University of Southern California Dissertations and Theses

Understanding the genetics, evolutionary history, and biomechanics of the mammalian penis bone

(USC Thesis Other)

Understanding the genetics, evolutionary history, and biomechanics of the mammalian penis bone

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Understanding the Genetics, Evolutionary History, and Biomechanics of the Mammalian Penis Bone by Nicholas G. Schultz A Dissertation presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY MOLECULAR BIOLOGY May 2019 Copyright 2019 Nicholas G. Schultz i Acknowledgements I would like to thank Matt and all Dean Lab members for their guidance and support through my doctoral work. I would like to acknowledge the impact that former Dean Lab member, friend, and scientific mentor, Brent Young had on my development as a scientist and my life. Finally, I would like to thank my dissertation committee members for their input in this process: Ian Ehrenreich, Sergey Nuzhdin, Adam Huttenlocker, and Biren Patel. ii Table of Contents Acknowledgements i Abstract 1 1 Introduction 2 1.1 Challenges in understanding the role of the baculum 2 1.2 Theories on genital evolution not related to sexual selection 2 1.2.1 The lock-and-key hypothesis 3 1.2.2 The pleiotropy hypothesis 6 1.3 Sexual selection strongly influences genital evolution 7 1.3.1 Genital evolution through sperm competition 8 1.3.2 Genital evolution through cryptic female choice 9 1.3.3 Genital evolution through sexual conflict 11 1.4 The current limitations of baculum research 12 1.5 Goals of this dissertation 13 1.6 Summary of the chapters 14 2 The Genetic Basis of Baculum Size and Shape Variation in Mice 16 2.1 Abstract 16 2.2 Introduction 17 2.3 Materials and Methods 20 2.3.1 Specimens 20 2.3.2 Micro-CT Scanning 22 2.3.3 3D transformation 22 2.3.4 Definining semi-landmarks 23 2.3.5 Size 25 2.3.6 Shape 25 2.3.7 Repeatability 26 2.3.8 Environmental input 27 2.3.9 Heritability 27 2.3.10 Mapping Quantitative Trait Loci 27 2.3.11 RNA sequencing 29 2.3.12 Bioinformatics 30 2.4 Results 31 2.4.1 BXD RILs, size 31 2.4.2 BXD RILs, Shape 32 2.4.3 BXD RILs, size and shape correlated 33 2.4.4 BXD RILs, RNA-seq and bioinformatics 34 2.4.5 LGxSM AILs, Size and Shape 34 2.4.6 Testing prediction that stem from rapid divergence 36 2.4.7 Rapid genital divergence versus conserved genital development pathways 38 2.4.8 Candidate genes 39 2.5 Conclusions 40 3 The baculum was gained and lost multiple times during mammalian evolution 43 3.1 Abstract 43 3.2 Introduction 44 3.3 Materials and Methods 47 3.3.1 Phylogeny 48 3.3.2 Baculum status 50 3.3.3 Testing evolutionary hypotheses 51 3.3.4 Bootstrapping 52 3.3.5 Correlating baculum presence to rates of diversification 54 3.4 Results 54 3.4.1 The baculum was gained and lost multiple times 54 3.4.2 Boostrapping 59 iii 3.4.3 Groups with a baculum diversify rapidly 59 3.5 Discussion 60 3.5.1 Multiple derivations 60 3.5.2 Sexual selection 62 3.5.3 Developmental biology 63 3.5.4 Functional biology 64 3.6 Conclusion 65 4 Understanding Mouse Baculum Biomechanics through an Investigation of Bone Histology 66 4.1 Abstract 66 4.2 Introduction 67 4.3 Materials and Methods 71 4.3.1 Experimental Design 71 4.3.2 Micro-CT Scanning of bacula 73 4.3.3 3D-Transformation and Defining semi-landmarks 73 4.3.4 Quantifying shape variation 74 4.3.5 Histology Preparation 75 4.3.6 Polarized Light Analysis 77 4.3.7 Calcein Quantification 78 4.4 Results 78 4.4.1 Histological features vary within a single slice and between regions of the bone 78 4.4.2 Shape did not vary by mating status, but mated bacula are larger than unmated bacula 82 4.4.3 Baculum growth did not differ between mated and unmated males 82 4.4.4 Collagen fiber orientation varies by bone region in both experiments, and between cortices in the short-term experiment 83 4.5 Discussion 84 4.5.1 Experimental Design 84 4.5.2 Baculum Size and Mating 85 4.5.3 Trends in Baculum Histology 86 4.6 Conclusion 91 5 Concluding remarks 96 5.1 Impact of this thesis 96 5.2 Future directions 97 References 103 1 Abstract For centuries, biologists have been captivated by the rapid evolution and extreme morphological diversity of external genitalia. One of the most mysterious and widely studied genital characters in mammals is the penis bone, or baculum. Numerous studies have sought to explain the function of the baculum, yet to date very little is known about its role in reproduction. Unveiling the baculums’ influence on male reproductive fitness requires investigations into the genetics, evolutionary history, and microstructure of this bone. In this thesis, I investigate baculum evolution from all three of the above perspectives. In chapter two, I use an established genetic resource in mice to uncover loci influencing shape and size variation in bacula. In chapter three, I show that the baculum has likely evolved multiple independent times using a phylogenetic approach. Finally, in chapter four, I characterize histological features of the mouse baculum, providing insight into the biomechanical function of the bone. 2 1 Introduction 1.1 Challenges in understanding the role of the baculum The rapid evolution of external genitalia is a well-documented biological phenomenon, but the ever-changing nature of these structures makes them incredibly difficult to study, and the baculum is no exception. Numerous studies have attempted to describe the function of the baculum (André et al. 2018; Patterson and Thaeler 1982; D. E. Tasikas et al. 2009a), yet no single function can be successfully applied across taxonomic groups (Brassey et al. 2018). In this chapter, I provide a brief explanation of common theories used to explain the rapid evolution of genitalia, how they have been applied to baculum research, and how I approached gaining a better understanding of this enigmatic bone. 1.2 Theories on genital evolution not related to sexual selection The primary role of male genitalia is to transfer sperm to females (Holwell et al. 2010), yet selection based on this function alone does not adequately explain the incredible morphological diversity observed in genitalia across the animal kingdom (D. J. Hosken and Stockley 2004). For instance, marsupials and snakes commonly have two penises (Birkhead 2000), certain slugs have penises longer than their entire bodies (Baur 1998), and some squid species have no penis at all, but instead transfer sperm to females using a modified arm called a hectocotylus (Hanlon and Messenger 2018). While these examples are striking, divergence in male genitalia is not limited to the size or number of copulatory organs, as male genitalia also vary in the types of structures that extend from genitalia itself. Hardened projections like keratinized penile spines are not only common in species like primates, bats, and rodents, but the number, size, and morphology of spines also varies widely between species (M. J. Anderson 2000; AF Dixson 1987b; Orr and Brennan 2016; Rocha-Barbosa et al. 2013). Researchers have 3 developed many explanations for this divergence (Langerhans et al. 2016; A. M. Shapiro and Porter 1989), and recent evidence supports the idea that sexual selection largely drives genital evolution (D. J. Hosken and Stockley 2004; Langerhans et al. 2016; Simmons 2014). However, this may not be the case across all taxa, and there are a number of alternative hypotheses that warrant further explanation. 1.2.1 The lock-and-key hypothesis The lock-and-key hypothesis suggests that genitalia rapidly diverge between species to avoid hybridization (Dufour 1844; William G Eberhard 1985). Accordingly, conspecific male and female genitalia co-evolve so that only males of the same species have the correct “key” to the accompanying “lock” of female genitalia. Only when this match is made can mating result in proper insemination and viable offspring (William G Eberhard 1985; Masly 2012). A critical assumption of this hypothesis is that hybridization imposes a large enough fitness cost to influence the divergence of genitalia between species (A. M. Shapiro and Porter 1989). Sexual reproduction comes with inherent costs, some of which are incurred by provisioning resources for offspring, primarily through eggs or placentas (Rosenheim 1999; Tunster et al. 2013), while other costs are associated with heightened immune responses related to genital damage (Crudgington and Siva-Jothy 2000; Eady et al. 2006) or increased infection risk (Zuk and Stoehr 2002). Incurring the costs of physical copulation can potentially reduce the lifelong reproductive output of an individual (Daly 1978; Pellati et al. 2008). Ultimately, these costs can be balanced if mating produces viable offspring, but if hybrid offspring are less fit, there should be strong selection for mates to avoid incompatible encounters (Spencer et al. 1986). In fact, most examples of reproductive isolation occur through pre-copulatory mechanisms (Pfennig and Simovich 2002), like species-specific shell patterns in snails (Johnson 4 1982), or specific pheromone profiles in moths (Löfstedt et al. 1991; Tumlinson et al. 1974), which minimize the probability of physical copulation and the potential energetic consequences. Not only must there be a cost to hybridizing, but male and female genitalia must be shown to co-evolve, which is an easier task in species with keratinized, or hardened genital structures (William G Eberhard 1992; Sota and Kubota 1998). These hardened structures allow researchers to more accurately identify where male and female genital contact may lead to incompatibilities. A clear example of this comes from a study of two beetle species with a narrow hybridization zone in Japan (Sota and Kubota 1998). After heterospecific matings, females from both species often die due to ruptured genitalia, and male genitalia typically break as well. This is a classic example of the lock-and-key hypothesis through a mechanical mismatch; there is a clear reduction in hybrid fitness, male and female genitalia co-evolve, and there are no observable pre-copulatory isolating mechanisms (Sota and Kubota 1998), indicating that the species-specific shape of genitalia is causing isolation. Reproductive isolation according to the lock-and-key hypothesis can develop through mechanical and/or sensory mismatches (William G Eberhard 1985; Masly 2012). In instances of mechanical dissimilarities, like the example in the previous paragraph, fitness costs are generally related to improper genital coupling leading to insufficient sperm transfer (William G Eberhard 1993; Holwell et al. 2010; Otronen and Siva-Jothy 1991) or damaged reproductive tracts (Holwell et al. 2010). Divergence in female genital sensory patterns may also lead to reproductive isolation (Masly 2012), since stimulation from males can influence phenotypes like sperm storage (Otronen and Siva-Jothy 1991). A recent study in damselflies illustrated that mechanical incompatibilities between heterospecific mating pairs did not entirely explain the observed patterns of reproductive isolation (Barnard et al. 2017). Specifically, they looked at the 5 interaction between male genital structures, called cerci, and female mesostigmal plates. The morphology of cerci and mesostigmal plates co-evolve, and mesostigmal plates have species- specific distributions of sensory receptors (Robertson and Paterson 1982). A mechanical mismatch occurs between these structures when heterospecific males and females cannot physically create a mating tandem. If genital coupling does occur and females still reject males, that may be the consequence of insufficient stimulation by the heterospecific male (Barnard et al. 2017) and thus support for sensory isolation. Barnard (Barnard et al.) showed that in crosses with partial mechanical incompatibilities, females still rejected heterospecific males frequently signifying that divergence in genital stimulatory capabilities influenced reproductive isolation between these damselfly species (Barnard et al. 2017). Ultimately, this form of reproductive isolation is more difficult to identify, in part because it requires mapping the species-specific distribution or sensory receptors (Barnard and Masly 2018; Masly 2012), which is often more complicated than identifying structural variance in genitalia (Barnard et al. 2017). Undoubtedly, the cost of hybridization leading to genital incompatibilities is a driving force in some taxa (C. M. Anderson and Langerhans 2015), but the lofty requirements to support the lock-and-key hypothesis makes it potentially more difficult to test than other hypotheses (Brennan and Prum 2015; A. M. Shapiro and Porter 1989). This is especially true in species with soft tissue genitals, like the hydrostatic genital systems in mammals. The increased flexibility of soft-tissue genitalia may enable mammals to tolerate divergent genital shapes more than organisms with hardened genitalia, as the added flexibility minimizes the risk of severe damage (William G Eberhard 1985). Additionally, the invagination of female mammal genitalia makes them more difficult to study (William G Eberhard 1985) since male-female tissue interactions in copula cannot be as easily visualized. Taken together, these factors limit the ability to detect the 6 influence of lock-and-key in mammals, and ultimately limit the potential influence on baculum evolution as well. 1.2.2 The pleiotropy hypothesis One of the oldest explanations for the rapid evolution of male genitalia is the theory of pleiotropy, which states that changes in genital morphology are the product of selection on related traits, not genitalia itself (Mayr et al. 1963). These morphological changes are tolerated as long as sperm can be successfully transferred to females (William G Eberhard 1985; Mayr et al. 1963). If genital morphology has no evolutionary consequence, then genitalia should reflect the phenotypic range of the connected trait, which is under direct selection (Göran Arnqvist 1997). However, this explanation does not account for the seemingly unmatched rapid evolution of external genitalia (William G Eberhard 1985). If genital shape evolves rapidly, then so should the morphology of the linked trait, but that does not seem to be the case (D. J. Hosken and Stockley 2004). Our understanding of genetic evolution and how it affects linked traits has dramatically changed since this theory was first developed. Over recent years, the connectivity of genetic networks, and how they operate on seemingly unrelated structures has come to the forefront of research (Molodtsova et al. 2014). For instance, many of the genes shown to influence limb development and morphology also affect genitalia (Infante et al. 2015) (Sears et al. 2015; József Zákány et al. 1997). These shared networks often use highly conserved genes, which cannot tolerate structural changes, but instead evolve cis-regulatory elements to modify tissue-specific expression of the same gene (Indjeian et al. 2016; Prud'homme et al. 2007). Evolution through cis-regulatory change is less vulnerable to the effects of pleiotropy (Prud'homme et al. 2007), thus potentially mitigating the influence pleiotropy has on genital evolution. I will discuss the 7 potential role of cis-regulatory evolution influencing bacula in greater depth in the final chapter of this dissertation. The pleiotropy hypothesis also underestimates the dynamic nature of sexual reproduction. While transferring sperm close to the site of fertilization is critical, it does not ensure paternity success (Simmons 2014). Females of most species mate with multiple males within a reproductive cycle, establishing competition between ejaculates (Renée C Firman et al. 2017). Males can improve their chances of fertilizing eggs through various mechanisms, including changes in genital morphology. For instance, male mosquitofish with longer genitalia fertilize more eggs than other males (Head et al. 2016). If males achieve those feats through changes in genital morphologies that suggests that those morphological features would be under direct selection (William G Eberhard 1985), thus discounting the influence of pleiotropy. 1.3 Sexual selection strongly influences genital evolution Sexual reproduction is not as harmonious as we commonly think. Males and females often have different reproductive goals, and males of most species compete with other males for fertilization success (Göran Arnqvist and Rowe 2013; D. J. Hosken and Stockley 2004). Certain individuals have sexual traits that enable them to produce more offspring than others. This predictably leads to frequency changes of certain phenotypes, and has been broadly defined as sexual selection (M. B. Andersson 1994). Although competition for mating success is observed prior to mating in pre-copulatory selection, male-female interactions and interactions between rival ejaculates after mating often have a stronger influence on genital evolution (Göran Arnqvist 1998; Goran Arnqvist and Danielsson 1999; William G Eberhard 1985; Langerhans et al. 2016). For this reason, I focus on common post-copulatory mechanisms of sexual selection in this section. Specifically, I will discuss the influence of sperm competition, cryptic female choice, 8 and sexual conflict (Göran Arnqvist 1997; D. J. Hosken and Stockley 2004). These explanations are not necessarily mutually exclusive, but all of them are relevant to baculum research. 1.3.1 Genital evolution through sperm competition Polyandry, when a female mates with multiple males during a reproductive cycle, is an incredibly common reproductive phenomenon (G. Parker 1984). This mode of reproduction introduces a scenario where sperm from numerous males co-exist in a females’ reproductive tract simultaneously. Competition between ejaculates ensues, and certain males will achieve higher paternity rates depending on sperm, ejaculate and genital phenotypes (M. B. Andersson 1994). Experiments comparing genital morphologies between monogamous and polyandrous mating systems have illustrated that genitalia of males from polyandrous systems diverge more rapidly (Göran Arnqvist 1998). Some of the common mechanisms for gaining higher paternity in sperm competition are transferring more of your own sperm and removing previously deposited ejaculates. Males can transfer more of their own sperm by maintaining copulation for longer periods of time, and many taxa have evolved genital ‘grasping’ structures to prolong copulatory duration (Kentaro M Tanaka et al. 2018; Tong et al. 2018). Male garter snakes have large spines on the base of their genitalia, and when these spines are removed, copulatory duration shortens and males transfer smaller copulatory plugs to females (Friesen et al. 2014), reducing the probability of successful fertilization. Males may also evolve genitalia equipped for removing rival sperm. For example, in an experiment with flour beetles, when a female is sequentially mated to two different males, the second male to mate fertilized the vast majority of eggs (Haubruge et al. 1999). The second male accomplished this by removing previously ejaculated sperm through spines on his genitalia (Haubruge et al. 1999). Genital characters, which remove rival sperm, like spines, are observed 9 across animals like rodents (Dewsbury 1981), damselflies (Waage 1979), and primates (Parga 2003). The prevalence of polyandry implies that sperm competition certainly influences genital evolution. Yet, these examples only consider interactions between rival males, and male-female genital interactions can also bias sperm usage. 1.3.2 Genital evolution through cryptic female choice Males seldom ejaculate directly on eggs (William G Eberhard 1985), leaving sperm to navigate some portion of the female reproductive tract to achieve successful fertilization. Cryptic female choice describes the ability of females to bias sperm usage based on input received from the male during the mating process (Renée C Firman et al. 2017). Female mediated control of fertilization can then drive the evolution of male phenotypes, since cryptic female choice is often based on the perceptions of male traits (Gasparini and Evans 2018). For example, females of many species will non-randomly eject sperm after mating (Renée C Firman et al. 2017) based on factors like genital stimulation (Davies 1983; Van der Schoot et al. 1992) and the males most capable of properly stimulating females will achieve a higher rate of paternity (Göran Arnqvist 1997; William G Eberhard 1985). This preference for certain male genital traits with stimulatory capabilities, leads to the rapid evolution and divergence of male genitalia (William G Eberhard 1985). An example of the stimulatory features of male genitalia was recently uncovered in bushcrickets (Wulff et al. 2017). Using live X-ray imaging, researchers were able to see the “tapping” motion of male genitalia during mating. These tapping structures called titilators, were moving along the female genital fold, potentially stimulating females to expose their genital opening during mating and allow for more sperm transfer (Wulff et al. 2017). This is where the lock-and-key hypothesis and sexual selection deviate; in the latter, females are discriminating 10 between conspecific males of variable quality, not males from different species (Brennan and Prum 2015; D. J. Hosken and Stockley 2004). A male’s ability to develop and maintain elaborate genital features may be a signal of that males’ genetic quality (William G Eberhard 1985; Renée C Firman et al. 2017). If his overall genetic quality is heritable, it may benefit females to mate with males with the most exaggerated genitalia, as their offspring will be of higher quality (Gustafsson and Qvarnström 2005). This is referred to as the ‘good genes’ hypothesis, and it has been used to explain how male genitalia evolve so rapidly (William G Eberhard 1985). A related hypothesis, commonly known as the ‘sexy sons’ hypothesis, states that females prefer to mate with males who possess a strong ability to manipulate fertilization with their genitalia (Cordero and Eberhard 2003). This genital feature can then be passed on to her sons, and they will be more successful breeders as well (Weatherhead and Robertson 1979). Furthermore, a Fisherian runaway process may occur if female preference for exaggerated genitalia is also genetically linked to loci causing the exaggeration of male features. Ultimately, this process will reinforce female preference and it may lead to male genitalia developing increased complexity (William G Eberhard 1985). This would be adaptive for females initially, since their sons would have a reproductive advantage over other males. Yet, over time Fisherian Runaway selection pushes the male trait to be so exaggerated that it becomes cumbersome to produce which then makes female preference for that trait costly, making this form of selection unstable (M. Andersson and Iwasa 1996; Göran Arnqvist 1997; Renée C Firman et al. 2017). Cryptic female choice can also be non-adaptive for females if males are simply hijacking pre-existing female sensory systems to bias sperm usage (Cordoba-Aguilar 1999; Peretti and Eberhard 2010). This is especially true if male genital characteristics are not an indication of high genetic quality (William G Eberhard 1985), and 11 males are using these mechanisms to bias paternity in their favor. A classic example of sensory exploitation has been observed in Túngara frogs (Ryan et al. 1990), where males have adapted their calling pattern to the pre-existing auditory frequency preference of females (Ryan and Rand 1990). Males who produce a mating call, which matches this pre-existing preference, will achieve higher paternity success, and there is no inherent benefit for females (Ryan and Rand 1990). The separation between male and female interest in reproduction creates inherent conflict in mating systems. 1.3.3 Genital evolution through sexual conflict Male and female reproductive interests are often unaligned. It is more beneficial for a male to fertilize all of the eggs available with every female he mates, while females gain a larger benefit from discriminating between males (Cordero and Eberhard 2003). These interactions have been hypothesized to lead to an evolutionary arms race, where males and females continuously adapt and counter-adapt to gain a momentary reproductive advantage (D. J. Hosken et al. 2018; Langerhans et al. 2016). A classic example of this conflict manifests in the seminal fluid of Drosophila melanogaster. In this species, males transfer a particular protein in their ejaculate called sex peptide. Sex peptide decreases the remating rate of females, reducing their overall fitness, while providing the male who released the sex peptide higher paternity success (Wigby and Chapman 2005). This is undoubtedly conflict between the reproductive interests of the male and female, and this conflict drives divergent morphologies in genitalia as well. Sexual conflict and genital evolution can also be observed in the traumatic insemination common to bedbugs (D. J. Hosken et al. 2018). Male bedbug genitalia have a sharp distal tip, which they use to puncture the female abdomen and ejaculate near her ovaries (Stutt and Siva-Jothy 2001). Such traumatic 12 inseminations provide the male higher paternity success, while coming at a cost to females (Stutt and Siva-Jothy 2001). Female bedbugs have counter adapted to traumatic insemination through anatomic features which may function to kill some of the sperm ejaculated before they can reach their ovaries (Carayon 1966). Less severe examples of sexual conflict have be observed in rats and primates, where mating events reduce the likelihood of female remating due to genital damage (Van der Schoot et al. 1992), possibly caused by penile spines (P Stockley 2002). Testing these hypotheses requires an understanding of male variation in genital morphology, male-female interactions, and the evolutionary history of shape changes (Langerhans et al. 2016). Even then, it is difficult to disentangle which sexual selection hypotheses are operating, or if they are solely responsible for the trends observed in genital change (Fricke et al. 2009; D. J. Hosken and Stockley 2004). Additionally, the same genital feature can serve distinct functions in different taxonomic groups (Orr and Brennan 2016), especially since mating systems evolve so rapidly. 1.4 The current limitations of baculum research Although bacula are a commonly studied mammalian genital character in sexual selection research, very little is known about the function of these bones. Functional hypotheses have been proposed (Larivière and Ferguson 2002; Miller and Burton 2001; Patterson and Thaeler 1982), but most of those studies have focused on baculum variation in wild populations where assessing reproductive fitness and heritability of baculum morphology are much more difficult. In order to empirically test if baculum evolution is influenced by sexual selection, research must investigate if baculum shape and size variation is associated with male reproductive success, and uncover if those shapes are heritable. Recent papers have shown a connection between mouse baculum shape and reproductive fitness (André et al. 2018; Simmons 13 and Firman 2014; Paula Stockley et al. 2013), specifically highlighting baculum width as a potential target of sexual selection. However, these results require cautious interpretation as males in a competitive mating experiment may produce more testosterone, which has been shown to not only influence ejaculate properties (Alan F Dixson and Anderson 2004), but also bone development (Vanderschueren et al. 2004). Regardless, these experiments still provide a strong foundation to further examine baculum evolution in response to sexual selection, but the use of outbred mice still limits our ability to uncover genomic regions influencing this phenotype (Solberg Woods 2013). Additionally, while shape changes influence bone function, looking at bone histology may uncover microstructural changes which further our understanding of baculum function (Bromage et al. 2003; McFarlin et al. 2016; Sun et al. 2018). 1.5 Goals of this dissertation The goal of this dissertation is to gain a better understanding of the factors influencing baculum evolution. As specified above, distinguishing between those factors requires knowledge of: 1) What, if any, genetic loci influence baculum shape and size variation? 2) Can inconsistencies in assigning baculum function be explained by the evolutionary history of the bone? , and 3) How does baculum shape respond to mating events? Because of previous mouse baculum research and the strength of mouse genomics, I first approached these questions by using the house mouse Mus musculus as a model species. I then investigated the evolution of bacula through a phylogenetic approach. Lastly, based on the results of the phylogenetic study, I focused again on mouse baculum evolution by examining shape and histological changes associated with mating. 14 1.6 Summary of the chapters In chapter two, I utilize an established inbred mouse genetic panel called the BXD recombinant inbred lines (RILs). This genetic panel has been successfully used to provide glimpses into genomic regions influencing variation of many phenotypes, and here I applied it to understand baculum morphological variation. Using geometric morphometrics, and the BXD RILs, I uncovered three quantitative trait loci (QTL) influencing baculum size and shape variation. In chapter three, I expanded the study to all mammalian species in order to better understand the evolutionary history of the baculum. Research to this point has failed to assign a singular function to these bones (Paula Stockley 2012), which may be the result of bacula evolving multiple independent times. If that is the case, then a universal function should not apply to all bacula since they would not be homologous structures. Researchers have looked at baculum evolution within mammal clades like carnivores (Larivière and Ferguson 2002) or primates (A. F. Dixson 1987a), but the phylogenetic distribution of the baculum across all mammal species has never been rigorously tested. Through a robust literature search for baculum presence combined with a phylogenetic assessment, I determined that the baculum likely evolved multiple independent times and was lost in some mammalian lineages. In chapter four, I attempted to glean information on the functionality of the mouse baculum through controlled mating experiments. Previous experiments have focused on how the shape of the mouse baculum changes related to mating success (André et al. 2018; Simmons and Firman 2014; Paula Stockley et al. 2013), but bone adaptations do not solely manifest through external shape changes. Changes in the internal structure of bone, like collagen fiber orientation and bone deposition rate, can also reflect adaptations to the external environment (Goldman et al. 2003; Martin et al. 1996). These histological features have never before been studied in a 15 baculum mating experiment. By assessing both, shape changes and changes in histological features, I identified potentially functionally important regions of bacula. In chapter five, I review the impact of this work and how it may be applied to the future of baculum, sexual selection, and bone evolution broadly. 16 2 The Genetic Basis of Baculum Size and Shape Variation in Mice This chapter appears principally as published in 2016 in G3: Genes, Genomes, and Genetics May 1, 2016 vol. 6 no. 5 1141-1151. 2.1 Abstract The rapid divergence of male genitalia is a pre-eminent evolutionary pattern. This rapid divergence is especially striking in the baculum, a bone that occurs in the penis of many mammalian species. Closely related species often display diverse baculum morphology where no other morphological differences can be discerned. While this fundamental pattern of evolution has been appreciated at the level of gross morphology, nearly nothing is known of the genetic basis of size and shape divergence. Quantifying the genetic basis of baculum size and shape variation has been difficult because these structures generally lack obvious landmarks, so comparing them in three dimensions is not straightforward. Here we develop a novel morphometric approach to quantify size and shape variation from three-dimensional micro-CT scans taken from 369 bacula, representing 75 distinct strains of the BXD family of mice. We identify two quantitative trait loci (QTL) that explain ~50% of the variance in baculum size, and a QTL that explains more than 20% of the variance in shape. Together, our study demonstrates that baculum morphology may diverge relatively easily, with mutations at a few loci of large effect that independently modulate size and shape. Based on a combination of bioinformatic investigations and new data on RNA expression, we prioritized these QTL to 16 candidate genes which have hypothesized roles in bone morphogenesis, and may enable future genetic manipulation of baculum morphology. 17 2.2 Introduction The rapid divergence of male genitalia is an almost universal evolutionary pattern among sexually reproducing organisms (William G Eberhard 1985; W. Eberhard 2009a; D. J. Hosken and Stockley 2004). Rapid evolution extends to the baculum, a bone in the penis of many mammalian species. Even a cursory examination of over 200 illustrations by Burt (1960a) reveals a bewildering array of baculum diversity, spanning diverse scales of size, shape, amorphousness, symmetry, and even the number of bones and patterns of branching. Mammalogists have long recognized the utility of the baculum to identify species that are otherwise morphologically indistinguishable (i.e., Herdina et al. 2014; Thomas 1915), even positing that baculum morphology reinforces reproductive isolation between species (Patterson and Thaeler 1982), but see (Good et al. 2003). The selective forces that drive evolutionary novelty of the baculum are likely to include a combination of female choice, male-male competition, and conflict between male and female reproductive interests (Göran Arnqvist 1998; G. Arnqvist and Rowe 2005; Breseño and Eberhard 2009; W. Eberhard 1996; W. G. Eberhard 2000; William G Eberhard 2009b; D. J. Hosken and Stockley 2004; C. M. House and Simmons 2003; C. House and Simmons 2005; Long and Frank 1968; Patterson and Thaeler 1982; Patterson 1983; Perry and Rowe ; S. A. Ramm 2007; Rönn et al. 2007; Rowe and Arnqvist 2012; Schulte-Hostedde et al. 2011; D. Tasikas et al. 2009b) . The enormous interspecific divergence in baculum morphology coupled with relatively low levels of intraspecific variation (e.g., Klaczko et al. 2015; S. Ramm et al. 2010) suggests that male genitalia are the targets of recurrent adaptive evolution. 