The Florida Panther (Puma Concolor Coryi) and the Arabian Oryx (Oryx Leucoryx) and the ARABIAN ORYX (Oryx Leucoryx)

Home , Al Areen Wildlife Park, Spotted linsang

Population Genomics/Genetics of Endangered and Vulnerable Wildlife: The Florida Panther (Puma concolor coryi) and the Arabian oryx (Oryx leucoryx) AND THE ARABIAN ORYX (Oryx leucoryx)

Item Type text; Electronic Dissertation

Authors Ochoa, Alexander

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 30/09/2021 10:55:08

Link to Item http://hdl.handle.net/10150/624483

POPULATION GENOMICS/GENETICS OF ENDANGERED AND VULNERABLE WILDLIFE: THE FLORIDA PANTHER (Puma concolor coryi) AND THE ARABIAN ORYX (Oryx leucoryx)

Alexander Ochoa

A Dissertation Submitted to the Faculty of the

SCHOOL OF NATURAL RESOURCES AND THE ENVIRONMENT

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY WITH A MAJOR IN NATURAL RESOURCES

In the Graduate College

THE UNIVERSITY OF ARIZONA

2017 1

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Alexander Ochoa, titled Population Genomics/Genetics of Endangered and Vulnerable Wildlife: The Florida Panther (Puma concolor coryi) and the Arabian Oryx (Oryx leucoryx) and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy.

______Date: February 13, 2017 Melanie Culver

______Date: February 13, 2017 John Koprowski

______Date: February 13, 2017 Katrina Dlugosch

______Date: February 13, 2017 David Onorato

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: February 13, 2017 Dissertation Director: Melanie Culver 2

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of the requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that an accurate acknowledgment of the source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.

SIGNED: Alexander Ochoa

TABLE OF CONTENTS

ABSTRACT …………………………………………………………………………… 5

APPENDIX A: The puma genome: insights into the evolution and demographic history of Florida panthers and Texas pumas ………………………. 6 TABLES……………………………………………………………………….. 20 FIGURES……………………………………………………………………… 25

APPENDIX B: Evolutionary and functional mitogenomics associated with the genetic restoration of the Florida panther ……………………….. 36 TABLES……………………………………………………………………….. 51 FIGURES…………………………………………………………………...... 52 SUPPLEMENTARY MATERIAL………………………………………...... 55

APPENDIX C: Can captive populations function as sources of genetic variation for reintroductions into the wild? A case study of the Arabian oryx from the Phoenix Zoo and the Shaumari Wildlife Reserve, Jordan ………………………………………………….. 62 TABLES……………………………………………………………………….. 83 FIGURES……………………………………………………………………… 86 SUPPLEMENTARY MATERIAL………………………………………...... 89

REFERENCES……………………………………………………………………….. 90

ABSTRACT

The most straightforward contribution of genetics to conservation has been the use of neutral markers, such as microsatellite loci, to estimate evolutionary and demographic processes (e.g. loss of genetic diversity, increase in genetic structure, population bottlenecks, migration rates) occurring in endangered and vulnerable wildlife. To this extent, next-generation sequencing technologies have enabled scientists to examine thousands or even millions of neutral loci with relative ease and at low cost in non-model species. Unlike classic population genetics studies, genomics has emerged as a reliable field to study selection and the distribution of loci affecting fitness across the entire genome. In this study, I used genomics and classic population genetics approaches to understand and/or to predict the evolutionary consequences of introductions and reintroductions in wildlife, with implications for its management and conservation. First, I employed whole-genome sequences from Florida panthers and Texas pumas to assemble and annotate the genome of the puma. In this regard, I detected genes under positive selection that could be associated with inbreeding depression traits observed in the Florida panther (e.g. heart failure, cryptorchidism, spermatozoal defects, low testosterone levels, immune deficiencies).

Second, I examined the evolutionary consequences of multiple introduction events in Florida panther mitochondrial genetic diversity. Finally, I used mitochondrial DNA and composite microsatellite data in order to develop a management strategy to reintroduce captive Arabian oryx into the wild.

APPENDIX A

This draft manuscript is distributed solely for purposes of scientific peer review. Its content is deliberative and predecisional, so it must not be disclosed or released by reviewers. Because the manuscript has not yet been approved for publication by the U.S. Geological Survey (USGS), it does not represent any official USGS finding or policy.

The puma genome: insights into the evolution and demographic history of Florida panthers and Texas pumas

Alexander Ochoa, David P. Onorato, Robert R. Fitak, Melody E. Roelke-Parker, and Melanie

Culver

From the School of Natural Resources and the Environment, University of Arizona, Tucson, AZ

85721 (Ochoa and Culver); Fish and Wildlife Research Institute, Florida Fish and Wildlife

Conservation Commission, Naples, FL 34114 (Onorato); Department of Biology, Duke

University, Durham, NC 27708 (Fitak); Leidos Biomedical Research, Inc., Frederick National

Laboratory of Cancer Research, Bethesda, MD 20892 (Roelke-Parker); and U.S. Geological

Survey, Arizona Cooperative Fish and Wildlife Research Unit, Tucson, AZ 85721 (Culver).

Address correspondence to Alexander Ochoa at the address above, or e-mail: [email protected].

Abstract

The puma (Puma concolor) is the terrestrial mammal with the largest geographic range in the

Americas. The Florida panther (Puma concolor coryi) represents the only puma population found east of the Mississippi River, specifically in South Florida, USA. As a consequence of the genetic introgression of female pumas from Texas in the mid-1990s, traits associated with inbreeding depression in the Florida panther (e.g. heart failure, cryptorchidism, spermatozoal defects, low testosterone levels, immune deficiencies) were reversed. In this study, we used whole-genome sequences from most parental lines of the admixed Florida panthers that resulted from the introduction of Texas pumas into South Florida in order to understand evolutionary and demographic processes occurring in the puma. We assembled a composite genome consisting of

~2.6 Gb. Repetitive regions accounted for 29.2% of the puma genetic background. Additionally, we identified 3,931 RNA loci and annotated 22,745 protein-coding genes, of which ~1,600 could be associated with the aforementioned inbreeding depression traits observed in the Florida panther. Some expanded gene families are associated with hippocampus development, embryo and multicellular organism development, and determination of adult lifespan. Many gene families that lost paralogs are associated with the sensory perception of smell. We characterized

> 30 positively selected genes that are related with visual perception. These results suggest that pumas have developed an acute vision at the expense of losing olfactory capabilities.

Demographic simulations are consistent with a recent colonization event in North America by a small number of founders from South America during the last glacial period.

Keywords: positive selection, inbreeding depression, next-generation sequencing, Puma concolor, founder effect. 7

Introduction

Next-generation sequencing (NGS) technologies have become an important and necessary tool for modern day research within the biological sciences. Although evolutionary biology literature has traditionally been dominated by model species (e.g. mouse, chicken, fruit fly, yeast) for which complete and well-annotated genomes are available, such technologies can be properly adapted to non-model species to reveal the underlying evolutionary forces shaping their genetic diversity. For example, restriction site associated DNA (RAD-Seq) markers and single nucleotide polymorphism (SNP) loci have been used extensively to examine the genetic structure and demographic history in Metazoans (Amato et al. 2009; Goodwin et al. 2016).

Perhaps one of the most interesting aspects of applying genomics tools to non-model species, particularly those of management and conservation concern, is the identification of genes responding to selection―and their impact on fitness―for future monitoring. In this regard, whole-genome and trancriptome sequencing technologies have become the platform of choice in determining the causal variants associated with long-term survival and inbreeding depression traits in endangered wildlife (DeSalle and Amato 2004; Allendorf et al. 2010). However, such inquiries may require properly assembled and annotated reference genomes beforehand.

The puma (Puma concolor), also known as mountain lion, cougar, or panther, occurs in a broad range of habitats such as deserts, coniferous and tropical forests, and grasslands from

British Columbia, Canada, to the Strait of Magellan, Chile (Sunquist and Sunquist 2002). Hence, it is the terrestrial mammal with the largest geographic range in the Western Hemisphere. Pumas share a recent common ancestor with jaguarundis (P. yagouaroundi) and cheetahs (Acinonyx jubatus), which dates back to the Early Pliocene (Johnson et al. 2006; Ochoa et al. 2017). During

the Late Pliocene pumas colonized South America from the north as a result of the geological joining of the Panamanian land bridge (Webb and Rancy 1996). Reduced mtDNA and nuclear genetic variation in current North American pumas indicate that the latter derive from a recent colonization event by a small number of founders from South America ca. 60,000−10,000 years before present (Culver et al. 2000; Ochoa et al. 2017).

The Florida panther (Puma concolor coryi) represents the only puma population found east of the Mississippi River, USA, having historically ranged from western Louisiana to the tip of peninsular Florida (Nowell and Jackson 1996; Onorato et al. 2010). However, by the mid-

1900s the Florida panther population was heavily reduced as a consequence of habitat loss and unregulated hunting, persisting only in patches of cypress forests, thicket swamps, saw palmetto woodlands, and pine flatwoods south of the Caloosahatchee River, Florida (McBride et al. 2008;

Onorato et al. 2010). Florida panthers were listed as endangered in 1967 by the U.S. Federal

Government and were granted protection under the U.S. Endangered Species Act in 1973

(Federal Register 1967; U.S. Fish and Wildlife Service 1973).

By the early 1990s, inbreeding and decreased levels of genetic variation within the small, remnant Florida panther population of < 30 adults likely resulted in several diagnostic phenotypic traits including kinked tails, dorsal cowlicks, atrial septal defects (i.e. the atria of the heart fail to close properly allowing oxygen-rich blood to leak into the oxygen-poor chambers), cryptorchidism (i.e. 90% of males born during this period had one or two undescended testicles), spermatozoal defects, low testosterone levels, and immune deficiencies (Roelke et al. 1993;

Barone et al. 1994; Cunningham et al. 1999; Hostetler et al. 2010, 2013; Johnson et al.2010;

Benson et al. 2011). In 1995, eight female pumas from Texas were introduced into habitat

occupied by Florida panthers as part of a genetic restoration program implemented to reverse trends associated with inbreeding depression (Seal 1994; Onorato et al. 2010). The breeding of five of these females produced a minimum of 20 offspring (Land et al. 2004; Johnson et al.

2010) and helped propel the increase in the Florida panther population size to 120‒230 individuals by 2017 (www.floridapanthernet.org).

The genetic introgression of Texas pumas contributed to reduce the occurrence of kinked tails, dorsal cowlicks, atrial septal defects, and cryptorchidism in the Florida panther population

(Land et al. 2004; Onorato et al. 2010). Even though neutral genetic diversity has doubled and fitness—survival of adults, sub-adults, and kittens—has increased in Florida panthers (Hostetler et al. 2010; Johnson et al. 2010; Benson et al. 2011), the impact of the Texas introgression event on their functional or coding genetic diversity remains undetermined.

In this study, we used whole-genome sequences from Florida panthers and Texas pumas to understand evolutionary and demographic processes occurring in the puma. We specifically characterized genes that could be associated with the inbreeding depression traits observed in the

Florida panther and genes that may have conferred selective advantages to the puma.

Materials and Methods

We collected 1 ml of whole blood from four male Florida panthers (FP16, FP45, FP60, and

FP79) and from the five female Texas pumas (TX101, TX105, TX106, TX107, and TX108) that bred with these males in the mid-1990s (Johnson et al. 2010; Ochoa et al. 2017). We also collected 1 ml of whole blood from a female Florida panther, FP73, which resulted from the

breeding between TX101 and an unsampled male Florida panther, CM7 (Johnson et al. 2010;

Ochoa et al. 2017).

We isolated genomic DNA from these samples using a phenol-chloroform extraction protocol and prepared paired-end libraries of 500-bp inserts from each sample and a mate-pair library of 5-kb inserts from sample FP16. We sequenced each DNA library on independent lanes on an Illumina HiSeq 2000/2500 system (Illumina Inc., San Diego, California) at the University of Arizona Genetics Core (UAGC, Tucson, Arizona).

We used Trimmomatic v0.35 (Bolger et al. 2014) to remove the adapters from the reads, eliminate leading and trailing bases with a Phred-scaled score of < 20, remove leading and trailing ‘N’ bases, scan the reads with a 4-base wide sliding window and cut them when the average Phred-scaled quality per base was < 20, and eliminate reads with an average Phred- scaled score of < 30 or with a length of < 50 bases. To improve data quality, we used Musket v1.1 (Liu et al. 2013) to count the number of occurrences of all non-unique k-mers (k = 21 and

23) from each paired-end library, estimate the coverage cutoff to differentiate spurious from authentic k-mers, correct substitution sequencing errors in the spurious k-mers, and trim ≤ 70 bases per read while keeping the longest error-free substring after error correction.

We produced multiple de novo genome assemblies with the validated and corrected paired-end reads in ABySS v1.3.6 (Simpson et al. 2009) using k-mer lengths of k = 45, 51, 55,

59, 61, 63, 65, 69, and 75 bases for each assembly. We used the trimmed mate-pair reads to perform the scaffolding of the resulting contigs while retaining only the scaffolds > 500 bases.

Using CEGMA v2.4 (Parra et al. 2007), we predicted the exon-intron structure of each assembly based on the mapping of a predefined set of mammalian core eukaryotic genes (CEGs).

Ultimately, we selected the assembly with roughly the greatest proportion of predicted CEGs, greatest N50, and least number of scaffolds altogether for downstream analyses.

We used RepeatMasker v4.0.7 (Smit et al. 2013-2015) to characterize repetitive elements in the selected genome assembly. We also identified structural RNAs with INFERNAL v1.1.2

(Nawrocki and Eddy 2013). In this regard, we conducted structure-based homology searches against the Rfam database (Griffiths-Jones et al. 2003; Nawrocki et al. 2015) and used a

‘gathering’ cut-off score of 85% for the covariance models and an e-value threshold of 10−9.

We performed gene annotations with MAKER v2.31.6 (Cantarel et al. 2008) using a two- step iterative procedure. The first iteration consisted of the annotation of genes with SNAP (Korf

2004) using the CEGs identified from CEGMA, ab initio predictions of genes from

GENEMARK-ES (Lomsadze et al. 2005), transcriptome sequences from the puma (Fitak et al.

2015a), and protein sequences from the cheetah, cat, tiger, and dog downloaded from

ENSEMBL (www.ensembl.org). The second iteration was similar to the first one, but for the former we performed the ab initio predictions of genes with AUGUSTUS v2.5.5 (Stanke et al.

2006). We then removed the gene predictions with an exon annotation edit distance (eAED) <

0.75.

We used the longest isoform from each predicted gene to functionally annotate them with

BLASTP v2.2.30 (http://ncbi.nlm.nih.gov/blast) and INTERPROSCAN v5.7.48 (Jones et al.

2014). We performed the BLASTP searches against the metazoan ‘nr’ database using an e-value threshold of 10−3 and retained the first 20 top hits from each search. We performed the

INTERPROSCAN searches to assign protein domains to the predicted genes using a variety of databases (i.e. GENE3D, COILS, HAMAP, PANTHER, PFAMA, PIRSF, PRINTS, PRODOM,

PROSITEPATTERNS, PROSITEPROFILES, SMART, SUPERFAMILY, TIGRFAM). We merged the BLASTP and INTERPROSCAN searches with Blast2GO v4.0.7 (Conesa et al. 2005) to obtain the Gene Ontology (GO) terms of each sequence.

To explore gene family expansions and contractions, we used OrthoMCL v1.0 (Li et al.

2003) to identify orthologs across the puma, cat, dog, panda, cow, human, and mouse. We downloaded the protein sequences of these species from the ENSEMBL database with the exception of the puma, which we had previously annotated. We clustered these sequences into protein families by performing all-against-all local alignments. We then used the DupliPHY-

Web server (http://www.bioinf.manchester.ac.uk/dupliphy/; Ames and Lovell 2014) to determine the evolutionary histories of gene family sizes based on the following phylogenetic tree:

((((puma:5.6, cat:5.6):50.6, (dog:46.6, panda:46.6):9.6):24.9, cow:81.1):13.9, (human:90.1, mouse:90.1):4.9). We determined the divergence times (scaled to million years before present) of this phylogenetic tree from www.timetree.org and Ochoa et al. (2017). For this analysis we excluded gene families absent in either the puma, cat, dog, panda, and cow clade or the mouse and human clade.

