Genetic Variability and Its Relationship to Acanthamoeba

Home , Acanthamoeba, Acanthamoeba infection, Legionella cherrii

Genetic Variability and its Relationship to

Acanthamoeba Pathogenesis

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Monica Jane Crary

**********

Graduate Program in Molecular Genetics

The Ohio State University 2012

Dissertation Committee:

Dr. Amanda Bird

Dr. Gregory C. Booton, Advisor

Dr. Susan Cole

Dr. David Denlinger

Dr. Paul A. Fuerst, Advisor

Monica Jane Crary

2012

ABSTRACT

Acanthamoeba is a pathogenic protist that causes a sight-threatening eye infection,

Acanthamoeba keratitis (AK). The central focus of this research is to analyze the genetic relationships between Acanthamoeba isolates and how that contributes to pathogenesis.

Acanthamoeba is a diverse genus with more than 18 genotypes based on the 18S ribosomal RNA gene. DNA barcoding has suggested that the mitochondrial cytochrome oxidase c subunit 1 (COI) gene can be accurately used to identify a eukaryotic organism.

As part of understanding more about the phylogenetic structure of the genus of

Acanthamoeba, representative isolates of most of the genotypes were analyzed using their

COI genes. The phylogenetic relationships based on the COI gene were nearly identical to those produced using the 18S rRNA gene. This allows COI to be used as a reasonable substitute to the 18S rRNA gene. This project found an alternative method of classifying

Acanthamoeba and allowed Acanthamoeba phylogenetics to be examined at more than one locus.

A majority of this research focused on the ongoing Acanthamoeba keratitis outbreak in Chicago, Illinois since 2002. The dramatic increase of AK cases has occurred following EPA mandated water treatment changes which have been hypothesized to have increased microorganisms in the water system. To understand the phylogenetic structure of Acanthamoeba within this outbreak, a multilocus sequence typing (MLST) protocol ii was created. This project involved identifying and developing sequencing methods for five highly conserved housekeeping genes in Acanthamoeba. Sequences from these genes were used to determine the degree of variability amongst Acanthamoeba as well as to further our understanding of sub-lineages within the most common genotype, T4. In addition to the Chicago AK review, three surveys were conducted to elucidate the diversity of environmental Acanthamoeba from Chicago, Sonora, Mexico and on an Ohio farm. These surveys combined with the MLST study demonstrate how diverse the genus of Acanthamoeba is, regardless of source.

Acanthamoeba are also known to contain potentially pathogenic endosymbionts including Legionella and Pseudomonas. The presence of endosymbionts has been shown to increase the pathogenicity of Acanthamoeba. To determine if endosymbionts in

Acanthamoeba could contribute to the Chicago outbreak, Acanthamoeba was analyzed for the presence of Legionella and Pseudomonas. More than 50% of Acanthamoeba from

Chicago contained one or both bacterial genera and these bacteria were located intracellularly in Acanthamoeba. We hypothesize that increases in microorganisms in the

Chicago water systems increased the opportunity for Acanthamoeba to associate with pathogenic bacteria. Future research will examine the mechanism that these bacteria use to increase Acanthamoeba’s virulence as well as monitoring if diseases associated with these bacteria are also increasing. Together, these projects are meant to further our understanding on the diversity of the genus Acanthamoeba and the underlying factors of

Acanthamoeba’s virulence.

iii

This dissertation is dedicated to my parents, Roger and Leota.

Acknowledgements

I would like to thank both of my advisors, Dr. Greg Booton and Dr. Paul Fuerst, for their guidance and patience. Thank you to my dissertation committee and fellow lab members,

Daryl and Mike for their technical assistance. To my parents, thank you for your unwavering encouragement and confidence. To Nil, thank you for showing me all the

How's and Why's in the lab. I would like to also thank my friends, Ashley, Kaylan,

Mindy and Bibi for their ears. Finally, thanks to my family, friends and fellow graduate students for their support.

Vita

2003 Madison Central School

2007 B.S. Biology, Ursinus College

2007 to present Graduate Teaching Associate, Department of Molecular Genetics,

The Ohio State University

Publications

Crary, M.; Narayanan, S.; Hopkins, A.; Booton, G.; and Fuerst, P. Phylogenetic analysis and Taxon Barcoding of the Opportunistically Pathogenic Protistan genera Acanthamoeba and Balamuthia. In preparation.

Crary, MJ.; Lares-Villa, F.; Booton, GC.; Pearlman, E.; Shoff, M.; Joslin, C.; Tu E.; Fuerst, PA. Acanthamoeba keratitis outbreak in Chicago, Illinois are Associated with the Presence of the Pathogenic Bacteria Legionella pneumophila and Pseudomonas aeruginosa. In preparation.

Shoff, M. Crary, MJ.; Joslin, C.; Tu, E.; Booton, GC.; Fuerst, PA. Analysis of Acanthamoeba isolates from the Greater Chicago Area Tap Water. In preparation

Cherukuri, NC.; Zhu, Y.; Wolf, JN.; Wu, Z.; Crary, M.; Buckley, K.; Bisaro, K.; and Parris, DS.; 2012 Characterization of the RNA Silencing Suppression Activity of the Ebola Virus VP35 Protein in Plants and Mammalian Cells. Journal of Virology. 86:(6)30-38.

Visvesvara, G.; Shoff, M.; Sriram, R.; Booton, G.; Crary, M.; Fuerst, P.; Hanley, C.; and Garner, M. 2010 Isolation, morphologic, serologic and molecular identification of Acanthamoeba T4 genotype from the liver of a Temminck's tragopan (Tragopan temminckii). Veterinary Parsitology. 170(3)194-200.

Fields of Study

Major Field: Molecular Genetics vi

Table of Contents

Page Abstract...... ii

Dedication...... iv

Acknowledgements...... v

Vita...... vi

List of Tables...... ix

List of Figures...... x

Chapters:

1: Introduction...... 1

2: Cytochrome Oxidase I Taxon Barcoding and Phylogenetic Analysis of the Pathogenic Protistan Genera Acanthamoeba, Protacanthamoeba, and Balamuthia...... 22 Methods...... 27 Results...... 29 Discussion...... 40

3: Environmental Acanthamoeba Isolates Exhibit High Levels of Genetic Diversity and Indicate Possible Means of Acanthamoeba-Human Interaction………...... 44 Methods...... 47 Results...... 49 Discussion...... 65

4: Acanthamoeba in Chicago Illinois are Associated with the Presence of the Pathogenic Bacteria Legionella pneumophila and Pseudomonas aeruginosa...... 71 Methods………………...... 72 vii

Results...... 79 Discussion...... 87

5: Multilocus Sequence Typing of Acanthamoeba Keratitis-Associated Clinical Isolates from a Chicago Outbreak...... 93 Methods...... 96 Results...... 100 Discussion...... 123

6: Discussion...... 126

Bibliography...... 134

Appendix...... 144

viii

List of Tables

Table 1. Acanthamoeba species names with associated genotypes based on the 18S rDNA gene...... 12 Table 2. Protists and their source used in the DNA barcode study...... 28 Table 3. Pairwise distances between Acanthamoeba genotypes at the COI locus...... 33 Table 4. Keratitis-causing and tap water Acanthamoeba isolates examined in the Chicago Tap Water survey...... 56 Table 5. Matched Acanthamoeba isolates from keratitis patients and tap water Acanthamoeba isolates from patient’s residence...... 60 Table 6. Dimensions and genotypes of various Mexican Acanthamoeba isolates...... 63 Table 7. Chicago Acanthamoeba isolates from tap water and keratitis cases used in the endosymbiont study ...... 75 Table 8. Acanthamoeba samples from Non-Chicago keratitis cases used in the endosymbiont study...... 78 Table 9.Presence of intracellular bacteria in regrown Acanthamoeba AK samples from Chicago that were positive for Legionella, Pseudomonas or both………………...... 86 Table 10. Acanthamoeba isolates used in the MLST study...... 99 Table 11. Variability of 5 loci used in MLST analysis of T4 genotype and primer function across the Acanthamoeba genus...... 105 Table 12. Pairwise distance comparisons between Acanthamoeba genotypes across the seven loci...... 122 Appendix Table 1. All primers used in this research with names, primer sequence and annealing temperatures...... 152

List of Figures

Figure 1. Representative trophozoites and cysts of the genus Acanthamoeba...... 3 Figure 2. Phylogenetic relationship of the 18 genotypes of the genus Acanthamoeba as determined by the 18S rDNA gene...... 15 Figure 3. Source of Acanthamoeba and relative frequency of genotypes based on the literature and OSU archived samples...... 17 Figure 4. Phylogenetic relationship between free living amoeba using the full length 18S rRNA gene...... 24 Figure 5. Phylogenetic relationship between free living amoeba using the full length COI gene...... 31 Figure 6. Phylogenetic tree of all the isolates of the genera Acanthamoeba, Balamuthia and Protacanthamoeba using the COI mitochondrial gene...... 32 Figure 7. Comparison of the phylogenetic trees of the 18S and COI genes of the genera Acanthamoeba using major genotypes...... 35 Figure 8.Phylogenetic relationship of the T4 sub-lineages using the 18S rRNA gene………………………………………………………………...... 36 Figure 9. Phylogenetic analysis of the three free living amoeba using the minibarcode amplimer...... 39 Figure 10. Phylogenetic relationship between the Ohio Farm Acanthamoeba isolates....51 Figure 11. Protist and Acanthamoeba presence in tap water by Chicago zip code and sampled households...... 53 Figure 12. Zip codes in the Greater Chicago Area positive for Acanthamoeba, Free- living amoeba or neither...... 54 Figure 13. Phylogenetic relationships between Acanthamoeba keratitis isolates from the Greater Chicago Area...... 58 Figure 14. A. Phylogenetic relationships between Acanthamoeba tap water isolates from the Greater Chicago Area...... 59 Figure 15. Phylogenetic relationship between the Ohio Farm Acanthamoeba isolates....61 Figure 16. Phylogenetic relationships of Acanthamoeba isolated from Mexico...... 64 Figure 17. Relative frequency of the Acanthamoeba from Chicago keratitis and tap water isolates which were positive for Legionella, Pseudomonas or both...... 80 Figure 18. Relative frequency of the Acanthamoeba from Chicago and Non-Chicago sources which were positive for Legionella, Pseudomonas or both...... 81 Figure 19. Prevalence of Legionella and Pseudomonas by species isolated from all Acanthamoeba isolates tested...... 83

Figure 20. A. In situ hybridization of Acanthamoeba with Pseudomonas specific probe. B. In situ hybridization of Acanthamoeba with Legionella specific probe...... 85 Figure 21. Gene structure of the five loci chosen for MLST analysis...... 102 Figure 22.. Neighbor-joining tree constructed using exon sequence from Beta Tubulin data across four genotypes...... 104 Figure 23. Neighbor-joining tree constructed using exon sequence from Elongation Factor 1 data across seven genotypes...... ………107 Figure 24. Neighbor-joining tree constructed using exon sequence from G3PD data across six genotypes…………………………………………………………………….109 Figure 25. Neighbor-joining tree constructed using exon sequence from Glycogen Phosphorlyase data across seven genotypes...... …...... 111 Figure 26. Neighbor-joining tree constructed using exon sequence from RasC data across three genotypes...... 113 Figure 27. Total data phylogenetic tree across the five housekeeping gene loci...... 116 Figure 28. Total data phylogenetic tree across the seven gene loci...... 118 Figure 29. BURST cluster analysis of 33 sequence types using allelic profiles created by five gene loci...... 119 Figure 30. Polymorphic sites of the five loci used in the MLST scheme...... 120

CHAPTER 1

INTRODUCTION

Acanthamoeba and its role in the environment

Acanthamoeba is an aerobic, mitochondriate, eukaryotic protist that is ubiquitous in the environment. It was given its name because of the spine-like structures on its surface, "Acantha" is the Greek word for spiky, spiny, thorny, etc. This genus is classified in Super Group Amoebozoa and traditionally classified in the class Lobosea,

(Bovee 1985; Page 1987) because of the amoeba's lobose psuedopods (Siddiqui & Khan

2012). More recently, Cavalier-Smith has placed Acanthamoeba and its sister genus

Balamuthia within the order Centramoebida, although the exact placement of this order within the Amoebozoa is unstable and depends on the molecular tool used for placement

(Cavalier-Smith et al,. 2004).

Originally observed in the 1600s by Antonio van Leeuwenhoek, it was not until the late 1900s that Acanthamoeba was conclusively linked to diseases in humans including Acanthamoeba keratitis (AK) and Granulomatous Amebic Encephalitis (GAE)

(Jager & Stamm et al., 1972; Nagington et al., 1974). Acanthamoeba were originally classified under genus Hartmanella and it was not until the 1980s that Hartmanella was

1 split into two genera, Hartmanella and Acanthamoeba (Page 1987). The confusion between different free-living amoeba was a result of both organisms having similar life cycles and plastic morphology. A bacterivore, Acanthamoeba contributes to the microbial loop by releasing bacterial biomass back into the environment. In the rhizophere, it is thought to play a role in optimizing the bacterial community composition through grazing (Kreuzer et al., 2006). This optimization improves root structure of various plants. Additionally Acanthamoeba increases phosphorus uptake in plants as well as liberates nitrogen, which had been previously sequestered in its various prey.

Acanthamoeba is found in environments beyond just soil. It has been found in air, sand, fresh and salt water, glaciers and sandstorm dust. In manmade environments, it has been isolated from swimming pools, tap water, bottled water, showerheads, faucets, and domestic water systems. Rarely, Acanthamoeba is an opportunistic pathogen of more complex organisms, including birds, fish, dogs, reptiles, and of course, humans

(Frank & Bosch et al., 1972; Dykova et al., 1999; Visvesvara et al., 2007; Kent et al.,

2011). Acanthamoeba is able to inhabit all these environments and organisms because of its ability to reversibly differentiate.

Acanthamoeba life cycle

Acanthamoeba has a biphasic life cycle: mobile trophozoites and dormant cysts

(Figure 1). As trophozoites, they are actively feeding by pinocytosis and phagocytosis and prefer to live at interfaces such as soil-water, water-air, etc (DeMores & Alfieri et al.,

2008). Under optimal growth conditions, trophozoites will grow and replicate by binary

A B C

D E F G H

Figure 1. Representative trophozoites (A,B, and C) and cysts (D-H) of the genus Acanthamoeba. A & D : 07-104, Acanthamoeba keratitis, T4. B: 03-001, Acanthamoeba castellanii Neff, T4. C, F, G & H: 05-022, environmental isolate, T17. E: 07-103, Acanthamoeba keratitis, T4.

3 fission. Acanthamoeba has no documented sexual reproduction. Acanthamoeba attaches to its prey through the Mannose-Binding protein (MBP), which binds to mannosylated glycoproteins on the surface of bacteria (Garate et al., 2004). MBP is thought to be responsible for all of Acanthamoeba’s receptor mediated endocytosis regardless of its prey. However, the unique features of Acanthamoeba are most noted in its alternate form, the cyst.

Encystment

Under stress conditions, the amoeba encases itself in a double layered cyst from which it will not emerge until conditions are again optimal for reproduction. Encystment is induced by desiccation, starvation, anoxia, osmolarity, extreme pH, temperature, or other environmental stresses (Lloyd et al., 2001; Aksozek et al., 2002). There are three stages of encystment: pre-encystment, cyst initiation and cyst wall synthesis (Khunkitti et al., 1998). During early stages of encystment, the trophozoites round up and withdraw their pseudopodia (Krishna & Shukla, 1984). Trophozoites store glucose in the form of glycogen, which is used to synthesize cellulose (Lorenzo et al., 2008). Glycogen is the most rapidly-degraded macromolecule during initial phase of Acanthamoeba encystment

(Lorenzo et al., 2008). This glucose is then used for the construction of the cyst walls.

There is an 80% decrease in cell volume and autolyosomes appear to remove mitochondria, excess food and other materials (Tomlinson, 1967). During encystment, there is a loss of DNA, RNA and protein, though encystment is repressed if protein and

RNA synthesis are inhibited (Krishna & Shukla et al., 1984). The outer ectocyst,

4 containing an acid-insoluble protein, is synthesized first and is continuous over the entire cell. The endocyst, made of cellulose, is created after the exocyst and becomes attached to the ectocyst at ostioles, which are closed by an operculum plug (Mattar & Byer et al.,

1971). The mechanism of encystment is not entirely understood but recent studies have found that encystment is dependent on genes responsible for glycogen breakdown, cellulose synthesis, autophagy proteins and cysteine proteases (Chambers & Thompson,

1974; Byers et al., 1991; Lorenzo et al., 2008; Bouyer et al., 2009). Very little research has been done on the cyst once fully formed, so little is known about how it is able to remain dormant for years or at what level it is still functioning metabolically.

There appears to be an unknown maturation process that occurs over time as newly made cysts are not capable of immediately excysting (Chambers & Thompson,

1974). Several studies have found that the age of the cyst corresponds with their ability to excyst which suggests that further mechanisms occur following visible encystment that lend to complete cyst maturation. Cysts can synchronously excyst when kept in cyst form for more than 800 hours (Chambers & Thompson, 1974). When kept in cyst form for less than 800 hours, not all cysts are capable of excystment. As a cyst, Acanthamoeba can remain dormant for more than twenty years and retain its ability to excyst and reproduce (Sriram et al., 2008). Cysts are capable of surviving low pHs, UV and gamma radiation, antibiotics and heavy metals (Aksozek et al., 2002). Cysts play a larger role in mammalian infections because they are largely invisible to the immune system.

Excystment

Very little is understood about what triggers excystment. While the presence of prey can stimulate excystment, Acanthamoeba will not always excyst even under apparent optimal conditions. The process of excystment has been divided into four stages: mature cyst, activated cyst, pre-emergence stage, and emergence (Mattar & Byers,

1971). Excystation is dependent on protein and RNA synthesis (Chambers & Thompson,

1974; Krishna & Shukla, 1984). However, only very low levels of RNA synthesis occurs during excystment, suggesting that the cyst is largely preloaded with the necessary RNA needed during excystment (Krishna & Shukla, 1984). The first clear sign of excystment is the movement of the trophozoite within the cyst walls. From there, the trophozoite emerges from an ostiole from which the operculum has been removed. Beyond the operculum, no other parts of the two walls are visually degraded. However, the endo- and ectocyst at the operculum is significantly thinner than elsewhere in the cyst.

Following excystment, the empty cyst walls persist in culture and are eventually broken down (Chambers & Thompson, 1974; Sriram et al., 2008). Together reversible differentiation makes Acanthamoeba a hardy organism that is difficult to treat but fascinating to study.

Acanthamoeba and human health: GAE

In 1972, Acanthamoeba was conclusively linked to the Granulomatous Amebic

Encephalitis(GAE) (Jager & Stamm, 1972). GAE is a brain infection with near 100% mortality. Its symptoms include headache, fever, aphasia, ataxia, vomiting, intracranial pressures, seizures and eventually death (Khan, 2007). GAE results when Acanthamoeba 6 crosses the blood brain barrier and begins to destroy the cells of the brain, resulting in brain lesions and ultimately death. GAE appears to occur only in chronically ill or immunocompromised individuals (Khan, 2007). It is believed that Acanthamoeba accesses the brain by either entering the blood stream through the skin or the lung then transversing the Blood Brain Barrier or by migrating up the nasal route and invading through the olfactory bulb. GAE usually leads to a disseminated infection throughout the body with Acanthamoeba pneumonitis in the pulmonary parenchyma, and skin lesions can also be observed (Telang et al., 1996; Martinez & Visvesvara, 1997; Aichelburg et al., 2008). While GAE is the most severe disease caused by Acanthamoeba, it is also the rarest.

Acanthamoeba and human health: AK

In 1973, Acanthamoeba was linked to an eye infection, Acanthamoeba keratitis

(AK) (Naginton et al., 1974; Jones et al., 1975). The most prominent infection caused by this organism, AK is characterized by inflammation of the cornea. Acanthamoeba attaches to the cornea via MBP, invades the stroma, and causes cytotoxicity to corneal cells through the release of various cysteine and serine proteases (Lorenzo-Morales et al.,

2005). This results in an intense inflammatory response followed by the necrosis of the stromal cells. The large number of cysts seen late in Acanthamoeba infection suggest that cell-mediated immune response to Acanthamoeba is primarily directed towards trophozoites as opposed to cysts, which results in a high incidence of persistent infections

(Aksozek et al., 2002). AK is classically identified by a ring in the stroma which is the result of neutrophils infiltrating the stroma in an attempt to clear the Acanthamoeba.

AK can result in blindness if not rigorously treated. Acanthamoeba is resistant to most antimicrobials and under certain drug conditions Acanthamoeba will encyst and become dormant within the cornea. This can result in a persistent infection that can last or reemerge years after the original infection. It is also a major concern in cornea transplants where infected donor corneas can be lost following transplant. Treatment typically of AK includes hourly topical applications of chlorhexidine or polyhexamethylene biguianide but relapse may occur (Chomicz et al., 2005). Treatment can last from weeks to months. While AK is exceptionally rare, cases of AK are steadily increasing.

AK is most prevalent among contact lens wearers, especially when contact lens are washed or stored in tap water (Stehr-Green et al., 1989). Bacteria found in tap water create a biofilm on contact lens and their cases that Acanthamoeba (also from tap water) can feed upon. AK is also associated with cornea trauma because injured corneas have increased expression of mannose-containing glycoproteins, which are the main receptors for Acanthamoeba (Jaison et al., 1998). It is thought that another risk factor is an immune deficiency that may protect most people from AK. Mannosylated glycoproteins that associate with IgA in tears ordinarily may protect against AK (Said et al., 2001).

People without these specific antibodies may be at higher risk for AK infection.

Humans are exposed to Acanthamoeba on a consistent basis. Up to 80% of unaffected individuals have been shown to harbor anti-Acanthamoeba antibodies, indicating widespread exposure to these organisms (Alizadeh et al., 2001; Chappel et al.,

2001). However, diagnosis of AK is generally very low. An estimated 700 cases were

8 reported in the United States between 1973 and 1996 (Telang et al., 1996; Martinez &

Visvesvara, 1997). In histological preparations, the amoebae look very similar to keratoplasts, neutrophils, or monocytes, and this can often lead to false negative and/or false positive results. It has been estimated that 70% of AK cases are misdiagnosed clinically as viral keratitis (Bacon et al., 1993; Goodall et al., 1996; Thebpatiphat et al.,

2007). Thus, many patients are initially treated with inappropriate drug therapies. This lag time hampers disease resolution as several studies have found that earlier diagnosis and treatment results in lower morbidity and better visual outcome (Bacon et al., 1993a;

Bacon et al., 1993b; Tan et al., 1993).

Acanthamoeba has gained public attention in recent years as a result of an outbreak of Acanthamoeba keratitis (AK) in Chicago, Illinois, USA and a subsequent increase in AK in several other locations (Thebpatiphat et al., 2007). Prior to 2002,

Chicago clinicians would document 2-3 cases of AK per year. Since 2003, they have been averaging 30 cases a year. These increases in AK incidences corresponded with

EPA-mandated water treatment changes that required US cities to reduce the amount of chlorine by-products in the water. This has potentially allowed the biofilm which

Acanthamoeba feeds upon to proliferate in the water system. In 2009, the AK outbreak resulted in the voluntary recall of Advanced Medical Optics Complete Moisture Plus lens care solution. However, no solution was found to be contaminated (Vernani et al 2009).

Acanthamoeba genomics

In 2008, the genome of the isolate Acanthamoeba castellanii Neff was sequenced with 0.5X coverage. The genome size is estimated to be 1 x 108 bp (Moon et al., 2009).

Using data from EST libraries, A. castellanii Neff was expected to have an average of 3.0 exons per gene. Since the first glimpse of the genome, numerous EST libraries of A. castellanii Neff as well as other isolates of Acanthamoeba have been examined. The usefulness of the genome sequencing is debatable due to the lack of annotation as well as the choice of Acanthamoeba strain used in the sequencing. Acanthamoeba castellanii

Neff is an unusual Acanthamoeba when compared to Acanthamoeba of its most similar genotype class. Isolates similar to A. castellanii Neff have very rarely been reisolated. A. castellanii Neff is nonpathogenic and has been grown in the lab since 1957, making it an unlikely candidate for representing a genus for which research is largely focused on the organism’s capacity to cause disease. As a consequence, a large portion of the research documented here is focused on understanding more about the genetics of the broader constituents of the genus Acanthamoeba and the challenges faced when so little is known about their genetic contents. At this point, even the ploidy level of members of

Acanthamoeba is speculative, with the number of chromosomes reported in the literature varying greatly. Depending on the Acanthamoeba strain and cell cycle stage, chromosome numbers have been reported to be between 9 to 21 (Rimm et al., 1988;

Matsunaga et al., 1998). The research that follows has suggested that Acanthamoeba is at least diploid and perhaps occasionally triploid. This large discrepancy is due to

Acanthamoeba's apparent destruction of up to 50% of its DNA during encystment (Byers et al., 1991) and the difficulty in synchronizing the cell cycle across a culture of

Acanthamoeba. Since Acanthamoeba was first discovered, attempts have been made to properly classify it and to understand differences between clinical and environmental

Acanthamoeba. The following is a review of the current methods of classification of

Acanthamoeba and the research that follows attempts to expand on that classification, offer alternatives to historically used classification and to draw some conclusions about the variable nature of Acanthamoeba's pathogenicity.