18 Several studies have linked baculum characteristics with male reproductive success. In semi-natural enclosures of multiple male and female house mice, males with a wider baculum sired both more and larger litters (Paula Stockley et al. 2013). Similarly, in wild populations and experimental populations subjected to higher levels of sperm competition, males developed bacula that were significantly wider than males from low sperm competition populations (Simmons and Firman 2014). Dominant males tend to develop wider bacula (Lemaître et al. 2012), suggesting additional links between baculum morphology and presumably adaptive behavioral phenotypes. Since sperm competition is a regular feature of mouse mating ecology (M. D. Dean et al. 2006; R. C. Firman and Simmons 2008), baculum size and shape are probably important components of fitness in natural populations. In mice the baculum resides in the distal portion of the glans penis which enters the vagina during copulation (Rodriguez et al. 2011). Intromission may be an important aspect of “copulatory courtship” which influences fertilization success beyond delivery of ejaculate (Adler and Toner 1986; Diamond 1970; Matthews Jr and Adler 1978; Toner and Adler 1986; Toner et al. 1987) and it is possible that the baculum plays a role in these processes. The puzzling diversity of baculum morphology, its link to male reproductive fitness, and our lack of understanding of the specific processes that influence baculum evolution demand more attention. A genetic dissection of ultimate sources of variation in baculum morphology has been hindered by two main obstacles. First, bacula lack true landmarks such as sutures, foramina, or processes. As a result, most studies summarize baculum complexity with simple length, width, and/or weight measurements (Long and Frank 1968). Some studies have captured baculum outlines using more modern computational techniques (Simmons and Firman 2014), but ideally, baculum size and shape would be captured in three dimensions. Here we overcome this 19 challenge using tools in computational geometry and geometric morphometrics to measure bacula without reliance on true landmarks, which should be applicable to many different biological structures. Second, although variation in baculum morphology is heritable (Simmons and Firman 2014), no study to date has mapped the genetic basis of variation in size or shape. Both genetic causes and molecular mechanisms underlying its rapid evolution are uncharted. The rapid evolutionary divergence discussed above predicts a relatively simple genetic architecture in which mutations in a few key loci lead to large changes in morphology. Furthermore, we might predict that size and shape would be modulated by different loci, so that divergence in one aspect is uncoupled from divergence in the other. Here we unite our newly developed morphometric methods with the power of mouse genetics in a quantitative genetics framework, using two genetic models: a family of recombinant inbred lines (RILs) known as the BXDs (Peirce et al. 2004; B. A. Taylor et al. 1973; Benjamin A Taylor et al. 1999), and a family of advanced intercross lines (AILs) known as the LGxSM (reviewed in Nikolskiy et al. 2015). Our study makes several advances towards understanding the genetic basis of baculum size and shape. We demonstrate that variation in baculum size and shape is heritable, and is controlled by a small number of quantitative trait loci (QTL) with comparatively large effects. We identified different QTL affecting size vs. shape, that when combined with new data on gene expression and bioinformatic filtering, highlight several compelling candidate genes. Future molecular analyses should eventually lead to a better understanding of genetic mechanisms of recurrent adaptive morphological evolution. 20 2.3 Materials and Methods 2.3.1 Specimens All protocols and personnel were approved by USC’s Institute for Animal Care and Use Committee (protocol #11394). Our study was based on two powerful mouse genetic models, the BXD RILs and the LGxSM AILs. Most mice were already euthanized and frozen as part of unrelated research programs; a few were raised in-house or ordered through common vendors, then euthanized via carbon dioxide exposure followed by cervical dislocation. The BXD RILs began in the 1970s with approximately 30 RILs (B. A. Taylor et al. 1973; Benjamin A Taylor et al. 1999) and were expanded to ~150 lines (Peirce et al. 2004, Williams unpublished; Benjamin A Taylor et al. 1999). All strains have been genotyped at over 3,000 informative markers (www.genenetwork.org). Each BXD RIL began as a cross between two classical inbred strains, C57BL/6J (B6) and DBA/2J (D2). BXD1 through BXD42 were then maintained via brother-sister mating. BXD43 and higher were maintained through non-sib matings until generation 8-12 to accumulate larger numbers of recombinations and then inbred, making them an advanced intercross-rederived RIL family (Peirce et al. 2004). To create homozygous strains that on average have 50% B6 and 50% D2 genome, but with a different 50% segregating across different BXD RILs dependent upon the randomness of recombination during inbreeding. Individuals from the same BXD RIL are essentially genetically identical, and can be considered biological replicates of the same genotype. We sampled males that were 60-200 days old since the baculum is fully developed at this age (Glucksmann et al. 1976). The genomes of B6 and D2 have been nearly completely sequenced (Keane et al. 2011; X. Wang et al. In Press; Waterston et al. 2002), providing power to follow-up investigations of QTL. 21 The LGxSM AILs began from crosses between the LG/J and SM/J strains, originally selected over many generations for large and small body size, respectively (Goodale 1938; MacArthur 1944). LGxSM AILs were not intentionally inbred, but rather maintained through random breeding of unrelated individuals. Approximately 100 families are created each generation by pairing males and females that are not full siblings. Each generation accumulates more recombination events that break up the genomes of the two parental strains while avoiding inbreeding. Published studies of the LGxSM AILs exist at the F2, F8, and F34 generations (Cheng et al. 2010; Norgard et al. 2011; C. Parker et al. 2011; Clarissa C Parker et al. 2012; Clarissa C. Parker et al. 2014). As part of unrelated research, several hundred individuals of the F43 and F44 generations were genotyped at thousands of markers (Cheverud, unpublished). These F43 and F44 mice were sampled here. Unfortunately, many of the frozen carcasses were missing the penis as a result of prior necropsy, and only 144 males could be sampled. The parental LG/J and SM/J genomes have been sequenced (Nikolskiy et al. 2015). For all specimens, the glans penis was cut proximal to the baculum, then placed in distilled water at room temperature for approximately one week, at which point tissue was easily teased away using forceps and pressure of liquid flow from a squeeze bottle of 70% ethanol. Bacula were soaked in 200 µL of 1% ammonia solution for a maximum of 2 hours to clean any remaining tissue and remove oils, then dried and packed into thin layers of Aquafoam (Kokomo, IN) for micro-CT scanning. There is a distinct mass of fibrocartilage distal to the bony baculum; our study did not include this structure. For the remainder of the manuscript, “baculum” refers to the single bone in the baculum, excluding the distal fibrocartilage (Fig. 2.1). 22 2.3.2 Micro-CT Scanning Aquafoam slices containing bacula were stacked in a 35mm cylindrical micro-CT sample holder. Micro-CT scans were acquired using a uCT50 scanner (Scanco Medical AG, Bruttisellen, Switzerland) at the USC imaging core facility, under the following settings: 90 kVp, 155 uA, 0.5 mm Al filter, 750 projections per 180 degrees (360 degree coverage), exposure time of 500ms and voxel size of 15.5 µm. 2.3.3 3D transformation Customized scripts were written in Python (www.python.org) and R (R Core Team 2014) to process microCT images, segment individual bacula, perform transformations, and call semi- landmarks; all scripts are available on G3 website. Each baculum was segmented out of the z- stack of micro-CT images and converted to a 3D point cloud of x-y-z points. Each pixel in the point cloud represents bone, and the images include internal structures, not just the surface. Point clouds were transformed in three main steps borrowing methodology from Dines et al. (2014) and are graphically illustrated in Fig. 2.1. First, the two points furthest apart in the point cloud were used to initially define a distal-proximal axis, which we consider a z-axis. Second, the 10% most proximal and the 10% most distal points were sliced out separately, and the centroids of their convex hulls calculated. The point cloud was then transformed such that the proximal centroid became 0, 0, 0 and the distal point cloud became 0, 0, z (where z was some positive value). This second transformation effectively controlled for slight variation associated with defining points in the first step. Third, we sliced out points that fell 15.00-15.25% along the length of the new z-axis and computed its minimum bounding rectangle (MBR) (Butterworth 2013). The entire point cloud was then re-transformed so that the long end of this MBR ran 23 parallel to a new x-axis, and the short end parallel to a new y-axis. Slight curvatures in the bone, especially towards the distal end were exploited to ensure the dorsal-ventral axis was correct. All bone transformations were visually confirmed using the rgl library in R with customized R scripts. 2.3.4 Defining semi-landmarks The baculum lacks any true landmarks, so we mathematically defined 802 points along each bone. These points can be thought of as “semi-landmarks” (F. Bookstein 1991; F. J. Bookstein 1997; Mitteroecker and Gunz 2009) since they correspond to regionally homologous regions across all specimens. Fifty point slices, each with thickness 0.25% the length of the z- axis, were evenly spaced along the anterior-posterior axis. One such slice is shown in detail in Fig. 2.1E. Each slice was divided along the x-axis by 7 lines running parallel to the y-axis. Points within 4% of the width of the slice to each line were projected onto that line, then the ventral most and dorsal most points defined as semi-landmarks, specifically labeled according to slice and position. The leftmost and rightmost points of each slice were also defined as semi- landmarks, for a total of 16 semi-landmarks per slice (Fig. 2.1E), and a total of 800 semi- landmarks across the 50 slices. The single anterior-most and single posterior-most points of the baculum were added, bringing the total to 802 (Fig. 2.1A-D). 24 Figure 2.1 Semilandmark Morphometrics Visual representation of our morphometric pipeline for defining semi-landmarks for quantification of baculum size and shape. Detailed methodology can be found at [Dryad]. Dorsal (A, B) and lateral (C, D) views a typical D2 (A, C) or B6 (B, D) individual (specimens #DBA_1 and #C57_1, respectively). The gray background of each bone represents approximately 175,000 x-y-z points segmented from micro-CT scans. Moving from proximal (left sides of A-D) to distal (right sides of A-D), we sampled 50 slices. One slice is shown in more detail (E), which is looking down the center of the bone, displaying an empty internal medullary cavity. Exactly 7 points were defined on the ventral and dorsal surfaces, as well as the leftmost and rightmost, totaling 16 semi-landmarks per slice, indicated by spheres. The colors of the spheres indicate their contribution to shape differences (LD1) between D2 and B6. Horizontal black axes indicate the z-axis (A, B, C, D) or the x-axis (E). Vertical black axes indicate the x-axis (A, B), or the y-axis (C, D, E). 25 2.3.5 Size Size was quantified as centroid size, the square root of the sum of squared distances of these 802 semi-landmarks from their centroid. Counting the number of pixels per scan, or even the density of pixels (number of pixels divided by centroid size) yielded identical results, but we present those based on centroid size for simplicity. 2.3.6 Shape We quantified shape difference between all possible pairs of bacula in a Generalized Procrustes framework, which standardized each set of 802 semi-landmarks to a common size, translated them to a common origin, then optimally rotated their coordinates to minimize their Procrustes distance (Rohlf and Bookstein 1990; Slice 2007). During Procrustes superimposition, semi-landmarks were allowed to “slide” along the bones’ surfaces using the function gpagen in the R package geomorph (D. Adams and Otárola-Castillo 2012), which improves alignment of corresponding anatomical regions lacking individual landmark homology (F. L. Bookstein 1996; F. J. Bookstein 1997; Gunz et al. 2005; Mitteroecker and Gunz 2009). In short, sliding semi- landmarks accompanies uncertainty in specific placement of landmarks. The gpagen analysis resulted in a pairwise distance matrix (in the case of the BXD RILs, a 369 x 369 matrix; in the case of the LGxSM AILs, a 144 x 144 matrix). Our goal was to define a single metric that summarized each specimen’s shape in the context of the parental strains, which we accomplished in two main steps. First, the pairwise distance matrix was converted to its Principal Components using the dudi.pca function in the R package ade4 (Chessel et al. 2004; Dray and Dufour 2007), with variables centered to have zero mean and scaled to have unit variance. In the case of the BXD RILs, each specimen was weighted by the inverse of the per- 26 strain sample size to account for biological replicates within a genetically homozygous strain. Second, using only specimens of the parental strains (in the case of BXD RILs, 27 B6 and 18 D2 individuals; in the case of LGxSM AILs, 10 LG and 9 SM individuals), we performed a linear discriminant analysis on the pairwise distance matrix, using the lda function in the R package mass (Ripley 2002) and parental strain as the separating factor. With two parental strains, a single linear discriminant function exists. We then projected the entire pairwise distance matrix into the space defined by this LD1, using the predict function. 2.3.7 Repeatability To assess repeatability of our micro-CT scanning and computational geometric manipulations, we scanned 34 bacula 81 times (32 bacula scanned twice, one baculum scanned 6 times, and one baculum scanned 11 times), each time removing the specimen and reloading it into Aquafoam and the micro-CT holder. For size, we calculated repeatability as the one minus the median coefficient of variation of centroid size (unbiased standard deviation divided by the mean) across specimens. For shape, we could not calculate repeatability of LD1 because the measure spans zero, meaning we could be dividing by zero to calculate coefficients of variation. Instead, we took the pairwise shape distance of each replicate bone to the other 369 bones in the dataset, then calculated the coefficients of variation per pairwise comparison, averaging across pairwise comparisons. 27 2.3.8 Environmental input To assess the effect of environment, we focused on 28 B6 individuals (10 from the Jackson Laboratory, 9 from the Levitt Lab at USC, 9 from the Williams/Lu Lab) and 17 D2 (7 from the Jackson Laboratory, 10 from the Levitt Lab at USC). We performed an ANOVA with lab origin nested within strain for these 45 parental individuals. There were no other strains for which we sampled substantial number of individuals from different labs. 2.3.9 Heritability Heritability, the proportion of phenotypic variance explained by genetic variance, of size and shape measurements was estimated using a one-way ANOVA to summarize the amount of variance explained by strain. Among our 73 BXD RILs, we collected at least three males for 60. Heritabilities were estimated by focusing on these 60 (though the QTL mapping described below included all 73 RILs). The proportion of variance explained by strain identity was taken as the heritability. 2.3.10 Mapping Quantitative Trait Loci We employed two main analyses to map loci affecting baculum size (centroid size) and shape (LD1). For the BXD RILs, phenotypic values were first averaged for any RILs for which we had sampled multiple individuals. Genotypes were downloaded from GeneNetwork, and included 3,805 markers (2.2% mean missing data per RIL) spaced at a mean 0.54 cM throughout the genome. We kept the 2,953 markers that were part of a relatively recent effort to update the mouse genetic map (Cox et al. 2009). After removing parental strains (which are uninformative 28 since they lack recombinant chromosomes), we employed the scanone function in the R package qtl, using Hayley-Knott regression to estimate the location and the effect of QTL (Broman and Sen 2009). To determine significance, we permuted phenotypes and genotypes 1000 times, and took the 95 th quantile of the 1000 maximum LOD scores as our empirical significance threshold. We estimated the confidence intervals of any significant QTL using the lodint function in the R package qtl, by dropping 1.5 LOD units from the maximal LOD unit on the chromosome. Age and weight could not be included as covariates in analyses of the BXD RILs because not all labs that donated carcasses collected this information, and where they did occur they often varied dramatically given the individuals derived from very diverse research programs. Ignoring age and weight should only add noise to our analysis and inflate Type II error. Unlike the BXD RILs, the LGxSM AILs were not maintained via brother-sister mating, so no two individuals were genetically identical but are also not equally related (A Darvasi and Soller 1995; Ariel Darvasi 1998). The LGxSM AILs were analyzed using the R package qtlrel (Cheng et al. 2011), which accounts for background genetic relatedness prior to scanning for QTL. Genetic relatedness was estimated using all markers except the markers on the same chromosome as the one being analyzed (the “marker” option). Unlike the BXD RILs, age and weight could be included as covariates in our analyses of LGxSM AILs, because all individuals derived from the Cheverud Lab and were reared under identical conditions. Individuals were genotyped at 4,716 markers (0.1% mean missing data per specimen) spaced at a mean 0.36 F2 cM (Cheverud, unpublished). Significance thresholds were estimated with 1000 permutations. Recombination distances were calculated from Cox et al. (2009). 29 2.3.11 RNA sequencing To potentially narrow down QTL identified in the BXD RILs to candidate genes, we generated RNA-seq data from ten 5-week-old B6 (N=5) and D2 (N=5) individuals. For logistical reasons, our B6 bacula derived from C57BL/6N, which is a substrain that is over 99.9% genetically identical to C57BL/6J. The baculum is still not fully ossified at this point (Glucksmann et al. 1976) so genes expressed may ultimately affect its size and shape. Focusing on 5-week-old males admittedly overlooks potentially important expression patterns that occur prior to or after this age. Therefore, we view these data as a means to generate hypotheses rather than definitively link genes under QTL to causality. Because we did not identify any significant QTL among the LGxSM AILs (see below) we only gathered RNA-seq data from the BXD parental strains. Bacula were homogenized in liquid nitrogen, then placed immediately into Trizol according to the Direct-zol RNA MiniPrep (Zymo Research R2052) protocol. RNA integrity was verified by Experion analysis (Bio-Rad). PolyA selection was carried out using Illumina Truseq V2 polyA beads. Libraries were prepared using Kapa Biosystems Stranded mRNA-Seq kit. We performed 12 PCR cycles to amplify the libraries, which were then visualized by Bioanalyzer analysis (Agilent) and quantified by qPCR (Kapa Biosystems Illumina Library Quantification Kit). Sequencing was performed on a NextSeq 500 using V2 chemistry. Paired-end 100 bp sequencing reads were mapped using Tophat (Trapnell et al. 2009) and aligned to the complete Mus musculus GRCm38.73 genome release available from Ensembl. Tophat v2.06 was run on a Linux x86 64-bit cluster with the following parameters: --read-edit- dist 8 --read-mismatches 8 --segment-mismatches 3 --min-anchor-length 12 --report-secondary- alignments. In addition, the GRCm38.73 General Transcript File (GTF file) was included with the “-G” and “–no-novel juncs” tags, ensuring that only known annotated exons were used. 30 These mappings of full length and junction reads were subsequently used by Cufflinks (Trapnell et al. 2009) to generate gene counts. Cufflinks v2.1.0 was run on a Linux x86 64-bit cluster with the following parameters "--multi-read-correct --upper-quartile-norm --compatible-hits-norm -- frag-bias-correct" and with the gene annotation included in the GTF file. Illumina sequencing generated 276 M paired end reads, of which 247 M (90%) mapped reads were used to access expression for the entire annotated transcriptome. A total of 13,926 genes had an FPKM value of at least 1 in at least one of the 10 specimens analyzed. Of these, 479 (610) were significantly differentially expressed at a corrected p=0.05 (0.10) between the D2 and B6 parents. Because our main interest with the RNA-seq data was to generate hypotheses, we were willing to accept increased Type I error (false rejection of a null hypothesis) in favor of reduced Type II error (false acceptance of null), so we used a cutoff of p=0.10 instead of the more traditional p=0.05. From the RNA-seq data, we identified differentially expressed genes (p=0.10 between parental strains after correction by Benjamini-Hochberg) as well as genes that were highly expressed in both parents (in the top 10% of genes expressed with FPKM at least 1). Even though this latter category may not show evidence of differential expression between parental strains, they could still lead to differential baculum development between parental strains, for example through post-translational modification. 2.3.12 Bioinformatics To further identify candidate genes under QTL, we characterized genetic variants that differed between the parental strains D2 and B6. Single nucleotide polymorphisms were downloaded using the Biomart tool from Ensembl version 81 (www.ensembl.org) and we identified genes with at least one nonsynonymous difference between B6 and D2 that also fell 31 under our QTL. We also downloaded their estimates of dN/dS from “one-to-one” orthologs between mouse and rat to scan for genes with unusual rates of nonsynonymous evolution. 2.4 Results We microCT-scanned 369 bacula representing 73 different BXD RILs and the two parental strains B6 and D2. Repeatability was 0.99 for centroid size, and 0.97 for shape divergence, respectively. Lab origin did not significantly influence shape (F 3,1 =1.3, p=0.28), but did contribute to size (F 3,1 =12.68, p<0.0001), explaining 18% of the variance in the latter case. In sum, our methods show good repeatability and any contribution arising from lab origin should not introduce systematic biases into our analyses. 2.4.1 BXD RILs, size Strain was a significant predictor of size variation (F 59,244 =14.8, p<0.01), accounting for 78% of the variance (heritability=0.78), with BXD RILs largely falling between the parental B6 and D2 strains indicative of mostly additive genetic variance (Fig. 2.2). Phenotype means that fell outside either parent suggested transgressive segregation or epistasis. From the 73 BXD RILs, we detected two QTL for size (Fig. 2.3A). One occurred on chromosome 1 between 136.3 and 150.1 Mb, with a maximum LOD score of 7.45, which was higher than all 1000 permutations (95% quantile = 3.63). The other occurred on chromosome 12, 58.0-76.4 Mb, with a maximum LOD score of 4.21, which was greater than 988 permutations (p = 0.012). The two markers closest to the maximal QTL peaks on chromosomes 1 or 12 showed a clear difference in centroid size among BXD RILs carrying a B6 allele versus D2 allele (Fig. 2.4A, B) 32 . Figure 2.2 Baculum Shape and Size Relationships Baculum size and shape variation among the 73 BXD RILs and the two parental strains D2 and B6. Each point represents the mean size and shape for each strain, bars indicate standard errors where possible. B and D indicate the two parental strains of the BXD RILs. B=B6, D=D2. 2.4.2 BXD RILs, Shape Strain was a significant predictor of LD1 variation (F 59,244 =18.1, p<0.01), accounting for 81% of the variance (heritability=0.81), with BXD RILs largely falling between the parental B6 and D2 strains indicative of mostly additive genetic variance (Fig. 2.2). From the 73 BXD RILs, we detected a single QTL on chromosome 2, 154.7-161.7 Mb, with a maximum LOD score of 4.32, which was higher than 988/1000 permutations (p=0.012) (Fig. 2.3B). At the marker closest 33 to the maximum QTL peak of chromosome 2, there was a clear difference in LD1 between the 34 BXD RILs with a B6 allele compared to the 39 with a D2 allele (Fig. 2.4C). 2.4.3 BXD RILs, size and shape correlated Interestingly, the mean size and shape per strain were correlated (Pearson’s correlation coefficient=-0.35, p=0.002, Fig. 2.2), even though major-effect QTL affecting size and shape were found on different chromosomes. This correlation suggests that small bacula tend to have the straight shape of D2, while large bacula tend to have the dorsoventral curvature of B6. Figure 2.3 Baculum Shape and Size QTL Results of scanone analyses for baculum size (centroid size, upper panel) and shape (LD1, lower panel) in the BXD RILs. Dashed line indicates significance threshold determined from 1000 permutations of genotype and phenotype, lines indicate LOD scores testing the null hypothesis of no QTL along 2,953 markers in the genome. 34 2.4.4 BXD RILs, RNA-seq and bioinformatics There were 111 (162) protein-coding genes that occurred under the size QTL identified on chromosome 1 (chromosome 12), of which 93 (145) were expressed, 3 (5) were highly expressed, 1 (2) was differentially expressed, and 11 (14) had at least one nonsynonymous variant (Refer to G3 website for supplementary data). There were 160 protein-coding genes that occurred under the shape QTL on chromosome 2, of which 140 were expressed, 12 highly expressed, 2 differentially expressed, and 27 had at least one nonsynonymous variant. (Refer to G3 website for supplementary data). Literature review of all protein-coding genes that fell under any one of our three QTL (Fig. 2.3) and were also highly expressed, differentially expressed, or had at least one nonsynonymous variant, revealed 16 genes with potentially interesting affects on bone morphology (Refer to G3 website for supplementary data). Four of these - Ptprc, 2700049A03Rik, Aspm, and Kif14 – showed relatively high pairwise dN/dS estimates when compared to their one-to-one ortholog in rat (all above the 75% genome-wide quantile). 2.4.5 LGxSM AILs, Size and Shape From the LGxSM AILs, we microCT-scanned 144 bacula, including 36 unrelated F43 and 89 F44 from 52 unique families, as well as 10 LG and 9 SM parental strains. We could not sample all F43 and F44 male carcasses that were used in previous studies (Rai et al. 2015), because many were missing their penises due to prior dissections. Although both centroid size and LD1 variance showed evidence of largely additive genetic contribution (Supplementary Fig. 2.1), no significant QTL were identified for either (Supplementary Fig. 2.2). Furthermore, there was no correlation between the proportion of an individual’s genome that was from the LG 35 parent and either baculum size (centroid size) or shape (LD1) (Pearson’s correlation coefficient, p=0.98, 0.67, respectively). This is probably an indication that our sample size for LGxSM was underpowered, a hypothesis supported by the lack of significant QTL or correlation between proportion LG/J alleles and body size (Pearson’s correlation coefficient, p=0.75), a trait that is highly heritable in larger studies of LGxSM intercrosses (Cheverud et al. 1996; Kramer et al. 1998). Figure 2.4 Parental Allele Distribution around QTL Baculum size (centroid size) and shape (LD1) from the 73 BXD RILs, separated according to which parental allele they carry at the three QTL identified in Fig. 2.3. (A) rs6202860, the marker closest to the size QTL on chromosome 1. (B) rs3716547, the marker closest to the size QTL on chromosome 12. (C) rs13476871, the marker closest to the shape QTL on chromosome 2. 36 2.4.6 Testing prediction that stem from rapid divergence Our finding of a few QTLs with large effects suggests that morphological divergence of the baculum may accumulate easily, acting via mutations in a few key targets. Several studies (all in insects, most in Drosophila) have found QTL of large effect (>10% of phenotypic variance) on aspects of the male genital apparatus. Many of these studies were initiated by crossing two closely related species, then backcrossing the F1s to either parental species, including two species of carabid beetles (Sasabe et al. 2007), Drosophila simulans and D. mauritiana (LeVasseur-Viens and Moehring 2014; Liu et al. 1996; Kentaro M. Tanaka et al. 2015; True et al. 1997; Zeng et al. 2000), D. simulans and D. sechellia (Macdonald and Goldstein 1999), and D. yakuba and D. santomea (Peluffo et al. 2015). Although large-effect QTLs were detected in all these studies, the extent of linkage that will exist in these backcross designs coupled with the limited number of markers used to interrogate the genome may underestimate the true number of QTL and thus inflate the percent variance explained by any single QTL. Furthermore, it is possible that interspecific studies are biased towards finding large-effect QTL since the different species employed were already known to harbor highly divergent male genitalia. Alternatively, we might predict a bias towards finding small-effect QTL if genetic effects are masked by disruption of normal development in hybrid males. In contrast, studies within species tend to find QTL with smaller effects. Three QTL contributed 4.7-10.7% to variance in the shape of the posterior lobe of genital arch within an advanced intercross panel of D. melanogaster (McNeil et al. 2011). Similarly, a genome-wide association test among 155 inbred strains of D. melanogaster strains of the Drosophila Genetic Resource Panel (Mackay et al. 2012) found multiple SNPs throughout the genome that correlated with small to moderate differences in the size and shape of the posterior lobe (Takahara and 37 Takahashi 2015). Another study in D. montana showed mostly small effect QTL explaining less than 10% of the variance (Schäfer et al. 2011). Taken together, QTL identified in these intraspecific studies are more numerous, with smaller effect sizes, than the interspecific studies discussed above. In this sense, our study is unusual in that it was an intraspecific study but identified a few QTL of large effect, although it should be noted that a small amount of interspecific introgression was detected in both B6 and D2 (Yang et al. 2011). In contrast, a recent morphometric study of mouse skulls found many QTL of very small effect (Maga et al. 2015). Perhaps bacula are unique in that a few major QTL enable their rapid divergence. In our study, QTL affecting size and shape were independent of each other, another genetic characteristic that might allow for more rapid morphological divergence, since mutations may be “less pleiotropic”. Size and shape QTL could not be separated in many of the studies mentioned above (Liu et al. 1996; Macdonald and Goldstein 1999), as the low number of markers employed precluded delineation of multiple QTL. Two studies within D. melanogaster found distinct QTL affecting size vs. shape of male posterior lobes (Schäfer et al. 2011; Takahara and Takahashi 2015). Interestingly, we found that the major QTL affecting size and shape of the baculum were independent of each other (Fig. 2.3), even though size and shape were correlated (Fig. 2.2). One explanation is that many QTL of small effect were not detected here but sufficient to drive an overall correlation between size and shape. Two additional predictions could not be properly evaluated here. First, we predicted that different genetic backgrounds of mice might yield distinct QTL, meaning there are multiple genetic pathways that can generate baculum variation. Some support for this hypothesis comes from a comparison of two quantitative genetic studies that mapped non-overlapping QTL in two different populations of D. melanogaster (McNeil et al. 2011; Takahara and Takahashi 2015). However, one study of a D. 38 mauritiana-D. simulans backcross identified identical QTL as found in a D. mauritiana-D. sechellia backcross (LeVasseur-Viens and Moehring 2014). In the current study, we found three large-effect QTLs in the BXD RILs (Fig. 2.3), none of which were found in the LGxSM AIL panel (Supplementary Fig. 2.2). However, the lack of QTL in the LGxSM AIL panel may simply be due to lack of power. Second, we predicted that baculum variation might arise via alterations in gene expression since both early genital development and bone morphogenesis proceed from highly conserved molecular pathways where protein-coding changes would be expected to have dire consequences (Carroll 2008). Our data do not yet speak to this question; across the 16 candidate genes identified under our three QTL (Refer to G3 website for supplementary data), 7 were highly or differentially expressed among parental strains of the BXD RILs, and 10 had at least one nonsynonymous mutation. Further evaluation of whether baculum variation arises from expression or structural prediction requires finer scale mapping, more detailed expression studies, and/or stronger evidence that certain genes are good candidates for baculum variation. 