatography matrix-assisted laser desorption/Ionization time-of-flight mass spectrometry peptide fingerprinting of tarantula venoms in the genus Brachypelma: Chemotaxonomic and biochemical applications. Rapid Communications in Mass Spectrometry, 11(17), 1891–1899. https://doi.org/10.1002/(SICI)1097- 0231(199711)11:17<1891::AID-RCM94>3.0.CO;2-X Escoubas, P., Corzo, G., Whiteley, B. J., Célérier, M.-L., & Nakajima, T. (2002). Matrix- assisted laser desorption/ionization time-of-flight mass spectrometry and high- performance liquid chromatography study of quantitative and qualitative variation in tarantula spider venoms. Rapid Communications in Mass Spectrometry, 16(5), 403– 413. https://doi.org/10.1002/rcm.595 Escoubas, P., Diochot, S., & Corzo, G. (2000). Structure and pharmacology of spider venoms neurotoxins. Biochimie, 82, 893–907. https://doi.org/10.1016/S0300- 9084(00)01166-4 Escoubas, P., & Rash, L. (2004). Tarantulas: Eight-legged pharmacists and combinatorial chemists. Toxicon, 43(5), 555–574. https://doi.org/10.1016/j.toxicon.2004.02.007 Escoubas, P., Sollod, B., & King, G. F. (2006). Venom landscapes: Mining the complexity of spider venoms via a combined cDNA and mass spectrometric approach. Toxicon, 47(6), 650–663. https://doi.org/10.1016/j.toxicon.2006.01.018 Evans, B. J. (2012). Coalescent-based analysis of demography: applications to biogeography on Sulawesi. In D. J. Gower, K. G. John-son, J. E. Richardson, B. R. Rosen, L. Rüber, & S. T. Williams (Eds.), Biotic Evolution and Environmental Change in Southeast Asia (pp. 270–289). Cambridge: Cambridge University Press. Ferretti, N., Pérez-Miles, F., & González, A. (2010). Mygalomorph spiders of the natural and historical reserve of Martín García Island, Río de la Plata River, Argentina. Zoological Studies, 49(4), 481–491. Foelix, R. (2011). Biology of Spiders (3rd ed.). New York: Oxford University Press. Folmer, O., Black, M., Hoeh, W., Lutz, R., & Vrijenhoek, R. (1994). DNA primers for 102

amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3(5), 294–299. https://doi.org/10.1371/journal.pone.0013102 Ford, J. (1986). Avian Hybridization and Allopatry in the Region of the Einasleigh Uplands and Burdekin-Lynd Divide, North-Eastern Queensland. Emu - Austral Ornithology, 86(2), 87–110. https://doi.org/10.1071/MU9860087 Framenau, V. W., Baehr, B. C., & Zborowski, P. (2014). A guide to the spiders of Australia. London: New Holland Publishers. Francis, R. M. (2017). Pophelper: an R package and web app to analyse and visualize population structure. Molecular Ecology Resources, 17(1), 27–32. https://doi.org/10.1111/1755-0998.12509 Fujisawa, T., & Barraclough, T. G. (2013). Delimiting species using single-locus data and the generalized mixed yule coalescent approach: A revised method and evaluation on simulated data sets. Systematic Biology, 62(5), 707–724. https://doi.org/10.1093/sysbio/syt033 Gentz, M. C., Jones, A., Clement, H., & King, G. F. (2009). Comparison of the peptidome and insecticidal activity of venom from a taxonomically diverse group of theraphosid spiders. Toxicon, 53(5), 496–502. https://doi.org/10.1016/j.toxicon.2009.01.025 Glez-Pena, D., Gomez-Blanco, D., Reboiro-Jato, M., Fdez-Riverola, F., & Posada, D. (2010). ALTER: program-oriented conversion of DNA and protein alignments. Nucleic Acids Research, 38, W14–W18. https://doi.org/10.1093/nar/gkq321 Gibbs, H. L., Sanz, L., Chiucchi, J. E., Farrell, T. M., & Calvete, J. J. (2011). Proteomic analysis of ontogenetic and diet-related changes in venom composition of juvenile and adult Dusky Pigmy rattlesnakes (Sistrurus miliarius barbouri). Journal of Proteomics, 74(10), 2169–2179. https://doi.org/10.1016/j.jprot.2011.06.013 Gillespie, R. G., Croom, H. B., & Palumbi, S. R. (1994). Multiple origins of a spider radiation in Hawaii. Proceedings of the National Academy of Sciences, 91(6), 2290– 2294. https://doi.org/10.1073/pnas.91.6.2290 González-Orozco, C. E., Ebach, M. C., Laffan, S., Thornhill, A. H., Knerr, N. J., Schmidt- Lebuhn, A. N., … Miller, J. T. (2014). Quantifying phytogeographical regions of Australia using geospatial turnover in species composition. PLoS ONE, 9(3). https://doi.org/10.1371/journal.pone.0092558 Gray, M. R. (2010). A revision of the Australian funnel-web spiders (Hexathelidae: Atracinae). Records of the Australian Museum, 62(2–3), 285–392. https://doi.org/10.3853/j.0067-1975.62.2010.1556 103

Guadanucci, J. P. L. (2014). Theraphosidae phylogeny: relationships of the “Ischnocolinae” genera (Araneae, Mygalomorphae). Zoologica Scripta, 43(5), 508– 518. https://doi.org/10.1111/zsc.12065 Guette, C., Legros, C., Tournois, G., Goyffon, M., & Célérier, M. L. (2006). Peptide profiling by matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry of the Lasiodora parahybana tarantula venom gland. Toxicon, 47(6), 640–649. https://doi.org/10.1016/j.toxicon.2006.01.017 Hall, R. (2012). Sundaland and Wallacea: geology, plate tectonics and palaeogeography. In D. J. Gower, K. G. John-son, J. E. Richardson, B. R. Rosen, L. Rüber, & S. T. Williams (Eds.), Biotic Evolution and Environmental Change in Southeast Asia (pp. 32–78). Cambridge: Cambridge University Press. Hamilton, C. A. (2009). Determining the phylogeographic dynamics of the edge relationships between Aphonopelma hentzi (Girard) and its neighbors along the Colorado river basin (Araneae, Mygalomorphae, Theraphosidae). The University of Texas at Arlington. Hamilton, C. A., Formanowicz, D. R., & Bond, J. E. (2011). Species delimitation and phylogeography of Aphonopelma hentzi (Araneae, Mygalomorphae, Theraphosidae): Cryptic diversity in north american tarantulas. PLoS ONE, 6(10), 12–16. https://doi.org/10.1371/journal.pone.0026207 Hamilton, C. A., Hendrixson, B. E., & Bond, J. E. (2016). Taxonomic revision of the tarantula genus Aphonopelma Pocock, 1901 (Araneae, Mygalomorphae, Theraphosidae) within the United States. ZooKeys, 560, 1–340. https://doi.org/10.3897/zookeys.560.6264 Hamilton, C. A., Hendrixson, B. E., Brewer, M. S., & Bond, J. E. (2014). An evaluation of sampling effects on multiple DNA barcoding methods leads to an integrative approach for delimiting species: A case study of the North American tarantula genus Aphonopelma (Araneae, Mygalomorphae, Theraphosidae). Molecular Phylogenetics and Evolution, 71(1), 79–93. https://doi.org/10.1016/j.ympev.2013.11.007 Hammer, Ø., Harper, D. A. T., & Ryan, P. . (2001). PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4(1). Harrison, S. E., Harvey, M. S., Cooper, S. J. B., Austin, A. D., Rix, M. G., Roderick, G., … Gray, M. (2017). Across the Indian Ocean: A remarkable example of trans-oceanic dispersal in an austral mygalomorph spider. Plos One, 12(8), e0180139. https://doi.org/10.1371/journal.pone.0180139 Harvey, M. S., Main, B. Y., Rix, M. G., & Cooper, S. J. B. (2015). Refugia within refugia: In 104

situ speciation and conservation of threatened Bertmainius (Araneae:Migidae), a new genus of relictual trapdoor spiders endemic to the mesic zone of south-Western Australia. Invertebrate Systematics, 29(6), 511–553. https://doi.org/10.1071/IS15024 Hausdorf, B. (2011). Progress toward a general species concept. Evolution, 65(4), 923– 931. https://doi.org/10.1111/j.1558-5646.2011.01231.x Hawkeswood, T. J. (2003). Spiders of Australia: An introduction to their Classification, Biology and Distribution. Bulgaria: Pensoft Publishers. Hebert, P. D. N., Cywinska, A., Ball, S. L., & DeWaard, J. R. (2003). Biological identifications through DNA barcodes. Proceedings of the Royal Society B: Biological Sciences, 270, 313–321. https://doi.org/10.1098/rspb.2002.2218 Hebert, P. D. N., Ratnasingham, S., & DeWaard, J. R. (2003). Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270, S96– S99. https://doi.org/10.1098/rsbl.2003.0025 Hedin, M., & Bond, J. E. (2006). Molecular phylogenetics of the spider infraorder Mygalomorphae using nuclear rRNA genes (18S and 28S): Conflict and agreement with the current system of classification. Molecular Phylogenetics and Evolution, 41(2), 454–471. https://doi.org/10.1016/j.ympev.2006.05.017 Hedin, M., Carlson, D., & Coyle, F. (2015). Sky island diversification meets the multispecies coalescent - Divergence in the spruce-fir moss spider (Microhexura montivaga, Araneae, Mygalomorphae) on the highest peaks of southern Appalachia. Molecular Ecology, 24(13), 3467–3484. https://doi.org/10.1111/mec.13248 Hedin, M., Derkarabetian, S., Ramírez, M. J., Vink, C., & Bond, J. E. (2018). Phylogenomic reclassification of the world’s most venomous spiders (Mygalomorphae, Atracinae), with implications for venom evolution. Scientific Reports, 8(1), 1636. https://doi.org/10.1038/s41598-018-19946-2 Heled, J., & Drummond, A. J. (2010). Bayesian Inference of Species Trees from Multilocus Data. Molecular Biology and Evolution, 27(3), 570–580. https://doi.org/10.1093/molbev/msp274 Hendrixson, B. E., & Bond, J. E. (2004). A new species of Stasimopus from the Eastern Cape Province of South Africa (Araneae, Mygalomorphae, Ctenizidae), with notes on its natural history. Zootaxa, 619, 1–14. Retrieved from http://www.mapress.com/zootaxa/2004f/z00619f.pdf Hendrixson, B. E., DeRussy, B. M., Hamilton, C. A., & Bond, J. E. (2013). An exploration of species boundaries in turret-building tarantulas of the Mojave Desert (Araneae, 105