Acanthamoeba classification: Morphology and other methods

Castellani observed an amoeba in cultures of Cryptococcus pararoseus in 1930, and Volnonsky used this strain as the basis of the first description of the genus

Acanthamoeba (Castellani, 1930; Volkonsky, 1931). Historically, Acanthamoeba was classified based on morphology largely focused on characters related to the mitotic spindle, cyst features, and trophozoite size and shape (Visvesvara, 1991). The genus

Acanthamoeba has been divided into more than 25 named species, which were placed into three morphological groups based on cell morphology (Table 1). These groups were designated Groups I, II, and III. Group I is the strongly supported by molecular methods

(see below), but the support for a separation between Groups II and III are less clear. The characteristic features of Group I organisms are extremely large Acanthamoeba trophozoites with cysts that either have 4 or less arms or more than 6 arms. The plasticity of this organism is very impressive. Figure 1 shows Group I Acanthamoeba in panels C,

F, G and H. Interestingly, all three cysts are from the same Acanthamoeba, showing that even clonal Acanthamoeba isolates can vary in their cyst structure. Acanthamoeba

Group I contained named species A. astronyxis, A. tubashi, A. comondani and A. echinulata. Acanthamoeba Group II is the most common morphological group within

Acanthamoeba, having cysts with rounded arms as seen in Figure 1 panel E. It includes

Table 1. Acanthamoeba species names with genotype based of the 18S rDNA gene.

T-type Species Name T1 A. species T2 A. palestinensis; A. pustulosa T3 A. griffin; A. pearcei T4 A. castellanii, A. lugdunensis, A. polyphaga, A. rhysodes, A. divionensis, A. mauritaniensis, A. quina, A. terricola, and A. triangularis, A. royreba T5 A. lenticulata T6 A. palestinensis, A. hatchetti T7 A. astronyxis T8 A. tubiashi T9 A. comandoni T10 A. culbertsoni T11 A. hatchetti; A. stevensoni T12 A. healyi T13 A species T14 A. species T15 A. jacobsi T16 A species T17 A. species T18 A. species

12 species names A. castellanii, and A. polyphaga (Table 1). Finally, Group III consists of smaller cysts <15um, with 3-5 points, and includes A. palestinensis and A. lenticulata, among others. In this dissertation, Group designations and species names will not usually be used beyond the introduction. Acanthamoeba will be identified solely by their DNA- based genotype due to the considerable misunderstanding that results from species names that are largely useless and morphological characteristics that are plastic. Acanthamoeba trophozoite size is an imprecise method of classification due to Acanthamoeba size differences based on method of culturing (personal observation).

A variety of other methods of classification have emerged in Acanthamoeba research. Isoenzyme patterns have been used for identification but largely did not agree with species names and thus this method of identification failed to become standard

(DeJonckheere & Michael, 1988). Later molecular work made species names invalid as many names corresponded with different genotypes (Table 1). The use of RFLP techniques failed to become adopted as a method of identification because it could not distinguish between Acanthamoeba castellanii and Acanthamoeba polyphaga (Kong &

Chung, 1996). Interestingly, those two species names can refer to isolates from disparate genotypes as well as nearly identical Acanthamoeba isolates. Finally, the use of nuclear

18S ribosomal RNA gene (18S rDNA) for genotyping became the standard method of identifying of Acanthamoeba.

18S nuclear small subunit ribosomal DNA gene

Over the last two decades, advances in molecular techniques have led to the development of methods for genotyping of Acanthamoeba isolates. A fast, reliable, and

13 repeatable identification method based on the nuclear 18S small subunit ribosomal RNA gene (rDNA) is now used by investigators worldwide to identify Acanthamoeba isolates.

A genotype in Acanthamoeba was originally arbitrarily defined as including all strains whose 18S rRNA gene sequences are less than 6% divergent from one another in a standard alignment of the sequence (Gast et al., 1996). This criterion was adjusted to divergences less than 5% in a subsequent expanded study, and has remained at this value since (Stothard et al., 1998). Thus, different strains within a “genotype” may have different 18S rDNA sequences. In Acanthamoeba, genotypes have come to be referred to as “T-types”, which has resulted in genotypes being designated T1-T18 (Table 1; Figure

2).

More than a thousand isolates of Acanthamoeba have been characterized using the 18S rDNA gene, and several dozen isolates have also been examined using mitochondrial 16S gene sequences, which supports the phylogenetic conclusions based on the 18S rDNA data. At least eighteen different Acanthamoeba genotypic groups have been identified, beginning with twelve first identified in our lab (Gast et al., 1996;

Stothard et al., 1998). This DNA classification system is now used globally, permitting any researcher to clearly identify an Acanthamoeba isolate and allowing the accurate description of those isolates associated with disease. A phylogenetic tree of the various genotypes demonstrates the diversity of the genus Acanthamoeba (Figure 2). The different genotypes differ in their prevalence, and in their association with different clinical presentations. The following presents a review of some of the characteristics of different genotypes.

91 T6 T2 100 T4 68 96 T4-Neff T11 58 T3 92 T18

84 T1

99 T14 90 T12 96 T10 53 75 T16 T13 75 T15 T5 T9 87 T8 100 T7 T17 Protacanthamoeba

0.02

Figure 2. Phylogenetic relationship of the 18 genotypes of the genus Acanthamoeba using the full length 18S rDNA gene. Tree was constructed using neighbor-joining and bootstrap values (1000 replicates) are indicated next to nodes. The scale bar indicates the number of nucleotide changes per inch. Protacanthamoeba was used as an outgroup to show relative variation within the genus Acanthamoeba.

T2 & T6

T2 and T6 are largely environmental isolates that are phylogenetically close to one another and were often confused prior to, and since the establishment of the 18S genotyping. T2 has been isolated from soil, which could be a logical source of human contact. Interestingly, both have been isolated from clinical AK cases (Walochnik et al.,

2000; Maghsood et al., 2005).

T3, T4, T11

The T4 genotype is of special interest because most keratitis isolates belong to this single lineage (Stothard et al., 1998; Schroeder et al., 2001). T3 and T11 are closely related to T4 and have also been found to be responsible for multiple cases of keratitis

(Ledee et al., 1996). The close genetic similarity relationship may explain why these three genotypes have all been observed in AK infections. From more than 1400 samples of all Acanthamoeba collected regardless of source, 84% are either T3, T4 or T11 (Figure

3). In addition to its prevalent involvement in AK, genotype T4 is capable of causing every type of Acanthamoeba infection (Figure 3). T4 Acanthamoeba have been found to cause CNS infections, specifically GAE (Martinez et al., 1993; Booton et al., 2005).

Acanthamoeba samples obtained from HIV-infected patients found T4 isolates in skin and lung samples (Booton et al., 2005).

Genotype T4 is the most common type of Acanthamoeba encountered in both clinical and environmental settings. It is also the most diverse of the genotypes of

Acanthamoeba It is not clear if this is simply because of the overwhelming number of this genotype that have been isolated compared to other genotypes, or if this genotype is 16

17 simply genetically more diverse. Examination of several genetic systems indicates that the T4 genotype can be divided into several different subclades, as will be shown in this thesis. However, whether or not a particular sub-lineage of T4 is more pathogenic has not been investigated. The genetic relationships among isolates within the T4 genotype will be examined in depth later in this document.

T1, T10, T12, T14, T18

Genotypes T1, T10, T12, T14 and T18 form a cluster of phylogenetically somewhat similar Acanthamoeba. Although they represent some of the rarest

Acanthamoeba, they are nearly always isolated from a clinical setting. T1, T10, T12, and

T18 have been isolated from CNS infections, as well as disseminated infections (Stothard et al., 1998; Crary unpublished data). It is currently unknown what makes these four genotypes capable of the most deadly instances of Acanthamoeba disease; possibly they have some unique properties that allow them to specifically target the CNS and spread throughout the body. The T18 is of interest in the following studies as it is a reliable outgroup for the T3-T4-T11 lineage while still being phylogenetically close to draw conclusions about genetic relationships.Given how rarely they have been isolated from the environment, it naturally raises questions as to their natural habit, and if they are more suited to a role as a parasite than as a free-living amoeba in the environment. T14 has only been isolated in two clinical AK cases (Di Cave et al., 2009). With such rare observation, it is unknown whether it is capable of producing other more deadly

Acanthamoeba infections.

T13 & T16

Very little is known about these two genotypes. They are likely the rarest and the most recently identified genotypes in Acanthamoeba. T13 and T16 have only been isolated from environmental sources (Corsaro & Venditti, 2010). Nearly simultaneous identification of individual Acanthamoeba isolates has resulted in multiple isolates being designated as T16. With divergent Acanthamoeba both being named T16, it has caused considerable confusion about how to designate new genotypes of Acanthamoeba. The

“T18” mentioned above is a newly identified genotype from the OSU archive which had previously been identified as T4 (Visvesvara et al., 2007). It is likely that once the T16 misnaming has been corrected that it will become T19 or even T20. For sake of avoiding confusion, T16 will be grouped together and only the 18 genotypes reliably described will be discussed.

T5 & T15

T5 is the second most prevalent Acanthamoeba genotype found (Booton et al.,

2005). T5 has been isolated from numerous environmental samples. It is often associated with polluted or sewage dump locations. Until recently with the identification of a genotype T5 isolated from AK infection, and another from a fatal infection in a heart transplant patient, genotype T5 had not been observed in an infection (Spanakos et al.,

2006; Barete et al., 2007). An argument can be made that T5 Acanthamoeba are found in environmental locations that are unlikely to permit exposure to humans. However, T5 isolates have been found strains in beach and tap water surveys in close proximity to T4

(Booton et al., 2004). Therefore, at least in some cases, T5 strains are in locations where 19 interaction with humans is possible. T15 is another Acanthamoeba that has been found both in the environmental and AK infections (Horn et al., 1999). T15 has been found worldwide in the environment both in soil and in water.

T7, T8, T9, T17

The Group I Acanthamoeba comprise genotypes T7, T8, T9, and T17. They consist of Acanthamoeba that are larger in size than the other Acanthamoeba species.

These genotypes are genetically very divergent from the rest of the genus Acanthamoeba and consistently resolve phylogenetically as an outgroup to the rest of the genus (Figure

2). Interestingly, these Acanthamoeba have never been isolated in an infection and are designated as strictly nonpathogenic. While they have been shown to be cytopathic in vitro, they have also failed to express critical proteases that are required for

Acanthamoeba infection of humans. Prior to the following studies, the only genes sequenced from this group were from the mitochondrial 16S and nuclear18S rRNA genes.

SUMMARY

The use of the 18S rDNA genotyping has been beneficial in identifying new genotypes. However, as discussed above, the pitfall is how a new genotype is defined.

The sequential numbering system can be problematic when Acanthamoeba isolates are being described at a high rate. Additionally, some scientists continue to use morphology for classification, or apply species names when referring and classifying Acanthamoeba, causing considerable confusion about what precise genotype or isolate of Acanthamoeba is being used in a particular study. Even more challenging is when new genotypes are 20 characterized that are less than 6% different from previously described genotypes.

However, the widespread use of the genotypic identification method based on the 18S rRNA gene has shown that at least one isolate from 14 of the ~18 genotypes have the capacity to cause disease in humans, making the genus Acanthamoeba as a whole a serious public health risk (Figure 3).

CHAPTER 2

CYTOCHROME C OXIDASE SUBUNIT I

TAXON BARCODING AND PHYLOGENETIC

ANALYSIS OF THE PATHOGENIC

PROTISTAN GENERA ACANTHAMOEBA,

PROTACANTHAMOEBA AND BALAMUTHIA

INTRODUCTION

Acanthamoeba, Balamuthia and Protacanthamoeba are free-living protists found in the environment (Dykova et al., 2005; Visvesvara et al., 2007). Both Acanthamoeba and Balamuthia have been found to cause a devastating brain infection, Granulomatous

Amebic Encephalitis (GAE) (Visvesvara et al., 2007). Protacanthamoeba has been isolated from fish, showing that they too can infect a multicellular host (Dykova et al.,

2005). At present, no human diseases have been associated with Protacanthamoeba.

While Acanthamoeba can only cause GAE in the immunocompromised, such as AIDS or

22 transplant patients, Balamuthia can cause this infection in any human, regardless of their immune status. Historically, both genera have been identified in part by using the sequences in the 18S rRNA gene. As previously discussed, Acanthamoeba has a large amount of genetic variability found in the 18S rRNA gene between genotypes as well as within the genotype T4. Unlike Acanthamoeba, Balamuthia has very little sequence variability in the 18S rRNA gene differentiating isolates, regardless of source (clinical or environmental) (Booton et al., 2003). This causes serious questions about the overall pathogenicity of the genus Balamuthia, considering the low level of genetic diversity found in the nuclear 18S gene. With such low levels of genetic variability, it would suggest that all isolates of this genus would contain virulence factors with a similar ability to cause disease. Analysis of sequences from 18S rDNA shows Balamuthia,

Protacanthamoeba and Acanthamoeba to be phylogenetically close to each other, especially when compared to hypothesized close neighbors of free-living amoeba, specifically Hartmannella (Figure 4). Of note, based on the 18S rRNA gene, Balamuthia and Acanthamoeba are more closely related than are Acanthamoeba and

Protacanthamoeba. Pairwise distances between Balamuthia and Acanthamoeba T4 wer

8.8% compared to 11% between Acanthamoeba T4 and Protacanthamoeba.

98 Acanthamoeba T4 84 Acanthamoeba T17 81 Balamuthia mandrillaris Protacanthamoeba bohemica Hartmannella vermiformis Dictyostelium discoideum

0.05

Acanthamoeba P. Acanthamoeba B. D.

T4 bohemica T17 mandrillaris discoideum

P. bohemica 0.118 — — — — Acanthamoeba T17 0.095 0.153 — — — B. mandrillaris 0.088 0.128 0.130 — — D. discoideum 0.299 0.326 0.329 0.305 — H. vermiformis 0.159 0.168 0.208 0.165 0.330

Figure 4. Phylogenetic relationship of the free-living amoeba, Acanthamoeba, Balamuthia, Hartmannella and Protacanthamoeba using the full-length 18S rDNA gene. The slime mold Dictyostelium is used as an out-group. This tree was constructed using the neighbor-joining algorithm. Pairwise distances between the various genera are shown below. Genbank Accession numbers for free-living amoeba: Acanthamoeba T4: AY549558; Protacanthamoeba: AY960120; Hartmannella: GU001158; Dictyostelium: AM168040; Acanthamoeba T1718S rRNA gene will be submitted to Gebank.

Acanthamoeba genotypes were previously established using the 18S ribosomal

RNA gene. Two isolates were found to be the same genotype if they were less than 6% divergent using their full length 18S rRNA gene. However, to confirm the phylogenetic relationships within the genus Acanthamoeba, an additional locus was used to examine

Acanthamoeba with previously established genotypes. The mitochondrial 16S-like rDNA gene was chosen as the second loci because it has a similar function to 18S rDNA, but it was expected to evolve at a different, faster rate due to being a mitochondrial encoded gene. Acanthamoeba mitochondrial 16S-like rDNA sequences were used to determine if mitochondrial DNA data also supported the proposed nuclear 18S rDNA phylogeny between Acanthamoeba genotypes. Previous studies in our lab that used 16S- like rRNS gene sequences for analysis of Acanthamoeba showed that the mitochondrial gene showed greater divergence than the nuclear 18S rRNA gene (Ledee et al. 2003).

This rapid rate of evolution might strengthen and resolve relationships between

Acanthamoeba genotypes as well as the variable T4. The genetic lineages identified by

16S-like rDNA sequences were consistent with the lineages identified using the nuclear

18S rRNA gene, thus increasing our confidence in the 18S rDNA phylogeny (Ledee et al.

2003).

In addition to the 16S-like rRNA gene found on the mitochondrial genome, other genes in this organelle have been actively examined for their potential phylogenetic utility in many other organisms. One of the genes, cytochrome c oxidase subunit I (COI), has been screened recently in a great number of organisms. COI encodes an enzyme in the electron transport chain. This gene is evolutionarily conserved and can be examined

25 in all eukaryotic organisms. In addition, due to the multiple copies of the mitochondrial genome in eukaryotes, the amplification of this gene by standard PCR methods using conserved COI primers is easily performed. Further, the regions of high DNA sequence conservation permit the design of conserved primers for use in many animal species.

Despite the presence of conserved sequences, this gene also contains regions of relatively high DNA variability due to the more rapid accumulation of sequence substitutions found in the mitochondrial versus the nuclear genome. For these reasons it has been proposed that the COI gene sequence is a convenient target to establish a DNA "barcode" for each eukaryotic species.

A DNA barcode is envisioned as a biological equivalent to the Universal Product

Code (UPC) used to unambiguously identify a specimen. Theoretically the information content in an ~650 bp region of the COI is sufficient for individual identification of all eukaryotic species on earth (Hebert et al., 2003; Marshall et al., 2005). An online database (DNA Working Group, Consortium for the Barcode of Life) has been established to collect COI sequences determined by researchers worldwide

(http://www.barcodeoflife.org/). Most of the COI data collected to date has been focused on higher animal species, such as arthropods, birds, fishes, and mammals (Hebert et al.,

2004; Ward et al., 2005; Hajibabaei et al., 2006; Hajibabaei et al., 2007).

There are a number of protistan COI sequences currently available in the

GenBank database, but there have been no large scale "barcoding" projects reported in protistan genera, to our knowledge. In the current study we have screened a large number of opportunistically pathogenic protistan organisms of the genera Acanthamoeba,

Balamuthia and Protacanthamoeba to determine the utility of COI "barcoding" for lower eukaryotes. In addition, previous barcoding results have found that the full length COI amplimer is often not required for identification of a species. To determine if a

“minibarcode” also functions as an identifier of Acanthamoeba, as well as examining its ability to predict genotypes, the phylogenetic analysis of the minibarcode for

Acanthamoeba was also performed. This shorter minibarcode would potentially allow a more cost effective method of identifying Acanthamoeba as well as determining genotypes (Meusnier et al., 2008).

METHODS

Acanthamoeba isolates to be examined were selected based on previously established genotypes identified by using the 18S rRNA gene (Table 2). A random sampling of Acanthamoeba isolates from genotype T4 was selected from a variety of sources (environmental, AK, non-AK). Additional isolates were selected as representatives of all major Acanthamoeba genotypes, as well as several rare ones.

Acanthamoeba isolates were grown in amoeba saline (Appendix) seeded with

Enterbacter aerogenes. When trophozoites reached a density of 1x104 cells/ml, DNA was extracted using standard methods (Appendix). The Balamuthia mandrillaris isolates were also examined using DNA extracted for previous studies for the current COI barcoding analyses.

Two universal primers for the COI gene were used in these amplifications:

LCO1490 and HCO2198 (Appendix Table 1). PCR was performed using standard

ThermopolTaq (Appendix). 25µl PCR reactions were performed for each sample under 27

Table 2. Protists and their source sued in DNA Barcode Study. Acanthamoeba genotypes are identified as appropriate. AK: Acanthamoeba keratitis; ENV:Environmental; DI:Disseminated Infection; GAE:Granulomatous Amebic Encephalitis; NI:Nasal Infection; SC:Subcutaneous.

OSU ID Genotype Origin OSU ID Genotype Origin A. spp. 02-003 T5 ENV A. spp. 07-029 T10 AK A. spp. 02-031 T4 AK A. spp. 07-032 T4 AK A. spp. 02-042 T11 ENV A. spp. 07-033 T4 SC A. spp. 02-043 T5 ENV A. spp. 07-041 T7 ENV A. spp. 02-044 T5 ENV A. spp. 07-042 T3 ENV A. spp. 03-001 T4 ENV A. spp. 07-047 T4 AK A. spp. 03-023 T10 GAE A. spp. 07-069 T4 AK A. spp. 03-027 T4 SC A. spp. 07-095 T4 AK A. spp. 03-039 T4 AK A. spp. 08-002 T1 DI A. spp. 04-020 T18 DI A. spp. 08-005 T4 AK A. spp. 05-003 T4 DI A. spp. 08-016 T4 DI A. spp. 05-009 T4 AK A. spp. 09-006 T2/6 ENV A. spp. 05-011 T4 AK A. spp. 10-025 T4 AK A. spp. 05-020 T3 AK A. spp. 106096 T4 AK A. spp. 05-022 T17 ENV A. spp. 162999 T4 AK A. spp. 05-023 T4 AK A. spp. 2802 T6 ENV A. spp. 06-004 T4 AK A. spp. 30732 T15 ENV A. sp 06-005 T4 AK ATCC 30976 Protacanthamoeba DI A. spp. 06-024 T4 AK A. spp. 50239 T4 AK A. spp. 06-025 T4 AK A. astro T7 ENV A. spp. 06-033 T4 AK A. tubiashi T8 ENV A. spp. 06-051 T4 AK Ambiente Balamuthia GAE A. spp. 06-071 T4 AK A. spp. V006 T1 GAE A. spp. 07-003 T4 AK A. spp. V013 T12 GAE A. spp. 07-004 T4 AK A. spp. V017 T4 NI A. spp. 07-008 T1 NI A. spp. MEX7 T4 ENV A. spp. 07-011 T1 GAE V039 Balamuthia GAE A. spp. 07-012 T1 GAE V118 Balamuthia GAE A. spp. 07-019 T4 AK V194 Balamuthia GAE A. spp. 07-020 T10 NI V416 Balamuthia GAE A. spp. 07-025 T1 AK V437 Balamuthia GAE A. spp. 07-027 T4 AK V451 Balamuthia GAE

28 the following conditions: an initial denaturing step of 5 min at 94° C; three cycles of 30s

@ 94°C, 1min 30s @ 45°C, and 1min 30s @ 72°C, followed by 35 cycles of the following:

30s @ 94°C, 1min 30s @ 50°C, and 1min @ 72°C. A final extension step of 5min @ 72°C completed the amplification protocol. Positive and negative controls were included in each amplification protocol. The expected amplimer size of 715bp is based upon the size of this product predicted from the Acanthamoeba castellanii Neff mitochondrial genome reference data (Accession number NC_001637). A sample of 4-6µl of each PCR reaction was screened for successful amplification by agarose gel electrophoresis. The remainder of the PCR reactions from samples yielding the appropriate band of ~715bp were purified using a PEG extraction as described in the Appendix. A sample of 4µl of the purified

PCR reactions were then sequenced using automated fluorescent sequencing as described in the Appendix using 2pm of the LCO1490 and HCO2198 primers. Sequences were aligned using MUSCLE as implemented in MEGA5.0, and phylogenetic analysis and tree building was done through this program. All phylogenetic trees were constructed using the neighbor-joining algorithm. The consistency of data with the neighbor-joining tree were examined using bootstrapping. Bootstrap values (1000 replicates) that were greater than 50% are shown next to the nodes. All sequences were submitted to

Genbank.

RESULTS

The expanding use of the COI gene in DNA “barcoding” studies led us to examine the utility of this gene in this important group of opportunistic pathogens. The

29 traditional interpretation of barcode results would use DNA from phylogenetically identified reference standards to determine the barcode associated with a “species.”

However, the use of species names in Acanthamoeba is problematic, since “species” are based on plastic morphological characters. If the COI gene is to be used as a legitimate barcode, the phylogenetic reconstructions based on this gene must be comparable to some recognized standard. At present, that standard is provided by the genotypes (T types) from the 18S rDNA gene sequence data.

To answer the question of whether COI sequences seem to provide a legitimate barcode, we have determined the full COI sequences, and used these sequences for DNA

“barcoding” in 56 isolates of Acanthamoeba spp., representing fifteen Acanthamoeba genotypes, in seven isolates of Balamuthia mandrillaris, and in one isolate of

Protoacanthamoeba caledonica (Table 2). The COI gene shows a high amount of diversity between the five genera. Protacanthamoeba, Balamuthia and Acanthamoeba were compared to Hartmannella and Dictyostelium. The differentiation between these genera is significant, with pairwise differences between the three genera protists being greater than 25% (Figure 5). Again, Balamuthia and Acanthamoeba are closer phylogenetically than to other protists. Hartmannella is closer to Acanthamoeba than

Protacanthamoeba, unlike the 18S rRNA gene.

Diversity within a genus was considerably less. The average pairwise difference between Balamuthia isolates was 2.4%, showing the low diversity between individual isolates of Balamuthia (Figure 5, 6). Unlike Balamuthia, Acanthamoeba shows much higher level of diversity, where the average pairwise difference between isolates was

Acanthamoeba T17 Acanthamoeba T4 Balamuthia mandrillaris Hartmannella vermiformis Protacanthamoeba caledonica Dictyostelium discoideum

0.05

Acanthamoeba Acanthamoeba B. D. P. bohemica T4 T17 mandrillaris discoideum

P. bohemica 0.320 — — — — Acanthamoeba T17 0.206 0.262 — — — B. mandrillaris 0.341 0.365 0.313 0.024 — D. discoideum 0.409 0.425 0.353 0.479 — H. vermiformis 0.280 0.374 0.301 0.352 0.444

Figure 5. Phylogenetic relationship between the free living amoeba, Acanthamoeba, Balamuthia, and Protacanthamoeba using COI gene. Dictyostelium is used as an outgroup. Pairwise distances between the various genera are shown below. Acanthamoeba, Protacanthamoeba and Balamuthia COI sequences will be submitted to Genbank. Accession numbers: Hartmannella: GU828005; Dictyostelium: AB000109.

07-032-T4-A 89 08-005-T4-B 06-051-T4-A 97 05-011-T4-A 93 06-005-T4-B 78 07-004-T4-B V017-T4-B 92 07-047-T4-A 07-003-T4-B 98 07-027-T4-A 100 07-033-T4-A 03-027-T4-A

100 06-024-T4-A 55 73 08-016-T4-A 100 06-025-T4-A 68 06-071-T4-A 05-003-T4-A 02-031-T4-A

96 05-023-T4-B 100 07-019-T4-B 100 106096-T4-B 100 03-001-Neff Neff Reference Strain 50 100 06-033-T4-E 50 07-095-T4-E 100 06-004-T4-E

99 03-039-T4-E 78 98 07-069-T4-E 162999-T4-C

100 10-025-T4-C 99 Mex7-T4-C 05-009-T4-D 99 50239-T4-D 02-042-T11 05-020-T3 100 07-042-T3 V013-T12 04-020-T18 100 08-002-T1 50 V006-T1 07-008-T1 100 07-011-T1 100 07-012-T1 07-025-T1 99 02-043-T5 100 02-044-T5 68 02-003-T5 30732-T15 03-023-T10 07-020-T10 78 100 07-029-T10 09-006-T2 84 2802-T6 05-022-T17

98 07-041-T7 51 A. astro-T9 80 A. tubiashi-T8 30976-Proto V451-Bala 52 Ambiente-Bala 100 V194-Bala 100 V039-Bala V188-Bala

51 V416-Bala 81 V437-Bala

0.05

Figure 6. Phylogenetic reconstruction of the genera Acanthamoeba, Balamuthia and Protacanthamoeba using the COI mitochondrial gene. 18S rRNA gene sub-clades are identified next to OSU ID# for the T4 genotype. All other genotypes are indicated. 32

19%, with the maximum observed being 26% (Table 3). While most pairwise distances between genotypes are greater than 19%, certain genotypes specifically the Group I (T7-

T9,T17) show significantly small genotypes between 10-13% (Table 3). However, the average pairwise distance between T4 is 12.1% with the highest distance being 18%.

Thus, there is no easily defined pairwise difference in the COI gene that can be used as the cutoff for a new genotype like the 18S rRNA gene where the divergence must be greater than 5% to determine a new genotype. New genotypes will still require the use of the 18S to determine if they are indeed a novel genotype. However, COI gene is able to distinguish previously identified genotypes based on the 18S rRNA gene. This will allow

COI as a reasonable alternative to the 18S, as long as the Acanthamoeba is phylogenetically similar to one of the genotypes already described by COI.

When compared to the phylogeny of the genus Acanthamoeba obtained using the

18S rRNA gene sequences, evolutionary relationships resolve using the COI sequences are fairly similar (Figure 6, 7). The T4 genotype remains very diverse; sub-clades within the genotype can be identified (Figure 7). T4 sub-clades that have been suggested using

18S rRNA sequences (Fuerst, unpublished) are identified as A-E next to the OSU ID for the T4 genotype. No Acanthamoeba isolate from the 18S rDNA clade F was available for the Barcode study due to its rarity and lack of occurrence in the OSU archive. Several

18S rDNA sub-clades resolve as monophyletic clades using COI, as seen in the 18S rRNA gene (Figure 6; 8). Specifically, sub-clade C and D resolve outside of the rest of the T4 genotype. Sub-clade E has two separate clusters that resolve together. However,

18S rDNA sub-clades A and B do not resolve with COI in the manner seen for the 18S

Figure 7. Comparison of the phylogenetic trees of the 18S rRNA gene (right) of the genera Acanthamoeba using major genotypes and the COI gene tree (left).

Figure 8. Phylogenetic relationship of Acanthamoeba Neff with other members of genotype T4 based on the 18S rRNA gene. The size of the triangle represents the relative number of strains in a subgroup of T4. Analysis used only sequences that exceeded 2000 basepairs and all sequences show less than 5% divergence between genotypes.

rRNA gene. These two clades are largely mixed together at the COI loci and isolates of sub-clade A and B are identical in several cases at the COI locus. T4 Acanthamoeba with identical COI sequences are infrequent in our data.

The relationships between other genotypes are maintained at the COI locus. The

T11 and T3 genotypes resolve closely with the T4 genotype as seen on the 18S rRNA gene. The pairwise distance between T4 to T3 and T11 are 19.2% and 19.5%, respectively. However, between T3 and T11, it is only 17.2% (Table 3). Within T1 or

T10 genotypes, isolates show little divergence in sequence when compared to isolates of the same genotype, which is also similar to the 18S rRNA gene. As per the 18S rDNA, the group I Acanthamoeba, T7, T8, T9 and T17 resolve outside the other genotypes of

Acanthamoeba (Figure 6). It is important to analyze a genus at multiple loci to confirm that the resulting phylogenetic tree is not skewed by some unknown variable at a single locus. The strong similarities between the 18S and COI cladograms suggest that the relationships between Acanthamoeba genotypes are well defined and consistent regardless of the gene analyzed. Together, the COI sequences show a level of diversity within Acanthamoeba that, given our current state of knowledge of other protists, would allow for proper identification of members of the genus Acanthamoeba, as well as providing information at the genotype level. However, COI cannot yet be used to establish novel genotypes without additional information from the 18S rRNA gene.

One difference between 18S rDNA and the COI phylogenetic relationships is in how the relationship between the five genera of Acanthamoeba, Balamuthia,

Hartmannella and Protacanthamoeba resolve when an out-group (Dictyostelium) is used.

In the 18S rDNA tree, Acanthamoeba and Balamuthia are sister taxa, with

Protacanthamoeba as an out-group and Hartmannella phylogenetically further from

Protacanthamoeba. However, in the COI gene, Hartmannella is phylogenetically closer to Acanthamoeba than Protacanthamoeba (Figure 5). This is likely due to the different evolutionary rates of the two genes, as well as the number of informative sites in sequences being analyzed. The 18S rRNA gene has 133 informative sites whereas the

COI gene has 178 informative sites. These additional informative sites in COI may be the reason for the change in relationship as well as the effect of using mitochondrial genes which evolve at a faster rate. Also of note is the fact that the COI sequence of

Protacanthamoeba indicates the possibility of RNA editing of the transcript of the COI gene, which could also influence how the different genera relate.

Finally, analysis of the minibarcode was done by examining the sequences between the original 5’ end of the sequence and the location of the 3’ primer that was used in the new minibarcode protocol. This resulted in 114 basepairs that was used in the following analyses. The phylogenetic relationships between the Acanthamoebidae protist genera do no change using the minibarcode, (Figure 9), as was also seen in the analyses of 18S rDNA. Additionally, several Acanthamoeba genotypes are no longer maintained as monophyletic groups. While certain genotypes (T1, T7, T8, T9, T10 and T17) are still resolving in a manner similar to the full length COI and 18S rDNA, other genotypes, especially T3, T4, T5 and T11 no longer resolve in the manner expected, most likely due to their close phylogenetic relationships and the lower information content of the

09-006-T2 05-023-T4-B 55 06-005-T4-B 07-004-T4-B 05-011-T4-A 73 06-051-T4-A 63 07-032-T4-A 08-005-T4-B 06-024-T4-A 95 08-016-T4-A 03-027-T4-A 05-003-T4-A

65 06-025-T4-A 99 06-071-T4-A V017-T4-B 07-047-T4-A 07-003-T4-B 07-027-T4-A 98 07-033-T4-A 03-001-Neff 100 Neff Reference Strain 10-025-T4-C 100 162999-T4-C Mex7-T4-C 02-042-T11 98 06-033-T4-E 73 07-095-T4-E 06-004-T4-E

84 03-039-T4-E 50 75 07-069-T4-E 05-020-T3 100 07-042-T3 99 07-019-T4-B 73 106096-T4-B 02-031-T4-A 57 05-009-T4-D 50239-T4-D 02-003-T5

100 02-043-T5 75 02-044-T5 04-020-T18 2802-T6 99 08-002-T1 V006-T1 07-008-T1 97 07-011-T1 98 07-012-T1 07-025-T1 05-022-T17

88 07-041-T7 60 A. astro-T9 A. tubiashi-T8 30732-T15 V013-T12 03-023-T10 07-020-T10 100 07-029-T10 V451-Bala Ambiente-Bala

99 V039-Bala V188-Bala 99 V194-Bala V416-Bala V437-Bala 30976-Proto

0.05 Figure 9. Phylogenetic analysis of the three protists using the minibarcode amplimer.

39 minibarcode sequence. Another major phylogenetic change using the minibarcode is how genotypes T7, T8, T9 and T17 are no longer an outgroup to the rest of the

Acanthamoeba genus. This is a major disruption to the Acanthamoeba phylogenetics as established by the 18S rRNA gene, the 16S-like rRNA gene and the full length COI gene.

As such, though the minibarcode would identify an individual protist, it may not properly identify the genotype to which an Acanthamoeba isolate belongs. Interestingly, the slight variation seen in Balamuthia when considering the full length COI sequence is largely removed when using the minibarcode. This is likely due to the decrease in informative sites to 70 when using the minibarcode. This reduction in informative sites result in both the differences in the phylogenetic relationships of the genera but also the Acanthamoeba genotypes.

DISCUSSION

With the recent advances in DNA sequencing, specifically the use of field PCR machines and NextGen sequencing, the ability to gain access to large amounts of DNA data has improved and become significantly cheaper. Such advancements allow researchers to identify samples in the field, which sparked the idea of a universal method of identification. The COI gene was chosen after analyses revealed its ability to differentiate between different eukaryotes from fish to fungi (Hebert et al., 2003).

However, little work has been done in protists, especially medically relevant protists.

Often, Acanthamoeba and Balamuthia are diagnosed by a sample of tissue (cornea, brain etc) being sent to a specific lab that cultures the protist and then uses specific primers for the 18S rDNA to determine the identity of the pathogen. This process is time consuming 40 and requires specialized knowledge. The utility of the COI gene as an alternate to the

18S rDNA sequence could have several advantages. First, as COI can determine not only the genera but also the genotype (in the case of Acanthamoeba) of the sample, it successfully substitutes for the 18S rDNA gene. However, more importantly, it requires the sequencing of only 630 bases versus over two thousand for proper identification of a genotype. This allows genotypic identification of a sample with a single read (i.e., one direction; one primer) versus a minimum of four as is needed in the 18S rDNA to get a complete sequence.

The universal primers used in the analysis of COI allow any researcher to correctly identify their sample. This will allow a larger number of protists, especially

Acanthamoeba, to be identified. A larger sampling of Acanthamoeba from all sources

(environmental vs. clinical) could allow new conclusions to be drawn about the overall pathogenicity of the genus Acanthamoeba. Specifically, since the analysis of COI sequences also indicates the presence of sub-lineages within the genotype T4, an increase in isolates from T4 being sequenced would allow for comparative analysis of clinical isolates in terms of their membership in the same T4 sub-lineages or in different lineages.

Such information is crucial in understanding the overall ability of any Acanthamoeba to cause disease, but is specifically important of T4 isolates.

Finally, unlike the 18S rDNA, the COI gene sequences can be obtained with relative ease. The 18S rRNA gene is notorious for producing poor sequencing reads as a result of PCR errors due to secondary structure of the original genomic ribosomal DNA or the presence of introns. COI has no observed introns and all Acanthamoeba and

Balamuthia isolates contain the same number of nucleotides when sequenced. This is not necessarily true for the case of Protacanthamoeba, where RNA editing appears to be occurring. The 18S rRNA genes of different isolates vary widely between genotype in the total number of nucleotides, leading to gaps in the alignments. However, unlike the

18S rDNA, COI genes are protein coding and are easily aligned, making phylogenetic analysis simpler and requiring no specialized knowledge of protist genetics. Together, these results have made COI a reasonable alternative to the 18S rDNA for identifying protists and may increase the number of these pathogens being identified in the clinical setting.

Prior to the sequencing of COI, most of the genetic variability in Acanthamoeba and Balamuthia was known through the sequencing of the 18S and 16S-like rRNA genes.

Protacanthamoeba had only had its 18S rRNA gene sequenced. The addition of COI permits Acanthamoeba and Balamuthia to be analyzed using three loci with largely different functions and evolutionary rates. Balamuthia’s COI gene results are consistent with previous data showing that these organisms show very little genetic diversity, at least within their core metabolism genes. This does lend credence to the hypothesis that all Balamuthia are potentially pathogenic, given how they are virtually identical. With the additional confirmation of the COI gene, it is likely that the 18S rDNA evolutionary analysis for these three protists is the correct one, with Balamuthia and Acanthamoeba being sister taxa and Protacanthamoeba is the out-group. Regardless, there are still some unanswered questions about how these protists are related. Sequences from more loci are going to be required before definite conclusions can be made about how the genera

Acanthamoeba, Balamuthia and Protacanthamoeba relate, as well as providing a more complete understanding about the genetic variability within Acanthamoeba and how it corresponds to pathogenicity.

CHAPTER 3

ENVIRONMENTAL ACANTHAMOEBA

ISOLATES EXHIBIT HIGH LEVELS OF

GENETIC DIVERSITY AND INDICATE

POSSIBLE MEANS OF

ACANTHAMOEBA-HUMAN INTERACTION

INTRODUCTION

Acanthamoeba has been largely researched because of its ability to become an opportunistic pathogen of humans. However, the role of Acanthamoeba in the environment is often understated. Acanthamoeba is responsible for releasing bacterial biomass back into the microbial loop. Sampling Acanthamoeba in the environment is an excellent way to begin to estimate the diversity of the genus Acanthamoeba. Up until this point, only single genotypes/isolates of Acanthamoeba have been isolated from infections. It is unknown if multiple types of Acanthamoeba infect humans simultaneously, but due to the culturing method, only a single type of Acanthamoeba is

44 identified as the causative agent. Acanthamoeba culturing from environmental samples is significantly more challenging than clinical samples. Clinical samples are largely devoid of other containments such as bacteria and fungi, especially in the case of Acanthamoeba keratitis. This absence of other single celled organisms is due to the numerous antibiotic and antifungal drugs that are used in AK cases, resulting in a nearly pure culture of

Acanthamoeba. However, when a cornea scrape is cultured for Acanthamoeba, the total sample is plated on Non-Nutrient Amoeba Saline Agar plates (Appendix). The resulting culture of Acanthamoeba is nearly always a pure clonal culture of the isolate most able to grow in the lab setting. An environmental sample (soil, water, etc) will contain numerous single celled organism which may prevent the isolation of Acanthamoeba DNA. In order to end up with a pure culture of Acanthamoeba DNA, single Acanthamoeba are removed from plates and allowed to grow clonally. The serial culturing of environmental samples results in a pure culture of Acanthamoeba but again, only the one most suited for lab culturing. This method of culturing Acanthamoeba severely hampers our ability to determine the diversity of Acanthamoeba within a sample, regardless of source. Possible alternate methods of establishing diversity outside this clonal culturing will be discussed later. Regardless, diversity of the genus can be established more easily by examining isolates in the environment.

As Acanthamoeba can inhabit most environments, it is difficult to pinpoint exactly where Acanthamoeba-human interaction occurs. It has long been believed that tap water is an important factor in the etiology of Acanthamoeba keratitis infections

(Kilvington et al., 2004). It is generally accepted that contact lens wearers who rinse

45 their lenses in tap water can introduce amoeba onto the eye (Kilvington et al., 2004).

Additionally, there is an increased AK risk associated with wearing contact lenses while showering (Joslin et al., 2007). Previous studies have shown the presence of

Acanthamoeba in domestic water supplies in Japan, Korea, Jamaica, Mexico and United

Kingdom (Kilvington et al., 2004; Jeong & Yu, 2005; Lorenzo-Morales et al., 2005;

Edagawa et al., 2009; Bonilla-Lemus et al., 2010). Small scale studies examining tap water in south Florida and Columbus, OH (unpublished data) found Acanthamoeba in the domestic water supply (Shoff et al., 2008).

Historically, our lab has been sent Acanthamoeba isolates for genotyping, resulting in a large archive of samples. While these Acanthamoeba were largely clinical isolates, recently, the lab has participated in a series of environmental surveys for

Acanthamoeba. The first survey analyzed soil samples collected from a central Ohio farm for the presence of and diversity of Acanthamoeba. The second survey was a wide spread tap water study of the Greater Chicago area, related to residences associated with known AK cases. The aim of this survey was to determine the prevalence of

Acanthamoeba and other amoebae in the greater Chicago domestic water supply and determine if Acanthamoeba isolates cultured from tap water were related genotypically to those obtained from corneal scrapings of Chicago AK patients. Finally, a survey was performed of Acanthamoeba isolates to compare with the isolates from a clinical case of

AK that was observed in Sonora, Mexico. Samples were obtained of both soil and water in Cuidad Obregon, Sonora to isolate Acanthamoeba.

METHODS

Collection of Acanthamoeba from central Ohio farm

Soil samples were collected from hay fields of an 80 acre central Ohio farm at 39°

59’ 26.174” N, -82° 19’ 0.732” W as part of an Ohio Department of Agriculture soil- testing program. Sampling occurred in December 2007, when the ambient daytime temperature was approximately 4C. Portions of the soil samples were sent to the

Department of Agriculture to be tested for pH and phosphorus concentrations. Soil samples were stored at 4C until inoculated onto non-nutrient agar plates seeded with

Enterobacter aerogenes. Acanthamoeba visibly feeding on the bacteria were transferred to a fresh plate. Samples were cultured in this manner until free of other bacterial and fungal contaminants and Acanthamoeba DNA could be extracted.

Chicago Tap Water and Acanthamoeba keratitis sampling

Corneal scrapings from the University of Illinois-Chicago (UIC) were stored in amoeba saline. Water samples were collected from sites in the greater Chicago area.

Sample sites were determined based on previous UIC AK cases since 2003 (Joslin et al.,

2006; Joslin et al., 2007). The water sampling site for each home was the cistern tank on the toilet. This site is constantly in contact with cold water supplied by the municipal water mains for that area. The storage tank is rarely (if ever) cleaned and can contain a biofilm, perfect for harboring Acanthamoeba (Shoff et al., 2008). Water was sampled using sterile swabs from the inside surface film of the lavatory cistern reservoir tank. In addition, 50 mL of tank water serving the lavatory was also sampled. 47

To culture Acanthamoeba, the swab tips were placed on non-nutrient amoeba saline (NNAS) agar plates seeded with E. aerogenes (Appendix). The 50 ml water samples were passed through a 0.45µm filter and filters were placed on NNAS plates with E. aerogenes. For AK isolates, aliquots of saline were placed in amoeba saline flasks with bacteria. After two weeks of incubation, the plates were rinsed with amoeba saline to dislodge amoebae and the washings were examined by light microscopy.

Cultures positive for Acanthamoeba were transferred to amoeba saline with bacteria.

Once cultures had reached a density of at least 104 cells/ml, genomic DNA was extracted as described in the Appendix.

Mexico environmental sampling

Water and soil samples were collected from 18 urban and suburban areas of Cd.

Obregon in the semi-desert Mexican state of Sonora located in the northwestern region of

Mexico. Acanthamoeba were isolated from these samples as previously described and

DNA collected. Additionally, a single clinical sample (10-025) from a 16-year-old male contact lenses user diagnosed with herpetic keratitis but no optimal response to medical treatment from Culiacán, Sinaloa was received and processed as was done with the environmental samples.

Genotyping of Acanthamoeba and Analysis of Diversity

Acanthamoeba DNA was amplified for sequences within the 18S rRNA gene with the genus-specific primers JDP1 and JDP2 (Appendix Table 1). Positive amplimers were sequenced using either primers JDP1, JDP2 or 892c. Sequences were aligned using

MUSCLE in MEGA5.0. The sequences from Chicago tap water samples that were positive for Acanthamoeba in households with AK cases were compared to homologous sequences from isolates from the patient’s eye, to determine if the Acanthamoeba were genetically related. For each study, phylogenetic trees were created to compare diversity of the isolates using the neighbor-joining method. Bootstrap values (1000 replicates) are shown next to nodes when greater than 50% . All Acanthamoeba 18S rRNA sequences were submitted to Genbank.

RESULTS

Central Ohio Farm Study

We recovered 6 isolates of Acanthamoeba from 6 separate soil samples from around the Ohio Farm. Amoebae were genotyped using the 18S ribosomal subunit RNA gene. Of these 4 isolates were identified as belonging to the T4 genotype, one isolate was a T2 genotype and one isolate is most closely related to an as yet unidentified group of putative Acanthamoeba-related forms referred to as “Acanthamoeba emeriidae”, which have only been seen in environmental samplings (Figure 10). Acanthamoeba genotypes were identified by comparing with previously genotyped Acanthamoeba from our lab or other reference sequences that have been deposited in the DNA database GenBank

(Figure 10). Interestingly while four isolates were determined to belong to genotype T4, and therefore are closely related, none were identical. The T4 genotype is more diverse than many of the other genotypes, although this diversity may be associated with the fact that genotype T4 is the most prevalent type within all Acanthamoeba samples. Genotype

T2 is typically associated with environmental isolates though it rarely has been found in infection. No culture of the environmental form “Acanthamoeba emeriidae” is known to exist. When we returned to the culture from which our DNA sample had been obtained to further investigate, no viable Acanthamoeba-like form remained.

There were no specific genotypic differences associated with various soil conditions. The pH of these soil samples ranged from 5.4 to 6.8. For phosphorous, standard soil levels are consistently low if they are less than 16 and very high if they are greater than 50 (Espinosa et al.). Acanthamoeba were found both in low and high phosphorus conditions.

98 09-004 77 09-010

99 09-007 99 A. sp. ATCC 30868 09-002 09-001 100 Acanthamoeba Eimeriidae 09-006 100 A. palestinensis 2802

0.01

09-001 09-002 09-004 09-006 09-007 09-010 Range Genotype T4 T4 T4 T2/6 T4 T4 - Soil pH 6.4 6.8 5.4 6.1 5.7 5.5 5.4-6.8 Phosphorus (ppm) 40 8 49 10 141 45 8-141

Figure 10. Ohio Farm Acanthamoeba genetic relationships. Soil pH and phosphorus content of soil samples associated with individual Acanthamoeba isolates.

Chicago Tap water Survey Water samples from 228 households in 143 zip codes in the Greater Chicago area were analyzed. Amoebae (regardless of genus) were present in 117 (51.9%) of these homes sampled. Thirty-seven of the zip codes (25.9%) had samples positive for

Acanthamoeba, with 85 of the 143 zip codes (59.4%) being positive for Acanthamoeba, other amoebae, or both (Figure 11). Acanthamoeba were found in 46 (20.2 %) households (Figure 11). All zip codes tested for the presence of protists are indicated in

Figure 12. Zip codes that were positive for either protists or Acanthamoeba are shown geographically (Figure 12).

Figure 11: Protist and Acanthamoeba presence in tap water by Chicago zip code and sampled households.

Figure 12. Zip codes positive for Acanthamoeba (Lines), Free-living amoeba (Black) or neither (Light grey); zip codes not tested (white).

Of the tap water samples that came from homes of AK patients, 21.2% were positive for the presence of Acanthamoeba in the tap water. However, 39.4% of AK patients lived in zip codes that had positive tap water samples for Acanthamoeba. Eight

AK patients had residences that also tested positive for Acanthamoeba. For these residences, the Acanthamoeba from AK cornea samples and from residence tap water were compared to determine if the Acanthamoeba isolates were identical.

Acanthamoeba isolates were genotyped using the 18S ribosomal DNA subunit. A total of 56 samples were genotyped, 26 tap water samples and 30 AK samples (Table 4,5;

Figure 13,14). Of these samples, 51 were determined to be from genotype T4 (22 tap water and 29 AK) and 5 were from genotype T3 (4 tap water and 1 AK), as seen in Table

4,5 No other genotypes were found in any of the tap water or AK samples. The matched samples from homes of AK patients with positive Acanthamoeba cultures from their tap water were identical in sequence in three of eight cases (Table 5, Figure 15).

Table 4. Summary of keratitis-causing Acanthamoeba isolates and tap water Acanthamoeba isolates examined in the Chicago tap water survey.

OSU ID Source Genotype 05-009 AK T4 05-011 AK T4 05-020 AK T3 06-001 AK T4 06-002 AK T4 06-004 AK T4 06-016 AK T4 06-024 AK T4 06-025 AK T4 06-027 AK T4 06-028 AK T4 06-033 AK T4 06-034 AK T4 06-053 AK T4 06-057 AK T4 06-059 AK T4 06-069 AK T4 07-047 AK T4 07-069 AK T4 07-072 AK T4 07-075 AK T4 07-076 AK T4 07-095 AK T4 08-004 AK T4 08-007 AK T4 08-008 AK T4 W06-003 Tap water T4 W06-050 Tap water T4 W06-062 Tap water T4 W06-087 Tap water T4 W06-131 Tap water T4 W06-140 Tap water T4 W06-146 Tap water T4 W06-168 Tap water T4 Continued Continued 56

Table 4 continued W06-174 Tap water T4 W06-194 Tap water T4 W06-201 Tap water T4 W06-207 Tap water T4 W06-212 Tap water T4 W06-221 Tap water T4 W06-222 Tap water T3 W06-264 Tap water T4 W06-269 Tap water T3 W07-024 Tap water T4 W07-036 Tap water T4 W07-070 Tap water T4 W07-131 Tap water T4

45 05-009 33 07-095 06-033 14 100 06-034 Acanthamoeba sp. ATCC 30871

1 72 06-004 99 07-069 05-011

2 06-016 3 21 06-027 19 08-007 42 07-072 15 08-004 07-075 69 06-057 2 26 06-069 29 06-024

84 06-028 45 06-059 100 07-047 08-008

40 06-001 40 06-053 57 07-076 35 06-002 06-025 05-020 70 Acanthamoeba giffini CCAP 1501/4

0.01

Figure 13. Genetic relationship among Acanthamoeba keratitis isolates from Chicago.

66 W06-174 34 W06-221 4 W06-201 4 W06-003

3 W06-146 53 W06-264 4 W06-140 species 2 Acanthamoeba sp. ATCC 50497 8 66 W06-062 W06-207

26 W06-194 56 W06-212 W06-140 species 1 W06-050 18 W07-036 7 W06-087 7 W06-168 14 20 W07-024 60 W07-070 62 W07-131 Acanthamoeba griffini CCAP 1501/4

89 W06-269 93 W06-222 42 W06-131

0.01

Figure 14. Genetic relationship among Acanthamoeba tap water isolates from Chicago.

Table 5. Matched Acanthamoeba isolates from keratitis patients and tap water Acanthamoeba isolates from patient’s residence.

OSU ID # Source Genotype (MATCH) 05-013 AK T4 W06-129 Tap water T4 05-014 AK T3 W06-101 Tap water T4 05-023 AK T4 W06-131 Tap water T3 06-005 AK T4 W07-073 Tap water T4 (MATCH) 06-050 AK T4 W08-029 Tap water T4 06-061 AK T4 W06-229 Tap water T4 06-071 AK T4 W06-100 Tap water T4 (MATCH) 07-073 AK T4 W07-135 Tap water T4

75 A. sp. ATCC 50496 56 W06-101 Set B W07-073 Set D 05-013 Set A MATCH 96 W06-129 Set A 05-023 Set C 65 06-005 Set D 05-014 Set B

100 A. griffini T5 89 W06-131 Set C

0.02

99 06-050 Set E MATCH 80 W08-029 Set E 53 A. sp. ATCC 50496 07-073 Set H W06-100 Set G W07-135 Set H 06-071 Set G 56 W06-229 Set F MATCH 76 06-061 Set F

0.002

Figure 15. Acanthamoeba keratitis isolates and matched Acanthamoeba tap water isolates phylogenetic relationships.

Sonora, Mexico Acanthamoeba survey

A total of 17 locations around Sonora, Mexico were sampled for Acanthamoeba, yielding 11 Acanthamoeba isolates. Among the new strains, 6 samples were isolated from water, and 5 isolates were obtained from soil samples (Table 6; Figure 16). These

Acanthamoeba were genotyped using the diagnostic fragment (DF3) of the 18S rRNA gene and compared to a clinical case of AK from Sonora (OSU10-025), which was identified as a member of genotype T4(Booton et al., 2002). Eleven environmental isolates were genotyped, resulting in seven being identified as genotype T5, and 3 identified as genotype T4. The remaining isolate, MEX-FA7, does not resolve with any known genotype. Instead, it resolves between T3 and T11 but is genetically different from both of them. Further analysis will have to be done before it can be confirmed that this is a novel Acanthamoeba genotype.

Morphological measurements of Acanthamoeba from Mexico were collected to determine if any morphological features were useful to distinguish different genotypes.

While the T5 genotype isolates averaged larger in both cyst and trophozoite size than T4, the range of sizes for T4 cysts were 13.16-17.10 µm compared to T5:10.71-23.31 µm

(Table 6). Acanthamoeba trophozoite volume for T5 isolates was also slightly larger at

520.6 µm3 compared to the average of T4’s 494.2 but the ranges also overlapped, indicating that Acanthamoeba cell size cannot be used to correctly identify a genotype

(Table 6). If anything, it only shows that genotype T5 has a more variable size than genotype T4.

Table 6. Dimensions and genotype of various Mexican Acanthamoeba isolates.

Sample Cyst (µm) Trophozoite (µm) Genotype MEX-1 13.16 26.32x19.79 T4 MEX-FA7 11.56 21.15x11.56 ? MEX-FA9 10.71 21.62x18.33 T5 MEX-3 11.28 38.54x20.68 T5 MEX-7 17.10 27.73x10.99 T4 MEX-8 15.22 21.62x18.33 T5 MEX-13 23.31 38.54x18.33 T5 10-025 13.53 30.08x11.75 T4 MEX-17 12.87 28.67x16.92 T5

MEX-28F 17.91 29.14x11.75 T5

MEX-30ª 15.98 28.76x27.73 T4

71 A. sp. L1629/99 T4 94 MEX-7 10-025

50 A. sp ATCC 50370 T4 MEX-1 89 MEX-30A Acanthamoeba sp. T11 MEX-FA7 64 Acanthamoeba pearcei T3 MEX-28F A. lenticulata ATCC 30841 T5 98 MEX-8 100 MEX-3 100 MEX-13 MEX-17

53 MEX-8A 63 MEX-FA9

0.005

Figure 16. Phylogenetic relationship among Acanthamoeba isolates from Mexico. 10- 025 is an Acanthamoeba isolate from a Mexican AK patient.

DISCUSSION

A major difficulty when determining the genotypic diversity of Acanthamoeba within a soil sample is that standard methods of isolating Acanthamoeba which rely on slow culturing of organisms from the environment material often result in ultimate overgrowth by a pure culture progeny of a single clonal organism. This means that while numerous different Acanthamoeba may exist in a single soil sample, only the clonal lineage that is most successful under the culture conditions, either through faster excystment or a higher rate of division, will be isolated. Thus, estimating the diversity of a genus within the original sample can be very difficult using standard culture methods.

In the case of the samples from the Ohio farm, all amoebae in the soil sample were likely encysted due to the ambient temperature on the collection date. Since samples were stored at 4C until plated there were no active amoebae prior to culturing.

The six distinct Ohio farm isolates recovered in this study represent the individual

Acanthamoeba within the soil sample that was able to most rapidly excyst in a controlled laboratory setting, the most able to respond to the presence of the bacterial prey provided, and thus to rapidly propagate. Essentially, the best adapted Acanthamoeba to this environment were found. The fact that such a variety of Acanthamoeba strains was isolated from a similar geographical area shows how diverse the microenvironment in the soil can be. In theory, slowly replicating, less active Acanthamoeba could never be cultured in the laboratory because they are consistently outcompeted. This may be one explanation of the observation that certain genotypes are very rare.

Forest soil samples have up to 107 active protists per gram of soil (Adl & Gupta,

2006). This study was designed to examine the diversity of the genus Acanthamoeba within a small geographical location. It has been observed that within the genus

Acanthamoeba nearly identical genetic isolates can be found on different continents, but that, nevertheless, Acanthamoeba obtained from the same location may be extremely divergent in genotype. For the studies summarized in this chapter, though the sampling done was often from a small geographical location (the Ohio farm study), the dominant

Acanthamoeba for each soil sample was different. Thus, the true extent of

Acanthamoeba diversity cannot be properly determined without using more sensitive methods such as real-time PCR or targeted PCR based on genotype specific primers.

Acanthamoeba has an important role in agriculture because of its interactions with both bacteria and plants. Mineralization of nutrients such as phosphorus is increased in the presence of protists and Acanthamoeba increases the nitrogen intake of plants (Adl &

Gupta, 2006). Thus, Acanthamoeba are actually able to improve the health and quality of the soil and plants respectively in agricultural soil. Though the Ohio farm study involved samples from the same geographical focus of about 1 km2, the soil conditions from which these amoebae were isolated were extremely varied. It does suggest that a large amount of diversity may be lost by culturing methods and many types of Acanthamoeba may remain unidentified because they do not grow well in culture. Further, more sensitive tests need to be performed to determine the true diversity of Acanthamoeba within a single microenvironment.

There were no specific genotypic differences associated with various soil conditions in either Ohio or Mexico. However, Acanthamoeba has been shown to have a very high tolerance to pH variability (Chomics et al., 2010). Additionally,

Acanthamoeba in Ohio were tolerant of various levels of phosphorus regardless of genotype. Together, these results suggest that a high level of variability in soil organisms can be found even within a small geographical location such as a single farm, which may contribute to the overall health of the soil.

In studies of Acanthamoeba related to tap water, it is believed that the recent increase in AK cases in several US metropolitan areas is linked to the reduction of allowable biocides in tap water as mandated by EPA regulations (Joslin et al., 2007;

Shoff et al., 2008; Booton et al., 2009). It was therefore expected that not only would there be a high percentage of water samples in the Chicago area with culturable

Acanthamoeba, but that many of these samples would correlate to actual AK infections and show the same geographical clustering as has been shown in the AK cases (Joslin et al., 2006). As expected, most genotypes for the tap water and AK cases were of genotype T4, with sequences similar to those previously seen. These results are consistent with previous studies that found that 72% of environmental isolates were genotype T4, and with several studies that have shown that most AK cases are also of the

T4 genotype (Schroeder et al., 2001; Booton et al., 2002; Booton et al., 2005; Yera et al.,

2008). It is likely that it is the true high prevalence of Acanthamoeba with genotype T4 in nature (and in tap water) that makes it the predominate genotype seen in both eye and brain infections (Booton et al., 2005; Yera et al., 2008). It is also possible that

Acanthamoeba of genotype T4 are just more capable of surviving different or harsher conditions that allow them to be found in such a variety of locations. They may also be able to respond faster to changing conditions that allow them to have a competitive advantage in causing disease.

Three of the eight matched samples of keratitis-causing Acanthamoeba and tap water Acanthamoeba were identical in sequence. Given the diversity of Acanthamoeba within the Chicago water system, this suggests that the AK patients came into contact with Acanthamoeba contaminated water in their residence and contracted Acanthamoeba keratitis as a result. Few studies have successfully matched Acanthamoeba isolated from an infection to Acanthamoeba from the environment due to the high level of genetic diversity of Acanthamoeba (Seal et al., 1999). These matched samples highlight the hazard of Acanthamoeba-contaminated water as a serious risk factor for contracting

Acanthamoeba keratitis. The remaining five matched sets that contained non-identical

Acanthamoeba isolates from AK and water could simply be the result of tap water that is colonized with multiple types of Acanthamoeba. Due to the clonal nature of the culture method, it is possible that the tap water from these residences could have contained numerous Acanthamoeba, but only a single strain was isolated. It is also important to note that sampling for Acanthamoeba in patients' home occurred months or years after initial diagnosis, and even later after initial infection. This means that significant changes in the population of Acanthamoeba colonizing a patient's water system could have occurred between infection and sampling. Future methods of comparing clinical and

68 environmental isolates will require more sensitive methods of identification so that all

Acanthamoeba within an environmental sample are properly identified.

Geographical analysis of Acanthamoeba and other protists in the tap water suggest a random distribution of zip codes that are contaminated. The question of whether the distance from water treatment plants for these zip codes is related to the degree of the colonization of the water system by both protists and other microorganisms that they can feed on remains unanswered at this time due to limited availability of information on the structure of the water systems in the Chicago area. Regardless, the wide spread contamination of the water supply around Chicago by protists shows that it is a risk across this geographical area, since the occurrence of Acanthamoeba shows no pattern of increased prevalence in a particular geographical area, but rather shows fairly uniform distribution of Acanthamoeba over a widespread area in Greater Chicago. Further studies will be needed to draw more accurate conclusions about prevalence in geographical locations around Chicago and the risk for microbial contamination of the water system.

These three surveys of Acanthamoeba diversity were meant to address very different questions. The Ohio farm survey was meant to determine the diversity of the genus Acanthamoeba within a small geographical area while the Mexico survey was attempting to determine the genotypes of Acanthamoeba found in Sonora and how they relate to the Acanthamoeba isolated from a keratitis patient. The Chicago survey was meantto determine prevalence of Acanthamoeba in the domestic water supply and relating them to previous AK cases. However, all three surveys had several features in common. First, diverse environmental sources will yield Acanthamoeba with alarming

69 frequency. Second, that the Acanthamoeba found in the environment can be genotypically related to those found in human diseases and that genotype T4 is overwhelmingly the most common genotype isolated. Finally, that the T4 genotype is exceptionally diverse, which may lend to its ability to adapt to numerous environments and be the cause of its consistent isolation. The diversity of T4 will be discussed in

Chapter 5. These surveys demonstrate that humans are probably in consistent contact with Acanthamoeba but only perfect conditions will actually result in infection.

CHAPTER 4

ACANTHAMOEBA FROM CHICAGO

ASSOCIATED WITH THE PRESENCE OF THE

PATHOGENIC BACTERIA LEGIONELLA

PNEUMOPHILA AND PSEUDOMONAS

AERUGINOSA

INTRODUCTION

A dramatic increase of AK in conjunction with the discovery that Acanthamoeba can harbor pathogenic bacteria as potential endosymbionts has heightened public health concerns about this protist. Acanthamoeba may act as a “trojan horse” for many different types of bacteria including Legionella, Chlamydia, Mycobacteria and Pseudomonas

(Siddiqui, 2012). In Acanthamoeba, these bacteria can multiply and be released into the environment, facilitating transmission to humans. Acanthamoeba is thus a unique microbial vector, associated with disease both as a proximal agent and as a mediator of disease due to its potential to host other pathogens. It has been previously established

71 that Acanthamoeba plays a crucial role in Legionella’s life cycle allowing it to propagate in the protist and then be packaged into vesicles that when deposited outside

Acanthamoeba are the inoculums responsible for human infections (Berk et al., 1998;

Bouyer et al., 2007). In addition, both Acanthamoeba and Legionella have been shown to increase their virulence following association with the other (Fritsche et al., 1998;

Cirillo et al., 1999). Finally, environmental Acanthamoeba have been shown to harbor a plethora of intracellular bacteria giving credence to the possibility that keratitis-causing

Acanthamoeba could potentially also protect other bacteria, which could increase their capacity to cause disease (Schmitz-Esser et al., 2008; Choi et al., 2009). A recent examination of several clinical samples of Acanthamoeba found bacteria species such as

Legionella cherrii and Pseudomonas aeruginosa (Iovieno et al., 2010). In light of the recently mandated changes in water treatment (EPA, 1998), it is likely that there is an overall increase in the bacterial concentration in the water system, which could increase the opportunity for Acanthamoeba and pathogenic bacteria to interact. Together, these observations led us to question whether recent Chicago Acanthamoeba keratitis isolates from an ongoing AK outbreak harbored pathogenic bacteria, and if that relationship could exacerbate Acanthamoeba keratitis. We hypothesized that Acanthamoeba from Chicago will be associated with a high level of intracellular bacteria, which may contribute to the

Acanthamoeba infections.

METHODS

From Chicago, keratitis-causing Acanthamoeba isolates isolated from clinical cases in 2005-2010 and Tap water Acanthamoeba collected between 2006-2008 were 72 used in this study (Table 7). Acanthamoeba from keratitis cases outside of Chicago that were available in the OSU isolate collection were also examined (Table 8).

Acanthamoeba was cultured on nonnutrient agar plates seeded with Escherichia coli. Genomic DNA was extracted as described in the Appendix. Samples were tested for both Legionella and Pseudomonas using PCR with genus specific 16S rDNA primers

(Appendix Table 1). PCR reactions were performed using either OneTaq or PhusianTaq

(Appendix). For each polymerase, 25µl PCR reactions were performed as described in the Appendix. Positive and negative controls were included in each amplification protocol. 4-6µl of each PCR reaction was screened for successful amplification by gel electrophoresis. The PCR reactions from samples yielding the appropriate band of

~500bp were purified using a PEG extraction as described in the Appendix. Species of

Legionella and Pseudomonas were determined from positive samples that were successfully sequenced. Sequences of the 16S rRNA gene for each sample were aligned in MEGA5 (Tamura et al., 2011) and phylogenetic relationships were inferred using the neighbor-joining algorithm. Bootstrap values (1000 replicates) greater than 50% are listed beside the nodes.

The long-term viability of Legionella and Pseudomonas due to their association with Acanthamoeba was examined by re-growing stock Acanthamoeba cultures that had been previously shown to be positive for bacteria. Cultures of 13 Acanthamoeba isolates from Chicago were examined (Table 9). The Acanthamoeba isolates had been stored in original patient stock vials as cysts for up to seven years. For all samples, cysts were placed on non-nutrient agar plates with E. coli and grown to 105 cells (Appendix). Whole

73 sample DNA was extracted and tested for the presence of Legionella and Pseudomonas as described above.

Finally, to confirm that the bacteria were intracellular in the Acanthamoeba, in situ hybridization was performed on intact Acanthamoeba. Cysts and trophozoites were fixed to slides using 4% paraformaldyhyde at room temperature, and then incubated in hybridization buffer containing 100ng of a genus-specific 16S rDNA probe for 1 hour at

45 °C (Appendix Table 1). Slides were washed and examined for fluorescence. A scrambled probe with 6FAM was used as a control.

Table 7. Acanthamoeba keratitis and tap water isolates from Chicago used in the endosymbiont study with 18S rDNA genotypes. Samples that were positive for bacteria are indicated. Where applicable, species identification of bacterial endosymbiont has been listed (Acanthamoeba keratitis: AK; Tap Water: TW) Chicago AK Acanthamoeba OSU ID Source Genotype Legionella Species Pseudomonas Species 05-003 AK T4 Legionella spp. Pseudomonas aeruginosa 05-009 AK T4 ― Pseudomonas aeruginosa 05-011 AK T4 ― ― 05-012 AK T4 ― ― 05-013 AK T4 ― ― 05-014 AK T3 ― ― 05-019 AK T3 ― ― 05-020 AK T3 ― ― 05-021 AK T3 ― ― 05-023 AK T4 Legionella pneumophila Pseudomonas putida 06-001 AK T4 ― ― 06-002 AK T4 ― Pseudomonas putida 06-004 AK T4 ― ― 06-005 AK T4 ― ― 06-016 AK T4 Legionella pneumophila Pseudomonas putida 06-024 AK T4 ― ― 06-025 AK T4 Legionella pneumophila Pseudomonas aeruginosa 06-027 AK T4 ― Pseudomonas aeruginosa 06-033 AK T4 Legionella pneumophila Pseudomonas aeruginosa 06-034 AK T4 ― ― 06-035 AK T4 ― ― 06-049 AK T4 Legionella spp. Pseudomonas putida 06-050 AK T4 Legionella spp. ― 06-051 AK T4 Legionella spp. Pseudomonas putida 06-052 AK T4 Legionella spp. Pseudomonas aeruginosa 06-053 AK T4 ― ― 06-057 AK T4 Legionella spp. ― 06-059 AK T4 ― ― 06-061 AK T4 Legionella spp. Pseudomonas aeruginosa 06-069 AK T4 Legionella spp. Pseudomonas putida 06-071 AK T4 ― Pseudomonas aeruginosa 06-073 AK T4 Legionella spp. Pseudomonas putida 07-047 AK T4 ― ― 07-070 AK T4 ― ― Continued 75

Table 7 continued 07-072 AK T4 ― ― 07-073 AK T4 Legionella pneumophila Pseudomonas aeruginosa 07-075 AK T4 Legionella pneumophila Pseudomonas spp. 07-076 AK T4 ― ― 07-095 AK T4 ― ― 08-004 AK T4 ― ― 08-005 AK T4 Legionella spp. Pseudomonas aeruginosa 08-007 AK T4 ― ― 08-008 AK T4 Legionella spp. ― Legionella pneumophila 08-014 pneumophila & Legionella Pseudomonas putida & AK T4 pneumphila parceusis Pseudomonas aeruginosa 08-017 AK T4 ― Pseudomonas putida 10-001 AK T4 ― ― 10-002 AK T4 Legionella pneumophila Pseudomonas aeruginosa 10-003 AK T4 ― ― 10-004 AK T4 ― ― 10-005 AK T4 Legionella spp. ―

Tap Water Acanthamoeba

OSU ID # Source Genotype Legionella Pseudomonas W06-003 TW T4 ― Pseudomonas aeruginosa W06-054 TW Legionella pneumophila Pseudomonas aeruginosa W06-062 TW T4 ― Pseudomonas stutzeri W06-064 TW T4 ― Pseudomonas aeruginosa W06-072 TW Legionella spp. Pseudomonas aeruginosa W06-084 TW Legionella pneumophila Pseudomonas aeruginosa W06-087 TW T4 ― ― W06-100 TW T4 ― Pseudomonas aeruginosa W06-101 TW T4 ― Pseudomonas aeruginosa W06-103 TW Legionella spp. P. plecoglossicida W06-129 TW T4 ― Pseudomonas aeruginosa W06-131 TW T3 Legionella pneumophila ― W06-140 TW T4 ― ― W06-145 TW T4 ― Pseudomonas aeruginosa W06-163 TW Legionella spp. P. putida W06-168 TW T4 ― ― W06-174 TW T4 ― ― Continued 76

Table 7 continued

W06-194 TW T4 Legionella spp. P. putida W06-201 TW T4 ― ― W06-202 TW T4 ― P. plecoglossicida W06-207 TW T4 ― ― W06-212 TW T4 ― ― W06-216 TW T4 ― P. putida W06-221 TW T4 Legionella pneumophila ― W06-222 TW T3 Legionella pneumophila P. putida W06-229 TW T4 ― Pseudomonas aeruginosa W06-253 TW ― ― W06-268 TW T4 ― Pseudomonas toyotomiensis W06-269 TW T3 ― Pseudomonas aeruginosa W07-003 TW L. drozanskii Pseudomonas aeruginosa W07-006 TW T4 ― Pseudomonas aeruginosa W07-015 TW T4 ― ― W07-024 TW T4 ― P. plecoglossicida W07-036 TW T4 ― Pseudomonas aeruginosa W07-039 TW T4 Legionella pneumophila Pseudomonas aeruginosa W07-065 TW ― ― W07-070 TW T4 ― Pseudomonas aeruginosa W07-073 TW T4 ― ― W07-087 TW Legionella spp. Pseudomonas aeruginosa W07-131 TW T4 ― ― W07-133 TW ― Pseudomonas putida W07-135 TW T4 ― Pseudomonas aeruginosa W08-007 TW ― Pseudomonas aeruginosa W08-009 TW ― Pseudomonas aeruginosa W08-021 TW ― Pseudomonas aeruginosa W08-029 TW T4 Legionella spp. Pseudomonas aeruginosa W08-035 TW Legionella pneumophila Pseudomonas aeruginosa W08-047 TW Legionella pneumophila Pseudomonas aeruginosa

Table 8. Acanthamoeba keratitis isolates from outside Chicago used in the endosymbiont study with 18S rDNA genotypes. Samples that were positive for bacteria are indicated. Where applicable, species identification of bacterial endosymbiont has been listed (Acanthamoeba keratitis: AK)

Non-Chicago Acanthamoeba OSU ID Source Genotype Legionella Species Pseudomonas Species 03-012 AK T4 ― ― 03-018 AK T4 ― ― 03-039 AK T4 ― ― 07-003 AK T4 ― ― 07-007 AK T4 ― ― 07-016 AK T4 ― ― 07-018 AK T4 ― ― 07-024 AK T4 ― ― 07-025 AK T1 ― ― 07-026 AK T4 ― ― 07-027 AK T4 ― ― 07-028 AK T4 ― ― 07-029 AK T10 ― Pseudomonas putida 07-031 AK T4 ― ― 07-032 AK T4 ― ― 07-091 AK T4 Legionella pneumophila ― 07-103 AK T4 ― ― 10-025 AK T4 ― Pseudomonas viridilflava 1629/99 AK T4 ― ― A. quina AK T4 ― ―

RESULTS

A total of 50 isolates of keratitis-causing Acanthamoeba from Chicago were tested for the presence of Legionella and Pseudomonas. Of the 50 samples, 26 (32%) tested positive for both Pseudomonas and Legionella, 20 (40%) tested positive for

Legionella and 21 (42%) tested positive for Pseudomonas (Table 7; Figure 17). For comparison, 20 Acanthamoeba associated with keratitis from outside Chicago were tested for these bacteria. None were positive for both bacteria, only 1 (5%) was positive for

Legionella, and 2 (10%) were positive for Pseudomonas (Table 8, Figure 18). A 2- sample Z test was used to test the significance of differences between the two samples.

This test showed a significantly higher prevalence of either bacterial genera in Chicago

Acanthamoeba compared to Acanthamoeba from outside of Chicago (p<0.001) (Figure

18).

To determine whether intracellular bacteria was a feature only of clinical

Acanthamoeba from Chicago, tap water Acanthamoeba as described in Chapter 3 were tested for the presence of Legionella and Pseudomonas. 48 tap water Acanthamoeba were tested for intracellular bacteria, 15 (31%) were positive for Legionella, 34 (71%) were positive for Pseudomonas and 12 (25%) were positive for both (Table 7; Figure 17).

A 2-sample Z test was used to test the significance of differences between the Chicago

AK and Tap water isolates. There was no statistical difference between the prevalence of

Legionella in these samples, however there was a statistically higher prevalence of

Pseudomonas in the tap water Acanthamoeba compared to the keratitis-causing

Acanthamoeba (p<0.001). 79

Figure 17. Percentage of Acanthamoeba from Chicago that were positive for Legionella, Pseudomonas or both. All comparisons are statistically significant (p<0.001) using a 2- sample Z test.

Figure 18. Percentage of Acanthamoeba from Chicago and Non-Chicago sources that were positive for Legionella, Pseudomonas or both. All comparisons are statistically significant (p<0.001) using a 2-sample Z test.

Analysis of the 16S rRNA bacterial sequences indicated the presence of

Legionella pneumophila and Pseudomonas aeruginosa as the most prevalent forms associated with Acanthamoeba from all sources. While Legionella pneumophila was overwhelmingly the predominant species identified in Acanthamoeba, Pseudomonas aeruginosa was found in 63% of samples and Pseudomonas putida was found in 26% of samples (Figure 19). In addition, sequences from other species of these genera were identified (Table 7, 8; Figure 19). One Acanthamoeba isolate contained two distinct

Legionella species, L. pneumophila pnuemophila and L. pneumophila parceusis and two distinct Pseudomonas species, P. putida and P. aeruginosa (Tables 7).

Figure 19. Prevalence of Legionella and Pseudomonas by species isolated from all Acanthamoeba isolates tested.

In situ hybridization confirmed the presence of both Legionella and Pseudomonas intracellularly in the Acanthamoeba (Figure 20). Legionella occurred intracellularly in both cysts and trophozoites of Acanthamoeba. In trophozoites, Legionella are found in the vacuoles in the cytoplasm. Legionella were also packaged into vesicles and exocytosed from the Acanthamoeba as the amoeba encysted (not shown). The results indicate that Legionella found in cultures of Acanthamoeba are replicating intracellularly.

Pseudomonas was observed by FISH in Acanthamoeba cysts between the two walls of

Acanthamoeba (Figure 20). Pseudomonas was concentrated at the ostioles, which are discontinuous portions of the Acanthamoeba cysts inner wall where the amoeba exits during excystment.

By examining the occurrence of active bacteria obtained from isolates of

Acanthamoeba that had been stored as cysts for various time periods and then recultured, the long term viability of the bacteria with Acanthamoeba cysts can be evaluated. Long- term viability studies found that Legionella remained present in Acanthamoeba and was viable after Acanthamoeba excystment for up to seven years (Table 9). In contrast viable

Pseudomonas was only obtained from Acanthamoeba isolates that had been stored as cysts for three years or less after original isolation (Table 9).

A B Figure 20 A. In situ hybridization of Acanthamoeba with Psuedomonas specific probe (red). B. In situ hybridization of Acanthamoeba with Legionella specific probe (green).

DISCUSSION

The dramatic increase in Acanthamoeba keratitis cases around metropolitan

United States has raised serious questions about the potential health risk of

Acanthamoeba. Recently, an Acanthamoeba keratitis outbreak investigation was launched by the CDC, showing the importance of understanding more about the potential risks associated with Acanthamoeba (Verani et al., 2009). This results presented in this chapter show that not only are Acanthamoeba public health threats in their own right, but in addition they have the potential to contribute to the spread of other pathogens.

Acanthamoeba is a host to many bacterial pathogens including Legionella pneumophila

(Rowbotham, 1980), Pseudomonas aeruginosa (Michel et al., 1995), Vibrio chlolerae

(Abd et al., 2007), Mycobacterium (Sandstrom et al., 2010) and Shigella (Sandstrom et al., 2010).

The nature of the relationship between Acanthamoeba and its intracellular parasites has not been clearly established. On the one hand, Acanthamoeba can provide intracellular bacteria protection from numerous antibiotic drugs to which they would otherwise be susceptible. This is due to Acanthamoeba’s ability to encyst into a double- walled dormant cell in the presence of drug therapies. Numerous bacteria are capable of surviving Acanthamoeba’s reversible differentiation including Protochlamydia (Kennedy et al., 2012), Shigella (Sandstrom et al., 2010) and V. cholerae (Saeed et al., 2009).

However, many bacteria are not only able to resist digestion by Acanthameoba but also multiply intracellularly in the amoeba as well.

Legionella is able to grow in a multitude of amoebae in a temperature dependent nature (Nakamura et al., 2010). In Acanthamoeba, Legionella can cause cytotoxicty in

87 the amoeba at higher temperatures through the activation of certain virulence factors

(Abd et al., 2007). However, Acanthamoeba can escape Legionella’s cytotoxicity at high temperatures by encysting. Prior to encysting, the Acanthamoeba releases vesicles containing live Legionella (Buse & Ashbolt, 2011). Acanthamoeba can also endocytose such vesicles after excystment to reintroduce Legionella to their cytoplasm, continuing the cycle. Acanthamoeba, or some other amoebae, represent a necessary step in

Legionella pathogenesis, where the vesicles released by Acanthamoeba can be inhaled by mammalian hosts (Ohno et al., 2008). Additionally, Acanthamoeba has been found to increase the virulence of Legionella following passage through the protozoa, when compared to Legionella cultured alone (Cirillo et al., 1999). Finally, previously nonculturable Legionella can be resuscitated after passage through Acanthamoeba (Berk et al., 1998). Virulence of Legionella pneumophila is maintained after resuscitation in

Acanthamoeba (Berk et al., 1998). Together, the relationship between Acanthamoeba and Legionella has the potential to be beneficial to both organisms, where Acanthamoeba provides protection to Legionella, and Legionella increases Acanthamoeba’s virulence, potentially by stimulating its metabolism.

The relationship between Acanthamoeba and Legionella is not unique. Other bacteria are capable of interacting with Acanthamoeba in an endosymbiotic relationship.

V. cholerae and Shigella can multiply in vacuoles localized in the cytoplasm of

Acanthamoeba (Saeed et al., 2009; Sandstrom et al., 2010). Like Legionella, Shigella is able to kill Acanthamoeba at 37ºC because of virulence factors, which activate at higher temperatures (Sandstrom et al., 2010). It is interesting to note that these bacterial virulence factors activate at temperatures consistent with being in a larger mammalian

88 host. This lends credence to the idea that bacteria could potentially contribute to the progression of Acanthamoeba keratitis by being released into the cornea by the amoeba.

At higher temperatures, the activated virulence factors of the bacteria could force

Acanthamoeba encyst, releasing bacteria into the stroma, which could enhance immune response of the mammalian host. This could lead to more severe loss of visual acuity following infection with Acanthamoeba carrying pathogenic bacteria as supported by previous studies (Iovieno et al., 2010).

This study is the first to establish that Legionella pneumophila is found in clinically relevant Acanthamoeba isolates. However, numerous other studies have found other potential endosymbionts associated with keratitis-causing Acanthamoeba including a patient with intracellular Chlamydia and Legionella, as well as clinical Acanthamoeba containing various Proteobacteria (Stone & Abu, 1998; Iovieno et al., 2010).

Additionally, Acanthamoeba isolated from a contact lens case was found to harbor

Mycobacterium (Steinert et al., 1997). Environmental Acanthamoeba can contain more than one bacterial species making Acanthamoeba a potential vector for numerous pathogens simultaneously (Xuan et al., 2007).

The clinical relevance of these potential endosymbionts beyond their own capacity to cause disease is important. Acanthamoeba’s ability to cause disease is also increased by the presence of endosymbionts (Fritsche et al., 1998). A previous study documented a trend of greater impairment of visual acuity in patients with Acanthamoeba containing endosymbionts (Iovieno et al., 2010). Some Acanthamoeba keratitis patients also have infectious crystalline keratopathy (ICK) and historically have the poorest outcomes of both diseases (Yu et al., 2007). Five Chicago AK patients showed bacterial

89 keratitis as an additional secondary opportunistic infection. Though the sources of these infections were not determined for every patient, Streptococcus oralis was found in two of the patients (Yu et al., 2007). It is possible that S. oralis was in Acanthamoeba, which was then released when the Acanthamoeba lysed following effective treatment. Location of the infection in the deep stromal area suggests this may be the result of internalized bacteria that were protected from other antimicrobials and carried by Acanthamoeba into the stroma (Yu et al., 2007). In a South African study, a keratitis patient was found to have both Acanthamoeba and Pseudomonas keratitis showing that dual Acanthamoeba- bacteria cornea infections are possible (Heinz et al., 2007).

In this study, 40% of keratitis-causing Acanthamoeba isolated from Chicago were positive for Legionella. Differences in relative prevalence of Legionella found in

Acanthamoeba could be due to the source of isolates. Tap water Acanthamoeba were positive for Legionella in 32% of amoeba tested. Manmade water systems often contain warmer water temperatures compared to natural freshwater samples, which could explain the discrepancy in Acanthamoeba harboring Legionella. Legionella has a higher prevalence in warmer water temperatures versus colder (Dini et al., 2000; Abd et al.,

2007; Tu et al., 2009). Another study of both environmental and clinical samples showed that nearly 60% of Acanthamoeba isolates contained at least one potential endosymbiont

(Iovieno et al., 2010). Previously, this study had reported the highest level of intracellular bacteria, as well as being the first report to examine bacteria from clinically relevant Acanthamoeba. It is interesting that Iovieno et al., 2010 only identified one species of Legionella in their collection of clinically relevant Acanthamoeba. These discrepancies between the Chicago data and Iovieno et al., 2010 could be due to the

90 current screening using a more sensitive test. Alternately, Iovieno found a high level of

Mycobacterium in the Acanthamoeba isolates. The Chicago Acanthamoeba have yet to be tested for this bacteria and it is possible that Acanthamoeba has a limit for the number or type of intracellular bacteria that could skew the results toward a particular bacteria. It is equally likely that the Chicago Acanthamoeba will have a high level of Mycobacterium as well which may be contributing to the amoeba's virulence. At the same time, the potential role of endosymbionts in Acanthamoeba pathogenesis is a new theory so the prevalance of bacteria in historical Acanthamoeba samples are impossible to predict.

Since this study is one of the first to examine clinical Acanthamoeba for Legionella and

Pseudomonas specifically as intracellular bacteria, it is not a surprise that the results are startling.

The previous data shows a statistically higher prevalence of pathogenic bacteria in the Chicago Acanthamoeba compared to that found in a random sampling of

Acanthamoeba isolated from other clinical and geographical settings. In addition, our results show a high prevalence of bacteria associated tap water Acanthamoeba from the same geographical area. While the relative prevalence of Legionella between Chicago tap water and keratitis-causing Acanthamoeba was not significant, the prevalence of

Pseudomonas was significantly higher in tap water amoeba. Our regrowth study suggests that the relationship between Pseudomonas and Acanthamoeba is more transient than with Legionella. All keratitis-causing Acanthamoeba would have been subjected to numerous drug therapies prior to isolation from the cornea. If the relationship between

Pseudomonas and Acanthamoeba was less stable, it is possible that these drug treatment resulted in the clearing of Pseudomonas from AK isolates, as a result of encystment.

Either way, all Chicago Acanthamoeba showed a high prevalence of the bacteria tested. suggesting a potential role for pathogenic bacteria in the virulence of Acanthamoeba.

Together, the data presented in this chapter suggest two alternate hypotheses: The increase in AK cases during the last decade in Chicago could be a result of

Acanthamoeba containing pathogenic bacteria that increase the virulence of the

Acanthamoeba itself or could result from a process in which the pathogenic bacteria themselves exacerbate an existing Acanthamoeba infections by increasing innate immune responses in the cornea. The role of intracellular pathogenic bacteria in Acanthamoeba requires further investigation before any definitive conclusions can be drawn.

CHAPTER 5

MULTILOCUS SEQUENCE TYPING OF

ACANTHAMOEBA KERATITIS-ASSOCIATED

CLINICAL ISOLATES FROM A CHICAGO

OUTBREAK

INTRODUCTION

T4 is the most commonly found genotype of environmental and clinical

Acanthamoeba isolates (Booton et al., 2005). It is unknown whether T4 is naturally more pathogenic or just more prevalent. The phylogenetic relationship between the

Acanthamoeba samples that are found in clinical AK, especially those cases involved in an AK outbreak, has not been clearly established. Unlike some Acanthamoeba genotypes in which different isolates are nearly identical, T4 is very heterogenous, and it is unknown some sub-lineages are more common, or more prone to causing disease.

For many isolates and genotypes of Acanthamoeba, the nuclear 18S rRNA or mitochondrial COI genes are the only sequences available for genetic identification.

Over the past 15+ years, ribosomal DNA has usually provided the only phylogenetic

93 identification. Genotypes defined by rDNA may resolve differently when more conserved genes, such as housekeeping genes are compared to rDNA genes that harbor more variability and are under different selective pressures. By examining a variety of genes that code for proteins required in basic cellular functions, relationships between genotypes can be reconstructed, and this may provide new insights into how the genotypes evolved from a single common ancestor into multiple strains with varying levels of pathogenicity.

Multilocus sequence typing (MLST) is an approach often employed to study bacterial strains associated with an outbreak, and is a valuable tool to examine the evolution of a pathogen and identify clonal complexes. MLST was first described in a study of Neisseria meningitis using sequences of 6-8 housekeeping genes (Maiden et al.,

1998). Each unique sequence is assigned an allelic number with each nucleotide difference generating a new allelic number for that specific gene. Each isolate will have a specific allelic designation for each gene. Distinct alleles form a haplotype or sequence type of the specific isolate. Housekeeping genes are used in MLST because all isolates of a species are likely to have these genes due to their essential nature in metabolism. The analysis involves the analysis of coding regions of housekeeping genes to establish sequence types. This permits the identification of clonal complexes within an outbreak based on sequence types, which can be used to draw conclusions about the lineage of an organism and its relationship with other disease-causing organisms within the species or genus (Maiden et al., 1998). Epidemic clones generally describe bacterial lineages that are involved in outbreaks, which are considered very closely related and descended from a common ancestor.

Any phylogenetic reconstruction method that assigns isolates to “types” must provide enough information to distinguish one “type” from another but not be so discriminating that every isolate is unique (Turner et al., 2007). Clonal complexes are created based on distinct clusters of closely related isolates in a population. Isolates exhibiting similar or identical MLST genotypes are very closely related. A clonal complex as defined by MLST is a group of isolates with a common ancestor that evolved clonally where recombination is not present (Fraser et al., 2007).

MLST has been applied to all major human bacterial pathogens but rarely to protists (Glaberman et al., 2001; Waki et al., 2007). However, MLST has recently been applied to the protozoa Cryptospordium, the causative agent of enteric diseases worldwide (Glaberman et al., 2001). More importantly, there have been several recent studies that utilized MLST for diploid organisms, specifically the diploid pathogen protist

Trypanosoma (Prowell et al., 2004; Yeo et al., 2011). With housekeeping genes, there are relatively few heterozygous sites where two alleles are clearly present at a single nucleotide site because such changes could result in a nonsynonymous amino acid change. Changes in these proteins would affect fundamental metabolic processes and thus would generally be selected against by evolution. For the purpose of MLST, these heterozygous alleles are removed from the final dataset (Yeo et al., 2011).

Our previous studies have established that high levels of sequence variability exist between and within the genotypes of the genus Acanthamoeba (Gast et al., 1996; Stothard et al., 1998; Schroeder et al., 2001; Booton et al., 2009). These genotypes were initially designated as those groups of isolates that showed less than six percent divergence in their respective full-length nuclear 18S rDNA gene sequences. Subsequent, and ongoing,

95 studies on additional nuclear and mitochondrial genes have generally supported the phylogenetic structure established in the initial 18S rDNA studies. Based upon this phylogenetic framework and advances in technology that now permits rapid accumulation of sequence data, this study examined the utility of MLST to establish the levels of genetic variability and structure in a number of previously identified

Acanthamoeba nuclear genes in the widely distributed Acanthamoeba genotype T4 from

Chicago Illinois.

METHODS

Selection of suitable gene loci

Acanthamoeba genes used for multilocus sequence typing were chosen from a number of previously sequenced nuclear genes available on Genbank. Only genes involved in basic cellular functions were considered. Wherever possible, genes used in standard bacterial MLST analysis (G3PD, Elongation Factor 1 alpha) were used.

Originally eight genes were selected as potential MLST candidates. Five housekeeping genes were selected based on functional primers and the placement of introns (Appendix

Table 1). Primers were designed based on sequence data from, but not limited to the following sources: NCBI Genbank, Acanthamoeba EST database, Taxomically Broad

EST Database, and the Acanthamoeba Neff Genome Project. Wherever possible, sequences from multiple genotypes were used to develop primers.

Acanthamoeba isolates

Representative Acanthamoeba isolates from our studies in Chicago were used, and were supplemented by some isolates that 18S rRNA sequences belonged to known

96 genotypes that were underrepresented in the Chicago samples (Table 10). All of the

Chicago isolates except one (T3) belong to genotype T4. Only isolates that had been previously identified using the 18S ribosomal RNA gene were used.

DNA extraction, PCR amplification and Sequencing

Acanthamoeba isolates were cultured and genomic DNA was extracted as discussed in the Appendix. PCR was performed using gene-specific annealing temperatures and elongation steps adjusted for amplimer length (Appendix Table 1).

Positive samples were determined using gel electrophoresis and PCR amplimers were purified as described in the appendix. Housekeeping Genes: Multilocus Sequence Analysis

Forward and reverse DNA sequences were aligned using MUSCLE and analyzed using MEGA 5.0 (Tamura et al., 2011). Polymorphic nucleotide sites were identified in alignments for each loci as well as for alignments that including genotypes outside of T4.

Sequence variability was analyzed at the DNA level (including identification of exon and intron regions). All heterozygous sites were examined using the chromatographs to verify sequence. For each heterozygous site, it was determined that all Acanthamoeba isolates had two alleles at that site and the variable site was replaced as the same nucleotide in all isolates. Multilocus sequence types were analyzed by concatenating all gene loci and subsequently analyzing the multilocus sequences for nucleotide diversity, gene diversity and phylogenetic structure. Single nucleotide polymorphisms were used to identify subtypes. Phylogeny was inferred by constructing neighbor-joining trees.

For each unique allelic variant at a particular locus, an arbitrary number was assigned. Numbers for all five loci were combined to create a sequence type. Structural

97 analysis of MLST was performed using the eBURST algorithm, which allows the analysis and identification of the ancestral strains that give rise to a clonal cluster.

MLST combined Analysis: Housekeeping, COI and 18S rRNA genes

For 22 isolates, including isolates assigned by 18S rRNA typing to genotypes T2,

T3, T4 and T18, the sequences from the 5 housekeeping genes, the COI gene and the 18S rRNA gene were used to create a total data neighbor-joining tree as previously described.

Several subclades within the T4 genotype have previously been proposed based on the sequence of the 18S rRNA genes (represented by A-F) (Fuerst, unpublished data). The

19 T4 Acanthamoeba isolates used were assigned subclade designations based on the sequence of Diagnostic Fragment 3 (Booton et al., 2002) that had been previously used to identify these Acanthamoeba as belonging to genotype T4. Full-length 18S rRNA sequence data for the T4 genotypes were not collected, and the sequences of the DF3 fragment differ in size, depending on the quality of the sequencing read. To compensate for this, representative full-length 18S gene sequences from Acanthamoeba isolates that were identical (or nearly so) at the Diagnostic fragment (DF3) to the specific isolate were used in this analysis. Additionally, representative Clade C sequences were used for isolates 06-002 and 10-005 at the COI locus, since no COI sequence had been collected for these isolates. Finally, the isolate 04-020, newly classified as the first clear representative of a new genotype, designated here T18, failed to amplify at the RASC locus, so it was only analyzed at 6 loci. These seven genes were then used to examine the phylogenetic relationship between the Acanthamoeba isolates, with specific emphasis on the strength of support for the T4 subclades.

Table 10.Acanthamoeba isolates used in the MLST study, including source and genotype.

OSU ID# Origin Source Genotype 03-001 Neff ENV T4 03-023 CDC GAE T10 04-020 CDC GAE T18 05-003 UIC AK T4 05-009 UIC AK T4 05-011 UIC AK T4 05-013 UIC AK T4 05-020 UIC AK T3 05-021 UIC AK T3 05-022 OSU ENV T17 05-023 UIC AK T4 06-001 UIC AK T4 06-002 UIC AK T4 06-004 UIC AK T4 06-005 UIC AK T4 06-024 UIC AK T4 06-025 UIC AK T4 06-033 UIC AK T4 06-034 UIC AK T4 06-035 UIC AK T4 06-039 UIC ENV T4 06-047 UIC ENV T4 06-049 UIC AK T4 06-051 UIC AK T4 06-057 UIC AK T4 06-059 UIC AK T4 06-061 UIC AK T4 06-069 UIC AK T4 06-071 UIC AK T4 06-073 UIC AK T4 07-020 CDC GAE T10 07-027 CDC AK T4 07-029 CDC GAE T10 07-047 UIC AK T4 07-069 UIC AK T4 07-070 UIC AK T4 07-072 UIC AK T4 07-075 UIC AK T4 07-076 UIC AK T4 07-095 UIC AK T4 08-004 UIC AK T4 08-005 UIC AK T4 08-007 UIC AK T4 08-008 UIC AK T4 09-006 OSU ENV T2/6 10-001 UIC AK T4 10-003 UIC AK T4 10-005 UIC AK T4 MEX-FA7 Mexico ENV T5 V006 CDC GAE T1

RESULTS

Candidate gene determination

Eight potential candidates genes were selected for use in a multilocus sequence analysis. These including Elongation Factor 1 alpha (ELF1), Glyceraldehyde-3- phosphate dehydrogenase (G3PD), Corinin (COR), Beta Tubulin (BETA), Glycogen

Phosphorylase (GLYP), Laminin Binding Protein (LBP), Nicotinamide adenine dinucleotide (NADH), and RASC. Primers were developed based on EST sequences from various sources including Genbank and TBESTDB. Wherever possible, primers were developed to function in multiple genotypes.

Five genes were successfully amplified using the primers developed in this study, and could be sequenced at a quality level necessary to complete the analysis: BETA,

ELF1, G3PD, GLYP, and RASC. All primers were capable of functioning in all genotype T4. Once their ability to amplify in T4 was established, amplification was attempted in neighboring genotypes for all primers in the following order: T3, T2, T18,

T1, T10 and T17.

Beta Tubulin (BETA) is used in the formation of microtubules. Elongation Factor

1 alpha (ELF1) is responsible for delivery of aminoacyl tRNAs to the ribosome.

Glycerol-3-Phosphate Dehydrogenase (G3PD) catalyzes the sixth step in glycolysis.

Glycogen phosphorylase (GLYP ) breaks down glycogen into free glucose for use in building the cyst wall. RasC is a putative gene in Acanthamoeba, potentially involved in quorum sensing. Together, these genes represent a variety of nuclear genes that have basic cellular functions in metabolism, motility and encystment.

100

For each gene, exons and introns were identified between the two primers.

Putative splice sites were determined (GTACG) and introns removed from sequence alignments used for the MLST analyses (Figure 21). All introns were easily identified using electropherograms due to the development of diploid sites for the length of the introns. Variability of each locus was determined by identifying sites with two or more nucleotide states in the collection of isolates. While isolates that belong to genotype T4 are the most likely genotype to be involved in an outbreak, the ability of the primers

(Appendix) developed here to amplify the homologous gene in other genotypes was examined. All primers were functional in genotypes T2, T3 and T4. Additionally, with the exception of RasC, all genes were also amplified in the T18 genotype, which resolves outside the T2/6-T3-T4 cluster. Isolates for genotype T1 and T10 were successfully amplified and sequenced for ELF1, G3PD, GLYP. Isolates from none of the remaining genotypes were successfully amplified for at more than two of the five MLST loci. The success of amplification of the five genes in various genotypes can be seen in Table 11.

101

Figure 21. Gene structure between primers of the five loci chosen for MLST analysis. Red line indicates sequences used in MLST analysis.

102

Beta Tubulin

Beta Tubulin is one of the only protein coding genes in Acanthamoeba that has previously been studied in more than one isolate (Henriquez et al., 2008). This is largely due to use of Acanthamoeba ability as a model system for macrophages and for research on the underlying mechanisms involved in Acanthamoeba motility. Sequencing primers functioned in T2, T3, T4 and T18 (Table 11), but not in other types. Phylogenetic relationships seen at the BETA locus are largely similar to that seen in the 18S rRNA and

COI genes (Figure 22). 20 haplotypes involving 23 variable sites were identified within the 35 T4 genotype isolates, (Table 11; Figure 30). Many of the T4 subclades established by the 18S rRNA gene are maintained at the Beta Tubulin loci. Interestingly, the T4 genotype had a variability of 4.8% at the Beta Tubulin locus, which is similar to that level seen in the 18S rDNA. It may suggest that a certain level of variability must exist in a gene for the separation of clades in the genotype T4 to appear. Unlike for the case of 18S rDNA, T2/6 resolves more closely to T4 than T3 (Figure 22).

103

50 07-027-A 07-076-A 08-007-A 52 08-008-A 08-005-B 06-073-B 06-051-A 06-049-A 06-024-A 08-004-B 10-003-A 06-057-A 10-001-A 07-047-A 05-003-A 06-059-A 06-005-B 07-072-B 05-009-D 55 06-033-E 06-034-E 62 07-069-E

88 07-070-E 06-004-E 74 07-095-E 05-011-A 06-039-A 84 05-023-B 54 06-001-B 07-075-A 06-025-A 06-035-A 05-013-A 06-047-A 63 06-061-A 58 06-071-A 06-069-A 03-001-Neff 06-002-C 10-005-C 04-020-T18 09-006-T2 05-020-T3 100 05-021-T3

0.005 Figure 22. Neighbor-joining tree constructed using exon sequence from Beta Tubulin data across four genotypes. Bootstrap values (1000 replicates) greater than 50% are shown next to nodes. All isolates are T4 unless otherwise indicated.

104

105

Elongation Factor 1 alpha

Elongation Factor 1 alpha primers worked in all genotypes across the genus

Acanthamoeba and overall phylogenetic relationships between the genotypes were basically consistent with the 18S rDNA relationships, even though EF1α is a highly conserved gene (Table 11; Figure 23). Of the five genes examined for MLST, EF1α was the most conserved in the T4 genotype, with only four variable sites in 507 nucleotides

(0.79%) (Table 11; Figure 30). While EF1α was highly conserved, eight distinct haplotypes were still observed in the 35 isolates from the T4 genotype (Figure 23, 30;

Table 11), but they provide little resolution among T4 subtypes, As seen at other locus,

T17 remains a well-supported outgroup of the other Acanthamoeba genotypes. T18

(representing the only group I Acanthamoeba that was examined for the MLST analysis) resolves outside of the T2-T3-T4 genotypes as observed in the analysis of 18S rDNAand the phylogenetic relationships between T1, T10 and T18 are similar to that seen in the

18S rDNA. Again, T2/6 resolves more closely to the T4 genotype than seen in the 18S rRNA gene.

106

07-072-B 07-076-A 07-047-A 07-027-A 06-073-B 06-071-A 06-069-A 06-061-A 06-059-A 06-057-A 06-051-A 06-049-A 06-039-A 06-035-A 06-034-E 06-033-E 06-025-A 06-005-B 05-011-A 05-003-A 08-004-B 08-005-B 08-007-A 08-008-A 10-001-A 10-003-A 06-002-C 06-024-A 06-047-A 10-005-C 05-013-A 05-023-B 06-001-B 07-075-A 05-009-D 95 03-001-Neff 06-004-E 93 07-069-E 07-070-E 99 07-095-E 09-006-T2 05-020-T3 100 05-021-T3 04-020-T18 53 v006-T1 03-023-T10 07-020-T10 100 07-029-T10 05-022-T17

0.02 Figure 23. Neighbor-joining tree constructed using exon sequence from Elongation Factor 1 data across seven genotypes. Bootstrap values (1000 replicates) greater than 50% are shown next to nodes.

107

G3PD

G3PD was the most variable of the five loci with 37 variable sites or 7.3% within the T4 genotype. Despite the high number of variable sites, only 22 haplotypes were seen

(Table 11; Figure 30). Again, T4 sub-clades for C, D and E were largely maintained but not for sub-clades A and B (Figure 24). The primers worked to amplify the gene in genotypes, T1, T2, T3, T4, T10, and T18, and all phylogenetic relationships among genotypes were identical to those seen using the COI gene (Table 11; Figure 6; 24). The only difference from the patterns seen with the 18S rDNA gene was that the T2/6 again was phylogenetically closer to T4 than T3 was. Overall, for a gene that is involved in such a basic cellular process, it was unexpected to see such a high level of genetic diversity at this locus.

108

06-025-A 06-069-A 06-035-A 06-071-A 06-005-B 07-072-B 07-076-A 65 05-013-A 07-075-A 05-023-B 99 06-001-B 05-003-A 06-047-A 06-057-A 06-039-A 06-049-A 07-047-A 06-059-A 06-061-A 10-003-A 06-051-A 53 07-027-A 10-001-A 08-004-B 82 08-005-B 08-007-A 68 08-008-A 03-001-Neff 05-009-D 87 05-011-A 98 06-033-E 06-034-E 07-095-E 55 06-004-E 99 06-024-A 54 07-069-E 99 70 07-070-E 06-002-C 93 97 10-005-C

99 06-073-B 09-006-T2 05-020-T3 99 05-021-T3 V006-T1 04-020-T18 03-023-T10 07-020-T10 99 07-029-T10

0.02 Figure 24. Neighbor-joining tree constructed using exon sequence from G3PD data across six genotypes. Bootstrap values (1000 replicates) greater than 50% are shown next to nodes. 109

Glycogen Phosphorylase

Glycogen Phosphorylase is a required enzyme for the process of encystment.

Since all known Acanthamoeba encyst, the presence of GLYP is expected in all

Acanthamoeba. The primers designed here worked to amplify the sequences of T1, T2,

T3, T4, T5, T10 and T18. Glycogen Phosphorylase had sixteen variable sites in the 35

T4 genotype isolates but this only resulted in 10 allelic profiles at the locus compared to

EF1α, which had 8 haplotypes from only 4 variable sites (Table 11; Figure 30). Possibly as a result of the relatively small number of haplotypes, the phylogenetic relationships between certain genotypic isolates are very different from what was observed in results from the 18S rDNA and COI genes (Figure 25). While there is very little variation seen within the T4 genotype at this locus, it is interesting to note that the T4 sub-clades C, D, and E are identifiable at this locus (Figure 25). The T5 and T10 genotypes are placed within the major clade containing T4 isolates, whereas both of these genotypes diverge outside T1-T2-T3-T4-T18 at both the COI and 18S loci. Overall, glycogen phosphorylase's very unusual phylogenetic tree suggests that Glycogen Phosphorylase may be under different selective pressures than other genes like EF1α, 18S rDNA, 16S- like rDNA, and COI.

110

07-047-A 08-007-A 07-076-A 07-027-A 06-059-A 06-027-A 06-001-B 05-003-A 05-023-B 06-024-A 06-051-A 06-073-B 07-075-A 06-071-A 08-005-B 05-013-A 06-005-B 06-049-A 06-061-A 07-072-B 08-004-B 08-008-A MEX-FA7-T5 06-047-A 10-001-A 50 66 10-003-A 05-011-A 06-025-A

61 06-035-A 06-039-A 06-057-A 06-069-A 03-023-T10 05-009-D 81 06-002-C 10-005-C 06-004-E 84 87 06-033-E 86 06-034-E 07-069-E 60 96 07-070-E 07-095-E 03-001-Neff 09-006-T2 V006-T1 04-020-T18

67 05-020-T3 99 05-021-T3

0.005 Figure 25. Neighbor-joining tree constructed using exon sequence from Glycogen Phosphorlyase data across seven genotypes. Bootstrap values (1000 replicates) greater than 50% are shown next to nodes.

111

RasC

RasC was the only putative gene from Acanthamoeba showing RasC motifs and a high level of amino acid similarity to two Ras genes in Dictyostelium (Crary, unpublished data). Unlike the other MLST genes, RasC was only able to amplify sequences from genotypes T2, T3, and T4 (Table 11). While it appeared to have some amplification in genotype T18, the sequence isolated could not be conclusively identified as the homologous Ras gene that was amplified in the other genotypes. RasC had 25 variable sites and 23 haplotypes within genotype T4(Table 11; Figure 30). Overall, it had a variability of 5.4%. While using a putative gene may be risky for use in MLST, it was less variable than G3PD, which has been a standard MLST gene used in bacterial outbreaks (Table 11). This low variability as well as RasC's presence in EST databases suggested RasC is a transcribed gene and its low variability qualifies it as a housekeeping gene and kept it in the dataset. Most T4 subclades were maintained within the T4 genotype, with the exception of sub-clades A and B. Phylogenetic relationships involving T2-T3-T4 were identical to those seen from sequences of the other housekeeping genes (Figure 26).

112

08-005-B 08-008-A 08-004-B 06-059-A 07-047-A 07-076-A 06-049-A 07-027-A 08-007-A 05-003-A 05-011-A 05-023-B 58 06-001-B 05-013A 54 06-024-A 10-001-A 60 06-005-B 07-072-B 06-039-A 06-073-B 06-051-A 06-025-A 77 06-035-A 06-069-A 06-057-A 10-003-A 03-001-Neff 05-009-D 07-075-A 72 87 06-047-A 73 06-061-A 06-071-A 06-004-E 06-033-E 99 06-034-E 06-002-C 10-005-C 64 07-095-E

100 07-069-E 90 07-070-E 09-006-T2

98 05-020-T3 100 05-021 T3

0.01 Figure 26. Neighbor-joining tree constructed using exon sequence from RasC data across seven genotypes.

113

Total Data: Housekeeping genes

Table 11 shows the number of STs (sequence types) for each gene locus as well as the relative variability of that gene. A total of 105 nucleotide polymorphisms were found across the five loci from sequences totaling 2673 nucleotides (Table 11; Figure

30). Only 21 of those changes resulted in a nonsynonymous change (Figure 30). MLST analysis was able to resolve genotype T3 and genotype T2 as significantly different from and outside of the T4 genotype, as has been seen when using full-length sequences of the

18S rRNA gene (Figure 27). However, unlike the 18S rDNA, genotype T2 was found to be branch between genotypes T4 and T3. For 18S rDNA, genotype T3 is the sister group to T4 which was also seen in the COI gene. Together, the loci had an average variability of 3.9% (Table 11). A phylogenetic tree was created for all five loci by concatenating the five sequences and running the phylogenetic analysis. The relationships between T2-T3-

T4 were identical to those seen using the individual genes. T4 sub-clade C, D and E were phylogenetically separate from each other (Figure 27). Note that what isolates classified in T4 sub-clade E based on 18S rDNA appear to form two separate phylogenetic clusters using MLST data. It is possible that the 06-033/06-034 isolates could be considered a separate sub-clade from all other sub-clades. Unlike in 18S rDNA, sub-clade B is interspersed amongst sub-clade A. As a result, the sub-structuring within T4 genotype remains unresolved.

As a test of whether MLST analysis provides any indication of specific epidemic spread, eBURST analysis was applied to the sequence types created at the five loci within the T4 genotype. This will help to establish the possibility of common ancestors or clusters within the genotype (Figure 29). A single sequence type (#28, which include

114 isolate(s) OSU-07-047) form the nexus of the relationship among isolates. While not necessarily a true common ancestor, other isolates are closely related, and some small potential clades become evident in the T4 genotype, although only one other sequence type (#19, including isolate(s) OSU-06-025) may form the origin of several other sequence types.

115

08-004-B 72 08-005-B 66 08-007-A 92 08-008-A 07-027-A 64 06-051-A 07-047-A 10-001-A 81 10-003-A 05-003-A 06-049-A 06-059-A 07-076-A 06-073-B 06-071-A

57 06-047-A 67 06-061-A 06-005-B 88 07-072-B 06-057-A 06-069-A 06-025-A 06-035-A 89 06-039-A 05-013-A

62 07-075-A 05-023-B 56 100 06-001-B 05-011-A 82 03-001-Neff 88 05-009-D 06-024-A 06-033-E 100 100 06-034-E 100 06-002-C 10-005-C 06-004-E

99 07-095-E 94 07-069-E 99 07-070-E 09-006-T2 05-020-T3 100 05-021-T3

0.005

Figure 27.. Total data from all five loci concatenated to produce a neighbor-joining tree. Genotypic outgroups T2 and T3 were used as roots. Bootstrap values (1000 replicates) which are greater than 50% are indicated next to nodes.

116

Seven gene MLST analysis

Finally, all seven genes (MLST, COX and 18S rDNA) were examined in 22 isolates which represented all the genotype T4 clades seen in the Chicago samples as well as the T2 and T3 genotypes. The addition of the 18S rDNA sequence data did not significantly change the phylogenetic relationships between T2, T3 and T4, however low bootstrap values between these genotypes prevent definitive conclusions about their overall relatedness to be drawn (Figure 28). Sub-clade C and D remain separate from other sub-clades within T4 (Figure 28). Acanthamoeba Neff may be phylogenetically separate from all other genotype T4 subclades. Sub-clade E again resolves into two subgroups though the two subgroups cluster together using all the loci, unlike what was observed using single loci or the housekeeping 5-gene analysis. The sub-structuring within the T4 genotype is clearly very complex and the existence of significant sub- clades may be clearer only with the addition of more isolates.

117

100 06-005-B 08-005-B 05-011-A 93 07-027-A 06-024-A 55 06-051-A 07-047-A 05-003-A 99 06-025-A 100 83 06-071-A 05-023-B 95 03-001-Neff 05-009-D 100 06-004-E 07-069-E 100 100 06-033-E 98 07-095-E 06-002-C 100 10-005-C 09-006-T2 05-020-T3 04-020-T18

0.01

Figure 28. Total data from all seven loci to produce a neighbor-joining tree. T18 was used as an outgroup. Bootstrap values (1000 replicates) greater than 50% are indicated next to nodes.

118

119

120

Variability between genotypes at the different loci

The 19 isolates used in the seven gene loci MLST analysis were used to calculate sequence divergence between genotypes at individual loci (Table 12). At the 18S rDNA,

T3 shows the smallest divergence (only 4.6% different) from T4, compared to the differences between T2 and T4 (8.1%). T2 is also more divergent from T4 than T3 for

COI. However, the pattern is different for housekeeping genes, which show only a 5% difference between T2 and T4, versus 6.1% difference between T3 and T4. COI showed the greatest rate of divergence, with a mean pairwise distance between the three genotypes of 14.6% (Table 12). This is consistent with the expectations for a protein coding gene in the mitochondrial genome. The nuclear housekeeping genes showed the lowest average distance between all T4 isolates at 2.5%. It is possible that the differences between the variability at each loci, as well as the variable phylogenetic relationships, are the result of different selective pressures. For example, the 18S rRNA gene is not a protein coding gene and the COI gene is in the mitochondrial genome. These could have significant effects on how isolates compare to each other when using these individual loci to determine relatedness.

121

Table 12. Pairwise distances between the 19 Acanthamoeba isolates used to analyze phylogenetic relationships between seven loci. Distances between genotypes were determined at individual loci and using total data.

18S rDNA T3 T4 T18 T2 8.30% 8.10% 12.10% T4 4.60% 2.70% ▬ T18 9.40% 9.10% ▬ COI T3 T4 T18 T2 20.00% 22.00% 23.90% T4 17.40% 12.50% ▬ T18 24.90% 22.30% ▬ Housekeeping genes T3 T4 T18 T2 6.60% 5.00% 8.20% T4 6.10% 1.70% ▬ T18 7.60% 7.70% ▬ All seven loci T3 T4 T18 T2 8.70% 8% 11.70% T4 6.80% 2.60% ▬ T18 10.40% 9.90% ▬

122

DISCUSSION

Multilocus Sequence Typing allows a highly reproducible genetic analysis of pathogens and potential pathogens. Typing of pathogenic isolates increases our understanding of the epidemiology of infections and the evolution of pathogens. The primary strength of MLST analysis is the highly reproducible nature of the data, that can be accessed on a global level, allowing researchers to compare results worldwide. While

MLST is largely used to analyze bacterial populations, generally associated with epidemic outbreaks, recently MLST has been shown to be highly resolutive for epidemiological and population structure for several fungi, such as Candida. It has also been used to investigate Fusarium, Aspergillus and Cryptosporidium.

Here, PCR primers that would allow MLST analysis in Acanthamoeba have been developed. MLST analysis is demonstrated to be a practical method to show the diversity of the genus Acanthamoeba, especially in its most prominent genotype, T4.

MLST was originally designed for analyzing haploid genomes. Heterozygosity can be clearly seen in electropherograms of the genes of Acanthamoeba by observation of a double peak at a particular nucleotide site. This causes ambiguous allelic sequences within a locus, and may result in combined alleles. However, much of the problem caused by the presence of heterozygosity can be overcome by using highly conserved genes like housekeeping genes.

Diploid sequence typing of 40 T4 sequences indicates that the five genes were highly discriminatory as 33 of the 40 resolved in a unique sequence type. Since many of these isolates come from the Chicago outbreak, our results further support the suggestion that our lab and collaborators have previously made that the outbreak is not the result of

123 epidemic spread of a new genotype (or sequence type). The results instead suggest that many genetically different isolates have the potential for causing AK.

This study found complex sub-structuring potentially exists within the

Acanthamoeba genotype T4. We propose that five genes (Appendix Table 1) be used for characterization of new isolates, especially those potentially involved in an outbreak.

These loci were all capable of distinguishing between T2, T3, and T4, which are very closely related genotypes when using standard 18S rDNA methods of genotyping. The number of sequence types for each gene locus varied greatly. Distance based cladograms correlated well with clonal clusters created by eBURST. Despite eBURST being the typical method of analyzing MLST data, until a larger number of Acanthamoeba isolates have had MLST applied to it, it is difficult to use this tool for predicting founders or clusters. While 18S rDNA data suggests the presence of up to six sub-clades within the

T4 genotype, counting the Neff clade, the MLST analysis does not provide strong support for some of the substructuring seen in the 18S rDNA results. The data reflected by this

MLST analysis would require either the possibilities of further subdivision of clades

(specifically within sub-clades A and B) or a combining of sub-clades, namely A and B.

Additionally, other potential sub-clades could exist but without additional samples, it is difficult to determine how robust these sub-clades are. Analysis of a larger number of

Acanthamoeba isolates from environmental and clinical sources as well as different geographic regions will be necessary to determine if the sequence types found in Chicago are specific to this area or if they represent most of the possible haplotypes possible within the T4 genotype. Additional isolates will also be required in order to draw firm conclusions about how the substructure in Acanthamoeba T4 relates to the overall

124 pathogenicity of this genotype, as well as shedding light on hypotheses about whether certain sub-lineages are responsible for infections whereas other lineages may only be found in environmental sources.

This study demonstrated that MLST is an excellent method to gain fine-scale resolution of the population structure of Acanthamoeba, particularly in the diverse T4 genotype. MLST confirms previous general observations based on ribosomal RNA and mitochondrial genes that indicated high levels of diversity within the T4 genotype.

MLST has the potential in the future to answer questions about the overall pathogenicity of the T4 genotype and to increase our understanding about Acanthamoeba isolated from the environment compared to those isolated from clinical infections.

125

CHAPTER 6

DISCUSSION

Acanthamoeba is a dynamic organism. For every bit of knowledge we uncover about it, a hundred new questions arise. It is difficult even to pinpoint an aspect of

Acanthamoeba biology that is the most interesting. Its ability to phenotypically switch between active and dormant phases raises many issues about environmental sensors, cyst wall composition and the metabolic burden of maintaining genes for cyst wall construction, and of course, drug resistance. Genetically, even its most basic adhesion proteins have little homology to known proteins from other organisms, while the ploidy of Acanthamoeba remains hotly debated, even in our laboratory. Finally,

Acanthamoeba’s ability to cause disease through the contribution to virulence by nongenetic factors, like endosymbionts, has added a whole new layer of complexity to an organism whose risk to public health is increasing. As a result, it is difficult to focus research on a particular aspect of this organism.

For the body of research summarized in this document, it comes down to two main points. It is essential to understand the genetic similarity of different

Acanthamoeba and how those similarities and differences affect their capacity to cause disease. Understanding the phylogenetics of Acanthamoeba will ultimately allow us to

126 correctly identify newly discovered Acanthamoeba as well as begin to predict their risk of causing disease. Additionally, it is crucial to understand the relationship between

Acanthamoeba and their potential endosymbionts, both in the endosymbionts’ ability to enhance Acanthamoeba’s virulence as well as Acanthamoeba’s role in the pathogenesis of the bacteria. The potential hazard associated with Acanthamoeba toting around these deadly pathogens is understated especially in light of Acanthamoeba’s drug resistance.

In fact, it is impossible to speculate how much of a role Acanthamoeba may have had previously as a vector of these pathogens since the prevelance of the interrelationships between Acanthamoeba and bacteria has only recently been investigated and its presence has surely been extremely underestimated.

The method of taxonomic identification of Acanthamoeba has been a consistant issue in this field. The use of morphology as the basis for species names is questionable and has led to considerable misinformation in the literature. The use of 18S rDNA to identify Acanthamoeba has largely replaced morphological methods but many researchers still resist the use of genetic identification. This may be due to the difficulty inherent to using the 18S rDNA as an identifier, as it requires expert knowledge to make correct identification. It also relies on the assumption that previously sequenced genes were properly genotyped. However, it is possible that DNA barcoding could be a reasonable additional identifier or even substitutefor information on 18S rDNA, in light of its ability to properly predict genotypes in a manner similar to the 18S rRNA gene.

While DNA barcoding requires less specialized knowledge and equipment to identify Acanthamoeba, that is not the most important potential advantage. The current

127 method of culturing and identifying Acanthamoeba results in a single clonal isolate being found. However, it is unknown whether multiple populations of genetically different

Acanthamoeba may be contributing to an infection. Additionally, no research has been done to study the interaction between genetically different Acanthamoeba, either in the environment or in vitro. DNA barcoding offers a unique opportunity to ask whether or not an Acanthamoeba infection is caused by one or numerous types of Acanthamoeba.

Through the use of RT-PCR with genotype specific or even sub-clade specific primers, researchers could identify not only if different Acanthamoeba exist within a sample, say from a cornea scrape, but also the relative level of those different populations. Since

DNA barcoding requires much less sequence to properly identify Acanthamoeba, such primers could be created to identify numerous types of Acanthamoeba simultaneously.

This also would remove the need to use single pure isolates in order to identify

Acanthamoeba. Thus, the relative diversity of the Acanthamoeba from an environmental sample could be determined through this method. Finally, with faster and easier methods of identification, more definitive answers about particular Acanthamoeba genotypes or clones being associated with particular diseases (ie GAE vs AK), as well as any correlation between geographical location and the relative frequency of a particular

Acanthamoeba type could be investigated. An easier method of identification will hopefully encourage Acanthamoeba researchers to properly identify their samples and may lend clarification to the contradictory results in the literature about which

Acanthamoeba are pathogenic, as well as prevent misusage of species names as identifiers for what are often remarkably different organisms.

128

It is obviously important to understand the relative risk of Acanthamoeba to the public health. Thus it has been important to understand the frequency of Acanthamoeba in the environment as well as the relative distribution of particular genotypes of

Acanthamoeba. The frequency of Acanthamoeba in the environment is probably much higher than previously estimated and it is likely that humans come into contact with

Acanthamoeba much more often than the the level of disease attrinuted to Acanthamoeba would suggest. It seems likely that human-Acanthamoeba interaction is fairly common and it is a combination of numerous factors that results in an actual infection. These factors could include the existence of a particular virulent Acanthamoeba, either due to genetics or to the presence of endosymbionts, though probably both. Additionally, it seems likely that some level of immunodeficiency in the host is required, perhaps a very specific suceptibility to Acanthamoeba in an otherwise normal immunocompetent individual. Finally, a method of entry into the host body either through an microabrasion caused by a contact lens, the inhalation of contaminated water or a skin wound is probably necessary to start an initial infection.

The consistent rarity of infection in spite of rising contact between the amoeba and its host does suggest a perfect storm of conditions are necessary for infection.

Additionally, the prevalence of Acanthamoeba cysts in the stroma of the eye during infection (Hsia, unpublished data) suggests that Acanthamoeba does not find a mammalian host to be a particularly hospitable environment. In the instance of keratitis, it is usually the immune cells that cause a majority of the damage, not the Acanthamoeba themselves. However, these unusual conditions must not suggest that Acanthamoeba is

129 not a risk. The rising use of contact lens and the changes in water treatment which are increasing the density of microorganisms in the water system are only going to increase the number of cases of AK. Additionally, the role of Acanthamoeba as a vector for other pathogens is an additional risk of public health. There must be more efforts to understand

Acanthamoeba, how it causes disease, and its relationship with other pathogens. More monitoring of the frequency of Acanthamoeba as well as other bacterial pathogens in the water system is necessary. More environmentally friendly options of water treatment must still be tested for the efficiency against Acanthamoeba. New methods such as UV irradiation must take into account how hardy Acanthamoeba is. If Acanthamoeba is able to survive treatment and continues to carry pathogenic bacteria, then the treatment itself is useless. Much more focus needs to be placed on in reducing Acanthamoeba in the water supply as well as reducing its opportunity to interact with pathogenic bacteria, for which it can act as a motile vector while bacteria simultaneously increase the virulence of

Acanthamoeba. Finally, more research will be necessary to develop better drug treatment plans for Acanthamoeba. With end results like blindness, corneal transplants and death, better therapies are required to improve patient outcomes. Such therapies will only come with better understanding of Acanthamoeba pathogenicity, how the amoeba interacts with humans and its relative contribution to other diseases like Legionnaire’s.

The question of how potentially pathogenic Acanthamoeba can be devoid of any endosymbionts is something that also needs to be explored. It seems unlikely that a bacterivore is ever truly free of intracellular bacteria, especially since a wide variety of bacteria can survive the encystment process in Acanthamoeba. It is also interesting to

130 question whether a particular species or genus of endosymbiont may have a more potent effect on Acanthamoeba’s virulence than another bacteria. For example, Legionella has been shown to change gene expression in Dictyostelium and Hartmanella (abu Kwaik et al., 1994; Farbrother et al., 2006) following uptake into the amoeba. Legionella highjacks the amoeba's cellular function in order to replicate. In Dictyostelium, 24 hours post infection, major transcriptional changes had occurred in 10% of probed genes

(Farbrother et al., 2006). Immediately following infection, there is an overall increase of transcription of genes associated with translation or other metabolic processes, suggesting the energy needs of the amoeba would be required to increase to keep up with increased protein synthesis. This could easily be translated into an increased virulence in an organism like Acanthamoeba. If Legionella increases the demands on Acanthamoeba metabolically, then this would increase the need of Acanthamoeba to consume other bacteria or in the case of human infection, cause Acanthamoeba to consume corneal cells.

It will be important to understand what virulence factors are upregulated in

Acanthamoeba following Legionella infection. This could lead to novel targets against

Acanthamoeba pathogenicity and help understand the process by which Legionella infects macrophages in humans.

It is very likely, but also underappreciated, that the endosymbionts of

Acanthamoeba play a role in Acanthamoeba keratitis. In vivo studies have shown that a

Legionella-infected Acanthamoeba results in a more severe keratitis infection versus

Legionella or Acanthamoeba alone (Hsia, unpublished data). It is possible that the

Acanthamoeba provide a mode of entry into the eye and the release of Legionella into the

131 stroma enhances damage caused by the immune response to the Acanthamoeba. As stated previously, Acanthamoeba cysts are often visible in the stroma. It is equally possible that the encystment process is stimulated by the temperature dependent virulence factors in Legionella and as Acanthamoeba encyst, it expels Legionella into the stroma.

Either way,the role of Legionella role in Acanthamoeba pathogenesis still has many unanswered questions.

The alternative aspect of Acanthamoeba pathogenesis concerns the role

Acanthamoeba may play in the pathogenesis of bacteria. While Acanthamoeba can apparently harbor endosymbionts for years, it has also been show that Acanthamoeba can resuscitate and increase the virulence of Legionella. It stands to reason that the role of

Acanthamoeba as a vector may be even more substantial than that. Previous studies found that when inhaled in the murine model Legionella-infected Hartmanella were able to cause more severe pneumonia when compared to Legionella or amoeba alone.

Additionally, the development itself of pneumonia demonstrates that bacteria harbored in an amoeba are capable of causing disease in a multicellular host (Brieland et al., 1996;

Brieland et al., 1997). Equivalent experiments have not been performed with

Acanthamoeba. This raises questions about the role of Acanthamoeba in Legionnaire's disease and whether or not inhalation of Acanthamoeba that contain Legionella could result in someone contracting Legionnaire’s disease. It has been shown that

Acanthamoeba can be airborne, and is generally more capable of surviving adverse conditions than Legionella. Legionella is notoriously difficult to culture in vitro and is better adapted to surviving in a host like an amoeba than out in the naked environment.

132

However, given the difficulties of diagnosing an infection by Legionella, let alone one involving Acanthamoeba, it would be difficult to link Acanthamoeba conclusively to

Legionnaire’s disease. Regardless, this relationship needs significantly more scrutiny in order to understand the relative risks of these organisms individually and separately.

Acanthamoeba has been both a challenge and a joy to work with. Its morphological plasticity and motility make it fascinating to view, and its ability to overcome multicellular hosts is nothing short of impressive. Only time will tell how this emerging pathogen will play a role in the public health arena. Much more research needs to be dedicated in understanding how it causes disease, and an increased surveillance of its frequency in the US domestic water supply must be considered a priority. Further investigation is required to clarify its role as a vector for numerous bacterial pathogens as well as the role of those bacteria in Acanthamoeba-related disease. Hopefully, we will continue to make great strides in our understanding of this opportunistic pathogen and develop new strategies to combat the risk of Acanthamoeba diseases.

133

BIBLIOGRAPHY

1. Abd H, Saeed A, Weintraub A, Nair GB, Sandström G. (2007) Vibrio cholerae O1 strains are facultative intracellular bacteria, able to survive and multiply symbiotically inside the aquatic free-living amoeba Acanthamoeba castellanii. FEMS Microbiol Ecol. 60(1):33-9. 2. abu Kwaik Y, Fields BS, Engleberg NC. (1994) Protein expression by the protozoan Hartmannella vermiformis upon contact with its bacterial parasite Legionella pneumophila. Infect. Immun. May 1994 62:1860-1866 3. Adl, MS, Gupta, VVS. (2006) Protists in soil ecology and forest nutrient cycling. Source: Canadian Journal of Forest Research, 36(7):1805-1817. 4. Aichelburg AC, Walochnik J, Assadian O, Prosch H, Steuer A, Perneczky G, Visvesvara GS, Aspöck H, Vetter N. (2008) Successful treatment of disseminated Acanthamoeba sp. infection with miltefosine. Emerg Infect Dis. 14(11):1743-6. 5. Aksozek A, McClellan K, Howard K, Niederkorn JY, Alizadeh H. (2002) Resistance of Acanthamoeba castellanii Cysts to Physical, Chemical, and Radiological Conditions. Journal of Parasitology. 88(3):621-623. 6. Alizadeh H, Apte S, El-Agha MS, Li L, Hurt M, Howard K, Cavanagh HD, McCulley JP, Niederkorn JY. (2001) Tear IgA and serum IgG antibodies against Acanthamoeba in patients with Acanthamoeba keratitis. Cornea. 20(6):622-7. 7. Bacon AS, Dart JK, Ficker LA, Matheson MM, Wright P. (1993) Acanthamoeba keratitis: The value of early diagnosis. Ophthalomol. 100:1238- 1243. 8. Bacon AS, Frazer DG, Dart JK, Matheson M, Ficker LA, Wright P. (1993) A review of 72 consecutive cases of Acanthamoeba keratitis. Eye. 7:719-725. 9. Barete S, Combes A, de Jonckheere JF, Datry A, Varnous S, Martinez V, Ptacek SG, Caumes E, Capron F, Francès C, Gibert C, Chosidow O. (2007) Fatal disseminated Acanthamoeba lenticulata infection in a heart transplant patient. Emerg Infect Dis. 13(5):736-8. 10. Berk SG, Ting RS, Turner GW, Ashburn RJ. (1998) Production of respirable vesicles containing live Legionella pneumophila cells by two Acanthamoeba spp Appl Environ Microbiol. 64(1):279-86. 11. Bonilla-Lemus P, Ramírez-Bautista GA, Zamora-Muñoz C, Ibarra-Montes Mdel R, Ramírez-Flores E, Hernández-Martínez MD. (2010) Acanthamoeba spp. in domestic tap water in houses of contact lens wearers in the metropolitan 134

area of Mexico City. Exp Parasitol. 126(1):54-8. 12. Booton GC, Carmichael JR, Visvesvara GS, Byers TJ, Fuerst PA. (2003) Genotyping of Balamuthia mandrillaris based on nuclear 18S and mitochondrial 16S rRNA genes. Am J Trop Med Hyg. 68(1):65-9. 13. Booton GC, Joslin CE, Shoff M, Tu EY, Kelly DJ, Fuerst PA. (2009) Genotypic identification of Acanthamoeba sp. isolates associated with an outbreak of Acanthamoeba keratitis. Cornea. 28(6):673-6. 14. Booton GC, Kelly DJ, Chu YW, Seal DV, Houang E, Lam DS, Byers TJ, Fuerst PA. (2002) 18S ribosomal DNA typing and tracking of Acanthamoeba species isolates from corneal scrape specimens, contact lenses, lens cases, and home water supplies of Acanthamoeba keratitis patients in Hong Kong. J Clin Microbiol. 40(5):1621-5. 15. Booton GC, Rogerson A, Bonilla TD, Seal DV, Kelly DJ, Beattie TK, Tomlinson A, Lares-Villa F, Fuerst PA, Byers TJ.(2004) Molecular and physiological evaluation of subtropical environmental isolates of Acanthamoeba spp., causal agent of Acanthamoeba keratitis. J Eukaryot Microbiol. 51(2):192- 200. 16. Booton GC, Visvesvara GS, Byers TJ, Kelly DJ, Fuerst PA. (2005) Identification and distribution of Acanthamoeba species genotypes associated with nonkeratitis infections. J Clin Microbiol. 43(4):1689-93. 17. Bouyer S, Imbert C, Rodier MH, Héchard Y. (2007) Long-term survival of Legionella pneumophila associated with Acanthamoeba castellanii vesicles. Environ Microbiol. 9(5):1341-4. 18. Bouyer S, Rodier MH, Guillot A, Héchard Y. (2009) Acanthamoeba castellanii: Proteins involved in actin dynamics, glycolysis, and proteolysis are regulated during encystation, Experimental Parasitology. 123(1):90-94. 19. Bovee EC. (1985) Class Lobosea Carpenter 1861 in J. J. Lee, S. H. Hutner, and E. C. Bovee, eds. An illustrated guide to the protozoa. Society of Protozoologists. Lawrence, Kans. pp. 158–211 20. Brieland J, McClain M, Heath L, Chrisp C, Huffnagle G, LeGendre M, Hurley M, Fantone J, Engleberg C. (1996) Coinoculation with Hartmannella vermiformis enhances replicative Legionella pneumophila lung infection in a murine model of Legionnaires' disease. Infect. Immun. 64:2449-56 21. Brieland JK, Fantone JC, Remick DG, LeGendre M, McClain M, and Engleberg NC. (1997) The role of Legionella pneumophila-infected Hartmannella vermiformis as an infectious particle in a murine model of Legionnaire's disease. Infect. Immun. 65:5330-.3 22. Buse HY, Ashbolt NJ. (2011) Differential growth of Legionella pneumophila strains within a range of amoebae at various temperatures associated with in- premise plumbing. Lett Appl Microbiol. 53(2):217-24. 23. Byers TJ, Kim BG, King LE, Hugo ER. (1991) Molecular Aspects of the Cell Cycle and Encystment of Acanthamoeba. Clin Infect Dis. 13(5): S373-S384. 24. Castellani A. (1930). An amoeba found in culture of yeast: preliminary note. J. Trop. Med. Hyg. 33:160. 135

25. Cavalier-Smith, T., Chao, EEY. & Oates, B. (2004) Molecular phylogeny of Amoebozoa and the evolutionary significance of the unikont Phalansterium. Eur. J. Protistol., 40:21–48. 26. Chambers JA & Thompson JE. (1974) Age-dependent Excystment of the Protozoan Acanthamoeba castellanii. J Gen Microbiol. 80:375-380. 27. Chappell CL, Wright JA, Coletta M, Newsome AL. (2001) Standardized method of measuring Acanthamoeba antibodies in sera from healthy human subjects. Clin Diagn Lab Immunol. 8(4):724-30. 28. Choi SH, Cho MK, Ahn SC, Lee JE, Lee JS, Kim DH, Xuan YH, Hong YC, Kong HH, Chung DI, Yu HS. (2009) Endosymbionts of Acanthamoeba isolated from domestic tap water in Korea. Korean J Parasitol. 47(4):337-44. 29. Chomicz L, Padzik M, Graczyk Z, Starosciak B, Graczyk TK, Naprawska A, Oledzka G, Szostakowska B. (2010) Acanthamoeba castellanii: in vitro effects of selected biological, physical and chemical factors. Exp Parasitol. 126(1):103-5. 30. Chomicz L, Zebrowska J, Piekarczyk J, Staroseciak B, Myjak P, Walski M and Kazimierczuk Z. (2005) In vitro studies on susceptibility of Acanthamoeba castellanii to selected chemical agents. Acta Parasitologica. 50(1), 25–31 31. Cirillo JD, Cirillo SL, Yan L, Bermudez LE, Falkow S, Tompkins LS. (1999) Intracellular growth in Acanthamoeba castellanii affects monocyte entry mechanisms and enhances virulence of Legionella pneumophila. Infect Immun. 67(9):4427-34. 32. Corsaro D, Venditti D. (2010) Phylogenetic evidence for a new genotype of Acanthamoeba (Amoebozoa, Acanthamoebida). Parasitol Res. 107(1):233-8. 33. De Jonckheere JF, Michel R. (1988) Species identification and virulence of Acanthamoeba strains from human nasal mucosa. Parasitol Res. 74(4):314-6. 34. De Moraes J & Alfieri SC. (2008) Growth, encystment and survival of Acanthamoeba castellanii grazing on different bacteria. FEMS Microbiology Ecology. 66:221–229. 35. Di Cave D, Monno R, Bottalico P, Guerriero S, D'Amelio S, D'Orazi C, Berrilli F. (2009) Acanthamoeba T4 and T15 genotypes associated with keratitis infections in Italy. Eur J Clin Microbiol Infect Dis. 28(6):607-12. 36. Dini LA, Cockinos C, Frean JA, Niszl IA, Markus MB. (2000) Unusual case of Acanthamoeba polyphaga and Pseudomonas aeruginosa keratitis in a contact lens wearer from Gauteng, South Africa. J Clin Microbiol. 38(2):826-9. 37. Dyková I, Lom J, Schroeder-Diedrich JM, Booton GC, Byers TJ. (1999) Acanthamoeba strains isolated from organs of freshwater fishes. J. Parasitiol. 85:1106-1113. 38. Dykova I, Veverkova-Fialova M, Fiala I, Dvorakova H. (2005) ProtAcanthamoeba bohemica sp. n., Isolate from the Liver of Tench, Tinca tinca (Linnaeus, 1758). Acta Protozoologica. 44:369-376. 39. Edagawa A, Kimura A, Kawabuchi-Kurata T, Kusuhara Y, Karanis P. (2009) Isolation and genotyping of potentially pathogenic Acanthamoeba and Naegleria species from tap-water sources in Osaka, Japan. Parasitol Res. 105(4):1109-17. 136

40. Espinoza, L, Slaton, N, Mozaffari, M. Understanding the Numbers on Your Soil Test Report. Agriculture and Natural Resources. Available: http://www.uaex.edu/Other_Areas/publications/pdf/FSA-2118.pdf Accessed 2012 Feb 13. 41. Farbrother P, Wagner C, Na J, Tunggal B, Morio T, Urushihara H, Tanaka Y, Schleicher M, Steinert M, Eichinger L. (2006), Dictyostelium transcriptional host cell response upon infection with Legionella. Cellular Microbiology, 8: 438– 456. 42. Frank W & Bosch I. (1972) Isolation of amoebae of the type Hartmannella- Acanthamoeba and Naelgeria from cold-blooded vertebrates. Z Parasitenkd. 40(2):139-50. 43. Fraser C, Hanage WP, Spratt BG. (2007) Recombination and the nature of bacterial speciation. Science. 315(5811)476-480. 44. Fritsche TR, Sobek D, Gautom RK. (1998) Enhancement of in vitro cytopathogenicity by Acanthamoeba spp. following acquisition of bacterial endosymbionts. FEMS Microbiol Lett. 166(2):231-6. 45. Garate M, Cao Z, Bateman E, Panjwani N. (2004) Cloning and characterization of a novel mannose-binding protein of Acanthamoeba. J. Biol. Chem. 279:29849-29856. 46. Gast RJ, Ledee, DR, Fuerst PA, Byers TJ. (1996). Subgenus Systematics of Acanthamoeba: Four Nuclear 18S rDNA Sequence Types. Journal of Eukaryotic Microbiology, 43:498–504. 47. Glaberman S, Sulaiman IM, Bern C, Levior J, Peng MM, Morgan U, Gilman R, Lal AA, Xiao L. (2001) A Multilocus Genotypic Analysis of Cryptosporidium meleagridis. Journal of Eukaryotic Microbiology. 48: 19–22. 48. Goodall K, Brahma A, Ridgeway A. (1996) Acanthamoeba keratitis: Masquerading as adenoviral keratitis. Eye. 10:643-644. 49. Hajibabaei M, Janzen DH, Burns JM, Hallwachs W, Hebert PD. (2006) DNA barcodes distinguish species of tropical Lepidoptera. ProcNat Acad Sci Unit Stat Amer. 103(4):968-971 50. Hajibabaei M, Singer GA, Clare EL, Hebert PD. (2007) Design and applicability of DNA arrays and DNA barcodes in biodiversity monitoring. BMC Biol. 5:24 51. Hebert PD, Cywinska A, Ball SL, deWaard JR. (2003) Biological identifications through DNA barcodes. Proc Biol Sci. 270(1512):313-321 52. Hebert PD, Stoeckle MY, Zemlak TS, Francis CM. (2004) Identification of birds through DNA barcodes. PLoS Biology. 2(10):E312 53. Heinz E, Kolarov I, Kästner C, Toenshoff ER, Wagner M, Horn M. (2007) An Acanthamoeba sp. containing two phylogenetically different bacterial endosymbionts. Environ Microbiol. 9(6):1604-9. 54. Heitman J, Filler SG, Edwards JE, Mitchell AP. (2006)Fungal Molecular Pathogenesis: Evolution of Human Pathogenic Fungi: Phylogenies and Species In Molecular Principles of Fungal Pathogenesis. Washington, DC: ASM Press;113– 162. 137

55. Henriquez FL, Ingram PR, Muench SP, Rice DW, Roberts CW. (2008) Molecular basis for resistance of Acanthamoeba tubulins to all major classes of antitubulin compounds. Antimicrob. Agents Chemother. 52(3):1133-1135. 56. Horn M, Fritsche TR, Gautom RK, Schleifer KH, Wagner M.(1999) Novel bacterial endosymbionts of Acanthamoeba spp. related to the Paramecium caudatum symbiont Caedibacter caryophilus. Environ Microbiol. 1(4):357-67. 57. Hsu BM, Chen CH, Wan MT, Cheng HW. (2006) Legionella prevalence in hot spring recreation areas of Taiwan. Water Res. 40(17):3267-73. 58. Iovieno A, Ledee DR, Miller D, Alfonso EC. (2010) Detection of bacterial endosymbionts in clinical Acanthamoeba isolates. Ophthalmology. 117(3):445- 452. 59. Jager, BV & Stamm, W. (1972) Brain abscesses caused by free-living amoeba probably of the genus Hartmannella in a patient with Hodgkin's disease. Lancet. 2:1343-1345. 60. Jaison PL, Cao Z, Panjwani N. (1998) Binding of Acanthamoeba to 23 mannose-glycoproteins of corneal epithelium: effect of injury. Current Eye Research. 17:8, 770-776. 61. Jeong HJ, Yu HS. (2005) The role of domestic tap water in Acanthamoeba contamination in contact lens storage cases in Korea. Korean J Parasitol. 43(2):47-50. 62. Jones BR, McGill JI, Steele AD. (1975) Recurrent suppurative kerato-uveitis with loss of eye due to infection by Acanthamoeba castellanii. Trans Ophthalmol Soc U K. 95(2):210-3. 63. Joslin CE, Tu EY, McMahon TT, Passaro DJ, Stayner LT, Sugar J. (2006) Epidemiological characteristics of a Chicago-area Acanthamoeba keratitis outbreak. Am J Ophthalmol. 142(2):212-7. 64. Joslin CE, Tu EY, Shoff ME, Booton GC, Fuerst PA, McMahon TT, Anderson RJ, Dworkin MS, Sugar J, Davis FG, Stayner LT. (2007) The association of contact lens solution use and Acanthamoeba keratitis. Am J Ophthalmol. 144(2):169-180. 65. Kennedy GM, Morisaki JH, Champion PA. (2012) Conserved mechanisms of Mycobacterium marinum pathogenesis within the environmental amoeba Acanthamoeba castellanii. Appl Environ Microbiol.78(6):2049-52. 66. Kent M, Platt SR, Rech RR, Eagleson JS, Howerth EW, Shoff M, Fuerst PA, Booton G, Visvesvara GS, Schatzberg SJ. (2011) A multi-systemic infection with an Acanthamoeba species in a dog. J. Amer. Vet. Med. Assoc. 238:1476- 1481. 67. Khan NA. (2007) Acanthamoeba invasion of the central nervous system, International Journal for Parasitology. 37(2):131-138. 68. Khunkitti W, Lloyd D, Furr JR, Russell AD. (1998) Acanthamoeba castellanii: Growth, encystment, excystment and biocide susceptibility. Journal of Infection. 36(1):43-48. 69. Kilvington S, Gray T, Dart J, Morlet N, Beeching JR, Frazer DG, Matheson M. (2004) Acanthamoeba keratitis: the role of domestic tap water contamination 138

in the United Kingdom. Invest Ophthalmol Vis Sci. 45(1):165-9. 70. Kong HH, Chung DI. (1996) PCR and RFLP variation of conserved region of small subunit ribosomal DNA among Acanthamoeba isolates assigned to either A. castellanii or A. polyphaga. Korean J Parasitol. 34(2):127-34. 71. Kreuzer K, Adamczyk J, Iijima M, Wagner M, Scheu S, Bonkowski M. (2006) Grazing of a common species of soil protozoa (Acanthamoeba castellanii) affects rhizosphere bacterial community composition and root architecture of rice (Oryza sativa L.). Soil Biology and Biochemistry. 38(7):1665-1672. 72. Krishna Murti CR & Shukla OP. (1984) Differentiation of pathogenic amoebae: encystation and excystation of Acanthamoeba culbertsoni — A model. Journal of Biosciences. 6(4):475-489. 73. Ledee DR, Booton GC, Awwad MH, Sharma S, Aggarwal RK, Niszl IA, Markus MB, Fuerst PA, Byers TJ. (2003) Advantages of using mitochondrial 16S rDNA sequences to classify clinical isolates of Acanthamoeba. Invest Ophthalmol Vis Sci. 44(3):1142-9. 74. Ledee DR, Hay J, Byers TJ, Seal DV, Kirkness CM. (1996) Acanthamoeba griffin: Molecular characterization of a new corneal pathogen. Invest Ophthalmol Vis Sci. 37(4):544-50. 75. Lloyd D, Turner NA, Khunkitti W, Hann AC, Furr JR, Russell AD. (2001) Encystation in Acanthamoeba castellanii: development of biocide resistance. J Eukaryot Microbiol. 48(1): 11-16. 76. Lorenzo-Morales J, Kliescikova J, Martinez-Carretero E, De Pablos LM, Profotova B, Nohynkova E, Osuna A, Valladares B. (2008) Glycogen Phosphorylase in Acanthamoeba spp.: Determining the Role of the Enzyme during the Encystment Process Using RNA Interference. Eukaryotic Cell. 7:509- 517. 77. Lorenzo-Morales J, Lindo JF, Martinez E, Calder D, Figueruelo E, Valladares B, Ortega-Rivas A. (2005) Pathogenic Acanthamoeba strains from water sources in Jamaica, West Indies. Ann Trop Med Parasitol. 99(8):751-8. 78. Lorenzo-Morales J, Ortega-Rivas A, Foronda P, Abreu-Acosta N, Ballart D, Martínez E, Valladares B. (2005) RNA interference (RNAi) for the silencing of extracellular serine proteases genes in Acanthamoeba: Molecular analysis and effect on pathogenicity, Molecular and Biochemical Parasitology. 144(1)10-15. 79. Maghsood AH, Sissons J, Rezaian M, Nolder D, Warhurst D, Khan NA. (2005) Acanthamoeba genotype T4 from the UK and Iran and isolation of the T2 genotype from clinical isolates. J Med Microbiol. 54(8):755-9. 80. Maiden CJM, Bygraves JA, Feil E, Morelli G, Russell JE, Urwin R, Zhang Q, Zhou J, Zurth K, Caugant DA, Feavers IM, Achtman M, Spratt BG. (1998) Multilocus sequence typing: A portable approach to the identification of clones within populations of pathogenic microorganisms. PNAS. 95(6)3140- 3145. 81. Marshall E. (2005) Taxonomy: Will DNA bar codes breathe life into classification? Science. 307(5712):1037 82. Martinez AJ & Visvesvara GS. (1997) Free-living, Amphizoic and 139

Opportunistic Amebas. Brain Pathology. 7:583–598. 83. Martinez AJ. (1993) Free-living amebas: infection of the central nervous system. Mt Sinai J Med. 60(4):271-8. 84. Matsunaga S, Endo T, Yagita K, Hirukawa Y, Tomino S, Matsugo S, Tsuruhara T. (1998) Chromosome Size Polymorphisms in the Genus Acanthamoeba Electrokaryotype by Pulsed-Field Gel Electrophoresis. Protist. 149(4):323-340. 85. Mattar FE & Byers TJ. (1971) Morphological Changes and the Requirements for Macromolecule Synthesis during Excystment of Acanthamoeba castellanii. J Cell Biol. 49:507-519. 86. Meusnier I, Singer GAC, Landry JF, Hickey DA, Hebert PDN, Hajibabaei M. (2008) A universal DNA mini-barcode for biodiversity analysis. BMC Genomics. 9:214 87. Michel R, Burghardt H, Bergmann H. (1995) Acanthamoeba, naturally intracellularly infected with Pseudomonas aeruginosa, after their isolation from a microbiologically contaminated drinking water system in a hospital. Zentralbl Hyg Umweltmed. 196(6):532-44. 88. Moon EK, Chung DI, Hong Y, Kong HH. (2011) Expression levels of encystation mediating factors in fresh strain of Acanthamoeba castellanii cyst ESTs. Experimental Parasitology. 127(4):811-816. 89. Moon EK, Kim JO, Xuan YH, Yun YS, Kang SW, Lee YS, Ahn TI, Hong YC, Chung DI, Kong HH. (2009) Construction of EST database for comparative gene studies of Acanthamoeba. Korean J Parasitol. 47(2):103-7. 90. Nagington J, Watson PG, Playfair TJ. (1974) Amoebic infection of the eye. Lancet. 2:1547-1550. 91. Nakamura S, Matsuo J, Hayashi Y, Kawaguchi K, Yoshida M, Takahashi K, Mizutani Y, Yao T, Yamaguchi H. (2010), Endosymbiotic bacterium Protochlamydia can survive in Acanthamoeba following encystation. Environmental Microbiology Reports, 2:611–618. 92. Ohno A, Kato N, Sakamoto R, Kimura S, Yamaguchi K. (2008) Temperature- dependent parasitic relationship between Legionella pneumophila and a free- living amoeba (Acanthamoeba castellanii). Appl Environ Microbiol. 74(14):4585-8. 93. Page, FC. (1987) The classification of ‘naked’ amoebae (Phylum Rhizopoda). Archiv Fur Protistenkunde 133:199–217. 94. Palmer CJ, Tsai YL, Paszko-Kolva C, Mayer C, Sangermano LR. (1993) Detection of Legionella species in sewage and ocean water by polymerase chain reaction, direct fluorescent-antibody, and plate culture methods. Appl Environ Microbiol. 59(11):3618-24. 95. Pawlowski, J & Burki, F. (2009) Untangling the Phylogeny of Amoeboid Protists. J Eukaryot Microbiol. 56:16–25 96. Prowell DP, McMichael M, Silvain JF. (2004) Multilocus Genetic Analysis of Host Use, Introgression, and Speciation in Host Strains of Fall Armyworm (Lepidoptera: Noctuidae) Annals of the Entomological Society of America. 140

97(5):1034-1044. 97. Rimm DL, Pollard T D, Hieter, P. (1998) Resolution of Acanthamoeba castellanii chromosomes by pulsed field gel electrophoresis and construction of the initial linkage map. Chromosoma. 97(3): 219-223. 98. Rowbotham TJ. (1980) Preliminary report on the pathogenicity of Legionella pneumophila for freshwater and soil amoebae. J Clin Pathol. 33(12):1179-83. 99. Saeed A, Abd H, Edvinsson B, Sandström G. (2009) Acanthamoeba castellanii an environmental host for Shigella dysenteriae and Shigella sonnei. Arch Microbiol. 191(1):83-8. 100. Said NA, Shoeir AT, Panjwani N, Garate M, Cao Z. (2004) Local and systemic humoral immune response during acute and chronic Acanthamoeba keratitis in rabbits. Current Eye Research. 29:6, 429-439 101. Sandström G, Saeed A, Abd H. (2010) Acanthamoeba polyphaga is a possible host for Vibrio cholerae in aquatic environments. Exp Parasitol. 126(1):65-8. 102. Schmitz-Esser S, Toenshoff ER, Haider S, Heinz E, Hoenninger VM, Wagner M, Horn M. (2008) Diversity of bacterial endosymbionts of environmental Acanthamoeba isolates. Appl Environ Microbiol. 74(18):5822-31. 103. Schroeder JM, Booton GC, Hay J, Niszl IA, Seal DV, Markus MB, Fuerst PA, Byers TJ. (2001) Use of subgenic 18S ribosomal DNA PCR and sequencing for genus and genotype identification of Acanthamoeba from humans with keratitis and from sewage sludge. J Clin Microbiol. 39(5):1903-11. 100. Schroeder JM, Booton GC, Hay J, Niszl IA, Seal DV, Markus MB, Fuerst PA, Byers TJ. (2001) Use of subgenic 18S ribosomal DNA PCR and sequencing for genus and genotype identification of Acanthamoeba from humans with keratitis and from sewage sludge. J Clin Microbiol. 39(5):1903-11. 104. Seal DV, Kirkness CM, Bennett HGB, Peterson M. (1999) Acanthamoeba keratitis in Scotland: risk factors for contact lens wearers, Contact Lens and Anterior Eye. 22(2):58-68. 105. Shoff ME, Joslin CE, Tu EY, Kubatko L, Fuerst PA. (2008) Efficacy of contact lens systems against recent clinical and tap water Acanthamoeba isolates. Cornea. 27(6):713-9. 106. Shoff ME, Rogerson A, Kessler K, Schatz S, Seal DV. (2008) Prevalence of Acanthamoeba and other naked amoebae in South Florida domestic water. J Water Health. 6(1):99-104. 107. Siddiqui, R & Khan, NA. (2012) Biology and pathogenesis of Acanthamoeba. Parasites & Vectors. 5(6):1-13. 108. Spanakos G, Tzanetou K, Miltsakakis D, Patsoula E, Malamou-Lada E, Vakalis NC. (2006) Genotyping of pathogenic Acanthamoeba isolated from clinical samples in Greece--report of a clinical isolate presenting T5 genotype. Parasitol Int. 55(2):147-9. 109. Sriram R, Shoff M, Booton GC, Fuerst PA, Visvesvara GS. (2008) Survival of Acanthamoeba cysts after desiccation for more than 20 years J. Clin. Microbiol. 46(12):4045-4048. 110. Stehr-Green JK, Bailey TM, Visvesvara GS. (1989) The epidemiology of 141

Acanthamoeba keratitis in the United States. Am J Ophthalmol. 107(4):331-6. 111. Steinert M, Emödy L, Amann R, Hacker J. (1997) Resuscitation of viable but nonculturable Legionella pneumophila Philadelphia JR32 by Acanthamoeba castellanii. Appl Environ Microbiol. 63(5):2047-53. 112. Stone BJ, Abu Kwaik Y. (1998) Expression of multiple pili by Legionella pneumophila: identification and characterization of a type IV pilin gene and its role in adherence to mammalian and protozoan cells. Infect Immun. 66(4):1768- 75. 113. Stothard DR, Schroeder-Diedrich JM, Awwad MH, Gast RJ, Ledee DR, Rodriguez-Zaragoza S, Dean CL, Fuerst PA, Byers TJ. (1998) The evolutionary history of the genus Acanthamoeba and the identification of eight new 18S rRNA gene sequence types. J Eukaryot Microbiol. 45(1):45-54. 114. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, and Kumar S. (2011) MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Molecular Biology and Evolution. 28: 2731-2739. 115. Tan B, Weldonlinne CM, Rhone DP, Penning CL, & Visvesvara GS. (1993) Acanthamoeba infection presenting as skin lesions in patients with the acquired- immunodeficiency-syndrome. Arch. Path. Lab. Med. 117:1043-1046. 116. Tay-Kearney ML, McGhee CN, Crawford GJ, Trown K. (1993) Acanthamoeba keratitis: A masquerade of presentation in six cases. Aust NZ J Ophthalmol. 21:237-245. 117. Telang GH, Scola F, Kantor GR, Stieritz D, Reboli AC. (1996) Disseminated Acanthamoeba Infection in a Patient with AIDS. Infectious Diseases in Clinical Practice. 5(6):387-388. 118. Thebpatiphat N, Hammersmith KM, Rocha FN, Rapuano CJ, Ayres BD, Laibson PR, Eagle RC, Jr., Cohen EJ. (2007) Acanthamoeba keratitis: a parasite on the rise. Cornea. 26(6):701-706. 119. Tomlinson G. (1967) The Glyoxylate Pathway in Acanthamoeba spp. Journal of Eukaryotic Microbiology. 14:114–116. 120. Tu EY, Joslin CE, Nijm LM, Feder RS, Jain S, Shoff ME. (2009) Polymicrobial keratitis: Acanthamoeba and infectious crystalline keratopathy. Am J Ophthalmol. 148(1):13-9.e2. 121. Turner KME, Hanage WP, Fraser C, Connor TR and Spratt BG. (2007) Assessing the reliability of eBURST using simulated populations with known ancestry. BMC Microbiology, 7:30. 122. U.S. EPA. (1998) National Primary Drinking Water Regulations: Disinfectants and Disinfection. Byproducts. Federal Register 63(241):69389-69476. 123. Verani, JR, Lorick SA, Yoder JS, Beach MJ, Braden CR, Roberts JM, Conover CS, Chen S, McConnell KA, Chang DC, Park BJ, Jones DB, Visvesvara GS, Roy SL. (2009) National Outbreak of Acanthamoeba Keratitis Associated with Use of a Contact Lens Solution, United States. Emerg. Infect. Dis. 15: 1236-1242. 124. Visvesvara GS, Booton GC, Kelley DJ, Fuerst P, Sriram R, Finkelstein A, 142

Garner MM. (2007) In vitro culture, serologic and molecular analysis of Acanthamoeba isolated from the liver of a keel-billed toucan (Ramphastos sulfuratus). Vet Parasitol. 143(1):74-8. 125. Visvesvara GS, Moura G, Schuster FL. (2007) Pathogenic and opportunistic free-living amoebae: Acanthamoeba spp., Balamuthia mandrillaris, Naegleria fowleri, and Sappinia diploidea. FEMS Immunol Med Microbiol. 50(1):1-26. 126. Visvesvara GS. (1991) Classification of Acanthamoeba. Rev Infect Dis. 13(5):S369-S372. 127. Volkonsky, M. (1931) Hartmannella castellanii Douglas et classification des Hartmannelles. Arch. Zool. Exp. Gén. 72, 317-39 128. Waki K, Dutta S, Ray D, Kolli BK, Akman L, Kawazu SI, Lin CP, Chang KP. (2007) Transmembrane Molecules for Phylogenetic Analyses of Pathogenic Protists: Leishmania-Specific Informative Sites in Hydrophilic Loops of Trans- Endoplasmic Reticulum N-Acetylglucosamine-1-Phosphate Transferase. Eukary Cell 6(2):198-210. 129. Walochnik J, Obwaller A, Aspöck H. (2000) Correlations between morphological, molecular biological, and physiological characteristics in clinical and nonclinical isolates of Acanthamoeba spp. Appl Environ Microbiol. 66(10):4408-13. 130. Ward RD, Zemlak TS, Innes BH, Last PR, Hebert PD. (2005) DNA barcoding Australia's fish species. Philos Trans R Soc Lond B Biol Sci. 360(1462):1847- 1857. 131. Xuan YH, Yu HS, Jeong HJ, Seol SY, Chung DI, Kong HH. (2007) Molecular characterization of bacterial endosymbionts of Acanthamoeba isolates from infected corneas of Korean patients. Korean J Parasitol. 45(1):1-9. 132. Yeo M, Mauricio IL, Messenger LA, Lewis MD, Llewellyn MS. (2011)Multilocus Sequence Typing (MLST) for Lineage Assignment and High Resolution Diversity Studies in Trypanosoma cruzi.PLoS Negl Trop Dis 5(6):1049. 133. Yera H, Zamfir O, Bourcier T, Viscogliosi E, Noël C, Dupouy-Camet J, Chaumeil C. (2008) The genotypic characterization of Acanthamoeba isolates from human ocular samples. Br J Ophthalmol. 92(8):1139-41. 134. Yu HS, Jeong HJ, Hong YC, Seol SY, Chung DI, Kong HH. (2007) Natural occurrence of Mycobacterium as an endosymbiont of Acanthamoeba isolated from a contact lens storage case. Korean J Parasitol. 45(1):11-8.

143

APPENDIX

AI. DNeasy Kit: Protocol for DNA Purification of Animal Tissues All solutions except Acanthamoeba were provided by the kit. Solutions were prepared as instructed. 1. Acanthamoeba was grown to a density of 105 cells/ml. Acanthamoeba was collected, spun down, and resuspended in 180 μl Buffer ATL.

2. 20 μl proteinase K, was added, mixed by vortexing, and incubated at 55 C for 3- 24 h.

3. 200 μl Buffer AL was added to the sample, mixed by vortexing and then incubated at 70 C for 10 min.

4. 200 μl ethanol (96-100%) was added to the sample. Entire mixture was applied to a DNeasy spin column and centrifuged at 8000 rpm for 1 minute. Flow-through was discarded.

5. 500 μl Buffer AW1 was added to the column and centrifuged for 1 min at 8000 rpm. Flow through was discarded.

6. 500 μl Buffer AW2 was added to the column and centrifuged for 3 min at full speed to dry the DNeasy membrane. Flow-through was discarded.

7. DNA was eluted with 200 μl Buffer AE added directly onto the DNeasy membrane. The column was incubate at room temperature for 1 min, and then centrifuged for 1 min at 8000 rpm.

144

A2. PCR Purification using the PEG Extraction Protocol with minor modifications

20% PEG Solution 10.0 g Polyethylene glycol 7.3 g ddH2O up to 50ml

1) 25 μL of 20% PEG was added to each positive PCR tube. total mixture (25 μL of PEG and 21 μL of PCR mixture) was transferred to a 1.5ml tube. 2) Mixture was incubated at 4°C for one hour. 3) Mixture was spun at maximum speed (~15,000 x g) for 15 minutes at room temperature. 4) Supernatant was removed and 200 μL of cold 100% ethanol added to the tube. 5) Tube was spun again at maximum speed for 7 minutes. Supernatant was removed. 6) 250 μL of cold 75% ethanol added to the tube. Tube was spun again at maximum speed for 7 minutes. 7) Ethanol was poured off and tube dried at room temperature for 30 minutes. 8) The PCR product was resuspended by 15μL of water.

145

A3. QIAquick Gel Extraction Kit Protocol with minor modifications

All solutions were provided by the kit except the water, solutions were prepared as instructed on the kit. 1. DNA fragment was excised from the agarose gel and placed in a clean microfuge tube. 2. 300ul of Buffer QG was added to the fragment and incubated at 50°C for 10 minutes. 3. Mixture was then added to a QIAquick spin column and centrifuged for 1 minute to bind DNA to the column at 8000 rpm. 4. Flow-through was discarded and the column was washed with 500ul of Buffer QG. 5. The column was centrifuged for 1 minute at 8000 rpm, and then washed with 750ul of Buffer PE. The column was centrifuged for 1 minute at 8000 rpm. Flow through was discarded and the centrifuge was spun again for two minutes at the highest speed. 6. DNA was eluted by placing QIAquick column into a clean 1.5 ml microcentrifuge tube and 30 ul of H20 was added to the center of the membrane. The column was allowed to stand for one minute then centrifuged at the highest speed for one minute. Gel Extraction

146

A4. Media Recipes

Amoeba Saline (AS) 1. 5 Stock Solutions were prepared: a. NaCl : 1.20 g/100 ml b. CaCl2 : 0.04 g/100 ml c. MgSO4 : 0.04 g/100 ml d. Na2HPO4 : 1.42 g/100 ml e. KH2PO4 : 1.36 g/100 ml 2. 10 ml of each stock solution was added to 950 ml dH2O.

Non-Nutrient Amoeba Saline Agar (NNAS) 1. Bacteriological agar 15 g 2. Amoeba saline 1 L

ATCC medium: PYG 712 Basal Medium Proteose Peptone (BD 211684) 20.0 g Yeast extract 1.0 g Distilled water to 950.0 ml Autoclave 25 minutes at 121C. Aseptically added: 2 M Glucose (filter-sterilized) 50.0 ml

Bacto-Casitone Medium Formula Final concentration Bacto-Casitone (Difco) 2% Fetal Bovine Serum* 10% Penicillin 200IU/mL Streptomycin 200µg/mL Distilled water --

Prepared media with Bacto-Casitone and distilled water. Autoclaved media and added antibiotics and FBS when cooled.

1X Phosphate Buffered Saline 1. The following was dissolved in 800ml distilled H2O. 8g of NaCl 0.2g of KCl 1.44g of Na2HPO4 0.24g of KH2PO4 2. pH was adjusted to 7.4. 3. Volume was adjusted to 1L with additional distilled H2O. 4. Sterilized media by autoclaving

147

A5. Ethanol Preparation of DNA for Sequencing

Each sequencing PCR sample received the following procedure: 1. Added 8μl of deionized water and 32μl of non-denatured 95% ethanol and moved total volume to a 1.5ml tube. Incubated at room temperature for ten minutes. 2. Tubes were spun at 13,000 RPM for 20 minutes. Supernatant was removed and 250ul of 70% ethanol was added. 3. Tubes were spun for 10 minutes at 13,000 RPM. Supernatant was poured off and remaining ethanol evaporated by incubating the tubes at 70° C for ~30 min. 4. 15ul of Hi-Di formamide was added to each tube. Samples were vortexed vigorously and spun down. Samples were added to 96 well plates for sequencing.

148

A6. Fluorescent In Situ Hybridization 1. Acanthamoeba was harvested from axenic cultures and washed with AS. Acanthamoeba was spun down and resuspected in 100ul of AS. 2. 20ul aliquots of amebic suspension were incubated on poly-L-lysine slides for 20 minutes at 45 degrees Celsius. 3. Cells were fixed with 20ul of 4% paraformaldehyde for 20 minutes at room temperature 4. Slides were washed three times in 1X PBS. 5. The slides were dehydrated using 50%, 70%, 96% Ethanol for 3 minutes each time. 6. Slides were incubated in 20µL hybridization buffer with 100 ng of probe (1µL), for 1 hour at 45 °C. 7. Slides washed gently with washing buffer. 8. The slides reincubated for 15 minutes at 45 °C covered with 500ul washing buffer. 9. Slides were then washed with distilled water, dried at room temperature and mounted with mounting media.

In situ Solutions Mounting media 0.1M TRIS buffer (pH 9.0): 10 ml Glycerol: 90 ml

In situ Hybridization buffer NaCl 0.9M Tris-HCl 20mM Formamide 20% SDS 0.01% Water

In situ hybridization washing buffer NaCl 0.9M Tris-HCl 20mM SDS 0.01% EDTA .01M Water

149

A7. Polymerase Chain Reaction Formulas

NEB Taq DNA Polymerase DNA (2-3 ul) 10mM Nucleotides 4ul 20 pm primer 5’ 1ul 20pm primer 3’ 1ul Taq 0.3ul ThermoPol Buffer 2.5ul Water to total volume of 25ul Elongation temperature: 72°C

Phusion Taq Polymerase DNA (2-3 ul) 10mM Nucleotides 4ul 20 pm primer 5’ 1ul 20pm primer 3’ 1ul Taq 0.3ul Phusion Buffer 5ul Water to total volume of 25ul Elongation temperature: 68°C

OneTaq DNA (2-3 ul) 10mM Nucleotides 4ul 20 pm primer 5’ 1ul 20pm primer 3’ 1ul OneTaq 0.3ul GC Reaction Buffer 5ul Water to total volume of 25ul Elongation temperature: 68°C

Titanium Taq DNA (2-3 ul) 10mM Nucleotides 4ul 20 pm primer 5’ 1ul 20pm primer 3’ 1ul Titanium Taq 0.3ul 10X TitTaq Buffer 2.5ul Water to total volume of 25ul Elongation temperature: 68°C

150

A8. Polymerase Chain Reaction Cycle Procedures:

1. Initial denature step: 95 degrees for 5 min 2. Denature step: 95 degrees for 45 sec 3. Annealing step: Primer specific temperature as seen in Table 9 for 45 sec. 4. Elongation step: Polymerase specific temperature as listed above with a elongation length changed based off amplimer length where the assumption is a speed of 1Kb/min. Repeat steps 2-4 between 35-40 times. Final elongation step for 15 minutes at polymerase specific temperature.

Sequencing PCR Purified PCR product 2ul 5X BigDye Buffer 2ul BigDye Terminator v3.1 0.5ul 2pm primer 1ul Water 4.5ul Total volume 10ul

Sequencing PCR Cycle Step 1. 95 degrees for 30 sec Step 2. 50 degrees for 15 sec. Step 3. 60 degrees for 4 minutes. Repeat 25 cycles.

151

152