2.4.7 Rapid genital divergence versus conserved genital development pathways One of the paradoxes of the rapid evolution of male genitalia, at least in vertebrates, is the highly conserved set of genes that appear to be expressed during early genital development, including Sonic Hedgehog (Shh), various fibroblast growth factor receptors (Fgfr’s), and various homeobox-containing genes of the D cluster (Hoxd) (Cohn 2011; Gredler et al. 2015; Haraguchi et al. 2000; Haraguchi et al. 2001; Infante et al. 2015; Klonisch et al. 2004; Miyagawa et al. 2009; Perriton et al. 2002; Sanger et al. 2015; Seifert et al. 2009). In fact, the outgrowth of external genitalia shares many genetic pathways with the development of digits (Cobb and 39 Duboule 2005; Dollé et al. 1991; Haraguchi et al. 2000; Kondo et al. 1997a; Lonfat et al. 2014; Perriton et al. 2002; Tschopp et al. 2014; Warot et al. 1997; J. Zákány and Duboule 1999) or gut (Cohn 2011), although genital-specific enhancers (Lonfat et al. 2014) may enable genetic- specific modifications. Hoxd-13 mutant mice have smaller bacula than wild type (Hérault et al. 1997; J. Zákány and Duboule 1999), and differently shaped bacula (Peichel et al. 1997), and Hoxd genes have enhancers that drive expression specific to external genitalia (Dollé et al. 1991; Lonfat et al. 2014). When Hoxd-13, Hoxd-11 and Hoxd-12 were all experimentally made nonfunctional, no genitals developed (Kondo et al. 1997a). However, all these genetically manipulated individuals show extreme dysmorphia in body plan, and many are lethal before birth. Importantly, none of the QTL that we observed here overlap with Shh, Fgfr’s, or various Hoxd genes, suggesting the variation we observe is not due to genetic variation in these conserved pathways, although it is formally possible that something under our QTL affects expression of these genes in trans. 2.4.8 Candidate genes Our study offers the first set of candidate genes to explain variation in baculum size and shape. Although systematically testing which (if any) of these protein-coding genes explain baculum variance remains outside the scope of the current study, we prioritized genes based on three criteria: 1) they fell under QTL identified above (Fig. 2.3) 2) they were highly expressed or differentially expressed in 5-week-old bacula, or had at least one nonsynonymous variant between parental strains, and 3) literature searching revealed a potential link to bone or genital morphogenesis. The baculum forms after birth, through ossification of hyaline cartilage as the animal approaches sexual maturity (Murakami and Mizuno 1984), so bone-related phenotypes 40 are potentially interesting. By these criteria, we identified 16 candidate genes (Refer to G3 website for supplementary data). Four genes under the size QTL of chromosome 1 are potentially interesting candidate genes. Kif14 and Aspm have been shown to cause cranial deformities and reduced body size (Fujikura et al. 2013; Pulvers et al. 2010). A third, Ptprc, alters bone morphology when knocked- out (Shivtiel et al. 2008). Finally, mice deficient in Mgat2 exhibit reduced body size and reduced bone density due to hyperactive osteoclasts (Yan Wang et al. 2001). Under the size QTL of chromosome 12, Dact1 has been shown to negatively regulate Wnt signaling (Wen et al. 2010), which in turn affects the development of bone and genitals (Berendsen and Olsen 2015; Haraguchi et al. 2000; Haraguchi et al. 2001; Hu et al. 2005; Qin et al. 2012; Yamaguchi et al. 1999). Wnt signaling has also been implicated in the divergence of male genitalia in flies (Kentaro M. Tanaka et al. 2015). Mice missing a functional Dact1 allele lacked external genitals and displayed numerous skeletal abnormalities (Wen et al. 2010). Another gene under this QTL, Hif1a, also affects bone density (Ying Wang et al. 2007). Two genes under the shape QTL of chromosome 2 influence bone development. Rbl1 is a key regulator of cell development that if inactivated results in shortened limbs, defective endochondral ossification, altered chondrocyte growth (Cobrinik et al. 1996), and reduced body size (Scimè et al. 2005). A second gene, Tgm2, plays a role in chondrogenesis and ultimately bone formation (reviewed in Iismaa et al. 2009). 2.5 Conclusions Our study provides the first candidate genes to explain variation in baculum size and shape, a structure with an astonishing rate of evolutionary divergence. Our study provides 41 evidence that baculum variation is explained by a few QTL of large effect that independently affect size and shape with apparently minimal pleiotropic side-effects, and helps reconcile the paradox of rapid morphological divergence coupled with conserved developmental pathways. Future experiments should focus on testing candidate genes identified here as a means to genetically dissect its function, and to provide a deeper understanding of the rapid evolution of this amazing structure. Supplementary Figure 2.1 LGxSM AIL Shape and Size Distribution Phenotypic distribution of baculum size (centroid size, upper panel) and shape (LD1, lower panel) from LGxSM AILs. Green indicates parental strains (in upper panel SM on left, LG on right; in lower panel, LG on left, SM on 42 right), red indicates F44 individuals and blue indicates F44 individuals. Error bars are the standard error (unbiased standard deviation divided by the mean) for any families where multiple individuals were phenotyped. Supplementary Figure 2.2 LGxSM AIL QTL Scans for QTL in the LGxSM for size (centroid size) and shape (LD1) failed to yield any significant correlations. 43 3 The baculum was gained and lost multiple times during mammalian evolution This chapter appears principally as published in 2016 in Integrative and comparative biology 56.4: 644- 656. 3.1 Abstract The rapid evolution of male genitalia is a nearly ubiquitous pattern across sexually reproducing organisms, likely driven by the evolutionary pressures of male-male competition, male-female interactions, and perhaps pleiotropic effects of selection. The penis of many mammalian species contains a baculum, a bone that displays astonishing morphological diversity. The evolution of baculum size and shape does not consistently correlate with any aspects of mating system, hindering our understanding of the evolutionary processes affecting it. One potential explanation for the lack of consistent comparative results is that the baculum is not actually a homologous structure. If the baculum of different groups evolved independently, then the assumption of homology inherent in any comparative studies is violated. Here, we specifically test this hypothesis by modeling the presence/absence of bacula of 954 mammalian species across a well-established phylogeny and show that the baculum evolved a minimum of 9 times, and was lost a minimum of 8 times. Furthermore, groups with a baculum show evidence of higher rates of diversification. Our study offers an explanation for the inconsistent results in the literature, and provides insight into the evolution of this remarkable structure. 44 3.2 Introduction Understanding the evolutionary forces that drive rapid divergence of morphological structures is a fundamental part of understanding adaptation, speciation, and the diversity of life. Across nearly all sexually reproducing organisms, male genital anatomy evolves more rapidly than other known morphological structures (William G Eberhard 1985; Klaczko et al. 2015; Romer and Parsons 1986). In fact, male genitals diverge so rapidly that taxonomists often use them to distinguish closely related species that are otherwise morphologically indistinguishable (D. R. Adams and Sutton 1968; Hamilton 1949; Patterson and Thaeler 1982; Simson et al. 1993). The baculum is a bone that occurs in the penis of many mammal species, and they display astonishing morphological diversity (W. Burt 1960a; Chaine 1925; AF Dixson 1995; Eadie 1947; Romer and Parsons 1986; Weimann et al. 2014). Qualitatively, interspecific divergence exceeds intraspecific polymorphism, a classic signature that suggests the baculum is a target of recurrent adaptive evolution. Several hypotheses for the function of the baculum have been proposed which lead to testable predictions in a comparative framework. Unfortunately, comparative studies have failed to yield general and consistent results (summarized in Table 3.1). One hypothesis is that the baculum protects the urethra and provides mechanical support during copulation (Dyck et al. 2004; Long and Frank 1968; Oosthuizen and Miller 2000). Dixson found that in pinnipeds and primates, species with prolonged intromission tended to have more elongate bacula (AF Dixson 1987b, 1995; Alan Dixson 1998), However, across 52 species of carnivores, baculum length did not covary with intromission duration (Larivière and Ferguson 2002). Although Dixson (1995) concluded that carnivores with increased intromission length had longer bacula, that study did not incorporate modern phylogenetic control. 45 Another set of hypotheses maintain the baculum functions in different aspects of the mating system, for example by stimulating the female in a way that biases paternity towards a particular male (male-female interactions), assisting in the removal of sperm from prior males (sperm competition to increase “offensive” strategies of males), or by inducing damage to the female to inhibit remating (sexual conflict to increase “defensive” strategies of males). These latter two hypotheses predict a correlation between baculum morphology and the inferred risk or intensity of sperm competition. In an experimental evolution study within a single species of mouse, males evolved relatively wider bacula when subjected to high levels of sperm competition compared to enforced monogamy (Simmons and Firman 2014), lending support to the hypothesis that bacula function in the context of sperm competition. Across rodents and carnivores, baculum length increased with the inferred intensity of sperm competition (S. A. Ramm 2007), but no such correlation was found in bats or primates (D. J. Hosken, K.E. Jones, K. Chipperfield, and A. Dixson 2001; S. A. Ramm 2007). A corollary prediction is that baculum features in a trade-off with sexual dimorphism, since very strong dimorphism often indicates that males are investing disproportionately in precopulatory rather than postcopulatory competition (James P. Dines et al. 2015; Stefan Lüpold et al. 2014; G. A. Parker et al. 2013). In some pinnipeds, there was a negative correlation between sexual size dimorphism and baculum size (Fitzpatrick et al. 2012), but this pattern was not observed across other carnivores (Larivière and Ferguson 2002). If large bacula somehow signal indicate male quality (S Lüpold et al. 2004; Miller and Burton 2001), then relatively fit males should invest disproportionately in ever larger bacula, and a positive allometric relationship should arise (but see Bonduriansky 2007). The relationships between baculum size and body size are inconsistent, ranging from positive allometry (Miller et 46 al. 1999; Miller and Burton 2001; D. Tasikas et al. 2009b), to isometry or even negative allometry (S Lüpold et al. 2004; S. Ramm et al. 2010; Schulte-Hostedde et al. 2011). In sum, there are no consistent relationships between features of the baculum and organismal biology (Table 3.1), hindering our understanding of the evolutionary forces affecting their morphological diversity. There are at least five potential explanations to reconcile inconsistencies from the literature. First, nearly all studies in Table 3.1 focus on the length of the baculum, and shape may be a more important parameter to test some of these hypotheses (Baryshnikov et al. 2003; Paula Stockley et al. 2013). Unfortunately, modern morphometric techniques have only recently begun being applied to studies of baculum morphology (N. G. Schultz et al. 2016b). Second, the baculum may function in distinct biological processes across species, so that correlations to one group’s biology need not apply to another’s (D. Kelly 2000; Patterson and Thaeler 1982). Third, the baculum may evolve so rapidly that it outpaces evolutionary correlation to other characters. Fourth, heterogeneity in the morphological placement of the baculum implies the baculum might function in different contexts. For example, in some groups the baculum is at the distal extreme of the glans, while in others it is more deeply embedded proximally (Patterson 1983). Fifth, and the main topic of the current study, is that the baculum evolved more than once. If the baculum has similarly evolved multiple times, then the baculum should not be considered a homologous structure, in that it was not inherited through common descent from a mammalian ancestor with modification. Instead, multiple derivations would suggest that the baculum evolved in different biological contexts, possibly to solve different evolutionary challenges, confounding any straightforward correlations to aspects of organismal biology. Most importantly, multiple derivations would violate the fundamental assumption of homology that is 47 inherent in any comparative analysis. Through intensive literature review and phylogenetic analysis, we provide strong evidence that the baculum has been gained and lost multiple times through mammalian evolution, and evidence suggests that groups which have evolved a baculum diversify more rapidly than those without. Our study helps explain inconsistencies observed in the literature and provides valuable insight into the evolution of this astonishing structure. Table 3.1 Hypothesized Bacula Functions a Dixson (1987a); b Dixson (1987b); c Dixson (1995); d Lariviere and Ferguson (2002); e Ramm (2007); f Hosken et al. (2001); g Fitzpatrick et al.(2012); h Stockley et al. (2013); i Simmons and Firman (2014); j Tasikas et al. (2009); k Miller and Burton (2001); l Miller et al. (1999); m Schulte-Hostedde et al. (2011); n Ramm et al. (2010); o Lupold et al. (2004). 3.3 Materials and Methods Any phylogenetic analysis will be sensitive to the exact taxa sampled. For example, non- randomly including more species with bacula would bias the estimated rate of baculum gain upwards, since more evolutionary time would be spent with a baculum. To avoid such biases, we included as many species as possible in a large mammalian phylogeny, without a priori knowledge of their baculum status. It is possible that all previously published phylogenies are inherently biased towards including species with a baculum, since taxonomists often use this structure to delineate species when other morphologies fail to distinguish them. However, we expect such a bias to add relatively short external branches - for example to separate extremely closely related species - and therefore would not compromise analyses over deeper evolutionary 48 time. After describing our analytical pipeline, we introduce three different bootstrapping strategies to account for potential taxonomic and phylogenetic bias. 3.3.1 Phylogeny Six phylogenies were merged to create a tree of 3,707 mammal species. The first, from Meredith et al. (2011), was a family-level phylogeny across mammals and served as the scaffold to which five more taxonomically focused studies were added. Meredith et al. (2011) used a likelihood framework to analyze a supermatrix that included 164 mammal species plus five outgroups, taken from 26 gene fragments consisting of 35,603 base pairs (bp) and 11,010 amino acids. Although 164 species is a small fraction of the over 5,000 mammalian species (Nowak 1999; Wilson and Reeder 2005), they included at least one representative from nearly every mammalian family, providing a reasonable foundation for phylogenetic inference. We then replaced specific nodes of the Meredith et al. (2011) with larger phylogenies described below, in all cases renormalizing branching times to the Meredith et al. (2011) scale. Shi and Rabosky (2015) used a likelihood framework to analyze a supermatrix that included 812 bat species, gathered from 29 loci of 20,376 bp. All 20 bat families were represented by at least one species. We replaced the single bat clade from the Meredith et al. (2011) phylogeny with the single clade of 812 bat species from Shi and Rabosky (2015). Fabre et al. (2012) used a maximum likelihood framework to analyze a supermatrix that included 1,265 rodent species, which represents more than 80% of known generic diversity in rodents, utilizing 11 loci. Bacular diversity has been widely studied in rodents in particular (W. Burt 1960a; Wade and Gilbert 1940), and they represent the most speciose order of mammals, so the addition of a separate rodent phylogeny was deemed necessary for our analysis. The Fabre et al. (Fabre et al. 2012) phylogeny replaced the single rodent clade in Meredith et al. (2011). 49 McGowen et al. (2009) developed a molecular phylogeny through a Markov Chain Monte Carlo Bayesian analysis using 45 nuclear loci, transposons, and mitochondrial genomes from 87 Cetacean species. Again, the McGowen et al. (2009) phylogeny replaced the single Cetacean clade in Meredith et al. (2011). Nyakatura and Bininda-Emonds (Nyakatura and Bininda-Emonds 2012)built a supertree of 286 carnivore species using matrix representation parsimony from existing phylogenetic hypotheses and molecular data. The dataset included 114 phylogenetic hypotheses as well as 74 novel trees derived from 45,000 bp of sequence data. This phylogeny replaced the carnivore clade in Meredith et al. (2011). Perelman et al. (Perelman et al. 2011) amplified 34,927 bp sequenced from 54 homologous genomic regions of primate species representing 186 species, then built a phylogeny using maximum likelihood. This primate phylogeny was the only one of the additional five that was not published as ultrametric. We therefore converted this phylogeny to an ultrametric tree using the chronos function in the R package ape (Paradis et al. 2004; Popescu et al. 2012), using 8 fossil calibration dates provided in the legend of Fig. 3.1 of Perelman et al.(Perelman et al. 2011). The overall combined phylogeny contained 3,707 species, and was normalized to the scale of Meredith et al. (2011) to make it ultrametric. Uncertainty in the branching patterns was not considered, but any uncertainty is unlikely to greatly alter our main conclusions. The reason for this is that the bacula arise either in localized groups of related species (or single lineages) or at deeper nodes that unite species at approximately the family level (Fig. 3.1). It is difficult to envision how minor branch swapping could affect our main conclusion, which is that the baculum evolved multiple times. 50 3.3.2 Baculum status Through literature searching, we were able to score the presence/absence of the baculum in 1,028 species (925 with a baculum, 103 without) (Refer to ICB website for supplemental data). Most of these were represented in the phylogeny we created above so that patterns of evolution could be evaluated. The primary sources were the Index for Mammalian Species (Hayssen), and Asdell’s Patterns of Mammalian Reproduction (Asdell and Hubbs 1964). Additional data came from searches in Google Scholar (www.scholar.google.com), with the phrases baculum, bacula, os penis, os priapi (Latin, penis bone), l’os pénien (French, penis bone), penisknochen (German, penis bone), báculo (Portuguese, baculum), and various species names, museum databases iDigBio (www.idigbio.org) and Morphobank (www.morphobank.com), as well as personal communication with taxonomic experts. All primary sources are listed on the ICB website. Any mention of ossification in the penis was considered a baculum, however sources were used only if they included images or measurements of bacula, and listed the exact species names. If a source did not indicate species name for the baculum, the closest taxonomic level was used. For example, Cratogeomys castanops was listed in Fabre (2012), but we could only find information on baculum presence for one of its congeners, Cratogeomys merriami (W. Burt 1960a). In this case, we scored C. castanops as having a baculum. Methodologically, this makes one of two assumptions: 1) species within the same genus share baculum state, or 2) species from the same genus can be swapped out in the phylogeny (in this case, we could have replaced Cratogeomys castanops in the phylogeny with Cratogeomys merriami). Of the 1,028 species scored for baculum presence, 993 had species-level evidence (i.e., species names matched in both the phylogeny and the baculum literature), and 35 had genus-level evidence (e.g., the Cratogeomys example just discussed). We excluded any species for which the 51 evidence was above the genus level, and we also re-ran our analyses below after excluding the 35 species with genus-level evidence. The literature is rife with claims of baculum presence/absence without data, which we excluded here. For example, it is often stated that no cetaceans have a baculum, when in fact only a few cetacean studies specifically report on the baculum (Refer to ICB website for supplemental data). Similarly, two studies that simply mention the possibility that moon rats (Podogymnura) have a baculum (Gerhardt 1909; Kaudern 1907) are commonly mis-cited as providing evidence the baculum exists. 3.3.3 Testing evolutionary hypotheses There was an overlap of 954 taxa between the 3,707 species in the phylogeny and the 1,144 species scored for baculum presence/absence. With these 954 species, we estimated the number of times bacula have been independently gained and lost using stochastic mapping as implemented in the function make.simmap of the R package phytools (Revell 2012). Given an observed phylogenetic tree and distribution of character states, stochastic mapping generates multiple iterations of character evolution that are consistent with the observed character states, using a continuous time-reversible Markov model. There are two main stages in stochastic mapping (Bollback 2006; Huelsenbeck et al. 2003; Nielsen 2002). First, the probabilities of possible ancestral states at all interior nodes are calculated (Felsenstein 1981). A collection of ancestral states is then sampled according to their state probabilities at each node. Every branch then starts at state i and ends at state j, with character states at the tips simply the observed baculum presence/absence for each species. 52 Second, potential character transitions are placed over each branch. For our particular study, as there are only two character states (baculum present or absent), there are two corresponding transition rates, which equal 1/mean time to switch out of each respective state. In short, stochastic mapping produces randomly sampled character state histories that are consistent with the states at the tips of the tree by estimating transition rates and sampling ancestral states at internal nodes. We summarized baculum gains and losses from 1000 such histories. Visual representations were made using the densityMap function of phytools (Revell 2012). We considered branches where at least 50% of the iterations showed a gain or loss as “high confidence”. Since any cutoff is admittedly arbitrary, we also present the number of gains and losses observed in at least 95% of the iterations. Stochastic mapping has a number of advantages over more traditional parsimony methods. Parsimony underestimates both the mean and variance of the number of character state changes because it allows at most a single transition along a branch for characters with only two states (Bollback 2006). With stochastic mapping, characters are allowed to change multiple times along a single branch, which allows uncertainty in the exact reconstruction of ancestral states to be accounted for (Nielsen 2002). Nevertheless, for comparative purposes we also evaluated transitions in a strict parsimony framework, using the ancestral.pars function in the R package phangorn (Schliep 2010). 3.3.4 Bootstrapping To evaluate the robustness of our results, we repeated the stochastic mapping procedure under three different bootstrapping regimes, all written with customized scripts in R (available 53 from the authors at request). Under each regime, we subsampled 50%, 60%, 70%, 80%, or 90% of the species and repeated the stochastic mapping procedure 100 times each. The first and most straightforward version of bootstrapping is to subsample species uniformly, with each species in the dataset equally likely of being excluded from any one bootstrap iteration. We refer to this first version as “uniform bootstrapping”. Although computationally easy to implement, this version of boostrapping ignores potential taxonomic and phylogenetic biases, which the next two versions of boostrapping address. Second, in order to address potential taxonomic bias, we repeated the bootstrap, but species from mammalian clades that were sampled more (relative to the number of known extant species) had higher probabilities of being excluded during any one bootstrap iteration. For example, 95 of 294 (32%) known carnivore species, and 4 of 452 (0.009%) of known Eulipotyphla species, could be included in our current dataset (Supplementary Table 2). Therefore, carnivore species would be 32 / 0.009 = 4000 times more likely to be excluded in any one bootstrap replicate compared to Eulipotyphla species. We refer to this second version as “proportional bootstrapping”. Third, to account for potential phylogenetic bias, we weighted each bootstrap replicate according to the average relationship of each species to all other species in the phylogeny, calculated using the cophenetic function in the R package ape (Paradis et al. 2004; Popescu et al. 2012). Species that were more closely related to other species on the phylogeny had a higher probability of being excluded from bootstrap replicates, in an attempt to more evenly sample the mammal phylogeny. For example, the sister group Oryzomys rostratus + Oryzomys melanotis was separated by a cophenetic distance of 10 -5 , while the sister group `was separated by 1.33 (with cophenetic distance in units of 100 million years). In this example, one of the Oryzomys 54 species would be 1.33/10 -5 = 13,000 times more likely of being excluded in a bootstrap iteration. This bootstrap had to be recursive, where we excluded one species at a time, then re-evaluated the cophenetic matrix for the remaining (n-1) species before excluding the next, until the appropriate number of species was dropped. We refer to this third version as “cophenetic bootstrapping”. 3.3.5 Correlating baculum presence to rates of diversification We tested whether groups with or without a baculum differed in estimated rates of diversification using methods of Binary State Speciation and Extinction (BiSSE). BiSSE was implemented using the R package diversitree, by creating a likelihood function with the make.bisse function (FitzJohn 2012). This function takes on 6 parameters: the speciation and extinction rates in groups with vs. without a baculum, plus the transition rates from baculum absent->present and present->absent. Initial starting points for these six parameters and the necessary likelihood function were estimated using the starting.point.bisse function, and the likelihood of the model estimated using the find.mle function (FitzJohn 2012). The null model constrained the two speciation rates to be equal, using the constrain function, then a likelihood ratio test (LRT) performed to infer whether the two speciation rates were significantly different from each other (FitzJohn 2012). 3.4 Results 3.4.1 The baculum was gained and lost multiple times Across 1000 iterations of stochastic mapping, the baculum evolved an average of 9.5 times and was lost an average of 11.5 times (Fig. 3.1). Gains and losses clustered along 19 branches in the phylogeny (Fig. 3.1). Specifically, 9 branches gained a baculum, and 10 branches 55 lost a baculum in at least 50% of the 1000 iterations of stochastic mapping, with 5 gains and 6 losses occuring in at least 95% of the iterations. We thus conclude that the baculum evolved a minimum of 9 times and was lost a minimum of 10 times throughout mammalian evolution. We now discuss these 19 gains and losses in more depth, in general from “top to bottom” of the phylogeny (Fig. 3.1). The baculum is only found among Eutherian mammals, and absent in basal species of Metatherians. Therefore, the ancestor of mammals lacked a baculum. Four baculum gains occurred in single lineages on our phylogeny – the hedgehog tenrec (Echinops telfairi, 99% of iterations, Fig. 3.1B), the American pika (Ochotona princeps, 83%, Fig. 3.1E), Spix’s disc- winged bat (Thyroptera tricolor, 97%, Fig. 1F), and the European mole (Talpa europaea, 100%, Fig. 1G). In primates, the ancestral state remains uncertain with our data, but there were at least two independent gains, one in lemur-like primates (Strepsirhini, 66%) and another leading to a subset of monkeys and apes (Simiiformes, 55%) (Fig. 3.1C). The latter gain was followed by three independent losses, one in the Cacajao + Chiropotes clade (88%), one in the Lagothrix + Ateles clade (89%), and one in humans (Homo sapiens, 100%) (Fig. 3.1C). We are unaware of any scientific publication to suggest that extinct Homo species, including H. neanderthalensis, had a baculum. The lack of a baculum in Tarsius syrichta (Fig. 3.1C) was not strongly resolved to a loss in that lineage given the uncertainty in transitions at the base of the primate clade. The common ancestor of all rodents gained a baculum (82%), followed by the maintenance of a baculum in all rodent species that we could include here (Fig. 3.1D). This was somewhat surprising, and implies that the incredible morphological diversity of rodent bacula (W. Burt 1960a) occurs against a backdrop of evolutionary constraint maintaining the structure. 56 Bats showed somewhat complicated patterns (Fig. 3.1F). The baculum was gained in the common ancestor of all bats (99%), followed by five independent losses in 1) the ancestor of the 42-species clade that includes Myzopoda aurita (77%), 2) the big-crested mastiff bat (Promops centralis, 97%), 3) the dwarf dog-faced bat + southern dog-faced bat (Molossops temminckii + Cynomops planirostris, 76%), 4) the western mastiff bat (Eumops perotis, 99%), and 5) the two Miniopterus species (Miniopterus schreibersii + Miniopterus minor, 96%). Thyroptera tricolor represents the only lineage in the mammal phylogeny that traces through two independent gains (one in the ancestor of all bats, followed by loss in the ancestor of the 42-species clade that includes Myzopoda aurita, followed by a second independent gain in Thyroptera tricolor). The common ancestor of all carnivores gained a baculum (99%), followed by two independent losses: one in the ancestor of aardwolf and two hyaena species (Proteles cristata + Hyaena hyaena + Crocuta crocuta, 100%) and one in the bearcat (Arctictis binturong, 100%) (Fig. 3.1G). It is interesting that the spotted hyaena lost the baculum, as this species has famously high levels of circulating androgens, even among females (Glickman et al. 1987). We re-ran the stochastic mapping procedure after excluding the 35 species for which baculum presence/absence was inferred from a congeneric species. Our results changed very little, and we inferred a minimum of 8 independent gains and 10 independent losses. The reason we have one fewer gain species is because Tarsius syrichta was one of the 35 species removed in this follow-up analysis, which leads to inference of a single, high confidence gain in the primate ancestor, supported by 67% of iterations. As comparison, we also ran a strict parsimony reconstruction of ancestral states, which revealed a minimum of 5 independent gains and 10 independent losses. 57 Figure 3.1 Stochastic Mapping Results A) High confidence gains and losses of the baculum derived from stochastic mapping. Red branches indicate species with a baculum, blue without; numbers in red (blue) boxes indicate the proportion of iterations that mapped a baculum gain (loss) to those particular branches. Branches where a transition occurred in at least 50% of the iterations are highlighted in B-H. 58 Figure 3.2 Bootstrapping Techniques Three different bootstrapping techniques were applied (see Methods). Following the color scheme in Fig. 3.1, red indicates baculum gain and blue baculum loss. Boxplots show the results of 100 iterations at each level of resampling. Red and blue diamonds indicate the number of 9 high confidence gains and 10 high confidence losses observed in the full dataset. A) Uniform bootstrapping, B) Proportional bootstrapping, C) Cophenetic bootstrapping. 59 3.4.2 Boostrapping The three versions of boostrapping – uniform, proportion, and cophenetic – revealed our general conclusions that the baculum has been gained and lost multiple times during mammalian evolution (Fig. 3.2). We therefore pooled the bootstrapping results. After subsampling 50%, 60%, 70%, 80%, and 90% of the species in our dataset, we observed a median (2.5%-97.5% quantile) of 6 (1-10), 6 (2-11), 7 (3-11), 8 (5-11), and 9 (6-11) high confidence gains, respectively, and 7 (2-13), 8 (4-13), 8 (5-12), and 9 (6-11), 10 (7-10) high confidence losses, respectively. These numbers are close to those inferred from the whole dataset, where the baculum was gained a minimum of 9 times and lost a minimum of 10 times. In sum, the inference that the baculum has been gained and lost multiple times during mammalian evolution is robust to our exact sampling, and to unknown biases in taxonomic or phylogenetic studies. 