Characterization of Early Biofilm Formation and Physiology in <I>Neisseria Gonorrhoeae</I>

City University of New York (CUNY) CUNY Academic Works

All Dissertations, Theses, and Capstone Projects Dissertations, Theses, and Capstone Projects

9-2019

Characterization of Early Biofilm ormationF and Physiology in Neisseria gonorrhoeae

Kelly Eckenrode The Graduate Center, City University of New York

How does access to this work benefit ou?y Let us know!

More information about this work at: https://academicworks.cuny.edu/gc_etds/3466 Discover additional works at: https://academicworks.cuny.edu

This work is made publicly available by the City University of New York (CUNY). Contact: [email protected] CHARACTERIZATION OF EARLY BIOFILM FORMATION AND PHYSIOLOGY IN NEISSERIA GONORRHOEAE

Kelly Eckenrode

A dissertation submitted to the Graduate Faculty in Biology in partial fulfillment of the requirements for the degree of Doctor of Philosophy, The City University of New York

2019

Kelly Eckenrode

ii CHARACTERIZATION OF EARLY BIOFILM FORMATION AND PHYSIOLOGY IN NEISSERIA GONORRHOEAE

Kelly Eckenrode

This manuscript has been read and accepted for the Graduate Faculty in Biology in satisfaction of the dissertation requirement for the degree of Doctor of Philosophy.

Date Nicolas Biais

Chair of Examining Committee

Date Christine Li

Executive Officer

Supervisory Committee:

Lars Dietrich Luis Quadri

Peter Lipke Theodore Muth

THE CITY UNIVERSITY OF NEW YORK

iii ABSTRACT AND AIMS

CHARACTERIZATION OF EARLY BIOFILM FORMATION AND PHYSIOLOGY IN NEISSERIA GONORRHOEAE

By Kelly Eckenrode

Advisor: Dr. Nicolas Biais

Many bacteria rely on the dynamics of their extracellular appendages to perform important tasks, like motility and biofilm formation. Interestingly, these dynamics have been linked to physiological responses in some pathogenic bacteria; therefore, it is important to understand more about the role of physical forces in bacteria. I used the causative agent of the human disease gonorrhea, Neisseria gonorrhoeae, as a model system to study the role of physical force on early biofilm formation. The advantage of this system is that cell-cell interactions are controlled by extracellular filaments called type IV pili (tfp). Tfp is composed of monomers that give bacteria the ability to produce a dynamic filament undergoing cycles of elongations and retractions, and thus to exert forces on their surroundings.

Through experiments and modeling, I demonstrated that pilus interactions produce motility gradients in microcolonies potentially establishing a force gradient across the microcolonies. I was interested in testing the biological implications of those motility and force gradients, so I utilized an established genetic mutant, ∆pilT, which lacks the pilus retraction motor pilT (Merz, So, and Sheetz 2000). A ∆pilT mutant allowed us to measure physiological response in cells that do not produce retractive force from its pilus. I measured

the level of gene expression of seven pilus-related genes in two backgrounds: WT and a pilus retraction-deficient mutant, ∆pilT. I found that some WT microcolonies express pilus-related genes in a heterogeneous fashion, while others are homogeneous. Spatiotemporal patterns in the microcolony are modified in a ∆pilT background. The presence or absence of retraction forces between bacteria have a profound impact on bacterial physiology: the WT and ∆pilT background do not survive in a classical static biofilm assay at the same rate. Together these results point toward a fundamental role for intracellular forces in shaping bacteria physiology.

The work of biologists has been dominated by a biochemical perspective. Although biochemical processes, like metabolism and information transfer, are certainly essential in all hierarchical levels of life, there is growing evidence that physical forces may provide an alternate physiological mechanism. The introduction in Chapter 1 provides context for understanding the role of force pattern formation in multicellular structures, in the hopes to extend this line of thinking to microbial communities.

The development of microbial communities relies on self-assembly of single cells. The development of Neisseria gonorrhoeae cellular aggregates rely exclusively on type IV pili interactions (Taktikos et al. 2015a). In Chapter 2 is a transcription of the publication where

I explore the dynamics of the microcolonies (W. Pönisch et al. 2018a). We found that cells have differential motility depending where in the microcolony cells are located. Differential motility is a result of fewer pili-pili interactions on the perimeter of the microcolony, and more pili-pili interactions closer to the center. Therefore, due to frequency of pili-pili interactions, a gradient of motility produces heterogeneous behavior in the microcolony.

To investigate whether heterogenous behavior is extended beyond motility, I investigated whether there is a connection between retraction force and the physiology of microcolonies. In Chapter 3 I used a quantitative approach to analyze seven pilus-related genes using fluorescent reporters. Using fluorescence and confocal microscopy, I quantified fluorescence intensity within space and time in microcolonies. Here, I provide evidence that physical intracellular cues in a three-dimensional bacterial aggregate provide context for spatial organization, since spatiotemporal patterning and survival in ∆pilT background are compromised in comparison to WT microcolonies. This suggests the important role PilT retraction force plays in regulating spatiotemporal patterning during early biofilm development.

Lastly, in Chapter 4 I characterized some physical features of microcolonies. I measured the formation size and survival rates of microcolonies when exposed to a range of osmotic pressures. These experiments were motivated by my interest in understanding the native context of developing microcolonies. Microcolonies inhabit the viscous mucosal membranes of epithelial cells; therefore, I measured one aspect of the environmental effects of microcolony when exposed to similar osmotic pressure created by mucus. I also measured the plasticity of WT and ∆pilT microcolonies through squeezing microplate experiments.

The overall aim of this work is to understand the role of physical force on microbial development. I largely focused on role of tfp forces on Neisseria gonorrhoeae microcolony formation. Characterizing gene expression in microcolonies provided key evidence for spatiotemporal heterogeneity in developing WT microcolonies. Heterogeneity was minimized without pilus retraction forces, which suggests that retraction forces play a role in the early development of biofilm formation.

ABSTRACT AND AIMS

CHARACTERIZATION OF EARLY BIOFILM FORMATION AND PHYSIOLOGY IN NEISSERIA GONORRHOEAE

By Kelly Eckenrode

Advisor: Dr. Nicolas Biais

vii

viii

ACKNOWLEDGEMENTS

I am grateful for so much and to so many people. Having the privilege to do a PhD program is a gift for which I will forever be thankful. In an era where truth is compromised, I am thankful for the City University of New York, Graduate Center and Brooklyn College for accepting me and helping me develop the intellectual tools required to navigate research and beyond. I will share what I have learned at CUNY to help others also navigate their future.

Thank you to CUNY Macaulay Honors College, the Instructional Technology Fellowship program and Lisa Brundage, for giving me an amazing opportunity to fund my PhD and develop new instructional tools for my future.

My advisor, Nicolas Biais was at the front lines of helping me develop into an independent thinker. My learning curve felt rather steep, so I really appreciate the patience and honest feedback. I like that I never had to guess what he really thought about something. ;) Most importantly, he showed me that there are as many ways to be a scientist as there are scientists, so my voice is as valuable as anyone else. Thank you, Nicolas, for sharing your contagious curiosity and unparalleled work ethic.

My thesis committee, Lars Dietrich, Peter Lipke, Anu Janakiraman, T.R Muth, Luis Quadri, who have been instrumental in developing my thinking and my research. They have collectively provided invaluable feedback to help me grow, especially as a writer. Thank you for your time serving on my committee and all the other budding graduate students in your career. You are helping shape the future of science even beyond your own amazing research.

My brilliant and supportive lab mates! The MMBL (mechano-micro-biology lab) is my home away from home, where a family of incredible lab mates work together, laugh together, and cry together (especially after the 2016 election). Jingbo Kan, thank you for helping me and sharing your thoughtful insights into my work. Thank you for being a great lunch buddy too!

John Eugenis, I’ll see you at the frontiers of mannose research. Olga Tsygelnytska and Aleks

Ratkiewicz, you give me so much hope for the future. Whatever you end up doing, I have complete faith that you will be the very best at it. And Mafer Seow, welcome to our lab. Please take good care of it and enjoy making it your own. <3

I want to extend many personal thank you’s to my friends and family. My husband, Ari

Winkleman who helped support my ambitions by being a super human. My mother, Judy, my father, Bob, my brother, Dan, and my nephew, Andrew who have shaped me into the person

I am today. Thank you for your support throughout my life. The Winklemans: Roberta,

Joseph, Dov, Freda (and Dan, Atria, Maor, Naava), Abby (and Reuven, Binyamin, Yaakov,

Tehilla, Shmuel, Ayelet) and Chicky (and Margaret), who have all been such positive examples of fierce individuality. All of you are inspirations to always be ourselves. My closest and dearest friends, Kim DeBarge and Emily Potter. I would not have gotten through my life and these last 7 years without your sisterhood. And lastly, my dog, Pepper who is the best ever.

DEDICATION

I wish to dedicate this thesis in memory of Joseph Winkleman. His spirit was unapologetically fierce, and I am thankful I was able to experience his determination to live fully and serve his fellow humans. May his memory be a blessing.

I also wish to dedicate this work to a set of influential teachers and past mentors:

Theresa Bidle of Hagerstown Community College. After I had taken every biology class the department offered over my 4-year stay in a 2-year degree, she told me, “I might like science more than I think”. I had never considered a career in science. I had no idea what a career in science would even mean. Frankly, I still feel like I’m learning what it means. But, I appreciate her insight in helping direct my passion for exploration.

My first mentor, Angela DePace at Harvard Medical School, who (I still can’t believe) let me into her new lab as her research technician. She established my absolute wonder for biology and research. Her leadership instilled the value of ‘giving down’ in research because research can only progress if we insist on training the next generation of scientists. Thank you from the very bottom of my heart. I am indebted to your generosity for including me in your work.

Lastly, Uri Alon at the Weizmann Institute of Science in Israel who told me that getting a PhD is a gift you give to yourself. I trusted you, even though I did not understand what that meant, but thank you for telling me because I get it now.

xii

Table of Contents ABSTRACT AND AIMS ...... iv ABSTRACT AND AIMS ...... vii ACKNOWLEDGEMENTS...... x DEDICATION ...... xii FIGURES ...... xv CHAPTER 1 ...... 1 Introduction and Background ...... 1 1: Introduction and background...... 2 1-1: Brief history of microbiology...... 2 1-2 Formation of heterogeneity in multicellular structures ...... 6 1-2.1: Central dogma of molecular biology ...... 6 1-2.2: Canonical models of pattern formation ...... 8 1-2.3: Mechanotransduction: a new model for pattern formation ...... 12 1-3: Neisseria biology ...... 15 1-3.1: Neisseria gonorrhoeae physiology ...... 15 1-3.2: Neisseria gonorrhoeae as a pathogen ...... 16 1-3.3: Surface appendages: Type IV Pilus biology ...... 19 1-3.4: Single cell to multicellular structures ...... 24 1-4: Significance ...... 27 CHAPTER 2 ...... 31 Pili mediated intercellular forces shape heterogeneous bacterial microcolonies prior to multicellular differentiation ...... 31 2-1: ABSTRACT ...... 32 2-2: INTRODUCTION ...... 33 2-3: RESULTS ...... 35 2-3.1: Neisseria gonorrhoeae microcolonies merge with dynamics consistent with a heterogeneous composition ...... 35 2-3.2: Neisseria gonorrhoeae microcolonies have an outer layer of motile cells and a core of far less motile cells thus presenting a heterogeneous motility behavior ...... 39 2-4: DISCUSSION ...... 47 2-5: METHODS...... 49 2-5.1: Bacteria strains and growth conditions ...... 49 2-5.2: Microcolony formation ...... 49 2-5.3: Microscopy and microcolony merger ...... 50 2-5.4: Scanning Electron Microscopy ...... 50 2-5.5: Image analysis of experimental data ...... 51 2-5.6: Simulations ...... 53 CHAPTER 3 ...... 56 Type IV retraction forces shape spatiotemporal physiology in Neisseria gonorrhoeae...... 56 3-1: ABSTRACT ...... 57 3-1: INTRODUCTION ...... 58 3-2: RESULTS ...... 61 3-2.1: Type IV pilus genes are expressed spatiotemporally in microcolonies ...... 61 3-2.2: Retraction-deficient microcolonies gene expression...... 75 3-2.3: Validation of mCherry fluorescence quantification ...... 81

xiii

3-2.4: Fluorescent in situ hybridization probes show a different pattern in comparison to genetic reporter pattern ...... 83 3-2.5: Artificial force applied to developing microcolonies alters differential spatiotemporal expression ...... 88 3-3.8: Survival assay of WT and ΔpilT ...... 92 3-2.8: Antibiotic resistance sensitivity in WT and retraction-deficient microcolonies ...... 97 3-2.9: RNA-Seq demonstrate a global differential expression in WT and ΔpilT ...... 99 3-3: DISCUSSION ...... 102 3-4: METHOD ...... 105 3-4.1: Culture conditions ...... 105 3-4.2: Promoter reporter construction ...... 105 3-4.3: Primers used for Neisseria gonorrhoeae pilus-related promoter amplification ...... 106 3-4.4: FISH probe sequence and fluorophore information...... 106 3-4.6: Microcolony assays ...... 108 3-4.8: Microscopy ...... 111 3-4.9: Survival and antibiotic resistance assays ...... 112 3-4.10: RNA purification for RNA-seq...... 112 3-4. 11: Principal Component Analysis for RNA-SEQ ...... 113 CHAPTER 4 ...... 116 Efforts to characterize physical properties of Neisseria gonorrhoeae microcolonies ...... 116 4-1: EFFECTS OF OSMOTIC PRESSURE ON DEVELOPING MICROCOLONIES...... 117 4-1.1: INTRODUCTION ...... 117 4-2: RESULTS ...... 119 4-2.1: Osmotic forces negatively effects microcolony formation in WT, but not ΔpilT ...... 119 4-2.2: Survival of WT is sensitive to high forces and ΔpilT is less sensitive ...... 121 4-2.3: Survival of WT and ΔpilT in Dextran-100, PEG-10 and sorbitol does not show any effect ...... 123 4-2.4: Pilin production decreased in higher osmotic pressure ...... 124 4-3: DISCUSSION ...... 126 4-4: MICROPLATE SQUEEZING EXPERIMENT TO MEASURE MICROCOLONY PLASTICITY ...... 127 4-4.1: INTRODUCTION ...... 127 4-5.2: RESULTS ...... 128 4-5.2.A: ΔpilT microcolonies are extremely plastic...... 128 4-5.3: DISCUSSION ...... 131 4-6: METHODS...... 132 4-6.1: Dextran microcolony assay ...... 132 4-6.2: Image analysis ...... 132 4-6.3: Western Blot ...... 133 4-6.4: Coomassie Stain ...... 137 4-6.5: Microplate squeezing experiments ...... 138 CHAPTER 5 ...... 139 Conclusions and future directions ...... 139 5-1: CONCLUSIONS ...... 140 Bibliography...... 145

xiv

FIGURES Figure 1: Leeuwenhoek's microscopy drawings ...... 3 Figure 2: Spatiotemporal patterns models in eukaryotes ...... 11 Figure 3: Schematic of a simplified type IV pilus in Ng ...... 21 Figure 4: Tfp mediated microcolony mergers are a common mode of microcolony formation ...... 37 Figure 5: Merging of Neisseria gonorrhoeae Microcolonies ...... 38 Figure 6: Motility of single cells inside a microcolony ...... 42 Figure 7: Demixing of Neisseria gonorrhoeae Microcolonies ...... 44 Figure 8: Heterogeneous genetic expression within a microcolony ...... 46 Figure 9A: Neisseria gonorrhoeae early microcolony development...... 60 Figure 9B: Genomic insertion region for transcription reporters…………………………………….60 Figure 10: Promoter assay demonstrates spatiotemporal gene expression patterns in Neisseria gonorrhoeae microcolonies ...... 63 Figure 11: Additional promoter assays demonstrates spatiotemporal gene expression patterns in Neisseria gonorrhoeae microcolonies ...... 66 Figure 12A: LacUV5-mCherry consensus reporter ...... 68 Figure 12B: pilA-mCherry reporter ...... 69 Figure 12C: pilB-mCherry reporter...... 70 Figure 12D: pilC1-mCherry reporter ...... 71 Figure 12E: pilC2-mCherry reporter ...... 72 Figure 12F: pilE-mCherry reporter ...... 73 Figure 12G: pilT-mCherry reporter ...... 74 Figure 12H: pilU-mCherry reporter...... 75 Figure 13A: ΔpilT-promoter assay demonstrates spatiotemporal gene expression patterns in Neisseria gonorrhoeae microcolonies ...... 77 Figure 13B: Minor ATPase, pilT2 and pilU, mutant promoter reporters at 3-hour and 7-hour time points ...... 78 Figure 14A: ΔpilT pilE-mCherry reporter spatiotemporal gradient analysis method (top row) ...... 79 Figure 14B: ΔpilT pilT- mCherry reporter spatiotemporal gradient analysis method (top row) ...... 80 Figure 15: Fluorescence intensity of LacUV5-mCherry (top) and pilE-mCherry (bottom) single cells (black), whole microcolony (red), non-motile single (blue), and fixed (yellow) 82 Figure 16A: FISH of pilE 373 at 7-hour...... 85 Figure 16B: FISH of pilT 347 at 7-hour ...... 86 Figure 16C: FISH of ΔpilT pilTE343 at 7-hour ...... 87 Figure 16D: FISH of ΔpilT pilT 347 at 7-hour ...... 88

Figure 17: Relative fluorescence integrated intensity of vibrated and not vibrated microcolonies ...... 90 Figure 18: Spatiotemporal patterning of pilE-mCherry is disrupted when exposed to continuous vibration for 7 hours ...... 91 Figure 19: Size of microcolonies (diameter) of vibrated and not vibrated in WT and ΔpilT background ...... 91 Figure 20A: Growth and survival of WT, ΔpilT, and mixture of WT, ΔpilT. CFU, colony forming units per mL plotted over time in hours ...... 93 Figure 20B: Growth and survival assay for wild type Ng, Ng ∆PilT and 1:1 mixture of both on KAN plate……………………………………………………………………………………………………….……...……………………...93 Figure 20C: Growth and survival assay for wild type Ng, complemented-Ng ∆PilT and 1:1 mixture on GCB….…………….…………………………………………………………………………………………….………………………94 Figure 21: Artificial external force applied assay early increases survival rate of WT and ΔpilT, while sustained force increases ΔpilE, mutant lacking tfp, survival ...... 97 Figure 22: Centripetal force exposure and no force exposure. Cell survival is measured by colony forming units ...... 98 Figure 23: Antibiotic exposure 24-hour survival assay ...... 99 Figure 24: DEG analysis of Neisseria gonorrhoeae WT compared to ΔpilT at 0, 12, and 24- hour incubation (left) ...... 100 Figure 25: FPKM of the top three differential expressed genes at all time points ...... 101 Figure 26: Global transcription fold change Neisseria gonorrhoeae WT compared to ΔpilT at 0, 12, and 24-hour incubation ...... 102 Figure 12I: Inhomogeneity indices of colonies with an exponential profile given by equation 1 for different value intensity ...... 111 Figure 27: Log Expression box plot of each sample demonstrates normalization after sequencing...... 114 Figure 28: Distribution after normalization of samples and replicates ...... 115 Figure 29: Microcolony morphology after 24-hour exposure to pressurized growth medium...... 120 Figure 30: Osmotic pressure negatively impacts microcolony colony size of WT, but with little impact on ΔpilT ...... 121 Figure 31: 24-hour survival in different osmotic pressure solutions of WT and ΔpilT ...... 122 Figure 32: Growth of WT at 7 hours ...... 123 Figure 33: Osmotic pressure survival assay with PEG-8000, Dextran-100, and Sorbitol ... 124 Figure 34: Coomassie gel of Neisseria gonorrhoeae samples after osmotic pressure exposure ...... 125 Figure 35: Western Blot of EFTU (top band) and SM-1 (pilin)...... 125 Figure 36: Experimental set-up of growth media containment chamber on the microscope stage (illuminated by green light) ...... 129 Figure 37: Plasticity of WT and ΔpilT ...... 130

xvi

CHAPTER 1

Introduction and Background

1: Introduction and background

1-1: Brief history of microbiology

The history of microbiology provides important context for the field’s past and present development. The humble origins of microbiology emerged from Robert Hooke’s technological advances and the curiosity of Anton van Leeuwenhoek in 1674. Van

Leeuwenhoek observed a world unseen to the naked eye by using Hooke’s primitive microscopy tools. The first microscopes were made of a magnifying lens crafted from soda lime glass and brass stage to position a water specimen. Found in fresh pond water, van

Leeuwenhoek discovered single-cell animals he described as animalcules, now known as protists. Intrigued by his initial discoveries, he found new liquid samples, like rain water, drain pipes, ocean water, and continued observing. When he sampled plaque from human mouths, he found hidden orders of being, even smaller than protists: bacteria (Lane 2015).

In exquisite detail, he documented the breadth of diversity of bacterial cell shape and motility (Figure 1) (Gest 2004).

Van Leeuwenhoek’s bacterial discovery caught the attention of the Royal Society of

London in 1683 (Leeuwenhoek 1683). The discovery of microbes challenged the current understanding of life: is it possible that larger organisms are built from smaller and smaller creatures, akin to Matryoshka nesting dolls? The idea that life is composed of smaller biotic forms fueled the spontaneous generation theory debate; a theory that proposed life can materialize without a biotic origin source. Van Leeuwenhoek was staunchly opposed to the spontaneous generation theory (Lane 2015).

The discovery of bacteria was not quickly, nor warmly, accepted by the scientific community. Van Leeuwenhoek refused to share his microscopes or lens-making techniques with others; researchers were unable to reproduce his findings with the lower quality lens available at the time (Ford 1993). In addition, his discovery was suspect due to his

‘unschooled origins’ and colloquial language in publications, which hurt his credibility (Lane

2015).

Figure 1: Leeuwenhoek's microscopy drawings. a) Rotifers, hydra and vorticellids associated with a duckweed root, from a Delft canal, Netherlands. b) Bacteria from Leeuwenhoek's mouth; the dotted line portrays movement. Copyright © The Royal Society (Lane 2015).

There is no doubt that bacteria exist, and not only do they exist, but exist in staggering proportions. Van Leeuwenhoek’s usage of microscope led to discoveries that prompted new questions about life at the microscale and still remain unanswered. What is the origin of bacteria? What is the cause of all the diversity? I will address aspects of the latter in section

1-2.3 “Mechanotransduction: a new model for pattern formation”. But, even today,

researchers lack resources to pinpoint the full scale of diversity of bacteria. Like. is it even possible to organize all of the diverse forms of bacteria into discrete categories?

A natural philosopher, Carolus Linnaeus, provided the first attempt of organizing organisms into a unified classification system. Linnaeus; however, did not attempt to classify microbes. He lumped all of microbes into the phylum Vermes (‘worms'), genus Chaos

(‘formless’) (Lane 2015). In his 1735 publication, Systema Natura, Linnaeus established a standardized system where organisms are classified into hierarchical categories by physical and metabolic similarities. Each category is a classification based on hierarchical levels: kingdom, phylum, class, order, genus, and species. For example, the human species are organized as: Animalia (Kingdom), Chordata (Phylum), Mammalia (Class), Primates (Order),

Hominidae (Family), Homo (Genus), sapiens (species).

Although, Linneaus was not the first to attempt organizing animals into a hierarchical tree of life (Aristotle, Historion Animalium), but the Linnaeus taxonomic system became the standard and largely remains today. The Linnaeus taxonomy system did not account for the origins of life and the creation of all its diversity. This required a new era of classification that emerged from the development of evolutionary phylogeny by Charles Darwin and

Ernest Haeckel. Evolutionary phylogeny asks questions regarding origins of ancestry, which ultimately facilitated the re-construction of the tree of life.

In 1977, the tree of life received its largest revision. Carl Woese emphasized the separation of Bacteria and Archaea as two kingdoms from Eukaryotes, to produce a three- domain tree of life. It was later refined as three domains by Woese, Otto Kandler and Mark

Wheelis by using sequencing of 16S rRNA. Due to the slow evolution rate of the 16S rRNA sequence region, categorizing distinct domains was created by shotgun 16S rRNA

sequencing. The addition of bacteria contributes an enormous number of branches and established a new view of life. As metagenomics and deep sequencing technology progress, the current taxonomic classification will be challenged and adapted as new information unfolds. In 2016, the most current version of the tree of life highlights underrepresented microbial lineages and identifies candidate radiations, which will be key for evolutionary analysis (Hug et al. 2016).

Bacteria contribute to the majority of biodiversity, yet our understanding of their life remains elusive. There are 500 times more bacteria in one human body than human beings on Earth, but researchers are only beginning to understand the role of human-host bacteria that composes the diverse and dynamic microbiome. During the first sequencing of the human genome in 2001, sequencing of oral cavity and gut bacteria identified potential commensal activity, hallmarking the beginning of microbiome research (Duncan, 2003).

Today, the National Institutes of Health have begun phase 2 of the Integrative Human

Microbiome Project, which was initiated in 2007 as a response to centralized methodologies and multi-omic data, specifically data from ‘healthy’ and ‘specific disease state’ populations

(“The Integrative Human Microbiome Project” 2019). The goal is to answer the question: how does the microbiome and their metabolites influence human health and disease? The recent emphasis on microbiota research has promise to uncover the roles of bacterial populations as a coordinated system in humans. However, there is still so much researchers do not understand about the formation of bacterial populations, which is the key step in understanding how populations are established.

Because of bacteria’s simplistic nature, they were an ideal first model organism for studying molecular biology processes beginning in the 1950s. The study of bacteria has crossed disciplines from biology to physics and back many times. Richard Feynman, physicist and educator, gave a (now famous) lecture in 1959 entitled “Plenty of Room at the

Bottom” (Feynman 1959). Feynman explored new avenues for information processing and different methods of data storage by looking at biological systems. He gave examples where nature had taken advantage of optimizing space through nano-scale data storage methods. Biological information systems, like DNA and RNA, are perfect examples of successful ‘miniaturized’ storage. Feynman believed there was much to learn about the world at the micro and nano scales. 60 years later, I believe there is still plenty of room to understand these ‘miniaturized’ lives, and the different embedded systems that encode information.

1-2 Formation of heterogeneity in multicellular structures

1-2.1: Central dogma of molecular biology

The central dogma of biology describes how hereditary information is stored and processed in cells. Although cells contain many forms of encoded language, the central dogma specifically refers to the chemical components. Nucleic acids, DNA

(deoxyribonucleotides) and RNA (ribonucleotides), act as the physical genetic information which encodes for mRNA and proteins. Decoding the genetic code is a dynamic interplay of cellular information flow, which begins with transcribing RNA from a DNA template to then translate into a functional protein composed of amino acids. Although a simplification, the central dogma of molecular biology provides the first model of the encryption events. The

central dogma was first published by Francis Crick in 1958 and re-phrased by James Watson ideas in the first edition of The Molecular Biology of the Gene in 1965.

While the central dogma explains a general flow of cellular information, it begs the question: How is the flow of information regulated? All organisms are the products of interwoven perturbations and parameters, but there is control of the information flow.

Genes and the associated regulatory regions can be compared to an electrical circuit: An input signal produces an output product where different forms of feedback can occur, either positive or negative feedback loops (Jacob and Monod 1959). An input signal is necessary to initiate transcription of a gene or gene set. Input signal source can be internal like, proteins, metabolites, or signal sources can be external sources like temperature, pH, cellular density, and in the case of this thesis, mechanical stimuli (see 1-2.3).

The information architecture of DNA provides insight into regulation. Promoters are stretches of DNA located directly upstream of a gene composed of sequence motifs coding for sequence-specific DNA-binding proteins called transcription factors. The presence or absence of particular transcription factors will regulate transcription by activating or repressing RNA polymerase recruitment. There are other DNA regulatory elements like, enhancers, silencers, insulators, 3’ and 5’ UTR (untranslated regions) were initially classified only in eukaryotes However, there is mounting evidence that bacteria contain as diverse regulatory elements (Maas 1991; Bervoets and Charlier 2019). In addition, regulation of protein activity also occurs during and after translation called post-translational modifications.

Understanding gene regulation is complicated by the diversity of input signals and by intrinsic variability of biological systems. While the genetic circuitry of a single cell is difficult to understand, the complexity is compounded when considering multicellular structures.

The next section couples molecular biology with the canonical models of eukaryotic heterogeneity that arise in multicellular systems during development.

1-2.2: Canonical models of pattern formation

The development of biological forms inspired a fundamental question in biology: How is diversity produced across life; while maintaining reproducibility within species? The body axis of humans is consistent, yet not a single person is identical to each other. This paradox led researchers to hypothesize about a cell-based developmental program, thus began efforts to understand pattern formation mechanisms during embryonic development.

Historically, the origin of pattern formation research with modeling the diffusion of a chemical throughout a multicellular structure. The two canonical eukaryotic pattern formation models arose from Alan Turing, reaction-diffusion model, and Lewis Wolpert, the positional information, i.e. “French Flag” model (Figure 2).

Alan Turing, an extraordinary mathematician, generated a model using first principles to derive the simplest self-organizing system possible. The initial assumption was all cells in a space field are homogeneous, then spatial coordination follows imparted by a chemical concentration creating a heterogeneous pattern. Turing explained pattern formation by diffusion, which was controversial since diffusion is characteristically attributed for uniformity. Yet, the interaction between an auto-inducing activator and a faster diffusing inhibitor can theoretically produce periodic phase of spots or stripes. Turing

also suggested that self-organizing systems are not created by only diffusing chemical concentrations, but by mechanical cues (Green and Sharpe 2015).

In contrast, Lewis Wolpert pitted the Turing model against his own positional information model. The positional information model assumes that all cells in a space field are not homogeneous (Wolpert and Sharpe 1989). Wolpert suggests that complex patterns are produced by inherent space field asymmetries. The asymmetry created a positional pole to initiate a space field coordinate system interpreted in a series of hierarchical chemical gradients named morphogens--a term coined by Turing.

An ideal biological discovery strengthened acceptance of Wolpert’s model: Early developmental spatiotemporal morphogen gradients were discovered in Drosophila melanogaster blastoderm embryos (Campos-Ortega and Hartenstein 1985). Maternally deposited mRNA, Bicoid, in the anterior end of Drosophila embryos diffused through the system and established asymmetrical positional information for development (Bejsovec and

Wieschaus 1993). The fruit fly embryo develops from one very large cell that is later divided into approximately 3000 cells by the invagination of membranes. These spatiotemporal morphogen gradients are composed of mainly transcription factors modulating gene activity by repression or activation or both, and generate the body axis: anterior to posterior, dorsal to ventral, and later further body segmentation (Nüsslein-Volhard, Wieschaus, and Kluding

1984).

While Turing proposed that long-range chemical information is dispersed somewhat evenly throughout the tissue via diffusion of molecules, Turing’s model was largely neglected until the last 30 years. Emerging work from developmental systems in mice demonstrate examples of Turing’s model. Reaction-diffusion systems are found during palatal shelf

(mouth roof ridges) development, in digit formation and left-right patterning. Based on these examples, there is reason to believe that the reaction-diffusion and positional information model are not mutually exclusive, but various systems for developing patterns (Economou et al. 2012).

Diffusion of molecules accounts for many developmental processes (Müller et al.

2013). Yet, diffusion cannot be the only transportation mechanism in development, since the diffusion rate is too slow (Howard, Grill, and Bois 2011b). Alone, reaction-diffusion and the positional information model fail to explain how the diffusion of morphogens induces cellular movement involved in morphogenesis (Milo and Phillips 2006.; Howard, Grill, and

Bois 2011b). Currently, a purely chemical model lacks the full narrative to the pattern formation story.

Figure 2: Spatiotemporal patterns models in eukaryotes. Reaction-diffusion proposed by Turning on the left (i-iii). Wolpert’s proposed positional information model on the right. Re-designed from original source.

While pattern formation was modeled exclusively in eukaryotes, microbial communities also demonstrate pattern formation. Bacterial colonies grown on soft-solid media demonstrate a range of diverse morphology: Color, texture, opacity, form, elevation, margins, and even smell. Diffusion, akin to Turing's reaction-diffusion model, is credited for the metabolite oscillations seen in bacterial colonies produced from single progenitor cells like Bacillus subtilis (Süel et al. 2006). However, not all microbial colonies are developed from one progenitor cells. Some microbial communities are composed through a repetitive process of adhering individual cells together.

Many bacteria communities are adhered to either air-solid surface, or a liquid-solid surface. For more information on microbial multicellular development, see chapter 1-3.4,

Single cell to multicellular structures. The physical properties required to produce an aggregate in liquid may differ from the air-solid interface; therefore, the differentiation process may require an alternate model to diffusion. In the next section, I will introduce an alternate self-organization model for producing differentiation: mechanotransduction.

1-2.3: Mechanotransduction: a new model for pattern formation

At planetary or atomic scales, movement dictates fate. Physical forces cause the displacement of mass to change its velocity. In biology, prime examples of physical forces are seen in motility, cell division, and morphogenesis. Is physical force merely a consequence of a series of actions, or do physical forces play a deeper biological role? Does physical force encode biological information in a similar way as DNA by providing an additional set of cues?

Here, I will introduce an emerging field, mechanobiology, while also illustrating prominent examples in the field.

Mechanobiology is the study of how cells detect, transmit, and respond to mechanical force (Thompson 1945; Paluch et al. 2015). In order detect a signal, a sensory system must be present, which consists of three components: a signal, transduction and receptor.

Mechanobiology considers all three sensory components. The signal can be a variety of physical forces, like stress, strain, movement. There are a wide range of mechanisms responsible for transducing signals, like membrane tension, pressure, protein conformational change. The signals are transduced to downstream receptors, which may ultimately play a role in molecular pathways. However, it is still quite difficult to parse

mechanical signaling mechanisms due to interplay of signals, transduction and receptors

(Schwartz and DeSimone 2008).

It is unknown whether different kinds of force inputs, like mechanical shear or stress, produce different phenotypic outputs. There are examples in eukaryotic cells that sense local forces, which produce systemic phenotypic changes. For example, soft tissues, like human lungs, respond to the rigidity of extracellular matrix. The tensile strength of the matrix determines the amount of cellular pulling and stretching on the substrate (Cui et al. 2015).

More critically, the amount of pulling experienced by the cell will dictate cellular fate in terms of growth, form and function (Engler et al. 2006; Nava, Raimondi, and Pietrabissa 2012).

Mechanical forces are ubiquitous phenomena experienced by all developing systems; therefore, I believe that mechanobiology is a reasonable approach for investigating pattern formation during multicellular development. The mechanical morphogenesis of gastrulation during embryogenesis is one of the most visually dramatic forms of mechanical influences in biology.

Even at the single cell level, eukaryotic stem cells demonstrate this principle powerfully during cell lineage differentiation (Le et al. 2016). Pluripotent stem cells differentiate by reading the physical features of the extracellular matrix, meaning mechanical forces alone can influence phenotypic on individual cell fates by the tensile strength of their surface medium (Trepat, Lenormand, and Fredberg 2008; Davidson, von

Dassow, and Zhou 2009). If a pluripotent stem cell is dividing on a hard surface, then the cell fate will likely be bone; intermediate rigidity will result in muscle; and soft substrate often give rise to fat and neural tissues (Engler et al. 2006). Essentially, stem cells will respond to physical stimuli without the initial assistance of a chemical cue. In reality, chemical

interactions are always present in living structures, so it can be difficult to parse the physical and chemical contributions. Regardless, physical forces play a key role in determining cellular fate because mechanotransduction pathways demonstrate the ability to regulate gene expression.

While mechanical forces on single cells translate into differential gene expression, what about in multicellular structures? In previous sections, I have established that multicellular structures differentiate through the pattern formation of morphogens. Do physical forces influence pattern formation? One theory, the differential adhesion theory, suggests that cells position themselves by cells with similar adhesive strength to maximize the intracellular bond, which will produce a more thermodynamically stable structure

(Davis, Phillips, and Steinberg 1997). Based on this principle, if various adhesion properties are present, then a heterogeneous multicellular structure will be produced.

Most importantly for this thesis, there is mounting evidence of mechanosensitive pathways in bacteria. In single cells of Pseudomonas aeruginosa, the retraction force of the type IV pilus (tfp) can signal differential gene expression. Physical interactions of the tfp with the surface transmits a stimulus via chemosensory system, Chp, a chemotaxis-like sensory system that regulates cAMP production with downregulates pilus expression and upregulates virulence genes (Persat 2017). The tfp retraction initiates the differential expression of hundreds of genes, including virulence factors (Persat, Inclan, et al. 2015b).

However, in Pseudomonas aeruginosa, it is difficult to parse the role of mechanical response from the chemosensory system. Understanding the role of mechanical signaling separate from the chemosensory system poses experimental challenges because it is difficult to measure an effect from the mechanical system apart from the chemical sensation.

Therefore, researchers have exploited the mechanical nature of Neisseria gonorrhoeae extracellular appendage tfp to study the role of force because Neisseria gonorrhoeae has no one chemosensory system. In addition, Neisseria gonorrhoeae constitutively expresses tfp as opposed to Pseudomonas aeruginosa, which expresses tfp only prior to biofilm development.

Past studies have suggested that Neisseria gonorrhoeae tfp may behave as a mechanosensitive channel by producing mechanical stress (Howie, Glogauer, and So 2005).

In Neisseria gonorrhoeae tfp retraction-deficient mutant, ΔpilT, downregulation of genes associated with stress–signaling pathways that alter epithelial cell gene expression were found in a transcriptional microarray assay (Dietrich et al. 2009). This evidence suggests that tfp retraction force powered by the ATPase motor, PilT, generates a stimulus to activate

Neisseria gonorrhoeae genes to enhance its infection pathway. How physical signals are converted into downstream pathways requires deeper study.

1-3: Neisseria biology

1-3.1: Neisseria gonorrhoeae physiology

The genus Neisseria was discovered by Albert Neisser in 1879. Neisseria are Gram negative proteobacteria live in a range of environmental niches and have diverse nutrient requirements. The Neisseria genus are mostly diplococcus, coffee bean, shaped; however, there are exceptions, such as the rod-shaped Neisseria bacilliformis and Neisseria elongata

(Bovre and Holton 1970). There are ten Neisseria species associated with commensalism in humans, and two pathogenic species meningitidis and gonorrhoeae. Neisseria gonorrhoeae is hypothesized to have evolved from a commensal Neisseria ancestor that colonizes mucous epithelial membranes of the oral cavity and the upper respiratory tract. A translocation event

from oral to genital cavity required evolving new mechanisms to survive its new ecological niche. Neisseria gonorrhoeae pathogenic nature is highlighted by its physiological and metabolic requirements. Neisseria gonorrhoeae, a fastidious facultative anaerobe, colonizes a range of human niches, yet the exact nutrient composition of each niche is unknown.

However, in the lab, Neisseria gonorrhoeae grow at specific conditions: 100% humidity at

37°C with 5% CO2 in complex media with supplemented source of glucose, glutamine, thiamine, phosphate, and iron (Jones and Talley 1977; Wong, Shockley, and Johnston 1980).

Iron is an essential element for metabolism and needs to be obtained extracellularly.

Since free iron is toxic, Neisseria gonorrhoeae has developed mechanisms for sequestering iron from its host and suppress iron redox reactivity. Neisseria gonorrhoeae incorporates iron intracellularly in proteins such as ferritin and hemoproteins, while also binding iron extracellularly by lactoferrin and transferrin (McKenna et al. 1988). Sequestering iron and other nutrients from a host contributes to Neisseria gonorrhoeae successful survival strategies as a pathogen. Largely, the physiology Neisseria gonorrhoeae is understood through the lens of its pathogenicity in humans.

1-3.2: Neisseria gonorrhoeae as a pathogen

Neisseria gonorrhoeae is a leading sexually transmitted infectious (STI) pathogenic bacterium, after human papillomavirus (HPV) and Chlamydia trachomatis, responsible for approximately 350,000 cases a year in the United States and 62 million cases worldwide

(CDC, 2012; Initiative for Vaccine Research (IVR) Sexually Transmitted Diseases—

Gonorrhea; Satterwhite et al. 2013). The first treatment for Neisseria gonorrhoeae infection was established in 1937, after the advent of bacteriostatic sulfonamides (Unemo and Shafer

2014). The incidence of Neisseria gonorrhoeae infection in both men and women has increased from 100.2 cases in 2010 to 107.4 cases in 2014 per 100,000 people. Along with the rise of infection incidences, antimicrobial drugs are becoming less effective due to increasing populations of drug resistant Neisseria gonorrhoeae strains. Nearly 30% of clinical isolates are resistant to some antibiotics according to the CDC, which poses an imminent health threat (“Gonorrhea - STD Information from CDC” 2012). As of 2011, a combination therapy consisting of a single intramuscular injection dose of ceftriaxone and oral azithromycin, is required to slow down the development of drug resistance (Tuddenham and

Ghanem 2015). No vaccine for Neisseria gonorrhoeae currently exists because of the lack of a reliable vaccine target, although researchers are still pursuing new targets (Cohen et al.

1994).

Like most pathogenic bacteria, there are many strains of Neisseria gonorrhoeae found in clinics. FA1090 and MS11 are two of the most commonly studied in the laboratory. It is important to note that the physiology of these strains might differ. In this work, I exclusively studied the MS11 strain because it is associated with high virulence due to the strong tfp retraction force previously characterized (Biais et al. 2008).

Neisseria gonorrhoeae pathogenesis occurs in four main steps: transmission, colonization, adaptation to the environment, and host-immune evasion. Since Neisseria gonorrhoeae cannot survive outside of a host, successful transmission is key for its survival.

The bulk of transmission of Neisseria gonorrhoeae occurs through unprotected sexual contact or maternal transmission during vaginal child birth. Although the precise transmission mechanism of Neisseria gonorrhoeae is unknown, one hypothesis suggests that

Neisseria gonorrhoeae are engulfed by host polymorphonuclear leukocytes (PMNs) and

transmitted during sexual intercourse in male seminal fluid. Although this does not explain how females transmit Neisseria gonorrhoeae to males (Quillin and Seifert 2018).

Colonization of Neisseria gonorrhoeae requires adherence to epithelium of the mucosa, which is coordinated with various bacterial surface adhesins (see 1-3.3 for more detail). The dynamics of the ubiquitously expressed tfp are the most critical for establishing residency.

Tfp are extracellular polymeric filament responsible for motility, cell-cell adherence, DNA uptake (natural transformation competence), and host-immune evasion via antigenic variation (Craig, Pique, and Tainer 2004; Higashi et al. 2007). Neisseria gonorrhoeae adapts to its new environment by acquiring host nutrients, in particular, siphoning iron. Other nutrients may be acquired as a result of the tissue damage causing an influx of during colonization. Neisseria gonorrhoeae encounter high amounts of various environmental stress, and different kinds of stress, like low oxygen concentrations, reactive-oxidative species, membrane stress, and antimicrobial efflux (Quillin and Seifert 2018). Each of these factors are correlated to transcriptional regulatory programs, which aid in the adaptation process. Lastly, Neisseria gonorrhoeae evades host immunity by two mechanisms. First,

Neisseria gonorrhoeae dismantles the human complement system by preventing the insertion of transmembrane membrane attack complexes, C3, within the bacterial surface.

Secondly, Neisseria gonorrhoeae protects itself by present host proteins, Factor H, which marks a cell as self. Neisseria gonorrhoeae survival is due to its ability to colonize and adapt and avoid the human immune system (Ngampasutadol et al. 2008).

Neisseria gonorrhoeae infection symptoms in males tends to be obvious due to painful urination and discharge from the penis. However, female infections are readily overlooked due to a pervasive misunderstanding of female symptoms. Researchers are only beginning

to appreciate and understand how sex differences shape infection responses (Edwards and

Apicella 2004). Neisseria gonorrhoeae infection can be asymptomatic and depending on the site of exposure, urogenital, rectal or pharyngeal, where mucous epithelial membranes are found. The major symptoms of urogenital infection in men include urethral discharge and dysuria; other complications include penile inflammation, abscesses, acute and seminal vesiculitis. The primary site for infection in women is at the mucosal columnar epithelium in the endocervix (Wang et al. 2017). Due to overlooked symptoms in women, improper diagnosis can lead to untreated chronic infections. From such infections, Neisseria gonorrhoeae can cause cervicitis, endometritis or inflammation in the fallopian tube.

Complicated ascending gonococcal infection can lead to pelvic inflammatory disease (PID), ectopic pregnancy and infertility.

1-3.3: Surface appendages: Type IV Pilus biology

Extracellular appendages aid in survival by helping microorganisms adapt to complex environments (Savage 1977; Aagaard et al. 2012; Cantarel, Lombard, and Henrissat 2012).

Probably due to the lack of organelle specialization inside of bacteria, prokaryotes have developed external specialized structural appendages. Bacterial appendages perform a wide range of tasks, like motility, secretion, adhesion, natural transformation, and virulence. The most commonly known appendage, the flagellum, is used for only a few of the aforementioned functions, while it also has the capacity for unique metabolic functions, like the oxidation of sulfur in hydrothermal vents miles from the ocean’s surface in some bacteria

(Sievert et al. 2008; Cooper 2000). Pili, sometimes known as fimbriae, accomplish some of the same functions as flagellum; however, their physiology and evolutionary path are vastly

different and make them distinct appendages by expanding their functional repertoire.

Specifically, the genetic occurrence of type IV pilus (tfp) is similar in nearly all bacteria and even archaea (Peabody et al. 2003). Tfp’s ubiquitous nature suggests a conserved role in the survival of microorganisms.

Tfp are dynamic appendages that perform tasks such as cell-cell adhesion, DNA uptake, and twitching motility, all of which are properties that contribute to Neisseria gonorrhoeae’s virulent capacity (Muschiol et al. 2015; Berry and Pelicic 2015). Tfp are helical filaments composed of protein subunits, pilE, requiring many other accessory proteins for proper biogenesis (Figure 3; Eriksson 2012). Neisseria gonorrhoeae are associated with 23 pilus-related proteins and 15 of those proteins are essential for functional biogenesis

(Goosens et al. 2017). The Tfp filaments are composed of modular ~15-20 kDa amphipathic protein subunits called pilins, produced from the pilE gene (Giltner, Nguyen, and Burrows

2012). Tfp is subcategorized into tfpa and tfpb, distinguished by the main pilin subunit, pilE, leader peptide sequence. Neisseria gonorrhoeae have tfpa pilus proteins. Structurally, pilin subunits are held together by tight-packing of the N-terminal α-helices, but loose packing of the C-terminal globular domains leaves substantial gaps on the filament surface. These gaps expose a glycine-rich, amphipathic segment of the N-terminal α-helix (Li et al. 2008).

An ATP-dependent polymerization and depolymerization process of pilin subunits is required for pilus extension and retraction. PilF and PilT are ATPase proteins located in the inner cytoplasmic membrane that are responsible for powering extension and retraction.

PilF supports pilin extension and PilT aids in pilus retraction. PilT depolymerizes PilE subunits from the extracellular space into the periplasmic membrane; however, both mechanisms are currently unknown (Merz, So, and Sheetz 2000; Satyshur et al. 2007; Misic,

Satyshur, and Forest 2010). PilT2 and PilU are also minor ATPase motors known to assist in fine-tuning of retraction speeds (Eriksson, Eriksson, and Jonsson 2012; Kurre et al. 2012).

The complete biogenesis mechanism of Tfp is unclear because the functions of all Tfp related proteins have not been resolved. Many of the accessory proteins, including PilE, PilF, and PilT, from the Tfp system have structural and functional homology with the type II secretion system (T2SS), which make pseudopilins in archaea species (Jarrell et al. 2013).

PilT sequence is not strictly conserved among archaea and bacteria, but structurally, FlaI is a homologous protein to PilT (Merz, So, and Sheetz 2000; Berry and Pelicic 2015). The shared homology suggests a common evolutionary origin, which aids researchers in understanding the unknown functions of proteins involved in producing a functional pilus

(Donlan 2002).

Figure 3: Schematic of a simplified type IV pilus in Ng. OM: Outer membrane; IM: inner membrane. Light blue ball: ComP; Magenta ball: PilX; Green ball: PilV. Rendered by Jens Eriksson, 2012

To identify which genes are necessary for tfp assembly, a reductionist experiment was performed where T2SS or Tfpb genes are transferred with entire corresponding operons to a surrogate non-piliated organism to determine which subset of genes are sufficient (d’Enfert, Ryter, and Pugsley 1987; Sohel et al. 1996; Goosens et al. 2017). Though the T2SS will not be discussed any further here, it is important to recognize the widespread homology of the pilus-like systems outside of Neisseria gonorrhoeae (Berry and Pelicic

2015). In addition, not all essential pilus genes are tfp structural genes. PilA and PilB are two- component histidine kinase essential for regulating pilE expression (Taha et al. 1990; Taha et al. 1991). pilA is also implicated in regulating cell growth (Taha et al. 1992).

Tfp can form bundles in the extracellular space by binding laterally to other pili subunits. Because of this lateral binding, Tfp are considered ‘sticky’ organelles. The protein filaments can extend their length by 100 times than the cell body, while only being 6nm wide

(Craig, Pique, and Tainer 2004). Additionally, in Neisseria gonorrhoeae, when Tfp is under tension the filament becomes 40% narrower than the unstretched conformation (Biais et al.

2010). Filament bundling and dynamic polymers of Tfp can enhance the cell’s interaction with the extracellular world. Tfp is essential and require extracellular protein machinery for microcolony formation in Neisseria gonorrhoeae. The mechanical forces exerted from the tfp from the cell body allows for a dynamic range of behavior. Tfp in Neisseria gonorrhoeae reveal a reversible force-dependent hidden epitope inside pilus subunits. When the pilus polymerizes, binds and stretches under tension, a hidden epitope is revealed in the hydrophobic region of the PilE monomer (Biais et al. 2010). This suggests the conformational change induced by mechanical force may uncover new tfp structural diversity. Tfp is the best

candidate for mechanosensing due to its physiology: external polymerization, persistence length, and dynamic nature.

Some extracellular appendages are more pervasive across microbes than others.

Nearly all Gram-negative and Gram-positive bacteria possess tfp filaments (Gurung et al.

2016). Tfp mediate an extraordinary array of functions, including adhesion, motility, microcolony formation and secretion of proteases and colonization factors. They play a vital role in virulence, and their ability to elicit an immune response make Tfp structures particularly relevant for study as targets for component vaccines and therapies, especially due to the ubiquitous nature of the tfp across microorganisms (Craig, Pique, and Tainer

2004).

It is important to add; Neisseria gonorrhoeae have outer membrane adhesion proteins beside tfp. A diverse set of other adhesion molecules such as lipopolysaccharides (LOS), opacity-associated proteins (OpA and OpC), porins (Por), and autotransporters (App and

MspA/AusI). Opa proteins are responsible for binding to CEACAM (carcinoembryonic antigen-related cell adhesion molecule) receptor, after the initial tfp adherence. During expression, OpA proteins demonstrate antigenic and phase variation and are structural similar to NspA (Neisseria surface protein A). OpC is only expressed in Neisseria meningitidis

(Zhu, Morelli, and Achtman 1999). The roles of the other adhesion protein are not well characterized, yet there is evidence they aid in establishing infection pathways (Hung and

Christodoulides 2013). Since, these proteins are not required for biofilm formation, they are not considered in this study. The expression of Opa proteins have a yellow-gold phenotype in the MS11 strain, the strain used in this thesis, when grown on an agar plate; therefore, colonies expressing Opa proteins were avoided from studies.

1-3.4: Single cell to multicellular structures

A perspective that bacteria live as single cells originates from Robert Koch’s pure- culture “golden criteria” methods used to study bacterial physiology (Davey and O’toole

2000). However, researchers now accept that most bacteria found in nature undergo communal development from single free-living cells to interdependent communities called biofilms (Costerton et al. 1995; Costerton, Stewart, and Greenberg 1999; R.M. Donlan and

Costerton 2002). Some bacteria form biofilms under environmental stress, while others form biofilms as default. Although there is no clear answer to when and why bacteria form biofilms, which occur in nearly all bacterial species.

Biofilms are formed using a variety of mechanisms depending on the specific bacterial species. Biofilms are surface, attached communities physically supported by an extracellular polymeric substrate (EPS) matrix. An EPS matrix is composed of mostly polysaccharides, lipids, eDNA (extracellular DNA). and extracellular appendages, like pilus. Some bacteria, like

Pseudomonas aeruginosa and Bacillus subtilis, form biofilms by regulating the production of exopolysaccharides (EPS). When EPS is secreted into the extracellular space it acts as molecular scaffolding to provide structure for cell-cell and cell-surface attachment.

Pseudomonas aeruginosa alter EPS chemical composition to produce small colony variants, which increases their ability to adapt (Mann and Wozniak 2012).

Exactly how biofilms are formed is complex and different between species (Davey and

O’Toole 2000). Microbial communities, biofilms, sustain individual bacteria adhered together as a collective structure either as a single species or multiple species. As a collective, these structures can form on biotic or abiotic surfaces at either submerged in liquid or at the

liquid-air interface. Across bacterial species, biofilm formation occurs in several transition states: Planktonic state, attach to surface, microcolony, biofilm, and dispersal. Biofilms structures are known to help protect the cells against the host defenses and also confer resistance to antibiotics (Lewis 2001). Microorganisms transition from a planktonic state to diverse attached community to confer a fitness advantage, like competition between other species, defense against antimicrobials and resource sharing (R.M. Donlan and Costerton

2002; Oliveira et al. 2015). Bacterial aggregates have higher tolerance for stress than freely suspended bacteria, which may be why cells produce microcolonies and biofilms (Haaber et al. 2012).

The physical parameters of single cells contribute to the cellular arrangement of biofilms. For instance, single cell morphology contributes to the entire arrangement of biofilms. In a recent study, a range of E. coli shapes were mixed together, coccus and rod shape cells, to observe whether shape influenced colony and biofilm packing arrangements.

Indeed, cell morphology drove various arrangements. All coccus, all rod, or mixed will drive different layered structures. Considering cell shape in addition to nutrient access, whether from above or below the biofilm, was also a key factor in shape arrangement (Smith et al.

2017).

Relative to this work, Neisseria gonorrhoeae also produce biofilms. Planktonic cells adhere together, without the excretion of EPS, by the adhesion of extracellular protein appendages by tfp. While some bacteria use quorum sensing to build microcolonies,

Neisseria gonorrhoeae relies solely on tfp to form microcolonies, and do not rely on chemosensory pathways, like cAMP, to induce tfp expression, as seen in Pseudomonas aeruginosa (Persat 2017). Microcolonies form within 90 minutes of interaction in liquid

culture, where, biofilms form after 48 hours, provided the appropriate nutrient conditions.

Currently, very little research is available on Neisseria gonorrhoeae’s various community structures.

Neisseria gonorrhoeae microcolonies are aggregates of 10s to 1000s bacterial cells typicall ~20µm in diameter (Taktikos et al. 2015b). It is unknown why the average packing size of cells in microcolonies is ~20µm. A Neisseria gonorrhoeae microcolony can be characterized into two morphotypes: spherical or asymmetric. The shape of the colony is directly correlated to the packing density of the cells. The density is determined by several factors: the number of cells, the amino acid sequence of extracellular pilin proteins, and the dynamics of the tfp (Craig, Pique, and Tainer 2004). Tfp filaments extend into extracellular space to bind to a variety of surfaces: Plastic, glass, or in the native conditions, epithelial lining. If a tfp binds to another bacterial tfp and it retracts, then bacteria will move more closely together. This “cast and pull” mechanism is repeated continuously until a microcolony is formed (Higashi et al. 2007). Tfp promotes bacteria-bacteria interaction from pilus-pilus contacts that can lead to microcolonies (Baker, Biais, and Tama 2013). Neisseria gonorrhoeae microcolony formation can be seen with light microscopy after approximately

3 hours incubation at 37°C (Taktikos et al. 2015b). Any time after 24 hours, at proper conditions, Neisseria gonorrhoeae biofilm formation may occur. Microcolony formation is the necessary for colonization. In addition, the activity of Tfp retraction triggers host cell response (Higashi et al. 2007).

Even as late as 1987, biofilms were considered slabs of homogeneous material without any specialization (Costerton et al. 1987). Cellular differentiation in biofilms does occur, and cell fate is predicted to establish prior to biofilm maturation (Kouzel, Oldewurtel,

and Maier 2015). Not until recently were biofilms demonstrated not to be homogenous, yet heterogeneous mixture of specialized cells with a variety of phenotypes (Sauer et al. 2002;

Berk et al. 2012; Chimileski, Franklin, and Papke 2014). Gene regulation of cells aid in differentiation to perform unique tasks requiring multicellularity. Most notable, Myxococcus spp. differentiate during starvation to form fruiting bodies, Bacillus subtilis can transform into spores, and Streptomyces coelicolor develop varying morphologies in response to nutritional conditions (Straight, Willey, and Kolter 2006). At first, it is not obvious that microorganisms could exploit intercellular interactions and communication to coordinate collective behavior. It is my assertion, and others, that pattern formation can be measured in biofilms to understand the development of heterogeneity (Vlamakis et al. 2008).

Although the role of physical forces in microbial life cycle is not well understood at this time, I have provided evidence where physical force influence microbial development.

Physical forces are crucial for the motility, formation of multicellular aggregates, and initiation of pathogenic response during colonization. All of these aspects are imperative for the survival of microbes; therefore, it is urgent to understand the mechanosensitive mechanisms, so researchers can develop new antimicrobial targets.

1-4: Significance

Since the advent of antibiotics as medicine, in 1928 by Alexander Fleming, their overuse has created a new public health problem: Antibiotic resistant bacteria. In 2012, the

Center of Disease Control (CDC) reported antibiotic resistant bacteria were responsible for the death of 23,000 patients a year in the United States. With the global emergence of more

antibiotic resistant bacteria strains, the pattern suggests this rate will continue (“Antibiotic

/ Antimicrobial Resistance | CDC” 2017).

The first commercially available antibiotic, sulfonamides, became available in 1935.

Within ten years, the first sulfonamide resistant bacteria were reported. This was the beginning of a cyclic problem. After penicillin, tetracycline, streptomycin, chloramphenicol, and ampicillin were introduced commercially, antibiotic resistant strains began to emerge within 10 years (Davies and Davies 2010). Today, some antibiotic resistant strains emerge less than 2 years into production. Due to the high rate of antibiotic resistance emergence, researchers are fighting against time to understand more about the physiology of bacteria.

Encountering an antibiotic-resistance bacterium is not a rare event. One in 31 patients are diagnosed with a hospital acquired infection (HAI), which makes hospitals the highest-risk venue for encountering an antibiotic resistant bacteria (“Antibiotic /

Antimicrobial Resistance | CDC” 2017). In addition, since the 1990s, there has been a global increase of antibiotic resistance bacteria found in food-borne bacterial infections, especially

Salmonella enterica and Campylobacter. One study found that food from animal origin contained 40% of bacteria that are resistant to one or more antibiotic (Ventola 2015). The high incidence of food-borne antibiotic bacteria is caused from the frequent exposure to antibiotics during agricultural food processing. Resistant bacteria carry antibiotic resistant genes, which could be passed to the human intestinal microbiota.

Perhaps the production of more antibiotics is not the solution against antibiotic resistant bacteria. It appears introducing more antibiotics into the environment is only encourages microbes to evolve their antibiotic defenses. It is imperative to understand more

about relationship between bacteria and humans, and more broadly, bacterial physiology, in order to get ahead of the antibiotic resistance crisis.

Recently, microbiologists have begun researching the origins and evolutionary path of microbes and humans. One example is the evolutionary path of pathogenic Neisseria species evolving from an ancestral commensal Neisseria species. Neisseria have co-evolved with human, and along Neisseria’s evolutionary path, some species adapted the capacity to cause host damage, while others did not adapt pathogenicity (Seifert 2019). Specifically, two pathogenic species of the Neisseria genus, gonorrhoeae and meningitidis, adapted pathogenicity abilities from its original commensal origin during a translocation event from oral to genital region, which changed the bacteria’s environmental context. So, what induced the pathogenicity? Was it the translocation event to a new environment that induced evolution? Or did a particular commensal strain already have the pathogenic ability prior to translocation?

This is not a new dilemma. In the 1860s, Louis Pasteur and Antoine Béchamp debated between Pasteur’s “Germ Theory of disease” versus Béchamp’s “Terrain Theory of disease”.

While germ theory is the modern consensus that bacteria are pathogenic agents; terrain theory states the environment selects for pathogenicity, which has a new relevance today.

Modern geneticists, like Rob DeSalle, claim different bacterial communities live in different places (DeSalle and Wynne 2018.). Context of where bacteria’s geography guides their survival and adaptation path, similar to the Neisseria commensal to pathogen example.

There is an evolutionary relationship between the bacteria and their environmental context.

Environmental context, both the chemical and mechanical nature, informs bacterial

physiology. It is my opinion that mechanical cues from the bacteria’s niche may inform its evolution.

Biofilms grow on a wide variety of both abiotic and biotic surfaces. As a result, clinicians routinely fight against biofilm-forming bacteria on implanted medical devices, like catheters and plastic implants. Therefore, investigating the mechanical events, like the formation of microcolonies and biofilms, are key to understanding the physiology of bacteria.

The result of this work provides a new relationship between type IV pilus and physiology bacterial biofilm architecture. Understanding the physiology of bacteria biofilms provides alternative treatments for bacterial infections. Mechanomicrobiology is paving new frontiers which will hopefully lead to new therapeutic targets against pathogens.

CHAPTER 2

Pili mediated intercellular forces shape heterogeneous bacterial microcolonies prior to multicellular differentiation

Wolfram Pönisch, Kelly B. Eckenrode, Khaled Alzurqa, Hadi Nasrollahi, Christoph Weber, Vasily Zaburdaev & Nicolas Biais in Scientific Reports Volume 8, Article number: 16567 (2018).

This chapter is transcription of the paper written by Ponisch, Zaburdaev, and Biais

Author Contributions: Experiments performed by W. P, K.B.E, K.A and H.N. Analysis performed by W.P. I assisted minorly in the final editing. I assisted in the final editing and the design of the experiments

2-1: ABSTRACT

Microcolonies are aggregates of a few dozen to a few thousand cells exhibited by many bacteria. The formation of microcolonies is a crucial step towards the formation of more mature bacterial communities known as biofilms, but also marks a significant change in bacterial physiology. Within a microcolony, bacteria forgo a single cell lifestyle for a communal lifestyle hallmarked by high cell density and physical interactions between cells potentially altering their behavior. It is thus crucial to understand how initially identical single cells start to behave differently while assembling in these tight communities. Here we show that cells in the microcolonies formed by the human pathogen Neisseria gonorrhoeae present differential motility behaviors within an hour upon colony formation. Observation of merging microcolonies and tracking of single cells within microcolonies reveal a heterogeneous motility behavior: cells close to the surface of the microcolony exhibit a much higher motility compared to cells towards the center. Numerical simulations of a biophysical model for the microcolonies at the single cell level suggest that the emergence of differential behavior within a multicellular microcolony of otherwise identical cells is of mechanical origin. It could suggest a route toward further bacterial differentiation and ultimately mature biofilms.

2-2: INTRODUCTION

It is now broadly accepted that bacteria principally exist as surface-associated communities called biofilms (Stoodley et al. 2002; G. A. O’Toole and Kolter 1998) Formation and survival of biofilms are a major concern, both in a medical and industrial context (R.M. Donlan and

Costerton 2002; Costerton, Stewart, and Greenberg 1999; Tan et al. 2014). On the other side, biofilms can also provide useful applications for wastewater treatment (Nicolella, van

Loosdrecht, and Heijnen 2000) and are important for the proper functioning of many ecosystems (Hunter 2008). The early stages of biofilm development are usually characterized by the formation of tethered small aggregates, so-called microcolonies, either by successive recruitments of new bacteria from the surrounding bulk fluid, multiplication of adhered bacteria or aggregation of bacteria actively moving on a surface (G. A. O’Toole and

Kolter 1998). Early microcolonies are comprised of dozens to thousands of cells, are often assembled in matter of hours and have been observed in many different bacteria species

(Thomason 1971; Thormann et al. 2004). Microcolonies represent the first stage of a usually complex development into mature differentiated multicellular biofilms (Stoodley et al.

2002). However, microcolonies are also commonly found by themselves in vivo (Thomason

1971; Zijnge et al. 2010; J L Edwards et al. 2000; Steichen et al. 2008). For instance, a recent study has mapped out the advantage that microcolonies could represent for infection of the human pathogen Neisseria meningitidis (Bonazzi et al. 2018). Microcolonies thus also play an important role beyond that of an intermediate towards biofilm maturation. The first measurable step towards obtaining cells with different biological functions, the hallmark of differentiation, will be heterogeneous gene expression among the cell population.

Understanding the nature of this transition from unicellular to multicellular lifestyle holds

the promise for unraveling some of the mechanisms leading to differentiation and controlling biofilm development.

While eukaryotes and prokaryotes are obviously dissimilar, a parallel can be drawn between the development of bacterial biofilm and the development of complex multicellular organisms (G. O’Toole, Kaplan, and Kolter 2000). The cues that dictate the cascade of signaling that governs multicellular development have long been thought to be of chemical origin but recent studies have brought to light the importance of the mechanical interactions between eukaryotic cells in the differentiation process (Howard, Grill, and Bois 2011a;

(Farge 2003; Mitrossilis et al. 2017). With the knowledge accumulated in the dynamics of eukaryotic cell aggregates in mind, we will explore here the internal dynamics of bacterial microcolonies.

In many bacterial species, the events leading to microcolonies are highly regulated with ultimately many different interactions governing the adhesion of bacterial cells to the substrate and between bacterial cells (Zhao et al. 2013; G. A. O’Toole and Kolter 1998). Often biofilms are held together by an extracellular matrix composed of DNA, excreted polysaccharides and various bacterial appendages (Drescher et al. 2016; Klausen et al. 2003;

Berk et al. 2012). In contrast, the formation of microcolonies of Neisseria gonorrhoeae is solely relying on the interactions mediated by a ubiquitous appendage, the Type IV pilus

(Tfp)(Higashi et al. 2007; Berry and Pelicic 2015). Mutants lacking Tfp are not able to form microcolonies(Taktikos et al. 2015b). The unique reliance on Tfp makes Neisseria gonorrhoeae an ideal model system to fully understand the dynamics of formation of bacterial microcolonies. In this study, we look experimentally at the dynamics of formation

of Neisseria gonorrhoeae microcolonies and highlight the crucial role of the mechanical forces generated by retractile Tfp in this process. Our central result is the discovery of emerging heterogeneous behavior within bacterial microcolonies within the first hours of formation. We observe a sharp gradient of bacterial motility from mobile surface layer towards nearly immobile bulk of the microcolony. These results are corroborated by experiments with bacteria incapable of Tfp retraction and comparison with the predictions of the in silico model we recently developed (W. Pönisch et al. 2017). Ultimately, we see that heterogeneous gene expression follows the heterogeneous motile behavior.

2-3: RESULTS

2-3.1: Neisseria gonorrhoeae microcolonies merge with dynamics consistent with a heterogeneous composition

Tfp are retractile bacterial appendages whose cycles of elongation and retraction enable bacteria to exert forces on their surroundings (Berry and Pelicic 2015; Merz, So, and Sheetz

2000). These polymers have a diameter of molecular size (below 10 nm) and length exceeding the size of the bacteria body (several microns)(Berry and Pelicic 2015) An average

Neisseria gonorrhoeae cell has 10–20 Tfps. Tfp can generate forces up to the nanonewton range when in bundles (Biais et al. 2008). In the case of Ng, Tfp are the only motility appendage that the bacteria possess. This leaves the cycles of elongation and retraction of

Tfp and the forces that Tfp can exert on their surroundings as the principal agents of microcolony formation. Neisseria gonorrhoeae bacteria can form nearly spherical microcolonies of upward to thousands of cells within a few hours, which greatly facilitates their study (Fig. 4a,). The active merging of smaller microcolonies into a larger one is the central mechanism responsible for microcolony growth (Taktikos et al. 2015b) (Fig. 4a–c).

We took advantage of the fact that the merger of microcolonies necessitates a complex rearrangement of cells and thus will inform us on the internal dynamics of bacterial microcolonies. To this end, we studied in detail the dynamics of two merging microcolonies.

Microcolonies were self-assembled by letting bacteria interact with each other on a surface.

Microcolonies of the desired size could be subsequently retrieved and brought into close vicinity and let to interact under a microscope. To quantify the transition of two interacting colonies towards a spherical shape we used the images at the midplane cross section. By fitting an ellipse to the shape of the cross section we measured the aggregate’s short and long symmetry axis. Their ratio approaches 1 from below as the colony rounds up (See Figs 5a).

Additionally, we measured the height of the “bridge” – a contact area forming between the two touching microcolonies. Our analysis shows that the merger occurs with three different dynamical regimes. In the first rapid regime (a few seconds) the two microcolonies are pulled together by retracting pili (retracting speed of Tfp is of the order of 1 μm/s and hence the time scale). In the intermediate regime (a few minutes) the two microcolonies smoothen the gap in the contact area and attain an ellipsoidal shape. Finally, the slowest regime (half an hour to more than one hour) is where the two microcolonies round up to a sphere. To follow the mixing of cells during merger we used fluorescently labeled bacteria creating microcolonies with two different colors (expressing either YFP or tdTomato fluorescent proteins). The merger showed a mostly flat contact region (See Fig. 5c). These findings would be consistent with a heterogeneous microcolony where an envelope of mostly motile cells drives the relaxation in the area with highest surface curvature (at the contact line between microcolonies) and is resisted by a core of many fewer motile cells responsible for the longer relaxation time. To probe directly for the existence of such “liquid-like” envelope and a more

“solid-like” core, we decided to follow the motility of single cells as a function of their position in a microcolony.

Figure 4: Tfp mediated microcolony mergers are a common mode of microcolony formation. (a) DIC micrographs of a time series showing two microcolonies merging among many interacting bacteria. See also Supplementary Movie S1. Scale bar = 8 µm. (b) Scanning Electron Micrograph of two merging microcolonies. Scale bar = 8 µm. (c) Scanning Electron Micrograph of the bridge formed between two merging microcolonies. Scale bar = 4 µm.

Figure 5: Merging of Neisseria gonorrhoeae Microcolonies. (a) Merger of Neisseria gonorrhoeae microcolonies recorded with a DIC microscope (Scale bar = 10 µm). The red line highlights the detected edges. (b) In order to estimate the time scales of the merging of two colonies the time- dependent bridge height and the symmetry axis ratio of a fitted ellipse were measured from the binary images. By fitting a function of the form h(t)=h0⋅(1−α⋅exp−t/τ1−(1−α)⋅exp−t/τ2) to the bridge height we were able to estimate the first time scale corresponding to the initial approach of the two colonies and the time scale characterizing the closure of the bridge. The third time scale resulted from a fit to the aspect ratio of the short axis and the long axis of the ellipse and corresponds to the

relaxation of the ellipsoidal colony to a spherical shape. (c) Merger of two fluorescently labeled microcolonies.

2-3.2: Neisseria gonorrhoeae microcolonies have an outer layer of motile cells and a core of far less motile cells thus presenting a heterogeneous motility behavior

To look at the motility of single cells within microcolonies we used a mixture of wild-type

(WT) non-labeled cells mixed with a small percentage (5 to 10%) of fluorescently labeled bacteria. We used fluorescent and bright field channels to simultaneously record bacterial tracks and determine their location within the microcolony. To quantify the motility of cells in the microcolonies without being hindered by the global motion of the microcolony, we evaluated the changes in relative distance for pairs of cells at a given depth from the colony surface. The mean squared relative distance increments (MSRD) as a function of time provide information about the motility of cells (Fig. 6a,b, Methods section)(W. & Z. Pönisch

2018). For short time scales (<1 minute) and for all depths the MSRD grows linearly with time, consistent with normal diffusive behavior of cells. However, the intensity of the random motion (quantified by the value of the diffusion constant) declines sharply for bacteria at different depth away from the microcolony surface (deeper into the microcolony) (Fig. 6c,d).

These results indicate that Neisseria gonorrhoeae bacteria as they assemble into microcolonies create a microenvironment with strikingly different motility behaviors. In order to investigate the nature of the emergence of differential motility within a microcolony we turned to computer modeling. We use a numerical simulation of a bacterial microcolony at the single cell level of detail with explicit pili-pili interactions (Pönisch et al. 2017). This model accounts for known Tfp dynamics and force generation, along with the physical shape

of cells and reaches the limits of current computational time constraints (see Methods). The simulations recapitulate the motility behavior observed experimentally: the mean square displacements of bacteria pairs show diffusive motility throughout a microcolony at short time scales, but the value of the associated diffusion constant diminishes greatly for the bacteria tracked deeper towards the core of the colony (Fig. 6e,f). Importantly, since all cells in our computational study have identical features, the emergence of the heterogeneous behavior highlights the importance to take a dynamical approach when analyzing any aggregation of cells. An energetic approach that would only take into account the difference in adhesion properties between the cells will most likely miss the complex reality that a more precise and dynamical approach might uncover (Pawlizak et al. 2015). In the simulations we can correlate the heterogeneous motility of cells with an average number of Tfp bound to pili of other cells and the times of presence and attachment of each pilus. The number of attached pili slowly decreases going from the core of the microcolony to the outer layer as the fluctuations in the number of pili (attached or not) increases. Additionally, in the simulations the attachment times rapidly decrease when going from the core to the outer layer. For the cells deep in the core of the colony, even if one of their pili detaches, the multitude of remaining contacts holds the cell in place long enough for the detached pilus to bind again, thus leading to an effectively very slow dynamics in the core. Due to a higher cell density compared to the colony surface, within the bulk the pili density is increased, thus increasing also the re-binding rate of a pilus. In other words, cells in the core are more likely to be bound to each other than cells on the outside. The qualitative agreement between experiments and simulations indicates that the occurrence of different motility behaviors within a microcolony can be explained solely by the intercellular forces exerted by Tfp. If we use the

same model to determine the dynamic of microcolonies merger we obtain similarly good qualitative agreement with the experiments. Our model contains just a few free parameters with the other parameters provided by literature, however, while reproducing the data qualitatively it still deviates in numerical values. One observation is that our in silico colonies are less heterogeneous as compared to experiments with a sharper transition. Most likely the discrepancy is due to the simplification of the model allowing for one pilus to have at maximum one contact point with any other pili of other cells. While dictated by computer feasibility it leads to more “dynamic” microcolonies as compared to experiments. Multiple pili contact points in the real setting lead to the formation of active pili network with higher gradient in motility and less dynamic core of the cells. It is an interesting direction of further research to understand the biophysical properties of such a network. Recently a different modeling approach averaging the contribution of all pili to an intermittent attractive force similarly led to the existence of a smooth diffusion gradient emphasizing the importance to be able to take into account multiple pili interactions in the future (Bonazzi et al. 2018).

Importantly, we see that the heterogeneous behavior in the microcolonies emerges during the assembly process of physically identical bacterial cells and this phenomenon is very robust for a wide range of model parameters (Pönisch et al. 2017).

Figure 6: Motility of single cells inside a microcolony. (a) Representation of the detection of fluorescently labeled cells. The left image highlights the detection the fluorescently labeled cells. The right image shows the position of individual cells relative to microcolony. Scale bar = 10 µm. (b) In order to be able to reduce the effect of rotation of the microcolonies on the trajectories of single cells, we computed the mean squared relative distance of cell pairs. Both cells were defined to be a pair if they could be found in a similar region, defined by their distance from the surface. (c,d) Diffusion coefficient D from the experimental data as a function of the distance Rs from the surface and MSRD as a function of time. (e,f) Diffusion coefficient D from the simulation data as a function of the distance Rs from the surface and MSRD as a function of time.

2-3.3: Non-retracting cells are excluded on the outside of microcolonies

If the forces due to retracting Tfp play a central role in shaping the microcolony, it would be predicted that cells deprived of the ability to generate those forces would alter the process of colony formation. To test this in the model, we performed simulations of a 1:1 mixture of pulling wild-type cell and cells that possess pili but are unable to retract them. We observed that the non-pulling cells were excluded to the outside of the microcolonies (Fig. 7c). To find the actual experimental confirmation of this prediction, we used the fact that retraction of

Tfp in Neisseria gonorrhoeae is under the control of a AAA ATPase pilT. Bacteria deprived of this molecular motor (Ng ΔpilT) still have Tfp and thus can interact through Tfp-Tfp interactions with other bacteria but they do not retract their pili and thus cannot exert forces on their surroundings (Higashi et al. 2007). A 1:1 mixture of Ng WT cells and Ng ΔpilT

(fluorescently labeled) led, in agreement with the prediction of the computational model, to the sorting of cells that cannot exert forces to the outside of the microcolonies during the self-assembly process (Fig. 7a,b). Similar sorting of two types of bacteria were observed previously in the case of bacteria with different types of pili and thus different interaction forces between them (Oldewurtel et al. 2015). Here, the pili are made of the same major pilin in both, pulling and non-pulling bacteria. Thus, the origin of the sorting in our experiments is the absence of retraction forces, not the difference in the interaction forces between different types of pili. A similar discrepancy exists in eukaryotic systems where the difference of adhesion forces between cells has been first postulated to be a driving force in early embryogenesis (Steinberg 2007). More recent theories take also into account the ability of eukaryotic cells to exert forces on their surrounding through contraction of their skeleton. In this case the dynamics of the cells seem crucial to the biological outcome

(Pawlizak et al. 2015; Schötz et al. 2013). In a control experiment, when the fluorescent markers were used to label two NgWT populations with different fluorophores in a 1:1 mixture, the corresponding forming microcolonies showed a homogeneous repartition of both types of bacteria (Fig. 7d,e). The homogeneous repartition is also obtained in the case of the simulations (Fig. 7f). These results exemplify the crucial role that Tfp contractile forces have in shaping Neisseria gonorrhoeae microcolonies and demonstrate the predictive power of the computational model.

Figure 7: Demixing of Neisseria gonorrhoeae Microcolonies. (a) DIC and fluorescence images allowed to detect the positions of the WT cells (right) and the ΔpilT mutants (center). (b) Intensity profile of fluorescently labeled WT (YFP) and ΔpilT mutant (mCherry) cells along a line in the midplane going through the center of the colony (a–c) Simulated assembly of a mixture of WT and ΔpilT mutant cells. The inset shows the midplane of a colony. The green cells are WT cells, the red cells are ΔpilT mutants. (d–f) Same data as a-c for a mixture of differently labeled WT cells.

2-3.4: Heterogeneous gene expression follows heterogeneous motility behavior in microcolonies We have demonstrated that Tfp retraction forces in a set of initially identical cells can lead to the formation of a heterogeneous motility behavior as they form a microcolony. To probe whether this motility behavior is associated with differential gene expression across a microcolony we have generated a gene reporter for the pilin gene: the promoter of the pilE gene was incorporated heterologous in the genome of Neisseria gonorrhoeae driving a mCherry fluorescent protein (Ng PpilE-mCherry) to get an idea of the expression of pilE

across a microcolony. The genomic site of incorporation has been shown to not modify the behavior of the bacteria and the fluorescence will be a proxy for the expression of the pilE gene. When these microcolonies were treated in similar conditions as the ones we used to study the dynamics of microcolonies, the fluorescence reaches the edge of the microcolony and the fluorescence is consistent with a homogeneous expression across the microcolony

(Fig. 8a). Our measurements of metabolic levels among cells within microcolonies during the first few hours of formation show also homogeneous profiles. As we have seen previously in the case of Tfp numbers, life and attachment times, the computer model enables us to access quantities that are difficult to measure experimentally. Importantly, we can also compute the mechanical forces within the microcolony. According to the model the dynamical formation of bacterial microcolonies via Tfp interactions will lead not only to a gradient of numbers of

Tfp and times of attachments, but also to the appearance of a force gradient across a microcolony (Fig. 8c). So, the heterogeneous motility within microcolonies is linked to a force gradient. After a few hours of interactions, the force gradient is present, but the gene expression and metabolism are homogeneous (Figs 8a). All small diffusible chemical cues are most likely also spatially homogeneous due to the small size of the microcolonies. For bacterial aggregates below 40 microns in diameter diffusion is a very efficient process and both simulations and experiments indicates that oxygen, nutrients and waste products will be homogeneously distributed (Stewart 2003; Wessel et al. 2014).

Figure 8: Heterogeneous genetic expression within a microcolony. (a) Brightfield (left), fluorescence (center) and overlayed image (right) of a Neisseria gonorrhoeae PpilE-mCherry microcolony after 3 hours of formation. (Scale bar = 10 µm.) (b) Brightfield (left), fluorescence (center) and overlayed image (right) of a Neisseria gonorrhoeae PpilE-mCherry microcolony after 7 hours of formation. (Scale bar = 10 µm.) Note that (a,b) represent two microcolonies imaged in the same conditions showing the spatial heterogeneity of expression in (b) and not in (a,c) Forces as a function of the distance of the center (COM) of an in-silico colony.Fij are the excluded volume force acting on a cell due to neighboring cells (blue) and the absolute values of pili forces acting on one cell (red) and where i and j are pili indices.

When microcolonies are left to develop for 7 hours in liquid the pattern of gene expression of pilE becomes heterogeneous as shown by the fluorescence across the microcolony (See Fig. 8b). The pattern of gene expression parallels the pattern of differential motility that we observed. But importantly the pattern of gene expression appears multiple hours after the observation of the motility pattern and the metabolic activity is still homogeneous at that later 7 hour time point. These results are indicative that an initial change of motility behavior precedes a differential gene expression within the microcolony.

The pursuit of the mechanisms relating these gradients to differential gene expression in a microcolony is beyond the scope of the present study. Nevertheless, keeping in mind the idea from eukaryotic mechanobiology that a gradient of mechanical forces can trigger differences

in gene expression, it is tantalizing that the gradient of forces within a bacterial microcolony triggered by the dynamics of Tfp can be the first step towards differentiation as it precedes heterogeneity in gene expression.

2-4: DISCUSSION

The development of multicellular eukaryotic organisms has been a long standing biological question with an overwhelming focus on the presence of gradients of different molecules until recently, when the role of physical forces in development was acknowledged (Howard,

Grill, and Bois 2011a). Bacterial biofilms have been recognized as being similar to multicellular entities and being able to develop differentiated states usually over multiple days (Stoodley et al. 2002; G. O’Toole, Kaplan, and Kolter 2000). The mechanisms at play in this development are the subject of intense scientific inquiry due to the health-related importance and ubiquity of biofilms. The role of mechanics in bacteria lifestyle, in particular within biofilm, is being recently appreciated, whether it is the role of motility, hydrodynamical flow or internal forces building within biofilms (Persat, Nadell, et al. 2015).

Besides, there is now mounting evidence that bacteria have the ability to sense mechanical forces and those forces can in turn change genetic expression (Belas 2014; Persat, Inclan, et al. 2015a).

In the case of Neisseria gonorrhoeae, we have shown that a group of identical cells powered by cycles of extensions and retractions of Tfp can self-organize in microcolonies with heterogeneous motility pattern. Such heterogeneous motility behavior correlates with a gradient of mechanical forces within a microcolony. If these mechanical forces can be sensed by bacteria, a feedback mechanism could provide the ability to these forces to take over the

control of the spatial development of the microcolony by triggering heterogeneous gene expressions. Neisseria gonorrhoeae microcolonies are akin to early developing embryos and eukaryotic cell spheroids, where the forces between cells, both adhesive and contractile, play a crucial role (Farge 2003; Schötz et al. 2013; Montel et al. 2011; Desprat et al. 2008; Maitre et al. 2012). What might unify these biologically so different systems is the common biophysical mechanism where the mechanical forces generated on the surface of aggregates drive the shape transformation which is resisted by the viscoelastic response of the bulk.

This mechanism can apply generally despite the difference in origin of these aggregates whether it is motile aggregation of different cells or successive divisions of a starting cell or a combination of those two modalities. The simple system of Neisseria gonorrhoeae bacteria, accompanied by the in silico model, not only enables us to understand better the physiology of an important human disease but it could also give a new insight into the earliest steps of genetic differentiation within a group of identical bacterial cells and ultimately the evolution of multicellularity (Lyons and Kolter 2015).

2-5: METHODS

2-5.1: Bacteria strains and growth conditions

The wild-type (WT) strain used in this study is MS11 (Schoolnik et al. 1984; Meyer et al.

1984). The ΔpilT mutant was obtained by an in-frame allelic replacement of the pilT gene by a Kanamycin resistance cassette. Fluorescent proteins (YFP, mCherry or tdTomato) driven by a consensus promoter were incorporated by allelic replacement together with an antibiotic marker (either Kanamycin or Chloramphenicol). Similarly, mCherry driven by the reporter of the pilin gene (370 bp before the beginning of the starting ATG of the pilin ORF) was incorporated by allelic replacement together with a Chloramphenicol marker. Bacteria were grown on GCB-medium base agar plates supplemented with Kellogg’s supplements at

37 °C and 5% CO2. 80 μg/ml of Kanamycin or 7 μg/ml of Chloramphenicol were added when growing mutants with the corresponding antibiotic resistance cassette. Cells were streaked from frozen stock allowed to grow for 24 hours and then lawn onto identical agar plates and used after a 16 to 20 hour growth period.

2-5.2: Microcolony formation

Bacteria from lawns on agar plates were resuspended in 1 ml of GCB medium at an optical density of O.D. = 0.7. 100 μl of the suspension was added in the well of the 6 well plate containing 2 ml of GCB medium with a BSA coated cover glass (round 25 mm diameter coverglasses (CS-25R) Warner Instruments) at the bottom or without. The 6 well plate was centrifuged at 1600 × g in a swinging bucket in an 5810 R centrifuge (Eppendorf) for

5 minutes resulting in single bacteria uniformly coating the bottom of the well. For direct imaging the coverglass were transferred to an observation chamber (Attofluor cell chamber,

Thermo Fisher Scientific). In the case of mixture the suspension at O.D. = 0.7 of both

components of the mixture were prepared and a new 1 ml was prepared by the proper ratio of the two suspension. 100 μl of that new suspension was used similarly to what was described previously.

2-5.3: Microscopy and microcolony merger

All movies were obtained on a Nikon Ti Eclipse inverted microscope equipped for epifluorescence and DIC microscopy and with an optical tweezer setup all under an environmental chamber maintaining temperature, humidity and CO2 concentration. The objective used is a 60X plan Apo objective. The camera used were either a sCMOS camera

(Neo, Andor) or a CMOS USB camera (DCC1240M, Thorlabs). 1 Hz fluorescent movies and

0.1 Hz DIC movies of either microcolony merger or follow up of single cell motility were taken for further analysis. In the case of microcolony merger, microcolonies were performed and were brought into contact either by optical tweezers or hydrodynamical flow.

2-5.4: Scanning Electron Microscopy

Bacteria microcolonies on a glass coverglass were fixed with 3.7% formaldehyde in PBS pH

7.4 for one hour. The microcolonies were subsequently washed 3 times with PBS and then dehydrated by step in ethanol (50%, 70%, 80%, 90% and 100% ethanol). The samples were then critical point dried and imaged on a Hitachi S-4700 Scanning Electron microscope.

2-5.5: Image analysis of experimental data

General information

We analyzed in total 28 merger events and 40 individual colonies for the tracking of individual cells (with at least 42 trajectories used for the computation of the spatially dependent MSRD) inside of microcolonies. Matlab R2015b was used for edge detection of

DIC movies and tracking of individual cells inside of colonies.

Edge detection

In order to detect the edges of single colonies and the merger from DIC data the same algorithm was used. We computed the first derivatives of intensities in x- and y- direction of a Gaussian filtered image and thresholded its absolute value. Afterwards we dilated and eroded the binary image, filled all remaining holes and removed small objects. For all steps internal functions of Matlab were used.

Single Cell Tracking inside of Colonies

To track single cells from the fluorescence images, we first computed the center of mass

(COM) of the binary shape resulting from the edges and corrected the fluorescence data in order to reduce the effects of the translations of the colony on the tracking algorithm. We interpolated the position of the COM from the DIC data recorded with frequency of 0.1 Hz to match that of the fluorescence data with 1 Hz resolution. A cubic spline data interpolation was applied on the x- and y-component on the COM. We next used the detection and tracking algorithm developed by Blair et al (Blair, D. & Dufresne 2017). For original images we computed the background by applying a large-scale Gaussian filter and subtracted its values from the images. Additionally, we used a smaller Gaussian filter for smoothing.

Estimating the MSRD and the diffusion coefficient of cells inside of colonies

The measurements of displacement of individual cells is strongly affected by translation and rotation of the microcolony as a whole. This effect can be circumvented by measuring the absolute value of the distance vector of two individual cells and computing the time- averaged second moment of the relative distance increments, as a function of the time lag.

One can show that this quantity is equal to the sum of the mean squared displacements of the two individual cells in the colony [Note: The ensemble averaged mean squared relative displacement coincides with the time averaged quantity as long as the individual cells displacements are smaller than the initial distance between the cells, which is the case for most of pairs of cells in our measurements.]. It thus grows linearly with the lag time and the prefactor of this linear growth quantifies the diffusion of the cells. To identify the location of cells in the microcolony, we computed the distance from the surface by finding the edge shape from DIC images taken in parallel to the fluorescent images and picking the cell pairs that move in a region. By assuming that cells belonging to the same region have the same diffusion coefficient we can read out of individual cells from the fit. The constant offset b accounts for the measurement noise. This method allows us to find the diffusivity of cells as a function of the distance from the surface.

Ellipse fitting

The binary images of the merger movies allowed us to fit an ellipse by computing the central moments (Burger, W. & Burge 2009). This algorithm also allowed us to compute the orientation of a non-spherical colony, the length of the short and long axis of the ellipse and their axis ratio (Fig. 5b).

Bridge height measurement

To compute the height of the bridge forming between the two regions of the binary images we rotated the image so that the colonies were oriented along the x-axis and calculated the

COM of the combined regions. Then we moved a line of length L centered around the COM and perpendicular to the axis connecting the two colonies. The bridge was defined as the range for which the whole line could be found inside of the colony region. L was chosen to be small enough to be not affected by the elliptical shape of the colonies. For the late coalescence the results were compared to the short axis of the ellipse.

2-5.6: Simulations

General information

We implemented the simulation of the model (see Supplementary Information, not provided in this thesis) in C++ and used the package OpenMP for parallelization on up to eight CPUs.

The simulations were performed on the local computing cluster of the Max Planck Institute for the Physics of Complex Systems consisting of x86-64 GNU/Linux systems.

Single colony modeling

For the simulation of the dynamics of individual cells inside of a colony we randomly initialized 1700 cells inside of a sphere without cell-substrate interactions. First, we allowed the cells to repel each other until there was no overlap between neighboring cells. Next, we introduced pili and their dynamics, causing the motion of cells. In order to reproduce the experimental results, we only monitored the cells being positioned <1 µm above and below the midplane and computed the second moment of the relative distance increments (as for the experimental data) of cell pairs, projected to the midplane.

Merger modeling

For the merging simulations we again neglected cell-substrate interactions and initialized two separate colonies each consisting of 1000 randomly distributed cells. The spherical shape of microcolonies indicates cell-substrate interactions are negligible as, if not, they would deform the microcolony towards the substrate (Pönisch et al. 2017). For the first 100 s the pili of the colonies were not allowed to interact with pili of the other colony. This way we allowed for the formation of stable individual microcolonies with a radius of ~7 µm and initial separation of 16 µm between their centers. To compute the bridge height and the eccentricity we created the binary image from the projection of the cell positions and their shapes onto a plane tangential to the axis between the two colonies.

Assembly modeling

For the assembly experiments we distributed 1500 cells on a substrate of the size

98.56 × 98.56 µm2 with periodic boundary conditions. The pili of a fraction of cells were not able to retract, modeling the ΔpilT mutant.

CHAPTER 3

Type IV retraction forces shape spatiotemporal physiology in Neisseria gonorrhoeae

Kelly B. Eckenrode, Wolfram Pönisch, Hadi Nasrollahi, Mahmound Nametalla, Brian Ford, Vasily Zaburdaev & Nicolas Biais

Upcoming manuscript in preparation

Author contributions: K.B.E prepared writing, reporter experiments, and figures. W.P, V.Z, N.B developed image analysis technique. W.P performed reporter analysis. H.N and B.F performed survival experiments, M.N performed some ATPase reporter experiments

3-1: ABSTRACT

Multicellular structures integrate both chemical and mechanical signals in order to coordinate development. Understanding the role of mechanical signaling poses experimental challenges, so we exploited the mechanical nature of Neisseria gonorrhoeae’s extracellular appendage type IV pilus (tfp) to study the role of force during the multicellular microcolony formation. Here, we demonstrate through fluorescence transcription reporters that Neisseria gonorrhoeae generate early gene expression spatiotemporal patterns of pilus- related genes. In the absence of tfp depolymerization induced retraction forces these patterns are disrupted. Not only are patterns disrupted, but survival is negatively impacted by the loss of tfp retraction force. In addition, the fact that application of external artificial forces disrupts gene expression patterns in WT and in a tfp retract-deficient mutant, pilT.

These results suggest a global sensitivity to mechanical forces. Even further, by accelerating tfp engagement with nearby cells, decreases antibiotic susceptibility after exposure, in comparison to cell that engage tfp dynamics at a later time point. Collectively, these results suggest that tfp retraction forces influence the physiology of developing microbial multicellular structures through a mechanosensitive mechanism.

3-1: INTRODUCTION

The physiology of bacteria has traditionally been studied as an entire population during growth phases, and recent technological advances in high-throughput screening provides insight into single cells physiology. However, these approaches miss how the majority of bacteria actually live, as adhered multicellular structures (Costerton et al. 1987;

G. O’Toole, Kaplan, and Kolter 2000). Researchers accept that most bacteria found in nature undergo communal development from single free-living cells to interdependent- communities called biofilms (Stoodley et al. 2002). It is important to understand cells in their native context, as individual cellular aggregates, in order to comprehend the physiology of multicellular development.

Microbial multicellular structures are known to help protect the cells against the host defenses and also confer resistance to antibiotics, which provides fitness advantage, like defense against antimicrobials and resource sharing (Donlan and Costerton 2002; Oliveira et al. 2015). Some bacterial species use dynamic extracellular adhesion filaments help single cells develop into multicellular structures, like microcolonies or biofilms. The causative agent of gonorrhea, Neisseria gonorrhoeae produce biofilm precursors called microcolonies

(Higashi et al. 2007). The only requirement for proper microcolony formation is through the physical interaction of their ubiquitously expressed type IV pilus (tfp) (Figure 9) (Taktikos et al. 2015b; Craig, Pique, and Tainer 2004; Craig, Forest, and Maier 2019). Without tfp, aggregation does not occur and cells remain planktonic.

Microcolonies assemble into spherical ~20µm in diameter aggregates within 3 hours from single cells interacting with a surface, abiotic or biotic (Higashi et al. 2007). Tfp dynamics are reminiscent of a repetitive “cast and pull” mechanism. Tfp coordinate mechanical activity through a multimeric pilus molecular machinery composed of approximately 15 proteins (Goosens et al. 2017). The main tfp protein filament is composed of a main subunit, PilE. PilE monomers polymerize and extend into extracellular space by

AAA-ATPase, PilF. Once the pilus is extended, it will eventually retract by the depolymerization fueled by another AAA-ATPase, PilT, from within the periplasmic membrane. In the case of Ng, retraction from the PilT motor can amount to forces up to 1 nN,

100,000 times stronger that the cells own weight, which makes Neisseria gonorrhoeae the strongest microbe measured to date (Merz, So, and Sheetz 2000; Maier et al. 2002; Biais et al. 2008). While Tfp is essential for microcolony formation, it has other versatile functions, including DNA horizontal transfer and twitching motility among others, which are all powered by PilT retraction force (Beatson et al. 2002; Berry and Pelicic 2015; Craig, Forest, and Maier 2019).

Since force produced from tfp retraction plays a critical role in microcolony formation

(Howie, Glogauer, and So 2005; Lee et al. 2005 Higashi et al. 2007), our main hypothesis is that tfp retraction forces play a role in shaping Neisseria gonorrhoeae microcolony physiology. To understand the connection between retraction force and physiology, we utilize an established genetic retraction-deficient mutant, ΔpilT, in which pili are present and unchanged chemically, do not retractile and thus exert no force (Dietrich et al. 2009). Both retraction-prone and retraction-deficient bacteria can form microcolonies; however, since the ΔpilT is no longer motile, it is incapable of progressing from the initial cellular adherence

stage and does not progress to tightly formed microcolony, even though there is an overall increase of cell aggregation (Brown et al. 2010).

The role of the physical forces in multicellular development, in both eukaryotes and prokaryotes, is not well understood. In our previous work, we have demonstrated that cells in microcolonies have differential motility depending on their location in the aggregate

(Pönisch et al. 2018b). We showed that Neisseria gonorrhoeae microcolonies have an outer perimeter layer of more motile cells and less motile cells in the core, which demonstrates heterogeneous motility. Therefore, we explored the relationship between tfp activity and microcolony physiology (Oldewurtel et al. 2015).

To understand more closely the role of tfp physical forces in bacterial multicellular development, we examined how tfp retraction forces effect microcolony physiology. We

Figure 9A: Neisseria gonorrhoeae early microcolony development. A) Schematics demonstrating the single cell-cell interactions induced by tfp dynamics over 3 hours. B). Live fluorescent Neisseria gonorrhoeae cells demonstrating microcolony formation. Images taken from time lapse movie.

measured transcriptional activity within microcolonies by novel quantification methods. We find that transcription of tfp genes is organized and coordinated during early biofilm development. In addition, we tracked growth, survival, and antibiotic susceptibility as physiological responses in the presence and absence of tfp retraction forces.

3-2: RESULTS

3-2.1: Type IV pilus genes are expressed spatiotemporally in microcolonies

To measure the spatiotemporal transcriptional activity of developing MS11 Neisseria gonorrhoeae microcolonies, we designed a set of genetic constructs where pilus-related promoters drive an mCherry reporter cassette in live cells. At 3-hour and 7-hour microcolony developmental time points, mCherry fluorescent emission intensity of individual microcolonies was captured by epifluorescence imaging and was used as a proxy for transcriptional activity. The rationale behind observing these two time points is that 3- hour microcolonies achieve average size (~20µm in diameter) not vastly different from later time points, and at 7-hours we avoid complications of growth and DNA secretion seen at later time points. We focus on microcolonies with symmetrical spheres, which was important for our measurement consistency. As microcolonies can merge together and form asymmetrical shapes, the majority are round. To get a consistent image plane, it is sufficient to image the mid-section of each microcolony.

To locate pilus-related promoters, we used a bacterial genome analysis program,

SoftBerry Bacterial Promoter Prediction Program (BPROM), which predicts regulatory elements associated with promoter regions (Solovyev 2011). Since the regulatory regions are upstream of genes, we measured the likelihood of promoter elements were present at

various sequence lengths. BPROM uses genome databases to predict transcriptional start sites, transcriptional binding sites, -10 box and -35 box. Sequences used to drive mCherry expression in these experiments range from 163bp to 934bp long (Methods for more details).

A genetic cassette composed of pilus-gene promoter fused to mCherry gene, which was chromosomally inserted to the same loci (Figure 9B). As a non-pilus related control construct, we used an E. coli LacUV5 consensus sequence promoter driving the mCherry expression.

Figure 9B: Genomic insertion region for transcriptional reporters. GenBank: CP003909.1. Numbers on ends of schematic represent base pair number. Asterix indicates insertion site at 1489951bp.

Figure 10: Transcription assay demonstrates spatiotemporal gene expression patterns in Neisseria gonorrhoeae microcolonies. LacUV5, WT, pilE, pilT-mCherry expression are shown in the TexasRed channel, and the merged DIC image in composite. Microcolony diameters range 10-20um. Each microcolony was imaged in identical conditions.

This experimental design demonstrated a range of spatiotemporal intensities and patterns in developing microcolonies (Figure 10). Pilus-related promoters demonstrate variable activity across the microcolony spatial plane and time. We plotted the intensity profile of fluorescence across the microcolony to visualize the shape of spatiotemporal patterns

(Figure 10, 11). Here we consider the different length scales of the variation of microcolony sizes, which provides detailed information on gradient shape, then fit to an exponential

푑 퐼(푑) = 퐼0 − 퐼1 exp (− ). In addition, we created a quantification method to provide a simple 푑푐 and intuitive value to measure the degree of inhomogeneity (Figure 12 A-H) (see Methods for more information).

The pilE promoter drove mCherry expression to the highest intensity, consistent with pilE (tfp major pilin subunit) being one of the most expressed proteins in Neisseria gonorrhoeae. pilE-mCherry shows a homogeneous expression across a microcolony at 3- hour. At 7-hour, the pattern shifted to an inhomogeneous expression by decreasing fluorescence along the perimeter and increasing proximal expression to the center. Some constructs showed less promoter activity as demonstrated by low fluorescence intensities levels (>500 A.U). These ‘low-level expressers’, like pilA-mCherry, pilB-mCherry, and pilT- mCherry expressed homogeneously at both 3-hour and 7-hour (Figure 12B, 12C, and 12G, top left two charts).

The pilC1 and pilC2 are duplicate genes regulated by different promoters at separate loci. pilC1 and pilC2 share some nucleotide sequence homology, but are not identical.

Interestingly, pilC genes are phase variable due to a frameshift mutation in a G-residue stretch at the beginning of the open reading frame (Jonsson, Rahman, and Normark 1995). pilC2 is also considered to be the only copy that is transcriptionally active in our strain. From

our promoter reporter assay, we measured promoter activity of both pilC1-mCherry and pilC2-mCherry, which also expressed different patterns. pilC1-mCherry is expressed homogeneously while becoming heterogeneous at 7 hours. pilC2-mCherry does not express nearly at the same intensity level as pilC1-mCherry, and remains homogeneous in both time points (Figure 11).

Figure 11: Additional promoter assays demonstrates spatiotemporal gene expression patterns in Neisseria gonorrhoeae microcolonies. pilA, pilB, pilC1, and pilC2-mCherry expression are shown in the TexasRed channel, and the merged DIC image in composite. Microcolony diameters range 10- 20um. Each microcolony was imaged at identical conditions

The expression of both pilT and pilU have been considered to be transcribed from one

promoter upstream of pilT, which transcribes both pilT and pilU (Park et al., 2002). As an

alternative, we also found a pilU promoter located immediately after the pilT stop codon, and

upstream of pilU. The 163bp sequence contains a -10 box at 126bp (CCGCAAACT), -35 box at

107bp (TTCCTG), and sigma70 binding site; however, no other known transcription binding sites were predicted using BPROM. UniPro UGENE, an open-source bioinformatics software,

SITECON transcription binding site prediction software, predicts 83% probability of a DnaA binding sites at 89-116bp and 87% probability of CpxR binding sites at 136-151bp. pilU- mCherry microcolonies expression is higher than pilT-mCherry, yet it is also homogeneously expressed like pilT-mCherry.

Here, we provide a novel analysis method to measure fluorescence inhomogeneity measured from promoter reporter microcolonies for consensus LacUV5-mCherry, pilA- mCherry, pilB-mCherry, pilC1-mCherry, pilC2-mCherry, pilE-mCherry, pilT-mCherry, and pilU-mCherry (Figure 12A-H. For more details on inhomogeneity analysis see Methods and

Figure 12I). This analysis demonstrates that pilE (Figure 12E, bottom row right most chart) is expressed in an heterogeneous fashion at 7-hour with an inhomogeneous index, H = 0.45. pilT is expressed homogeneously at both 3-hour and 7-hours with an H = 0.23 (see methods for analysis details and reference image of an inhomogeneous index) (Figure 12G, bottom row right most chart). Interestingly, the consensus LacUV5-mCherry reporter demonstrates a slight heterogeneous expression pattern at well at 7-hour of H = 0.33 from H = 0.26 at 3- hour (Figure 12A, bottom row right most chart). In summary, the promoter-reporter analysis shows that WT Neisseria gonorrhoeae microcolonies express genes in spatiotemporal patterns.

Figure 12A: LacUV5-mCherry consensus reporter. Spatiotemporal gradient analysis method (top row). Profile 3-hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H : 3-hr = 0.26, 7-hr =0.31. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12B: pilA-mCherry reporter. Spatiotemporal gradient analysis method (top row). Profile 3- hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H : 3-hr = 0.21, 7-hr =0.31. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12C: pilB-mCherry reporter. Spatiotemporal gradient analysis method (top row). Profile 3- hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H : 3-hr = 0.26, 7-hr =0.31. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12D: pilC1-mCherry reporter. Spatiotemporal gradient analysis method (top row). Profile 3- hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.21, 7-hr =0.32. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12E: pilC2-mCherry reporter. Spatiotemporal gradient analysis method (top row). Profile 3- hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.31, 7-hr =0.32. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12F: pilE-mCherry reporter. Spatiotemporal gradient analysis method (top row). Profile 3- hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.20, 7-hr =0.45. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12G: pilT-mCherry reporter. Spatiotemporal gradient analysis method (top row) Profile 3- hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index analysis method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.23, 7-hr =0.23. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 12H: pilU-mCherry reporter. Spatiotemporal gradient analysis (top row). Profile 3-hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. The inhomogeneity index (H) method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.255, 7-hr =0.258. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

3-2.2: Retraction-deficient microcolonies gene expression

PilT has previously been demonstrated to initiate the expression of several groups of pilT-responsive genes that are predicted to aid in initiating the infection pathway (Dietrich et al. 2009, 2011). To determine whether retraction force plays a role establishing spatiotemporal gene expression patterns, we used the same set of transcriptional mCherry reporters, but deleted the pilT gene and replaced it with Kanamycin resistance gene for selection. This set of pilT::Kan reporter genetic constructs allowed us to measure if microcolonies regulate pilE and pilT expression differently than in WT background.

We measured an overall decrease in fluorescence intensity in the pilT::Kan- mCherry reporters. A dramatic decrease of pilT expression in the ΔpilT background and a more modest effect in ΔpilT::Kan pilE-mCherry (Figure 13A; Figure 14A-B). The inhomogeneous index method measured fluorescence patterns in ΔpilT::Kan reporters

(Figure 14A-B), and ΔpilT::Kan pilE-mCherry with H: 3-hr = 0.21,is similar to the WT pilE- mCherry (H: 3-hr = 0.2). At 7 hours, ΔpilT::Kan pilE-mCherry reaches H: 7-hr =0.37, which is lower than the WT pilE-mCherry (H: 7-hr 0.45). In Figure 14A, the gradient analysis (top row, left two charts) measure a repressed spatiotemporal pattern of the ΔpilT::Kan pilE-mCherry, where total fluorescence is low, but the H-index remains similar to WT. We believe inhomogeneity persists due to a few bright fluorescent cells in the center, but the intensity does not broaden beyond a few cells, like in the WT pilE-mCherry reporter. In addition, another striking result was the initial variability of intensity present in ΔpilT::Kan pilE- mCherry (Figure 14A, top row left two charts). In addition, pilE intensity signal decreased at the 7-hour time point, instead of increasing like the WT pilE-mCherry.

The ΔpilT::Kan pilT-mCherry reporter intensity continues to increase from 3-hour to

7-hour similarly as the WT pilT-mCherry; however, the total fluorescence of ΔpilT::Kan pilT- mCherry does not reach the same intensity as the WT pilT-mCherry at 7-hours (Figure 14B, top row, left two charts). Inhomogeneity remains low (H: 3-hr = 0.25, 7-hr =0.25), similar to

WT pilT-mCherry at both time points (Figure 14B, bottom row, right two charts). Our analysis suggests that microcolonies lacking pilT cannot fully produce a WT spatiotemporal pattern.

It had been previously reported the ΔpilT::Kan mutant is responsible for upregulation of pilE to produce hyperpiliated cells. Yet, in our ΔpilT::Kan pilE-mCherry reporters, we see

a local decrease of mCherry fluorescence, suggesting a downregulation of pilE in microcolonies (Figure 13A, bottom row). We believe this disparity reflects differences in the growth conditions on solid agar in comparison our method of incubating cells in liquid culture (Dietrich et al. 2009).

Figure 13A: ΔpilT::Kan promoter assay demonstrates spatiotemporal gene expression patterns in Neisseria gonorrhoeae microcolonies. ΔpilT::Kan pilE and ΔpilT::Kan pilT-mCherry expression are shown in the TexasRed channel, and the merged DIC image in composite. Microcolony diameters range 10-15um. Each microcolony was imaged at identical conditions.

Figure 13B: Minor ATPase, pilT2 and pilU, mutant promoter reporters at 3-hour and 7-hour time points. Left, 3-hour; Right 7-hour microcolony.

The PilT ATPase is not the only pilus motor in Neisseria gonorrhoeae. Minor ATPases,

PilT2 and PilU, are thought to be responsible for the fine-tuning of pilus retraction: pilU deletion has been linked to an infection efficiency only in mice, and pilT2 deletion alters the speed of retraction (Kurre et al. 2012; Eriksson, Eriksson, and Jonsson 2012). To control for the impact of other pilus retraction motors, we replaced gene with Kanamycin antibiotic resistance gene in the pilE-mCherry and pilT-mCherry reporter background. We observe an extreme morphology change in ΔpilU::Kan, while ΔpilT2::Kan maintains a WT morphology

(Figure 13B). We see also there is in minimal decrease of fluorescence in both constructs. In both the ΔpilU::Kan and ΔpilT2::Kan pilE-mCherry, the inhomogeneity is increased. However,

the changes in spatiotemporal patterns are not as dramatic as seen in the ΔpilT reporters

(Figure 13B). This data is consistent with the major role that PilT has been demonstrated to play in tfp retraction forces in Neisseria gonorrhoeae.

Figure 14A: ΔpilT pilE-mCherry reporter spatiotemporal gradient analysis method (top row). Profile 3-hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.21, 7-hr =0.37. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 14B: ΔpilT pilT- mCherry reporter spatiotemporal gradient analysis method (top row). Profile 3-hour and profile 7 hour show fluorescence intensity of entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.25, 7-hr =0.25. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

In summary, ΔpilT reporters do not recapitulate the same spatiotemporal patterns as

WT. Consistently between the two ΔpilT reporters, total fluorescence does not achieve the similar intensities as WT reporters, and spatiotemporal patterns are disrupted. In addition, by removing minor ATPase motors, we also see reduced total fluorescence. Collectively, these data suggest that tfp retraction force contributes to microcolony gene expression patterns. With the loss of pilT retraction force, spatiotemporal patterns are perturbed; therefore, pilT may play a signal role in microcolony gene expression and suggests a mechanosensitive role of tfp in Neisseria gonorrhoeae physiology.

3-2.3: Validation of mCherry fluorescence quantification

When quantifying fluorescence in live cells, it is important to consider alternate possibilities for changes in intensities. For instance, the differential fluorescence signal we measured in our transcriptional promoters could be from sources other than differential gene regulation. Cell growth within the microcolony could create different intensity patterns; however, if this were true, then we would see the same fluorescence pattern in all the microcolonies. Several transcriptional promoters were constructed and analyzed to control for a cell growth pattern.

Another possibility for measuring inhomogeneity in microcolonies is the dilution of mCherry protein in the cytoplasm. mCherry has a maturation time of 15 minutes and is the fastest maturing red fluorescent protein in bacteria; therefore, it is recommended for gene expression measurements (Beilharz et al. 2015). Along with the fast maturation time coupled with speedy transcription time (~10nt/sec), we can distinguish between mCherry protein accumulation and protein dilution during cell division (Proshkin et al. 2010). I designed an experiment to distinguish between dilution or accumulation during division. I measured fluorescence intensity during a live time lapse of four cell conditions, single cells

(fixed, non-motile, dividing) and a whole colony, of both the consensus lacUV5-mCherry and pilE-mCherry reporter. I measured individual fluorescence intensity using the ROI function in FIJI. The whole colony was measured with a different sized ROI as single cells. Background noise was removed from intensity measurements.

This experiment demonstrates a few findings (Figure 15). First, a time lapse every 10 minutes allows us to prevent photobleaching in pilE-mCherry because the intensity of the fixed cell is stable over time. LacUV5-mCherry did decrease by half of the total initial

intensity. This suggests that photobleaching may be variable between constructs. There is a population of single cells that are either dead or in a non-active state (blue line). The lacUV5- mCherry non-motile cells do not seems to be as sensitive to photobleaching as the fixed cells.

Secondly, tracking a single cell in pilE-mCherry showed variability over time, yet clearly have a more active promoter than lacUV5-mCherry because the intensity decreased over time.

From this experiment alone, it is difficult to parse between promoter activity in a single cell and photobleaching in different constructs; however, I believe this experiment shows the pilE-mCherry cells do not dilute completely during division.

Figure 15: Fluorescence intensity of LacUV5-mCherry (top) and pilE-mCherry (bottom) single cells (black), whole microcolony (red), non-motile single (blue), and fixed (yellow). Plotted over time by fluorescence intensity (A.U).

3-2.4: Fluorescent in situ hybridization probes show a different pattern in comparison to genetic reporter pattern

As an alternate to the genetic reporters, I also measured mRNA transcripts of pilE and pilT microcolonies using FISH, fluorescent in situ hybridization. Genetic reporters are a great proxy for capturing transcriptional and translational activity, and FISH allowed us to use mRNA concentration as a proxy for gene expression. This method also allows for direct quantitative measurements at the transcript. Each single strand DNA (ssDNA) probe is decorated with one fluorophore at the 5’ end. ssDNA probes are used to directly detect RNA transcripts. The genetic promoter reporter method detailed above may be biased due to locus-specific positional effects of the transcription-fluorescence reporter cassette. In addition, the total amount of fluorescence produced in live cells by gene reporters is the accumulation of protein over the course of the cell’s life cycle, and thus also depends on protein degradation. However, FISH requires the fixation of cells, which limits retrieving live time lapse data. Each technique is a proxy for transcriptional activity, so trying two methods may provide a broader understanding of gene expression’s stochastic behavior (Paulsson

2005).

I measured pilE and pilT transcription activity. pilE is especially interesting because it is one of the 72 antigenic and phase variable genes in Neisseria gonorrhoeae (Stuart A Hill and Davies 2009). The field’s current understanding of when cells change the variable genes is unclear. Being able to target spatiotemporal expression during 3 and 7-hour microcolony development will give insight into transcriptional activity. Since pilE is a genetically variable gene, I was careful to test probes for both the non-variable region (NVR) and the

‘hypervariable region’ (HVR). The HVR is located between 121- 152 amino acids in GC MS11

(Craig, Pique, and Tainer 2004).

DNA and RNA sequences can fold into secondary, and less commonly tertiary, structures. When designing probes, it is necessary to determine whether the probes can produce secondary structure probabilities. We measured the gradient intensity profile of both pilE373-A488 (Alexa488 dye label) and pilT347-A555 (Alexa555 dye label) (Figure

16A-5B). In Figure 16A-b, pilE inhomogeneity index at 7-hour, H = 0.28, is lower than the pilE-mCherry genetic reporter, H = 0.45. This discrepancy may highlight the different activity between mCherry protein dynamics and reading directly from mRNA. Interestingly, pilT inhomogeneity index at 7-hour, H = 0.35, in figure 16B is slightly higher than the pilT- mCherry genetic reporter, H = 0.23. While an increased inhomogeneity in pilT, and a decrease in pilE is present, these genes are expressed in a spatiotemporal fashion in WT microcolonies.

Composite DIC GFP488 channel

Figure 16A: FISH of pilE 373 at 7-hour. A) Epifluorescence microscopy images (100X) in two channels: GFP-488 and DIC. B) MATLAB gradient scale analysis (top) of a pilE 373 hybridization. Profile 3-hour and profile 7 hour show fluorescence intensity of an entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.22, 7-hr =0.28. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Composite DIC TxRed555 channel

Figure 16B: FISH of pilT 347 at 7-hour. A) Epifluorescence microscopy images (100X) in two channels: TexasRed-555 and DIC. B) MATLAB gradient scale analysis of a pilT 347 hybridization (top). Profile 3-hour and profile 7 hour show fluorescence intensity of an entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.31, 7-hr =0.35. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed

When FISH was repeated in ΔpilT, and the inhomogeneity index analysis determines the index for both pilE and pilT, the index unchanged from WT (Figure 16C-D). ΔpilT pilE has an

H = 0.27, in comparison to H = 0.28 in WT, at 7-hours (Figure 16C). However, ΔpilT pilT =

2.9, in comparison to H = 0.35 in WT, at 7-hours (Figure 16D). I consider these results inconclusive. Mainly due to the low statistics for ΔpilT analysis because experimental sample preparation for ΔpilT was difficult due to microcolonies fragile morphology.

Figure 16C: FISH of ΔpilT pilTE343 at 7-hour. A) Epifluorescence microscopy images (100X) in two channels: TexasRed-555 and DIC. B) MATLAB gradient scale analysis of a pilE 373 hybridization (top). Profile 3-hour and profile 7 hour show fluorescence intensity of an entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.21, 7-hr =0.27. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

Figure 16D: FISH of ΔpilT pilT 347 at 7-hour. A) Epifluorescence microscopy images (100X) in two channels: TexasRed-555 and DIC. B) MATLAB gradient scale analysis of a pilE 373 hybridization (top). Profile 3-hour and profile 7 hour show fluorescence intensity of an entire microcolonies, plotted individually over distance from the perimeter into the bulk center. Length scale and relative difference show the average diameter size over time. Inhomogeneity index method (bottom row). Inhomogeneity index vs. radius compared index by microcolony radius to control for morphology size. Inhomogeneity index vs. Axis Ratio compares index by morphology shape. Blue bar chart demonstrates range of index, with a red horizontal line at the average, H: 3-hr = 0.30, 7-hr =0.29. Lastly, the right most chart tests whether the distribution of the inhomogeneity index at each time is normally distributed with help of the Lilliefors test. If p<0.05 it is not normally distributed.

3-2.5: Artificial force applied to developing microcolonies alters differential spatiotemporal expression

During microcolony formation, we found that tfp genes are expressed in spatiotemporal patterns observed with transcriptional fluorescence reporters. The signaling pathways that regulate expression of pilus genes are not well understood. In previous work we measured single cell motility in microcolonies, and found cells on the perimeter of the microcolony have increased motility in comparison to cells inside of the bulk of the microcolony. This result led us to theorize that cells on the perimeter experience less tfp

retraction force then cells in the bulk due to the less pilus-pilus binding interactions producing a motility gradient (Ponisch, 2018). As an extension of this work, we hypothesize that spatiotemporal patterning of the pilus genes could be regulated via a mechanosensitive pathway; therefore, patterning may reflect the motility gradient. Therefore, I assessed the effect of artificial external forces on transcriptional activity of promoter reporters in both the WT and ΔpilT background. External forces by vibration (~60Hz) were applied to live developing microcolonies for 7 hours incubation. Then, microcolonies were imaged using fluorescence microscopy. The integrated fluorescence intensities of vibrated and non- vibrated microcolonies were compared (Figure 17).

pilE-mCherry and lacUV5-mCherry reporters both demonstrated significant sensitivity (p-value <0.0001) to vibration in comparison to the non-vibrated control. pilE- mCherry in the WT background increases fluorescence intensity 3.6 times, while doubles in the ΔpilT background. pilE-mCherry vibrated (n=58) with a mean intensity of 1,366,187, and pilE-mCherry non-vibrated (n=63) with a mean intensity of 375,524. This suggests the pilE transcriptional activity could be sensitive to mechanical stimulation, and in the absence of pilT, the transcriptional response is damped. LacUV5-mCherry fluorescence increase 4.1 times in the WT background, vibrated (n=85) with a mean intensity of 296,505 and non- vibrated (n=65) with a mean intensity of 72,569. While no significant difference between vibrated and not vibrated in the ΔpilT background was measured. Lastly, pilT-mCherry in the

WT and ΔpilT background also was not significantly different (Figure 17).

Although there is a global increase of fluorescence in pilE-mCherry vibrated conditions, we predicted that if a promoter is sensitive to mechanical stimulation, then perhaps the spatiotemporal pattern would be disrupted. In the WT pilE-mCherry strain, I

observe a loss of the differential fluorescence pattern in vibrated condition (Figure 18). pilE- mCherry and lacUV5-mCherry demonstrates significant sensitivity to vibration in WT and

ΔpilT background.

The size of the WT and ΔpilT microcolonies, vibrated and not vibrated, was measured by using the line tool in FIJI (Figure 19). There was a significant difference in size between vibrated and not vibrated mCherry constructs (WT-mCherry: p-value 0.0003, ΔpilT- mCherry: p-value 0.0329). There was no significant difference in size between vibrated and not vibrated on pilE and pilT reporters in both WT and ΔpilT (WT pilE-mCherry: p-value

0.993, ΔpilT pilE-mCherry: p-value >0.999, WT pilT-mCherry: p-value 0.202, ΔpilT pilT- mCherry: p-value 0.874).

Figure 17: Relative fluorescence integrated intensity of vibrated and not vibrated microcolonies. Y- axis is fluorescence intensity (A.U). X-axis is strains and vibration condition. dT: delta pilT, mC: mCherry, E: pilE-mCherry reporter, T: pilT-mCherry reporter, NV: not vibrated, V: Vibrated, ns: not significant, **** : highly significant p-value <0.0001

Figure 18: Spatiotemporal patterning of pilE-mCherry is disrupted when exposed to continuous vibration for 7 hours. Other reporters, pilT-mCherry, are not affected similarly. Scale bar is 10µm.

Figure 19: Size of microcolonies (diameter) of vibrated and not vibrated in WT and ΔpilT background. WT-mcherry and ΔpilT-mCherry microcolony size are significantly affected by vibration.

In summary, this line of evidence suggests that in the presence crude external forces can disrupt gene expression spatiotemporal patterns in microcolonies. External forces can modulate gene expression, which is independent of microcolony size.

3-3.8: Survival assay of WT and ΔpilT

PilT activity promotes microcolony formation by powering retraction dynamics of the tfp.

We hypothesize that PilT-generated force may behave as a global physiological cue in a mechanosensitive pathway. To address this hypothesis, we measured the growth and survival of WT, ΔpilT::Kan and 1:1 mixture in a liquid GCB media. All conditions grew at similar rates in the first 4 hours during exponential growth; however, ΔpilT::Kan populations sharply decreased at 8 hours in comparison to WT (Figure 20A). After 24 hours, ΔpilT::Kan survival was approximately 9.5 times lower on average in comparison to WT. While, this significant decrease is partially compensated in the mixture where WT consist half of the mixture. The average of counts after growth and survival for 24 hours for wild type Ng on

GCB culture plate was 7.992 × 107, ∆pilT::Kan was 0.834 × 107 and for 1:1 mixture of them was 1.262 × 107 based on the counts of colony forming units (CFU). After 48 hours all strains show no viable cells (Figure 20A). ∆pilT::Kan has a Kanamycin selection, therefore, a control set of WT, ∆pilT::Kan , and 1:1 mixture was plated on Kanamycin agar plates (Figure 20B).

Time (hours)

Figure 20A: Growth and survival of WT, ΔpilT, and mixture of WT, ΔpilT. CFU, colony forming units per mL plotted over time in hours. Blue bars, WT (wild-type) Ng; Green bars, Neisseria gonorrhoeae ΔpilT.; Red bars, 1:1 mixture of Neisseria gonorrhoeae WT and Neisseria gonorrhoeae ΔpilT.

In order to control for ∆pilT::Kan mutant activity, we complemented ∆pilT::Kan with a functional copy of pilT inserted at a neutral loci (the same loci as the transcriptional reporters). The complement-∆pilT::Kan survived at the same rate as WT, which demonstrates a rescued phenotype (Figure 20C). In addition, we also controlled for antibiotic selection during survival by plating on no selection (GCB), Kanamycin, and

Kanamycin-Chloramphenicol plates (for complement-∆pilT::Kan)

Figure 20B: Growth and survival assay for wild type Ng, Ng ∆PilT and 1:1 mixture of both on KAN plate. Wild type Ng cannot grow on KAN plate therefore value for blue bar is zero. Samples grow to the same range as it can be seen in bars of time point 0 to 8 hours then Ng ∆PilT dies faster than the mixture. Remember that in mixture, wild type Ng cells are dead but cell content and pili are present in mixture which could be the reason for the boost in survival. Ng stands for Neisseria gonorrhoeae and WT stands for wild type. Error bars are standard error of the mean (SEM)

Figure 20C: Growth and survival assay of complemented-∆pilT Ng strains in GCB for 24 hours. In order to count viable cells, samples were on GCB culture plates to quantify whole cells counts of all strains. GCB selective plates to quantify role mixtures (KAN plate is GCB selective plate for Ng ∆PilT and Chl + KAN plate is GCB selective plate for Ng complemented PilT and both plates do not allow the growth of wild type Ng). The ∆PilT complement, the survival rate is rescued to WT rates. The WT - ∆PilT (1:1) mixture, survival of ∆PilT increases from the GCB and GCB + KAN plate counts. For comparison of WT with ∆PilT, related bars are colored in red and for comparison of WT with PilT complemented (Blue bars). Ng stands for Neisseria gonorrhoeae and WT stands for wild type. Standard error of the mean (SEM).

To corroborate the theory that physical retraction forces are a critical developmental cue during microcolony formation, another 24-hour survival assay was performed with WT,

ΔpilT::Kan, and ΔpilE::Kan, while applying artificial force (same method as the previous vibration experiment). A ΔpilE::Kan genetic mutant where pilE gene was replaced with

Kanamycin cassette. PilE is the main pilus subunit, so without PilE a pilus is not produced.

Vibrations were applied at three stages: First 3 hours followed by 21 hours of no vibrations

(labeled: 3), first 3 hours without vibrations followed by 21 hours with vibrations (labeled:

21), and continuous vibrations for 24 hours (labeled: 24). We observed that the rate of survival for WT was comparable, except 21 under VCN selection showed a 10-fold decrease in comparison to the control (no vibrations) (Figure 21). We noted the rate of survival for

ΔpilT::Kan was not comparable across vibration time. In comparison to the control,

ΔpilT::Kan increased survival by 10-fold in both 3 and 24 condition, while decreasing dramatically by 500-fold in the 21 condition. Lastly, in ΔpilE::Kan the survival rates increased 100-fold under sustained and long-term vibrations of condition 21 and 24. In summarization, these results suggest that external physical forces applied to developing microcolonies is important in early development, especially shown in ΔpilT::Kan. This insight compounded with the ΔpilE::Kan survival increases implies the importance of physical force even in the absence of pili.

Figure 21: Artificial external force applied assay early increases survival rate of WT and ΔpilT, while sustained force increases ΔpilE, mutant lacking tfp, survival. WT GC, wild-type Ng; (0), no vibrations; (3), vibrations applied for first 3 hours, then 21 hours without; (21), no vibrations applied for the first 3 hours, then 21 hours of vibrations; (24) vibrations applied for continuous 24 hours.

3-2.8: Antibiotic resistance sensitivity in WT and retraction-deficient microcolonies

The emergence of multiple antibiotic-resistant Neisseria gonorrhoeae strains is a clear and imminent threat (“Antibiotic / Antimicrobial Resistance | CDC” 2017).

Microcolonies are precursors to biofilm formation and, as biofilms have been implicated in both antibiotic resistance and pathogen persistence in the environment, could be an important step in the gain of antibiotic resistance. Here we assessed whether accelerating microcolony formation plays a role in direct physiological role when exposed to antibiotics.

By spinning planktonic cells in a centrifuge, we can accelerate the interaction of cells in liquid media; therefore, spun cells form microcolonies sooner than unspun because of the immediate surface contact with other cells, and thus a higher local concentration of bacteria.

Spun (accelerated) WT cells and unspun cells were incubated for 1-hour in the presence of either growth media or growth media supplemented with one of six different antibiotics:

Erythromycin (Erm; 50µg/ml), Chloramphenicol (Chl; 250µg/ml), Nalidixic acid (NA;

30µg/ml), Kanamycin (Kan; 80 µg/ml), Ciprofloxacin (Ci; 0.12 µg/ml) and Cephalexin (Ce;

50µg/ml); the three being bacteriostatic and the latter three bactericidal at the minimum inhibitory concentration (MIC). We potentially modified the forces of interactions between the bacterial cells by spinning down of single cells prior to the 2.5 hours of microcolony formation, and subsequently assessed for viability after incubation with antibiotics (Figure

22). These findings suggest that tfp generated mechanical forces do indeed play an important role in modulating bacterial antibiotic susceptibility: even though the microcolonies have the

same size whether spun or not, their susceptibility to the antibiotics can be orders of magnitude apart. The most likely candidate is the difference of tfp retraction forces experiences by the bacteria (Persat, Inclan, et al. 2015b; Persat 2017).

Figure 22: Centripetal force exposure and no force exposure. Cell survival is measured by colony forming units. Erythromycin, Erm; Chloramphenicol, Chl; Nalidixic acid, NA; Kanamycin, Kan; Ciprooxacin, Ci; and Cephalexin; Ce.

ΔpilT bacteria cannot be spun because cells will adhere to the cell culture plate surface and not form any aggregates as they are non-motile. Although ΔpilT is incapable of producing mechanical forces, cells still can produce loose microcolony aggregates when let to settle by gravity. Antibiotic exposure survival assay demonstrates that both WT and ΔpilT are sensitive, which leads to an increase in cell death under all tested antibiotic selection.

Here, we see that despite the difference in microcolony formation dynamics, there is no difference in survival for WT and ΔpilT (Figure 23). Thus, the extra protection shown in spun

WT bacteria might indeed be due to the extra interbacterial forces acting during microcolony formation.

Figure 23: Antibiotic exposure 24-hour survival assay. Cell survival is measured by colony forming units. Erythromycin, Erm; Chloramphenicol, Chl; Nalidixic acid, NA; Kanamycin, Kan; Ciprofloxacin, Ci; and Cephalexin; Ce.

3-2.9: RNA-Seq demonstrate a global differential expression in WT and ΔpilT

Finally, we evaluated the global molecular activity WT microcolonies and ΔpilT to determine which genes are differentially gene expression due to tfp forces by RNAseq analysis. WT and ΔpilT cells were removed from a 16-hour overnight plate and dilated into growth media incubated in cell culture plates at 37C at 5% CO2. The 12 and 24-hour time points were prepared in duplicates for measuring purification variability. Total RNA was extracted at specific time points (Qiagen): 0-hour, 12-hour and 24-hour incubation. Samples were reverse transcribed to cDNA and prepared for sequencing according to the Illumina

Miseq Library Preparation Kit protocol. Sequencing was in Illumina 150bp PE 30X coverage

(Girihlet Inc.). Ribosomal RNA was removed from analysis. Normalization was performed by

Illumina by quantile normalization (see methods for more details). RNA-seq processing and quality control was performed using BioConductor with Biobase and edgeR package in R.

First, fragments per kilobase of exon per million reads (FPKM), or also called read counts.

These read counts are converted into gene IDs. DEGSeq2/limma program allows for measuring fold changing and also obtain a gene list of differential expressed targets.

From DEG2 analysis, 1782 genes were not differentially expressed. There was the most change detected at time 0-hour when comparing WT and ΔpilT. The 0-hour timepoint did not have a duplicate, therefore, quality of these samples was difficult to control; additionally, many of the differentially expressed were associated with metabolism. An early time point is difficult to establish experimentally, therefore, some flaws in this design may be present. Regardless, it would be interesting to perform a gene enrichment analysis of the

0-hour point, due to the sheer abundance of hits.

Additionally, I found there are 7 differentially expressed genes when all time points of WT were compared to ΔpilT (Figure 24). Unfortunately, 4 out of the 7 were 5s ribosomal

RNA. The other 3 were genes which are thought to

share an operon with pilT, CDS:NGFG_01978 (pilT

gene), CDS:NGFG_01979 (pilU gene), and

CDS:NGFG_01980 (FUSC-like gene) (Figure 25)

Figure 24: DEG analysis of Neisseria gonorrhoeae WT compared to ΔpilT at 0, 12, and 24-hour incubation (left)

100

Figure 25: FPKM of the top three differential expressed genes at all time points. Transcript read counts are plotted over time. Error bars are calculated as standard deviation.

Next, we measured of fold change between all conditions and found that there is a large shift between WT and ΔpilT at the 0-hour time point in comparison to 12 and 24-hour (Figure

26). Strikingly, this result shows approximately less than half of the genome is downregulated between WT and ΔpilT at 0-hour, while the other half was upregulated.

Interesting, those opposite set of genes are upregulated/downregulated at 24-hours (Figure

26). This demonstrates a global shift in expression over time. The bifurcation seen at 0-hours suggests the potential of a global role in ΔpilT regulatory pathways, and perhaps more global variability seen in gene expression at early time points.

101

Figure 26: Global transcription fold change Neisseria gonorrhoeae WT compared to ΔpilT at 0, 12, and 24-hour incubation. Blue, 4-fold decrease; Purple, 4-fold increase. Black, 0-fold change.

3-3: DISCUSSION

Gene expression in developing multicellular structures in animals have been understood to be controlled by many genetic regulatory functions, such as enhancers, promoters, and associated regulatory elements like transcription factors. However, gene expression in bacterial multicellular structures is still not well understood. Many different mechanisms might exist for the initiation and formation of biofilms across species. What is common among bacteria during biofilm formation is the similar transitional states: Planktonic state, microcolony, biofilm, and ultimately dispersal (Stoodley et al. 2002). Independently of the specific molecular mechanisms at play, all transitional states require a change in physical interactions between bacterial cells and between cells and their substrate (Drescher et al.

102

2016). In Ng, the physical process of developing a multicellular state is solely mediated by tfp making Neisseria gonorrhoeae an important model system to study the mechanical forces at play in biofilm formation. There is evidence to suggest that local tfp retraction force initiates a mechanosensitive signal transduction pathway opening the possibility of a feedback loop during biofilm formation (Paluch EK et al., 2015; Shivashankar et al.,2015; Cui et al., 2015).

We revealed that spatiotemporal gene expression patterns are established early in microcolony formation by using promoter fluorescence reporters as a proxy to measure transcriptional activity. pilE-mCherry produced the most extreme spatiotemporal pattern at

7-hours, as quantified by its high inhomogeneous index (H = 0.45). Some patterns are also homogeneous, like pilT, pilA, pilB, pilC2, and pilU, at both time points with ranging inhomogeneous index (0.2-0.3). Differential promoter activity suggests cooperation of cells in the microcolony; however, the global mechanism is unclear. The modification of the spatiotemporal patterns measured in ΔpilT::Kan reporters suggests that tfp retraction force modulate a mechanosensitive network responsible for inducing signaling pathways.

We also measured the gene expression and cell survival when exposed to modulating physical forces. By exploring the physiological response to continuous artificial forces, we found that continuous vibrations applied to developing microcolonies did not affect morphology; however, vibrations did disrupt gene expression spatiotemporal patterns. In pilE-mCherry and consensus sequence, lacUV5, sensitivity to vibration, suggests an induction of either a global stress response or revealing a mechanosensitive pathway.

Additionally, survival rates of retraction-deficient and pilus-deficient mutants were considerably rescued in the presence of artificial vibration forces. This evidence is

103

corroborated with the 1:1 WT-ΔpilT survival mixture, which showed a modest rescues of survival rate in comparison to the ΔpilT alone.

Moreover, by accelerating the time bacteria experience intercellular tfp forces during microcolony formation reduces sensitivity to a wide spectrum of antibiotics. In other words, when planktonic cells engage the retraction dynamics of tfp, seem to initiate a different physiological state which protects them against small molecular exposure. When unspun WT and unspun retraction-deficient, ΔpilT, are exposure to the same antibiotic regime as spun

WT, unspun WT and ΔpilT have the same survival rate. Collectively, these data add credence to the theory that mechanics play a crucial role in physiological development of microbial structures.

The pilE-mCherry reporter demonstrated the most robust gene expression pattern changed when exposed to modulating forces. The pilE promoter suggests that PilE is regulated by a diverse set of transcription factors, Ihf, PhoB, ArcA, and σ70, reported from previous transcriptional protein binding site analysis (Hill et al. 1997; Hill and Davies 2009;

Masters et al. 2016). Therefore, further study of the pilE regulation may lead to new mechanosensitive targets. We believe that the results of this study lay the fundamental knowledge of early biofilms physiology and the role of tfp retraction force. Since tfp-bearing bacteria are common among pathogens, it is important to investigate more about microcolony physiology and the role of tfp retraction force in other pathogenic and non- pathogenic bacteria.

104

3-4: METHOD

3-4.1: Culture conditions

Neisseria gonorrhoeae was cultured on Gonococcal Broth (GCB) with VCN (Nystatin, Colistin, and Vancomyocin) agar plates with Kellogg’s supplements (Dextrose, L-Glutamine,

Cocarboxylase, ferric nitrate) (Schoolnik et al. 1984). Incubated at 37°C and 5% CO2. In order to have the most live cells, it is advised that the plate be incubated no longer than 18 hours.

Frozen bacterial stocks are stored in freezing media (20% glycerol with 80% GCB) stored at

-80°C.

3-4.2: Promoter reporter construction

Gibson Assembly method (NEB #E2611S) ligated DNA fragments for 60 minutes at 50°C.

Genetic cassettes of promoter, mCherry and antibiotic resistance genes were inserted into locus CP003909 at 1489951bp (Dillard 2011). Primer design using Geneious 8.1.6 software.

Verify ligation by TAE ethidium bromide gel electrophoresis (RE Biotechnology Grade,

Amresco). Ligate desired DNA fragments for 60 minutes at 50°C. Amplify ligated construct using GoTaq polymerase master mix (Promega #M7122) along with 10µM forward and reverse primer. To transform the construct in Ng, spot 10µl of verified ligated construct

(~100 ng/µl) onto GCB agar plate, and inoculate with Neisseria gonorrhoeae WT bacteria.

Incubate overnight at 37°C 5% CO2, then select for positive colonies with construct.

Sequenced all constructs using GeneWiz.

105

3-4.3: Primers used for Neisseria gonorrhoeae pilus-related promoter amplification

Promoter Forward Primer Reverse Primer Promoter length (bp) pilA gggtatagagcagaacggatAAA cctcgcccttgctcaccatTGAATTT 400 CATAAAGGACACCAGC TCCTTTTAATTTT pilB gggtatagagcagaacggatACT cctcgcccttgctcaccatTTTGATG 238 TCAGATTCTACTTTTG TTTGTGTGTGG pilC1 gggtatagagcagaacggatATT cctcgcccttgctcaccatAGTGTGT 589 TTATTCCTTAAGTTTGCC ATCCGAGTTTGTGC TTC pilC2 gggtatagagcagaacggatGCC cctcgcccttgctcaccatAACCTGT 934 CGGTATGGGAAAACAT TTTCCTTGCAGCAAAC pilE gggtatagagcagaacggatCCC cctcgcccttgctcaccatCATCCGT 393 CACCCAACCCACCCG TCTGCTCTATACC pilT gggtatagagcagaacggatTCC cctcgcccttgctcaccatCATCCGT 368 GCCAATTCCTCCGTTTTG TCTGCTCTATACCC pilU gggtatagagcagaacggatCAC cctcgcccttgctcaccatCGTTAGC 163 ACAACCGCC TTCTTTTCGG

LacUV5 consensus sequence: Cgctgctgccgctgtaatacggcttgacactttatgcttccggctcgtataatgtgtggatagtgggaggaaagc

3-4.4: FISH probe sequence and fluorophore information mRNA target Sequence Fluorophore pilE-131 GAC CTT CGG CCA AAA GGA Alexa555 pilE-373 CTG TCC GCA GAA CCA TTT Alexa555 pilT-347 GCT GGG TTT CGT TGA TGT Alexa488 16sRNA CCT GTG TTA CGG CTC CCG Alexa488

3-4.5: Fluorescent In Situ Hybridization (FISH)

RNAfold WebServer to predict the secondary structures in the RNA. RNA parameters for incorporating chemical modification constraints into a dynamic programming algorithm for prediction of RNA secondary structure (Mathews et al. 2004). https://www.biosearchtech.com/support/education/stellaris-rna-fish provides secondary structure information helped me choose one probe over another one.

106

Fixation: The microcolonies were fixed the day before. All microcolonies were raised directly in 6-well plates at OD600 0.07. Each sample was fixed with 4% formaldehyde in PBS (total

2% formaldehyde). I did not remove growth medium during fixation. Samples were fixed for

1 hour, dehydrated with 50%. 80%, 100% EtOH for 3 min each. Each well was air dried for

2 minutes, then washed quickly 2X with PBS. Stored with 500ul of PBS at 4C.

Hybridization step: Plasma cleaned round glass coverslips and were decorated with 100ul of poly-lysine for 1-hour Rinse with 1ml PBS and placed into clean observation chamber.

Poly-lysine side should be facing upward. Look at fixed microcolonies in 6-well plate, just to make sure the morphology is good. With a 1ml pipette, scrap microcolonies off the bottom of the 6-well plate. Scraping is required. Add 1ml of microcolonies PBS onto poly-lysine glass in observation chamber. Allow the microcolonies to slowly fall to the glass in the liquid (~30 minutes). Check to see if the microcolonies are on the cover glass before you begin hybridization. Slowly removed with a pipette the total volume of PBS (should be 1ml) out of the chamber. Add 100ul of 2.5ul 50ng/µl + 100µl warm hybridization buffer* in the chamber on top of the glass. Hybridization for 2 hours at 46C. Wash 3X with hybridization wash solution. Image using 100X DIC on Rogue.

*mix together probe and buffer in a microfuge tube ahead of time. Use 20% formaldehyde for these probes. Remember, the concentration of formaldehyde may differ between probes.

107

3-4.6: Microcolony assays Filter GCB with 0.2 µm syringe filter (VWR #28145•477). Following 14-16 hours of growth, remove cells with polyester swabs (VWR #22222-046) and resuspend in 1 mL GCB plus supplement. Vigorously shake bacterial suspensions for 2 minutes using ‘Disruptor Genie’.

Vortexing ensures single non-aggregated cells. Dilute cultures to OD600 0.7 in GCB. Add

100μl of dilution to 900μl GCB supplement in 6-well plate (Costar #3516) and incubate at

37°C at 5% CO2. 500μl of culture was transferred to Attofluor chambers (Life Technologies

#A7816).

3-4.7: Image analysis

General information: For each colony images of two channels were used: DIC channel allowed for the identification of the colony boundaries, the fluorescent channel allowed for the analysis of the gene expression profile. MatLab R2017b & R2019a were used for edge detection of DIC and the analysis of fluorescent patterns within colonies.

Edge detection: To detect the edges from the DIC images we tried two different methods for each colony and manually choose the method that better describes the edge. In the first method we apply Otsu’s method to threshold the DIC image. In the second method we first compute the absolute value of the first derivatives in x- and y-direction of the image and then apply Otsu’s method for thresholding. We then apply morphological operators (dilation, erosion and hole filling) to generate the binary image of the colonies and do not consider colonies at the edge of the DIC image. We manually disregarded colonies that formed a recent coalescence event and that typically have a dumbbell shape or a clear signature of the two previous colonies in the fluorescence channel. We only picked colonies with a solidity > 0.8

108

(ratio of the area and the convex area of the binary shape of the colony) and an axis ratio >

0.8 (ratio of short and long axis of an ellipse fitted to the binary colony shape). Additionally, we computed the radius 푅 of the circle with the same area as the binary shape of the colony and only pick those colonies with 4 µm < 푅 < 10 µm.

Background subtraction: To estimate the background of the fluorescent channel, we manually choose a rectangle in the image and without any cell and calculate the mean intensity. This mean intensity is subtracted from the fluorescence image.

Estimating the length scale of the fluorescent patterns: We perform a distance transform of the binary image computed from the DIC channel. We then pick pixels in a certain interval of distances from the colony edge and compute their mean intensity. We choose interval sizes of 0.1 µm. We then compute the mean intensities for all intervals between 0 µm and the 30

µm and create the intensity profile. To this profile we fit a function of the form 퐼(푟) = 퐼0 −

퐼1 exp(− 푟⁄푑푐 ), allowing us to estimate the gradient length scale 푑푐.

Inhomogeneity index: To explain the behavior of the inhomogeneity index and to what type of profiles it corresponds, we study its behavior for a profile of the form

푑 퐼(푑) = 퐼0 − 퐼1 exp (− ) 푑푐 with the intensity 퐼0 − 퐼1 at 푑 = 0, the intensity 퐼0 for 푑 → ∞, the characteristic length scale of the gradient 푑푐 and the distance from the surface of the colony 푑. Additionally, we assume that the colony has the radius 푅. The normalized profile, fulfilling

109

푅 ∫ 퐼푛(푑) d푑 = 1 0 is then given by

푑 1 − 훼 exp (− ) 푑 퐼 (푑) = 푐 푛 푅 푅 + 훼푑푐 [exp (− ) − 1] 푑푐

퐼 With the intensity ratio 훼 = 1. The inhomogeneity index is defined as the absolute difference 퐼0 1 of this profile with the homogeneous and normalized profile 퐼 (푑) = for 0 ≤ 푑 < 푅. It is ℎ 푅 calculated by

푅 1 퐻 = ∫ |퐼푛(푑) − | d푑 0 푅 and given by

1 1 1 2훼푑̃ [−푑̃ + [푑̃ − 1] exp ( ) − 푑̃ [exp ( ) − 1] ln (푑̃ − 푑̃ exp (− ))] 푐 푐 푐 푑̃ 푐 푑̃ 푐 푐 푑̃ 퐻 = 푐 푐 푐 ̃ ̃ 1 −훼푑푐 + [훼푑푐 − 1] exp ( ̃ ) 푑푐

푑 with the rescaled gradient with 푑̃ = 푐. Homogeneity indices for different values of 훼 and 푑̃ 푐 푅 푐 are given in the Figure. Importantly, the inhomogeneity index does not increase

̃ ̃ monotonically for increasing characteristic colony lengths 푑푐 since, if 푑푐 is too small, the

̃ profile quickly relaxes to a constant profile. If 푑푐 is too large, the profile is only slowly increasing and as a result the inhomogeneity index will become smaller.

110

Figure 12I: Inhomogeneity indices of colonies with an exponential profile given by equation 1 for different value intensity.

3-4.8: Microscopy

All images were obtained on a Nikon Ti Eclipse inverted microscope equipped for epifluorescence and DIC microscopy all under an environmental chamber maintaining temperature, humidity and CO2 concentration. The objective used is an oil immersion Plan

Apo VC 100X objective, N.A 1.4, refraction index 1.51. DIC images were captured at 100ms exposure, gain 2.3X, binning 1x1. TexasRed images were excited at 555nm, captured at

111

500ms exposure, gain 1.5X, binning 1x1, and laser power of 50%. The camera used was digital SLR CMOS Nikon DS-Qi2, 16.25 megapixel.

3-4.9: Survival and antibiotic resistance assays

Bacteria were streaked from frozen stocks and incubated overnight at 37°C, 5% CO2 and humidity. At 24 H, colonies were lawned and incubated for 18 H at 37°C, 5% CO2 and humidity. 18H lawns were resuspended in 1 ml GCB, shaken 3 min on a disruptor, and assayed for OD600nm. Bacteria were then normalized to an OD600 0.7 and 100ul used to inoculate 1 ml of GCB (~5x107 cells). Plates were either spun down at 3500rpm, or not, and bacteria were then incubated at 37°C, 5% CO2 with humidity for 2-hour before addition of the selective agent. 1 ml of 2X concentration antibiotics were added following a pre- incubation period and left for either 1 hour. Following incubation under antibiotic selection, cells were scraped, disrupted for 3 min, diluted in a 1:10 dilution series, plated on agar and grown overnight at 37°C, 5% CO2 and humidity. Control for WT, ∆pilT::Kan, ∆pilT- complement, and WT, ∆pilT::Kan mixture in growth and 48-hour survival assay.

3-4.10: RNA purification for RNA-seq

Performed with Qiagen RNAeasy mini kit protocol. Mix 2 volumes (1000ul) of RNAprotect with 500ul of media. Pipet solution onto agar plate for 0-hour time point. Carefully remove lawn and add to a tube. Immediately mix by vortex and incubate at RT for 5 minutes.

Centrifuge 10 min at 5000 gX (Eppendorf 5424, F-45-24-11 Set to 7297rpm) Decant supernatant. Remove all the liquids. Add at 200ul TE buff containing lysozyme (1mg/ml,

112

make fresh). Vortex for 10s. Incubate at RT for 5 min. During incubation--vortex every 2 min for 10 sec. Add the appropriate amount (700ul) of Buffer RLT and vortex with vigor. If particulate material is visible, centrifuge it and add supernatant to next step. Add appropriate amount (500ul) of 96-100% EtOH. Mix by pipetting. Add lysate to spin column placed in a 2 ml collection. Centrifuge for 15sec at 8000xG (9230rpm). DNAse treatment: Add

350ul of RW1 buffer and centrifuge for 15s @9230rpm. Discard flow through. In a separate tube: Add 10ul of DNAseI stock to 70ul Buffer RDD Mix gently in tube. Add the DNase I treatment directly to spin column membrane and incubate at room temperature for 15 min.

Add 350ul of Buffer RW1 to column, wait for 5 minutes and then centrifuge for 15 s at

9230rpm. Discard flow through. Place column in a new tube. Add 500ul RPE to column and spin for 15 s @ 9230 rpm. Repeat, except centrifuge for 2 minutes. Discard flow through.

Centrifuge at full speed for 1 min. Place column in RNAse-free tube and add 50ul of RNAse- free water and elute for 1 min @9230rpm.

3-4. 11: Principal Component Analysis for RNA-SEQ

Normalized sequences from the RNAseq was performed to read the transcript counts. Based on the file name (qnorm) it is thought these sequences were normalized using quantile normalization. Quantile normalization requires binning data into divisions; therefore, outlier numbers can be removed or adjusted in comparison to the remaining data. The schematic represents binning of data based on transcript frequency (figure below). “x” # of divisions are made to ensure the mean and all divisions are the same distributions. This is an extra normalization, and could lose small significant effects. Upper quantile: “x”/# reads;

113

however, it can be dominated by high expression genes. The upper quantile is represented by the bold 75%-100% . The box plot to the left shows the data, which was normalized of each sample (Figure 27). The mean of the distribution is the same between all samples, which does verify that the data was normalized. FPKM stands for Fragments Per Kilobase of transcript per Million mapped reads. In RNA-Seq, the relative expression of a transcript is proportional to the number of cDNA fragments that originate from it.

Figure 27: Log Expression box plot of each sample demonstrates normalization after sequencing.

We can see in the Figure 28 shows similar strains and replicated group together, which is what we would expect if our replicates are accurate. The x and y-axis are the LogFC, the log fold change between all conditions and replicates. Replicates are close together indicating similarity between sample preparations. Only one sample exists for WT and dpilT at 0-hour time point.

114

Figure 28: Distribution after normalization of samples and replicates. Each replicate shows close proximity to each other, indicating the samples have similar quality.

115

CHAPTER 4

Efforts to characterize physical properties of Neisseria gonorrhoeae microcolonies

This section as two parts:

I. EFFECTS OF OSMOTIC PRESSURE ON DEVELOPING MICROCOLONIES

II. MICROPLATE SQUEEZING EXPERIMENT TO MEASURE MICROCOLONY PLASTICITY

Author Contributions: All experiments, writing, and analysis is by Kelly Eckenrode

116

4-1: EFFECTS OF OSMOTIC PRESSURE ON DEVELOPING MICROCOLONIES

4-1.1: INTRODUCTION

Mucus is a complex material that lines surfaces like lungs, gastrointestinal tract, vagina, and eyes. Mucus acts as a physical barrier between epithelial cells to protect against foreign particles, as well as providing lubrication. Mucus is mainly composed of gel-forming glycoproteins and mucins (Lai et al. 2009). In vertebrates, gel-forming mucin genes are well- conserved. Interestingly, mucin protein domains have evolutionary origins in bacteria and can be detected in phyla of Flavobacteria, Deltaproteobacteria, Sphingobacteriia, Cytophaga, and Gammaproteobacteria (Lang et al. 2016).

Recent advances in rheology technology has offered a deeper understanding of the physiology of mucus. Mucins, gel-forming proteins, produce interesting physical properties that are defined by rheology methodologies measuring responses to shear rate and shear stress. Mucus is a non-Newtonian gel at a macroscale; however, at the nanoscale, mucus behaves like a low viscosity fluid (Lai et al. 2009).

Although microscale organisms, like bacteria, may not succumb to the effects of high osmotic pressures, there is a connection between changes in mucosal environments and the virulence of a pathogen. Studies of the mucus viscoelastic properties, sputum microbiology, and mucus clearance rates show that increased mucus concentrations in chronic bronchitis patients were positively correlated with mucus partial osmotic pressure and negatively correlated with mucus clearance rates. In both chronic bronchitis and cystic fibrosis, the hyper concentration of mucin is associated with pathogen persistence. In addition, there is

117

an association with specific mucin proteins and colon cancer (Hollingsworth and Swanson

2004).

What interests me about mucus is the dynamic range of its physical properties. The gel-like substance can create different rigidities based on its viscosity. Viscosity of a liquid is the measurement of its resistance to deformation over time, or more informally known as the ‘thickness’ of a liquid. This means viscosity is measured under flow. Mucus is not a static substance because epithelial cilia beat which generates flow adding another dynamic component to bacteria colonization (Fliegauf et al. 2013). Separately, mucus can also contribute to the osmotic pressure, which is the number of molecules in a liquid, by changing the amount of mucin glycoproteins present in a liquid matrix. Both of these physical factors produce a complex architectural landscape that can either select for or against bacterial colonization. In respiratory organ culture, Neisseria meningitidis are known to colonize non- ciliated separated cells, unlike other pathogens, like Pseudomonas aeruginosa, where cells colonize exclusively at ciliated and mucosal regions (Dowling and Wilson 1998). In addition, mucins have been noted for triggering biofilm dispersal of Pseudomonas aeruginosa (Co et al. 2018). As a result, I believe it is imperative to understand the physiological relationship between the mucus and bacteria due to the nature of bacteria colonization in mucosal membranes.

Without access to human mucus samples, I designed an artificial system to test the effect of osmotic pressure typical of mucus on bacterial biofilm formation. Previous studies used a non-metabolic thickening agent, dextran (MW 100kDa), to test the effect of osmotic pressures on tumor spheroids (Montel et al. 2011). The pressure of mucus gel layer typically ranges from 150 Pa to 5000 Pa, respectively, so I added specific concentrations of dextran to

118

Neisseria gonorrhoeae growth media to test the effect of microcolony formation in a range of geometric constraints similar to the native Neisseria gonorrhoeae environment. I measured

24-hour survival, live/dead cell arrangement in microcolonies, pilin production, and size of microcolony formation of Neisseria gonorrhoeae WT and ΔpilT::Kan in a range of osmotic pressures.

4-2: RESULTS

4-2.1: Osmotic forces negatively effects microcolony formation in WT, but not ΔpilT

I wanted to measure whether Neisseria gonorrhoeae microcolony formation in the presence of a range of osmotic pressures would affect assembly. I designed an experiment where planktonic Neisseria gonorrhoeae cells were exposed to a range of osmotic pressures in growth media. To make osmotic pressure solution, I used dextran-20 (molecular weight:

15,000-20,000 Da) to produce a range of pressures representative of mucosal osmotic pressure: 100Pa, 500Pa, 1000Pa, 2500Pa, and 5000Pa. Dextran is used in eukaryotic methods to test the effects of osmotic pressures because it is known not to be metabolized.

In addition, the chemical structure of dextran is ideal for producing a ‘caging’ effect to create an osmotic pressure difference. Neisseria gonorrhoeae also does not have any known metabolic pathways for degrading dextran; therefore, it was used for these experiments.

Single Neisseria gonorrhoeae cells were incubated for 24 hours in pressurized dextran-20 growth media (see Methods for more detail). To first measure the effect of the pressure exposure, I observed microcolony morphology. I compared the surface area of both

WT and 훥pilT::Kan aggregates in GCB, without dextran, as a control. Microscope images

(20X) demonstrate a representative image of the morphology observed (Figure 29). I

119

measured the surface area of microcolonies by thresholding microscope images into a binary by using FIJI (IMAGEJ). Then, I could measure the area of by ‘Analyzing Particles’ by binning the pixel size from 5 to 5000. I plotted the area of WT and 훥pilT:Kan by pixel size (Figure

30). From this analysis, I can deduce that dextran-20 inhibits WT microcolony formation at

500Pa. At the highest concentration of dextran-20, 5000Pa, is the same size as all ΔpilT::Kan data points.

Figure 29: Microcolony morphology after 24-hour exposure to pressurized growth medium. A. Neisseria gonorrhoeae WT; B. Neisseria gonorrhoeae ΔpilT::Kan..

120

Figure 30: Osmotic pressure negatively impacts microcolony colony size of WT, but with little impact on ΔpilT. (Pa, Pascal; dT, ΔpilT). Error bars show standard error of the mean (SEM).

4-2.2: Survival of WT is sensitive to high forces and ΔpilT is less sensitive

Due to the complete loss of WT microcolony formation at 5000Pa, I had hypothesized the pressure either killed or slowed down the growth of WT cells. Therefore, I performed a

24-hour survival (Figure 31). I measured a slight decrease of WT survival (100- 2500Pa), while at 5000Pa, survival was reduced by 1-fold. In WT, at 2500Pa, there appears to be an over-compensation of growth. I do not have an explanation for this effect. Overall, ΔpilT survival was less affected by osmotic pressure than WT.

In addition, I measured the amount of growth at 7-hours in all pressure conditions in only WT. I measured at 7-hours because this is a critical time point for microcolony formation. I measured a linaer decrease in respect to increasing osmotic pressure (Figure

32). I repeated survival assays and microcolony formation assays (N = 2). The result shows

121

that as dextran-20 concentration increases, there is a decrease in WT survival. Interestingly,

ΔpilT is less sensitive to dextran-20 than WT.

Figure 31: 24-hour survival in different osmotic pressure solutions of WT and ΔpilT. CFUs (colony forming units) are reported as cell per mL.

122

Figure 32: Growth of WT at 7 hours. Optical density plotted at different dextran-20 osmotic pressures. Red dotted line signifies starting concentration of growth assay. The two lines represent two trials (purple and black).

4-2.3: Survival of WT and ΔpilT in Dextran-100, PEG-10 and sorbitol does not show any effect

I wished to parse whether the effect I measured previously in dextran-20 was not a chemical effect, but was a physical effect created by osmotic pressure. Therefore, I choose two more osmotic pressure solutions that are not known to be metabolized by Neisseria gonorrhoeae:

PEG and sorbitol. In addition, I tested a larger molecular weight dextran, Dextran-100 with a M.W of 100,000 - 250,000). Survival is 5-10 times increase to the dextran used in the previous experiments. When preparing solutions, I assumed the relationship of pressuring agent in growth media had a linear relationship at small pressures; therefore, I could use the ideal gas law to derive how much reagent is required (see Methods). From this survival experiment, I measured nearly no effect of survival using other solutions, except at high pressures of 50,000Pa with PEG-8000 (Figure 33).

123

Figure 33: Osmotic pressure survival assay with PEG-8000, Dextran-100, and Sorbitol. The survival of both WT and ΔpilT. Colony forming units were plotted as cell per mL. Pa is pascal.

4-2.4: Pilin production decreased in higher osmotic pressure

Due to the negative effect seen in WT microcolony formation using Dextran-20, I wanted to measure whether tfp production is impacted by osmotic pressure. Therefore, after exposure to the range of osmotic pressures after 24hrs. Whole cell lysates were normalized, prepared, and run by electroporation in SDS-PAGE gel. The gel was stained with Coomassie brilliant blue to evaluate a crude comparison of relative protein concentrations per reaction

(Figure 34). Each lane appears to stain equally across all samples; therefore, I conclude that each sample have whole cell lysate was roughly equivalent. Ladder is PageRuler™ Prestained

Protein Ladder, 10 to 180 kDa (left two lanes and ninth lane in Figure 34).

124

Figure 34: Coomassie gel of Neisseria gonorrhoeae samples after osmotic pressure exposure. (left- right) Ladder, ladder (duplicate), WT GCB, WT 100Pa, WT 500Pa, WT1000Pa, WT2500Pa, WT5000Pa, ladder, ΔpilT GCB, ΔpilT 100Pa, ΔpilT 500Pa, ΔpilT 1000Pa, ΔpilT 2500Pa, ΔpilT 5000Pa.

Figure 35: Western Blot of EFTU (top band) and SM-1 (pilin). SM1 targets Neisseria gonorrhoeae pilin. 0, 100, 500, 1000, 2500, 5000 represent osmotic pressures (Pa) from exposure to dextran-20 for 7 hours. dpilT is ΔpilT.

125

Measuring the amount of pilin produced seemed to be dramatically decreased in WT under any pressure beyond GCB. The results showed decreased amount of pilin purified as pressure increases in WT, but not in ΔpilT. The top band, EFTU, is the positive control for protein concentration (Figure 35). Because the retraction-deficient mutant, ∆pilT, does not demonstrate the same sensitivity to osmotic pressure as WT, the null effect seen in ∆pilT suggest a mechanosensitive role PilT plays in signaling (Dietrich et al. 2009).

4-3: DISCUSSION

This preliminary research into the effect of Neisseria gonorrhoeae microcolony formation demonstrated some key findings. First, WT microcolony formation was negatively impacted by the presence of dextran-20, which is assumed to create osmotic pressure around single cells. Secondly, the amount of WT pilin production is dramatically decreased in the presence of osmotic pressure in comparison to ΔpilT. In the western blot results, pilin protein is not present in any of the osmotic pressure conditions, while the ΔpilT maintains the same intensity. Since this result was dramatic, it is important for me to highlight a potential experimental flaw. From the two western blots I ran, I had an increased intensity and cleaner bands in the pilin prep when I added the Laemmli buffer directly into the 6-well plate, instead of removing the cells first, then adding the Laemmli buffer. However, this method did not allow me to normalize the cells, so it is possible that more cells were added in the ΔpilT samples than the WT. Lastly, it is difficult to interpret the results of the other osmotic pressure reagents, PEG and sorbitol, since there was nearly no effect on survival, in comparison to the dextran-20 results. I conclude that the effect of dextran-20 might be a result not of osmotic pressure, but of altering the motility and pilus production. Dextran-20

126

may serve as an important reagent for studying the altered tfp motility. As a result, I recommend a future experiment measuring the tfp retraction force using the micropillar analysis under in 5000Pa of dextran-20.

4-4: MICROPLATE SQUEEZING EXPERIMENT TO MEASURE MICROCOLONY PLASTICITY

4-4.1: INTRODUCTION

The mechanical properties of cells can provide an understanding about physiology, structure and aggregate formation. Change in mechanical properties, like rigidity, can mark important physiological events in cells (i.e. pathogenicity of human disease). Think about squeezing a foam stress ball. If you squeeze it, you can get a sense for “how squishy is this material?”. If we record the compression of the stress ball, we can calculate the deformation and the elasticity of the material. We can apply this principle to biological materials.

In chapter 2, our microcolony motility model based on experimental motility data can simulate a symmetric spherical structure of a microcolony based solely on the physical interaction of the type IV pilus. This suggests less binding of the type IV pili on the outside of the microcolony than inside. This result predicts that perimeter cells are more motile than cells within the bulk of the colony; therefore, the heterogeneous pattern of force in WT microcolonies. However, the differentiation of motility is lost in ΔpilT mutant microcolonies.

As a result of those data, I wanted to measure whether motility played a role in the overall plasticity of the microcolony structure. If the microcolony were deformed, would the microcolonies reform, or stay permanently deformed?

127

4-5.2: RESULTS

4-5.2.A: ΔpilT microcolonies are extremely plastic

Neisseria gonorrhoeae were prepared for a 3-hour microcolony assay in growth media. 3- hour microcolonies were transferred inside a plastic containment chamber on the light microscope (Figure 36). The containment chamber had a glass coverslip adhered to the bottom, which allowed for imaging. In addition, a small piece of PDMS was also adhered to the glass bottom. Essentially, the method was designed to squeeze living microcolonies between a ~500 µm thin glass microplate and a PDMS block (Figure 37). The squeeze and reformation events were captured using time lapse imaging (see methods for details). As a result of this experiment, I measured the size of microcolony and plotted the size (pixels) over time. The analysis allows us to get a sense of the rate of reformation after deformation, i.e., plasticity. From this analysis, I measured that WT microcolony mostly reforms within 5 minutes (WT takes 30 minutes to completely reform, data not shown). However, ΔpilT microcolonies instantaneously reform shape 1 second after deformation with microplate

(Figure 37).

128

Figure 36: Experimental set-up of growth media containment chamber on the microscope stage (illuminated by green light). Microplate is attached ‘hockey-stick’ (vertical center) held by the micromanipulator. ‘Hockey-stick’ is lowered into the containment chamber and into the growth media.

129

Figure 37: Plasticity of WT and ΔpilT. A: Stills from Neisseria gonorrhoeae WT time lapse, B: Stills from Neisseria gonorrhoeae ΔpilT time lapse. The 1 second time point shows the moment where microcolonies were squeezed (A, B). C: Diameter of microcolony (in pixels) measured and plotted over time. N = 1

130

4-5.3: DISCUSSION

Characterizing the behavior of a material allows for a deeper understanding of the structures which influence its properties and performance. By applying compression stress via a microplate onto a microcolony provides information about the bulk modulus, or the plasticity, of this active biomaterial. Cells in a microcolony are adhered together by active polymers, tfp, so we had hypothesized that the material property of a WT microcolony would be different from a non-active retraction-deficient, ΔpilT, microcolony. I observed that WT re-formed its circular shape after approximately 20 minutes, but ΔpilT instantaneously sprung back after deformation. These results suggest a difference in the material phase, where ΔpilT behaves more like a liquid, where WT behaves more like an elastic-solid. The

ΔpilT result also suggests that PilT plays a role in decreasing the rate of formation after deformation.

131

4-6: METHODS

4-6.1: Dextran microcolony assay

Filter GCB with 0.2 µm syringe filter (VWR #28145•477). Following 14-16 hours of growth, polyester swabs (VWR #22222-046) remove cells and resuspend in 1 mL GCB plus supplement. Vigorously shake bacterial suspensions for 2 minutes using ‘Disruptor Genie’.

Vortexing ensures single non-aggregated cells. Dilute cultures to OD600 0.7 in GCB. Add

100μl of dilution to 900μl into polystyrene 6-well cell culture dish (Costar #3516). Incubate at 37°C at 5% CO2 for desired time, either 7 hour or 24 hours. To make dextran pressure solutions, I used ideal gas law to measure the volume of dextran required to produce a desired force. As an example: P = nrt ; Pi = MRT ; M = pi / RT

if R = 0.0826L atm mol-1 K-1

then, M = (5 x 103 / 101325) * (1 / 0.08206)

= 5 x 103 / 103 x 300 x 8.2 x10-2

=5 / 300 * 8.2 = 2mM

4-6.2: Image analysis

Measured diameters of individual microcolonies using the line measurement tool in FIJI, FIJI is just ImageJ. Schindelin, J.; Arganda-Carreras, I. & Frise, E. et al. (2012), "Fiji: an open-source platform for biological-image analysis", Nature methods, 9(7): 676-682, PMID 22743772 (on

Google Scholar). Measurements were reported in pixels because a micrometer to calibrate the stage was unavailable at the time.

132

4-6.3: Western Blot

SDS-PAGE GEL Reference: Mini-Protean Tetra Cell Instruction Manual

SDS-PAGE gel buffer I 1.5 M Tris-HCl (for separating gel) 118.2 g of Tris-HCl in H2O, pH 8.8 Final volume 500 mL Filter and degas

SDS-PAGE gel buffer II 1 M Tris-HCl (for stacking gel) 78.8 g of Tris-HCl in H2O, pH 6.8 Final volume 500 mL Filter and degas *Careful about the use of either Tris-HCL or Tris Base here you should use TrisBase and follow the instructions that are on the booklet nearby the gel making dock

SDS-PAGE 10X gel running buffer 248 mM Trisma (60 g) 1.92 M glycine (288 g) 1% w/v SDS (20 g) Final volume 2 L No need to pH, filter, or degas Dilute to 1X for running SDS-PAGE gels

10X transfer buffer 288 g glycine 60.4 Tris base 1.8 L dH20

Making the Gel ● Prepare 30% Acrylamide/Bis Solution Stock: ● 40% Acrylamide ● bis-acrylamide powder

Set up the glass to fit in the holder so that it is sealed and polymer solution will not leak out.

Prepare the 15% gel using the resolving gel buffer of manual. Add enough gel so that there is room to add the comb (leave a 1cm space in addition to the height of the comb for the stacking gel). Carefully pipette 50% isopropanol onto the top of the Acrylamide polymer

133

solution to help flatten the gel and stock oxygen from poisoning polymerization. Let sit for 1 hr. Pour off the isopropanol. Wash remaining isopropanol with DI water and then prepare the stacking gel (2ml) (4% gel) and add the comb. Let sit for 1hr. Store gel in sealed humid environment at 4C for up to 2 months.

Running the Samples Prepare Laemmli loading buffer Stock: Bio-Rad Laemmli Sample Buffer (950ul) and β- mercaptoethanol - (50ul). Wash the gel with water and then rinse out the wells with the running buffer that is added to the top of the gel in the holder and the bottom, so that the Pt electrode is covered complete. Prepare the Gel so that it fits into a holder that goes into the electrophoresis chamber. The gel stays sandwiched between the glass slides! Prepare the samples 1:1 with Laemmli loading buffer stock. Heat at 90-95C for 10 minutes on dry heat block. Load 10ul per well. (mark which wells used in lab notebook). Load 2µl the Protein

Ladder (PAGE ruler + Prestain Pro ladder, ThermoCycle #26616). Run at 120V for ~1hr or till bands hit the bottom of the gel.

Transferring to nitrocellulose membrane

Equipment 1. 1X transfer buffer 2. Black Gauze pad 3. Western blot sandwich holder 4. Trans Blot Paper 5. Nitrocellulose Membrane paper

134

Transfer Buffer (for Western blotting) - 2 L 25 mM Tris, 192 mM glycine, 10% methanol

1X Transfer Buffer 10X Transfer Buffer

28.8 g glycine 288 g glycine

6.04 g Tris base 60.4 g Tris base

200 ml methanol - methanol

1.6 L ddH2O 1.8 L ddH2O

NOTE: for the proper transfer of large proteins, up to 0.5% SDS may need to be added to 1X Transfer Buffer. **

Dissolve Tris base and glycine together in 1.6 L of ddH2O. Add methanol and mix. Add ddH2O to a final volume of 2 L.

Transfer methods After performing an SDS page gel electrophoresis to completion (laemmli buffer has reached the end of the polyacrylamide gel), prepare for a western blot. [Remember to load the SDS page gel asymmetrically so that it is easier to distinguish which side of the gel you are looking at.] Rinse the gel holder (with gel in it). This removes any carcinogenic non-polymerized acrylamide. Prepare 1x transfer buffer from the 10x stock. Dilute with distilled water. Cut out nitrocellulose paper using the transfer blot papers to measure the size. Create a pool of

1X transfer buffer and soak two black gauze pads on both sides. While still working in the transfer buffer, put one black gauze rectangle on each side of the holder (clear side and black side). Put black side in the buffer. Soak two pieces of trans blot paper in the transfer buffer on both sides and then place them on each black gauze. Using a plastic wedge open the SDS page gel glass casing carefully (making sure not to rupture the gel). While still working in

135

the transfer buffer place the gel in the buffer and then in the trans blot paper that is on the

BLACK side of the holder. The gel should remain in the same orientation as it was when running in SDS-page. Remove the protective paper on each side of the nitrocellulose membrane and soak the membrane in the transfer buffer and place the membrane on top of the gel. Press gently on the membrane to remove any bubbles that may be between it and the SDS-page polyacrylamide gel. Close the sandwich to ensure that every part is perfectly stacked on the next and then lock it using the white clip attachment. The white piece should face the red side of the chamber (the positive electrode). Orient the sandwich in the charge casing so that the red (+) attracts the proteins (-) from the gel onto the membrane. match the black side of the holder closer to the black electrode (-). Pour transfer buffer to fill the apparatus. Use ice or to cool in the tub. Replace the ice pack roughly 45 minutes into the run.

Run for 70 volts for 1.5 hrs. in TRANSFER BUFFER 1x. Can store nitrocellulose in Transfer

Buffer at 4C.

Staining a Western Blot Equipment ● petri dish ● 1x TBS ● 1x TBST (tbs with 0.1% tween) ● 2% non-fat milk powder TBS solution (~5ml mark powder and rest TBS in 50ml conical tube) ● primary antibody and secondary antibody

Cut the edges of the nitrocellulose with transferred ladder and bands (unseen). Careful to not touch the bands with the scissors. Place the 5% non-fat milk powder (2.5g in 50ml)

TBS1x as a blocking agent. Replace TBS in petri dish with 10ml (enough to cover the nitrocellulose paper) of this and let rotate for 15 minutes to coat the protein surface. Prepare

5% TBS/milk solution with primary antibody. IF this is just a test, use older mixtures of this.

136

Make a 1:2,000 dilution of the antibody. So in 10 ml, 5ul-3 of antibody.

● SM1 (yellow) for GC in NBA Ab box (18-22kDa) ○ Mouse = 2nd antibody (brown) alpha mouse HRP ● A2 for elongata in NBA pillin elongata box ○ secondary is rabbit and is called alpha R HRP and is stored in Glycerol ● Ef-Tu antibody (in A fridge in red box) HM6010 from hycult biotech ○ Ef-TU is mouse, so antibody is anti-mouse (~43kDa)

Replace the original tbs/milk solution with this in each petri dish. Let it sit overnight in fridge

(4C) on rotator. COVER so that the liquid does not evaporate. Wash the nitrocellulose with

TBST (TBS1x with 0.5% tween 3 times for 5 minutes each. No need to remove previous antibody solution. After the last wash prepare the secondary antibody solution with TBST in the same way (carefully label and confirm that you are using the correct antibody). (1:2000 dilution, 5ul) Again let this sit for 1.5hrs. Wash the sample with TBST twice and for the last time only with 1x TBS for 5 minutes at a time. Can store nitrocellulose after staining in 1X

TBS at 4C. Image with LAS-4000 (let cool to -30C)

4-6.4: Coomassie Stain

Run normalized samples on SDS-PAGE gel prepared as instructed above. After electroporation is complete, stain gel in Coomassie Brilliant Blue R-250 overnight at room temperature on rotation. Destain and rinse with ddH2O 3 times and image using digital camera.

137

4-6.5: Microplate squeezing experiments

Capillary-like tubes were used as the microplate instrument. A thin microplate edge was created using a needle puller (Narishige company). Microplate needles are melted onto another capillary tube to form a ‘hockey-stick’ like shape. The hockey stick microplate was placed under the microscope attached to a micromanipulator (Sutter Instrument, MP-285).

Separately, a plastic rectangular liquid containment chamber was adhered using melted parafilm to a glass coverslip (24x60mm, # 1.5). A 1cm x 0.5cm piece of cured PDMS was adhered to coverslip inside containment chamber. This piece of PDMS is where the squeezing of the microcolony occurred, between the PDMS and the microplate. Containment structure was placed on the microscope (Olympus IX70) stage and hockey stick microplate was lower into the containment chamber. Images were taken on 40X UPlanFI Olympus objective and captured by Micro-Manager 1.2 Software powered by ImageJ.

138

CHAPTER 5

Conclusions and future directions

139

5-1: CONCLUSIONS

Neisseria gonorrhoeae is the causative agent of the sexually transmitted infection, gonorrhea. Neisseria gonorrhoeae form multicellular aggregates called microcolonies on human epithelial cells. Microcolonies are composed of single cells binding together within an hour of contact with other cells (Higashi et al. 2007). It is thought that microcolonies are the preferred structure of Niesseria gonorrhoeae, as opposed to large biofilms or single cells.

My thesis work examined the physiology of Neisseria gonorrhoeae microcolonies.

Specifically, I investigated the relationship between microcolony physiology of Neisseria gonorrhoeae and its dynamic type IV pilus (tfp).

My research was initiated at a time when there was little literature available about

Neisseria gonorrhoeae microcolony physiology. Additionally, there is momentum developing in tfp research in Neisseria gonorrhoeae and other microbes because of tfp’s ubiquitous evolutionary nature (Berry and Pelicic 2015). Therefore, I seized the opportunity to connect the role of tfp in microcolony formation and how it may influence microcolony physiology.

From other work studying Neisseria gonorrhoeae tfp, there is evidence that tfp acts as a mechanosensitive signal transducer; therefore, it is critical to understand the role of force separately from the chemical composition of tfp. I utilized a tfp retraction-deficient mutant,

ΔpilT, to understand how force is connected to physiology.

The research surrounding cellular development is dominated by a biochemical perspective. At all hierarchical levels of life, biochemical processes, like metabolism and information transfer, are certainly critical. Yet, there is growing evidence that physical forces may provide a deeper understanding of well characterized biochemical mechanisms.

140

Chapter 1 provides an overview of Neisseria gonorrhoeae and tfp physiology; plus, a foundation of microbial multicellular communities with an emphasis on the emerging roles of physical forces on pattern formation. This introduction demonstrates the breadth and depth required to study mechanomicrobiology.

In Chapter 2 is a transcription of my first manuscript, “Pili mediated intercellular forces shape heterogeneous bacterial microcolonies prior to multicellular differentiation”.

In this work, I collaborated with others to quantify differential behaviors observed in microcolonies. My collaborator measured single cell motility within the microcolony and measured differential motility based on location in the microcolony. I performed an additional experiment to test whether retraction force in pili-pili interactions affected geography of a single cell. By mixing retraction-deficient mutants with Neisseria gonorrhoeae WT cells, we saw that retraction-deficient cells are unable to incorporate themselves into the microcolony; therefore, remaining at the perimeter. This suggests that pilus retraction force is a key element in the composition of early Neisseria gonorrhoeae microcolonies.

To extend this thinking, I was motivated to investigate more about the physiology of microcolony architecture, in order to understand the influence of tfp retraction force. In

Chapter 3 I used a quantitative approach to analyze seven pilus-related genes using fluorescent reporters. Using epifluorescence and confocal microscopy, I quantified fluorescence intensity over time in microcolonies. Microcolonies demonstrate spatiotemporal organization due to the differential patterning observed of the pilus-related genes. Most critically, spatiotemporal patterning in retraction-deficient ∆pilT background is downregulated in individual microcolonies in comparison to WT. I provided evidence that

141

physical force by tfp retraction and artificial forces can disrupt the WT gene expression.

These experiments suggest an important role of physical forces in establishing and maintaining spatiotemporal patterning during microcolony development as well defining long-term biological fate. The next steps of this work would be to take a deep dive into metabolism of the microcolony, specifically, oxygen usage, catalase activity and redox reactions (Wessel et al. 2014). Redox reactions, such as denitrification and iron oxide formation, will be important to analyze specifically in Neisseria gonorrhoeae, in particular when Neisseria gonorrhoeae microcolonies mature to form larger scaled biofilms.

Lastly, in Chapter 4, I provide a novel physical characterization study of Neisseria gonorrhoeae microcolonies. I was interested in studying the physiology of cells in their native context, which motivated the work in this chapter. Since the native context of Neisseria gonorrhoeae is the mucosal epithelial lining of human cells, I measured the physiology of

Neisseria gonorrhoeae under a range of osmotic pressures resembling typical mucosal pressures secreted by human epithelial cells. I measured the formation size and survival rates of microcolonies exposed to a range of osmotic pressures. I conclude that WT microcolony development is sensitive to pressure; however, WT can survive under a wide range (100Pa-1000Pa). The retraction-deficient mutant, ∆pilT, however, does not demonstrate the same sensitivity to osmotic pressure as WT. The null effect seen in ∆pilT provides more evidence of the mechanosensitive pathways must play in signaling.

Alternatively, it is very well possible that my osmotic pressure study was not measuring the effects of pressure, but the effects of pilE (tfp main subunit) production when exposed to dextran. From western blots, I measured a sharp decrease of PilE in WT as pressure increased while observing a null effect in ∆pilT. Lastly, I measured the plasticity of WT and

142

∆pilT microcolonies via a de novo instrument, a microplate apparatus, to squeeze live microcolonies. Microplate squeezing was recording using time lapse light microscopy. This experiment shows that WT microcolonies are plastic, but not nearly as plastic as ∆pilT because the microcolony reformation was nearly instantaneous after deformation.

The overall aim of this work is to begin to understand the role of physical forces during different life stages of microbes. In this thesis, I determined that type IV pilus retraction forces within Neisseria gonorrhoeae influence microcolony formation at single cell resolution. Characterizing gene expression in nascent microcolonies provided key evidence for spatiotemporal heterogeneity in developing microcolonies, which has not been demonstrated in any bacterium previously. In addition, I continued a line of research developed in the lab by connecting pilus retraction forces with organization development in early biofilm formation.

Based on my thesis focus, I wish to recommend a change of perspective concerning how bacteria are viewed. I believe considering bacteria as a multicellular organism, as opposed to an exclusive single cell, is an important perspective shift for researchers. I do believe there is a slow shift beginning in the microbiology field to consider bacteria as a multicellular structure. We need to understand more about biofilm development in its early stages because I believe variability is induced early in biofilm formation. As a result, the broad future directions of this work are to understand more about the variability between individual microcolonies. My work demonstrated the inherent variability of gene expression in nascent microcolonies, so understanding how variability is induced (or how homogeneity is maintained, perhaps) will lead to deeper knowledge of biofilms, which is crucial for human pathogenic bacteria.

143

144

Bibliography

Aagaard, Kjersti, Kevin Riehle, Jun Ma, Nicola Segata, Toni-Ann Mistretta, Cristian Coarfa, Sabeen Raza, et al. 2012. “A Metagenomic Approach to Characterization of the Vaginal Microbiome Signature in Pregnancy.” Edited by Adam J. Ratner. PLoS ONE 7 (6): e36466. https://doi.org/10.1371/journal.pone.0036466.

“Antibiotic / Antimicrobial Resistance | CDC.” n.d. Accessed August 13, 2019. https://www.cdc.gov/drugresistance/index.html.

Baker, Joseph L., Nicolas Biais, and Florence Tama. 2013. “Steered Molecular Dynamics Simulations of a Type IV Pilus Probe Initial Stages of a Force-Induced Conformational Transition.” Edited by Michael Nilges. PLoS Computational Biology 9 (4): e1003032. https://doi.org/10.1371/journal.pcbi.1003032.

Beatson, Scott A, Cynthia B Whitchurch, Jennifer L Sargent, Roger C Levesque, and John S Mattick. 2002. “Differential Regulation of Twitching Motility and Elastase Production by Vfr in Pseudomonas Aeruginosa.” Journal of Bacteriology 184 (13): 3605–13. https://doi.org/10.1128/jb.184.13.3605-3613.2002.

Beilharz, Katrin, Renske van Raaphorst, Morten Kjos, and Jan-Willem Veening. 2015. “Red Fluorescent Proteins for Gene Expression and Protein Localization Studies in Streptococcus Pneumoniae and Efficient Transformation with DNA Assembled via the Gibson Assembly Method.” Applied and Environmental Microbiology 81 (20): 7244–52. https://doi.org/10.1128/AEM.02033-15.

Bejsovec, Amy. Wieschaus, Eric. 1993. “Segment Polarity Gene Interactions Modulate Epidermal Patterning in Drosophila Embroys.” Development 119: 501–17. https://dev.biologists.org/content/develop/119/2/501.full.pdf.

Belas, Robert. 2014. “Biofilms, Flagella, and Mechanosensing of Surfaces by Bacteria.” Trends in Microbiology 22 (9): 517–27. https://doi.org/10.1016/j.tim.2014.05.002.

Berk, V., J. C. N. Fong, G. T. Dempsey, O. N. Develioglu, X. Zhuang, J. Liphardt, F. H. Yildiz, and S. Chu. 2012. “Molecular Architecture and Assembly Principles of Vibrio Cholerae Biofilms.” Science 337 (6091): 236–39. https://doi.org/10.1126/science.1222981.

Berry, Jamie-Lee, and Vladimir Pelicic. 2015. “Exceptionally Widespread Nanomachines Composed of Type IV Pilins: The Prokaryotic Swiss Army Knives.” FEMS Microbiology Reviews 39 (1): 134–54. https://doi.org/10.1093/femsre/fuu001.

Bervoets, Indra, and Daniel Charlier. 2019. “Diversity, Versatility and Complexity of

145

Bacterial Gene Regulation Mechanisms: Opportunities and Drawbacks for Applications in Synthetic Biology.” FEMS Microbiology Reviews 43 (3): 304–39. https://doi.org/10.1093/femsre/fuz001.

Biais, Nicolas, Dustin L Higashi, Jasna Brujic, Magdalene So, and Michael P Sheetz. 2010. “Force-Dependent Polymorphism in Type IV Pili Reveals Hidden Epitopes.” Proceedings of the National Academy of Sciences of the United States of America 107 (25): 11358–63. https://doi.org/10.1073/pnas.0911328107.

Biais, Nicolas, Benoît Ladoux, Dustin Higashi, Magdalene So, and Michael Sheetz. 2008. “Cooperative Retraction of Bundled Type IV Pili Enables Nanonewton Force Generation.” Edited by Manfred Schliwa. PLoS Biology 6 (4): e87. https://doi.org/10.1371/journal.pbio.0060087.

Blair, D. & Dufresne, E. 2017. “The {M}atlab {P}article {T}racking {C}ode {R}epository.”

Bonazzi, Daria, Valentina Lo Schiavo, Silke Machata, Ilyas Djafer-Cherif, Pierre Nivoit, Valeria Manriquez, Hirokazu Tanimoto, et al. 2018. “Intermittent Pili-Mediated Forces Fluidize Neisseria Meningitidis Aggregates Promoting Vascular Colonization.” Cell 174 (1): 143-155.e16. https://doi.org/10.1016/j.cell.2018.04.010.

BOVRE, K., and E. HOLTEN. 1970. “Neisseria Elongata Sp.Nov., a Rod-Shaped Member of the Genus Neisseria. Re-Evaluation of Cell Shape as a Criterion in Classification.” Journal of General Microbiology 60 (1): 67–75. https://doi.org/10.1099/00221287-60-1-67.

Brown, Daniel R, Sophie Helaine, Etienne Carbonnelle, and Vladimir Pelicic. 2010. “Systematic Functional Analysis Reveals That a Set of Seven Genes Is Involved in Fine- Tuning of the Multiple Functions Mediated by Type IV Pili in Neisseria Meningitidis.” Infection and Immunity 78 (7): 3053–63. https://doi.org/10.1128/IAI.00099-10.

Burger, W. & Burge, M. J. 2009. Principles of Digital Image Processing. Springer.

Campos-Ortega, José A., and Volker Hartenstein. 1985. “The Pattern of Embryonic Cell Divisions.” In The Embryonic Development of Drosophila Melanogaster, 165–71. Berlin, Heidelberg: Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-662-02454- 6_5.

Cantarel, Brandi L., Vincent Lombard, and Bernard Henrissat. 2012. “Complex Carbohydrate Utilization by the Healthy Human Microbiome.” Edited by Vasu D. Appanna. PLoS ONE 7 (6): e28742. https://doi.org/10.1371/journal.pone.0028742.

Chimileski, Scott, Michael J Franklin, and R Thane Papke. 2014. “Biofilms Formed by the Archaeon Haloferax Volcaniiexhibit Cellular Differentiation and Social Motility, and

146

Facilitate Horizontal Gene Transfer.” BMC Biology 12 (1): 65. https://doi.org/10.1186/s12915-014-0065-5.

Co, Julia Y., Gerardo Cárcamo-Oyarce, Nicole Billings, Kelsey M. Wheeler, Scott C. Grindy, Niels Holten-Andersen, and Katharina Ribbeck. 2018. “Mucins Trigger Dispersal of Pseudomonas Aeruginosa Biofilms.” Npj Biofilms and Microbiomes 4 (1): 23. https://doi.org/10.1038/s41522-018-0067-0.

Cohen, M. S., J. G. Cannon, A. E. Jerse, L. M. Charniga, S. F. Isbey, and L. G. Whicker. 1994. “Human Experimentation with Neisseria Gonorrhoeae: Rationale, Methods, and Implications for the Biology of Infection and Vaccine Development.” Journal of Infectious Diseases 169 (3): 532–37. https://doi.org/10.1093/infdis/169.3.532.

Cooper, Geoffrey M. 2000. The Cell : A Molecular Approach. ASM Press.

Costerton, J. W., P S Stewart, and E P Greenberg. 1999. “Bacterial Biofilms: A Common Cause of Persistent Infections.” Science 284 (5418): 1318–22. https://doi.org/10.1126/science.284.5418.1318.

Costerton, J W, K J Cheng, G G Geesey, T I Ladd, J C Nickel, M Dasgupta, and T J Marrie. 1987. “Bacterial Biofilms in Nature and Disease.” Annual Review of Microbiology 41 (1): 435– 64. https://doi.org/10.1146/annurev.mi.41.100187.002251.

Costerton, J W, Z Lewandowski, D E Caldwell, D R Korber, and H M Lappin-Scott. 1995. “Microbial Biofilms.” Annual Review of Microbiology 49 (1): 711–45. https://doi.org/10.1146/annurev.mi.49.100195.003431.

Craig, Lisa, Katrina T. Forest, and Berenike Maier. 2019. “Type IV Pili: Dynamics, Biophysics and Functional Consequences.” Nature Reviews Microbiology 17 (7): 429–40. https://doi.org/10.1038/s41579-019-0195-4.

Craig, Lisa, Michael E. Pique, and John A. Tainer. 2004. “Type IV Pilus Structure and Bacterial Pathogenicity.” Nature Reviews Microbiology 2 (5): 363–78. https://doi.org/10.1038/nrmicro885.

Cui, Yidan, Feroz M. Hameed, Bo Yang, Kyunghee Lee, Catherine Qiurong Pan, Sungsu Park, and Michael Sheetz. 2015. “Cyclic Stretching of Soft Substrates Induces Spreading and Growth.” Nature Communications 6 (1): 6333. https://doi.org/10.1038/ncomms7333. d’Enfert, C, A Ryter, and A P Pugsley. 1987. “Cloning and Expression in Escherichia Coli of the Klebsiella Pneumoniae Genes for Production, Surface Localization and Secretion of the Lipoprotein Pullulanase.” The EMBO Journal 6 (11): 3531–38. http://www.ncbi.nlm.nih.gov/pubmed/3322811.

147

Davey, M E, and G A O’toole. 2000. “Microbial Biofilms: From Ecology to Molecular Genetics.” Microbiology and Molecular Biology Reviews : MMBR 64 (4): 847–67. https://doi.org/10.1128/mmbr.64.4.847-867.2000.

Davidson, Lance, Michelangelo von Dassow, and Jian Zhou. 2009. “Multi-Scale Mechanics from Molecules to Morphogenesis.” The International Journal of Biochemistry & Cell Biology 41 (11): 2147–62. https://doi.org/10.1016/j.biocel.2009.04.015.

Davies, Julian, and Dorothy Davies. 2010. “Origins and Evolution of Antibiotic Resistance.” Microbiology and Molecular Biology Reviews : MMBR 74 (3): 417–33. https://doi.org/10.1128/MMBR.00016-10.

Davis, G S, H M Phillips, and M S Steinberg. 1997. “Germ-Layer Surface Tensions and "Tissue Affinities" in Rana Pipiens Gastrulae: Quantitative Measurements.” Developmental Biology 192 (2): 630–44. https://doi.org/10.1006/dbio.1997.8741.

DeSalle, Rob, and Patricia Wynne. n.d. Our Senses : An Immersive Experience. Accessed August 13, 2019. https://books.google.com/books?id=vDlFDwAAQBAJ&pg=PA8&lpg=PA8&dq=rob+des alle+bacteria+location&source=bl&ots=s3InuJ5OAG&sig=ACfU3U0hI9tDZ9YHGJzgQLY bHmrELUUm_Q&hl=en&sa=X&ved=2ahUKEwiMpq_m6YDkAhUmm- AKHWNiCPkQ6AEwFnoECAoQAQ#v=onepage&q=rob desalle bacteria location&f=false.

Desprat, Nicolas, Willy Supatto, Philippe-Alexandre Pouille, Emmanuel Beaurepaire, and Emmanuel Farge. 2008. “Tissue Deformation Modulates Twist Expression to Determine Anterior Midgut Differentiation in Drosophila Embryos.” Developmental Cell 15 (3): 470–77. https://doi.org/10.1016/j.devcel.2008.07.009.

Dietrich, Manuela, Sina Bartfeld, Rebekka Munke, Claudia Lange, Lesley A. Ogilvie, Alexandra Friedrich, and Thomas F. Meyer. 2011. “Activation of NF-ΚB by Neisseria Gonorrhoeae Is Associated with Microcolony Formation and Type IV Pilus Retraction.” Cellular Microbiology 13 (8): 1168–82. https://doi.org/10.1111/j.1462- 5822.2011.01607.x.

Dietrich, Manuela, Hans Mollenkopf, Magdalene “Maggie” So, and Alexandra Friedrich. 2009. “Pilin Regulation in the PilT Mutant of Neisseria Gonorrhoeae Strain MS11.” FEMS Microbiology Letters 296 (2): 248–56. https://doi.org/10.1111/J.1574- 6968.2009.01647.X.

Dillard, Joseph P. 2011. “Genetic Manipulation of Neisseria Gonorrhoeae.” Current Protocols in Microbiology Chapter 4 (November): Unit4A.2. https://doi.org/10.1002/9780471729259.mc04a02s23.

148

Donlan, R.M., and J.W. Costerton. 2002. “Biofilms: Survivalmechanisms of Clinically Relevant Microorganisms.” Clin.Microbiol. Rev. 15 (2): 167–19. https://doi.org/10.1128/CMR.15.2.167.

Donlan, Rodney M. 2002. “Biofilms: Microbial Life on Surfaces.” Emerging Infectious Diseases 8 (9): 881–90. https://doi.org/10.3201/eid0809.020063.

Donlan, Rodney M, and J William Costerton. 2002. “Biofilms: Survival Mechanisms of Clinically Relevant Microorganisms.” Clinical Microbiology Reviews 15 (2): 167–93. https://doi.org/10.1128/cmr.15.2.167-193.2002.

Dowling, R B, and R Wilson. 1998. “Bacterial Toxins Which Perturb Ciliary Function and Respiratory Epithelium.” Journal of Applied Microbiology. Vol. 85. https://onlinelibrary.wiley.com/doi/pdf/10.1046/j.1365-2672.1998.0840s1138S.x.

Drescher, Knut, Jörn Dunkel, Carey D. Nadell, Sven van Teeffelen, Ivan Grnja, Ned S. Wingreen, Howard A. Stone, and Bonnie L. Bassler. 2016. “Architectural Transitions in Vibrio Cholerae Biofilms at Single-Cell Resolution.” Proceedings of the National Academy of Sciences 113 (14): E2066–72. https://doi.org/10.1073/pnas.1601702113.

Duncan, Margaret J. 2003. “Genomics of Oral Bacteria,” 175–87. www.hgsc.tmc.edu/.

Economou, Andrew D, Atsushi Ohazama, Thantrira Porntaveetus, Paul T Sharpe, Shigeru Kondo, M Albert Basson, Amel Gritli-Linde, Martyn T Cobourne, and Jeremy B A Green. 2012. “Periodic Stripe Formation by a Turing Mechanism Operating at Growth Zones in the Mammalian Palate.” Nature Genetics 44 (3): 348–51. https://doi.org/10.1038/ng.1090.

Edwards, J L, J Q Shao, K A Ault, and M A Apicella. 2000. “Neisseria Gonorrhoeae Elicits Membrane Ruffling and Cytoskeletal Rearrangements upon Infection of Primary Human Endocervical and Ectocervical Cells.” Infection and Immunity 68 (9): 5354–63. https://doi.org/10.1128/iai.68.9.5354-5363.2000.

Edwards, Jennifer L, and Michael A Apicella. 2004. “The Molecular Mechanisms Used by Neisseria Gonorrhoeae to Initiate Infection Differ between Men and Women.” Clinical Microbiology Reviews 17 (4): 965–81, table of contents. https://doi.org/10.1128/CMR.17.4.965-981.2004.

Engler, Adam J., Shamik Sen, H. Lee Sweeney, and Dennis E. Discher. 2006. “Matrix Elasticity Directs Stem Cell Lineage Specification.” Cell 126 (4): 677–89. https://doi.org/10.1016/j.cell.2006.06.044.

Eriksson, Jens. 2012. AVHANDLINGAR.SE: Virulence Factors and Motility Mechanisms of

149

Pathogenic Neisseria. http://www.avhandlingar.se/avhandling/f7535faa71/.

Eriksson, Jens, Olaspers Sara Eriksson, and Ann-Beth Jonsson. 2012. “Loss of Meningococcal PilU Delays Microcolony Formation and Attenuates Virulence In Vivo.” Edited by B. A. McCormick. Infection and Immunity 80 (7): 2538–47. https://doi.org/10.1128/IAI.06354-11.

Farge, Emmanuel. 2003. “Mechanical Induction of Twist in the Drosophila Foregut/Stomodeal Primordium.” Current Biology : CB 13 (16): 1365–77. https://doi.org/10.1016/s0960-9822(03)00576-1.

Feynman, Richard P. n.d. “There’s Plenty of Room at the Bottom.” Accessed August 13, 2019. http://www.zyvex.com/nanotech/feynman.htmlwith.

Fliegauf, Manfred, Andreas F-P Sonnen, Bernhard Kremer, and Philipp Henneke. 2013. “Mucociliary Clearance Defects in a Murine in Vitro Model of Pneumococcal Airway Infection.” PloS One 8 (3): e59925. https://doi.org/10.1371/journal.pone.0059925.

Ford, BJ. 1993. “The Leeuwenhoek Legacy.” https://philpapers.org/rec/FORTLL.

Gest, H. 2004. “The Discovery of Microorganisms by Robert Hooke and Antoni van Leeuwenhoek, Fellows of The Royal Society.” Notes and Records of the Royal Society of London 58 (2): 187–201. https://doi.org/10.1098/rsnr.2004.0055.

Giltner, Carmen L., Ylan Nguyen, and Lori L. Burrows. 2012. “Type IV Pilin Proteins: Versatile Molecular Modules.” Microbiology and Molecular Biology Reviews 76 (4): 740–72. https://doi.org/10.1128/MMBR.00035-12.

“Gonorrhea - STD Information from CDC.” n.d. Accessed August 13, 2019. https://www.cdc.gov/std/gonorrhea/default.htm.

Goosens, Vivianne J., Andreas Busch, Michaella Georgiadou, Marta Castagnini, Katrina T. Forest, Gabriel Waksman, and Vladimir Pelicic. 2017. “Reconstitution of a Minimal Machinery Capable of Assembling Periplasmic Type IV Pili.” Proceedings of the National Academy of Sciences 114 (25): E4978–86. https://doi.org/10.1073/PNAS.1618539114.

Green, Jeremy B A, and James Sharpe. 2015. “Positional Information and Reaction- Diffusion: Two Big Ideas in Developmental Biology Combine.” Development (Cambridge, England) 142 (7): 1203–11. https://doi.org/10.1242/dev.114991.

Gurung, Ishwori, Ingrid Spielman, Mark R. Davies, Rajan Lala, Peter Gaustad, Nicolas Biais,

150

and Vladimir Pelicic. 2016. “Functional Analysis of an Unusual Type IV Pilus in the Gram-Positive Streptococcus Sanguinis.” Molecular Microbiology 99 (2): 380–92. https://doi.org/10.1111/mmi.13237.

Haaber, Jakob, Marianne Thorup Cohn, Dorte Frees, Thorbjørn Joest Andersen, and Hanne Ingmer. 2012. “Planktonic Aggregates of Staphylococcus Aureus Protect against Common Antibiotics.” Edited by Michael Otto. PLoS ONE 7 (7): e41075. https://doi.org/10.1371/journal.pone.0041075.

Higashi, Dustin L, Shaun W Lee, Aurelie Snyder, Nathan J Weyand, Antony Bakke, and Magdalene So. 2007. “Dynamics of Neisseria Gonorrhoeae Attachment: Microcolony Development, Cortical Plaque Formation, and Cytoprotection.” Infection and Immunity 75 (10): 4743–53. https://doi.org/10.1128/IAI.00687-07.

Hill, S A, D S Samuels, J H Carlson, J Wilson, D Hogan, L Lubke, and R J Belland. 1997. “Integration Host Factor Is a Transcriptional Cofactor of PilE in Neisseria Gonorrhoeae.” Molecular Microbiology 23 (4): 649–56. http://www.ncbi.nlm.nih.gov/pubmed/9157237.

Hill, Stuart A, and John K Davies. 2009. “Pilin Gene Variation in Neisseria Gonorrhoeae: Reassessing the Old Paradigms.” FEMS Microbiology Reviews 33 (3): 521–30. https://doi.org/10.1111/j.1574-6976.2009.00171.x.

Hollingsworth, Michael A., and Benjamin J. Swanson. 2004. “Mucins in Cancer: Protection and Control of the Cell Surface.” Nature Reviews Cancer 4 (1): 45–60. https://doi.org/10.1038/nrc1251.

Howard, Jonathon, Stephan W. Grill, and Justin S. Bois. 2011a. “Turing’s next Steps: The Mechanochemical Basis of Morphogenesis.” Nature Reviews Molecular Cell Biology 12 (6): 392–98. https://doi.org/10.1038/nrm3120.

Howard, Jonathon, Stephan W Grill, and Justin S Bois. 2011b. “Turing’s next Steps: The Mechanochemical Basis of Morphogenesis.” https://doi.org/10.1038/nrm3120.

Howie, Heather L, Michael Glogauer, and Magdalene So. 2005. “The N. Gonorrhoeae Type IV Pilus Stimulates Mechanosensitive Pathways and Cytoprotection through a PilT- Dependent Mechanism.” PLoS Biology 3 (4): e100. https://doi.org/10.1371/journal.pbio.0030100.

Hug, Laura A., Brett J. Baker, Karthik Anantharaman, Christopher T. Brown, Alexander J. Probst, Cindy J. Castelle, Cristina N. Butterfield, et al. 2016. “A New View of the Tree of Life.” Nature Microbiology 1 (5): 16048. https://doi.org/10.1038/nmicrobiol.2016.48.

151

Hung, Miao-Chiu, and Myron Christodoulides. 2013. “The Biology of Neisseria Adhesins.” Biology 2 (3): 1054–1109. https://doi.org/10.3390/biology2031054.

Hunter, Philip. 2008. “The Mob Response. The Importance of Biofilm Research for Combating Chronic Diseases and Tackling Contamination.” EMBO Reports 9 (4): 314– 17. https://doi.org/10.1038/embor.2008.43.

JACOB, F, and J MONOD. 1959. “[Genes of Structure and Genes of Regulation in the Biosynthesis of Proteins].” Comptes Rendus Hebdomadaires Des Seances de l’Academie Des Sciences 249 (October): 1282–84. http://www.ncbi.nlm.nih.gov/pubmed/14406327.

Jarrell, Ken, Yan Ding, Divya Nair, and Sarah Siu. 2013. “Surface Appendages of Archaea: Structure, Function, Genetics and Assembly.” Life 3 (1): 86–117. https://doi.org/10.3390/life3010086.

Jones, R T, and R S Talley. 1977. “Simplified Complete Medium for the Growth of Neisseria Gonorrhoeae.” Journal of Clinical Microbiology 5 (1): 9–14. https://jcm.asm.org/content/5/1/9?ijkey=af21ffb551eaa50a2b7b4e3dab82c7f111a8 2e94&keytype2=tf_ipsecsha.

Jonsson, a. B., M. Rahman, and S. Normark. 1995. “Pilus Biogenesis Gene, PilC, of Neisseria Gonorrhoeae: PilC1 and PilC2 Are Each Part of a Larger Duplication of the Gonococcal Genome and Share Upstream and Downstream Homologous Sequences with Opa and Pil Loci.” Microbiology 141 (10): 2367–77. https://doi.org/10.1099/13500872-141- 10-2367.

Klausen, Mikkel, Anders Aaes-Jørgensen, Søren Molin, and Tim Tolker-Nielsen. 2003. “Involvement of Bacterial Migration in the Development of Complex Multicellular Structures in Pseudomonas Aeruginosa Biofilms.” Molecular Microbiology 50 (1): 61– 68. https://doi.org/10.1046/j.1365-2958.2003.03677.x.

Kouzel, Nadzeya, Enno R. Oldewurtel, and Berenike Maier. 2015. “Gene Transfer Efficiency in Gonococcal Biofilms: Role of Biofilm Age, and Architecture, and Pilin Antigenic Variation.” Journal of Bacteriology, no. May: JB.00171-15. https://doi.org/10.1128/JB.00171-15.

Kurre, Rainer, Andrea Höne, Martin Clausen, Claudia Meel, and Berenike Maier. 2012. “PilT2 Enhances the Speed of Gonococcal Type IV Pilus Retraction and of Twitching Motility.” Molecular Microbiology 86 (4): 857–65. https://doi.org/10.1111/mmi.12022.

Lai, Samuel K., Ying-Ying Wang, Denis Wirtz, and Justin Hanes. 2009. “Micro- and

152

Macrorheology of Mucus.” Advanced Drug Delivery Reviews 61 (2): 86. https://doi.org/10.1016/J.ADDR.2008.09.012.

Lane, Nick. 2015. “The Unseen World: Reflections on Leeuwenhoek (1677) ‘Concerning Little Animals’.” Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 370 (1666). https://doi.org/10.1098/rstb.2014.0344.

Lang, Tiange, Sofia Klasson, Erik Larsson, Malin E V Johansson, Gunnar C Hansson, and Tore Samuelsson. 2016. “Searching the Evolutionary Origin of Epithelial Mucus Protein Components-Mucins and FCGBP.” Molecular Biology and Evolution 33 (8): 1921–36. https://doi.org/10.1093/molbev/msw066.

Le, Huy Quang, Sushmita Ghatak, Ching-Yan Chloé Yeung, Frederik Tellkamp, Christian Günschmann, Christoph Dieterich, Assa Yeroslaviz, et al. 2016. “Mechanical Regulation of Transcription Controls Polycomb-Mediated Gene Silencing during Lineage Commitment.” Nature Cell Biology 18 (8): 864–75. https://doi.org/10.1038/ncb3387.

Lee, Shaun W., Dustin L. Higashi, Aurelie Snyder, Alexey J. Merz, Laura Potter, and Magdalene So. 2005. “PilT Is Required for PI(3,4,5)P3-Mediated Crosstalk between Neisseria Gonorrhoeae and Epithelial Cells.” Cellular Microbiology 7 (9): 1271–84. https://doi.org/10.1111/j.1462-5822.2005.00551.x.

Leeuwenhoek, Antonie. n.d. “Antonie van Leeuwenhoek and His Little Animals.” New York, NY.

Lewis, K. 2001. “Riddle of Biofilm Resistance.” Antimicrobial Agents and Chemotherapy 45 (4): 999–1007. https://doi.org/10.1128/AAC.45.4.999-1007.2001.

Li, Juliana, Mindy S. Lim, Sheng Li, Melissa Brock, Michael E. Pique, Virgil L. Woods, and Lisa Craig. 2008. “Vibrio Cholerae Toxin-Coregulated Pilus Structure Analyzed by Hydrogen/Deuterium Exchange Mass Spectrometry.” Structure 16 (1): 137–48. https://doi.org/10.1016/j.str.2007.10.027.

Lyons, Nicholas A, and Roberto Kolter. 2015. “On the Evolution of Bacterial Multicellularity.” Current Opinion in Microbiology 24 (April): 21–28. https://doi.org/10.1016/j.mib.2014.12.007.

Maas, W K. 1991. “The Regulation of Arginine Biosynthesis: Its Contribution to Understanding the Control of Gene Expression.” Genetics 128 (3): 489–94. http://www.ncbi.nlm.nih.gov/pubmed/1874410.

Maier, Berenike, Laura Potter, Magdalene So, Cynthia D Long, Hank S Seifert, and Michael P

153

Sheetz. 2002. “Single Pilus Motor Forces Exceed 100 PN.” Proceedings of the National Academy of Sciences of the United States of America 99 (25): 16012–17. https://doi.org/10.1073/pnas.242523299.

Maitre, J.-L., H. Berthoumieux, S. F. G. Krens, G. Salbreux, F. Julicher, E. Paluch, and C.-P. Heisenberg. 2012. “Adhesion Functions in Cell Sorting by Mechanically Coupling the Cortices of Adhering Cells.” Science 338 (6104): 253–56. https://doi.org/10.1126/science.1225399.

Mann, Ethan E, and Daniel J Wozniak. 2012. “Pseudomonas Biofilm Matrix Composition and Niche Biology.” FEMS Microbiology Reviews 36 (4): 893–916. https://doi.org/10.1111/j.1574-6976.2011.00322.x.

Masters, Thao L., Shaun Wachter, Jenny Wachter, and Stuart A. Hill. 2016. “H-NS Suppresses PilE Intragenic Transcription and Antigenic Variation in Neisseria Gonorrhoeae.” Microbiology 162 (1): 177–90. https://doi.org/10.1099/mic.0.000199.

Mathews, D. H., M. D. Disney, J. L. Childs, S. J. Schroeder, M. Zuker, and D. H. Turner. 2004. “Incorporating Chemical Modification Constraints into a Dynamic Programming Algorithm for Prediction of RNA Secondary Structure.” Proceedings of the National Academy of Sciences 101 (19): 7287–92. https://doi.org/10.1073/pnas.0401799101.

McKenna, W R, P A Mickelsen, P F Sparling, and D W Dyer. 1988. “Iron Uptake from Lactoferrin and Transferrin by Neisseria Gonorrhoeae.” Infection and Immunity 56 (4): 785–91. http://www.ncbi.nlm.nih.gov/pubmed/3126143.

Merz, Alexey J., Magdalene So, and Michael P. Sheetz. 2000. “Pilus Retraction Powers Bacterial Twitching Motility.” Nature 407 (6800): 98–102. https://doi.org/10.1038/35024105.

Meyer, T. F., E. Billyard, R. Haas, S. Storzbach, and M. So. 1984. “Pilus Genes of Neisseria Gonorrheae: Chromosomal Organization and DNA Sequence.” Proceedings of the National Academy of Sciences 81 (19): 6110–14. https://doi.org/10.1073/pnas.81.19.6110.

Milo, Ron, and Rob Phillips. n.d. Cell Biology by the Numbers. Accessed August 13, 2019. https://books.google.com/books/about/Cell_Biology_by_the_Numbers.html?id=9NPR CgAAQBAJ.

Misic, Ana M, Kenneth A Satyshur, and Katrina T Forest. 2010. “P. Aeruginosa PilT Structures with and without Nucleotide Reveal a Dynamic Type IV Pilus Retraction Motor.” Journal of Molecular Biology 400 (5): 1011–21. https://doi.org/10.1016/j.jmb.2010.05.066.

154

Mitrossilis, Démosthène, Jens-Christian Röper, Damien Le Roy, Benjamin Driquez, Aude Michel, Christine Ménager, Gorky Shaw, et al. 2017. “Mechanotransductive Cascade of Myo-II-Dependent Mesoderm and Endoderm Invaginations in Embryo Gastrulation.” Nature Communications 8 (1): 13883. https://doi.org/10.1038/ncomms13883.

Montel, Fabien, Morgan Delarue, Jens Elgeti, Laurent Malaquin, Markus Basan, Thomas Risler, Bernard Cabane, et al. 2011. “Stress Clamp Experiments on Multicellular Tumor Spheroids.” Physical Review Letters 107 (18): 188102. https://doi.org/10.1103/PhysRevLett.107.188102.

Müller, Patrick, Katherine W Rogers, Shuizi R Yu, Michael Brand, and Alexander F Schier. 2013. “Morphogen Transport.” Development (Cambridge, England) 140 (8): 1621–38. https://doi.org/10.1242/dev.083519.

Muschiol, Sandra, Murat Balaban, Staffan Normark, and Birgitta Henriques-Normark. 2015. “Uptake of Extracellular DNA: Competence Induced Pili in Natural Transformation of Streptococcus Pneumoniae.” BioEssays 37 (4): 426–35. https://doi.org/10.1002/bies.201400125.

Nava, Michele M., Manuela T. Raimondi, and Riccardo Pietrabissa. 2012. “Controlling Self- Renewal and Differentiation of Stem Cells via Mechanical Cues.” Journal of Biomedicine and Biotechnology 2012: 1–12. https://doi.org/10.1155/2012/797410.

Ngampasutadol, Jutamas, Sanjay Ram, Sunita Gulati, Sarika Agarwal, Chongqing Li, Alberto Visintin, Brian Monks, Guillermo Madico, and Peter A. Rice. 2008. “Human Factor H Interacts Selectively with Neisseria Gonorrhoeae and Results in Species-Specific Complement Evasion.” The Journal of Immunology 180 (5): 3426–35. https://doi.org/10.4049/jimmunol.180.5.3426.

Nicolella, C, M C van Loosdrecht, and J J Heijnen. 2000. “Wastewater Treatment with Particulate Biofilm Reactors.” Journal of Biotechnology 80 (1): 1–33. http://www.ncbi.nlm.nih.gov/pubmed/10862983.

N�sslein-Volhard, C., E. Wieschaus, and H. Kluding. 1984. “Mutations Affecting the Pattern of the Larval Cuticle InDrosophila Melanogaster.” Wilhelm Roux’s Archives of Developmental Biology 193 (5): 267–82. https://doi.org/10.1007/BF00848156.

O’Toole, George A., and Roberto Kolter. 1998. “Flagellar and Twitching Motility Are Necessary for Pseudomonas Aeruginosa Biofilm Development.” Molecular Microbiology 30 (2): 295–304. https://doi.org/10.1046/j.1365-2958.1998.01062.x.

O’Toole, George, Heidi B. Kaplan, and Roberto Kolter. 2000. “Biofilm Formation as Microbial Development.” Annual Review of Microbiology 54 (1): 49–79.

155

https://doi.org/10.1146/annurev.micro.54.1.49.

Oldewurtel, Enno R, Nadzeya Kouzel, Lena Dewenter, Katja Henseler, and Berenike Maier. 2015. “Differential Interaction Forces Govern Bacterial Sorting in Early Biofilms.” ELife 4 (September). https://doi.org/10.7554/eLife.10811.

Oliveira, Nuno M., Esteban Martinez-Garcia, Joao Xavier, William M. Durham, Roberto Kolter, Wook Kim, Kevin R. Foster, and Kevin R Foster. 2015. “Biofilm Formation As a Response to Ecological Competition.” Edited by Michael T. Laub. PLOS Biology 13 (7): e1002191. https://doi.org/10.1371/journal.pbio.1002191.

Paluch, Ewa K., Celeste M. Nelson, Nicolas Biais, Ben Fabry, Jens Moeller, Beth L. Pruitt, Carina Wollnik, Galina Kudryasheva, Florian Rehfeldt, and Walter Federle. 2015. “Mechanotransduction: Use the Force(S).” BMC Biology 13 (1): 47. https://doi.org/10.1186/s12915-015-0150-4.

Paulsson, Johan. 2005. “Models of Stochastic Gene Expression.” Physics of Life Reviews 2 (2): 157–75. https://doi.org/10.1016/j.plrev.2005.03.003.

Pawlizak, Steve, Anatol W Fritsch, Steffen Grosser, Dave Ahrens, Tobias Thalheim, Stefanie Riedel, Tobias R Kießling, et al. 2015. “Testing the Differential Adhesion Hypothesis across the Epithelial−mesenchymal Transition.” New Journal of Physics 17 (8): 083049. https://doi.org/10.1088/1367-2630/17/8/083049.

Peabody, C. R., Yong Joon Chung, Ming-Ren Yen, Dominique Vidal-Ingigliardi, Anthony P. Pugsley, and Milton H. Saier Jr. 2003. “Type II Protein Secretion and Its Relationship to Bacterial Type IV Pili and Archaeal Flagella.” Microbiology 149 (11): 3051–72. https://doi.org/10.1099/mic.0.26364-0.

Persat, Alexandre. 2017. “Bacterial Mechanotransduction.” Current Opinion in Microbiology 36 (April): 1–6. https://doi.org/10.1016/J.MIB.2016.12.002.

Persat, Alexandre, Yuki F. Inclan, Joanne N. Engel, Howard A. Stone, and Zemer Gitai. 2015a. “Type IV Pili Mechanochemically Regulate Virulence Factors in Pseudomonas Aeruginosa.” Proceedings of the National Academy of Sciences 112 (24): 7563–68. https://doi.org/10.1073/pnas.1502025112.

Persat, Alexandre, Yuki F Inclan, Joanne N Engel, Howard A Stone, and Zemer Gitai. 2015b. “Type IV Pili Mechanochemically Regulate Virulence Factors in Pseudomonas Aeruginosa.” Proceedings of the National Academy of Sciences of the United States of America 112 (24): 7563–68. https://doi.org/10.1073/pnas.1502025112.

Persat, Alexandre, Carey D. Nadell, Minyoung Kevin Kim, Francois Ingremeau, Albert

156

Siryaporn, Knut Drescher, Ned S. Wingreen, Bonnie L. Bassler, Zemer Gitai, and Howard A. Stone. 2015. “The Mechanical World of Bacteria.” Cell 161 (5): 988–97. https://doi.org/10.1016/j.cell.2015.05.005.

Pönisch, W. & Zaburdaev. 2018. “Relative Distance between Tracers as a Measure of Diffusivity within Moving Aggregates.” Eur. Phys. J 91.

Pönisch, Wolfram, Kelly B. Eckenrode, Khaled Alzurqa, Hadi Nasrollahi, Christoph Weber, Vasily Zaburdaev, and Nicolas Biais. 2018a. “Pili Mediated Intercellular Forces Shape Heterogeneous Bacterial Microcolonies Prior to Multicellular Differentiation.” Scientific Reports 8 (1): 16567. https://doi.org/10.1038/s41598-018-34754-4.

———. 2018b. “Pili Mediated Intercellular Forces Shape Heterogeneous Bacterial Microcolonies Prior to Multicellular Differentiation.” Scientific Reports 8 (1): 16567. https://doi.org/10.1038/s41598-018-34754-4.

Pönisch, Wolfram, Christoph A Weber, Guido Juckeland, Nicolas Biais, and Vasily Zaburdaev. 2017. “Multiscale Modeling of Bacterial Colonies: How Pili Mediate the Dynamics of Single Cells and Cellular Aggregates.” New Journal of Physics 19 (1): 015003. https://doi.org/10.1088/1367-2630/aa5483.

Proshkin, S., A. R. Rahmouni, A. Mironov, and E. Nudler. 2010. “Cooperation Between Translating Ribosomes and RNA Polymerase in Transcription Elongation.” Science 328 (5977): 504–8. https://doi.org/10.1126/science.1184939.

Quillin, Sarah Jane, and Steven Seifert. 2018. “Neisseria Gonorrhoeae Host-Adaptation and Pathogenesis HHS Public Access.” Nat Rev Microbiol 16 (4): 226–40. https://doi.org/10.1038/nrmicro.2017.169.

Satterwhite, Catherine Lindsey, Elizabeth Torrone, Elissa Meites, Eileen F. Dunne, Reena Mahajan, M. Cheryl Bañez Ocfemia, John Su, Fujie Xu, and Hillard Weinstock. 2013. “Sexually Transmitted Infections Among US Women and Men.” Sexually Transmitted Diseases 40 (3): 187–93. https://doi.org/10.1097/OLQ.0b013e318286bb53.

Satyshur, Kenneth A., Gregory A. Worzalla, Lorraine S. Meyer, Erin K. Heiniger, Kelly G. Aukema, Ana M. Misic, and Katrina T. Forest. 2007. “Crystal Structures of the Pilus Retraction Motor PilT Suggest Large Domain Movements and Subunit Cooperation Drive Motility.” Structure 15 (3): 363–76. https://doi.org/10.1016/j.str.2007.01.018.

Sauer, Karin, Anne K Camper, Garth D Ehrlich, J William Costerton, and David G Davies. 2002. “Pseudomonas Aeruginosa Displays Multiple Phenotypes during Development as a Biofilm.” Journal of Bacteriology 184 (4): 1140–54. https://doi.org/10.1128/jb.184.4.1140-1154.2002.

157

Savage, D C. 1977. “Microbial Ecology of the Gastrointestinal Tract.” Annual Review of Microbiology 31 (1): 107–33. https://doi.org/10.1146/annurev.mi.31.100177.000543.

Schoolnik, G. K., R Fernandez, J Y Tai, J Rothbard, and E C Gotschlich. 1984. “Gonococcal Pili. Primary Structure and Receptor Binding Domain.” Journal of Experimental Medicine 159 (5): 1351–70. https://doi.org/10.1084/jem.159.5.1351.

Schötz, Eva-Maria, Marcos Lanio, Jared A. Talbot, and M. Lisa Manning. 2013. “Glassy Dynamics in Three-Dimensional Embryonic Tissues.” Journal of The Royal Society Interface 10 (89): 20130726. https://doi.org/10.1098/rsif.2013.0726.

Schwartz, Martin A, and Douglas W DeSimone. 2008. “Cell Adhesion Receptors in Mechanotransduction.” Current Opinion in Cell Biology 20 (5): 551–56. https://doi.org/10.1016/j.ceb.2008.05.005.

Seifert, H Steven. 2019. “Location, Location, Location—Commensalism, Damage and Evolution of the Pathogenic Neisseria.” Journal of Molecular Biology, April. https://doi.org/10.1016/J.JMB.2019.04.007.

Sievert, Stefan M., Michael Hügler, Craig D. Taylor, and Carl O. Wirsen. 2008. “Sulfur Oxidation at Deep-Sea Hydrothermal Vents.” In Microbial Sulfur Metabolism, 238–58. Berlin, Heidelberg: Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-540- 72682-1_19.

Smith, William P J, Yohan Davit, James M Osborne, Wook Kim, Kevin R Foster, and Joe M Pitt-Francis. 2017. “Cell Morphology Drives Spatial Patterning in Microbial Communities.” Proceedings of the National Academy of Sciences of the United States of America 114 (3): E280–86. https://doi.org/10.1073/pnas.1613007114.

Sohel, I, J L Puente, S W Ramer, D Bieber, C Y Wu, and G K Schoolnik. 1996. “Enteropathogenic Escherichia Coli: Identification of a Gene Cluster Coding for Bundle- Forming Pilus Morphogenesis.” Journal of Bacteriology 178 (9): 2613–28. https://doi.org/10.1128/jb.178.9.2613-2628.1996.

Solovyev, A Salamov V. 2011. “BPROM - Prediction of Bacterial Promoters.” Nova Sciences. 2011. http://www.softberry.com/berry.phtml?topic=bprom&group=programs&subgroup=g findb.

Steichen, Christopher T., Jian Q. Shao, Margaret R. Ketterer, and Michael A. Apicella. 2008. “Gonococcal Cervicitis: A Role for Biofilm in Pathogenesis.” The Journal of Infectious Diseases 198 (12): 1856–61. https://doi.org/10.1086/593336.

158

Steinberg, Malcolm S. 2007. “Differential Adhesion in Morphogenesis: A Modern View.” Current Opinion in Genetics & Development 17 (4): 281–86. https://doi.org/10.1016/j.gde.2007.05.002.

Stewart, Philip S. 2003. “Diffusion in Biofilms.” Journal of Bacteriology 185 (5): 1485–91. https://doi.org/10.1128/jb.185.5.1485-1491.2003.

Stoodley, P., K. Sauer, D. G. Davies, and J. W. Costerton. 2002. “Biofilms as Complex Differentiated Communities.” Annual Review of Microbiology 56 (1): 187–209. https://doi.org/10.1146/annurev.micro.56.012302.160705.

Straight, Paul D, Joanne M Willey, and Roberto Kolter. 2006. “Interactions between Streptomyces Coelicolor and Bacillus Subtilis: Role of Surfactants in Raising Aerial Structures.” Journal of Bacteriology 188 (13): 4918–25. https://doi.org/10.1128/JB.00162-06.

Süel, Gürol M., Jordi Garcia-Ojalvo, Louisa M. Liberman, and Michael B. Elowitz. 2006. “An Excitable Gene Regulatory Circuit Induces Transient Cellular Differentiation.” Nature 440 (7083): 545–50. https://doi.org/10.1038/nature04588.

Taha, M. K., B. Dupuy, W. Saurin, M. So, and C. Marchal. 1991. “Control of Pilus Expression in Neisseria Gonorrhoeae as an Original System in the Family of Two-Component Regulators.” Molecular Microbiology 5 (1): 137–48. https://doi.org/10.1111/j.1365- 2958.1991.tb01834.x.

Taha, M K, M Larribe, B Dupuy, D Giorgini, and C Marchal. 1992. “Role of PilA, an Essential Regulatory Gene of Neisseria Gonorrhoeae, in the Stress Response.” Journal of Bacteriology 174 (18): 5978–81. http://www.ncbi.nlm.nih.gov/pubmed/1522071.

Taha, Muhamed K, Christian Marchal, J Boslego, J Ciak, P Hitchcock, J Swanson, E Tramont, and J Sadoff. 1990. “Conservation of Neisseria Gonorrhoeae Pilus Expression Regulatory Genes PilA and PilB in the Genus Neisseria.” INFECTION AND IMMUNITY. Vol. 58. http://iai.asm.org/.

Taktikos, Johannes, Yen Ting Lin, Holger Stark, Nicolas Biais, and Vasily Zaburdaev. 2015a. “Pili-Induced Clustering of N. Gonorrhoeae Bacteria.” PLoS ONE 10 (9): 1–16. https://doi.org/10.1371/journal.pone.0137661.

———. 2015b. “Pili-Induced Clustering of N. Gonorrhoeae Bacteria.” Edited by Stuart Humphries. PLOS ONE 10 (9): e0137661. https://doi.org/10.1371/journal.pone.0137661.

Tan, Seth Yang-En, Su Chuen Chew, Sean Yang-Yi Tan, Michael Givskov, and Liang Yang.

159

2014. “Emerging Frontiers in Detection and Control of Bacterial Biofilms.” Current Opinion in Biotechnology 26 (April): 1–6. https://doi.org/10.1016/j.copbio.2013.08.002.

“The Integrative Human Microbiome Project.” 2019. Nature 569 (7758): 641–48. https://doi.org/10.1038/s41586-019-1238-8.

Thomason, B M. 1971. “Rapid Detection of Salmonella Microcolonies by Fluorescent Antibody.” Applied Microbiology 22 (6): 1064–69. http://www.ncbi.nlm.nih.gov/pubmed/4944805.

Thompson, D’Arcy Wentworth. 1945. On Growth and Form (1945 Edition) | Open Library. Cambridge, NY: University Press, Macmillan. https://openlibrary.org/books/OL13506098M/On_growth_and_form.

Thormann, K. M., R. M. Saville, S. Shukla, D. A. Pelletier, and A. M. Spormann. 2004. “Initial Phases of Biofilm Formation in Shewanella Oneidensis MR-1.” Journal of Bacteriology 186 (23): 8096–8104. https://doi.org/10.1128/JB.186.23.8096-8104.2004.

Trepat, Xavier, Guillaume Lenormand, and Jeffrey J. Fredberg. 2008. “Universality in Cell Mechanics.” Soft Matter 4 (9): 1750. https://doi.org/10.1039/b804866e.

Tuddenham, Susan, and Khalil G Ghanem. 2015. “Delaying the Widespread Emergence of Cephalosporin-Resistant Gonorrhoea: What Is the Best Target?” Sexually Transmitted Infections 91 (4): 232–33. https://doi.org/10.1136/sextrans-2015-052009.

Unemo, M., and W. M. Shafer. 2014. “Antimicrobial Resistance in Neisseria Gonorrhoeae in the 21st Century: Past, Evolution, and Future.” Clinical Microbiology Reviews 27 (3): 587–613. https://doi.org/10.1128/CMR.00010-14.

Ventola, C Lee. 2015. “The Antibiotic Resistance Crisis: Part 1: Causes and Threats.” P & T : A Peer-Reviewed Journal for Formulary Management 40 (4): 277–83. http://www.ncbi.nlm.nih.gov/pubmed/25859123.

Vlamakis, H., C. Aguilar, R. Losick, and R. Kolter. 2008. “Control of Cell Fate by the Formation of an Architecturally Complex Bacterial Community.” Genes & Development 22 (7): 945–53. https://doi.org/10.1101/gad.1645008.

Wang, Liang-Chun, Qian Yu, Vonetta Edwards, Brian Lin, Jessica Qiu, Jerrold R Turner, Daniel C Stein, and Wenxia Song. 2017. “Neisseria Gonorrhoeae Infects the Human Endocervix by Activating Non-Muscle Myosin II-Mediated Epithelial Exfoliation.” PLoS Pathogens 13 (4): e1006269. https://doi.org/10.1371/journal.ppat.1006269.

160

Wessel, Aimee K., Talha A. Arshad, Mignon Fitzpatrick, Jodi L. Connell, Roger T. Bonnecaze, Jason B. Shear, and Marvin Whiteley. 2014. “Oxygen Limitation within a Bacterial Aggregate.” Edited by Vanessa Sperandio. MBio 5 (2): e00992. https://doi.org/10.1128/mBio.00992-14.

Wolpert, L, and James Sharpe. 1989. “Positional Information Revisited.” Development (Cambridge, England) 107 Suppl (Supplement): 3–12. https://doi.org/10.1242/dev.114991.

Wong, T P, R K Shockley, and K H Johnston. 1980. “WSJM, a Simple Chemically Defined Medium for Growth of Neisseria Gonorrhoeae.” Journal of Clinical Microbiology 11 (4): 363–69. http://www.ncbi.nlm.nih.gov/pubmed/6768768.

Zhao, Kun, Boo Shan Tseng, Bernard Beckerman, Fan Jin, Maxsim L. Gibiansky, Joe J. Harrison, Erik Luijten, Matthew R. Parsek, and Gerard C. L. Wong. 2013. “Psl Trails Guide Exploration and Microcolony Formation in Pseudomonas Aeruginosa Biofilms.” Nature 497 (7449): 388–91. https://doi.org/10.1038/nature12155.

Zhu, Peixuan, Giovanna Morelli, and Mark Achtman. 1999. “The OpcA and PsiopcB Regions in Neisseria: Genes, Pseudogenes, Deletions, Insertion Elements and DNA Islands.” Molecular Microbiology 33 (3): 635–50. https://doi.org/10.1046/j.1365- 2958.1999.01514.x.

Zijnge, Vincent, M. Barbara M. van Leeuwen, John E. Degener, Frank Abbas, Thomas Thurnheer, Rudolf Gmür, and Hermie J. M. Harmsen. 2010. “Oral Biofilm Architecture on Natural Teeth.” Edited by Stefan Bereswill. PLoS ONE 5 (2): e9321. https://doi.org/10.1371/journal.pone.0009321.

161

Characterization of Early Biofilm Formation and Physiology in &lt;I&gt;Neisseria Gonorrhoeae&lt;/I&gt;

Characterization of Early Biofilm Formation and Physiology in <I>Neisseria Gonorrhoeae</I>