We used the nuclear (i.e. non-mitochondrial) single-copy orthologs present in the puma, cat, dog, panda, cow, human, and mouse to detect genes under positive selection in the puma.

First, we aligned each single-copy ortholog across species with PRANK v131204 (Löytynoja et al. 2005). Second, we converted each alignment of amino acid sequences into an alignment of codon sequences using PAL2NAL v14 (Suyama et al. 2006). Third, we eliminated poorly aligned codons with Gblocks v0.91b (Castresana 2000; Talavera and Castresana 2007). We then conducted likelihood ratio tests with PAML v4.8a (Yang 1997, 2007) by comparing the M1a

branch model, in which lineages evolve neutrally, against the M2a branch model, in which we assumed that the puma lineage evolved under positive selection. We computed the P-values from each likelihood ratio test using the χ2-statistic adjusted by the false discovery rate method.

We mapped the error-correct paired-end reads from each puma sample against the de novo genome assembly using BWA v0.7.9 (Li and Durbin 2009). We used SAMtools v0.1.19 (Li et al. 2009; Li 2011) to define variants―single nucleotide polymorphisms (SNPs) and indels―across the puma samples after removing potential PCR duplicates, eliminating reads with a mapping quality of ≤ 20, and removing variants with a Phred-scaled score of < 20 for the alternate alleles. We further intersected these variants with another set of variants previously defined with Platypus v2.6 (Rimmer et al. 2014). Using VCFtools v4.2 (Danecek et al. 2011), we excluded variants (three or more) clustered in 10-bp windows, with a strand bias P < 0.0001, or with a combined depth coverage of < 30 or > 180 .

We examined historical changes in effective population sizes of the puma with PSMC v0.6.4 (Li and Durbin 2011). We performed this analysis with the variants defined from each sample (lenient conditions). Simulations consisted of 25 iterations using the –p and –t parameters, an initial theta/rho ratio (–r) of 5, and 100 bootstrap replicates for each run. We scaled the final effective population size estimates to a generation time of three years and a mutation rate of 2.5 10−8 substitutions site−1 generation−1 (Nachman and Crowell 2000).

Results and Discussion

We created independent paired-end DNA libraries of 500-bp size inserts from five Florida panthers (samples FP16, FP45, FP60, FP73, and FP79) and five Texas pumas (samples TX101,

TX105, TX106, TX107, TX108). Likewise, we created a mate-pair DNA library of 5-kb size inserts from sample FP16. Overall, NGS produced ~1.3 billion paired-end reads and ~136 million mate-pair reads (Table A1). After applying various quality control steps to the sequence data, we retained 90.9% of the paired-end reads and 44.7% of the mate-pair reads (Table A1).

For the processed paired-end reads we used k-mer sizes of k = 45, 51, 55, 59, 61, 63, 65,

69, and 75 to obtain multiple assemblies. We used the processed mate-pair reads to create scaffolds from each assembly. We found that k = 51 produced the assembly with the greatest proportion of retained CEGs (97.2%), that k = 59 produced the assembly with the greatest scaffold N50 (195,096), and that k = 65 produced the assembly with the least number of scaffolds (147,881) (Figure A1). Nevertheless, we selected the assembly based on a k-mer size of

61 bases for downstream analyses because it roughly presented the greatest proportion of retained CEGs (96.4%), the greatest scaffold N50 (193,863), and the least number of scaffolds

(152,283) altogether (Figure A1). The latter assembly had an estimated genome size of ~2.6 Gb

(GC content = 41.0%) with a cumulative sequence coverage of 74.2 (Table A2; Figure A2).

Repetitive regions accounted for 29.2% of the puma genome; these regions included

SINEs (2.6%), LINEs (16.1%), LTRs (4.4%), DNA elements (2.7%), and simple repeats (2.2%)

(Table A3). We also identified 3,931 loci representing 636 different RNA families (Table A4).

These results are consistent across different mammalian species. For example, Fitak et al.

(2015b) reported a repeat content of 33.7% and 3,691 loci distributed along 643 RNA families in the dromedary genome.

We structurally annotated 22,926 proteins (eAED < 0.75) in the puma genome (Table

A5; Figure A3). We then defined 22,745 genes, of which 177 presented multiple isoforms (Table

A5). Using the longest isoform from each predicted gene, we functionally annotated 20,561

(90.4%) proteins with Blast2GO (Figure A4). Overall, ~1,600 genes were related with heart development, male gonad and germ cell development, testosterone secretion and steroid hormone signaling, and immune response (Supplementary Table A1).

We characterized 17,131 orthologous gene families across the puma, cat, dog, panda, cow, human, and mouse genomes (Figure A5). Of these, pumas presented 540 gene family expansions and 2,613 gene family contractions with respect to the puma-cat most recent common ancestor (Figure A6). These results are similar to those reported for the cheetah (Dobrynin et al.

2015); this species presented 814 gene family expansions and 2,169 gene family contractions with respect to the cheetah-cat most recent common ancestor.

Biological processes related with the expanded gene families in the puma include hippocampus development, embryo and multicellular organism development, and determination of adult lifespan. For example, the PLXNA gene family―which acquired three paralogs―is necessary for the signaling of semaphorins and for the subsequent remodeling of the cytoskeleton; it also plays a key role in axon guidance, invasive growth, and cell migration in the hippocampus for spatial memory that enables orientation and navigation (Cheng et al. 2001; Suto et al. 2005; Ben-Zvi et al. 2008; He et al. 2009). We also found that the HSPA gene family

(involved in the metabolism of fatty acids) acquired eight paralogs. Given that pumas are unable to synthesize certain fatty acids (e.g. arachidonic acid) due to low delta-6-desaturase, Montague et al. (2014) have suggested that felids may use alternate pathways to generate these compounds for reproduction. Likewise, fatty acid metabolism is essential for an obligatory carnivorous diet

(Irizarry et al. 2012; Cho et al. 2013).

On the other hand, ~40 of the contracted gene families in the puma are associated with the sensory perception of smell. The latter are members of a larger family of G-protein-coupled olfactory receptors coded by single-exon genes (Young et al. 2002; Zhang and Firestein 2002;

Malnic et al. 2004). The puma genome lost 14 paralogs from one of these gene families alone

(OR1) since the puma-cat split.

Furthermore, we defined 8,192 single-copy gene families present in the puma, cat, dog, panda, cow, human, and mouse genomes. Most of these gene families exhibited a mean dN/dS <

1.0 across species (Figure A7). Nevertheless, we found 1,345 single-copy gene families under significant positive selection (false discovery rate < 0.05) in the puma lineage only (Figure A8).

Genes shaped by positive selection in pumas include those associated with the sensory perception of light, sound, smell, taste, and pain; behavioral response to fear and stress; fatty acid metabolism; and muscle contraction and actin filament organization and polymerization

(Supplementary Table A2). We specifically detected > 30 positively selected genes related with visual perception and learning, eye and eyelid morphogenesis, optic placode formation, and lens and retina development (Supplementary Table A2). These results are similar to those observed in the tiger and in the cat, and they suggest that felids have developed an acute vision―perhaps as a result of crepuscular activity and hunting―at the expense of losing genes associated with the perception of smell (Cho et al. 2013; Montague et al. 2014). Still, other genes in the puma that have been accumulating a significant amount of nonsynonymous mutations (e.g. SEMA7A) are involved in the development of sensory axons located in the vomeronasal organ; these axons project to the nasal mesenchyme and target the amygdala and bed nucleus of the stria terminalis, which in turn project to the hypothalamus (Pasterkamp et al. 2003, 2007; Messina et al. 2011;

Supplementary Table A2). The vomeronasal organ, which is present in felids and other mammalian species, plays an important role in behavioral traits such as territory ownership (e.g. detection of urine) and sexual activity (e.g. detection of pheromones) (Corbett 1954; Cho et al.

2013).

We defined ~15.4 million alleles distributed along ~7.7 million SNPs in the puma genome (only 144 SNPs contained three or more alleles) (Figure A9). Of these alleles, ~41.5% were contributed solely by a non-canonical Florida panther (sample FP16) and by Texas pumas

(samples TX101, TX105, TX106, TX107, and TX108) (Figure A9). Sample FP16 from the

Everglades National Park carried the greatest amount of SNPs at the genome level, perhaps because this Florida panther had Costa Rican and Panamian ancestry (i.e. non-North-American ancestry) (O’Brien et al. 1990; O’Brien and Roelke 1990; Roelke et al. 1993; Ochoa et al. 2017;

Figure A10). Furthermore, Texas pumas (samples TX101, TX105, TX106, TX107, and TX108) carried a greater amount of SNPs than ‘canonical’ Florida panthers (samples FP45 and FP60)

(Figure A10). Admixed panthers FP73 and FP79 (Texas + Florida ancestry) carried a comparable amount of SNPs than the Texas pumas (Figure A10), which suggests that most of the historical

Florida panther genetic diversity could have been restored in just one generation as a result of the introduction of Texas pumas into South Florida in 1995.

Simulations with PSMC (Figure A11) suggest that pumas reached a maximum effective population size of ~45,000−80,000 individuals during the Late Pleistocene (ca. 30,000−40,000 years before present); after this period of time, a sharp decline in the effective population size of pumas is observed. This result is consistent with a recent colonization event in North America by

a small number of founders from South America during the last glacial period (Culver at al.

2000; Ochoa et al. 2017).

Funding

This project was funded by the William A. Calder III Memorial Scholarship from the

Department of Ecology and Evolutionary Biology of the University of Arizona. Consejo

Nacional de Ciencia y Tecnología and National Science Foundation-Integrative Graduate

Education and Research Traineeship scholarships were awarded to A. Ochoa.

Acknowledgments

We thank the Florida Fish and Wildlife Conservation Commission and the National Park Service for collecting and providing the biological samples used in this study. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S.

Government.

Data Availability

Supplementary Tables A1 and A2 are available by direct request to the authors.

TABLES

Table A1. Summary statistics of the raw and processed (trimmed and error-corrected) reads used for this study.

Table A2. Summary statistics of the mapped paired-end reads to the puma genome (~2.6 Gb).

Table A3. Repetitive elements characterized in the puma genome (~2.6 Gb).

Table A4. RNA loci and families identified in the puma genome (~2.6 Gb).

Table A5. Summary statistics of the proteins and genes found across different mammalian species.

*Exon annotation edit distance (eAED) < 0.75

Mammalian genomes other than the puma were downloaded from ENSEMBL

(www.ensembl.org).

FIGURES

Figure A1. (a) Proportion of core eukaryotic genes (CEGs), (b) scaffold N50, and (c) number of scaffolds for each puma assembly based on a different k-mer size (45, 51, 55, 59, 61, 63, 65, 69, and 75 bases). 25

Figure A2. Cumulative length of the puma genome assembly. Scaffolds are sorted from longest to smallest along the horizontal axis. The vertical dotted line indicates the number of scaffolds containing 95% of the total assembly.

Figure A3. Proteins annotated in the puma genome assembly. The vertical dotted line indicates an exon annotation edit distance (eAED) of 0.75.

Figure A4. Summary statistics of the genes annotated with Blast2GO v4.0.7 (Conesa et al.

2005).

Figure A5. Venn diagram representing shared orthologs across the puma, cat, dog, panda, cow, human, and mouse.

Figure A6. Gene family expansions (green) and contractions (red) across the puma, cat, dog, panda, cow, human, and mouse.

Figure A7. Mean dN/dS values for nuclear single-copy gene families present in the puma, cat, dog, panda, cow, human, and mouse.

Figure A8. Single-copy gene families under significant positive selection (false discovery rate <

0.05; red bar) in the puma lineage only.

Figure A9. Distribution of ~15.4 million alleles across ~7.7 million SNPs in three lineages:

FL+CRI/PAN (sample FP16), FL (samples FP45 and FP60), and TX (TX101, TX105, TX106,

TX107, TX108). Circles are proportional to size.

Figure A10. Heterozygous sites characterized in each puma sample used in this study.

Figure A11. Historical changes in effective population sizes of the puma as examined for each puma sample used in this study.

APPENDIX B

Accepted Manuscript: Journal of Heredity

Evolutionary and functional mitogenomics associated with the genetic restoration of the

Florida panther

Alexander Ochoa, David P. Onorato, Robert R. Fitak, Melody E. Roelke-Parker, and Melanie

Culver

From the School of Natural Resources and the Environment, University of Arizona, Tucson, AZ

85721 (Ochoa and Culver); Fish and Wildlife Research Institute, Florida Fish and Wildlife

Conservation Commission, Naples, FL 34114 (Onorato); Department of Biology, Duke

University, Durham, NC 27708 (Fitak); Leidos Biomedical Research, Inc., Frederick National

Laboratory of Cancer Research, Bethesda, MD 20892 (Roelke-Parker); and U.S. Geological

Survey, Arizona Cooperative Fish and Wildlife Research Unit, Tucson, AZ 85721 (Culver).

Address correspondence to Alexander Ochoa at the address above, or e-mail: [email protected].

Abstract

Florida panthers are endangered pumas that currently persist in reduced patches of habitat in

South Florida, U.S. We performed mitogenome reference-based assemblies for most parental lines of the admixed Florida panthers that resulted from the introduction of female Texas pumas into South Florida in 1995. With the addition of two puma mitogenomes, we characterized 174 single nucleotide polymorphisms (SNPs) across 12 individuals. We defined five haplotypes

(Pco1‒Pco5), one of which (Pco1) had a geographic origin exclusive to Costa Rica and Panama and was possibly introduced into the Everglades National Park, Florida, prior to 1995. Haplotype

Pco2 was native to Florida. Haplotypes Pco3 and Pco4 were exclusive to Texas, whereas haplotype Pco5 had an undetermined geographic origin. Phylogenetic inference suggests that haplotypes Pco1‒Pco4 diverged ~202,000 (95% HPDI = 83,000‒345,000) years ago and that haplotypes Pco2‒Pco4 diverged ~61,000 (95% HPDI = 9,000‒127,000) years ago. These results are congruent with a south-to-north continental expansion and with a recent North American colonization by pumas. Furthermore, pumas may have migrated from Texas to Florida no earlier than ~44,000 (95% HPDI = 2,000‒98,000) years ago. Synonymous mutations presented a greater mean substitution rate than other mitochondrial functional regions: nonsynonymous mutations, tRNAs, rRNAs, and control region. Similarly, all protein-coding genes were under predominant negative selection constraints. We directly and indirectly assessed the presence of potential deleterious SNPs in the ND2 and ND5 genes in Florida panthers prior to and as a consequence of the introduction of Texas pumas. Screenings for such variants are recommended in extant Florida panthers.

Keywords: negative selection, introduction, deleterious SNP, Puma concolor, endangered, felid. 37

Introduction

Mitochondria are organelles found in most eukaryotic cells, where their primary function is the production of energy in the form of adenosine 5’-triphosphate via the citric acid cycle (Mishra and Chan 2014). Other functions of mitochondria include the storage and signaling of calcium ions and the synthesis of steroid hormones such as testosterone and estrogen (Miller 1998;

Nicholls 2005; Miller 2013). In mammals, the mitogenome consists of 13 protein-coding genes,

22 tRNA genes, two rRNA genes, the origin of replication of the H-strand along with adjacent promoters for transcription (known collectively as the control region or CR), and the origin of replication of the L-strand (Anderson et al. 1981; Wallace 1986).

Negative selection in mitochondrial genes is likely to be strong and rapid, as their transcription products are critical for life (Hill et al. 2014). For instance, the protein-coding genes are transcribed into mRNAs and eventually translated into subunits of protein complexes that participate in the oxidative phosphorylation process in the mitochondrial inner membrane (Blier et al. 2001). Nevertheless, mitochondrial protein-coding genes can still carry deleterious mutations. In humans, ~275 mutations in these genes have been associated with different diseases (http://www.mitomap.org/MITOMAP). Moreover, evolutionary forces such as genetic drift may enhance the accumulation of slightly deleterious mutations in the mitochondria, particularly in endangered species characterized by small population sizes―i.e. Muller’s ratchet

(Muller 1963; Felsenstein 1974; Lynch and Gabriel 1990; Loewe 2006).