3.4.3 Groups with a baculum diversify rapidly Groups with a baculum diversify more rapidly than groups without (LRT=58.7, df=1, p<10 -13 ). The full model shows that the estimated rate of diversification in groups with a baculum is more than 3 times higher than those without (lambda1=0.071 vs. lamdba0=0.022 new species per million years). However, there are two important caveats. First, the baculum might be correlated to the overall number of known species not because those groups diversify more rapidly, but simply because species with a baculum are easier for taxonomists to describe. One way to test this caveat is to score the number of species descriptions that rely on baculum morphology. Second, these methods assume that the phylogeny represents a random sample of all known extant species, which is almost certainly not the case. For example, we only included 3 of the roughly 100 species of cetaceans because although it is generally believed that cetaceans do not have a baculum, it was only specifically reported in three species. Adding 100 cetacean 60 species without a baculum to our analysis would obviously increase the estimated rate of diversification among groups without a baculum. Therefore, we cautiously suggest that groups with a baculum might diversity more rapidly, but future expansion of our datasets are required to understand this pattern. 3.5 Discussion The baculum is an extremely diverse morphological feature and there have been many attempts to uncover correlates between aspects of baculum morphology and mating ecology, with a roughly equal number of positive and negative results (Table 3.1). Due to publication bias, there are probably relatively more negative results that remain unknown, suggesting baculum characteristics are not strongly or consistently correlated with aspects of organismal biology. The studies highlighted in Table 3.1 include bats, rodents, carnivores, and primates. Our study shows that these four groups evolved their bacula independently, potentially reconciling an inconsistent literature. Namely, comparisons across groups assume that the structure being studied is homologous, inherited via common descent with modification, but our study clearly demonstrates this assumption is violated in studies of the baculum. Instead the baculum appears to have evolved under many different ecological contexts. In addition, likelihood analyses suggest that rates of diversification are elevated among groups with a baculum, although caveats suggest a cautious interpretation. 3.5.1 Multiple derivations Specific details of our conclusions will probably be amended as more species are examined for the presence/absence of bacula and phylogenetic hypotheses expand. The baculum 61 can be easily overlooked – for example, the tiny baculum of the American pika (Ochotona princeps) was only recently characterized through scanning electron microscopy, mass spectrometry, and cross-sectional histology (Weimann et al. 2014). However, given the phylogenetic distribution of bacula (Fig. 3.1), it is difficult to imagine that including more species in the future would overturn our main conclusion that the baculum evolved multiple times. In fact, our study may underestimate the number of times that bacula have evolved. In the present study “baculum” refers to a bone in the penis, but this necessary simplification hides important heterogeneity that hint at additional independent origins. Bacula differ in their location in the penis and the distribution and type of ossification. Some bacula are expansive (Sharir et al. 2011), covering more than 75% of total penis length (Sinha 1976), while others are not (Hooper 1960; Rodriguez et al. 2011). Even the placement varies; while most bacula are dorsal to the urethra, and attach proximally to the corpora cavernosa (Evans and de Lahunta 2013; Rodriguez et al. 2011), the giant panda (Ailuropoda melanoleuca) baculum lies ventral to the urethra and does not attach to the distal aspect of the corpus cavernosum (Davis 1964). The distribution of woven or lamellar bone in the baculum also varies across species. Lamellar bone has a more organized structure than woven bone, which enables it to be mechanically stronger (Bonewald et al. 2009). The mid-shaft of the rat baculum is mostly composed of dense lamellar bone, and shows signs of active bone remodeling, suggesting a role in load bearing (D. Kelly 2000). The shaft of the baculum in some bats is composed of lamellar bone surrounded by woven bone (Herdina et al. 2015b). Structural support of the baculum may further be altered by the expanse of medullary cavities within the baculum, as some extend into the shaft (Herdina et al. 2015b), while other medullary cavities remain in the proximal region 62 (Rodriguez et al. 2011). In addition to the distribution of bone, the type of ossification in the baculum varies. Intramembranous, or direct, ossification occurs when undifferentiated mesenchyme is ossified, while endochondral ossification requires a cartilage intermediate (De Crombrugghe and Akiyama 2009). Both types of ossification have been observed in distinct regions of the rat baculum during development (Murakami and Mizuno 1984), while only endochondral ossification is observed in other species (Evans and de Lahunta 2013; Smirnov and Tsytsulina 2003). Even patterns of ossification vary, ranging from simultaneous ossification arising from two distinct zones (Evans and de Lahunta 2013), in two zones at separate developmental stages (Murakami and Mizuno 1984; Yoon et al. 1990) or more numerous ossification centers (Callery 1951). Developmental timing of ossification also varies; the mouse baculum is barely visible in neonates (Glucksmann et al. 1976), while some bat bacula develop by late embryonic stages (Smirnov and Tsytsulina 2003). Taken together, these studies reveal heterogeneity in baculum development which may in turn suggest additional derivations. Unfortunately, the detailed developmental data required to assess this hypothesis are simply lacking for most species. 3.5.2 Sexual selection One model that is often invoked to explain the diversity of bacula is one of sexual conflict, whereby the baculum is a male “offensive” trait that continuously evolves to counteract female “defenses”, which could lead to recurrent adaptive evolution of both male and female traits. The morphological diversity of the baculum may in fact fit such a model, but it does not seem like the presence/absence of the baculum itself does. Instead, a gain or loss (Fig. 3.1) is often followed by extreme amounts of evolutionary time without another transition. For 63 example, the astonishing morphological diversity found in rodent bacula (W. Burt 1960a) appears to have arisen against a backdrop of selective constraint maintaining its presence of the bone, as no rodents have lost it. 3.5.3 Developmental biology Our study brings up several important questions. First, do independent derivations of bacula proceed via switching on/off of conserved genetic pathways, or through the recruitment of novel molecular pathways? In sticklebacks, loss of pelvic girdles occurred via multiple independent mutations that affect expression of Pitx1 (Bell 1987; Chan et al. 2010; M. D. Shapiro et al. 2004). Many genes involved in growth and patterning are shared between limbs and genital tubercles (Cobb and Duboule 2005; Infante et al. 2015; Kondo et al. 1997a), and it is possible that some of these genes also affect baculum development. There is evidence that species which have lost a baculum retain the developmental pathways to develop on. Hershkovitz (Hershkovitz) identified vestiges of embryonic bacula in adult Cacajao (Primate, New World Monkey) specimens, the adults of which do not have bacula. Thus, bacula may be similar to other cases of arrested development followed by degeneration, including hind-limb regression in cetaceans (Thewissen et al. 2006), phallus regression in cloacal birds (Herrera et al. 2013), and the loss of teeth in birds (Harris et al. 2006). These examples show how loss of a trait as an adult is sometimes accompanied by early embryonic development of the trait, suggesting that conserved genetic pathways may be poised for subsequent rederivation. Clearly, more studies are needed to elucidate the genetic basis of baculum variation. 64 3.5.4 Functional biology Even a basic understanding of the function of the baculum will never be complete without knowledge of its precise interactions with the female during copulation. This is perhaps the largest obstacle to testing hypotheses of baculum function and evolution, as it requires observations of internal anatomy in naturally behaving, copulating animals. All bacula are thought to reside in the glans penis, which enters the female’s reproductive tract during copulation (Hooper 1960). Male genital morphology has been shown to evolve in response to complexity of the female reproductive tract (Brennan et al. 2007; Brennan et al. 2010; Higginson et al. 2012), with dramatic effects on paternity (Goran Arnqvist and Danielsson 1999; Dougherty et al. 2015; C. M. House and Simmons 2003; Paula Stockley et al. 2013). Herdina (2015a) showed that artificial inflation of the corpora cavernosa greatly altered the relative orientation of the bones to variable degrees in all three bat species tested, demonstrating that an understanding of the baculum must also take into account the effect of surrounding tissue and state of the penis. The number of erectile tissues and the degree to which they contribute to erections also varies in species with bacula (Christensen 1954; Davis 1964; Rodriguez et al. 2011), potentially adding even more diversity to baculum function. Many, but not all, bacula have cartilaginous extensions on their most distal tip, some of which extend past the most distal aspect of the penis and also show tremendous morphological diversity (Herdina 2008; Herdina et al. 2015a; Hooper 1960; Rodriguez et al. 2011). Such cartilaginous structures may interdigitate with the female reproductive tract (Meczyński 1974), or alter orientation of the penis to align properly during copulation (Evans and de Lahunta 2013). While these hypotheses remain untested, the variability in distal cartilage adds another layer of complexity to understanding the functional role of the baculum. 65 3.6 Conclusion Most mammals have bacula, but we still understand very little about the function, development, and origin of this bone. The lack of consistent correlations to aspects of organismal biology leaves many questions unanswered. In the current study we demonstrate that the baculum can no longer be considered a homologous structure in the traditional sense. Rather, multiple gains and losses of the bone suggest species-specific responses to species-specific challenges. To reveal these biological challenges, future studies should focus on developmental and functional dissection of this remarkable structure. 66 4 Understanding Mouse Baculum Biomechanics through an Investigation of Bone Histology 4.1 Abstract Male genitalia are some of the most rapidly evolving and morphological diverse structures in nature, a pattern exemplified by the mammalian penis bone. A tremendous amount of research has been conducted in pursuit of understanding the function of these bones in reproduction, with very limited success. The majority of baculum studies focus on correlating the gross anatomy of these bones to measures of mating ecology. This strategy ignores the influence of potentially relevant histological features of bones, like collagen fiber orientation, which is commonly used to infer bone function. In this study, we provide a novel approach to studying the function of bacula through a combination of controlled mating experiments with laboratory mice, followed by detailed shape and histological assessments of these bones. Specifically, we analyzed the overall shape and size of bacula along with collagen fiber orientation and bone deposition patterns over a 24-hour experiment (short-term) and a seven-week long mating experiment (long-term). No discernable patterns emerged in shape, or any histological feature between mated and unmated males in either experiment, but mated bacula were larger in the long-term experiment. Our data does supports clear differences in collagen fiber orientation between the proximal and distal regions and between the ventral and dorsal cortices of bacula. These results suggest that baculum histology may be developmentally constrained within a genotype, and could reflect an evolutionary adaptation to the mechanical strains involved with copulation. 67 4.2 Introduction The rapid evolution and remarkable morphological diversity of male genitalia is a widely recognized phenomenon (William G Eberhard 1985). In species with internal fertilization this trend is probably influenced by a combination of post-copulatory competition between males for paternity and the physical interactions between male and female genitalia (Clarissa M House et al. 2013). Some male genital features have evolved to influence paternity success through mechanisms like removing rival sperm (Hotzy et al. 2012) and stimulating female genitalia to bias sperm usage (Wulff et al. 2017). In addition to these specialized functions to gain a competitive advantage, male genitalia must also withstand the biomechanical requirements of copulation (Brennan 2016b). Specifically, intromitent organs must be adapted to resist species- specific compressive and bending forces associated with mating (W. W. Schultz et al. 1999) which may be accounted for by increasing the stiffness of genital tissues (D. A. Kelly 2016). Given the rampant diversity of genital structures, a number of strategies could be employed to reinforce genitalia in order to accommodate these forces, including modifications to boney or cartilaginous fins as seen in the intromitent organs in teleost fish (Turner 1950) (O’Shaughnessy et al. 2015), and potentially the mammalian penis bone, or baculum (W. H. Burt 1960b). Bacula are one of the most fascinating and commonly studied genital features in the animal kingdom (W. H. Burt 1960b; Patterson and Thaeler 1982). These bones evolve so rapidly that they are often used to distinguish closely related species (Patterson 1984), and in spite of decades of research, the function of these bones is still unknown. Many functional hypotheses have been suggested, including that bacula may stimulate females (AF Dixson 1995; Larivière and Ferguson 2002), provide structural support during long copulatory bouts (AF Dixson 1995; Larivière and Ferguson 2002; Romer and Parsons 1986), or signal male quality (Csanády et al. 68 2019; Miller and Burton 2001; D. E. Tasikas et al. 2009a). While there is evidence for certain functional hypotheses within certain taxonomic groups, patterns quickly dissolve when these hypotheses are applied across broader mammalian lineages (A. F. Dixson 1987a; Larivière and Ferguson 2002; Schulte ‐Hostedde et al. 2011; D. E. Tasikas et al. 2009a). For example, longer bacula may allow males to deposit sperm closer to the site of fertilization (G. Parker 1984) or stimulate females and affect female reproductive physiology to bias paternity (AF Dixson 1987b). Therefore, longer bacula may confer a reproductive advantage in mating systems with heightened post-copulatory competition (AF Dixson 1987b). However, when tested in a phylogenetic context, baculum length correlates with heightened sperm competition in rodents and carnivores, but not in bats and primates (S. A. Ramm 2007). The absence of a general functional hypothesis for bacula stems largely from our limited knowledge of the physical properties of these bones, and what forces are imposed on them during mating. Many studies survey baculum morphology in wild populations (Čanády 2013; Dyck et al. 2004), where variance in mating histories makes it almost impossible to accurately assess how the physical demands of copulation may influence bacula. Moreover, since mammalian copulation occurs internally any mechanical strains transferred to bacula during mating are unobservable; therefore researchers must rely on anatomical changes after mating as proxies of baculum function. A series of controlled mating experiments in mice and histological examinations in a number of other species have shed light on potentially critical morphological features of these bones. Recently, three independent studies investigated the effects of post-copulatory selection on mouse baculum morphology (André et al. 2018; Simmons and Firman 2014; Paula Stockley et al. 2013). One in particular examined the influence of perceived sperm competition on baculum morphology by introducing bedding of rival males (André et al. 2018). Over the course 69 of seven weeks, males exposed to rival bedding had significantly wider proximal bacula than control males, suggesting a relatively plastic response to environmental cues (André et al. 2018). Collectively, these studies hint at the potential influence of sexual selection shaping baculum morphology, particularly increasing proximal baculum width (André et al. 2018; Simmons and Firman 2014; Paula Stockley et al. 2013). Other studies have also identified interesting changes in the proximal baculum, which is the attachment site of the corpus cavernosum to the bone (Herdina 2008; Phillips et al. 2015). Artificial inflation of these erectile bodies in three bat species illustrated clear changes in baculum orientation at the proximal bone and along the dorsoventral axis (Herdina et al. 2015a), and a recent finite element analysis of carnivoran bacula indicated that species with longer intromissions evolved baculum morphologies that are more resistant to dorsoventral bending (Brassey et al. 2018). These two studies in particular, imply that there is a critical function of the proximal baculum and that these bones may be adapted in general to resist strains along the dorsoventral axis. However, these studies only capture a portion of the potential changes associated with mating, as bone shape is only one of many biomechanically relevant traits. The rigid nature of skeletal structures can potentially lead one to believe that bones are static tissues, which minimally respond to outside influences. However, in order to maintain their structural integrity and to support a number of complex movements (Riggs et al. 1993) bones are required to be sensitive to changes in their mechanical environment (Doherty et al. 2015). Forces experienced by bones can vary in their magnitude and directionality, which can be accounted for through a combination of altering bone shape and microstructural features (Pearson and Lieberman 2004). For instance, experimentally changing the locomotor patterns of mice from primarily walking along a linear path to restricting movement to a series of repeated turns 70 resulted in the redistribution of bone mass in limbs (Carlson and Judex 2007). Mice in the non- linear experiment shifted bone mass to the medial and lateral aspects of femora resulting in a distinct elliptical mid-shaft shape compared to linear walking animals (Carlson and Judex 2007). Experimental manipulations of locomotor patterns have also resulted in histological changes in bones. Specifically, a study in lambs illustrated that surgically altering the amount of rotational forces experienced during locomotion resulted in increased bone deposition along the lateral bone cortex and a dramatic shift in collagen fiber orientation (Volpon et al. 2014). While the baculum does not function in locomotion, its shape and microstructure may be plastic to the biomechanical influence of mating, and thus assessing those features may reveal new insights into the function of these bones. To date, no study has systematically tested if baculum histology changes as a function of mating behavior. To explore the effects of copulation related mechanical loading on mouse bacula we conducted two main experiments. First, we investigated the immediate response to mating through a 24-hour mating experiment. We then followed with a longer term mating experiment where males mated with different females continuously for ~7 weeks. By analyzing bone tissue distribution and collagen fiber orientation through the baculum, we identified clear regional differences in baculum histology that lend insight into the functional stresses endured by the baculum. Outside of the increased size of mated bacula in the long-term experiment, no other differences were observed between the bacula of mated and unmated males. We discuss potential reasons for this lack of difference, including developmental influences, and levels of plasticity and its interaction with genotype. 71 4.3 Materials and Methods 4.3.1 Experimental Design To assess the responsiveness of the baculum to any mechanical stresses imposed by mating, we conducted a short-term and long-term experiment. The short-term experiment lasted 24 hours and allowed us to evaluate how sensitive bacula are to the forces of a single mating event. The long-term experiment was designed to understand the effects of habitual loading on the baculum as a result of chronic mating. All experimental methods were conducted according to University of Southern Institutional Animal Care and Use Committee, protocol #11394. In the short term mating experiment males were weaned between 21-28 days, and housed alone in cages until they were 9-10 weeks old. At that point, males were randomly assigned to the “mated” (N=9) or “unmated” (N=6) groups, and were either paired with a female or remained singly housed. To increase the probability of mating over such a short period of time, estrus was induced in females via an intraperitoneal (IP) injection of 5U pregnant mare serum 48 hours before, and an injection of 5U human chorionic gonadotropin immediately before being paired with a male (Nagy et al. 2003). Successful mating was scored either by the presence of a copulatory plug, by clear indications of pregnancy, or newborn pups in the weeks following the experiment. Mating pairs were separated after 24 hours or when a copulatory plug was observed in female genitalia, whichever came first. Three of the nine males in the “mated” group did not successfully mate according to any of the criteria listed above, but were still included in the analysis as successful mating may still occur even if a copulatory plug is not deposited (Mangels et al. 2016). Bone regions experiencing heightened local stress will actively deposit more mineralized tissue (Ducy et al. 2000). This change in bone deposition can be visualized using the green 72 fluorochrome calcein (Lanyon et al. 1982), with binds to calcium ions as they are integrated into newly formed bone (Du et al. 2001). This is commonly accomplished through double-injections where the distance between two calcein bands represents the amount of bone deposited over a given time (Erben and Glösmann 2012). Here we use calcein double-injections to detect any differences in bone deposition patterns as a result of mating in the short-term experiment. Briefly, all males received a 200uL injection of a 6% calcein solution (mated males: N= 9, mean weight= 26.06 ± 0.64; unmated males: N= 6, mean weight = 25.72 ± 0.68) 13 and 3 days before euthanasia. The first injection occurred immediately before males were paired with a female. Unmated control males were injected at the same time but instead of being paired with a female, they continued to be housed alone. In this long term mating experiment, each group was comprised of three sibling males, referred to as “sib-groups” from here on out. Two males from each sib-group were randomly assigned as unmated controls, and the third male was designated for the mating experiment. In total, 7 sib-groups were included in the long-term experiment (mated males = 7; unmated males = 14). All males were weaned between 21-28 days old, and placed in separate cages at 28 days. The two control males were housed together for the duration of the experiment, while the mated male was immediately housed with his sister. Each week the mated male received a new female to mate with for the span of 7 days, until the end of the experiment when males were between 11-13 weeks old. Overall, mated males were paired with 7 females and the same criteria were used to confirm successful mating as in the short-term experiment. On average, males successfully mated with 4 (± 0.31) females over the duration of the long-term mating experiment. 73 At the conclusion of both experiments, bacula were removed following previous methods (N. G. Schultz et al. 2017). To account for any influence of body size or testosterone we took standard body measurements and the weights of male accessory organs (seminal vesicles, preputial glands, and testes), which are known to be highly sensitive to testosterone (Bartke 1974; Brain et al. 1983). 4.3.2 Micro-CT Scanning of bacula Bacula were microCT scanned following methods outlined in (N. G. Schultz et al. 2017). Samples were micro-CT scanned at the USC Molecular Imaging Center using a uCT50 scanner (Scanco Medical AG, Bruttisellen, Switzerland) under the following settings: 90 kVp, 155 uA, 0.5 mm Al filter, 750 projections per 180 (360 coverage), exposure time of 500 msec, and voxel size of 15.5 mm. 4.3.3 3D-Transformation and Defining semi-landmarks Resulting microCT images of bacula were segmented and converted to a 3D point cloud consisting of x-y-z points using custom Python (www.python.org) and R (R Core Team 2014) scripts. Every pixel in the point cloud is interpreted as bone, including the surface and internal structures. Point clouds were transformed following the methods (N. G. Schultz et al. 2016b) (and Ch.2 of this dissertation), with a few modifications. Unlike in Ch. 2, the bacula analyzed in this experiment were very similar to each other, since they all derived from the same genotype. Preliminary analyses suggested that placing too many semi-landmarks on highly similar bones inflated false positive rate. Therefore, we reduced the number of semi-landmarks used in this 74 chapter. The number of point slices, lines dividing slices, and the number of semilandmarks per slice were altered. Specifically, instead of using fifty point slices as previously described, here we analyzed fifteen slices. Each slice was defined by 12 semilandmarks instead of 16. In addition to these points, the anterior-most and posterior-most points were included bringing the total number of semilandmarks used to define shape and size to 182. 4.3.4 Quantifying shape variation Shape differences were quantified between all possible pairs of bacula using a Generalized Procrustes framework. This process involved standardizing each set of 182 semilandmarks to a common size, translating them to a common origin, followed by rotating semilandmark coordinates to minimize their Procrustes distance (Rohlf and Bookstein 1990; Slice 2007). Following previous methods in Ch. 2 and (N. G. Schultz et al. 2016b), semilandmarks were allowed to “slide” along bones during Procrustes superimposition using the GPAGEN function in the R package GEOMOPRH (D. C. Adams and Otárola ‐Castillo 2013). The GPAGEN analysis resulted in a pairwise distance matrix for the long-term (21·21 matrix, 7 sib-groups of 3 males each) and short-term (15·15 matrix, 6 unmated and 9 mated males) experiments respectively. Previously, in (N. G. Schultz et al. 2016b) we used a Linear Discriminate Analysis to analyze differences in shape and size between bacula. This approach worked well for that genetic system since the parental strains of the recombinant inbred line panel (BXD RILs) were so divergent in shape. When applied to the current study, the same strategy failed. Again, this was likely because the bones presently being analyzed are too similar in shape. As an alternative, we employed a multidimensional scaling approach using the CMDSCALE function in R (R Core Team 75 2014). Unlike LDA, multidimensional scaling (MDS) is blind to group identity and simply yields a visual representation of similarities between points (Sturrock and Rocha 2000), in our case overall bone shape. MDS is similar to Principal Components Analysis in that pairwise shape differences are projected into a space that maximizes the variance among bacula (Kruskal 1964). A pairwise Euclidean distance matrix between bones was calculated and used as input for the MDS, separately for the short-term and long-term experiments. After projection into the MDS space, we calculated the mean distance between mated and unmated bacula, then permuted mating status to generate a null distribution. For the short-term experiment, we analyzed 9 mated and 6 unmated bacula, and we generated all choose(9, 15)=5005 ways to permute those assignments. For the long-term experiment, we only permuted mating status within a sib-group. There were 7 sib-groups, each with two ways to permute mating status, leading to 7 2 = 128 possible permutations. 4.3.5 Histology Preparation To gain a better sense of the abundance and distribution of relevant bone microstructural features, primarily collagen fiber orientation, histologic slides of undecalcified, ground bone were prepared. Following microCT scanning, bones were fixed in 10% paraformaldehyde, dehydrated through a series of graded ethanol (30%, 50%, 70%, 100%) and cleared using xylenes (Comelis et al. 2015; Lee and Simons 2015). After histological prep, bones were left to dry overnight, and then embedded in an epoxy resin mixture at a weight ratio of two parts epoxy resin (AeroMarine 300) to one part hardener (AeroMarine 21). All samples were oriented in the same anatomical position, and resin blocks with bones were placed in a vacuum chamber to 76 remove air bubbles. Samples were left to dry in a desiccator chamber for 24 hours before being cut. Each block containing a bone sample was cut using a Buehler IsoMet low speed saw (Buehler, Lake Bluff, Illinois) and a Buehler Isomet 15HC diamond wafering blade. Bacula were cut from the proximal end to the distal end of the bone in the smallest possible increments yielding ~4 - 6 cross-sectional slides per sample. Wafers from each cut were gently soaked in absolute ethanol for 10 seconds, and then dried over night before being mounted on glass slides. All wafers were mounted in a consistent orientation, with the distal side of the bone facing toward the slide. After mounting, each specimen was coarsely ground to ~150µm using a Buehler MetaServ 250 tabletop grinder-polisher, and then finely hand polished using silicon carbide paper until each slide was within the standard thickness range of 100µm ± 5µm (Bromage et al. 2003), as determined by a handheld micrometer (Mitutoyo IP65). Since cuts between bones do not perfectly correspond to one another, each histological slide was placed into one of three categories based on known mouse baculum anatomy (Rodriguez et al. 2011). A proximal region was characterized by an extensive medullary cavity. A mid-region was characterized by the tapering of the bone width and significant reduction in the medullary cavity, and thicker cortices. A distal region was characterized by a minimal medullary cavity and an overall more circular shape of bone. These regions were then labeled A, B, and C from proximal to distal. Examples of these regions are highlighted in Figure 4.1. Circular polarized light (CPL) images were obtained at 10X magnification using a Leica DM 2700P (Leica Microsystems, Baunnockburn, IL) microscope equipped with two quarter wave plates, one above and the other below the sample (Bromage et al. 2003). Quarter wave plates were oriented to maximize background extinction, and those orientations and all other 77 microscope conditions were held constant throughout imaging. As a control, a single bone sample was imaged each day that experimental images were taken to assess repeatability. Images were obtained using an HD digital camera (Leica MC 170 HD), and saved as TIFF files using Leica Application Suite X software under standardized imaging conditions (exposure time= 9.8ms, gain=3). The exact same slides used for CPL imaging, were also used for calcein imaging for the short-term mating experiment. Calcein slides were imaged under blue light (~488nm) emitted from a ZEISS Axio Zoom.V16 microscope (ZEISS Manufacturing company) with an AxioCam MR3 fluorescent camera (ZEISS Manufacturing company). Images were taken under standardized conditions (100x magnification, 4.8ms exposure time) and saved as TIFF files in the Zeiss Zen Pro 2012 software. 