Novel Molecular and Genomic Aspects of Elephant Endotheliotropic Herpesvirus (Eehv)

NOVEL MOLECULAR AND GENOMIC ASPECTS OF ELEPHANT ENDOTHELIOTROPIC HERPESVIRUS (EEHV)

by Simon Y. Long, MLAS, VMD

A dissertation submitted to John Hopkins University in conformity with the requirements for the degree of Doctor of Philosophy

Baltimore, Maryland

December, 2016

Abstract

Two variants of elephant endotheliotropic herpesviruses (EEHV) 1A and 1B cause a fatal hemorrhagic disease that affects young Asian elephants (Elephas maximus) compromising the future long-term survival of this endangered species both in captive and wild populations worldwide. Since its discovery in 1999, more than 100 young captive and wild Asian elephants are known to have died from EEHV. For reasons that are not yet understood, approximately 20% of Asian elephant calves appear to be susceptible to the disease when primary infections are not controlled by normal innate cellular and humoral immune response. Recent crossed EEHV1 virus from African elephant hosts (Loxodonta cyclotis and Loxodonta africana) to Asian elephants in captivity was previously suggested as the cause of disease. This argument was disproved by our field work in India demonstrating the high level of EEHV1 subtype genetic diversity found among the orphan camp and wild Asian elephant calf cases showing the

Indian strains matches that among over 30 EEHV1 strains that have been evaluated from

Europe and North America. Over the years, multiple species of herpesviruses were discovered with three different lineages forming the Proboscivirus genus (EEHV1/6,

EEHV2/5, and EEHV3/4/7) infecting both Asian and African elephants with lethal hemorrhagic disease confined to EEHV1 in Asian elephant calves. Milder disease caused by EEHV5 and EEHV4 were increasingly being recognized during EEHV1 surveillance in Asian elephants. The complete 180-kb genomes of prototype strains from three AT- rich branch viruses, EEHV1A, EEHV1B, and EEHV5, have been published with hints of the second major branch of GC-rich viruses (EEHV3, EEHV4, and EEHV7). The complete 206-kb genome of EEHV4 was determined directly from a trunk wash sample

ii using next-generation sequencing and de novo assembly procedures revealing a genome composed of 119 genes with an overall collinear organization similar to the AT-rich

EEHV with major features including a family of 26 paralogous 7xTM and vGPCR-like genes plus 25 novel or missing genes. Many fundamental questions about the mechanism of pathogenesis and appropriate effective drug treatment for EEHVs have remained poorly understood because of the inability of scientists to grow the virus in cell culture.

Successfully propagation of EEHV in laboratory conditions will permit critical future experimental research for progress towards understanding the pathology and cell biology of this virus, as well as for any hoped for possibility of generating live attenuated vaccines. I report here the first elephant umbilical venule endothelial cell (EUVEC) cultures overlay with fresh minced Asian elephant necropsy tissue samples from a confirmed EEHV1A associated death with persistence EEHV1A viral DNA PCR detection for several months in spite of multiple cell culture passages and medium changes. All these aspect of research in EEHV adds to the general knowledge in hopes to help recover the 20% of all captive bred calves that are currently lost from this disease, thus increasing the chances of survival of this species considering all of the enduring conservation associated pressures it is under.

iii Thesis advisor: Gary S. Hayward, PhD

Professor, Dept. of Oncology

Second reader: Joseph L. Mankowski, DVM, PhD, Dipl. ACVP

Professor, Dept. of Molecular and Comparative Pathobiology

Committee: Joseph L. Mankowski, DVM, PhD, Dipl. ACVP (Chair)

Wade Gibson, PhD

J. Marie Hardwick, PhD

iv Acknowledgements

First and foremost, immense thanks to my thesis advisor Dr. Gary Hayward for taking me on as a PhD student under a challenging thesis research theme studying elephant endotheliotropic herpesvirus combining both herpesvirus virology and veterinary pathology. I would like to sincerely thank my thesis committee chair and director of Molecular Comparative Pathobiology (MCP), Dr. Joseph Mankowski for keeping me on track not only for my PhD training but also my veterinary pathology postdoctoral fellowship under MCP. I am thankful for the rest of my thesis committee including Dr. Wade Gibson and Dr. Marie Hardwick, for having been a huge help. I appreciate all the advice and encouragement that they have provided to me over the last few years. I am also grateful to the Cellular and Molecular Medicine Graduate program especially to Dr. Rajini Rao, Colleen Graham, and last but not least Leslie Lichter who all helped me get through some difficult times during my PhD training. Without their consistent support, I’m not sure I would have completed my training.

The entire elephant community is so supportive of this research and I thank all the zookeepers, veterinarians, and veterinary pathologists that have contributed time and samples that have be invaluable to this research. I would also like to thank the folks at the National Zoo for travel support and especially to Erin Latimer of the National

Elephant Herpesvirus Laboratory for samples and technical support. The following laboratory members of the Hayward laboratory have been invaluable resources in terms of advice and technical support: Sarah Heaggans, Colette apRhys, Jianchao Zong, Yong-

Gu Chi, and Chuang-Jiun Chiou. Most importantly, I would like to thank the Morris

v Animal Foundations (D14ZO-411), International Elephant Foundation, and NIH NIAID for funding this project and fellowship.

I would also like to extend my thanks to all the faculty members and trainees in the MCP department in playing a major role in my training to become a veterinary pathologist. Again, I am deeply thankful for Dr. Mankowksi’s advice and guidance both to me as a PhD student and as a budding veterinary pathologist. To my veterinary pathology mentor, Dr. Cory Brayton, for keeping me on track, constantly reminding me about the American College of Veterinary Pathology board exam, and keeping me focused on my career goals. To the rest of the faculty including Dr. Christine Zink, Dr.

Richard Montali, Dr. Baktiar Famir, Dr. Kathleen Gabrielson, and the late Dr. David

Huso. Many thanks to the pathology postdoctoral fellows Katie Kelly, Gillian Shaw, and

Sarah Beck who trained me when I first started the residency program. Also to all the junior pathology postdoctoral fellows and residents after me.

vi Table of Contents

Title Page…………………………………………………………………………………..i

Abstract …………………………………………………………………………………...ii

Acknowledgments ………………………………………………………………………...v

Table of Contents...………………………………………………………………………vii

List of Tables …………………………………………………………………………...viii

List of Figures ……………………………………………………………………………ix

List of Abbreviations …………………………………………………………………….xi

I. Introduction...... 1

II. Fatal herpesvirus hemorrhagic disease in wild and orphan……………………….8

Asian elephants in southern India

III. Complete genome sequence of EEHV4 as the first example……………………35

of a GC-rich branch Proboscivirus.

IV. Propagating elephant endotheliotropic herpesvirus 1A (EEHV1A)……………..83

in primary and immortalized elephant umbilical venule endothelial cells

(EUVEC)

References………………………………………………………………………………105

Curriculum Vitae……………………………………………………………………….109

vii List of Tables

Chapter II

Table 2-1. Summary of features and PCR results for nine positive……………...24

cases of EEHV disease from India.

Chapter III

Table 3-1. Gene content and major features of the complete……………………67

205,896-bp EEHV4 (Baylor) genome.

viii List of Figures

Chapter II

Figure 2-1. Pathological changes of EEHV-associated disease………………….25

found during field necropsy.

Figure 2-2. Representative agarose gel-ethidium bromide PCR………………...26

band results obtained after first round (520-bp) and second round semi-nested

(250-bp) amplification with EEHV PAN-POL primers from 12 samples of

necropsy tissue DNA.

Figures 2-3. Phylogenetic distance based dendrograms showing………………..27

patterns of DNA level divergence and subtyping amongst the Indian EEHV1

cases.

Figure 2-4. Graphical presentation generated from Geneious……………...……28

of the patterns of DNA sequence polymorphisms across the four most variable

PCR loci.

Figure 2-5. Comparison of the major EEHV1 gene subtype………………….…30

distribution patterns amongst Indian, North American and European hemorrhagic

disease cases at the two most variable PCR loci examined U51/vGPCR1 and

U48/gH-TK.

Figure 2-6. Direct side-by-side nucleotide level differences………………….....31

observed amongst the nine Indian EEHV1 cases at all six PCR loci.

ix Chapter III

Figure 3-1. Annotated physical gene map of the complete…………………..….71

EEHV4 (Baylor) genome.

Figure 3-2. Global alignment patterns for the intact EEHV4……………………73

genome compared to EEHV1 and HCMV.

Figure 3-3. EEHV4 encodes a very large family of distantly……………………75

related paralogous 7XTM and vGPCR-like genes.

Figure 3-4. Codon-specific scanning GC-content panels………………………..77

showing the wobble codon GC-bias effect across selected representative segments

of the EEHV4 (Baylor) genome.

Figure 3-5. Complex repeat and inverted repeat patterns………………………..79

within the predicted Ori-Lyt domain of EEHV4.

Figure 3-6. Positions and sizes of three identified……………………………….81

EEHV4A-EEHV4B chimeric domains and boundaries relative to those of

EEHV1A-EEHV1B.

Chapter IV

Figure 4-1. Microphotograph of pEUVEC……………………...……………...101

Figure 4-2 Agarose gel-ethidium bromide of PAN-HEL PCR…………...…….102

Figure 4-3. DNA sequence alignment between detected EEHV…………..…...103

HEL from heart iEUVEC and whole blood necropsy sample.

Figure 4-4. Levels of VGE in heart pEUVEC between estimated……………..104

compared to qPCR detected based on the number of passages.

x List of Abbreviations

CD Chimeric domain

CRE Cyclic AMP-response elements

E34 Open reading frame C

EEHV Elephant endotheliotropic herpesvirus

EUVEC Elephant umbilical venule endothelial cells gB Glycoprotein B gH Glycoprotein H gM Glycoprotein M gN Glycoprotein N gO Glycoprotein O

HCMV Human cytomegalovirus

Hel Helicase hTERT Human telomerase reverse transcriptase

HSV1 Herpes simplex virus 1 iEUVEC Immortalized EUVEC

IFA Immunofluorescence assay

KSHV Kaposi’s sarcoma-associated herpesvirus

LANA Latency associated nuclear antigen

LM-PCR Ligation-mediated polymerase chain reaction

MCP Major capsid protein

MDBP Major DNA binding protein

MIE1 Major immediate-early enhancer domain 1

xi MIE2 Major immediate-early enhancer domain 2

MTA mRNA transcript accumulation

NAP North American Proboscivirus nBA n-butyric acid

OBP Origin binding protein

OBS OBP-binding sites

ORF Open reading frame

PBMC Peripheral blood mononuclear cell

PBS Phosphate-buffered saline pEUVEC Primary EUVEC

PEL Primary effusion lymphoma

PCR Polymerase chain reaction

POL DNA Polymerase qPCR Quantitative polymerase chain reaction

RAIP3 Retinoic acid induced protein 3

RCMV Rat cytomegalovirus

RRB Ribonucleotide kinase B

SSP Single-stranded DNA-binding protein

SV40TL Simian vacuolating virus 40 large T antigen

TER Terminase

TK Thymidine kinase

TPA Tetradecanoyl phorbol acetate vCXCL1 Viral C-X-C motif chemokine ligand 1

xii vECTL Viral C-type lectin vFUT9 Viral fucosyl transferase 9

VGE Viral genome equivalent vGCNT1 Viral Glucosaminyl (N-acetyl) transferase 1 vGPCR Viral G-protein-coupled receptor vICA Viral inhibitor of caspase-8-induced apoptosis protein vMIP Viral macrophage inflammatory protein vOGT Viral O-linked acetyl glucosamine transferase

xiii

I. Introduction

1 EEHV Background

Wildlife conservationists are concerned about the critically endangered Asian

(Elephas maximus) and African elephant (Loxodonta cyclotis and Loxodonta africana) species due to their alarming decline over the last decade. These declines arise from both old and current challenges to these species as a whole. Each species faces similar yet very different unique problems including habitat loss and human encroachment into natural ranges resulting in human-elephant conflict. In the last several years, recent surges of highly organized crime syndicates poaching for ivory in Africa has severely depleted the population of African elephants overall. All these factors threaten elephant populations, forcing wildlife conservationists to look for alternative avenues including captive breeding programs to promote species survival. Elephant endotheliotropic herpesviruses

(EEHV) that cause a fatal hemorrhagic disease that affects young Asian elephant calves compromises the survival of this species both in captivity and in the wild populations worldwide and hinders these conservation efforts.

Since the report of the discovery of the first novel elephant endotheliotropic herpesvirus 1A (EEHV1A) in 19951, eight different species of EEHV within the novel

Proboscivirus genus have been identified between Asian and one of the two African elephant species 2-7. The viruses under this genus can be divided based on AT-rich

(EEHV1A/B, EEHV2, EEHV5 and EEHV6) vs. GC-rich (EEHV3, EEHV4, and EEHV7) sequences. The majority of these virus species appear to be common endemic subclinical universal infections regardless of location. However, EEHV1A and EEHV1B are highly pathogenic species that cause over 90% of the 100 known disease cases8 exclusively in

Asian elephants with the original hypothesis suggesting a recent crossing from African

2 elephants in captivity. The remaining 10% of EEHV associated disease cases are from other virus species.

After more than two decades of study, numerous fundamental questions still remain unanswered. Initially, it was hypothesized that cross-species transmission of

EEHV1 from African elephants to Asian elephants in captivity scenarios may explain the severe pathology1,9. Herpesviruses generally coevolve with their hosts to be asymptomatic in immune competent native hosts while cross-species transmission can result in pathology in the non-native host. This hypothesis offered an explanation for the severe fatal hemorrhagic pathology exhibited by young Asian elephants. Full genomic sequence of EEHV1A/B and EEHV5 representing the AT-rich branch of the

Proboscivirus is available yet the full genomic sequence of a GC-rich example is still lacking and represents a gap in knowledge which may provide further insight into a role for this unique group of viruses in the pathology. The inability to grow EEHV in cell culture is another major challenge in EEHV research; past attempts to cultivate the virus using standard methods in established cell lines have been unsuccessful.

Detection of EEHV1 in free ranging Asian elephants in India.

It was hypothesized that the fatal hemorrhagic disease caused by EEHV1 in Asian elephants may be due to cross species transmission from African elephants to Asian elephants in captivity. Using conventional PCR techniques, EEHV1 was previously detected in an archived nodule from an African elephant trunk, but these results could not be duplicated and thus were suggestive of a false positive due to contamination. Recent attempts to detect EEHV1 from additional African elephant samples with PCR have also

3 failed to detect EEHV1, arguing against the cross species transmission hypothesis.

Detection of EEHV1 in wild Asian elephants in their natural range would further disprove the hypothesis that EEHV1 only causes pathology following cross species transmission. Prior to our published 2012 fieldwork10, there has only been one single report of a PCR positive EEHV1 associated death in Cambodia by Reid11 in 2006 with additional anecdotal case reports all throughout the natural range of Asian elephants. The published work that I present in chapter 2 reports 9 different EEHV1-associated deaths in free-ranging Asian elephant calves in southern India. These deaths are the result of 8 unique strains of EEHV1A with wide genetic diversity in two hypervariable genetic loci that is similar to the diversity noted in both North American and European cases of fatal

EEHV1 infection. This finding disproves the cross species transmission hypothesis by confirming the endemic nature of EEHV1 in wild Asian elephants with an additional later report in Laos12, and unpublished lethal cases confirmed by PCR DNA sequencing in

Sumatra (two) and Nepal (one).

Complete genome sequence of EEHV4 as the first example of a GC-rich branch proboscivirus.

Multiple species of herpesviruses from three different lineages of the

Proboscivirus genus (EEHV1/6, EEHV2/5, EEHV3/4/7) infect both Asian and African elephants, but lethal hemorrhagic disease is largely confined to Asian elephant calves and is predominantly associated with EEHV1. EEHV2, EEHV3, EEHV6, and EEHV7 have been identified in African elephants7. EEHV1A, EEHV1B, EEHV3, EEHV4 and EEHV5 have been responsible for the deaths of Asian elephant calves3,4,13. Mild disease caused

4 by EEHV52,4,14 or EEHV415 is increasingly identified in Asian elephants, but little is known about the latter. EEHV4 is estimated to have diverged at least 35 million years ago from the others within a distinctive GC-rich branch of the Probosciviruses. The complete 180-kb genomes of prototype strains from three AT-rich branch viruses,

EEHV1A16,17, EEHV1B16, and EEHV514, have been published with hints of the second major branch of GC-rich viruses (EEHV3, EEHV4, and EEHV7)6,7. The death of a Thai

Asian elephant from EEHV4 hemorrhagic disease18 and disease episodes in the USA involving EEHV3 and EEHV43, as well as detection of EEHV3B and EEHV7B in skin nodules, emphasizes the need for a greater understanding of other species within the highly diverged GC-rich subgroup of EEHVs. The very high level of divergence of both

EEHV3 and EEHV4 compared to the other EEHVs deserves a further study so that we can gain understanding of this group of viruses.

Recent reports in Asian elephant calves at the Houston Zoo noted mild symptoms with detection of EEHV4 using PCR from blood and later trunk wash samples15. In collaboration with Paul Ling from Baylor College of Medicine, next generation Illumina sequencing data directly yielded contigs representing 80% of the genome from a trunk wash pellet from two different Asian elephant calves who were infected with the same strain of EEHV4 at a Houston Zoo in 2014. The remaining 20% of the genome was filled and joined by me using standard Sanger PCR amplification and cycle sequencing on the same trunk wash pellet sample. Chapter 3 documents the unique features of this GC-rich

EEHV4 genome includes 119 encoded genes arranged in similar co-linear organization to that of the AT-rich branch of viruses, a family of 26 paralogous 7Xtransmembrane and viral G-coupled protein receptor-like genes, 25 novel or missing genes with an unusual

5 distribution of tracts of 5-11 successive A or T nucleotides in intergenic domains between the GC-rich protein coding regions with a high GC-rich bias in the third wobble positions of codons for many proteins, and two novel captured cellular genes. Comparison of the sequence of this pathogenic subspecies of EEHV4 with the prototype EEHV1A genome from the AT-rich branch will lead to important insights into the common proteins shared between these two anciently diverged lineages of elephant herpesviruses, and may lead to the discovery of novel virulence factors.

Attempts to propagate elephant endotheliotropic herpesvirus 1A (EEHV1A) in primary and immortalized elephant umbilical venule endothelial cells (EUVEC).

The inability to grow the virus in cell culture is the major roadblock in EEHV research. If successful viral propagation in cell culture could be achieved, researchers would gain the ability to (a) accumulate large quantities of virus for easier molecular analysis of the novel genes and proteins that these viruses encode, (b) carry out pharmacologic studies to evaluate the efficacy of human anti-herpesviral drugs in inhibiting the replication of elephant herpesviruses (whether currently in use or in development); and finally (c) passage attenuated virus promote a vaccine development.

Within the last few years, our lab has established techniques to culture primary elephant umbilical venule endothelial cells (pEUVEC) and immortalized EUVEC (iEUVEC) from

Asian elephant umbilical cords. A recent experiment that I performed with fresh necropsy tissue samples from an elephant who died due to confirmed EEHV1 infection resulted in detection of EEHV1A DNA in both pEUVEC and iEUVEC cultures that persisted for 54 and 122 days, respectively. This finding is suggestive of viral cellular entry and

6 persistence retention of virus in latent phase in culture rather than in lytic replicating phase. Chapter 4 reports the findings of this attempt to propagate EEHV1A in primary and immortalized EUVEC cultures.

II. Fatal herpesvirus (EEHV) hemorrhagic disease in wild and orphan Asian elephants in southern India.

This chapter has been previous published as:

Zachariah A, Zong JC, Long SY, Latimer EM, Heaggans SY, Richman LR, Hayward GS. Fatal herpesvirus (EEHV) hemorrhagic disease in wild and orphan Asian elephants in southern India. Journal of Wildlife Diseases. 2013. 49(2):381-393

Reprint with permission of the publisher, Wildlife Disease Association.