Pumas (Puma concolor), also known in different regions as panthers, mountain lions, or cougars, have a broad distribution across a vast range of habitats from southwestern Canada to southern Chile (Sunquist and Sunquist 2002). However, their range has been severely reduced

over the past century as a result of habitat destruction and/or unregulated hunting (Laundré and

Hernández 2010; Castilho et al. 2011). In the U.S. alone, pumas were extirpated east of the

Mississippi River, except for a viable population of panthers that persisted in reduced patches of habitat in South Florida (Nowell and Jackson 1996; Onorato et al. 2010).

Florida panthers have been classified as endangered since 1967 and have been afforded protection under the U.S. Endangered Species Act since 1973 (Federal Register 1967; U.S. Fish and Wildlife Service 1973). In the early 1980s, < 30 Florida panthers were documented across their entire range (McBride et al. 2008). Long-term isolation and inbreeding in Florida panthers resulted in the expression of phenotypic traits (e.g. spermatozoal defects, cryptorchidism, deficiency in testosterone, atrial septal defects) with deleterious fitness consequences (Roelke et al. 1993; Barone et al. 1994; Cunningham et al. 1999; Hostetler et al. 2010, 2013; Johnson et al.

2010; Benson et al. 2011). Concerns about the possible extinction of Florida panthers motivated the introduction of eight female pumas from southwestern Texas into South Florida in 1995 (Seal

1994; Onorato et al. 2010). Parturition was documented in five of the introduced Texas females and, subsequently, deleterious correlates of inbreeding depression in Florida panthers were drastically reduced (Land et al. 2004; Johnson et al. 2010).

Given their high mutation rate and their non-recombinant, haploid, and uniparental (e.g. maternal) inheritance nature, mitogenomes have been used extensively in the field of molecular phylogenetics to describe evolutionary histories, resolve taxonomic uncertainties, and determine divergence times between lineages. Yet only a handful of studies has also focused on the selective processes that have shaped mitochondrial genetic diversity in natural populations. This study offers new insight into the evolution of the mitogenome in pumas; at the same time it also

addresses, directly and indirectly, the presence of potential deleterious single nucleotide polymorphisms (SNPs) in Florida panthers prior to and as a consequence of the introduction of

Texas pumas.

Materials and Methods

We obtained whole blood samples from tissue archives of the Florida Fish and Wildlife

Conservation Commission (FWC) that included puma samples collected by the FWC and

National Park Service staff. We used 1 ml of whole blood from one female (sample FP73) and four male (samples FP16, FP45, FP60, and FP79) Florida panthers captured in South Florida, and from five female pumas (samples TX101, TX105, TX106, TX107, and TX108) captured in southwestern Texas (Johnson et al. 2010; Supplementary Table B1). These Texas females constitute individuals that were documented to have produced litters of kittens with Florida panthers. There are five known paternal links between male Florida panthers and offspring of the introduced Texas pumas: FP16, FP45, FP60, FP79, and CM7 (not sampled). Individuals FP73 and FP79 are admixed Florida panthers that resulted from the mating between CM7 and TX101

(Johnson et al. 2010; Supplementary Figure B1).

We extracted genomic DNA from each sample using a phenol-chloroform protocol at the

University of Arizona Genetics Core (UAGC, Tucson, Arizona). We prepared paired-end libraries of 500-bp inserts for each DNA sample and sequenced them on independent lanes on an

Illumina HiSeq 2000/2500 system (Illumina, Inc., San Diego, California) at the UAGC. We used

Trimmomatic v0.35 (Bolger et al. 2014) to remove the adapter sequences from the paired-end reads (~100 bp  2), eliminate leading and trailing ‘N’ bases, eliminate leading and trailing

bases with a Phred-scaled score of < 20, trim reads when a 4-base wide sliding window averaged a Phred-scaled score of < 20, and remove reads with an average Phred-scaled score of < 30 or with a length of < 50 bases.

Using Musket v1.1 (Liu et al. 2013), we detected and corrected sequencing errors in the trimmed reads based on their k-mer abundances. In this regard, we assessed the abundances of each non-unique 21-mer and 23-mer, estimated the coverage cutoff value to differentiate spurious from true k-mers, and corrected substitution sequencing errors in the spurious k-mers.

We mapped the resulting paired-end and single-end reads (i.e. unmated processed reads) against a complete puma mitogenome (GenBank Acc. No. JN999997; https://www.ncbi.nlm.nih.gov/genbank/; collected from a rescued animal housed at In-Sync

Exotics in Wylie, Texas―D. McCulloch pers. comm.) with BWA v0.7.9 (Li and Durbin 2009).

We used SAMtools v0.1.19 (Li et al. 2009; Li 2011) to remove potential PCR duplicates with the lowest mapping qualities, eliminate reads with a mapping quality of ≤ 20, and define

SNPs with a Phred-scaled quality of ≥ 30 for the alternate alleles. We then performed a multiple sequence alignment (MSA) with SATé v2.2.7 (Liu et al. 2009, 2012) using the MAFFT aligner

(Katoh et al. 2002) and the MUSCLE merger (Edgar 2004). In addition to the reference-based assemblies and GenBank Acc. No. JN999997, we included the complete mitogenome of a jaguarundi (Puma yagouaroundi; GenBank Acc. No. KP202279; Li et al. 2016), cheetah

(Acinonyx jubatus; GenBank Acc. No. KP202271; Li et al. 2016), domestic cat (Felis catus;

GenBank Acc. No. U20753; Lopez et al. 1996), spotted linsang (Prionodon pardicolor;

GenBank Acc. No. KJ636050; Hassanin and Veron 2016), and another male Florida panther captured in South Florida (GenBank Acc. No. KP202261; sample FP12―W. Murphy pers.

comm.; Li et al. 2016; Supplementary Table B1) in the MSA for comparative purposes. We replaced ambiguous bases with ‘N’ bases in the MSA, defined puma mitochondrial haplotypes with PopART v1.7 (http://popart.otago.ac.nz), and created consensus sequences for each haplotype with BioEdit v7.2.5 (Hall 1999).

We estimated intra and interspecific divergence times with PAML v1.3.1 (Yang 2007) and multidivtime v09.25.03 (Thorne and Kishino 2002) using MSA partitions of each protein- coding gene and, ultimately, a concatenated MSA of the protein-coding sites (Rutschmann

2005). For each MSA, we used MEGA v6.06 (Tamura et al. 2007) to determine the best nucleotide substitution model based on the Bayesian information criterion and to create maximum-likelihood (ML) tree topologies as input data for further analyses. We obtained nucleotide frequency, transition/transversion ratio, heterogeneity rate, and α-shape parameter estimates from the ‘baseml’ function as implemented in PAML. Likewise, we used the

‘estbranches’ function in multidivtime to obtain ML estimates of branch lengths and to obtain a variance-covariance matrix. We ran five independent Markov chain Monte Carlo simulations, each consisting of 105 initial generations as burn-in and then 106 generations sampled every 100 generations. We assumed the following calibration nodes scaled to million years before present

(MYBP): puma-jaguarundi (4.2; 95% highest posterior density interval or HPDI = 3.2‒6.0), puma-cheetah (4.9; 95% HPDI = 3.9‒6.9) and puma-domestic cat (6.7; 95% HPDI = 5.3‒9.2)

(Johnson et al. 2006); and puma-spotted linsang (33.3; 95% HPDI = 28.9‒39.1) (Eizirik et al.

2010). For the phylogenetic inference we also assumed a relaxed molecular clock model. Using

Tracer v.1.6 (Rambaut et al. 2014), we assessed convergence within and between runs and summarized the posterior samples of the concatenated MSA of the protein-coding sites.

We employed the same procedure as above to calculate mean substitution rates for sets of concatenated sequences representing diverse mitochondrial functional regions: nonsynonymous and synonymous mutations along the same MSA of protein-coding genes, tRNAs, rRNAs, and

CR. In order to calculate the mean substitution rates for the nonsynonymous and synonymous mutations, we excluded the synonymous and nonsynonymous sites between the puma and the jaguarundi from the MSA of protein-coding genes, respectively. Therefore, we only considered the mean substitution rate estimates for the puma-jaguarundi split for each mitochondrial functional region.

We tested for selection in the protein-coding genes of the puma-jaguarundi with

KaKs_Calculator v2.0 (Wang et al. 2010). We estimated the Ka (i.e. rate of nonsynonymous substitutions per nonsynonymous site), Ks (i.e. rate of synonymous substitutions per synonymous site), and Ka/Ks parameters using the γ-NG, γ-LWL, γ-MLWL, γ-LPB, γ-MLPB, γ-YN, and γ-

MYN approximate methods (Wang et al. 2009a, 2009b). In general, Ka/Ks = 1 suggests neutral evolution, Ka/Ks < 1 suggests negative selection, and Ka/Ks > 1 suggests positive selection.

We predicted the impact of the amino acid mutations (i.e. nonsynonymous SNPs) in the puma using the PROVEAN web server (http://provean.jcvi.org/index.php; Choi 2012; Choi et al.

2012; Choi and Chan 2015). This program performs BLAST searches for query proteins across the NCBI ‘nr’ database. Homologs are then clustered using CD-HIT, and PROVEAN scores (i.e. effect on fitness) are computed by averaging ‘delta alignment scores’ within and between the top

30 clusters of closely related sequences. Typically, PROVEAN scores ≤ −2.5 are considered to be deleterious, whereas PROVEAN scores > −2.5 are considered to be neutral. We inferred the

immediate ancestral amino acid sequence for each nonsynonymous SNP from the ML phylogenetic trees of protein-coding gene partitions produced in MEGA.

We further used TreeSAAP v3.2 (Woolley et al. 2003) to determine the function of protein regions carrying the predicted deleterious amino acid mutations. This program identifies the specific physicochemical properties of these regions (e.g. hydropathy, isoelectric point, polar requirement) by means of a Z-test. ‘Radical’ quantitative classes for each physicochemical property (i.e. categories 6−8) with a Z-score > 3.09 are regarded as highly significant (P <

0.001). We used the nucleotide substitution models and tree topologies from the ML inference of gene partitions produced in MEGA, and sliding windows of 20 amino acids per gene for the calculations in TreeSAAP.

Results and Discussion

In this study we produced novel mitogenome reference-based assemblies from five Florida panthers (samples FP16, FP45, FP60, FP73, and FP79) and five Texas pumas (samples TX101,

TX105, TX106, TX107, and TX108) at sequence coverages of 317‒1,433 (Supplementary

Table B2). We then defined five haplotypes across 12 puma mitogenomes: Pco1 (sample FP16),

Pco2 (samples FP12, FP45, and FP60), Pco3 (samples FP73, FP79, TX101, and TX106), Pco4

(samples TX105, TX107, and TX108), and Pco5 (GenBank Acc. No. JN999997). Collectively, these haplotypes contained 174 SNPs after we replaced ambiguous bases with ‘N’ bases in the reference-based assemblies and after we removed two repetitive sequence loci from an intraspecific MSA matrix consisting of 17,153 bp (Supplementary Figures B2 and B3).

Ambiguous bases in the reference-based assemblies were distributed along the 12S rRNA, tRNA-Val, 16S rRNA, and ND1 genes (Supplementary Table B3), and they could have resulted from the mapping of numt pseudogenes to the puma mitogenome reference sequence

(GenBank Acc. No. JN999997) or from heteroplasmic mutations (Lebon et al. 2003; Li et al.

2016). The repetitive sequence loci, known as RS2 (sites 16,531‒16,903) and RS3 (sites

274‒618), were located on opposite sides of the central conserved region of the CR and were characterized by 80-bp and 8-bp consensus sequence patterns, respectively (Lopez et al. 1996;

Jae-Heup et al. 2001). We also found 1,174‒2,450 fixed differences between the puma and the jaguarundi, cheetah, domestic cat, and spotted linsang after we removed the RS2 and RS3 loci from a combined intra and interspecific MSA matrix consisting of 17,208 bp (Supplementary

Figure B3).

Previously, Culver et al. (2000) characterized 14 mitochondrial haplotypes (designated by the letters A‒N) for pumas across the Americas based on a concatenated sequence of 891 bp from the 16S rRNA, ATP8, and ND5 genes. Comparisons between this study and that of Culver et al. (2000) revealed that haplotypes Pco2‒Pco4 have a Central and North American ancestry

(haplotype M; Supplementary Table B4). In contrast, haplotype Pco1 has a geographic origin exclusive to Costa Rica and Panama (haplotype C; Supplementary Table B4). Haplotype Pco1 possibly derived from a captive-bred stock―of Costa Rican and Panamanian ancestry at the very least―that was introduced into the Everglades National Park, Florida, in 1956‒1966 (O’Brien et al. 1990; O’Brien and Roelke 1990; Roelke et al. 1993). Haplotype Pco5 displayed two unique

SNPs with respect to haplotypes A‒N (although it presented the greatest sequence identity to

haplotype M; Supplementary Table B4). For this reason, we were unable to successfully assign haplotype Pco5 to a specific geographic location.

Phylogenetic inference suggests that haplotypes Pco1‒Pco4 (Central and North America) diverged ~202,000 (95% HPDI = 83,000‒345,000) years ago and that haplotypes Pco2‒Pco4

(Florida and Texas, specifically) diverged ~61,000 (95% HPDI = 9,000‒127,000) years ago

(Figure B1). In connection with this estimate, Culver et al. (2000) have suggested that range- wide pumas―South America included―shared a common mitochondrial ancestry ~390,000 years ago and that North American pumas shared a common mitochondrial ancestry < 20,000 years ago. Therefore, our results are also congruent with a south-to-north continental expansion and with a recent North American colonization by pumas, conceivably during the Middle and

Late Pleistocene, respectively.

Since Texas haplotype Pco3 is basal to Texas haplotype Pco4 and to Florida haplotype

Pco2 (Figure B1), we can further hypothesize that pumas migrated from Texas to Florida once initial North American colonizations had taken place. Phylogenetic inference suggests that this migration event occurred no earlier than ~44,000 (95% HPDI = 2,000‒98,000) years ago (Figure

B1). Lastly, we estimated divergence times of ~3.8 (95% HPDI = 3.0‒4.8) MYBP for the puma- jaguarundi, ~4.0 (95% HPDI = 3.2‒5.0) MYBP for the puma-cheetah, and ~5.6 (95% HPDI =

5.0‒6.8) MYBP for the puma-domestic cat (Figure B1). These results reflect more recent―but not statistically different―felid interspecific divergence time estimates than those reported by

Johnson et al. (2006) (see Materials and Methods).

Mean evolutionary rates (expressed as 10‒9 nucleotide substitutions site‒1 year‒1) for diverse mitochondrial functional regions in the puma-jaguarundi were: nonsynonymous

substitutions (k = 2.7 ± 0.5), synonymous substitutions (k = 24.5 ± 3.7), tRNAs (k = 4.4 ± 1.5), rRNAs (k = 5.5 ± 1.5), and CR (k = 13.9 ± 3.7). These estimates are similar to those reported by

Pesole et al. (1999) for closely related mammalian species with divergence times of 2.5‒6.0

MYBP. For example, Pesole et al. (1999) calculated mean mitochondrial evolutionary rates (also expressed as 10‒9 nucleotide substitutions site‒1 year‒1) for the human-bonobo as follows: nonsynonymous substitutions (k = 2.0 ± 0.3), synonymous substitutions (k = 28.0 ± 4.0), tRNAs

(k = 3.6 ± 1.0), 12S rRNA (k = 2.7 ± 0.9), 16S rRNA (k = 5.6 ± 1.4), and CR (k = 13.7 ± 3.4).

Synonymous mutations in the puma-jaguarundi presented a high substitution rate relative to the other mitochondrial functional regions. We regard such mutations as a reliable indicator of neutral expectations. Hence, the tRNAs and rRNAs could be subject to particular selective constraints, considering that the former function as carriers of specific amino acids and that the latter act as complementary structural components of ribosomes during the synthesis of mitochondrial protein subunits. A case in point is the discovery of highly disruptive base changes in the RNAs―perhaps to a greater degree in the tRNAs (http://www.mitomap.org/MITOMAP); these are known to negatively affect human health and are therefore likely to be purged or kept at low frequencies in a population in response to selection. Moreover, Pesole et al. (1999) have proposed that many changes in the tRNA and rRNA genes are compensatory in nature to maintain critical secondary and tertiary structures. Selective processes may have also shaped the genetic diversity of the CR, as this functional region plays a major role in accelerating the synthesis of mitochondrial DNA to satisfy cellular energy demands in mammals (Fish et al.

2004).