4.3.6 Polarized Light Analysis Previous research indicated dramatic orientation changes along the dorsoventral axis of penis bones during artificial erection (Herdina et al. 2015a), implying that there may be different mechanical influences on these cortices. Therefore, we decided to analyze collagen fiber orientation differences between dorsal and ventral cortices using polarized light microscopy. Collagen fiber orientation can be visualized under polarized light, since the light will be refracted to certain degrees depending on the alignment of collagen fibers (Bromage et al. 2003). Specifically, transverse fibers relative to the light source appear bright and longitudinal fibers appear darker when imaged (Bromage et al. 2003; Martin and Ishida 1989). CPL images were analyzed in Photoshop (Adobe Inc.) by first converting images from TIFF files to 8-bit grayscale images. These files have a grayscale pixel value range from 0-255, where 0 represents the 78 darkest pixels and 255 is bright white (Goldman et al. 2003). Dorsal and ventral cortices were independently segmented resulting in distinct histograms of pixel values in grayscale. To detect any regional differences in the amount of bright pixels, we calculated two metrics: 1) the overall proportion of bright pixels (grayscale values between 209-255) (Goldman et al. 2003) in each region of the baculum, and 2) the proportion of bright pixels that were on the ventral versus dorsal side of the bone. We tested for differences due to mating status and baculum region using a two-way ANOVA, with bone region and mating status as factors and the non-parametric equivalent of the Friedman Test. 4.3.7 Calcein Quantification Resulting calcein images were converted to text files using custom R (R Core Team 2014) scripts. TIFF images were converted to a RGB scale using the RGB function in the package TIFF (Urbanek 2013), where red, green and blue channels each range from 0 to 1 (instead of the normal 0 to 255). We calculated the proportion of green pixels on the ventral and dorsal half of each baculum; as such asymmetry could provide insights into differential growth patterns and stresses endured by bacula. We again tested for differences due to mating status and baculum region using a two-way ANOVA and the non-parametric equivalent of the Friedman Test. 4.4 Results 4.4.1 Histological features vary within a single slice and between regions of the bone Our study provides the most detailed histological insights into the mouse baculum to date. Using both polarized light and calcein slides, we have inferred general patterns of bone growth and physiology in these bones. The key histological features mentioned here correspond 79 to the lower panels of Figure 4.1. The first crucial observation is the unequal distribution of bone types both within a given Region and between different regions of the bone. There are two discernible bone types in these slides, woven bone (wb) and lamellar bone (lb). Woven bone has is more disorganized in structure, while lamellar bone is highly organized into discrete layers. In Region A, there is a clear transition from lateral woven bone, to a highly organized lamellar bone structure in the middle of the bones. This is exemplified in the polarized light image, where clear layers of bright bands are observed on the ventral side of the bone (Figure 4.1: red arrow). However, this lamellar patterning is not commonly represented in the dorsal cortex of Region A bones, which still have bright transverse collagen fibers, but the overall organization of the dorsal bone is much less structured. The ventral lamellar bone in Region A is primarily constructed of bright transverse collagen fibers. The lateral aspects of the bone have a higher density of osteocyte lacunae, which is typical of woven bone and may reflect the metabolic capacity of that tissue (Hernandez et al. 2004). The lamellar pattern of bone in the ventral cortex is recapitulated in the calcein banding pattern at the same region. Two distinct green bands appear on the ventral cortex of Region A bone (Figure 4.1: yellow arrow) and this pattern abruptly stops when moving laterally along the bone in cross-section. Furthermore, the proximal end of the baculum appears to be highly developmentally active as indicated by the abundance of calcein signal in Region A. In Region B, the lamellar bone structure is maintained primarily around the medullary cavity (MC), and woven bone persists around the entire outer layer of bone. There is overall a higher concentration of bright collagen fibers in this section, although they seem to be equally distributed between ventral and dorsal cortices around the medullary cavity. There is a clear thickening of bone cortices in Region B compared to the relatively thin cortex typical of Region 80 A bone sections. There is almost no calcein signal in Region B outside slight the deposition around the medullary cavity. This indicates that the mid-section of bacula may primarily grow via endosteal bone deposition, from the inside out. Region C is the most uniform bone section in terms of shape. Generally, this region of the baculum is more round in shape, and the dorsal and ventral extremes of the bone have woven bone and only the very center of the bone around the medullary cavity still maintains a lamellar bone structure. There are very few transverse collagen fibers in this bone region, with the exception of the occasional brighter ring around the medullary cavity. In addition to representing the most uniform overall bone shape, Region C bones have the least evidence for bone deposition as indicated by the calcein slide in the bottom panel. Also of note, is the amount of soft-tissue around the baculum in polarized light slides. Bacula occupy only a small fraction of penis cross-sectional anatomy, which is most apparent in the Region C slides. This anatomical observation may be critical to understanding the biomechanics of mouse bacula, discussed further in the Discussion. 81 Figure 4.1 CPL and Calcein Histology The top section of the figure represents the point cloud renderings of microCT scanned bacula in dorsal (top) and lateral (bottom) views respectively. The black lines over these images represent the approximate sections that histological slides were cut in, representing regions A, B, and C. Underneath are the cross sections of CPL slides (top panel) and calcein slides (bottom panel). Each of these slides is oriented in the same direction, where the top of the slide represents the dorsal bone, and the bottom aspect of the slide is the ventral part of the bone. Bacula are embedded in the soft tissue anatomy of male genitalia in these images, and arrows indicate key trends described in the text. Histology images are not to scale as represented here. (White lettering MC = medullary cavity, lb = lamellar bone, wb= woven bone). 82 4.4.2 Shape did not vary by mating status, but mated bacula are larger than unmated bacula Baculum shape did not differ between mated and unmated males, in either the short-term or long-term experiment. In the short-term experiment, mean Euclidean distance between mated and unmated bacula was greater than 2746 of 5005 possible permutations, translating to p=0.45. In the long-term experiment, the mean Euclidean distance between mated vs. unmated bacula was greater than 94 of 128 permutations (p=0.26). Relationships between mating status and baculum size were also assessed based on microCT data. Among multiple tests, which are summarized in Supplemental tables 4.1 and 4.2, only one measure of baculum size correlated with mating status in the long-term experiment, while no significant relationships were detected in the short-term experiment. In the long-term experiment, mated males had a higher mean number of pixels compared to unmated males (mated = 159570.75, unmated = 149511.31; t=2.35, df=9.75, p= 0.04), but centroid size was not significantly different. Given the small effect size and multiple comparisons made, this result should be taken with caution, but suggest that mated males have larger bacula. 4.4.3 Baculum growth did not differ between mated and unmated males Calcein deposition was calculated by quantifying the percent of green pixels in the ventral cortex of a bone for a given slide. The percent of green pixels in the ventral bone does not change by region (F 2,36 =1.37, P=0.27) or mating status (F 1,36 =0.03, P=0.87), and there was no interaction detected between region and mating status. 83 4.4.4 Collagen fiber orientation varies by bone region in both experiments, and between cortices in the short-term experiment In the short-term experiment the proportion of bright pixels varied by the region of the bone (F 2,29 =6.98, P=0.003) but not based on mating status (F 1,29 =0.014, P=0.9) (Table 4.1). The proportion of bright pixels in the ventral cortex of the bone varied by region (F 2,29 =5.88, P=0.007) but again, not by mating status (F 1,29 =0.050, P=0.82) (Table 4.2). In the long-term experiment the proportion of bright pixels did not vary by either region of the bone (F 2,38 =2.26, P=0.11) or mating status (F 1,38 =1.09, P=0.30). However, the proportion of bright pixels in the ventral cortex varied along the bone by region (F 2,38 =4.45, P=0.018) and not based on mating status (F 1,38 =1.97, P=0.16) (Table 4.3). Table 4.1 Short-term Proportion of Bright Pixels (CPL) Df SumSq MeanSq F value Pr(>F) Region 2 0.33331 0.166657 6.98 0.003 ** Mated 1 0.00035 0.000348 0.01 0.9 Residuals 29 0.69213 0.023867 Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Table 4.2 Short-term Proportion of Bright Pixels in Ventral Cortex (CPL) Df SumSq MeanSq F value Pr(>F) Region 2 0.09 0.049 5.88 0.007 ** Mated 1 0.0004 0.0004 0.05 0.82 Residuals 29 0.24 0.008 Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Table 4.3 Long-term Proportion of Bright Pixels in Ventral Cortex (CPL) Df SumSq MeanSq F value Pr(>F) Region 2 0.073 0.036 4.45 0.018 * Mated 1 0.016 0.016 1.97 0.16 Residuals 38 0.309 0.008 Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 84 4.5 Discussion The overarching goal of this experiment was to address any potential physiological effects of bacula in response to mating. Of particular interest was assessing if bacula seem to be responsive to mating mechanics, and if so, if those changes alter shape, bone growth, or microstructural features. This approach was novel, and could potentially lead to expansive knowledge about the role of these bones in sexual selection. With the exception of increased bone size, no features measured in these experiments corresponded to mating status. A thorough explanation of potential influences need to be discussed, including limitations of the current study, insights into the dynamic relationship of bacula in the context of penis soft-tissue anatomy, and an understanding of the regulation of bone microstructural properties. 4.5.1 Experimental Design It is quite possible that we fail to see a strong effect of mating on baculum physiology here because of experimental design issues. The goal of our long-term mating experiment was to gauge the effects of repeated mating on baculum morphology. It is possible that the average of four successful matings per male was insufficient to elicit a strong response in bacula, and pairing a male with a novel female more frequently than every week could have increased mechanical stresses. This is especially important given that it is typical for males to only mate once when paired with a female over that period of time, as the first mating induces physiological effects in the female to resist remating for between 10-12 days (Mcgill et al. 1968). Therefore the amount of mating over a 7-day period with the same female may have been equivalent to one mating event, and thus did not adequately amplify the effects of mating as 85 intended. An alternative approach would have been to introduce novel females every 2-3 days, which according to previous reports should be enough time for a male to mate once with each female (Mcgill et al. 1968). While this approach may increase the effect of mating on baculum development, our results suggest that effect may be minimal, and thus does not justify increasing the number of animals to be used in such an experiment. Our original objective was to control genetic influences on mouse baculum development by focusing on a single genotype, 6J (C57BL/6J). However, if this genotype is not plastic, we may have inadvertently limited our ability to detect responses to mating. Various bone phenotypes, including overall stiffness and the ability to withstand external forces as determined through mechanical test, vary across mouse genotypes (Beamer et al. 1999; Jepsen et al. 2015). For example, Carlson (Carlson and Judex 2007) detected a plastic response in BALB/cByJ mouse limb bones as a result enforced non-linear movement, most specifically through changes in cross-sectional geometry. Other experiments indicate that genotype is the most accurate predictor of measured bone strength, and more specifically, exercise treatment did not affect bone strength or shape in 6J mice (Peacock et al. 2018). It seems possible that different mouse genotypes respond to bone loading regimes to varying degrees, and our 6J genotype may not be as receptive to those changes. 4.5.2 Baculum Size and Mating We did however detect a difference in bone size between mated and unmated males in the long-term mating experiment as determined by the number of bone pixels resulting from microCT scanning. The number of bone pixels is a fundamentally different measurement than centroid size, which is sensitive to the number of semilandmarks imposed through our geometric 86 morphometrics. Based on the distribution of semilandmarks, it is possible that areas of the bone that differ in size are not described by the overall shape, and thus are not included in the centroid size measurement (F. L. Bookstein 1989, 1996). Given that bones respond locally to loading patterns (Burr et al. 2002), it is possible that the location of landmarks imposed on bacula did not include the areas where mating related loading influenced baculum physiology, and our signal of bone size differences in the long-term experiment is heavily influenced by those regions. Another explanation for this trend is that our microCT analysis picks up on both external and internal bone pixels (N. G. Schultz et al. 2017). It is possible that internal shape changes, including changes in cortical thickness, may be driving the noted difference in the overall number of bone pixels between mated and unmated males. Increased cortical thickness is a commonly observed adaptation to mechanical strain in limb bones (Montañez ‐Rivera et al. 2018; Montoya ‐Sanhueza and Chinsamy 2016; Pflanz et al. 2017), and this may be an adaptation to mating in mouse bacula. We measured thickness (Table S4.1) at 25, 50, and 75% the length of bacula, and did not detect a difference between mated an unmated males. However, it is possible that our thickness measurements did not capture the exact regions that vary based on mating status, and a more comprehensive survey along the bone may provide further insight. 4.5.3 Trends in Baculum Histology While this study did not uncover a robust influence of mating history on baculum morphology, several important histological insights can be used to generate hypotheses about how these bones function. The most prominent histological changes occurred in the proximal baculum, primarily Region A, which is the attachment site for the corpus cavernosum in all species with bacula (N. G. Schultz et al. 2016a). Accordingly, the proximal baculum is 87 consistently noted as the widest part of the bone across bat (Herdina et al. 2010; Yoon et al. 1990), rodent (Layne 1960; Simson et al. 1993), and carnivore species (W. H. Burt 1960b; Miller and Burton 2001), and in outbred mice, the proximal baculum was wider among males that sired more offspring under competitive conditions (Simmons and Firman 2014; Paula Stockley et al. 2013). One possible explanation is that the corpus cavernosum transfers substantial forces to the proximal baculum during erections, which could stimulate increased regional bone deposition (M. Dean and Shahar 2012). Via calcein injections, our analyses indicated that calcium was more actively incorporated at the proximal end of the baculum. Yet, we did not observe a difference between mated and unmated males, so any pattern in the proximal bone is not likely due to mating related effects. Bone deposition patterns also reflect developmental trajectories, not just responses to mechanical stimulus (Lieberman et al. 2003). Therefore, it is possible that the regional differences in calcein signal in bacula reflect developmental variation along the bone. Mouse bacula develop through a combination of endochondral and intramembranous ossification (Murakami 1987), similar to limb bones (Carter et al. 1996). Endochondral bone growth allows for the rapid lengthening of skeletal structures (Kronenberg 2003), and is observed in the proximal baculum (Glucksmann et al. 1976) where we detected a strong calcein signal. In contrast, the middle and distal regions of the mouse baculum develop directly from genital mesenchyme through intramembranous ossification (Glucksmann et al. 1976), a process that is initiated prior to endochondral development in mice, and does not continue for nearly as long (Glucksmann et al. 1976). Based on age, the bacula included in our experiments were likely actively growing via endochondral bone deposition (Glucksmann et al. 1976), which may have hindered our ability to detect bone growth differences in response to mating. In the future, it may 88 be wise to test mating related responses in older males after endochondral baculum growth has ceased. Within Region A, the ventral and dorsal regions clearly differ in a number of important phenotypes; most notably the line pattern of calcein deposition. A very organized, two-banded deposition of calcein extended through most of the ventral side of Region A bone, compared to the relatively disorganized calcein incorporation on the dorsal side. This difference correlated with higher concentrations of lamellar bone in the ventral cortex, compared to higher proportions of woven bone in the dorsal and lateral sides. These bone type differences have both developmental and structural relevance. Woven bone can be laid down initially to accommodate rapid growth rates at the detriment of mechanical integrity, and is often replaced by the stronger lamellar bone when growth rates decrease (Currey 2003; McFarlin et al. 2016; Riggs et al. 1993). The uniform organization of lamellar bone enhances structural support compared to the haphazard organization of woven bone (Currey 2003), suggesting that baculum strength on the ventral side is important. Additionally, collagen fiber orientation commonly corresponded with bone type in our study. Regions of the baculum with more lamellar bone tended to have more transverse collagen fibers, while regions of woven bone had more longitudinally oriented collagen. The most notable regions that deviated from this pattern were the lateral and dorsal aspects of Region A, which while mostly comprised of woven bone, still had an abundance of transverse fibers. These regional variations in collagen fiber orientation may provide valuable insight into the biomechanics of mouse bacula. Both mechanical stress tests and observational data illustrate that collagen fiber orientation varies in a predictable manner based on local mechanical needs (Carando et al. 1991; Kalmey and Lovejoy 2002; Reilly and Currey 1999). Bone regions known to be loaded primarily in compression have predominately transverse 89 collagen fibers and regions loaded in tension have more longitudinal collagen fibers when viewed in cross-section (Carando et al. 1991; Kalmey and Lovejoy 2002; J. G. Skedros and Hunt 2004). This suggests that the proximal baculum may experience more compressive loading than the mid or distal regions of the bone. Again, it is important to emphasize that these regional differences were not the result of mating status in our experiment. Microstructural features like collagen fiber orientation and the distribution of woven versus lamellar bone types are commonly used by paleontologists to infer the locomotor patterns, taxonomic relationships, and life history traits of organisms (Houssaye 2013; Kolb et al. 2015; Montoya ‐Sanhueza and Chinsamy 2016). Collagen fiber orientation can be inherited and thus developmentally constrained (Maggiano et al. 2015; Montoya ‐Sanhueza and Chinsamy 2016) in addition to reflecting the experienced loading regime of a bone during an animals’ lifetime (Currey 2003; J. Skedros et al. 1996). Our study did not indicate a change in collagen fiber orientation as a result of mating experience. Instead, the proximal ends of all bacula showed more transverse fibers on the ventral side of the bone compared to the dorsal side. This pattern suggests that bacula might develop to resist bending forces that generate compression ventrally and tension dorsally. Part of this pattern may be explained by the attachment of the corpus cavernosum to the proximal baculum. The corpus cavernosum transmits forces to the proximal baculum during an erection (Herdina et al. 2015a) however; the connection of this erectile tissue may not be equally distributed along the proximal baculum. For instance, the corpus cavernosum in all three bat species tested by Herdina (Herdina et al. 2015a) attached to more of the dorsal baculum surface than the ventral baculum surface. Subsequent artificial inflation caused the baculum to straighten by pulling the dorsal cortex of the proximal bone down, and seemingly pushing the ventral cortex forward toward the distal tip of the bone (Herdina et al. 2015a). These 90 changes in orientation may create tension along the dorsal cortex and compression along the ventral cortex of the proximal bat baculum (Herdina et al. 2015a), a mechanical change that coincides with the collagen fiber orientation patterns observed in our data. Nevertheless, this relationship is speculative, as the artificial inflation experiments have not been replicated in mice. Additional evidence for baculum resistance to dorsoventral bending comes from recent theoretical material testing. Based on theoretical models, the carnivore baculum appears to resist similar dorsoventral bending, perhaps as a mechanism to protect the urethra during longer copulatory bouts (Brassey et al. 2018; AF Dixson 1995). Repeating these theoretical experiments in mice is more difficult since they have an irregular teardrop shape and thus cannot be modeled like a simple cantilevered beam for finite element analysis as in carnivores (Brassey et al. 2018). However, if the material properties of mouse bacula were known, then it is possible that physical models can be developed and tested. Three-point bending tests have been performed on small bones like mouse metatarsals (Schriefer et al. 2005), so future experiments could incorporate such data with bacula. Finally, understanding the function of bacula requires a deeper understanding of the interactions between bacula and surrounding soft-tissue structures. The proportion of the penis occupied by the baculum will have a large influence on the types of forces that it endures during copulation. Except for felids, carnivore bacula tend to occupy most of the volume of the penis (W. H. Burt 1960b; Sharir et al. 2011), much more than in rodents (Hooper 1961; Murakami and Mizuno 1986). Therefore, the robust and long carnivore baculum may add to penile rigidity compared to a much more reduced structure as seen in rodents (D. Kelly 2000). 91 4.6 Conclusion This study characterized the microstructural properties of the baculum, a bone that exhibits astonishing levels of divergence among species. Our goal was to incorporate methods in bone physiology to generate hypotheses of mechanical stresses that a baculum endures during mating. Although we did not detect strong differences in bacula of mated versus unmated males, our data provides the most comprehensive view of the microstructural details of mouse bacula to date. The patterns of bone microstructure observed here provide valuable insights into the potential biomechanical role of the mouse baculum, and lays the groundwork for future functional experiments. As live imaging technology continues to improve, direct imaging of genitalia in naturally copulating individuals might yield important insights into the function of this bone. 92 Exp. Phenotype T.test T T.tes tP Wilcox P Mean mated Mean unmated Median mated Median unmated Short- term index_thickest_sli ce -0.35 0.735 0.721 17.778 18.167 18.000 19.000 Short- term dorsalthickness_0 .25 1.08 0.302 0.607 6.369 5.782 6.304 6.056 Short- term ventralthickness_ 0.25 -1.52 0.152 0.388 3.549 4.310 3.720 4.261 Short- term medullarydist_0.5 -0.08 0.936 0.776 16.702 16.776 16.448 17.152 Short- term outermostdist_0.5 -0.2 0.843 0.328 26.619 26.868 26.104 27.673 Short- term dorsalthickness_0 .75 0.61 0.553 0.607 5.099 4.729 5.042 4.401 Short- term ventralthickness_ 0.75 -0.55 0.602 0.955 3.786 4.044 3.720 3.812 Short- term medullarydist_0.2 5 -1.28 0.224 0.328 8.489 9.716 8.050 9.223 Short- term outermostdist_0.2 5 -1.63 0.129 0.181 17.374 18.490 17.502 19.018 Short- term dorsalthickness_0 .5 0.6 0.568 0.529 6.969 6.497 7.131 5.759 Short- term ventralthickness_ 0.5 -1.89 0.090 0.114 4.293 5.123 4.190 5.139 Short- term medullarydist_0.7 5 -0.58 0.581 0.955 15.274 15.903 15.211 15.500 Short- term outermostdist_0.7 5 -1.41 0.193 0.224 26.535 27.523 26.373 27.467 Short- term max_height -1.32 0.213 0.224 34.004 35.310 34.051 35.403 Short- term max_width -1.37 0.195 0.328 95.079 97.309 94.482 97.095 Short- term ratio_maxheight_ maxwidth -0.51 0.619 0.689 0.358 0.363 0.352 0.366 Short- term ratio_midheight_ maxwidth 0.37 0.722 0.776 0.280 0.276 0.286 0.284 Short- term dorsal_ventral_as ymmetry -2.74 0.021 0.012 0.387 0.443 0.392 0.428 Long- term index_thickest_sli ce -0.3 0.766 0.631 18.000 18.188 18.000 18.000 Long- term dorsalthickness_0 .25 1.13 0.278 0.569 7.060 6.523 6.797 6.777 Long- term ventralthickness_ 0.25 0.42 0.684 0.383 4.848 4.664 5.264 4.886 Long- term medullarydist_0.5 -0.02 0.981 0.742 14.617 14.638 14.275 14.969 Long- term outermostdist_0.5 1.01 0.331 0.350 26.524 25.825 26.191 25.640 Long- term dorsalthickness_0 .75 0.42 0.677 0.452 5.993 5.780 5.980 5.513 Long- term ventralthickness_ 0.75 -0.15 0.879 0.697 4.969 5.024 5.028 5.151 Long- term medullarydist_0.2 5 -1.71 0.102 0.192 5.987 6.812 5.886 7.100 93 Long- term outermostdist_0.2 5 -1.03 0.321 0.569 16.949 17.615 17.228 17.257 Long- term dorsalthickness_0 .5 1.33 0.207 0.320 7.814 7.226 7.842 7.051 Long- term ventralthickness_ 0.5 -0.13 0.898 0.928 5.238 5.298 5.212 5.171 Long- term medullarydist_0.7 5 -0.92 0.375 0.291 12.962 13.908 12.129 14.061 Long- term outermostdist_0.7 5 -0.41 0.693 0.881 26.014 26.432 26.705 26.153 Long- term max_height 0.07 0.945 0.976 33.980 33.911 33.704 33.341 Long- term max_width 0 0.998 0.788 91.528 91.533 89.683 91.043 Long- term ratio_maxheight_ maxwidth 0.07 0.949 0.976 0.371 0.371 0.368 0.370 Long- term ratio_midheight_ maxwidth 0.92 0.373 0.417 0.290 0.283 0.287 0.284 Long- term dorsal_ventral_as ymmetry -0.8 0.441 0.742 0.419 0.434 0.434 0.433 Supplementary Table 4.1 List of baculum shape and size phenotypes based on microCT data. Starting with the short-term experiment, each row represents a test of the identified variable followed by the T.test T value, T.test P value, Wilcox P value, mean mated, mean unmated, median mated, and median unmated values. 94 Exp. Phenotype T.test P Wilcox P Mean Mated Mean Unmated Median Mated Med Un- mated Short-term residuals_centroid_siz e_by_BodyWeight_g 0.07 0.06 7.28 22.95 2.84 25.27 Short-term residuals_centroid_siz e_by_bodylength_min us_tail 0.37 0.49 15.70 27.57 8.37 22.23 Short-term residuals_n_bone_pixe ls_by_BodyWeight_g 0.28 0.35 -6427.06 4155.99 -5807.25 2529.67 Short-term residuals_n_bone_pixe ls_by_bodylength_min us_tail 0.39 0.49 -4491.96 5341.92 -5443.45 2053.72 Short-term residuals_density_by_ BodyWeight_g 0.38 0.41 -7.48 0.29 -8.00 -0.78 Short-term residuals_density_by_ bodylength_minus_tail 0.42 0.49 -6.87 0.76 -8.05 -0.63 Short-term residuals_Testes_mas s_g_by_BodyWeight_g 0.78 0.66 0.00 0.00 0.00 0.01 Short-term residuals_Testes_mas s_g_by_bodylength_mi nus_tail 0.30 0.41 0.00 0.01 0.00 0.01 Short-term residuals_SV_mass_g _by_BodyWeight_g 0.78 1.00 0.00 0.01 0.00 0.00 Short-term residuals_SV_mass_g _by_bodylength_minus _tail 0.98 0.94 0.01 0.01 0.01 0.01 Short-term residuals_PG_mass_g _by_BodyWeight_g 0.43 0.57 0.00 -0.01 0.00 0.00 Short-term residuals_PG_mass_g _by_bodylength_minus _tail 0.53 0.49 0.00 0.00 0.01 0.00 Short-term centroid_size 0.28 0.39 1004.51 1019.78 1001.19 1013.83 Short-term n_bone_pixels 0.32 0.33 144982.8 9 156640.0 0 145189.0 0 152415.00 Short-term SV_mass_g 0.86 0.83 0.19 0.19 0.19 0.19 Short-term Testes_mass_g 0.92 0.75 0.18 0.18 0.18 0.19 Short-term BodyWeight_g 0.66 0.48 26.15 25.72 25.68 24.89 Short-term bodylength_minus_tail 0.32 0.30 97.50 93.83 97.50 95.00 Long-term residuals_centroid_siz e_by_BodyWeight_g 0.24 0.25 2.58 -15.29 4.19 -16.17 Long-term residuals_centroid_siz e_by_bodylength_min us_tail 0.63 0.86 -7.75 -16.92 -17.87 -22.25 Long-term residuals_n_bone_pixe ls_by_BodyWeight_g 0.01 0.00 8666.44 -2441.75 12339.48 -3952.88 Long-term residuals_n_bone_pixe ls_by_bodylength_min us_tail 0.07 0.08 6223.09 -2834.11 7435.26 -2934.13 Long-term residuals_density_by_ BodyWeight_g 0.03 0.06 8.36 -0.03 6.23 -1.25 Long-term residuals_density_by_ bodylength_minus_tail 0.06 0.07 7.54 -0.17 6.25 -0.85 Long-term residuals_Testes_mas 0.72 0.49 -0.01 0.00 0.01 0.01 95 s_g_by_BodyWeight_g Long-term residuals_Testes_mas s_g_by_bodylength_mi nus_tail 0.61 0.58 -0.01 0.00 0.00 0.00 Long-term residuals_SV_mass_g _by_BodyWeight_g 0.51 0.40 0.00 0.00 0.00 -0.01 Long-term residuals_SV_mass_g _by_bodylength_minus _tail 0.93 0.97 -0.01 -0.01 -0.01 -0.01 Long-term residuals_PG_mass_g _by_BodyWeight_g 0.47 0.80 0.00 0.00 0.00 0.00 Long-term residuals_PG_mass_g _by_bodylength_minus _tail 0.36 0.79 -0.01 0.00 0.00 -0.01 Long-term centroid_size 0.68 0.70 986.01 979.27 984.10 976.00 Long-term n_bone_pixels 0.04 0.03 159570.7 5 149511.3 1 160579.5 0 149383.00 Long-term SV_mass_g 0.94 1.00 0.17 0.18 0.17 0.17 Long-term Testes_mass_g 0.62 0.91 0.17 0.18 0.18 0.18 Long-term BodyWeight_g 0.38 0.44 24.47 25.24 23.79 25.00 Long-term bodylength_minus_tail 0.94 0.52 96.00 96.21 98.00 95.50 Supplementary Table 4.2 List of body size, sex gland, and baculum phenotypes based on dissection measurements and microCT data. Starting with the short-term experiment, each row represents a test of the identified variable followed by the T.test P value, Wilcox P value, mean mated, mean unmated, median mated, and median unmated values. 96 5 Concluding remarks 5.1 Impact of this thesis The second chapter of my dissertation represents the first quantitative investigation into the genetics of baculum morphology, providing the first insights into the genetic basis of this extremely rapidly evolving structure. This experiment allowed us to highlight that baculum morphology is most likely influenced by a small number of genes with large affects on baculum shape and size in mice. This finding might explain how these bones can evolve so rapidly, as only a few genomic alterations may be necessary to produce dramatic morphological changes. Understanding the genetic mechanisms of baculum diversity is a critical first step toward the potential manipulation of the bone and understanding of its function in reproduction. In my third chapter, I demonstrated that the baculum was gained and lost multiple times during mammalian evolution. This is an important finding for a number or reasons. First, it suggests that a single functional explanation for bacula cannot be applied to penis bones across mammalian taxa (A. F. Dixson 1987a; Patterson and Thaeler 1982). Second, it can serve as groundwork for understanding the molecular mechanisms influencing the independent gains and losses of bone in mammalian penises. This is an especially remarkable evolutionary trend because the amniote phallus has evolved once (Brennan 2016a; Gredler 2016), meaning that these independent gains and losses occur in homologous tissue. Overall, this work establishes the mammalian penis bone as a legitimate model for understanding rapid evolutionary change, and bone development. There are at least two genetic mechanisms to explain this pattern. First the “light switch” model, which suggests that all species have the potential to develop bacula using the same collection of genes, but that bone growth potential is “flipped” on-or-off in different lineages. The second explanation is termed the “novel pathways” model. Here, we suggest that 97 bacula developed not through the modification of pre-existing genetic systems, but instead each lineage may have employed a lineage specific network of genes to initiate bone development. Inherent to both of these models is the role of testosterone in shaping baculum development (Murakami 1987), especially since the female analog, the baubellum holds a very different pattern of development and evolution compare to the baculum (Lough ‐Stevens et al. 2018). Both of these models, and the potential influence of testosterone will be discussed in greater detail in the Future Directions section of this chapter. Finally, in chapter four, I applied traditional bone histology methods to controlled mating experiments in order to make indirect inferences about forces experienced during mating. This represents the first attempt to visualize biomechanically relevant bone microstructural features like collagen fiber orientation, which have long been used as a proxy of biomechanics in limb bones (Keenan et al. 2017; J. G. Skedros et al. 2006). While we did not detect significant differences in shape or any microstructural feature between mated and unmated males, this experiment did identify regional differences in baculum histology, which were never before detected. Together, this chapter enriches our understanding of the biomechanics of the baculum, and establishes a method for assessing baculum function on a species to species basis. 