8 Abstract

Up to 65% of the deaths of young Asian elephants between 3 months and 15 years of age in Europe and North America over the past twenty years have been attributed to hemorrhagic disease associated with a novel DNA virus called Elephant Endotheliotropic

Herpesvirus (EEHV). To evaluate the potential role of EEHV in suspected cases of a similar lethal acute hemorrhagic disease occurring in Southern India, we studied pathological tissue samples collected from field necropsies. Nine convincingly positive cases amongst both orphaned camp and wild Asian elephants were identified by diagnostic PCR. These were then subjected to detailed gene subtype DNA sequencing at multiple PCR loci, which revealed seven distinct strains of EEHV1A and one of

EEHV1B. Two orphan calves that died within three days of one another at the same training camp proved to have identical EEHV1A DNA sequences to one another indicating a common epidemiological source. However, the high level of EEHV1 subtype genetic diversity found amongst the other Indian strains matches that amongst over 30

EEHV1 strains that have been evaluated from Europe and North America. These results argue strongly against the previously suggested notion that this is just a disease of captive elephants and that the EEHV1 virus has crossed recently from African elephant

(Loxodonta africana) hosts to Asian elephants (Elephas maximus). Instead, both the virus and the disease are evidently widespread in Asia and, despite the disease severity, Asian elephants themselves appear to be the ancient endogenous hosts of both EEHV1A and

EEHV1B.

9 Introduction

An acute hemorrhagic disease with 85% fatality rate was originally described in captive Asian elephants in both North American and European zoos 1,19 ,20-22. Amongst 52 suspected cases since 1988, the presence of high levels of a novel type of herpesvirus called Elephant Endotheliotropic Herpesvirus or EEHV has been confirmed by DNA

PCR tests carried out at the National Elephant Herpesvirus Laboratory in Washington,

DC. on blood or necropsy tissue from all 28 North American cases and from 10 of the 24

European cases (3,4, Richman LK et al unpublished data). However, these viruses have never been detected in numerous random unrelated elephant necropsy tissue samples nor in blood from healthy asymptomatic elephants. The PCR test results show that EEHV disease occurs predominantly in Asian juveniles between one and eight years of age and that here have been only ten known survivors, all of which received timely and aggressive treatment with the human anti-herpesvirus drugs famciclovir or ganciclovir

23,24. Initial mild symptoms of facial edema and tongue cyanosis progress rapidly to systemic viremia followed by hemorrhaging in all major internal organs 9. Nuclear inclusion bodies and enveloped virions can be observed by light and electron microscopy respectively in microvascular endothelial cells and extremely high levels of viral DNA are found in white blood cells, serum and necropsy tissue (4; Zong J-C et al, unpublished data).

Most herpesviruses have become highly adapted to the natural hosts in which they have evolved during mammalian evolution and usually cause only mild localized disease in healthy animals. Therefore, the original description of a herpesvirus associated with rapid acute hemorrhagic disease in juvenile Asian elephants was unexpected 1. However,

10 the idea that it might involve a cross-species transfer of a virus that was adapted to

African elephants into naïve Asian elephants seemed to make sense of the unexpectedly severe pathology. Originally, two distinct species of EEHV were described: EEHV1 was associated with 11 cases in captive Elephas maximus, whereas EEHV2 was associated with the deaths of two young Loxodonta africana 1. The EEHV1 viruses found in both those and all subsequent European and North American cases fall into two subgroups referred to as EEHV1A and EEHV1B 21,25, which are now known to represent two partially chimeric sub-species 22. EEHV1 and EEHV2 differ by an average of 25% at the nucleotide level all the way across their genomes, whereas EEHV1A and EEHV1B differ by between 15 and 55% at several small loci, but by no more than 1 to 2% over other large segments of their genomes 25,26. Two more highly diverged species EEHV3 and

EEHV4 (35% different) were later identified in two other lethal cases of hemorrhagic disease in North American Asian elephant calves 3. Together with EEHV5 or EEHV6 4 as well as EEHV7 (Zong J-C et al, unpublished data), which have not yet been associated with any lethal disease cases, these multiple species of elephant herpesviruses form a novel Proboscivirus genus that is most closely related to the Roseoloviruses within the

Betaherpesvirinae sub-family27. The Probosciviruses evidently diverged from all other mammalian herpesviruses over 100 million years ago when the ancestors of modern elephants diverged from all other placental mammals. All of the positive captive elephant cases have been subjected to EEHV gene subtyping analysis, which has identified more than 30 distinct strains of EEHV1A and six of EEHV1B 22,28 (Zong J-C et al, unpublished data), indicating that this is primarily a sporadic not epidemic disease. Those studies have also shown that there has been no direct chain of transmission between cases at different

11 elephant housing facilities, nor in most cases even between multiple afflicted progeny of a single breeding bull or cow elephant. Furthermore, recent qualitative real-time PCR assays for EEHV1 have revealed that many asymptomatic Asian elephant herdmates of calves that died or survived viremic disease at the same North American housing facilities carry and periodically shed either EEHV1A, EEHV1B or EEHV5 in trunk secretions 29,30.

We and others originally speculated that EEHV disease in Asian zoo elephants might be associated with opportunities for direct or indirect contact with and transmission from African zoo elephants 1,20. However, some cases occurred at facilities that had never co-housed African elephants, and there have also been a number of anecdotal suggestions of a similar disease occurring in range countries including India, Thailand, Cambodia,

Myanmar and Nepal, with one preliminary published report of an EEHV1A-positive case in Cambodia 11. Furthermore current evidence indicates that only EEHV2, EEHV3,

EEHV6 and EEHV7, but not EEHV1, are found in healthy asymptomatic African elephants (Zong JC et al, unpublished data), whereas only EEHV1A, EEHV1B and

EEHV5 have so far been found in healthy asymptomatic Asian zoo elephants2,29,31,32.

Therefore, we wished to (a) Confirm the presence of EEHV disease in the wild in an

Asian range country; (b) Address the question of which EEHV species might be causing disease there; and (c) Reassess the origins of pathogenic EEHV1.

Materials and Methods

Tissue and DNA Recovery from Field Necropsies

12 Amongst the 15 South India cases examined, two were captive-born privately- owned elephants, six were wild-born juveniles being reared as orphans at working or training camps or in one case at the Arigna Anna Zoological Park, Chennai, and seven were from free-ranging wild herds. The physical locations of the wild cases were distributed relatively uniformly across the entire South Western Ghats, including on both sides of the Palakhad gap which separates two major sub-populations of the elephants in

Southern India, the largest remaining wild population of Elephas maximus in Asia. In most cases, liver or heart necropsy tissue was collected and preserved in alcohol or directly frozen for long-term storage at -45°C. In one positive case, only a viremic blood sample was available without any necropsy tissue and in another only lymph node tissue was available. DNA was extracted from stored frozen necropsy tissue samples after mincing using standard procedures that have been described previously 4.

PCR Amplification

PCR amplification and gel purification procedures and most of the PCR primers used for the U38/POL, U60/TERex3, U71/gM, U77/HEL and U51/vGPCR loci were described in Stanton et al 29,30 or Latimer et al 4. Primers used for the U48/gH-TK locus encompassing the N-terminal part of U48/glyH and the C-terminal end of U48.5 (TK) were as follows:

L1 LGH7981 5’-CT[A/G]CATT[T/G][A/C]CCAAAGTATGGAAGTA-3’

L2 LGH7982 5’-C[G/A]T[C/T]TATATCATCAAA[A/G]AC[C/T]TCACA-3’

R2 LGH7984 5’-CAGCCTTCAAGCGGCATACACTG-3’

R1 LGH7985 5’-GGTAGGTTCACCTACATGGAACTTC-3’

13 First round R1/L1 = 1080-bp; second round A L1/R2 = 920-bp; second round B L2/R1

= 1040-bp; third round A R2/L2 = 860-bp.

Gene Subtype DNA Sequence Analysis

All DNA samples were initially subjected to PCR amplification with first or second-round primers for both PAN-POL and EEHV1-specific DNA polymerase (U38/

POL). DNA samples from the nine cases that consistently gave correctly-sized PCR products were then all further characterized by additional first, second or third-round semi-nested PCR analysis at up to five other previously well-characterized EEHV1 gene loci. We consider that the other six potential cases are still inconclusive at present, possibly because the necropsy tissue DNA available was degraded and of low quality, or because those calves instead died from non-EEHV related causes. Correct sized PCR products were purified after agarose gel electrophoresis with a Qiagen Gel Extraction kit

(Qiagen, Valencia, CA) then subjected to direct cycle sequencing on both strands with the ABI PRISM DigDye Terminator v3.1 cycle sequencing kit and analyzed on an ABI

310 DNA Sequencer (Life technologies Inc, Carlsbad, CA) or at Macrogen, Inc,

Rockville, MD. All DNA sequence editing, analysis and manipulation was performed using Assemblalign, Clustal-W and distance based phylogenetic trees under a neighbor- joining tree program with Tamura Nei logic as implemented in MacVector vers 7

(Symantec Corp. Mountain View, CA). Similar trees prepared after alignment in

MUSCLE and tree building by maximum likelihood in MEGA5 had the identical topography (not shown). Graphical alignments comparing the patterns of sequence

14 polymorphisms were prepared in Geneious, where deviations from consensus are shown in green (A), blue (C), black (G) or red (T).

All of the Indian EEHV1 PCR DNA sequence results described here have been deposited in Genbank under accession Nos. JX011032-JX011038 (U38/POL), JX011039-

JX011046 (U48/gH-TK), JX011047-JX011055 (U51/vGPCR1), JX011056 -JX011062

(U60/TERex3), JX011063-JX011071 (U71/gM), JX011072-JX011079 (U77/HEL). The comparative data used in the Figure 2-3 phylogenetic trees and Figure 2-4 graphical alignments from prototype North American EEHV cases EEHV1A (NAP18, NAP11),

EEHV1B (NAP14, NAP19), EEHV2 (NAP12) and EEHV6 (NAP35) comes from

Genbank Accession Nos HM568528, HM568515, HM568541,HM568556, HM568564 and JN983113 (U71-gM), HM568525, HM568513, HM568538, HM568552 and

JN983120 (vGPCR1), HM568529, HM568515, HM568542, HM568557, HM568564 and

JN983126 (HEL) plus HM568525, HM568513, HM568538, HM568552, HM568561 and

JN983119 (gH-TK) as well as JN633913 (NAP33, vGPCR1), JN633921 (NAP45, vGPCR1) and JQ300058 (NAP42-D, gH-TK). After manual sequence read editing, the assembly, Clustal-W or MUSCLE alignments and distance-based Neighbor-Joining or

Maximum Likelihood phylogenetic tree dendrograms were generated as implemented in

MacVector vers 7.0 or Geneious vers 5.4.6.

Result

The deaths of up to twenty young Asian elephant calves within South India since

1997 have been suspected of being caused by an acute disseminated viral hemorrhagic disease first described in captive zoo and circus elephants 1. Field necropsy examinations

15 were carried out on 15 suspected cases of sudden hemorrhagic deaths in juveniles or sub- adults that came to the attention of the Kerala Forest Department in Southern India between 2007 and 2011. Most revealed symptoms of edema and tongue cyanosis as well as typical hemorrhagic features in the heart and liver consistent with those described previously in young captive Asian elephants that had died suddenly of EEHV disease 9.

Representative examples of gross pathological features, including (A) tongue cyanosis,

(B, C) cardiac and (D) hepatic hemorrhages and (E, F) vascular endothelial cell inclusion bodies from free-range cases of Indian elephant hemorrhagic disease are pictured in

Figure 2-1.

The National Elephant Herpesvirus Laboratory at the National Zoological Park in

Washington DC and the Johns Hopkins University in Baltimore, MD working with the

Kerala Forest Department, have established a diagnostic EEHV laboratory in Wayanad

Wildlife Sanctuary in Kerala. Initial PCR analysis carried out there using both PAN POL and specific EEHV1 POL primers on DNA samples from stored frozen tissues identified nine of the 15 suspected cases tested that that were sufficiently well-preserved and abundant to yield high levels of PCR products that were positive for EEHV1 after either first or second round amplification (Figure 2-2). Features and sources of these nine positive cases are listed in Table 2-1. From our previous semi-quantitative analyses and plasmid DNA reconstruction experiments using either limiting dilution steps with conventional PCR 4 or real-time PCR analysis2010 29, we have determined that strong first round positive bands with PAN-POL EEHV primers from PBMC DNA represent levels of viral DNA ranging from at least 1 million on up to 75 million copies or genome equivalents / ml. Blood and necropsy DNA samples from healthy and known latently

16 infected elephants not undergoing primary or reactivated infections have always been negative even after third round conventional PCR. Our best estimates of the detection limits for first, second, and third round PCR as used here are 375,000, 15,000 and 600 genome equivalents per ml of blood). For whole blood DNA with strong first round PCR products this represents 1 to 75 million genome equivalents per 5 million PBMCs. Whilst we did not have blood samples from the India cases, the necropsy tissue DNA samples used here averaged the same 50 to 75 ug/ml as for the quantitated USA PBMC DNA samples. Therefore, when combined with the evidence for viral inclusion bodies within a small subset of the vascular endothelial cells (assume 1%) in sufficiently well-preserved

Indian case necropsy tissue, we can estimate the total amount of viral DNA per infected endothelial cell of between 20 to 1500 viral genome equivalents per cell, essentially the same as in the captive elephant cases. This value is only compatible with active acute lytic viral replication not latent state infection.

Detailed gene subtype DNA sequencing analysis at multiple PCR loci was then performed yielding unambiguous DNA sequence data for at least four and up to six tested

PCR loci in each case as summarized in Table 2-1. Amongst the first three PCR loci used, U38/POL, U60/TER and U76-U77/HEL all represent well conserved core genes that serve primarily to distinguish unambiguously between the EEHV1A and EEHV1B sub-species, which differ at these loci by 15/480, 11/360 and 20/640 common nucleotides respectively (about 3% each), but otherwise display relatively few additional characteristic polymorphisms. The primers at these loci and U71-gM do also have the potential to identify DNA from most other species of EEHV in the absence of EEHV1, but only EEHV1 was detected.

17 Subsequently, the positive samples were also analyzed with EEHV1-specific primer sets for three strain variable PCR gene loci, encompassing parts of the

U71/myristylated tegument protein plus U72/glycoprotein M region (U71-gM), the G- protein coupled receptor (U51/vGPCR1) and an overlapping glycoprotein-H (U48/gH) plus thymidine kinase (U48.5/TK) region (gH-TK). These latter three loci are much further diverged than the others and each cluster into multiple unlinked subtypes. The results obtained were then compared to those from similar extensive analyses of EEHV genomes carried out previously on most known European and North American hemorrhagic disease cases (Figure 2-3). For U71-gM, the A and B subtypes differ at the

DNA level by an average of 39/750 nucleotides (= 5.3 %) plus some deletions/insertions and there is a rare third intermediate subtype (C) as well. In the case of the vGPCR1 protein, there are five major EEHV1 subtype clusters known (A, B, C, D and E) plus several additional chimeric forms E/A, D2/A or other distinctive variants (A1, A2, B1,

B2, B3, D1 and D2) encompassing a hotspot with up to nine adjacent variable amino acids, including deletions and insertions. Within the hypervariable gH coding segment of the gH-TK locus, all EEHV1B strains differ from the EEHV1As by 35% at the amino acid level, with the EEHV1A strains themselves splitting further into five more subtype clusters (A, C, D, E and F) that differ by between 7 and 15% at the amino acid level.

Overall, the DNA sequencing results obtained revealed that eight of the nine positive Indian cases involved EEHV1A strains, whereas one was an EEHV1B strain.

Furthermore, all but two were readily distinguishable as independent unrelated strains

(Table 2-1). The exceptions were from two orphan calves at Kodanad Camp that died

18 within three days of one another in Dec 2007. These proved to have identical DNA sequences to one another at all six PCR loci tested (totaling over 4000-bp), including over 1500-bp tested of tissue from each of five different organs sampled for calf-1 and in both the heart and liver for calf-2. Several other instances of two afflicted calves at the same location being infected almost simultaneously with identical EEHV1 strains have also been observed in Europe and North America. Remarkably, amongst the two most hypervariable PCR loci tested, the Indian cases included four of the five known major vGPCR1 subtypes and five of the six known major glycoprotein-H subtypes, thus encompassing almost all of the overall genetic range of EEHV1 variants observed to occur amongst the 31 Asian elephant cases analyzed in a similar way from Europe and

North America. Phylogenetic comparisons of the viruses from all nine Indian cases with selected prototype EEHV1A, EEHV1B, EEHV2 and EEHV6 cases at the U71-gM, vGPCR1, HEL and gH-TK loci are presented in the dendrograms in Figure 2-3, panels

A-D and graphical alignment presentations of the nucleotide polymorphism patterns for the same four variable loci are presented in Figure 4, panels A-D. Although a small number of novel polymorphisms were detected, all nine Indian examples proved to be closely related to already known subtypes at each locus examined. Comparisons of the overall distribution of EEHV1 subtypes for the vGPCR1 and gH-TK loci between Indian,

North American and European cases are illustrated in Figure 2-5, panels A and B. The actual numbers of nucleotide polymorphisms and deletion/insertion differences found amongst the individual Indian EEHV1 strains at all six PCR loci are presented in Figure

2-6, panels A-F.

19 Discussion

We conclude that sporadic EEHV1-associated hemorrhagic disease is distributed across the entire Southern India elephant population and that it displays gross and microscopic pathological features that are indistinguishable from those found in the highly lethal disease documented previously in captive Asian elephants in Europe and

North America. Most importantly, the very wide range of genetic diversity observed amongst these nine Indian EEHV1 cases is not compatible with an earlier model suggesting that this disease resulted from cross-species transmission of EEHV1 from

African to Asian elephants. The latter interpretation was largely based on the apparent finding of EEHV1 viral DNA in archival paraffin block specimens of skin nodules from several African elephant calves that were imported from Zimbabwe to Florida in the mid-

1980s. However, we have not been able to reproduce that result upon re-examination of the same archival skin nodule specimens, and no other examples of EEHV1 DNA have ever been detected in necropsy samples from African elephants. Instead, multiple examples of EEHV2, EEHV3, EEHV6 and EEHV7 have now all been found in both lung and skin nodules from eight different asymptomatic adult African elephants examined from South Africa, Kenya and the USA 7 (Pearson V. et al, pers comm.). Similarly, only

EEHV2 or EEHV6 have been detected in blood or necropsy samples from just three known rare cases of hemorrhagic or viremic disease in African elephants in North

America. Therefore, we interpret that the earlier result was most likely a contamination error and that both EEHV1A and EEHV1B are natural endogenous viruses that co- evolved with Asian elephant hosts rather than in African elephants. Relatively recent introduction (even several thousand years ago) from just one or a few exogenous sources

20 would have resulted in a very limited number of strains and a very restricted level of genetic variation. Current evidence suggests that EEHV4 and EEHV5 4 may also be endogenous to Asian elephants, but whether they are distributed much less abundantly or are just less pathogenic than EEHV1A and EEHV1B is not yet known. Just a single example of the death of a zoo Asian elephant calf with hemorrhagic disease caused by

EEHV3 has been documented 3, but otherwise EEHV3 has been found on multiple occasions only in benign lung and skin nodules obtained from African elephants, and cross-species infections would thus appear to be extremely rare.

Herpesvirus genomes evolve relatively slowly and are generally highly stable, yet individual human cytomegaloviruses (HCMV) for example display considerable hypervariability in certain genes that evidently represent relics of very ancient genomic divergence that has subsequently become scrambled through extensive recombination events. The vGPCR1 and glycoprotein-H subtype hypervariability in EEHV1 appears to indicate a similar ancient species population drift effect. We had anticipated that the multiple different subtypes of EEHV1 originated in and might still be distributed non- uniformly in different Asian elephant sub-populations, especially considering that the

Nilgiri Eastern Ghat Elephas maximus population itself displays relatively narrow nuclear and mitochondrial DNA haplotype diversity compared to that of other Asian elephant sub-populations 33. However, this was not the case; instead most known subtypes proved to be present, even within just this small number of cases from Southern India, representing almost the entire known genetic diversity range of EEHV1. This result implies that EEHV1 is likely to be an ancient infection carried by, transmitted between, and widespread amongst all Asian elephant populations.

21 Evidently, both the A and B chimeric variants of EEHV1, but seemingly not the other EEHV2 to EEHV7 species of Probosciviruses, commonly produce severe hemorrhagic pathology upon primary infection within all present day Asian elephant populations, whether born and reared in the wild or under human care. Since the late

1980s, acute primary EEHV1viremic disease has afflicted more than 20% of all Asian elephant calves born in North America, with an observed mortality rate of over 80%.

Why a virus that would normally be expected to be well-adapted to and benign in its natural hosts displays such devastating fulminant pathology remains to be explained, and the possibility that some other agent or co-factor contributes to the disease cannot be ruled out. The rate of disease in the wild is unknown, but if it is anywhere near as high as observed in Europe and North America, then the effects on future reproductive success and survival of this highly endangered species could be devastating. Additional studies on the prevalence of infections by the EEHV1, EEHV4 and EEHV5 viruses and of

EEHV-associated hemorrhagic disease, especially in free-ranging wild elephants, and in all Asian range countries would appear to be urgently needed.

