Investigating the Genetic Basis of Plumage Variation in the Australian Magpie, Cracticus Tibicen

Home , Australian magpie, Black butcherbird, Cracticus, Kermode bear

Investigating the Genetic Basis of Plumage Variation in the Australian Magpie, Cracticus tibicen

Author Dobson, Ana Elizabeth

Published 2017

Thesis Type Thesis (PhD Doctorate)

School Griffith School of Environment

DOI https://doi.org/10.25904/1912/275

Downloaded from http://hdl.handle.net/10072/365743

Griffith Research Online https://research-repository.griffith.edu.au Investigating the genetic basis of plumage

variation in the Australian magpie,

Cracticus tibicen

Ana E. Dobson B.Sc. (Hons)

Griffith School of Environment Griffith Sciences Group Griffith University

Submitted in fulfilment of the requirements of the degree of Doctor of Philosophy

December 2016 SUMMARY

Many species exhibit some form of colour variation between groups or individuals, and studies of such colour variation, its genetic basis and functional importance can contribute unique insights into the evolutionary, ecological, demographic and genetic processes affecting the diversity of life. Through the examination of the genetics underlying colour traits, the relative importance of different selective and demographic processes in the evolution and maintenance of a trait can be inferred, enabling scientists to gain a better understanding of what drives evolutionary change in different organisms and environments. Such studies also assist in bridging the gaps in our knowledge about the interactions between genetics and ecology that are involved in phenotypic variation. Understanding the pathways linking genetic and regulatory changes to phenotypic change and evolutionary change through time contributes to our knowledge about how species develop and evolve.

The Australian magpie (Cracticus tibicen), is a group-living and territorial passerine bird with a characteristic pied appearance of strongly contrasting patches of black and white plumage. The species has several different plumage forms, delineated by the colour and/or pattern of back feathers; these plumage forms hybridise where their distributions overlap. Distributions of plumage forms are highly discordant with patterns of phylogeographic structure found in this species, making a scenario of allopatric divergence followed by secondary recontact at hybrid zones highly unlikely. Similar plumage colour forms seem likely to have evolved in situ within both eastern and western Australian magpie lineages. A number of previous studies have proposed that different forms of selection may be acting on plumage colour phenotypes either side of hybrid zones to generate and maintain these polymorphic forms in the face of gene flow. In such scenarios, black backs are putatively favoured in the north of the continent, by better crypsis where vegetation is more open-canopied, and white backed forms are favoured by sexual selection for ‘brightness’ in males, in the closed-canopy style vegetation dominating the far south of the continent, in which crypsis is less important. While experiments have found evidence to support the idea black backed magpies may be favoured via crypsis north of hybrid zones, as yet a number of studies looking for evidence of sexual selection favouring white backed males have failed to find strong evidence of such an effect, and it is currently unclear why white backed forms (including varied backs) dominate the south of the continent.

Previous phylogeographic analyses of the magpie suggest Pleistocene-era arid barriers to dispersal led to the genetic divergence of eastern and western groups II around 36,000 years ago, and also present some evidence for range expansion of eastern magpies following this period of aridity. Across all five of the nuclear candidate colour genes assayed for this study, tests of neutrality and selection did not identify signals consistent with selection acting on the gene(s), but did add significant support to the hypothesis of past population expansions within the recent evolutionary history of this species. It is suggested that this signature of population expansion may be linked to changes in population size and distribution of magpies during the late Pleistocene, as aridity decreased following the Last Glacial Maximum, allowing refugial populations to recolonise the interior of the continent.

A long-term study site in the eastern back colour hybrid zone, in which magpies have been progressively colour banded and DNA sampled over a 25 year period, was used to investigate several aspects of the quantitative genetics of magpie back colour variation. Microsatellite parentage analyses were first used in 35 territories with reliable sighting data and in which all, or nearly all adult members had been sampled, in order to establish genetic parentage of fledglings (as this species has been shown to have high levels of extra-pair and extra-group paternity). On average, data from each territory spanned a nine year period. Observations of individuals’ back colours, scored on a five-point scale, were averaged over the lifetime of the study, and this data was used to examine both the heritability of this back colour variation and the patterns of inheritance within family groups.

Both linear-mixed model and regression-based methods independently produced exceptionally similar estimates for the heritability of back colour variation in magpies, and these both indicated that this heritability is quite high: 0.92 using mixed models and 0.94 using corrected midparent/midoffspring regression. These large heritability estimates are comparable to those estimated for other melanin-based plumage traits across a wide range of bird species, and is at least partially a reflection of the need for melanic pigments to be manufactured endogenously, rather than obtained from environmental sources. This idea is reinforced by the non-significant contribution of patch quality and cohort to heritability models, indicating environmental contributions related to natal territories such as the differing quantity or quality of feeding sites between territories, and differing abundance of resources between years have little effect on the heritability of variation in this trait.

Regressions of offspring back colour on parents by sexes suggest neither maternal nor paternal non-genetic effects (e.g. parent body condition) are likely to play a large role in determining offspring back colour. There was no pattern of either positive or negative

III assortative mating with respect to parental back colours, which were not significantly correlated in mating pairs. Such assortative mating could have at least partially explained the maintenance of hybrid zones in the face of gene flow. Patterns of inheritance in familial groups suggest there may be some effect of dominance of black backed alleles, and these patterns together with the continuous nature of back colour phenotypes indicate that a minimum of two genes are likely to play a role in determining magpie back colour, whilst the upper limit of number of genes involved remains as yet undetermined. The high heritability estimate for back colour variation also indicates that the correlation between genotype and phenotype is very strong, and consequently future gene-mapping studies investigating this trait have a reasonable likelihood of pinning down genes of large effect that may play a role in magpie back colour.

Previous studies using a candidate gene approach to identify the genetic basis of traits, such as intraspecific colour variation, have yielded a large number of genotype- phenotype associations across a wide range of animal species. The melanocortin-1 receptor (MC1R) plays an important role in the synthesis of different types of melanin, and mutations in this gene have been associated with melanin-based pigmentation variation across a wide range of vertebrate taxa, including many bird species. The majority of the MC1R gene, comprising 861 nucleotides was sequenced in 98 magpie individuals, mapped onto the chicken genome and structural features of this transmembrane receptor were identified in magpies. Although four coding region mutations in magpies led to amino acid changes, none of these were partitioned between back colour forms, subspecies or previously delineated eastern and western clades. At sites of amino acid substitutions associated with colour variation in other bird species, magpies were consistently invariable.

A total of 3758 nucleotides from coding and non-coding regions of four additional candidate colour genes involved in pigmentation and melanocortin pathways were also sequenced: the Tyrosinase-related protein 1 (TYRP1), Tyrosinase (TYR), Dopachrome Tautomerase (DCT), and the Agouti-related protein (AGRP), in search of possible associations with plumage colour differences in this polymorphic species. For these four candidate genes, 30 magpie individuals, including representatives from all plumage colour forms (black back, white back and varied), both sexes, and each of eight recognised subspecies from 15 sites across the Australian continent were used to evaluate associations between genetic variation and colour variation using a series of inverse ordinal and univariate logistic regressions.

IV Of the 67 variable mutations identified, 64 were not significantly associated with plumage variation or geographic or subspecies groupings, and sites of mutations important in colour variation in other species were invariable in magpies. Of the four mutations identified by regression analyses as statistically significant, only two (both located in intron c of the TYRP1 gene) were associated with back colour variation, and neither of these mutations perfectly delineated different plumage forms. One of these was a deletion only found in some varied back individuals, whilst the other was an insertion separating most varied back birds and one white back bird from all other individuals. It is suggested that phylogeographic structure could be driving this significance, as these mutations were also significantly associated with previously delineated phylogeographic clades. This varied plumage form is restricted to the south-west of the continent, and previous microsatellite analyses has indicated a higher level of genetic diversity in magpie populations from this south-western region, compared to populations in other parts of the continent.

Due to the ontogenetic plumage development of magpies, and the continuous nature of magpie back colour variation in hybrid zones, it is suggested future studies of magpie pigmentation will benefit from a focus on regulatory elements and transcription factors of candidate colour genes. Genes involved in ‘patchy’ colour expression or pattern- based expression of melanins may also prove important in magpie back colour, along with splicing errors or variants that change the protein product of such genes. As results of this study suggest that more than one gene is involved, future studies will also need to carefully consider interactions between mutations, both within and between genes, and thus no gene should be considered in isolation. Pleiotropic genes may be involved in magpie colour variation, and it is suggested genes involved in thermoregulation and/or metabolism may be particularly relevant in magpies, as it is suggested that these physiological traits and their interaction(s) with back colour may explain an important part of the hypotheses of differential divergent selection on either side of back colour hybrid zones.

The missing puzzle piece from the broader question as to what is driving the creation and maintenance of magpie plumage colour forms and maintaining stable hybrid zones, is currently why white back forms are seemingly favoured in the south. This study outlines a mechanism that could putatively explain this, in which selection favours white back forms as a result of their more effective and efficient thermoregulatory abilities in regions (such as the south of the Australian continent) subject to high wind speeds during colder winter months.

V With recent sequencing of the draft genomes of many bird species, it is suggested targeted high-density genome scanning approaches, coupled with downstream expression and targeted candidate gene analyses of genes and genomic regions identified by these high-density scans have the greatest potential to reveal even further insights into the genetics underlying plumage colour variation in the Australian magpie, and thus also the evolutionary forces shaping the diversity of form and function across all species.

VI Statement of Originality

This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.

Ana E. Dobson

VII ACKNOWLEDGEMENTS

To my primary supervisor, Prof. Jane Hughes – I would like to thank you for your guidance, advice and the many critical discussions that have shaped both this thesis, and me as a scientist. You have been so incredibly supportive, encouraging and patient throughout this long, and at times, frustrating journey: after each of my numerous diagnoses, operations and setbacks, you were always ready to help me keep moving forward, solving problems and clearing roadblocks with kindness and a sensible plan of action. Thank you for believing in me.

A big thank you to Dr Dan Schmidt, my secondary supervisor. Your office door was always open to me, even when you were rushed off your feet with your own work, and I really enjoyed our problem-solving sessions. Your sense of perspective and dry humour helped get me through some difficult days.

I would like to acknowledge and thank all those who collected samples used in this study: Jane Hughes, Peter Mather, Alicia Toon, Kate Durrant, Andrew Baker, Corinna Lange, and the large number of volunteers who have helped with fieldwork at the Seymour study site over the last 25 years. That 14 km stretch of dirt road, its magpie residents and all the crazy shenanigans of the trips I took down there will always hold a very special place in my heart.

Thank you to Alicia Toon for her work on the Seymour parentage project, and patience in training me to be the ‘next magpie girl’. I have learnt so much from you. My thanks also to Steve Smith, for his involvement in early MC1R labwork.

This research was financially supported by Griffith University, through a Griffith School of Environment Postgraduate Research Scholarship. Samples were collected under Western Australian Department of Conservation and Land Management permit SF003903, Queensland Parks and Wildlife permit W4/002670/01/SAA, Northern Territory Parks and Wildlife Commission permit 18145, Victorian Wildlife permit 10003058 and animal ethics protocol AES/16/04/AEC.

VIII Thank you to James Fawcett, Daniel Schmidt, Brendan Cox and Bobbi Dobson- Patterson for laboriously and meticulously going through my drafts, as well as two anonymous reviewers who greatly improved the MC1R paper. I am particularly grateful to Petney Dixon for always being in my corner, and helping me cut through all the red tape after my illness.

To my fellow ‘Geeks’ from the Molecular Ecology lab: thank you all for creating a culture and environment that pushed me to grow as a scientist, and made work a place I couldn’t wait to get to every morning. Thank you for all the insightful discussions, rigorous debate, squash and football games, cut-throat card games, costume parties, camping trips and stupid bets. So many of you have become family, and I feel incredibly lucky to be part of such a supportive, generous, intelligent tribe. Special thanks to Joel Huey for always wanting to debate anything and everything about science and music, Jemma Somerville for her incredible patience, and Kathryn Real for introducing me to my husband.

My partner-in-crime, Sofie Bernays, thanks for being you: a small part crazy and a large part awesome. Doing my PhD alongside my best friend was so much fun. To my Jimmy, thanks for being the older brother I never had, taking me under your wing and letting me annoy you; your friendship means the world to me.

I am so grateful for the amazing support of my family. Thank you to my best friend and mother, Bobbi Dobson-Patterson; watching you get your PhD probably should have dissuaded me, but only inspired me more- you are my hero! Thanks to my little brother Andy Patterson, for putting up with me and helping out so much when I couldn’t.

Most of all, I want to thank my better half and husband Brendan Cox. You have been there for me every single step of the way, and sacrificed so much so that I could do what I love- it’s your turn next. Through it all you never doubted me, even when I doubted myself. Your love, support, understanding and patience run fathoms deep. I never could have done this without you.

IX TABLE OF CONTENTS

CHAPTER 1: GENERAL INTRODUCTION ...... 1

1.1 CONNECTING PHENOTYPE AND GENOTYPE IN ANIMALS ...... 1 1.2 PLUMAGE VARIATION IN BIRDS ...... 2 1.3 STUDY SPECIES: THE AUSTRALIAN MAGPIE ...... 6 1.3.1 Life history, variation and distribution ...... 6 1.3.2 Phylogeography of magpies ...... 11 1.4 QUANTITATIVE GENETICS OF MAGPIE BACK COLOUR ...... 13 1.4.1 Trait inheritance and pedigree analyses ...... 13 1.5 INVESTIGATING THE GENETIC BASIS OF MORPHOLOGICAL TRAITS ...... 15 1.5.1 Using candidate colour genes to examine the genetic basis of magpie plumage variation ...... 16 1.6 SELECTION AND MAGPIE PLUMAGE VARIATION ...... 22 1.7 RESEARCH QUESTIONS ...... 28 1.7.1 Heritability of plumage colour in magpies ...... 28 1.7.2 Magpie phylogeography ...... 28 1.7.3 Candidate colour genes and magpie plumage colour ...... 29 1.8 ORGANISATION AND STRUCTURE OF THIS THESIS ...... 30

CHAPTER 2: HERITABILITY AND PATTERNS OF INHERITANCE OF PLUMAGE VARIATION IN THE AUSTRALIAN MAGPIE, CRACTICUS TIBICEN ...... 31

2.1 INTRODUCTION ...... 31 2.2 METHODS ...... 41 2.2.1 Study site and field methods ...... 41 2.2.2 Genetic methods ...... 44 2.2.3 Genetic analyses ...... 47 2.2.4 Heritability analyses ...... 49 2.3 RESULTS ...... 52 2.3.1 Parentage analysis ...... 52 2.3.2 Back colour metrics, assortative mating and inheritance patterns ...... 55 2.3.3 Regressions ...... 58 2.3.4 Animal models ...... 62 2.4 DISCUSSION ...... 64 2.5 CONCLUSIONS ...... 72

CHAPTER 3: MC1R AND MAGPIE PHYLOGEOGRAPHY ...... 75

SEQUENCE VARIATION IN THE MELANOCORTIN-1 RECEPTOR (MC1R) DOES NOT EXPLAIN CONTINENT-WIDE PLUMAGE COLOUR DIFFERENCES IN THE AUSTRALIAN MAGPIE (CRACTICUS TIBICEN)

3.1 ABSTRACT ...... 75 3.2 INTRODUCTION ...... 76 3.3 MATERIAL AND METHODS ...... 85 3.3.1 Sampling ...... 85 3.3.2 MC1R genotyping and analysis ...... 86 3.4 RESULTS ...... 90 3.4.1 Analysis of genetic variation at MC1R ...... 90 3.4.2 Phylogeography ...... 90 3.4.3 MC1R as a candidate gene for magpie plumage differences ...... 93 3.5 DISCUSSION ...... 96 3.5.1 MC1R and plumage variation ...... 96 3.5.2 Phylogeography and selection at MC1R ...... 99 3.6 FUNDING ...... 102 3.7 ACKNOWLEDGEMENTS ...... 102

CHAPTER 4: CANDIDATE GENES FOR PLUMAGE DIFFERENCES IN THE AUSTRALIAN MAGPIE, CRACTICUS TIBICEN ...... 103

4.1 INTRODUCTION ...... 103 4.2 CANDIDATE COLOUR GENES ...... 109 4.2.1 Tyrosinase ...... 109 4.2.2 Tyrosinase-related protein 1 ...... 109 4.2.3 Dopachrome tautomerase ...... 111 4.2.4 Agouti-related protein ...... 112 4.3 MATERIALS AND METHODS ...... 113 4.3.1 Sampling ...... 113 4.3.2 Genotyping and analysis ...... 114 4.3.3 Bioinformatics analyses ...... 117 4.4 RESULTS ...... 119 4.4.1 Tyrosinase-related protein-1 ...... 120 4.4.2 Tyrosinase ...... 124 4.4.3 Dopachrome Tautomerase ...... 127 4.4.4 The agouti-related protein ...... 128 4.5 DISCUSSION ...... 131

CHAPTER 5: GENERAL DISCUSSION ...... 141

5.1 EXPLORING THE GENETIC BASIS OF INTRASPECIFIC COLOUR VARIATION ..... 141 5.2 ADDING TO THE PHYLOGEOGRAPHIC STORY OF THE MAGPIE ...... 141 5.3 HERITABILITY AND INHERITANCE PATTERNS OF MAGPIE BACK COLOUR ...... 143 5.4 SCREENING OF CANDIDATE COLOUR GENES ...... 146 5.5 DOES IT MATTER IF IT’S BLACK OR WHITE?: SELECTION AND MAGPIE BACK COLOUR VARIATION ...... 156 5.6 CONCLUSIONS ...... 166

REFERENCES ...... 168

APPENDICES ...... 185

APPENDIX I: FREQUENCY DISTRIBUTION OF MAGPIE BACK COLOURS WITHIN 15 KM OF THE

SEYMOUR STUDY SITE IN 1976. DATA FROM TRANSECTS CARRIED OUT IN 1976 BY BURTON &

MARTIN (1976) ...... 185

APPENDIX II: MIDPARENT – MIDOFFSPRING BACK COLOUR REGRESSION BASED ON THE ‘STRICT’

DATA SUBSET ...... 185

APPENDIX IV: SAMPLING LOCATIONS AND BACK COLOUR SAMPLING OF MAGPIE INDIVIDUALS

SCREENED AT TYRP1, TYR, DCT AND AGRP CANDIDATE GENES ...... 186

XII

LIST OF FIGURES

Figure 1.1: Distribution of Cracticus tibicen in Australia...... 8

Figure 1.2: Plumage forms of the Australian magpie (C. tibicen) and their sexual dimorphism ...... 10

Figure 1.3: Hybrid forms of black back and white back males in the eastern hybrid zone ...... 11

Figure 2.1: The three main back colour variants of Cracticus tibicen ...... 35

Figure 2.2: Distribution of C. tibicen and sub-species in Australia, and location of the long-term study site.

...... 36

Figure 2.3: Back colour scoring schematic ...... 43

Figure 2.4: Frequency distribution of magpie back colours in analysed territories of the Seymour study population ...... 56

Figure 2.5: Pairing patterns with relation to back colour of genetically verified mating pairs in the Seymour study population ...... 57

Figure 2.6: Offspring back colour ranges for different combinations of parental back colours ...... 58

Figure 2.7: Regressions of mid parent-mid offspring back colour calculated based on the complete dataset ...... 61

Figure 2.8: Regressions of parent-offspring back colour, based on the complete dataset ...... 62

Figure 2.9: Posterior density of heritability estimates for different animal models tested using GLMMs. .. 64

Figure 3.1: Magpie plumage variants and their distribution across the Australian continent ...... 80

Figure 3.2: Parsimony network of MC1R sequence data generated in TCS ...... 92

Figure 3.3: Partial amino acid sequence of the magpie MC1R gene ...... 95

Figure 4.1: Magpie plumage variants and their distribution across the Australian continent ...... 114

Figure 4.2: Three mutations in intron c of the C. tibicen TYRP1 gene found to be statistically significant in regressions of back colour, subspecies or phylogeographic grouping...... 124

Figure 5.1: Wind speed across the Australian continent ...... 163

Figure 5.2: Distribution of C. tibicen in Australia ...... 163

XIII

LIST OF TABLES

Table 2.1: Primer sequences and annealing temperatures of eight microsatellite loci developed for parentage in the Australian magpie ...... 46

Table 2.2: Summary statistics of microsatellite loci used in parentage analysis ...... 53

Table 2.3: Samples sizes used in complete dataset analyses of heritability ...... 54

Table 2.4: Average back colours of sampled birds in the Seymour population ...... 55

Table 2.5: Kin regressions using complete dataset ...... 60

Table 2.6: Kin regressions using strict data subset ...... 60

Table 2.7: Generalised linear mixed models (GLMMs) (animal models) of magpie back colour...... 63

Table 3.1: Number of sampled haplotypes, according to sub-species, site location and back colour ...... 86

Table 3.2: Analysis of molecular variance (AMOVA) results and tests of neutrality and selection ...... 93

Table 3.3: Comparison of MC1R amino acids across a number of bird species at SNP sites associated with plumage colour change ...... 95

Table 4.1: Primer sequences and annealing temperatures ...... 116

Table 4.2: Summary statistics of genetic variation and statistical tests of neutrality ...... 120

Table 4.3 : P-values of SNPs within candidate colour genes identified by inverse ordinal regressions... 120

Table 4.4: P-values of SNPs within candidate colour genes identified by univariate logistic regressions120

Table 4.5: Amino acid polymorphisms in the tyrosinase related-1 (TYRP1) protein associated with pigmentation changes in selected animal species ...... 122

Table 4.6: Amino acid polymorphisms in the Tyrosinase (TYR) gene associated with pigmentation changes in selected animal species ...... 126

Table 4.7: Amino acid polymorphisms and SNPs in the Dopachrome Tautomerase (DCT/ TYRP2) gene associated with pigmentation changes in selected animals ...... 128

Table 4.8: Amino acid polymorphisms in the agouti-related protein (AGRP) gene associated with pigmentation changes in selected animal species ...... 130

XI V XV General Introduction GENERAL INTRODUCTION

1.1 CONNECTING PHENOTYPE AND GENOTYPE IN ANIMALS

Morphological variation within species can take many forms. Attributes such as colouration, pattern, body and limb shape and size vary subtly and sometimes even substantially within many species, and this variation can be partitioned across a range of scales, from broad geographic divides, to sexual dimorphism, variation between populations and even individuals. Such variation has been the focus of a great deal of scientific research for the last few hundred years

(Beebee and Rowe, 2008), and as technology has advanced, the ways in which evolutionary questions about intraspecific variation have been approached have changed dramatically. Several different mechanisms have been invoked by researchers over the years to explain how and why intraspecific variation arises and/or is maintained, as well as the geographical distribution of such morphological variants.

Some morphological variation may not have a genetic basis, and can represent plastic responses to local environments or seasonal changes (Campbell and

Reece, 2002). Traits like body size or colouration may in some cases be environmentally plastic and partially altered by the diet, habitat quality or climate of an organism rather than its genotype (West-Eberhard, 1989, Price, 2006,

Norris et al., 2007). However it has also been suggested that traits of phenotypic plasticity themselves are likely to be the result of natural selection

(West-Eberhard, 1989), and more recent work has pointed to a molecular basis for some phenotypic plasticity, through altered patterns of expression of particular genes (Pigliucci, 1996, Beldade et al., 2011).

General Introduction

1.2 PLUMAGE VARIATION IN BIRDS

When distinct morphological forms of the same species coexist and interbreed in the same population, the term of polymorphism can be applied (Ford, 1945).

In some cases it can be difficult to distinguish between other forms of intraspecific variability and polymorphism. However, true polymorphism requires that traits have a genetic basis, intermediate forms between morphs are not present, and different morphs occur together in the same locations

(Ford, 1945). Whilst some researchers use the term polymorphic only to describe discrete rather than continuous traits (Buckly 1987, in (Roulin, 2004), it has also been convincingly argued that the distinction between continuous and discrete colour morphs of birds is scale-dependant, and morph-based species classifications are often relatively inconsistent (Roulin, 2004).

A review by Galeotti et al. (2003) of colour polymorphism in birds, identified 334 species which exhibit plumage colour variation, which is approximately 3.5% of all bird species. The occurrence of intraspecific colour variation is not exclusive to a particular taxonomic group of birds; in fact more than half of all bird orders contain at least one colour polymorphic species, indicating that this trait seems to have evolved independently on a large number of occasions throughout the evolutionary history of birds (Galeotti et al., 2003).

The problem of why some species exhibit morphological polymorphisms whilst others do not has puzzled biologists for years (Ford, 1945, Cain and Sheppard,

1954, Clarke, 1962, Galeotti et al., 2003). Whilst some polymorphisms may be transient and lost in relatively short time periods, others can be stable over long time periods and large geographic distances, whilst morphs in some species

General Introduction may meet at a ‘ratio cline’ of intermediate forms (e.g. Mundy et al., 2004). Many attempts have been made to explain how polymorphisms evolve and are maintained within different bird species.

Different plumage colours may be a reflection of past historical processes which have affected the species, and thus its evolutionary history. These include, but are not limited to: population expansions and constrictions, population bottlenecks, and historical barriers to gene flow, such as the Pleistocene arid barriers (Toon et al., 2007). A scenario of allopatric evolution has often been put forward as a likely mechanism for the development and/or maintenance of bird plumage polymorphisms (e.g. Avise et al., 1992, Zink, 1994, Bensch et al.,

1999, Kearns et al. 2008, Prkye & Griffith, 2009), but a review of many such studies by Roulin (2004) indicates there are only few examples that have demonstrated direct evidence supportive of such a scenario (e.g. Cooke et al.,

1988, Zink, 2012). The most common scenario suggested is that populations of the same species become isolated, plumage colour of one or both groups diverges and is fixed as a result of natural selection, random mutations and/or population bottlenecks or founder effects, and (but not always) secondary recontact between the groups further down the track may create a zone of introgression.

Neutral hypotheses suggest a balance between mutation and genetic drift may produce and/or maintain plumage variants within bird species. Variable plumage colour could also be determined by a pleiotropic gene also involved in fitness variation in differing environments, or a result of genetic hitchhiking by colour-determining genes with neighbouring genes with some selective

General Introduction advantage (Barton, 2000). The isolation followed by the secondary recontact and introgression scenario outlined above could theoretically occur in the absence of any selective pressures acting upon colour traits, as a result of stochastic genetic processes and population dynamics.

A number of specific forms of selection, primarily disruptive selection (also referred to as local adaptation), apostatic selection and sexual selection, have been highlighted as mechanisms that are particularly likely to be involved in development and maintenance of bird plumage colour variation (Galeotti et al.,

2003).

Apostatic, or frequency-dependant selection, has often been suggested as a mechanism that could maintain colour polymorphisms in prey species (Clarke,

1969, Allen, 1974, Allen, 1976, Shigemiya, 2004). This type of selection may occur when the relative frequency of a morph in a population affects its fitness.

Apostatic selection could potentially influence bird colour plumage in two different ways: Rare colour variants of a species may be less recognisable to species on which they prey, and thus have greater hunting success (Paulson,

1973); whilst as prey, rare colour morphs of a bird may themselves be hunted at a frequency disproportionately lower than that morphs’ frequency within the population (Fitzpatrick et al., 2009).

Sexual selection has been one of the most intensively studied mechanisms of intraspecific plumage variation. Since Darwin (1871) first suggested that sexual selection via female mate choice for exotic male plumage variants was likely to be responsible for the brightly coloured male plumage variants of many bird

General Introduction species, many researchers have investigated the case for sexual selection of plumage traits (see Andersson and Iwasa, (1996)). Sexual selection can take many forms; however, the most relevant to plumage colour variation are direct mate choice and intrasexual social interactions involved in breeding and territory dominance.

Disruptive selection in a heterogeneous environment, also known more simply as local adaptation, is where different morphs are favoured in different subsets of the environment because of selection differences between these subsets, can eventually lead to a loss of polymorphism within the individual populations present in each environmental subset (Levens & MacArthur (1966), cited in

(Schmidt, 1998)). However, when gene flow between these environmental subsets occurs at an appropriate level, a balance between this directional selection within microhabitats and the gene flow between them, may help to maintain stable colour polymorphisms (Slatkin, 1987). This type of selection has been suggested to play a role in plumage colour in a few bird species

(Kallioinen et al., 1995, Greene et al., 2000, Hughes et al., 2002), and seems an especially relevant hypothesis for species that have clinal intermediate zones between colour phenotypes, which are unlikely to be the product of secondary recontact after allopatric divergence.

As more taxa are studied, it is now becoming more widely accepted that none of these processes exclusively determines such variability; rather a combination of these processes contributes differently to variability in different species.

General Introduction

The study of intraspecific plumage colour variation is important in a variety of different ways. Many species of animals and plants display such variation in colouration between individuals, and investigations of the mechanisms by which such traits evolve and are maintained, as well as their genetic basis and functional importance, will further our understanding of the demographic, ecological and evolutionary forces that have produced, and continue to shape all living organisms. Plumage colour traits are a fascinatingly diverse system; studies of diversity of colour and pattern in birds, with their huge amount of variation both within and between species, offers researchers the opportunity to directly study the relative importance of different drivers of evolutionary change across different environments. Research upon the genetic basis of such plumage traits will help further our understanding of the sometimes complex relationships between genotype and phenotype, the diverse pathways and processes involved in pigmentation, and potentially even the process of speciation itself.

1.3 STUDY SPECIES: THE AUSTRALIAN MAGPIE

1.3.1 Life history, variation and distribution

As this thesis is presented as a series of papers, detailed descriptions of magpie life history, distribution, subspecies and plumage forms, and phylogenetic histories are presented in the relevant chapters. For brevity, only basic summaries of these are included here as part of the general introduction.

The Australian magpie (Cracticus tibicen), hereafter referred to simply as

‘magpie’, is a medium-sized black, white and (sometimes) grey passerine bird

General Introduction

(family Artamidae), that primarily lives in sedentary territorial groups. Previously given their own genus, Gymnorhina (Schodde and Mason, 1999, Christidis and

Boles, 2008), the species has now been included in the Australian butcherbird genus, Cracticus, on the basis that morphological differences between the magpie and butcherbirds, and the adaptation to ground-living are not enough to designate the groups into two separate genera (Christidis and Boles, 2008).

Recent molecular work using both mitochondrial and nuclear markers has verified this re-classification, finding magpies are indeed a sister taxon to the black butcherbird (Cracticus quoyi), with the two species likely to have split during the Pliocene (3.0-5.8Ma) (Kearns et al., 2013).

The species’ distribution spans the majority of the Australian continent (Figure

1.1), but densities generally decrease towards the arid interior of the continent

(Schodde and Mason, 1999). They are also found across small regions of

Papua New Guinea (Black, 1986) and Tasmania (Schodde and Mason, 1999).

Humans have transplanted the species further afield to New Zealand

((Thomson, 1922), cited in (Veltman and Hickson, 1989)) and Fiji in attempts to help control pasture pests.

General Introduction

Figure 1.1: Distribution of Cracticus tibicen in Australia; adapted from Toon, 2007 and Schodde and Mason, 1999.

Open woodland, pastoral land and urban environments are preferred habitats, with low magpie densities, or absence, in very densely forested or highly arid regions. Magpies are highly territorial, and aggressively defend their territories against intruders (Carrick, 1972, Kallioinen et al., 1995). The number of birds within territories varies both geographically and between populations, ranging between two and 26 individuals (Robinson, 1956, Hughes and Mather, 1991,

Hughes et al., 1996, Baker et al., 2000), while mostly non-geographically defined (territory-less) and usually large flock groups have been observed in sub-optimal habitat areas surrounding territorial groups (Carrick, 1972, Durrant and Hughes, 2005).

Juvenile dispersal levels and timing vary widely between populations (Carrick,

1972, Baker et al., 2000, Hughes et al., 2003), and dispersal seems to be male- biased (Veltman and Carrick, 1990, Toon, 2007), a generally unusual trait within 8

General Introduction passerine birds (Greenwood, 1980). Magpies have a longer lifespan than the majority of passerine birds, and most magpies that survive the first few years have a lifespan of around 25 years (Kaplan, 2004). Cooperative breeding and helping behaviour have been observed across a number of populations

(Hughes et al., 1996, Finn and Hughes, 2001, Hughes et al., 2003). Extra- group and extra-pair paternity levels are high for a passerine species, and also vary significantly between populations; rates of extra-group paternity have been found to range between 26% and 82% in different populations (Hughes et al.,

2003, Durrant and Hughes, 2005, Hughes et al., 2011).

Magpies have been observed to make up a small portion of the diet of several visually-hunting birds of prey, including peregrine falcons (Carrick, 1972), wedge-tailed eagles (Leopole and Wolfe, 1970, Brooker and Ridpath, 1980) and the little eagle (Debus, 1984). However, magpies have also been observed to be a target of small mammals, such as foxes and feral cats (Carrick, 1972), as well as monitor lizards ((Bravery, 1970), cited in (Koboroff et al., 2013)). Eggs and nestlings may also be targeted by snakes, although this has not been recorded in the literature (Koboroff et al., 2013).

Eight different sub-species of magpie are delineated, based on wing, bill, and overall body size, as well as plumage variation (Figure 1.1) (Schodde and

Mason, 1999). Three broad plumage forms are recognised: black backed, white backed and the varied (sometimes called western) form (Figure 1.2).

General Introduction

Figure 1.2: Plumage forms of the Australian magpie (C. tibicen) and their sexual dimorphism.

The white backed form is only found in the south of the continent, the varied form is restricted to the far south-west of Australia, and the black back form occurs across most suitable habitat in the northern and middle latitudes of the continent (Figure 1.1) (Schodde and Mason, 1999). The nature of this back colour variation is discussed in depth in each chapter. Sexual dimorphism occurs in all populations, and once sexual maturity is reached, male white backs and black backs display bright white napes and the females grey-ish white napes. Varied back males closely resemble white back males, while varied females have back feathers that are individually black with white edging, giving a scalloped appearance to their back area (Figure 1.2).

Sub-species and plumage forms freely interbreed at overlaps in their distribution, and presumably produce the hybrid plumage forms observed in these areas (Burton and Martin, 1976, Hughes, 1982). These hybrid forms have back patches composed of intermediate proportions of black feathers vs. white/grey/mottled feathers. The hybrid forms are not discrete morphs, but rather the back colour proportions vary along a continuous spectrum (see

Figure 1.3). Magpie back colour has been observed to be temporally stable at the level of the individual once sexual maturity is attained, and has not been observed to vary plastically with diet, season or dominance status (Hughes,

General Introduction

1982, Hughes et al., 2001, Hughes, unpublished data). Considered together, these attributes indicate back colour is very likely a heritable trait in magpies.

Figure 1.3 Hybrid forms of black back and white back males in the eastern hybrid zone. White napes and back feathers are instead grey in females in black backed and white backed populations, and females are grey/ mottled in varied populations. This morphological variation is continuous in nature.

1.3.2 Phylogeography of magpies

An early study suggested the strong pattern of dark back colour in the north of the continent, and lighter back colour in the south, was likely a result of the classical story observed in many bird species: historical isolation leading to morphological divergence, followed by secondary recontact between colour forms at the observed hybrid zones (Burton and Martin, 1976). However, studies using molecular markers have shown the history of this trait to be especially interesting: mitochondrial variation indicates a strong divergence between eastern and western populations, instead of the north-south divergence expected under the allopatric divergence and secondary re-contact hypothesis (Baker et al., 2000, Hughes et al., 2001, Toon et al., 2003, Toon et al., 2007). Baker et al. (2000) also found moderate levels of gene flow among eastern populations, regardless of back colour.

General Introduction

Mitochondrial DNA data indicate that several different arid barriers seem likely to have restricted dispersal between eastern and western populations, leading to their divergence during the Pleistocene (approximately 36 000 years ago, according to molecular clock estimates), as well as suggesting contractions of populations to refugial areas of suitable climate and vegetation during this dry period (Toon et al., 2007). Molecular clock estimates place the divergence of

Tasmanian populations from eastern populations more recently, at about 16

000 years ago, when changes to sea levels resulted in disconnection of the land bridge that had linked Tasmania to the mainland during that glacial cycle

(Chappell and Shackleton, 1986, Hughes et al., 2001, Toon et al., 2007). More recently than the divergence between eastern and western populations, mtDNA indicate a possible expansion by eastern populations, tracking further inland and north (Toon et al., 2007).

Nuclear data, in the form of microsatellites, have revealed a less clear picture of this divergence between eastern and western populations, and support the idea that in the north of the continent, eastern and western populations seem likely to have had secondary recontact (Toon et al., 2007). This is likely to be a rare, but not unprecedented example of nuclear gene flow preceding mtDNA gene flow, thus enabling the detection of dispersal events with nuclear DNA before a clear signal is detectable in mtDNA data. Generally this is unlikely due to the fourfold smaller effective population size of mtDNA; however this can be observed when dispersal patterns are heavily male-biased, as has been observed in magpie populations. This expansion of eastern populations northwards and inland has been suggested as possibly related to the expansion of suitable magpie habitat since European settlement, through the creation and continual expansion of

General Introduction pastoral and urban landscapes in areas previously too arid, or otherwise unsuitable, for magpie settlement (Campbell, 1929, Schodde and Mason,

1999). The weaker support of nuclear DNA for the historical isolation of eastern and western populations (confounded somewhat unexpectedly by more recent gene flow events) indicates phylogeographic analyses of further neutral nuclear genes could be useful to further elucidate historical population processes in this species.

1.4 QUANTITATIVE GENETICS OF MAGPIE BACK COLOUR

1.4.1 Trait inheritance and pedigree analyses

The rapid pace of technological advances in genetic research since the cloning of the first pigmentation gene in the mid-1980’s (Shibahara et al., 1986) means that now experimental manipulation of the genome itself is possible. However, it still remains important to also study colour variation in natural populations, rather than just in laboratory-based studies, for a number of reasons. Firstly, laboratory manipulations may create colour phenotypes not able to survive outside laboratory conditions, and secondly, colour phenotypes created in the laboratory may correspond to phenotypes that in the wild would be a product of the combination and/or interaction of several different genetic mutations (Protas and Patel, 2008).

Investigations of the nature of the genetic inheritance and heritability of plumage variation within wild populations of bird species still extensively rely upon pedigree-type data and relatedness to draw inferences about the trait heritability, the number of genes involved, types of dominance relationships

General Introduction between genotypes and the potential for sex-linkage of these genotypes

(Charmantier et al., 2014). Experimentally manipulating crosses in a wild, free- living bird species is often difficult and/or impossible, and essentially makes the subject no longer free-living. In such cases, genetic relationships within family groups can be used together with their plumage phenotypes, and this pedigree used to make some of the same inferences that could be experimentally determined using controlled breeding experiments and crosses in a laboratory- based setting.

Extra-pair and extra-group paternity (EPP & EGP) can lead to significant underestimations of the heritability of a trait of interest; and when these

EPP/EGP are high (e.g. > 25%), heritability estimates can be off by up to 40%

(Charmantier and Reale, 2005). However, when both EPP rates and actual heritability of the trait itself is low, Firth et al. (2015) suggest the use of social pedigrees alone may be adequate to estimate heritability. The heritability of melanic plumage traits is generally relatively high (Roulin and Ducrest, 2013), and the magpie is known to have high-to-very-high levels of EPP/EGP (Hughes et al., 2003, Durrant and Hughes, 2005, Hughes et al., 2011). Thus social parentage-based methods are unlikely to yield accurate estimates of a melanic plumage trait, such as the back colour variation observed within this species, and genetic parentage analyses must first be carried out to ensure heritability estimates are robust.

General Introduction

1.5 INVESTIGATING THE GENETIC BASIS OF

MORPHOLOGICAL TRAITS

Observational studies of distribution and morphological patterns, field and laboratory manipulations looking at the influence of environmental factors, and pedigrees and artificial cross-breeding experiments have all contributed a great deal to the current understanding of the nature of the development and/or inheritance of morphological traits. Over the last few decades, scientists have been able to approach the problem of intraspecific variation more directly, and huge advances in DNA technologies and techniques have enabled closer examination of the relationships between phenotype and genotype (Doerge,

2002).

Several different approaches can be taken when investigating the genotypes that determine a particular phenotype. However, the choice of method is crucially dependent upon the extent to which the species and trait of interest have been previously studied, as well as the nature of the trait itself and the resources available to researchers.

Studies of model organisms, such as mice, fruit flies, and chickens, as well as extensive medical research on human phenotypes (e.g. diseases) have led to the identification of a range of genes and/or sections of genomes involved in a range of traits e.g. beach mouse fur colour variation (Hoekstra, 2006), extended black colour in chickens (Kerje et al., 2003), and oculocutaneous albinism in humans (Oetting and King, 1993). Such studies have used a combination of methods, including Quantitative trait locus mapping (QTL) and microarray assessment of gene expression to pinpoint specific gene regions involved in

General Introduction phenotypic traits in such species. Such methods are costly, and some require specialised equipment currently still beyond the financial and technical reach of many molecular laboratories. A few well-studied non-model bird species have relatively recently benefited from the application of these methods e.g. QTL mapping for clutch size in the collared flycatcher (Husby et al., 2015), and microarray expression analysis focussing on migratory behaviour in the willow warbler (Boss et al., 2016). However, largely, it is still the characterisation of important genes for specific traits in model species which has provided a template enabling researchers to examine the genetic basis of functionally important traits in non-model species in the wild (Vasemagi and Primmer,

2005).

1.5.1 Using candidate colour genes to examine the genetic basis of magpie plumage variation

Candidate genes that have been identified as being involved in phenotypic variation within model species, can then be screened for variations associated with the relevant similar phenotypic variation in non-model species, an approach which in many cases requires substantially less fiscal investment than starting from scratch with microarrays and large-scale linkage mapping projects, which can impose substantial financial burdens (Wheelan et al., 2008).

The candidate gene approach has been successfully utilised to identify associations between phenotype and particular genotypes across a wide range of species and traits (Hoekstra and Coyne, 2007). Although this approach is more often used to investigate simple Mendelian traits thought to be determined

General Introduction by a single gene, studies of complex, multi-gene based traits may also benefit by using this approach to investigate the possible underlying pathways that link genetic variants to complex traits (Tabor et al., 2002).

Although a great deal of research has established clear associations between specific genotypes and phenotypic variants within species, a new field known as evolutionary developmental biology (evo-devo) has emerged in the last few decades that challenges a number of traditional evolutionary theories (Hoekstra and Coyne, 2007). The most controversial of these is the long-held idea that structural changes in coding regions, in the form of mutations, are the primary source of phenotypic variation upon which natural selection may act and thus lead to evolutionary change. Evo-devo advocates emphasise the importance of alterations in gene regulation, rather than structure, as the basis of evolutionarily important changes (Carroll, 2000).

On balance it seems likely that both structural changes in coding regions and the effects of cis-regulatory and transcription factors are involved in the evolution and expression of morphological variation within species (Hoekstra and Coyne, 2007). Whilst the regulatory aspect is deserving of a great deal of research investment over the coming decades, it seems unwise to altogether abandon the traditional pursuit of structural genetic variants to explain morphological variation, when such methods have already revealed a great deal about the underlying mechanisms that determine a large number of physiological and morphological traits. This is primarily because regulatory cis- and trans- regulatory element sequence regions for candidate colour genes have only recently been investigated in any detail in model species, and it

General Introduction appears likely wild species will benefit from such research, although the utility of doing so is currently hampered by practical considerations, such as a lack of published primer sets easily transferrable between species.

Pigmentation-related candidate genes are numerous; many hundreds of colour genes have already been described for model species such as mice and chickens, thus selecting which to target depends on the nature of colour variation being investigated. The plumage forms of the magpie vary in their distribution and absence of black pigment, therefore, genes involved in melanin production, distribution and regulation pathways in other model species (Ito and

Wakamatsu, 2011) were selected to screen for variation that may explain the genetic basis of magpie plumage colour differences.

1.5.1.1 The melanocortin-1 receptor (MC1R)

The melanocortin-1 receptor (MC1R) has been one of the most utilised and scrutinised candidate colour genes for melanic-based pigmentation variation, for several different reasons. A large number of early candidate gene studies found positive associations between MC1R variation and pigmentation variation across a wide range of animal species, including the horse (Marklund et al.,

1996), fox (Vage et al., 1997), arctic fox (Vage et al., 2005), pig (Kijas et al.,

1998), sheep (Vage et al., 1999), dog (Newton et al., 2000), black bear (Ritland et al., 2001), cattle (Klungland et al., 1995), jaguar and jaguarundis (Eizirik et al., 2003), pocket mouse (Nachman et al., 2003), domestic rabbit (Fontanesi et al., 2006), humans (Valverde et al., 1995), lesser earless lizard, and the little striped whiptail (Rosenblum et al., 2004). In particular, a number of well-studied avian taxa, many with subjectively striking plumage variants that were found to

General Introduction be associated with mutations in this gene, including chickens, Japanese quails, red-footed boobies, white-winged fairy-wrens, lesser snow geese, arctic skuas and bananaquits (Takeuchi et al., 1996, Theron et al., 2001, Andersson, 2003,

Doucet et al., 2004, Mundy et al., 2004, Nadeau et al., 2006, Baiao et al., 2007) brought the MC1R gene into focus as an important ‘must-screen’ candidate pigmentation gene.

The MC1R gene plays an important role in the initiation of the melanin- production process and sometimes additionally, the dispersal of melanosomes through cells (Jackson, 1997). It is expressed in specialised pigmentation cells, melanocytes, where it acts as a receptor, and is located in a transmembrane region (Mountjoy et al., 1992). When the receptor binds with hormones, such as variants of the melanocyte-stimulating hormone (MSH), MC1R is ‘activated’, resulting in production of black and/or brown eumelanin instead of ‘default’ production of red or yellow pheomelanin. The receptor can also be antagonised by peptide-signalling molecules such as the agouti signalling peptide (ASIP), causing it to revert to production of red or yellow pheomelanin despite stimulation by hormones. Mutations in the MC1R coding region can cause constant, or even excessive production of dark eumelanins, even when the receptor is not being stimulated by hormones (gain-of-function mutations), or conversely reduce or eliminate production of dark eumelanins (loss-of-function mutations), even when the appropriate melanocortin hormones are binding with the receptor (Robbins et al., 1993).

General Introduction

1.5.1.2 The tyrosinase gene (TYR)

The TYR gene encodes a melanogenic enzyme known to play an important role in pigmentation pathways. Binding of melanocortin hormones to the MC1R receptors indirectly leads to activation of the TYR enzyme, and it is this enzyme that then initiates production of eumelanin pigments; thus tyrosinase activity levels determine whether eumelanins or pheomelanins are produced within the melanocyte at a given point in time (Burchill et al., 1986).

Pigmentation variation across a wide range of animal species has been associated with mutations in the TYR gene, including albinism, hyperpigmentation and melanomas in humans (Oetting and King, 1993, Spritz and Hearing Jr, 1994), Burmese and Siamese cat breeds (Schmidt-Kuntzel et al., 2005), albino and Himalayan mice (Kwon et al., 1989, Yokoyama et al.,

1990), albino cattle (Schmutz et al., 2004), and albino and recessive white chickens (Tobita-Teramoto et al., 2000, Chang et al., 2006).

1.5.1.3 Dopachrome tautomerase (DCT)

DCT plays several differing and distinct roles in pigmentation pathways. It helps regulate levels of different melanin pigments, and in birds, the protein functions as an enzyme (April et al., 1998) which stabilises tyrosinase, increasing overall

TYR activity, and indirectly increasing the production of darker eumelanin pigments (Manga et al., 2000). Additionally, the melanin production pathway produces a number of different toxic by-products which must be removed for the continued functioning of normal pigmentation reactions and processes, and the

DCT enzyme plays an essential role in metabolising some of these by-products

(Jackson et al., 1992, Costin et al., 2005).

General Introduction

Mutations in the DCT gene, also known as the tyrosine-related protein 2

(TRP2), have been associated with a handful of pigmentation variations in animals. These include three different types of slaty mice (Jackson et al., 1992,

Kroumpouzos et al., 1994, Budd and Jackson, 1995, Guyonneau et al., 2004), albino rainbow trout (Boonanuntanasarn et al., 2004), blue iris colour in humans

(Frudakis et al., 2003), and possibly dominant and recessive white chickens

(April et al., 1998).

1.5.1.4 The tyrosinase-related protein 1 (TYRP1 or TRP1)

In birds, the TYRP1 gene occurs on the Z chromosome and is therefore sex- linked (Backström et al., 2010). The enzyme produced by this gene, like DCT, is thought to play an important role in the stabilisation of tyrosinase (Kobayashi and Hearing, 2007).

Pigmentation variants have been associated with mutations in the TYRP1 gene in a wide range of mammal species, and a few bird species. In mammals these include: brown oculocutaneous albinism (OAC3) (Boissy et al., 1996) and rufous albinism in humans (Manga et al., 1997), brown (Zdarsky et al., 1990) and light and white-based brown mice (Johnson and Jackson, 1992, Javerzat and Jackson, 1998), brown Dexter cattle (Berryere et al., 2003), brown Soay sheep (Gratten et al., 2007), brown Chinese pigs (Ren et al., 2011), cinnamon and chocolate cats (Schmidt-Kuntzel et al., 2005), and brown coats, nose and footpads in dogs (Schmutz et al., 2002). Within birds, the roux Japanese quail

(Nadeau et al., 2007), and the brown pied flycatcher (Buggiotti, 2007) have

General Introduction been linked to a mutation and a splicing error (respectively) within the TYRP1 gene.

1.5.1.5 The agouti-related protein (AGRP)

The agouti-related protein (AGRP) acts as an antagonist of MC3R and MC4R receptors, and in mammals it is almost exclusively expressed in the brain, where it is thought to play a significant role in feeding behaviour and weight regulation (Ollmann et al., 1997). However, in chickens, the protein is also expressed in other body tissues, including skin (Takeuchi et al., 2000).

Pigmentation studies of two species of willow grouse and two species of flycatcher found no mutations in the fragments of AGRP screened explained colour differences in these species (Buggiotti, 2007, Skoglund and Hoglund,

2010). However, a gene expression study by Li et al. (2011) has suggested the gene may be involved in the hyper pigmentation of a chicken breed, the silky fowl.

1.6 SELECTION AND MAGPIE PLUMAGE VARIATION

The dichotomy between plumage pattern and mitochondrial genetic structure in the magpie is striking and supports a primary intergradation hypothesis for the origin of the plumage hybrid zones in both the east and west of the continent.

As differentiation into dark plumage forms in the north of Australia, and lighter plumage forms in the south is likely to have occurred in situ, and at least twice independently in both the east and west of the continent, it seems reasonable to assume that an adaptive model, with some differential selection on plumage

General Introduction may be responsible for the creation of different plumage forms and most likely for the maintenance of plumage hybrid zones (Toon et al., 2003).

Hughes et al. (2001) proposed that different plumage forms may be favoured by differing selective pressures in differing habitats. Similar to the mechanism first outlined in Kallioinen et al. (1995), the authors suggested black backed birds, and thus their nests and offspring, may be less conspicuous to predators than white backed birds in the more open vegetation inhabited by magpies in the north. Whilst in the more heavily wooded south of the continent, ‘brighter’ white back males may be favoured by sexual selection once released from constraints related to conspicuousness/predation by the higher levels of canopy cover (Hughes et al., 2001). Together these local selective pressures could theoretically produce a stable hybrid zone, and the distribution of magpie plumage forms observed in both the east and west of Australia, despite ongoing gene flow, as both hybrid zones straddle clines between temperate, more heavily forested southern regions, and the more sparsely vegetated (even semi- arid/arid) northern regions (Hughes et al., 2001, Toon et al., 2003, Kallioinen et al., 1995). White backed birds observed well outside their normal distribution have also been observed exclusively in areas with dense forest and high canopy cover (Kallioinen et al., 1995).

Greater levels of aggression have been observed towards white backed intruder males than black backed males, by resident birds in both black backed and hybrid zone populations (Kallioinen et al., 1995); as would be expected if white backed males were favoured by sexual selection, and thus posed a greater threat to resident males. This aggression from conspecifics could also translate

General Introduction to white backed males being intrinsically better able to defend their territories and their social standing within territories against intruding and competing males, and thus preserve their ability to retain social ownership of, and breed in that territory.

‘Brighter’ plumage morphs may be favoured in males, if this brightness is used by females as a proxy for quality in making mate choice decisions, in order to maximise the potential fitness of her offspring and ensure access to adequate resources to successfully produce and rear said offspring (Hill, 1991). This may occur because plumage brightness (honestly or not) advertises good condition of the males’ qualities, such as low internal and external parasite loads

(Hamilton and Zuk, 1982), (e.g. satin bowerbirds (Doucet and Montgomerie,

2003)), youthfulness, high contribution to parental care (Linville et al., 1998), access to adequate nutrition, or the ability to defend good quality territory.

Alternatively, and possibly additionally, if brightness of males is perceived as directly indicating a fundamentally higher fitness and quality of genetic material

(‘good genes’ hypothesis), female mate choice may lead to sexual selection of brighter males. Within the eastern Australian magpie hybrid zone, Hughes et al.

(2011) found no evidence to support female preference for brighter males, and demonstrated that, in this population, no particular back colour of males were statistically more likely to be cuckolded by their female social partner.

White backed magpies appear more conspicuous than black backed forms to human observers, an effect which may be even more pronounced in bird-bird interactions, as many birds have at least four retinal cones and can see in the ultra-violet (UV) spectrum (Vorobyev et al., 1998). Under UV light, white

General Introduction magpie feathers have been observed to glow more brightly than either black or grey magpie feathers to a human observer (Durrant, 1998).

Generally there are two distinct types of colour vision in birds: violet sensitive

(VS) and ultra-violet sensitive (UVS). These differ in the sensitivity of their most shortwave cones, SWS1; while VS species have a wavelength maximum sensitivity of approximately 402-426 nm, UVS species see more of the ultraviolet spectrum (355-380 nm), seeing a broader range of wavelengths of light, and hence colour across all their visual cones (Hart and Hunt, 2007). Only one amino acid substitution (in the cone opsin protein) seems to differentiate these UVS and VS vision systems (Ödeen et al., 2009), and within passerine birds, for which it is unclear if the ancestral state was UVS or VS, the mapping of these amino acid states onto consensus phylogenetic trees indicate the colour vision system seems to have changed in both directions, a significant number of times as the group radiated (Ödeen et al., 2011).

Hence it is unclear if the Australian magpie is more likely to possess a VS or

UVS vision system, as the frequent state shifts in closely related taxa mean either VS or UVS systems may be equally plausible, and the relevant amino acid remains to be assayed in the species, and preferably should be examined also within each plumage form to rule out the possibility of intraspecific variation in magpie vision systems.

Whether more brightly coloured white backed magpies appear significantly more conspicuous to potential predators than black backed forms, when viewed by potential predators, with regard to the UV spectrum is also somewhat

General Introduction unclear. Within the last 20 years, it has been discovered that birds of prey possess VS rather than UVS vision systems (Ödeen and Håstad, 2003). The pigmented oil droplet that alters the sensitivity of the UVS/VS cone and changes the transmittance of light wavelengths via filtration of light wavelengths may further reduce short-wavelength UV perception in birds of prey (Lind et al.,

2013, Lind et al., 2014). It has been suggested by a number of authors that the

VS vision systems and low ocular media transmission of UV light found in birds of prey, contrasted with the UVS vision systems present in many passerine birds, may mean that many passerine species are effectively accessing a

‘secret’ channel of signalling and communication unbeknownst to their primary predators (birds of prey), by displaying plumage colours outside of the range able to be perceived by predators, but which may serve as sexual or social signals in intraspecies interactions (Håstad et al., 2005, Lind et al., 2013).

Thus the cost of such signalling in terms of increased predation risk to an individual or its offspring may be reduced (or removed) if the wavelength of the plumage colour used in such signalling is short enough to go undetected by visual predators.

Studies of magpies looking for evidence to support the hypothesis that different plumage forms may be favoured by differing selective pressures in differing habitats have found no evidence of sexual selection favouring white backed males, some evidence of differential reproductive success of back colour forms in differing habitats, and no evidence of asymmetrical mating patterns between different back colours that could maintain the hybrid zone in the absence of selective pressures (Hughes et al., 2002, Hughes et al., 2011).

General Introduction

In light of this apparent lack of assortative mating and sexual selection of white back colour forms, Hughes et al. (2011) suggest different back colours could vary in bacterial resistance due to differences between black and white feathers. The distribution of magpie back colour forms does conform to

Gloger’s rule, with darker forms found in locations of higher humidity closer to the equator, and lighter forms further from the equator (Zink and Remsen Jr,

1986). Alternatively, different colours of feathers may have different inherent

‘costs’ of production (Tickell, 2003), with the more costly black feathers only expressed in magpies in environments where such costs are outweighed by benefits, e.g. increased fitness. Different colour forms may vary in their intrinsic ability to efficiently thermoregulate, or their resistance to oxidative stress

(Galeotti et al., 2003, Roulin et al., 2011, Hung and Li, 2015).

Moreover, if genes involved in magpie back colour variation have pleiotropic effects, back colour distributions may not reflect selection acting directly on the trait of back colour itself, but on the other traits these gene(s) act upon. Several examples of pigmentation variation associated with physiological traits have been documented, and include metabolism, vision, immune function, and resistance to parasites (Roulin et al., 2001, Dreiss et al., 2010, Saino et al.,

2013a, Reissmann and Ludwig, 2013).

General Introduction

1.7 RESEARCH QUESTIONS

1.7.1 Heritability of plumage colour in magpies

In the eastern hybrid zone between black back and white back plumage forms, a study site has been established in which magpie individuals from several hundred territorial groups have been DNA sampled, colour banded and observed regularly over the last three decades. Using samples from a large number of territories at this site, this study aims to generate pedigree-like data using parentage analyses. This will then be used to evaluate several aspects of the quantitative genetics underlying the variation in magpie back colour. At this study site, the entire range of eastern back colours are present and are known to interbreed. The following hypotheses were investigated:

- Back colour variation is strongly additively heritable in magpies.

- Maternal or paternal effects will influence back colour variation.

- There will be no or weak evidence for either positive or negative

assortative mating patterns between back colours.

- A small number of genes are likely to best explain back colour variation

in magpies.

1.7.2 Magpie phylogeography

Different types of genes with differing demographic and evolutionary histories, such as maternal mtDNA, nuclear non-coding regions, and functional gene coding regions are likely to exhibit different patterns of genetic variation and structure. In the MC1R gene, which plays an important role in pigmentation and particularly melanogenesis, and is potentially under selection, such variation

General Introduction could reveal different phylogeographic signal and structure than that already found using mtDNA and putatively selectively neutral nuclear microsatellites repeat markers. This study aims to characterise the MC1R gene in the

Australian magpie, as well as investigate the following hypotheses:

If MC1R is under selection in the magpie-

- Patterns of genetic variation in the MC1R gene will exhibit different

structure than putatively neutral markers, with genetic variation

partitioned between plumage forms, rather than between broad eastern

and western clades previously indicated by the neutral gene markers.

If MC1R is not under selection in the magpie-

- Phylogeographic structure in MC1R will add support to the theory that

populations have recently undergone expansions in this species.

1.7.3 Candidate colour genes and magpie plumage colour

Using a number of different candidate genes for pigmentation, this study aims to sequence and screen relevant coding regions and SNPs in these genes that have been previously linked to pigmentation variation in other animal species.

These will be tested for associations with magpie back colour variation, as well as checked for possible interactions between any variable sites (mutations) found that may contribute to back colour phenotype in the Australian magpie.

In order to do this, the following hypotheses were tested:

General Introduction

- Genetic variation and/or interactions between genetic variants in one or

several of the five candidate colour genes screened will be associated

with back colour plumage variation in the magpie, and differentiate

magpie individuals of different back colours.

- If so, these mutations are likely to be similar, or the same (in terms of

location in the gene or the functional effect) as mutations found to be

associated with colour variation in other animal taxa.

1.8 ORGANISATION AND STRUCTURE OF THIS THESIS

This thesis is presented as a series of papers, bookended by a general introduction and a general discussion linking these studies together, and demonstrating how they collectively advance our current understanding of intraspecific plumage colour variation in the Australian magpie. Therefore, there will be some unavoidable repetition between chapters, as data chapters are designed to stand alone, as well as being part of a whole. These papers and their respective experiments were not necessarily conducted in the order in which they are presented. One of these data chapters (Chapter 3: MC1R and phylogeography) has been published in the Journal of Heredity prior to the submission of this thesis (and is also attached in the appendix in this format), and the other two data chapters shall shortly be submitted for review.

30 Heritability

CHAPTER 2: HERITABILITY AND PATTERNS OF INHERITANCE OF PLUMAGE VARIATION IN THE AUSTRALIAN MAGPIE, CRACTICUS TIBICEN.

2.1 INTRODUCTION

Many species exhibit some form of morphological polymorphism, and colour polymorphic species in particular offer biologists the chance to directly study the

‘engine room’ of evolution. Such species often present a ‘ready-made experimental design’ which can be exploited to investigate the functional and/or evolutionary significance of such variation, often yielding significant insights, theories and principles that can then be extended to a macro-evolutionary scale

(e.g. Cain and Sheppard, 1954, Clarke, 1969).

The mechanisms most generally invoked to explain how such polymorphisms evolve and are maintained can be widely grouped into neutral and selective in nature. Neutral mechanisms that have been suggested to at least partially explain some colour polymorphisms, include past historical processes such as bottlenecks and barriers to gene flow, allopatric isolation with or without secondary recontact, as well as a balance between mutation and genetic drift

(Kimura, 1984, Cook, 1992, Marshall and Ritland, 2002, Roulin, 2004, Hoffman et al., 2006). Genetic processes may also play a role, via genetic hitchhiking of pigmentation genes or even pleiotropic effects of other genes linked to fitness

(Barton, 2000). Demonstrating conclusively that such neutral processes are not the agent responsible for geographically-delineated polymorphisms is, however, difficult, but can be done utilising approximate Bayesian computation (ABC) simulations (i.e. Antoniazza et al., 2014). However, such analyses require a

Heritability large number of putatively neutral genetic markers to be genotyped across a relatively large number of individuals distributed throughout the geographic region of interest.

A great deal of colour polymorphism in birds has been linked to selective mechanisms, such as apostatic selection, sexual selection and disruptive selection/ local adaptation (Galeotti et al., 2003). Birds are particularly polymorphic: approximately 3.5% of extant bird species exhibit some form of intraspecific plumage variation, and phylogenetic analyses indicate that this trait of colour polymorphism seems to have evolved again and again throughout the history and radiation of avian taxa (Galeotti et al., 2003)

A large body of evidence has been building to support the idea that many different forms of selection, together with non-selective processes such as allopatric isolation are largely responsible for generating and maintaining a great deal of the intraspecific colour variation observed in birds (Roulin, 2004).

Genetic theory, and now empirical evidence (Hugall and Stuart-Fox, 2012) suggest such colour polymorphism could accelerate speciation, and morphs may represent an early or intermediate stage in this process, through a variety of mechanisms, including pre zygotic and post zygotic isolation. Indeed, a study of vertebrate species by Cattin et al. (2016) suggests that their findings of a significantly greater ‘age’ of polymorphic species, compared to monomorphic species, may indicate species with such intraspecific colour variation have an intrinsically higher capacity for long-term resilience to environmental changes.

Alternatively, this greater ‘age’ of polymorphic species may simply reflect that their persistence over long time scales has enabled more random mutation(s) to

Heritability occur throughout the history of such species, and some mutations may have resulted in colour polymorphism. Thus studies of this variation, its genetic basis and architecture represent important steps towards furthering our understanding of the evolutionary forces that shape life on earth.

Plumage colours are a result of pigmentation, refraction of light by structural arrangements within feathers, or a combination of both pigment and structure.

Structural colour is often seen in the form of iridescent (but not always) blues, greens, yellows and sometimes reds. This form of colour is not restricted to

Aves, and is also found across many Lepidopteran and Cephalopod species.

Melanins, carotinoids and porphyrines are the primary pigment types found in birds. Carotinoids are responsible for a great deal of the red, yellow and orange plumage found across bird species, and are obtained primarily via diet; thus carotinoid-based colour can often reflect condition and be plastic in expression

(McGraw, 2006, Price, 2006). Porphyrines are rarer, and are modified amino acids that fluoresce when exposed to UV light, primarily found in downy feathers of some owl and bustard species. Melanins account for a great deal of colouration in Aves, and are responsible for most black/brown/red-brown/yellow and grey colour variation (Fox, 1976). Production of melanins in animals is thought to be predominantly genetically determined and highly heritable

(Buckley, 1987, Hearing and Tsukamoto, 1991, Majerus, 1998, Roulin and

Ducrest, 2013), and these pigments can be expressed in the feathers, skin, beaks and claws of birds. White pigmentation in birds is most commonly a product of a lack of pigment and/or feather structure, and white feathers are generally weaker and more vulnerable to bacterial and pathogenic infection

Heritability than pigmented feathers (Burtt, 1986, Bonser, 1995, Mackintosh, 2001, Burtt Jr and Ichida, 2004), but not necessarily lice infestation (Bush et al., 2006).

Prior to the technological advances enabling affordable and direct analysis of

DNA, the genetics underlying pigmentation- such as heritability, dominance and the putative loci involved- were studied primarily through the development of pedigrees and breeding studies involving specific crosses over a number of generations (Hoekstra, 2006b). Such studies remain a powerful way to dissect the genetics underlying traits or phenotypes, especially as part of a multi- disciplinary approach, utilising candidate gene analyses, genome-wide association studies (GWAS), quantitative trait locus (QTL) mapping, or microarray expression analyses (Hoekstra, 2006b, Piertney and Webster,

2010).

An interesting case of intraspecific plumage variation is found in the Australian magpie (Cracticus tibicen), a medium sized passerine that lives in permanent territorial groups. Although eight sub-species are currently recognized based on small differences in morphological characters (e.g. wingspan) (Schodde and

Mason, 1999), three main plumage forms are found across the Australian continent. Of these, two are referred to as white backed, and one black backed

(Figures 2.1 and 2.2), with all forms displaying sexually dimorphism (Schodde and Mason, 1999).

Black backed forms that inhabit the northern parts of the continent have white napes, shoulders and rumps, and a white tail terminated by a black band, whilst all other feathered body parts (including their backs) are black. In south-eastern

Heritability white backed forms, the back is white instead of black, joining the white nape to the white rump and giving an appearance of an entirely white dorsal surface when bird is flying and viewed from above. The south-western white backed plumage form, also referred to as the ‘varied’ form, has males with the same plumage as white backed south-eastern forms, but females have a shorter white nape that joins a back composed of black feathers that are all edged with white, giving their backs a ‘scalloped’ or ‘mottled’ appearance, as well as a white rump and tail that is also terminated by a black band (for illustrations of plumage see Figure 2.1). The distribution of these plumage forms and subspecies across the Australian continent is presented in Figure 2.2.

The sexual dimorphism found in this species manifests as body parts (i.e. nape, shoulder, back and rump) that are white in males appearing pale grey or slate grey in females (with the exception of the backs of western white ‘varied’ females) (Figure 2.1). Sub-adults also appear grey in these white body parts, and severe infestations of lice can make the normally brilliantly white nape of males appear superficially grey at times (J. Hughes, personal communication).

Figure 2.1: The three main back colour variants of Cracticus tibicen.

Genetic inheritance of back colour has been proposed based on theory and anecdotal observation, but is not yet supported by empirical data (Hughes,

1982, Hughes et al., 2001), However it is clear that back colour does not

Heritability change temporally within individual birds once adult plumage is attained, and observational data rule out plastic, seasonal or dietary back-colour variation in a long-term study site in the eastern Australian hybrid zone (J. Hughes, unpublished data).

Where distributions of the different plumage forms overlap, interbreeding is unrestricted (Burton and Martin, 1976, Hughes, 1982, Hughes et al., 2011) and produces a range of different intermediate back colour forms; see Figure 2.3 for illustration of intermediate plumage forms in the eastern hybrid zone. These zones of intergradation, hereafter referred to as hybrid zones, differ in their width, and the western hybrid zone between white backs and black backs is wider (~500 km) and further north (Schodde and Mason, 1999) than the eastern hybrid zone between black backs and white backs, which is estimated to be 200 km wide (Burton and Martin, 1976), (see Figure 2.2).

Figure 2.2: Distribution of C. tibicen and sub-species in Australia, and location of the long-term study site; adapted from Schodde and Mason (1999) and Toon (2007).

Heritability

Magpies occur across the majority of the Australian continent (Schodde and

Mason, 1999) and inhabit open woodland and urban environments, and are uncommon or even absent in dense forest or extremely arid regions (e.g. the arid interior of the continent) (Figure 2.2). Magpies defend their territories vigorously, not only during breeding periods, but also throughout the year

(Carrick, 1972). Magpies from other territories or flocks trespassing in, or entering another territory are attacked aggressively by resident adults

(Kallioinen, et al., 1995).

Territory size varies between populations, from pairs in the northeast (Hughes et al., 1996), to up to 26 birds in the southwest (Robinson, 1956, cited in Baker et al., 2000). Non-territorial flocks have been described within several populations (Carrick, 1972, Durrant and Hughes, 2005), and are thought to be composed of young birds, dispersed from their natal territories but yet to establish a new territory or join an existing territory (Carrick 1972, Veltman,

1989 in (Veltman and Hickson, 1989)).

Different populations and sub-species have different patterns and timing of juvenile dispersal (Carrick, 1972, Baker et al., 2000, Hughes et al., 2003).

Genetic and observational data also indicate that dispersal is male-biased

(Veltman and Carrick, 1990, Toon, 2007), a trait not generally associated with passerine birds (Greenwood, 1980). Evidence of cooperative breeding and helping behaviour has been documented in a number of populations of magpies across Australia (Hughes et al., 1996, Finn and Hughes, 2001, Hughes et al.,

2003). Rates of extra-group paternity found in a number of magpie populations are high in comparison to the majority of passerine species (Hughes et al.,

37 Heritability

2003). The level of extra-group paternity seems to be highly variable between different populations, ranging from 26% - 88% (Durrant and Hughes, 2005,

Hughes et al., 2011).

Genetic variation in the mtDNA control region of magpies has been found to be partitioned into distinct eastern and western groups, separated by at least four mutations (Hughes et al., 2001, Toon et al., 2003), in direct contrast to variation in back colour forms, which is partitioned into northern (black backed) and southern (white backed) groupings. To explain this apparent dichotomy,

Hughes et al. (2001) have suggested a mechanism, similar to that first outlined in Kallioinen et al. (1995), in which natural selection favours black backs, while sexual selection favours the more conspicuous white backs.

Within one of the hybrid zones of C. tibicen’s distribution, where black-backed

(BB) and white-backed (WB) sub-species intergrade into a full range of intermediate back-colour forms, a large population outside the town of Seymour in Victoria has been the subject of a long-term study. At this study site, a mix of

BB, WB and intermediate plumage forms coexist and interbreed (Hughes et al.,

2002). Magpie territories at this site have been progressively observed, banded and had DNA samples collected since the early 1990’s onward (Finn and

Hughes, 2001).

The site is composed of several pastoral properties, with scattered tall trees and some small stands of uncleared or replanted immature native trees. As noted in

Hughes et al. (2002), magpies at this site breed cooperatively, with a large proportion of young in most territories remaining in their natal territory helping to

Heritability raise subsequent broods. Others have been observed joining non-territorial flocks in the area, and some flock members have been observed joining territories (usually when no birds of their sex were left in that territory), as well as attempting to start new territories (Hughes et al., 2002). Within each territory more than one female will, when possible, attempt to nest each breeding season; often this is done in order of social hierarchy, with the dominant female nesting first (Hughes et al., 2002); multiple mothers from the same territory have been shown to successfully produce fledglings in the same breeding season

(Carrick, 1972, Hughes et al., 2002, Hughes et al., 2003). The average number of individuals belonging to a territory, at five to seven individuals per territory (Baker et al., 2000; Hughes et al., 1996) is much larger than most other eastern populations sampled, although not as large as populations in the southwest region (Baker et al., 2000). Territories are strongly defended in a cooperative manner, with most adults actively defending their territory all year

(Hughes et al., 1996), and a previous study by Hughes et al. (2011), has estimated an extra-pair paternity rate of 36%, and extra-group paternity rate of

26% in this Seymour population.

The intermediate hybrid zones within C. tibicen’s distribution, such as the one within which the Seymour study site is located, present an ideal setting in which to develop a model of back-colour inheritance patterns within this species.

Whilst a large number of colour trait inheritance studies in animals have utilised researcher-controlled breeding and specific crosses (Majumdar et al., 1997,

Yurtsever, 2000, Luttikhuizen and Drent, 2008, Renieri et al., 2008), it is not always possible or practical to conduct captive or controlled breeding experiments. However, the Seymour study described earlier comes close to the

Heritability best feasible design for the study of trait inheritance in a wholly wild bird population.

Analysis of parent-offspring inheritance patterns should help to present a more definitive picture of the number of loci involved in back-colour variation, and it seems reasonable to hypothesise that only a small number of loci may be directly implicated, as Hughes (1982) posited, given that black-back genes seem to be dominant over white-back genes (Hughes and Mather, unpublished data, in Hughes et al., 2001). In the study population of magpies at Seymour, back colour varies continuously between black backs and white backs. While traditionally continuous traits have been thought to be a product of a large number of genes with small effect, this has not held up, and it seems just as likely such traits can be determined by only a few genes via differential gene regulation and/or expression, codominance, or epistatic interactions (Buckley,

1987).

Roulin (2004) argues that the designation of bird species’ colour variation, as either discrete or continuous, is often inconsistent, highlighting several species in which colour morphs previously identified as discrete (putatively due to convenience and human capabilities in a field situation) are in actuality continuous when examined at an appropriate scale (e.g. spectrophotometer analysis). Additionally, a large number of studies have found that variation at a single or relatively modest number of genes seem to determine colour phenotypes in a range of animals, including horses, cats, chickens and

Japanese quails, arctic skua and lesser snow geese, red-footed boobies, fairy wrens and bananaquits (Brunberg et al., 2006); (Schmidt-Kuntzel et al., 2005);

Heritability

(Kerje et al., 2004, Gunnarsson et al., 2007, Hiragaki et al., 2008, Nadeau et al.,

2008); (Mundy et al., 2004); (Baiao et al., 2007); (Doucet et al., 2004); (Theron et al., 2001).

This study will attempt to answer a number of questions with regard to back- colour inheritance in magpies, utilising pedigree-like data, in which familial relationships have been genetically verified, to estimate what proportion of back colour variability is genetically determined in a hybrid-zone population, and whether paternal, maternal effects or shared environmental effects are significantly influencing the narrow-sense heritability of back colour. This dataset will also be used to examine any evidence for assortative mating in this hybrid zone, and make inferences concerning the number of genes likely to be involved in this back colour variation from the observed patterns of inheritance in this trait.

2.2 METHODS

2.2.1 Study site and field methods

The Seymour study site consisted of a north-south transect of approximately 14 km along a road in which magpie territories either side of the road have been studied. In some places, side roads and private property access have enabled expansion of the study area width, to half a dozen territories deep in several places. Over the 24 years of this longitudinal study, more than 140 territories have been surveyed and observed, with well over 2000 birds banded and DNA sampled. Field methods presented here vary little from those outlined in Hughes et al. (2011). These territories were not static, but have formed, disappeared,

Heritability shifted borders and composition over the years, so that at a given point in time there were not likely to be more than 90 territories under observation and/or in existence. Three times each year data collecting trips of one to two weeks were undertaken by four people. As part of this PhD, the candidate contributed to these field trips over five years, averaging 1.5 weeks of intensive field-work per year. Extant territories were watched in a number of standardised 20 minute blocks and the bands and back-colours of each bird present in each territory recorded. On two of these trips, timed to coincide with fledging season, birds in each active territory were also trapped, colour and number banded, and a DNA sample taken by clipping a non-dominant toenail and collecting 1-3 drops of blood into 1 mL of lysis buffer solution (0.1 M Tris-HCl pH 8.0, 0.1 M EDTA,

0.5% SDS, 0.01 M NaCl).

Colour banding enabled subsequent confident identification of individual birds at a distance, using telescopes and binoculars. Back colours were recorded at time of banding, as well as each time an individual’s back-colour was confidently observed during watches. The assignment of back colours was done on a 0-4 scale (0 being completely white-backed; 4 being completely black- backed; see Figure 2.3), and each observer over the 24 years of the study was trained by the same researcher, and monitored to assure continued accuracy in scoring. These scores were most commonly integers, but not exclusively, especially in cases where back colour is asymmetrical, and this effectively measured a continuous trait as semi-continuous. In each extant territory, individuals’ presence and back colours, as well as the presence of any unbanded birds, were recorded at between six and eight (20 minute) watches

Heritability annually. This has enabled fairly confident delineation of territory membership at a given point in time.

Figure 2.3: Back colour scoring schematic (males only illustrated; females at the Seymour study site have grey, rather than completely white backs and napes, see Figure 2.1, centre).

This study utilised observational data and DNA samples from 35 territories in the Seymour population, and the data represent all the fledglings and adults caught, banded and/or observed in these territories, for a time period over which there was reliable sighting data and the majority of adult birds banded and bled. This varied between 4 and 14 years between territories (mean=8.97 yrs; SD=3.27), during the period spanning 1993 to 2010.

Territories were selected based on several factors: completeness of watching data, proportion of all adult birds banded, and production of fledglings. Back- colour observations for the 538 individual birds were calculated as the average of all recorded back-colour sightings, helping to make back colour a somewhat continuous trait again. The number of back-colour observations per individual varied between 1 and 42 (average=6.89), and fledglings dominated the lower end of that spectrum, as many dispersed from their natal territory or died within 43

Heritability the first year of being banded. During sightings, researchers utilised a semi- blind procedure in which the person actively scoping for magpies independently called band colours, sex, age status and back colour of birds in a territory whilst a second person recorded data, and cross-checked with previous sightings data in that territory. The data recorder forced the observer to state a definitive observation before confirming each bird, although, in the (relatively rare) cases of asymmetrical back-colours this was sometimes relaxed, as less experienced observers found these individuals harder to call accurately.

2.2.2 Genetic methods

After digestion of whole blood (stored in buffer at -80  C) in 2x CTAB buffer overnight at 65 0C, total genomic DNA was isolated from each sample using a modified version of the CTAB/phenol–chloroform DNA extraction protocol

(Doyle and Doyle, 1987).

The high levels of extra-pair/ extra-group paternity previously found in magpie populations necessitates genetic parentage analyses, rather than simply using long-term observational data gathered for this Seymour population. Eight variable microsatellite loci isolated specifically for C. tibicen in the Molecular

Ecology Laboratory of Griffith University (see Table 2.1) (Hughes et al., 2003,

Durrant and Hughes, 2005), and previously used for parentage analyses in magpie populations (Hughes et al., 2003, Durrant and Hughes, 2005, Durrant and Hughes, 2006) were utilised.

Heritability

PCR reactions were performed in 12.5 μl volumes for each primer, which contained 1.25 μl of 10 x Taq polymerase buffer, 0.75 μl of 25 mM magnesium chloride, 0.5 μl of 10 mM dNTP’s, 0.5 μl of forward and reverse primers, 0.1 μl of Biotech Taq polymerase, 1 μl of DNA template and 7.9 μl double-distilled water. PCR thermal cycling conditions were: one 5-min cycle at 94 0C, then 30 cycles of 94 0C for 30 seconds, primer-specific annealing temperature (see

Table 2.1) for 30 seconds and 72 0C for 30 seconds, followed by a final extension of 72 0C for 7 minutes, after which it was held at 4 0C.