5.2 Future directions Given the frequency of independent derivations of bacula (N. G. Schultz et al. 2016a), it is imperative to understand the gene expression patterns surrounding the development of this bone. The timing of mouse baculum development is already established (Murakami 1987), but studies have yet to thoroughly investigate the genetic pathways involved in that development. Below I propose a number of studies to address this lack of knowledge, including leveraging 98 histological data to assess gene expression profiles, and the potential of genetic engineering strategies to manipulate baculum morphology to understand its functional relevance. Previous studies have reported that mouse bacula develop in early neonatal mice via the aggregation of mesenchymal tissue (Murakami 1987). Unfortunately, collecting robust gene expression data was not feasible at the time of initial experiments, so our understanding of the pathways utilized to initiate baculum development in mice is limited. Now a number of molecular techniques could be applied to understanding the development of bacula, but the first step in that process is repeating previous histological studies (Murakami 1987) in mice. In particular, focusing on the very first signs of pre-baculum mesenchyme aggregations is critical, since this is the developmental period where those cells receive input to differentiate into bone. Once the time point for that aggregation is confirmed through histology, aggregating cells can then be isolated and used for RNA sequencing to identify the collection of genes employed to initiate mouse baculum development. RNA-sequencing is a necessary step toward understanding the development of bacula, but by itself this technique does not provide ample evidence for either the “light-switch” or “novel pathways” models. In order to investigate either of these developmental models, comparisons need to be made between baculum development and another bone. A logical comparison for gene expression patterns in developing bacula are hindlimbs given that mutants for key hind limb developmental genes have been shown to cause baculum deformities (Infante et al. 2015; Kondo et al. 1997b; József Zákány et al. 1997). Therefore, comparing gene expression patterns at analogous developmental periods between hindlimbs and genitalia may provide insights into shared or divergent pathways. To fully understand the ways that bacula can develop, expression data from species of different baculum lineages, like rabbits, mice and ferrets can be compared. According to our 99 phylogenetic results (N. G. Schultz et al. 2016a) these species represent two independent gains of bacula (mice and ferrets) and a third species without a penis bone entirely (rabbits). By conducting this experiment, one could elucidate if mice and ferrets use similar genes (light- switch) or divergent genetic networks (novel pathways) to initiate baculum development, and if those genes are altered or missing in rabbits. Two major outcomes need to be considered with these studies. One, and potentially the least likely outcome, is that each species investigated uses a unique combination of genes to initiate baculum development. The more likely outcome, given the deep homology of limbs (Sears et al. 2015) and the known similarities between limb and genital development (Cohn 2011) is that different mammal lineages manipulate a similar genetic network to initiate baculum development. If a common collection of genes is used to develop bacula across species, then a careful investigation into the manipulation of expression patterns is required, principally examining cis-regulatory evolution. Cis-regulatory evolutionary allows deeply conserved genetic networks, like those critical for skeletal development to be structurally unaltered (Prud'homme et al. 2007; Tarazona et al. 2016), while dramatically shifting gene expression patterns (Carroll 2005; Marlétaz et al. 2018). For example, the evolution of two separate enhancer elements for the bone morphogenetic protein GDF6 has been shown to be responsible for the change in toe length between chimps and humans and the number of spines between freshwater and marine sticklebacks (Indjeian et al. 2016). This example illustrates how divergent animal lineages can target the same gene to change bone phenotypes, without altering the core function of the highly conserved gene. Cis-regulatory evolution can also drive divergent gene expression patterns in different tissues within a single organism. Enhancers commonly affect morphological evolution through the timing of gene expression, and they can be restricted to a certain part of the body (Rice and 100 Rebeiz 2019). For instance, the insertion of different tissue-specific gene enhancers from snakes into mice drove the expression of a core skeletal development gene TBX4 (Infante et al. 2015). Specifically, the hindlimb enhancer in snakes influenced both hindlimb and genital development, including the baculum, in mice, but the forelimb enhancer for the same gene did not show an effect on those tissues (Infante et al. 2015). The regional influence of certain enhancers provides support for cis-regulatory evolution influencing baculum development, as bacula are known to be affected by core skeletal genes that cannot be fundamentally altered (Kondo et al. 1997b). Investigations into the influence of cis-regulation on baculum development could be performed via ChIP-seq (chromatin immunoprecipitation sequencing) to look for associations between proteins known to influence gene expression, and their DNA targets (Visel et al. 2009). If a baculum specific enhancer were discovered, reducing the bone and directly testing its function could be possible using transgenics or other gene replacement technologies like CRISPR-Cas9 (Nakamura et al. 2016). If there is progress made on the front of understanding the genetic influence of bacula, then it may be possible to manipulate the bone and see how mating functions without the baculum. The quick transitions in baculum evolution can also result from changes in primordial tissue boundaries and communication, specifically between mesenchyme and epithelium. These interactions are generally known to be important for skeletal development (Hall and Miyake 2000), and they have been specifically investigated in baculum morphogenesis (Williams- Ashman and Reddi 1991). Two studies in particular sought to better understand this relationship in rat baculum development, through transplantation of rat baculum mesenchyme into other locations of the body to see if the bone continued to grow (Beresford and Clayton 1977; Murakami and Mizuno 1986). When mesenchyme was grafted by itself no bone formed, but 101 when co-grafted with genital epithelium, bone and cartilaginous structures appeared indicating that this tissue interaction is critical for baculum development (Murakami and Mizuno 1986). Research on mesenchyme-epithelium interactions in tooth development have illustrated that perturbations in this dynamic often function via β-catenin Wnt-signaling pathways (Harris et al. 2006) (Järvinen et al. 2006), which are also essential for bone and genital development (Berendsen and Olsen 2015; Cohn 2011). Some of the genes shown to be differentially expressed in mouse baculum development (Dact1) are modulators of the Wnt-signaling pathway (N. G. Schultz et al. 2016b), and thus may be implicated in baculum evolution. Changes in mesenchyme-epithelium interactions can also be modulated by cis-regulatory evolution (Stemmler et al. 2005; Venkov et al. 2011), further reinforcing the need to investigate cis- regulatory dynamics. Testosterone is also known to play a significant role in baculum development, as mutant male mice lacking androgen receptors develop significantly reduced bacula (Murakami 1987), and baculum density decreases in castrated dogs (Sharir et al. 2011). Androgen regulation of skeletogenesis may also be important for understanding the divergent evolution between bacula and baubella, as female rats only develop the analog to bacula when treated with androgens (Glucksmann et al. 1976). While these studies highlight a pattern of androgen regulation in baculum development, it remains unknown exactly how androgens manipulate this process. Androgens are known to play an important role in human skeletal diseases like osteoporosis (Vanderschueren et al. 2004), but the exact genomic targets of androgen receptors are relatively unknown. Given the heightened sensitivity of bacula to androgens, mapping the genomic targets of the transcription factor androgen receptor (AR) in male genitalia could be incredibly valuable. Again, a reasonable strategy here would be to perform ChIP-seq against AR followed by RNA- 102 seq to detect the expression levels of AR targets in bacula compared to another bone. Studying the phenomenon of androgen sensitivity in penis bones could provide valuable insights into new dynamics of androgens in skeletal development and tissue maintenance. Overall, advances in our understanding of baculum evolution could potentially be applied to translational research. Bone is the second most transplanted tissue in the world, and bone- engineering techniques still require optimization (Mishra et al. 2016). Current research on bone development is primarily focused on the in vitro manipulation of developmental pathways understood through skeletal structures with deep homology (Sears et al. 2015). Applying insights from baculum development research could uncover alternative mechanisms, specifically related to the influences of cell fate decisions. A tremendous amount of bone engineering research focuses on mesenchymal stem cells (MSCs) (Barry and Murphy 2004; Murphy et al. 2013), and there is strong evidence that the location in which MSCs are derived largely dictates their ability to differentiate into bone (da Silva Meirelles et al. 2006). Moreover, MSCs from different species are capable of differentiating in vitro to varying degrees (Barry and Murphy 2004). These are the same types of cells that eventually develop into bacula (Murakami and Mizuno 1986), therefore understanding what factors manipulate those cell fate decisions in penis bones may provide useful information for the bone engineering world. A combination of histology, cell culturing, and sequencing approaches can be used to describe this dynamic. Evolution has already established a clear pattern of rapid modulation of bone development in bacula, and a more complete understanding of the mechanisms underlying those changes could potentially enhance our ability to grow and harvest bone for clinical applications. Ultimately, the baculum represents a unique opportunity to uncover novel mechanisms of bone development that traditional bone models may not be capable of discovering. 103 References Adams, DC and Otárola-Castillo, E (2012), 'Package ‘geomorph’: Geometric morphometric analysis of 2d/3d landmark data', R package version, 1. Adams, Dean C and Otárola ‐Castillo, Erik (2013), 'geomorph: an R package for the collection and analysis of geometric morphometric shape data', Methods in Ecology and Evolution, 4 (4), 393-99. Adams, Donald R and Sutton, Dallas A (1968), 'A description of the baculum and os clitoris of Eutamias townsendii ochrogenys', Journal of Mammalogy, 764-68. Adler, Norman T. and Toner, James Patrick (1986), 'The effects of copulatory behavior on sperm transport and fertility in rats', Ann N Y Acad Sci, 474 (1), 21-32. Anderson, Christopher M and Langerhans, R Brian (2015), 'Origins of female genital diversity: Predation risk and lock ‐and ‐key explain rapid divergence during an adaptive radiation', Evolution, 69 (9), 2452-67. Anderson, Matthew J (2000), 'Penile morphology and classification of bush babies (subfamily Galagoninae)', International journal of primatology, 21 (5), 815-36. Andersson, Malte and Iwasa, Yoh (1996), 'Sexual selection', Trends in ecology & evolution, 11 (2), 53-58. Andersson, Malte B (1994), Sexual selection (Princeton University Press). André, Gonçalo I, Firman, Renée C, and Simmons, Leigh W (2018), 'Phenotypic plasticity in genitalia: baculum shape responds to sperm competition risk in house mice', Proc. R. Soc. B, 285 (1882), 20181086. Arnqvist, G. and Rowe, L. (2005), Sexual conflict, eds J. R. Krebs and T. H. Clutton-Brock (Monographs in behavior and ecology; Princeton, New Jersey: Princeton University Press). Arnqvist, Goran and Danielsson, Ingela (1999), 'Copulatory behavior, genital morphology, and male fertilization success in water striders', Evolution, 147-56. Arnqvist, Göran (1997), 'The evolution of animal genitalia: distinguishing between hypotheses by single species studies', Biological Journal of the Linnean Society, 60 (3), 365-79. --- (1998), 'Comparative evidence for the evolution of genitalia by sexual selection', Nature, 393 (6687), 784-86. Arnqvist, Göran and Rowe, Locke (2013), Sexual conflict (Princeton University Press). Asdell, Sydney Arthur and Hubbs, Carl Leavitt (1964), Patterns of mammalian reproduction (Cornell University Press Ithaca, New York). Barnard, Alexandra A and Masly, John P (2018), 'Divergence in female damselfly sensory structures is consistent with a species recognition function but shows no evidence of reproductive character displacement', Ecology and Evolution. Barnard, Alexandra A, et al. (2017), 'Mechanical and tactile incompatibilities cause reproductive isolation between two young damselfly species', Evolution, 71 (10), 2410-27. Barry, Frank P and Murphy, J Mary (2004), 'Mesenchymal stem cells: clinical applications and biological characterization', The international journal of biochemistry & cell biology, 36 (4), 568-84. Bartke, A (1974), 'Increased sensitivity of seminal vesicles to testosterone in a mouse strain with low plasma testosterone levels', Journal of Endocrinology, 60 (1), 145-48. Baryshnikov, Gennady F, Bininda-Emonds, Olaf RP, and Abramov, Alexei V (2003), 'Morphological variability and evolution of the baculum (os penis) in Mustelidae (Carnivora)', Journal of Mammalogy, 84 (2), 673-90. Baur, Bruno (1998), 'Sperm competition in molluscs', Sperm competition and sexual selection. Beamer, Wesley G, et al. (1999), 'Quantitative trait loci for bone density in C57BL/6J and CAST/EiJ inbred mice', Mammalian Genome, 10 (11), 1043-49. 104 Bell, Michael A (1987), 'Interacting evolutionary constraints in pelvic reduction of threespine sticklebacks, Gasterosteus aculeatus (Pisces, Gasterosteidae)', Biological Journal of the Linnean Society, 31 (4), 347-82. Berendsen, Agnes D and Olsen, Bjorn R (2015), 'Bone development', Bone, 80, 14-18. Beresford, WA and Clayton, SP (1977), 'Intracerebral transplantation of the genital tubercle in the rat: the fate of the penile bone and cartilages', Journal of anatomy, 123 (Pt 2), 297. Birkhead, Tim (2000), 'Genitalia', Promiscuity: an evolutionary history of sperm competition (Harvard University Press), 58-105. Bollback, Jonathan P (2006), 'SIMMAP: stochastic character mapping of discrete traits on phylogenies', BMC bioinformatics, 7 (1), 88. Bonduriansky, Russell (2007), 'Sexual selection and allometry: a critical reappraisal of the evidence and ideas', Evolution, 61 (4), 838-49. Bonewald, LF, Dallas, SL, and Gorski, JP (2009), 'Bone mineralization', The Skeletal System, Cold Spring Harbor Laboratory Press, NY, USA, 277-95. Bookstein, F. (1991), Morphometric tools for landmark data: geometry and biology (New York: Cambridge University Press). Bookstein, F. J. (1997), 'Landmark methods for forms without landmarks: morphometrics of group differences in outline shape', Medical Image Analysis, 1, 225-43. Bookstein, Fred L (1989), '" Size and shape": a comment on semantics', Systematic zoology, 38 (2), 173-80. --- (1996), 'Biometrics, biomathematics and the morphometric synthesis', Bulletin of mathematical biology, 58 (2), 313-65. Brain, Paul F, Homady, Merza H, and Mainardi, Marisa (1983), 'Preputial glands, dominance and aggressiveness, in mice', Italian Journal of Zoology, 50 (3-4), 173-87. Brassey, Charlotte A, Gardiner, James D, and Kitchener, Andrew C (2018), 'Testing hypotheses for the function of the carnivoran baculum using finite-element analysis', Proc. R. Soc. B, 285 (1887), 20181473. Brennan, Patricia LR (2016a), 'Evolution: one penis after all', Current Biology, 26 (1), R29-R31. --- (2016b), 'Studying genital coevolution to understand intromittent organ morphology', Integrative and comparative biology, 56 (4), 669-81. Brennan, Patricia LR and Prum, Richard O (2015), 'Mechanisms and evidence of genital coevolution: the roles of natural selection, mate choice, and sexual conflict', Cold Spring Harbor perspectives in biology, 7 (7), a017749. Brennan, Patricia LR, Clark, Christopher J, and Prum, Richard O (2010), 'Explosive eversion and functional morphology of the duck penis supports sexual conflict in waterfowl genitalia', Proceedings of the Royal Society of London B: Biological Sciences, 277 (1686), 1309-14. Brennan, Patricia LR, et al. (2007), 'Coevolution of male and female genital morphology in waterfowl', PLoS one, 2 (5), e418-e18. Breseño, R. D. and Eberhard, W. G. (2009), 'Experimental modifications imply a stimulatory function for male tsetse fly genitalia, supporting cryptic female choice theory', Journal of Evolutionary Biology, 22 (7), 1516-25. Bromage, Timothy G, et al. (2003), 'Circularly polarized light standards for investigations of collagen fiber orientation in bone', The Anatomical Record Part B: The New Anatomist, 274 (1), 157-68. Broman, Karl W. and Sen, Saunak (2009), A guide to QTL mapping with R/qtl (New York: Springer). Burr, DB, Robling, Alexander G, and Turner, Charles H (2002), 'Effects of biomechanical stress on bones in animals', Bone, 30 (5), 781-86. Burt, W (1960a), 'Bacula of North American mammals', Misc Publ Mus Zool Univ Mich 113, 1– 75. 105 Burt, William Henry (1960b), 'Bacula of North American mammals'. Butterworth, David 'https://github.com/dbworth/minimum-area-bounding-rectangle', <https://github.com/dbworth/minimum-area-bounding-rectangle>, accessed. Callery, Roberdeau (1951), 'Development of the os genitale in the golden hamster, Mesocricetus (Cricetus) auratus', Journal of Mammalogy, 204-07. Čanády, Alexander (2013), 'Variability of the baculum in the red fox (Vulpes vulpes) from Slovakia', Zoology and Ecology, 23 (3), 165-70. Carando, S, et al. (1991), 'Macroscopic shape of, and lamellar distribution within, the upper limb shafts, allowing inferences about mechanical properties', Bone, 12 (4), 265-69. Carayon, J (1966), 'Traumatic insemination and the paragenital system', Monograph of Cimicidae (Hemiptera, Heteroptera). College ParkMd: Entomological Society of America, 81-166. Carlson, Kristian J and Judex, Stefan (2007), 'Increased non-linear locomotion alters diaphyseal bone shape', Journal of Experimental Biology, 210 (17), 3117-25. Carroll, Sean B (2005), 'Evolution at two levels: on genes and form', PLoS biology, 3 (7), e245. --- (2008), 'Evo-devo and an expanding evolutionary synthesis: a genetic theory of morphological evolution', Cell, 134 (1), 25-36. Carter, DR, Van der Meulen, MCH, and Beaupre, GS (1996), 'Mechanical factors in bone growth and development', Bone, 18 (1), S5-S10. Chaine, J. (1925), 'L'os pénien: étude descriptive et comparative', Actes de la Société linnéenne de Bordeaux, 78, 1-195. Chan, Yingguang Frank, et al. (2010), 'Adaptive evolution of pelvic reduction in sticklebacks by recurrent deletion of a Pitx1 enhancer', science, 327 (5963), 302-05. Cheng, Riyan, et al. (2011), 'QTLRel: an R package for genome-wide association studies in which relatedness is a concern', BMC genetics, 12 (1), 66. Cheng, Riyan, et al. (2010), 'Genome-wide association studies and the problem of relatedness among advanced intercross lines and other highly recombinant populations', Genetics, 185 (3), 1033-44. Chessel, Daniel, Dufour, Anne B, and Thioulouse, Jean (2004), 'The ade4 package-I-One-table methods', R news, 4 (1), 5-10. Cheverud, James M, et al. (1996), 'Quantitative trait loci for murine growth', Genetics, 142 (4), 1305-19. Christensen, George C (1954), 'Angioarchitecture of the canine penis and the process of erection', American Journal of Anatomy, 95 (2), 227-61. Cobb, John and Duboule, Denis (2005), 'Comparative analysis of genes downstream of the Hoxd cluster in developing digits and external genitalia', Development, 132 (13), 3055- 67. Cobrinik, David, et al. (1996), 'Shared role of the pRB-related p130 and p107 proteins in limb development', Genes & development, 10 (13), 1633-44. Cohn, Martin J (2011), 'Development of the external genitalia: conserved and divergent mechanisms of appendage patterning', Developmental Dynamics, 240 (5), 1108-15. Comelis, Manuela T, et al. (2015), 'Penile histomorphology of the neotropical bat Eptesicus furinalis (Chiroptera: Vespertilionidae)', Zoologischer Anzeiger-A Journal of Comparative Zoology, 258, 92-98. Cordero, Carlos and Eberhard, William G (2003), 'Female choice of sexually antagonistic male adaptations: a critical review of some current research', Journal of evolutionary biology, 16 (1), 1-6. Cordoba-Aguilar, Alex (1999), 'Male copulatory sensory stimulation induces female ejection of rival sperm in a damselfly', Proceedings of the Royal Society of London B: Biological Sciences, 266 (1421), 779-84. 106 Cox, Allison, et al. (2009), 'A new standard genetic map for the laboratory mouse', Genetics, 182 (4), 1335-44. Crudgington, Helen S and Siva-Jothy, Mike T (2000), 'Genital damage, kicking and early death', Nature, 407 (6806), 855-56. Csanády, Alexander, Duranková, Silva, and Labancová, Erika (2019), 'Are baculum size and allometry a response to post-copulatory sexual selection in promiscuous males of the house mouse?', Zoomorphology, 1-10. Currey, JD (2003), 'The many adaptations of bone', Journal of biomechanics, 36 (10), 1487-95. da Silva Meirelles, Lindolfo, Chagastelles, Pedro Cesar, and Nardi, Nance Beyer (2006), 'Mesenchymal stem cells reside in virtually all post-natal organs and tissues', Journal of cell science, 119 (11), 2204-13. Daly, Martin (1978), 'The cost of mating', The American Naturalist, 112 (986), 771-74. Darvasi, A and Soller, M (1995), 'Advanced intercross lines, an experimental population for fine genetic mapping', Genetics, 141 (3), 1199-207. Darvasi, Ariel (1998), 'Experimental strategies for the genetic dissection of complex traits in animal models', Nature genetics, 18 (1), 19-24. Davies, NB (1983), 'Polyandry, cloaca-pecking and sperm competition in dunnocks', Nature, 302 (5906), 334. Davis, D Dwight (1964), 'The Giant Panda: A morphological Study of Evolutionary Mechanisms (Fieldiana Zoology Memoirs Vol 3)'. De Crombrugghe, Benoit and Akiyama, Haruhiko (2009), '5 Transcriptional Control of Chondrocyte Differentiation', Cold Spring Harbor Monograph Archive, 53, 147-70. Dean, M. D., Ardlie, K. G., and Nachman, M. W. (2006), 'The frequency of multiple paternity suggests that sperm competition is common in house mice (Mus domesticus)', Mol Ecol, 15, 4141-51. Dean, MN and Shahar, R (2012), 'The structure ‐mechanics relationship and the response to load of the acellular bone of neoteleost fish: a review', Journal of Applied Ichthyology, 28 (3), 320-29. Dewsbury, Donald A (1981), 'On the function of the multiple-intromission, multiple-ejaculation copulatory patterns of rodents', Bulletin of the Psychonomic Society, 18 (4), 221-23. Diamond, M. (1970), 'Intromission pattern and species vaginal code in relation to induction of pseudopregnancy', Science, 169 (949), 995-97. Dines, James P, et al. (2014), 'Sexual selection targets cetacean pelvic bones', Evolution, 68 (11), 3296-306. Dines, James P., et al. (2015), 'A trade-off between precopulatory and postcopulatory trait investment in male cetaceans', Evolution, 69 (6), 1560-72. Dixson, A. F. (1987a), 'Baculum length and copulatory behavior in primates', American Journal of Primatology, 13 (1), 51-60. Dixson, AF (1987b), 'Observations on the evolution of the genitalia and copulatory behaviour in male primates', Journal of Zoology, 213 (3), 423-43. --- (1995), 'Baculum length and copulatory behaviour in carnivores and pinnipeds (Grand Order Ferae)', Journal of Zoology, 235 (1), 67-76. Dixson, Alan (1998), Primate sexuality (Wiley Online Library). Dixson, Alan F and Anderson, Matthew J (2004), 'Sexual behavior, reproductive physiology and sperm competition in male mammals', Physiology & Behavior, 83 (2), 361-71. Doherty, Alison H, Ghalambor, Cameron K, and Donahue, Seth W (2015), 'Evolutionary physiology of bone: bone metabolism in changing environments', Physiology, 30 (1), 17- 29. Dollé, P., et al. (1991), 'HOX-4 genes and the morphogenesis of mammalian genitalia', Genes & Development, 5 (10), 1767-76. 107 Dougherty, Liam R, et al. (2015), 'Experimental reduction of intromittent organ length reduces male reproductive success in a bug', Proceedings of the Royal Society B: Biological Sciences, 282 (1808), 20150724. Dray, Stéphane and Dufour, Anne-Béatrice (2007), 'The ade4 package: implementing the duality diagram for ecologists', Journal of statistical software, 22 (4), 1-20. Du, Shao Jun, et al. (2001), 'Visualizing normal and defective bone development in zebrafish embryos using the fluorescent chromophore calcein', Developmental biology, 238 (2), 239-46. Ducy, Patricia, Schinke, Thorsten, and Karsenty, Gerard (2000), 'The osteoblast: a sophisticated fibroblast under central surveillance', Science, 289 (5484), 1501-04. Dufour, Léon (1844), 'Anatomie générale des Dipteres', Ann. Sci. Nat, 1, 244-64. Dyck, Markus G, Bourgeois, Jackie M, and Miller, Edward H (2004), 'Growth and variation in the bacula of polar bears (Ursus maritimus) in the Canadian Arctic', Journal of Zoology, 264 (1), 105-10. Eadie, W Robert (1947), 'The accessory reproductive glands of Parascalops with notes on homologies', The Anatomical Record, 97 (2), 239-51. Eady, Paul E, Hamilton, Leticia, and Lyons, Ruth E (2006), 'Copulation, genital damage and early death in Callosobruchus maculatus', Proceedings of the Royal Society B: Biological Sciences, 274 (1607), 247-52. Eberhard, W. G. (2000), 'Criteria for demonstrating postcopulatory female choice', Evolution Int J Org Evolution, 54 (3), 1047-50. Eberhard, William (1996), Female control: sexual selection by cryptic female choice (Princeton University Press). --- (2009a), 'Evolution of genitalia: theories, evidence, and new directions', Genetica, 138 (1), 5- 18. Eberhard, William G (1985), Sexual selection and animal genitalia (Harvard University Press). --- (1992), 'Species isolation, genital mechanics, and the evolution of species ‐specific genitalia in three species of Macrodactylus beetles (Coleoptera, Scarabeidae, Melolonthinae)', Evolution, 46 (6), 1774-83. --- (1993), 'Evaluating models of sexual selection: genitalia as a test case', American Naturalist, 564-71. --- (2009b), 'Postcopulatory sexual selection: Darwin's omission and its consequences', Proceedings of the National Academy of Sciences, 106 (Supplement 1), 10025-32. Erben, Reinhold G and Glösmann, Martin (2012), 'Histomorphometry in rodents', Bone Research Protocols (Springer), 279-303. Evans, Howard E and de Lahunta, Alexander (2013), Miller's Anatomy of the Dog (Elsevier Health Sciences). Fabre, Pierre-Henri, et al. (2012), 'A glimpse on the pattern of rodent diversification: a phylogenetic approach', BMC Evolutionary Biology, 12 (1), 88. Felsenstein, Joseph (1981), 'Evolutionary trees from DNA sequences: a maximum likelihood approach', Journal of molecular evolution, 17 (6), 368-76. Firman, R. C. and Simmons, L. W. (2008), 'The frequency of multiple paternity predicts variation in testes size among island populations of house mice', J Evol Biol, 21, 1524-33. Firman, Renée C, et al. (2017), 'Postmating female control: 20 years of cryptic female choice', Trends in ecology & evolution, 32 (5), 368-82. FitzJohn, Richard G (2012), 'Diversitree: comparative phylogenetic analyses of diversification in R', Methods in Ecology and Evolution, 3 (6), 1084-92. Fitzpatrick, John L, et al. (2012), 'Male contest competition and the coevolution of weaponry and testes in pinnipeds', Evolution, 66 (11), 3595-604. Fricke, Claudia, et al. (2009), 'The conditional economics of sexual conflict', Biology Letters, 5 (5), 671-74. 108 Friesen, Christopher R, et al. (2014), 'Sexual conflict over mating in red-sided garter snakes (Thamnophis sirtalis) as indicated by experimental manipulation of genitalia', Proceedings of the Royal Society B: Biological Sciences, 281 (1774), 20132694. Fujikura, Kohei, et al. (2013), 'Kif14 mutation causes severe brain malformation and hypomyelination', PloS one, 8 (1), e53490. Gasparini, Clelia and Evans, Jonathan P (2018), 'Female control over multiple matings increases the opportunity for postcopulatory sexual selection', Proc. R. Soc. B, 285 (1888), 20181505. Gerhardt, Ulric (1909), 'Die gegenwärtige Stand der Kenntnisse von den Copulationsorganen der Wirbeltiere, insbesondere der Amnioten', Ergebnisse und Fortschritte der Zoologie, 1, 307-402. Glickman, Stephen E, et al. (1987), 'Androstenedione may organize or activate sex-reversed traits in female spotted hyenas', Proceedings of the National Academy of Sciences, 84 (10), 3444-47. Glucksmann, A, et al. (1976), 'The effect of neonatal treatment of male mice with antiandrogens and of females with androgens on the development of the os penis and os clitoridis', Journal of anatomy, 121 (Pt 2), 363. Goldman, Haviva M, et al. (2003), 'Preferred collagen fiber orientation in the human mid ‐shaft femur', The Anatomical Record, 272 (1), 434-45. Good, Jeffrey M., et al. (2003), 'Phylogeography and introgressive hybridization: chipmunks (genus Tamias) in the northern rocky mountains', Evolution, 57 (8), 1900-16. Goodale, HD (1938), 'A study of the inheritance of body weight in the albino mouse by selection', Journal of Heredity, 29 (3), 101-12. Gredler, Marissa L (2016), 'Developmental and evolutionary origins of the amniote phallus', Integrative and Comparative Biology, icw102. Gredler, Marissa L, Seifert, Ashley W, and Cohn, Martin J (2015), 'Tissue-specific roles of Fgfr2 in development of the external genitalia', Development, 142 (12), 2203-12. Gunz, P., Mitteroecker, P., and Bookstein, F. (2005), 'Semilandmarks in three dimensions', in Dennis E Slice (ed.), Modern morphometrics in physical anthropology (New York: Kluwer Press), 73-98. Gustafsson, Lars and Qvarnström, Anna (2005), 'A test of the “sexy son” hypothesis: sons of polygynous collared flycatchers do not inherit their fathers’ mating status', The American Naturalist, 167 (2), 297-302. Hall, Brian K and Miyake, Tsutomo (2000), 'All for one and one for all: condensations and the initiation of skeletal development', Bioessays, 22 (2), 138-47. Hamilton, W.J. (1949), 'The bacula of some North American vespertilionid bats', Journal of Mammalogy, 30, 97-102. Hanlon, Roger T and Messenger, John B (2018), 'Reproductive Behaviour', Cephalopod behaviour (Cambridge University Press), 148-205. Haraguchi, Ryuma, et al. (2001), 'Unique functions of Sonic hedgehog signaling during external genitalia development', Development, 128 (21), 4241-50. Haraguchi, Ryuma, et al. (2000), 'Molecular analysis of external genitalia formation: the role of fibroblast growth factor (Fgf) genes during genital tubercle formation', Development, 127 (11), 2471-79. Harris, Matthew P, et al. (2006), 'The development of archosaurian first-generation teeth in a chicken mutant', Current Biology, 16 (4), 371-77. Haubruge, Eric, et al. (1999), 'Fertilization by proxy: rival sperm removal and translocation in a beetle', Proceedings of the Royal Society of London B: Biological Sciences, 266 (1424), 1183-87. Hayssen, Virginia 'Mammalian Species Complete Systematic List of Accounts', (updated 29 August 2014) <http://www.science.smith.edu/msi/>, accessed. 109 Head, Megan, et al. (2016), 'Sexual selection on body size, genitals and heterozygosity: effects of demography and habitat complexity', bioRxiv, 045724. Hérault, Yann, et al. (1997), 'Ulnaless (Ul), a regulatory mutation inducing both loss-of-function and gain-of-function of posterior Hoxd genes', Development, 124 (18), 3493-500. Herdina, Anna Nele (2008), Light microscopy of the penis bone (baculum) in the Plecotus species (Chiroptera) in Austria (na). Herdina, Anna Nele, et al. (2010), 'Histomorphology of the Penis Bone (Baculum) in the Gray Long ‐Eared Bat Plecotus austriacus (Chiroptera, Vespertilionidae)', The Anatomical Record, 293 (7), 1248-58. Herdina, Anna Nele, et al. (2015a), 'Testing hypotheses of bat baculum function with 3D models derived from microCT', Journal of anatomy, 226 (3), 229-35. Herdina, Anna Nele, et al. (2014), 'MicroCT imaging reveals morphometric baculum differences for discriminating the cryptic species Pipistrellus pipistrellus and P. pygmaeus', Acta Chiropterologica, 16 (1), 157-68. Herdina, Anna Nele, et al. (2015b), 'Correlative 3D ‐imaging of Pipistrellus penis micromorphology: Validating quantitative microCT images with undecalcified serial ground section histomorphology', Journal of morphology, 276 (6), 695-706. Hernandez, CJ, Majeska, RJ, and Schaffler, MB (2004), 'Osteocyte density in woven bone', Bone, 35 (5), 1095-99. Herrera, Ana M, et al. (2013), 'Developmental basis of phallus reduction during bird evolution', Current Biology, 23 (12), 1065-74. Hershkovitz, Philip (1993), 'Male external genitalia of non-prehensile tailed South American monkeys. pt. 1. Subfamily Pitheciinae, family Cebidae', Fieldiana (USA). Publication/Field Museum of Natural History (USA). no. 1451. Higginson, Dawn M, et al. (2012), 'Female reproductive tract form drives the evolution of complex sperm morphology', Proceedings of the National Academy of Sciences, 109 (12), 4538-43. Holwell, Gregory I, et al. (2010), 'Genital shape correlates with sperm transfer success in the praying mantis Ciulfina klassi (Insecta: Mantodea)', Behavioral ecology and sociobiology, 64 (4), 617-25. Hooper, Emmet Thurman (1960), 'The glans penis in Neotoma (Rodentia) and allied genera'. --- (1961), 'The glans penis in Proechimys and other caviomorph rodents'. Hosken, D.J., K.E. Jones, K. Chipperfield, and A. Dixson (2001), 'Is the bat os penis sexually selected?', Behavioral Ecology and Sociobiology, 50, 450-60. Hosken, David J and Stockley, Paula (2004), 'Sexual selection and genital evolution', Trends in Ecology & Evolution, 19 (2), 87-93. Hosken, David J, et al. (2018), 'Penis evolution across species: divergence and diversity', Nature Reviews Urology, 1. Hotzy, Cosima, et al. (2012), 'Phenotypic engineering unveils the function of genital morphology', Current Biology, 22 (23), 2258-61. House, C. M. and Simmons, L. W. (2003), 'Genital morphology and fertilization success in the dung beetle Onthophagus taurus: an example of sexually selected male genitalia', Proc Biol Sci, 270 (1514), 447-55. House, Clarissa M, et al. (2013), 'Sexual and natural selection both influence male genital evolution', PLoS One, 8 (5), e63807. House, CM and Simmons, LW (2005), 'The evolution of male genitalia: patterns of genetic variation and covariation in the genital sclerites of the dung beetle Onthophagus taurus', Journal of evolutionary biology, 18 (5), 1281-92. 