Acknowledgements

Financial support for these studies, the setting-up and operation of the diagnostic laboratory in Wayanad, Kerala and for travel of team members from NEHL and Johns

Hopkins University to India came from the International Elephant Foundation, the

Smithsonian Institution and the Kerala Forest Department. Studies at Johns Hopkins

University were supported in part by NIH research grant R01 AI24576 to GSH. Field research by AZ was supported in part by the U.S. Fish and Wildlife Services Asian

Elephant Conservation Fund. We thank Dr Anilkumar for the inclusion body

22 photomicrographs and Dr R Thirumurugan of the Arigna Anna Zoological park in

Chennai for providing tissue samples.

23 Tables

24 Figures A B C

D E F

25 µm 25 µm

Figure 1. (A-D) Pathologic changes of EEHV-associated disease found during field necropsy. (A) Asian elephant with cyanosis of the tongue a ributed to EEHV disease; (B) Epicardial surface of the heart (apex view) from an Asian elephant Figureshowing s2e-v1.ere Pathologicalextensive hemorrh achangesge a ribute ofd to EEHV EEHV dise-aassociatedse; (C) Ventricu ldiseasear endocar dfoundial surfa cduringe of the h efieldart fro m an Asian elephant showing mul focal areas of ecchymo c hemorrhage a ributed to EEHV disease; (D) Serosal membrane surface of the liver showing diffusely sca ered petechial hemorrhage a ributed to EEHV disease; (E,F) Photomicrograph necropsy.of two capill a(Ary e)n dAsianothelial celephantells containi nwithg typic acyanosisl basophilic i noftra ntheucle atonguer viral inc luattributedsion bodies fr otom nEEHVecropsy l ivdisease;er ssue. (B) Hematoxylin and eosin stain, bar = 25 µm. 15 Epicardial surface of the heart (apex view) from an Asian elephant showing severe extensive hemorrhage attributed to EEHV disease; (C) Ventricular endocardial surface of the heart from an Asian elephant showing multifocal areas of ecchymotic hemorrhage attributed to EEHV disease; (D) Serosal membrane surface of the liver showing diffusely scattered petechial hemorrhage attributed to EEHV disease; (E, F) Photomicrograph of two capillary endothelial cells containing typical basophilic intranuclear viral inclusion bodies from necropsy liver tissue. Hematoxylin and eosin stain, bar = 25 µm.

Figure 2-2. Representative agarose gel-ethidium bromide PCR band results obtained after first round (520-bp) and second round semi-nested (250-bp) amplification with

EEHV PAN-POL primers from 12 samples of necropsy tissue DNA. Lanes 1, 2 and 12, different tissues from case IP07; 3, case IP10 (negative); 4, 5, 6, 7 & 8 multiple tissue samples from case IP06; 9, case IP01; 10, case IP04 (negative); 11, case IP11.

Figure 2-3. Phylogenetic distance based dendrograms showing patterns of DNA level divergence and subtyping amongst the Indian EEHV1 cases (boxed in orange) in comparison with prototype examples of EEHV1A, EEHV1B, EEHV6 and EEHV2 at the four most variable PCR sequence loci (A) U71/gM; (B) U77/HEL; (C) U51/vGPCR1 and

(D) U48/gH-TK. Bootstrap values are indicated at branch points.

) ) (A) (B) (C (D

28 Figure 2-4. Graphical presentation generated from Geneious of the patterns of DNA sequence polymorphisms (including gaps) across the four most variable PCR loci (A)

U71-gM; (B) U77/HEL; (C) U51/vGPCR and (D) U48/gH-TK for all Indian cases (IP#) compared to aligned prototype examples of EEHV1A, EEHV1B, EEHV6 and EEHV2.

Figure 2-5. Comparison of the major EEHV1 gene subtype distribution patterns amongst

Indian, North American and European hemorrhagic disease cases at the two most variable PCR loci examined (A) U51/vGPCR1 and (B) U48/gH-TK.

30 (A) U38/POL Locus (480-bp)

IP01 IP05 IP06 IP07 IP11 IP91 IP93 Subtype IP01 - A IP05 1 - A IP06 1 0 - A IP07 1 0 0 - A IP11 0 1 1 1 - A IP91 0 1 1 1 0 - A IP93 15 14 14 14 15 15 - B Polymorphisms

(B) U60/TERex3 Locus (360-bp)

IP01 IP05 IP06 IP07 IP11 IP91 IP93 Subtype IP01 - A1 IP05 6 - B2/A IP06 4 8 - A2 IP07 4 8 0 - A2 IP11 0 4 4 4 - A1 IP91 0 4 4 4 0 - A1 IP93 11 15 15 15 11 11 - B1 Polymorphisms

31 IP0 IP0 IP0 IP0 IP1 IP4 IP6 IP9 IP9 Subtyp 1 5 6 7 1 3 0 1 3 e IP0 - 6 3 3 0 0 3 0 12 A 1 IP0 1 - 9 9 6 6 9 6 6 A 5 IP0 1 2 - 0 3 3 0 3 9 A 6 IP0 1 2 0 - 3 3 0 3 9 A 7 IP1 0 1 1 1 - 0 3 0 12 A Ins/De 1 l IP4 0 1 1 1 0 - 3 0 12 A 3 IP6 2 3 3 3 2 2 - 3 9 A 0 IP9 0 1 1 1 0 0 2 - 12 A 1 IP9 36 37 37 37 36 36 38 36 - B 3 Polymorphisms

32 (D) U77/HEL Locus (952-bp)

IP01 IP05 IP06 IP07 IP11 IP43 IP60 IP91 Subtype IP01 - A IP05 3 - A IP06 0 3 - A IP07 0 3 0 - A IP11 3 2 3 3 - A IP43 0 3 0 0 3 - A IP60 2 1 2 2 1 2 - A IP91 3 2 3 3 2 3 1 - A Polymorphisms

(E) U51/vGPCR1 Locus (885-bp)

IP0 IP0 IP0 IP0 IP1 IP4 IP6 IP9 IP9 Subtyp 1 5 6 7 1 3 0 1 3 e IP0 - 3 3 3 3 0 0 0 12 A2 1 IP0 31 - 0 0 0 3 3 3 9 E 5 IP0 31 0 - 0 0 3 3 3 9 E 6 IP0 31 0 0 - 0 3 3 3 9 E Ins/De 7 l IP1 23 6 6 6 - 3 3 3 9 E/A 1 IP4 45 27 27 27 29 - 0 0 12 D1 3 IP6 26 28 28 28 26 25 - 0 12 D2/A 0 IP9 2 31 31 31 25 46 28 - 12 A2 1 IP9 29 38 38 38 39 45 33 29 - B1 3 Polymorphisms

(F) U48/gH-TK Locus (807-bp)

33 IP01 IP05 IP06 IP07 IP11 IP43 IP60 IP91 Subtype IP01 - A1 IP05 26 - F IP06 53 56 - C3 IP07 58 56 0 - C3 IP11 57 64 74 74 - E IP43 51 52 14 14 70 - C1 IP60 65 66 74 74 48 71 - D** IP91 63 66 72 72 48 67 4 - D* Polymorphisms

Figure 2-6. Direct side-by-side nucleotide level differences observed amongst the nine

Indian EEHV1 cases at all six PCR loci (A) U38/POL; (B) U60/TERex3; (C) U71/gM;

(D) U77/HEL; (E) U51/vGPCR1 and (F) U48/gH-TK. Nucleotide insertion/deletion differences found at the U71/gM and U51/vGPCR1 loci are also indicated.

III. Complete genome sequence of elephant endotheliotropic herpesvirus 4 (EEHV4) the first example of a GC-rich branch Proboscivirus

This chapter has been previous published as:

Ling PD, Long SY, Fuery A, Peng RS, Heaggans SY, Qin X, Worley KC, Dugan S, Hayward GS. Complete genome sequence of elephant endotheliotropic herpesvirus 4 (EEHV4) the first example of a GC-rich branch proboscivirus. mSphere. 2016. June 15;1(3) . pii: e00081-15. doi: 10.1128/mSphere.00081-15

Reprint with permission of the publisher, American Society for Microbiology. Abstract

35 A novel group of mammalian DNA viruses called Elephant Endotheliotropic

Herpesviruses (EEHVs) belonging to the Proboscivirus genus has been associated with nearly 100 cases of highly lethal acute hemorrhagic disease in young Asian elephants worldwide. The complete 180-kb genomes of prototype strains from three AT-rich branch viruses EEHV1A, EEHV1B and EEHV5 have been published. However, less than

6-kb of DNA sequence each from EEHV3, EEHV4 and EEHV7 showed them to be a hugely diverged second major branch with GC-rich characteristics. Here we determined the complete 206-kb genome of EEHV4 (Baylor) directly from trunk wash DNA by next generation sequencing and de novo assembly procedures. Amongst a total of 119 encoded genes with a similar overall co-linear organization to those of the AT-rich EEHVs, major features of EEHV4 include a family of 26 paralogous 7xTM and vGPCR-like genes plus

25 novel or missing genes. It also contains an unusual distribution of tracts of 5 to 11 successive A or T nucleotides in intergenic domains between the mostly much higher

GC-content protein coding regions. Furthermore, an extremely high GC-rich bias in the third wobble position of codons clearly delineates the coding regions for many but not all proteins. There are also two novel captured cellular genes, including a C-type lectin

(vECTL) and an O-linked acetyl glucosamine transferase (vOGT) as well as an unusual large and complex Ori-Lyt dyad symmetry domain. Finally, 30-kb from a second strain proved to include three small chimeric domains indicating the existence of distinct

EEHV4A and EEHV4B subtypes.

Introduction

36 The deaths of 62 young Asian elephants with acute hemorrhagic disease in North

America and Europe have been attributed to rapidly developing uncontrolled primary systemic infections by members of a novel genus of mammalian herpesviruses, designated Elephant Endotheliotropic Herpesviruses or EEHVs 1,3,8,19,34. At least 43 additional lethal cases have also been confirmed by DNA tests within Asian range countries, including 16 published examples from India, Thailand, Cambodia and Laos 10-

12. Seven major genotypes named EEHV1 to EEHV7 that all qualify as distinct species within the Proboscivirus genus have been identified within North American elephants

3,4,6,20,34,35, although EEHV1A has been associated with the majority of lethal cases.

Overall, some 46 highly divergent strains of EEHV1A have been identified by selective

PCR sequencing based gene subtyping at multiple loci. Only ten examples of EEHV1B plus a smaller number of EEHV4 and EEHV5 viruses have also all been associated with systemic disease in Asian elephants (Elephas maximus), whereas three others EEHV2,

EEHV3 and EEHV6 were detected in the few rare disease cases in zoo African elephant

(Loxodonta africana) calves 4,6,20. The latter three viruses as well as EEHV7 have also all been detected as quiescent infections in lung nodules from healthy adult African elephants 35. Just 12 young captive Asian elephants with confirmed EEHV1 DNA- positive systemic disease that had been treated with human anti-herpesvirus drugs are considered survivors 9,20,23.

Until recently, it seemed that EEHV1A and its less common partially chimeric variant EEHV1B 5,21,25,26 were the predominant viruses that Asian elephants might be encountering. However, routine testing between 2011 and 2015 within the most closely monitored USA zoo herd (which consists of just eight individual Asian adults and calves)

37 have now also detected episodes in which multiple herdmates underwent sequential mild primary viremic infections with subsequent trunk wash shedding by first EEHV5 2 and then later EEHV4 15. Because at least four different EEHV1A strains had been documented in lethal cases at this same facility within the previous 15 years, and most of this herd had already been observed to undergo sub-clinical infections by strains of

EEHV1A or EEHV1B or both either several years earlier or later 15,29,30, we conclude that it is likely that most Asian elephants eventually become infected with multiple EEHV species and subtypes. The relative timing and order of these primary EEHV infections are expected to have major impacts on the levels of immune protection to disease caused by the others.

Although no Probosciviruses have yet been grown in cell culture, the complete genomes of four reference strains of AT-rich branch Asian elephant EEHVs have been determined previously directly from necropsy tissue, including two of EEHV1A and one each of EEHV1B and EEHV5A 14,16 ,17. Therefore, we wished to take the opportunity that this latest EEHV4 episode provided to learn more about the genetic relationships amongst the different EEHV lineages and species by determining the complete genomic DNA sequence of EEHV4 strain Baylor as the prototype example of a GC-rich branch

Proboscivirus. In the accompanying paper 36 we compare and contrast the genomes of these two major branches of the Proboscivirus as well as describe a number of additional characteristic novel features of the entire group.

Materials and Methods

Source

38 The sequenced DNA came from a trunk wash sample collected from a four-year old Asian elephant that had experienced a mildly symptomatic episode of EEHV4 PCR- positive viremia starting Sept 2nd 2014, then presented with a transient high level trunk wash shedding beginning in Oct 2014. Clinical details of the diagnosis, case history and treatment of the case have been published under animal no. 2 in Feury et al 15.

Library Preparation and DNA Sequencing

Illumina paired-end libraries were prepared according to the manufacturer’s protocol (Illumina Multiplexing_SamplePrep_Guide_1005361_D) with modifications as described in the BCM-HGSC protocol (https://www.hgsc.bcm.edu/content/protocols- sequencing-library-construction). Briefly, 190 ng of DNA was sheared into fragments of approximately 200-300 base pairs with the Covaris E210 system followed by end-repair,

A-tailing, and ligation of the Illumina multiplexing PE adaptors. Ligation-mediated PCR

(LM-PCR) was performed for 6 to 8 cycles using the Library Amplification Readymix, which contains KAPA HiFi DNA Polymerase, and universal primer IMUX-P1.0 and

IMUX-P3.0. Purification was performed with Agencourt AMPure XP beads after enzymatic reactions. Size distribution of the LM-PCR products was determined using the

LabChip GX electrophoresis system (Perkin Elmer), and quantification was performed by gel analysis using AlphaView SA Version 3.4 software.

Library templates were prepared for sequencing using Illumina TruSeq reagents and protocols. Briefly, the libraries (or library pools) were denatured with sodium hydroxide, diluted to 6-9 pM in hybridization buffer, and then loaded on a lane of a

HiSeq flow cell. The libraries then underwent bridge amplification to form clonal

39 clusters, followed by hybridization with the sequencing primer. Sequencing runs were performed on the Illumina HiSeq 2000 platform using the 2x100 run format.

Illumina Read Processing and Assembly

A total of 416 million reads (42 Gb) were produced using the Illumina HiSeq sequencing technology. These raw reads were processed using SeqPrep program

(https://github.com/jstjohn/SeqPrep) to remove adapter sequences. Low quality sequences (with quality less than or equal to 2) at the ends of the reads were trimmed using an in-house script. The trimmed reads were mapped to the African elephant reference genome (Loxafr3.0) using BWA 37 to detect sequences with greater than 95% match, and such mapped reads were assumed to be of host origin and omitted from later assembly steps. The 212 million reads remaining after these processes were assembled using Velvet 38 with various kmer sizes of either 29, 45, 63 or 75 and these parameters: - exp_cov auto -cov_cutoff 20 -min_contig_lgth 400. The processed reads were also sub- sampled into smaller data sets of 20, 60 and 100 million reads. In the end, Velvet assemblies using all processed reads or using 100 million reads with the highest kmer sizes each yielded a total of between two and five contigs that aligned to EEHV1A

(Kimba) genomic DNA sequences in BLAST-N searches and added up to between 205 and 206-kb.

DNA Sequence Analysis and Comparisons

After filling and joining across the several remaining small gaps between adjacent contigs by standard Sanger PCR amplification and cycle sequencing approaches, a single final consensus contig file of 205,896-bp was constructed. The assigned gene

40 nomenclature and annotations were generated initially by BLASTP and BLASTX searches of the Genbank database for every potential ORF of greater than 80-aa that did not substantially overlap with another already identified ORF. There was just a single exception to the overlap rule (E44A) within the final assignments. Subsequently, each

ORF was confirmed directly for both orthologues and paralogues by amino acid identity comparisons with the corresponding ORFs in the EEHV1A (Kimba), EEHV1B (Emelia) and EEHV5A(Vijay) files, as well as with BLASTX searches of the intact

EEHV4(Baylor) genome file itself. Clustal alignments and phylogenetic trees were generated in MEGA5 based on MUSCLE alignments or in MacVector 12 as described previously 5. Dot matrix diagrams showing global nucleotide alignments were generated as implemented at http://blast.ncbi.nlm.nih.gov/Blast.cgi. Simplot software 5,39 was used to display nucleotide identity and divergence comparisons.

Database Accession Numbers

The final completed annotated 205,896-bp genomic DNA sequence file of

EEHV4 (Baylor) is deposited at NCBI GenBank under accession number KT832477.

Previous PCR data for this strain totaling 1,828-bp of unique sequence has accession numbers KR781023-KR781037 15. For comparative purposes an amended version of the

178-kb EEHV1A (Kimba) genome with updated annotations has accession number

KC618427 17. Thirteen new DNA sequence files totaling close to 24.3-kb of extended data from Sanger cycle sequencing of amplified PCR loci from the prototype

EEHV4(NAP22) genome have accession numbers KT832478-KT832490 and

KU147235.

Result

Assembly of the Complete 206-kb EEHV4 (Baylor) Genome Sequence

The intact genomes of four prototype Asian strains have all been assembled by random high throughput and de novo assembly approaches directly from high quality necropsy tissue DNA obtained from young elephants that died of acute hemorrhagic disease 14,16,17. For EEHV4 (Baylor), we instead used a trunk wash pellet DNA sample with a high measured ratio of viral to host cell DNA. From a total of 420 million 100-bp long runs of raw data, close to half resembled African elephant DNA and were filtered out before the remainder were assembled de novo by the Velvet program using a variety of different K-mer size parameters. Four independent assembled contig libraries that were searched for matches to EEHV1 (Kimba) genomic DNA produced results of between one and four contigs each with the largest being 202,155-bp. There were three small gaps in the data overall each of which was repaired by Sanger PCR sequencing and proved to be located in internal repetitive regions either within E34 (ORF-C) or in the predicted Ori-

Lyt locus between U41(MDBP) and U42(MTA) or close to the right-hand-side terminus.

Because of this being intracellular and not virion derived DNA, the final results in each case were identical contiguous circular genomes of 205,896-bp (average coverage of 110- fold).

To generate a linear version, we arbitrarily defined a G10-tract at the beginning of the original largest contig as the left-hand-side end of the EEHV4 (Baylor) genome.

Importantly, two regions mapping very close to the right-hand-side end contained distinct sets of potential packaging signal motifs of which the first matched at 45 out of 54

42 nucleotides (83%) to both copies of the terminal repeats of EEHV1A (Emelia) and

EEHV1A (Raman), and the second matched at 103 of 144 positions (72%) to all three copies of the terminal repeat “a” sequence of HSV1 (KOS). The extreme left-hand-side of the genome (outside gene E1) contains several segments including a set of 17-bp tandem-repeats as well as a cluster of 8-bp palindromic cyclic AMP-response elements

(CREs) that have high level homology to repetitive sequences present within the terminal repeats of the other EEHV genomes. Therefore, there is clear evidence that both ends of our assembled EEHV4 genome map within the predicted terminal repeat and lie close to the legitimate physical ends of the EEHV1A, EEHV1B and EEHV5 genomes as determined by Wilkie et al 14,16. Furthermore, on the far right-hand-side of the EEHV4

(Baylor) genome there is a 5.3-kb non-coding segment lying outside and immediately upstream of the gene encoding the U44 (ORF-L) transcription factor-like protein. This region is similar in size to the nearly 7-kb of non-coding DNA (which includes all of one copy of the 2.9-kb terminal repeat) at that position in the linear AT-rich branch genomes of Wilkie et al 14,16. Based on the fact that we could assemble a single intact contig with terminal repeat sequences near both expected ends, we deduced that this must represent the entire EEHV4 (Baylor) genome.