Evolutionary rates differed considerably among the protein-coding genes in the puma- jaguarundi (Figure B2). With respect to nonsynonymous substitutions, the COI gene evolved at the slowest rate, whereas the ATP8 gene evolved ~15 times faster (Figure B2a). In general, the cytochrome c oxidase (complex IV) subunits encoded by the mitochondria (COI, COII, and

COIII) could be conserved as a result of co-adaptation with proteins from the electron transport system encoded by the nuclear genome (Adkins et al. 1996; Zhen et al. 1999). Essentially, the nuclear encoded cytochrome c of the ubiquinol-cytochrome c oxidoreductase (complex III) is expected to dock with the cytochrome c oxidase (complex IV) to transfer electrons between these complexes (Blier et al. 2001). On the other hand, da Fonseca et al. (2008) have shown that high amounts of amino acid variation in the ATP synthase (complex V) are not uncommon in mammalian species. In this regard, it is possible for these amino acid changes to be associated with the mitochondrial coupling efficiency and with adaptations to environmental changes

(Mishmar et al. 2003; Wallace 2007). All 13 mitochondrial protein-coding genes in the puma- jaguarundi were under significant negative selection constraints (Ka/Ks < 1; P < 0.01 for each method for each gene; Figure B2c). Such patterns of mitochondrial evolution are consistent across vertebrates and invertebrates (Pesole et al. 1999; Ballard 2000a, 2000b; Bazin et al. 2006;

Hassanin et al. 2009).

We identified four nonsynonymous SNPs in the puma data set and resolved two of these to be deleterious using PROVEAN (Table B1). The latter SNPs corresponded to a P327S amino acid mutation in the ND2 gene (present only in haplotype Pco4 and possibly introduced into

South Florida by Texas pumas in 1995) and to a P478L amino acid mutation in the ND5 gene

(present only in haplotype Pco2, which is native to Florida) (Table B1).

Functional analyses performed with TreeSAAP suggest that the P478L amino acid mutation in the ND5 gene is found in a protein region associated with ‘alpha-helical tendencies’ and with ‘power to be at the middle of an alpha-helix’ (Figure B3). Hence, if deleterious, this mutation is likely to negatively affect alpha-helical protein properties. The protein encoded by the ND2 gene displayed no regions with highly significant (Z-score > 3.09; P < 0.001) physicochemical properties (data not shown). Mutations along the ND2 and ND5 genes in humans are reported to cause Leber hereditary optic neuropathy and Leigh syndrome, among many other diseases (http://www.mitomap.org/MITOMAP). The predicted deleterious SNPs in pumas, however, have not yet been associated with a particular disease and have not been documented in related felid species or in humans.

New challenges may arise in determining the evolutionary processes that allowed the continued existence of the predicted deleterious SNPs in pumas. Perhaps such mutations are mildly deleterious and/or too recent for selection to have purged them (Nachman 1998; Rand and

Kann 1998; James et al. 2016). Furthermore, a sudden and recent population bottleneck―and thus genetic drift―could have prompted a random increase in their frequency (Culver et al.

2008). There is compelling evidence suggesting that Florida panthers have greatly benefited from the 1995 genetic restoration program (Land et al. 2004; Hostetler et al. 2010; Johnson et al.

2010; Benson et al. 2011). Nevertheless, this study warrants the proper screening for and monitoring of potential deleterious variants in extant Florida panthers.

Funding

This project was funded by the William A. Calder III Memorial Scholarship from the

Department of Ecology and Evolutionary Biology of the University of Arizona. Consejo

Nacional de Ciencia y Tecnología and National Science Foundation-Integrative Graduate

Education and Research Traineeship scholarships were awarded to A. Ochoa.

Acknowledgments

We thank the Florida Fish and Wildlife Conservation Commission and the National Park Service for collecting and providing the biological samples used in this study. E. Vaughn and three anonymous reviewers provided useful comments and revisions to the manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Data Availability

We deposited the puma mitogenome reference-based assemblies produced in this study in

GenBank (https://www.ncbi.nlm.nih.gov/genbank/) under Acc. Nos. KX808222‒KX808231.

TABLES

Table B1. Predicted effect on fitness of the nonsynonymous single nucleotide polymorphisms

(SNPs) characterized in the puma mitogenome.

FIGURES

Figure B1. Mitochondrial Bayesian phylogenetic tree showing intra (haplotypes Pco1‒Pco5: geographic origin) and interspecific (jaguarundi, cheetah, and domestic cat) divergence times relative to the puma. Gray horizontal bars on each node represent 95% highest posterior density intervals (HPDIs). Point divergence time estimates (years before present) for each puma haplotype are also indicated. FL = Florida; TX = Texas; CRI = Costa Rica; PAN = Panama.

Figure B2. (a) Ka, (b) Ks, and (c) Ka/Ks estimates for the mitochondrial protein-coding genes in the puma-jaguarundi. Bottom, middle, and top horizontal lines of each box indicate lower quartile, median, and upper quartile estimates for each statistic (Ka, Ks, and Ka/Ks) across seven approximate methods: γ-NG, γ-LWL, γ-MLWL, γ-LPB, γ-MLPB, γ-YN, and γ-MYN (Wang et al. 2009a, 2009b). Horizontal bars at the end of each whisker represent minimum and maximum values within 1.5 deviations from the interquartile range. Circles represent numerical outliers.

Genes are ordered from smallest to greatest Ka values. 53

Figure B3. Physicochemical properties of the protein encoded by the ND5 gene, as examined across the puma, jaguarundi, cheetah, domestic cat, and spotted linsang. Protein regions are defined by sliding windows of 20 amino acids each. The P478L amino acid mutation in the puma is present in a protein region (bracket at the bottom of the graph) with highly significant (Z-score

> 3.09; P < 0.001; dashed gray line) ‘alpha-helical tendencies’ (black line) and ‘power to be at the middle of an alpha-helix’ (solid gray line).

SUPPLEMENTARY MATERIAL

Supplementary Table B1. Puma samples with known location used in this study (modified from

Johnson et al. 2010).

*The mitogenome of sample FP12 was downloaded from GenBank (Acc. No. KP202261; Li et al. 2016).

FP = Florida panther; TX = Texas puma.

Supplementary Table B2. Summary statistics of the trimmed and error-corrected reads used for the puma mitogenome reference-based assemblies.

FP = Florida panther; TX = Texas puma.

Supplementary Table B3. Sites with ambiguous bases in the puma mitogenome reference-based assemblies (17,153 bp).

R = A and G; Y = C and T; W = A and T; M = A and C.

FP = Florida panther; TX = Texas puma.

Supplementary Table B4. Single nucleotide polymorphisms (SNPs) across 891 bp of the 16S rRNA, ATP8, and ND5 genes for haplotypes Pco1−Pco5 and haplotypes A−N (modified from

Culver et al. 2000).

16S rRNA = sites 2,951−3,332; ATP8 = sites 8,676−8,866; ND5 = sites 12,679−12,996 of the puma mitogenome (17,153 bp).

ARG = Argentina; BOL = Bolivia; BRA = Brazil; CRI = Costa Rica; ECU = Ecuador; FL =

Florida, U.S.; GUF = French Guiana; GUY = Guyana; NA = North America; PAN = Panama;

PER = Peru; PRY = Paraguay; RCH = Chile; TX = Texas, U.S.; VEN = Venezuela; WA =

Washington, U.S.

Base pairs identical to haplotype Pco1 are indicated by a period.

Haplotype Pco1 corresponds to haplotype C.

Haplotypes Pco2−Pco4 correspond to haplotype M.

Haplotype Pco5 was downloaded from GenBank (Acc. No. JN999997; D. McCulloch unplubl. data). 58

Supplementary Figure B1. Florida panther sire, Texas puma dam, and offspring trios from the

1995 Florida panther genetic restoration program (modified from Johnson et al. 2010). Squares represent males and circles represent females. Individual IDs are indicated inside each square or circle. Complete mitogenome sequences were assembled for individuals colored in gray. FP =

Florida panther; CM7 = unsampled Florida panther sire; TX = Texas puma.

Supplementary Figure B2. Single nucleotide polymorphisms (SNPs) characterized in the puma mitogenome (haplotypes Pco1−Pco5). Each SNP is indicated as a gray bar within the inner circle. Locations for each gene and the control region (CR) are indicated in the outer circle. Gray arrows represent the direction of the transcription of genes located in the H- and L-strands. Black arrows represent the origin and direction of the replication of the H- and L-strands. OHR = origin of the H-strand replication; OLR = origin of the L-strand replication. 60

Supplementary Figure B3. Counts of mitochondrial single nucleotide polymorphisms (SNPs) in the puma (haplotypes Pco1−Pco5) and fixed differences between the puma and the jaguarundi, cheetah, domestic cat, and spotted linsang. Statistics are shown at the genome level and for each functional region: protein-coding genes (nonsynonymous and synonymous mutations), tRNAs, rRNAs, and control region (CR).

APPENDIX C

Published Manuscript: Conservation Genetics

Can captive populations function as sources of genetic variation for reintroductions into the wild? A case study of the Arabian oryx from the Phoenix Zoo and the Shaumari

Wildlife Reserve, Jordan

Alexander Ochoa, Stuart A. Wells, Gary West, Ma’en Al-Smadi, Sergio A. Redondo, Sydnee R.

Sexton, and Melanie Culver

From the School of Natural Resources and the Environment, University of Arizona, Tucson, AZ

85721 (Ochoa, Redondo, Sexton, and Culver); Arizona Center for Nature Conservation/Phoenix

Zoo, Phoenix, AZ 85008 (Wells and West); Royal Society for the Conservation of Nature,

Jubeiha, JO 1215 (Al-Smadi); Department of Biology, Stanford University, Stanford, CA 94305

(Redondo); and U.S. Geological Survey, Arizona Cooperative Fish and Wildlife Research Unit,

Tucson, AZ 85721 (Culver).

Address correspondence to Alexander Ochoa at the address above, or e-mail: [email protected].

Abstract

The Arabian oryx (Oryx leucoryx) historically ranged across the Arabian Peninsula and neighboring countries until its extirpation in 1972. In 1963‒1964 a captive breeding program for this species was started at the Phoenix Zoo (PHX); it ultimately consisted of 11 animals that became known as the ‘World Herd’. In 1978‒1979 a wild population was established at the

Shaumari Wildlife Reserve (SWR), Jordan, with eight descendants from the World Herd and three individuals from Qatar. We described the mtDNA and nuclear genetic diversity and structure of PHX and SWR. We also determined the long-term demographic and genetic viability of these populations under different reciprocal translocation scenarios. PHX displayed a greater number of mtDNA haplotypes (n = 4) than SWR (n = 2). Additionally, PHX and SWR presented

nuclear genetic diversities of NA = 2.88 vs. 2.75, HO = 0.469 vs. 0.387, and H E = 0.501 vs.

0.421, respectively. Although these populations showed no signs of inbreeding ( FIS ~ 0), they were highly differentiated (G''ST = 0.580; P < 0.001). Migration between PHX and SWR (Nm =

1, 4, and 8 individuals/generation) increased their genetic diversity in the short-term and substantially reduced the probability of extinction in PHX during 25 generations. Under such scenarios, maximum genetic diversities were achieved in the first generations before the effects of genetic drift became predominant. Although captive populations can function as sources of genetic variation for reintroduction programs, we recommend promoting mutual and continuous gene flow with wild populations to ensure the long-term survival of this species.

Keywords: translocations, admixture, mtDNA control region, microsatellites, population viability analysis, Oryx leucoryx.

Introduction

Captive breeding programs have become an important resource for the conservation of wildlife that has suffered sharp declines and near extinctions as a result of factors such as habitat destruction and overharvesting (Witzenberger and Hochkirch 2011). In this regard, zoos have played a pivotal role in the conservation of endangered species not only by means of scientific research and public education, but also through the preservation of genetic material (e.g. frozen sperm), management policies that favor the partial recovery of population sizes, and the reintroduction of captive-bred animals into their natural habitat (Pearce-Kelly et al. 1995; Fickel et al. 2007; Clarke 2009; Xia et al. 2014).

Integral management and conservation strategies for captive populations also include maintaining the physical health of individuals, reducing the likelihood of behavioral anomalies associated with enclosure and human contact, and preserving the genetic variability required for future adaptation to environmental stressors in the wild (Lacy 1997; Kirkwood 2003; Pelletier et al. 2009). With respect to the preservation of genetic variability, however, zoos oftentimes face the reality of reproducing only a reduced number of founder individuals from one or a few remaining wild populations (Hedrick and Fredrickson 2008; Tzika et al. 2009; Schulte-Hostedde and Mastromonaco 2015).

Although knowledge of the place of origin of individuals and the careful examination of pedigrees and genetic variation of lineages raised in captivity can reduce inbreeding and the expression of recessive deleterious traits (i.e. inbreeding depression; Frankham 1995; Hedrick and Kalinowski 2000), captive populations will inevitably lose evolutionary potential owing to small effective sizes (i.e. genetic drift) and genetic isolation (Allendorf and Luikart 2007;

Willoughby et al. 2015). Furthermore, maintenance of genetic variation in captivity can be challenging for species with low reproductive rates (Marker and O’Brien 1989) or with mating systems where not all individuals contribute their genes to the next generation (e.g. polygyny;

Renan et al. 2015). Thus, an increase in effective population sizes and a management strategy which includes long-term gene flow with other populations are required to balance the effects of genetic drift in captive animals, especially when they function as sources of genetic variation for reintroduction programs.

The Arabian oryx (Oryx leucoryx) historically ranged across the Arabian Peninsula, inhabiting the gravel plains, shallow wadis, sand dunes, rocky hillsides, and thick brushes of

Saudi Arabia, Kuwait, the United Arab Emirates (UAE), Oman, and Yemen (Nowak 1999).

They also extended west to Egypt (Sinai Peninsula) and north to Israel, Syria, Jordan, and Iraq

(west of the Euphrates) (Wilson and Reeder 1993). It is the smallest of four antelope species in the genus Oryx and the only one native to Asia. Arabian oryx are gregarious and polygynous, typically forming herds of 10 individuals with a dominant adult male and several adult females with their offspring (Stanley-Price 1989). Sexual maturity is reached after two years and females usually give birth to a single calf (Stewart 1963).

In 1972 the Arabian oryx was extirpated from the wild due to overhunting (Henderson

1974). In 1963‒1964 a captive breeding program for this species was started at the Phoenix Zoo with nine founders: two males and one female captured in Yemen and donated by the Fauna and

Flora Preservation Society, one female captured in Oman and donated by the Zoological Society of London, one female donated by H. H. Sheikh Jaber bin Abdullah Al-Sabah of Kuwait, and two breeding pairs donated by H. M. King Saud bin Abdul Aziz of Saudi Arabia. These nine

founders became known as the ‘World Herd’ and, eventually, seven individuals from this collection produced offspring (the female from Kuwait and one of the males from Saudi Arabia never bred). In 1967 the World Herd was supplemented with two animals from the Los Angeles

Zoo—one male and one female pregnant by a different male, both of which originally came from

Saudi Arabia—to serve as the founder population for future captive-breeding and reintroduction programs in other US facilities (mainly the San Diego Wild Animal Park and Gladys Porter

Zoo), Europe, the Middle East, and Morocco (Homan 1988; Stanley-Price 1989; Marshall 1998).

As a result of a successful management plan, eight Arabian oryx descendants from the

World Herd (four pairs from the San Diego Wild Animal Park) and one male and two females from a private farm, Al Wukayr, in Qatar, were transferred to the Shaumari Wildlife Reserve,

Jordan, in 1978‒1979 to establish a population to further translocate individuals to captive breeding facilities and other reserves in the Middle East (Stanley-Price 1989; Marshall 1998).

The Shaumari Wildlife Reserve population had expanded to ~270 Arabian oryx by the beginning of the 2000s (Harding et al. 2007). However, epizootic diseases, depredation by jackals, floods, and translocations to other populations in the Middle East have currently reduced the Arabian oryx population of the Shaumari Wildlife Reserve to 50 individuals (Marshall 1998; Harding et al. 2007; M. Al-Smadi pers. comm.). Arabian oryx from the Shaumari Wildlife Reserve are regarded as wild: they occupy a protected, fenced area of 22 km2, but here the habitat is native and human interference is minimal (Abu-Jafar and Shays-Shahin 1988; Harding et al. 2007).