110 Houssaye, Alexandra (2013), 'Bone histology of aquatic reptiles: what does it tell us about secondary adaptation to an aquatic life?', Biological Journal of the Linnean Society, 108 (1), 3-21. Hu, Hongliang, et al. (2005), 'Sequential roles of Hedgehog and Wnt signaling in osteoblast development', Development, 132 (1), 49-60. Huelsenbeck, John P, Nielsen, Rasmus, and Bollback, Jonathan P (2003), 'Stochastic mapping of morphological characters', Systematic Biology, 52 (2), 131-58. Iismaa, Siiri E, et al. (2009), 'Transglutaminases and disease: lessons from genetically engineered mouse models and inherited disorders', Physiological reviews, 89 (3), 991- 1023. Indjeian, Vahan B, et al. (2016), 'Evolving new skeletal traits by cis-regulatory changes in bone morphogenetic proteins', Cell, 164 (1), 45-56. Infante, Carlos R, et al. (2015), 'Shared Enhancer Activity in the Limbs and Phallus and Functional Divergence of a Limb-Genital cis-Regulatory Element in Snakes', Developmental cell, 35 (1), 107-19. Järvinen, Elina, et al. (2006), 'Continuous tooth generation in mouse is induced by activated epithelial Wnt/β-catenin signaling', Proceedings of the National Academy of Sciences, 103 (49), 18627-32. Jepsen, Karl J, et al. (2015), 'Establishing biomechanical mechanisms in mouse models: practical guidelines for systematically evaluating phenotypic changes in the diaphyses of long bones', Journal of Bone and Mineral Research, 30 (6), 951-66. Johnson, Michael S (1982), 'Polymorphism for direction of coil in Partula suturalis: behavioural isolation and positive frequency dependent selection', Heredity, 49 (2), 145. Kalmey, JK and Lovejoy, C Owen (2002), 'Collagen fiber orientation in the femoral necks of apes and humans: do their histological structures reflect differences in locomotor loading?', Bone, 31 (2), 327-32. Kaudern, Walter (1907), 'Beiträge zur kenntnis der männlichen geschlechtsorgane bei insectivoren', Zoologische Jahrbücher. Abteilung für Anatomie und Ontogenie der Tiere, 24 (4), 521-52. Keane, Thomas M., et al. (2011), 'Mouse genomic variation and its effect on phenotypes and gene regulation', Nature, 477 (7364), 289-94. Keenan, Kendra E, Mears, Chad S, and Skedros, John G (2017), 'Utility of osteon circularity for determining species and interpreting load history in primates and nonprimates', American journal of physical anthropology, 162 (4), 657-81. Kelly, DA (2000), 'Anatomy of the baculum-corpus cavernosum interface in the Norway rat (Rattus norvegicus), and implications for force transfer during copulation', Journal of Morphology, 244 (1), 69-77. Kelly, Diane A (2016), 'Intromittent organ morphology and biomechanics: defining the physical challenges of copulation', Integrative and comparative biology, 56 (4), 705-14. Klaczko, J, Ingram, T, and Losos, J (2015), 'Genitals evolve faster than other traits in Anolis lizards', Journal of Zoology, 295 (1), 44-48. Klonisch, Thomas, Fowler, Paul A, and Hombach-Klonisch, Sabine (2004), 'Molecular and genetic regulation of testis descent and external genitalia development', Developmental biology, 270 (1), 1-18. Kolb, Christian, et al. (2015), 'Mammalian bone palaeohistology: a survey and new data with emphasis on island forms', PeerJ, 3, e1358. Kondo, Takashi, et al. (1997a), 'Of fingers, toes and penises', Nature, 390 (6655), 29-29. Kondo, Takashi, et al. (1997b), 'Of fingers, toes and penises', Nature, 390 (6655), 29-29. Kramer, Melissa G, et al. (1998), 'Genetic variation in body weight gain and composition in the intercross of Large (LG/J) and Small (SM/J) inbred strains of mice', Genetics and Molecular Biology, 21 (2), 211-18. 111 Kronenberg, Henry M (2003), 'Developmental regulation of the growth plate', Nature, 423 (6937), 332. Kruskal, Joseph B (1964), 'Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis', Psychometrika, 29 (1), 1-27. Langerhans, R Brian, Anderson, Christopher M, and Heinen-Kay, Justa L (2016), 'Causes and consequences of genital evolution', Integrative and comparative biology, 56 (4), 741-51. Lanyon, Lance E, et al. (1982), 'Mechanically adaptive bone remodelling', Journal of biomechanics, 15 (3), 141-54. Larivière, S and Ferguson, SH (2002), 'On the evolution of the mammalian baculum: vaginal friction, prolonged intromission or induced ovulation?', Mammal Review, 32 (4), 283-94. Layne, James N (1960), 'The glans penis and baculum of the rodent Dactylomys dactylinus Desmarest', Mammalia, 24 (1), 87-92. Lee, Andrew H and Simons, Erin LR (2015), 'Wing bone laminarity is not an adaptation for torsional resistance in bats', PeerJ, 3, e823. Lemaître, Jean-François, et al. (2012), 'Genital morphology linked to social status in the bank vole (Myodes glareolus)', Behavioral Ecology and Sociobiology, 66 (1), 97-105. LeVasseur-Viens, Hélène and Moehring, Amanda J (2014), 'Individual Genetic Contributions to Genital Shape Variation between Drosophila simulans and D. mauritiana', International journal of evolutionary biology, 2014. Lieberman, Daniel E, et al. (2003), 'Optimization of bone growth and remodeling in response to loading in tapered mammalian limbs', Journal of Experimental Biology, 206 (18), 3125- 38. Liu, J., et al. (1996), 'Genetic analysis of a morphological shape difference in the male genitalia of Drosophila simulans and D. mauritiana', Genetics, 142 (4), 1129-45. Löfstedt, Christer, Herrebout, Wim M, and Menken, Steph BJ (1991), 'Sex pheromones and their potential role in the evolution of reproductive isolation in small ermine moths (Yponomeutidae)', Chemoecology, 2 (1), 20-28. Lonfat, Nicolas, et al. (2014), 'Convergent evolution of complex regulatory landscapes and pleiotropy at Hox loci', Science, 346 (6212), 1004-06. Long, Charles A and Frank, Theodore (1968), 'Morphometric variation and function in the baculum, with comments on correlation of parts', Journal of Mammalogy, 49 (1), 32-43. Lough ‐Stevens, Michael, Schultz, Nicholas G, and Dean, Matthew D (2018), 'The baubellum is more developmentally and evolutionarily labile than the baculum', Ecology and evolution, 8 (2), 1073-83. Lüpold, S, McElligott, AG, and Hosken, DJ (2004), 'Bat genitalia: allometry, variation and good genes', Biological Journal of the Linnean Society, 83 (4), 497-507. Lüpold, Stefan, et al. (2014), 'Female monopolization mediates the relationship between pre- and postcopulatory sexual traits', Nat Commun, 5, 1-8. MacArthur, John W (1944), 'Genetics of body size and related characters. I. Selecting small and large races of the laboratory mouse', American Naturalist, 142-57. Macdonald, Stuart J and Goldstein, David B (1999), 'A quantitative genetic analysis of male sexual traits distinguishing the sibling species Drosophila simulans and D. sechellia', Genetics, 153 (4), 1683-99. Mackay, Trudy F. C., et al. (2012), 'The Drosophila melanogaster Genetic Reference Panel', Nature, 482 (7384), 173-78. Maga, A. Murat, et al. (2015), 'Quantitative trait loci affecting the 3D skull shape and size in mouse and prioritization of candidate genes in-silico', Frontiers in Physiology, 6. Maggiano, Corey M, et al. (2015), 'Methods and theory in bone modeling drift: comparing spatial analyses of primary bone distributions in the human humerus', Journal of anatomy. Mangels, Rachel, et al. (2016), 'Copulatory plugs inhibit the reproductive success of rival males', Journal of evolutionary biology, 29 (11), 2289-96. 112 Marlétaz, Ferdinand, et al. (2018), 'Amphioxus functional genomics and the origins of vertebrate gene regulation', Nature, 1. Martin, RB and Ishida, J (1989), 'The relative effects of collagen fiber orientation, porosity, density, and mineralization on bone strength', Journal of biomechanics, 22 (5), 419-26. Martin, RB, et al. (1996), 'Collagen fiber organization is related to mechanical properties and remodeling in equine bone. A comparsion of two methods', Journal of biomechanics, 29 (12), 1515-21. Masly, John P (2012), '170 years of “lock-and-key”: genital morphology and reproductive isolation', International Journal of Evolutionary Biology, 2012. Matthews Jr, Michael K. and Adler, Norman T. (1978), 'Systematic interrelationship of mating, vaginal plug position, and sperm transport in the rat', Physiol Behav, 20 (3), 303-09. Mayr, Ernst, et al. (1963), Animal species and evolution (797: Belknap Press of Harvard University Press Cambridge, Massachusetts). McFarlin, Shannon C, et al. (2016), 'Primary bone microanatomy records developmental aspects of life history in catarrhine primates', Journal of Human Evolution, 92, 60-79. Mcgill, Thomas E, Corwin, David M, and Harrison, David T (1968), 'Copulatory plug does not induce luteal activity in the mouse Mus musculus', Reproduction, 15 (1), 149-51. McGowen, Michael R, Spaulding, Michelle, and Gatesy, John (2009), 'Divergence date estimation and a comprehensive molecular tree of extant cetaceans', Molecular phylogenetics and evolution, 53 (3), 891-906. McNeil, Casey L., Bain, Clint L., and Macdonald, Stuart J. (2011), 'Multiple Quantitative Trait Loci Influence the Shape of a Male-Specific Genital Structure in Drosophila melanogaster', G3: Genes, Genomes, Genetics, 1 (5), 343-51. Meczyński, S (1974), 'Morphohistological structure of female genital organs in sousliks', Acta theriologica, 19 (1-13), 91-106. Meredith, Robert W, et al. (2011), 'Impacts of the Cretaceous Terrestrial Revolution and KPg extinction on mammal diversification', Science, 334 (6055), 521-24. Miller, Edward H and Burton, Lauren E (2001), 'It's all relative: allometry and variation in the baculum (os penis) of the harp seal, Pagophilus groenlandicus (Carnivora: Phocidae)', Biological Journal of the Linnean Society, 72 (3), 345-55. Miller, Edward H, Jones, Ian L, and Stenson, Garry B (1999), 'Baculum and testes of the hooded seal (Cystophora cristata): growth and size-scaling and their relationships to sexual selection', Canadian Journal of Zoology, 77 (3), 470-79. Mishra, Ruchi, et al. (2016), 'The potential impact of bone tissue engineering in the clinic', Regenerative Medicine, 11 (6), 571-87. Mitteroecker, Philipp and Gunz, Philipp (2009), 'Advances in geometric morphometrics', Evolutionary Biology, 36 (2), 235-47. Miyagawa, Shinichi, et al. (2009), 'Dosage-dependent hedgehog signals integrated with Wnt/β- catenin signaling regulate external genitalia formation as an appendicular program', Development, 136 (23), 3969-78. Molodtsova, Daria, et al. (2014), 'Pleiotropy constrains the evolution of protein but not regulatory sequences in a transcription regulatory network influencing complex social behaviors', Frontiers in genetics, 5, 431. Montañez ‐Rivera, Irene, Nyakatura, John A, and Amson, Eli (2018), 'Bone cortical compactness in ‘tree sloths’ reflects convergent evolution', Journal of anatomy. Montoya ‐Sanhueza, Germán and Chinsamy, Anusuya (2016), 'Long bone histology of the subterranean rodent Bathyergus suillus (Bathyergidae): ontogenetic pattern of cortical bone thickening', Journal of anatomy. Murakami, RYUTARO (1987), 'A histological study of the development of the penis of wild-type and androgen-insensitive mice', Journal of anatomy, 153, 223. 113 Murakami, Ryutaro and Mizuno, Takeo (1984), 'Histogenesis of the os penis and os clitoridis in rats', Development, growth & differentiation, 26 (5), 419-26. --- (1986), 'Proximal—distal sequence of development of the skeletal tissues in the penis of rat and the inductive effect of epithelium', Journal of embryology and experimental morphology, 92 (1), 133-43. Murphy, Matthew B, Moncivais, Kathryn, and Caplan, Arnold I (2013), 'Mesenchymal stem cells: environmentally responsive therapeutics for regenerative medicine', Experimental & molecular medicine, 45 (11), e54. Nagy, Andras, et al. (2003), Manipulating the mouse embryo: a laboratory manual (Firefly Books). Nakamura, Tetsuya, et al. (2016), 'Digits and fin rays share common developmental histories', Nature. Nielsen, Rasmus (2002), 'Mapping mutations on phylogenies', Systematic biology, 51 (5), 729- 39. Nikolskiy, Igor, et al. (2015), 'Using whole-genome sequences of the LG/J and SM/J inbred mouse strains to prioritize quantitative trait genes and nucleotides', BMC genomics, 16 (1), 415. Norgard, Elizabeth A, et al. (2011), 'Genetic factors and diet affect long-bone length in the F34 LG, SM advanced intercross', Mammalian Genome, 22 (3-4), 178-96. Nowak, Ronald M (1999), Walker's Mammals of the World (1: JHU Press). Nyakatura, Katrin and Bininda-Emonds, Olaf RP (2012), 'Updating the evolutionary history of Carnivora (Mammalia): a new species-level supertree complete with divergence time estimates', BMC biology, 10 (1), 12. O’Shaughnessy, Katherine L, Dahn, Randall D, and Cohn, Martin J (2015), 'Molecular development of chondrichthyan claspers and the evolution of copulatory organs', Nature communications, 6, 6698. Oosthuizen, WH and Miller, EH (2000), 'Bacular and testicular growth and allometry in the Cape fur seal Arctocephalus p. pusillus (Otariidae)', Marine Mammal Science, 16 (1), 124-40. Orr, Teri J and Brennan, Patricia LR (2016), 'All features great and small—the potential roles of the baculum and penile spines in mammals', Integrative and Comparative Biology, icw057. Otronen, M and Siva-Jothy, MT (1991), 'The effect of postcopulatory male behaviour on ejaculate distribution within the female sperm storage organs of the fly, Dryomyza anilis (Diptera: Dryomyzidae)', Behavioral Ecology and Sociobiology, 29 (1), 33-37. Paradis, Emmanuel, Claude, Julien, and Strimmer, Korbinian (2004), 'APE: analyses of phylogenetics and evolution in R language', Bioinformatics, 20 (2), 289-90. Parga, Joyce A (2003), 'Copulatory plug displacement evidences sperm competition in Lemur catta', International Journal of Primatology, 24 (4), 889-99. Parker, Clarissa C, et al. (2012), 'Genome ‐wide association for methamphetamine sensitivity in an advanced intercross mouse line', Genes, Brain and Behavior, 11 (1), 52-61. Parker, Clarissa C., et al. (2014), 'High-Resolution Genetic Mapping of Complex Traits from a Combined Analysis of F2 and Advanced Intercross Mice', Genetics, 198 (1), 103-16. Parker, ClarissaC, et al. (2011), 'Fine-mapping alleles for body weight in LG/J × SM/J F2 and F34 advanced intercross lines', Mammalian Genome, 22 (9-10), 563-71. Parker, GA (1984), 'Sperm competition and the evolution of animal mating strategies', Sperm competition and the evolution of animal mating systems, 1-60. Parker, Geoff A., Lessells, C. M., and Simmons, Leigh W. (2013), 'Sperm competition games: a general model for pre-copulatory male-male competition', Evolution, no-no. Patterson, Bruce D (1983), 'Baculum-body size relationships as evidence for a selective continuum on bacular morphology', Journal of Mammalogy, 64 (3), 496-99. 114 --- (1984), 'Geographic variation and taxonomy of Colorado and Hopi chipmunks (genus Eutamias)', Journal of Mammalogy, 65 (3), 442-56. Patterson, Bruce D and Thaeler, Charles S (1982), 'The mammalian baculum: hypotheses on the nature of bacular variability', Journal of Mammalogy, 63 (1), 1-15. Peacock, Sarah J, et al. (2018), 'Predicting the bending properties of long bones: Insights from an experimental mouse model', American journal of physical anthropology, 165 (3), 457- 70. Pearson, Osbjorn M and Lieberman, Daniel E (2004), 'The aging of Wolff's “law”: ontogeny and responses to mechanical loading in cortical bone', American journal of physical anthropology, 125 (S39), 63-99. Peichel, Catherine L, Prabhakaran, Bindu, and Vogt, Thomas F (1997), 'The mouse Ulnaless mutation deregulates posterior HoxD gene expression and alters appendicular patterning', Development, 124 (18), 3481-92. Peirce, J. L., et al. (2004), 'A new set of BXD recombinant inbred lines from advanced intercross populations in mice', BMC Genet, 5, 7. Pellati, Donatella, et al. (2008), 'Genital tract infections and infertility', European Journal of Obstetrics & Gynecology and Reproductive Biology, 140 (1), 3-11. Peluffo, Alexandre E, et al. (2015), 'A major locus controls a genital shape difference involved in reproductive isolation between Drosophila yakuba and Drosophila santomea', G3: Genes, Genomes, Genetics, g3. 115.023481. Perelman, Polina, et al. (2011), 'A molecular phylogeny of living primates', PLoS genetics, 7 (3), e1001342. Peretti, Alfredo V and Eberhard, William G (2010), 'Cryptic female choice via sperm dumping favours male copulatory courtship in a spider', Journal of evolutionary biology, 23 (2), 271-81. Perriton, Claire L, et al. (2002), 'Sonic hedgehog signaling from the urethral epithelium controls external genital development', Developmental biology, 247 (1), 26-46. Perry, Jennifer C. and Rowe, Locke (2012), 'Sexual conflict and antagonistic coevolution across water strider populations', Evolution, 66, 544-57. Pfennig, Karin S and Simovich, Marie A (2002), 'Differential selection to avoid hybridization in two toad species', Evolution, 56 (9), 1840-48. Pflanz, David, et al. (2017), 'Sost deficiency led to a greater cortical bone formation response to mechanical loading and altered gene expression', Scientific reports, 7 (1), 9435. Phillips, Tiffany R, et al. (2015), 'A comprehensive atlas of the adult mouse penis', Sexual Development, 9 (3), 162-72. Popescu, Andrei-Alin, Huber, Katharina T, and Paradis, Emmanuel (2012), 'ape 3.0: New tools for distance-based phylogenetics and evolutionary analysis in R', Bioinformatics, 28 (11), 1536-37. Prud'homme, Benjamin, Gompel, Nicolas, and Carroll, Sean B (2007), 'Emerging principles of regulatory evolution', Proceedings of the National Academy of Sciences, 104 (suppl 1), 8605-12. Pulvers, Jeremy N, et al. (2010), 'Mutations in mouse Aspm (abnormal spindle-like microcephaly associated) cause not only microcephaly but also major defects in the germline', Proceedings of the National Academy of Sciences, 107 (38), 16595-600. Qin, A, et al. (2012), 'V-ATPases in osteoclasts: structure, function and potential inhibitors of bone resorption', The international journal of biochemistry & cell biology, 44 (9), 1422-35. R Core Team (2014), 'R: A language and environment for statistical computing', R Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org/. Rai, Muhammad Farooq, et al. (2015), 'Genetic loci that regulate ectopic calcification in response to knee trauma in LG/J by SM/J advanced intercross mice', Journal of Orthopaedic Research. 115 Ramm, Steven, Khoo, Lin, and Stockley, Paula (2010), 'Sexual selection and the rodent baculum: an intraspecific study in the house mouse (Mus musculus domesticus)', Genetica, 138 (1), 129-37. Ramm, Steven A (2007), 'Sexual selection and genital evolution in mammals: a phylogenetic analysis of baculum length', The American Naturalist, 169 (3), 360-69. Reilly, GWENDOLEN C and Currey, JOHN D (1999), 'The development of microcracking and failure in bone depends on the loading mode to which it is adapted', Journal of Experimental Biology, 202 (5), 543-52. Revell, Liam J (2012), 'Phylogenetic Tools for comparative biology (and other things)'. Rice, Gavin and Rebeiz, Mark (2019), 'Evolution: How Many Phenotypes Do Regulatory Mutations Affect?', Current Biology, 29 (1), R21-R23. Riggs, CM, Lanyon, LE, and Boyde, A (1993), 'Functional associations between collagen fibre orientation and locomotor strain direction in cortical bone of the equine radius', Anatomy and embryology, 187 (3), 231-38. Ripley, Brian D (2002), Modern applied statistics with S (Springer). Robertson, Hugh M and Paterson, Hugh EH (1982), 'Mate recognition and mechanical isolation in Enallagma damselflies (Odonata: Coenagrionidae)', Evolution, 36 (2), 243-50. Rocha-Barbosa, O, et al. (2013), 'Penial morphology in three species of Brazilian Tuco-tucos, Ctenomys torquatus, C. minutus, and C. flamarioni (Rodentia: Ctenomyidae)', Brazilian Journal of Biology, 73 (1), 201-09. Rodriguez, Esequiel, et al. (2011), 'New insights on the morphology of adult mouse penis', Biology of reproduction, 85 (6), 1216-21. Rohlf, F James and Bookstein, Fred L (1990), Proceedings of the Michigan morphometrics workshop (University of Michigan Museum of Zoology). Romer, AS and Parsons, TS (1986), 'The Vertebrate Body, Saunders', (Philadelphia, Saunders College Publishing). Rönn, J., Katvala, M., and Arnqvist, G. (2007), 'Coevolution between harmful male genetalia and female resistance in seed beetles', Proc Natl Acad Sci U S A, 104, 10921-25. Rosenheim, Jay A (1999), 'The relative contributions of time and eggs to the cost of reproduction', Evolution, 53 (2), 376-85. Rowe, Locke and Arnqvist, Göran (2012), 'Sexual selection and the evolution of genital shape and complexity in water striders', Evolution, 66 (1), 40-54. Ryan, Michael J and Rand, A Stanley (1990), 'The sensory basis of sexual selection for complex calls in the túngara frog, Physalaemus pustulosus (sexual selection for sensory exploitation)', Evolution, 44 (2), 305-14. Ryan, Michael J, et al. (1990), 'Sexual selection for sensory exploitation in the frog Physalaemus pustulosus', Nature, 343 (6253), 66. Sanger, Thomas J, Gredler, Marissa L, and Cohn, Martin J (2015), 'Resurrecting embryos of the tuatara, Sphenodon punctatus, to resolve vertebrate phallus evolution', Biology letters, 11 (10), 20150694. Sasabe, M, Takami, Y, and Sota, T (2007), 'The genetic basis of interspecific differences in genital morphology of closely related carabid beetles', Heredity, 98 (6), 385-91. Schäfer, M. A., et al. (2011), 'Multiple quantitative trait loci influence intra-specific variation in genital morphology between phylogenetically distinct lines of Drosophila montana', Journal of Evolutionary Biology, 24 (9), 1879-86. Schliep, Klaus Peter (2010), 'phangorn: phylogenetic analysis in R', Bioinformatics, 27 (4), 592- 93. Schriefer, Jennifer L, et al. (2005), 'A comparison of mechanical properties derived from multiple skeletal sites in mice', Journal of biomechanics, 38 (3), 467-75. 116 Schulte-Hostedde, Albrecht I., Bowman, Jeff, and Middel, Kevin R. (2011), 'Allometry of the baculum and sexual size dimorphism in American martens and fishers (Mammalia: Mustelidae)', Biological Journal of the Linnean Society, 104 (4), 955-63. Schulte ‐Hostedde, Albrecht I, Bowman, Jeff, and Middel, Kevin R (2011), 'Allometry of the baculum and sexual size dimorphism in American martens and fishers (Mammalia: Mustelidae)', Biological Journal of the Linnean Society, 104 (4), 955-63. Schultz, Nicholas G, Otarola-Castillo, Erik, and Dean, Matthew D (2017), 'Dissection, MicroCT scanning and morphometric analyses of the Baculum', Journal of visualized experiments: JoVE, (121). Schultz, Nicholas G, et al. (2016a), 'The Baculum was Gained and Lost Multiple Times during Mammalian Evolution', Integrative and comparative biology, icw034. Schultz, Nicholas G, et al. (2016b), 'The Genetic Basis of Baculum Size and Shape Variation in Mice', G3: Genes| Genomes| Genetics, g3. 116.027888. Schultz, Willibrord Weijmar, et al. (1999), 'Magnetic resonance imaging of male and female genitals during coitus and female sexual arousal', Bmj, 319 (7225), 1596-600. Scimè, Anthony, et al. (2005), 'Rb and p107 regulate preadipocyte differentiation into white versus brown fat through repression of PGC-1α', Cell metabolism, 2 (5), 283-95. Sears, Karen E, Capellini, Terence D, and Diogo, Rui (2015), 'On the serial homology of the pectoral and pelvic girdles of tetrapods', Evolution, 69 (10), 2543-55. Seifert, Ashley W, Yamaguchi, Terry, and Cohn, Martin J (2009), 'Functional and phylogenetic analysis shows that Fgf8 is a marker of genital induction in mammals but is not required for external genital development', Development, 136 (15), 2643-51. Shapiro, Arthur M and Porter, Adam H (1989), 'The lock-and-key hypothesis: evolutionary and biosystematic interpretation of insect genitalia', Annual review of entomology, 34 (1), 231-45. Shapiro, Michael D, et al. (2004), 'Genetic and developmental basis of evolutionary pelvic reduction in threespine sticklebacks', Nature, 428 (6984), 717-23. Sharir, A, et al. (2011), 'The canine baculum: The structure and mechanical properties of an unusual bone', Journal of structural biology, 175 (3), 451-56. Shi, Jeff J and Rabosky, Daniel L (2015), 'Speciation dynamics during the global radiation of extant bats', Evolution. Shivtiel, Shoham, et al. (2008), 'CD45 regulates retention, motility, and numbers of hematopoietic progenitors, and affects osteoclast remodeling of metaphyseal trabecules', The Journal of experimental medicine, 205 (10), 2381-95. Simmons, Leigh W (2014), 'Sexual selection and genital evolution', Austral Entomology, 53 (1), 1-17. Simmons, Leigh W and Firman, Renée C (2014), 'Experimental evidence for the evolution of the mammalian baculum by sexual selection', Evolution, 68 (1), 276-83. Simson, S, Lavie, B, and Nevo, E (1993), 'Penial differentiation in speciation of subterranean mole rats Spalax ehrenbergi in Israel', Journal of Zoology, 229 (3), 493-503. Sinha, P. (1976), 'Bacula of Rajasthan bats', Mammalia, 40, 97-103. Skedros, JG, et al. (1996), 'Evidence of structural and material adaptation to specific strain features in cortical bone', The Anatomical Record: An Official Publication of the American Association of Anatomists, 246 (1), 47-63. Skedros, John G and Hunt, Kenneth J (2004), 'Does the degree of laminarity correlate with site ‐specific differences in collagen fibre orientation in primary bone? An evaluation in the turkey ulna diaphysis', Journal of anatomy, 205 (2), 121-34. Skedros, John G, et al. (2006), 'Dissociation of mineral and collagen orientations may differentially adapt compact bone for regional loading environments: results from acoustic velocity measurements in deer calcanei', Bone, 39 (1), 143-51. 117 Slice, Dennis E (2007), 'Geometric morphometrics', Annu. Rev. Anthropol., 36, 261-81. Smirnov, Dmitry G and Tsytsulina, Katerina (2003), 'The ontogeny of the baculum in Nyctalus noctula and Vespertilio murinus (Chiroptera: Vespertilionidae)', Acta chiropterologica, 5 (1), 117-23. Solberg Woods, Leah C (2013), 'QTL mapping in outbred populations: successes and challenges', Physiological genomics, 46 (3), 81-90. Sota, Teiji and Kubota, Kohei (1998), 'Genital lock ‐and ‐key as a selective agent against hybridization', Evolution, 52 (5), 1507-13. Spencer, Hamish G, McArdle, Brian H, and Lambert, David M (1986), 'A theoretical investigation of speciation by reinforcement', The American Naturalist, 128 (2), 241-62. Stemmler, Marc P, Hecht, Andreas, and Kemler, Rolf (2005), 'E-cadherin intron 2 contains cis- regulatory elements essential for gene expression', Development, 132 (5), 965-76. Stockley, P (2002), 'Sperm competition risk and male genital anatomy: comparative evidence for reduced duration of female sexual receptivity in primates with penile spines', Evolutionary Ecology, 16 (2), 123-37. Stockley, Paula (2012), 'The baculum', Current Biology, 22 (24), R1032-R33. Stockley, Paula, et al. (2013), 'Baculum morphology predicts reproductive success of male house mice under sexual selection', BMC biology, 11 (1), 66. Sturrock, Kenneth and Rocha, Jorge (2000), 'A multidimensional scaling stress evaluation table', Field methods, 12 (1), 49-60. Stutt, Alastair D and Siva-Jothy, Michael T (2001), 'Traumatic insemination and sexual conflict in the bed bug Cimex lectularius', Proceedings of the National Academy of Sciences, 98 (10), 5683-87. Sun, David, et al. (2018), 'Evaluation of loading parameters for murine axial tibial loading: Stimulating cortical bone formation while reducing loading duration', Journal of Orthopaedic Research®, 36 (2), 682-91. Takahara, Baku and Takahashi, Kazuo H. (2015), 'Genome-Wide Association Study on Male Genital Shape and Size in <italic>Drosophila melanogaster</italic>', PLoS ONE, 10 (7), e0132846. Tanaka, Kentaro M, Kamimura, Yoshitaka, and Takahashi, Aya (2018), 'Mechanical incompatibility caused by modifications of multiple male genital structures using genomic introgression in drosophila', Evolution. Tanaka, Kentaro M., et al. (2015), 'Genetic Architecture and Functional Characterization of Genes Underlying the Rapid Diversification of Male External Genitalia Between Drosophila simulans and Drosophila mauritiana', Genetics, 200 (1), 357-69. Tarazona, Oscar A, et al. (2016), 'The genetic program for cartilage development has deep homology within Bilateria', Nature, 533 (7601), 86-89. Tasikas, Diana E, et al. (2009a), 'Baculum variation and allometry in the muskrat (Ondatra zibethicus): a case for sexual selection', Evolutionary Ecology, 23 (2), 223-32. Tasikas, DianaE, et al. (2009b), 'Baculum variation and allometry in the muskrat (Ondatra zibethicus): a case for sexual selection', Evolutionary Ecology, 23 (2), 223-32. Taylor, B. A., Heiniger, H. J., and Meier, H. (1973), 'Genetic analysis of resistance to cadmium- induced teticular damage in mice ', Proc Soc Exp Biol Med, 143, 629-33. Taylor, Benjamin A, et al. (1999), 'Genotyping new BXD recombinant inbred mouse strains and comparison of BXD and consensus maps', Mammalian genome, 10 (4), 335-48. Thewissen, JGM, et al. (2006), 'Developmental basis for hind-limb loss in dolphins and origin of the cetacean bodyplan', Proceedings of the National Academy of Sciences, 103 (22), 8414-18. Thomas, Oldfield (1915), 'XXXIV.—The penis-bone, or “Baculum,” as a guide to the classification of certuin squirrels', Journal of Natural History, 15 (88), 383-87. 118 Toner, James P. and Adler, Norman T. (1986), 'The pre-ejaculatory behavior of male and female rats affects the number of sperm in the vagina and uterus', Physiol Behav, 36 (2), 363-67. Toner, James P., Attas, Amy I., and Adler, Norman T. (1987), 'Transcervical sperm transport in the rat: the roles of pre-ejaculatory behavior and copulatory plug fit', Physiol Behav, 39 (3), 371-75. Tong, Xin, Zhong, Wen, and Hua, Bao ‐Zhen (2018), 'Copulatory mechanism and functional morphology of genitalia and anal horn of the scorpionfly Cerapanorpa dubia (Mecoptera: Panorpidae)', Journal of Morphology. Trapnell, C., Pachter, L., and Salzberg, S. L. (2009), 'TopHat: discovering splice junctions with RNA-Seq', Bioinformatics, 25 (9), 1105-11. True, John R, et al. (1997), 'Quantitative genetic analysis of divergence in male secondary sexual traits between Drosophila simulans and Drosophila mauritiana', Evolution, 816- 32. Tschopp, Patrick, et al. (2014), 'A relative shift in cloacal location repositions external genitalia in amniote evolution', Nature, advance online publication. Tumlinson, JH, et al. (1974), 'Sex pheromones and reproductive isolation of the lesser peachtree borer and the peachtree borer', Science, 185 (4151), 614-16. Tunster, Simon James, Jensen, Anna B, and John, Rosalind Margaret (2013), 'Imprinted genes in mouse placental development and the regulation of fetal energy stores', Reproduction, 145 (5), R117-R37. Turner, CL (1950), 'The skeletal structure of the gonopodium and gonopodial suspensorium of Anableps anableps', Journal of Morphology, 86 (2), 329-65. Urbanek, Simon (2013), 'tiff: Read and write TIFF images', R package version 0.1-5. Available at https://CRAN. R-project. org/package= tiff (accessed December 10, 2016). Van der Schoot, P, Van Ophemert, J, and Baumgarten, R (1992), 'Copulatory stimuli in rats induce heat abbreviation through effects on genitalia but not through effects on central nervous mechanisms supporting the steroid hormone-induced sexual responsiveness', Behavioural brain research, 49 (2), 213-23. Vanderschueren, Dirk, et al. (2004), 'Androgens and bone', Endocrine reviews, 25 (3), 389-425. Venkov, Christo, et al. (2011), 'Transcriptional networks in epithelial-mesenchymal transition', PLoS One, 6 (9), e25354. Visel, Axel, et al. (2009), 'ChIP-seq accurately predicts tissue-specific activity of enhancers', Nature, 457 (7231), 854. Volpon, Jose Batista, et al. (2014), 'Torsional force applied to the tibia of living lambs in an attempt to change the bone rotational axis', Acta Cirurgica Brasileira, 29 (3), 193-200. Waage, Jonathan K (1979), 'Dual function of the damselfly penis: sperm removal and transfer', Science, 203 (4383), 916-18. Wade, Otis and Gilbert, Paul T (1940), 'The baculum of some Sciuridae and its significance in determining relationships', Journal of Mammalogy, 21 (1), 52-63. Wang, Xusheng, et al. (In Press), 'Joint mouse–human phenome-wide association to test gene function and disease risk', Nature communications. Wang, Yan, et al. (2001), 'Modeling human congenital disorder of glycosylation type IIa in the mouse: conservation of asparagine-linked glycan-dependent functions in mammalian physiology and insights into disease pathogenesis', Glycobiology, 11 (12), 1051-70. Wang, Ying, et al. (2007), 'The hypoxia-inducible factor α pathway couples angiogenesis to osteogenesis during skeletal development', Journal of Clinical Investigation, 117 (6), 1616. Warot, X., et al. (1997), 'Gene dosage-dependent effects of the Hoxa-13 and Hoxd-13 mutations on morphogenesis of the terminal parts of the digestive and urogenital tracts', Development, 124 (23), 4781-91. 119 Waterston, R. H., et al. (2002), 'Initial sequencing and comparative analysis of the mouse genome', Nature, 420 (6915), 520-62. Weatherhead, Patrick J and Robertson, Raleigh J (1979), 'Offspring quality and the polygyny threshold:" the sexy son hypothesis"', The American Naturalist, 113 (2), 201-08. Weimann, Bill, Edwards, Mark A, and Jass, Christopher N (2014), 'Identification of the baculum in American pika (Ochotona princeps: Lagomorpha) from southwestern Alberta, Canada', Journal of Mammalogy, 95 (2), 284-89. Wen, Jun, et al. (2010), 'Loss of Dact1 disrupts planar cell polarity signaling by altering dishevelled activity and leads to posterior malformation in mice', Journal of Biological Chemistry, 285 (14), 11023-30. Wigby, Stuart and Chapman, Tracey (2005), 'Sex peptide causes mating costs in female Drosophila melanogaster', Current Biology, 15 (4), 316-21. Williams-Ashman, HG and Reddi, AH (1991), 'Differentiation of mesenchymal tissues during phallic morphogenesis with emphasis on the os penis: roles of androgens and other regulatory agents', The Journal of steroid biochemistry and molecular biology, 39 (6), 873-81. Wilson, Don E and Reeder, DeeAnn M (2005), Mammal species of the world: a taxonomic and geographic reference (3rd edn., 12: JHU Press). Wulff, Nadja C, et al. (2017), 'Copulatory courtship by internal genitalia in bushcrickets', Scientific reports, 7, 42345. Yamaguchi, T. P., et al. (1999), 'A Wnt5a pathway underlies outgrowth of multiple structures in the vertebrate embryo', Development, 126 (6), 1211-23. Yang, Hyuna, et al. (2011), 'Subspecific origin and haplotype diversity in the laboratory mouse', Nat Genet, 43 (7), 648-55. Yoon, Myung Hee, Ando, Koichi, and UCHIDA, Teru Aki (1990), 'Taxonomic validity of scientific names in Japanese Vespertilio species by ontogenetic evidence of the penile pseudobaculum', Journal of the Mammalogical Society of Japan, 14 (2), 119-28. Zákány, J. and Duboule, D. (1999), 'Hox genes in digit development and evolution', Cell Tissue Res, 296 (1), 19-25. Zákány, József, et al. (1997), 'Regulation of number and size of digits by posterior Hox genes: a dose-dependent mechanism with potential evolutionary implications', Proceedings of the National Academy of Sciences, 94 (25), 13695-700. Zeng, Z. B., et al. (2000), 'Genetic architecture of a morphological shape difference between two Drosophila species', Genetics, 154 (1), 299-310. Zuk, Marlene and Stoehr, Andrew M (2002), 'Immune defense and host life history', the american naturalist, 160 (S4), S9-S22.

