Viewed on a Zeiss Supra 35 FEG-VP Scanning Electron Microscope

MIAMI UNIVERSITY

The Graduate School

CERTIFICATE FOR APPROVING THE DISSERTATION

We hereby approve the Dissertation of

Dominika Angelika Jurkovic

Candidate for the Degree:

Doctor of Philosophy

______Dr. Mitchell F. Balish, Director

______Dr. Eileen Bridge, Reader

______Dr. Gary R. Janssen, Reader

______Dr. Luis Actis

______Dr. David G. Pennock, Graduate School Representative ABSTRACT

STRUCTURE, ORGANIZATION, AND FUNCTION OF THE TERMINAL ORGANELLE IN MYCOPLASMA PENETRANS

by Dominika Angelika Jurkovic

Bacteria utilize cytoskeletal elements to confer both morphological and functional polarity, which are necessary for growth and survival in their native environments. Some bacteria utilize polar appendages for attachment and motility to and within a host. Additionally, some bacteria utilize functional polarity for the subcellular localization of certain molecules to a particular area of bacterium. In most bacteria the cell wall plays a critical role in generating polarity, which poses a problem for the members of the Mycoplasma genus, whose members lack cell walls. Despite reductive evolution from gram-positive bacteria, mycoplasmas harbor a wide variety of morphological and functional complexity. Polarized species of mycoplasmas utilize polar organelles conferred by unique cytoskeletal elements utilized in attachment and motility, two processes associated with pathogenic species. One of these, Mycoplasma penetrans, is a potential opportunist usually isolated from human immunodeficiency (HIV)-infected individuals, proposed to play a role in AIDS progression. Previously, it has been demonstrated that M. penetrans attach to epithelial cells by a polar tip structure, yet nothing is known about the components involved in the architecture and functioning of the tip structure. We used time-lapse microcinematography to characterize gliding motility in both M. penetrans and its relative, Mycoplasma iowae, a poultry pathogen. Gliding speeds observed among strains in both species correlated positively with cytadherence differences by hemadsorption assay, suggesting that attachment and motility are both mediated by the same cellular components. During cell division, attachment to a surface and motility were critical for cytokinesis. We used scanning electron microscopy (SEM) and electron cyro-tomography to characterize the internal cellular organization and of the Triton X-100 (TX)-insoluble, cytoskeletal structure underlying the terminal organelle. The results from the different microscopy techniques are consistent with the cytoskeletal structure being artifactually dehydrated during processing for SEM, suggesting the cytoskeletal structure to be a proteinaceous gel found in varying amounts at both cell poles. Based on the importance motility plays in proper cell division, and the internal composition and organization of M. penetrans, we propose a cell cycle model in which a pole develops by accumulating enough TX-insoluble material in order to function in attachment and/or motility, and therefore power cytokinesis. To further identify the components of the TX-insoluble structure, we sought to identify proteins enriched in the TX-insoluble fraction compared to whole cells. We identified two proteins, MYPE1560 and MYPE1570, two of six proteins with similar features found in a putative transcriptional unit. Our observations of morphology and organization of M. penetrans and the TX-insoluble material demonstrate a novel cellular organization for generating polarity. We observed continued motility in the presence of an ATP depletion agent, a proton motive force inhibitor, and a sodium motive force inhibitor, raising the possibility that M. penetrans does not use chemically-derived energy source for motility. However, analysis of gliding under different temperature and pH conditions led us to conclude that thermal energy plays an important role in gliding motility. We propose a model for M. penetrans gliding motility wherein thermal energy is converted to unidirectional motility by means of a Brownian ratchet. Additionally, we provide evidence supporting the hypothesis that terminal organelles in distantly related mycoplasma species have evolved independently of each other, rather than diverging from a common ancestor. STRUCTURE, ORGANIZATION, AND FUNCTION OF THE TERMINAL ORGANELLE IN MYCOPLASMA PENETRANS

A DISSERTATION

Submitted to the faculty of Miami University in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Microbiology

Dominika Angelika Jurkovic Miami University Oxford, OH 2012

Dissertation Director: Dr. Mitchell F. Balish

TABLE OF CONTENTS

INTRODUCTION...... 1 A. Mycoplasma penetrans ...... 1 B. Bacterial polarity: morphology and function ...... 1 C. Polarity in the Mycoplasma genus ...... 2 D. Mycoplasma cytoskeletons ...... 6 E. Proposed mechanism of gliding motility in Mycoplasma genus ...... 12 F. Significance of studying M. penetrans ...... 15 G. Hypotheses ...... 15

CHAPTER 1: Conserved terminal organelle morphology and function in Mycoplasma penetrans and Mycoplasma iowae ...... 19 Abstract ...... 20 Introduction ...... 21 Materials and Methods ...... 22 Results ...... 24 Discussion ...... 42

CHAPTER 2: Structure and composition of the terminal organelle cytoskeleton of Mycoplasma penetrans ...... 45 Abstract ...... 46 Introduction ...... 47 Materials and Methods ...... 48 Results ...... 52 Discussion ...... 64

CHAPTER 3: Analysis of the energy source for Mycoplasma penetrans gliding motility... 76 Abstract ...... 77 Introduction ...... 78

Materials and Methods ...... 79 Results ...... 80 Discussion ...... 88 SUMMARY AND CONCLUDING REMARKS ...... 93 REFERENCES ...... 105

LIST OF TABLES

Table 1 Gliding motility parameters and Triton X-100 insoluble 26

structure dimensions of M. penetrans and M. iowae.

iii

LIST OF FIGURES

Figure 1 Cell cycle of Caulobacter crescentus. 4

Figure 2 Range of morphologies within the Mycoplasma genus. 8

Figure 3 Polarity-generating cytoskeletons of Mycoplasma genus. 10

Figure 4 Schematic of M. mobile gliding machinery. 14

Figure 5 Transmission electron microscopy of M. penetrans cells 17

attached to epithelial cells.

Figure 6 Consecutive phase-contrast images of M. penetrans. 28

Figure 7 Distribution of mycoplasma gliding velocities. 30

Figure 8 Colony HA of M. penetrans and M. iowae. 32

Figure 9 SEM of M. penetrans and M. iowae cells. 35

Figure 10 Time-lapse images of dividing M. penetrans cells. 37

Figure 11 Membrane and nucleoid staining in M. penetrans. 39

Figure 12 SEM of M. penetrans and M. iowae cells attached to glass, 41

treated with TX detergent.

Figure 13 Distribution of M. penetrans gliding speed. 54

Figure 14 Colony HA of M. penetrans strain HP88. 56

Figure 15 SEM of M. penetrans strain HP88 attached to glass. 59

Figure 16 Internal organization observed M. penetrans strain HP88. 61

Figure 17 Cytoskeletal filaments observed in M. penetrans str. HP88. 63

Figure 18 SDS-PAGE profiles and comparison of whole cell and 66

TX-insoluble proteins of M. penetrans strain HP88.

Figure 19 Protein characteristics of MYPE1560 and MYPE1570. 68

Figure 20 Putative transcriptional unit of M. penetrans cytoskeletal 70

protein-encoding genes.

Figure 21 Models of M. penetrans whole cell morphology observed 74

by ECT and SEM.

Figure 22 M. penetrans relative gliding speed in the presence and 83

absence of arsenate, CCCP, and amiloride.

Figure 23 Effect of sodium arsenate and sodium phosphate on the 85

growth of M. penetrans.

Figure 24 M. penetrans relative gliding speed under conditions 87

of varying temperature and pH.

Figure 25 Fitted response surface of M. penetrans motility generated 90

by the temperature and pH motility experiment.

Figure 26 Diversity of polarity-generating cytoskeletons of Mycoplasma 95

genus visualized by SEM.

Figure 27 Proposed model of cellular development in M. penetrans. 98

Figure 28 Proposed model of M. penetrans gliding motility. 102

ACKNOWLEDGEMENTS

Graduate school has been the most rewarding and challenging period of my life. My success wouldn’t be possible without the teaching, encouragement, patience, support, and friendship from my advisor, Dr. Mitchell Balish. Thank you for believing in and supporting me when others may have thought science was not for me. Thank you to my committee members, Dr. Luis Actis, Dr. Eileen Bridge, Dr. Gary Janssen, and Dr. David Pennock, for teaching me to accept the fact that, in research, I will never know all the answers. I’d like to thank my undergraduate research advisor, Dr. Steven Spilatro, for opening my eyes to a career in research and providing me with my first research experience. Thank you to my fellow Balish lab members, Jenn, Ryan, Rachel, Steven, Natalie, and Nicole, who made the lab a great place to be – even if my experiments were failing, horribly. I’m thankful to Jacqueline, whose support and friendship was invaluable when going through the ups and downs of graduate school and life. I’d like to thank my walking buddy and fellow Pioneer, Racheal, our hikes throughout Oxford were very therapeutic during the writing process and I hope I can return the favor as you approach the light at the end of the tunnel. To my other walking buddy and lab twin Natalie, thank you for your unconditional friendship and reminding me that is okay to be passionate about more than just science. To the entire Department of Microbiology, thank you for the camaraderie, support and knowledge you have given me. To my family, thank you from the bottom of my heart. To my parents, Ivan and Andjelka, thank you for working tirelessly to make sure your children could accomplish their dreams. Your sacrifices and support have never gone unnoticed and I am so blessed to have parents like you. To my sister, Viktoria, I am so glad we are sisters, because I don’t know how I would have survived some of these experiences without my best friend by my side. Thank you for being my pseudo-psychologist and life coach, even when you had a huge project due the next morning or an early day at work. To my brother, Michael, thank you for having patience with me when I act too much like a parent, so you can learn from my experiences, rather than just being your sister. Know that I only want my brother and sister to succeed as much as you two want the same for me. Even if you don’t want to embrace your “nerdiness,” thank you for loving science, so I’m not the only nerd in the family. Thank you, all, for helping me. I couldn’t have done this without you!

INTRODUCTION A. Mycoplasma penetrans The Mycoplasma genus of bacteria is comprised of both pathogens and commensals of vertebrate hosts, including humans. One of these species, Mycoplasma penetrans, is a potential opportunist, usually isolated from human immunodeficiency virus (HIV)-infected individuals, which is proposed to play a role in the progression of AIDS. M. penetrans str. GTU-54-6A1 was isolated from the urine of an HIV-infected patient (Lo et al., 1991) and most research on this strain has focused on its association with AIDS. Among HIV-positive homosexual males, 20-40% are seropositive for M. penetrans, and those patients may exhibit an enhanced progress of AIDS (Wang et al., 1992). Although the role of M. penetrans in progress of AIDS in unknown, it is believed to be a cofactor in its progression by stimulating the replication and expression of HIV (Shimizu et al., 2004). M. penetrans str. HF-2, isolated from the respiratory tract and blood of an HIV-negative infected patient with antiphospholipid syndrome, was used for the M. penetrans genome sequence project (Sasaki et al., 2002). M. penetrans is likely to be an opportunist, suggesting that it could potentially affect a variety of immunocompromised individuals. Additionally, studying M. penetrans can also provide a useful mechanism for controlling the progress of AIDS, which continues to spread worldwide. Understanding how M. penetrans attaches to and moves within its host will help in the development of therapeutic agents that may slow down the progression of AIDS while research continues. Additionally, deriving a more complete picture of how mycoplasmas generate, maintain, and use polarity will help increase knowledge of bacterial polarity and cytoskeletons. B. Bacterial polarity: morphology and function The control of bacterial cell shape is critical for nutrient acquisition, cell division, and protection from hostile environments ensuring survival of the bacteria (Young, 2006). Attachment, motility, and internal and external differentiation are all closely linked to cell shape (Young, 2006). Of the many shapes found among prokaryotic cells, some groups of bacteria have asymmetric, or polar, morphology (Brown et al., 2011). However, polarity is not limited only to shape, as cells utilize functional, or molecular, polarity for certain activities. Functional polarity may take the form of the subcellular localization of certain molecules, and may involve cytoskeletal elements, which also may contribute to generating morphological polarity. As an example, the life cycle of Caulobacter crescentus depends on both morphological and functional polarity. Events that

1 occur during the asymmetric division of C. crescentus establish the proper localization of proteins to distinguish its cell poles for the proper development of its flagellum, stalk, and pili, resulting in an adherent stalked cell and a motile swarmer cell at the end of division (Figure 1) (Ausmees and Jacobs-Wagner, 2003). Before division, specific proteins responsible for polar localization of the flagellum, pili, and stalk each use cues from the septum and the cell poles for proper localization (Kirkpatrick and Viollier, 2012). Bacterial cytoskeletal structures help to organize the cell by ensuring proper localization of proteins. MreB, a homolog of the eukaryotic cytoskeletal protein actin, is responsible for the rod shape of cells by directing peptidoglycan synthesis to the side walls during growth (White and Gober, 2012), giving rise to poles with distinct curvature, a cue recognized by the protein PopZ, which serves as a target for proteins associated with chromosome segregation and stalk biosynthesis (Ebersbach et al., 2008). The cell division protein FtsZ, a tubulin homolog, not only provides the force for cell constriction during division but also directs the activation of the peptidoglycan synthesis machinery at the division septum (Adams and Errington, 2009) and provides a target for proteins involved in flagellar biosynthesis following division (Kirkpatrick and Viollier, 2012). Crescentin, a homolog of eukaryotic intermediate filaments, mediates the curved cellular morphology characteristic of C. crescentus by restricting the amount of peptidoglycan added to one side of the cell (Cabeen et al., 2009); this is important for guiding motile cells to a surface for adherence, an essential feature of this organism's nutrient acquisition. C. Polarity in the Mycoplasma genus The contribution of the cytoskeleton and the cell wall in providing cues utilized by proteins that generate polarity presents a problem for members of the genus Mycoplasma, which lack a cell wall as a result of reductive evolution from gram-positive bacteria. Yet, despite their small genome size, a wide variety of morphological and functional complexity is present in this genus. Polarized species of mycoplasmas, which include important pathogens like Mycoplasma pneumoniae, Mycoplasma genitalium, and Mycoplasma gallisepticum, have polar organelles that contain unique cytoskeletal elements. These elements help to mediate attachment and motility, two processes that play key roles in establishing a host infection. Members of the genus Mycoplasma are pleomorphic, due to the lack of a cell wall, and exhibit a variety of morphologies (Figure 2) ranging from coccus-shaped (Mycoplasma gallinarum; J. Norton & M. Balish, unpublished) to rod-shaped (Mycoplasma insons; May et al.,

Figure 1: Cell cycle of Caulobacter crescentus. TipN (blue) localizes to the pole opposite the pili and flagellum in a swarmer cell. As the transition from swarmer cell to stalked cell occurs, TipN serves a polar landmark for the proper localization of TipF (yellow) and PflI (purple), proteins that regulate flagellar assembly and positioning, respectively. MreB (green) has a dynamic organization that alternates between a spiral-like and ring-like confirmation. At cytokinesis, TipN relocalizes to the divisome, remaining localized with FtsZ (red) so it can mark the new pole after division. To maintain its curved morphology, crescentin (orange) compresses the cell wall causing reducing the strain between peptide bridges of the cell wall. The reduction in strain is distributed in a gradient causing the cell curved morphology by distributing new cell wall growth along the gradient. Adapted from Mignot and Shaevitz, 2008.

2007) to the flask-shaped (M. pneumoniae and Mycoplasma mobile; Bredt, 1968; Kirchhoff and Rosengarten, 1984). The rod-shaped M. insons is an example of functional, rather than morphological polarity, as cells glide in a unidirectional manner (Relich et al., 2009), likely as a result of adhesins localized to one otherwise indistinct pole. The flask-shaped morphologies of M. pneumoniae and its phylogenetic relatives, as well as M. mobile, are conferred by the presence of a polarized terminal organelle (Seto & Miyata, 2003; Nakane & Miyata, 2007). The terminal organelle is a prosthecal structure continuous with the cell body, similar to the stalk of C. crescentus, rather than a proteinaceous appendage, such as flagella or pili in many other bacterial species. In the human pathogen M. pneumoniae, the terminal organelle mediates both cytadherence and gliding motility. Analysis of M. pneumoniae mutants with impaired attachment and/or gliding has led to the identification and characterization of several proteins localized to the M. pneumoniae attachment organelle, including the adhesins P1 and P30. The P1 adhesin (Feldner et al., 1982; Hu et al., 1982; Krause et al., 1982) is required for adherence and virulence demonstrated by avirulence of M. pnuemoniae mutants that fail to localize P1 properly. (Baseman et al., 1982). The P30 adhesin (Baseman et al., 1982; Morrison-Plummer et al., 1986) is responsible for both adherence and motility, demonstrated by the lack of either function in the mutant M. pneumoniae II-3. This mutant demonstrates profound pleomorphic morphologies (Romero-Arroyo et al. 1999; Hasselbring et al., 2005) that appear to result from ineffective cell division, suggesting a relationship between proper attachment organelle functioning and division. In the fish-associated M. mobile, gliding motility and adherence are mediated by its polar head-like structure. M. mobile exhibits the fastest gliding speed observed among mycoplasmas; it has no rest periods, and an average velocity of 2.0-4.0 µm/s (Rosengarten and Kirchoff, 1987; Miyata et al., 2002), which is ten-fold faster than the speed observed in M. pneumoniae (Bredt, 1968). Gliding-deficient mutants were used to determine the functions of the proteins localized to the neck region of the M. mobile head-like structure (Uenoyama & Miyata, 2005). Gli349 was determined to be an adhesin based on its role in hemadsorption and glass binding, similar to the M. pneumoniae P1 adhesin, although the two proteins do not share any homology (Uenoyama et al., 2004). Within the Mycoplasma genus, most lineages do not have a terminal organelle. Including the Mycoplasma muris cluster, whose members include M. penetrans, five lineages have

5 terminal organelles with four of the lineages having structurally and compositionally distinct terminal organelles. Two possibilities exist to explain the evolution of terminal organelles, divergent evolution from a common ancestor with terminal organelles, or convergent evolution from a common ancestor without a terminal organelle. In a divergent evolution scenario, two events would have to occur, the independent loss of a terminal organelle in multiple lineages and replacement of the original terminal organelle components in either all or all but one of the lineages with terminal organelles. Alternatively, in a convergent evolution scenario, independent gains of terminal organelles in a small number of lineages would result in the formation of terminal organelles that look similar and function similarity, but have different components and mechanisms. A convergent evolution model is more parsimonious if terminal organelles from different mycoplasma lineages are truly distinct from each other. The work presented in this dissertation test whether the terminal organelles in species from the M. muris cluster are compositionally and mechanistically distinct from terminal organelles found in other polarized lineages. D. Mycoplasma cytoskeletons Detergent extraction and visualization of cells by electron microscopy has identified three types of mycoplasma cytoskeletons, the electron-dense core of M. pneumoniae and its phylogenetic relatives, the jellyfish cytoskeletal structure of M. mobile, and a set of parallel, anastomosing filaments in M. insons (Figure 3). M. pneumoniae and M. mobile each have unique cytoskeletal structures that mediate the proper development of their terminal organelles, the attachment organelle and head-like structure, respectively. M. pneumoniae and M. mobile each have unique cytoskeletal structures that provide the framework by which they maintain their polar morphology and development of their terminal organelles. In M. pneumoniae, the electron-dense core within the attachment organelle is the cytoskeletal structure that confers polar morphology. The core contains three visually distinct substructures, the terminal button, rod, and the base (Hatchel and Balish, 2008). In M. mobile, polar morphology is conferred by the jellyfish-like structure within the polar head-like structure, which consists of a bell located at the attachment organelle tip, and multiple tentacle-like filaments extending into the cell body (Nakane and Miyata, 2007). The two species have very different cytoskeletal structures associated with polarity, with no homology in the protein components of either the electron-dense core or the jellyfish structure. The cytoskeleton of M. insons is unique among mycoplasmas in that it

Figure 2: Range of morphologies within the Mycoplasma genus. Mycoplasmas display diversity of cellular morphologies including spherical, rod-shaped, and cells with obvious polar organelles. A, M. gallinarum (J. Norton & M. Balish, unpublished), B, M. insons (adapted from Relich et al., 2009), C, M. pneumoniae, and D, M. mobile (adapted from Balish, 2006). Black arrow, attachment organelle; White arrow, head-like structure. Scale bar, 250 nm.

Figure 3: Polarity-generating cytoskeletons of Mycoplasma genus. Mycoplasmas utilize cytoskeletal structures varying in size, organization, and composition to generate polarity. A, M. insons (Relich et al., 2009), B, M. pneumoniae (Hatchel & Balish, 2008), and C, M. mobile (Nakane & Miyata, 2007). Scale bar, 250 nm.

10 extends the length of the entire cell, distinguishing it from the cytoskeletons observed in other mycoplasmas (Relich et al., 2009), which are concentrated at the attachment organelle. The components of the filaments have not yet been identified, nor has it been determined whether phylogenetic relatives of M. insons share similar cytoskeletal organization. In M. pneumoniae, the combination of visualization of the electron-dense core (Hatchel and Balish, 2006) and localization studies of cytadherence-associated proteins has resulted in a better understanding of how the proteins of the cytoskeleton are organized. The terminal button, associated with the most distal portion tip of the terminal organelle, is proposed to contain proteins HMW3 (Stevens and Krause, 1992) and P65 (Balish and Krause, 2005). The rod is suggested to be predominantly composed of the alpha-helical coiled-coil protein HMW2 (Balish et al., 2003), stabilized by protein HMW1 (Balish et al., 2001), and is composed of two parallel substructures (Henderson and Jensen, 2006). The base of the electron dense core contains proteins P24, P28, and P41 (Krause et al., 1997; Jordan et al., 2001; Seto and Miyata, 2003). Protein P200 is part of the attachment organelle and is proposed to be required for gliding and host colonization (Jordan et al., 2007). TopJ is proposed to function as a co-chaperone along with the ubiquitous chaperone DnaK for proper attachment organelle assembly (Cloward and Krause, 2010). Although the specific functions of all the cytoskeletal proteins remain unknown, without the proper functioning of all the proteins during assembly of the core, the attachment organelle does not form properly. As a result, adhesins fail to localize to the tip of the cell for efficient attachment and gliding motility to occur, thus rendering mutants avirulent (Balish and Krause, 2002; Seto and Miyata, 2003; Szczepanek et al., 2012 ). In M. mobile, proper formation of the jellyfish cytoskeleton, within the head-like structure, depends on proper functioning of its gliding proteins located at the cell neck near the cell membrane. Gli123, the mount protein (Uenoyama & Miyata, 2007), is required for proper location of both the leg protein Gli349, and the gear protein Gli521 (Seto et al., 2005). Gli349 adhesin are the “spikes” or “legs” observed when attached to a glass surface by rapid-freeze-and- fracture electron microscopy (Miyata & Peterson, 2004). These gliding proteins are crucial for functional polarity by mediating attachment during motility and are also required for the proper formation of the cytoskeleton, making them important in establishing the polar cellular morphology of M. mobile (Nakane & Miyata, 2007). Transmission electron microscopy of M. mobile cells treated with Triton X-100 reveal the bell and tentacle components of the

11 cytoskeleton (Nakane & Miyata, 2007). The bell of the jellyfish is a hemispherical structure located at the tip of the cell, connected to dozens of filaments located at the neck region that extends partially into the cell body (Nakane & Miyata, 2007). Six candidate proteins have been identified in the Triton X-100-insoluble fraction of M. mobile cells, and three have been observed to localize to the jellyfish structure, MMOB1630 and MMOB8460 at the bell and MMOB1670 within the tentacles. Six of the identified Triton X-100-insoluble proteins are coded tandemly in a genetic locus, suggesting that the entire locus may be responsible for the jellyfish structures (Nakane & Miyata, 2007). Although much work remains toward identifying the roles of the candidate proteins, much more is understood regarding a possible motility mechanism in M. mobile than in M. pneumoniae. E. Proposed mechanism of gliding motility in Mycoplasma genus Although many components and functions remain undiscovered in the polarity-driving mechanisms in the Mycoplasma genus, two mechanisms for motility have been proposed. The first, more speculative model is the inchworm model for M. pneumoniae. In this model, it is proposed that the conformational changes observed by electron cryotomography of M. pneumoniae may propel cells forward in an inchworm-like manner (Henderson and Jensen, 2006). The second model is the centipede model for M. mobile (Figure 4), which is better elucidated and supported by more experimental data. The centipede model requires a gliding mechanism that depends on ATP hydrolysis for the repeated binding of proteins to a substrate (Miyata, 2010). The effects of various drugs on M. mobile gliding have suggested that gliding mechanism of M. mobile to be coupled to ATP hydrolysis as cells are unable to move in the presence of arsenate, a known inhibitor of ATP hydrolysis (Jaffe et al., 2004). The addition of ATP directly to cells rendered immobile by Triton X-100 detergent permeabilization, causes the reactivation of motility to normal speeds, suggesting strongly that ATP directly powers the motor (Jaffe et al., 2004). The experimental data is further supported by the identification of a novel ATPase found in the same operon as in the gliding proteins (Ohtani & Miyata, 2007). In the centipede model, M. mobile is propelled by 400 “legs” composed of Gli349, which repeatedly bind and rerelease from sialic acid on surfaces in vtiro (Miyata, 2010). After Gli349 binds tightly to the surface, the force applied by the adhesin triggers a conformational change in the arm of the leg, resulting in a pull which uses ATP produced by the ATPase, P42, transmitted by Gli521 (Miyata, 2010). The continual movement of the cell caused by the pull of the other

Figure 4: Schematic of M. mobile gliding machinery. P42 protein utilizes ATP hydrolysis to produce energy (A) that is predicted to be transmitted to the gliding leg protein Gli349 through Gli521 (B). The sequential attachment-and-release cycles of Gli249 (C), pulls cells across a sialyl oligosaccharide-coated surface. Adapted from Miyata, 2010.

14 legs on the cell pulls the gliding machinery forward and the uppermost portion of the leg protein back to its initial confirmation. Once the leg has returned to its original conformational state, Gli349 can once again tightly bind substrate, repeating this process. Significant roles of adhesins P1 (Seto et al., 2005) and P30 (Hasselbring et al., 2005) in M. pneumoniae motility suggest that a related mechanism might occur in that species as well, possibly arguing against the inchworm model as a major mechanism for gliding. Experiments using α-P1 antibodies to moving M. pneumoniae cells reduced gliding and caused eventual detachment, while no effect is observed on nonmotile cells (Seto et al., 2005) The observation that α-P1 antibody can only inhibit attachment and motility of moving cells, indicates P1 is responsible for attachment during motility (Seto et al., 2005). Analysis of M. pneumoniae mutants lacking P30 do not glide, and mutants with truncated P30 exhibit reduced gliding speed, indicating P30 plays an important role in gliding. (Hasselbring et al., 2005). F. Significance of studying M. penetrans With most knowledge of cytadherence and gliding motility in mycoplasmas being based on observations from M. pneumoniae and M. mobile, it is important to study these processes in other pathogenic species of mycoplasmas. With M. penetrans, it is important to characterize cytadherence, which was previously observed to occur by a polar tip structure (Figure 5). Identifying the proteins involved in cytadherence and motility will allow for characterization of the mechanism(s) associated with these processes, as these processes almost certainly contribute to virulence in their host. The studies in this dissertation utilize microscopy, biochemistry, and cell biology to characterize cytadherence and gliding motility of a species that has not been studied in this capacity. This is partially due to the lack of molecular tools available for genetic manipulation of this organism. These studies will enhance our understanding of cytadherence and motility of mycoplasmas and elucidate the molecular mechanisms of these processes specifically in M. penetrans. With such diversity among motile mycoplasmas, it is becoming apparent that it is important to study more species individually and to understand what are likely to be varied mechanisms of adherence and motility, rather than focusing on a poorly representative model species for cytadherence and motility in all species. G. Hypotheses We propose that the mechanisms used to mediate cytadherence and gliding motility have evolved independently among different phylogenetic groups in the Mycoplasma genu; therefore,

Figure 5: Transmission electron microscopy of M. penetrans cells attached to epithelial cells. (Y. Sasaki; http://www0.nih.go.jp/Mypet/) Arrows, polar tip structure.

17 the mechanisms used to carry out the functions mediated by mycoplasma attachment organelles most likely vary species to species. To test whether the polar organelles of M. penetrans and the closely related avian pathogen Mycoplasma iowae function similarly to attachment organelles in better understood species, we characterized cytadherence and gliding motility in these species by hemadsorption and time-lapse microcinematography using different strains of M. penetrans and M. iowae. To better characterize the organization of the polar organelle in these two species, scanning electron microscopy of whole cell and Triton X-100 treated cells was used to compare the morphology and cytoskeletal structures of the M. penetrans and M. iowae cells. Since there is a relationship between motility and cell division is some mycoplasmas, we also observed cell division in M. penetrans to determine whether the polar organelle is important for this function. To determine whether M. penetrans gliding motility is mediated by mechanisms used by M. pneumoniae or M. mobile, we sought to identify the energy source required for M. penetrans motility. We observed the effect of temperature and/or pH, as well as known inhibitors of ATP hydrolysis, sodium motive force, and proton motive force on M. penetrans gliding speed. Finally, to test whether the components of the M. penetrans cytoskeleton are novel compared to members of the M. pneumoniae cluster and M. mobile, we characterized the organization of the M. penetrans internal structure by electron cryotomography. To identify the components of the M. penetrans cytoskeleton, we analyzed proteins enriched in the Triton X-100-insoluble fraction compared to whole cell lysates to identify the cytoskeletal proteins. The data obtained from these experiments will further our understanding of mycoplasma terminal organelle evolution. Furthermore, understanding how the attachment organelle of M. penetrans mediates cytadherence and motility can be valuable in understanding how this potential opportunistic human pathogen causes infection, and could lead to the development of more efficient therapeutic strategies in the future.

CHAPTER 1

Conserved terminal organelle morphology and function in Mycoplasma penetrans and Mycoplasma iowae

Dominika A. Jurkovic, Jaime T. Newman, and Mitchell F. Balish Journal of Bacteriology. 2012. 194:2877-2883

ABSTRACT Within the genus Mycoplasma are species whose cells have terminal organelles, polarized structures associated with cytadherence and gliding motility. Mycoplasma penetrans, found mostly in HIV-infected patients, and Mycoplasma iowae, an economically significant poultry pathogen, are members of the Mycoplasma muris phylogenetic cluster. Both species have terminal organelles that interact with host cells, yet the structures in these species, or any in the M. muris cluster, remain uncharacterized. Time-lapse microcinematography of two strains of M. penetrans, GTU-54-6A1 and HF-2, and two serovars of M. iowae, K and N, show the terminal organelles of both species to play a role in gliding motility, with differences in speed within and between the two species. The strains and serovars also differed in their hemadsorption ability that positively correlated with differences in motility speed. No morphological differences were observed between M. penetrans and M. iowae by scanning electron microscopy (SEM). SEM and light microscopy of M. penetrans and M. iowae showed the presence of membranous filaments connecting pairs of dividing cells. Breaking of this filament during cell division was observed for M. penetrans by microcinematography and suggests a role for motility during division. The Triton X-100-insoluble fraction of M. penetrans and M. iowae consisted of similar structures that were unique compared to those identified in other mycoplasma species. Like other polarized mycoplasmas, M. penetrans and M. iowae have terminal organelles with cytadherence and gliding functions. The difference in function and morphology of the terminal organelles suggests mycoplasmas have evolved terminal organelles independently of one another.

INTRODUCTION Reductive evolution of the bacterial genus Mycoplasma from gram-positive bacteria has resulted in cell wall-less organisms with both reduced genome size and reduced biosynthetic ability. This genus includes both commensals and pathogens and is associated with vertebrate hosts, including humans. For many pathogenic mycoplasmas, a first step in causing disease is adherence to host cells, often by adhesins associated with a polarized organelle (Balish, 2006). The interactions of mycoplasmas with host cells are diverse and complex. Mycoplasma pneumoniae, a human pathogen, requires a polar terminal organelle for cytadherence and gliding motility (Balish, 2006), which may be required for colonization of host tissue (Jordan et al., 2007). Mycoplasma mobile, a fish pathogen, also has a terminal organelle, but with distinct morphology as compared with M. pneumoniae (Balish, 2006). Although both species have a terminal organelle, the proteins involved in biogenesis and function of these polar structures are completely different. For example, M. pneumoniae adhesins P1 and P30 are located at the tips and/or sides of the terminal organelle (Seto and Miyata, 2003), whereas the M. mobile adhesin Gli349 is found at the base, or neck, of its terminal organelle (Uenoyama et al., 2004). The two species also have very different gliding characteristics (Jaffe et al., 2004a; Radestock and Bredt, 1977), and distinct internal organizations of their respective polar organelles (Biberfeld and Biberfeld, 1970; Nakane et al., 2007). Mycoplasma penetrans and Mycoplasma iowae are both found in the Mycoplasma muris cluster within the pneumoniae group. M. penetrans is found commonly in the urogenital tract of HIV-positive individuals (Lo et al., 1992; Wang et al., 1992), but has also been isolated from the blood of a non-HIV-infected individual with antiphospholipid syndrome (Yanez et al., 1999). Previous research has focused on the role of M. penetrans in AIDS. Its lipoproteins are mitogenic (Feng et al., 1994) and stimulate transcription of the HIV genome in HIV-infected cells in vitro via toll-like receptors (Shimizu et al., 2004), suggesting that M. penetrans may play a role in expediting the progression of AIDS. M. iowae is a poultry pathogen isolated from the respiratory tract of turkeys and chickens (Yoder et al., 1964) and the gastrointestinal tract of turkeys (Mirsalimi et al., 1989). In experimental infections, M. iowae causes embryo death in turkey (Yoder et al., 1964) and chicken eggs (Bradbury et al., 1983). M. iowae commercial infections have been identified in turkey poults, resulting in a variety of leg abnormalities (Ley et al., 2010, Trampel et al., 1994), which can have a negative impact upon meat production.

Both M. penetrans and M. iowae exhibit polarity, with a terminal polar organelle mediating adherence to human epithelial cells (Lo et al., 1992) and colonization of intestinal mucosa in turkeys (Mirsalimi et al., 1989), respectively. The proteins involved in the organization and function of the polar organelle remain unknown in both M. penetrans and M. iowae; the M. penetrans genome lacks homologs of genes involved in terminal organelle structure in M. pneumoniae and M. mobile (Himmelreich et al., 1996, Jaffe et al., 2004a ; Sasaki et al., 2002). Electron microscopy reveals that both species have two distinct cytoplasmic components. The polar organelles of M. penetrans and M. iowae are densely packed with material of a fine granular structure, as compared with the cell body, whose appearance is more typical of bacterial cytoplasm (Lo et al., 1991; Mirsalimi et al., 1989). Although it is likely that these organisms utilize a polar organelle to mediate attachment and invade their respective hosts (Giron et al., 1996), it is unknown whether features typically associated with this structure in other polarized mycoplasmas, including gliding motility and an internal cytoskeleton, are present. Improved understanding of the mechanisms that M. penetrans and M. iowae use to interact with host cells is hampered by the lack of a genetic manipulation system for use in M. penetrans, and the lack of a completed genome sequence for M. iowae. Despite the availability of a genome sequence for M. penetrans, it is unclear which proteins are associated with its terminal organelle. Reasoning that, like in other mycoplasmas, polarity in M. penetrans and M. iowae might be associated with gliding motility and a cytoskeletal structure, and that the elements responsible for these properties are likely to be novel, we investigated strains of M. penetrans and M. iowae by detergent treatment, microscopy, and microcinematography. Both M. penetrans and M. iowae were found to be motile, but with properties that distinguish both organisms from other mycoplasmas. In addition, they share novel, detergent-insoluble structures consistent with a terminal organelle-associated cytoskeletal element. This study expands the known diversity of mycoplasma cell structure and provides insight into interactions between these two species and their respective host cells.

MATERIALS AND METHODS Bacterial culture conditions. Mycoplasma penetrans strains GTU-54-6A1 (Lo et al., 1992) and HF-2 (Yanez et al., 1999) and Mycoplasma iowae serovars N and K (Barber and Fabricant, 1971) were cultured at 37˚C in SP-4 broth (Tully et al., 1979) or on SP-4 agar plates.

Hemadsorption (HA). Plates containing mycoplasma colonies were incubated with sheep red blood cells (SRBC) in Alsever’s solution (Cleveland Scientific) and washed with phosphate- buffered saline as previously described (Hatchel et al., 2006). Colonies were observed and photographed at 400X magnification on a Leica DM IRB inverted phase- contrast/epifluorescence microscope. Time-lapse microcinematography. M. penetrans and M. iowae cells from frozen, mid-log phase stocks were filtered through a 0.45-µm filter and incubated 3 h at 37˚C in glass chamber slides (Nunc) in SP-4 broth supplemented with 3% gelatin. In each analysis, 27 images were captured at 1000X magnification on a Leica DMIRM inverted phase-contrast/epifluorescence microscope at intervals ranging from 0.5 to 1 s. Images were merged and analyzed as previously described (Hatchel et al., 2006). Gliding speeds of individual cells were calculated by dividing the distance traveled by the number of frames between which cells moved. For each strain, at least 100 motile cells were analyzed. Scanning electron microscopy (SEM). Cells were prepared for SEM as previously described (Hatchel et al., 2006). Briefly, they were grown for 6 h to 1 d in SP-4 broth supplemented with 3% gelatin at 37˚ C. To analyze Triton X-100-insoluble structures, Triton X-100 in 20 mM Tris- HCl, pH 7.5, 150 mM NaCl was added to whole cells to a final concentration of 1%, and the coverslips were incubated for 30 min at 37˚C. Coverslips were fixed 30 min in 1.5% glutaraldehyde, 1% paraformaldehyde, 0.1 M sodium cacodylate, pH 7.2, rinsed with buffer, and dehydrated through a series of ethanol washes from 25% to 100%. Following dehydration, the coverslips were critical point-dried and then viewed on a Zeiss Supra 35 FEG-VP scanning electron microscope. Membrane and DNA staining. To visualize cell membranes, cells were grown overnight in 24- well plates on glass coverslips in SP-4 broth supplemented with 3% gelatin at 37°C. Cells were stained with 3,3′-dihexyloxacarbocyanine iodide (DiOC6; Sigma-Aldrich) for 10 min at a final concentration of 5 µg/ml and then fixed directly in the media using the same fixative as for SEM. To visualize the nucleoid, cells were grown overnight. The media and fixative were removed and 4',6-diamidino-2-phenylindole (DAPI; Sigma Aldrich) solution was added to the coverslips for 30 min in the dark at a final concentration of 1 µg/ml. Cells were visualized using phase-contrast microscopy and DiOC6 and DAPI by epifluorescence using rhodamine and DAPI filter sets, respectively.

RESULTS Gliding speed and HA. Because M. penetrans and M. iowae attach via a polar structure (Lo et al., 1991, Mirsalimi et al., 1989), we tested whether these organisms exhibited gliding motility, which is characteristic of other mycoplasma species bearing similar structures (Balish, 2006). M. penetrans strains GTU-54-6A1 and HF-2 and M. iowae serovars N and K were all found to be motile, albeit at distinctly different speeds (Table 1). Consecutive phase-contrast images of M. penetrans gliding showed that cells moved in a unidirectional manner, led by one end of the cell (Figure 6). Cells moved in both straight and curved paths but never reversed direction. In both species, average gliding speeds were within the range of extremes previously observed for mycoplasma motility, < 30 nm/s for Mycoplasma pirum and Mycoplasma insons and > 2500 nm/s for Mycoplasma testudinis and Mycoplasma mobile (Hatchel et al., 2006; Kirchhoff and Rosengarten, 1984; Relich et al., 2009). On average, M. penetrans strain GTU-54- 6A1 glided 10 times faster than strain HF-2, and M. iowae serovar N glided 3 times faster than serovar K. Unlike other motile species in the pneumoniae group, both M. penetrans and M. iowae exhibited a large range in their average gliding speeds (Table 1; Figure 7); however, like M. insons, gliding frequency was low. Motility of M. pneumoniae ceases around the time of cytokinesis (Hasselbring et al., 2006), raising the possibility that the low percentage of moving M. penetrans and M. iowae cells might be attributable to engagement in the cell division process. Indeed, paired cells were frequently observed (see Figure 10); individual stationary cells might also have been near cytokinesis. Cells that were clearly dividing (see below) and aggregates or microcolonies of cells were neither analyzed for speed nor enumerated among the total number of cells analyzed. To test whether isolates differed in cytadherence, we performed HA assays on both M. penetrans strains and both M. iowae serovars. Both M. penetrans strains were HA-positive, as previously described for strain GTU-54-6A1 (Lo et al., 1992), but there was a consistent difference in the coverage of colonies by SRBC. GTU-54-6A1 colonies were completely covered by SRBC (Figure 8A), but HF-2 colonies were more sparsely covered, with most SRBC binding the colony periphery (Figure 8B). M. iowae serovars N and K were also both HA-positive, as previously described (Lo et al., 1991), and also exhibited differences in SRBC coverage. M. iowae serovar N colonies (Figure 8C) were completely covered with SRBC, whereas serovar K

Table 1: Gliding motility parameters and Triton X-100 insoluble structure dimensions of M. penetrans and M. iowae.

Width of Cells Greatest insoluble Length of Mean Range of moving width of structure insoluble Strain speed speeds per insoluble across structure (nm/s) (nm/s) frame structure narrow (nm) (%) (nm) portion (nm)

M. penetrans str. GTU-54-A1 548 ± 217 151-1284 23 449 ± 79 187 ± 60 127 ± 25

M. penetrans str. HF-2 54 ± 18 25 -111 30 ND ND ND

M. iowae N 797 ± 427 200-2150 13 442 ± 54 237 ± 71 109 ± 18

M. iowae K 270 ± 125 95-621 8 ND ND ND

ND= Not determined

Figure 6: Consecutive phase-contrast images of M. penetrans motility. Images of M. penetrans incubated in a chamber slide at intervals of 5-6 s at 37˚C. Four representative cells are indicated by arrows. In the first frame (0 s), the carets point to the leading ends of gliding cells. In the last frame, the paths of the gliding cells are represented, with arrows indicating the direction of gliding during the observation period. Bar: 1 µm.

Figure 7: Distribution of mycoplasma gliding velocities about the mean. Gliding speeds of the individual motile cells for each strain or serovar were grouped into bins of 0.2x the mean, designated x, for each species. The percentage of cells with speeds in a given range out of the total population of motile cells was plotted against the speed as a fraction of the mean speed for each isolate. A, M. iowae. Black bars, serovar K, grey bars, serovar N. B, M. penetrans. Black bars, strain GTU-54-6A1, grey bars, strain HF-2.

Figure 8: Colony HA of M. penetrans and M. iowae. Mycoplasma colonies on SP-4 were incubated with SRBC for 30 min at 37°C, washed with warmed PBS, and visualized at 40X to observe coverage. A, M. penetrans strain GTU-54-6A1. B, M. penetrans strain HF-2. C, M. iowae serovar K, and D. M. iowae serovar N. Arrow, individual SRBC.

32 colonies were only covered with SRBC at the colony periphery (Figure 8D). Interestingly, for both species it was the faster-gliding variety that exhibited greater coverage by SRBC, but all isolates were nonetheless HA-positive. Morphology and internal organization of Mycoplasma penetrans and Mycoplasma iowae. The cell morphologies of glass-attached M. penetrans and M. iowae cells were observed by SEM. M. penetrans (Figure 9A, 9B) and M. iowae (Figure 9C, 9D) had a similar appearance. Though pleomorphic, cells consisted of an attachment organelle and a cell body. In both species, paired cells were frequently connected by a filament that was continuous with the cell membrane (data not shown) which varied considerably in length. Unfiltered cells had a propensity to form microcolonies, consistent with the aggregates observed during time-lapse microcinematography. To test whether the filaments connecting cell bodies represented a cytokinesis intermediate, we observed the division of paired M. penetrans strain GTU-54-6A1 cells under the same conditions used for analyzing gliding motility (Figure 10). The filament routinely increased in length and became thinner as the two cells moved away from each other. Eventually the filament snapped, irrespective of either the passage of a constant amount of time or achievement of a particular length. After division, only one cell glided immediately, whereas the second cell remained stationary for a period of time, as was previously observed for M. pneumoniae (Hasselbring et al., 2006). Staining of cells with DAPI revealed that the nucleoid was present in the cell body but absent from the terminal organelle and from the filaments connecting paired cells (Figure 11). The edge of the nucleoid adjacent to the terminal organelle was very flat, suggesting the presence of a physical boundary. The average length of the nucleoid-free space in the region of the terminal organelle was 470 ± 60 nm and the average width adjacent to the nucleoid was 310 ± 80 nm (n=32). Triton X-100-insoluble structure of M. penetrans and M. iowae. Since M. penetrans and M. iowae are both motile and share similar cell morphology, we treated cells with Triton X-100 to test for the presence of a cytoskeletal structure such as those found in the terminal organelles of M. pneumoniae (Meng et al., 1980) and M. mobile (Nakane et al., 2007). Treatment of both M. penetrans (Figure 12A) and M. iowae (Figure 12B) with 1% Triton X-100 revealed that both species harbored cylindrical or pear-shaped objects, studded by projections, which dominated the images. These structures were indistinguishable between strains within species, although

Figure 9: SEM of M. penetrans and M. iowae cells attached to glass. A, M. penetrans strain GTU-54-6A1; B, M. penetrans strain HF-2; C, M. iowae serovar K; D, M. iowae serovar N. White arrows, filamentous structures connecting individual cell bodies; black arrows, attachment organelles. Scale bar, 1 µm.

Figure 10: Time-lapse images of dividing M. penetrans str. GTU-54-6A1 cells. Seconds since leftmost panel are indicated in the upper right corner of each panel. Individual members of each dividing pair are numbered in the first and last panels. Arrows, filamentous structures prior to division. Scale bar, 1 µm.

Figure 11: Membrane and nucleoid staining in M. penetrans strain GTU-54-6A1. A. DiOC6 fluorescence; B, corresponding phase-contrast image. Arrows, filamentous structures connecting individual cell bodies. C. White DAPI fluorescence is overlaid onto a phase-contrast image of a pair of dividing cells. Arrows, connecting filament; arrowheads, terminal organelle lacking DNA. Scale bar, 1 µm.

Figure 12: SEM of attached M. penetrans and M. iowae cells treated with TX detergent. A, M. penetrans str. GTU-54-6A1; B, M. iowae serovar N. Scale bar, 1 µm.

41 somewhat distinctive between the two species. There was no significant difference in length of the most consistent structures between the two species, at 440-450 nm (Table 1), corresponding to that of the nucleoid-free space in the terminal organelle (see above). Because of the shape of the insoluble structures, width measurements were taken at both the widest point and the point at which the structure begins to narrow. For both width measurements, there was a statistically significant difference (one-way ANOVA, p < 0.05; n = 25) between the two species, with M. iowae exhibiting greater extremes (Table 1). In addition to these predominant structures, both species also contained Triton X-100-insoluble structures that were distinctly shorter or longer, which might represent intermediates in their assembly.

DISCUSSION Ultrastructure. The role of terminal organelles in mediating both cytadherence and gliding motility has been characterized in several members of the Mycoplasma genus (Hatchel et al., 2008; Miyata and Uenoyama, 2002). M. penetrans and M. iowae are the first motile species identified in the M. muris cluster, and increase the number of described motile mycoplasma species to twelve, eleven of which have terminal organelles. Despite the close relationship between the M. muris and M. pneumoniae clusters, the overall ultrastructure of the M. iowae and M. penetrans terminal organelles differ substantially compared to M. pneumoniae. The novel Triton X-100-insoluble structures of M. penetrans and M. iowae represent a third type of cytoskeletal organization associated with mycoplasma terminal organelles, with M. pneumoniae and M. mobile representing the first two. The length of the Triton X-100-insoluble structure of M. penetrans is consistent with that of the nucleoid-free space of the terminal organelle, consistent with a model in which it fills that space. The width of the structure, on the other hand, is less than that of the nucleoid-free space, suggesting that either this structure is at the core of a larger structure whose exterior is solubilized during extraction, or that dehydration reduces the width artifactually. It is likely that the densely packed material within the M. penetrans polar organelle observed in thin sections (Lo et al., 1991; Neyrolles et al., 1998) is actually this Triton X-100-insoluble structure. Similar material is observed in electron micrographs of M. iowae polar organelles (Mirsalimi et al., 1989). The uniform appearance of the densely packed material in these images argues against a second type of structure that was not recovered in our experiments. Further studies are necessary to identify the proteins that comprise both structures as well as their development and function.

Gliding motility. The gliding characteristics of M. penetrans and M. iowae are distinct from those of other mycoplasma species in particular ways. One is the discontinuity of their movement as compared with M. mobile cells, which rarely rest during motility (Rosengarten and Kirchoff, 1987), and with species of the M. pneumoniae phylogenetic cluster, which rest more frequently than M. mobile (Radestock and Bredt, 1977), but not to the degree observed in M. penetrans and M. iowae. As a measure of this phenomenon, on average, individual M. penetrans strain GTU-54-6A1 cells moved for ~70% of the observation period (data not shown), whereas M. genitalium strain G37 cells moved for ~90% of the period (J.M. Hatchel and M.F. Balish, unpublished data), in both cases discounting cells that appeared to be impeded. A second distinctive characteristic is the existence of substantial gliding speed differences among isolates of M. penetrans and M. iowae, not observed in other species (data not shown). It is unclear whether the differences observed in gliding speed and coverage of colonies by SRBC in the HA assay have any relationship to the sites from which the organisms were isolated. M. penetrans str. GTU-54-6A1, isolated from the urogenital tract of HIV-infected individuals (Lo et al., 1991; Wang et al., 1992), glides faster than strain HF-2, which was isolated from the respiratory tract of an HIV-negative person with primary antiphospholipid syndrome (Yanez et al., 1999). Additionally, M. iowae serovar N, isolated from the air sac of turkeys (Barber and Fabricant, 1971), glides faster than serovar K, isolated from the oviduct of chickens. Further comparison of both strains and serovars among M. penetrans and M. iowae may be helpful in identifying the molecular components associated with adherence and motility in these two species. Cell division. In M. penetrans, we observed a clear relationship between motility and cell division. Dividing cells were connected by cell membrane filaments, which increased in length as the distance between both daughter cells increased over time. The straightness of filaments connecting the cell pairs suggests that the filaments experience tension, most likely the result of the gliding-associated forces generated by the terminal organelles of one or both cells of the pair. The lack of DAPI fluorescence in the filaments of paired cells suggests that the filaments are free of DNA and are formed after nucleoid segregation, in agreement with images in which non- filamentous individual cells appear to have two distinct nucleoids (not shown). We observed breakage of filaments at the culmination of cell division, one cell remaining stationary and the other gliding away. This observation is identical to observations of dividing M. pneumoniae cells

(Hasselbring et al., 2006). Like other mycoplasmas, M. penetrans may rely at least in part on the force generated by gliding motility for cell division (Lluch-Senar et al., 2010). Independent evolution of mycoplasma terminal organelles. Although terminal organelles in M. mobile, the M. pneumoniae cluster, and the M. muris cluster facilitate the same sets of functions, there are significant disparities in overall terminal organelle organization and protein composition in each of these groups. We believe that these data are irreconcilable with a model in which these structures spring from a common evolutionary origin. The observations herein regarding M. penetrans and M. iowae demonstrate not only the distinct characteristics of members of the M. muris cluster, but also that comparing more Mycoplasma species is critical for gaining a more complete understanding of both the unifying and diverse elements of terminal organelle function and its role in disease caused by any single species.

ACKNOWLEDGMENTS This work was supported by the National Institutes of Health (Public Health Service grant R15 A1073994) and by the Miami University Doctoral-Undergraduate Opportunities in Scholarship program. We gratefully acknowledge the following for providing mycoplasma strains: Z. Raviv (The Ohio State University, M. iowae serovars N and K); M. Davidson (Mollicutes Culture Collection, Purdue University, M. penetrans str. GTU-54-6A1); L. Duffy (University of Alabama-Birmingham, M. penetrans str. HF-2). We thank C. M. Fullem for early work on DAPI staining of M. penetrans. We are also grateful to D. Krause (University of Georgia) for reading the manuscript.

CHAPTER 2

Structure and composition of the terminal organelle cytoskeleton of Mycoplasma penetrans

Dominika A. Jurkovic1, Jian Shi2, Grant J. Jensen2, and Mitchell F. Balish1

1 Department of Microbiology, Miami University, Oxford, OH 45056 USA

2 Division of Biology, California Institute of Technology, Pasadena, CA 91125 USA

ABSTRACT Mycoplasma penetrans, detected mainly in HIV-infected individuals, has a terminal organelle that is utilized in adherence and gliding motility and contains a novel Triton X-100 (TX) insoluble cytoskeletal structure whose components are unknown. Characterization of hyperadherent, fast-gliding, strain HP88 revealed no significant differences in morphology of whole cells or cytoskeletons as compared to the parent strain, GTU-54-6A1. However, electron cryotomography (ECT) suggested the lack of a distinctive tip structure, making it possible that the tip structure observed by other techniques may be an artifact of dehydration of a gel-like matrix during preparation. The increased ability of str. HP88 to adhere to surfaces allowed for analysis of the proteins of the TX-insoluble fraction, leading to the identification of proteins MYPE1560 and MYPE1570 as two components. These acidic, coiled-coil proteins are the last two proteins encoded in a putative transcriptional unit containing a total of six genes for proteins with similar characteristics. The novel organization and composition of the M. penetrans cytoskeleton supports a hypothesis that dissimilarity of terminal organelles in different phylogenetic clusters of mycoplasmas reflects independent evolution of terminal organelles.

INTRODUCTION Some species within the cell wall-less Mycoplasma genus of bacteria utilize polarized tip structures to attach to host cells. Although not all mycoplasmas cause disease, the first step of virulence in pathogenic species often involves adherence to host cells, mediated by adhesins associated with a polarized tip structure (Balish, 2006). Mycoplasma pneumoniae, a human pathogen associated with atypical pneumonia and asthma (Atkinson et al., 2008), has a polar terminal organelle required for cytadherence and gliding motility (Balish, 2006), processes linked to dissemination during infections (Jordan et al., 2007). Mycoplasma mobile, a possible fish pathogen (Stadtlander and Kirchhoff, 1995), has a terminal organelle used for similar cytadherence and gliding motility, but with different adhesins, morphology, and cytoskeletal organization (Balish, 2006). Genomic sequences of the M. pneumoniae and M. mobile reveal an absence of homologs of characterized organelle proteins between the two species. M. pneumoniae adhesins P1 and P30 are localized to both the tip and the sides of the attachment organelle (Seto and Miyata, 2003), whereas the unrelated M. mobile adhesin Gli349 is found at the base of the terminal organelle (Uenoyama et al., 2004). Extraction of both mycoplasmas with Triton X-100 (TX) detergent reveals cytoskeletal structures that provide the scaffolding for formation of functional terminal organelles. However, the two species have very different structures, which is not surprising given the disparate cytoskeletal proteins that the two species use. The main organization of the M. pneumoniae attachment organelle consists of a striated rod with a distal terminal button and a proximal bowl structure (Henderson and Jensen, 2006; Seybert et al., 2006). In contrast, the M. mobile terminal organelle cytoskeleton consists of a cluster of material with numerous strands extending into the cell body (Nakane and Miyata, 2007). The differences in the cytoskeleton organization and composition suggest that these two species utilize different mechanisms to assemble polar tip structures used in cytadherence and gliding motility. Mycoplasma penetrans str. GTU-54-6A1 is a putative human pathogen isolated from the urogenital tract of HIV-positive individuals (Lo et al., 1991; Lo et al., 1992; Wang et al., 1992). Its lipoproteins are mitogenic toward B and T lymphocytes (Feng et al., 1994, Sasaki et al., 1995) and stimulate transcription of the HIV genome in HIV-infected cells in vitro via toll-like receptors (Shimizu et al., 2004), suggesting a role for M. penetrans in accelerating the progression of AIDS. Previously, we have shown M. penetrans to have a polar tip structure

47 similar to the M. pneumoniae attachment organelle and the M. mobile terminal organelle (Jurkovic et al., 2012). The polar tip structure of M. penetrans mediates both cytadherence and gliding motility, but with differences in the degree of adherence and the speed of movement between strains (Jurkovic et al., 2012). M. penetrans has a TX-insoluble cytoskeletal structure that is distinct from its M. pneumoniae and M. mobile counterparts (Jurkovic et al., 2012), suggesting yet another mechanism for generating mycoplasma polarity and mediating pole- associated functions. Progress in identifying the proteins that compose the novel cytoskeleton of M. penetrans has lagged in the absence of molecular tools available for genetic manipulation, despite the availability of a complete genome sequence (Sasaki et al., 2002). The M. penetrans genome reveals the absence of clear homologs of polar organelle-associated proteins of M. pneumoniae and M. mobile, which correlates with the different organization observed among the three species. We characterized the terminal organelle associated properties in a hyperadherent M. penetrans strain, HP88, by microscopy, microcinematography, and detergent treatment. M. penetrans str. HP88 was found to attach to sheep red blood cells (SRBC), glide two times faster than M. penetrans str. GTU-54-6A1, from which it was derived by passaging, and have a TX- insoluble structure similar to that in both M. penetrans strains HF-2 and GTU-54-6A1. We used M. penetrans str. HP88 to identify two major TX-insoluble proteins by matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS). This study expands our knowledge of the M. penetrans cytoskeleton and the known diversity of mycoplasma cytoskeletons, which will provide important information about the interactions between M. penetrans and host cells.

MATERIALS AND METHODS Bacterial culture conditions and passaging. M. penetrans strain HP88 was obtained through a series of passages of M. penetrans strain GTU-54-6A1 (Lo et al., 1992) in SP-4 motility media (SP-4 broth supplemented with 3% gelatin) to select for isolates with different cytadherence or motility characteristics. A 100-µL aliquot of M. penetrans str. GTU-54-6A1 was added to 2 mL of SP-4 motility media in a 24-well plate (TPP Techno Plastic Products AG). Upon a color change in the medium from red to yellow, indicative of late-log phase growth, a 100-µL aliquot of the passaged M. penetrans was taken from the top of the well and transferred to fresh 2 mL of SP-4 motility media in the adjoining well. This process was repeated 75 times, generating M.

48 penetrans strain HP88. Subsequently, HP88 was cultured at 37°C in SP-4 broth (Tully et al., 1979) or on SP-4 agar plates.

Hemadsorption (HA). SP-4 agar plates containing M. penetrans str. HP88 were incubated with SRBC in Alsever's solution (Cleveland Scientific) and washed with phosphate-buffered saline

(PBS; 150 mM NaCl, 3.2 mM NaH2PO4, 13.6 mM Na2HPO4, pH 7.2) as previously described (Hatchel et al., 2006). Mycoplasma colonies were observed and photographed at 400X magnification on a Leica DM IRB inverted phase-contrast/epifluorescence microscope.

Time-lapse microcinematography. M. penetrans strain HP88 cells from frozen, mid-log phase stocks were filtered through a 0.45-µm filter and incubated for 3 h at 37°C in glass chamber slides (Nunc) in SP-4 motility media. For motility analysis, 27 images were captured at 1000X magnification on a Leica DMIRM inverted phase-contrast/epifluorescence microscope at 0.250 s intervals. Images were merged and analyzed as previously described (Hatchel et al., 2006). Gliding speeds of individual cells were calculated by dividing the distance traveled by the number of frames between which cells moved, a total of 100 motile cells were analyzed to determine the average gliding speed of M. penetrans str. HP88, excluding rest periods.

SEM. Cells were prepared for SEM as previously described (Hatchel et al., 2006). Briefly, M. penetrans str. HP88 was grown for 6 to 12 h in SP-4 motility media at 37°C. To analyze TX- insoluble structures, TX in TN buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl) was added directly to whole cells for a final TX concentration of 1% and the coverslips were incubated for 30 min at 37°C. Coverslips were fixed 30 min in 1.5% glutaraldehyde/1% paraformaldehyde/0.1 M sodium cacodylate, pH 7.2, rinsed with 0.1 M sodium cacodylate buffer, and dehydrated through a series of ethanol washes ranging from 25% to 100%. Upon dehydration, coverslips were critical point-dried and then viewed on a Zeiss Supra 35 FEG-VP scanning electron microscope.

Electron cryotomography. M. penetrans str. HP88 was grown on Quantifoil Au-finder grids (Quantifoil Micro Tools GmbH) in SP-4 broth at 37˚C until mid-log growth, indicated by a color change in the media from red to orange. The EM grids with attached M. penetrans cells were removed from the media, loaded onto tweezers, and washed with fresh media. Colloidal gold (10 nM) was added to the grids and they were blotted (Whatman, grade 40) prior to being plunge-

49 frozen in liquid ethane on a gravity plunger freezer. Grids were loaded into a 300 kV FEI Polara G2 electro cryo-transmission electron microscope equipped with a field emission gun (FEG), a lens-coupled 4k x 4k Gatan UltraCam, and a Gatan energy filter (GIF). Samples were maintained in liquid nitrogen temperature as tilt series were captured of whole cells. The tilt-series was recorded from -60° to +60° with 1° increments at a 10 µm defocus using Leginon (Suloway et al., 2009). All tilt-series were collected through GIF around zero-loss energy with a slit-width of 20 keV for a cumulative dose of 180 e-/Å2 used for each tilt series.

TX extraction. M. penetrans str. HP88 cells was grown in SP-4 broth at 37°C for 48 h in 175- cm2 tissue culture flasks (Thermo Fisher). To perform a TX extraction on attached cells only, unattached cells in the SP-4 broth were decanted from the tissue culture flask. The attached M. penetrans str. HP88 cells were washed twice with warm PBS to remove weakly attached cells or microcolonies from the flask surface. M. penetrans cells were scraped into 50 mL cold PBS and centrifuged for 20 min at 17,400 x g at 4°C. Upon removal of the supernatant, the remaining cell pellet was washed twice with PBS via centrifugation as described above and the M. penetrans str. HP88 cells were resuspended in 25 mL of cold PBS. For detergent extraction, TX in TN buffer was added to the M. penetrans cells and incubated at 37° C for 1 h, mixing periodically during the incubation period. To isolate the TX-insoluble fraction, M. penetrans cells were centrifuged for 20 min at 17,400 x g at 4°C and resuspended in 250 µL PBS. Whole cell lysate of M. penetrans was obtained by decanting SP-4 broth containing unattached M. penetrans cells from the tissue culture flask, scraping the attached cells into cold PBS, washing the cells in PBS by centrifugation, and resuspending the cells in 1 mL cold PBS. The protein concentrations of both whole-cell lysate and TX-insoluble proteins were determined using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific). For initial analysis, 10 µg of both whole cell lysate and TX-insoluble proteins were electrophoresed through 10% SDS-polyacrylamide gels (Laemmli, 1970) and stained with Sypro® Ruby Protein Gel Stain (Molecular Probes). To compare the protein profiles between the cell lysate and TX- insoluble fraction, VersaDoc™ software (Bio-Rad) was utilized to identify proteins enriched at least two-fold in the TX-insoluble fraction compared to the whole cell lysate for further analysis. Trypsin digestion of candidate proteins. In order to make certain an appropriate protein concentration sufficient for MALDI-TOF MS analysis was recovered for each candidate band, 50 µg TX-insoluble protein was electrophoresed on a 10% SDS-polyacrylamide gel and stained

50 with Coomassie Blue for visualization. After washing the stained gel 3 times with deionized water for 10 min, gel slices containing 4 candidate bands, myosin (control), and an empty gel slice were excised, chopped into pieces, and placed in Eppendorf tubes pretreated by rinsing the tube with 50% acetonitrile, 0.1% trifluoracetic acid (TFA). Gel pieces were washed with 50% acetonitrile/50 mM ammonium bicarbonate for 15 min with gentle shaking. Next, the gel pieces were washed in 50% acetonitrile/10 mM ammonium bicarbonate for 15 min with gentle shaking, followed by a wash in 50% acetonitrile for an additional 15 min while shaking. The gel pieces were then washed in 100% acetonitrile for 5 min, the liquid portion removed, and the gel pieces were allowed to air dry for 15 min. After drying, 10 mM ammonium bicarbonate containing 1 µg trypsin (Roche Diagnostics) was added to rehydrate the gel pieces, and the rehydrated pieces were incubated at 37˚C for 18 h with constant rocking. After trypsin digestion, the liquid portion was placed in a pretreated, as above, Eppendorf tube and the remaining gel pieces were incubated at room temperature with 250 µL of water containing 0.1% TFA for 15 min while shaking. The liquid portion was added to the previously collected liquid sample above and the gel slices were washed again with 30% acetonitrile/0.1% TFA and incubated at room temperature for 15 min while shaking. The liquid portion was removed and pooled with the previously collected liquid samples, and the gel pieces were washed with 60% acetonitrile/0.1% TFA solution followed by 90% acetonitrile/0.1% TFA. After both washes, the liquid portions were pooled with the previously collected liquid samples. All the liquid collected for each individual candidate band were pooled together and dried in a Speed-Vac until approximately 5 µL of liquid remained in each tube. Further clean-up of the samples was performed by passing the samples through C18 ZipTip™ pipette tips (Millipore) for a higher quality mass spectrometry. The samples were submitted to the Center for Bioinformatics and Functional Genomics (CBFG) at Miami University for MALDI-TOF analysis. Protein Analysis. MS-FIT (http://prospector.ucsf.edu/prospector/cgi- bin/msform.cgi?form=msfitstandard) was used to analyze data from MALDI-TOF mass spectrometry. Mass spectrometry results that were not identified as trypsin peaks were entered into MS-FIT with the maximum number of missed cleavages set to 1. The NCBI database was used to identify any potential protein matches between the mass spectrometry data and known mycoplasma proteomes. The predicted coiled-coil motifs were identified using MultiCoil (http://groups.csail.mit.edu/cb/multicoil/cgi-bin/multicoil.cgi).

RESULTS Gliding speed and HA M. penetrans strains GTU-54-6A1 and HF-2 adhere too weakly to plastic for biochemical analysis of their cytoskeletons (Jurkovic et al., 2012; unpublished data). In an effort to obtain a strain of M. penetrans with altered gliding and adherence properties¸ M. penetrans str. GTU-54- 6A1 was passaged in SP-4 motility media 75 times, and the resulting strain was designated HP88. M. penetrans str. HP88 glided in one direction with an average speed of 1201 ± 326 nm/s (n=103). This speed was 2 times faster than strain GTU-54-6A1 and >20 times faster than strain HF-2 (Jurkovic et al., 2012). Although significantly faster than other M. penetrans strains, strain HP88 was still slower than the extremely fast gliding of >2500 nm/s observed for M. mobile (Kirchoff and Rosengarten, 1984) and Mycoplasma testudinis (Hatchel and Balish, 2008). M. penetrans str. HP88 gliding speeds spanned a range of 158-2115 nm/s, corresponding to 0.2-1.8 times the average gliding speed (Figure 13). Compared to strains GTU-54-6A1 and HF-2, the range of speeds was slightly narrower, yet still large compared to other mycoplasma species. M. penetrans strains GTU-54-6A1 and HF-2 and M. iowae serovars K and N exhibit differences in gliding speeds that correlate positively with differences in cytadherence (Jurkovic et al., 2012). To test whether this was also the case for M. penetrans str. HP88 we performed HA assays. Colonies were completly covered by sheep erythrocytes (Figure 14), similar to the coverage observed of colonies of strain GTU-54-6A1 (Jurkovic et al., 2012). Despite the twofold difference in gliding speed, there was no obvious difference in erythrocyte coverage between M. penetrans strains HP88 or GTU-54-6A1 (Figure 8A). Nonetheless, M. penetrans HP88 was observed to withstand additional rinses throughout SEM processing to a greater extent than strains GTU-54-6A1 and HF-2, suggesting that this strain is better able to remain attached to glass during processing. This observation suggest that the HA assay is limited in its ability to distinguish among strains whose adherence capability is above a certain threshold. Morphology of M. penetrans strain HP88 The cell morphology of glass-attached M. penetrans str. HP88 was observed by SEM (Figure 15A) to be indistinguishable from that of M. penetrans strs. GTU-54-6A1 and HF-2 (Jurkovic et al., 2012). Although pleomorphic, cells consisted of a tip structure and a cell body. Additionally, paired cells were observed to be connected by a filament, which often varied in length, consistent with the observation that gliding motility is utilized during M. penetrans cell division (Jurkovic

Figure 13: Distribution of M. penetrans gliding speeds about the mean. Gliding speeds of the individual motile cells for M. penetrans str. HP88 were grouped into bins of 0.2x of the mean, designated by x. Grouping shown for M. penetrans HF-2 and M. penetrans str. GTU-54- 6A1 (Jurkovic et al., 2012) are included for comparison to M. penetrans str. HP88. Blue, strain HP88, black, strain GTU-54-6A1, grey, strain HF-2.

Figure 14: Colony HA of M. penetrans strain HP88. M. penetrans str. HP88 colonies grown on SP-4 agar plates were incubated with SRBC, rinsed with room temperate PBS, and visualized at 40X magnification. Arrow, individual SRBC.

56 et al., 2012). The increased ability to attach to surfaces allowed for visualization of HP88 by ECT, enabling examination of the internal cell morphology and organization without artifacts caused by the fixation and dehydration used in other techniques. Cryotomographic images of M. penetrans str. HP88 sections revealed no cell constrictions at either pole (Figure 16A, 16B), which is at odds with results from other types of imaging (Lo et al., 1991; Giron et al., 1996; Jurkovic et al., 2012). Two different types of internal composition were observed at the cell pole and within the cell body of whole cells. Areas free of ribosomes were present at both cell pole regions, and were often of different dimensions. These ribosome exclusion zones, which have previously been seen in other strains (Lo et al., 1991), are likely to correspond to the TX- insoluble material previously identified in strain GTU-54-6A1 (Jurkovic et al., 2012). Treatment of M. penetrans str. HP88 (Figure 15B) with 1% TX confirmed the presence of a TX-insoluble structure indistinguishable from that observed in other strains (Jurkovic et al., 2012). Occasionally remnants of the connecting filament observed in SEM were found at one pole in sectioned cells by ECT, enabling identification of this pole as the one not involved in attachment and motility (Figure 16C). Additionally, striations in the ribosome exclusion zone were often observed parallel to the long axis of the cell, converging at the poles (Figure 17). The striations were visible only in sections subtending the cell membrane, suggesting that they were located on the surface of the internal cytoskeletal structure. TX-insoluble structure of M. penetrans HP88 In order to identify the protein components of the TX-insoluble material, detergent extraction of M. penetrans cells was performed to compare the TX-insoluble protein profile to whole cell lysate. Efforts to obtain highly enriched structures by TX extraction of cell pellets resulted in incompletely extracted cells with membrane contamination, as determined by SEM visualization (data not shown). This contamination likely resulted from incomplete extraction caused by inaccessibility to TX of cells found in pelleted material. Therefore we extracted M. penetrans cells attached to a surface. After TX extraction of M. penetrans str. HP88 cells grown in tissue culture flasks, adherent material was scraped, collected, and separated on a 10% SDS polyacrylamide gel and compared with whole cell lysate (Figure 18A). Several bands in the TX- insoluble fraction were enriched at least 2-3 fold compared to whole cell lysate (Figure 18B).

Figure 15: SEM of M. penetrans strain HP88 attached to glass. A: M. penetrans str. HP88 whole cells. B: M. penetrans str. HP88 cells treated with TX detergent. Black arrow, tip structure; Arrowhead, cell body; Red arrow, non-uniform TX-insoluble structures. Scale bar, 1 µm.

Figure 16: Internal organization of M. penetrans strain HP88 observed by ECT. A , B, and C, sections of different whole cells, D, E, and F, schematic of respective whole cell sections. Remnant of connecting filament between paired cells observed at one pole of sectioned cells (C and F) Black line, cell membrane; yellow line, border of ribosome-filled region; green circle, ribosome. Scale bar, 100 nm.

Figure 17: Cytoskeletal filaments observed in M. penetrans strain HP88 cell pole by ECT. A-G, thin section series of one cell pole, from inside the cell out to its cell membrane, H, merged schematic of filaments observed in sections (A-G). Unaltered images provided of innermost section A (I) and the section closest to the cell membrane in section G (J). Blue lines, filaments. Scale bar, 100 nm.

Four of these candidate bands were chosen for MALDI-TOF analysis. The peptide sequences acquired were compared to the M. penetrans str. HF-2 genome, since it is the only complete genome sequence currently available for M. penetrans. Three of the bands gave no clear matches. The candidate band migrating at 102 kDa was identified as a mixture of MYPE1560 and MYPE1570 (Sasaki et al., 2002). The sequences identified for MYPE1560 and MYPE1570 covered 26% (Figure 19B) and 44% of the respective predicted amino acid segments (Figure 19D), leaving their identification as these proteins unambiguous. Further analysis of the MYPE1560 and MYPE1570 genome sequences revealed the presence of predicted alpha-helical coiled coil motifs (Figure 19A, 19C). The genes encoding MYPE1560 and MYPE1570 are immediately downstream of four other genes encoding similar proteins (Figure 20). The short intergenic spaces (<30 nt) between all six genes indicates that they are most likely transcribed polycistronically, under the control of one promoter. All six genes of this putative transcriptional unit have with similarly low pI values and predicted coiled-coil structures (not shown).

DISCUSSION

Gliding motility and cytadherence. Like M. penetrans strains GTU-54-6A1 and HF-2, the high-passage M. penetrans strain HP88 was HA-positive and motile, albeit twice as fast as the faster of the other two strains (Jurkovic et al., 2012). In addition to the faster gliding speed, M. penetrans strain HP88 had a slightly smaller range in gliding speeds, with twice as many cells gliding close to the average speed compared to strains GTU-54-6A1 and HF-2 (Jurkovic et al., 2012). Strain HP88 was observed to adhere more tightly to both plastic and glass surfaces than the other two M. penetrans strains. The lack of difference in HA between strains HP88 and GTU-54-6A1 may reflect an upper detection limit to the HA assay. The increase in both speed and adherence reduces background noise in assays of terminal organelle-associated functions by providing a higher baseline, making strain HP88 particularly suitable for further characterization of motility, adherence, and terminal organelle features of M. penetrans. To passage M. penetrans strain HP88, cells were inoculated in SP-4 motility media in plastic wells. The passaging conditions might have created an environment in which better adhering cells outcompeted other cells by taking up available surface area. Since M. penetrans cells incubated in shaking culture are unable to grow (D. Jurkovic & M. Balish, unpublished), cells unable to attach to the plastic surfaces would be expected to remain in the broth and

Figure 18: SDS-PAGE profiles and comparison of whole cell and TX-insoluble proteins of M. penetrans strain HP88. A, Whole cell lysate and TX-insoluble proteins were separate by SDS-PAGE to identify proteins enriched in the TX-insoluble portion, Arrow, candidate band chosen for MALDI-TOF MS. B, Comparison of protein SDS-PAGE profiles of both whole cell lysate and TX-insoluble proteins. Arrow, peaks corresponding to candidate bands identified in panel A. Blue line, lysate; red line, TX-insoluble fraction.

Figure 19: Protein characteristics predicted for MYPE1560 and MYPE1570. Coiled-coil motifs predicted by MULTICOIL for MYPE1560 (A) and MYPE1570 (C) depict multiple coils in both proteins. Amino acid sequence for MYPE1560 (B) and MYPE1570 (D). Red, acidic amino acids; underlined, peptide matches from MALDI-TOF MS results.

Figure 20: Putative transcriptional unit of M. penetrans cytoskeletal protein-encoding genes. Predicted open reading frames for genes encoding proteins MYPE1520, 1530, 1540, 1550, 1560, and 1570. Proteins identified within the TX-insoluble fraction by MALDI-TOF analysis are indicated in red, proteins with similar characteristics to MYPE1560 and MYPE1570 are indicated in grey. Values under gene name are length (nucleotides).

70 consequently be unable to proliferate. Therefore, random mutations affecting adhesins might have resulted in a better-adhering M. penetrans cell becoming dominant in the culture. Passaging of two isolates of the close M. penetrans relative Mycoplasma iowae resulted in two strains with different motility speeds, one faster and one slower than its parent strain (J. Newman & M. Balish, unpublished), suggesting that passaging is a good method of obtaining mutants in adherence, gliding, or both. The positive correlation between adherence strength and motility speed between strains HF-2 and GTU-54-6A1 (Jurkovic et al., 2012) is further supported by the observation of stronger adherence correlating with faster gliding in strain HP88. The relationship between the two functions supports a model in which adherence and motility are mediated by the same cellular components. Biogenesis of TX-insoluble structure of M. penetrans strain HP88 A ribosome exclusion zone is observed at one of the cell poles of Caulobacter crescentus. Its size dramatically increases upon overexpression of the cytoskeletal protein PopZ, which adopts a highly cross-linked polymeric structure (Ebersbach et al., 2008). The presence of a similar region at the pole of M. penetrans is therefore consistent with a proteinaceous, cross-linked structure whose components are restricted to this region. The observation of ribosome exclusion zones at both poles and TX-insoluble structures of similar size distribution are further evidence that they are one and the same. A model in which TX-insoluble material is found at both poles could explain the variation in TX-insoluble structure length observed after cells attached on glass coverslips are extracted with detergent (Jurkovic et al., 2012). The accumulation of TX-insoluble material would contribute to acquisition of terminal organelle functions at the poles, with adhesins and motility-associated molecules only being recruited when a certain size threshold is reached, providing a timing mechanism for terminal organelle activation. TX-insoluble structures of other mycoplasmas also contain proteins, some acidic and coiled coil-rich, that form complex structures (Henderson & Jensen, 2006; Seybert et al., 2006; Nakane & Miyata, 2007). Morphology of M. penetrans HP88 As visualized by SEM, M. penetrans str. HP88 cells had a protruding tip structure. In contrast, no protruding tip structure was observed by ECT. Rather, the only indication of a differentiated cell pole was the different internal composition observed between the pole regions and cell body. One of the benefits of ECT is the ability to obtain images of cells in a nearly native state (Li and Jensen, 2009), without the artifacts caused by dehydration and fixation of cells visualized by

SEM or TEM. The fact that discrete, relatively narrow tip structures were observed in M. pneumoniae prepared by ECT (Henderson & Jensen, 2006) and in M. mobile prepared by rapid- freeze-and-fracture EM (Miyata and Peterson, 2004) underscores the structural and compositional novelty of the M. penetrans terminal organelle. A possible explanation for the different appearance of the M. penetrans poles when prepared for ECT and by conventional methods is that the M. penetrans cytoskeleton consists of a gel-like substance, associated with the cell membrane, constructed largely from TX-insoluble proteins arranged parallel to the long axis of the cell. In this model (Figure 21), the parallel cytoskeletal components move closer together upon dehydration, resulting in a dramatic narrowing of the cell in that region. Thus, we propose that in its native state, M. penetrans lacks a protruding tip structure. A defined tip structure is not a requirement for mycoplasma attachment or motility, as previous work has shown that Mycoplasma insons, whose morphology is characteristic of a rod-like bacterium, is polarized, with motility in one direction and adherence often at one pole (Relich et al., 2009). The observation that motile mycoplasma cells do not all have a defined tip structure suggests that their common ancestor did not have a polarized tip structure, and that there are at least four distinct ways for establishing functional polarity, modeled by M. pneumoniae, M. mobile, M. insons, and M. penetrans. Alternatively, mycoplasma terminal organelles may have divergently evolved from a common ancestor with a polarized tip structure. This is more likely not to have happened, as it would require non-polarized mycoplasmas to lose polar morphology, and the polarize mycoplasmas would have to replace the terminal of the common ancestor with the distinct cytoskeletal structures observed in the Mycoplasma genus. MYPE1560 and MYPE1570 Proteins MYPE1560 and MYPE1570 matched the peptide data obtained from TX-insoluble proteins isolated from strain HP88. The fact that matches were obtained even though the HP88 strain is not sequenced suggests that the proteins are well-conserved across strains of M. penetrans. MYPE1560 and MYPE1570 are the last two genes found in a cluster of six genes with similar characteristics. All six genes in the putative operon encode coiled-coil proteins with low pI values, features shared by M. pneumoniae cytoskeletal proteins, including the acidic TopJ (Cloward and Krause, 2010), P200 (Balish and Krause, 2002), HMW1, HMW3, P65, and the coiled-coil proteins P41 (Kenri et al., 2004), HMW1, P28, and P65. Attempts to localize MYPE1570 by immunofluorescence microscopy were unsuccessful (not shown), implying that it

Figure 21: Models of M. penetrans whole cell morphology observed by ECT and SEM. When processing cells for ECT (A), no dehydration occurs, keeping the TX-insoluble structure intact; therefore, the width of cell membrane at the pole is maintained. The ethanol washes used to dehydrate samples for visualization by SEM (B) remove water from the cytoskeletal material, causing the width of structure to shrink and the attached cell membrane to become tapered at the end with the fully developed polar organelle. Red, TX-insoluble protein; blue, chromosome; green, ribosomes.

74 might be buried within the TX-insoluble structure, and therefore not accessible by the antisera. Different microscopy techniques will have to be used to observe localization of MYPE1560 and MYPE1570 in the TX-insoluble structure, as well as obtaining antisera to the other proteins of the putative transcriptional unit to observe their localization. The lack of sequence homology among the cytoskeletal proteins of M. pneumoniae, M. mobile, and M. penetrans is consistent with a model of convergent evolution of terminal organelles. With the diversity observed among the identified mycoplasma cytoskeletons, it is reasonable to propose that the different structures might also employ different underlying molecular mechanisms for adherence and motility, which needs to be further investigated.

CHAPTER 3

Analysis of the energy source for Mycoplasma penetrans gliding motility

D.A. Jurkovic1, M. R. Hughes2, and M.F. Balish1

1 Department of Microbiology, Miami University, Oxford, OH 45056 2 Statistical Analysis Center, Miami University, Oxford, OH 45056

ABSTRACT Mycoplasma penetrans, a potential human pathogen found mainly in HIV-infected individuals, uses a tip structure for both adherence and gliding motility. To improve our understanding of the molecular mechanism of M. penetrans gliding motility we generated and characterized the hyperadherent, fast-gliding strain HP88. M. penetrans gliding motility was not eliminated in the presence of an ATPase inhibitor, a proton motive force inhibitor, or a sodium motive force inhibitor. At physiological pH, gliding speed increased as temperature increased. The absence of a clear chemical energy source for gliding motility and a positive correlation between speed and temperature suggest that energy derived from heat provides the major source of power for the gliding motor of M. penetrans.

INTRODUCTION Cellular motility is important for a variety of processes, including obtaining nutrients, evading threats, organizing cells for developmental processes, and cell division. Both thermal and chemical energy are employed as direct sources of energy for cellular motility. For example, many eukaryotic cells are driven forward by the formation of membrane protrusions through localized polymerization of actin, powered principally by thermal energy in the form of a Brownian ratchet (Peskin et al., 1993). Bacterial twitching motility is powered by ATP hydrolysis, which powers extension and retraction of type IV pili attached to a surface (Burrows, 2005). Rotation of bacterial flagella, which drive swimming and swarming movements, is powered by proton motive force (PMF) (Berg &Anderson, 1973), or rarely by sodium motive force (SMF) (McCarter, 2004). In both Flavobacterium johnsoniae and Myxococcus xanthus, gliding motility, the smooth movement of cells over a surface, is powered by PMF (Liu et al., 2007; Nan et al., 2010; Sun et al., 2011). Since gliding motility is carried out among diverse bacterial groups and uses diverse mechanisms (McBride, 2004), no single organism can be used to model a molecular mechanism for this process. Several mycoplasmas exhibit gliding motility, enabling these bacteria to colonize and cause infection in their hosts (Jordan et al., 2007; Szczepanek et al., 2012). Among these species, only Mycoplasma mobile has been studied in depth to identify its motility energy source. Arsenate, a phosphate analogue that causes depletion of cellular ATP, rapidly and potently inhibits motility of M. mobile (Jaffe et al., 2004), and Triton X-100 (TX)-permeabilized cells resume movement when ATP is added directly to the cells, demonstrating that the motor is directly dependent on ATP hydrolysis (Uenoyama et al., 2005). Little is known about the energy source necessary for gliding motility in other mycoplasmas. However, it is well-established that different mycoplasma species use compositionally dissimilar tip structures for gliding motility (Relich et al., 2009; Miyata, 2010; Jurkovic et al., 2012), making it impossible to generalize the motility mechanisms they use. One mycoplasma species whose gliding mechanism is unknown is Mycoplasma penetrans, a putative human pathogen originally isolated from the urogenital tract of HIV- positive patients (Lo et al., 1991, 1992; Wang et al., 1992). Its lipoproteins are mitogenic toward B and T lymphocytes (Feng et al., 1994; Sasaki et al., 1995) and stimulate transcription of the HIV genome in vitro via toll-like receptors (Shimizu et al., 2004), implying a role for M.

78 penetrans in the accelerated progression of AIDS. M. penetrans has a polar terminal organelle that leads during gliding motility and whose TX-insoluble cytoskeleton is distinct from those of most other species, including M. mobile (Jurkovic et al., 2012). Genomic analysis reveals the absence of clear homologs of terminal organelle-associated proteins of other species (Sasaki et al., 2002). The present study aims to identify potential sources of energy for gliding motility of M. penetrans, examining both chemical and thermal contributions to the movement of cells of this species.

MATERIALS AND METHODS Bacterial culture conditions and passaging. M. penetrans strain HP88 was cultured at 37°C in SP-4 broth (Tully et al., 1979) or SP-4 motility medium (SP-4 broth supplemented with 3% gelatin). As a control, M. mobile strain 163K (Kirchoff and Rosengarten, 1984) was cultured at room temperature in SP-4 broth or SP-4 motility media.

Energy source assays. To determine the effects of inhibitors of ATP metabolism and ion motive force on M. penetrans motility, cells were analyzed in buffers with or without the test reagent. M. penetrans motility stocks were incubated in SP-4 motility medium for 3 h at 37°C in a glass chamber slide. M. mobile cells from frozen mid-log phase growth were syringed 10 times to break apart aggregates before incubation in SP-4 motility media for 1 h at 25°C. For both species, the medium was then removed and each chamber was rinsed 5 times with the control or test buffer, incubated in the control or test buffer for 1 h, and analyzed for motility as described above. The following buffers were used: phosphate-buffered saline supplemented with gelatin 2 and glucose (PBS-G ; 150 mM NaCl, 32 mM NaH2PO4, 136 mM Na2HPO4, 10 mM glucose, 3% gelatin, pH 7.2); arsenate-buffered saline supplemented with gelatin and glucose (ArBS-G2K; 140 mM NaCl, 75 mM KCl, 10 mM glucose, 2.5 mM potassium arsenate, 4.75 mM sodium arsenate, 3% gelatin, pH 7.2); PBS-G2 supplemented with potassium (PBS-G2K; 140 mM NaCl, 10 mM KCl, 10 mM glucose, 50 mM sodium phosphate, pH 7.2); PBS-G2 supplemented with carbonyl cyanide m-chlorophenyl hydrazone (CCCP) [C3PBS-G2; 150 mM NaCl, 3.2 mM

NaH2PO4, 13.6 mM Na2HPO4, 10 mM glucose, 3% gelatin, 10 µM CCCP in dimethyl sulfoxide (DMSO), pH 7.2]; and PBS-G2 supplemented with amiloride (APBS-G2; 150 mM NaCl, 3.2 mM

NaH2PO4, 13.6 mM Na2HPO4, 10 mM glucose, 3% gelatin, 10 µM amiloride, pH 7.2). All reagents were purchased from Sigma Aldrich.

Effect of arsenate on growth of M. penetrans. To determine the effect of arsenate on the growth of M. penetrans, cells were grown in 10 mL SP-4 broth or SP-4 broth supplemented with either 10 mM sodium arsenate (pH 7.2) or sodium phosphate (pH 7.2) for approximately 2 d, until the pH indicator in the SP-4 broth control changed from red to yellow indicating the cells are in late-log phase of growth.

Temperature/pH study. Motility stocks were incubated in SP-4 motility medium with the desired pH (5.8, 6.8, 7.8, 8.8) and temperature (30°C, 37°C, 40°C) in glass chamber slides. For motility analysis, 18 images were captured at 1000X magnification on a Leica DM IRB inverted phase-contrast/epifluorescence microscope at approximately 0.25-s intervals. Images were merged and analyzed for 20-25 motile cells as previously described (Hatchel et al., 2006).

Statistical analysis. The temperature and pH data were analyzed using two-factor factorial analysis of variance (ANOVA) to examine the effects of both temperature and pH on motility speed. To determine the temperature and pH associated with maximal gliding speed, a statistical response surface model was fit to the data with an accompanying canonical analysis to determine which of the two factors optimizes motility. The effects of energy source inhibitors on motility were analyzed by ANOVA. All statistical analyzes were performed using SAS version 92 for Windows.

RESULTS Effect of arsenate on motility and growth. Arsenate was added to M. penetrans cells to determine whether ATP hydrolysis by a motor-associated component directly provides energy for gliding, as proposed for M. mobile (Jaffe et al., 2004). Arsenate enters prokaryotic and eukaryotic cells via phosphate transporters (Rosen, 2002) and inhibits many reactions involving phosphate, including substrate-level phosphorylation events leading to ATP synthesis via the glycolysis (Warburg & Christian, 1939) and arginine dihydrolase (Knivett, 1953) pathways, the only two means of ATP synthesis available to M. penetrans (Lo et al., 1992; Sasaki et al., 2002). In M. mobile, arsenate has an immediate negative impact on gliding motility (Jaffe et al., 2004). In contrast, M. penetrans continued to glide in the presence of 50 mM arsenate (2.5 mM potassium arsenate, 47.5 mM sodium arsenate) at incubation times ranging from 1 to 8 h. To determine the change in motility speed, M. penetrans cells were analyzed after 1 h incubation in 50 mM or 250 mM arsenate. In 50 mM arsenate, the gliding speeds of both M. mobile and M.

80 penetrans were significantly reduced (p<0.0003 for both). However, the 37% decrease for M. penetrans was much smaller than that of M. mobile, which exhibited an 89% decrease speed (Figure 22), in agreement with the observations of an absence of M. mobile cells moving faster than 10% of normal gliding speed after 10 min under similar conditions (Jaffe et al., 2004). Although the change in speed of M. penetrans was statistically significant, the moderate value of the decrease and its continued movement after 8 h (not shown) suggest that direct inhibition of the motor was unlikely. Increasing the arsenate concentration fivefold had a negligible effect on M. penetrans (Figure 22). After 2 d of incubation at 37°C, growth of M. penetrans was observed with the addition of 10 mM sodium phosphate, pH 7.2, but not 10 mM sodium arsenate, pH 7.2 (Figure 23), confirming that M. penetrans takes up arsenate and its growth is inhibited at much lower arsenate concentrations than were used in the motility experiments. Thus, ATP hydrolysis is an unlikely energy source for gliding by M. penetrans. Effects of CCCP and amiloride on motility. The presence of membrane potential has been reported in a variety of mycoplasma species (Benyoucef et al., 1981; Schiefer & Schummer, 1982). To determine whether PMF supplies the energy needed for M. penetrans gliding motility, we observed motility in the presence of the ionophore CCCP, which collapses proton gradient. Cells were incubated for 1 h in the presence of 10 mM CCCP in DMSO and in PBS-G2K containing the same volume of DMSO used in the test buffer. After 1 h, M. penetrans gliding speed actually increased by 29% compared to the control buffer (p<0.0001) (Figure 22). To test SMF as a potential energy source for M. penetrans gliding, cells were observed in the presence of amiloride, an inhibitor of the Na+/H+ antiporters and sodium channels which competes with Na+ in the medium (Benos, 1982). M. penetrans gliding speed was not significantly affected by 1 h of incubation in amiloride (p=0.6) (Figure 22), ruling out SMF as an energy source. Effects of pH and temperature on gliding motility. Mycoplasma gliding motility is routinely analyzed at their optimum growth temperature (Radestock and Bredt, 1977). To determine the effects of temperature and pH on the gliding speed of M. penetrans, motility was analyzed at temperatures ranging from 30°C to 40°C and pH levels ranging from 5.8 to 8.8 (Figure 24). Speed increased with temperature, but at acidic or alkaline pH. the trend was less distinct. Additionally, speed was greater closer to neutral pH. The interaction between temperature and pH was significant (p<0.0001), suggesting that the effects of temperature depend on the pH. To determine the temperature and pH parameters for maximal speed, a statistical response surface

Figure 22. Affect of chemical energy inhibitors on M. penetrans gliding motility. M. penetrans relative gliding speed in the presence and absence of arsenate, CCCP, and amiloride compared to the average gliding speed of M. mobile in the presence of 10 mM arsenate. Values were normalized to respective control buffers. 1x ArBS, arsenate-buffered saline with 50 mM arsenate; 5x ArBS, arsenate-buffered saline with 250 mM arsenate; CCCP, 10 µM CCCP in DMSO; amiloride, 10 µM in phosphate-buffered saline.

Figure 23. Effect of sodium arsenate and sodium phosphate on the growth of M. penetrans. M. penetrans cells were grown in 10 mL SP-4 broth supplemented with sodium arsenate (10 mM) or sodium phosphate (10 mM) and compared to cells grown only in SP-4 broth. The pH indicator in the SP-4 broth causes a color change from red to yellow when the culture becomes acidic, the result of metabolic activity of a growing culture. When the pH indicator remains red, no cell growth has occurred.

Figure 24. Affect of temperature and pH on M. penetrans motility speed. M. penetrans relative gliding speed under conditions of varying temperature and pH. Values were normalized to mean gliding speed at 37°C and pH 7.8. Asterisks, significant difference (p<0.0001) determined by two-factor factorial analysis of variance between adjoining bars.

87 model was fitted to the data obtained from the temperature and pH assays, along with accompanying canonical correlation analysis (Figure 25). There were highly significant linear and curvilinear effects, as well as a marginally significant interaction effect of both temperature and pH, and both were found to be significant contributors to gliding speed. The surface revealed a rising ridge along the temperature gradient, suggesting that maximal speed occurs at a temperature higher than 40°C. Ridge analysis suggests that maximal speed is well maintained near neutral pH levels, and is found on a strong linear trajectory in increasing temperature. At 45°C almost no cells adhered (data not shown), marking 40°C as an upper limit to the experiment. These data suggest that thermal energy is limiting for gliding speed as long as the adherence and motility machinery is capable of functioning, most likely at near-neutral pH, rather than the pH extremes analyzed in our experiment.

DISCUSSION Energy source of M. penetrans str. HP88 gliding motility. The molecular mechanism of M. penetrans gliding motility is unknown and no homologs of known motility proteins in the better-characterized species, M. pneumoniae and M. mobile, are present. In an effort to identify the energy source used to power gliding, the motility behavior of M. penetrans was observed in the presence of chemical inhibitors previously used to characterize motility energetics in other mycoplasma and bacterial species. Arsenate did not have the same degree of impact on M. penetrans gliding, as it did on M. mobile, with a much smaller reduction in speed. Furthermore, M. penetrans cells were still able to glide well after 8 h in the presence of arsenate, and at concentrations fivefold greater than those tested for M. mobile, both of which are conditions under which ATP ought to be nearly completely depleted through inhibition of the reactions catalysed by glyceraldehyde 3-phosphate dehydrogenase (Warburg & Christian, 1939) and ornithine carbamoyltransferase (Knivett, 1954). As mycoplasma membrane ATP synthase actually operates in reverse to maintain a proton gradient functioning in sodium extrusion and cell volume maintenance (Linker & Wilson, 1985), and is therefore not involved in ATP synthesis, it is overwhelmingly likely that ATP is depleted under our experimental conditions, which include incubation in 25 times the concentration of arsenate that prevents growth. These data suggest that ATP hydrolysis is at best an indirect source of energy for motility in M.

Figure 25: Response surface plot predicting effect of temperature and pH on M. penetrans motility speed. To determine the effect of temperature (temp) and (pH) on motility response (predicted value) a response surface model was fit to the motility data obtained during the temperature and pH experiment. Canonical analysis of the curve was used to locate the conditions that result in an optimal result, if an optimum was reached in the analysis. A saddle point, the location on the model where the response is maximum in one direction, yet a minimum in another direction, was identified at pH 7.085, 30.55°C (asterisk). The identification of a saddle point suggests that in our temperature and pH experiments, the true optimum calculated by the response surface model was not obtained. To better determine what pH and temperature result in an optimum speed, ridge analysis was performed on the rising edge that exists along the temperature gradient (arrow). Ridge analysis found maximal motility to be well maintained at a pH level of approximately 7.3, but the maximal motility is found on a linear trajectory in increasing temperature.

90 penetrans, perhaps only providing the energy necessary to make and maintain the molecular components of the motor. Furthermore, neither PMF inhibitor nor an SMF inhibitor decreased gliding speed. We therefore could not identify a convincing source of chemically derived energy for gliding. Temperature and pH effects on gliding motility. To examine the possibility of a thermal component to the energy source for gliding, motility was observed under different temperature and pH conditions. We found that at physiologically relevant pH, an increase in temperature caused an increase in gliding speed, but fitted response surface analysis suggested that the maximum speed was mostly likely not observed in the experiments, even though normal physiological temperature was exceeded at 40°C. Therefore, at physiological pH, there is a linear relationship between temperature and speed. These data suggest that thermal energy may be a substantive energy source for M. penetrans gliding motility, whereas a chemical energy source typically observed for bacterial motility was not identified. Mechanism of M. penetrans gliding motility. Two models have been proposed for gliding motility in M. mobile and M. pneumoniae, the centipede and inchworm model, respectively (Miyata, 2010). In the better elucidated centipede model, adhesins reversibly bind substrate in a manner dependent upon ATP hydrolysis. There is no direct evidence in support of a particular motility model in M. pneumoniae, but the inchworm model has been proposed based on electron cryotomography data. In this model, flexing of the cytoskeleton within the attachment organelle causes the displacement and association of adhesins to the cell surface, moving the cell forward (Henderson & Jensen, 2006). In our Brownian ratchet model, M. penetrans motility had similarities with motility based on actin polymerization at the cell membrane. Thermal energy leads to insertion of monomers at the growing end of the filament adjacent to the membrane. Depolymerization is based on the identification of the older part of the filament by slow ATP-hydrolysis resulting in an ADP-actin molecule that is removed, therefore promoting polymerization. In contrast to actin polymerization, in M. penetrans dislocation of the filaments depends on the movement of the cytoskeletal template. Although it remains unclear whether either of these occurs in M. penetrans, our data indicate that the mechanism of motility has an important thermal component. M. mobile speed also correlates positively with temperature (Miyata & Uenoyama, 2002), but in that organism

ATP hydrolysis is absolutely required for movement (Jaffe et al., 2004; Uenoyama et al., 2004), unlike in M. penetrans. If M. penetrans gliding motility is in fact driven by a mechanism that converts thermal energy into forward movement, then this is unique among prokaryotes, and suggests the existence of a yet uncharacterized cytoskeletal component capable of polarized polymerization and depolymerization. Further investigation of the structure and composition of the M. penetrans motor is warranted.

SUMMARY AND CONCLUDING REMARKS Cellular polarity is critical for proper functioning and development of eukaryotes and prokaryotes. Despite the recent increase in study of prokaryotic cytoskeletal elements, understanding of the subject pales in comparison to those found in eukaryotic cells. Members of the Mycoplasma genus lack those prokaryotic cytoskeletal elements that interact with the cell wall and generate morphological and functional polarity. Some mycoplasmas have developed novel cytoskeletal structures to perform tasks mediated by other prokaryotic cytoskeletal filaments. For example, in some species the action of FtsZ is supplemented (Lluch-Senar et al., 2010) or possibly replaced in M. mobile by gliding motility, which can provide the force necessary for cytokinesis. In these species, terminal organelles are formed by novel cytoskeletal proteins. No models explaining the evolution of mycoplasma polarity have been advanced in the literature. We propose that terminal organelles have evolved independently of each other in multiple lineages of the Mycoplasma genus. In each of these lineages, a non-polarized, spherical, ancestor localized its adhesins to one area of the cell, which was favored by preventing these antigenic molecules from being accessed by immune cells. The coupling of these polarized adhesins with novel gliding motility machinery would have improved cell division efficiency, which is demonstrably low in M. gallinarum (Jordan and Balish, unpublished data), a non- polarized species, perhaps due to the absence of a cell wall. M. penetrans and its phylogenetic relative M. iowae utilize their polar tip structure for the same functions mediated by the terminal organelles of other mycoplasma species, despite the major compositional and mechanistic differences. The conserved morphology and organization between the two species supports the relationship between terminal organelle morphology and phylogeny, rather than host (Hatchel and Balish, 2008). The identification of a fourth mycoplasma lineage with distinct cell polarity-generating components supports the hypothesis that terminal organelles have evolved independently from a non-polarized common ancestor. To date, four distinct cytoskeletons are found within four lineages of the Mycoplasma genus to generate polarity and facility polarity-associated functions such as attachment and motility (Figure 26). The M. pneumonia, M. sualvi, M. muris, and M. fastidiosum clusters modeled by M. pneumoniae, M. mobile, M. penetrans and M. insons, respectively. The M.

Figure 26: Diversity of polarity-generating cytoskeletons of Mycoplasma genus visualized by SEM. Four distinct types of cytoskeletons, differing in size and shape, have been identified in the Mycoplasma genus, utilized by cells in polarity-associated functions. A, M. pneumoniae (Hatchel & Balish, 2008), B, M. penetrans, C, M. mobile (Nakane & Miyata, 2007), and D. M. insons (Relich et al., 2009). Scale bars, 250 nm.

95 pneumoniae electron-dense core (Figure 26A; Hatchel & Balish, 2008), located within the attachment organelle, consists of three distinct substructures, the terminal button, rod, and base. In contrast, the M. mobile jellyfish structure (Figure 26C; Nakane & Miyata, 2007) consists of two substructures, the “head” located within the terminal organelle, and the “tentacles” that extend into the cell body from the terminal organelle. In M. penetrans, the cytoskeleton that confers polarity is likely a proteinaecous gel (see Chapter 2) which appears as a cylindrical structure studded with smaller projections (Figure 26B) when dehydrated during SEM, the same method used to visualize the other mycoplasma cytoskeletons. Lastly, the cytoskeleton of M. insons (Figure 26D; Relich & Balish, 2009) is unlike the other three, comprised of parallel, anastomosing filaments that extend the entire length of the cell, consistent with the lack of a polar organelle in this species. In chapter 1 we described the functions associated with the M. penetrans terminal organelle, namely cytadherence, gliding motility, and cell division. To begin characterizing the roles of cytoskeletal proteins in M. penetrans attachment and motility, we began identifying the components of the M. penetrans Triton X-100 (TX)-insoluble structure through examination of its internal composition and organization in chapter 2. Electron cryotomography of whole cells attached to surfaces were observed to not have differentiated tip structures, suggesting that the distinct tip structure observed by SEM is the artifactual result of dehydration during sample preparation. This would explain certain seemingly anomalous features of the TX-insoluble structure as visualized by SEM. First, the average width of the most consistently observed type of structure was smaller than the nucleoid-free space observed by DAPI fluorescence of paired cells. Second, shorter and possibly undeveloped structures were often observed (Figure 15B) suggesting that the material that constitutes these structures is found at both poles in varying amounts. Taken together with the absence of a differentiated tip structure in cells under more native conditions, the TX-insoluble material may be a proteinaceous gel composed of parallel filaments, which constricts along its long axis during dehydration for SEM preparation. We propose a model to explain M. penetrans cell development in which only one pole is completely functional in adherence, motility, and cell division (Figure 27). In this model, it is assumed that a cell pole is functional only when it has acquired an amount of cytoskeletal material above a certain threshold. However, the possibility exists that the filaments observed at the cell poles

Figure 27: Proposed model of cellular development in M. penetrans. A motile M. penetrans cell begins cell division by replicating its genome and making the components necessary to create another cell. As division occurs, motility driven by the activated cell tip (green) continues to glide, causing the cell to become elongated. Since the anchor pole (red) cannot glide, it remains stationary as the active pole continues to glide away creating a connecting filament (CF) between the two cells. As the CF becomes thinner and longer, TX-insoluble material begins to accumulate at what will be the new poles in the daughter cells, including the space within in the CF. Upon cytokinesis, the CF breaks, resulting in two cells with superficially identical polar morphology, but only one of which can glide away as it has the active pole from the previous generation. As a result, the second cell remains stationary until its old pole accumulates and organizes the TX-insoluble material to become active in motility. Once activation of the old pole occurs, the cell is able to glide and start another round of division. Green, active pole; red, anchor pole.

98 may also participate in the activation of the motility machinery at the motile pole. At this point, it is not clear if the striations are found at both poles, further analysis by ECT would be required to better understand the role of the filaments in the activation process. The observation of cells moving in one direction supports the assumption that only one pole can function in motility. Additionally, upon cytokinesis only one daughter cell is motile and the other remains stationary, presumably because neither pole has obtained and organized the TX-insoluble cytoskeletal for functioning to be activated. As a cell prepares for division it begins to elongate due to the force generated by the motility-active pole. The opposite pole acts as an anchor, suggesting that the adhesins are able to keep the anchor pole attached to the surface because the incomplete or inactivated motility machinery is not yet functional. Presumably after chromosome replication and segregation, the distance between the two nascent cells increases, and the connecting filament (CF) between the cells becomes longer and thinner. The lack of DNA observed in the CF suggests that the TX- insoluble material is present within the CF, serving as a nucleation site for addition of cytoskeletal material at the new poles after division. After cytokinesis, the CF breaks, creating two daughter cells. The daughter cell that contains the pole that was already motility-active, which had generated the force for division, glides away, whereas the daughter cell with the anchor pole remains immotile. During this period of immobility, we propose that the anchor pole is acquiring and organizing the cytoskeletal material necessary for complete terminal organelle functionality. The motile cell can begin a new round of cell division immediately upon division. In contrast, the anchor cell must develop a functional polar organelle so that it can begin another round of cell division. In order to better elucidate a model of cell development in M. penetrans, the localization of cytoskeletal proteins would allow for an improved understanding of the cytoskeletal organization and how and when during cell development the components organized themselves for functionality. Although antisera against cytoskeletal proteins could be generated and used to observe localization of the proteins in the cytoskeletal structure, our attempts to make antisera to MYPE1560 and MYPE1570 were unsuccessful (data not presented). Furthermore, the inability to generate specific antisera would not allow us to ascertain where proteins are located or the amount of protein needed to activate a pole for cytadherence and motility. Alternatively, cytoskeletal proteins fused to fluorescent proteins could be used to observe localization of these

99 proteins during cytadherence and motility. Although genetic manipulation is still not possible in M. penetrans, recent data suggest that M. iowae may be genetically manipulated by electroporation (Newman and Balish, unpublished). Preliminary analysis of the M. iowae genome identified homologs of the M. penetrans cytoskeletal proteins (M. Balish and Z. Raviv, unpublished data), supporting the morphological and functional similarities characterized between the two species. Genetic manipulation of M. iowae could therefore allow for characterization of both M. iowae and M. penetrans cytoskeletal components by characterizing the properties of individual cytoskeletal proteins in M. iowae. Similar studies have been used to characterize the functions of homologous proteins within the M. pneumoniae cluster to better elucidate the functioning of attachment organelle associate proteins in their respective species and phylogenetic relatives, such as studies of P30 homologs shared between M. pneumoniae and Mycoplasma genitalium (Relich and Balish, 2011). To analyze the mechanisms responsible for motility in M. penetrans we tested the effects of chemicals and environmental conditions on gliding motility to determine the energy source required for motility, discussed in chapter 3. We propose a model for M. penetrans gliding motility based on our characterization of attachment organelle function and organization, and the energetic studies of M. penetrans gliding motility (Figure 28). The positive correlation between temperature and gliding speed at biologically relevant pH levels supports a model in which thermal energy provides the direct energy needed for motility in the form of a "Brownian ratchet" (Peskin, 1993). In such a model, a growing polymer can exert force by adding monomers to a growing tip located behind a load, for example, actin polymerization at the cell membrane in lamellipodia of crawling cells (Peskin, 1993). In our system, we envision that the filaments observed by ECT consist of linear polymers which interact loosely with both the large TX-insoluble cytoskeletal structure and the cell membrane. In our model, the bonds between monomers in the filament are stable only if they are associated with the surface of the large cytoskeletal structure. To begin a cycle of motility (Figure 28, step 1), thermal energy stochastically causes the cell membrane to vibrate outward from the cytoskeletal structure and the filaments, establishing a new contact between a membrane-associated adhesin molecule, not yet identified, and its unidentified ligand on the gliding substrate. This step constitutes the forward motion, which changes the angles at which the other adhesins are engaged with the substrate. This angle favors detachment by the rearmost

100

Figure 28: Proposed model of M. penetrans gliding motility. Step 1: During M. penetrans gliding motility thermal energy is required for the random vibration of the cell membrane outward from the cell, allowing for an adhesin to make a new contact with an unidentified substrate molecule, which pulls the cell membrane forward and causes release of an adhesin- substrate bond at the rear. The forward movement of the cell creates a space between the cytoskeletal template and the cell membrane. Step 2: Monomers fill the space adjacent to the cell membrane at the tip. Step 3: Thermal energy randomly causes reverse vibration of the cell membrane, pushing the filaments back inward, causing the monomers at the leading end of the cell to be positioned in such a way that the cytoskeletal template promotes their addition to the filaments, and causing monomers at rear of the cytoskeleton to be positioned in such a way as to dissociate from the polymers.

101

102 adhesin, ensuring that motility is unidirectional. New filament monomers fill the space between the cytoskeletal structure and the cell membrane, but the monomers are not yet incorporated into the filament because they are not on the surface of the template (Figure 28, step 2). When thermal energy stochastically causes the cell membrane to vibrate backwards (Figure 28, step 3), these new monomers are pushed to a position over the cytoskeletal structure and therefore incorporated into the filaments. The addition of the monomers at this end pushes the rear ends to a position that is no longer associated with the cytoskeletal structure, causing the rearmost monomers to dissociate. Thus, the filaments experience no net growth. This model could be tested by identifying the proteins that constitute the filament and the large cytoskeletal template to characterize the interaction of the filament with the cytoskeletal template. Purified protein(s) of the filament could be used to observe filament formation in the presence or absence of the cytoskeletal template in vitro. Observing whether or not a filament can form properly in the absence or presence of the cytoskeletal template would provide evidence of the physical interaction between the two components, supporting our model that depends on this interaction. In our model, the M. penetrans attachment organelle resembles the belt and wheels of a tanker. The adhesins act like the belt of a tanker truck and the filaments act as the wheels of the truck. As the filaments (wheels) move forward, the adhesins (belt) continuously move in one direction. The difference in gliding speeds among M. penetrans strains might be caused by differences in the strength or number of adhesins, changing the likelihood of a new adhesin making contact with the substrate when the membrane stochastically vibrates forward. Alternatively, speed differences could be caused by differences in the concentration of monomers or in their affinity for the polymers. Both the monomers and the adhesins must be identified and characterized. To better understand the components and mechanism of the motility motor in M. penetrans, further energetic studies of gliding motility are necessary. The possibility remains that M. penetrans could contain ATP even after treatment with arsenate, so a series of ATP hydrolysis inhibitors could be used to observe their effects on motility. However, based on the inability of M. penetrans to grow in the presence of sodium arsenate, we predict that other inhibitors that interfere with ATP hydrolysis will have the same effect as arsenate. Although it is difficult to imagine significant amounts of cellular ATP after lengthy incubation in high concentrations of sodium arsenate, it could also be advantageous to formally measure the amount of ATP in M. penetrans before and after the addition of arsenate to demonstrate the ability of

103 sodium arsenate to effectively remove ATP and any reserves from M. penetrans during motility analysis. Identification of the M. penetrans adhesins utilized for attachment and motility will also be crucial in characterizing their functioning and activity during these processes, which can further improve our understanding of the mechanism that drives M. penetrans gliding. The information obtained from the experiments of this dissertation has increased our understanding of mycoplasma cytoskeletal structure. Previous research on M. penetrans focused on its clinical relevance with AIDS. While this topic is crucial for understanding the role of M. penetrans on the AIDS pandemic, it does not improve our understanding of the basic functions M. penetrans utilizes to infect and survive in its host. M. penetrans has been observed to attach to human epithelial cells (Lo et al., 1991), and the relationship between cytadherence and motility with virulence has been established in other pathogenic mycoplasma species (Jordan et al., 2007). Therefore, understanding cytadherence, motility, and cell division in the potential human opportunist M. penetrans has implications for the development of novel therapeutic strategies while AIDS research continues. Our characterization of M. penetrans internal composition and cytoskeleton is novel among the known Mycoplasma cytoskeletons. It is likely with this discovery of a fourth structure for generating mycoplasma polarity that other novel structures remain to be identified within the Mycoplasma genus. This work is further evidence that one Mycoplasma species cannot serve as a model for all mycoplasmas. Continued studies of other species in more poorly characterized phylogenetic clusters within the Mycoplasma genus will further our understanding of how mycoplasmas generate polarity and how the polarity-generating cytoskeletal structures have developed over the course of evolution. Additionally, studying polarity and its associated functions in pathogenic mycoplasma species will contribute to a better understanding of virulence mechanisms used by this genus to cause disease. Increasing the amount of information available in regard to how bacteria cause disease improves the likelihood of developing novel therapies as antibiotics become increasingly ineffective against many pathogenic bacteria, including mycoplasmas (Bebear et al.¸ 2011).

104

REFERENCES

Adams, D.W. & Errington, J. (2009). Bacterial cell division: assembly, maintenance and disassembly of the Z ring. Nat. Rev. Microbiol. 7, 642-653.

Atkinson, T.P., Balish M.F. & Waites, K.B. (2008). Epidemiology, clinical manifestations, pathogenesis and laboratory detection of Mycoplasma pneumoniae infections. FEMS Microbiol. Rev. 32, 956-973.

Ausmees, N. & Jacobs-Wagner, C. (2003). Spatial and temporal control of differentiation and cell cycle progression in Caulobacter crescentus. Cell. 115, 705-713.

Balish, M. F., Hahn, T.W., Popham, P.L. & Krause, D.C. (2001). Stability of Mycoplasma pneumoniae cytadherence-accessory protein HMW1 correlates with its association with the triton shell. J. Bacteriol. 183, 3680-3688.

Balish M.F. & Krause, D.C. (2002). Cytadherence and the cytoskeleton, p. 491-518. In S. Razin and R. Herrmann (ed.), Molecular biology and pathogenicity of mycoplasmas. Kluwer Academic/Plenum Publishers, New York, NY.

Balish, M.F., Ross, S.M., Fisseha, M. & Krause, D.C. (2003). Deletion analysis identifies key functional domains of the cytadherence-associated protein HMW2 of Mycoplasma pneumoniae. Mol. Microbiol. 50,1507-1516.

Balish, M.F. (2006). Subcellular structures of mycoplasmas. Front. Biosci. 11, 2017-2027.

Balish, M.F. & Krause, D.C. (2006). Mycoplasmas: a distinct cytoskeleton for wall-less bacteria. J. Mol. Microbiol. Technol. 11, 244-255.

Barber, T.L. & Fabricant, J. (1971). A suggested reclassification of avian Mycoplasma serotypes. Avian Dis. 15, 125-138.

Baseman, J.B., Cole, R.M., Krause, D.C. & Leith, D.K. (1982). Molecular basis for cytadsorption of Mycoplasma pneumoniae. J. Bacteriol. 151, 1514-1522.

Bebear, C., Pereyre, S. & Peuchant, O. (2011). Mycoplasma pneumoniae: susceptibility and resistance to antibiotics. Future Microbiol. 6, 423-431.

105

Benos, D.J. (1982). Amiloride: a molecular probe of sodium transport in tissues and cells. Am. J. Physiol. 242, 131-145.

Benyoucef, M., Rigaud, J.L. & Leblanc, G. (1981).The electrochecmial proton gradient in Mycoplasma cells. Eur. J. Biochem. 113, 491-498.

Berg, H.C. & Anderson, R. A. (1973). Bacteria swim by rotating their flagellar filaments. Nature. 19, 380-382.

Biberfeld, G. & Biberfeld, P. (1970). Ultrastructural features of Mycoplasma pneumoniae. J. Bacteriol. 102, 855-861.

Bradbury, J. M. & McCarthy, J.D. (1983). Pathogenicity of Mycoplasma iowae for chick embryos. Avian Pathol. 12, 483-496.

Bray, D. (2001). Cell movements: from molecules to motility, 2nd ed. Garland Publishing, New York, NY.

Bredt, W. (1968). Motility and multiplication of Mycoplasma pneumoniae. A phase contrast study. Pathol. Microbiol. 32, 321-326.

Brown, P.J.B., Kysela, D.T. & Brun, Y.V. (2011). Polarity and the diversity of growth mechanisms in bacteria. Semin. Cell Div. Biol. 22, 790-798.

Burrows, L.L. (2005). Weapons of mass retraction. Mol. Microbiol. 57, 878-888.

Cabeen, M.T., Charbon, G., Vollmer, W., Born, P., Ausmees, N., Weibel, D.B. & Jacobs- Wagner, C. (2009). Bacterial cell curvature through mechanical control of cell growth. EMBO J. 6, 1208-1219.

Cloward, J.M. & Krause, D.C. (2010). Functional domain analysis of the Mycoplasma pneumoniae co-chaperone TopJ. Mol. Microbiol. 77, 158-169.

Ebersbach, G., Briegel, A., Jensen, G.J., & Jacobs-Wagner, C. (2008). A self-associating protein critical for chromosome attachment, division, and polar organization in caulobacter. Cell. 19, 956-968.

106

Feldner, J., Gobel, U. & Bredt. W. (1982). Mycoplasma pneumoniae adhesin localized to tip structure by monoclonal antibody. Nature. 298, 765-767.

Feng, S.H. & Lo, S.C. (1994). Induced mouse spleen B-cell proliferation and secretion of immunoglobulin by lipid-associated membrane proteins of Mycoplasma fermentans incognitus and Mycoplasma penetrans. Infect. Immun. 62, 3916-3921.

Giron, J.A., Lange, M., & Baseman, J.B. (1996). Adherence, fibronectin binding and induction of cytoskeleton reorganization in cultured human cells by Mycoplasma penetrans. Infect. Immun. 64, 197-208.

Hasselbring, B.M., Jordan, J.L., & Krause, D.C. (2005). Mutant analysis reveals specific requirement for protein P30 in Mycoplasma pneumoniae gliding motility. J. Bacteriol. 187, 6281-6289.

Hasselbring, B.M., Jordan, J.L., Krause, R.W., & Krause, D.C. (2006). Terminal organelle development in the cell wall-less bacterium Mycoplasma pneumoniae. Proc. Natl. Acad. Sci. USA. 103, 16478-16483.

Hatchel, J.M., Balish, R.S., Duley, M.L., & Balish, M.F. (2006). Ultrastructure and gliding motility of Mycoplasma amphoriforme, a possible human respiratory pathogen. Microbiology. 152, 2181-2189.

Hatchel, J.M. & Balish, M.F. (2008). Attachment organelle ultrastructure correlates with phylogeny, not gliding motility properties, in Mycoplasma pneumoniae relatives. Microbiology. 154, 286-295.

Henderson, G.P. & Jensen, G.J. (2006). Three-dimensional structure of Mycoplasma pneumoniae’s attachment organelle and a model for its role in gliding motility. Mol. Microbiol. 60, 376-385.

Himmelreich, R., Hilbert, H., Plagens, H., Pirkl, E., Li, B., & Herrmann, R. (1996). Complete sequence analysis of the genome of the bacterium Mycoplasma pneumoniae. Nucleic Acids Res. 22, 4420-4449.

107

Hu, P.C., Cole, R.M., Huang, Y.S., Graham, T.A., Gardner, D.E., Collier, A.M., & Clyde, W.A. (1982). Mycoplasma pneumoniae infection: role of a surface protein in the attachment organelle. Science. 216, 313-315.

Jaffe, J.D., Miyata, M., & Berg, H.C. (2004a). Energetics of gliding motility in Mycoplasma mobile. J. Bacteriol. 186, 4254-4261.

Jaffe, J.D., Stange-Thomann, N., Smith, C., DeCaprio, D., Fisher, S., Butler, J., Calvo, S., Elkins, T., FitzGerald, M.G., Hafez, N., Kodira, C.D., Major, J., Wang, S., Wilkinson, J., Nicol, R., Nusbaum, C., Birren, B., Berg H.C., & Church, G.M. (2004b). The complete genome and proteome of Mycoplasma mobile. Genome Res. 8, 1447-1461.

Jordan, J.L., Berry, K.M, Balish, M.F., & Krause, D.C. (2001). Stability and subcellular localization of cytadherence-associated protein P65 in Mycoplasma pneumoniae. J. Bacteriol. 183, 7387-7391.

Jordan, J.L., Chang, H.Y., Balish, M.F., Holt, L.S., Bose, S.R., Hasselbring, B.M., Waldo III, R.H., Krunkosky, T.M., & Krause, D.C. (2007). Protein P200 is dispensable for Mycoplasma pneumoniae hemadsorption but not gliding motility or colonization of differentiated bronchial epithelium. Infect. Immun. 75, 518-522.

Jurkovic, D.A., Newman, J.T., & Balish, M.F. (2012). Conserved terminal organelle morphology and function in Mycoplasma penetrans and Mycoplasma iowae. J. Bacteriol. 194, 2877-2883

Kenri, T., Seto, S., Horina, A., Sasaki, Y., Sasaki, T., & Miyata, M. (2004). Use of fluorescent-protein tagging to determine the subcellular localization of Mycoplasma pneumoniae proteins encoded by the cytadherence regulatory locus. J. Bacteriol. 186, 6944-6955.

Kirchhoff, H., & Rosengarten, R. (1984). Isolation of motility mycoplasma from fish. J. Gen. Microbiol. 130, 2439-2445.

Kirkpatrick, C.L., & Viollier, P.H. (2011). Decoding Caulobacter development. FEMS Microbiol. Rev. 36, 193-205.

108

Knivett, V.A. (1954). The effect of arsenate on bacterial citrulline breakdown. Biochem. J. 56, 606-610.

Kojima, S., Yamamoto, K., Kawagishi, I. & Homma, M. (1999). The polar flagellar motor of Vibrio cholerae is driven by an Na+ motive force. J. Bacteriol. 181, 1927-1930.

Krause, D.C., Leith, D.K., Wilson, R.M. & Baseman, J.B. (1982). Identification of Mycoplasna pneumoniae proteins associated with hemadsorption and virulence. Infect. Immun. 35, 809-817.

Krause, D.C., & Baseman, J.B. (1983). Inhibition of Mycoplasma pneumoniae hemadsorption and adherence to respiratory epithelium by antibodies to a membrane protein. Infect. Immun. 39, 1180-1186.

Krause, D.C., Proft, T., Hedreyda, C.T., Hilbert, H., Plagens, H. & Hermann, R. (1997). Transposon mutagenesis reinforces the correlation between Mycoplasma pneumoniae cytoskeletal protein HMW2 and ctyadherence. J. Bacteriol. 179, 2668-2677.

Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 227, 680-685.

Ley, D.H., Marusak, R.A., Vivas, E.J., Barnes H.J., & Fletcher, O.J. (2010). Mycoplasma iowae associated with chondrodystrophy in commercial turkeys. Avian Pathol. 39, 87-93.

Linker, C., & Wilson, T.H. (1985). Sodium and proton transport in Mycoplasma gallisepticum. J. Bacteriol. 163, 1250-1257.

Lluch-Senar, M., Querol, E., & Piñol, J. (2010). Cell division in a minimal bacterium in the absence of ftsZ. Mol. Microbiol. 78, 278-289.

Lo, S.C., Hayes, M.M., Wang, R.Y., Pierce, P.F., Kotani, H., & Shih, J.W. (1991). Newly discovered mycoplasma isolated from patients infected with HIV. Lancet. 338, 1415-1418.

Lo, S.C., Hayes, N.M., Tully, J.G., Wang, R.Y., Kotani, H., Pierce, P.F., Rose D.L., & Shih, J.W. (1992). Mycoplasma penetrans sp. nov., from the urogenital tract of patients with AIDS. Int. J Syst. Bacteriol. 42, 357-364.

109

Luciano, J., Agrebi, R., LeGall A.V., Wartel, M., Fiegna, F. Ducret, A., Brochier-Armanet, C., & Mignot, T. (2001). Emergence of modular evolution of a novel motility machinery in bacteria. PLoS Genet. 7, e1002268.

McBride, M.J. (2004). Cytophaga-flavobacterium gliding motility. J. Mol. Microbiol. Biotechnol. 7, 63-71.

McCarter, L.L. (2004). Dual flagellar systems enable motility under different circumstances. J Mol. Microbiol. Biotechnol. 7, 18-29.

Meng, K.E. & Pfister, R.M. (1980). Intracellular structures of Mycoplasma pneumoniae revealed after membrane removal. J. Bacteriol. 144, 390-339.

Mignot, T., Shaevitz, J.W., Hartzell, P.L. & Zusman, D.R. (2007). Evidence that focal adhesion complexes power bacterial gliding motility. Science. 215, 853-856

Mignot, T. & Shaevitz, J.W. (2008). Active and passive mechanisms of intracellular transport and localization in bacteria. Curr. Opin. Microbiol. 11, 580-585.

Mirsalimi, S.M., Rosendal, S. & Julian, R.J. (1989). Colonization of the intestine of turkey embryos exposed to Mycoplasma iowae. Avian Dis. 33, 310-315.

Miyata, M. & Uenoyama, A. (2002). Movement on the cell surface of the gliding bacterium, Mycoplasma mobile, is limited to its head-like structure. FEMS Microbiol. Lett. 215, 285-289.

Miyata, M. & Peterson, J. (2004). Spike structures at interface between gliding Mycoplasma mobile cell and glass surface visualized by rapid-freeze and fracture electron microscopy. J Bacteriol. 186, 4382-4386.

Miyata, M. (2010). Unique centipede mechanism of Mycoplasma gliding. Annu. Rev. Microbio. 64, 519-537.

Morrison-Plumner, J., Leith, D.K. & Baseman, J. B. (1986). Biological effects of anti-lipid and anti-protein monoclonal antibodies on Mycoplasma pneumoniae. Infect. Immun. 53, 398- 403.

110

Nakane, D. & Miyata, M. (2007). Cytoskeletal "jellyfish" structure of Mycoplasma mobile. Proc. Natl. Acad. Sci. USA 104, 19518-19523.

Nan, B., Mauriello, E.M.F., Sun, I., Wong, A. & Zusman, D.R. (2010). A multi-protein complex from Mycococcus xanthus required for bacterial gliding motility. Mol. Microbiol. 76, 1539-1554.

Neyrolles, O., Brenner, C., Prevost, M., Fontaine, T., Montaginer, L., & Blanchard, A. (1998). Identification of two glycosylated components of Mycoplasma penetrans: a surface- exposed capsular polysaccharide and a glycolipid fraction. Microbiology. 144, 1247-1255.

Ohtani, N. & Miyata, M. (2007). Identification of a novel nucleoside triphosphatase from Mycoplasma mobile: a prime candidate motor for gliding motility. Biochem. J. 403, 71-77.

Peskin, C.S., Odell, G.M. & Oster, G.F. (1993). Cellular motions and thermal fluctuations: the Brownian ratchet. Biophys. J. 65, 316-324.

Radestock, U. & Bredt, W. (1977). Motility of Mycoplasma pneumoniae. J. Bacteriol. 129, 1495-1501.

Relich, R.F., Friedberg, A.J., & Balish, M.F. (2009). Novel cellular organization in a gliding mycoplasma, Mycoplasma insons. J. Bacteriol. 191, 5312-5314.

Relich, R.F., & Balish, M.F. (2011). Insights into the function of Mycoplasma pneumoniae protein P30 from orthologous gene replacement. Micrbiol. 157, 2862-2870.

Romero-Arroyo, C.E., Jordan, J., Peacock, S.J., Willby, M.W., Farmer, M.A., & Krause, D.C. (1999). Mycoplasma pneumoniae protein P30 is required for cytadherence and associated with proper cell development. J. Bacteriol. 181, 1071-1079.

Rosen, B.P. (2002). Biochemistry of arsenic detoxification. FEMS Lett. 529¸ 86-92.

Rosengarten, R., & Kirchhoff, H. (1987). Gliding motility of Mycoplasma sp. nov. strain 163K. J. Bacteriol. 169, 1891-1898

Sasaki, Y., Blanchard, A., Watson, H.L., Garcia, S. Dulioust, A., Montagnier, L., & Gougeon M.L. (1995). In vitro influence of Mycoplasma penetrans on activation of peripheral T

111 lymphocytes from healthy donors or human immunodeficiency virus-infected individuals. Infect. Immun. 63, 4277-4283.

Sasaki Y, Ishikawa, J., Yamashita, A., Oshima, K., Kenri, T., Furuya, K., Yoshino, C., Horino, A., Shiba, T., Sasaki, T., & Hattori, M. (2002). The complete genomic sequence of Mycoplasma penetrans, an intracellular bacterial pathogen in humans. Nucleic Acids Res. 23, 5293-5230.

Schiefer, H.G. & Schummer, U. (1982). The electrochemical potential across mycoplasmal membranes. Rev. Infect. Dis. 4, S65-S70.

Seto, S. & Miyata, M. (2003). Attachment organelle formation represented by localization of cytadherence proteins and formation of the electron-dense core in wild-type and mutant strains of Mycoplasma pneumoniae. J. Bacteriol. 185, 1082-1091.

Seto, S., Uenoyama, A., & Miyata, M. (2005). Identification of a 521-kilodalton protein (Gli521) involved in force generation or force transmission for Mycoplasma mobile gliding. J. Bacteriol. 187, 3502-3510.

Seybert, A., Herrmann, R., & Frangaskis, A.S. (2006). Structural analysis of Mycoplasma pneumoniae by cryo-electron tomography. J. Struct. Biol. 156, 342-354.

Shimizu, T., Kida, Y., & Kuwano, K. (2004). Lipid-associated membrane proteins of Mycoplasma fermentans and M. penetrans activate human immunodeficiency virus long- terminal repeats through Toll-like receptors. Immunology. 113, 121-129.

Stadtlander, C.T., & Kirchhoff, H. (1995). Attachment of Mycoplasma mobile 163 K to piscine gill arches and rakers—light, scanning and transmission electron microscopic findings. Br. Vet. J. 15, 89-100.

Stevens, M.K. & Krause, D.C. (1992). Mycoplasma pneumoniae cytadherence phase-variable protein HMW3 is a component of the attachment organelle. J. Bacteriol. 174, 4265-4274.

Sun, M., Wartel, M., Cascales, E., Shaevitz, J.W. & Mignot, T. (2011). Motor-driven intracellular transport powers bacterial gliding motility. Proc. Natl. Acad. Sci. USA. 18, 7559- 7564.

112

Szczepanek, S. M., Majumder, S., Sheppard, E.S., Liao, X., Rood, D., Tulman, E.R., Wyand, S., Krause, D.C., Silbart, L.K., & Geary, S.J. (2012). Vaccination of BALB/c mice with an avirulent Mycoplasma pneumoniae P30 mutant results in disease exacerbation upon challenge with a virulent strain. Infect. Immun. 80, 1007-1014.

Trampel, D.W. & Frederick G., Jr. (1994). Outbreak of Mycoplasma iowae infection in commercial turkey poults. Avian Dis. 38, 905-909.

Tully, J.G., Rose, D.L., Whitcomb, R.F., & Wenzel, R.P. (1979). Enhanced isolation of Mycoplasma pneumoniae from throat washings with a newly-modified cultured medium. J. Infect Dis. 139, 478-482.

Uenoyama, A., Kusumoto, A., & Miyata, M. (2004). Identification of a 349-kilodalton protein (Gli349) responsible for cytadherence and glass binding during gliding of Mycoplasma mobile. J. Bacteriol. 186, 1537-1545.

Uenoyama, A., & Miyata, M. (2005). Identification of a 123-kilodalton protein (Gli123) involved in machinery for gliding motility of Mycoplasma mobile. J. Bacteriol. 187, 5578-5584.

Wang, R.H., Shih, J.W., Grandinetti, T., Pierce, P.F., Hayes, M.M., Wear, D.J., Alter, H.J., and Lo, S.C. (1992). High frequency of antibodies to Mycoplasma penetrans in HIV-infected patients. Lancet. 340, 1312-1316.

Warburg, O. & Christian, W. (1939). Isolierung und Krystallisation des Proteins des oxydierenden Garungsferments. Biochem. Z. 303, 132-144.

White, C.L. & Gober, J.W. (2012). MreB: pilot or passenger of cell wall synthesis? Trends Microbiol. 20, 74-79.

White, D. (2007). The physiology and biochemistry of prokaryotes, 3rd ed. Oxford University Press, New York, NY.

Yañez, A., Cedillo, L., Neyrolles, O., Alonso, E., Prévost, M.C., Rojas, J., Watson, H.L., Blanchard, A., & Cassell, G.H. (1999). Mycoplasma penetrans bacteremia and primary antiphospholipid syndrome. Emerg. Infect. Dis. 5, 164-167.

113

Yoder H.W. & Hofstad, M.S. (1964). Characterization of avian mycoplasma. Avian Dis. 8, 481-512.

Young, K.D. (2006). The selective value of bacterial shape. Microbiol. Mol. Biol. Rev. 70, 660- 703.

114