Mygalomorphae, Theraphosidae, Aphonopelma). Molecular Phylogenetics and Evolution, 66(1), 327–340. https://doi.org/10.1016/j.ympev.2012.10.004 Hendrixson, B. E., Guice, A. V., & Bond, J. E. (2015). Integrative species delimitation and conservation of tarantulas (Araneae, Mygalomorphae, Theraphosidae) from a North American biodiversity hotspot. Insect Conservation and Diversity, 8(2), 120–131. https://doi.org/10.1111/icad.12089 Hennig, W. (1966). Phylogenetic Systematics. University of Illinois Press. Henning, F., Weberz, M., Hinkel, H., & Schottler, B. (2002). Defensive hissing display of African baboon spiders (Araneae: Theraphosidae). Bioacoustics, 13(1), 99. Herzig, V. (2010). Ontogenesis, gender, and molting influence the venom yield in the spider Coremiocnemis tropix (Araneae, Theraphosidae). Journal of Venom Research, 1, 76–83. Herzig, V., & Hodgson, W. C. (2008). Neurotoxic and insecticidal properties of venom from the Australian theraphosid spider Selenotholus foelschei. NeuroToxicology, 29(3), 471–475. https://doi.org/10.1016/j.neuro.2008.03.002 Herzig, V., & Hodgson, W. C. (2009). Intersexual variations in the pharmacological properties of Coremiocnemis tropix (Araneae, Theraphosidae) spider venom. Toxicon, 53(2), 196–205. https://doi.org/10.1016/j.toxicon.2008.11.002 Herzig, V., Ward, R. J., & Dos Santos, W. F. (2004). Ontogenetic changes in Phoneutria nigriventer (Araneae, Ctenidae) spider venom. Toxicon, 44(6), 635–640. https://doi.org/10.1016/j.toxicon.2004.07.020 Herzig, V., Wood, D. L. A., Newell, F., Chaumeil, P. A., Kaas, Q., Binford, G. J., … King, G. F. (2011). ArachnoServer 2.0, an updated online resource for spider toxin sequences and structures. Nucleic Acids Research, 39 (SUPPL. 1), 653–657. https://doi.org/10.1093/nar/gkq1058 Hirst, A. S. (1907). On two spiders of the genus Selenocosmia. Annals and Magazine of Natural History, 19, 522–524. Hogg, H. R. (1901). On Australian and New Zealand Spidera of the Suborder Mygalomorphae. Proceedings of the Zoological Society of London, 71(1), 218–284. https://doi.org/10.1111/j.1469-7998.1901.tb08176.x Hogg, H. R. (1902). On some additions to the Australian spiders of the suborder Mygalomorphae. Proceedings of the Zoological Society of London, 2, 121–142. Hoskin, C. J., & Couper, P. J. (2014). Two new skinks (Scincidae: Glaphyromorphus) from rainforest habitats in north-eastern Australia. Zootaxa, 3869(1), 1–16. https://doi.org/10.11646/zootaxa.3869.1.1 106

Huber, K. C., Haider, T. S., Mueller, M. W., Huber, B. A., Schweyen, R. J., & Barth, F. G. (1993). DNA sequence data indicates the polyphyly of the family Ctenidae (Araneae). Journal of Arachnology, 21(3), 194–201. https://doi.org/10.2307/3705482 Hugall, A. F., Moritz, C., Moussalli, A., & Stanisic, J. (2002). Reconciling paleodistribution models and comparative phylogeography in the Wet Tropics rainforest land snail Gnarosophia bellendenkerensis (Brazier 1875). Proceedings of the National Academy of Sciences, 99(9), 6112–6117. https://doi.org/10.1073/pnas.092538699 Hugall, A. F., & Stanisic, J. (2011). Beyond the prolegomenon: A molecular phylogeny of the Australian camaenid land snail radiation. Zoological Journal of the Linnean Society, 161(3), 531–572. https://doi.org/10.1111/j.1096-3642.2010.00644.x Hurd, L. E., & Fagan, W. F. (1992). Cursorial spiders and succession: age or habitat structure? Oecologia, 92(2), 215–221. https://doi.org/10.1007/BF00317367 Hutchison, C. A. (2007). DNA sequencing: bench to bedside and beyond. Nucleic Acids Research, 35(18), 6227–6237. https://doi.org/10.1093/nar/gkm688 Hwang, U.-W., & Kim, W. (1999). General properties and phylogenetic utilities of nuclear ribosomal DNA and mitochondrial DNA commonly used in molecular systematics. The Korean Journal of Parasitology, 37(4), 215. https://doi.org/10.3347/kjp.1999.37.4.215 Ibler, B., Michalik, P., & Fischer, K. (2013). Factors affecting lifespan in bird-eating spiders (Arachnida: Mygalomorphae, Theraphosidae) - A multi-species approach. Zoologischer Anzeiger, 253(2), 126–136. https://doi.org/10.1016/j.jcz.2013.09.004 Jackson, T. N. W., Koludarov, I., Ali, S. A., Dobson, J., Zdenek, C. N., Dashevsky, D., … Fry, B. G. (2016). Rapid radiations and the race to redundancy: An investigation of the evolution of Australian elapid snake venoms. Toxins, 8(11), 1–24. https://doi.org/10.3390/toxins8110309 James, C. H., & Moritz, C. (2000). Intraspecific phylogeography in the sedge frog Litoria fallax (Hylidae) indicates pre-Pleistocene vicariance of an open forest species from eastern Australia. Molecular Ecology, 9(3), 349–358. https://doi.org/10.1046/j.1365- 294x.2000.00885.x Jiang, L., Zhang, D., Zhang, Y., Peng, L., Chen, J., & Liang, S. (2010). Venomics of the spider Ornithoctonus huwena based on transcriptomic versus proteomic analysis. Comparative Biochemistry and Physiology - Part D: Genomics and Proteomics, 5(2), 81–88. https://doi.org/10.1016/j.cbd.2010.01.001 Joseph, L., & Moritz, C. (1994). Mitochondrial-Dna Phylogeography of Birds in Eastern Australian Rain-Forests - First Fragments. Australian Journal of Zoology, 42(3), 385. https://doi.org/10.1071/ZO9940385 107

Kapli, P., Lutteropp, S., Zhang, J., Kobert, K., Pavlidis, P., Stamatakis, A., & Flouri, T. (2017). Multi-rate Poisson Tree Processes for single-locus species delimitation under Maximum Likelihood and Markov Chain Monte Carlo. Bioinformatics. https://doi.org/10.1093/bioinformatics/btx025 King, G. F. (2004). The wonderful world of spiders: Preface to the special Toxicon issue on spider venoms. Toxicon, 43, 471–475. https://doi.org/10.1016/j.toxicon.2004.02.001 King, G. F. (2011). Venoms as a platform for human drugs: translating toxins into therapeutics. Expert Opinion on Biological Therapy, 11(11), 1469. King, G. F., & Hardy, M. C. (2013). Spider-venom peptides: structure, pharmacology, and potential for control of insect pests. Annual Review of Entomology, 58(1), 475–96. https://doi.org/10.1146/annurev-ento-120811-153650 Kirchner, W. H., & Röschard, J. (1999). Hissing in bumblebees: An interspecific defence signal. Insectes Sociaux, 46(3), 239–243. https://doi.org/10.1007/s000400050140 Koch, L. (1874). Die Arachniden Australiens. Nürnberg, 1, 473–576. Kotzman, M. (1990). Annual activity patterns of the Australian tarantula Selenocosmia stirlingi (Araneae, Theraphosidae) in an arid area, 18(2), 123–130. Kumar, S., Stecher, G., & Tamura, K. (2016). MEGA7: Molecular Evolutionary Genetics Ananlysis version 7.0 for bigger dataset. Molecular Biology and Evolution, 33, 1870– 1874. Lampe, R. A., Defeo, P. A., Davison, M. D., Young, J., Herman, J. L., Spreen, R. C., … Keith, R. A. (1993). Isolation and pharmacological characterization of omega- grammotoxin SIA, a novel peptide inhibitor of neuronal voltage-sensitive calcium channel responses. Molecular Pharmacology, 44(2), 451–460. Retrieved from http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citat ion&list_uids=8394998 Lanfear, R., Frandsen, P. B., Wright, A. M., Senfeld, T., & Calcott, B. (2016). PartitionFinder 2: New Methods for Selecting Partitioned Models of Evolution for Molecular and Morphological Phylogenetic Analyses. Molecular Biology and Evolution, 11(11), msw260. https://doi.org/10.1093/molbev/msw260 Lemmon, A. R., & Moriarty, E. C. (2004). The importance of proper model assumption in bayesian phylogenetics. Systematic Biology, 53(2), 265–277. https://doi.org/10.1080/10635150490423520 Lenth, R. V. (2016). Least-Squares Means: The R Package lsmeans. Journal of Statistical Software, 69(1). https://doi.org/10.18637/jss.v069.i01 Liao, Z., Cao, J., Li, S., Yan, X., Hu, W., He, Q., … Liang, S. (2007). Proteomic and 108

peptidomic analysis of the venom from Chinese tarantula Chilobrachys jingzhao. Proteomics, 7(11), 1892–1907. https://doi.org/10.1002/pmic.200600785 Lin, C.-P., & Danforth, B. N. (2004). How do insect nuclear and mitochondrial gene substitution patterns differ? Insights from Bayesian analyses of combined datasets. Molecular Phylogenetics and Evolution, 30(3), 686–702. https://doi.org/10.1016/S1055-7903(03)00241-0 Lin, Y. P., Edwards, R. D., Kondo, T., Semple, T. L., & Cook, G. (2017). Species delimitation in asexual insects of economic importance : The case of black scale (Parasaissetia nigra), a cosmopolitan parthenogenetic pest scale insect, PLoS ONE, 12 (5), 1–22. https://doi.org/10.1371/journal.pone.0175889 Lin, Y. P., Kondo, T., Gullan, P., & Cook, L. G. (2013). Delimiting genera of scale insects: Molecular and morphological evidence for synonymising Taiwansaissetia Tao, Wong and Chang with Coccus Linnaeus (Hemiptera: Coccoidea: Coccidae). Systematic Entomology, 38(2), 249–264. https://doi.org/10.1111/j.1365-3113.2012.00664.x Longhorn, S. J., Nicholas, M., Chuter, J., & Vogler, A. P. (2007). the Utility of Molecular Markers From Non-Lethal Dna Samples of the Cites II Protected “Tarantula” Brachypelma vagans (Araneae, Theraphosidae). Journal of Arachnology, 35(2), 278– 292. https://doi.org/10.1636/S05-62.1 Machkour‐M’Rabet, S., Hénaut, Y., Dor, A., Pérez-Lachaud, G., Pélissier, C., Gers, C., & Legal, L. (2009). ISSR (Inter Simple Sequence Repeats) as molecular markers to study genetic diversity in tarantulas (Araneae, Mygalomorphae). Journal of Arachnology, 37(1), 10–14. https://doi.org/10.1636/A08-27.1 Main, B. Y. (1981). Eco-evolutionary radiation of mygalomorph spiders in Australia. Biogeography and Ecology of New Guinea, 2, 853–872. Main, B. Y. (1982). Some zoogeographic considerations of families of spiders occurring in New Guinea. Monographiae Biologicae, 42(2), 583–602. Main, B. Y. (1984). Spiders. Sydney: William Collins. Main, B. Y. (1985). Mygalomorphae. Zoological Catalogue of Australia, 3, 1–48. Main, B. Y. (1987). Ecological disturbance and conservation of spiders: implications for biogeographic relics in southwestern Australia. In J. Majer (Ed.), The role of invertebrates in conservation and biological surveys. (pp. 89–98). Perth: Western Australian Department of Conservation and Land Management Report. Mason, L. D., Tomlinson, S., Withers, P. C., & Main, B. Y. (2013). Thermal and hygric physiology of Australian burrowing mygalomorph spiders (Aganippe spp.). Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology, 109

183(1), 71–82. https://doi.org/10.1007/s00360-012-0681-8 Maxam, A., & Gilbert, W. (1977). A new method for sequencing DNA. Proceedings of the Academy of Natural Sciences of Philadelphia, 74, 560–564. Mayden, R. L. (1997). A Hierarchy of Species Concepts: The Denouement in the Saga of the Species Problem. In M. F. Claridge, H. A. Dawah, & M. R. Wilson (Eds.), Species: The units of diversity, (pp. 381–423). Chapman & Hall. Mendoza, J., & Francke, O. (2017). Systematic revision of Brachypelma red-kneed tarantulas (Araneae: Theraphosidae), and the use of DNA barcodes to assist in the identification and conservation of CITES-listed species. Invertebrate Systematics, 31, 157–179. https://doi.org/10.1071/IS16023 Minch, E. W. (1978). Daily activity patterns in the tarantula Aphonopelma chalcodes Chamberlin. Bulletin of the British Arachnological Society, 4(5), 231–237. Montes de Oca, L., D’Elía, G., & Pérez-Miles, F. (2016). An integrative approach for species delimitation in the spider genus Grammostola (Theraphosidae, Mygalomorphae). Zoologica Scripta, 45(3), 322–333. https://doi.org/10.1111/zsc.12152 Mourão, C. B. F., Oliveira, F. N., E Carvalho, A. C., Arenas, C. J., Duque, H. M., Gonçalves, J. C., … Schwartz, E. F. (2013). Venomic and pharmacological activity of Acanthoscurria paulensis (Theraphosidae) spider venom. Toxicon, 61(1), 129–138. https://doi.org/10.1016/j.toxicon.2012.11.008 Mourão, C. B. F., Silva, L. P., Block Jr, C., & Schwartz, E. F. N. (2007). Peptide profiling by matrix-assisted laser desorption-ionization time-of-flight mass spectrometry of the Acanthoscurria atrox tarantula venom: Evidences for sex-linked venom variation. In IX Pan-American Section Congress of the IST (p. 136). Juriquilla. Nagaraju, S., Mahadeswaraswamy, Y. H., Girish, K. S., & Kemparaju, K. (2006). Venom from spiders of the genus Hippasa: Biochemical and pharmacological studies. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology, 144, 1–9. https://doi.org/10.1016/j.cbpc.2006.04.016 Nelsen, D. R. (2013). Ontogeny of venom use and venom composition in the western widow spider Latrodctus hesperus. Loma Linda University. Retrieved from http://search.proquest.com/docview/1417778029?accountid=14723. Nentwig, W., Friedel, T., & Manhart, C. (1992). Comparative investigations on the effect of the venoms of 18 spider species onto the cockroach Blatta orientalis (Blattodea). Zoologische Jahrbücher : Abteilung Für Allgemeine Zoologie Und Physiologie Der Tiere. 110

Nicolas, V., Schaeffer, B., Missoup, A. D., Kennis, J., Colyn, M., Denys, C., … Laredo, C. (2012). Assessment of three mitochondrial genes (16S, Cytb, CO1) for identifying species in the Praomyini tribe (Rodentia: Muridae). PLoS ONE, 7(5), 1–11. https://doi.org/10.1371/journal.pone.0036586 Nunn, S. C., West, R. C., & Von Wirth, V. (2016). A revision of the Selenocosmiinae tarantula genus Phlogiellus Pocock 1897 (Araneae: Theraphosidae), with description of four new species. International Journal of Zoology, 2016 (Pocock 1897). https://doi.org/10.1155/2016/9895234 Odell, G. V., Ownby, C. L., Cabblness, S. G., Hudlburg, S. A., & Herrero, M. V. (1979). Venom components of Brachypelma emilia, the mexican red-legged tarantula. 8th World Congress on Animal, Plant and Microbial Toxins, 7(1974), 600. https://doi.org/10.1016/0041-0101(85)90301-0 Opatova, V., Bond, J. E., & Arnedo, M. A. (2013). Ancient origins of the Mediterranean trap-door spiders of the family Ctenizidae (Araneae, Mygalomorphae). Molecular Phylogenetics and Evolution, 69(3), 1135–1145. https://doi.org/10.1016/j.ympev.2013.08.002 Ortiz, D., & Francke, O. F. (2016). Two DNA barcodes and morphology for multi-method species delimitation in Bonnetina tarantulas (Araneae: Theraphosidae). Molecular Phylogenetics and Evolution, 101, 176–193. https://doi.org/10.1016/j.ympev.2016.05.003 Ortiz, D., & Francke, O. F. (2017). Reconciling morphological and molecular systematics: revision of the Mexican endemic tarantulas of the genus Bonnetina (Araneae: Theraphosidae). Zoological Journal of the Linnean Society, 1–68. Osteen, J. D., Herzig, V., Gilchrist, J., Emrick, J. J., Zhang, C., Wang, X., … Julius, D. (2016). Selective spider toxins reveal a role for the NaV1.1 channel in mechanical pain. Nature, 534, 494–499. Ostrow, K. L., Mammoser, A., Suchyna, T., Sachs, F., Oswald, R., Kubo, S., … Gottlieb, P. A. (2003). cDNA sequence and in vitro folding of GsMTx4, a specific peptide inhibitor of mechanosensitive channels. Toxicon, 42(3), 263–274. https://doi.org/10.1016/S0041-0101(03)00141-7 Ownby, C. L., & Odell, G. V. (1983). Pathogenesis of skeletal muscle necrosis induced by tarantula venom. Experimental and Molecular Pathology, 38(3), 283–296. https://doi.org/10.1016/0041-0101(85)90120-5 Palagi, A., Koh, J. M. S., Leblanc, M., Wilson, D., Dutertre, S., King, G. F., … Escoubas, P. (2013). Unravelling the complex venom landscapes of lethal Australian funnel-web 111

spiders (Hexathelidae: Atracinae) using LC-MALDI-TOF mass spectrometry. Journal of Proteomics, 80, 292–310. https://doi.org/10.1016/j.jprot.2013.01.002 Pascarel, C., Cazorla, O., Le Guennec, J. Y., Orchard, C. H., & White, E. (1997). The effect of the venom of a Chilean tarantula, Phrixotrichus spatulatus, on isolated guinea pig ventricular myocytes. Toxicology and Applied Pharmacology, 147, 363–371. https://doi.org/S0041-008X(97)98291-0 [pii]\r10.1006/taap.1997.8291 Pebesma, E. J., & Bivand, R. S. (2005). Classes and methods for spatial data in {R}. R News, 5(2), 9–13. Retrieved from https://cran.r-project.org/doc/Rnews/ Pérez-Miles, F. (2000). Iracema cabocla new genus and species of a theraphosid spider from Amazonic Brazil (Araneae, Theraphosinae). Journal of Arachnology, 28(2), 141– 148. https://doi.org/10.1636/0161-8202(2000)028[0141:ICNGAS]2.0.CO;2 Perret, B. A. (1977a). Proteolytic activity of tarantula venoms due to contamination with saliva. Toxicon : Official Journal of the International Society on Toxinology, 15(6), 505–510. https://doi.org/10.1016/0041-0101(77)90101-5 Perret, B. A. (1977b). Venom regeneration in tarantula spiders. I. Analysis of venom produced at different time intervals. Comparative Biochemistry and Physiology, 56(4A), 607–613. Pichardo, M. (1978). Phenetic vs. cladistic systematics: A response to Thoma. Journal of Human Evolution, 7(2), 151–158. https://doi.org/10.1016/S0047-2484(78)80007-4 Piser, T. M., Lampe, R. A., Keith, R. A., & Thayer, S. A. (1994). ω-Grammotoxin blocks action-potential-induced Ca 2+ influx and whole-cell Ca 2+ current in rat dorsal-root ganglion neurons. Europian Journal of Physiology, 426, 214–220. Pocock, R. I. (1895). On a new and natural grouping of some of the Oriental genera of Mygalomorphae, with descriptions of new genera and species. Annals and Magazine of Natural History, 6(15), 165–184. Potter, S., Cooper, S. J. B., Metcalfe, C. J., Taggart, D. A., & Eldridge, M. D. B. (2012). Phylogenetic relationships of rock-wallabies, Petrogale (Marsupialia: Macropodidae) and their biogeographic history within Australia. Molecular Phylogenetics and Evolution, 62(2), 640–652. https://doi.org/10.1016/j.ympev.2011.11.005 Powell, A. (2012). The discovery and naming of Port Darwin. Journal of Northern Territory History, 23, 77–93. Puillandre, N., Lambert, A., Brouillet, S., & Achaz, G. (2012). ABGD, Automatic Barcode Gap Discovery for primary species delimitation. Molecular Ecology, 21(8), 1864–1877. https://doi.org/10.1111/j.1365-294X.2011.05239.x Punzo, F., & Henderson, L. (1999). Aspects of the natural history and behavioural ecology 112

of the tarantula spider Aphonopelma hentzi (Girard, 1854) (Orthognatha, Theraphosidae). Bulletin of the British Arachnological Society, 11(4), 121–128. R Core Team. (2017). R: A language and environment for statistical computing. Vienna; Austria: R Foundation for Statistical Computing. Retrieved from https://www.r- project.org/ Rainbow, W. J. (1899). Contribution to a knowledge of Papuan Arachnida. Records of the Australian Museum, 3(5), 108–118. Raizer, J., Japyassú, H. F., Indicatti, R. P., & Brescovit, A. D. (2005). Comunidade de aranhas (Arachnida, Araneae) do pantanal norte (Mato Grosso, Brasil) e sua similaridade com a araneofauna amazônica. Biota Neotropica, 5(1a), 125–140. https://doi.org/10.1590/S1676-06032005000200011 Rambaut, A., & Drummond, A. J. (2010). TreeAnnotator. Retrieved from http://beast.bio.ed.ac.uk Rambaut, A., Suchard, M., Xie, D., & Drummond, A. (2014). Tracer v1.6. Retrieved from http://beast.bio.ed.ac.uk/Tracer Rash, L. D., & Hodgson, W. C. (2002). Pharmacology and biochemistry of spider venoms. Toxicon, 40, 225–254. https://doi.org/10.1016/S0041-0101(01)00199-4 Raven, R. J. (1982). Systematics of the Australian mygalomorph spider genus Ixamatus Simon (Diplurinae: Dipluridae: Chelicerata). Australian Journal of Zoology, 30, 1035– 1067. https://doi.org/10.1071/ZO9821035 Raven, R. J. (1985). The spider infraorder Mygalomorphae (Araneae): cladistics and systematics. Bulletin of the American Museum of Natural History, 182, 1–180. Retrieved from http://hdl.handle.net/2246/955 Raven, R. J. (2000). Taxonomica Araneae I: Barychelidae, Theraphosidae, Nemesiidae and Dipluridae (Araneae). Memoirs of the Queensland Museum, 45(2), 569–575. Raven, R. J. (2005). A new tarantula species from northern Australia (Araneae, Theraphosidae). Zootaxa, 1004(May), 15–28. Raven, R. J., & Hebron, W. (2018). A review of the Water spider family Pisauridae in Australia and New Caledonia with descriptions of four new genera and 23 new species. Memoirs of the Queensland Museum, 60, 233–381. Reichling, S. B., & West, R. C. (1996). A New Genus and Species of Theraphosid Spider from Belize (Araneae, Theraphosidae). The Journal of Arachnology, 24(3), 254–261. Ridpath, M. G., Williams, M. A. J., & Haynes, C. D. (1991). A torrid land. In Monsoonal Australia: landscape, ecology and man in the northern lowlands (pp. 207–221). Rotterdam: A.A. Balkema. 113

Rix, M. G., Bain, K., Main, B. Y., Raven, R. J., Austin, A. D., Cooper, S. J. B., & Harvey, M. S. (2017). Systematics of the spiny trapdoor spiders of the genus Cataxia (Mygalomorphae: Idiopidae) from south-western Australia: documenting a threatened fauna in a sky-island landscape. Journal of Arachnology, 45(3), 395–423. https://doi.org/10.1636/JoA-S-17-012.1 Rix, M. G., Cooper, S. J. B., Meusemann, K., Klopfstein, S., Harrison, S. E., Harvey, M. S., & Austin, A. D. (2017). Post-Eocene climate change across continental Australia and the diversification of Australasian spiny trapdoor spiders (Idiopidae: Arbanitinae). Molecular Phylogenetics and Evolution, 109, 302–320. https://doi.org/10.1016/j.ympev.2017.01.008 Rix, M. G., Main, B. Y., Raven, R. J., & Harvey, M. S. (2018). Systematics of the spiny trapdoor spiders of the genus Eucanippe (Mygalomorphae: Idiopidae: Aganippini) from south-western Australia: documenting a poorly-known lineage from Australia’s biodiversity hotspot. Journal of Arachnology, 46(1), 133–154. https://doi.org/10.1636/JoA-S-17-030.1 Rocha-e-Silva, T. a a, Sutti, R., & Hyslop, S. (2009). Milking and partial characterization of venom from the Brazilian spider Vitalius dubius (Theraphosidae). Toxicon, 53(1), 153– 161. https://doi.org/10.1016/j.toxicon.2008.10.026 Ronquist, F., & Huelsenbeck, J. P. (2003). MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics, 19(12), 1572–1574. https://doi.org/10.1093/bioinformatics/btg180 Roth, M. (1993). Investigations on lead in the soil invertebrates of a forest ecosystem. Pedobiologia, 37(5), 270–279. Rozas, J., Ferrer-Mata, A., Sánchez-DelBarrio, J. C., Guirao-Rico, S., Librado, P., Ramos- Onsins, S. E., & Sánchez-Gracia, A. (2017). DnaSP 6: DNA sequence polymorphism analysis of large data sets. Molecular Biology and Evolution, 34(12), 3299–3302. https://doi.org/10.1093/molbev/msx248 Saez, N. J., Senff, S., Jensen, J. E., Er, S. Y., Herzig, V., Rash, L. D., & King, G. F. (2010). Spider-venom peptides as therapeutics. Toxins, (2), 2851–2871. Sanderson, M. J., & Shaffer, H. B. (2002). Troubleshooting Molecular Phylogenetic Analyses. Annual Review of Ecology and Systematics, 33(1), 49–72. https://doi.org/10.1146/annurev.ecolsys.33.010802.150509 Sandoval, C. P. (1994). Plasticity in Web Design in the Spider Parawixia bistriata: A Response to Variable Prey Type. Functional Ecology, 8(6), 701. https://doi.org/10.2307/2390229 114

Sanger, F., Nicklen, S., & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences of the United States of America, 74(12), 5463–5467. Sannaningaiah, D., Subbaiah, G. K., & Kempaiah, K. (2014). Pharmacology of spider venom toxins. Toxin Reviews, 33(4), 206–220. https://doi.org/10.3109/15569543.2014.954134 Santana, R. C. (2015). Community structure and composition of litter spiders (Arachnida: Araneae) and influence of macro-climatic factors on Parque Ecológico Jatobá Centenário, Morrinhos, Goiás, Brazil. Journal of Threatened Taxa, 7(10), 7612–7624. https://doi.org/10.11609/JoTT.o4030.7612-24 Santana, R. C., Perez, D., Dobson, J., Panagides, N., Raven, R., Nouwens, A., … Fry, B. (2017). Venom profiling of a population of the theraphosid spider Phlogius crassipes reveals continuous ontogenetic changes from juveniles through adulthood. Toxins, 9(4), 116. https://doi.org/10.3390/toxins9040116 Santorum, J. M., Darriba, D., Taboada, G. L., & Posada, D. (2014). jmodeltest.org: selection of nucleotide substitution models on the cloud. Bioinformatics, 30(9), 1310– 1311. https://doi.org/10.1093/bioinformatics/btu032 Satler, J. D., Carstens, B. C., & Hedin, M. (2013). Multilocus species delimitation in a complex of morphologically conserved trapdoor spiders (Mygalomorphae, Antrodiaetidae, Aliatypus). Systematic Biology, 62(6), 805–823. https://doi.org/10.1093/sysbio/syt041 Savel-Niemann, a, & Roth, D. (1989). Biochemical analysis of tarantula venom (Eurypelma californicum). Die Naturwissenschaften, 76, 212–213. https://doi.org/10.1007/BF00627688 Schanbacher, F. L., Lee, C. K., Hall, J. E., Wilson, I. B., Howell, D. E., & Odell, G. V. (1973). Composition and properties of tarantula Dugesiella hentzi (Girard) venom. Toxicon, 11(1). https://doi.org/10.1016/0041-0101(73)90147-5 Schanbacher, F. L., Lee, C. K., Wilson, I. B., Howell, D. E., & Odell, G. V. (1973). Purification and characterization of tarantula, Dugesiella hentzi (Girard) venom hyaluronidase*. Comparative Biochemistry and Physiology, 44(B), 389–396. Schmidt, G. (1995). Gehören “Selenocosmia” crassipes (L. Koch, 1873) und “Selenocosmia” stirlingi Hogg, 1901 (Araneida: Theraphosidae: Selenocosmiinae) wirklich zu Selenocosmia Ausserer, 1871? Arachnologisches Magazin, 3(11), 1–12. Schmidt, G. (2002). Selenocosmia lanipes Ausserer, 1875 gehört zur Gattung Phlogius Simon, 1887 (Araneae: Theraphosidae: Selenocosmiinae). Arachnologisches 115

Magazin, 10, 1–5. Schmidt, G. (2003). Die Vogelspinnen: Eine weltweite Übersicht. Neue Brehm-Bücherei Hohenwarsleben. Schmidt, G. (2015). Warum Chilocosmia arndsti (Schmidt & Von Wirth, 1992), Chilocosmia barensteinerae Schmidt, Hettegger & Matthes, 2010, Chilocosmia peerboomi Schmidt, 1999, Selenopelma kovariki Schmidt & Krause, 1995 und Phlogius Simon, 1887 (Araneae: Theraphosidae: Selenocosmiinae). Tarantulas of the World, 143, 32– 38. Schmitz, A. (2013). Tracheae in Spiders: Respiratory Organs for Special Functions. In W. Nentwig (Ed.), Spider Ecophysiology (pp. 29–39). Berlin, Heidelberg: Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-33989-9 Schroeder, F. C., Taggi, A. E., Gronquist, M., Malik, R. U., Grant, J. B., Eisner, T., & Meinwald, J. (2008). NMR-spectroscopic screening of spider venom reveals sulfated nucleosides as major components for the brown recluse and related species. Proceedings of the National Academy of Sciences, 105(38), 14283–14287. https://doi.org/10.1073/pnas.0806840105 Sela, I., Ashkenazy, H., Katoh, K., & Pupko, T. (2015). GUIDANCE2: accurate detection of unreliable alignment regions accounting for the uncertainty of multiple parameters. Nucleic Acids Research, 43(W1), W7–W14. https://doi.org/10.1093/nar/gkv318 Shillington, C., & McEwen, B. (2006). Activity of Juvenile Tarantulas in and around the Maternal Burrow. Journal of Arachnology, 34(1), 261–265. Retrieved from http://www.jstor.org/stable/4489070 Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., & Flook, P. (1994). Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the Entomological Society of America, 87, 651–701. https://doi.org/00138746/94/0651 Simon, E. (1887). Observation sur divers arachnides: synonymies et descriptions. Annales de La Société Entomologique de France, 6(7). Sinsheimer, R. (1959). A single-stranded DNA from bacteriophagephi X174. Journal of Molecular Biology, 1, 43. Skinner, W. S., Dennis, P. a., Lui, a., Carney, R. L., & Quistad, G. B. (1990). Chemical characterization of acylpolyamine toxins from venom of a trap-door spider and two tarantulas. Toxicon, 28(5), 541–546. https://doi.org/10.1016/0041-0101(90)90298-L Smith, A. (1986). The Tarantula: Classification and Identification Guide. London: Fitzgerald Publishing. 116

Smith, A. (1987). The Tarantula: Classification and Identification Guide (Second). London, UK: Fitzgerald Publishing. Smith, A. M., & Jacobi, M. A. (2015). Revision of the genus Phormingochilus (Araneae, Theraphosidae, Ornithoctoninae) with the description of three new species from Sulawesi and Sarawak and notes on the placement of the genera Cyriopagopus, Lampropelma and Omothymus. Journal of the British Tarantula Society, 30(3), 25–48. Smith, J. J., Lau, C. H. Y., Herzig, V., Ikonomopoulou, M. P., Rash, L. D., & King, G. F. (2015). Chapter 8. Therapeutic Applications of Spider-Venom Peptides. In Venoms to Drugs - Venom as a Source for the Development of Human Therapeutics (pp. 221– 244). London, UK: Royal Society of Chemistry. https://doi.org/10.1039/9781849737876-00221 Stradling, D. J. (1994). Distribution and behavioral ecology of an arboreal `tarantula’ spider in Trinidad. Biotropica, 26(1), 84. https://doi.org/10.2307/2389113 Strand, E. (1907). Aviculariidae und Atypidae des Kgl. Naturalienkabinetts in Stuttgart. Jahreshefte Des Vereins Für Vaterländische Naturkunde in Württemberg, 63, 1–100. Talavera, G., & Castresana, J. (2007). Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Systematic Biology, 56, 564–577. Tedford, H. W., Sollod, B. L., Maggio, F., & King, G. F. (2004). Australian funnel-web spiders: Master insecticide chemists. Toxicon, 43, 601–618. https://doi.org/10.1016/j.toxicon.2004.02.010 Théry, M., & Casas, J. (2002). Predator and prey views of spider camouflage. Nature, 415(6868), 133–133. https://doi.org/10.1038/415133a Thomson, R. C., Wang, I. J., & Jhonson, J. R. (2010). Genome-enabled development of DNA markers for ecology, evolution and conservation. Molecular Ecology, 19(11), 2184–2195. https://doi.org/10.1111/j.1365-294X.2010.04650.x Thorell, T. (1881). Studi sui Ragni Malesi e Papuani. III. Ragni dell’Austro Malesia e del Capo York, conservati nel Museo civico di storia naturale di Genova. Annali Del Museo Civico Di Storia Naturale Di Genova, 17, 1–727. Trifinopoulos, J., Nguyen, L.-T., von Haeseler, A., & Minh, B. Q. (2016). W-IQ-TREE: a fast online phylogenetic tool for maximum likelihood analysis. Nucleic Acids Research, 44(W1), W232–W235. https://doi.org/10.1093/nar/gkw256 Tsuyama, N., Mizuno, H., & Masujima, T. (2012). Molecular and Functional Analysis of Cellular Phenomena Using Single-Cell Mass Spectrometry. Biological & Pharmaceutical Bulletin, 35(9), 1425–1431. https://doi.org/10.1248/bpb.b212012 117

Uetz, G. W. (1991). Habitat structure and spider foraging. In Habitat Structure (pp. 325– 348). Dordrecht: Springer Netherlands. https://doi.org/10.1007/978-94-011-3076-9_16 Uetz, G. W. (1992). Foraging strategies of spiders. Trends in Ecology & Evolution, 7(5), 155–159. https://doi.org/10.1016/0169-5347(92)90209-T Vassilevski, A. A., Kozlov, S. A., & Grishin, E. V. (2009). Molecular diversity of spider venom. Biochemistry (Mosc), 74(13), 1505–34 ST–Molecular diversity of spider venom. https://doi.org/BCM74131505 Vavrek, M. J. (2011). Fossil: palaeoecological and palaeogeographical analysis tools. Palaeontologia Electronica. Retrieved from http://palaeo- electronica.org/2011_1/238/index.html Venables, W. N., & Ripley, B. D. (2002). Modern Applied Statistics with S. New York: Springer. Retrieved from http://www.stats.ox.ac.uk/pub/MASS4 Vink, C. J., Derraik, J. G. B., Phillips, C. B., & Sirvid, P. J. (2011). The invasive Australian redback spider, Latrodectus hasseltii Thorell 1870 (Araneae: Theridiidae): current and potential distributions, and likely impacts. Biological Invasions, 13(4), 1003–1019. https://doi.org/10.1007/s10530-010-9885-6 von Dohlen, C. D., & Moran, N. A. (1995). Molecular phylogeny of the homoptera: a paraphyletic taxon. Journal of Molecular Evolution, 41(2), 211–223. https://doi.org/10.1007/BF00170675 Walckenaer, C. A. (1837). Histoire naturelle des insectes. (1st ed.). Paris: Aptères. Wang, Z. L., Yang, X. Q., Wang, T. Z., & Yu, X. (2017). Assessing the effectiveness of mitochondrial COI and 16S rRNA genes for DNA barcoding of farmland spiders in China. Mitochondrial DNA Part A: DNA Mapping, Sequencing, and Analysis, 0(0), 1– 8. https://doi.org/10.1080/24701394.2017.1350949 Warren, D., Geneva, A., & Lanfear, R. (2016). Comparing RWTY plots : what am I looking at ? Warren, D. L., Geneva, A. J., & Lanfear, R. (2017). RWTY (R We There Yet): An R package for examining convergence of Bayesian phylogenetic analyses. Molecular Biology and Evolution, msw279. https://doi.org/10.1093/molbev/msw279 West, R. C., & Nunn, S. C. (2010a). A taxonomic revision of the tarantula spider genus Coremiocnemis Simon 1892 (Araneae, Theraphosidae), with further notes on the Selenocosmiinae. Zootaxa, 2443, 1–64. West, R. C., & Nunn, S. C. (2010b). A taxonomic revision of the tarantula spider genus Lyrognathus Pocock 1895 (Araneae, Theraphosidae), with notes on the Selenocosmiinae. Zootaxa, 2362, 1–43. 118

West, R. C., Nunn, S. C., & Hogg, S. (2012). A new tarantula genus, Psednocnemis, from West Malaysia (Araneae: Theraphosidae), with cladistic analyses and biogeography of Selenocosmiinae Simon 1889. Zootaxa, 43(3299), 1–43. von Dohlen, C. D., & Moran, N. A. (1995). Molecular phylogeny of the homoptera: a paraphyletic taxon. Journal of Molecular Evolution, 41(2), 211–223. https://doi.org/10.1007/BF00170675 Walckenaer, C. A. (1837). Histoire naturelle des insectes. (1st ed.). Paris: Aptères. Warren, D., Geneva, A., & Lanfear, R. (2016). Comparing RWTY plots : what am I looking at ? Warren, D. L., Geneva, A. J., & Lanfear, R. (2017). RWTY (R We There Yet): An R package for examining convergence of Bayesian phylogenetic analyses. Molecular Biology and Evolution, msw279. https://doi.org/10.1093/molbev/msw279 Whiting, M. F., Carpenter, J. C., Wheeler, Q. D., & Wheeler, W. C. (1997). The Strepsiptera problem: phylogeny of the holometabolous insect orders inferred from 18S and 28S ribosomal DNA sequences and morphology. Systematic Biology, 46(1), 1–68. https://doi.org/10.1093/sysbio/46.1.1 Wilkins, J. S. (2009). Species: The History of an Idea. Berkeley, CA: University of California Press. Wilmer, J. W., & Couper, P. (2015). Phylogeography of north-eastern Australia’s Cyrtodactylus radiation: a habitat switch highlights adaptive capacity at a generic level. Australian Journal of Zoology, 63(6), 398. https://doi.org/10.1071/ZO15051 Wilson, J. D., Hughes, J. M., Raven, R. J., Rix, M. G., & Schmidt, D. J. (2018). Spiny trapdoor spiders (Euoplos) of eastern Australia: Broadly sympatric clades are differentiated by burrow architecture and male morphology. Molecular Phylogenetics and Evolution, 122(January), 157–165. https://doi.org/10.1016/j.ympev.2018.01.022 Wilson, J. S., Gunnell, C. F., Wahl, D. B., & Pitts, J. P. (2013). Testing the species limits of the tarantulas (Araneae: Theraphosidae) endemic to California’s Southern Coast Ranges, USA. Insect Conservation and Diversity, 6(3), 365–371. https://doi.org/10.1111/icad.12000 Wishart, G., & Rowell, D. M. (2008). Trapdoor spiders of the genus Misgolas (Mygalomorphae: Idiopidae) from eastern New South Wales, with notes on genetic variation. Records of the Australian Museum, 60(1), 45–86. https://doi.org/10.3853/j.0067-1975.60.2008.1495 Woinarski, J., Mackey, B., Nix, H., Traill, B., & McMillan, L. (2007). The Nature of Northern Australia. ANU Press. Retrieved from http://www.jstor.org/stable/j.ctt24h8n3 119

World Spider Catalog. (2018). World Spider Catalog. Retrieved March 19, 2018, from http://wsc.nmbe.ch Wright, S. (1943). Isolation by Distance. Genetics, 28(2), 114–138. https://doi.org/Article Zhang, D.-X., & Hewitt, G. M. (2003). Nuclear DNA analyses in genetic studies of populations: practice, problems and prospects. Molecular Ecology, 12(3), 563–584. https://doi.org/10.1046/j.1365-294X.2003.01773.x Zhang, Y.-Y., Huang, Y., He, Q.-Z., Luo, J., Zhu, L., Lu, S.-S., … Liang, S.-P. (2015). Structural and Functional Diversity of Peptide Toxins from Tarantula Haplopelma hainanum (Ornithoctonus hainana) Venom Revealed by Transcriptomic, Peptidomic, and Patch Clamp Approaches. Journal of Biological Chemistry, 290(22), 14192– 14207. https://doi.org/10.1074/jbc.M114.635458

120

Appendices

121

Supplementary Material

Chapter Two

Table of contents Supplementary figure 2.1. 16S gene MCC tree of AMC...... 123 Supplementary figure 2.2. 16S gene MCC tree of AEC ...... 123 Supplementary figure 2.3. 16S gene MCC tree of CYQC ...... 124 Supplementary figure 2.4. 16S gene MCC tree of TKC ...... 124 Supplementary figure 2.5. 18S gene NJ tree of AMC ...... 125 Supplementary figure 2.6. 28S gene NJ tree of AMC ...... 125 Supplementary figure 2.7. EF1γ gene NJ tree of AMC ...... 126 Supplementary figure 2.8. 18S gene NJ tree of AEC ...... 126 Supplementary figure 2.9. 28S gene NJ tree of AEC ...... 127 Supplementary figure 2.10. EF1γ gene NJ tree of AEC ...... 127 Supplementary figure 2.11. 18S gene NJ tree of CYQC ...... 128 Supplementary figure 2.12. 28S gene NJ tree of CYQC ...... 128 Supplementary figure 2.13. EF1γ gene NJ tree of CYQC ...... 129 Supplementary figure 2.14. 18S gene NJ tree of TKC ...... 129 Supplementary figure 2.15. 28S gene NJ tree of TKC ...... 130 Supplementary figure 2.16. EF1γ gene NJ tree of TKC ...... 130 Supplementary figure 2.17. PCA including 20 environmental variables collected from ALA ...... 131 Supplementary Table 2.1. Environmental variables...... 131

122

Supplementary figure 2.1. 16S gene MCC tree of AMC. Values corresponds to posterior probabilities. GMYC and mPTP species recovery results on the right.

Supplementary figure 2.2. 16S gene MCC tree of AEC. Values corresponds to posterior probabilities. GMYC and mPTP species recovery results on the right.

123

Supplementary figure 2.3. 16S gene MCC tree of CYQC. Values corresponds to posterior probabilities. GMYC and mPTP species recovery results on the right.

Supplementary figure 2.4. 16S gene MCC tree of TKC. Values corresponds to posterior probabilities. GMYC and mPTP species recovery results on the right. 124

Supplementary figure 2.5. 18S gene NJ tree of AMC.

Supplementary figure 2.6. 28S gene NJ tree of AMC.

125

Supplementary figure 2.7. EF1γ gene NJ tree of AMC.

Supplementary figure 2.8. 18S gene NJ tree of AEC.

126

Supplementary figure 2.9. 28S gene NJ tree of AEC.

Supplementary figure 2.10. EF1γ gene NJ tree of AEC.

127

Supplementary figure 2.11. 18S gene NJ tree of CYQC.

Supplementary figure 2.12. 28S gene NJ tree of CYQC.

128

Supplementary figure 2.13. EF1γ gene NJ tree of CYQC.

Supplementary figure 2.14. 18S gene NJ tree of TKC.

129

Supplementary figure 2.15. 28S gene NJ tree of TKC.

Supplementary figure 2.16. EF1γ gene NJ tree of TKC.

130

Supplementary figure 2.17. PCA including 20 environmental variables collected from ALA. AEC on blue, TKC on green, AMC on purple and CYQC on red.

Supplementary Table 2.1. Environmental variables used on biogeographic analysis of Australian tarantulas. The first three represent the variables that explained most of variation on Component 1 and 2. Environmental variable PC 1 PC 2 Precipitation - warmest quarter (Bio18) 0.42 0.85 Precipitation - autumn 0.43 -0.01 Precipitation - summer 0.79 -0.42 Precipitation - wettest period (Bio13) 0.07 0.09 Precipitation - driest period (Bio14) 0.00 0.00 Precipitation - coldest quarter (Bio19) 0.00 0.09 Precipitation - seasonality (Bio15) 0.04 0.00 Temperature - coldest period min (Bio06) 0.01 0.01 Temperature - driest quarter mean (Bio09) 0.00 0.00 Temperature - seasonality (Bio04) 0.00 0.00 Temperature - warmest period max (Bio05) 0.00 -0.01 Temperature - wettest quarter mean (Bio08) 0.00 0.00 Precipitation - spring 0.09 -0.15 Precipitation - winter 0.00 0.10 Temperature - coldest month min 0.01 -0.01 Temperature - month hottest maximum 0.00 0.00 Temperature - annual max mean 0.00 -0.01 Soil Moisture Index - annual mean (Bio28) 0.00 0.00 Precipitation - equinox seasonality ratio 0.00 0.00 Precipitation - solstice seasonality ratio 0.07 -0.22

131

Supplementary Material

Chapter Three

Table of contents Supplementary figure 3.1. ABGD of 16S and CO1 from Phlogius samples ...... 133 Supplementary figure 3.2. IBD of 18S and 28S of Phlogius samples ...... 133 Supplementary figure 3.3. IBD of EF1γ sequences of Phlogius samples ...... 134 Supplementary figure 3.4. ML tree of COI and 16S genes ...... 134 Supplementary figure 3.5. ML tree of 18S and 28S genes ...... 135 Supplementary figure 3.6. ML tree of EF1γ gene ...... 135 Supplementary figure 3.7. ML tree of 5 genes from Phlogius ...... 136 Supplementary figure 3.8. Haplotype network of 18S and 28S on QLD map ...... 136 Supplementary figure 3.9. PCA of morphometric values from Phlogius...... 137 Supplementary figure 3.10. LDA of morphometric values from Phlogius ...... 137 Supplementary figure 3.11. PCA of morphometric values from males and females Phlogius ...... 138 Supplementary table 3.1. Morphometric values of Phlogius ...... 138

132

Supplementary figure 3.1. ABGD generated using 16S and CO1 from Phlogius samples. Left picture analysed using K2P distance and right picture using p distance.

Supplementary figure 3.2. IBD of concatenated matrix of 18S and 28S of Phlogius samples. Genetic distance is p distance and geographic distance in kilometres.

133

Supplementary figure 3.3. IBD of EF1γ sequences of Phlogius samples. Genetic distance is p distance and geographic distance in kilometres.

Supplementary figure 3.4. ML tree of concatenated matrix of COI and 16S genes. No sequences for Coremiocnemis are included. Numbers correspond to bootstrap values.

134

Supplementary figure 3.5. ML tree of concatenated matrix of 18S and 28S genes. Coremiocnemis specimens are nested in between Phlogius clades. All nodes present BS > 70%.

Supplementary figure 3.6. ML tree of EF1γ gene. Coremiocnemis specimens are nested in between Phlogius clades. Numbers correspond to bootstrap values.

135

Supplementary figure 3.7. ML tree of concatenated matrix of COI, 16S, 18S, 28S and EF1γ genes from Phlogius. No sequences for Coremiocnemis are included. Numbers correspond to bootstrap values.

Supplementary figure 3.8. Haplotype network of 18S and 28S on Queensland map. Different colours represent different alleles.

136

Supplementary figure 3.9. PCA of morphometric values from Phlogius females (left) and males (right). Blue polygon represents southern species, green polygon central species and red polygon represents northern species

Supplementary figure 3.10. LDA of morphometric values from Phlogius females (left) and males (right). Blue polygon represents southern species, green polygon central species and red polygon represents northern species.

137

Supplementary figure 3.11. PCA of morphometric values from Phlogius females (dots) and males (triangles). Blue polygon represents southern species and red polygon represents northern species.

Supplementary table 3.1. Morphometric values of Phlogius specimens

Carapace

Ventral

4th leg 4th

1st leg 1st

Male

Number

Fovea

Eye tubercle width Eye tubercle

Metatarsus lengthMetatarsus lengthMetatarsus

Metatarsus width Metatarsus width Metatarsus Sex

Carapace length Carapace

Carapace width Carapace

Sternum length Sternum

Sternum width Sternum

Labium length Labium

Maxilla length Maxilla

Patella length length Tarsus Patella length length Tarsus

Femur length Femur length Femur width Labium

Caput

Tibia length Tibia length Tibia

Bulblength

Tibia width Tibia width Tibia

-

Eyes tubercle

width

S10507

17.84 12.52 12.67 18.94 10.91 16.55 12.77 16.05 13.77

6.26 3.22 7.59 3.11 2.52 8.28 16.6 2.42 1.69 9.74 8.57 3.69

7.8 8.8

20

M

3

S10616

18.47 11.36 11.15 17.42 15.41 15.74 12.47

5.71 15.9 2.58 9.72 11.8 6.91 2.46 2.25 7.55 15.9 1.47 7.49 5.99 2.03 2.82 8.26

7.1

M

2

S10690

15.92 14.04 10.52 10.06 16.59 10.21 14.97 11.94 15.05 12.96 15.33

5.01 2.44 6.94 2.56 2.17 6.73 6.54 1.89 1.49 7.31 6.39 2.15 2.83 7.68

M

138

S10673

19.93 17.84 12.67 12.26 18.98 10.41 16.19 11.38 16.86 12.83 16.24

3.32 7.02 3.46 2.57 7.78 7.56 2.43 1.63 8.43 7.58 2.56 2.77 8.91

M

6

S10465

18.87 11.91 11.62 17.45 10.16 15.58 12.55 15.91 13.14 16.75

5.79 16.9 2.91 7.13 3.07 2.46 7.68 7.62 2.33 1.53 7.46 2.48 3.24 8.92

8.9

M

S105049

19.71 11.51 11.89 17.51 11.33 16.39 13.31 15.78

5.72 2.87 10.4 6.16 3.07 2.27 7.85 6.74 1.55 8.62 2.63 3.36

2.4 8.4 8.5

17 15

M

S34610

19.19 17.36 11.52 19.19 11.19 16.37 12.39 16.86 12.76 16.94

5.74 3.05 7.35 2.84 7.58 7.57 2.27 1.64 2.98 3.34 8.92

2.5

12

M

? ?

CN08

16.8 14.4 2.16 9.87 10.7 15.7 7.35 13.1 15.9 6.08 2.06 1.51 7.68 7.04 2.24 3.05

7.7

M

5 ? ? ? ? ? ? ?

S105049

15.52 13.05 14.68 12.55 11.33 13.63

4.96 2.45 9.14 9.13 8.41 9.69 5.28 2.12 6.38 5.57 1.75 1.29 6.41 6.47 2.53 7.03

1.7 1.8

14

M

S10502

20.31 16.62 12.11 11.16 19.52 11.06 17.18 13.24 17.61 14.37 17.22

5.61 2.58 3.08 2.44 7.89 7.04 2.06 1.46 8.44 7.82 2.72 8.75

M

7 3

S10695

18.09 15.59 10.77 11.16 17.78 10.26 14.92 11.39 15.79 12.72 15.71

2.52 7.05 2.96 2.14 7.65 2.18 1.37 7.86 7.57 2.18 8.52

6.9 3.3

M

6

S10704

19.54 17.47 11.21 12.26 19.67 11.41 17.74 13.92 17.38 14.57 17.73

5.86 3.11 2.35 8.27 6.83 2.31 1.48 9.44 8.43 2.58 3.92 8.87

3.1

M

7

S10704

18.48 15.82 10.41 18.25 16.58 16.95 13.69 16.75

5.55 2.63 10.5 13.2 6.93 2.06 7.69 1.94 1.41 7.95 7.57 2.48 3.45 7.89

2.3 6.7

11

M

S4553

16.77 15.26 10.15 11.37 10.05 15.25 11.48 15.09 15.39

5.51 2.77 17.3 7.44 2.62 2.18 7.18 12.5 1.91 1.27 8.06 7.78 3.29 8.05

2.4

M

7

S67299

17.14 10.06 10.09 15.21 13.19 10.15 14.46 11.91 14.11

5.35 2.63 8.98 5.74 2.65 1.78 6.93 6.12 1.77 1.27 7.61 7.34 2.35 3.04 7.35

15

M

S10496

19.64 16.94 11.71 12.45 16.28 10.13 13.74 10.32 14.41 11.64 13.44

2.74 6.05 3.53 2.81 7.38 6.23 2.59 1.73 9.17 8.54 2.38 3.74 8.92

F

S10674

22.92 14.26 11.34 12.18 13.96

3.26 14.3 15.1 6.83 4.12 3.31 16.1 8.26 6.87 9.48 3.38 4.45 9.78

2.9 1.9

19 18 11 10

F

S105043

24.08 20.18 14.75 17.18 11.47 13.74 10.93 16.65 12.81 14.82 10.63

3.15 14.5 6.16 3.72 8.87 7.08 3.06 9.84 3.64 4.26 9.88

2.8 1.8

F

139

S10631

13.93 11.66 10.22 10.49 10.22

2.22 8.08 8.37 6.81 8.49 6.58 3.66 1.99 1.44 5.38 1.57 1.05 6.12 5.98 1.97 2.92 6.13

8.4 4.6

F

S56432

22.73 19.94 15.05 15.82 18.35 15.18 12.57 15.52 12.16 15.51 10.06

3.17 6.51 3.15 2.81 6.47 2.61 1.73 3.29 4.65

12

F

9 ? ?

S10489

19.74 16.85 12.37 12.58 16.32 10.56 13.23 10.76 14.41 11.11 13.11

2.62 6.42 3.15 2.71 6.52 1.71 2.74 3.75 8.33

7.8 2.4

F

? ?

S10661

18.81 16.79 12.31 13.41 17.05 14.04 11.04 14.97 11.85 13.91

2.89 9.82 2.93 2.78 7.65 5.96 2.07 1.75 7.71 7.16 1.92 3.16 8.45

F

6

S48240

16.05 13.55 10.78 11.09 11.73 10.07

2.54 9.75 7.66 9.06 6.77 4.56 2.64 1.76 8.11 4.39 2.09 1.35 6.62 6.43 1.85 2.88 6.22

6.7

F

S10466

20.85 17.46 12.37 13.45 16.07 10.24 12.68 14.38 10.46 13.26

3.16 6.23 3.12 2.67 7.62 6.17 9.69 8.38 3.12 3.93 9.05

2.4 1.7

11

F

S10474

22.48 17.89 14.56 13.25 17.97 11.52 14.42 11.44 15.93 11.44 15.47 10.03

2.91 6.62 3.77 3.12 8.59 5.95 2.87 2.95 4.26

1.9

F

? ?

S10491

21.26 17.52 13.44 13.54 10.55 15.52 13.47

2.76 17.3 14.2 3.41 2.84 8.13 11.6 6.06 2.57 1.72 9.19 8.51 2.92 3.84

6.2 9.8

11

F

S10687

17.57 12.96 13.57 11.18 14.21 11.41 15.25 11.63 14.33

20.6 2.99 17.4 6.12 3.12 2.58 8.11 6.26 7.83 3.62 9.07

2.5 1.6 9.2 3.1

F

S10482

11.06 11.43 10.56 10.69

13.7 2.12 8.49 8.54 6.67 9.77 7.79 5.21 2.46 1.99 8.63 5.12 1.72 5.46 5.53 1.63 2.45 5.99

5.6 1.4

F

S10749

22.84 19.83 14.09 14.36 17.93 11.47 11.51 13.52

2.68 13.3 5.93 3.61 15.5 6.09 2.56 1.89 2.93

8.5 9.7 9.6 4.1

11 10

F

3

S97175

23.19 19.86 15.94 14.84 18.07 11.96 12.24 11.72 15.07 10.61

3.24 15.3 3.82 3.16 8.99 5.41 1.89 9.69 9.43 2.73 4.07

6.6 2.7

F

?

S63729

24.17 21.45 15.34 17.19 11.23 13.91 11.64 15.44 10.86 10.72 10.34

4.41 15.9 6.41 3.72 8.74 15.3 2.66 2.09 3.55 4.35 9.77

5.8

F

3

S10682

16.85 14.35 10.35 10.25 14.63 12.55 11.25 10.48 11.72

2.39 8.92 5.18 2.71 2.22 5.93 1.26 7.56

9.5 6.4 1.7

F

? ? ? ?

S10470

22.06 13.02 14.16 17.86 11.56 15.41 12.95 16.71

3.27 6.93 3.43 2.65 12.5 15.6 6.17 2.72 1.42 9.74 3.32 3.97 9.45

8.3 8.8

19

F

140

JRF0

22.4 18.4 3.11 13.2 13.6 18.2 11.5 15.1 11.9 6.44 2.71 16.3 8.43 12.6 15.1 6.38 2.41 1.74 9.51 3.29 4.14 9.64

3.3

F

9

SIWR2

15.62 12.94 13.76 11.57 12.95 10.56 11.84

9.15 9.66 8.22 9.49 2.13 1.76 6.22 5.63 1.74 1.17 6.74 6.55 2.24 2.64

2.2 5.5 6.6

F

S35388

21.08 17.43 13.21 13.09 16.09 12.96 10.61 15.09 13.74

3.02 9.59 5.49 2.68 10.8 5.55 2.46 1.63 9.14 7.91 3.43 8.72

3.4 7.9 2.9

F

MAREEBA1

21.93 19.44 13.85 14.27 19.28 12.14 16.21 13.08 16.97 15.55

3.39 3.45 3.11 8.59 12.4 6.26 2.29 1.72 9.53 8.45 2.98 4.12 9.95

F

6

S10683

11.77 11.93 10.06 11.16 10.67

13.6 2.23 8.06 8.68 7.27 7.76 4.58 1.67 5.16 4.88 1.62 1.18 5.54 5.61 1.78 2.18 5.79

8.9

F

2

WEIPA4

23.23 18.76 14.53 19.49 11.46 15.33 12.88 16.79 16.62 10.79 10.22

3.19 14.2 6.12 3.46 2.91 12.6 6.38 2.95 1.79 9.51 2.88 4.17

F

9

JRF6

19.6 17.1 3.05 13.2 13.2 16.3 10.2 13.7 10.9 6.13 3.48 7.92 11.1 13.2 6.19 2.29 1.85 8.47 6.94 2.85 3.05 8.91

2.8

15

F

S10509

20.78 18.81 12.89 14.22 14.36 11.65 12.39

3.58 17.3 6.38 3.31 2.54 15.8 7.97 14.6 6.35 2.27 1.63 8.95 3.04 3.85

9.4 9.2

11

F

S62159

18.15 15.53 10.55 10.15 13.63 14.61 11.39 13.61

2.62 11.2 10.8 6.46 2.89 7.12 2.07 1.47 7.61 2.48 7.97

2.6 5.9 7.9

16

F

3

JRF4

10.7 1.93 7.79 11.8 7.23 9.75 7.33 4.63 1.96 1.57 11.1 4.93 8.93 10.4 5.17 1.55 1.16 5.36 5.63 1.84 2.48 5.75

7.6

13

F

S10698

18.23 16.31 12.81 12.03 16.94 14.22 11.47 15.03 11.37 14.03

2.62 9.86 6.34 2.95 2.56 7.69 6.63 2.27 1.54 8.24 8.08 2.35 8.74

F

3

JRC1

16.7 14.9 2.34 10.3 11.1 13.5 8.97 12.9 10.3 5.37 2.12 1.79 12.8 6.27 10.8 5.55 1.73 1.17 7.42 6.77 2.54 3.28 7.88

13

F

S91202

18.68 16.24 11.44 12.32 16.05 10.13 13.49 10.21 14.95 11.55

2.52 6.34 3.21 2.46 7.44 6.22 2.46 1.52 7.54 7.92 2.61 3.12 8.49

14

F

S91202

17.47 14.78 10.11 10.72 15.11 13.87 10.91 13.04

2.57 8.96 12.7 9.52 5.66 2.89 2.23 6.66 2.14 1.45 7.18 6.88 2.28 3.07

7.6

F

?

S91202

13.95 12.36 12.42 11.72 10.64

8.51 8.76 10.2 7.85 4.95 2.18 1.77 5.54 9.16 5.35 1.76 1.14 5.79 6.57 1.98 2.43 6.34

2.1 7.3

F

141

S10698

16.85 15.33 11.22 11.43 15.91 13.33 14.46 11.09 12.92

9.31 5.52 2.76 2.28 6.94 6.13 1.95 1.43 7.56 6.41 2.67 3.03

2.5 9.9 7.8

F

S10698

15.62 13.31 12.18 10.26 15.04 12.78 13.72 10.61 13.14

8.19 5.36 2.35 2.06 6.38 5.63 1.91 1.43 7.26 6.43 2.77 7.41

2.4 9.9 2.2

F

S10683

17.71 15.49 10.56 10.56 18.29 15.31 12.21 16.71 13.02 16.17

5.44 6.81 3.13 7.35 6.59 2.15 1.52 7.89 7.52 3.02 7.82

2.4 2.2 2.4

10

M

S91202

10.46 10.44 17.56 15.19 12.04 12.57 15.54

4.84 15.5 2.65 9.89 6.34 2.86 2.04 7.58 2.18 1.35 7.56 7.22 2.04 2.92 7.93

18 16

M

7

S91202

15.83 14.13 15.97 13.74 10.91 14.71 11.84 14.14

4.68 2.38 9.34 9.62 8.88 6.56 2.25 1.85 6.12 1.81 1.24 7.05 6.34 2.19 2.71 7.16

6.8

M

S91202

14.14 16.45 14.15 11.25 11.91 14.81

4.67 2.24 9.16 9.16 9.05 5.83 2.52 1.95 6.19 1.89 6.72 6.46 2.16 7.08

7.1 1.3 2.7

16 15

M

S91202

15.92 14.12 14.73 11.92 14.52

16.2 14.5 2.31 9.47 9.12 10.5 5.95 2.47 6.66 6.11 1.95 1.28 2.45 7.15

4.8 9.2 1.9

M

7 7 3

S91202

13.34 10.85 13.92

4.66 15.7 13.7 2.22 9.34 15.3 5.46 2.29 1.79 14.1 6.29 11.4 5.72 1.84 1.17 6.71 5.76 1.96 2.49

8.8

M

9 7

4.62 15.8 13.7 2.46 9.14 8.82 13.2 5.62 2.34 13.9 6.42 11.2 1.81 1.22 6.73 6.25 2.35 2.79 6.84

9.9 1.9

P3

15 10 13

M

6

5.07 15.8 13.4 8.85 9.47 16.3 8.87 13.9 10.7 6.16 1.82 14.5 11.9 14.2 6.13 1.54 1.26 6.55 6.19 2.02 2.98 7.03

CL1

2.4 2.3 6.6

M

S10467

19.92 17.43 11.43 11.49 18.93 10.39 12.91 17.24 13.63 16.57

5.58 3.12 16.7 6.83 7.98 7.12 2.27 1.53 7.86 2.05 3.16 8.24

2.3 8.5

M

3

S10698

15.18 10.15 13.94 14.51 11.47 14.01

4.64 14.1 9.78 15.7 5.65 2.21 6.25 6.91 1.21 6.35 5.64 1.86 2.33 7.15

2.3 9.1 9.9 1.9 1.8

M

S20389

18.72 15.45 10.97 10.71 15.54 12.37 13.29 16.05

2.57 17.7 10.5 6.39 2.42 16.8 7.75 6.23 2.43 1.48 8.15 2.05 2.89 8.24

5.5 3.2

M

8

S21324

14.44 10.69 12.25 13.72 10.69 12.47

16.7 2.49 10.2 14.7 9.35 5.98 2.83 2.23 5.85 1.94 1.37 6.89 6.98 2.41 7.43

9.8

F

7 3

S38795

16.83 14.75 16.28 10.03 15.34 11.56 15.43 12.93 15.12

5.31 2.48 9.88 9.56 6.27 2.36 1.98 7.28 1.89 1.23 2.19 3.26 7.63

6.4 7.4 6.5

M

142

S10677

21.63 13.69 13.56 16.75 10.61 13.45 10.22 16.34 14.78

19.2 2.92 3.79 2.79 8.86 5.14 2.87 1.87 2.77 3.95 9.22

9.4 9.2

12

F

5

S105877

18.65 16.88 11.99 11.97 15.12 11.87 14.72 10.83

2.81 9.38 8.62 5.86 3.28 2.35 13.1 2.47 1.62 8.27 2.68 3.44 8.52

7.6 6.2

F

8

S10716

17.833

15.79 10.79 10.77 16.18 13.61 10.54 15.04 12.17 14.41

5.65 2.81 9.69 6.43 2.73 2.11 7.59 5.83 2.32 1.43 7.73 7.44 7.85

2.1 3.2

M

S50862

16.12 13.92 13.86 14.77 11.74 14.34

4.95 9.24 9.53 10.8 5.73 2.28 6.11 6.06 1.86 1.29 6.39 1.93 2.82 7.17

2.2 9.1 1.8 7.1

16

M

S10510

21.36 18.43 13.53 13.64 15.74 10.12 13.45 11.66 13.87

3.13 9.78 5.49 2.55 15.5 8.19 5.53 1.85 8.81 2.61

3.5 2.9 8.9 3.4 9.1

F

S63893

24.23 19.91 15.59 15.45 11.07 10.36 10.72

3.27 17.4 13.3 10.6 5.97 3.88 2.75 14.3 1.87 9.79 2.68

8.7 5.7 4.4

16 12

F

3

S17729

15.94 11.64 16.58 17.16 13.94 17.23

6.07 18.4 2.54 10.7 12.8 6.96 2.12 7.04 1.77 1.34 7.44 3.29 8.15

2.5 7.7 8.4 2.9

11 19

M

S105045

19.37 16.13 12.34 11.62 14.88 13.37

2.92 11.6 9.51 5.85 3.22 2.28 7.33 11.4 5.55 2.42 1.58 8.33 7.86 2.47 3.52 8.54

9.5

15

F

S6801

16.49 13.64 10.36 10.67 10.26 11.72

2.32 9.87 13.3 4.75 4.73 1.77 6.94 2.32 2.92 7.05

8.4 8.1 2.7 6.4 1.3 7.2

13

F

2

S10510

19.33 16.43 12.42 12.45 11.96 11.59 13.28

5.57 14.8 9.65 9.54 5.46 3.22 2.17 14.5 5.98 2.32 1.57 8.12 2.38 3.36 8.62

7.4 7.7

F

S87217

12.31 10.15 13.15 10.27 10.82 13.08

18.7 16.1 2.56 11.6 5.28 14.4 5.73 2.12 1.55 8.37 8.33 2.54 3.37 8.45

3.2 2.5 7.5

16

F

S10468

20.82 12.38 12.95 20.82 12.03 18.24 14.32 17.75

6.39 18.8 3.07 17.7 13.1 7.52 3.52 2.56 8.58 7.66 8.87 9.13 2.53 3.55 9.18

2.6 1.5

M

NO:0011

17.16 14.72 11.23 10.71 15.12 12.82 13.74 10.75 13.47

2.62 9.63 9.65 4.88 2.64 2.18 6.56 5.06 2.05 1.46 7.58 6.88 2.06 2.88 7.71

F

Joshgreen

16.26 13.96 11.35 10.39 12.17 10.15 10.86 12.09 11.43

2.45 7.53 4.89 2.72 2.11 6.25 9.85 4.69 2.12 1.42 1.95 2.55

6.3 7.1

F

?

143

S87203

17.61 14.65 11.95 11.88 16.02 13.79 11.06 14.37 11.53

2.71 9.96 5.96 2.75 2.35 13.8 5.67 2.21 1.59 1.92 7.97

F

7 ? ? ?

All#2335

17.85 15.96 11.63 16.09 14.02 10.75 12.47 15.22

2.53 9.69 2.96 2.07 15.4 6.16 2.19 1.39 7.62 6.67 1.99 2.66 8.07

4.8 5.1 7.6

11

M

S105880

14.75 12.25 11.72 11.32 11.77

4.16 9.55 8.82 6.48 9.93 8.15 4.77 2.13 1.66 5.92 9.82 5.15 2.06 6.43 5.25 1.35 1.97 6.55

2.1 1.4

M

S10715

17.15 16.29 11.12 11.34 17.91 10.31 16.08 12.56 13.34 16.61

5.35 6.48 2.86 15.7 6.28 2.25 7.36 7.04 1.99 2.34 8.08

2.7 2.2 7.4 1.4

M

144

Supplementary Material

Chapter Four

Table of contents Supplementary table 4.1. Sample information of tarantulas ...... 146 Supplementary figure 4.1. PCA analysis of all spider venom ...... 147 Supplementary figure 4.2. PCA of all venom samples excluding Clade 1 ...... 148 Supplementary figure 4.3. Chromatogram group A ...... 148 Supplementary figure 4.4. Chromatogram group B ...... 149 Supplementary figure 4.5. Chromatogram group C ...... 149 Supplementary figure 4.6. Chromatogram group D ...... 149 Supplementary figure 4.7. Chromatogram group E ...... 150 Supplementary figure 4.8. Chromatogram group F ...... 150 Supplementary figure 4.9. Chromatogram group G ...... 150 Supplementary figure 4.10. Chromatogram group H ...... 151 Supplementary figure 4.11. Chromatogram group I ...... 151 Supplementary figure 4.12. Chromatogram group J ...... 151 Supplementary figure 4.13. Chromatogram group K ...... 152 Supplementary figure 4.14. Chromatogram group L ...... 152 Supplementary figure 4.15. Chromatogram group M ...... 152 Supplementary figure 4.16. Chromatogram group N ...... 153 Supplementary figure 4.17. Chromatogram group O ...... 153 Supplementary figure 4.18. Chromatogram group P ...... 153 Supplementary figure 4.19. Chromatogram group Q ...... 154 Supplementary figure 4.20. Chromatogram group R ...... 154 Supplementary figure 4.21. Chromatogram group S ...... 154 Supplementary figure 4.22. Chromatogram group T ...... 155 Supplementary figure 4.23. Chromatogram group U ...... 155 Supplementary figure 4.24. Chromatogram group V ...... 155 Supplementary figure 4.25. Chromatogram group W ...... 156 Supplementary figure 4.26. Chromatogram group X ...... 156 Supplementary figure 4.27. Chromatogram group Y ...... 156 Supplementary figure 4.28. Chromatogram group Z ...... 157

145

Supplementary table 4.1. Sample information of each phylogenetic group of tarantulas. Including quantity, location and sex. Phylogenetic Sample Sex and Location group Quantity Maturity Kings Plains, Laura and 4 females A 5 Mareeba (QLD) 1 male The Tip of Cape York, Dixie, 15 females B 21 Weipa, Quinkan and Steve 4 males Irwin Natural Reserve (QLD) 2 juveniles 1 female C 2 Iron Range (QLD) 1 male Prosepine, Airlie Beach, 9 females D 11 Mackay, Magnetic Island, 1 male Townsville, Wallaman (QLD) 1 juvenile E 1 Cairns (QLD) 1 female F 1 Coombah (NSW) 1 female 1 female G 2 Glennie Tablelands (QLD) 1 male H 3 Camooweal (QLD) 3 females I 5 Barcaldine (QLD) 5 females J 1 Muttaburra (QLD) 1 female K 2 Townsville (QLD) 2 females L 1 Goodiwindi (QLD) 1 female 2 females M 3 Mt Bundey (NT) 1 juvenile N 3 Middle Point (NT) 3 females O 2 Adelaide River (NT) 2 females P 3 Jugbarra (NT) 3 females Kings Plain and Laura Q 2 2 females (QLD) R 1 Townsville (QLD) 1 female Avon Downs and S 2 2 females Barkly Tablelands (NT)

146

Winton and Muttaburra T 2 2 females (QLD) U 1 Darwin (NT) 1 female V 3 Bradshaw (NT) 3 females W 5 Bradshaw (NT) 5 females X 1 Tennant Creek (NT) 1 juvenile Y 3 Burt Plains (NT) 3 females Z 4 Alice Springs (NT) 4 females

Supplementary figure 4.1. PCA analysis of all spider venom using principal components 1 and 2. Each dot represents a specimen of Australian Theraphosidae spiders collected and the colours of the dots indicate phylogenetically close samples.

147

Supplementary figure 4.2. PCA analysis of all venom samples excluding phylogenetic groups A, B, C, D and E using principal components 1 and 2 to highlight differences between the venom of the remaining groups. Each dot represents a specimen of Australian Theraphosidae spiders collected and the colours of the dots indicate phylogenetically close samples.

Supplementary figure 4.3. Chromatogram overlapping of all samples included on group A.

148

Supplementary figure 4.4. Chromatogram overlapping of all samples included on group B.

Supplementary figure 4.5. Chromatogram overlapping of all samples included on group C.

Supplementary figure 4.6. Chromatogram overlapping of all samples included on group D.

149

Supplementary figure 4.7. Chromatogram sample of group E.

Supplementary figure 4.8. Chromatogram sample of group F.

Supplementary figure 4.9. Chromatogram overlapping of all samples included on group G.

150

Supplementary figure 4.10. Chromatogram overlapping of all samples included on group H.

Supplementary figure 4.11. Chromatogram overlapping of all samples included on group I.

Supplementary figure 4.12. Chromatogram sample of group J. 151

Supplementary figure 4.13. Chromatogram overlapping of all samples included on group K.

Supplementary figure 4.14. Chromatogram sample of group L.

Supplementary figure 4.15. Chromatogram overlapping of all samples included on group M. 152

Supplementary figure 4.16. Chromatogram overlapping of all samples included on group N.

Supplementary figure 4.17. Chromatogram overlapping of all samples included on group O.

Supplementary figure 4.18. Chromatogram overlapping of all samples included on group P. 153

Supplementary figure 4.19. Chromatogram sample of group Q.

Supplementary figure 4.20. Chromatogram sample of group R.

Supplementary figure 4.21. Chromatogram overlapping of all samples included on group S. 154

Supplementary figure 4.22. Chromatogram overlapping of all samples included on group T.

Supplementary figure 4.23. Chromatogram sample of group U.

Supplementary figure 4.24. Chromatogram overlapping of all samples included on group V. 155

Supplementary figure 4.25. Chromatogram overlapping of all samples included on group W.

Supplementary figure 4.26. Chromatogram sample of group X.

Supplementary figure 4.27. Chromatogram overlapping of all samples included on group Y. 156

Supplementary figure 4.28. Chromatogram overlapping of all samples included on group Z.

157

Supplementary Material

Chapter Five

Table of contents Supplementary Tables ...... 159

158

Supplementary tables. http://www.mdpi.com:8080/2072-6651/9/4/116/s1

159