Global Features and Initial Comparisons with Other EEHV Genomes

The genome of EEHV4 (Baylor) contains 119 open reading frames (ORFs) arranged as shown in the map in Figure 3-1. A full listing of the names, sizes and map coordinate positions for all designated protein coding ORFs arranged from left to right across the EEHV4 (Baylor) genome oriented in the same direction as in our prototype

43 EEHV1 (Kimba) is presented in Table 3-1. This data includes comparative information relative to the intact genomes of EEHV1A (Kimba and Raman), EEHV1B (Emelia) and

EEHV5A (Vijay) to indicate all of the genes present in EEHV4 that are not found in the others, as well as all genes present within the others that are missing from EEHV4. Table

3-1 also shows the GC-content of each ORF in EEHV4 and whether the gene is assigned to a gene family and its status as a common core herpesvirus gene or is shared between subsets of the Alpha (α), Beta (β), Gamma (γ) sub-families or is unique to the

Proboscivirus genus. Because we are proposing here that the Probosciviruses have numerous novel biological properties and genetic and evolutionary features that may justify their future designation as a new Delta (δ) subfamily of mammalian herpesviruses, to reduce possible confusion later on we have adopted an interim dual nomenclature of either p or δ when referring to unique features of the Proboscivirus gene and protein sets in a phylogenetic context.

The gene-ORF-protein nomenclature used here for EEHV4 (Baylor) is also based on that used originally for EEHV1B (Kiba) by Ehlers et al 26 and expanded upon for

EEHV1A (Kimba) by Ling et al 17 to include an E-series numbering system for all novel

Proboscivirus-specific proteins. All proteins with identifiable orthologues in EEHV1A

(Kimba) retain those same numbers, whereas all newly assigned proteins that are unique to EEHV4 have been given distinctive E number descriptors.

The most obvious feature about the EEHV4 genome is that the overall orientation and gene content (especially within the central core gene segment) are essentially conserved and collinear with those of the three AT-rich branch genomes, although with considerable divergence towards both ends as revealed in a full length genomic dot

44 matrix comparison (Figure 3-2a). In particular, the large 40-kb inversion of the conserved core blocks I, II and III between U27 and U44 in Betaherpesviruses, as well as a second smaller inversion of a weakly conserved 24-kb gene block, are both retained in common with the organization found in EEHV1A, EEHV1B and EEHV5. Part of the core gene region involved in the first of these two genomic inversions is clearly revealed on the left-hand side within a dot matrix comparison with the prototype human

Betaherpesvirus genome HCMV (Merlin) (Figure 3-2b), although the second inverted region (further towards the left-hand side) is not sufficiently conserved to be detectable in the diagram. The other half of the conserved core segment (blocks IV, V, VI and VII) produces the largest visible signal located further towards the right-hand-side, but is not inverted. No other mammalian herpesviruses have this same type of overall gene organization as found within both major branches of the EEHVs.

In common with the pattern found for the other Proboscivirus genomes, the central core segment of EEHV4 (Figure 3-1) retains the three signature optional shared

Alpha-Gamma (αδ) and Alpha-Beta2 (αβ2) class genes U27.5 (RRB), U48.5 (TK) and

U73 (OBP), plus two other Betaherpesvirus-like genes U47 (gO) and U51 (vGPCR1) mapping together with the 35 obligatory true core genes found in all mammalian herpesviruses, and the six other shared Beta-Gamma (βγ) class genes that are absent from all Alphaherpesviruses. Although 15 other genes in the HCMV US22, vMIP, vICA and mIE gene block from UL23 through to UL43 are clearly absent and the whole 24-kb genome segment from E32 (U14.5) to E35 in EEHV4 (Baylor) at coordinates 60.5 to

81.6 lies in an inverted orientation compared to that of the adjacent blocks in all

Betaherpesvirus genomes, some of the genes designated U14.5, U14, U13.5, U12

45 (vGPCR2), U4 and U4.5 (ORF-B) in EEHV1A (Kimba) have low level resemblance to a subset of their Betaherpesvirus orthologues in this region. Because of the evidently shared common evolutionary origin within a predicted ancestor of both sub-families, these six genes in both EEHV1A and EEHV4 have been assigned Beta-Proboscivirus

(βp) or Beta-delta (βδ) class status 17 (Table 3-1).

A dramatic feature of the EEHV4 (Baylor) genome is the presence of 26 paralogous members of an ancient and highly diverged family of 7xTM anchored transmembrane proteins (designated the p3 or δ3 family) within the left-hand side novel segment of the genome. The linear distance-based protein level phylogenetic tree given in

Figure 3-3 shows relationships and sub-grouping amongst this protein family, of which the closest viral or cellular homologue is Retinoic acid induced protein 3 (RAIP3) an orphan group G-protein coupled receptor or GPCR. Although they are very highly diverged, a combination of the phylogenetic tree branching patterns together with

PBLAST domain-match identity values (not shown) allowed us to loosely classify them into five subfamilies, based on closer relationships to the prototype examples of E3, E6,

E14, E15 and E18, of which there are seven (E1, E3, E3.1, E3.1, E3.3, E3.4 and E26), six

(E6, E7, E9, E11, E12, E13), four (E14, E14.1, E14.2. and E16), three (E15, E20 and

E21) and two (E18 and E28) paralogous copies respectively in EEHV4. Only the adjacent pair of E14.1 and E14.2 display a much closer identity than any others, suggesting that they are the most recently duplicated and most likely both arose from the next adjacent gene E14. More detailed descriptions and comparisons of the multiple vGPCR-like proteins as well as several other smaller gene families and some captured cellular genes

46 from both major branches of the Probosciviruses and related mammalian herpesvirus proteins are presented in the accompanying manuscript by Ling et al 36.

Another striking feature of the EEHV4 (Baylor) genome assembly is the evident absence of the entire 10.5-kb block encoding the ten to 12 highly variable genes mapping between E47 to E55 of EEHV1A (Kimba), or of the rearranged versions of this block found in EEHV1B (Emelia) and EEHV5 (Vijay). Those genes include between three and five members of the CD48-related vIgFam gene family plus several vGPCRs, as well as the E47 (vFUT9) and E54(vOX2-1) genes. All six copies of the signature 10- to 35-bp long alternating CA-repeats found in EEHV1A (Kimba) are also missing in EEHV4

(Baylor), although that was already known to be the case for EEHV5 (Vijay) as well.

Although the intact genomes of EEHV4 (Baylor) and EEHV1A (Kimba) are too far diverged for accurate global determinations of the overall DNA sequence identity level between them, a numerical value of 49.6% was measured by the Emboss DNA

Stretcher program for the most conserved 97-kb core genomic segment from U44

(82,500) to U77 (179,500). Distortions to the linearity of the alignments preclude obtaining meaningful results for the other non-core outer parts of the EEHV4 genome.

This result compares to values of 63% identity for the intact EEHV1B (Emelia) genome versus EEHV5 (Vijay) and 92.3% for EEHV1B (Emelia) versus EEHV1A (Raman), which increase to 64.2% and 94.2% respectively when the equivalents of the 10.5-kb non-linear vIgFam plus vGPCR block between E47 and E55 in Kimba are omitted.

Unusual Patterns and Distribution of GC-Rich Sequence Blocks and A or T-Tracts

47 The several small previously analyzed PCR segments of EEHV3 and EEHV4

DNA (totaling 4 to 5-kb each) were recognized to have a distinctive much higher overall

GC-content of 63% compared to 43% for EEHV1A, 1B and 5 6,15. However, the overall base composition of the intact EEHV4 genomic DNA proved to be just 57% GC-content.

Much of this difference results from the curious finding that, unlike the adjacent protein coding regions, the majority of the interspersed non-protein coding regions in EEHV4 are highly conspicuous by the inclusion of numerous successive A or T-tracts of between five and 11 nucleotides in length. These features also apply within the relatively small number of clearly identifiable introns.

Despite the lower than expected overall GC-content, the level does increase to an average of 62% (range 57 to 66%) for the 17 largest protein coding regions, which are all over 2,500-bp in length and occupy 67.6-kb. In fact, all but 11 of the 119 recognized

ORFs display a GC-content of greater than 48% (Table 3-1), whether they map within the novel left-hand-side part of the genome or within the core conserved region. The only exceptions to this pattern occur across six small dispersed locations (bolded in Table 3-1) encompassing just 10.8-kb in total length and encoding genes that are either very highly diverged from their counterparts in EEHV1 and EEHV5 or entirely novel. Five of these genes, including the additional second captured acetyl glucosamine transferase E9A

(vOGT), lie adjacent to one another between map coordinates 20.8 to 25.2-kb. The second largest such locus covers the E37 (ORF-O) and adjacent novel E39A (ORF-R) glycoprotein between map coordinates 186 to 189.7. The third locus encodes two genes, including a captured E16D (vECTL) protein mapping between map coordinates 35.6 to

36.8. Finally, E4A at coordinates 15.2 to 15.7 and E30A mapping between coordinates

48 55.5 to 56.2 also fall into this category. Perhaps reflecting a relatively more recent acquisition of at least one of the genes at each of these five AT-rich ORF locations, they appear to be much more similar to host cell DNA than to the rest of the viral genome, with ORF base compositions ranging from only 27 to 45% GC-content.

Novel EEHV4 Type-Specific Genes

The currently updated annotation of the EEHV1A (Kimba) genome contains 118 identified genes. Amongst these, a total of 26 are evidently missing within the EEHV4

(Baylor) genome), whereas a total of 25 additional apparently novel genes are present

(Table 3-1). The missing genes in EEHV4 (Baylor) relative to EEHV1 (Kimba) include, in addition to the entire 10.5-kb ten gene block from E47 to E55, those designated E2,

E5A, E7A, E8, E10, E16A/B, E17, E18A, E18B, E24, E25, E31, E36A (vCXCL1),

E38(ORF-P) and E39(ORF-Q). The extra genes in EEHV4 (Baylor) compared to

EEHV1A (Kimba) include E3.1, E3.2, E2A, E3.3, E3.4, E4A, E4B, E4C, E6B, E7B,

E9A (vOGT), E9B, E9C, E10A, E12A, E14.1, E14.2, E16D (vECTL), E17A, E18C,

E20B, E24B, E27ex2, E30Aex1/2 and E39A (ORF-R). There are also two genes EE23A

(vOX2-V) and EE44A that are unique to EEHV5 (Vijay) and that are absent from both

EEHV1 and EEHV4, and E39 (ORF-Q) is unique to EEHV1A and EEHV1B, but absent from EEHV2, EEHV4 and EEHV5, whereas E33A and E36A (vCXCL1) are unique to

EEHV1A although absent from EEHV1B, EEHV2, EEHV4 and EEHV5.

Six of the new genes in EEHV4 represent recognizable highly-diverged duplicated paralogues of genes already present in both EEHV1 and EEHV4 that could presumably have been deleted from some ancestral genome as the AT-rich branch

49 evolved separately from the GC-rich branch. All of the latter are members of the p3 (or

δ3) 7xTM family described above (E3.1, E3.2, E3.3, E3.4 and E14.1 and E14.2). E2A which maps amongst the new duplicated versions of the vGPCR6 cluster, is also clearly a member of this 7xTM family, but this protein has no recognizable orthologues or paralogues in either EEHV1 or EEHV5. A similar situation applies for E7A and E10A. In contrast, the following remaining 16 members of the 7xTM family E1, E3, E7, E9, E11,

E12, E13, E14, E15, E16, E18, E20, E21, E26, E28 and E29 all have direct orthologues in EEHV4 as recognized by both positional and protein homology criteria.

EEHV4 E37 (ORF-O) and E39A (ORF-R) are both predicted to be spliced potential envelope glycoproteins mapping between gL and ORF-K at coordinates 186 to

189.6 at the same location and orientation as the similarly spliced E37 (ORF-O) plus E38

(ORF-P) and E39 (ORF-Q) ser plus thr-rich glycoproteins of EEHV1A (Kimba).

However, although quite plausibly these proteins all had common evolutionary origins,

ORF-R now lacks sufficient residual homology to be designated as an unambiguous orthologue of either ORF-P or ORF-Q. Interestingly, all three proteins have two exons and residual identity between ORF-P and ORF-Q (especially within exon1) suggests that

ORF-Q is a highly diverged tandem-duplicated version of ORF-P that is present only in

EEHV1A and EEHV6, but is missing from EEHV1B, EEHV2 and EEHV5. However, both ORF-P and ORF-Q are absent from EEHV4 and have seemingly been replaced by

(or evolved into) the much shorter ORF-R glycoprotein that now displays no detectable identity to either of them. In contrast, the adjacent predicted to be three exon ORF-O glycoprotein of EEHV4 still retains significant identity (37% across all of exon-3,

50 although barely detectable in exons 1 and 2) to its orthologues in all of the AT-rich

Probosciviruses.

Varying Patterns of Base Composition Bias within Wobble Codon Positions

We previously pointed out that the coding regions within each of the four to five small segments sequenced by Sanger PCR approaches from both prototype genomes

EEHV3 (NAP27) and EEHV4 (NAP22) within the GC-rich branch of the

Probosciviruses displayed extraordinarily high GC-content bias at the third nucleotide or wobble codon position 6. For five of the seven short ORF fragments analyzed there in each virus the wobble position GC-content was between 86 and 99%. This compares with codon wobble position GC-contents ranging from just 41 to 50% for the orthologous

ORF segments in EEHV1 and EEHV5.

Evaluation of the complete 206-kb EEHV4 (Baylor) genome shows that a very high level of GC-content bias at the third nucleotide or wobble codon position applies across close to 90% of the entire genome involving as many as 105 of the 119 total annotated genes. Remarkably, a global plot comparison of the GC-content of each of the three EEHV4 reading frames as an indirect measure of the wobble position effect produces such a dramatic pattern that the position of virtually every ORF (except those in the six minority AT-rich regions) are easily recognized and defined. However, there is very little similar frame-specific demarkation when this type of plot is applied to

EEHV1A (Kimba). Four selected examples of these codon-specific plot panels scanning

18-kb segments each of the EEHV4 (Baylor) genome are shown in Figure 3-4a-d, two from novel areas of the genome on the left-hand side, one showing a conserved core gene

51 segment surrounding Ori-Lyt near the center of the genome and one encompassing the presumed transcriptional regulators ORF-K and ORF-L at the extreme right-hand-side.

Typical examples of large genes with a high GC-wobble bias in EEHV4 (Baylor) include

U57 encoding the major capsid protein (MCP), plus U41 encoding the major single- stranded DNA-binding protein (MDBP or SSP) (Figure 3-4c) and also the novel EEHV- specific gene E4 encoding the captured vGCNT1 protein (Figure 3-4a). These proved to have wobble-position GC-contents compared to overall ORF GC-contents of 96.6(63)%,

89(65)% and 86(61)% respectively. Omitting 318-aa at the ends of the U57 (MCP) protein from the analysis leaves a central 3,000-bp segment with an extreme wobble position GC-content of 99.5%. The other two proteins also have a typical tendency towards deviations from the wobble position GC-bias near one or both ends. Almost all of the well-conserved common herpesvirus core proteins, as well as a subset of the more novel and diverged group-specific proteins in EEHV4 also follow this trend.

The major exceptions to this very high wobble position GC-content bias pattern include all 14 genes mentioned above that map within the six small well-dispersed loci with 48% or less overall GC-content (bolded in Table 3-1). Ten of these genes display wobble position GC-contents very close to their overall GC-content as follows: E4A =

46(44)%, E7B = 29(27)%, E9 = 31(27)%, E9A (vOGT) = 46(42)%, E9B = 26(23)%, E9C

= 25(27)%, E16D (vECTL) = 45(45)%, E17 = 53(45)%, E30A = 39(42)% and E37

(ORF-O) = 37 (41)%. In contrast, the E39A (ORF-R) glycoprotein shows an apparently enhanced bias in the opposite direction (ie towards even lower GC-richness) of 26% wobble GC-content compared to 41% overall GC-content. The first of those aberrant

52 genes is included in the three-frame GC-scan diagram shown in Figure 3-4a and the next seven in Figure 3-4b.

Interestingly, although 11 genes that do not follow the typical EEHV4 pattern of high-GC bias neither represent core genes nor are shared with any other virus groups, they do include a curious mixture of both obvious novel captured cell genes (vOGT and vECTL), plus the unique ORF-R glycoprotein and several other novel genes that are also not shared with EEHV1 and EEHV5 (E4A, E7B, E9C and E30A), together with some adjacent but highly diverged genes that are shared (E9, E9B, E17 and ORF-O). They even include two apparent members of the p3 or δ3 7xTM multi-gene family (E7B, E9), although the first of these has no direct matching identity to any other EEHV4 7xTM protein. Therefore, although some of the AT-rich genes may indeed be relatively newly- captured cellular genes, it would be hard to argue that they were all acquired by any kind of common mechanism or event.

The two adjacent large 1,465-aa and 2,065-aa leftwards-oriented potential transcriptional regulatory proteins E40 (ORF-K) and E44 (ORF-L) of EEHV4 are especially intriguing in this regard with unusual mixed bias patterns (Fig 3-4d, inverted orientation). Both have a total wobble position GC-content of just 69 or 63% matching closely to their overall GC-contents of 66 and 64%. However, this is not distributed uniformly, with just small 230-aa and 244-aa C-terminal domains in both (matching the segments that are conserved within the otherwise highly-diverged EEHV1 and EEHV5 orthologues) having wobble position GC-contents of 86% and 90%. In contrast, another segment of ORF-L that overlaps a potential second much smaller 296-aa coding region within a different reading frame here 14, known as ORF-EE1A, E44A or ORF-S, has a

53 wobble position GC-content of just 43%. However, this novel internal overlapping ORF-

S protein of EEHV4 itself (which is conserved at the 35% identity level in EEHV1 and

EEHV5) has a typical wobble position GC-content of 89% (overall 65%). Unexpectedly, in contrast to the situation in the AT-rich branch Probosciviruses EEHV1A (Kimba),

EEHV1A (Raman), EEHV1B (Emelia) and EEHV5 (Vijay) 5,16, where the entire ORF-L coding region and adjacent upstream region display highly localized CpG-suppression similar to that seen in the MIE1 (UL123) gene region of Cytomegalovirus and many other Betaherpesviruses 40, this feature is completely absent from the EEHV4 version of

ORF-L.

Complex Enlarged Ori-Lyt Dyad-Symmetry Domain

Similar to the other EEHVs, EEHV4 encodes a U73 orthologue of the alphaherpesvirus HSV UL09-type of origin-binding protein (OBP), but the expected matching dyad-symmetry Ori-Lyt region mapping between the U41 (MDBP) and

U42(MTA) genes of EEHV4 is much larger and more complex than those of EEHV1 and

EEHV5 and was initially very difficult to sequence because of unusual features attributed to several internal stem-loop structures. The entire combined inverted and direct repeat arrangement encompassing the predicted EEHV4A (Baylor) 1,180-kb Ori-Lyt domain from map coordinates 92,120 to 93,324 is illustrated in the dot matrix self-comparison plot presented in Figure 3-5a. A cartoon representation of the structure is also shown in

Figure 3-5b for comparison with the just 75 and 192-bp versions from HSV Ori-S, HSV

Ori-L, HHV6 Ori-Lyt and EEHV1/5 Ori-Lyt. The EEHV4 locus has a total of seven copies of the consensus OBP-binding sites (OBS) sequences (GAG)GGTGGAACG

54 present compared to just three to four in the other viruses. Four of these OBS motifs in

EEHV4 are arranged as pairs of direct head-to-head palindromic copies, whereas in the other viruses (as well as in a third pair in EEHV4) they mostly have five to seven bp gaps between the head-to-head binding site pairs.

Most of the other herpesvirus Ori-Lyt regions of this type 5,16,17,41 have a single central very AT-rich loop between the inverted repeated stems containing the OBP- binding sites (although this is duplicated in EEHV1 and EEHV5 compared to in HHV6), but the EEHV4 version instead consists predominantly of two large loops with 22 and 15 multimerized copies of 20 to 22-bp long AT-rich direct tandem-repeat elements bounded by three nearly identical 60 to 70-bp inverted repeats of the binding site containing stem- loops. There is also an additional pair of 40-bp inverted repeats encompassing this whole

800-bp block, with a few additional repeated elements lying within an apparent degenerate area beyond that on the left-hand-side.

Clustered Palindromic CREB-Binding Site Motifs

A very characteristic feature of primate cytomegaloviruses is the large and complex major immediate-early (MIE) enhancer domain mapping directly upstream from the genes encoding the spliced overlapping MIE1 (UL121) and MIE2 (UL123) transcriptional regulatory proteins. In particular, the 750-bp core HCMV MIE enhancer contains several sets of dispersed multimerized cis-acting elements encompassing consensus binding sites for the cellular transcription factors AP1, NFkB and CREB located upstream of the TATATAA-box motif. In particular, there are eight copies of the very high-affinity doubled-up 8-bp palindromic versions of cyclic AMP response

55 elements (CREs) that have in isolation been shown to provide both powerful basal transcriptional effects and responsiveness to cAMP, Ca2+-forskolin and TPA 42. These same multimerized cis-acting elements are also present in rearranged patterns within

AGM, rhesus and baboon CMV MIE enhancers, together with additional accessory control elements including tandemly-repeated further upstream NF1-binding site clusters and in some cases a large stretch of bent DNA 43-45. Well characterized repetitive control elements of this type are not present upstream of the equivalent MIE genes of

Muromegaloviruses nor Roseoloviruses, therefore it was surprising to find a closely spaced cluster of both consensus 8-bp (TGACGTCA) and lower affinity 5-bp (TGACG) half-site CRE motifs within the terminal repeat regions in all three of the EEHV1,

EEHV5 and EEHV4 genomes (Table 3-1). In EEHV1 there are six copies of the palindromic 8-bp CREs within a 153-bp segment and eight within an extended 590-bp segment, but no others at all within the rest of the entire 180-kb genome, whereas in

EEHV4 whilst four of the total of six 8-bp CREs occur within this cluster there are also six more of the 5-bp CREs within the same 540-bp block. By stochastic chance, a single

8-bp CRE motif should occur just once every 64,000-bp of average GC-content DNA and a second one within the same 540-bp region would occur about 150-fold less frequently.

Therefore, these are hardly random occurrences. Intriguingly, in the circularized form of the EEHV genomes the CRE clusters would all lie within a large non-coding region between 2.5-kb and 6.5-kb directly upstream of the candidate E44 (ORF-L) transcriptional transactivator protein gene at one end, as well as in EEHV1 and EEHV5 only just a few hundred bp upstream from the E47 (vFUT9) gene at the other end of the genome. A pair of these same 8-bp CRE motifs are also key functional elements in the

56 LTR enhancers of both HTLV1 and HTLV2. Therefore, based on the astronomical odds against them occurring by chance, together with the apparent strategical location, it seems reasonable to speculate that these clusters could play some important role in either packaging events or in transcriptional regulation in the EEHVs.

Chimeric Subtype Level Divergence Patterns from the Prototype EEHV4 (NAP22)

Genome

To address whether or not there may be chimeric domains of the CD-I, CD-II and

CD-III type described previously that distinguish EEHV1A from EEHV1B 5 and

EEHV5A from EEHV5B 5,9 selectively targeted PCR loci totaling 24.3-kb were Sanger sequenced from necropsy heart tissue DNA of the prototype EEHV4(NAP22) genome in addition to the 5.7-kb already available from that strain 3-5. Whilst most loci showed minimal differences at the nucleotide level (0.3 to 2.5%), three gene blocks encompassing glycoproteins U39 (gB), plus U46 (gN)-U47 (gO)-U48 (gH) as well as the

O-linked acetyl glucosamine transferase E9A (vOGT) instead proved to be highly diverged. The results across a contiguous 4.9-kb segment including gN to gH revealed a clearly defined contiguous chimeric domain of nearly 3.7-kb at Baylor equivalent coordinates 131,750 to 135,400, which almost exactly matches the same position and size as CD-II of EEHV1B compared to EEHV1A 5. This complete CD-II-like block displayed

26.3% total nucleotide divergence with the individual protein (and DNA) level differences being 9(11)%, 23(31)%, 32(37)%, 26(27)% and 13(13)% respectively for the intact E35A (ORF-J), U46 (gN), U47 (gO), U48 (gH) and nearly intact U48.5 (TK) genes. For the 1.6-kb segment of the U39 (gB) gene evaluated, the results for the EEHV4

57 (NAP22) version revealed a much smaller version of CD-I than in EEHV1 with 189 nucleotide polymorphisms over an internal 1,150-bp segment mapping between EEHV4

(Baylor) coordinates 99,993 to 101,150 representing 16.4% DNA level and 14% aa divergence with sharp chimeric boundaries on both sides. Similarly, because we could not amplify any of the expected gene block between UDG and the C-terminus of ORF-K for EEHV4 (NAP22) using EEHV4 (Baylor) based primers, we deduce that this latter region encompassing UDG, gL, ORF-O and ORF-R most likely represents a highly- diverged chimeric domain of a similar size and location as CD-III of EEHV1. Finally, a

6.6-kb sequenced block surrounding the novel captured vOGT gene proved to encompass a new fourth 4,660-bp chimeric domain designated CD-IV with sharp chimeric boundaries mapping at coordinates 20,541 to 25,200. This domain encompasses six genes including part of E7 plus all of E7B, E9, E9A (vOGT), E9B and E9C and displays 741- bp (16.0%) overall nucleotide polymorphism. The five intact proteins involved diverge by 28%, 22%, 13%, 18% and 28% respectively, whereas the values for the adjacent E7

ORF, all of E10A and part of E11 analyzed are instead just 3%, 2% and 0%. Intriguingly, in addition, the E9C and E9B ORFs have become fused in-frame in EEHV4A (NAP22) rather than occupying two different reading frames as in EEHV4B (Baylor).

A pictorial illustration of the high (= species-level) patterns of divergence between these first two known examples of EEHV4 strains within the equivalents of the

CD-I and CD-II chimeric blocks are presented in the Simplot comparisons given in

Figures 3-6a and 3-6b. These diagrams include matching to-scale comparisons with the equivalent CD-I and CD-II chimeric blocks of EEHV1B (Emelia) versus EEHV1A

(Kimba) as described by Richman et al 5. The apparent new CD-IV chimeric domain in

58 EEHV4 across the E7 to E9C gene block and including the novel captured E9A (vOGT) gene is also shown as a Simplot comparison in Figure 3-6c, but there is no known equivalent at this position in EEHV1A-1B or in EEHV5A-5B. In addition, an intermediate level of nucleotide variability of 11 to 12% was found at three other smaller loci for E4A, E31B and U50-U51, although 40% of the latter occurs within non-coding regions, whereas for the vGPCR1 ORF itself the DNA and protein variability levels only reach 7.8% and 6.7%, respectively. Some additional EEHV4 (NAP22) PCR sequencing was also carried out to confirm predicted ORFs and consensus splicing patterns at loci such as E4A, E4B, E4C, E6B, E12A, E16D (vECTL), E17, E17A, E18C, E20B, E23B and E31B that were somewhat ambiguous from the EEHV (Baylor) data alone. Overall, it seems amply justified to refer to EEHV4 (NAP22) as the prototype example of an

EEHV4A subtype, whereas EEHV4 (Baylor) would be the prototype of an EEHV4B subtype. The total of 30-kb of DNA sequence now available for EEHV4 (NAP22) represents nearly 15% of the total expected genome size, and gives an overall measured strain divergence from EEHV4B (Baylor) of 8.8% (2,684-bp polymorphisms). However, that value drops to just 1.9% when the 8.4-kb of data from the three observed CD-like chimeric domains and the variable vGPCR1 locus are omitted. Finally, it is also very evident when comparing the two chimeric subtypes of EEHV4 that the intergenic domains represent rapidly changing sequences.

Accuracy of the Predicted Spicing Patterns

Of necessity, because of the inability to grow EEHVs in cell culture, all predicted splicing patterns in EEHV genes are provisional. However there are multiple levels of

59 expected accuracy involved. All of our predictions are based on finding typical patterns of herpesvirus splicing involving strong to medium consensus donor and acceptor sites, generally small 60 to 150-bp introns and the need to jump across frameshifts in most of the introns to generate a logical intact protein. In addition, for EEHV1 we had relied on consistent conserved splicing signal locations and patterns across numerous distinct strains with sometimes otherwise quite variable sequences. For four of the predicted spliced proteins of EEHV4, namely U12 (vGPCR2ex1,2), U60-66(TERex1,2,3),

U42(MTAex1,2) and U37(ORF-Oex1,2,3), not only do the consensus motifs and patterns match those in multiple strains of EEHV1A5,17, but they each also agree with the predictions made by at least one other group for EEHV1A, EEHV1B or EEHV514,16,26.

The U12 pattern also matches that found in HCMV and the U42 (MTA) does so for the

EBV orthologue. For EEHV4 (ORF-Rex1,2) although it lacks homology with the positional equivalents ORF-P and ORF-Q of EEHV1A, EEHV1B or EEHV5, the predicted splicing is consistent with the similar interpretations for them by both our group5,17 and the Wilkie et al group14,16. Several other genes including E17 and E27 also match our predicted splicing patterns from multiple strains of the EEHV1 versions.

However, the situation for EEHV4 E24B (vOX2-B) just as for the highly diverged equivalents in EEHV1A, EEHV1B and EEHV5 is much more complex and speculative.

In all three of the latter, the authors of the complete genome papers made logical predictions of multiple spliced exons and proteins, which are conserved across many highly diverged strains (and even include a partial match to a host OX2 splicing motif), but they have no predictive value for the EEHV4 version. The simplest plausible interpretation that we could make was the presence here of two partially overlapping

60 proteins namely E24Bex1, 2 and E23B of which only the former has amino acid identity to viral and cellular OX2 proteins. However, the presence of multiple additional consensus splicing signals here could also be interpreted to generate four separate alternatively spliced proteins, including one that joins parts of both E24 and E23 from a total of six different exons plus four donor sites and four acceptor sites mapping in frame across the region between positions 40,982 and 50,001 in EEHV4B (Baylor).

Furthermore, all of those same signals and putative alternative forms are fully conserved in EEHV4A (NAP22).

Discussion

There were three major goals in determining the complete genome DNA sequence of EEHV4. The first was to learn more about the nature of this novel class of elephant herpesviruses and the range of genes and genetic variation that the different EEHV species and subtypes display. The second was to further address the question of whether the entire Proboscivirus genus would be best classified as just an outlier member of the

Betaherpesviruses or instead as the prototype of a distinct sub-family of the mammalian herpesviruses (the Deltaherpesvirinae) separate from the Alpha, Beta and

Gammaherpesvirus sub-families 5,6,17,46. Thirdly, questions had arisen about not only whether the AT-rich and GC-rich branches of the EEHVs might also be sufficiently diverged to justify separate genus status, but also more generically about the nature, origins and significance of this commonly encountered tendency amongst some other herpesvirus groups as well of trending towards extremely high GC-content.

61 The intact EEHV4 genome did prove to have many differences from the AT-rich branch viruses in a manner that is fully consistent with the predicted 30 to 35 million years since their last common ancestor as judged from the initial 5-kb of data 6. Some major features of these differences, especially relating to the extraordinarily high GC-bias found in many coding regions and the novel enlarged repetitive Ori-Lyt domain are described in detail here, whereas several other unique features relating especially to the enriched A plus T tracts in non-coding domains and the genus or sub-family specific sets of novel gene families and captured host genes, some of which are common to both the

AT-rich and GC-rich branches whereas others represent characteristic differences between them, will be described in detail in the accompanying paper by Ling et al 36.

Finally, we also showed here that even just the first two independent strains of EEHV4 examined display a significant level of localized EEHV4A-4B chimerism, with at least two of the apparent four such domains (CD-I, CD-II, CDIII and CD-IV) being quite similar to the patterns described previously for both EEHV1A-1B and EEHV5A-5B subtype pairs. Understanding such subtype and strain divergence patterns will be a critical factor if vaccination by non-pathogenic strains and routine monitoring of active infections are to become useful future management tools for EEHV hemorrhagic disease.

Our findings here of an extraordinary high GC-rich bias within the wobble codon positions of most but not all EEHV4 genes and the contrasting feature of numerous AT- rich sequence tracts lying within the intergenic domains in all of the EEHVs, raises the question of whether other herpesvirus genomes with similar overall GC-contents might show similar features. In fact HCMV, MCMV and the two highly diverged RCMV(M) and RCMV(E) genome types do all have a similar overall base composition of around 57

62 to 60% GC-content. The only previous attempt that we are aware of to evaluate GC- content patterns in those genomes was the work of Brocchieri at al 47, which used a measure referred to as S-content differences in the three different reading frames to attempt to identify or clarify additional potential coding genes in MCMV and RCMV

(M). As shown by Geyer et al 48 a small sub-set of those proposed corrected annotations were indeed later found to also be conserved in RCMV(E).

However, the major features of the S-profiles shown in the Brocchieri paper show similarity to the GC-wobble position biases that we described here for the EEHV4 genome. In fact, although those authors did not expressly say so, the S-values do serve as a measure of wobble position GC-biases, which are also fairly common in many of the core genes of those Muromegalovirus genomes. For example, in MCMV, a total of 40 genes mostly located within the central core domains between M33 to M105, plus M23,

M27, M115 and M139 to M143 show this feature with wobble position GC-biases approaching or exceeding 90%, and up to a dozen more genes either do so over at least half of their protein coding regions or exceed 75 to 80% bias. On the other hand, more than 70 other genes mostly mapping towards the ends of the genome instead show much less or no evidence for wobble-position bias. Notable examples of the latter include

M122 (IE1), M123 (IE2), M74 (gO) and the N-terminal half of M55 (gB).

The MCMV genome is also very similar to HCMV by having at least three large

AT-rich non-coding domains located around the Ori-Lyt, 5-kb stable intron and upstream

MIE loci, but with no more than five or so smaller intergenic domains that resemble the very large set in EEHV4 that have GC-contents below 50% with a surfeit of AT-rich tracts. Overall, these results engender some suspicion that in high-GC herpesvirus

63 genomes, the genes with high GC-rich wobble position bias generally tend to be well- established lytic cycle viral genes that function during late-lytic stages of virus infection, whereas those genes without any significant GC-bias may tend to be those that function at immediate-early times and during latency for example and therefore need to more closely resemble host genes, or alternatively may be genes that have been acquired relatively recently or evolved unusually rapidly.

A significant issue about EEHV pathogenicity that is not yet resolved is knowing whether the apparent observed much greater involvement of EEHV1A rather than

EEHV1B, EEHV4 or EEHV5 in lethal disease in Asian elephant calves reflects different pathogenesis mechanisms or efficiency per se, or perhaps simply reflects a much greater prevalence and universality of the former over the latter. It is also possible that the other

EEHVs including EEHV4 are just as ubiquitous in Asian elephant hosts as is EEHV1A, but instead exhibit a tighter and less frequently reactivated shedding relationship (thus appearing to be less abundant overall). Furthermore, whereas the overall detection rates for EEHV1B and especially EEHV4 and EEHV5 in asymptomatic Asian elephants have been quite low in random trunk wash shedding studies, the fact that all four viruses have nevertheless swept through most of both the adults and juveniles present in the closely monitored Texas zoo herd over a five-year testing period does tend to imply that infection by EEHV1B, EEHV4 and EEHV5 might also be just as ubiquitous in all Asian elephant populations as is EEHV1A. Once the differences or otherwise in pathogenesis of EEHV1,

EEHV5 and EEHV4 and their respective A and B subtypes are better understood, detailed comparisons of the sequences and gene content of all six genome types will hopefully provide important and useful information for future diagnosis, serological

64 evaluation and potential vaccine-based or other antiviral approaches to monitoring and controlling EEHV-associated hemorrhagic disease.

Acknowledgment

We thank Lauren Howard of the Houston Zoo for coordinating the routine blood and trunk wash screening of the herd, as well as the clinical care and treatment of the case. PDL received research funding from the Houston Zoo. SYL is the recipient of a

Morris Animal Foundation Postdoctoral Fellowship (D14ZO-411). Partial funding support for the studies at Johns Hopkins University was obtained by research grants to

GSH from the International Elephant Foundation. Both the PDL and GSH groups were also supported by subcontracts within a Leadership Grant (MG-30-130086-13) under the

Collections Stewardship Program awarded to Lauren Howard at the Houston Zoo by the

Institute of Museum and Library Services. The authors are grateful for the sequencing and analysis contributions of the production, Next-Gen, LIMS, Library, and systems teams in the Human Genome Sequencing Center (HGSC) led by Operations Director

Donna M. Muzny and Directed by Richard A. Gibbs. The HGSC is funded by the

National Human Genome Research Institute, National Institutes of Health grant U54

HG003273 to Richard Gibbs.

Author Contributors

AF carried out initial diagnostic real-time PCR analyses, and identified and prepared the best DNA sample for analysis. SYL carried out initial PCR sequencing at selected EEHV4B (Baylor) gene loci and to bridge the gaps between contigs, as well as

65 all of the EEHV4A (NAP22) PCR sequencing work. RSP carried out PCR sequencing to complete the terminal repeat regions of EEHV4B (Baylor). XQ, SD and KW coordinated

Illumina library preparation and sequencing. XQ performed the de novo genome assembly and analysis to identify viral contigs. GSH and SYH carried out all of the DNA sequence difference analyses and comparisons for identifying ORFs and generated the gene annotations for Genbank as well as the phylogenetic trees. GSH wrote several drafts and revisions of the manuscript. PDL coordinated the overall project, including the approaches to identifying the most suitable sample for the work and plans for generation of the sequence data. All authors read and provided input to the manuscript.

66 Tables

70 Figures

71 Figure 3-1. Annotated Physical Gene Map of the Complete EEHV4(Baylor)

Genome. The intact 206-kb EEHV4B (Baylor) genome determined here (Genbank

Accession No. KT832774) is depicted to scale. The relative sizes and orientations of all predicted open reading frames (ORFs) are indicated by horizontal arrows. Gene nomenclature is shown below each of the ORFs. The color key indicates groups of ORFs shared between all herpesviruses or subsets of herpesvirus sub-families or multiple paralogues of repetitive gene families. Grey arrows indicate captured cellular genes and white arrows denote novel genes that do not have obvious orthologues in EEHV1 or

EEHV5. Thin lines connecting arrows indicate introns. The position of the putative lytic replication origin is marked by a black rectangle.

Figure 3-2. Global Alignment Patterns For the Intact EEHV4 Genome Compared to

EEHV1 and HCMV. The dot matrix diagrams showing direct linear nucleotide alignments were generated as implemented at http://blast.ncbi. nlm.gov./Blast.cgi. Panel

73 (a): Comparison across the intact 206-kb genome of EEHV4 (Baylor, KT832477) from the GC-rich branch of the Proboscivirus genus versus the intact 180-kb genome of

EEHV1A (Kimba, KC618527) from the AT-rich branch of the Proboscivirus genus derived from Ling et al 17 when aligned in the same orientation. Panel (b): Comparison across the intact 206-kb genome of EEHV4 (Baylor, KT832477) versus the intact 235-kb genome of HCMV (Merlin, AY446834.2) in the Cytomegalovirus genus of the mammalian Betaherpesvirus sub-family derived from Dolan et al 49, with the latter aligned in the standard orientation.

75 Figure 3-3. EEHV4 Encodes a Very Large Family of Distantly Related Paralogous

7XTM and vGPCR-like Genes. Linear distance-based Bayesian bootstrap phylogenetic tree comparisons for all 26 members of the 7xTM containing multigene family from the

EEHV4 (Baylor) GC-rich branch Proboscivirus compared to their nearest host cell analogue Lox RAIP3 as the outgroup. The entire family is loosely divided into five subgroups whose designated prototypes of E15, E3, E18, E14 and E6 are indicated. Note that as indicated by the distance values all of these paralogues are very highly diverged from one another, with the exception of E14.1 and E14.2 which most likely represent the most recent duplication event. A subset of the genes in this family (p3 or δ3) exhibit features of GPCR genes as described in greater detail in the accompanying paper 16.

Figure 3-4. Codon-Specific Scanning GC-Content Panels Showing the Wobble

Codon GC-Bias Effect Across Selected Representative Segments of the EEHV4

(Baylor) Genome. Diagrams showing the percentage G plus C-content of each of the three potential translated codon frames across four selected 18-kb segments of the

EEHV4B (Baylor) genomic DNA sequence as implemented under the codon-specific G plus C% toolbox item in MacVector 12. Short vertical bars indicate forwards direction terminators. Annotated ORF positions and sizes are denoted by open arrows. Highly GC-

77 biased wobble position blocks with average values between 80 to 100% are marked with solid bars. For a hypothetical ORF with an initiator codon beginning in frame 1 at position x in the diagram the wobble position codons are represented by the succeeding frame 3 line (or frame 1 for a frame 2 initiator and frame 2 for a frame 3 initiator). Panel

(a): Forwards-directed strand across coordinates 1 to 18,000 at the extreme left-hand side encompassing 10 out of 11 rightwards-oriented genes from E1 to E4C including E4

(vGCNT1) with high wobble position GC-bias. Panel (b): Inverted segment of the complementary strand across coordinates 37,000 to 19,000 encompassing predominantly leftwards-oriented genes (16x between E6 and E16D) and two rightwards oriented genes

(E9A and E17), including 11 genes displaying uniformly high wobble codon GC-bias plus seven genes in two blocks E7B-E9 (vOGT)-E9A-E9B-E9C and E16D (vECTL)-

E17-E17A that do not display wobble GC-bias (all of the latter are labeled with asterisks*). Panel (c): Forwards strand across coordinates 86,000 to 104,000 encompassing seven rightwards-oriented core region genes U43 (PRI) to U38 (POL) with high wobble position GC-bias on either side of the predicted novel Ori-Lyt domain.