Two different methods were used to resolve the alleles, as technology progressed. Initially PCR product was then denatured at 95 0C for five minutes with formaldehyde and temporarily stored on ice, before being resolved on 5% polyacrylamide gels on a Gelscan 2000 (Corbett Research Pty. Ltd). A commercial standard, ABI PRISM Tamera 350, and two reference samples of known genotype were used to size alleles within and between gels. The ONE-

Dscan software package was used to score genotypes (Scanalytics Inc.). From

2009 onwards, PCR product was resolved on an ABI 3130 Genetic Analyser with a commercial standard (GeneScan -500LIZ, Applied Biosystems).

Genotypes were then scored using Genemapper v4.0 (Applied Biosystems).

In order to maintain consistency between the two different methods of microsatellite resolution, adjustment factors for comparing Gelscan data to ABI

3130 data were calculated; these were due to the addition of a fluorescent tag to PCR products during the PCR reaction in products resolved on the ABI 3130, and these tags differed in length between different loci (Real et al., 2009).

Heritability

Two magpie individuals of known allele sizes were included as calibration standards in every PCR on Gelscans, and then in every run on the ABI 3130 for at least the first 200 individuals to ensure that the transition between the two genotyping methods was accurate and consistent. Commercial size standards were also used in all runs for both genotyping methods. In addition to this, a number of individuals from a wide range of territories (at least 50 individuals) were also genotyped and scored across both methods. Scoring of at least 10% of all individuals was done independently by different researchers, and compared for consistency and accuracy.

The scoring of alleles of all individuals, within each territory, was re-checked as a group to ensure small differences between conditions in different PCRs and runs/gels were not misinterpreted as true variation. Mismatches between likely parents and fledglings were also reviewed and/or rescored at least twice.

Table 2.1: Primer sequences and annealing temperatures for eight microsatellite loci developed for parentage in the Australian magpie (Hughes et al., 2003, Durrant and Hughes, 2005).

Locus Repeat Sequence Primer Sequence Annealing Reference Temp. (0C) Gt43a (AAG)18 F: 5′-GCTACCCGTAAATAAACAAACC 54 Hughes et al. 2003 R: 5′-GAGATGGCAGTGTACAATAAC Gt112 (CA)19 F: 5′-GATGCCTGCATCAGCCACAAG 54 Hughes et al. 2003 R: 5′-AATCTTTTTGCCCTCCTGATC Gt67c (GCA)2AGACCCA(GCA)6 F: 5′-GTCAAATGTCTATTTAAACAGG 52 Hughes et al. 2003 R: 5′-CACAGGAATATCTTGTTACTTC Gt115a (CAAA)4(CAA)10 F: 5′-GTAGTTCTCACTATGGATAAC 52 Hughes et al. 2003 R: 5′-CTGCAATGTTATCAGTTTGCT Gt206b (TC)C(TC)19CACATTCT F: 5′-CAAGCTCAGCCTACAAGATTC 54 Hughes et al. 2003 (TC)2C(TC)2 R: 5′-ATCATTCAGTGCTCGCCGTGG Gt201a (AG)4GGGA(AG)4(GG) F: 5′-CTGAAATCTCAAGCATCTTCC 52 Hughes et al. 2003 (AG)5TCA(AAG)3 R: 5′-TGTCCTGATACCTCTAGCCAA Gt115b (CAAA)4(CAA)10 F: 5′-TGTTCCTAGAGATGACATTTC 54 Hughes et al. 2003 R: 5′-TGCTCTCAGGAGATAAATAAC Gt208 (CT)13G(CT)10GT(CT)6 F: 5’-TC AGA AAG ACC TAG TTG GTG C 52 Durrant & Hughes 2005 R: 5’-GCC TAGTTG AGG TTT TCA AAT G

Heritability

2.2.3 Genetic analyses

Microsatellite alleles were tested for the presence of null alleles using the

MICRO-CHECKER software (Van Oosterhout et al., 2004) and exact tests were performed on each locus in GENEPOP (Raymond and Riousset, 1995) to test for deviation from Hardy-Weinberg proportions. The paternity analysis software package CERVUS (Marshall et al., 1998) was then used to statistically evaluate parentage exclusion probabilities for each possible fledgling-parent trio or pair.

In CERVUS analyses, simulation parameters assumed 85% of candidate mothers and 60% of candidate fathers in the population were sampled, to account for known high levels of extra-group paternity in this population

(Hughes et al., 2003, Durrant and Hughes, 2005), as well as unsampled birds in surrounding groups and flocks. A typing error of 0.05% was assumed to account for observations that allelic dropout was occurring at a slightly higher level than has been previously assumed. A higher incidence of allelic dropout in magpie

PCRs at these loci (than is generally assumed) was uncovered when the change-over between allele resolution methods resulted in the independent duplication of the PCR amplification and allele resolution of a substantial number of magpie individuals. Selected individuals previously identified as homozyotes at a given locus (any) were found to exhibit heterozygosity (and visa-versa) in separate duplicate PCR reactions, both using identical template, primers and PCR conditions (Heterozygote scores were used in downstream analyses of these individuals) Previous magpie parentage studies have assumed a rate of allelic dropout of approximately 1% (Durrant and Hughes,

2005), however, in independently replicated magpie PCRs, this rate was

47 Heritability observed to be closer to 2.5% across all individuals genotyped as part of this study.

Confidence levels of 80% are considered sufficient to accurately estimate parentage using the CERVUS package, even if putative fathers are closely related (Slate et al., 2000), when combined with behavioural data. In the

CERVUS package, confidence levels reflect the “tolerance of levels for false- positive” parentage assignments (Marshall et al., 1998). When using parentage assignment to look at extra-pair and extra-group paternity and maternity, it is generally considered more conservative to accept social parents as true parents, despite a few loci mismatches, in order to avoid inflating estimates of extra-pair/group parentage. This is a consequence of the level of genotyping error inherent in microsatellite analyses, introduced by allelic dropout, microsatellite stutter, null alleles, contamination, and human error, especially using earlier methods and technologies (Kalinowski et al., 2007).

However, in this study, parentage analyses are a vehicle to ensure that confident parental-offspring relationships are used to examine colour heritability.

Therefore, the more serious error becomes falsely assigning the wrong candidate parentage (false positives), as false negatives will only result in the reduction in the total number of data points. Thus a strict confidence level of

90% and a maximum of one mismatching locus between parent and offspring, were used to reduce the false positive rate, at the expense of inflating the number of false negatives and consequent loss of data.

Initially, within each of the territories, all females and males older than one year of age that were present in a territory within the previous or following two years,

Heritability were tested as potential genetic parents of fledglings born in that territory.

However, if maternity or paternity could not be assigned due to low confidence, fledglings were tested against adults from the surrounding 10-20 territories.

Paternity and maternity were only assigned if confidence levels (C.L.) were at or above 90%, and parent-fledgling trio or pairs mismatched at a maximum of one locus. Mismatches were rechecked exhaustively for scoring errors, especially in cases of one microsatellite repeat unit difference. In cases where more than one potential mother or father matched at an equal number of loci, parentage was only assigned if a putative parents’ LOD score was above the 90% confidence level and/or positive assignment of the other parent sex enabled discrimination between the most likely of these putative parents using trio LOD scores.

2.2.4 Heritability analyses

Heritability generally refers to the proportion, or ratio, of variation in a given phenotypic trait that is explained by genetic variation. Broad-sense heritability is the proportion of this trait variation that is a result of all genetic factors, including gene interactions and dominance effects; while narrow-sense heritability is the heritability metric estimated in this chapter, and represents the proportion of trait variation due to additive genetic factors, and is often explained more simply as the resemblance between relatives.

2.2.4.1 Regressions

To estimate heritability of back colour in this population, several different regressions were undertaken to best utilise all identified genetic relationships. 49

Heritability

Parent-offspring regressions are often utilised for the estimation of the narrow sense/additive heritability (h2) of a phenotype (Lynch and Walsh, 1998). These are robust even under conditions of assortative mating, and single parent- offspring regressions (measuring half the heritability) can be used to identify possible maternal or paternal effects (Falconer, 1975). Excel software and the

Real Statistics Resource Pack software (Release 4.3) (Zaiontz, 2014) were used for simple linear mid-parent/mid-offspring regressions, calculated on families in which parent pairs had had more than one offspring together

(broods). Mother/mid-offspring, father/mid-offspring as well as lightest parent/mid-offspring and darkest parent/mid-offspring regressions were also calculated to examine maternal, paternal, and dominance effects.

As a large number of parent pairs had only one offspring, these same regressions were repeated, but on back-colours of individual offspring (as opposed to mid-offspring values) for all parent combinations above. All parent/mid-offspring heritability estimates were weighted with an ICC (intra- class correlation coefficient), using iterative reweighting procedures until h2 values converged, in order to account for variance in the number of offspring within different broods, which ranged from 2-9 (mean=3.63) following Lynch &

Walsh (1998).

An even more restricted subset of the data, composed of only families in which paternity and maternity confidence levels (C.L.) were above 90%, and parent- fledgling trio or pairs did not mismatch at any loci, was also examined using the same set of regressions in order to assess confidence in, and congruence with the larger dataset.

Heritability

2.2.4.2 Animal models

A mixed-effects model, known as the animal model, was also used to analyse heritability of back-colour. Animal models have become an increasingly popular means of estimating the heritability of traits (e.g. Morales et al., 2010, Lane et al., 2011, Losdat et al. 2014), in part because they cope well with unbalanced designs and missing data, an often inherent feature of long-term studies of natural breeding populations (Kruuk, 2004). These models also importantly make use of all of the complex familial relationships common in the pedigrees of natural breeding populations, increasing overall sample size. Additionally, they enable the estimation of the effect of environmental components of variance

(e.g. cohort, nest, birth year) by including them as random effects in the model, which could artificially inflate h2 estimations if left unaccounted for. Such environmental effects and/or their covariances are often themselves key factors of interest when investigating ecological questions about phenotypic variance

(Wilson et al., 2010). Animal models are also relatively robust to bias from inbreeding or assortative mating (Lynch and Walsh, 1998, Kruuk, 2004).

The package MCMCglmm (Version 2.22.1) (Hadfield, 2010), implemented in R

(Version 3.2.3) (R Development Core Team, 2013) through R-Studio (Version

0.99.891) (R Studio Team, 2015) was used to generate an additive genetic relationship matrix from genetically verified pedigree data (strictest confidence dataset), comprising 406 individual magpies (289 offspring, 58 fathers, 70 mothers; 11 offspring later produced their own broods). This matrix was then used to fit a series of animal models with back colour phenotypes, including and excluding the random effects of territory of birth and/or year of birth. These models were run for 100 million iterations (burnin: 1 x 106, thin: 3000) to bring

Heritability autocorrelation down to negligible levels, and were evaluated carefully to ensure convergence, good sampling and adequate effective sample size.

This data was then used to calculate the posterior distribution of the heritability and its upper and lower 95% credible intervals, as well as Deviance information criterion (DIC) in order to compare model fit and complexity and determine if the inclusion of either or both random effects was justified. The DIC is a Bayesian metric similar to the Akaike information criterion (AIC) (Burnham and Anderson,

1998) that quantifies the balance between model fit and complexity

(Spiegelhalter et al., 2002); small values of DIC are generally preferable when selecting the best model ((Wilson et al., 2010) Supplementary online file: File

S5. Tutorial for MCMCglmm). Each model was run for 100 million iterations, discarding the first one million iterations (burnin) and keeping one in every 1500 iterations (thin). Models were checked to ensure adequate mixing, low levels of autocorrelation and effective sampling size (Hadfield, 2010, Wilson et al., 2010).

2.3 RESULTS

2.3.1 Parentage analysis

The number of microsatellite alleles ranged from 4 to 24 per locus, while heterozygosity varied between loci, from 0.40 to 0.877 (Table 2.2). One locus

(Gt206b) differed significantly from Hardy-Weinberg proportions after sequential

Bonferroni correction for multiple tests; however, this locus has been previously used in parentage analyses in this exact population (Durrant and Hughes, 2005,

Hughes et al., 2011), and other populations of the Australian magpie (Hughes et al., 2003, Durrant and Hughes, 2006) without issue. The very high levels of

Heritability relatedness within several sampled magpie territories (A. Dobson, unpublished data), are considered to have led to this aberrant result for Gt206b. Hughes et al. (2011), utilising these same microsatellite loci for the same Seymour population, used a subset of their data consisting of putatively unrelated individuals to conclusively confirm none of these eight loci were significantly out of Hardy-Weinberg equilibrium, and thus Gt206b was not excluded from further analyses.

The first assigned parent had an exclusion probability of 0.9992 from eight loci combined, and where both parents were assigned, this rose slightly to 0.9999

(Table 2.2).

Table 2.2: Summary statistics of microsatellite loci used in parentage analysis.

Locus Number HW p-value Exclusion probability Exclusion probability Null allele of alleles (1st parent) (parent pair) frequency estimate Gt43a 18 0.025 0.614 0.762 0.014 Gt112 24 0.177 0.764 0.866 0.031 Gt67c 4 0.940 0.9085 0.227 -0.003 Gt115a 16 0.066 0.449 0.623 0.026 Gt206b 15 0.002* 0.418 0.594 -0.023 Gt201a 16 0.470 0.388 0.571 -0.023 Gt115b 14 0.005 0.602 0.753 0.087 Gt208 22 0.502 0.582 0.737 0.009 Combined 4 to 24 0.999 0.999

Two parentage datasets from the Seymour population were used together in further analyses of back colour: the first dataset encompassed 124 offspring from 17 territories at Seymour (1995 to 2005) for which both parents had been assigned in Hughes et al. (2011), which was combined with a second dataset composed of an additional 193 offspring, from 19 additional territories (1993-

Heritability

2009) with at least one assigned parent. Sampling sizes are further detailed in

Table 2.3.

Of the 24 offspring from these (second dataset) 19 territories that were unable to be assigned to a parent (C.L ≥90%, ≤1 loci mismatch), and which were discarded from further analyses, only five were cases in which no likely parent could be found within, or outside of that territory. The remaining unassigned offspring were split fairly evenly between cases in which highly likely true parents came in just under our defined confidence levels, and cases where more than one potential mother or father were identified as equally likely (or very nearly equally likely) to be a true mother or father. Offspring with multiple highly likely parents, based on parentage analysis, were born in territories in which earlier offspring of either or both sexes had remained in their territory of birth, well past sexual maturity living (and putatively breeding) alongside their genetic parents, or territories with siblings from the same original brood

(putatively) breeding alongside each other within the territorial group.

Table 2.3: Summary of magpie samples sizes used in complete dataset analyses of heritability; combines birds from the 17 territories sampled in Hughes et al., (2011) together with the 19 territories sampled f this study.

Sampling group Number

INDIVIDUAL OFFSPRING OFFSPRING with both parents assigned 238 with mother assigned 292 with father assigned 262 with only mother assigned 55 with only father assigned 24

FAMILY GROUPS (≥ 2 offspring) FAMILIES with both parents assigned 54 with mother assigned 66 with father assigned 58 with only mother assigned 12 with only father assigned 4

Heritability

2.3.2 Back colour metrics, assortative mating and inheritance patterns

Back colour scores at the Seymour population are not uniformly distributed, with darker backed individuals occurring in higher frequencies than lighter-backed individuals, at least within the last 25 years (Table 2.4). Transects done in the early 1970s of magpie back colour frequencies, in close proximity to the

Seymour study site (Burton and Martin, 1976) indicate slightly different frequencies of the different back colours, with less of a skew towards darker back colours (see frequency distribution in Appendices, Appendix I). However it should be noted that this could simply be due to either the slightly different back colour scoring scheme used in that study, their smaller sampling, or a result of slight geographical distances between the two sampling locations, as the proportions of different back coloured magpies shifts rapidly over small distances in this part of the hybrid zone (Jane Hughes, personal communication). Additionally, the criterion used to select territories to analyse was biased in terms of back colour, with territories chosen influenced by the research questions, rather than representing a truly random sample from this population.

Table 2.4: Average back colours of sampled birds in the Seymour population, calculated using averaged observations of (1993 to 2014) back colour scores for each individual bird.

N Mean SD Offspring 317 2.28 1.37 Mothers 293 2.13 1.47 Fathers 262 2.38 1.31 Microsatellite typed birds (19 territories) 528 2.46 1.18

Heritability

200 180 160

p 140

a 120

i 100

er 80

b i 60

u 40

N 20 0 0 1 2 3 4 Back colour

Figure 2.4: Frequency distribution of magpie back colours in analysed territories of the Seymour study population.

Back colours of female and male magpie pairs that produced fledglings sampled in this study were examined for evidence of assortative mating. The correlation value (0.12) between female and male back colours of these mating pairs was not statistically significant (t= 0.11, p= 0.92). There was no apparent trend toward either positive or negative assortative mating patterns (Figure 2.5).

The distribution of the back colour pairings reflect frequencies of back colours observed in the sampled population, with more dark backed individuals breeding overall.

Heritability

2 ack

M 1

0 0 1 2 3 4 Female back colour

Figure 2.5: Pairing patterns with relation to back colour, of genetically verified mating pairs in the Seymour study population; N=55.

The range of offspring produced by different combinations of parental colours indicate a pattern of like producing like; whilst opposite colour pairings, and pairs both of mid back colour produce an almost Gaussian-like distribution of all back colours, centred near the most intermediate phenotype (Figure 2.6).

Heritability

Figure 2.6: Offspring back colour ranges for different combinations of parental back colours, y- axis values indicate the number of offspring. The designation ‘light’ includes bird with back colours from 0-1.5, ‘mid’ includes back colours 1.5-3, and ‘dark’ includes back colours 3-4. Note the top histogram (parents both dark) is scale differently on the y-axis.

2.3.3 Regressions

Two subsets of the data were used to calculate kin regressions of back colour:

a complete dataset with a parentage assignment C.L ≥90%, ≤1 loci mismatch,

Heritability and a subset of this data, comprised only of kin relationships confirmed under even stricter criteria (C.L ≥90%, 0 loci mismatching). In all kin back colour regressions calculated, heritability estimates were highly significant (p≤0.001)

(Tables 2.5 and 2.6; Figure 2.7).

When only kin relationships with extremely high confidence (C.L ≥90%, 0 loci mismatches) were used in regressions, estimates of h2 were almost identical to those obtained using the complete dataset (Table 2.6). All were still very highly significant at p≤0.001 (Table 2.6). Thus we concluded that using the complete dataset to accurately assess h2 is reasonable, and present regression figures derived from this dataset in this study (Figures 2.7 and 2.8). Regression figures derived from the strict subset are included in the Appendices as appendices II and III.

Using the complete dataset, regressions of individual offspring back colour, on one or both parents, indicated a range of h2 from 0.34 to 0.6 (Table 2.5), while regressions of lifetime offspring average back colour upon single parents or parent pairs ranged from 0.48-1 (Table 2.5). Regressions of mid offspring colour on the darkest and the lightest parent colour yielded heritability estimates of 0.48 and 0.61 respectively (Table 2.5, Figure 2.8 a & b).

Once weighted using intra-class correlation, the estimated heritability of back colour variation increased markedly and similarly across all regression types.

Parent midpoint-offspring midpoint regression (corrected for unequal family size) is considered the most robust regression-based measure of narrow sense heritability out of all the types of kin relationships calculated in this study (Lynch

Heritability and Walsh, 1998), and after ICC weighting, this was calculated to be 0.94 in the study population, indicating a high level of heritability of magpie back colour variation (Table 2. 5, Figure 2.7).

Although heritability estimates obtained from mother-offspring regressions were lower than those obtained from father-offspring estimates, this difference virtually disappeared once regressions were weighted to account for differing family sizes. Back colour h2 based on corrected mother-offspring regression was estimated at 0.86, and h2 based on corrected father-offspring regressions estimated at 0.92 (Table 2.5, Figure 2.8 c and d), indicating this was likely to have been an artefact of the tendency of father-offspring ‘families’ to be larger; a consequence of the social structure of magpie territories in this population.

Table 2.5: Kin regressions using complete dataset: heritability estimates for C. tibicen back- colour variability in the Seymour population. Values in brackets refer to total numbers of offspring within all families. Back colour regression N h2 SE p-value ICC weighted h2 SE weighted Father - individual offspring 262 0.60 1.17 6.60E-22 Mother - individual offspring 292 0.34 1.25 2.00E-13 Parent midpoint - individual offspring 238 0.44 1.04 1.27E-31 Darkest parent - mid offspring 54 (196) 0.48 0.83 4.82E-09 Lightest parent - mid offspring 54 (196) 0.62 0.71 2.05E-12 Father - mid offspring 58 (210) 1.00 0.79 1.18E-10 0.92 4.55 Mother - mid offspring 66 (235) 0.54 0.96 7.84E-06 0.86 4.73 Parent midpoint - mid offspring 54 (196) 0.70 0.63 5.77E-33 0.94 1.47

Table 2.6: Kin regressions using strict data subset: heritability estimates for C. tibicen back- colour variability in the Seymour population. Values in brackets refer to total numbers of offspring within all families.

Back colour regression N h2 SE p-value ICC weighted h2 SE weighted Father - individual offspring 233 0.65 1.16 2.00E-21 Mother - individual offspring 243 0.34 1.28 2.46E-11 Parent midpoint - individual offsprin 194 0.47 1.05 3.79E-28 Father - mid offspring 39 (194) 1.00 0.77 4.01E-08 0.90 0.93 Mother - mid offspring 51 (217) 0.48 0.90 2.46E-04 0.88 11.48 Parent midpoint - mid offspring 43 (153) 0.73 0.64 2.30E-13 0.94 1.53

Heritability

Figure 2.7: Regressions of mid parent-mid offspring back colour in a population of Australian magpies at Seymour, calculated based on the complete dataset. Data are for 54 broods (brood size μ=3.64, σ=1.77) comprising 196 offspring and 86 parents, from 31 territorial groups during the period 1993-2009.

Heritability

Figure 2.8: Complete dataset regressions of parent-offspring back colour, in a population of Australian magpies at Seymour calculated based on the complete dataset. Birds were from 30- 32 territorial groups during the period 1993-2009. (a-d) refer to individual regressions of mid offspring back colour onto different parents by sex and by back colour.

2.3.4 Animal models

Heritability estimates of magpie back colour variation, obtained using animal models (h2 = 0.88-0.94) (Table 2.7) substantially agreed with ICC-weighted regression estimates (h2 = 0.84-0.92) (Table 2.5). Inclusion of the different additional random effects of year of birth and territory of birth within animal

Heritability models affected estimates of additive heritability in different ways (Table 2.7,

Figure 2.9).

Fitting the most complex model (model 4) reduced the h2 estimate and broadened its 95% confidence interval substantially, but increased the information criterion (DIC). Model 2 was found to be most parsimonious; the addition of the random factor of year of birth substantially improved model fit for magpie back colour, with the lowest DIC value by a large margin. Therefore, to avoid biased parameter estimation, model 2 was chosen to estimate heritability of this trait.

Table 2.7: Generalised linear mixed models (GLMMs) (animal models) of magpie back colour with and without the random effect of year of birth and territory of birth included. h2 values include 95% credible intervals in parenthesis. Analyses included data from 406 birds, including 288 offspring from 35 territories between 1993 and 2009. Some individuals were included as both offspring and parents. .

Random effect(s) h2 DIC Model 1: ~ back colour 0.91 (0.80-0.99)** 382.51 Model 2: ~ back colour + year of birth 0.92 (0.80-0.99)** 317.65 Model 3: ~ back colour + territory of birth 0.89 (0.75-0.99)** 500.35 Model 4: ~ back colour + year of birth + territory of birth 0.84 (0.65-0.99)** 444.27 ** significant at p< 0.001

Heritability

Figure 2.9: Posterior density of heritability estimates for different animal models tested using GLMMs.

2.4 DISCUSSION

Comparisons between the complete dataset and a sub-sample, utilising only perfectly-matched parentage assignments, indicated that the criterion used to delineate true genetic parents for the complete dataset was sufficiently robust, as very similar estimates of heritability between kin groups were recovered from both datasets. While microsatellite-based parentage analyses used to investigate levels of extra-pair and extra-group paternity and maternity are more conservative when they confirm social parents as true parents in cases of highly likely (but imperfect) offspring-parent matches, the use of parentage analyses to inform genetic models of trait inheritance must conversely be more conservative and assign only the most stringent of pairings between parents and offspring.

Heritability

Only a small proportion of fledglings (11%) typed for this analysis could not be assigned any confident parentage under the criterion used for the complete dataset. Whilst it is possible more loci, or more diversity within loci may have been able to recover these ‘lost’ data points, at least half of these were from territories with very high levels of relatedness, and/or with closely related females simultaneously (putatively) breeding in territories. In such cases it seems reasonable to suggest that only significantly higher numbers of loci may disentangle true genetic parentage, at marginal gains to the overall size of the complete dataset.

Assortative mating may artificially inflate heritability estimates using single parent-offspring regressions (Falconer, 1975), but not heritability estimates using mid-parent/ mid-parent values (Lynch and Walsh, 1998), whereas the animal model purports that non-random mating should not unduly influence its estimation of trait heritability due to its inclusion of multi generational data

(Kruuk, 2004). The very weak trend towards more mating pairs in which both individuals are black backed, is suggested to be an effect of the higher frequency of darker backed individuals overall at the study site, rather than any assortative mating pattern.

If some form of even weak assortative mating had been observed, it could have potentially played a large role in the maintenance of this hybrid zone by limiting gene flow across it to some extent (Barton and Hewitt, 1985). It has been proposed that the hybrid zone between hooded and carrion crows in Europe is largely maintained by positive assortative mating within the zone (Brodin and

Haas, 2009, Haas et al., 2010), and it is interesting to note that these birds

Heritability display pied-like black, white and grey plumage variants strikingly similar to those found in the Australian magpie.

Traditionally, one of the most difficult aspects of obtaining good estimates of genetic heritability of a phenotypic or behavioural trait has been disentangling additive heritability from the potentially confounding effects of common environment and/or maternal effects (Lynch and Walsh, 1998). Simply comparing mother-offspring regression h2 estimates to midparent-offspring estimates is a good way to exclude strong maternal effects: if maternal effects are an important source of the variation, mother-offspring estimates should be higher than midpoint estimates (Lynch and Walsh, 1998). Initial analyses of magpie back colour variation indicated higher father-offspring estimates than midparent-offspring estimates, with mother-midoffspring estimates much lower still. However, unequal offspring number distributions, within and between sexes, were found to be the driver of this, and once intra-class correlation- based weightings were applied to regressions, this effect disappeared entirely.

Thus mother-midoffspring regressions were consistently much lower (weighted h2=0.88) than midparent estimates (weighted h2=0.94), as were father- midoffspring h2 estimates (weighted h2=0.92), indicating it is unlikely that maternal or paternal effects play a large role in back colour heritability of magpies in this population.

As magpie back colour variation also has a sexually dimorphic component, it is plausible that this trait may have a component of sex-linked inheritance, and if so, animal model heritability estimates of autosomal additive variance presented here could potentially be inflated by the effect of sex-linked inheritance. 66

Heritability

Methods developed to model and test for sex-linked inheritance using modified animal models have been used in a number of bird species (Husby et al., 2012,

Larson et al., 2014). It is suggested future studies of magpie back colour with access to genetically determined sex data for individuals observed only as juveniles, as well as a significantly larger pedigree will benefit from the estimation of the partitioning of heritability into autosomal and sex-linked components, especially if sex-specific selection is indeed involved in this back colour trait.

Cross-fostering experiments have also been utilised in many bird species to separate genetic from environmental sources of co variation (Merilä and

Sheldon, 2001). In cases where sample sizes are low, immigration high, and suspected common environmental effects are high, Kruuk & Hadfield (2007) advocate the use of both cross-fostering manipulations and animal models together, but suggest that if common environmental effects are likely to be low, animal models alone are likely to give reasonably unbiased estimates of h2.

Cross-fostering experiments are largely impractical in wild magpie populations, for reasons mentioned above, and discussed more extensively in Durrant et al.

(2005), and thus the possible environmental effects of patch quality and cohort were investigated by evaluating h2 using animal models with and without the inclusion of territory and year of birth as random effects. With both environmental factors included in the model, h2 estimates did fall slightly, with wider 95% confidence intervals (Table 2.7).

Neither territory nor year of birth contributed much to the model based on posterior means (although territory did contribute more than birth year), and a

67 Heritability much larger DIC value indicates the complete model is likely to be less well- fitting and/or over-parameterised. If we accept the model with the lowest DIC,

(Model 2: back colour and year of birth), the estimated h2 value of 0.92 (95% CI

0.8-0.99) is indeed highly congruent with h2 estimates obtained using weighted regressions of midparent on midoffspring of 0.94.

These values indicate a very high level of back colour heritability in this population of magpies, although it is not dissimilar to melanin-based trait heritability estimates in other bird species, or entirely unexpected. First, magpie back colour has been observed to be stable at the level of individual, both spatially and temporally, and does not materially change across seasons or moults. Secondly, melanin-based traits have often been shown to follow

Mendelian laws of inheritance across a large number of bird species, indicating these traits are under strong genetic control, rather than environmental or body- condition-dependant (Roulin, 2004). This makes sense in the light of the fact that melanins are for the most part endogenously produced in the body, in contrast to the majority of carotinoid-based plumage traits, which are primarily ingested and as such, are heavily dependent upon environmental sources of carotinoids and the relative ability of individuals to obtain them (Fox, 1976,

McGraw, 2006, Roulin and Ducrest, 2013). Thirdly, a large number of other bird species that also express different morphs of melanic-based phenotypic traits have been found to have estimated heritability levels of a similarly high magnitude, including tawny owls (0.93) (Gasparini et al., 2009), barn owls (0.81)

(Roulin and Dijkstra, 2003), barn swallows (0.8) (Saino et al., 2013b), alpine swifts (0.78) (Bize et al., 2006), pied flycatchers (0.31-0.93) (Potti and Canal,

2011), and common kestrels (0.67-0.83) (Kim et al., 2013). These contrast

Heritability dramatically with heritability estimates for some non-melanin based variation in other species, such as the carotinoid chest (0.07) and structurally-based crown

(0.10) in blue tits (Hadfield et al., 2007), and the carotinoid based ventral patch in the great tit (0.29) (Evans and Sheldon, 2012).

Heritability levels of traits do not necessarily indicate the number of loci that contribute to its genetic variation: traits with high heritability can potentially be complex traits with many loci each contributing a small amount to its variation, while conversely traits with low heritability can potentially be determined by just a single mutation (Visscher et al., 2008). Human height is a good illustration of this; while height is highly heritable (0.8) (Silventoinen et al., 2003), hundreds of thousands of SNPs together are thought to represent the genetic basis of this trait (Yang et al., 2010). The common misconception that high heritability values indicate that few genes of large effect may be responsible likely stems from two different factors. Firstly, traits that have Mendelian (or close to) patterns of inheritance, and are determined by a single gene generally have heritability values of 1, or very close to 1 (Visscher et al., 2008). Secondly, gene-mapping studies of traits with high heritability are more likely to be able to detect genes of large effect, due to the stronger correlation between genotype and phenotype (Risch, 2000, Visscher et al., 2008).

Thus, while the heritability by itself cannot be used directly to reveal the genetic architecture underlying a trait, it can inform and focus efforts to find the genetic basis of such traits. Investigations of traits with heritability estimates close to 1, that also have Mendelian patterns of inheritance are probably good candidates for candidate gene analyses, and broader gene mapping studies may be an

Heritability unnecessary use of resources in such cases, although such gene-mapping studies are likely to be more efficient for such traits. The high heritability of back colour variation in magpies is relatively close to 1, but the patterns of inheritance do not indicate it is inherited in a strictly Mendelian fashion, suggesting the minimum number of genes likely to be involved in determining magpie back colour is likely to be at least two. Although data from this study cannot suggest more detailed information about the genetic architecture of this trait than this two-gene minimum, Hughes and Mather (1980) suggested a small number of genes may determine magpie back colour, and Hughes (1982) further expanded on this theory, putting forward a model based on two genes, with black back colour being the dominant allele in both genes. They demonstrated how such a model could explain the previously observed asymmetry in the frequencies of back colours across the eastern hybrid zone

(Burton and Martin, 1976).

The range of offspring back colours produced by different crosses of parental phenotypes in this study do show some support for a dominance of black back alleles, at least at one locus (if not more), as two black backed parents could produce a white backed offspring, but two white backed parents have yet to be found to produce completely black backed offspring (across 238 offspring born across 35 territories in a 16 year period). Incomplete dominance effects seem less likely, given intermediate colour forms are not true ‘mixes’ of ‘blends’ of black and white phenotypes (i.e. grey or similar), but express both white and black phenotypes over varying proportions of their backs, which itself is more suggestive of some type of co-dominance effects between phenotypes. It also remains quite possible that more complex mechanisms involving incomplete

Heritability penetrance and/or differing levels of expressivity could potentially be involved in magpie back colour variation. The continuous nature of magpie back colour does, however, indicate that in the case of a hypothetical two-gene model of inheritance (as suggested by Hughes, 1982), at least one gene would likely need to be a modifier of some type i.e. control the width of the black band on the back, as was first suggested in Hughes (1982).

Heritability estimates can also be a useful tool with which to compare the relative contributions of genes and environment to a trait, or type of trait. A great example of this is carotinoid-based plumage traits, which have consistently been found to have low-to-moderate heritability (Hadfield et al.,

2006, Hadfield et al., 2007, Quesada and Senar, 2009, Evans and Sheldon,

2012) ostensibly due to the fact that such pigments cannot be produced endogenously by birds, and this pigment must be sourced from their environment, often through dietary intake. The low environmental contribution to magpie back colour variation in the study population, implied by the high heritability estimates, may indeed be different in different populations or timeframes. This is because heritability can vary within and between populations, as well as temporally, and a pattern of higher heritability of morphological traits in ‘better’ or more stable environments is becoming apparent in the literature, with this increased heritability likely to be a result of the decreased environmental variance in such environments (Visscher et al.,

2008).

It would be interesting to compare heritability estimates for magpie back colour in poorer-quality habitats within a hybrid zone, but the study population is the

Heritability only long-term study site in a hybrid zone, and the data needed to estimate heritability using a pedigree approach would be disproportionately costly, time- consuming and impractical to collect, especially in poorer habitats with lower magpie densities. However, alternative methods of estimating heritability that do not require large pedigrees within populations are promising (Ritland, 1996,

Lynch and Walsh, 1998, Thomas and Hill, 2000, Thomas et al., 2002), and use relatedness estimates from a large number of genetic markers. However due to the large number of markers required, this method may prove prohibitively costly unless the data can be used for more than one purpose, and some studies indicate this method may still be less likely to yield heritability estimates as accurate as those obtained using traditional pedigree-based methods

(Thomas et al., 2002). In addition, heritability estimates in this population are very high, suggesting that environment is likely to have minimal effect on back colour.

2.5 CONCLUSIONS

Evidence presented here indicates heritability of back colour variation is high in this population of magpies within the eastern hybrid zone, and that this high heritability is consistently estimated by two different analytical methods. This result is not unexpected for a melanic plumage trait, which generally show higher heritability than carotinoid or structurally-based plumage colour traits.

Single-parent heritability estimates indicate that neither maternal nor paternal non-genetic effects (e.g. parent body condition), play a large role in determining offspring back colour, and animal models indicate the environmental effects of patch quality and cohort also contribute very little to the heritability of this trait in

Heritability magpies. Patterns of inheritance in familial groups suggest there may be some effect of dominance of black backed alleles, and that at least two genes are likely to play a role in determining magpie back colour. The high heritability of back colour variation found in this study also indicates that gene-mapping studies may be an effective approach for further studies of the genetic basis of plumage colour variation in this species.

73 Chapter 3 is a modified version of a pre-copyedited, author-produced version of an article accepted for publication in the Journal of Heredity following peer review.

Dobson, A.E., Schmidt, D.J. and Hughes, J.M., 2012. Sequence variation in the melanocortin-1 receptor (MC1R) does not explain continent-wide plumage color differences in the Australian magpie (Cracticus tibicen). Journal of Heredity, 103(6), pp.769-780.

This record is available online at:

URL: https://academic.oup.com/jhered/article/103/6/769/891034/Sequence-Variation-in-the- Melanocortin-1-Receptor

DOI: https://doi.org/10.1093/jhered/ess053

74 MC1R & Phylogeography

CHAPTER 3: MC1R AND MAGPIE PHYLOGEOGRAPHY SEQUENCE VARIATION IN THE MELANOCORTIN-1 RECEPTOR (MC1R) DOES NOT EXPLAIN CONTINENT- WIDE PLUMAGE COLOUR DIFFERENCES IN THE AUSTRALIAN MAGPIE (CRACTICUS TIBICEN)

3.1 ABSTRACT

Whilst the genetic basis of plumage colour variation has already been determined for many model species, the genetic mechanisms responsible for intraspecific colour variation in the majority of wild bird species are yet to be uncovered. The Australian magpie (Cracticus tibicen) is a large black and white passerine which is widely distributed across the Australian continent. The proportion of melanised back plumage varies between regionally delineated subspecies; where back-colour forms overlap, intermediate colour phenotypes are produced. This study examined the majority (861 base pairs) of the coding region of the melanocortin-1 receptor (MC1R), a candidate gene for plumage colour differentiation in 98 magpies from across the Australian continent in order to determine if the gene is associated with magpie back colour variation and explore phylogeographic signal within the gene. Neutrality and selection tests

(Tajima’s D, Fu’s Fs, MKT) indicate the gene is unlikely to be currently under selection pressure and together with other lines of evidence, suggest a past demographic expansion event within the species congruent with the results of previous mitochondrial phylogeographic work on this species. None of the fifteen synonymous and four nonsynonymous substitutions within MC1R were found to be associated with plumage variation. Our results suggest that genes or regulatory elements other than MC1R may determine back-colour variation in

C. tibicen. 75 MC1R & Phylogeography

3.2 INTRODUCTION

Plumage colour variation in bird species has long fascinated both scientists and amateurs alike. These often striking, but sometime subtle differences in plumage colour and pattern provide a natural experimental design ideal for investigating questions about morphological diversity and the forces that generate and shape this diversity.

Understanding the underlying genetic and regulatory mechanisms responsible for such variation is a first step towards addressing these important questions.

However until recent decades, the construction of large pedigrees of domesticated and laboratory animals was the only way to investigate the genetics of plumage variation. As technology has advanced, the ways in which evolutionary questions about intraspecific variation have been approached have changed dramatically; genetic sequencing advances have now made it possible to screen well-characterised candidate genes relatively affordably and efficiently

(Piertney and Webster, 2010).

To date, investigations of colour variation within animal species have focussed primarily on structural coding regions, and these have revealed a number of different candidate genes that are thought to be involved in colour variation within different species. Studies of the genetic basis of pigmentation have been dominated by one gene in particular over the last 20 years: the melanocortin-1- receptor (MC1R), a gene that encodes a seven-transmembrane domain G- protein coupled receptor (Mountjoy et al., 1992). This protein is expressed in specialised pigmentation cells known as melanocytes, where it plays a role in 76 MC1R & Phylogeography the dispersal of melanosomes through cells and/or initiation of the melanin- production process (Jackson, 1997). Activation of the receptor by melanocyte- stimulating hormone (MSH) has been shown to lead to an increase in the production of black and brown eumelanin in melanosomes (Robbins et al.,

1993). Mutations in the receptor that lead to activation of MC1R and increased synthesis of eumelanin are known as gain-of-function mutations, whilst loss-of- function mutations in MC1R often are associated with the production of red or yellow pheomelanin (Robbins et al., 1993).

MC1R has been implicated in intraspecific pigmentation variation across a wide range of mammal and reptile species, including the horse (Marklund et al.,

1996), fox (Vage et al., 1997), arctic fox (Vage et al., 2005), pig (Kijas et al.,

1998), sheep (Vage et al., 1999), dog (Newton et al., 2000), black bear (Ritland et al., 2001), cow (Klungland et al., 1995), jaguar and jaguarundis (Eizirik et al.,

2003), pocket mouse (Nachman et al., 2003), domestic rabbit (Fontanesi et al.,

2006), human (Valverde et al., 1995), lesser earless lizard and little striped whiptail (Rosenblum et al., 2004). Birds are no exception to this widespread

MC1R association with colour phenotype, and specific mutations within the

MC1R coding region have been found to be associated with plumage colour in chickens (Andersson, 2003, Takeuchi et al., 1996), Japanese quails (Nadeau et al., 2006), red-footed boobies (Baiao et al., 2007), lesser snow geese (Mundy et al., 2004), arctic skuas (Mundy et al., 2004) chestnut-bellied monarch (Uy et al.,

2009) and bananaquits (Theron et al., 2001), although several species with plumage variation in the form of fine-scale patterning have been shown to have no association with variation in this gene, including old world leaf-warblers

(MacDougall-Shackleton et al., 2003), blue-crowned manakins (Cheviron et al.,

77 MC1R & Phylogeography

2006), rosy-finches (Drovetski et al., 2008), and carrion/hooded crows (Haas et al., 2008).

A candidate gene approach has been successfully used to identify associations between phenotypes and genotypes across a wide range of species and traits

(see Hoekstra and Coyne 2007). Although this approach is more often used to investigate simple Mendelian traits; studies of complex, multi-gene based traits may also benefit by using this approach to investigate the possible underlying pathways that link genetic variants to complex traits (Tabor et al., 2002).

The Australian magpie Cracticus tibicen (formerly Gymnorhina tibicen) is a sedentary, group-living passerine belonging to the family Artamidae (Order

Passeriformes). Eight subspecies of magpie are currently recognised based on morphological traits including size, bill length, wingspan and also plumage colour, which varies considerably across its distribution (Schodde and Mason,

1999); see Figure 3.1 for illustrations and distribution. Three of these subspecies (C.t. telonocua, C.t. tyrannica and C t. hypoleuca) are white-backed

(WB) plumage forms and are restricted to south-eastern Australia and

Tasmania. These WB plumage forms have a white back which joins their white nape and extends down to their black rump (Schodde and Mason, 1999).

Black-backed (BB) plumage forms dominate northern parts of the continent, and comprise four different sub-species: C.t. longirostris in the north-west, C.t. eylandtensis in a belt across the central north of the continent, C.t. terraereginae throughout most of Queensland and New South Wales, and C.t. tibicen along the New South Wales coastline. Instead of a white saddle, these

BB birds have a black saddle extending from their white nape down to their

MC1R & Phylogeography rumps. All of the seven sub-species of black and white-backed forms exhibit a sexually dimorphic plumage pattern, in which the white plumage areas on males are grey or partially grey on females (although heavy parasite loads can make males appear somewhat grey in the nape area (J. Hughes, personal communication)).

A third plumage group exists in the form of the varied magpie, C.t. dorsalis, a subspecies found only in the extreme south-west of the continent. The males of this sub-species resemble male WB forms, in terms of plumage. Females of this sub-species have a black back which begins at their shortened white nape and extends down to the rump, but these black feathers are all edged in white, giving a ‘scalloped’ or ‘mottled’ appearance to their back (Schodde and Mason,

1999). All plumage forms seem to interbreed where their distributions overlap

(Hughes, 1982, Burton and Martin, 1976), producing a range of different intermediate forms. In eastern Australia, BB and WB plumage forms intergrade in a 200 km wide belt across the south-east (Burton and Martin, 1976). The area over which varied western magpies intergrade with western BBs is even larger and was estimated by Schodde and Mason (1999) to be up to 500 km wide. It is predominately the length of the colours in the saddle which mark intermediate forms; half way between a black back and a white back male produces a bird in which the top half of the back saddle is black whilst the bottom half of the saddle is white. The size of these ‘bands’ of black and white on the saddle seem to be a continuous trait, and patterns of inheritance of this trait are currently being investigated in a long-term study site in the eastern hybrid zone.

MC1R & Phylogeography

Figure 3.1: Magpie plumage variants and their distribution across the Australian continent. Black backed forms inhabit black shaded areas, white backed forms pale grey areas, and intermediate forms are found in the medium grey shaded areas. The varied plumage form inhabits the shaded south-west region of the continent. White areas of the continent are not necessarily limits of the distribution and may simply represent regions of very low magpie density. Letters indicate subspecies: (A) C.t. longirostris, (B) C.t. eylandtensis, (C) C.t. terrareginae, (D) C.t. tibicen, (E) Hybrid C.t. terrareginae and C.t. tyrannica, (F) C. t. tyrannica, (G) C.t. hypoleuca, (H) C.t. telonocua, (I) C.t. dorsalis, (J) Hybrid C.t. dorsalis and C.t. longirostris. Map adapted from Toon and Hughes (2008) after Schodde and Mason (1999).

Recent molecular phylogenies based on several nuclear genes indicate the species is more closely related to Currawongs (Strepera) and Butcherbirds

(Cracticus) than any other bird species (Barker et al., 2004). Magpies are widely distributed across the Australian mainland (Schodde and Mason, 1999).

They inhabit a range of woodland environments and are uncommon in dense forest or extremely arid regions

In approaching the investigation of the genetic basis of back-colour variation in magpies, the nature of the phenotypic variation was considered. A number of studies have implied that back-colour seems to be a heritable, non-plastic genetic trait within this species (Hughes et al., 2001, Hughes, 1982). 80 MC1R & Phylogeography

Observations at a long-term study site indicate that individuals seem to maintain the same back-colour throughout their adult lifetime and do not exhibit seasonal or dietary variation in plumage colour (unpublished data). It seems likely that only a small number of genes are involved in this back-colour variation (Hughes and Mather, (1980) in Hughes, 1982)), and this indicates that a candidate gene approach should be well-suited to this study, especially as a number of candidate colour genes have been successfully associated with plumage variation in a number of bird species in recent decades.

Two studies of magpie phylogeography have demonstrated that genetic variation in the mtDNA control region is not concordant with back colour across the eastern (Hughes et al., 2001) and western (Toon et al., 2003) areas of their distribution, inclusive of intermediate zones. Hughes et al (2001) have suggested a mechanism, similar to that first outlined in Kallioinen et al (1995), in which natural selection for different back colours in different habitats counteracts effects of gene flow between BB and WB populations to explain the lack of partitioning of neutral gene markers between back colours. Hughes et al. (2001) suggested that assortative mating, or a preference for a particular back colour in a mate, together with differential success across different habitats could drive divergent selection on either side of back colour hybrid zones, such that natural selection may favour black backs, whilst sexual selection may favour white backed forms.

Black backed magpie populations tend to inhabit relatively open woodlands, whilst southern white backed populations are generally found in more thickly vegetated areas. White backs are more conspicuous in open woodland,

81 MC1R & Phylogeography especially in the UV spectrum, and it has been suggested they are more vulnerable to predation in such environments than black backed individuals

(Hughes et al., 2001). A study by Hughes et al. (2002) showed that in territories with nests in forested areas, white backed males produced more fledglings than black backed males, while black backed males produced more fledglings in territories where nests were located in more open woodland environments. The same paper found no evidence for assortative mating, however relied only on social parentage observations. A more recent study which genetically determined parentage also found little evidence for assortative mating and no significant preference for brighter white backed males, as well as no discernible increase or decrease in extra-pair fledglings produced by females with high relatedness to their social males (Hughes et al., 2011). In this paper, Hughes et al.(2011) propose an alternative hypothesis to sexual selection: that the higher bacterial resistance of black feathers in the hotter and more humid northern parts of the continent may give these individuals an advantage in these regions, and white backs occur in the south simply as the weather renders this higher bacterial resistance unnecessary and melanin may be costly to produce

(Hughes et al., 2011).

In direct contrast with the strongly northern and southern groupings of plumage forms, magpie populations from eastern Australia and western Australia have been found to be strongly divergent from one another, both in the south (Baker et al., 2000, Hughes et al., 2001, Toon et al., 2003) and the north (Toon, 2007) based on mtDNA data. Molecular clock estimates place the divergence of eastern and western clades during the Pleistocene, approximately 36, 000 years ago (Toon et al., 2007). These genetic clades geographically and

MC1R & Phylogeography temporally correspond with arid barriers which may have restricted dispersal between eastern and western populations and/or restricted populations to refugial areas during these periods (Toon et al., 2007). Of these arid barriers, the Northern Carpentarian and Canning seem to be implicated in population structuring in the north, whilst in the south of the continent, the Nullabor-Eyrean arid barrier is likely to have significantly restricted gene flow and dispersal between eastern and western populations of the magpie (Toon et al., 2007).

Tasmanian populations seem to have diverged only relatively recently from mainland populations, well after the western and eastern populations had begun to diverge (Hughes et al., 2001, Toon et al., 2007). It is estimated that the isolation of Tasmanian from mainland populations occurred approximately

16,000 years ago, and this timing neatly dovetails with known geological changes which may explain this isolation- it was approximately around this time when sea levels rose and filled Bass Strait, cutting off the land bridge which had temporarily linked Tasmania to the mainland during the last glacial cycle (Toon et al., 2007, Chappell and Shackleton, 1986).

Mitochondrial DNA also indicates that eastern magpie populations may have undergone an expansion more recently than the east-west divergence, with eastern populations spreading further inland and north (Toon et al., 2007).

Whilst mtDNA data has consistently supported an east-west split between lineages, nuclear DNA reveals a slightly different and interesting story: microsatellite analysis has indicated secondary recontact between eastern and western populations in northern Australia (Toon et al., 2007). Male-biased dispersal, such as that observed in magpies by Veltman and Carrick (1990)

MC1R & Phylogeography should theoretically result in nuclear gene flow preceding mtDNA gene flow; this would account for nuclear markers detecting an event of secondary recontact before the signature is detected in mtDNA. The mtDNA divide of eastern and western populations was only weakly supported by this microsatellite data, indicating further investigation of these groupings is desirable (Toon et al.,

2007).

Investigating the differing roles both selective and historical demographic processes may play in shaping variation within the MC1R gene underlies the main thrust of this study. Genetic structure is generally considered as the distribution of genetic variance that results from a range of factors, including genetic drift, mutation, migration, and selection. Phylogeographic analyses examining the magpie’s distribution have, to date, used only putatively neutral genes including the mitochondrial control region and a number of microsatellite loci (Hughes et al., 2001, Toon et al., 2003, Toon et al., 2007). These types of markers are useful for inferring demographic processes and history which may have led to the current distribution of a particular species. However, utilising a gene that is putatively under some form of selection pressure offers the added opportunity of exploring a marker that may be responsible for an adaptive response to differing environments in the form of plumage colouration

(Hoffmann and Willi, 2008). Different haplotype sets, allele frequencies or some form of genetic structure might be expected in a given candidate gene subject to selection pressures for different variants of a phenotypic trait, and the phylogeographic metrics from variation in such a gene can be compared to neutral markers to scrutinise selection processes further (Piertney and Webster,

2010).

MC1R & Phylogeography

It is expected that patterns of major genetic groupings inferred from gene regions of disparate evolutionary histories (maternal, neutral and functional) ought to have significantly different topologies. The MC1R gene is so inextricably involved in pigmentation pathways that regardless of whether or not structural changes in the coding region of the gene directly determine plumage morphs in the magpie, phylogeographic analyses of this gene has the potential to unearth genetic structure that may expose patterns of pigmentation selection history in this species.

In this study we examined sequence variation in the coding region of the MC1R gene to test the hypothesis that back colour variation is associated with MC1R variation in C. tibicen, in that an amino acid change(s) may lead to further melanisation of certain patches on their back, as well as characterise the gene in magpies, and answer questions about phylogeographic structure in a nuclear gene potentially under selection pressure. It was hypothesised that MC1R variation would be associated with magpie back colour variation, and if so, this mutation or set of mutations may be similar to others found to be associated with pigmentation changes in other animal species in terms of their location within the MC1R gene.

3.3 MATERIAL AND METHODS

3.3.1 Sampling

Magpie DNA samples used in this project were collected by a number of different researchers over the last few decades. These samples are both whole

MC1R & Phylogeography blood and blood in a range of different buffers and are frozen at -80 0C. The samples have been collected from across most of the Australian continent, and include individuals of each sub-species, back-colour, and sex from most of the

Australian magpies’ current distribution. Both forward and reverse MC1R sequences from 100 magpie individuals were sequenced and included approximately even numbers of representatives of all back-colour variants and sub-species, from 20 sites around Australia (Table 3.1).

Table 3.1: Number of sampled haplotypes, according to sub-species, site location and back colour.

Number haplotypes Site name Sub-species Back colour morph Coordinates sampled Charters Tow ers C. tibicen terraereginae Black back 10 20°04'S 146°15'E

Hydeaw ay Bay C. tibicen terraereginae Black back 10 20°05'S 148°29'E Dubbo C. tibicen terraereginae/ tibicen Black back 8 32°14'S 148°36'E

Brisbane C. tibicen tibicen Black back 8 27°28'S 153°01'E

Grafton C. tibicen tibicen Black back 8 29°40'S 152°56'E Ouyen C. tibicen tyrannica/terraereginae White back/Hybrid 16 35°04'S 142°21'E

Seymour C. tibicen tyrannica/terraereginae Black back/White back/Hybrid 6 37°01'S 145°09'E Horsham C. tibicen tyrannica/terraereginae Hybrid 4 36°42'S 142°13'E Row sley C. tibicen tyrannica White back 14 37°43'S 144°24'E

Phillip Island C. tibicen tyrannica White back 10 38°27'S 145°15'E Tasmania C. tibicen hypoleuca White back 8 41°26'S 147°08'E Nullabor C. tibicen telonocua White back 12 32°07'S 133°40'E

Esperance C. tibicen dorsalis Varied form 16 33°51'S 121°53'E Albury C. tibicen dorsalis Varied form 10 35°00'S 117°53'E

Busselton C. tibicen dorsalis Varied form 10 33°39'S 115°21'E

Mandurah C. tibicen dorsalis Varied form 8 32°31'S 115°44'E Pilbara C. tibicen dorsalis/longirostris Black back/Varied 16 27°05'S 116°09'E & 26°53'S 115°57'E

& 25°33'S 116°39'E & 27°26'S 117°53'E Nth. WA C. tibicen longirostris Black back 6 22°41'S 117°47'E & 24°18'S 116°54'E Kimberley C. tibicen eylandtensis Black back 6 16°49'S 124°55'E

Northern Territory C. tibicen eylandtensis Black back 10 14°55'S 133°04'E & 19°42'S 135°49'E

3.3.2 MC1R genotyping and analysis

Whole genomic DNA was extracted from blood in buffer using a modified version of the simplified Rapid Method (RM) described in (Lahiri and Schnabel,

1993) using NP-40 as the detergent and a MgCl2 concentration of 4 mM. A

MC1R & Phylogeography reduced volume of 50 μl of blood in buffer yielded an adequate volume of DNA for subsequent analysis.

An 861 bp fragment of the MC1R gene that encompassed all sites known to be associated with colour variation in birds was amplified using the primer

MSHR72 (ATGCCAGTGAGGGCAACCA) (Mundy et al., 2004) and a reverse primer designed specifically for C. tibicen, Ana-R

(TGTAGAGCACCAGCATGAGG) developed to overcome the problem of non- specific amplification.

PCR reactions were carried out in 10 μl reaction volumes of 0.3 mM each of forward and reverse primers, 0.2 mM of dNTPs (Bioline), 2 mM MgCl2 (Fisher

Biotech), 2 μl 10 x reaction buffer (Fisher), 3.76 μl ddH2O, 3.0 μl of extracted template DNA and 0.2U Thermus aquaticus DNA Taq polymerase (Fisher).

Cycling conditions were as follows: initial denaturation at 94 °C for 5 minutes,

42 cycles of 30 s at 94 °C, 45 s at 67 °C, and 45 s at 72 °C, followed by a final extension step of 30 min at 72 °C before samples were held at 4 °C. The amplified product was then purified with exo-sap (Fermentas)

Purified PCR products were used in a sequencing reaction of 10 μl volume which contained: 2.0 μl of 5x sequencing buffer (Applied Biosystems), 2.0 μl Big

Dye Terminator Mix 3.1 (Applied Biosystems), 4.68 μl ddH2O, 0.32 mM of the respective primer, and 1.0 μl of purified PCR product. These reactants were subjected to a thermal cycling program consisting of a hold at 96 °C for 1 minute, followed by 30 cycles of 96 °C for 10 seconds, 50 °C for 5 seconds, and

87 MC1R & Phylogeography

60 °C for 4 minutes, before being held at 4 °C until clean-up of the sequencing reaction was carried out.

Sequencing reaction products were washed twice with 70% ethanol before the pellet was dried out completely in a vacuum bell. Both strands were then directly sequenced with PCR primers using BigDye 3.1 terminator chemistry on an automated sequencing machine (Applied Biosystems 3130x1).

Forward and reverse sequences were aligned and manually edited in

Sequencher 4.1 (Gene Codes Corporation 2000) and sequences deposited in

Genbank (Accession numbers JN172943-JN172967). Heterozygous bases were common and a double chromatograph peak of equal height on both strands was considered sufficient evidence to deem an individual heterozygous at that site. Phase ver. 2.1.1 (Stephens et al., 2001, Stephens and Scheet,

2005) was used to statistically infer haplotype phase within a Bayesian framework; those individuals for which phase could not be determined with high levels of confidence (>75%) were then experimentally cloned to unambiguously assign haplotypes. Phase thresholds >60% have been shown to be relatively robust; lowering thresholds from a highly stringent 90% to 60% has been found to reduce the number of unresolved haplotype pairs with very little increase in false positives in a number of studies (Garrick et al., 2010, Harrigan et al.,

2008). Gel-purified PCR product was cloned using the TOPO TA cloning kit

(Invitrogen). After colonies were cultured overnight in LB broth, 24 colonies of each individual were randomly picked and amplified with M13F and M13R primers (Invitrogen), then run on 0.8% agarose gels to detect successful MC1R inserts. Eight of these inserts were then sequenced in both directions for each

MC1R & Phylogeography individual. Multiple inserts were sequenced for each individual as PCR error is common when amplifying from a single clone, and both the misincorporation of nucleotides and PCR recombination can lead to inaccuracies in resultant sequences (Paabo and Wilson, 1988). Comparison of sequences of multiple inserts and directly sequenced genomic DNA, as in Harrigan et al. (2008), enabled distinction of misincorporated nucleotides and sites of PCR recombination.

Magpie sequences were aligned to bird MC1R sequences from ten additional species obtained on Genbank (details and GenBank accession numbers in

Table 3.3) and SNPs found to be associated with melanism in other bird species were scrutinised in magpies. This alignment also enabled sequences to be translated into amino acid sequences and checked for unexpected indels and stop codons that often flag the unintentional amplification of psuedogenes, using DnaSP version 4.50 (Rozas et al., 2003).

Putative transmembrane helices were calculated in TMHMM 2.0 and drawn to map the magpie MC1R gene and the relative position of amino acid changes

(Figure 3.3), and a haplotype network was constructed in TCS (Clement et al.,

2000). A number of separate AMOVAs were carried out to test which factors best explained the genetic variation observed: back-colour, populations or the historical east-west divide suggested by mitochondrial data of previous studies

(Baker et al., 2000, Toon et al., 2007, Toon et al., 2003). Fu’s Fs and Tajima’s

D were calculated to examine the possibility of selection at the MC1R locus. A

McDonald-Kreitman test for selection was also carried out utilising MC1R

89 MC1R & Phylogeography sequences of another passerine bird, the carrion crow (Corvus corone corone) for comparison (GenBank Accession no. EU348721-630).

3.4 RESULTS

3.4.1 Analysis of genetic variation at MC1R

A total of 861 base pairs of the MC1R gene were sequenced for 100 magpies, from which 83 individuals were initially able to be confidently assigned haplotypes using the statistical Bayesian method implemented in Phase 2.1.1.

The haplotypes of a further 11 individuals were determined experimentally by cloning, and the addition of these known haplotypes into the Phase analysis improved the confidence in the estimation of haplotypes of the remaining six individuals to the extent that an additional four of these were then at sufficient confidence levels (>75%) to be included, yielding a total of 98 individuals utilised in downstream analyses. No unexpected indels or stop codons were encountered; it is therefore likely MC1R sequences were not confounded by the amplification of pseudogenes. Within these 196 haplotype sequences, 19 sites were variable, of these, 15 were synonymous substitutions and four were non- synonymous substitutions.

3.4.2 Phylogeography

The nature of selection at MC1R was examined using several different tests of neutrality and selection. Tajima’s D value was significantly negative at -1.78

(P=0.016). A negative Tajima’s D can be indicative of purifying selection or population size expansions, both of which characteristically produce an excess

MC1R & Phylogeography of low frequency polymorphisms. The Fs value for magpie MC1R was also negative (-24.22) and highly significant (p<0.0001). Fu’s Fs has been shown to have a great deal of power to detect recent demographic expansions (Ramos-

Onsins and Rozas, 2002).

A McDonald-Kreitman test comparing MC1R sequences of 10 haplotypes of carrion crow (Corvus corone corone) with magpie haplotypes was not statistically significant (p=0.18), and the Neutrality Index score of 0.45 sat firmly in the range expected under neutrality, providing no evidence of selection on

MC1R in magpies. This test uses the ratio of synonymous to non-synonymous mutations within a species of interest and between that species and another closely related species at the same loci to evaluate the loci’s selective neutrality

(Egea et al., 2008, McDonald and Kreitman, 1991)

The 861 bp fragment of MC1R screened across 196 magpie alleles identified 25 unique haplotypes (Figure 3.2). The haplotype network revealed a star-shaped phylogeny, featuring a dominant central haplotype. This central and common haplotype (57% of individuals) occurred in 19 of the 20 sampled sites from all around the Australian continent, and was also the most abundant haplotype at all but three of these sites. This dominant haplotype was common to every subspecies and back-colour sampled.

MC1R & Phylogeography

Figure 3.2: Parsimony network of MC1R sequence data generated in TCS, showing relationships between 25 haplotypes. Frequency of haplotypes is indicated by circle size; circle fill colour indicates sub-species. Connecting lines indicate a mutation of one base pair and dashed thick black lines mark sites of non-synonymous amino acid substitutions. Back colour of each sub-species is indicated in legend: BB=black back, WB=white back, Var=western varied back.

The haplotype network (Figure 3.2) of MC1R illustrates no readily apparent geographic structure: no particular population, region or geographically bound subspecies was restricted to a specific group of haplotypes. Global AMOVAs testing for structure between both populations and eastern and western groups found a statistically significant level of genetic structure among geographically delineated groups. Fst values of these global AMOVAs indicated differences between populations and east/west accounted for 8.11% and 4.81% of genetic variation in MC1R respectively (Table 3.2). These results indicate that the

MC1R is weakly geographically structured in magpies, and this geographical structure accounts for slightly more of the genetic variance in this gene than plumage differences or subspecies designations (2.67% and 4.05%

MC1R & Phylogeography respectively), though these were also found to have statistically significant levels of genetic structure at the 1% level.

Table 3.2: Analysis of molecular variance (AMOVA) results and tests of neutrality and selection.

AMOVA tests Pairwise FST

by population 0.081 P = 0.000 by back colour (WB v BB v Varied) 0.026 P = 0.002 by East-West mtDNA division 0.048 P = 0.000 by sub-species 0.040 P = 0.000 by East-West mtDNA division (Sth only) 0.051 P = 0.001 by population without Tasmania 0.060 P = 0.000

Neutrality and selection tests

Fu's Fs Fs = -24.222 P = 0.000 Tajima's D D = -1.775 P = 0.016 McDonald-Kreitman Neutrality P = 0.18 (Fisher's exact 2-tailed) Index = 0.45

3.4.3 MC1R as a candidate gene for magpie plumage differences

The 861 base pairs sequenced represent the majority of the MC1R gene, which varies slightly in length between species but is generally close to 954 bp in vertebrates (Wlasiuk and Nachman, 2007). The magpie MC1R gene is highly similar to MC1R regions characterised in other birds, with 7 trans-membrane domains. The 861 base pairs correspond to amino acids 21-306 of the Gallus gallus MC1R gene, and all amino acids discussed in this study are numbered after this model species for comparability and convenience.

MC1R & Phylogeography

Magpie non-synonymous substitutions Q N S S A G V T A Change to more melanic phenotype S G N G30S C Q V M R G V L PS L G I H N P D V P T F I E R C C T P L H T C N40S L Chestnut- N M bellied V F E L D monarch C S N I F D119N L M S Y T L F L F L M I I Y I L I L H S Y E M D T L L L F N V L L I C F G A L C I L C F L I L S N S V F I G F L V I S S A G P I S S S V T F W I V L V V V S F L G192S C I C E L S S A F F N N L M L M L V S S D L I W V D V I L V L S S A L G L P V A F L G M M L L I L A A V C V T L V T L Y I I C I F I V Y I T A F L A V V A H red- V footed A K Y banananaquit D R M boobies R F H207R N Y chicken R Q G S L A K Q R T Japanese quail Y L L P E92K T L E N I S A227T T L H S T M A G L R T A R I R I R G R H P lesser snow goose L Y H S H articskua red-footed boobies F Y A P R230H V85M H K L S Q Q H S I S

Figure 3.3: Partial amino acid sequence of the magpie MC1R gene (numbered after Gallus gallus). SNPs associated with plumage colour changes in other bird species are indicated by black circles representing eumelanism. Boxed amino acids indicate non-synonymous changes in magpies. The shaded area represents the transmembrane region: above this is extracellular and below is intracellular.

Amino acid substitutions within MC1R that are known to be associated with plumage variation in other bird species, including the well-known Glu92Lys change linked to plumage changes in chickens, quails and bananaquits, were scrutinized in magpie sequences (Table 3.3). Across all of these ‘candidate amino acids’ all magpies assayed were invariable. The amino acids present in magpies at these positions in the gene were a mix of those associated with both melanic types and non-melanic types in other bird species. MC1R haplotype(s) did not show a perfect or strong association with back-colour phenotype.

Instead, all back-colours were dominated by the two most common haplotypes in similar frequencies.

MC1R & Phylogeography

Table 3.3: Comparison of MC1R amino acids across a number of bird species at SNP sites associated with plumage colour change. Data for species other than magpie sourced from sequences deposited in Genbank by a number of authors. Dashes indicate missing data; asterisks indicate consensus with the top row. Gene numbered after Gallus gallus.

Amino acid Common name Species Plumage form Accession # 85 92 119 207 230 Chicken Gallus gallus Red Junglefowl, e+ AY220303 Val Glu Asp His Arg White Leghorn, E AY220304 *** Lys *** *** *** Buttercup, ebc AY220305 *** Lys *** *** *** Japanese quail Coturnix japonica Extended brown, E DQ395089 *** Lys *** *** *** Wild-type, e+ DQ395091 *** *** *** *** *** Bananaquit Coereba flaveola Melanic morph AF362605 *** Lys *** *** *** Yellow morph AF362576 *** *** *** *** *** Lesser snow geese Anser c. caerulescens Blue phenotype AY521183 Met *** *** *** *** White phenotype AY521184 *** *** *** *** *** Red-footed boobies Sula sula Brown N/A Met - *** Arg - White N/A *** - *** His - Arctic skua Stercorarius parasiticus Pale AY521215 *** *** *** *** *** Dark AY521217 *** *** *** *** His Chestnut-bellied monarch Monarcha castaneiventris Chestnut-bellied N/A - - Asp - - Black/blue N/A - - Asn - - Old World leaf warbler Phylloscopus collybita AY308747 *** *** *** *** *** Blue-crowned manakin Lepidothrix coronata DQ388310 *** *** *** *** His Rosy finch Leucosticte arctoa FJ547769 *** *** *** *** *** Carrion/Hooded crow Corvus corone EU348730 *** *** *** *** *** Australian magpie Cracticus tibicen Black-back JN172954 *** *** *** *** *** White-back JN172965 *** *** *** *** *** Varied form JN172957 *** *** *** *** ***

The results of a number of separate AMOVAs suggest that more MC1R variation was explained by variation within individual populations than any other geographical, taxonomic or phenotypic grouping (Table 3.2). In pairwise estimates among back colours, only 2.67% of variance was explained by differences between different plumage types, while differences between populations, sub-species and eastern and western mtDNA clades accounted for

8.11%, 4.05% and 4.81% respectively.

The four non-synonymous substitutions identified (Gly30Ser, Asn40Ser,

Gly192Ser, Ala227Thr) were mapped onto the haplotype network (Figure 3.2), as they have the potential to be of functional significance, regardless of their lack of association with back-colour. These changes do not seem to have

95 MC1R & Phylogeography particular associations with sub-species or geographic location (excepting the singletons).

3.5 DISCUSSION

3.5.1 MC1R and plumage variation

Although variation in the MC1R gene has been linked with plumage variability across many other bird species (Andersson, 2003, Baiao et al., 2007, Mundy et al., 2004, Nadeau et al., 2006, Takeuchi et al., 1996, Theron et al., 2001,

Doucet et al., 2004, Uy et al., 2009), back-colour plumage in Australian magpies does not seem likely to be determined by variation in the coding region of this gene. No single allele or set of alleles was found to be exclusive to any plumage type; likewise no geographically defined group or sub-species could be delineated based on MC1R variation.

A Glu92Lys mutation has been implicated in pigmentation changes in bananaquits (Theron et al., 2001), Japanese quail (Nadeau et al., 2006), chickens (Kerje et al., 2003, Takeuchi et al., 1996) and has been shown to lead to constitutive activation of MC1R during in vitro experiments of both chickens and mice (Robbins et al., 1993, Ling et al., 2003). At this position, none of the sampled magpies had this mutation, but rather both alleles in all individuals coded for Glutamic acid, the amino acid of less melanic forms of Japanese quail and bananaquit. Other mutations in MC1R posited to lead to plumage changes in other bird species were all non-variable across all magpie individuals, and the amino acid present in magpies at each of these sites was not consistently either the supposed melanic or non-melanic form (Table 3.3).

MC1R & Phylogeography

The sequenced fragment encompassed all but one of the sites that have been associated with colour variation in birds, and includes all transmembrane domains (Figure 3.3). As the entire length of the MC1R gene has not been sequenced, it remains possible that structural variation within these unsequenced portions of MC1R may be associated with the back-colour of magpies. However this is less likely, given most SNPs associated with colour variation in birds have been mapped to transmembrane and cytoplasmic regions of MC1R (Cheviron et al., 2006).

This finding is interesting in the context of the numerous other studies of MC1R and colour variation in bird species. A large number of bird species have been found to have associations between MC1R variation and colour phenotype

(Andersson, 2003, Baiao et al., 2007, Mundy et al., 2004, Nadeau et al., 2006,

Takeuchi et al., 1996, Theron et al., 2001, Uy et al., 2009). However as more wild species of birds with variable plumage colours are now being screened for segregating mutations in the MC1R gene, a number of species have been identified in which no association between phenotype and MC1R genotype has been detected. These include old world leaf-warblers (MacDougall-Shackleton et al., 2003), blue-crowned manakins (Cheviron et al., 2006), rosy-finches

(Drovetski et al., 2008) and carrion/hooded crows (Haas et al., 2008). Plumage variants within the white-winged fairy-wren were initially thought to be associated with MC1R variants (Doucet et al., 2004), however a more recent and much broader study conclusively demonstrated the variation in question in the gene was not correlated with colour variation in this and several other fairy- wren species (Driskell et al., 2010). All of these studies suggest that further

MC1R & Phylogeography investigation of other structural candidate colour genes is needed, and some also propose that regulatory processes affecting the expression of MC1R and other colour genes are deserving of further research, especially when plumage colours are variable seasonally or across life-history stages of a species

(Cheviron et al., 2006, MacDougall-Shackleton et al., 2003, Drovetski et al.,

2008).

Haas et al. (2008) sequenced MC1R in the hooded crow, carrion crow and a range of hybrids that occur along a long corridor where their distributions overlap in Europe, and found no association between MC1R variation and plumage variation. The plumage of the carrion crow is entirely black, whilst the hooded crow only has a black head, neck, wings, thighs and tail feathers, the remainder of their feathers are a slate grey, resulting in a sharp contrast between areas of grey and black. Plumage of hybrids between the two species varies along a continuum between the two extremes (Haas et al., 2008). Both

Haas et al. (2008) and MacDougall-Shackleton et al. (2003) discuss the nature of the colour variation as a likely important factor; they suggest that MC1R is less likely to be the genetic basis for fine-scale patterning changes and/or discrete plumage patterning, as is the case in magpies. Rather, variation in

MC1R seems to be associated with whole-body colour changes or gradational patterning of colour. Australian magpies bear a striking resemblance to hooded crows in terms of the discrete nature of the colour variation on their bodies, and the lack of association between MC1R variation and back-colour in magpies found in this study may add further weight to this hypothesis.

MC1R & Phylogeography

A number of other structural and regulatory candidate colour genes, such as

TYRP1, ENDRB2, PMEL17, ASIP and SLC45A2 have also been linked to plumage colour variation in other bird species (Kerje et al., 2004, Gunnarsson et al., 2007, Miwa et al., 2007, Nadeau et al., 2007, Hiragaki et al., 2008), and it is suggested future studies of magpie pigmentation will benefit from further investigation of such genes.

3.5.2 Phylogeography and selection at MC1R

Phylogeographic structure ought to be substantially weaker in nuclear genes than mitochondrial markers, a consequence of the lower effective population size (Birky et al., 1989, Moore, 1995), thus it is not remarkable that MC1R shows weak structuring, despite the fact that high levels of structure have been found in the mitochondrial control region of this species (Toon et al., 2007). In the same study, nuclear microsatellite markers revealed only weak structuring, a level of magnitude below that observed in the mitochondrial control region, although there was enough power to detect the divergence between eastern and western haplotypes first delineated by mtDNA (Toon et al., 2007).

Structural coding regions such as MC1R might be expected to have inherently less variability and evolve more slowly than microsatellites, as they are constrained by selective pressures; mutations in such regions can be lethal if gene function is disrupted e.g. lethal yellow mutation in mice (Michaud et al.,

1993), whereas the majority of mutations in microsatellite regions should have no functional consequences and be selectively neutral (Schlotterer, 2000).

Star-like haplotype networks are generally recognized as predictive of either an expansion event following a bottleneck or a selective sweep in the recent 99

MC1R & Phylogeography evolutionary history of the species (Slatkin and Hudson, 1991). Thus the star- like MC1R network observed (Figure 3.2) may indicate one or both of these processes have influenced the trajectory of Australian magpie populations.

The significant negative Tajima’s D value, indicating an excess of low frequency polymorphisms compared to neutral expectations, is also suggestive of either a demographic population expansion or past selective sweeps (Tajima, 1989), or alternatively may indicate that purifying selection has acted on the MC1R gene.

This negative Tajima’s D value also indicates the gene is unlikely to have undergone a recent bottleneck event for which a significantly positive value of this test would be expected. Fu’s Fs value was also highly significant, a result strongly suggestive of recent population expansion or genetic hitchhiking (Fu,

1997).

The weight of evidence indicates magpies experienced a population expansion at some point in the species’ recent past: The significant Tajima’s D value, highly significant Fu’s FS value and star-shaped haplotype network all indicate either a population expansion or selective process has occurred in the recent evolutionary history of magpies. Mitochondrial DNA markers also show some evidence of this past expansion (Toon et al., 2007). However, as the results of a McDonald-Kreitman test could not reject neutrality of MC1R, it seems this gene is unlikely to be currently under selection.

The balance of evidence indicates that magpies within Australia seem to have undergone a change in population size which is more likely to have been an expansion than a contraction. A number of authors have suggested that

100

MC1R & Phylogeography

European settlement may have led to an increase and/or expansion of magpies as the creation of pastoral and urban lands transformed large tracts of Australia into high quality magpie habitat (Campbell, 1929, Schodde and Mason, 1999).

Toon et al. (Toon, 2007) found evidence to support this contemporary gene flow in north-western Australia with microsatellites, which are rapidly evolving nuclear markers (Ellegren, 2007). However, the patterns observed in MC1R gene sequences will pre-date this anthropogenic influence significantly.

Climatic fluctuations during the Pleistocene have been linked to geographically delineated genetic structure between magpie populations e.g. the eastern and western divergence dated at c. 36,000 years ago (Toon et al., 2007), putatively through periodic restrictions to dispersal and gene flow between groups posed by arid barriers such as the Nullabor-Eyrean in South Australia.

It is suggested that the signature of population expansion found in both this study of a nuclear gene, and that found in a mitochondrial marker (though less significantly) by Toon et al. (2007), seem likely to be related to the increase in available habitat suitable for magpies following the Last Glacial Maximum in the late Pleistocene, and the subsequent expansion of magpies out of refugial habitats.

Our findings indicate that MC1R is not likely to be under selection in the

Australian magpie, and implicate an expansion event in the recent evolutionary history of the species. Sequence variation in the majority of coding region of the MC1R gene screened was not found to be associated with back colour variation, and it is anticipated that future studies which screen other likely

101

MC1R & Phylogeography candidate colour genes and regulatory elements will help further our understanding of plumage colour evolution in this species.

3.6 FUNDING

This work was supported by Griffith University. A. E. Dobson was supported by a Griffith School of Environment Postgraduate Research Scholarship.

3.7 ACKNOWLEDGEMENTS

Sampling was carried out by Jane Hughes, Peter Mather, Alicia Toon, Andrew

Baker, Kate Durrant, and a large number of volunteers. We would like to thank

Steve Smith for his contribution to laboratory work, and two anonymous reviewers for their helpful comments on this manuscript. Samples were collected under Western Australian Department of Conservation and Land

Management permit SF003903, Queensland Parks and Wildlife permit

W4/002670/01/SAA, Northern Territory Parks and Wildlife Commission permit

18145, Victorian Wildlife permit 10003058 and animal ethics protocol

AES/16/04/AEC.

102

Candidate colour genes

CHAPTER 4: CANDIDATE GENES FOR PLUMAGE DIFFERENCES IN THE AUSTRALIAN MAGPIE, CRACTICUS TIBICEN

4.1 INTRODUCTION

Studies of animal pigmentation are among the oldest of scientific endeavours

(Hooke et al., 1665, Darwin and Wallace, 1858). The incredible diversity of colour and pattern across the animal kingdom is compelling in itself, yet the biological questions posed by such conspicuous variation are even more interesting. The similarity of colour and pattern observed between some distantly related species is at odds with the striking differences that often exist between different populations, or sexes within the same species.

Examining the genetics underlying such colour variation was, for the last century, limited to breeding experiments and the construction of large pedigrees of domestic and laboratory animals. Studies of this period concentrated on heritability, inheritance patterns, gene linkage, correlations between phenotype and environment, and mutation rates (Hoekstra, 2006b). Current molecular technology has progressed to the point where regular sequencing of the entire genome of non-model species of interest is often feasible (Ellegren, 2008,

Kircher and Kelso, 2010, Moura et al. 2014). These genetic advances mean interesting questions about the underlying basis of pigmentation traits can now be addressed, such as: Which genes are involved? How many different genes are involved? What is the mechanism by which these genetic changes result in pigmentary changes? Do such genes have pleiotropic effects? We can now examine on a molecular scale linkages between pigmentation genes and genes

103

Candidate colour genes involved in traits potentially under heavy selection pressure (e.g. immunity, physiochemical tolerances), and pick apart the confounding effects of selection on one or the other of these traits.

This type of work also allows comparisons of the pigmentation genes involved and their mechanisms of action, within and across species and animal groups to investigate evolutionary questions about traits which may be shared by descent or a result of parallel/convergent evolution, and establish if such traits are determined by structural genetic variation, or are regulatory or developmentally controlled.

The pigmentation pathway is highly complex; over 171 mouse coat colour genes have already been cloned and described; and another 207 remain to be cloned (As of October 2011 http://www.espcr.org/micemut/). Genes directly involved in melanin production and/or regulation of the relative amounts of eumelanin and pheomelanin, of which there are approximately 20 (Ito and

Wakamatsu, 2011) seem most promising as candidates to explain melanin- based colour variation in non-model species.

The Australian magpie (Cracticus tibicen) is a large passerine bird widely distributed across mainland Australia, Tasmania and a number of neighbouring islands, including Papua New Guinea, and New Zealand (Schodde and Mason,

1999). The species prefers open woodland and pastoral environments, and is also common in urbanised areas throughout the continent (Campbell, 1929).

However, extremely arid regions, dense forests and rainforests tend to have low magpie densities. The species is group-living, and whilst helping behaviours

104

Candidate colour genes have been extensively documented within familial territories (Hughes et al.,

1996, Finn and Hughes, 2001), rates of extra-group paternity measured in several populations have been found to be relatively high (Hughes et al., 2003,

Durrant and Hughes, 2005, Hughes et al., 2011), an interesting dichotomy, particularly as groups defend their territories from interlopers vigorously

(Farabaugh et al., 1992). The species was previously known as Gymnorhina tibicen; however, recent molecular work has indicated a very close relationship with butcherbirds (Cracticus) and they have been renamed accordingly (Barker et al., 2004, Kearns, 2011).

A number of different plumage morphs were previously designated as separate species based on plumage, bill length, wingspan and size variation; however, these are now grouped into one species comprising eight sub-species, and genetic work confirms neither plumage morphs nor sub-species are genetically divergent enough to be classified as separate species (Hughes et al., 2001, A.

Dobson, unpublished data). Plumage morphs differ primarily in the quantity and distribution of black and white dorsal pigmentation, and have been broadly designated as “white backed”, “black backed” and “varied” forms, although varied forms are often included in the “white backed” category (see Figure 4.1).

Different sub-species intergrade where their distributions overlap, producing intermediate plumage forms with differing proportions of black and white dorsal pigmentation (Burton and Martin, 1976, Hughes, 1982). These ‘hybrid’ zones contain both parental and intermediate back colour forms and are relatively wide zones: Schodde and Mason (1999) estimate that western black backs intergrade with western varied forms over a distance of 500 km, while eastern

105

Candidate colour genes black backs and white backs hybridise across a belt approximately 200 km wide

(Burton and Martin, 1976).

The back colour variation demonstrated in the Australian magpie is thought to be genetically inherited based on social observational data (Hughes, 1982,

Hughes et al., 2001); more recently the trait has been demonstrated to have a high level of heritability in an eastern hybrid zone (Chaper 2, this study), and adult birds remain the same back colour throughout their lives. This plumage colour has not been observed to change plastically, or as a consequence of season or diet at a long term study site (J. Hughes, unpublished data).

Whilst plumage forms segregate into roughly northern and southern forms, with black backs across northern latitudes of the continent and white/varied backs in the south, several phylogeographic studies have revealed an interesting dichotomy: eastern and western populations most strongly divergent from one another based on mtDNA data (Baker et al., 2000, Hughes et al., 2001, Toon et al., 2003, Toon, 2007). This seems to indicate different back colours either evolved in situ, separately in eastern and western populations, or back colour differences pre-date this east-west divergence and are maintained by some

(likely selective) mechanism. Indeed, Hughes et al (2001) have suggested a mechanism, similar to that first outlined in Kallioinen et al. (1995), in which natural selection for different back colours in different habitats counteracts effects of gene flow between BB and WB populations to explain the lack of partitioning of neutral gene markers between back colours. The divergence of eastern and western groups in mtDNA is estimated to have occurred during the

106

Candidate colour genes

Pleistocene, approximately 36 000 years ago, according to molecular clock estimates (Toon et al., 2007).

Long lists of candidate colour genes have been identified in birds, many of which are involved in the synthesis, suppression or regulation of melanins.

When trying to elucidate the genetic basis of colour variation in a non-model species such as the Australian magpie, for which the variation seems likely to be under the control of a small number of genes (Hughes and Mather, 1980 in

Hughes, 1982), it seems probable that screening a range of candidate pigmentation genes may yield further insights. In non-model species such as the magpie, for which high quality annotated genomes are not yet available, resequencing is very resource-intensive, and candidate gene approaches offer an efficient way to investigate the genetics of coloration.

Due to the black and white nature of plumage variation in the magpie, the obvious candidate colour gene is one that has been extensively demonstrated to be associated with melanism across a wide range of animal species: the melanocortin-1 receptor (MC1R). MC1R encodes a seven-transmembrane domain G-protein coupled receptor (Mountjoy et al., 1992) expressed in melanocytes, where it plays a major role in initiation of the melanin production process and the dispersal of melanosomes through cells (Jackson, 1997).

Mutations in the coding region of this gene have a significant potential to disrupt or interfere with various aspects of pigmentation, and both gain-of-function and loss-of-function mutations associated with phenotypic colour variations identified across broad range of animal, and specifically, bird species (Roulin and

Ducrest, 2013).

107 Candidate colour genes

However, the majority of the MC1R coding region (over 90%) has been screened across a large sample of magpies from continental Australia and adjacent islands, including all sub-species, plumage morphs and sexes

(Dobson et al., 2012). No variation in this gene was found to explain the back colour variation observed in magpies. Variation in the magpie MC1R gene also did not delineate sub-species or geographically defined groups, although it did provide further support for a population expansion of magpies within Australia in the recent evolutionary history of the species (Dobson et al., 2012).

In this study we examined sequence fragments from a number of genes that had previously been identified as candidates for plumage colour differentiation, in order to test for associations with magpie back colour. The four genes screened include all three members of the tyrosinase family (TYR, TYRP1,

DCT) which all play key roles in the pigmentation and/or melanocortin pathways

(Hoekstra, 2006a), as well as the agouti-related protein (AGRP). Mutations in the tyrosinase genes have been shown to be associated with colour variation in a range of animals, and the role of AGRP in avian pigmentation is still being investigated (extensively reviewed below). It is suggested that some mutation(s) in one or more of these candidate pigmentation genes may help explain back colour variation in the Australian magpie. Such mutations, if discovered, may be similar to those identified in other bird species, both in terms of their location on the gene and the functional basis for the phenotype change they confer.

108

Candidate colour genes

4.2 CANDIDATE COLOUR GENES

4.2.1 Tyrosinase-related protein 1

The tyrosinase-related protein 1 (TYRP1 or TRP1) gene is an 8 exon melanosomal enzyme involved in the production of eumelanin. The TYRP1 gene is sex-linked in birds, occurring on the Z chromosome (Backström et al.,

2010), and the enzyme plays an important role in the stabilisation of tyrosinase

(Kobayashi and Hearing, 2007). As this gene is sex-linked, it could potentially be involved in the sexual dimorphism (back colour/pattern) exhibited across all plumage forms of the magpie, especially if magpies are not mating at random with respect to back colour. Indeed, Buggiotti (2007) found that splicing variants of the TYRP1 gene in Fidecula flycatcher species were associated with a sexually-selected plumage trait in males

A significant number of studies have identified TYRP1 associations with pigmentation variation across a wide range of animals, including humans

(Boissy et al., 1996, Manga et al., 1997), mice (Zdarsky et al., 1990, Johnson and Jackson, 1992, Javerzat and Jackson, 1998), dogs (Schmutz et al., 2002,

Schmidt-Kuntzel et al., 2005), cattle (Berryere et al., 2003), pigs (Ren et al.,

2011), Soay sheep (Gratten et al., 2007), pied and collared flycatchers

(Buggiotti, 2007), and Japanese quail (Nadeau et al., 2007).

4.2.2 Tyrosinase

The tyrosinase gene (TYR) encodes a rate-limiting melanogenic enzyme of 529 amino acids, spread across five exons in the chicken genome (Mochii et al.,

109 Candidate colour genes

1992, Sato et al., 2007). This enzyme is heavily involved in pigmentation pathways, and catalyses a suite of reactions between important proteins involved in pigmentation. Levels of tyrosinase activity are directly related to the type of melanin produced; melanocytes generally produce pheomelanin unless signalled otherwise, however, when αMSH binds to the melanocortin-1 receptor

(MC1R), elevating intracellular cAMP levels, tyrosinase is activated in certain melanocytes, initiating the synthesis of eumelanin instead of pheomelanin in these cells (Burchill et al., 1986).

SNPs within the tyrosinase gene have been associated with a large number of pigmentation changes across a range of animal species, including albino and

Himalayan mice (Kwon et al., 1989, Yokoyama et al., 1990), Burmese and

Siamese cat breeds (Schmidt-Kuntzel et al., 2005), albino and recessive white chickens (Tobita-Teramoto et al., 2000, Chang et al., 2006) and albino cattle

(Schmutz et al., 2004). Mutations in this gene cause two different forms of albinism in humans, and others are implicated in hyper pigmentation and a number of melanomas (Oetting and King, 1993, Spritz and Hearing Jr, 1994).

However, other pigmentation studies have found no association between the

TYR gene and pigmentation variants, including Scandinavian vs. Scottish willow grouse phenotypes (Skoglund and Hoglund, 2010), cream coats in domestic dogs (Schmutz and Berryere, 2007), and several cattle coat colours (Gan et al.,

2007).

110

Candidate colour genes

4.2.3 Dopachrome tautomerase

The dopachrome tautomerase (DCT) gene is also known as dopachrome delta- isomerase or the tyrosine-related protein 2 (TRP2). The DCT gene is of uncertain structure in birds, but in mice it is made up of 8 exons spread over 50 kilobases (kb) (Budd and Jackson, 1995), in humans 8 exons over 55 kb (Sturm et al., 1995), whilst in Fugu fish it has 8 exons spread over just 5.7 kb (Wang et al., 2007). In chickens, this gene is 516 amino acids in length and codes for a

Type 1 membrane protein that functions as a regulatory enzyme (April et al.,

1998). The DCT enzyme plays several important roles in melanin synthesis and the regulation of eumelanin and pheomelanin levels, and is also involved in metabolising toxic by-products of the melanin production pathway (Jackson et al., 1992, Costin et al., 2005). The DCT gene can also increase overall tyrosinase activity through its role as a stabiliser of tyrosinase (Manga et al.,

2000).

Mutations in the DCT gene have been found to be associated with specific colour changes in three slaty phenotypes of mice linked to premature melanocyte death (Jackson et al., 1992, Budd and Jackson, 1995, Guyonneau et al., 2004), iris pigmentation in humans (Frudakis et al., 2003), and albinism in rainbow trout (Boonanuntanasarn et al., 2004). DCT has also been loosely associated with the recessive white chicken phenotype, but further evidence is still required to confirm this relationship (April et al., 1998). Black-boned sheep have been screened at the DCT gene, but no mutations were found that differentiated these from wild type sheep (Deng et al., 2009).

111

Candidate colour genes

4.2.4 Agouti-related protein

The agouti-related protein (AGRP) gene encodes a three-exon protein that varies in length between 154 and 165 amino acids in chickens (Takeuchi et al.,

2000). The protein is a known antagonist of the MC3R and MC4R receptors, and is heavily involved in feeding behaviour and weight regulation across a range of animals (Ollmann et al., 1997, Zimanyi and Pelleymounter, 2003).

The gene is also a part of the melanocortin pathway, and whilst mice-based studies have found the gene is predominantly expressed in the brain of these mammals, AGRP mRNA is also expressed in a wide range of different tissues in chickens, including skin (Takeuchi et al., 2000).

A functional avian homologue of the mammalian agouti gene was not thought to exist until late 2009, in the main because of repeated early failures to clone the

ASIP gene in chickens (Nadeau et al., 2008). The (then assumed) absence of an avian homologue of an agouti gene, together with the expression of AGRP in chicken skin led to a widely-held assumption that the AGRP gene must play an important role in avian pigmentation (Takeuchi et al., 2000, Jawor and

Breitwisch, 2003, Boswell and Takeuchi, 2005). The discovery of an avian

ASIP homologue in the Japanese quail does not, however, render this theory irrelevant, as the fact that AGRP mRNA is expressed in chicken skin suggests a significantly larger role for the AGRP gene in avian pigmentation than has been documented in the mammalian pigmentation system.

Fragments of the AGRP gene have since been screened across a few bird species, including the Scandinavian and Scottish Willow grouse, Lagopus

112 Candidate colour genes lagopus (Skoglund and Hoglund, 2010), pied flycatcher, Ficedula hypoleuca and collared flycatcher, F. Albicollis (Buggiotti, 2007). Neither of these studies found an AGRP association with pigmentation, but this does not preclude a role for this gene in colour determination across other bird species. Indeed, Li et al.

(2011) suggest from gene expression data that AGRP may be involved in excessive melanin synthesis in silky fowl chickens. Boswell and Takeuchi

(2005) suggested a role for AGRP in birds similar to that of Agouti in mammals, where Agouti (ASIP) acts as an antagonist on MC1R signalling and thus plays a part in the regulation of skin/coat pigmentation.

The aim of this paper is to sequence and screen relevant coding regions and

SNPs in these four candidate colour genes in magpies. These will be tested for associations with magpie back colour variation as well as checked for possible interactions between the variable sites (mutations) found that may contribute to back colour phenotype in the Australian magpie. If any of these mutations do segregate different magpie back colours, it is hypothesised that such mutation(s) may be similar, or the same as mutations linked to colour variation in other animal taxa, in terms of their location in the gene and/or their functional effect.

4.3 MATERIALS AND METHODS

4.3.1 Sampling

Magpie DNA samples from around mainland Australia and Tasmania have been collected by a number of researchers in the Hughes laboratory at Griffith

University (Queensland, Australia) over the last two decades. These samples

113

Candidate colour genes span the majority of the species’ distribution and take the form of whole blood, as well as blood in a range of buffer solutions, stored at -80 0C. Blood samples of 30 individuals sourced from 15 populations, spread across the continent were screened for this study, and include representatives from all sub-species, back colours and both sexes (see Figure 4.1 for sampling locations, and Appendix IV for further sampling details).

Black backed magpie

eylandtensis

longirostris terraereginae

dorsalis

tyrannica Varied magpie White backed magpie hypoleuca

Black backed

Hybrid zone

White backed Sampled population

Figure 4.1: Magpie plumage variants and their distribution across the Australian continent. Black backed forms inhabit black shaded areas, white backed forms pale grey areas, and intermediate forms are found in the medium grey shaded areas. The varied plumage form inhabits the shaded south-west region of the continent. White areas of the continent are not necessarily limits of the distribution and may simply represent regions of very low magpie density. Subspecies names are overlaid on their distributions.

4.3.2 Genotyping and analysis

From both whole blood and blood in buffer samples, whole genomic DNA was extracted using the simplified rapid method developed by Lahari and Schnabel

(1993) with several modifications. These included a ten-fold reduction in the volume of blood (50 μl from 500 μl), which was found to yield an ample quantity of high-quality DNA for downstream analyses, reducing sample wastage. 114 Candidate colour genes

Additionally, nonidet-40 was used as the detergent, and the MgCl2 concentration was 4 mM.

Primers used to amplify fragments of AGRP, DCT, TYR and TYRP1 and their sources are presented in Table 4.1. These gene fragments were selected based on several factors: they were in sections of these pigmentation genes found to be associated with colour in other animal species or, in the case of the

AGRP fragment, contained a specialised motif thought to be particularly important in pigmentation processes. Additionally, these gene fragments could be amplified utilising a combination of existing primer sets, gene-walking techniques and the development of specialised primer sets, a notable technical advantage in the absence of a detailed annotated genome, as is not yet available for the magpie.

Internal primers and separate amplification of different gene sections were used in some genes to capture the maximum read lengths possible. Primers designed specifically for this study were designed in several ways. TYR primers TYRZeb1F and TYRZeb1R were based on a published zebra finch TYR mRNA sequence (XM_002186685). DCT primer DCTaltR was designed from magpie sequences which were originally amplified using published primers

(Nadeau et al., 2007). AGRP primer AGRP_FY4R was designed from alignments of magpie AGRP sequences (amplified with primers from Buggiotti,

2007) with published sequences from chicken (AB029443, NM001031457), quail (AB489990) and duck (AY635903). The programs Primer3 (Rozen and

Skaletsky, 1999) and AutoDimer (Vallone and Butler, 2004) were used as part of the primer design and refinement processes.

115

Candidate colour genes

The PCR reactions of 10 μl contained 1-2 μl of extracted DNA template, 1 μl

10x reaction buffer (Fisher), 1 mM MgCl2 (Fisher Biotech), 0.8 mM each of forward and reverse primers, 0.2 mM of dNTPs (Bioline), 0.1 U Thermus aquaticus DNA Taq polymerase (Fisher), and the remaining volume was made up to 10 μl with ddH2O.

Table 2.1: Primer sequences and annealing temperatures used to amplify fragments from four candidate colour genes

Gene Primer Direction Primer name Primer sequence (5'-3') Annealing Source location Temp. (°C)

Agouti-related protein exon 1 F AGRP_GG1F AGGACCATGCTGAACGCG 58 Buggiotti et al. unpublished (AGRP) exon 3 R AGRP_FY4R TTCCTGCAGTAGCAAAAGGC This study intron 3 Internal AGRP_GG2F GCAGAAAGGCCTTGAGACAC Buggiotti et al. unpublished Dopa-chrometautomerase exon 2 F DCTF2 TGCTGGCTATAACTGTGGTGA 60 Nadeau et al. 2007 (DCT) exon 2 R DCTaltR CTGACWGAGTAGTAATGAAGCCARA This study Tyrosinase exon 1 F TYRZeb1F CTATTTGCCCTGGGTTTTCT 63 This study (TYR) exon 1 R TYRZeb1R CCTGCCATGAGGAGAAAATTGA This study Tyrosinase related exon 1 F TP1e1F3 CTCAGTTCCCTCGCCAGT 60 Nadeau et al. 2007 protein 1 exon 1 R TP1e1R1 GATTTGCTGGCTACAGGTAGGTC Nadeau et al. 2008 (TYRP1) exon 3 F TYRP1_GG1F CCTACCCTACTGGAACTTTG 60 Buggiotti et al. unpublished exon 3 R TYRP1_Fic1R TACACCAACTTCCAAACACTG Buggiotti et al. unpublished exon 5 F TYRP1_Fic3F GGTACAGTGATCCTTCAGGA 60 Buggiotti et al. unpublished exon 5 R TYRP1_Fic2R CGTCTCAGCCACTCGTCAA Buggiotti et al. unpublished

The cycling protocol was as follows: 94 °C of initial denaturation for 5 minutes, followed by 35 cycles of 30 seconds at 94 °C, 60 seconds at differing annealing temperatures specified for each gene in Table 4.1, and 70 seconds at 72 °C, after which a final 6 minute extension step was carried out and the samples were held at 4 °C until further processing. This PCR product was then purified in a reaction with exo-sap (Fermentas), and visualised on a 0.6% agarose gel stained with GelRed (Biotium).

Sequencing reactions were carried out in 10 μl volumes consisting of 1.0 μl purified PCR product, 0.32 mM of the relevant primer, 2.0 μl Big Dye Terminator

116

Candidate colour genes

Mix 3.1 (Applied Biosystems), 2.0 μl of 5X sequencing buffer (Applied

Biosystems), with the remainder of the 10 μl volume made up by ddH2O.

Sequencing thermal cycling conditions were as follows: 1 minute hold at 96 °C,

30 cycles of 96 °C for 10 seconds, 50 °C for 5 seconds, and 60 °C for 4 minutes, and subsequent storage at 4 °C.

Products from sequencing reactions were subjected to two 70% ethanol washes and the DNA pellet dried in a vacuum bell. An automated sequencing machine

(Applied Biosystems 3130x1) was then used to sequence both strands with

BigDye 3.1 terminator chemistry.

4.3.3 Bioinformatics analyses

Raw magpie sequences were aligned and preliminary edits made manually in

Sequencher 4.9 (Gene Codes Corporation 2000). Heterozygous bases were accepted if double peaks of similar heights were present on both strands.

Within each gene, haplotype phases were estimated using Phase version 2.1.1

(Stephens et al., 2001, Stephens and Scheet, 2005). Haplotypes were calculated for later use in downstream association analyses that utilized the R package MultiPhen (Coin et al., 2013). In cases where a large number of individuals were unable to be confidently assigned haplotype phases, sequences from the same gene fragment in closely related species were taken from Genbank and added into the analysis (as individuals of known phase) to improve estimation. At a Phase threshold of 60% (considered relatively robust, see Harrigan et al., 2008, Garrick et al., 2010) this out-group species inclusion increased the number of individuals useable in downstream analyses between 0 and 43%, depending on the gene.

117

Candidate colour genes

Magpie sequences were aligned with sequences from multiple other bird species to infer reading frame, calculate gene coverage and similarity between species, identify intron/exon boundaries, identify SNPs between magpie individuals, and compare SNPs thought to be involved in pigmentation changes in these other species. This was done using a combination of Sequencher version 4.9 (Gene Codes Corporation 2000), Bioedit version 7.0.9.0 (Hall, 1999) and Mafft (Katoh and Standley, 2013). Alignments were manually scanned for

SNPs that delineated back colour, subspecies or eastern vs. western populations, both in coding and non-coding regions. Basic population differentiation, neutrality, and linkage disequilibrium tests were carried out in

DnaSP version 5.1 (Rozas et al., 2003).

The relationship between magpie SNPs and different phenotypes and groupings were then analysed more thoroughly with a series of standard univariate regressions, multivariate regressions, and inverse ordinal regressions implemented in the R (Team, 2015) package MultiPhen (Coin et al., 2013).

Multiphen enables overlapping and correlated phenotypes to be modelled jointly in regressions using independent SNPs as dependant ordinal variables, to increase the power to detect associations between phenotypes and genotypes.

The analyses are relatively robust to non-normality of phenotypes (Wang, 2014).

In order to ensure multiple comparisons did not inflate type 1 error, a simple

Bonferroni correction for the number of SNPs analysed (67) was made, and

SNP-phenotype associations were only considered significant at p ≤ 0.00074

(Miller, 1981). Genetic groupings tested as psuedo-“phenotypes” in

118

Candidate colour genes

MultiPhen analyses were based on divergent mtDNA groups of populations and nuDNA groups delineated in Toon (2007).

4.4 RESULTS

A number of variable sites were found in each of the four genes (see Table 4.2) across the 30 magpie individuals, and these were used to calculate nucleotide diversity, neutrality measures and population parameters. Populations defined by plumage form showed little to moderate genetic differentiation (all FST values

≤ 0.15) between black back, white back and varied back individuals screened in this study (Table 4.2), with no statistically significant genetic structure between plumage forms at any of the four genes. Further details of sampling and these geographic and plumage groupings are contained in Appendix IV. Nucleotide diversity within all four gene fragments was in line with general estimates of this measure in bird species. Tajima’s D, Fu and Li’s D* and Fay and Wu’s H were non-significant across all four genes, indicating it is unlikely these genes are evolving non-neutrally (Table 4.2). Fu’s FS values were highly significant and negative in the TYRP1, TYR and AGRP genes, indicating an excessive number of alleles; a characteristic generally associated recent population expansions or genetic hitchhiking (Fu, 1997, Ramos-Onsins and Rozas, 2002) (Table 4.2).

Kelly’s (1997) ZnS tests indicated low levels of linkage disequilibrium across all four genes (Table 4.2).

119

Candidate colour genes

Table 4.2: Summary statistics of genetic variation and statistical tests of neutrality for phased sequence data used in this study. Statistical significance is indicated by * (p<0.05, or p<0.02 for Fu's FS). n is the number of phased haplotypes, S is the number of segregating sites, π is the pairwise nucleotide diversity, D is Tajima`s D, D* is Fu and Li's D*, FS is Fu's FS, H is Fay and Wu's H, ZnS is Kelly's ZnS test of linkage disequilibrium, and FST is Wright's fixation index of genetic differentiation between plumage forms (i.e. black backed, white backed, or varied form; hybrid individuals were excluded). No FST values were significant at p<0.05.

n S ∏ D D* Fu's Fs H ZnS FST (between plumage forms) TYRP1 60 12 0.03 -0.64 0.44 -21.34* 0.22 0.034 0.098 TYR 60 14 0.002 -1.28 -0.23 -12.04* -2.82 0.032 0.005 DCT 60 6 0.004 0.07 -1.50 0.01 -1.68 0.090 -0.030 AGRP 60 30 0.004 -0.87 0.53 -24.38* -3.19 0.039 0.030

Table 4.3 : P-values of SNPs within candidate colour genes identified by inverse ordinal regressions (calculated in the R package MultiPhen) as statistically significantly (p< 7.4E-04) associated with colour, subspecies or phylogeographic groupings. SNPs denoted by * are statistically significant at this level.

SNPs TYRP1 intb309 TYRP1 intb402 TYRP1 intb487 AGRP ex449

Back colour 4.14E-04* 6.45E-02 2.12E-06* 5.32E-01 Subspecies 4.14E-04* 6.25E-02 2.71E-05* 5.83E-01 mt E-W clades 2.50E-03 1.81E-04* 3.99E-05* 4.54E-02 nuE-W clades 4.14E-04* 6.48E-04* 3.51E-04* 1.16E-03* Joint Model 1.42E-02 3.13E-03 3.17E-05* 5.62E-04*

Table 4.4: P-values of SNPs within candidate colour genes identified by univariate logistic regressions (calculated in the R package MultiPhen) as statistically significantly (p< 7.4E-04) associated with colour, subspecies or phylogeographic groupings. SNPs denoted by * are statistically significant at this level.

SNPs TYRP1 intb309 TYRP1 intb402 TYRP1 intb487

Back colour 1.37E-02 2.05E-02 1.21E-04* Subspecies 5.80E-03 1.75E-02 5.88E-05* mt E-W clades 2.50E-03 1.60E-04* 3.99E-04* nu E-W clades 4.61E-04* 8.47E-04 1.73E-04*

4.4.1 Tyrosinase-related protein-1

TYRP1 is spread over 8 exons and 531 amino acids in the chicken. We sequenced 1522 base pairs of the magpie TYRP1 gene, which, numbered according to a reference chicken sequence (NM_205045), was comprised of: the 23rd- 125th amino acid (80% exon 1), the 263rd -303rd amino acid (60% exon

120

Candidate colour genes

3), the entire intron c (875 bp), the 304th- 318th amino acid (28% exon 4), and the 357th- 413th amino acid (93% exon 5). This sequenced region makes up just over 40% of the TYRP1 coding regions, plus intron c. Concatenated magpie exonic TYRP1 regions sequenced showed high congruence with analogous regions in other bird species: 88% with chicken, 88% with Japanese quail and 94% with zebra finch (Genbank accessions in Table 4.5).

121

Candidate colour genes

Table 4.5: Amino acid polymorphisms in the tyrosinase related-1 (TYRP1) protein associated with pigmentation changes in selected animal species. Amino acid positions numbered according to Gallus gallus sequence AF003631. Unshaded boxes indicate non-synonymous mutations in magpie coding regions.

Amino acid Common name Species Phenotype/Strain Allele Genbank accession 30 37 41 99 109 282 287 289 362 367

Australian magpie Cracticus tibicen pending T S C R C F A/V C S/I - Cracticus tibicen pending A S C R C F V C S - Cracticus tibicen pending A/T S C R C F V C S - Pied flycatcher Ficedula hypoleuca Long transcript Buggiotti, 2007 - - - - - F V C S - Zebra finch Taeniopygia guttata Black17 XM002193778 A S C R C F V C S - Chicken Gallus gallus Black Australorp X NHR BR+ AF003631 A S C R C F V C S - Gallus gallus Australorp NM205045 A S C R C F V C S - Japanese quail Coturnix japonica Wildtype BR+ AB005228 A S C R C F, F/S V C S - Coturnix japonica Roux brr EU046599 A S C R C S (F282S) V C S - Dog Canis familaris Black Large Munsterlander B+ AY052751 A N H C F V C S - C Canis familaris brow n b1 Schmutz et al. 2002 ------S (S41C) Cat Felis catus Wildtype B+ AY956310 A N F V C S - C R C Felis catus Chocolate b AY965744 A N F V C S - Felis catus Cinnamon b1 AY965746 V N C R C - - - - - Mice Mus musculus w ildtype C57BL/6 B+ NM031202 A R C X (R100X)Q - F V C S -

Mus musculus brow n b Zdarsky et al. 1990 - - - - C - - - - - Mus musculus Light Blt Johnson & Jackson 1992 - C (R38C) - - Y (C86Y) - - - - - Soay sheep Ovis aries Dark B+ EF102110 A N C R C- F V C S - Ovis aries Light b EF102109 A N C R C F V F (C290F) S -

Human Homo sapien B+ BC052608 A S C H C F V C S - Homo sapien Ruf ous OCA3 ------deletion

122

Candidate colour genes

Of the 40 SNPs found across the TYRP1 regions sequenced, 16 were in coding regions, and three of these were non-synonymous mutations that did not show any clear association with a particular back colour, subspecies or east-west grouping.

The first was a first position mutation at the 30th amino acid in exon 1, that separated the majority of magpie individuals and nearly all other animal species compared (alanine homozygotes) (see Table 4.5) from one white-backed C. t.hypoleuca individual (threonine homozygote) and one varied backed C. t. dorsalis individual (alanine/threonine heterozygote). The notable exception to this conservation of threonine at this SNP across species is the cinnamon phenotype of cat, which possesses a homozygous valine at this amino acid.

The second SNP at amino acid 287 (exon 3) differentiated one black-back C. t. terraereginae (valine/alanine heterozygote) from all other magpies and all other species compared (valine/valine homozygote). The third at amino acid 326

(exon 5) made a seemingly random 9/30 magpies (all back colours, subspecies, geographic groups represented) serine/isoleucine heterozygotes instead of the putatively ancestral serine/serine homozygote state seen in the rest of magpies, and all other compared species.

Both conventional and inverse regression analyses identified three SNPs in

TYRP1 as being significantly associated with different phenotypes (see Tables

4.3 and 4.4), and all three of these were located in the middle of intron c of the

TYRP1 gene. Mutations in varied-back individuals (C.t.dorsalis) from the southwest region of Australia seem to be the key (but not exclusive) driver of this effect (Figure 4.2). One of these was a one base-pair (bp) deletion, one an

123

Candidate colour genes

A-G substitution, and the third a 1 bp insertion. All magpies sequenced at

TYRP1 consistently possessed wildtype amino acids at sites thought to differentiate colour variants in other animal species (Table 4.5).

Figure 4.2: Three mutations in intron c of the C. tibicen TYRP1 gene found to be statistically significant in regressions of back colour, subspecies or phylogeographic grouping.

4.4.2 Tyrosinase

The TYR gene comprises 529 amino acids across five exons in the chicken

(Genbank accession: NM_204160), and over half of the coding region is contained within the first exon. We sequenced 713 base pairs of the TYR gene in magpies, comprising the majority (86%) of this first exon. This region corresponds to amino acids 23-259 of the chicken TYR gene (amino acids numbered after Genbank accession: NM_204160) and represents 44% of the gene coding regions. The sequenced region displayed 89%, 87% and 94% similarity to the corresponding regions in chicken, willow grouse and zebra finch respectively.

13 SNPS were identified in the fragment of the magpie coding region sequenced and, of these, three resulted in a change to an amino acid. The first, 124

Candidate colour genes at the 27th amino acid, split magpies into methionine/valine heterozygotes and methionine and valine homozygotes, but these groupings were seemingly not defined by back colour, subspecies or east/west grouping, while most other bird species aligned (chicken, quail, black cussarow, red legged partridge and great argus) possessed threonine at this site, the exception being the zebra finch

(valine). The second SNP, at the 152nd amino acid, was a change from alanine

(remainder of magpies and all other birds species mentioned above) to alanine/valine heterozygous state in one of each of the three magpie back- colour forms. The third non-synonymous change was in a single individual at the 193rd amino acid that made it an asparagine/serine heterozygote, whilst all other magpies and aligned bird species were homozygous for asparagine

(Table 4.6).

125

Candidate colour genes

Table 4.6: Amino acid polymorphisms in the Tyrosinase (TYR) gene associated with pigmentation changes in selected animal species. Amino acid positions numbered according to Gallus gallus sequence D88349. Unshaded boxes indicate non-synonymous mutations in magpie coding regions.

Amino acid Common name Species Phenotype/Strain Allele Genbank 27 77 103 146 152 193 227 237 238 294

Australian magpie Cracticus tibicen M R C S A N G D W - Cracticus tibicen V R C S A N G D W - Cracticus tibicen M/V R C S A/V N/S G D W - Zebra finch Taeniopygia guttata Black17 XM002186685 V R C S A N G D W E Chicken Gallus Gallus White leghorn albino ca AB023291 T R C S A N G ΔD237 ΔW238 E Chicken Gallus Gallus White leghorn D88349 T R C S A N G D W E Chicken Gallus Gallus OAC1A NM_204160 T R C S A N G D W E Japanese quail Coturnix japonica Wildtype AB024278-9 T R C S A N G D W E Great argus Argusianus argus Assumed w ildtype EF571148 T R C S A N G D W E Willow grouse Lagopus lagopus Assumed w ildtype EF571131 T R C S A N G D W E Red-legged partridge Alectoris rufa Assumed w ildtype EF571144 T R C S A N G D W E Black cussarow Crax alector Assumed w ildtype EF571142 T R C S A N G D W E Human OCA Homo sapien NM00372 S R C S P E G D W E Mouse Mus musclus Jackson Tyr c-2J ------L (R77L) Mus musclus Classical albino Tyr c X51743 ------S (C85S) - Mus musclus Dark-eyed albino Tyr c-44H ------I (S128I) Mus musclus C57BL/6J black Tyrs-J D00440 S R C S P E G D W E Cat Felis catus Burmese cb AY743344 S R C S P E W (G227W) D W E Felis catus Wildtype C AY959314 S R C S P E G D W E Rabbit Oryctolagus cuniculus Wildtype-German giant AF210660 S R C S P E G D W E Oryctolagus cuniculus Chinchilla & Californian ------G (E294G)

126

Candidate colour genes

No TYR SNPs were identified as significantly associated with any individual phenotype considered, or in a joint phenotype model, including back colour, subspecies, and east-west grouping, across a number of different univariate, multivariate and inverse ordinal regressions (Tables 4.3 and 4. 4). Magpie TYR sequences consistently possessed the wildtype amino acid at sites thought to be functionally important in colour changes in other species (Table 4.6).

4.4.3 Dopachrome Tautomerase

In chickens the DCT gene spans between 15 and 30 kb of the genome, and is made up of 516 amino acids spread across an unknown number of exons (April et al., 1998). For this study, the majority of the second exon was sequenced, comprising the 104th- 196th amino acid of the gene (numbered according to chicken reference NM204935). This sequenced region also showed high levels of similarity with equivalent regions in other bird species of 87%, 89% and 95% with chicken, duck and zebra finch respectively (Genbank accessions in Table

4.7).

Four SNPs were found in the sequenced region of DCT, of these three were synonymous substitutions, while one SNP led to the change of an arginine for a histidine at amino acid 149 (R149H). This non-synonymous substitution occurred in only one white-backed C. t. telonocua individual, and in only one allele in this individual (Table 4.7). Magpies of the same back colour, subspecies and other individuals sampled in close geographical proximity all retained the (presumed) ancestral arginine at this position that is shared in all five of the other bird species compared. (Table 4.7). 127

Candidate colour genes

Of the SNPs found to be associated with pigmentation in other animal species, one lies within the sequenced region: the R194Q mutation that causes a slaty phenotype in mice (Budd and Jackson, 1995), and magpies retain the wildtype amino acid at this site.

Table 4.7: Amino acid polymorphisms and SNPs in the Dopachrome Tautomerase (DCT/ TYRP2) gene associated with pigmentation changes in selected animals. Amino acids positions numbered according to Gallus gallus sequence NM204935.

Amino acid / intron SNP Common name Species Phenotype Genbank intron a 149 194 intron b SNP rs1325611 SNP rs1407995 Australian magpie Cracticus tibicen - R R - Cracticus tibicen - R/H R - Cracticus tibicen - R R - Zebra finch Taeniopygia guttata XM002199349 - R R - Chicken Gallus gallus NM204935 - R R - Japanese quail Coturnix japonica AB081466 - R R - Indian peaf ow l Pavo cristatus EF571054 - R R - Edw ard's pheasant Lophura edwardsi EF571062 - R R - Mice Mus musculus w ildtype - L R - Mus musculus slaty - - Q - Mus musculus slaty2J - - - - Mus musculus slaty light - - - - Black-boned sheep Ovis aries - L R - Human Homo sapien increased blue iris c L R c Human Homo sapien decreased brow n iris t L R t Rainbow trout Oncorhynchus mykiss w ildtype - L R - Oncorhynchus mykiss albino - - - -

4.4.4 The agouti-related protein

The AGRP gene is has three exons, and ranges between 154 and 165 amino acids in length in the chicken (Takeuchi et al., 2000). We sequenced 1247 base pairs of the AGRP gene in the magpie, which correspond to amino acids

32 -154 of the chicken AGRP gene (White leghorn chicken Genbank accession:

AB029443, protein ID BAA82256.1). (Magpie AGRP sequences are hereafter numbered after this record). This region encompasses partial exon 1, complete

128

Candidate colour genes exon 2 and partial exon 3, as well as the entirety of both intervening introns.

These introns were markedly different in size from those of the reference chicken sequences: intron a is 114 bp in chicken but 268 bp in magpies, whilst intron b is 980 bp in chickens but only 613 bp in magpies. The sequenced region thus covered around 79% of the putative coding region of AGRP, assuming the ends of exons 1 and 3 are relatively similar in length to those of chickens. In coding regions, amino acids showed moderate similarity to those sequenced in other bird species: Willow grouse (Skoglund and Hoglund, 2010), quail (Genbank accession: AB489990), chicken (Genbank accession:

AB029443) and duck (Genbank accession: AY635903) were found to have

79%, 81%, 82% and 83% identity to magpie sequences respectively.

Of the seven SNPs identified in exons of the magpie AGRP gene, five were synonymous changes and two led to changes in the amino acid. These two

SNPS were located in the first and second exons. The first led to a serine- proline change at the 54th amino acid in one black-backed magpie, whilst other bird species mentioned above had a leucine at this site, while the second, at the

81st amino acid resulted in a alanine-asparagine change in two varied back magpies and a alanine-valine change in a BB individual. Other bird species had an alanine at this site, like the majority of magpies sequenced (Table 4.8).

None of the seven SNPS were segregated by back-colour or sub-species, and the locations of the non-synonymous changes in magpies were also non- overlapping with the AGRP fragment (83- 132 aa) that has been suggested to suppress the dispersion of pigment in melanophore cells through α-MSH in humans (Quillan et al., 1998), (Table 4.8). The C-terminal domain of AGRP that is highly conserved across many animal species was found to be highly

129

Candidate colour genes conserved in magpies, and was identical to chicken, Japanese quail and duck sequences compared (Table 4.8). The peptide within AGRP suggested by

Thirumoorthy et al. (2001) as a possible antagonist of MC1R in human skin tissues was also highly conserved in magpies, and this peptide fragment in magpies was also identical to chicken, Japanese quail, duck and human sequences (Table 4.8).

An inverse ordinal regression jointly using back colour, subspecies, and east- west groupings as predictor variables identified one SNP in the third position of the 91st amino acid (Table 4.3), close to the end of the second exon, as significant at the p ≤0.000746 level. This synonymous mutation changes the third nucleotide in two white-backed C.t. hypoleuca and two black-backed C.t. tibicen individuals to a heterozygous (C/T) at this position instead of a homozygous (T/T) which does not alter the amino acid leucine present at this position in all other magpies and all other species compared.

Table 4.8: Amino acid polymorphisms in the agouti-related protein (AGRP) gene associated with pigmentation changes in selected animal species. Amino acid positions are numbered after Gallus gallus AB029443.

Amino acid Common name Species Phenotype Genbank 54 81 109-145 132-139 C- terminal domain Peptide antagonist of MC1R? (humans) Australian magpie Cracticus tibicen S A CRV →RKI CRFFNAF Cracticus tibicen P D CRV →RKI CRFFNAF Cracticus tibicen V CRV →RKI CRFFNAF Chicken Gallus Gallus White leghorn AB029443 L A CRV →RKI CRFFNAF Japanese quail Coturnix japonica AB489990 L A CRV →RKI CRFFNAF Willow grouse Lagopus lagopus L A - CRFFNAF Duck Anas platyrhynchos Mallard AY635903 L A CRV →RKI CRFFNAF Human Homo sapien - - - CRFFNAF

130

Candidate colour genes

4.5 DISCUSSION

Tests of neutrality failed to identify any signs of positive selection in any of the four candidate colour genes, although the very large and highly significant negative Fs values obtained for TYRP1, TYR and AGRP suggest either genetic hitchhiking, or recent population expansions in magpies. As neither Fu and Li’s

D* statistic nor Fay and Wu’s H statistic showed any evidence supporting selective sweeps in these genes, and Kelly’s ZnS did not find evidence of linkage disequilibrium within any of these three genes, it seems less likely that genetic hitchhiking has left this signal in TYRP1, TYR and AGRP. Instead it is suggested that these highly negative Fs values reflect recent population expansions in this species. Tajima’s D values of these three genes were not statistically significant, but were, however, negative, which also (albeit weakly) adds support for a hypothesis of past population expansions. This same signal was found in analyses of the MC1R gene in magpies (Dobson et al., 2012), as well as (less significantly) in the mitochondrial control region (Toon et al., 2007).

This signature of population expansion seems likely to be a result of posited expansions of magpie populations out of refugial environments in the late

Pleistocene, discussed in more depth in Dobson et al. (2012), (Chapter 3) and

Toon (2007).

Across the four pigmentation genes investigated, although significant variation was uncovered (67 SNPs, 51 in coding regions out of 3758 nucleotides), no variation was discovered that wholly or even partially explained colour variation in the Australian magpie. Regression analyses (that were carried out in order to sift through fine-scale associations that may play a small role in pigmentation), failed to identify SNPs that may play a small role in magpie pigmentation, but

131

Candidate colour genes rather likely reflected some characteristics of different populations. For example, three TYRP1 SNPs in intron c were identified as significant in regression analyses (two for colour phenotype) (Tables 4.3 and 4.4). However, upon closer dissection, this effect could be ascribed to the higher genetic variability within and between south-western white backed C.t. dorsalis populations (as opposed to all other sampled subspecies and populations) that has been observed previously in a putatively neutral mitochondrial marker

(Toon, 2007). It seems likely this same signal is being uncovered here, in a non- coding region unconstrained by functional importance or possible deleterious outcomes of variation. The three mutations were clustered in the middle of intron c (875 bp long in magpies), and thus it seems unlikely they would affect splicing during transcription, as they are located far from splicing sites.

Interestingly, splicing variation in the TYRP1 gene that was found to be associated with plumage colour variation in pied flycatchers (Buggiotti, 2007) was not a result of mutations in splicing sites. This splicing error resulted in the transcription of a premature stop codon in an intron (e), and subsequent loss of exons 6 and 7 in final protein products. The mechanism underlying this splicing error remains unclear, although Buggiotti (2007) suggested mutations in regulatory regions or the action of regulatory elements such as enhancers.

High-density genome scanning of the Z chromosome in two recently-diverged sister flycatcher species (Backström et al., 2010) also suggests the TYRP1 gene may play a role in diversifying selection leading to sexual isolation and ultimately speciation, putatively related to plumage colour. Thus it is suggested that these mutations in intron c of the TYRP1 gene in magpies, which may yet simply reflect phylogeographic history, indicate this TYRP1 gene and its protein

132

Candidate colour genes products are deserving of further investigation with regard to magpie back colour variation.

One AGRP SNP (AGRPex449) was found to be significant in an inverse ordinal regression analysis for the joint model, in which ‘phenotypes’ of back colour, subspecies, and phylogeographic groupings, as well as their interactions, were considered together to explain magpie SNP genotypes. This SNP was found to be non-significant for each ‘phenotype’ individually in univariate analyses. The magpie individuals with this SNP were two white-backed C. t. hypoleuca

(Tasmania) and two black-backed C.t. tibicen (Grafton) birds, and both of these populations belong to the eastern Australia grouping, for both mitochondrial and nuclear DNA (Toon, 2007), and these groupings were two of the ‘phenotypes’ in the joint model. Thus it may be this relationship driving the significant result, or alternatively, due to the fact that sequences from both Tasmania and Grafton populations contained consistently more variation than those from other populations across all four candidate colour genes (A. Dobson, unpublished data). Tasmanian magpies have been suggested to have diverged more recently than other subspecies, likely from a population of eastern white backed magpies that became isolated when sea levels rose towards the end of the

Pleistocene (Hughes et al., 2001, Toon, 2007).

Whilst this study screened a large proportion of the four pigmentation genes, it remains possible that SNPs in parts of the genes not sequenced could be associated with magpie colour variation. However, with the possible exception of the TYRP1 gene, they now seem unlikely candidates, as a large proportion of the SNPs in these genes that have been identified as functionally important in

133

Candidate colour genes pigmentation across other animal species were targeted, and found to be non- variable within magpies. Indeed, at all the assayed SNPs important in colour variation in at least one other species (Tables 4.5 - 4.8), magpies retained the conserved (presumed ancestral) wild type allele, as do most other species compared, indicating that these sites are likely to be functionally important across a large range of species. The structure and content of the four genes examined in magpies appear to be very similar to analogous regions in other bird species, ranging between 79-95% similarity (between genes) between magpie sequences and those of duck, chicken, Japanese quail, willow grouse and zebra finches.

The range of magpie back-colour and subspecies, as well as geographic regions has been well-sampled, and therefore increased sampling is unlikely to uncover additional variation that explains pigmentation variation.

While the sequenced coding regions of the AGRP, TYRP1, TYR and DCT genes seem highly unlikely to be involved in magpie pigmentation, it is quite possible that other mechanisms that are involved with these genes may alter colour. Mutations in coding regions of candidate colour genes are far from the only mechanism thought to determine colour variation within animal species: regulatory elements and genes are also thought to play a role, and may be solely responsible for such variation in some organisms (Linnen et al., 2009,

Manceau et al., 2010). The number of genes controlling magpie variation could be relatively small, despite its continuous nature in hybrid zones, if regulatory genes are also involved; for instance a regulatory mutation potentially controlling the width of the black saddle on magpie backs could produce a

134

Candidate colour genes range of back colours with a continuous rather than discrete number of morphs or forms.

Cheviron et al. (2006) suggest the ontogenetic pattern of plumage development in the blue-crowned manakin (in which the plumage of males resembles females until they develop definitively male plumage colour in their second year of life) may point towards an underlying mechanism within regulatory genes, rather than structural coding regions. Magpies have a similar plumage developmental stage, in which sub-adults are grey in their napes and backs; in sub-adult white backed birds this grey is a substantially paler shade of grey than the napes and backs of birds that grow into black backed forms, when viewed side-by-side in a hybrid zone (Hughes et al., 2002, personal observation).

A growing body of literature linking pigmentation variation to regulatory mechanisms in animal species is emerging from the contentious debate (e.g. see Hoekstra and Coyne, 2007, Carroll, 2008) over the relative importance of mutations in structural coding regions vs. regulatory regions. A deletion in

Agouti, which antagonises the MC1R gene, was found to be perfectly associated with melanism in Peromyscus mice (Kingsley et al., 2009), while the dark brown (DB) mutation in chickens has been found to been associated with a deletion upstream of the SOX10 transcription factor, in a region thought to be an enhancer of the gene (Gunnarsson et al., 2011). The homozygous lethal yellow mutation in the Japanese quail has also been linked to regulatory mutations, in which deletions upstream of the agouti signalling protein (ASIP), in two promoter regions, affect transcription of the ASIP gene (Nadeau et al.,

2008). Additionally, an isoform of the transcription factor MITF has also been

135

Candidate colour genes suggested to play a regulatory role in white spotting in domestic dogs (Körberg et al., 2014). Clearly such regulatory regions, although generally harder to identify, locate and screen (Hoekstra and Coyne, 2007) are worthy of further research in magpie pigmentation, especially in light of their ontogenetic plumage development and the continuous nature of magpie back colour variation in hybrid zones.

In the tawny owl, Emaresi et al. (2013), found that expression levels of both the

MC1R and TYRP1 genes were positively correlated with increasing ‘darkness’ of melanic nestling plumage, and offspring of ‘darker’ parents expressed TYR,

TYRP1 and MC1R at higher levels than those with parents of lower plumage

‘darkness’. This study also suggests expression patterns of melanogenesis genes indicate plumage colour in this species is likely to be associated with a number of pleiotropic genes involved in other morphological, physiological, and even life-history or social/behavioural traits (Emaresi et al., 2013). Indeed, San-

Jose et al. (2017) even linked mutations in the coding region of MC1R

(associated with white and rufous forms of the barn owl) to differential expression levels of other melanogenic genes, notably TYR, TYRP1 and DCT; the white allele (associated with lighter coloration) showed significantly lower expression of these genes than the rufous allele. Such studies of the expression levels of different pigmentation genes offer a window into the more detailed picture of the regulation of pigmentation genes, and it is becoming increasingly clear that a range of highly complex interactions and regulatory mechanisms may play a significant role in determining pigmentary phenotypes in some bird species.

136

Candidate colour genes

Splicing errors have also been implicated in a number of examples of animal patterning, including sabino white spotting in horses (Brooks and Bailey, 2005) and male pied flycatchers (Buggiotti, 2007). Many examples of colour variation within species may also be due to pleiotrophic effects of certain genes, and thus may be linked to selection acting upon other non-pigmentary traits (Barton,

2000, Ducrest et al., 2008). Other pigmentation variation traits may be determined by multigenic effects, where variation across several or many genes may interact to alter pigmentation. A good example of such variation is found in the domesticated dog, where interactions between SNPs in MC1R, ASIP and

TYRP1 determine the colour of nose and paw pads in many breeds (Schmutz et al., 2002, Berryere et al., 2005, Dreger and Schmutz, 2010, Dreger and

Schmutz, 2011). The non-discrete nature of back colour variation in magpies may indicate a polygenic basis for this trait, as Wlasiuk and Nachman (2007) suggested may be the case in gophers with continuous coat colour variation.

Back colour variation in magpies is thought to be heritable and non-plastic

(Hughes, 1982, Hughes et al., 2001), and early work on the species suggested it seemed more likely only a small number of genes are involved in this pigmentary variation (Hughes and Mather, 1980 in Hughes, 1982), as black- back genes seem to be dominant over white-back genes (Hughes and Mather, unpublished data, in Hughes et al., 2001). Both the high heritability of back colour variation and hypothesised involvement of only a few genes suggest candidate gene screening as a reasonable approach for attempting to pinpoint the genetic mechanism behind colour variation in this species. However, between this study and Dobson et al. (2012) who examined MC1R for

137

Candidate colour genes associations with back-colour in magpies, five of the most promising pigmentation genes have now been partially screened.

Studies using the candidate gene approach to identify the genetic basis of traits such as intraspecific colour variation have yielded a large number of genotype- phenotype associations across a wide range of animal species (Piertney and

Webster, 2010). One of the advantages of such a candidate gene approach is that it is ‘hypothesis driven’, in that researchers can utilise a prori knowledge or hypotheses about the nature of the trait of interest to inform selection of relevant candidate genes (Tabor et al., 2002). However, some authors have argued this may reduce their efficiency when the biology of the trait of interest is complex and/or poorly understood. The nature of scientific publication means that many studies that have not found associations between the candidate gene(s) screened and their trait of interest may never be published, unless the gene is of particular importance, or the trait is of intrinsic interest (e.g. MC1R or breast cancer). In some cases the search for the genetic basis of a trait that seems initially tractable using candidate genes may require substantial investment and exhaustive research far beyond the screening of candidate genes to pin down; the plumage variation between the hooded and carrion crow, finally linked to a particular region of DNA using genomic methods, following exhaustive candidate gene screens is one such example (Haas et al.,

2008, Poelstra et al., 2013, Poelstra et al., 2014).

Many of the ‘low hanging fruit’; (to borrow a phrase from Risch (2000)), in terms of linking animal phenotypes to genotypes have now been ‘picked’, as traits with simple inheritance patterns, strong genotype-phenotype associations

138

Candidate colour genes and large effect have been tackled first by researchers, alongside model animal species. While many of these studies used linkage analyses, for other, more complex and non-Mendelian traits with low displacement in non-model species, these linkage analysis approaches seem less likely to prove as reliable and efficient as they have been for the ‘low-hanging fruit’ (Risch, 2000).

Risch and Merikangas (1996) and Risch (2000) suggest the candidate gene approach, discarded by many researchers in the excitement of early linkage analysis successes, may prove a vital part in the unravelling of the genetics underlying traits, especially if efforts are targeted in a measured way. Tabor

(2002) argue that by targeting known polymorphisms in relevant genes, weighted by the type of variation (e.g. coding sequence, splice site boundary), their potential to cause changes to phenotype, their frequency in the genome as well as the predicted functional effect for the gene product, offers an approach that may dramatically increase the efficiency of candidate gene studies. Such prioritisation of SNPs by type was also suggested by Risch (2000) to tackle the genetic basis of complex and non-Mendelian diseases in humans.

High-density genome scans utilising linkage disequilibrium metrics have also been suggested as a fruitful approach when searching for determinants of genetic trait(s) (Collins et al., 1997). Very recent sequencing and assembly of a draft genome of the magpie as part of the “Bird 10,000 Genomics Project”

(Zhang, 2015) claims coverage of approximately 44% of the magpie genome, with more than one billion nucleotides sequenced. Using this database, the utility of high-density genome scans for narrowing down genomic regions involved in a trait of interest is greatly increased. Downstream candidate gene

139

Candidate colour genes association studies will benefit from the narrowed and better targeted focus of relevant genes and genomic regions for screening revealed by such high- density genome scans, as well as the ability to easily design primers from such draft genomes, specific to the species of interest, to screen the relevant candidate genes. It is envisioned these advances will enable much more cost and time-effective screening of candidate genes for trait associations into the future for non-model species with interesting plumage colour variation such as the Australian magpie.

140

General Discussion

CHAPTER 5: GENERAL DISCUSSION

5.1 EXPLORING THE GENETIC BASIS OF INTRASPECIFIC

COLOUR VARIATION

Investigations of the genetic basis of morphological variation within a species are an important step towards understanding the evolutionary forces that shape life on earth. Plumage colour forms of birds are fascinatingly diverse; often striking similarities in plumage colour and pattern are observed in very distantly related taxa, and conversely sister species may have very different plumage colours and patterns. Colour polymorphisms within species have been suggested to represent an intermediate step in the process of speciation, and such intraspecific variation offers researchers the exciting opportunity to directly study evolution in action. Investigations of the genetic mechanisms behind such variation, when paired together with the wealth of knowledge from past studies examining historical and contemporary processes (both neutral and adaptive) that shape species and their morphological traits (such as colouration), give us the ability to ask deeper questions about the origin, function and maintenance of such variation.

5.2 ADDING TO THE PHYLOGEOGRAPHIC STORY OF THE

MAGPIE

Phylogeographic analyses of MC1R sequences from populations of magpies around Australia were used to further build upon the interesting history of changes in demography and geographic distribution that previous studies have suggested for this species. It seems unlikely that the MC1R gene is currently 141 General Discussion under selection in magpies, denying us the opportunity of examining genetic structure directly shaped by different selective pressures/processes in different plumage forms. Instead, the genetic structure in this gene was used to further investigate putative historical expansions of the species (Toon et al., 2007).

Several lines of evidence were found in this study that support the idea of a historical expansion(s) of magpies in Australia, including a star-like haplotype network, significantly negative Tajimas’ D value, a highly significant Fu’s Fs value, and the distinctive shape of the mismatch distribution. The weak genetic structure observed within MC1R is likely to reflect that it is a nuclear gene in a structural coding region, in which, even if not currently under selection, mutations have the potential to have deleterious effects and alter or destroy the functionality of the gene, its products and/or processes downstream from the gene.

The MC1R gene is generally highly conserved between species (Hubbard et al.,

2010), and therefore, it is not remarkable that the weak genetic structure found in the gene failed to echo the strong divergences between eastern and western magpie populations found in mtDNA control region data (Toon et al., 2007). The

MC1R gene is highly likely to be evolving at a slower rate than the mtDNA or non-coding regions of nuclear DNA that have been examined previously in the species (Baker et al., 2000, Hughes et al., 2001, Toon et al., 2003, Toon et al.,

2007). Thus it seems unlikely the expansion signature observed in MC1R is the result of relatively recent expansions of magpies (post European settlement), into previously unsuitable habitat altered by pastoralisation and urbanisation, as suggested by Campbell (1929) and Shodde and Mason (1999), and supported

142

General Discussion by evidence of contemporary gene flow between north-eastern and north- western populations (Toon et al., 2007). Instead, it is suggested that this signature of population expansion in MC1R may be linked to older (but still relatively recent in evolutionary timescales) changes in population size and distribution in magpies during the late Pleistocene, as suggested by Toon et al.

(2007), enabled as aridity decreased following the Last Glacial Maximum, allowing refugial populations to recolonise the interior of the continent.

5.3 HERITABILITY AND INHERITANCE PATTERNS OF MAGPIE

BACK COLOUR

A review by Visscher et al. (2008) compared heritability values across a range of animal taxa, and suggested that heritability values are generally higher for morphological traits (e.g. colour, bill length) than they are for fitness traits (e.g. brood size, offspring survival rates). Of the traits they reviewed, morphological trait heritability (h2) estimates ranged between 0.3 and 0.8 (µ=0.43), while fitness traits ranged between 0.05 and 0.3 (µ=0.18) (Visscher et al., 2008).

Within birds, heritability of colour trait variation seems to be strongly related to the type of pigmentation. Melanin-based traits generally have much higher heritability estimates than carotinoid or structurally-based pigmentation, and melanic trait heritability estimates ranging between 0.8 and 0.95 across a wide range of bird species tend to be the rule rather than the exception (see discussion in Chapter 2). Therefore, the high heritability estimates for magpie back colour variation found in this study (weighted midparent regression: h2=0.94; animal model: h2=0.92, C.I. 0.80-0.99) seem likely to be an accurate

143

General Discussion reflection of a high genetic contribution to back colour variation in the study population. This conclusion is further supported by the fact that both traditional regression-based estimates and linear mixed model-based estimates independently produced very congruent estimates of heritability.

Assortative mating, maternal and paternal effects, and shared environmental effects can all potentially bias heritability estimates. However, previous work has suggested magpies do not mate assortatively in this study population

(Hughes et al., 2011). This is supported by data in this study that shows a slightly skewed distribution towards mating pairs in which both sexes possess dark plumage. This is almost certainly an effect of the higher frequency of darker back colour in this study site rather than assortative mating patterns.

Maternal and paternal effects also seem very unlikely to have significantly influenced heritability estimates, as both maternal-offspring regressions and paternal-offspring back colour regressions estimated lower heritability values than midparent-offspring back colour regressions. The inclusion of shared environmental effects, in the form of patch quality (territory of birth) and cohort

(year of birth), in the animal model indicated these may account for a very small proportion of the variance in this trait, but measures of model fit indicated the inclusion of patch quality led to a lack of fit between models and data as well as over-parameterisation of the model, and indicated only the inclusion of cohort in the most statistically supported model. Generally, it is suggested that not accounting for shared environmental effects artificially inflates measures of heritability (Lynch and Walsh, 1998). However, the inclusion of cohort as an effect resulted in a slightly higher h2 estimate (h2=0.92) than a model with no random effects (h2=0.91), indicating that not including the cohort term very

144

General Discussion slightly increased the error term of the model rather than artificially inflating the heritability estimate.

The high heritability of magpie back colour variation does not necessarily imply that the trait is more likely to be determined by a small number of genes, although this assumption has been erroneously made by many researchers (i.e.

Karell et al., 2011); although high heritability has sometimes been found in traits determined by a small number of genes, this is not always, or even predominantly true (Visscher et al., 2008). The heritability metric itself does not reveal much about the genetic architecture of magpie back colour; however, if the patterns of inheritance observed in magpie family groups were more

Mendelian in nature (i.e. discrete rather than continuous variation of back colour in hybrid zones), then the high heritability of this trait may have been indicative of a single-gene genetic basis. This high heritability does indicate that the correlation between genotype and phenotype is very strong, and consequently future gene-mapping studies investigating this trait have a reasonable likelihood of pinning down genes of large effect that may play a role in magpie back colour.

It seems likely that environmental effects play a very limited role in magpie back colour determination, and although heritability estimates obtained in this study are reflective of the heritability of this trait in this study population over the last few decades, it seems reasonable to suggest that heritability of back colour variation is also likely to be high in other variable magpie populations across the range of the species (Visscher et al., 2008).

145

General Discussion Future studies comparing heritability of back colour variation in other magpie populations may benefit from the use of methods that do not require extensive genetic pedigrees, but instead utilise large number of genetic markers to estimate heritability (Ritland, 1996, Lynch and Walsh, 1998, Thomas and Hill,

2000). Patterns of back colour inheritance within family groups, together with the continuous nature of back colour phenotypes indicates that this trait is likely to be determined by a minimum of two genes in magpies, and it remains possible that the number of genes involved may be very large indeed.

The high heritability of magpie back colour variation found in this study is very comparable to those estimated for melanin-based plumage traits in other bird species and it is suggested that this is at least partially a reflection of the need for melanic pigments to be manufactured endogenously, rather than obtained from environmental sources. Many melanic plumage variations have been found to have a conserved genetic basis, e.g. the MC1R gene (Mundy, 2005) and many such traits have been convincingly linked to selective processes in other taxa. It is suggested that further research into the genetic architecture and genetic basis of this trait in magpies may benefit from gene-mapping methods, given the high heritability estimates found in this study.

5.4 SCREENING OF CANDIDATE COLOUR GENES

Five genes (MC1R, TYRP1, DCT, TYR and AGRP) that each play different roles in pigmentation pathways were screened for variation that could explain, either wholly or partially, back colour variation in magpies. No SNP or group of

SNPs was found to be meaningfully significant when tested for associations with any magpie back colour, although two SNPs in intron b of the TYRP1 gene 146

General Discussion did show some imperfect partitioning between varied back colour forms vs. white and black backed forms. It is suggested this was likely to be an effect of phylogeographic historical processes. However given the unknown mechanism of splicing variation in the TYRP1 gene linked to plumage colour variation in the pied flycatcher (Buggiotti, 2007) and the association of several intronic SNPs in the TYRP1 gene with iris colour variation in humans (Frudakis et al., 2003), it is evident this gene is deserving of focus in future studies of magpie pigmentation.

All SNPs previously found to be associated with colour variation in other bird species that were screened for this study, were invariable across nearly every magpie individual. Interestingly, in all five genes, at these candidate SNPs, magpie sequences always coded for the conserved (and presumably ancestral) wildtype amino acid. Most often, these wildtype amino acids seem to be highly conserved across a wide range of species, indicating they may play an important functional role across a broad range of taxa.

The sequenced portion of four of the five genes (MC1R, TYRP1, TYR, DCT) showed high levels of similarity (>90%) with a wide range of other bird species, while magpie AGRP gene sequences were less congruent with other bird species tested (79-83%), indicating it may be under less functional constraint.

None of the five genes were screened across the entirety of their coding regions, and it is possible that unassayed variation in these genes could be associated with magpie back colour variation. However a large number of

SNPs identified as important in colour variation in other bird species were specifically targeted by this study and found to be invariable in magpies.

Coverage of putative coding regions varied between genes (all of which varied

147

General Discussion in size and exon/ intron structure), from 90% of MC1R, 79% of AGRP, 44% of

TYR, 40% of TYRP1, and 18% of DCT.

A large number of the genetic variants found to be associated with melanic intraspecific colour variation in animals have been located within structural coding regions of pigmentation genes (Roulin and Ducrest, 2013), but this is not necessarily reflective of the only, or indeed even the dominant, mechanism underlying the genetic basis of such traits. Assaying structural coding regions of relevant candidate genes has been the first port of call in most studies of pigmentation genetics in non-model bird species, as a relatively affordable process able to be easily carried out in most laboratories with the capability to extract and amplify genomic DNA. Additionally, amplification of genomic coding regions are generally successful using easily obtained or pre-existing samples

(e.g. blood in buffer stored at -80 ºC) from the species of interest, and primers can be designed within more conserved exonic regions. Early studies of candidate colour genes had a number of high-profile and/or biologically compelling success stories linking variation in candidate gene coding regions to pigmentation variants, many within the same few genes e.g. MC1R. As such genes become the intense focus of research, disproportionate selection of such

‘popular’ candidate genes for screening can lead to sampling bias that over- represents the importance of such genes in pigmentation processes, relative to other candidate genes in the literature, resulting in a potentially “biased view of evolutionary forces” (Roulin and Ducrest, 2013). However, the MC1R gene does seem to represent a truly conserved genetic source of melanism-related mutations across many taxa, and thus is suggested to be deserving of its

‘superstar’ status as a key protein in melanogenesis (Mundy, 2005).

148 General Discussion Together, these factors have resulted in the identification of a plethora of SNPs in coding regions of candidate genes thought to be involved in colour variation, but as this field of research advances, it is becoming evident that other genetic mechanisms may also play important roles. Splicing errors, mostly caused by mutations in introns which commonly lead to the skipping of the following exon during transcription, have been associated with pigmentation variation, and specifically patterning in a number of species, including horses (Brooks and

Bailey, 2005), chickens (Gunnarsson et al., 2007) and pied flycatchers

(Buggiotti, 2007). Discovering such splicing errors is markedly more resource- intensive than screening simple candidate genes coding regions, and requires analyses at the cDNA level to study transcription products of the gene of interest. RNA work is very sensitive to contamination, and thus laboratory workspaces and equipment may need to be purchased or thoroughly decontaminated and then reserved exclusively for such work, imposing a not insubstational resource burden upon older laboratories.

The relative importance of mutations in regulatory regions (i.e. promoters and transcription factors) versus structural mutations in coding regions in determining morphological variation, is an ongoing and hotly debated topic

(Hoekstra and Coyne, 2007, Carroll, 2008). This idea that the evolution of form finds its basis in cis- and trans- regulatory elements (which regulate the expression of genes), rather than structural changes in DNA, stems primarily from the expectation that if structural changes were responsible, a great deal more variation in gene content should be observed between divergent taxa

(King and Wilson, 1975, Carroll, 2000). However as pointed out by Hoekstra and Coyne (2007), even the relatively small percentage of their respective

149

General Discussion genomes that differs between chimpanzees and humans is still a large total number of genetic changes (approximately 60 000), of which many have the potential to determine meaningful physiological and morphological variation between these species.

One of the main criticisms levelled at evo-devo is that the majority of studies cited in support of this concept of the dominance of regulatory changes have yet to demonstrate that cis mutations modify particular traits, and instead rely on correlations between cis variants and phenotypes (Hoekstra and Coyne, 2007).

Yet much of the work on non-model animals that has concluded particular structural coding changes are responsible for phenotypic changes has also relied upon genotype-associations; although direct causal mechanisms have been demonstrated for some model taxa and genes, especially in the field of human disease research.

The insistence that the underlying genetic basis of evolutionary change should fundamentally differ between physiological and morphological traits is another aspect of the regulatory idea which has divided researchers. This delineation between form and function has been dismissed by some as a way to ignore the large amount of evidence already gathered that structural genetic variation determines specific physiological traits (Hoekstra and Coyne, 2007). However, a review of the literature conducted by Stern and Orgogozo (2008) concluded that cis-regulatory mutations, although only accounting for approximately one- fifth of trait mutations catalogued, composed a disproportionate proportion

(74%) of mutations implicated in morphological traits.

150

General Discussion A growing number of studies have now linked regulatory mutations to pigmentation variation across a range of animal species (Nadeau et al., 2008,

Kingsley et al., 2009, Manceau et al., 2010, Gunnarsson et al., 2011). Although they can be harder to identify than structural gene mutations owing to their less conserved nature, smaller size, and their often large and variable distance from the gene(s) they regulate (Hoekstra and Coyne, 2007), it is apparent that such mutations are likely be a component of the genetic basis of pigmentation variation deserving of further research across both model and non-model species. The identification of the functional roles of a few transcription factors involved in melanocyte development, including the Microphthamia-associated transcription factor (Mitf) provides a starting point for such research (Hornyak et al., 2001); and variation in the MITF-M isoform has already been linked to regulation of white spotting in dogs (Körberg et al., 2014).

Indeed, scrutiny of regulatory gene regions for associations with magpie back colour may prove fruitful due to the nature of magpie back colour variation; magpie back colour differences could reflect differential distribution and/ or expression of pigments via regulatory mechanisms such as the differential pigmentation distribution along individual hair shafts of deer mice, ascribed to cis-acting mutation(s) in and upstream of the Agouti gene (Linnen et al., 2009).

It is suggested that future studies of magpie pigmentation will benefit from investigating regulatory elements and transcription factors as potential sources of phenotypic variation, once these are more practically achievable with limited resources in a non-model species such as the magpie.

151

General Discussion Pleiotropic genes that have an effect upon more than one trait may mean that colour variation within a species is not a direct reflection of selection upon the colour trait itself, but rather selection acting upon the other trait(s) linked to that gene (Ducrest et al., 2008). An example of this is the lethal yellow mutation, that causes a completely yellow coat in mice, that has also been linked to obesity, diabetes and tumours (Michaud et al., 1994, Miltenberger et al., 1997).

These diverse phenotypic and physiological effects are thought to occur because the ASIP gene which harbours the lethal yellow mutation antagonises not only the MC1R gene (affecting pigmentation), but also the MC2R, MC3R and MC4R genes, expressed predominantly in the brain, that are associated with metabolism and feeding behaviours (Nadeau et al., 2008). Pigmentation- determining genes that co-segregate with, or are tightly linked to other genes under selection, or located in areas of low recombination on chromosomes, may also result in colour phenotypes that are not reflective of direct selection acting upon the colour trait itself (Roulin, 2004).

In humans and mice, the MC1R gene, which is generally thought to have very few pleiotropic effects (Mundy, 2005), has been shown to influence pain tolerance and response to analgesics (Mogil et al., 2005), and a study of

Elanora’s falcons has suggested the gene may to be involved in immune responses in this species (Gangoso et al., 2011). In a similar way, magpie back colour variation could be a result of the effects of a pleiotropic gene(s) affecting both back colour and another morphological or physiological trait such as energy balance, immune factors or even temperature regulation, leading to segregation and distribution of different back colour types across environments

152

General Discussion due to selective pressures acting on the other trait, or on both the other trait and back colour together.

Previous studies of magpie pigmentation have often noted that dark back colour forms and light back colour forms are generally separated by environments with differing levels of canopy cover (Kallioinen et al., 1995, Hughes et al., 2001,

Hughes et al., 2011). The distribution of magpie back colour variants is also fairly congruent with humidity patterns, and seemingly follows Gloger’s Rule

(Zink and Remsen Jr, 1986), with darker morphs closer to the equator (more humid) and lighter morphs further from the equator (less humid). Burtt Jr and

Ichida (2004), in a study of song sparrows in the United States, suggest that lighter feathers may be less resistant than dark feathers to the bacteria that thrive on (and degrade) feathers in high-humidity environments. The genetics of immune factors could also be potentially linked with pigmentation in magpies; in barn swallows Saino et al. (2013a) found belly feather darkness covaried with immune response, with darker individuals showing a lower immune response to a novel antigen than lighter individuals, indicating these traits are likely linked and may even share a pleiotropic genetic basis. Similarly, female barn owls with more and larger black spots in their plumage have been found to have higher resistance to ectoparasitic flies than those with smaller and lighter plumage spotting (Roulin et al., 2001). Interestingly, Hughes (1983) found that loads of an ectoparasitic feather-chewing lice, Brueelia semiannulata (which is thought to exclusively parasitize the Australian magpie) were different across different magpie back colour forms: white backed and varied back birds were found to have significantly fewer infestations than black backed individuals.

153

General Discussion Metabolic factors could also potentially be genetically linked to magpie back colour. Food intake and energy homeostasis have been linked to the same mutation which causes ‘lethal yellow’ pigmentation in mice (discussed earlier), and manipulation studies of tawny owls revealed the ‘darkness’ of nestlings was correlated with their response to food restriction, with darker nestlings losing less weight under food scarcity conditions (Dreiss et al., 2010).

A possible candidate for investigation of such genes of pleiotropic action involved in pigmentation in birds is the agouti-related protein (AGRP), (a portion of which was screened in this study), as well as its closely related peptides.

The AGRP gene is predominantly expressed in the brain of mammals, where it is a known antagonist of the MC3R and MC4R receptors involved in the regulation of food intake and appetite (Zimanyi and Pelleymounter, 2003); and

SNPs in the AGRP gene have been linked to anorexia in humans (Vink et al.,

2001). In the chicken, the AGRP gene has been shown to be expressed across a wide range of tissues, including skin (Takeuchi et al., 2000) where it has been suggested it may be involved in melanin synthesis (Li et al., 2011), putatively via antagonism of the MC1R gene in these tissues (Boswell and Takeuchi,

2005). The attractin (ATRN) gene, a C-type lectin with several isoforms, seems to affect the regulation of both pigmentation (via ASIP) and energy homeostasis (via AGRP) as well as possibly immune function (via T-cell interactions), through pathways involving complicated interactions with the α melanocyte-stimulating hormone in mice, and certain isoforms are a candidate for the mouse mahogany (mg) mutation (Jackson, 1999). Such complex gene interactions may be more commonly involved in animal pigmentation and the selective forces shaping it than previously thought: many pigmentation genes

154

General Discussion have been found to have pleiotropic effects on physiological processes (e.g.

TYR’s involvement in both albinism and ocular disorders) and these are extensively reviewed (for mammals) in Reissmann and Ludwig (2013). It is suggested that the further investigation of such genes may be of great utility in pinning down the genetic basis of colour variation in species such as the magpie, in which pigmentation may conceivably be linked to pleiotropic genes also affecting physiological processes such as metabolism or thermoregulation.

It has been suggested by a number of authors that the efficiency of the candidate gene approach to identifying the underlying genetic basis of traits can be greatly increased through the careful targeting of known polymorphisms, prioritised by their type, location, frequency and predicted functional effects

(Risch and Merikangas, 1996, Risch, 2000, Tabor et al., 2002). Excitingly, large collaborative efforts enabled by rapid technological advances have led to the very recent availability of draft genomes for a large and steadily increasing number of bird taxa (Zhang, 2015), including the Australian magpie, and this will greatly accelerate advances in this field of research. This ‘Bird 10,000

Genomes (B10K) Project’ aims to sequence draft genomes of all living bird species by 2020 (Zhang, 2015) and has already sequenced representatives from every extant bird order. High-density genome scans of candidate regions

(selected based on relevance or prior knowledge e.g. the Z chromosome in flycatchers), or to look for relevant genes and genomic regions involved in a trait of interest should thus become significantly more resource efficient, and in turn this should enable more finely targeted screening of large numbers of candidate genes simultaneously, a process itself made significantly more cost and time- effective by the availability of draft genomes of the species of interest. Such

155

General Discussion methods also offer a way to identify novel and interesting mechanisms of species divergence and evolution, such as the ‘narrow genomic islands’ of genetic divergence that may facilitate reproductive isolation and ultimately speciation in allopatry (Poelstra et al., 2014), or ‘supergenes’ in which tightly linked neighbouring loci are inherited together in a polymorphic manner (e.g. the large-scale chromosomal inversion ‘haplotypes’ in avian ruffs that produce discrete male mating morphs within the species (Kupper et al., 2016)).

5.5 DOES IT MATTER IF IT’S BLACK OR WHITE?:

SELECTION AND MAGPIE BACK COLOUR VARIATION

While the main thrust of this study was to further the search for the genetic basis of magpie back colour variation, the broader contextual question remains as to what is driving the creation and maintenance of these colour variants. The historical genetic split between eastern and western populations (Hughes, 1982,

Hughes et al., 2001, Toon et al., 2003, Toon et al., 2007), lack of restriction to gene flow between black backed north-eastern and white backed south-eastern populations (Baker et al., 2000), genetic structuring ruling out allopatry with secondary recontact in hybrid zones (Hughes et al., 2001), together with a behavioural study that found resident male magpies were more aggressive toward a white backed intruder than a black backed intruder (Kallioinen et al.,

1995) led to the development of an adaptive hypothesis for this plumage variation. Hughes et al. (2001) proposed that divergent selection for crypsis in different environments, alongside sexual selection of white backed males may have created and be maintaining these back colour forms. Some evidence has been found to support the idea that different back colours may have differential

156

General Discussion survival/fitness in different environments. In an eastern hybrid zone population,

Hughes et al. (2002) showed that territories with white backed males produced more fledglings than black backed male-led territories in forest areas, and in areas with more open vegetation, territories with black backed males produced more fledglings. However, the same study could not find significant differences between white back and black back male held territories, in the proportion of territories held, the size of territories, or the number of adult females in territories, or any indication of positive assortative mating (Hughes et al., 2002).

A further study utilised genetically-determined parentage, rather than social parentage, to search for assortative mating patterns (to correct any effects of extra-pair and extra-group paternity), but still did not uncover any sign of female preference for light backed males or assortative mating of any kind (Hughes et al., 2011). Data from the present study also do not support a hypothesis of assortative mating in the eastern hybrid zone. Thus, while it seems plausible that some form of selection linked to vegetative cover and predation may favour black backed birds in the north of the continent, where canopy cover is more open, there is no evidence for sexual selection favouring white backed birds in the south.

All of the studies discussed above have only been able to evaluate metrics based on fledgling production, as nests are generally inaccessible to researchers due to their great height from the ground. Therefore, perhaps a piece of the puzzle relating the advantage of white backs in the south could be due to this. According to Carrick (1972), the highest mortality rates in magpies are in their first year of life, with an average survival rate of 33% of offspring from a nest surviving their first year. Prior to fledgling, offspring may fail to

157

General Discussion hatch, be eaten by predators as eggs or nestlings, starve from insufficient parental feeding, succumb to bacterial, viral or parasitic infection or infestation, or even get irretrievably stuck in branches as they learn to fly. Perhaps the inability to observe offspring in these earlier developmental stages, before they leave the nest, could be obscuring true patterns and differences in fecundity and/or productivity between back colours.

The majority of studies suggesting a balance between crypsis and sexual selection to explain magpie back colour have alluded to, or implicitly suggested, a scenario in which a visually-hunting predator finds white backed individuals, and thus their nests, easier to locate. Although a number of birds of prey have been shown to consume magpies (Leopole and Wolfe, 1970, Carrick, 1972,

Brooker and Ridpath, 1980, Debus, 1984), gut studies generally show that magpies make up a very small proportion of the diet of these species (e.g.

Debus et al., 2007). It may be that other predators of magpies, including monitor lizards, feral cats and foxes, or even snakes (Carrick, 1972, Koboroff et al., 2013) contribute equally or even more significantly to the predation of magpies. Monitor lizards use both sight and smell to hunt, but have particularly powerful abilities to track or locate prey using smell, via chemoreception, using their forked tongue (Wilson, 2012), as do snakes. Cats and foxes hunt primarily at night and both have highly developed senses of hearing and smell that are used in conjunction with sight to hunt. While all these putative predators use vision to hunt, for most it is not their most highly-developed sense; additionally the relative contribution of primarily visually-hunting predators (raptors) to overall magpie predation may have been overstated in the past. In light of this, it seems reasonable to suggest the visual component of predation upon

158

General Discussion magpies may be less dominating, or important in terms of selective pressure than has been previously assumed.

The idea that white backed magpies may be particularly conspicuous in the ultra-violet (UV) spectrum to visual predators (Hughes et al., 2011) also seems less likely with recent work finding most raptor species have very little vision in the UV spectrum (Ödeen and Håstad, 2003, Lind et al., 2013, Lind et al., 2014), and instead are only violet-sensitive. Rather it is more likely that violet-sensitive

(as opposed to UV sensitive) predators perceive magpie back colour forms and contrasts similarly to humans (Seddon et al., 2010) in that white backed individuals stand out more, especially in open environments, but this effect is not amplified further by UV visible wavelengths. Whether magpies themselves have the capacity to perceive UV wavelengths of light is deserving of further investigation; if they can, it may be a situation under which magpies can signal each other (sexually and/or socially) in a matter that is effectively ‘hidden’ from any visual predators, as suggested in Hasted et al. (2005) and Lind et al.

(2013).

Additionally, predation may not be an important enough source of mortality in magpies to drive selection for different back colour types. In early work by

Carrick (1972), the dominant causes of death he observed were: human mediated (road death, poison, shooting) 38%, succumbing to injuries (e.g. from fighting) 11%, disease and infection 9%, and direct predation 8%. While these figures reflect human effects that were not present until the last few hundred years, and are based only on a few years observation of one population, it is clear that predation is only one part of the story. Carrick also observed that

159

General Discussion weather conditions had a large impact upon mortality rates, with fatal bacterial and fungal infections spreading quickly through a magpie population when winters were particularly wet or frosty. Such conditions force many territories and flock groups to come into frequent contact in the search for scarce food, and can lead to waves of high mortality over short time periods (Carrick, 1972).

Selection acting indirectly upon magpie back colour via physiological traits could potentially explain the distribution of magpie back colours and maintenance of seemingly stable hybrid zones, if the genetic basis for these traits were the same or linked to those responsible for back colour. As discussed earlier, a number of physiological traits related to immune function, resistance to parasites and pathogens, and thermoregulation have been found to be associated with pleiotropic genes that are also involved in pigmentation processes in a range of other animal taxa.

If the driver behind magpie back colour variation and hybrid zones is really divergent selection on either side of these hybrid zones, it seems less likely, in the face of current knowledge, that the process favouring white backed birds in the south is sexual selection (through mate choice or another mechanism) for more ‘brightly’ coloured white back males. However, if this sexual selection scenario is indeed true, two main arguments could be made to explain the lack of evidence for sexual selection found thus far. Firstly, as all studies have used fledgling data to investigate signs of sexual selection, it remains possible (but seems unlikely) that some real effect of back colour could be being obscured by the inability to obtain data at the egg or hatchling developmental stages.

Secondly, due to the nature of the long-term study site flock males are generally

160

General Discussion unsampled; thus if, as has been suggested, extra-group matings occur in a pre- dawn environment (Hughes et al., 2003), with females covertly leaving their territories of residence to seek mating opportunities, white blacked males from the flock may simply be easier to visually locate in this low-light environment.

Galeotti et al. (2003) has suggested that “detectability affected by differing light conditions” may be linked to selective mechanisms that drive the evolution of colour polymorphisms in a large number of bird species. However, the data from this study do not support such an idea, as offspring with extra-group paternity were not dominated by light backed individuals, and instead the mean back colour of these ‘fatherless’ offspring was slightly darker than the mean of all genetically typed birds, and nearly 80% of these fatherless offspring had a darker back colour than their genetic mother (data not presented here). This indicates a white or light backed father would be quite unlikely given the high heritability of back colour variation also found in this study. Hughes et al. (2011) posited that a possible mechanism for the “apparent advantage of white backed birds in the south” could simply be that white feathers may cost less to produce in terms of energy expenditure, or that, as suggested by Tickell (2003) feathers may simply be white by default.

Magpies have been observed to display symptoms of heat stress on hot days

(Carrick, 1972), and increase heat dissipating behaviours in temperatures above 27 ºC, causing a decline in foraging effort on hot days (Edwards et al.,

2015). Thus another possibility is that magpie back colour could be linked to adaptation for thermoregulation. Darker back colours should theoretically absorb more visible light, increasing the temperature of the bird, while light back colours should reflect some of this light (and thus heat) away from the bird

161

General Discussion (Heppner, 1970). This would translate to white backed birds having to spend less energy to avoid heat distress in hotter environments, and black backed birds spending less energy maintaining body temperatures in colder environments; in magpies this should theoretically lead to a completely reversed distribution to that actually observed across the Australian continent.

However early work by Walsberg (1983) indicate this may be over-simplistic, as the radiant absorption of light by feather colours is affected by wind levels: in still conditions white feathers heat up less, while in high winds black feathers heat up less. This is primarily a result of the greater penetration of radiation deeper through less-pigmented white feather layers, meaning that in white feathered regions/ birds, less solar heat may be lost to the environment in windy conditions than in darkly pigmented feather regions or individuals (Wolf and

Walsberg, 2000). Walsberg (1983) calculated from data in Walsberg et al.

(1978), that under certain wind conditions, short wave radiation increases thermal heat gain of pigeons of different plumage colours very differently: by up to 30% for black plumage forms, and up to 600% in white plumage forms.

Across the Australian continent, average wind speeds are highest in environments inhabited by white backed (including varied form) magpies and lowest in parts of the continent inhabited by black backed magpies. In fact, areas of change in average wind speeds from windy to fairly calm correspond geographically with the magpie back colour hybrid zones, in both the east and the west of Australia (Mills, 2001) (see Figures 5.1 and 5.2).

162

General Discussion

Figure 5.1: Wind speed across the Australian continent. From the atlas of wind resources for Australia, Mills (2001). Wind speed measured at 70 m between May 1997- April 1999.

Figure 5.2: Distribution of C. tibicen in Australia; adapted from Schodde and Mason, 1999 and Toon, 2007.

Assuming Walsbergs’ (1983) findings are applicable to magpies, this could explain the apparent (and heretofore unexplained) advantage of white backed and varied backed forms in the generally colder southern section of the continent. The southern regions of Australia experience particularly windy conditions during winter, as westerly winds from the roaring 40’s influence these regions (Coppin et al., 2003).

163

General Discussion Passerine bird species generally maintain high core body temperatures of 41 ºC to 42 ºC and have relatively high basal metabolic rates (BMR) (Schmidt-

Nielsen, 1997). In Bech et al.’s (2016) study of BMR in many Australian passerine birds, the two closest relatives of the magpie, the grey butcherbird and the pied currawong, were both found to have high basal metabolic rates of

-1 2.13 and 5.65 ml O2 min respectively. Even though the grey butcherbird is more closely related to the magpie (Kearns et al., 2013), BMR is scaled by body size i.e. Kleiber’s law, thus the BMR of the pied currawong is likely to be a closer reflection of BMR in magpies, as magpies and pied currawongs are both three to four times heavier than grey butcherbirds. When compared to other

-1 birds of a similar size range, 5.65 ml O2 min is a high BMR; in a study by

Londono et al. (2015) the nine birds tested in the 270-400 g weight class had basal metabolic rates of between 0.88 to 2.12 W. If this currawong BMR can be loosely assumed similar to that of magpies, it is apparent that, even in optimal temperatures, the energetic needs of magpies are likely to be very high.

In the cold and windy conditions seen in the south of Australia in winter months, black backed magpies may be at a disadvantage in a thermoregulatory sense.

Maintenance of a constant body temperature composes a significant amount of the energy budget of birds; thermoregulation and its metabolic and behavioural energy costs (e.g. muscle shivering, panting) need to be accounted for before birds can allocate energy to maintenance of their body condition, social interactions and parental tasks or costs (Speakman and Król, 2010). In cold and windy conditions, magpies may have decreased access to food sources, especially when the soil effectively freezes, and both territorial and flock individuals are often forced to leave their territories and/or cover large distances

164

General Discussion daily to forage for food in more sheltered areas (Carrick, 1972). Birds have less daylight hours to forage in winter, and decreased access to food, as well as a need to maintain a high metabolic rate simply to maintain a constant body temperature, in the face of the prolonged cold and wind that characterises

Australia’s southern winter. This may result in selection for white backed magpies in the south of the continent, whom (presumably) have a much greater ability to maximise thermal heat gains from sun exposure, through deeper penetration of light through white feathers to the skin in windy conditions. The location of this ‘patch’ of plumage of varying darkness on the backs of magpies also seems logical under this hypothesis, given magpies are uniquely adapted to ground-feeding, and spend very large portions of their day foraging in a body position that maximises the exposure of this patch to sunlight. Magpies are also frequently observed sunning themselves, often in positions that maximise the exposure of their white feathers to sunlight (Kaplan, 2004, personal observation).

Thus it is suggested that selection on back colour (favouring white backed magpies in the south of Australia) directly or indirectly via selection for genes

(structural or regulatory) involved in thermoregulation, pigmentation or both, could explain the apparent advantage of white backed forms in the south.

Together with the already established hypotheses to explain the advantages of black backed forms in the north of the continent (e.g. humidity/bacterial loads; increased crypsis to aerial predators in more open northern vegetation), this could result in a pattern of differential divergent selection on either side of back colour hybrid zones, as well as the continued maintenance of these hybrid zones (in the face of moderate gene flow) at geographical locations where

165

General Discussion gradients in the relevant environmental variables exist (e.g. vegetation type change, wind speed and minimum temperature).

5.6 CONCLUSIONS

A multigenic basis for magpie back colour seems likely given several aspects of the nature of the traits’ inheritance patterns. These genes may act in tandem or interact to produce the pigmentation variation observed in the Australian magpie. The big question of genetic architecture remains; do few mutations of large effect (structural or regulatory) or many mutations of small effect determine the differences in back colour, or indeed, is the picture complicated by differential selection acting on pleiotropic genes that affect back colour as well as some physiological trait such as metabolism or thermoregulation? It is suggested that genes known to be involved in both pigmentation and feeding, energy balance or especially temperature regulation (and antagonists of such genes) may be fruitful hunting grounds for further studies of the genetic basis of pigmentation in this species. This may be particularly so in light of the scenario outlined in this study, that suggests wind and temperature could play a role in the apparent selective advantage of white backed birds in southern regions of the Australian continent. Regulatory genes, or genes thought to be involved in

‘patchy’ or spotted melanisation or differential patterning of melanic pigments, such as KIT, SLC45A2 and EDNRB2 are also suggested to be deserving of focus in future studies.

While none of the five candidate colour genes screened in this study associated with magpie back colour, the high heritability estimate for back colour variation in an eastern hybrid zone population indicate that environmental contributions to 166

General Discussion variation in this trait are low, and back colour is likely to be predominantly determined by genetic variation. This also indicates that future gene-mapping studies examining back colour variation in this species are likely to have sufficient power to efficiently detect putative loci or gene regions for further investigation. However, the utility of QTL mapping techniques is somewhat restricted in studies of wholly wild populations with uncontrolled crosses, such is likely to remain the case for non-model species such as the magpie. Rather, with recent sequencing of the draft genomes of many bird species, it is suggested targeted high-density genome scanning approaches, coupled with downstream expression and candidate gene analyses of genes and genomic regions identified by these high-density scans have the potential to greatly accelerate progress in this field of research in coming years.

It seems that the genetic basis of magpie back colour variation, as well as the selective and/or neutral forces driving this intraspecific variation in pigmentation, are both complicated and nuanced, and it is suggested that further carefully targeted research investment may yield additional significant insights into the origin, function and maintenance of such pigmentary diversity, and ultimately the evolutionary forces that shape the diversity of life on earth.

167

References

REFERENCES

Allen, J. A. 1974. Further evidence for apostatic selection by wild passerine birds: training experiments. Heredity, 33, 361-372. Allen, J. A. 1976. Further evidence for apostatic selection by wild passerine birds - 9:1 experiments. Heredity, 36, 173-180. Andersson, L. 2003. Melanocortin Receptor Variants with Phenotypic Effects in Horse, Pig, and Chicken. Annals of the New York Academy of Sciences, 994, 313-318. Andersson, M. & Iwasa, Y. 1996. Sexual selection. Trends in Ecology & Evolution, 11, 53-58. Antoniazza, S., Kanitz, R., Neuenschwander, S., Burri, R., Gaigher, A., Roulin, A. & Goudet, J., 2014. Natural selection in a postglacial range expansion: the case of the colour cline in the European barn owl. Molecular ecology, 23, 5508-5523. April, C. S., Jackson, I. J. & Kidson, S. H. 1998. Molecular cloning and sequence analysis of a chicken cDNA encoding tyrosinase-related protein-2/DOPAchrome tautomerase. Gene, 219, 45-53. Avise, J. C., Alisauskas, R. T., Nelson, W. S., & Ankney, C. D. (1992). Matriarchal population genetic structure in an avian species with female natal philopatry. Evolution, 46, 1084-1096. Backström, N., Lindell, J., Zhang, Y., Palkopoulou, E., Qvarnström, A., Sætre, G. P. & Ellegren, H. 2010. A high-density scan of the Z chromosome in Ficedula flycatchers reveals candidiate loci for diversifying selection. Evolution, 64, 3461- 3475. Baiao, P. C., Schreiber, F. A. & Parker, P. G. 2007. The genetic basis of the plumage polymorphism in red-footed boobies (Sula sula): a Metanocortin-I Receptor (MCIR) analysis. Journal of Heredity, 98, 287-292. Baker, A. M., Mather, P. B. & Hughes, J. M. 2000. Population genetic structure of Australian magpies: evidence for regional differences in juvenile dispersal behaviour. Heredity, 85, 167-176. Barker, F. K., Cibois, A., Schikler, P., Feinstein, J. & Cracraft, J. 2004. Phylogeny and diversification of the largest avian radiation. Proceedings of the National Academy of Sciences of the United States of America, 101, 11040-11045. Barton, N. H. 2000. Genetic hitchhiking. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 355, 1553-1562. Barton, N. H. & Hewitt, G. M. 1985. Analysis of hybrid zones. Annual Review of Ecology and Systematics, 113-148. Bech, C., Chappell, M. A., Astheimer, L. B., Londoño, G. A. & Buttemer, W. A. 2016. A ‘slow pace of life’in Australian old-endemic passerine birds is not accompanied by low basal metabolic rates. Journal of Comparative Physiology B, 186, 503- 512. Beebee, T. J. C. & Rowe, G. 2008. An Introduction to Molecular Ecology, Oxford, Oxford University Press. Beldade, P., Mateus, A. R. A. & Keller, R. A. 2011. Evolution and molecular mechanisms of adaptive developmental plasticity. Molecular Ecology, 20, 1347- 1363. Bensch, S., Andersson, T., & Akesson, S. (1999). Morphological and molecular variation across a migratory divide in willow warblers, Phylloscopus trochilus. Evolution, 53, 1925-1935. Berryere, T. G., Kerns, J. A., Barsh, G. S. & Schmutz, S. M. 2005. Association of an Agouti allele with fawn or sable coat color in domestic dogs. Mammalian Genome, 16, 262-272.

168

References Berryere, T. G., Schmutz, S. M., Schimpf, R., Cowan, C. & Potter, J. 2003. TYRP1 is associated with dun coat colour in Dexter cattle or how now brown cow? Animal Genetics, 34, 169-175. Birky, C. W., Fuerst, P. & Maruyama, T. 1989. Organelle Gene Diversity under Migration, Mutation, and Drift - Equilibrium Expectations, Approach to Equilibrium, Effects of Heteroplasmic Cells, and Comparison to Nuclear Genes. Genetics, 121, 613-627. Bize, P., Gasparini, J., Klopfenstein, A., Altwegg, R. & Roulin, A. 2006. Melanin-based coloration is a nondirectionally selected sex-specific signal of offspring development in the alpine swift. Evolution, 60, 2370-2380. Black, A. 1986. The Taxonomic Affinity of the New Guinean Magpie Gymnorhina tibicen papuana. Emu, 86, 65-70. Boissy, R. E., Zhao, H., Oetting, W. S., Austin, L. M., Wildenberg, S. C., Boissy, Y. L., Zhao, Y., Sturm, R. A., Hearing, V. J. & King, R. A. 1996. Mutation in and lack of expression of tyrosinase-related protein-1 (TRP-1) in melanocytes from an individual with brown oculocutaneous albinism: a new subtype of albinism classified as "OCA3". American Journal of Human Genetics, 58, 1145. Bonser, R. H. 1995. Melanin and the abrasion resistance of feathers. The Condor, 97, 590-590. Boonanuntanasarn, S., Yoshizaki, G., Iwai, K. & Takeuchi, T. 2004. Molecular cloning, gene expression in albino mutants and gene knockdown studies of tyrosinase mRNA in rainbow trout. Pigment Cell Research, 17, 413-421. Boss, J., Liedvogel, M., Lundberg, M., Olsson, P., Reischke, N., Naurin, S., Åkesson, S., Hasselquist, D., Wright, A., Grahn, M. and Bensch, S., 2016. Gene expression in the brain of a migratory songbird during breeding and migration. Movement ecology, 4, 4. Boswell, T. & Takeuchi, S. 2005. Recent developments in our understanding of the avian melanocortin system: Its involvement in the regulation of pigmentation and energy homeostasis. Peptides, 26, 1733-1743. Bravery, J. 1970. The birds of Atherton shire, Queensland. Emu, 70, 49-63. Brodin, A. & Haas, F. 2009. Hybrid zone maintenance by non-adaptive mate choice. Evolutionary Ecology, 23, 17-29. Brooker, M. G. & Ridpath, M. 1980. The diet of the Wedge-Tailed Eagle, Aquila audax, in western Australia. Wildlife Research, 7, 433-452. Brooks, S. A. & Bailey, E. 2005. Exon skipping in the KIT gene causes a Sabino spotting pattern in horses. Mammalian Genome, 16, 893-902. Brunberg, E., Andersson, L., Cothran, G., Sandberg, K., Mikko, S. & Lindgren, G. 2006. A missense mutation in PMEL17 is associated with the Silver coat color in the horse. Bmc Genetics, 7. Buckley, P. 1987. Mendelian genes. Avian genetics: A population and ecological approach, 1, 44. Budd, P. S. & Jackson, I. J. 1995. Structure of the mouse tyrosinase-related protein- 2/dopachrome tautomerase (Tyrp2/Dct) gene and sequence of two novel slaty alleles. Genomics, 29, 35-43. Buggiotti, L. 2007. Avian Evolutionary Genomics: Studies of Ficedula Flycatchers. PhD, University of Turku. Burchill, S., Thody, A. & Ito, S. 1986. Melanocyte-stimulating hormone, tyrosinase activity and the regulation of eumelanogenesis and phaeomelanogenesis in the hair follicular melanocytes of the mouse. Journal of Endocrinology, 109, 15-21. Burnham, K. P. & Anderson, D. R. 1998. Model Selection and Inference. Springer. Burton, T. C. & Martin, A. A. 1976. Analysis of hybridization between Black-backed and White-backed Magpies in South-eastern Australia. Emu, 76, 30-36. Burtt, E. H. 1986. An analysis of physical, physiological, and optical aspects of avian coloration with emphasis on wood-warblers. Ornithological Monographs, iii-126. Burtt Jr, E. H. & Ichida, J. M. 2004. Gloger's rule, feather-degrading bacteria, and color variation among song sparrows. The Condor, 106, 681-686.

169

References Bush, S. E., Kim, D., Moyer, B. R., Lever, J., Clayton, D. H. & Fleischer, R. 2006. Is melanin a defense against feather-feeding lice? The Auk, 123, 153-161. Cain, A. J. & Sheppard, M. 1954. Natural selection in Cepaea. Genetics, 39, 89-116. Campbell, A. G. 1929. Australian Magpies of the Genus Gymnorhina. Emu, 28, 165- 176. Campbell, N. A. & Reece, J. B. 2002. Biology, Benjamin Cummings. Carrick, R. 1972. Population Ecology of the Australian black-backed magpie, royal penguin and silver gull. Population Ecology of Migratory Birds: A Symposium. Washington: US Department of the Interior Wildlife Research Report. Carroll, S. B. 2000. Endless forms: the evolution of gene regulation and morphological diversity. Cell, 101, 577-580. Carroll, L., Schuerch, J.,Salamin, N., & Dubey, S.. 2016. Why are some species older than others? A large-scale study of vertebrates. BMC Evolutionary Biology, 16, 90. Cattin, S. B. 2008. Evo-devo and an expanding evolutionary synthesis: a genetic theory of morphological evolution. Cell, 134, 25-36. Chang, C.-M., Coville, J.-L., Coquerelle, G., Gourichon, D., Oulmouden, A. & Tixier- Boichard, M. 2006. Complete association between a retroviral insertion in the tyrosinase gene and the recessive white mutation in chickens. BMC Genomics, 7, 1. Chappell, J. & Shackleton, N. J. 1986. Oxygen Isotopes and Sea-Level. Nature, 324, 137-140. Charmantier, A., Garant, D. & Kruuk, L. E. 2014. Quantitative genetics in the wild, Oxford, Oxford University Press. Charmantier, A. & Reale, D. 2005. How do misassigned paternities affect the estimation of heritability in the wild? Molecular Ecology, 14, 2839-2850. Cheviron, Z. A., Hackett, S. J. & Brumfield, R. T. 2006. Sequence variation in the coding region of the melanocortin-1 receptor gene (MC1R) is not associated with plumage variation in the blue-crowned manakin (Lepidothrix coronata). Proceedings of the Royal Society B-Biological Sciences, 273, 1613-1618. Christidis, L. & Boles, W. 2008. Systematics and taxonomy of Australian birds, Csiro Publishing. Clarke, B. 1962. Balanced polymorphism and the diversity of sympatric species. Taxonomy and Geography, 4, 47-70. Clarke, B. C. 1969. The evidence for apostatic selection. Heredity, 24, 237-352. Clement, M., Posada, D. & Crandall, K. A. 2000. TCS: a computer program to estimate gene genealogies. Molecular Ecology, 9, 1657-1659. Coin, L., O'Reilly, P., Pompyen, Y., Hoggart, C., Calboli, F., HardyWeinberg, I. & Calboli, M. F. 2013. Package ‘MultiPhen’. Collins, F. S., Guyer, M. S. & Chakravarti, A. 1997. Variations on a theme: cataloging human DNA sequence variation. Science, 278, 1580-1581. Cook, L. 1992. The neutral assumption and maintenance of colour morph frequency in mangrove snails. Heredity, 69, 184-189. Cooke, F., Parkin, D., & Rockwell, R. (1988). Evidence of former allopatry of the two color phases of lesser snow geese (Chen caerulescens caerulescens). The Auk, 105, 467-479. Coppin, P., Ayotte, K. & Steggel, N. 2003. Wind resource assessment in Australia: a planners guide, CSIRO Wind Energy Research Unit. Costin, G.-E., Valencia, J. C., Wakamatsu, K., Ito, S., Solano, F., Milac, A. L., Vieira, W. D., Yamaguchi, Y., Rouzaud, F. & Petrescu, A.-J. 2005. Mutations in dopachrome tautomerase (Dct) affect eumelanin/pheomelanin synthesis, but do not affect intracellular trafficking of the mutant protein. Biochemical Journal, 391, 249-259. Darwin, C. 1871. The descent of man, and selection in relation to sex, London, Murray. Darwin, C. & Wallace, A. 1858. On the tendency of species to form varieties; and on the perpetuation of varieties and species by natural means of selection. Journal of the proceedings of the Linnean Society of London. Zoology, 3, 45-62. 170

References Debus, S., Hatfield, T., Ley, A. & Rose, A. 2007. Breeding biology and diet of the Wedge-tailed Eagle Aquila audax in the New England region of New South Wales. Australian Field Ornithology, 24, 93. Debus, S. J. 1984. Biology of the little eagle on the northern tablelands of New South Wales. Emu, 84, 87-92. Deng, W., Tan, Y., Wang, X., Xi, D., He, Y., Yang, S., Mao, H. & Gao, S. 2009. Molecular cloning, sequence characteristics, and polymorphism analyses of the tyrosinase-related protein 2/DOPAchrome tautomerase gene of black-boned sheep (Ovis aries). Genome, 52, 1001-1011. Dobson, A. E., Schmidt, D. J. & Hughes, J. M. 2012. Sequence variation in the melanocortin-1 receptor (MC1R) does not explain continent-wide plumage color differences in the Australian magpie (Cracticus tibicen). Journal of Heredity, 103, 769-780. Doerge, R. W. 2002. Mapping and analysis of quantitative trait loci in experimental populations. Nature Reviews Genetics, 3, 43-52. Doucet, S. M. & Montgomerie, R. 2003. Multiple sexual ornaments in satin bowerbirds: ultraviolet plumage and bowers signal different aspects of male quality. Behavioral Ecology, 14, 503-509. Doucet, S. M., Shawkey, M. D., Rathburn, M. K., Mays, H. L. & Montgomerie, R. 2004. Concordant evolution of plumage colour, feather microstructure and a melanocortin receptor gene between mainland and island populations of a fairy- wren. Proceedings of the Royal Society of London Series B-Biological Sciences, 271, 1663-1670. Doyle, J. J. & Doyle, J. L. 1987. A rapid DNA isolation procedure from small quantities of fresh leaf tissues. Phytochemistry Bulletin, 19, 11-15. Dreger, D. L. & Schmutz, S. M. 2010. A New Mutation in MC1R Explains a Coat Color Phenotype in 2 “Old” Breeds: Saluki and Afghan Hound. Journal of Heredity, 101, 644-649. Dreger, D. L. & Schmutz, S. M. 2011. A SINE insertion causes the black-and-tan and saddle tan phenotypes in domestic dogs. Journal of Heredity, 102, S11-S18. Dreiss, A., Henry, I., Ruppli, C., Almasi, B. & Roulin, A. 2010. Darker eumelanic barn owls better withstand food depletion through resistance to food deprivation and lower appetite. Oecologia, 164, 65-71. Driskell, A. C., Prum, R. O. & Pruett‐Jones, S. 2010. The evolution of black plumage from blue in Australian fairy‐wrens (Maluridae): genetic and structural evidence. Journal of Avian Biology, 41, 505-514. Drovetski, S. V., Zink, R. M. & Mode, N. A. 2008. Patchy distributions belie morphological and genetic homogeneity in rosyfinches. Molecular Phylogenetics and Evolution. Ducrest, A. L., Keller, L. & Roulin, A. 2008. Pleiotropy in the melanocortin system, coloration and behavioural syndromes. Trends in Ecology & Evolution, 23, 502- 510. Durrant, K. L. & Hughes, J. M. 2005. Differing rates of extra-group paternity between two populations of the Australian magpie (Gymnorhina tibicen). Behavioral Ecology and Sociobiology, 57, 536-545. Durrant, K. L. & Hughes, J. M. 2006. Are there correlates of male Australian Magpie Gymnorhina tibicen reproductive success in a population with high rates of extra-group paternity? Ibis, 148, 313-320. Edwards, E. K., Mitchell, N. J. & Ridley, A. R. 2015. The impact of high temperatures on foraging behaviour and body condition in the Western Australian Magpie Cracticus tibicen dorsalis. Ostrich, 86, 137-144. Egea, R., Casillas, S. & Barbadilla, A. 2008. Standard and generalized McDonald- Kreitman test: a website to detect selection by comparing different classes of DNA sites. Nucleic Acids Research, 36, W157-W162. Eizirik, E., Yuhki, N., Johnson, W. E., Menotti-Raymond, M., Hannah, S. S. & O'Brien, S. J. 2003. Molecular genetics and evolution of melanism in the cat family. Current Biology, 13, 448-453. 171

References Ellegren, H. 2007. Molecular evolutionary genomics of birds. Cytogenetic and Genome Research, 117, 120-130. Ellegren, H. 2008. Sequencing goes 454 and takes large‐scale genomics into the wild. Molecular Ecology, 17, 1629-1631. Emaresi, G., Ducrest, A.L., Bize, P., Richter, H., Simon, C.. 2013. Sequencing Pleiotropy in the malanocortin system: expression levels of this system are associated with melanogenesis and pigmentation in the tawny owl. Molecular Ecology, 22, 4915-4930. Evans, S. R. & Sheldon, B. C. 2012. Quantitative genetics of a carotenoid-based color: heritability and persistent natal environmental effects in the great tit. The American Naturalist, 179, 79-94. Falconer, D. S. 1975. Introduction to quantitative genetics, Pearson Education India. Farabaugh, S. M., Brown, E. D. & Hughes, J.M. 1992. Cooperative Territorial Defense in the Australian Magpie, Gymnorhina tibicen (Passeriformes, Cracticidae), a Group-living Songbird. Ethology, 92, 283-292. Finn, P. G. & Hughes, J. M. 2001. Helping behaviour in Australian magpies, Gymnorhina tibicen. Emu, 101, 57-63. Firth, J. A., Hadfield, J. D., Santure, A. W., Slate, J. & Sheldon, B. C. 2015. The influence of nonrandom extra-pair paternity on heritability estimates derived from wild pedigrees. Evolution, 69, 1336-1344. Fitzpatrick, B. M., Shook, K. & Izally, R. 2009. Frequency-dependent selection by wild birds promotes polymorphism in model salamanders. BMC Ecology, 9, 12. Fontanesi, L., Tazzoli, M., Beretti, F. & Russo, V. 2006. Mutations in the melanocortin 1 receptor (MC1R) gene are associated with coat colours in the domestic rabbit (Oryctolagus cuniculus). Animal Genetics, 37, 489-493. Ford, E. B. 1945. Polymorphism. Biological Reviews, 20, 73-88. Fox, D. L. 1976. Animal biochromes and structural colours: physical, chemical, distributional & physiological features of coloured bodies in the animal world, Univ of California Press. Frudakis, T., Thomas, M., Gaskin, Z., Venkateswarlu, K., Chandra, K. S., Ginjupalli, S., Gunturi, S., Natrajan, S., Ponnuswamy, V. K. & Ponnuswamy, K. 2003. Sequences associated with human iris pigmentation. Genetics, 165, 2071-2083. Fu, Y. X. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics, 147, 915-925. Galeotti, P., Rubolini, D., Dunn, P. O. & Fasola, M. 2003. Colour polymorphism in birds: causes and functions. Journal of Evolutionary Biology, 16, 635-646. Gan, H. Y., Li, J. B., Wang, H. M., Gao, Y. D., Liu, W. H., Li, J. P. & Zhong, J. F. 2007. Allele frequencies of TYR and MC1R in Chinese native cattle. Animal Science Journal, 78, 484-488. Gangoso, L., Grande, J. M., Ducrest, A. L., Figuerola, J., Bortolotti, G. R., AndrÉS, J. A. & Roulin, A. 2011. MC1R-dependent, melanin-based colour polymorphism is associated with cell-mediated response in the Eleonora’s falcon. Journal of Evolutionary Biology, no-no. Garrick, R. C., Sunnucks, P. & Dyer, R. J. 2010. Nuclear gene phylogeography using PHASE: dealing with unresolved genotypes, lost alleles, and systematic bias in parameter estimation. Bmc Evolutionary Biology, 10. Gasparini, J., Bize, P., Piault, R., Wakamatsu, K., Blount, J. D., Ducrest, A. L. & Roulin, A. 2009. Strength and cost of an induced immune response are associated with a heritable melanin‐based colour trait in female tawny owls. Journal of Animal Ecology, 78, 608-616. Gratten, J., Beraldi, D., Lowder, B., McRae, A., Visscher, P., Pemberton, J. & Slate, J. 2007. Compelling evidence that a single nucleotide substitution in TYRP1 is responsible for coat-colour polymorphism in a free-living population of Soay sheep. Proceedings of the Royal Society of London B: Biological Sciences, 274, 619-626.

172 References Greene, E., Lyon, B. E., Muehter, V. R., Ratcliffe, L., Oliver, S. J. & Boag, P. T. 2000. Disruptive sexual selection for plumage coloration in a passerine bird. Nature, 407, 1000-1003. Greenwood, P. J. 1980. Mating systems, philopatry and dispersal in birds and mammals. Animal Behaviour, 28, 1140-1162. Gunnarsson, U., Hellstrom, A. R., Tixier-Boichard, M., Minvielle, F., Bed'hom, B., Ito, S., Jensen, P., Rattink, A., Vereijken, A. & Andersson, L. 2007. Mutations in SLC45A2 cause plumage color variation in chicken and Japanese quail. Genetics, 175, 867-877. Gunnarsson, U., Kerje, S., Bed’hom, B., Sahlqvist, A. S., Ekwall, O., Tixier‐Boichard, M., Kämpe, O. & Andersson, L. 2011. The Dark brown plumage color in chickens is caused by an 8.3‐kb deletion upstream of SOX10. Pigment cell & melanoma research, 24, 268-274. Guyonneau, L., Murisier, F., Rossier, A., Moulin, A. & Beermann, F. 2004. Melanocytes and pigmentation are affected in dopachrome tautomerase knockout mice. Molecular and Cellular Biology, 24, 3396-3403. Haas, F., Knape, J. & Brodin, A. 2010. Habitat preferences and positive assortative mating in an avian hybrid zone. Journal of Avian Biology, 41, 237-247. Haas, F., Pointer, M. A., Saino, N., Brodin, A., Mundy, N. I. & Hansson, B. 2008. An analysis of population genetic differentiation and genotype-phenotype association across the hybrid zone of carrion and hooded crows using microsatellites and MC1R. Molecular Ecology. Hadfield, J., Nutall, A., Osorio, D. & Owens, I. 2007. Testing the phenotypic gambit: phenotypic, genetic and environmental correlations of colour. Journal of Evolutionary Biology, 20, 549-557. Hadfield, J. D. 2010. MCMC methods for multi-response generalized linear mixed models: the MCMCglmm R package. Journal of Statistical Software, 33, 1-22. Hadfield, J. D., Burgess, M. D., Lord, A., Phillimore, A. B., Clegg, S. M. & Owens, I. P. 2006. Direct versus indirect sexual selection: genetic basis of colour, size and recruitment in a wild bird. Proceedings of the Royal Society of London B: Biological Sciences, 273, 1347-1353. Hall, T. A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic acids symposium series., 41. Hamilton, W. & Zuk, M. 1982. Heritable true fitness and bright birds: a role for parasites? Science, 218, 384-387. Harrigan, R. J., Mazza, M. E. & Sorenson, M. D. 2008. Computation vs. cloning: evaluation of two methods for haplotype determination. Molecular Ecology Resources, 8, 1239-1248. Hart, N. S. & Hunt, D. M. 2007. Avian visual pigments: characteristics, spectral tuning, and evolution. the american naturalist, 169, S7-S26. Håstad, O., Victorsson, J. & Ödeen, A. 2005. Differences in color vision make passerines less conspicuous in the eyes of their predators. Proceedings of the National Academy of Sciences of the United States of America, 102, 6391- 6394. Hearing, V. J. & Tsukamoto, K. 1991. Enzymatic control of pigmentation in mammals. The FASEB Journal, 5, 2902-2909. Heppner, F. 1970. The metabolic significance of differential absorption of radiant energy by black and white birds. The Condor, 72, 50-59. Hill, G. E. 1991. Plumage coloration is a sexually selected indicator of male quality. Nature, 350, 337-339. Hiragaki, T., Inoue-Murayama, M., Miwa, M., Fujiwara, A., Mizutani, M., Minvielle, F. & Ito, S. 2008. Recessive black is allelic to the yellow plumage locus in Japanese quail and associated with a frameshift deletion in the ASIP gene. Genetics, 178, 771-775. Hoekstra, H. 2006a. Genetics, development and evolution of adaptive pigmentation in vertebrates. Heredity, 97, 222-234.

173

References Hoekstra, H. E. 2006b. Genetics, development and evolution of adaptive pigmentation in vertebrates. Heredity, 97, 222-234. Hoekstra, H. E. & Coyne, J. A. 2007. The locus of evolution: Evo devo and the genetics of adaptation. Evolution, 61, 995-1016. Hoffman, E. A., Schueler, F. W., Jones, A. G. & Blouin, M. S. 2006. An analysis of selection on a colour polymorphism in the northern leopard frog. Molecular Ecology, 15, 2627-2641. Hoffmann, A. A. & Willi, Y. 2008. Detecting genetic responses to environmental change. Nature Reviews Genetics, 9, 421-432. Hooke, R., Allestry, J. & Martyn, J. 1665. Micrographia, or, Some physiological descriptions of minute bodies made by magnifying glasses :with observations and inquiries thereupon London :Jo. Martyn and Ja. Allestry Hornyak, T. J., Hayes, D. J., Chiu, L. Y. & Ziff, E. B. 2001. Transcription factors in melanocyte development: distinct roles for Pax-3 and Mitf. Mechanisms of Development, 101, 47-59. Hubbard, J. K., Uy, J. A. C., Hauber, M. E., Hoekstra, H. E. & Safran, R. J. 2010. Vertebrate pigmentation: from underlying genes to adaptive function. Trends in Genetics, 26, 231-239. Hugall, A. F. & Stuart-Fox, D. 2012. Accelerated speciation in colour-polymorphic birds. Nature, 485, 631-634. Hughes, J. 1983. Distribution of Mallophaga on the Australian Magpie. Aust. Zool, 21. Hughes, J. & Mather, P. 1980. A note on magpie back colours. Sunbird: Journal of the Queensland Ornithological Society, The, 11, 42. Hughes, J. & Mather, P. 1991. Variation in the size of territorial groups in the Australian magpie, Gymnorhina tibicen. Proceedings of the Royal Society of Queensland, 101, 13-19. Hughes, J. M. 1982. An Explanation for the Asymmetrical Hybrid Zone between White- Backed and Black-Backed Magpies. Emu, 82, 50-53. Hughes, J. M., Baker, A. M., De Zylva, G. & Mather, P. B. 2001. A phylogeographic analysis of southern and eastern populations of the Australian magpie: evidence for selection in maintenance of the distribution of two plumage morphs. Biological Journal of the Linnean Society, 74, 25-34. Hughes, J. M., Hesp, J. D. E., Kallioinen, R., Kempster, M., Lange, C. L., Hedstrom, K. E., Mather, P. B., Robinson, A. & Wellbourn, M. J. 1996. Differences in social behaviour between populations of the Australian Magpie Gymnorhina tibicen. Emu, 96, 65-70. Hughes, J. M., Lange, C. L., Mather, P. B. & Robinson, A. 2002. A comparison of fitness components among different plumage morphs of the Australian Magpie, Gymnorhina tibicen. Emu, 102, 331-338. Hughes, J. M., Mather, P. B., Toon, A., Ma, J., Rowley, I. & Russell, E. 2003. High levels of extra-group paternity in a population of Australian magpies Gymnorhina tibicen: evidence from microsatellite analysis. Molecular Ecology, 12, 3441-3450. Hughes, J. M., Toon, A., Mather, P. B. & Lange, C. L. 2011. Maintenance of a hybrid zone: the role of female mate choice. The Auk, 128, 688-695. Hung, H.-Y. & Li, S.-H. 2015. Brightness of melanin-based plumage coloration is a cue to oxidative stress in Himalayan Black Bulbuls (Hypsipetes leucocephalus nigerrimus). Avian Research, 6, 1. Husby, A., Kawakami, T., Rönnegård, L., Smeds, L., Ellegren, H. and Qvarnström, A., 2015. Genome-wide association mapping in a wild avian population identifies a link between genetic and phenotypic variation in a life-history trait. Proceedings of the Royal Society B: Biological Sciences. 282, 20150156. Ito, S. & Wakamatsu, K. 2011. Human hair melanins: what we have learned and have not learned from mouse coat color pigmentation. Pigment cell & melanoma research, 24, 63-74. Jackson, I. J. 1997. Homologous pigmentation mutations in human, mouse and other model organisms. Human Molecular Genetics, 6, 1613-1624. 174

References Jackson, I. J. 1999. The mahogany mouse mutation: further links between pigmentation, obesity and the immune system. Trends in Genetics, 15, 429- 431. Jackson, I. J., Chambers, D. M., Tsukamoto, K., Copeland, N., Gilbert, D. J., Jenkins, N. & Hearing, V. 1992. A second tyrosinase-related protein, TRP-2, maps to and is mutated at the mouse slaty locus. The EMBO journal, 11, 527. Javerzat, S. & Jackson, I. J. 1998. White-based brown (Tyrp1 Bw) is a dominant mutation causing reduced hair pigmentation owing to a chromosomal inversion. Mammalian Genome, 9, 469-471. Jawor, J. M. & Breitwisch, R. 2003. Melanin ornaments, honesty, and sexual selection. The Auk, 120, 249-265. Johnson, R. & Jackson, I. 1992. Light is a dominant mouse mutation resulting in premature cell death. Nature Genetics, 1, 226-229. Kalinowski, S. T., Taper, M. L. & Marshall, T. C. 2007. Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Molecular Ecology, 16, 1099-1106. Kallioinen, R. U. O., Hughes, J. M. & Mather, P. B. 1995. Significance of back colour in territorial interactions in the Australian magpie. Australian Journal of Zoology, 43, 665-673. Kaplan, G. 2004. Australian magpie: biology and behaviour of an unusual songbird, Collingwood, CSIRO publishing. Karell, P., Ahola, K., Karstinen, T., Valkama, J. & Brommer, J. E. 2011. Climate change drives microevolution in a wild bird. Nature Communications, 2, 208. Katoh, K. & Standley, D. M. 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution, 30, 772-780. Kearns, A. 2011. Systematics, phylogeography and evolutionary history of the Australo-Papuan butcherbirds (Artamidae: Cracticus). PhD, University of Queensland. Kearns, A. M., Joseph, L., Edwards, S. V., & Double, M. C. (2009). Inferring the phylogeography and evolutionary history of the splendid fairy‐wren Malurus splendens from mitochondrial DNA and spectrophotometry. Journal of Avian Biology, 40, 7-17. Kearns, A. M., Joseph, L. & Cook, L. G. 2013. A multilocus coalescent analysis of the speciational history of the Australo-Papuan butcherbirds and their allies. Molecular Phylogenetics and Evolution, 66, 941-952. Kelly, J. K. 1997. A test of neutrality based on interlocus associations. Genetics, 146, 1197-1206. Kerje, S., Lind, J., Schutz, K., Jensen, P. & Andersson, L. 2003. Melanocortin 1- receptor (MC1R) mutations are associated with plumage colour in chicken. Animal Genetics, 34, 241-248. Kerje, S., Sharma, P., Gunnarsson, U., Kim, H., Bagchi, S., Fredriksson, R., Schutz, K., Jensen, P., von Heijne, G., Okimoto, R. & Andersson, L. 2004. The Dominant white, Dun and Smoky color variants in chicken are associated with insertion/deletion polymorphisms in the PMEL17 gene. Genetics, 168, 1507- 1518. Kijas, J. M. H., Wales, R., Tornsten, A., Chardon, P., Moller, M. & Andersson, L. 1998. Melanocortin receptor 1 (MC1R) mutations and coat color in pigs. Genetics, 150, 1177-1185. Kim, S., Fargallo, J., Vergara, P. & Martínez-Padilla, J. 2013. Multivariate heredity of melanin-based coloration, body mass and immunity. Heredity, 111, 139-146. Kimura, M. 1984. The neutral theory of molecular evolution, Cambridge University Press. King, M. C. & Wilson, A. C. 1975. Evolution at two levels in humans and chimpanzees. Science, 188, 107-116.

175

References Kingsley, E. P., Manceau, M., Wiley, C. D. & Hoekstra, H. E. 2009. Melanism in Peromyscus is caused by independent mutations in Agouti. PLoS ONE, 4, e6435. Kircher, M. & Kelso, J. 2010. High-throughput DNA sequencing–concepts and limitations. Bioessays, 32, 524-536. Klungland, H., Vage, D. I., Gomezraya, L., Adalsteinsson, S. & Lien, S. 1995. The Role of Melanocyte-Stimulating Hormone (Msh) Receptor in Bovine Coat Color Determination. Mammalian Genome, 6, 636-639. Kobayashi, T. & Hearing, V. J. 2007. Direct interaction of tyrosinase with Tyrp1 to form heterodimeric complexes in vivo. Journal of Cell Science, 120, 4261-4268. Koboroff, A., Kaplan, G. & Rogers, L. 2013. Clever strategists: Australian Magpies vary mobbing strategies, not intensity, relative to different species of predator. PeerJ, 1, e56. Körberg, I. B., Sundström, E., Meadows, J. R., Pielberg, G. R., Gustafson, U., Hedhammar, Å., Karlsson, E. K., Seddon, J., Söderberg, A. & Vilà, C. 2014. A simple repeat polymorphism in the MITF-M promoter is a key regulator of white spotting in dogs. PLoS ONE, 9, e104363. Kroumpouzos, G., Urabe, K., Kobayashi, T., Sakai, C. & Hearing, V. J. 1994. Functional-Analysis of the Slaty Gene-Product (TRP2) as Dopachrome Tautomerase and the Effect of a Point Mutation on Its Catalytic Function. Biochemical and Biophysical Research Communications, 202, 1060-1068. Kruuk, L. & Hadfield, J. 2007. How to separate genetic and environmental causes of similarity between relatives. Journal of Evolutionary Biology, 20, 1890-1903. Kruuk, L. E. B. 2004. Estimating genetic parameters in natural populations using the 'animal model'. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 359, 873-890. Kupper, C., Stocks, M., Risse, J. E., dos Remedios, N., Farrell, L. L., McRae, S. B., Morgan, T. C., Karlionova, N., Pinchuk, P., Verkuil, Y. I., Kitaysky, A. S., Wingfield, J. C., Piersma, T., Zeng, K., Slate, J., Blaxter, M., Lank, D. B. & Burke, T. 2016. A supergene determines highly divergent male reproductive morphs in the ruff. Nat Genet, 48, 79-83. Kwon, B. S., Halaban, R. & Chintamaneni, C. 1989. Molecular basis of mouse Himalayan mutation. Biochemical and Biophysical Research Communications, 161, 252-260. Lahiri, D. K. & Schnabel, B. 1993. DNA Isolation by a Rapid Method from Human Blood-Samples - Effects of Mgcl2, Edta, Storage Time, and Temperature on DNA Yield and Quality. Biochemical Genetics, 31, 321-328. Lane, J.E., Kruuk, L.E.B., Charmantier, A., Murie, J.O., Coltman, D.W., Buoro, M., Raveh, S. and Dobson, F.S. 2011. A quantitative genetic analysis of hibernation emergence date in a wild population of Columbian ground squirrels. Journal of Evolutionary Biology, 24, 1949-1959. Leopole, A. S. & Wolfe, T. 1970. Food habits of nesting Wedge-tailed Eagles, Aquila audax, in south-eastern Australia. Wildlife Research, 15, 1-17. Li, Y., Zhu, X., Yang, L., Li, J., Lian, Z., Li, N. & Deng, X. 2011. Expression and network analysis of genes related to melanocyte development in the Silky Fowl and White Leghorn embryos. Molecular Biology Reports, 38, 1433-1441. Lind, O., Mitkus, M., Olsson, P. & Kelber, A. 2013. Ultraviolet sensitivity and colour vision in raptor foraging. The Journal of Experimental Biology, 216, 1819-1826. Lind, O., Mitkus, M., Olsson, P. & Kelber, A. 2014. Ultraviolet vision in birds: the importance of transparent eye media. Proceedings of the Royal Society B: Biological Sciences, 281. Ling, M. K., Lagerstrom, M. C., Fredriksson, R., Okimoto, R., Mundy, N. I., Takeuchi, S. & Schioth, H. B. 2003. Association of feather colour with constitutively active melanocortin 1 receptors in chicken. European Journal of Biochemistry, 270, 1441-1449. Linnen, C. R., Kingsley, E. P., Jensen, J. D. & Hoekstra, H. E. 2009. On the origin and spread of an adaptive allele in deer mice. Science, 325, 1095-1098. 176

References Linville, S. U., Breitwisch, R. & Schilling, A. J. 1998. Plumage brightness as an indicator of parental care in northern cardinals. Animal Behaviour, 55, 119-127. Londono, G. A., Chappell, M. A., Castañeda, M. d. R., Jankowski, J. E. & Robinson, S. K. 2015. Basal metabolism in tropical birds: latitude, altitude, and the ‘pace of life’. Functional Ecology, 29, 338-346. Losdat, S., Helfenstein, F., Blount, J.D. and Richner, H. 2014. Resistance to oxidative stress shows low heritability and high common environmental variance in a wild bird. Journal of Evolutionary Biology, 27, 1990-2000. Luttikhuizen, P. C. & Drent, J. 2008. Inheritance of Predominantly Hidden Shell Colours in Macoma Balthica (L.) (Bivalvia:Tellinidae). Journal of Molluscan Studies, 74, 363-371. Lynch, M. & Walsh, B. 1998. Genetics and analysis of quantitative traits, Sinauer Sunderland, MA. MacDougall-Shackleton, E. A., Blanchard, L. & Gibbs, H. L. 2003. Unmelanized plumage patterns in old world leaf warblers do not correspond to sequence variation at the melanocortin-1 receptor locus (MC1R). Molecular Biology and Evolution, 20, 1675-1681. Mackintosh, J. A. 2001. The antimicrobial properties of melanocytes, melanosomes and melanin and the evolution of black skin. Journal of Theoretical Biology, 211, 101-113. Majerus, M. E. 1998. Melanism: evolution in action, Oxford University Press. Majumdar, K. C., Nasaruddin, K. & Ravinder, K. 1997. Pink body colour in Tilapia shows single gene inheritance. Aquaculture Research, 28, 581-589. Manceau, M., Domingues, V. S., Linnen, C. R., Rosenblum, E. B. & Hoekstra, H. E. 2010. Convergence in pigmentation at multiple levels: mutations, genes and function. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 365, 2439-2450. Manga, P., Kromberg, J., Box, N., Sturm, R., Jenkins, T. & Ramsay, M. 1997. Rufous oculocutaneous albinism in southern African Blacks is caused by mutations in the TYRP1 gene. The American Journal of Human Genetics, 61, 1095-1101. Manga, P., Sato, K., Ye, L., Beermann, F., Lamoreux, M. & Orlow, S. J. 2000. Mutational Analysis of the Modulation of Tyrosinase by Tyrosinase‐Related Proteins 1 and 2 In Vitro. Pigment Cell Research, 13, 364-374. Marklund, L., Moller, M. J., Sandberg, K. & Andersson, L. 1996. A missense mutation in the gene for melanocyte-stimulating hormone receptor (MC1R) is associated with the chestnut coat color in horses. Mammalian Genome, 7, 895-899. Marshall, H. & Ritland, K. 2002. Genetic diversity and differentiation of Kermode bear populations. Molecular Ecology, 11, 685-697. Marshall, T. C., Slate, J., Kruuk, L. E. B. & Pemberton, J. M. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Molecular Ecology, 7, 639-655. McDonald, J. H. & Kreitman, M. 1991. Adaptive protein evolution at the Adh locus in Drosophila. Nature, 351, 652-654. McGraw, K. J. 2006. Bird coloration: mechanisms and measurements, Harvard University Press. Merilä, J. & Sheldon, B. C. 2001. Avian quantitative genetics. Current Ornithology. Springer. Michaud, E. J., Bultman, S. J., Klebig, M. L., Van Vugt, M. J., Stubbs, L. J., Russell, L. B. & Woychik, R. P. 1994. A molecular model for the genetic and phenotypic characteristics of the mouse lethal yellow (Ay) mutation. Proceedings of the National Academy of Sciences, 91, 2562-2566. Michaud, E. J., Bultman, S. J., Stubbs, L. J. & Woychik, R. P. 1993. The embryonic lethality of homozygous lethal yellow mice (A(y)/A(y)) is associated with the disruption of a novel RNA-binding protein. Genes & Development, 7, 1203- 1213. Miller, R. G. J. 1981. Simultaneous statistical inference, N.Y., McGraw Hill.

177

References Mills, D. 2001. Assessing the potential contribution of renewable energy to electricity supply in Australia. PhD, University of Queensland. Miltenberger, R. J., Mynatt, R. L., Wilkinson, J. E. & Woychik, R. P. 1997. The role of the agouti gene in the yellow obese syndrome. The Journal of nutrition, 127, 1902S-1907S. Miwa, M., Inoue-Murayama, M., Aoki, H., Kunisada, T., Hiragaki, T., Mizutani, M. & Ito, S. 2007. Endothelin receptor B2 (EDNRB2) is associated with the panda plumage colour mutation in Japanese quail. Animal Genetics, 38, 103-108. Mochii, M., Iio, A., Yamamoto, H., Takeuchi, T. & Eguchi, G. 1992. Isolation and characterization of a chicken tyrosinase cDNA. Pigment Cell Research, 5, 162- 167. Mogil, J. S., Ritchie, J., Smith, S. B., Strasburg, K., Kaplan, L., Wallace, M., Romberg, R., Bijl, H., Sarton, E. & Fillingim, R. 2005. Melanocortin-1 receptor gene variants affect pain and μ-opioid analgesia in mice and humans. Journal of Medical Genetics, 42, 583-587. Moore, W. S. 1995. Inferring phylogenies from mtDNA variation: mitochondrial-gene trees versus nuclear-gene trees. Evolution, 49, 718-726. Morales, J., KIM, S.Y., Lobato, E., Merino, S., Tomás, G., Martinez‐de la Puente, J. and Moreno, J., 2010. On the heritability of blue‐green eggshell coloration. Journal of Evolutionary Biology, 23,1783-1791. Mountjoy, K. G., Robbins, L. S., Mortrud, M. T. & Cone, R. D. 1992. The Cloning of a Family of Genes That Encode the Melanocortin Receptors. Science, 257, 1248- 1251. Moura, A.E., van Rensburg, C.J., Pilot, M., Tehrani, A., Best, P.B., Thornton, M., Plön, S., de Bruyn, P.N., Worley, K.C., Gibbs, R.A. & Dahlheim, M.E.2014. Killer whale nuclear genome and mtDNA reveal widespread population bottleneck during the last glacial maximum. Molecular biology and evolution, 31, 1121- 1131. Mundy, N. I. 2005. A window on the genetics of evolution: MC1R and plumage colouration in birds. Proceedings of the Royal Society B-Biological Sciences, 272, 1633-1640. Mundy, N. I., Badcock, N. S., Hart, T., Scribner, K., Janssen, K. & Nadeau, N. J. 2004. Conserved genetic basis of a quantitative plumage trait involved in mate choice. Science, 303, 1870-1873. Nachman, M. W., Hoekstra, H. E. & D'Agostino, S. L. 2003. The genetic basis of adaptive melanism in pocket mice. Proceedings of the National Academy of Sciences of the United States of America, 100, 5268-5273. Nadeau, N. J., Minvielle, F., Ito, S. i., Inoue-Murayama, M., Gourichon, D., Follett, S. A., Burke, T. & Mundy, N. I. 2008. Characterization of Japanese Quail yellow as a Genomic Deletion Upstream of the Avian Homolog of the Mammalian ASIP (agouti) Gene. Genetics, 178, 777-786. Nadeau, N. J., Minvielle, F. & Mundy, N. I. 2006. Association of a Glu92Lys substitution in MC1R with extended brown in Japanese quail (Coturnix japonica). Animal Genetics, 37, 287-289. Nadeau, N. J., Mundy, N. I., Gourichon, D. & Minvielle, F. 2007. Association of a single-nucleotide substitution in TYRP1 with roux in Japanese quail (Coturnix japonica). Animal Genetics, 38, 609-613. Newton, J. M., Wilkie, A. L., He, L., Jordan, S. A., Metallinos, D. L., Holmes, N. G., Jackson, I. J. & Barsh, G. S. 2000. Melanocortin 1 receptor variation in the domestic dog. Mammalian Genome, 11, 24-30. Norris, D. R., Marra, P. P., Kyser, T. K., Ratcliffe, L. M. & Montgomerie, R. 2007. Continent-wide variation in feather colour of a migratory songbird in relation to body condition and moulting locality. Biology Letters, 3, 16-19. Ödeen, A., Hart, N. S. & Håstad, O. 2009. Assessing the use of genomic DNA as a predictor of the maximum absorbance wavelength of avian SWS1 opsin visual pigments. Journal of Comparative Physiology A, 195, 167-173.

178 References Ödeen, A. & Håstad, O. 2003. Complex distribution of avian color vision systems revealed by sequencing the SWS1 opsin from total DNA. Molecular Biology and Evolution, 20, 855-861. Ödeen, A., Håstad, O. & Alström, P. 2011. Evolution of ultraviolet vision in the largest avian radiation - the passerines. Bmc Evolutionary Biology, 11, 1-8. Oetting, W. S. & King, R. A. 1993. Molecular basis of type I (tryrosinase-related) oculocutaneous albinism: Mutations and polymorphisms of the human tyrosinase gene. Human Mutation, 2, 1-6. Ollmann, M. M., Wilson, B. D., Yang, Y. K., Kerns, J. A., Chen, Y. R., Gantz, I. & Barsh, G. S. 1997. Antagonism of central melanocortin receptors in vitro and in vivo by Agouti-related protein. Science, 278, 135-138. Paabo, S. & Wilson, A. C. 1988. Polymerase Chain-Reaction Reveals Cloning Artifacts. Nature, 334, 387-388. Paulson, D. R. 1973. Predator Polymorphism and Apostatic Selection. Evolution, 27, 269-277. Piertney, S. B. & Webster, L. M. I. 2010. Characterising functionally important and ecologically meaningful genetic diversity using a candidate gene approach. Genetica, 138, 419-432. Pigliucci, M. 1996. How organisms respond to environmental changes: From phenotypes to molecules (and vice versa). Trends in Ecology & Evolution, 11, 168-173. Poelstra, J. W., Ellegren, H. & Wolf, J. 2013. An extensive candidate gene approach to speciation: diversity, divergence and linkage disequilibrium in candidate pigmentation genes across the European crow hybrid zone. Heredity, 111, 467- 473. Poelstra, J. W., Vijay, N., Bossu, C. M., Lantz, H., Ryll, B., Müller, I., Baglione, V., Unneberg, P., Wikelski, M. & Grabherr, M. G. 2014. The genomic landscape underlying phenotypic integrity in the face of gene flow in crows. Science, 344, 1410-1414. Potti, J. & Canal, D. 2011. Heritability and genetic correlation between the sexes in a songbird sexual ornament. Heredity, 106, 945-954. Price, T. D. 2006. Phenotypic plasticity, sexual selection and the evolution of colour patterns. Journal of Experimental Biology, 209, 2368-2376. Protas, M. E. & Patel, N. H. 2008. Evolution of Coloration Patterns. Annual Review of Cell and Developmental Biology, 24, 425-446. Pryke, S. R., & Griffith, S. C. (2009). Postzygotic genetic incompatibility between sympatric color morphs. Evolution, 63, 793-798. Quesada, J. & Senar, J. 2009. Cross-fostering experiments to compare carotenoid-and melanin-based plumage traits and long-term parental effects in post-moulted great tits. Behaviour, 146, 1235-1251. Quillan, J. M., Sadée, W., Wei, E. T., Jimenez, C., Ji, L. & Chang, J.-K. 1998. A synthetic human Agouti‐related protein‐(83–132)‐NH2 fragment is a potent inhibitor of melanocortin receptor function. FEBS Letters, 428, 59-62. R Development Core Team 2013. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. R Studio Team 2015. RStudio: integrated development for R. RStudio, Inc., Boston, MA. URL http://www.RStudio.com/ide Ramos-Onsins, S. E. & Rozas, J. 2002. Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution, 19, 2092-2100. Raymond, F. & Riousset, M. 1995. Population genetics software for exact tests and ecumenism. Journal of Heredity, 86, 248-249. Real, K. M., Schmidt, D. J. & Hughes, J. M. 2009. Mogurnda adspersa microsatellite markers: multiplexing and multi-tailed primer tagging. Conservation Genetics Resources, 1, 411-414. Reissmann, M. & Ludwig, A. 2013. Pleiotropic effects of coat colour-associated mutations in humans, mice and other mammals. Seminars in Cell & Developmental Biology, 24, 576-586. 179

References Ren, J., Mao, H., Zhang, Z., Xiao, S., Ding, N. & Huang, L. 2011. A 6-bp deletion in the TYRP1 gene causes the brown colouration phenotype in Chinese indigenous pigs. Heredity, 106, 862-868. Renieri, C., Valbonesi, A., La Manna, V., Antonini, M. & Lauvergne, J. J. 2008. Inheritance of coat colour in Merino sheep. Small Ruminant Research, 74, 23- 29. Risch, N. & Merikangas, K. 1996. The future of genetic studies of complex human diseases. Science, 273, 1516-1517. Risch, N. J. 2000. Searching for genetic determinants in the new millennium. Nature, 405, 847-856. Ritland, K. 1996. A marker-based method for inferences about quantitative inheritance in natural populations. Evolution, 1062-1073. Ritland, K., Newton, C. & Marshall, H. D. 2001. Inheritance and population structure of the white-phased "Kermode" black bear. Current Biology, 11, 1468-1472. Robbins, L. S., Nadeau, J. H., Johnson, K. R., Kelly, M. A., Rosellirehfuss, L., Baack, E., Mountjoy, K. G. & Cone, R. D. 1993. Pigmentation Phenotypes of Variant Extension Locus Alleles Result from Point Mutations That Alter Msh Receptor Function. Cell, 72, 827-834. Robinson, A. 1956. The annual reproductory cycle of the Magpie, Gymnorhina dorsalis Campbell, in south-western Australia. . Emu, 56, 235-336. Rosenblum, E. B., Hoekstra, H. E. & Nachman, M. W. 2004. Adaptive reptile color variation and the evolution of the Mc1r gene. Evolution, 58, 1794-1808. Roulin, A. 2004. The evolution, maintenance and adaptive function of genetic colour polymorphism in birds. Biological Reviews, 79, 815-848. Roulin, A., Almasi, B., Meichtry‐Stier, K. & Jenni, L. 2011. Eumelanin and pheomelanin-based colour advertise resistance to oxidative stress in opposite ways. Journal of Evolutionary Biology, 24, 2241-2247. Roulin, A. & Dijkstra, C. 2003. Genetic and environmental components of variation in eumelanin and phaeomelanin sex-traits in the barn owl. Heredity, 90, 359-364. Roulin, A. & Ducrest, A.-L. 2013. Genetics of colouration in birds. Seminars in Cell & Developmental Biology, 24, 594-608. Roulin, A., Riols, C., Dijkstra, C. & Ducrest, A.-L. 2001. Female plumage spottiness signals parasite resistance in the barn owl (Tyto alba). Behavioral Ecology, 12, 103-110. Rozas, J., Sanchez-DelBarrio, J. C., Messeguer, X. & Rozas, R. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics, 19, 2496-2497. Rozen, S. and Skaletsky, H., 1999. Primer3 on the WWW for general users and for biologist programmers. In Bioinformatics methods and protocols: methods in molecular biology, pp.365-386. New Jersey, United States. Saino, N., Canova, L., Costanzo, A., Rubolini, D., Roulin, A. & Møller, A. P. 2013a. Immune and stress responses covary with melanin-based coloration in the barn swallow. Evolutionary Biology, 40, 521-531. Saino, N., Romano, M., Rubolini, D., Teplitsky, C., Ambrosini, R., Caprioli, M., Canova, L. & Wakamatsu, K. 2013b. Sexual Dimorphism in Melanin Pigmentation, Feather Coloration and Its Heritability in the Barn Swallow (Hirundo rustica). PLoS ONE, 8, e58024. San-Jose, L.M., Ducrest, A.L., Ducret, V., Simon, C., Richter,H., Wakamatsu, K., Roulin, A.. 2017.MC1R varients affect the expression of melanocortin and melanogenic genes and the association between melanocortin genes and coloration.. Molecular Ecology, 26, 259-276. Sato, S., Otake, T., Suzuki, C., Saburi, J. & Kobayashi, E. 2007. Mapping of the recessive white locus and analysis of the tyrosinase gene in chickens. Poultry Science, 86, 2126-2133. Schlotterer, C. 2000. Evolutionary dynamics of microsatellite DNA. Chromosoma, 109, 365-371.

180

References Schmidt-Kuntzel, A., Eizirik, E., O'Brien, S. J. & Menotti-Raymond, M. 2005. Tyrosinase and tyrosinase related protein 1 alleles specify domestic cat coat color phenotypes of the albino and brown loci. Journal of Heredity, 96, 289-301. Schmidt-Nielsen, K. 1997. Animal physiology: adaptation and environment, Cambridge, Cambridge University Press. Schmidt, S. K. 1998. The role of natural selection in the maintenance of the shell colour polymorphism in mangrove snails of the genus Littoraria.. Doctor of Philosophy, Griffith University. Schmutz, S. M. & Berryere, T. G. 2007. The Genetics of Cream Coat Color in Dogs. Journal of Heredity, 98, 544-548. Schmutz, S. M., Berryere, T. G., Ciobanu, D. C., Mileham, A. J., Schmidtz, B. H. & Fredholm, M. 2004. A form of albinism in cattle is caused by a tyrosinase frameshift mutation. Mammalian Genome, 15, 62-67. Schmutz, S. M., Berryere, T. G. & Goldfinch, A. D. 2002. TYRP1 and MC1R genotypes and their effects on coat color in dogs. Mammalian Genome, 13, 380-387. Schodde, R. & Mason, I. J. 1999. The Directory of Australian Birds: Passerines, Canberra, CSIRO. Seddon, N., Tobias, J. A., Eaton, M. & Odeen, A. 2010. Human Vision Can Provide a Valid Proxy for Avian Perception of Sexual Dichromatism The Auk, 127, 283- 292. Shibahara, S., Torruta, Y., Sakakura, T., Nager, C., Chaudhuri, B. & Müller, R. 1986. Cloning and expression of cDNA encoding mouse tyrosinase. Nucleic Acids Research, 14, 2413-2427. Shigemiya, Y. 2004. Reversible frequency-dependent predation of a puffer, Takifugu niphobles (Pisces: Tetreodontidae), related to spatial distribution of colour- polymorphic prey. Biological Journal of the Linnean Society., 81. Silventoinen, K., Sammalisto, S., Perola, M., Boomsma, D. I., Cornes, B. K., Davis, C., Dunkel, L., De Lange, M., Harris, J. R. & Hjelmborg, J. V. 2003. Heritability of adult body height: a comparative study of twin cohorts in eight countries. Twin research, 6, 399-408. Skoglund, P. & Hoglund, J. 2010. Sequence Polymorphism in Candidate Genes for Differences in Winter Plumage between Scottish and Scandinavian Willow Grouse (Lagopus lagopus). PLoS ONE, 5. Slate, J., Marshall, T. & Pemberton, J. 2000. A retrospective assessment of the accuracy of the paternity inference program CERVUS. Molecular Ecology, 9, 801-808. Slatkin, M. 1987. Gene flow and geographic structure of natural populations. Science, 236, 787-792. Slatkin, M. & Hudson, R. R. 1991. Pairwise comparisons of mitochondrial-DNA sequences in stable and exponentially growing populations. Genetics, 129, 555- 562. Speakman, J. R. & Król, E. 2010. Maximal heat dissipation capacity and hyperthermia risk: neglected key factors in the ecology of endotherms. Journal of Animal Ecology, 79, 726-746. Spiegelhalter, D. J., Best, N. G., Carlin, B. P. & van der Linde, A. 2002. Bayesian Measures of Model Complexity and Fit. Journal of the Royal Statistical Society. Series B (Statistical Methodology), 64, 583-639. Spritz, R. A. & Hearing Jr, V. J. 1994. Genetic disorders of pigmentation. Advances in human genetics. Springer. Stephens, M. & Scheet, P. 2005. Accounting for decay of linkage disequilibrium in haplotype inference and missing-data imputation. American Journal of Human Genetics, 76, 449-462. Stephens, M., Smith, N. J. & Donnelly, P. 2001. A new statistical method for haplotype reconstruction from population data. American Journal of Human Genetics, 68, 978-989. Stern, D. L. & Orgogozo, V. 2008. The loci of evolution: How predictable is genetic evolution? Evolution, 62, 2155-2177. 181

References Sturm, R. A., O'Sullivan, B. J., Box, N. F., Smith, A. G., Smit, S. E., Puttick, E. R., Parsons, P. G. & Dunn, I. S. 1995. Chromosomal structure of the human TYRP1 and TYRP2 loci and comparison of the tyrosinase-related protein gene family. Genomics, 29, 24-34. Tabor, H. K., Risch, N. J. & Myers, R. M. 2002. Candidate-gene approaches for studying complex genetic traits: practical considerations. Nature Reviews Genetics, 3, 391-A396. Tajima, F. 1989. Statistical-Method for Testing the Neutral Mutation Hypothesis by DNA Polymorphism. Genetics, 123, 585-595. Takeuchi, S., Suzuki, H., Yabuuchi, M. & Takahashi, S. 1996. A possible involvement of melanocortin 1-receptor in regulating feather color pigmentation in the chicken. Biochimica Et Biophysica Acta-Gene Structure and Expression, 1308, 164-168. Takeuchi, S., Teshigawara, K. & Takahashi, S. 2000. Widespread expression of Agouti-related protein (AGRP) in the chicken: a possible involvement of AGRP in regulating peripheral melanocortin systems in the chicken. Biochimica Et Biophysica Acta-Molecular Cell Research, 1496, 261-269. Team, R. C. 2015. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2014. R Foundation for Statistical Computing. ISBN 3-900051-07-0. http://www.R-project.org. Theron, E., Hawkins, K., Bermingham, E., Ricklefs, R. E. & Mundy, N. I. 2001. The molecular basis of an avian plumage polymorphism in the wild: A melanocortin- 1-receptor point mutation is perfectly associated with the melanic plumage morph of the bananaquit, Coereba flaveola. Current Biology, 11, 550-557. Thirumoorthy, R., Holder, J. R., Bauzo, R. M., Richards, N. G., Edison, A. S. & Haskell- Luevano, C. 2001. Novel agouti-related-protein-based melanocortin-1 receptor antagonist. Journal of Medicinal Chemistry, 44, 4114-4124. Thomas, S., Coltman, D. & Pemberton, J. 2002. The use of marker‐based relationship information to estimate the heritability of body weight in a natural population: a cautionary tale. Journal of Evolutionary Biology, 15, 92-99. Thomas, S. C. & Hill, W. G. 2000. Estimating quantitative genetic parameters using sibships reconstructed from marker data. Genetics, 155, 1961-1972. Thomson, G. M. 1922. The Naturalisation of Animals and Plants in New Zealand, Cambridge, Cambridge University Press. Tickell, W. L. N. 2003. White plumage. Waterbirds, 26, 1-12. Tobita-Teramoto, T., Jang, G., Kino, K., Salter, D., Brumbaugh, J. & Akiyama, T. 2000. Autosomal albino chicken mutation (ca/ca) deletes hexanucleotide (- ΔGACTGG817) at a copper-binding site of the tyrosinase gene. Poultry Science, 79, 46-50. Toon, A. 2007. Reconstructing the evolutionary history of the Australian magpie (Gymnorhina tibicen): patterns of molecular variation in a widespread passerine and two species of obligate feather ectoparasites. Ph.D., Griffith University. Toon, A., Hughes, J., Baker, A. & Mather, P. 2003. Discordance between morphology and genetic structure among three plumage forms of the Australian Magpie. Emu, 103, 337-343. Toon, A. & Hughes, J. M. 2008. Are lice good proxies for host history? A comparative analysis of the Australian magpie, Gymnorhina tibicen, and two species of feather louse. Heredity. Toon, A., Mather, P. B., Baker, A. M., Durrant, K. L. & Hughes, J. M. 2007. Pleistocene refugia in an arid landscape: analysis of a widely distributed Australian passerine. Molecular Ecology, 16, 2525-2541. Uy, J. Â AlbertÂ C., Moyle, RobertÂ G., Filardi, ChristopherÂ E. & Cheviron, ZacharyÂ A. 2009. Difference in Plumage Color Used in Species Recognition between Incipient Species Is Linked to a Single Amino Acid Substitution in the Melanocortinâ€•1 Receptor. The American Naturalist, 174, 244-254. Vage, D. I., Fuglei, E., Snipstad, K., Beheirn, J., Landsem, V. M. & Klungland, H. 2005. Two cysteine substitutions in the MC1R generate the blue variant of the arctic 182

References fox (Alopex lagopus) and prevent expression of the white winter coat. Peptides, 26, 1814-1817. Vage, D. I., Klungland, H., Lu, D. & Cone, R. D. 1999. Molecular and pharmacological characterization of dominant black coat color in sheep. Mammalian Genome, 10, 39-43. Vage, D. I., Lu, D. S., Klungland, H., Lien, S., Adalsteinsson, S. & Cone, R. D. 1997. A non-epistatic interaction of agouti and extension in the fox, Vulpes vulpes. Nature Genetics, 15, 311-315. Valverde, P., Healy, E., Jackson, I., Rees, J. L. & Thody, A. J. 1995. Variants of the Melanocyte-Stimulating Hormone-Receptor Gene Are Associated with Red Hair and Fair Skin in Humans. Nature Genetics, 11, 328-330. Van Oosterhout, C., Hutchinson, W. F., Wills, D. P. M. & Shipley, P. 2004. MICRO- CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes, 4, 535-538. Vasemagi, A. & Primmer, C. R. 2005. Challenges for identifying functionally important genetic variation: the promise of combining complementary research strategies. Molecular Ecology, 14, 3623-3642. Vallone, P.M. and Butler, J.M., 2004. AutoDimer: a screening tool for primer- dimer and hairpin structures. Biotechniques, 37, 226-231. Veltman, C. J. & Carrick, R. 1990. Male-biased dispersal in Australian magpies. Animal Behaviour, 40, 190-192. Veltman, C. J. & Hickson, R. E. 1989. Predation by Australian Magpies (Gymnorhina- tibicen) on Pasture Invertebrates - Are Non-Territorial Birds Less Successful. Australian Journal of Ecology, 14, 319-326. Vink, T., Hinney, A., Van Elburg, A., Van Goozen, S., Sandkuijl, L., Sinke, R., Herpertz- Dahlmann, B., Hebebrand, J., Remschmidt, H. & Van Engeland, H. 2001. Association between an agouti-related protein gene polymorphism and anorexia nervosa. Molecular Psychiatry, 6, 325-328. Visscher, P. M., Hill, W. G. & Wray, N. R. 2008. Heritability in the genomics era— concepts and misconceptions. Nature Reviews Genetics, 9, 255-266. Vorobyev, M., Osorio, D., Bennett, A. T., Marshall, N. & Cuthill, I. 1998. Tetrachromacy, oil droplets and bird plumage colours. Journal of Comparative Physiology A, 183, 621-633. Walsberg, G., Campbell, G. & King, J. 1978. Animal coat color and radiative heat gain: a re-evaluation. Journal of comparative Physiology, 126, 211-222. Walsberg, G. E. 1983. Coat color and solar heat gain in animals. Bioscience, 33, 88- 91. Wang, J., Hou, L., Zhang, R., Zhao, X., Jiang, L., Sun, W., An, J. & Li, X. 2007. The tyrosinase gene family and albinism in fish. Chinese Journal of Oceanology and Limnology, 25, 191-198. Wang, K. 2014. Testing genetic association by regressing genotype over multiple phenotypes. PLoS ONE, 9, e106918. West-Eberhard, M. J. 1989. Phenotypic Plasticity and the Origins of Diversity. Annual Review of Ecology and Systematics, 20, 249-278. Wheelan, S. J., Murillo, F. M. & Boeke, J. D. 2008. The incredible shrinking world of DNA microarrays. Molecular Biosystems, 4, 726-732. Wilson, A. J., Reale, D., Clements, M. N., Morrissey, M. M., Postma, E., Walling, C. A., Kruuk, L. E. B. & Nussey, D. H. 2010. An ecologist's guide to the animal model. Journal of Animal Ecology, 79, 13-26. Supplementary files available at http://onlinelibrary.wiley.com/doi/10.1111/j.1365-2656.2009.01639.x/full Wilson, S. 2012. Australian lizards: a natural history, Collingwood, CSIRO Publishing. Wlasiuk, G. & Nachman, M. W. 2007. The genetics of adaptive coat color in gophers: Coding variation at Mc1r is not responsible for dorsal color differences. Journal of Heredity, 98, 567-574. Wolf, B. O. & Walsberg, G. E. 2000. The role of the plumage in heat transfer processes of birds. American Zoologist, 40, 575-584.

183

References Yang, J., Benyamin, B., McEvoy, B. P., Gordon, S., Henders, A. K., Nyholt, D. R., Madden, P. A., Heath, A. C., Martin, N. G., Montgomery, G. W., Goddard, M. E. & Visscher, P. M. 2010. Common SNPs explain a large proportion of the heritability for human height. Nat Genet, 42, 565-569. Yokoyama, T., Silversides, D. W., Waymire, K. G., Kwon, B. S., Takeuchi, T. & Overbeek, P. A. 1990. Conserved cysteine to serine mutation in tyrosinase is responsible for the classical albino mutation in laboratory mice. Nucleic Acids Research, 18, 7293-7298. Yurtsever, S. 2000. Inheritance of the two dorsal colour/pattern phenotypes in New Zealand populations of the polymorphic meadow spittlebug Philaenus spumarius (L.) (Homoptera : Cercopidae). Journal of the Royal Society of New Zealand, 30, 411-418. Zaiontz, C. 2014. Real statistics using Excel. Internet: http://www real-statistics. com/regression/power-regression/[Última fecha de acceso: Diciembre, 14, 2015]. Zdarsky, E., Favor, J. & Jackson, I. J. 1990. The molecular basis of brown, an old mouse mutation, and of an induced revertant to wild type. Genetics, 126, 443- 449. Zhang, G. 2015. Genomics: Bird sequencing project takes off. Nature, 522, 34-34. Zimanyi, I. & Pelleymounter, M. 2003. The role of melanocortin peptides and receptors in regulation of energy balance. Current Pharmaceutical Design, 9, 627-641. Zink, R. M. (1994). The geography of mitochondrial DNA variation, population structure, hybridization, and species limits in the fox sparrow (Passerella iliaca). Evolution, 48, 96-111. Zink, R. M. (2012). The Geography of Speciation: Case Studies from Birds. Evolution: Education and Outreach, 5, 541-546. Zink, R. & Remsen Jr, J. 1986. Evolutionary processes and patterns of geographic variation in birds. In: Johnston, R. F. (ed.) Current Ornithology. New York: Plenum Press.

184

Appendices APPENDICES

al 30

i 25

i 20

of 15

r e 10 5

Numb 0 0 1 2 3 4 Adjusted back colour score

APPENDIX I: Frequency distribution of magpie back colours within 15 km of the Seymour study site in 1976. Data from transects carried out in 1976 by Burton & Martin (1976).

APPENDIX II : Midparent – midoffspring back colour regression based on the ‘strict’ data subset. Magpies sampled from the Seymour study site, in the eastern hybrid zone. Data are for 43 broods (brood size μ=2.56, σ=1.61) comprising 155 offspring and 70 parents, from 27 territorial groups during the period 1993-2009. This ‘strict’ subset contains only kin relationships with confidence levels ≥90% in the program CERVUS, for which offspring and parents did not mismatch at any microsatellite loci.

185 Appendices

APPENDIX III: (a) mother-offspring and (b) father-offspring back colour regressions based on the ‘strict’ data subset.

Magpies sampled from the Seymour study site, in the eastern hybrid zone from territorial groups during the period 1993-2009. This ‘strict’ data subset contains only kin relationships with confidence levels ≥90% in the program CERVUS, for which offspring and parents did not mismatch at any microsatellite loci.

APPENDIX IV: Sampling locations and back colour sampling of magpie individuals screened at TYRP1, TYR, DCT and AGRP candidate genes.

Site name Sub-species BB WB Hybrid Varied TOTAL Sex Coordinates

Charters Tow ers C. tibicen terraereginae 2 2 F & M 20°04'S 146°15'E Hydeaw ay Bay C. tibicen terraereginae 1 1 U 20°05'S 148°29'E Grafton C. tibicen tibicen 3 3 U 29°40'S 152°56'E Ouyen C. tibicen tyrannica/terraereginae 1 1 U 35°04'S 142°21'E Horsham C. tibicen tyrannica/terraereginae 1 2 3 U 36°42'S 142°13'E Row sley C. tibicen tyrannica 1 1 M 37°43'S 144°24'E Phillip Island C. tibicen tyrannica 1 1 U 38°27'S 145°15'E Tasmania C. tibicen hypoleuca 2 2 U 41°26'S 147°08'E Nullabor C. tibicen telonocua 3 3 F & M 32°07'S 133°40'E Esperance C. tibicen dorsalis 3 3 F 33°51'S 121°53'E Albury C. tibicen dorsalis 3 3 U 35°00'S 117°53'E Busselton C. tibicen dorsalis 2 2 U 33°39'S 115°21'E Nth. WA C. tibicen longirostris 2 2 U 22°41'S 117°47'E & 24°18'S 116°54'E Kimberley C. tibicen eylandtensis 1 1 M 16°49'S 124°55'E Northern Territory C. tibicen eylandtensis 2 2 F & M 14°55'S 133°04'E & 19°42'S 135°49'E

TOTALS 11 9 2 8 30

186