Given recent cooperative efforts between the Phoenix Zoo and the Shaumari Wildlife

Reserve for the conservation of this species, the goals of this study were to describe the genetic diversity and structure of these populations and to determine the effect of migration on their

long-term demographic and genetic viability. Based on these results, we evaluated if captive populations can function as sources of genetic variation for reintroductions into the wild.

Materials and Methods

Samples and laboratory experiments

We collected contemporary Arabian oryx fecal samples from the Phoenix Zoo (PHX; n = 12) and tissue biopsies from the Shaumari Wildlife Reserve (SWR; n = 21) (Supplementary Table

C1). We used the DNeasy Tissue Kit (Qiagen Co., Valencia, CA, USA) to isolate genomic DNA from these samples following the manufacturer’s recommendations. For the fecal samples, however, the first step of the extraction process consisted of soaking the external surface of five pellets from each individual with 1 ml of ATL buffer and adding the resulting solution into a 1.5 ml microtube containing 20 μl of proteinase K for cellular lysis.

We amplified the mtDNA control region using the same polymerase chain reaction

(PCR) conditions and primers (MFR-F and MFR-R) tested by El Alqamy et al. (2012). We performed PCR in 20 μl reaction volumes containing 3 μl DNA template, 0.5 mM dNTPs, 1 μM forward primer, 1 μM reverse primer, 1 unit of Taq polymerase, and 1 PCR buffer (Qiagen

Co.). We sequenced the amplicons from each individual in both forward and reverse directions at the University of Arizona Genetics Core (UAGC, Tucson, AZ, USA) using an ABI3730XL automatic sequencer (Applied Biosystems, Foster City, CA, USA). We then assembled the

forward and reverse sequences from each individual in SEQUENCHER v5.3 (Gene Codes Co.,

Ann Arbor, MI, USA).

We also amplified eight nuclear DNA bovid microsatellite loci [BM3501, MB066,

MCM38, MCMAI, MNS64 (MacHugh et al. 1997; Zhou et al. 2007); IOBT395 (Nijman et al.

1996); OarCP26 (Ede et al. 1995); and OarFCB304 (Buchanan and Crawford 1993)] for each individual using PCR. We performed PCR in 10 μl volumes containing 1 μl DNA template, 0.25 mM dNTPs, 0.5 μM fluorescently labeled M13 primer, 0.5 μM forward primer, 0.5 μM reverse primer, 0.5 units of Taq polymerase, and 1 PCR buffer (Qiagen Co.). PCR conditions consisted of 25 cycles of 45 s at 95 °C, 45 s at 55 °C, and 45 s at 72 °C, with an initial denaturation step of

95 °C for 4 m and final extension of 72 °C for 4 m. We analyzed the amplicon lengths on an

ABI3730 automatic sequencer (Applied Biosystems) at the UAGC and scored alleles using

Genotyper v2.1 (Applied Biosystems). We repeated this procedure three times for every DNA sample to control for allele dropout and ambiguous genotypes.

mtDNA statistical analyses

We downloaded Arabian oryx mtDNA control region sequences from the GenBank nucleotide database (Acc. Nos. AJ235326, FJ797434, FJ821297‒FJ821313, FJ860220‒FJ860225,

GU230150‒GU230156, and JN632679) and aligned them against the mtDNA sequences produced in this study using SEQUENCHER. We used the nomenclature proposed by El

Alqamy et al. (2012) to label mtDNA haplotypes. We then built a median-joining network

(Bandelt et al. 1999) in PopART v1.7 (http://popart.otago.ac.nz) to infer intraspecific relationships between Arabian oryx mtDNA lineages.

Nuclear genetic diversity and structure

We tested for linkage disequilibrium between pairs of loci for each population and assessed significant departures from Hardy-Weinberg equilibrium (HWE) per locus and population in

GENEPOP v4.2 (Raymond and Rousset 1995; Rousset 2008) using a Markov chain method with a dememorization of 10,000 steps and 100 batches of 5,000 iterations each. For these analyses we used a Bonferroni adjustment for multiple tests to account for false positives (Rice 1989). We ran 1,000 Monte Carlo simulations in MICRO-CHECKER v2.2.3 (van Oosterhout et al. 2004) to detect the presence of null alleles that could also cause departures from HWE.

We used GENEPOP to estimate the number of alleles (NA), the observed heterozygosity

(HO), the expected heterozygosity (HE), and the inbreeding coefficient (FIS) per locus and population. We performed unpaired t-tests to determine statistical differences in the means of these parameters between populations after confirming that observations followed a normal distribution (Shapiro-Wilk test) and variances were equal (Levene’s test) at the P = 0.05 level.

We also estimated genetic differentiation between populations (FST and G''ST) with 1,000 permutations in GenAlEx v6.5 (Peakall and Smouse 2006, 2012). We used G''ST as a complementary population structure statistic to FST because the former was developed for highly polymorphic markers (e.g. microsatellites), is standardized with respect to the maximum within-

population HE (scale 0‒1), and is best suited for a small number of populations (Meirmans and

Hedrick 2011).

We used STRUCTURE v2.3.2.1 (Pritchard et al. 2000; Falush et al. 2003; Hubisz et al.

2009) to determine individual assignment probabilities to putative populations assuming admixture and correlated allele frequencies between populations. The number of putative populations (K) ranged from 1 to 10, with 1,000,000 Markov chain Monte Carlo generations and

20 independent simulations per K. We eliminated the first 200,000 generations of each simulation as burn-in. We then used the CLUMPAK server (clumpak.tau.ac.il; Kopelman et al.

2015) to summarize simulations for each K and to identify the most likely number of populations based on the maximum value of ΔK (Evanno et al. 2005). We estimated mean effective population sizes using approximate Bayesian computation in ONeSAMP (Tallmon et al. 2008).

Simulations consisted of 50,000 iterations, each with an effective population size range of 2–500 individuals.

Population viability analysis and migration

We used Vortex v.10.1.5.0 (Lacy and Pollack 2014) to examine the long-term demographic and genetic viability of the PHX and SWR populations assuming no migration and different reciprocal translocation scenarios. We simulated 1,000 PHX and SWR populations for each run consisting of Nm = 0, 1, 4, and 8 migrants exchanged between populations each generation (~8 years/generation; Stanley-Price 1989) for 25 generations.

We considered a census size of 16 individuals for the PHX population and a census size of 50 individuals for the SWR population. We assumed a maximum lifespan of 18 years for

Arabian oryx and used the default, stable age distributions from Vortex to define the number of individuals in each age-sex class for each population. Female and male reproductive ages were

2–16 and 2–12 years, respectively. We set the number of breeding adult females in each population to 65% and the mate monopolization ratio to four dams per sire, with each successful mating resulting in one calf per year. We considered a mortality rate of 20% for individuals < 1 year old. The annual mortality rate for individuals ≥ 1 year old was 5%.

We bred adults to maintain populations at current census sizes. We used the dynamic mean kinship (mk) selection method to minimize inbreeding in the PHX population only, as animals from the SWR population are not thoroughly managed. This method ranks animals according to their mk values and selects the breeding pair with the lowest mks; prospective offspring are added to the population and mk values are recalculated for overlapping generations

(Pollack et al. 2002; Ivy and Lacy 2012). We assumed a genetic load of 6.29 total lethal equivalents―combined mean effect of inbreeding on fecundity and first year survival―per individual; of these total lethal equivalents, 50% were due to recessive lethals (O’Grady et al.

2006).

We limited migrants between populations to individuals with an age of 2–10 years and used the empirical allele frequency distributions from each population (along with a mutation

-4 rate of μ = 510 per locus per individual; Garza and Williamson 2001) to estimate the NA ,

HO , and H E per generation across extant populations (i.e. populations where both sexes remained). Extinct populations (i.e. populations where only one sex remained) could be

‘recolonized’ only in those scenarios where migration was allowed. Hence, we defined

‘probability of extinction’ as the proportion of simulated PHX and SWR populations that went extinct at least once during 25 generations for each scenario. We implemented the demographic parameters for this analysis according to our experience.

Results

mtDNA statistical analyses

We obtained 11 mtDNA haplotypes (Hap A‒K) as a result of analyzing the 33 mtDNA control region sequences produced in this study (12 from PHX and 21 from SWR) along with the 33 mtDNA sequences downloaded from the GenBank (Table C1; Supplementary Table C1). These haplotypes were defined by 23 polymorphic sites (either C↔T or A↔G transitions) within a

638-bp alignment (Table C1).

The PHX population presented four mtDNA haplotypes (Hap A, E, F, and K) and the

SWR population presented two mtDNA haplotypes (Hap A and G) (Figure C1). Only one haplotype, Hap A, was shared between these populations. Hap K (GenBank Acc. No.

KU985184) from PHX had not been documented before in Arabian oryx and was observed for the first time in these data (Table C1).

The median-joining network (Figure C1) shows that Arabian oryx mtDNA haplotypes are separated from each other by a maximum of 16 mutational steps (Hap C and K). At most, there

are 14 mutational steps between haplotypes from PHX (Hap F and K), whereas there are only six mutational steps between the haplotypes from SWR (Hap A and G).

Nuclear genetic diversity and structure

Of the total dataset (n = 33), we identified five alleles at locus OarCP26; four alleles at locus

MCMAI; three alleles at loci MCM38, MNS64, and MB066; and two alleles at loci IOBT395,

BM3501, and OarFCB304. PHX (n = 12) had two unique alleles at locus OarCP26 and SWR (n

= 21) had one unique allele at this same locus (Supplementary Table C1).

We assumed independence among loci as no tests of linkage disequilibrium were significant at the P = 0.05 level after a sequential Bonferroni adjustment. Likewise, all loci from

PHX and SWR were in HWE after a Bonferroni correction for multiple tests. Only locus

IOBT395 in PHX displayed a high frequency of null alleles (P < 0.05; Table C2). Nonetheless, we included this locus for the comparative analyses because results were not substantially modified when excluding it from our study.

Though not statistically different at the P = 0.05 level, PHX had a greater mean number

of alleles ( NA = 2.88 vs. 2.75), observed heterozygosity ( HO = 0.469 vs. 0.387), and expected

heterozygosity ( H E = 0.501 vs. 0.421) than SWR (Table 2). We found no evidence of

inbreeding in either population across loci ( FIS ~ 0; Table C2).

PHX and SWR displayed highly significant genetic differentiation values (FST = 0.202 and G''ST = 0.580; P < 0.001 in both cases). The Bayesian population assignment algorithm implemented in the program STRUCTURE further confirmed subdivision between populations.

In this analysis, the maximum ΔK value was recorded at K = 2. Individuals from PHX and SWR were assigned to opposite gene pools with average memberships of 94% and 92% to the predominant genetic cluster, respectively, and without any prior information about their origin

(Figure C2). We estimated a mean effective population size of 17 individuals (95% CL = 13–28) for PHX and a mean effective population size of 22 individuals (95% CL = 16–40) for SWR across iterations using the approximate Bayesian computation method implemented in

ONeSAMP.

Population viability analysis and migration

In the absence of migration (Nm = 0), 99% of the simulated PHX populations and 3% of the simulated SWR populations went extinct at any time during 25 generations (Table C3).

Migration between populations (Nm = 1, 4, and 8) reduced the probability of extinction in PHX to 13–69% and in SWR to 1% (Table C3).

Assuming no migration between populations (Nm = 0), genetic diversities decreased 56–

81% in PHX and 39–50% in SWR after 25 generations (Table C3). Under this model, PHX

displayed maximum genetic diversities of NA = 2.79, HO = 0.480, and H E = 0.464, while

SWR presented maximum genetic diversities of = 2.72, = 0.410, and = 0.405

(Figure C3).

In general, scenarios where populations exchanged migrants (Nm = 1, 4, and 8) were characterized by an abrupt increase in genetic diversity in the first generations followed by a gradual decrease in genetic diversity in later generations (Figure C3). Under these scenarios,

genetic diversities decreased 29–47% in PHX and 22–33% in SWR after 25 generations (Table

C3). PHX presented maximum genetic diversities of NA = 2.85, HO = 0.510, and H E = 0.537, while SWR displayed maximum genetic diversities of = 2.93, = 0.492, and = 0.487 as a result of migration between populations (Figure C3).

Discussion

Genetic diversity and structure

The PHX population represents an important reservoir of mtDNA genetic diversity for the

Arabian oryx. Despite having a reduced number of individuals, this population presented four mtDNA haplotypes (Hap A, E, F, and K), most of which are highly diverse, as indicated by the number of mutational steps (14 in the most extreme case) linking these haplotypes. Furthermore, one haplotype from this population, Hap K, is novel and undocumented in other Arabian oryx populations. On the other hand, the SWR population displayed only two mtDNA haplotypes,

Hap A and G, which are separated by six mutational steps and have been documented in the Al

Bahia population of the UAE (El Alqamy et al. 2012). Previously, Khan et al. (2011) reported seven mtDNA haplotyopes (Hap B, C, D, E, H, I, and J) for populations reintroduced in Saudi

Arabia (Mahazat as-Sayd Protected Area and National Wildlife Research Center), whereas El

Alqamy et al. (2012) reported the same number of mtDNA haplotypes (Hap A, B, C, D, E, F, and G) for populations reintroduced in the UAE (Sir Bani Yas Island, Al Bahia, and Al Ain

Zoo). The median-joining network suggests that there are as many as 18 unsampled Arabian

oryx mtDNA haplotypes, some of which could have been lost as a consequence of a recent population bottleneck.

Both Arabian oryx populations examined in this study presented comparable (not

significantly different) nuclear genetic diversities. Minimal variations in NA and H E between populations are attributed to the presence of private alleles (two in PHX and one in SWR) at locus OarCP26 and to differences in allele frequencies at each locus. Similarly, Marshall et al.

(1999) reported a H E = 0.509 across six microsatellite loci for the captive Arabian oryx

population of the San Diego Wild Animal Park and a H E = 0.470–0.626 for populations reintroduced in Oman (Arabian Oryx Sanctuary) and Saudi Arabia (King Khalid Wildlife

Research Center and National Wildlife Research Center). Arif et al. (2010) reported a combined

NA = 3.00 and = 0.565 across seven microsatellite loci for populations reintroduced in

Saudi Arabia (Mahazat as-Sayd Protected Area and National Wildlife Research Center), whereas

El Alqamy et al. (2012) reported a = 2.25–2.92 and = 0.337–0.384 across 12 microsatellite loci for populations reintroduced in the UAE (Sir Bani Yas Island and Al Bahia).

Despite having a polygynous mating system, Arabian oryx from PHX and SWR

displayed no evidence of inbreeding (populations are in HWE and FIS ~ 0). In the Phoenix Zoo management strategies are aimed at removing dominant males with long breeding periods from their herds, since this significantly reduces the number of father-daughter matings. In addition, groups with distinct origins and parentage are kept in separate enclosures and interbred to equalize the genetic contributions of the founders (Stanley-Price 1989). In contrast, human- mediated rotation of males is not implemented in the Shaumari Wildlife Reserve. Here, animals occupy a single large enclosure: subordinate males may find opportunities to reproduce with 76

receptive females by chance, or challenge the dominant males and form their own herds every mating season. Consequently, the likelihood of finding alleles that are identical by descent in the next generation is reduced.

Genetic differentiation between PHX and SWR is attributed to genetic isolation―no

Arabian oryx born in the Phoenix Zoo or in the Shaumari Wildlife Reserve have been translocated between these populations since 1990—coupled with the random drift of alleles in either or both populations at any given time. Nevertheless, it is also feasible that the observed genetic structure is the result of independent admixture between distinct lineages occurring in

PHX and SWR since the foundation of these populations. For instance, the SWR population was founded with eight descendants from the World Herd (Studbook Nos. 55, 76, 81, 92, 109, 110,

111, and 125; see Goodwin 2014) and three animals from the Al Wukayr group (Studbook Nos.

188, 189, and 190) in 1978‒1979. Once initial admixture between the World Herd and the Al

Wukayr lineages took place, five male descendants from the World Herd [three from the Zurich

Zoo (Studbook Nos. 513, 579, and 600) and two from the San Diego Wild Animal Park

(Studbook Nos. 1051 and 1055)] were imported to SWR between 1984 and 1990.

Similarly, in 1992 and 1997 two male Arabian oryx (Studbook Nos. 1349 and 1361) from the Al Areen Wildlife Park, Bahrain, were transferred to the PHX population. Ancestors of these males possibly included two pairs from a private farm, Al Sulaimi, in Qatar, and two females from a collection of the late H. H. Sheikh Zayed bin Sultan Al Nahyan of the UAE (Marshall

1998). Little information is available about the geographic origin of the founders of the Qatar (Al

Wukayr and Al Sulaimi) and UAE collections, but these individuals were possibly captured in

the southern Rub' al Khali desert, perhaps near the border of Saudi Arabia and Oman, between

1967 and 1972 (Henderson 1974; Jones 1990; Cribiu et al. 1991; Marshall 1998).

We can further hypothesize that the PHX and SWR populations represent two different, admixed gene pools: World Herd-Al Sulaimi-UAE and World Herd-Al Wukayr, respectively.

Moreover, discordant levels of shared ancestries are observed at the mitochondrial level. In this sense, all mtDNA haplotypes found in PHX are of World Herd origin, as no females of Qatar or

UAE ancestry have ever been introduced into this population. Hap G, which is found in 81% of the individuals from SWR and is absent in PHX, was most likely introduced into SWR by at least one of the female founders of Al Wukayr ancestry. Due to either inaccuracies or lack of information in the studbook record, we were unable to successfully assign all mtDNA haplotypes to the female founders of the PHX and SWR populations.

Though reproduction between unrelated individuals within PHX and SWR can increase

their HO and can counteract inbreeding depression, these populations will not be able to regain genetic variation lost by genetic drift in the short-term, let alone acquire novel alleles introduced

independently in both populations since their foundation. Under this breeding scheme, NA and

H E will most likely decrease in PHX and SWR as a function of time and of their limited effective population sizes. This premise holds true for mtDNA genetic diversity as well, particularly for number of haplotypes, provided that mitochondrial genomes have a smaller effective size than nuclear genomes (i.e. mitochondrial genomes are haploid and inherited only through maternal lineages). Given that the genetic differentiation statistics and the Bayesian population assignment test suggest that PHX and SWR are highly structured, there is a better

chance of significantly increasing their NA , HO , and H E if gene flow between these two populations is promoted.

Population viability analysis and migration

Our results suggest that the PHX and SWR populations can benefit from mutual translocations.

From a demographic perspective, reciprocal translocations of individuals contribute to balancing the number of adult males and females, further granting these populations the ability to produce offspring each generation. From an evolutionary perspective, bidirectional migration allows populations to acquire novel genetic variation and to increase the frequency of rare alleles. Also, migration between populations can alleviate their genetic loads and, thus, can minimize the risk of inbreeding depression (Schwartz and Mills 2005; Whiteley et al. 2015).

Given that PHX has more private alleles than SWR, the latter has a greater potential of increasing its genetic diversity from translocation events and subsequent gene flow. For instance,

SWR can increase its by 7%, by 20%, and by 20%, while PHX can only increase its by 2%, by 6%, and by 15% as a result of comparing maximum genetic diversities with respect to initial genetic diversities. Furthermore, SWR will retain gained genetic variation for a longer number of generations because it has a greater effective population size than PHX. Still, maximum and values are observed in PHX since rare alleles in this population are found at higher relative frequencies in those particular generations.

Simulations considering mutual translocations between PHX and SWR suggest that these management strategies are only viable in the short-term, as loss of genetic diversity becomes 79

predominant after a few generations even if migration between these populations remains constant. In this regard, both populations have smaller effective sizes than those required for the long-term retention of genetic diversity (Ne = 500‒1,000; Franklin and Frankham 1998).

Implications for management and conservation

In this study we have demonstrated that captive Arabian oryx populations can function as sources of genetic variation for reintroductions into the wild: these populations may have novel mtDNA haplotypes reflecting unique evolutionary trajectories, or they may have greater nuclear genetic diversities than wild populations. Nevertheless, mutual and continuous translocations between captive and wild populations are desirable since overall gains in genetic diversity could be more significant. Likewise, captive and wild populations may benefit from assisted breeding programs (e.g. artificial insemination; Comizzoli et al. 2000; Schook et al. 2013) where transcontinental transportation of individuals and quarantine confinements can be evaded.

Though gene flow may carry the risk of outbreeding depression (Marshall and Spalton

2000; Tallmon et al. 2004; Edmands 2007), it is unlikely that Arabian oryx populations have accumulated adaptations to local environments prior to extirpation and since reintroduction—or as a result of enclosure—because populations have not been isolated for the last 500 years and because there are no clear taxonomic differences between them (Frankham et al. 2011; El

Alqamy et al. 2012). However, matings between individuals with different chromosomal arrangements could further lead to a decreased fitness in the progeny (Pagacova et al. 2011). A case in point is a Robertsonian translocation (a fusion between chromosomes 17 and 19 from a

standard 2n = 58) of Al Wukayr origin that was unknowingly introduced into the Shaumari

Wildlife Reserve and then was possibly spread to other populations in the Middle East (Cribiu et al. 1990, 1991; Marshall and Spalton 2000). Therefore, karyotype screenings should be performed in Arabian oryx populations prior to exchanging migrants.

Cryopreservation of sperm and oocytes could be useful for Arabian oryx managers to maintain the genetic diversity of their stocks, given the limited effective sizes these populations may have (Roth et al. 1999; Frankham et al. 2010; Boutelle et al. 2011). Population genomics studies are also essential to developing management strategies aimed at increasing the fitness of reintroduced animals into their natural habitat (Gompert 2012; Harrisson et al. 2014; McMahon et al. 2014). Furthermore, such studies may shed new light on determining the ancestries of current Arabian oryx (vonHoldt et al. 2011; Der Sarkissian et al. 2015). Perhaps implementing an Arabian oryx metapopulation in the wild along with captive populations that can function as genetic reservoirs for future reintroduction programs will ensure the long-term survival of this species.

Funding

This project was funded by the Arizona Center for Nature Conservation/Phoenix Zoo and by the

Consejo Nacional de Ciencia y Tecnología and the National Science Foundation-Integrative

Graduate Education and Research Traineeship scholarships awarded to A. Ochoa.

Acknowledgments

We thank D. Subaitis and J. Swenson from the Arizona Center for Nature Conservation/Phoenix

Zoo and A. H. Eljarah and A. Elhala from the Royal Society for the Conservation of Nature for collecting the Arabian oryx biological samples used in this study. R. Fitak, T. Edwards, and two anonymous reviewers provided useful comments and revisions to the manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Data Availability

We deposited the Arabian oryx mtDNA control region sequence corresponding to Hap K in

GenBank (https://www.ncbi.nlm.nih.gov/genbank/) under Acc. No. KU985184.

TABLES

Table C1. Distribution of 23 polymorphic sites defining 11 Arabian oryx haplotypes across 638 bp of the mtDNA control region. GenBank Acc. Nos. for each mtDNA haplotype are also provided.

Table C2. Genetic diversity of Arabian oryx from the Phoenix Zoo (PHX) and the Shaumari

Wildlife Reserve (SWR) across eight nuclear microsatellite loci.

n = number of successful genotypes; NA = number of alleles; HO = observed heterozygosity; HE

= expected heterozygosity; FIS = inbreeding coefficient; *presence of null alleles (P < 0.05).

Table C3. Population viability analysis and changes in genetic diversity of Arabian oryx from the Phoenix Zoo (PHX) and the Shaumari Wildlife Reserve (SWR) after 25 generations assuming Nm = 0, 1, 4, and 8 migrants exchanged between populations each generation.

N = census size; Prob. of extinction = proportion of simulated populations that went extinct at

least once during 25 generations; NA = mean number of alleles across 1,000 simulated

populations; HO = mean observed heterozygosity across 1,000 simulated populations; H E = mean expected heterozygosity across 1,000 simulated populations; Initial = genetic diversity in generation 0; Final = genetic diversity in generation 25; Decline = rate of loss of genetic diversity after 25 generations.

FIGURES

Figure C1. Median-joining network of 11 Arabian oryx haplotypes across 638 bp of the mtDNA control region. Dark gray circles represent haplotypes found in the Phoenix Zoo (PHX). Light gray circles represent haplotypes found in the Shaumari Wildlife Reserve (SWR). The size of these circles is proportional to the number of individuals displaying the haplotype. White circles represent previously documented Arabian oryx mtDNA haplotypes not found in PHX and SWR.

Small black circles represent vertices connecting multiple branches. Horizontal small lines on branches represent mutational steps between haplotypes and vertices. 86

Figure C2. Genetic structure of Arabian oryx from the Phoenix Zoo (PHX) and the Shaumari

Wildlife Reserve (SWR). Each individual is represented by a vertical line divided into dark and light gray segments that represent the proportion of ancestry to each of K = 2 genetic clusters.

Figure C3. Changes in (a) NA , (b) H O , and (c) H E across 1,000 simulated Phoenix Zoo (PHX; black lines) and Shaumari Wildlife Reserve (SWR; gray lines) populations for 25 generations.

Solid, long-dashed, short-dashed, and dotted lines indicate Nm = 0, 1, 4, and 8 migrants exchanged between populations each generation, respectively. 88

SUPPLEMENTARY MATERIAL

Supplementary Table C1. mtDNA haplotypes and nuclear genotypes (amplicon fragment sizes) of Arabian oryx from the Phoenix Zoo (PHX) and the Shaumari Wildlife Reserve (SWR).

REFERENCES

Abu-Jafar MZ, Hays-Shahin C. 1988. Re-introduction of the Arabian oryx into Jordan. In: Dixon

A, Jones D, editors. Conservation and biology of desert antelopes. London: Christopher

Helm Ltd. p. 35–40.

Adkins RM, Honeycutt RL, Disotell TR. 1996. Evolution of eutherian cytochrome c oxidase

subunit II: heterogeneous rates of protein evolution and altered interaction with

cytochrome c. Mol Biol Evol 13:1393–1404.

Allendorf F, Luikart G. 2007. Conservation and the genetics of populations. Oxford: Blackwell

Publishing.

Allendorf FW, Hohenlohe PA, Luikart G. 2010. Genomics and the future of conservation

genetics. Nature Rev Genet 11:697–709.

Amato G, DeSalle R, Ryder OA, Rosenbaum HC. 2009. Conservation genetics in the age of

genomics. New York: Columbia University Press.

Ames RM, Lovell SC. 2015. DupliPHY-Web: a web server for DupliPHY and DupliPHY-ML.

Bioinformatics 31:416–417.

Anderson S, Bankier AT, Barrell BG, De Bruijn MHL, Courlson AR, Drouin J, Eperon IC,

Nierlich DP, Roe BA, Sanger F, Schreier PH, Smith AJH, Staden R, Young IG. 1981.

Sequence and organization of the human mitochondrial genome. Nature 290:457–465.

Arif IA, Khan HA, Shobrak M, Al Homaidan AA, Al Sadoon M, Al Farhan AH. 2010.

Measuring the genetic diversity of Arabian oryx using microsatellite markers: implication

for captive breeding. Genes Genet Syst 85:141–145.

Ballard JWO. 2000a. Comparative genomics of mitochondrial DNA in members of the

Drosophila melanogaster subgroup. J Mol Evol 51:48–63.

Ballard JWO. 2000b. Comparative genomics of mitochondrial DNA in Drosophila simulans. J

Mol Evol 51:64–75.

Bandelt H, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific

phylogenies. Mol Biol Evol 16:37–48.

Barone M, Roelke M, Howard J, Brown J, Anderson A, Wildt D. 1994. Reproductive

characteristics of male Florida panthers: comparative studies from Florida, Texas,

Colorado, Latin America, and North American zoos. J Mammal 75: 150–162.

Bazin E, Glémin S, Galtier N. 2006. Population size does not influence mitochondrial genetic

diversity in animals. Science 312:570–572.

Benson JF, Hostetler JA, Onorato DP, Johnson WE, Roelke ME, O’Brien SJ, Jansen D, Oli MK.

2011. Intentional genetic introgression influences survival of adults and sub-adults in a

small, inbred felid population. J Anim Ecol 80:958–967.

Ben-Zvi A, Manor O, Schachner M, Yaron A, Tessier-Lavigne M, Behar O. 2008. The

semaphorin receptor PlexinA3 mediates neuronal apoptosis during dorsal root ganglia

development. J Neurosci 28:12427–12432.

Blier PU, Dufresne F, Burton RS. 2001. Natural selection and the evolution of mtDNA-encoded

peptides: evidence for intergenomic co-adaptation. Trends Genet 7:400–406.

Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence

data. Bioinformatics 30:2114–2120.

Boutelle S, Lenahan K, Krisher R, Bauman KL, Asa CS, Silber S. 2011. Vitrification of oocytes

from endangered Mexican gray wolves (Canis lupus baileyi). Theriogenology 75:647–

654.

Buchanan FC, Crawford AM. 1993. Ovine microsatellites at the OarFCB11, OarFCB128,

OarFCB193, OarFCB266 and OarFCB304 loci. Anim Genet 24:145.

Cantarel BL, Korf I, Robb SMC, Parra G, Ross E, Moore B, Holt C, Sánchez Alvarado A,

Yandell M. 2008. MAKER: An easy-to-use annotation pipeline designed for emerging

model organism genomes. Genome Res 18: 188–196.

Castilho CS, Marins-Sá LG, Benedet RC, Freitas TO. 2011. Landscape genetics of mountain

lions (Puma concolor) in southern Brazil. Mamm Biol 76:476–83.

Castresana J. 2000. Selection of conserved blocks from multiple alignments for their use in

phylogenetic analysis. Mol Biol Evol 17:540–552.

Cheng HJ, Bagri A, Yaron A, Stein E, Pleasure SJ, Tessier-Lavigne M. 2001. Plexin-A3

mediates semaphoring signaling and regulates the development of hippocampal axonal

projections. Neuron 32:249–263.

Cho YS, Hu L, Hou H, Lee H, Xu J, Kwon S, Oh S, Kim H-M, Jho S, Kim S, Shin Y-A, Kim

BC, Kim H, Kim C-U, Luo S-J, Johnson WE, Koepfli K-P, Schmidt-Küntzel A, Turner

JA, Marker L, Harper C, Miller SM, Jacobs W, Bertola LD, Kim TH, Lee S, Zhou Q,

Jung H-J, Xu X, Gadhvi P, Xu P, Xiong Y, Luo Y, Pan S, Gou C, Chu X, Zhang J, Liu S,

He J, Chen Y, Yang L, Yang Y, He J, Liu S, Wang J, Kim CH, Kwak H, Kim J-S,

Hwang S, Ko J, Kim C-B, Kim S, Bayarlkhagva D, Paek WK, Kim S-J, O’Brien SJ,

Wang J, Bhak J. 2013. The tiger genome and comparative analysis with lion and snow

leopard genomes. Nat Commun 4:2433.

Choi Y, Chan AP. 2015. PROVEAN web server: a tool to predict the functional effect of amino

acid substitutions and indels. Bioinformatics 31: 2745–2747.

Choi Y, Sims GE, Murphy S, Miller JR, Chan AP. 2012. Predicting the functional effect of

amino acid substitutions and indels. PLoS One 7:e46688.

Choi Y. 2012. A fast computation of pairwise sequence alignment scores between a protein and a

set of single-locus variants of another protein. Proceedings of the ACM Conference on

Bioinformatics, Computational Biology and Biomedicine (BCB ‘12). New York: ACM.

p. 414–417.

Clarke AG. 2009. The Frozen Ark Project: the role of zoos and aquariums in preserving the

genetic material of threatened animals. Int Zoo Yb 43:222–230.

Comizzoli P, Mermillod P, Mauget R. 2000. Reproductive biotechnologies for endangered

mammalian species. Reprod Nutr Dev 40:493–504.

Conesa A, Götz S, García-Gómez JM, Terol J, Talón M, Robles M. 2005. Blast2GO: a universal

tool for annotation, visualization and analysis in functional genomics research.

Bioinformatics 21:3674–3676.

Corbett J. 1954. The temple tiger and more man-eaters of Kumaon. London: Oxford

University Press.

Cribiu EP, Asmondé JF, Durand V, Greth A, Anagariyah S. 1990. Robertsonian chromosome

polymorphism in the Arabian oryx (Oryx leucoryx). Cytogenet Cell Genet 54:161–163.

Cribiu EP, Vassart M, Durand V, Greth A, Asmodé J-F, Claro F, Anagariyah S. 1991.

Distribution of the 17;19 Robertsonian translocation in a herd of Arabian oryx.

Mammalia 55:121–126.

Culver M, Hedrick PW, Murphy K, O’Mrien SJ, Hornocker MG. 2008. Estimation of the

bottleneck size in Florida panthers. Anim Conserv 11:104–110.

Culver M, Johnson WE, Pecon-Slattery J, O’Brien SJ. 2000. Genomic ancestry of the American

puma (Puma concolor). J Hered 91:186–197.

Cunningham M, Dunbar M, Buergelt C, Homer B, Roelke-Parker M, Taylor S, King R, Citino S,

Glass C. 1999. Atrial septal defects in Florida panthers. J Wildl Dis 35: 519–530. da Fonseca RR, Johnson WE, O’Brien SJ, Ramos MJ, Antunes A. 2008. The adaptive evolution

of the mammalian mitochondrial genome. BMC Genomics 9:119.

Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, Handsaker RE, Lunter G,

Marth GT, Sherry ST, McVean G, Durbin R, 1000 Genomes Project Analysis Group.

2011. The variant call format and VCFtools. Bioinformatics 27: 2156–2158.

Der Sarkissian C, Ermini L, Schubert M, Yang MA, Librado P, Fumagalli M, Jónsson H, Kahila

Bar-Gal G, Albrechtsen A, Vieira FG, Petersen B, Ginolhac A, Seguin-Orlando A,

Magnussen K, Fages A, Gamba C, Lorente-Galdos B, Polani S, Steiner C, Neuditschko

M, Jagannathan V, Feh C, Greenblatt CL, Ludwig A, Abramson NI, Zimmermann W,

Schafberg R, Tikhonov A, Sicheritz-Ponten T, Willerslev E, Marques-Bonet T, Ryder

OA, McCue M, Rieder S, Leeb T, Slatkin M, Orlando L. 2015. Evolutionary genomics

and conservation of the endangered Przewalski’s horse. Curr Biol 25:2577–2583.

DeSalle R, Amato G. 2004. The expansion of conservation genetics. Nature Rev Genet 5:702–

712.

Dobrynin P, Liu S, Tamazian G, Xiong Z, Yurchenko AA, Krasheninnikova K, Kliver S,

Schmidt-Küntzel A, Koepfli K-P, Johnson W, Kuderna LFK, García-Pérez R, de Manuel

M, Godinez R, Komissarov A, Makunin A, Brukhin V, Qiu W, Zhou L, Li F, Yi J,

Driscoll C, Antunes A, Oleksyk TK, Eizirik E, Perelman P, Roelke M, Wildt D,

Diekhans M, Marques-Bonet T, Marker L, Bhak J, Wang J, Zhang G, O’Brien SJ. 2015.

Genome Biol 16:277. doi: 10.1186/s13059-015-0837-4.

Ede AJ, Pierson CA, Crawford AM. 1995. Ovine microsatellites at the OarCP9, OarCP16,

OarCP20, OarCP21, OarCP23 and OarCP26 loci. Anim Genet 26:129–130.

Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accuracy and high

throughput. Nucleic Acids Res 32:1792–1797.

Edmands S. 2007. Between a rock and a hard place: evaluating the relative risks of inbreeding

and outbreeding for conservation and management. Mol Ecol 16:463–475.

Eizirik E, Murphy WJ, Koepfli KP, Johnson WE, Dragoo JW, Wayne RK, O’Brien SJ. 2010.

Pattern and timing of diversification of the mammalian order Carnivora inferred from

multiple nuclear gene sequences. Mol Phylogenet Evol 56:49–63.

El Alqamy H, Senn H, Roberts M-F, McEwing R, Ogden R. 2012. Genetic assessment of the

Arabian oryx founder population in the Emirate of Abu Dhabi, UAE: an example of

evaluating unmanaged captive stocks for reintroduction. Conserv Genet 13:79–88.

Evanno G, Regnaut S, Goudet J. 2005. Detecting the number of clusters of individuals using the

software STRUCTURE: a simulation study. Mol Ecol 14:2611–2620.

Falush D, Stephens M, Pritchard JK. 2003. Inference of population structure using multilocus

genotype data: linked loci and correlated allele frequencies. Genetics 164:1567–1587.

Federal Register. 1967. Native Fish and Wildlife: Endangered Species. Washington, D.C.:

Federal Register-Department of the Interior-Fish and Wildlife Service.

Felsenstein J. 1974. The evolutionary advantage of recombination. Genetics 78:737–756.

Fickel J, Wagener A, Ludwig A. 2007. Semen cryopreservation and the conservation of

endangered species. Eur J Wildl Res 53:81–89.

Fish J, Raule N, Attardi G. 2004. Discovery of a major D-loop replication origin reveals two

modes of human mtDNA synthesis. Science 306:2098–2101.

Fitak RR, Naidu A, Thompson RW, Culver M. 2015a. A new panel of SNP markers for the

individual identification of North American pumas. J Fish Wildl Manag 7:13–27.

Fitak RR, Mohandesan E, Corander J, Burger PA. 2015b. The de novo genome assembly and

annotation of a female domestic dromedary of North African origin. Mol Ecol Resour

16:314–324.

Frankham R, Ballou JD, Briscoe DA. 2010. Introduction to conservation genetics, 2nd edition.

Cambridge: Cambrigde University Press.

Frankham R, Ballou JD, Eldridge MDB, Lacy RC, Ralls K, Dudash MR, Fenster CB. 2011.

Predicting the probability of outbreeding depression. Conserv Biol 25:465–475.

Frankham R. 1995. Conservation genetics. Ann Rev Gen 29:305–327.

Franklin IR, Frankham R. 1998. How large must populations be to retain evolutionary potential?

Anim Conserv 1:69–70.

Garza JC, Williamson EG. 2001. Detection of reduction in population size using data from

microsatellite loci. Mol Ecol 10:305–318.

Gompert Z. 2012. Population genomics as a new tool for wildlife management. Mol Ecol

21:1542–1544.

Goodwin I. 2014. International studbook for Arabian oryx (Oryx leucoryx). Winchester: Marwell

Wildlife.

Goodwin S, McPherson JD, McCombie WR. 2016. Coming of age: ten years of next-generation

sequencing technologies. Nature Rev Genet 17:333–351.

Griffiths-Jones S, Bateman A, Marshall M, Khanna A, Eddy SR. 2003. Rfam: an RNA family

database. Nucleic Acids Res 31:439–441.

Grimwood I. 1967. Operation Oryx. Oryx 9:110–122.

Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis

program for Windows 95/98/NT. Nucl Acids Symp Ser 41:95–98.

Harding LE, Abu-Eid OF, Hamidan N, al Sha’lan A. 2007. Reintroduction of the Arabian oryx

Oryx leucoryx in Jordan: war and redemption. Oryx 41:478–487.

Harrisson KA, Pavlova A, Telonis-Scott M, Sunnucks P. 2014. Using genomics to characterize

evolutionary potential for conservation of wild populations. Evol Appl 7:1008–1025.

Hassanin A, Ropiquet A, Couloux A, Cruaud C. 2009. Evolution of the mitochondrial genome in

mammals living at high altitude: new insights from a study of the tribe Caprini (Bovidae,

Antilopinae). J Mol Evol 68:293–310.

Hassanin A, Veron G. 2016. The complete mitochondrial genome of the Spotted Linsang,

Prionodon pardicolor, the first representative from the family Prionodontidae

(Mammalia, Carnivora). Mitochondrial DNA 27:912–913.

He H, Yang T, Terman JR, Zhang X. 2009. Crystal structure of the plexin A3 intracellular region

reveals an autoinhibited conformation through active site sequestration. Proc Natl Acad

Sci USA 37:15610–15615.

Hedrick PW, Fredrickson RJ. 2008. Captive breeding and the reintroduction of Mexican and red

wolves. Mol Ecol 17:344–350.

Hedrick PW, Kalinowski ST. 2000. Inbreeding depression in conservation biology. Ann Rev

Ecol Syst 31:139–162.

Henderson DS. 1974. Were they the last Arabian oryx? Oryx 12:347–350.

Hill JH, Chen Z, Xu H. 2014. Selective propagation of functional mitochondrial DNA during

oogenesis restricts the transmission of a deleterious mitochondrial variant. Nat Genet

46:389–392.

Homan WG. 1988. The establishment of the World Herd. In: Dixon A, Jones D, editors.

Conservation and biology of desert antelopes. London: Christopher Helm Ltd. p. 9–13.

Hostetler JA, Onorato DP, Jansen D, Oli MK. 2013. A cat’s tale: the impact of genetic

restoration on Florida panther population dynamics and persistence. J Anim Ecol 82:608–

620.

Hostetler JA, Onorato DP, Nichols JD, Johnson WE, Roelke ME, O’Brien SJ, Jansen D, Oli MK.

2010. Genetic introgression and the survival of Florida panther kittens. Biol Conserv

143:2789–2796.

Hubisz MJ, Falush D, Stephens M, Pritchard JK. 2009. Inferring weak population structure with

the assistance of sample group information. Mol Ecol Resour 9:1322–1332.

Irizarry KJ, Malladi SB, Gao X, Mitsouras K, Melendez L, Burris PA, Brockman JA, Al-

Murrani SW. 2012. Sequencing and comparative genomic analysis of 1227 Felis catus

cDNA sequences enriched for developmental, clinical and nutritional phenotypes.

BMC Genomics 13:31.

Ivy JA, Lacy RC. 2012. A comparison of strategies for selecting breeding pairs to maximize

genetic diversity retention in managed populations. J Hered 103:186–196.

Jae-Heup K, Eizirik E, O’Brien SJ, Johnson WE. 2001. Structure and patterns of sequence

variation in the mitochondrial DNA control region of the great cats. Mitochondrion

3:279–292.

James JE, Piganeau G, Eyre-Walker A. 2016. The rate of adaptive evolution in animal

mitochondria. Mol Ecol 25:67–78.

Johnson WE, Eizirik E, Pecon-Slattery J, Murphy WJ, Antunes A, Teeling E, O’Brien SJ. 2006.

The late Miocene radiation of modern Felidae: a genetic assessment. Science 311:73–77.

Johnson WE, Onorato DP, Roelke ME, Land ED, Cunningham M, Belden RC, McBride R,

Jansen D, Lotz M, Shindle D, Howard J, Wildt DE, Penfold LM, Hosttler JA, Oli MK,

O’Brien SJ. 2010. Genetic restoration of the Florida panther. Science 329:1641–1645.

Jones DM. 1990. Arabian oryx in Qatar. Unpublished report. London: Zoological Society of

London.

Jones P, Binns D, Chang H-Y, Fraser M, Li W, McAnulla C, McWilliam H, Maslen J, Mitchell

A, Nuka G, Pesseat S, Quinn AF, Sangrador-Vegas A, Scheremetjew M, Yong S-W,

Lopez R, Hunter S. 2014. InterProScan 5: genome-scale protein function classification.

Bioinformatics 30:1236–1240.

Katoh M, Misawa K, Kuma K, Miyata T. 2002. MAFFT: a novel method for rapid multiple

sequence alignment based on fast Fourier transform. Nucleic Acids Res 30:3059–3066.

Khan HA, Arif IA, Shobrak M, Homaidan AA, Farhan AH, Sadoon MA. 2011. Application of

mitochondrial genes sequences for measuring the genetic diversity of Arabian oryx.

Genes Genet Syst 86:67–72.

Kirkwood JK. 2003. Welfare, husbandry and veterinary care of wild animals in captivity:

changes in attitude, progress in knowledge and techniques. Int Zoo Yb 38:124–130.

Kopelman NM, Mayzel J, Jakobsson M, Rosenberg NA, Mayrose I. 2015. CLUMPAK: a

program for identifying clustering modes and packaging population structure inferences

across K. Mol Ecol Resour 15:1179–1191.

Korf I. 2004. Gene finding in novel genomes. BMC Bioinformatics 5:59.

Lacy RC, Pollak JP. 2014. Vortex: a stochastic simulation of the extinction process. Version

10.0. Brookfield: Chicago Zoological Society.

Lacy RC. 1997. Importance of genetic variation to the viability of mammalian populations. J

Mammal 78:320–335.

Land ED, Shindle D, Cunningham M, Lotz M, Ferree B. 2004. Florida Panther Genetic

Restoration and Management. Annual Report 2003–2004. Tallahassee: Florida Fish and

Wildlife Conservation Commission.

100

Laundré JW, Hernández L. 2010. What we know about pumas in Latin America. In: Hornocker

M, Negri S, editors. Cougar Ecology and Conservation. Chicago: University of Chicago

Press. p. 76–90.

Lebon S, Chol M, Benit P, Mugnier C, Chretien D, Giurgea I, Kern I, Girardin E, Hertz-Pannier

L, de Lonlay P, Rötig A, Rustin P, Munnich A. 2003. Recurrent de novo mitochondrial

DNA mutations in respiratory chain deficiency. J Med Genet 40:896–899.

Li G, Davis BW, Eizirik E, Murphy WJ. 2016. Phylogenomic evidence for ancient hybridization

in the genomes of living cats (Felidae). Genome Res doi: 10.1101/gr.186668.114.

Li H, Durbin R. 2009. Fast and accurate short read alignment with Burrows-Wheeler transform.

Bioinformatics 25:1754–1760.

Li H, Durbin R. 2011. Inference of human population history from whole genome sequence of a

single individual. Nature 475: 493–496.

Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R

and 1000 Genome Project Data Processing Subgroup. 2009. The Sequence

alignment/map (SAM) format and SAMtools. Bioinformatics 25:2078–2079.

Li H. 2011. A statistical framework for SNP calling, mutation discovery, association mapping

and population genetical parameter estimation from sequencing data. Bioinformatics

27:2987–93.

Li L, Stoeckert CJ, Roos DS. 2003. OrthoMCL: identification of ortholog groups for eukaryotic

genomes. Genome Res 13:2178–2189.

Liu K, Raghavan S, Nelesen S, Linder CR, Warnow T. 2009. Rapid and accurate large scale

coestimation of sequence alignments and phylogenetic trees. Science 324:1561–1564.

101

Liu K, Warnow TJ, Holder MT, Nelesen S, Yu J, Stamatakis A, Linder CR. 2012. SATé-II: very

fast and accurate simultaneous estimation of multiple sequence alignments and

phylogenetic trees. Syst Biol 61:90–106.

Liu Y, Schröder J, Schmidt B. 2013. Musket: a multistage k-mer spectrum-based error corrector

for Illumina sequence data. Bioinformatics 29:308–315.

Loewe L. 2006. Quantifying the genomic decay paradox due to Muller's ratchet in human

mitochondrial DNA. Genet Res 87:133–159.

Lomsadze A, Ter-Hovhannisyan V, Chernoff Y, Borodovsky M. 2005. Gene identification in

novel eukaryotic genomes by self-training algorithm. Nucleic Acids Res 33:6494–6506.

Lopez JV, Cevario S, O’Brien SJ. 1996. Complete nucleotide sequences of the domestic cat

(Felis catus) mitochondrial genome and a transposed mtDNA tandem repeat (Numt) in

the nuclear genome. Genomics 33:229–246.

Löytynoja A, Goldman N. 2005. An algorithm for progressive multiple alignment of sequences

with insertions. Proc Natl Acad Sci USA 102:10557–10562.

Lynch M, Gabriel W. 1990. Mutation load and the survival of small populations. Evolution 44:

1725–1737.

MacHugh DE, Shriver MD, Loftus RT, Cunningham P, Bradley DG. 1997. Microsatellite DNA

variation and the evolution, domestication and phylogeography of taurine and zebu cattle

(Bos taurus and Bos indicus). Genetics 146:1071–1086.

Malnic B, Godfrey PA, Buck LB. 2004. The human olfactory receptor gene family. Proc Natl

Acad Sci 101:2584–2589.

102

Marker L, O’Brien SJ. 1989. Captive breeding of the cheetah (Acinonyx jubatus) in North

American zoos (1871–1986). Zoo Biol 8:3–16.

Marshall TC, Spalton JA. 2000. Simultaneous inbreeding and outbreeding depression in

reintroduced Arabian oryx. Anim Conserv 3:241–248.

Marshall TC, Sunnucks P, Spalton JA, Greth A, Pemberton JM. 1999. Use of genetic data for

conservation management: the case of the Arabian oryx. Anim Conserv 2:269–278.

Marshall TC. 1998. Inbreeding and ﬁtness in wild ungulates. PhD thesis, University of

Edinburgh.

McBride RT, McBride RT, McBride RM, McBride CE. 2008. Counting pumas by categorizing

physical evidence. Southeast Nat 7:381–400.

McMahon BJ, Teeling, EC, Höglund J. 2014. How and why should we implement genomics into

conservation? Evol Appl 7:999–1007.

Meirmans PG, Hedrick PW. 2011. Assessing population structure: FST and related measures.

Mol Ecol Resour 11:5–18.

Messina A, Ferraris N, Wray S, Cagnoni G, Donohue DE, Casoni F, Kramer PR, Derijck AA,

Adolfs Y, Fasolo A, Pasterkamp RJ, Giacobini P. 2011. Dysregulation of

Semaphorin7A/b1-integrin signaling leads to defective GnRH-1 cell migration, abnormal

gonadal development and altered fertility. Hum Mol Gen 20:4759–4774.

Miller RJ. 1998. Mitochondria – the kraken wakes! Trends Neurosci 21: 95–97.

Miller WR. 2013. Steroid hormone synthesis in mitochondria. Mol Cell Endocrinol 379:62–73.

103

Mishmar D, Ruiz-Pesini E, Golik P, Macaulay V, Clark AG, Hosseini S, Brandon M, Easley K,

Chen E, Brown MD, Sukernik RI, Olckers A, Wallace DC. 2003. Natural selection

shaped regional mtDNA variation in humans. Proc Natl Acad Sci USA 100:171–176.

Mishra P, Chan DC. 2014. Mitochondrial dynamics and inheritance during cell division,

development and disease. Nat Rev Mol Cell Biol 15:634–46.

Montague MJ, Li G,Gandolfi B, Khan R, Aken BL, Searle SMJ, Minx P, Hillier LW, Koboldt

DC, Davis BW, Driscoll CA, Barr CS, Blackistone K, Quilez J, Lorente-Galdos B,

Marques-Bonet T, Alkan C, Thomas GWC, Hahn MW, Menotti-Raymond M, O’Brien

SJ, Wilson RK, Lyons LA, Murphy WJ, Warren WC. 2014. Comparative analysis of the

domestic cat genome reveals genetic signatures underlying feline biology and

domestication. Proc Natl Acad Sci 111:17230–17235.

Muller HJ. 1963. The need for recombination to prevent genetic deterioration. Genetics 48:903–

903.

Nachman MW, Crowell SL. 2000. Estimate of the mutation rate per nucleotide in humans.

Genetics 156:297–304.

Nachman MW. 1998. Deleterious mutations in animal mitochondrial DNA. Genetica 102:61–69.

Nawrocki EP, Burge SW, Bateman A, Daub J, Eberhardt RY, Eddy SR, Floden EW, Gardner

PP, Jones TA, Tate J, Finn RD. 2015. Rfam 12.0: updates to the RNA families database.

Nucleic Acids Res 43:130–137.

Nawrocki EP, Eddy SR. 2013. Infernal 1.1: 100-fold faster RNA homology searches.

Bioinformatics 29:2933–2935.

Nicholls DG. 2005. Mitochondria and calcium signaling. Cell Calcium 38:311–317.

104

Nijman IJ, Lenstra JA, Schwerin M, Olsaker I. 1996. Polymorphisms and physical locations of

three bovine microsatellite loci: IOBT395, IOBT528, IOBT1401. Anim Genet 27:221–

222.

Nowak R. 1999. Walker’s mammals of the world, Vol II, 6th edition. Baltimore and London:

The John Hopkins University Press.

Nowell K, Jackson P. 1996. Wild Cats. Status Survey and Conservation Action Plan. Gland:

IUCN/SSC Cat Specialist Group.

O’Brien SJ, Roelke M. 1990. What does geology have to do with panther conservation? Natl

Geogr Res 7: 117–119.

O’Brien SJ, Roelke ME, Yuhki N, Richards KW, Johnson WE, Franklin WL, Anderson AE,

Bass OL, Belden RC, Martin JS. 1990. Genetic introgression within the Florida panther

Felis concolor coryi. Natl Geogr Res 6:485–494.

O’Grady JJ, Brook BW, Reed DH, Ballou JD, Tonkyn DW, Frankham R. 2006. Realistic levels

of inbreeding depression strongly affect extinction risk in wild populations. Biol Conserv

133:42–51.

Ochoa A, Onorato DP, Fitak R, Roelke-Parker ME, Culver M. 2017. Evolutionary and functional

mitogenomics associated with the genetic restoration of the Florida panther. J Hered 1–7.

doi:10.1093/jhered/esx015.

Onorato D, Belden C, Cunningham M, Land D, McBride R, Roelke M. 2010. Long-term

research on the Florida panther (Puma concolor coryi): historical findings and future

obstacles to population persistence. In: Macdonald D, Loveridge A, editors. Biology and

Conservation of Wild Felids. Oxford: Oxford University Press. p. 453–469.

105

Pagacova E, Cernohorska H, Kubickova S, Vahala J, Rubes J. 2011. Centric fusion

polymorphism in captive animals of family Bovidae. Conserv Genet 12:71–77.

Parra G, Bradnam K, Korf I. 2007. CEGMA: a pipeline to accurately annotate core genes in

eukaryotic genomes. Bioinformatics 23:1061–1067.

Pasterkamp RJ, Kolk SM, Hellemons AJCGM, Kolodkin AL. 2007. Expression patterns of

semaphoring 7A and plexin C1 during rat neural development suggest roles in axon

guidance and neuronal migration. BMC Dev Biol 7:98.

Pasterkamp RJ, Peschon JJ, Spriggs MK, Kolodkin AL. 2003. Semaphorin 7A promotes axon

outgrowth through integrins and MAPKs. Nature 424:398–405.

Peakall R, Smouse PE. 2006. GENALEX 6: genetic analysis in Excel. Population genetic

software for teaching and research. Mol Ecol Notes 6:288–295.

Peakall R, Smouse PE. 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic

software for teaching and research–an update. Bioinformatics 28:2537–2539.

Pearce-Kelly P, Mace GM, Clarke MD. 1995. The release of captive bred snails (Partula

taeniata) into a semi-natural environment. Biodivers Conserv 4:645–663.

Pelletier F, Reale D, Watters J, Boakes EH, Garant D. 2009. Value of captive populations for

quantitative genetics research. Trends Ecol Evol 24:263–270.

Pesole G, Gissi C, De Chirico A, Saccone C. 1999. Nucleotide substitution rate of mammalian

mitochondrial genomes. J Mol Evol 48:427–434.

Pollak JP, Lacy RC, Ballou JD. 2002. Population management 2000, version 1.16. Brookfield:

Chicago Zoological Society.

106

Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using multilocus

genotype data. Genetics 155:945–959.

Rambaut A, Suchard MA, Xie D, Drummond AJ. 2014. Tracer v1.6. Available at

http://beast.bio.ed.ac.uk/Tracer.

Rand DM, Kann LM. 1998. Mutation and selection at silent and replacement sites in the

evolution of animal mitochondrial DNA. Genetica 102:393–407.

Raymond M, Rousset F. 1995. GENEPOP (version 1.2): population genetics software for exact

tests and ecumenicism. J Heredity 86:248–249.

Renan S, Greenbaum G, Shahar N, Templeton AR, Bouskila A, Bar-David S. 2015. Stochastic

modelling of shifts in allele frequencies reveals a strongly polygynous mating system in

the re-introduced Asiatic wild ass. Mol Ecol 24:1433–1446.

Rice WR. 1989. Analyzing tables of statistical tests. Evolution 43:223–225.

Rimmer A, Phan H, Mathieson I, Iqbal Z, Twigg SRF, WGS500 Consortium, Wilkie AOM,

McVean G, Lunter G. 2014. Integrating mapping-, assembly- and haplotype-based

approaches for calling variants in clinical sequencing applications. Nature Genet 46:912–

918.

Roelke ME, Martenson JS, O’Brien SJ. 1993. The consequences of demographic reduction and

genetic depletion in the endangered Florida panther. Curr Biol 3:340–350.

Roth TL, Bush LM, Wildt DE, Weiss RB. 1999. Scimitar-horned oryx (Oryx dammah)

spermatozoa are functionally competent in a heterologous bovine in vitro fertilization

system after cryopreservation on dry ice, in a dry shipper, or over liquid nitrogen vapor.

Biol Reprod 60:493–498.

107

Rousset F. 2008. GENEPOP’007: a complete reimplementation of the GENEPOP software for

Windows and Linux. Mol Ecol Resour 8:103–106.

Rutschmann F. 2005. Bayesian Molecular Dating Using PAML/multidivtime. A Step-By-Step

Manual. Zurich: University of Zurich. Available at www.plant.ch.

Schook MW, Wildt DE,Weiss RB, Wolfe BA, Archibald KE, Pukazhenthi BS. 2013.

Fundamental studies of the reproductive biology of the endangered Persian onager

(Equus hemionus onager) result in first wild equid offspring from artificial insemination.

Biol Reprod 89:1–13.

Schulte-Hostedde AI, Mastromonaco GF. 2015. Integrating evolution in the management of

captive zoo populations. Evol Appl 8:413–422.

Schwartz MK, Mills LS. 2005. Gene flow after inbreeding leads to higher survival in deer mice.

Biol Conserv 123:413–420.

Seal US. 1994. A plan for genetic restoration and management of the Florida panther (Felis

concolor coryi). Report to the Florida Game and Freshwater Fish Commission. Apple

Valley: Conservation Breeding Specialist Group.

Simpson JT, Wong K, Jackman SD, Schein JE, Jones SJM, Birol I. 2009. ABySS: A parallel

assembler for short read sequence data. Genome Res 19:1117–1123.

Smit AFA, Hubley R, Green P. 1996–2010. RepeatMasker Open-3.0.

http://www.repeatmasker.org.

Stanke M, Tzvetkova A, Morgenstern B. 2006. AUGUSTUS at EGASP: using EST, protein and

genomic alignments for improved gene prediction in the human genome. Genome Biol 7

Suppl 1: S11.1–18.

108

Stanley-Price MR. 1989. Animal re-introductions: the Arabian oryx in Oman. Cambridge:

Cambridge University Press.

Stewart DRM. 1963. The Arabian oryx (Oryx leucoryx Pallas). Afr J Ecol 1:103–117.

Sunquist M, Sunquist F. 2002. Wild Cats of the World. Chicago: University of Chicago Press.

Suto F, Ito K, Uemura M, Shimizu M, Shinkawa Y, Sanbo M, Shinoda T, Tsuboi M, Takashima

S, Yagi T, Fujisawa H. 2005. Plexin-A4 mediates axon-repulsive activities of both

secreted and transmembrane semaphorins and plays roles in nerve fiber guidance. J

Neurosci 25:3628–3637.

Suyama M, Torrents D, Bork P. 2006. PAL2NAL: robust conversion of protein sequence

alignments into the corresponding codon alignments. Nucleic Acids Res 34:609–612.

Talavera G, Castresana J. 2007. Improvement of phylogenies after removing divergent and

ambiguously aligned blocks from protein sequence alignments. Syst Biol 56:564–577.

Tallmon DA, Koyuk A, Luikart G, Beaumont MA. 2008. ONeSAMP: a program to estimate

effective population size using approximate Bayesian computation. Mol Ecol Res 8:299–

301.

Tallmon DA, Luikart G, Waples RS. 2004. The alluring simplicity and complex reality of

genetic rescue. Trends Ecol Evol 19:489–496.

Tamura K, Dudley J, Nei M, Kumar S. 2007. MEGA4: Molecular Evolutionary Genetics

Analysis (MEGA) software version 4.0. Mol Biol Evol 24:1596–1599.

Thorne JL, Kishino H. 2002. Divergence time and evolutionary rate estimation with multilocus

data. Syst Biol 51:689–702.

109

Tzika AC, Remy C, Gibson R, Milinkovitch MC. 2009. Molecular genetic analysis of a captive-

breeding program: the vulnerable endemic Jamaican yellow boa. Conserv Genet 10:69–

77.

U.S. Fish and Wildlife Service. 1973. U.S. Endangered Species Act, FWS-F-037. van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. 2004. MICRO-CHECKER: software

for identifying and correcting genotyping errors in microsatellite data. Mol Ecol

Notes 4:535–538. vonHoldt BM, Pollinger JP, Earl DA, Knowles JC, Boyko AR, Parker H, Geffen E, Pilot M,

Jedrzejewski W, Jedrzejewska B, Sidorovich V, Greco C, Randi E,Musiani M, Kays K,

Bustamante CD, Ostrander EA, Novembre J, Wayne RK. 2011. A genome-wide

perspective on the evolutionary history of enigmatic wolf-like canids. Genome Res

21:1294–1305.

Wallace DC. 1986. Mitochondrial genes and disease. Hosp Pract 21:77–92.

Wallace DC. 2007. Why do we still have a maternally inherited mitochondrial DNA? Insights

from evolutionary medicine. Annu Rev Biochem 76:781–821.

Wang DP, Wan HL, Zhang S, Yu J. 2009a. γ-MYN: a new algorithm for estimating Ka and Ks

with consideration of variable substitution rates. Biol Direct 4:20.

Wang D, Zhang S, He F, Zhu J, Hu S, Yu J. 2009b. How do variable substitution rates influence

Ka and Ks calculations? Genomics Proteomics Bioinformatics 7:116–127.

Wang D, Zhang Y, Zhang Z, Zhu J, Yu J. 2010. KaKs_Calculator 2.0: a toolkit incorporating

gamma-series methods and sliding window strategies. Genomics Proteomics

Bioinformatics 8:77–80.

110

Webb SD, Rancy A. 1996. Late Cenozoic evolution of the neotropical mammal fauna. In:

Jackson JBC, Budd AF, Coates AG (editors). Evolution and environment in tropical

America. Chicago: University of Chicago Press.

Whiteley AR, Fitzpatrick SW, Funk WC, Tallmon DA. 2015. Genetic rescue to the rescue.

Trends Ecol Evol 30:42–49.

Willoughby JR, Fernandez NB, Lamb MC, Ivy JA, Lacy RC, Dewoody JA. 2015. The impacts

of inbreeding, drift and selection on genetic diversity in captive breeding populations.

Mol Ecol 24:98–110.

Wilson D, Reeder D. 1993. Mammal species of the world: a taxonomic and geographic

reference, 2nd edition. Washington and London: Smithsonian Institution Press.

Witzenberger KA, Hochkirch A. 2011. Ex situ conservation genetics: a review of molecular

studies on the genetic consequences of captive breeding programmes for endangered

animal species. Biodivers Conserv 20:1843–1861.

Woolley S, Johnson J, Smith MJ, Crandall KA, McClellan DA. 2003. TreeSAAP: selection on

amino acid properties using phylogenetic trees. Bioinformatics 19:671–672.

Xia C, Cao, J, Zhang H, Gao X, Yang W, Blank D. 2014. Reintroduction of Przewalski’s horse

(Equus ferus przewalskii) in Xinjiang, China: the status and experience. Biol Conserv

177:142–147.

Yang Z. 1997. PAML: a program package for phylogenetic analysis by maximum likelihood.

Bioinformatics 13:555–556.

Yang Z. 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol 24:1586–

1591.

111

Young JM, Friedman C, Williams EM, Ross JA, Tonnes-Priddy L, Trask BJ. 2002. Different

evolutionary processes shaped the mouse and human olfactory receptor gene families.

Hum Mol Genet 11:535–546.

Zhang X, Firestein S. 2002. The olfactory receptor gene superfamily of the mouse. Nature

Neurosci 5:124–133.

Zhen Y, Hoganson CW, Babcock GT, Ferguson-Miller S. 1999. Definition of the interaction

domain for cytochrome c on cytochrome c oxidase. I. Biochemical, spectral, and kinetic

characterisation of surface mutants in subunit II of Rhodobacter sphaeroides cytochrome

aa3. J Biol Chem 274:38032–38041.

Zhou H, Li D, Zhang Y, Yang T, Liu Y. 2007. Genetic diversity of microsatellite DNA loci of

Tibetan antelope (Chiru, Pantholops hodgsonii) in Hoh Xil National Nature Reserve,

Qinghai, China. J Genet Genomics 34:600–607.

112