Panel (d): Inverted segment of the complementary strand from the extreme right-hand- side across coordinates 187,867-205,894 encompassing U44A (ORF-S), U44 (ORF-L) and U40 (ORF-K). The only three high GC-bias wobble codon blocks found within this region occur in ORF-S and in the conserved C-terminal domains of ORF-L and ORF-K

(marked with solid bars).

79 Figure 3-5. Complex Repeat and Inverted Repeat Patterns within the Predicted Ori-

Lyt Domain of EEHV4. Panel (a): Dot matrix self-comparison of the DNA from coordinates 92,021 to 93,860 encompassing the entire intergenic region between the N- terminus of U41 (MDBP) and the C-terminus of U42 (MTA) from EEHV4 (Baylor). The proposed 1.2-kb Ori-Lyt domain spans from coordinates 92,120 to 93,325. Direct tandemly-repeated structure is indicated by additional lines parallel to the main diagonal, whereas inverted repeats are indicated by additional lines perpendicular to the main diagonal. Panel (b): Features of the Unusual Expanded Dyad Symmetry Ori-Lyt Region of EEHV4 Compared to those of EEHV1 and Other Alphaherpesvirus-like Dyad-

Symmetry Type Origins. Cartoon diagram comparing the sizes and major structural features of the predicted dyad symmetry domains of EEHV4 (Baylor) and EEHV1A

(Kimba) with those of the HHV6 version and with both Ori-L and Ori-S of HSV1. Filled circles denote alternating A plus T-dinucleotide runs. Short horizontal pointed bars represent copies of the consensus OBP binding site motif. Other larger sets of arrows designate various types of inverted repeats as well as the 30x and 20x copies of the 20-bp

AT-rich direct tandem-repeats in the EEHV4 (Baylor) version.

Figure 3-6: Positions and Sizes of Three Identified EEHV4A-EEHV4B Chimeric

Domains and Boundaries Relative to Those of EEHV1A-EEHV1B Chimeric

Domains. The diagrams show SimPlot (30) comparisons of the nucleotide identity

81 patterns between EEHV4A (NAP22) and EEHV4B (Baylor) across three mapped chimeric domains CD-I (1.1-kb), CD-II (3.7-kb) and CD-IV (4.7-kb) shown as blue lines in comparison to superimposed data for CD-I (3.2-kb) and CD-II (3.7-kb) of EEHV1A

(Kimba) versus EEHV1B (Emelia) shown as red lines. CD-IV of EEHV4 has no equivalent in EEHV1 and there is no data available for the presumed region CD-III of

EEHV4. Panel (a): CD-I Chimeric Region within U39 (gB) of EEHV4A versus

EEHV4B at EEHV4 (Baylor) map coordinates 99,993 to 101,150 compared to the much larger overlapping CD-II of EEHV1A versus EEHV1B which encompasses part of U40, all of U39 (gB) and part of U38 (POL). Vertical arrows denote the positions of the

EEHV1A-1B chimeric domain boundaries. Panel (b): CD-II Chimeric region encompassing part of ORF-J, all of gN-gO-gH and part of TK of EEHV4A versus

EEHV4B at EEHV4 (Baylor) map coordinates 131,750 to 135,400 compared to the nearly equivalent superimposed CD-II of EEHV1A versus EEHV1B. Vertical arrows denote the positions of the EEHV1A-1B chimeric domain boundaries. Panel (c): CD-IV chimeric domain of EEHV4A versus EEHV4B mapping between EEHV4 (Baylor) coordinates at map coordinates 20,541 to 25,210. Vertical arrows denote the positions of the EEHV4A-4B chimeric domain boundaries that encompass part of E7 and all of E7B,

E9, E9A (vOGT), E9B and E9-C, but end before E10A.

IV. Attempts to propagate elephant endotheliotropic herpesvirus 1A (EEHV1A) in elephant umbilical venule endothelial cells (EUVEC)

83 Abstract

A highly lethal bleeding disease caused by a novel herpesvirus species designated as elephant endotheliotropic herpesvirus 1A (EEHV1A) is posing a major threat to the successful breeding of future generations of highly endangered Asian elephants (Elephas maximus) worldwide both in the wild and in ex situ breeding programs. Efforts to understand and control these viruses depend greatly upon scientists being able to generate large quantities of virus in the laboratory. The ability to grow and propagate this virus in the laboratory conditions is critical for the progress of future research towards understanding the biology and pathology of this virus, as well as developing assays to test drug inhibitors and the possibility of generating live attenuated vaccines, making this the top priority goal of EEHV research. Here we report overlaying of fresh minced up necropsy heart, tongue and lung tissues from a 5-year-old female Asian elephant that succumbed to an EEHV1A infection onto both primary EUVEC (pEUVEC) and immortalized EUVEC (iEUVEC) cultures for three days. After the removal of necropsy tissue and despite the routine cell culture medium replacement and passaging of cell cultures, EEHV1A DNA polymerase (POL) and/or helicase (HEL) gene DNA was persistently detectable with first round conventional polymerase chain reaction (PCR) up to 54 and 122 days in the iEUVEC and pEUVEC cultures, respectively. Quantitative

PCR (qPCR) for EEHV1A DNA in the pEUVEC cultures overlay with heart tissue revealed decreasing viral DNA over time. These findings suggest viral entry into both the

EUVEC cultures followed by a decrease of viral DNA attributed to passaging and/or die off of EUVEC with no evidence of viral DNA increase. Further work is needed to confirm viral entry and latency after viral introduction to the EUVEC cultures.

84 Introduction

The lethal acute hemorrhagic disease caused by elephant endotheliotropic herpesviruses (EEHV) has been responsible for 60% of all deaths among Asian elephant calves between 4 months and 15 years of age that were born in zoological parks in North

America8. More than 100 cases with an 80% mortality rate have been confirmed in

Europe and North America (representing 20% of all live births)8 and as many as 20 plus cases have also been confirmed or suspected among both orphan and free ranging calves in Asia. Despite some Asian elephant calves surviving after clinical intervention, there were still 14 Asian elephant calf deaths in Europe caused by EEHV within the last 6 years and 4 recent confirmed North America EEHV1 associated deaths within the last 2 years.

Several previous attempts to propagate EEHV in laboratory cell culture by standard virological approaches using clinical tissue or blood samples collected from

Asian elephant calves with systemic hemorrhagic disease have not been successful, despite being effective with other well-characterized herpesviruses. Previous attempts in our group to propagate virus from infected samples included using and passaging standard primary (HFF, MEF, REF), established human (Hela, K562, U373, BJAB, 143), mouse (NIH-and BALB/c3T3, Ltk-), and hamster (BHK) cell lines, as well as obtaining primary cultures of both Asian and African elephant fibroblast cells from the San Diego

Zoo’s collection. Attempts to grow EEHV in embryonated chicken eggs and inoculated laboratory mice were also unsuccessful (Laura Richman, unpublished data), although this is a well-established technique for propagating human herpes simplex virus (HSV)50-52.

85 Our lab previously found some success using an alternative approach that was first used to demonstrate infectious Kaposi’s sarcoma-associated herpesvirus (KSHV) derived from primary effusion lymphoma (PEL) cells lines in contact-inhibited primary human DMVEC cultures53. Using this technique, we successfully produced latency associated nuclear antigen (LANA) positive colonies that later turned lytic either spontaneously or after tetradecanoyl phorbal acetate (TPA) or n-butyric acid (nBA) induction treatment, as demonstrated by immunoflurorescence assay (IFA) with rabbit antibodies against viral lytic cycle proteins. Nevertheless, using similar techniques for

EEHV propagation was still not achieved within a laboratory setting.

Further experiments were also conducted in primary elephant peripheral blood mononuclear cell (PBMC) cultures by adding supernatant or live cells from EEHV positive blood, including attempts to induce with TPA and other agents and after isolating adherent monocytes, differentiating monocytes into macrophages and after several weeks of propagation of T-cells and or B-cells in the presence of IL-2, IL-4, TNF alpha and other agents. None of these approaches succeeded in producing lytic cycle EEHV progression at the level of visible cytopathic effects or positive IFA with antibodies to early lytic cycle nuclear antigens MDBP or OBP. Nevertheless, the ability to grow and propagate this virus is so critical for future progress towards understanding the biology and pathology of this virus, as well as for the possibility of generating live attenuated vaccines, that additional attempts are planned and necessary.

If successful viral propagation in cell culture can be achieved, researchers will then gain the much needed ability to (a) accumulate large quantities of virus for easier molecular analysis of all the novel genes and proteins these viruses encode, (b) study

86 which cell types these viruses can infect, (c) study the virus life cycle including how

EEHV establishes and reactivates from latency, (d) make progress towards understanding the basis and mechanisms involved in the unusual pathology of EEHVs in vascular endothelial and white blood cells, (e) carry out pharmacological studies evaluating the effectiveness of the human anti-herpesvirus inhibitory drugs and/or screen for additional drug candidates, and finally, (f) develop virologic methods to passage attenuated virus for potential vaccine development. Currently, there has been great progress in our ability to monitor and quantitate EEHV infections in blood and trunk wash secretions, as well as to assess the genetic relationships between the multiple different species and strains of

EEHV. However, without the ability to successfully cultivate and propagate the virus in the research laboratory, all progress in understanding the nature of these viruses and ultimately gaining ability to control disease progression will be distressingly slow.

Here we report persistence EEHV1A viral DNA PCR detection in EUVEC cultures overlay with fresh minced Asian elephant necropsy tissue samples from a confirmed EEHV1A associated death for several months in spite of multiple cell culture passages and medium changes.

Materials and Methods pEUVEC cultures

pEUVEC were established from fresh unfrozen Asian elephant umbilical cord harvested immediately after birth. Umbilical cord was collected in 20 to 30 cm segments stored in sterile phosphate-buffered saline (PBS) with 1X concentration of Antibiotic

Antimycotic solution (Sigma-Aldrich Co., St. Louis, MO), and was transported on

87 ice/refrigerator packs within 24-48 hours as listed in the EEHV Research and Tissue

Request Protocol.

EUVEC were isolated following the protocol for “Isolation and culture of human umbilical vein endothelial cells” 54 with the following modifications. All procedures were performed in a biological safety level 2 cabinet under sterile and aseptic techniques.

Umbilical cord veins were cannulated with a blunt 10-gauge needle and perfused with 50 ml of sterile PBS twice at both ends to wash out blood. Each vein was then infused with

10 mL warm (37°C) PBS with 0.1% collagenase (Invitrogen, Carlsbad, CA) with the other end clamped off with a hemostat. Umbilical veins were incubated for 10 min at

37°C. Immediately following incubation, the clamped umbilical cords were briefly massaged to loosen the cells, the cord was unclamped, and the collagenase solution was flushed from the cord with 30 ml of PBS. The recovered umbilical extract containing umbilical vein endothelial cells were collected into a 50 ml conical polypropylene tube containing Clonetics EBM-2 medium (Lonza, Walkesville, MD) supplemented with

EGM-2 singleQuots (Lonza, Walkersville, MD), 1X concentration of Antibiotic

Antimycotic, and 2% fetal bovine serum. The cells were pelleted at 250 x g for 10 minutes at room temperature. The supernatant was decanted and the pellet was resuspended in 5 ml of Clonetics EBM-2 medium supplemented with EGM-2 singleQuots, 1X concentration of Antibiotic Antimycotic solution, and 2% fetal bovine serum and plated into a T25 cell culture flask. Cells were observed daily for 1 week or until confluent. 1X Antibiotic Antimycotic PBS solution was used to rinse cells with fungal or bacterial contamination. All persistently contaminated cultures were discarded after a week. pEUVEC cultures were maintained in T25 cell culture flasks with Clonetics

88 EBM-2 medium supplemented with EGM-2 singleQuots and 2% fetal bovine serum at

37°C with 5% CO2. pEUVEC retained the ability to divide and form cobblestone-like contact-inhibited monolayers (Figure 4-1).

Creation of iEUVEC cultures

Proviral vectors were received from Addgene (Cambridge, MA). The lentiviral transfer plasmid pBABE-zero largeTcDNA (Addgene plasmid # 1779) and pBABE-neo- hTERT (Addgene plasmid # 1774) were gifts from Bob Weinberg. pEUVEC were transduced with plasmids for both large T antigen (SV40TL) and human telomerase

(hTERT) using a vesicular stomatitis virus-G protein pseudotyped retrovirus. Transduced cells were selected for drug resistance with 200-250 μg/ml of gentamycin and 50 μg/ml of zeocin. iEUVEC were maintained in T25 cell culture flasks with Clonetics EBM-2 medium (Lonza, Walkersville, MD) supplemented with EGM-2 singleQuots (Lonza,

Walkersville, MD) and 2% fetal bovine serum. iEUVEC retained the ability to divide and form cobblestone-like contact-inhibited monolayers.

Necropsy samples

A 5-year-old female Asian elephant (Diazy) from the Albuquerque BioPark, New

Mexico died from an acute hemorrhagic disease on May 9, 2015. PCR confirmed

EEHV1A detected in blood, serum, and other visceral organs. The necropsy was performed on May 11, 2015 by the zoological staff. Samples were collected according to the EEHV Research and Tissue Request Protocol. Received samples included frozen whole blood, serum, and fresh organ tissue samples. 2 ml of frozen whole blood and 2 ml of serum were shipped in an EDTA tube and screw-top tube, respectively. Fresh unfrozen

89 heart, liver, tongue, spleen, and lung tissue samples were aseptically placed into 50 ml conical polypropylene tubes with sterile PBS with 1X concentration of Antibiotic

Antimycotic solution shipped on ice. All samples were received at Johns Hopkins

University School of Medicine on May 12, 2015.

All fresh tissue samples were grossly normal. Tissue samples were minced following the standard procedure reported by Latimer 2011 4. Approximately 20 mg of heart, liver, and spleen were individually minced then resuspended in 1 ml of cold sterile

PBS. Approximately 10 mg of tongue and 10 mg of lung tissue samples were minced together then resuspended in 1ml of cold sterile PBS. 200-μl of each fresh tissue sample suspension were overlay into each T25 cell culture flasks with either pEUVEC or iEUVEC for 3 days with Clonetics EBM-2 medium with supplemented with EGM-2 singleQuots, 1X concentration of Antibiotic Antimycotic, and 2% fetal bovine serum.

After 3 days, medium and fresh tissue sample suspension mixture were removed followed by a 5 ml sterile 37°C PBS wash and replenished with fresh medium.

After overlay with fresh tissue samples, both the infected pEUVEC and iEUVEC culture flasks were allowed to become confluent with no passaging for one month maintaining a monolayer from contact inhibition. Further culture steps include multiple successive passaging with added fresh uninfected cells.

DNA extraction from fresh initial samples

200-μl of each initial fresh tissue PBS suspensions were used for DNA extractions using the Qiagen Generation Capture kit (Qiagen, Valenica, CA) with the DNA eluted into 100-μl elution buffer.

DNA extraction from EUVEC cultures

When passaging pEUVEC and iEUVEC with trypsin/EDTA (Lonza,

Walkersville, MD), a third (6.6x105 EUVEC) of a confluent T25 cell culture flask (2x106

EUVEC) of EUVEC were collected for DNA extraction. DNA extractions were performed using the Qiagen Generation Capture kit listed above. The only exception from this procedure was the first passaging (32 days later) of the iEUVEC cultures with either heart or tongue/lung where two-thirds of a confluent T25 cell culture flask (1.3x106

EUVEC) were used for DNA extraction. On daily observation of cultures, medium with noted cellular debris or cloudiness was collected, spun down to separate cellular pellet and supernatant for DNA extraction using the same procedure stated above.

PCR amplification

PCR amplification and the PCR primers for EEHV1 POL and HEL were as previously described 4, 29.

DNA sequencing

All DNA sequencing was carried out by direct cycle sequencing on both strands of purified PCR DNA products from first round PCR amplification. The correct sized

PCR product bands were purified after agarose gel electrophoresis with a Qiagen II Gel

Extraction kit (Valenica, CA). Purified PCR products were sequenced at the Johns

Hopkins University School of Medicine Genetic Resources Core facility. All DNA sequence editing, analysis and manipulation was performed using Assmeblialign and

91 CLustal-W nearest neighbor joining as implemented for MacVector vers 7, together with

BLASTX or TBLASTX comparison programs provided online at NCBI.

qPCR for EEHV1

qPCR for viral DNA levels of EEHV1 based on the EEHV1 major DNA-binding protein (MDBP) were performed as previously reported 29 and performed at the National

Elephant Herpesvirus Laboratory at the Smithsonian National Zoological Park. 5-μl of template DNA was used for each qPCR reaction performed in duplicates with the exception of the heart necropsy DNA extraction using only 1-ul per reaction. The limit of detection is established as 10 copies/reaction with 100% sensitivity as previously reported 29.

Results

EUVEC cultures with fungal contamination or nonviable EUVEC

The pEUVEC heart, iEUVEC heart, and iEUVEC tongue/lung cultures were the only viable or uncontaminated EUVEC cultures after the overlay with fresh necropsy samples. All other cultures were disposed of due to nonviable EUVEC cells or due to persistent fungal contamination within the 3 days overlay period. Source of the fungal contamination is attributed to the non-sterile tissue samples collected during the necropsy process.

Detection of viral EEHV1A DNA in iEUVEC for 54 days after overlay with fresh heart and tongue/lung necropsy tissue

92 The iEUVEC tongue/lung and heart cultures were first passaged for 32 days keeping a third of the confluent T25 cell culture EUVEC to reseed the cultures while using two-thirds of the remaining cells (approximately 1.3x106 EUVEC) for DNA extraction. Subsequent passages of cultures only utilized a third of a confluent T25 cell culture flask (6.6x105 EUVEC) for DNA extraction. First round conventional PCR for

EEHV1 POL yielded the proper-estimated sized 520 bp band for both the tongue/lung and heart cultures. First round PCR for EEHV1A HEL amplified the estimated sized 980 bp band (Figure 4-2) and was submitted for sequencing with results consistent with the same EEHV1A strain detected in necropsy samples containing the unique polymorphism

(Figure 4-3). Detection of EEHV1A POL continued until day 54 with failure to detect on first round conventional PCR in iEUVEC heart and tongue/lung cultures at 61 days. A total of 6 passages were performed after the overlay with both fresh heart and tongue/lung by the last PCR detection (32 days, 40 days, 45 days, 47 days, 52 days, and

54 days). Each passaging resulting in an approximately 1/3 dilution with an estimated calculated 1/729 or a 1.3x10-3 overall dilution of the original input viral DNA amount of the overlay even by the last passage on 64 days.

Detection of EEHV1 DNA in pEUVEC for 122 days after overlay with fresh heart tissue

The pEUVEC heart culture was maintained for 47 days then split into two T25 cell culture flasks with the remaining third of the EUVEC collected for DNA extraction.

First round PCR for EEHV1A HEL DNA was amplified from this DNA extraction. Last detection with first round PCR for EEHV1 POL was on pEUVEC heart passaged at 122 days but failed to amplify on the following passage at 126 days. A total of 7 passages of

93 the pEUVEC heart flask (54 days, 66 days, 78 days, 91 days, 119 days, 122 days) were performed. Each passaging step resulted in a 1/3 dilution giving in a 1/2187 or a 4x10-4 dilution of the total original input of viral DNA amount of the overlay event by the last passage 122 days later (Figure 4-4).

Initial qPCR measurement in pEUVEC heart cultures

qPCR detected 1.3 X 105 viral genome equivalents (VGE) in 1-μl of extracted

DNA from the fresh heart PBS suspension with calculations estimating approximately

1.3x107 VGE overlay onto 2x106 EUVEC cells per flask with a viral to cellular ratio of

6:1. qPCR quantitation of DNA extracts for heart pEUVEC culture passaged for 47 days,

54 days, 66 days and 73 days yielded 100, 250, 60, and lesser than 60 VGE/reaction respectively with total VGE in the final culture estimated to be 6x103, 1.5x104, 3.6x103, and lesser than 60 VGE per T25 cm2 flask respectively (Figure 4-4).

Discussion

The continued detection of persisting EEHV1A DNA using either cPCR or qPCR in a laboratory cell culture system has not been reported previously to our knowledge.

Sequencing of the amplified EEHV1A DNA from both iEUVEC and pEUVEC using conventional PCR revealed the same novel HEL gene polymorphism as detected in the original necropsy tissue samples confirming that it did not represent contamination from any other viral genome source. First round cPCR for EEHV1A POL revealed persistence of viral DNA in the pEUVEC heart culture for 122 days and similarly for 54 days in the iEUVEC heart and lung/tongue culture after overlay.

94 With the removal of the tissue sample PBS suspension overlay after 3 days, plus routine medium changing (2-3 days intervals), and with continuous passaging of cultures over the course of maintenance and monitoring these EUVEC cultures, viral EEHV1A

DNA was detected at multiple successive passaging steps. After all these procedures for several weeks to months, the source of viral DNA would be assumed limited to within cells as a possible circularized viral genome into a nuclear episome. The basic maintenance of cell culture procedures should remove any free virus or viral DNA within the supernatant. Surface viral attachment to EUVEC should be considered but it seems highly unlikely that the viability would remain for such an extended duration.

Although it is not known how much of the input viral DNA was lost at the first wash step, the simple direct 3X dilution factor occurring during each of the subsequent passaging steps would be expected to reduce the overall DNA level (if it just persisted throughout) by 2,500x fold (i.e. down to 1x103 VGEs). Therefore, there was clearly no overall increase in total viral DNA over the course of the experiment. However, at 4th passage, the residual amount of viral DNA present appears to exactly match the expected dilution factor if all the input DNA was adsorbed or attached to the EUVEC monolayer with no overall gain or loss. Therefore, theoretically some combination of the following scenarios might have occurred: (1) All of the input DNA was in live tissue cells (likely either macrophages or vascular endothelial cells) that became attached and survived without dividing throughout the full three months; (2) Just a small fraction of the input cells containing viral DNA became attached and either these cells divided or the viral genomes replicated to some extent to maintain the levels seen; (3) The same occurred as in (1) and (2) above except that the viral DNA was transferred to the new uninfected

95 cultured EUVEC and replicated at low levels in them; or (4) Free virions possibly present within the input hemorrhagic tissue truly established new low level infections in the cultured EUVEC cells. Nevertheless, because no cytopathic effects nor positive IFA with any of the three different rabbit antibodies against EEHV lytic cycle nuclear antigens were ever detected (as done in other similar experiments), we concluded that the persisting viral DNA genomes whatever cells they had entered (or attached to) were probably maintained in some form of latent rather than lytic cycle state but that even this was not permanent.

qPCR quantification of DNA from limited passages from pEUVEC heart cultures

(47 days, 54 days, 66 days and 73 days) revealed a trend for decreasing amounts of

EEHV1A DNA based on the EEHV1A MDBP real-time assay described above (Figure

4-4). A decrease in any persisting original input viral DNA levels should be assumed based on each passaging event initially reducing the number of cells in the culture to approximately a third of those in the previous flask (although the overall number of live

EUVEC cells clearly increased about three-fold again before the next passage). Estimated calculations of 1.3x107 VGE in the heart tissue PBS suspension overlay was determined or a viral to cellular ratio of 6:1. A measurement of VGE in the removed tissue sample

PBS and medium mixture after the initial 3 day overlay period was not obtained but would have given an accurate number of the amount of VGE/virus remaining in the flask and possible levels of cellular entry. Without knowing this remaining VGE amount after overlay, we used the beginning heart input 1.3x107 VGE as a starting point and applied a rough calculation that only a maximum of one third of the input unreplicated VGE could remain after every passage event yielding 5.9x103 and 1.8x104 VGE for pEUVEC (7 total

96 passages) (Figure 4-4) and iEUVEC (6 total passages). Two qPCR measurement from the pEUVEC heart culture from passage 54 days and 66 days extrapolated to be 1.5x104 and 3.6x103 VGE, respectively showing a four fold decrease compared to a three fold decrease. Cell death should be considered as a possible cause for the additional decrease.

These numbers indicate that there was no overall increase in viral DNA copy number or that of there was replication of input genomes that was only enough to replace the genome lost because of to presumed cell death.

The inconsistency between first round cPCR with the qPCR results puts into question if EEHV1A DNA truly persisted till 122 days in the pEUVEC heart culture or was it more similar to iEUVEC cultures which were limited to 54 days consistent with the lack of detection of viral DNA at 73 days. The lowest limit for first round conventional PCR band detection is 1000 VGE with the ability to perform additional second and third rounds for lower limit detection. qPCR measurement is the preferred total method for quantitation with lower limits of 10 viral copies/reaction 29 or 600 viral copies per T25 flask. Limited qPCR measurements (47 days - 73 days) failed to detect any viral copies from passaged iEUVEC heart passaged DNA after 73 days in contrast to the last conventional PCR detection at 122 days in the iEUVEC heart culture. Additional qPCR measurements of both iEUVEC and pEUVEC from multiple passage dates with multiple replicates are needed to obtain an accurate measurement of the viral levels and degree of decrease to attribute to passage dilution, and/or die off.

At this point, we can only assume that EEHV1A, like all herpesviruses, exhibits two distinct phases of infection: latency and lytic replication. During primary infection, herpesviruses access the cell followed by entering a latency phase where the viral

97 genome is circularized and exist as a nuclear episome through multiple host cell divisions and expressing a subset of latent specific viral genes. Shifting from latency to lytic replication producing infectious virion particles generally requires chemical agents or environmental stress. Detection of viral EEHV1A DNA persisting over time and even after routine cell culture maintenance (changing medium, washing with PBS, passaging, etc) is suggestive of viral entry of EEHV1A into the EUVEC. Virus or viral DNA is unlikely to remain stable without cellular entry over this period of time. EEHV1A latent in the EUVEC cells would explain the persistence of detection of viral DNA after removal of tissue sample PBS suspension removal and routine cell culture maintenance.

In spite of the decreasing trend of viral EEHV1A DNA directly due to passage dilution and possible die off, the PCR data does not display any increase in viral DNA suggesting

EEHV1A latency in these EUVEC cells. Future work is needed to definitively confirm the presence of the nuclear episome within cells.

Potential sources of EEHV1A viral DNA in the cultures are viable cells that adhered to the cell culture flask during the overlay period. We believe this is unlikely, as we received these necropsy tissue samples 3 day postmortem, well beyond point of viability of these tissue cells. The only EUVEC cultures that either were viable with detectable EEVH1A viral DNA with no fungal contaminations were the pEUVEC heart, iEUVEC heart, and iEUVEC lung/tongue cultures. Daily visual inspection of the cultures revealed a monolayer of cobblestone EUVEC with no visual detection of primary cells from the tissue sample suspensions. We believe the potential source of EEHV1A DNA from viable cells from necropsy tissue is small to none.

98 For all except the first passaging step for both the “infect” pEUVEC and iEUVEC cultures that were positive for viral EEHV1A DNA, a third of the EUVEC were cryopreserved and stored in -80°C. However, multiple attempts to resurrect confirmed positive cells at 91 and 119 days for pEUVEC failed to detect EEHV1A DNA based on conventional PCR once cultures reached confluence. This is consistent with the qPCR results where iEUVEC heart cells after 73 days later did not contain any EEHV1A DNA.

Future directions include additional quantifications of viral EEHV1 DNA with qPCR for both pEUVEC and iEUVEC, for different passage dates, and from the removed tissue sample/medium mixture after overlay period to measure an accurate viral DNA input into each culture. Detection of viral nuclear episome with EUVECs positive for viral EEHV1A DNA confirming viral entry and latency is also planned. Frozen samples of the same original heart tissue suspensions have been stored at -80°C for additional cell culture experiments, which should include adding chemical activators such as TPA or nBA to induce lytic replication of herpesvirus in endothelial cell cultures 53.

In summary, this is the first report of persistence EEHV1A DNA PCR detection within a cell culture system in a laboratory setting infected with necropsy tissue samples from a confirmed EEHV1A associated death. Additional work is needed to confirm viral entry and to progress into viral propagation to generate large quantities of virus for research to develop successful treatment methods for EEHV1A disease.

Acknowledgement

I would like to thank our colleague Sarah Y. Heaggans for assisting me with the cell culture work and for carrying out the PCR phylogenetic analyses. I would also like to

99 thank Colette apRhys for helping me establish the pEUVEC and aiding me in creating the iEUVEC. The qPCR analysis were performed at the NEHL and were supported by research grants to Erin Latimer and Timothy Walsh from the Smithsonian Institution,

National Zoo, the International Elephant Foundation, the Morris Animal Foundation, and the Ringling Bros and Barnum and Bailey Center for Elephant Conservation. We are especially grateful to the zoo veterinarians and staff of the Albuquerque BioPark who provided clinical and pathological samples for this research. I would like to thank the

Morris Animal Foundation for the postdoctoral fellowship D14ZO-411. Studies at Johns

Hopkins University were supported at various times by research grant R01 AI24576 to

G.S.H. from NIAID, National Institutes of Health, DHEW, the International Elephant

Foundation, and a subcontract to Lauren Howard at the Houston Zoo under an IMLS

National Leadership Collections Stewardship Program. And special thanks to Sarah Beck from the Molecular and Comparative Pathobiology department at Johns Hopkins

University School of Medicine for initial revisions.

100 Figures

Figure 4-1. Microphotograph of pEUVEC. 10X microphotograph of pEUVEC retaining the cobblestone-like contact-inhibited monolayers within a T25 flask.

101 1 2 3 4 5 6 7

980 bp

Figure 4-2. Agarose gel-ethidium bromide of PAN-HEL PCR. Agarose gel-ethidium bromide PCR band results obtained after first-round (980 bp) amplification with elephant endotheliotropic herpesvirus PAN-HEL primers from DNA extracted from iEVUEC, cell culture supernatant and necropsy blood. Lane 1, tongue/lung iEUVEC; Lane 2, heart iEUVEC; Lane 3, supernatant from tongue/lung iEUVEC culture; Lane 4, supernantant from heart iEUVEC culture; Lane 5, Diazy whole blood (positive control); Lane 6, iEUVEC (tissue negative control); Lane 7, blanket (negative control).

102

Figure 4-3. DNA sequence alignment between detected EEHV HEL from heart iEUVEC and whole blood necropsy sample. DNA sequence alignment for first-round

EEHV PAN-HEL (980 bp) between EEHVSLDAIZYINFHEL719-2.SEQ (heart iEUVEC) vs. EEHVSLDAIZYWBHEL719-5.SEQ (Diazy whole blood) generated using

MacVector with Assember v12.0.6.

103

Figure 4-4. Levels of VGE in heart pEUVEC between estimated compared to qPCR detected based on the number of passages. For the estimated pEUVEC VGE, flask was reseeded with 1/3 of the total flask cells resulting in a 2/3-fold decrease in VGE for each consecutively passage.

104 References

1. Richman LK, Montali RJ, Garber RL, et al. Novel endotheliotropic herpesviruses fatal for Asian and African elephants. Science (New York, NY). 1999;283(5405):1171-1176. 2. Atkins L, Zong JC, Tan J, et al. Elephant endotheliotropic herpesvirus 5, a newly recognized elephant herpesvirus associated with clinical and subclinical infections in captive Asian elephants (Elephas maximus). J Zoo Wildl Med. 2013;44(1):136- 143. 3. Garner MM, Helmick K, Ochsenreiter J, et al. Clinico-pathologic features of fatal disease attributed to new variants of endotheliotropic herpesviruses in two Asian elephants (Elephas maximus). Vet Pathol. 2009;46(1):97-104. 4. Latimer E, Zong JC, Heaggans SY, Richman LK, Hayward GS. Detection and evaluation of novel herpesviruses in routine and pathological samples from Asian and African elephants: identification of two new probosciviruses (EEHV5 and EEHV6) and two new gammaherpesviruses (EGHV3B and EGHV5). Vet Microbiol. 2011;147(1-2):28-41. 5. Richman LK, Zong JC, Latimer EM, et al. Elephant endotheliotropic herpesviruses EEHV1A, EEHV1B, and EEHV2 from cases of hemorrhagic disease are highly diverged from other mammalian herpesviruses and may form a new subfamily. J Virol. 2014;88(23):13523-13546. 6. Zong JC, Latimer EM, Long SY, Richman LK, Heaggans SY, Hayward GS. Comparative genome analysis of four elephant endotheliotropic herpesviruses, EEHV3, EEHV4, EEHV5, and EEHV6, from cases of hemorrhagic disease or viremia. J Virol. 2014;88(23):13547-13569. 7. Zong JC, Heaggans SY, Long SY, et al. Detection of Quiescent Infections with Multiple Elephant Endotheliotropic Herpesviruses (EEHVs), Including EEHV2, EEHV3, EEHV6, and EEHV7, within Lymphoid Lung Nodules or Lung and Spleen Tissue Samples from Five Asymptomatic Adult African Elephants. J Virol. 2015;90(6):3028-3043. 8. Long SY, Latimer EM, Hayward GS. Review of Elephant Endotheliotropic Herpesviruses and Acute Hemorrhagic Disease. ILAR J. 2016;56(3):283-296. 9. Richman LK, Montali RJ, Cambre RC, et al. Clinical and pathological findings of a newly recognized disease of elephants caused by endotheliotropic herpesviruses. J Wildl Dis. 2000;36(1):1-12. 10. Zachariah A, Zong JC, Long SY, et al. Fatal herpesvirus hemorrhagic disease in wild and orphan asian elephants in southern India. J Wildl Dis. 2013;49(2):381- 393. 11. Reid CE, Hildebrandt TB, Marx N, et al. Endotheliotropic elephant herpes virus (EEHV) infection. The first PCR-confirmed fatal case in Asia. Vet Q. 2006;28(2):61-64. 12. Bouchard B, Xaymountry B, Thongtip N, Lertwatcharasarakul P, Wajjwalku W. First reported case of elephant endotheliotropic herpes virus infection in Laos. J Zoo Wildl Med. 2014;45(3):704-707.

105 13. Fickel J, Richman LK, Montali R, et al. A variant of the endotheliotropic herpesvirus in Asian elephants (Elephas maximus) in European zoos. Vet Microbiol. 2001;82(2):103-109. 14. Wilkie GS, Davison AJ, Kerr K, et al. First fatality associated with elephant endotheliotropic herpesvirus 5 in an Asian elephant: pathological findings and complete viral genome sequence. Sci Rep. 2014;4:6299. 15. Fuery A, Browning GR, Tan J, et al. Clinical Infection of Captive Asian Elephants (Elephas Maximus) with Elephant Endotheliotropic Herpesvirus 4. J Zoo Wildl Med. 2016;47(1):311-318. 16. Wilkie GS, Davison AJ, Watson M, et al. Complete genome sequences of elephant endotheliotropic herpesviruses 1A and 1B determined directly from fatal cases. J Virol. 2013;87(12):6700-6712. 17. Ling PD, Reid JG, Qin X, et al. Complete Genome Sequence of Elephant Endotheliotropic Herpesvirus 1A. Genome announcements. 2013;1(2):e0010613. 18. Sripiboon S, Tankaew P, Lungka G, Thitaram C. The occurrence of elephant endotheliotropic herpesvirus in captive Asian elephants (Elephas maximus): first case of EEHV4 in Asia. J Zoo Wildl Med. 2013;44(1):100-104. 19. Ossent P, Guscetti F, Metzler AE, Lang EM, Rubel A, Hauser B. Acute and fatal herpesvirus infection in a young Asian elephant (Elephas maximus). Vet Pathol. 1990;27(2):131-133. 20. Richman LK, Montali RJ, Hayward GS. Review of a newly recognized disease of elephants caused by endotheliotropic herpesviruses. Zoo Biology. 2000;19(5):383- 392. 21. Fickel J, Lieckfeldt D, Richman LK, Streich WJ, Hildebrandt TB, Pitra C. Comparison of glycoprotein B (gB) variants of the elephant endotheliotropic herpesvirus (EEHV) isolated from Asian elephants (Elephas maximus). Vet Microbiol. 2003;91(1):11-21. 22. Zong J-C LE, Heaggans SY, Richman LK, Hayward GS. Viral gene subtyping of eighteen North American cases of EEHV hemorrhagic disease. . Paper presented at: Proceedings of the International Elephant Conservation and Research Symposium; 2008; Pattaya, Thailand. 23. Schmitt DL, Hardy DA, Montali RJ, et al. Use of famciclovir for the treatment of endotheliotrophic herpesvirus infections in Asian elephants (Elephas maximus). J Zoo Wildl Med. 2000;31(4):518-522. 24. Isaza R, Hunter R, Richman L, et al. Famciclovir pharmacokinetics in young Asian elephants (Elephas maximus). Paper presented at: JOINT CONFERENCE- AMERICAN ASSOCIATION OF ZOO VETERINARIANS2003. 25. Richman LK. Pathological and Molecular Aspects of Fatal Endotheliotropic Herpesviruses of Elephants: [Ph.D Thesis]: Ph.D. Thesis2003. 26. Ehlers B, Dural G, Marschall M, Schregel V, Goltz M, Hentschke J. Endotheliotropic elephant herpesvirus, the first betaherpesvirus with a thymidine kinase gene. Journal of general virology. 2006;87(10):2781-2789. 27. Davison AJ, Eberle R, Ehlers B, et al. The order Herpesvirales. Archives of virology. 2009;154(1):171-177.

106 28. Zong J-C LE, Heaggans SY, Richman LK, Hayward GS. Viral Gene Subtyping of Eighteen North American Cases of EEHV Hemorrhagic Disease. Pro Internat Elephant Cons and Res Symp; Nov 2008, 200; Bangkok, Thailand. 29. Stanton JJ, Zong JC, Latimer E, et al. Detection of pathogenic elephant endotheliotropic herpesvirus in routine trunk washes from healthy adult Asian elephants (Elephas maximus) by use of a real-time quantitative polymerase chain reaction assay. American journal of veterinary research. 2010;71(8):925-933. 30. Stanton JJ, Zong J-C, Eng C, et al. Kinetics of viral loads and genotypic analysis of elephant endotheliotropic herpesvirus-1 infection in captive Asian elephants (Elephas maximus). Journal of Zoo and Wildlife Medicine. 2013;44(1):42-54. 31. Schaftenaar W, Reid C, Martina B, Fickel J, Osterhaus AD. Nonfatal clinical presentation of elephant endotheliotropic herpes virus discovered in a group of captive Asian elephants (Elephas maximus). J Zoo Wildl Med. 2010;41(4):626- 632. 32. Hardman K, Dastjerdi A, Gurrala R, et al. Detection of elephant endotheliotropic herpesvirus type 1 in asymptomatic elephants using TaqMan real-time PCR. Vet Rec. 2012;170(8):205. 33. Vidya T, Fernando P, Melnick D, Sukumar R. Population differentiation within and among Asian elephant (Elephas maximus) populations in southern India. Heredity. 2005;94(1):71-80. 34. Hayward GS. Conservation: clarifying the risk from herpesvirus to captive Asian elephants. The Veterinary record. 2012;170(8):202. 35. Zong J-C, Heaggans SY, Long SY, et al. Detection of Quiescent Infections with Multiple Elephant Endotheliotropic Herpesviruses (EEHVs), Including EEHV2, EEHV3, EEHV6, and EEHV7, within Lymphoid Lung Nodules or Lung and Spleen Tissue Samples from Five Asymptomatic Adult African Elephants. Journal of virology. 2016;90(6):3028-3043. 36. Ling PD, Long SY, Zong JC, Heaggans SY, Qin X, Hayward GS. Comparison of the Gene Coding Contents and Other Unusual Features of the GC-Rich and AT- Rich Branch Probosciviruses. mSphere. 2016;1(3). 37. Li H, Durbin R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics. 2009;25(14):1754-1760. 38. Zerbino DR, Birney E. Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome research. 2008;18(5):821-829. 39. Lole KS, Bollinger RC, Paranjape RS, et al. Full-length human immunodeficiency virus type 1 genomes from subtype C-infected seroconverters in India, with evidence of intersubtype recombination. Journal of virology. 1999;73(1):152-160. 40. Honess R, Gompels U, Barrell B, et al. Deviations from expected frequencies of CpG dinucleotides in herpesvirus DNAs may be diagnostic of differences in the states of their latent genomes. Journal of general virology. 1989;70(4):837-855. 41. Krug LT, Inoue N, Pellett PE. Differences in DNA binding specificity among Roseolovirus origin binding proteins. Virology. 2001;288(1):145-153. 42. Chang Y-N, Crawford S, Stall J, Rawlins D, Jeang K, Hayward G. The palindromic series I repeats in the simian cytomegalovirus major immediate-early promoter behave as both strong basal enhancers and cyclic AMP response elements. Journal of virology. 1990;64(1):264-277.

107 43. Chan Y-J, Tseng W-P, Hayward GS. Two distinct upstream regulatory domains containing multicopy cellular transcription factor binding sites provide basal repression and inducible enhancer characteristics to the immediate-early IES (US3) promoter from human cytomegalovirus. Journal of virology. 1996;70(8):5312-5328. 44. Chang Y-N, Jeang K, Chiou C-J, Chan Y, Pizzorno M, Hayward G. Identification of a large bent DNA domain and binding sites for serum response factor adjacent to the NFI repeat cluster and enhancer region in the major IE94 promoter from simian cytomegalovirus. Journal of virology. 1993;67(1):516-529. 45. Jeang K, Rawlins D, Rosenfeld PJ, Shero J, Kelly T, Hayward G. Multiple tandemly repeated binding sites for cellular nuclear factor 1 that surround the major immediate-early promoters of simian and human cytomegalovirus. Journal of virology. 1987;61(5):1559-1570. 46. Pellett PE. Trunkloads of viruses. J Virol. 2014;88(23):13520-13522. 47. Brocchieri L, Kledal TN, Karlin S, Mocarski ES. Predicting coding potential from genome sequence: application to betaherpesviruses infecting rats and mice. Journal of virology. 2005;79(12):7570-7596. 48. Geyer H, Ettinger J, Möller L, et al. Rat cytomegalovirus (RCMV) English isolate and a newly identified Berlin isolate share similarities with but are separate as an anciently diverged clade from Mouse CMV and the Maastricht isolate of RCMV. Journal of General Virology. 2015;96(7):1873-1882. 49. Dolan A, Cunningham C, Hector RD, et al. Genetic content of wild-type human cytomegalovirus. Journal of General Virology. 2004;85(5):1301-1312. 50. Cox HR. Growth of viruses and rickettsiae in the developing chick embryo. Annals of the New York Academy of Sciences. 1952;55(2):236-247. 51. Nahmias AJ, Dowdle WR, Naib ZM, Highsmith A, Harwell RW, Josey WE. Relation of pock size on chorioallantoic membrane to antigenic type of herpesvirus hominis. Proceedings of the Society for Experimental Biology and Medicine Society for Experimental Biology and Medicine (New York, NY). 1968;127(4):1022-1028. 52. Akter T, Tabassum S, Jahan M, Nessa A, Islam MN, Giasuddin M. Isolation of herpes simplex viruses by chick embryo culture. Mymensingh medical journal : MMJ. 2013;22(2):365-369. 53. Ciufo DM, Cannon JS, Poole LJ, et al. Spindle cell conversion by Kaposi's sarcoma-associated herpesvirus: formation of colonies and plaques with mixed lytic and latent gene expression in infected primary dermal microvascular endothelial cell cultures. J Virol. 2001;75(12):5614-5626. 54. Cheung AL. Isolation and culture of human umbilical vein endothelial cells (HUVEC). Current protocols in microbiology. 2007;Appendix 4:Appendix 4B.

108 Simon Y. Long 108 South Wolfe Street, Baltimore, MD 21231 Email address: [email protected] Mobile phone: (610) 299-6637

PERSONAL STATEMENT: Research veterinary pathologist seeking a position with an emphasis on molecular and pathology techniques utilizing animal models and molecular biology to develop therapeutic approaches for patients with unmet medical needs.

PROFILE  Strong research and laboratory animal pathology background that includes phenotyping/comparative pathology study experience. Research experience includes academic and pharmaceutical settings.  Extensive research experience including molecular virology investigating disease pathogenesis, molecular cellular pathways in cell cycling and toxicological evaluations in both academic and pharmaceutical settings.  Strong laboratory animal pathology and science background in laboratory animal species including rodents, rabbits, canines, and nonhuman primates in both academic and pharmaceutical settings for both basic research and safety assessment including GLP standards.

EDUCATION JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE Baltimore, MD PhD Candidate- Cellular and Molecular Medicine PhD Program Sept. 2010- June 2016

JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE Baltimore, MD Pathology Postdoctoral Fellow-Pathology Residency July 2009- June 2013

UNIVERSITY OF PENNSYLVANIA, SCHOOL OF VETERINARY MEDICINE Philadelphia, PA Veterinary Medicine Doctorate (VMD)-MIXED ANIMAL MAJOR Graduated May 2009

DREXEL UNIVERSITY, BIOSCIENCE AND BIOTECHNOLOGY Philadelphia, PA Graduate Student Sept. 2003-June 2004

DREXEL UNIVERSITY COLLEGE OF MEDICINE MCP HAHNEMANN Philadelphia, PA Master in Laboratory Animal Science (MLAS) Graduated May 2003

SUNY COLLEGE OF ENVIRONMENTAL SCIENCE AND FORESTRY Syracuse, NY Bachelor of Science Environmental Forest Biology/Pre-veterinary Medicine Graduated May 2001

PROFESSIONAL CERTIFICATIONS NATIONAL VETERINARY BOARD EXAMINATION: passed December 2008 Pennsylvania State Veterinary Board Licensed: 2009-present

Eligible for the board certification examination for the American College of Veterinary Pathologists with passing Phase I General Pathology in March 2015 and continuation with Phase II.

109 RESEARCH AND PROFESSIONAL EXPERIENCE Johns Hopkins University: Dr. Gary S. Hayward’s Herpesvirus Laboratory PhD Graduate Student January 2011-Present  Thesis project on dissecting the molecular pathogenesis of the Elephant Endotheliotropic Herpesvirus (EEHV) in Asian elephant (Elephas maximus).  Developed a primary and immortalized elephant umbilical vascular endothelial cell lines for EEHV viral propagation.  Genetic analysis/comparison between different GC-rich and AT-rich EEHV viruses obtained from clinical and necropsy samples.  Development of a next generation proteomic chip serology assay using cloned GST fusion EEHV viral proteins.  Detecting viral genomic copies using PCR to determine viral levels and virus subtypes in necropsy tissues blood samples, and biopsy samples.  Familiar with mammalian cell culture, PCR, western blot, proteinomic cloning and serological cHip assay development.

Johns Hopkins University: Department of Molecular and Comparative Pathobiology Postdoctoral Pathology Fellow July 2009- June 2013  Performed necropsies and histopathological evaluation on small animals, lab animals and wildlife species.  Performed histopathological evaluation on biopsy specimens.  Performed mouse phenotyping studies for principle investigators under the mentorship of Dr. Cory Brayton, DVM, Dipl. ACLAM, Dipl. ACVP.  Trained and supervised veterinary students in summer veterinary pathology externships.  Attended weekly slide conference and board training sessions.  Performed human autopsy rotation with completion of 6 human autopsy cases from June to July 2010.  Passed American College of Veterinary Pathologist Board Certification Phase I General Pathology section in March 2015.

Children’s Hospital of Philadelphia: Dr. Jake Kushner’s Diabetes Cell Cycling Laboratory Senior Research Technician II/Laboratory Manager March 2004-March 2008  Hired to start up a diabetes cell cycling laboratory and managed a molecular laboratory, animal colony and 2-3 junior research technicians.  Involved in leading two major research projects through animals experiments, molecular, biochemistry techniques plus guiding and assisting other research projects.  Responsibilities included genotyping, characterize mice for diabetic phenotypes, cell culture, gene expression analysis, immunohistochemistical slide development, fluorescent microscopy, and islet morphometrics.

NIH/Merck Summer Research Program: Dr. Susan Weiss’s Coronavirus Laboratory Summer Research Veterinary Student Summer of 2006  Researched the role of the N protein in coronaviruses in pathogenesis through using PCR to design and create chimeric viruses using constructs and cell cultures and observing their affects in mice.

Bristol-Myers Squibb: Drug Metabolism and Pharmacokinetics Biological Technician August 11- October 2003  Biological technician in the Technical Support Unit of Drug Metabolism and Pharmacokinetics Department, Lawrenceville New Jersey.  Assisted and lead pharmacokinetic studies on mice, rats, dogs and primates. Assisted in cannulation surgery on rodents.

110 Neuronyx, Inc. Assistant Researcher December- May 2003  Assisted in a GLP FDA rodent model of spinal cord injury recovery facilitated by transplant of adult bone marrow stem cells as a potential therapeutic.  Responsibilities included basic husbandry, quality assurance, veterinary care and surgical preparation.

Medical College of Pennsylvania Hahnemann University: Dr. Libby Blankerhorn’s Laboratory Assistant Molecular Biologist December - April 2002  Assisted graduate students in daily gel electrophoresis and PCR techniques of genetic materials of colony of inbred rats for insulin study.  Responsibilities include gel electrophoresis, PCR, proper usage of radioactive material and rodent handling.

SUNY College of Environmental Science and Forestry: Dr. Chun June Wang’s Laboratory Assistant Molecular Biologist and Mycologist September- May 2001  Involved in isolating and genetically sequencing a new species of Exophila sp. fungi pathogenic to lumpfish.  Responsibilities included mycological techniques (aseptic techniques), DNA extraction, sequencing and PCR.

SUNY College of Environment Science and Forestry: Dr. Scott O. Rogers’ Laboratory Assistant Molecular/Mycological Biologist September 1997-May1999  Assisted Dr. Scott O. Rogers and Dr. Chun June Wang with numerous molecular projects.  Responsibilities included DNA extraction, PCR, gel electrophoresis, maintaining fungi colonies with aseptic mycological techniques, micro-cross-section techniques, slide preparation processes and basic photo-development processes.

SUNY College of Environment Science and Forestry: Dr. Larry Smart’s Laboratory Assistant Molecular Biologist September 1998-May1999  Involved in a two-year project with Dr. James Gibbs and Eriko Motegi in College of Environmental Science and Forestry as a laboratory assistant in the “UROP Project: The Genetic Diversity of Sugar Maples”.  Responsibilities included DNA extraction, PCR and gel electrophoresis.

EXTERNSHIP Wildlife Conservation Society, Bronx Zoo: Department of Pathology Student Externship January 2008  Performed a 4 week externship at the Department of Pathology.  Responsibilities included performing daily necropsies, writing necropsy reports, attending weekly pathology rounds, weekly Armed Forces Institute of Pathology (AFIP) slide conferences and presentation of zoonotic diseases of non-human primates.

Merck & Co., Inc. West Point, PA: Laboratory Animal Resource Department Student Externship February 2008  Performed a 4 week externship at the Laboratory Animal Resource Department.  Daily rotations through the different groups and services of LAR providing veterinary services to investigators and working on rodents, canines, guinea pigs, and non-human primates models.

Merck & Co., Inc. West Point, PA: Safety Assessment-Pathology Merck/Merial Veterinary Summer Externship Summer of 2007  Performed a 10 week pathology internship in the Merck West Point Safety Assessment Department.

111  Training included weekly AFIP slide conferences, weekly peer-review case studies, necropsy training, tissue trimming, histo-technology, attending pathology rounds, presentation of summer research project on hyperplasia in rat pituitaries.

MERCK & Co., Inc. Rahway, NJ: Laboratory Animal Resource Department Co-op Intern January 13- May 2, 2003  Intern in the comparative medicine department.  Trained in all the basic functions of running and maintaining an animal facility.  Training includes management skills, operations of both barrier and conventional animal facilities, clinical care of laboratory animals, pharmacokinetic and transgenic studies.

TEACHING EXPERIENCE Johns Hopkins University School of Medicine: Mouse Pathobiology & Phenotyping Short Course Necropsy Instructor 2010, 2011, and 2015  Introduced and guided participants in performing a full mouse necropsy for phenotyping analysis.  Introduced and guided participants in trimming mouse tissues for phenotyping analysis.

Johns Hopkins University School of Medicine: Toxicology Pathology Endocrinology Lecturer April 2013  Lectured on the toxicological principals centered on the endocrine system.

Drexel University: Developmental Biology Laboratory Teaching Assistant March- June 2004  Taught college students the basic concepts of applied biology under Dr. Jeremy Lee.  Responsibilities included teaching and managing undergraduate level laboratory course.

Drexel University: Applied Biology Teaching Assistant January-March 2004  Taught undergraduate students the basic concepts of applied biology under Dr. Jennifer Quinlan.  Responsibilities included teaching and managing undergraduate level laboratory course.

Drexel university: Organismal Physiology Teaching Assistant Sept. -December 2003  Taught undergraduate students the basic concepts of organismal physiology under Dr. Jeremy Lee.  Responsibilities included teaching and managing undergraduate level laboratory course.

SUNY College of Environmental Science and Forestry Teaching Assistant January 1999- May 2000  Taught undergraduate students the basic concepts of herpetology under the guidance of Dr. James Gibbs.  Responsibilities included teaching and managing undergraduate level college laboratory course.  Teaching included lecturing, managing laboratories, setting up a laboratory, and lead field trips.

COMMITTEE Johns Hopkins School of Medicine: Molecular and Comparative Pathobiology Department Admissions Committee Member 2009-2014  Served as an admissions committee member for screening of candidates applying for residency and postdoctoral fellow positions for both the anatomic pathology and laboratory animal positions.

112 Drexel College of Medicine: Institutional Animal Care and Use Committee Alternative IACUC Member January- December 2002  Served as an alternative IACUC member to review animal protocols for IACUC board approval for animal studies at Drexel University, Drexel College of Medicine Hahnemann University.

PUBLICATIONS Ling PD, Long SY, Zong JC, Heaggans SY, Qin X, Hayward GS. Comparison of the gene coding context and other unusual features of the the GC-rich and AT-rich branch probosciviruses. mSphere. 2016; 1(3): e00091-16. doi:10.1128/mSphere.00091-16.

Ling PD, Long SY, Fuery A, Peng RS, Heaggans SY, Qin X, Worley KC, Dugan S, Hayward GS. Complete genome sequence of elephant enodtheliotropic herpesvirus 4 (EEHV4) the first example of a GC- rich branch proboscivirus. mSphere. 2016; 1(3): e00081-15. doi:10.1128/mSphere.00081-15.

Long SY. Approach to reptile emergency medicine. Vet Clin North Am Exot Anim Pract. 19(2): 567-90. May 2016. PMID:27131162

Fuery A, Browning GR, Tan J, Long SY, Hayward GS, Cox SK, Flanagan JP, Tocidlowski ME, Howard LL, Ling PD. Clinical infections of captive asian elephant (Elephas mamimus) with elephant endotheliotropic herpesvirus 4. Journal of zoo and wildlife medicine. 47(1): 311-318. March 2016. PMID: 27010293

Long SY, Latimer EM, Hayward GS. Review of elephant endotheliotropic herpesviruses and acute hemorrhagic disease. Institue for Laboratory Animal Research Journal 56(3): 283-96. Feburary 2016. PMID: 26912715

Zong JC, Heaggans SY, Long SY, Latimer EM, Nofs SA, Bronson E, Casares M, Fouraker MD, Pearson VR, Richman LK, Hayward GS. Detection of quiescent infections with multiple elephant endotheliotropic herpesviruses EEHV2, EEHV3, EEHV6 and EEHV7 within lymphoid lung nodules or lung and spleen tissue samples from five asymptomatic adult African elephants. Journal of Virology 90(6): 3028-43. December 2015. PMID: 26719245

Zong JC, Latimer EM, Long SY, Richman LK, Heaggans SY, Hayward GS. Comparative genome anaylysis of four elephant endotheliotropic herpesvirus, EEHV3, EEHV4, EEHV5, and EEHV6, from cases of hemorrhagic disease or viremia. Journal of Virology 88(23):13547-69. December 2014. PMID: 25231309

Sarotir DJ, Wilbur CJ, Long SY, Rankin MM, Li C, Bradfield JP, Hakonarson H, Grant SF, PU WT, Kushner JA. GATA factors promte ER integrity and beta-cell survivial and contribute to type 1 diabetes risk. Molecular Endocrinology 28(1):28-39. January 2014. PMID: 24284823

Zachariah A, Zong JC, Long SY, Latimer EM, Heaggans SY, Richman LK, Hayward GS. (2013). Fatal Herpesvirus (EEHV) Hemorrhagic Disease in Wild and Orphan Asian Elephants in India. J. Wildl. Dis. 49, 381-393. April 2013 PMID: 23568914

Tuttle AH, Rankin MM, Teta M, Sartori DJ, Stein GM, Kim GJ, Virgillio C, Granger A, Zhou D, Long SY, Schiffman AB, Kushner JA. Immunofluorescent detection of two thymidine analogues (CldU and IdU) in primary tissue. Journal of Visualized Experiments (46). December 2010 PMID: 21178965

Cowley TJ, Long SY, Weiss SR: The murine coronavirus nucleocapsid gene is a determinant of virulence. Journal of Virology 84(4):1752-63. February 2010 PMID: 20007284

113 He LM, Sartori DJ, Teta M, Opare-Addo LM, Rankin MM, Long SY, Diehl JA, Kushner JA: Cyclin D2 protein stability is regulated in pancreatic beta-cells. Molecular Endocrinology 23(11): 1865-1875. November 2009 PMID: 19628581

Teta M, Rankin MM, Long SY, Stein GM, Kushner JA: Growth and regeneration of adult beta cells does not involve specialized progenitors. Developmental Cell 12(5): 817-826, May 2007 PMID: 17488631

Teta M, Long SY, Wartschow LM, Rankin MM, Kushner JA: Very slow turnover of beta cells in aged adult mice. Diabetes 54:2557-2567, September 2005 PMID: 16123343

Kushner JA, Ciemerych MA, Sicinska E, Wartschow LM, Teta M, Long SY, Sicinski P, White MF: Cyclins D2 and D1 are essential for postnatal pancreatic beta-cell growth. Molecular and Cellular Biology 25(9436): 3752-3762, May 2005 PMID: 15831479

PRESENTATIONS Long SY, Zachariah A, Zong JC, Latimer EM, Heaggans SY, and Hayward GS. High level genetic variability amongst nine cases of fatal elephant endotheliotropic herpesvirus hemorrhagic disease in wild and orphan Asian elephants in southern India. Oral presentation at the 9th Annual International Elephant Endotheliotropic Herpes virus (EEHV) workshop. January 2013

Long SY, Ap Rhys CJ, Heaggans SY, Hayward GS. Immortalization of elephant umbilical vascular endothelial cells (EUVEC) for elephant endotheliotropic herpesviruses (EEHV) infection. Poster presentation at the American College of Veterinary Pathologist Annual Meeting. December 2011.

Long SY, Hadfield CA, Clayton LA, Huso DL, Montali RJ. Disseminated chromoblastomycosis in White’s Tree frog (Litoria caerulea). Poster presentation at the American College of Veterinary Pathologist Annual Meeting. December 2010.

Long SY, Li C, Matschinsky FM, Stanley CA, Kushner JA. GATA-4 is essential for normal beta cell function. Oral presentation at the Mid-Atlantic Diabetes Research Session at National Institute for Health. October 1, 2005

Long SY, Cowley TJ, Weiss SR. Understanding the role of the nucleocapsid gene in the mouse hepatitis virus using chimeric viruses. Poster presentation at the 7th Annual Merck-Merial/NIH Veterinary Scholars Symposium at Louisiana State University. August 3-6, 2006.

Long SY, Cowley TJ, Weiss SR. Understanding the role of the nucleocapsid gene in the mouse hepatitis virus using chimeric viruses. Poster presentation at the Penn Vet Student Research Day. University of Pennsylvania School of Veterinary Medicine March 22, 2007.

EXTRAMURAL FUNDING Morris Animal Foundation: Wildlife Studies Fellowship D14ZO-411 (S.LONG) Fellowship Award June 2014-June 2016  Wildlife fellowship salary funding support for 2 years for “Innovative attempts to propagate elephant endotheliotropic herpesvirus in cell culture”.

National Institution of Health: 2T32RR007002-35 (C.ZINK) Postdoctoral training June 2009-June 2013  Training Veterinarians for Careers in Biomedical research  PI: Christine Zink, DVM, PhD, Dipl. ACVP  Mentor: Cory F. Brayton, DVM, Dipl. ACVP, Dipl. ACLAM  Role: Veterinary Pathology Postdoctoral fellow

114 AWARDS Delaware Valley Branch of American Association of Laboratory Animal Science Scholarship Award 2002  Recipient of the 2002 J.J Noonan Scholarship award for laboratory animal science.

PROFESSIONAL ASSOCIATIONS  Member of the Wildlife Disease Association  Member of Society of Toxicologic Pathology  Member of Society of Study of Amphibians and Reptiles

115