Abstract (if available)

Abstract For centuries, biologists have been captivated by the rapid evolution and extreme morphological diversity of external genitalia. One of the most mysterious and widely studied genital characters in mammals is the penis bone, or baculum. Numerous studies have sought to explain the function of the baculum, yet to date very little is known about its role in reproduction. Unveiling the baculum’s influence on male reproductive fitness requires investigations into the genetics, evolutionary history, and microstructure of this bone. ❧ In this thesis, I investigate baculum evolution from all three of the above perspectives. In chapter two, I use an established genetic resource in mice to uncover loci influencing shape and size variation in bacula. In chapter three, I show that the baculum has likely evolved multiple independent times using a phylogenetic approach. Finally, in chapter four, I characterize histological features of the mouse baculum, providing insight into the biomechanical function of the bone.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The evolution of gene regulatory networks

PDF

Investigating the potential roles of three mammalian traits in female reproductive investment

PDF

Studies in bivalve aquaculture: metallotoxicity, microbiome manipulations, and genomics & breeding programs with a focus on mutation rate

PDF

From gamete to genome: evolutionary consequences of sexual conflict in house mice

PDF

Exploring the genetic basis of quantitative traits

PDF

Genetic architectures of phenotypic capacitance

PDF

Male-female conflict after mating: function and dynamics of the copulatory plug in mice (Mus domesticus)

PDF

Evolutionary genomic analysis in heterogeneous populations of non-model and model organisms

PDF

Complex mechanisms of cryptic genetic variation

PDF

The complex genetic and molecular basis of oxidative stress tolerance

PDF

Genome sequencing and transcriptome analysis of the phenotypically plastic spadefoot toads

PDF

Applications of next generation sequencing in sessile marine invertabrates

PDF

Understanding genetics of traits critical to the domestication of crops using Mixed Linear Models

PDF

Genetic and molecular insights into the genotype-phenotype relationship

PDF

Genome-scale insights into the underlying genetics of background effects

PDF

Robustness and stochasticity in Drosophila development

PDF

Genetic engineering of fungi to enhance the production and elucidate the biosynthesis of bioactive secondary metabolites

PDF

Evolutionary mechanisms responsible for genetic and phenotypic variation

PDF

The evolution and functional significance of the cetacean pelvic bones

PDF

Developing genetic tools to assist in the domestication of giant kelp

Asset Metadata

Creator Schultz, Nicholas Gene (author)

Core Title Understanding the genetics, evolutionary history, and biomechanics of the mammalian penis bone

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Publication Date 04/25/2019

Defense Date 03/05/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag bacula,baculum,bone,Evolution,OAI-PMH Harvest,sexual selection

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Dean, Matthew (committee chair), Ehrenreich, Ian (committee member), Huttenlocker, Adam (committee member), Nuzhdin, Sergey (committee member), Patel, Biren (committee member)

Creator Email ngschult@usc.edu,nickschultzusc@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-144030

Unique identifier UC11676860

Identifier etd-SchultzNic-7249.pdf (filename),usctheses-c89-144030 (legacy record id)

Legacy Identifier etd-SchultzNic-7249.pdf

Dmrecord 144030

Document Type Dissertation

Format application/pdf (imt)

Rights Schultz, Nicholas Gene

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA