Viral Host Shifts: Ecological Dynamics, Cross-Species Transmission

VIRAL HOST SHIFTS: ECOLOGICAL DYNAMICS, CROSS-SPECIES TRANSMISSION

AND HOST ADAPTATION IN BAT RABIES

DANIEL GREGORY STREICKER

(Under the Direction of Sonia Altizer and Pejman Rohani)

ABSTRACT

Newly emerging infectious diseases present some of the most pressing challenges facing human health, wildlife conservation and veterinary medicine. A disproportionate number of emerging pathogens are RNA viruses that originate in other species through a process termed ‘cross- species transmission.’ Mechanisms underlying viral emergence in new species remain poorly understood, despite the importance of emergence events for human and wildlife health. This dissertation explores the ecological and evolutionary factors that contribute to the emergence of rabies virus in new host species using bats as a model system. By constructing large datasets of viral sequences from North and South American bat species, I first examined the effects of host ecology and phylogeny on rates of cross-species transmission and viral evolution. Next, I combined evolutionary and demographic inference to demonstrate links between the extent of adaptive evolution associated with the establishment of rabies virus in new bat species and the speed of emergence in new bat species. Through a field study, I examined the transmission dynamics of rabies within populations of a single species, common vampire bats, to enhance prospects for rabies prevention in humans and domesticated animals. Together, these findings develop a framework for dissecting viral host shifts into quantifiable sequential processes and

constitute a step towards the ultimate goal of predicting which host shifts are most likely to occur and what measures can be taken to prevent them.

INDEX WORDS: Host Shift, Anthropogenic Change, Phylogenetic Signal, Rabies Virus, Chiroptera, Common Vampire Bat, Wildlife, Adaptive Evolution

VIRAL HOST SHIFTS: ECOLOGICAL DYNAMICS, CROSS-SPECIES TRANSMISSION

AND HOST ADAPTATION IN BAT RABIES

DANIEL GREGORY STREICKER

B.A., University of Virginia, 2004

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2011

Daniel Gregory Streicker

VIRAL HOST SHIFTS: ECOLOGICAL DYNAMICS, CROSS-SPECIES TRANSMISSION

AND HOST ADAPTATION IN BAT RABIES

DANIEL GREGORY STREICKER

Major Professors: Sonia Altizer Pejman Rohani

Committee: Steven Castleberry Charles Rupprecht John Wares

Electronic Version Approved:

Maureen Grasso Dean of the Graduate School The University of Georgia December 2011

DEDICATION

For mom, dad, sarah and pin.

ACKNOWLEDGEMENTS

I am extremely grateful to my colleagues, collaborators and fellow graduate students not only for their technical critiques of my work, but also to their contributions to my professional and personal development during my time at UGA. I express most sincere thanks to Janis

Antonovics, Amy Pedersen, Michael Hood and the other members of the Antonovics lab at the

University of Virginia for first exposing me to many of the ideas in infectious disease ecology that have evolved into central themes of my dissertation and for providing a social and intellectual standard for academia that I will strive to replicate and build upon throughout my career. I owe tremendous thanks also to Charles Rupprecht and the members of the CDC rabies team for welcoming an ecologist into their fold and for providing a diverse environment and freedom to explore new ideas. At UGA, I wish to thank the members of the Altizer, Rohani and

Ezenwa lab groups for providing a constructive, friendly and stimulating environment. Much of this work involved international collaboration with members of both governmental and academic institutions. Without the help and support of Victor Pacheco, Jorge Gomez and William

Valderrama none of my work in Peru would have been possible. I am indebted particularly to the numerous field assistants, including Fernando Pancorbo, Carlos Tello, Oscar Centty, Jorge

Carrera and Liz Huamani, who worked tremendously for many late nights and early mornings in the most difficult of field conditions. Finally, I thank my committee, which includes Charles

Rupprecht, Steven Castleberry, John Wares, Pej and Sonia for providing just the right ratio of support and independence.

This work was funded by a National Science Foundation Graduate Research Fellowship and a University of Georgia Dissertation Completion award; small grants from National

Geographic, the American Philosophical Society, the Odum School of Ecology and the Latin

American and Caribbean Studies Institute; a seed grant from the CDC/UGA Research

Collaboration in Infectious Disease (FID 908) and a research grant from the National Science

Foundation (DEB 1020966).

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... v

CHAPTER

1 INTRODUCTION AND LITERATURE REVIEW ...... 1

2 HOST PHYLOGENY CONSTRAINS CROSS-SPECIES EMERGENCE AND

ESTABLISHMENT OF RABIES VIRUS IN BATS ...... 8

3 RATES OF VIRAL EVOLUTION ARE LINKED TO HOST LIFE HISTORY IN

BAT RABIES ...... 19

4 SELECTION LINKED TO WITHIN HOST SPREAD DRIVES RABIES VIRUS

EMERGENCE IN NEW HOST SPECIES ...... 39

5 ECOLOGICAL AND ANTHROPOGENIC DRIVERS OF RABIES EXPOSURE IN

VAMPIRE BATS: IMPLICATIONS FOR TRANSMISSION AND CONTROL .....66

6 CONCLUDING REMARKS AND FUTURE DIRECTIONS ...... 92

LITERATURE CITED ...... 97

APPENDICES

A SUPPORTING INFORMATION FOR CHAPTER 2 ...... 113

B SUPPORTING INFORMATION FOR CHAPTER 3 ...... 147

C SUPPORTING INFORMATION FOR CHAPTER 4 ...... 156

vii

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

“At the most fundamental level, the dynamics of infectious diseases are determined by

transmission between infected and susceptible individuals: who infects who, and how often.

Ironically, this is a level of organization at which studying epidemiology is hard.”

Dan Haydon and Louise Matthews (2007)

The majority of animal pathogens infect multiple host species. Recognition of this phenomenon has transformed the study of infectious disease ecology from its single-host single-pathogen roots, and has simultaneously uncovered clear links between the health of humans, domesticated animals and wildlife (Anderson & May 1991; Cleaveland, Laurenson & Taylor 2001; Aguirre

2002; Fenton & Pedersen 2005; Pedersen et al. 2005). It is now well understood that most newly emerging infectious diseases originate in other host species (Daszak, Cunningham & Hyatt 2000;

Woolhouse & Gaunt 2007). The transmission of a pathogen from one host species to another

(‘cross-species transmission’) is therefore a defining process of disease emergence and one that has critical importance to human health and wildlife conservation. Despite this growing awareness, studies that integrate across the entire process of emergence, from the initial cross- species transmission to the ecological and evolutionary dynamics that ensue in the recipient species remain “dismayingly rare” and a “black box” in the study of disease emergence (Lloyd-

Smith et al. 2009; Daszak 2010). Consequently, factors that determine pathogen transmission

between species and the evolutionary forces that allow some pathogens to overcome barriers to establish in new host species remain untested.

In spite of these empirical gaps, the theoretical underpinnings of disease emergence by cross-species transmission have a solid foundation. First, it is critical to recognize that disease emergence encompasses a spectrum of possible outcomes following cross-species transmission

(Wolfe, Dunavan & Diamond 2007). These outcomes depend on how much transmission occurs within populations of the recipient species, with some pathogens typically causing only a single

“spillover” infection (e.g., rabies virus or West Nile virus in humans), others causing short chains of transmission that are bound for extinction (e.g., human monkeypox) and others successfully establishing in new host species (e.g., HIV) (Lloyd-Smith et al. 2009).

To date, no clear set of rules has been established to predict the occurrence and outcome of cross-species transmission, but modeling work and case studies point to biological, ecological and evolutionary barriers to emergence (Woolhouse, Haydon & Antia 2005; Parrish et al. 2008).

Biologically, the pathogen must be capable of infecting and replicating within a new host species. This may involve a variety of barriers including incompatible receptors for cell entry or exit, failure to replicate and spread between tissues, evolution of optimal virulence for onward transmission, or host-specific intracellular host defenses such as interferon responses (Kuiken et al. 2006; Parrish et al. 2008). From an ecological standpoint, the donor host species must overlap with a putative recipient species and a mechanism for cross-species transmission (e.g., environmental contamination, vector-borne transmission, direct contact) must exist (Pulliam

2008). The relevance of ecological opportunities for cross-species transmission implies a potential role for human activity in shaping disease emergence (Daszak, Cunningham & Hyatt

2001). Processes such as urbanization, deforestation and agricultural intensification can break

down historical barriers between species to generate novel opportunities for cross-species transmission and can homogenize host communities in ways that alter patterns of pathogen transmission within the donor species (Daszak, Cunningham & Hyatt 2001; Wright & Gompper

2005; Bradley & Altizer 2007). Finally, pathogens must maintain chains of transmission within populations of the new host species, a process that is likely to involve a delicate balance between evolutionary host-adaptation and pathogen population dynamics (Kuiken et al. 2006). Both theory and empirical studies indicate that rapid evolution can allow pathogens that initially have low fitness to establish long-term or permanent transmission cycles in new host species (Antia et al. 2002; Shackelton et al. 2005; Song et al. 2005; Anishchenko et al. 2006). Nevertheless, the extent of genomic change and the physiological characteristics of infection that must be manipulated to enable host adaptation remain a mystery in the vast majority of actualized and potential host shift events.

Of all pathogens, RNA viruses are the most notoriously successful at crossing host species barriers (Woolhouse 2002). Several high profile examples of RNA virus emergence in humans include SARS and avian influenza, but more generally, these viruses comprise over half of new human pathogens since 1980 (Guan et al. 2003; Kuiken et al. 2006; Woolhouse & Gaunt

2007). Much of the success of RNA viruses has been attributed their capacity for rapid molecular evolution, a product of their large within-host population sizes and error-prone RNA polymerases (Moya, Holmes & Gonzalez-Candelas 2004). Rates of RNA virus evolution are almost universally orders of magnitude greater than those of their hosts with many viruses accumulating 0.1 – 1 mutations per genome per replication (Drake & Holland 1999; Duffy,

Shackelton & Holmes 2008). This extreme genomic plasticity plays a major role in generating the phenotypic variation that allows viruses to evade host immunity, resist anti-viral drugs and

vaccines and overcome the defenses of new host species (Holmes 2009; Holmes 2011). Hence, understanding the determinants of the tempo of RNA virus evolution has been a topic of increasing interest as a factor that may contribute to the outcome of viral emergence (Duffy,

Shackelton & Holmes 2008). Work to date has focused on the effects of genomic features of viruses (e.g., genome length, single vs. double stranded RNA) and aspects of viral transmission cycles (e.g., latency, transmission route) on evolutionary rates, with surprisingly few efforts to understand how hosts and viruses may interact to set viral evolutionary rates (Jenkins et al. 2002;

Chen & Holmes 2006). Given that some of the most important human pathogens (e.g., SARS, influenza, HIV) have close relatives in not one, but several wildlife reservoirs, understanding how the ecological differences among these hosts are translated into the evolutionary dynamics of their viruses could provide insights into the key question of which host of a multi-host virus may trigger cross-species emergence.

Bats (Chiroptera) are one group of host species that has received a great deal of attention as source of emerging zoonotic RNA viruses. Within the last decade, the emergence of highly pathogenic viruses such as SARS coronavirus, Nipah virus, Ebola virus and Marburg virus in humans and/or domesticated animals have each been attributed to this group (Leroy et al. 2005;

Li et al. 2005; Breed et al. 2006; Swanepoel et al. 2007). Importantly, fundamental gaps remain in our understanding of virus transmission in wild bat populations and the process of cross- species pathogen transmission among bats and from bats to other species (Calisher et al. 2006;

Omatsu et al. 2007). One recurrent theme in the emergence of viruses from bats is the involvement of anthropogenic change. For some viruses, such as SARS and Ebola, emergence is likely the product of human encroachment into pristine environments and the corresponding greater contact rates with wildlife through trade or hunting (Leroy et al. 2005; Li et al. 2005;

Wolfe et al. 2005). In other instances, agricultural intensification has been implicated as a force that might both increase contact rates between bats, humans and domesticated animals and fundamentally alter patterns of virus transmission within bat populations. The best understood example of the interaction between bats and agricultural intensification is the emergence of

Nipah virus in Malayasia (Field et al. 2001). There, planting of mango trees near pig farms promoted the aggregation large numbers of bats in close proximity to pigsties, eventually enabling transmission from bats to pigs to humans (Pulliam et al. 2011). The emergence of

Hendra virus in Australia may represent an analogous case of emergence driven by the ability of bats to exploit human driven environmental change in the form of urbanization (Plowright et al.

2011). In both cases, shifts in transmission dynamics following exploitation of human resources are plausible, but this remains a key area for research.

Rabies virus (Lyssavirus, Rhabdoviridae), common in wild bat populations throughout the Americas, is the best-studied and arguably most important zoonotic virus of bats, annually causing thousands of costly human exposures, untold numbers of domesticated animal deaths, and increasingly frequent human infections (Messenger, Smith & Rupprecht 2002; WHO 2005;

Schneider et al. 2009). In bats and all other mammals, untreated infection causes an acute encephalitis with a case fatality ratio approaching 1, the highest of any infectious disease

(Rupprecht, Hanlon & Hemachudha 2002). In addition to causing sporadic spillover infections, bat rabies has demonstrated a remarkable ability to seed transmission cycles new species, both within the Chiroptera and to other mammalian orders such as Carnivora (Leslie et al. 2006;

Arechiga-Ceballos et al. 2010; Streicker et al. 2010). Bat rabies therefore represents a uniquely suited system for elucidating the fundamental processes involved in cross-species viral emergence and for clarifying the general maintenance mechanisms of bat-borne viruses, but also

one where improved epidemiological and evolutionary understanding may substantially benefit human and animal health (Calisher et al. 2006).

As outlined above, scientific investigation of cross-species transmission and disease emergence for rabies and other pathogens must integrate across the ecological and evolutionary sciences. Specifically, viral emergence in new species is a three stage process that is mediated by: (i) the frequency of virus transmission between different species, (ii) the pre-existing viral factors that contribute to the probability of ongoing transmission in new host species and (iii) the joint evolutionary and epidemiological dynamics that allow viral adaptation and permanent establishment in new species. A fourth stage in this process for some viral emergence events is the preventative or reactionary response of humans to lessen the burden of infectious disease.

The structure of this dissertation largely parallels the different stages of viral emergence.

Chapter 2 explores the influence of host ecology and evolutionary relatedness on the frequency of rabies virus transmission between 15 North American bat species. This work provides a novel framework to estimate rates of cross-species transmission in complex host communities using ecological and genetic data. Results demonstrate that host genetic relatedness predicts both initial spillover transmission and the probability of establishment in new host species. Chapter 3 considers how differences among many bat reservoir species can influence rates of molecular evolution in rabies virus. These results establish that viral evolutionary rates differ among hosts, and that bat life history (as driven by environmental variation) can influence transmission dynamics. Chapter 4 focuses more explicitly on rabies virus as a case study of repeated host shift events by the same virus into different host species to identify evolutionary mechanisms that allow successful establishment. As a step towards linking the adaptive evolutionary and epidemiological dynamics of host shifts, Chapter 4 also lays a foundation for understanding how

emergence in a new host species depends on the extent of viral genomic change. Chapter 5 examines how more detailed understanding of the transmission dynamics within a single key reservoir species can help predict and prevent future cross-species transmission to humans and domesticated animals. Here, I report results from the first 4 years of an ongoing field study of rabies in common vampire bats (Desmodus rotundus) in Peru. This study identifies population and individual-level factors that contribute to rabies exposure bats, with the ultimate goal of using this information to develop effective control measures to mediate the growing problem that this disease is presenting in Latin America.

CHAPTER 2

HOST PHYLOGENY CONSTRAINS CROSS-SPECIES EMERGENCE AND

ESTABLISHMENT OF RABIES VIRUS IN BATS 1

______

1 Streicker, D.G., Turmelle, A.S., Vonhoff, M.J., Kuzmin, I., McCracken, G.F. and Rupprecht, C.E. 2010. Science. 329: 676-679. doi: 10.1126/science.1188836. Reprinted here with permission of the publisher.

ABSTRACT

For RNA viruses, rapid viral evolution and the biological similarity of closely related host species have been proposed as key determinants of the occurrence and long-term outcome of cross-species transmission. Using a dataset of hundreds of rabies viruses sampled from 23 North

American bat species, we present a general framework to quantify per capita rates of cross- species transmission and reconstruct historical patterns of viral establishment in new host species using molecular sequence data. These estimates demonstrate diminishing frequencies of both cross-species transmission and host shifts with increasing phylogenetic distance between bat species. Evolutionary constraints on viral host range indicate that host species barriers may trump the intrinsic mutability of RNA viruses in determining the fate of emerging host-virus interactions.

INTRODUCTION, RESULTS AND DISCUSSION

In recent decades, cross-species transmission (CST) of RNA viruses has resulted in a range of disease emergence outcomes (Lloyd-Smith et al. 2009), from single infection “spillover” events such as rabies virus infections in humans (Rupprecht, Hanlon & Hemachudha 2002), to transient outbreaks bound for extinction such as Nipah virus (Hsu et al. 2004), to sustained epidemics with potential for endemic establishment such as SARS Coronavirus (Riley et al. 2003). Though critical to anticipating the impact of viral emergence on human and animal health, the factors that determine the frequency and outcome of CST remain obscure. In RNA viruses, evidence for high mutation rates and occasional human epidemics originating from distantly related species have popularized the view that rapid evolution allows these viruses to overcome host-specific barriers in cellular, molecular or immunological defenses (Moya, Holmes & Gonzalez-Candelas 2004).

Consequently, it has been argued that RNA viruses emerge primarily between species with high contact rates (Song et al. 2005; Anishchenko et al. 2006; Parrish et al. 2008). An alternative explanation posits that innate similarity in the defenses of closely related species may favor virus exchange by flattening the fitness valley that viruses traverse during adaptation to new hosts

(Kuiken et al. 2006).

Identifying the most important determinants of viral emergence requires considering how the ecological dynamics of CST interact with evolutionary factors to shape replicated patterns of viral establishment in natural communities. Rabies, a ubiquitous, multi-host viral zoonosis, provides this opportunity. In the USA, bats (Chiroptera) are the most common source of indigenously acquired human rabies infections, and approximately 2,000 rabies positive bats are collected annually following human or domesticated animal exposures (Blanton et al. 2009).

Transmission occurs mainly by bite, and infection causes encephalitis with behavioral and motor abnormalities prior to death (Brass 1994a). Importantly, the phylogeny of rabies virus in North

American bats is structured by host species, reflecting an evolutionary history of host shifts followed by predominately within-species transmission (Smith, Orciari & Yager 1995; Hughes,

Orciari & Rupprecht 2005). This species-association of viral lineages enables identification of the species origins of relatively rare CST events from bats to humans or domesticated animals or within the bat community (Blanton et al. 2009). Because North American bats span evolutionary divergences of approximately 3 to 60 million years, a substantial range of ecological and physiological differences exist among species that might influence viral emergence (Bininda-

Emonds et al. 2007).

We sequenced the nucleoprotein gene of 372 rabies viruses from 23 bat species collected across the continental USA over a 10-year period (Figure 2.1A, Appendix A Table 1). Bayesian

and maximum likelihood (ML) analyses revealed 18 phylogenetic lineages of rabies virus that were statistically compartmentalized to particular bat taxa (Figure 2.1B, Appendix A Table 2).

New viral lineages were discovered in Lasiurus intermedius floridanus (LiV), L. seminolus

(LsV) and Myotis yumanensis (MyV) establishing each as an independent rabies virus reservoir.

The host-specificity of most viral lineages allowed us to infer the species origin of 360 infections in the dataset after confirming the taxonomic identities of bats with mitochondrial DNA sequencing (Appendix A Table 3). Forty-three unambiguous CST events were observed, involving 15 bat species and 26 unique species pairs. Nearly all viruses from cross-species infections were tightly nested within source clades and no more genetically divergent than donor lineage viruses (Appendix A Table 4), suggesting that they were more likely dead-end infections than infections occurring within stuttering chains of transmission in the recipient species.

We applied Markov chain Monte Carlo (MCMC) simulation to viral genetic data to compare four models of the strength and direction of CST between species pairs: symmetrical, bidirectional transmission; asymmetrical, bidirectional transmission; and each case of unidirectional transmission. Models selected by Akaike’s information criterion were exclusively asymmetrical and predominately unidirectional (21/26), suggesting unequal probability of infection for a given inter-specific contact rate. Using parameters estimated from MCMC simulations (Appendix A Table 5), we quantified the expected number of infections in species i resulting from a single infected individual of species j (the per capita CST rate, Rij) and visualized these in a “transmission web” (Figure 2.2). Depending on species, a single rabid bat may infect between 0 and 1.9 heterospecifics, and on average, CST occurs once for every 72.8 within-species transmission events.

We next explored the intensity of CST between bat species pairs as a function of their ecological overlap (i.e., similarity in foraging behavior, roosting strategy and body length), geographic range overlap and phylogenetic relatedness using host trait values estimated from our data and the literature (Appendix A Table 6). The intensity of Rij declined continuously with the genetic distance between donor and recipient species and increased to a lesser extent with the amount of geographic overlap between species (Figure 2.3A); however, our ecological proxies of inter-species contact failed to predict CST (Appendix A Tables 7 and 8). Results were robust to exclusion of several viruses for which the taxonomic identity of the host was based on morphology alone (F2,25 = 9.38, P < 0.001; Appendix A Table 4 lists exclusions). Finally, a reanalysis of the transmission web using a novel metric of connectance from food web theory (the proportion of realized inter-specific connections in a food web) illustrated that rates of CST were highest to and from bat species that are sympatric with many closely related species but independent of the viral genetic diversity within the donor clade and the sampling effort for each

2 bat species (supporting online text, F2,13 = 12.24, r = 0.67, P = 0.001). These results suggest that initial infection of a new species is facilitated by evolutionary conservation of the cellular, immunological or metabolic traits of hosts, with secondary effects of probabilistic factors, perhaps including exposures involving high viral load, that increase with species’ range overlap.

In light of the host specificity implied by compartmentalization tests and phylogenetic analyses (Appendix A Table 2, Figure 2.1B), the high rates of CST shown here indicate that the vast majority of cross-species infections are evolutionary dead ends. Nevertheless, it is clear that rabies virus has successfully established repeatedly in North American bat species (Hughes,

Orciari & Rupprecht 2005). This observation prompts the critical question of what determines whether CST causes a dead-end infection or sustained transmission in recipient species? We

tested whether historical host shifts share a common phylogenetic constraint to present-day CST using Bayesian ancestral state estimation of the host species origin of viral lineages. Strikingly, nearly all (22/23) host shifts occurred between bat species that were more closely related than the median pair, and 66% of host shifts occurred within the top 25% of the most closely related

North American bats, consistent with a lack of sustained transmission in distantly related species

(Figure 2.3B).

Phylogenetic signal in pathogen host range has been observed in fungal infections of heterospecific plants and in a database study of parasite community similarity in wild primates

(Gilbert & Webb 2007; Davies & Pedersen 2008; de Vienne, Hood & Giraud 2009). Although the consistency of host phylogeny as a predictor of emergence has been questioned for RNA viruses because of their potential for rapid within-host adaptation (Parrish et al. 2008), sufficient data to test this hypothesis have been unavailable until now. Our study demonstrates that rapid evolution can be insufficient to overcome phylogenetic barriers at two crucial stages of viral emergence: initial infection and sustained transmission.

The decline in CST that we observed among more distantly related bat species might result from lower inter-specific contact rates or a reduced probability of infection upon exposure.

Although we could only examine a small number of species traits for which data were available, we found no effect of ecological proxies of inter-specific contact on CST. This result is surprising given the infectiousness of rabies virus across mammals and abundant opportunities for CST among bats that share roosting and foraging sites. One explanation is that the disorientation and indiscriminate aggression caused by rabies infection could limit the selectivity of inter-specific contacts, causing their occurrence to depend on the frequency of host species sympatry (Brass 1994a). Importantly, our analysis supported both geographic overlap and host

phylogenetic distance as strong predictors of CST: these two factors probably determine the frequency of exposure and the likelihood of infection following exposure, respectively.

Two explanations could account for the elevated frequency of host shifts among closely related bats. First, similarity in the biological barriers and social structure of closely related species could minimize the amount of evolution required to achieve an optimal balance of within-host replication and viral shedding (Kuiken et al. 2006). Although rabies virus uses evolutionarily conserved receptors for cell entry, receptor density on susceptible cells varies among species, leading to variable resistance to infection that maladapted viruses must overcome

(Baer et al. 1990). Evolution of optimal virulence through modulation of transcription, gene expression and replication might also be needed to balance entry into the central nervous system

(ultimately leading to host death) with the timing and intensity of viral excretion from the salivary glands (necessary for transmission) to ensure sustained transmission within new host species (Finke & Conzelmann 2005). As a second limiting factor, even if the likelihood of viral establishment is independent of host phylogeny, viruses might still shift disproportionately between close relatives because of the greater frequency of CST. Although our results imply that rabies virus host shifts followed common, rather than rare CST, the overwhelming support for host phylogenetic distance as the principal predictor of initial infection argues more strongly for intrinsic features of the host-virus interaction as the primary barrier to emergence.

The repeated failure of a notoriously generalist virus to colonize bat species that are capable of enzootic maintenance highlights the limitation of viral evolution to overcome host species barriers within a mammalian order. Similar effects could be critical determinants of the host range of other infections of public health or veterinary concern such as lentiviruses in primates or morbilliviruses in carnivores. Nonetheless, the ultimate goal for predicting viral

emergence is to understand drivers across varying taxonomic scales. Future studies of viral host range could examine whether the phylogenetic barriers evident at relatively shallow evolutionary distances dissipate for more distantly related taxa, where all emergence events might be equally improbable and driven by the frequency of inter-specific contact.

Finally, we outlined a general framework to identify the origins of host shifts and quantify CST in complex, multi-host communities. A similar approach could be applied to any host-associated pathogen for which molecular sequence data are attainable. Quantification of per capita rates of pathogen transmission between species will be particularly useful to parameterize predictive models of viral emergence, which have traditionally ignored the process of cross- species transmission despite its importance as the defining feature of zoonoses (Lloyd-Smith et al. 2009). Models incorporating such information will be critical to test the efficacy of specific disease prevention strategies applied not only within donor and recipient communities, but also in the realm where they intersect.

FIGURES

LcV

●/* ●/* LsV .85/* LbV1 LbV2

●/*

.53/.88 ●/* ●/* .54/.89 LxLiV LnV .87/* ●/* ●/ PsV .84/ * .46/.71 * B ●/* LiV

EfV1 ●/* ●/* MyV ●/*

TbV .80/.96

●/* ●/* .76 ●/* .58 /* MspV .71 /.84 ●/* /* PhV E. fuscus (Ef) M. austroriparius (Ma) EfV2 L. seminolus (Ls) M. lucifugus complex (Ml) ●/* L. borealis (Lb) M. californicus complex (Mc) L. xanthinus (Lx) M. yumanensis (My) L. cinereus (Lc) N. humeralis (Nh) L. intermedius (Li) P. hesperus (Ph) EfV3 L. blossevillii (Lbl) P. subﬂavus (Ps) ●/* T. brasiliensis (Tb) L. noctivagans (Ln) A. pallidus (Ap) C. townsendii (Ct) 0.5

Figure 2.1. Geographic origins, phylogenetic relationships and host range of viral lineages. (A) Collection localities for 347/372 rabies virus samples; diamonds jittered randomly for visualization. (B) Bayesian phylogenetic tree with viral lineages labeled by donor host (Appendix A Table 3 contains full species names). MspV was associated with various Myotis species in the northwestern USA; LxLiV was associated with the western yellow bat (L. xanthinus) and the northern yellow bat (L. i. intermedius). Pie charts show the host species composition of lineages found in multiple species; pie diameter proportional to the number of bats sampled. ML bootstrap values (BV) > 0.50 and Bayesian posterior probability (PP) values > 0.70 are shown to the lineage level (BV/PP). Filled circles are BV ≥ 0.90; asterisks are PP ≥ 0.98. Root branch removed for clarity; dashed line indicates trunk.

Rij Mc < 0.01

0.01 - 0.05

0.05 - 0.10 Ml Ln 0.10 - 0.5

> 0.5 My

No. bats sampled Tb Ef

> 50

Lc Lb

Ap Ps 26-50

16-25 Li Lbl Nh 5-15

< 5 Lx Ls

Figure 2.2 Transmission web for 15 bat species. Pie charts describe the observed proportion of each species infected by cross-species transmission. Arrows show the direction of transmission between species; arrow width indicates per capita transmission rate (Rij). Abbreviations for bat species names follow Figure 2.1.

Per capita CST A (log10[Rij]+2.5) 1.5

2.0

1.0 1.5

1.0 0.5

0.5 Geographic range overlap (asin(sqrt[prop.])) overlap range Geographic

0.2 0.4 0.6 0.8 1.0 B 3.0 [BF]) 10 2.0

1.0 0.0 0.5 1.0 1.5

0.0

-1.0

-2.0

-3.0 Phylogenetic support for host shift support (log Phylogenetic

0.0 0.5 1.0 1.5 Phylogenetic distance between bats

Figure 2.3. Predictors of two stages of viral emergence. (A) Per capita cross-species transmission declines with the genetic distance (substitutions per site) between bat species (slope = -3.93, F1,28 = 21.89, P < 0.001) and increases with the proportion of their shared geographic 2 range (slope = 1.21, F1,28 = 8.22, P = 0.008; Full model: n = 31, r = 0.44; P < 0.001). Surface generated by bivariate interpolation. (B) The Bayes factor (BF) statistic describes the relative support for models containing versus lacking epidemiological linkages (i.e., historical host shifts) between each pair of viral lineages. Red circles indicate host shifts supported by ancestral state estimations (BF > 3), and open circles indicate host shifts that were inconsistent with phylogenetic data (BF < 3). Vertical lines show the median (blue) and lower and upper 25% limits (black) of the distribution of pairwise genetic distances between bats (blue histogram). Inset densities show the distributions of pairwise genetic distances for bat species implicated (red) or not implicated (black) in host shifts (t-test for difference of means: t = 4.57, P < 0.0001).

CHAPTER 3

RATES OF VIRAL EVOLUTION ARE LINKED TO HOST LIFE HISTORY IN BAT

RABIES 2

______

2 Streicker, D.G., Lemey, P., Velasco-Villa, A. and Rupprecht, C.E. Submitted to PLoS Biology, 11/30/11.

ABSTRACT

Rates of evolution span several orders of magnitude among RNA viruses and improved understanding of this variability can enhance efforts to mitigate viral emergence in new species, counter immune evasion and slow the rate of drug resistance. While differences in rates of evolution among viral families are typically ascribed to viral features such as mutation rates and transmission mode, these factors cannot explain variation at shallower taxonomic levels, where host biology might operate more strongly on viral evolution. Here, we analyze sequence data from 648 rabies viruses collected from 21 bat taxa from throughout the Americas over a 38-year time period to describe dramatic variation in the evolutionary rate of a single virus when isolated from different host species. We use newly developed Bayesian techniques for phylogenetic hypothesis testing to demonstrate that viral evolutionary rates are labile following host jumps between bat species and accelerated by aspects of bat life history predicted to enhance virus replication or transmission. Specifically, the viral molecular clock ticked nearly four times faster in tropical and subtropical bats relative to temperate species, an acceleration likely reflecting reduced seasonality in bat activity and virus transmission. These findings imply geographically variable maintenance strategies in a bat-borne zoonosis, offer host biology as a general factor to explain rates of viral evolution and suggest an evolutionary mechanism by which host identity might determine the emergence potential of infectious diseases.

INTRODUCTION

RNA viruses display exceptionally variable rates of molecular evolution, with up to 6 orders of magnitude in nucleotide substitution rates observed among viral species (Duffy, Shackelton &

Holmes 2008). Because of the importance of genetic and phenotypic evolution for colonizing

new host species, evading immune responses and obstructing successful pharmaceutical development, understanding the factors that drive rates of viral evolution is critical for mitigating viral emergence (Holmes 2009). To date, explanations for evolutionary rate heterogeneity have been predominately virus-oriented. These have focused on features of genomic architecture that determine underlying mutation rates, and aspects of the virus life cycle, such as latency and transmission mode, that can influence replication rate within hosts, generation time between hosts and diversification through positive selection (Jenkins et al. 2002; Woelk & Holmes 2002;

Hanada, Suzuki & Gojobori 2004; Holmes 2009). Less well understood are the determinants of evolutionary rate variation among closely related viral groups or species (Salemi et al. 1999;

Kurath et al. 2003). Because viral genomic features and replication mechanisms are minimally variable at such shallow taxonomic levels, aspects of host biology that influence rates of transmission and replication may be more likely to control the tempo of viral evolution.

Consistent with this hypothesis, several human viruses (HTLV, HIV and Chikungunya virus) that exploit multiple modes of transmission or experience variable immunological pressures within-hosts demonstrated accelerated molecular evolution in conditions associated with enhanced transmission and replication (Salemi et al. 1999; Edo-Matas et al. 2010; Volk et al.

2010).

The propensity of many RNA viruses to ‘jump’ between host species presents an intriguing natural experiment to test whether viral evolutionary rates change according to traits of host species that influence viral replication and transmission or remain evolutionarily conserved along the virus ancestral history, reflecting intrinsic biological features of viruses (Woolhouse &

Gowtage-Sequeria 2005; Kitchen, Shackelton & Holmes 2011). Despite the implications of accelerated viral evolution in certain host species for predicting the geographic and species

origins of viral emergence events, these ideas remain largely untested. In two recent analyses, influenza A viruses infecting wild and domesticated birds and infectious haematopoietic viruses of wild and farmed fish each showed intra-specific variation in viral evolutionary rates.

However, host species identity failed to explain these differences, perhaps because high rates of transmission between host groups in each system dilute the effects of any single species on virus evolution (Kurath et al. 2003; Chen & Holmes 2006). Elucidating the influence of host life history on viral evolution therefore requires large datasets of closely related viruses from multiple ecologically distinct host species that are capable of independent viral maintenance.

Rabies virus (Lyssavirus, Rhaboviridae) is a globally distributed and lethal zoonotic agent that infects over 80 bat species, representing 4 chiropteran families throughout the Americas

(Constantine 2009). Phylogenetic analyses indicate that bat rabies viruses have established numerous host species-associated viral lineages through a process of sequential host shifts through the bat community followed by predominantly within-species transmission (Hughes,

Orciari & Rupprecht 2005; Streicker et al. 2010). Coupled with the diverse life history strategies of bats, rabies therefore provides a unique opportunity to explore the effects of host ecology and behavior on virus evolution while explicitly considering the historical effects of viral ancestry on evolutionary rate. Moreover, because many American bat species and genera are broadly distributed with distinctive viral lineages in different parts of their geographic range, ecological effects that reflect geographic variation in life history strategies can be distinguished from taxonomic effects that arise from the physiological similarity of closely related host species.

Bats represent an especially pertinent taxonomic group for exploring the effects of host biology on viral evolution because of growing interest in how bat ecology influences zoonotic agents such as SARS virus, Nipah virus and Ebola virus (Calisher et al. 2006). If the life history

and ecological traits of bats that are hypothesized to influence the maintenance and emergence of pathogens affect either virus replication within hosts or the rate of transmission between hosts, they might also have consequences for viral evolution. For example, overwintering of temperate- zone bats through hibernation or extended bouts of torpor is thought to cause a seasonal pause in transmission and/or decelerated disease progression within hosts, perhaps due to metabolic down regulation of cellular processes or reduced contact rates while bats are inactive (Sadler & Enright

1959). These climate-mediated mechanisms might slow evolution in viruses associated with temperate-zone bats compared to tropical species, where year-round food availability and milder temperatures extend bat and virus activity through all seasons. Next, high contact rates in colonial bats may promote infections with greater virulence and reduced incubation periods, increasing the number of viral generations per unit time and speeding viral evolution (Brown et al. 2001; George et al. 2011). Finally, long distance migration, a relatively common strategy in bats, may slow viral evolution by homogenizing viral populations or by reducing transmission if the physiological stress from migration removes infected hosts from the population, i.e.,

‘migratory culling’ (Altizer, Bartel & Han 2011).

Here, we used a large dataset of bat rabies virus sequences collected from throughout the

Americas to quantify variation in the evolutionary rate of rabies virus when associated with different host species. We ask whether the tempo of evolution undergoes episodic shifts when rabies virus colonizes new species, implicating host biology as a key driver of viral evolution, or evolves gradually along its ancestral history, reflecting the greater importance of conserved viral features in controlling rates. Next, we used newly developed Bayesian hierarchical phylogenetic models and traditional comparative analysis to identify the life history traits of bats that shape the tempo of viral evolution.

RESULTS

Variation in substitution rates across bat rabies lineages. We employed maximum likelihood

(ML) and Bayesian phylogenetic analysis to define 21 subspecies, species or genus specific lineages of rabies virus for comparative analysis of evolutionary rates (see Materials and

Methods for analytical details and operational definitions of lineages). A relaxed molecular clock analysis indicated that viral lineages were relatively young, ranging in age from 83 – 305 years, with a most recent common ancestor of all bat lineages occurring in 1585 (95% Highest

Posterior Density, HPD: 1493 - 1663; Appendix AI Table 1). To describe the evolutionary rate variation among viral lineages, we focused on substitution rates in the third codon position (CP3), as these predominately synonymous substitutions can indicate more clearly how viruses respond to processes affecting their generation time (Nei 1987). Average substitution rates estimated for each lineage independently (Independent Lineage Models, ILM) and by the hierarchical phylogenetic model (HPM) each spanned approximately one order of magnitude among viral lineages compartmentalized to different host species (ILM range: 8.31x10-5 – 2.08x10-3; HPM range: 2.16x10-4 – 1.07x10-3 substitutions/site/year). This indicates that select lineages exhibit approximately 5 - 22 fold acceleration of evolution relative to the slowest evolving viruses. The

HPM substantially improved the precision of parameter estimates relative to the ILMs, with only negligible difference in point estimates for most lineages, as described by others (Suchard et al.

2003; Edo-Matas et al. 2010) (Figure 3.1). Notably, the more extreme values estimated by the

ILMs, typically from viral lineages with less informative datasets, were drawn closer to the population mean, suggesting less susceptibility of the HPM to stochastic noise introduced by sampling error and potentially more accurate estimates.

Virus evolutionary rates are labile following host shifts. If evolutionary rate is a relatively static trait of viruses, it should be conserved in novel environments and would be expected to reflect viral evolutionary history, with closely related viral lineages having similar rates, regardless of their contemporary host environment. We tested whether the evolutionary rates of bat rabies virus lineages are conserved or labile following host shifts by viruses between bat species by quantifying values of Blomberg’s K (a common statistic for diagnosing phylogenetic non-independence in comparative analysis) for viral substitution rates in the bat rabies phylogeny

(Blomberg, Garland & Ives 2003; Freckleton 2009) (Figure 3.2A). We detected very low and non-significant values of K for both rates estimated under the ILMs (K = 0.36, P = 0.58) and the

HPM (K = 0.39, P = 0.40) and these estimates of K did not differ significantly from expected values given our phylogenetic tree and the observed rates under a null model of randomly distributed rates (Figure 3.2B). This result was corroborated by our Bayesian phylogenetic analysis which, accounting for uncertainty in the evolutionary history of bat rabies lineages, found no correlation in the evolutionary rates along consecutive branches in the bat rabies virus phylogeny (covariance = 0.005, 95% HPD: -0.051 – 0.058). These analyses indicate that rates of viral evolution are capable of dramatic shifts following colonization of new host species. Using a phylogeny of bat hosts from mitochondrial sequence data, we found that viral evolutionary rates were similarly unconstrained by host evolutionary relatedness (ILMs: K = 0.07, P = 0.18; HPM:

K = 0.21, P = 0.08), such that viruses associated with closely related bat species or sub-species often had dissimilar evolutionary rates (Figure 3.1, Appendix B Figure 1).

Local host environment determines virus evolutionary rates. Because the evolutionary rates of rabies virus lineages varied among host species independently of reservoir host or virus

phylogenetic constraints, we tested whether the physiological, ecological or environmental traits of hosts (Appendix B Table 2) could instead determine viral evolutionary rates using a generalized linear model (GLM) comparison approach. This analysis identified the climatic region of bat taxa as the single strongly supported predictor of viral evolution in the confidence set of GLMs (Akaike importance weight = 1.0), alone explaining 66% of the variance in viral

2 evolutionary rates (F1,19 = 37.2, R = 0.66, P < 0.0001; Table 3.1). However, this traditional statistical approach could not account for the often-substantial uncertainty in point estimates of evolutionary rate from the ILMs (Figure 3.1). We therefore incorporated the same categorical and continuous effects directly into a Bayesian HPM that allowed us to simultaneously quantify the distribution of the rate of evolution for each viral lineage from the molecular sequence data, estimate model parameters and conduct model comparison using Bayes factors (BF). The

Bayesian model echoed the strong support found in the GLM analysis for accelerated viral evolution in the tropics and subtropics relative to viruses restricted to the temperate zone (log effect size, β = 1.24 [95% highest posterior density = 0.72 – 1.76]; BF = 466.54) with negligible support for all other predictors (BF ≤ 1; Figure 3.3A). On average, rabies viruses found in tropical or subtropical bat species accumulated 9.44 x 10-4 (ILM: 1.15 x 10-3) substitutions per site per year (subs/site/year), while viruses in temperate bats accumulated only 2.53 x 10-4 (ILM:

2.92 x 10-4) subs/site/year (Figure 3.3A,B) – a nearly fourfold deceleration of viral evolution in temperate bats.

DISCUSSION

By applying a comparative phylogenetic approach to a unique dataset spanning hundreds of viruses isolated from many host species, our study demonstrates strong effects of host biology on

the tempo of evolution in a zoonotic RNA virus of bats. First, we observed approximately one order of magnitude of variation in rates of viral evolution among rabies viruses isolated from different bat species, indicating that certain viruses evolve up to 22 times faster than others

(Figure 3.1). Such rate heterogeneity within a single species is exceptional given that similar variation is more commonly observed among different viral families and comparable to the variation in mitochondrial DNA divergence rates between vertebrate taxa genetically isolated for millions of years (e.g., whales versus rodents) (Martin & Palumbi 1993). Moreover, our phylogenetic analysis indicates that these shifts in evolutionary rate can occur within decades, consistent with rapid rate adjustment after colonization of new host species (Appendix B Table

1).

Because the genomic structure, transmission route and replication mechanisms are not known to vary among rabies virus lineages, the differences that we observed in evolutionary rate can only arise from epizootiological differences among viral lineages found in different reservoir species. The extraordinary plasticity of rabies virus to successfully maintain itself in numerous bat species therefore represents a striking example of viral robustness in vastly different host environments. In previous work, the existence of species-specific maintenance strategies was implied by mathematical models of rabies virus dynamics parameterized by detailed field studies in a temperate bat, Eptesicus fuscus. In that system, host-virus co-existence was impossible without a transmission-free hibernation period, yet rabies virus is clearly maintained in many bat species that do not hibernate (George et al. 2011). The strong host barriers to the establishment of sustained chains of transmission in new host species despite frequent cross-species transmission among bats similarly argue against a universal strategy for maintenance (Streicker et al. 2010). Our finding that evolutionary rates undergo dramatic shifts upon colonization of

new species (Figure 3.2) is consistent with the idea that adaptation to each species might be enabled by adjusting features of the host-virus interaction that influence the generation time between infections, such as incubation period or replication rate.

Because the tempo of viral evolution was unconstrained by viral evolutionary history, we sought to identify the traits of hosts that accelerate or decelerate viral evolution. Strikingly, we found that the viral molecular clock ticked nearly four times slower in rabies viruses in temperate bat species compared to tropical and subtropical species (Figure 3.3B). This pattern could not be explained by geographic structuring of bat diversity because several widely distributed bat species and genera supported disparate rates of viral evolution in lineages circulating in different climatic zones (Figure 3.1) resulting in a lack of phylogenetic signal of viral evolution in the bat phylogeny (Appendix B Figure 1). Similarly, although we faced limited resolution in the assignment of other host species traits, we found no evidence for evolutionary conserved traits of bats such as metabolic rates, migratory behavior or coloniality on setting rates of viral evolution

(Table 3.1, Appendix B Table 3).

Consequently, we suggest that geographical variation in bat life history and seasonality that is largely independent of the bat phylogeny underlie the observed effect of climatic region on viral evolution. Specifically, reduced seasonality in the activity patterns of tropical and subtropical bats likely allows year-round replication within and transmission between bats, resulting in a net increase in the number of viral generations per year relative to seasonal pulses of transmission occurring during periods of high bat activity in the temperate zone. Indeed, when we conditioned our Bayesian analysis on 231 generations of the HPM that lacked the climatic region term, we found that seasonal inactivity was the only predictor that gained strong statistical support (BF = 36) with significantly faster viral evolution for bats that remain active year-round

relative to species that hibernate or use prolonged torpor during winter (β = 1.01 [95% highest posterior density = 0.39 - 1.52]). The selection of climatic region as a surrogate for seasonal activity rather than the records of overwintering behavior themselves appears somewhat counterintuitive. One explanation may be that the occurrence and duration of seasonal inactivity and torpor is poorly understood for many bat species, particularly those that span large geographic areas where overwintering activity can differ among climatic zones (McNab 1974;

Dunbar & Brigham 2010). Because assignment of overwintering records often required generalization of a few observations to an entire species range, we suspect that climatic zone may be a more accurate descriptor of seasonal activity, especially for species with geographically variable life history strategies.

Regardless of the exact mechanisms involved, the possibility of accelerated viral evolution in tropical environments is a topic of general interest for understanding the maintenance and emergence of many viral infections that occur across climatic zones or experience altered transmission dynamics as a result of anthropogenic environmental change. For example, lineages of Chikungunya virus evolve more slowly in seasonal African environments where mosquito populations and transmission dynamics are more variable relative to urban transmission cycles in Asia, where consistently large human and mosquito populations may shorten times between infections and support epidemic maintenance over multiple years (Volk et al. 2010). Similarly, influenza shows reduced seasonality in the tropics relative to temperate zones of the Americas (Viboud, Alonso & Simonsen 2006). Our results would predict that this sustained transmission might accelerate influenza evolution in the tropics, and more broadly, motivate the need to consider not only functional traits of viruses, but also the seasonality and epidemiological dynamics of the host-virus interaction for a more complete understanding of the

tempo of viral evolution.

Although we focused on a proxy for neutral evolution as a marker of viral replication and transmission, accelerated substitution rates also increase the genetic variance available to natural selection, thereby potentially increasing the rate of adaptive evolution (Fisher 1930). Ultimately, this genetic variability is critical for pathogen adaptation to new host species, evasion of host defenses and resistance to medical prophylaxis and treatment (Holmes 2009). In this context, our finding that host biology explains nearly 70% of variation in viral evolution indicates a role for the host in influencing the genetic diversity of viral populations. Through generating a broader spectrum of fitness genotypes, more rapidly evolving viruses might both be more difficult to manage in their reservoirs and predisposed to overcome species barriers to emergence.

Therefore, hosts may be variable not only in the ecological patterns of inter-specific contacts that provide opportunities for cross-species transmission and in the infectiousness of their viruses to host species of variable phylogenetic distance, but also in their likelihood of producing ‘pre- adapted’ viruses for emergence - based solely on the evolutionary dynamics that occur within the reservoir species (Parrish et al. 2008).

The role of life history in explaining variation in rates of molecular evolution has been previously demonstrated for variety of free-living plants and animals (Martin & Palumbi 1993;

Bromham 2002; Smith & Donoghue 2008). To date, such an analysis has not been extended to host-pathogen interactions. Our results demonstrate the existence of a previously unexplored cross-kingdom effect of host biology that can accelerate the evolution of RNA viruses in ways that may have important implications for understanding the species origins and control of emerging diseases. The broad host range and increasing availability of sequence data from many rapidly evolving viruses makes them an ideal, real-time system to reveal how host biology and

behavior shape the epizootiology and evolution of pathogen ensembles, while providing general insights into the factors that influence the tempo of molecular evolution.

MATERIALS AND METHODS

Rabies virus sampling and selection of lineages for hypothesis testing. Complete and partial rabies virus Nucleoprotein gene sequences that were associated with bats from North and South

America and contained information on sampling date to year were downloaded from GenBank (n

= 650). New sequences (n = 18) were generated for relatively poorly sampled viral lineages for which additional virus samples were available within the collection of the CDC Rabies Program

(see Appendix B). To define phylogenetic lineages to be included in subsequent analyses, ML and Bayesian phylogenetic analyses were performed using Garli v.0.96b and BEAST v.1.6.1, respectively (Zwickl 2006; Drummond & Rambaut 2007). The ML analysis used the General

Time Reversible (GTR) model of nucleotide substitution with invariant sites (I) and Γ distributed rate variation among sites as selected by Akaike’s information criterion corrected for small samples size (AICc) in JModeltest (Posada 2008). The ML tree was estimated by 5 independent searches with random starting trees, followed by 5 additional searches using the best tree from the previous set of searches as the starting tree. For the Bayesian analysis, we linked substitution rates for the first and second codon positions (CP12) and allowed independent rates in CP3.

Separate substitution models were selected for CP12 and CP3 in jModeltest using AICc after partitioning aligned sequences by codon position. The BEAST analysis therefore applied the

TIM1ef+I+Γ substitution model to CP12 and the TVM+Γ substitution model to CP3. We used the

Bayesian skyride model as a flexible demographic prior for viral effective population size and an uncorrelated lognormal relaxed molecular clock to accommodate rate variation among lineages.

Five independent Markov Chain Monte Carlo (MCMC) analyses were run for 50 million generations each, with samples from the posterior drawn every 50,000 generations following variable burn-in periods based on convergence of likelihood values and model parameters as indicated in Tracer (Rambaut & Drummond 2007). The results from the five runs were combined to generate a maximum clade credibility tree and divergence time summaries. Lineages for subsequent analyses of substitution rates included those that (i) contained at least 8 sequences

(mean = 30.9 sequences), (ii) were supported by Bayesian posterior probabilities of > 0.9 and

(iii) were sampled over a minimum time span of 4 years (mean = 19.4 years). Of the 28 viral lineages identified in the initial phylogenetic analyses, 21 fit our criteria for inclusion in subsequent analyses, amounting to a final dataset of 648 sequences collected between 1972 and

2009 from 21 bat species or sub-species.

Independent estimation of nucleotide substitution rates in bat rabies lineages. Sequence alignments were constructed for each viral lineage and nucleotide substitution models were selected for CP12 and CP3 using AICc, as described above. For each lineage, the substitution rate was estimated in BEAST assuming an uncorrelated lognormal relaxed molecular clock and the

Bayesian skyline model of demographic growth. The evolutionary rate in CP3 was calculated by multiplying the mean substitution rate by the relative rate parameter for that partition. Each simulation was run for at least 100 million generations, with parameters sampled every 5,000-

10,000 generations. The first 10% of each run was discarded prior to the construction of the posterior probability distributions of parameters. Each analysis was run sufficiently long that effective sample sizes for parameters were > 200 and results of several independent runs were combined when necessary. Analyses of evolutionary rates focused on substitution rates in CP3

since these largely synonymous substitutions reflect differences in evolution associated with generation time (Nei 1987). However, rates in CP12 were closely correlated with CP3 rates (r =

0.87, P < 0.0001).

Bayesian hierarchical inference of nucleotide substitution rates and hypothesis testing. The separate estimation of molecular substitution rates for each rabies virus lineage assumes complete independence of parameters across viral lineages, but this is unlikely the case given the close evolutionary relationships among lineages and biological similarities of the processes of infection and replication among lineages. Because the quantity of data varied among lineages

(the number and temporal range of sequences), estimation in sparsely sampled lineages may lack power, resulting in imprecise estimates in the ILMs. HPMs have been proposed as a method to improve the precision of parameter estimates for partially independent datasets such as these

(e.g., populations of HIV within different patients) by assuming that individual lineage parameters vary around a shared unknown, but estimable population mean (Suchard et al. 2003).

More recently, tools have been developed within BEAST to incorporate fixed effects into HPMs and to select among candidate models via Bayes factors (Edo-Matas et al. 2010). These models take the general form of:

logθ = β0 + δP1 βP1 P1 + … + δPn βn Pn , (1) where θ is the evolutionary response variable of interest (here, the rate of molecular evolution in

CP3), β0 is an unknown grand mean, δ is a binary indicator that tracks the posterior probability of the inclusion of predictor, P, in the model and β is the estimated effect size of predictor P. The use of binary indicator variables within the MCMC search allows for a Bayesian stochastic search variable selection approach that simultaneously estimates the posterior probabilities of

parameters for all possible combinations of predictors and allows for calculation of the Bayes factor support for individual predictors as the ratio of the posterior odds to the prior odds of each predictor in the model.

We constructed a HPM for the 21 bat rabies virus lineages that assigned separate strict molecular clocks to CP12 and CP3 of each viral lineage and included fixed effect predictors of the evolutionary rate in CP3. For each CP, the evolutionary rate parameter of the molecular clock, the parameters of the GTR substitution model and the shape parameter of the discrete Γ distribution were modeled hierarchically across lineages, with all other parameters varying independently across data partitions. The fixed effects in the full model included climatic region

(temperate vs. tropics/subtropics), mass-independent basal metabolic rate (BMR), mass- independent torpid metabolic rate (TMR), coloniality (solitary vs. colonial), seasonal activity

(non seasonal vs. hibernation/periodic seasonal torpor) and long-distance migration (migrants vs. non-migrants) (Appendix B, Appendix B Table 2). Climatic region was condensed to two categories in the final HPM based on exploratory analyses that demonstrated no difference in evolutionary rate between tropical and subtropical viral lineages. All continuous variables were log transformed. The HPM was implemented in BEAST using four independent MCMC searches of 150 million generations each, with the posterior sampled every 5,000 generations.

Results from the four runs were combined after discarding the first 10% of each. Effect sizes of predictors reported from the HPM were calculated conditionally on the portion of the posterior distribution for which the respective effects were included in the model (i.e., βi | δEffect i = 1).

Similarly, evolutionary rate estimates from the HPM were calculated conditionally on samples of the posterior in which the statistically supported predictor was included in the model

(107,773/108,004 samples). Source files for the BEAST analysis and R scripts for conditional

effect size and parameter estimation are available from the corresponding author upon request.

In addition to the Bayesian hierarchical hypothesis testing framework described above, we conducted a more traditional GLM analysis of host predictors of mean viral evolutionary rates and a phylogenetic least squares (PGLS) regression analysis, the latter designed to test effects while controlling for potential effects of virus phylogeny on viral evolutionary rates.

These analyses contained the factors included in the HPM above, but because these methods do not account for uncertainty in the estimation of evolutionary rates (Figure 3.1), we also included several factors to identify the influence of estimation error: the number of years spanned and the number of sequences that comprised each dataset. For the GLM analysis, an initial model containing all terms was simplified using an exhaustive search of possible models using AICc in the glmulti package of R (Calcagno 2011; R Development Core Team 2011). Models with

Akaike weights within 10% of the highest were retained in the confidence set shown in Table

3.1. The PGLS regression was conducted in the caper package of R, using the Pagel’s λ statistic to account for phylogenetic non-independence of viral evolutionary rates (Orme et al. 2011). In our analysis, the ML value of λ was estimated as 0 in all models, indicating independence of evolutionary rates from the virus phylogeny, hence the results mirrored those of the GLM analysis (Appendix B, Table 4).

Phylogenetic signal in viral substitution rates and host traits. Blomberg’s K measures the degree of phylogenetic non-independence of species traits, with values ranging from 0 to infinity

(Blomberg, Garland & Ives 2003; Freckleton 2009). Values of K < 1 indicate less phylogenetic signal (more trait lability) than expected under a Brownian motion model of evolution and K > 1 indicating more correlation with phylogeny than expected. The K statistic was calculated for the

ILM and HPM sets of rate estimates using the topology of the rabies virus phylogeny in Figure

2A in the picante package of R (Kembel et al. 2010; R Development Core Team 2011). The statistical significance of estimates of K was tested by comparing the expected distribution of values on our phylogenetic tree to 5,000 randomizations of observed rates along the tips of the tree. A similar analysis using the ML phylogeny of bats estimated from published mitochondrial cytochrome oxidase I sequences was undertaken to assess whether the ecological and physiological similarity of closely related host species promotes evolution towards similar rates of viral evolution (Appendix B Figure 1).

FIGURES AND TABLE

Hierarchical phylogenetic model Independent lineage models 3.5e-3

3.0e-3

2.5e-3

2.0e-3

1.5e-3

1.0e-3 iral evolutionary rate evolutionary iral substitutions/site/year V 5.0e-4

0.0 V V V LiV NlV DrV LcV LsV LxV TbV LnV PsV PhV EfV2 EfV3 LbV1 LbV2 MyV1 MyV2 EfV1a EfV1b TbSA MySA EfSaA Tadarida Eptesicus Lasiurus Myotis Perimyotis Desmodus Nyctinomops Lasionycteris Parastrellus

Rabies virus lineage

Figure 3.1. Heterogeneity in evolutionary rates of host-associated bat rabies viruses. Median substitutions per site per year in the 3rd codon position of the nucleoprotein gene for estimates generated by the HPM (filled circles) and ILMs (open circles). Colors and dashed grey lines distinguish bat genera as indicated below the x-axis. Credible intervals represent the 95% highest posterior density on evolutionary rate.

A LcV LxV 0.71 LbV1 1.00 LsV 0.90 LbV2 1.00 1.00 LnV PsV LiV 1.00 EfV1b 0.90 EfV1a MV1 0.70 NlV 1.00 TbV 1.00 DrV TbSAV 1.00 1.00 EfV3 EfV2 0.86 PhV 0.91 MV2 0.89 EfSAV MSAV 50 years

9.31e-5 - 4.24e-4 4.25e-4 - 7.55e-4 7.56e-4 - 1.09e-3 1.10e-3 - 1.42e-3 1.43e-3 - 1.75e-3 1.76e-3 - 2.08e-3 B 1200

1000

800 RRM: K = 0.37 HPM: K = 0.39 ILMs: K = 0.36 600

Frequency 400

200

0.2 0.4 0.6 0.8 1.0 Blomberg's K

Figure 3.2. Evolutionary lability of viral substitution rates across host shifts. (A) Bayesian phylogenetic tree of bat rabies viruses with viral lineage names denoted in black. Host-associated lineages are condensed to triangles connecting the most recent common ancestor to the sampled branches. Lineages are colored along a blue (slowest) to red (fastest) continuum according to evolutionary rate in CP3 using estimates from the Independent Lineage Models (ILMs). Bayesian posterior support values ( > 0.70) are given above branches to the lineage level only. All colored lineages received Bayesian posterior support values of ≥ 0.91. (B) Frequency histogram of expected values of Blomberg’s K from 5000 random assignments of substitution rates estimated from the ILM to lineages. Dashed lines indicate 95% bounds of the null distribution and diamonds denote the median values of K for the randomized rate model, RRM (grey), the ILMs (blue), and the HPM (red).

A B Predictor Bayes factor ! (95% HPD) | " = 1 1.0e-3 Climatic region 466.54 Basal metabolic rate 0.82 8.0e-4 Torpid metabolic rate 1.00 6.0e-4 Coloniality 0.46 Seasonal inactivity 0.46 4.0e-4 Long-distance migration 0.69 rate Evolutionary Substitutions/site/year 2.0e-4

-1 -0.5 0 0.5 1 1.5 Temperate Subtropics/tropics Climatic region

Figure 3.3. Predictors of viral evolutionary rate from the Bayesian hierarchical phylogenetic model. (A) Effect sizes (β) on a log scale for each predictor were conditioned on the inclusion of that term in the model (i.e., β | δ = 1). Climatic region, coloniality, seasonal inactivity and long distance migration were treated as categorical variables. Horizontal lines are the 95% highest posterior density intervals on conditional effect sizes and squares (median effect sizes) are proportional to effect sizes. (B) Violin plot showing the effect of climatic region on viral evolutionary rate. White points, black boxes and whiskers indicate the median, inter- quartile range and the total range of values for that group, respectively. The grey shading shows the probability density of evolutionary rate at different values.

Table 3.1. The confidence set of generalized linear models examined to explain viral evolutionary rate.

Predictors* AICc Δ AICc w† r2 Climate 40.297 0.000 0.370 0.66 Climate + migration 42.710 2.413 0.111 0.67 Climate + nyrs 42.823 2.526 0.105 0.67 Climate + log(BMR) 43.173 2.876 0.088 0.67 Climate + seasonal 43.185 2.888 0.087 0.67 Climate + log(TMR) 43.311 3.014 0.082 0.66 Climate + log(n) 43.379 3.082 0.079 0.66 Climate + colony 43.384 3.087 0.079 0.66

* Shown is the confidence set of models from the GLM analysis (see Materials and Methods for details of model selection). All GLMs included an intercept term and had P values < 0.0001. Abbreviated terms are defined as follows: BMR = mass-independent basal metabolic rate; TMR = mass-independent torpid metabolic rate; n = number of sequences per lineage; nyrs = range of years spanned per lineage. † AIC weights (w) describe the relative likelihood for each model given the set of models considered.

CHAPTER 4

SELECTION LINKED TO WITHIN HOST SPREAD DRIVES RABIES VIRUS

EMERGENCE IN NEW HOST SPECIES 3

______

3 Streicker, D.G., Altizer, S., Rohani, P., Velasco-Villa, A. and Rupprecht, C.E. To be submitted to PLoS Pathogens.

ABSTRACT

Understanding the genetic basis for viral adaptation to new host species is crucial to predict the extent of disease emergence resulting from cross-species transmission. We examined a unique data set on rabies virus host shifts among multiple bat species to explore their dynamical outcomes and quantify the extent of adaptive evolution. Using sequence data for 78% of the rabies virus genome, we identified repeated bursts of positive selection on several viral proteins that coincided with extended periods of low viral fitness before endemic establishment in a new host. By combining these results with coalescent reconstructions of epizootics in individual viral lineages, we show that the number of positively selected amino acid changes predicts demographic patterns of emergence in recipient host species. Finally, we demonstrate a positive relationship between the genetic distance between donor and recipient hosts and the amino acid distance between their associated viruses at positively selected sites, hinting that a flattening of the viral “fitness valley” between closely related hosts might facilitate emergence. Our results provide a set of candidate amino acid sites for dissecting the molecular basis for host adaptation in rabies virus and demonstrate the critical roles for both chance transmission of pre-adapted variants and adaptive evolution within recipient species for viral emergence.

INTRODUCTION

Many newly emerging infectious diseases affecting humans, domesticated animals and wildlife are viruses that originate from other host species (Daszak, Cunningham & Hyatt 2000;

Woolhouse, Haydon & Antia 2005; Wolfe, Dunavan & Diamond 2007). The impacts of viral emergence on hosts depend in large part on the extent of sustained transmission in recipient species after cross-species transmission (Lloyd-Smith et al. 2009). Therefore, understanding how

viruses increase their fitness in new hosts is paramount for predicting and mitigating disease emergence. For RNA viruses, the large genetic diversity generated by low fidelity of error-prone viral RNA polymerases provides genetic and phenotypic plasticity for establishment in new host environments. Indeed, rapid adaptive evolution is a key mechanism characterizing several well- studied viral emergence events such as parvovirus in carnivores, SARS in humans and

Venezuelan equine encephalitis in horses (Shackelton et al. 2005; Song et al. 2005; Anishchenko et al. 2006). Because host shifts are rare events, this limits insights into the generality of the evolutionary processes associated with cross-species emergence. Key questions that could be addressed using data from multiple host shifts include: Is adaptive evolution a universal feature of viral emergence? Do host shifts tend to utilize the same genomic pathways or does adaptation to different host species require changes in different genomic regions? How does adaptive evolution in viruses affect the speed of viral emergence?

The evolutionary dynamics and epidemiological outcome of viral emergence likely depend on how the donor-recipient species context interacts with viral evolutionary history. This concept has been described previously in terms of a viral fitness landscape, wherein each potential host species can be viewed as a distinct fitness peak, separated by greater or fewer adaptive mutations depending on the strength of host barriers to emergence (Kuiken et al. 2006).

With a shallower fitness valley between peaks, viral establishment should occur more easily and at an accelerated pace, so that only evolutionary fine-tuning is required in the new host species.

Implications for lowering the risk of disease emergence in this case would focus more on limiting initial cross-species transmission than on post-emergence control. On the other hand, extended periods of low viral fitness prior to epidemic emergence would provide a greater window for intervention within the novel host species. Empirical support exists for both of these

pathways to viral establishment in new hosts. Pathogens such as dengue virus in humans and

Venezuelan equine encephalitis in horses required very few adaptive changes in the recipient species, yet other host shifts (e.g., avian influenza and SARS in humans) seem to have required more extensive genetic change (Song et al. 2005; Anishchenko et al. 2006; Smith et al. 2006;

Vasilakis et al. 2007). Because it has rarely been possible to compare the emergence outcomes of closely related viruses in multiple host species, relating the extent of genomic change to the temporal dynamics of viral emergence remains limited (Haagmans, Andeweg & Osterhaus

2009).

Rabies virus (RV, Lyssavirus, Rhabodoviridae) is a globally distributed zoonotic pathogen responsible for over 50,000 human deaths annually (WHO 2005). Despite the fact that rabies can infect virtually all mammal species, the virus typically persists via independent, species-specific transmission cycles, principally in carnivores and bats (Rupprecht, Hanlon &

Hemachudha 2002). This present day distribution of rabies reflects an evolutionary history of host shifts between species, including several recent introductions such as from dogs to foxes in

Europe in the 1930s, from bats to skunks in the United States in 2001 and from bats to coatis and/or foxes in Mexico in 2008 (Bourhy et al. 1999; Badrane & Tordo 2001; Leslie et al. 2006;

Arechiga-Ceballos et al. 2010). Importantly, the evolutionary mechanisms that underlie rabies virus adaptation to new host species remain poorly understood. Broad, comparative studies across the genus Lyssavirus suggest that changes in the amino acid sequence of the glycoprotein

(G) gene, a major antigenic region responsible for receptor recognition, membrane fusion and pathogenicity, may drive adaptation to new host species (Badrane & Tordo 2001). Changes in G also have been observed in passaging studies of RV in novel environments; however, because passaging inherently relaxes the selective constraints imposed by natural transmission, the

importance of the G relative to other genes is uncertain (Kissi et al. 1999; Holmes et al. 2002).

Other studies analyzed the nucleoprotein (N), the major capsid-forming gene, which also holds antigenic sites and plays a role in transcription in replication, but found no strong signals of host adaptation (Holmes et al. 2002; Wunner 2007).

Another potentially important genomic region for rabies virus adaptation to new hosts is the virus dependent RNA polymerase gene (or large gene, L). As the catalytic component of the

RNA polymerase, this gene regulates viral transcription and replication (Poch et al. 1990).

Consistent with the idea that the L gene has major effects on within-host dynamics and virulence, mutations in the L genes of structurally similar avian metapneumoviruses and paramyxoviruses have been linked to changes in virulence and replication in new host species. As of yet, however, no studies have investigated the role of L in the adaptation of RV to new hosts (Poch et al. 1990;

Brown et al. 2011; Dortmans et al. 2011).

Bats are an excellent host group to examine the evolutionary processes associated with the cross-species emergence of rabies. Over 80 bat species are known to be infected with rabies and many of these maintain species-associated viral lineages that derive from a relatively recent common ancestor (Constantine 2009; Streicker et al. submitted). The occurrence of repeated host shifts by genetically similar viruses among ecologically and evolutionarily distinct host species makes bat rabies a unique system for understanding the processes underlying viral emergence.

Here, we estimate the evolutionary history of bat RV to identify plausible host shifts between bat species using a dataset comprised of 78% of the viral genome from most known bat RV lineages.

By combining estimates of donor-recipient relationships with an analysis of selective pressures on viral genes and reconstructions of past viral demography, we identify genomic signatures of

adaptation and ask how these evolutionary changes relate to dynamical patterns of viral emergence between host species.

RESULTS

Evidence for selection along the rabies virus genome. One measure of the adaptive evolutionary pressure exerted on genes is the ratio of non-synonymous (dN) to synonymous (dS) substitutions, with dN/dS > 1 or dN - dS > 0 indicative of positive or diversifying selection. We measured the dN/dS ratio across along the branches of maximum likelihood (ML) phylogenetic trees inferred for the N, G and L genes of bat RV lineages. Datasets comprised 30 viral lineages

(184 unique sequences) for N, 26 lineages (68 unique sequences) for G and 23 lineages (48 unique sequences) for L. The overall dN/dS ratio averaged across all sites in the N, G and L genes was 0.05, 0.15 and 0.05, respectively, indicating purifying selection on most sites. Selection on specific amino acid positions was investigated using a fixed effects likelihood (FEL) analysis, which averages rates of dN and dS among branches, and by mixed effects models of episodic selection (MEME) which aim to identify sites undergoing temporally varying positive selection

(Kosakovsky Pond & Frost 2005; Kosakovsky Pond et al. 2011). For all three genes, both methods identified amino acid sites putatively evolving under positive selection; all sites detected by FEL were also supported by MEME (Figure 4.1). The overall proportion of sites putatively under positive selection was slightly higher in G (0.02 of 524 total sites) than N and L

(0.01 of 450 total sites and 0.009 of 2128 total sites, respectively), but this difference was only marginally significant (χ2 = 5.47, P = 0.07).

In N, sites evolving under positive selection were located predominately within the first

160 codons of the coding region (5/6 sites). In G, 9 sites suggested to be evolving under positive

selection were found in the ectodomain (sites G34, G68, G139, G264, G331, G357, G373, G389 and G448; site numbers indicate position on G including the signal peptide unless otherwise noted), which forms the spikes in the virus envelope responsible for host cell fusion and entry.

Only one of these sites, G357, was located in the main antigenic sites or on the described epitopes of G (Wunner & Dietzschold 1987; Prehaud et al. 1988; Dietzschold et al. 1990;

Benmansour et al. 1991). Two other sites under positive selection were found in the endodomain, which is responsible for the interaction of G with internal viral proteins (sites G493 and G515) (Figure 4.1 B). Although not implicated as evolving under positive selection by our criteria of statistical significance, we did find a greater rate of dN than dS in site 333 of the ectodomain, a key virulence site (dN/dS = 1.189, PFEL = 0.81, PMEME = 0.07) (Dietzschold et al.

1983; Tuffereau et al. 1989). Specifically, we found substitutions to glutamine in several lineages perpetuated by bats of the genus Myotis in the western United States, while other bat

RVs had the typical argenine or lysine residues associated with virulence. Although we expected to find evidence for positive selection at the putative toxic loop site in G (266) owing to past work implicating this site in differences between RVs of silver-haired bats (Lasionycteris noctivagans) and terrestrial carnivores in North America (Morimoto et al. 1996), this site was, in fact, totally invariant among the bat RVs surveyed here (with all viruses containing cysteine at this site, the same residue found in terrestrial carnivore RVs). No other sites within the toxic loop of G were found to be under positive selection. In L, sites under positive selection were dispersed along the gene and occurred in all of the putative functional regions described by Poch et al.

(1990) except region II (Figure 4.1 C). Positive selection appeared to be particularly frequent in regions III (n = 3; 27.3% of selected sites) and IV (n = 4; 36.4%), and to a lesser degree VI (n =

2; 18.2%). Three sites were located in the catalytic domain (sites L637, L638 and L643) and 6

were found in the region that interacts with phosphoprotein to form the RNA polymerase complex (sites L1620, L1742, L1743, L1840, L2090 and L2093).

Timing of positive selection on different genes. We mapped the non-synonymous substitutions in positively selected sites to the maximum likelihood phylogenies of each gene to describe the intensity and timing of selective pressures associated with host shifts. Changes at positively selected sites in N were rarer than in the other two genes, and occurred almost exclusively on the tips of the tree, suggesting a lesser involvement in host-species adaptation (Figure 4.2 A,D). In contrast, changes in G occurred predominately on deeper nodes associated with viral establishment in new host groups (Figure 4.2B,E) and changes in L accumulated at a fairly constant frequency throughout the viral phylogeny (Figure 4.2 C,F). Several sites, most notably in the G and L genes, showed high frequencies of non-synonymous substitutions along the viral genealogy with strikingly different patterns of amino acid use (Figure 4.2 B,C). For example site

G493 switched 9 times between 6 different amino acid residues. In contrast, site G357 was altered 7 times, but all substitutions occurred in a flip-flop manner between valine and isoleucine. A similar pattern was observed in L1620, which switched repeatedly between glycine and 4 other residues; suggesting positive selection on evolutionarily constrained sites (Appendix

C Figure 1).

Variable amounts of adaptive evolution accompany rabies virus emergence events. To better understand the amount of adaptive evolution associated with each shift into a new bat species, we needed to know in which bat species positively selected substitutions occurred. To do this, we used a phylogenetic ancestral state estimation model to assign host species to

branches while integrating information across all three genes and accounting for phylogenetic uncertainty through hierarchical Bayesian inference. Using a Bayes factor cutoff of 3 for statistical significance of transitions between host states, we identified 64 potential host shifts. In some cases, host shifts were inferred in both directions for a given pair of bat species, indicating a limitation in the data to discriminate the donor versus recipient species for some host shifts

(Appendix C Table 2). In some cases, supported host shifts from deeper nodes in the tree to only one of a pair of hosts aided in the assignment of host species to branches. Nevertheless, we treated these assignments with caution and subsequent analyses were performed with and without assuming knowledge of ancestral host states.

Mapping the amino acid changes onto the consensus phylogeny revealed that different numbers of positively selected sites were associated with different host shift events. At the extreme, some lineages sustaining up to 6 changes in putatively positively selected sites from the original virus associated with the host shift to the consensus sequence of present-day isolates

(Appendix C Figure 1). Substitutions associated with host shifts occurred predominately in the L and G genes. In most cases, substitutions occurred in both genes along the branches leading to new lineages, but several host shifts appeared to involve changes in only G or only L and some host shift events showed no evidence for changes in positively selected sites (Appendix C Figure

2 B,C).

Adaptive evolution and the speed of viral emergence. The variation in numbers of amino acid changes associated with the establishment of rabies virus in different bat species led us to ask whether this apparent heterogeneity in the strength of host barriers to emergence might cause variable patterns of emergence in recipient species. Using viral genetic data to characterize the

demographic history of RV in each new host species for which we had adequate data, we found that bat RVs shared amazingly similar three-phase demographic histories (Figure 4.3).

Specifically, most lineages showed evidence for a lag phase during which the effective number of cases was low, an expansion phase where the number of cases increased, and an endemic phase whereby the number of cases plateaued. Only RVs found in common vampire bats (DrV,

Desmodus rotundus) deviated from this pattern, with a precipitous decline in the third phase, perhaps associated with the extensive culling of that species in Latin America.

Comparison across viral trajectories revealed differences in the duration of the initial lag period and in the rate of increase during the second phase, allowing us to ask how the number of amino acid changes in positively selected sites covaried with the timing of emergence. In particular, we hypothesized that less adaptive evolutionary change might be associated with more rapid emergence and greater viral fitness on average, whereas emergence events that required greater evolutionary change would be accompanied by a longer lag phase (Kuiken et al.

2006). When we considered only amino acid changes on the branch of the phylogeny immediately leading to each terminal lineage (i.e., ignoring inferred ancestral host states), we found a marginally significant positive relationship between the number of changes in L and the

2 duration of the lag phase (Figure 4.4 D; F1,10 = 3.758, r = 0.27, P = 0.08); but no relationship between lag time and the total number of substitutions in G (Figure 4.4 C, P = 0.65). When we added the host species assignments from the ancestral state estimation to both the number of positively selected substitutions and the inferred dates of host shifts, a positive non-linear relationship between the total number of substitutions since each host shift and the duration of the lag phase became apparent (Figure 4.4 A). Closer inspection suggested that this relationship was mainly driven by substitution in L; however, this pattern was only marginally significant

2 (Figure 4.4 B, r = 0.28, F1,10 = 3.85, P = 0.078). In contrast, neither the rate of increase nor the net gain in the effective number of infections during the expansion phase were associated with the number of viral substitutions in any gene.

Effects of host relatedness on the depth of the viral fitness valley. We expected that viruses shifting between biologically similar host species would require fewer genetic changes to establish in a new host. Therefore, we tested the effects of two variables hypothesized to influence the depth of the fitness valley that RV must cross to establish transmission in new species: difference in social group size and the genetic distance between donor-recipient pairs

(Hughes, Orciari & Rupprecht 2005; Streicker et al. 2010). Neither factor explained the number of amino acid changes from ancestrally reconstructed viral sequences to the contemporary consensus sequences of each lineage. Notably, these analyses were restricted only to the small subset of possible host shifts that were inferred by ancestral state estimation. They therefore only reflected patterns among actualized host shifts and would fail to detect host differences that totally inhibit viral establishment. We therefore conducted a similar analysis including all pairs of bat species and found that increasing genetic distance between bat species was associated with a greater number of amino acid differences across positively selected sites in the G of their respective viruses (Figure 4.5; Mantel r = 0.15 [95% CI = 0.09 - 0.22], P = 0.045).

DISCUSSION

By examining changes in the RV genome across multiple successful host shifts, our study provides evidence for adaptive changes in RV to new host species and constitutes a first step towards understanding how the intensity of barriers to host shifts could affect viral emergence.

Importantly, our results support past episodes of positive selection on all three genes examined

(Figure 4.1). Although previous studies have found mixed evidence for positive selection at some regions on G, our study is the first to implicate the N and the L genes as possible factors in host adaptation (Badrane & Tordo 2001; Holmes et al. 2002; Hughes, Orciari & Rupprecht

2005). Because our data set was larger than those used in past studies, encompassing more host species, more viral lineages, and a greater number of genes, we likely had greater power in identifying amino acid changes associated with host adaptation relative to data sets used in previous studies. Moreover, we used biologically realistic models for host switching that allowed selection pressures to vary across genomic space and time (Kosakovsky Pond et al. 2011).

When we mapped AA substitutions along the branches of phylogenetic trees (Figure 4.2

A-C), we found that changes in N were mainly singleton substitutions that occurred along the tips of the tree (Figure 4.2 A,D). This suggests that while there are many lineage specific amino acids in N, these likely arose through a combination of neutral genetic drift and purifying selection rather than adaptive evolution. Patterns of positive selection in G and L showed a dramatically different pattern, with amino acid changes commonly occurring on branches preceding viral establishment in new host species (Figure 4.2 B,C). This pattern is consistent with the idea that episodic bursts of positive selection follow cross-species transmission to enable viral adaptation to new hosts, whereas subsequent evolutionary dynamics are dominated by purifying selection. Temporal shifts in selection pressure are further supported by the striking consistency of three-phase demographic histories in our Bayesian skyline plots, although as noted below, these patterns can also be influenced by host population characteristics (Figure 4.3).

Mapping the amino acid sites under selection to functional regions of genes provided some clues to the mechanisms of host-mediated selection. First, most sites evolving under

positive selection in G were found in the ectodomain, which holds the main antigenic sites and mediates interaction with host cells. The inefficacy of the immune response against rabies after entry into the central nervous system argues against selection by host mediated immunity, and the infectiousness of RVs across mammal species suggests a lack of cell receptor limitation.

However, several regions of the ectodomain, particularly antigenic sites II and III, hold critical residues for virulence (Dietzschold et al. 1983; Seif et al. 1985; Prehaud et al. 1988). Although we only found one site under selection within antigenic site III, a cluster of adjacent sites was also under positive selection (Figure 4.1 B). Moreover, we found an elevated rate of non- synonymous substitutions at the key virulence site 333 of the ectodomain, with changes including substitutions away from arginine and lysine (Figure 4.1 B). These substitutions, also described recently in several Brazilian bat variants, reduce or eliminate pathogenicity in laboratory RV strains and can influence cell-to-cell spread, the viral pathway to the central nervous system and the distribution of virus in the brain (Dietzschold et al. 1983; Dietzschold et al. 1985; Kucera et al. 1985; Yan et al. 2002; Sato et al. 2009). This implies that processes critical for within-host development of RV could be directly affected by these changes or indirectly by compensatory mutations that maintain virulence.

Another line of evidence for selection operating on within-host progression comes from changes in genomic regions that affect internal proteins and on portions of proteins that have minimal interaction with the host environment. In particular, positive selection was common in conserved functional regions IV and VI and in the catalytic domain of L (Figure 4.1 C).

Although the precise functional roles of these regions have yet to be delineated, substitutions in these sites might directly influence rates of transcription and replication, which would in turn, influence rates of within-host spread. Positively selected sites were also common in regions of

interaction between viral proteins. These included the region of L that interacts with the phosphoprotein to form the RNA polymerase and the endodomain of G, which is responsible for the interaction of G with other viral proteins inside the virion (Figure 4.1 B,C). One possible explanation for selection on these regions is that changes in protein interactions could alter transcription or replication to modulate viral replication within the host, analogous to the hypothesized role of defective interfering virions in regulating the progression of viral infections

(Holland 1987; Bangham & Kirkwood 1993). Collectively, changes in positively selected sites in

G and L that influence transcription, replication and pathogenicity, may regulate viral spread to and from the central nervous system and salivary glands in ways that enhance transmission before host death. More generally, the possibility that RVs adapted to different bat species utilize specialized within-host infection strategies could also affect the outcome of cross-species exposures to humans or other terrestrial mammals. This may help explain the disproportionate involvement of certain bat RV variants in human rabies cases, and our results provide a set of candidate genomic sites for exploring differences in pathogenesis and infectiousness among viral strains (Morimoto et al. 1996; Messenger et al. 2003).

Although the concepts of fitness valleys experienced by pathogens during the early phase of a host shift are frequently discussed in a theoretical context, our comparison of viruses across a large number of hosts linked the depth of fitness valleys to the transmission dynamics of host shifts (Antia et al. 2002; Kuiken et al. 2006). Specifically, we found that greater numbers of amino acid changes in positively selected sites were linked to longer lag times preceding viral establishment in a new host (Figure 4.4 A,C). We suspect that this relationship was stronger when we took into account ancestrally inferred host states because estimating the year of cross- species transmission as the date of the most recent common ancestor of sampled viruses almost

surely underestimates the age of each lineage, and subsequently, the lag time until viral growth.

From the viral perspective, the lags that we observed likely represent extended periods of low fitness during which adaptive changes in L and G can arise and increase in frequency, facilitating large-scale emergence and demographic growth. From a disease control perspective, long delays between cross-species transmission and host adaptation represent a key window of opportunity for public health or veterinary intervention.

The ability to predict how many amino acid changes separate existing reservoir species from potential new hosts is paramount to understand the speed of pathogen emergence and host adaptation. Our previous work showed that host shifts occurred disproportionately among closely related bat species, and analogous results have been observed in other host-virus systems

(Charleston & Robertson 2002; Streicker et al. 2010; Longdon et al. 2011). This suggests that the phylogenetic distance between possible host species may determine the depth of the fitness valley separating them. Here, we tested this hypothesis directly and found a positive correlation between the genetic distance between host species and the amino acid distance between their viral glycoproteins (Figure 4.5). This implies that the disproportionate frequency of actualized host shifts between closely related bat species might reflect the lesser need to alter sites associated with viral replication and virulence, though this remains an area in need of further investigation (Streicker et al. 2010). Curiously, no relationship with host genetic distance was found for L (Figure 4.5). This suggests that other ecological factors, perhaps such as differences in population density, colony size, migratory or overwintering behavior may be more important in setting the number of adaptive changes in L needed for establishment.

It is also noteworthy that we found that no viral genetic predictor of the rate, duration or extent of the epizootic growth phase in Bayesian skyline plots (the slope, length and height of the

shaded portions in Figure 4.3). One plausible explanation follows our argument above that most evolutionary adaptation occurs before epizootic growth begins. In this case, the details of epizootic growth might depend more heavily on the spatial spread of RV through host populations, as was suggested for raccoon RV (Biek et al. 2007). Factors such as the geographic range size, population density, social structure or dispersal ability of bats would likely be important for understanding the epizootic dynamics after evolutionary adaptation. More generally, the prospect that the majority of adaptive evolution occurring during times of minimal viral transmission provides a strong evolutionary incentive that reinforces calls for greater early pathogen detection through enhanced surveillance of wildlife (Kuiken et al. 2005).

In conclusion, our results show evidence for episodic positive selection on several RV genes during extended periods of low viral fitness in the early phases of host shifts. However, the extent and epizootiological consequences of adaptive evolution varied among host shifts, with some viruses emerging rapidly and with very few genetic changes, and others experiencing a prolonged lag phase during which multiple evolutionary changes occurred. Better understanding of the role of host genetics and ecology in driving this variability should be a key goal for future studies, and we highlight the utility of a comparative approach that analyzes multiple host shift events for exploring these effects.

MATERIALS AND METHODS

Collection and sequence amplification of bat rabies viruses. L and G sequences were generated for representative bat RV lineages using samples available in the collection of the

CDC Rabies Program. Total RNA was extracted directly from naturally infected rabid bat brains using TRIzol (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. For the G,

reverse-transcription polymerase chain reaction (RT-PCR) was used to amplify two overlapping fragments of 1078 base pair (bp) and 1205 bp, which encompassed 136 bp of the intergenic space between the matrix protein and G, the complete coding sequence of G (1575 bp) and 487 bp of the G-L intergenic region (Appendix C Table 1). For the L sequences, 5 overlapping RT-

PCR reactions of approximately 1000-1400 bp each were used to amplify the complete 6,387 bp coding region using a variety of primer sets (Appendix C Table 1). Thermocycling conditions followed Smith (2002). PCR amplicons were purified using ExoSAP-IT® (USB Corporation) or from bands excised from low melting agarose gels using Wizard PCR Clean-up kits (Promega).

Sequencing was carried out using the forward and reverse PCR primers from Table S1 as well as internal sequencing primers when needed using an Applied Biosystems 3730 capillary DNA sequencer. Chromatograms were edited by eye, aligned via Clustal W and assembled as contigs using Geneious Pro v.5.1.6 (Larkin et al. 2007; Drummond et al. 2010). The sequences generated herein will be deposited into Genbank upon submission. Additional RV nucleoprotein

(N) (n = 625), G (n = 39) and L (n = 3) sequences associated with bats were collected from

GenBank.

Estimating selection pressures along the viral genome. To identify the genomic location and timing of positive selection associated with host shifts, we estimated the non-synonymous (dN) and synonymous substitution (dS) rates for each of the N, G and L datasets, with dN/dS > 1 or dN - dS > 0 indicative of positive selection. For each gene, datasets were assembled that contained a maximum of 10 randomly selected, but unique sequences per lineage (N: n = 184; G: n = 68; L: n = 48). For each dataset, we estimated a phylogentic tree using 5 replicate ML searches with random starting trees in Garli v.096b8 under substitution models selected by jModeltest, using a

raccoon RV sequence as an outgroup (Zwickl 2006; Posada 2008). The tree with the highest log likelihood was used in analyses of selection after removing the outgroup sequence, except in ancestral sequence reconstruction. The selection pressures at specific codon sites were estimated using the fixed effects likelihood model which independently fits dN and dS to each codon position and compares the fit of these models to a null model assuming dN = dS via a likelihood ratio test with 1 degree of freedom (Kosakovsky Pond et al. 2006). Because the FEL method assumes the same dN/dS across all branches in the phylogeny for each site, an unlikely scenario in the case of viruses experiencing sudden shifts in host environment, we also used a mixed effects model of episodic selection (MEME) to relax the assumption of constant substitution rates through time. This model considered the dN/dS at each site as a fixed effect, while allowing for two categories of branches, those with dN/dS ≤ 1 and those with dN/dS > 1, which was treated as a random effect. This model was tested against a null that constrained all branches to have dN/dS <

1 (Kosakovsky Pond et al. 2011). Significant positive and purifying selection on codon positions were indicated by P values < 0.05 in the MEME analysis and < 0.1 in the FEL analysis, because the more conservative nature of the FEL test reduces the probability of type 1 errors.

Bayesian estimation of donor-recipient host species relationships for bat rabies lineages. To better understand in which host species positive selection occurred, we inferred the most probable donor host species of presently circulating viral lineages using Bayesian phlyogenetic ancestral state estimation. We built upon our previous analysis by allowing host shifts to occur asymmetrically between host species, by including lineages perpetuated by Central and South

American bat species and, most importantly, by integrating information from the N, G and L genes through hierarchical phylogenetic modeling (Suchard et al. 2003; Streicker et al. 2010).

For each of the 30 host-associated lineages, we selected a maximum of 3 representative isolates, preferentially choosing those isolates for which we had sequences from multiple genes. Thus, our dataset comprised 87 isolates with a total of 86, 58 and 44 sequences for the N, G and L genes respectively. Each alignment was treated as a separate data partition in a combined analysis that allowed us to estimate a consensus evolutionary history of bat rabies, while accounting for variable patterns of molecular evolution across genes. Specifically, our analysis linked the parameters describing the topology of the phylogenetic tree, the past viral demography and the ancestral host states across partitions. For each gene, molecular evolution was modeled with a separate general time reversible substitution model (including gamma distributed rates and a proportion of invariant sites; GTR + I + Γ) for the first and second codon positions (CP12) and for the third codon position (CP3), respectively. Each gene was also assigned a separate uncorrelated lognormal relaxed molecular clock to accommodate evolutionary rate differences among genes. However, because different genes from the same virus isolate are not evolutionarily independent, we hierarchically linked the parameters of the nucleotide substitution and molecular clock models by assuming that parameters for each gene vary around a shared unknown, but estimable population mean (Suchard et al. 2003). The hierarchical phylogenetic modeling substantially improved the marginal likelihood of the model relative to assuming independence of the evolutionary parameters across genes (Bayes Factor, BF = 72.55).

The ancestral state estimation model was implemented in BEAST v.1.6.1, which estimates genealogies and evolutionary parameters using a Bayesian Markov Chain Monte Carlo

(MCMC) simulation approach (Drummond & Rambaut 2007). Four replicate runs of 50 million generations each were conducted with genealogies and parameters sampled every 5,000 generations. The results of all four runs were combined after discarding the first 20% of each as a

burn-in. This search strategy yielded effective sample sizes > 200 for all parameters. To identify the donor-recipient relationships for each viral lineage, a Bayesian stochastic search variable selection procedure was implemented to allow discrete state change rates (i.e., the rate describing host shifts) to be zero, enabling the use of a BF test to identify non-zero rates (Suchard, Weiss &

Sinsheimer 2005; Lemey et al. 2009). Transition rates supported by a BF > 3 were considered significant support for a host shift between bat species. When multiple transitions to single viral lineages were supported by the BF analysis, we deferred to the host species with the greatest posterior probability on the branch leading to the new lineage as the most probable donor.

Calculating the amino acid changes since host shifts. To estimate the number of positively selected amino acid changes that occurred since the introduction of each virus into its present day host species, we reconstructed the historical nucleotide sequences of each gene along the branches of the phylogenetic tree using the marginal ML method implemented on the datamonkey webserver of the HyPhy software package (Yang, Kumar & Nei 1995; Pond &

Muse 2005; Delport et al. 2010). The ancestral sequence reconstruction assumed the GTR + Γ substitution model and the raccoon RV sequence was used as an outgroup for each ML tree.

Amino acid sequences were translated from the ancestral nucleotide sequences and the number of changes at positively selected sites since each host shift occurred was calculated in two ways.

First, we calculated the number of substitutions from the hypothetical lineage ancestor (the node of the branch preceding the most recent common ancestor of each lineage) and the consensus sequences of sampled viruses. Because this approach makes no assumptions about the directionality of host shifts, the substitutions that occurred in the donor host species prior to the most recent host shift are ignored. Therefore, using the statistically supported host states

assigned by the Bayesian ancestral state estimation as a guide, we calculated the number of amino acid changes from the nodes where host shifts were predicted to have occurred until the consensus sequence of presently circulating viruses.

Estimation of viral demographic histories since host shifts. To estimate the demographic histories of RV lineages since host shifts occurred, we constructed 13 datasets of N gene sequences for which a minimum of 20 sequences (mean = 45, range: 20 - 82) sampled over at least 10 years (mean = 22.3, range: 10-30 years) were available. Bayesian skyline plots were estimated simultaneously for all lineages in BEAST using a model that hierarchically linked evolutionary model parameters across viral lineages (considered as data partitions), while allowing for independent genealogical topologies and demographic histories (Drummond &

Rambaut 2007). Similar hierarchical approaches have proven useful for improving the precision of parameter estimation in analogous datasets that display partial dependence across viral lineages, while reducing the susceptibility of parameter estimates to the stochastic noise inherent in less informative datasets (Edo-Matas et al. 2010; Streicker et al. submitted). As above, our analysis assigned separate GTR substitution models to CP12 and CP3 and separate uncorrelated lognormal molecular clocks to each dataset, which were hierarchically linked across data partitions. All other parameters were allowed to vary independently across datasets. Four replicate MCMC runs were conducted for 100 million generations each. Trees and parameter estimates were sampled every 5000 generations and the results of all 4 runs were combined after discarding 20% of each as a burn-in period. Source files for all BEAST analyses are available from the corresponding author upon request.

Characterizing epizootic dynamics for bat rabies lineages. We characterized the amount of time between the introduction of each virus into its present day host species and the period of epizootic growth (‘epizootic lag time’) and the dynamics of that growth once initiated using the characteristics of each Bayesian skyline plot (Figure 4.3). The epizootic lag time was calculated as the difference between the year when epizootic growth began and estimated year of each host shift. The dates of host shifts were calculated in two ways depending on assumptions of when host shifts occurred. First, the origin of each host shift was conservatively estimated as the year of the most recent common ancestor of each lineage according to the Bayesian skyline models of the N gene. In the second calculation, we took advantage of the full phylogenetic information from the hierarchical analysis of N, G and L and included the stem branches leading to each clade inferred from the ancestral state estimation model, recognizing that the true date of each host shift is likely somewhere in between these extremes. To estimate the population growth rate of each lineage, r, we defined the growth period in each skyline plot as annual periods of ≥ 2.5% growth Neτ and calculated the slope of a linear regression through that segment of the demographic history. The duration and total fitness gain of the epizootic growth period were calculated as the length and height of the growth period, respectively.

FIGURES

A B

SP II a III TM 1 1 I/IV IV Ectodomain Endo

0 0

-1 -1

Normalized dN-dS Normalized -2 dN-dS Normalized -2

-3 -3

0 100 200 300 400 0 100 200 300 400 500

C Codon position Codon position

Catalytic domain ATP binding P interaction I II III IV V VI 0

-5

-10 Normalized dN-dS Normalized

-15

0 500 1000 1500 2000 Codon position

Figure 4.1. Site-specific selection pressures on the rabies virus (A) nucleoprotein, (B) glycoprotein and (C) polymerase genes. Black and red points indicate sites evolving under purifying (black) and positive (red) selection according to FEL models. Grey points show neutrally evolving sites. Red triangles denote sites under episodic positive selection only according to MEME models. For glycoprotein (B), structural regions are indicated by colored lines and text: SP = signal peptide, TM = transmembrane region, Endo = endodomain, and grey boxes indicate antigenic regions. The green asterisk is site 333 in the ectodomain. For polymerase (C), the catalytic domain, the potential ATP nucleotide binding region and the region of interaction with the phosphoprotein gene (P) are marked with colored lines and text, and the putative functional regions described by Poch (1990) are indicated by grey bars.

A. Nucleoprotein B. Glycoprotein C. Polymerase NhV EfV2 EfV2 EfV2 EfV3

PhV EfV3 McV

EfV3 MaV McV ApV MyV MaV PhV PhV MV2

M2V MySAV MaV MyV MySAV EfSAV MV2

EfSAV MV1 MV1

TbSAV MmV EfV1a EfV1b DrV TbSAV

TbV TbSAV TbV DrV HmV TbV

I DrV 0.01 NlV 0.01 0.01 PtV

MV1 LiV2 EfV1b EfV1b LiV1 EfV1a EfV1a PtV PsV MmV

LiV1/LiV2 LnV1 LeSAV

LnV1 LiV1/LiV2 96 LbV2 LnV2 PtV 204 PsV PsV 222 LxV LxV LnV1 236 LsV LsV LsV 637 638 LcV LxV 1319 643 1620 LcV 4 34 LbV1 61 68 373 888 1742 LbV1 LbV2 921 1743 107 139 389 970 1840 LbV1 108 264 448 160 LbV2 331 493 LcV 1012 2090 437 357 515 1019 2093

internal internal D E F 8 internal 8 8

6 tips 6 tips 6 tips

4 4 4

2 2 site per Substitutions Substitutions per site per Substitutions Substitutions per site per Substitutions

0 0 0 Site Site Site Figure 4.2. Episodic positive selection in the bat rabies virus phylogeny. Maximum likelihood topologies of the bat rabies virus N, G and L genes are shown in panels A, B and C respectively. Symbols indicate amino acid changes in positively selected sites only. Color spectra of points follow the relative position along each gene. Bar charts in panels D, E and F show the frequency of amino acid changes for each site with colors as in A-C. The hollow portions of bars indicate substitutions that occurred in single isolates and pie charts show the proportion of internal to tip substitutions for all sites in each gene.

DrV EfV1a EfV1b EfV2 EfV3 4 4 4 4 4

3 3 3 3 3

2 2 2 2 2

1 1 1 1 1

0 0 0 0 0

1940 1970 2000 1940 1970 2000 1940 1970 2000 1940 1970 2000 1940 1970 2000

dr$TimeLbV2 ef1a$TimeLcV ef1b$TimeLnV ef2$TimeM2V ef3$TimeNlV 4 4 4 4 4

3 3 3 3 3

2 2 2 2 2

1 1 1 1 1

0 0 0 0 0

1940 1970 2000 1940 1970 2000 1940 1970 2000 1940 1970 2000 1940 1970 2000 Effective number of rabies infections of rabies Effective number

lb2$TimePhV lc$TimePsV ln1$TimeTbV m2$Time nl$Time fective number of rabies infections of rabies fective number

Ef 4 4 4

3 3 3

2 2 2

1 1 1

0 0 0

1940 1970 2000 1940 1970 2000 1940 1970 2000 Year

Figure 4.3. Demographic histories of 13 bat rabies virus lineages. Each panel shows a Bayesian skyline plot estimated from the hierarchical phylogenetic model. The effective number of infections is the product of the effective population size and the average time between infections (Neτ). Red dashed lines are the 95% highest posterior density. The shaded gray region denotes the epizootic growth phase of each lineage. Time scales are limited to 1940 for uniformity across panels, but the effective number of infections remained constant for those lineages with older origins. Viral lineage labels as in Figure 4.2. Host species for each lineage are given in Appendix C, Table 2).

A B

400 400

350 350

300 300

250 250

200 200 Epizootic lag time (years) lag Epizootic Epizootic lag time (years) lag Epizootic 150 150 Glycoprotein

100 100 Polymerase

0 1 2 3 4 5 6 0 1 2 3 4 5 6 Total positively selected AA changes Total positively selected AA changes

C D

2.0 2.0

1.5 1.5

1.0 1.0 Log10[epizootic lag time] lag Log10[epizootic time] lag Log10[epizootic

0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 Positively selected AA changes Positively selected AA changes

Figure 4.4. Epizootic consequences of viral adaptation. Each panel plots the number of amino acid changes at positively selected sites against the lag time to epizootics inferred from Bayesian skyline plots alone (C,D) and using branches inferred from ancestral host state estimation models (A,B). Panels A and C show the combined number of non-synonymous substitutions in G and L. Panels B and D show the same relationships for each gene, with points jittered along the x axis for visualization. N was omitted because nearly all points fell on zero. Dashed lines show the fit of statistically significant or marginally significant models (see text for details). For panel A, an asymptotic model provided a slightly better fit than a linear model, which was also statistically 2 significant (linear: r = 0.34, F1,10 = 5.21, P = 0.046; asymptotic versus linear: P = 0.057).

0.6 Glycoprotein Polymerase

0.4

0.2

0.0 Log10[Amino acid distance at positively selected sites] selected at positively distance acid Log10[Amino

0.0 0.1 0.2 0.3 0.4 Genetic distance between bat species

Figure 4.5. Relationship between host genetic distance and the amino acid distance between associated viral variants at putatively positively selected sites. Points for the N gene were omitted because limited amino acid differences at positively selected sites caused most distances to be zero. Genetic distances between bats were estimated from partial cytochrome oxidase I sequences retrieved from Genbank. Mantel tests indicated a significant positive relationship with host genetic distance for G (Mantel r = 0.15, P = 0.045), and no relationship for L (Mantel r = - 0.07, P = 0.506).

CHAPTER 5

ECOLOGICAL AND ANTHROPOGENIC DRIVERS OF RABIES EXPOSURE IN VAMPIRE

BATS: IMPLICATIONS FOR TRANSMISSION AND CONTROL 4

______

4 Streicker, D.G., Recuenco, S., Valderrama, W., Gomez-Benavides, J., Vargas, I., Pacheco, V., Condori, E., Montgomery, J., and Rupprecht, C.E., Rohani, P. and Altizer, S. To be submitted to Proceedings of the Royal Society – B series.

ABSTRACT

Common vampire bats (Desmodus rotundus) are arguably the most important wildlife rabies reservoir in the Americas. Although this species is culled extensively, lethal human outbreaks are increasingly common and livestock mortality continues. We conducted a 4-year, capture- recapture study of 20 vampire bat colonies spanning 4 regions of Peru to identify the individual and population level factors that drive rabies transmission in wild vampire bats. Serology demonstrated the circulation of rabies virus antibodies among bats in all regions in all years.

Seroprevalence ranged from 3 to 28% and varied over time and space, perhaps reflecting many independent transmission cycles within regions. Rabies antibody prevalence was independent of bat colony size, suggesting that population thresholds for viral invasion and extinction are unlikely for this system. Finally, rates of rabies exposure were highest in juvenile and sub-adult bats and tended to increase following culling by humans, perhaps due to the targeted removal of adults, but greater importance of juvenile and sub-adult bats for rabies transmission. These findings provide insights into the mechanisms of rabies maintenance in vampire bat populations and suggest new directions to develop ecologically informed strategies for rabies prevention and control.

INTRODUCTION

Bats (Chiroptera) have been increasingly recognized as important reservoirs of emerging zoonotic RNA viruses including SARS coronavirus, Nipah virus, Marburg virus and Ebolavirus

(Chua et al. 2000; Leroy et al. 2005; Li et al. 2005; Swanepoel et al. 2007). A prerequisite to predicting the spatial and temporal dynamics of cross-species transmission risk is fundamental understanding of how highly lethal viruses are maintained within wild bat populations over space

and time (Calisher et al. 2006). Of particular relevance is the identification of the natural and anthropogenic factors that contribute to the prevalence of infection in bat populations and the subsets of bat populations and communities that are most important for long-term viral maintenance. In spite of their importance for understanding maintenance mechanisms and anticipating zoonotic risk, long-term field studies of virus dynamics in wild bat populations have been exceedingly rare (Amengual et al. 2007; Plowright et al. 2008; Wacharapluesadee et al.

2010). This substantially limits the development of informed strategies for controlling viruses within bat populations and predicting transmission risk to humans and domesticated animals.

Rabies virus, common in bat populations throughout the Americas, is the best-studied and arguably most important zoonotic virus of bats (WHO 2005). In bats and all other mammals, untreated infection causes an acute encephalitis with a case fatality ratio approaching 1

(Rupprecht, Hanlon & Hemachudha 2002). Of all bat rabies reservoirs, the common vampire bat

(Desmodus rotundus) causes the greatest human rabies burden due to its unique habit of feeding exclusively on mammalian blood, an ideal behavioral mechanism for transmission by biting events (Brass 1994b). Consequently, with greater human encroachment in Amazonian regions of

South America, lethal vampire bat transmitted human rabies outbreaks have become increasingly common the and now surpass the annual number of cases caused by dogs (Schneider et al. 2009).

In the agricultural zones of Latin America, the number of vampire bat transmitted rabies cases in livestock appears to have declined from the greater than 500,000 annual cases that were reported throughout the 1960’s, but cases remain commonplace (Baer 1991). Modern estimates from

Brazil alone indicate thousands of vampire-bat transmitted rabies cases in cattle every year and underreporting is very likely extensive in much of Latin America (Mayen 2003; Belotto et al.

2005). For example, 4% of 1000 cattle slaughtered for human consumption in Mexico City were

rabies positive (reviewed in Constantine 1988). Because rabies virus transmission either does not occur or is extremely rare within livestock and human populations, most of these infections result from direct contact with bats (Delpietro 1996).

Strategies to control vampire bat transmitted rabies in Latin America include vaccination of humans and livestock and reduction of bat populations by culling. Vaccination is biologically effective, but poses economic and logistical challenges that limit its practical efficacy in the developing Latin American countries where the vampire bat rabies problem is most severe.

Moreover, the anticipated growth of vampire bat populations with the novel food source provided by the expansion and intensification of the livestock rearing industry suggests that these costs will only become more prohibitive in the future (Koopman 1988; Delpietro, Marchevsky &

Simonetti 1992; FAO 2006). For humans, although bat bites are common in many parts of the

Amazon jungle, vaccination is predominately a reactive public health effort that is initiated only after cases are reported (Schneider et al. 2001; ProMed-mail 2010). Unfortunately, the geographic isolation of human settlements in the Amazon where rabies outbreaks are common can cause substantial delays in vaccine delivery that provide a strong human health incentive for educational or otherwise preventative approaches for reducing the risk of bat bites (Schneider et al. 1996; ProMED-mail 2007).

Large-scale culling of vampire bat populations for rabies control began in the early

1970’s following the development of “vampiricide,” an anticoagulent paste applied to captured bats and spread to other members of the colony by allogrooming after treated bats return to the roost (Linhart, Mitchell & Crespo 1972). The paste can also be placed on bite wounds of livestock to kill bats that return to feed from the same wound (Arellano-Sota 1988). Culling via vampiricide paste is most effective at killing adult bats that groom other adults and juveniles, but

is unlikely to kill younger bats that do not groom adults and have less interaction with livestock

(Wilkinson 1986; Greenhall 1988; DeNault & McFarlane 1995; Gomes, Uieda & Latorre 2006).

Vampiricide is used across Latin America including Brazil, Mexico and Peru; however, culling as currently implemented has proven insufficient for rabies virus elimination on even a regional scale (Arellano-Sota 1988). Moreover, culling in wildlife disease systems can actually increase disease prevalence when it stimulates the recruitment of susceptible individuals or increases host dispersal (Barlow 1996; Choisy & Rohani 2006; Donnelly et al. 2006; Woodroffe et al. 2006;

Bolzoni & De Leo 2007; Holt & Roy 2007). The persistence of vampire bat rabies despite culling could therefore reflect immigration from neighboring colonies to fill empty roosts, removal of adults that perhaps have protective immunity, or an increase in new births of susceptible bats following the relaxation of density dependent constraints.

Implementing successful control strategies and anticipating the frequency, intensity and duration of rabies outbreaks requires understanding the drivers of rabies transmission within vampire bat populations (Goncalves, Sa-Neto & Brazil 2002). To date, most knowledge of the disease in vampire bats has been inferred from patterns of rabies cases in livestock or experimental infections in captive bats (Lord et al. 1975; Moreno & Baer 1980; Delpietro &

Russo 1996; Aguilar Setien et al. 1998). This has left critical gaps in understanding viral transmission in natural populations. For example, the theoretical basis underpinning vampire bat culling to prevent rabies rests on two untested assumptions: first, that adult bats are important for transmission, and second, that virus transmission scales positively with bat density (McCallum,

Barlow & Hone 2001; Lloyd-Smith et al. 2005). The question of density dependent transmission is also a major limiting factor for predicting the frequency of rabies outbreaks in humans and lifestock. In the absence of density dependence, population thresholds for viral invasion would

not be expected to exist, so outbreaks would depend more heavily on stochastic factors and detectable rates of transmission to humans or livestock might occur according to the frequency of infection in the local bat population (McCallum, Barlow & Hone 2001).

Here, we report the results of a four year, capture-recapture field study of rabies exposure across 20 geographically widespread vampire bat colonies in Peru. Our specific goals were (i) to identify the population and individual-level factors that influence rates of rabies exposure in wild vampire bats, (ii) to test the relationship between bat density and virus transmission and (iii) to compare rates of exposure in bat populations that are affected by contrasting anthropogenic forces of culling and resource supplementation through livestock rearing. We predicted that seroprevalence would increase with host colony size (assuming a density-dependent component to rabies transmission) and that seroprevalence would be higher among bats captured near areas with high livestock densities (assuming that resource supplementation leads to higher host birth rates and recruitment of susceptibles). We further predicted that seroprevalence would be lower in regions where culling is commonly practiced owing to the removal of infected hosts and the overall lowering of host population density.

MATERIALS AND METHODS

Field sites and sampling design. Between July 2007 – October 2010, we sampled 20 vampire bat colonies in 4 departments of Peru: Apurimac, Cajamarca, Lima and Madre de Dios (Figure

5.1). Colonies were selected to represent the three major geographic regions of Peru that are inhabited by vampire bats (coastal deserts, Andean valleys and the Amazon jungle) and to span a spectrum of livestock and bat densities. The Peruvian Ministry of Health or the Peruvian

Ministry of Agriculture had previously located most colonies sampled here. Colonies in the

Amazon were located by capturing bats near livestock corrals or human settlements where bat bites had been reported and radio-tracking bats back to their roosts via airplane, motorcycle and foot (Radiotag model BD-2, 1.48g, Holohil Inc.). Colonies inhabited both natural (trees and caves) and man-made structures (mines, tunnels) and were separated by a minimum distance of

10 km.

Capture and sampling of wild vampire bats. Pilot studies carried out in 2007 sampled each colony for a single day or night. From 2008-2010, we conducted 3 - 6 night capture-recapture studies per year at most colonies to estimate colony sizes and sample individuals arriving from neighboring roosts. During these sessions, bats were captured in mist nets or harp traps placed at the exit of roosts from approximately 18:00 until 05:00. An additional 2 - 5 mist nets were placed within 50 meters of each roost over trails, streams and other likely flyways. In 2007 and 2010, diurnal captures were undertaken for colonies where it was possible to enter roost sites to increase the sampling of juvenile bats. During these sessions, bats were captured using aerial insect nets, and mist nets were placed at the cave or tunnel exit to catch escaping bats.

Upon capture, bats were temporarily stored in individual cloth bags and hung from a clothesline in the order of capture. Bats were weighed in the cloth bags and each bat was issued one or two unique 4-digit incoloy bat bands (3.5mm, Porzana Inc.). Age was classified as juvenile, sub-adult or adult based on the degree of fusion of the phalangeal epiphyses (Anthony

1988). Juveniles included volant individuals that had a gap of 0.5 mm or greater in the phalangeal epiphyses, corresponding to an age of approximately 2 – 6 months. Sub-adults were classified by lack of complete fusion of the phalangeal epiphyses (< 0.5 mm), but otherwise had adult sized forearms, weight and pelage, corresponding an age of approximately 6 - 9 months

(Delpietro & Russo 2002). The classification of juveniles and sub-adults was confirmed by two independent observers. Reproductive activity in adults was indicated in males by the presence of scrotal testes and in females by pregnancy or lactation.

For serological assays, a maximum of 250 µl of whole blood was collected by lancing the propatagial vein with a sterile 23 gauge needle. Blood was collected with heparinized capillary tubes, centrifuged in the field using serum separator tubes and stored on cold packs until it could be frozen (typically 0-3 days). When necessary, styptic powder was applied to cease bleeding.

After sample collection, bats were released at the site of capture. Individuals that were recaptured in the same night were released immediately and excluded from capture-recapture studies. The

University of Georgia’s Institutional Animal Care and Use Committee approved protocols for the capture and handling of bats (AUP # A2009-10003-0) and collection permits were obtained from the Peruvian government (103-2008-INRENA-IFFS-DCB; RD-222-2009-AG-DGFFS-DGEFFS;

RD-0299-2010-AG-DGFFS-DGEFFS). Samples were shipped to the CDC rabies laboratory

(Atlanta, GA, USA) on dry ice under export permits 003851-AG-DGFFS, 004692-AG-DGFFS and 005216-AG-DGFFS.

Detection of rabies virus neutralizing antibodies in bat sera. The rapid fluorescent focus inhibition test (RFFIT) is the WHO laboratory standard method for detecting the presence of rabies virus neutralizing antibodies (Smith, Yager & Baer 1996; WHO 2005). This test compares the ability of serial dilutions of serum to inhibit infection of mouse neuroblastoma cells with a constant dose of challenge virus. We used the modified RFFIT described by Kuzmin et al.

(2008) using four-well (6mm) Teflon-coated glass slides to accommodate the small volumes of serum that could be collected from bats. Briefly, a 3.5 µl aliquot of serum was diluted to 1:10

and 1:25 in minimum essential medium supplemented with 10% fetal calf serum, inoculated with

12.5 µl of rabies virus (laboratory strain CVS-11) and incubated on mouse neuroblastoma cells for 20 hours. Cells were then fixed in acetone and stained with a fluorescein isothiocyanate- labeled anti-rabies virus conjugate (Fujirebio Diagnostics Inc.) and 10 separate fields dilution were read under a florescent microscope. Samples were considered positive if they showed 100% neutralization of at least 5 fields at the 1:10 dilution or if they showed > 50% reduction of the area of infected cells relative to that of the negative control. To calculate the area of infected cells, we captured 4-5 digital images at the 1:10 dilution and counted the number of fluorescent green (infected cell area) and red (healthy cell area) pixels using using Adobe Photoshop software with Fovea Pro plugins (Reindeer Graphics, Inc.), The proportion of infected cell area was calculated for each image and for the negative control slides. Samples were considered positive if the 95% confidence interval of the proportion of infected cells fell outside that of the negative control x 0.5. Since antibody titers decline to undetectable levels within 5 months in experimentally inoculated bats (Eptesicus fuscus), we assumed that the presence of virus neutralizing antibodies reflected recent exposure to rabies virus (Jackson et al. 2008). Therefore, observations across years were treated as independent, except where noted otherwise.

Estimation of the size of bat colony sizes. We estimated the size of each bat colony using capture-recapture models. Because multiple day sampling periods for most sites took place only in 2009 and 2010, we restricted inference to those years, with the exception of one site in Madre de Dios that was also intensively sampled in 2008. We treated each 3-6 day sampling period per site per year as an independent population and used the loglinear closed population models in the rcapture package of R v.2.12 to estimate capture probabilities and colony sizes (Baillargeon &

Rivest 2009; R Development Core Team 2011). For each sampling session, a set of 8 models with different assumptions on the sources of variation in capture probabilities (capture date, individual heterogeneity and behavioral responses to previous captures) were fit to the data and compared by Akaike’s information criterion (AIC). Confidence intervals on colony size estimates were generated by profile likelihood from the model with the lowest AIC score.

Data on livestock rearing and culling activity. To test the effects of livestock rearing on bat colony sizes and rabies exposure, we used data from the 1994 Agricultural Census of Peru

(CENAGRO III) to calculate livestock densities as the number of cows, sheep, pigs and goats within 5km2 of each colony using district level data (approximately equivalent to US districts).

Culling history was described at two different time scales. First, a three-level variable categorized colonies that (i) had never been culled, (ii) were periodically culled during the study and (iii) were regularly culled between 2007 and 2010. Second, we recorded whether colonies were culled during the 12-month period immediately prior to sampling. Colonies were categorized using data and interviews of personnel from the regional offices of the Ministry of

Health and/or the Ministry of Agriculture, the main groups that carry out vampire bat control.

Colonies that were destroyed or directly impacted by unsanctioned activities (e.g., logging of roost trees, lighting fires in caves and direct capture/killing of bats with nets) were included in the periodic culling category.

Statistical analysis. Spatial and temporal variation in seroprevalence among and within sites was assessed using χ2 tests for count data. Additional statistical analyses used generalized linear mixed models and generalized linear models to identify the population level factors that

contribute to vampire bat colony sizes and the individual and population-level factors that influence exposure to rabies virus in individual bats.

Anthropogenic effects on vampire bat colony size. We tested the expected opposing effects of culling and livestock rearing on vampire bat colony sizes estimated from capture-recapture studies using a generalized linear mixed modeling (GLMM) approach. The full model included fixed effects describing the density of livestock within 5km of bat roosts, whether culling had occurred within the 4 years prior to sampling, the departmental region of Peru, and the interaction between culling and livestock density. Colony size and livestock density were square root transformed to fit model assumptions. Site was included as the only random effect after comparing the AIC scores of competing random effects components under the full fixed effects model (Zuur et al. 2009). Other random effects components considered included hierarchical nesting of site within department and department alone. The fixed effects portion of the model was simplified by stepwise removal of model terms, followed by model comparison using nested likelihood ratio tests. Term removals that resulted in significant reduction in the explanatory power of the model (p = 0.05) were retained in the minimal adequate model. The 95% highest posterior densities (HPD) on effect sizes were generated using Markov Chain Monte Carlo

(MCMC) sampling of the posterior distribution of the minimal adequate model using the mcmcsamp function in the lme4 package of R (Bates, Maechler & Bolker 2011; R Development

Core Team 2011).

Individual and population level risk factors for rabies exposure. We assessed the factors contributing to rabies exposure in vampire bats using a binomial GLMM (logit link). The full fixed effects component contained age, sex, reproductive status, forearm length, colony size, livestock density and the two measures of culling history along with biologically meaningful

two-way interactions (age by sex, sex by reproductive status). As above, we retained site as a random effect after testing several competing random effects structures (combinations of year, site, department and site nested within department) by AIC (Zuur et al. 2009). Because variance component analysis indicated that none of the random effects explained a substantial proportion of variance in rabies exposure in the full fixed effects model, the retention of site was mainly based on our study design. We confirmed the non-significance of other random effects by testing their influence on the minimal model when treated as fixed effects. The minimal adequate fixed effect model structure was identified using a truncated dataset (N = 816 observations) to accommodate missing values for some explanatory values. As above, we used stepwise removal of non-significant fixed effects, starting with interactions. The effect sizes of the simplified model were then estimated from full dataset for which there were no missing observations in significant terms (N = 1040 observations); however, it was not possible to estimate HPDs on effect sizes by MCMC sampling because these methods are not yet implemented for binomial response variables.

RESULTS

Colony size and population dynamics of vampire bats. Over 4 years, we captured 1436 unique vampire bats. Recaptures were relatively common both within (N = 216) and between years (N =

132), with individuals recaptured up to 3 years after initial capture (the maximum duration permitted by our study design). Vampire bats demonstrated a remarkable regional fidelity with no instances of roost switching observed over the study period. Colony sizes estimated by capture-recapture models varied from 16 to 444 individuals and generally remained stable over years, with the exception of one statistically significant increase and 2 significant decreases

(Table 5.1). Both significant decreases in colony size (AP13, MD134) involved intentional human destruction of roosts. The significant increase in LM4 in 2010 may indicate re- colonization of that cave after the cessation of mining activity on the outer wall.

Surprisingly, the GLMM analysis found no effect of culling history on bat colony size (χ2

= 1.91, df = 2, p = 0.38). However, bat colony size differed among departments (χ2 = 10.8, df =

3, p = 0.01) and according to livestock density (χ2 = 4.75, df = 1, p = 0.03). On average, colonies were smaller in the Amazonian region of Madre de Dios (effect size, β = -12.89 [95% HPD = -

22.48 – -3.89]) relative to the Andean regions of Cajamarca (β = 7.68 [95% HPD = 1.13 –

14.03]) and Apurimac (baseline effect size). Colony sizes in the coastal department of Lima overlapped with both Andean and Amazon regions (β = -2.47 [95% HPD = -7.77 – 3.31]).

Livestock densities varied by department in the same pattern as colony sizes (Cajamarca >

Apurimac > Lima > Madre de Dios) leading to an overall positive relationship between livestock density and bat density. However, because department explained much of the variance in livestock density, when both were included in the model, livestock density actually had a significant negative effect on colony size (β = -0.41 [95% HPD = -0.82 – 0.014]). This negative effect appeared to be driven by a single large colony in Lima (LM6) found in an area of relatively low livestock density.

Spatial and temporal patterns of rabies exposure in vampire bats. Analysis of 1086 serum samples revealed a global seroprevalence of 10.2% across all years and sites. The vast majority of individuals that were captured and sampled over multiple years were seronegative in all years

(N = 78/86). Seroconversion, loss of detectable antibody titer and maintenance of antibodies across years occurred in 2, 5 and 1 individuals, respectively.

Across years, seropositive vampire bats were found in all sites where greater than 4 individuals were sampled and prevalence in these colonies varied from 3.3 – 28.6%.

Seroprevalence was relatively stable over time in most sites; however, significant shifts in seroprevalence or putative disease extinctions and recolonizations occurred in AP13, LM10 and

LM6 (χ2 test: p < 0.05 for each; Figure 5.1). Within years, sites within same department had significantly different levels of rabies exposure in Lima in 2009 (χ2 = 25.63, p < 0.001), and in

Apurimac and Cajamarca in 2010 (Apurimac: χ2 = 10.56, p = 0.014; Cajamarca: χ2 = 12.99, p =

0.005), but other department-year combinations could not be statistically differentiated.

Individual and population level predictors of rabies exposure in vampire bats. The minimal adequate GLMM retained only two factors: age and culling history within the 4-year study period, with very little support for effects of reproductive status, body size (as measured by forearm length), colony size, or the density of livestock on exposure to rabies virus (Table 5.2).

Notably, rabies exposure did not increase linearly with age, but instead was elevated in juveniles relative to adults (odds ratio, OR = 3.01), peaked in sub-adults (OR = 3.58), then declined in adult bats (Figure 5.2 A). Inspection of the culling effect showed that rabies exposure was higher in bat colonies that were subjected to periodic (OR = 2.03) and regular culling (OR = 1.43) compared to those that were never culled during the 4 years of the study; although the increase in regularly culled colonies was not statistically different from undisturbed or periodically culled colonies (Figure 5.2B).

We also tested the robustness of the GLMM to different model structures. Results remained unchanged when year and department, which were excluded from the random effects component, were included as fixed effects and neither of these was significant (department: χ2 =

2.71, d.f. = 3, p = 0.44; year: χ2 = 1.17, d.f. = 2, p = 0.55). Similarly, although both our inter- annual recapture data and past experimental infections suggested that the serological status of bats over long time spans is largely driven by recent exposures, we repeated the GLMM analysis using only a single randomly selected blood sample per recaptured bat to account for possible non-independence of samples from the same individual. We also repeated the analysis excluding

6 colonies that contained other bat species because cross-species exposures might cause seroconversion in D. rotundus. In each of the latter cases, age and culling retained statistical support; however, in the analysis using roosts with D. rotundus only, females had significantly higher seroprevalence than males (OR = 1.98, χ2 = 5.0, d.f. = 1, p = 0.025). Notably, females also had slightly higher rates of exposure to rabies virus than males in the original GLMM (11.43% versus 9.16%); however, that difference was only marginally significant (Table 5.2).

Because the relationship between colony size and seroprevalence was central to our study but not supported in the GLMM analysis, we conducted a separate logistic regression analysis to confirm the absence of a relationship after removing potentially correlated factors. This analysis suggested a very weak positive association between colony size and seroprevalence, but this trend only was statistically significant when excluding two site/year combinations that had unusually high seroprevalence (Figure 5.3). Importantly, both models had a significantly positive y-intercept, suggesting the absence of lower population threshold for viral invasion; although stochastic viral extinctions might be more likely in smaller colonies.

DISCUSSION

Culling vampire bats to reduce their population size is a principal tool used by many Latin

American countries to control the problem of vampire bat transmitted rabies in humans and

livestock. The fundamental assumption behind culling to reduce wildlife disease is that pathogen transmission increases with host population size. We tested this assumption by surveying rabies exposure in Peruvian vampire bat colonies that showed natural variation in size and in colonies with and without a history of culling. Our results demonstrated that exposure to rabies virus was ubiquitous across geographically widespread vampire bat populations, was at best, only weakly related to bat colony size, and tended to increase following culling activity.

The most powerful predictor of rabies exposure based on our analyses was host age class, with higher rates of exposure in juvenile and sub-adult bats relative to adults. This was somewhat surprising because we expected that the feeding and social behavior of juveniles would expose them less frequently to rabies infections. Nevertheless, we argue that this pattern reflects early- life exposures rather than maternally derived antibodies because average seroprevalence was higher in juveniles than adult females and tended to increase rather than decrease during the first year of life (Figure 5.2 A). Frequent exposure of young bats to rabies virus was previously suggested in serological studies of the Mexican free-tailed bat (Tadarida brasiliensis) (Steece &

Altenbach 1989; Turmelle et al. 2010a). Similarly, a disproportionate role for juveniles in rabies transmission was indicated in a survey of rabies positive bats submitted for diagnostic testing in

Canada, where and juveniles comprised 38.5 – 84.2% of rabies positive bats, depending on species (Dorward, Schowalter & Gunson 1977). It therefore appears that rabies exposure and infection is common in young bats, and this might contribute to the high juvenile mortality rates observed in vampire bats (> 50%) (Wilkinson 1984). High infection rates in juvenile bats may also facilitate long-term maintenance at the population level, as was suggested in mathematical models of rabies in an insectivorous bat, Eptesicus fuscus, by providing annual pulses of susceptible individuals (George et al. 2011).

Elevated rates of rabies exposure in young bats could also explain the positive association between periodic culling and seroprevalence (Figure 5.2 B, Table 5.2). Specifically, because the vampiricide paste used for culling disproportionately targets adult bats, which also have lower seroprevalence, this selective removal might numerically increase population level seroprevalence. A shift in age structure could in turn affect transmission dynamics because as social groups reform after disturbance, a greater proportion of contacts from infected individuals are with highly susceptible animals. Demographic mechanisms acting over longer time scales, such as increased immigration from neighboring colonies (the ‘vacuum effect’) or increased survival or births following a sudden release from density dependent constraints, could further increase transmission by augmenting the susceptible bat population. At some locations, it is possible that bats were killed in response to the detection of rabies cases in neighboring livestock or other animals. This may explain the high seroprevalence at site AP13 in September 2011, where farmers captured bats in nets and lit fires in caves after two cows died of rabies in August

2010 (SENASA Weekly Epidemiological Reports 32-2010 and 37-2010). However, if this were more generally the case, we would have expected culling during the year prior to sampling, rather than the general culling strategy over the 4-year period, to have been the better predictor of rabies exposure in bats (Table 5.2).

Regardless of whether culling increased rabies transmission, numerically increased seroprevalence by disproportionately killing adults, or occurred in response to rabies outbreaks in livestock, it is apparent that reducing vampire bat populations failed to reduce or eliminate regional rabies virus circulation (Figure 5.1, Figure 5.2B). This failure of culling to reduce the risk of infection in a wildlife reservoir has also been observed for other systems, most notably bovine tuberculosis in badgers in the UK, where reactive culling of badgers after cases were

detected in neighboring cattle actually increased tuberculosis prevalence (Donnelly et al. 2006)

Among vampire bats, two mechanisms might explain the persistence of rabies despite culling.

First, it is possible that culling was simply not implemented at a frequency or scale sufficient to have a demographic effect on the overall size of bat colonies. Indeed, we found that regularly culled colonies were not significantly smaller than undisturbed colonies and that culling in the year prior to our capture-recapture surveys only caused significant reductions in colony size when it involved total or near total destruction of roost sites (Table 5.1). In this case, sub-optimal culling might be sufficient to shift the age structure of bat colonies without having dramatic effects on total colony size. While this effect might have the unfortunate consequence of increasing rabies transmission at low levels of culling, it leaves the open possibility that more intense culling or coordinated campaigns within and among regions might reach a goal of rabies elimination. A second explanation for the failure of culling to reduce rabies seroprevalence is the lack of a strong relationship between bat density and rabies transmission (Table 5.2, Figure 5.3).

Thus, contact rates between susceptible and infected bats might be independent of colony size, such that the per capita force of infection depends more heavily on the prevalence of infection than overall population density (i.e., frequency dependent transmission) (Begon et al. 2002).

Although we know of no other study that has examined the relationship between bat colony size and virus transmission, the absence of such a relationship for a directly transmitted pathogen such as rabies virus is perhaps not surprising given that even in large colonies, any single animal will have a limited number of neighbors to bite. For vampire bats, infectious contacts may be even less homogenized because of the complex social networks that exist within colonies

(Wilkinson 1988).

An important consequence of the uncoupling of transmission from population density is that bat population thresholds for rabies virus invasion or persistence are unlikely (McCallum,

Barlow & Hone 2001). Our findings of seropositive animals across a range of colony sizes from tens to hundreds of bats, and significant, positive y-intercept terms for the relationship between colony size and seroprevalence are consistent with the absence of bat population thresholds for viral invasion (Table 5.2, Figure 5.3). This implies a serious challenge for rabies management because no reasonable level of culling will be sufficient to eliminate rabies or prevent re-invasion after stochastic viral extinction from the bat population. Indeed, even when vampire bat populations in northern Argentina were experimentally reduced by 95% by gassing roosts with cyanide, rabies cases in bats and livestock were still reported as close as 1 km from the area of elimination (Fornes et al. 1974). The feasibility of achieving this degree of bat population reduction over even modest spatial scales is at best daunting in flat agricultural landscapes such as the Argentinian chaco and likely impossible in mountainous or jungle landscapes such as those in the Andes and Amazon jungle, where many roost sites are unknown or inaccessible.

Important questions remain about the spatial scale of vampire bat rabies maintenance.

Previous studies have described vampire bat rabies as a slowly migrating epizootic, where the virus spreads from colony to colony, killing some bats and immunizing others, but preventing reinvasion until a threshold number of susceptible bats is reached (Fornes et al. 1974; Delpietro

& Russo 1996). Although detailed studies of livestock rabies cases supported the slow spread of epizootics, several lines of evidence cast doubt on whether viral extinction occurs. First, our serological data indicated the continual presence of rabies virus in most sites across a several year period (Figure 5.1). This is more likely to reflect the local circulation of rabies virus than long term maintenance of antibodies because (i) we observed several instances of antibody loss

in recaptured bats and (ii) controlled infections of other bat species have shown that antibody titers decline to undetectable levels in approximately 5 months after initial infection and do not recover without subsequent exposures (Jackson et al. 2008; Turmelle et al. 2010b). Second, in regions of Peru as well as other countries in Latin America, livestock rabies mortality is recurrent over many years within confined geographic regions, and phylogenetic studies of vampire bat rabies virus have revealed that these represent many independently circulating viral lineages

(Velasco-Villa et al. 2006; Kobayashi et al. 2008).

We suggest that the apparent conflict between hypotheses of viral perpetuation by a migrating epizootic versus enzootic local maintenance can be explained by the existence of two fundamentally different epidemic phases, such that an initial epizootic wave into naïve populations is followed by local enzootic cycles that persist indefinitely. Such dynamics have been described most clearly for rabies in North American raccoons using case report and phylogenetic data. After the epizootic spread of raccoon rabies virus through the eastern United

States, enzootic maintenance was achieved by recurrent small outbreaks that did not require immigration of infected individuals (Biek et al. 2007). A re-assessment of livestock rabies mortality data suggests that analogous dynamics may occur in vampire bats. When patterns of vampire bat-transmitted livestock cases from the same region of Argentina were analyzed by

Fornes et al. (1974) from 1959 – 1974 and by Delpietro and Russo (1996) from 1984 – 1993, the spatial signatures of spread that were apparent in the first epizootic had disappeared in later years, consistent with local enzootic maintenance. Rabies control programs would greatly benefit from acknowledging these alternative scenarios because local enzootic maintenance implies that cases in livestock or humans will occur sporadically, making them far less predictable than in the scenario of slow viral spread across a landscape.

Finally, we note that our results could be affected by sympatric bat species also maintain independent viral transmission cycles (Kobayashi et al. 2007). Although cross-species transmission from other bats to D. rotundus has never been reported and vampire bats aggressively segregate themselves from other bat species in mixed species roosts, cross-species exposures could cause seroconversion in vampire bats and thereby influence our estimates of seroprevalence (Turner 1975; Wohlgenant 1994). We minimized the potential for cross-species exposures by focusing our capture efforts on roosts that were occupied only by D. rotundus and by verifying the robustness of our results to the exclusion of those sites where other species were present. A second limitation arose from uncertainty in the scale of vampire bat population sizes

(single roost site to regional population size) that would be most relevant for modeling rabies transmission. We attempted to generate an intermediate estimate by (i) always working in the largest known colony in each region (ii) by placing arrays of mist nets in the area surrounding roost sites in order to survey passing individuals and (iii) by working at each roost for multiple nights to catch individuals arriving from nearby roosts. Although we observed no roost switching between our relatively distant sites (>10km), previous work suggests that dispersal occurs between closer roosts (Trajano 1996). Thus, genetic estimates of effective population size could be used to confirm the absence of density dependence in rabies transmission that we observed.

In conclusion, we provided evidence that rabies virus is maintained at a regional scale in vampire bat populations over multiple years, and this is perhaps mediated through frequency dependent transmission with a key role for juvenile and sub-adult bats. The absence of population thresholds for rabies virus invasion and persistence is consistent with the observed inefficacy of culling to eliminate viral circulation in bats and transmission to humans and domesticated animals. Although culling bat populations benefits livestock through alleviating bat

bites and may reduce the total number of infected bats and subsequently instances of cross- species transmission, the apparent positive effect of culling on seroprevalence, coupled with demographic responses that might increase the proportion of susceptible bats, might have unforeseen consequences for transmission. Future work should experimentally identify the mechanistic basis for the relationship between culling and seroprevalence and use that information to inform epidemiological models of optimal rabies control strategies.

FIGURES AND TABLES

CA1

CA4

CA3

2010

CA2 2009

2007

LM5 2008 MD2

MD130

LM6 LM11 MD134

AP4

AND13 LM10

LM8 AP140

LM4 AP138 AP1 AP3 AP9

Figure 1. Map of study sites in Peru with spatio-temporal patterns of rabies exposure. White points show sampling locations. Colored regions indicate the governmental departments that were sampled (red = Apurimac, green = Madre de Dios, blue = Lima, orange = Cajamarca). Pie charts show the proportion of seropositive (white) vampire bats in each site in each year,

with pie diameter proportional to sample size (range = 6 - 102). Colonies with ≤ 5 samples in a single year were classified as seropositive (open circle) or seronegative (filled circle). Dashed lines connect sites across years and orbits around Peru group sites by years.

0.5 0.20 A B AB 0.4 0.15 A 0.3 A 0.10

0.2 Serorevalence Seroprevalence

B 0.05 0.1

0.0 0.00

Juveniles Subadults Adults Never Periodic Regular N=85 N=31 N=923 N=371 N=523 N=192 Age Class Culling history during study

Figure 2. Effects of age and culling history on rabies exposure in vampire bats. Error bars are 95% confidence intervals. Letters indicate statistically significant differences among groups (p < 0.05) in generalized linear mixed models. Sample sizes for groups are indicated below bar labels.

0.4

2008 0.3 2009 2010 0.2

Seroprevalence 0.1

0.0

0 100 200 300 400 500

Colony size

Figure 3. The relationship between vampire bat colony size and seroprevalence. The size of symbols is proportional to the number of bats sampled in each site (range = 2 – 102, mean = 35.7), with different shapes for each year that capture-recapture studies were conducted. Colors indicate the departmental region as in Figure 1. Consistent with the GLMM analysis, a logistic regression found no significant association between colony size and seroprevalence (z = 0.71, d.f. = 1, p = 0. 479). When the two sites with unusually high seroprevalence were removed (see top left portion of figure), this relationship became weakly, but significantly positive (z = 2.138, d.f. = 1 p = 0. 031); however, the y-intercept remained significantly greater than zero suggesting the absence of a strict population threshold for viral invasion. The dotted and dashed lines show model predictions of prevalence for the full and partial datasets, respectively.

Table 5.1. Capture-recapture abundance estimates for vampire bat colonies in Peru

2009 2010 Model ✪ Site Region Department District § CE N (95% CI) Model CE N (95% CI)

AP1 Andes Apurimac Huaccana Mt 4 132 (62-416) Mt 4 227 (62-723) † AP13 Andes Apurimac Abancay Mb 4 303 (224-446) Mt 3 43 (15-150) AP9 Andes Apurimac Pacucho Mt 4 50 (34-97) Mt 4 164 (88-420) CA1 Andes Cajamarca Chota M0 4 191 (146-269) Mt 4 224 (178-299) CA2 Andes Cajamarca Cutervo Mt 4 408 (253-767) Mt 3 392 (314-515) CA3 Andes Cajamarca Cutervo Mt 4 371 (98-1170) M0 3 95 (39-304) LM10 Coast Lima Chancay Mt 3 75 (59-109) M0 4 105 (76-163) † LM4 Coast Lima Mala Mb 4 22 (16-29) Mt 3 112 (36-373) 352 (154- LM6 Coast Lima Huacho Mt 3 Mth 4 444 (319-670) 1086) LM8 Coast Lima Huaral Mb 3 16 (14-24) ND 1 ND Madre de Madre de MD130 Amazon Mt 6 50 (26-160) M0 4 19 (7-68) Dios Dios Madre de † MD134* Amazon Inambari M0 6 17 (14-24) M0 4 19 (7-68) Dios

§ Capture-recapture models are defined as follows: M0 = equal capture probability across nights; Mt = temporal heterogeneity in capture probability; Mth = capture probability is variable among individuals and across nights; Mb = capture probability differs after first capture (i.e., a behavioral effect). ✪ CE: capture effort in number of nights. † Statistically significant increase or decrease in prevalence according to non-overlap of 95% confidence intervals. * Colony size in 2008 by M0 = 98 (50 - 289); 7 nights of capture.

Table 5.2. Generalized linear mixed model analysis of the individual and population level factors influencing rabies exposure in vampire bats. Bold text indicates inclusion in the minimal model. Significance was tested by likelihood ratio tests comparing the minimal model to models including each term (or excluding in the case of age and 4 year culling). Two non- significant two-way interactions were omitted from the table. All models contained site as a random effect and a significant y intercept term.

factor χ2 d.f. p age 16.41 2 < 0.001 culling (4 year) 6.72 2 0.035 sex 3.39 1 0.066 reproductive 2.39 1 0.121 colony size 2.02 1 0.225 livestock density 0.26 1 0.607 forearm 0.005 1 0.943 culling (1 year) 0.0005 1 0.982

CHAPTER 6

CONCLUDING REMARKS AND FUTURE DIRECTIONS

The overarching goal of this dissertation research was to dissect the complex and multi-faceted process of viral host shifts into their ecological and evolutionary components to test fundamental ideas of how and when viruses emerge. The upwelling of newly emerging and re-emerging diseases in humans, domesticated animals and wildlife has made these some of the most critical questions to modern public health, veterinary medicine and wildlife conservation (Daszak,

Cunningham & Hyatt 2000; Lloyd-Smith et al. 2009). Yet, I chose to focus much of my work on understanding seemingly inconsequential host shifts between bat species rather than say, host shifts from bats to domesticated animals or terrestrial wildlife that might have more direct implications. In some sense, this was a pragmatic decision based on the samples and animals that

I was able to get my hands on. However, as I progressed through my research, I began to understand that taking this approach allowed me to answer a number of questions that simply could not be answered in many other systems. The key problem with studying host shifts, especially dramatic ones, is that most of the time they only happen once. As a consequence, it is difficult to draw general conclusions about virus ecology and evolution from a sample size of a single event, which might actually have been a very unlikely event in the first place. Answering broad questions about a phenomenon that cannot be replicated required taking a comparative approach, and rabies virus provided a unique perspective as a system where the behavior of a similar virus could be assessed in variety of ecologically and evolutionarily distinct host settings.

Specifically, my research was structured around 3 central themes: (i) the influence of inter-specific host heterogeneity on disease transmission (Chapters 2 and 3), (ii) identifying the genomic mechanisms involved in host adaptation and accessing how the extent of that genomic change determines the epidemiological dynamics of viral emergence (Chapter 4), and (iii) using better understanding of viral maintenance within the reservoir to improve strategies to mitigate future disease emergence (Chapter 5). In each case, I made an attempt to tackle general questions in disease ecology and evolution, while contributing to understanding of a specific disease system of relevance to human and animal health.

Chapter 2 (Streicker et al. 2010) will probably be the most scientifically influential contribution of this dissertation. That chapter described the first empirical test of the popular hypothesis that because of their potential for rapid evolution, the emergence of RNA viruses is limited more by ecological constraints than by genetic similarity between donor and recipient hosts. My results showed that despite their potential for rapid evolution, rabies viruses are more likely to jump between closely related than anciently diverged bat species. By illustrating that the origins of emerging viruses may be both constrained and broadly predicted by the evolutionary relatedness of host species, these results prompted reconsideration of factors driving the emergence of RNA viruses. Moreover, this chapter presented a novel “transmission web” framework based on food web theory, which may be influential in understanding transmission dynamics in complex multi-host systems. However, this work was rather limited in the breadth of the evolutionary relatedness of species tested. An important area for future work will be to extend similar analyses to other systems to test the relationship between host phylogeny and disease emergence across larger phylogenetic distances.

Although Chapter 2 suggested that the capacity for rapid evolution did not guarantee a successful outcome of cross-species transmission for rabies virus, it was clear from other systems that genetic change is often associated with RNA virus establishment in new host species (Song et al. 2005; Anishchenko et al. 2006). Similarly, Chapter 4 provided evidence that specific genetic changes were associated with the adaptation of rabies virus to new host species.

Therefore, in Chapter 3, I explored the influence of host ecology and seasonality on the tempo of rabies virus molecular evolution with the underlying thought that more rapidly evolving viruses might have an advantage for emergence in new host species. This study demonstrated very well that viral evolutionary rates are malleable in different host species. Interestingly, no correlation between viral evolutionary rates and host phylogeny was observed, reinforcing the argument presented in Chapter 2 that rapid evolution per se does not obviate host barriers to emergence.

Synthesizing these two chapters, it appears that the rate of viral evolution may be more a function of post-establishment host-virus ecology than an intrinsic feature of viruses that makes some more adept at crossing the species barrier than others, at least in relatively rapidly evolving viruses such as rabies virus. From a more general standpoint, Chapter 3 was an exploration of the mechanisms driving pathogen molecular evolution. If we are to fully grasp the variability that exists in rates of RNA virus evolution in multi-host systems, it may be necessary to extend exclusively virus-oriented explanations to include the role of host biology.

Chapters 2 and 3 took a very broad view of how host ecology and evolution influence viral transmission; however, in both cases the comparative nature of the work took center stage and the mechanistic explanations for patterns were to me, rather vague and unsatisfying.

Accordingly, Chapter 4 represented an attempt to integrate more detailed understanding of the molecular biology of rabies virus infection with the epidemiological processes that characterize

cross-species emergence. I am not aware of previous analyses that have taken a similar approach, perhaps because of the problems mentioned above that arise from the singular nature of most host shifts. The clearest result that arose was that even for the same virus, the evolutionary route to host adaptation can be quite context dependent. However, what actually defined that context and hence, the degree of difficulty for a host shift, again returned to the realm of the vague and unsatisfying. The contribution of Chapter 4 to the biology of rabies virus was more redeeming.

Viral self-regulation of within host spread appeared to be a key component of host adaptation, which perhaps was not so surprising given that immunological and cellular constraints were expected to be minimal in this system. Nevertheless, the work presented a large new dataset and from it, a new suite of specific genomic regions that pave a way forward to uncover the specific mechanisms of host adaptation.

Although Chapter 5 was the most preliminary and perhaps the least scientifically novel of the research chapters presented in this dissertation, it may ultimately prove to be the most impactful for human and animal health. Vampire bat-transmitted rabies is a growing problem in

Latin America that can only be expected to worsen with the expansion of human and livestock populations. However, the sheer commonness of vampire bat-transmitted rabies cases in humans and livestock (thousands every year) seems to have contributed to its largely neglected status relative to other zoonotic viruses of bats, such as SARS, Nipah or Ebola, whose more rare, but explosive outbreaks cause far less human and animal mortality on an annual basis. Consequently, research on the transmission dynamics of rabies virus in vampire bats has lagged behind even that on rabies in insectivorous bats despite its greater public health and veterinary importance.

The work presented in Chapter 5 challenged two of the canonical assumptions about rabies in vampire bats that arose from studies in the 1970s and had since grown to become

conventional wisdom (Fornes et al. 1974; Lord et al. 1975; Delpietro & Russo 1996; Mayen

2003). First, that bat population thresholds for viral invasion and maintenance exist and second, that spatial migration is a key mechanism for producing outbreaks. The correctness of these assumptions forms the theoretical foundation for the use of vampire bat population control programs to prevent rabies throughout Latin America. My results, though far from conclusive, were inconsistent with these contentions. The epidemiological situation of vampire bat rabies appeared to be a more complicated story than one of spatial spread alone, and ironically, seemed more consistent with notions of rabies maintenance in most other reservoir species (i.e., an initial epizootic period followed by enzootic maintenance mediated by low prevalence of infection and a variable incubation period) (Biek et al. 2007; George et al. 2011). Interestingly, the early observations of massive outbreaks and epizootic spread from the 1930s – 1970s might then reflect the original introduction of rabies virus into vampire bat populations as was suggested by the molecular data in Chapters 3 and 4, and consequently a relatively recent origin for bat rabies in the Americas (Pawan 1936; Lord et al. 1975).

Clearly, vampire bat rabies remains an area where much research is desperately needed.

It will be exciting to continue to collect data on rabies virus in wild vampire bat populations and synthesize that information through phylogenetic and mathematic inference to gain a better understanding of this important zoonosis. Vampire bat rabies is also an encouraging system from an applied standpoint because control programs already exist, indicating a widespread acknowledgement of the problem and an existing resource base for large scale implementation of new ideas. Close collaboration with governmental ministries in Peru has laid a strong foundation for the possibility of implementing more developed research findings. Accordingly, a key future

direction for this work will be the development of ecologically and epidemiologically informed strategies for rabies prevention and evaluating these through large-scale field experiments.

LITERATURE CITED

Aguilar Setien, A., Brochier, B., Tordo, N., de Paz, O., Desmettre, P., Peharpre, D. & Pastore, P.P. (1998) Experimental rabies infection and oral vaccination in vampire bats (Desmodus rotundus). Vaccine, 16, 1122-1126.

Aguirre, A.A. (2002) Conservation medicine: ecological health in practice. Oxford University Press, USA.

Altizer, S., Bartel, R. & Han, B.A. (2011) Animal migration and infectious disease risk. Science, 311, 296-302.

Amengual, B., Bourhy, H., Lopez-Roig, M. & Serra-Cobo, J. (2007) Temporal dyamics of European Bat Lyssavirus Type 1 and survival of Myotis myotis in natural colonies. Plos One, 2, E566.

Anderson, R.M. & May, R.M. (1991) Infectious diseases of humans - dynamics and control. Oxford University Press, Oxford, U.K.

Anishchenko, M., Bowen, R.A., Paessler, S., Austgen, L., Greene, I.P. & Weaver, S.C. (2006) Venezuelan encephalitis emergence mediated by a phylogenetically predicted viral mutation. Proc. Nat. Acad. Sci. U.S.A., 103, 4994-4999.

Anthony, E.L.P. (1988) Age determination in bats. Ecological and behvioral methods for the study of bats (ed. T.H. Kunz), pp. 47-58. Smithsonian Institution Press, Washington D.C.

Antia, R., Regoes, R.R., Koella, J.C. & Bergstrom, C.T. (2002) The role of evolution in the emergence of infectious diseases. Nature, 246, 658-661.

Arechiga-Ceballos, N., Velasco-Villa, A., Shi, M., Flores-Chavez, S., Barron, B., Cuevas- Dominguez, E., Gonzalez-Origel, A. & Aguilar-Setien, A. (2010) New rabies virus variant found during an epizootic in white-nosed coatis from the Yucatan Peninsula. Epidemiology and Infection, 138, 1586-1589.

Arellano-Sota, C. (1988) Biology, ecology, and control of the vampire bat. Review of Infectious Diseases, 10, S615.

Badrane, H. & Tordo, N. (2001) Host switching in Lyssavirus history from the chiroptera to the carnivora orders. Journal of Virology, 75, 8096-8104.

Baer, G.M., Shaddock, J.H., Quirion, R., Dam, T.V. & Lentz, T.L. (1990) Rabies susceptibility and the actylcholine receptor. Lancet, 335, 664-665.

Baillargeon, S. & Rivest, L.P. (2009) Rcapture: Loglinear Models for CaptureñRecapture Experiments. R package version, 1.2-0.

Bangham, C. & Kirkwood, T. (1993) Defective interfering particles and virus evolution. Trends in Microbiology, 1, 260.

Barlow, N.D. (1996) The ecology of wildlife disease control: Simple models revisited. Journal of Applied Ecology, 33, 303-314.

Bates, D., Maechler, M. & Bolker, B.M. (2011) lme4: Linear mixed-effects models using S4 classes.

Begon, M., Bennett, M., Bowers, R.G., French, N.P., Hazel, S.M. & Turner, J. (2002) A clarification of transmission terms in host-microparasite models: numbers, densities and areas. Epidemiology and Infection, 129, 147-153.

Belotto, A., Leanes, L., Schneider, M., Tamayo, H. & Correa, E. (2005) Overview of rabies in the Americas. Virus Research, 111, 5-12.

Benmansour, A., Leblois, H., Coulon, P., Tuffereau, C., Gaudin, Y., Flamand, A. & Lafay, F. (1991) Antigenicity of rabies virus glycoprotein. Journal of Virology, 65, 4198.

Biek, R., Henderson, J.C., Waller, L.A., Rupprecht, C.E. & Real, L.A. (2007) A high-resolution genetic signature of demographic and spatial expansion in epizootic rabies virus. Proceedings of the National Academy of Sciences of the United States of America, 104, 7993-7998.

Bininda-Emonds, O.R.P., Cardillo, M., Jones, K.E., MacPhee, R.D.E., Beck, R.M.D., Grenyer, R., Price, S.A., Vos, R.A., Gittleman, J.L. & Purvis, A. (2007) The delayed rise of present-day mammals. Nature, 446, 507-512.

Blanton, J.D., Robertson, K., Palmer, D. & Rupprecht, C.E. (2009) Rabies surveillance in the United States during 2008. Journal of the American Veterinary Medical Association, 235, 676-689.

Blomberg, S.P., Garland, T. & Ives, A.R. (2003) Testing for phylogenetic signal in comparative data: Behavioral traits are more labile. Evolution, 57, 717-745.

Bolzoni, L. & De Leo, G.A. (2007) A cost analysis of alternative culling strategies for the eradication of classical swine fever in wildlife. Environment and Development Economics, 12, 653-671.

Bourhy, H., Kissi, B., Audry, L., Smreczak, M., Sadkowska-Todys, M., Kulonen, K., Tordo, N., Zmudzinski, J.F. & Holmes, E.C. (1999) Ecology and evolution of rabies virus in Europe. Journal of General Virology, 80, 2545.

Bradley, C.A. & Altizer, S. (2007) Urbanization and the ecology of wildlife diseases. Trends in Ecology & Evolution, 22, 95-102.

Brass, D.A. (1994a) Clinical manifestations of rabies in insectivorous bats. Rabies in Bats: Natural History and Public Health Implications (ed. D.A. Brass), pp. 151-162. Livia Press, Ridgefield.

Brass, D.A. (1994b) Epizootiology of vampire-bat rabies. Rabies in Bats: Natural History and Public Health Implications (ed. D.A. Brass), pp. 85-100. Livia Press, Ridgefield, Connecticut.

Breed, A.C., Field, H.E., Epstein, J.H. & Daszak, P. (2006) Emerging henipaviruses and flying foxes - Conservation and management perspectives. Biological Conservation, 131, 211- 220.

Bromham, L. (2002) Molecular clocks in reptiles: Life history influences rate of molecular evolution. Molecular Biology and Evolution, 19, 302-309.

Brown, C.R., Komar, N., Quick, S.B., Sethi, R.A., Panella, N.A., Brown, M.B. & Pfeffer, M. (2001) Arbovirus infection increases with group size. Proceedings of the Royal Society of London Series B-Biological Sciences, 268, 1833-1840.

Brown, P.A., Lupini, C., Catelli, E., Clubbe, J., Ricchizzi, E. & Naylor, C.J. (2011) A single polymerase (L) mutation in avian metapneumovirus increased virulence and partially maintained virus viability at an elevated temperature. Journal of General Virology, 92, 346.

Calcagno, V. (2011) glmulti: GLM model selection and multimodel inference made easy. R package version 1.0.1, http://CRAN.R-project.org/package=glmulti.

Calisher, C.H., Childs, J.E., Field, H.E., Holmes, K.V. & Schountz, T. (2006) Bats: important reservoir hosts of emerging viruses. Clin Microbiol Rev, 19, 531-545.

Charleston, M.A. & Robertson, D.L. (2002) Preferential host switching by primate lentiviruses can account for phylogenetic similarity with the primate phylogeny. Systematic Biology, 51, 528-535.

Chen, R.B. & Holmes, E.C. (2006) Avian influenza virus exhibits rapid evolutionary dynamics. Molecular Biology and Evolution, 23, 2336-2341.

Choisy, M. & Rohani, P. (2006) Harvesting can increase severity of wildlife disease epidemics. Proceedings of the Royal Society B-Biological Sciences, 273, 2025-2034.

Chua, K., Bellini, W., Rota, P., Harcourt, B., Tamin, A., Lam, S., Ksiazek, T., Rollin, P., Zaki, S. & Shieh, W.J. (2000) Nipah virus: a recently emergent deadly paramyxovirus. Science, 288, 1432.

Cleaveland, S., Laurenson, M.K. & Taylor, L.H. (2001) Diseases of humans and their domestic mammals: pathogen characteristics, host range and the risk of emergence. Philos. Trans. R. Soc. London Ser. B, 356, 991-999.

Constantine, D.G. (1988) Transmission of Pathogenic Microorganisms by Vampire Bats. The Natural History of Vampire Bats (eds A.M. Greenhall & U. Schmidt), pp. 167-190. CRC Press, Boca Raton.

Constantine, D.G. (2009) Bat rabies and other lyssavirus infections. U.S. Geological Survey Circular 1329, Reston, VA.

Daszak, P. (2010) Bats, in Black and White. Science, 329, 634.

Daszak, P., Cunningham, A.A. & Hyatt, A.D. (2000) Emerging infectious diseases of wildlife: threats to biodiversity and human health. Science, 287, 443-449.

Daszak, P., Cunningham, A.A. & Hyatt, A.D. (2001) Anthropogenic environmental change and the emergence of infectious diseases in wildlife. Acta Tropica, 78, 103-116.

Davies, T.J. & Pedersen, A.B. (2008) Phylogeny and geography predict pathogen community similarity in wild primates and humans. Proc. R. Soc. London Ser. B, 275, 1695-1701. de Vienne, D.M., Hood, M.E. & Giraud, T. (2009) Phylogenetic determinants of potential host shifts in fungal pathogens. Journal of Evolutionary Biology, 22, 2532-2541.

Delpietro, H.A. (1996) Ecological and epidemiological aspects of attacks by vampire bats in relation to paralytic rabies in Argentina, and an analysis of proposals for control. Rev. sci. tech. Off. int. Epiz., 15, 971- 984.

Delpietro, H.A., Marchevsky, N. & Simonetti, E. (1992) Relative Population-Densities and Predation of the Common Vampire Bat (Desmodus-Rotundus) in Natural and Cattle- Raising Areas in North-East Argentina. Preventive Veterinary Medicine, 14, 13-20.

Delpietro, H.A. & Russo, R.G. (1996) Ecological and epidemiological aspects of attacks by vampire bats in relation to paralytic rabies in Argentina, and an analysis of proposals for control. Revue Scientifique Et Technique De L Office International Des Epizooties, 15, 971-984.

Delpietro, H.A. & Russo, R.G. (2002) Observations of the common vampire bat (Desmodus rotundus) and the hairy-legged vampire bat (Diphylla ecaudata) in captivity. Mammalian Biology - Zeitschrift fur Saugetierkunde, 67, 65-78.

100

Delport, W., Poon, A.F.Y., Frost, S.D.W. & Kosakovsky Pond, S.L. (2010) Datamonkey 2010: a suite of phylogenetic analysis tools for evolutionary biology. Bioinformatics, 26, 2455.

DeNault, L.K. & McFarlane, D.A. (1995) Reciprocal altruism between male vampire bats, Desmodus rotundus. Animal Behaviour, 49, 855-856.

Dietzschold, B., Gore, M., Marchadier, D., Niu, H., Bunschoten, H., Otvos Jr, L., Wunner, W., Ertl, H., Osterhaus, A. & Koprowski, H. (1990) Structural and immunological characterization of a linear virus-neutralizing epitope of the rabies virus glycoprotein and its possible use in a synthetic vaccine. Journal of Virology, 64, 3804.

Dietzschold, B., Wiktor, T.J., Trojanowski, J.Q., Macfarlan, R., Wunner, W., Torres-Anjel, M. & Koprowski, H. (1985) Differences in cell-to-cell spread of pathogenic and apathogenic rabies virus in vivo and in vitro. Journal of Virology, 56, 12.

Dietzschold, B., Wunner, W.H., Wiktor, T.J., Lopes, A.D., Lafon, M., Smith, C.L. & Koprowski, H. (1983) Characterization of an antigenic determinant of the glycoprotein that correlates with pathogenicity of rabies virus. Proceedings of the National Academy of Sciences, 80, 70.

Domingo, E. & Holland, J.J. (1997) RNA virus mutations and fitness for survival. Annual Review of Microbiology, 51, 151-178.

Donnelly, C.A., Woodroffe, R., Cox, D.R., Bourne, F.J., Cheeseman, C.L., Clifton-Hadley, R.S., Wei, G., Gettinby, G., Gilks, P., Jenkins, H., Johnston, W.T., Le Fevre, A.M., McInerney, J.P. & Morrison, W.I. (2006) Positive and negative effects of widespread badger culling on tuberculosis in cattle. Nature, 439, 843-846.

Dortmans, J.C.F.M., Rottier, P., Koch, G. & Peeters, B.P.H. (2011) Passaging of a Newcastle disease virus pigeon variant in chickens results in selection of viruses with mutations in the polymerase complex enhancing virus replication and virulence. Journal of General Virology, 92, 336.

Dorward, W., Schowalter, D. & Gunson, J. (1977) Preliminary studies of bat rabies in Alberta. The Canadian Veterinary Journal, 18, 341.

Drake, J.W. & Holland, J.J. (1999) Mutation rates among RNA viruses. Proceedings of the National Academy of Sciences of the United States of America, 96, 13910-13913.

Drummond, A., Ashton, B., Buxton, S., Cheung, M., Cooper, A., Heled, J., Kearse, M., Moir, R., Stones-Havas, S. & Sturrock, S. (2010) Geneious v5. 1. Biomatters Ltd., Auckland, New Zealand.

Drummond, A.J. & Rambaut, A. (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology, 7, 214.

101

Duffy, S., Shackelton, L.A. & Holmes, E.C. (2008) Rates of evolutionary change in viruses: patterns and determinants. Nature Reviews Genetics, 9, 267-276.

Dunbar, M.B. & Brigham, R.M. (2010) Thermoregulatory variation among populations of bats along a latitudinal gradient. Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology, 1-9.

Edo-Matas, D., Lemey, P., Tom, J.A., Serna-Bolea, C.l., van den Blink, A.E., van ‚Äòt Wout, A.l.B., Schuitemaker, H. & Suchard, M.A. (2010) Impact of CCR5delta32 host genetic background and disease progression on HIV-1 intra-host evolutionary processes: efficient hypothesis testing through hierarchical phylogenetic models. Molecular Biology and Evolution, 28, 1605-1616.

FAO (2006) Livestock policy brief 03: cattle ranching and deforestation.

Fenton, A. & Pedersen, A.B. (2005) Community epidemiology framework for classifying disease threats. Emerg. Infect. Dis., 11, 1815-1821.

Field, H., Young, P., Yob, J.M., Mills, J., Hall, L. & Mackenzie, J. (2001) The natural history of Hendra and Nipah viruses. Microbes and Infection, 3, 307-314.

Finke, S. & Conzelmann, K.K. (2005) Replication strategies of rabies virus. Virus Research, 111, 120-131.

Fisher, R.A. (1930) The genetical theory of natural selection. Clarendon Press, New York: Dover, Oxford.

Fornes, A., Lord, R.D., Kuns, M.L., Larghi, O.P., Fuenzalida, E. & Lazara, L. (1974) Control of bovine rabies through vampire bat control. Journal of Wildlife Diseases, 10, 310.

Freckleton, R. (2009) The seven deadly sins of comparative analysis. Journal of Evolutionary Biology, 22, 1367-1375.

George, D.B., Webb, C.T., Farnsworth, M.L., O'Shea, T.J., Bowen, R.A., Smith, D.L., Stanley, T.R., Ellison, L.E. & Rupprecht, C.E. (2011) Host and viral ecology determine bat rabies seasonality and maintenance. Proceedings of the National Academy of Sciences, 108, 10208.

Gilbert, G.S. & Webb, C.O. (2007) Phylogenetic signal in plant pathogen-host range. Proc. Nat. Acad. Sci. U.S.A., 104, 4979-4983.

Gomes, M.N., Uieda, W. & Latorre, M. (2006) Influence of sex differences in the same colony for chemical control of vampire Desmodus rotundus (Phyllostomidae) populations in the state of Sao Paulo, Brazil. Pesquisa Veterinaria Brasileira, 26, 38-43.

102

Goncalves, M.A.S., Sa-Neto, R.J. & Brazil, T.K. (2002) Outbreak of aggressions and transmission of rabies in human beings by vampire bats in northeastern Brazil. Revista da Sociedade Brasileira de Medicina Tropical, 35, 453-460.

Greenhall, A.M. (1988) Feeding Behavior. Natural History of Vampire Bats (eds A.M. Greenhall & R.K. Schmidt), pp. 7-17. CRC Press, Inc., Boca Raton, FL.

Guan, Y., Zheng, B.J., He, Y.Q., Liu, X.L., Zhuang, Z.X., Cheung, C.L., Luo, S.W., Li, P.H., Zhang, L.J., Guan, Y.J., Butt, K.M., Wong, K.L., Chan, K.W., Lim, W., Shortridge, K.F., Yuen, K.Y., Peiris, J.S.M. & Poon, L.L.M. (2003) Isolation and characterization of viruses related to the SARS coronavirus from animals in Southern China. Science, 302, 276-278.

Haagmans, B.L., Andeweg, A.C. & Osterhaus, A. (2009) The Application of Genomics to Emerging Zoonotic Viral Diseases. PLoS Pathogens, 5, 5.

Hanada, K., Suzuki, Y. & Gojobori, T. (2004) A large variation in the rates of synonymous substitution for RNA viruses and its relationship to a diversity of viral infection and transmission modes. Molecular Biology and Evolution, 21, 1074-1080.

Holland, J.J. (1987) Defective interfering rhaboviruses. Plenum Press, New York.

Holmes, E.C. (2009) The evolution and emergence of RNA viruses. Oxford University Press, USA.

Holmes, E.C. (2011) The comparative genomics of viral emergence. Proceedings of the National Academy of Sciences of the United States of America, 107, 1742-1746.

Holmes, E.C., Woelk, C.H., Kassis, R. & Bourhy, H. (2002) Genetic constraints and the adaptive evolution of rabies virus in nature. Virology, 292, 247-257.

Holt, R.D. & Roy, M. (2007) Predation can increase the prevalence of infectious disease. American Naturalist, 169, 690-699.

Hsu, V.P., Hossain, M.J., Parashar, U.D., Ali, M.M., Ksiazek, T.G., Kuzmin, I., Niezgoda, M., Rupprecht, C., Bresee, J. & Breiman, R.F. (2004) Nipah virus encephalitis reemergence, Bangladesh. Emerging Infectious Diseases, 10, 2082-2087.

Hughes, G.J., Orciari, L.A. & Rupprecht, C.E. (2005) Evolutionary timescale of rabies virus adaptation to North American bats inferred from the substitution rate of the nucleoprotein gene. Journal of General Virology, 86, 1467-1474.

Jackson, F.R., Turmelle, A.S., Farino, D.M., Franka, R., McCracken, G.F. & Rupprecht, C.E. (2008) Experimental rabies virus infection of big brown bats (Eptesicus fuscus). Journal of Wildlife Diseases, 44, 612-621.

103

Jenkins, G.M., Rambaut, A., Pybus, O.G. & Holmes, E.C. (2002) Rates of molecular evolution in RNA viruses: A quantitative phylogenetic analysis. Journal of Molecular Evolution, 54, 156-165.

Kembel, S.W., Cowan, P.D., Helmus, M.R., Cornwell, W.K., Morlon, H., Ackerly, D.D., Blomberg, S.P. & Webb, C.O. (2010) Picante: R tools for integrating phylogenies and ecology. Bioinformatics, 26, 1463-1464.

Kissi, B., Badrane, H., Audry, L., Lavenu, A., Tordo, N., Brahimi, M. & Bourhy, H. (1999) Dynamics of rabies virus quasispecies during serial passages in heterologous hosts. Journal of General Virology, 80, 2041-2050.

Kitchen, A., Shackelton, L.A. & Holmes, E.C. (2011) Family level phylogenies reveal modes of macroevolution in RNA viruses. Proceedings of the National Academy of Sciences of the United States of America, 108, 238-243.

Kobayashi, Y., Sato, G., Kato, M., Itou, T., Cunha, E.M.S., Silva, M.V., Mota, C.S., Ito, F.H. & Sakai, T. (2007) Genetic diversity of bat rabies viruses in Brazil. Archives of Virology, 152, 1995-2004.

Kobayashi, Y., Sato, G., Mochizuki, N., Hirano, S., Itou, T., Carvalho, A., Albas, A., Santos, H., Ito, F. & Sakai, T. (2008) Molecular and geographic analyses of vampire bat-transmitted cattle rabies in central Brazil. BMC Veterinary Research, 4, 44.

Koopman, K.F. (1988) Systematics and Distribution. Natural History of Vampire Bats (eds A.M. Greenhall & R.K. Schmidt), pp. 7-17. CRC Press, Inc., Boca Raton, FL.

Kosakovsky Pond, S.L. & Frost, S.D.W. (2005) Not so different after all: a comparison of methods for detecting amino acid sites under selection. Molecular Biology and Evolution, 22, 1208.

Kosakovsky Pond, S.L., Frost, S.D.W., Grossman, Z., Gravenor, M.B., Richman, D.D. & Brown, A.J.L. (2006) Adaptation to Different Human Populations by HIV-1 Revealed by Codon- Based Analyses. PLoS Comput Biol, 2, e62.

Kosakovsky Pond, S.L., Murrell, B., Fourment, M., Frost, S.D.W., Delport, W. & Scheffler, K. (2011) A random effects branch-site model for detecting episodic diversifying selection. Molecular Biology and Evolution.

Kucera, P., Dolivo, M., Coulon, P. & Flamand, A. (1985) Pathways of the early propagation of virulent and avirulent rabies strains from the eye to the brain. Journal of Virology, 55, 158.

Kuiken, T., Holmes, E.C., McCauley, J., Rimmelzwaan, G.F., Williams, C.S. & Grenfell, B.T. (2006) Host species barriers to influenza virus infections. Science, 312, 394-397.

104

Kuiken, T., Leighton, F., Fouchier, R., LeDuc, J., Peiris, J., Schudel, A., Stˆhr, K. & Osterhaus, A. (2005) Pathogen surveillance in animals. Science, 309, 1680.

Kurath, G., Garver, K.A., Troyer, R.M., Emmenegger, E.J., Einer-Jensen, K. & Anderson, E.D. (2003) Phylogeography of infectious haematopoietic necrosis virus in North America. Journal of General Virology, 84, 803.

Kuzmin, I.V., Niezgoda, M., Franka, R., Agwanda, B., Markotter, W., Beagley, J.C., Urazova, O.Y., Breiman, R.F. & Rupprecht, C.E. (2008) Lagos bat virus in Kenya. Journal of Clinical Microbiology, 46, 1451.

Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J. & Higgins, D.G. (2007) Clustal W and clustal X version 2.0. Bioinformatics, 23, 2947- 2948.

Lemey, P., Rambaut, A., Drummond, A.J. & Suchard, M.A. (2009) Bayesian Phylogeography Finds Its Roots. PLoS Computational Biology, 5, Article No.: e1000520.

Leroy, E.M., Kumulungui, B., Pourrut, X., Rouquet, P., Hassanin, A., Yaba, P., Delicat, A., Paweska, J.T., Gonzalez, J.P. & Swanepoel, R. (2005) Fruit bats as reservoirs of Ebola virus. Nature, 438, 575-576.

Leslie, M.J., Messenger, S., Rohde, R.E., Smith, J., Cheshier, R., Hanlon, C. & Rupprecht, C.E. (2006) Bat-associated rabies virus in skunks. Emerging Infectious Diseases, 12, 1274- 1277.

Li, W.D., Shi, Z.L., Yu, M., Ren, W.Z., Smith, C., Epstein, J.H., Wang, H.Z., Crameri, G., Hu, Z.H., Zhang, H.J., Zhang, J.H., McEachern, J., Field, H., Daszak, P., Eaton, B.T., Zhang, S.Y. & Wang, L.F. (2005) Bats are natural reservoirs of SARS-like coronaviruses. Science, 310, 676-679.

Linhart, S.B., Mitchell, G.C. & Crespo, R.F. (1972) Control of vampire bats by topical application of an anticoagulent (chlorophacinone). Boletin De La Oficina Sanitaria Panamericana, 73, 100-&.

Lloyd-Smith, J.O., Cross, P.C., Briggs, C.J., Daugherty, M., Getz, W.M., Latto, J., Sanchez, M.S., Smith, A.B. & Swei, A. (2005) Should we expect population thresholds for wildlife disease? Trends in Ecology & Evolution, 20, 511-519.

Lloyd-Smith, J.O., George, D., Pepin, K.M., Pitzer, V.E., Pulliam, J.R.C., Dobson, A.P., Hudson, P.J. & Grenfell, B.T. (2009) Epidemic Dynamics at the Human-Animal Interface. Science, 326, 1362-1367.

105

Longdon, B., Hadfield, J.D., Webster, C.L., Obbard, D.J. & Jiggins, F.M. (2011) Host Phylogeny Determines Viral Persistence and Replication in Novel Hosts. PLoS Pathogens, 7, e1002260.

Lord, R.D., Fuenzalida, E., Delpietro, H., Larghi, O.P., de Diaz, A.M. & Lazaro, L. (1975) Observations on the epizootiology of vampire bat rabies. Bull Pan Am Health Organ. , 9, 189-195.

Martin, A.P. & Palumbi, S.R. (1993) Body size, metabolic rate, generation time, and the molecular clock. Proceedings of the National Academy of Sciences of the United States of America, 90, 4087-4091.

Mayen, F. (2003) Haematophagous bats in Brazil, their role in rabies transmission, impact on public health, livestock industry and alternatives to an indiscriminate reduction of bat population. Journal of Veterinary Medicine Series B-Infectious Diseases and Veterinary Public Health, 50, 469-472.

McCallum, H., Barlow, N. & Hone, J. (2001) How should pathogen transmission be modelled? Trends in Ecology & Evolution, 16, 295-300.

McNab, B.K. (1974) The behavior of temperate cave bats in a subtropical environment. Ecology, 55, 943-958.

Messenger, S.L., Smith, J.S., Orciari, L.A., Yager, P.A. & Rupprecht, C.E. (2003) Emerging pattern of rabies deaths and increased viral infectivity. Emerging Infectious Diseases, 9, 151-154.

Messenger, S.L., Smith, J.S. & Rupprecht, C.E. (2002) Emerging epidemiology of bat-associated cryptic cases of rabies in humans in the United States. Clinical Infectious Diseases, 35, 738-747.

Moreno, J.A. & Baer, G.M. (1980) Experimental Rabies in the Vampire Bat. American Journal of Tropical Medicine and Hygiene, 29, 254-259.

Morimoto, K., Patel, M., Corisdeo, S., Hooper, D.C., Fu, Z.F., Rupprecht, C.E., Koprowski, H. & Dietzschold, B. (1996) Characterization of a unique variant of bat rabies virus responsible for newly emerging human cases in North America. Proc. Nat. Acad. Sci. U.S.A., 93, 5653-5658.

Moya, A., Holmes, E.C. & Gonzalez-Candelas, F. (2004) The population genetics and evolutionary epidemiology of RNA viruses. Nature Reviews: Microbiology, 2, 279-288.

Nei, M. (1987) Molecular Evolutionary Genetics. Columbia University Press, New York, Ny.

106

Omatsu, T., Watanabe, S., Akashi, H. & Yoshikawa, Y. (2007) Biological characters of bats in relation to natural reservoir of emerging viruses. Comparative Immunology Microbiology and Infectious Diseases, 30, 357-374.

Orme, C.D.L., Freckleton, R.P., Thomas, G., Petzoldt, T., Fritz, S.A. & Isaac, N. (2011) caper: Comparative Analysis of Phylogenetics and Evolution in R.

Parrish, C.R., Holmes, E.C., Morens, D.M., Park, E.C., Burke, D.S., Calisher, C.H., Laughlin, C.A., Saif, L.J. & Daszak, P. (2008) Cross-species virus transmission and the emergence of new epidemic diseases. Microbiology and Molecular Biology Reviews, 72, 457-470.

Pawan, J.L. (1936) The transmission of paralytic rabies in Trinidad by the vampire bat (Desmodus rotundus). Annals of Tropical Medicine and Parasitology, 30, 101.

Pedersen, A.B., Altizer, S., Poss, M., Cunningham, A.A. & Nunn, C.L. (2005) Patterns of host specificity and transmission among parasites of wild primates. International Journal for Parasitology, 35, 647-657.

Plowright, R.K., Field, H.E., Smith, C., Divljan, A., Palmer, C., Tabor, G., Daszak, P. & Foley, J.E. (2008) Reproduction and nutritional stress are risk factors for Hendra virus infection in little red flying foxes (Pteropus scapulatus). Proceedings of the Royal Society B: Biological Sciences, 275, 861.

Plowright, R.K., Foley, P., Field, H.E., Dobson, A.P., Foley, J.E., Eby, P. & Daszak, P. (2011) Urban habituation, ecological connectivity and epidemic dampening: the emergence of Hendra virus from flying foxes (Pteropus spp.). Proceedings of the Royal Society B: Biological Sciences.

Poch, O., Blumberg, B.M., Bougueleret, L. & Tordo, N. (1990) Sequence comparison of five polymerases (L proteins) of unsegmented negative-strand RNA viruses: theoretical assignment of functional domains. Journal of General Virology, 71, 1153.

Pond, S. & Muse, S. (2005) HyPhy: hypothesis testing using phylogenies. Statistical methods in molecular evolution, 125-181.

Posada, D. (2008) jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution, 25, 1253-1256.

Prehaud, C., Coulon, P., Lafay, F., Thiers, C. & Flamand, A. (1988) Antigenic site II of the rabies virus glycoprotein: structure and role in viral virulence. Journal of Virology, 62, 1.

ProMED-mail (2007) Rabies, Human, Vampire Bat - Peru (03). ProMED-mail 2007, 6 March: 20070306.0786.

ProMed-mail (2010) Rabies, Vampire bat, Human - Peru (04): (Amazonas).

107

Pulliam, J.R.C. (2008) Viral host jumps: Moving toward a predictive framework. EcoHealth, 5, 80-91.

Pulliam, J.R.C., Epstein, J.H., Dushoff, J., Rahman, S.A., Bunning, M., Jamaluddin, A.A., Hyatt, A.D., Field, H.E., Dobson, A.P. & Daszak, P. (2011) Agricultural intensification, priming for persistence and the emergence of Nipah virus: a lethal bat-borne zoonosis. Journal of the Royal Society Interface.

Rambaut, A. & Drummond, A.J. (2007) Tracer v1.4, Available from http://beast.bio.ed.ac.uk/Tracer.

Riley, S., Fraser, C., Donnelly, C.A., Ghani, A.C., Abu-Raddad, L.J., Hedley, A.J., Leung, G.M., Ho, L.-M., Lam, T.-H., Thach, T.Q., Chau, P., Chan, K.-P., Lo, S.-V., Leung, P.-Y., Tsang, T., Ho, W., Lee, K.-H., Lau, E.M.C., Ferguson, N.M. & Anderson, R.M. (2003) Transmission Dynamics of the Etiological Agent of SARS in Hong Kong: Impact of Public Health Interventions. Science, 300, 1961-1966.

Rupprecht, C.E., Hanlon, C.A. & Hemachudha, T. (2002) Rabies re-examined. Lancet Infectious Diseases, 2, 327-343.

Sadler, W.W. & Enright, J.B. (1959) Effect of Metabolic Level of the Host Upon the Pathogenesis of Rabies in the Bat. Journal of Infectious Diseases, 106, 267-273.

Salemi, M., Lewis, M., Egan, J.F., Hall, W.W., Desmyter, J. & Vandamme, A.M. (1999) Different population dynamics of human T cell lymphotropic virus type II in intravenous drug users compared with endemically infected tribes. Proceedings of the National Academy of Sciences of the United States of America, 96, 13253.

Sato, G., Kobayashi, Y., Motizuki, N., Hirano, S., Itou, T., Cunha, E.M.S., Ito, F. & Sakai, T. (2009) A unique substitution at position 333 on the glycoprotein of rabies virus street strains isolated from non-hematophagous bats in Brazil. Virus Genes, 38, 74-79.

Schneider, M.C., Aron, J., Santos-Burgoa, Uieda, W. & Ruiz-Velazco (2001) Common vampire bat attacks on humans in a village of the Amazon region of Brazil. Cad, Saude Publica, Rio de Janeiro, 17, 1531-1536.

Schneider, M.C., Romijn, P.C., Uieda, W., Tamayo, H., da Silva, D.F., Belotto, A., da Silva, J.B. & Leanes, L.F. (2009) Rabies transmitted by vampire bats to humans: An emerging zoonotic disease in Latin America? Rev. Panam. Salud Publica, 25, 260-269.

Schneider, M.C., SantosBurgoa, C., Aron, J., Munoz, B., RuizVelazco, S. & Uieda, W. (1996) Potential force of infection of human rabies transmitted by vampire bats in the Amazonian region of Brazil. American Journal of Tropical Medicine and Hygiene, 55, 680-684.

108

Seif, I., Coulon, P., Rollin, P. & Flamand, A. (1985) Rabies virulence: effect on pathogenicity and sequence characterization of rabies virus mutations affecting antigenic site III of the glycoprotein. Journal of Virology, 53, 926.

Shackelton, L.A., Parrish, C.R., Truyen, U. & Holmes, E.C. (2005) High rate of viral evolution associated with the emergence of carnivore parvovirus. Proceedings of the National Academy of Sciences of the United States of America, 102, 379-384.

Smith, G., Naipospos, T., Nguyen, T., De Jong, M., Vijaykrishna, D., Usman, T., Hassan, S., Nguyen, T., Dao, T. & Bui, N. (2006) Evolution and adaptation of H5N1 influenza virus in avian and human hosts in Indonesia and Vietnam. Virology, 350, 258-268.

Smith, J., Yager, P. & Baer, G. (1996) A rapid fluorescent focus inhibition test (RFFIT) for determining rabies virus-neutralizing antibody. Laboratory techniques in rabies, 4, 181- 192.

Smith, J.S. (2002) Molecular Epidemiology. Rabies (eds A.C. Jackson & W.H. Wunner), pp. 79- 111. Academic Press, San Diego.

Smith, J.S., Orciari, L.A. & Yager, P.A. (1995) Molecular epidemiology of rabies in the United States. Seminars in Virology, 6, 387-400.

Smith, S.A. & Donoghue, M.J. (2008) Rates of molecular evolution are linked to life history in flowering plants. Science, 322, 86-89.

Song, H.D., Tu, C.C., Zhang, G.W., Wang, S.Y., Zheng, K., Lei, L.C., Chen, Q.X., Gao, Y.W., Zhou, H.Q., Xiang, H., Zheng, H.J., Chern, S.W.W., Cheng, F., Pan, C.M., Xuan, H., Chen, S.J., Luo, H.M., Zhou, D.H., Liu, Y.F., He, J.F., Qin, P.Z., Li, L.H., Ren, Y.Q., Liang, W.J., Yu, Y.D., Anderson, L., Wang, M., Xu, R.H., Wu, X.W., Zheng, H.Y., Chen, J.D., Liang, G.D., Gao, Y., Liao, M., Fang, L., Jiang, L.Y., Li, H., Chen, F., Di, B., He, L.J., Lin, J.Y., Tong, S.X., Kong, X.G., Du, L., Hao, P., Tang, H., Bernini, A., Yu, X.J., Spiga, O., Guo, Z.M., Pan, H.Y., He, W.Z., Manuguerra, J.C., Fontanet, A., Danchin, A., Niccolai, N., Li, Y.X., Wu, C.I. & Zhao, G.P. (2005) Cross-host evolution of severe acute respiratory syndrome coronavirus in palm civet and human. Proc. Nat. Acad. Sci. U.S.A., 102, 2430-2435.

Steece, R.R. & Altenbach, J.J.S. (1989) Prevalence of rabies specific antibodies in the Mexican free-tailed bat (Tadarida brasiliensis mexicana) at Lava Cave, New Mexico. Journal of Wildlife Diseases, 25, 490-496.

Streicker, D.G., Lemey, P., Velasco-Villa, A. & Rupprecht, C.E. (submitted) Rates of viral evolution are linked to host life history in bat rabies.

Streicker, D.G., Turmelle, A.S., Vonhof, M.J., Kuzmin, I.V., McCracken, G.F. & Rupprecht, C.E. (2010) Host phylogeny constrains cross-species emergence and establishment of rabies virus in bats. Science, 329, 676-679.

109

Suchard, M.A., Kitchen, C.M.R., Sinsheimer, J.S. & Weiss, R.E. (2003) Hierarchical phylogenetic models for analyzing multipartite sequence data. Systematic Biology, 52, 649-664.

Suchard, M.A., Weiss, R.E. & Sinsheimer, J.S. (2005) Models for estimating Bayes factors with applications to phylogeny and tests of monophyly. Biometrics, 61, 665-673.

Swanepoel, R., Smit, S.B., Rollin, P.E., Formenty, P., Leman, P.A., Kemp, A., Burt, F.J., Grobbelaar, A.A., Croft, J., Bausch, D.G., Zeller, H., Leirs, H., Braack, L.E.O., Libande, M.L., Zaki, S., Nichol, S.T., Kslazek, T.G. & Paweska, J.T. (2007) Studies of reservoir hosts for Marburg virus. Emerging Infectious Diseases, 13, 1847-1851.

R Development Core Team (2011) R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna.

Trajano, E. (1996) Movements of cave bats in southeastern Brazil, with emphasis on the population ecology of the common vampire bat, Desmodus rotundus (Chiroptera). Biotropica, 28, 121-129.

Tuffereau, C., Leblois, H., Benejean, J., Coulon, P., Lafay, F. & Flamand, A. (1989) Arginine or lysine in position 333 of ERA and CVS glycoprotein is necessary for rabies virulence in adult mice. Virology, 172, 206-212.

Turmelle, A.S., Allen, L.C., Jackson, F.R., Kunz, T.H., Rupprecht, C.E. & McCracken, G.F. (2010a) Ecology of Rabies Virus Exposure in Colonies of Brazilian Free-Tailed Bats (Tadarida brasiliensis) at Natural and Man-Made Roosts in Texas. Vector-Borne and Zoonotic Diseases, 10, 165-175.

Turmelle, A.S., Jackson, F.R., Green, D., McCracken, G.F. & Rupprecht, C.E. (2010b) Host immunity to repeated rabies virus infection in big brown bats. Journal of General Virology, 91, 2360-2366.

Turner, D.C. (1975) The vampire bat: a field study in behavior and ecology. The Johns Hopkins University Press, Baltimore, Maryland.

Vasilakis, N., Shell, E.J., Fokam, E.B., Mason, P.W., Hanley, K.A., Estes, D.M. & Weaver, S.C. (2007) Potential of ancestral sylvatic dengue-2 viruses to re-emerge. Virology, 358, 402- 412.

Velasco-Villa, A., Orciari, L.A., Juarez-Islas, V., Gomez-Sierra, M., Padilla-Medina, I., Flisser, A., Souza, V., Castillo, A., Franka, R., Escalante-Mane, M., Sauri-Gonzalez, I. & Rupprecht, C.E. (2006) Molecular diversity of rabies viruses associated with bats in Mexico and other countries of the Americas. Journal of Clinical Microbiology, 44, 1697- 1710.

110

Viboud, C., Alonso, W.J. & Simonsen, L. (2006) Influenza in tropical regions. Plos Medicine, 3, 468-471.

Volk, S.M., Chen, R., Tsetsarkin, K.A., Adams, A.P., Garcia, T.I., Sall, A.A., Nasar, F., Schuh, A.J., Holmes, E.C. & Higgs, S. (2010) Genome-scale phylogenetic analyses of chikungunya virus reveal independent emergences of recent epidemics and various evolutionary rates. Journal of Virology, 84, 6497.

Wacharapluesadee, S., Boongird, K., Wanghongsa, S., Ratanasetyuth, N., Supavonwong, P., Saengsen, D., Gongal, G. & Hemachudha, T. (2010) A longitudinal study of the prevalence of Nipah virus in Pteropus lylei bats in Thailand: evidence for seasonal preference in disease transmission. Vector-Borne and Zoonotic Diseases, 10, 183-190.

WHO (2005) WHO expert consultation on rabies. Tech. Rep. Ser., 931, 1-88.

Wilkinson, G.S. (1984) Reciprocal food sharing in the vampire bat. Nature, 308, 181-184.

Wilkinson, G.S. (1986) Social Grooming in the Common Vampire Bat, Desmodus-Rotundus. Animal Behaviour, 34, 1880-1889.

Wilkinson, G.S. (1988) Social Organization and Behavior. Natural history of vampire bats (eds A.M. Greenhall & U. Schmidt), pp. 85-97. CRC Press, Boca Raton, FL.

Woelk, C.H. & Holmes, E.C. (2002) Reduced positive selection in vector-borne RNA viruses. Molecular Biology and Evolution, 19, 2333.

Wohlgenant, T.J. (1994) Roost Interactions between the Common Vampire Bat (Desmodus- Rotundus) and 2 Frugivorous Bats (Phyllostomus-Discolor and Sturnira-Lilium) in Guanacaste, Costa-Rica. Biotropica, 26, 344-348.

Wolfe, N.D., Daszak, P., Kilpatrick, A.M. & Burke, D.S. (2005) Bushmeat hunting deforestation, and prediction of zoonoses emergence. Emerging Infectious Diseases, 11, 1822-1827.

Wolfe, N.D., Dunavan, C.P. & Diamond, J. (2007) Origins of major human infectious diseases. Nature, 447, 279-283.

Woodroffe, R., Donnelly, C.A., Jenkins, H.E., Johnston, W.T., Cox, D.R., Bourne, F.J., Cheeseman, C.L., Delahay, R.J., Clifton-Hadley, R.S., Gettinby, G., Gilks, P., Hewinson, R.G., McInerney, J.P. & Morrison, W.I. (2006) Culling and cattle controls influence tuberculosis risk for badgers. Proceedings of the National Academy of Sciences of the United States of America, 103, 14713-14717.

Woolhouse, M. & Gaunt, E. (2007) Ecological origins of novel human pathogens. Critical Reviews in Microbiology, 33, 231-242.

111

Woolhouse, M.E.J. (2002) Population biology of emerging and re-emerging pathogens. Trends in Microbiology, 10, s3-s7.

Woolhouse, M.E.J. & Gowtage-Sequeria, S. (2005) Host range and emerging and reemerging infectious diseases. Emerging Infectious Diseases, 11, 1842-1847.

Woolhouse, M.E.J., Haydon, D.T. & Antia, R. (2005) Emerging pathogens: the epidemiology and evolution of species jumps. Trends in Ecology & Evolution, 20, 238-244.

Wright, A.N. & Gompper, M.E. (2005) Altered parasite assemblages in raccoons in response to manipulated resource availability. Oecologia, 144, 148-156.

Wunner, W. & Dietzschold, B. (1987) Rabies virus infection: genetic mutations and the impact on viral pathogenicity and immunity. Contributions to microbiology and immunology, 8, 103.

Wunner, W.H. (2007) Rabies virus, 2 edn. Elsevier, Oxford.

Yan, X., Mohankumar, P.S., Dietzschold, B., Schnell, M.J. & Fu, Z.F. (2002) The rabies virus glycoprotein determines the distribution of different rabies virus strains in the brain. Journal of Neurovirology, 8, 345-352.

Yang, Z., Kumar, S. & Nei, M. (1995) A new method of inference of ancestral nucleotide and amino acid sequences. Genetics, 141, 1641.

Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A. & Smith, G.M. (2009) Mixed effects models and extensions in ecology with R. Springer Verlag.

Zwickl, D.J. (2006) Genetic algorithim approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. PhD, The University of Texas at Austin.

112

APPENDIX A

SUPPORTING INFORMATION FOR CHAPTER 2

MATERIALS AND METHODS

Bats and rabies viruses included in the study. We acquired 372 rabid bats of 23 species from

14 public health laboratories across the USA (Figure 2.1 B, Table 2.1). Bats were originally collected following human or domesticated animal exposures. Upon receipt at the Rabies

Laboratory of the USA Centers for Disease Control and Prevention, brains were confirmed positive for rabies virus antigen by the Direct Fluorescent Antibody test or by reverse transcriptase-polymerase chain reaction (RT-PCR) when infected tissues were limited or degraded.

Phylogenetic species identification of bats. Incorrect species identification of rabid bats could lead to inaccurate estimation of cross-species transmission (CST). To confirm morphological species identities of bats, we undertook phylogenetic species identification of a large subset of all bats (n = 243) including 34/43 individuals infected by CST. Genomic DNA was isolated from bat brain, liver or wing tissue using DNeasy Tissue Kits (Qiagen) following the manufacturer’s protocol. A 636 base pair (bp) fragment of the mitochondrial cytochrome oxidase subunit I

(COI) gene was amplified by polymerase chain reaction (PCR) using primers and protocols from ref. (1) for bat species other than Eptesicus fuscus, with modified primers for E. fuscus. PCR amplicons were purified with ExoSAP-IT ® (USB Corporation) and sequenced on an ABI 3100 capillary sequencer at the University of Tennessee Molecular Biology Resource Facility

113

(Knoxville, TN). Forward and reverse sequences were aligned and edited by eye using

Sequencher 4.7 (Gene Codes Corporation, Ann Arbor, MI) and were compared with reference

COI sequences from 32 USA bat species using a neighbor-joining algorithm in PAUP* v.4.0b10 with distances calculated according to the GTR+I+Γ model of evolution as selected by Modeltest v.3.7 (2, 3). The reference COI database included vouchered museum specimens for all species except Idionycteris phyllotis, Myotis grisescens, M. leibii and M. sodalis (Table A3). Reference sequences from museum specimens have been deposited into GenBank under accession numbers

GU723168 - GU723257.

Sequence amplification and phylogenetic analysis of rabies viruses. Total RNA was extracted directly from naturally infected bat brains using TRIzol™ (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. An 852 bp fragment comprising the last 759 bp of the nucleoprotein (N) gene, a non-coding region following the 3’ end of N and a small fragment of the phosphoprotein gene was amplified by RT-PCR using the oligonucleotide primers described in Table A9. The first 165 nucleotides and the regions immediately following the 3’ end of N were deleted prior to analysis due to incomplete sequencing of some samples, resulting in a 594 bp fragment. Complete N gene sequences (1353 bp) were generated for 148 representative isolates. Thermocycling was performed following ref. (4). PCR amplicons were purified using

ExoSAP-IT® (USB Corporation) or from bands excised from low melting agarose gels using

Wizard PCR Clean-up kits (Promega). Sequencing was carried out using an Applied Biosystems

Prism 377 automated DNA sequencer. Chromatograms were edited by eye in Bioedit and sequences were aligned using MAFFT (5, 6).

114

The Maximum likelihood (ML) and Bayesian phylogenetic analyses were performed using Garli v.0.96b and MrBayes v.3.1.2, respectively, using the General Time Reversible + I +

Γ model of nucleotide substitution selected by AIC in Modeltest v.3.7 (2, 7, 8). The ML tree was estimated by 10 independent searches with random starting trees, followed by 2 sets of 5 additional searches using the best tree from the previous set of searches as the starting tree.

Support for each node was estimated from 2000 ML bootstrap searches using Grid computing available through the Lattice project (9, 10). For the Bayesian analysis, we used a codon model with substitution rates for positions 1 and 2 linked and independent of 3rd position sites. Four simultaneous Metropolis Coupled Markov Chain Monte Carlo (MC3) estimations were run for

4.5 million generations, with trees sampled every 100 generations and the first 5000 trees burned in prior to generation of a consensus tree. The burn-in period was determined by accessing convergence of the likelihood and model parameters in preliminary runs. Run lengths were determined by examination of the standard deviation of split frequencies and convergence of parameters across chains was confirmed by potential scale reduction factors approaching 1 (11).

Bayesian and ML phylogenetic trees were rooted to a rabies virus sequence from North

American raccoons (Procyon lotor) from GenBank (Accession no: AF351826).

Rabies virus lineages were operationally defined as monophyletic groups that (i) were supported by Bayesian posterior probabilities > 0.98 and ML bootstrap values > 0.70, (ii) contained at least 3 unique sequences and (iii) were statistically compartmentalized to a particular bat taxa. Compartmentalization to host species was tested using the BaTS package

(12), which accesses phylogeny-trait associations while accounting for phylogenetic uncertainty by comparing observed associations in the posterior distribution of phylogenetic trees from a

Bayesian search to a large number of trees with randomly assigned traits. We compared 1000

115

trees from the posterior distribution of the Bayesian phylogenetic analysis to a null distribution of 100 trait-randomized trees, with bat taxa assigned as the character trait of interest.

Genetic divergence of rabies viruses from cross-species infections. Substantial genetic divergence of rabies viruses sampled from cross-species infections relative to within-species infections would suggest limited, ongoing transmission within populations of the recipient species. We assessed whether the minimal genetic divergence of rabies viruses from each individual bat infected by CST was greater than those infected by within-species transmission in the same lineage using an outlier criterion. Outlying sequences were identified as those with a minimum pairwise genetic distance to the donor lineage greater than 1.5 times the inter-quartile range above the third quartile (Table A4). Outlier cutoffs were calculated using only sequences from the donor host species (within-species transmission). Genetic distances were generated in

PAUP* v.4.0b10 using models of nucleotide substitution specific to the rabies virus lineage selected by AIC in Modeltest v.3.7 (2, 3).

Maximum likelihood coalescent estimation of cross-species transmission. We quantified the transmission of rabies virus between bat species using the computer software Migrate v.3.0.3, which uses Metropolis Coupled Markov Chain Monte Carlo (MC3) coalescent simulation to estimate θ (proportional to effective population size) and migration rates (β) between an arbitrary number of populations (13). We first estimated θj for the viruses associated with each bat species j under an n-island model with unequal effective population sizes. Multiple viral lineages associated with single bat species were analyzed as independent loci and averaged. For each pair of species that was infected by a common viral lineage, we partitioned viral sequences by bat

116

species and constructed transmission models describing 4 hypotheses on the directionality of

CST: asymmetrical, bi-directional transmission (H0: βij > 0, βji > 0, βij ≠ βji); symmetrical, bidirectional transmission (H1: βij > 0, βji > 0, βij = βji); and each case of unidirectional transmission, (H2: βij > 0, βji = 0; H3: βji > 0, βij = 0). The best transmission model for each species pair was selected by AIC. CST was thus defined as the “migration” of rabies virus from one bat species to another. One exception was allowed for some Myotis species in the western

USA because their phylogenetic relationships remain unresolved in 2 distinct species complexes

(see Table A3). We took a conservative approach to estimate transmission between but not within different species complexes.

3 MC simulations in Migrate used 4 chains with the following static heating scheme: 1,

1.3, 3, and 100,000, where the temperature 1 indicates the cold chain with increasingly hotter chains exploring parameter space more freely. Simulations were run for 30 million generations with sampling broken into 10 short and 3 long chains. Genealogies were sampled every 1000 generations in the long chain for a total of 30,000 sampled genealogies per run. At least 3 fully replicated simulations were run for each species pair, with parameters averaged across runs using the reverse logistic regression method of Geyer (14) for a total of at least 90 million generations per species pair. Simulations were carried out using the parallel version of Migrate on the Linux cluster at the University of Georgia’s Research Computing Center. Input and parameter files are available from the corresponding author upon request.

Quantifying the per capita cross-species transmission rate (Rij). Rij is the expected number of infections in species i resulting from a single infected individual of species j. This is the between-species analog of the effective reproductive number of a pathogen (Re), the average

117

number of secondary cases resulting from a single infected individual in a heterogeneous population comprised, susceptible, infected and immune individuals. While the existence of an immune class in bat rabies is uncertain, many studies have demonstrated substantial numbers of wild bats with naturally acquired rabies virus neutralizing antibodies, suggestive of immune

%1 protection following abortive infections (15, 16). We estimated Rij as "ij# j$ , where τ, generation time is the sum of the incubation and infectious periods, and βij and θj are derived ! from genetic data. Experimental rabies virus infections in insectivorous bats indicate a mean τ of

29 days (17).

Quantification and analysis of asymmetries in the transmission web. In ecological food webs, connectance describes the proportion of realized trophic connections between sympatric species (18). We hybridized food web connectance with measures of the strength and direction of CST into a measure termed “transmission asymmetry,” which describes whether bat species were disproportionately donors or recipients of CST. Specifically, transmission asymmetry is the difference between the transmission connectance of each species as a donor:

1 % ( CD = ' L j $Rij * , (1) n "1& i#n ) or recipient:

! 1 % ( (2) CR = ' Li $Rij * n "1& j #n ) of CST. Here, Lj is the number of other species infected by species j, Li is the number of other species that infected species! i and n is the number of sympatric hosts in the transmission web.

118

Positive values indicate an excess of transmission to other species; negative values indicate an excess of infections by other species.

Auxiliary data and statistical analyses of cross-species transmission. Statistical analyses of

CST rates sought to distinguish whether ecological similarity, evolutionary relatedness or probabilistic factors that increase with geographic range overlap best predict per capita rates of

CST (see Tables 7, 8). Foraging niche overlap was approximated using three morphological measurements: wing aspect ratio, wing loading and body length, which are associated with foraging behavior in bats (19). The absolute value of differences between species served as a proxy for similarity. Roost site overlap was classified by two categorical descriptions of the likelihood that bats roost in the same structures: the first (potential overlap, R1), based on typical summer roost type, and the second (documented overlap, R2) based on historical records of roost sharing (Table A6). The phylogenetic distance between bat species was estimated from COI sequences using substitution models selected by AIC in Modeltest as above (2, 3). We estimated the geographic overlap between species as the proportion of the recipient species range within the continental USA where it co-occurs with the donor species using range maps from

Natureserve (20) under the Albers equal-area projection.

Two sets of generalized linear models (GLM) were constructed using (i) only species pairs implicated in CST to describe the intensity of Rij and (ii) all pairwise combinations of species in the dataset to describe the probability and intensity of CST. To account for non- independence of data from repeated comparisons in dataset ii, we used GLMs with normal error distributions to select candidate models and confirmed significance with partial Mantel tests.

This approach was used because formal model selection procedures are unavailable for partial

119

Mantel tests. In both datasets, minimal adequate models were selected by forward and reverse stepwise AIC model selection from full models containing all ecological, genetic and geographic variables as well as a term describing the sample size of recipient species in dataset i (see

Supporting Text). Mantel tests assessed statistical significance with 1000 matrix permutations.

When appropriate, terms were transformed to meet model assumptions. Statistical analyses were performed in R v. 2.7.2 (21).

Bayesian estimation of the origins of established host shifts. Historical and phylogenetic evidence indicate that rabies spread relatively recently through the North American bat community through a process of sequential host shifts, rather than ancient co-divergence (22-24).

Because rabies virus lineages are compartmentalized to bat species (Table A2), host species may be modeled as a discrete trait of each viral lineage using ancestral state estimation. We identified pairs of viral lineages that were strongly linked by host shifts using the models developed by ref.

(25) for the software BEAST v. 1.5.4. These models estimate changes between discrete states, in our case, host species or sub-species, while controlling for uncertainty in the phylogenetic relationships of viral lineages and variation in substitution rates among branches and incorporating temporal information from the sampling date of viruses (26).

The dataset was constructed with a maximum of 15 complete N gene sequences (1353 bp) per lineage to minimize skew in the sample size of lineages and reduce run time. We relaxed the criterion for lineage definition relative to the CST analysis to include (i) previously described rabies viruses that were represented by fewer than 3 sequences in our dataset (27, 28) and (ii) rabies virus lineages which were compartmentalized to genetically distinct subspecies of

Eptesicus fuscus and Lasiurus intermedius in different parts of their geographic range. In

120

addition, we included sequences from a viral lineage associated with the common vampire bat

(Desmodus rotundus, GenBank accession number for COI: EF080321), due to evidence that the ancestor of this virus is evolutionarily linked to rabies viruses from North American insectivorous bats (22, 29). The rabies virus lineages found only in Central and South America were excluded because hosts of many sequences are not defined to the species level and their geographic isolation from North American species makes host shifts relatively unlikely (29, 30).

Previously published N gene sequences from GenBank (n = 87) were included to increase the temporal breadth of sampling, for a total of 236 sequences, distributed over 23 viral lineages, that spanned the time period from 1972-2006. GenBank accession numbers for previously published sequences included are: AB083805, AB083817, AB297633-AB297636, AB297645, AB297646,

AF045166, AF351827, AF351828-AF351848, AF351852-AF351862, AF394868-AF394887,

AY039224-AY039229, AY170397-AY170401, AY170404, AY170405, AY170412-AY170417,

AY854587, AY854588, AY854592, AY877433, AY877435 and EF363727.

Analyses in BEAST used the TVM+Γ model of nucleotide substitution selected by AIC in Modeltest with separate substitution parameters assigned to the first two versus the third codon positions based on preliminary analyses. We used the Bayesian skyline model of demographic growth to minimize assumptions on the demographic history of rabies virus in bats throughout the time period spanned by the genealogy. An uncorrelated lognormal relaxed molecular clock model was used to accommodate rate variation among lineages. Four independent MCMC analyses were run for 25 million generations each with samples from the posterior drawn every 1000 generations following burn-in periods of 4 million generations based on convergence of likelihood and model parameters as indicated in Tracer (31). The Bayesian stochastic search variable selection procedure described by ref. (25) was implemented to allow

121

discrete state change rates (i.e., the rate describing host shifts) to be zero, enabling the use of a

Bayes factor (BF) test to identify non-zero rates (32). The results from the four runs were combined for final analysis and BF support for host shifts was calculated using the software provided by ref. (25). Transition rates supported by a BF > 3 were considered significant support for a host shift between bat species. Beast input files are available from the corresponding author upon request.

SUPPORTING TEXT

Asymmetry in the transmission web. The transmission web (Figure 2.2) revealed that CST is often unidirectional and always asymmetrical, leading to large variation among species as donors or recipients of CST (Figure A1). For example, L. cinereus, L. borealis and the M. californicus species complex were hubs of CST, but were less often recipients of infection, while M. lucifugus, M. yumanensis, L. xanthinus and L. blossevillii were disproportionately recipients of

CST. We analyzed the transmission web to determine which host or viral factors influence transmission asymmetry, focusing on: the nucleotide diversity (π) of each rabies virus lineage, the amount of geographic overlap with all other bat species, the average phylogenetic similarity of each bat to other species within its geographic range or the sampling effort (defined as the number of bats analyzed) for each species. The average phylogenetic similarity to other bat

"1 species was estimated as 1" n #dij , where dij is the genetic distance (from COI data) from i species j to a sympatric species i, and n is the number of sympatric species. Geographic overlap

! was calculated as "gij , where gij is the proportion of the range of species j where it overlaps j with species i. We used DnaSP (33) to calculate π for each rabies virus lineage and used the

122

arithmetic mean for species that maintained multiple rabies virus lineages. Forward and reverse

AIC model selection was used to select from a starting GLM with all 4 factors assuming normal error distributions. Sampling effort and π were log10 transformed prior to analysis. Although we found no reliable predictor of the magnitude and direction of transmission asymmetry among the

15 species involved in CST (GLM: Full model: n = 15, F3,11 = 1.44, P = 0.28), the magnitude of transmission asymmetry (the absolute value of CD-CR) increased with the average phylogenetic similarity of bats to others in their geographic distribution and with the summed proportion of

2 geographic overlap with other species (GLM: n = 15, F2,12 = 12.24, r = 0.67, P = 0.001). In essence, bats that overlap with many genetically similar neighbors were more likely to be involved in CST than bats that are evolutionarily distinct in their communities; however, whether a particular bat species acts as a donor or recipient remains unclear. An important avenue for future research will be to determine how host characteristics, such as variation in susceptibility, behavioral responses to infection and intra-specific contact rates, or viral factors, such as infectiousness or stage of adaptation to the donor host (e.g., recent versus older host-virus association), influence transmission rates.

Unequal sampling, unequal prevalence and detection of cross-species transmission. Despite the large number of viruses included in this study, only a fraction of the CST events that occur in this system are observable. We attempted to control for unequal sampling effort across species in our estimate of per-capita CST rates by using population genetics software that can accommodate this potential bias. Still, it is possible that the detection of rare CST depends on the number of individuals of the recipient species that were examined, causing higher detection of

CST in very well sampled species than in less-sampled species. Examination of the data revealed

123

that the number of individuals infected by CST was not strongly associated with the number of individuals examined (GLM: F1,16 = 3.88, P = 0.07) and visual inspection of the transmission web (Figure 2.2) shows very few cases of transmission from poorly-sampled to well-sampled species. Similarly, as discussed above, sample size was not retained as a term in minimal adequate models describing the magnitude of transmission asymmetry or pairwise rates of CST

(Table A8).

A second potential driver of CST that we were not able to measure directly was the prevalence of rabies virus infection in different bat species. Available evidence from extensive surveys of free-living bats indicates that prevalence is low (<1-3%) and does not differ substantially among North American insectivorous bat species (15, 34). Nevertheless, we attempted to control for the possibility of unequal prevalence in our calculation of Rij by scaling this statistic by θ for each viral lineage. It should be noted that while incorporating θ into Rij would correct our estimates of per-capita CST, AIC selection of transmission models in Migrate considered only β, raising the possibility that unequal rabies prevalence in donor and recipient species might bias selection of asymmetrical transmission models. Examination of the data revealed no relationship between prevalence (approximated with θ) and β (Pearson product moment correlation: r = 0.047, P = 0.802). In light of the biological arguments for low variation in the prevalence of bat rabies among species and the lack of a relationship between θ and β in our dataset, the potential for unequal prevalence of rabies to bias our estimates of CST is minimal.

124

FIGURES AND TABLES

A B Lc Lc Mc Mc Lb Lb Ln Ln Tb Tb Ps Ps Nh Nh Ls Ls Ap Ap Li Li Ef Ef Lbl Lbl Lx Lx My My Ml Ml

-0.5 -0.4 -0.2 0.0 0.2 0.4 0.6 0.7 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Transmission asymmetry (C - C ) D R Per capita transmission balance

Figure A1. Heterogeneities among bat species as donors or recipients of cross-species transmission. (A) Distribution of transmission asymmetry among bat species. Species with positive values infect more often than they are infected, and species with negative values are infected more often than they infect other species. (B) Raw scores of per capita transmission (Rij) to/from any species. Positive bars (black) indicate the average number of individuals of all other species infected by a single infected individual. Negative bars (white) indicate the average number of individuals infected by cross-species transmission by a single infected individual of each sympatric species. Abbreviations are bat species names as in Table A3. Dashed lines indicate the transition from positive to negative transmission asymmetry.

125

Table A1. Species, date, location and GenBank accession numbers for sequences generated in this study.

ID Species Year State N accession no. COI accession no. AZ4030 * Antrozous pallidus 2005 Arizona GU644641 GU723167 AZ1968 * Eptesicus fuscus 2004 Arizona GU644642 - AZ7590 * Eptesicus fuscus 2005 Arizona GU644643 - CA237 * Eptesicus fuscus 2002 California GU644644 GU723065 CA29 * Eptesicus fuscus 2002 California GU644645 - CA9242 * Eptesicus fuscus 2002 California GU644646 GU723066 CAO120 * Eptesicus fuscus 2002 California GU644647 GU723067 CA0253 * Eptesicus fuscus 2003 California GU644648 - CA148 * Eptesicus fuscus 2004 California GU644649 - CA6860 * Eptesicus fuscus 2004 California GU644650 GU723062 CA0100 * Eptesicus fuscus 2005 California GU644651 GU723064 GA31940 * Eptesicus fuscus 2004 Georgia GU644652 - GA36568 * Eptesicus fuscus 2004 Georgia GU644653 GU723031 IA042 * Eptesicus fuscus 2005 Iowa GU644654 GU723020 IA381 * Eptesicus fuscus 2005 Iowa GU644655 - IA543 * Eptesicus fuscus 2005 Iowa GU644656 GU723117 MI1251* Eptesicus fuscus 2003 Michigan GU644657 GU723036 MI1271 * Eptesicus fuscus 2003 Michigan GU644658 GU723034 MI1406 * Eptesicus fuscus 2003 Michigan GU644659 GU723033 MI1586 * Eptesicus fuscus 2003 Michigan GU644660 GU723042 MI1672 * Eptesicus fuscus 2003 Michigan GU644661 GU723043 MI1833 * Eptesicus fuscus 2003 Michigan GU644662 GU723044 MI782 * Eptesicus fuscus 2003 Michigan GU644663 GU723038 MI784 * Eptesicus fuscus 2003 Michigan GU644664 GU723035 MI1209 * Eptesicus fuscus 2005 Michigan GU644665 - MI1328 * Eptesicus fuscus 2005 Michigan GU644666 - MI1865 * Eptesicus fuscus 2005 Michigan GU644667 - MI1905 * Eptesicus fuscus 2005 Michigan GU644668 - MI399 * Eptesicus fuscus 2005 Michigan GU644669 - MI596 * Eptesicus fuscus 2005 Michigan GU644670 - NJ104 * Eptesicus fuscus 2005 New Jersey GU644671 GU723046 NJ1049 * Eptesicus fuscus 2005 New Jersey GU644672 - NJ1212 * Eptesicus fuscus 2005 New Jersey GU644673 - NJ511 * Eptesicus fuscus 2005 New Jersey GU644674 - NJ949 * Eptesicus fuscus 2005 New Jersey GU644675 GU723047 VA1623 * Eptesicus fuscus 2004 Virginia GU644676 GU723057 VA2057 * Eptesicus fuscus 2004 Virginia GU644677 GU723051 WA0355 * Eptesicus fuscus 1999 Washington GU644678 - WA369 * Eptesicus fuscus 1999 Washington GU644679 GU723070 WA052 * Eptesicus fuscus 2000 Washington GU644680 - WA173 * Eptesicus fuscus 2000 Washington GU644681 GU723076 WA267 * Eptesicus fuscus 2000 Washington GU644682 GU723071 WA1043 * Eptesicus fuscus 2003 Washington GU644683 - WA1087 * Eptesicus fuscus 2003 Washington GU644684 GU723075 WA1159 * Eptesicus fuscus 2003 Washington GU644685 GU723079 WA016 * Eptesicus fuscus 2004 Washington GU644686 - WA1455 * Eptesicus fuscus 2004 Washington GU644687 - WA1586 * Eptesicus fuscus 2004 Washington GU644688 GU723069 WA1596 * Eptesicus fuscus 2004 Washington GU644689 -

126

WA1625 * Eptesicus fuscus 2004 Washington GU644690 GU723078 WA50 * Eptesicus fuscus 2004 Washington GU644691 - WA1770 * Eptesicus fuscus 2005 Washington GU644692 - WA1833 * Eptesicus fuscus 2005 Washington GU644693 - WA1858 * Eptesicus fuscus 2005 Washington GU644694 - WA2017 * Eptesicus fuscus 2005 Washington GU644695 GU723077 CA16461 * Lasiurus blossevillii 2002 California GU644696 GU723019 CA0077 * Lasiurus blossevillii 2003 California GU644697 GU723018 FL701 * Lasiurus borealis 2003 Florida GU644698 GU722986 FL854 * Lasiurus borealis 2003 Florida GU644699 - GA60243 * Lasiurus borealis 2005 Georgia GU644700 GU722982 MI1625 * Lasiurus borealis 2005 Michigan GU644701 - NJ2262 * Lasiurus borealis 2005 New Jersey GU644702 GU722988 TN132 * Lasiurus borealis 2004 Tennessee GU644703 GU722994 TN269 * Lasiurus borealis 2004 Tennessee GU644704 GU722992 TN33 * Lasiurus borealis 2004 Tennessee GU644705 GU722993 TX4843 * Lasiurus borealis 2003 Texas GU644706 - TX6070 * Lasiurus borealis 2003 Texas GU644707 - TX2356 * Lasiurus borealis 2004 Texas GU644708 - TX5276 * Lasiurus borealis 2004 Texas GU644709 GU723003 TX5751 * Lasiurus borealis 2004 Texas GU644710 GU722999 TX5976 * Lasiurus borealis 2004 Texas GU644711 GU722997 AZ1838 * Lasiurus cinereus 2005 Arizona GU644712 GU722954 AZ5392 * Lasiurus cinereus 2005 Arizona GU644713 GU722955 AZ7771 * Lasiurus cinereus 2005 Arizona GU644714 - ID7227 * Lasiurus cinereus 2005 Idaho GU644715 - ID7232 * Lasiurus cinereus 2005 Idaho GU644716 - TN183 * Lasiurus cinereus 2004 Tennessee GU644717 GU722969 TN410 * Lasiurus cinereus 2004 Tennessee GU644718 GU722970 WA0524 * Lasiurus cinereus 1998 Washington GU644719 GU722979 WA1617 * Lasiurus cinereus 2004 Washington GU644720 GU722981 WA2085 * Lasiurus cinereus 2005 Washington GU644721 - FL1024 * Lasiurus intermedius 2001 Florida GU644722 GU722927 FL845 * Lasiurus intermedius 2001 Florida GU644723 GU722985 FL905 * Lasiurus intermedius 2001 Florida GU644724 GU722925 FL978 * Lasiurus intermedius 2001 Florida GU644725 GU722928 FL1165 * Lasiurus intermedius 2004 Florida GU644726 GU722929 TX4904 * Lasiurus intermedius 2002 Texas GU644727 GU722935 TX5433 * Lasiurus intermedius 2003 Texas GU644728 GU722932 TX4350 * Lasiurus intermedius 2004 Texas GU644729 GU722933 ID7376 * Lasionycteris noctivagans 2005 Idaho GU644730 GU722947 FL769 * Lasiurus seminolus 2003 Florida GU644731 - GA7034 * Lasiurus seminolus 2003 Georgia GU644732 GU723012 TX6197 * Lasiurus seminolus 1998 Texas GU644733 GU723015 TX5565 * Lasiurus seminolus 2000 Texas GU644734 GU723017 TX5512 * Lasiurus seminolus 2002 Texas GU644735 GU723154 TX5850 * Lasiurus seminolus 2002 Texas GU644736 GU723013 TX6127a * Lasiurus seminolus 2003 Texas GU644737 - AZ2953 * Lasiurus xanthinus 2004 Arizona GU644738 GU722939 CA2070 * Lasiurus xanthinus 2003 California GU644739 GU722938 CA06 * Lasiurus xanthinus 2004 California GU644740 GU722940 FL1078 * Myotis austroriparius 2001 Florida GU644741 GU723112 FL331* Myotis austroriparius 2001 Florida GU644742 GU723113 ID7198 * Myotis californicus 2005 Idaho GU644743 -

127

ID7261 * Myotis californicus 2005 Idaho GU644744 GU723118 WA1502 * Myotis californicus 2004 Washington GU644745 GU723121 ID7233 * Myotis evotis 2005 Idaho GU644746 GU723124 WA2020 * Myotis evotis 2005 Washington GU644747 GU723122 MI1100 * Myotis lucifugus 2005 Michigan GU644748 GU723127 MI1367 * Myotis lucifugus 2005 Michigan GU644749 GU723129 TN39 * Myotis lucifugus 2004 Tennessee GU644750 GU723130 AZ2857 * Myotis yumanensis 2004 Arizona GU644751 GU723139 CA828 * Myotis yumanensis 2004 California GU644752 GU723134 CA957 * Myotis yumanensis 2004 California GU644753 GU723131 FL1384 * Nycticeius humeralis 2001 Florida GU644754 GU723164 AZ1258 * Parastrellus hesperus 2004 Arizona GU644755 GU723158 CA2167 * Parastrellus hesperus 2003 California GU644756 GU723160 IN1657 * Perimyotis subflavus 2004 Indiana GU644757 GU723149 TX5168 * Perimyotis subflavus 2004 Texas GU644758 GU723156 CA178 * Corynorhinus townsendii 2003 California GU644759 - AZ2405 * Tadarida brasiliensis 2005 Arizona GU644760 GU723080 AZ2579 * Tadarida brasiliensis 2005 Arizona GU644761 GU723085 AZ3086 * Tadarida brasiliensis 2005 Arizona GU644762 GU723083 AZ6914 * Tadarida brasiliensis 2005 Arizona GU644763 GU723082 AZ982 * Tadarida brasiliensis 2005 Arizona GU644764 GU723081 CA0408 * Tadarida brasiliensis 2002 California GU644765 - CA0052 * Tadarida brasiliensis 2003 California GU644766 - CA0299 * Tadarida brasiliensis 2003 California GU644767 - CA15822 * Tadarida brasiliensis 2003 California GU644768 GU723094 CA15824 * Tadarida brasiliensis 2003 California GU644769 GU723093 CA1984 * Tadarida brasiliensis 2003 California GU644770 GU723092 CA2132 * Tadarida brasiliensis 2003 California GU644771 GU723088 CA268 * Tadarida brasiliensis 2003 California GU644772 GU723087 CA3054 * Tadarida brasiliensis 2003 California GU644773 - CA46 * Tadarida brasiliensis 2003 California GU644774 GU723090 CA1032 * Tadarida brasiliensis 2004 California GU644775 - CA96 * Tadarida brasiliensis 2004 California GU644776 - FL148 * Tadarida brasiliensis 2004 Florida GU644777 GU723101 GA093 * Tadarida brasiliensis 2004 Georgia GU644778 - GA112 * Tadarida brasiliensis 2004 Georgia GU644779 GU723105 GA135 * Tadarida brasiliensis 2004 Georgia GU644780 GU723104 GA187 * Tadarida brasiliensis 2003 Georgia GU644781 GU723106 MS076 * Tadarida brasiliensis 2004 Mississippi GU644782 - MS079 * Tadarida brasiliensis 2004 Mississippi GU644783 GU723107 TX5775 * Tadarida brasiliensis 2003 Texas GU644784 - TX5922 * Tadarida brasiliensis 2003 Texas GU644785 GU723110 TX6344 * Tadarida brasiliensis 2003 Texas GU644786 - TX7244 * Tadarida brasiliensis 2003 Texas GU644787 - TX5603 * Tadarida brasiliensis 2004 Texas GU644788 GU723109 AZ2609 Antrozous pallidus 2004 Arizona GU644789 GU723166 CA141a Tadarida brasiliensis 2003 California GU644790 - AZ2408 Eptesicus fuscus 2005 Arizona GU644791 GU723061 AZ2472 Eptesicus fuscus 2005 Arizona GU644792 GU723059 CA09 Eptesicus fuscus 2002 California GU644793 GU723063 CA0115 Eptesicus fuscus 2003 California GU644794 - CA6795 Eptesicus fuscus 2003 California GU644795 - CA0515 Eptesicus fuscus 2004 California GU644796 - GA31217 Eptesicus fuscus 2004 Georgia GU644797 -

128

GA31320 Eptesicus fuscus 2004 Georgia GU644798 GU723032 GA31454 Eptesicus fuscus 2004 Georgia GU644799 - GA32492 Eptesicus fuscus 2005 Georgia GU644800 - ID7318 Eptesicus fuscus 2005 Idaho GU644801 - IN1128 Eptesicus fuscus 2004 Indiana GU644802 - IN141 Eptesicus fuscus 2004 Indiana GU644803 GU723027 IN144 Eptesicus fuscus 2004 Indiana GU644804 - IN1515 Eptesicus fuscus 2004 Indiana GU644805 GU723026 IN833 Eptesicus fuscus 2004 Indiana GU644806 - IN845 Eptesicus fuscus 2004 Indiana GU644807 GU723028 IN121 Eptesicus fuscus 2005 Indiana GU644808 GU723024 IN133 Eptesicus fuscus 2005 Indiana GU644809 GU723025 IN1363 Eptesicus fuscus 2005 Indiana GU644810 - IN272 Eptesicus fuscus 2005 Indiana GU644811 - IN313 Eptesicus fuscus 2005 Indiana GU644812 GU723029 IN646 Eptesicus fuscus 2005 Indiana GU644813 - IN902 Eptesicus fuscus 2005 Indiana GU644814 GU723030 MI1957 Eptesicus fuscus 2003 Michigan GU644815 GU723037 MI1966 Eptesicus fuscus 2003 Michigan GU644816 GU723039 MI2114 Eptesicus fuscus 2003 Michigan GU644817 - MI2155 Eptesicus fuscus 2003 Michigan GU644818 GU723040 MI738 Eptesicus fuscus 2003 Michigan GU644819 GU723045 MI639 Eptesicus fuscus 2004 Michigan GU644820 GU723041 NJ1510 Eptesicus fuscus 2005 New Jersey GU644821 - NJ1548 Eptesicus fuscus 2005 New Jersey GU644822 - NJ1609 Eptesicus fuscus 2005 New Jersey GU644823 - NJ1681 Eptesicus fuscus 2005 New Jersey GU644824 - NJ1772 Eptesicus fuscus 2005 New Jersey GU644825 - NJ1793 Eptesicus fuscus 2005 New Jersey GU644826 - NJ1916 Eptesicus fuscus 2005 New Jersey GU644827 - NJ2027 Eptesicus fuscus 2005 New Jersey GU644828 - NJ2091 Eptesicus fuscus 2005 New Jersey GU644829 - NJ2282 Eptesicus fuscus 2005 New Jersey GU644830 - NJ2304 Eptesicus fuscus 2005 New Jersey GU644831 - NJ2305 Eptesicus fuscus 2005 New Jersey GU644832 - NJ2367 Eptesicus fuscus 2005 New Jersey GU644833 - NJ2405 Eptesicus fuscus 2005 New Jersey GU644834 - NJ2452 Eptesicus fuscus 2005 New Jersey GU644835 - NJ2605 Eptesicus fuscus 2005 New Jersey GU644836 - NJ2938 Eptesicus fuscus 2005 New Jersey GU644837 - TN103 Eptesicus fuscus 2005 Tennessee GU644838 - TN126 Eptesicus fuscus 2005 Tennessee GU644839 - TN259 Eptesicus fuscus 2005 Tennessee GU644840 - TN822 Eptesicus fuscus 2005 Tennessee GU644841 - VA1716 Eptesicus fuscus 2002 Virginia GU644842 GU723053 VA964 Eptesicus fuscus 2002 Virginia GU644843 GU723052 VA245 Eptesicus fuscus 2003 Virginia GU644844 GU723058 VA579 Eptesicus fuscus 2003 Virginia GU644845 - VA3448 Eptesicus fuscus 2004 Virginia GU644846 GU723056 VA1548 Eptesicus fuscus 2005 Virginia GU644847 - VA1597 Eptesicus fuscus 2005 Virginia GU644848 GU723055 VA190 Eptesicus fuscus 2005 Virginia GU644849 - VA2275 Eptesicus fuscus 2005 Virginia GU644850 - WA1086 Eptesicus fuscus 2003 Washington GU644851 GU723074

129

WA1459 Eptesicus fuscus 2004 Washington GU644852 - WA2028 Eptesicus fuscus 2005 Washington GU644853 GU723073 WA2102 Eptesicus fuscus 2005 Washington GU644854 - CA33 Lasiurus blossevillii 2004 California GU644855 - GA32657 Lasiurus borealis 2005 Georgia GU644856 GU722987 IN362 Lasiurus borealis 2005 Indiana GU644857 GU722983 IN427 Lasiurus borealis 2005 Indiana GU644858 GU722984 TN272 Lasiurus borealis 2004 Tennessee GU644859 - TN209 Lasiurus borealis 2005 Tennessee GU644860 - TN80 Lasiurus borealis 2005 Tennessee GU644861 - TX5798 Lasiurus borealis 2003 Texas GU644862 GU723001 TX5943 Lasiurus borealis 2003 Texas GU644863 - TX5987 Lasiurus borealis 2003 Texas GU644864 GU722995 TX6151 Lasiurus borealis 2003 Texas GU644865 - TX4881 Lasiurus borealis 2004 Texas GU644866 GU722998 TX5739 Lasiurus borealis 2004 Texas GU644867 GU723000 TX5742 Lasiurus borealis 2004 Texas GU644868 GU723002 TX5975 Lasiurus borealis 2004 Texas GU644869 GU722996 TX6083 Lasiurus borealis 2004 Texas GU644870 - TX5275 Lasiurus borealis NA Texas GU644871 - VA1333 Lasiurus borealis 2000 Virginia GU644872 GU723009 VA1340 Lasiurus borealis 2001 Virginia GU644873 GU723004 VA399 Lasiurus borealis 2001 Virginia GU644874 GU723007 VA679 Lasiurus borealis 2001 Virginia GU644875 GU723008 VA1787 Lasiurus borealis 2002 Virginia GU644876 GU723005 VA1924 Lasiurus borealis 2002 Virginia GU644877 - VA200 Lasiurus borealis 2002 Virginia GU644878 GU723006 VA1198 Lasiurus borealis 2004 Virginia GU644879 - VA1973 Lasiurus borealis 2005 Virginia GU644880 - VA2002 Lasiurus borealis 2005 Virginia GU644881 - VA546 Lasiurus borealis 2005 Virginia GU644882 - AZ2944 Lasiurus cinereus 2005 Arizona GU644883 GU722953 CA32 Lasiurus cinereus 2001 California GU644884 GU722966 CA030 Lasiurus cinereus 2002 California GU644885 GU722965 CA068 Lasiurus cinereus 2002 California GU644886 GU722959 CA217 Lasiurus cinereus 2002 California GU644887 GU722964 CA25 Lasiurus cinereus 2002 California GU644888 GU722961 CA3872 Lasiurus cinereus 2002 California GU644889 GU722962 CA545 Lasiurus cinereus 2002 California GU644890 GU722960 CA2586 Lasiurus cinereus 2003 California GU644891 - CA281 Lasiurus cinereus 2003 California GU644892 GU722958 CA48 Lasiurus cinereus 2003 California GU644893 - CA7979 Lasiurus cinereus 2003 California GU644894 GU722963 CA0579 Lasiurus cinereus 2004 California GU644895 - CA335 Lasiurus cinereus 2004 California GU644896 - CA47 Lasiurus cinereus 2004 California GU644897 - CA4287 Lasiurus cinereus 2005 California GU644898 - IN520 Lasiurus cinereus 2004 Indiana GU644899 GU722957 IN669 Lasiurus cinereus 2004 Indiana GU644900 GU722956 TN310 Lasiurus cinereus 2004 Tennessee GU644901 GU722971 TN59 Lasiurus cinereus 2004 Tennessee GU644902 GU722968 TN898 Lasiurus cinereus 2005 Tennessee GU644903 - TN977 Lasiurus cinereus 2005 Tennessee GU644904 - TX6362 Lasiurus cinereus 2003 Texas GU644905 GU722977

130

TX6464 Lasiurus cinereus 2003 Texas GU644906 GU722974 TX2130 Lasiurus cinereus 2004 Texas GU644907 GU722975 TX4688 Lasiurus cinereus 2004 Texas GU644908 GU722972 TX5370 Lasiurus cinereus 2004 Texas GU644909 GU722976 TX5427 Lasiurus cinereus 2004 Texas GU644910 - TX5695 Lasiurus cinereus 2004 Texas GU644911 GU722973 TX6295 Lasiurus cinereus 2004 Texas GU644912 GU722978 VA484 Lasiurus cinereus 2004 Virginia GU644913 - FL1010 Lasiurus intermedius 2002 Florida GU644914 - FL1191 Lasiurus intermedius 2002 Florida GU644915 - FL527 Lasiurus intermedius 2002 Florida GU644916 - FL879 Lasiurus intermedius 2002 Florida GU644917 - FLLi04 Lasiurus intermedius 2004 Florida GU644918 GU722926 TX5960 Lasiurus intermedius 2002 Texas GU644919 GU722937 TX6706 Lasiurus intermedius 2002 Texas GU644920 GU722931 TX4500 Lasiurus intermedius 2003 Texas GU644921 GU722936 TX3054 Lasiurus intermedius 2003 Texas GU644922 GU722934 ID7275 Lasionycteris noctivagans 2005 Idaho GU644923 GU722948 ID7282 Lasionycteris noctivagans 2005 Idaho GU644924 GU722949 ID7285 Lasionycteris noctivagans 2005 Idaho GU644925 GU722951 ID7360 Lasionycteris noctivagans 2005 Idaho GU644926 - WA0323 Lasionycteris noctivagans 1997 Washington GU644927 - WA0410 Lasionycteris noctivagans 1997 Washington GU644928 GU722946 WA341 Lasionycteris noctivagans 1997 Washington GU644929 - WA580 Lasionycteris noctivagans 1998 Washington GU644930 GU722943 WA0353 Lasionycteris noctivagans 1999 Washington GU644931 GU722952 WA161 Lasionycteris noctivagans 2000 Washington GU644932 GU722944 WA1066 Lasionycteris noctivagans 2003 Washington GU644933 GU722950 WA1185 Lasionycteris noctivagans 2003 Washington GU644934 GU722942 WA1347 Lasionycteris noctivagans 2004 Washington GU644935 - WA1400 Lasionycteris noctivagans 2004 Washington GU644936 GU722945 WA1955 Lasionycteris noctivagans 2005 Washington GU644937 - WA1991 Lasionycteris noctivagans 2005 Washington GU644938 - FL732 Lasiurus seminolus 2001 Florida GU644939 - FL673 Lasiurus seminolus 2003 Florida GU644940 - FL792 Lasiurus seminolus 2004 Florida GU644941 GU723010 FL793 Lasiurus seminolus 2004 Florida GU644942 GU723011 FL942 Lasiurus seminolus 2004 Florida GU644943 - TX5419 Lasiurus seminolus 2003 Texas GU644944 - TX6265 Lasiurus seminolus 2003 Texas GU644945 GU723014 AZ6954 Lasiurus xanthinus 2005 Arizona GU644946 GU722941 CA120 Lasiurus xanthinus 2005 California GU644947 - CA6054 Myotis californicus 2000 California GU644948 GU723119 CA3974 Myotis californicus 2001 California GU644949 - CA2306 Myotis californicus 2004 California GU644950 GU723116 CA4262 Myotis californicus 2004 California GU644951 GU723114 CA6846 Myotis californicus 2004 California GU644952 GU723115 CA0044 Myotis californicus 2005 California GU644953 - WA1316 Myotis californicus 2004 Washington GU644954 GU723120 CA7629 Myotis evotis 2003 California GU644955 GU723123 WA1890 Myotis evotis 2005 Washington GU644956 GU723126 IN90 Myotis lucifugus 2005 Indiana GU644957 GU723128 AZ2991 Myotis yumanensis 2004 Arizona GU644958 - CA62 Myotis thysanodes 2002 California GU644959 GU723125

131

AZ2407 Myotis velifer 2004 Arizona GU644960 GU723140 AZ4490 Myotis yumanensis 2005 Arizona GU644961 GU723138 CA286 Myotis yumanensis 2002 California GU644962 GU723136 CA3481 Myotis yumanensis 2002 California GU644963 - CA414 Myotis yumanensis 2002 California GU644964 GU723133 CA114 Myotis yumanensis 2003 California GU644965 GU723135 CA49 Myotis yumanensis 2004 California GU644966 GU723137 CA9999 Myotis yumanensis 2004 California GU644967 GU723132 VA2446 Myotis lucifugus 2003 Virginia GU644968 - FL724 Nycticeius humeralis 2003 Florida GU644969 GU723163 TX5929 Nycticeius humeralis 2003 Texas GU644970 GU723165 CA2149 Parastrellus hesperus 2003 California GU644971 GU723159 CA2022 Parastrellus hesperus 2004 California GU644972 GU723161 CA164 Parastrellus hesperus 2005 California GU644973 GU723162 GA31526 Perimyotis subflavus 2004 Georgia GU644974 GU723145 GA5162 Perimyotis subflavus 2005 Georgia GU644975 - IA429 Perimyotis subflavus 2005 Iowa GU644976 GU723146 IN279 Perimyotis subflavus 2004 Indiana GU644977 - IN304 Perimyotis subflavus 2004 Indiana GU644978 GU723148 IN1487 Perimyotis subflavus 2005 Indiana GU644979 GU723150 TN169 Perimyotis subflavus 2004 Tennessee GU644980 GU723142 TN107 Perimyotis subflavus 2005 Tennessee GU644981 GU723153 TN186 Perimyotis subflavus 2005 Tennessee GU644982 GU723152 TX3775 Perimyotis subflavus 2002 Texas GU644983 GU723155 VA1615 Perimyotis subflavus 2005 Virginia GU644984 - VA1700 Perimyotis subflavus 2005 Virginia GU644985 GU723157 AZ360 Tadarida brasiliensis 2005 Arizona GU644986 GU723084 AZ468 Tadarida brasiliensis 2005 Arizona GU644987 GU723086 AZ7773 Tadarida brasiliensis 2005 Arizona GU644988 - CA141 Antrozous pallidus 2002 California GU644989 - CA0030 Tadarida brasiliensis 2004 California GU644990 GU723089 CAO42 Tadarida brasiliensis 2004 California GU644991 - CA0057 Tadarida brasiliensis 2005 California GU644992 - FL1079 Tadarida brasiliensis 2001 Florida GU644993 GU723100 FL1318 Tadarida brasiliensis 2002 Florida GU644994 GU723099 FL2369 Tadarida brasiliensis 2003 Florida GU644995 GU723098 FL385 Tadarida brasiliensis 2003 Florida GU644996 - FL1002 Tadarida brasiliensis 2004 Florida GU644997 - FL2835 Tadarida brasiliensis 2004 Florida GU644998 GU723096 FL333 Tadarida brasiliensis 2004 Florida GU644999 GU723095 FL412 Tadarida brasiliensis 2004 Florida GU645000 GU723097 FL542 Tadarida brasiliensis 2004 Florida GU645001 - FL967 Tadarida brasiliensis 2004 Florida GU645002 - TX5638 Tadarida brasiliensis 2003 Texas GU645003 - TX6224 Tadarida brasiliensis 2003 Texas GU645004 - TX6820 Tadarida brasiliensis 2003 Texas GU645005 - TX6989 Tadarida brasiliensis 2003 Texas GU645006 - TX5218 Tadarida brasiliensis 2004 Texas GU645007 - TX5573 Tadarida brasiliensis 2004 Texas GU645008 - TX6127b Tadarida brasiliensis 2004 Texas GU645009 - TX6604 Tadarida brasiliensis 2004 Texas GU645010 GU723108 TX7040 Tadarida brasiliensis 2004 Texas GU645011 - TX3545 Tadarida brasiliensis 2004 Texas GU645012 - TX6902 Lasiurus intermedius 2002 Texas - GU722930

132

GA31565 Lasiurus cinereus 2004 Georgia - GU722967 WA0453 Lasiurus cinereus 1997 Washington - GU722980 TN98 Lasiurus borealis 2004 Tennessee - GU722989 TN89 Lasiurus borealis 2004 Tennessee - GU722990 TN321 Lasiurus borealis 2004 Tennessee - GU722991 TX5499 Lasiurus seminolus 2002 Texas - GU723016 IA145 Eptesicus fuscus 2005 Iowa - GU723021 IA344 Eptesicus fuscus 2005 Iowa - GU723022 IA253 Eptesicus fuscus 2005 Iowa - GU723023 VA1290 Eptesicus fuscus 2005 Virginia - GU723048 VA1323 Eptesicus fuscus 2005 Virginia - GU723049 VA1268 Eptesicus fuscus 2005 Virginia - GU723050 VA1337 Eptesicus fuscus 2005 Virginia - GU723054 AZ4050 Eptesicus fuscus 2005 Arizona - GU723060 CA0229 Eptesicus fuscus 2002 California - GU723068 WAef07 Eptesicus fuscus 2004 Washington - GU723072 CA408 Tadarida brasiliensis 2002 California - GU723091 GA213 Tadarida brasiliensis 2005 Georgia - GU723102 GA226 Tadarida brasiliensis 2004 Georgia - GU723103 FL704 Myotis austroriparius 2004 Florida - GU723111 TN145 Myotis septentrionalis 2004 Tennessee - GU723141 GA36645 Perimyotis subflavus 2004 Georgia - GU723143 GA35162 Perimyotis subflavus 2005 Georgia - GU723144 PAps Perimyotis subflavus Unknown Pennsylvania - GU723147 FL710 Perimyotis subflavus 2000 Florida - GU723151 * Specimens for which the complete nucleoprotein gene was sequenced

133

Table 2. Phylogenetic compartmentalization of rabies viruses to bat species.

Statistic * No. individuals Observed mean (95% HPD) Null mean (95% HPD) P-value AI 372 8.93 (7.97-9.87) 32.85 (0.62-33.80) < 0.001 PS 372 67.84 (66.00-69.00) 236.40 (0.82-241.14) < 0.001 MCS (Ap) 3 2.0 (2.00-2.00) 1.02 (0.00-1.00) < 0.01 MCS (Ef) 118 26.25 (25.00-43.00) 3.25 (0.48-4.19) < 0.01 MCS (Lb) 41 7.0 (7.00-7.00) 1.81 (0.24-2.34) < 0.01 MCS (Lbl) 3 2.0 (2.000-2.000) 1.0 (0.00-1.00) < 0.01 MCS (Lc) 41 4.20 (3.00-6.00) 1.75 (0.30-2.18) < 0.01 MCS (Li) 17 12.0 (12.00-12.00) 1.21 (0.00-2.00) < 0.01 MCS (Ls) 14 3.61 (2.00-5.00) 1.12 (0.00-1.76) < 0.01 MCS (Lx) † 5 1.22 (1.00-2.00) 1.02 (0.00-1.05) 1 MCS (Ln) 17 3.52 (2.00-6.00) 1.20 (0.00-2.00) < 0.01 MCS (Ma) 2 2.0 (2.00-2.00) 1.0 (0.00-1.00) < 0.01 MCS (Mc) † 10 1.25 (1.00-2.00) 1.19 (0.00-1.51) 1 MCS (Ml) † 10 1.0 (1.00-1.00) 1.04 (0.00-1.28) 1 MCS (My) 12 3.36 (2.00-5.00) 1.07 (0.00-1.39) < 0.01 MCS (Nh) 3 2.0 (2.00-2.00) 1.0 (0.00-1.00) < 0.01 MCS (Ph) 5 5.5 (5.00-5.00) 1.01 (0.00-1.01) < 0.01 MCS (Ps) 14 4.66 (2.00-9.00) 1.14 (0.00-2.00) < 0.01 MCS (Ct) 1 1.0 (1.00-1.00) 1 (0.00-1.00) 1 MCS (Tb) 56 29.87 (28.00-33.00) 2.05 (0.73-2.53) < 0.01

* The association index (AI) and parsimony score (PS) tests compartmentalization across the entire phylogenetic tree, while the maximum monophyletic clade statistic (MC) assesses the correlation between each species and virus phylogeny. Abbreviations are bat species names as in Table A3. † MspV and LxLiV were statistically compartmentalized to several sympatric species within the genera Myotis and Lasiurus, respectively (MspV: Maximum clade size statistic, MCS = 3.83, P < 0.01; LxLiV: MCS = 2.83, P < 0.01).

134

Table A3. Coverage and haplotypic diversity for bat species with reference cytochrome oxidase subunit I sequence data. The number of reference sequences is shown with the number of unique haplotypes listed in parentheses. Genetic data from bats in this study is shown by number of bats identified in a phylogenetic species, with the number of incorrect morphological diagnoses in parentheses.

Text Reference This Reference Taxonomy Common name abbreviation sequences study * accession no. † Vespertilionidae Antrozous pallidus Pallid bat Ap 7 (6) 2 (0) MVZ 155864 Corynorhinus Rafinesque's big-eared Cr 7 (5) 0 LSUMZ M119 rafinesquii bat Corynorhinus Townsend's big-eared Ct 4 (4) 0 MVZ 146694 townsendii bat Eptesicus fuscus Big brown bat Ef 15 (8) 60 (4) MVZ 198298 Euderma maculatum Spotted bat Em 2 (2) 0 ROM 692 Idionycteris phyllotis Allen’s big-eared bat Ip 2 (1) 0 - Lasionycteris Silver-haired bat Ln 12 (6) 11 (0) MVZ 192695 noctivagans Lasiurus blossevillii Western red bat Lbl 7 (7) 2 (0) CIB 11649 Lasiurus borealis Eastern red bat Lb 20 (14) 28 (1) CM 102846 Lasiurus cinereus Hoary bat Lc 9 (3) 29 (0) CM 82018 ASNHC 1408, Lasiurus intermedius ‡ Northern yellow bat Li 5 (5) 13 (0) LSUMZ M352 LSUMZ Lasiurus seminolus Seminole bat Ls 6 (4) 8 (0) M6259 Lasiurus xanthinus ‡ Western yellow bat Lx 11 (4) 4 (1) CIB 10085 Myotis austroriparious Southeastern Myotis Ma 6 (2) 3 (0) USNM 568964 Myotis auriculus Southwestern Myotis Mau 1 (1) 0 UAM 66652 Myotis californicus § California Myotis Mc 12 (9) 8 (2) ASNHC 11511 Western small-footed Myotis ciliolabrum § Mc 9 (5) 0 ASNHC 12082 Myotis Western long-eared Myotis evotis || Ml 10 (7) 0 MVZ 201321 Myotis Myotis grisescens Gray bat Mg 6 (3) 0 - Myotis keenii || Keen’s Myotis Ml 2 (1) 0 UAM 23338 Eastern small-footed Myotis leibii § Mc 22 (3) 0 - Myotis Myotis lucifugus || Little brown bat Ml 83 (32) 9 (1) CM 102862 Northern long-eared Myotis septentrionalis Ms 19 (4) 1 (0) CM 82047 Myotis Myotis sodalis Indiana bat Mso 56 (15) 0 - Myotis thysanodes || Fringed bat Ml 7 (5) 0 ASNHC 12975 MVZ 146766, Myotis velifer Cave Myotis Mv 3 (3) 2 (0) ASNHC 8265 Myotis volans Long-legged bat Mvo 81 (28) 0 MVZ 198299

135

Myotis yumanensis Yuma Myotis My 25 (12) 8 (1) MVZ 155859 Nycticeius humeralis Evening bat Nh 11 (9) 3 (0) ASNHC 12501 Parastrellus hesperus Canyon bat Ph 3 (3) 5 (0) MVZ 198302 Perimyotis subflavus Tri-colored bat Ps 20 (11) 16 (6) CM 106677 Molossidae Tadarida brasiliensis Mexican free-tailed bat Tb 2 (2) 31 (1) ASNHC 12083 Phyllostomidae Desmodus rotundus Common vampire bat Dr 1 (1) 0 ROM 113682

* Some submissions were called Myotis sp. and were excluded in the overall statistic of mis-identification for the phylogenetic species. † Codes indicate location of vouchered specimens as follows: ASNHC: Angelo State Natural History Collection; CIB: Centro de Investigaciones Biológicas del Noroeste; CM: Carnegie Museum of Natural History; LSUMZ: Louisiana State University, Museum of Natural Science; MVZ: Museum of Vertebrate Zoology, University of California, Berkeley; ROM: Royal Ontario Museum; UAM: University of Alaska Museum; USNM: National Museum of Natural History. ‡ A clade of L. intermedius, sister to L. xanthinus, may be L. ega, which once had subspecific placement in xanthinus; however, no reference sequences were available for L. ega. § Myotis californicus, M. ciliolabrum and M. leibii and are not reciprocally monophyletic with COI data, and are called M. californicus complex in this analysis. || M. lucifugus, M. evotis, M. thysanodes and M. keenii are not reciprocally monophyletic with the COI data, and are called M. lucifugus complex in this analysis.

136

Table A4. Genetic divergence of 43 individuals infected by cross-species transmission relative to individuals infected by within-species transmission within the same rabies virus lineage. Genetic divergence from donor lineage in excess of outlier cutoff (see Supporting Materials and Methods) highlighted in boldface type.

Viral Mean minimum distance to donor Minimum distance to Bat ID Donor Recipient lineage for WST (outlier cutoff) donor for CST * AZ2991† E. fuscus M. yumanensis EfV1 1.90E-03 (5.32E-03) 5.15E-03 AZ2944 E. fuscus L. cinereus EfV1 1.90E-03 (5.32E-03) 5.15E-03 CA9242 T. brasiliensis E. fuscus TbV 1.88E-03 (4.31E-03) 0.00 VA2446 E. fuscus M. lucifugus EfV3 2.22E-03 (8.58E-03) 0.00 IN90 E. fuscus M. lucifugus EfV3 2.22E-03 (8.58E-03) 1.70E-03 VA484 † E. fuscus L. cinereus EFV3 2.22E-03 (8.58E-03) 3.43E-03 TX6127 † L. borealis L. seminolus LbV1 7.17E-03 (9.39E-03) 3.39E-03 TX5419 L. borealis L. seminolus LbV1 7.17E-03 (9.39E-03) 8.52E-03 TX3775 L. borealis P. subflavus LbV2 4.26E-03 (1.03E-02) 5.08E-03 TX5929 L. borealis N. humeralis LbV2 4.26E-03 (1.03E-02) 5.13E-03 TN39 L. borealis M. lucifugus LbV2 4.26E-03 (1.03E-02) 4.49E-03 TX5512 † L. borealis L. seminolus LbV2 4.26E-03 (1.03E-02) 0.00 TX5565 L. borealis L. seminolus LbV2 4.26E-03 (1.03E-02) 0.00 TX5850 L. borealis L. seminolus Lb2V 4.26E-03 (1.03E-02) 2.23E-03 TX6197 L. borealis L. seminolus LbV2 4.26E-03 (1.03E-02) 3.38E-03 FL793 L. borealis L. seminolus LbV2 4.26E-03 (1.03E-02) 4.48E-03 TX6265 L. borealis L. seminolus LbV2 4.26E-03 (1.03E-02) 1.38E-03 MI738 L. borealis E. fuscus LbV2 4.26E-03 (1.03E-02) 0.00 CA16461 L. borealis L. blossevillii LbV2 4.26E-03 (1.03E-02) 3.38E-03 CA2306 L. borealis M. californicus LbV2 4.26E-03 (1.03E-02) 2.87E-03 CA0077 L. borealis L. blossevillii LbV2 4.26E-03 (1.03E-02) 5.10E-03 CA141 L. cinereus A. pallidus LcV 1.71E-03 (4.28E-03) 1.70E-03 CA6054 L. cinereus M. californicus LcV 1.71E-03 (4.28E-03) 0.00 TX3054 L. cinereus L. intermedius LcV 1.71E-03 (4.28E-03) 1.70E-03 TX6706 L. cinereus L. intermedius LcV 1.71E-03 (4.28E-03) 1.70E-03 ID7318 † L. cinereus E. fuscus LcV 1.71E-03 (4.28E-03) 1.70E-03 CA33 † L. cinereus L. blossevillii LcV 1.71E-03 (4.28E-03) 5.15E-03 CA286 L. cinereus M. yumanensis LcV 1.71E-03 (4.28E-03) 3.42E-03 AZ2472 L. cinereus E. fuscus LcV 1.71E-03 (4.28E-03) 3.42E-03 ID7376 L. cinereus L. noctivagans LcV 1.71E-03 (4.28E-03) 1.70E-03 AZ6954 L. cinereus L. xanthinus LcV 1.71E-03 (4.28E-03) 0.00 CA62 L. cinereus M. lucifugus LcV 1.71E-03 (4.28E-03) 1.70E-03 CA09 L. cinereus E. fuscus LcV 1.71E-03 (4.28E-03) 0.00 CA141 L. cinereus T. brasiliensis LcV 1.71E-03 (4.28E-03) 0.00 CA2070 L. cinereus L. xanthinus LcV 1.71E-03 (4.28E-03) 0.00 MI2114 † L. cinereus E. fuscus LcV 1.71E-03 (4.28E-03) 0.00 MI1100 L. noctivagans M. lucifugus LnV 3.06E-03 (5.99E-03) 1.70E-03

137

WA1316 L. noctivagans M. californicus LnV 3.06E-03 (5.99E-03) 1.70E-03 FL701 L. seminolus L. borealis LsV 1.13E-03 (4.24E-03) 0.00 L. xanthinus or AZ7590 † E. fuscus LxLiV 4.29E-03 (3.61E-03) 1.70E-03 L. intermedius CA6846 M. yumanensis M. californicus MyV 8.09E-03 (1.60E-02) 5.10E-03 VA1597 P. subflavus E. fuscus PsV 2.35E-03 (7.44E-03) 1.69E-03 GA6568 P. subflavus E. fuscus PsV 2.35E-03 (7.44E-03) 5.09E-03

* Pairwise genetic distances calculated using substitution models selected independently for each rabies virus lineage by AIC in Modeltest. † Identity of recipient species confirmed by morphology alone.

138

Table A5. Maximum likelihood parameter estimates from program Migrate.

Donor (j) Recipient (i) θj (95% CI) βij (95% CI) Rij L. cinereus A. pallidus 0.038 (0.029 - 0.050) 82.736 (11.225 - 594.789) 0.108 L. borealis E. fuscus 0.023 (0.016 - 0.035) 17.489 (7.854 - 33.503) 0.014 L. cinereus E. fuscus 0.038 (0.029 - 0.050) 24.015 (11.103 - 43.395) 0.031 L. xanthinus E. fuscus 0.007 (0.004 - 0.014) 38.207 (17.775 - 70.613) 0.009 M. lucifugus E. fuscus 0.118 (0.073 - 0.209) 2.951 (2.214 - 11.074) 0.012 complex P. subflavus E. fuscus 0.017 (0.011 - 0.029) 33.270 (13.122 - 68.340) 0.019 T. brasiliensis E. fuscus 0.042 (0.033 - 0.054) 5.249 (0.589 - 17.528) 0.008 L. seminolus L. borealis 0.004 (0.002 - 0.009) 57.842 (20.454 - 154.635) 0.008 L. borealis L. borealis 0.023 (0.016 - 0.035) 130.272 (34.720 - 411.162) 0.103 L. cinereus L. borealis 0.038 (0.029 - 0.050) 131.663 (19.211 - 712.183) 0.172 E. fuscus L. cinereus 0.011 (0.009 - 0.013) 7.134 (1.445 - 21.516) 0.003 L. intermedius L. cinereus 0.009 (0.006 - 0.015) 12.023 (2.764 - 34.192) 0.004 L. cinereus L. intermedius 0.038 (0.029 - 0.050) 40.377 (7.586 - 97.895) 0.053 L. xanthinus L. intermedius 0.007 (0.004 - 0.014) 55.034 (16.730 - 136.634) 0.013 L. cinereus L. noctivagans 0.038 (0.029 - 0.050) 25.685 (19.263 - 94.340) 0.033 L. borealis L. seminolus 0.023 (0.016 - 0.035) 92.996 (34.053 - 120.715) 0.073 L. cinereus L. xanthinus 0.038 (0.029 - 0.050) 417.934 (85.766 - 743.914) 0.545 M. californicus L. cinereus 0.038 (0.029 - 0.050) 44.881 (13.759 - 140.019) 0.059 complex M. californicus L. noctivagans 0.014 (0.009 - 0.022) 45.932 (11.109 - 219.268) 0.022 complex M. californicus M. yumanensis 0.011 (0.006 - 0.021) 143.013 (95.302 - 214.130) 0.053 complex E. fuscus M. lucifugus complex 0.011 (0.009 - 0.013) 32.839 (16.795 - 60.534) 0.012 L. borealis M. lucifugus complex 0.023 (0.016 - 0.035) 30.169 (12.013 - 160.066) 0.024 L. cinereus M. lucifugus complex 0.038 (0.029 - 0.050) 41.039 (10.757 - 202.283) 0.054 L. noctivagans M. lucifugus complex 0.014 (0.009 - 0.022) 95.290 (9.969 - 227.220) 0.046 M. californicus M. lucifugus complex 0.118 (0.073 - 0.209) 362.787 (104.644 - 2599.60) 1.474 complex L. cinereus M. yumanensis 0.038 (0.029 - 0.050) 26.853 (2.145 - 93.244) 0.035 M. californicus M. yumanensis 0.118 (0.073 - 0.209) 115.499 (86.624 - 169.494) 0.469 complex M. lucifugus M. yumanensis 0.118 (0.073 - 0.209) 46.234 (15.732 - 119.358) 0.188 complex L. borealis N. humeralis 0.023 (0.016 - 0.035) 47.409 (12.491 - 216.604) 0.037 L. borealis P. subflavus 0.023 (0.016 - 0.035) 41.417 (12.628 - 99.823) 0.033 L. cinereus T. brasiliensis 0.038 (0.029 - 0.050) 18.291 (8.125 - 34.223) 0.024

139

Table A6. Morphological and ecological traits of bat species and data sources.

Wing Wing Body Forearm Summer Known to No. Bat species aspect loading length length Source roost type † roost with no. ratio (g/cm2) * (mm) (mm) Antrozous 2,3,12,16,18, 1 6.1 0.082 135 60 D,F (35-39) pallidus 20,22 Corynorhinus (35, 37, 40- 2 5.9 0.073 90 39.2 D,E 1,16,17,22 townsendii 42) Eptesicus 1,13,15,17 (35, 37, 43, 3 6.4 0.096 138 54 A,D fuscus 18,22 44) Lasionycteris (35, 37, 45- 4 6.6 0.083 115 44 A,B,F 15, 21 noctivagans 47) Lasiurus 5 6.7 0.143 108.9 40.6 C - (35, 37, 48) borealis Lasiurus 6 7.5 ‡ 0.153 ‡ 112 40 C - (19, 35, 49) blossevillii Lasiurus 7 8.1 0.168 150 57 C - (35, 37) cinereus Lasiurus (19, 35, 50, 8 7.5 ‡ 0.153 ‡ 131.5 48.1 C - intermedius 51) Lasiurus (19, 35, 52, 9 7.5 ‡ 0.153 ‡ 97.7 40.2 C - seminolus 53) Lasiurus 10 7.5 ‡ 0.153 ‡ 146 50 C - (19, 35, 54) xanthinus Myotis (19, 35, 55, 11 6 ‡ 0.067 ‡ 83.7 36 A,D,E 13,22 austroriparius 56) Myotis (35, 37, 57- 12 5.6 0.049 95 35 B,D,F 1 californicus 59) Myotis 13 6 0.076 102 41 A,B,D,E,F 3,11,15,17,18 (35, 37, 60) lucifugus - (35, 37, 61, 14 Myotis keenii 5.8 0.069 94 36.4 A,D 62) Myotis (19, 35, 63, 15 6 ‡ 0.067 ‡ 95 36.4 A,B,D 3,4,13,21 septentrionalis 64) Myotis (35, 37, 65, 16 6.1 0.063 77 40 B,D,E 1,2,17 thysanodes 66) 2,3,13,16,18, (35, 37, 67, 17 Myotis velifer 6.2 0.064 99.5 46 D,E 22 68) Myotis (35, 37, 69- 18 6.3 0.080 84 37 D 1,3,13,17,22 yumanensis 71) Nycticeius (35, 37, 50, 19 6.8 0.109 93 38 A,D 22 humeralis 72-74) Parastrellus (35, 37, 75, 20 5.7 0.070 86 33 F 1 hesperus 76) Perimyotis (35, 37, 77, 21 6.2 0.057 89 34.1 A,D,E,F 4,15 subflavus 78) Tadarida 1,3,11,13,17, (35, 37, 70, 22 8.2 0.118 109 42.2 D,E brasiliensis 18,19 79, 80) * Converted to g/cm2 using wing area corrections of ref. (19). † Typical summer roosting structures are coded as follows: A: tree hollow, B: tree bark, C: tree foliage, D: man- made structures, E: cave, F: rock crevice. ‡ No published estimates available; mean value from North American representatives of the genus used in analyses.

140

Table A7. Summary of partial Mantel tests comparing predictors of cross-species rabies transmission (Rij) among 17 bat species. Models describe principal effects controlled for other terms. Terms included were: the phylogenetic distance between hosts (PD), the proportion geographic overlap between hosts (Range) and the absolute value of body length differences between donor and recipient species (Body). Shown are statistical significance and explanatory power according to partial Mantel tests for terms retained in initial generalized linear models (see Materials and Methods).

Model r (95% CI) P-value

Rij = PD|Range,Body -0.418 (-0.502, -0.271) 0.001

Rij = PD|Range -0.415 (-0.50, -0.244) 0.001

Rij = PD -0.406 (-0.511, -0.246) 0.001

Rij = PD|Body -0.406 (-0.508, -0.239) 0.001

Rij = Body|PD,Range -0.191 (-0.285, -0.128) 0.003

Rij = Body|Range -0.185 (-0.233, -0.152) 0.007

Rij = Body -0.174 (-0.216, -0.14) 0.01

Rij = Body|PD -0.172 (-0.239, -0.123) 0.013

Rij = Range|PD,Body 0.143 (0.072, 0.256) 0.064

Rij = Range|PD 0.115 (0.048, 0.214) 0.134

Rij = Range|Body 0.092 (0.011, 0.216) 0.186

Rij = Range 0.065 (-0.018, 0.213) 0.28

Table A8. Summary of best 5 GLMs examined to explain variation in cross-species transmission between bat species. Models were constructed with the following terms: phylogenetic distance between hosts (PD), proportion of geographic overlap (Range), potential roost overlap (R1), documented roost overlap (R2), the sample size of the recipient species (N) and the absolute values of species differences in wing loading (WL), wing aspect ratio (WAR) and body length (Body). Terms absent from best 5 models are omitted from the table.

2 Model r AIC Δ AIC wi *

Rij = PD + Range 0.443 98.598 0 0.378

Rij = PD + Range + N 0.451 100.128 1.53 0.176

Rij = PD + Range + Body 0.445 100.456 1.858 0.149

Rij = PD + Range + WAR 0.443 100.459 1.861 0.149

Rij = PD + Range + R2 0.445 100.472 1.874 0.148

* wi is the AIC weight, the probability that the model is the best one of those tested

141

Table A9. Oligonucleotide primers used in RT-PCR and sequencing of the rabies virus nucleoprotein gene.

Primer ID Sense Sequence SAD NT position † 001 Forward 5’ ACGCTTAACGAMAAA 3’ 1-15 550 Forward 5’ ATGTGYGCTAAYTGGAGYAC 3’ 647-666 1066 * Forward 5’ AGAAGATTCTTCAGGGA 3’ 1136-1155 930 Reverse 5’ CATCCAACAAAGTGAATGAGG 3’ 1000-1020 114 Reverse 5’ CCCATATAGCATCCTAC 3’ 1013-1030 1066 * Reverse 5’ TCCCTRAAGAATCTYCTYTC 3’ 1136-1155 304 * Reverse 5' TTGACGAAGATCTTGCTCAT 3' 1514-1533

* Primers adapted from ref. 4. † Positions given relative to the complete genome sequence of the SAD B19 vaccine strain (GenBank accession no. M31046).

LITERATURE CITED

1. P. D. N. Hebert, A. Cywinska, S. L. Ball, J. R. DeWaard, Proc. R. Soc. London Ser. B 270, 313 (2003).

2. D. D. Posada, K. A. Crandall, Bioinformatics 14, 817 (1998).

3. D. L. Swofford. (Sinauer Assoc., Sunderland, Mass., 2002).

4. J. S. Smith, in Rabies, A. C. Jackson, W. H. Wunner, Eds. (Academic Press, San Diego, 2002), pp. 79-111.

5. K. Katoh, K.-I. Kuma, H. Toh, T. Miyata, Nucl. Acids Res. 33, 511 (2005).

6. T. A. Hall, Nucleic Acids Symp. Ser. 41, 95 (1999).

7. J. P. Huelsenbeck, F. Rhonquist, Bioinformatics 17, 754 (2001).

8. D. J. Zwickl, The University of Texas at Austin (2006).

9. M. P. Cummings, J. C. Huskamp, Educause Review 40, 116 (2005).

10. A. L. Bazinet, M. P. Cummings, in Distributed & Grid Computing - Science Made Transparent for Everyone. Principles, Applications and Supporting Communities., M. H. W. Weber, Ed. (Tectum Publishing, Marburg, 2007).

11. A. Gelman, D. B. Rubin, Statistical Science 7, 457 (1992).

142

12. J. Parker, A. Rambaut, O. G. Pybus, Infect., Genet. Evol. 8, 239 (2008).

13. P. Beerli, J. Felsenstein, Proc. Nat. Acad. Sci. U.S.A. 98, 4563 (2001).

14. C. J. Geyer, Technical Report 568 School of Statistics, (1991).

15. R. R. Steece, J. J. S. Altenbach, J. Wildl. Dis. 25, 490 (1989).

16. V. V. Shankar, R. A. Bowen, A. D. Davis, C. E. Rupprecht, T. J. O'Shea, J. Wildl. Dis. 40, 403 (2004).

17. F. R. Jackson et al., J. Wildl. Dis. 44, 612 (2008).

18. D. L. Deangelis, Ecology 56, 238 (1975).

19. U. M. Norberg, J. M. V. Rayner, Proc. R. Soc. London, Ser. B 316, 337 (1987).

20. B. D. Patterson et al. (Natureserve, Arlington, Virginia, USA, 2005).

21. R Development Core Team, in R Foundation for Statistical Computing. ( Vienna, 2008).

22. G. J. Hughes, L. A. Orciari, C. E. Rupprecht, J. Gen. Virol. 86, 1467 (2005).

23. P. L. Davis, H. Bourhy, E. C. Holmes, Infect., Genet. Evol. 6, 464 (2006).

24. J. E. Scatterday, J. Am. Vet. Med. Assoc. 124, 125 (1954).

25. P. Lemey, A. Rambaut, A. J. Drummond, M. A. Suchard, PLoS Comp. Biol. 5, Article No.: e1000520 (2009).

26. A. J. Drummond, A. Rambaut, BMC Evol. Biol. 7, 214 (2007).

27. S. A. Nadin-Davis et al., Virus Res. 74, 139 (2001).

28. J. S. Smith, L. A. Orciari, P. A. Yager, Semin. Virol. 6, 387 (1995).

29. A. Velasco-Villa et al., J. Clin. Microbiol. 44, 1697 (2006).

30. S. A. Nadin-Davis, E. Loza-Rubio, Virus Res. 117, 215 (2006).

31. A. Rambaut, A. J. Drummond. (2007).

32. M. A. Suchard, R. E. Weiss, J. S. Sinsheimer, Biometrics 61, 665 (2005).

33. P. Librado, J. Rosas, Bioinformatics 25, 1451 (2009).

143

34. D. G. Constantine, Pub. Hlth. Repts. 82, 867 (1967).

35. A. Lollar, B. French, Captive care and medical reference for the rehabilitation of insectivorous bats. (Bat World Publication, Mineral Wells, 2002).

36. J. W. Hermanson, T. J. O'Shea, Mamm. Spec. 213, 1 (1983).

37. J. Farney, E. D. Fleharty, J. Mammal. 50, 362 (1969).

38. T. A. Vaughan, T. J. Oshea, J. Mammal. 57, 19 (1976).

39. R. Davis, J. Mammal. 50, 729 (1969).

40. T. H. Kunz, R. A. Martin, Mamm. Spec. 175, 1 (1982).

41. S. R. Humphrey, T. H. Kunz, J. Mammal. 57, 470 (1976).

42. W. W. Dalquest, J. Mammal. 28, 17 (1947).

43. A. Kurta, R. H. Baker, Mamm. Spec. 356, 1 (1990).

44. S. J. Agosta, Mamm. Rev. 32, 179 (2002).

45. T. H. Kunz, Mamm. Spec. 172, 1 (1982).

46. T. A. Mattson, Great Basin Nat. 56, 247 (1996).

47. H. J. Parsons, D. A. Smith, R. F. Whittam, J. Mammal. 67, 598 (1986).

48. K. A. Shump, A. U. Shump, Mamm. Spec. 183, 1 (1982).

49. H. H. Genoways, R. J. Baker, Tex. J. Sci. 40, 111 (1988).

50. M. A. Menzel, D. M. Krishon, T. C. Carter, J. Laerm, Fla. Sci. 62, 185 (1999).

51. W. D. Webster, J. K. Jones, R. J. Baker, Mamm. Spec. 132, 1 (1980).

52. C. D. Hein, S. B. Castleberry, K. V. Miller, J. Mammal. 89, 964 (2008).

53. K. T. Wilkins, Mamm. Spec. 280, 1 (1987).

54. J. L. Higginbotham, L. K. Ammerman, M. T. Dixon, Southwest. Nat. 44, 343 (1999).

55. C. Jones, R. W. Manning, Mamm. Spec. 332, 1 (1989).

144

56. K. M. Mirowsky, P. A. Horner, R. W. Maxev, S. A. Smith, Southwest. Nat. 49, 294 (2004).

57. M. R. Simpson, Mamm. Spec. 428, 1 (1993).

58. R. M. Brigham, M. J. Vonhof, R. M. R. Barclay, J. C. Gwilliam, J. Mammal. 78, 1231 (1997).

59. P. H. Krutzsch, J. Mammal. 35, 539 (1954).

60. M. B. Fenton, M. R. Barclay, Mamm. Spec. 142, 1 (1980).

61. J. H. Fitch, K. A. Shump, Mamm. Spec. 121, 1 (1979).

62. C. G. V. Dejong, D. W. Nagorsen, Can. J. Zool.-Rev. Can. Zool. 72, 1069 (1994).

63. M. C. Caceres, M. R. Barclay, Mamm. Spec. 634, 1 (2000).

64. J. C. Timpone, J. G. Boyles, K. L. Murray, D. P. Aubrey, L. W. Robbins, Am. Midl. Nat. 163, 115 (2010).

65. M. J. O'Farrell, E. H. Studier, Mamm. Spec. 137, 1 (1980).

66. P. M. Cryan, M. A. Bogan, G. M. Yanega, Acta Chiropt. 3, 43 (2001).

67. J. H. Fitch, K. A. Shump, A. U. Shump, Mamm. Spec. 149, 1 (1981).

68. T. H. Kunz, Ecology 55, 693 (1974).

69. M. J. Evelyn, D. A. Stiles, R. A. Young, Biol. Conserv. 115, 463 (2004).

70. K. Geluso, J. N. Mink, Southwest. Nat. 54, 421 (2009).

71. W. W. Dalquest, Am. Midl. Nat. 38, 224 (1947).

72. C. D. Hein, K. V. Miller, S. B. Castleberry, J. Wildl. Manage. 73, 511 (2009).

73. L. C. Watkins, K. A. Shump, Am. Midl. Nat. 105, 258 (1981).

74. L. C. White, Mamm. Spec. 23, 1 (1972).

75. S. R. Hoofer, R. A. Van den Bussche, I. Horacek, J. Mammal. 87, 981 (2006).

76. R. Barbour, W. Davis, Bats of America. (The University Press of Kentucky, Lexington, Kentucky, 1969).

145

77. M. S. Fujita, T. H. Kunz, Mamm. Spec. 228, 1 (1984).

78. J. O. Whitaker, J. Mammal. 79, 651 (1998).

79. K. T. Wilkins, Mamm. Spec. 331, 1 (1989).

80. M. Betke et al., J. Mammal. 89, 18 (2008).

146

APPENDIX B

SUPPORTING INFORMATION FOR CHAPTER 3

MATERIALS AND METHODS

RT-PCR amplification and sequencing of rabies viruses. For sequences generated in this study, total RNA was extracted directly from naturally infected rabid bat brains using Trizol

(Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. A 903 bp fragment comprising the last 687 bp of the N gene, a non-coding region following the 3’ end of N and a small fragment of the phosphoprotein gene was amplified by reverse transcription-polymerase chain reaction using oligonucleotide primers 550F and 304R, as described previously (1).

Amplicons were purified using ExoSAP-IT® (USB Corporation) and sequencing was carried out using an Applied Biosystems 3730 automated DNA sequencer. Chromatograms were edited by eye in Bioedit and sequences were aligned using Clustal W2 (2). Sequences generated herein have been deposited into Genbank under accession numbers JN594500 - JN594503 and

DQ445318 - DQ445330, DQ445352 (updated sequences).

Auxiliary life history and ecological data. We collected information on the overwintering activity patterns, migratory status and metabolic rates of the bat species that serve as reservoir hosts for the rabies viruses included here from the primary literature and existing databases

(Table B2). Long distance migration was defined as seasonal movements of individual bats of at least 1000 km (3). Data on coloniality and basal metabolic rate (BMR) where extracted from the

147

Pantheria database when available and from the primary literature otherwise (4). For 4 species for which BMR data were unavailable, we borrowed values from species within the same genus or family that had similar body mass. Notably, body mass explains > 92% of variation in BMR in bats and phylogeny explains much of the residual variation (5). Data on torpid metabolic rate

(TMR) were collected from the primary literature or estimated herein. When TMR estimates spanned a range of temperatures or spatial locations, rates were selected to match the conditions that bats are likely to experience in northern latitudes of their range where hibernation/torpor is most important. To calculate TMR for species that lacked values in the literature, we estimated the relationship between BMR and TMR for 9 species in our dataset for which both values were reported in the literature. This relationship was remarkably consistent across species (TMR = 2.2

- 3.2% of BMR, mean = 2.8%), with the exception of Tadarida brasiliensis, for which TMR was

12.6% of BMR. Because the reported estimate of TMR for that species (and for the other subtropical and tropical species in our study) likely represented a daily torpor rather than longer- duration, seasonal torpor, it was excluded from the calculation of the average mentioned above

(6). Overwintering activity was challenging to classify because the frequency and duration of bat activity during winter are poorly understood for many temperate bat species and can vary substantially throughout geographic ranges (7, 8). Consequently, whether records of bat activity in winter, particularly in southern regions of the United States that experience mild winters with year-round food availability, indicate occasional arousal from a predominate state of torpor or true year-round activity, is uncertain. Therefore, we classified species as inactive during winter if extended bouts of seasonal torpor or hibernation were reported in any part of their geographic range, recognizing that this classification may be overly conservative. As a potentially more geographically explicit proxy of year-round activity, the climatic region (tropical, subtropical,

148

temperate) of the center of the geographic range of each viral lineage was also recorded. North

American lineages circulating between 35° and 23.5° latitude and South American lineages circulating south of -23.5° latitude that support mean winter temperatures of ≥ 10°C were considered subtropical and lineages found towards and away from the equator relative to these latitudes were classified as tropical and temperate, respectively (9).

FIGURES AND TABLES

Myotis yumanensis (MyV1) Myotis nigricans (MySAV) Myotis californicus (MyV2) Perimyotis subflavus (PsV) Parastrellus hesperus (PhV) Eptesicus furinalis (EfSAV) Eptesicus fuscus (EfV3) Eptesicus fuscus (EfV2) Eptesicus fuscus (EfV1a) Eptesicus fuscus (EfV1b) Lasionycteris noctivagans (LnV) Lasiurus seminolus (LsV) Lasiurus borealis (LbV2) Lasiurus borealis (LbV1) Lasiurus intermedius (LiV) Lasiurus xanthinus (LxV) Lasiurus cinereus (LcV) Nyctinomops laticaudatus (NlV) Tadarida brasiliensis (TbV) Tadarida brasiliensis (TbSAV) Desmodus rotundus (DrV) 0.1 9.31e-5 - 4.24e-4 4.25e-4 - 7.55e-4 7.56e-4 - 1.09e-3 1.10e-3 - 1.42e-3 1.43e-3 - 1.75e-3 1.76e-3 - 2.08e-3

Figure B1. Maximum likelihood phylogenetic tree describing the phylogenetic relationships among bat lineages in the study and the corresponding rates of viral evolution for their associated rabies virus lineages. The figure shows the phylogenetic tree with the highest log likelihood after five replicate ML searches in Garli using the TIM1+I+Γ model of nucleotide substitution suggested by AICc in jModeltest (33, 35). A single outgroup sequence from Emballonura alecto (Genbank Accession No. HM540244) was pruned from the tree prior to the analysis of phylogenetic signal. The scale bar indicates substitutions per site. The rates shown are estimates from the independent lineage models, and neither these rates nor the hierarchical phylogenetic model rates showed significant association with the bat phylogeny (see Chapter 3).

149

Table B1. Host species, divergence time summaries and sampling information for rabies virus lineages.

Date TMRCA TMRCA range of Number of Lineage Host species (95% HPD) † (95% HPD) ‡ sequences sequences DrV Desmodus rotundus 202 (146-308) 171 (122-229) 22 60 TbSAV§ Tadarida b. brasiliensis 304 (227-407) 32 (18-48) 12 10 Tadarida brasiliensis TbV mexicana; T. b. 195 (139-261) 65 (44-95) 26 81 cynocephala Nyctinomops NlV 155 (81-240) 55 (29-101) 16 17 laticaudatus EfSAV Eptesicus furinalis 280 (188-356) 34 (22-49) 10 14 Eptesicus fuscus EfV1a 185 (107-271) 52 (32-85) 10 20 bernardinus EfV1b Eptesicus fuscus fuscus 185 (107-271) 127 (74-202) 19 26 Eptesicus fuscus fuscus; EfV2 243 (171-317) 80 (54-107) 22 32 E. f. bernardinus Eptesicus fuscus EfV3 243 (171-317) 145 (103-204) 34 69 bernardinus LbV1 Lasiurus borealis 84 (62-108) 62 (46-83) 20 17 LbV2 Lasiurus borealis 103 (70-134) 44 (34-57) 25 46 LcV Lasiurus cinereus 87 (64-111) 47 (37-60) 28 72 Lasiurus intermedius LiV 147 (82-224) 46 (31-69) 17 11 floridanus Lasionycteris LnV 98 (72-129) 51 (37-77) 30 40 noctivagans LsV Lasiurus seminolus 88 (63-111) 41 (31-66) 27 8 LxV Lasiurus xanthinus 85 (56-113) 33 (19-50) 11 9 MSAV§ Myotis nigricans 281 (193-363) 161 (95-233) 4 12 * MV1 Myotis yumanensis 305 (223-395) 103 (63-172) 13 11 * MV2 Myotis californicus 261 (182-330) 216 (154-288) 20 25 PhV Parastrellus hesperus 255 (182-335) 205 (131-288) 25 38 PsV Perimyotis subflavus 83 (61-107) 43 (31-60) 17 30

* Viruses circulate in several species of Myotis in the western United States, see Streicker et al. (2010) for details. † Estimate of the time since the most recent common ancestor (TMRCA) including the stem branch leading to existing lineage diversity. Theoretically, this value represents the number of years since host shift into each bat taxon, but is sensitive to incomplete knowledge of existing viral lineage diversity. ‡ Estimate of the TMRCA for existing viral genetic diversity, not including the leading stem branch. § “SAV” notation refers to lineages found in South American bat populations/species that contain a congeneric North American member in the dataset.

150

Table B2. Life history traits of bat species and data sources.

Long distance Seasonal Climatic ‡ Lineage Host species BMR TMR migrant Colonial activity region Sources † DrV Desmodus rotundus 2.777 0.078 N Y Y TR (1-3) † † EfSAV Eptesicus furinalis 2.872 0.080 N Y Y TR (1, 4, 5) Eptesicus fuscus EfV1a pallidus 4.576 0.145 N Y N ST (2, 6, 7) Eptesicus fuscus EfV1b bernardinus 4.576 0.145 N Y N TE (2, 6-8) Eptesicus fuscus EfV2 fuscus 4.576 0.145 N Y N TE (2, 6-8) Eptesicus fuscus fuscus; E. f. EfV3 bernardinus 4.576 0.145 N Y N TE (2, 6-8) LbV1 Lasiurus borealis 3.264 0.099 Y N N TE (2, 6-8) LbV2 Lasiurus borealis 3.264 0.099 Y N N TE (2, 6-8) † LcV Lasiurus cinereus 2.226 0.062 Y N N TE (2, 6, 8) Lasiurus intermedius (2, 6, 8- † LiV floridanus 2.266 0.063 N Y Y ST 11) Lasionycteris (6, 12, † LnV noctivagans 2.864 0.115 Y Y N TE 13) (2, 6, 8, † LsV Lasiurus seminolus 2.927 0.082 N N N ST 14, 15) (8, 16- † † LxV Lasiurus xanthinus 2.845 0.079 Y N Y ST 19) † MSAV Myotis nigricans 1.535 0.043 N Y Y TR (1, 2, 20) (2, 6, 8, * † MV1 Myotis yumanensis 2.891 0.081 N Y N ST 21) (2, 6, 8, * MV2 Myotis californicus 2.651 0.057 N Y N TE 22, 23) Nyctinomops (2, 24, † NlV laticaudatus 1.887 0.053 N Y Y TR 25) (2, 6, 8, † † PhV Parastrellus hesperus 3.572 0.100 N Y N ST 22, 26) (2, 6, 8, † † PsV Perimyotis subflavus 3.130 0.087 N Y N TE 27) Tadarida b. (2, 28- TbSAV brasiliensis 2.663 0.337 Y Y Y TR 31) Tadarida brasiliensis mexicana; T. b. (2, 8, 22, TbV cynocephala 2.663 0.337 Y Y Y ST 28, 29)

* Viruses circulate in several species of Myotis in the western United States. Representative species trait values are reported. † Rate estimated using available information on body mass and metabolic rates (see Materials and Methods). ‡ Climatic regions abbreviated as follows: TR = tropical; ST = subtropical; TE = temperate.

151

Table B3. Phylogenetic signal in bat life history traits

Blomberg’s K* Pagel’s λ† Trait K P λ P Basal metabolic rate 0.9 <0.001 1 <0.00001 Torpid metabolic rate 1 <0.001 1 <0.00001 Range of years sampled 0.011 0.720 1.00E-07 1 No. sequences 0.004 0.996 1.00E-07 1 Climate‡ ND ND 5.00E-05 1 Colony ND ND 1 0.005 Seasonal activity ND ND 1 0.016 Migration ND ND 1 0.051

ND = not done. * Values of Blomberg’s K were estimated only for continuous traits with significance tested by 5000 randomizations of trait values on the phylogenetic tree of bats from COI sequences (Figure B1). † Maximum likelihood (ML) values of λ were estimated using the Geiger package of R with significance determined by likelihood ratio tests comparing models assuming the ML estimate of lambda to models assuming no phylogenetic signal (λ = 0) with 1 degree of freedom. ‡ When only two categories of climatic zone were included (tropical and subtropical versus temperate) the ML value of λ was 1; however, this estimate was only marginally significantly better than the model assuming λ = 0 (P = 0.06).

Table B4. Phylogenetic least squares regression support for the confidence set of models found in the generalized linear model analysis.

Predictors* AICc Δ AICc w† λ‡ Climate 28.308 0.000 0.314 0 Climate + migration 30.283 1.975 0.117 0 Climate + nyrs 30.200 1.892 0.122 0 Climate + log(BMR) 30.347 2.039 0.113 0 Climate + seasonal 30.281 1.973 0.117 0 Climate + log(TMR) 31.233 2.925 0.073 0 Climate + log(n) 31.252 2.944 0.072 0 Climate + colony 31.252 2.944 0.072 0

* Shown is the confidence set of models from the GLM analysis (Table 3.1). Abbreviated terms are defined as follows: BMR = mass-independent basal metabolic rate; TMR = mass- independent torpid metabolic rate; n = number of sequences per lineage; nyrs = range of years spanned per lineage. † AIC weights (w) describe the relative likelihood for each model given the models considered. ‡ λ, a statistic of phylogenetic signal varying from 0 (phylogenetic independence) to 1 (covariance of trait and phylogeny expected under Brownian motion model of evolution), is estimated using the viral phylogenetic tree (topology from Figure 2a) and observed trait data.

152

LITERATURE CITED

1. Streicker DG, et al. (2010) Host phylogeny constrains cross-species emergence and establishment of rabies virus in bats. Science 329(5992):676-679.

2. Larkin MA, et al. (2007) Clustal W and clustal X version 2.0. Bioinformatics 23(21):2947-2948.

3. Fleming TH & Eby P (2003) Ecology of Bat Migration. Bat Ecology, ed Kunz TH (University of Chicago Press, Chicago), pp 156-208.

4. Jones KE, et al. (2009) PanTHERIA: a species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology 90(9):2648-2648.

5. Speakman JR & Thomas DW (2003) Physiological Ecology and Energetics of Bats. Bat Ecology, eds Kunz TH & Fenton MB (The University of Chicago Press, Chicago), pp 430-490.

6. Geiser F & Stawski C (2011) Hibernation and Torpor in Tropical and Subtropical Bats in Relation to Energetics, Extinctions, and the Evolution of Endothermy. Integr Comp Biol.

7. Boyles JG, Dunbar MB, & Whitaker J, J.O. (2006) Activity following arousal in winter in North American vespertilionid bats. Mamm Rev 36(4):267-280.

8. Dunbar MB & Brigham RM (2010) Thermoregulatory variation among populations of bats along a latitudinal gradient. Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology:1-9.

9. Kalnay E, et al. (1996) The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society 77(3):437-471.

LITERATURE CITED IN TABLE 2

1. Borisenko AV, Lim BK, Ivanova NV, Hanner RH, & Hebert PDN (2008) DNA barcoding in surveys of small mammal communities: a field study in Suriname. Molecular Ecology Resources 8(3):471-479.

2. Jones KE, et al. (2009) PanTHERIA: a species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology 90(9):2648-2648.

3. Greenhall AM, Joermann G, & Schmidt U (1983) Desmodus rotundus. in Mammalian Species (The American Society of Mammalogists), pp 1-6.

4. Mies R, Kurta A, & King DG (1996) Eptesicus furinalis. Mammalian Species (526):1-7.

153

5. Myers P (1977) Patterns of reproduction of four species of vespertilionid bats in Paraguay (University of California Press).

6. Boyles JG, Dunbar MB, & Whitaker J, J.O. (2006) Activity following arousal in winter in North American vespertilionid bats. Mamm Rev 36(4):267-280.

7. Dunbar MB & Brigham RM (2010) Thermoregulatory variation among populations of bats along a latitudinal gradient. Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology:1-9.

8. Streicker DG, et al. (2010) Host phylogeny constrains cross-species emergence and establishment of rabies virus in bats. Science 329(5992):676-679.

9. Webster WD, Jones Jr JK, & Baker RJ (1980) Lasiurus intermedius. Mammalian Species:1-3.

10. Humphrey SR (1975) Nursery roosts and community diversity of Nearctic bats. J Mammal:321-346.

11. Kunz TH (1982) Lasionycteris noctavagans. Mammalian Species 172:1-5.

12. Vonhof MJ & Barclay RMR (1996) Roost site selection and roosting ecology of forest dwelling bats in southern British Columbia. Canadian Journal Of Zoology-Revue Canadienne De Zoologie 74(10):1797-1805.

13. Dunbar MB (2007) Thermal energetics of torpid silver-haired bats Lasionycteris noctivagans. Acta Theriol 52(1):65-68.

14. Hein CD, Castleberry SB, & Miller KV (2008) Sex-specific summer roost-site selection by seminole bats in response to landscape-level forest management. J Mammal 89(4):964-972.

15. Wilkins KT (1987) Lasiurus seminolus. Mammalian Species 280:1-5.

16. O'Farrell MJ, Williams JA, & Lund B (2004) Western Yellow Bat (Lasiurus xanthinus) in Southern Nevada. The Southwestern Naturalist 49(4):514-518.

17. Higginbotham JL, Dixon MT, & Ammerman LK (2000) Yucca provides roost for Lasiurus xanthinus (Chiroptera: Vespertilionidae) in Texas. The Southwestern Naturalist 45(3):338-340.

18. Carter TC, Menzel JM, Lacki M, Hayes J, & Kurta A (2007) Behavior and day-roosting ecology of North American foliage-roosting bats. Bats in forests: conservation and management:61-81.

154

19. Bisson IA, Safi K, & Holland RA (2009) Evidence for repeated independent evolution of migration in the largest family of bats. PLoS One 4(10):e7504.

20. Wilson DE & LaVal RK (1974) Myotis nigricans. Mammalian Species (39):1-3.

21. Dalquest WW (1947) Notes on the natural history of the bat, Myotis yumanensis, in California, with a description of a new race. Am Midl Nat 38(1):224-247.

22. Geluso K (2007) Winter activity of bats over water and along flyways in New Mexico. The Southwestern Naturalist 52(4):482-492.

23. Speakman JR & Thomas DW (2003) Physiological Ecology and Energetics of Bats. Bat Ecology, eds Kunz TH & Fenton MB (The University of Chicago Press, Chicago), pp 430-490.

24. Avila-Flores R, Flores-MartÌnez JJ, & Ortega J (2002) Nyctinomops laticaudatus. Mammalian Species:1-6.

25. Clare EL, Lim BK, Engstrom MD, Eger JL, & Hebert PDN (2007) DNA barcoding of Neotropical bats: species identification and discovery within Guyana. Mol Ecol Notes 7(2):184-190.

26. Cross SP (1965) Roosting habits of Pipistrellus hesperus. J Mammal:270-279.

27. Fujita MS & Kunz TH (1984) Pipistrellus subflavus. Mammalian Species 228:1-6.

28. Wilkins KT (1989) Tadarida brasiliensis. Mammalian Species (331):1-10.

29. Herreid C & Schmidt-Nielsen K (1966) Oxygen consumption, temperature, and water loss in bats from different environments. American Journal of Physiology--Legacy Content 211(5):1108.

30. Clare EL, Lim BK, Fenton MB, & Hebert PDN (2011) Neotropical Bats: Estimating Species Diversity with DNA Barcodes. PLoS One 6(7):e22648.

31. Romano MC & Maidagan JI (1999) Behavior and demography in an urban colony of Tadarida brasiliensis (Chiroptera: Molossidae) in Rosario, Argentina. Rev Biol Trop 47(4):1121-1127.

155

APPENDIX C

SUPPORTING INFORMATION FOR CHAPTER 4

FIGURES AND TABLES

Branch Substitutions (N) Substitutions (G) Substitutions (L) 3 N389H G1620N 58 TbSA 54 4 G1620S 1.0 57 DrV 6 I357V, H373P, 55 1.0 G448D, L493H 56 TbV 8 G331V*, S34I V222A*, L222V, 53 MmV P1742Q*, L1743Y* 49 0.34 10 F643N 52 NlV 50 1.0 12 P373H D1012A* 51 HmV 13 I2093L 48 PhV 14 T107A*, T107F*, 46 0.67 S107T, L108V* S34T, D448G* 47 MaV 44 0.93 15 D4E*,L108V* L2093I 31 0.51 42 45 M2V 20 D448E, H493R 1.0 43 ApV 22 N389T, L493H 1.0 36 1.0 23 K1019E* 41 Ef3V 39 0.99 24 G331V, I357V T204A, K1019R 40 Ef2V 37 1.0 26 32 1.0 S1620G 25 38 NhV 0.96 27 D4N H139N, L493P L222E, I2093V 35 MySAV 29 T204S, G1620C 33 1.0 34 EfSAV 30 G68E*, H139S T204V 34 N389H 30 EfV1a 28 1.0 35 I357V 29 EfV1b 26 0.71 39 N389H, S515Y 27 M1V 40 N160I* A96P*, S637L*, 24 LiV2 V638L*, K1019Y*, 1 0.85 G1620V*, G1620S*, 22 1.0 23 LiVa I2093V 41 I970V*, G1620V, 21 LeSAV L493P* D1840E, N1319T* 4 0.95 20 LnV2 44 I357V, N389S G1620S, D1840E 18 0.67 45 D4G*, S437W* 19 PsV V357I*, L493P I970Y*, D1012K* 5 1.0 16 1.0 47 A888P*, E921Q*, 17 LnV T264A N1319G 15 Lb1V 53 I357V, L493F 2 6 0.95 13 1.0 54 L493S 0.92 14 LcV 11 55 G1620C, V2090R 0.61 9 12 LbV2 56 S493L* G236S 0.41 7 10 LxV 57 N61E* T264H*, H373P 1.0 58 S637A 8 LsV Other H373N*, V357I, N236G, V638L*, 3 PtV 10 S515L N1319S* years

Figure C1. Map of amino acid substitutions along the evolutionary history of bat rabies virus. Shown is the maximum clade credibility tree from the hierarchical phylogenetic model using N, G and L gene data, with lineages pruned to a single representative branch. Branch

156

widths are proportional to the number of substitutions at positively selected sites across all genes; grey indicates no observed substitutions. Circled numbers are branch numbers referring to the adjacent table. Node labels are posterior probabilities (PP) of branch partitions (all terminal clades had PP ≥ 0.99). Substitutions at positively selected sites are given for each branch of the phylogenetic tree, using standard amino acid abbreviations and codon positions relative to the coding region of each gene (e.g., in row 1, N389H indicates a change from asparagine to histidine at codon 389 of the glycoprotein). Asterisks indicate substitutions that occurred on the tips of the viral phylogeny. Substitutions that occurred on branches that were supported by individual gene trees, but not in the hierarchical analysis are listed as “other”.

A. Nucleoprotein B. Glycoprotein C. Polymerase

58 TbSA 58 TbSA 58 TbSA 54 54 54 57 DrV 57 DrV 57 DrV 55 55 55 56 TbV 56 TbV 56 TbV

53 MmV 53 MmV 53 MmV 49 49 49 52 NlV 52 NlV 52 NlV 50 50 50 51 HmV 51 HmV 51 HmV

48 PhV 48 PhV 48 PhV 46 46 46 47 MaV 47 MaV 47 MaV 44 44 44 31 31 31 42 45 M2V 42 45 M2V 42 45 M2V 43 ApV 43 ApV 43 ApV 36 36 36 41 Ef3V 41 Ef3V 41 Ef3V 39 39 39 40 Ef2V 40 Ef2V 40 Ef2V 37 37 37 32 32 32 25 38 NhV 25 38 NhV 25 38 NhV

35 MySAV 35 MySAV 35 MySAV 33 33 33 34 EfSAV 34 EfSAV 34 EfSAV

30 EfV1a 30 EfV1a 30 EfV1a 28 28 28 29 EfV1b 29 EfV1b 29 EfV1b 26 26 26 27 M1V 27 M1V 27 M1V

24 LiV2 24 LiV2 24 LiV2 1 22 1 22 1 22 23 LiVa 23 LiVa 23 LiVa

21 LeSAV 21 LeSAV 21 LeSAV 4 4 4 20 LnV2 20 LnV2 20 LnV2 18 18 18 19 PsV 19 PsV 19 PsV 5 16 5 16 5 16 17 LnV 17 LnV 17 LnV

15 Lb1V 15 Lb1V 15 Lb1V 2 6 13 2 6 13 2 6 13 14 LcV 14 LcV 14 LcV 11 11 11 9 12 LbV2 9 12 LbV2 9 12 LbV2 7 10 LxV 7 10 LxV 7 10 LxV 8 LsV 8 LsV 8 LsV

3 PtV 3 PtV 3 PtV

Figure C2. Frequency of amino acid substitutions by gene along the evolutionary history of bat rabies virus. The phylogenetic tree is derived from the hierarchical phylogenetic model using N, G and L gene data, with tips pruned to a single representative per viral lineage. Branch widths are proportional to the total number of substitutions observed on each branch for each gene: nucleoprotein (A), glycoprotein (B) and polymerase (C). Grey branches indicate no non- synonymous substitutions. Dashed branches indicate lack of data for that gene. See Figure C1 for the complete details of substitutions along each branch.

157

Table C1. Oligonucleotide primers used for amplification and sequencing of the rabies virus glycoprotein and polymerase genes.

Target Primer ID gene Sense Sequence * Position † LAS Mf G Forward ATGGGCTTGACTCCAACCTTCGG 3,158 – 3,180 750G G Forward CAYGACTTTCRCTCRGATGA 4,154 – 4,173 989G G Reverse GTRAGYTTYAGACGTCTMAG 4,259 – 4,278 A2R G Reverse ATTRRCAACACTTCTCMTCCT 5,379 – 5,399 EGLF L Forward CATCCCGATAARRTGTGCWTARCT 5,295 – 5,318 DSL A2F L Forward ATTRRCAACACTTCTCMTCCT 5,379 – 5,399 L1460 L Reverse GATGAYAAATCACAYTCTTTCAC 6,849 – 6,871 L1085 L Forward TCGACAACATACAYGAYTTRGT 6,478 – 6,499 L1918 L Reverse GAGGATGTATTYTCTGTVCT 7,308 – 7,327 L1790 L Forward GATTGACAGAGTTACTGGACAGG 7,202 – 7,224 L1939 L Forward GARTCRACMGAGGATGTATT 7,299 – 7,318 RV C3 L Forward GACTGTVGCVCAACACTCTCA 7,919 – 7,939 RV C4 L Forward ATGTCAGTVCAAGCYGTYTT 7,974 – 7,993 L3206 L Reverse GAACRATAMGAAGRCAGTTT 8,566 – 8,585 L3500 L Reverse ATGTCRACCCAGCTATTCCA 8,928 – 8,947 L3000 L Forward ATCCTSTTGTCYAAGACCCATAG 8,421 – 8,443 L4000 L Reverse ATATCYAGAATGGTYTCTGG 9,444 – 9,463 DSL E3F L Forward CTCCTCTCTCAYATCTCTGTCAG 9,192 – 9,214 L9320 L Forward CGACTGAAAGACTCTACGTTTCATTGG 9,327 – 9,353 DSL E3R L Reverse GCAATCATGAGYGGRGGAGA 10,650 – 10,669 L5680 L Reverse GAGCCTTGTRTTATTCAAYTGTAGC 11,093 – 11,117 L5060 L Forward CCTATTCTDGATGATCTCAATGT 10,476 – 10,498 DSL F2F L Forward GCAATCATGAGYGGRGGAGA 10,650 – 10,669 Lysend L Reverse TTTWTTTGTTAAGCGT 11,914 – 11,929

* All sequences given in 5’ to 3’ orientation † Positions relative to the complete genome sequence of the SAD B19 vaccine strain (GenBank accession no. M31046)

158

Table C2. Putative host shifts identified in the Bayesian ancestral state estimation analysis

Other possible donors Lineage Host species Most probable donor BF (BF > 3) ApV Antrozous pallidus MaV 4.40 PhV, MV2, NhV DrV Desmodus rotundus TbV 10.32 TbSAV EfV1a Eptesicus fuscus EfV1b 19.85 EfV1b Eptesicus fuscus EfV1a 10.80 EfV2 Eptesicus fuscus EfV3 7.25 NhV EfV3 Eptesicus fuscus EfV2 7.33 NhV EfSA Eptesicus furinalis MySAV 5.81 HmV Histiotus montanus NlV 7.46 LbV2 Lasiurus borealis LbV1 7.48 LcV, LsV, LxV LbV1 Lasiurus borealis LcV 9.41 LbV2, LsV, LxV LcV Lasiurus cinereus LbV1 11.67 LbV2, LxV, LsV LeSAV Lasiurus sp. LnV1 3.36 LnV2 LiV1 Lasiurus intermedius LiV2 10.60 LiV2 Lasiurus intermedius LiV1 12.07 Lasionycteris LnV1 noctivagans LnV2 10.57 PsV, LeSAV Lasionycteris LnV2 noctivagans PsV 12.03 LnV1, LeSAV LsV Lasiurus seminolus LbV1 7.29 LxV, LbV2, LcV LxV Lasiurus xanthinus LsV 6.95 LbV2, LbV1, LcV MV1 Myotis spp. MmV 3.08 MV2 Myotis spp. MaV 6.08 PhV, ApV MaV Myotis austroriparious PhV 7.69 MV2, ApV MySAV Myotis nigricans EfSAV 5.12 NhV Nycticeius humeralis EfV3 5.07 EfV2 Nyctinomops NlV laticaudatus HmV 9.78 PhV Parastrellus hesperus MaV 6.29 MV2, ApV PsV Perimyotis subflavus LnV2 15.77 LnV1 TbV Tadarida brasiliensis DrV 13.39 TbSAV TbSAV Tadarida brasiliensis DrV 3.80 TbV

159